ISBN 0-9 11 239-44-8 QL 737 .156 A39 1994X MAMM ADVANCES IN THE BIOLOGY OF SHREWS edited by Joseph F, Merritt Gordon L, Kirkiand, Jr. Robert K. Rose CARNEGiE MUSEUM OF NATURAL HISTORY SPEOIAL PUBLiOATiON NO. 18 PITTSBURGH, PA 1994 SPECIAL PUBLICATION of CARNEGIE MUSEUM OF NATURAL HISTORY ADVANCES IN THE BIOLOGY OF SHREWS edited by JOSEPH F. MERRITT Powdermill Biological Station GORDON L. KIRKLAND, JR. Shippcnshurg Uni varsity ROBERT K. ROSE Old Dominion University NUMBER 18 PITTSBURGH, 1994 SPECIAL PUBLICATION OF CARNEGIE MUSEUM OF NATURAL HISTORY Number 18, pages i-x + 1-458 Issued 30 September 1994 James E. King, Director Series Editors: C. J. McCoy, Editor Mary Ann Schmidt, ELS, Assistant Editor ® 1994 by Carnegie Institute, all rights reserved ISBN 0-911239-44-8 THE CARNEGIE MUSEUM OF NATURAL HISTORY Contents •u 3 0 2CC2 „ i UBRARIES-/ Preface v Acknowledgments v Participants vii Latitudinal variation in the life histories of Sorex araneus and S. caecutiens in Finland— Kaikusalo and Johan Tost . . 1 Population biological consequences of body size in Sorex — Ilkka Hanski 15 Population dynamics of the short-tailed shrew, Blarina brevidauda— Lowell L. Getz 27 A live-trapping study of two syntopic species of Sorex, S. cinereus and S. fumeus — J. Michelle Cawthorn 39 Spatial distribution of nine species of shrews in the central Siberian taiga— I. Sheftel 45 Community organization of shrews in temperate zone forests of northwestern Russia— A. Shvarts and Dmitry V. Demin 57 Territoriality in juveniles of the common shrew {Sorex araneus) in prepeak and peak years of population density — Natalia Moraleva and Alexandra Telitzina 67 Foraging strategies of shrews, and the evidence from field studies — Sara Churchfield 77 The territorial and demographic structures of a common shrew population — Ernest V. Ivanter, Tatiana V. Ivanter, and Alexander M. Makarow 89 Parasitism by gastrointestinal helminths in the shrews Sorex araneus and S. caecutiens — Voitto Haukisalmi , Heikki Henttonen, and Taina Mikkonen 97 The masked shrew, Sorex cinereus, in a relictual habitat of the southern Appalachian Mountains — John F. Pa gels, Kristen L. Uthus, and Henry E. Duval 103 Life histories of the Soricidae; A review — Duncan G. L. Innes Ill Shrews as indicators of heavy metal pollution — Erkki Pankakoski , Ilkka Koivisto, Heikki Hyvdrinen, and Juhani Terhivuo 137 Histopathology of the common shrew Sorex araneus in Finland— T Soveri, E. Rudback, and H. Henttonen 151 Variation in brain morphology of the common shr&w— Vladimir A. Yaskin 155 Thermal biology of free-ranging shrews as revealed by computer-facilitated radiotelemetry: Energetic implications — Joseph F. Merritt and Francisco Bozinovic 163 The development of the skull of Suncus murinus — Yayoi Masuda and Takeshi Yohro 171 Characteristics of the breeding season in the common shrew {Sorex araneus): Male sexual maturation, morphology, and mobility — Paula Stockley and Jeremy B. Searle 181 Visual and hearing biology of shrews— Branis and Hynek Burda 189 Relationship of mandibular morphology to relative bite force in some Sorex from western North America — Leslie N. Carraway and B. J. Verts 201 Ultrastructure of the olfactory epithelium of the short-tailed shrew, Blarina brevicauda — Keith A. Carson, Joan L. Cole, and Robert K. Rose 211 Comparison of pigment and other dental characters of eastern Palearctic Sorex (Mammalia: Soricidae) — Erland Dannelid 217 Function of the feeding apparatus in red-toothed and white-toothed shrews (Soricidae) using electromyography and cineradiography — Christel Dotsch 233 Comparative embryonic development of the Soricidae — Kerry R. Foresman 241 Brown fat and the wintering of shrews — Heikki Hyvdrinen 259 iii Effects of melatonin on the chronobiology of the least shrew, Cryptotis parva — Orin B. Mock 267 Captive breeding of the common shrew (Sorex araneus) for chromosomal analysis — S. J. Mercer and J. B. Searle .... 271 Proposed standard protocol for sampling small mammal communities — Gordon L. Kirkland, Jr. and Patricia Krim Sheppard 277 The trapline concept applied to pitfall arrays — Charles O. Handley, Jr. and Merrill Yarn 285 Albumin evolution in the Soricinae and its implications for the phylogenetic history of the Soricidae — Sarah B. George and Vincent M. Sarich 289 The Sorex of the araneus-arcticus group (Mammalia: Soricidae): Do they actually speciate? — Jacques Hausser 295 The Plio-Pleistocene patterns of distribution of the Soricidae in Poland— Rzebik-Kowalska 307 Comparative cytogenetics and systematics of .Sorer: A cladistic approach— £. Yu. Ivanitskaya 313 The evolution of the Soricidae as shown by the variability of cranial morphology — V. A. Dolgov 325 Chromosomal evolution in the genus Crocidura (Insectivora: Soricidae)— T. Maddalena and M. Ruedi 335 Phylogeny and distribution of the Crocidosoricinae (Mammalia: Soricidae) — Jelle W. F. Reumer 345 A preliminary analysis of biogeography and phylogeny of Crocidura from the Philippines — Lawrence R. Heaney and Manuel Ruedi 357 Evolution and phylogenetic affinities of the African species of Crocidura, Suncus, and Sylvisorex (Insectivora: Soricidae)— Lnwra J. McLellan 379 Identification of the Carolinean shrews of Bachman Charles O. Handley, Jr. and Merrill Yarn 393 Shrews of ancient Egypt: Biogeographical interpretation of a new species — Rainer Hutterer 407 Progesterone (P4) and Estradiol (E,) secretion by Suncus murinus ovaries and adrenals in ViUo—Stasia Stoklosowa, Janice Bahr, and Gil Dryden 415 Metabolic rates and regulation of cardiac and respiratory function in European shrews — Alfred Nagel 421 The white-toothed shrew Crocidura russula nionacha: How does its physiology fit mammalian allometry? — Haya Mover, Amos Ar, and Salo Hellwing 435 The structure and adaptive peculiarities of pelage in soricine shrews — Ernest V. Ivanter 441 Soricid biology: A summary and look ahead — Robert K. Rose 455 IV Preface Acknowledgments In 1988 we decided to organize an international colloquium on the biology of the Soricidae at Powdermill Biological Station of the Carnegie Museum of Natural History, the site of a number of previous scientific meetings. It was clear from past conferences at Powdermill that the number of participants had to remain small to enhance communication and productivity. Participants were selected to represent the nearly worldwide distribution of soricids. The International Colloquium on the Biology of the Soricidae, held 8-14 October 1990, welcomed 55 biologists representing 14 countries— Canada, Czechoslovakia, Finland, France, Germany, Israel, Japan, The Netherlands, Poland, Russia, South Africa, Sweden, the United Kingdom, and the United States. The colloquium encompassed many fields of soricid biology, including ecology, anatomy, physiology, behavior, biogeography, and evolution and systematics. A session dedicated to field and laboratory methods was also included in the program. Topics presented were diverse, ranging from physiological ecology and population dynamics of Eurasian shrews to comparative allozyme and albumin evolution in the Soricidae. The colloquium was declared a great success by all participants, largely due to the high quality of presentations and the relaxed atmosphere of the Powdermill setting which fostered considerable informal discussion between sessions and in the evenings. The intellectual stimulation of the gathering was further heightened by the crisp autumn days and colorful foliage of the Appalachian Mountains of western Pennsylvania. Our goal as conveners and editors was to produce a volume of proceedings indicative of the current research and state of knowledge in soricid biology, and to stimulate additional research on this group of mammals. We believe these objectives were met at the colloquium and are formally summarized in the contents of this volume. In order to conserve costs of publication, figures and tables appear at the end of each article. The International Colloquium on the Biology of the Soricidae was made possible by funds from the Carnegie Museum of Natural History, Pittsburgh, Pennsylvania. We are indebted to James E. King, Director, for his support of the colloquium and publication of this volume. This colloquium and its success also depended on the support of many people. We are deeply indebted to our friends, Ingrid and Bill Rea and Theresa and Tom Nimick for hosting receptions, and to M. Graham Netting and Ingrid and Bill for providing lodging in their homes for several of the participants, who will never forget their gracious and kind hospitality. The colloquium ran smoothly due to the expert preparation of the facilities by the Powdermill maintenance staff — Gilbert Lenhart, Albert Lenhart, and Lloyd Moore. Participants were housed in cabins at Powdermill and in cottages provided by Donald Ankney and the Laurel Mountain Camp. Robert Leberman and Robert Mulvihill kindly gave ad libitum bird-banding demonstrations for the participants. Terri Kromel and Kathy Matt worked long hours in a variety of roles ranging from tour guides and travel agents to medical technicians. Theresa Gay Rohall, the projectionist, kept all speakers on schedule. Dyana Kessel of Ligonier Travel and Tours assisted participants with domestic and international travel arrangements. Individual participants increased their body mass by upwards of 3 kg during the week due to the copious amounts of delicious food provided by Pat Piper and the staff of Ligonier Country Catering and by our hosts for the evening receptions, Ingrid and Bill Rea and Theresa and Tom Nimick. Transportation between the colloquium site and Pittsburgh and Dulles International Airports was accomplished through help from Gordon L. Kirkland, Jr., Kathy Matt, Joseph Merritt, Santiago Reig, Bob Rose, Duane Schlitter, and Jeff Wilcox. We are indebted to the creative talents of Nancy Perkins of the Section of Exhibit Design and Production, Carnegie Museum of Natural History, for designing the colloquium logo. We thank the participants for their cooperation in submitting manuscripts and revisions in a timely fashion. We thank the many reviewers of individual manuscripts, whose suggestions for revision contributed significantly to the quality of papers in this volume. We are most grateful to Robert S. Hoffmann and Jane Junge, who although unable to attend the colloquium, did yeoman service through their meticulous editing of several manuscripts. We offer special thanks to Colleen Hannakan and Julia Zeyzus for reworking some of the illustrations. Lastly, special thanks go to the late C. J. McCoy, Editor, and Mary Ann Schmidt, Assistant Editor of scientific publications for Carnegie Museum of Natural History, for preparing this volume for publication. JFM, GLK, RKR V .-:n ■‘ ‘" : 'liyT :- ■'i^ “ f" ■ 1,1 If lii I ■::Y,;^ ■ ® ,, ^ ' ' ' ■:» ^ ’ ■> • ■ - ■. O'f#?;!';' i -:S4E^’ •. .,V3 -H 5, I. ffeJw-L'Mi • . irtutfull in ; ,.. ,":a Uv.i .a-?i ;»'I, , kynrlitiy^ ., :■ •% '..../: «W?*: V, ; \ #.{ U|(i«?V ..1^-1 -VI r ■ '^ •' -•ivU;!.-, . ••.«• r '},•> ... •I *v T|p.^0iJ|>.4»l>'i’#l »H)t 'i I '•n,. '. ’‘V .. .‘^s?*-- 4 f 'IC / • J i ., . !(' '. ■'! • •' ,,«•' r- •: ■ -' i'''j •'. "■ • '" I'.-l' ' ’ - -_■ '. ' ! • , .^ • . ' I •' ; l' J . ■ V ■ ■'' . , 14.- 1 s. ,: f ■■ ' > f-’l( ’ .'(I' ll'i ! , I'i ,'H^i ■-.»;);■ ^\r, |?l^l •*j}- nr' j ,^'j. fjf< JBSqtA ■f • >■•'’■( ,y,<}|i(jiirf]’ ■, .,;;v' I ’<(■!( (,/..-lti| bn.-i i>f!>S„.0'i lNSiler.>t{>sfj rtoi^^irw. ..M-' '■■ ■ ",•*■>.'■■''. bah il;,!/! •;? iifVl' ^>‘ivU''r!i -. ff,M"ii.‘-r}fi. L «■.•■., ;• ■ .iil* n. Wi;v. 'iMBtajSWlJ.to ir. '-I' •>•,■. .■■..(: S ‘oY-4t!<,-Vi VM' '.-^u i''‘> .‘;‘.:i^*vi'''tr;:,;tff ,] ;,iib \i^rix\ t/ .', ■-•i.tf'j-!.-; '\(> /.ic.r ..' ' •'1';'^' A<;;?vii-‘'i^-it4 fftCiCUiW ^;H£1i.lfcii'£Sj''r^ •' ' -faiHo' T»t fWWiiniijtJ- WW-Mff 'j^' .livn», >-:\:;h /rif'fiji’. ,0 l«r.M =tl.-i •■., -= V a • w/<^ ?// h> •■• u^‘'p'‘ p,?‘ijp''- ai:ii,n)c't(J ■'yv-{l'ii'> ••'■'' .••uniif y riY^>3^r Ir ('rH,i=/< ~,/i b,U' llliiilipafi VM }!s! U. i-H!^ ,v0-y.V ' ■; Ifo •■•, I n-tJlr\3a; ■ ;,■ «.i - .i natiJ^jii'n. •] ^'/ iVai'.’-} *>' I .r, vMr.h ■• '■ ■v.; p/lp + (l-p)g], where T is the starv’ation time in hours, p is the probability that food availability w'ili return to a high level in the next hour, and q is the probability that a foraging shrew' will die in the next hour when food availability is low. The left side of this inequality gives the probability of surviving the period of food shortage if the shrew rests, and the right side gives the corresponding probability if the shrew continues to forage. This model makes two testable predictions: 1) because T increases with body size (Table 1), larger species are more likely than small' ones to decrease activity during periods of low food availability; and 2) the shorter the average length of food shortage (larger p), the more likely the response will be decreased activity. In an experiment comparing small and large species under the same conditions, it was found that, as predicted, small shrews increased activity while large species decreased activity in response to short-term food shortages (Fig. 2). The average duration of food shortages had also the predicted effect on activity of S. araneus. Foraging bouts were longer when food availability was constantly rather than temporarily low (Hanski, 1992; J. Saarikko and I. Hanski, personal observation). One option not considered in the above model, but which is available to shrews in nature, is food caching. Both short-term (Crowcroft, 1957; Goulden and Meester, 1978) and long-term (seasonal) caching (Platt, 1976; Martin, 1984) have been observed in shrews. Short-term caching is especially common in small species of shrews (Hanski, 1989a). Food caching may replace the function of body energy reserves in small species, and bridge short-term gaps in food availability. Food caching may also decrease the frequency of interaction of small species with larger and competitively superior species. Comparative studies of food caching in small and large species are nmied. Population Dynamics and Body Size Comparati ve studies of colonization and extinction dynamics of three species {S. araneus, S. caecutiens and S. minutus) have been conducted on small islands in lakes in eastern Finland since 1982 (Hanski, 1986; Hanski and Kuitunen, 1986; Hanski and Peltonen, 1988; Peltonen and Hanski, 1991). The islands vary in size from less than 1 ha to hundr^ls of hectares. Populations of shrews occupying smaller islands have substantial risk of extinction (Peltonen and Hanski, 1991). However, extinctions are compensated for by the establishment of new populations, and species occur on the islands in a dynamic equilibrium between extinction and colonization (Hanski, 1986). These field studies provide quantitative data on dispersal, colonization, and extinction rates in the three species (Table 2). Dispersal rate is measured as the fraction of individuals in the mainland population which disperse to islands. Colonization and extinction rates refer to the fractions of previously empty and occupied islands which became occupied and empty, 1994 HANSKI— Consequences of Body Size in Sorex 17 respectively, during one year. TTie term colonization ability refers to the capacity of an individual or a group of individuals to establish a new local population on an empty island, conditional on arrival at the island. Dispersal Larger species show a somewhat higher rate of dispersal to islands (Hanski and Peltonen, 1988) than smaller species, as expected from their longer starvation times (Table 1) and faster swimming rates (Skaren, 1980; Hanski, 1986), but the difference is not significant (Table 2). Unfortunately, results on dispersal rates confound the dispersal ability of a species with the inclination to start overwater dispersal in the first place. Michielsen (1966) found in a comparative study that S. mi nut us had greater dispersal tendency than S. araneus, which may have compensated (in our study) for its lower overwater dispersal ability. Studies have revealed that S. araneus generally avoids entering water (Hanski and Peltonen, 1988). In many, but not all years, dispersers are smaller and probably socially subordinate individuals, which are apparently forced out of the mainland population by stronger conspecifics (Hanski et a!., 1991). Colonization There were no interspecific differences in colonization rates, which are determined by dispersal rate and colonization ability (Table 2). Sorex minutus may have a somewhat lower dispersal rate than the two larger species, but unexpectedly the results indicate that it has a significantly higher colonization ability than the other species (Table 2). Why should small shrews make good colonizers? These results should be confirmed with studies on other populations and other species. Three observations could be tested or could lead to testable predictions. First, during the five years of our study, we observed only one pregnant shrew dispersing to an island. This shrew was S. minutus (Hanski, 1986). Pregnant females are potentially good colonizers, as there is then no need to find a mate before founding a colony. Second, the low per capita food requirements of S. minutus (Table 1) and other small species may improve their colonization ability by enhancing survival of newly-dispersed individuals on islands (the same applies to dispersers crossing unfamiliar terrain on land). This could be tested by determining survival rates of shrews experimentally introduced on islands. Third, Michielsen (1966) found in a population ecology study of S. araneus and S. minutus that the latter species, in spite of smaller size, had a substantially larger home range. This was unexpected, because generally home range size is positively correlated with body size (McNab, 1963). Michielsen (1966) suggested that differences in the diets of the two species might explain these anomalous results. An alternative explanation is that smaller species are generally less territorial and have less well-defined home ranges than large species. Territoriality may be less profitable for small than large species because the food resources of the former are scattered and hence less defendable (Davies and Houston, 1984). Hawes (1977) also reported extensive movements in two small Sorex species in British Columbia. Weak and poorly defined home ranges should increase the probability of individuals wandering to new areas. Extinction Extinction rate in the three species increases with decreasing body size, and the interspecific differences are highly significant (Table 2). Such a relationship between body size and extinction rate was expected because of the shorter starvation times of the smaller species (Table 1), which make them more susceptible than large species to temporal variation in food availability. Furthermore, an analysis of species’ incidence functions, which describe the pattern of occurrence on islands of different size, strongly suggested that extinction rates decrease much faster with increasing island area (and hence with increasing population size) in large species than in small species (Fig. 3; Hanski, 1991). This result also indicates a greater role of environmental stochasticity in small species than in large species (Hanski, 1991). Michielsen’s (1966) observation that the average rate of mortality is higher in S. minutus than in S. araneus is consistent with this conclusion. To summarize, the smallest species (S. minutus) had the highest rate of extinction but also the best colonization ability and probably a high dispersal tendency. The extinction proneness of the small species is probably due to short starvation time and hence great sensitivity to temporal changes in food availability, whereas its good colonization ability may be due to small per capita food requirements and possibly to dispersal of pregnant females. Colonization ability is associated with high extinction rates of local populations in this species because high extinction rates select for increased colonization rates (Brown, 1951; Southwood, 1962). Only highly dispersive species may survive regionally if local populations frequently go extinct (Hanski, 1992). Community Structure Size Distributions in Coexisting Species Assemblages of coexisting shrews show good separation in body size. Principal component analyses of measurements of the skull and postcranial skeleton in a five-species assemblage from Finland and a four-species assemblage from Alberta, Canada, are given in Fig. 4. In both assemblages, there is little overlap in size (PC I in Fig. 4), and adjacent species in a size ranking show almost equal size ratios, or “community -wide character displacement” (Strong et al., 1979). Such body-size differences reduce interspecific competition. Experiments have demonstrated interspecific competition between species with similar body sizes (Hawes, 1977; Neet and Hausser, 1990), but not between species with different body sizes (Ellenbroek, personal communication). The distribution of body sizes in coexisting shrews is comparable to patterns found in granivorous, desert-dwelling rodents (Bowers and Brown, 1982; Hopf and Brown, 1986; Brown, 1987) and mustelids (Dayan et al., 1989). Nonetheless, exactly how body-size differences facilitate coexistence in shrews is not entirely understood (Hanski and Kaikusalo, 1989; Hanski, 1992). In the skull measurements, there are pairs of 18 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 morphologically similar species, such as isodon-arcticus, caecutiens-monticolus , and minutus-dnereus (Fig. 4). The North American species deviate in skull shape from the Eurasian species, as indicated by the position of the species along PC II (Fig. 4). Such differences may reflect phylogenetic divergence, as most of the species on the two continents belong to different subgenera (George, 1988). The even spacing of sp^ies differing in size, which is evident in the assemblages of extant shrews (Fig. 4), may also have been a common feature of shrew assemblages in the past. A well-studied late Pliocene assemblage from Poland has five species, Sorex minutus, S. bor (an extinct species), and three other extinct Sorex species (Rzebik-Kowalska, 1994). Sorex bor was a small species, intermediate in size between S. minutus and S. caecutiens. The three remaining species included a species comparable in size to S. araneus, and tw'o larger species, roughly the size of S. mirabilis (B. Rzebik-Kowalska, personal communication). Habitat Selection Large species of shrews have greater per capita food requirements than small species (Table 1). Assuming that the larger species are not more efficient in foraging than the smaller species (Fig. 1), there must be habitats in which food availability is so low that large shrews cannot survive and reproduce but small species can. Therefore, the numerical dominance of large shrews should increase with increasing habitat productivity. Habitat selection of small, medium, and large species of shrews in an extensive set of data from Eurasia is summarized in Fig. 5. These results support the idea that small species dominate in the least productive habitats, whereas large species dominate in the most productive habitats. The absolute density of small species varied little among habitat types (B. Sheftel and I. Hanski, personal observation), which has two possible explanations. Either the more productive habitats are not intrinsically better for small species than the less productive habitats, or small species are competitively excluded or suppressed in numbers by larger species in the more productive habitat types. This question cannot be settled without experiments. However, previous studies have demonstrated that interspecific competition is common in shrews (Michielsen, 1966; Hawes, 1977; Malmquist, 1985; Neet and Hausser, 1990). Many large species of shrews prefer to feed on earthworms if available (Rudge, 1968; Okhotina, 1974; Pemeita, 1976). Earthworms often have a much higher biomass than all other prey types combined (Terhivuo, 1988), but they are typically either very scarce or entirely absent from more barren habitat types, making the existence of large shrew species in these habitats more difficult. Biogeography and Body Size The assemblages of Sorex inhabiting coniferous forests in north-temperate Eurasia and North America show one major difference; the dominant species in Eurasia is large (S. araneus), whereas the dominant species in North America is small (S. cinereus. Fig. 4). The presence of Blarina, a very large shrew, in North America is not a likely explanation because Blarina occurs only in the southern part of the coniferous forest zone and is very different ecologically from Sorex (van Zyll de Jong, 1983, and references therein). I have suggested that this difference may be a consequence of the generally greater productivity of coniferous forests in most of Eurasia compared to those of North America (Hanski, 19891?). Such a difference in productivity would favor larger species in Eurasia and smaller species in North America, in the same way as differences in productivity of habitat types favor small or large sp^ies (Fig. 5). Less productive forest types probably tend to have lower availability of large prey items than more productive forest types, which would also be a disadvantage to large shrews (Fig. 1). One particular consideration is the scarcity of earthworms in northern coniferous forests in most of North America, due to their almost complete disappearance during Pleistocene glaciations. The above hypothesis is supported by the exceptions. In eastern Siberia, where the landscape is barren and dominated by larch forests away from river valleys, the dominant shrew is a middle-sized species (5. caecutiens), w'hicli is not very different in size from S. cinereus in North America (Fig. 4). On the other hand, in coniferous forest regions of eastern North America, including Nova Scotia and New Brunswick, where rainfall is high and forests presumably productive, the dominant species is S. fumeus, which is only slightly smaller than S. araneus (Kirkland and Schmidt, 1982). Latitudinal changes in the size of shrews do not agree with Bergmann’s rule. On the contrary, the size of species and the average size of shrews in local assemblages both tend to decrease with increasing latitude (Hanski, 19891?, has analyzed geographical changes in the size of Sorex caecutiens as an example). I suggest that the reason for the generally smaller size of shrews in the north is decreasing food availability with increasing latitude. Latitudinal changes in climate may be relatively unimportant for shrews because they spend the winter under snow cover, where temperatures are relatively constant regardless of surface ambient temperature. Character Displacement? Sorex araneus, the dominant sp«:ies in Europe and western Siberia, is absent from eastern Siberia. In eastern Siberia, the dominant species is S. caecutiens, which is larger there than in Europe. However, another transcontinental species, S. isodon, is smaller in eastern Siberia than in Europe (Fig. 6). These changes in body size may be due to interspecific interactions, S. isodon and S. caecutiens possibly shifting toward the size of S. araneus in eastern Siberia in the absence of that species. That S. isodon and S. caecutiens change in size in opposite directions supports this interpretation. However, in view of the difficulties of demonstrating character displacement (e.g., Grant, 1975), the above example is cited not as conclusive proof but as a demonstration that shrew's provide suitable material to study character displacement. Discussion I have attempted to show that comparisons between small 1994 HANSKI— Consequences of Body Size in Sorex 19 and large species of shrews can be rewarding in the study of behavioral ecology and population dynamics. In explaining behavioral and ecological differences among congeneric species, expectations based on body size differences comprise a useful “null hypothesis. ” One field not covered here is the life histories of shrews, for which there are relatively few data. Available data suggest relatively little variation in life histories among species varying in body size. All Sorex species are basically “biennials,” maturing in their second summer and almost never surviving a second winter. All shrews may have several litters per year, the number being limited by the length of the season. Variation in litter size appears not to be related to body size (e.g., Sheftel, 1989). Sheftel (1989) suggested, from extensive study of eight coexisting Sorex species in central Siberia, the hypothesis that competitively inferior species have a greater intrinsic rate of increase than competitively superior species, which facilitates the coexistence of these species. The main evidence was that species that attain high abundance early each season also peak earlier during the general four-year cycle of small mammals in central Siberia (Sheftel, 1989). This hypothesis is worth further study, although surprisingly the competitive ranking of species in Sheftel’s hypothesis was not correlated with body size. The main theme of this paper, which I advance as a working hypothesis, is that small shrews live a more precarious life than large ones. Table 3 summarizes the key elements of this hypothesis. At the individual level, a crucial difference is the shorter starvation times of small shrews, which is reflected in the individuals’ behavioral responses to spatially and temporally varying food abundance. At the population level, smaller species are expected to show more temporal variation in population size and a higher rate of extinction of local populations. Hanski (1989/j) demonstrated that temporal variation in population size was indeed greater in small (S. mi nut us and S. caecutiens) than in large {S. araneus) species. The results reviewed in this paper conclusively demonstrate the greater extinction-proneness of local populations of small species. Finally, at the metapopulation level (Gilpin and Hanski, 1991), the hypothesis predicts higher turnover in populations of small species, with populations of small species going extinct relatively frequently but new ones being established at a correspondingly high rate. There is direct evidence for the good colonization ability of Sorex minutus, a small species. In agreement with these results, it has been found that S. minutus, which is typically less abundant than S. araneus in most of Europe, is nonetheless more frequent on large islands in the Baltic and elsewhere in northwest Europe (Williamson, 1981; Malmquist, 1985; Peltonen et al., 1989). The hypothesis of high turnover rate in small species of Sorex provides a fresh perspective to the population ecology puzzle of Sorex minutissimus, the smallest species of Sorex and one of the smallest extant mammals. Sorex minutissimus is widely distributed in the Palearctic region but is apparently very rare throughout its entire range. The rarity of S. minutissimus is unlikely to be a trapping artifact, because pitfall traps suitable for small Sorex are regularly used to catch shrews in Europe and Siberia. Applying the present hypothesis to S. minutissimus. I suggest that it occurs as small and very ephemeral local populations, continuously shifting from one place to another. If this is correct, S. minutissimus must have exceptionally high dispersal and colonization rates. Perhaps the same traits which make S. minutus a good colonizer are even more strongly expressed in and enhance the colonization rate of S. minutissimus. However, practically nothing is known about the population biology of this species. The first extant species of Sorex to appear in the fossil record are S. minutus and S. minutissimus (B. Rzebik- Kowalska, personal communication), which are the two smallest species in Europe. The great evolutionary age of the smallest Sorex is in striking contrast with the hypothesized precarious life of individuals and populations. The solution to this apparent paradox may lie in the same factor that allows metapopulations to survive in spite of frequent local extinctions: high rates of dispersal and colonization. High dispersal rate has two consequences that tend to preserve the status quo of the species. First, other things being equal, the geographical range of the species increases with increasing dispersal rate. Sorex minutus and especially S. minutissimus have large geographical ranges. The probability of extinction of a species is expected to decrease with increasing size of its geographical range for two reasons: the total number of individuals, and the diversity of environmental conditions under which these individuals live, increase with increasing geographical range. Second, a species with high dispersal rate is probably less likely to evolve than a species with restricted dispersal, because high dispersal rate increases gene flow between populations. In summary, high dispersal rate increases gene flow and thereby decreases speciation rate, and it increases the size of the geographical range of the species, which decreases the risk of extinction. I suggest that these factors help explain the relatively great age of the smallest species of Sorex. Acknowledgments I thank C. R. Chandler, R. K. Rose, and an anonymous referee for useful comments on the manuscript. R. K. Rose forced me to try my best by his scrupulous editing of the manuscript. Literature Cited Bowers, M. A, and J. H. Brown. 1982. Body size and coexistence in desert rodents: Chance or community structure. Ecology, 63:391-400. Brown, E. S. 1951 . The relation between migration rate and type of habitat in aquatic insects, with special reference to certain species of Corixidae. Proceedings of the Zoological Society of London, 121:539-545. Brown, J. H. 1987. Variation in desert rodent guilds: Patterns, processes, and scales. Pp. 185-203, in Organization of Communities: Past and Present (J. H. R. Gee and P. S. Giller, eds.), Blackwell, Oxford, 576 pp. Buckner, C. H. 1964. Metabolism, food consumption, and feeding behaviour in four species of shrews. Canadian Journal of Zoology, 42:259-279. Crowcroft, P. 1957. The Life of the Shrew. Reinhardt, London, 166 pp. 20 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 Davies, N. B., and A. I. Houston. 1984. Territory economics. Pp. 148-169, in Behavioural Ecology: An Evolutionary Approach. (J. R. Krebs and N. B. Davies, eds.), Blackwell Scientific Publications, Oxford, 493 pp. Dayan, T., D. Simberloff, E. Tchernov, and Y. Yom-Tov. 1989. Inter- and intraspecific character displacement in mustelids. Ecology, 70:1526-1539. Diamond, J. M. 1975. Assembly of species communities. Pp. 342-344, in Ecology and Evolution of Communities (M. L. Cody and J. M. Diamond, eds.). Harvard University Press, Cambridge, Massachusetts, 545 pp. Diamond, J. M., D. Phan, and F. L. Carpenter. 1986. Hummingbird digestive physiology, a determinant of foraging bout frequency. Nature, 320:62-63. Eisenberg, j. F. 1981. The Mammalian Radiations. An Analysis of Trends in Evolution, Adaptation, and Behavior. The University of Chicago Press, Chicago, 610 pp. George, S. B. 1988. Systematics, historical biogeography, and evolution of the genus Sorex. Journal of Mammalogy, 69:443-461 . Gilpin, M. E., and I. Hanski. 1991. Metapopulation Dynamics. Academic Press, London, 336 pp. Goulden, E. A, AND J. MEESTER. 1978. Notes on the behaviour of Crocidura and Myosorex (Mammalia: Soricidae) in captivity. Mammalia, 42:197-207. Grant, P. R. 1975. The classical case of character displacement. Evolutionary Biology, 8:237-337. Hanski, I. 1984. Food consumption, assimilation and metabolic rate in six species of shrew (Sorex and Neomys). Annales Zoology Fennici, 21:157-165. _. 1985. What does a shrew do in an energy crisis? Pp. 247-252, in Behavioural Ecology (R. M. Sibly and R. H. Smith, eds), Blackwell, Oxford, 620 pp. 1986. Population dynamics of shrews on small islands accord with the equilibrium model. Biological Journal of the Linnean Society, 28:23-36. 1989a. Habitat selection in a patchy environment: Individual differences in common shrews. Animal Behaviour, 38:414-422. 1989Z?. Population biology of Eurasian shrews: Towards a synthesis. Annales Zoologici Fennici, 26:469-479. 1991. Single-species metapopulation dynamics: Concepts, models, and observations. Pp. 17-38, in Metapopulation Dynamics (M. E. Gilpin and 1. Hanski, eds.). Academic Press, London, 336 pp. 1992. Insectivorous mammals. Pp. 163-187, in Natural Enemies: The Population Biology of Predators, Parasites and Diseases. (M. J. Crawley, ed.), Blackwell, Oxford, 576 pp. Hanski, I., and A. KaKUSALO. 1989. Distribution and habitat selection of shrews in Finland. Annales Zoologici Fennici, 26:331-479. Hanski, I., and J. Kuitunen. 1986. Shrews on small islands: Epigenetic variation elucidates population stability. Holarctic Ecology, 9:193-204. Hanski, I., and A. Peltonen. 1988. Island colonization and peninsulas. Oikos, 51:105-106. Hanski, I., A. Peltonen, and L. Kaski. 1991. Natal dispersal and social dominance in the common shrew Sorex araneiis. Oikos, 62:48-58. Harvey, P. H., and M. D. Pagel. 1989. Comparative studies in evolutionary ecology: Using the data base. Pp. 209-230, in Toward a More Exact Ecology (P. J. Grubb and J. B. Whittaker, eds.), Blackwell, Oxford, 468 pp. Harvey, P. H., and A. F. Read. 1988. How and why do mammalian life histories vary? Pp. 213-232, in Evolution of Life Histories of Mammals (M. S. Boyce, ed.), Yale University Press, New Haven, Connecticut, 373 pp. Hawes, M. L. 1977. Home range, territoriality and ecological separation in sympatric shrews Sorex vagrans and Sorex obscurus. Journal of Mammalogy, 58:354-367. Hope, F. A., and J. H. Brown. 1986. The bull’s-eye method for testing randomness in ecological communities. Ecology, 67:1139-1155. IVANTER, E. V. 1976. On specific diagnostics and intraspecific taxonomy of the shrews of Karelia. Pp. 59-68, in Ecology of Birds and Mammals in the North-West of USSR (E. V. Ivanter, ed.), Karelian Branch of Academy of Sciences of USSR, Petrozavodsk, 149 pp. Karasov, W. H., D. Phan, J. Diamond, and F. L. Carpenter. 1986. Food passage and intestinal nutrient absorption in hummingbirds. Auk, 103:453-464. Kirkland, G. L., Jr., and D. F. Schmidt. 1982. Abundance, habitat, reproduction and morphology of forest dwelling small mammals of Nova Scotia and southeastern New Brunswick, Canada. The Canadian Field-Naturalist, 96:156-162. MacArthur, R. H. 1972. Geographical Ecology. Harper and Row, New York, 269 pp. Malmquist, M. G. 1985. Character displaeement and biogeography of the pygmy shrew in northern Europe. Ecology, 66:372-377. Martin, I. G. 1984. Factors affecting food hoarding in the short- tailed shrew Blarina brevicaiida. Mammalia, 48:65-71. McNab, B. K. 1963. Bioenergetics and the determination of home range size. The American Naturalist, 97:133-140. Michielsen, N. C. 1966. Intraspecific and interspecific competition in the shrews Sorex araneiis L. and Sorex minutus L. Archives Neerlandaises de Zoologie, 17:73-174. Neet, C. R, and j. Hausser. 1990. Habitat selection in zones of parapatric contact between the common shrew Sorex araneus and Millet’s shrew S. coronal us. Journal of Animal Ecology, 59:235-250. Okhotina, M. V. 1974. Morpho-ecological features making possible joint habitation for shrews of the genus Sorex (Sorex, Insectivora). Pp. 42-57, in Fauna and Ecology of the Terrestrial Vertebrates of the Southern Part of the Soviet Far East (M. V. Okhotina, ed.). New Series Volume 17 (120), Academy of Sciences of the USSR, Vladivostok, 205 pp. Peltonen, A., and I. Hanski. 1991. Patterns of island occupancy explained by colonization and extinction rates in three species of shrew. Ecology, 72:1698-1708. Peltonen, A., S. Peltonen, P. Vilpas, and A. Beloff. 1989. Distributional ecology of shrews in three archipelagoes in Finland. Annales Zoologici Fennici, 26:381-388. Pernetta, j. C. 1976. Diets of the shrews Sorex araneus L. and Sorex minutus L. in Wytham grasslands. Journal of Animal Ecology, 45:899-912. Peters, R. H. 1983. The Ecologieal Implications of Body Size. Cambridge University Press, Cambridge, 329 pp. Platt, W. J. 1976. The social organization and territoriality of short- tailed shrew Blarina brevicauda populations in old field habitats. Animal Behaviour, 24:305-318. Rudge, M. R. 1968. The food of the common shrew Sorex araneus (Insectivora, Soricidae) in Britain. Journal of Animal Ecology, 37:565-581. Rzebik-KowalSKA, B. 1994. The Plio-Pleistocene patterns of distribution of the Soricidae in Poland. Pp. 307-312, in Advances in the Biology of Shrews (J. F. Merritt, G. L. Kirkland, Jr., and R. K. Rose, eds.), Carnegie Museum of Natural History Special Publication 18, x -f 458 pp. 1994 HANSKI— Consequences of Body Size in Sorex 21 SaarIKKO, J., and I. Hanski. 1990. Timing of rest and sleep in foraging shrews. Animal Behaviour, 40:861-869. Schmidt-Nielsen, K. 1984. Sealing. Why is animal size so important? Cambridge University Press, Cambridge, 241 pp. ShefteL, B. I. 1989. Long-term and seasonal dynamics of shrews in central Siberia. Annales Zoologici Fennici, 26:357-370. Skaren, U. 1980. Vesi Sorex-paastaisten leviamisesteena. Savon Luonto, 12:44-47. SOUTHWOOD, T. R. E. 1962. Migration of terrestrial arthropods in relation to habitat. Biologieal Review, 37:171-214. Strong, D. R., L. A. Szyska, and D. Simberloff. 1979. Tests of community-wide charaeter displaeement against null hypotheses. Evolution, 33:897-913. Terhivuo, J. 1988. The Finnish Lumbricidae (Oligochaeta) fauna and its formation. Annales Zoologici Fennici, 25:229-247. VAN Zyll DE Jong, C. G. 1983. Handbook of Canadian Mammals. 1. National Museums of Canada, Ottawa, 216 pp. Vogel, P. 1980. Metabolic levels and biological strategies of shrews. Pp. 170-180, in Comparative Physiology: Primitive Mammals (K. Schmidt-Nielsen, L. Bolis, and C. R. Taylor, eds), Cambridge University Press, Cambridge, 416 pp. Williamson, M. 1981. Island Populations. Oxford University Press, 286 pp. Table 1. — Body mass, metabolic rate, food consumption, mass-specific food requirement , and starvation time in five species of European Sorex. Food consumption was measured in carbon. Measurements were made at 23 °C. Starvation time is defined as the time during which a starving shrew is expected to lose 20% of its initial carbon content. The values given here are mean values (for sample sizes and other details, see Hanski, 1984). Species Mass g Metabolic Rate J/h Food Requirement for 24 h Period mg C % body mass Starvation Time in Hours S. minutissimus 2.5 778 410 94 5.1 S. minutus 2.7 813 420 89 5.4 S. caecutiens 4.9 1148 600 70 6.9 S. araneus 8.9 1626 850 55 8.8 S. isodon 11.1 1845 960 49 9.7 Table 2.— Comparison between Sorex araneus, S. caecutiens, and S. minutus in dispersal rate, colonization ability, colonization rate, and extinction rate (from Peltonen and Hanski, 1991 , and 1. Hanski, personal observation). The figures are numbers of individuals (dispersal rate) or colonization/extinction events (the other three variables). The expected numbers are given in brackets. The test result is from test or from Monte Carlo (MC) randomization test. P gives the significance level (NS stands for nonsignificant result at 5% level). Variable araneus caecutiens minutus Test P Dispersal rate® 41 (36.4) 13 (12.1) 4 (9.5) 3.83 NS Colonization ability’’ 5 (7.4) 2 (3.5) 5 (1.1) MC 0.004 Colonization rate® 5 (3.0) 2 (4.5) 5 (4.5) 2.78 NS Extinction rate‘’ 1 (5.6) 3 (1.8) 6 (2.6) MC 0.004 ^Expected values were calculated by multiplying the observed number of dispersers by the proportions of the species on the mainland (pooled data for five mainland study sites located around the lake). *The success of individuals that have reached an empty island in establishing a new population (the expected figures were calculated from the products of the numbers of dispersers times the numbers of empty islands). ‘^The rate of establishment of new populations (the expected figures were calculated from the numbers of empty islands). ‘’The rate of disappearance of existing populations (the expected figures were calculated from the numbers of island populations). Data were available for five years, and the colonization and extinction events were scored from one year to another. Data on dispersal rate were obtained by trapping shrews on small islets without local populations. Dispersal may also occur in winter over ice, but we have no data on winter dispersal. For further explanation see Peltonen and Hanski (1991). 22 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 Table 3. — The working hypothesis is that small species of Sorex have relatively unstable local populations but sur\’ive regionally as metapopulations. The arrow means “contributes to. ” Level of Organization Traits and Variables Stability of local populations is decreased by... — Small body size low interference competitive ability -* low density in productive habitats where superior competitors abundant -> high risk of extinction — Small body size and high metabolic rate -»■ short starvation time high risk of extinction of local populations due to environmental stochasticity Stability of metapopulations is increased by... — Small body size -* small per capita food requirement individuals may establish local populations also in unproductive habitat patches — Good colonization ability (reasons discussed in the text) -» high colonization rate Stability of species over evolutionary time is increased by... — Unstable local dynamics -> low rate of speciation — Stable metapopulation dynamics ^ low rate of species extinction h- z LU Q. (/) LU >" LU GC Q. LU I h- Li. O H Z LU O tr LU Q. o Q z < I GC o LL 40 n 30- 20- 5 10 15 SIZE OF THE SHREW (g) Fig. 1. — The percentage of 24 hours spent handling (including killing and ingesting) prey items to satisfy daily energy requirements in different-sized shrews (on the horizontal axis; the same species as in Table 1, all roughly equally represented). The small prey were Cercyon spp. (Coleoptera, fresh weight 1-2 mg), and the large prey were Sphaeridium spp. and Aphodius fossor (Coleoptera, fresh weight 15-35 mg). Each point gives the mean value for ten trials. The regression lines have significantly different slopes (small prey: t = 2.69, P < 0.05; large prey: t = -6.10, P < 0.001). Results from Hanski (1991a, personal observation). 23 1994 HANSKI— Consequences of Body Size in Sorex 1.0- 0.9- X 0.8- Q 0.7- ^ 0.6- ^ 0.5- > n /I 0.2 - 0.1 - 5 10 15 HOURS SINCE THE BEGINNING OF FOOD SHORTAGE Fig. 2. — Activity of small (5. minutus and S. caecutiens\ three individuals, seven experiments) and large shrews {S. araneus and S. isodon; five individuals, eight experiments) during an “energy crisis,” a period of time during which an individual’s energy budget was made experimentally negative by regulating the amount of food to a level 5% less than the requirement, both monitored continuously using an infrared gas analyzer and a computer-controlled feeder. Activity index measures the level of movement activity on an arbitrary scale but adjusted to 0.5 for each individual at the beginning of the experiment to eliminate individual differences. The shaded regions indicate the range of observations for small and large species. For further details see Hanski (1985). SMALL SPECIES LARGE SPECIES CARRYING CAPACITY (K) Fig. 3. — The relationship between mean time to extinction (T, in generations) and the environmental carrying capacity (K) in S. araneus, S. caecutiens, and S. minutus, based on an analysis of their incidence functions on islands in two lakes (Hanski, 1991). The carrying capacity is assumed to be proportional to island area, and it corresponds to the equilibrium population size which the species may attain on the island. PC2 24 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 SKULL BODY PC1 Fig. 4. — Principal component analysis of skull and postcranial skeletal measurements of Sorex species from comparable regions of coniferous forest in Finland and Alberta, western North America. The first principal component (horizontal axis) reflects the general size of the species. The Finnish species are (from the largest species on the left to the smallest species on the right) S. isodon, S. araneus, S. caecutiens, S. minutus, and S. minutissimus\ the North American species are S. arclicus, S. monticolus, S. cinereus, and S. hoyi (H. Virtanen and I. Hanski, personal observation). The ellipses were drawn to enclose all conspecific individuals (ten young individuals per species). The species abbreviations consist of the first three letters of the species’ name (excepting mss for minutissimus). 1994 HANSKI— Consequences of Body Size in Sorex 25 LOG POOLED DENSITY LOG POOLED DENSITY LOG POOLED DENSITY Fig. 5. — Percentages of small, medium, and large species of Sorex against the logarithm of their pooled density in 16 habitat types in Eurasia (the pooled density is used as a measure of habitat productivity). The data come from 38 boreal forest localities distributed throughout Russia. The material includes 11 species, which were divided among small (adult weight less than 4 g; S. minutus, S. minutissimus, S. gracillimus, and S. cinereus), medium (adult weight 5-6 g: S. caecutiens, S. tundrensis, and S. daphaenodon), and large species (adult weight more than 8 g: S. araneus, S. isodon, S. roboratus, and S. unguiculatus\ from B. Sheftel and I. Hanski, in preparation). LOG WEIGHT (G) 26 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 1.0 0.5 mirabilis unguicuiatus isodon araneus caecutiens minutus graciilimus minutissimus WEST EAST RELATIVE ABUNDANCE Fig. 6. — Body sizes (fresh mass) and relative abundances of Sorex in Karelia (west, dark shading) and the Russian Far East (east, light shading), in the western and eastern edges of the Eurasian coniferous forest, respectively. Data for Karelia are from Ivanter (1976) and for the Far East from Okhotina (1974). POPULATION DYNAMICS OF THE SHORT-TAILED SHREW, BLARINA BREVICAUDA Lowell L. Getz Department of Ecology, Ethology, and Evolution, University of Illinois, Urbana, Illinois 61801 Abstract Short-tailed shrew, Blarina hrevicauda, populations in east-central Illinois displayed relatively uniform low amplitude annual population cycles in three habitat types over 18 years of continuous study. There was no indication of multiannual or erratic high-amplitude population fluctuations. Fluctuations of shrew populations were not synchronous with those of two species of microtine rodents, both of which underwent erratic high-amplitude fluctuations in abundance. Generalist predators appear to be the primary source of shrew mortality; mortality rates were relatively constant throughout the year. Snakes do not appear to be major predators of nestling juvenile shrews, as they are for voles in the region. Blarina hrevicauda does not appear to be a major predator on voles, including nestling juveniles. Annual population fluctuations are only weakly correlated with precipitation. Precipitation appears to have a greater influence on adult mortality than on reproduction. Introduction Most species of shrews, including the short-tailed shrew Blarina hrevicauda, appear to display either annual fluctuations in population density (Blair, 1940, 1948; Getz, 1989) or irregular high-amplitude population fluctuations (Harper, 1929; Manville, 1949; Buckner, 1966; Smith et al., 1974; Grant, 1976; Yahner, 1983; Henttonen et al., 1989; Korpimaki and Norrdahl, 1989n). However, shrew populations in boreal regions of Fermoscandia and Siberia display multiannual fluctuations, sometimes in synchrony with those of microtine rodent populations. Specialist predators (e.g., weasels) have been suggested to be responsible for the multiannual shrew-rodent population fluctuations (Hansson and Henttonen, 1985; Henttonen, 1985; Henttonen et al., 1987, 1989; Sonerud, 1988; Korpimaki and Norrdahl, l9S9h). Korpimaki (1986) further concluded that where only the spring decline phases of vole-shrew populations were in synchrony, nomadic avian predators were responsible for the declines. Sheftel (1989) on the other hand, suggested that cyclic fluctuations of shrews observed in central Siberia may result from intrinsic regulatory factors, rather than from predation effects. In more southern regions of Scandinavia both microtine rodent and shrew population fluctuations are annual (Hansson, 1984). The above observations have been used to support the hypothesis that presence of fewer species of generalist predators in northern regions is responsible for the more pronounced multiannual population fluctuations of small mammals commonly observed in these regions (Hansson and Henttonen, 1985; Henttonen et al., 1987). In southern regions more species of generalist predators produce annual population cycles. Elsewhere in Europe and in North America microtine rodents have been observed to undergo both erratic and multiannual fluctuations in population density (Hansson and Henttonen, 1985; Taitt and Krebs, 1985; Getz et al., 1987). However, no evidence has been presented regarding presence or absence of synchrony between shrew and microtine rodent population fluctuations. Such evidence would provide a further test of the influence of generalist and specialist predators on population fluctuations of small mammals. Blarina hrevicauda has been proposed as an important predator on microtine rodents (Eadie, 1944, 1948, 1952). Although the efficiency of B. hrevicauda as a predator on free- living adult voles has been questioned (Barbehenn, 1958; Lomolino, 1984), the shrews may feed on nestling juveniles. If B. hrevicauda does feed extensively on nestling voles, one might expect fluctuations of sympatric populations of these shrews and voles to resemble typical predator-prey interaction curves. Blarina hrevicauda has high moisture requirements (Chew, 1951) and is usually associated with mesic habitats (Pruitt, 1953, 1959; Getz, 1961). One would therefore expect population densities in drier habitats to be positively correlated with precipitation. During the course of a continuing long-term study of prairie vole, Microtus ochrogaster, and meadow vole, M. pennsylvanicus, population fluctuations in east-central Illinois (Getz et al., 1987), data are being obtained regarding population fluctuations of B. hrevicauda. Detailed data are presented for the period of January 1972 through December 1989. These data have been analyzed to determine the pattern of population fluctuations of B. hrevicauda (multiannual or annual), and to determine factors responsible for the observed fluctuations. Based on (1) observations of shrew-vole population fluctuations in southern Scandinavia, (2) the potential for predator-prey interactions between B. hrevicauda and voles, and (3) the existence of high-amplitude fluctuations of vole populations in the study region, I predicted B. hrevicauda would display high-amplitude population fluctuations in synchrony with those of the voles. Study Areas All study sites were located in the University of Illinois Biological Research Area (Phillips Tract) and in Tr el ease Prairie, both 6 km NE Urbana, Illinois (40°15’N, 88°28’W). Populations of B. hrevicauda were monitored in bluegrass (Poa pratensis), alfalfa (Medicago saliva), and restored tallgrass 27 28 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 prairie habitats. Three to eight different study sites were monitored at a time (Getz et al., 1987). The four bluegrass sites (0.5 ha, 0.8 ha, and two each 1.0 ha) were released from domestic grazing in June 1971; by August 1971 there was a dense cover of bluegrass. The bluegrass study sites were mowed approximately 25 cm above the surface during the summer at 2-3 year intervals to control weed growth. Vegetation in the bluegrass sites varied little throughout the study. Relative abundance of plants in the bluegrass habitat was bluegrass (70%); dandelion. Taraxacum officinale (14%); wild parsnip, Pastinaca sativa (4%); and goatsbeard, Tragopogon sp. (3%); and approximately 20 other species with relative abundances of less than 1% (Getz et al., 1979). Four different alfalfa sites were used (three each 1.0 ha, one 1.4 ha); all initially included at least 75% alfalfa. Other species gradually increased in prominence; during the last year of use of each site the common species included, in addition to alfalfa, bluegrass, goldenrod, timothy (Phleum pratense), brome grass (Bromus sp.), clover {Trifolium repens and T. pratense), and plantain (Plantago sp.). Four tallgrass sites were studied: two 0.5 ha each, one 0.7 ha, and one 2.0 ha. The relative abundances of plants in the latter two sites were big bluestem {Atuiropogon gerardi, 17%), bush clover (Lespedeza cuneata, 16%), ironweed (Vernonia sp. , 12%), Indian grass {Sorghastrum nutans, 10%), milkweed {Asclepias sp., 9%), goldenrod (Solidago sp., 9%), bluegrass (5%), switch grass {Panicum sp., 5%), little bluestem {Andropogon scoparius, 2%), and approximately ten other species with relative abundances of less than 1 %. Vegetation in the two 0.5-ha areas was similar to that in the other two, except that Indian grass was the predominant grass ( «50%) and big bluestem less common ( =®5%). The tallgrass study areas were burned in May 1974, 1979, 1984, and 1987 to control invading shrubs and weeds, and to maintain the prairie grasses. In the bluegrass study sites vegetative cover 0. 1-1.0 m above the surface and a dead-litter mat at the soil surface were dense throughout the year. Although herbaceous cover (1. 0-2.0 m above the surface) was dense year-round in tallgrass, the soil surface was more open in this habitat than in bluegrass. The mat of dead vegetation at the soil surface in tallgrass was not as dense as that in bluegrass. However, it should provide shrews protection from avian and large mammalian predators. Vegetative cover was dense in alfalfa from April through December. Because of the growth form of the forbs, the soil surface was relatively open. Owing to dominance of forbs, which lost their leaves in winter, the alfalfa habitat provided poor cover during winter in comparison with bluegrass and tallgrass habitats. Approximately 75% of the soil surface of alfalfa sites was exposed from January to March, and dead vegetation over the remaining surface area rarely was higher than 10 cm. The only other small mammals that occurred more than sporadically in the study areas were prairie and meadow voles, and the western harvest mouse, Reithrodontomys megalotis. Of the two other species of shrews present in the region, less than 25 individuals of the least shrew, Cryptotis parva, and only three southeastern shrews, Sorex longistrostris, were captured. Mammalian predators occurring on the study areas were raccoons, Procyon lotor, mink, Mustela vison\ least and long- tailed weasels, Mustela nivalis and M. frenata respectively; striped skunks. Mephitis mephitis-, red and gray foxes, Vulpes vulpes and Urocyon cinereoargenteus respectively; and domestic cats, Felis cat us. Other predators included Great Homed Owls, Bubo virginianus-. Eastern Screech Owls, Otus asio-. Red-tailed Hawks, Buteo jamaicensis-. Rough-legged Hawks, B. lagopus'. Northern Harriers, Circus cyaneus; and fox snakes, Elaphe vulpina. Methods Trap stations were spaced at 10-m intervals; each had one wooden multiple-catch live trap (Burt, 1940) baited with cracked com. Traps were examined in early morning and late afternoon for three days each month. Traps were covered with vegetation during summer. Owing to the effectiveness of the insulation provided by wood (1.25 cm-thick redwood), bedding material was not placed in traps in winter. All live B. brevicauda were toe-clipped for individual identification when first captured. Owing to difficulty in determining sex and reproductive condition of B. brevicauda by external examination, sex and reproductive condition of live shrews usually were not recorded. From 1977 to 1990 shrews that died in traps were necropsied in the field for sex and reproductive condition (males, testis size; females, size of utems or presence of embryos). Testis length (mm) was visually estimated; uterine size was visually estimated as less than or greater than 2 mm in diameter. Some B. brevicauda can escape from the Burt multiple-catch trap by lifting the door with their nose (Getz, 1961). Accordingly, our data are conservative estimates of population densities of B. brevicauda. Population densities were based on the minimum number known alive during each trapping session. Monthly weather data were obtained from the Illinois State Water Survey weather station located 6 km SW of the study areas. Results Trap Mortality Although many shrews escaped from our traps, mortality was still a serious problem during the study. A total of 5,981 individuals were captured, of which 4,256 (71.2%) were alive at first capture; 796 of these eventually died in the traps. Thus, trap mortality accounted for 42.2% of the total losses from the population. The three habitats differed in overall trap mortality: bluegrass, 43.4% {n — 3,952); alfalfa, 28.3% (n = 1,047); tallgrass, 51.8% (n = 982). There was no evidence that trap mortality in a specific study site differed significantly from that of other sites in the same habitat type. Likewise, trap mortality did not differ seasonally. Despite having a significant impact upon the overall shrew populations and with considerably fewer losses in alfalfa than in tallgrass and bluegrass, trap mortality did not appear to have imposed a serious bias on conclusions. 1994 GETZ — Blarina Populations 29 Population Fluctuations Blarina brevicauda populations in all three habitat types displayed essentially an annual cycle of abundance; there was no indication of multiannual cycles (Fig. 1). Populations in individual study sites within each habitat type also displayed annual population cycles. The mean annual peak densities for the 18 years of the study were 25.6 ± 2.08 (10.0-37.1), 20. 1 ± 2.61 (4.3-54.0), and 15.6 + 1.47 (2.0-18. 4)/ha in bluegrass, alfalfa, and tallgrass, respectively. The mean annual low densities for the three habitat types were 2.6 ± 0.48 (0-8.0), 0.1 ± 0.00 (0-5.0), and 0.4 ± 0.10 (0-3.2)/ha, respectively. The mean annual amplitude of fluctuations (mean of the difference between the annual peak and low densities for each year) were 23.1 + 1.89 (9.0-39.2), 21.0 ± 2.61 (4.3-54.0), and 15. 1 ± 1.48 (3.0-28. 8)/ha, respectively. Peak densities deviated little from year to year from the 18- year mean in each habitat: bluegrass, 7.1 + 1.16/ha (27.7 % of the mean peak density); alfalfa, 7.5 ± 1 . 86 /ha (35.5%); tallgrass, 4.8 ± 0.89/ha (30.8%). The only years of unusually high peak densities were 1981-83 in bluegrass and 1987 in alfalfa (Fig. 1). The annual peak density exceeded the mean peak density eight of the 18 years in bluegrass and nine years in alfalfa and tallgrass. The modal month of annual peak densities was July for bluegrass and October for alfalfa and tallgrass. The month of the peak density deviated from the modal month by two or more months 15 times in 18 years: bluegrass, 1973 (September), 1981 (December), and 1989 (November); alfalfa, 1972 (August), 1974 (July), 1975 (August), 1983 (July), 1984 (August), 1986 (August), 1987 (July), 1988 (May), and 1989 (August); tallgrass, 1975 (August), 1979 (December), and 1983 (August). There was no significant synchrony in terms of deviation of the annual peaks from the 18-year mean peaks (higher, >5 /ha above the mean peak; lower, >5 /ha below; or no deviation, <5 /ha above or below) among the three habitats (Table 1). Similarities among all three habitats existed for only three years: 1975 and 1979, no deviation; 1980, below. There were five years of similarities between bluegrass and alfalfa (three, no deviation; one, lower; one, higher), seven between bluegrass and tallgrass (six, no deviation; one, higher), and three between alfalfa and tallgrass (two, no deviation; one, higher) (Table 1). Overall population densities and amplitudes of fluctuation were higher in bluegrass than in alfalfa or tallgrass. The one major exception was May-September 1987 when population densities in alfalfa were more than twice those in bluegrass (Fig. 1). The generalized annual population cycle differed among the three habitats (Fig. 2). Overwintering densities in bluegrass were approximately 2.5 times those in the other two habitats at the annual low density in early March. Thereafter, population densities in all three habitats increased, bluegrass and alfalfa at essentially the same rate, through June. Actual population densities in bluegrass were higher during the increase phase than in the other two habitats during these months. The bluegrass populations continued to increase to the annual peak in July, followed by a gradual decline to October and November after which numbers declined rapidly to the winter low. Alfalfa densities stabilized from June through September with an increase to the annual peak in October. Thereafter the numbers declined rapidly to the winter low. Tallgrass population densities increased gradually from April through October and then declined rapidly to the winter low. Mean January-March population density correlated significantly with the annual peak density in bluegrass and tallgrass (Spearman’s rank correlation test: = 0.52349, P = 0.0258 and = 0.53120, P = 0.0282, respectively). January-March density and the annual peak in alfalfa did not correlate significantly (r^ = —0.03204, P = 0.9028). Comparisons were made between annual peak population densities and total precipitation for the following periods: (1) previous September-December, (2) January- August of the same year, (3) April-August of the same year, and (4) April-August of the previous year. January-August precipitation correlated significantly with peak annual density the same year in bluegrass (r^ = 0.54724, P = 0.0230), but not in alfalfa or tallgrass. Annual peak densities did not correlate with the amount of precipitation during any of the other time periods. Comparisons were also made of deviations in annual peak densities from the 18 -year mean peak density with deviation of total January-August precipitation from the 70-year mean for the study region (Table 1). Deviations of more than 5 /ha, above or below the mean peak density for the habitat were considered to be significant. In two of the seven years when precipitation was below average, the peak density in bluegrass was more than 5 /ha below the mean peak (Table 1). The peak density was higher than the mean peak for the 18 years in three of the nine years when precipitation was higher than the mean. Peak densities in alfalfa were significantly lower than the mean in four of the seven years with less than average precipitation. There was no consistent relationship between deviation of the annual peak density in alfalfa in those years when precipitation was above average. Likewise, there was no consistent relationship between variation in the annual peak densities and precipitation, whether above or below the mean, in tallgrass sites. Precipitation during the period April-August 1988 was the lowest on record for the region of the study areas (27.7 cm below the mean of 48.6 cm for this period). The alfalfa population peaked in May-June (27/ha, 6.9/ha above the 18- year mean) and in June in bluegrass and tallgrass, 2.6 and 10.6/ha below the mean peaks, respectively. Unlike most other years, population densities in all three habitats declined to very low levels in July-August. Except for a slight recovery in alfalfa in October-November 1988, populations in all three habitats remained very low through July 1989 (Fig. 1). Annual peak densities of B. brevicauda were compared with mean vole population densities of M. ochrogaster and M. pennsylvanicus combined (both species are potential prey) during April-July, the period when most female B. brevicauda were pregnant or lactating. Food, in the form of nestling voles, during this time period would have the greatest potential to result in higher annual B. brevicauda population densities. Vole population density in April-July did not correlate significantly with annual peak B. brevicauda densities in any of the three 30 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 habitats (Kendall’s Tau = —0.1184, —0.3629, and 0.0075, bluegrass, alfalfa, and tallgrass, respectively; significance at 0.05 = 0.468). Survival Overall 30-day survival rates (those animals alive one month that survived until the next month) were highest in bluegrass (47.5%) and lowest in alfalfa (33.4%); survival in tallgrass (42.5%) was only slightly lower than that in bluegrass. Mean monthly survival rates were consistently higher in both bluegrass and tallgrass than in alfalfa, except for September when survival in tallgrass was 3.4% lower in tallgrass than alfalfa (Table 2). Survival curves for the three habitat types (Fig. 3) indicated slightly, but nonsignificantly, higher (x^ = 5.63, d.f. = 2, P > 0.05) survival the first month following initial capture in tallgrass than in alfalfa and bluegrass. The second month following the initial capture survival was significantly lower in alfalfa than in the other two habitats (x^ = 11.6, d.f. = 2, P < 0.01). Thereafter, survival did not differ among the three study areas. The mean persistence times (considering an individual to have entered the population halfway between the previous trapping period and that in which first caught, and to have survived 0.5 month following the last capture) were 1.76 (n = 2.844), 1.55 (n = 830), and 1.69 (n = 567) months for bluegrass, alfalfa, and tallgrass respectively. The mean persistence time for all three habitats combined was 1.71 months (n — 4,241). Fourteen individuals survived 12 months; one survived 17 months. Only 6.0% survived at least four months. Reproduction Males were considered to be reproductively active if the testes were three or more mm long. Testes usually were one or less mm in length during presumed nonreproductive periods. A three-fold increase in size therefore was used as an indicator of reproductive activity. Females were considered reproductively active if embryos were present or if horns of the uteri were two or more mm in diameter. Because relatively few females were observed to have embryos, enlarged uteri was the more common indicator of reproductive activity. Uteri of presumed nonreproductive females were less than 0.25 mm in diameter. Although these are arbitrary indicators, they provide the best evidence of reproductive activity obtainable under field conditions. Blarina brevicauda did not enter live traps until at least three-fourths grown; thus, young of the year normally could not be distinguished from older adults. This may have resulted in underestimation of the proportion of the adult population that was reproductive during late summer-early autumn. Neither could recruitment of young into the population be used as an indicator of reproductive activity. Data from all three habitats have been grouped together for analysis. The two sexes of the population were considered to be reproductive when approximately 40% of the individuals of each sex displayed evidence of reproduction. Reproductive activity of males increased to more than 40% approximately two months prior to that of females (January and March, respectively) and declined to less than 40% two months prior to that of females (July and September) (Table 3). There were no differences in the male or female breeding periods among the three habitats. Reproductive activity of both males and females was lowest in December (11.7 and 7.4%, respectively). There were four years in which precipitation deviated markedly from the 70-year mean during January-August (January- August 1980, 24% lower; May-August 1982, 13% higher; May-August 1983, 13% higher; and April-August 1988, 57 % lower). There was no evidence of major differences in reproductive activity (sexes combined) associated with amount of precipitation during these periods: 1980, 72.9%, n = 59; 1982 and 83 (combined) 76.7%, « = 146; 1988, 50.0%, n = 32. Discussion Blarina brevicauda populations in east-central Illinois displayed a distinct annual cycle in bluegrass, alfalfa, and tallgrass habitats. Annual peak densities were higher in bluegrass than in the other two habitats. In none of the habitats did the amplitude of population fluctuation exceed ten-fold for any year, the minimum amplitude normally ascribed to multiannual population cycles (Taitt and Krebs, 1985). The annual peak densities for each habitat varied from the mean for that habitat by only 27.7-35.5 %. Although population fluctuations in all three habitat types were annual and there was relatively little annual variation in amplitudes of fluctuation in each, annual population cycles among the three habitats were not synchronous. The annual peak normally occurred in July in bluegrass and in October in alfalfa and tallgrass. Synchrony was also compared in terms of deviations of more than 5 /ha above or below the mean peak for each habitat. All three populations were in synchrony in only three years; synchrony between any two habitats occurred during only 5-7 of the 18 years of the study. All but four involved peak densities which did not deviate from the mean peaks for each habitat. Higher overwinter and resultant peak summer densities in bluegrass than in alfalfa and tallgrass most likely result from a combination of differences in winter cover and food availability in the three habitat types. The denser surface vegetation cover afforded by bluegrass, combined with the potential for more ready availability of overwintering invertebrates, favors winter survival. Survival was approximately 20% higher in bluegrass than in alfalfa and tallgrass during the months of December and January. Food availability would also be high early in the spring in bluegrass, supporting rapid population growth. Alfalfa plants started growing in early March; thus, insect and other invertebrate herbivore populations, and in turn food availability for shrews, would also increase in early spring. Rapid population growth in alfalfa, starting from a low density, therefore would be expected. Tallgrass vegetation does not begin its major growth until mid-May. Later food availability, combined with a low winter population density, may be the reason for the slower growth of B. brevicauda populations in tallgrass than in the other two habitats. Given the higher starting 1994 GETZ — Blarina Populations 31 density and rapid rate of increase, population densities in bluegrass would be expected to achieve higher peak densities before the annual decline in reproduction slowed or stopped population growth. Owing to lower starting population densities (both alfalfa and tallgrass) and/or slower population growth rates (tallgrass), population densities were lower in alfalfa and tallgrass than in bluegrass at the end of the breeding period, when population growth stopped. Survival rates did not change during the spring, summer, and autumn in a way that could account for the summer halt of population growth and the autumn decline in numbers. Although the indicator of female reproductive activity (enlarged uteri) remained high until September, actual production of young declined to low levels in August (see below). Thus, the primary reason for termination of annual population growth and the annual late summer-autumn decline of B. brevicauda populations appears to be a decline in reproduction; increased mortality does not appear to be involved. Eventual death of most of the breeding adults in midsummer may have resulted in a large proportion of the population being comprised of young of the year at this time. Thus, a change in the age structure and the resultant influence on reproduction may be a factor in the late autumn population decline of B. brevicauda. There was increased reproductive activity of both males and females in September and October. This may represent reproductive activation of young of the year. Confirmation of these predictions will require detailed analysis of reproductive activity of young of the year and of the timing of disappearance of the spring-early summer breeding adults from the population. Survival of B. brevicauda was lower in alfalfa than in the other two habitats throughout the year. Even though vegetation cover is much less in winter in alfalfa than in bluegrass and tallgrass, mortality rates in alfalfa during December- M arch did not differ from those during spring-autumn, when vegetation cover was more dense. The sparse vegetation cover near the ground surface in alfalfa, as contrasted to the more dense cover at the ground surface of bluegrass and tallgrass, appears a likely factor in the year-round higher mortality rates in alfalfa. Avian and large mammalian predators probably would be more efficient in alfalfa during summer and autumn than in bluegrass or tallgrass. During spring-early autumn small mammalian predators (e.g., weasels) and perhaps snakes may also be more efficient predators in alfalfa than in the other two habitats owing to the more open unobstructed ground surface in the former. Overall persistence (in situ mortality and emigration from the study site were not separated) on the study areas ranged from 1.55 months in alfalfa to 1.76 months in bluegrass; mean persistence for all habitats combined was 1.71 months. This is approximately the same persistence times for the two species of voles in the same habitats (Getz et al., 1979). Although differential food availability among the habitats may also be a factor in both vole and shrew survival (Cole and Batzli, 1979), it appears that voles and shrews are equally subject to mortality and emigration; predation within the study site is presumed to be the primary source of disappearance of voles. Approximately 3.3% of M. ochrogaster emigrated from a study site (Getz et al., 1990a); data are not available regarding emigration of B. brevicauda. Appearance of a few new adult B. brevicauda during the winter suggests that at least some emigration does take place. There was no correlation between the annual peak shrew densities and April-July population densities of voles. Three of the four years (1981, 1982, and 1987) of highest shrew population densities were during periods of relatively low spring-early summer vole population densities (Getz et al., 1987; unpublished data). Vole populations were relatively high during spring-early summer 1983 when shrew densities were also very high. Furthermore, there were no concurrent rapid declines in B. brevicauda and vole populations. Most of the distinct population declines of M. ochrogaster in the various study sites were during late winter (28); eight were during spring, and five during the summer. Declines of M. pennsylvanicus also occurred primarily in late winter (17); seven were in spring, three in summer, and only one in autumn. All B. brevicauda declines occurred in autumn-early winter (Fig. 2). Microtus ochrogaster never declined in autumn-early winter and M. pennsylvanicus only once. I conclude from these observations that: (1) B. brevicauda is not a significant predator on voles, including juveniles in the nest. At least voles do not constitute a major food source for B. brevicauda during the breeding period. (2) Population fluctuations of voles and B. brevicauda in east-central Illinois are not regulated in the same manner. Although there is no evidence for multiannual population cycles of voles in east- central Illinois (Getz et al., 1987), both M. ochrogaster and M. pemusylvanicus undergo high-amplitude, but erratic, fluctuations in numbers. During these periods of high-amplitude fluctuations in vole densities B. brevicauda populations displayed relatively uniform, low-amplitude fluctuations in abundance. There is anecdotal evidence of least weasel involvement in some of the vole population declines (Getz et al., unpublished data). However, because voles and B. brevicauda do not decline concurrently, variation in predation by these specialist predators does not appear to be involved in population regulation of B. brevicauda. The above observations agree with those of Korpimaki and Norrdal (1989/?), who concluded that least weasels are not important predators on shrews. Neither was there evidence to suggest that avian predators contributed to either vole or B. brevicauda declines. Hawks were rarely observed flying over the study areas, including during periods of population decline, and few owls were known to roost within the vicinity of the study areas. Pellets of owls and hawks were not examined for prey utilization. The results of this study also support the conclusions of Hansson (1984) and Henttonen et al. (1989), that where voles do not display multiannual cycles, generalist predators are more important sources of mortality than where distinct multiannual population cycles occur. In the latter regions specialist predators are involved in generating multiannual vole cycles and as a consequence shrew populations also display distinct multiannual cycles. Generalist predators on voles in east-central Illinois include house cats, large avian predators, and snakes. Snakes are especially important sources of mortality on nestling juvenile voles (Getz et al., 1990b). However, snakes do not 32 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 appear to be a primary source of shrew mortality. Snakes are active until mid-October, at which time both juvenile vole survival and the general population density increase. Although B. brevicauda populations begin declining in August in bluegrass, when snakes are active, the declines in alfalfa and tallgrass do not begin until November, after snakes have entered hibernation. If snakes are a major source of mortality on shrews, they take adults throughout the spring, summer, and early autumn and do not feed selectively on nestling juveniles. January-August precipitation and peak B. brevicauda densities correlated only in bluegrass. Peak densities tended to be below the 18-year mean in both bluegrass and alfalfa during those years when precipitation was below the January-August mean. However, there was no consistent relationship between peak densities in these two habitats and precipitation above the January-August mean, nor between deviation in precipitation above or below the mean and peak shrew densities in tallgrass. Thus, the relationship between precipitation and annual population cycles of B. brevicauda in east-central Illinois is weak. The extreme drought of 1988 afforded additional opportunity to evaluate the effects of precipitation on B. brevicauda populations. During the January-May period (before the summer drought), 43.8% of the females (n = 16) were reproductive, but during June-September 92.3 % (« = 13) were reproductive. Comparable data for these periods over the entire study are 53.9 (n = 191) and 60.9% (n = 409), respectively. Although sample sizes for 1988 were small, the drought apparently had no major impact on reproduction. Survival data were analyzed in terms of adults present in June 1988 (start of the drought) that survived to July (thereafter sample sizes were too small to be meaningful). For all habitats combined 9.3 % (n = 75) survived from June to July. This is much less than the 41.1% (n = 200) June-July survival recorded over the entire study. Thus, increased adult mortality appears to be the major response to the drought conditions resulting in the early population decline in the summer of 1988. Reasons for high mortality were not obvious. The vegetation, including grasses and forbs, in both bluegrass and tallgrass was extremely dry and decimated by the drought. Very little green vegetation was present in either habitat from late June through July. In contrast, the alfalfa plants were robust and succulent throughout the drought. Although invertebrate abundance was not sampled, insect populations most likely would have remained high in alfalfa. Although no data are available, it is difficult to perceive the other two habitats as supporting equally high insect populations. The upper 5-10 cm of soil in all three habitats was extremely dry and hard-packed. June-July mortality was higher in alfalfa than in the other two habitats; none of the 24 individuals present in June survived to July (9.8% of the 51 animals present in June survived to July in bluegrass and tallgrass). I suggest that the reasons for the early rapid decline in B. brevicauda populations during the drought of 1988 resulted from adult mortality influenced by factors other than food availability. Water deprivation, combined with low humidities, may have been responsible for the increased mortality at this time (Pruitt, 1953, 1959; Getz, 1961). Reproduction in B. brevicauda populations in east-central Illinois is limited primarily to March- July. Males become reproductive in January, but females do not do so until March, at which time food resources are assumed to be sufficient to support the demands of production of young. At this time vegetation growth resumes and invertebrates presumably become more available. Although testes start regressing in July, with reproductive activity becoming very low by August, female reproductive activity (mostly enlarged uteri rather than pregnancy) extends into September. There is undoubtedly a period of time following birth of the last litter before uterine size regresses to the nonreproductive state. Further, males probably would not successfully mate following regression of their testes in August. As indicated above, I could not separate the effects of loss of older breeding animals from the population on the perceived overall proportion of the population in reproductive condition in late summer. A greater proportion of the population being comprised of young of the year may be responsible for lower reproductive activity within the population at this time. Regardless of the reason, I conclude the main breeding period of B. brevicauda in east-central Illinois to be spring-early surruner (March-July). Reproduction apparently was not influenced by variation in precipitation; this differs from the results of Pankakoski (1985), who found a positive relationship between reproductive success and precipitation for Sorex araneus in Finland. There is no evidence for major autumn breeding success. However, approximately 21% of the females and 29% of the males appear to be reproductive into October. Owing to the short life span (Fig. 3), approximately 2% of the young bom in October would be expected to survive to the spring breeding period. Thus, even though reproduction is low in autumn, young produced at this time may become the breeders the next spring. Overall survival rates are too low for many young produced during the spring-early summer breeding period to survive until the next year. In summary, B. brevicauda displays low-amplitude annual population fluctuations in east-central Illinois. Population fluctuations of B. brevicauda differ, and are influenced by different factors, from those of voles in east-central Illinois; the latter displayed erratic multiarmual fluctuations. There is no indication of predator-prey interactions between voles and B. brevicauda sufficient to affect population fluctuations of either the shrew or voles. Although several generalist predators feed on both voles and shrews in east-central Illinois, snakes do not appear to influence B. brevicauda population fluctuations as they do those of voles. The annual peak density and population cycle of B. brevicauda are assumed to be influenced primarily by (1) population density at the begirming of the spring breeding period, (2) timing of the annual increase in food availability, and (3) a midsummer decline in reproduction. Population density at the beginning of the spring breeding period will be higher in those habitats providing good overwinter vegetation cover and food availability than where such conditions do not exist. Further, population growth will be more rapid where and when food availability increases early in the breeding period 1994 GETZ — Blarina Populations 33 than where such increases in food availability occur later. The combined effects of population density at the beginning of the breeding period and timing of food availability determine the peak density achieved before reproduction declines, thereby stopping population growth for the year. The annual nature of the population cycle appears to be generated by the marked decline in reproduction begirming in July. Variation in mortality is not involved. I propose that most of the overwintering breeding adults may finally have been lost from the population by July. Further, young of the year may not become reproductive until autumn. If so, a decline in food availability and/or the early onset of winter may result in such a brief autumn breeding period that population growth cannot be maintained. Thus, it is possible that the annual population cycle of B. brevicauda results from (1) the normal mortality of overwintering adults, (2) delay in reproductive maturity of young of the year, and (3) conditions unfavorable to sustain high levels of reproduction during the winter. Acknowledgments This study was supported in part by grants NSF DEB 78- 25864 and NIH HD 09328. I thank the following individuals for their assistance with the field work: J. Hofmann, P. Mankin, L. Vemer, R. Cole, B. Klatt, B. McGuire, T. Pizzuto, M. Snarski, D. Tazik, M. Schmierbach, D. Avalos, J. Edgington, B. Frase, L. Schiller, S. Vanthoumout, and the 796 undergraduate “Mouseketeers” who have participated in the field program. Literature Cited Barbehenn, K. R. 1958. Spatial and population relationships between Microlus and Blarina. Ecology, 39:293-304. Blair, W. F. 1940. Notes on home ranges and populations of the short-tailed shrew. Ecology, 21:284-288. 1948. Population density, life-span, and mortality rates of small mammals in the blue-grass meadow and blue-grass field associations of southern Michigan. The American Midland Naturalist, 40:395-418. Buckner, C. H. 1966. Populations and ecological relationships of shrews in tamarack bogs of southeastern Manitoba. Journal of Mammalogy, 47:181-194. Burt, W. H. 1940. Territorial behavior and populations of some small mammals in southern Michigan. Miscellaneous Publication of the Museum of Zoology, University of Michigan, 45:1-58. Chew, R. W. 1951. The water exchanges of some small mammals. Ecological Monographs, 21:215-225. Cole, F. R., and G. O. Batzli. 1979. Nutrition and population dynamics of the prairie vole, Microlus ochrogaster, in central Illinois. The Journal of Animal Ecology, 48:455-470. Eadie, W. R. 1944. The short-tailed shrew and field mouse predation. Journal of Mammalogy, 25:359-364. 1948. Shrew-mouse predation during low mouse abundance. Journal of Mammalogy, 29:35-37. 1952. Shrew predation and vole populations on a localized area. Journal of Mammalogy, 33:185-189. Getz, L. L. 1961. Factors influencing the local distribution of shrews. The American Midland Naturalist, 65:67-88. 1989. A 14-year study of Blarina brevicauda populations in east- central Illinois. Journal of Mammalogy, 70:58-66. Getz, L., L. Verner, F. Cole, J. Hofmann, and D. Avalos. 1979. Comparisons of population demography of Microlus ochrogasler and M. pennsylvanicus. Acta Theriologica, 24:319-349. Getz, L., J. Hofmann, B. Klatt, L. Verner, R. Cole, and R. Lindroth. 1987. Fourteen years of population fluctuations of Microlus ochrogasler and M. pennsylvanicus in east-central Illinois. Canadian Journal of Zoology, 65:1317-1325. Getz, L., B. McGuire, J. Hofmann, T. Pizzuto, and B. Frase. 1990a. Social organization and mating system of the prairie vole Microlus ochrogasler. Pp. 69-80, in Social Systems and Population Cycles in Voles (R. H. Tamarin, R. S. Ostfeld, S. R. Pugh, G. Bujalska, eds.), Birkhauser Verlag, Basel. Getz, L., N. G. Solomon, and T. M. Pizzuto. \990h. The effects of predation of snakes on social organization of the prairie vole, Microlus ochrogasler. The American Midland Naturalist, 123:365-371. Grant, P. R. 1976. An 1 1 -year study of small mammal populations at Mont St. Hilaire, Quebec. Canadian Journal of Zoology, 54:2156-2173. Hansson, L. 1984. Predation as the factor causing extended low densities in microtine cycles. Oikos, 43:255-256. Hansson, L., and H. Henttonen. 1985. Gradients in density variations of small rodents: the importance of latitude and snow cover. Oecologia, 67:394-402. Harper, F. 1929. Notes on the mammals of the Adirondacks. Handbook of the New York State Museum, 8:51-118. Henttonen, H. 1985. Predation causing extended low densities in microtine cycles: Further evidence from shrew dynamics. Oikos, 45:156-157. Henttonen, H., T. Oksanen, A. Jortdcka, and V. Haukisalmi. 1987. How much do weasels shape microtine cycles in the northern Fennoscandian taiga? Oikos, 50:353-365. Henttonen, H., V. Haukisalmi, A. Kaikusalo, E. Korpimaki, K. Norrdahl, and U. a. P. Skaren. 1989. Long-term population dynamics of the common shrew, Sorex araneus. Annales Zoologici Fennici, 26:349-355. Korpimaki, E. 1986. Predation causing synchronous decline phases in microtine and shrew populations in western Finland. Oikos, 46:124-127. Korpimaki, E., and K. Norrdahl. 1989a. Avian predation on mustelids in Europe 2: Impact on small mustelid and microtine dynamics — a hypothesis. Oikos, 55:273-276. 1989fc. Avian and mammalian predators of shrews in Europe: regional differences, between-year and seasonal variation, and mortality due to predation. Annales Zoologici Fennici, 26:389-400. Lomolino, N. V. 1984. Immigrant selection, predation, and the distributions of Microlus pennsylvanicus and Blarina brevicauda on islands. The American Naturalist, 123:468-483. Manville, R. H. 1949. A study of small mammal populations in northern Michigan. Miscellaneous Publications of the Museum of Zoology, University of Michigan, 73:1-83. Pankakoski, E. 1985. Relationship between some meteorological factors and population dynamics of Sorex araneus in southern Finland. Acta Zoologica Fennica, 173:287-289. Pruitt, W. O. 1953. An analysis of some physical factors affecting the local distribution of the short-tail shrew (Blarina brevicauda) in the northern part of the Lower Peninsula of Michigan. Miscellaneous Publications of the Museum of Zoology, University of Michigan, 79:1-39. 1959. Microclimates and local distribution of small mammals on the George Reserve, Michigan. Miscellaneous Publications of the Museum of Zoology, University of Michigan, 109: 1-27. 34 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 Sheftel, B. 1. 1989. Long-term and seasonal dynamics of shrews in central Siberia. Annales Zoologici Fennici, 26:357-369. Smith, M. H., J. B. Gentry, and J. Finder. 1974. Annual fluctuations in small mammal populations in an eastern hardwood forest. Journal of Mammalogy, 55:231-234. SONERUD, G. A. 1988. What causes extended lows in microtine cycles? Analysis of fluctuations in sympatric shrew and microtine populations in Fennoscandia. Oecologia, 76:37-42. Taitt, M. j., and C. j. Krebs. 1985. Population dynamics and cycles. Pp. 567-620, in Biology of New World Microtus (R. H. Tamarin, ed.). Special Publication, The American Society of Mammalogists, 8:1-893. Yahner, R. H. 1983. Population dynamics of small mammals in farmstead shelterbelts. Journal of Mammalogy, 64:380-386. Table 1. — Synchrony in the deviation of the annual peaks t^Blarina brevicauda from the 18-year mean peak for each habitat type. L, >5/ha below the mean peak; H, >5 /ha above the mean peak; 0 <5/ha above or below the mean peak. Relationship between actual deviation (in parentheses) in the annual peak population density (n/ha) and deviation of the January- August precipitation from the 70-year mean total (65 cm) for those months. Year Bluegrass Alfalfa Tallgrass Precipitation {% Deviation) 1972 0(-5) L(-15) 0(-2) -30 1973 0( + 3) H( + 7) 0(0) + 33 1974 0( + 2) 0(0) L(-6) + 4 1975 0( + 4) 0(0) 0( + 3) + 1 1976 0( + 3) L(-17) 0( + l) -19 1977 0(-4) L(-IO) H(-b8) -26 1978 0(-2) H( + 7) 0(+l) -16 1979 0(-2) 0(-3) 0(-4) + 8 1980 L(-IO) L(-8) L(-7) -24 1981 H(+17) 0( + 2) 0(-4) + 9 1982 H(+13) 0( + 3) H( + 7) + 13 1983 H(+16) H(+10) 0( + l) + 13 1984 H( + 6) 0(-5) L(-8) 0 1985 L(-6) L(-7) 0( + 4) + 4 1986 L(-12) 0(0) 0( + 3) -16 1987 L(-6) H(-b33) H(+14) + 5 1988 0(-3) H( + 6) L(-IO) -30 Table 2.— Mean monthly survival (% present that month surviving to next month) r^Blarina brevicauda, 1972-1989. Sample sizes in parentheses. Habitat Type Month Bluegrass Alfalfa Tallgrass Total January 57.9(154) 27.8(18) 30.6(49) 49.3(221) February 45.1(153) 28.6(14) 47.6(21) 44.1(188) March 58.6(116) 41.7(12) 44.4(18) 55.5(146) April 47.5(202) 22.6(31) 47.6(21) 44.5(254) May 41.6(320) 27.3(77) 38.3(47) 38.7(444) June 44.5(474) 28.1(153) 49.0(49) 41.1(676) July 50.9(556) 36.4(154) 57.5(80) 48.7(790) August 54.6(533) 39.8(161) 58.3(127) 52.2(821) September 48.4(481) 43.9(132) 40.5(153) 46.1(766) October 46.4(470) 32.0(128) 37.1(170) 41.9(768) November 39.5(435) 25.6(78) 35.0(123) 37.0(636) December 42.6(258) 22.8(35) 29.2(65) 38.3(358) Mean 47.5 33.4 42.5 44.4 1994 GETZ — Blarina Populations 35 Table 3.— Percent male and female Blarina brevicauda displaying evidence of reproductive activity (see text), as determined by necropsied animals from all study sites combined. Sample size in parentheses. Month Males Percent Reproductive Females Combined January 45.0(20) 5.9(51) 16.9(71) February 62.5(24) 5.3(19) 37.2(43) March 92.3(13) 78.5(14) 85.2(27) April 85.7(28) 77.4(31) 81.4(59) May 83.3(72) 84.2(76) 83.8(148) June 62.5(72) 74.2(128) 70.0(200) July 38.5(91) 55.5(117) 48.1(208) August 15.4(65) 42.5(87) 30.9(152) September 26.5(49) 66.2(77) 50.8(126) October 28.7(108) 21.4(145) 24.5(253) November 18.6(140) 13.0(154) 15.6(294) December 11.7(77) 7.4(95) 9.3(172) 36 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 25 20 15 10 5 0 Fig. 1.— Population fluctuations of Blarina brevicauda in three habitat types in east-central Illinois and total monthly precipitation (stippled bar graphs). Precipitation (cm) 1994 GETZ — Blarina Populations 37 Fig. 2. — Generalized annual population cycles of Blarina brevicauda in three habitat types in east-central Illinois. Data represent monthly mean ( + SEM) population densities from all study sites in each habitat for 1972-1989. Percent Surviving 38 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 Fig. 3.— Survival of Blarina brevicauda following first capture for individuals alive at first capture. Sample sizes in parentheses. A LIVE-TRAPPING STUDY OF TWO SYNTOPIC SPECIES OF SOREX, S. CINEREUS AND S. FUMEUS, IN SOUTHWESTERN PENNSYLVANIA J. Michelle Cawthorn Department of Biological Sciences, Bowling Green State University, Bowling Green, Ohio 43403; current address: Department of Biology, Ball State University, Muncie, Indiana 47306-0440 Abstract This study examined the population biology of two syntopic soricids, Sorex cinereus (masked shrew) and S. fume us (smoky shrew), in southwestern Pennsylvania. I used dry pitfalls as live-traps; two 1-ha areas were trapped (100 traps each) for three summers. Traps were opened for 6-8 h once a week and checked every 1 .5 h. Fifty-six individual shrews were captured, marked , and released. More shrews were caught during the first summer. No individual was recaptured more than four times and only two individuals were recaptured between summers. Low capture rates were due both to inefficiency of the pitfall traps and to a real decline in density. Introduction Although many species of Sorex are found in a variety of habitats, and may constitute a substantial portion of small mammal communities in terms of species diversity, their high metabolic rates, small size, and concomitant low survival rates in live traps have limited studies of their population biology. Shrews rarely survive more than 2-3 hours in a trap and many shrews are not heavy enough to be reliably trapped in most traditional box traps, thus sample sizes are small. Most of what is known about shrews comes from studies in which shrews were kill -trapped. Although such information is valuable, it conveys little about the behavior of shrews. Many studies of shrews have been designed to define microhabitat and diet of sympatric species to examine the importance of intra- and interspecific competition in structuring insectivore assemblages (Spencer and Pettus, 1966; Brown, 1967; Wrigley et al., 1979; Churchfield, 1984; French, 1984; Whitaker and French, 1984; MacCracken et al., 1985; Ryan, 1986). It has been suggested that competition may be important in structuring insectivore communities, based on the absence of habitat and dietary overlap (e.g., Whitaker and French, 1984) and competitive release on islands (Ellenbroek, 1980; Malmquist, 1986). However, important questions of spatial and temporal activity patterns have not been addressed. Some investigators have successfully examined insectivore communities using innovative live-trapping methods, including the Longworth trap (Chitty and Kempson, 1949), and dry pitfall traps (Buckner, 1966) combined with frequent trap checks (every 1.5-3 hours). However, most (Croin-Michelson, 1966; Ellenbroek, 1980; Churchfield, 1984) live-trapping studies of shrews have been conducted in Britain and Europe. The objectives of this study were to examine the ecology of two sympatric species of Sorex, S. cinereus, (masked shrew), and S. fumeus, (smoky shrew) using dry pitfall traps as live traps. In western Pennsylvania, these two species are sympatric with two or three other species of Sorex, Blarina brevicauda, and several other species of small mammals. My primary study objectives were to collect data on changes in population density, reproductive parameters, home range size and exclusivity, and temporal activity patterns. Another goal of this study was to quantify the efficiency of pitfall traps as live traps, as it is known that pitfall trapping is the most effective method of collecting shrews (e.g., Williams and Braun, 1983). The relationship of these two species to their microhabitat and to each other was further examined by quantifying microhabitat parameters. Methods Study Species. — Sorex cinereus is a small shrew which rarely weighs over 5 g, whereas S. fumeus weighs between 5-10 g. Sorex fumeus is found throughout northeastern North America primarily in forests. Sorex cinereus is more catholic in habitat requirements and can be found in a variety of habitats from forests to bogs (Merritt, 1987). Both species are sympatric over much of their ranges with other Sorex species. Breeding begins in March and continues through July or August. Juveniles may breed in their first summer. Home ranges of individuals (5. cinereus) overlap only slightly and do not change with season (Buckner, 1966). Population density varies from year to year, ranging from 1/ha to over 124/ha (Merritt, 1987). Most shrews live less than one month. Those that survive the juvenile and subadult stages live 13-18 months, breed during the spring and summer, and die before the end of summer (Hamilton, 1940; Buckner, 1966). Individuals are active day and night, and moderate rainfall increases activity significantly (Doucet and Bider, 1974; Vickery and Bider, 1977). Trapping Procedure. — The study was conducted during 1985 through 1987 at Powdermill Biological Station, Carnegie Museum of Natural History’s field station in southeastern Westmoreland County, Pennsylvania. The study area is heavily wooded with second growth deciduous forest, dominated by sugar maple {Acer saccharum), tulip poplar (Liriodendron tulipifera), and American beech {Fagus grandifolia). Two study grids, located approximately 1 km apart on an east-facing slope, were similar in plant species composition. However, one grid (Moul Spring) was 30 m higher in elevation than the other (Calverley Lodge) and had steeper topography. Shrews were trapped in dry pitfall traps placed 10 m apart in a 10 X 10 array (two 1-ha grids). Pitfall traps were made from #10 size cans (22.86 cm x 12.3 cm), buried up to the rim 39 40 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 near a log or pile of rocks within 1 m of the measured trap station. Traps were closed between trapping sessions by placing a section of log, slightly smaller in both diameter and height, in the can and placing a square of roofing shingle (12.7 x 12.7 cm) over the top. Each can was punctured through the bottom to allow water drainage. In some areas where traps were near streams, water collected in the cans during nonsampling periods and had to be removed. Trapping was conducted on the two grids 1 day/week for 6-8 h/day during May through August. During March, April, September, October, and November of these years, trapping was conducted 1 day /month for 6-8 h/day. Trapping was conducted only during daylight hours, as it has been shown by previous investigators that diurnal trapping is successful (Churchfield, 1980, 1984). Open traps were provided with ground beef, which presumably acted as a food source (not a bait) once the animal was trapped, and a piece of moss which provided both cover and moisture. Traps were checked approximately every 1.5 h. Using this regime, only two shrews died in traps during the study. All shrews were individually marked at first capture by toe clipping. Species determination, body mass, time and location of capture, sex, reproductive condition, and tooth pigmentation were recorded for every capture. Males were identified either by the presence of visible testes or an everted penis, and females were identified by the presence of visible nipples. Individuals that were not sexually active could not be sexed. Males were classified as having either visible or nonvisible testes, and nipple size in females was described as small, medium, or large. Pregnancy and lactation in females also were recorded when apparent. Soricid teeth are pigmented, but tooth wear gradually removes the pigment. Therefore, approximate age can be determined by examining the amount of pigmentation on the teeth (Hamilton, 1940; Rudd, 1955). In this study, shrews with heavy pigmentation were assumed to be young of the year, and those with medium and light pigmentation were older (animals that had overwintered). Juveniles were also distinguished from adults by body mass: S. cinereus that weighed less than 3 g and S. fumeus that weighed less than 5 g were classified as juveniles. Microhabitat Analysis. — Microhabitat data were collected at each of the 200 trapping stations using a method similar to that of Dueser and Shugart (1978) and Porter and Dueser (1982) and described in detail in Chandler (1989). Data from the two grids were pooled for the microhabitat analysis. When appropriate, transformed (log or square root transformations) variables were used in order to meet the assumption of normality. Stations where shrews occurred and stations where shrews were absent, and stations where only S. cinereus occurred and stations where only S. fumeus occurred were compared using Discriminant Function Analysis (DFA). All statistical analysis were performed using SAS (SAS Institute Inc., 1985). Results and Discussion Trapping. — Three species of shrews, S. cinereus, S. fumeus, and B. hrevicauda were caught on each grid during the three years of the study (Table 1). Trap success was low (0.5%); during 5,301 trap days, 1 12 Sorex and 32 Blarina were captured. Trap success on both grids was highest at the beginning of the study, declined through the summer of 1985, and remained low during 1986 and 1987 (Fig. 1). Of the 144 captures, 46 were recaptures. Of those, 32 were recaptures of Sorex and most were recaptures of S. cinereus. Shrews were captured simultaneously in a trap four times during the study, resulting in the death of two shrews. No individual was recaptured more than four times. Many of the recaptures occurred in the same trap or traps adjacent to the original capture site, suggesting that shrews had regular runways. Trap mortality in this study compared favorably to that reported in previous live-trapping studies of Sorex. Estimates of trap mortality have been between 0.04% (Hawes, 1977) and 8% (Churchfield, 1980). In previous studies, traps were checked frequently (every 1.5-3 h) or an appropriate food source was left in the trap (Buckner, 1966; Croin-Michielsen, 1966; Hawes, 1977; Churchfield, 1980; Ellenbroek, 1980). Furthermore, several investigators have opened traps only during daylight hours, thus avoiding the problem of other nocturnal small mammals occupying the traps (Croin- Michielsen, 1966; Hawes, 1977; Churchfield, 1980; Ellenbroek, 1980). Density was estimated by the minimum number known to be alive method (MNA; Krebs et al., 1969). Initially, density in June exhibited a high of 13 5. cinereus and 6 S. fumeus lha, but declined through the summer of 1985 and remained relatively constant during 1986 and 1987 (Fig. 1). Density was 3 /ha for S. cinereus and 1-3/ha for S. fumeus. This estimate of density should be considered conservative, as the recapture rate was low and because there was incomplete sampling of individuals. Reported shrew densities range from 3 /ha to 124/ha (Hamilton, 1940) and changes in density between and within years are well-documented (e.g., Buckner, 1966). The high density recorded at the beginning of this study, followed by lower densities in subsequent years, is consistent with the notion that shrew densities change between years. It seems unlikely that individuals remembered the location of traps between years, but this factor may have contributed to the initial decline in capture success during the summer of 1985. Churchfield (personal communication) and Crowcroft (1957) have suggested that shrews regularly shift runways, and that this may influence trap success when traps cannot be moved. The age structure of the population changed during each summer, as evidenced by changes in mass and tooth pigmentation. During May and June most animals captured were adults, based on both weight of animals and tooth pigmentation. During this period, animals were relatively heavy and few had completely pigmented teeth. However, during July and August the population was dominated by subadults and juveniles, as indicated by lower average weights, and fully- pigmented teeth in most individuals. Weights for S. cinereus ranged from 2. 2-5. 6 g, and for S. fumeus from 5. 0-9.0 g. Most reproduction occurred in May, June, and July, because it was only during these months that shrews could be sexed by enlarged nipples in females or enlarged testes in males. Juveniles first appeared in June and were present through 1994 CAWTHORN— Population Dynamics of Shrews 41 August, which indicates late spring or early summer reproduction and supports the conclusions of previous investigators (e.g., Owen, 1984). Diel activity, measured by time of day when shrews were captured, occurred throughout the day, although there were periods of greater and lesser activity. Because trapping effort after dark was minimal, no observation on differences in activity during day and night could be made within or between the grids. There was no significant correlation between activity of the two species (Grid 1, r = 0.155, NS; Grid 2, r = 0.033, NS). Both species were most active during the morning and late afternoon and early evening hours. This is similar to the findings of Buckner (1966), Croin-Michielsen (1966), and Pemetta (1977) for syntopic species. It may be that activity in shrews is constrained because of their high metabolic rates, so that a difference in activity periods is unlikely to be detected. On the other hand, given the low population density, the potential for direct interaction between these species was probably slight in most years. Microhabitat Analysis. — The grids were established in sites having similar plant species composition (Jacard’s similarity index = 0.83), and density of shrews on the two grids was similar, so data from both grids were combined for the microhabitat analyses. Data were partitioned into two groups for discriminant function analysis (DFA): stations where shrews were captured versus stations where shrews were not captured, and stations where only S. cinereus were captured versus stations where only S. fumeus were captured. In the first analysis, there was significant discrimination between sites where shrews were captured and sites where shrews were not captured (Wilks’ Lambda = 0.70, d.f. = 27, F = 1.64, P < 0.03; Fig. 2). In this case the canonical variate described a gradient of stations with low tree densities (poor habitat) to stations with high woody species richness and higher woody stem density (good habitat). Both species were caught more often in areas with high woody stem density, and less often in open areas with low woody stem density. In the second analysis, there was no significant discrimination between traps where only S. cinereus occurred and traps where only S. fumeus occurred (Wilks’ Lambda = 0.58, F = 1.33, d.f. = 27, NS). None of the variables which have traditionally been described as important components of shrew habitat (e.g., fallen logs, Hamilton, 1940) were important in the DFA. This lack of variation in microhabitat is not surprising considering the history of the region. Both study areas were selectively cut approximately 50 years ago, and there are no differences in soil types or slopes on the grids. Consequently, the vegetation on the grids is of uniform age and type. Both species of shrews in this study are primarily forest dwellers, although S. cinereus has more catholic habitat preferences than S. fumeus (Merritt, 1987). Of course, the relatively small sample size of shrews captured in this study hampered the detection of differences in microhabitat use. Acknowledgments I thank Joseph Merritt for making the work at Powdermill Biological Station possible. J. Cawthom, R. Chandler, M. Gromko, H. Korber, G. Lenhart, V. Mosca, J. Rotenberry, and S. Vessey all provided valuable assistance. This research was supported by the Roosevelt Memorial Fund of the American Museum of Natural History and Bowling Green State University. Literature Cited Brown, L. N. 1967. Ecological distribution of six species of shrews and eomparison of sampling methods in the eentral Roeky Mountains. Journal of Mammalogy, 48:617-623. Buckner, C. H. 1966. Populations and ecologieal relationships of shrews in tamarack bogs of southeastern Manitoba. Journal of Mammalogy, 47:181-194. Chandler, M. C. 1989. The population eeology and temporal and spatial activity of three species of shrews (Sorex cinereus, S. fumeus, and Blarina brevicauda) in southwestern Pennsylvania. Unpublished Ph.D. dissert.. Bowling Green State University, Bowling Green, Ohio, 124 pp. Chitty, D., and D. A. Kempson. 1949. Prebaiting small mammals and a new design of live traps. Eeology, 30:536-542. Churchfield, S. 1980. Population dynamies and the seasonal fluctuations in numbers of the common shrew in Britain. Aeta Theriologiea, 25:415-424. 1984. Dietary separation in three species of shrew inhabiting water-eress beds. Journal of Zoology (London), 204:211-228. Croin-Michielsen, N. 1966. Intraspecific and interspecific competition in the shrews Sorex araneus L. and S. minutus L. Archives Neerlandaises de Zoologie, 17:73-174. Crowcroft, P. 1957. The Life of the Shrew. Max Reinhardt, London, 166 pp. Doucet, G. j., and j. R. Bider. 1974. The effects of weather on the activity of the masked shrew. Journal of Mammalogy , 55:348-363. Dueser, R. D., and H. H. S hug art, Jr. 1978. Microhabitats in a forest-floor small mammal fauna. Eeology, 59:89-98. Ellenbroek, F. j. M. 1980. Interspecifie competition in the shrews Sorex araneus and Sorex minutus (Soricidae, Inseetivora): A population study of the Irish pygmy shrew. Journal of Zoology (London), 192:119-136. French, T. W. 1984. Dietary overlap of Sorex longirostris and S. cinereus in hardwood floodplain habitats in Vigo County, Indiana. American Midland Naturalist, 111:41-46. Hamilton, W. J., Jr. 1940. The biology of the smoky shrew {Sorex fumeus fumeus Miller). Zoology, 25:473-491. Hawes, M. L. 1977. Home range, territoriality, and ecological separation in sympatric shrews, Sorex vagrans and Sorex obscurus. Journal of Mammalogy, 58:354-367. Krebs, C. J., B. L. Keller, and R. H. Tamarin. 1969. Microtus population biology: Demographie ehanges in fluctuating populations of M. ochrogaster and M. pennsylvanicus in southern Indiana. Eeology, 50:587-607. MacCracken, j. G., D. W. Uresk, and R. M. Hansen. 1985. Habitat use by shrews southeastern Montana. Northwest Science, 59:24-27. Malmquist, M. 1986. Density compensation in allopatrie populations of the pygmy shew Sorex minutus on Gotland and the outer Hebrides: Evidence for the effect of interspecific competition. Oecologia, 68:344-346. Merritt, J. F. 1987. The Mammals of Pennsylvania. The University of Pittsburgh Press, Pittsburgh, Pennsylvania, 408 pp. Owen, J. G. 1984. Sorex fumeus. Mammalian Species, 215:1-8. Pernetta , J. C. 1977. Population ecology of British shrews in grassland. Acta Theriologiea, 22:279-296. 42 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 Porter, J. H., and R. D. Dueser. 1982. Niche overlap and competition in an insular small mammal fauna: A test of the niche overlap hypothesis. Oikos, 39:228-236. Rudd, R. L. 1955. Age, sex, and weight comparisons in three species of shrews. Journal of Mammalogy, 36:323-345. Ryan, J. M. 1986. Dietary overlap in sympatric populations of pygmy shrews, Sorex hoyi, and masked shrews, Sorex cinereus, in Michigan. The Canadian Field-Naturalist, 100:225-228. SAS Institute, Inc. 1985. SAS User’s Guide: Statistics. Fifth Edition. SAS Institute Inc., Cary, North Carolina. Spencer, A. W., and D. Pettus. 1966. Habitat preferences of five sympatric species of long-tailed shrews. Ecology, 47:677-683. Vickery, W. L., and J. R. Bider. 1977. The effect of weather on Sorex cinereus aetivity. Canadian Journal of Zoology, 56:291-297. Whitaker, J. O., Jr., and T. W. French. 1984. Foods of six speeies of sympatric shrews from New Brunswick. Canadian Journal of Zoology, 62:622-626. Williams, D. F., and S. E. Braun. 1983. Comparison of pitfall and conventional traps for sampling small mammal populations. The Journal of Wildlife Management, 47:841-845. Wrigley, R. E., j. E. Dubois, and H. W. R. Copland. 1979. Habitat, abundance, and distribution of six species of shrews in Manitoba. Journal of Mammalogy, 6:505-520. Table 1 . — Number of captures by sex ofSorex cinereus, S. fumeus, and Blarina brevicauda during a three-year live-trapping study in western Pennsylvania. Values in parentheses indicate number of recaptures. Gender Species Males Female Unknown S. cinereus Grid 1 5(3) 15(8) 29(12) Grid 2 2(1) 11(2) 24(6) Total S. cinereus Total 86(32) 7(4) 26(10) 53(18) S. fumeus Grid 1 5(0) 2(1) 8(1) Grid 2 4(0) 0(0) 7(1) Total S. fumeus Total 26(3) 9(0) 2(1) 15(2) B. brevicauda Grid 1 0(0) 0(0) 5(1) Grid 2 2(0) 9(7) 16(3) Total B. brevicauda Total 32(1 1) 2(0) 9(7) 21(4) 1994 CAWTHORN — Population Dynamics of Shrews 43 1985 1986 1987 Fig. 1. — Percent trap success on each grid during each month of the study. Trap success was calculated based on the number of captures/total trap hours for the month. 00 CU 00 0- H O 2 g P Pi o 0- O a ec, 0 shrews abseni low woody vegetation density high woody vegetation density and the presence of pennanent water within 50 m of the trap Fig. 2.— Stations where shrews were captured and those where shrews were not as they occur along the discriminant axis. There was significant separation of these two groups along this axis (Wilks’ Lambda = 0.792; d.f. = 27, F = 1.638, P < 0.03). /■' * efi T. ■' * *** ^wu.4 ^ »< ^.^i *^'ii. '’■v i>^', W.49S,,. . . .^.0 ; ^ ‘ f ■' ‘■'^ ■' ''t* ‘■'^'l- I (tir.,, |...| !'■ ■ '.’'V 'V I ' ' >1 i, •^.'0’ - .'.jj •'-»] ,.i 'v.' w , ■ ,)|' M>t .('i' ‘tphVy' i*i'K V- ’> V)ni;'«40%) was also high in the boreal coniferous (taiga) forest (Fig. 1). Sorex tundrensis was dominant in riparian willow thickets and in floodplain forest with bushes (35.7 % and 25.7 % of shrews, respectively). No other species were dominant and their fraction was always less then 16%. Thus, in anthropogenic grassy habitats S. araneus prevailed; in natural habitats with moss cover, S. caecutiens; and in most typical floodplain habitats, S. tundrensis. In the herbaceous spruce forest, marshy alder thickets, and hillock meadow, S. araneus dominated, but the degree of dominance was less pronounced here due to high species richness (Table 1). To find the ecological valency of species, we calculated the width of the spatial ecological niche according to the formula: a where n is the number of individuals of /th species captured in habitat a and y is the total number of individuals of /th species captured (Levins, 1968). The ecological niche width was calculated separately for each year. Table 2 gives the averaged values for this parameter. On the basis of the calculated spatial niche widths, species can be conditionally divided into two groups: eurytopic (niche width > 10) and stenotopic (niche width < 10). Differences in the ecological niche width between species belonging to different groups in most cases were statistically significant. It is noteworthy that within both eurytopic and stenotopic groups, species exhibited considerable variation in abundance. The ecological distributions of the shrews were plotted in two-dimensional ecological space with the Ramensky scales of moisture and richness as coordinates. The detailed geobotanical descriptions of all 30 capture sites permitted calculation of values for moisture and richness; after that, all 30 points were mapped on the described two-factor space (richness as abscissa, moisture as ordinate). The area delimited by coordinates of soil richness and moisture show ecological habitats occupied by shrews (Fig. 2) Analysis of distribution of shrews revealed that the most abundant species, such as S. araneus, occupied practically all ecological space. Accordingly, I marked only those points where the abundances of such species exceeded mean values over the year. This procedure was used for all nine species of shrews. As an example. Fig. 3 illustrates the distribution of shrews in 1981. There is broad overlap in the contours for different species. This pattern of overlap was evident in the other years. Then for each species, I plotted contours for each of the eight years of observations (1976-1983) and separated their common part (ecological optimum), which contained habitats with stable high preferences by each species, independent of the population dynamics phase, weather conditions, etc. (Fig. 4). If representatives of given species were not observed in a particular year (e.g., S. daphaenodon in 1978) or the number of catches was quite small and animals were caught only in one or two ditches, then I used data for fewer years. That is why the distribution of S. daphaenodon is based on only six years of data and those of S. roboratus and N. fodiens on data for seven years. When the ecological optima for all species were mapped onto common ecological space (Fig. 5), they were found to be largely discrete. Partial overlap was found only between S. araneus and S. minutus and also between 5. caecutiens and S. minutissimus. The majority of species had one highly preferred habitat. Only the smallest species (S. minutissimus) and the 48 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 largest (N. fodiens) had more than one preferred zone, each with two. Habitats supporting maximum annual abundance of shrews are considered to represent the ecological niche space for each species. In Fig. 5, species having one optimum are aligned in the following order: Sorex caecutiens {#2), S. araneus (#1), S. minutus (#3), S. isodon (#5), S. tutidrensis (#6), S. roboratus (#7), and S. daphaenodon (#8). Optima for S. caecutiens, S. araneus, and S. minutus are similar for moisture but show a progressive increase in soil richness. The optima of the remaining species differ little in soil richness but show a progressive increase in moisture from S. isodon to S. daphaenodon (Fig. 5). In the analysis of these species sequences (Fig. 5), an interesting pattern is evident: each pair of species neighboring in the ecological space is very different in size. For example, relatively small S. caecutiens is a neighbor of large S. araneus, which in turn has a slightly overlapping space optimum with very small S. minutus. Then follows large S. isodon, followed by relatively small S. tundrensis. Similar conditions are evident for the large S. roboratus and, at the end of the chain, the small S. daphaenodon. If we compare the length of the upper tooth row for the neighboring species (Table 3), we find that their ratio is at least equal to or somewhat bigger than 1:1.2. This is slightly less than observed by Hutchinson (1959), although shrews with such dimensional differences nevertheless prefer different habitats. A considerable difference in size is observed in the pair S. araneus-S. minutus, and this was the only pair with partial overlap of spatial optima. Discussion In this paper I have attempted to combine the continuum concept of plant community ecology with the theory of ecological niches in animal community studies. The possibilities and prospects of using ordination methods in such investigations are reviewed in detail in Austin (1985). Previous attempts to apply such methods to study the spatial distribution of animals were analyzed in Shvartz and Sheftel (1990), thus I will not use ecological ordination. I will only stress, that, using the parameters of soil moisture and richness as spatial coordinates in studying the ecological distributions of nine shrew species, all of which are terrestrial, yielded better results than have been obtained for other groups, such as birds, whose distributions are strongly influenced by other factors (Kozlenko, 1987). Using ordination methods demonstrates that all observed shrew species have preferred sites in ecological space, which either do not overlap, or overlap only slightly, between different species. Such sites correspond to the center of the spatial ecological niche (Schoener, 1970). The majority of species have one such center. Only S. minutissimus and N. fodiens had two. In the first case this can be explained by the preference of S. minutissimus for ecotonal habitats. This species is quite common at the borders of taiga and swamp, taiga and meadow, etc. As these habitats are quite separated on the ordination map, S. minutissimus has several sites with consistently high density. It is interesting that the smallest North American shrew Sorex (Microsorex) hoyi also prefers edge habitats (Spencer and Pettus, 1965). For N. fodiens, the two preferred habitats are bottomland rich meadow and weedy places in former vegetable gardens. For these two places the common factors are the presence of long-stemmed grasses and the proximity of the Yenisei River. Distance from water is likely an important factor for location of preferred N. fodiens habitat. Species having one optimum are aligned practically along one line in the following order from poor, dry soils to moist, rich soils: S. caecutiens, S. araneus, S. minutus, S. isodon, S. tundrensis, S. roboratus, S. daphaenodon. The ecological optima for the first three species are situated at about the same conditions of moisture along the increasing soil richness axis. The others prefer sites with relatively high soil richness but differ in terms of increasing moisture. Species whose centers of optima are at low soil richness exhibit maximum width of the ecological niche space, while those who prefer rich, moist biotopes have minimal width of ecological niche space (Table 2). For this reason inhabitants of poor habitats more often can be detected in rich habitats than vice versa. In studying the spatial distribution of shrews, the majority of authors have arrived at the conclusion that preferred habitats differ among species. If body size of one species is larger than that of another, then it is supposed that the larger species forces the smaller into poorer habitat (Dickman, 1988). In the case of similar-sized species, e.g., S. araneus and S. coronatus (Neet and Hausser, 1990) or S. vagrans and S. obscurus [=monticolusJ (Hawes, 1977) their ecological distributions are believed to be determined by competition. Neet and Hausser (1990) support this statement by their experiments with species removal. On the other hand, Hutchinson (1957) concludes that each species must have a refugium, where it is a favored competitor in relation to other species. In these studies only pairs of species were considered. In reality shrew communities are often composed of several species. Those investigators who worked with such communities (Getz, 1961; Brown, 1967; Spencer and Pettus, 1965) found that all species are mainly separated in space. Wrigley et al. (1979) noted that shrew species of the same community differ in size. Among species of similar size, abundance of one species exceeds other species abundances considerably. Churchfield (1984), Whitaker and French (1984), and Shvartz and Demin (1986) studied the feeding habits of different species in multispecies Sorex communities, and found that the differences in animal sizes are quite important for division of food resources, an essential condition for species coexistence. In the present work I have arrived at the conclusion that, in the case of mutual coexistence of congeneric species of Sorex, on the one hand each species has a preferred habitat that is unique for this species; on the other hand neighboring habitats are occupied by Sorex species of different body size. The question of the role of interspecies competition in determination of such ecological distributions remains open. To solve it I analyzed published data on the ecological distribution of shrews in Siberia (Yudin, 1980; Yudin et al., 1976; Yudin et al., 1979; 1994 SHEFTEL — Distribution of Nine Species of Shrews in Siberia 49 Shvetzov, 1977; Sheftel, 1983). In general, the ecological distributions of species in the present study agree quite well with published data. As a rule, exceptions are noted in northern communities with low species richness. Thus, S. roboratus in the northern Yenisei River taiga in the absence of S. araneus occupies small-leaved and coniferous forests with developed herbaceous cover, while it normally did not occupy these habitats in the middle Yenisei River taiga region in the presence of S. araneus (Sheftel, 1983). In the same way, S. tundrensis in the northern forest and southern tundra in the absence of S. caecutiens occupied wet, mossy habitats (Yudin, 1980). Moraleva (1987) observed that, with increasing abundance of S. caecutiens (in years of peak abundance), S. tundrensis was forced out from habitats where this species was common in the prepeak year. During peak population growth of the dominant species S. araneus and S. caecutiens, the density of species with a narrow-space ecological niche (e.g., S. roboratus, S. tundrensis, S. daphaenodon) decreased. Conclusions Multispecies communities of shrews include coexisting species of both similar and different sizes. Each species has its own spatial optimum (center of ecological niche space) which is nearly always separated from that of other species. Neighboring habitats are occupied by species with different body sizes. Other studies have shown that species of small body size are forced by larger species into poorer habitats, where particular dimensional features of food objects enhance the survival of the larger species (Demin and Glazov, 1990). When species with similar body sizes occur in the same area, this can lead to parapatry with a narrow contact zone and pronounced segregation between habitats (Neet and Hausser, 1990; Hawes, 1977). In the case where ranges of similar-sized species overlap significantly, one species can live only in a suppressed state in the most productive habitats, which represents a kind of ecological refiigium. Acknowledgments I express my gratitude to N. V. Moraleva, E. M. Grigorjev, M. A. Zhukov, A. Yu. Alexandrova, and D. Yu. Alexandrov, who took active part in collecting material used in this work; to O. V. Boursky, who participated in the discussions of the results; and to Prof. E. E. Syroechkovsky, who supported these investigations. Drs. R. S. Hoffmarm, W. E. Wrigley, and G. L. Kirkland, Jr., provided valuable assistance with revision of the manuscript. Literature Cited Austin, M. P. 1985. Continuum concept, ordination methods and niche theory. Annual Review of Ecology and Systematics, 16:39-61. Belgard, a. L. 1950. Forest vegetation of south-east Ukrainian SSR. Kiev, 264 pp. (in Russian). Brown, L. N. 1967. Ecological distribution of six species of shrews and comparison of sampling methods in the central Rocky Mountains. Journal of Mammalogy, 48:617-623. Churchfield, S. 1984. Dietary separation in three species of shrews inhabiting water-cress beds. Journal of Zoology (London), 204:211-228. Demin, D. V., and M. V. Glazov. 1990. The dependence of shrew community structure on soil vertebrate community structure. Pp. 272-273, in 5 sjezd Vsesouznogo Teriologicheskogo obschestva Akademii nauk USSR (29 Jan. -2 Feb. 1990, Moscow), v. 2, Moscow, 320 pp. (in Russian). Dickman, C. R. 1988. Body size, prey size and community structure in insectivorous mammals. Ecology, 69:569-580. Ellenberg, H. 1952. Weisen und Weiden und ihre standortliche Bevertung. Landwirtschaftliche Pflanzencociologie II Stuttgart, 352 pp. Getz, L. L. 1961. Factors influencing the local distribution of shrews. The American Midland Naturalist, 65:67-68. Hawes, M. L. 1977. Home range, territoriality and ecological separation in sympatric shrews. Journal of Mammalogy, 58:354-367. Hutchinson, G. E. 1957. Concluding remarks. Cold Spring Harbor Symposium on Quantitative Biology, 22:415-427. 1959. Homage to Santa Rosalia, or why are there so many kinds of animals? The American Naturalist, 93:137-145. Kozlenko, a. B. 1987. Ecological estimation of distribution of birds over taiga habitats (on the basis of Ramensky scales). Pp. 142-150, in Fauna i ekologiya ptits i mlekopitauschih Centralnoi Sibiri (E. E. Syroechkovsky, ed.), Nauka, Moscow, 231 pp. (in Russian). Levins, R. 1968. Evolution in changing environments. Princeton University Press, Princeton, New Jersey, 120 pp. Moraleva, N. V. 1987. On the question of interspecies relations of neighboring shrew species (Insectivora, Sorex). Pp. 213-228, in Fauna i ekologia ptits i mlekopitauschih Centralnoi Sibiri (E. E. Syroechkovsky, ed.), Nauka, Moscow, 231 pp. (in Russian). Mras, K., and V. Samek. 1966. Lesny rostliny. Praha, 246 pp. Neet, C. R., and J. Hausser. 1990. Habitat selection in zones of parapatric contact between the common shrew Sorex araneus and Millet’s shrew S. coronatus. Journal of Animal Ecology, 59:235-250. Ramensky, L. C. 1925. The main interrelation on vegetation cover and their study. Voronezh, 37 pp. (in Russian). Ramensky, L. C., 1. A. Tsatsenkin, O. P. Chizhicov, and N. A. Antipin. 1956. Ecological estimation of pasture vegetation cover. Selhozgiz, Moscow, 472 pp. (in Russian). Schoener, T. W. 1970. Nonsynchronous spatial overlap of lizards in patchy habitats. Ecology, 54:408-418. Sheftel, B. 1. 1983. Zonal features of an insectivorous community in the Yenisei taiga and forest-tundra regions. Pp. 184-203, in Zhivotnyi mir Yeniseiskoi taigi i lesotundry i prirodnaya zonalnost (E. E. Syroechkovsky, ed), Nauka, Moscow, 233 pp. (in Russian). 1989. Long-term and season dynamics of shrews in central Siberia. Annales Zoologici Fennici, 26:357-369. Shvarts, E. a. 1989. Formation of small mammal and insectivorous fauna in the taiga of Euro-Asia. Pp. 1 15-143, in Fauna i ekologia gryzunov (V. V. Kucheruk, ed.), Moscow University Press, 17:1-223 (in Russian). Shvarts, E. A., and D. V. Demin. 1986. Factors in the coexistence of congeneric species in location within their distributional areas of sympatry (as exemplified by Soricidae). Doklady of Academy of Science USSR, Biological Science Section (English translation), Doklady Academii Nauk USSR, 289:412-415. Shvarts, E. A., and B. 1. Sheftel. 1990. Ecological ordination in geozoological investigation . Pp. 3-17, in Ecological Ordination and Community (G. Dlussky, ed.), Nauka, Moscow, 196 pp. (in Russian). Shvetsov, Yu. G. 1977. Small mammals of Baikal hollow. Nauka, 50 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 Novosibirsk, 158 pp. (in Russian). Spencer, A. W., and D. Pettus. 1965. Habitat preference of five sympatric species of long-tailed shrews. Ecology, 47:677-683. Tupkova, N. V., V. A. Zaklinskaja, and V. S. Evseeva. 1961. Drift fences as method of estimation of abundance and massive capture of small mammals. Pp. 78-80, in Voprosy organizacii i metody ucheta resursov fauny nazemnyh pozvonochnyh (Yu. A. Isakov, ed.), Moscow, 152 pp. (in Russian). VOROBJEV, D. V. 1959. Methods of research of forest types. Kharkov, 144 pp. (in Russian). Whitaker, J. O., Jr., and T. W. French. 1984. Foods of six species of sympatric shrews from New Brunswick. Canadian Journal of Zoology, 62:622-626. Whittaker, R. H. 1967. Gradient analysis of vegetation. Biological Review, 42:207-264. Wrigley, R. E., J. E. Dubois, and H. W. R. Copland. 1979. Habitat, abundance and distribution of six species of shrews in Manitoba. Journal of Mammalogy, 60:505-520. Yudin, B. S. 1980. Zonal and landscape group of small mammals (Micromammalia) on the Taimyr Peninsula. Pp. 5-31, in Trudy Biologicheskogo instituta Sibirskogo otdeleniya Akademii nauk USSR (B. S. Yudin, ed.), Nauka, Novosibirsk, 44:1-456 (in Russian). Yudin, B. S., V. G. Krlvosheev, and V. G. Belyaev. 1976. Small mammals of the northern part of the Far East. Nauka, Novosibirsk, 272 pp. (in Russian). Yudin, B. S., L. I. Galkina, and A. F. Potapkina. 1979. Mammals of Altai-Sayan mountain country. Nauka, Novosibirsk, 296 pp. (in Russian). Table 1 . — Distribution of shrew species in 1 1 different habitats during one abundance cycle 1978-1981 (individuals per 100 pitfall- nights). Species Habitats Sorex araneus Sorex caecutiens Sorex minutiis Sorex miniitissimus Sorex isodon Sorex tundrensis Sorex Sorex roboratus daphaenodon Neomys fodiens Total Bog forest 8.8 19.8 0.8 0.6 1.2 0.5 0.2 0.1 32.0 Boreal coniferous forest 34.1 29.4 1.6 0.7 1.1 0.9 0.8 0.1 0.1 68.8 Grass-moss small-leafed forest 43.7 29.8 5.6 0.8 1.7 1.5 1.5 0.1 0.1 84.8 Border of taiga forest 65.1 18.1 5.1 0.4 2.9 1.5 0.2 0.3 93.6 Dry meadow 21.3 7.6 2.0 0.5 1.1 2.3 0.1 — 0.2 35.1 Herbaceous riparian spruce forest 49.4 36.1 8.2 0.4 3.5 2.5 2.0 0.3 0.4 102.8 Riparian willow thickets 20.2 7.9 4.6 0.6 4.0 23.8 4.0 0.3 1.3 66.7 Floodplain forest with bushes 15.5 14.8 4.5 0.5 4.4 18.1 10.6 1.0 0.9 70.3 Seasonally flooded hillock meadow 14.7 11.8 5.7 0.4 4.5 7.0 5.0 1.1 0.7 50.9 Riparian high grass meadow 42.0 12.5 12.1 0.5 2.9 7.3 — — 2.3 79.6 Marshy alder thickets 48.2 24.9 16.9 1.0 13.4 5.3 — — 1.7 111.4 Table 2. — Size of spatial ecological niche and average annual shrew abundance. Data represent an average for an eight-year period. Species Size of Spatial Niche Mean Relative Abundance Sorex araneus 20.7 ± 1.18 33.7 Sorex caecutiens 19.4 ± 1.64 22.0 Sorex minutus 14.6 ± 1.25 4.2 Sorex minutissimus 10.4 + 1.50 0.4 Sorex isodon 13.7 + 1.00 3.1 Sorex tundrensis 8.3 + 0.89 3.8 Sorex roboratus 4.9 ± 0.89 1.1 Sorex daphaenodon 3.7 ± 0.64 0.3 Neomys fodiens 5.0 ± 0.82 0.4 1994 SHEFTEL— Distribution of Nine Species of Shrews in Siberia 51 Table 3. — Relation between the length of upper intermediate teeth in Sorex species (40 specimens for each species). Species Length of Upper Teeth (mm) Ratio of Length araneus / caecutiens 2.82 ± 0.018 / 2.38 ± 0.018 1.18 ± 0.012 araneus / minutus 2.82 + 0.018 / 1.93 ± 0.018 1.43 + 0.016 isodon / minutus 2.85 ± 0.022 / 1.93 ± 0.018 1.48 ± 0.018 isodon / tundrensis 2.85 ± 0.022 / 2.30 ± 0.024 1.24 + 0.016 roboratus / tundrensis 2.70 ± 0.022 / 2.30 ± 0.024 1.17 + 0.055 roboratus / daphaenodon 2.70 ± 0.022 / 2.23 ± 0.018 1.21 ± 0.014 // / odien t |lliP!!!i! d.faphden'jdart S.t-oborcdus I d . iuhdt-enu: d isoa'a/! SHi . mmuT/ ;;itvu; WIlFlIh f. hiinutw; d. caGCutien? 0 Qt^QJl9u! HABITAT TYPE Fig. 1.— Relative abundance (%) of shrew species in the different habitats. 1, bog forest; 2, boreal coniferous forest; 3, grass-moss small-leaved forest; 4, border of taiga forest; 5, dry anthropogenic meadow; 6, herbaceous riparian spruce forest; 7, riparian willow thicket; 8, floodplain forest with bushes; 9, seasonally flooded hillock meadow; 10, riparian high grass meadow; 11, alder thickets. 52 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 SOIL RICHNESS o^irch meadow ^wamp pine, wi llo'j/ l^prucS. ^ i oeriart pine Li'rch alder Fig. 2. — Observed ecological space with coordinates of soil richness and moisture according to Ramensky et al. (1956). (The order number of specific habitat is the same as in Fig. 1). 1994 SHEFTEL — Distribution of Nine Species of Shrews in Siberia 53 aanisioiAi nos rn ci) iZ Map of the ecological distribution of nine species of shrews in 1980. 54 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 Optimal ecological conditions of soil moisture and soil richness for Sorex minutus for the years 1976-1979. 1994 SHEFTEL — Distribution of Nine Species of Shrews in Siberia 55 Fig. 5. — Map of ecological distribution of nine species of shrews in two-dimensional ecological space. 1, Sorex araneus\ 1, S. caecutiens; 3, S. minutus; 4, S. minutissimus\ 5, S. isodon\ 6, S. tundrensis; 7, S. roboratus; 8, S. daphaenodorv, 9, Neornys fodiens. COMMUNITY ORGANIZATION OF SHREWS IN TEMPERATE ZONE FORESTS OF NORTHWESTERN RUSSIA Evgeny A. Shv arts' and Dmitry V. Demin“ 'institute of Geography, Russian Academy of Sciences, Moscow, Russia ^Institute of Developmental Biology, Russian Academy of Sciences, Moscow, Russia Abstract The ecological distributions of six species of Sorex were studied during 12 years in the Novgorod region of northwestern Russia. Data on the biotopic distribution and estimated pattern of spatial association in multispecies communities of shrews were analyzed using cluster analysis. Results concerning the coexistence of closely-related species and the permissible limits of size similarities for such species are consistent with studies of Cause and Hutchinson. Species that were most similar in spatial distribution differed substantially in body size. Analysis of soil invertebrates revealed a relationship between population structure of soil invertebrates and biotopic distribution of shrews. Introduction Shrews (Soricidae) are small insectivorous mammals which often occur in assemblages of five or more species (Kirkland, 1991). Within the study area in northwestern Russia, as well as in regions of Siberia and the Far East, soricid species are rather uniform in morphology and ecology, which theoretically should hinder their coexistence. However, the species diversity of soricid assemblages usually is substantially higher than that of associated groups of small rodents (Sheftel, 1994). Most shrew species share a similar mode of life; they forage for invertebrates in the upper soil layers and forest litter. Our attention therefore was focused on the food resources of shrews as a basic factor determining the structural organization of multispecies communities. Materials and Methods Data on the biotopic distribution of six species of Sorex (Table 1) were collected from 1975 to 1987 at a research station near Valdai (57°59’N, 33°10’E), Novgorod region, Russia. Specimens were caught with snap traps and in ditches (20 m in length). Trapping was concentrated in three periods: May-June, July, and August-September. Sampling effort totalled 37,632 trapdays and 1,259 ditchdays. Geobotanical characteristics of the vegetation cover (ca 2,300 descriptions) of almost all traplines and ditches were recorded using the succession classification system of Razumovsky (1981) for the Moscow district. Razumovsky’s classification scheme for describing native vegetation cover is a system of communities defined by floristic criteria (similar to the Braun-Blanquet method) and the ordination in space of moisture and soil richness factors. We used indicator species of plants for classifying plant communities in accordance with Razumovsky’s system; however, we adapted this scheme slightly to meet the specific conditions of the study area and to facilitate data processing. Soil invertebrates were collected near the trap stations. Samples of forest litter and soil layers were taken from 25 X 25 cm squares. Ten to 12 samples were taken in each habitat. Three hundred twenty samples yielded ca 10,000 invertebrates. which were scored, weighed, and measured in the laboratory. Eleven size classes of invertebrates were defined on the basis of 2-mm size intervals. The median size class was calculated for all habitats. Skull morphology was examined in six species of soricids: Neomys fodiens {n = 20), Sorex araneus (n = 69), S. caecutiens (n = 39), S. isodon (n = 16), S. minutus (n = 45), S. minutissimus {n — 31). Only animals bom in the same year were used, and the sexes were pooled. The following measurements were taken with an ocular micrometer (magnification 16X) to nearest 0.05 mm: length of mandibular symphysis (Fig. la), length of “arm” of mandible (from tip of Ij to anterior edge of coronoid process. Fig. lb), tip of first incisor to posterior tip of angular process of mandible (Fig. Ic), distance between anterior edge of Mj and lower condylar facet (Fig. Id), length of lower toothrow (Fig. le), and length of upper toothrow. All measurements were taken by the second author. Means and standard deviations were calculated. Cluster analysis was used to present data on the spatial distribution of shrews. A value equal to 1.0 minus the correlation coefficient for the pairs of species was taken as a measure of the coupling. This measure helped to compare the ecological distributions of shrew species and did not depend on their abundance. Results Abundance and Size Distribution of Soil Invertebrates Samples of soil invertebrates tended to be dominated by smaller size classes ( <4 nun; Fig. 2). The dominance of small forms was pronounced in habitats with low primary productivity (Fig. 2 a,b) such as pine forest on peat-moss bogs (oligotrophic marshes) and dry spruce forest (Fig. 2 a,b). Larger size classes were better represented in richer habitats such as taiga forests (Fig. 2 c,d). The biomass of invertebrates approximated a normal distribution with very small and very large types constituting only a minor portion of the entire biomass. The distribution of biomass among size classes was more balanced in low productivity habitats (Fig. 3 a,b). In more productive plant 57 58 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 communities, the distribution of invertebrate biomass featured a peak of medium-sized items (Table 2, Fig. 3 c,d). Biotopic Distribution of Shrews Cluster analysis of snap trap data grouped four shrew species having similar ecological distributions (Neomys fodiens, Sorex isodon, S. minutus, S. araneus. Fig. 4a). These are listed in order of decreasing moisture of their preferred habitats. These species preferred rich, humid habitats dominated by grassy plant associations such as floodplains and drained lowland bogs, or European broad-leaved and coniferous-broad-leaved forests. Sorex isodon and S. minutus, which were most similar in their biotopic distribution, differ considerably in size (Fig. 5). The distributions of S. caecutiens and S. minutissimus differed from this group. They preferred typical taiga forests; however, the small sample size of S. minutissimus (Table 1) limited comparisons with other species. Cluster analysis of the data obtained from ditches showed very similar results (Fig. 4b). Three clusters are recognizable: two species of floodplain habitats {N. fodiens, S. isodon), two species of European broad-leaved and coniferous-broad-leaved forests (5. araneus, S. minutus), and two species of typical taiga (S. caecutiens, S. minutissimus). The correlation coefficients for the first two species pairs were 0.89 and 0.69, respectively, and differed from zero (P < 0.003 and P < 0.07). In the second cluster, S. minutus grouped closer to S. araneus than to S. isodon. This may be due to the lesser resolution of captures in ditches and to the complicated conditions of the narrow floodplain of the Valdaika River, resulting in a combination of the data obtained close to the water and in the moister plant associations. Similar results were also obtained when we analyzed the combined data from snap traps and ditches of Putchkovsky (1969fl). There was negative correlation between the pairs of typical taiga shrews and shrews inhabiting richer biotopes. The distribution of S. araneus differed substantially from both taiga species and species inhabiting humid and rich habitats (Fig. 4c). To conclude, the cluster analysis of biotopic distribution of shrews identified two groups of biotopically similar species: one group inhabiting rich, humid, grassy associations and European broad-leaved and mixed coniferous-broad-leaved forests (European faunal element with S. araneus, S. minutus, N. fodiens, and an eastern element, S. isodon [Shvarts, 1989]), and the other group inhabiting the taiga-type habitats {S. caecutiens, S. minutissimus, both representatives of the eastern element). Species with the most similar biotopic distributions were characterized by considerable difference in body size. Morphometric Analysis of Feeding Apparatus of Coexisting Species of Shrews The divergence of coexisting species in size of preferred food items is apparently related to differences in size of parts of the feeding apparatus (Hutchinson, 1959). Simberloff and Boecklen (1981) raised doubts regarding the significance of Hutchinsonian ratios; however, subsequent statistical analysis by Losos et al. (1989) lend support to the significance of these interspecific size ratios. The ranges of ratios for five dimensions of the feeding apparatus (Fig. 1, b-e and length of upper toothrow) for pairs of species within the same faunal elements were: N. fodiens IS. araneus = 1. 1-1.24; S. araneus IS. minutus = 1.32-1.39; S. isodonIS. caecutiens = 1.18-1.21; S. caecutiensIS. minutissimus = 1.16-1.29 (Shvarts and Demin, 1986). The ratios for pairs of species from different faunal elements were markedly smaller: S. isodonIS. araneus = 0.99-1.04; S. minutus! S. minutissimus = 1.04-1.12. The differences between species in pairs, S. isodon-S. araneus and S. minutus-S. minutissimus, for all jaw characters were not significant (P > 0. 1). The coexistence of such species within individual habitats should be limited in accordance with ideas of Hutchinson (1959). In the larger shrews {N. fodiens, S. isodon, S. araneus), the dimensional ratios were smaller than in the smaller species. This is, perhaps, connected with the fact that in case of shrews feeding on larger items, it is not the increase in the absolute length of jaws that is of primary importance but the increase in the power of the masticatory musculature. This is amply confirmed by the higher values for the ratios of dimensions of symphysis in the large shrews by comparison with the small species: N. fodiens IS. isodon = 1.3; S. araneus IS. caecutiens - 1.3; S. caecutiensIS. minutus = 1.14; S. minutusIS. minutissimus = 0.97. The analysis of the skull measurements showed also that the maximum size differences were between coexisting species, particularly between S. isodon and S. minutus, and S. caecutiens and S. minutissimus (Fig. 5). Discussion We compared the results of the morphometric analysis with data on density of individual shrew species in different habitats. The abundant species were S. isodon and S. minutissimus, two members of the Asian faunal element whose existence in multi species soricid communities may have been influenced by competition with similar-sized species of the European faunal element (e.g., S. araneus and S. minutus, respectively). Of the three most numerous species of shrews, S. araneus and S. minutus are members of the European faunal element and differ considerably in size (Fig. 5). This size difference apparently is sufficient for S. caecutiens, a member of the Asian faunal element, to coexist. Average ratios of jaw elements of the three species were S. araneus! S. caecutiens = 1.16-1.21, and S. caecutiensIS. minutus = 1. 12-1. 18. These are on the low end of “acceptable” Hutchinsonian ratios (Hutchinson, 1959). It is not known whether the size interval between S. araneus and S. minutus existed prior to the invasion of S. caecutiens or whether it represents an evolutionary response as has been demonstrated for Pleistocene Sylvaemus (= Apod emus) of the Near East (Tchemov, 1979). Based on the numbers of individuals collected, the data from our study suggest that S. caecutiens is more successful as a member of the Valdai Hills shrew community that either of the other two members of the Asian faunal element, S. isodon and S. minutissimus (Table I). 1994 SHVARTS AND DEMIN — Temperate Forest Shrew Communities 59 The most closely associated species biotopically were S. araneus and S. minutus, which differ considerably in size. These two species of the European faunal element were clearly associated with nemoral or broad-leaved deciduous forest habitats. Unlike S. minutus, S. caecutiens was associated with taiga habitat and clearly preferred sites with continuous moss cover. Sorex araneus was considerably less abundant in such habitats. Despite the small number of S. minutissimus trapped, the species appeared to be closely associated with S. caecutiens (Fig. 4b). The spatial distributions of these two species in the Valdai Hills was thus similar to that reported by Putchkovsky (1969fl) for these species in the taiga of the Onega Peninsula, approximately 770 km NE. Sorex isodon was the most stenotopic shrew in the Valdai Hills. Its distribution was clearly restricted to humid, rich floral communities in grass bogs and meadow stages of the floodplains. Being a bit larger than S. araneus, S. isodon is adapted to coexistence with S. minutus and both had their highest densities in floodplain habitats. In regions of extensive floodplain habitats, S. isodon, which is usually rare in Europe, is numerically superior to S. araneus (Putchkovsky, 1969/?). The data suggest that optimal habitats for S. isodon are humid and rich communities with well-developed grass cover, whereas S. araneus prefers drier broad-leaved deciduous (nemoral) forest communities (Sheftel, 1994). Neomys fodiens is semiaquatic, inhabiting floodplains and seldom occurring elsewhere, except when displaced by rising flood waters. About 50% of the food items consumed by N. fodiens are aquatic organisms, which are less available to other species (Churchfield, 1984). The biomass of aquatic invertebrates must not be great or opportunities for water shrews to forage in aquatic habitats must be limited in the study area. How else can one explain that small number of N. fodiens in the floodplain? According to our data, taking into account all three study periods, there were 0.4/5. 2/4. 4 animals per ten ditches/days for N. fodiens and 0.7/9.8/11.2 for S. araneus. The absence of a satisfactory explanation for this case has allowed some authors to consider as peculiar the fact that the water shrew is not the most numerous shrew species in habitats adjacent to lakes (Churchfield, 1984). The results of our analysis of organization of the shrew community agree with those of Sheftel (1994). It is important to note the fact that his work established not only the same patterns of organization in a nine-species shrew community, but also similar species optima in ecological space represented by soil humidity and soil richness. The results of our morphometric and ecological analyses did not reveal the reasons for the clear numerical superiority of S. araneus over other species in our study (Table 1). In addition, the causes of changes in the role of one or other species in different habitats should be analyzed in more detail. Frequently, the abundance of species can differ substantially in adjacent habitats. For example, in taiga forests of the European part of Russia, S. araneus frequently is replaced by S. caecutiens. We asked, what are the factors that determine the numerical domination of one or other species of shrews? In attempting to answer this question, we analyzed the numerical and biomass distributions of size classes of invertebrates in four habitats. We compared habitats on the basis of the median size class for biomass and number of invertebrates. There was a shift in the main biomass and number of invertebrates from smaller to larger size classes with increasing soil richness (Fig. 2, 3). Over the range of the poorest to the richest habitats there was a shift towards the larger items (Table 2). The distribution of shrew species by habitat mirrored shifts in resource abundance. The proportion of smaller-sized species of shrews was greater in poorer habitats, whereas larger species were dominant in richer habitats with more abundant prey items (Table 2). This may explain the phenomenon of different species of shrews being numerically dominant in adjacent habitats. Two species {S. araneus and S. caecutiens) tend to dominate soricid communities in the region of the Valdai Station and in adjacent regions of European Russia. In this study, their numerical relationship is changed as invertebrate resources increased in biomass and shifted towards larger-sized prey. Thus the ratio between the numbers of S. araneus and S. caecutiens was 0.77 in dry pine forest, 2.28 in dry Picea abies forests, 3.65 in temporarily wet Picea forests, and 16.08 in forests of the river floodplains (Table 2). Sorex araneus was numerically dominant in the richest and medium-rich habitats but was substantially less abundant and comprised a smaller portion of the shrew community in poorer areas, where S. caecutiens frequently was the numerically dominant species (Table 2). Our data suggest that characteristics of the food resource base (number and biomass of different-sized invertebrates) govern the numerical superiority of individual shrew species. A Model of Organization of Multispecies Community of Shrews A relationship between abundance and body size of soil invertebrates was shown by Gilarov (1944) (Fig. 6a). Furthermore, Tseitlin (1985) demonstrated that the slope of the regression line was very similar for soil invertebrates of the tundra, coniferous forests, oak woods, and other zonal habitats. We may therefore assume that the size distribution of invertebrates shows a similar pattern in all north temperate habitats. If we propose (Fig. 6a) that Ig N = b — a Ig L, and the biomass of each size class B = k N (where L is the linear size of invertebrates; B its biomass; b, free term in the equation; and a and k, proportionate factors), then Ig B = Ig k L'* N = 3 Ig L -I- Ig k — Ig N, or Ig B = Ig N + 3/a(b — Ig N) -I- Ig k = 3b/a -I- Ig k + [1 — (3/a)]lg N. The dependence of the individual mass distribution on the size of the objects is shown in Fig. 6b. On the basis of graphs and equations, the relationship between the biomass and body size of invertebrates was figured (Fig. 6c). As mentioned earlier, the highly abundant small-sized invertebrates have a very small total biomass. Large-sized invertebrates occur in low density but have a high individual mass. While a small-sized prey has a low energy value, a large- sized prey is probably rarely encountered as a result of its low abundance. As can be seen from the relationship between total 60 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 biomass and body size of invertebrates, the most economical strategy for predators such as shrews should be feeding on medium-sized invertebrates. However, if the same food source allows more than one species of shrews to coexist in the same biotope, then specialization of different species on different portions of the food resource spectrum must be expected. As the total biomass of both very small and very large invertebrates is relatively low, the number of predator species specialized on these prey groups will be also lower than the number of those feeding on abundant medium-sized prey. The model allows some generalizations concerning the formation of shrew community and the rules of its organization. 1 . The spatial distribution of shrews and their average annual population levels should be the result of a coevolutionary process within a community of several species of different sizes. The amount of competition between species is a function of overlap in the size of preferred food items, which is apparently closely correlated with jaw size (Pianka, 1978). Similar-sized members of the community should therefore occupy different habitats. But different-sized members of a community may coexist due to divergence in different size classes of food items. 2. In shrew communities with three or more species of shrews, certain species will be mainly adapted to feeding on invertebrates of the medium-sized classes. The greater abundance of such species may produce the impression of ecological plasticity in these shrews. In the case of two syntopic species, the one with a prey-size distribution corresponding to the main invertebrate biomass reserves will be dominant in number. 3. In the case of a secondary contact between species originating from different faunal provinces, those species for which no competitors of corresponding size are present should be the most successful. Acknowledgments We are grateful to N. Chernyshev, M. Glazov, I. Popov, and A. Tishkov for the providing unpublished data on the ecology of shrews and soil invertebrates. We thank B. Sheftel, K. Rogovin, G. Shenbrot, and E. Ivanitskaya for useful discussion and comments during the preparation of this article. We are particularly grateful to D. Zamolodchikov for his help with the statistical treatment of our data. We especially wish to thank G. L. Kirkland, Jr., R. S. Hoffmann, and two anonymous reviewers for extensive and helpful editing of the manuscript. Literature Cited Churchfield, S. 1984. Dietary separation in three species of shrews inhabiting water-cress beds. Journal of Zoology (London), 204:211-228. Gilarov, M. S. 1944. The relationship between sizes and number of soil animals. Doklady Academii Nauk, 43:283-285 (in Russian). Hutchinson, G. H. 1959. Homage to Santa Rosalia, or why are there so many kinds of animals. The American Naturalist, 93:145-159. Kirkland, G. L., Jr. 1991. Competition and coexistence in shrews (Insectivora: Soricidae). Pp. 15-22, in The Biology of the Soricidae. (J. S. Findley and T. L. Yates, eds.). Special Publication of the Museum of Southwestern Biology, University of New Mexico, 1:1-91. Loses, J. B., S. Naeem, and R. K. Colwell. 1989. Hutchinsonian ratios and statistical power. Evolution, 43:1820-1826. Pianka, E. R. 1978. Evolutionary Ecology. 2nd ed. Harper and Row, New York, 397 pp. Putchkovsky, S. V. 1969a The pattern of shrews distribution (Insectivora, Soricidae) on the biotopes of the Onega Peninsula taiga. Pp. 100-109, in Fauna, Ecology, and Geography of Animals (S. P. Naumov, ed.). Research Notes of Moscow Teachers Institute, Moscow. 362:1-335 (in Russian). \969b. Migrations and structure of shrews’ population (Soricidae, Insectivora) in the taiga of the Onega Peninsula. Zoologitcheskii Zhumal, 48:1544-1551 (in Russian). RazuMOVSKY, S. M. 1981. Regularities of biocenose dynamics. Nauka, Moscow, 231 pp. (in Russian). Sheftel, B. I. 1994. Spatial distribution of nine species of shrews in the central Siberian taiga. Pp. 45-55, in Advances in the Biology of Shrews (J. F. Merritt, G. L. Kirkland, Jr., and R. K. Rose, eds.), Carnegie Museum of Natural History Special Publication No. 18, x -F 458 pp. Shvarts, E. a. 1989. The fauna formation of small rodents and insectivores in the taiga of Eurasia. Pp. 1 15-143, in Fauna and Ecology of the Rodents. University Press, Moscow, 17:1-223 (in Russian). Shvarts, E. A., and D. V. Demin. 1986. Factors in the coexistence of related species in loeations within their distribution areas where they are sympatric (as exemplified by the Soricidae). Doklady Biological Sciences (English translation from Doklady Academii Nauk), 289:412-415. SIMBERLOFF, D., AND W. BOEKLEN. 1981. SanU Rosalia reconsidered: Size ratios and competition. Evolution, 35:1206-1226. Tchernov, E. 1979. Polymorphism, size trends, and Pleistocene paleoclimatic response of the subgenus Sylvaemus (Mammalia: Rodentia) in Israel. Israel Journal of Zoology, 28:131-159. Tseitlin, V. B. 1985. Distribution of organism by size in different ecosystems. Doklady Biological Sciences (English translation from Doklady Academii Nauk), 285:693-696. 61 1994 SHVARTS AND DEMIN— Temperate Forest Shrew Communities Table 1 . — Number of shrews caught in different habitats near the research station of the Institute of Geography, Russian Academy of Sciences. Species Captured in Snap Traps Captured in Ditches Total Sorex araneus 1,060 632 1,692 Sorex caecutiens 83 258 341 Sorex minutus 112 143 255 Sorex isodon 10 43 53 Sorex minutissimus 3 6 9 Neomys fodiens 17 103 120 Table 2. — Relationship in structure of the population of soil mesofauna and shrews. Small species are Sorex minutus, a/ki S. minutissimus,’ average number data indicate the number of mammals per ten pitfall-days. caecutiens, S. Type of Habitat Biomass Median (mm) Number Median (mm) Share of Small Species (%) Average Number S. caecutiens S. araneus Pinus sylvestris- Vaccinium vitis-idaea-Pleurozium schreberi 8.8 2.4 60.8 2.00 1.53 Picea abies-Pleurozi um schreberi 9.9 2.6 56.9 1.03 2.35 Picea abies-Oxalis acetosella 10.4 4.9 42.4 0.66 2.53 Picea abies (+Alnus incana)-Geum ri vale + FilipetuJula ulmaria 11.3 5.3 16.5 0.25 4.02 Fig. 1. — Measurements of the lower jaw of a shrew used for morphological analysis, a, length of lower jaw symphysis; b, length of the arm of the lower jaw (from the top of the incisor of the lower jaw to the anterior edge of the coronoid process); c, tip of first incisor to posterior tip of angular process of the mandible; d, distance between the front border of the molars of the lower jaw and the lower tip of the coronoid process; e, length of the lower toothrow. 62 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY O o o S 2 ® ® 1IOS 2 lAl / SN3l^l03dS 30 !33aiMnN E 03 i OS z ! > i Q o m I'c ni0S2l^/SN31AII03dS 30 UdaiAJON o NO. 18 Fig. 2. — Histograms of number/size class of soil invertebrates in the plant associations studies: a, Pinus sylvestris. Sphagnum fuscum\ b, Pinus sylvestris, Vaccinium vitis-idaea, Pleurozium schreberi; c, Picea abies, Oxalis acetosella (taiga type); d, Picea abies + Alnus incana, Geum rivale + Filipendula ulmaria. 1994 SHVARTS AND DEMIN — Temperate Forest Shrew Communities 63 ( 2 uj/6 ) S31Vfcia31d3ANI HOS 30 SSVAI019 ( S31VUe31U3ANI IIOS 30 SSVIAIOIS ct5 ( 2LU/6 ) S31Vaa31H3ANI IIOS 30 SSVl^OI9° (2LU/6) S31Vd9aiU3ANI IIOS 30 SSVl^019 Fig. 3. Histograms of biomass/size class of soil invertebrates in the plant associations studies: a, Pinus sylvestris, Sphagnum fuscum; b, Pinus sylvestris, Vacciniurn vitis-idaea, Pleurozium schreberi; c, Picea abies, Oxalis acetosella (taiga type); d, Picea abies + Alnus incana, Geum rivale + Filipendula ulmaria. 64 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY 1.M ■ O OD LL O Z! LU ! — 1.0- OB- O O 0£ LU cr Z) I — o QV 0.2- QQl 1 3 A 6 2 5 b 1 3 6 2 5 C NO. 18 5 Fig. 4. — Cluster analysis of the data on the biotopic distribution of shrews in the study area, a, data from snap trapping; b, data from ditches; c, data of Putchkovsky (1969a). Numbers refer to the following species: 1, Sorex araneus; 2, S. caecutiens; 3, S. minutus; 4, S. isodon-, 5, S. minutissimus; 6, Neomys fodietis. 1994 SHVARTS AND DEMIN — Temperate Forest Shrew Communities 65 ( UJm ) SISAHdIAlAS dO H10N31 NV31AI — O. oj fvi c- 3"cOcJ O cn <33 — O O H- 1 — : — ! i 1 J — t- H . — — p ill. CD O I .re CO OJ cn -J 1 OO CO ( LUUI ) SU310VBVH0 Mvr dO H10N31 NV31AI Fig. 5.— Means (±1 SD) of jaw characters in the shrews studied. 1, Neomys fodiens; 2, Sorex isodon; 3, S. araneus; 4, S. caeculiens; 5, S. minutus; 6, S. minutissimus. Characters: A, length of the arm of the lower jaw (from the tops of the incisor of the lower jaw to the anterior edge of the coronoid process); b, tip of first incisor to posterior tip of the angular process of the mandible; c, distance between the front border of the molars of the lower jaw and the lower tip of the coronoid process, d, length of the lower toothrow; e, length of the upper toothrow; f, length of the lower jaw symphysis. NUMBER OF INDIVIDUALS 66 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 BODY SIZE ( mm ) Fig. 6. — Numbers and individual mass of invertebrates in relation to body size and biomass distribution. TERRITORIALITY IN JUVENILES OF THE COMMON SHREW (SOREX ARANEUS) IN PREPEAK AND PEAK YEARS OF POPULATION DENSITY Natalia Moraleva* and Alexandra Telitzina' 'institute of Evolutionary Morphology and Ecology of Animals, Russian Academy of Sciences, Leninsky Prospect 33, Moscow, 117071 Russia Abstract The results of the study of the territoriality and spacing behavior of immature common shrews Sorex araneus in the middle Yenisei taiga are presented and discussed for prepeak (1984) and peak (1985) years. Using information obtained from 100 live traps and 60 pitfalls on a square grid, shrews were mobile in summer, but became sedentary by autumn. Three groups of animals — transients, settlers (short-term occupants of the grid), and residents — were delimited. The percentages of these groups remained constant irrespective of the increase of the total population density. The decrease of the home-range size in August in the peak year was accounted for mainly by settlers reducing their exploratory behavior. The decrease of the residents’ home ranges in September may be one reason for the dramatic population crash in the winter period. The change in territoriality observed during the study is likely to be one of the main prerequisites for regular four-year cycling in population density. Introduction Many alternate causal mechanisms have been hypothesized to account for population fluctuations of small mammals. Four behavioral hypotheses were reviewed by Gaines and McClenaghan (1980): social subordination (Christian, 1970), genetic-behavioral polymorphism hypothesis (Chitty, 1967; Krebs, 1978, 1979), presaturation-saturation dispersal hypothesis (Lidicker, 1975), and social cohesion hypothesis (Bekoff, 1977). A resident fitness hypothesis (Anderson, 1989) and social fence hypothesis (Hestbeck, 1982, 1988) are also worthy of mention. All hypotheses of population dynamics have been based on investigations of small rodents, mainly Microtinae (Microtus and Clethrionomys). Studies of the dispersal, territoriality, and spacing behavior of Soricidae are numerous, but most do not concern the variation of those parameters in relation to long- term population dynamics (Shillito, 1963; Michelsen, 1966; Hawes, 1977; Churchfield, 1980; Ellenbroek, 1980; Malmquist, 1986). Such research, especially in cyclic populations, is of great interest due to special biological features of the Soricidae. One of the most important features of Sorex shrews is that they usually do not breed in the year of their birth (Pucek, 1959; Kaikusalo and Hanski, 1985; Sheftel, 1989). Therefore, until the next spring all immatures are equal in their social status in the population, which usually is associated with sexual maturation (Gaines and McClenaghan, 1980; Boyce and Boyce, 1988). The other characteristic feature of the soricid shrew is its solitary and aggressive nature (Croweroft, 1957; Eisenberg, 1964; Moraleva, 1989). Stability of social interaction occurs only for mother and young. But at the end of lactation relations between mother and young and between siblings become aggressive (Dehnel, 1952; Moraleva and Pavlova, 1983; Michalak, 1988) and most immature animals disperse from their natal home ranges. In populations of common shrews, specific interaction between adults and immature animals was observed (Moraleva, 1989). Although breeding adult rodents achieve dominant rank (Gaines and McClenaghan, 1980), immature common shrews are able to pressure adult males and cause their emigration (Moraleva, 1989). Interaction between immature animals and adult females was more complicated. Immature shrews avoided encounters with adult females but at the same time dispersing females avoided the home ranges of immatures (Moraleva, 1989). Old adults disappeared from the population between July and October (Croweroft, 1956; Michelsen, 1966; Pemetta, 1977; Churchfield, 1979, 1980; Moraleva, 1983, 1989). The influence of adult shrews on the dispersal and spacing behavior of immatures is less compared with rodents. That is the reason why it is possible to consider intraspecific interactions separately for sex and age groups. In this study, the interaction between adult and immature common shrews, considered previously (Moraleva, 1989), is examined critically during the summer period, when the territorial population structure is forming. It is known that the shrew population of the present study possesses a four-year density cycle (Sheftel, 1989). We studied the spacing behavior during the summer of peak and prepeak years, keeping in mind that these data could be of some significance for the understanding of the general mechanism of population dynamics. Methods Studies were conducted at the Northern Ecological Station, Mimoe, of the A. N. Severtsov Institute of Evolutionary Morphology and Ecology of Animals, Russian Academy of Sciences. The station is located on the eastern bank of the Yenisei River, at the eastern edge of the western Siberian plain. The study area has a typical taiga landscape and is characterized by a continental climate (Sheftel, 1989). Shrews were live trapped on the eastern bank of the Yenisei River, about 1.5 km from the river. The live-trapping area is typified by a mixed forest with Pinus sibirica. Pice a obovata, and Populus tremula as the dominant tree species. The understory is dominated by Vaccinium myrtillus, Equisetum sylvaticum, and Equisetum pratense. Mosses, mainly Pleurozium schreberi and Hylocomium splendens, cover 67 68 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 40-50% of the area. Live trapping was carried out with handmade, box-type live traps and pitfall traps with food and moss provided to increase the survival of shrews in the traps. A piece of porous, spongy material (Porolon) was placed on the bottom of the pitfall to absorb extra moisture. The pitfalls were covered to prevent rain water from entering. Within the study area, 100 live traps were arranged in a square grid with a 15-m interval between traps (ten traps in a line). In addition, 60 pitfall traps were placed in seemingly favorable spots about 30 m apart. The live traps and pitfalls were examined every six to nine hours, and the shrews captured were marked by toe clipping. In 1984, the live-trapping periods were 23 June to 3 July, 6-15 August, and 14-19 September. In 1985, the trapping periods were 7-17 July, 1-12 August, and 14-19 September. We replaced traps in September 1985 to determine the boundaries of small home ranges. The date, time, and point of capture of each shrew were recorded. Shrews were classified as juveniles (sexually immature, current-year animals) or adults (sexually mature, over-wintered animals — Dehnel, 1949); gender was determined only for adult shrews. Dead shrews were autopsied. For statistical analysis we employed the Student r-tests and test. Results All marked animals were divided into two groups — those never recaptured, and those captured more than one time. These categories were considered a reasonable approximation to evaluate dispersing and residential animals (Table 1). The difference in values obtained for peak and prepeak years was not significant (P > 0.05; T^^ = 0.42). The number of newcomer animals sharply decreased in September and all (four and two) were recaptured. That is evidence for the end of dispersal at this time. Some dispersing animals marked on the grid were recaptured outside the grid in the trapping ditches (Sheftel, 1989). These data give us an opportunity to estimate the distance of the movements of shrews. In 1984, two animals were captured in five and eight days after the first capture at 1.5 and 2 km away from the grid. In 1985, one animal was found 1.5 km away after 44 days. These results are in agreement with those previously reported (Moraleva, 1983). For both prepeak and peak years the majority of animals marked in July disappeared gradually by September and the base of the shrew population in September was composed of animals marked in August (Table 2). The difference in the proportion of animals marked in July that remained until September for different years may be affected by the fact that the term of trapping in 1984 occurred two weeks earlier than in 1985. Several variations of spacing behavior of immature individuals can be recognized (Fig. la). There were some animals in the population that moved widely across the study grid (Fig. la). Most such animals usually were recorded only in one trapping period (e.g., nos. 216, 295) and single individuals (e.g., no. 179), during two trapping periods (July and August). This shows that such animals can occupy an area beyond the grid. The other type of spacing behavior is shown in Fig. lb. These animals also moved widely but mainly within the grid. It is typical for them to investigate the site more thoroughly after crossing some distance (ca 80-100 m). Figure Ic depicts animals that appeared on the study grid in June (no. 52) or in August (nos. 226, 327). At the beginning they moved more widely and were likely to search the area in which to establish a home range. By the next trapping period in August or September they settled down in small home ranges. Figure Id shows the sites of captures and borders of home ranges of resident shrews which lived there for a long time and never were recorded beyond these boundaries. Such animals usually were recorded all through the trapping periods in summer 1984 (Fig. Id, lower portion), or in 1985 (Fig. Id, upper portion). One animal (no. 81) survived the winter and was recaptured in the summer of 1985 within the same home range. Thus, changes in the type of spatial behavior could be recognized through the transformation from wide, active movements to the restricted movements within the single home range. This pattern prompted us to use the maximum distance between capture points instead of an area estimate to express home range size. In this analysis we used animals which were captured more than three times in August and two times in September. The animals captured only on the last line of traps were not included (Table 3). The decrease of the home range size reflects the end of the exploratory period by September and the stabilization of the home ranges in both years (1984: P < 0.01, Tj, = 2.7; 1985: P < 0.001, = 5.3). The home range sizes in August {P < 0.01, = 2.9) and September {P < 0.01, Tjj = 3.1) decreased in the peak year in comparison with the prepeak year. Numbers of individuals with different spacing behavior types are evaluated in Fig. 2. The analysis of histograms (x^ test) showed a significant difference {P < 0.05, x^ = 13.6) between Augusts in prepeak and peak years. The differences in the distribution can be related to the increasing proportion of animals with a home range size of 20-40 m during the peak year. In the prepeak year, the proportion of animals with a home range size of 40-80 m was higher than in the peak year. For the calculation of population density a strip half the width of the mean specific range of movement was added to the area of the grid. Chitty (1948) proposed the formula A = L“ -I- 4 (Lr -I- %7rr^) where A is the effective or true trapping area, L is the length of a side of the square trapping grid (the spaced-out area, within the confines of which the traps are arranged), and r is the radius of mean territory when it is assumed to be circular. The surface of the practical trapping area was 18,225 m^ (135 m X 135 m). The means of territory size were calculated as a maximum distance between two points of recapture (Table 3). Half of this distance was considered as a radius (r) of activity ranges. Population density was expressed as the number of individuals per ha (Table 4). For purposes of density calculations we used the total number of individuals present on the grid more than three days. The maximum density of shrews occurred in early August, so the competition for space at that time should be the most 1994 MORALEVA AND TELITZINA — Territoriality and Population Density of Sorex araneus 69 intensive. Interaction between shrews in prepeak and peak years can be considered as competition between the “old” resident shrews marked during a previous trapping period (June or July) with established home ranges, and a second group of individuals appearing and settling down in August. In the prepeak year the density of “old” residents in August was 3.6 shrews/ha. Their home ranges were exclusive and a considerable portion (about 80%) of the study grid was unoccupied. In contrast to that in August of the peak year, the density of “old” residents was four times more, 14.2 shrews/ha. Their home ranges covered 54% of the area of the study grid and the overlap was about 17 % of the whole territory occupied. During both years, “old” residents disappeared gradually during August. Some of them showed dispersal trends because their last captures were outside of their home ranges near the border of the study grid. For a comparison of the average numbers of recaptures per individual in different years, we used the animals which were caught in the first five days of the trapping period in August (Table 5). In the prepeak year (1984) the total number of recaptures per individual (7.9) for August was significantly greater {P < 0.001) than this parameter for “old” residents (2.9). However, there was no difference of these values (4.5 vs 3.8) for the peak year (P > 0.05, T^^ = 1.60). The total number of recaptures per individual in the prepeak was significantly greater than the total in the peak {P < 0.001, = 3.7), but at the same time there was no significant difference between “old” residents (P > 0.05, = 0.74). Thus residential animals were captured in the same frequency in prepeak and peak years, but the frequency of recaptures was reduced in the peak years for those animals which appeared just in August. The proportion of animals that persisted on the grid in August is shown in Fig. 3. The animals staying on the grid longer than ten days (to September or to the next year) are considered as residents. Animals of the intermediate type, called settlers, searched for territories and attempted to settle down. The differences in proportions of those types in the prepeak and peak years was not statistically significant (P > 0.05, x" = 2.2). There were no significant differences in home range sizes in August between prepeak and peak years for residential animals (Table 6). At the same time the home range size of settlers decreased significantly in the peak year (P < 0.05, = 2.51). Hence, the high density of “old” residents leads to reducing search activity of settlers. On the other hand, settler shrews in August exhibited overlapping home ranges to a greater extent with “old” residents than did residents (those shrews remaining until September). The settlers, together with “old” residents, were captured in 41 % of the traps, and residents in 26%. Intraspecific competition forced the shrews to avoid the home ranges of “old” residents and the most successful tactic was to settle in any unoccupied space between the home ranges of residents. The final position of home ranges in September is shown in Fig. 4. In the prepeak year four new animals appeared in our grid. Two of them (nos. 259 and 279) settled in small home ranges. The home range of no. 279 overlapped with the home range of no. 182. Two other shrews (nos. 282 and 277) moved all over the grid as they tried to establish their own home ranges. In September of the peak year two new animals appeared— nos. 2511 and 2312. They both lived near the borders of the grid and were residents (Fig. 4). All animals of that time were strictly resident and stayed within small areas. Only two individuals (nos. 92 and 269) went beyond their home ranges a short distance; two others (nos. 189 and 212) moved widely enough and came to neighboring home ranges. In September, when the border of home ranges stabilized, we calculated the area of home ranges (Table 7) using the exclusive boundary method (Ward, 1984). The size of home ranges varied between 440 and 1013 m^ in 1984 and between 225 and 900 m“ in 1985. The average size was smaller in the peak year (P < 0.05). These data are in agreement with those of Michelsen (1966), who found the winter home ranges of shrews to vary between 370 and 630 m^. Discussion Our data show that the population structure of immature shrews of S. araneus in the Yenisei taiga was similar to other populations investigated and also labile on a seasonal basis (Shillito, 1963; Michelsen, 1966; Buckner, 1969; Pemetta, 1977; Churchfield, 1980, 1984). Shrews were mobile in summer and sedentary in autumn. However, the seasonal changes of population structure occurred more rapidly in this study than in European populations, probably due to the shorter warm period. In this study, the first litters began dispersing as late as 20 June (Sheftel, 1989), and all animals became residents by the middle of September. Aggressive interaction in the natal nest stimulates dispersal of young (Dehnel, 1952; Moraleva and Pavlova, 1983; Michalak, 1988). As Michelsen believed, the young (at least of the first litters) occupy the optimal home ranges during the first weeks following dispersal. The investigation of interactions between adult females and immatures shows that the young of the first litters avoided the home ranges of breeding females and preferred to settle in unoccupied territories (Moraleva, 1989). By the time of the dispersal of the second litter (the end of June to the beginning of August), the majority of adult females have abandoned their home ranges and become transients. By early August, the population density as well as competition for space by immature shrews reached maximum levels and simultaneously the final rearrangement of individual territories of immatures occurred. Some residential animals disappeared from the grid. Probably some shrews perished, while others were forced to move as they were replaced by immigrants. Thus, the residential status of shrews did not determine the “winner” in competition for space. Individual features are likely to determine the result of interaction. The successful result of the competition for space may be dependent, for example, on the type of nervous system, body mass, or some other features. Hanski et al. (1991) suggested that the individual’s competitive ability may be related to the biting strength of the jaw rather than body mass. 70 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 As a result of the sharp competition for space in August, territorial structure without overlapping home ranges was formed by September. By that time the dispersal process had actually finished and only scattered individuals kept searching for home ranges (Fig. 4). In September, the population consisted predominantly of shrews, mostly of the second litter, that had settled the grid in August. Hence, the actual role of the early- and late-born young in the maintenance of the population could be different. This evidence is in agreement with the idea (Pemetta, 1977) that the late-bom young are the most important cohort in the population structure. Individual variation in behavior is the rule rather than the exception in most species of animals (Slater, 1981; Ims, 1989), and should be especially important in shrew populations where individuals tend to be equal in age and sexual maturity, and prevailing ecological conditions seem uniform. Differences in spacing behavior of immature common shrews during summer is demonstrated here. Some shrews chose home ranges rapidly and maintained them without change throughout the year until the end of the next period of reproduction. Those individuals are similar to residents which were described by Inoue (1988) for Sorex unguiculatus. The opposite type of behavior was shown by transient animals, often moving a considerable distance. The maximum distance moved was about two km in our study but S. araneus is known to move even greater distances (Tegelstrom and Hanssen, 1987) and even to disperse over water (Hanski et al., 1991). Ail other kinds of territorial behavior in summer, demonstrated in Fig. 1, could be considered as variants of exploratory behavior, when animals tried to establish their home ranges. At the same time these different behavioral patterns could be imagined as a succession of actions; hence, different movement patterns reflect the different stages of settlement. At first, the shrews could move extensively like a transient, but subsequently the radii of their trips become shorter and at the end they choose their home ranges and become residents. On the other hand, some residents could be forced to disperse and become transient. Frequently distribution diagrams show the relationship between home range length (Fig. 2) and the type of territorial behavior (Fig. 1). Home range length for residents (Fig. Id) extends to 40 m, supporting the results of other authors: 17.8 m (Dickman, 1980), 22-24 m (Michelsen, 1966), and 11.4 and 30 m in different types of habitat (Yalden, 1974). Animals whose home ranges are 40 m to 80-100 m in length can be considered to be settlers searching for new home ranges in which to settle. This type of territorial behavior is shown in Fig. lb, c. The animals moving wider than 80-100 m (nos. 216, 268 — Fig. la) seems to be intermediate between transients and settlers. Dispersal and social behavior in rodent populations are functions of age, reproductive status, and ecological conditions (Christian, 1970; Fleming, 1979). The dispersal of young is greatly influenced by philopatry as well as dominant rank and aggressiveness of adults, competition for breeding success, avoidance of inbreeding, and so on (for review, see Gaines and McClenaghan, 1980; Stenseth, 1983; Anderson, 1989). In contrast to rodents, all offspring of shrews are equal in breeding and social conditions and are able to force out adult males. In early August, offspring become independent from the influence of adult females (Moraleva, 1989), so dispersal of immatures was induced by internal factors only. The differences in dispersal trends of individual animals can be genetic in nature (Krebs et al., 1973; Krebs, 1978, 1979) or nongenetic, including maternal effects (Hilbom, 1975; Kawata, 1987) or developmental stability (Zakharov et al., 1991). In this study density fluctuated from 7 to 36 shrews and correlated approximately with published results. For August-September, Shillito (1963) found density of the common shrew 24-28 shrews/ha, Yalden (1974) 42 shrews/ha, and Dickman (1980) 39 shrews/ha. In contrast, Michelsen (1966) calculated only the total number of shrews of territories present. Hence, our results for September are well-suited for comparison with Michelsen (1966), as at that time shrew dispersal became reduced and total number of territories could be actually evaluated. Our parameters for August were higher at the expense of the animals temporarily occupying the study grid. Population density (about 18 shrews/ha) observed by Michelsen (1966) actually did not change for the two-year investigation and correlated with the density characteristics of our population in the peak year (Table 4). The difference in population structure for the prepeak and peak years can be considered a special subject. In the peak year the population density increased approximately two times in comparison with the prepeak year. The proportion of animals with different trends of dispersal remains constant irrespective of the variation in total population density (Table 1, Fig. 3). This is evidence that there is no correlation between the proportion of animals dispersing and the density of the population for the prepeak and peak years. As we later found, there was such a correlation in the years of low population density (Moraleva, unpublished data). The increase in population density in the peak year led to reduction of home range size (Table 3). The decrease of the home range size in August as well as number of recaptures per individual (Table 5) was due mainly to settlers, which can be considered an indication of inhibition of exploratory behavior. This agrees with the Johnson (1988) model, predicting the decrease in exploration when density increased. The size decrease of the residential home ranges in September may be one of the reasons for the dramatic population crash in the winter period. Because the region studied is characterized by significant spatial heterogeneity (Sheftel, 1989, 1994), models based on this factor (Lidicker, 1975, 1985; Hestbeck, 1982, 1988) are of interest for the analysis of the population dynamics. Our data suggest that in the prepeak year the home ranges of the first generation animals do not cover the space completely. There are vacant areas remaining that permit other shrews to move and establish home ranges. The possibility of avoiding extra contacts as vacant areas are occupied decreases the stress level in the population. This pattern of distribution allows more “experienced” animals to find and occupy the best habitats. The same type of spacing (when more experienced animals were located in the optimal habitats) was found for the white-footed mouse (Morris, 1989). On the conditions of such 1994 MORALEVA AND TELITZINA — Territoriality and Population Density of Sorex araneus 71 distribution, relatively few animals (in the first instance from poor habitats) are thought to perish in the extreme winter period. Intensive reproduction during the next spring produced a first generation of shrews that numbered about two times more in peak than in prepeak years. As a result of four times greater coverage by home ranges, exploration behavior was inhibited in the peak year. The most successful strategy is the fast selection of any vacant space in any type of habitat. In the peak year young shrews occupied all kinds of habitats, including suboptimal ones. According to the social fence hypothesis (Hestbeck, 1982, 1988), when population density rises above the carrying capacity, further population regulation is achieved through resource exhaustion. The reduction of the home range is believed to lead to resource exhaustion, making it impossible for some individuals to survive winter. When environmental resources become depleted in winter, animals probably begin to move. Obviously, shrews from the poorest habitats or those with smallest home ranges or the weakest individuals were the first to leave their home ranges (“saturation dispersal” according to Lidicker, 1975). Those individuals usually must cross the home ranges of residential animals, resulting in some extra stress and depletion of energy. The mobile animals are the first to die. Step by step, more and more animals perish during winter and density depression occurs. This study of immature common shrews in prepeak and peak years supports the viewpoint that change in territoriality and spacing behavior can be one of the reasons for regular four-year cycling in the population dynamics of S. araneus. Acknowledgments We are indebted to B. Sheftel and O. Bourski for their helpful suggestions and constructive criticism during the preparation of this paper. Critical comments by R. Rose improved the manuscript and are gratefully acknowledged. Our special thanks are due to the Carnegie Museum of Natural History and J. F. Merritt of Powdermill Nature Reserve for organizing this fine colloquium. We are grateful to the following individuals for their help in the field: V. Lebedev, A. Panaiotidy, and E. Semiochina. We reserve our deepest gratitude for V. M. Moralev, K. G. Cheshihina, and V. Miklaev for the help in translation and drawing the figures. Literature Cited Anderson, P. K. 1989. Dispersal in rodents: A resident fitness hypothesis. Special Publication, The American Society of Mammalogists, 9:1-141. Bekoff, M. 1977. Mammalian dispersal and the ontogeny of individual behavioral phenotypes. American Naturalist, 111:715-732. Boyce, C. C. K., and J. L. Boyce 111. 1988. Population biology of Microtus arvalis. II. Natal and breeding dispersal of females. Journal of Animal Ecology, 57:723-736. Buckner, C. H. 1969. Some aspects of the population ecology of the common shrew Sorex araneus, near Oxford, England. Journal of Mammalogy, 50:326-332. Chitty, D. 1948. An estimate of the true density and mean range movement of moles. Appendix I, in The Ecology of Mole (Talpa europaea L.) Populations (P. A. Larkin). Unpublished Ph.D thesis. University of Oxford. 1967. The natural selection of self-regulatory behaviour in animat populations. Proceedings of the Ecological Society of Australia, 2:51-78. Christian, J. J. 1970. Social subordination, population density, and mammalian evolution. Science, 168:84-90. Churchfield, j. S. 1979. Studies on the ecology and behaviour of British shrews. Unpublished Ph.D. thesis. University of London. 1980. Population dynamics and seasonal fluctuations in numbers of the common shrew in Britain. Acta Theriologica, 25:415-424. 1984. An investigation of the population ecology of syntopic shrews inhabiting watercress beds. Journal of Zoology (London), 204:229-240. Crowcroft, P. 1956. On the life span of the common shrew (Sorex araneus L.). Proceedings of the Zoological Society of London, 127:285-292. 1957. The Life of the Shrew. Reinhart, London, 166 pp. Dehnel, A. 1949. Studies on the genus Sorex L. Annals of the University of M. Curie-Sklodowska, Section C, 4: 17-102 (in Polish). 1952. The biology of breeding of the common shrew in laboratory conditions. Annals of the University of M. Curie- Sklodowska, Section C., 6, 11:359-376. Dick man, C. R. 1980. Estimation of population density in the common shrew Sorex araneus from a conifer plantation. Journal of Zoology (London), 192:550-552. Eisenberg, j. F. 1964. Studies on the behavior of Sorex vagrans. The American Midland Naturalist, 72:417-425. Ellenbroek, F. j. M. 1980. Interspecific competition in the shrews Sorex araneus and S. minutus (Soricidae, Insectivora): A population study of the Irish pygmy shrew. Journal of Zoology (London), 192:119-136. Fleming, T. H. 1979. Life history strategies. Pp. 1-6, in Ecology of Small Mammals (D. M. Stoddard, ed.), Chapman and Hall, London, England. Gaines, M. S., and L. R. McClenaghan, Jr. 1980. Dispersal in small mammals. Annual Review of Ecology and Systematics, 11:163-196. Hanski, L, a. PELTONEN, and L. Kaski. 1991. Natal dispersal and social dominance in the common shrew Sorex araneus. Oikos, 62:48-58. Hawes, M. L. 1977. Home range, territoriality and ecological separation in sympatric shrews, Sorex vagrans and Sorex obscurus. Journal of Mammalogy, 58:354-367. Hestbeck, J. B. 1982. Population regulation of cyclic mammals: The social fence hypothesis. Oikos, 39:157-163. 1988. Population regulation of cyclic mammals: A model of the social fence hypothesis. Oikos, 52:156-168. Hilborn, R. 1975. Similarities in dispersal tendency among siblings in four species of voles (Microtus). Ecology, 56:1221-1225. IMS, R. A. 1989. Kinship and origin effects on dispersal and sharing in Cletlirionomys rufocanus. Ecology, 70(3):607-616. Inoue, T. 1988. Territory establishment of young bi-clawed shrew, Sorex unguiculatus (Dobson) (Insectivora, Soricidae). Researches in Population Ecology, 30(l):83-93. Johnson, M. L. 1988. Exploratory behavior and dispersal: A graphical model. Canadian Journal of Zoology, 67:2325-2328. Kaikusalo, A., AND I. Hanski. 1985. Population dynamics of Sorex araneus and S. caecutiens in Finnish Lapland. Acta Zoologici Fennici, 173:283-285. Kawata, M. 1987. The effect of kinship on spacing among female red-back voles Clethrionomys rufocanus beldfordiae. Oecologia (Berlin), 72:115-122. 72 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 Krebs, C. J. 1978. A review of the Chitty hypothesis of population regulation. Canadian Journal of Zoology, 56:2463-2480. 1979. Dispersal, spacing behavior and genetics in relation to population fluctuations in the \o\eMicrotus townsendii. Fortschritte der Zoologie, 25:61-77. Krebs, C. J., M. S. Gaines, B. L. Keller, J. H. Meyers, and R. H. Tamarin. 1973. Population cycles in small mammals. Science, 179:35-41. Lidicker, W. Z., Jr. 1975. The role of the dispersal in the demography of small mammal populations. Pp. 103-28, in Small Mammals: Their Production and Population Dynamics (F. B. Golley, K. Petrusewicz, and L. Ryszkowski, eds.), Cambridge University Press, London, 451 pp. 1985. Dispersal. Pp. 420-254, in Biology of New World Microtus (R. H. Tamarin, ed.). Special Publication, American Society of Mammalogists, 8:1-893. Malmquist, M. G. 1986. Density compensation in allopatric populations of the pygmy shrew Sorex minutus on Gotland and the Outer Hebrides: Evidence for the effect of interspecific competition. Oecologia (Berlin), 68:344-346. Michalak, J. 1988. Behavior of young Neomys fodiens in captivity. Acta Theriologia, 33:487-504. Michelsen, N. C. 1966. Intraspecific and interspecific competition in the shrews Sorex araneus L. and S. minutus L. Archives Neerlandaises de Zoologia, 17:73-174. Moraleva, N. V. 1983. Territorial relationships of shrews in the Yenisei taiga (using individual tagging). Pp. 215-230, in Zhivotnyi mir eniiiseiskoi taiga i lesotundra i prirodnaya zonalnost (E. E. Syroechkovskii, ed.), Nauka, Moscow. 1989. Intraspecific interactions in the common shrew Sorex araneus in central Siberia. Annales Zoologici Fennici, 26:425-432. Moraleva, N. V., and E. J. Pavlova. 1983. Breeding in captivity of Sorex tundrensis and the behavior of young in the first month of life. Pp. 128-130, in Mekhanismy Povedeniiu Zhivotnikh 1. Nauka, Moscow. Morris, D. W. 1989. Density-dependent habitat selection: Testing the theory with fitness data. Evolutionary Ecology, 3:80-94. Pernetta, j. C. 1977. Population ecology of British shrews in grassland. Acta Theriologica, 22:279-296. PUCEK, Z. 1959. Sexual maturation and variability of the reproductive system in young shrews {Sorex L.) in the first calendar year of life. Acta Theriologica, 3:269-296. Sheftel, B. I. 1989. Long-term seasonal dynamics of shrews in central Siberia. Annales Zoologica Fennici, 26:357-369. 1994. Spatial distribution of nine species of shrews in the central Siberian taiga. Pp. 45-55, in. Advances in the Biology of Shrews (J. F. Merritt, G. L. Kirkland, Jr., and R. K. Rose, eds.), Carnegie Museum of Natural History Special Publication no. 18, X + 458 pp. Shillito, j. F. 1963. Observation on the range and movements of a woodland population of the common shrew Sorex araneus L. Proceedings of the Zoological Society of London, 140:533-546. Slater, P. J. B. 1981 . Individual differences in animal behavior. Pp. 35-49, in Perspectives in Ethology 4 (W. Bateson and P. Klopfer, eds.). Plenum Press, New York. Stenseth, N. C. 1983. Causes and consequences of dispersal in small mammals. Pp. 63-101, in The Eeology of Animal Movements (J. Swingland and P. Grenwood, eds.), Oxford University Press, Oxford, England. Tegelstrom, H., and L. Hanssen. 1987. Evidence of long distance dispersal in the common shrew (Sorex araneus). Zeitschrift fur Saugetierkunde, 52:52-54. Ward, G. D. 1984. Comparison of trap and radiorevealed home ranges of the brushtailed possum (Trichosurus vulpecula Kerr) in New Zealand lowland forest. New Zealand Zoology, 11:85-92. Yalden, D. W. 1974. Population density in the common shrew, Sorex araneus. Journal of Zoology (London), 173:262-264. Zakharov, V. M., E. Pankakoski, B. 1. Sheftel, A. Peltonen, AND I. Hanski. 1991. Developmental stability and population dynamics in the common shrew Sorex araneus. The American Naturalist, 138:797-810. Table 1. — Number and percentage of capture-recaptured Sorex araneus based on the period in which first trapped and marked in prepeak (1984) and peak (1985) years. 1984 1985 Trapping period 6123-113 8/5-8/15 9/14-9/19 7/7-7/17 8/1-8/13 9/14-9/19 Captured one time 19 (49%) 44 (44%) — 42 (45%) 79 (49%) — Recaptured 20 (51%) 55 (56%) 4 (100%) 52 (55%) 82 (51%) 2 (100%) Total 39 99 4 94 161 2 Table 2. — Number and percentage of Sorex araneus in September in prepeak (1984) and peak (1985) years divided according time of marking. 1984 1985 Total in September 25 40 Marked in July 2 (8%) 13 (32.5%) Marked in August 19 (76%) 25 (62.5%) Marked in September 4 (16%) 2 (5%) 1994 MORALEVA AND TELITZINA — Territoriality and Population Density of Sorex araneus 73 Table 3. — Average home range size of Sorex araneus (in m between extreme capture points) in prepeak (1984) and peak (1985) years. 1984 1985 August September August September X 65.6 32.5 44.4 16.5 SD 38.8 33.8 34.1 10.6 SE 6.2 9.4 4.2 1.9 n 39 13 67 31 Table A.— Population density of Sorex araneus (number of individuals per ha) in prepeak (1984) and peak (1985) years. 1984 1985 July August September July August September Radius 19.0 32.9 16.3 25.3 22.3 7.4 True trapping area (m“) 29,619 39,466 27,669 33,688 31,626 22,159 Number of individuals 20 60 25 50 115 40 Density 6.8 15.2 9.0 14.8 36.4 18.1 Table 5. — Number of recaptures per individual for Sorex araneus caught in first five days of trapping period in August in prepeak (1984) and peak (1985) years. 1984 1985 Total “Old" Residents Total “Old" Residents X 7.9 2.9 4.5 3.8 SD 6.50 2.42 2.64 3.99 SE 0.99 0.9 0.34 0.69 n 44 8 60 35 Table 6.— Average home range size of Sorex araneus in August for resident and settlers in prepeak (1984) and peak (1985) years. 1984 1985 Residents Settlers Residents Settlers X 52.4 89.3 44.0 48.4 SD 19.9 47.3 30.0 30.0 SE 6.0 13.1 8.3 6.3 n 12 14 14 24 Table 7 .—Average home range size of Sorex araneus in September in prepeak (1984) and peak (1985) years (nr). 1984 1985 r 613.8 417.6 SD 203.7 204.3 SE 64.5 38.6 n 11 29 74 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 Fig. 1.— Variation in spatial behavior (a-d) of individually marked common shrews. The circled numbers indicate the points of first capture, arrows indicate the subsequent captures. Empty symbols indicate the points of capture in July, black symbols indicate August, symbols with a dot inside indicate September, solid lines indicate borders of home ranges. 1994 MORALEVA AND TELITZINA — Territoriality and Population Density of Sorex araneus 75 -+-J SIVnaiAIQNl dO UBaiAIHN CN IE Home range sizes of common shrews in August and September in prepeak (1984) and peak (1985) years. SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 76 (f) < Q > Q U. O CC UJ m D Fig. 3. — The duration (in days) spent by common shrews within the grid in August in prepeak (1984) and peak (1985) years. 1984 Fig. 4.— Movements and home ranges of individually marked common shrews in September of prepeak (1984) and peak (1985) years. All symbols, lines, and numbers as in Fig. 1. Dashed line indicates borders of supposed home ranges of animals captured in August 1984 and then 1985. Shaded areas indicate home ranges of animats marked in June and July. FORAGING STRATEGIES OF SHREWS, AND THE EVIDENCE FROM FIELD STUDIES Sara Churchfield Life Sciences, King’s College, Campden Hill Road, London, England W8 7AH Abstract This paper outlines aspects of the feeding ecology of ten soricid species in an attempt to elucidate the foraging strategies of wild shrews using information gathered from field studies. Shrews are shown to be wide-spectrum feeders but the diet of each species is dominated by a few major prey taxa which are common and abundant. Shrews exhibit quantitative rather than qualitative specialization. While each species predominantly uses a single foraging mode with respect to prey location, this is not exclusive. Within the constraints of body size each species exploits a wide range of prey sizes which is attributed to encounter rates with invertebrates. Prey selection was related more to availability than to size or food value as a guide to profitability. While shrews take certain prey in proportion to their abundance, other prey appear to be underutilized, and the reasons for this are explored. Evidence of prey-switching and differential patch use is investigated. Shrews are found to be excellent opportunists with elements of dietary specialization and partial selection in addition to generalization. The feeding and foraging strategies that they employ allow exploitation of a full range of terrestrial and semiterrestrial habitats, and permit rapid adaptation to spatial and temporal changes in prey availability. Introduction With their high metabolic rates, voracious appetites, and continuous daily and year-round activity, shrews are of considerable interest with respect to their feeding ecology and foraging strategies. Questions arise about the ability of these small predators to locate sufficient food to satisfy their daily requirements, their selection of prey, their response to changing availability of prey, and their relationships with competitors. Shrews are diverse in form and in body size, and they have a wide geographical distribution, occurring in a range of terrestrial and semi terrestrial habitats. So, aside from their intrinsic interest they serve as useful models in the study of predator-prey interactions. There have been numerous studies of the diets of soricids and we have considerable knowledge of the feeding habits of many species, such as Sorex mi nut us (Grainger and Fairley, 1978); Sorex araneus (Rudge, 1968); S. cinereus, S.fumeus, S. palustris and Blarina brevicauda (Hamilton, 1930); Neomys fodiens (Wolk, 1976; Churchfield, 1985; Kuvikova, 1985); and Crocidura russula (Bever, 1983). Fewer studies have investigated their feeding habits in relation to prey availability (e.g., Pemetta, 1976; Churchfield, 1982). Increasing interest is being shown in the community ecology of shrews and ecological separation between species, using diet and habitat use as indicators of niche overlap (Terry, 1981; Churchfield, 1984, 1991; French, 1984; Whitaker and French, 1984; Ryan, 1986). But there has been little effort to answer many basic questions about the foraging behavior and prey selection of wild shrews, and to develop theories about their foraging strategies based on data from field studies. This is in contrast to the growing literature on optimal foraging theory involving laboratory-based experiments on shrews (e.g., Barnard and Brown, 1981, 1985; Pierce, 1987). These studies suggest that shrews are more selective in their feeding habits than field studies have previously indicated. Using data on the diets of ten soricid species from different geographic regions and habitats, and of different body sizes, this paper attempts to formulate some general theories about the feeding ecology and foraging strategies of shrews which can be derived from field studies of wild populations. In particular, it investigates prey selection by shrews. Methods Study Animals Ten species of shrews, representative of different geographic ranges, body sizes, and foraging modes were studied (Table 1). They included three European species from Britain, three species from North America, two from subtropical Africa, and two large species from tropical Africa. Eight of these species are truly terrestrial and two species (Neomys fodiens and Sorex palustris) possess adaptations for a semiaquatic mode of life. While the shrews came from a diversity of habitats (grassland, scrub-grassland and forest), each species was collected from a single habitat type using live-trapping techniques. The aim of the investigation was not to provide basic data on the diets of these soricids but to explore trends within these species which might provide insight into the foraging strategies of shrews, making use of data available from field-based studies. Diet Analysis The feeding habits of each species were studied by stomach and/or fecal analysis. For seven of the species, alimentary tracts were available from 86 specimens ranging from 8 to 16 individuals per species (Table 1). From these specimens, stomachs and intestines were dissected and the complete contents removed for analysis of prey remains. The remaining three species (Sorex araneus, S. minutus, and N. fodiens) were part of longer-term population studies and, in order to avoid kill-trapping, their diets were studied through fecal analysis. The feces produced by a shrew captured in a live-trap constituted a single sample. A mean of 14 pellets per sample was obtained from S. araneus, 12 from N. fodiens, and six from S. minutus. Further details of the technique of fecal 77 78 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 analysis and a critique of the method can be found in Churchfield (1982). Reference collections of potential prey items were used to facilitate identification of invertebrate remains. From these it was possible not only to identify the types of prey consumed, but also their size ranges since, even in fecal samples, sufficiently large fragments of most prey were found. Large numbers of diet samples were available from S. araneus, S. minutus, and N. fodiens (Table 1), covering different seasons. Fewer were available from the other species and coverage was more limited. The number of different prey taxa found will be affected by the number of samples examined, but sample number need not be great to provide a reliable indication of dietary diversity. Figure 1 shows the relationship between the number of fecal samples examined in random order and the cumulative number of prey taxa found for four species of shrews where dietary diversity was particularly high. Ninety percent of the prey types were recorded by the fifth sample in Myosorex cafer, by the seventh sample in Neomys fodiens, by the ninth sample in Sorex araneus, and the eleventh in Crocidura hirta. However, in view of the number of samples available, only Sorex araneus and S. minutus were selected for more detailed consideration. Both stomach and fecal analyses are subject to criticism on the basis that the samples collected represent only a small component of the total diet, and that quantitative assessment is difficult or impossible when only small fragments of prey remains are found. Criticisms of the technique are discussed in Churchfield (1982). A major problem is the disparity between the number of prey taken and the volume they represent. Small prey, such as Formicidae and Isoptera, may have a high encounter rate and be eaten in large numbers, but their individual energy content is low compared with a large prey type, such as an orthopteran which has a lower encounter rate and is probably more difficult to catch and handle. Using stomach and fecal analysis and assessing the diet in terms of the frequency of occurrence of different prey items clearly creates a bias towards small prey items and does not give a true reflection of which prey types constitute the bulk of the diet. Therefore, the relative volume of each prey identified in stomach and fecal samples was assessed by eye and recorded. Results were expressed in terms of percentage composition by volume and percentage of dietary occurrences (the number of occurrences of a named prey item as a proportion of all occurrences). Results and Discussion Dietary Diversity and Specialization Shrews feed on a wide range of invertebrates and in all but one of the ten species examined, the number of different prey taxa identified in the diet exceeded 12. The relatively low dietary diversity of some species, notably C poensis, may be related to sample sizes. With the exception of a single incidence of bird remains in N. fodiens, no vertebrate parts were found. Plant material occurred in small amounts in some samples but will not be considered further here. The greatest dietary diversity, with 36 different prey taxa, was found in N. fodiens, which exploits both terrestrial and aquatic prey (Fig. 2). Most invertebrate taxa are consumed by shrews, and the many detailed diet studies of other soricid species by various workers confirm this (e.g., Hamilton, 1930; Rudge, 1968; Whitaker and Mumford, 1972; Whitaker and Maser, 1976; Grainger and Fairley, 1978; Bever, 1983; Churchfield, 1984; Whitaker and French, 1984). This suggests that shrews are wide-spectrum feeders, exhibiting little selection for prey type. However, the many different prey recorded were not consumed in equal proportions and the bulk of the diet of each species comprised between two and five dominant prey types, each of which contributed at least 10% of occurrences and together made up at least 50% of dietary occurrences (Fig. 2). Despite differences in sample size, the number of dominant taxa was remarkably consistent. So, some degree of specialization or selection did occur. The identity of these major prey types differed according to the species of shrew, habitat and geographic location. Coleoptera and Araneae were particularly important prey for many temperate soricids, whereas Coleoptera, Orthoptera, Isoptera, Formicidae, and Diplopoda were dominant prey for the tropical and subtropical species. Individual species tended, then, to take large proportions of a few prey taxa and smaller but fairly constant amounts of other prey. While the actual value of these components of the diet may differ from location to location, each species appeared to have its own typical prey specialties. For example, Coleoptera, Lumbricidae, insect larvae (Coleoptera, Diptera, and Lepidoptera), and Araneae together comprised a mean of 77% (SE = 9.1) of occurrences in the diet of S. araneus during three years of sampling in a grassland habitat. The observation that each species has prey specialties is confirmed by studies of coexisting shrews, such as S. araneus, S. minutus, and N. fodiens, where each species has its own distinct dietary profile in the proportions of different prey taxa eaten (Churchfield, 1984, 1991). Foraging Mode Prey specialization among individual species may be due to the adoption of a particular foraging mode which affects their encounters with certain prey types. Shrews are mainly active on or just below the ground surface, and this is reflected in the foraging modes identified for the eight terrestrial species studied here (Fig. 3). Most terrestrial species were found to be both epigeal and hypogeal, exploiting soil-dwelling invertebrates as well as surface-dwelling prey. However, the proportions of prey taken in each mode differed. Some species were almost exclusively epigeal (e.g., C. viaria, C. poensis) whereas others showed increasing subterranean activity, culminating in S. araneus. Temperate-zone species tended to show a greater degree of subterranean foraging than tropical or subtropical species. It is possible that differences in soil compaction, organic content, and/or abundance of soil-dwelling invertebrates, or a combination of factors, between tropical and temperate soils affect foraging modes of shrews. Caution must be taken in interpretation of such results because of the disparity between the frequency of occurrence of different prey and their volume. The former takes into account the encounter rate of prey in that each occurrence is recorded, but small, frequently-eaten prey are likely to be 1994 CHURCHFIELD— Foraging Strategies of Shrews 79 overemphasized. The latter takes into account the bulk of the prey, so large prey which may be encountered and eaten comparatively infrequently will be overemphasized. The disparity between the two was not found to be significant except in those species which had a particular foraging mode or dominant prey type, notably S. araneus, which consumed large quantities of earthworms. If the volume as well as the incidence of these prey is taken into account, then 74% of the diet of this shrew comprised subterranean prey, compared with only 42% in terms of the total occurrences. The predominance of particular foraging modes within individual species can be associated with body size and anatomical adaptations (Hutterer, 1985). Sorex minutus is a small, lightly-built species, best suited for foraging on the ground surface and among vegetation. Sorex araneus is larger, more robust and better adapted for pushing into the soil. Myosorex cafer, another hypogeal, earthworm-eating soricid, is very similar. Despite their size, the two large tropical species (C. viaria and C. poensis) took few soil-dwelling invertebrates. This may reflect the scarcity of such prey relative to surface- dwelling prey, or the unsuitability of this foraging mode in hard tropical soils in the absence of special adaptations for burrowing, or both. More obvious specializations are possessed by soricid species, including N. fodiens and S. palustris, which have evolved a semiaquatic existence and whose diets include a range of freshwater prey, mostly invertebrates (e.g., Hamilton, 1930; Wolk 1976; Churchfield, 1985). Nevertheless, N. fodiens and S. palustris, at least, still retain the epigeal and hypogeal foraging modes to a greater or lesser extent (Fig. 4). Sorex palustris appeared to be more of a specialist aquatic forager than N. fodiens since over 80% of its diet (both in terms of percentage dietary occurrences and prey volume) were aquatic in origin, compared with some 50% for N. fodiens. So, although dietary specialization occurs within species with respect to foraging mode, it is not complete or exclusive. Prey Size Selection Optimal foraging theory predicts that a predator should feed more selectively when profitable prey are abundant, and ignore unprofitable prey (e.g., Cowie, 1977; Krebs et al, 1977). Laboratory experiments on S. araneus suggest that size is a guideline of prey profitability, and that there is clear selection for larger prey (Barnard and Brown, 1981). Shrews preferred larger prey, but this depended on the encounter rate. If encounter rates for large prey were low, shrews became unselective. In the wild, shrews are presented with a great diversity of prey types and sizes. Is there any evidence that prey is selected because of size? In the wild, all shrews catch and eat invertebrates of a wide range of body sizes. Even very small shrews take prey 30 mm or more in length, such as lumbricids and the larger Lepidoptera and Diptera larvae, in addition to tiny prey 3 mm in length. Some of the largest shrews may feed extensively on tiny prey 3-5 mm in length (e.g., Isoptera, Formicidae). Figure 5 shows the percentage composition by volume of prey of different size ranges found in the diets of the eight species of terrestrial soricids. Although there is no clear relationship between the size of the shrew and the prey taken, the bulk of the diet of the smallest shrews (S. minutus, S. vagrans, and S. monticolu.s) comprised small prey 3-10 mm in body length. Sorex minutus, with a mean body mass of 3.5 g, clearly preferred smaller prey. The largest shrews, exemplified by C. poensis with a mean body mass of 16.0 g, took increasing quantities of larger prey. Medium-sized shrews 8-1 1 g took a more even spread of prey sizes. Clearly, body size (or, more appropriately, jaw size) influences the prey which can be tackled, but considerable variation occurs between species in the same size range. For example, C. viaria, despite being a large shrew of about 14 g, took great quantities of very small prey (Formicidae and Isoptera), in contrast to C. poensis. Unlike S. minutus, which rarely if ever eats earthworms, S. vagrans was found to take these large prey, despite the small size of those studied here (4.3 g body mass). Thus, it is difficult to make generalizations concerning prey size selection since certain species clearly have a propensity for particular prey types, regardless of their size and their individual energetic considerations. The consumption of Formicidae by C. viaria suggests selection for these prey since C. poensis, which came from a similar area where these invertebrates are also extremely abundant, consumed few of them. Similarly, the consumption of large quantities of earthworms, which are relatively large prey, by S. araneus suggests a preference since they were not eaten to the same extent by S. minutus or N. fodiens living in the same habitat. So, despite the choices available, wild shrews do not select the largest prey they can tackle but take greater numbers of apparently less profitable prey, such as formicids. Some even seem to specialize in less profitable prey. This suggests either that shrews do not relate profitability with prey size or that availability or encounter rate are key factors in prey consumption. Prey Selection and Availability The feeding habits of shrews vary according to habitat and location with respect to the identity of the major prey types consumed and their relative contributions to the diet (Rudge, 1968; Whitaker et al., 1983). There are also seasonal oscillations, although these are often of small magnitude (Pemetta, 1976; Churchfield, 1982, 1984). But, for a given location and habitat, different prey taxa assume a characteristic proportion of the diet of each soricid species. Figure 6 shows the mean percentage contribution of each major prey type to the diet of syntopic populations of S. araneus and S. minutus over a three-year period. Despite some seasonal differences in their relative importance, Lumbricidae, Coleoptera, Diptera larvae, and Araneae remained the four dominant prey types in the diet of S. araneus, and Araneae, Coleoptera, Hemiptera, and Lepidoptera larvae remained the major prey of S. minutus. Other taxa were secondary in importance. Is this an indication of prey selection or merely a reflection of availability and encounter rate? There is some evidence to suggest that availability (and hence encounter rate) has an important influence on the incidence of certain prey in the diet. For example, changes in 80 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 seasonal abundance of adult Coleoptera, reflected in numbers captured in pitfall traps, were closely accompanied by changes in their composition in the diet of S. araneus (Fig. 7). In fact, there is a significant correlation between dietary incidence and the abundance of these prey (Fig. 8). Despite signs of seasonal changes in the abundance of other prey and attempts to monitor their occurrence in field samples and in diets, no other positive, statistically significant correlations were found (Churchfield et al., personal observation). For example, Araneae were common prey of shrews throughout the year but they had a highly seasonal occurrence in field samples, with low numbers in winter and high numbers in summer. However, this was not reflected in the diet of S. araneus. Regardless of changing abundance, this species took a fixed proportion of these prey, around 11.0% of dietary occurrences. This may reflect the inability of S. araneus to catch these athletic prey, many of which are out of reach among the vegetation. Araneae featured much more prominently in the diet of S. minutus which is smaller and more agile on and above the ground surface than is S. araneus, but sample sizes were insufficient to permit further analysis of correlation for this species. Similarly, Isopoda varied in availability according to season and habitat. Although they were certainly eaten more frequently by S. araneus in habitats where they were abundant than where they were scarce, there was little correlation between diet and availability because this shrew tended to take a fixed proportion of these prey (mean 5%, maximum 10%, of dietary occurrences). These prey had a clumped distribution and should be easy to catch and so could be highly profitable, and they were eaten in large quantities by S. minutus and N. fodiens in the same habitat. Clearly S. araneus is discriminating against them. It has been shown that certain isopod species are not favored by S. araneus (Crowcroft, 1957). Sorex araneus appears to avoid isopods when other prey are abundant. Neomys fodiens, with its larger jaws, may be better adapted for eating these prey. Sorex minutus may overcome these problems by selecting smaller species and/or individuals which are not so heavily chitinized. It thus appears that feeding habits of shrews are not simply a matter of availability, for there are also elements of palatability which affect prey choice, and these differ among soricid species. For instance, Diplopoda were rarely eaten by Sorex species, and yet they were readily taken by Neomys, Crocidura, and Myosorex. Prey Selection, Profitability, and Encounter Rate Since the bulk of prey eaten by shrews are locally common and abundant invertebrates, why don’t shrews specialize further and restrict their feeding habits to the most profitable prey? The consumption of Isopoda and Diplopoda, which have among the lowest energy values and the highest ash contents of any invertebrates (Cummins and Wuycheck, 1971) seems to contradict optimal foraging theory. Of all the major prey taxa commonly eaten by shrews, Coleoptera are among the most profitable in terms of high energy but low ash and water content (Cummins and Wuycheck, 1971; Churchfield, 1991). WTiile they ranked within the first three most frequently consumed prey in the diets of the eight terrestrial shrews studied here, and were preferred by S. araneus (Fig. 8), no species fed exclusively on these prey. The answer must lie in the encounter rate of such prey and the daily energy requirements of shrews, which dictate not only that a target number of prey must be captured daily in order to survive, but also that fr^uent meals are required. Profitable prey such as Coleoptera cannot be sufficiently abundant or catchable to enable them to serve as the sole dietary item. Although the probability of locating such a preferred prey may be high, the risk of not doing so may still be too great when starvation is imminent. Despite clumping of prey, shrew's are likely to encounter invertebrate taxa one at a time and thus not be in a position to choose between them. While they are searching for the most profitable prey, they encounter other invertebrates which may be worth eating as a short-term solution. Small invertebrates, for instance, are far more numerous than large ones. The mere abundance (and encounter rate) of Formicidae and Isoptera in tropical regions may explain why they are dominant prey for some soricids. Although individually they may be classed as unprofitable prey, collectively they have higher energy contents per gram than Isopoda or Diplopoda, and much lower ash contents (Cummins and Wuych«:k, 1971). Difficulty of capture may also be an important factor in prey selection. Many important and profitable prey, such as carabid and staphylinid beetles and Araneae, are strong, agile and fast- running. Following detection, successful capture rates may be low, and so shrews may have to rely on easier prey much of the time. Given their energetic constraints, shrews simply cannot afford to be too selective. Nevertheless, selection can still operate after a prey is captured, by eating only the most profi.tabIe parts of it. Observations of wild and captive shrews show that hungry individuals feed rapidly and fail to eat entire prey, discarding portions of legs, head capsules, and chitinous exoskeleton, and pressing on quickly to locate the next prey. This suggests that handling cost and profitability may indeed be taken into account, as optimal foraging theory suggests. Whether selectivity increases with increasing satiation in wild shrews is not known, but this seems likely. Less desirable prey may be eaten early in a foraging bout when the shrew is very hungry, but ignored in favor of more preferred prey as the risk of starvation recedes. All these factors would explain why shrews do not specialize more, and why shrew's of different body sizes do not exhibit more selection in their prey sizes, particularly the larger shrews. Wild shrews seem to operate on a strategy of partial selection: they eat more of the most profitable prey if they are available but do not ignore other prey. Prey Switching How do shrews react to declining availability of major prey items? Because shrews feed on common and usually abundant prey it is rare for a particular prey taxon not to feature in the diet at any one time. Nevertheless, seasonal changes in 1994 CHURCHFIELD — Foraging Strategies of Shrews 81 abundance do occur, and it may then be necessary for shrews to switch prey. An example of this is provided by the changing abundance of Coleoptera and the effect that this produced on the feeding habits of S. araneus. Density of these prey declined in winter. Their declining prominence in the diet was compensated for by an increase in the importance of certain other prey, namely Diptera larvae. Gastropoda, and Myriapoda. Individually none of these prey types compensated for the decline in Coleoptera, but collectively there was correlation between changing dietary occurrence of Coleoptera and the importance of these three alternative prey (Fig. 9). Again, no other prey type showed such a relationship with changing abundance. For example, Lumbricidae always featured prominently in the diet of S. araneus and, although they showed some seasonal variation in dietary occurrence, this was not consistent with changing importance of particular alternative prey. Again, Araneae were fairly consistent in the diet and their incidence was not related to other prey. By retaining a diverse diet the amount of prey-switching need only be small. A decrease of 25% in the dietary occurrence of Coleoptera as a result of declining abundance required a change of only 3 % in each of the eight other major prey taxa consumed to compensate for it, and this was hardly noticeable in diet samples. In the three taxa where a change was detected, only an 8% difference in each was required to compensate. Response to Changing Population Density and Competition Optimal foraging theory predicts that predators should become less selective in the presence of competition from conspecifics, and laboratory experiments have demonstrated this (Barnard and Brown, 1981). But there was no evidence of this in wild S. araneus in response to increasing population density and competition. The highly diverse diets of all shrews tend to obscure any possible differences in dietary composition within the same species, although changes in dietary diversity and niche overlap between species in different communities have been described (Churchfield, 1991). However, the abundance of invertebrates in many habitats suggests that intra- and interspecific competition may not occur (Churchfield, 1982; Churchfield and Brown, 1987). There is evidence that the availability of certain prey types affects the relative abundance of different shrew species. For example, S. araneus is rare in acid moorland while S. minutus is relatively abundant. This has been attributed to the dearth of earthworms, a major prey for S. araneus. Sorex minutus is able to subsist on small arthropods which are abundant in this habitat (Butterfield et al., 1981). Patch Use Most soricids appear to have stable home ranges which they occupy for most, if not all, their lives, although males may vacate their normal home ranges in response to the breeding season and the need to locate females (Pemetta, 1977; Churchfield, 1980fl). Social interactions, including competition for food, and habitat characteristics may also affect the use of home ranges (Platt, 1976; Hawes, 1977; Neet and Hausser, 1990). But there is no evidence that shrews move home ranges solely in response to food supply. For example, in a recent field study in which abundances of major prey types differed between neighboring areas, and where S. araneus and S. minutus were free to move among these areas, there was no evidence that the shrews were attracted to sites where Coleoptera, for example, were most abundant (Churchfield et al., personal observation). However, there is evidence that shrews use parts of their home ranges differentially in response to changing abundance and distribution of prey. They certainly respond to clumps of prey. This is clear enough during live-trapping studies when resident shrews can be attracted time after time to a food source in a trap. Experimental studies in outdoor enclosures with simulated natural environments also show that shrews will revisit or concentrate on sites where prey have recently been found (Churchfield, 19806). Even diet studies of wild shrews show this to be the case. Dipteran larvae, such as Bibionidae, have a clumped distribution in the soil around plant roots where they feed. Fecal analyses of S. araneus showed that these prey were eaten in considerable numbers whenever a patch was located. Up to 15 larvae were found per fecal sample. Attraction to clumped prey may also occur in tropical soricids, such as C. viaria, which feeds extensively on Isoptera and Formicidae. However, in the absence of field experiments to investigate patch or home range use in response to food distribution little more can be concluded. Conclusions All soricid species examined exhibited high dietary diversity, but with some specialization with respect to the dominant prey types exploited. While each species showed some specialization for a particular foraging mode, this was not exclusively used. Each species took a wide range of prey sizes. There was some selection for prey size but this was not strictly related to the body mass of the shrew. Prey selection was related more to availability and encounter rate than to size or food value as a guide to profitability. Elements of catchability and palatability also influence prey choice. Shrews are true opportunists. They show elements of specialization and partial selection for prey on the basis of dietary composition, foraging mode, size, and profitability, in addition to generalization. This permits rapid adaptation to spatial and temporal changes in prey availability. Acknowledgments I am grateful to Mr. J. Hollier, Dr. S. B. George, and Mr. R. Ford for their assistance in the field, and to Dr. T. Maddalena for the provision of diet samples from C. viaria. Financial assistance was provided by the University of London Central Research Fund and NERC. Literature Cited Barnard, C. J., and C. A. J. Brown. 1981. Prey size selection and competition in the common shrew Sorex araneus. Behavioral Ecology and Sociobiology, 8:239-243. 1985. Risk-sensitive foraging in common shrews (Sorex araneus). Behavioral Ecology and Sociobiology, 16:161-164. Bever, K. 1983. Zur nahrung der hausspitzmaus, Crocidura russula (Hermann, 1780). Saugetierkundliche Mitteilungen, 31:13-26. Butterfield, J., J. C. Coulson, and S. Wanless. 1981. Studies 82 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 on the distribution, food, breeding biology and relative abundance of the pygmy and common shrews {Sorex minutus and S. araneus) in upland areas of northern England. Journal of Zoology, (London), 195:169-180. Churchfield, S. 1980a. Population dynamics and the seasonal fluctuation in numbers of the common shrew in Britain. Acta Theriologica, 25:415-424. 1980fe. Subterranean foraging and burrowing activity of the common shrew. Acta Theriologica, 25:451-459. 1982. Food availability and the diet of the common shrew, Sorex araneus, in Britain. The Journal of Animal Ecology, 51:15-28. 1984. Dietary separation in three sjjecies of shrew inhabiting water-cress beds. Journal of Zoology (London), 204:211-228. 1985. The feeding ecology of the European water shrew. Mammal Review, 15:13-21. 1991. Niche dynamics, food resources and feeding strategies in multi-species communities of shrews. Pp. 23-34, in The Biology of the Soricidae (S. B. George, G. L. Kirkland, Jr., and J. S. Findley, eds.). Special Publication 1, Museum of Southwestern Biology, University of New Mexico Press, Albuquerque, 91 pp. Churchfield, S., and V. K. Brown. 1987. The trophic impact of small mammals in successional grasslands. Biological Journal of the Linnean Society, 31:273-290. COWIE, R. J. 1977. Optimal foraging in great tits (Parus major). Nature, 268:137-139. Crowcroft, P. 1957. The Life of the Shrew. Max Reinhardt, London. 166 pp. Cummins, K. W., and J. C. Wuycheck. 1971. Calorific equivalents for investigations in ecological energetics. Mitteilungen Internationale Vereinigung fiir Theoretische und Angewandte Limnologie, 18:1-158. French, T. W. 1984. Dietary overlap of Sorex longirostris and S. cinereus in hardwood floodplain habitats in Vigo County, Indiana. The American Midland Naturalist, 111:41-46. Grainger, J. P., and J. S. Fairley. 1978. Studies on the biology of the pygmy shrew, Sorex minutus, in the west of Ireland. Journal of Zoology (London), 186:109-141. Hamilton, W. J., Jr. 1930. The food of the Soricidae. Journal of Mammalogy, 11:26-39. Hawes, M. L. 1977. Home range, territoriality and ecological separation in sympatric shrews, Sorex vagrans and Sorex ohscurus. Journal of Mammalogy, 58:354-367. Hutterer, R. 1985. Anatomical adaptations of shrews. Mammal Review, 15:43-55. Krebs, J. R., J. T. Erichsen, M. 1. Webber, and E. L. Charnov. 1977. Optimal prey selection in the great tit (Parus major). Animal Behaviour, 25:30-38. Kuvikova, a. 1985. Zur nahrung der wasserspitzmaus, Neomys fodiens (Pennant, 1771) in der Slowakei. Biologia (Bratislava), 40:563-572. Neet, C. R., and j. Hausser. 1990. Habitat selection in zones of parapatric contact between the common shrew Sorex araneus and Millet’s shrew S. coronatus. The Journal of Animal Ecology, 59:235-250. Pernetta, j. C. 1976. Diets of the shrews Sorex araneus L. and Sorex minutus L. in Wytham grassland. The Journal of Animal Ecology, 45:899-912. 1977. Population ecology of British shrews in grassland. Acta Theriologica, 22:279-296. Pierce, G. J. 1987. Search paths of foraging common shrews Sorex araneus. Animal Behaviour, 35:1215-1224. Platt, W. J. 1976. The social organisation and territoriality of short- tailed shrew (Blarina brevicauda) populations in old-field habitats. Animal Behaviour, 24:305-318. RUDGE, M. R. 1968. Food of the common shrew Sorex araneus in Britain. The Journal of Animal Ecology, 37:565-581. Ryan, J. M. 1986. Dietary overlap in sympatric populations of pygmy shrews, Sorex hoyi, and masked shrews, Sorex cinereus, in Michigan. The Canadian Field-Naturalist, 100:225-228. Terry, C. J. 1981. Habitat differentiation among three species of Sorex and Neurotrichus gibbsi in Washington. The American Midland Naturalist, 106:119-125. Whitaker, J. O., Jr., S. P. Cross, and C. Maser. 1983. Food of vagrant shrews (Sorex vagrans) from Grant County, Oregon, as related to livestock grazing pressure. Northwest Science, 57:107-111. Whitaker, J. O., Jr., and C. Maser. 1976. Food habits of five western Oregon shrews. Northwest Science, 50:102-107. Whitaker, J. O., Jr., and T. W. French. 1984. Foods of six species of sympatric shrews from New Brunswick. Canadian Journal of Zoology, 62:622-626. Whitaker, J. O., Jr., and R. E. Mumford. 1972. Food and ectoparasites of Indiana shrews. Journal of Mammalogy, 53:329-335. Wolk, K. 1976. The winter food of the European water shrew. Acta Theriologica, 21:117-129. Table 1. — The soricid species examined in the present study including body sizes and geographic origins. Species Mean Body Mass (g) Collection Site Sample Size Sorex araneus 8.2 Southern England 240 Sorex minutus 3.5 Southern England 35 Neomys fodiens 11.9 Southern England 169 Sorex vagrans 4.3 Sierra Nevada, California 16 Sorex monticolus 5.5 Sierra Nevada, California 10 Sorex palustris 11.7 Sierra Nevada, California 15 Myosorex cafer 10.0 Eastern Zimbabwe 13 Crocidura hirta 10.0 Northern Zimbabwe 14 Crocidura viaria 14.3 Burkina Faso 10 Crocidura poensis 16.0 Southeast Nigeria 8 83 1994 CHURCHFIELD — Foraging Strategies of Shrews No. of samples successively examined Fig. 1. — The relationship between the number of diet samples examined and the number of prey taxa found. S.araneus S.minutus S. Vagrans S.monticolus S.palustris N.fodlens M.cafer C.hlrta C.viaria C.poensis 0 1 0 2 0 3 0 — I 4 0 No. of prey taxa Fig. 2.— Dietary diversity and the number of dominant prey types in ten species of shrews. Hatched bars represent the total number of prey taxa identified; solid bars represent the number of dominant taxa. 84 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 S.araneus S.mlnutus S.vagrans S.monticolus M.cafer C.hirta C.viarla '/////////////////y/z/A ^^^7777777777777777777777^ y/////////////////////////////'//////A Y777777777ZA77777777Z^aA7777. y7/////7////7////////////7/A '/////////////////////////////////A C.poensis V/////////y/A/7////////////////A//////A — I ' ' 1 — 2 0 4 0 6 0 8 0 1 00 % Dietary occurrences Fig. 3. — Foraging modes of terrestrial shrews; the percentage dietary occurrences of surface-dwelling (epigeal: hatched bars) and soil-dwelling (hypogeal: solid bars) prey in the diets of eight terrestrial soricids. Aquatic Epigeal Hypogeal Fig. 4. — Foraging modes of two semiaquatic soricids, Neomys fodiens and Sorex palustris. % Dietary composition Dietary composition % Dietary composition 1994 CHURCHFIELD— Foraging Strategies of Shrews 85 Small shrews (3-6g) M S.minutus ■ S. Vagrans Q S.montlcolus Medium shrews (8-11 g) 0 S.araneus ■ M.cafer □ C.hirta Body length of prey (mm) Large shrews (14-17g) 0 C.viaria ■ C.poensfs Body length of prey (mm) Fig. 5.— The percentage composition by volume of prey of different body lengths in the diets of small, medium, and large shrews. SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 86 (U o c (U 3 o o o >. L. nj ■a 40 -I A Coleoptera B Hemiptera C Diptera larvae D Lepidoptera larvae E Myriapoda F Araneae G Isopoda H Gastropoda J Lumbricidae Prey type Fig. 6.— The mean percentage dietary occurrences of nine major prey types in the diets of S. araneus (shaded bars) and S. minutus (hatched bars) over a three-year period, with standard errors. • Diet 0--0 Pitfalls Fig. 7.— The percentage dietary occurrences of adult Coleoptera in the diet of S. araneus, and their incidence in pitfall samples. No. in pitfalls 1994 CHURCHFIELD — Foraging Strategies of Shrews 87 Fig. 8. — The relationship between the availability of Coleoptera and their occurrence in the diet of S. araneus. Fig. 9. — Prey switching in S. araneus: declining occurrences of Coleoptera in the diet were compensated for by increases in other prey (Diptera larvae, Gastropoda, and Myriapoda). -I:- .■7<''iyW'.)i ..*i v'*'* .v: O I I THE TERRITORIAL AND DEMOGRAPHIC STRUCTURES OF A COMMON SHREW POPULATION Ernest V. Ivanter^ Tatjana V. Ivanter’, and Alexander M. Makarow* ’Department of Zoology, University of Petrovodsk, 185035 Petrovodsk, Russia Abstract Territorial and demographic structure (age and sex) were determined in a long-term population study of the common shrew {Sorex araneus L.) using live-trapping, mark, and recapture techniques. Territorial behavior among overwintered females varied based on age and reproductive activity. Home ranges of overwintered females averaged 1300 m^ (range, 800-1700 m^). After the reproductive period, females began to move and home range structure disappeared. Home ranges of settled juveniles (young-of-the-year) overlapped significantly; their mean area in different years varied from approximately 360 to 500 m^. Concomitantly, there were young transient individuals in the population without home ranges. Contrary to previous observations that shrews move randomly, Sorex araneus in this study had distinct home ranges but these often overlapped. Introduction Rodents and shrews inhabit forests and play essential but different roles in terrestrial ecosystems. Territorial behavior, although often studied in rodents, has been investigated very little in shrews. Data on shrews are scarce and often contradictory. Karasewa (1955) and Nikitina and Korchagina (1966) described shrews as wandering animals, not associated with any territory, and having no defined home ranges. More recently, other investigators using live traps have concluded that shrews do have home ranges and defend territories (Shillito, 1963; Michielsen, 1966; Yalden, 1974; Pemetta, 1977; Ellenbroek, 1980). Radioactive marking has been used to evaluate daily activity of shrews only with respect to short-term stays within areas. More recently, studies of territoriality of shrews in the Enisey taiga (Moraleva and Sheftel, 1980; Moraleva, 1983) have provided information on the influence of season and age on this phenomenon. The aim of the present work was to investigate territoriality in a population of the common shrew {Sorex araneus L.) in the taiga of northwestern Russia. Materials and Methods This study was conducted from 1986 to 1989 on the northeastern shore of Ladoga Lake in southern Karelia near the Russia-Finland border (61°22’N, 31®58’E). The territorial behavior of small mammals was evaluated based on live- trapping of marked animals in experimental enclosures (Ivanter and Makarow, 1988). Study sites were secondary forests near forest Lake Kurkunlampi. Plant cover was a mosaic of spruce, birch, and mixed forest associations: '"Piceetum oxalidosum," ''Betuletum myrtilloso-mixtoherbosum," “Betuletum graniinoso- myrtillosum," and “'Piceeto-pinetum myrtilloso-herbosum." Traps were open from June to September during dry weather and were closed during rain and very cold weather to prevent mortality of shrews. Animals were sampled in the periods 26-30 June, 10-19 July, and 1 1-22 August 1986; 1 8-24 August 1987; 16 June-20 August 1988; 21-26 June, 13-18 July, 17-22 August, and 8-13 September 1989. Spring-loaded live traps were used and baited with rye bread and sunflower oil. The study plots were quadrats of 1 ha (1986-1987) and 2 ha (1988-1989). Traps were set in a grid with 10 m intervals between traps. Traps were checked every two hours during daylight hours and were closed at night. Captured animals were marked by toe-clipping. Records were made of sex, age, weight, reproductive condition, date, and time and point of capture. The inclusive boundary strip method (Burt, 1943; Evans and Holdenreid, 1943; Stickel, 1954) was used to determine the home ranges. Single captures of the animals outside their own ranges, and captures of dispersing individuals, were not taken into account. Live-trapping data were supplemented by data on relative abundance from snap-trap lines and pitfall traps operated concurrently in the study area. Results and Discussion General Characteristics of the Population The number of Sorex araneus marked varied greatly from year to year (Table 1). Because of the border effect, the total number of animals marked in the study area was greater than the actual population density. The proportion of animals caught several times varied from year to year, ranging from 58 to 77%. Thus, temporary inhabitants of the trapping area should not have greatly influenced estimates of abundance. Using the number of shrews in the sampling grid in August as an index of shrew density, we concluded that population density exhibited a 12-fold change during the five years of observations (Table 1). Shrew captures were highest in 1986 and 1989 and lowest in 1987; 1988 marked the beginning of an increase in shrew abundance. Differences in the seasonal dynamics of shrews between the two peak years (1986 and 1989) are explained by unusual weather conditions in 1989; an early and warm spring in 1989 resulted in earlier reproduction. Data on abundance obtained from live-trapping were consistent with data from pitfall and snap-trap sampling (Table 2). The population of shrews changed gradually during the summer-autumn period (Fig. 1). Distinct differences were obvious between July and August and were related to high mortality of adults marked in June. The proportion of newly- caught animals decreased gradually, reaching a minimum in September after the end of reproduction. The sex and age structure of the population varied only 89 90 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 slightly by year (Fig. 2). In general, the proportion of animals that overwintered did not exceed 25% of the total number of marked animals. The habitat distribution of animals did not vary significantly through the years. The mosaic nature of the vegetation on the study area smoothed differences in numbers, because home ranges of many animals were located in two or three neighboring habitats. Thus, estimation of habitat differences on the basis of number of captures was more illustrative than by number of individuals marked in each habitat. The greatest number of captures was in the meadow-sweet brush woods (in August 1989 with an average of 3.1 captures per trap). There were fewer shrews in the mixed forest (1.3 captures per trap). Ferny birch forest and thick spruce forest with needle fall on the ground were poorly settled (0.5 captures per trap in both forest types). Occupation of the latter habitat started only in July, and in June there were only 0.2 captures per trap. Increased abundance of shrews was accompanied by increased numbers of animals captured in every habitat, but to a greater extent in preferred ones. Variation of Territorial Behavior in Animals of Different Sex and Age The majority of overwintered shrews disappeared from the study site during June and July, and the number of overwintered shrews did not exceed ten specimens per hectare in August. The behavior of overwintered males was very different from that of overwintered females. Overwintered Females. — Movements of overwintered females were less than movements of any other age and sex group. Overwintered females had well-defined home ranges which averaged 1300 m^ (range 800-1700 m“, n = 9). Home ranges were observed only in the most highly populated areas, and home ranges rarely overlapped (Fig. 3). As a rule, only one female was caught at a given trap. Traps visited by two females never exceeded 10% of all traps that caught overwintered females. Only once were three different females caught in the same trap. The majority of adult females spent all the reproductive period in the same home range, but the configuration of home ranges varied. Some females changed their home range during the interval between litters, but new home ranges were in a state of dynamic equilibrium, that is, as one range changed, the borders of the surrounding ranges also changed. Consequently, minimal overlap of ranges was maintained, suggesting that the home ranges functioned as territories. After the reproductive season, some females left their ranges and either moved randomly over surrounding territories or left the area entirely (Fig. 4). Overwintered Males. — Overwintered males were the most mobile members of the population. The majority did not have a consistent home range and moved randomly over the study plot, often through the home ranges of several females. Habitat preference was not well -ex pressed, although capture sites were often in low-lying areas. The distance between points of sequential captures was often 100 m and more. Territorial behavior of males ranged between two extremes: 1) chaotic movements not connected with specific home ranges. with animals present on the study grid for very limited periods of time (Fig. 5); and 2) animals limited to large but defined home ranges. Males basically were segregated from each other and seldom shared home ranges, once again suggesting the presence of territoriality. In 1989, of 40 traps in which males were captured, 33 traps caught only one individual, four caught two different individuals, one caught three, and only two caught four different individuals. The-Young-of-the-Year (Subadults). — Movements were variable among subadult shrews. Some had small, well-defined home ranges; others moved randomly within the study grid. Analysis revealed a relationship between these two patterns. Young bom early in the year tended to occupy distinct home ranges. Specimens initially captured early in summer established their home ranges almost at once and occupied them throughout the period of observation. Sometimes individuals left their home range briefly but always returned (see Fig. 6, specimen no. 80). Occasionally, a home range changed, but the major part of it remained constant. In contrast, individuals that appeared on the study grid later in the summer did not have as much of a choice of home ranges. During years of high populations, ail suitable habitat was occupied by resident shrews which forced young shrews to move actively in search of unoccupied habitat in which to establish home ranges. There are three possible methods for acquiring a home range: 1) find and occupy a free space in favorable habitat, 2) occupy an established home range (see Fig. 6, animal no. 142), or 3) establish a home range in suboptimal habitat. Young shrews searching for unoccupied space generally moved through the most favorable habitats. As a result of differences in movements between resident and dispersing individuals, there were differences in numbers of shrews in various habitats. The occurrence of an animal in suboptimal habitat (e.g., the fem-birch forest) usually did not result in establishment of a home range. Young of the year found moving through the study site were those who did not find unoccupied habitat during the period of observation (Fig. 6, specimen no. 270). During the gradual occupation of the study site, all favorable habitats became occupied and were divided into individual home ranges. The mean area of the home ranges in different years ranged from 360 m^ to 500 m^. In years of high shrew density, establishment of home ranges occurred earlier, usually in July (Fig. 7), and in years of low numbers, it occurred in August and September. Sometimes the full occupation of all suitable habitats did not take place due to extremely low numbers of animals (Fig. 8). The number of young-of-the-year inhabiting the study area changed with time (Fig. 1). The change occurred because of mortality of animals or their disappearance from the study area. With the appearance of new individuals, the borders of previously established home ranges changed. This produced the impression of substantially overlapping home ranges of neighbors, whereas in reality the home range of the earlier occupant was compressed. Conclusions The results of the study suggest that a population of the common shrew, Sorex araneus, is a complex and dynamic 1994 IVANTER ET AL. — Territoriality and Demography of Common Shrews 91 system, in which groups of individuals of different ages and sexes coexist and exhibit differential territorial behavior. The population of the common shrew in southern Karelia in this study was less stable than populations investigated in Holland (Michielsen, 1966) and England (Shillito, 1963; Buckner, 1969; Pemetta, 1977). However, animals in the present study were more closely associated with apparent territories than were those in the taiga forests of Siberia (Moraleva, 1983, 1988, 1989). The unique characteristic of the population investigated here was that home ranges of overwintered females and young- of-the-year in the same preferred habitats were distributed independently and overlapped greatly. Literature Cited Buckner, C. H. 1969. Some aspects of the population ecology of the common shrew, Sorex araneus, near Oxford, England. Journal of Mammalogy, 50:326-332. Burt, W. H. 1943. Territoriality and home range concepts as applied to mammals. Journal of Mammalogy, 24:346-352. Ellenbroek, F. J. 1980. Interspecific competition in the shrews Sorex araneus and Sorex minulus (Soricidae, Insectivora): A population study of the Irish pygmy shrew. Journal of Zoology (London), 192:119-136. Evans, F. C., and R. Holdenreid. 1943. Beechey ground squirrel. Journal of Mammalogy, 24:63-64. IVANTER, E. V., AND A. M. M AKA ROW. 1988. Territorial interactions in populations of shrews. Pp. 81-83, in Ecology of Populations (no editor. Symposium in Novosibirsk, USSR). Nauka, Moscow, 250 pp. (in Russian). Karasewa, E. V. 1955. Marking of the ground mammals in USSR. Soviet Bulletin of Moscow Society of Investigators of Nature, Biology, 60:32-42 (in Russian). Michielsen, N. C. 1966. Intraspecific and interspecific competition in the shrews Sorex araneus L. and Sorex minulus L. Archives Neerlandaises de Zoologie, 17:73-174. Moraleva, N. V. 1983. Relation to territory of shrews in Enisej taiga (by data of individual marking). Pp. 215-230, in Animals in Enisej Taiga (E. E. Shroechkovsky , ed.), Nauka, Moscow, 232 pp. (in Russian). 1988. Some peculiarities of territorial structure in population of common shrew in the years of high number. Pp. 48-49, in Species and Its Productivity in the Distribution Area (R. S. Volskis, ed.), Lithuanian Academy of Sciences, Vilnius, 308 pp. (in Russian). 1989. Intraspecific interactions in the common shrew Sorex araneus in central Siberia. Annales Zoologica Fennici, 26:425-432. Moraleva, N. V., and B. I. Sheftel. 1980. Some peculiarities of territorial structure of shrew population in middle Enisei taiga. Pp. 9-13, in Some Aspects of Study of Flora and Fauna in USSR, Nauka, Moscow, 210 pp. (in Russian). Nikitina, N. A., and L. D. Korchagina. 1966. Use of territory in shrews characterize by marking. Soviet Bulletin of Moscow Society of Investigators of Nature, Biology, 71:26-31 (in Russian). Pernetta, I. C. 1977. Population of ecology of the British shrews in grassland. Acta Theriologica, 22:279-296. Shillito, J. F. 1963. Observation on the range and movements of a woodland population of the common shrew, Sorex araneus L. Proceedings of the Zoological Society of London, 140:533-546. Stickel, L. F. 1954. A comparison of certain methods of measuring ranges of small mammals. Journal of Mammalogy, 35:1-15. Yalden, D. W. 1974. Population density of the common shrew Sorex araneus L. Journal of Zoology (London), 173:262-264. Table 1. — Number of common shrews marked on the study grid. The study area was 1 ha in 1986-1987 and 2 ha in 1988-1990. x = no sampling. Year June July August September 1986 8 48 44 X 1987 X 1 3 X 1988 X 22 39 X 1989 79 113 73 60 1990 15 X 25 X Table 2.— Mean number of common shrews in forests of the Ladoga Lake region per 100 snap trap nights and 10 pitfall nights in 1985-1990. Method of Sampling 1985 1986 1987 1988 1989 1990 Snap trap lines 3.6 6.5 2.2 2.3 10.0 2.2 Pitfall traps 9.2 10.4 3.2 4.2 9.0 no trapping 92 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 Fig. 1. — Sex and age structure of the common shrew population from 1986-1989. 3 O. O CL LJL O LJJ CC 3 f— o 3 CC i — OD a = 79 n=83 n^59 June Jufy Jla^ust Sepiem6eR Fig. 2.— Change of structure of the common shrew population on the study grid from June through September, 1989 (in all age groups). The individuals were marked in: 1) June, 2) July, 3) August, 4) September. 1994 IVANTER ET AL. — Territoriality and Demography of Common Shrews 93 Fig. 3. — Home ranges of overwintered females in June (left) and July 1989. 1) stations of capture, 2) boundaries of home ranges. Fig. 4. — Home range of overwintered female no. 5 in June-July and its movements in August 1986. See the legend of Fig. 3. 94 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 Fig. 5. — Two types of territorial behavior of males (22 June- 18 July 1989). 1) random movements (no. 55), 2) extended stay within definite territory (no. 28). Fig. 6. — Location of capture of young-of-the-year marked in June (no. 80, 23.06-11.09), July (no. 142, 16.07-14.09), and August (no. 270. 21.08-8.09) 1989. 1994 IVANTER ET AL. — Territoriality and Demography of Common Shrews 95 j^Om Fig. 7.— Home ranges of young-of-the-year in July 1989 (high shrew density). 1) imaginary overlap of home ranges caused by the disappearance of animal no. 78 during study, 2) space occupied by animals. Fig. 8. — Home ranges of young-of-the-year in June (left) and July (right) 1988 (low shrew density). 1) imaginary overlap of home ranges as result of the compression of the home range of animal no. 17. I-Vf V ^' / • >■ ■'t'^ '‘^jr-; 1 i ■»;■' }."&'■ ■‘■’■■‘ ':i ■ ' • ■r-*i^ V ■■ .r*'" n' .»> ,• If’- "a »> i** • - 5^j t 0< . •' -••• 'I ;. ■- .1-, ' 'i' &T Of; ^ > - , PARASITISM BY GASTROINTESTINAL HELMINTHS IN THE SHREWS SOREX ARANEUS AND S. CAECUTIENS Vonro Haukisalmi\ Heikki Henttonen“, and Taina Mikkonen* ’ Department of Zoology, Division of Ecology, P. O. Box 17 (P. Rautatiekatu 13), FIN-00014 University of Helsinki, Finland; ^ Finnish Forest Research Institute, Department of Forest Ecology, P. O. Box 18, SF-01301 Vantaa, Finland Abstract We studied parasitism by gastrointestinal helminths (trematodes, cestodes, and nematodes) in Sorex araneus (« = 1 14) and Sorex caecutiens (n = 105) in Finnish Lapland. The large-sized S. araneus had significantly higher overall infection levels than the smaller species 5. caecutiens, even when the size difference of shrews was taken into account. Similarly, most helminth species were significantly more prevalent in S. araneus than in S. caecutiens. Adult males tended to have higher overall infection levels than juvenile males, but significant differences between the sexes were few. Infection level/body size ratios, possible indicators of parasite pathogenicity, did not differ consistently between the age groups or sexes of shrews. Several helminth species, especially cestodes, showed significant differences in prevalence between age groups, but not between sexes of shrews. We were unable to find a clear effect on weight of shrews due to heavy helminth infections. Introduction The most conspicuous features of shrews of the genus Sorex are their small body size and high metabolic rate (Vogel, 1976). Since shrews also have small energy reserves (Vogel, 1980), the high metabolic rate inevitably implies severe energetic constraints and high risk of starvation. One of the effects of gastrointestinal parasites on hosts may be decreased efficiency of absorption and digestion, which can be compensated for by increased food consumption or use of energy reserves (Munger and Karasov, 1989). Adverse effects of parasites could be especially severe for shrews, as shrew populations seem to be largely food limited (Kaikusalo and Hanski, 1985) and probably experience frequent energy crises. Furthermore, the common shrew Sorex araneus, which has unusually high burdens of gastrointestinal helminths (Haukisalmi, 1989), is expected to be more severely affected than other species of Sorex. Hanski (1989) suggested that heavy investment in reproduction may increase the vulnerability of adult male shrews to biotic and abiotic factors. For example, adult males seem to have a low social position in the population (Moraleva, 1989). Increased susceptibility to parasitism or increased pathogenicity of parasites may be another consequence of intense reproductive effort. As an extreme example, Lee and Cockbum (1985) have shown that because of very intense reproduction, males of some Antechinus species (insectivorous marsupials) are attacked by various pathogens and parasites, which contributes to the disappearance of males after the mating season. Female investment in pregnancy and lactation increases their energy requirements (Glazier, 1985) and food consumption, which may also lead to increased levels of parasitism. We describe herein parasitism by gastrointestinal helminths in Sorex araneus and S. caecutiens in Finnish Lapland. Specifically, we examine differences in infection levels between the species (Haukisalmi, 1989), the level of parasitism in relation to age and sex of shrews, and the potential of endoparasites to affect the condition of shrews. In addition to the number of parasite species and individuals, we present data on helminth biomass; such data could elucidate the energetic constraints of parasites on shrews. Materials and Methods One hundred fourteen common shrews (Sorex araneus) and 105 masked shrews (S. caecutiens) were trapped in 1988-1990 at Pallasjarvi (68°03’, 24°09’), western Finnish Lapland. Most of the specimens (S. araneus, 78%; S. caecutiens, 70%) were obtained between June and October; only a few were caught in midwinter (December-February). The shrews were collected in old taiga forests characterized by a thick moss layer and dominance of spruce (Picea abies) and blueberry ( Vaccinium myrtillus) (for details, see Henttonen et al., 1987, 1989). The shrews were usually found dead in live-traps, which were primarily used for catching voles as the interval between checking traps was mostly six and sometimes eight hours. After capture, the shrews were frozen for later examination. The contents of stomach and intestines were searched for helminths under a binocular microscope. Identification of parasites was based on the monographs of Zamowski (1960), Vaucher (1971), Vaucher and Durette-Desset (1973), and Genov (1984). Earlier data on shrew helminths in Finland was provided by Vaucher (1971), Vaucher and Durette-Desset (1973), and Haukisalmi (1989). Multiway contingency tables (log-linear models, Fienberg, 1970) were used to analyze dependence between the occurrence of the common helminths (H), and age (A) and sex (S) of shrews. For each parasite species we selected the best ( = simplest) model that fit the observed data (x^ test, P > 0.05). Fienberg (1970) and Harris (1984) described the process of selecting the best model in multiway contingency tables. In log- linear models interactions are indicated by combined variables (e.g., AH), and lack of interactions by separating the variables with a comma (e.g., A,H). The highest order model (SAH), which includes all possible interactions, is accepted if its fit to the data is significantly (P < 0.05) better than the fit of model SA, AH, SH, which includes all but the highest order effect (Harris, 1984). The number and volume of all helminths and the number of helminth species in each host were used to describe overall 97 98 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 infection levels. Ratios of these parameters to weight (without alimentary tract) of shrews were also calculated. Davidson et al. (1980) have shown these ratios to be good indicators of parasite pathogenicity in white-tailed deer (Odocoileus virginiana). Significance of differences between and within shrew species in overall infection levels was checked by Kruskal-Wallis test. The index of helminth volume was obtained as a product of mean length and width of helminths, based on the measurements of ten individual worms or on published data. This measure of volume is obviously crude but is satisfactory for the present purpose. Because of their highly non-normal distributions, logarithmic transformations were performed on the number and volume of parasites in partial correlation analyses. Results Differences Between Species Seventeen species of gastrointestinal helminths parasitized Sorex araneus and S. caecutiens at Pallasjarvi, including four species of trematodes, nine of cestodes, and four of nematodes (Table 1). All 17 species parasitized S. araneus, whereas only 13 species were found in S. caecutiens. The nematodes of the genus Longistriata were the most prevalent helminths in both shrew species. Most of the helminth species had significantly higher prevalence in S. araneus than in S. caecutiens. The cestode Hymenolepis schaldybini was the only species having higher prevalence in S. caecutiens (Table 1). The overall infection levels also tended to be higher in S. araneus, especially in juveniles (Table 2). In adult males, some parameters were higher in S. araneus, but in adult females there were no significant interspecific differences. Differences Between Age Groups and Sexes The number of helminth species (5. caecutiens), number of individual parasites (both shrew species) and volume of helminths (S. caecutiens) per host were significantly higher in adult males than in juvenile males (Table 2). However, when infection indices were expressed as ratios of host body weight, adult males rarely showed higher infection levels than juveniles. Furthermore, the number of parasite species per host weight was significantly higher in juvenile males and females of S. araneus than in adults; such differences were not observed in the original values. In females there were no significant differences between age groups in the original infection indices. Adult males showed consistently higher overall infection levels than adult females, but significant differences between the sexes were rare. The log-linear models showed that occurrence of the common cestodes depends on the age, but usually not on the sex, of shrews (Table 3). Choanotaenia crassiscolex and Hymenolepis sp. had higher prevalence in juveniles, whereas H. schaldybini , H. scutigera, and H. infirma were more prevalent in adults (Fig. 1). The full-order model SAH for H. infirma (host S. araneus) was accepted, since it gave a significantly (x“ = 8.3, d.f. = 1, < 0.01) better fit to the data than the model SA, AH, SH. In H. infirma, the age-dependence differed between male and female shrews so that adult males, but not females, had higher prevalence than juveniles. The occurrence of nematodes was independent of age and sex of the host, with the exception of Longistriata pseudodidas in S. caecutiens (Table 3, Fig. 1). Infection Parameters and Weight of Shrews Because infection level may be determined by the size of the host, comparisons between weight and parasite burden could be confounded by the effect of varying size of shrews. Therefore, body length of shrews was controlled in correlation analyses between the various infection parameters and weight of shrews (partial correlation). Since the weight and body length of males and females did not differ significantly in either of the shrew species, sexes were pooled. Significant correlations between overall infection parameters and weight of shrews were few (Table 4), i.e., no consistent evidence of impaired condition of shrews due to gastrointestinal parasites was found. We also performed partial correlation analyses for the weight of juvenile shrews and various helminth species with at least ten occurrences (zero observations were excluded), but none of the correlations was found to be significant. Discussion The results confirm earlier observations by Haukisalmi (1989) that Sorex araneus has much higher infection levels of gastrointestinal helminths than S. caecutiens. The larger size of S. araneus implies higher absolute food requirements (Hanski, 1984), which should increase the colonization rate of helminths. In addition, S. araneus is expected to have higher numbers of helminths because of its long and voluminous intestinal tract. However, when the size difference between the two shrew species is taken into account, S. araneus still shows significantly higher infection levels than S. caecutiens, especially in juveniles. This suggests that high infection levels in S. araneus are due to either its generalized diet or high abundance (Haukisalmi, 1989), rather than to large body size. In addition, harmful effects of parasites may be greater in S. araneus than in S. caecutiens (c.f. Davidson et al., 1980). Soveri et al. (1994) have shown that histopathological changes, some of which could be caused by helminths, are very common in S. araneus. Unfortunately, we do not know the prevalence of such changes in S. caecutiens. Hanski (1989) raised the question of the role of parasites in permitting the coexistence of competing shrew species. Our findings favor the idea that coexistence is facilitated by large species of shrew being more severely affected by parasites. This conclusion is supported by the fact that interspecific differences in infection levels were most pronounced in juveniles, which show strict interspecific territoriality (Groin Michielsen, 1966; Hawes, 1977). On the other hand, S. araneus is expected to be better adapted to helminths than the small species since the bulk of the helminth population circulates through it (Hanski, 1989; Haukisalmi, 1989). Adult males, especially of S. caecutiens, had higher overall infection levels than juvenile males, but in females there were 1994 HAUKISALMI ET AL. — Parasitism of Shrews by Gastrointestinal Helminths 99 no significant differences between age groups. This seems to support the idea that the intense reproductive effort of adult males is accompanied by a high risk of being affected negatively by biotic factors (Hanski, 1989). The data presented by Kisielewska (1961) and Erkinaro and Heikura (1977) also showed that adult males are especially susceptible to infection by intestinal and extra-intestinal endoparasites, respectively. If the high overall infection levels of adult males were due to intense reproduction resulting in impaired resistance, there should be increased infection levels in most helminth species, including species with direct (nematodes) and indirect (cestodes) transmission. The patterns of age dependence of parasite species were variable: some cestodes were most prevalent in adult shrews, some in juveniles, and the occurrence of nematodes usually did not depend on the age of host. High infection levels thus seem to reflect specialization of particular helminth species on adult shrews, rather than general impairment in resistance of adults. For example, high overall infection in adult males (Table 2) seems to be due primarily to a single cestode species, H. infinna, which occurs in very high numbers in both shrew species (Table 1) and which is most prevalent in adult males (Fig. 1). Factors not directly related to reproductive effort, i.e., diet, longer exposure time, and greater amounts of food consumed, could also contribute to high infection levels in adult shrews. Kisielewska (1961) suggested that susceptibility to endoparasite infections could contribute to the rapid disappearance of males of S. araneus after the breeding season (Moraleva, 1989; see also Lee and Cockbum, 1985). However, the infection level /body weight ratios, which are possible indicators of parasite pathogenicity (Davidson et al., 1980), suggest that harmful effects of helminths are not more severe in adult shrews than in juvenile shrews. Studies of the effect of helminths on the condition and weight of hosts have yielded contradictory results. Low body weight of snowshoe hares (Keith et al., 1985, 1986) and rabbits (Yuill, 1964; Jacobsen et al., 1978; Dunsmore, 1981) were associated with the presence of high numbers of helminths. On the other hand, some studies report only minor changes in physiological parameters of hosts (cotton rat, Briese and Smith, 1980; white-footed mouse, Munger and Karasov, 1989) or practically no effect at all (mountain hare, lason and Boag, 1988; willow ptarmigan, Thomas, 1986). Our analysis showed that within shrews of equal size the body weight generally did not correlate significantly either with the overall infection levels or infection levels of various helminth species. We could not find any support for the assumption that biomass of parasites has more severe energetic effects on shrews than the number of parasite species and individuals. We conclude that gastrointestinal helminths do not have detectable effects on the condition of shrews. Acknowledgments We thank I. Hanski, J. Heikkila, and E. Pankakoski for commenting on an earlier draft of the manuscript, and G. L. Kirkland, Jr., and J. O. Whitaker, Jr., for critically reviewing the final version. Our helminth studies in Finland have been financed primarily by the Emil Aaltonen Foundation, the Research Council for Natural Sciences in Finland, and the Oskar Oflund Foundation. Literature Cited Briese, L. A., and M. H. Smith. 1980. Body condition, elemental balance, and parasitism in cotton rats. Journal of Mammalogy, 61:760-763. Croin Michielsen, N. 1966. Intraspecific and interspecific competition in the shrews Sorex araneus L. and S. minutus L. Archives Neerlandaises de Zoologie, 17:73-174. Davidson, W. R., M. B. McGhee, V. F. Nettles, and L. C. Chappel. 1980. Haemonchosis in white-tailed deer in the southeastern United States. Journal of Wildlife Diseases, 16:499-508. Dunsmore, J. D. 1981. The role of parasites in the population regulation of the European rabbit {Oryctolagus cuniculus) in Australia. Proceedings of the Worldwide Furbearer Conference, 2:654-669. Erkinaro, E., and K. Heikura. 1977. Dependence of Porrocaecum sp. (Nematoda) occurrences on the sex and age of the host (Soricidae) in northern Finland. Aquilo Serie Zoologica, 17:37-41. FieNBERG, S. E. 1970. The analysis of multidimensional contingency tables. Ecology, 51:419-433. Genov, T. 1984. Helminths of Insectivorous Mammals and Rodents in Bulgaria. Publishing House of the Bulgarian Academy of Sciences, Sofia, 348 pp. (in Bulgarian, English summary). Glazier, D. S. 1985. Energetics of litter size in five species of Peromyscus with generalizations for other mammals. Journal of Mammalogy, 66:629-642. Hanski, I. 1984. Food consumption, assimilation and metabolic rate in six species of shrew (Sorex and Neomys). Annales Zoologici Fennici, 21:157-165. 1989. Population biology of Eurasian shrews: Towards a synthesis. Annales Zoologici Fennici, 26:469-479. Harris, J. H. 1984. An experimental analysis of desert rodent foraging ecology. Ecology, 65:1579-1584. Haukisalmi, V. 1989. Intestinal helminth communities of Sorex shrews in Finland. Annales Zoologici Fennici, 26:401-409. Hawes, M. L. 1977. Home range, territoriality and ecological separation in sympatric shrews, Sorex vagrans and Sorex obscurus. Journal of Mammalogy, 58:354-367. Henttonen, H., T. Oksanen, a. Jortikka, and V. Haukisalmi. 1987. How much do weasels shape microtine cycles in the northern Fennoscandian taiga? Oikos, 50:353-365. Henttonen, H., V. Haukisalmi, A. Kaikusalo, E. Korpimaki, K. Norrdahl, and U. a. P. SkareN. 1989. Long-term population dynamics of the common shrew Sorex araneus in Finland. Annales Zoologici Fennici, 26:349-355. Iason, G. R., and B. Boag. 1988. Do intestinal helminths affect condition and fecundity of adult mountain hares? Journal of Wildlife Diseases, 24:599-605. Jacobson, H. A., R. L. Kirkpatrick, and B. S. McGinnes. 1978. Disease and physiologic characteristics of two cottontail populations in Virginia. Wildlife Monographs, 60:1-53. Kaikusalo, A., and I. Hanski. 1985. Population dynamics of Sorex araneus and S. caecutiens in Finnish Lapland. Acta Zoologica Fennica, 173:283-285. Keith, L. B., J. R. Cary, T. M. Yuill, and I. M. Keith. 1985. Prevalence of helminths in a cyclic snowshoe hare population. Journal of Wildlife Diseases, 21:233-253. Keith, I. M., L. B. Keith, and J. R. Cary. 1986. Parasitism in a 100 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 declining population of snowshoe hares. Journal of Wildlife Diseases, 22:349-363. Kisielewska, K. 1961. Circulation of tapeworms of Sorex araneus araneus L. in biocenosis of Bialowieza National Park. Acta Parasitologica Polonica, 9:331-369. Lee, K. L., and A. Cockburn. 1985. Evolutionary ecology of marsupials. Cambridge University Press, Cambridge, 274 pp. Moraleva, N. V. 1989. Intraspecific interactions in the common shrew Sorex araneus in central Siberia. Annales Zoologici Fennici, 26:425-432. Munger, J. C., and W. H. Karasov. 1989. Sublethal parasites and host energy budgets: Tapeworm infections in white-footed mice. Ecology, 70:904-921. Soveri, T., E. Rudback, and H. Henttonen. 1994. Pp. 151-153, in Advances in the Biology of Shrews (J. F. Merritt, G. L. Kirkland, Jr., and R. K. Rose, eds.), Carnegie Museum of Natural History Special Publication 18, i-x + 458 pp. Thomas, V. G. 1986. Body condition of willow ptarmigan parasitized by cestodes during winter. Canadian Journal of Zoology, 64:251-254. Vaucher, C. 1971. Les cestodes parasites des Soricidae d’Europe, etude anatomique, revision taxonomique et biologic. Revue Suisse de Zoologie, 78:1-113. Vaucher, C., and M.-C. Durette-Desset. 1973. Nematodes heligmosomes parasites d’insectivores soricides de la region Holarctique. Annales de Parasitologic Humaine et Comparee, 48:135-167. Vogel, P. 1976. Energy consumption of European and African shrews. Acta Theriologica, 21:195-206. 1980. Metabolic levels and biologic strategies of shrews. Pp. 170-180, in Comparative Physiology: Primitive Mammals (K. Schmidt-Nielsen, L. Bolis, and C. R. Taylor, eds.), Cambridge University Press, Cambridge, 338 pp. Yuill, T. M. 1964. Effects of gastrointestinal parasites on cottontails. The Journal of Wildlife Management, 28:20-26. ZarNOWSKI, E. 1960. Parasitic worms of forest micromammalians (Rodentia and Insectivora) of the environment of Pulawy (district Lublin). II. Trematoda. Acta Parasitologica Polonica, 8:127-167. Table 1. — Gastrointestinal helminths recovered from Sorex araneus (n = 114) and S. caecutiens (n = 105) at Pallasjdrvi, optimum microhabitats , median volume index, number of infected hosts (n), and mean (± SD) number of parasites in infected hosts. stomach; three parts of the intestines; *^**^***^ if the number of infected hosts differs between the shrew species, the higher value is marked with asterisks (\~ test): *, P < 0.05; P < 0.01; ***, P < 0.001. S. araneus S. caecutiens Helminth species Volume n X ± SD n X + SD Trematoda Brachylaimus fulvus^^ 0.7 4 1.0 ± 0.0 0 — — Opisthioglyphe sobolevi“ 0.1 2 2.0 ± 0.0 2 1 ± 0.0 Rubenstrema opisthioglyph^ 3.9 2 23.0 ± 22.0 0 — — Pseudoleucoch loridium sori cis^ 0.5 3 29.0 ± 28.3 0 — — Cestoda Choanotaenia crassiscolex * 7.0 44=1=** 5.3 + 9.4 3 1.3 + 0.6 Hymenolepis fur cat a^ 3.6 6 3.0 ± 4.9 1 1 — H. schaldybini~'^ 1.5 37 6.8 + 8.0 48* 11.1 ± 12.0 H. singular!^ 3.6 10 4.9 ± 6.9 6 6.2 ± 6.5 H. scutigera"’^ 0.4 72*** 28.6 ± 30.0 13 4.0 + 4.1 Hymenolepis sp.^ 0.3 45* 28.5 + 34.7 25 11.7 ± 16.2 H. infirma^ 0.0 30 84.5 + 102.1 29 43.3 ± 71.5 H. globosoideP' 18.0 11* 1.8 ± 1.3 1 1 — Dilepis undula^'^'^ 0.4 11* 1.5 + 0.7 0 — — Nematoda Capillaria sp.®* 0.5 15* 1.0 ± 0.0 4 1.0 ± 0.0 Longistriata depressa^ 0.1 84* 18.8 ± 20.0 54 7.5 ± 6.6 L. pseudodidas^ 0.1 80* 9.1 ± 7.7 54 5.8 + 5.7 Parastrongyloides winchesv' 0.1 22* 4.4 ± 3.5 6 2.7 ± 2.1 1994 HAUKISALMI ET AL. — Parasitism of Shrews by Gastrointestinal Helminths 101 Table 2. — Infection indices describing the total helminth burden of shrews. Values for number of species and ratio of number of species to host weight are arithmetic means ( ± SD), other values are medians. All indices show significant (? < 0. 01) differences among the eight categories of shrews (Kruskal-Wallis test). if infection parameters differ significantly (P < 0.05) between the age groups (a), sexes (s), or host species (h), the higher value is marked with the respective symbol. s. araneus s. caecutiens Males Females Males Females Juveniles Adults Juveniles Adults Juveniles Adults Juveniles Adults n 47 21 37 9 46 19 29 13 Number of species 4.1*' 4.6*' 4.3*' 3.7 1.9 3.4® 2.1 3.0 (1.3) (2.0) (1.7) (1.2) (1.1) (1.0) (1.2) (1.6) Species/weight (X 10) 8.2^ 6.4 8.3®*' 5.2 6.1 7.3 6.6 6.8 (2.5) (2.7) (3.4) (2.1) (3.8) (2.2) (3.7) (3.8) Number of helminths 59** 107ash 52*' 38 12 33® 16 20 Helminths/weight 10.3*' 15. 4"*' 10.0*' 4.9 3.7 4.9 5.3 5.0 Volume of helminths 26*' 29 22*' 11 5 27® 4 16 Volume/weight (X 1000) 5.2*' 4.0 4.0*' 2.2 1.6 5.6® 1.3 3.3 Table 3. — Best log-linear models for dependence between the occurrence of helminths (H), and sex (S) and age (A) of shrews (see text). full-order model; goodness-of-fit test is not possible. Helminth Species S. araneus S. caecutiens Model d.f. ■7 r P Model d.f. P C. crassiscolex S,AH 3 3.1 0.38 H. schaldybini S,AH 3 3.2 0.36 S,AH 3 3.1 0.37 H. scutigera S,AH 3 3.8 0.29 — — — — Hymenolepis sp. S,AH 3 5.3 0.15 S,AH 3 0.8 0.84 H. infirma SAH® S,AH 3 4.2 0.24 L. depressa S,A,H 4 5.0 0.28 S,A,H 4 6.5 0.70 L. pseudodidas S,A,H 4 3.8 0.43 S,AH 3 0.4 0.94 P. winchesi S,A,H 4 3.1 0.53 — — — — Table 4. — Partial correlations between three infection indices and weight (without alimentary tract) of juvenile and adult shrews, controlled for the effect of body length. Logarithmic transformations were performed on the number and volume of helminths. *, P < 0.05. No. of No. of Helminth Species Helminths Volume S. araneus Juvenile 1988 (« = 29) 1989 (n = 34) Adult (n =18) S. caecutiens Juvenile 1988 (n = 25) 1989 (n = 22) Adult (n = 22) 0.36 -0.00 0.18 -0.24 -0.37* -0.28 -0.14 0.27 0.23 -0.30 -0.30 -0.09 0.20 -0.015 -0.07 0.41 0.26 0.51* 102 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 % 100-|- 80- 60-- C. crassiscolex ^JJl M F araneus ( n M F caecutiens araneus caecutiens % lOOj 80-- H. scutigera 60-1- H mL M F araneus ■ I M F caecutiens % 100 80- Hymenolepis sp. 60 -t- :Lli M F araneus Lk M F caecutiens H. infirma L. depressa araneus caecutiens araneus caecutiens L. pseudodidas 1 1 M F caecutiens % lOOn 80 60- 40 P. winchesi :ljiI .n M F araneus M F caecutiens Fig. 1. — Prevalence of common helminths in the shrews Sorex araneus and S. caecutiens. M, males; F, females. Black bars, juveniles; white bars, adults. THE MASKED SHREW, SOREX CINEREUS, IN A RELICTUAL HABITAT OF THE SOUTHERN APPALACHIAN MOUNTAINS John F. Pagels', Kristen L. Uthus’, and Henry E. Duval* ’Department of Biology, Virginia Commonwealth University, Richmond, Virginia 23284-2012 Abstract We analyzed habitat features in relation to the abundance of Sorex cinereus at ten high-elevation sites (1082-1524 m) in Virginia, USA. Each site contained red spruce {Picea rubens), an indicator of boreal habitat. S. cinereus comprised 89.4% of 434 shrews captured, and it was the only species captured at all sites. Captures of S. cinereus were significantly correlated (P < 0.05) with soil moisture-holding capacity, soil organic matter, and total understory vegetation, but otherwise no suite of habitat characteristics was present that could be correlated with the abundance of S. cinereus. Characteristics that provided cover, e.g., rocks, stumps, and fallen trees, were important, but features that provided suitable habitat at one site were often replaced by other features at other sites. Our data indicate that habitat characteristics that promote shaded, moist habitat are critical to S. cinereus, and that such features may be especially important in a relictual forested habitat in the southern Appalachian Mountains. Introduction The late John E. Guilday (1972:233) observed that the Appalachian Mountains provide “...a tongue of ‘more northerly’ environment into the Carolinean lowlands of the South.” The northern conditions present at high altitudes in the southern Appalachians allow boreal species, such as Sorex palustris, Glaucomys sabrinus, and Microtus chrotorrhirius, to exist far south of their centers of distribution. These species had greater, more continuous ranges in the Appalachians during periods of glacial cooling, but in the southern Appalachians all now have highly disjunct, fragmented ranges (Hall, 1981). Both the populations and the habitats in which they occur are considered relicts of the ice age. In contrast, other northern species that reach into the southern Appalachians have broader distributions. The range of the masked shrew, Sorex cinereus, the most widely distributed member of the genus in North America, is centered in the transcontinental coniferous forest. In addition to southern extensions into montane forests of the Appalachian and Rocky mountains, it also occurs northward into the tundra (Junge and Hoffmann, 1981). Indicative of its boreomontane distribution, the masked shrew has not been taken below 610 m elevation in Virginia (Pagels and Handley, 1989). Numerous studies have described S. cinereus as a habitat generalist. In Canada, van Zyll de Jong (1983) reported S. cinereus from alder and willow thickets; along margins of marshes, bogs, and streams; and in deciduous and coniferous woods up to the timberline. In Manitoba, Wrigley et al. (1979) captured S. cinereus in herbaceous and shrubby areas, deciduous and coniferous forests, dry prairie, sparse weeds on sand dunes, and spruce-aspen-juniper savannas. In Wyoming, S. cinereus is known from willow-sedge savannas, and wet, grass-sedge meadows (Negus and Findley, 1959; Clark, 1973), in spruce-fir forests (Raphael, 1988), and in numerous other habitats from sagebrush to an alpine rockslide (Brown, 1967). Additional habitat types include hardwood floodplains (French, 1984) and old fields (French, 1984; Whitaker, personal communication) in Indiana; old fields, marshes, hardwood swamps, spruce swamps, spruce bums, and bogs in northern Michigan (Pruitt, 1953, 1959; Getz, 1961; Ryan, 1982); swamp hardwood forests in the Northeast (Hill, 1982); farmstead shelterbelts in Minnesota (Yahner, 1982); and a high-altitude grassy area in Tennessee (Tuttle, 1964). The masked shrew is also a generalist in terms of its diet, which consists of adults and larvae of numerous insect taxa and other invertebrates such as earthworms, sowbugs, centipedes, spiders, mollusks, and vertebrates, including young mice and salamanders (van Zyll de Jong, 1983; French, 1984). Getz (1961) suggested that because of its small size S. cinereus is able to forage effectively for invertebrates that are present in almost all areas, and that food availability may not be a critical factor in its local distribution. Several studies have suggested that moisture is the most important factor influencing the local distribution of S. cinereus (Pruitt, 1953, 1959; Getz, 1961; Spencer and Pettus, 1966; Wrigley et al., 1979; Hill, 1982). Pruitt (1953, 1959) observed that S. cinereus can only permanently inhabit those microhabitats in which the humidity approaches saturation. Getz (1961) concluded that the primary importance of cover, whether it be leaf litter, moss, or vegetation, was its effect on humidity. Our objectives were to characterize habitat features in relictual habitat of the southern Appalachian Mountains and to relate these features to the abundance of S. cinereus. We expected that S. cinereus would occur at all sites. However, because all sites were forested, we hypothesized that if differences did exist in the abundance of S. cinereus, our analyses would indicate a suite of habitat features, notably those that promote high moisture, associated with high S. cinereus abundance. Methods The ten study sites in western Virginia possessed vegetation features, for example Picea rubens and northern hardwood species, usually associated with a boreal fauna (Payne et al., 1989). All sites exceeded 1000 m elevation. Mean elevation of the eight Highland County sites was 1 160 m (Hi-12, 1097 m; Hi-12 + , 1082 m; Hi-13, 1127 m; Hi-14, 1127 m; Hi-14 + , 1158 m; Hi-16, 1188 m; Hi-18, 1219 m; Hi-21, 1280 m), and 103 104 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 1501 m (CC, 1478 m; and WT, 1524 m) at the two Grayson County sites. Mammal sampling and habitat measurements were centered on points that appeared to best represent the habitat. Small mammals were trapped with arrays of pitfall traps and terrestrial drift fences as described by Bury and Com (1987) and similar to those of Kirkland and Sheppard (1994). The pitfall arrays at each site consisted of two triads of pitfalls spaced 20-25 m apart. A triad consisted of three 5-m strips of aluminum flashing forming drift fences 24.5 cm high radiating approximately 120° outward from each other, beginning 2-3 m from a central point. Two #10 tin cans were buried to ground level at each end of the drift fence, one on each side. Each pair of cans was considered one trap. A plastic insert fitted with a wire handle was placed in each can to facilitate removal of specimens. The pitfalls were filled with 15% formalin solution to drown and preserve captured animals. Traps were checked at approximately three-week intervals from August to November 1988, and from March to November 1989. All specimens were deposited in the Virginia Commonwealth University Mammal Collection (VCU). The line-intercept method (Canfield, 1941) was used to measure most habitat features. Transects for all habitat measurements were centered on pitfall arrays, but transects were not necessarily the same for different sets of data collected on different dates. One hundred measurements of understory features were taken at 1-m intervals along transects that extended 25 m in the four major compass directions. Measurements of the understory recorded at the 100 points included surface and underground rock, moss, lichens, ferns, herbs, tree seedlings, and dead wood (fallen trees, stumps). A pole with a spike was driven into the ground to a depth of 10 cm to determine the presence or absence of underground rock. A rock was counted as a surface rock if its greatest length was at least 10 cm. Soil pH, field capacity, or the moisture-holding capacity of the soil (after Salter and Williams, 1967), and percent organic matter (after Ball, 1964) were determined for ten soil samples taken randomly from each site. Observations of the overstory included presence, species composition, and layering of the canopy and subcanopy. All trees, including standing dead trees, with a diameter at breast height (dbh) of at least 10 cm were counted within a circumscribed area (60 m diameter) at each site to determine density of the forest stand. Stand age was determined from 40 increment cores per site taken from trees with a dbh of 10 cm or greater which were closest to 3-m points along 30-m transects, again directed in the four major compass directions. Canopy openness was estimated using a tube held overhead. Average linkage cluster analysis (Ludwig and Reynolds, 1989) was used to determine communities based on the following habitat measurements: soil features, tree species and density, understory features, and understory vegetation. Analyses of variance and Tukey multiple range tests were used to determine if significant differences in habitat variables existed among the sites. Regression analyses were used to determine if correlations existed between habitat features and S. cinereus captures. Results A total of 434 shrews representing six species was captured. Sorex cinereus comprised 89.4% of shrew captures and was the only species taken at all sites (Table 1). In addition, seven rodents were captured: Peromyscus maniculatus (2), Synaptomys cooperi (I), Microtus pennsylvanicus (I), and Clethrionomys gapperi (3). Five overstory species accounted for over two-thirds of the 1,590 trees counted: Picea rubens (31%), Acer rubrum (11%), Fagus grandifolia (9%), Betula lutea (9%), and Betula lenta (8%) (Table 2). Picea rubens was present at all sites, A. rubrum and F. grandifolia at eight sites, and B. lenta and Prunus serotina at seven sites. Overstory diversity (Shannon index, H') was generally high at all sites except Hi-14 and WT which were characterized by an abundance of P. rubens (Table 2). Soils from Hi-16, CC, and WT had similar features including low pH, high field capacity, and high organic matter content (Table 3). The field capacities at CC and WT were significantly different from those at other sites (P < 0.05). Three communities were identified in a cluster analysis of the sites using soil features. Sites CC and WT were each established as separate communities. The other sites were clustered as a third community at 50% dissimilarity. Percent organic matter of soils, field capacity of soils, and total understory vegetation were the only individual microhabitat features significantly correlated to captures of S. cinereus (P < 0.05, r" — 0.72, 0.52, and 0.43, respectively). Shrubs were absent at four of the ten sites (Table 4). Kalmia latifolia was the most prevalent shrub species, especially at Hi- 16 and Hi-18, sites with relatively high understory diversity. Herbs were present at all sites, with a notable abundance of Houstonia sp. at WT and grasses at Hi-21. Surface rocks were present at six sites, but they were particularly large at CC and WT, two of the sites with significantly higher numbers of S. cinereus (Table 4). Total vegetation (total understory variables minus rock, fallen trees, and stumps) at the sites ranged from i 41 at Hi-14 -I- to 279 at WT. A cluster analysis of sites using understory features (Table 4) identified three communities. Sites Hi-16 and Hi-18 were paired as one community, and WT, the most dissimilar site, was established as a separate community. ' The remaining sites were clustered as a third community. The > communities showed no relation to those based on S. cinereus captures. Mean stand age ranged from 47.7 years at Hi-14-l- (mean dbh = 0.30 m) to 108.0 years at WT (mean dbh = 0.28 m) (Table 5), but stand age was not correlated to mean dbh. Analysis of variance determined that there was no significant difference in dbh among the sites; however, WT was significantly older than all other stands (P < 0.001). All sites had at least three trees 70 years of age or older and one tree 80 years of age or older. Total tree density ranged from 312 trees/ha at Hi-13 to 734 trees/ha at Hi-14 (Table 5). No relationship was observed between the total density of trees and the diversity of the stands 1994 PAGELS ET AL. — Sorex cinereus in a Relictual Habitat 105 or tree density and the relative abundance of S. cinereus. Basal area, the product of tree density and dbh, ranged from 15.3 m^/ha at Hi-13 and CC to 30.5 m^/ha at Hi-14. No relationship was observed between total tree density and basal area of stands or the relative abundance of shrews. All sites possessed a layered canopy; however, the extent of layering and number of layers varied among sites (Table 6). A high overstory layer was present at almost all of the sampling points, but the canopy was not completely closed at any site. Mean canopy openness ranged from 7.9% at Hi-13 to 44.0% at Hi-21. Sites Hi-21 and WT were significantly more open than other sites (P < 0.05). Foliage height diversity (FHD) (Aber, 1979; McPeek et al., 1983) showed little variation among the sites, with the exception of WT which had a significantly lower FHD than other sites (Table 6). Not surprisingly, FHD was found to be negatively correlated to understory vegetation (P < 0.05, = 0.57). As noted, however, only total understory vegetation, not FHD, was significantly correlated with S. cinereus captures. Measurements of selected structural features of the understory of all sites are given in Table 7. Of the sites with greatest S. cinereus captures, Hi-16 had relatively large numbers of stumps and fallen trees, and CC and WT had many large surface boulders that approached one m in their greatest dimension (Table 7). Discussion The influence of habitat on populations of small mammals has received much attention in recent years (August, 1983; McPeek et al., 1983; Seagle, 1985; Adler, 1985, 1987; Hanski and Kaikusalo, 1989). Many studies have focused on the effects of habitat structure and the roles of both microhabitat and macrohabitat (Price, 1978; Yahner, 1982; Morris, 1987; Bowers and Flanagan, 1988). Price (1978) and Yahner (1982) noted that the availability of suitable microhabitats may determine abundances of species on a local scale. Other studies (Morris, 1987; Bowers and Flanagan, 1988) have suggested that the density of small mammal populations may more closely reflect habitat variability at the macrohabitat level. Adler and Wilson (1987) noted that habitat generalists utilize different habitat types over their entire distributional range, but on a local scale, they may be particularly efficient at exploiting diverse resources and responding readily to environmental fluctuations. The variation in abundance of S. cinereus among our sites in basically a single type of habitat supported these views. Further, as Getz (1961) surmised, amount of cover is more important than type of cover in determining the distribution and abundance of S. cinereus. Our data suggest that various features of a habitat, both micro- and macro-, are important in supplying suitable habitat for S. cinereus. Our results support the findings of Getz (1961) and Pruitt (1953, 1959) that greater numbers of S. cinereus are found in association with features that promote shaded, moist conditions, i.e., features that one would hypothesize to be important to a northern species. Our findings concur with Snyder and Best (1988) and Yahner (1982), who reported a positive relationship between vegetation density and the abundance of S. cinereus in diverse habitats in Minnesota. Importantly, however, as we also found in this relictual forested habitat, features that provide suitable habitat are variable, and not all such variables must be present at the same time or in equal abundances. Illustrating this point, several cluster analyses were performed using different groups of habitat variables; however, the clusters showed no concordance with S. cinereus captures or with one another. These analyses indicate that no single set of habitat variables, such as soil features, understory features, or vegetation, was responsible for the differences in S. cinereus densities among the sites. Although the three sites with greatest S. cinereus captures — Hi-16, CC, and WT — were each separated into distinct soil and understory vegetation communities, they were similar in that they had structural features that provided suitable habitat. Site Hi-16 had few rocks, but a high number of fallen trees and stumps, and relatively heavy shrub cover (Table 4, 7). Site WT had the lowest FHD, but had large boulders and an abundance of herbs and evergreen tree seedlings. Site CC did not have a particularly heavy understory; however, there were several large boulders at the site (Table 7). Site CC also had dense thickets of rhododendron which were counted as part of the subcanopy due to their height. Rhododendron undoubtedly had the same effect as shrubs on the microhabitat in terms of shading and cover. In summary, because all of our sites were in a single kind of habitat, i.e., a relictual forest, we cannot suggest that the sites with the greatest numbers of S. cinereus provided better habitat than we would have found had we sampled other habitats. In the forest type we studied both micro- and macrohabitat features were important in contributing to a suitable environment for S. cinereus. Such features, although they demonstrated much variability among sites, were those that promoted shaded, moist conditions. Acknowledgments We thank T. Brody, L. Brody, and T. McBride for use of their properties; M. Fies, R. Glasgow, and D. Young for assistance in the selection of sites; D. Young for advice on habitat measurements and data analyses; M. Whittemore for development of software for data entry; and personnel of both the Thomas Jefferson National Forest and the George Washington National Forest for assistance and cooperation. J. Baker and J. Haulsee assisted in trap placement, and they and Fies checked the traps in Grayson County. This study was supported by funds to Pagels from the Virginia Department of Game and Inland Fisheries Nongame Wildlife and Endangered Species Program. Literature Cited Aber, J. D. 1979. Foliage-height profiles and succession in northern hardwood forests. Ecology, 60:18-23. Adler, G. H. 1985. Habitat selection and species interactions: An experimental analysis with small mammal populations. Oikos, 45:380-390. 1987. Influence of habitat structure on demography of two rodent species in eastern Massachusetts. Canadian Journal of 106 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 Zoology, 65:903-912. Adler, G. H., and M. L. Wilson. 1987. Demography of a habitat generalist, the white footed mouse, in a heterogenous environment. Ecology, 68:1785-1796. August, P. V. 1983. The role of habitat complexity and heterogeneity in structuring tropical mammal communities. Ecology, 64:1495-1507. Ball, D. F. 1964. Loss-on-ignition as an estimate of organic matter and organic carbon in non-calcareous soils. Journal of Soil Science, 15:84-92. Bowers, M. A., and C. A. Flanagan. 1988. Microhabitat as a template for the organization of a desert rodent community. Pp. 300-312, in Management of Amphibians, Reptiles and Small Mammals in North America (R. C. Szaro, K. E. Severson, and D. R. Patton, tech, coords.), USDA Forest Service General Technical Report RM-166, 458 pp. Brown, L. N. 1967. Ecological distribution of six species of shrews and comparison of sampling methods in the central Rocky Mountains. Journal of Mammalogy, 48:617-623. Bury, R. B., and P. S. Corn. 1987. Evaluation of pitfall trapping in northwestern forests: Trap arrays with drift fences. Journal of Wildlife Management, 51:112-119. Canfield, R. 1941. Application of the line interception method in sampling range vegetation. Journal of Forestry, 39:388-394. Clark, T. W. 1973. Distribution and reproduction of shrews in Grand Teton National Park, Wyoming. Northwest Science, 47:128-131. French, T. W. 1984. Dietary overlap of Sorex longirostris and S. cinereus in hardwood floodplain habitats in Vigo County, Indiana. The American Midland Naturalist, 111:41-46. Getz, L. L. 1961. Factors influencing the local distribution of shrews. The American Midland Naturalist, 65:67-88. Guilday, J. E. 1972. The Pleistocene history of the Appalachian mammal fauna. Pp. 233-262, in The Distributional History of the Biota of the Southern Appalachians, Part III: Vertebrates (P. C. Holt, R. A. Paterson, and J. P. Hubbard, eds.). Research Division Monograph 4, Virginia Polytechnic Institute and State University, Blacksburg, 306 pp. Hall, E. R. 1981. The Mammals of North America. 2nd cd. John Wiley and Sons, New York, 1:1-600-1-90. Hanski, I., AND A. Kaikusalo. 1989. Distribution and habitat selection by shrews in Finland. Annales Zoologici Fennici, 26:339-348. Hill, B. J. 1982. Small mammal habitat associations in selected timbered community types on the White Mountain National Forest, New Hampshire. Unpublished M.S. thesis. University of Massachusetts, Amherst, 140 pp. JUNGE, J. A., AND R. S. Hoffmann. 1981. An annotated key to the long-tailed shrews (genus Sorex) of the United States and Canada, with notes on the middle American Sorex. Occasional Papers of the Museum of Natural History, University of Kansas, 94:1-48. Kirkland, G. L., Jr., and P. K. Sheppard. 1994. Proposed standard protocol for sampling small mammal communities. Pp. 277-283, in Advances in the Biology of Shrews (J. F. Merritt, G. L. Kirkland, Jr., and R. K. Rose, eds.), Carnegie Museum of Natural History Special Publication no. 18, x + 458 pp. Ludwig, J. A., and J. F. Reynolds. 1989. Cluster analysis. Pp. 189-202, in Statistical Ecology: A Primer on Methods and Computing, John Wiley and Sons, New York, 337 pp. McPeek, M. a., B. L. Cook, and W. C. McComb. 1983. Habitat selection by small mammals in an urban woodlot. Transactions of the Kentucky Academy of Science, 44:68-73. Morris, D. W. 1987. Tests of density-dependent habitat selection in a patchy environment. Ecological Monographs, 57(4):269-281. Negus, N. C., and J. S. Findley. 1959. Mammals of Jackson Hole, Wyoming. Journal of Mammalogy, 40:371-381. Pagels, j. F., and C. O. Handley, Jr. 1989. Distribution of the southeastern shrew, Sorex longirostris Bachman, in western Virginia. Brimleyana, 15:123-131. Payne, J. L., D. R. Young, and J. F. Pagels. 1989. Plant eommunity characteristics associated with the endangered northern flying squirrel, Glaucomys sabrinus, in the southern Appalachians. The American Midland Naturalist, 121:285-292. Price, M. V. 1978. The role of microhabitat in structuring desert rodent communities. Ecology, 59(5):910-921. Pruitt, W. O., Jr. 1953. An analysis of some physical factors affecting the local distribution of the shorttail shrew (Blarina brevicauda) in the northern part of the lower peninsula of Michigan. Miscellaneous Publications of the Museum of Zoology, University of Michigan, 79:1-39. 1959. Microclimates and local distribution of small mammals on the George Reserve, Michigan. Miscellaneous Publications of the Museum of Zoology, University of Michigan, 109:1-27. Raphael, M. G. 1988. Habitat associations of small mammals in a subalpine forest, southeastern Wyoming. Pp. 359-367, in Management of Amphibians, Reptiles and Small Mammals in North America (R. C. Szaro, K. E. Severson, and D. R. Patton tech, coords.), USDA Forest Service General Technical Report RM-166, 458 pp. Ryan, J. M. 1982. Distribution and comparative ecology of the pygmy (Sorex hoyi) and masked (Sorex cinereus) shrews in northern lower Michigan. Unpublished M.S. thesis. University of Michigan, Ann Arbor, 56 pp. Salter, P. J., and J. B. Williams. 1967. The influence of texture on the moisture characteristics of soils. Journal of Soil Science, 18:174-181. S EAGLE, S. W. 1985. Patterns of small mammal microhabitat utilization in cedar glade and deciduous forest habitats. Journal of Mammalogy, 66:22-35. Snyder, E. J., and L. B. Best. 1988. Dynamics of habitat use by small mammals in prairie communities. The American Midland Naturalist, 119:128-136. Spencer, A. W., and D. Pettus. 1966. Habitat preferences of five sympatric species of long-tailed shrews. Ecology, 47:677-683. Tuttle, M. D. 1964. Observation of Sorex cinereus. Journal of Mammalogy, 45:148. VAN Zyll DE Jong, C. G. 1983. Sorex cinereus Kerr. Pp. 65-71, in Handbook of Canadian Mammals, Volume 1. Marsupials and Insectivores, National Museum of Natural Sciences, Ottawa, Canada, 210 pp. Wrigley, R. E., j. E. Dubois, and H. W. R. Copland. 1979. Habitat, abundance, and distribution of six species of shrews in Manitoba. Journal of Mammalogy, 60:505-520. Yahner, R. H. 1982. Microhabitat use by small mammals in farmstead shelterbelts. Journal of Mammalogy, 63:440-445. 1994 PAGELS ET AL. — Sorex c'mereus in a Relictual Habitat 107 Table 1. — Captures of shrews at each site. Relative abutuJance (captures per 100 trapnights) is indicated under numbers captured. Total trapnights were 3,564 at CC and WT, and 3,660 at all other sites. Site Species 12 12 + 13 14 14 + 16 18 21 CC WT Total Sorex cinereus 17 28 24 36 14 68 18 43 64 76 388 0.46 0.77 0.66 0.98 0.38 1.86 0.49 1.17 1.80 2.13 5.37 S. fumeus 1 3 2 1 1 2 — — 4 6 20 0.03 0.08 0.05 0.03 0.03 0.05 — — 0.11 0.17 0.28 S. dispar — — — — — — — — — 1 1 0.03 0.01 S. hoyi — — — — — 1 3 3 1 — 8 0.03 0.08 0.08 0.03 0.11 Cryptotis parva — — — — — — — — — 1 1 0.03 0.03 Blarina brevicauda 1 1 1 — 4 3 1 2 3 — 16 0.03 0.03 0.03 0.11 0.08 0.03 0.05 0.08 0.22 Total 19 32 27 37 19 74 22 48 72 84 434 Richness 3 3 3 2 3 4 3 3 4 4 6 Evenness 0.37 0.41 0.38 0.17 0.64 0.26 0.52 0.37 0.33 0.28 0.26 Diversity (H') 0.18 0.19 0.18 0.05 0.30 0.16 0.25 0.18 0.20 0.17 0.20 Table 2. — Density of trees per hectare (first line) and relative frequency (total of one species/total of all species; second line) of live overs tory species. Site Species 12 12 + 13 14 14 + 16 18 21 CC WT Picea rubens 39 21 109 595 198 117 106 71 64 448 0.07 0.05 0.35 0.81 0.33 0.25 0.27 0.21 0.13 0.91 Acer rubrum 85 28 7 64 25 195 209 11 0 0 0.14 0.07 0.02 0.09 0.04 0.42 0.53 0.03 — — Fagus grandifolia 28 110 11 43 237 0 0 4 60 14 0.05 0.27 0.04 0.06 0.39 — — 0.01 0.13 0.03 Betula lenta 209 103 7 0 0 103 53 21 0 7 0.35 0.25 0.02 — — 0.22 0.13 0.07 — 0.01 Betula lutea 103 0 14 0 0 18 4 0 326 11 0.17 — 0.04 — — 0.04 0.01 — 0.70 0.02 Tsuga canadensis 85 74 135 0 0 7 11 0 0 0 0.04 0.18 0.43 — — 0.02 0.03 — — — Prunus serotina 21 57 11 14 92 0 11 4 0 0 0.14 0.14 0.04 0.02 0.15 — 0.03 0.01 — — Robinia pseudoacacia 0 0 0 0 35 0 0 71 0 14 — — — — 0.06 — — 0.22 — 0.03 Crataegus sp. 0 0 0 0 0 0 0 106 4 0 — — — — — — — 0.34 0.01 — Acer saccharum 21 11 7 0 0 0 0 14 0 0 0.04 0.03 0.02 — — — — 0.04 — — Quercus rubrum 0 0 0 18 11 0 0 14 0 0 — — — 0.02 0.02 — — 0.04 — — Amelanchier arborea 0 0 0 0 4 7 4 0 7 0 — — — — 0.01 0.02 0.01 — 0.01 — 108 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 Table 2 (cont.) Site Species 12 12 + 13 14 14 + 16 18 21 cc WT Acer pennsylvanicus 0 0 11 0 0 4 0 0 11 0 — — 0.04 — — 0.01 — — 0.02 — Abies fraseri 0 0 0 0 0 0 0 0 7 0.01 0 0 Magnolia acuminata 0 o 1 0 o 1 o 1 7 0.02 4 1 o 0 0 Cary a sp. 0 4 o 1 0 0 o 1 0 0 0 — 0.01 — — — 0.01 — — — Sorb us americana 0 0 0 0 0 0 0 0 4 0.01 4 0.01 0 Acer spicatum 0 0 1 o 1 0 0 0 0 1 o 1 0 Total density 591 408 312 734 602 459 398 316 487 488 Dead trees 88 78 14 21 103 152 149 46 99 110 Richness 8 8 9 5 7 9 7 9 9 5 Evenness 0.86 0.85 0.67 0.44 0.73 0.66 0.64 0.78 0.51 0.27 Diversity (H') 2.58 2.55 2.13 1.03 2.05 2.09 1.80 2.48 1.61 0.63 Table 3. — Mean soil pH, mean field capacity (amount of water expressed as percent of soil dry weight), and mean percent organic matter at each site. Numbers below are one standard error of the mean. Site 12 12 + 13 14 14 + 16 18 21 cc WT X pH 3.7 4.4 4.0 4.1 3.8 3.7 3.9 4.1 3.5 3.3 3.9 0.07 0.07 0.06 0.08 0.05 0.03 0.06 0.08 0.05 0.02 0.03 Field capacity 34.9 0.9 12.5 1.2 41.4 1.7 8.3 0.9 50.2 2.9 13.1 2.1 41.9 1.5 10.2 0.3 39.6 1.3 10.2 1.0 47.3 7.5 26.0 6.4 40.7 4.7 11.1 2.8 29.5 1.9 9.3 0.8 103.0 7.9 20.3 1.4 176.5 19.8 35.3 5.1 60.5 4.9 15.6 1.2 Organic matter 1994 PAGELS ET AL. — Sorex cinereus in a Relictual Habitat 109 Table 4. — Occurrences of understory variables at 100 points at each site. Site 12 12 + 13 14 14 + 16 18 21 cc WT Surface rock 0 0 0 0 1 1 20 15 12 7 Underground rock 7 7 17 4 16 16 57 52 21 21 Fallen trees 8 9 9 15 7 13 13 7 4 9 Stumps 1 0 1 3 1 5 1 2 0 0 Shrubs 0 0 0 17 1 33 69 10 0 7 Evergreen tree seedlings 15 3 5 1 2 38 1 7 3 42 Deciduous tree seedlings 11 16 8 6 22 18 19 13 15 9 Herbs 37 35 23 9 5 4 7 51 30 78 Ferns 44 24 27 5 8 19 18 44 52 44 Lycopodium 7 0 5 0 0 1 4 0 0 35 Mosses 6 4 27 9 4 19 16 9 15 64 Total vegetation 120 82 95 47 41 132 127 ]34 115 279 Total understory 136 98 122 57 67 167 225 210 152 316 Diversity (H') 0.78 0.72 0.83 0.86 0.78 0.89 0.87 0.83 0.76 0.86 Table 5. — Mean dbh, mean age, and estimated basal area calculated from measurements made on 40 trees per site. Total density was derived from total tree counts at each site. Numbers below are one standard error of the mean. Site 12 12 + 13 14 14 + 16 18 21 cc WT X DBH (m) 0.23 0.24 0.25 0.23 0.22 0.22 0.27 0.30 0.20 0.28 0.24 0.017 0.019 0.021 0.021 0.021 0.017 0.014 0.025 0.014 0.019 0.004 Age (yrs) 59.7 71.73 77.9 53.2 47.7 66.7 62.0 48.8 64.4 108.0 66.0 3.12 4.05 4.69 1.88 2.00 3.47 2.08 3.45 5.42 5.33 5.6 Density Basal area 591 408 312 734 602 459 398 316 487 488 480 (m^/ha) 24.6 18.5 15.3 30.5 22.9 17.4 22.8 22.3 15.3 30.0 21.7 Table 6.— Occurrence of subcanopy and canopy layers, and mean canopy openness from 100 points at each site. Foliage height diversity (FHD) indices (see Aber, 1979) were calculated from the number of layers of the subcanopy and canopy. Site 12 12 + 13 14 14 + 16 18 21 cc WT Subcanopy High 46 51 15 23 18 56 7 35 61 21 Low 2 0 0 1 1 8 0 0 5 1 Overstory High 98 100 99 100 100 100 100 67 98 88 Medium 50 59 52 47 61 28 30 14 23 3 Low 9 16 6 7 11 5 5 2 0 0 FHD 0.52 0.54 0.44 0.48 0.49 0.52 0.37 0.44 0.46 0.28 Canopy openness 12.8 15.4 7.9 14.4 10.1 13.3 16.0 44.0 20.2 33.2 =i(: . I ■ :'i M * ■V V^ ' ■ '1 * ' r *< ■ ■ -^IfL^ 6^' :•■ .■*.. • f; • . . * ,_ . ^- ,: IV. i^rt- lj» ^\^\^0'A i)Cj\ .Vi .lvSUWV»T')t'^ .01 i-f u - 1 . \ i 1 .1 U '"S - • #(|: ' tv ^ 11 I \1'.' •'!, • 'W ^ ■ •n ^ “€6ws■ •:*'. *f'>.v.*\ Vv»> WrtVj, ,,'.-,j. ..p\. > ,,, .. \.: *Vw^?yi ■ v'Ji 'i'' '•' 'v^U 5^?') r.4'« -p,V iNiw«~i.l^.l*^ - tfT ' rjt ‘ ^ ' -7- ♦ ' 1 ' 1 ' • 1. ?i '■ rtt « ; / ‘ 'i ^- ^iotfSV0'4> .■! i«v» 'All In *ti i ■ ‘ . ' . ■>' Q r am 4 .•J >1 V/ It tr (i. i^?.0 V I 't. » -I - t •:( T ■% ■. . '^ • ' 3 LIFE fflSTORIES OF THE SORICIDAE: A REVIEW Duncan G. L. Innes Faculty of Dentistry, University of Western Ontario, London, Ontario, Canada N6A 5C1 Abstract Data on adult body mass, litter size, gestation length, neonate mass, age when eyes open, age and mass at weaning, growth rate to weaning, maximum life span, and length of the breeding season were compiled primarily from the published literature for 93 soricid species representing 12 genera. Body mass and litter size were the most frequently reported traits, and most information was found on Crocidura and Sorex species. Both litter size and the length of the breeding season differed significantly among all species, all genera, and between subfamilies. However, when the analyses were restricted to Crocidura and Sorex, Crocidura were larger as adults and neonates, had smaller litters, opened their eyes earlier, and had longer breeding seasons. Significant correlations between female adult body mass and other traits were not common. However, a number of developmental traits eonsistently eovaried with one another. Litter size was negatively correlated with length of the breeding season among populations, among species, and among genera, but not within any one species. Overall, the life histories of the Soricidae are diverse. For example, across all populations average male body mass and litter size ranged from 1.7 g to 117.0 g, and 1.2 to 9.8, respectively. Introduction Species of the family Soricidae occur almost worldwide and in a variety of habitats ranging from deserts to tropical rain forests (Hutterer, 1985). Thus, one might expect diverse life history tactics to have evolved to cope with different environmental conditions. There have been numerous reviews on the life histories of mammals. Most of these have examined large data bases and have emphasized differences and covariation among traits at higher taxonomic levels (families, orders; e.g., Eisenberg, 1981; Steams, 1983). Most have found that body size is a key trait because many other traits vary with it. However, other traits can covary if body size effects are removed (e.g., Harvey et al., 1989). Although differences among taxa and allometric relationships seem well-established at higher taxonomic levels, the ecological correlates for differences among groups are not always evident. For example, Gittleman (1986) found only a few dietary and vegetational effects that could explain life-history variation among carnivore families. Life-history traits may also be influenced by the degree of seasonality. Boyce (1979) stated that seasonality is important in explaining: 1) fat and resource storage mechanisms, 2) geographic variation in litter size, 3) rapid somatic growth patterns, and 4) the evolution of large body size, especially in homeothermic vertebrates. Some support for these postulates has been reported by Boyce (1978; 1988), May and Rubenstein (1984), Cameron and McClure (1988), and Zeveloff and Boyce (1988). These studies have demonstrated that mammalian litter size or body size, or both, increase with increasing latitude or highly seasonal evapotranspiration rates and temperature regimes. However, Millar (1984) found that adult mass and neonate mass, litter size, age at weaning, and developmental rates in mice of the genus Peromyscus were not significantly correlated with the length of the breeding season. Rather, he found that survival increased as the length of the breeding season decreased. Most comparative studies on life histories in the Soricidae have focused on a few species in a few genera (e.g., Vogel, 1972a, \912b). Recent studies have emphasized differences in longevity, litter size, reproductive effort, and postnatal development in species of Crocidura and Sorex (Vogel, 1980; Genoud, 1988; Genoud and Vogel, 1990). These studies have suggested that these differences may apply broadly to the two extant subfamilies (Crocidurinae and Soricinae). In this paper, I review the life histories of the Soricidae by examining 1 1 life-history traits and the length of the breeding season. Each variable was tested for differences at three levels: among species, among genera, and between subfamilies. Similarly, I examined how the traits covaried with one another, and with the length of the breeding season within species, among populations, among species, and among genera. Other ecological correlates were not considered because most species for which life-history data were available appear to fit into the “terrestrial” habitat grouping as tentatively classified by Hutterer (1985). Also, many species appear to have similar diets. Methods Life-History Traits The literature was searched for data on adult body mass, litter size, gestation length, neonate mass, age when the eyes open, age and mass at weaning, growth rate to weaning, maximum life span, and length of the breeding season. When possible, I consulted original sources, but some secondary sources were used. Unpublished data were also included and are cited as such. Data on adult body mass were compiled for males, females, or both sexes combined having a minimum sample size of two per population. Most are means (although a few median and modal values are included), primarily representing adults caught during the breeding season. Some authors were not explicit on how data were presented, and some samples may have included juveniles and pregnant females. Measures of combined mass for males and females were available for Crocidura cross ei, C. flavescens, and Suncus murinus, but were not used because males are over 33% heavier than females. Most other species were much less sexually size dimorphic, usually exhibiting less than a 20% difference between sexes. Ill 112 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 Litter size estimates are primarily mean values from embryo counts, with an n of two or more from each population. These estimates were supplemented with a few neonate counts from laboratory matings for species that had little or no other litter size data. Two additional studies (Johnston and Rudd, 1957; Forsyth, 1976) gave estimates based on counting neonates in natural nests. Four other studies (Hoffmeister and Goodpaster, 1962; Hutterer, 1976; Baxter and Lloyd, 1980; Michalak, l9Slb) that determined litter size from counting neonates conceived in the wild, but bom in the laboratory, were also included. Litter size estimates from placental scar counts were very rare and only two studies (Hamilton, 1940; Connor, 1966), which combined placental scar counts and embryo counts, were used. The length of gestation was usually determined under laboratory conditions, where it was measured as the number of days between a mating and subsequent parturition. In at least one species, Suncus etruscus, gestation can be longer if it is coincident with lactation (Vogel, 1970), so such estimates were not used. In two additional studies (Pearson, 1944; Crowcroft, 1957), minimum gestation length was determined by noting the time between capture of a wild gravid female and the birth of her litter in the laboratory. However, in many cases, the methodology used to determine the length of gestation was not described. All neonate weights are mean values for one or more litters measured in the laboratory, except for the data of Forsyth (1976) who inspected nests in the wild. Most young soricids open their eyes over several days and therefore a range was usually reported. Median values for this trait were compiled. Weaning is a gradual, transitional period for most mammals and has been measured in a variety of ways. Most investigators who described their methodology have defined the age at weaning as two or more days after the young were first seen eating solid food. Only two studies (Dryden, 1968; Michalak, 1987^) used the quantitative method of King et al. (1963). Other studies probably gave ages when weaning was forced and the young survived. Body mass at weaning and growth rate to weaning also represent values from laboratory studies. If not given, growth rates were calculated by subtracting neonate mass from mass at weaning and then dividing by the number of days of lactation. Estimates of maximum life span were compiled mainly from mark-recapture studies, although a few were based on toothwear indices (e.g., Dapson, 1968). Other traits such as age at maturity, survival, and number of litters per season were not compiled, primarily because they are rarely reported in the literature. Length of the breeding season was defined as the number of months in which pregnant or lactating females were caught under natural conditions. Statistical Analysis Most analyses were performed using SPSS-X (Statistical Package for the Social Sciences-Extension; SPSS, Inc., 1988). Means are given + 1 standard error (SE). Species means were calculated from population values, generic means were calculated from the constituent species means, and subfamily means were calculated from the constituent generic means, following Harvey et al. (1989). This averaging procedure minimizes bias if, for example, some genera are species-rich while others are species-poor. However, 70% of the species used here were from the genera Crocidura or Sorex. For this reason, I analyzed the data for all species, as well as those only from the above two genera. Analysis of life-history data at lower taxonomic levels (below the family level) is usually not recommended because taxa are closely related genetically and, therefore, do not serve as independent points (Harvey and Glutton-Brock, 1985). In a bivariate analysis, this would increase the level of significance. I analyzed the data at various levels to find differences and correlations that were common to all levels, reasoning that if the relationship occurred at all levels, it was likely real. All data were log (base 10) transformed before analyses, except growth to weaning, which as a proportion was arcsine transformed (Sokal and Rohlf, 1981). For correlation analyses, a minimum n of five was required. Also, significant correlations between growth rate to weaning and the traits it was derived from (neonate mass, mass at weaning, and age at weaning) should be viewed with caution because of possible autocorrelation. Nomenclature Taxonomically, the Soricidae is one of the most difficult of all mammalian families (Pruitt, 1957; Groin Michielsen, 1966) and the number of included species varies with the authority. In this study, a species name was retained if it was recognized by Honacki et al. (1982), Nowack and Paradiso (1983), or Gorbet and Hill (1986). Using the above authorities, the following generic names were changed: Megasorex gigas was placed in Notiosorex, Microsorex hoyi was placed in Sorex, and Surdisorex species were placed in Myosorex. Similarly, the following species names were changed: Crocidura occidentalis to C. flavescens, C. hildegarde to C. gracilipes, C. nigeriae to C. poensis, C. jouvenetae to C. crossei, C. deserti to C. hirta, Cryptotis floridanus to C. parva, Sorex personatus to S. cinereus, S. obscurus to S. monticolus, and S. roboratus to S. vir. Recently, three new species pertinent to this review have been recognized: Crocidura canariensis (Hutterer et al., 1987), C. zimmermanni (Vogel et al., 1986), and Sylvisorex vulcanorum (Hutterer and Verheyen, 1985). These new names are recognized herein. The results of studies by Hellwing (1970; 1971; 1973a; 1973ft) and Mover et al. (1988) on ^Crocidura russula” were attributed to C. suaveolens following Gatzeflis et al. (1985). Results More than 260 species of Soricidae are recognized, but life- history data were found for only 93 (Table 1). The data were highly heterogeneous, with 35 species represented by estimates for only one trait, whereas others had estimates for two or more traits. Most data were for species of the genera Crocidura and Sorex. Adult mass and litter size were the most frequently reported traits. 1994 INNES — Soricid Life Histories 113 Differences Among All Species A one-way analysis of variance showed that mean adult male mass differed among species (F = 52.5, P < 0.0001, n = 51) and ranged from 1.8 g {Suncus etruscus) to 65.2 g (S. murinus) (Table 1). Similarly, adult female mass differed among species {P - 46.4, P < 0.001, n = 45) and ranged from 2.0 g {S. etruscus) to 42.8 g (S. murinus). Combined body mass for both sexes also differed among species (F = 27.0, P < 0.0001, n = 54) and ranged from 2. 1 g {S. etruscus) to 22.0 g {Crocidura hirta). Depending on which adult mass category was used, 27 to 37 species were under 10 g, 14 or 15 species were between 10 and 20 g, and only 4 or 5 species were greater than 20 g. Mean litter size differed among species {F — 10.5, P < 0.001, n = 59) and ranged from 1.2 (Crocidura grayi) to 8.7 (Sorex tundrensis). Only 20 of 59 species (33.9%) had mean litter sizes of five or more. Mean gestation length differed among species (F = 30. 1, P < 0.0001, n — 14) and ranged from 18.0 days (Sorex isodori) to 31.5 days (Crocidura canariensis). Gestation lengths were more or less evenly distributed over that 14-day range. Neonate mass differed among species (F = 20.0, P < 0.0001, n = 22) and ranged from 0.24 g (Sorex minutus and Suncus etruscus) to 2.84 g (Suncus murinus). Only six of 22 species (27.3%) produced neonates that weighed one gram or more. The age at which young opened their eyes differed among species (F = 9.1, P < 0.0001, n = 17) and ranged from 8.6 days (Suncus murinus) to 22.0 days (Neomys anomalus and N. fodiens). Ages in other species were more or less evenly distributed across this range. The age at weaning differed among species (F = 2.8, P < 0.005, n = 21) and ranged from 18.0 days (Crocidura hirta and C. suaveolens) to 30.0 days (Sorex cornatus). Only six of 21 species (28.6%) weaned their offspring on or before 20 days of age. Body mass at weaning differed among species (F 1 2.0, P < 0.001, n — 17) and ranged from 2.1 g (Crocidura bicolor) to 30.4 g (Suncus murinus). Thirteen of the 21 species (76.5 %) weaned their young when they were less than 10 g. Growth rates to weaning differed among species (F = 3.1, P < 0.05, n — 17) and ranged from 0.06 g/day (C. bicolor) to 1.39 g/day (Suncus murinus). Fourteen of 17 species (82.4%) had growth rates of 0.5 g/day or less. Maximum life spans did not differ significantly among species and averaged 16.8 months for all species. Length of the breeding season differed among species (F = 3.4, P < 0.0005, n = 21) and ranged from 4.4 months (Sorex caecutiens) to year-round (Scutisorex somereni). Other species were more or less evenly distributed across this range. Differences Among Crocidura and Sorex Species When only Crocidura and Sorex species were considered, there were fewer differences than when all 93 species were considered. Maximum life span was not significantly different among species as before. Age at weaning and growth rate to weaning also were not significantly different. Despite the exclusion of the lightest and heaviest species (Suncus etruscus and S. murinus, respectively), male mass, female mass, and the combined mass of both sexes all showed differences among species (F = 36.0, n = 42; F — 30.9, n = 35; F = 26.5, n = 41; F < 0.0001, respectively). Similarly, neonate mass and mass at weaning differed among species (F = 4.7, n = 14; F = 6.0, n = 12; P < 0.01, respectively). Litter size, gestation length, age when eyes open, and length of the breeding season were also different among species (F = 9.8, n = 43, P<0.0001; F = 23.7, n = 9, P < 0.0001; F = 3.8, n = 10, 7’<0.05; F = 2.4, n = 16, P < 0.05, respectively). Differences Among Genera Mean male mass differed among genera (F = 2.4, n = 8, P < 0.05) and ranged from 4.7 g (Crypt ot is) to 33.5 g (Suncus). Litter size also differed among genera (F = 11.3, n = 11, F < 0.0001) and ranged from 1.9 (Scutisorex) to 6.8 (Neomys). Only Blarina, Neomys, and Sorex had mean litter sizes exceeding five. Gestation length differed among genera (F = 6.9, n = 6, P < 0.01) and ranged from 20.3 days (Blarina) to 29.7 days (Crocidura). Age when the eyes opened differed among genera (F = 6.0, n = 1, P < 0.01) and ranged from 11.6 days (Suncus) to 22.0 days (Neomys). Age at weaning differed among genera (F = 3.9, n = 1 , P < 0.05) and ranged from 19.2 days (Suncus) to 28.7 days (Neomys). Maximum life span differed among genera (F = 7.9, n — 1, P < 0.01), ranging from 13.5 months (Neomys) to 27.0 months (Crocidura). Length of the breeding season differed among genera (F = 2.8, n = 10, P < 0.05) and ranged from 5.0 months (Notiosorex) to year-round (Scutisorex). The following traits did not differ among genera: female mass, combined mass of both sexes, neonate mass, mass at weaning, and growth rate to weaning. Differences Between Crocidura and Sorex Adult mass of males and females showed that Crocidura were larger than Sorex (Table 2). Sorex females gave birth to larger litters of smaller neonates after a shorter gestation period compared to Crocidura females. Crocidura juveniles opened their eyes and are weaned earlier than Sorex juveniles. Crocidura lived longer and had longer breeding seasons than Sorex. Combined mass, body mass at weaning, and growth rate to weaning did not differ between genera. Differences Between the Two Subfamilies Of the 12 variables only two showed significant differences between the two subfamilies (Table 3). The Crocidurinae produced smaller litters but had longer breeding seasons compared to the Soricinae. Correlations Between Length of the Breeding Season and Life-History Traits Within Species.— Since the data were heterogeneous, only a few species and only female mass and litter size were examined in relation to the length of the breeding season within species. Within Blarina brevicauda, Sorex araneus, S. cinereus, or Suncus murinus, neither female mass nor litter size was 114 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 significantly correlated with the length of the breeding season. Among All Populations and Only in Crocidura and Sorex Populations. — Among all populations, both male and female weights were significantly correlated with the length of the breeding season r = 0.60, n = 36; r = 0.56, n = 39, respectively; P < 0.001), whereas combined weights were not. Litter size was negatively correlated with the length of the breeding season (r = -0.66, n = 80, P < 0.001) (Fig. 1). The only other significant correlation was between age at weaning and length of the breeding season (r = -0.88, n = 6, P < 0.05). Among Crocidura and Sorex populations, the only trait that was correlated with the length of the breeding season was litter size (r = -0.57, n = 54, P < 0.001). Among All Species and Only in Crocidura atul Sorex Species. — Among all species, length of the breeding season was negatively correlated with both litter size and the age when eyes open (r = -0.80, n = 27, P < 0.001; r = -0.61, n =13, P < 0.05, respectively). Male mass, female mass, gestation length, and neonate weight were positively correlated with the length of the breeding season (r = 0.50, n = 23; r = 0.43, n = 22; r = 0.61, n = 12; r = 0.53, n = 15; P < 0.05, respectively). Using Crocidura and Sorex species only, length of the breeding season was negatively correlated with litter size (r = -0.79, n = 16, P < 0.001) but positively correlated with gestation length (r = 0.76, n = 1 , P < 0.05). Among All Genera. — Among all genera, the length of the breeding season was correlated only with litter size (r — -0.83, n = 10, P < 0.01). Samples were too small for analysis by subfamily. Correlations Among Life-History Traits Within Species. — In Blarina brevicauda, Sorex araneus, S. cinereus, and Suncus murinus, no significant relationships were found between female mass and litter size. However, in S. murinus there was a negative correlation between litter size and gestation length (r = -0.78, n = 9; P < 0.05). Among All Populations and Only Among Crocidura atul Sorex Populations. — Of the possible 36 correlations, only 16 were significant among all populations (Table 4). Female mass was associated with only three other traits, but many developmental traits were correlated with one another. Larger females appeared to produce smaller litters (Fig. 2), but gave birth to and weaned larger young. Larger litters were associated with smaller young at birth and weaning, shorter gestation lengths, and later ages when eyes open. Gestation length was positively associated with neonate mass, but negatively associated with age at weaning and age when eyes open. Neonate mass was piositively correlated with mass at and growth to weaning, but negatively correlated with age at weaning and age when eyes open. Earlier age when eyes open was associated with earlier age at weaning, and greater mass at weaning was positively correlated with growth to weaning. Ten correlations were significant when only Crocidura and Sorex species were used (Table 4). The patterns were similar to those above except that the following relationships were no longer significant: litter size and neonate mass, litter size and mass at weaning, female mass and neonate mass, female mass and mass at weaning, gestation length and neonate weight, gestation length and age at weaning, and neonate mass and age at weaning. Also, in this comparison litter size was negatively correlated with maximum life span. Among All Species and Only Among Crocidura and Sorex Species.— Oi the 36 possible correlations, only 12 were significant among all species (Table 5). Greater female mass was associated with larger neonates, which grew more rapidly and weighed more at weaning. Litter size was negatively correlated with both gestation length and neonate mass, but positively correlated with both the age when eyes open and age at weaning. Gestation length was negatively correlated with the age when eyes open. Neonate mass was positively correlated with both mass at weaning and growth rate to weaning. The later the young opened their eyes, the later they were weaned. Also, the more rapid the growth to weaning, the heavier the young were when weaned. Using only Crocidura and Sorex species, 13 correlations were significant (Table 5). The patterns were the same as above with the following exceptions: litter size was negatively correlated with both growth rate to weaning and maximum life span. Also, age at weaning was not significantly correlated with litter size. Among Genera. — Eight correlations were significant among genera (Table 6). Female mass was positively correlated with neonate mass, mass at weaning, and growth to weaning. Litter size was negatively correlated with gestation length. Neonate mass was positively correlated with both mass at weaning and growth to weaning. Age at weaning was positively correlated with the age when eyes open. Also, mass at weaning was positively correlated with growth to weaning. Samples were too small for analyses within each subfamily. Influences of Adult Female Mass. — Ideally, the effects of body size should be controlled for, because previous studies have shown that it covaries with many other traits. In this study, body mass was usually correlated with one or three other traits (Tables 4, 5, and 6). Due to the structure of the data, adequate sample sizes for a partial correlation analysis (holding female mass constant) could be done using only species means with all species (Table 5). Controlling for female mass did alter some of the covariation among traits (Table 7). For example, the positive correlations between neonate mass and both mass at weaning and growth to weaning were no longer significant. Also, five new relationships became apparent: gestation length was positively correlated with both neonate mass and mass at weaning, neonate mass was negatively associated with both age when eyes open and age at weaning, and growth to weaning was positively correlated with age when eyes open. When correlations with length of the breeding season were considered, a number of changes occurred. Length of the breeding season was still negatively correlated with litter size and positively correlated with gestation length. However, the correlations between length of the breeding season, female mass, neonate mass, and age when eyes open were no longer significant. Also, length of the breeding season was negatively associated with growth to weaning. 1994 INNES — Soricid Life Histories 115 Discussion Quality of the Data Base This study has many of the drawbacks inherent in this type of review. First it must be assumed that the 93 species examined are representative of the approximately 260 soricid species. Also, sample sizes were often small and some traits were measured in slightly different ways in different studies. Other difficulties with this type of review have been addressed by P. Harvey and colleagues (Clutton-Brock and Harvey, 1984; Harvey and Mace, 1982; Harvey and Clutton-Brock, 1985). Among other things, they have pointed out that rarely are numbers of mammalian species evenly distributed across genera. Most of the life-history data collected were from Crocidura (39 species) and Sorex (32 species). Other genera were represented by one to four species. There is also a geographic bias. Species found in Europe and North America tended to be overrepresented relative to species in other areas. Another bias which has large statistical implications involves the genus Suncus. This genus is represented in this study by two species: S. etruscus and S. murinus. Both male and female adult weights show that they are the smallest and largest species, respectively, in the family. Adult males of S. etruscus weigh as little as 1.3 g (Fons, 1970), but S. murinus males can reach 177.0 g (Louch et al., 1966). In the latter species, body mass varies considerably from area to area. Those from Bangladesh are heavier than those from Sri Lanka which in turn are heavier than those from Japan (Tsubota et al., 1986). Suncus murinus is also unusual because growth and development of the young appear to be phenotypically plastic. Dryden and Ross (1971) found that a higher quality diet increased the growth rate and accelerated some development changes compared to young on a lower quality diet. Length of Breeding Season Length of the breeding season ranged from four months to year-round among all populations. Most breeding seasons are continuous, although in at least one species (Sori cuius caudatus) it is bimodal, occurring before and after the monsoon in Nepal (Mitchell, 1977). Considering all species, length of the breeding season was significantly different and common to all three taxonomic levels (among species, among genera, and between subfamilies). Similarly, when only Crocidura and Sorex species were considered, length of the breeding season was different among species as well as between genera. On average, members of the Crocidurinae have longer breeding seasons than members of the Soricinae. In Blarina brevicauda, Sorex araneus, S. cinereus, and Suncus murinus, length of the breeding season was not correlated with either female mass or litter size. Four traits among all populations, six among all species, and three among all genera were significantly correlated with the length of the breeding season. Common to all levels was a negative relationship between litter size and length of the breeding season. At the species level, this negative correlation occurs even if the data are adjusted for the effects of female mass. Among all populations of Crocidura and Sorex, and among all Crocidura and Sorex species, one and three traits were correlated with the length of the breeding season, respectively. Common to both levels was a negative relationship between litter size and length of the breeding season. The relationship between litter size and length of the breeding season in small mammals remains unclear. A number of studies have shown that litter size increases with increasing latitude (or altitude) within species (Spencer and Steinhoff, 1968; limes, 1978; McLaren and Kirkland, 1979; Halfpenny, 1980). Usually, the negative relationship between the two variables is much stronger among species than within species (Lord, 1960; limes, 1978; McLaren and Kirkland, 1979). However, whether latitude (or altitude) accurately predicts the length of the breeding season is unknown. Millar (1984) found a significant negative correlation between litter size and length of the breeding season in 15 species of Peromyscus, but not in P. maniculatus . He also found that many other traits did not vary with the length of the breeding season. Thus, soricids appear to fit a similar pattern in that litter si2K and length of the breeding season are negatively correlated, but only at or above the species level, and few other traits covary with length of the breeding season. Differences in Life-History Traits Considering all species, only litter size was different and common to all three taxonomic levels. The number of traits that were different decreased as the taxonomic level increased (ten among species, six among genera, one between subfamilies). This result is not surprising because the amount of variation decreased as data were averaged over each successively higher taxonomic level. Since litter size was the only trait that differed at all levels, this suggests that there is a considerable amount of variation in the other traits that are independent of the two subfamilies. In a broader mammalian survey. Steams (1983) found that ten traits were significantly different among four orders as well as among eight families. When only Crocidura and Sorex species were used, eight traits differed among species and seven traits between the two genera. Traits common to both levels indicate that Crocidura are larger as adults (both males and females) and as neonates, have smaller litters and open their eyes earlier than Sorex. This suggests that these two groups have evolved distinct life histories. Earlier, Vogel (1972n, \912b) pointed out similar differences in a comparison (mainly) of C. russula and S. araneus. The contrasting life histories of Crocidura and Sorex species, as well as between crocidurine and soricine species, appear to be related to their metabolic rates (Vogel, 1976, 1980; Genoud, 1988). These authors found that the Crocidurinae have lower metabolic rates than the Soricinae and this may be related to the respective Paleotropic versus Holarctic origins of the two subfamilies. They also suggested that this is related to differences in reproduction, development, and longevity. For example, Crocidura species usually have litter sizes less than five, whereas Sorex species usually have litter sizes greater than five. These patterns may have ultimately resulted from differences in climate (warm versus cold) or resource 116 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 availability (unpredictable versus seasonally predictable) (Genoud, 1988). Correlations Among Life-History Traits Considering all species, the number of significant correlations decreased with increasing taxonomic level. Among populations, among species, and among genera, 16, 12, and 8 correlations were significant, respectively. Seven relationships were common to all three levels. Larger females produced larger neonates and weaned larger young compared to smaller females. These relationships have been found in broader surveys (Millar, 1977; Read and Harvey, 1989). Litter size was negatively correlated with gestation length and this pattern is also evident across all eutherian orders (Read and Harvey, 1989). Larger neonates grew more rapidly and were weaned at a greater weight than smaller neonates. Also, age when eyes open was positively correlated with age at weaning, and weight at weaning was positively correlated with growth to weaning. These latter relationships have not been examined in broader surveys, but Modi (1984) found no significant relationships between either growth rate to or age at weaning with age when eyes open in 15 taxa of Perotnyscus. When only Crocidura and Sorex were used, three more significant correlations were found and common to both the population and species levels. Litter size was negatively correlated with maximum life span, but positively correlated with age when eyes open. Gestation length was also negatively correlated with age when eyes open. The negative tradeoff between litter size and life span has been found across eutherian orders (Read and Harvey, 1989). Modi (1984) found a negative correlation between litter size and age when eyes open. A number of the patterns of covariation were found among species whether the effects of female mass were controlled for or not (Tables 5 and 7). This suggests that these relationships occur independently of adult mass. For example, there was a negative correlation between litter size and neonate mass whether the influence of female mass was adjusted for or not. This situation is evident across all eutherian orders (Read and Harvey, 1989). However, when female mass was adjusted for, a number of new relationships were evident. For example, gestation length became significantly correlated with neonate mass. Thus, among soricids, covariation among other traits may or may not be independent of adult female mass. Future Research Needs and Summary A number of important traits were not quantified in this review. For example, age at sexual maturity can have a profound effect on the rate of increase of a population (Cole, 1954). Under laboratory conditions, young females of Blarina brevicauda, Crocidura russula, Cryptotis parva, and Suncus rnurinus can conceive at approximately 30 days of age (Blus, 1971; Hellwing, 1971; Mock and Conaway, 1976; Dryden, 1969, respectively). Mock and Conaway (1976) also reported that C. par\’a males could breed at 36 days of age. In a review of the literature, Jeanmarie-Besan^on (1988) found that the number of females maturing in the year of their birth was highly variable under natural conditions, especially among Sorex species. He concluded that Crocidura species were more likely to breed as young-of-the-year than Sorex species. This situation may be the result of their respective Paleotropic versus Holarctic origins. However, not all European populations of Crocidura breed in the year of their birth (Bishop and Delany, 1963), and all studies on age at sexual maturation in Crocidura have been conducted in temperate regions. Similar studies are needed in tropical regions. Another trait that has been relatively neglected is the number of litters per season or per lifetime. Necropsies of snap- or pitfall-trapped specimens have led most investigators to conclude that females produce at least two litters each season. This is based on the observation that females can be pregnant and lactating simultaneously and thus undergo a postpartum estrus. A postpartum estrus has been documented in seven soricid genera: Blarina (Lutz, 1964), Cryptotis (e.g., Blair, 1938), Crocidura (e.g., Hutterer et al., 1987), Myosorex (Baxter and Lloyd, 1980), Neomys (Michalak, 1987a), Sorex (e.g., Skaren, 1979), and Suncus (Dryden, 1969). However, determining the number of litters based on necropsied samples has drawbacks and this trait may be best measured by mark- recapture studies (Innes and Millar, 1987). Buckner (1966) used the latter technique and found that female Blarina brevicauda, Sorex arcticus, and S. cinereus averaged about two litters per lifetime (range = 1-5). Additional live-trapping studies to quantify this trait as well as survival rates could prove useful. Generalizing across all species, adult soricids are small, with most weighing < 10.0 g. Neonates are small, usually weighing < 1.0 g. Juveniles are usually weaned when body mass is only 2.0 g below adult mass. Growth rate to weaning is usually <0.3 g/day. Gestation length and age at weaning are variable, with both traits ranging from about 18 to 30 days. The age when eyes open ranges from 8 to 22 days. Litter size is extremely variable, but most species have litters of less than five. Maximum life span averaged 17 months across all species. Length of the breeding season is variable, but usually exceeds four months. Acknowledgments I thank J. Ryan for informing me of this Colloquium and sending me many useful references. I express my gratitude to J. Merritt and staff of the Powdermill Nature Reserve for hosting the Colloquium. I am greatly appreciative of the help of Z. Pucek, who kindly took the time to extract pertinent data for me from several Polish papers. The following people are sincerely thanked for sending me unpublished data: T. French, S. Hellwing, A. Kaikusalo, G. Kirkland, Jr., H. Mover, C. Nores, E. Pankakoski, and E. Wolk. I also thank the following people who responded to my letters, sent reprints, or helped in some way: L. Carraway, M. Cawthom, S. Churchfield, L. Getz, E. Gould, I. Hanski, R. Hoffmann, R. Hutterer, Y. Kaneko, J. Layne, C. Long, J. Meester, S. Mercer, S. Oda, W. Podushka, R. Rose, R. Rudd, B. Sheftel, W. Sheppe, C. van Zyll de Jong, P. Vogel, and J. Whitaker, Jr. M. Ossenkopp extracted pertinent data from most of the papers that were in German. Two anonymous reviewers provided useful comments. 1994 INNES — Soricid Life Histories 117 This research was supported by a grant from the Natural Sciences and Engineering Research Council of Canada to M. Kavaliers (University of Western Ontario) and a grant from Foundation Western. Literature Cited Baxter, R. M., and C. N. V. Lloyd. 1980. Notes on the reproduction and postnatal development of the forest shrew. Acta Theriologica, 25:31-38. Bishop, I. R., and M. J. Delany. 1963. Life histories of small mammals in the Channel Islands in 1960-61. Proceedings of the Zoological Society of London, 141:515-526. Blair, W. F. 1938. Ecological relationships of the mammals of the Bird Creek region, northeastern Oklahoma. The American Midland Naturalist, 20:473-526. Blus, L. J. 1971. Reproduction and survival of short-tailed shrews (Blarina brevicauda) in captivity. Laboratory Animal Science, 21:884-891. Boyce, M. S. 1978. Climatic variability and body size variation in the muskrats {Ondatra zibethicus) of North America. Oecologia, 36:1-19. 1979. Seasonality and patterns of natural selection for life histories. The American Naturalist, 1 14:569-583. 1988. Evolution of life histories: Theory and patterns from mammals. Pp. 3-30, in Evolution of Life Histories of Mammals (M. S. Boyce, ed.), Yale University Press, New Haven, 337 pp. Buckner, C. H. 1966. Populations and ecological relationships of shrews in tamarack bogs of southeastern Manitoba. Journal of Mammalogy, 47:181-194. Cameron, G. N., and P. A. McClure. 1988. Geographic variation in life history traits of the hispid cotton rat (Sigmodon hispid us). Pp. 33-64, in Evolution of Life Histories of Mammals (M. S. Boyce, ed.), Yale University Press, New Haven, 337 pp. Catzeflis, F., T. Maddalena, S. Hellwing, and P. Vogel. 1985. Unexpected findings on the taxonomic status of East Mediterranean Crocidura russula auct. (Mammalia, Insectivora). Zeitschrift fur Saugetierkunde, 50:185-201. Clutton-Brock, T. H., and P. H. Harvey. 1984. Comparative approaches to investigating adaptation. Pp. 7-29, in Behavioural Ecology: An Evolutionary Approach. 2nd ed. (N. B. Davies and J. R. Krebs, eds.), Blackwell Scientific Publications, Oxford, 493 pp. Cole, L. C. 1954. The population consequences of life history phenomena. The Quarterly Review of Biology, 29:103-137. Connor, P. F. 1966. The mammals of the Tug Hill Plateau, New York. Bulletin of the New York State Museum and Science Service, 406:1-82. Corbet, G. B., and J. E. Hill. 1986. A World List of Mammalian Species. University Printing House, Oxford, 254 pp. Croin Michielsen, N. 1966. Intraspecific and interspecific competition in the shrews Sorex araneus L. and Sorex minutus L. Archives neerlandaises de Zoologie, 17:73-174. Crowcroft, P. 1957. The Life of the Shrew. Reinhart Publishing Company, London, 166 pp. Dapson, R. W. 1968. Reproduction and age structure in a population of short -tailed shrews, Blarina brevicauda. Journal of Mammalogy, 49:205-214. Dryden, G. L. 1968. Growth and development of Suncus murinus in captivity on Guam. Journal of Mammalogy, 49:51-62. 1969. Reproduction in Suncus murinus. Journal of Reproduction and Fertility, Supplement, 6:377-396. Dryden, G. L., and J. M. Ross. 1971. Enhanced growth and development of captive musk shrews, Suncus murinus, on an improved diet. Growth, 35:31 1-325. Eisenberg, j. F. 1981 . The Mammalian Radiations: An Analysis of Trends in Evolution, Adaptation, and Behavior. The University Chicago Press, Chicago, 610 pp. Fons, R. 1970. Contribution a la connaissance de la musaraigne etrusque Suncus etruscus (Savi, 1822) (Mammifere Soricidae). Vie et Milieu, 21C:209-218. Forsyth, D. J. 1976. A field study of growth and development of nestling masked shrews (Sorex cinereus). Journal of Mammalogy, 57:708-721. Genoud, M. 1988. Energetic strategies of shrews: Ecological constraints and evolutionary implications. Mammal Review, 18:173-193. Genoud, M., and P. Vogel. 1990. Energy requirements during reproduction and reproductive effort in shrews (Soricidae). Journal of Zoology (London), 220:41-60. Gittleman, j. L. 1986. Carnivore life history patterns: Allometric, phylogenetic , and ecological associations . The American N aturalist , 127:744-771. Halfpenny, J. C. 1980. Reproductive strategies: Intra and interspecific comparisons within the genus Peromyscus. Unpublished Ph D. dissert.. The University of Colorado, Boulder, 160 pp. Hamilton, W. J., Jr. 1940. The biology of the smoky shrew (Sorex fiimeus fumeus Miller). Zoologica, 25:473-492. Harvey, P. H., and T. H. Clutton-Brock. 1985. Life history variation in primates. Evolution, 39:559-581. Harvey, P. H., and G. M. Mace. 1982. Comparisons between taxa and adaptive trends: Problems and methodology. Pp. 343-361 , in Current Problems in Sociobiology (Kings College Sociobiology Group, eds.), Cambridge University Press, Cambridge, 394 pp. Harvey, P. H., D. E. L. Promislow, and A. F. Read. 1989. Causes and correlates of life history differences among mammals. Pp. 305-318, in Comparative Socioecology. The Behavioural Ecology of Humans and Other Mammals (V. Standen and R. A. Foley, eds.), Blackwell Scientific Publications, Oxford, 519 pp. Hellwing, S. 1970. Reproduction of the white-toothed shrew Crocidura russula monacha Thomas in captivity. Israel Journal of Zoology, 19:177-178. 1971. Maintenance and reproduction in the white-toothed shrew, Crocidura russula monacha Thomas, in captivity. Zeitschrift fur Saugetierkunde, 36:103-113. 1973a. Husbandry and breeding of white-toothed shrews (Crocidurinae) in the research zoo of the Tel-Aviv University. International Zoo Yearbook, 13:127-134. \913b. The postnatal development of the white-toothed shrew Crocidura russula monacha in captivity. Zeitschrift fur Saugetierkunde, 38:257-270. Hoffmeister, D. F., and W. W. Goodpaster. 1962. Life history of the desert shrew, Notiosorex crawfordi. The Southwestern Naturalist, 7:236-252. Honaki, j. H., K. E. Kin man, and J. W. Koeppl (eds.). 1982. Mammal Species of the World. Allen Press, Inc., and The Association of Systematics Collections, Lawrence, Kansas, ix -f 694 pp. Hutterer, R. 1976. Beobachtungen zur Geburt und Jungendentwicklung der Zwergspitzmaus, Sorex minutus L. (Soricidae-Insectivora). Zeitschrift fiir Saugetierkunde, 41:1-22. 1985. Anatomical adaptations of shrews. Mammal Review, 15:43-55. Hutterer, R., and W. Verheyen. 1985. A new species of shrew, genus Sylvisorex, from Rwanda and Zaire (Insectivora: Soricidae). 118 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 Zeitschrift fiir Saugetierkunde, 50:266-271. Hutterer, R., L. F. Lopez-Jurado, and P. Vogel. 1987. The shrews of the eastern Canary Islands: A new species (Mammalia: Soricidae). Journal of Natural History, 21:1347-1357. INNES, D. G. L. 1978. A reexamination of litter size in some North American microtines. Canadian Journal of Zoology, 56:1488-1496. Innes, D. G. L., and J. S. Millar. 1987. The mean number of litters per breeding season in small mammal populations: A comparison of methods. Journal of Mammalogy, 68:675-678. Jeanmarie-Besanqon, F. 1988. Sexual maturity of Crocidura russula (Insectivora: Soricidae). Acta Theriologica, 33:477-485. Johnston, R. F., and R. L. Rudd. 1957. Breeding of the salt marsh shrew. Journal of Mammalogy, 38:157-163. Kemg, J. A., J. C. Deshaies, and R. Webster. 1963. Age of weaning in two subspecies of mice. Science, 139:483-484. Lord, R. D. 1960. Litter size and latitude in North American mammals. The American Midland Naturalist, 64:488-499. LoucH, C. D., A. K. Ghosh, and B. C. Pal. 1966. Seasonal changes in weight and reproductive activity of Suncus murinus in West Bengal, India. Journal of Mammalogy, 47:73-78. Lutz, J. E. 1964. Natural history of the short-tailed shrew in southeastern Michigan. Unpublished Ph.D. dissert.. The University of Michigan, Ann Arbor, 341 pp. May, R. M., and D. I. RUBENSTEIN. 1984. Reproductive strategies. Pp. 1-23, in Reproduction in Mammals. Book 4. Reproductive Fitness (C. R. Austin and R. V. Short, eds.), Cambridge University Press, Cambridge, 241 pp. McLaren, S. B., and G. L. Kirkland, Jr. 1979. Geographic variation in litter size of small mammals in the central Appalachian region. Proceedings of the Pennsylvania Academy of Science, 53:123-126. Michalak, I. 1987a. Growth and postnatal development of the European water shrew. Acta Theriologica, 32:261-288. \9%lb. Keeping and breeding the Eurasian water shrew Neomys fodiens under laboratory conditions. International Zoo Yearbook, 26:223-228. Millar, J. S. 1977. Adaptive features of mammalian reproduction. Evolution, 31:370-386. 1984. Reproduction and survival of Peromyscus in seasonal environments. Pp. 253-266, in Winter Ecology of Small Mammals (J. F. Merritt, ed.), Carnegie Museum of Natural History Special Publication, 10:ix + 380. Mitchell, R. M. 1977. Accounts of Nepalese mammals and analysis of the host-ectoparasite data by computer techniques. Unpublished Ph.D. dissert., Iowa State University, Ames, 558 pp. Mock, O. B., and C. H. Conaway. 1976. Reproduction in the least shrew (Cryptotis parva) in captivity. Pp. 59-74, in The Laboratory Animal in the Study of Reproduction (T. Antikatzides, S. Erichsen, and A. Spiegel, eds.), Gustav Fisher Verlag, Stuttgart, 186 pp. Modi, W. S. 1984. Reproductive tactics among deer mice of the genus Peromyscus. Canadian Journal of Zoology, 62:2576-2581. Mover, H., S. Hellwing, and A. Ar. 1988. Energetic costs of gestation in the white-toothed shrew Crocidura russula monacha (Soricidae, Insectivora). Physiological Zoology, 61:17-25. Nowack, R. M., and j. L. Paradiso. 1983. Walker’s Mammals of the World. 4th ed. The Johns Hopkins University Press, Baltimore, Maryland, 568 pp. Pearson, O. P. 1944. Reproduction in the shrew (Blarina brevicauda Say). The American Journal of Anatomy, 75:39-93. Pruitt, W. 1957. A survey of the mammalian family Soricidae (shrews). Saugetierkundliche Mitteilungen, 5:18-27. Read, A. F., and P. H. Harvey. 1989. Life history differences among the eutherian radiations. Journal of Zoology (London), 219:329-353. SPSS, Inc. 1988. SPSS-X User’s Guide. 3rd ed. SPSS, Inc., Chicago, Illinois, 1072 pp. S KAREN, U. 1979. Variation, breeding and moulting in Sorex isodon Turov. Acta Zoologica Fennici, 159:1-30. SOKAL, R. R., AND F. J. Rohlf. 1981. Biometry: The principles and practice of statistics in biological research. 2nd ed. W. H. Freeman and Company, San Francisco, California, 859 pp. Spencer, A. W., and H. W. Steinhoff. 1968. An explanation of geographic variation in litter size. Journal of Mammalogy, 49:281-286. Stearns, S. C. 1983. The influence of size and phylogeny on patterns of covariation in mammals. Oikos, 41:173-187. Tsubota, Y., T. Namikawa, T. Nishida, A. Adachi, and H. W. Cyril. 1986. Morphology and reproduction of wild musk shrews, Suncus murinus, in Sri Lanka. Report of the Society for Researchers on Native Livestock of Japan, 11:241-250 (not seen, cited in Ishikawa and Namikawa, 1987, see under S. murinus). Vogel, P. 1970. Biologische Beobachtungen an Etruskerspitzmausen (Suncus etruscus Savi, 1832). Zeitschrift fiir Saugetierkunde, 35:173-185. 1972a. Vergleichende Untersuchung zum Ontogenesemodus einheimischer Soriciden (Crocidura russula, Sorex araneus und Neomys fodiens). Revue Suisse de Zoologie, 79:1201-1332. 1972/j. Beitrag zur Fortpflanzungsbiologie der Gattungen SorcY, Neomys und Crocidura (Soricidae). Verhandlungen der naturforschenden Gesellschaft in Basel, 82:165-192. 1976. Energy consumption of European and African shrews. Acta Theriologica, 21: 195-206. 1980. Metabolic levels and biological strategies in shrews. Pp. 170-180, in Comparative Physiology: Primitive Mammals (K. Schmidt-Nielsen, L. Bolis, and C. R. Taylor, eds.), Cambridge University Press, Cambridge, 338 pp. Vogel, P., T. Maddalena, and F. Catzeflis. 1986. A contribution to the taxonomy and ecology of shrews (Crocidura zimmermanni and C. suaveolens) from Crete and Turkey. Acta Theriologica, 31:537-545. Zeveloff, S. L, and M. S. Boyce. 1988. Body size patterns in North American mammal faunas. Pp. 123-146, in Evolution of Life Histories of Mammals (M. S. Boyce, ed.), Yale University Press, New Haven, Connecticut, 337 pp. Appendix Sources for Table 1 Blarina brevicauda: Bailey, B. 1929. Mammals of Sherburne County, Minnesota. Journal of Mammalogy, 10:153-164. Blair, W. F. 1948. Population density, life span, and mortality rates of small mammals in the blue-grass meadow and blue- grass field associations of southern Michigan. The American Midland Naturalist, 40:395-419. Blus, L. J. 1971 (see Literature Cited). Blus, L. j., and D. a. Johnson. 1969. Adoption of a nestling house mouse by a female short-tailed shrew. The American 1994 INNES— Soricid Life Histories 119 Midland Naturalist, 81:583-584. Bowles, J. B. 1975. Distribution and biogeography of mammals of Iowa. Special Publications, The Museum, Texas Tech University, 9:1-184. Brimley, C. S. 1923. Breeding dates of small mammals at Raleigh, North Carolina. Journal of Mammalogy, 4:263-264. Buckner, C. H. 1966 (see Literature Cited). Campbell, C. A., and A. I. Dag. 1972. Mammals of Waterloo and South Wellington Counties. Published privately, Waterloo, Ontario, 130 pp. Connor, P. F. 1960. The small mammals of Otsego and Schohaire counties. New York. Bulletin of the New York State Museum and Science Service, 382:1-84. 1966 (see Literature Cited). Crile, G., and D. P. Quiring. 1940. A record of the body weight and certain organ and glands weight of 3690 animals. Ohio Journal of Science, 40:219-259. Dapson, R. W. 1968a (see Literature Cited). 1968&. Growth patterns in a post-juvenile population of short-tailed shrews (Blarina brevicauda). The American Midland Naturalist, 79:118-129. Dice, L. R. 1923. Notes on some mammals of Riley County, Kansas. Journal of Mammalogy, 4:107-112. Fowle, C. D., and R. Y. Edwards. 1955. An unusual abundance of short-tailed shrews, Blarina brevicauda. Journal of Mammalogy, 36:36-41. French, T. W. Unpublished data. Getz, L. L. 1989. A 14-year study of Blarina brevicauda populations in east-central Illinois. Journal of Mammalogy, 70:58-66. Gifford, C. L., and R. Whitebread. 1951. Mammal Survey of South Central Pennsylvania. Pennsylvania Game Commission, Harrisburg, 75 pp. Gould, E. 1969. Communication in three genera of shrews (Soricidae): Suncus, Blarina and CryptoHs. Communications in Behavioral Biology, A, 3:11-31. Grimm, W. C., and H. A. Roberts. 1950. Mammal Survey of Southwestern Pennsylvania. Pennsylvania Game Commission, Harrisburg, 99 pp. Grimm, W. C., and R. Whitebread. 1952. Mammal Survey of Northeastern Pennsylvania. Pennsylvania Game Commission, Harrisburg, 82 pp. Guilday, J. E. 1957. Individual and geographic variation in Blarina brevicauda from Pennsylvania. Annals of Carnegie Museum, 35:41-68. Hamilton, W. J., Jr. 1929. Breeding habits of the short-tailed shrew, Blarina brevicauda. Journal of Mammalogy, 10:125-134. 1931. Habits of the short-tailed shrew, Blarina brevicauda (Say). Ohio Journal of Science, 31:97-106. 1943. The Mammals of Eastern United States. Comstock Publishing Company, Ithaca, New York, 432 pp. 1949. The reproductive rates of some small mammals. Journal of Mammalogy, 30:257-260. Harper, F. 1929. Notes on mammals of the Adirondacks. New York State Museum Handbook, 8:51-118. Hoffmeister, D. F. 1989. Mammals of Illinois. University of Illinois Press, Urbana, 348 pp. Innes, D. G. L., and j. F. Bendell. Unpublished data. Iverson, S. L., and B. N. Turner. 1976. Small Mammal Radioecology: Natural Reproductive Patterns of Seven Species. Atomic Energy of Canada Limited, AECL-5393, 53 pp. Jones, J. K., Jr. 1964. Distribution and taxonomy of mammals of Nebraska. University of Kansas Publications, Museum of Natural History, 16:1-356. Kirkland, G. L., Jr. 1978. The short-tailed shrew, Blarina brevicauda (Say), in the central mountains of West Virginia. Proceedings of the Pennsylvania Academy of Sciences, 52:126-130. Unpublished data. Komarek, E. V., AND R. Komarek. 1938. Mammals of the Great Smoky Mountains. Bulletin of the Chicago Academy of Sciences, 5:137-162. Layne, j. N. 1958. Notes on mammals of southern Illinois. The American Midland Naturalist, 60:219-254. Lindsay, D. M. 1960. Mammals of Ripley and Jefferson counties, Indiana. Journal of Mammalogy, 41:253-262. Linzey, D. W., and a. V. Linzey. 1968. Mammals of the Great Smoky Mountains National Park. Journal of the Elisha Mitchell Scientific Society, 84:384-414. Lowery, G. H., Jr. 1974. The Mammals of Louisiana and Its Adjacent Waters. Louisiana State University Press, Baton Rouge, 565 pp. Lutz, J. E. 1964 (see Literature Cited). Manville, R. H. 1949. A study of small mammal populations in northern Michigan. Miscellaneous Publications of the Museum of Zoology, University of Michigan, 73:1-83. Martin, I. G. 1982. Maternal behavior of a short-tailed shrew {Blarina brevicauda). Acta Theriologica, 27: 153-156. Merritt, J. F. 1986. Winter survival adaptations of the short- tailed shrew (Blarina brevicauda) in an Appalachian montane forest. Journal of Mammalogy, 67:450-464. MillER-Ben-ShauL, D. 1962. Short-tailed shrews (Blarina brevicauda) in captivity. International Zoo Yearbook, 4:121-123. Mumford, R. E. 1969. Distribution of the Mammals of Indiana. Indiana Academy of Science, Indianapolis, 1 14 pp. Mumford, R. E., and J. O. Whitaker, Jr. 1982. Mammals of Indiana. Indiana University Press, Bloomington, 537 pp. O’Farrel, M. j., D. W. Kaufman, J. B. Gentry, and M. H. Smith. 1977. Reproductive patterns of some small mammals in South Carolina. Florida Scientist, 40:76-84. Pearson, O. P. 1944 (see Literature Cited). 1945. Longevity of the short-tailed shrew. The American Midland Naturalist, 34:531-546. Rand, A. L., and P. Host. 1942. Results of the Archbold expeditions. Number 45. Mammal notes from Highland County, Florida. Bulletin of the American Museum of Natural History, 80:1-21. Richmond, N. D., and. H. R. Roslund. 1949. Mammal Survey of Northwestern Pennsylvania. Pennsylvania Game Commission and United States Fish and Wildlife Service, Harrisburg, 67 pp. Rood, J. P. 1958. Habits of the short-tailed shrew in captivity. Journal of Mammalogy, 39:499-507. Saunders, W. E. 1932. Notes on the mammals of Ontario. Transactions of the Royal Canadian Institute, 18:271-309. Timm, R. M. 1975. Distribution, natural history and parasites of mammals of Cook County, Minnesota. Occasional Papers of the Bell Museum of Natural History, The University of Minnesota, 14:1-56. 120 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 Townsend, M. T. 1935. Studies on some of the small mammals of central New York. Roosevelt Wildlife Annals, 4:6-120. B. carolinensis: Hoffmeister, D. F. 1989 (see under B. brevicauda). Crocidura attenuata: Newton, P. N., M. R. W. Rands, and C. G. R. Bowden. 1990. A collection of small mammals from eastern Nepal. Mammalia, 54:239-244. C. beatus: Heany, L. a., and E. a. Rickart. Unpublished data. Maddalena, T., and C. Maddalena. Unpublished data. C. bicolor: Ansell, W. F. H. 1964. Captivity behaviour and postnatal development of the shrew Crocidura bicolor. Proceedings of the Zoological Society of London, 142:123-127. Dippenaar, N. J. 1979. Notes on the early post-natal development and behaviour of the tiny musk shrew, Crocidura bicolor Bocage, 1889 (Insectivora: Soricidae). Mammalia, 43:83-91. Sheppe, W. a. 1973. Notes on Zambian rodents and shrews. Puku, 7:167-190. Smithers, R. H. N. 1971. The Mammals of Botswana. National Museums of Rhodesia, Museum Memoir 4, 340 pp. Watson, C. R. B., and R. T. Watson. 1986. Observations on the postnatal development of the tiny musk shrew, Crocidura bicolor. South African Journal of Zoology, 21:352-354. C. bottegi: Happold, D. C. D. 1987. The Mammals of Nigeria. Claredon Press, Oxford, 402 pp. Heim de Balsac, H., and V. Aellen. 1958. Les Soricidae de basse Cote-dTvoire. Revue Suisse de Zoologie, 65:921-956. Maddalena, T., and P. Vogel. Unpublished data. C. canariensis: Hutterer, R., L. F. Lopez, and P. Vogel. 1987 (see Literature Cited). Martin, A., R. Hutterer, and G. B. Corbet. 1984. On the presence of shrews in the Canary Islands. Bonner Zoologische Beitrage, 35:5-14 (not seen, cited in Molina and Hutterer, 1989, see under C. osorio). C. crossei: Happold, D. C. D. 1987 (see under C. bottegi). Heim de Balsac, H., and V. Aellen. 1958 (see under C. bottegi). C. dolichura and C. douceti: Happold, D. C. D. 1987 (see under C. bottegi). C. flavescens: Bauchot, R., and H. Stephan. 1966. Donnees nouvelles sur I’encephalisationdes insectivoresetdes prosimiens. Mammalia, 30:160-196. Delany, M. J. 1964. An ecological study of the small mammals in the Queen Elizabeth Park, Uganda. Revue de Zoologie et de Botanique Africaines, 70:129-147. Dieterlen, F., and H. Heim de Balsac. 1979. Zur Okologie und Taxonomie der Spitzmause (Soricidae) des Kivu-Gebietes. Saugetierkundliche Mitteilungen, 27:241-287. Happold, D. C. D. 1987 (see under C. bottegi). Marlow, B. J. G. 1955. Observations on the herero musk shrew, Crocidura flavescens hereto St. Leger, in captivity. Proceedings of the Zoological Society of London, 124:803-808. Rowe-Rowe, D. T., and j. Meester. 1982. Population dynamics of small mammals in the Drakensberg of Natal, South Africa. Zeitschrift fiir Saugetierkunde, 47:347-356. Sheppe, W. 1972. The annual cycle of small mammal populations on a Zambian floodplain. Journal of Mammalogy, 53:445-460. C. foxi: Happold, D. C. D. 1987 (see under C. bottegi). C. fuscomurina: Happold, D. C. D. 1987 (see under C. bottegi). Maddalena, T. Unpublished data. C. gracilipes: Bauchot, R., and H. Stephan. 1966 (see under C. flavescens). Hollister, N. 1918. East African mammals in the United States National Museum. I. Insectivora, Chiroptera and Carnivora. Bulletin of the American Museum of Natural History, 35:663-680 (not seen, cited in Dieterlen and Heim de Balsac, 1979, see under C. flavescens). C. grandiceps: Hutterer, R. 1983. Crocidura grandiceps, eine neue Spitzmaus aus Westafrika. Revue Suisse de Zoologie, 90:699-707. C. grayi: Heany, L. R., and E. A. Rickart. Unpublished data. C. hirta: Meester, J. 1960. Shrews in captivity. African Wild Life, 14:57-63. Sheppe, W. 1972 (see under C. flavescens). _. 1973 (see under C. bicolor). Sheppe, W., and P. Haas. 1981. The annual cycle of small mammal populations along the Chobe River, Botswana. Mammalia, 45:157-176. Shortridge, G. C. 1934. The Mammals of South West Africa. Windmill Press, Kingswood, 1:1-437. Smithers, R. H. N. 1971 (see under C. bicolor). C. horsfieldi: Mitchell, R. M. 1977 (see Literature Cited). C. jacksoni : Bauchot, R., and H. Stephan. 1966 (see under C. flavescens). C. jouveneta: Maddalena, T., and P. Vogel. Unpublished data. C. kivuana and C. lanosa: Dieterlen, F., and H. Heim de Balsac. 1979 (see under C. flavescens). C. leucodon: Frank, F. 1953. Beitrag zur Biologie, insbesondere Paarungsbiologie der Feldspitzmaus, {Crocidura leucodon). Bonner Zoologische Beitrage, 4:187-194. 1954. Zur Jugendentwicklung der Feldspitzmaus {Crocidura leucodon Herm.). Bonner Zoologische Beitrage, 5:173-178. Gaffrey, G. 1961. Merkmale der wildlebenden Saugetiere Mitteleuropas. Geest und Portig, Leipzig (not seen, cited in Vogel, \912b, see Literature Cited). 1994 INNES — Soricid Life Histories 121 Herter, K. 1957. Das Verhalten der Insektivoren. Pp. 1-50, in Handbuch der Zoologie (W. Kukenthal, ed.), Walter de Gruyter, Berlin, Vol. 10 (not seen, cited in Dryden, 1968, and Vlasak, 1972, see Literature Cited and under C. suaveolens, respectively). Meylan, a. 1967. Les petits mammiferes terrestries du Valais Central (Suisse). Mammalia, 31:225-245. Wahlstrom, a. 1929. Beitrage zur Biologie von Crocidura leucodon (Herm.). Zeitschrift fiir Saugetierkunde, 4:157-185 (not seen, cited by Frank, 1954, and Michalak, 1987a, see under C. leucodon and Literature Cited, respectively). Yalden, D. W., P. a. Morris, and J. Harper. 1973. Studies on the comparative ecology of some French small mammals. Mammalia, 37:257-276. C. littoralis: Dieterlen, F., and H. Heim de Balsac. 1979 (see under C. flavescens). C. luna: Maddalena, T. Unpublished data. C. lusitania: Maddalena, T., and C. Maddalena. Unpublished data. C. mariquensis: Goulden, E. a., and J. Meester. 1978. Notes on the behaviour of Crocidura and Myosorex (Mammalia: Soricidae) in captivity. Mammalia, 42:197-207. Sheppe, W. 1973 (see under C. bicolor). Sheppe, W., and P. Haas. 1981 (see under C. hirta). Smithers, R. H. N. 1971 (see under C. bicolor). C. nanilla: Maddalena, T., and P. Vogel. Unpublished data. C. nigricans: Sheppe, W. 1973 (see under C. bicolor). C. nigrofusca: Maddalena, T. Unpublished data. C. niobe: Bauchot, R., and H. Stephan. 1966 (see under C. flavescens). C. olivieri: Genoud, M., and P. Vogel. 1990 (see Literature Cited). GrOnwald, a., and F. P. Mohres. 1974. Beobachtugen zur Jugendentwicklung und Karawanenbildung bei WeiBzahnspitzmausen (Soricidae-Crocidurinae). Zeitschrift fiir Saugetierkunde, 39:321-337. Maddalena, T. Unpublished data. Maddalena, T., and C. Maddalena. Unpublished data. Maddalena, T., and P. Vogel. Unpublished data. Verschuren, j., and j. Meester. 1977. Note sur les Soricidae (Insectivora) du Nimba liberien. Mammalia, 41:291-299. C. osorio: Molina, O. M., and R. Hutterer. 1989. A cryptic new species of Crocidura from Gran Canaria and Tenerife, Canary Islands (Mammalia: Soricidae). Bonner Zoologische Beitrage, 40:85-97. C. poensis: Happold, D. C. D. 1987 (see under C. bottegi). Heim de Balsac, H., and V. Aellen. 1958 (see under C. bottegi). Maddalena, T., and P. Vogel. Unpublished data. C. russula: Bauchot, R., and H. Stephan. 1966 (see under C. flavescens). Besan^ON, F. 1984. Contribution a I’etude de la Biologie et de la Strategie de Reproduction de Crocidura russula (Insectivora, Soricidae) en Zone Temperee. Unpublished Ph.D. dissert.. University of Lausanne, 184 pp. (not seen, cited in Hutterer et al., 1987, see Literature Cited). Bishop, I. R., and M. J. Delany. 1963 (see Literature Cited). Fayard, a., and G. Erome. 1977. Les micromammiferes de la bordure orientale du Massif Central. Mammalia, 41:301-319. Fons, R. 1972. La Musaraigne musette Crocidura russula (Herman, 1780). Science et Nature, Paris, 1 12:23-28. Genoud, M., and P. Vogel. 1981. The activity of Crocidura russula (Insectivora, Soricidae) in the field and in captivity. Zeitschrift fiir Saugetierkunde, 46:222-232. 1990 (see Literature Cited). GrOnwald, A., and F. P. Mohres. 1974 (see under C. olivieri). Hutterer, R. 1986. The species of Crocidura (Soricidae) in Morocco. Mammalia, 50:521-534. 1 EANMARIE-Bes ANCON , F. 1986. Estimation de I’age et de la longevite chez Crocidura russula (Insectivora: Soricidae). Acta Oecologia, Oecologia Applicata, 7:355-366. Kahmann, H., and E. Kahmann. 1954. La musareigne de Corse. Mammalia, 18:129-158. Lopez-Fuster, M. j., j. Gosalbez, and V. Sans-Coma. 1985. Uber die Fortpflanzung der Hausspitzmaus {Crocidura russula Herman, 1780) im Ebro-Delta (Katalonien, Spanien). Zeitschrift fiir Saugetierkunde: 50: 1-6. Niethammer, G. 1950. Zur Jungenpflege und Orientierung der Hausspitzmaus {Crocidura russula Herm.). Bonner Zoologische Beitrage, 1:117-125. Niethammer, J. 1970. Uber Kleinsauger aus Portugal. Bonner Zoologische Beitrage, 21:89-1 18. PoiTEviN, F., J. Catalan, R. Fons, and H. Croset. 1987. Biologie evolutive des populations ouest-europeennes de Crocidures (Mammalia, Insectivora). II. Ecologie comparee de Crocidura russula Hermann, 1780 et de Crocidura suaveolens Pallas, 1811 dans le midi de la France et en Corse: Role probable de la competition dans le partage des milieux. Revue Ecologie (Terre Vie), 42:39-58. Sans-Coma, V., and S. Mas-Coma. 1978. Uber die Kleinsaugetiere, ihre Helminthen und die Schleiereule auf der Insel Meda Grossa (Katalonien: Spanien). Saugetierkundliche Mitteilungen, 26:139-150. Sans-Coma, V., I. Gomez, and J. Gosalbez. 1976. Eine Untersuchung an der Hausspitzmaus {Crocidura russula Hermann, 1780) auf der Insel Meda Grossa (Katalonien, Spanien). Saugetierkundliche Mitteilungen, 24:279-288. Vogel, P. 1972a (see Literature Cited). \912b (see Literature Cited). Yalden, D. W., P. A. Morris, and J. Harper. 1973 (see under C. leucodon). C. suaveolens: Hanzak, j. 1966. Vyvoj mladat belozubky sede, Crocidura suaveolens (Pallas) (1821). Lynx, 6:67-74 (not seen, cited in Hellwing, 1973/j, see Literature Cited). 122 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 Hellwing, S. 1970 (see Literature Cited). 1971 (see Literature Cited). 1973a (see Literature Cited). 1973Z? (see Literature Cited). Mover, H., and S. Hellwing. Unpublished data. Mover, H., S. Hellwing, and A. Ar. 1988 (see Literature Cited). Ondrias, J. C. 1979. Contribution to the knowledge of Crocidura suaveolens (Mammalia, Insectivora) from Greece, with a description of a new subspecies. Zeitschrift fur Saugetierkunde, 35:371-381. Rood, J. P. 1965. Observations on population structure, reproduction and molt of the Scilly shrew. Journal of Mammalogy, 46:426-433. Vasarhelyi, S. 1929. Beitrage zur Kenntnis der Lebenweise zweier Kleinsauger. Allattani Kozlemenyek, 26:90-91 (not seen, cited in Dryden, 1968, see Literature Cited). Vlasak. P. 1972. The biology of reproduction and post-natal development of Crocidura suaveolens Pallas, 1811, under laboratory conditions. Acta Universitatis Carolinae-Biologica, 1970:207-292. 1973. Vergleich der postnatalen Entwicklung der Arten Sorex araneus L. und Crocidura suaveolens (Pall.) mit Bemerkungen zur Methodik der Laborzucht (Insectivora: Soricidae). Vestnfk Ceskoslovenske Spolecnosti Zoologicke, 37:222-233. Vogel, P., T. Maddalena, and F. Catzeflis. 1986. A contribution to the taxonomy and ecology of shrews {Crocidura zitnmermanni and C. suaveolens) from Crete and Turkey. Acta Theriologica, 31:537-545. C. tarfayensis: Hutterer, R. 1986 (see under C. russula). C. there sae: Maddalena, T., and C. Maddalena. Unpublished data. Verschuren, j., and j. Meester. 1977 (see under C. olivieii). C. viaria: Genoud, M., and P. Vogel. 1990 (see Literature Cited). Happold, D. C. D. 1987 (see under C. bottegi). Hutterer, R. 1986 (see under C russula). Maddalena, T., and C. Maddalena. Unpublished data. C. whitakeri: Hutterer, R. 1986 (see under C. russula). C. zinmiermani: Vogel, P., T. Maddalena, and F. Catzeflis. 1986 (see under C. suaveolens). Crypt Otis magna: Robertson, P. B., and E. A. Richart. 1975. Cryptotis magna. Mammalian Species, 61:1-2. C. parva: Alvarez, T. 1963. The Recent mammals of Tamaulipas, Mexico. University of Kansas Publications, Museum of Natural History, 14:363-473. Blair, W. F. 1938 (see Literature Cited). Brimley, C. S. 1923 (see under B. brevicauda). Choate, J. R. 1970. Systematics and zoogeography of middle American shrews of the genus Cryptotis. University of Kansas Publications, Museum of Natural History, 19:195-317. ! Conaway, C. H. 1958. Maintenance, reproduction, and growth of the least shrew in captivity. Journal of Mammalogy, [ 39:507-512. ‘ Davis, W. B., and L. Joeris. 1945. Notes on the life history of i the little short-tailed shrew. Journal of Mammalogy, 26:136-138. Gould, E. 1969 (see under B. brevicauda). Hamilton, W. J., Jr. 1934. Habits of Cryptotis parva in New York. Journal of Mammalogy, 15:154-155 1944. The biology of the little short-tailed shrew, Cryptotis parva. Journal of Mammalogy, 25:1-7. Hoffmeister, D. F. 1989 (see under B. brevicauda). Kirkland, G. L., Jr. Unpublished data. Komarek, E. V., AND R. Komarek. 1938 (see under B. brevicauda). Layne, j. N. 1974. Ecology of small mammals in a flatwoods habitat in north-central Florida, with emphasis on the cotton rat (Sigmodon hispidus). American Museum Novitates, 2544:1-48. Layne, J. N., and J. R. Redmond. 1959. Body temperature of the least shrew. Saugetierkundliche Mitteilungen, 7: 169-172. Lindsay, D. M. 1960 (see under B. brevicauda). Linzey, D. W., and a. V. Linzey. 1968 (see under B. brevicauda). Lowery, G. H., Jr. 1974 (see under B. brevicauda). Mock, O. B. 1982. The least shrew {Cryptotis parva) as a laboratory animal. Laboratory Animal Science, 32: 177-179. Mock, O. B., and C. H. Conaway. 1976 (see Literature Cited). Mumford, R. E. 1969 (see under B. brevicauda). Mumford, R. E., and j. O. Whitaker, Jr. 1982 (see under B. brevicauda). O’Farrel, M. j., D. W. Kaufman, J. B. Gentry, and M. H. Smith. 1977 (see under B. brevicauda). [ POURNELLE, G. H. 1950. Mammals of a north Florida swamp. I Journal of Mammalogy, 31:310-319. | Walker, E. P. 1954. Shrews is shrews. Nature Magazine, 47:125-128. I Diploniesodon pulchellum: Heptner, V. G. 1939. The Turkestan desert shrew, its biology and adaptive peculiarities. Journal of Mammalogy, 20: 139-149. | Myosorex baboulti: Dieterlen, F., and H. Heim de Balsac. 1979 (see under C. flavescens). M. norae and M. pollulus: Duncan, P., and R. W. Wrangham. 1971. On the ecology and distribution of subterranean insectivores in Kenya. Journal of Zoology (London), 164:149-163. M. varius: Baxter, R. M., and C. N. V. Lloyd. 1980 (see Literature Cited). Goulden, E. a., and j. Meester. 1978 (see under C. mariquensis). Rowe-Rowe, D. T., and j. Meester. 1982 (See under C. flavescens). 1985. Biology of Myosorex varius in an African montane region. Annales Zoologici Fennici, 173:271-273. 1994 INNES— Soricid Life Histories 123 Neomys anomalus: BOROWSKI, S., AND A. Dehnel. 1952. Materialy do biologii Soricidae. Annales Universitatis Marie Curie-Sklodowska, Sectio C, 7: 305-448 (not seen, Pucek, Z., in litt.). Gebczynski, M. 1977. Body temperatures in five species of shrews. Acta Theriologica, 22:521-530. Michalak, I. 1982. Reproduction and behaviour of the Mediterranean water shrew under laboratory conditions. Saugetierkundliche Mitteilungen, 30:307-3 10. NiETHAMMER, J. 1977. Ein syntopes Vorkommen der Wasserspitzraause Neomys fodiens und N. anomalus. Zeitschrift fur Saugetierkunde, 42:1-6. 1978. Weitere Beobachtungen iiber syntope Wasserspitzmause der Arten Neomys fodiens und N. anomalus. Zeitschrift fiir Saugetierkunde, 43:313-321. N. fodiens-. Barrett-Hamilton, G. E. H. 1910. A History of British Mammals. Land Mammals. Gurney and Jackson, London, 2:1-748. Bauchot, R., and H. Stephan. 1966 (see under C. flavescens). Borowski, S., and a. Dehnel. 1952 (see under N. anomalus). Cantuel, P. 1940. Poids moyen de quelques micromammiferesde la faune fran^aise. Mammalia, 4:113-117. 1946. Periode de reproduction et nombre de foetus de quelque micromammiferes de France. Mammalia, 10: 140-144. Churchfield, S. 1984. An investigation of the population ecology of syntopic shrews inhabiting water-cress beds. Journal of Zoology (London), 204:229-240. Crowcroft, P. 1957 (see Literature Cited). Gauckler, a. 1962. Beitrag zur Fortflanzungsbiologie der Wasserspitzmaus, (Neomys fodiens). Bonner Zoologische Bietrage, 13:321-323. Gebczynski, M. 1977 (see under N. anomalus). Herter, K. 1957 (see under C. leucodon; cited in Vogel, 1912b, see Literature Cited). IVANTER T. V., E. V. IVANTER, AND E. J. TERNOUSKO. 1974. Biology of reproduction of the shrews (Soricidae) of Karelia. Pp. 95-143, in Problems of Animal Ecology (E. V. Ivanter, ed.), Karelian Branch of the Academy of Sciences of the USSR Institute of Biology, Petrozavodsk, 192 pp. (not seen, cited in Michalak, 1983). Kohler, D. 1984. Zum Pflegeverhalten und Verhaltensontogenese von Neomys fodiens (Insectivora: Soricidae). Zoologischer Anzeiger, 213:275-290. Michalak, I. 1983. Reproduction, maternal and social behaviour of the European water shrew under laboratory conditions. Acta Theriologica, 28:3-24. 1987a (see Literature Cited). \9%lb (see Literature Cited). Myrcha, a. 1969. Seasonal changes in caloric value, body water and fat in some shrews. Acta Theriologica, 14:211-227. NiETHAMMER, G. 1950 (see under C. russula). Popov, V. A. 1960. Mlekopitajuscie Volzsko-kamskogo Kraja. Nasekomojadnye, rukokrylye, gryzuny. Akademiya Nauk, SSSR, Kazanskij Filial, Kazan. 467 pp. (not seen, cited in Michalak, 1983, see under N. fodiens). Price, M. 1953. The reproduction cycle of the water shrew, Neomys fodiens bicolor Shaw. Proceedings of the Zoological Society of London, 123:599-621. Roben, P. 1969. Die Spitzmause (Soricidae) der Heidelberger Umgebung. Saugetierkundliche Mitteilungen, 17:42-62. Sheftel, B. I. 1989. Long-term and seasonal dynamics of shrews in central Siberia. Annales Zoologici Fennici. 26:357-369. Vogel, P. 1972a (see Literature Cited). WOLK, E. Unpublished data. Notiosorex crawfordi: Armstrong, D. M., and J. K. Jones, Jr. 1971. Mammals of the Mexican state of Sinaloa. 1 . Marsupialia, Insectivora, Edentata, Lagomorpha. Journal of Mammalogy, 52:747-757. Dixon J. 1924. Notes on the life history of the gray .shrew. Journal of Mammalogy, 5:1-6. Hoffmeister, D. F., and W. W. Goodpaster. 1962 (see Literature Cited). Lindstedt, S. L. 1980. Energetics and water economy of the smallest desert shrew. Physiological Zoology, 53:82-97. Preston, J. R., and R. E. Martin. 1963. A gray shrew population in Harmon County, Oklahoma. Journal of Mammalogy, 44:268-270. N. gigas: Armstrong, D. M., and J. K. Jones, Jr. 1971 (see under N. crawfordi). Jones, J. K., Jr. 1966. Recent records of the shrew, Megasorex gigas (Merriam), from western Mexico. The American Midland Naturalist, 75:249-250. Scutisorex somereni : Dieterlen, F., and H. Heim de Balsac. 1979 (see under C. flavescens). Sorex alpinus: Kirkland, G. L., Jr. Unpublished data. S. araneus: Adams, L. E. 1910. A hypothesis as to the cause of the autumnal epidemic of the common and lesser shrew, with some notes on their habits. Memoirs and Proceedings of the Manchester Literary and Philosophical Society, 54:1-13. Barrett-Hamilton, G. E. H. 1910 (see under A. fodiens). Bauchot, R., and H. Stephan. 1966 (see under C. flavescens). Borowski, S., and A. Dehnel. 1952 (see under A. anomalus). Brambell, F. W. R. 1935. Reproduction in the common shrew (Sorex araneus Linnaeus). Philosophical Transactions of the Royal Society of London, Series B, 225:1-62. Branis, M. 1985. Postnatal development of the eye of Sorex araneus. Acta Zoologica Fennici, 173:247-248. Cantuel, P. 1940 (see under A. fodiens). Churchfield, S. 1980. Population dynamics and the seasonal fluctuations in numbers of the common shrew in Britain. Acta Theriologica, 25:415-424. 1981. Water and fat contents of British shrews and their role in the seasonal changes in body weight. Journal of Zoology (London), 194:165-173. 1984 (see under A. fodiens). Croin Michielsen, N. 1966 (see Literature Cited). Crowcroft, P. 1957a (see Literature Cited). 1957/7. On the life span of the common shrew (Sorex araneus L.). Proceedings of the Zoological Society of London, 127:285-292. Dehnel, A. 1949. Badania nad rodzajem Sorex L. Annales Universitatis Marie Curie-Sklodowska, Sectio C, 4: 17-102 (not 124 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 seen, cited in Kowalska-Dyrcz, 1961). _. 1952. Biologia rozmnazania S. araneus w warunkach laboratory) nych. Annales Universitatis Marie Curie- Sklodowska, Sectio C, 6:359-376 (not seen, Pucek, Z., in litt.). GgBCYZYNSKI, M. 1965. Seasonal and age changes in the metabolism and activity of Sorex araneus Linnaeus (1758). Acta Theriologica, 10:303-331. 1977 (see under N. anomalus). Godfrey, G. K. 1979. Gestation period in the common shrew, Sorex coronatus {araneus) fretalis. Journal of Zoology (London), 189:548-551. Gorecki, a. 1965. Energy values of body fat in small mammals. Acta Theriologica, 10:333-352. Handwerk, J. 1987. Neue Daten zur Morphologic, Vertbreitung und Okologie der Spitzmause Sorex araneus und S. coronatus im Rheinland. Bonner Zoologische Beitrage, 38:273-297. Hanski, I. 1989a. Habitat selection in a patchy environment: Individual differences in common shrews. Animal Behavior, 38:414-422. Heikura, K. 1984. The population dynamics and the influence of winter on the common shrew {Sorex araneus L.). Pp. 343-361, in Winter Ecology of Small Mammals (J. F. Merritt, ed.), Carnegie Museum of Natural History Special Publication 10, ix + 380 pp. Hissa, R., and H. Tarkkonen. 1969. Seasonal variations in brown adipose tissue in two species of voles and the common shrew. Annales Zoologici Fennici, 6:444-447. Hyvarinen, H. 1969. On the seasonal changes in the skeleton of the common shrew {Sorex araneus L.) and their physiological background. Aquilo, Serie Zoologica, 7:1-32. .. 1984. Wintering strategy of voles and shrews in Finland. Pp. 139-148, in Winter Ecology of Small Mammals (J. F. Merritt, ed.), Carnegie Museum of Natural History Special Publication 10, ix + 380 pp. Hyvarinen, H., and K. Heikura. 1971. Effects of age and seasonal rhythm on the growth patterns of some small mammals in Finland and in Kirkenes, Norway. Journal of Zoology (London), 165:545-556. Kaikusalo, V. A. Unpublished data. Kirkland, G. L., Jr. Unpublished data. Kowalska-Dyrcz, A. 1961. Seasonal variations in Sorex araneus Linnaeus 1758 in Poland. Acta Theriologica, 4:268-273. Malzahn, E., and E. Wolk. 1977. Oxidation-reduction activity of the blood of the common shrew and the bank vole. Comparative Physiology and Ecology, 2:168-175. Meylan, a. 1967 (see under C. leucodon). Michalak, I. 1987a (see Literature Cited). Middleton, A. D. 1931. A contribution to the biology of the common shrew, Sorex araneus Linnaeus. Proceedings of the Zoological Society of London, 1931:133-143. Myrcha, a. 1969 (see under N. fodiens). Niethammer, G. 1950 (see under C. russula). Pankakoski, E. 1979. The cone trap — A useful tool for index trapping of small mammals. Annales Zoologici Fennici, 16:144-150. 1989. Variation in the tooth wear of the shrews Sorex araneus and S. mi nut us. Annales Zoologici Fennici, 26:445-457. Pankakoski, E., and K. M. Tahka. 1982. Relation of adrenal weight to sex, maturity and season in five species of small mammals. Annales Zoologici Fennici, 19:225-232. Passarge, H. 1984. Sorex isodon marchicus ssp. nova in Mitteleuropa. Zeitschrift fiir Saugetierkunde, 49:278-28. Pernetta, j. C. 1977. Population ecology of British shrews in grassland. Acta Theriologica, 22:279-296. Pucek, Z. 1955. Untersuchungen iider die Veranderlichkeit des Schadels in Lebenszyklus von Sorex araneus araneus L. Annales Universitatis Marie Curie-Ski odowska, Sectio C, 9:163-212 (not seen, cited in Kowalska-Dyrcz, 1961; see under S. araneus). 1960. Sexual maturation and variability of the reproductive system in young shrews {Sorex L.) in the first calendar year of life. Acta Theriologica, 3:269-296. Roben, P. 1969 (see under N. fodiens). Saint -Girons, M.-C., and V. Mazak. 1971. Donnees morphologiques sur quelque micromammiferes en Laponie. Zeitschrift fiir Saugetierkunde, 36:179-190. Sans-Coma, V. 1979. Beitrag zur Kenntnis der Waldspitzmaus, Sorex araneus Linne, 1758, in Katalonien, Spanien. Saugetierkundliche Mitteilungen, 27:96-106. SCHUBARTH, H. 1958. Zur Variabilitat von Sorex araneus L. Acta Theriologica, 2: 175-202. Searle, j. B. 1990. Evidence for multiple paternity in the common shrew {Sorex araneus). Journal of Mammalogy, 71:139-144. Serafinski, W. 1955. Badania morfologiczne i ecologiczne nad polskimi gatunkami rodzaju Sorex L. (Insectivora; Soricidae). Acta Theriologica, 1:27-86. Sheftel, B. 1. 1989 (see under N. fodiens). Shillito, j. F. 1963. Field observations on the growth, reproduction and activity of a woodland population of the common shrew Sorex araneus L. Proceedings of the Zoological Society of London, 140:99-114. SirvoNEN, L. 1954. liber die Grossenvariationen der Saugetiere und die Sorex macropygmaeus Mill. — Frage in Fennoskandien. Annales Academiae Scientiarum Fennici, A, V, 21:1-24 (not seen, cited in Skaren, 1964). Skaren, U. 1964. Variation in two shrews, Sorex unguiculatus Dobson and S. a. araneus L. Annales Zoologici Fennici, 1:94-124. 1973. Spring moult and onset of the breeding season of the common shrew {Sorex araneus L.) in central Finland. Acta Theriologica, 18:443-458. 1979 (see Literature Cited). Tarkowski, a. K. 1957. Studies on reproduction and prenatal mortality of the common shrew {Sorex araneus L.). II. Reproduction under natural conditions. Annales Universitatis Marie Curie-Sklodowska, Sectio C, 10: 177-244 (not seen, Pucek, Z., in lift.). Vogel, P. 1972a (see Literature Cited). 1912b (see Literature Cited). Vlasak, P. 1973 (see under C. suaveolens). WoLK, E. 1969. Body weight and daily food intake in captive shrews. Acta Theriologica, 14:35-47. Unpublished data. WoLSKA, J. 1952. Rozwdj aparatu plciowego w cyklu zyciowym Sorex araneus L. (I). Annales Universitatis Marie Curie- Sklodowska, Sectio C, 7:497-539. Yudin, B. S. 1962. Ecology of shrews (genus Sorex) of western Siberia, U.S.S.R. Pp. 33-134, in Voprosy Ekologii, 1994 INNES — Soricid Life Histories 125 Zoogeograffi, i Sistematiki Zhivotnykh (A. I. Cherepanov, ed.), Trudy Biologichoskogo Instituta Akademiya Nauk SSSR, Sibirskoye Otdelaniyo, Novosibirsk 8:1-192. ZiPPELlUS, H.-M. 1958. Zur Jugendentwicklung der Waidspitzmaus, Sorex araneus. Eine verhaltenskundliche Studie. Bonner Zoologische Beitrage, 9:120-129. S. arcticus: Bailey, B. 1929 (see under B. brevicauda). Baird, D. D., R. M. Timm, and G. E. Nordquist. 1983. Reproduction in the arctic shrew, Sorex arcticus. Journal of Mammalogy, 64:298-301. Buckner, C. H. 1959. Studies of the feeding habits and populations of shrews in southeastern Manitoba. Unpublished Ph.D. dissert.. University of Western Ontario, London, 207 pp. 1966 (see under B. brevicauda). Clough, G. C. 1963. Biology of the arctic shrew. The American Midland Naturalist, 69:69-81. Doyle, D. M. 1979. Population dynamics and ecology of subarctic shrews. Unpublished Master’s thesis. University of Alberta, Edmonton, 118 pp. Iverson, S. L., and B. N. Turner. 1976 (see under B. brevicauda). Jackson, H. H. T. 1961. Mammals of Wisconsin. University of Wisconsin Press, Madison, 504 pp. Jones J. K., Jr., D. M. Armstrong, R. S. Hoffmann, and C. Jones. 1983. Mammals of the Northern Great Plains. University of Nebraska Press, Lincoln, 379 pp. Smith. R. W. 1940. The land mammals of Nova Scotia. The American Midland Naturalist, 24:213-241. S. bendirii: Pattie, D. L. 1973. Sorex bendirii. Mammalian Species, 27:1-2. S. caecutiens: Barowoski, S., and a. Dehnel. 1952 (see under N. fodiens). Dokuchaev, N. E. 1989. Population ecology of Sorex shrews in North-East Siberia. Annales Zoologici Fennici, 26:371-379. Gebczynski, M. 1977 (see under N. anomalus). Hanski, I. 1989ft. Population biology of Eurasian shrews: Towards a synthesis. Annales Zoologici Fennici, 26:469-479. Kaikusalo, V. A. Unpublished data. PUCEK, Z. (ED.). 1984. Klucz do oznaczania ssakdw Polski. Panstwowe Wydawnictus, Naukowe, Warszawa, 388 pp. (not seen, Pucek, Z., in litt.). Sheftel, B. I. 1989 (see under N. fodiens). Walton, D. W., and H. K. Hong. 1975. Sorex caecutiens annexus Thomas from Korea. Mammalia, 39:325-327. Yudin, B. S. 1962 (see under S. araneus). Yudin, B. S., L. I. Galkina, and A. F. Potapkina. 1979. Mammals of the mountainous region of Altai-Sajan. Izdatelstvo Nauka Sibirskoe Otdelaniyo Novosibirsk (not seen, cited in Hanski, 1989ft, see under S. caecutiens). S. cinereus: Bailey, B. 1929 (see under B. brevicauda). Bole, B. P., Jr. 1939. The quadrat method of studying small mammal populations. Scientific Publications of the Cleveland Museum of Natural History, 5:15-77 (not seen, cited in Mohr, C. O., 1943. A comparison of North American small-mammals censuses. The American Midland Naturalist, 29:545-587). Buckner, C. H. 1959 (see under S. arcticus). 1966 (see Literature Cited). Campbell, C. A., and A. I. Dag. 1972 (see under B. brevicauda). Connor, P. F. 1960 (see under B. brevicauda). 1966 (see Literature Cited). Doyle, D. M. 1971 (see under S. arcticus). Folinsbee, J. D. 1971. Ecology of the masked shrew, Sorex cinereus, on the island of Newfoundland. Unpublished Master’s thesis. University of Manitoba, Winnipeg, 132 pp. Forsyth, D. J. 1976 (see Literature Cited). French, T. W. 1980a. Ecological relationships between the southeastern shrew (Sorex longirostris Bachman) and the masked shrew (S. cinereus Kerr) in Vigo County, Indiana. Unpublished Ph.D. dissert., Indiana State University, Terre Haute, 54 pp. Unpublished data. Gifford, C. L., and R. Whitebread. 1951 (see under B. brevicauda). Grimm, W. C., and H. A. Roberts. 1950 (see under B. brevicauda). Grimm, W. C., and R. Whitebread. 1952 (see under B. brevicauda). Harper, F. 1929 (see under B. brevicauda). iNNES, D. G. L., AND J. F. Bendell. 1990. High densities of the masked shrew, Sorex cinereus, in jack pine plantations in northern Ontario. The American Midland Naturalist, 124:330-341. Unpublished data. Iverson, S. L., and B. N. Turner. 1976 (see under B. brevicauda). Jackson, H. H. T. 1928. A taxonomic review of the American long-tailed shrews. North American Fauna, 51:1-238. Jones, J. K., Jr., D. M. Armstrong, R. S. Hoffmann, and C. Jones. 1983 (see under S. arcticus). Kirkland, G. L., Jr. Unpublished data. Komarek, E. V., AND R. Komarek. 1938 (see under B. brevicauda). Lindsay, D. M. 1960 (see under B. brevicauda). Linzey, D. W., and a. V. Linzey. 1968 (see under B. brevicauda). Long, C. A. 1965. The mammals of Wyoming. University of Kansas Publications, Museum of Natural History, 14:493-758. Martell, a. M., and a. M. Pearson. 1978. The small mammals of the Mackenzie Delta Region, Northwest Territories, Canada. Arctic, 31:475-488. Moore, J. C. 1949. Notes on the shrew, Sorex cinereus, in the southern Appalachians. Ecology, 30:234-237. Mumford, R. E. 1969 (see under B. brevicauda). Mumford, R. E., and j. O. Whitaker, Jr. 1982 (see under B. brevicauda). Murie, O. j. 1959. Fauna of the Aleutian Islands and Alaska Peninsula. North American Fauna, 61:1-364. Okhotina, M. V. 1977. Palaearctic shrews of the subgenus Otisorex: Biotopic preference, population number, taxonomic revision and distribution history. Acta Theriologica, 22:191-206. Opper, C. L., G. C. Lorenz, and G. W. Barrett. 1988. The masked shrew, Sorex cinereus, in southwestern Ohio. Ohio Journal of Science, 88: 170-171. Palmer, R. S. 1947. Notes on some Maine shrews. Journal of 126 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 Mammalogy, 28:13-16. Preble, E. A. 1908. A biological investigation of the Athabasca- Mackenzie region. North American Fauna, 27:1-574. Richmond, N. D., and W. C. Grimm. 1950. Ecology and distribution of the shrew Sorex dispar in Pennsylvania. Ecology, 31:279-282. Richmond, N. D., and H. R. Roslund. 1949 (see under B. brevicauda). Saunders, W. E. 1932 (see under B. brevicauda). Teferi, T. 1988. Reproduction, age structure and movement of Sorex cinereus Kerr on Bon Portage Island. Unpublished Mas- ter’s thesis, Acadia University, Wolfville, Nova Scotia, 98 pp. Timm, R. M. 1975 (see under B. brevicauda). Townsend, M. T. 1935 (see under B. brevicauda). Turner, R. W. 1974. Mammals of the Black Hills of South Dakota and Wyoming. Miscellaneous Publications, Museum of Natural History, University of Kansas, 60:1-178. Van Zyll DE Jong, C. G. 1980. Systematic relationships of woodland and prairie forms of the common shrew, Sorex cinereus cinereus Kerr and S. c. haydeni Baird, in the Canadian prairie provinces. Journal of Mammalogy, 61:66-75. Youngman, P. M. 1975. Mammals of the Yukon Territory. Publications in Zoology, National Museums of Canada, 10:1-192. S. coronatus: Genoud, M. 1984. Activity of Sorex coronatus (Insectivora, Soricidae) in the field. Zeitschrift fur Saugetierkunde, 49:74-78. 1988 (see Literature Cited). Genoud, M., and P. Vogel. 1990 (see Literature Cited). Godfrey, G. K. 1979 (see under S. araneus). Lopez-Fuster, M. J., E. Castien, and J. Gosalbez. 1988. Reproductive cycle and population structure of Sorex coronatus Millet, 1828 (Insectivora: Soricidae) in the northern Iberian Peninsula. Bonner Zoologische Beitrage, 39:163-170. Nores, C. 1989. Some basic biological features of Sorex coronatus. Pp. 678-679, in Abstracts, Fifth International Theriological Congress, Rome. Unpublished data. S. daphaenodon: Sheftel, B. I. 1989 (see under N. fodiens). Yudin, B. S. 1962 (see under S. araneus). S. dispar. Conaway, C. H., and D. W. Pettzer. 1952. Sorex pal usttis and S. dispar from the Great Smoky Mountains National Park. Journal of Mammalogy, 33:106-108. Connor, P. F. 1960 (see under B. brevicauda). Grimm, W. C., and H. A. Roberts. 1950 (see under B. brevicauda). Harper, F. 1929 (see under B. brevicauda). Kirkland, G. L., Jr. 1981. Sorex dispar and Sorex gaspensis. Mammalian Species, 155:1-4. Unpublished data. Kirkland, G. L., Jr., and H. M. van Deusen. 1979. The shrews of the Sorex dispar group: Sorex dispar Batchelder and Sorex gaspensis Anthony and Goodwin. American Museum Novitates, 2675:1-21. Linzey, D. W., and a. V. Linzey. 1968 (see under B. brevicauda). Richmond, N. D., and W. C. Grimm. 1950 (see under S. cinereus). S. Jumeus: Campbell, C. A., and A. I. Dag. 1972 (see under B. brevicauda). Connor, P. F. 1960 (see under B. brevicauda). 1966 (see Literature Cited). French, T. W. Unpublished data. Gifford, C. L., and R. Whitebread. 1951 (see under B. brevicauda). Hamilton, W. J., Jr. 1940 (see Literature Cited). 1943 (see under B. brevicauda). INNES, D. G. L., AND J. F. Bendell. Unpublished data. Kirkland, G. L., Jr. Unpublished data. Komarek, E. V., AND R. Komarek. 1938 (see under B. brevicauda). Linzey, D. W., and A. V. Linzey. 1968 (see under B. brevicauda). Owen, J. G. 1984. Sorex fumeus. Mammalian Species, 215:1-8. Richmond, N. D., and W. C. Grimm. 1950 (see under S. cinereus). Richmond, N. D., and H. R. Roslund. 1949 (see under B. brevicauda). Whitaker, J. O., Jr., G. S. Jones, and D. D. Pascal, Jr. 1975. Notes on the mammals of the Fires Creek Area, Nantahala Mountains, North Carolina, including their ectoparasites. Journal of the Elisha Mitchell Scientific Society, 91:13-17. S. gaspensis: Dalton, M., and B. A. Sabo. 1980. A preliminary report on the natural history of the Gaspe shrew. The Atlantic Center for the Environment, Ipwich, Massachusetts, 29 pp. (not seen, cited in van Zyll de Jong, C. G., 1983, Handbook of Canadian Mammals. Vol. 1. Marsupials and Insectivores. National Museums of Canada, Ottawa, 210 pp. French, T. W. Unpublished data. Kirkland, G. L., Jr., and H. M. van Deusen. 1979 (see under S. dispar). Saunders, W. E. 1932 (see under B. brevicauda). S. granarius: Gisbert, j., M. j. Lopez-Fuster, R. Garci'a-Perea, and J. Ventura. 1988. Distribution and biometry of Sorex granarius (Miller, 1910) (Soricinae: Insectivora). Zeitschrift fur Saugetierkunde, 53:267-275. S. hoyi: Bailey, B. 1929 (see under B. brevicauda). Diersing, V. E. 1980. Systematics and evolution of the pygmy shrews (subgenus Microsorex) of North America. Journal of Mammalogy, 61:76-101. French, T. W. Unpublished data. Heaney, L. R., and E. C. Birney. 1975. Comments on the distribution and natural history of some mammals in Minnesota. The Canadian Field-Naturalist, 89:29-34. Innes, D. G. L., and j. F. Bendell. Unpublished data. Kirkland, G. L., Jr. Unpublished data. Kirkland, G. L., Jr.. A. M. Wilkinson, J. V. Planz, and J. E. Maldonado. 1987. Sorex {Microsorex) hoyi in 1994 INNES — Soricid Life Histories 127 Pennsylvania. Journal of Mammalogy, 68:384-387. Long, C. A. 1972a. Taxonomic revision of the mammalian genus Microsorex Coues. Transactions of the Kansas Academy of Science, 74:181-196. 1972Z). Notes on habitat preference and reproduction in pygmy shrews, Microsorex. The Canadian Field-Naturalist, 86:155-160. 1974. Microsorex hoyi and Microsorex thompsoni. Mammalian Species, 33:1-4. 1976. Notes on reproduction in pygmy shrews and observed ratios of mammae to body weights. Report, Museum of Natural History, Wisconsin State University, 11:5-6. Miller, D. H. 1964. Pygmy shrew in Vermont. Journal of Mammalogy, 45:651-652. S. isodon: Kaikusalo, V. A. Unpublished data. Passarge, H. 1984 (see under S. araneus). Sheftel, B. I. 1989 (see under N. fodiens). Skaren, U. 1979 (see Literature Cited). 1982. Intraspecific aggression and postnatal development in the shrew Sorex isodon Turov. Annales Zoologici Fennici, 19:87-91. S. longirostris: French, T. W. 1980a (see under S. cinereus). 1980Zj. Natural history of the southeastern shrew, Sorex longirostris Bachman. The American Midland Naturalist, 104:13-31. Hoffmeister, D. F. 1989 (see under B. brevicauda). Lowery, G. H., Jr. 1974 (see under B. brevicauda). Mumford, R. E. 1969 (see under B. brevicauda). O’Farrel, M. J., D. W. Kaufman, J. B. Gentry, and M. H. Smith. 1977 (see under B. brevicauda). S. macrodon: He ANY, L. R. and E. C. Birney. 1977. Distribution and natural history notes of some mammals from Puebla, Mexico. The Southwestern Naturalist, 21:543-545. S. merriami: Hudson, G. E., and M. Bacon. 1956. New records of Sorex merriami for eastern Washington. Journal of Mammalogy, 37:436-438. Johnson, M. L., and C. W. Clanton. 1954. Natural history of Sorex merriami in Washington state. The Murrelet, 35:1-4. S. milleri: Baker, R. H. 1956. Mammals of Coahuila, Mexico. University of Kansas Publications, Museum of Natural History, 9:125-335. S. minutissimus: Kaikusalo, V. A. 1967. Beobachtungen an gekafigten Knirpsspitzmaussen Sorex minutissimus Zimmermann, 1780. Zeitschrift fur Saugetierkunde, 32:301-306. Unpublished data. S. minutus: Adams, L. E. 1910 (see under S. araneus). Barrett-Hamilton, G. E. H. 1910 (see under N. fodiens). Bauchot, R., and H. Stephan. 1966 (see under C. flavescens). Borowski, S., and a. Dehnel. 1952 (see under N. anomalus). Brambell, F. W. R. 1935 (see under S. araneus). Brambell, F. W. R., and K. Hall. 1936. Reproduction of the lesser shrew (Sorex minutus Linnaeu.s). Proceedings of the Zoological Society of London, 1936:957-969. Butterheld, j., j. C. Coulson, and S. Wanless. 1981. Studies on the distribution, food, breeding biology and relative abundance of the pygmy and common shrews (Sorex minutus and S. araneus) in upland areas of northern England. Journal of Zoology (London), 195:169-180. Churchfield, S. 1981 (see under S. araneus). CORKE, D., R. A. D. COWLIN, and W. W. Page. 1969. Notes on the distribution and abundance of small mammals in south-west Ireland. Journal of Zoology (London), 158:216-221. Croin Michielsen, N. 1966 (see under S. araneus). Crowcroft, P. 1957 (see Literature Cited). Gebezynski, M. 1971. The rate of metabolism of the lesser shrew. Acta Theriologica, 16:329-339. 1977 (see under N. anomalus). Genoud, M., and P. Vogel. 1990 (see Literature Cited). Grainger. J. P., and J. S. Fairley. 1978. Studies on the biology of the pygmy shrew Sorex minutus in the west of Ireland. Journal of Zoology (London), 186:109-141. Hutterer, R. 1976 (see Literature Cited). Kaikusalo, V. A. Unpublished data. Kirkland, G. L., Jr. Unpublished data. Middleton, A. D. 1931 (see under S. araneus). Myrcha, a. 1969 (see under N. fodiens). Pankakoski, E. Unpublished data. Pankakoski, E., and K. M. Tahka. 1982 (see under S. araneus). Pernetta, j. C. 1977 (see under S. araneus). PUCEK, Z. 1960 (see under S. araneus). Roben, P. 1969 (see under N. fodiens). Serafinski, W. 1955 (see under S. araneus). Sheftel, B. I. 1989 (see under N. fodiens). Yudin, B. S. 1962 (see under S. araneus). S. mirabilis: Okhotina, M. V. 1969. Some data on ecology of Sorex (Ognevia) mirabilis Ognev, 1937. Acta Theriologica, 14:273-284. S. monticolus: Doyle, A. T. 1990. Use of riparian and upland habitats by small mammals. Journal of Mammalogy, 71:14-23. Hawes, M. L. 1977. Home range, territoriality, and ecological separation in sympatric shrews, Sorex vagrans and Sorex obscurus. Journal of Mammalogy, 58:354-367. Hoffmeister, D. F. 1986. Mammals of Arizona. University of Arizona Press, Tucson, 602 pp. Jackson, H. H. T. 1928 (see under S. cinereus). Rausch, R. 1951. Notes on the Nunamiut Eskimo and mammals of the Anaktuvuk Pass region. Brooks Range, Alaska. Arctic, 4:146-195. Youngman, P. M. 1975 (see under S. cinereus). S. nanus: Hoffmann, R. S., and J. G. Owen. 1980. Sorex tenellus and Sorex nanus. Mammalian Species, 131:1-4. Jones, J. K., Jr., D. M. Armstrong, R. S. Hoffmann, and C. Jones. 1983 (see under S. arctic us). S. ornatus: Owen, J. G., and R. S. Hoffmann. 1983. Sorex ornatus. Mammalian Species, 212: 1-5. 128 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 Rudd, R. L. 1955. Age, sex, and weight comparisons in three species of shrews. Journal of Mammalogy, 36:323-339. S. pacificus: Carraway, L. N. 1988. Records of reproduction in Sorex pacificus. The Southwestern Naturalist, 33:479-501. S. palmtris: Bailey, V. 1936. The mammals and life zones of Oregon. North American Fauna, 55:1-416. Conaway, C. H. 1952. Life history of the water shrew (Sorex palustris navigator). The American Midland Naturalist, 48:219-248. Conaway, C. H., and D. W. Pfitzer. 1952 (see under S. dispar). Connor, P. F. 1966 (see Literature Cited). French, T. W. Unpublished data. Grmm, W. C., and R. Whitebread. 1952 (see under B. brevicauda). Grinnell, J., J. Dixon, and J. M. Linsdale. 1930. Vertebrate natural history of a section of northern California through the Lassen Peak region. University of California Publications in Zoology, 35:1-594. Hall, E. R. 1946. Mammals of Nevada. University of California Press, Berkeley, 710 pp. Jackson, H. H. T. 1928 (see under S. cinereus). Kirkland, G. L., Jr. Unpublished data. Linzey, D. W., and a. V. Linzey. 1968 (see under 5. cinereus). Negus, N. C., and J. S. Findley. 1959. Mammals of Jackson Hole, Wyoming. Journal of Mammalogy, 40:371-381. Whitaker, J. O., Jr., G. S. Jones, and D. D. Pascal, Jr. 1975 (see under S. fumeus). S. sinuosus: Newman, J. R., and R. L. Rudd. 1978. Minimum and maximum metabolic rates of Sorex sinuosus. Acta Theriologica, 23:371-380. Rudd, R. L. 1955 (see under S. ornatus). Rust, A. K. 1978. Activity rhythms in the shrews, Sorex sinuosus Grinnell and Sorex trowbridgii Baird. The American Midland Naturalist, 99:369-382. S. trowbridgii: Doyle, A. T. 1990 (see under S. monticolus). Gashwiler, j. S. 1976. Notes on the reproduction of Trowbridge shrews in western Oregon. The Murrelet, 57:58-62. Jackson, H. H. T. 1928 (see under S. cinereus). Jameson, E. W., Jr. 1955. Observations on the biology of Sorex trowbridgei in the Sierra Nevada, California. Journal of Mammalogy, 36:339-345. Rust, A. K. 1978 (see under S. sinuosus). Scheffer, V. B., and W. W. Dalquest. 1942. A new shrew from Destruction Island, Washington. Journal of Mammalogy, 23:333-335. S. tundrensis: Martell, a. M., and a. M. Pearson. 1978 (see under S. cinereus). Sheftel, B. I. 1989 (see under N. fodiens). S. unguiculatus: Skaren, U. 1964 (see under S. araneus). S. vagrans: Bailey, V. 1936 (see under S. palustris). Clothier, R. R. 1955. Contribution to the life history of Sorex vagrans in Montana. Journal of Mammalogy, 36:214-221. Grinnell, I., J. Dixon, and J. M. Linsdale. 1930 (see under S. palustris). Hall, E. R. 1946 (see under S. palustris). Hawes, M. L. 1977 (see under S. monticolus). Hooven, E. F., R. F. Hoyer, AND' R. M. STORK4. 1975. Notes on the vagrant shrew, Sorex vagrans, in the Willamette Valley of western Oregon. Northw'est Science, 49:163-173. Jackson, H. H. T. 1928 (see under S. cinereus). Johnston, R. F., and R. L. Rudd. 1957. Breeding of the salt marsh shrew. Journal of Mammalogy, 38:157-163. Newman, J. R. 1976. Population dynamics of the wandering shrew Sorex vagrans. The Wasmann Journal of Biology, 34:235-250. Rudd, R. L. 1955 (see under S. ornatus). Vaughan, T. A. 1969. Reproduction and population densities in a montane small mammal fauna. Miscellaneous Publications, Museum of Natural History, University of Kansas, 51:51-74. S. vir: Sheftel, B. I. 1989 (see under V. fodiens). Yudin, B. S. 1962 (see under S. araneus). Soriculus caudatus: Mitchell, R. M. 1977 (see Literature Cited). Newton, P. N., M. R. W. Rands, and C. G. R. Bowden. 1990 (see under C. attenuata). S. nigrescens: Mitchell, R. M. 1977 (see Literature Cited). Newton, P. N., M. R. W. Rands, and C. G. R. Bowden. 1990 (see under C. attenuata). Suncus etruscus: DELigonnes, P. 1965. Captures de pachyures estrusques, Suncus etruscus (Savi, 1822) en Lozere. Mammalia, 29:620-622. Fons, R. 1970 (see Literature Cited). 1973. Modalities de la reproduction et development postnatal en captivite chez Suncus etruscus (Savi, 1822). Mammalia, 37:288-324. 1975. Premieres donnees sur I’ecologie de la pachyure etrusque Suncus etruscus (Savi, 1822) et comparison avec deux autres Crocidurinae: Crocidura russula (Herman, 1780) et Crocidura suaveolens (Pallas, 1811) (Insectivora Soricidae). Vie et Milieu, 25C:3 15-360. 1979. Duree de vie chez la Pachure etrusque, Suncus etruscus (Savi, 1822) en captivite (Insectivora, Soricidae). Zeitschrift flir Saugetierkunde, 44:241-248. Fons, R., and M.-C. Saint Girons. 1975. Notes sur les maminiferes de France. XIV.— Donnees morphologiques concernant la pachyure etrusque, Suncus etruscus (Savi, 1822). Mammalia, 39:685-688. Spitzenberger, V. F. 1970. Erstnachweise der Wimperspitzmaus (Suncus etruscus) fur Kreta and Kleinasien und die Verbreitung der Art im sudwestasiatischen Raum. Zeitschrift fiir Saugetierkunde, 35:107-113. Vogel, P. 1970 (see Literature Cited). 1994 INNES — Soricid Life Histories 129 S. murinus: Barbehenn, K. R. 1962. The house shrew on Guam. Pp. 247-256, in Pacific Island Rat Ecology (T. I. Storer, ed.), Bulletin of the Bernice Pauahi Bishop Museum, 225:1-274. Bauchot, R., and H. Stephan. 1966 (see under C. flavescens). Brooks, J. E., P. T. Htun, D. W. Walton, H. Naing, and M. M. Tun. 1980. The reproductive biology of Suncus murinus L. in Rangoon, Burma. Zeitschrift fur Saugetierkunde, 45:12-22. Dryden, G. L. 1968 (see Literature Cited). 1969 (see Literature Cited). 1970. Post-parturitional conception in captive musk shrews, Suncus murinus. Journal of Reproduction and Fertility, 23:493-495. Dryden, G. L., and R. R. Anderson. 1978. Milk composition and its relation to growth rate in the musk shrews, Suncus murinus. Comparative Biochemistry and Physiology, 60A:213-216. Dryden, G. L., and J. M. Ross. 1971 (see Literature Cited). Furumura, K., R. Kuriki, K. Ota, and A. Yokoyama. 1985. Reproduction. Pp. 126-139, in Suncus murinus. Biology of the Laboratory Shrew (S. Oda, J. Kitoh, K. Ota, and G. Isomura, eds.), Japan Scientific Societies Press, Tokyo, 523 pp. Harrison, J. L. 1955. Data on reproduction of some Malayan mammals. Proceedings of the Zoological Society of London, 125:445-460. Hasler, M. j., and a. V. NaLBANDOV. 1978. Pregnancy maintenance and progesterone concentrations in the musk shrew, Suncus murinus (Order: Insectivora). Biology of Reproduction, 19:407-413. Hasler, M. J., J. F. Hasler, and A. V. Nalbandov. 1977. Comparative biology of musk shrews (Suncus murinus) from Guam and Madagascar. Journal of Mammalogy, 58:285-290. Inouye, M., S. Oda, K. Shimamura, and Y. Kameyama. 1985. Embryonic and fetal development and teratogenic susceptibility. Pp. 140-154, in Suncus murinus. Biology of the Laboratory Shrew (S. Oda, J. Kitoh, K. Ota, and G. Isomura, eds.), Japan Scientific Societies Press, Tokyo, 523 pp. ISHIKAWA, A., AND T. Namikawa. 1987. Postnatal growth and development in laboratory strains of large and small musk shrews (Suncus murinus). Journal of Mammalogy, 68:766-774. ISHIKAWA, A., Y. TSUBOTA, AND T. NaMIKAWA. 1987. Morphological and reproductive characteristics of musk shrews (Suncus murinus) collected in Bangladesh, and development of the laboratory line (BAN line) derived from them. Experimental Animals, 36:253-260. ISHIKAWA, A., I. AKADAMA, T. NaMIKAWA, AND S. ODA. 1989. Development of a laboratory line (SRI line) derived from the wild house musk shrew, Suncus murinus, indigenous to Sri Lanka. Experimental Animals, 38:231-237. Louch, C. D.. a. K. Ghosh, and B. C. Pal. 1966. Seasonal changes in weight and reproductive activity of Suncus murinus in West Bengal, India. Journal of Mammalogy, 47:73-78. Medway, L. 1978. The Wild Mammals of Malaya (Peninsular Malaysia) and Singapore. Oxford University Press, Kuala Lumpur, 128 pp. Mitchell, R. M. 1977 (see Literature Cited). Newton, P. N., M. R. W. Rands, and C. G. R. Bowden. 1990 (see under C. attenuata). Rana, B. D., and I. Prakash. 1979. Reproductive biology and population structure of the house shrew, Suncus murinus sindensis, in western Rajasthan. Zeitschrift fur Saugetierkunde, 44:333-343. Rissman, E. F., R. j. Nelson, J. L. Black, and F. H. Bronson. 1987. Reproductive response of a tropical mammal, the musk shrew (Suncus murinus), to photoperiod. Journal of Reproduction and Fertility, 81:563-566. Shigehara, N. 1980. Epiphyseal union and eruption of the Ryukyu house shrew, Suncus murinus, in captivity. Journal of the Mammalogical Society of Japan, 8: 151-159 (not seen, cited in Ishikawa and Namikawa, 1987, and Michalak, 1987a, see under S. murinus and Literature Cited, respectively). Stine, C. J., and G. L. Dryden. 1977. Lip-licking behavior in captive musk shrews, Suncus murinus. Behavior, 62:298-313. Tsubota, Y., T. Namikawa, T. Nishida, A. Adachi, and H. W. Cyril. 1986 (see Literature Cited; not seen, cited in Ishikawa and Namikawa, 1987, see under S. murinus). Sylvisorex granti and S. lunaris: Dieterlen, F., and H. Heim DE Balsac. 1979 (see under C. flavescens). S. megalura: Bauchot, R., and H. Stephan. 1966 (see under C. flavescens). Dieterlen, F., and H. Heim de Balsac. 1979 (see under C. flavescens). Happold, D. C. D. 1987 (see under C. bottegi). S. vulcanorum: Hutterer, R., and W. Verheyen. 1985 (see Literature Cited). 130 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 Table 1. — Life-history traits and length of the breeding season for 93 species of the Soricidae. Means are given +1 standard error (SE) Species Male Mass (g) Female Mass (g) Combined Mass (g) Litter Size Gestation Length (Days) Blarina brevicauda 18.1 + 0.5 (25) 17.0 ± 0.5 (27) 16.9 ± 1.4 (11) 5.4 ± 0.2 (24) 20.3 ±0.5 (4) B. carolinensis 5.4 (1) Crocidura attenuata 7.4 (1) C. beatus 10.5 (1) 10.5 (1) C. bicolor 4.7 ± 1.2 (2) 5.7 (1) 4.0 (1) C. bottegi 2.0 (1) 3.2 (1) C. canariensis 7.5 (1) 2.1 (1) 30.0 (1) C. crossei 9.8 (1) 6.3 (1) 3.0 (1) C. dolichura 5.5 (1) C. douceti 4.3 (1) C. flavescens 45.4 + 9.8 (2) 24.8 (1) 3.5 ± 0.5 (5) C. foxi 21.0 (1) C. fuscomurina 5.0 (1) 3.0 (1) C. gracilipes 12.2 (1) 10.8 (1) 3.0 (1) C. grandiceps 21.9 (1) C. grayi 11.1 (1) 10.5 (1) 1.2 (1) C. hirta 16.9 (1) 15.1 (1) 22.0 (1) 3.9 ± 0.2 (6) C. horsfieldi 3.3 (1) C. jacks oni 12.7 (1) C. jouveneta 9.8 (1) 2.5 (1) C. kivuana 1.7 (1) C. lanosa 3.3 (1) C. leucodon 7.9 (1) 4.0 ±0.8 (3) 31.5 ± 0.5 (2) C. littoralis 21.0 (1) 3.0 (1) C. lusitania 4.1 (1) C. luna 9.5 (1) C. mariquensis 11.0 (1) 9.0 (1) 11.0 (1) 3.6 ± 0.2 (3) C. nanilla 3.4 (1) 2.3 (1) C. nigricans 21.0 (1) C. nigrofusca 3.3 (1) C. niobe 12.5 (1) C. olivieria 66.4 + 16.0 (3) 34.9 ± 7.4 (2) 3.1 ± 0.3 (6) C. osorio 5.7 (1) C. poensis 14.2 + 3.0 (2) 14.6 (1) 19.8 ± 3.3 (2) 2.3 (1) C. russula 10.6 ± 0.6 (7) 11.2 ± 0.7 (5) 13.0 ± 1.5 (3) 3.8 ± 0.2 (13) 29.7 ± 0.3 (3) C. suaveolens 7.3 + 0.3 (3) 7.7 ± 0.2 (4) 7.9 + 0.8 (3) 4.0 ± 0.7 (4) 27.6 ± 0.5(5) C. tarfayensis 6.5 (1) C. theresae 3.7 ± 1.2 (2) C. viaria 17.7 (1) 16.5 ± 0.5 (2) 3.5 (1) C. whitakeri 5.5 (1) C. zinimermani 7.7 (1) Cryptotis tnagna 3.0 (1) C. parva 4.7 + 0.02 (8) 5.0 ± 0.2 (6) 4.4 + 0.1 (3) 4.2 ± 0.3 (11) 22.0 ± 0.4 (4) Diplomesodon pulchellum 8.9 (1) Myosorex baboulti M. norae 23.7 (1) 25.1 (1) M. polulus 19.5 (1) 17.6 (1) M. varius 12.5 ± 1.1 (4) 2.9 ± 0.1 (3) Neotnys anomalus 11.9 ± 1.0 (2) 14.0 ± 0.7 (2) 10.9 ± 0.7 (2) 7.2 ± 1.7 (3) N. fodiens 14.6 ± 1.2 (5) 14.9 ± 0.9 (4) 16.4 ± 1.5 (5) 6.3 ± 0.6 (11) 20.8 ± 0.9 (5) Notiosorex crawfordi 5.4 (1) 4.1 ± 0.4 (2) 3.9 ± 0.2 (2) N. gigas 13.1 ± 2.2 (2) 1994 INNES— Soricid Life Histories 131 and the number of populations are in parentheses. See Methods for further explanation. Sources for each species appear in Appendix. Neonate Age Eyes Age at Mass at Growth Life Breeding Mass (g) Open (Days) Weaning (Days) Weaning (g) Rate (g/day) Span (Months) Season (Months) 0.8 (1) 12.0 (1) 20.6 ± 1.5 (5) 17.2 + 1.5 (6) 6.6 ± 0.3 (10) 7.0 (1) 0.8+ 0.6 (2) 13.0 ± 2.1 (3) 19.3 ± 3.2 (3) 2.1 ± 0.9 (2) 0.06 ± 0.20 (2) 0.8 (1) 9.0 ± 3.0 (2) 1.0 (1) 13.0 (1) 18.0 (1) 0.8 ± 0.1 (2) 11.7 ± 1.3 (3) 20.5 ± 0.5 (2) 7.0 (1) 0.30 (1) 8.0 (1) 2.1 ± 0.2 (2) 15.0 (1) 23.5 + 3.5 (2) 16.4 ± 4.4 (2) 0.65 ± 0.28 (2) 1.0 ± 0.1 (4) 10.7 ± 1.2 (3) 20.0 + 0.6 (3) 8.2 + 0.9 (2) 0.37 ± 0.04 (2) 27.0 (1) 7.5 ± 0.5 (2) 0.7 ± 0.1 (3) 10.8 + 1.4 (4) 18.0 ± 1.1 (4) 6.5 ± 2.1 (2) 0.32 ± 0.13 (2) 7.0 (1) 1.4 (1) 20.0 (1) 13.7 (1) 0.62 (1) 0.3 ± 0.1 (6) 14.3 ± 0.3 (4) 20.0 ± 0.6 (3) 3.2 (1) 0.15 (1) 11.0 (1) 7.5 ± 0.5 (2) 2.5 (1) 1.0 (1) 17.0 (1) 22.0 (1) 16.0 (1) 9.0 (1) 0.6 (1) 22.0 (i) 28.0 (1) 9.3 (1) 0.31 (1) 5.0 (1) + 0.1 (2) 22.0 ± 0.6 (3) 29.3 ± 1.3 (4) 11.2 ± 0.8 (3) 0.50 ± 0.07 (2) 13.5 + 5.5 (2) 6.5 ± 0.5 (2) 5.0 (1) 132 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 Table 1 (cont.) Species Male Mass (g) Female Mass (g) Combined Mass (g) Litter Size Gestation Length (Days) Scutisorex somereni 1.9 (1) Sorex alpinus 7.7 (1) 8.5 (1) S. araneus 9.8 ± 0.4 (14) 9.9 ± 0.4 (16) 9.2 ± 0.3 (32) 6.7 ± 0.2 (22) 21.5 ± 0.7 (6) S. arcticus 8.0 + 0.4 (4) 8.0 ± 0.4 (4) 6.7 ± 0.3 (10) S. betidirri 15.3 + 0.8 (2) S. caecutiens 6.5 + 0.6 (2) 6.0 (1) 5.9 + 0.1 (2) 7.4 ± 0.4 (8) S. dnereus 4.2 + 0.1 (22) 4.4 ± 0.2 (24) 3.9 ± 0.3 (15) 6.5 ± 0.3 (31) S. coronatus 8.6 + 0.3 (3) 9.4 ± 0.3 (3) 4.9 ± 0.5 (5) 24.0 (1) S. daphaenodon 6.2 + 0.4 (3) S. dispar 5.0 ± 0.4 (4) 4.8 ± 0.5 (2) 4.9 ± 0.2 (6) 3.5 ± 1.5 (2) S. fumeus 7.7 ± 0.3 (15) 7.2 ± 0.3 (12) 7.6 ± 0.1 (2) 4.6 + 0.4 (9) S. gaspensis 3.2 (1) 4.0 + 1.1 (2) 5.7 + 0.0 (2) S. granarius 6.6 (1) S. hoyi 3.2 + 1.0 (4) 3.1 ± 0.7 (3) 3.8 ± 0.7 (4) 5.7 (1) S. isodon 12.3 ± 0.7 (3) 11.9 ± 0.6 (3) 6.7 ± 0.6 (3) 18.0 (1) S. longirostris 3.1 (1) 2.6 (1) 3.1 ± 0.2 (9) 4.0 ± 0.3 (4) S. macrodon 10.6 (1) S. meniami 5.0 (1) 5.9 (1) 6.0 (1) S. milleri 4.1 (1) 3.5 (1) S. minutisslmus 2.0 ± 0.3 (2) 2.5 (1) S. minutus 4.5 ± 0.2 (7) 4.4 ± 0.2 (8) 3.9 ± 0.2 (9) 6.1 + 0.4 (13) 25.0 (1) S. mirabilis 15.0 (1) S. monticolus 6.9 ± 0.2 (3) 5.9 (1) 6.4 ± 1.1 (4) S. nanus 2.6 (1) 6.5 (1) S. ornatus 5.2 (1) 5.0 (1) 5.1 (1) S. padficus 4.2 (1) S. palustris 13.3 ± 0.8 (9) 11.7 ± 0.6 (8) 12.7 ± 1.5 (3) 5.5 ± 0.3 (5) S. sinuosus 6.1 + 0.1 (2) 5.6 ± 0.2 (2) 5.8 + 0.5 (2) S. trowbridgii 4.7 ± 0.2 (3) 5.2 (1) 7.4 ± 0.2 (2) 4.1 ± 0.5 (3) S. tundrensis 6.3 (1) 8.7 ± 1.2 (2) S. unguiculatus 13.2 (1) S. vagrans 6.4 ± 0.4 (4) 6.6 ± 0.7 (3) 6.9 (1) 5.9 + 0.2 (9) 20.0 (1) S. vir 7.5 ± 0.1 (2) Soriculus caudatus 4.8 (1) S. nigrescens 15.2 (1) 4.9 (1) Suncus etruscus 1.8 + 0.0 (2) 2.0 ± 0.1 (2) 2.1 ± 0.3 (2) 3.9 ± 0.2 (2) 28.0 ± 0.0 (2) S. rnurinus 65.2 ± 10.2 (9) 42.8 ± 6.3 (10) 2.9 + 0.2 (18) 29.6 ± 0.5(11) Sylvisorex granti 3.8 (1) 1.6 (1) S. lunaris 2.6 (1) S. megalura 5.3 (1) 4.0 (1) 1.8 (1) S. vulcanorum 3.5 (1) 1994 INNES — Soricid Life Histories 133 Neonate Age Eyes Age at Mass at Growth Life Breeding Mass (g) Open (Days) Weaning (Days) Weaning (g) Rate (g/day) Span (Months) Season (Months) 12.0 (1) 0.4 ± 0.1 (4) 193.5 ± 0.5 (4) 23.4 + 0.8 (7) 7.6 ± 0.5 (4) 0.31 + 0.01 (4) 14.9 ± 0.8 (7) 6.1 ± 0.3 (12) 16.5 + 1.5 (2) 5.3 ± 0.5 (4) 16.0 (1) 4.4 ± 0.3 (5) 0.3 (1) 18.0 (1) 20.0 (1) 3.5 (1) 0.16 (1) 20.5 ± 2.5 (2) 5.3 ± 0.4 (8) 0.6 (1) 30.0 (1) 8.6 (1) 0.27 (1) 20.0 ± 1.0 (1) 8.0 ± 2.0 (2) 17.0 (1) 5.7 + 0.7 (3) 0.9 (1) 16.0 (1) 22.0 (1) 7.3 (1) 0.29 (1) 16.0 (1) 5.0 ± 0.0 (2) 6.0 ± 0.0 (2) 0.2 + 0.1 (2) 26.3 + 2.0 (3) 3.2 (1) 0.10 (1) 16.3 ± 1.2 (3) 6.4 + 0.5 (7) 21.0 (1) 16.0 (1) 16.0 (1) 7.0 (1) 16.0 (1) 18.0 (1) 8.0 ± 1.0 (2) 0.4 ± 0.1 (2) 21.0 (1) 22.0 + 1.5 (3) 3.9 + 0.4 (2) 0.15 ± 0.0 (2) 19.0 + 3.0 (3) 6.5 + 1.5 (2) 6.0 (1) 0.2 ± 0.1 (2) 14.5 + 0.5 (2) 19.5 ± 0.5 (2) 2.2 + 0.1 (2) 0.10 + 0.01 (2) 17.0 (1) 6.0 (1) 2.8 ± 0.2 (8) 8.6 ± 0.2 (5) 18.9 ± 1.0 (8) 30.4 ± 5.5 (6) 1.39 ± 0.26 (6) 10.4 ± 1.0 (7) 134 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 Table 2.— Life-history traits arid length of the breeding season compared between Crocidura and Sorex. Data were derived from Table 1 by averaging species means for each genus. Means are given ±1 SE and the number of species are in parentheses. ° P < 0.05; * P < 0.01; ^ P < 0.001. Variable Crocidura Sorex t-value Male mass (g) 15.1 ± 3.4(20) 6.3 ± 0.6 (22) 3.5*’ Female mass (g) 12.6 ± 2.2(14) 6.7 ± 0.7 (22) 2.9“ Combined mass (g) 10.6 + 1.5(21) 6.9 ± 0.8 (20) 1.9 Litter size 3.1 ± 0.2 (21) 5.9 ± 0.3 (22) bo o Gestation length (days) 29.7 ±0.8 (4) 21.7 ± 1.3 (5) 4.4*’ Neonate mass (g) 1.1 ± 0.2 (8) 0.5 ± 0.1 (6) 3.7*’ Age eyes open (days) 12.4 ± 0.7 (6) 18.6 ± 1.1 (4) 5.1*’ Age at weaning (days) 19.9 ± 0.7 (7) 23.5 ± 1.3 (7) 2.6“ Mass at weaning (g) 9.0 ± 2.1 (6) 5.7 ± 1.0 (6) 1.0 Growth rate (g/day) 0.39 ± 0.1 (6) 0.21 ± 0.1 (6) 1.6 Life span (months) 27.0 (1) 17.1 ± 1.7 (13) 4.1“ Breeding season (months) 7.9 ± 0.4 (4) 6.1 + 0.3 (12) 2.7“ Table 3. — Life-history traits and length of the breeding season compared between the Crocidurinae and Soricinae. Data were derived from Table 2 by averaging genus means within each subfamily. Means are given ±1 SE and the number of genera are given in parentheses. Mann-Whitney U test; ^ P < 0.01. Variable Crocidurinae Soricinae Ug-Value^ Male mass (g) 23.4 ± 5.4 (3) 9.6 ± 2.6 (5) 14.0 Female mass (g) 15.4 ± 4.0 (4) 10.8 ± 2.9 (4) 5.0 Combined mass (g) 7.6 ± 2.0 (5) 10.9 ± 2.0 (6) 21.0 Litter size 2.7 ± 0.3 (5) 5.1 ± 0.5 (6) 30.0*’ Gestation length (days) 29.3 ± 0.5 (2) 21.2 ± 0.4 (4) 11.0 Neonate mass (g) 1.4 ± 0.2 (3) 0.6 ± 0.1 (4) 12.0 Age eyes open (days) 13.7 ± 1.7 (3) 16.7 ± 2.2 (2) 9.0 Age at weaning (days) 20.4 ± 0.8 (3) 23.2 ± 2.0 (3) 10.0 Mass at weaning (g) 12.7 ± 3.7 (2) 6.4 ± 2.1 (3) 5.0 Growth rate (g/day) 0.57 ± 0.2 (2) 0.26 ± 0.1 (3) 5.0 Life span (months) 20.0 ± 3.5 (3) 14.7 ± 1.5 (2) 8.0 Breeding season (months) 9.3 ± 0.9 (4) 6.2 ± 0.4 (6) 24.0*’ Table 4.— Correlations among life-history traits (excluding male mass and combined mass) among all populations (lower triangle) and only among Crocidura and Sorex populations (upper triangle). Only significant populations appears in parentheses. " P < 0.05; * P < 0.01; P < 0.001. T-values are reported and the number of (1) (2) (3) (4) (5) (6) (7) (8) (9) (1) Female mass (2) Litter size (3) Gestation length (4) Neonate mass (5) Age eyes open (6) Age at weaning (7) Mass at weaning (8) Growth to weaning (9) Life span — -0.43*’ (58) -0.45"^ (80) — -0.77*’ (12) -0.8L (30) - 0.91*’ (8) -0.56*’ (31) 0.43“ (24) 0.73‘’ (22) -0.76^= (22) -0.51*’ (29) 0.95*’ (7) -0.51“ (24) 0.74*’ (11) -0.73“ (10) - -0.52“ (18) -0.60^^ (31) — 0.64*’ (19) -0.33“ (38) 0.7U (34) — 0.86*’ (32) 0.74^^ (27) -0.50“ (17) 0.68*’ (20) 0.68*’ (20) — 0.90'-' (20) 0.9U (27) — 1994 INNES — Soricid Life Histories 135 Table 5. — Correlations among life-history traits (excluding male mass and combined mass) among all species (lower triangle) and only among Crocidura and Sorex species (upper triangle). Only significant r-values are reported and the number of species appears in parentheses. P < 0.05; * P < 0.01; ^ P < 0.001. (1) (2) (3) (4) (5) (6) (7) (8) (9) (1) Female mass (2) Litter size (3) Gestation length (4) Neonate mass (5) Age eyes open (6) Age at weaning (7) Mass at weaning (8) Growth to weaning (9) Life span 0.90‘' (17) 0.94‘= (15) 0.89^= (14) -0.79*’ (14) -0.52“ (21) 0.68** (17) 0.49“ (21) -0.78“ (9) -0.71“ (11) 0.86*’ (11) -0.66“ (14) 0.82^^ (17) 0.78^^ (16) 0.74“ (10) -0.84“ (6) 0.76"^ (17) 0.89*’ (10) 0.92"^ (10) -0.58“ (12) -0.67“ (12) 0.70“ (12) 0.76*’ (12) 0.65“ (10) — 0.96^= (12) 0.95*= (16) — 6. —Correlations among life-history traits (excluding male mass and combined mass) among genera. Only significant r-values are reported and the number of genera are given in parentheses. (Missing rows and columns are the result of nonsignificant values.) “ P < 0.05; ^ P < 0.01. (1) (2) (4) (5) (7) (1) Female mass (2) Litter size (3) Gestation length (4) Neonate mass 0.92*’ (7) -0.82“ (6) (5) Age eyes open (6) Age at weaning (7) Mass at weaning 0.99*^ (5) 0.92*’ (7) 0.93“ (5) (8) Growth to weaning 0.96“ (5) 0.92“ (5) 0.94“ (5) Table 7. — Correlations among life-history traits controlling for female body mass among all species. Only significant r-values are reported and the number of species occurs in parentheses. (Missing rows and columns are the result of nonsignificant values.) Compare with Table 5 (lower triangle). " P < 0. 05; * P < 0.01; ^ P < 0.001. (1) (2) (3) (4) (6) (7) (1) Litter size (2) Gestation length (3) Neonate mass -0.82*’ (9) -0.85'’ (14) 0.63“ (9) (4) Age eyes open 0.67” (12) -0.75“ (8) -0.63“ (12) — (5) Age at weaning 0.52“ (15) -0.69*’ (14) 0.82” (12) (6) Mass at weaning (7) Growth to weaning (9) Breeding season -0.82” (19) 0.63“ (9) 0.71“ (9) 0.65“ (10) 0.76*’ (11) — -0.77*’ (9) 136 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 Fig. 1.— Relationship between litter size and length of the breeding season among populations. Symbols for each genus are as follows: Blarina ♦; Crocidura ■; Cryptotis O; Myosorex a; Neomys *; Notiosorex □; Scutisorex O; Soricuius ©; Sorex i • ; Suncus a. Fig. 2. — Relationship between litter size and adult female mass among populations. Symbols for each genus are as follows: Blarina ♦ ; Crocidura ■ ; Cryptotis O ; Neomys Sorex • ; Suncus a . Note that the scale of the x-axis changes after 25 g. SHREWS AS INDICATORS OF HEAVY METAL POLLUTION Erkki Pankakoski*, Ilkka KorvisTo^, Heikki Hyvarinen^, and Juhani Terhivuo'^ 'Department of Zoology, P. O. Box 17 (P. Rautatiekatu 13), FIN-00014 University of Helsinki, Finland; ^Helsinki Zoo, FIN-00570 Helsinki, Finland; ^Department of Biology, University of Joensuu, Box 111, FIN-80101 Joensuu, Finland; ‘'Zoological Museum, P. O. Box 17 (P. Rautatiekatu 13), FIN-00014 University of Helsinki, Finland Abstract Heavy metal (Cu, Ni, Zn, Cd, Cr, Hg, Pb) concentrations were analyzed in small mammals trapped near the sources of pollution and on control sites in southern Finland. Shrews (Sorex and Neomys) had higher heavy metal concentrations in their livers and kidneys than voles. In S. araneus, the most common species of shrew in southern Finland, heavy metal concentrations decreased with increasing distance from the source of pollution. Some heavy metals (Cd, especially) accumulated with increasing age of individual shrews. Heavy metal accumulation in shrew tissues tended to increase with increasing soil acidity. The biological effect of heavy metal pollution was studied by assessing developmental stability in skull morphology of S. araneus. Developmental stability was usually reduced in the polluted areas, possibly due to toxic effects of heavy metals. S. araneus may be a useful bioindicator of pollution in terrestrial ecosystems because it is a relatively localized animal with high potential of reproduction and also is abundant in areas intensively affected by human activities. Introduction Small mammals have been used in field studies to indicate accumulation of toxic compounds such as heavy metals. For example, the effect of automobile exhaust lead residues along roadsides has been studied by analyzing lead concentrations in small mammals (Quarles et al., 1974; Getz et al., 1977; Goldsmith and Scanlon, 1977). Interspecific differences among small mammals in the accumulation efficiency of heavy metals have been found. Herbivorous or granivorous voles and mice accumulate smaller amounts of toxic compounds in their bodies than predatory mammals such as shrews (Goldsmith and Scanlon, 1977; Hunter and Johnson, 1982; Andrews et al., 1984, 1989; Forsyth and Peterle, 1984; Beyer et al., 1985; Hunter et al., 1987; Hegstrom and West, 1989; Ma, 1989), although some contradictory results exist (Roberts and Johnson, 1978). Small mammals can achieve high population densities, thereby yielding adequate sample sizes for studies of heavy metal bioaccumulation with reasonable work effort. Shrews may he particularly useful in this respect because they have high consumption rates due to their high metabolic rates (Hanski, 1984). Some species, such as the common shrew, Sorex araneus, eat large quantities of earthworms (Lumbricidae; e.g., Pemetta, 1976), which in turn accumulate high concentrations of heavy metals (Ireland, 1983; Morgan and Morgan, 1988; Ma, 1989). The aim of this study was to evaluate the usefulness of shrews in studying heavy metal pollution. We compare shrews and rodents living in the same polluted area and then focus on different species of shrews to find the most effective indicator. We also examine differences between sexes and age groups of shrews in this respect. Heavy metal accumulation in shrews evidently depends on the amount of metals in both soil and food resources, which again may be affected by physical factors, such as the distance from the point source of pollution and soil pH. We present some data for these relationships here. Negative detrimental effects attributed to high heavy metal concentrations in the tissues of small mammals have been reported, including shrews. These include histopathological changes of the kidneys due to Cd and Pb, and increased relative organ weights and decreased body size (Goyer et al., 1970; Nicholson et al., 1983; Andrews et al., 1984; Ma, 1989). However, the polluted environment may stress an animal population, even if acute poisoning effects are not observed. The level of this stress can be assessed by analyzing the developmental stability of the population (Yablokov, 1986; Zakharov, 1989). In stressing environments, developmental stability of the population is reduced, which results in an increased level of fluctuating asymmetry in morphological traits (Leary and Allendorf, 1989), increased number of phenodeviants (exceptional forms of some meristic character; Zakharov, 1989), and in some cases in increased amounts of total phenotypic variability (Soule, 1982; Zakharov, 1984; Pankakoski et al., 1987). Reduced developmental stability due to toxic compounds in the environment has been shown in fish and seal populations (Valentine and Soule, 1973; Valentine et al., 1973; Jagoe and Haines, 1985; Zakharov, 1989; Zakharov et al., 1989). In our earlier study (Pankakoski et al., 1992) we demonstrate that heavy metal pollution may also affect the developmental stability of shrew populations. We summarize the results of that earlier paper here, along with some new analyses. When assessing developmental stability in shrew populations our hypothesis is that the progeny of female shrews exposed to heavy metals has a decreased level of developmental stability. The change in level of developmental stability implies the biological meaning of pollution for that shrew population. Methods We analyzed concentrations of several heavy metals (Cu, Ni, Zn, Cd, Cr, Hg, and Pb) in the livers and kidneys of small mammals trapped near sources of pollution and in control sites in southern Finland. We present here results of the five most common species in southern Finland. Three of these are insectivorous species, namely the common shrew S. araneus, pygmy shrew S. minutus, and water shrew Neomys fodiens\ the two others are herbivorous rodents, viz. , the field vole Microtus 137 138 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 agrestis, and red-backed vole Clethrionomys glareolus. Field Studies and Descriptions of Study Sites The trapping was conducted (August 1987 and 1988) on polluted sites near the metal works and on control sites. The control sites have the average background level of heavy metals in south Finland; completely clean areas are not found anywhere in Finland (Riihling et al., 1987), but these were known to be much less contaminated than the polluted sites. The three main study areas were as follows: 1) Tikkurila. — A lead smeltery in the city of Vantaa (60°16’N, 25°03’E) and its control area in northern part of the city of Espoo, ca 20 km west of Tikkurila. From the late 1920s until the early 1980s the lead smeltery in Tikkurila heavily polluted the soil (Ervio and Lakanen, 1973); high levels of lead have been recorded in mushrooms there (Liukkonen-Liljaet al., 1983). In Tikkurila the traps were set on verges between a grain field and a mixed forest of Populus tremula, Betula spp., Salix spp., Picea abies, and Pinus silvestris. The dense ground vegetation layer was comprised of several species of grasses and herbs. The control area in Espoo was a forested area, where soil analyses and accumulation of metals in moss bags (for the latter method, see Makinen, 1977) has shown low levels of heavy metals (Marttinen and Edgren, 1984; Kinnunen et al., 1985). In Espoo the moist mixed forest was dominated by Picea abies, Betula spp., and Populus tremula-, the undergrowth was mostly Vaccinium my rt Ulus, mosses, and different species of grasses. 2) Harjavalta. — An industrial town in southwestern Finland (61°19’N, 22°07’E) and control sites at 10 and 20 km distances in rural areas, in Panelia and Eurajoki communes. According to earlier studies, the Haijavalta works (copper, sulfuric acid, and fertilizer factories) spread several heavy metals (especially Cd, Cu, and Ni) over a wide area (Hynninen, 1986). By using the moss bag method, Hynninen (1986) showed significantly elevated concentrations of several heavy metals as far as 9 km from the works. The rate of accumulation is very high around the works, but it rapidly decreases with increasing distance (Hynninen, 1986; Hynninen and Lodenius, 1986). Small mammals were trapped at a distance of 1-2 km from the factories, as closer to the works the undergrowth was too sparse due to pollution to give vegetative cover for mammals. Trapping was performed on several restricted areas, including shrubby sites (Betula and Salix spp.) along the edges of fields and along a small creek, abandoned old fields, and moist mixed forest (dominated by Betula spp., Pinus silvestris and Picea abies) with several species of herbs and grasses in the ground vegetation layer. The two control sites were situated against the prevailing winds that transport the metals (Hynninen, 1986; Hynninen and Lodenius, 1986) southwest of Harjavalta in mixed forests dominated by Picea abies, Betula, and Salix. In the undergrowth Vaccinium myrtillus and V. vitis-idaea dominated, along with grasses and herbs. 3) Koverhar.— An iron and steel works in Hanko, in southern Finland (59°53’N, 23°13’E) and its control area 4 km away in Tvarminne. Iron and lead concentrations in lichens are clearly increased at the distance of 2 km and slightly increased still at 5 km from the works (Helminen et al., 1986). Soil is [ much more alkaline in Koverhar than is common in Finnish forests due to the addition of lime in the iron smelting process (Fritze, 1991). The trapping habitats were similar both in ■ Koverhar and its control area: moist mixed forest situated at the seashore and dominated by Alnus glutinosa, Betula spp., and j Pinus silvestris. The soil was wet and the undergrowth was comprised of several species of moist habitat grasses and herbs, such as Filipendula ulmaria, Phragmites australis, and Equisetum palustris. Ten soil samples each from Koverhar and its control area were analyzed for heavy metals and pH. Small mammals were trapped with metal or plastic pitfall I traps which were partly filled with water to kill the animal { (Pankakoski, 1979). Snap traps baited with bread or cheese also 1 were used in Tikkurila and its control area in Espoo. The small |i mammals caught were put in plastic bags and deep-frozen. Soil acidity (pH) was measured in samples (usually ten samples/area) from the lower part of humus layer (above the t leaching layer). The pH was measured at -f-23°C in a mixture [ of 15 ml soil and 25 ml distilled water using a Knick Co. analyzer. In Tikkurila 16 sample plots (area 50 X 25 cm, 30 cm deep) were dug up and hand sorted for earthworms in the field. The numbers of species and individuals were counted later. Soil samples were taken at 3 cm depth in each plot and analyzed for | lead. Eight sample plots were similarly studied in Pomainen (60°28’N, 25°20’E) 25 km northeast from Tikkurila in a habitat similar to that in Tikkurila. Laboratory Studies The small mammals were weighed and sexed, and the livers | and kidneys were removed for analysis of heavy metals. These | organs usually accumulate more heavy metals than other soft tissues. The age grouping of shrews was determined by the i general size of the individual, the color and condition of the | coat, especially the coat of the hind feet and tail (Crowcroft, 1957), and by the level of tooth wear (Pankakoski, 1989). Juvenile shrews were defined as individuals trapped in their | year of birth; adults were bom in the previous summer (Crowcroft, 1957; Pankakoski, 1989). Juveniles were almost ii always immature, all adults were mature. j Heavy metals in small mammals were analyzed using a | SpectraMetrics SpectraSpan HIB plasmaemission > spectrophotometer (Cd, Cr, Ni, Pb), atomic absorption spectrophotometer (AAS; Cu, Zn), or gold film mercury I analyzer (Hg; Jerome Instrument Corporation, Model 511). Tissues were dried overnight at + 105°C, weighed, and digested j in 10 ml of a 1:4 mixture of HCIO4 and HNO3 under a reflux \ cap. In order to obtain sufficient quantities of S. minutus kidney for analyses, it was necessary to pool organs from 3-4 individuals of the same age and sex. Heavy metals in soil samples and earthworms were analyzed using atomic absorption spectrophotometer. Heavy metal concentrations (mg/kg) are expressed on a dry weight basis. Earthworms sampled in Tikkurila and Pomainen were kept alive on moist absorbent paper in a refrigerator for some days to make them empty their guts. Samples comprising 1-19 1994 PANKAKOSKI ET AL. — Shrews as Indicators of Heavy Metals 139 individuals of Aporrectodea caliginosa, the commonest lumbricid there, were deep-frozen in small plastic tubes and later analyzed for lead (AAS). Statistical comparisons of heavy metal concentrations were based on median values and nonparametric tests (Mann-Whitney U-test, Kruskal-Wallis test, and Spearman rank correlation coefficient), because the data mostly were not normally distributed. There were usually some individuals with very high concentrations, the fact that violates the assumptions of parametric tests and artificially raises the arithmetic mean. Developmental stability of S. araneus was assessed by using the numbeis of small paired foramina in the skull and from mandible measurements (for the method, see Pankakoski and Hanski, 1989). The foramina were counted as a blind test without knowing the trapping place of the animal. The metrical measurements of both mandibles were taken by using a video camera connected to a digitizer and a microcomputer (Pankakoski and Hanski, 1989). In developmental stability analyses we used Friedman’s nonparametric test to compare variances of several traits. Numbers of asymmetric traits or phenodeviants were compared with one-way analysis of variance. For the details of analyzing developmental stability, see Pankakoski et al., (1992). Results Inter- and Intraspecific Comparisons Differences in heavy metal concentrations between shrews and voles in Tikkurila and its control area in northern Espoo are presented in Table 1. In both areas the concentrations of Cd, Pb, and Hg were higher in S. araneus and N. fodiens than in C. glareolus and M. agrestis. Copper concentrations seemed to be higher in S. araneus than in the other species, but this difference was usually not significant. Neomys fodiens may accumulate high amounts of Hg, as observed in the small sample from the control area in Espoo; the median of Hg concentrations in liver was 4.47 mg/kg (Table 1). Excepting Ni, heavy metal concentrations in liver did not differ significantly between S. araneus and S. minutus (Haijavalta sample, juveniles; Table 2). Concentrations of all metals but Zn, however, were significantly higher in the kidneys of S. araneus. Sorex araneus adults tended to have higher concentrations of heavy metals than juveniles (Table 3). The difference was greatest in Cd (both liver and kidney), but significant also in Zn, Cu, and Hg (liver). Chromium and Pb did not exhibit higher concentrations in adults, but Pb in kidneys was significantly higher in juveniles. No differences were observed in heavy metal accumulation between the sexes of shrews. Ihe Effect of Distance from the Pollution Source In S. araneus trapped less than I km from the lead smeltery in Tikkurila the concentration of Pb in the liver decreased as the distance from the pollution source increased (Fig. 1). The Pb in both soil and earthworms {Aporrectodea caliginosa) also decreased according to increasing distance in the same area (Fig. 2A). The total number of earthworms, however, increased by increasing distance from the lead smeltery in Tikkurila (Fig. 2B). Some sample plots close to the smeltery were devoid of earthworms. The median for Pb concentrations in 16 soil samples from Tikkurila was ten times higher than the corresponding median for the eight samples taken in similar habitat at a rural site (Table 4). Lead concentrations were almost 20 times higher in A. caliginosa from Tikkurila than in control samples (Table 4). Because earthworms are an important food resource for the common shrew, our results imply one pathway of Pb, from soil through earthworms to shrews. On a larger scale, the effect of emissions from the metal industry in Haijavalta was demonstrated in tissues of S. araneus sampled at 10 and 20 km from Haijavalta. However, the decreasing trend was evident; hence, the lowest concentrations usually were from the control site at 20 km from the factories (Fig. 3). Heavy Metal Bioaccumulation and Soil pH Bioavailability of several heavy metals is higher in acidified soils than in soils with high pH (e.g., Andersson and Nilson, 1974; Roberts et al., 1978; Beyer et al., 1987; Scheuhammer, 1991). In Tikkurila, its control area in Espoo, and in the Harjavalta area soil pH ranged from 4.2 to 5.4 (Table 5), which is not beyond normal variation in Finnish soils. The range is due mainly to differences in soil types: forest soils are more acidic than old field and arable soils. On the other hand, as seen in Table 5, the mean (±SD) soil pH values had much greater range in Koverhar and its control area (7.60 + 0.28 and 4.59 ± 0.29, respectively; P < 0.001, t-test). Due to addition of lime during the smelting process at the Koverhar steel factory, much calcium is distributed to the surroundings and, consequently, the soil pH was highly significantly elevated near the works (see also Fritze, 1991). If soil pH affects heavy metal concentrations, the lower pH values at the control sites than at the polluted sites of Tikkurila and Harjavalta should increase heavy metal concentrations of shrews there. However, the concentrations show the opposite trend, at least in Haijavalta, i.e., lower values were from control sites (Fig. 3). There was much more Pb in shrews in Tikkurila than in the control area in Espoo, but the Cd and Zn concentrations were higher in Espoo (Table 6). The strong influence of the heavy metal pollution source may overcome the impact of pH in these study sites. Top soil layers near the Koverhar metal works contained higher concentrations of heavy metals than the soil in the control area (significant for Pb and Zn, ten samples on both sites; Cd concentrations were below the detection level in both sites; Fig. 4). However, shrews on the control site had a significantly greater burden of Pb in liver and kidneys, and of Zn, Cu, and Ni in the kidneys (Fig. 4; for the corresponding medians, see Table 7). The accumulations of Cr in all tissues and of Cd and Zn in the livers of 5. araneus from Koverhar were greater than in the control area (Fig. 4; for Cd, see Table 7). The fact that Pb, especially, had higher concentrations in the control area is perhaps related to the differences in the soil pH of the two sites. 140 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 Heavy Metals and Developmental Stability of Shrew Populations The biological effect of heavy metal pollution was assessed by comparing developmental stability in morphological skull characters of S. araneus in polluted and control areas. The stress of heavy metal pollution should decrease the level of developmental stability in shrew populations (see Pankakoski et al., 1992). Figure 5 summarizes the data from the three main study areas, Tikkurila (A), Harjavalta (B), Koverhar (C), and their control areas. Comparisons of heavy metal concentrations among these sites are presented in Tables 6 and 7 and Fig. 3 and 4. Developmental stability was usually reduced in areas where shrews have high heavy metal concentrations in their tissues (Fig. 5). This was indicated by the tendency towards higher levels of fluctuating asymmetry and total phenotypic variability, as well as the higher numbers of asymmetric traits and phenodeviants in the polluted areas (Tikkurila and Harjavalta, Fig. 5A, B). In the Haijavalta area there was a parallel trend in most characteristics; the columns in Fig. 5B are highest in Haijavalta (indicating high level of disturbances during development, i.e., low developmental stability) and they decrease with increasing distance, as was the case also in heavy metal concentrations. In Koverhar (Fig. 5C), the result of developmental stability in “polluted” and control areas is controversial (reduced developmental stability at the control site), as were also the concentrations of most heavy metals in the shrews (more heavy metals at the control site; see Fig. 4). When there was a significant difference between Koverhar and its control area, developmental stability was lower in the control area (high columns in Fig. 5C: FA of foramina and PV of mandibular measurements). Discussion Searching for an Indicator Species A good indicator species for heavy metal pollution accumulates high concentrations of different metals in its tissues. Such a species is also useful in slightly polluted areas. The indicator species should also occupy a wide range of hab- itats and be abundant enough to provide adequate sample sizes. Shrews Versus Rodents. — In our study, shrews had clearly higher concentrations of the most toxic heavy metals (Cd, Pb, and Hg) than the rodents in the same area. This agrees with the studies by Goldsmith and Scanlon (1977), Hunter and Johnson (1982), Andrews et al. (1984, 1989), Beyer et al. (1985), and Hunter et al. (1987). Shrews and voles have similar Zn concentrations. In mammalian tissues, Zn is usually effectively controlled by regulatory systems, which may be violated by high levels of Cd (Bremner, 1974; Czamowska and Gworek, 1980; Wlostowski, 1987). At least when Cd, Pb, or Hg are considered, shrews are better indicators than the rodents. The higher concentrations are doubtless due to the higher amount of these metals in the largely animal foods eaten by shrews (Hunter et al., 1989). The Pb concentration in soil, earthworms, and shrews in Tikkurila demonstrate the probable pathway of Pb into the shrew. Earthworms are regarded as the main heavy metal source for shrews (Ma, 1989; Ma et al., 1991). Different Species of Shrews. — Dissimilar heavy metal concentrations in S. araneus and S. minutus perhaps reflect differences in the diets between the two species. Earthworms accumulating high concentrations of heavy metals are more important food items for S. araneus than for S. minutus (Pemetta, 1976; Butterfield et al., 1981; Bauerova, 1984). Furthermore, the kidneys of one S. minutus specimen usually were too small for an adequate analysis of heavy metals, making the pooling of samples of the same age group necessary. This increased the number of individuals needed in the study. Because most heavy metal concentrations were lower in the kidneys of S. minutus than in S. araneus, the latter species is a more appropriate indicator of heavy metal pollution. Neomys fodiens may be an effective indicator of Hg, as indicated by high concentrations in the Espoo sample (Table 1). A similar result was obtained by Bergbom (1987) in south Finland. However, N. fodiens is a habitat specialist and is usually too rare for adequate sample size. We conclude that S. araneus is a better bioindicator for heavy metal pollution of terrestrial ecosystems than other species of small mammals in south Finland. It is also abundant in areas with intensive human impact. Accumulation of Heavy Metals in S. araneus. — The dependence of heavy metal accumulation on age in S. araneus accords well with earlier studies. Accumulation in several species of small mammals is most evident for Cd (McKinnon et j al., 1976; Johnson et al., 1978; Way and Schroeder, 1982; Wlostowski, 1987; Hunter et al., 1989). The increased level of Zn in older individuals may be connected with increased Cd in ! them (see above). In our study Pb concentrations did not increase according to the age of the individual, as is also the case in studies by McKinnon et al. (1976), Anderson et al. (1981), and Cloutier et al. (1986). Contradictory results have j been presented by Way and Schroeder (1982) and Scanlon et al. (1983). Because concentrations of most heavy metals in adult shrews were at least as high as in juveniles, they may be more | appropriate indicators for low heavy metal concentrations in the | environment than juveniles. On the other hand, juvenile shrews have smaller home ranges and usually reach higher densities ‘ than adults (Croin Michielsen, 1966), both characteristics being ' useful for an indicator animal of heavy metal pollution. For comparisons of heavy metal concentrations on several sites, the | age groups must be treated separately. :> Shrews are short-lived animals with high potential of ‘ reproduction. Their relatively small home ranges (compared to those of larger mammals or birds) exposes them to local environmental conditions. In the present study, shrews indicated ! changes in heavy metal concentrations occurring at relatively short distances, as was found in the Tikkurila Pb area (Fig. 1). In earlier studies a similar decrease has been observed in Pb ■. concentrations of soil and mushrooms there (Ervio and Lakanen, 1973; Liukkonen-Lilja et al., 1983). It is noteworthy that the decrease in Pb accumulation with increasing distance was demonstrated also in shrews, more mobile animals than earthworms. 1994 PANKAKOSKI ET AL. — Shrews as Indicators of Heavy Metals 141 Heavy Metal Accumulation and Soil pH Several heavy metals move into organisms more readily from acidified soils than from soils with higher pH, as shown in plants (Andersson and Nilson, 1974; Nuorteva et al., 1986), in earthworms (Ma, 1982; Ma et al., 1983; Beyer et al., 1987; Morgan and Morgan, 1988) and in small mammals (Roberts et al., 1978; Ma, 1989). This effect is especially evident in Pb, where concentrations are higher in organisms living in places with low soil pH and/or low concentration of Ca ions (Johnson et al., 1978; Roberts et al., 1978; Andersen, 1979; Beyer et al, 1987; Scheuhammer, 1991). In Koverhar both high pH and high amount of Ca-ions in the soil may be responsible for the low rate of accumulation of several heavy metals in S. araneus, although these metals show high concentrations in the soil there. In the control area at the distance of 4 km, where the soil pH is “normal” (i.e., much lower) and the soil is less polluted, heavy metal accumulation (especially Pb) in shrews was more intensive. In the other study areas (Tikkurila, Espoo, and Haijavalta area), where the differences in soil pH between the polluted and control sites were smaller, the effect of pH on heavy metal accumulation could not be demonstrated. Developmental Stability In the present study developmental stability was reduced in those shrew populations which showed highest heavy metal concentrations. This suggests that their toxic effects may stress the shrew populations. It is possible that Pb (high concentrations both in Tikkurila and Koverhar) is especially important in this respect. Other differences in habitat quality may also affect developmental stability. TTie high level of developmental stability in the shrews from control area of Espoo perhaps partly indicates the benefits of a forest habitat for S. araneus. However, the final cause-and-effect relationship between heavy metals and developmental stability cannot be proved by field studies alone. Although the concentrations of heavy metals in natural small mammal populations are relatively easy to analyze, it is difficult to indicate changes in viability or breeding success caused by their toxic effects. If the concentrations are too low to increase mortality or to cause histopathological tissue alterations, their negative effects cannot be demonstrated by traditional methods. By analyzing developmental stability it seems possible to assess the biological effect of toxic compounds on small mammal populations also in moderately polluted areas (Pankakoski et al., 1992). Acknowledgments We thank P. Karvinen, K. Koivisto, T. Landen-Leino, M. Mononen, L. Rahunen, and H. Virtanen for assistance in the field and laboratory. We are also grateful to V. Haukisalmi and to two unknown referees for useful comments on the manuscript. Grants from the Academy of Finland to EP and IK and from the Maj and Tor Nessling Foundation to EP are gratefully acknowledged. Literature Cited Andersen, C. 1979. Cadmium, lead and calcium, number and biomass in earthworms from sewage sludge treated soils. Pedobiologia, 19:309-319. Anderson, K. L., F. A. Servello, and P. F. Scanlon. 1981. Heavy metal concentrations of tissues of pine voles from four Virginia orchards. Virginia Journal of Science, 32:87. Andersson, A., and K. O. Nilsson. 1974. Influence of lime and soil pH on Cd availability to plants. Ambio, 3:198-200. Andrews, S. M., M. S. Johnson, and J. A. Cooke. 1984. Cadmium in small mammals from grassland established on metalliferous mine waste. Environmental Pollution (Series A), 33:153-162. 1989. Distribution of trace element pollutants in a contaminated grassland ecosystem established on metalliferous fluorspar tailings 1: Lead. Environmental Pollution, 58:73-85. BaUEROVA, Z. 1984. The food eaten by Sorex araneus and Sorex minutus in a spruce monoculture. Folia Zoologica, 33:125-132. Berg BOM, K. 1987. Forekomsten av tungmetaller i smadaggdjur. Unpublished M. Sc. Thesis, Department of Environmental Science, University of Helsinki, 69 pp. (in Swedish). Beyer, W. N., O. H. Pattee, L. Sileo, D. J. Hoffman, and B. M. Mulhern. 1985. Metal contamination in wildlife living near two zinc smelters. Environmental Pollution (Series A), 38:63-86. Beyer, W. N., G. Hensler, and J. Moore. 1987. Relation of pH and other soil variables to concentrations of lead, copper, zinc, cadmium and selenium. Pedobiologia, 30:167-172. Bremner, I. 1974. Heavy metal toxicities. Quarterly Reviews of Biophysics, 7:75-124. Butterfield, J., J. C. Coulson, and S. Wanless. 1981. Studies on the distribution, food, breeding biology and relative abundance of the pygmy and common shrews (Sorex minutus and S. araneus) in upland areas of northern England. Journal of Zoology (London), 195:169-180. Cloutier, N. R., F. V. Clulow, T. P. Lim, and N. Dave. 1986. Metal (Cu, Ni, Fe, Co, Zn, Pb) and Ra-226 levels in tissues of meadow voles Microtus pennsylvanicus living on nickel and uranium mine tailings in Ontario, Canada: Site, sex, age and season effects with calculation of average skeletal radiation dose. Environmental Pollution (Series A), 41:295-314. Croin Michielsen, N. 1966. Intraspecific and interspecific competition in the shrews Sorex araneus L. and S. minutus L. Archives Neerlandaises de Zoologie, 17:73-1740. Crowcroft, P. 1957. The life of the shrew. Max Reinhardt, London, 166 pp. Czarnowska, K., and B. Gworek. 1980. Heavy metals in the liver of Apodemus agrarius from urban areas. Bulletin de I’Academie Polonaise des Sciences, Serie des sciences biologiques. Cl. II, 27:505-510. Ervio, R., and E. Lakanen. 1973. Maan lyijysaastuminen sulattamon ymparistossa Tikkurilassa. Annales Agriculturae Fenniae, 12:200-206 (in Finnish with English summary). Forsyth, D. J., and T. J. Peterle. 1984. Species and age differences in accumulation of ^^Cl-DDT by voles and shrews in the field. Environmental Pollution (Series A), 33:327-340. Fritze, H. 1991. Forest soil microbial response to emissions from an iron and steel works. Soil Biology and Biochemistry, 23:151-155. Getz, L. L., L. Verner, and M. Prather. 1977. Lead concentrations in small mammals living near highways. Environmental Pollution, 13: 151-157. Goldsmith, C. D., Jr., and P. F. Scanlon. 1977. Lead levels in small mammals and selected invertebrates associated with highways 142 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 of different traffic densities. Bulletin of Environmental Contamination and Toxicology, 17:311-316. Coyer, R. A., D. L. Leonard, J. F. Moore, B. Rhyne, and M. R. Krigman. 1970. Lead dosage and the role of the intranuclear inclusion body. Archives of Environmental Health, 20:705-711. Hanski, 1. 1984. Food consumption, assimilation and metabolic rate in six species of shrews from Finland (Sorex and Neomys). Annales Zoologici Fennici, 21:157-165. HEGSTROM, L. J., and S. D. West. 1989. Heavy metal accumulation in small mammals following sewage-sludge application to forests. Journal of Environmental Quality, 18:345-349. Helminen, O., M. Laakso, and R. Holmberg. 1986. Hangon ilmansuojelun perusselvitys. TutkimusjuUcaisu 46, Lansi- Uudenmaan vesiensuojeluyhdistys r.y., 59 pp. (in Finnish). Hunter, B. A., and M. S. Johnson. 1982. Food chain relationships of copper and cadmium in contaminated grassland ecosystems. Oikos, 38:108-117. Hunter, B. A., M. S. Johnson, and D. J. Thompson. 1987. Ecotoxicology of copper and cadmium in a contaminated grassland ecosystem. 111. Small mammals. The Journal of Applied Ecology, 24:601-614. 1989. Ecotoxicology of copper and cadmium in a contaminated grassland ecosystem. IV. Tissue distribution and age accumulation in small mammals. The Journal of Applied Ecology 26:89-99. Hynninen, V. 1986. Monitoring of airborne metal pollution with moss bags near an industrial source at Haijavalta, southwest Finland. Annales Botanic! Fennici, 23:83-90. Hynninen, V., and M. Lodenius. 1986. Mercury pollution near an industrial source in South West Finland. Bulletin of Environmental Contamination and Toxicology, 36:294-298. Ireland, M. P. 1983. Heavy metal uptake and tissue distribution in earthworms. Pp. 247-265, in Earthworm Ecology (J. E. Satchell, ed.). Chapman and Hall, London, 495 pp. Jagoe, C. H., and T. A. Haines. 1985. Fluctuating asymmetry in fishes inhabiting acidified and unacidified lakes. Canadian Journal of Zoology, 63:130-138. Johnson, M. S., R. D. Roberts, M. Hutton, and M. J. Inskip. 1978. Distribution of lead, zinc and cadmium in small mammals from polluted environments. Oikos, 30:153-159. Kinnunen, j., H. Veuola, M.-T. Seppanen, and U.-M. Holopainen. 1985. Lyijyn ja kadmiumin leviaminen Espoossa. Espoon ymparistonsuojelulautakunnan julkaisu 6/1985, 45 pp. (in Finnish). Leary, R. F., and F. W. Allendorf. 1989. Fluctuating asymmetry as an indicator of stress: Implications for conservation biology. Trends in Ecology and Evolution, 4:214-217. Liukkonen-Liua, H., T. Kuusi, K. Laaksovirta, M. Lodenius, AND S. PlEPPONEN. 1983. The effect of lead processing works on the lead, cadmium and mercury contents of fungi. Zeitschrift fiir Lebensmittel-Untersuchung und -Forschung, 176:120-123. Ma, W. 1982. The influence of soil properties and worm-related factors on the concentration of heavy metals in earthworms. Pcdobiologia, 24:109-119. 1989. Effect of soil pollution with metallic lead pellets on lead bioaccumulation and organ/body weight alterations in small mammals. Archives of Environmental Contamination and Toxicology, 18:617-622. Ma, W., T. Edelman, I. vanBeersum, andT. Jans. 1983. Uptake of cadmium, zinc, lead and copper by earthwonns near a zinc smelting complex: Influence of soil pH and organic matter. Bulletin of Environmental Contamination and Toxicology, 30:424-427. Ma, W., W. Denneman, and J. Faber. 1991. Hazardous exposure of ground-living small mammals to cadmium and lead in contaminated terrestrial ecosystems. Archives of Environmental Contamination and Toxicology, 20:266-270. Makinen, a. 1977. Moss- and peat-bags in air pollution monitoring. Suo, 28:79-88. Marttinen, M., and a. Edgren. 1984. Viljelyalueiden raskasmetallikartoitus Espoon kaupungin alueella. Espoon ympari- stonsuojelulautakunnan julkaisu 2/1984, 30 pp. (in Finnish). McKinnon, J. G., G. L. Hoff, W. J. Bigler, and E. C. Prather. 1976. Heavy metal concentrations in kidneys of urban gray squirrels. Journal of Wildlife Diseases, 12: 367-371. Morgan, J. E., and A. J. Morgan. 1988. Earthworms as biological monitors of cadmium, copper, lead and zinc in metalliferous soils. Environmental Pollution, 54:23-138. Nicholson, J. K., M. D. Kendall, and D. Osborn. 1983. Cadmium and mercury nephrotoxicity. Nature, 304:633-635. Nuorteva, P., S. Autio, j. Lehtonen, A. Lepisto, S. Ojala, a. j Seppanen, E. Tulisalo, P. Veide, J. Vupuri, and R. Willamo. i 1986. Levels of iron, aluminum, zinc, cadmium and mercury in j plants growing in the surroundings of an acidified and a non- acidified lake in Espoo, Finland. Annales Botanici Fennici, 23:333-340. Pankakoski, E. 1979. The cone trap — A useful tool for index trapping of small mammals. Annales Zoologici Fennici, I 16:144-150. I 1989. Variation in the tooth wear of the shrews Sorex araneus and S. minutus. Annales Zoologici Fennici, 26:445-457. Pankakoski, E., and 1. Hanski. 1989. Metrical and non-metrical skull traits of the common shrew Sorex araneus and their use in ; population studies. Annales Zoologici Fennici, 26:433-444. Pankakoski, E., R. A. Vaisanen, and K. Nurmi. 1987. Variability of muskrat skulls: Measurement error, environmental modification and size allometry. Systematic Zoology, 36:35-51. j Pankakoski, E., I. Koivisto, and H. Hyvarinen. 1992. Reduced j developmental stability as an indicator of heavy metal pollution in j the common shrew Sorex araneus. Acta Zoologica Fennica, 191:137-144. PernettA, j. C. 1976. Diets of shrews Sorex araneus L. and Sorex minutus L. in Wytham grassland. The Journal of Animal Ecology, 45:899-912. Quarles, H. D., Ill, R. B. Hanawalt, and W. E. Odum. 1974. Lead in small mammals, plants and soil at varying distances from a highway. The Journal of Applied Eeology, 11:937-949. Roberts, R. D., and M. S. Johnson. 1978. Dispersal of heavy metals from abandoned mine workings and their transference through terrestrial food chains. Environmental Pollution, 16:293-310. ' Roberts, R. D., M. S. Johnson, and M. Hutton. 1978. Lead contamination of small mammals from abandoned metalliferous i mines. Environmental Pollution, 15:61-69. Ruhling, a., L. Rasmussen, K. Pilegaard, A. Makinen, and E. i Steinnes. 1987. Survey of Atmospheric Heavy Metal Deposition I* in the Nordic Countries in 1985 — Monitored by Moss Analyses. , Nord 1987, 21 . The Nordic Council of Ministers, Goteborg, 44 pp. Scanlon, P. F., R. J. Kendall, R. L. Lochmiller II, and R. L. | Kirkpatrick. 1983. Lead concentrations in pine voles from two Virginia orchards. Environmental Pollution (Series B), 6:157-160. i Scheuhammer, a. M. 1991. Effects of acidification on the availability of toxic metals and calcium to wild birds and mammals. Environmental Pollution, 71:329-375. Soule, M. E. 1982. Allomeric variation. 1. The theory and some consequences. The American Naturalist, 120:751-764. Valentine, D. W., and M. Soule. 1973. Effect of p,p’-DDT on developmental stability of pectoral fin rays in the grunion. 1994 PANKAKOSKI ET AL. — Shrews as Indicators of Heavy Metals 143 Leuresthes tenuis. Fishery Bulletin, 71:921-926. Valentine, D. W., M. E. Soule, and P. Samollow. 1973. Asymmetry analysis in fishes: A possible statistical indicator of environmental stress. Fishery Bulletin, 71:357-370. Way, C. a., and G. D. SchROEDER. 1982. Accumulation of lead and cadmium in wild populations of the commensal rat, Rattus norvegicus. Archives of Environmental Contamination and Toxicology, 11:407-417. Wlostowski, T. 1987. Heavy metals in the liver of Clethrionomys glareolus (Schreber, 1780) and Apodemus agrarius (Pallas, 1771) from forests contaminated with coal-industry fumes. Ekologia Polska, 35:115-129. Yablokov, a. V. 1986. Population Biology. Progress and Problems of Studies on Natural Populations. Advances in Science and Technology in the USSR. Biology Series. MIR Publishers, Moscow, 303 pp. Zakharov, V. M. 1984: Analysis of homeorhesis in ontogenetic, population and evolutionary aspects. Pp. 509-519, in Thermodynamics and Regulation of Biological Processes (1. Lamprechtand A. 1. Zotin, eds.), Walter de Gruyter, Berlin., 573 pp. 1989. Future prospects for population phenogenetics. Soviet Scientific Reviews, Section F, 4:1-79. Zakharov, V. M., M. Olsson, A. V. Yablokov, and A. G. Esipenko. 1989. Does environmental pollution affect the developmental stability of the Baltic grey seal (Halichoerus grypus)? F*p. 96-108, in Influence of Human Activities on the Baltic Ecosystem (A. V. Yablokov and M. Olsson, eds.). Proceedings of the Soviet-Swedish Symposium. Leningrad Gidrometeoizdat, 149 pp. Table 1. — Comparison of heavy metal concentrations (mg/kg) in the livers of four small mammal species from Tikkurila (lead smeltery area) and Espoo (control area) populations. All age groups are combined, atuJ median and maximum (highest concentration in the sample; in parentheses) values are given. In Tikkurila, the animals were trapped within a radius of 500 m from the lead smeltery, n, number of individuals. Statistical significance in all tables and figures is indicated by the following symbols: o, P < 0.1; *, P < 0. 05; **, P < 0.01; ***, P < 0. 001. The first column of asterisks (a) after the concentrations of Neomys fodiens, Clethrionomys glareolus, and Microtus agrestis indicates significant difference from those of Sorex araneus,* the second column of asterisks (b) indicates significant difference ofC. glareolus and M. agrestis from those o/’N. fodiens (Mann- Whitney U-tests). Because the age structure in samples of S. araneus differed between Tikkurila and Espoo, see Table 6 for comparison of concentrations between these localities. S. araneus N. fodiens C. glareolus M. agrestis Tikkurila Cd 2.90 (29.20) 0.74 (13.42) (a) Pb 15.95 (233.4) 12.80 (45.62) Zn 86.44 (130.1) 86.42 (131.3) Cu 23.41 (45.79) 16.02 (37.20) ** Hg 0.10 (0.29) 0.13 (0.20) n 24 13 Espoo Cd 6.77 (18.00) 6.26 (6.97) Pb 4.15 (7.55) 2.67 (3.03) Zn 102.21 (130.1) 78.90 (85.22) Cu 24.13 (30.30) 10.89 (11.30) Hg 0.13 (0.29) 4.47 (5.94) n 28 (a) (b) (a) (b) 0.29 (1.71) ** 0.09 (0.69) * 3.64 (9.43) ** 1.69 (6.78) *** 95.71 (107.0) o 89.86 1 (100.0) 19.54 (39.53) o 18.58 I (32.67) 0.03 (0.07) + *** 0.00 (0.09) ** 8 11 0.00 (0.87) 0.00 (0.00) 0.00 (2.73) *** 0.00 (1.22) 109.33 (132.9) 105.28 (116.3) 17.64 (28.51) 13.67 (18.63) 0.02 (0.07) *** 0.06 (0.07) 3 22 3 144 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 Table 2. — Heavy metal concentrations (mg/kg) in two shrew species, S. araneus and S. minutus, given as median and (maximum) values. Only juveniles trapped near the works in Harjavalta were included. If the difference between the species was significant (Mann-Whitney U-tests, P, asterisks) the greater value is in italics. Sample sizes: S. araneus n — 24 (except in zinc, in which liver n = 50, kidney n = 33); S. minutus liver n = 25, kidney n = 5 (pooled samples). Heavy Metal s. araneus s. minutus P Liver Cd 5.55 (23.04) 5.51 (10.43) Cr 0.00 (4.79) 0.00 (13.40) Pb 1.41 (5.14) 0.00 (17.61) o Zn 76.94 (310.1) 82.98 (176.5) Cu 20.76 (30.27) 21.43 (44.44) Ni 0.26 (7.24) 3.39 (68.08) Kidneys Cd 7.49 (24.83) 0.00 (2.86) *** Cr 10.00 (31.30) 0.99 (2.37) *** Pb 13.76 (26.67) 9.74 (13.42) * Zn 71.43 (329.9) 54.63 (64.29) o Cu 38.11 (67.74) 19.43 (25.71) *** Ni 1.30 (23.00) 0.00 (0.71) ** Table 3. — Heavy metal concentrations in the two age groups ofS. araneus from Harjavalta (medians, mg/kg). If the difference between the species was significant (Mann-Whitney U-tests, P, asterisks), the greater value is in italics. Heavy Metal Juveniles Adults Liver Cd 5.55 29.09 *** Cr 0.00 0.23 Pb 1.41 1.79 Zn 71.25 78.81 Cu 20.76 31.81 *** Ni 0.26 0.64 Hg 0.06 0.20 Kidneys Cd 7.49 46.66 Cr 10.00 8.53 Pb 13. 76 7.53 Zn 67.86 90.59 Cu 38.11 40.84 Ni 1.30 2.92 n 24 18 o 1994 PANKAKOSKI ET AL.— Shrews as Indicators of Heavy Metals 145 Table 4. — Lead concentrations (mg/kg) in soil and in the earthworm Aporrectodea caliginosa in Tikkurila (lead smeltery area) and Pornainen (control area). Median values and ranges (in parentheses) of 16 plots in Tikkurila (in A. caliginosa only ten plots, as earthworms were absent in the six plots closest to the smeltery) and eight plots in Pornainen are presented. P = significance in Mann-Whitney tests. Soil A. caliginosa Tikkurila Pornainen P 1781 (411-79960) 18.2 (13-542) 925.5 (554-2200) 50.0 (0-139) Table 5. — Soil pH (mean, standard deviation, and sample size) at the study areas. The pH of the trapping points that yielded the greatest catches of shrews is presented in Espoo and Harjavalta (the latter consists of three points, ten analyses in each). Locality Distance from the Source of Pollution (km) Mean pH ±SD n Tikkurila <0.1 5.29 + 0.107 10 Control in Espoo 20 4.21 ±0.586 10 Haijavalta 1-2 5.36 ±0.360 30 Control 10 10 4.72 ±0.418 10 Control 20 20 4.18 ±0.172 10 Koverhar 0.5 7.60 ±0.284 10 Koverhar 1.5 5.54 ±0.777 10 Control 4 4.59 ±0.293 10 Table 6.— Heavy metal concentrations (mg /kg) in the liver of juvenile S. araneus around a lead smeltery (Tikkurila) and a rural control area (Espoo), given as median and (maximum) values. The greater concentration is in italics, if the difference was statistically significant (Mann-Whitney U-tests, P, asterisks). Metal Tikkurila Espoo P Cd 2.64 (7.1) 4.97 (18.0) * Pb 15.55 (233.4) 4.66 (7.6) Zn 85.08 (128.7) 102.57 (130.1) Cu 23.15 (36.8) 23.50 (30.3) Hg 0.10 (0.3) 0.10 (0.3) n 22 18 146 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 Table 7. — Heavy metal concentrations (mg/kg) in the liver and kidneys of juvenile S. araneus in Koverhar and its control area at a distance of 4 km, given as median and (maximum) values. If the difference is significant (Mann-Whitney U-tests, P, asterisks), the greater value is in italics. Heavy Metal Koverhar Control P Liver Cd 1.47 (4.23) 1.10 (3.02) Cr 0.00 (1.80) 0.00 (0.55) * Pb 1.57 (4.68) 3.92 (5.25) Zn 89.82 (97.45) 79.89 (131.3) * Cu 19.01 (23.74) 20.27 (32.93) 0 Ni 0.00 (0.00) 0.00 (0.00) Kidneys Cd 2.07 (6.18) 0.00 (7.50) ** Cr 1.82 (8.57) 0.00 (4.29) Pb 0.00 (8.82) 9.31 (30.00) *** Zn 109.1 (134.5) 144.0 (225.0) *** Cu 38.71 (57.69) 50.09 (120.0) *** Ni 0.00 (4.29) 0.00 (37.50) * n 21 16 Pb log mg/kg Distance from lead smeltery (m) Fig. 1. — The dependence of lead concentration in the liver of S. araneus on the trapping distance from the lead smeltery in Tikkurila. a, four individuals with concentration under the determining level (ca 0.05 mg/kg Pb); r^, Spearman rank correlation coefficient. Notice that the vertical axis is in log scale. 1994 PANKAKOSKI ET AL. — Shrews as Indicators of Heavy Metals 147 A B 79960 Fig. 2. — The effect of distance from the smeltery on soil and earthworm {Aporrectodea caliginosa) lead concentrations (A) and on the total number of earthworms (all species; B). Tikkurila, May-June 1990: from 16 plots measuring 25 X 50 cm. The lead concentrations of earthworms represent median values of 1-9 samples analyzed in each plot. Each sample comprised of 1-19 individuals of A. caliginosa. This species was present only in ten plots at the greatest distance from the smeltery (A). Spearman rank correlation coefficients with distance: soil Pb —0.80** (16 plots), A. caliginosa Pb —0.47 (ten plots), total number of earthworms r^= +0.86*** (16 plots). In A. caliginosa the Pb concentration value (705 mg/kg) at the nearest distance (35 m) was based on only one small juvenile individual; excluding this individual the Spearman rank correlation coefficient with distance is r,= —0.78* (nine plots). Fig. 3.— Relative concentrations of heavy metals in tissues of S. araneus (only juveniles) at three distances from Haijavalta metal works. To make different metals (with different scales) comparable, the concentrations are presented in percentages; 1(X)% is the sum of concentrations (median values) in the three sites: at the Haijavalta works (0 km), control area at 10 km, and control area at 20 km. If the concentrations are equal in each trapping area, the relative value is 33.3 % in each. The asterisks indicate sigmficant differences among the three areas (Kruskal-Wallis tests). Number of individuals: Haijavalta works, n — 24; control 10 km, n - 19; control 20 km, n — 20. 148 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 SOIL % 100 75 50 25 0 Cr Pb Zn Cu Ni LIVER % Cr Pb Zn Cu Ni KIDNEY QQ * *** *** *** * 75 50 25 0 Cr Pb Zn Cu Ni Fig. 4. — Comparison of relative concentrations of heavy metals in soil and tissues of S. araneus (only juveniles) between Koverhar works {n = 21) and its control area {n = 16). To make different metals (with different scales) comparable, the concentrations are presented in percentages; 100% is the sum of concentrations at Koverhar works and at its control area. If the concentrations are equal on both sites, the black (Koverhar) and white (control) areas of the bar are equally high (50%). The asterisks indicate significant differences between the two areas (Mann-Whitney tests). The relative values are based on mean values of the concentrations, as several of the median values (shown in Table 7) were 0. 1994 PANKAKOSKI ET AL.— Shrews as Indicators of Heavy Metals 149 A Tikkurila Espoo C as (U E (-> o C 2 lU Foramina FA PV Mandibles B c: ea (U s Si o c 2 (U :s m Harjavalta f%| Control 10 I I Control 20 FA NA PV NP Foramina FA PV Mandibles Fig. 5. — Comparison of developmental stability in the morphometrical traits of Sorex araneus between the polluted and control areas. High columns indicate low developmental stability. A, Tikkurila lead area and control in Espoo; B, Haijavalta and two control areas at 10 and 20 km distances; C, Koverhar and control. Developmental stability was measured by fluctuating asymmetry (FA), number of asymmetric traits per individual (NA), number of phenodeviants per individual (NP) and by total phenotype variability (PV). For the details, see text. The values on the y axis are medians (FA, PV; mandibular PV median X 0. 1) or means (NA, NP; mean X 0. 1). The asterisks indicate significant differences between the areas (FA, PV: Friedman tests; NA, NP: t-tests or ANOVAs). Number of individuals: Tikkurila n = 22, Espoo n = 18; Haijavalta n = 50, control 10 km n = 19, control 20 km n = 20; Koverhar n = 46, control n = 53. i-r-i. ? ^ ^ ' ;' # ■ ■ 1 f-4 ^U'-. v’C Ai? ■:■■ i \ •''' "J'-IL' ',, •• ' .'Ur '■ a. ' " •ar ■'■'®- ..j o 58 Jlrl :V|. i£r ii,k» iA nwnfiqcnfA.'-- £,.ti .immm . ••■■' Mit- ;i‘uy i\j '■I* ■* A i . : t ■/» ' ■#{??»«<;■■ •'■ r'l < ;,!• '■•: .:v „'(.' '■ t \) (> . (' i •vv -ii.>’,i-«.i- ■ni., (I I taD^i(p^ i;./«A,V' , • If V.V^>''\- .H, .-I, ,i(! i;, -'Iwf.i'' ’ •>(i ?.'■' 'i * ■ ■ ■ l' >. A. '-\ ^ ‘ ! 1 ,■ • ■ ‘' ./, - rV, ., .■ 'V -1. V, r^,.j ' '_ •.■■■•' ri’jvi '' - •' til 11). .- ti; y. f\ ■■Mmvi;/. a ,1 , i.:-i'“itni» ■■ ■'"'.VtO? '* , *{.1/ 1 ^1' ' 'j* \ ti f(iS5P?*^ \'<'R,iiSUl*f ,4^*'-lV ■■<<*' I" VV^J >0. V Hi' i.’. ' • • ! * . r-: ^ , Mi ■'■» ■' i‘'4\ ui.-i'^'icj -R*} t ■uia-iuuf u* ^ iS‘-»"-'U'f;AiinC&.A,W^ li' -ii' 'JVl’^~»,47^iirrf 4Aij i\x!flt' .'■: >0- ••■••.■• • •'1 .»‘jci»J' :h '' ' ' ^.. .'■' Pi(:«tair S.’foi’^S I,- . "i( j>,ir. ^ 1!) 1 "V iiit^. r .1 r.; 1 , , H,,", , ,^s;.:V..^(p ^^.r'!/> ', . ., ■ ^ K‘: ').! •<'■ . .>,ti.‘.tV’.% 1 I' ■ ■(. ■ 1, I ',' 'i«U- I,-.. ,.• y>, (< t .u, I" vu ■■ ' ■» ■ ■* 1 .' (^.l>^ VARIATION IN BRAIN MORPHOLOGY OF THE COMMON SHREW VLADIMIR A. YASKIN Department of Vertebrate Zoology and General Ecology, Faculty of Biology, Moscow State University, Moscow 119899, Russia Abstract Common shrews (Sorex araneus) were trapped on study grids located in mixed deciduous forests of western Siberia and 60 km west of Moscow. The patterns of seasonal changes in masses of body, brain, and brain regions were similar in both populations, but there were some differences in absolute values. The regions of the brain differed in their responses to winter and drought conditions by showing greater variation in the phylogenetically recent regions (such as neocortex) than in the phylogenetically ancient regions of the brain, such as olfactory bulbs and myelencephalon. Water content decrease was not the only reason leading to brain size decrease over winter. Dry mass of the brain declined in this period, primarily because forebrain mass declined by 21 % (neocortex especially, 28%). During a period of mass gain in spring the most considerable increase was observed in the hippocampus ( + 33% of wet mass, +26% of dry mass). The most noticeable mass changes occurred in this region and only as reproductive activity began. My research indicates broad macromorphological variability occurs in the brain of the common shrew. This variation is associated with age, sex, and different environmental factors such as geographic, seasonal, and extreme climatic (summer drought) parameters. The possibility is explored that seasonal (and other) variation in brain mass (previously reported by this author for voles) may represent a more generalized mammalian pattern of adaptive mechanism of winter-active small mammals employed to conserve energy during harsh climatic conditions. Introduction The study of seasonal morphological changes of nonhibemating small mammals at intermediate and high latitudes is important in understanding the adaptive strategies and environmental relations of these animals. Dehnel (1949) was the first to find seasonal changes of the skull height in Sorex araneus. Seasonal changes found in the volume of the braincase (Pucek, 1955) suggested that analogous variations should take place in the volume of the brain. This was demonstrated for mass and volume of the brain in Sorex minutus (Gabon, 1956) and in S. araneus (Bielak and Pucek, 1960). The data on seasonal changes of brain size were surprising because of the widely adopted viewpoint of the stability and sufficient protection of the brain from extreme environmental factors. The complex of seasonal changes in shrews, and of skull and brain changes especially, has been considered by investigators as a specific adaptation to winter conditions of this group of animals only. But detailed investigations of brain variability in voles have shown the existence both of seasonal changes in brain size (similar to that of shrews) and of significant alterations in brain structure and proportions (Yaskin, 1980, 1984, 1989). Histological changes in the skulls of rodents similar to those of shrews were also recorded (Quay, 1984). Seasonal changes in brain morphology, accompanied by the changes in its size and cranial capacity, are considered to be a general pattern for a large group of Palearctic and Nearctic species of small mammals (Yaskin, 1984). This report deals with original data on seasonal, annual, and geographic variability in brain size and brain macromorphology of the common shrew, Sorex araneus. Materials and Methods Two populations of Sorex araneus were sampled circumannually. The study areas were situated in the floodplain biotope of the Pyshma River (western Siberia) and in mixed forest near the Moskwa River 60 km west from Moscow. Perhaps the most significant climatic characteristic of the western Siberian site was the high variability between the seasons. The locality near Moscow features a milder climate with smaller seasonal temperature gradients. Animals were obtained through frequent inspection of live traps and metal cylinders. All captured shrews were measured, weighed, and then dissected as soon as possible after death. The braincase height was taken to the nearest 0. 1 mm using a dial caliper. The brain mass was determined after its separation from the spinal cord at the posterior attachment of the pyramids. After being weighed to the nearest 1 mg, the brain was immersed in 10% formalin in which it was kept at room temperature for a period of more than one month. The method of determination of the mass of brain regions (Latimer, 1950; Dmitrieva, 1969) was used for quantitative investigation of peculiarities of brain structure. The brain and principal brain regions of 217 (Siberian locality) and 53 (Moscow region) individuals of common shrews were weighed. To convert brain regions after formaldehyde fixation into their respective fresh masses, a conversion index was calculated (k = fresh brain mass/fixed brain mass). The index diverged in animals of different ages, so it was calculated for each individual separately. Weighed wet brain parts were transferred to weighing glasses and evaporated in an oven at a temperature of 60 °C for 48 h. This proved to be sufficient time to dry the tissue to a constant mass. After drying, the brain parts were weighed to an accuracy of 0.2 mg. The absolute water mass of the fresh brain was determined as the difference of the fresh brain and the sum of dry masses of brain regions. Water content was calculated as the percentage of absolute water mass to fresh brain mass. The significance of the differences was assessed by the Student’s t-test of the differences in mean values for two independent groups. Results The observation of winter reduction of skeleton size in 155 156 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 f shrews (Dehnel, 1949) has been confirmed for different species and populations, although some geographic peculiarities of the processes have been found (Z. Pucek, 1965; Hyvarinen and Heikura, 1971; Mock, 1994). According to my data the mean height of the braincase in young common shrews from the Moscow region was 6.01 + 0.048 mm in summer. The reduction in the braincase during the autumn and winter was expressed by a decrease in its height in February (X = 4.94 ± 0.040 mm) of 17.8% (P < 0.001) in relation to summer. During the spring “jump in growth” mean height of braincase increased by 8.5% (P < 0.001), and in overwintered adult shrews was 5.36 + 0.056 mm. A comparison of these data with some other populations has indicated that the population under study was characterized by a great winter regression in the height of the braincase. Figure 1 shows the changes in the mean body and brain masses for both sexes in different seasons during the life cycle of common shrews from western Siberia. These curves agree with data on the Bialowieza population (Poland) obtained previously by Z. Pucek (1965). The reduction in the brain mass during the winter was expressed by a decrease in its mass in February (Fig. 1) of 21.2% (P < 0.001) relative to July. From October to February the brain mass decreased by 17% (P < 0.001). The corresponding values obtained by Z. Pucek (1965) in Bialowieza were 23.6% and 15.0%. The mass of the brain increased in the spring; however, the summer brain mass in overwintered adults was almost 16% (P < 0.001) lower relative to that of young from the same season. A similar value (14.6%) was given by Z. Pucek (1965). As can be seen from Fig. 1, the brain tended to lose mass before the decrease in body mass. In the period from June to September, brain mass decreased by 11% (P < 0.001). In contrast, the body mass loss began in September. The relative brain mass (Fig. 1) maintained approximately the same level in young shrews from summer and winter. The fundamental changes in the relative brain mass took place from March, since the brain mass gain within the spring-to-summer period was smaller than the corresponding spring-to-summer body mass increase. A comparison of seasonal body and brain mass changes in the two populations under study (Table 1) shows that animals from the Moscow region were characterized by more pronounced winter regression in brain mass than were those from western Siberia. No differences in body mass were found. There were no regional differences in brain mass of young shrews from summer; however, the winter brain masses in young and overwintered adults from the Moscow region were statistically lower (by 6. 3-6. 8%; P < 0.01) than those of western Siberian individuals of corresponding age. The water content of an organism tends to decrease with age. There is also information on the lowering water content of tissues in small mammals during a winter period (M. Pucek, 1965; Sawicka-Kapusta, 1974). Seasonal variations in the brain mass of shrews are due to changes in water content and dry matter (M. Pucek, 1965). My data indicated that both dry brain mass and water content decreased during autumn and winter. In shrews of six to ten months of age in March, the mean dry mass of the brain was 13.0% (P < 0.001) lower than in animals one to three months old in summer. The absolute water mass was 25.4% lower {P < 0.001) and water content was 2.36% lower (P < 0.01). During the period of rapid growth in i spring, the dry brain mass in shrews aged seven to 11 months increased by 10.8% (P < 0.001) (Fig. 3), and the water mass increased by 16.2% {P < 0.001) during this period. From the winter to summer, the brain water content increased by 0.15 fo (P < 0.05). The present results indicated that the decrease in | the brain mass of common shrews during the winter was caused • both by water and dry matter loss. Water loss was more pronounced, but even more importantly, the dry matter, whose ! mass was absolutely less in winter than in summer, was also : subject to seasonal variation. Although many investigations have been dedicated to studies of age-dependent variability of brain region ratios in mammals, the problem of variation in the morphology of the brain . according to fluctuations in environmental parameters is still ! obscure. Seasonal changes in brain region ratios previously < were observed in rodents in the genera Clethrionomys and : Microtus (Yaskin, 1980, 1984). The patterns of seasonal changes in the masses of different i regions of the brain in common shrews were diverse (Fig. 2; , Table 2). The most pronounced winter loss of mass occurred in the forebrain ( — 31.5%; P < 0.001), and especially in the neocortex ( — 37.4%; P < 0.001). The masses of the hippocampus , paleocortex , striatum, and ! mesencephalon + diencephalon decreased for this period by j 20-29 % . On the contrary, no noticeable winter regression in r the masses of such regions as olfactory bulbs, myelencephalon, and cerebellum were observed. During a period of gain in mass in spring, the rate of mass increase was different in different i brain structures. The most considerable increase was observed in the hippocampus (-1-32.6%; P < 0.001). This region showed the most noticeable sexual differences in its mass that occurred ! only as reproductive activity began. Although the seasonal variations in the absolute masses of ^ different brain regions were substantial, the seasonal changes in \ the ratio of brain regions were also appreciable. During winter i the relative mass of the forebrain declined; the most intensive | decrease was observed in the neocortex ( — 18.4%; P < 0.001). By contrast, the relative mass of myelencephalon increased by ■ 32.6% (P < 0.001). The relative mass of the olfactory bulbs and cerebellum increased significantly simultaneously. The highest relative mass increase during the winter-to-summer , period was found in the hippocampus (-1-15.2%; P < 0.001). | The quantitative ratio of different regions of the brain is altered during the winter and does not return in overwintered adults to the state observed in young. The proportions of the ; brain were considerably different. In overwintered shrews, the relative masses of the forebrain, neocortex, paleocortex, and j striatum were lower than in young nonwintered shrews (during | the spring mass gain, the mass of these regions did not reach |l the autumnal level), while the relative masses of the cerebellum 1 and myelencephalon were higher. There were no noticeable | differences in masses of the hippocampus, mesencephalon -f diencephalon, and bulbus olfactorius. 1994 YASKIN — Brain Morphology of Sorex araneus 157 The pattern of seasonal dry mass changes of different brain regions (Fig. 3, Table 3) did not exactly coincide with the pattern of that before drying (Fig. 2, Table 2), but the general picture was similar. Seasonal changes were comparatively more pronounced in the dry mass of forebrain (—21.2%; P < 0.001), and the neocortex lost as much as 27.6% {P < 0.001) of its dry matter. Winter regression in dry mass of the bulbus olfactorius, myelencephalon, and cerebellum was not observed. Dry masses of the forebrain and such parts as the neocortex, paleocortex, and striatum in overwintered adults did not reach the level characterized for young before winter, while the dry mass of another forebrain region, the hippocampus, was similar in both age groups. In overwintered adults the dry matter mass of the cerebellum and myelencephalon was appreciably higher compared to that in the young. Some sexual differences in these variations were observed. Brain mass of females was lower than in males, but usually only in summer months. The hippocampus showed the most substantial sexual differences in its size that occurred only as reproductive activity began. The absolute and relative masses of this region in males were higher by 5-7 % (P < 0.05) than in females. In addition, certain changes of brain mass and brain structure in common shrews were correlated with effects of the 1975 summer drought (western Siberia locality) when the body masses of young shrews in June- August were on the average lower by 8% than in other years. The brain mass was lower by 5% in July 1975 (P < 0.05). The differences disappeared in October. Some differences were also found in proportions of the brain. The ratio of forebrain to whole brain mass in young shrews from the summer drought was 6.3 % lower (P < 0.01) with respect to other years. Discussion Both wet and dry masses of some brain regions in overwintered shrews were higher than in young ones of the previous summer, while the total brain mass of adults was significantly lower. This indicates that the brain regions have different regularity in age-dependent patterns of development. Some parts, such as the forebrain, decreased in size with age, whereas others (the cerebellum and myelencephalon) increased significantly. Therefore, Dehnel’s phenomenon (Dehnel, 1949) does not affect some of the brain regions. The present data on seasonal variations of brain morphology in common shrews agree to a considerable extent with the results obtained in the bank vole, Clethrionomys glareolus (Yaskin, 1980, 1984). The comparison of the age-related mass variability of different brain regions in shrews with known data on alterations of learning abilities of mammals as aging progresses (Obrazcova, 1964) leads to the suggestion that the greater dimensions of the forebrain and neocortex in young shrews may be related to their higher need to react to environmental perturbations. Lxiwering of the vital activities may be accompanied by consequent cessation of some organismal function, the phylogenetically younger functions ending first (Astvazaturov, 1939). My results illustrated such a fundamental strategy concerning morphophysiological alterations in the brain. The forebrain suffers the most impressive mass regression during the winter (and drought also), the phylogenetically youngest part of the brain (neocortex) in particular. The mass of the relatively phylogenetically ancient bulbus olfactorius and myelencephalon remained constant during the winter period. My results revealed the broad morphological variability of such a highly specialized and relatively stable organ as the brain. This variation is correlated with age, sex, and different environmental factors such as geographic location, season, and extreme climatic parameters, such as drought conditions. The similarity in morphological brain changes in winter and during drought conditions suggests that these variations represent a more generalized mammalian pattern of adaptive mechanisms. It seems that these morphological changes are advantageous as a nonspecific adaptive mechanism by conferring an adaptive advantage of conservation of energy during the harsh winter and summer drought periods. Literature Cited Astvazaturov , M. I. 1939. On the nature of the stomach reflexes. Proceedings of the Military-Medical Academy, 20:169-172 (in Russian). Bielak, T., and Z. Pucek. 1960. Seasonal changes in the brain weight of the common shrew (Sorex araneus araneus Linnaeus, 1758). Acta Theriologica, 3:297-300. Gabon, K. 1956. Untersuchungen iiber die saisonale Veranderlichkeit des Gehimes bei der kleinen Spitzmaus (Sorex minutus minutus L.). Annales of the University of M. Curie-Sklodowska, Section C, 10:93-115. Dehnel, A. 1949. Studies on the genus Sorex L. Annales of the University of M. Curie-Sklodowska, Section C, 4:17-102. Dmitrieva, N. I. 1969. Growth of brain and spinal cord in postnatal ontogeny in the white rat. Pp. 132-144, in Brain Development in Animals, Nauka, Leningrad (in Russian). HyvarINEN, H., and K. Heikura. 1971. Effect of age and seasonal rhythm on the growth patterns of some small mammals in Finland and in Kirkenes, Norway. Journal of Zoology (London), 165:545-556. Latimer, H. B. 1950. The weight of the brain and its parts and the weight and length of the spinal cord in the adult male guinea pig. Journal of Comparative Neurology, 93:37-51. Mock, O. B. 1994. Effects of melatonin on the chronobiology of the least shrew, Cryptotis parva. Pp. 267-269, in Advances in the Biology of Shrews (J. F. Merritt, G. L. Kirkland, Jr., and R. K. Rose, eds.), Carnegie Museum of Natural History Special Publication no. 18, x + 458 pp. Obrazcova, G. A. 1964. On the periodism of higher nerve activity development in the early postnatal orthogenesis of the rabbit and dog. Journal of Higher Nerve Activity, 4:644-651 (in Russian). Pucek, M. 1965. Water contents and seasonal changes of the brain weight in shrews. Acta Theriologica, 10:353-367. Pucek, Z. 1955. Untersuchungen iiber die Veranderlichkeit des Schadels in Lebenszyklus von Sorex araneus araneus L. Annales of the University of M. Curie-Sklodowska, Section C, 9:113-211. 1965. Seasonal and age changes in the weight of the internal organs of shrews. Acta Theriologica, 10:369-438. 1970. Seasonal and age change in shrews as an adaptive process. Symposium of the Zoological Society of London, 26:189-207. Quay, W. B. 1984. Winter tissue changes and regulatory mechanisms 158 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 in nonhibemating small mammals: A survey and evaluation of adaptive and non-adaptive features. Pp. 149-163, in Winter Ecology of Small Mammals (J. F. Merritt, ed.), Carnegie Museum of Natural History Special Publication No. 10, ix + 380 pp. Sawicka-Kapusta, K. 1974. Changes in the gross body composition and energy value of the bank vole during their postnatal development. Acta Theriologica, 19:27-54. Yaskin, V. A. 1980. Seasonal changes of brain morphology, main morphological indices and behavior in bank voles. Pp. 152-159, in Animal Adaptations to Winter Conditions, Nauka, Moscow (in Russian). 1984. Seasonal changes in brain morphology in small mammals. Pp. 183-191, in Winter Ecology of Small Mammals (J. F. Merritt, ed.), Carnegie Museum of Natural History Special Publication No. 10, ix + 380 pp. _. 1989. Comparative analysis of seasonal changes in brain and skull morphology in small mammals from geographically distant populations. Pp. 882-883, in Proceedings of the Fifth International Theriological Congress, Rome, Italy, (Abstract). Table 1. — Seasonal and age changes in brain and body mass of Sorex araneus from western Siberia and Moscow region. Values given are mean ±SE. Season Site Sample Size Body Mass (g) Brain Mass (mg) Summer (VI- VIII) Siberia 103 7.06 + 0.09 258 + 1.9 Moscow 10 7.36 ± 0.12 262 ± 3.6 Autumn (X) Siberia 32 6.47 ± 0.19 236 ± 2.5 Moscow 6 6.02 + 0.25 235 + 3.7 Winter (II-III) Siberia 18 5.33 ± 0.17 207 ± 3.8 Moscow 18 5.23 ± 0.08 193 ± 2.4 Summer (VI- VIII) Siberia 62 9.27 + 0.28 221 ± 2.2 Moscow 9 19.91 ± 0.31 207 ± 2.1 Table 2. — Seasonal and age changes (in percent) in the wet mass of the brain regions in common shrews, Sorex araneus, yrow western Siberia. 1 , autumn-winter regression; 2, spring “jump of growth 3, differences between nonwintered and overwintered animals; im, immature; m, mature. Brain Regions Summer (im)-* Winter (im) 1 Winter (im)-» Summer (m) 2 Summer (im)-* Summer (m) 3 % P % P % P forebrain -31.5 <0.001 + 17.1 <0.001 -19.8 <0.001 neocortex -37.4 <0.001 + 18.4 <0.01 -25.9 <0.001 hippocampus -29.2 <0.001 + 32.6 <0.001 -6.1 >0.05 paleocortex -28.1 <0.001 + 12.5 <0.05 -19.18 <0.05 striatum -27.6 <0.001 + 5.7 >0.05 -23.4 <0.001 bulbus olfactorius -4.2 >0.05 -2.8 >0.05 -10.1 >0.05 mesencephalon + diencephalon -20.1 <0.05 + 22.3 <0.01 -2.3 >0.05 cerebellum -8.3 >0.05 + 22.8 <0.01 + 12.6 <0.05 myelencephalon + 1.9 >0.05 + 10.0 <0.05 + 12.2 <0.05 1994 YASKIN — Brain Morphology of Sorex araneus 159 Table 3. — Seasonal and age changes (in percent) in the dry mass of the brain regions in common shrews from western Siberia. Designations as in Table 2. Brain Regions Summer (im)-* Winter (im) 1 Winter (im)-» Summer (m) 2 Summer (im)-» Summer Cm) 3 % P % P % P forebrain -21.2 <0.001 + 15.7 <0.001 -8.8 <0.05 neocortex -27.6 <0.001 + 18.2 <0.01 -14.3 <0.01 hippocampus -19.7 <0.01 + 25.6 <0.001 + 0.9 >0.05 paleocortex -19.5 <0.001 + 12.7 <0.05 -9.3 <0.05 striatum -20.9 <0.01 + 9.5 <0.05 -8.3 <0.05 bulbus olfactorius -5.1 >0.05 + 9.1 >0.05 + 3.4 >0.05 mesencephalon + diencephalon -16.9 <0.01 + 18.1 <0.05 -1.9 >0.05 cerebellum -4.8 >0.05 + 21.5 <0.01 + 15.6 <0.05 myelencephalon + 4.6 >0.05 + 13.2 <0.01 + 18.4 <0.01 JJASONDJFMAMJJAS Fig- 1-— Seasonal and age changes in the body mass (g). a, absolute (rag); and b, relative (mg/g) brain mass of the common shrew {Sorex araneus) from western Siberia. At the beginning of the curves, the shrews are young nonwintered animals and at the end of the curves, adult wintered animals. 160 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 S I ' • 'I 1 I' j - Fig. 2. — Seasonal and age changes in the wet mass (mg) of the brain and its regions in common shrews (western Siberia). At the beginning of the curves, the shrews are young nonwintered animals and at the end of the curves, adult wintered ajiimals. BR, brain; FBR, forebrain; NC, neocortex; HI, hippocampus; PC, paleocortex; ST, striatum; B.O., bulbus olfactorius; ME+DI, mesencephalon + diencephalon. The ends of the vertical lines indicate ±SE. YASKIN — Brain Morphology of Sorex araneus 161 1994 (mg) I — I I 1 1 1 1 [ 1 I r— — r 1 r- — t \ , ^ / U I I ^ I I i I I I I I I I J J JASONDJ FMAMJJ Fig 3. — Seasonal and age changes in the dry mass (mg) of the brain and its regions in common shrews (western Siberia). Designations as in Fig. 2. 13' THERMAL BIOLOGY OF FREE-RANGING SHREWS AS REVEALED BY COMPUTER-FACILITATED RADIOTELEMETRY: ENERGETIC IMPLICATIONS Joseph F. Merritt* and Francisco Bozinovic^ * Powdermill Biological Station, Carnegie Museum of Natural History, Star Route South, Rector, PA 15677, USA; ^ Departamento de Ciencias Ecologicas, Facultad de Ciencias, Universidad de Chile, Casilla 653, Santiago, Chile Abstract Seasonal changes in core body temperature of northern short-tailed shrews (Blarina brevicauda) residing in a natural outdoor enclosure were monitored by radiotelemetry techniques. The study was conducted at the biological station of the Carnegie Museum of Natural History, southwestern Pennsylvania, during autumn, winter, spring, and summer of 1989-1990. Radiotransmitters were surgically implanted within the peritoneal cavity of shrews. Each monitoring period lasted five days, during whieh time body temperature readings were recorded every 15 min using a computer subsystem. A temperature “pulse” of the implanted shrew was relayed from an outdoor enclosure to a lab-based receiver interfaced with a hardware-software DATACOL-AVM subsystem. At the laboratory, instantaneous body temperature readings were automatically tabulated by an Apple 11 -I- computer. Body temperature of shrews was maintained eontinuously at euthermy, and averaged 38.3°C (n = 1,470 readings) ranging from 34.4-42.2°C during the year-long study. Minimum body temperatures occurred during summer while maximum temperatures occurred during autumn {P = 0.001). Shrews compensated for the high cost of continuous euthermy by minimizing thermal conductance through an increase in pelage insulation. Soricids respond to the winter physical environment by coupling of anatomical, behavioral, and physiological modifications of their thermoregulatory constitution. Introduction The overwinter survival of small, nonhibemating mammals is largely contingent on their ability to cope with extreme cold coupled with a reduced availability of food and water. In north- temperate and boreal regions of North America, the winter period may last up to 7 'A months with minimum ambient temperatures frequently reaching — 40°C (Merritt, 1984). In response to these selection pressures, small mammals have evolved behavioral, anatomical, and physiological adjustments which permit their overwinter survivorship (Wang and Hudson, 1978; Vogel, 1980; Feist, 1984; Hanski, 1984, 1989; Hyvarinen, 1984; Merritt, 1984, 1986; Wunder, 1984; Hutterer, 1985; Heller et al., 1986; Zegers and Merritt, 1988a, imb). Wunder (1984) postulated that during winter, small mammals exhibit a variety of behavioral, physiological, and anatomical compensatory mechanisms to cope with reduced food and increased cold. These mechanisms included avoidance behavior (i.e., migration, use of ameliorating microclimates, torpidity, increased thermal insulation) and resistance (i.e., increased thermogenesis by changes in basal metabolic rate, nonshivering thermogenesis, shivering, and increases in level of activity). Blarina brevicauda, the largest North American shrew, is distributed throughout much of the eastern half of North America, where it is one of the most abundant mammals. This semifossorial soricid is most common in mesic forests possessing a well-developed layer of leaf litter and humus, but can be found in a diverse array of plant assemblages (George et al., 1986). Research on the population dynamics of B. brevicauda derived from field studies in Ontario, Manitoba, Minnesota, and Pennsylvania reveals good overwinter survivorship in regions characterized by ambient temperatures reaching — 40°C (Buckner, 1966). The apparent success of B. brevicauda in coping with harsh winter climates derives from a complex suite of behavioral, anatomical, and physiological mechanisms which contribute to their thermoregulatory constitution (Merritt, 1986; Merritt and Adamerovich, 1991; Bozinovic and Merritt, 1992). Merritt (1986) described two resistance mechanisms (sensu Wunder, 1984) in B. brevicauda — increased resting metabolic rate and nonshivering thermogenesis during winter. Also, by employing radiotelemetry techniques on free-ranging shrews, Merritt and Adamerovich (1991) demonstrated that B. brevicauda did not possess the capability to undergo torpor, and did not display communal nesting as a means of energy conservation during winter. Randolph (1973), and more recently Bozinovic and Merritt (1992), reported a lower rate of heat loss in B. brevicauda during winter, compared with summer-caught shrews. This increased insulation during winter is reported as a plausible avoidance mechanism to cold by small mammals (Wunder, 1984). In order to assess accurately the ecology and behavior of free-ranging species of mammals, researchers have employed radiotelemetry techniques. Such applications have been applied to many mammals in order to understand their activity, movements, survival, and thermal physiology (Amlaner and MacDonald, 1979; Cochran, 1980; Mech, 1983; Kenward, 1987). However, radiotelemetry is limited in use for many species of small mammals due to the comparatively large size of transmitters. Because of this limitation in technology, only two studies have been conducted on members of the family Soricidae. Cawthom (1989) detailed the ecology and activity patterns, and Merritt and Adamerovich (1991) examined temperature regulation in the largest North American soricid, B. brevicauda. Several questions concerning the thermal biology of shrews have surfaced as a result of these initial studies, and prompted further refinement in techniques designed to elucidate daily body temperature fluctuations of the northern short-tailed shrew. Thus, the objective of our study was to determine the thermoregulatory patterns of B. brevicauda and assess the 163 164 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 energetic implications of the maintenance of continuous euthermy in this soricid. The thermal biology of free-ranging shrews was elucidated during autumn, winter, spring, and summer of 1989-1990 at Powdermill Biological Station by employing computer-facilitated radiotelemetry techniques. Methods Northern short-tailed shrews were live-trapped at Powdermill Biological Station, Carnegie Museum of Natural History, southeastern Westmoreland County, Pennsylvania (Merritt, 1986). The collection site (elevation, 400 m) encompassed a secondary growth forest of hawthorn (Crataegus sp.), crab apple (Pyrus coronaria), black locust (Robinia pseudoacacia), sugar maple (Acer saccharum), and black cherry (Prunus serotina). Upon capture, shrews were marked by toe clipping, sexed, weighed, and their reproductive status recorded if evidenced by external criteria. Candidates for surgery were maintained in the animal facility for two days on Tenebrio larvae, cat food, and water ad libitum. Shrews were then anesthetized by using methoxyflurane (Pitman-Moore, Inc.) inhalation therapy. A radiotransmitter battery package (AVM Instrument Company, Inc., Livermore, California) encapsulated in “Elvax” paraffin coating material (The Mini-Mitter Company, Inc., Sunriver, Oregon) was surgically implanted within the peritoneal cavity of the shrew by access through an incision in the ventro-lateral abdominal wall. Each 2. 6-3.0 g encapsulated radiotransmitter battery package was pretuned to a specific frequency ranging between 150-151 MHz. To ensure that frequencies did not drift during the study, temperature-sensitive transmitters were calibrated before and after use by immersion in a water bath of known temperature (ranging from 22-42 °C) as recorded by a YSI 2100 T elethermometer . Following surgical implantation, the peritoneum and outer skin were closed using 4-0 Ethicon silk. The shrew was retained in the animal facility for two days following surgery, and then released in an outdoor enclosure. The body mass of the shrew was recorded before release. The purpose of the outdoor enclosure was to restrict the movements of shrews to facilitate telemetry measurements and also to minimize loss of implanted animals due to dispersal or predation. The outdoor enclosure, placed in a maple-cherry-locust forest, measured 4.8 X 2.5 X 1.4 m; it was constructed of 14 " clear plexiglass framed by “2 X 4” outdoor lumber (Merritt and Adamerovich, 1991). The interior of the enclosure consisted of a layer of soil 1.5 m deep provided with rocks, logs, sticks, ferns, and herbs, to simulate the natural environment. The enclosure was covered with 2 cm^ polypropylene net to restrict predators, while permitting normal precipitation to reach the inside environment. The bottom of the enclosure consisted of small-gauge fiberglass screening set upon a 40-cm deep base of no. 2B gravel to permit drainage and prohibit burrowing activities of shrews and other small mammals. A soil embankment graded from the surrounding ground to the side of the enclosure in order to maintain a soil temperature inside the enclosure that approximated that of the outside soil. Within the enclosure, three wooden nest boxes were positioned along three walls and supplied with grasses and leaves. Nest boxes were buried slightly below ground level, and a feeding station sheltered within a plastic canister was positioned adjacent to each box. Three different temperature regimes were recorded in the enclosure during the study periods by a three-point thermograph ; (Model 4030; Qualimetrics, Inc., Sacramento, California). ; Temperatures were recorded 1.5 m above ground surface ? (ambient), on ground surface, and in a subsurface tunnel. Snow ! depth was also measured within the enclosure. Shrews residing ' in the enclosure were provided with 5.0 g of Tenebrio larvae equivalent to ca 31.5 Kcal per day (Merritt and Adamerovich, 1991). I Each seasonal monitoring period lasted from 5-6 days. Within the enclosure, pulse signals from shrews with transmitters were received and conveyed at 15 -min intervals via a Yagi antenna to an LA 12-DS portable telemetry receiver in line with a two-stage repeater transmitter. The signals then were transmitted and received in the laboratory (160 m from the enclosure) by another LA 12-DS receiver. This lab-based receiver w'as facilitated by a DATACOL-AVM hardware-software subsystem (AVM Instrument Company, Inc.) that converted pulse interval signals into instantaneous body temperature readings and automatically displayed and tabulated these data on an Apple II + computer. Results ! Seasonal Changes in Body Temperature The highest core body temperature for B. brevicauda i occurred during autumn. Body temperature of one adult B. brevicauda (22.6 g) monitored from 6-10 November 1989 averaged 41.2°C and ranged from 40.0-42.4°C (n = 362 readings; Fig. 1). Temperatures at ground level ranged from n — 1 to 3°C during this time. The range in body temperatures of this shrew represented the greatest fluctuation, and also the > highest level of body temperature of a shrew for all seasons j studied. Of the 362 body temperature readings sparming five j days, only three were above 42.0°C, indicating such temperatures were probably spurious readings, and thus were ;i declared as outliers in the sample. j Body temperatures of B. brevicauda (adult, 23.9 g) during the winter period (10-15 January 1990) averaged 38.4°C and u ranged from 37.2-40.0°C {n = 423 readings; Fig. 2). |j Temperatures on the ground surface within the enclosure ranged ii from — 1 to — 0.5°C for the six-day study period and snow ; cover was intermittent reaching a maximum depth of only 30 ' cm. This shrew exhibited a conservative thermoregulatory ; budget with 96% of the readings between 38.0-39.0°C. Body. temperatures for B. brevicauda (adult, 23.9 g) during the spring period (5-10 April 1990) were similar to those of winter and averaged 38.5°C (range, 37.0-39.6°C, n = 322 ' readings; Fig. 3.). Temperatures on the ground surface of the i enclosure ranged from 9-15°C. As was the case with the shrew i during the winter trial (Fig. 2), body temperature fluctuations were minimal, with 93 % of the readings between 38.0-39.0°C. Body temperature fluctuations for B. brevicauda (subadult, 18.15 g) during the summer period (7-11 August 1990) averaged 35.4°C and ranged from 34.4-36.0°C (n = 363 1994 MERRITT AND BOZINOVIC — Thermal Biology of Blarina brevicauda 165 readings; Fig. 4). Ground temperature during this period ranged from 17.5-23.0°C. This shrew exhibited the most conservative level of core body temperatures of all shrews tested, with only 3% (12/363) of the readings outside of 35.0-36.0°C. In contrast to B. brevicauda monitored during the autumn trial (Fig. 1), this shrew exhibited the lowest core body temperature of all shrews tested during the study. A one-way ANOVA revealed significant differences in body temperature through the year [F (3,146.9) = 14,165.77; P < 0.0001]. The a posteriori Student-Newman-Keuls test revealed significant differences in body temperature between seasons {P < 0.05), except between winter and spring (38.4 vs 38.5 respectively, P > 0.05). Thus, a comparison between body temperature and time of day for each seasonal trial was clearly random. Results should be interpreted with some caution due to the fact that each seasonal trial, although composed of many data points (temperature readings), consisted of only one shrew. Further, shrews were commonly derived from different cohorts, thus age of a given shrew would likely influence the thermal biology of a seasonal trial. All shrews were judged to be in a nonreproductive state during the trial periods. Evaluation of the Technique Results of the present study indicate that seasonal changes in core body temperature of B. brevicauda can be evaluated satisfactorily by use of computer-facilitated radiotelemetry techniques. All shrews recovered well from the methoxyflurane inhalation therapy and the surgical procedure as evidenced by a 100% survivorship. The intraperitoneal positioning of a radiotransmitter (less than 15% of body mass) caused no apparent short-term aberrations in behavior or physiology of northern short-tailed shrews. Postoperative gross inspection of the peritoneum and associated organs revealed no pathological effects attributable to the presence of the implant. Body temperatures of implanted shrews, confined to a natural outdoor enclosure, are realistic approximations of those core body temperatures exhibited by shrews residing in the natural environment. This fact was confirmed by periodic recording of body temperatures by use of rectal probes during the period in which shrews were implanted with transmitters. Major shortcomings of the technique described herein are associated with the cost of temperature transmitters (ca $380 each). In most cases, the cost of employing this technology prohibits large sample sizes, thus restricting the data base and compromising the validity of results. Further, this “economic factor” contributes to a reluctance of investigators to use this technology in an “unrestrictive” field situation. Technological advances in the field of radiotelemetry coupled with use of outdoor enclosures may result in increased use of this technique with members of the family Soricidae. Discussion The family Soricidae is composed of some 266 species belonging to 20 different genera (Churchfield, 1990). This family is distributed on all continents except Australia, Antarctica, and central and southern South America. Of the family Soricidae, the subfamily Crocidurinae exhibits a tropical distribution centered in Africa and Asia, while the subfamily Soricinae exhibits a Holarctic and northern Neotropical distribution (Corbet and Hill, 1980; Nowak and Paradiso, 1983). Members of the Soricinae inhabiting northern regions are faced with extreme cold and scarcity of food and water during the long winter period (Merritt, 1986; Sheftel, 1989). Members of this subfamily possess high surface area/volume ratios, and are typified by elevated mass-independent metabolic rates. Small endotherms equipped with the ability to exhibit physiological heterothermy would surely possess an adaptive edge for coping with harsh climatic perturbations. However, physiological heterothermy is uncommon within the Soricinae (Dawson, 1973; Vogel, 1976, 1980; Genoud, 1985, 1988; Merritt and Adamerovich, 1991). Results of our study and those of an earlier work (Merritt and Adamerovich, 1991) clearly indicate that during autumn, winter, and spring, B. brevicauda (a member of the subfamily Soricinae) did not exhibit physiological heterothermy as do various members of the more southerly subfamily Crocidurinae (Genoud, 1988). Instead, body temperatures recorded at 15-min intervals ranged from 37.4-41.8°C during autumn, winter, and spring. Maximum body temperature occurred during the autumn period. This increase in thermogenic capacity represented a response to cold cues received at ground surface (the foraging zone of the shrew) and is due, in part, to nonshivering thermogenesis mediated by brown adipose tissue (Merritt, 1986). Body temperature of B. brevicauda recorded during the summer period did show a slight departure from euthermy, ranging from 34.2-36.0°C. This slight decline in core body temperature may represent a thermogenic response to warm temperatures encountered by the shrew on the ground surface while foraging. Conclusions and implications of this occurrence are tenuous, complicated by a low sample size coupled with the fact that the single individual monitored during the summer period was a subadult. Our results demonstrated that B. brevicauda did not undergo a state of torpor or heterothermy as a means of energy conservation during winter. Here we evaluate the role of resistance (increased activity) and avoidance (changes in pelage insulation) as compensatory mechanisms influencing the amount of energy allocated to maintenance or respiration in the short- tailed shrew during winter. We compare our results with the general model proposed by Wunder (1975). This model predicts the amount of energy used for maintenance (R), given body weight (W) in grams, ambient temperature (T^) in °C and the degree of running activity in km/h. The model is expressed as: R = a Mbasal + Mjr + Mgctivity where a is a coefficient that modifies the metabolic rate for the posture associated with activity, is the basal metabolic rate, Mjp is the metabolic rate associated with temperature regulation below the thermoneutral zone, and Ma^bvity the metabolic rate due to the level of activity. In a mathematical form the model is expressed as: 166 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 T = a(3.8 W-O-25) + 1.05W-°-5 [(38 - 4W«'25) _ ^ (g 46 V where V (km/h) is the velocity of running. Assuming B. brevicauda with a weight of W = 20.0 g under conditions of resting metabolic rate (i.e., a = 1.0, V = 0, winter T^ = — 1.0°C, and summer T^^ = 20°C [see Results]), the model predicts that during summer, R = 54.65 cal/g h (1.09 Kcal/animal h), while during winter, R = 155.90 cal/g h (3. 12 Kcal/animal h). Using the values of M},asai reported by Merritt (1986), i.e., 13.74 cal/g h during summer, and 16.63 cal/g h during winter, the values of thermal conductance (TC) documented by Bozinovic and Merritt (1992), i.e., 2.117 cal/g h °C during summer, and 1.532 cal/g h °C during winter, a mean value of body temperature (T3) = 38.6°C (see Results). Using the model of Wunder (1975) in the form: R = a Mjjagai + TC (T3 — T^), the amount of energy used for maintenance during summer is 53.1 1 cal/g h (1.06 Kcal/animal h), while during winter the energy devoted to maintenance is 77.29 cal/g h (1.50 Kcal/animal). The amount of energy used for maintenance (R) measured during winter was 48 % lower than predicted by the model (Wunder, 1975). During summer energy used for maintenance (R) was practically the same as predicted by the model (Fig. 5). Results of our radiotelemetry research revealed that B. brevicauda was a continuously euthermic species (Fig. 1-4). Our analysis of energy allocated to maintenance indicated that winter individuals compensate for the large difference between body temperature and temperatures encountered during foraging by morphological, physiological, and behavioral compensatory mechanisms. These changes are especially important for species of soricids that possess small body sizes because the cost of continuous endothermy is extremely high (Vogel, 1980; Genoud, 1988; McNab, 1991; Merritt and Adamerovich, 1991). Our research indicates that the avoidance and resistance tactics delimited herein act synergistically to enhance winter survivorship for B. brevicauda and may explain in part the evolutionary success of shrews inhabiting cold regions of the world. Acknowledgments We are indebted to T. A. Kromel for assistance with research activities. We thank I. Benzinger, T. A. Kromel, and S. McCollam for help in preparing the figures, and the staff at AVM Instrument Company for assisting with radiotelemetry technology. F.B. acknowledges a fellowship from the Fundacion Andes (Chile) and the International Program of the Carnegie Museum of Natural History. Literature Cited Amlaner, C. J., Jr., and D. W, MacDonald. 1979. A Handbook on Biotelemetry and Radio Tracking. Pergamon Press, Oxford, 804 pp. Bozinovic, F., and J. F. Merritt. 1992. Summer and winter thermal conductance of Blarina brevicauda (Mammalia: Insectivora: Soricidae) inhabiting the Appalachian Mountains. Annals of Carnegie Museum, 61:33-37. Buckner, C. H. 1966. Populations and ecological relationships of shrews in tamarack bogs of southeastern Manitoba. Journal of Mammalogy, 47:181-194. Cawthorn, j. M. 1989. The population ecology and temporal and spatial activity of three species of shrews (Sorex cinereus, S. fumeus, and Blarina brevicauda) in southwestern Pennsylvania. Unpublished Ph.D. dissert.. Bowling Green State University, Bowling Green, Ohio, 124 pp. Churchfield, S. 1990. The Natural History of Shrews. Christopher Helm, London, 178 pp. Cochran, W. W. 1980. Wildlife telemetry. Pp. 507-520, in Wildlife Management Techniques Manual, 4th ed. (S. D. Schemnitz, ed.), The Wildlife Society, Washington, D.C., 686 pp. Corbet, G. B., and J. E. Hill. 1980. A World List of Mammalian Species. British Museum of Natural History, 813 pp. Dawson, T. J. 1973. Primitive mammals. Pp. 1-46, in Comparative Physiology of Thermoregulation (G. C. Whittow, ed.). Academic Press, New York, 278 pp. Feist, D. D. 1984. Metabolic and thermogenic adjustments in winter acclimatization of subarctic Alaskan red-backed voles. Pp. 131-137, in Winter Ecology of Small Mammals (J. F. Merritt, ed.), Carnegie Museum of Natural History Special Publication no. 10, 380 pp. Genoud, M. 1985. Ecological energetics of two European shrews: Crocidura russula and Sorex coronatus (Soricidae: Mammalia). Journal of Zoology, (London), 207:63-85. 1988. Energetic strategies of shrews: Ecological constraints and evolutionary implications. Mammal Review, 18:173-193. George, S. B., J. R. Choate, and H. H. Genoways. 1986. Blarina brevicauda. Mammalian Species, 261:1-9. Hall, E. R. 1981. Mammals of North America. 2nd ed. John Wiley and Sons, New York, 2 volumes, 1,373 pp. Hanski, I. 1984. Food consumption, assimilation and metabolic rate in six species of shrew (Sorex and Neomys). Annales Zoologici Fennici, 21:157-161. 1989. Population biology of Eurasian shrews: Towards a synthesis. Pp. 469-479, in Population Biology of Eurasian Shrews (I. Hanski and E. Pankakoski, eds.), Annales Zoologica Fennici, 26:1-479. Heller, H. C., X. J. Musacchai, and L. C. H.Wang (eds.). 1986. Living in the Cold: Physiological and Biochemical Adaptations. Elsevier, New York, 587 pp. HUTTERER, R. 1985. Anatomical adaptation of shrews. Mammal Review, 15:43-55. Hyvarinen, H. 1984. Wintering strategy of voles and shrews in Finland. Pp. 139-148, in Winter Ecology of Small Mammals (J. F. Merritt, ed.), Carnegie Museum of Natural History Special Publication no. 10, 380 pp. Kenward, R. 1987. Wildlife Radio Tagging. Academic Press, New York, 222 pp. McNab, B. K. 1991. The energy expenditure of shrews. Pp. 35-45, in The Biology of the Soricidae (J. S. Findley and T. L. Yates, eds.). Special Publication, The Museum of Southwestern Biology, 1:1-91. Mech, L. D. 1983. Handbook of Animal Radio-Tracking. University of Minnesota Press, Minneapolis, 107 pp. Merritt, J. F. (ed.). 1984. Winter Ecology of Small Mammals. Carnegie Museum of Natural History Special Publication no. 10, 380 pp. 1986. Winter survival adaptations of the short-tailed shrew (Blarina brevicauda) in an Appalachian montane forest. Journal of Mammalogy, 67:450-464. Merritt, J. F., and A. Adamerovich. 1991. Winter 1994 MERRITT AND BOZINO VIC— Thermal Biology of Blarina brevicauda 167 thermoregulatory mechanisms of Blarina brevicauda a.s revealed by radiotelemetry. Pp. 47-64, in Biology of the Soricidae (J. S. Findley and T. L. Yates, eds.). Special Publication, The Museum of Southwestern Biology, 1:1-91. Nowak, R. M., and J. L. Paradiso. 1983. Walker’s Mammals of the World. 4th ed. Johns Hopkins University Press, Baltimore, 1:1-568, 2:569-1362. Randolph, J. C. 1973. Ecological energetics of a homeothermic predator, the short-tailed shrew. Ecology, 54:1166-1187. Sheftel, B. I. 1989. Long-term and seasonal dynamics of shrews in central Siberia. Pp. 357-369, in Population Biology of Eurasian Shrews (I. Hanski and E. Pankakoski, eds.), Annales Zoologici Fennici, 26:1-479. Vogel, P. 1976. Energy consumption of European and African shrews. Acta Theriologica, 21:195-206. 1980. Metabolic levels and biological strategies in shrews. Pp. 170-180, in Comparative Physiology: Primitive Mammals (K. Schmidt-Nielsen, L. Bolis, and C. R. Taylor, eds.), Cambridge University Press, Cambridge, England, 338 pp. Wang, L. C. H., and J. W. Hudson (eds.). 1978. Strategies in Cold: Natural Torpidity and Thermogenesis. Academic Press, New York, 715 pp. WUNDER, B. A. 1975. A model for estimating metabolic rate of active or resting mammals. Journal of Theoretical Biology, 49:345-354. 1984. Strategies for and environmental cueing mechanisms of seasonal changes in thermoregulatory parameters of small mammals. Pp. 165-172, in Winter Ecology of Small Mammals (J. F. Merritt, ed.), Carnegie Museum of Natural History Special Publication no. 10, 380 pp. Zegers, D. a., ANDJ. F. Merritt. 1988a Effect of photoperiod and ambient temperature on nonshivering thermogenesis of Peromyscus maniculatus. Acta Theriologica, 33:273-281. 1988fe. Adaptations of Peromyscus for winter survival in an Appalachian montane forest. Journal of Mammalogy, 69:516-523. o o Lll DC Z) h- < DC LU Q_ LU f— > Q O m AUTUMN ^ DAY 1 42.0^ 41.0 40.0^ 39.0 I ' ' ' I ' ' ' I " ' I ' ' ' I ' ' ' I ' ' ' I ' ' ' I ' " I ' DAY 2 J 42.0 Y 41.0Y 40.0 -a 39.0 • I I ' ' I M I I I I I I I I I I I I I I I I I I I I I 4 I ' " I ' " I " ' I " ' I ' " I " ' I " ' I ' ' I I I, “n"'"' 1 1 ' I i 1 M i I E-42,0 41.0 ^40.0 39.0 ' D ' ' ' I ' " I " ' I " , I , , 1 I I I . I I . I I I , , I I , , , I , . , , , Y42.0 -41.0 r40.0 39.0 42.0 41.0 40.0 39.0 TIME (HOURS) Fig- 1- — Core body terapterature of one adult Blarina brevicauda (22.6 g) monitored by radiotelemetry techniques during autumn (6-10 November 1989) in a natural outdoor enclosure at Powdermill Biological Station. Average body temperature was 41 .2°C based on 362 readings. Dashed lines represent period of malfunctioning transmitter or incomplete transmission of signal. 168 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 WINTER O LU DC Z) h- < CE Ol 0. LU I- > Q o m 40 Ol 39.0 - i 38.0- 37 0- DAY 1 1 1 r400 -39 0 -380 -37.0 400“ 39.0- 38.0- 37.0“ 1 ' ' ■ 1 ■ ■ ■ 1 ■ ■ ■ 1 ■ ■ • 1 ■ ■ ’ 1 ■ • ■ 1 ' " 1 ■ " 1 " ' 1 " ' 1 " ' 1 ri ' 1 1 M 1 " 1 1 1 1 . 1 1 " 1 1 1 1 1 " 1 1 . 1 I 1 I M 1 " 1 1 M 1 1 " 1 1 1 1 1 1' DAY 2 1 ! f 40.0 -39.0 -380 -370 ■ 40.0- 39.0“ 38.0“ 370- DAY 3 ■ \ i " ■ r ■ ■ 1 ■ ‘ ■ ! ■ ■ ■ 1 ■ ■ ■ 1 ■ - ■ 1 ■ ■ ■ 1 ■ ' ■ I ■ ■ ■ 1 ■ ■ ■ ■ 1 ■ ■ ‘ 1 ■ ■ ■ 1 ■ ■ ■ 1 . . ■ 1 . ■ ■ I . . 1 . ■ ■ , 1 , . . 1 ( , .'i . . . -400 -390 r38 0 r37 0 40.0“ 39.0“ 380- 37 0“ DAY 4 'l ' 1' ' ‘ 1- ■ T ■ ■ r " r " r " r " 1 " T " r " r " i ■ ■ T "1 ■ " 1 " T " 1" ■ 1 ' ■ '1 " ' 1 ' ‘ ■ r " 1 ' ' ' f 40.0 -390 -380 r370 400- 39.0- 38 0- 37.0- -Av- 1 ' 1 " ■ r ■ ' r ■ ■ 1 " ' 1 ' " 1" ' 1 " M ' ’ ' 1' ‘ ' 1 ' ■ ' r ‘'''l ■ " i ' ] 1 ' M" ' 1' ' ■ 1 ’ “ 1 ’ ' ‘I' ' M ’ ■ ■ 1 ■ ■ -1 r40O -39.0 -38.0 -370 400- 390 T 380- 370 -i DAY 6 I r400 -390 -38.0 -370 0 1 2 3 4 5 6 7 8 9 10 1 1 12 13 14 15 16 17 18 19 20 21 22 23 24 TIME (HOURS) Fig. 2. — Core body temperature of one adult Blarina brevicauda (23.9 g) monitored by radiotelemetry techniques during winter (10-15 January 1990) in a natural outdoor enclosure at Powdermill Biological Station. Average body temperature was 38.4°C based on 423 readings. Dashed lines represent period of malfunctioning transmitter or incomplete transmission of signal. SPRING 40.0z,| 39.1 380- 37.0- M[M t I It » I M l| TM| ■ 'l| I i I) I M I ■ M| M ■[ M ■ I M ■ I M l[ .1 l| » M| 1 M| I M| Ml | ■ M [ ~ 6 7 8 9 10 11 12 t3 14 15 16 17 18 19 20 21 22 23 24 400 39 0 38.0 37.0 TIME (HOURS) Fig. 3.— Core body temperature of one adult Blarina brevicauda (23,9 g) monitored by radiotelemetry techniques during spring (5-10 April 1990) in a natural outdoor enclosure at Powdermill Biological Station. Average body temperature was 38,5°C based on 322 readings. Dashed lines represent period of malfunctioning transmitter or incomplete transmission of signal. 1994 MERRITT AND BOZINOVIC— Thermal Biology of Blarina brevicauda 169 SUMMER DAY 1 o LU CC h- < CC LU Q. 2 UJ h- > Q o m 37 0- 36.0-1 35.0^ 34,0 i yi-rr|irr| i n|Ttr| nTyrTtyr 37 0- „ i DAY 2 36,0 : 35 0 "• • 34 0 ; -V 'TI '' 'I r'-T" I ’ " r' "I’ 37.0- 36 0 DAY 3 35.0 34 0 ° 5DAY 4 36° i. .. 35.0-4 34 0-a I ... U360 i ^35 0 h340 rr-> Ti pmpiT] rrr ( ri-T-j rr-rpT n-pi-i'i p i r]-r-n p-n-prrrpf f37 0 , p36 0 ' ' •' - "-35 0 • ^34 0 i I M |i M| - ' |M ipTT, prp-n rpnyi TTprT l|-| . iirrT|-^ r37.0 - ^36 0 , n-p,-..p 34 0 p36.0 ^ r C . ^35.0 1-34.0 37.0- DAY 5 xpn r|-n-rTn-rp-T I 'n , . | rr r f rn pT-q tt T -p-n-pn r n-n-T| . .r|T-r,p, I p , I |- p37,Q =-36,0 35 0 34.0 'III l| II 1| I IT|ir'(TIn M I l|" I I" n '■I'TI rq-n i { ii ipTn'ITl-rri ri | rn P'l . r, i > > | :i i ; i m 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 TIME (HOURS) Fig. 4. — Core body temperature of one subadult Blarina brevicauda (18.15 g) monitored by radiotelemetry techniques during summer (7-11 August 1990) in a natural outdoor enclosure at Powdermill Biological Station. Average body temperature was 35.4°C based on 363 readings. Dashed lines represent period of malfunctioning transmitter or incomplete transmission of signal. .c Fig. 5. — Predicted and observed energy used for maintenance during summer and winter in Blarina brevicauda. Values for predicted energy use are according to the model of Wunder (1975). THE DEVELOPMENT OF THE SKULL OF SUNCUS MURINUS Yayoi Masuda’ and Takeshi Yohro” * Museum of Comparative Zoology, Harvard University, Cambridge, Massachusetts 02138. Current address: Department of Anatomy, Jichi Medical School, Tochigi 329-04, Japan; ^ Department of Anatomy, Faculty of Medicine, University of Tokyo, Hongo, Bunkyo, Tokyo 1 13, Japan Abstract The ontogenetic changes of the shape of skulls of Suncus murinus were described using 20- and 25-day embryos, newborn, and adult shrews. The length, width, and height of skulls and distances between specific points were measured in four embryonic stages (20-, 25-, 27-, and 29-day embryos) and four postnatal stages (newborn, 9-, and 14-day infants and adults). Most linear measurements of the skull reached 20%, 60%, and 90% of the adult value in 20-day embryos, neonates, and 14-day infants, respectively. The eye and the cochlear capsule developed early, whereas the growth of the mandible was retarded. The postglenoid grows profoundly after 14 days. Introduction The head of Suncus murinus, the musk shrew, has several unusual characters including a long and movable snout, small eyes, primitive ossified otic capsules, no zygomatic arch, double jaw articulation, and extraordinarily well-developed temporal muscles. An adult Suncus has a flat skull that is long and narrow (Fig. 1), whereas an early embryo has fundamentally different features (Fig. 3), such as a spherical head with relatively large eyes. We described skulls of three ontogenetic stages and adult Suncus murinus, and examined the development in five embryonic and three postnatal stages by analyzing the linear measurements of the cartilaginous and bony skulls. Materials and Methods Relatively few studies on skull development of shrews have been reported (Parker, 1885; Amback Christie-Linde, 1907; Roux, 1907; de Beer, 1929), largely because the shrew is difficult to breed and maintain in captivity. We used musk shrews supplied by the Central Institute for Experimental Animals (Kanagawa, Japan), which keeps a colony of animals originally from Okinawa, Japan, and Tainan, Taiwan. The gestation period of S. murinus is 30 or 31 days. One 20-day and three 25 -day embryos, one newborn, one 9-day, and one 14-day infants were fixed in 4% formaldehyde, stained with 0.01 % Alcian Blue and 1 % Alizarine red-S, and then cleared with 0.5-1 % KOH and glycerol (Kelly and Bryden, 1983). Salivary glands, brown fat, and viscera, except eyes, were removed from skinned animals before staining and clearing. Two other embryos, which were larger than the 25-day embryo, were obtained from pregnant shrews captured in Taiwan. We estimated that the ages of these two embryos were 27 and 29 days. Five macerated skeletons were made from adult specimens which were trapped in Taiwan. The exact ages of these adults were unknown, but we plotted the adult measurements against 68 days post fertilization because Suncus is reported to reach adult weight between 30 and 40 days after birth (Shigehara, 1985). Records and images of specimens were retained as a database (Masuda and Yohro, 1990) in a microcomputer. To quantify the growth of the skeletal system in S. murinus, a series of linear measurements was taken on lateral- and dorsal-plane photographs of cleared specimens and of dried skulls. To minimize distortions of measurements from photographs due to three-dimensional specimens, three photographs were taken in a stable condition at different times, dimensions were measured five times for each photograph, and the averages were calculated. Skull measurements, defined below and illustrated in Fig. 1, were defined in lateral view unless otherwise noted. Abbreviations are: GL, the greatest length of skull (distance between anterior tip of the snout and posterior end of the occipital bones); CBL, condylobasal length (distance from anterior edge of premaxilla to the most posterior projection of the occipital condyles); CB, cranial breadth (the greatest width of braincase, as seen in dorsal view); H, height (maximum height); E, diameter of an eye ball; CC, diameter of the cochlear capsule, as seen in ventral view; PG, postglenoid length (distance between posterior end of the occipital bones and glenoid fossa); FL, bony facial length (distance between anterior tip of the snout and posterior edge of the infraorbital foramen); ML, the distance between the tip of the lower incisor and posterior end of the angular process; HP, the greatest height of the mandible. More precise measurements could be taken on cleared skulls during the prenatal period by identifying six specific points (Fig. 2): AT, anterior tip of the snout; ET, the posterior end of the ethmoturbinale; AH, ala hypochiasmatica; PA, processus alaris; E, portion of the basal plate where the distance between the cochlear capsules is the narrowest; PE, posterior end. Results Skeletal Development Twenty-Day Embryo. — In the 20-day embryo (Fig. 3), the basic cranial structures have already been chondrified. The central stem has already chondrified with no visible separation between parts. The anterior nasal septum unites posteriorly with the parachordal plate, which is perforated by the foramen hypophyseos. The nasal septum is narrow and the parachordal region broadened in width posteriorly. The medial end of the ala hypochiasmatica, which will join the ala orbitalis in later 171 172 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 stages, and the processus alaris project on each side of the central stem. The medial end of the ala orbitalis has two branches, pila praeoptica and pila metoptica, which end freely. The former is short and relatively broad, and the latter is slender and directed to the ala chiasmatica. However, a considerable gap remains between the pila metoptica and the ala chiasmatica. Spherical cartilago pterygoideus is situated anterior to the processus alaris on both sides. The ala temporalis is chondrified anterolaterally to the processus alaris separately as a Y-shaped slender structure, but there remains a wide gap between the arms of the “Y.” The cochlear capsule connects anteriorly with the commissura alicochlearis, which unites with the processus alaris to form the lateral border of the carotid foramen. In lateral view, the lamina parietalis lies dorsal to the otic capsule and connects anteroventrally to the otic capsule with the associated commissura parietocapsularis and posteriorly to the pila occipitale with its commissura parieto-occipitalis. However, there is no connection between the lamina parietalis and the ala orbitalis in the 20-day embryo. The tectum nasi (nasal roof) has already chondrified and fully continued ventrally to the paries nasi, whereas the paries nasi (wall of the nasal capsule) and the solum nasi (nasal floor) are not completely developed. In the solum nasi, the lamina transversalis anterior, cartilago paraseptalis, and the lamina transversal is posterior are visibly separated from anterior to posterior. The lamina transversalis anterior connects anteriorly to the nasal septum, whereas the lamina transversalis posterior connects laterally to the paries nasi. The middle portion of the nasal skeleton has practically no floor, thus resulting in an extensive fenestra basalis. Posteroventral to the nares, three pairs of fringes project from the midventral line. On the dorsal aspect, the processus alaris superior projects ventrocaudally from the cupula anterior. The paries nasi connect to the ala orbitalis and with the associated commissura orbitonasalis. The cochlear capsule is almost chondrified, but the canalicular part has chondrified only the portion surrounding the semicircular canals. The basal plate and the cochlear capsule are connected to the thin commissura basicapsularis anterior and the commissura basicapsularis posterior. The caudolateral comer of the parachordal plate gives rise to the pilae occipitale. Relatively broad cartilaginous plates lie on the upper ends of the pila occipitale. We have not identified a separate chondrification center of this structure, tentatively identified as the supraoccipital cartilage, which is separated from the otic capsule by the fissura occipito-capsularis superior. The supraoccipital cartilage on both sides is medially suspended by a slender cartilaginous thread, which originates at the flexion between hindbrain and midbrain. This slender bridge, which can be observed only for a short period, soon disappears at the middle portion but remains as a slender process on both sides from the supraoccipital cartilage. The premaxilla, the maxilla, the frontal , the parietal, and the squamosal are already present. The squamosal has begun ossifying from the anterior portion, which contributes to the articular facets. The lower jaw is first formed by the Meckel’s cartilage, which articulates with a small incus, then the second, osseous jaw forms laterally to the Meckel’s cartilage, starting anteriorly. The Meckel’s cartilages unite anteriorly into a pointed synchondrosis. The stapes is chondrified in most parts except at the basis stapedius. A slender secondary cartilage is present at the site of the coronoid process of the mandible. [ Twenty-Five-Day Embryo.— 'Qy 25 days, further chondrification and ossification of the embryo has occurred (Fig. 4). A pair of spherical cartilages is found between the pila metoptoica and the central stem, and the processus alaris has grown laterally to join with the commissura alicochlearis, the i processus ascendens, and the processus pterygoideus of the ala temporalis. A small slit, which does not stain well with Alcian blue, can be seen between the ala temporalis and the other two : cartilaginous elements. The median head of the ala temporalis gives rise to a tiny forward projection. The cartilago ' pterygoideus attaches dorsally to the ossified pterygoid dorsally. A small secondary cartilage is evident at the anterior portion of the pterygoid. The commissura basicapsularis is separated by a foramen into commissura basicapsularis anterior and commissura basicapsularis medialis. The premaxilla and maxilla expand their ossifications and the > underlying cartilaginous paries nasi are becoming resolved. The ; sulcus of the nasolacrimal duct remains between the lamina ' transversalis anterior and the paries nasi. The ethmoid plate • starts to chondrify from the tectum nasi downwards, and the i turbinals have begun developing. All dermal bones, nasal, palatine, vomer, lacrimal, and . pterygoid, are ossifying. Chondral ossification has begun in the ■ exotympanic, basioccipital, and exoccipital. The supraoccipitale ; has already ossified by 25 days; however, whether it ossifies ; cartilaginously or dermally is unclear. In mammals the i supraoccipital is supposed to be ossified from a broad ^ cartilaginous band which makes the dorsal border of the ? foramen magnum, whereas in Suncus a series of cartilaginous j islands is seen in the 20-day but not in the 25 -day embryo. The ' ossification of the exoccipitale extends to the lateral edge of the ' hypoglossal foramen. Slender secondary cartilages are seen at ' the tip of coronoid process, condylaris, and angularis of the mandible. Upper and lower incisor development are visible as calcification of dentin. Newborn.— \n the skull of newly bom musk shrew (Fig. 5), most parts have ossified and the cartilaginous portion has degenerated, but borders of most bones still can be distinguished. The following structures remain cartilaginous in this stage: nasal skeleton, interorbital septum, parietal plate, canalicular part of the otic capsule, commissura basicapsularis medialis, most of the styloid process, connection between the ! processus alaris and the alisphenoid, part of the basal plate ■ between the basisphenoid and the basioccipital, commissura [ basicapsularis posterior, and base of the pilae occipitale between j the basioccipital and the exoccipital. The cartilaginous nasal capsule is resolved on its dorsal surface because of the ossification of the frontal and the parietal, whereas the turbinal area and the anterior rostrum remain cartilaginous. The parietal plate degenerates anteriorly and ossifies independently to form a part of the occipital bones (indicated by an asterisk in Fig. 6), | which is covered laterally by the squamosal and visible only in the separated occipital bones of a macerated skull. The j 1994 MASUDA AND YOHRO — Development of Skull of Suncus murinus 173 Meckel’s cartilage is still present and articulates with the incus; however, anterior to the tooth row the cartilaginous bar has disappeared from the lingual surface of the mandible. Secondary cartilages of the mandible broaden and form the distal ends of the three processes. I Ossifications are seen in the basisphenoid, alisphenoid from the ala temporalis, orbitosphenoid from the ala orbitalis, cochlear capsule, goniale, crus longum of incus and collum ! mallei, and the tips of turbinals. The long and slender hamulus i of the pterygoid projects posteriorly. The ethmoid plate begins I ossifying at the middle. The squamosal has two facets to I articulate with the coronoid process of the mandible. TTie I goniale is a small spicule of bone lying close to the ventral border of the Meckel’s cartilage. The ossification of the dorsal border of the foramen magnum, presumably the supraoccipital, proceeds radially and anteriorly. The interparietal is absent, as is any remnant of the jugal. Adult. — Only features that differ between embryos and adults will be mentioned here because Sharma (1958), Dotsch (1982, 1983fl, 1983/?) and Inamura et al. (1984) have already described the adult skull of Suncus murinus in great detail. In the adult skull most bones are fused and only a few sutures can be recognized, such as the spheno-occipital, the parieto-occipital, I and the parieto-squamosal sutures. The general shape of the skull is flat and long (Fig. 6), the bones are thick, and the sagittal crest, the lambdoid crest, the lateral crest, the zygomatic process of the maxilla, and the paroccipital process are well-developed. There is a great gap between the squamosal and the zygomatic process of the maxilla, and the orbital fossa ; and the temporal fossa are confluent. The jugal, the zygomatic I arch, and the auditory bulla are absent, and the tympanic I remains a ring throughout life. The first upper incisor develops 1 a large, hook-like crown and the comparable lower incisor is also well-developed and curved upward. A part of the cartilaginous nasal skeleton projects anterior to the premaxilla. The prominent coronoid process of the mandible flares posterodorsally and laterally, whereas the angular process extends posteriorly as a delicate spicule. The condylar process is short with ventral and dorsal facets. Differential Growth Rate During Ontogeny Most measurements of the skull increased linearly with age, and reached 90% of the adult value by 14 days after birth. Measurements were plotted against age in Fig. 7. In the 20-day embryo and in the newborn, most cranial values (greatest length, condylobasal length, cranial breadth, postglenoid length, and facial length) were about 20% and 55 % of adult values, respectively. Regressions of cranial measurement (CBL, CB, H) against age showed that condylobasal length increased with age with a slope of 0.80, whereas the cranial breadth and cranial height showed much lower slopes of 0.34 and 0. 14, respectively. The diameter of the eye ball achieved 59% and 80% of the adult size in the 20-day embryo and in the newborn respectively, whereas auditory organs (CC and TT) showed about 37 % and 88% of development by that time. The slope of the regression line of the eye was very small (0.017), but the eye continued to increase in size until adulthood. The mandibular measurements (ML and HP) were 10% and 50% in the 20-day embryo and the newborn, respectively. The slope of the regression line of the length of the mandible against age was 0.37, which was the same as that of the bony facial length of the skull. The height of the mandible had a slope of 0.27, which was twice that of the cranial height. The postglenoid length was only 63 % of the adult value in the 14-day infant, although it had showed similar percentages as the other skull measurements in the 20-day embryo and newborn. The proportion of PG to the condylobasal length decreased from 50% in the 20-day embryo to 25% in the 14- day infant, but then increased again to 38% in the adult. In the postglenoid region, the proportion behind the commissura basicapsularis medialis decreased from 16.3 % to 1 1.9% between the 25 -day embryo and the newborn (Fig. 8). Growth rates of different parts of the cartilaginous basicranial axis varied extensively (Fig. 8). Between the 20-day embryo and the 25-day embryo, every part except the facial region, which is the distance between AT and ET, increased in proportion to total length, whereas parts of the facial region decreased. However, between the 25-day embryo and the newborn, only the distance between AT and ET increased its proportion (from 0.4% to 1.1%) and contributed to the lengthening of the skull. By contrast, the proportion of bony facial length to condylobasal length consistency remained about 50% during development. Discussion According to previous studies on the growth of S. murinus (Dryden, 1968; Shigehara, 1980, 1985), early postnatal development is rapid, but this animal is thought to be bom in a relatively undeveloped condition (Shigehara, 1985). We have shown here that most linear measurements have reached 60% of adult values in newborn and 90% values by postnatal day 14. However, not all parts of the skull appear to grow at an equivalent rate. Sensory organs such as the eyeball and cochlear capsule began developing early and also stopped growing earlier than other skull features. In the adult, eye balls are small (diameter of 1.64 mm in the adult), the auditory region lacks bony bullae, and the tympanic rings are exposed without any bony supports. By contrast, the mandible starts its growth later than the skull, but develops well. Markedly long and slender angular processes and well-developed coronoid processes were observed in the adult. Among the measurements, the height of the mandible did not reach 90% of the adult value by 14 days after birth, which means that the coronoid process continues to increase its size after postnatal day 14. This observation agrees with the experimental data, which showed that the development of the coronoid process depends on the external stress of the masticatory muscles (Washburn, 1947). The retardation of the growth of the mandible in early postnatal stages can be explained by the retardation of the development of the masticatory muscles, because in newborn Suncus, ratios of the masticatory muscles against those of the adult are very low (Yamada and Yohro, 1988); 1.46% (masseter muscle), 1.43% 174 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY (temporal muscle), 3.16% (medial pterygoid muscle), 7.53% (lateral pterygoid muscle), 2.74% (digastric muscle), compared with those of Mwi' (house mouse); 8.68, 10.76, 10.55, 13.59, and 21.00%, respectively. Feamhead et al. (1955) showed that the ratio of the postglenoid length against the condylobasal length is markedly high in the Soricidae, with an average of 41.84%. In this study, we showed that the ratio of the postglenoid length against the condylobasal length decreased from the 20-day embryo to the 14-day infant but increased thereafter. Postnatal growth of the postglenoid length is more pronounced than that for the face. The increase in length of the postglenoid beyond 14 days after birth does not depend on the elongation of the cranial base, but on the growth of the postcochlear region. The paroccipital process developed as an insertion for the pars cephalica of the trapezius muscle. The squamosal and the parietal are elongated and contributed to increase the area for insertion of temporal and masseter muscles. Consequently, different parts of the skull develop in different stages — the sensory organs developed in the prenatal stages, the skull height and the basicranial region elongated around birth, and bony additions occurred in the postglenoid region and mandible after postnatal day 14. In newborn animals, skull length, skull height, and postglenoid length have reached about 50% of adult values. The width and facial length, in contrast, were 61 % and 68% of their adult size, respectively. Growth in the transverse direction is limited after birth and, as a result, the skull gains a more slender appearance as development progresses toward adulthood. Acknowledgments We thank T. Hijikata (Department of Anatomy, Tokyo University) for providing cleared postnatal specimens, S. Kuratani (Department of Biochemistry, Baylor University) for corrunents on the description of the chondrocranium of the musk shrew, T. Sakai (Department of Anatomy, Juntendo University) for suggestions in the early stages of this study, and A. W. Crompton (Museum of Comparative Zoology, Harvard University) for comments. Literature Cited ARNBACK CHRISTIE-Linde, A. 1907. Der Bau der Soriciden und ihr Beziehungen zu anderen Saugetiere. Gegenbaurs morphologisches jahrbuch, 36:463-514. NO. 18 ? i DE Beer, G. R. 1929. The development of the skull of the shrew. Philosophical Transactions of the Royal Society of London, Series B. Biological Sciences, 217:411-480. j; Dotsch, C. 1982. Der Kauapparat der Soricidae (Mammalia, Insectivora). Funktionsmorpholgische Untersuchungen zur ' Kaufunktion bei Spitzmausen der Gattungen Sorex Linnaeus, ! Neomys Kaup und Crocidura Wagler. Zoologische Jahrbucher, ■ Abteilung fiir Anatomie und Ontogenie der Tiere, 108:421-484. 1983a. Das Kiefergelenk der Soricidae (Mammalia, Insectivora). Zeitschrift fiir Saugetierkunde, 48:65-77. 1983i. Morphologische Untersuchungen am Kauapparat der Spitzmaus Suncus murinus L., Soriculus nigrescens G. und Soriculus caudatus H. (Soricidae). Saugetierkundliche Mitteilungen, 31:27-46. j Dryden, G. L. 1968. Growth and development of Suncus murinus in captivity on Guam. Journal of Mammalogy, 49:51-62. Fearnhead, R. W., C. C. D. Shute, and a. D. a. Bellaris. j 1955. The temporo-mandibular joint of shrews. Proceedings of the Zoological Society of London, 125:795-806. Inamura, G., M. Yoshizawa, and S. Oda. 1984. Osteology of the musk shrew (Suncus murinus). Anatomischer Anzeiger, 155:131-141. ( Kelly, W. L., and M. M. Bryden. 1983. A modified differential j! stain for cartilage and bone in whole mount preparations of i mammalian fetuses and small vertebrates. Stain Technology, | 58:131-139. Masuda, Y., and T. Yohro. 1990. A database of embryological i terms and figures of mammalian skulls. Unpublished master’s I thesis. University of Tokyo, Tokyo, Japan, 25 pp. Parker, W. K. 1885. On the structure and development of the skull | in the Mammalia II & III. Philosophical Transactions of the Royal i Society of London, Series B. Biological Sciences, 176:121-275. Roux, G. 1907. The cranial development of certain Ethiopian 1 1 “insectivores” and its bearing on the mutual affinities of the group. ‘ Acta Zoologica, 28:165-397. Sharma, D. R. 1958. Studies on the anatomy of the Indian i insectivore, Suncus murinus. Journal of Morphology, 102:427-541. Shigehara, N. 1980. Epiphyseal union and tooth eruption of the i' Riukiu house shrew, Suncus murinus, in captivity. Journal of Mammalogical Society of Japan, 8:151-159. 1985. Some characteristics of growth and development of the | .1 Soricidae. Pp. 68-76, in Suncus murinus, Biology of the j Laboratory Shrew (K. Kondo, S. Oda, J. Kitoh, K. Ohta, and G. , Isomura, eds.), Japan Scientific Societies Press, Tokyo, Japan. Washburn, S. L. 1947. The relation of the temporal muscle to the \ form of the skull. Anatomical Record, 99:239-348. Yamada, K., and T. Yohro. 1988. Comparative observations on the ij development of the masticator muscles. Acta Anatomica Nippon, [ 63:384. ' 1994 MASUDA AND YOHRO— Development of Skull of Suncus murinus 175 CQ O a. -s Fig. 1.— An adult skull of Suncus murinus from lateral view with specific points for measurements. See text for abbreviations. Fig. 2. — A chondrocranium of the 20-day embryo of Suncus murinus from dorsal view with specific points for measurements. See text for abbreviations. 176 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 o Fig. 3. Skull of a 20-day embryo of Suncus murinus. Drawings of a cleared specimen in lateral (A), dorsal (B), and ventral (C) I aspects. 1994 MASUDA AND YOHRO— Development of Skull of Suncus murinus 111 fiss. basicaos. ant. fiss. basicaps. post. Fig. 4.— Skull of a 25-day embryo of Suncus murinus. Drawings of a cleared specimen in lateral (A), dorsal (B), and ventral (C) aspects. Right parietal and the anterior part of Meckel’s cartilage are removed. 178 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 Fig. 5.— Skull of a newborn Suncus murinus. Drawings of a cleared specimen in lateral (A), dorsal (B), and ventral (C) aspects. Dorsal bones except a part of left side are removed. 1994 MASUDA AND YOHRO— Development of Skull of Suncus murinus 179 E Fig. 6. — Skull of an adult male Suncus murinus. Drawings of a macerated skull in lateral (A), dorsal (B), ventral (C), and anterior (D) aspects. E; Posterolateral aspect of the right side of an occipital complex in a young Suncus. The asterisk indicates the anterior wing of an occipital bone, which is covered by the squamosal laterally and cannot be seen from outside. The anterior of the occipital bone can be observed only in the macerated skull of a young animal. 180 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 •Q* C.R. a -0- CBL days -o- E CC -a- TT Fig. 7. — Linear measurements (mm) of skulls of Suncus murinus against age (gestation day). CB, condylobasal length; CC, diameter of a cochlear capsule; CBL, condylobasal length; C.R., cranio-rostral length; FL, facial length; GL, greatest length; E, diameter of the eye; HP, mandibular length; ML, mandibular height; PG, postglenoid length; TT, diameter of a tympanic ring. B AT-ET ET-AH m AH-PA PA-E □ E-PE 20 day emb. 25 day emb. Stage newborn Fig. 8.— Proportions long axial segments in Suncus chondrocranium. AH, ala hypochiasmatica; AT, anterior tip; E, eye; ET, ethmoturbinal; PA, processus alaris; PE, posterior end. CHARACTERISTICS OF THE BREEDING SEASON IN THE COMMON SHREW (SOREX ARANEUS): MALE SEXUAL MATLHATION, MORPHOLOGY, AND MOBILITY Paula Stockley'’^ and Jeremy B. Searle^’^ 'Department of Zoology, University of Oxford, South Parks Road, Oxford 0X1 3PS, United Kingdom; ^Department of Environmental and Evolutionary Biology, University of Liverpool, P. O. Box 147, Liverpool L69 3BX, United Kingdom; ^Department of Biology, University of York, York YOl 5DD, United Kingdom Abstract A population of Sorex araneus near Oxford, England, was studied during the breeding season of 1990. Several phenotypic traits that may potentially influence male mating success were monitored among marked individuals. These included morphological characteristics such as body size, testes size, sperm count, seminal vesicle size, and lateral gland length; also, timing of sexual maturation, relative mobility, and parasite load were examined. Male body mass increased significantly between March and April, and was strongly correlated with body length in May, when both body mass and length were also correlated with seminal vesicle mass (an indicator of androgen activity). Body length in May was negatively correlated with gut parasite load. Early maturing males were found to develop the largest testes but had relatively small lateral glands in May. Male shrews were significantly more mobile than females during the breeding season; individual mobility was associated with body size in both sexes, and with testes size among males. Introduction Little is known about the mating system of the common shrew (Sorex araneus). Behavioral studies in the field are essentially limited to an old fashioned live-trapping methodology (the species is too small for radio collars), and in captivity, common shrews are not easily maintained and bred (Mercer and Searle, 1994). Nevertheless, common shrews are unlikely to be more difficult to study than any other shrews of the subfamily Soricinae, which comprises some 102 species (Hutterer, 1985). The common shrew itself is one of the most abundant small mammals in the northern Palearctic region. Furthermore, this i species is one of the few mammals for which it has been j demonstrated unambiguously that, in nature, multiple paternity : can occur within a single litter (Searle, 1990; Tegelstrom et al. , ‘ 1991). I In this paper we review what is known about the breeding I season of the conunon shrew, based largely on long-term mark- release-recapture studies (for behavioral analysis), and i systematic snap-trap collections (for studies of reproductive I physiology). Attributes of the breeding season in this species vary somewhat according to geographical location; the emphasis in this paper, however, is on common shrews in southern Britain. We also present the results of our own initial field studies which are more specifically directed towards an understanding of the mating system of the common shrew. I Given that multiple paternity occurs in this species, we are I particularly interested in competition among males for mates ! and the morphological and behavioral characteristics which may I maximize male mating success. i Characteristics of the Breeding Season in the Common Shrew The common shrew, a seasonal breeder, has a long immature phase when males and females are extremely similar in their behavior and morphology (Michielsen, 1966; Buckner, 1969; Searle, 1985), followed by a relatively short adult phase. when sexual dimorphism becomes apparent. Thus, in Britain, almost all individuals become sexually mature in the spring following their birth, and die within the same year (e.g., Adams, 1910; Middleton, 1931; Brambell, 1935). With a gestation period of 20 days, a lactation period of about 23 days, and a postpartum estrus (e.g., Dehnel, 1952; Searle, 1984a), females could potentially produce numerous litters of up to ten young (see Mercer and Searle, 1994) during the March-October breeding season (Middleton, 1931; Brambell, 1935; Crowcroft, 1957; Michielsen, 1966). In reality, however, it is the first two or three litters which are the most important; adult females become scarce and erratic breeders thereafter (Brambell, 1935; Tarkowski, 1957). Onset of Sexual Maturity.— It is generally agreed that male common shrews attain sexual maturity (as defined by presence of sperm in the testes) three to four weeks before females have their first recognized ovulation (Brambell, 1935; Crowcroft, 1957; Michielsen, 1966; Skaren, 1973). In southern Britain, some males reach maturity in March (Brambell, 1935), and first conceptions occur in late April or May (Brambell, 1935; Crowcroft, 1957; Searle, 1984J>) with a high degree of synchrony of the first estrus within a particular locality (Brambell, 1935; Tarkowski, 1957). Features associated with the onset of sexual maturity include an increase in body size. (Adams, 1910; Brambell, 1935; Michielsen, 1966; Shillito, 1963a), and onset of the spring molt (Crowcroft, 1957; Shillito, 1963a; Skaren, 1973). In addition, sexual maturity of males is characterized by rapid growth of the testes and accessory sexual organs (Middleton, 1931; Brambell, 1935) and maximal development of the lateral glands (Crowcroft, 1957; Michielsen, 1966; Searle, 1985). The body weight at which sexual maturity is reached has been reported as ranging from 7.7 g to 9 g (Middleton, 1931; Brambell, 1935; Crowcroft, 1957; Shillito, 1963a; Michielsen, 1966). Adult males and nonpregnant parous females ultimately attain a body mass of around 12 g (Brambell, 1935). Adult 181 182 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 body length ranges from about 70-85 mm among common shrews in Britain (Crowcroft, 1957). From birth until the end of February the testes are minute, weighing less than 5 mg combined (Brambell, 1935). They develop rapidly, however, and become evident externally as a bulge at the base the tail by March or April (Crowcroft, 1957). By the end of April the testes are fully grown (Brambell, 1935) with a combined fresh weight in excess of 200 mg (Middleton, 1931; Garagna et al., 1989). The increased androgen production associated with development of mammalian testes is known also to influence development and activity of sebaceous scent glands (Clarke and Frearson, 1972; Yahr et al., 1979). In the common shrew, lateral glands (situated on either side of the body) develop into large active structures at sexual maturity in the male (but not the female: Searle, 1985), producing a characteristic odor (Crowcroft, 1957; Michielsen, 1966; Skaren, 1973). In older males the skin around the lateral glands and over the testes may become bare (Searle, 1985; Mercer and Searle, 1994). Range and Movements. — In addition to these physical changes associated with sexual maturity, there are also modifications of behavior. In particular, changes in the range and movements of individuals have been revealed by live- trapping studies. Over the winter months, immature animals generally have nonoverlapping and stable home ranges of approximately 30 m diameter (Shillito, 1963/j; Michielsen, 1966; Buckner, 1969). The dimensions of the home range may vary according to vegetation type, but within one population, males and females have ranges of similar size (Buckner, 1969; Michielsen, 1966). With the onset of sexual maturity, females do not drastically modify their movements, but males become very active and may abandon their original home ranges (e.g., Shillito, 1963^; Michielsen, 1966; Churchfield, 1980). During spring and summer the male shrews are generally regarded as nomadic. Shillito (1963Z?), for example, reports a range of up to 144 m in diameter, and the majority of males on her study site disappeared from the area, temporarily or permanently, as a result of their ranging behavior. It is uncertain, however, whether males are truly nomadic, or merely establish large, stable home ranges. Determinants of Male Mating Success On the basis of current understanding of morphology, development, and behavior of the common shrew during the breeding season as reviewed above, several characteristics emerge as potentially important in determining male mating success. One such factor is individual timing of maturation relative to other males. Among other species, advantages established among early maturing individuals (for example, with respect to body size) may be maintained or exaggerated, as competition commonly prevents compensatory growth among those maturing later (Clutton-Brock, 1988). Relative adult body size may be an important determinant of mating success among common shrews if the aggressive behavior between adult males in captivity (Moraleva, 1989) is expressed during competition for mates. Although less well-studied among mammals than body size, variation in testes size may also influence the outcome of male competition for mates. It is now realized, for example, that male fertilizing capacity may be limited under certain circumstances (Dewsbury, 1982; Small, 1988) and that sperm competition is widespread among mammals (Ginsberg and . Huck, 1989; Moller and Birkhead, 1989). The testes of the common shrew are considerably larger than expected for a typical mammal of its size, implying that sperm competition has been influential in their evolution (Kenagy and Trombulak, 1986). Testes size or sperm count (or both) may therefore correlate with insemination success. The activity of the lateral glands may also have some ' influence on the mating success of male common shrews. ■ Although the function of the lateral glands is uncertain in j shrews, sexual selection has apparently influenced the relative size and activity of particular (sexually dimorphic) sebaceous i glands in other mammals (Jarmett, 1986). Male mobility may ; also be a sexually selected trait when females are widely > distributed in space (Schwagmeyer, 1988), as is typically the ; case among populations of the common shrew. , Here, in a step toward a more detailed understanding of the mating system of the common shrew, we explore the way in , which these variables are related, based on our preliminary study of common shrews in a population in southern Britain. Study Area and Methods The study site, a 2-ha strip of rough grassland on Little Wittenham Nature Reserve, Oxfordshire, England (Ordnance Survey grid reference: SU567/932), was bounded by a river to ' the north and a road to the west, grazed meadow to the south and a larger area of mixed woodland to the east. In addition to common shrews, several species of small mammals were recorded at the site: moles {Talpa europaea), pygmy shrews (Sorex minutus), woodmice (Apodemus sylvaticus), and bank voles (Clethrionomys glareolus). Likely predators at the site included a kestrel (Falco tinnunculus), seen regularly over the area, and domestic cats from nearby houses. The first period of live trapping was conducted between ' 19-30 March 1990. The study area was divided into two | overlapping areas, each of which was trapped for five * consecutive days using 96 numbered Longworth traps provided with hay bedding and set in a grid at approximately 10-m intervals. Puparia of Calliphora (killed by freezing, see Little and Gumell, 1989) were used as bait. Traps were locked open at night, except for one night during each five-day trapping period, when traps were baited additionally with approximately 15 g of moist minced ox heart. During the day, traps were visited at 1-2 h intervals between 0700 h and 1900 h. When set overnight, traps were emptied at 0530 h. No trap deaths occurred with this regime. All common shrews were weighed, sexed, and the body length and tail length measured. These latter measurements were difficult to make on live animals under field conditions, and although results were consistent, our measurements of body length of live animals are not necessarily directly comparable with those of dead animals. The degree of sexual maturation 1994 STOCKLEY AND SEARLE — Breeding Season in the Common Shrew 183 and stage of molt were noted. Each animal was individually marked by toe clipping and released. Males were subsequently ranked with respect to timing of sexual maturation. Ranking was based on relative size of the testes bulge, stage of spring molt, and body size during late March. In general, these parameters were closely associated with one another. Shared ranks were assigned where no clear differences were evident. There were insufficient data to calculate accurate home-range areas but the relative mobility of shrews captured four or more times was estimated using three parameters; the farthest distance between any two (of four or more) capture points (FD), the farthest distance between any of the first four capture points (FD[4]), and the mean distance moved between four or more consecutive captures (MDM). In addition, distances moved between consecutive captures were classified either as zero (animal recaptured in same trap as previous capture), short (animal recaptured 10-19 m from previous capture point), medium (animal recaptured 20-29 m from previous capture point), or long (animal captured 30 m or more from previous capture point). Traps were removed on 30 March and the population left undisturbed until around the time of first estrus, which was determined by the first occurrence of nape scars (an indication of mating, see Crowcroft, 1957). Previous studies in Oxfordshire indicate a high level of synchrony among females within a site with respect to first conception (Searle, 1984ft). During April (23-27), traps were replaced in their original positions over four consecutive nights, again baited with moist minced ox heart and opened at 0530 h. All male common shrews captured were weighed, their position of capture noted, and then transferred to captivity. These animals were sacrificed between 28 April and 7 May to provide detailed information on their reproductive condition, body size, lateral gland size, and parasite load. Each was weighed, and measurements of body length, tail length, combined lateral gland length, and fresh mass of the testes were recorded. Sperm counts were made on j the right epididymis by a method derived from Searle and I Beechey (1974). The sperm count is given in terms of the total J number of sperm in the caput epididymis. Fresh mass of the seminal vesicle was recorded as an indicator of androgen activity (Grocock and Clarke, 1974). Males were ranked according to relative number and size of cestodes and I nematodes in the intestine. j Statistical Analysis I Parametric distributions were assumed for all characters 1 except maturity and parasite load. Product-moment correlation I coefficients, regression equations, and t-tests were used for I parametric data; Spearman rank correlation coefficients were * used for nonparametric data; and two-tailed tests were used throughout. Results j Morphology During the Breeding Season j During March, 26 common shrews were captured and individually marked. Mean male body mass ( + SE), 8.5 ± 0.2 g (rt = 11), and mean female body mass, 7.5 + 0. 1 g (n = 15), differed significantly at this time (t = 5.11, d.f. = 24, P < 0.001), while body length, 71.2 ±1.8 mm and 71.3 ± 1.5 mm for males and females, respectively, did not. There was no significant correlation between body mass and body length during March for either sex. In April, body masses had increased significantly (males: t = 5.52, d.f. = \1, P < 0.001; females: t = 31.21, d.f - 14, P < 0.001). Furthermore, male and female body masses remained significantly different (r = 3.10, d.f. = 15, P < 0. 01). The mean body mass increase for nine recaptured males over this period was 1.5 ± 0.3 g and the mean body mass overall 10.2 ± 0.3 g (n = 1 1). The mean female mass increase was 1.4 ± 0.2 g and the mean body mass 9.0 ± 0.3 g(n = 9). Correlations between male morphological characters measured throughout the breeding season are presented in Fig. 1 . There are three values for body mass; those recorded in late March, late April, and early May, respectively. The May values (recorded after a short period of maintenance in captivity) are significantly correlated with body mass measured during March (r = 0.76, n = 9, P < 0.05), but not during April (note, however, that the mean body mass values during late April, 9.56 ± 0.22 g, and early May, 10.20 ± 0.33 g, were not significantly different). In contrast with the March data, however, male body masses in May were strongly correlated with their body lengths (r — 0.81, « = 11, P < 0.005). Both body mass and length in May were significantly correlated with seminal vesicle mass (body mass: r = 0.68, n = 11, P < 0.05; body length: r = 0.89, n = 11, P < 0.0005). Testes mass in May was correlated with body mass in April at the time of recapture (r = 0.70, n = 9, P < 0.05) but not with body mass in May after a short period of maintenance in captivity. Testes mass of these sexually mature individuals was not correlated with seminal vesicle mass or sperm count. Lateral gland length was not correlated with adult body size or seminal vesicle mass, but was negatively correlated with body mass increase between March and April (r = 0.80, n = 11, P < 0.05) and positively correlated with sperm count (r = 0.60, n = 11, P < 0.05). Mobility During the Breeding Season Individual shrews were caught between one and 14 times each. The number of captures per trap hour did not differ significantly for males and females. Measures of mobility are based on 16 individuals that were captured on four or more occasions during late March. During March, male shrews were found to be significantly more mobile than females. Mean farthest distance moved between any two of the first four captures (FD[4]) for males was 38 ± 5 m, and for females 20 ± 4 m (r = 2.98, d.f. 1 4, P < 0.01). Mean distance moved between consecutive captures was 25 ± 3 m for males, and 12 ± 2 m for females, {t = 3.65, d.f. = 108, P < 0.001). Mean farthest distance moved between any two capture points (FD) for males was 59 ± 12 m and for females 25 ± 7 m (r = 1.89, d.f. = 14, P < 0.1). Maximum recorded distance moved between consecutive captures for males was 100 m, and for females 43 m. In 184 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 f general, males made a greater proportion of long distance movements (>30 m) between consecutive captures, whereas females were more likely to be captured in the same trap on consecutive occasions (Fig. 2a, b; = 13.27, d.f. = 3, P < 0.01). The mean distance moved between the March and April trapping periods was also calculated for the 18 marked shrews that were recaptured. Again males were found to have moved considerably greater distances than females. The mean distance moved from last capture in March to the first capture in April was 58 ± 1 1 m for males and 20 + 6 m for females {t = 2.66, d.f. = 16, P < 0.02). All females recaptured in April were found within 30 m of the area in which they had been captured during March (although several females were absent in April and may therefore have moved longer distances off the study site). Of nine recaptured males, three were found within 30 m of the area where they were last captured in March. Two of these males were unusual in several respects. The first was the latest male to reach maturity and showed no mass increase over April. In May, this male was relatively small, had relatively small seminal vesicles and testes, but had well-developed lateral glands. The second male was also relatively small, with small seminal vesicles; additionally, this male had the highest gut parasite load among those examined. The greatest distance moved by any recaptured male from its last point of capture was 110 m. Two marked males were not recovered from the study site. Timing of Sexual Maturation Early maturing males were found to have attained the greatest testes masses in May (Spearman rank correlation coefficient — 0.83, n — 9, P < 0.02) but they also had the smallest lateral glands (r, = 0.71, n = 9, P < 0.05). Male maturity rank was not correlated with mobility scores during March or with distance moved between the March and April trapping sessions. Morphology and Mobility There was a strong trend during March for the heaviest males to move the greatest distances between any two of their first four capture points (FD[4]: n = 8, =0.477, = 5.47, P = 0.06). No such trend was found among females. Male body mass (March, April, May) was not correlated with the distance moved between March and April trapping sessions. There was, however, a strong trend for males with the largest testes to move the greatest distances during this period {n = 8, = 0.412, Fj g = 4.93, P = 0.06). Body length was cor- related with farthest distance moved during March (FD) for all individuals (rt = 16, = 0.350, Fj g = 7.52, P < 0.02) and for both males and females when considered separately (Fig. 3a; males: n = %, = 0.544, Fj g = 7. 16, P < 0.05. Fig. 3b; females: ti = S, t^ = 0.784, F, g = 21.74, P < 0.005). Parasite Load No relationship was found between gut parasite load and body mass, testes mass, seminal vesicle mass, or lateral gland length. When one heavily infested shrew was removed from the analysis, however, gut parasite load was negatively correlated with body length (r^ = 0.73, n = 9, P < 0.05). We found no significant correlations between gut parasite load and measures : of male mobility during the breeding season. Discussion Morphological Relationships The values obtained for morphological variables such as body mass, body length, and testes mass in general agree well with those described by previous authors (Adams, 1910; Middleton, 1931; Brambell, 1935; Shillito, 1963a; Michielsen, i 1966). The relationships between seminal vesicle mass and body size suggest that the degree of androgen production may influence growth rates of maturing males and hence adult body size. The lack of significant correlation between testes and seminal vesicle masses, however, indicates that large testes do not necessarily secrete the most androgens. Studies on i chromosomally abnormal shrews also suggest that testes size is a poor indicator of androgen production (Searle, 1984c; Searle ! and Wilkinson, 1986). | Testes mass (measured in May) was related to body mass at the time of recapture (April), but was apparently unrelated to I sperm count. Sperm output is related to testes size in several mammalian species (rams, Abdou et al., 1978; mice. Must musculus, H. Hauffe, personal communication; kangaroo rats, ij Dipodomys ordii, Hoditschek and Best, 1983; rats. Lino, 1972), although this is not always the case (e.g.. Carter et al., 1980). ; Sexual Maturation Individual timing of maturity at the onset of the breeding season was significantly correlated with two morphological s characters of potential importance in determining reproductive success among adult male shrews. Early maturing males tended to develop larger testes but had relatively small lateral glands in May. The significance of testes size has already been discussed. The size of the lateral glands was inversely proportional to i. weight increase between March and April, such that large lateral glands were associated with both late development and ij slow growth rate. Interestingly, lateral gland length was also positively correlated with sperm count in these males. It is difficult to speculate on the significance of results relating to the lateral glands without knowing their function. However, in view of the positive correlation found between early development and I adult testes size, these results were unexpected and as such warrant further investigation. > Relationships Between Morphology and Mobility Our results agree with those of previous studies with respect to general sex differences in mobility during the breeding season. During March, male shrews were found to be generally more mobile than females. Similarly, those males recaptured in April were generally found to have moved greater distances from their previous points of capture than had recaptured i females. I 1994 STOCKLEY AND SEARLE — Breeding Season in the Common Shrew 185 The spatial distribution of female common shrews during the breeding season, coupled with relatively short periods of sexual receptivity (Crowcroft, 1957), suggest mobility to be an important correlate of male mating success. During March we found a strong trend for heavier males to have greater mobility scores (farthest distance moved between any of first four capture points, FD[4]), whereas no such trend was found among females. Furthermore, those males which moved the farthest distances in April had, in general, developed the largest testes. While body mass was associated with mobility during March among male shrews only, another measure of mobility (the farthest distance moved between any two [of four or more] capture points, FD) was significantly correlated with body length among both males and females (body length was unrelated to body mass at this time). Hence, regardless of sex or stage of maturation, longer-bodied subadults ranged farther distances overall than did shorter-bodied individuals (male and female body lengths were not significantly different at this time). If subadults are actively territorial, then such a difference between large and small individuals might feasibly be brought about by dominance relations, raising the question of cause and effect (i.e., do large individuals become dominant and monopolize larger territories, or do dominant individuals become larger as a result of gaining access to larger territories?). Alternatively, there may be no strict territoriality as such, and differences in ranging behavior may simply result from larger individuals having increased foraging requirements. In future studies we aim to measure the reproductive success of individual common shrews in natural populations using the I method of DNA fingerprinting. This information will be related ' to various individual characteristics such as relative mobility, i body size, and timing of maturity, in order to obtain a more I detailed picture of the mating system of the common shrew. j Acknowledgments I This work was supported by the Natural Environment Research Council (P. Stockley) and The Royal Society of I London (J. B. Searle). We thank Dr. R. Buxton for access to I Little Wittenham Nature Reserve; D. Ferrier and C. Deubelbeiss for fieldwork assistance; Dr. D. Macdonald, Dr. J. Clarke, and Professor A. Anwar for advice; and Dr. D. Macdonald, Dr. J. Clarke, and Dr. A. Douglas for comments on the manuscript. Literature Cited Abdou, M. S. S., T. M. Hass UN, and S. A. El-Sauaf. 1978. Testicular and epididymal sperm numbers and related parameters in the developing Awassi ram. Australian Journal of Biological Science, 31:257-266. Adams, L. E. 1910. A hypothesis as to the cause of the autumnal epidemic in the common and the lesser shrew, with some notes on their habits. Memoirs and Proceedings of the Manchester Literary and Philosophical Society, 54:1-13. Brambell, F. W. R. 1935. Reproduction in the common shrew (Sorex araneus Linnaeus). Philosophical Transactions of the Royal Society of London, B225:l-62. Buckner, C. H. 1969. Some aspects of the population ecology of the common shrew Sorex araneus, near Oxford, England. Journal of Mammalogy, 50:326-332. Carter, A. P., P. D. P. Wood, and P. A. Wright. 1980. Association between scrotal circumference, live-weight and sperm output in dairy cattle. Journal of Reproduction and Fertility, 59:447-451 Churchfield, S. 1980. Population dynamics and the seasonal fluctuations in numbers of the common shrew in Britain. Acta Theoriologica, 25:415-424. Clarke, J. R., and S. Frearson. 1972. Sebaceous glands on the hind quarters of the vole, Microtus agrestis. Journal of Reproduction and Fertility, 31:477-481. Clutton-Brock, T. H. 1988. Reproductive Success. Pp. 472-485, in Reproductive Success: Studies of Individual Variation in Contrasting Breeding Systems (T. H. Clutton-Brock, ed.). University of Chicago Press, Chicago, 538 pp. Crowcroft, P. 1957. The Life of the Shrew. Reinhardt, London, 166 pp. Dehnel, a. 1952. The biology of breeding of the common shrew (Sorex araneus L.) in laboratory conditions. Annales Universitatis Mariae C u rie-Sklodo wska , C6:359-376. Dewsbury, D. A. 1982. Ejaculate cost and mate choice. American Naturalist, 119:601-610. Garagna, S., M. Zuccotti, j. B. Searle, C. A. Redi, and P. J. Wilkinson. 1989. Spermatogenesis in heterozygotes for Robertsonian chromosomal rearrangements from natural populations of the common shrew, Sorex araneus. Journal of Reproduction and Fertility, 87:431-438. Ginsberg, J. R., and U. W. Huck. 1989. Sperm competition in mammals. Trends in Ecology and Evolution, 4:74-79. Grocock, C. a., and j. R. Clarke. 1974. Photoperiodic control of testis activity in the vole, Microtus agrestis. Journal of Reproduction and Fertility, 39:337-347. Hodischek, B., and T. L. Best. 1983. Reproductive biology of Ord’s Kangaroo rat (Dipodomys ordii). Journal of Mammalogy, 64:121-127. Hutterer, R. 1985. Anatomical adaptations of shrews. Mammal Review, 15:43-55. Jannett, F. j., Jr. 1986. Sexual selection suggested by relative scent gland development in representative voles (Microtus). Pp. 541-550, in Chemical Signals in Vertebrates 4: Ecology, Evolution and Comparative Biology (D. Duvall, D. M uller-Schwarze , and R. M. Silverstein, eds.). Plenum Press, London, 742 pp. K ENA GY, G. J., AND S. C. Trombulak. 1986. Size and function of mammalian testes in relation to body size. Journal of Mammalogy, 67:1-22. Lino, B. F. 1972. The output of spermatozoa in rats. II. Relationship to scrotal circumference, testis weight, and the number of spermatozoa in different parts of the urogenital tract. Australian Journal of Biological Science, 25:351-358. Little, J. L., and J. Gurnell. 1989. Shrew captures and rodent field studies. Journal of Zoology (London), 218:329-331. Mercer, S. J., and J. B. Searle. 1994. Captive breeding of the common shrew (Sorex araneus) for chromosomal analysis. Pp. 271-276, in Advances in the Biology of Shrews (J. F. Merritt, G. L. Kirkland, Jr., and R. K. Rose, eds.), Carnegie Museum of Natural History Special Publication No. 18, i-x -F 458 pp. Michielsen, N. C. 1966. Intraspecific and interspecific competition in the shrews Sorex araneus L. and Sorex minutus L. Archives Neerlandaises de Zoologie, 17:73-174. Middleton, A. D. 1931. A contribution to the biology of the common shrew, Sorex araneus Linnaeus. Proceedings of the Zoological Society of London, 1931 : 133-143. 186 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 Moller, a. P., and T. R. Birkhead. 1989. Copulation in mammals: Evidence that sperm competition is widespread. Biological Journal of the Linnean Society, 38:119-131. Moraleva, N. V. 1989. Intraspecific interactions in the common shrew Sorex araneus in central Siberia. Annales Zoologici Fennici, 26:425-432. SchwagmeyeR, P. L. 1988. Scramble competition polygyny in an asocial mammal: Male mobility and mating success. The American Naturalist, 131:885-892. Searle, a. G., and C. V. BEECHEY. 1974. Sperm-count, egg fertilization and dominant lethality after X-irradiation of mice. Mutation Research, 22:63-72. Searle, J. B. 1984a. Breeding the common shrew (Sorex araneus) in captivity. Laboratory Animals, 18:359-363. 1984Zj. Nondisjunction frequency in Robertsonian heterozygotes from natural populations of the common shrew, Sorex araneus L. Cytogenetics and Cell Genetics, 38:265-271. 1984c. A wild common shrew (Sorex araneus) with an XX Y sex chromosome constitution. Journal of Reproduction and Fertility, 70:353-356. 1985. Methods for determining the sex of common shrews (Sorex araneus). Journal of Zoology (London), 206:279-282. 1990. Evidence for multiple paternity in the common shrew (Sorex araneus). Journal of Mammalogy, 71:139-144. Searle, J. B., and P. J. Wilkinson. 1986. The XYY condition in a wild mammal: An XY/XYY mosaic common shrew (Sorex araneus). Cytogenetics and Cell Genetics, 41:225-233. Skaren, U. 1973. Spring moult and the onset of the breeding season of the common shrew (Sorex araneus) in central Finland. Acta Theriologica, 18:443-458. Shillito, J. F. 1963a. Field observations on the growth, reproduction and activity of a woodland population of the common shrew, Sorex araneus L. Proceedings of the Zoological Society of London, 140:99-114. 1963(7. Observations on the range and movements of a woodland population of the common shrew, Sorex araneus L. Proceedings of the Zoological Society of London, 140:533-546. Small, M. F. 1988. Female primate sexual behaviour and conception: Are there really sperm to spare? Current Anthropology, 29:81-100. TarKOWSKI, a. K. 1957. Studies on reproduction and prenatal mortality of the common shrew (Sorex araneus L.) II. Reproduction under natural conditions. Annales Universitatis Mariae Curie-Sklodowska, CIO: 177-244. Tegelstrom, H., j. B. Searle, J. Brookfield, and S.. Mercer. 1991. Multiple paternity in wild common shrews (Sorex araneus) is confirmed by DNA fingerprinting. Heredity, 66:373-379. Yahr, P. a., D. Newman, and R. Stephens. 1979. Sexual behaviour and scent marking in male gerbils after castration and testosterone replacement. Hormones and Behavior, 13:175-184. Body mass (April) Body mass (May) Body mass increase (March-April) Body length (March) Body length (May) Testes mass (May) Sperm count (May) Seminal vesicle mass (May) Lateral gland length (May) Body mass (March) Body n.s. mass (April) Body Body n.s. mass (May) mass increase n.s. n.s. (March- April) Body length (March) n.s. n.s. n.s. n.s. Body length (May) n.s. n.s. n.s. n.s. Testes mass (May) n.s. n.s. n.s. n.s. n.s. Sperm n.s. n.s. n.s. n.s. n.s. n.s. count (May) Seminal vesicle m2 (May) n.s. n.s. n.s. n.s. n.s. n.s. n.s. n.s. n.s. n.s. n.s. n.s. n.s. Fig. 1.— Correlations between male morphological characters. Abbreviations: n.s., P > 0.05; *, P < 0.05; **, P < 0.005; ***, P < 0.0005. 1994 STOCKLEY AND SEARLE — Breeding Season in the Common Shrew 187 Fig. 2. — The relative frequency of movements of different distances made by common shrews between consecutive captures in March: a, males and b, females (x^ = 13.27; d.f. = 3, P < 0.01). a Male body length (mm) b Fig. 3. — Relationship between body length of common shrews and farthest distance (FD) moved between capture points (>4) during March: a, males (y = 3.98x - 227, r = 0.544, Fj g = 7.16, P < 0.05), and b, females (y - 2.21x - 134, r = 0.784, Fi g = 21.74, P < 0.005). IfB . ... 'r ■uK . ! - i '.Jr. fp '• -) . ttti :f> Y.ciiswjjen't ' '' /■ ' 1 ,T^.‘ > ^ ., 1986) and (prenatal development only) in Suncus murinus (Sprando et al., 1989). The eye primordium was first indicated by the presence of optic vesicles broadly continuous with the third brain ventricle in the Sorex embryo aged about 1 1 days (gestation lasts 21 days). The optic cup with a closed lens vesicle was present in embryos 13-15 days old. In 1 6-day -old embryos, the retina had begun to differentiate into the primitive neuroblastic and marginal layers. The retinal pigment epithelium layer contained densely packed melanin granules. A small number of nerve fibers appeared in the optic nerve primordium. The , hyaloid artery and the annular vessel system were highly developed at that stage. The inside of the lens vesicle became gradually occluded by extending primary lens fibers. By days 17 or 18, the anterior segment of the eye formed and the eyelids developed. The lens vesicle was filled with primary lens fibers, and secondary lens fibers appeared laterally. Before ; birth, at gestational day 20, the eyelids were fused, the eyeball wall was organized, and the retina had differentiated into three layers. ! The whole eye differentiates very quickly during the first three postnatal weeks. The diameter of the eyeball grows rapidly up to the tenth day after birth, when it reaches about j 80% of the adult ocular dimensions. The choroidal pigment emerges first on day 8, and the adultlike pigmentation appears ij on postnatal days 18-20. The cornea becomes stratified at the | age of 12-16 days after birth. The retinal layers develop progressively during the initial postnatal weeks. On day 13, the il two types of visual receptors are discernible at the light l! microscopic level. At the time of eyelid opening, postnatal days 18-19, the retina is fully formed. The first myelinated axons in | the optic nerve are visible at ten days after birth. At the time of I eyelid opening, only about 25 % of the axons are myelinated. In general, the eye in both Sorex araneus and Suncus \ murinus develops normally; no part of it appears to he i developmentally retarded. Auditory System I Hearing Physiology , No audiological (behavioral or electrophysiological) studies i have been conducted to directly and explicitly ascertain the i characteristics of hearing (frequency tuning, hearing range, i sensitivity, resolution, and sound localization capabilities) in any soricid species. In the absence of direct data, some characteristics of hearing may be deduced from general bioacoustic principles, i characteristics of vocalization, behavioral ecology, and i functional morphology of the auditory apparatus. This deduction " is based upon the generally accepted correlation between tuning of the auditory system and all of the aforementioned aspects. | These factors will be discussed sequentially. Physical Constraints What an animal should hear, and what it can hear are determined and constrained by the physical principles of acoustics. For production of high tones, small resonance cavities (larynx, mouth, nose) are needed, whereas voluminous organs are required to produce low tones. Similarly, smaller and lighter receiving structures, including the ear canal, middle i j ear cavity, ear drum, and auditory ossicles, should resonate at |i higher frequencies than the larger structures of larger animals. |i Small mammals like shrews would be expected to have their smaller auditory as well as vocalization organs tuned to higher ! frequencies. 1994 BRANIS AND BURDA — Vision and Hearing in Shrews 191 Small mammals have small heads with a short interaural distance. Consequently, they are typically not able to localize sound due to the negligible difference in time of arrival of a stimulus at the two ears, and therefore make use of the difference in frequency-intensity spectra of a sound reaching the two ears (Heffner and Masterton, 1980). The frequency- intensity difference cue is available even to a small mammal provided that it can perceive frequencies high enough to be effectively shadowed by the head and pinnae (Heffner and Masterton, 1980). High frequencies are, however, more attenuated in the environment than low frequencies. Hence it is advantageous for mammals to have their hearing tuned to the lowest frequency that can be accurately localized. For example, in Sorex araneus, with an average interaural distance (measured around the head) of about 22 mm, the frequency of a corresponding wave length which could be effectively shadowed by the head and thus localized, is about 15 kHz (Table 2). Vocalization The vocalizing and auditory systems in each species have coevolved and are tuned to each other. This also must be true to permit active acoustic orientation or echolocation. Consequently, frequency characteristics of calls used for intraspecific communication and echolocation should provide us with information on the frequency tuning of hearing. Hutterer and Vogel (1977) demonstrated in some crocidurine shrews that the species-specific main frequency of defensive, or fright calls (32-14 kHz) was inversely related to the body size (9-120 g). It is not known whether these calls are aimed at conspecifics only, and therefore whether such calls serve as a correlate of hearing capabilities. However, there is no such problem with courtship calls. The main frequency of these calls is about 10 kHz in Neomys fodiens, and about 15 kHz in Crocidura russula (Hutterer, 1978). Unfortunately some earlier papers (e.g., Hutterer, 1976), though valuable, cannot be considered here since ultrasonic calls were not recorded. Frequencies of sounds considered to be echolocation calls ranged from 30-60 kHz in three Sorex species (Gould et al., 1964), 18-60 kHz with the majority of the energy between 20-40 kHz in S. vagrans (Buchler, 1976), and about 30-50 kHz in Blarina brevicauda (Gould et al., 1964; Tomasi, 1979). Griinwald (1969) found two Crocidura species vocalizing up to 46 kHz, but the main frequencies were about 20 kHz. Acoustic Behavior Shrews are relatively vocal animals (Gould, 1969; Hutterer, 1978) and there is no doubt that vocalization and hearing are important for intraspecific communication during activities such as courtship, mother-offspring contact, and agonistic behavior. This notwithstanding, the vocal repertoire of shrews is not as complex as in higher mammals (Hutterer and Vogel, 1977). It has been suggested that Crocidura, in contrast to Sorex, may use hearing to locate prey (Burda and Bauerova, 1985), but this hypothesis has not been experimentally tested. It is doubtful whether hearing plays a crucial role to warn of predators, and surely not at a great distance. Higher frequency sounds at natural intensities are soon attenuated in the environment; low- frequency sounds apparently are not perceived, and in any case they cannot be localized by shrews. Shrews are able to produce and perceive high-frequency sounds. Echolocation would certainly be of great use for these mammals with poor eyesight. It will be argued that due to the short basilar membrane and small numbers of receptors, the hearing of shrews must limit discrimination capacities. The analysis of echoes would be thus too crude for echolocation to be useful in hunting. Gould et al. (1964), Buchler (1976), and Tomasi (1979) demonstrated echolocation during orientation by the soricine shrews Sorex and Blarina. According to these authors the primary use of echolocation by shrews was for exploration of the environment; Blarina in Tomasi’s experiments were able to distinguish by means of echolocation between closed and open turmels. Griinwald (1969) was not able to document echolocation in Crocidura. The usual interpretation of these conflicting results is that soricine shrews echolocate and crocidurine shrews do not. We do not question the results, but we feel the experiments should be repeated using representatives of both subfamilies under the same conditions, preferably in the same laboratory. Examination of the ears of many wild-caught shrews of several species did not reveal diseases or disorders which could substantially affect the hearing function (Burda, 1978). Either these conditions are very rare or such cases are immediately eliminated by natural selection. This observation may be considered indirect evidence that a functioning auditory system is important for survival and success in shrews. Functional Morphology Outer ear.— The structure of the outer ear of shrews has been described by Boas (1912) and Burda (1980). The auricle of shrews (Fig. 6) approximates the hypothetical prototype of a primitive mammalian auricle (see Boas, 1912). The relative length and surface area, degree of divergence from the head, intensity of hair cover, and resulting exposure of the auricle are genus-specific. Crocidura and Suncus have relatively larger, less hairy, and more diverted and exposed auricles than Sorex or Neomys (Burda, 1980). The difference may be explained by Allen’s rule (Burda, 1980; see also Nagel, 1994): the Crocidurinae originated in warmer southern climates whereas the Soricinae originated in cooler northern areas. Desert shrews of the genus Notiosorex (Soricinae) also have conspicuous auricles (Walker, 1975). In addition, the crocidurine shrews are surface-dwelling whereas the soricine shrews are more fossorial (Burda and Bauerova, 1985). Fossorial and subterranean mole shrews {Anourosorex squamipes) have auricles completely concealed in the fur (Walker, 1975). Yet it would be hasty to conclude that the pinnae in these forms became small and hidden in the fur primarily to avoid interference with burrowing so as not to “act as shovels collecting the dirt.” We suggested that a prerequisite for the reduction of the auricles in subterranean mammals is abandonment of the necessity for keeping the auricles for sound collecting, amplifying, and binaural localization (Burda et al., 1990), and perhaps even for thermoregulation. The cutaneous muscles of the head and neck and the 192 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 auricular muscles exhibit a primitive arrangement in shrews which approximates an archetypal mammalian condition (Burda, 1979a). The auricular muscles are somewhat larger in Crocidura suaveolens than in Sorex araneus, which may be related to the larger size and greater deflection of auricles in the former species (see also Burda and Bauerova, 1985). It is not clear whether the auricles of shrews are movable. The position of the auricles at the sides of the head, the relatively small, undeveloped auricular muscles, and the steady, lively, fast, and violent motion of the head characteristic of shrews, are conditions not fiilly consistent with movable auricles. The outer earcanal (external meatus) in shrews is cartilaginous along its entire course. It is relatively long: 4.5-8 mm in Sorex minutus and Neomys fodiens, respectively (Burda, 1980). For comparison, the meatus length in the Norway rat is 5.7 mm (Plassmann and Brandle, 1992). If the ear canal is considered a resonating pipe, the meatus in S. araneus, for example, would resonate at about 13 kHz (see Table 2). However, the length of the meatus is primarily determined and constrained by the distance between the outer ear canal opening (auricle) and the ear drum (tympanic bone). Since shrews retain the primitive condition of tympanic bones being situated on the ventral side of the head, the meatus has to be relatively long (Burda, 1979h, 1980). Middle ear. --Detailed morphological descriptions of the middle ear structures in shrews have been provided by Burda (1979h), Fleischer (1973), and Henson (1961), (see also literature cited therein). The tympanic bulla in shrews (at least in Sorex, Neomys, and Crocidura) is largely ligamentous and nonossified. The only bone it contains, excluding the paries cochlearis, is the tympanic ring or annulus tympanicus, which makes a frame for the tympanic membrane or ear drum. The areas of the pars tensa of the ear drum and the estimated volumes of the middle ear cavity in some shrew species are given in Table 2 (based on Burda, 1979h). Using the formulas of Plassmann and Brandle (1992) the resonance frequency of these structures was estimated and included in Table 2. Although the computed values are only rough approximations (Plassmann and Brandle, 1992), it is remarkable how similar these values are for different parts of the ear in each species. The mean resonance frequencies vary predictably when a size- graded series of shrew species is compared (Table 2). We assume that the ear structures have the lowest impedance at the resonance frequency, therefore the resonance frequency and the frequency range of best hearing sensitivity should correlate. The estimated values also agree with frequency characteristics of vocalizations used for intraspecific communication. The vibration of the ear drum is transferred to the cochlea by the auditory ossicles. The vibratory amplitude is increased and the impedance from air to cochlear fluid is matched by the acoustic lever action of the ossicles, expressed by the “lever ratio,” and by condensation of energy from the larger ear drum to the smaller footplate of the stapes, expressed by the “area ratio.” These ratios in some shrew species are given in Table 2. The values of the area ratio in shrews are the highest known among mammalian species (see Burda 1919b; Plassmann, 1989). The lever-ratio values were determined according to Plassmann (1989) (see Fig. 7). The “final transformation ratio” is defined as the product of the two ratios (Moller, 1974). ; Theoretically, a value of about 63 indicates efficient transmission of energy (Moller, 1974). Taking into account that i this value ( = 63) is based on approximations, and that we did not measure the actual ear drum area but the area enclosed by , the tympanic ring, one conclusion is that the middle ears of shrews fit the theory remarkably well. The auditory ossicles in shrews (Fig. 7) are firmly connected with each other and, through the gonial, with the tympanic ring. * Ossicles of similar form and arrangement are also found in ^ other small and “ultrasonic” forms like bats and mice (Fleischer, 1973). Cochlea of the inner ear. —The structure of the cochlea of ' shrews has been described from different points of view and • using different methods of study by Platzer (1964), Platzer and i Firbas (1966), Firbas and Platzer (1969), Fleischer (1973), Burda (1978, 1979h), Sigmund (1985), and Walther (1987), (see also references therein). The cochlea of shrews is low and flat. It has 1.5 turns in : Sorex and Neomys, and 1.75 turns In Crocidura. The secondary | spiral lamina is well -developed. The low number of cochlear I coils is considered a primitive trait in mammals. A low and flat I cochlea with well -developed secondary spiral laminae is characteristic of mammals with good high-frequency hearing j (Fleischer, 1973). | The length of the basilar membrane varies from 2. 9-4. 5 mm i in four soricid species (Table 3) and may be correlated with the size of the pinna in the examined Soricidae and with body size I at the subfamily level. Crocidura has a longer basilar membrane and a larger pinna than would be found in a soricine shrew of I comparable body size. An analogous situation was found among I acoustically unspecialized mice and rats (Muridae) and discussed in detail (Burda et al., 1988). Parameters of many elements of the cochlear partition, such ■ as thickness of the basilar membrane, height of the organ of ' Corti, and length of the stereocilia of hair cells, change from 5 the base toward the apex of the cochlea in Sorex araneus and Crocidura russula in the same predictable manner as in other acoustically unspecialized mammals (Walther, 1987). Apart from some differences in the structure of the basilar membrane between S. araneus and C. russula, the functional significance of which is not clear, the cochleas of both species are structurally unspecialized organs. Local specializations or discontinuities in cochlear morphometrical baso-apical gradients, which would account for extension of the region where the best frequency is represented and thus for better hearing resolution in the respective frequency range, have been reported in bats and gerbils but were not observed in shrews (Walther, 1987; Walther and Bruns, 1987). The density of cochlear receptors in the species of shrews examined amounted to about 390 outer hair cells and 1 10 inner hair cells per 1 mm (Table 3). These densities are somewhat higher than the “mammalian average” (Burda and B ranis, unpublished data). This may be due to the smaller size of cells in smaller mammals compared to larger ones (see Burda et al.. 1994 BRANIS AND BURDA — Vision and Hearing in Shrews 193 1988). While the density of outer hair cells increased from the base to the apex of the basilar membrane, the density of inner hair cells reached maximum at about 60% of the basilar membrane length from the basal end in Sorex araneus (Burda and Branis, unpublished data). A similar pattern of receptor distribution has been found in house mice and many other mammals (Burda and Branis, unpublished data; Burda et al., 1988). The total number of cochlear receptors in mammals is determined primarily by the length of the basilar membrane, which shows greater interspecific variation than does the mean density of receptors. The absolutely shorter basilar membrane in shrews can accommodate only a smaller number of receptors (Table 3). This may have functional consequences: either the dynamic range of the shrew ear may be relatively limited, or frequency and intensity discrimination (sensitivity) of hearing may be relatively poor (cf. Wever, 1974; Burda et al., 1988). In shrews, the cuticular plates of the cochlear hair cells are arranged in a conspicuous geometric pattern on the reticular lamina along the entire organ of Corti (Fig. 8). Such a geometrically regular arrangement is typical of basal parts of the Corti organs in many mammalian species and of the entire Corti organs of bats. This is a feature typical of Corti organs tuned to high frequencies (Burda, \919b). Considering some additional features, like the presence of the secondary spiral lamina and the massive spiral ligament throughout the cochlear duct, we conclude that even the apical parts of the Corti organ are tuned to relatively high frequencies. The hearing range of shrews is apparently shifted to higher frequencies. The spiral cochlear ganglia in Sorex minutus and S. araneus contain 6,440 and 7,090 neuron cells, respectively (Sigmund, 1985). The ratios between the numbers of cochlear neurons and counts of cochlear receptors are within the range of values found in murids (Burda et al., 1988). The postnatal development of the organ of Corti has been studied in Sorex araneus (Branis and Burda, unpublished data). The reticular lamina of the Corti organ in this species matures in a way similar to that of murids (Burda, 1985; Burda and Branis, 1988). At birth, the triad of cuticular plates of outer hair cells is narrow at the apex and wide at the base of the cochlear spiral. During maturation it grows at the apex and decreases at the base to reach its mature dimensions. At a certain point along the organ of Corti the mature values are present at birth, and the dimensions of this area do not change significantly during further development. In murids there is an exact correlation between the location of this point and the region where the maximum density of inner hair cells is found. This area also corresponds to the cochlear region activated by the frequency range of greatest auditory acuity (Burda, 1985; Burda and Branis, 1988). This developmental “turning point” in S. araneus also was localized at the place of the maximum density of inner hair cells. The width of the triad of outer hair cells was about 17 /rm at this place. The region of basilar membrane with a comparable dimension is where the frequencies (octave) of 16-32 kHz are represented in the rat, the house mouse, and the cat (Burda, 1985; Burda and Branis, 1988, and unpublished data for morphometry; Ehret, 1975; Liberman, 1982; and Muller, 1988, for tonotopy). Consequently, based on assessment of only this parameter, we assume that the cochlear partition in S. araneus is tuned to frequencies of about 16-32 kHz. Conclusions The components of the visual system in the shrews examined appear to be structurally normally developed. The retina is of a diurnal type with both rods and cones. The ontogenetic development of the eye appears to be normal and no part of the eye can be considered developmen tally retarded. The morphological substrate for functional light perception is provided. The small size of the eyeball and consequently small retinal surface accommodates a limited number of receptors. There is neither a fovea nor an area centralis in the retina of shrews. The accommodating apparatus of the eye is very simple and probably nonfunctional. These quantitative aspects indicate that visual acuity may be very limited in shrews. Shrews are capable of light-dark discrimination on the basis of long- and short-term stimulation. They are unable visually to discern moving objects or sudden brief light stimuli. The controversy may be explained in terms of functional morphology. The cone-rich retinas of shrews provide a morphological basis for color discrimination. The ability to distinguish colors has been demonstrated in behavioral tests. Due to the small size of the perceptive retinal surface, shrews probably have only the ability to perceive color changes in the environment as stimuli affecting circadian and/or circannual rhythms. Binocular fixation of color images seems to be unlikely. The structure of the outer, middle, and inner ear in shrews generally conforms to a primitive eutherian pattern. The size of ear structures in soricine shrews follows a slightly different, yet parallel, allometric curve to that of crocidurine shrews. Assuming that the resonance frequency of ear structures is determined by their dimensions, frequency tuning was calculated for each structure independently according to physical formulas and formulas given in the literature. It was shown that all structures in each particular species are tuned to roughly the same frequency. The dimensions of auditory system components remained proportional and follow the general mammalian plan. Neither the high-frequency hearing of shrews nor the structure of their ear which makes high-frequency hearing possible should be considered evolutionary adaptations (sensu Sluys, 1988). In summary, hearing in shrews is correlated with vocalization and both are tuned to the same frequencies. The tuning is determined by the dimensions of auditory components and those structures in turn are determined by overall size and strict conformity to a general pattern. Crocidura can theoretically enhance auditory sensitivity in the range of higher frequencies by changes in structural dimensions without abandoning the general pattern. These changes may have occurred due to environmental factors encountered by living in warmer habitats (Allen’s rule), or by a less fossorial existence, more open habitats (with other acoustic properties), and by different hunting strategies. Hearing undoubtedly is important for shrews as evidenced by behavioral observations and by the 194 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 fact that no pathologic cases were found in nature. Echolocation cannot be ruled out; it may be useful, but its discrimination capacities would be relatively limited. The actual role of hearing and its significance for shrews have yet to be thoroughly tested in controlled experiments. Generally, the qualitatively normal structure of the eye and ear predispose shrews to normal vision and audition. Their small size, however, presumably reduces and constrains the range or sensitivity or resolution capabilities of visual and acoustic perception. No structural or dimensional modifications of the respective sensory organs which would adaptively enhance their functional capacities or change their tuning were noted. Shrews can thus be considered visually and acoustically unspecialized (generalized) marrunals. The ecological niche of shrews apparently exerts little selective pressure for vision and audition, therefore the eye and ear have retained small dimensions and limited functional capacities. Acknowledgments We are grateful to G. L. Kirkland, Jr., and J. F. Merritt for their invitation to write this review. We thank L. Sigmund, L. Voldrich, and F. Vrabec (Prague) for all the support during the research work. HB appreciates the support of J. Winckler (Frankfurt) which made it possible to complete some parts of the study and to participate in writing this paper. V. Bruns (Frankfurt) and J. Winckler as well as all the reviewers are thanked for commenting on the manuscript. Stimulating discussions with W. Plassmann (Frankfurt) are highly acknowledged. This paper is dedicated to our wives. Literature Cited Boas, J. E. V. 1912. Ohrknorpel und ausseres Ohr der Saugetiere. Nielsen and Lydiche, Copenhagen, 250 pp. BranIS, M. 1981. Morphology of the eye of shrews (Soricidae, Insectivora). Aeta Universitatis Carolinae-Biologica, 1979(1 1):409-445. 1985a. The optic nerve in shrews (Soricidae, Insectivora). Fortschritte der Zoologie, 30:715-717. 1985t. Postnatal development of the eye of Sorex araneus. Acta Zoologica Fennica, 173:247-248. 1986. Eye ontogeny in the common shrew {Sorex araneus). Unpublished Ph.D. dissert., Charles University, Prague, Czechoslovakia, 127 pp. (in Czech). 1988. Light perception in the white-toothed shrew, Crocidura suaveolens (Mammalia, Insectivora). Vestnik ceskoslovenske Spolecnosti zoologicke, 52:1-6. _. 1989. Morphology of the cornea in shrews (Soricidae, Insectivora) and its relation to burrowing habits. Abstract, Fifth International Theriological Congress, Rome, 653 pp. Buchalczyk, a. 1972. Seasonal variation in the activity of shrews. Acta theriologica, 17(17):221-243. Buckler, E. R. 1976. The use of echolocation by the wandering shrew {Sorex vagrans). Animal Behaviour, 24:858-873. Burda, H. 1978. Population der Haarzellen des Cortischen Organs der Spitzmause. Zeitschrift fur mikroskopischanatomische Forschung, 92:514-528. 1979a. Muscles of the external ear of Czechoslovak shrews (Soricidae). Lynx, 20:1-48. 19797>. Morphology of the middle and inner ear in some species of shrews (Insectivora, Soricidae). Acta Scientiarum Naturalium Brno, 13(4): 1-46. 1980. Morphologic des ausseren Ohres der einheimischen Arten der Familie Soricidae (Insectivora). Vestnik ceskoslovenske Spolecnosti zoologicke, 44:1-15. 1985. Quantitative assessment of postnatal maturation of the organ of Corti in two rat strains. Hearing Research, 17:201-208. Burda, H., and Z. Bauerova. 1985. Sensory adaptations and feeding ecology in shrews (Soricidae). Acta Zoologica Fennica, 173:253-254. Burda, H., and M. Branis. 1988. Postnatal development of the organ of Corti in the wild house mouse, laboratory mouse, and their hybrid. Hearing Research, 36:97-106. Burda, H., L. Ballast, and V. Bruns. 1988. Cochlea in the Old World mice and rats (Muridae). Journal of Morphology, 198:269-287. Burda, H., V. Bruns, and M. MOller. 1990. Sensory adaptations in subterranean mammals. Pp. 269-293, in Evolution of Subterranean Mammals at the Organismal and Molecular Levels (E. Nevo and O. A. Reig, eds.). Progress in Clinical and Biological Research, Wiley-Liss, New York, 335:1-422. Cei, G. 1946. Morfologia degli organ! della vista negli insettivori. Archivio Italiano della Anatomia ed Embriologia, 52:1-42. Crowcroft, P. 1964. Note on the sexual maturation of the shrews {Sorex araneus, Linnaeus 1758) in captivity. Acta theriologica, 8(5):89-94. Duke-Elder, S. 1958. The Eye in Evolution. Volume 1 in System of Ophthalmology (S. Duke-Elder, ed.), Henry Kimpton, London, 843 pp. Ehret, G. 1975. Masked auditory thresholds, critical ratios, and scales of the basilar membrane of the house mouse {Mus musculus). Journal of Comparative Physiology, A. Sensory, Neural, and Behavioral Physiology, 103:329-341. Firbas, W., and W. Platzer. 1969. Uber die Cochlea der Insectivoren. Anatomischer Anzeiger, 124:233-243. Fleischer, G. 1973. Studien am Skelett des Gehororgans der Saugetiere einschliesslich des Menschen. Saugetierkundliche Mitteilungen, 21:131-239. Gebczynski, M. 1965. Seasonal and age changes in the metabolism and activity of Sorex araneus, Linnaeus 1758. Acta Theriologica, 10(22):303-331. Gould, E. 1969. Communication in three genera of shrews (Soricidae): Suncus, Blarina, and Cryptotis. Communications in Behavioral Biology, A, 3:11-31. Gould, E., N. C. Negus, and A. Novick. 1964. Evidence for echolocation in shrews. Journal of Experimental Zoology, 156:19-38. Grun, G., and K. H. Schwammberger. 1980. Ultrastructure of the retina in the shrew (Insectivora, Soricidae). Zeitschrift fiir Saugetierkunde, 45:207-216. Grltnwald, A. 1969. Untersuchungen zur Orientierung der Weisszahnspitzmause (Soricidae, Crocidurinae). Zeitschrift fur vergleichende Physiologic, 65:191-217. Heffner, H., and B. Masterton. 1980. Hearing in Glires: Domestic rabbit, cotton rat, feral house mouse, and kangaroo rat. Journal of Acoustical Society of America, 68(6): 1584-1599. Henson, O. W. 1961. Some morphological and functional aspects of certain structures of the middle ear in bats and insectivores. The University of Kansas Science Bulletin, 42:151-255. Hutterer, R. 1976. Beobachtung zur Geburt und J ugendentwicklung von Sorex minutus L. (Soricidae, Insectivora). Zeitschrift fiir Saugetierkunde, 41:1-22. 1978. Paarungsrufe der Wasserspitzmaus {Neomys fodiens) und 1994 BRANIS AND BURDA — Vision and Hearing in Shrews 195 verwandteLaute weitererSoricidae. Zeitschrift fur Saugetierkunde, 43:330-336. HUTTERER, R., AND P. VoGEL. 1977. Abwehrlaute afrikanischer Spitzmause der Gattung Crocidura Wagler, 1832 und ihre systematische Bedeutung. Bonner zoologische Beitrage, 28:218-227. KodeJSOVA, V. 1989. An ethological study of color vision in Neomys fodiens. Unpubl. M.S. thesis, Charles University, Prague, Czechoslovakia, 38 pp. (in Czech). Kolmer, W., and H. Lauber. 1936. Haut und Sinnesorgane. Chapter 1 in Handbuch der mikroskopischen Anatomie des Menschen (W. v. Mollendorf, ed.). Springer, Berlin, 782 pp. Liberman, M. C. 1982. The cochlear frequency map for the cat: Labeling auditory nerve fibers of known characteristic frequency. Journal of Acoustical Society of America, 72(5): 1441-1449. M0LLER, A. R. 1974. Function of the middle ear. Pp. 491-517, in Handbook of Sensory Physiology. Volume V/1. Auditory System. Anatomy, Physiology (ear) (W. D. Keidel and W. D. Neff, eds.). Springer, New York, 736 pp. Muller, M. 1988. Frequency place map of the rat cochlea and brainstem. P. 184, in Sense Organs. Interfaces Between Environment and Behavior (N. Eisner and F. G. Barth, eds.), Thieme, Stuttgart. Nagel, A. 1994. Metabolic rates and regulation of cardiac and re- spiratory function in European shrews. Pp. 421-434, in Advances in the Biology of Shrews (J. F. Merritt, G. L. Kirkland, Jr., and R. K. Rose, eds.), Carnegie Museum of Natural History Special Pubication no. 18, x + 458 pp. Plassmann, W. 1989. Ubertragungseigenschaften des peripheren Horsystems von Nagem. Unpubl. habilitation thesis. J. W. Goethe University, Frankfurt am Main, Germany, 123 pp. Plassmann, W., and K. Brandle. 1992. A functional model of the peripheral auditory system in mammals and its evolutionary implications. Pp. 637-653 in The Evolutionary Biology of Hearing (D. B. Webster, R. R. Fay, and A. N. Popper, eds.), Springer, New York. Platzer, W. 1964. Zur vergleichenden Anatomie der Cochlea bei Talpa europaea, Sorex araneus und Sorex alpinus. Anatomischer Anzeiger, 115:113-118. Platzer, W., and W. Firbas. 1966. Die Cochlea der Soricidae (ihre Windungszahl und absolute Grosse). Anatomischer Anzeiger, 117:101-113. Rissman, E. F., R. j. Nelson, J. L. Blank, and F. H. Bronson. 1987. Reproductive response of a tropical mammal, the musk shrew (Suncus murinus), to photoperiod. Journal of Reproduction and Fertility, 81:563-566. Rochon-DuvigneauD, a. 1943. Les yeux et la vision des vertebres. Mason et Cie, Paris, 719 pp. Rood, J. P. 1958. Habits of the short tailed shrew in captivity. Journal of Mammalogy, 39:449-507. Sato, Y. 1977. Comparative morphology of the visual system of some Japanese species of Soricoidea (superfamily) in relation to life habits. Zeitschrift fiir Himforschung, 6(18):531-546. Schwartz, S. 1935. Uber das Mausauge, seine Akkommodation und uber das Spitzmausauge. Jena. Zeitschrift fur Naturwissenschaften , 70:113-158. Sharma, D. R. 1957. Studies on the anatomy of the Indian insectivore Suncus murinus. Journal of Morphology, 102:405-591. SlEGMUND, R., AND L. SIGMUND. 1983. Circadian oscillation of locomotor activity in Crocidura suaveolens (Soricidae, Insectivora, Mammalia). Zeitschrift fiir Saugetierkunde, 48(3): 185-187. Sigmund, L. 1985. Anatomy, morphometry and function of sense organs in shrews (Soricidae, Insectivora, Mammalia). Fortschritte der Zoologie, 30:661-665. Sigmund, L., R. Druga, and R. Siegmund. 1984. Retinal projections in Crocidura suaveolens (Soricidae, Insectivora, Mammalia). The primary optic pathway. Vestnik ceskoslovenske Spolecnosti zoologicke, 48:296-301. Sigmund, L., R. Siegmund, and C.-P. Claussen. 1987. Structure and function of the photoreceptors in Crocidura suaveolens and Sorex araneus (Soricidae, Insectivora, Mammalia) and their relation to the locomotor behaviour. Zoologisches Jahrbuch, Physiologie, 91:63-78. Sigmund, L., R. Dandova, V. Kodejsova, and R. Siegmund. 1989. Das Licht als Umweltfaktor im Leben der Wasserspitzmaus (Neomys fodiens). Zeitschrift fiir Saugetierkunde 63, Supplement 7. Sluys, R. 1988. On adaptation, the assessment of adaptations, and the value of adaptive arguments in phylogenetic reconstruction. Zeitschrift fiir zoologische Systematik und Evolutionsforschung, 26:12-26. Sprando, R. L., M. B ranis, and G. L. Dryden. 1989. Prenatal development of the eye of the Asian musk shrew, Suncus murinus (Mammalia, Insectivora). Vestnik ceskoslovenske Spolecnosti zoologicke, 53:7-16. Tomasi, T. E. 1979. Echolocation by the short-tailed shrew Blarina brevicauda. Journal of Mammalogy, 60:751-759. Walker, E. P. 1975. Mammals of the World. 3rd Edition. The Johns Hopkins University Press, Baltimore, 1:1-644. Walls, C. L. 1942. The vertebrate eye. Cranbrook Institute of Science, Bloomfield Hills, Michigan, 785 pp. Walther, Y. 1987. Quantitative morphologische Analyse des Cortischen Organs von Rattus norvegicus, Tupaia belangeri, Crocidura russula, Sorex araneus. Unpubl. M.S. thesis, J. W. Goethe University, Frankfurt am Main, Germany, 150 pp. Walther, Y., and V. Bruns. 1987. Das Innenohr von Sorex araneus und Crocidura russula (Soricidae). Zeitschrift fiir Saugetierkunde, 61 , supplement:51-52. Wever, E. G. 1974. The evolution of vertebrate hearing. Pp. 423-454, in Handbook of Sensory Physiology. Volume V/1. Auditory System. Anatomy, Physiology (ear) (W. D. Keidel and W. D. Neff, eds.). Springer, New York, 736 pp. 196 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 Table 1. — Morphometry of the eye in shrews. Species Eyeball Diameter® Retinal Thickness*’ Retinal Area*’ Rod/Cone Ratio Receptor Density** Axons Number' S. minutus 0.93 119 1.2 7/1 245,000 S. alpinus 0.72 138 0.7 8/1 221,000 — S. araneus 1.14 139 1.6 6/1 220,000 6,000 C. suaveolens 1.14 145 1.4 12/1 300,000 3,800 C. leucodon 1.20 140 1.4 15/1 390,000 — N. anomalus 1.29 137 1.8 16/1 410,000 — N. fodiens 1.49 140 2.0 17/1 465,000 — “ Equatorial diameter of eyeball (mm). ** Thickness of central retina Retinal area (mm^). Number of receptors per 1 mm^. ® Number of axons in the optic nerve. Table 2. — Morphometrical and functional characteristics of the outer attd middle ears of shrews. Species a Body Mass isY b Inter- aural Distance (mm)** c f (kHZ)*’ d Meatus Length (mm)** e f (kHZ)*’ f Meatus Area (mm^)* g Middle Ear Volume (mm^)® h f (kHz)' i Mean Eardrum Radius ’ (mm)‘ j f (kHz)i k f (kHz)** 1 Area Ratio* m Lever Ratio™ n Final Ratio” Sorex minutus 4 15 22.9 4.5 19.6 0.95 3.13 23 0.68 19.6 21.3 (1.9) 49:1 1.33:1 65.2:1 Crocidura suaveolens 5 19 18 5.2 17 1.41 6.48 18.7 0.84 15.9 17.4 (1.2) 55:1 1.37:1 75.3:1 Sorex araneus 8 22.5 15.2 6.8 13 1.33 7.64 16.9 0.81 16.4 15.4 (1.7) 51:1 1.17:1 59.7:1 Neomys fodiens 13 27 12.7 8.0 11 1.77 11.15 15.5 0.94 14.2 13.3 (1.9) 46:1 1.44:1 66.2:1 ^ Based on Burda (1979/?). ** Measured around the head; mean values based on measurements of five specimens in each species. Frequency effectively shadowed by the head: f = c/d where c = 343m/s, d = interaural distance (column b). Based on Burda (1980). ® Resonance frequency of the meatus: f = c/4-( where c = 352.9 m/s, f = meatus length (column d). Mean values of measurements made in five specimens in each species. ® Calculated from dimensions given in Burda (1979/?). ^ Resonance frequency of the outer and middle ear computed according to Plassmann and Brandle (1992): m \ VL where x = constant = 56, 166, m = meatus area (column g), V = middle ear volume (column g), L = 1 .5r + 0.96 where r = radius of the meatus. ‘ Based on Burda (1979b). J Resonance frequency of the eardrum computed according to Plassmann and Brandle (1992): f = Y/r where y = constant = 13.32, r = radius of the eardrum. Mean frequency (and standard deviation) of values in columns b, e, h, j. ' Eardrum area : stapedial footplate area; see Fig. 7 (this paper) and Burda (1979b). ™ Lever ratio of the malleus : incus; calculated from dimensions given in Burda (1979b), see Fig. 7 (this paper). " Final ratio sensu Moller (1974) (theoretically expected value = 63); product of the lever and area ratios (columns 1, m). 197 1994 BRANIS AND BURDA — Vision and Hearing in Shrews Table 3.— Length of the basilar membrane, density, and total counts of cochlear receptors in shrews (based on Burda, 1979h). Sorex minutus Crocidura suaveolens Sorex araneus Neomys fodiens Length of basilar membrane (mm) 2.9 4.0 3.95 4.5 Density of outer hair cells per 1 mm of basilar membrane length 400 380 396 384 Density of inner hair cells per 1 mm of basilar membrane length 109 107 112 106 Total number of outer hair cells 1,161 1,522 1,563 1,734 Total number of inner hair cells 315 430 442 479 Fig. 1.— Sagittal section through the eyeball of Sorex araneus. C, cornea; Cb, ciliary body; L, lens; R, retina; Su, sclerouveal complex. Arrow points to sphincter pupillae muscle. Semithin section; toluidin blue; bar = 0.2 mm. Fig. 2.— Sagittal section of the cornea of Sorex araneus. En, corneal endothelium; Ep, corneal epithelium; K, keratinized layer; Sp, corneal matrix (substantia propria comeae). Semithin section; toluidin blue; bar = 30 fim. 198 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 Fig. 3. — Electromicrograph of the comeal surface in Sorex araneus. E, epithelium-cell layer; K, keratinized lamaellae. Bar 1 0 fim. Fig. 4.— Tangential section of the retinal receptor layer (LM). a, Sorex araneus. b, Crocidura suaveolens. R, rods; C, cones. Note the difference in the density of rods and cones in both species. Paraffin section 6 /xm thick; Held hematoxyline; bar = 25 /xm. 1994 BRANIS AND BURDA— Vision and Hearing in Shrews 199 Fig. 5. — The retina of Sorex araneus. P, pigmented epithelium; R, receptor layer; On, outer nuclear layer; Op, outer plexiform layer; In, inner nuclear layer; Ip, iimer plexiform layer; G, ganglion cell layer. Note both types of visual cells, rods and cones, in the receptor layer. Semithin section; toluidin blue; bar = 30 jim. Fig. 6.— The auricle of a shrew (semischematic drawing). AP, apex auriculae; AT, antitragus; CC, cavum conchae; HE, helix; PP, plica principalis; SC, scapha; TR, tragus. 200 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 I Fig. 7. — The sound-conducting apparatus of the middle ear of a shrew (Sorex araneus) (semischematic drawing). I, incus; G, goniale; M, malleus; S, stapes; T, tympanic ring. Indicated is the lever system of the ossicles: IL, incudial lever; ML, mallear I lever; dotted sections indicate the ear drum and the stapedial footplate. f Fig. 8. — Surface specimen of the Corti organ in Sorex araneus. ihc, one row of inner hair cells; ohc, three rows of outer hair .. cells. Note a disturbance of a regular geometric pattern caused by a supernumerary outer hair cell. Stained in toto by toluidin - blue and Ehrlich hematoxyline; bar = 30 ^m. ; RELATIONSfflP OF MANDIBULAR MORPHOLOGY TO RELATIVE BITE FORCE IN SOME SOREX FROM WESTERN NORTH AMERICA Leslie N. Carraway' and B. J. Verts' 'Department of Fisheries and Wildlife, 104 Nash Hall, Oregon State University, Corvallis, Oregon 97331 Abstract Based on the reported decreasing size of the median tine on l' from north to south, combined with the progressive enlargement of certain cranial characters along that gradient, we hypothesized that bite force among several taxa of Sorex along the Pacific Coast should be related inversely to latitude and should be greatest in taxa without a median tine. We tested these hypotheses with 12 taxa of Sorex from western North America. Condylobasal length was highly correlated with mass of the mastieatory musculature. Thus, bite force in shrews can be described by cos © AIB, where © = 90° — a (a. = the angle subtended by a line from the apex of the coronoid process through the distalmost extension of the lower condylar facet and a line parallel to the ventral edge of the dentary), A is the eoronoid-condyloid length, and B is the length from the lower condyloid process to any point of interest on the dentary. Based on this formula, variation in bite force at the tip of I; and at the metaconid of Mj, accounted for by factors other than body size, was relatively small across the taxa. However, differences in bite force among taxa, other than that related to size, are related largely to eoronoid-condyloid length and to angle ©. Bite force per se seems correlated positively with absence or reduced size of the tine on r, greater appression of shrews. 1 Introduction Mammals possess a wide variety of cranial morphologies and dental batteries usually closely tied with their foraging ecologies. Differences in jaw morphology and associated musculature commonly are sufficient to explain differences in diets and foraging behavior even among closely related taxa (Freeman, 1979, 1981; Kiltie, 1982, 1984; Humphrey et al., 1983; Herring, 1985) including insectivores (Nikolskaya, 1965). Because of their high rates of metabolism and nearly continuous activity, shrews (Soricidae) require a constant supply of food (Genoud, 1988). These demands “should exert relatively strong selection for food gathering and processing” (Pearson, 1988: 1). Some dentary specialization has occurred among soricids as they have evolved large and procumbent I's (first upper incisors) and some have red iron-bearing enamel (secondarily ! lost in others) that by differential erosion produces sharp cutting edges on the teeth (Dotsch and Koenigswald, 1978; Vogel, 1984). Median tines on I*, first recorded by Baird (1858) who considered them of taxonomic importance, also may have a dietary function (possibly related to the hardness of food items — Carraway, 1990). Presence and position of tines have been used to separate strongly similar syntopic shrews (Hoffmann, 1971; Hennings and Hoffmann, 1977; Carraway, 1990). Carraway (1990) found considerable variation in the size, shape, and position of the tines within and among some taxa that possessed them; there was some correlation with latitude. Several closely related taxa that possess relatively large tines, posteromedial ridges, or closely appressed I's without tines or ridges are syntopic in western Oregon. If tines have a i dietary function, such a function may be manifested through j differences in bite force, thus permitting use of different food 1 resources and avoidance of excessive competition among these taxa. We considered this hypothesis by measuring bite force in j several taxa of Sorex without tines and with tines of different j morphologies, and we tested the hypothesis that bite force , relative to the size of the animal was not related to the s, greater aridity of habitats, and greater hardness of foods eaten by presence, size, or position of median tines. Subsequently, we analyzed jaw mechanics in these taxa in an attempt to explain observed differences in bite force and we attempted to relate our findings to information available on food items used by these shrews. Materials and Methods Specimens and Measurements. — The tongue, hyoid apparatus, and associated musculature; salivary glands; and eyes were dissected free, and the brain aspirated from freshly skinned skulls of frozen specimens referable to Sorex monticolus setosus (n = 54), S. s. sonomae {n = 28), S. s. tenelUodus (n = 88), S. p. pacificus (n = 20), and S. vagrans (n = 25 — sensu Carraway, 1990); and S. m. alascensis (n 1 2 ) and S. trowbridgii (n — 29— sensu Jackson, 1928) from Alaska, Washington, or Oregon. Prepared skulls were weighed to the nearest 0.0001 g on a Mettler AE 240 digital balance, placed in a dermestid colony for 2-5 days for removal of soft tissues, cleaned of frass, and reweighed. We considered the net value between the two weights to be an index to the mass of the temporalis muscle because the temporalis constitutes 61-70% of the total masticatory musculature (Dotsch, 1983, 1985). All specimens ultimately were prepared as complete skeletons or as skins and partial skeletons for deposit in the Oregon State University, Department of Fisheries and Wildlife mammal collection (OSUFW). Condylobasal length was measured with a Fowler Max-Cal electronic caliper to the nearest 0.01 mm. Length of the mandible (parallel to the ventral edge of the right dentary) from the lower condylar facet (the pivot during closure of the soricine jaw — Feamhead et al., 1955) to the tip of Ij (first lower incisor) and to the metaconid of Mj (first lower molar), eoronoid-condyloid length, and the angle (a) subtended by a line from the apex of the coronoid process through the distalmost extension of the lower condylar facet and a line along the ventral edge of the right dentary (Fig. la) were measured 201 202 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 with an ocular micrometer or ocular protractor mounted in a Bausch and Lomb binocular microscope. The same linear and angular dimensions were measured on skulls of museum specimens for S. b. bairdii (« = 16 — sensu Carraway, 1990); and S. preblei (n = 22), S. palustris (n = 13), S. betidirii palmeri (n = 21), S. merriami (n = 13 — sensu Jackson, 1928), and additional S. m. alascensis (n = 17) on deposit in the Oregon State University, Department of Fisheries and Wildlife mammal collection. All 12 taxa were classified according to the presence, size, and position of the median tine on I* by LNC. Jaw Mechanics and Bite Force. — The soricine jaw (Fig. lb) may be considered a type I lever (MacDonald and Bums, 1975) with the lower condylar facet serving as the fulcrum (Feamhead et al., 1955) and with the muscle moment arm (coronoid- condyloid length) set at an acute angle to the resistance moment arm (lower condylar facet to the tip of Ij or to the metaconid of Mj). This can be represented by the simple formula: bite force = cos gAIB (1) where, e = 90° — a, cos e is the proportion of the force vector directed at a right angle to the muscle moment arm, A is the length of the muscle moment arm, and B is the length of the resistance moment arm. We regarded values calculated by use of this formula to be an index to bite force, because we considered the force applied in the closure of the jaw to be that supplied solely by the temporalis muscle and the direction of that force to be parallel to the ventral edge of the dentary. Thus, if the force applied to the muscle moment arm remains constant, bite force may be increased by lengthening the muscle moment arm in relation to the length of the resistance moment arm, by decreasing the length of the resistance moment arm in relation to the length of the muscle moment arm, or by altering e such that it becomes more acute (a becomes more obtuse). Analysis. — Means of bite force, and means of linear and angular variables were calculated for each taxon. From these summary statistics (Table 1) linear regressions in STATGRAPHICS (Statistical Graphics Corporation, 1987) were performed to ascertain the degree of relationship between bite force and other variables as a method of evaluating the contribution of each of the variables to bite force. Deviations (Fons et al., 1984) from the “average” shrew were calculated to provide an index to the degree of allometry of the jaw among shrew taxa; this was presented as the percentage of the expected “average” shrew. Results and Discussion Morphology of the Median Tine on 7^. — Sorex sonomae (both subspecies) and S. merriami have no median tines (Fig. 2a, b, c); the anterior cusps of I's usually are parallel and appressed (for a greater proportion of their length in S. sonomae). Sorex pacificus has parallel cusps on I^s separated by a ridge that extends along as much as 50% of the posteromedial edge of each cusp (Fig. 2d). Sorex bendirii and S. palustris have median tines (contrary to previously reported absence of a tine in the latter— Verts and Carraway, 1984; Jameson and Peeters, 1988) set within the pigmented area between divergent I*s (Fig. 2e, f)- Sorex trowbridgii and S. vagrans have short, somewhat obtuse tines set above the pigmented area at about 20-30° to the long axis of I^s; in both species the tines commonly are not in direct opposition (Fig. 2g, h). Tine morphology is extremely variable in S. vagrans, but I's usually are parallel and not greatly divergent (Carraway, 1990); in S. trowbridgii, I*s are widely divergent (Carraway, 1987). Sorex monticolus (both subspecies), S. bairdii, and S. preblei have relatively long, acutely pointed tines set within the pigmented area on I's (Fig. 2i, j, k, 1). The coastal forms S. monticolus and S. bairdii exhibit a dine in angle of the tine from about 15-20° in S. m. alascensis from Alaska to about 15° in S. m. setosus from Washington to about 10° in S. bairdii from Oregon. Thus, a dine from most-divergent to least-divergent I*s is apparent. Mass of the Masticatory Musculature. — Mean (+ SE) proportions of the weight of prepared skulls composed of soft tissues removed by dermestids ranged from 70.9 ± 0.3% to 76.6 ± 0.3% for the seven taxa analyzed. Also, the relationship between the weight of soft tissues removed to condylobasal length (Fig. 3) was significant (r^ = 0.9613; P < 0.0001). The narrow range of proportions and this strong relationship indicate that the weight of the masticatory musculature is proportional to the size of the animal, a situation identical to that found for European soricids (Dotsch, 1985). Turnbull (1970:243-244) opined that although “muscle pull is not necessarily always directly proportional to muscle weight (or mass)... in general there is an approach to direct proportionality between muscle mass and force and that this is a reasonable first approximation of the actual condition.” Consequently, we chose to ignore muscle pull in considering other factors that affected bite force and to consider condylobasal length as an index to muscle mass to examine other relationships. Because our findings paralleled those of Dotsch (1985) we considered it appropriate to extrapolate the direct relationship of muscle mass to condylobasal length to taxa for which we had no direct measure of muscle mass, thereby permitting use of prepared museum specimens in subsequent analyses. Kiltie (1982:189) believed that “reasonable predictions of relative bite force can be made if the parameters involved are estimated by the same criteria for the species being compared.” Relative Bite Force Among 12 Taxa q/" Sorex from Western North America. — Bite force at the tip of Ij (Fig. 4a) was poorly correlated with condylobasal length (r^ = 0.2775; P > 0.07), but at the metaconid of Mj (Fig. 4b) the correlation was stronger (r“ = 0.6118; P < 0.05). For shrew taxa that do not possess a median tine on I* and in which I's are parallel and appressed (5. s. sonomae, S. s. tenelliodus, and S. merriami) bite force at Ij was 106-112% (104-111% at Mj) of that expected on the basis of size (as indexed by condylobasal length). For the taxon that possesses a posteromedial ridge on I* and in which I's are parallel but slightly separated (5. pacificus) bite force at Ij was 104% (104% at Mj) of expected. For the taxon with a rather blunt, high-set tine on I^ and separated but parallel I's (5. vagrans), bite force at Ij was 99% (99% at Mj) of expected. For taxa with a long, low-set tine and slightly divergent I’s (S. bairdii, S. m. alascensis, S. m. setosus, and S. preblei), bite force at Ij was 93-103% (94-103 % at Mj) of expected (Fig. 4a, b). Lastly, for taxa with a low-set tine and widely divergent I*s or with a blunt tine and 1994 CARRAWAY AND VERTS-Bite Force in Sorex 203 widely divergent I^s (S. betidirii, S. palustris, and S. trowbridgii), bite force at Ij was 92-94% (93-96% at Mj). Thus, except for S. bairdii, it appears that shrews with the more parallel and less divergent I^s (Fig. 2) have progressively greater bite force. Also, there was a north-south increase in bite force among S. m. alascensis, S. m. setosus, S. bairdii, S. pacificus, S. s. tenelliodus, and S. s. sonomae (Fig. 4), taxa distributed along the Pacific Coast. These trends support those predicted by Carraway (1990). Mechanisms Responsible for Differences in Bite Force. — If the mass of masticatory musculature for all 12 taxa we examined is correlated as strongly with condylobasal length as it was for the seven taxa for which we have a direct measure of muscle mass, then the relationship between condylobasal length and the variables described in equation 1 should provide insight into morphological differences among taxa that produced differences in bite force. Disproportionately shorter resistance lever arms were not responsible for greater bite forces calculated for species without a median tine, with a posteromedial ridge, or with a small tine, because the distance from the lower condylar facet to neither the tip of I, nor the metaconid of Mj deviated from expected based on condylobasal length by more than 3%, most by no more than 1% (Fig. 5a, b). In only S. palustris and S. bendirii was the resistance arm shorter than expected by as much as 3% (Fig. 5a, b). In S. merriami and S. s. sonomae, both without median tines, the coronoid-condyloid length was 1 10% and 109% of expected on the basis of condylobasal length, respectively (Fig. 5c). In both S. s. tenelliodus and S. pacificus the length was 103 % of expected. In S. bairdii, the geographical intermediate between the latter three taxa and the long-tined taxa to the north, the coronoid-condyloid length was 101 % of expected. In S. bendirii, S. palustris, S. m. setosus, S. m. alascensis, S. trowbridgii, S. vagrans, and S. preblei, the coronoid-condyloid length was only 92-100% of expected (Fig. 5c). Thus, those species that have the greatest bite force may have developed it by evolving longer coronoid processes. However, if all else remains equal, an increase in the height of the coronoid process would not only increase the length of the muscle moment arm, but it also would reduce angle e so that the force vector to the muscle moment arm would be more acute. Indeed, angle e was 90-96 % of expected for S. merriami, S. bairdii, S. s. tenelliodus, S. s. sonomae, and S. pacificus, taxa (except S. bairdii) without tines or with only a ridge, and 102-110% of expected for the remaining taxa (Fig. 5d). However, on the basis of the coronoid-condyloid length, angle © was only 92-96% of expected for S. merriami, S. bairdii, S. s. tenelliodus, and S. pacificus (Fig. 5e). Thus, in these four taxa, bite force was increased not only by lengthening the muscle moment arm but also by further reducing angle e. In S. s. sonomae, bite force was enhanced by the longer muscle moment arm, but reduced slightly by the small increase in angle © relative to that expected on the basis of greater length of the muscle moment arm (Fig. 5e). Except for S. bendirii, angle © in the remaining species was 101-105% of expected on the basis of the length of the muscle moment arm (Fig. 5e). Sorex bendirii, the largest member of the genus in North America, not only has a muscle moment arm only 96% of expected on the basis of size (Fig. 5c), but angle © is 110% of expected on the basis of length of the muscle moment arm (Fig. 5e), explaining the least relative bite force at Ij (Fig. 4a) among the 12 taxa examined. The somewhat greater relative bite force at Mj for S. bendirii may be explained by a shorter length of mandible to Mj than expected (Fig. 5b). Because of the shorter-than- expected coronoid-condyloid length and greater-than-expected angle ©, bite force in S. palustris (Fig. 4) no doubt would be even less than observed had not the length of both resistance moment arms been less than expected on the basis of condylobasal length (Fig. 5a and 5b). Bite Force, Tine Morphology, Diet, and Habitat of Shrews. — Sorex merriami is a shrew of the sagebrush (Artemisia) desert and short -grass prairie (Brown, 1967; George, 1990) and S. sonomae is a shrew of the fogbelt of the Pacific Coast (Carraway, 1990). Foods eaten by S. merriami consist largely of Lepidoptera larvae, beetles (Coleoptera: Carabidae and Tenebrionidae), and cave crickets (Orthoptera: Ceuthophilus — Johnson and Clanton, 1954). Because of the small size of this shrew (Armstrong and Jones, 1971) and its relatively large hard-bodied prey, possession of greater bite force than a similar-sized shrew that occupies more mesic habitats seemingly would be a distinct advantage. Snails and slugs (Gastropoda) and large beetles (Coleoptera: Cicindelidae), abundant in coastal habitats occupied, are primary foods of S. sonomae (Maser, 1973; Whitaker and Maser, 1976). Greater bite force in this species likely is an advantage in piercing the calcareous shells of snails and the tough skin of slugs, and immobilization of beetles. In both S. merriami and S. sonomae, I's are appressed for some or most of their length because of the absence of encumbering tines, thereby permitting the rapid piercing of prey to subdue it. Although the function of the I's in piercing is attributed to shrews in general (Pemetta, 1977), such would seem a particularly desirable attribute for shrews that feed on large or especially hard or tough-skinned prey. Sorex pacificus (sensu Carraway, 1990), the shrew with the next strongest bite force, is often found in moist, wooded areas with fallen decaying logs and brushy vegetation (Bailey, 1936). Its diet consists of 29 of the 47 food items found in stomachs of five species of shrews examined, with internal organs of adult insects comprising 28.6% by volume of its stomach contents (Whitaker and Maser, 1976). It also immobilizes and caches cicindelids then later consumes their softer parts (Maser, 1973). This diet, combined with a greater-than-expected bite force (Fig. 4), suggests that the parallel, only slightly separated I*s are an adaptation for capturing, immobilizing, killing, and eviscerating hard-bodied prey. Sorex trowbridgii usually is considered a forest-dwelling species; it occurs in all stages of the sere, and is most abundant near logs surrounded with large quantities of ground litter where other species of shrews are absent (Dalquest, 1948; Jameson, 1955; Whitaker and Maser, 1976; Terry, 1981). It is considered to have the least specialized diet of shrews in western Oregon; it consumed all 47 food items found in stomachs of five species of shrews. Centipedes (Chilopoda) were the most common food (Whitaker and Maser, 1976). 204 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 Sorex vagrans is associated with grasslands and meadows (Terry, 1981) and, within these habitats, also tends to be a food generalist with 30 of the 47 food items occurring in its diet (Whitaker and Maser, 1976); no identified item comprised > 14% by volume of stomach contents. Also, S. vagrans is able to shift its diet depending on habitat conditions (Whitaker et al. , 1983). The strongly divergent I*s (possibly related to the obtuse tines set at about 20-30° to the long axis of the I’s) in S. trowbridgii, the extremely variable tine morphology observed in S. vagrans (Carraway, 1990), and the less-than-expected bite force in both species (Fig. 4) may be related to their generalized diets. Habitat relationships of the relatively rare S. preblei are poorly understood. Specimens have been collected in a wide variety of habitats ranging from marshes and riparian areas to sagebrush-steppe (Bailey, 1936; Verts, 1975; Hoffmann and Fisher, 1978; Larrison and Johnson, 1981; Tomasi and Hoffmaim, 1984; Williams, 1984; Ports and George, 1990). The remaining long-tined shrews (5. bairdii and both subspecies of S. monticolus) usually are species of montane and coastal forests (Bailey, 1936; Dalquest, 1948; Williams, 1955; Hennings and Hoffmann, 1977; Wrigley et al., 1979; Terry, 1981). The diet of S. monticolus consisted of 58.8% by volume of insect larvae and soft-bodied invertebrates (S. Churchfield, personal corrununication) as would be expected based on its less-than-expected bite force and mesic habitat affinity. Apparently, nothing is known regarding the diets of S. preblei and S. bairdii. However, based on the less-than-expected bite force and probably mesic habitat of S. preblei, we suggest that its diet may consist of soft-bodied prey. Conversely, the greater-than-expected bite force of S. bairdii suggests harder prey despite its association with habitats subject to high levels of precipitation. We are unable to offer a reasonable explanation of how the elongate tine, set nearly parallel to the long axis of the anterior cusp of I* (Fig. 2), might function in capturing and dispatching prey. At the opposite end of the bite-force spectrum from shrews without median tines are S. bendirii, a shrew of marsh areas with standing water (Pattie, 1969), and S. palustris, a shrew of riparian zones (Beneski and Stinson, 1987). Of five species of shrews examined in western Oregon by Whitaker and Maser (1976), S. bendirii had the most specialized diet; at least 25% of the prey consumed were aquatic with mayfly naiads (Ephemeroptera) and earthworms (Oligochaeta: Lumbricidae) the top foods. Aquatic organisms were primary food items in stomachs of S. palustris (Conaway, 1952; Linzey and Linzey, 1973; Whitaker and French, 1984). The soft-bodied prey likely require little effort either to subdue or to dispatch; thus, specializations of jaw mechanics to enhance bite force seemingly are unnecessary. In contrast, the widely divergent I's (reminiscent of those of S. trowbridgii) suggest a more generalized diet than reported. Conclusions Because mass of the masticatory musculature is so strongly correlated with size (as indexed by condylobasal length), and because muscle mass can be used as an index to force applied. differences in mandibular morphology seem nearly wholly responsible for differences in bite force among taxa of Sorex in western North America. Differences in bite force among taxa, other than that related to size, are related largely to coronoid- condyloid length and to angle e. Bite force per se seems correlated positively with the absence or reduced size of the tine on I* and the corresponding greater appression of the I^s (accounting at least in part, for the observed north-south increase in bite force among several taxa), greater aridity of habitats, and greater hardness of foods eaten by shrews. We suggest that both the characters that caused us to investigate the role of bite force in avoidance of competition among syntopic shrews (appressed I's and lack of tines on I^s) and the characters primarily responsible for an increase in bite force (a more acute angle e, a longer muscle moment arm, and shorter resistance moment arms) are derived morphological characters. Phytogenies based on morphometric (Findley, 1955) and allozymic electrophoretic (George, 1988) studies are congruent, with S. trowbridgii and S. merriami placed in an unnamed subgenus (George, 1988) and the remaining taxa that we investigated placed in the subgenus Otisorex. Therefore, we offer as evidence of the derived condition, the character states of presence of appressed I*s and lack of tines on I^s in members of two subgenera {S. merriami in the unnamed subgenus and S. sonomae in Otisorex). In addition, the north-south dine in tine morphology in closely related taxa (large, acutely angled tines in S. monticolus to short, less acutely angled tines in S. bairdii to posteromedial ridges in S. pacificus to complete absence of tines in S. sonomae) along the Pacific Coast (Carraway, 1990) adds support to our contention. Finally, for characters responsible for an increase in relative bite force, we offer as supportive evidence of the derived condition, the occurrence of a more acute angle e, a longer muscle moment arm, and shorter resistant moment arms in some but not all taxa of both subgenera (Fig. 5). The derived condition, and the apparent relationship among bite force, tine morphology, diet, and conditions of the physical environment, suggest that the state of these characters among the several taxa examined are adaptive to current ecological conditions. We believe that the correlations of bite force and tine morphology with ecological conditions were sufficient for us to reject our hypothesis that these characters and factors were unrelated. Consequently, bite force, as it affects diet, probably is sufficient to explain avoidance of excessive competition among the several syntopic taxa of Sorex in western Oregon. We suggest that the imprecision of the correlations we found was related more to the lack of knowledge of diets and habitat affinities of the specific populations of shrews that we studied than to the crudeness of our measures of bite force or our classification of tine morphology. We also suggest that precise and species-specific information on shrew behavior relative to their capturing, dispatching, and dissecting prey may provide explanations for discrepancies we observed in the postulated relationships. Acknowledgments Frozen specimens used in this research were provided by J. 1994 CARRAWAY AND VERTS — Bite Force in Sorex 205 E. Comely (U. S. Fish and Wildlife Service, formerly of William L. Finley National Wildlife Refuge), J. Rozdilsky (Burke Memorial Washington State Museum, University of Washington), R. G. Anthony (U. S. Fish and Wildlife Service, Oregon Cooperative Wildlife Research Unit), A. Hansen (Coastal Oregon Productivity Enhancement Program, Oregon State University), C. Maguire (formerly of U. S. Forest Service, Forestry Sciences Laboratory, Olympia, Washington), and L. J. Blus and R. A. Grove (U. S. Fish and Wildlife Service, Pacific Northwest Field Station, Corvallis, Oregon). E. B. Kritzman, Museum of Natural History, University of Puget Sound (PSM) and A. M. Rea, San Diego Natural History Museum (SDNHM) loaned specimens in their care. We thank W. T. Verts and E. Fichter for discussions regarding bite force. B. R. Stein commented on an earlier draft of this manuscript. This is Technical Paper No. 9375, Oregon Agricultural Experiment Station. Literature Cited Armstrong, D. M., and J. K. Jones, Jr. 1971. Sorex merriami. Mammalian Species, 2:1-2. Bailey, V. 1936. The mammals and life zones of Oregon. North American Fauna, 55:1-416. Baird, S. F. 1858. Mammals. In Reports of explorations and surveys for a railroad route from the Mississippi River to the Pacific Ocean. Beverley Tucker, Printer, Washington, D.C., 8(l):l-757 + 43 pis. Beneski, j. T., and D. W. Stinson. 1987. Sorex palustris. Mammalian Species, 296:1-6. Brown, L. N. 1967. Ecological distribution of six species of shrews and comparison of sampling methods in the central Rocky Mountains. Journal of Mammalogy, 48:617-623. Carraway, L. N. 1987. Analysis of characters for distinguishing Sorex trowbridgii from sympatric S. vagrans. The Murrelet, 68:29-30. _. 1990. A morphologic and morphometric analysis of the “Sorex vagrans species complex” in the Pacific Coast region. Special Publications, The Museum, Texas Tech University, 32:1-76. Conaway, C. H. 1952. Life history of the water shrew (Sorex palustris navigator). The American Midland Naturalist, 48:219-248. Dalquest, W. W. 1948. Mammals of Washington. University of Kansas Publications, Museum of Natural History, 2:1-444. Dotsch, C. 1983. Morphologische Untersuchungen am Kauapparat der Spitzmause Suncus murinus (L.), Sorinculus nigrescens (Gray) und Soriculus caudatus (Horsfield) (Soricidae). Saugetierkundliche Mitteilungen, 31:27-46. 1985. Masticatory function in shrews (Soricidae). Acta Zoologica Fennica, 173:231-235. Dotsch, von C., and W. v. Koenigswald. 1978. Zur Rotfarbung von Soricidenzahnen. Zeitschrift fiir Saugetierkunde, 43:65-70. Fearnhead, R. W., C. C. D. Shute, and A. d’A. Bella irs. 1955. The tempo ro-mandibular joint of shrews. Proceedings of the Zoological Society of London, 125:795-806. Findley, J. S. 1955. Speciation of the wandering shrew. University of Kansas Publications, Museum of Natural History, 9:1-68. Fons, R., H. Stephan, and G. Baron. 1984. Brains of Soricidae. I. Encephalization and macromorphology, with special reference to Suncus etruscus. Zeitschrift fiir Zoologische Systematik und Evolutionsforschung, 22:145-158. Freeman, P. W. 1979. Specialized insectivory: Beetle-eating and moth-eating molossid bats. Journal of Mammalogy, 60:467-479. 1981. Correspondence of food habits and morphology in insectivorous bats. Journal of Mammalogy, 62:166-173. Genoud, M. 1988. Energetic strategies of shrews: Ecological constraints and evolutionary implications. Mammal Review, 18:173-193. George, S. B. 1988. Systematics, historical biogeography, and evolution of the genus Sorex. Journal of Mammalogy, 69:443-461 . 1990. Unusual records of shrews in New Mexico. The Southwestern Naturalist, 35:464-465. Hennings, D., and R. S. Hoffmann. 1977. A review of the taxonomy of the Sorex vagrans species complex from western North America. Occasional Papers of the Museum of Natural History, The University of Kansas, 68:1-35. Herring, S. W. 1985. Morphological correlates of masticatory patterns in peccaries and pigs. Journal of Mammalogy, 66:603-617. Hoffmann, R. S. 1971 . Relationships of certain Holarctic shrews, genus Sorex. Zeitschrift fiir Saugetierkunde, 36:193-200. Hoffmann, R. S., and R. D. Fisher. 1978. Additional distributional records of Preble’s shrew (Sorex preblei). Journal of Mammalogy, 59:883-884. Humphrey, S. R., F. J. Bonaccorso, and T. L. Zinn. 1983. Guild structure of surface-gleaning bats in Panama. Ecology, 64:284-294. Jackson, H. H. T. 1928. A taxonomic review of the American long- tailed shrews. North American Fauna, 51:1-238. Jameson, E. W., Jr. 1955. Observations on the biology of Sorex trowbridgii in the Sierra Nevada, California. Journal of Mammalogy, 36:339-345. Jameson, E. W., Jr., and H. J. Peeters. 1988. California Mammals. University of California Press, Berkeley, 403 pp. Johnson, M. L., and C. W. Clanton. 1954. Natural history of Sorex merriami in Washington State. The Murrelet, 35:1-4. Kiltie, R. A. 1982. Bite force as a basis for niche differentiation between rain forest peccaries (Tayassu tajacu and T. pecari). Biotropica, 14:188-195. 1984. Size ratios among sympatric neotropical cats. Oecologia (Berlin), 61:411-416. Larrison, E. j., and D. R. Johnson. 1981. Mammals of Idaho. University Press of Idaho, Moscow, 166 pp. Linzey, D. W., and A. V. Linzey. 1973. Notes on food of small mammals from Great Smoky Mountains National Park, Tennessee-North Carolina. The Journal of the Elisha Mitchell Scientific Society, 89:6-14. MacDonald, S. G. G., and D. M. Burns. 1975. Physics for the life and health sciences. Addison-Wesley Publishing Company, Reading, Massachusetts, 750 pp. Maser, C. 1973. Notes on predators upon cicindelids. Cicindela, 5:21-23. Nikolskaya, V. N. 1965. A comparative morphological survey of the masticatory musculature of Insectivora of the USSR fauna. Zoologicheskii Zhumal, 44: 1228-1237 (in Russian, English summary). Pattie, D. L. 1969. Behavior of captive marsh shrews (Sorex bendirii). The Murrelet, 50:28-32. Pearson, D. W. 1988. Jaw mechanics and masticatory apparatus of three Florida shrews (Soricidae). Unpubl. M.S. thesis. University of Florida, Gainesville, 47 pp. Pernetta, j. C. 1977. Anatomical and behavioural specialisations of shrews in relation to their diet. Canadian Journal of Zoology, 55:1442-1453. Ports, M. A., and S. B. George. 1990. Sorex preblei in the 206 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 northern Great Basin. The Great Basin Naturalist, 50:93-95. Statistical Graphics Corporation. 1987. Statgraphics: Statistical Graphics System. User’s Guide. Statistical Graphics Corporation, Rockville, Maryland (looseleaf, chapters paged separately). Terry, C. J. 1981. Habitat differentiation among three species of Sorex and Neurotrichus gibbsi in Washington. The American Midland Naturalist, 106:119-125. Tomasi, T. E., and R. S. Hoffmann. 1984. Sorex prehlei in Utah and Wyoming. Journal of Mammalogy, 65:708. Turnbull, W. D. 1970. Mammalian masticatory apparatus. Fieldiana, Geology, 18:149-356. Verts, B. J. 1975. New records for three uncommon mammals in Oregon. The Murrelet, 56:22-23. Verts, B. J., and L. N. Carraway. 1984. Keys to the mammals of Oregon. O. S. U. Book Stores, Inc., Corvallis, Oregon, 178 pp. Vogel, P. 1984. Verteilung des roten Zahnschmelzes im Gebiss der Soricidae (Mammalia, Insectivora). Revue suisse Zoologie, 91:699-708. Whitaker, J. O., Jr., S. P. Cross, and C. Maser. 1983. Food of vagrant shrews (Sorex vagrans) from Grant County, Oregon, as related to livestock grazing pressures. Northwest Science, 57:107-111. Whitaker, J. O., Jr., and T. W. French. 1984. Foods of six species of sympatric shrews from New Brunswick. Canadian Journal of Zoology, 62:622-626. Whitaker, J. O., Jr., and C. Maser. 1976. Food habits of five western Oregon shrews. Northwest Science, 50:102-107. Williams, D. F. 1984. Habitat associations of some rare shrews (Sorex) from California. Journal of Mammalogy, 65:325-328. Williams, O. 1955. Distribution of mice and shrews in a Colorado montane forest. Journal of Mammalogy, 36:221-231. Wrigley, R. E., j. E. Dubois, and H. W. R. Copland. 1979. Habitat, abundance, and distribution of six species of shrews in Manitoba. Journal of Mammalogy, 60:505-520. 1994 CARRAWAY AND VERTS-Bite Force in Sorex 207 0 1 3 S O t>3 s 5 b 0 1 I S s a -V ^ 0) oo c; ^ o ^ on 3 t? 0« ^ fN I? C vP lii ^ •5 ?> r ^ r s ^ ^ -a CO -o i[ 2 S ^ c o 3 ^ s I 13 1 ^ S. 3 -g -2 s '■s I CO 2 S| C "O t 2 ^ C s. o ^ o — 3 S i> s o ft, o. ^ H r- •s o an 1' ^ o U o q O 8 q q O s 8 d o 8 o o 8 d d d d d d d d d d d d +1 +1 +1 +1 +1 +1 +1 +1 +1 +1 +1 +1 00 (N •'I' >o C7N lO NO 00 NO cs m >o >o q Tj- ■Nl" q q q q d d d d d d d d d d d d o o o o o o o o o o o o O O o o o 8 o o O o o O o o q q q q o O o o o d d d d d d d d d d d d +1 +! +1 +1 +1 +1 +1 +1 +1 +1 +1 +1 CN ON 00 r- 00 00 r- o NO cn m fO m «N (N q q (N q m q d d d d d d d d d d d d (N 1 1 1 ■o OJ q q o o o o q d d d d d d d +1 +1 +1 +1 +1 +1 +1 o NO (N o (7n CO >r» O') CN CN Ol d d d d d d d r-- ro o 00 m VO 00 »o q q q q r~ m m CS 00 VO d d d d d d d d d d d d +1 +1 +1 +1 +1 +1 +1 +1 +1 +1 +1 +1 o o in Q NO o NO NO NO 00 00 VO lO VO O 00 q r~- m m q fO q 00 oo' o6 d /^ Tf (N (S r-- o O q o O q q q q q q o d d d d d d d d d d d d +1 +1 +1 +1 +1 +1 +1 +1 +1 +1 +1 +1 CT) (s Tt NO r' r) On r«^ ON (S VO r-j •q fO q q q q q q d d d d d d d d d d VO CO >o O') m fO O o IT) NO ■'t NO r- o ’— 1 q q q o (N o d d d d d d d d d d d d +1 +i +1 +1 +1 +1 +1 +1 +1 +1 +! +1 Tf o lO r- On o in 00 q q q r' q q q in \o d d ON m On d NO d d 00 CN T— < 00 00 ro o m On VO rt ON NO fNJ fS 00 fs O'! (N fS VO CN - Co 5 .S :5 •-- 1^ ■?- „ - c c: 3 8 •C ^ ^ § I g M 1 3 I ^ I S 3, -Q 3, C > S •2 2 8 •5 5 '2 3 g -3 - ^3 c: S -3 3, o a> cx V5 s D dJ tn =3 s 0^ &/) c:: e 1-1 c« QQ «/5 -\ |7 > o o II g (u o 0.01 S. merriami • pocificas^ (112) C04) 5. bairdii • (103) a S. vograns^ (99)^ • S. s. sonomae (110) S. 5. teneHiodus i (106) S. bendirii • (92) 5. pa!ustris(SVt 'S. m. setosus (98) I • S. m. a/ascensis (93) \ 5. trowbridgi! S. preblei (93) (95) JL I _L _L _L 12.32 16.83 21.34 CONDYLOBASAL LENGTH (mm) 25.85 Fig. 4. — The relationship between condylobasal length and a, bite force at tip of Ij and b, bite force at the metaconid of Mj for 12 taxa of Sorex from western North America. Percent deviation from “average” shrew (value expected based on condylobasal length) in parentheses. CORONOID ~ CONDYLOID LENGTH (mm) LENGTH OF MANDIBLE TO TIP OF II (mm) 210 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 CONDYLOBASAL LENGTH (mm) CONDYLOBASAL LENGTH (mm) ^5. preblai (103) 12,32 16,83 21.34 25.85 CONDYLOBASAL LENGTH (mm) CORONOID - CONDYLOID LENGTH (mm) Fig. 5. — The relationship between condylobasal length and a, length of mandible to tip of Ij; b, length of mandible to metaconid of Mj; c, coronoid-condyloid length; d, angle e; and e, between coronoid-condyloid length and angle e. Percent deviation from “average” shrew (value expected based on condylobasal length) in parentheses. ULTRASTRUCTURE OF THE OLFACTORY EPITHELIUM OF THE SHORT-TAILED SHREW, BLARINA BREVICAUDA Keith A. Carson’’^, Joan L. Cole*, and Robert K. Rose* 'Department of Biological Sciences, Old Dominion University, Norfolk, Virginia 23529-0266; ^Department of Anatomy and Neurobiology, Eastern Virginia Medical School, P. O. Box 1980, 700 West Olney Road, Norfolk, Virginia 23501 Abstract The ultrastructure of the olfactory epithelium has been investigated in many vertebrates including mammals of several orders. Although these reports indicate that there are basic similarities as well as significant differences, epithelial organization does not appear to correlate with evolutionary development or ecological niche, especially among mammals. Few reports on the ultrastructure of insectivore olfactory epithelium are available; therefore, we investigated the fine structure of the olfactory epithelium of Bkirina brevicauda, the short-tailed shrew. This species was especially appropriate because its brain has relatively well-developed olfactory regions. The olfactory epithelium from several male shrews was prepared for transmission electron microscopy and investigated in detail. The basic structure of the shrew olfactory epithelium is similar to that reported for other mammals. A distinctive supporting cell with a different ultrastructure was noted among the more numerous and typical supporting cells. The shrew olfactory epithelium is not distinctive in its organization or complexity compared to that of other mammals, but it does appear to provide the anatomical substrate for significant olfactory sensation. Introduction Schultze (1856, 1862) provided the earliest histological description of the vertebrate olfactory epithelium and distinguished three types of cells — receptor, supporting, and basal cells. Since that time, numerous reports have appeared supporting those observations, and more detailed studies have been made using electron microscopy (Seifert, 1970; Graziadei, 1973). These basic cell types have similar morphology in a wide variety of vertebrates including the Insectivora (Wohrmann-Repenning, 1975). However, signi ficant di fferences have been detected between species even within orders of mammals (Gemmell and Nelson, 1988). There have been few studies of the ultrastructure of the olfactory epithelium in the Insectivora (Graziadei, 1966) and none of these have involved the Soricidae. Since morphometric studies of the olfactory system of shrews have supported the hypothesis that olfaction is of prime importance in these animals (Baron et al., 1983, 1987; Larochelle and Baron, 1989; Stephan et al., 1984), we supplemented previous investigations by examining the fine structure of the shrew olfactory epithelium. The purpose of our study was to make detailed observations on the olfactory epithelium of Blarina brevicauda and to relate that data to what is known about the fine structure of the olfactory epithelium in other mammals. Materials and Methods Ten Blarina brevicauda were collected in the vicinity of Norfolk, Virginia, by live-trapping. Only adult males weighing at least 16 g were used in these studies. Animals were maintained in the laboratory less than 24 hours prior to being sacrificed. During this period the shrews received canned cat food and water ad libitum. Sacrifice was by transcardiac perfusion of fixative in anesthetized animals. The shrews were briefly etherized to facilitate handling and then injected intraperitoneally with 0.2 ml of 2.5 % tribromoethanol. Two shrews were sacrificed without etherization to verify that this brief exposure to ether had no observable effect on olfactory epithelium morphology. A 22-gauge needle was inserted into the left ventricle and the vascular system was perfused with about 10 ml of a room temperature solution of 0.9% sodium chloride, 1% sodium nitrite, 5 mg heparin, and 0.05 M phosphate buffer pH 7.4, followed by 100 ml of room temperature 1.25% glutaraldehyde, 1 % formaldehyde (from paraformaldehyde) in 0. 1 M cacodylate buffer pH 7.4. The perfusion lasted about 30 minutes. Following removal of the brain, the roof of the nasal cavity was excised and turbinates covered with olfactory epithelium were gently removed and placed in fixative containing 2.5% glutaraldehyde, 2% formaldehyde in 0. 1 M cacodylate buffer pH 7.4. During this dissection of the turbinates the epithelium was kept covered with fixative. The olfactory epithelium was fixed overnight at 4°C. Following a 30-min rinse in 0. 1 M cacodylate buffer pH 7.4 with three changes, the tissues were post-fixed for 2 h at 4°C in 2% osmium tetroxide in 0.1 M cacodylate buffer pH 7.4. The tissues were then rinsed in buffer as before, dehydrated in 10-min steps through a graded ethanol series followed by 100% acetone, infiltrated for 6 h with a 1:1 mixture of epoxy resin and acetone, and for 12 h with 100% epoxy resin. The olfactory epithelium was embedded in flat molds oriented so that cross sections of the epithelium could be cut. Following 48-h polymerization of the epoxy resin (Polybed 812, Polysciences, Warrington, Pennsylvania) at 65°C, the blocks were trimmed for sectioning. One-micron sections were cut with glass knives, mounted on glass slides, and stained with methylene blue/azure II (Richardson et al., 1960). These sections were photographed with a Nikon Biophot microscope. Silver thin sections were cut with a diamond knife using an LKB III ultramicrotome, picked up on naked copper grids. 211 212 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 stained with uranyl acetate and lead citrate, and photographed with a JEOL 100 CXII transmission electron microscope. Results Light Microscopic Observations The olfactory epithelium of the short-tailed shrew varied in thickness from about 17 p to over 60 fx in different areas of the nasal cavity. In some instances, thick and thin epithelia, both containing olfactory receptor neurons, were located directly adjacent to each other, typically on opposite sides of the supporting vascular and connective tissue matrix of the turbinate (Fig. 1). The epithelial structure was pseudostratified columnar and the nuclei of the three basic cell types, supporting, receptor, and basal, formed distinct layers (Fig. 1). In some animals, many receptor cells were more darkly stained than supporting or basal cells. In other animals, all three cell types were stained to the same degree. Although no systematic study was done, it appeared that the dark receptors were generally not as well fixed as receptors in epithelial regions where all three cell types were the same density. The dark staining could be due to ischemic changes in the cells that occurred during the perfusion before fixative reached those areas. Dendrites topped by olfactory vesicles were common. Supporting cells had abundant cytoplasm. Their nuclei were uppermost in the epithelium and were rounded and lightly stained. In some animals, a relatively small percentage of supporting cells were distinctly pale. The apical surface of these pale cells typically protruded above the adjacent supporting cells. In other areas of the nasal epithelium or in other animals, all supporting cells had the same degree of staining, but there was a small number of cells identical in other respects to the pale cells. These distinctive supporting cells were designated type 2 and the more common supporting cells type 1. Pale supporting cells were more common in areas of epithelium where dark receptors were present, so it is possible that the pale appearance was also due to ischemic changes. No secretory granules were observed in supporting cells. Basal cells had little cytoplasm and also exhibited varying degrees of staining. Mitotic figures were present in this layer. Large groups of axons with surrounding Schwann cells were common in the lamina propria. Electron Microscopic Observations Receptor neurons varied in cytoplasmic density (Fig. 2) with some cells having more electron-dense cytoplasm than others. Dark cells were found together in various regions of the olfactory epithelium and often appeared to be less well fixed than less dense receptors. The olfactory vesicle was rounded or cylindrical with 10-15 cilia and contained numerous small, empty, membrane-bound structures of a vesicular or tubular nature. Ciliary microtubule structure had the common 9 -f 2 configuration. The dendrite contained many mitochondria and also often had basal body rootlets (Fig. 3). The perikaryal cytoplasm contained numerous mitochondria, sparse profiles of rough endoplasmic reticulum, and a prominent Golgi apparatus (Fig. 2, 4). Type 1 supporting cells had irregular, long, and often branched microvilli on their apical surfaces (Fig. 2, 3, 5). The apical cytoplasm was filled with smooth endoplasmic reticulum that was often aligned in parallel arrays (Fig. 2, 3). Occasional dense granules were observed in the apical cytoplasm. Adjacent supporting cells often appeared to share gap junctions (Fig. 4). Type 2 supporting cells had an apical projection with microvilli that extended above the adjacent supporting cells (Fig. 5). The apical cytoplasm of this type 2 supporting cell lacked the profuse smooth endoplasmic reticulum of the more numerous type 1 supporting cell. The type 2 cell also appeared to have more mitochondria. In some areas of the olfactory epithelium the type 2 cells had cytoplasm with very low electron density. This was common in areas where the receptors had increased density and may have been due to ischemia prior to fixation. In areas with optimal fixation, the type 2 cells had similar cytoplasmic density as the type 1 cells and the receptors. This type 2 cell is similar to a “pale” cell (Kratzing, 1978) and a “microvillar” cell (Moran et al., 1982) previously reported. Basal cells often had pale cytoplasm that contained many polyribosomes and little rough endoplasmic reticulum (Fig. 6). The nucleus had less heterochromatin than either the supporting or receptor cells. Some basal cells had darker cytoplasm, more nuclear heterochromatin, and processes that surrounded receptor cell axons (Fig. 7). Discussion Despite the observation that the olfactory structures constitute a very large proportion of the brain in Blarina (Baron et al., 1983, 1987; Stephan et al., 1984), there does not appear to be a corresponding increase in complexity of the organization of the olfactory epithelium. In its general features, the olfactory epithelium of Blarina brevicauda is similar to that of other mammals (Arstila and Wersall, 1967; Frisch, 1967; Andres, 1969; Graziadei, 1973; Seifert, 1970; Yamamoto, 1976; , Kratzing, 1978; Gemmell and Nelson, 1988). In the only previous ultrastructural study of the olfactory epithelium of an insectivore, Graziadei (1966) reported that the structure of the olfactory epithelium in the mole, Talpa europaea, was similar to that of other vertebrates. The structure of the olfactory epithelium of Blarina differs from that reported for the mole. The olfactory epithelium of the mole was only 30 n thick compared to a maximum of 60 fi in Blarina. The olfactory vesicle of the mole exhibited relatively few cilia, less than five, compared to as many as 15 observed in the shrew. No distinctive variation in supporting cell morphology, such as that of the type 2 cell, was reported for the mole. The significance of these differences with regard to the evolutionary development of the mammalian olfactory system is not apparent at this time. The organization of the olfactory epithelium of Blarina exhibits a complexity that appears to be a sufficient anatomical substrate to support significant olfactory sensitivity. Although it has been suggested that the degree of olfactory system development in shrews may not necessarily indicate a high level of olfactory acuity (Sigmund and Sedlacek, 1985), the basis for such a conclusion does not rest on sufficient behavioral data at the present time. The possible correlation of olfactory system structural complexity and level of olfactory acuity with an 1994 CARSON ET AL.— Olfactory Epithelium of Blarina brevicauda 213 animal’s ecological niche and degree of social interaction remains to be resolved by further investigation. Literature Cited Andres, K. H. 1969. Der olfaktorische Saum der Katze. Zeitschrift fiir Zellforschung und Mikroskopische Anatomic, 96:140-154. Arstila, a., and J. Wersall. 1967. The ultrastructure of the olfactory epithelium of the guinea pig. Acta Otolaryngologica, 64:187-204. Baron, G., H. D. Frahm, K. P. Bhatnagar, and H. Stephan. 1983. Comparison of brain structure volumes in Insectivora and Primates. III. Main olfactory bulb (MOB). Journal fiir Himforschung, 24:551-568. Baron, G., H. Stephan, and H. D. Frahm. 1987. Comparison of brain structure volumes in Insectivora and Primates. VI. Paleocortical components. Journal fiir Himforschung, 28:463-477. Frisch, D. 1967. Ultrastructure of the mouse olfactory mucosa. American Journal of Anatomy, 121:87-120. Gemmell, R. T., and J. Nelson. 1988. Ultrastructure of the olfactory system of three newborn marsupial species. Anatomical Record, 221:655-662. Graziadei, P. P. C. 1966. Electron microscopic observations of the olfactory mucosa of the mole. Journal of Zoology (London), 149:89-94. 1973. The ultrastructure of vertebrates olfactory mucosa. Pp. 267-305, in The Ultrastructure of Sensory Organs (1. Friedmann, ed.), Elsevier North-Holland, Amsterdam, 318 pp. Kratzing, j. E. 1978. The olfactory apparatus of the bandicoot {Isoodon macrourus): Fine structure and presence of a septal organ. Journal of Anatomy, 125:601-613. La ROCHELLE, R., AND G. Baron. 1989. Comparative morphology and morphometry of the nasal fossae of four species of North American shrews (Soricinae). American Journal of Anatomy, 186:306-314. Moran, D. T., J. C. Rowley III, B. W. Jafek, and M. A. Lovell. 1982. The fine structure of the olfactory mucosa in man. Journal of Neurocytology, 11:721-746. Richardson, K. C., L. Jarret, and E. H. Finke. 1960. Embedding in epoxy resins for ultrathin sectioning in electron microscopy. Stain Technology, 35:313. Schultze, M. 1856. Uber die Endigungsweise des Geruchsnerven und der Epithelialgebilde der Nasenschleimhaut. Monatsberichte der Deutschen Akademie der Wissenschaften, 21:504-515. 1862. Untersuchungen uber den Bau der N asenschleimhautm namentlich die Struktur und Endigungsweise der Geruchsnerven bei dem Menschen und den Wirbeltieren. Abhandlungen der Naturforschenden Gesellschaft zu Halle, 7:1-100. Seifert, K. 1970. Die Ultrastruktur des Riechepithels beim Makrosmatiker: Eine Elektronenmikroskopische Untersuchung. Georg Thieme Verlag, Stuttgart, vi + 99 pp. Sigmund, L., and F. SEDLACEK. 1985. Morphometry of the olfactory organ and olfactory thresholds of some fatty acids in Sorex araneus. Acta Zoologica Fennica, 173:249-251. Stephan, H., G. Baron, and H. D. Frahm. 1984. Brains of Soricidae. II. Volume comparison of brain components. Zeitschrift fur Zoologische Systematik und Evolutionsforschung, 22:328-342. Wohrmann-Repenning , A. 1975. Zur vergleichenden makro- und mikroskopischen Anatomic der Nasenhohle europaischer Insektivoren.GegenbaursMorphologischesJahrbuch, 121 :698-756. Yamamoto, M. 1976. An electron microscopic study of the olfactory mucosa in the bat and rabbit. Archivum Histologicum Japonicum, 38:359-412. Fig. 1. — Two areas of olfactory epithelium of different thickness are separated by the lamina propria and cartilaginous (C) supports of the turbinate. Cell nuclei of supporting cells (S), receptor cells (R), and basal cells (B) can be distinguished. Receptor cells, their dendrites, and olfactory vesicles are distinct (arrow). Bar equals 50 /x. 214 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 Fig. 2. — This receptor cell (R) has a prominent Golgi apparatus, numerous mitochondria in the dendrite, several cytoplasmic dense bodies, and several cilia emerging from the olfactory vesicle (V). Type 1 supporting cells (SI) with extensive smooth endoplasmic reticulum surround the receptor. A process resembling the axon emerges from the basal surface of the receptor (arrow). Bar equals 3 fi. Fig. 3.— The olfactory vesicle of this receptor has a more cylindrical, rather than spherical shape. A basal body rootlet (arrow) is present in the dendrite. The adjacent type 1 supporting cells (SI) have profuse smooth endoplasmic reticulum and irregular, complex microvilli. Bar equals 3 n. 1994 CARSON ET AL. — Olfactory Epithelium of Blarina hrevicauda 215 Fig. 4.— The receptor cell (R) perinuclear cytoplasm contains a well-developed Golgi apparatus (G). There are apparent gap junctions between adjacent supporting cells (arrows). Bar equals 3 ii. Fig. 5.— This type 2 supporting cell (S2) lacks the profusion of smooth endoplasmic reticulum present in the adjacent supporting cells. The apical portion of the cell has microvilli and projects above the adjacent type 1 supporting cells. There is a centriole in the apical cytoplasm of the type 2 supporting cell (arrow). Bar equals 3 fi. 216 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 Fig. 6.— This basal cell (B) has many polyribosomes and few other cytoplasmic organelles. Receptor cells (R) and a supporting cell (SI) are nearby. Near the basal cell is the basal lamina (arrow) and a capillary (C) in the lamina propria. Bar equals 3 Fig. 7. — Several receptor cell axons (A) are surrounded by processes (arrows) of this distinctive basal cell (B) with denser cytoplasm and nucleus. An adjacent basal cell (BC) has the more typical morphology. Type 1 supporting cell processes (S) encompass the basal cell. The basal lamina (double arrow) is evident. Bar equals 3 fi. COMPARISON OF PIGMENT AND OTHER DENTAL CHARACTERS OF EASTERN PALEARCTIC SOREX (MAMMALIA: SORICIDAE) Erland Dannelid Department of Zoology, University of Stockholm, S-10691, Stockholm, Sweden Abstract Phylogeny of shrews is based mainly on morphological characters (chiefly cranial and dental characters), and more recently biochemical characters (electrophoretic and karyological). However, many dental characters that are useful for identification of different species are of tittle phylogenetic value. Parallel evolution appears to be frequent, probably often due to ecological adaptations. In this study skulls of eight species of shrews of the genus Sorex from the Palearctic were analyzed. Nine characters and 17 measurements were recorded from each skull. The characters and measurements were compared with those of European species of Sorex. Each character is discussed in terms of usefulness for identification, and phylogenetic and ecological relationships. The two most useful identification characters, the frontal shape of the upper incisors and the shape of the upper antemolars, were shown to have limited phylogenetic value, in many cases because of parallel evolution. Introduction Identification of shrew species is based mostly on quantitative and qualitative skull characters, particularly dental characters. Only qualitative dental characters will be considered in this work. However, to evaluate the correct number of species and of their relationships, it is necessary to combine skull characters (both metric and descriptive) with other morphological characters, and also with karyological and electrophoretic data. Although the number of European species of Sorex is well- known, this is not true for Asian members of the genus. However, many taxonomic publications (Yudin, 1971; Corbet, 1978; Hutterer, 1979; Dolgov, 1985; Hoffmann, 1987) have dealt with this problem. Corbet and Hill (1986) recognized 22 Asian species, whereas Honacki et al. (1982) recognized 24. The red-toothed shrews of the genus Sorex in the Palearctic may be classified into four geographic distribution groups (Dolgov, 1967; Corbet, 1978; Honacki et al., 1982; Corbet and Hill, 1986). Distribution group A includes species with wide distributions in the Palearctic taiga zone (S. araneus, S. caecutiens, S. daphaenodon, S. isodon, S. minutissimus , S. minutus, S. roboratus [includes S. vir, Hoffmann, 1985] and S. tundrensis). Six of these species occur in the Russia, whereas S. daphaenodon and S. roboratus are not known west of the Ural Mountains. Also, six of these species occur east to the Pacific Ocean, whereas S. araneus and S. minutus are not known east of the Yenisei River and Lake Baikal. Distribution group B includes species occurring only in far eastern Siberia (5. cinereus, [according to Ivanitskaya and Kozlovsky {1985}, a complex of three species], S. gracillimus, S. mirabilis, and S. unguiculatus). Distribution group C includes the Caucasian species S. raddei, S. satunini (replaces S. caucasicus Zaitsev, 1988) and S. volnuchini. Distribution group D includes species found only in the central mountain ranges of southern Siberia: S. asper from Tien Shan and S. buchariensis (should, according to Hoffmann [1987], be included in S. thibetanus) from Pamir. Sorex roboratus from Altai is a synonym of S. vir (here in group A), but the name roboratus, Hollister (1913) has precedence over the name vir, Allen (1914). Because S. alpinus occurs outside Europe only in the Carpathian Mountains, it is not considered in this study, which includes only species in groups A, B (excluding S. cinereus complex and S. mirabilis), and C. Methods Skulls of S. daphaenodon (n = 7), S. gracillimus (n = 7), S. raddei (n = 10), S. roboratus (n = 1), S. satunini {n = 7), S. tutuJrensis (n = 12), S. unguiculatus (n — 7), and S. volnuchini {n = 10) were examined under a dissecting microscope. Specimens of S. raddei, S. satunini, and S. volnuchini were obtained from the Zoological Institute, Russian Academy of Sciences, St. Petersburg. Specimens of the remaining species were obtained from the Swedish Museum of Natural History, Stockholm. In addition to qualitative characters, 17 cranial and mandibular characters were measured on each skull (Fig. 1) by use of an image analyzer (MOP Videoplan, Kontron image analysis division, Zeiss). These measurements were compared with the corresponding measurements of nine European species of Sorex (Dannelid, 1990) by using the principal component analysis and the PLS (partial least squares) discriminant analysis (Wold et al., 1983; Stable and Wold, 1988) of the SIMCA (soft independent modelling of class analogy) pattern recognition package (Wold et al., 1984). The qualitative characters examined were: A) the accessory medial tines on the upper incisors, which differ among species in size and position in the pigmented field and thus can be valuable as a species character (Heptner and Dolgov, 1967; Hoffmann, 1971; Diersing and Hoffmeister, 1977; Hennings and Hoffmann, 1977; Junge and Hoffmarm, 1981; Dannelid, 1989) as can the incisor tips (either parallel or divergent). These two characters are connected: species with medial tines high up in the pigmented area usually have divergent incisor tips, whereas those with medial tines placed lower usually have parallel incisor tips (Fig. 2); B) the five teeth posterior to the incisor, usually known as unicuspids, are herein termed antemolars (Reumer, 1984). The relative sizes of these teeth, compared to each other, are usually constant within the species and have been used in species identification (Fig. 3); C) the 217 218 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 hypocones are small cusps situated on the posterolingual parts of P^, M^ and M'. They may be either unpigmented or weakly pigmented (Fig. 4a). In S. daphaenodon they are heavily pigmented (Fig. 4/>). Usually, smaller species have unpigmented hypocones and larger species pigmented hypocones, but the large species S. roboratus has completely unpigmented hypocones; D) the position of the lacrimal foramen relative to the posterior cusps of M* (or rarely the anterior cusps of M^) which differs slightly among species, and is sometimes used as a taxonomic character (van Zyll de Jong, 1980); E) all Sorex species have four cusps (the tip of the incisor is here also regarded as a cusp) on the lower incisor. Sorex minutus and its possible ecological equivalents {S. gracillinius, S. volnuchini) have sharp cusps and little difference between the first depression and the others (Fig. 5a). However, most other species have blunt cusps and a more shallow depression between the first two cusps than between the rest (Fig. 5b)\ F) the second tooth in the lower jaw, herein called Aj (Reumer, 1984) is in S. araneus and some other species triangular in appearance, whereas in other species it has a more elongate shape. This character is visible only in young animals; in old animals with worn teeth the difference disappears; G) Dannelid (1989) recognized three pigment patterns on the first three teeth in the lower jaw: the araneus" pattern, where the border between pigmented and unpigmented areas runs straight and unbroken on Ij, Aj, and P4; the "minutus" pattern, where there is a pigment gap on P4; and the "alpinus" pattern, where the pigment on P4 starts at the base of the tooth. These pigment patterns are useful only for some species (e.g., S. minutus never shows an araneus pattern). None of the non-European species studied in this work showed an alpinus pattern; H) although the mental foramen has been used for taxonomic purposes, the position on the mandible varies not only interspecifically but also intraspecifically; in some instances there is variation in its position between the left and the right mandible of an animal. Most species have the mental foramen situated under the trigonid of Mj; and I) the intensity of color of the pigment on the teeth was examined. Most species have what herein is called medium pigmentation, whereas S. minutus, S. gracillinius, and to some extent S. volnuchini have lighter pigmentation and 5. minutissimus has darker pigmentation. Results Sorex daphaenodon A: The medial tines are small (smaller than in S. araneus) and positioned in the lower part of the pigmented field, close to the middle of this area; the tips of incisors are blunt and parallel (Fig. 2a). B: A'-A'* decrease in size evenly posteriorly (although sometimes A* and A" are of the same size); A^ is clearly smaller and always pigmented (Fig. 3a). C: In all other species the hypocones on the upper molars are either unpigmented or very weakly pigmented. In S. daphaenodon much of the occlusal area of those teeth is pigmented and the hypocones are as strongly pigmented as the other molar cusps. This is by far the easiest way to recognize this species (Fig. 4b). D: The lacrimal foramen is positioned over the metacone of M*. E: The lower incisor has blunt cusps. F: The lower antemolar is triangular (although not to the extent in S. araneus). G: The pigmentation pattern on I1-P4 is mostly an araneus pattern, but there is sometimes a small gap in the pigmentation between Aj and P4. H: The mental foramen is positioned centrally under the trigonid of Mj. I: The pigmentation is medium (as in S. araneus) or a little darker, although one aberrant individual has extremely light pigmentation. Sorex gracillimus A: The medial tines are small, situated in the uppermost part of the pigmented area; the tips of the incisors are sharp and divergent. Compared with S. minutus, the medial tines are smaller, situated higher in the pigment field, and the incisor tips are more divergent. Also, the upper border of the pigmented area is oblique (straight in the similar species S. minutus and S. volnuchini. Fig. 2b). B: The upper antemolars differ from those of S. minutus in that A'-A^ are approximately the same size (A^ is not larger than A^ as in S. minutus)', A“* and A^ are smaller, but the difference in size between these two teeth is not as great as in S. minutus (A^ is larger than in S. minutus and pigmented) (Fig. 3b). C: The hypocones of P^-M^ are unpigmented. D: The lacrimal foramen is over the metastyle of M’. E: The lower incisor is sharp-cusped, almost indistinguishable from that of S. minutus (Fig. 5a). F: The lower antemolar is elongate, nontri angular. G: The pigmentation pattern on I1-P4 is a minutus pattern. H: The mental foramen is positioned under the trigonid of Mj, mostly in an anterior position. I: The pigmentation is light (similar to S. minutus or a little darker). Sorex raddei A: The medial tines are small and situated high in the pigmented area; the incisor tips are slightly divergent (Fig. 2c). B: The upper antemolars are variable, with A*-A^ large, A^ thicker than A', A^-A^ gradually diminishing in size, and A^ unpigmented or weakly pigmented. However, one specimen has similar-sized A*-A^ and another has similar-sized A^-A^. These variations are probably not due to tooth wear, because the specimens are subadults with little wearing of the tooth (Fig. 3c). C: The hypocones are weakly pigmented on M*-M^, never on P*. D: The lacrimal foramen is over the metastyle of M* (somewhat farther back than in the sympatric S. satunini). E: The lower incisor is intermediate between blunt-cusped and sharp-cusped types. The cusps are mostly blunt, but some individuals have cusps almost as sharp as those of S. minutus, S. gracillimus, and S. volnuchini. F: The lower antemolar is triangular and sometimes weakly two-cusped. G: The pigmentation of I1-P4 is an araneus pattern. H: The mental foramen is under the anterior part of Mj, sometimes under P4-M1. I: The pigmentation is medium. Sorex roboratus A: The medial tines are fairly large to small, and very low in the pigmented field. The often longish appearance gives this 1994 DANNELID— Dental Characteristics of Sorex 219 species the appearance of a gigantic S. minutissimus; the tips of the incisors are blunt and parallel (Fig. 2d). B: The upper antemolars are grouped into three size classes (A*-A^, A^-A'*, and A^), much as in S. araneus (Fig. 3d). C: The hypocones of are unpigmented. D: The lacrimal foramen is over the metacone (rarely the metastyle) of M*. E: The lower incisor is blunt-cusped, the cusps sometimes higher and more distinct than in other blunt-cusped forms. F: The lower antemolar is elongate, nontriangular, sharp-cusped, and rarely weakly two- cusped. G: The pigmentation of Ij-P4 is an araneus pattern, but not so constant as in S. araneus and S. raddei. H: The mental foramen is under the trigonid of Mj, in a posterior position. I: The pigmentation is medium. Sorex satunini A: The medial tines are large and positioned in the lower half of the pigmented field, often close to the center of that field; the tips of the incisors are blunt and parallel. The incisors of S. satunini appear very similar to those of S. araneus in frontal view (Fig. 2e). B: The upper antemolars are similar to those of S. araneus, with antemolars in three size groups (A*-A^, A^-A'*, and A^). A” is often thicker than A’, A^ is larger than A'*, and A^ is large and pigmented. It was not always possible to distinguish S. satunini from the sympatric S. raddei (Fig. 3e). C: The hypocones are pigmented on M’-M", sometimes also on F*. D: The lacrimal foramen is over the metacone (rarely the metastyle) of M' (more anteriad than in the sympatric S. raddei). E: The lower incisor is blunt-cusped. F: The lower antemolar is weakly triangular. G: The pigmentation of I1-P4 is sometimes an araneus pattern, but this is not a consistent character. H: The mental foramen is positioned centrally under the trigonid of M,, sometimes a little farther back. I: The pigment color is medium. Sorex tundrensis A: The upper incisor is similar to those of S. araneus and S. satunini, differing only in that the medial tines are smaller, sometimes very small, and rarely absent (Fig. 2f). B: A*-A^ are of similar size, A^-A^ diminish gradually in size posteriorly, and A^ is large and pigmented, similar to the condition in S. araneus (Fig. 3f). C: The hypocones are unpigmented on P* and pigmented on D: The lacrimal foramen is over the metacone of M*. E: The lower incisor is blunt-cusped. F: The lower antemolar is triangular. G; The pigment pattern on I1-P4 is mostly an araneus pattern. H: The mental foramen is positioned posteriorly under the trigonid of Mj, or between the trigonid and the talonid of Mj. I; The pigment color is medium. Sorex unguiculatus A: The medial tines are absent (or in some cases rudimentary in the upper half of the pigmented area); the tips of the incisors are blunt and slightly divergent (Fig. 2g). B: The upper antemolars gradually decrease in size posteriorly, except for A^ which is larger than A^; A^ is large and pigmented (Fig. 3g). C: The hypocones are sometimes weakly pigmented on M* and M“. D: The lacrimal foramen is above the metastyle (sometimes the metacone) of M^. E; The lower incisor is blunt- cusped (Fig. 5b). F: The lower antemolar is triangular. G; An araneus pattern sometimes occurs on I1-P4. H: The mental foramen is under the trigonid of M,. I: The pigment color is medium. Sorex volnuchini A: The medial tines are situated lower in the pigmented field than in S. minutus and S. gracillimus (but still in the upper half); they are otherwise similar to those species, with relatively sharp, widely divergent incisor tips (Fig. 2h). B; The upper antemolars are almost indistinguishable from those of S. minutus. A*-A^ are large, with A^ larger than A^. A'* is smaller, and A^ is very small. However, A5 is mostly larger than in S. minutus (although not so large as in S. gracillimus) (Fig. 3h). C: The hypocones on P*-M“ are always unpigmented. D: The lacrimal foramen is above the metastyle of M*. E: The lower incisor is sharp-cusped and similar to the lower incisors of S. minutus and S. gracillimus, but the first cusp (= tip of incisor) is often stronger and more upturned than in those species. F: The lower antemolar is elongate and nontriangular. G: The pigment pattern on I1-P4 is variable; however, an araneus pattern never occurred. H: The mental foramen is under the trigonid of Mj, somewhat farther back than in S. minutus. I: The pigment color is light, slightly darker than in S. minutus and S. gracillimus. Measurements for 17 characters in the eight species examined are presented in Table 1. For comparison, the corresponding measurements (from Dannelid, 1990) for nine western European species are given in Table 2. A PCA on all 210 individuals of the 17 species resulted in separation of the shrews into two groups: one consisting of smaller species (5. caecutiens, S. gracillimus, S. minutissimus, S. minutus, and S. volnuchini), and another consisting of the remaining species of which S. roboratus and S. unguiculatus were separated as the largest. The first projection of the PCA (X = PC 1, Y = PC 2; Fig. 6a) explained 74.7% of the variance. Overall, the PCA was 81.6% of the total variance. A PLS discriminant analysis on the same matrix (Fig. 6b) explained 33.2% of variance and the separation of the species in the plot was similar to the separation in the PCA. Discussion Dental characters are usually regarded as “good” characters. Even if a skull is badly damaged or recovered from an owl pellet, it is usually possible to find at least part of the dentition intact. In old animals, however, the teeth are worn to a considerable degree, and many dental characters may be obscured. Furthermore, cranial and dental characters can be used not only for identification, but also for phylogenetic inference and for deduction of ecological valences. The characters of the anterior aspect of the upper incisors are useful for identification of species. The position of the medial tines in the pigmented field rarely varies, and the appearance of the incisor tips is also constant among species. 220 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 Species with the medial tines in the upper half of the pigmented area include S. minutus, S. gracillimus, S. volnuchini (the latter two species are probably ecological equivalents to S. minutus), S. raddei, S. samniticus, and S. isodon. In the last species, the medial tines are very small; they are smaller still in S. unguiculatus wherein the medial tines are rudimentary or absent (but situated in the upper half, when present). Finally, the European S. alpinus shows no trace of medial tines. Although the small species S. minutus, S. gracillimus, and S. volnuchini all have medial tines in the upper part of the pigmented area, the position of these structures is not identical. Sorex gracillimus has medial tines placed higher than S. minutus, whereas S. volnuchini has the tines in a lower position. Also, in S. gracillimus the upper border of the pigmented area is oblique, whereas this border is relatively straight in S. minutus and S. volnuchini. Species with medial tines in the lower half of the pigmented area include S. araneus, S. coronatus, S. granarius, S. daphaenodon, S. satunini, S. tumJrensis (all members of the S. araneus I arcticus group defined on karyological grounds, Hausser et al., 1985), S. minutissimus, S. roboratus, and S. caecutiens. The last species shows some variation in the position of the medial tines (Dannelid, 1989). The phylogenetic value of the characters of the anterior aspect of the upper incisors is limited, as all members of the monophyletic S. araneus! arcticus group have similar upper incisors. However, S. minutus and S. gracillimus have almost identical incisors, but are considered distantly related based on both karyological data (Tada and Obara, 1988) and the shape of the glans penis (Dolgov and Lukyanova, 1966). Therefore, this is probably an instance of parallelism. There also may be some ecological significance to the characters of the upper incisors. Some shrews dig in the ground and may construct burrows (Pelikan, 1960). If the incisors are involved in digging, then large medial tines situated low in the pigment field would adapt the incisors for digging. This hypothesis is supported by comparison between S. araneus and S. minutus. Sorex araneus spends a lot of time underground, whereas S. minutus functions more as an epi faunal predator (Croin-Michielsen, 1966). Sorex araneus also has larger medial tines than S. minutus which are situated in the lower part of the pigmented field, whereas in S. minutus they are situated in the upper half of the field. Moreover, the American S. vagrans, which does not make burrows (Terry, 1981), has minute medial tines situated very high on the incisors above the pigment field (Junge and Hoffmann, 1981), a condition that does not occur in Eurasian species. However, S. trowbridgii, which may burrow (Terry, 1981), has small medial tines in the upper part of the pigment field (Junge and Hoffmann, 1981), and S. unguiculatus, which is semifossorial (Yoshino and Abe, 1984; Hutterer, 1985), may have rudimentary medial tines. Thus, although digging may be related to the shape and position of the medial tines, other selective factors are obviously involved. The identification value of characters of the upper antemolars is good. This character is relatively constant, and only S. raddei has a distinct degree of variation. The upper antemolars gradually decrease in size posteriad without apparent gaps in size in S. alpinus, S. caecutiens, S. daphaenodon (A' and A^ often equal in size, might have a size gap between A^ and A^), S. gracillimus, S. isodon, and S. unguiculatus (A^ larger than A^). This situation is often combined with a (relatively) large A^. The upper antemolars are arranged in three size groups (A’-A^, A^-A^, and A^) in S. araneus, S. roboratus, S. satunini (A^ mostly larger than A'*), and sometimes S. raddei. The upper antemolars are arranged in three different size groups (A*-A^, A'*, and A^) in S. minutissimus, S. minutus, S. volnuchini, and sometimes S. raddei. The condition of A* and A^ being of similar size and the other antemolars decreasing posteriorly occurs in S. tundrensis and S. raddei. Tlie condition of A^ being larger than A? occurs in S. minutus (but not S. gracillimus), S. unguiculatus , and S. volnuchini. The phylogenetic value of the antemolar characters is probably not great. In the subgenus Otisorex, A^ is usually larger than or equal in size to A^; whereas, in the subgenus Sorex, A? is usually larger than A“* 1 (Diersing, 1980). However, that study chiefly concerned North ! American members of the genus, and Diersing (1980) also i included in the subgenus Sorex species such as S. merriami and S. trowbridgii which, on electrophoretic grounds, are not " considered to belong in this subgenus (George, 1988). No ji member of the subgenus Sorex has A‘* larger than A^; however, || several species have A^ and A'* of equal size. The ecological significance of antemolar characters is obscure. The antemolars j do not contact the underlying dentition of the lower jaw | (posterior part of Ij, Aj) when the jaw is occluded (Dotsch, n 1985). They might, however, come in direct contact with prey. ! The pigmentation on the hypocones of F*-M^ is excellent for i distinguishing between S. daphaenodon and other species, but is otherwise limited. Dannelid (1989) reported that, among the European species, only S. araneus and S. coronatus have pigmented hypocones on the upper molars; pigmented ’< hypocones on specimens of S. caecutiens and S. isodon from | Kamchatka were later observed. Sorex minutus and its possible ecological equivalents never have pigmented hypocones on | upper molars, and even a large species such as S. roboratus has completely unpigmented hypocones. Species of the chiefly I North American Otisorex complex have unpigmented hypocones on F*^-M^. This character probably has no phylogenetic value. The ecological significance of this character is obscure. The hypocones are small cusps and probably do not serve any major function. Many shrew species show reduction of the hypocones. The strong pigmentation of the hypocones and the strong pigmentation on the upper molars of S. daphaenodon is difficult to explain. The diet of this species includes a large proportion of beetles (Yudin, 1962). However, other species have the same dietary specialization without exhibiting the same pigmentation. Both S. araneus and S. alpinus eat a large proportion of land snails (Churchfield, 1984; Kuvikova, 1986) that might be even harder to crush than beetles; nevertheless, these species have normal pigmentation on the upper molars. As far as is known, the diet of S. daphaenodon is similar to the diets of S. araneus, S. alpinus, S. isodon, and S. tundrensis. The position of the lacrimal foramen has limited identification value. Most species have the lacrimal foramen situated over the metacone or the metastyle of M*, and only in 1994 DANNELID — Dental Characteristics of Sorex 221 S. alpinus is it in a more posteriad position. The phylogenetic value of this character is nil, and it probably has no ecological significance. The identification value of characters of the lower incisor is good within limits. Two distinct morphological types occur. Most species have blunt cusps and a much more shallow depression between first and second cusps than between the others. Sorex minutus and its possible ecological equivalents (5. gracilUmus, S. volnuchini) have sharper cusps and depressions between the cusps of almost equal size. This pattern is also sometimes found in S. raddei. The phylogenetic value of this character is probably low. Sharp cusps may be the result of at least two instances of parallel evolution. This character probably has some ecological significance. The sharp-cusped forms, S. minutus, S. gracilUmus, and S. volnuchini, are small shrews with lightly pigmented teeth (slightly darker in S. volnuchini). They also have l’ with small medial tines situated in the upper half of the pigment field. They probably all live as epi faunal predators, and the limitation of burrowing activities might make it possible for them to maintain sharp cusps on Ij. This does not mean that all species of Sorex that seldom burrow possess lower incisors with sharp cusps; however all burrowing (and some nonburrowing) species are characterized by lower incisors with blunt cusps. The characters of the lower antemolar have limited identification value. Some species, chiefly of the S. araneuslarcti cus group but also S. raddei and S. unguiculatus, have a triangular Aj when not worn; other species have a more elongated Aj. The phylogenetic value of this character is probably limited; e.g., not all members of the araneuslarcticus group have a triangular Aj. Based on karyology S. granarius is regarded as more primitive than S. araneus and S. coronal us (Volobouev and Catzeflis, 1989); however, it does not have a triangular Aj. It might be argued that a triangular Aj is a synapomorphy for more advanced members of the S. araneuslarcticus group, and that the similar shape of the tooth in S. raddei and S. unguiculatus is due to parallelism. This character has obscure ecological significance. Species with a triangular Aj usually eat a large proportion of lumbricids and it is possible that a triangular A^ gives a better hold on the prey if it is transversely placed in the mouth. However, lumbricids are also important food items in many species that have a nontriangular Aj. The highest proportion of lumbricids is eaten by S. mirabilis, occurring in 82.5 % of the specimens (Okhotina, 1969). Unfortunately the shape of the Aj of S. mirabilis is unknown to the author. The pigment pattern on I1-P4 has good identification value in some cases, such as when comparing S. araneus and S. minutus, but many species have an intermediate pattern. The phylogenetic value of this character is probably low, except that the S. alpinus pattern might be an autapomorphy for S. alpinus. This character probably has no ecological significance. The position of the mental foramen has limited identification value (many species have the mental foramen situated below the trigonid of Mj). This character probably has no phylogenetic or ecological significance. The identification value of the pigmentation intensity is limited; most species of Sorex have teeth with approximately the same pigmentation, and some show intraspecific variation. For the smaller species, however, it may be a useful character as S. minutus and S. gracilUmus have lighter pigmentation than other species (S. volnuchini is a little darker), whereas S. minutissimus has the darkest tooth pigmentation of the Eurasian species. The pigmentation character probably has no phylogenetic value. It may, however, have some ecological significance. Red tooth pigment contains iron (Dotsch and Koenigswald, 1978) which may make the teeth more resistant to wear (Selvig and Halse, 1975; Vogel, 1984). However, C. Dotsch (personal communication) believes that red-pigmented enamel is weaker than white enamel, in which case its function must be explained differently. The greatest amount of iron is present on the incisors and on the occlusal surfaces of the molars, especially so on the lower incisor of S. araneus (Dotsch and Koenigswald, 1978). If burrowing is responsible for part of the tooth wear on the incisors, it is conceivable that species that spend little time underground (like S. minutus and S. gracilUmus) would have lightly pigmented incisors. The overall lighter pigmentation of the teeth of these species also may reflect a diet of less sclerotized food. Crocidurines, which lack tooth pigmentation, do not, however, eat softer food than soricines; they may in some instances be adapted for even more powerful food crushing (Dotsch, 1985). But crocidurines have a lower metabolic rate than soricines, and thus probably consume less food, which could account for less tooth wear (Vogel, 1984). Also, carcass feeding (Schliiter, 1980) and burrowing might be responsible for tooth wear during the winter in soricines; the more austral crocidurines come in contact with partly frozen carcasses and soil to a lesser extent than soricines. The function of the red pigment might not be only resistance against abrasion. Different rates of tooth wear between the harder red enamel and the softer white enamel creates sharp cutting edges (Vogel, 1984). Tooth wear is not continuous throughout life, but is more pronounced in overwintered adults than in juveniles (Pankakoski, 1989). Shrews with worn teeth do not switch to a softer diet, at least not in Crocidura russula (Bever, 1983). If the same is true for Sorex, as seems likely, it is conceivable that chewing with teeth heavily worn after the winter causes an even higher rate of abrasion. For a discussion of habitat and food preferences, the size of the species is an important factor. I have therefore grouped the species studied according to size, based on cranial measurements (see tables 1 and 2). Because size relationships are important in interspecific comparisons, it is essential to compare sympatric populations whenever possible. For example, if measurements of a Siberian species such as S. tundrensis are compared to those of European populations of a widespread species such as S. araneus, measurements of S. tundrensis will not be compared with those of sympatric populations of S. araneus, which differ considerably in size from European populations (Hoffmann, 1985). The species studied (17 of the 30 Palearctic species) were divided into seven size classes. Sorex minutissimus, the smallest species, does not overlap much in size with other species. The second size class 222 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 consists of S. minutus. Sorex caecutiens, S. gracilUmus , and S. volnuchini make up the next size class. The two latter species are morphologically very similar to S. minutus-, as none of these three species overlap in distribution, they are considered possible ecological vicars. Sorex minutus and S. volnuchini probably are related, whereas S. gracilUmus might be the result of parallel evolution in eastern Siberia. Sorex caecutiens has smaller mandible measurements than S. volnuchini and in some cases also S. gracilUmus. Sorex tundrensis is usually a little smaller than S. araneus, as is also S. granarius of Spain; these two species constitute the next size class. Sorex araneus, S. coronatus, S. satunini, and S. daphaenodon are all related, not greatly different in size. Sorex samniticus and S. alpinus of Europe are also of the same size class, though the latter species is intermediate between this size class and the next which is made up of S. isodon and S. raddei. Sorex raddei has larger mandibular measurements than S. isodon from Europe. The final size class consists of S. roboratus and S. unguiculatus (5. mirabilis, not included in this study, is larger). A more useful division may be into two groups: “smaller species” (size classes 1-3) and ’’larger species” (size classes 4-7). Although many species of Sorex may occur in the same area, there is little distributional overlap among species in the same size class. The only overlaps found in the eastern Palearctic are as follows: S. caecutiens and S. gracilUmus (size class 2) occur sympatrically in eastern Siberia; S. araneus and S. daphaenodon (size class 5) occur sympatrically in Siberia between the Ural Mountains and Lake Baikal; and S. roboratus and S. unguiculatus (size class 7) occur sympatrically in eastern Siberia. Also, S. caecutiens overlaps in northeastern Siberia with members of the S. cinereus complex which may be approximately the same size. The diets of species of Sorex show some variation between size classes. Sorex minutissimus (size class 1) eats many different kinds of invertebrates but avoids lumbricids and gastropods (Skaren, 1978). Sorex minutus (size class 2) also avoids lumbricids and eats relatively few gastropods (Churchfield, 1984). Sorex caecutiens (and the North American S. cinereus) eats a small amount of lumbricids (Yudin, 1962; Whitaker and Mumford, 1972). Species in size classes 4 to 6 all have diets based largely on lumbricids, coleopterans, and in some cases terrestrial gastropods (Yudin, 1962; Pemetta, 1976; Skaren, 1979; Churchfield, 1984; Kuvikova, 1986). Of the larger forms, S. roboratus concentrates on coleopterans (Yudin, 1962), whereas S. unguiculatus and S. mirabilis feed heavily on lumbricids (Okhotina, 1969; Yoshino and Abe, 1984). Thus, two different feeding types can be recognized among the Eurasian species of Sorex. The smaller shrews (size classes 1-3) probably live as epifaunal predators, not burrowing much. They eat any animal matter they can overpower but avoid lumbricids and (class 1) gastropods. Two distinct complexes of ecological equivalents might be involved here. These are separated from each other by dental characters, particularly by the shapes of the upper and the lower incisors. Sorex minutissimus (and the American S. hoyi) makes up one of these complexes; S. minutus, S. gracilUmus, and S. volnuchini the other. These complexes are not equivalent to size classes 1 and 2 because the size of S. hoyi more closely resembles a small S. minutus than S. minutissimus. Larger shrews (classes 4-7) are often more or less semifossorial, and they often heavily rely on lumbricids and coleopterans in their diets. Sorex caecutiens (size class 3) and the American S. cinereus occupy an intermediate position. They eat few lumbricids, and may live as epigeal predators (Yoshino and Abe, 1984; Aitchison, 1987), at least in the absence of other small shrews. The qualitative characters discussed were all characters chosen with respect to their identification value. No single character separates all species, but a combination of characters is often very useful. All nine European species of Sorex (except S. araneus and S. coronatus) can be separated by the characters cited herein, if the skull is in good condition and the geographic origin of the specimen is known. However, all characters are not equally useful. The best identification characters are the shape of the upper incisors (including the medial tines) and the shape of the upper antemolars. Other characters, such as the pigmentation of the hypocones on the upper molars (for separating S. daphaenodon) and the shape of the lower incisors (for separating S. minutus and its possible ecological equivalents), are of more limited value. Finally some characters like the relative position of the lacrimal and mental foramina are of very limited usefulness and can only be used in combination with other characters. Diersing (1980) listed four qualitative cranial and mandibular characters for separation of the subgenera Sorex and Otisorex. However, none of these characters is constant and no single morphological character can be used for absolute separation of Sorex and Otisorex. Such characters seem to exist in electrophoretic studies (George, 1988). Unfortunately, the phylogenetic value of identification characters seems to be very limited. The characters useful for identification and the characters good for phylogenetic analysis are not the same. Parallel evolution appears to have taken place frequently, in many instances probably due to similar ecological forces. It is likely that the different shapes of the upper and lower incisors among different species of Sorex reflect different ways of living. Incomplete knowledge about the ecology of many species of Sorex prevents us, however, from relating different types of incisors to different ecological niches. Future work in this field, combined with studies of the relationships between different size classes of Sorex should help understanding of ecological niche separation within the genus. Acknowledgments I express gratitude to Dr. M. V. Zaitsev, Zoological Institute of the Russian Academy of Sciences, St. Petersburg, for kindly lending me material. I thank Prof. B. Femholm, Swedish Museum of Natural History, Stockholm, and Dr. M. Malmquist, Uppsala, for critical comments on the manuscript, and Ms. B. Mayrhofer, Stockholm, for assistance with the illustrations. Literature Cited Aitchison, C. W. 1987. Review of winter trophic relations of soricine shrews. Mammal Review, 17:1-24. 1994 D ANNELID— Dental Characteristics of Sorex 223 Allen, G. M. 1914. Notes on birds and mammals of the Arctic coast of east Siberia. Proceedings of the New England Zoological Club, 5:49-66. Bever, K. 1983. Zur Nahrung der Hausspitzmaus, Crocidura russula (Hermann, 1780). Saugetierkundliche Mitteilungen, 31:13-26. Churchfield, S. 1984. Dietary separation in three species of shrew inhabiting water-cress beds. Journal of Zoology (London), 204:211-228. Corbet, G. B. 1978. The Mammals of the Palaearctic Region, a Taxonomical Review. British Museum of Natural History, London, and Cornell University Press, Ithaca, 314 pp. Corbet, G. B., and J. E. Hill. 1986. A World List of Mammalian Species. Second edition. British Museum (Natural History), Comstock Publishing Association, a Division of Cornell University Press, London and Ithaca, 255 pp. Croin-Michielsen, N. 1966. Intraspecific and interspecific competition in the shrews Sorex araneus and S. minutus L. Archives Neerlandaises de Zoologie, 17:73-174 D ANNELID, E. 1989. Medial tines on the upper incisors and other dental features used as identification characters in European shrews of the genus Sorex (Mammalia, Soricidae). Zeitschrift fiir Saugetierkunde, 54:205-214. 1990. Principal component and PLS discriminant analyses applied on skulls of European Sorex (Mammalia, Soricidae). Bonner Zoologische Beitrage, 41:141-156. Diersing, V. E. 1980. Systematics and evolution of the pygmy shrews (subgenus Microsorex) of North America. Journal of Mammalogy, 61:76-101. Diersing, V. E., and D. F. Hoffmeister. 1977. Revision of the shrew Sorex merriami and a description of a new species of the subgenus Sorex. Journal of Mammalogy, 58:321-333. Dolgov, V. A. 1967. Distribution and number of Palaearctic shrews (Insectivora, Soricidae). Zoologicheskii Zhumal, 46: 1701-1712 (in Russian, English summary). 1985. Red-toothed shrews of the Old World. Moscow University, Moscow, 220 pp. (in Russian). Dolgov, V. A., and I. V. Lukyanova. 1966. On the structure of genitalia of Palaearctic Sorex sp. (Insectivora) as a systematic character. Zoologicheskii Zhumal, 45:1852-1861 (in Russian, English summary). Dotsch, C. 1985. Masticatory function in shrews (Soricidae). Acta Zoologica Fennica, 173:231-235. Dotsch, C., and W. v. Koenigswald. 1978. Zur Rotfarbung von Soricidenzahnen . Zeitschrift fiir Saugetierkunde, 43:65-78. George, S. B. 1988. Systematics, historical biogeography and evolution of the genus Sorex. Journal of Mammalogy, 69:443-461 . Hausser, J., F. Catzeflis, A. Meylan, and P. Vogel. 1985. Speciation in the Sorex araneus complex (Mammalia, Insectivora). Acta Zoologica Fennica, 170:125-130. Hennings, D., and R. S. Hoffmann. 1977. A review of the taxonomy of the Sorex vagrans species complex from western North America. Occasional Papers of the Museum of Natural History, The University of Kansas, 68:1-35. Heptner, V. G., AND V. A. Dolgov. 1967. Systematic position of Sorex mirabilis, Ognev, 1937 (Mammalia, Soricidae). Zoologicheskii Zhumal, 46:1419-1422 (in Russian, English summary). Hoffmann, R. S. 1971. Relationships of certain Holarctic shrews, genus Sorex. Zeitschrift fiir Saugetierkunde, 36:193-200. 1985. The correct name for the Palaearctic brown or flat-skulled shrew is Sorex roboratus. Proceedings of the Biological Society of Washington, 98:17-28. 1987. A review of the systematics and distribution of Chinese red-toothed shrews (Mammalia: Soricinae). Acta Theriologica Sinica, 7:100-139. Hollister, N. 1913. Two new mammals from the Siberian Altai. Smithsonian Miscellaneous Colleetions, 60:1-3. Honacki, j. H., K. E. Kin man, and J. W. Koeppl (eds.). 1982. Mammal Species of the World. Allen Press, Inc. and the Association of Systematics Collections, Lawrence, Kansas, 694 pp. Hutterer, R. 1979. Verbreitung und Systematik von Sorex minutus Linnaeus, 1766 (Insectivora; Soricinae) im Nepal-Himalaya und angrenzenden Gebieten . Zeitschrift fur Saugetierkunde, 44:65-80. 1985. Anatomical adaptations of shrews. Mammal Review, 15:43-55. Ivanitskaya, E. Y., and A. I. Kozlovsky. 1985. Karyotypes of Palaearctic shrews of the subgenus Otisorex with comments on taxonomy and phylogeny of the group “cinereus.” Zoologicheskii Zhumal, 64:950-953. Junge, j. a., and R. S. Hoffmann. 1981. An annotated key to the long-tailed shrews (genus Sorex) of the United States and Canada with notes on middle American Sorex. Occasional Papers of the Museum of Natural History, The University of Kansas, 94:1-48. Kuvikova, a. 1986. Nahrung und Nahrungsanspriiche der Alpenspitzmaus (Sorex alpinus. Mammalia, Soricidae) unter den Bed ingungen der T schechoslowakischen Karpaten . Folia Zoologica, 35:117-125. Okhotina, M. V. 1969. Some data on the ecology of Sorex (Ognevia) mirabilis Ognev, 1937. Acta Theriologica, 14:273-284. Pankakoski, E. 1989. Variation in the tooth wear of the shrews Sorex araneus and S. minutus. Annales Zoologici Fennici, 26:445-457. PeliKAN, j. 1960. A burrow constructed by the common shrew (Sorex araneus L.). Zoologicke Listy, 9:269-272. Pernetta, j. C. 1976. Diets of the shrews Sorex araneus L. and Sorex minutus L. in Wytham grasslands. The Journal of Animal Ecology, 45:899-912. Reumer, j. W. F. 1984. Ruscinian and early Pleistocene Soricidae (Insectivora, Mammalia) from Tegelen (The Netherlands) and Hungary. Scripta Geologica, 73:1-173. ScHLUTER, A. 1980. Waldspitzmaus (Sorex araneus) und Wasserspitzmaus (Neomys fodiens) als Aasfresser im Winter. Sagetierkundliche Mitteilungen, 28:45-54. Selvig, K. a., and a. Halse. 1975. The ultrastructural localization of iron in rat incisor enamel. Scandinavian Journal of Dental Research, 83:88-95. S KAREN, U. 1978. Feeding behavior, coprophagy and passage of foodstuffs in a captive least shrew. Acta Theriologica, 23:131-140. 1979. Variation, breeding and moulting in Sorex isodon Turov in Finland. Acta Zoologica Fennica, 159:1-30. StAhle, L., and S. Wold. 1988. Multivariate data analysis and experimental design in biomedical research. Progress in Medicinal Chemistry, 25:291-338. Tada, T., and Y. Obara. 1988. Karyological relationships among four species and subspecies of Sorex revealed by different staining techniques. Journal of the Mammalogical Society of Japan, 13:21-31. Terry, C. J. 1981. Habitat differentiation among three species of Sorex and Neurotrichus gibbsi in Washington. The American Midland Naturalist, 106:119-125. VAN Zyll de Jong, C. G. 1980. Systematic relationships of woodland and prairie forms of the common shrew Sorex cinereus cinereus Kerr and S. c. haydeni Baird in the Canadian Prairie Provinces. Journal of Mammalogy, 68:66-75. Vogel, P. 1984. Verteilung des roten Zahnschmelzes im Gebiss der Soricidae (Mammalia, Insectivora). Revue Suisse de Zoologie, 91:669-708. 224 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 VOLOBOUEV, V. T., AND F. Catzeflis. 1989. Mechanisms of chromosomal evolution in three European members of the Sorex araneus-arcticus group (Insectivora, Soricidae). Zeitschrift fiir Zoologische Systematik und Evolutionsforschung, 27:252-262. Whitaker, J. O., Jr., and R. E. Mumford. 1972. Food and ectoparasites of Indiana shrews. Journal of Mammalogy, 53:329-335. Wold, S., C. Albano, W. J. Dunn III, K. Esbensen, S. Hellberg, E. Johansson, and M. Sjostrom. 1983. Pattern recognition: Finding and using regularities in multivariate data. Pp. 1-36, in Food Research and Data Analysis (H. Martens and H. Russwurm, Jr., eds.). Applied Science Publishers, London, 535 PP- Wold, S., C. Albano, W. Lindberg, and M. Sjostrom. 1984. Modelling data tables by principal components and PLS: Class patterns and quantitative predictive relations. Analusis, 12:477-485. Yoshino, H., and H. Abe. 1984. Comparative study on the foraging habits of two species of soricine shrews. Acta Theriologica, 29:35-43. Yudin, B. S. 1962. Ecology of shrews (genus Sorex) in western Siberia. Trudy, Akademia Nauk SSSR, Siberian Section, Institut Biologii, 8:33-134 (in Russian). 1971. Insectivorous Mammals of Siberia (key). Akademia Nauk SSSR, Siberian Branch, Novosibirsk, 171 pp. (in Russian). Zaitsev, M. V. 1988. On the nomenclature of red-toothed shrews of the genus Sorex in the fauna of the USSR. Zoologicheskii Zhumal, 67:1878-1888 (in Russian, English summary). 1994 DANNELID— Dental Characteristics of Sorex 225 fn VO 00 cn CN so 00 00 0 r- (N ’3 o »— • n; fN 1-H q q q *— • •S c5 d d d d d d d d d d d d d d d d 3 +1 +1 +1 +1 +1 +1 +1 -hi -hi -hi -hi -hi -hi -hi -hi -hi -hi 00 so s Os rs Os so Os (N in in 00 CN in f T § (N 00 so oo in 9 00 q Tf q q q q q '-' 'Cf ni d d d d so d d cn d d J rN| g O os r- o Tf CN Os in so * in d in d oo d d d 00 d cd fd in •2? cn r- (N so O so Os 00 9 os so 8 00 *— ( o § ’o C 3 o c^ 0 +1 u> 3 O o 0 5 •5 « 1. Character d ob c Ji 13 c/5 C6 2, i-> 0> > O a 2 to o u, S»_ O X d •3 a: o -O >> cd 3 3 c« 0) U4 1 *3 3 -2P *S J3 3 a o G G i-i 0) «x Oh G O U—i ■l O a Ui 0) O. cu G O 3 3 G 0> (h o 3 3 G 3 a oo S *7? c/5 "O G G G G a 3 3 G 3 3 G B G G CU H 0 U u< m Vh u (-C U 3 Ph O a/ d: Q Ui 0 Table 2. —Means (in mm) + SD of cranial and mandibular characters recorded from skulls of nine species of Sor&x from Europe. SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 C c: cj § K .5, U oo o r^ o 00 VO NO O r*' oo o r4 1—^ q 1 kM o KM o d d d d d d d d d d d d d d d II +1 +1 +1 +1 -hi -hi -hi -hi dl 3 3 dl dl 3 dl dl oo o cn o KM o VO o ON -H nJ 00 r- VO D. Im 1 C/5 o Oh s Ui o t-i ■i 3 3 3 3 o o O o -o 3 C3 3 3 -S CdO a cd 'O s 2 V5 o u St-H O •3 O •S 4) O. Oh =3 u, ’3 3 cd u t-i iS *>< 3 0) 1=1 *3 u O Ut 3 •£f *S 43 3 3 u 4) a- a- :3 1+-. o 3 3) ntemolars 1h 4> Oi Oi :3 o 3 3i 3 4> S Ih 3 *c 3 3 3 3 cd 4) (-• X) 3 3 O O Ih a 4) 4) 4) X 4> O a CU c/5 c/5 4) O O u. Oh 3 *o o 3 o iX 1m 3 3 o 3 fi o o (m 3 3 X 3 c cc5 1h _o <4-1 o M X 3 o CC 6 o 3 3 3 4> O X 4> O 1m 3 C o 0> ci3 c cd c S cd *0 4) a O. c/5 o cd S a o 2 o U CQ 1-1 u «H u 3 o, o 3 X Q o d +1 (S o +1 cn s d +1 00 4W05».'jr.- ■ . i 1 ''T R I- W J > _ . ' i' f -r ■ * i . ■"'i'" t * " I ■i ^ \ I ■n»i - »1 ... ‘ V ,' . .. .A\ ' t rv ' iM& ■? fetfi’ Apl — -^ ' ' ^ ■ ’a '■ v>5,«: r.-ii .'i fit ♦'arj* W3'*a[ 3t!>7y{>^ 'i'" FUNCTION OF THE FEEDING APPARATUS IN RED-TOOTHED AND WHITE-TOOTHED SHREWS (SORICIDAE) USING ELECTROMYOGRAPHY AND CINERADIOGRAPHY Christel Dotsch Department of Anatomy and Embryology, University of Groningen, Oostersingel 69, 9713 EZ Groningen, The Netherlands; current address: Briickenstrafte 33, W-6238 Hofheim 7, Germany Abstract Electromyography and cineradiography of masticatory action in crocidurine and soricine shrews elucidated feeding strategies in relation to their unusual skull features. Multichannel EMG demonstrated differences in masticatory rates as well as in EMG activity levels of masticatory muscles between the two soricid groups. Parts 1 and 2 of M. temporalis primarily exerted pressure on food while the pinnated part 3 of the muscle produced a force concentration along one line of action and, thus, acted as a guiding muscle. M. masseter and M. pterygoideus medialis, as a complex, adducted the jaws with lateral components of movements. The dorsal part of M. pterygoideus lateralis supported the digastric as a jaw opener. The activity of M. pterygoideus medialis in an intermediate phase between the power phase and jaw opening was correlated with jaw tilting. Graphic presentation of masticatory orbits as well as distance calculations from markers in dorsoventral, lateral, and frontal planes illustrated jaw movements. Rotational movements at the inferior articulation of the double jaw joint during jaw opening were intensified in soricines, thus providing better alignment of the teeth for effective grinding movements. In crocidurines forceful transverse motions were gained due to the anatomy of the articulation and the masticatory muscles. Introduction Detailed experimental studies in the last two decades on mastication in mammals of various feeding types abound with electromyographical data and models of three-dimensional jaw movements (for reviews, see Cans et al., 1978; Gomiak, 1985). Based on these studies, many investigators have pointed out that one chewing cycle pattern is probably common to all mammals (Hiiemae, 1978; Butler, 1983). The skulls of soricids are in some respects unique among mammals; there is no zygomatic arch, and the lower jaws have a “double mandibular articulation” and a “fossa temporalis interna.” These features are related to masticatory muscle function (Dotsch, 1982, 1983a, 1986). Furthermore, striking differences in the anatomy of the double jaw joint between the shrew subfamilies (Dotsch, 1983^) make it clear that there is a need for additional information concerning the masticatory behavior of soricids. None of the previous anatomical studies of the feeding apparatus of shrews reported details of their chewing patterns (see review in Dotsch, 1982), and only two investigations mentioned jaw movements of soricids (Feamhead et al., 1954; Gasc, 1963). Feeding has been studied in the insectivorous bat Myotis lucifugus (Kallen and Gans, 1972) and in the insectivore Tenrec ecaudatus (Oron and Crompton, 1985). However, a member of the shrews (Insectivora, Soricidae) has never been studied electromyographically. A comparison of mastication between the insectivorous tenrecs and soricids also should provide an understanding of early mastication in mammals. Tenrecs as well as soricids have no zygomatic arch. It should then be determined if a “general mammalian chewing pattern” (Hiiemae, 1978) does exist. Comparative studies on the functional morphology of mastication in mammals (among others, Storch, 1968; Biihler, 1977) and papers on the philosophy of this subject (Herring, 1988) have shown that studies in mastication should be more than sampling experimental data across mammalian orders and families. Moreover, representatives of both soricid subfamilies (Crocidurinae, Soricinae) occupy a variety of different food niches and environments (Hutterer, 1985). Hence, it would be instructive to know both the way distinct forms of the shrew masticatory apparatus and its functions are linked, and whether there are differences in the chewing patterns of white-toothed (Crocidurinae) and red-toothed (Soricinae) shrews. Materials and Methods Twelve specimens of Suncus murinus (ranging from 30 to 50 g) and seven of Crocidura flavescens (ranging from 30 to 80 g), both members of the subfamily Crocidurinae, and four specimens of Blarina brevicauda (ranging from 21 to 27 g), a member of the subfamily Soricinae, were used for electromyography (EMG) and cineradiography. The white- toothed shrews were from breeding populations whereas the red-toothed shrews were wild-caught animals. The animals were trained to take food (mealworms, pieces of chicken breast and heart, and soft commercial cat food) during the experiments after a period of fasting. Electromyography Electrodes (up to 12, 37, or 50 pm bipolar nickel-chromium wires, Clark, Medical Instruments) were implanted under halothane anesthesia with a 2 1 -gauge hypodermic needle (for details see Dantuma and Dotsch, 1989a) to simultaneously record activities of right and left masticatory muscles (Dotsch and Dantuma, 1989a). The electrodes were placed in the following muscles: M. digastricus, M. masseter, M. pterygoideus medialis, and three parts of M. temporalis. The M. pterygoideus lateralis was too small for electrode implantation. The exact positions of the electrodes were later determined by dissection. 233 234 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 The EMG signals were amplified with a Princeton Applied Research 113 preamplifier at a frequency range of 30 Hz to 10 kHz. The synchronization pulses of the film frames were generated by a mechanism of a 16-mm camera (Arriflex) on Kodak Ektachrome film. The pulses and the EMG signals were displayed on a Tektronix R 5103 N oscilloscope and a multichannel chart recorder (Siemens Oscillomink B), and stored on a 14-channel tape recorder (Bell & Howell CEC/VR 3369) at a speed of 15 or 7.5 ips. A mirror placed at 45° to the lateral side of the animal allowed a split view (lateral and ventral) recording. The EMG signals were digitized with an A/D converter and then analyzed with an IBM computer using programs that determined the percentage distribution of muscle activity (written by R. Dantuma). Maximal jaw opening was determined by frame-by-frame analysis and served as a basis for the grouping of masticatory cycles. It was impossible to do EMG of the masticatory muscles of shrews in the same way as in larger animals (Weijs and Dantuma, 1975; De Gueldre and De Vree, 1988). The small size of the shrews restricted application of multipin connectors. Consequently, we developed an alternative system (Dantuma and Dotsch, 1989a). After implantation, the electrodes formed a 30-cm cable leading out at the dorsal region of the head. Their free ends were soldered to a connector. The cable was then protected with silicone rubber (Silastic 285R). The connector was suspended from a rotating unit when the shrews were not in experiments. This allowed the animals to move relatively freely in their cages. Cineradiography To follow the complicated jaw movements radiographically, dental parapulpar pins (TMS Link Series, self-shearing, gold- plated, stainless steel pins) were implanted in the skulls of the shrews (Dantuma and Dotsch, 1989/?). The standardized size and shape of the pins, and the stability of pin position made all marks comparable. Also, for any given pin two marks were defined, screw-head and end. This is especially important for small objects. The shrews were filmed during feeding with an Arriflex 16-mm camera at 50 frames/sec. The radiographic films (Agfa Gevapan 30) were taken with a Siemens X-ray apparatus equipped with a 70-mm lens (Tridoros optimatic 800, with a Sirecon-2 image intensifier, 48-56 kV, pulse duration 1 msec) in either a dorsoventral or lateral projection. The cineradiographic films were analyzed frame by frame with an “Old Delft” variable speed analytical projector. The X and Y coordinates of 20 marks on the film frames, i.e., those belonging to one masticatory cycle, were digitized with a Calcomp 2500 digitizer, and recorded on an IBM- compatible computer. Dorsoventral and lateral two-dimensional X-ray photographs (displaced at an angle of 90°) were also taken and their X and Y coordinates determined. A PC program then calculated the X/Y/Z coordinates using these coordinates, another (both programs were written by H. Amesz-Voorhoeve) rotated the three-dimensional coordinates onto the two- dimensional projection of the film frames and recalculated their X/Y/Z coordinates. Graphic presentation of the masticatory orbits as well as distance calculations from the markers in dorsoventral, lateral, and frontal planes illustrated jaw movements. Results and Discussion Morphology The masticatory apparatus of the white-toothed Crocidura flavescens and Suncus murinus and the red-toothed Blarina brevicauda differ with respect to the double mandibular articulation and masticatory musculature. Because the skull structures and masticatory muscles were described in Dotsch (1982, 1983a, 1983/?), only a brief summary of their essential features is included. The two condylar facets of the jaw articulation in crocidurines are connected by a bony ridge. The dorsal facet carries a large articular disc which extends to the ventral facet (Dotsch, 1983/?). The pterygoid bone does not contribute to the formation of the glenoid fossa. The condylar facets of soricines are clearly separated, and the disc covers only the dorsal facet. The jaw articulation with its glenoid fossa is relatively larger than that in crocidurines (Dotsch, 1983/?). In soricines, the pterygoid bone is part of the glenoid fossa. The mass of masticatory muscles relative to total body mass in white-toothed shrews is larger than in red-toothed shrews (Dotsch, 1982, 1983a). The percentage of M. pterygoideus lateralis of the total masticatory musculature is greater in Blarina than in Crocidura and Suncus. In contrast, M. masseter and M. pterygoideus medialis in the white-toothed shrews are somewhat larger than in Blarina. M. digastricus in the white- toothed shrews is slightly smaller. Morphologically, M. temporalis showed the greatest differences among the species investigated. Its three main parts, as well as M. suprazygomaticus and M. zygomaticomandibularis, are well- developed in crocidurines (Dotsch, 1983a). The characteristic pinnation of part 3 is particularly well-expressed in Suncus and Crocidura. The large part 1 of M. temporalis which attaches anteriorly at the coronoid process is, in these species, mainly longitudinally oriented whereas in Blarina it faces laterally to the coronoid process. In spite of differences in some aspects of tooth morphology among the species studied, in all shrews the upper dental arch is wider in the molar region than is the lower arch. This produces unilateral occlusion along labial parts of the molars at the ipsilateral side. According to the general anatomy in B. brevicauda described above, I predicted a greater variety of movements would be possible in this species than in the white- toothed shrews. Masticatory Patterns As in other mammals with anisognathous jaws, mastication is unilateral in shrews. A masticatory sequence included initial grasping of food, followed by chewing and repositioning cycles. The results indicate that the stage of the reduction sequence as well as food size and food consistency determined the lengths of these cycles. Mealworms were masticated with the most regular changing of the ipsilateral active side, in a course of five to seven cycles per side. The soft cat food mixed with 1994 DOTSCH — Function of Feeding Apparatus in Shrews 235 small resistant particles was chewed slightly more irregularly. Food of homogeneous, tough consistency such as pieces of chicken heart and chicken breast was preferentially chewed on one side of the jaws during the reduction sequence. Changing to the other side was infrequent. Masticatory patterns and chewing frequencies in mammals are determined essentially by the consistency of food and the individual’s bite (Herring and Scapino, 1973; Hiiemae, 1978; Thexton et al., 1980; Fish and Mendel, 1982). The masticatory cycle time of shrews varied with the type of food presented. For example, in C. flavescens, the mean chewing cycle duration was shortest on soft cat food (212.0 ± 18.4 msec). Chewing on mealworms averaged 242.8 + 7.6 msec, on pieces of chicken heart 238.4 ± 15.2 msec, and on pieces of chicken breast 231.0 + 20.2 msec. A comparison of the masticatory rates of several shrew species revealed correlation with subfamily, independent of body size. The white-toothed Crocidura flavescens masticated at a rate of 4.6 orbits per second, Suncus murinus at 5.5, and Crocidura russula at 5.9. The red-toothed Sorex araneus chewed at a rate of 7. 1 orbits per second, the European water shrew Neomys fodiens at 7.7, Blarina brevicauda at 7.9, and Cryptotis parva at 7.7. Thus, chewing rates were significantly higher in soricines than in crocidurines. Differences were tested by the Mann-Whitney U test of significance (a = 0.05). An association with the anatomical differences in the jaw articulation and the masticatory musculature as well as in the muscle activities (EMG) is possible. Additionally, Vogel (1981) has shown that the two subfamilies of the Soricidae have different metabolisms. Usually, masticatory cycles are divided into closing (fast and slow closing, FC and SC) and opening phases (slow and fast opening, SO and FO, Hiiemae, 1978). A “general mammalian chewing pattern” postulated by Hiiemae involves a closing stroke and an opening stroke. Subsequent studies (reviewed by Gomiak, 1985, except the article on Tenrec ecaudatus by Oron and Crompton, 1985) suggested deviations from the presumed generalized pattern owing to differences in dentition, jaw articulation, and jaw musculature (De Gueldre and De Vree, 1988). Because of the absence of a zygomatic arch, in both tenrecs and in soricids, it is of interest to determine the significance of muscle activity and jaw movements in these animals. In S. murinus (Dotsch, 1986), I found similar percentages for opening (61-66%) and closing phases (34-39%), compared to a cycle for shrews and tenrecs (Oron and Crompton, 1985). However, if these percentages are compared with those in other shrew species, differences between the two soricid subfamilies become apparent. Depending on the food that was chewed, in B. brevicauda percent time for opening (30.6-44.2%) and closing (55.8-69.4%) phases were reversed in contrast to those recorded for C. flavescens and S. murinus (opening 52.6-64.7%, closing 35.3-47.4%). Gamble (1979) reported a percentage of 62% for closing and power stroke (definition of terms in Gamble’s work) and 38% for opening in B. brevicauda. In the white-toothed shrews, the long opening phase is affected by prolongation of the slow opening phase (Dotsch, 1986). In cats, tree shrews (Tupaia), and fruit bats, the significance of the SO phase is related to food transport by the tongue (Thexton et al., 1980; Mendel et al., 1981; Fish and Mendel, 1982; De Gueldre and De Vree, 1988). If there is a long FO/FC complex (for example in cats, Thexton et al., 1980) and a short SO phase, the food is chewed with the cheek teeth. Meeting the conditions in shrews, the very long SO phase, especially in 5. murinus, is associated with movements at the mobile symphysis and of the tongue (Dotsch, 1986). Given the configuration of the jaw-joint in Blarina, the more obliquely-oriented masticatory muscles, and the occlusal surfaces of the cheek teeth, a greater variety of jaw movements is postulated in this genus. Additionally, food may be chewed better in red-toothed shrews (Dotsch, 1982). This might affect the longer time needed for jaw closing. Electromyography The stage of the reduction sequence and the type of food determined the length of masticatory cycles. These, in turn, affected the EMG pattern which generally varies with each cycle. This made it difficult to average the myograms. Figure 1 presents myograms of right and left masticatory muscles of C. flavescens during chewing on pieces of chicken heart in the middle of a chewing series. The EMG levels of the muscles on the ipsilateral (active) and the contralateral (balancing) sides were asymmetrical. A clear asynchronous EMG pattern occurred in the large M. temporalis. The right, relatively small, pinnated part 3 of M. temporalis (M.TE3R) fired in the power phase before the left muscle (M.TE3L) was activated. The large M. temporalis 2 started almost simultaneously on both sides in the closing phase. The EMG level, however, in the first four cycles was higher in the right portions of M. temporalis than in the left. This indicated that the right side was the ipsilateral chewing side. In the last three cycles, left M. temporalis parts (M.TE3L, M.TE2L) were highly active, expressing the fact that the left jaw was the active side. Another pattern occurred in M. pterygoideus medialis and in M. masseter. The right M. masseter (M.MASR) and M. pterygoideus medialis (M.PTMR) were in the closing phase longer than the contralateral muscles. Additionally, the EMG levels of M. masseter and M. pterygoideus medialis on the balancing side were higher than on the active side. M. digastricus showed main activity during the opening phase, which was slightly intensified on the active side. In two of the cycles (Fig.l), the left M. digastricus (M.DIGL) showed activity in an intermediate phase between the power phase and fast opening. The first of these cycles was a reversal of the ipsilateral side from right to left. The intermediate activity of M. digastricus occurred simultaneously with a strong action of the left M. pterygoideus medialis and a weaker action of the right M. pterygoideus medialis. In general, M. pterygoideus medialis may be firing in the power phase and in an intermediate phase during slow opening. These observations suggest that M. temporalis with its main parts 1 and 2 is the primary muscle exerting pressure on food. 236 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 The pinnated part 3 was mainly acting as a guiding muscle. M. masseter and M. pterygoideus medialis are primary adductors of the jaws. M. digastricus is a jaw opener with lateral components of movement. Table 1 summarizes the activity of the masticatory muscles in C. /laves cens. The number of pluses for the degree of muscular activity in a respective phase of a cycle showed the trends of muscle functions. M. pterygoideus lateralis is two-headed. The activity of the superior head was recorded by chance. Depending on the food that was chewed, it showed high levels of activity during the power phase and the fast opening phase. Previous observations on dried skulls and muscles (Dotsch, 1982) confirmed the presumption that the superior M. pterygoideus lateralis supports M. digastricus as a jaw opener. This also occurs in Tenrec ecaudatus (Oron and Crompton, 1985), although it is not clear if the superior or inferior head of the muscle was recorded. The M. pterygoideus lateralis muscle weight in soricines was greater than in crocidurines (Dotsch, 1983a). The inferior M. pterygoideus lateralis would primarily cause laterally directed jaw closing movements. If crocidurines and soricines are compared, M. pterygoideus lateralis presumably will be more active in movements with guiding components in Blarina than in Crocidura or Suncus. In this context, M. mylohyoideus of C. flavescens (by chance punctured) was highly active in the intermediate phase and during fast jaw opening. From the line of action observed, this muscle inverted the mandibles so they rotated along a longitudinal axis. The resulting outward tilting of the mandibles can be seen as movements at the loose symphysis separating the lower incisors (Dotsch, 1986). The simultaneous action of M. pterygoideus medialis and M. mylohyoideus in the intermediate phase are considered to be responsible for rotational movements of the jaws. Cineradiography Figure 2a shows the positions of the implanted pins that served as markers for following complex jaw movements cineradiographically. Number 1 is a calculated midpoint between points 21 and 22 in the centers of left and right articulations. These points, 21 and 22 (not presented in the figure), are not implanted markers but estimated points. Calculation of the X/Y/Z coordinates of the cineradiographic film frame markers allowed estimation of distance changes between the pin points (Dantuma and Dotsch, 19896). These changes showed the moving jaws during mastication and thus reproduced the movement pattern. Principally, each orbital movement varied from the other. Figure 2b shows the movement orbits of points on the tip of the coronoid process and at the lower jaws in frontal view from one masticatory cycle. This cycle was the second in a series of three successive cycles (Fig. 2c, d) taken while B. brevicauda masticated a mealworm. The rotation of both lower jaws about the midline of the head (Fig. 2b, frontal view, line 1/2) is clearly demonstrated in these figures. The mobile symphysis enables relatively independent rotational movement of the mandibles. Points 9 and 10 on the tip of the coronoid process of the ipsilateral left side moved much more medially than points 15 and 16 on the contralateral side (Fig. 2b, c). Larger rotation of the lower jaw at the inferior jaw articulation produced stronger outward deflection of the coronoid process on the working side during jaw opening. This was associated with movements of the various M. temporalis parts acting on the coronoid process of the lower jaw. The markers on the mandibular ramus, which lie anterior to the jaw joints, however, differed less in their movement pattern on left and right sides (Fig. 2d). The relatively weak M. masseter/M. pterygoideus medialis complex acting on the mandibles would effect slight differences of movement on the working and the balancing side. Conclusions Chewing in soricids involves a combination of jaw movements that includes shearing, crushing, and grinding functions. The movements of the lower jaws in shrews vary from cycle to cycle and are conditioned by the double jaw joint and the mobile symphysis. This is reflected in the complex EMG patterns of the masticatory muscles. Additionally, masticatory muscle functions as well as jaw movements do not suggest a simple classification into a carnivorous, omnivorous, or herbivorous feeding type. The dissimilarities in the masticatory patterns of white-toothed and red-toothed soricids are explained by differences in the structure of the jaw joint and of the chewing muscles. In all shrews, M. temporalis is the primary muscle exerting pressure on the food while M. masseter and M. pterygoideus medialis act as a complex to adduct the jaws with lateral components of movement. The superior part of M. pterygoideus lateralis supports M. digastricus as a jaw opener. M. pterygoideus medialis shows an immense repertory of muscle activity in most cycle phases. Its action in an “intermediate” phase is correlated with movements between the power phase and jaw opening which can be seen as tilting of the mandibles. A strong outward rotation of the coronoid process during opening is coupled with rotation at the inferior articulation of the double jaw joint. The large parts 1 and 2 of M. temporalis produce the main force that is concentrated in the crocidurines on the tip of the coronoid process. In soricines, the lines of force are laterally directed along the process. The pinnated part 3 of M. temporalis provides a force concentration along one line of action (Gamble, 1979). It functions most likely with M. suprazygomaticus and M. zygomaticomandibularis (the latter two muscles were too small for experiments) in stabilization of the lower jaw, especially in the area of the double joint. Given the anatomy, strong rotational movements will be prevented in crocidurines but forceful transverse motions will be enhanced. Rotational movements are intensified in soricines and will provide better alignment of the molars for effective grinding masticatory movements. The unique skull features of all shrews, such as the absence of a zygomatic arch, the existence of a double mandibular articulation, and a fossa temporalis interna, result in a variety of muscle activities and jaw movements. The variation of feeding behavior is greater in soricids than in other mammals, even than in the closely-related tenrecs (Oron and Crompton, 1994 DOTSCH — Function of Feeding Apparatus in Shrews 237 1985) which also have no zygomatic arch. Therefore, a “typical” mammalian chewing pattern (Hiiemae, 1978) is not common to all mammals. Finally, different metabolisms in the two soricid subfamilies (Vogel, 1981) indicate divergent biological strategies. The different feeding strategies of red-toothed and white-toothed shrews are reflected in anatomical differences of the masticatory apparatus. Considering the disparity and universal presence of soricines and crocidurines and their adaptation to different ecological niches (Hutterer, 1985), variation in food acquisition and masticatory behavior of these “primitive” insectivores was undoubtedly a basis for successful evolution. Acknowledgments The following persons are gratefully acknowledged: G. L. Dryden, J. F. Merritt, O. B. Mock, and P. Vogel for the gift of experimental animals; R. Dantuma and L. Koese for technical help in EMG work, and F. de Vree and collaborators in cineradiography work; R. Dantuma for EMG computer programs; H. Amesz- V oorhoeve and L. Stijnen for cineradiography computer programs. Part of this work was supported by DFG grant DO 299 from the Department of Anatomy and Embryology, University of Groningen, The Netherlands; and FKFO 2.9005.84 grant to F. de Vree, Antwerp, Belgium. Literature Cited Buhler, P. 1977. Comparative kinematics of the vertebrate jaw frame. Fortschritte der Zoologie, 24:123-138. Butler, P. M. 1983. Evolution and mammalian dental morjihology. Journal of Biology Buccale, 11:285-302. Dantuma, R., and C. Dotsch. 1989a. Technical aspects of electromyography of the masticatory apparatus in shrews (Soricidae). Progress in Zoology, 35:151-152. 1989fc. A new marker technique in cineradiography for the recording of movements in small vertebrates— Application to the study of jaw movements in soricids (Insectivora). Experientia, 45:702-705. De Gueldre, G., and F. De Vree. 1988. Quantitative electromyography of the masticatory muscles of Pteropus giganteus (Megachiroptera). Journal of Morphology, 196:73-106. Dotsch, C. 1982. The masticatory apparatus of Soricidae (Mammalia, Insectivora). An investigation of the functional morphology of masticatory action in shrews of the genera Sorex L., Neomys K. and Crocidura W. Zoologische Jahrbiicher, Anatomic, 108:421-484. 1983a. Morphologische Untersuchungen am Kauapparat der Spitzmause Suncus murinus L., Soriculus nigrescens G. und Soriculus caudatus H. (Soricidae). Saugetierkundliche Mitteilungen, 31:27-46. 1983fc. Das Kiefergelenk der Soricidae (Mammalia, Insectivora). Zeitschrift fiir Saugetierkunde, 48:65-77. 1986. Mastication in the musk shrew, Suncus murinus (Mammalia, Soricidae). Journal of Morphology, 189:25-43. Dotsch, C., and R. Dantuma. 1989. Electromyography and masticatory behavior in shrews (Insectivora). Progress in Zoology, 35:146-147. Fearnhead, R. W., C. C. D. Shute, and A. d’A. Bellairs. 1954. The temporomandibular joint of shrews. Proceedings of the Zoological Society of London, 125:795-806. Fish, D. R., and F. C. Mendel. 1982. Mandibular movement pattern relative to food types in common tree shrews (Tupaia glis). American Journal of Physical Anthropology, 58:255-269. Gamble, C. 1979. Function of the masticatory apparatus of the short- tailed shrew (Blarina brevicauda) with special reference to the evolution of the soricid temporo-mandibular joint. Unpublished M.S. thesis. University of Massachusetts, Amherst. Gans, C., F. De Vree, and G. C. Gorniak. 1978. Analysis of mammalian masticatory mechanisms: Progress and problems. Zentralblatt fiir Veterinar Medizin Reihe C. Anatomic Histologic Embryologie, 7:226-244. Gasc, J. P. 1963. La musculature cephalique chez Suncus Ehr., Crocidura Wag., Sylvisorex Thom., Myosorex Gr. Mammalia, 27:582-601. Gorniak, G. C. 1985. Trends in the actions of mammalian masticatory muscles. American Zoologist, 25:331-337. Herring, S. W. 1988. Introduction: How to do functional morphology. American Zoologist, 28:189-192. Herring, S. W., and R. P. Scapino. 1973. Physiology of feeding in miniature pigs. Journal of Morphology, 141:427-460. Hiiemae, K. M. 1978. Mammalian mastication: A review of the activity of the jaw muscles and the movements they produce in chewing. Pp. 359-398, in Development, Function and Evolution of Teeth (P. M. Butler and K. A. Joysey, eds.). Academic Press, New York and London, 523 pp. Hutterer, R. 1985. Anatomical adaptations of shrews. Mammal Review, 15:43-55. KaLLEN, F. C. , AND C. Gans. 1972. Mastication in the little brown bat, Myotis lucifugus. Journal of Morphology, 136:385-420. Mendel, F. C., W. Hicks, D. R. Fish, and F. C. Kallen. 1981. Oral and hyoid movement patterns related to mastication in giant fruit bats, Pteropus giganteus. Bat Research News, 22:47. Oron, U., and a. W. Crompton. 1985. A cineradiographic and electromyographic study of mastication in Tenrec ecaudatus. Journal of Morphology, 185:155-182. Storch, G. 1968. Funktionstypen des Riefergelenks bei Saugetieren. Natur und Museum, 98:41-46. Thexton, a. j., K. M. Hiiemae, and A. W. Crompton. 1980. Food consistency and bite size as regulators of jaw movement during feeding in the cat. Journal of Neurophysiology , 44:456-474. Vogel, P. 1981. Metabolic levels and biological strategies in shrews. Pp. 170-180, in Comparative Physiology of Primitive Mammals (K. Schmidt-Nielsen, ed.), Cambridge University Press, Cam- bridge, England, 338 pp. Weus, W. a., and R. Dantuma. 1975. Electromyography and mechanics of mastication in the albino rat. Journal of Morphology, 146:1-34. 238 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 Table 1. — Activity of the masticatory muscles during jaw closing and opening in Crocidura flavescens. For each muscle, symbob are as follows: First row, three pluses, main activity in a distinct phase; two pluses to one plus, high level to low level activity; plus/minus, occasional activity; minus, no activity. Second, row, I, ipsilateral, and C, contralateral activity; I~C, I synchronous C; I*^C, I later than C, I-^C, I before C. Third row, iEMG leveb of ipsilateral and contralateral muscles in the main activity phases; /< C, I lower than C; 1> C, I higher than C; I—C, 1 similar to C. Closing Opening 1 Masticatory Cycle Phases -I Fast 1 Slow Slow 1 Fast “Power" Phase Intermediate Opening M. digastricus + ± to - + + + KC I>C M. pterygoideus medialis + + + + + to + i I^ I = C KC I-C M. masseter + + 4- ± to - ~ I^ KC M. temf)oralis 1 + + + - - M. temporalis 2 I = C I>C M. temporalis 3 + + + - - I-C I>C (M. pterygoideus lateralis) + + + - + + TIME — [m- s.msl :| , M |l II ^ ^ ff—l I. PTMR Hjtfj — r' 1^ him .Jill T T' ni' rf)li 'In ! ■■ 1..1 i Tir~" 4|T(r i . 1,1 i 1 hil.. 111 iLn J ... .. Illil !. . . .iJl — ITf nri n 1* \ lL dll .1 djli. 1' f|Ti rr ^ ^ |ff|r rpp lyi , 1 'iWiWi- iliii Ml V -'ll ' ri ™ r w rtri' Fig. 1.— Crocidura flavescens. Original electromyograms from the middle of a chewing series, the animal was feeding on pieces of chicken heart. Reversal of ipsilateral side from right to left. M.MASR, right M. masseter; M.PTMR and PTML, right and left M. pterygoideus medialis; M.DIGL, left M. digastricus; M.TE2R and TE2L, right and left M. temporalis 2; M.TE3R and TE3L, right and left M. temporalis 3. 1994 DOTSCH — Function of Feeding Apparatus in Shrews 239 Fig. 2.— a. Positions of implanted parapulpar pins in the skull of C flavescens. 1 , calculated midpoint between the centers of right and left jaw articulations; 2, screw-head point of the pin on the midline of the skull; 5/7, left and right screw-head points of the pins in the rostrum; 9/15, left and right markers on the tip of the coronoid process; 11/17, anterior, and 13/19, posterior left and right lower jaws. b. Movement orbits of marker points in frontal view of Blarina brevicauda chewing a mealworm. Ipsilateral side is to the left. Left (9, 10) and right (15, 16) points in the coronoid process; left (11, 13) and right (17, 19) points in the mandibles; 21 and 22 are estimated markers in the centers of the articulations, c. Calculations of distance changes between points on the coronoid processes (9/15) and the midline of the skull (line 1/2), frontal view. d. Distance changes between point 2 and left and right mandible markers (11/7), frontal view. Movement orbits from cycle 2 are shown in b. , '., .. ifl,; ••: i/:mjf\'ifji^ hSi'iiU i i ■ ,i A, ' % ■ ^ B ■ U ' ^ j“ ■■-■•' ■ -' ■ - . :/ ,«„ ; ■ ‘IK ■‘.'(■.Utllf -•-'^T!v'*!.1nlf»Wi -iJK - v/fi*mct^--,. D'' 5 'V ■- ■■ '.Vi . ,vr ^ I.'- ; 1- \ ’ ' * • -f - ‘ ’. wi ■> ■ ■:• .ft- i'j ’A' ..Jitt.jaoi &rl) •■' <1 M* '^’ k'vf ^ '.it'Vf .»V>) ' :j!i ft) 5* twJ(:i /i> : '• ' "■■•••'' cdtilioq' XjSS',^»i^'- oi.- -Ttliir ' ■•- ''-ji Inti " /" - t*\ j'T \ < '■ ‘''iMwilt 1 ^ . COMPARATIVE EMBRYONIC DEVELOPMENT OF THE SORICIDAE Kerry R. Foresman Division of Biological Sciences, University of Montana, Missoula, MT 59812 Abstract Embryonic and fetal development was studied in Sorex vagrans and S. cinereus using embryos from wild-eaught, pregnant females. A total of 103 embryos and fetuses were classified to developmental stage based on external morphologieal features and crown-rump measurements; and developmental changes in skeletal, neural, respiratory, and digestive systems eharaeterized for each stage. These data were compared to the developmental pattern of the mouse Mus musculus. Differenees were observed in timing of neural tube closure, and skeletal and respiratory developmental events. All embryological changes occurred at a later developmental stage in Sorex than in Mus. Results substantiate allometric or evolutionary relationships between taxa. Introduction Members of the mammalian Order Insectivora are considered to be the most primitive living eutherians. Their primitive status is based on ancestral dental and skull characteristics retained by extant species. These mammals are insectivorous, as were early Cretaceous forms. Within the Order Insectivora, members of the Family Soricidae (Subfamily Soricinae) are considered the most primitive (Repenning, 1967; Vogel, 1980, 1981). Developmental studies of postnatal growth in one species, S. araneus, indicate that young are bom in an immature, nidicolous state, more similar to the developmental sequence of metatherians than to that of other eutherians (Vogel, 1972, 1980, 1981). This developmental evidence is interpreted as further support for the hypothesis that insectivores are evolutionarily very primitive mammals. The similarity in fetal development at birth between the Soricinae and marsupials represents a plesiomorphic, or primitive, relationship rather than an apomorphic, or derived, relationship (Lillegraven, 1976; Marshall, 1979). Members of the Subfamily Soricinae, therefore, may be the most primitive extant eutherians with which to study ontogenetic correlates to metatherians. Although a comparative atlas on the staging of mammalian embryos has recently been published (Butler and Juurlink, 1987), no insectivores were included among the species studied. Several studies have addressed prenatal development in insectivores (Sterba, 1977, 1985), yet to date, no comprehensive atlas of embryonic development has been produced on any member of the Genus Sorex. Initial studies are presented here for two members of this genus, S. vagrans and S. cinereus. Materials and Methods Shrews were collected in pitfall traps between April 1985 and August 1990 in western Montana. Reproductive tracts from all females were removed and immediately fixed in 10% calcium chloride-buffered formalin. Those containing obvious embryos were handled in one of two ways. The smallest swellings were separated and processed for histological analysis in their entirety, whereas larger swellings (> 3 X 5 mm) were opened to remove the fetuses prior to processing. Each larger embryo and fetus was photographed and classified by developmental stage based on external morphological features and crown-rump measurements, employing Theiler’s (1989) criteria for Mus musculus, Streeter’s (1942, 1945, 1948, 1949, 1951) and O’Rahilly’s (1973) for the human embryo, and Sterba’s work (1977, 1985) for insectivores. A total of 103 embryos and fetuses (74 S. vagrans and 29 S. cinereus) were classified. From these, two or three individuals representing each ontogenetic stage recognized by Theiler (1989) were dehydrated, embedded in paraffin, serially sectioned at either 3 or 8 fim, and stained with hematoxylin and eosin for histological examination. Developmental stages of skeletal, neural, respiratory, and digestive systems were characterized for each individual. These embryological data were compared with the ontogenetic time frame observed by Sterba (1977) for Sorex araneus and the developmental pattern exemplified by M. musculus as outlined by Rugh (1968) and Theiler (1989). Results The prenatal growth rates of S. vagrans and S. cinereus, expressed as crown-rump length as a function of developmental stage, closely parallel one another (Fig. 1). Ontogenetic changes described for each developmental stage depicted in Fig. 1, with the exception of Theiler Stage 12, are outlined below. The ontogenetic stage and approximate day of development follow Sterba (1977) for insectivores, and thus is not consistent with the developmental time frame of Theiler stages for Mus', Theiler stages were used solely as an index of morphological criteria for development. Certain ontogenetic stages observed fell between those described by Sterba and are thus approximated by indicating a “-I-” value (e.g.. Ontogenetic Stage 2 + ). Ontogenetic Stage 2 (Theiler Stage 13): [Approximately Day 10 post-coitus (p.c.); S. vagrans/ External features. — [Crown rump (CR) measurement of one embryo - 1.3 mm]. Neural folds are forming along the length of the embryo leaving the anterior and posterior regions open as an exposed neuropore. Somites are readily visible (Fig. 2a). Internal anatomy.— The forebrain anlage is thickening and flexing, bending the anterior portion of the embryo. Somites are readily apparent (a total of nine in the embryo sectioned), and 241 242 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 initial heart formation is occurring (Fig. 2a, b). The gut is developing, with an expansion in the foregut region (Fig. 2b). Ontogenetic Stage 2+ (Theiler Stage 14): [Approximately Day 11 p.c.; S. cinereusy External features. — [Mean crown-rump measurement (XCR) — 1.6 mm, range 1.4- 1.8 mm]. Two branchial bars are evident, and the forming heart bulges from the ventral surface. The forelimb bud is barely discernible, and the tail curves to the right (Fig. 2c). The anterior neuropore has closed, whereas the posterior neuropore remains open. Internal anatomy. — Nephric ducts and vesicles are forming in conjunction with the thoracic somites. In the hindgut region the aorta is paired. Ontogenetic Stage 3 (Theiler Stage 16): [Approximately Day 12 p.c.; S. vagransy External features.— [XCR = 3.0 mm, range = 2.9-3. 1 mm]. The posterior neuropore is still open. Front limb buds are prominent, and hind limb buds are evident as distinct bulges. The lens plate is indented to form a pit, whereas the otic vesicle has completely closed and is separated from the epidermis. Branchial bars 1 and 2 are enlarged and convex, whereas numbers 3 and 4 are concave and only apparent from the presence of the cervical sinus. The tail is beginning to elongate and terminates in an enlarged stump (Fig. 3a, b). Internal anatomy. — Bronchial buds are forming as branches from the laryngotracheal tube (Fig. 3c), and the thyroid diverticulum is prominent. Rathke’s pouch appears as a distinct evagination from the roof of the oral cavity. The digestive tract has formed as a narrow tube running the length of the embryo with a slight expansion forming as the stomach anlage. Mesonephric ducts and tubules are in evidence along the latter third somites of the embryo. Ontogenetic Stage 3 + (Theiler Stage 1 7): [Approximately Day 13 p.c.; S. vagrans and S. cinereusy External features. — [XCR; S. vagrans = 3.6 mm, range 3. 0-4. 3 mm; S. cinereus = 2.8 mm, range 2. 7-3.0 nun]. The eye lens vesicle is conspicuous, forming a deep pocket. Both forelimb and hindlimb buds are enlarging, and the tail is elongating (Fig. 3d). Internal anatomy. — The surface epithelial layer is invaginating to form a lens pocket (Fig. 3e). The liver anlage is forming blood sinuses, and there are a large number of mesonephric tubules. The heart has formed single atrial and ventricular chambers, but these are not subdivided (Fig. 3f). Ontogenetic Stage 4 (Theiler Stage 19): [Approximately Day 14 p.c.; S. vagransy External features.— [XCR = 4.4 mm, range 4.2-4. 8 mm]. The most characteristic feature of this developmental stage is the formation of a footplate on the forelimb. A distinct constriction marks the presumptive wrist. The hindlimb, slower to form, is still represented by an expanded bud. Three brain regions, telencephalon, mesencephalon, and rhombencephalon , are clearly visible, and the nostrils have formed adjacent to a nasolacrimal groove. Somites are distinct, particularly in the tail region which has elongated (Fig. 4a). Internal anatomy. — The tongue has not yet elevated from the floor of the mouth. Major bronchi of the lungs are forming lined with pseudostratified columnar epithelium (Fig. 4c). The liver is a diffuse organ consisting of large sinusoids filled with nucleated erythrocytes (Fig. 4f). The stomach is enlarging, lined internally with pseudostratified, columnar epithelium and externally with a single layer of cuboidal cells. The matrix of the stomach is densely packed with undifferentiated cells (Fig. 4f). Mesonephric development is marked by the appearance of glomerular bodies (Fig. 4b). The eye lens has detached from the surface ectoderm and has completed closure (Fig. 4e). All vertebrae are in the precartilage coalescence stage, and the ribs are begirming to chondrify (Fig. 4d, Table 1). Ontogenetic Stage 4+ (Theiler Stage 20): [Approximately Day 14 p.c.; S. vagransy External features. — [XCR = 6.4 mm, range 6. 1-6.6 mm]. The hindfoot plate has formed on an elongating limb being displaced from the body, and the forefoot plate has begun to show signs of digitation. Eye pigmentation is distinct, and the otic vesicle is deeply invaginated (Fig. 5a). Internal anatomy. — Semilunar ganglia are very prominent. The eye lens is forming, but there is only slight development of lens fibers (Fig. 5b). Mitoses are apparent in cells of the lens. The major bronchus is branching and rapidly elongating. There are numerous mesonephric tubules, and the metanephric tissue has begun to develop. Primary renal calyces are present, and metanephric ducts are forming (Fig. 5c). The gonad is at its earliest stage of differentiation; in the animals examined (all males) an outline of the seminiferous tubules is apparent. Vertebral tissue is condensing above and around the notochord. Ontogenetic Stage 5 (Theiler Stage 21): [Approximately Day 15 p.c.; S. vagrans/ External features. — [XCR = 6.6 mm, range 6. 3-7.0 mm]. Marked changes have occurred with further development of the eye, and conspicuous formation of the ear pinna. The forelimb has constricted further at the wrist, and hand rays are forming with distinct indentations demarcating the presumptive digits. The hindlimb has also further differentiated with a constricture at the ankle and a hint of digit formation. Somites are clearly visible in the tail region, and hair tracts, which are progenitors of whiskers, are noticeable on the snout (Fig. 6a, b). Internal anatomy.— A. small number of bronchioles are branching off of the bronchi , lined with pseudostratified, columnar epithelium (Fig. 6c). The stomach has continued expansion and is thickly lined with pseudostratified cells. Circular muscle layers are beginning to differentiate, but there is no glandular development. All valves are present in the heart, and the pituitary gland is differentiating (Fig. 6f). Gonadal development is proceeding rapidly with the appearance of discrete seminiferous tubules (Fig. 6d, e). Skeletal development 1994 FORESMAN — Embryonic Development in Sorex 243 has proceeded to varying stages of chondrification, with caudal vertebrae lagging behind (Fig. 6g, h, i; Table 1). Ontogenetic Stage 6 (Theiler Stage 22): [Approximately Day 17 p.c.; S. vagrans and S. cinereusy External features. — [XCR: S. vagrans = 8.4 mm, range 7. 5-8. 7 mm; S. cinereus — 7.6 mm, range 7.4-7. 8 mm]. Two features characterize this developmental stage. A prominent umbilical hernia appears, and the digitation of the foot plates has continued. The ear piima has expanded, and hair follicles cover most of the body. The tail still curves to the right (Fig. 7a, b, c). Internal anatomy. — Marked development has occurred in all internal organ systems by this stage. The tongue has risen up, and been undercut by the lower jaw, with concurrent formation of the epiglottis; the secondary palate is still open. The lung has become distinctly lobed, and extensive branching of the bronchioles has occurred, although the lung tissue remains compacted (Fig. 71). The bronchioles are lined by simple columnar epithelium. The esophagus, lined with a stratified layer of cuboidal cells with round nuclei, opens into an expanded stomach lined with tall, pseudostrati fied, columnar cells with very elongated nuclei. Two or three clearly distinct layers of circular smooth muscle cells are differentiating in the stomach wall, although no longitudinal layers are evident (Fig. 7o). The lining of the intestinal tract has an appearance similar to that of the stomach, though it is thrown into folds. A portion of the small intestine lies externally as a component of the umbilical hernia. The choroid plexus is well-developed, and projects deeply into the lateral ventricle (Fig. 7k). Further pituitary development and folding have reduced the size of Rathke’s pouch (Fig. 7h). Rapid development of the cervical (Fig. 7i) and cranial (Fig. 7j) nerves is evident, and anlagen of both the upper and lower incisors are visible. Gonadal development has proceeded with further rounding in the walls of the seminiferous tubules and the presence of large, prominent gonocytes within the lumina (Fig. 7m, n). Ovarian development has proceeded as well, although follicular development is not evident. Skeletal ossification has begun medially in the upper limbs, and in the cervical vertebrae, whereas other skeletal tissues are still in precartilaginous stages or are begirming to progress through stages of chondrification (Fig. 7d, e, f, g; Table 1). Ontogenetic Stage 7 (Theiler Stage 25); [Approximately Day 19 p.c.; S. vagrans/ External features. — [XCR = 12.0 mm, range 11.5-12.4 mm]. The eyelids are fused. The umbilical hernia has disappeared, and the skin is noticeably wrinkled. Hindlimbs are tightly curled into the body (Fig. 8a, b, c, h). Internal anatomy. — Palatine processes have fused resulting in formation of the secondary palate. Lung tissue has become less compacted as the terminal bronchioles begin opening into terminal saccules through transitory ducts (Fig. 8i, j, k). The upper and lower incisors have differentiated into precalcified enamel and dentin layers — the bell stage of development (Fig. 8f). Formation of the eye has progressed significantly; cells of the lens are rapidly laying down fibers, the ciliary body has formed, and the eyelids have closed (Fig. 81). The thymus has greatly enlarged with the proliferation of small, darkly stained lymphocytes in the cortical region. The medullary region is still reduced and contains lightly-staining epithelial reticular cells (Fig. 8g). By this stage all vertebral types (though not every vertebra) are undergoing ossification as are the proximal regions of long bones of upper and lower limbs, and bones of the skull (Table 1). Skeletal elements of the fore- and hindfeet are undergoing chondrification (Fig. 8d, e; Table 1). Ontogenetic Stage 7+ (Theiler Stage 26): [Approximately Day 20 p.c.; S. vagrans/ (This stage marks the final growth sequence before birth.) External features. — [XCR = 12 mm, range 1 1.4-12.6 mm]. The most diagnostic feature at this stage is the withdrawal of limbs away from the body. As in Stage 25 the eyes are essentially invisible with the closure of the eyelids (Fig. 9a, b). Internal anatomy. — Overall development is similar to that described for the previous stage. The gastrointestinal tract has differentiated providing many of the cellular characteristics seen in the adult. The esophagus is lined with stratified, squamous epithelial cells which give way to columnar cells in the stomach. Six to eight layers of circular muscle invest, and support, the stomach, and the first longitudinal folds (rugae) are apparent (Fig. 9j). Gastric pits are evident as are sporadic Goblet cells, although the submucosal glandular tissue has not yet developed. A proliferation of thin-walled villi lined with columnar epithelial cells, and surrounded by two or three layers of circular smooth muscle, predominate in the small intestine (Fig. 9k). However, no lacteals are visible at this stage. The lungs are becoming less compacted, but alveolar development is not yet evident. Branching of bronchioles ends in terminal saccules (Fig. 9n, o). There is continued development in all of the endocrine glands. The medulla of the thymus has enlarged, and a few of the thyroid follicles appear to be invested with colloid. Pancreatic islands are well-differentiated, and pituitary celts of the Pars distalis (adenohypophysis) are enlarging slightly, and becoming more rounded (Fig. 9f). Rathke’s pouch has been further reduced. Seminiferous tubules of the testis are encapsulated with a single layer of fibrocytes, and the basement membrane is forming. Large numbers of spermatogonia are evident within the tubules, and there is much interstitial tissue, although the interstitial cells are not easily distinguishable from support tissues (Fig. 91, m). A thick pad of multilocular fat tissue has formed in the intrascapular region. The eye appears similar to that of the previous stage; the iris rudiment is present, but has not differentiated further (Fig. 9c). Ear ossicles are well-formed, and vestibular sensory hairs are visible (Fig. 9i). All vertebral types are ossifying further as are the ribs, appendicular long bones, and bones of the skull. Chondrification is continuing in the fore- and hindfeet. (Fig. 9d, e, g, h; Table 1). 244 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 Discussion Both S. vagrans and S. cinereus are bom in an altricial condition similar to that described for S. araneus (Vogel, 1972). Although the timing of developmental changes and rate of growth in S. cinereus parallel that observed in S. vagrans (Fig. 1), overall fetal size attained at comparable stages is slightly smaller. This difference is most likely due to the smaller adult body size of S. cinereus. In wild populations adult S. cinereus weigh on average 39% less than adult S. vagrans (4.6 g vs. 7.6 g). The general ontogenetic pattern appears similar to that previously described for the laboratory mouse, M. musculus (Rugh, 1968; Theiler, 1989). When Sterba’s (1977, 1985) results on S. araneus are taken into consideration, the time at which comparable developmental stages are reached in these soricids is estimated to be later. However, if the degree of internal organogenesis is correlated with the stage of development, as exemplified by external body form characteristics, there is close correlation between the soricids and M. musculus with a few notable exceptions. Staging of soricid embryos in the present study was based on two criteria: first, comparison of external characteristics with those described for Mus (Theiler, 1989), and second, consideration of actual age attained as proposed by Sterba (1977). Although dates of conception were not known, the completeness of developmental stages available, and their close correlation with Theiler’s and Sterba’s described stages, suggest that this method can be used to accurately sequence soricid developmental stages. Use of developmental characteristics to classify embryos rather than relying solely on CR measurements is of critical importance. Embryos of the same age, particularly at the later stages of development, may vary several millimeters in length even within a litter. Also, it is recognized that at comparable stages of development embryo length may vary greatly between species (ten Donkelaar et al., 1979; Butler and Juurlink, 1987). Some variation in the age assigned to various stages in the present study may have occurred due to variability in developmental rates, particularly between littermates (ten Donkelaar et al., 1979; Juurlink and Fedoroff, 1980), but this is considered to be negligible. Given these assumptions, several developmental differences are apparent in soricids. Vogel (1972) described in detail the reduced skull ossification at birth in S. araneus and Neomys fodiens, a condition paralleling that observed in neonate marsupials. A similar situation is found in the present study. In comparison to Mus (Rugh, 1968), the timing of chondrification and/or ossification of a few cranial bones, vertebrae, long bones of the hind limb, and feet, occurs later in the soricids (Table 1). Other developmental events in Sorex occur at a later embryonic age as well. The posterior neuropore which closes in Mus at Theiler Stage 16 (ten days p.c.), remains open past Stage 17 in soricids, which approximates Day 13 p.c. In all cases closure is complete by Stage 19, or approximately Day 14 p.c. in soricids. Eye development also appears to lag behind that observed in Mus. Lens fiber formation and growth of the iris are retarded. A significant difference occurs in the development of the lungs. In soricids, alveolar growth does not occur by birth. The final stage of development in near-term fetuses consists of terminal saccules (Fig. 8j, 9o). Lung tissue remains more compacted and possesses fewer terminal saccules per area than in Mus. Although Theiler (1989) described an “alveolar explosion” in Mus at this stage (Stage 26, just prior to birth), more detailed descriptions of lung development in the mouse (Engel, 1953) and rat (Burri, 1974, 1985) suggest that true alveolar formation has not yet occurred. However, lung tissue in these species does become markedly less dense, and terminal saccules become numerous as the progenitors to alveoli (Burri, 1974). Sterba (1984) argued that the altricial ontogenetic pattern, exemplified by the genus Sorex, is the primary evolutionary pattern from which the precocial pattern, as exemplified by Mus, developed. This suggests that the difference in developmental rates between Mus and Sorex represents acceleration in development in Mus, rather than retardation in Sorex. This argument is convincing considering that in altricial species of the genus Sorex, considerable developmental maturation occurs during the postnatal period. A comparison of ontogenetic time patterns in other insectivores (e.g., Crocidura and Talpa: Sterba, 1977; Suncus: Vogel, 1972) further supports this thesis. Taken collectively these studies thus suggest that development in Sorex species falls between the Crocidurinae and the Muridae. The very altricial condition of newborn Sorex has also been suggested as supporting this group’s phylogenetic relationship to marsupials (Vogel, 1981). Developmental similarities between Sorex and Neomys, and a generalized marsupial, Didelphis, have been demonstrated (Vogel, 1972, 1981). Those ^ studies and the present one illustrate differences in many prenatal development stages relative to typical nidicolous eutherians such as Mus. However, the influence of allometric scaling on developmental processes should not be overlooked. Sterba (1984) proposed a thesis which addressed the limiting factors of adult body size and physiological longevity. He attempted to explain the occurrence of nidicolous and nidifugous (precocial) ontogenetic patterns by ecological selective pressures which influence the development of r- and K-strategies of reproduction. These selective pressures, he suggested, determine the ontogenetic pattern, which then secondarily becomes evolutionarily “fixed.” Much broader comparative embryological studies on additional soricids and similarly-sized advanced eutherians should further clarify this relationship. Acknowledgments This work was supported by grants from the University of Montana, the Division of Biological Sciences, and the Montanans on a New Trac for Science program. Many individuals helped in the collection and processing of animals, among them R. D. Long, K. McCracken, R. Jensen, D. Worthington, and S. Gillihan. M. Henry helped in the initiation of this study, supported by a University of Montana Watkin’s scholarship, and her efforts are warmly appreciated. R. Petty, D. Williams, and N. Baker produced the finished photographs and their efforts are recognized. Special gratitude is extended 1994 FORESMAN — Embryonic Development in Sorex 245 to R. D. Long who expended hundreds of uncompensated hours huddled over the microtome to produce superior tissue sections. Literature Cited Burri, P. H. 1974. The postnatal growth of the rat lung. II. Morphology. Anatomical Record, 180:77-98. 1985. Development and growth of the human lung. Pp. 1-46, in Handbook of Physiology, Section 3: The Respiratory System, Volume I (A. P. Fishman and A. B. Fisher, eds.), American Physiological Society, Maryland, 397 pp. Butler, H., and B. H. J. Juurlink. 1987. An Atlas for Staging Mammalian and Chick Embryos. CRC Press, Inc., 218 pp. Engel, S. 1953. The structure of the respiratory tissue in the newborn. Acta Anatomica, 19:353-365. Juurlink, B. H. J., and S. Fedoroff. 1980. The development of mouse spinal cord in tissue culture. I. Cultures of whole mouse embryos and spinal-cord primordia. In Vitro, 15:86-94. LILLEGRAVEN, J. A. 1976. Biological considerations of the marsupial- placental dichotomy. Evolution, 29:707-722. Marshall, L. G. 1979. Evolution of metatherian and eutherian (mammalian) characters: A review based on cladistic methodology. Zoological Journal of the Linnean Society, 66:369-410. O’Rahilly, R. 1973. Developmental stages in human embryos, including a survey of the Carnegie collections. Part A. Embryos of the first three weeks (Stages 1 to 9). Carnegie Institute of Washington, Washington, D.C. Repenning, C. A. 1967. Subfamilies and genera of the Soricidae. Geological Survey Professional Papers, 565:1-74. RUGH, R. 1968. The Mouse: Its Reproduction and Development. Burgess Publishing Co., Minneapolis, Minnesota, 430 pp. Sterba, O. 1977. Prenatal development of central European insectivores. Folia Zoologica, 26:27-44. 1984. Ontogenetic patterns and reproductive strategies in mammals. Folia Zoologica, 33:65-72. 1985. Ontogenetic levels in mammals. Pp. 567-571, in Evolution and Morphogenesis (J. Mlikovsky and V. J. A. Novak, eds.). Academia, Prague. Streeter, G. L. 1942. Developmental horizons in human embryos; description of age group XI, 13 to 20 somites, and age group XII, 21 to 29 somites. Contributions to Embryology, Carnegie Institution, 30:211-245. 1945. Developmental horizons in human embryos; description of age group XIII, embryos about 4 or 5 millimeters long, and age group XIV, period of indentation of the lens vesicle. Contributions to Embryology, Carnegie Institution, 31:27-63. 1948. Developmental horizons in human embryos; description of age groups XV, XVI, XVII, and XVIII, being the third issue of a survey of the Carnegie collection. Contributions to Embryology, Carnegie Institution, 32:133-203. 1949. Developmental horizons in human embryos (fourth issue) a review of the histogenesis of cartilage and bone. Contributions to Embryology, Carnegie Institution, 33:149-168. 1951. Developmental horizons in human embryos; description of age groups XIX, XX, XXI, XXII, XXIII, being the fifth issue of a survey of the Carnegie eollection. Contributions to Embryology, Carnegie Institution, 34:165-196. TEN Donkelaar, H. J., L. G. M. Geysberts, and P. j. W. Dederen. 1979. Stages in the prenatal development of the Chinese hamster (Cricetulus griseus). Anatomy and Embryology, 156: 1-28. Theiler, K. 1989. The House Mouse. Springer-Verlag, New York, 178 pp. Vogel, P. 1972. Vergleichende Untersuchung zum Ontogenesemodus einheimischer Soriciden (Crocidura russula, Sorex araneus, and Neomys fodiens). Revue Suisse de Zoologie, 79:1201-1332. 1980. Metabolic levels and biological strategies in shrews. Pp. 170-180, in Comparative Physiology: Primitive mammals (K. Schmidt-Nielsen, L. Bolis, and C. R. Taylor, eds.), Cambridge University Press, 338 pp. 1981. Occurrence and interpretation of delayed implantation in insectivores. Journal of Reproduction and Fertility, supplement, 29:51-60. Table 1. — Chondrification and/or ossification of tissues. “ Ontogenetic Stage (OS) and approximate developmental day (D) after Sterba (1977). Skull Bones: Basioccipital, basisphenoid, exoccipital, tympanic (endochondral bones which pass through a chondrification stage); maxillary, palatine, premaxillary , pterygoid (dermal bones which ossify directly from connective tissue) OS5(~D15)‘* chondrified — early /middle condensation stages OS6(~D17) - chondrified — middle/late condensation stages OS7(~D19) ossification occurring OS7-b(~D18) continued ossification Vertebrae: Cervical OS4(~D14) precartilage coalescence OSS chondrified — late condensation stage OS6 ossification— early stage OS7 continued ossification OS7 + continued ossification Thoracic OS4 precartilage coalescence OSS chondrified — early condensation stage OS6 chondrified — early condensation stage 246 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 OS7 OS7 + ossification occurring continued ossification Lumbar OS4 055 056 057 OS7 + precartilage coalescence chondrified— middle condensation stage chondrified— middle condensation stage ossification occurring continued ossification Sacral OS4 055 056 057 OS7 + precartilage coalescence chondrified — middle/late stages chondrified— late stage/early ossification continued ossification continued ossification Caudal OS4 055 056 057 OS7 + precartilage coalescence precartilage coalescence precartilage/early chondrification ossified — early stages continued ossification Ribs OS4 055 056 057 OS7 + chondrified — early stage chondrified — early stage chondrified — early stage ossification occurring continued ossification Appendicular Skeleton: Humerus OSS 056 057 OS7 + chondrified ossification occurring continued ossification continued ossification Ulna/Radius OSS 056 057 OS7 + chondrified ossification occurring continued ossification continued ossification Carpals, Metacarpals, Phalanges 056 057 OS7 + prechondral/earliest chondrification stages chondrification— early stages chondrification continuing Femur OSS 056 057 OS7 + chondrification— early stage chondrification — middle stage ossification occurring continued ossification Tibia/Fibula OSS 056 057 OS7 + chondrification — early stage chondrification— middle stage continued ossification continued ossification Tarsals, Metatarsals, Phalanges 056 057 OS7 + prechondral/earliest chondrification stages chondrification— middle stage; (epiphyses in prechondral stage) chondrification continuing 1994 FORESMAN — Embryonic Development in Sorex 247 Embryonic Development in Sorex Theiler Stage of Development Fig. 1. — Crown-rump measurement as a function of Theiler developmental stage in Sorex vagrans and Sorex cinereus. Fig. 2. — a, b, Ontogenetic Stage 2 in nine-somite embryo of Sorex vagrans. So, somite; Nf, neural folds; Fg, foregut. c. Ontogenetic Stage 2+ in embryo of Sorex cinereus, left side view. Bh, branchial bars; Rt, tail curved to right. 248 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 Fig. 3.— a-c, Ontogenetic Stage 3 in Sorex vagrans. a, b, Embryo, right side views. L, lens pit; Ms, mesencephalon; 1, first branchial bar, 2, second branchial bar; O, otic vesicle; T, tail; He, heart; FI, front limb bud; HI, hind limb bud; Rt, tail curvature to the right, c. Lung. B, bronchial bud. d, Ontogenetic Stage 3+ in Sorex vagrans, embryo, right side. L, lens pit; FI, front limb bud; HI, hind limb bud. e. Eye showing retinal layer (RL). f, heart. Ve, unpaired ventricle; A, incompletely partitioned atrial chambers. 1994 FORESMAN — Embryonic Development in Sorex 249 Fig. 4.— Ontogenetic Stage 4 in Sorex vagrans. a, Embryo, right side. Tel, telencephalon; Ms, mesencephalon; Rh, rhombencephalon; 1, first branchial bar; 2, second branchial bar; L, lens (L), Ng, nasolacrimal groove; N, nostrils; Fp, forelimb bud forming handplate; HI, hindlimb bud undifferentiated; So, tail somites, b. Me, mesonephric development; Li, liver, c. Lung. B, bronchi; V, thoracic vertebrae; E, esophagus, d, V, Thoracic vertebrae, e, eye. PL, pigment layer; IN, irmer neuroblastic layer; ON, outer neuroblastic layer, f, St, stomach; Li, liver. 250 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 I li :.i Fig. 5. — Ontogenetic Stage 4+ in Sorex vagrans. a, Embryo, left side. Ep, eye pigmented; Hf, hindfoot plate forming on elongating limb (arrow); O, otic vesicle, b, Eye. L, lens; Op, optic nerve, c. Kidney. Mt, metanephric ducts; C, renal calyx. 1994 FORESMAN — Embryonic Development in Sorex 251 Fig. 6.— Ontogenetic Stage 5 in Sorex vagrans. a, b, Embryo, right and frontal views. Ep, eye pigmented; Pi, ear pinna; W, whiskers; Fp, foot pads; Fr, forelimb rays; Tr, toe rays; I, footplate indentations; So, somites; N, nostrils, c. Lung. B, bronchi, d, K, kidney with nephric ducts developing; T, testis, e. Enlargement of testis section shown in d. St, developing seminiferous tubules, f. Pituitary gland. Pt, pars tuberalis; Ad, adenohypophysis, g, V, thoracic vertebrae; Sp, spinal cord, h. Cervical vertebrae, C4 identified, i. Enlargement of C4. Ch, chondrification center; No, notochord. 252 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 Fig. 7. — Ontogenetic Stage 6 in Sorex vagrans. a, b, c. Embryos, frontal view flanked by two views of right side. Pi, ear pinna; L, lens; W, whiskers; Fs, forefoot plate indented; I, hindfoot plate indenting; Vh, umbilical hernia; Ep, eye pigmentation; N, nostrils; Ts, toes separating at slightly later stage 22; H, hair follicles prominent, d. Cervical vertebrae, C3 identified. OC, ossification centers; Sp, spinal cord (Sp). e, V, thoracic vertebrae; Sp, spinal cord; Tr, trachea, f. Enlargement of ossification center (OC) of thoracic vertebrae, g, Hindlimb showing tibia (Ti), and fibula (Fi), undergoing chondrification. h. Pituitary folding upon itself. Pt, pars tuberalis; Ad, adenohypophysis; Co, cochlea of ear ossicles, i. Spinal nerves (Ne) developing in 1994 FORESMAN— Embryonic Development in Sorex 253 association with the spinal cord (Sp). j, Frontal head region. Ce, cerebrum; N, nose with hair follicles and well-developed ganglion Gasseri (Gg) with nerve tracts radiating distally (arrows), k, Choroid plexus of brain (CP), directed anteriorly from the cerebellum (Cb). I, Lung tissue illustrating branching of bronchioles (B). m, Kidney (K) with continued duct development, and testis (T). n. Enlargement of testis section shown in m illustrating discrete seminiferous tubules (ST), and the presence of gonocytes (G). o. Cross section of stomach (St) with development of circular muscle layers (arrow). 254 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 Fig. 8.— Ontogenetic Stage 7 in Sorex vagrans. a, b, c. Embryo; right side, frontal, and left side views. Es, eyelid fused; P, fingers and toes parallel, d. Junction of skull and spinal column. Bo, basioccipital bone ossifying; OC, ossification center of atlas; At, anterior arch of atlas, e. Lumbar vertebrae. OC, ossification center; No, notochord; Sp, spinal cord, f. Tooth formation showing development of enamel (E), dentin (D), pulp (Pv), and ameloblast layers (Am), g. Thymus illustrating 1994 FORESMAN — Embryonic Development in Sorex 255 medullary (Md) and cortical (Cor) regions, h, Thoracic region showing folding of outer skin surface (Sk) and underlying ribs (R). i, Complete respiratory tract. Tr, trachea; Lg, lung lobes; B, primary bronchus; Sp, spinal cord, j, Enlargement of lung lobe illustrating branching of secondary bronchus (Br). k, Enlargement of section shown in i showing transitory ducts (TD), terminal saccules (TS), and pleural cavity (PC). 1, Sagittal section through eye. L, lens; CB, ciliary body; Es, fused eyelid. 256 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 Fig. 9.— Ontogenetic Stage 7 + in Sorex vagrans. a, b, Embryo, right side and frontal views. Es, eyelid fused, c, Sagittal section through eye. L, lens; Ir, iris; Nu, nuclear layer of retina; Ga, ganglion layer of retina; Es, fused eyelid (Es). d, V, cervical vertebrae; Ne, spinal nerve, e. Enlargement of cervical vertebrae shown in d illustrating progressive ossification, f, Pituitary gland continuing to fold. Pt, pars tuberalis; Ad, adenohypophysis; S, sphenoid bone, g, Proximal region of tibia (Ti) and fibula (Fi) illustrating ossification, h. Thoracic vertebrae. OC, ossification centers; DA, dorsal aorta; Sp, spinal cord, i. Vestibule of inner ear illustrating sensory hair cells (H). j. Cross section of stomach; development of folds (Rg, rugae) and increased 1994 FORESM AN— Embryonic Development in Sorex 257 layering of circular smooth muscles (Ci). k, Cross section of small intestine. Vi, villus; Ci, circular smooth muscles. 1, Longitudinal section of testis (T) and epididymis (Ep); seminiferous tubules are distinct (arrows), m, Enlargement of seminiferous tubules shown in section /. TA, tunica albuginea; G, gonocytes. n, Lung lobe illustrating slightly more porous tissue. R, rib. o, Enlargement of lung section shown in n. TB, terminal bronchiole opening into transitory ducts (*); transition into smooth-walled channels of transitory ducts (arrows). BROWN FAT AND THE WINTERING OF SHREWS Heikki Hyvarinen Department of Biology, University of Joensuu, Box 111, SF-80101 Joensuu, Finland Abstract The relationship between mass of the interscapular brown fat and body mass was studied in six species of shrews and four species of rodents in Finland. Both anatomical location and changes in amount of brown fat were studied in Sorex araneus, Sorex minutus, and Neomys fodiens. The fatty acid composition of brown fat lipids was also measured in Sorex araneus and Neomys fodiens. The distribution of brown fat in shrews was very extensive, and nearly all the fat tissue was brown fat. For animals caught during the same season, logarithmic values for relative weights of interscapular brown fat of different species were strongly and negatively correlated with body weight. In Neomys fodiens the proportion of brown fat was higher than expected on the basis of body mass. In the smallest shrew species the mass of brown adipose tissue can be as much as 20% of the body mass. The proportion by weight of interscapular brown fat to total brown fat in spring was estimated to be about 30%. The fatty acid composition of brown fat lipids in shrews resembled the lipid composition in marine mammals. The very high proportion of 20- and 22-carbon fatty acids was assumed to be connected to the diet of shrews. Of special interest is the finding that brown fat is located between the muscles of the limbs. Warming of the blood coming from the extremities is assumed to be one explanation for the successful wintering of shrews. Introduction To survive extremely cold winters, shrews have developed means of adapting to low temperatures. Because shrews are very small and their insulation is poor (Hissa and Tarkkonen, 1969), their only effective means of thermoregulation in a cooling environment is to increase the metabolic rate. In Soricinae the metabolic rate per unit body mass is generally 2-3 times higher than that of a nonsoricine of the same size (Morrison et al., 1959; Vogel, 1976; Nagel, 1985). Furthermore, the metabolic potential of different tissues in shrews is very high compared to that in other small mammals (Hyvarinen and Pasanen, 1973). In shrews, elevation of a high basal metabolism is uneconomical but necessary. During the Finnish winter, shrews live continuously in temperatures below the thermoneutral zone. In these harsh conditions shivering thermogenesis is not a solution (see Wunder, 1984). Brown adipose tissue (BAT) is known to be the main factor responsible for heat production in small mammals (for reviews see Himms- Hagen, 1976; Rothwell and Stock, 1984). Previously studied seasonal variations in the amount or metabolism of BAT in shrews have been related only to interscapular brown fat (IBAT) (Buchalzyk and Korybska, 1964; Hissa and Tarkkonen, 1969; Pasanen and Hyvarinen, 1970; Pasanen, 1971), which is only part of the total BAT in small mammals. In Clethrionomys gapperi and Microtus pennsylvanicus, IBAT makes up only 16-25% of the total BAT (Anderson and Rauch, 1984). In this work, the distribution of BAT in the common shrew {Sorex araneus) was studied. To compare the role of BAT in shrews of different sizes, the weights of IBAT in several shrew species and in other small mammals were measured. The fatty acid composition of shrew BAT was compared with that of nonsoricine mammals. Materials and Methods Specimens were obtained by snap trapping or with pitfalls near Joensuu or in Rautjarvi, eastern Finland. Eleven S. araneus and four Sorex minutus were trapped in April-May 1984 and 13 5. araneus, three S. minutus, and one Neomys fodiens in J anuary- F ebruary 1985. Sixteen S. araneus, ten S. minutus, eight Sorex caecutiens, and two Sorex minutissimus were obtained in a pitfall catch in September 1989. In that same catch were also two Micromys minutus. Ten young S. araneus and six S. minutus were taken in July 1984 and July 1990; and two Neomys fodiens, two Sorex isodon, and three Myopus schisticolor were captured near Joensuu at the end of August 1989. Mass of IBAT was determined for all animals except the shrews captured in April-May 1984. An attempt was made to determine the weights of the other brown adipose tissues, but this was too inaccurate because those adipose tissues were intermixed with other tissues. For mapping the distribution of different types of BAT, animals captured in April-May 1984 were skinned, the internal organs, stomach and intestines removed, and the bodies were cut into 4-5 pieces from head to tail. The pieces were fixed in Bouin fixative, decalcified in 6% HNO3 (Romeis, 1948), and embedded in paraffin. The pieces were cut into 8 /xm transverse or sagittal sections and stained with Lillie’s (1951) allochrome method or with hematoxylin- eosin. Succinic dehydrogenase activity was demonstrated histochemically using freshly frozen cryostatmicrotomy slices (Burstone, 1962). Burstone’s method was used to determine the location of BAT. To measure fatty acid composition, the IBAT of four S. araneus and two N. fodiens captured in July was extracted with chloroform methanol (2:1). The lipids were hydrolyzed and the fatty acids methylated using the BF3 -method (Metcalfe et al., 1966). Samples were analyzed by gas chromatography on a Hewlett-Packard gas chromatograph (model 5890) using a fused silica OV-1 capillary column. Results The mass of IBAT of shrews captured in September in Rautjarvi was dependent on the size of the species, with the 259 260 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 highest relative weights of IBAT found in the smallest species (Table 1, Fig. 1). When correlations of the logarithmic values for the relative IBAT weights and the body weights of different shrew and microtine rodent species from the same time of year were calculated, the correlation coefficient was —0.94 (Fig. 1). Values for Clethrionomys glareolus and Microtus agrestis were taken from the work of Pasanen (1971). The proportion of the body mass made up by IBAT increased from autumn to winter in S. araneus (Fig. 2). There was only one N. fodiens specimen taken in winter, but for that animal the relative IBAT weight was 3%, compared to 1.3% for two young N. fodiens caught in late summer. In N. fodiens relative IBAT weight was also higher than in other small mammals of the same size (Fig. 1). The anatomical locations of BAT in S. araneus were (with one exception) typical for the other mammals studied (Fig. 3). Inter- and subscapular brown fat formed the major concentrations, but the amount of BAT around the kidneys, below the braincase, between the muscles of the neck, in the thorax area, in the iliac and inguinal areas, and in the limbs totaled about twice that in the interscapular area. Brown fat between the muscles of the limbs has not been observed in other mammals. Between the muscles of the thigh were especially large amounts of BAT (Fig. 4, 5) surrounding the large arteries, veins, and nerves like a jacket (see Smith, 1961). The proportion of total BAT to body weight was 4-10% in S. araneus and 6-12% in S. minutus. If the distribution of BAT is about the same in the smaller S. minutissimus, the proportion of the body mass made up by BAT would be 15-20%. According to histological structure, all adipose tissue of S. araneus and S. minutus is classified as BAT. Compared to other terrestrial mammals, the fatty acid composition of interscapular brown fat lipids in S. araneus and N. fodiens includes a very high proportion of 20- and 22-carbon fatty acids. Therefore, the amount of polyunsaturated fatty acids is very high and the number of double bonds per mole (A/mole) is about two in N. fodiens and about 1.6 in S. araneus (Table 2). Discussion The capacity for nonshivering thermogenesis has been shown to be inversely related to body mass, age, and acclimation temperature (Rothwell and Stock, 1984). Brown fat is mainly responsible for nonshivering thermogenesis (NST) in cold- adapted mammals. Rothwell and Stock (1984) estimated that the contribution of BAT to thermogenesis is 60% or more. Therefore, the negative correlation of relative IBAT mass and body mass is to be expected. It is possible, however, that in the smallest winter-active mammal of Europe, S. minutissimus, the proportion of BAT is so great (15-20% of body weight) that it may be one factor limiting the body size. Foster and Frydman (1978) demonstrated in the laboratory rat that BAT can produce heat at a rate equivalent to 500 W/kg. The aerobic power of muscle is about 60 W/kg during maximal exercise (Rothwell and Stock, 1984). Although shrews are capable of producing enough heat to live in cold environments, energy costs may become too high for survival. When non-BAT tissues of shrews reach a higher metabolic capacity than those of other small mammals (Hyvarinen and Pasanen, 1973), the amount of food that must be consumed in cold becomes too great. The animal carmot eat enough to provide the energy it is producing (see Hanski, 1984). The mass of BAT differs in members of the same species living at different latitudes (Pasanen, 1969) as well as at the same location at different times of the year. Furthermore, j winter conditions greatly influence BAT weights (Pasanen, 1971). For example, Anderson and Rauch (1984) recorded | higher relative IBAT mass in C. gapperi and M. pennsylvanicus f from Manitoba, Canada, than in voles from Finland (Hissa and ; Tarkkonen, 1969; Pasanen, 1971). The main reason for this | variation may be adaptation to the severity of the winter. Because geographic location and seasonal changes greatly l| influence relative IBAT weights, IBAT weights of different » species should be compared only for animals captured at the • same time of year in similar climatic conditions. In early autumn the IBAT weights are very similar between years (Pasanen, 1971). i The fatty acid composition of BAT lipids in shrews does not s resemble that of other terrestrial mammals, but is similar to that of marine mammals, wherein C20 ^nd C22 fatty acids make up more than 30% of the total fatty acids. In S. araneus this | proportion is over 20% and in the water shrew 30%. In the IBAT tissue of hamsters, for example, 95-99 % of the fatty acids belong to the C,g and Cjg series (Smalley, 1970). | Although the proportion of saturated fatty acids is high, the ,■ proportion of polyunsaturated fatty acids is also high. In ij| general, the brown fat of mammals contains more saturated ij fatty acids than the white fat (Smith and Horwitz, 1969). The j diet of shrews may affect the unusual fatty acid composition. j The proportion of the body mass of N. fodiens made up by 1 1 BAT is greater than expected on the basis of body mass. In one | | animal captured in February, IBAT made up 3% of the body 1 mass and the total BAT may have been about 10% of total || mass. Neomys fodiens is a diving mammal and it can be ; assumed that a high proportion of BAT is an adaptation for ||| diving in cold water. In the muskrat (Mac Arthur, 1986), also ! a diving mammal, total BAT is about 0.8% of body weight. Based on the body weight (about 1 kg), however, the proportion of BAT should be much smaller. Anderson and Rauch (1984) _ measured the weights of BAT from different parts of the body of red-backed and meadow voles. They found that the weight of IBAT is about 16-25% of total BAT. In S. araneus and S. minutus the proportion of IBAT was about 30%. This result, however, is an approximation made on the basis of histological |j| observations. ( The distribution of BAT between muscles and other tissues and the small size of shrews makes it impossible to accurately i estimate the weights of different types of BAT. In addition, BAT in shrews cannot be classified according to anatomical ,! location because BAT is found nearly everywhere. In fact, white adipose cells are uncommon and virtually all the adipose tissue in S. araneus and S. minutus is BAT. Especially unusual is the location of BAT in the limbs, surrounding blood vessels and nerves, where it warms blood coming from the thin uninsulated paws and legs. BAT is also abundant below the 1994 HYVARINEN— Brown Adipose Tissue in Shrews 261 braincase, and probably warms blood from the nose. This warming of blood returning from the extremities may facilitate successful overwintering of shrews. Because white adipose tissue is virtually absent in shrews, at least in S. araneus and S. minutus, I assume that brown adipose tissue also performs the functions of white fat. This assumption is supported by the finding of comparatively low GDP-binding rate of the brown fat of Sorex vagrans (Tomasi et al., 1987). Normally, GDP-binding of BAT is correlated with the capacity for nonshivering thermogenesis (see Horwitz, 1989). In Sorex species, BAT must also function as a site for lipid storage. Especially in overwintered adults and in young shrews at the time of weaning, the lipid content of BAT is very high compared with that of voles (Pasanen, 1971; Pasanen and Hyvarinen, 1970). In shrews, the metabolic capacity of BAT per unit mass is highest in autumn (Hyvarinen and Pasanen, 1973) when the lipid content and relative weight of IBAT are lowest (Pasanen, 1971). It is possible that during the winter fatty acids are liberated into the blood stream for use in other tissues. The activity of lipase-esterase enzyme is much higher in the BAT of S. araneus than in that of votes (Pasanen, 1971), although the cytochrome c content is the same or lower (Hyvarinen and Pasanen, 1973). In general, lipid metabolism seems to be more important in shrews than in voles during winter (Hyvarinen, 1984). In Blarina brevicauda, nonshivering thermogenesis increased 54% from August to January (Merritt, 1986), indicating the important role of BAT in the overwintering of that shrew species. Tomasi (1984) observed that the rate of thyroxine utilization is much higher in shrews than in rodents, and assumed that it plays a role in the high metabolism of shrews. According to histological results, the thyroid gland of S. araneus is very active during the autumn critical period but is inactive during overwintering (Hyvarinen, 1969). The only endocrine gland studied which is not inactive in winter is the adrenal medulla (Hyvarinen, 1984). During winter, all physiological functions of S. araneus are organized as economically as possible. BAT, adaptations of the sympathetic nervous system, and lipid metabolism in general provide the main physiological adaptations for shrews to the harsh winters of northern latitudes. Literature Cited Anderson M. J., and J. C. Rauch. 1984. Seasonal changes in white and brown adipose tissue in Clethrionomys gapperi (Red-backed vole) and in Microtus pennsylvanicus (Meadow vole). Comparative Biochemistry and Physiology, A. Comparative Physiology, 79:305-310. Buchalzyk, a., and Z. Korybska. 1964. Variation in the weight of the brown adipose tissue of Sorex araneus Linnaeus 1758. Acta Theriologica, 9:193-215. Burstone, M. S. 1962. Enzyme Histochemistry and Its Application in the Study of Neoplasms. Academic Press, New York. Foster, D. O., and M. L. Frydman. 1978. Nonshivering thermogenesis in the rat. II. Measurement of blood flow with microspheres point to brown adipose tissue as the dominant site of the calorigenesis induced by noradrenaline. Canadian Journal of Physiology and Pharmacology, 57:257-270. Hanski, I. 1984. Food consumption, assimilation and metabolic rate in six species of shrews from Finland (Sorex and Neomys). Annales Zoologici Fennici, 21:157-165. Himms-Hagen, j. 1976. Cellular thermogenesis. Annual Review of Physiology, 38:315-351. Hissa, R., and H. Tarkkonen. 1969. Seasonal variations in brown adipose tissue in two species of voles and the common shrew. Annales Zoologici Fennici, 6:443-447. Horwitz, B. A. 1989. Biochemical mechanisms and control of cold- induced cellular thermogenesis in placental mammals. Pp. 83-116, in Advances in Comparative and Environmental Physiology 4. Animal Adaption to Cold (L. C. H. Wang, ed.). Springer Verlag, Berlin, 441 pp. Hyvarinen, H. 1969. Seasonal changes in the activity of the thyroid gland and wintering problem at the common shrew (Sorex araneus). Aquilo Series Zoologica, 8:32-37. 1984. Wintering strategy of voles and shrews in Finland. Pp. 139-147, in Winter Ecology of Small Mammals (J. F. Merritt, ed.), Carnegie Museum of Natural History Special Publication 10, 380 pp. Hyvarinen, H., and S. Pasanen. 1973. Seasonal changes in cytochrome c content of some tissues in three small mammals active in winter. Journal of Zoology (London), 170:63-67. Lillie, R. D. 1951. The allochrome procedure. A differential segregating the connective tissue, collagen, reticulum and basement membranes into two groups. American Journal of Clinical Pathology, 21:484-488. MacArthur, R. M. 1986. Brown fat and aquatic temperature regulation in muskrats. Ondatra zibethicus . Physiological Zoology, 59:306-317. Merritt, J. F. 1986. Winter survival adaptations of the short-tailed shrew (Blarina brevicauda) in an Appalachian montane forest. Journal of Mammalogy, 67:450-464. Metcalfe, L. D., A. A. Schmitz, and J. R. Pelka. 1966. Rapid reparation of fatty acid esters from lipids for gas chromatographic analysis. Analytical Chemistry, 33:363-364. Morrison, P., F. A. Ryser, and A. R. Da we. 1959. Studies on the physiology of the masked shrew, Sorex cinereus. Physiological Zoology, 32:256-271. Nagel, A. 1985. Sauerstoffverbrauch, T emperaturregulation und HertzfrequenzbeieuropaischenSpitzmausen (Soricidae). Zeitschrift fiir Saugetierkunde, 50:249-266. Pasanen, S. 1969. On the seasonal variation of the weight and of the alkaline phosphatase activity of the brown fat in the common shrew (Sorex araneus L.). Aquilo Series Zoologica, 8:36-43. 1971. Seasonal variations in interscapular brown fat in three small mammals wintering in an active state. Aquilo Series Zoologica, 11:1-32. Pasanen, S., and H. Hyvarinen. 1970. Seasonal variation in the activity of phosphorylase in the interscapular brown fat of some small mammals active in winter. Aquilo Series Zoologica, 10:37-42. Rothwell, N. j., and M. J. Stock. 1984. Brown adipose tissue. Pp. 349-384, in Recent Advances in Physiology (P. F. Baker, ed.), Churchill Livingstone, Edinburg, 407 pp. Romeis, B. 1948. MikroskopischeTechnik. Leibniz Verlag, Munchen, 540 pp. Smalley, R. L. 1970. Developmental changes in adipose tissue. Pp. 73-96, in Brown Adipose Tissue (O. Lindberg, ed.). Academic Press, New York, 421 pp. Smith, R. E. 1961 . Thermogenic activity of the hibernating gland in the cold-acclimated rat. Physiologist, 4:113. Smith, R. E., and B. A. Horwitz. 1969. Brown fat thermogenesis. 262 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 Physiological Reviews, 49:330-425. Tomasi, T. 1984. Shrew metabolic rates and thyroxine utilization. Comparative Biochemistry and Physiology, A. Comparative Physiology, 78:431-435. Tomasi, T. E., J. S. Hamilton, and B. A. Horwitz. 1987. Thermogenie eapaeity in shrews. Journal of Thermal Biology, 12:143-147. Vogel, P. 1976. Energy consumption of European and African shrews. Acta Theriologica, 26:195-206. WUNDER, B. A. 1984. Strategies for, and environmental cueing mechanisms of, seasonal changes in thermoregulatory patterns of small mammals. Pp. 165-172, in Winter Ecology of Small Mammals (J. F. Merritt, ed.), Carnegie Museum of Natural History Special Publication 10, 380 pp. Table 1. — Relative IBAT weights (% of body mass) of nonwintered small mammals captured in late August or in September in eastern Finland, or in the vicinity of Oulu, and overwintering small mammals in January-February 1986 near Joensuu. September 1989 in Rautjdrvi (61°20'N, 29°20’E); * late August 1989 in Lieksa (63°N, 30°E); material of Pasanen (1971) captured in September 1966-70 near Oulu (65°N, 26°E). Species n August-September Body mass (g) % IBAT January-February Body mass a (g) % IBAT Sorex minutissimus T 1.9 5.20 Sorex minutus 10“ 2.9 2.56 + 0.22 3 2.2 2.67 ± 0.30 Sorex caecutiens 3“ 4.1 1.35 ± 0.10 Sorex araneus 16“ 6.7 1.23 ± 0.09 11 5.5 2.29 ± 0.22 Micromys minutus 2“ 8.05 1.15 Sorex isodon if 10.5 1.12 Neomys fodiens 2*’ 11.4 1.37 1 10.2 3.01 Myopus schisticolor 3b 13.7 0.65 +0.11 Clethrionomys glareolus 50" 15.7 0.51 Microtus agrestis 82" 25.8 0.38 Table 2. — Fatty acid composition (mean %) of interscapular brown fat of Neomys fodiens and Sorex araneus captured during summer. Fatty Acid Neomys fodiens n = 2 Sorex araneus n = 4 ^12:0 0.41 0.42 12:1 0.26 0.06 13:0 0.80 1.58 13:1 0.18 0.02 14:0 1.10 4.03 14:1 0.26 0.40 15:0 0.95 0.74 15:1 0.19 0.69 16:0 17.93 18.73 16:1 4.72 3.59 17:0 2.38 1.20 17:1 1.84 2.08 18:0 12.43 10.57 18:1 13.88 18.56 18:2 8.47 14.55 18:3 0.35 0.14 18:4 0.22 0.08 1994 HYVARINEN — Brown Adipose Tissue in Shrews 263 Table 2 (cont.) 19:0 0.33 19:1 0.22 20:0 0.04 20:1 0.78 0.81 20:2 0.40 0.48 20:3 1.36 0.69 20:4 and 20:5 17.97 8.68 22:4 0.16 22:5 4.80 2.31 22:6 8.52 6.91 Number of double bonds per mole (A/mole) 1.9-2. 1 1.4-1. 5 IBAT % Fig. 1. — The relationship of relative IBAT weight and body weight of ten small mammal species captured in August-September (material in Table 1). 264 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 Fig. 2. — Seasonal changes in IBAT weight of Sorex minutus and Sorex araneus in the vicinity of Joensuu (for material, see Table 1) and Sorex araneus in the vicinity of Oulu (+ SE; Pasanen and Hyvarinen, 1970). Number of individuals in parentheses. 1994 HYVARINEN — Brown Adipose Tissue in Shrews 265 Fig. 3.— Schematic diagram showing anatomical location of BAT in shrews. Dorsal view (A) and ventral view (B) of Neomys fodiens captured in February. Sagittal section of Sorex araneus showing distribution of BAT in trunk and hind foot (C). 266 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 Fig. 4. — Schematic diagram showing the location of BAT (stippled area) in transverse section of the thigh of Sorex araneus. Bone = black. Fig. 5. — Oblique section through the thigh of the common shrew. Brown fat (BAT) surrounding blood vessels and the sciatic nerve. M, muscle; A, artery; V, vein; N, nerve. Allochrome x 200. EFFECTS OF MELATONIN ON THE CHRONOBIOLOGY OF THE LEAST SHREW, CRYPTOTIS PARVA Orin B. Mock Kirksville College of Osteopathic Medicine, Kirksville, Missouri 63501 Abstract A principle obstacle to studies describing the mechanisms involved in Dehnel’s phenomenon is the difficulty in raising shrews of the genus Sorex in the laboratory. The least shrew, Cryptotis parva, thrives in captivity and serves as an animal model to investigate the role of melatonin in two chronobiologic events; 1) the capacity of selected age categories to respond to cold, and 2) the autumnal reduction in body length. This study did not confirm a role for melatonin in aiding pubertal (24-33 days old) and old-aged (>300 days) shrews’ ability to thermoregulate when exposed to cold. Shrews implanted with melatonin showed significantly smaller lumbar intervertebral discs than did their matched controls. Introduction Shrews of the genus Sorex often die in large numbers in the autumn (Adams, 1910). Survivors of this autumnal epidemic then show a decrease in body size (Dehnel, 1949). Understanding these seasonally related events, characterized as Dehnel’s phenomenon, offers a challenge with biological implications beyond the soricids. A principle obstacle to studies describing the mechanisms of this phenomenon is the difficulty in rearing Sorex shrews in the laboratory. The least shrew, Cryptotis parva, thrives in captivity (Mock, 1982), and may serve as an animal model for determining the physiological bases of these occurrences. This study was undertaken to determine the role of melatonin in two of these chronobiologic events — response to cold, and autumnal reduction in body length. Pearson (1945) thoroughly reviews the so-called autumnal epidemic in older shrews. He concedes that shrews are frequently found dead and discusses in some detail many of the reported causes for this occurrence. He then argues that there is little reason to believe that regulation of the life span of shrews differs significantly from that of other small, prolific mammals. I found Pearson’s explanation to be satisfactory until I provided least shrews used in a study of the response of brown fat to cold exposure (Chaffee et al., 1969). Old males died when they were cold stressed in the same marmer as other species used in the study. Subsequent experiments to determine the role of age in the least shrew’s response to cold found that animals over 300 days old showed significantly greater heat loss than did prepubertal (21-23 days) and young adult (60-90 days) shrews. Surprisingly, heat loss per unit time in pubertal animals (24-33 days) approached that of the older animals (Mock, 1985). Winter-induced decreases in shrews’ body lengths are caused primarily by a reduction in the volume of the nucleus pulposus. Some resorption in the intervertebral disc cartilage may also occur (Hyvarinen, 1969). Many hormones and enzymes have been investigated in an attempt to determine the seasonal acclimatization mechanism (Hyvarinen, 1984). Factors controlling what Hyvarinen and Heikura (1971) call the endogenous seasonal rhythm have not been elucidated. A scenario can be constructed in which melatonin is an agent in these reported chronobiologic events. Most mammals of temperate and arctic regions have seasonally limited and synchronized reproductive patterns. In many of these species, reproductive cycles depend on changes in photoperiod. The pineal gland and its primary hormone, melatonin, respond to changes in the light-dark cycle and act to inhibit reproduction during certain periods of the year (Reiter et al., 1983). Melatonin probably also functions in thermoregulatory adaptations of northern small mammals (Quay, 1984), and in events associated with the onset of puberty (Sizonenko et al., 1985). The possibility that melatonin is the agent controlling the endogenous seasonal rhythm of shrews serves as the basis for this study. Materials and Methods All the shrews used in the course of this study were maintained according to a regimen previously described (Mock, 1982) in a breeding colony at Kirksville College of Osteopathic Medicine. Melatonin for implanting was prepared by kneading one part melatonin (Sigma Chemical Co., St. Louis, Missouri) with four parts beeswax (Impression Wax, The Hygenic Corp., Akron, Ohio). Five-milligram lots of the mixture were then compressed into 2.5 mm diameter pellets. The 5 -mg control pellets contained beeswax only. The pellets were injected subcutaneously via needles designed for implanting medicated pellets into the pinnal cartilage of cattle. Dosages of 2 mg of melatonin were used in some of the initial cold-stress studies for pubertal animals. The duration of melatonin treatment prior to cold exposure was 8-10 days for old-aged shrews. Duration of treatment for pubertal males varied from 2-8 days and is so indicated in Table 1. Needle microprobes attached to a digital thermometer (BAT- 12, Bailey Instruments Inc., Saddle Brook, New Jersey) were placed subcutaneously along the flanks of restrained shrews of selected ages. Initial temperatures were recorded, and after 5 min, the animals were placed in a cold room (3-5°C) and body temperatures were recorded at regular intervals for 30 min. The temperature value reported was the difference between the initial and final readings. The intervertebral disc protocol involved allowing the implanted animals to survive for an eight- or ten-day treatment 267 268 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 time. Animals were sacrificed and their vertebral columns fixed in 10% buffered formalin. After proper fixation the tissue ventral to the vertebral column was carefully removed, exposing the lumbar vertebrae and intervertebral discs. A ventral view of the five lowest lumbar discs was measured using a calibrated ocular disc in a dissecting microscope. Agreement of two observers was required before any measurement was recorded. All vertebral columns were given an identification number by a technician who was not involved in the measuring. Observers were unaware of the source or treatment of vertebral columns examined. The matched animals used in the melatonin studies were sibling pairs with the exception of one pair of old females. These animals were one day apart in age and from the same lines with similar histories. Results and Discussion A role for melatonin in protecting pubertal and old age shrews from rapid heat loss during exposure to cold was not demonstrated by these experiments (Table 1, 2). However, these findings do reconfirm the previous report that least shrews of selected age classes show limited ability to thermoregulate (Mock, 1985). The age categories for which young shrews show an inability to properly respond to cold are expanded with this study. In fact, the term pubertal probably is inappropriate for describing the affected animals. Captive male least shrews are sexually mature prior to 55 days of age (Mock and Conaway, 1976). The evolutionary advantages of a factor that selectively removes older individuals from a population prior to periods of limited food availability are obvious. It is difficult, however, to envision the benefit of eliminating a major age category of young animals. It is possible that pubertal and young adult least shrews are not found in autumnal populations. Factors similar to those reported for voles may be present that retard sexual development in autumnal offspring exposed to short gestational or lactational photoperiods (Horton, 1984, 1985). Shrews implanted with melatonin showed significantly smaller intervertebral disc lengths than did their matched controls (Table 3). The primary change in disc size is the reduction in the volume of the nucleus pulposus (Hyvarinen, 1969). The major components of the nucleus pulposus are proteoglycan complexes and the water which they bind (Humzah and Soames, 1988). A likely mechanism to explain these observations is a direct or indirect effect of melatonin on glycosaminoglycans or the proteins to which they are linked. Acknowledgments This research was supported by American Osteopathic Association Grant number 81-04-306. Thanks are due to J. H. Shaddy and M. Tannenbaum, Northeast Missouri State University, for their assistance in computations and a critical review of the manuscript. I am particularly grateful to M. Mock, S. Catlett, M. Baumeier, N. Hoyal, M. Kubiak, G. Matlock, and W. Wobken for assistance in this task. Literature Cited Adams, C. F. 1910. A hypothesis as to the cause of the autumnal epidemic of the common and lesser shrew, with some notes on their habits. Manchester Memoirs, 54:1-13. Chaffee, R. R. J., J. C. Roberts, C. H. Conaway, M. W. Sorenson, and W. C. Kaufman. 1969. Comparative effects of temperature exposure on mass and oxidative enzyme activity of brown fat in insectivores, tupaiads and primates. Lipids, 5:23-29. Dehnel, A. 1949. Badania nad rodzajem Sorex L. Annales Universitaties Mariae Curie-Sklodowska, Section C, 4:17-102. Horton, T. H. 1984. Growth and maturation in Microtus montanus: Effects of photoperiods before and after weaning. Canadian Journal of Zoology, 62:1741-1746. 1985. Cross-fostering of voles demonstrates in utero effect of photoperiod. Biology of Reproduction, 33:934-939. Humzah, M. D., and R. W. Soames. 1988. Human intervertebral disc: Structure and function. Anatomical Record, 220:337-356. Hyvarinen, H. 1969. On the seasonal changes in the skeleton of the common shrew (Sorex araneus L.) and their physiological background. Aquilo, Serle Zoologica, 7:1-32. 1984. Wintering strategy of voles and shrews in Finland. Pp. 139-148, in Winter Ecology of Small Mammals (J. F. Merritt, ed.), Carnegie Museum of Natural History Special Publication no. 10, 380 pp. Hyvarinen, H., and K. Heikura. 1971. Effects of age and seasonal rhythm on the growth patterns of some small mammals in Finland and in Kirkenes, Norway. Journal of Zoology (London), 165:545-556. Mock, O. B. 1982. The least shrew (Cryptotis parva) as a laboratory animal. Laboratory Animal Science, 32:177-179. 1985. The effects of cold stress on Cryptotis parva. Acta Zoologica Fennica, 173:269-270. Mock, O. B., and C. H. Conaway. 1976. Reproduction of the least shrew (Cryptotis parva) in captivity. Pp. 59-74, in The Study of Reproduction (T. Antikatzides, S. Erichsen, and A. Spiegel, eds.), Gustav Fischer Verlag, Stuttgart, New York, 185 pp. Pearson, O. P. 1945. Longevity of the short-tailed shrew. American Midland Naturalist, 34:531-546. Quay, W. B. 1984. Winter tissue changes and regulatory mechanisms in nonhibemating small mammals: A survey and evaluation of adaptive and non-adaptive features. Pp. 149-163, in Winter Ecology of Small Mammals (J. F. Merritt, ed.), Carnegie Museum of Natural History Special Publication no. 10, 380 pp. Reiter, R. J., B. A. Richardson, and T. S. King. 1983. The pineal gland and its indole products: Their importance in the control of reproduction in mammals. Pp. 151-199, in The Pineal Gland (R. Reiter, ed.), Elsevier Biomedical, New York, 311 pp. SizoNENKO, P. C., U. Lang, R. W. Rivest, and M. L. Aubert. 1985. The pineal and pubertal development. Pp. 208-225, in Photoperiodism, Melatonin and the Pineal (D. Evered and S. Clark, eds.), Ciba, The Bath Press, Avon, 323 pp. 1994 MOCK — Chronobiology of the Least Shrew 269 Table 1. — Wilcoxon’s signed-rank test for evaluating the effects of melatonin on pair-matched, male, pubertal least shrews’ response to cold values show body temperature loss (°C) per 30 min at 3-5 °C. P > 0.05. Animal Numbers Age (Days) Dosage (mg) Duration of Treatment (Days) Control Melatonin Treated 13532-13534 40 1 8 13.6 22.2 14005-14013 55 1 3 15.1 13.0 14022-14021 36 2 4 18.1 12.1 14123-14124 36 2 4 10.2 7.0 14201-14202 37 2 3 3.9 9.2 14215-14220 36 2 2 15.2 6.7 14224-14221 38 2 4 11.0 22.5 14233-14232 33 2 3 19.1 14.4 14245-14252 31 2 4 12.1 16.0 Table 2. — Wilcoxon ’s signed-ranks test for comparing effects of cold treatment on melatonin implanted and control, pair-matched , old-aged least shrews values equal heat loss (°C) per 30 min at 3-5°C. P > 0.05. Animal Numbers Sex Age (Days) Control Melatonin Treated D (°C) 15005-15004 66 443 7.9 12.3 -4.4 15025-15024 66 411 16.9 11.5 5.4 14513-14505 66 531 17.0 13.6 3.4 14551-14554 66 464 14.0 6.2 7.8 14431-14430 99 736 18.6 11.1 7.5 14543-14542 99 501 18.9 21.2 -2.3 14424-14421 99 748 11.8 18.5 -6.7 15000-14553 C?(? All 20.6 21.8 -1.2 15020-15015 99 458 22.1 19.0 3.1 14541-14532 66 503 20.7 14.6 6.1 Table 3. — Paired comparisons t-test for comparing the effects of melatonin implants on the total length of the lower five lumbar intervertebral discs in 90-day-old least shrews. Values are in mm. Treatment duration was 8 or 20 days. P < 0.001. Animal Numbers Sex Control Melatonin Treated D D^ 12003-12004 66 2.950 2.475 .475 0.226 12305-12302 66 2.825 2.550 .275 0.076 12311-12310 66 2.250 2.225 .025 0.001 12314-12312 66 2.550 2.300 .250 0.062 12313-12315 99 2.550 2.350 .200 0.040 12324-12322 99 2.750 2.440 .310 0.096 12325-12323 dd 2.475 2.175 .300 0.090 12321-12320 99 2.325 2.250 .075 0.006 12502-12501 99 2.285 2.105 .180 0.032 12420-12415 66 2.350 2.260 .090 0.008 12425-12424 66 2.235 2.015 .220 0.048 12500-12503 99 2.315 2.105 .210 0.044 X = 2.488 X = 2.271 E = 2.135 0.503 „ *N :? Ss ““ V- " ' i» ^ ,^ • 3 ;J!}* •r ••;••.?*« it’.' S' ■ ■■ ■*/:• -> ■ :■. ,r J •• .M •?/» • i >H Y • •V. »2«l4< - ■- vC*” I ’^ • ' f*rr' ••>‘.Ti"aS4 *>v^ ' ■ ^'' ■■••MM*. '«■• 5MW«H|U.V -if Ifeii -.^ ^.soa >'• ?M*r 'f' -ij,.- ...■ , 13V^’ I '■ ■■ ■ 1 f .«*«l ti r ^ *•’ -■'r ill ?!&!:)« jf’.?!™' rTsV— '■" \nm ■* * f ■ r 'l.'.JiJM-ujA-j SJ'"/ r „'( •'l*.| ■ s-.r-'.^l •'■'■Ssl'i^t'i#!: * 5 hJ ' A" < '”" ■ )i: -:m. >.- ' I 1 •I • , J^'- ' ' * .' ■■v“: ■■ **u5 'fp • w'Sij'AiV'rt ,•.„- ..•injt.f.\ ■ ■* ■■’ ' ■" - K- - f ;'! ‘ ■" .-, i. • '.:♦/■&• 5.. ■''■’■' -f. ' lcf. P’ ’ ■ ■■ ^ V' ^ wf( K. >nyr , ■ . > >,• i'TIJS-1. _ 0.05, d.f. = 5), and the second relationship is therefore taken to suggest that heavier females produced larger litters. No relationship was found between litter size and mean weanling mass, indicating that smaller litter size is not compensated for by larger weanlings. Sex Ratio. — Of those offspring sexed on weaning, 57 % (119 of 209 animals) were males. Although the sex ratio among weanlings did not differ significantly from 50%, there was a clear tendency towards male bias, which agrees with the findings of Brambell (1935a), who gives an average sex ratio of 54% males for wild-caught animals throughout the year (n = 1,064). (Note that male bias determined in nature may partly reflect behavioral differences, particularly in the spring: Crowcroft, 1957; Skaren, 1973; Pucek, 1959.) There were no significant differences in sex ratio between litters conceived in nature and those conceived in captivity. Response to Photoperiod. — Thirteen males survived to yield data on responses to photoperiod (Table 3, 4), and on the basis of testis and seminal vesicle masses, males subjected to each of the three photoperiod regimes attained sexual maturity. As regards those individuals kept singly (Table 3), it is particularly striking that one individual each from regimes B (four months LD) and C (two months SD, two months LD) had a combined testis mass greater than 250 mg, comparable with males at the height of sexual development. Associated with testis growth, the skin over the testes became bare in some males, as has been recorded in wild-caught individuals (Searle, 1985a). One individual from regime C demonstrated regrowth of hair on this bare patch, also consistent with animals in nature (J.B.S., personal observation). Five animals were housed in groups of three and two animals respectively (see Table 4). Of the three males subjected to two months LD (regime A), one was of immature size and reproductive condition, while the others clearly showed testis and seminal vesicle development. In the other group (four months LD, regime B), one had barely begun to mature, while the other showed clear signs of adulthood, both in reproductive and body growth. Neither individual showed sexual development to the degree found in regime B animals housed singly (Table 3). Inhibition of sexual maturation due to the proximity of another animal is a well-documented phenomenon in both sexes (Lee and McDonald, 1985; Spears and Clarke, 1986), and it is interesting that this may occur in the common shrew despite its normally solitary nature (Croin-Michielsen, 1966). All animals were in good health, and it is considered unlikely that individuals remained immature due to insufficient food supply. Although no significant differences were found between animals kept singly under regimes B and C, and individual variation was large, some trends should be noted (Table 3). Seminal vesicle mass was similar in the two regimes (indicating similar androgen levels), but testis mass was higher on average under regime B (four months LD). In contrast, the animals under regime C (two months SD, two months LD) had a more adult body-to-tail ratio, with a mean (±S£) of 1.87 ± 0.13 (mean for regime B: 1.59 + 0.09). In both regimes there were males with adult testis mass and immature body proportions, suggesting sexual and physical maturation are, to an extent, independent processes, although in nature they usually occur concurrently. Pucek (1960) who clearly demonstrated that female common shrews of immature body proportions become sexually mature, made a similar postulation. The lateral scent gland, which in nature undergoes substantial development only in the male (Searle, 1985a), became active in individuals of all three photoperiodic regimes, in parallel with body size and testis development. Inherited Characters. — Breeding studies have demonstrated that alleles at the Mpi-1 , Pgm-2, and Pgm-3 enzyme loci segregate in a Mendelian fashion in the common shrew (Searle, 1983, 1985^). We report here on segregation studies of other polymorphic features. Various patterns of distribution of white fur have been noted on the body of the common shrew (Crowcroft, 1955). While white nape patches may be found in adult females as a result of damage during mating (Crowcroft, 1955) and a generalized white peppering is associated with old age (Searle, 1983), the incidence of white patches (commonly found on the ears, but also found in association with the feet, abdomen, and tip of the tail; Crowcroft, 1955) may have a genetic basis. If white spotting results from the expression of an allele at a particular locus, it would seem likely that the gene is analogous or homologous to the recessive, nonpleiotropic white-spotting gene of variable penetrance (j-), found in several other marrunals, e.g., guinea pigs (Searle, 1968). The results from breeding studies are consistent with the interpretation that white spotting is controlled by an J-type gene. A number of offspring from crosses between white-spotted shrews did not display any form of white spotting (data in Searle, 1983), consistent with expression of a recessive trait. The fact that within an individual one ear may have white hair and the other not (Crowcroft, 1955) suggests that the white- spotting allele is not fully penetrant. Breeding studies have also proved invaluable for cytogenetic analysis. In Britain, there are three karyotypic races of common shrew (“Oxford,” “Aberdeen,” and “Hermitage”), each with a distinctive chromosomal complement. All of these races can be crossed in captivity and it has been demonstrated that litter sizes are similar to those derived from intraracial crosses (Searle, 1984Z?). Furthermore, we have been able to bring Ox ford - Aberdeen race hybrids to maturity by photoperiod manipulation (as described above), and to study their fertility (Mercer et al., 1992). Searle (1986) has also examined transmission of variant chromosomes from chromosomal heterozygotes. The results from wild-caught females are unreliable due to multiple paternity (which has itself been demonstrated with the help of this breeding study — Tegelstrom et al., 1991; see also Searle, 1990), but those from crosses in captivity can be used. Conclusions The success of this breeding program may be measured in 274 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 terms of percentage of fertile crosses, survival of young to weaning, and postweaning survival. In all of these respects, and especially in terms of the number of young reared, the method has shown itself to be remarkably reliable. Other workers (M. Dodds-Smith, P. Stockley, personal communications) have successfully adopted the same protocol. Altogether 50% of attempted crosses were fruitful. Collecting adult females during their first pregnancy rather than before enhances the success of subsequent captive crosses. Of those offspring known to have been produced, 87 % survived to weaning (excluding those young deliberately killed before weaning), but postweaning mortality was sometimes high, especially if the young were exposed to excessive disturbance. The sample of animals subjected to photoperiodic manipulation was too small to allow thorough quantification of any of the observed effects. It seems certain, however, that prolonged periods of long days (16L:8D) stimulate sexual maturation (as indicated by testis size and seminal vesicle growth) in young common shrews. Under these conditions, sexual development may occur in the absence of body growth, and both may be inhibited when common shrews under long photoperiod are housed in groups. These observations deserve further study. We have demonstrated that captive maintenance of the common shrew is feasible at all stages of the life cycle, and most of the technical obstacles to the establishment of a breeding colony have been removed. Shrews can be kept in standard laboratory cages for several months at a time, and early sexual maturation can be induced. The principal problems remain those of the frequency of feeding and the level of mortality, but both of these are likely to be overcome through modification of existing techniques. The benefits of a permanent colony of common shrews for studies of behavior, traditional genetics, and reproductive biology would be substantial. In particular, such a colony would permit a more detailed experimental approach than hitherto possible for the study of the phenomenal chromosomal variation in this species. Acknowledgments We thank M. J. Amphlett and C. A. Everett for assistance in collection of animals. This work was supported by grants from the Royal Society of London (J.B.S.) and the Natural Environment Research Council (S.J.M.). Literature Cited Adams, L. F. 1912. The duration of life of the common and the lesser shrew, with some notes of their habits. Memoranda and Proceedings of the Manchester Literary and Philosophical Society, 56:1-10. Brambell, F. W. R. 1935a. Reproduction in the common shrew (Sorex araneus Linnaeus) I — The oestrous cycle of the female. Philosophical Transactions of the Royal Society of London, Series B, 255:1-50. 1935fc. Reproduction in the common shrew (Sorex araneus Linnaeus) II — Seasonal changes in the reproductive organs of the male. Philosophical Transactions of the Royal Society of London, Series B, 255:51-62. Chitty, D., and D. a. Kempson. 1949. Prebaiting small mammals and a new design of live trap. Ecology, 30:536-542. CROIN-MiCHlELSEiN, N. 1966. Intraspecific and interspecific competition in the shrews Sorex araneus L. and S. minutus L. Archives Neederlandaises de Zoologie, 17:73-174. Crowcroft, P. 1955. Remarks on the pelage of the common shrew (Sorex araneus L.). Proceedings of the Zoological Society of London, 123:715-729. 1956. On the life span of the common shrew (Sorex araneus L.). Proceedings of the Zoological Society of London, 127:286-292. 1957. The Life of the Shrew. Max Reinhardt, London, 166 pp. Dehnel, A. 1952. The biology of breeding of the common shrew S. araneus L. in laboratory conditions. Annales Universitatis Marie Curie Sklodowska, Series C, 6:359-376. Fedyk, S. 1980. Chromosome polymorphism in a population of Sorex araneus L. in Bialowieza. Folia Biologica Krakow, 28:83-120. Ford, C. E. 1966. The use of chromosome markers. Pp. 197-206, in Tissue Grafting and Radiation (H. S. Micklem and J. F. Loutit, eds.). Academic Press, New York, 228 pp. Genoud, M., and P. Vogel. 1990. Energy requirements during reproduction and reproductive effort in shrews (Soricidae). Journal of Zoology, 220:41-60. Grocock, C. A., AND J. R. Clarke. 1974. Photoperiodic control of testis activity in the vole, Microtus agrestis. Journal of Reproduction and Fertility, 39:331-347 . Hellwing, S. 1971. Maintenance and reproduction in the white- toothed shrew, Crocidura russula monarcha Thomas, in captivity. Zeitschrift fiir Saugetierkunde, 36:103-113. Jewell, P. A., and P. J. Fullagar. 1966. Body measurements of small mammals: Sources of error and anatomical changes. Journal of Zoology, 150:501-509. Lee, A. K., and I. R. McDonald. 1985. Stress and population regulation in small mammals. Pp. 261-304, in Oxford Reviews of Reproductive Biology 7 (J. R. Clarke, ed.), Oxford University Press, 409 pp. Mercer, S. J., B. M. N. Wallace, and J. B. Searle. 1992. Male common shrews (Sorex araneus) with long meiotic chain configurations can be fertile: Implications for chromosomal models of speciation. Cytogenetics and Cell Genetics, 60:68-73. Oda, S.-I., j. Kitoh, K. Ota, and G. Isomura (eds.). 1985. Suncus murinus. Biology of the Laboratory Shrew. Japan Scientific Societies Press, Tokyo, 522 pp. PUCEK, Z. 1959. Some biological aspects of the sex-ratio in the common shrew (Sorex araneus araneus L.) Acta Theriologica, 3:43-73 1960. Sexual maturation and variability of the reproductive system in young shrews (Sorex L.) in the first calendar year of life. Acta Theriologica, 3:269-296. Searle, A. G. 1968. Comparative Genetics of Coat Colour in Mammals. Logos Press/ Academic Press, London, 308 pp. Searle, J. B. 1983. Robertsonian chromosomal variation in the common shrew Sorex araneus L. Unpublished Ph.D. thesis. University of Aberdeen, Scotland, 177 pp. 1984a. Breeding the common shrew (Sorex araneus) in captivity. Laboratory Animals, 18:359-363. 1984/7. Hybridization between Robertsonian karyotypic races of the common shrew Sorex araneus. Experientia, 40:876-878. 1985a. Methods for determining the sex of common shrews (Sorex araneus). Journal of Zoology, 206:279-282. 1985/7. Isoenzyme variation in the common shrew (Sorex araneus) in Britain, in relation to karyotype. Heredity, 55:175-180. 1986. Preferential transmission in wild common shrews (Sorex araneus), heterozygous for Robertsonian rearrangements . Genetical Research, 47:147-148. 1994 MERCER AND SEARLE — Breeding the Common Shrew 275 . 1990. Evidence for multiple paternity in the common shrew (Sorex araneus). Journal of Mammalogy, 71:139-144. Skar^, U. 1973. Spring moult and onset of the breeding season of the common shrew (Sorex araneus L.) in central Finland. Acta Theriologica, 18:443-458. Spears, N., and J. R. Clarke. 1986. Effect of male presence and of photoperiod on the sexual maturation of field vole (Microtus agrestis). Journal of Reproduction and Fertility, 78:231-238. Stockley, P., and j. B. S EARLE. 1994. Characteristics of the breeding season in the common shrew (Sorex araneus): Male sexual maturation, morphology and mobility. Pp. 181-187, in Advances in the Biology of Shrews (J. F. Merritt, G. L. Kirkland, Jr., and R. K. Rose, eds.), Carnegie Museum of Natural History Special Publication no. 18, x + 458 pp. TEGELSTROM, H . , j . SEARLE, J . BROOKFIELD, AND S. MERGER. 1991. Multiple paternity in wild common shrews (Sorex araneus) is confirmed by DNA-fingerjirinting. Heredity, 66:213-319 . Vlasak, P. 1973. Vergleich der postnatalen Entwicklung der Arten Sorex araneus L. und Crocidura suaveolens (Pall.) mit Bemerkungen zur Method ik der Laborzucht (Insectivora: Soricidae). Vestnik Ceskoslovenske Spolecnosti Zoologicke, 37:222-233. Vogel, P. 1972. Vergleichende Untersuchungzum Ontogenesemodus einheimischer Soriciden (Crocidura russula, Sorex araneus und Neomys fodiens). Revue Suisse de Zoologie, 79: 1201-1332. Table 1. — Success of captive breeding in the common shrew. ^ minimum estimates; * data of Searle (1984a); ^ data of S.J.M.; ^ total of 23 attempted crosses. Three females were killed during gestation; ^ excludes crosses with nulliparous females (nine, all failed);^ includes only one complete litter (of a single animal). Preweaning At Weaning Where Failed Number of Animals (Litters) Mean Litter Proportion Conceived Crosses Bom“ Killed Died* Weaned Size Males Wild*’ — 80 (15) 8 (1) 29 (6) 43 (8) 5.4 55% Captive*’ 6** 69 (14) 7 (1) 4 (1) 58 (12) 4.8 Wild*’ — no (17) 0 3 (3)f 107 (16) 6.7 55% Captive*’ 10^ 35 (8) 0 0 35 (8) 4.4 66% WILD — 190 (32) 8 (1) 32 (9) 150 (24) 6.3 — CAPTIVE 16 104 (22) 7 (1) 4 (I) 93 (20) 4.7 — TOTAL 294 (54) 243 (44) 5.5 57% Table 2. — Body (including head) lengths and tail lengths for a representative sample of adult and immature common shrews. The adults were collected from the vicinity of Oxford (United Kingdom) during 10-25 May 1981. The immatures (the progeny of a variety of crosses involving common shrews from the vicinity of Oxford and Aberdeen, United Kingdom) were reared in captivity and measured within five days of weaning from 27 July to 11 August 1981. Age and Length (Mean + S.E., in mm) Body-to-Tail Sex n Body Tail Ratio Adult females 10 82.67 ± 0.69 39.92 ± 0.89 2.1 Adult males 10 83.03 ± 0.50 41.61 ± 0.71 2.0 Immature females Litter 1 1 66.5 42.5 1.6 Litter 2 4 75.25 ± 0.71 42.52 ± 0.24 1.8 Litter 3 2 71.65 ± 0.05 40.20 ± 0.20 1.8 Litter 4 2 71.85 ± 1.45 40.05 ± 2.75 1.8 Litter 5 4 69.05 ± 0.46 39.65 ± 0.76 1.7 Immature males Litter 1 7 70.96 + 0.37 43.04 ± 0.49 1.6 Litter 2 2 72.85 + 1.45 44.80 ± 0.60 1.6 Litter 3 3 72.93 + 0.57 41.53 ± 0.93 1.8 Litter 4 3 69.93 ± 1.09 37.37 + 1.10 1.9 Litter 5 4 69.68 ± 0.86 40.38 ± 1.28 1.7 276 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 Table 3. — The extent of sexual maturation in male shrews kept singly under different photoperiod regimes. age in days from weaning; * animal died prematurely, and therefore received approximately five weeks less long photoperiod than the other two regime B animals (86 days, as opposed to 123 and 124 days). Its loss of condition is reflected by its low body mass. Photoperiod Regime Age“ Combined Fresh Weights (mg) Testes Seminal Vesicles Body Weight (g) Body-to-Tail Ratio B 191 282 105 9.27 1.51 (4 months LD) 192 124 122 9.09 1.76 145^ 156 82 5.64 1.49 Means: lie 187 103 8.00 1.59 C 189 255 100 8.46 1.60 (2 months LD + 191 71 50 8.28 1.84 2 months SD) 191 127 135 10.09 2.32 192 79 115 8.65 1.60 182 88 136 8.48 1.97 Means: 189 124 107 8.79 1.87 Table 4. — The extent of sexual maturation in male shrews kept in groups under different photoperiod regimes. “ age in days from weaning; * too small to weigh accurately; sperm and meiotic division observed in testes; no sperm in testes, no indications of meiosis. Photoperiod Regime Age“ Combined Fresh Weights (mg) Testes Seminal Vesicles Body Weight (g) Body-to-Tail Ratio A 112 117 118 8.6 (2 months LD) 112 75 29 6.5 — 112 5 b 6.2 — B 194 54^= 12 7.8 1.82 (4 months LD) 194 20^* b 7.6 1.37 8- O 2- O O I J 1 — . 1 — 1 9.5 10 10.5 II 11.5 12 12.5 13 13.5 M 14.5 15 FEMALE BODY MASS (g) AT WEANING OF PREVIOUS LITTER Fig. 1. — Relationship between the size of the second or third litter and the weight of the mother at the weaning of her previous litter. For two females, the data for both the second and third litters are included on this graph, but for the statistical analysis (see text) single mean values were calculated for these females. PROPOSED STANDARD PROTOCOL FOR SAMPLING SMALL MAMMAL COMMUNITIES Gordon L. Kirkland, Jr.* and Patricia Krim Sheppard* * Vertebrate Museum, Shippensburg University, Shippensburg, Pennsylvania 17257 Abstract Aecurate estimates of small mammal eommunity structure are difficult to obtain due to numerous sources of variation, including trap type and method of deployment, duration of sampling, and weather conditions. This is particularly true in the case of shrews (Soricidae), which are underrepresented by traditional sampling methods. To foster research on the community ecology of small mammals, particularly shrews, we propose adoption of a standard protocol for sampling small mammals. This sampling protocol involves deploying Y-shaped arrays of ten pitfall traps and drift fences. Each arm, which is anchored on a central pitfall, consists of three pitfalls separated by 5-m sections of drift fence. Pitfalls, not less than 14 cm in diameter and 19 cm deep, should be filled approximately half full of water to quickly drown captured animals. Due to the significant influence of precipitation on sample data, we recommend that arrays be operated for ten consecutive days, which in temperate forest regions will usually encompass at least one precipitation event. Use of the proposed sampling protocol will reduce sources of error related to trap type, method of trap placement, and the trapping skill levels of individual investigators. This will facilitate comparisons of community data between sites and provide better estimates of abundance and structure of soricid communities than nonstandard ized sampling methods. Introduction it is often difficult to compare the structure of terrestrial small Shrews (Soricidae) are important constituents of small mammal communities in forested regions of the Holarctic and Old World tropics (Kirkland, 1991). Despite their importance in terms of numbers of species and individuals, shrews apparently are underrepresented in many studies of small mammal community structure. A key factor is trap type, which can influence the capture rate of individual species (Cockrum, 1947; Wiener and Smith, 1972; Pizzimenti, 1979; Williams and Braun, 1983) and thus perception of small mammal community structure. Live traps typically underestimate the abundance of shrews (Pucek, 1969), whereas pitfall traps are very efficient in capturing soricids, especially the smallest species (Prince, 1941; MacLeod and Lethiecq, 1963; Wolfe and Esher, 1981). Even old and new models of the Museum Special snap trap differ in their sensitivity to individual species of small mammals, including shrews (West, 1985). Added to variation attributable to trap type is the considerable difference in skill levels of individual investigators in deploying traps, particularly snap traps. Other potential sources of variation include the method of deployment of traps (e.g., randomly vs. systematically placed traps) and interstation distances in grids and transects. Another potentially important source of error between studies is variation due to weather and moon phase. Not only can precipitation have a significant influence on activity levels of small mammals (Burt, 1940; Sidorowicz, 1960; Mystkowska and Sidorowicz, 1961; Falls, 1968; Drickamer and Capone, 1977; Pankakoski 1979a), but hard rains can set off snap traps, particularly those placed in exposed sites, thus reducing probability of capture. Small mammals may respond to high ambient light during the full moon by decreasing activity (Blair, 1951) or by spatially shifting foraging patterns (Bowers, 1988, 1990). As a consequence, samples of small mammals obtained during periods of precipitation or full moon may differ quantitatively and qualitatively from samples obtained during periods of clement weather or new moon. One consequence of these various sources of variation is that mammal communities as revealed by different studies. In fact, potential sources of error and variability are so numerous that it seems impossible to expect that valid comparisons of community structure could be made between studies. Nevertheless, we believe extraneous variation between studies can be minimized thereby facilitating valid comparisons. An essential step in this direction would be adoption of a standard protocol for sampling small mammal communities. Use of a standard sampling protocol would alleviate some of the problems normally encountered in comparing estimates of small mammal conununity structure obtained by different methods, and thereby foster an understanding of regional variation in small mammal community structure. This would be especially important in the case of shrews, which are sensitive to sampling techniques and are often underrepresented in samples obtained by conventional trapping (Wolfe and Esher, 1981). The sampling protocol proposed herein, which was initially tested in five forest habitats in south-central Pennsylvania, USA, employs Y-shaped arrays of pitfalls and drift fences. Results of sampling from September to November 1987 demonstrated the effectiveness of the proposed standard sampling protocol in capturing shrews. In this paper, we describe the sampling protocol, present the results of initial sampling with the pitfall arrays, compare data obtained by pitfall and snap trap sampling in the same habitats, and provide an annotated list of equipment and supplies needed for the proposed protocol. Description of Pitfall Arrays and Proposed Protocol Arrays of ten pitfall traps (minimum size of 14 cm diameter X 19 cm deep) and drift fences (5-m long sections of 25-30 cm wide aluminum flashing, plastic sheeting, or similar material) should be arranged as illustrated in Fig. 1. Each arm of an array consists of three pitfalls connected by sections of drift fence, which extend to the central pitfall. The three arms should be separated by arcs of approximately 120°. This configuration 277 278 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 minimizes directional bias in sampling. Drift fencing should be buried to a depth of 3-5 cm to prevent small mammals from crawling or burrowing underneath. Pitfalls should be installed so that the rims are flush with or slightly below the soil surface. One problem frequently encountered when installing pitfalls is soil getting into the pitfalls as it is back-filled around the traps. If tapered plastic containers are used for pitfalls, the amount of soil getting into pitfalls during installation can be minimized if two traps are nested before the pitfall is placed in the ground. As soil is filled in around the pitfall, material will fall into the upper pitfall. When this pitfall is removed, the lower one will be clean and its rim will be slightly below the soil surface. Traps should be filled with water to a depth of approximately 9 cm (half full). There should be sufficient water in traps to prevent animals from touching the bottom and thereby jumping out, but not enough water to permit floating animals to reach the rim. Based on experience in the north temperate forest biome, we recommend operating pitfall arrays for ten-day periods, checking them daily. This totals 100 trapnights (TN) of sampling effort per array per sampling period and permits easy calculation of relative abundance as percentage capture success. Operating individual arrays for the same number of nights also facilitates comparison between arrays, either as whole numbers or as percentage capture success. The minimum ten-day sampling period is justified in part by the considerable labor needed to establish an array. However, a more important consideration is to ensure that each sampling period encompasses at least one precipitation event, given the significant influence of precipitation on small mammal activity. The ten-day sampling period is particularly applicable in temperate deciduous forest regions where precipitation is relatively high and evenly distributed throughout the year, and a seven-day cycle of storms or weather fronts occurs. If there is no precipitation during the first ten days of sampling, we recormnend extending the sampling period until there is a precipitation event, and perhaps one or two days after precipitation occurs, or until the capture rate returns to prerainfall levels. In regions of low or highly seasonal precipitation, adjustment in the length of sampling periods should be made to accommodate prevailing local climatic conditions. Methods and Materials Arrays of pitfalls and drift fences were initially tested in five forest habitats (two arrays per habitat) on South Mountain, Cumberland County, south-central Pennsylvania, USA. Habitats sampled were mature lowland mixed deciduous forest, midelevation chestnut oak (Quercus prinus) forest, ridge forest dominated by chestnut oak and black gum (Nyssa sylvatica), three-year-old clearcut, and nine-year-old clearcut. The two clearcuts were in predominantly chestnut oak stands. The ten arrays were run concurrently for 39 days between 22 September and 21 November 1987. For comparative data from snap trap sampling, we used data obtained by the first author at 18 localities in central Pennsylvania, including two recent (< 5-years -old) clearcuts. This sampling included both trapping grids and best site trapping (Kirkland, 1979; Kirkland and Hench, 1980; Combower and Kirkland, 1983). Results and Discussion Initial Sampling with Proposed Protocol and Comparisons } Initial sampling of small mammals using arrays of pitfalls and drift fences revealed soricids to be considerably more important constituents in the five habitats sampled than was expected based on snap trap sampling (Table 1). Not only did pitfall trapping reveal a significantly higher proportion of [• soricids in the sample, but there were significant shifts in the j; perceived structure of the soricid assemblage (Table 1). The j! snap trap sample was dominated by the large, nearly mouse-size ! short-tailed shrew (Blarina brevicauda), which comprised I 77.9% of the shrews taken, whereas this species comprised less j| than one-third of soricids taken in the pitfall arrays (Table 1). j This is consistent with the results of Mengak and Guynn (1987), | who found that in South Carolina Blarina carolinensis comprised 98 % of shrews taken in Museum Special snap traps t but only 36% of soricids taken with pitfalls and drift fences. ] Long-tailed shrews (Sorex spp.) were significantly more abundant in the small mammal community and in the soricid ij assemblage based on pitfall trapping (Table 1). j Noteworthy in pitfall sampling was capture of 37 pygmy ] shrews (Sorex hoyi). This species was first collected in j| Pennsylvania by pitfall sampling (Kirkland et al., 1987). ■ Previous live and snap trap sampling of small mammals on J South Mountain by the first author over a 17 -year period failed I to capture this species. Yet, S. hoyi comprised 10% of the | pitfall sample and was approximately evenly distributed in the | five habitats sampled (Table 2). These data indicate that I sampling with the proposed standard protocol yielded much j better estimates of soricid abundance and community j composition than conventional snap trap sampling. I As the proportion of soricids in the pitfall sample was larger, I the relative importance of rodents was proportionately reduced. When data for the most abundant rodent species were analyzed, the proportions of Peromyscus spp. (combined white-footed mouse, P. leucopus, and deer mouse, P. maniculatus), and southern red-backed vole (Clethrionomys gapperi) were significantly lower in the pitfall sample (Table 1). However, when compared in the context of their respective contributions to the rodent assemblage, proportions of these species did not differ between pitfall and snap trap samples (Table 1). This suggests that pitfall samples provide valid estimates of the relative abundance of these species within rodent assemblages but may underestimate their overall abundance. Part of this difference may be related to the behavior of individual species. For example, the tendency of forest-dwelling populations of Peromyscus maniculatus and P. leucopus to utilize downed trees ; as routes of travel (Plantz, 1987; Graves et al., 1988), may make them less subject to capture in pitfalls than in snap traps, which are often set at the bases of trees or on logs where these species are active. Pitfall arrays tend to be installed in open areas between trees. 1994 KIRKLAND AND SHEPPARD — Proposed Standard Sampling Protocol 279 Although total numbers of rodents may be underestimated in pitfall sampling, data for P. leucopus and C. gapperi (Table 2) indicate that pitfall sampling provides good estimates of relative abundance of rodent species in individual habitats. Peromyscus leucopus, an abundant habitat generalist, did not differ in abundance among the five habitats (x^ = 3.98, 0.25 < P < 0.50). Clethriomonys gapperi, which is not common in oak forests of south-central Pennsylvania and which evinces positive population responses to clearcutting (Kirkland, 1990), was not equally abundant in the five habitats (x^ = 45.09, P < 0.001), with 40 of 46 (87.0%) specimens captured in the two clearcut habitats. We did not capture any eastern chipmunks {Tamias striatus) during our initial sampling. Subsequently we have captured only five individuals (all juveniles) in approximately 13,000 TN of pitfall sampling effort in suitable habitats for this species. TTiis suggests that adult chipmunks are not subject to capture by our method of pitfall trapping, perhaps because they can escape from traps before drowning. No sampling protocol will be optimal for all taxa of small mammals. For example, although our protocol is particularly effective in capturing soricids, it is less efficient in taking small muroid rodents and is inefficient in capturing specimens weighing more than 60 g. One way to address these deficiencies would be to augment pitfall arrays with snap traps, which are more efficient in capturing rodents. Influence of Precipitation on Perceptions of Small Mammal Abundance and Community Structure Preliminary data confirmed the significant influence of precipitation on the overall capture rate of small mammals, and the differential responses of shrews and rodents to precipitation (Fig. 2). During periods of no rain, the average daily capture rate for small mammals was 7.4/100 TN, but when it rained (>0.3 cm), the rate increased to 18.9/100 TN (x^ = 76.79, P < 0.001). This increase was different for shrews and rodents. On nonrainy nights, capture rates for shrews and rodents (3.8/KX) TN and 3.7/100 TN, respectively) did not differ. However, when it rained, capture rates increased significantly to 13.4/100 TN for shrews (x^ -- 95.17, P < 0.001) and 5.4/100 TN for rodents (x^ = 4.12, P < 0.05). Although the capture rates for shrews and rodents did not differ on nonrainy nights, the capture rate for shrews was significantly higher when it rained (x^ = 34.63, P < 0.001). These results indicate that perception of relative abundance and community structure of small mammals can be significantly influenced by precipitation during sampling. Consequently, we believe that it is important to include at least one episode of rainfall (>0.3 cm) in each sampling period. Ideally, this precipitation should occur during the first hours after sunset when small mammals normally are most active. Problems; Rocky Substrates and Traps Filling with Water Use of pitfalls and drift fences may be limited by substrate conditions. For example, Pucek (1981) noted that pitfalls are “impossible to use in rock, gravel, or clay.” Nevertheless, Brown (1967) successfully employed pitfalls to sample soricid corrununities in rockslide habitats in the central Rocky Mountains (USA). However, he used single pitfalls, and thus was not constrained by the limitations imposed by arrays of pitfalls with drift fences. Pitfalls are also difficult to install in habitats with high water tables. At such sites, it may be necessary to anchor pitfalls to keep them from being forced out of the ground by subsurface water pressure. Metal or plastic tent stakes, or stakes cut from shrubs or trees hooked over the rims of pitfalls work well to secure pitfalls in the ground. Frequent, heavy rains will cause pitfalls to fill with water, which must be removed. One way to prevent this is to shelter the pitfalls (Handley and Vam, 1994). Another is to put holes in the walls of the pitfalls at the desired maximum depth of water. However, in saturated soil (e.g., bogs), such holes often will cause the trap to completely fill with water. Capture of Amphibians, Reptiles, and Invertebrates Drift fences with pitfalls have been used for many years to collect amphibians and reptiles (Gibbons and Semlitsch, 1981; Campbell and Christman, 1982). In fact, the inspiration for development of our pitfall arrays was the success of herpetologists at Carnegie Museum of Natural History’s Powdermill Biological Station in capturing large numbers of shrews incidental to sampling amphibians with pitfalls and drift fences (J. F. Merritt, personal communication). Because amphibians may serve as both prey for and competitors with shrews, we have preserved amphibians captured during pitfall sampling. During the initial 39 days of sampling in 1987, we collected 424 amphibians representing nine species. If removed daily, amphibians are in good shape and suitable for fluid preservation. Substantial numbers of epigeal invertebrates are also collected in pitfalls. Such material has been used to assess the food resource base of soricids (Pemetta, 1976; Churchfield, 1982). We collect invertebrates at the end of each ten-day sampling period by straining pitfall water through a sieve and then preserving them in 10% formalin or 70% ethanol. In summer, warm water may cause some deterioration of invertebrates during a ten-day sampling period, but this is not sufficient to preclude identification to family. Annotated List of Recommended Supplies AND Equipment Drift Fencing. — A variety of materials, including aluminum flashing, vinyl plastic, tarpaper, corrugated metal, and fiberglass panels, can be used for drift fencing, with factors such as price and local availability determining which material is selected. For example, in Czechoslovakia J. Gaisler (personal communication) uses 1 nun-thick plastic sheeting. This material is used by utility companies to cover buried electrical cables and by construction crews as barriers around excavations. Handley and Vam (1994) use salvaged pieces of aluminum siding for drift fencing. Whatever material is used, it is important that drift fencing be high enough to prevent mammals 280 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 from climbing or jumping over, and that it be buried to a sufficient depth to prevent them from crawling or tunneling under. We recommend using material that is at least 25 cm wide. Nine 5-m sections are needed for each standard array (see Fig. 1). We initially chose aluminum flashing, which although very durable, was expensive (approximately $1.95 per meter of 25.4 cm-wide material or nearly $90 per array in September 1987), somewhat bulky, and did not conform to irregularities in the ground surface. The latter problem was overcome by extensive trenching. The high cost of aluminum flashing also makes it attractive to thieves. More recently we have used heavy-duty (6 mil-thick) vinyl plastic sheeting as drift fencing. This material is sold in 30.5 m (100 ft) by 30 cm (12 in) wide rolls for approximately $10 per roll, making it considerably less expensive than aluminum flashing. Vinyl plastic is substantially less bulky than the aluminum flashing, is flexible so that it easily conforms to undulating topography, and is therefore considerably easier to install. We cut the vinyl plastic sheeting into lengths slightly longer than 5 m (5.3 m works well). The extra length permits some flexibility in avoiding obstacles while still maintaining the 5- m interval between pitfalls. Any excess is wound on the end stakes for greater stability. We use a staple gun to attach the plastic sheeting to wooden stakes for vertical support. The lower 5-8 cm should not be stapled so that it can be rolled to line the trench. Once inserted in a trench, the plastic is held in place by wooden blocks and nails (see below). Soil is then back-filled into the trench to provide greater stability and to prevent small mammals from crawling or turmeling underneath. Pitfalls. — Several types of containers may be used as pitfalls; however, these should be at least 14 cm in diameter and 19.5 cm deep. Plastic containers used to transport bulk dairy products (e.g., cottage cheese) are relatively inexpensive and often come with snap-on lids, which are useful in covering traps securely during nonsampling periods. Such plastic containers are slightly tapered so they stack and thus are convenient to carry. In contrast, #10 tin cans, which are approximately the same size, do not stack, lack convenient lids, and unlike plastic containers, do not deform to accommodate roots and rocks. Handley and Vam (1994) used 2-L plastic soda bottles with the tops cut off. European researchers have traditionally employed cones as pitfalls (Pankakoski, 19796). These vary in size but generally are about 14 cm in diameter and 38 cm deep. Staple Gun. — Needed to secure plastic sheeting to stakes. We have found that 10 mm (% in) staples are an optimal length for attaching 6-mil plastic sheeting. Blocks of Wood and Nails. — These are essential for securing plastic sheeting in trenches. At least five to eight pieces of wood (1.9 cm X 1.9 cm X 15.3-20.3 cm or ^4 in by % in by 6- 8 in) are needed for each 5-m section of drift fence to secure vinyl plastic fencing. Each block has a single hole drilled in the middle to accommodate snugly a large nail or spike which is driven into the ground through the bottom of the vinyl plastic. We recommend 15.3 cm (6 in) nails. Stakes. — Stakes are essential to the installation of plastic sheeting drift fences since they provide vertical support. They are also needed to stabilize aluminum flashing. At sites where wind is expected (e.g., recent clearcuts, grasslands, or deserts), a greater number of stakes will be required to stabilize drift fencing. Gibbons and Semlitsch (1981) recommended securing aluminum flashing to stakes with plastic cable ties. Narrow-Bladed Shovels or Drain Spades. — Drain spades are narrow shovels with rounded blades. The cross section of the drain spade approximates the curvature of the pitfalls recommended. As a result, four cuts, one on each side, will loosen a plug of soil which approximates the shape and volume of a pitfall. Wrecking/Pry Bars or Rock Picks. — Small wrecking/pry bars (46-6 1 cm long) are useful for prying loose rocks from pitfall excavations and for excavating drift fence trenches. Geologist’s rock picks can substitute for the wrecking/pry bars in excavating trenches for drift fences. Conclusions Our purpose in proposing a standard protocol for sampling small mammals is to foster research on small mammal community ecology and to facilitate comparison of data. The proposed protocol is designed to minimize a number of sources of error and variation that have made valid comparisons of such data difficult. We recognize that no sampling protocol is perfect. Each has strengths and weaknesses, and there may be conditions under which our proposed sampling protocol is either inappropriate or impractical. Nevertheless, we believe that if our proposed sampling protocol is adopted by a modest number of researchers, the resulting pool of data will lead to a significant enhancement of understanding of the structure of small mammal communities. Because of the efficiency of pitfalls with drift fences in capturing soricids, adoption of our protocol will yield better insight into the abundance and diversity of shrews in small mammal communities. Acknowledgments We thank many undergraduate and graduate students in biology at Shippensburg University, C. A. Klinedinst, and C. J. and S. G. Kirkland for assistance with the field testing. We also thank participants of the International Colloquium on the Biology of the Soricidae for useful comments, and the following individuals for information: Z. Pucek (Mammals Research Institute, Polish Academy of Science), T. Ireland (Wildlife Science Department, New Mexico State University), J. Y unger (Department of Biological Sciences, Northern Illinois University), and F. J. Dirrigl, Jr. (Natural Heritage Program, New Jersey Department of Environmental Protection). We thank T. French, J. Zejda, and R. Rose, who provided suggestions for revision of the manuscript. Literature Cited Blair, W. F. 1951. Population structure, social behavior, and environmental relations in a natural population of the beach mouse (Peromyscus polionotus leucocephalus). Contributions from the Laboratory of Vertebrate Biology, University of Michigan, 48:1-47. 1994 KIRKLAND AND SHEPPARD— Proposed Standard Sampling Protocol 281 Bowers, M. A. 1988. Seed removal experiments on desert rodents: The microhabitat by moonlight effect. Journal of Mammalogy, 69:201-204. . 1990. Exploitation of seed resources by Merriam’s kangaroo rat: Harvesting rates and predatory risk. Ecology, 71:2334-2344. Brown, L. N. 1967. Ecological distribution of six species of shrews and comparison of sampling methods in the central Rocky Mountains. Journal of Mammalogy, 48:617-623. Burt, W. H. 1940. Territorial behavior and populations of some small mammals in southern Michigan. Miscellaneous Publications of the Museum of Zoology, University of Michigan, 45:1-58. Campbell, H. W., and S. P. Christman. 1982. Field techniques for herpeto fauna community analysis. Pp. 193-200, in Herpetological Communities: A Symposium of the Society for the Study of Amphibians and Reptiles and the Herpetologists’ League, August 1977 (N. J. Scott, Jr., ed.), U. S. Department of the Interior, Fish and Wildlife Service, Wildlife Research Report 13, Washington, D. C. 239 pp. Churchfield, S. 1982. Food availability and the diet of the common shrew, Sorex araneus. Journal of Animal Ecology, 51:15-28. COCKRUM, E. L. 1947. Effectiveness of live versus snap traps. Journal of Mammalogy, 28:186. Cornbower, T. R., and G. L. Kirkland, Jr. 1983. Comparisons of pine vole {Pity my s pinetorum) populations from orchards and natural habitats in southcentral Pennsylvania. Proceedings of the Pennsylvania Academy of Science, 57:147-154. Drickamer, L. C., and M. R. Capone. 1977. Weather parameters, trapability and niche separation in two sympatric species of Peromyscus. The American Midland Naturalist, 98:376-381. Falls, J. E. 1968. Activity. Pp. 543-570, in Biology of Peromyscus (Rodentia) (J. A. King, ed.), Special Publication, The American Society of Mammalogists, 2:1-593. Gibbons, J. W., and R. D. Semlitsch. 1981. Terrestrial drift fences with pitfall traps: An effective technique for quantitative sampling of small mammal populations. Brimleyana, 7:1-17. Graves, S. J., J. E. Maldonado, and J. O. Wolff. 1988. Use of ground and arboreal microhabitats by Peromyscus leucopus and Peromyscus maniculatus. Canadian Journal of Zoology, 66:277-278. Handley, C. O., Jr., and M. Yarn. 1994. The trapline concept as applied to pitfall arrays. Pp. 285-287, in Advances in the Biology of Shrews (J. F. Merritt, G. L. Kirkland, Jr., and R. K. Rose, eds.), Carnegie Museum of Natural History Special Publication no. 18, + 458 pp. Kirkland, G. L., Jr. 1979. Initial responses of small mammals to clearcutting of Pennsylvania hardwood forests. Proceedings of the Pennsylvania Academy of Science, 52:21-23. 1990. Patterns of initial small mammal community change after clearcutting of temperate North American forests. Oikos, 59:313-320. 1991. Competition and coexistence in shrews (Insectivora: Soricidae). Pp. 15-22, in The Biology of the Soricidae (J. S. Findley and T. L. Yates, eds.). Special Publication, The Museum of Southwestern Biology, University of New Mexico, 1:1-91. Kirkland, G. L., Jr., and J. E. Hench. 1980. Notes on the small mammals of the Carbaugh Run Natural Area, Adams Co., Pennsylvania. Proceedings of the Pennsylvania Academy of Science, 54:31-35. Kirkland, G. L., Jr., A. M. Wilkinson, J. V. Planz, and J. E. Maldonado. 1987. Sorex (Microsorex) hoyi in Pennsylvania. Journal of Mammalogy, 68:384-387. MacLeod, C. F., and J. L. Lethiecq. 1963. A comparison of two trapping procedures for Sorex cinereus. Journal of Mammalogy, 44:277-278. Mengak, M. T., and D. C. Guynn, Jr. 1987. Pitfalls and snap traps for sampling small mammals and herpeto fauna. The American Midland Naturalist, 118:284-288. Mystkowska, E. T., and j. Sidorowicz. 1961. Influence of weather on captures of Micromammalia. II. Insectivora. Acta Theriologica, 5:263-273. Pankakoski, E. 1919a. The influence of weather on the activity of the common shrew. Acta Theriologica, 24:522-526. 1919b. The cone trap — A useful tool for index trapping of small mammals. Annales Zoologica Fennica, 16:144-150. Pern ETTA , J. C. 1976. Diets of the shrews Sorex araneus L. and Sorex minutus L. in Wytham grassland. The Journal of Animal Ecology, 45:899-912. Pizzimenti, j. j. 1979. The relative effectiveness of three trap types for small mammals in some Peruvian rodent communities. Acta Theriologica, 25:351-361. Planz, J. V. 1987. Activity patterns in the white-footed mouse (Peromyscus leucopus) — influence of woody litter. Unpublished master’s thesis, Shippensburg University, Shippensburg, Pennsylvania. 37 pp. Prince, L. A. 1941. Water traps capture the pygmy shrew (Microsorex hoyi) in abundance. The Canadian Field-Naturalist, 55:72. PUCEK, Z. 1969. Trap response and estimation of numbers of shrews in removal catches. Acta Theriologica, 14:403-426. 1981. Keys to the vertebrates of Poland — Mammals. Polish Scientific Publishers, Warsaw, 367 pp. (translated from Polish). Sidorowicz, J. 1960. Influence of weather on capture of Micromammalia. I. Rodents (Rodentia). Acta Theriologica, 4:139-158. West, S. D. 1985. Differential capture between old and new models of the Museum Special snap trap. Journal of Mammalogy, 66:798-800. Wiener, J. C., and M. H. Smith. 1972. Relative efficiencies of four small mammal traps. Journal of Mammalogy, 53:868-873. Williams, D. F., and S. E. Braun. 1983. Comparison of pitfall and conventional traps for sampling small mammal populations. The Journal of Wildlife Management, 47:841-845. Wolfe, J. L., and R. J. Esher. 1981. Relative abundance of the southeastern shrew. Journal of Mammalogy, 62:649-650. 282 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 Table \ .—Comparisons of results from sampling of small mammal communities in forested and clearcut habitats with arrays of pitfalls and drift fences at ten localities in south-central Pennsylvania and with Museum Special snap traps at 18 localities in Pennsylvania. * P < 0.001. Community Characteristic Pitfalls Snap Traps z-Value % Soricids 58.1 16.6 14.616* Blarina brevicauda as % of total soricids 28.8 77.9 -9.048* Blarina brevicauda as % of total small mammals 16.7 13.0 -1.704* Sorex spp. as % of total soricids 71.2 22.1 9.048* Sorex spp. as % of total small mammals 41.4 3.7 16.832* Peromyscus leucopus as % of total small mammals 24.1 47.6 -7.677* Peromyscus leucopus as % of rodents 57.4 57.1 0.077 Clethrionomys gapperi as % of total small mammals 12.4 30.2 -6.607* Clethrionomys gapperi as % of rodents 29.7 36.2 -1.535 Total Sampling Effort (TN) 3900 6748 Total Catch 370 841 Catch/ 100 TN 9.49 12.46 Table 2.— Ecological distribution and community structure of small mammals in five habitat types on South Mountain, Cumberland County, Pennsylvania, as revealed by 39 nights of sampling with ten arrays of pitfalls and drift fences (two per habitat) between 22 September and 21 November 1987. Distribution of Small Mammals by Habitat Type Mature Lowland Nine- Mid- Three- Deciduous Year-Old Altitude Year-Old Ridge Species Forest Clearcut Oak Forest Clearcut Forest Blarina brevicauda 17 25 5 9 6 Sorex fontinalis 21 24 15 27 12 S. fumeus 6 9 0 2 0 S. hoyi 6 9 8 10 4 Clethrionomys gapperi 4 17 2 23 0 Peromyscus leucopus 18 15 13 19 24 Microtus pinetorum 8 3 3 3 0 Microtus pennsylvanicus 0 1 0 2 0 Number of species 7 8 6 8 4 Number of specimens 80 103 46 95 46 1994 KIRXLAND AND SHEPPARD — Proposed Standard Sampling Protocol 283 Fig. 1. — Schematic representation of arrangement of pitfalls and drift fences in proposed standard sampling protocol. Fig. 2.— Pattern of captures of soricids and rodents during 39 nights of sampling with ten pitfall arrays between 22 September and 21 November 1987 in Michaux State Forest, South Mountain, Cumberland County, Pennsylvania. Precipitation indicated by horizontal bars. .r 1 “ >f. » ■ 'S'JKsr'' ~f»- < Ika^'^ro ■v^;:'; vJi_ v9' * .» A ‘ *’ • ^ ■ ■’^ • , in.. * 'l-i.'^.lO'*'' i>:a;4S*, , ,, i./.ls""-: i. i \l I ' - * 1; ''V. * fhf- 1 r ■ r\ ilf i-w Ai 0m „^|*>''|J— ll:-; 1 K ■ :*ct _ ‘^-.i! ...f. ■ I 'I «r- 1 ( . > ac. ' i ' I I I'H ^■: 'I. •* f y. -t ' I'f -T'." 7 “"“ r ^ . » 4- -Vf * i , T- ty- ' 4.«:'0fJ*-'l- ,.iUK)i' kn<<'< W i' ^.*1^ k:iacshci4',’it'I THE TRAPLINE CONCEPT APPLIED TO PITFALL ARRAYS Charles O. Handley, Jr.’ and Merrill Yarn’ 'Division of Mammals, National Museum of Natural History, Smithsonian Institution, Washington, D.C. 20560 Abstract We describe a small and easily set pitfall array for capturing and preserving shrews, other small vertebrates, and invertebrates. It is designed to be set in a transect. Because this array can be set quickly it is practical to set many, simulating a trapline. Like a trapline, a transect of pitfall arrays can easily and quickly sample all of the habitat types and give a quick estimate of diversity of species in an area. Introduction During 14 months of sampling the distribution, abundance, and habitat preferences of shrews in coastal South Carolina (July 1989 to September 1990), we successfully used a very compact pitfall array. The whole array fits into a triangle with 2.5-m sides. This small array is inexpensive, seldom vandalized (constructed entirely of salvaged, discarded, rigid materials), lightweight, and conveniently transportable. It can be set quickly (two people can set two arrays per hour in any but the rockiest soil), even in rough terrain and amidst obstacles, and it causes minimal disturbance to habitat. Because of their small size, our arrays even survived the calamity of falling trees with little or no damage when Hurricane Hugo passed through our study area, leveling the forest on 21 September 1989. Arranged in a transect, with 14 arrays (98 pitfalls) per 1000 m, our arrays are equivalent to a trapline, and are particularly good for quick estimates of habitat selection and presence or absence of species of shrews. Materials and Methods Our pitfalls were 2-liter, heavy-gauge plastic soft drink bottles with the tops cut off. Bottles with reinforced round bottoms, rather than those with rippled bottoms, maintained shape better in the ground, and removal of specimens from them was easier. The topless plastic bottles are 20 cm deep and 11 cm in diameter. At the center of each array we also used one 4-liter plastic container approximately 18 cm deep and 15 cm in diameter. Equivalent-sized metal cans might be used for short-term sampling, but they are not good for long-term projects because they rust. Also, their flat bottoms make removal of specimens more difficult. Containers were placed in arrays of seven pitfalls each (Fig. 1). The pitfalls were arranged in a three-leaf clover pattern (120° between arms), with the 4-liter container at the center and the 2-liter bottles on either side near the distal end of each arm (drift fence). The drift fences converged at the central pitfall. The fences were made of vinyl or aluminum siding, 1.2 m long by 30 cm high, inserted into a trench 2-5 cm deep to prevent burrowing through litter beneath the fence, and held upright by sticks cut from the forest. An array fits into a triangle a little less than 2.5 m from comer to comer. Containers were sunk into the ground with lips flush with the surface. Each pitfall was sheltered from rain, falling leaves, sun, and moonlight with a square of vinyl or aluminum siding 30 X 30 cm, leaned over the pitfall, against the drift fence. A larger square of plastic or aluminum, laid flat on the convergent drift fences, sheltered the central pitfall. In damp ground on swamp borders, we compensated for fluctuating water tables by pegging down the pitfalls. Otherwise, they popped out of the ground. We made hooked pegs from slender sprouts. One long peg on either side of a pitfall effectively held it in place, even when it was totally submerged. If preservative is not used pitfalls can be kept in wet ground by puncturing them so they will fill to the level of the water table. We filled the pitfalls to about half their depth with 10% formalin to preserve specimens. During sampling, we checked the pitfall traplines at 4-6 week intervals and had little trouble with specimens spoiling except occasionally in flooded pitfalls. At first, we put a skim of mineral oil on the formalin to reduce evaporation and odor. However, we soon eliminated the oil because it was expensive, coated the specimens, and evaporation proved not to be a problem in the humid climate of coastal South Carolina. In spite of the increased odor of the formalin, catch numbers did not decline. If arrays are checked daily (preferably morning and evening), water can be substituted for formalin. Shrews also can be captured in dry pitfalls, but in experiments in 1987 at Mountain Lake Biological Station in southwestern Virginia, we observed a higher capture rate in pitfalls containing liquid, even if no more than a few millimeters deep. Rodents, with the exception of baby voles, can easily jump out of dry pitfalls. We marked arrays with pink surveyor’s tape (highest visibility in forest). To deter vandalism we included with the array marker a waterproof label which stated, “THIS IS A PROJECT OF THE SMITHSONIAN INSTITUTION STUDYING POPULATIONS OF SMALL ANIMALS. FOR MORE INFORMATION PHONE (a number).” A local cooperator responded to telephone queries by reading a brief explanation of the project. Equipment needed to install the pitfalls and drift fences is minimal. However, two items not ordinarily found in the home or laboratory greatly simplify the job of installation: 1) drain spade with a 5 X 16-in (12.5 X 40-cm) blade, the edge sharpened for cutting roots in the pitfall hole; and 2) posthole digger with a lift diameter of 5 in (12.5 cm), only 1.5 cm greater in diameter than a 2-liter container. This tool expedites removal of soil after the hole has been cut with the drain spade. These tools may be obtained from a large hardware store or 285 286 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 from Forestry Suppliers, Inc. (Box 8397, Jackson, Mississippi 39284-8397). Other equipment will vary with the preference of individual pitfall trappers. We recommend: mattock with combination adz-shaped and ax blades (essential for digging in hard or rocky ground), trowel, pruning shears, gloves (in the South soil often contains roots of poison ivy, Rhus radicans), l-m“ plastic sheet for removing waste earth from pitfall holes, an extra 2-liter bottle for plugging pitfalls during installation to keep soil out, and a 5-gal (20-liter) plastic container for dispensing formalin into the pitfalls. Gear for collecting specimens from pitfalls varies with the preference of individual trappers. We used a bucket, 14-in diameter x 8 in deep (35 x 20 cm), for carrying specimens and supplies; plastic bags, '/^-gal, 5 X 15 in (2-liter, 12.5 X 37.5 cm); waterproof labels; #2 pencil or pen with waterproof permanent ink; latex gloves; containers of formalin and water; '4 -cup measure; trowel; and pruning shears. Wearing latex gloves, we scooped specimens from pitfalls by hand into a plastic collecting bag, prelabeled with the array number. Then we cleaned out leaves and debris, added water if necessary, and restored the strength of the preservative in the pitfall by adding 'A cup of 37 % formaldehyde solution. Results Where we used the 2.5-m arrays in coastal South Carolina, 40-60 km NE of Charleston, for 14 months in 1989 and 1990, flooding often drastically reduced catch during the 4-6 weeks between pitfall checks. Likewise, invertebrates in the pitfalls were a deterrent to capture of shrews. The mammals could use the invertebrates as a platform and scramble out of the pitfalls. Nevertheless, in 22 pitfall arrays we caught shrews: 53 Sorex longirostris, 32 Blarina carolinensis, and 1 1 Cryptotis parva, and a few rodents including: 7 Reithrodontomys humulis, 2 Oryzomys palustris, 1 Ochrotomys nuttalU, and 1 Microtus pinetorum. The miniature arrays were even more effective for capturing amphibians and reptiles. They caught 575 frogs and toads, 190 salamanders, 26 lizards, and 6 small snakes. The major fraction of the pitfall catch, however, was composed of a broad spectrum of invertebrates, including snails, earthworms, nematodes, centipedes, millipedes, spiders, harvestmen, crayfish, crabs, amphipods, isopods, Coleoptera, Hymenoptera, Orthoptera, Diptera, and even Lepidoptera. Except for Hymenoptera, the insects included both adult and larval stages. Sometimes pitfalls were jammed full to the top with a multitude of one kind of invertebrate such as crayfish, isopod, or beetle. In using pitfalls for invertebrates, entomologists at the Smithsonian Institution prefer as a preservative a solution of 50% antifreeze (ethylene glycol) and 50% water containing detergent or soap, or a solution of 50% alcohol and 50% water containing detergent or soap. For high- quality specimens, pitfalls filled with these weak preservative solutions must be checked at intervals of less than a week. There is a considerable amount of literature on pitfall trapping. Because of the scale of the pitfalls described, two papers are particularly appropriate to our discussion. Bury and Com (1987) concluded that pitfall arrays with short drift fences (2.5 m) are no more effective for capturing small mammals than pitfalls without fences. However, they did not distinguish shrews from other small mammals in their comparison. Small scale pitfalls are designed primarily for catching shrews and are relatively ineffective for other small mammals. Furthermore, in the array that Bury and Cora described was a 10-m gap between the fences at their inner ends which largely nullified the value of the drift fences. The resulting array differed little from an array of pitfalls without fences. Williams and Braun (1983) compared single pitfalls with 1.2 m- and 0.6 m-long drift fences radiating from them, with single pitfalls without drift fences. A pitfall with 1.2-m fences caught 2.5 times as many shrews as a pitfall without fences, and 3.3 times as many as a pitfall with 0.6-m fences. The use of the pitfall as a sampling tool has been developed, refined, and tested so thoroughly that it is desirable and feasible to propose standards (Kirkland and Sheppard, 1994). We do not pretend that the method we describe here is either new, refined, or tested as much as the larger pitfall arrays. However, it does catch shrews, and it offers a person who works alone or with limited resources a method of sampling shrews that is not labor- intensive and doesn’t require as much space as a large array. Acknowledgments We are grateful to G. Stapleton, District Ranger, and D. Carlson, Biologist, Wambaw Ranger District, Francis Marion National Forest, for permission to work in the National Forest and for helping us locate areas suitable for pitfall transects. We thank J. Cely, South Carolina Wildlife and Marine Resources Department, for collecting permits. Funds from the Smithsonian Office of Product Development and Licensing, L. Stevenson, Director, made frequent travel to South Carolina possible. Literature Cited Bury, R. B., and P. S. Corn. 1987. Evaluation of pitfall trapping in northwestern forests: Trap arrays with drift fences. The Journal of Wildlife Management, 51:112-119. Kirkland, G. L., Jr., and P. K. Sheppard. 1994. Proposed standard protocol for sampling small mammal communities. Pp. 277-283, in Advances in the Biology of Shrews (J. F. Merritt, G. L. Kirkland, Jr., and R. K. Rose, eds.), Carnegie Museum of Natural History Special Publication no. 18, x + 458 pp. Williams, D. F., and S. E. Braun. 1983. Comparison of pitfall and conventional traps for sampling small mammal populations. The Journal of Wildlife Management, 47:841-845. 1994 HANDLEY AND YARN— Pitfall Traplines 287 Fig. 1.— Plan for 2.5-m pitfall array, with a 4-liter container in the center and pairs of 2-liter pitfalls near the outer ends of 1.2 m drift fences. Not drawn to scale. f > j - Mi ■ i • V- . *'1 i*'.. S' ■■ 'v'ndi • - ., , ,-Xv »!''■>.' i>f 'IjfitjSfrfar:; ■' ' r iirw-A.&s&f, H •, ■ .. \V ■ 'iCYVi)) iPM^h ■■ . v,;?:.vi... , it>v. V it='t ^ . ) «>;iK ^i^..Jf•^^1n,?^■-^»•i^■tCw:- ' ■ ■ ,;^T)'v I, •.‘'»tfjJuufc* ,;hyr:'^> ^.i-' , ., -';A»rtlilSi»i‘. FUf^!'W4^S(V '■ ■„■ '.■•■ •;’ • sfrkC. k ij{|,.* lVuU%u • ir ■ , j.^.y iilifrrr'ijc!; ^n’i" jf-T-iM, ' ■" -.tu*. ly ',.s..' fh >u CfeljpIlV*. .. »' '■ !rV-?^'7.rt,,. ■ I t'V' •Ml ••( .'■ "/I . hi'rtjy v-v'>A'itkf Hl'i' \f .R?5t lytiiv .1 . •• .u'rtxy ; ivu |i i'lj'.s (■:P''.r?« i)rt4 R- . ,. I<> mv- :v' '; :'i r't ffitJ , H-" :. ■ M* ' . '■ r>',il I’,' sr'V' “ *-■"•'♦ Sil^sE ALBUMIN EVOLUTION IN THE SORICINAE AND ITS IMPLICATIONS FOR THE PHYLOGENETIC HISTORY OF THE SORICIDAE Sarah B. George’ and Vincent M. Sarich- ’ Section of Mammalogy, Natural History Museum of Los Angeles County, 900 Exposition Boulevard, Los Angeles, California 90007; current address: Utah Museum of Natural History, University of Utah, Salt Lake City, Utah 84112; ^ Department of Anthropology, University of California, Berkeley, California 94720 Abstract Results of albumin immunodistance analyses suggest that the Nearctic Otisorex and Sorex trowbruJgii are sister taxa, and Palearctic Sorex are more distantly related. The genera Notiosorex and Neomys are equidistant from the Soricini/Blarinini cluster. Shrews are monophyletic with respect to other Insectivora, although the Crocidurinae and Soricinae diverged soon after shrews diverged from moles. Divergence times estimated from the albumin immunodistance data are considerably older than those estimated from the fossil record, suggesting an Eocene rather than Miocene divergence of crocidurine and soricine shrews. Shrews, moles, tenrecs/golden moles/hedgehogs , and solenodons represent four relatively equidistant lineages of insectivores that diverged sometime in the late Paleocene or early Eocene. Introduction Living shrews of the family Soricidae are Holarctic and African in distribution, with a few species found in the northern Neotropics and the Orient. Because of their small size and secretive habits, few specimens existed in museums and little phylogenetic work was attempted until very recently. In the last 15 years or so, a virtual explosion of systematic studies has been published on shrews (a summary may be found in the introduction of George [1988] and additional citations in this volume). These new studies, based on allozyme and karyotypic data, reexaminations of fossil series, and morphometric analyses of osteological data, sometimes have yielded conflicting results. Herein we review the phylogenetic conclusions of previous studies, present new results from albumin immunodistance data, and attempt to estimate divergence times for lineages of shrews from these data. Using albumin immunodistance data, we address four issues in shrew systematics. First, we examine conflicting phylogenetic trees proposed for intrageneric relationships within the genus Sorex based on allozymes and morphology, and compare them to the tree generated from the immunodistance data. Allozyme data divide the genus into three major lineages comprising the following species groups (George, 1988): an unnamed subgenus including Sorex trowbridgii, S. merriami, and S. arizonae; the subgenus Sorex including Palearctic long- tailed shrews and S. arcticus and S. tundrensis from the Nearctic; and the subgenus Otisorex which consists solely of Nearctic species. These results conflict with previous classifications in several ways, most notably in that the Sorex trowbridgii clade had been classified within the subgenus Sorex with the Palearctic shrews (Findley, 1955). Second, we address the question of the phylogeny of tribes within the Soricinae. Allozyme results classified the tribe Neomyini as the sister taxon to the tribes Soricini and Blarinini. However, divergence dates for these taxa could not be calculated from the allozyme data because the data indicated that the genus Sorex has a more rapid rate of protein evolution than the other lineages, and the calculation of dates of divergence from genetic distances requires the assumption, or more preferably, the demonstration of roughly equal rates of change among the constituent lineages. Third, we compare the date of divergence of the Crocidurinae and Soricinae calculated from the immunodistance data to the date estimated from the fossil record. Finally, we examine evolutionary relationships of soricids to other insectivores and mammals, and compare the phylogeny based on i mmunod i stance data to other proposed higher-level mammalian phylogenies. Methodology Following methods described by Sarich (1969) and Maxson and Maxson (1990), antisera to Sorex vagrans, Blarina brevicauda, Neomys fodiens, and Suncus murinus albumins were generated in rabbits (Oryctolagus cunicularis). For family-level studies, previously prepared antisera to the albumins of Mogera wogura, Scapanus townsendii, Condylura cristata (moles, family Talpidae), Solenodon paradoxus (solenodon, family Solenodontidae), Erinaceus europaeus (hedgehog, family Erinaceidae), Eremitalpa granti (golden mole, family Chry sochloridae) , and Hemicentetes semispinosus (tenrec, family Tenrecidae) were used. Antisera from a large number of noninsectivores were also available. Additional antigens used were Sorex palustris, S. trowbridgii, S. araneus, S. caecutiens, Cryptotis parva, and Notiosorex crawfordii. With the exception of a few quantitative precipitin experiments to be discussed below, all the albumin comparisons were carried out by immunodiffusion. This apparently retrogressive decision requires some justification. The several methods (immunodiffusion, quantitative precipitin, microcomplement fixation, radioimmunoassay) that have been used to measure and compare immunological distances among mammalian taxa have long been known to give highly correlated results. In addition, there is much less variation in precision and ultimate phylogenetic resolving power among them than has been thought to be the case. This was demonstrated by the doctoral research of Elizabeth Pierson on higher-taxon (family and above) relationships among bats using immunological comparisons of their transferrins (Pierson, 289 290 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 1986). After preparing antisera, Pierson carried out a series of “preliminary” immunodiffusion comparisons over a few days, then over parts of the next three years she performed micro- complement fixation (MC’F) reactions. The phylogenetic trees resulting from the two data sets were congruent, and little, if any, resolving power was added by the MC’F comparisons. As Pierson wrote (1986:227): While immunodiffusion may not offer the quantitation of MC’F, it nevertheless proved to be an extremely useful tool in this study. One of the most striking advantages of the immunodiffusion technique is best illus- trated by telling a story. Three years ago when I began the analysis of family level relationships in bats, I spent four days running immunodiffusion experiments to get preliminary indications of results I might get from MC’F comparisons. Upon completion of these first experiments, I sketched the phylogeny suggested by the results. Three years, and many, many MC’F experiments later, the PHYLIP Fitch analysis yielded essentially the same tree. The truly robust MC’F results were equally apparent by immunodiffusion; the taxa that were difficult to place by immunodiffusion also caused problems in MC’F. Thus this quick, and very simple method provided, in general, the same branching order among higher taxa as the much more technically demanding and time consuming MC’F. There are additional advantages to the Ouchterlony procedure. First, it has greater phylogenetic scope.... Second, MC’F reactions have a very narrow window.... Third, immunodiffusion is an essentially foolproof technique, and will work with samples, like tissue extracts, that can present anti-complementarity problems with the more exacting MC’F procedure. . . . This is not to say that procedures such as quantitative precipitation, MC’F, and radioimmunoassay do not add resolving power at shorter distances, but we do not believe that they add much at most of the distances with which we are concerned in this study. In addition to these technical advantages, immunodiffusion comparisons have the unique advantage, for molecular data, of being able to be seen through photographs. Thus, it is difficult to consign immunodiffusion data, when presented as actual photographs, to the black box status accorded to immunological data in general. We have added only two technical irmovations to the method as originally presented by Goodman (1960). First, the agar gels contained 5 % polyethylene glycol (PEG, MW approximately 80(X)). This markedly increases the strength of more distant reactions, and adds sharpness to the precipitin lines (W. Rainey, personal communication). Second, we added a fourth component to the usual three, allowing the use of an outgroup in each experiment. For example, to test soricine relationships, we placed, in four wells located clockwise from 12 o’clock, an antiserum to Sorex albumin, Suncus serum, an antiserum to Mogera albumin, and Solenodon serum, at appropriate dilutions. Solenodon reacts more strongly with anix-Mogera than does Suncus; with anti- Sorex, the two sera react about equally well. Going from an anti-mole reference point to an anti-soricine point, the relative strength of the crocidurine reaction increases, indicating a phylogenetic association between the two shrews. What of the fact that Solenodon and Suncus react equally with anti-5orex? This is explained by replacing anti-Mogera in the pattern with antisera to albumins of other orders such as edentates, whereupon Solenodon consistently reacts more strongly than Suncus, indicating less change along its albumin lineage. Such arrangements then provide internal rate tests, and, by extension, the phylogenetic relationships of the taxa compared, given one assumption about where the root of the overall tree lies. The small number of precipitin comparisons (small because of the paucity of shrew material) was conducted using turbidimetric methods. For these comparisons, 0.15 ml of an appropriate antiserum concentration (so that the final optical density at 370 nm and a 1 cm path was < 1.0) was added to each of five tubes containing 0.15 ml of diluted serum (typically, 1:50 in the first tube, 1:100 in the second, and so on). After mixing, the reactions were allowed to equilibrate over 18 hours at 20° C. Isotonic tris-buffered saline (1 ml at pH 7.45) was then added to each tube, the mixture vortexed, and the optical density read in a Zeiss spectrophotometer. Large series of tests have shown that data obtained in this manner are quantitatively equivalent to those obtained using centrifugation and three washings of the precipitate, with final solution in dilute NaOH. The advantage of the turbidimetric procedure lies in its simplicity. Each reduction of I % in amount of turbidity given in the heterologous, as compared to the homologous, reaction is equivalent to two units of immunological distance (AID). Results Within-SoT&x Relationships To test alternative hypotheses of subgeneric relationships within (Findley, 1955; George, 1988), antibodies to vagrans were tested against antigens of four species of long- tailed shrews representing the three subgenera, S. araneus and S. caecutiens (subgenus Sorex), S. palustris (subgenus Ot isorex), and S. trowbridgii (unnamed subgenus). The allozyme data for these five species (using Cryptotis parva and Notiosorex crawfordii as outgroups) have indicated (with two synapomorphies) that Sorex trowbridgii is the sister taxon to the Palearctic Sorex and Nearctic Otisorex (Fig. la, modified from George, 1988). By contrast, the an\\-Sorex vagrans albumin comparisons show a slightly stronger reaction with S. trowbridgii than with the Palearctic Sorex (araneus and caecutiens; Fig. lb). The relationship among the three lineages is probably close to a trichotomy. From the immunodiffusion data, we estimate an AID value for the Sorex vagrans! trowbridgii branch point of +15. Using a calibration of 2.8 AID units per million years (Sarich, 1985), this suggests a divergence time of 5-6 MYBP or latest Miocene. The AID values for vagrans! araneus and vagrans! caecutiens are in the low 20s, which yields an approximate date for this branch point of 8 MYBP or late Miocene. The much higher rate of electrophoretic differentiation within Sorex compared to other soricines (George, 1986, 1988) makes 1994 GEORGE AND SARICH— Soricine Albumin Evolution 291 it difficult to use those data for any divergence time calculations within the genus. However, divergence must have been younger than the 14 MYBP figure obtained using the standard calibration (George, 1988:453). Thus, our molecular estimates for divergence times among the subgenera of Sorex slightly predate the earliest fossils assigned to the genus (5 MYBP, or Miocene/Pliocene boundary; Repenning, 1967; Savage and Russell, 1983). We see no inconsistencies among the three bodies of data, particularly considering that fossil-based dates are necessarily minimum ones, and that the early fossil record of the genus Sorex from the Miocene is scant and would benefit from additional paleontological work, especially in Asia (J. Reumer, personal communication). Soricine Tribal Relationships The next question to be addressed is tribal relationships within the Soricinae. Albumin immunodistance data were used to test the topology of the tree constructed from allozyme data (Fig. 2a; modified from George, 1986) and to calculate divergence dates for these lineages. Anti-iSorer albumin gave, using turbidimetric precipitin comparisons, the following distances: Blarina brevicauda, 42; Neomys fodiens, 80; Notiosorex crawfordii, 88; and Suncus murinus, 144. The smaller Neomys distance (relative to Notiosorex) indicates that the SorexIBlarina lineage and Neomys shared a brief period of common ancestry after the divergence of Notiosorex (Fig. 2b). This conclusion is also supported by the fact that Neomys albumin reacts better with znii-Blarina (in immunodiffusion comparisons) than does Notiosorex albumin. With &nii-Neomys , Sorex and Notiosorex are at the same distances, whereas with the anti -edentate/mole outgroup, Notiosorex reacts more strongly than Sorex. The Soricini/Blarinini AID value of 42 corresponds to a divergence time of approximately 15 MYBP, or mid-Miocene. For perspective, this is slightly more recent than the divergence of the great apes and Old World monkeys (Sarich and Cronin, 1976). The AID value for the trichotomy is ±85, which is approximately 30 MYBP or mid-Oligocene. Repenning (1967), using morphological analysis of living and fossil forms, classified soricines into three tribes: Soricini, Blarinini, and Neomyini. The allozyme data support this three- tribe classification (George, 1986). However, Reumer’ s (1984) assessment of the fossil evidence suggested that Notiosorex and Neomys should be classified in separate tribes, the Notiosoricini and Soriculini, respectively. The albumin data are consistent with this latter classification as they place Neomys and Notiosorex about equidistant from one another and from the soricine/blarinine unit. Crociduri nae-Sori cinae Relationsh ips Sorex, Cryptotis, Blarina, Notiosorex, Suncus, and Solenodon albumins were reacted with antisera to the albumins of Sorex, Neomys, Mogera, Hemicentetes, Erinaceus, Bradypus (Xenarthra: Bradypodidae), Tamandua (Xenarthra: Myrmecophagidae), and Cabassous (Xenarthra: Dasypodidae) in a series of 4-well immunodiffusion reactions. The decrease in the relative distance to Suncus when going from nonshrew to shrew comparisons is consistent but not very strongly marked. Thus, the shrews form a monophyletic group, but the distance between the Soricinae and Crocidurinae, relative to those between shrews and nonshrews, is unexpectedly large. Current interpretations of fossil data place the divergence date of crocidurine and soricine shrews in the early Miocene (Reumer, 1984, 1987, 1989), whereas our data suggest that the divergence took place as early as the Eocene. This conclusion derives from two separate lines of evidence. First, the AID of 144 for this divergence point converts to a time of ±50 MYBP. Second, soricine-crocidurine albumin cross reactions are only slightly stronger than those between shrews and moles (reactions among the three major mole albumin lineages are also only slightly stronger than those between shrews and moles; D. W. Moore, V. M. Sarich, and T. L. Yates, personal communication). Reactions between shrews and moles are not discemably stronger than those between them and other “insectivores.” If shrews and moles are to remain monophyletic units, then the crocidurine-soricine divergence must have followed closely in time the divergence of soricids and talpids, and those among the various “insectivoran” lineages. Interfamilial Relationships What about other taxa that have been placed among the insectivores? Most of the pertinent albumin studies have been published, and we briefly review those results. Dermopterans and tupaiids are, in that order, sister groups to the primates (Cronin and Sarich, 1980). Macroscelids align with lagomorphs, and the data indicate that the two groups of elephant shrews, pikas, and rabbits/hares, comprise four lineages that are essentially equidistant from one another (Sarich, 1985). What are the relationships among the remaining insectivoran taxa? A few conclusions may be drawn from existing data (see previous section for a list of sera and anti-sera compared). Golden moles, represented by Eremitalpa, join strongly with tenrecs (Hemicentetes), and the resulting clade then joins with hedgehogs (Erinaceus). These three taxa may be slightly closer to Solenodon than to the shrews and moles, but it is a weak grouping, based on the reactions observed. Conservatively, with dermopterans and tupaiids placed with primates and macroscelids with lagomorphs, four relatively equidistant “insectivoran” lineages remain: shrews; moles; Solenodon', and tenrecs/golden moles/hedgehogs (Sarich, 1993). These four lineages are no closer to one another than they are to the other probable members of one of the two major groups of mammals: lagomorphs/macroscelids, primates, bats, rodents and camivores/pangolins (ungulates and subungulates making up the other major group of mammals). TTiese probably represent old lineages that had begun diverging sometime in the Paleocene, and finished in the early Eocene. This lack of precise resolution might be taken as indicative of a basic weakness of the molecular approach. We believe this is an unrealistic conclusion. Quoting Sarich (1985:433): ...only three extant orders can be shown in the Paleocene, and it is not clear that for any of these (primates, carnivores, rodents) can the Paleocene lineages 292 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 be associated with extant ones. The real beginnings of diversification leading to extant intra-ordinal lineages would then appear to be no earlier than late Paleocene and early Eocene, and indeed most orders do not even appear in the fossil record until the Eocene If the picture just sketched is in fact realistic, it also goes a long way toward answering the question of why it has proved so difficult to determine interordinal affinities among the placentals — to, for example, determine the sister group to rodents. If we are now only allowing some 10 MY to proceed from the adaptive radiation which gave rise to placental orders to the late Paleocene to early Eocene interordinal radiations, then any periods of common ancestry among extant orders are going to be rather brief indeed, and one might argue that we are fortunate to be able to see any, and not worry about our inability to see more. This may be one of those questions that doesn’t have accessible answers; the answers (shared, derived characters linking orders) having been destroyed or distorted beyond recognition over time. A quantitative assessment of this point was made by Nei (1986:134-139), wherein he calculated how much DNA sequence was necessary to resolve, to a given level of confidence, a lineage of length PMY at a time depth of ZMY. For F = 5 and Z = 50 (time spans appropriate to the relationships involved here), 4,100 base pairs of DNA sequence are needed to achieve resolution at a 95% confidence level. In other molecular studies of higher-order relationships within the Mammalia, few authors have included more than one lipotyphlan insectivore taxon, and none has looked at as many taxa as we have. Therefore, there are difficulties in developing comparisons of our results with those of others. Amino acid sequence data (mainly globins) have been used to argue for a soricid/erinaceid unit which links to talpids (Miyamoto and Goodman, 1986). Those results are at variance with this study and also with morphological data (McKenna, 1975). Miyamoto and Goodman’s (1986) conclusion is equivocal because they had only two sequences for the shrew and mole, although their methodology required at least three. Miyamoto and Goodman (1986) and Shoshani et al. (1985) also suggested a shrew/mole/hedgehog unit as a sister group to carnivores and pangolins. Finally, Shoshani ’s (1986) immunodiffusion data produced some dubious groupings: hedgehogs with bats, primates, and tupaiids; tenrecs and shrews; edentates and lagomorphs. These results show little congruence with other morphological or molecular data and none with ours. Conclusions We believe that the problem with congruence lies in a lack of appropriate and extensive rate-testing such as that discussed in the methodology section. That judgment should not be accepted as suggesting that we have never read too much into our data. However, we have tried to be conservative in claims here. For clearer patterns of relationships to be discerned, future work must sample greater numbers of taxa and, if the primate data are instructive (Nei, 1986), examine thousands of DNA base pairs from each taxon. Acknowledgments This work was supported by the Taylor Science Fund of the Natural History Museum of Los Angeles County. We thank the late Allan Wilson for laboratory facilities. We thank R. K. Rose, C. L. Watts, and two anonymous reviewers for comments on drafts of the manuscript. Literature Cited Cronin, J. E., and V. M. Sarich. 1980. Tupaiid and Archonta phylogeny: The macromolecular evidence. Pp. 293-312, in Comparative Biology and Evolutionary Relationships of Tree Shrews (W. P. Luckett, ed.). Plenum Press, New York. Findley, J. S. 1955. Speciation of the wandering shrew. University of Kansas Publications, Museum of Natural History, 9:1-68. George, S. B. 1986. Evolution and historical biogeography of soricine shrews. Systematic Zoology, 35:153-162. 1988. Systematics, historical biogeography, and evolution of the genus Sorex. Journal of Mammalogy, 69:443-461. Goodman, M. 1960. The species specificity of proteins as observed in the Wilson comparative analysis plates. The American Naturalist, 94:184-186. Maxson, L. R., and R. D. Maxson. 1990. Proteins II: Immunological techniques. Pp. 127-155, in Molecular Systematics (D. M. Hillis and C. Moritz, eds.), Sinauer Associates, Inc., Sunderland, Massachusetts, xvi + 588 pp. McKenna, M. C. 1975. Toward a phylogenetic classification of the Mammalia. Pp. 21-46, in Phylogeny of the Primates (W. P. Luckett and F. S. Szalay, eds.). Plenum Press, New York, 483 pp. Miyamoto, M. M., and M. Goodman. 1986. Biomolecular systematics of Eutherian mammals: Phylogenetic patterns and classification. Systematic Zoology, 35:230-240. Nei, M. 1986. Stochastic errors in DNA evolution and molecular phylogeny. Pp. 133-147, in Evolutionary Perspectives and the New Genetics (H. Gershowitz, D. L. Rucknagel, and R. E. Tashian, eds.), Alan R. Liss Press, New York. Pierson, E. D. 1986. Molecular systematics of the Microchiroptera: Higher taxon relationships and biogeography. Unpublished Ph.D. dissert.. University of California, Berkeley. Repenning, C. A. 1967. Subfamilies and genera of the Soricidae. United States Geological Survey Professional Paper, 565:1-74. Reumer, j. W. F. 1984. Ruscinian and early Pleistocene Soricidae (Insectivora, Mammalia) from Tegelen (The Netherlands) and Hungary. Scripta Geologica, 73:1-173. .. 1987. Redefinition of the Soricidae and the Heterosoricidae (Insectivora, Mammalia), with the description of the Crocidosoricinae, a new subfamily of Soricidae. Revue de Paleobiologie, 6: 189-192. 1989. Speciation and evolution in the Soricidae (Mammalia: Insectivora) in relation with the paleoclimate. Revue de Suisse Zoologie, 96:81-90. Sarich, V. M. 1969. Pinniped origins and the rate of evolution of carnivore albumins. Systematic Zoology, 18:286-295. 1985. Rodent macromolecular systematics. Pp. 423-452, in Evolutionary Relationships Among Rodents: A Multidisciplinary Analysis (W. P. Luckett and J.-L. Hartenberger, eds.). Plenum Press, New York, xiv + 721 pp. 1993. Mammalian systematics: 25 years among their albumins and transferrins. Pp. 103-112, in Mammalian Phylogeny. Vol. 2. Placentals (F. S. Soule, M. J. Novacek, and M. C. McKenna, 1994 GEORGE AND SARICH — Soricine Albumin Evolution 293 eds.), Springer-Verlag, New York, 321 pp. Sarich, V. M., AND J. E. Cronin. 1976. Molecular systematics of the primates. Pp. 141-170, in Molecular Anthropology (M. Goodman and R. E. Tashian, eds.). Plenum Press, New York. Savage, D. E., and D. E. Russell. 1983. Mammalian paleofaunas of the world. Addison- Wes ley Publishing Company, Reading, Massachusetts, 432 pp. ShoshANI, J. 1986. Mammalian phylogeny: Comparison of morphological and molecular results. Molecular Biology and Evolution, 3:222-242. Shoshani, j., M. Goodman, J. Czelusniak, and G. Braunttzer. 1985. A phylogeny of Rodentia and other eutherian orders: Parsimony analysis utilizing amino acid sequences of alpha and beta hemoglobin chains. Pp. 191-210, in Evolutionary Relationships Among Rodents: A Multidisciplinary Analysis (W. P. Luckett and J.-L. Hartenberger, eds.). Plenum Press, New York, xiv -t- 721 pp. a — Sorex — Otisorex trowbridgii-group b .. S. (Otisorex) Vagrans palustris - S. trowbndgli S. (Sorex) araneus caecutiens “1 r 1 1 1 20 10 0 AID Fig. 1.— Relationships of three subgenera of Sorex. a. Tree based on allozymic data (George, 1988). b, Tree based on albumin immunodistance (AID) data from quantitative precipitin comparisons. a Soricini Blarinini Neomyini Soricini (Sorex) Blarinini (Btarina) Soriculini (Neomys) Notiosoricini (Notiosorex) r~ 100 -t 1 — —I 1 60 20 AID Fig. 2. — Phylogeny of soricine shrews, a, Tree based on allozymic data (George, 1986). b. Tree based on albumin immunodistance (AID) data from quantitative precipitin comparisons. ,.,f. '1 ,s >■■ ■>: -CU ^ ,.» - ,'■■ ,'yfi o^<|;J::«(..j5i. e^l ,:'>jdit«i f5i!mi)^/ftSi^ ib'-ztj.W ■■ ^ -■•■■»•' tv,'.. ■"' 1 ■ -1% .’T^H , y ‘.^,-fi^ 4p; i ,j...i»<. .,.(. K, jA.M.t»(*r; ,N'T

■\y ■Vl. A m4"-’’^v^V-')’H>^%iW^i'''. ■ ^ ''-r i\ I '■•’^' '■,. ?; ■ lUiTt’y f./4^«’vsi!5 <‘t;i,''y'if'/^i 'T'-- J,,':' 4‘V. ' ‘ ‘I’l rio rt ^ ! V •' -.fv-‘“-jli^t(yv-'. -dt . M: '’■V 1^-,.. '■'■■•) (\ i"'’> - .' ' . •' '.'M 4c»:»a4c3^^^ ..-,.1. -y 'f'-' •' ■•li*.' ,v.Wjr f V . A.' •■ . ' i; .C«»>v '^3| 0}.‘k I , ,1^'^ . ■^> '' " : . «>A. r f ,» ^»h4 ■ i)lr^ ‘«'.,'.l -^,1 ., .; T,_ V' -■¥*■ v;'’ IV? nu» 4,n»:y< *'<' »■• ■■•'-‘■’■^ie'ta'rf ''-A.'' -"'Y,, a •< ^ A TO ..vri,. '|P4t9lK». I , -<\ '^\0 ii i'. ’■ VT.Ilv'ji ,j ' i' I ,it!T jnilfJiii "•j'ti:.!'* (_ V i ■<■ < t ' ■'■ "V ■ V Tt-aS; I^Atj ' J «< »ii‘l TiA. ' I UU»"u . -5^ o. • .■■ '■? 3-- .-■ :* '■ .«'. ;V»4 -'J>.-. •■•.•Ml .t«J -t :'i II ■ ^~.A THE SOREX OF THE ARANEUS-ARCTICUS GROUP OlAMMALIA: SORICIDAE): DO THEY ACTUALLY SPECIATE? Jacques Hausser Institute of Zoology and Animal Ecology, University of Lausanne, Batiment de Biologic, CH-1015 Lausanne, Switzerland Abstract In shrews of the Sorex araneus-arcticus group, the efficiency with which Robertsonian fusions can induce genetic isolation between chromosomal races seems to be linked to the size of the chromosome arms involved as well as to the level of geographic fragmentation of suitable habitats. Only those populations differentiated by incompatible fusions of the longest autosomic arms have reached complete reproductive isolation. Some data suggest that ecological differentiation could be initiated afterwards by competition between the new species. Introduction The Sorex araneus-arcticus group is composed of eight species (Table 1) characterized by a common XX-XYjY2 sex chromosome system (bibliography compiled by Reumer and Meylan, 1986; see also Ivanitskaya et al., 1986; Voiobouev and van Zyll de Jong, 1988; Wojcik and Searle, 1988). Although they have different karyotypes, they are closely related species. Chromosomal phylogenies and mechanisms responsible for the chromosomal evolution of this group have been suggested (Searle, 1984; Hausser et al., 1985; Haikka et al., 1987; Voiobouev, 1989). The main mechanisms advocated are Robertsonian fusion, centromeric shift, and tandem fusion, the first being particularly important in the European species. The karyotype of the Iberian species, S. granarius, which is all acrocentric except for the sex chromosomes and a small polymorphic metacentric (2Na = 34, NFa = 34-36), is considered to represent the ancestral condition (Wojcik and Searle, 1988; Voiobouev, 1989; and Voiobouev and Catzeflis, 1989). Chromosomal polymorphism has been found in many species of the araneus-arcti cus group: S. daphaenodon and S. tundrensis (Ivanitskaya and Malygin, 1983; Ivanitskaya et al., 1986), S. arcticus (Voiobouev and van Zyll de Jong, 1988), and especially S. araneus, where 12 chromosome arms are involved in an impressive Robertsonian polymorphism, leading to more than 20 karyotypic races each characterized by distinctive metacentric sets. The exact number of races is still difficult to assess because a clear definition of a karyotypic race is lacking. This polymorphism has been thoroughly studied by numerous authors; the most recent review is by Zima et al. (1988). Robertsonian fusions are thought to be the primary mechanism of this karyotypic differentiation. Attempts have been made on this basis to retrace the evolution of these karyotypic races (Searle, 1984, 1988). Nevertheless, some authors consider that the inverse mechanism of fissions or reciprocal translocations also plays an important part in chromosomal evolution of this species (Haikka et al., 1987). The S. araneus-arcticus group may be involved in a continuous speciation process driven by chromosome rearrangements (Hausser et al., 1985), considering the existence of closely related species and the intraspecific chromosomal polymorphism in the most widely distributed species. However, in a recent review, Bengtsson and Frykman (1990) strongly contested the idea that such simple chromosome mutations as Robertsonian fusions play a major part in speciation. Considering the evidence for selection against genetic isolation of karyotypic races of S. araneus in hybrid zones (Fedyk, 1986; Searle 1986a; abstracts of unpublished papers of the international meeting “The Population and Evolutionary Cytogenetics of Sorex araneus,” Oxford, England, August 29-31, 1987, which will be referred here as “in litt.”), they suggested that, presently, these races are not undergoing speciation, but rather a “de-speciation” process which increases gene flow between chromosome races. The aim of this work is to mitigate these conclusions, which in my view do not account for the variety of situations encountered. The Size of the Robertsonian Metacentrics The difference among karyotypic races of S. araneus is due to Robertsonian fusions of primitive acrocentric chromosomes which correspond, with minor modifications (two centromeric shifts), to those of S. granarius. These arms or acrocentric chromosomes are labelled according to the usual letter nomenclature for S. araneus first proposed by Haikka et al. (1974) and (unwillingly) modified by Fredga and Nawrin (1977). In this nomenclature, arms are ordered according to their size, a being the largest, u the smallest (Fig. 1). Robertsonian polymorphism in S. araneus involves every arm from g to r. Robertsonian fusions are also responsible for a large part of the karyotypic difference between S. araneus and its sibling species S. coronatus, in addition to two centromeric shifts on the chromosome arms b and / (Voiobouev and Catzeflis, 1989). Zima et al. (1988) first demonstrated that fusions of arms g to r do not occur at random in Sorex araneus. Their demonstration can be extended to the 15 acrocentric autosomes of S. granarius, which show fusions in S. coronatus and 5. araneus. These autosomes can be ordered by size and grouped into three arbitrarily delimited categories: long {a to j), medium {g to /), and short (m to q). Documented fusions in which the longer arm belongs to each of those categories are listed in Fig. 2. Fusions involving long arms are few compared to their possible number (3/54). This does not mean that long arms fuse 295 296 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 infrequently or with difficulty, but merely reflects self-limitation of the process (Zima et al., 1988). In S. araneus and S. coronal us, the large metacentrics are spread everywhere and the autosomic arms a to f are never found in an acrocentric state; the metacentric of is common to both species, whereas be is species-specific for S. araneus and ci is species-specific for S. coronal us. Where long arms are fused into metacentrics, they can no longer contribute to new fusions. If fission or reciprocal translocation were generally tolerated in these species, as suggested by Halkka et al. (1987), other combinations of the long arms should be observed. In contrast to fusions involving long arms, almost every possible fusion involving the short arms was found (14/15). However, in populations of S. araneus, these arms frequently remain unfused, or show a high level of intrapopulation Robertsonian polymorphism. These polymorphisms usually characterize local races, e.g., the three such races described for Great Britain (Searle, 1984, 1988). The medium-sized metacentrics present an intermediate situation; 25 of 51 possible fusions are documented. Those involving relatively long arms are distributed through large groups of populations. For example, gm and hi are found everywhere in the west European phylogenetic group of karyotypic races (WEPG; Searle, 1984). These occur from the northern slope of the Alps to central Sweden and from Great Britain to Poland, and jl occurs in all known populations except in the Valais race (southern Alps and Italy). Why Have Robertsonian Metacentrics Spread? The pattern of geographic distribution of metacentric chromosomes suggests that Robertsonian fusions of long arms occurred earlier, or were more successful, than fusions of short arms. It is commonly believed that Robertsonian heterozygotes should have a disadvantage due to increased frequency of nondisjunction at meiosis. TTiey would suffer from partial sterility, which would seriously thwart expansion of metacentrics. Most models of chromosome evolution advocate small demes and inbreeding to explain the initial success of Robertsonian metacentrics (see for example Capanna, 1980, for the superficially similar case of Robertsonian evolution in Mus domeslicus). In Sorex araneus, wherein neither inbreeding nor small isolated demes are likely to occur (Hausser et al., 1985; Bengtsson and Frykman, 1990), meiotic drive, together with high fertility of simple Robertsonian heterozygotes (Garagna et al., 1989), is practically the only mechanism that can account for the success of the metacentrics. Searle (1986/j) provided support for such a process. Unfortunately, as pointed out by Bengtsson and Frykman (1990), his conclusions rest partially on reconstruction of the male karyotype by comparing karyotypes of females with those of their offspring, and are undermined by his own discovery of multiple paternity in common shrew litters (Searle, 1990). Even though it needs further confirmation, the meiotic drive hypothesis remains the most likely one. Hedrick (1981) demonstrated that even a very weak meiotic drive can have spectacular effects. In S. araneus, the link between the size of the metacentrics and their success strengthens the idea that a “mechanical” meiotic process is involved rather than some external selection pressure. Therefore, I suggest that meiotic drive efficiency is dependent to some extent on the size of the fused arms. This hypothesis needs to be tested by crossing heterozygotes for metacentrics of various sizes with homozygotes for the corresponding chromosomes. Hybrid Zones and Contact Zones Bengtsson and Frykman (1990) emphasized that selection in hybrid zones between metacentrics increases gene flow rather than increasing genetic isolation. The best documented example of this process is the contact zone between the Oxford race and the Hermitage race in England (Searle, 1986a). Both races belong to the WEPG, characterized by the gm, hi, and jl metacentrics (Searle, 1984). Characteristics for the Oxford race are the metacentrics kq, no, and pr, whereas the Hermitage race carries the metacentric ko. Metacentric pr, elements of which are acrocentric in the Hermitage race, seems to spread relatively freely throughout the 30-km wide hybrid zone, and a sharp decrease of other metacentrics is observed. Also, an increase of the so-called monobrachial hybrids (e.g., ko-kq) and of the corresponding acrocentric chromosomes k, n, o, and q characterizes populations at the center of the hybrid zone. In this case, the acrocentric state is selected for, because of compatibility with both chromosome races, whereas monobrachial hybrids apparently suffer severe loss of fertility. Thus, a tension zone is maintained wherein polymorphism allows genetic flow between karyotypic races. In northern Poland (Fedyk, 1986; Fedyk and Leniec, 1987), Sweden (K. Fredga, in litt., 1987), and Finland (Halkka et al., 1987), the contact zones involve metacentrics of different phylogenetic groups, which are defined by different fusions of medium-sized arms. In these cases, “hybrid races” frequently occur between the main phylogenetic groups. These “races” can show a mixture of compatible metacentrics issuing from each phylogenetic group. For example, the northern race of Sweden has both gm, which is characteristic of the WEPG, and hn, which originated in the east European phylogenetic group, EEPG (Fredga and Nawrin, 1977). This pattern suggests an independent pace of spread of metacentrics, which may be linked to their size. Additionally, “hybrid races” frequently carry local fusions which could have developed through hybridization (S. Fedyk, in litt., 1987). For example, a local race in northeastern Poland shows the metacentric hk, whereas the EEPG and WEPG carry ik and hi (Wojcik, 1986). These “hybrid races” could act as buffers and decrease the level of difference between forms in direct contact. The contact zones between such forms are sometimes, but not always, characterized by increase in acrocentric frequencies, which suggests the maintenance of genetic flow (Fedyk, 1986). A dine in allele frequencies of MPI through the hybrid zone between northern and central races in Sweden supports this hypothesis (Frykman and Bengtsson, 1984; Bengtsson and Frykman, 1990). Although a very complex structure of hybrid races developed between WEPG and EEPG in northeastern Poland together with an increase in the frequency of acrocentrics, these groups are in direct contact in eastern and southeastern Poland where they show a very narrow contact zone. Interracial 1994 HAUSSER— Speciation in Sorex 297 monobrachial hybrids have been found at only one locality, and increase in frequency of the concerned acrocentrics was not observed (J. Wojcik, in litt., 1987). Thus, the “de-speciation” process may be less general than postulated by Bengtsson and Frykman (1990). Contact Zones in the Alps The situation in the Alps is schematized in Fig. 3. A succession of karyotypic forms of S. araneus is found along the northern slope of the Alps. In the north, the Vaud race occurs as a typical race of WEPG, with metacentrics gm, hi, jl, kr, and no. In the south, the so-called “Acrocentric” form occurs, where all the arms g to r are present as acrocentric chromosomes. These populations are nevertheless polymorphic for the metacentric jl, which was discovered after their name was given. Between the Rhone and the Arve valleys, intermediate “Vaud- Acrocentric” populations occur, in which gm is always present, no always absent, and hi, Jl, and kr are polymorphic. These three forms are mutually isolated by the presence of S. coronatus in the lowest parts of the Rhone and Arve valleys. South of the Bernese Alps, another race is found, the Valais race, characterized by the metacentrics gi, hi, kn, and lo (Hausser et al., 1986). Contact zones have been found between the Valais and Vaud races and between the Valais race and the Acrocentric form (Hausser et al., 1991). These contact zones are different from the Oxford-Hermitage hybrid zone in that they are extremely narrow (less than one kilometer; sympatry was found at only one locality in each case). No hybrids were found either in the vicinity of the Acrocentric-Valais contact zone (25 individuals caught on a transect of about five km) or between Vaud and Valais races (13 individuals on a similar transect; Fig. 4). The arms g, h, i, and k, which are implied in metacentrics with monobrachial homologies in the Vaud and Valais races, were never found in an acrocentric condition in any individual of these races. Although samples are small, two comparisons can be made. First, the Intermediate Vaud- Acrocentric populations occupy an area about 40 km wide between the two parental forms (with 15 individuals out of 24 analyzed being heterozygotes for at least one Vaud metacentric), whereas Valais- Acrocentric intermediates have never been identified. Even if this zone has been accidentally widened by recent isolation of parental forms by S. coronatus, the contrast with the Valais- Acrocentric contact zone is striking. Secondly, in the Oxford-Hermitage hybridization zone, the frequency of monobrachial hybrids reaches 10% in the middle of the zone (which is approximately 30 km wide), whereas the frequency of acrocentrics k and o (arms involved in these monobrachial hybrids) reaches 80% in the same populations. If hybridization occurred, such a pattern should have been detected in the Vaud-Valais contact zone even in a very small sample, because the meiotic problems encountered by monobrachial hybrids are far more severe. A hypothetical hybrid between Vaud and Valais would face in meiosis at worst a multivalent of 1 1 elements at meiotic metaphase 1 , and at best (accounting for the known polymorphic elements of each race), a tetravalent and a multivalent of seven elements (Hausser et al., 1986). By comparison, the monobrachial hybrids between Oxford and Hermitage races only face a tetravalent or a pentavalent, which still is sufficient to induce selection against metacentrics (Searle, 1988). Thus, these very sharp replacements of one form by another strongly suggest genetic isolation. To test this hypothesis, Taberlet et al. (1991) examined a partial sequence (279 bp) of mtDNA (Cytochrome b) for 11 individuals of the various taxa found in the western Alps, especially near known or postulated contact zones (numbered localities in Fig. 3). The preliminary results are shown in Fig. 5. Individuals of the Acrocentric, Intermediate, and Vaud karyotypic forms belong to the same clone, along with two Valais individuals of Les Houches near Chamonix (locality 1, Fig. 3). The other Valais individuals, especially those collected near the contact zones with the Vaud race (localities 2 and 3, Fig. 3) belong to another well- differentiated clone. The most parsimonious interpretation of these data is that Valais metacentrics were able to cross the mtDNA clone boundary near Chamonix, where they did not encounter incompatible metacentrics. Congruence between the clone boundary and the karyotypic race limit in the Vaud-Valais contact zones, on the other hand, strengthens the hypothesis of genetic isolation of these races. The Vaud and Acrocentric populations are not genetically differentiated, as shown by the electrophoretic analysis of 30 loci (Fig. 6; Hausser et al., 1991). Hence, the only difference between the two types of contact zones is the presence of metacentrics with monobrachial homologies in one and not in the other. The chromosome arms involved in the Vaud-Valais differentiation (g, h, i, J, k, I, n, o) not only are more numerous than the ones implied in the Oxford-Hermitage hybrid zone (k, n, o, q), but also include longer elements. These contrasting situations suggest that the size of the fused arms can influence the structure of the contact zone. However, some long elements are involved in the contact zones detected in northern Sweden and in Poland, where genetic exchanges are likely to occur through local “hybrid” races. This apparent contradiction needs to be resolved. The Role of Geographic Barriers and Filters The key is in the presence, in northern Europe, of acrocentric chromosomes corresponding to the metacentrics with monobrachial homology in the involved races. These Robertsonian events are recent. Their present incompleteness and the existence of the Acrocentric form suggest that the population that first recolonized northern and central Europe after the last glaciation has mostly acrocentric karyotypes. I suggest that in northern Europe contact between carriers of incompatible metacentrics occurred in highly polymorphic populations, whereas in Switzerland and in southeastern Poland the present metacentric configurations were fully established before contact of the different races. The sharpness of the Valais- Acrocentric border supports this hypothesis. Despite this clearcut replacement of one form by the other, genetic exchange is attested by the presence of Valais metacentrics on a Vaud- Acrocentric mtDNA substrate. This apparent contradiction is easily resolved if one assumes that 298 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 each of the Valais metacentrics has progressed independently into the Acrocentric population through a broad polymorphic “front” similar to the present Vaud- Acrocentric intermediate zone. After glaciation, the upper Arve valley was frequently cut and reopened by extensions and regressions of glacier tongues from Mont Blanc. Temporary isolation would have cut the hypothetical Valais-Acrocentric intermediates from their Acrocentric relatives for a time long enough to allow Valais metacentrics to accumulate and reach a homozygous state (except for lo, which is still polymorphic in every Valais population). Presently, the polymorphic front no longer exists, and the border between the two forms is maintained by a swiftly flowing river, the Torrent des Griaz, that originates from a glacier two km above the trapping area. The torrent presents a lOO-m-wide stony and sandy bed, which is almost abiotic due to frequent bursts of water, ice, and mud, which is probably unattractive for shrews and thus rarely crossed by them. Thus, hybrids should be rare. Because they should present a quadruple Robertsonian heterozygosity and, for some of them, even a monobrachial homology (jl-lo), hybrids also should have reduced fertility. Nevertheless, one Valais individual was found on the “wrong” side of the river. Such crossings may eventually induce a new polymorphic zone, and the Valais metacentrics may further introgress into acrocentric populations. This situation indicates that strong geographical barriers such as the Bernese Alps, which have existed for a long time between the Vaud and Valais races, are not a prerequisite for a metacentric to reach homozygosity in a local population. Geographic filters, such as rivers or rocky cliffs, seem to be sufficient to ensure quick elimination of the corresponding acrocentrics. This process, which can be helped by local extinctions and recolonizations (Lande, 1985), is more likely if concerned metacentrics originated from the fusion of large arms which should be, as suggested above, strongly favored by meiotic drive. Geographic filters also help to coordinate the progression of various metacentrics differentially advantaged by meiotic drive. Thus, they may contribute to sharpening the boundary between contiguous races, which eventually leads to complete genetic isolation. The different behavior of metacentrics in the Alps compared to northern Europe may therefore be attributed to the stronger partitioning of habitats in the former region. Unquestionable evidence for the absence of hybrids between the Vaud and Valais karyotypic races of Sorex araneus is lacking. Our knowledge of the contact zones between S. araneus Vaud race and S. coronatus is far better. In that case, the lack of hybrids is well substantiated in the large sample analyzed from contact zones (331 individuals; Neet and Hausser, 1989). Thus, true specific status exists between these two forms. A prezygotic barrier presumably was developed (Neet and Hausser, 1989). It is very difficult in this case to decide which geographic barrier led to the independent differentiation of S. coronatus. Based on its present distribution, the species should have originated in southwestern France or in northwestern Spain (Hausser, 1978). As the Pyrenean glaciers never completely separated France from northwestern Spain during the Pleistocene (Nilsson, 1983), and as S. coronatus is not isolated from the diverse chromosomal races of S. araneus in biochemical phylogenies (Catzeflis, 1984), it is difficult to advocate a long geographical isolation to explain this speciation. What remains is a set of specific Robertsonian fusions, involving some of the longer arms, and some other chromosomal mutations. Non-Robertsonian Mutations Sorex coronatus is chromosomally distinguished not only by seven Robertsonian fusions, but also by two centromeric shifts (Volobouev and Catzeflis, 1989) and differences in the nuclear organizing regions (NORs; Olert and Schmid, 1978). The question remains whether these non-Robertsonian mutations are the threshold between mere intraspecific polymorphism and full speciation? For the centromeric shifts, the answer is definitely no: in Sorex araneus, a small secondary arm occurs on the j acrocentric chromosome when it is not fused into jl metacentric (Fig. 7a). A centromeric shift seems the best interpretation for this observation, also made in England (Searle, personal communication), but not in the intermediate Vaud-Acrocentric or in the Acrocentric populations (Fig. 7b). Since such a centromeric-shift polymorphism is widely tolerated in populations, it is surely no more efficient than Robertsonian fusions as an isolating mechanism. The data on the NORs of S. coronatus were obtained from one male. Halkka and Soderlund (1987) demonstrated that four to six of the eight potential NORs of S. araneus were randomly activated in individuals from Finland. Even if the difference in localization of NORs of each species is clear, it is premature to assume that it should have a negative effect on hybrid survival or fertility. Sorex tundrensis seems to have NORs similar to those of both S. araneus and S. coronatus (Ivanitskaya, 1989). Thus, it is hazardous to assign different importance to various types of chromosome mutations in speciation processes. Because Robertsonian fusions are still responsible for the most numerous differences between S. araneus and S. coronatus, the karyologic differentiation between these species is, in my view, of the same nature as the karyotype differentiations among karyotypic races of S. araneus. But, again, it involves longer metacentrics which should have been strongly favored by meiotic drive. Ecology From a strictly genetic point of view, the speciation of S. araneus and S. coronatus, and perhaps of S. araneus Vaud and Valais, is achieved with reproductive isolation. Unlike species which evolved through geographical isolation, they are poorly differentiated genetically (Catzeflis, 1984). Additionally, it has been shown that variation in mandible measurements, which provide the only way to distinguish these species morphologically (Hausser and Jammot, 1974), is correlated primarily with the habitat of these shrews. Thus, most of the morphological differentiation between S. araneus and S. coronatus is a by-product of their parapatric distribution (Hausser, 1984). Neet (1989a) found that the interspecific 1994 HAUSSER — Speciation in Sorex 299 morphological difference is greater between allopatric populations than in contact zones, wherein interspecific territoriality occurs. The contact zones studied consisted of a mosaic of individual territories occupied by one or the other species (Neet, 1989^). These data strengthen the idea that both species exploit the same ecological niche. Such pairs of species in competition should not expect a long future because slight climatic modification may be sufficient to end the equilibrium and lead one of them to extinction. Indeed, S. coronatus seems to be expanding at the expense of S. araneus (Hausser, 1978). Some slight ecological differences in microhabitat utilization were nevertheless noted in the contact zones studied, with S. coronatus living in significantly drier places (Fig. 8; Neet and Hausser, 1990). This differentiation may be maintained by competition pressure. In removal experiments, the remaining species quickly exploited all available habitats and significant differences were lost. In such a situation, selection should favor increasing ecological differentiation. The presence of sympatric species of the araneus-arcticus group in Siberia (Luk’yanova and Rafkin, 1974) suggests that this process occurred in previous stages of diversification of this group. Conclusions In the Sorex araneus-arcticus group, a succession of relationships between taxa differentiated primarily by Robertsonian fusions was observed. These fusions lead to the meeting of chromosome races bearing incompatible metacentrics (monobrachial homologies). When these incompatible metacentrics are small or few, and the corresponding acrocentrics are still present, selection acts against metacentrics and the resulting tension polymorphism allows genetic flow through the hybrid zone. When some of the incompatible metacentrics are medium-sized, genetic flow occurs in some homogenous and continuous habitats, whereas it seems to be interrupted in strongly partitioned habitats like the Alps. When metacentrics are composed of the longer arms, genetic flow is interrupted, although genetic and ecological differentiation is incomplete in the European species of this group. Thus, the combination of chromosome size-dependent meiotic drive and efficiency of geographic filters leads to a variety of situations, including true speciation. Acknowledgments This paper is for a great part a consequence of vivid arguing during the meetings organized by the International Sorex araneus Cytogenetics Committee (ISACC, Dr. J. B. Searle, Chairman) in Oxford (1987) and Lausanne (1990). I owe also many thanks to Dr. R. K. Rose and to the reviewers for helping me to improve both my English and my line of argument. Literature Cited Bengtsson, B. O., and 1. Frykman. 1990. Karyotype evolution: Evidence from the common shrew (Sorex araneus L.). Journal of Evolutionary Biology 3:85-101. Bovey, R. 1948. Un type nouveau d ’heterochromosome chez un mammifere: Le trivalent sexuel de Sorex araneus L. Archiv der J ulius-Klaus-Stiftung, 23:507-5 10. CapANNA, E. 1980. Chromosomal rearrangement and speciation in progress in Mus musculus. Folia Zoologica, 29:43-57. Catzeflis, F. 1984. Systematique biochimique, taxonomic et phylogenie des musaraignes d’Europe (Soricidae, Mammalia). These Universite de Lausanne, Switzerland, 164 pp. Dolgov, V. A. 1985. Soricines of the Palearctic. Moskowskogo Universiteta, 220 pp. (in Russian). FEDYK, S. 1986. Genetic differentiation of Polish populations of Sorex araneus L. II: Possibilities of gene flow between chromosome races. Bulletin of the Polish Academy of Sciences, Series in Biology, 34:161-172. Fedyk, S., and E. Yu. Ivanitskaya. 1972. Chromosomes of Siberian shrews. Acta Theriologica, 17:475-492. Fedyk, S., and H. Leniec. 1987. Genetic differentiation of Polish populations of Sorex araneus L. I. Variability of autosome arm combinations. Folia Biologica (Krakow), 35:57-68. Fredga, K., and j. Nawrin. 1977. Karyotype variability in Sorex araneus L. (Insectivora, Mammalia). Chromosomes Today, 6:153-161. Frykman, I., and B. O. Bengtsson. 1984. Genetic differentiation in Sorex. III. Electrophoretic analysis of a hybrid zone between two karyotypic races in Sorex araneus. Hereditas, 100:259-270. Garagna, S., M. Zuccotti, j. B. Searle, C. A. Redi, and P. J. Wilkinson. 1989. Spermatogenesis in heterozygotes for Robertsonian chromosomal rearrangements from natural populations of the common shrew, Sorex araneus. Journal of Reproduction and Fertility, 87:431-438. Halkka, L., and V. Soderlund. 1987. Random NOR-activation in polymorphic and stable chromosomes of S. araneus. Hereditas, 106:293-294. Halkka, L., O. Halkka, U. Skaren, and V. Soderlund. 1974. Chromosome banding pattern in a polymorphic population of Sorex araneus from northeastern Finland. Hereditas, 76:305-314. Halkka, L., V. Soderlund, U. Skar^, and J. Heikkila. 1987. Chromosomal polymorphism and racial evolution of Sorex araneus L. in Finland. Hereditas, 106:257-275. Hausser, J. 1978. Repartition en Suisse et en France de Sorex araneus L., 1758 et de Sorex coronatus Millet, 1828 (Mammalia, Insectivora). Mammalia, 42:330-341. 1984. Genetic drift and selection: Their respective weights in the morphological and genetic differentiation of four species of shrews in southern Europe (Insectivora, Soricidae). Zeischrift fiir zoologische Systematik und Evolutionsforschung, 22:302-320. Hausser, J., and D. Jammot. 1974. Etude biometrique des machoires chez les Sorex du groupe araneus en Europe continentale (Mammalia, Insectivora). Mammalia, 38:324-343. Hausser, J., D. Graf, and A. Meylan. 1975. Donnees nouvelles sur les Sorex d’Espagne et des Pyrenees (Mammalia, Insectivora). Bulletin de la Societe vaudoise des Sciences naturelles, 72:241-272. Hausser, J., F. Catzeflis, A. Meylan, and P. Vogel. 1985. Speciation in the Sorex araneus complex (Mammalia: Insectivora). Acta Zoologica Fennica, 170:125-130. Hausser, J., E. Dannelid, and F. Catzeflis. 1986. Distribution of two karyotypic races of Sorex araneus (Insectivora, Soricidae) in Switzerland and the post-glacial recolonization of the Valais: First results. Zeitung fur zoologische Systematik und Evolutionsforschung, 24:307-314. HauSvSER, j., F. Bosshard, P. Taberlet, and J. Wojcik. 1991. Relationships between chromosome races and species of Sorex of the araneus group in the western Alps. Pp. 79-95, in The Cytogenetics of the Sorex araneus Group and Related Topics (J. 300 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 Hausser, ed.), Proceedings of ISACC’s Second International Meeting, Memoires de la Societe vaudoise des Sciences naturelles, 19. Hedrick, P. V. 1981. The establishment of chromosomal variants. Evolution, 33:234-251. Ivanitskaya, E. Yu. 1989. Constitutive heterochromatin and nucleolar organizer regions in karyotypes of some shrews (Soricidae, Insectivora). Genetika (Moscow), 25:1188-1198. (In Russian, English summary). Ivanitskaya, E. Yu., and A. I. Kozlowsky. 1983. The karyo logical evidence of the absence of the Arctic shrew (Sorex arcticus) in the Palaearctic. Zoologicheskii Zhumal, 62:399-408 (in Russian, English summary). Ivanitskaya, E. Yu, and V. M. Malygin. 1983. Chromosome polymorphism in Sorex tundrensis from Mongolia. Populatsionie izmenenia vida i problemi ochrenenia gemnofondas mlekopiev, Akademia Nauk SSSR, Moskow, 72-73 (in Russian). Ivanitskaya, E. Yu., A. I. Kozlovskii, V. N. Orlov, Yu. M. Kovalska, and M. I. Basskevich. 1986. New data on karyotypes of shrews (Sorex, Soricidae, Insectivora) in the fauna of the USSR. Zoologicheskii Zhumal, 65:1228-1236 (in Russian, English summary). Kozlowsky, A. I. 1971. Karyotypes and systematics of some populations of shrews usually classed with Sorex arcticus (Insectivora, Soricidae). Zoologicheskii Zhumal, 50:756-762 (in Russian, English summary). 1973. Somatic chromosomes in two species of shrews in the Caucasus. Zoologicheskii Zhumal, 52:571-576 (in Russian, English summary). Lande, R. 1985. The fixation of chromosomal rearrangements in a subdivided population with local extinction and colonisation. Heredity, 54:323-332. Luk’yanova, I. V., AND Y. S. Ravkin. 1974. Quantitative characteristics of the shrew population of the southern taiga and subtaiga forests of the Ob region. The Soviet Journal of Ecology, 5:267-272. Meylan, A. 1968. Formules chromosomiques de quelques petits mammiferes nord-americains. Revue suisse de Zoologie, 75:691-696. Neet, C. R. 1989a. Ecologie comparee et biogeographie evolutive de deux especes parapatriques: Sorex araneus et Sorex coronalus (Mammalia, Insectivora, Soricidae). These Universite de Lausanne, Switzerland, 241 pp. 1989fc. Evaluation de la territorialite interspecifique entre Sorex araneus et S. coronatus dans une zone de syntopie (Insectivora, Soricidae). Mammalia, 53:329-335. Neet, C. R., and J. Hausser. 1989. Chromosomal rearrangements, speciation and reproductive isolation: The example of two karyotypic species of the genus Sorex. Journal of Evolutionary Biology, 2:373-378. 1990. Habitat selection in zones of parapatric contact between the common shrew Sorex araneus and Millet’s shrew S. coronatus. The Journal of Animal Ecology, 59:235-250. Nilsson, T. 1983. The Pleistocene. Geology and Life in the Quaternary Ice Age. F. Enke, Stuttgart, 651 pp. Olert, J., and M. Schmid. 1978. Comparative analysis of karyotypes in European shrew species. I. The sibling species Sorex araneus and S. gemellus: Q-bands, G-Bands and position of NORs. Cytogenetics and Cell Genetics, 20:308-322. Reumer, j. W. F., and A. Meylan. 1986. New developments in vertebrate cytotaxonomy. IX: Chromosome numbers in the order Insectivora (Mammalia). Genetica, 70:119-151. Searle, j. B. 1984. Three new karyotypic races of the common shrew Sorex araneus (Mammalia: Insectivora) and a phylogeny. Systematic Zoology, 33:184-194. 1986a. Factors responsible for a karyotypic polymorphism in the common shrew, Sorex araneus. Proceedings of the Royal Society London, B, 222:277-298. 1986fc. Preferential transmission in wild common shrews (Sorex araneus), heterozygous for Robertsonian rearrangements. Genetic Research, 47:147-148. 1988. Karyotypic variation and evolution in the common shrew, Sorex araneus. Pp. 97-107, in Kew Chromosome Conference 3 (P. E. Brandham, ed.), H. M. S. O. London. 1990. Evidence for multiple paternity in the common shrew, Sorex araneus. Journal of Mammalogy, 71:139-144. Sharman, G. B. 1956. Chromosomes of the common shrew. Nature, 177:941-942. Taberlet, P., L. Fumagalli, and J. Hausser. 1991. mtDNA comparison of the Alpine chromosomal races and species of the Sorex araneus group: Preliminary results. Pp. 107-118, in The Cytogenetics of the Sorex araneus Group and Related Topics (J. Hausser, ed.). Proceedings of ISACC’s Second International Meeting, Memoires de la Societe vaudoise des Sciences naturelles, 19. VAN Zyll de Jong, C. G. 1983. Handbook of Canadian Mammals 1. Marsupials and Insectivores. National Museum of Natural Sciences, Ottawa, Canada, 210 pp. VOLOBOUEV, V. T. 1989. Phylogenetic relationships of the Sorex araneus-arcticus species complex (Insectivora, Soricidae) based on high resolution chromosome analysis. Journal of Heredity, 80:284-290. VoLOBOUEV, V. T., AND F. Catzeflis. 1989. Mechanisms of chromosomal evolution in three European species of the Sorex araneus-arcticus group. (Insectivora, Soricidae). Zeitschrift fiir Zoologische Systematik and Evolutionsforschung, 27:252-262. VOLOBOUEV, V. T., AND C. G. VAN ZYLL DE JONG. 1988. The karyotype of Sorex arcticus maritimensis (Insectivora, Soricidae) and its systematic implications. Canadian Journal of Zoology, 66:1968-1972. WoJClK, J. M. 1986. Karyotypic races of the common shrew Sorex araneus from northern Poland. Experientia (Basel), 47:960-962. WOJCIK, J. M., AND J. B. Searle. 1988. The chromosome complement of Sorex granarius: The ancestral karyotype of the common shrew Sorex araneus. Heredity, 61:225-230. ZiMA, J., J. M. WOJCIK, AND M. HORAKOVA. 1988. The number of karyotypic variants in the common shrew Sorex araneus. Acta Theriologica, 33:467-476. 1994 HAUSSER— Speciation in Sorex 301 Table 1. — Species of the Sorex araneus-arcticus group, their karyotypes and distributions. 2Na: diploid autosomic number; NFa\ autosomic fundamental number. These species bear similar X Yj Y2 sex chromosomes. These data were compiled from various sources cited in Reumer and Meylan (1986). The geographic distribution follows van Zyll de Jong (1983) and Dolgov (1985). Species 2Na NFa Distribution First Description of Karyotype Sorex araneus 18-30 36 Pyrenees to Lake Baikal Sharman, 1956 Sorex arcticus 26 34 Yukon to Newfoundland Meylan, 1968 Sorex asper 30 52 E. Kazakhstan to W. Sinkiang Ivanitskaya and Kozlowsky, 1983 Sorex caucasicus 22 42 Northern Turkey, Caucasus Kozlowsky, 1973 Sorex coronatus 20 40 Northern Spain to Germany Bovey, 1948 Sorex daphaenodon 24-26 42 Urals to Kamchatka Fedyk and Ivanitskaya, 1972 Sorex granarius 34 34-36 Western Spain, Portugal Hausser et al., 1975 Sorex tundrensis 30-34 52-56 Urals to N.W. Canada Kozlowsky, 1971 302 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 Sorex araneus Vaud f ^ I 1 *i I I ^ f * r « *' o .. 5 i\np ■' A \ 4-'' IV Intermediate Vaud - Aero X Y, Y. f C % I ^ » ^ ^ 2^ > i5 ^ »r* h %f i n Z Z o ti L* ;? 4 // « *•**.,& <7 * ¥ % e « Acrocentric form i*i7 u:i ./ % UUh rt:, ^ gi / Sr ^ r::rn55 o « p iSl 4|f» *t t MiM s »* Tt M Valais X \ c f^A/ J n %-o /V,^2 -/ | fc I iW q ^ fi r " ^ It ^ f * ^ Sorex coronatus X Y, Y, 9s ;’•' U * - ' \W V n r;»^7 O P f* ft 1 _ ^ ^ • / • ♦ • * '•T. * r. * m * * * j X r. Ci Fig. 1. — Karyotypes of the species, chromosomal races, and forms of the Sorex araneus group in the western Alps. Nomenclature of the arms after Fredga and Nawrin, 1977. Asterisks: centromeric shifts in Sorex coronatus. Intermediate Vaud- Acrocentric are polymorphic for metacentrics jl, hi, and kr. The acrocentric form is polymorphic for the metacentric jl. i 1994 HAUSSER— Speciation in Sorex 303 smaller arm Fig. 2. — Theoretical and observed autosomic Robertsonian fusions in Sorex araneus and S. coronatus. A primitive acrocentric karyotype similar to that of S. granarius is postulated. WEPG: Western European Phylogenetic Group. S. araneus : Acrocentric Vaud - Aero Vaud Valais S. coronatus intermediates Fig. 3. — Distribution of the species, chromosomal races, and forms of the Sorex araneus group in the western Alps. For karyotypes, see Fig. 1. The numbered localities correspond to hypothetized (2) or known (1 and 3) contact zones. Black: above 2000 m; this area is only partially exploited by shrews (up to 2400-2800 m). 304 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 Fig. 4. — Contact zones between the Valais race (black circles) and the Acrocentric form (white circles) of Sorex araneus (a) and between the Valais race and the Vaud race (grey circles) of the same species (b). The numbers of karyotypically analyzed individuals are shown. These drawings correspond to localities 1 and 3 on Fig. 3. Valais 1 Vaud Acrocentric S. granarius Valais 3 Valais I Valais 1 Vaud Acrocentric S. coronatus S. coronatus S. granarius Valais 3 Valais 2 S. coronatus S. coronatus Fig. 5.— Most parsimonious phylogenetic tree obtained by comparison of partial sequences of mtDNA (cytochrome b) clones (Fitch parsimony). The length of branches is proportional to the number of substitutions. The tree was rooted with S. granarius as the ancestral species. The numbers for the Valais populations correspond to the numbered localities on Fig. 3. After Taberlet et al. (1991). Fig.-6.— UPGMA dendrogram computed from electrophoretic data (30 loci, 88 individuals) of the chromosomal forms presented in Fig. 1. After Hausser et al. (1991). 1994 HAUSSER — Speciation in Sorex 305 Fig. 7. — (a): Centromeric shift on the chromosome j of a heterozygous j, l-jl female of the Vaud race of Sorex araneus. (b) This feature does not exist in this male of the Acrocentric form. 306 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 D o • Q o Q ipl • • Q O O O o O O O O O O o o O O O O • O O o • • • o o • • • Q o Q O • Species captured Habitat type O S. araneus □ Cladium Q Both species Alnus • S. coronatus g Quercus Fig. 8. — Habitat segregation in a 90 x 70 m area of a contact zone between S. araneus, Vaud race, and S. coronatus. The Quercus habitat is drier, the Cladium wetter. After Neet and Hausser (1990), modified. THE PLIO-PLEISTOCENE PATTERNS OF DISTRIBUTION OF THE SORICIDAE IN POLAND Barbara Rzebik-Kowalska Institute of Systematics and Evolution of Animals, Polish Academy of Sciences, Slawkowska 17, 31-106 Krakow, Poland Abstract The study of the insectivore fauna from 26 Polish Pliocene, Pleistocene, and Holocene localities revealed that 15-23 species of shrews were present in the Early and Middle Pliocene, but beginning at the Late Pliocene their number never exceeded ten shrews. However, as genera disappeared from the record, the number of Sorex species increased. The greater diversity of shrews in Poland than in southwestern Europe in the past probably is due to the migrations of shrews from Asia. Introduction The origin of shrews (Soricidae) is far from being clear. According to Sige (1976), some of their morphological characters point to the Paleogene family Nyctitheriidae as their direct ancestor. The oldest true soricid yet known is Srinitium marteli Hugueney, 1976 (Reumer, 1987, 1989). Found in France in the sediments of the Middle Oligocene (about 30 mya), it is representative of the subfamily Crocidosoricinae. This subfamily was erected by Reumer (1987) for the most primitive Oligocene and Miocene forms. So far it is known only from Europe although the Early Miocene Antesorex Repenning, 1967 may also be a member of this subfamily. Crocidosoricinae did not survive beyond the M iocene-Pliocene boundary, but gave rise to the fossil American subfamily Limnoecinae as well as to still living subfamilies Crocidurinae and Soricinae (Reumer, 1987, 1989, 1994). Soricinae strongly radiated during the Pliocene, replacing Crocidosoricinae in Europe. In Poland, there are no materials of shrews from the Oligocene. The only Miocene locality (Belchatow) containing shrews has not yet been examined. On the other hand, the Pliocene and Pleistocene localities are numerous and very rich in the Soricidae. This paper is the result of my long, just finished study of the Plio-Pleistocene shrews of Poland (Rzebik-Kowal ska , 1975, 1976, 1981, 1989, 1990a, \990b, in press). I revised materials described by Kowalski (1956, 1958, 1960a, 1960/j) and Sulimski (1959, 1962) and examined new evidence. In all, about 26 localities were studied, representing the time from the Early Pliocene to the Holocene, but with a long gap in the Middle Pleistocene, when an important part of Poland was covered several times by the Scandinavian ice sheet (Fig. 1). The material studied derived from excavations of sediments situated in the karstic caves and channels of Krakdw-Wieluri Upland in southern Poland (see maps in Szyndlar, 1984; Woloszyn, 1987). Remains of shrews in fossil localities originate from owl pellets. The thanatocoenosis (an assemblage of organisms or their parts brought together in nature after death) of most Polish fossil localities also has this origin (Kowalski, 1964). Some owls nest and seek shelter in caves and devour shrews usually avoided by mammalian carnivores. This explains why a fossil mammalian fauna from a cave almost always includes some shrews. Contrary to diurnal raptors which digest most of their food, owls regurgitate the pellets consisting of fur and bones. These pellets are characteristic in shape and size of particular owl species. Pellets disintegrate quickly, but bones, mainly jaws and teeth, being the hardest and most resistant elements of the skeleton, remain in sediments (Andrews, 1990; Kowalski, 1990). Although fossil remains are incomplete, their size, morphology of jaws (especially of the coronoid and condyloid processes of the mandible), as well as the number and morphology of teeth, allow for their identification. The knowledge of the fauna of shrews can help, on the other hand, for determination of geologic age and paleoenvironment of the fossil fauna. Principal Localities The oldest studied locality, Podlesice, is dated to the Early Pliocene and contains an extremely diverse fauna of shrews. No less than 23 species belonging to about 14-15 genera were found there (Table 1, column A) . Such a large number can arouse doubts as possibly being too high and representing several periods, but studies of the rodents suggest that the fauna from Podlesice is uniform in age (Nadachowski, 1989). It is of course possible that some morphotypes have been determined as separate species, but it does not change the general picture of an extremely rich assemblage of the Soricidae. The fauna of Podlesice contains forms associated with different biotypes. Episoriculus, Deinsdoifia, and the talpid Desmana are typical of humid or aquatic environments. Blarinella, Sorex, and probably Paranourosorex are restricted to the forest, whereas several species of moles and maybe Mafia are inhabitants of open areas. Petenyia, Blarinella, and some other genera were probably eurytopic. The thanatocoenosis of Podlesice evidently derives from owl pellets. The high number of taxa must therefore reflect differentiated ecological conditions in an extensive hunting territory of those birds (Webster, 1973; Nilsson, 1984). The composition of the entire fauna points to a climate colder and drier than that of the Miocene, but definitely warmer than the present one. 307 308 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 The next rich locality, W^ze 1 (Table 1), contains the fauna of the early Middle Pliocene. It is also rich in shrews. I found 15 species belonging to 12 genera there. Environmental conditions were not very different from those at Podlesice, although the climate might have been more humid, perhaps similar to the present-day Mediterranean. A rich fauna of shrews is also present in a younger Middle Pliocene locality, R^bielice Krolewskie lA (Table 1, E). The fauna has a similar number of genera (13) and species (20) as are present at W$ze 1. Thirteen species are identical in both localities; in contrast only eight species known from W^ze 1 and nine from R^bielice Krolewskie lA were present at Podlesice. TTie shrew fauna from the Middle Pliocene of Poland suggests the same mosaic of biotopes as the Early Pliocene of Podlesice. Among other groups of mammals some at R^bielice Krolewskie lA point to a climate more humid and colder than in Early Pliocene, but still warmer than at present (Kowalski, 1960/7, 1989). The number of shrews did not drop to less than ten species until the Late Pliocene faunas at Kielniki 3B (eight) and Kadzielnia 1 (Table 1, F and G), dated to the Plio-Pleistocene boundary (nine). Both suggest a climate similar to the present. Kielniki 3B has six and Kadzielnia 1 seven species identical with those from W^ze 1 and R^bielice Krolewskie lA. Among them are Beremendia fissidens, Petenyia hungarica, and Sorex minutus. A new taxon at this time is Sorex (Drepanosorex) praearaneus. Early Pleistocene faunas are known from two localities in Poland, Kamyk and Kielniki 3 A (Table 1, H and I). Each of them contains only five species of Soricidae. In the slightly older Kamyk fauna, four ancient forms remain: Blarinoides marine, Beremendia fiss ideas, Petenyia hungarica, and Sorex bor (the absence of Sorex minutus, present in all older and younger localities, is undoubtedly accidental, the material being very limited). In the younger Kielniki 3A, Sorex minutus is present but Blarinoides mariae absent. Present in both localities was Sorex (Drepanosorex) praearaneus . A climate similar to the present is suggested by this assemblage. The fauna of Kozi Grzbiet (Table 1, J) lived in a mild phase of climate near the end of the Early Pleistocene. Shrews increased in diversity, ten species were present again, but their number did not reach that of the Pliocene. At Kozi Grzbiet, Beremendia was still present; however, seven or less taxa belonged to the genus Sorex, others being Neomys newtoni and Macroneomys brachygnatus . Woodland was the dominant environment at Kozi Grzbiet. After a long gap in the Middle Pleistocene, the fossil record begins in the Last Interglacial and continues until the Recent. During this time only shrews still living in Poland were present in our country, except for Sorex minutissimus which in the cool period of the Early Pleistocene reached as far as northern France (Heim de Balsac, 1940). Now S. minutissimus is restricted to northern Fennoscandia and Russia. Distribution of the Fauna of Soricidae From this survey, it is evident that during the Pliocene shrews were extremely diversified in Poland. This is in contrast with the picture known from southwestern Europe, where only the Middle Pliocene localities yielded more than ten species of Soricidae (Fejfar, 1966; Reumer, 1984). A part of the answer for this high number of species in the fossil faunas of Poland (in relation to those in southern and western Europe) may be the character of the localities. Some localities in Poland containing materials from the Early Pliocene (Zalesiaki IB or Zamkowa Dolna Cave B) yielded five or less and up to eight taxa of shrews. The absence of some species in a fossil assemblage does not necessarily mean their absence in the area; for example, selection by the predator that accumulated the fossil remains could produce such a result (Lopez-Cardo et al., 1977; Nilsson, 1984; Kowalski, 1990). The difference in diversity of shrews in the southwestern and northeastern parts of Europe may nevertheless be genuine. The source of new insectivore taxa in the European fauna was, as a rule, Asia. Today the family Soricidae is more diversified in eastern Asia, where nine genera are present (Honacki et al., 1982), than in Europe with only four genera (Niethanuner and Krapp, 1990) (see Table 2). Northeastern Europe was more open to the migration from Asia, but some forms probably did not reach the south or west of Europe. The same is true in the case of the genus Sorex. According to earlier publications (Kowalski, 1956; Sulimski, 1959, 1962; Repenning, 1967), Sorex was represented by numerous species in the Plio-Pleistocene. Recent research has demonstrated that many of the species considered to be Sorex belong to other genera (Reumer, 1984). When corrections are made, it becomes clear that the number of Sorex in Pliocene localities of southwestern Europe ranged from zero (Crochet, 1986) to three i (Reumer, 1984). At about the middle of the Pleistocene this number increased to four or five in Germany (Koenigswald, 1972, 1973). On the other hand, in the Pliocene of Poland, five localities for Sorex species are known from Podlesice, R^bielice Krolewskie lA and four from Kielniki 3B and Kad2delnia. The Kozi Grzbiet locality — at the end of the Early Pleistocene— had as many as seven species. In general, shrews prefer warm climatic conditions, but the genus Sorex is associated mainly with rather cold climates. Today Sorex is absent from the tropics, and in the Old World the greatest number of species is known from northeast Asia. Presently, there are no Sorex species in the southern part of the Iberian Peninsula, there are three of them in Spain south of the Pyrenees (Niethammer and Krapp, 1990), five in Poland (Pucek, 1984), five in Finland (Hanski and Kaikusalo, 1989), but 13 or less in Siberia (Judin, 1989) (see Table 3). It is therefore possible, even likely, that the dine in the number of Sorex species existed in the Plio-Pleistocene and that northeastern Europe was more suitable for these shrews than the western parts of the continent. The morphology of some Polish fossil shrews resembles eastern forms like Sorex unguiculatus and S. caecutiens rather than S. araneus, now widely distributed in Europe. This thesis is supported by the absence of Crocidura species in any of the fossil faunas of Poland before the Holocene. Another reason for the high number of shrew taxa in Poland may be the survival of some forms which lasted longer there 1994 RZEBIK-KOWALSKA— Plio-Pleistocene Distribution of the Soricidae 309 than in the west. Blarinella, Sulimskia, Zelceina, and Mafia became extinct in the south and west of Europe at the end of the Middle Pliocene, but survived in Poland until the Plio- Pleistocene boundary at Kadzielnia (Rzebik-Kowalska, 1989, 1990a, 1990^>). Neomys is first recorded at Kozi Grzbiet (end of the Early Pleistocene in Poland), which is later than its first appearance in western Europe. Crocidura, a genus now diversified in Africa but present in both fossil and extant faunas of southern Europe, does not appear as fossil materia! in Poland earlier than in the Holocene. Summary When comparing the presence of fossil shrews in Poland with the western and southern parts of Europe, it becomes evident that: 1) The explosion of Soricidae in Europe, through immigration and radiation, took place earlier than is generally recognized; diversification already had occurred by the start of the Early Pliocene. 2) TTie comparison of the contemporaneous Plio-Pleistocene localities of Poland and southwestern Europe indicates that the former fauna was more diversified. This was probably due to several introductions from Asia, some of which did not reach the west or south. Also, some species persisted longer in the east, contributing to greater diversity. 3) The abundance of species in northeastern Europe is particularly striking in the case of the genus Sorex, where many localities have four or five species, and sometimes as many as seven. In contrast, Crocidura was, in all probability, absent during the Plio-Pleistocene north of the Carpathians and its migration and colonization to the west went to the south of these mountains. Literature Cited Andrews, P. 1990. Owls, Caves and Fossils. Natural History Museum Publications, London, 231 pp. Crochet, J.-Y. 1986. Insectivores pliocenes du sud de la France (Languedoc-Rousillon) et du nord-est de I’Espagne. Palaeo verteb rata , 16:145-171. Feifar, O. 1966. Die plio-pleistozanen Wirbeltierfaunen von Hajnacka und Invanovce (Slowakei), CSSR. V. Allosorex stenodus n.g., n.sp. aus Invanovce A. Neues Jahrbuch fiir Geologic und Palaontologie Abhandlungen, 123:221-248. Hanski, I., AND A. Kaikusalo. 1989. Distribution and habitat selection of shrews in Finland. Annales Zoologici Fennici, 26:339-348. Heim de Balsac, H. 1940. Un Soricidae nouveau du Pleistocene. Comptes rendus des hebdomadaires seances 1’ Academic des Sciences, 211:808-810. Honacki, J. H., K. E. Kinman, and J. W. Koeppl (eds.). 1982. Mammal Species of the World. Allen Press and the Association of Systematics Collections, Lawrence, Kansas, 694 pp. Judin, B. S. 1989. Insectivores of Siberia. Nauka, Siber. Otdel., Novosibirsk, 360 pp. (in Russian). Koenigswald , W. VON. 1972. Sudmer-Berg-2, eine Fauna des friihen Mittelpleistozans aus dem Harz. Neues Jahrbuch fiir Geologic und Palaontologie Abhandlungen, 141:194-221. 1973. Husarenhof 4, eine alt- bis mittelpleistozane Kleinsaugerfauna aus Wiirtemberg mit Petauria. Neues Jahrbuch fiir Geologic und Palaontologie Abhandlungen, 143:23-38. Kowalski, K. 1956. Insectivores, bats and rodents from the Early Pleistocene bone breccia of Podlesice near Kroczyce (Poland). Acta Palaeontologica Polonia, 1:331-394. 1958. An Early Pleistocene fauna of small mammals from the Kadzielnia hill in Kielce (Poland). Acta Palaeontologica Polonia, 3:1-47. 1960a. An Early Pleistocene fauna of small mammals from Kamyk (Poland). Folia Quatemaria, 1:1-24. \960b. Pliocene insectivores and rodents from R^bielice Krolewskie (Poland). Acta Zoologica Cracoviensia, 5: 155-196. 1964. Palaeoecology of mammals from the Pliocene and Early Pleistocene of Poland. Acta Theriologica, 8:73-88 (in Polish). (ed.). 1989. History and Evolution of the Terrestrial Fauna of Poland. Folia Quatemaria, 59-60. 1990. Some problems of the taphonomy of small mammals. Pp. 285-296, in International Symposium on the Evolution, Physiology, and Biostratigraphy of the Arvicolids, Praha. Lopez-Gordo, j. L., E. Lazaro, and A. Fernandez-Jorge. 1977. Comparation de las dietas de Srix aluco, Asio otus y Tyto alba en un mismo biotope de la provincia de Madrid. Ardeola, 23:189-221. Nadachowski, A. 1989. Rodentia. Pp. 151-176, in History and Evolution of the Terrestrial Fauna of Poland (K. Kowalski, ed.). Folia Quatemaria, 59-60. Niethammer, j. and F. Krapp (EDS.). 1990. Handbuch der Saugetiere Europas. Aula-Verlag, Wiesbaden, 3/1, 524 pp. Nilsson, I. N. 1984. Prey weight, food overlap, and reproductive output of potentially competing long-eared and tawny owls. Omis Scandinavica, 15:176-182. PUCEK, Z. 1984. Insectivora. Pp. 48-85, in Key for Identification of Polish Mammals (Z. Pucek, ed.), P. W. N., Warszawa. Repenning, C. A. 1967. Subfamilies and genera of the Soricidae. U. S. Geological Survey Professional Paper, Washington, 565:1-74. Reumer, j. W. F. 1984. Ruscinian and Early Pleistocene Soricidae (Insectivora, Mammalia) from Tegelen (The Netherlands) and Hungary. Scripta Geologica, 73:1-173. 1987. Redefinition of the Soricidae and the Heterosoricidae (Insectivora, Mammalia), with the description of the Crocidosoricinae, a new subfamily of Soricidae. Revue de Paleobiologie, 6:189-192. 1989. Speciation and evolution in the Soricidae (Mammalia: Insectivora) in relation with the paleoclimate. Revue suisse de Zoologie, 96:81-90. 1994. Phylogeny and distribution of the Crocidosoricinae (Mammalia: Soricidae). Pp. 345-356, in Advances in the Biology of Shrews (J. F. Merritt, G. L. Kirkland, Jr., and R. K. Rose, eds.), Carnegie Museum of Natural History Special Publication no. 18, X + 458 pp. Rzebik-Kowalska , B. 1975. The Pliocene and Pleistocene insectivores (Mammalia) of Poland. II. Soricidae: Paranourosorex and Ainbfycoptus. Acta Zoologica Cracoviensia, 20:167-182. 1976. The Neogene and Pleistocene insectivores (Mammalia) of Poland. III. Soricidae: Beremendia and Blarinoides . Acta Zoologica Cracoviensia, 21:359-385. 1981 . The Pliocene and Pleistocene Insectivora (Mammalia) of Poland. IV. Soricidae: Neomysorex n.g. and Episoriculus Ellerman et Morrison-Scott, 1951 . Acta Zoologica Cracoviensia, 25:227-250. 1989. Pliocene and Pleistocene Insectivora (Mammalia) of Poland. V. Soricidae: Petenyia Kormos, 1934 and Blarinella Thomas, 1911. Acta Zoologica Cracoviensia, 32:521-546. 1990a. Pliocene and Pleistocene Insectivora (Mammalia) of Poland. VI. Soricidae: Deinsdorfia Heller, 1963 and Zelceina Sulimski, 1962. Acta Zoologica Cracoviensia, 33:45-77. 310 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 1990b. Pliocene and Pleistocene Insectivora of Poland. VII. Soricidae: Mafia Reumer, 1984, Sulimskia Reumer, 1984, and Paenelimnoecus Baudelot, 1972. Acta Zoologica Cracoviensia, 33:303-327. In press. Pliocene and Pleistocene Insectivora of Poland. VIII. Soricidae: Sorex Linnaeus, 1758, Neomys Kaup, 1829, Macroneomys Fejfar, 1966, Paenelimnoecus Baudelot, 1972 and Soricidae gen. et sp. indet. Acta Zoologica Cracoviensia. SiGE, B. 1976. Insectivores primitifs de I’Eocene superieur et Oligocene inferieur d’Europe occidentale. Memoires du Museum national d’Histoire naturelle, Serie C, 34:1-140. SULIMSKI, A. 1959. Pliocene insectivores from W^ze. Acta Palaeontologica Polonia, 4:119-173. 1962. Supplementary studies on the insectivores from W^ze 1 (Poland). Acta Palaeontologica Polonia, 7:441-502. SZYNDLAR, Z. 1984. Fossil snakes from Poland. Acta Zoologica Cracoviensia, 28:1-156. Webster, J. A. 1973. Seasonal variation in mammal contents of bam owl castings. Bird Study, 20:185-196. WOLOSZYN, B. W. 1987. Pliocene and Pleistocene bats of Poland. Acta Palaeontologica Polonia, 32:207-325. Table 1.— Fossil Soricidae from the principal Pliocene and Early Pleistocene localities of Poland. Localities: A, Podlesice; B, Zalesiaki IB; C, Zamkowa Dolna Cave B; D, W^ie 1; E, Relielice Krolewskie lA; F, Kielniki 3B; G, Kadzielnia; H, Kamyk; I, Kielniki 3A; J , Kozi Grzbiet; plus sign (+), present at a specific locality; minus sign (—), absent at a specific locality. Localities Species ABCDEFGH I J Paenelimnoecus pannonicus (Kormos, 1934) Paenelimnoecus sp. Paranourosorex gigas Rzebik-Kowalska, 1975 Amblycoptus cf. topali Janossy, 1972 Episoriculus gibberodon (Petenyi, 1864) Neomysorex alpinoides (Kowalski, 1956) Mafia dehneli (Kowalski, 1956) Mafia cf. csarnotensis Reumer, 1984 Sulimskia kretozoii (Sulimski, 1962) Blarinoides mariae Sulimski, 1959 Beremendia fissidens (Petenyi, 1864) Beremendia minor Rzebik-Kowalska, 1976 Zelceina podlesicensis Rzebik-Kowalska, 1990 Zelceina soriculoides (Sulimski, 1959) Blarinella dubia (Bachmayer and Wilson, 1970) Blarinella europaea Reumer, 1984 Petenyia robusta Rzebik-Kowalska, 1989 Petenyia hungarica Kormos, 1934 Deinsdorfia reumeri Rzebik-Kowalska, 1989 Deinsdorfia insperata Rzebik-Kowalska, 1989 Deinsdorfia hibbardi (Sulimski, 1962) Deinsdorfia cf. kordosi Reumer, 1984 Sorex minutus Linnaeus, 1766 Sorex bor Reumer, 1984 Sorex casimiri Rzebik-Kowalska n. sp. (in press) Sorex pseudoalpirius Rzebik-Kowalska n. sp. (in press) Sorex polonicus Rzebik-Kowalska n. sp. (in press) Sorex praealpinus Heller, 1930 Sorex subararieus Heller, 1958 Sorex runtonensis Hinton, 191 1 Sorex minutissirnus Zimmermann, 1870 Sorex (Drepanosorex) praearaneus Kormos, 1934 Sorex (Drepanosorex) savini Hinton, 191 1 Sorex (Drepanosorex) sp. Sorex sp. 1 Sorex sp. 2 Sorex sp. 3 Neomys newtorii Hinton, 191 1 -t -1- -1- + — — — — + — — — — — — — + — — — — — — — + — + + 7 - - - - + 7 -1- + - - -f + -1- - - -b -b - -b - - - -1- - -b - + 4- - - - -1- + + -b + -b -b -b — — — — -b — — — — — -1- — — — + -b — — — — — -1- -1- -1- -b - + -b -1- - + - - - + -b -b -b -b -b - -t- -1- - - - -b -b -b -b - - - — — — -b -b — — — — — -1- + - -b -b -b -b - + + + - - -b + + -b -b - - 4- + - + - - - - - - - - + - - - - - — - - - - -b - - - - - - - - - - -b - - - -b — — — — — — -b — — -b -b -b -b -b -b -b - -b -b + + - 1994 RZEBIK-KOWALSKA — Plio-Pleistocene Distribution of the Soricidae 311 Table 1 (cont.) Macroneomys brachyonatus Fejfar, 1966 Soricidae gen. et sp. indet. 1 + Soricidae gen. et sp. indet. 2 + Soricidae gen. et sp. indet. 3 + Soricidae gen. et sp. indet. 4 + Soricidae gen. et sp. indet. 5 - Soricidae gen. et sp. indet. 6 - Soricidae gen. et sp. indet. 7 - Table 2. — Number of Recent genera of Soricidae in Europe and eastern Asia. Eastern Asia Europe Sorex Sorex Neomys Neomys Crocidura Crocidura S uncus S uncus Soriculus Blarinella Anourosorex Chimarrogale Nectogale Table 3. — Number of Sore\ species in Recent fauna of Europe attd Siberia. Plus sign (+), present in the specific locality; minus sign (—), absent in the specific locality; asterisk (*), south of the Pyrenees. References: ^ Neithammer and Krapp, 1990; ^Pucek, 1984; ^Hanski and Kaikusalo, 1989; ^Judin, 1989. Species Spain*' Poland^ Finland^ Siberia'* Sorex coronatus Sorex granarius + - - - Sorex minutus + + + + Sorex araneus - + + + Sorex caecutiens - + + + Sorex alpinus - + - - Sorex isodon - - + + Sorex minutissimus - - + + Sorex cinereus - — — + Sorex mirabilis - - - + Sorex daphaenodon - - - + Sorex unguiculatus - - - + Sorex gracilimus - - - + Sorex roboratus - - - + Sorex tundrensis - — — + Sorex beringianus - - - + Number of species 3 4 5 13 312 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 Time in Millions of years Chronostratigrophy Global Biostratigraphy Mammal ages Localities - 0 -1 -2 -3 -4 -5 HOLOCENE UJ z LU U o I — (/) LU z LU O o Raj Cave ,Giebutt6w, Jozefow Upper TORINGIAN Middle a 0) Q. Q. Z) _o; "O ■g Z B I MARIAN VILLANYIAN o LU RUSCINIAN Koziarnia Cave , Mamutowa Cave Kozi Grzbiet Kielniki 1 Kielniki 3A, Zamkowa Dolna CaveC Kamyk Kadzielnia Kielniki 3B Zamkowa Dolna Cave A R^bielilice Krolewskie 1A and 2 Zamkowa Dolna Cave B W^ze 1 Zalesiaki 1B Podlesice z LU O O O) CL CL z TUROLIAN Fig. 1.— Correlation of the Pliocene to Holocene terrestrial faunas of Poland. COMPARATIVE CYTOGENETICS AND SYSTEMATICS OF SOREX: A CLADISTIC APPROACH E. Yu. Ivanitskaya Institute of Animal Evolutionary Morphology and Ecology, Russian Academy of Sciences, Moscow 117071, Russia; present address: Institute of Evolution, University of Haifa, Mount Carmel, Haifa 31905, Israel Abstract G- and C-banded chromosomes of 11 Sorex species were compared to determine the degree of chromosomal differentiation, typ)es of chromosomal rearrangements, direction of chromosomal evolution, and to propose an ancestral karyotype. Using data on 2N and NF for 30 Sorex species, a WISS cladogram was constructed. This phylogenetic tree shows that the genus Sorex is divided into three large polytypic groups, which may be regarded as the subgenera Sorex s. str. , Otisorex, and Homalurus. The subgenus Sorex includes S. araneus, S. coronatus, S. asper, S. tundrensis, S. daphaenodon, S. satunini, S. arctic us, and S. granarius. The subgenus Otisorex consists mainly of the Nearctic species S. vagrans and its allies, S. bendirii, S. ornatus, S. cinereus, S. leucogaster, and S. ugyunak. The subgenus Homalurus consists of the Palearctic species 5. alpinus, S. minutus, S. volnuchini, S. hucharensis, S. isodon, S. unguiculatus, S. roboratus, and S. caecutiens. The American species 5. fumeus and S. trowbridgii represent independent phyletic branches on the cladogram. The position of European S. samniticus is also rather isolated. Two methods of numerical analysis (UPGMA and WISS) of G-banded karyotypes were carried out for 24 chromosomal races of S. araneus. The WISS method allowed for identification of homoplastic events, synapomorphic characters, and the determination of the racial groups more definitely than has been produced by other methods. Introduction Cytogenetic investigations of shrews belonging to the genus Sorex have made valuable contributions to the systematics of this group. Karyotypic analyses support the recognition of S. isodon (Kozlovsky and Orlov, 1971), S. satunini (syn. S. caucasicus), S. volnuchini (Kozlovsky, 1973a, 19731?), S. gracillimus (Tsuchiya, 1979), S. tundrensis (Ivanitskaya and Kozlovsky, 1983), S. asper (Ivanitskaya et al., 1986), S. leucogaster, and S. ugyunak (Ivanitskaya and Kozlovsky, 1985), S. coronatus (Olert and Schmid, 1978), and S. samniticus (Graf et al., 1979). At the same time, application of cytogenetic data to questions of Sorex subgeneric taxonomy stilt is relatively under- utilized, although there are many unsolved problems to which it could be applied. Most systematists of recent decades have followed Findley (1955), who recognized two subgenera within Sorex: the nominate subgenus, which includes most Palearctic and some Nearctic species, and Otisorex, including most species from North America. However, this classification does not take into account Stroganov’s (1957) recognition of the subgenus Eurosorex for S. hucharensis and its allies, and Heptner and Dolgov’s (1967) recognition of the monotypic subgenus Ognevia for S. mirabilis. An earlier proposition to place S. alpinus in a separate subgenus, Homalurus Schultze, 1890, also should be mentioned. Contrary to this “splitting” tradition, Gureev (1971) argued that it is impossible to divide Sorex into subgenera on the basis of traditionally-used characters, such as morphology of the postmandibular channel. Attempts to classify the genus on the basis of different macromorphological features led to conflicting systems, most of which, however, did not gain nomenclatural recognition. For instance, Hoffmann (1971), used the informal category “species group” to divide the two traditional subgenera into three groups each, many of them monotypic. Some years earlier another system of four groups was proposed by Dolgov and Lukjanova (1966) for Palearctic Sorex on the basis of glans penis morphology. The first attempt to revise the subgeneric division of Sorex with cytogenetic data was done by Vorontsov and Krai (1986). Those authors created 13 subgenera reflecting proposed genealogical relationships. Unfortunately, all their subgeneric names are formally unavailable. For example, S. hucharensis in that scheme was recognized as the subgenus Yudinia, but this species is the type species for the subgenus Eurosorex (Stroganov, 1957). Also, the closely related species S. isodon and S. unguiculatus were placed in different subgenera. All of these schemes of classification, although differing in composition of proposed subgenera and species groups, are similar in one important way: they are not (save Vorontsov and Krai’s [1986] system, as these authors thought) cladistic. And it is evident that cladistic analysis of relationships within such a large and complicated group as the genus Sorex is urgently desirable. In the present paper two cladistic interpretations of relationships within Sorex based on karyotypic data are considered. The first is an analysis of relations of karyotypic races in S. araneus. The second is an analysis of relationships among the Sorex species for which data are available. Phylogenetic Relationships of Karyotypic Races IN THE Superspecies Sorex araneus Karyotypic races of Sorex araneus differ by combinations of autosomal arms arising as a result of different orders of acrocentric chromosome fusions. The first two and the last bi armed autosomal pairs are identical in all races, whereas the other autosomes are acrocentrics or meta-submetacentrics with race-specific combinations of the arms. Twelve karyotypic races of S. araneus were known by 1984. At that time the first cladistic analysis of the group was undertaken by Searle (1984) who presented four cladograms. Two alternative cladograms (Searle, 1984, fig. 3) were constructed using the principle of 313 314 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 minimal number of evolutionary steps (Camin and Sokal, 1965). These two schemes differ in the position of race G: in one case it is included in a group with races I, J, K, A, B, and F; in the other case it is included with races C and H. In addition, both schemes unite race K with J and I, and race F with A and B on the basis of only plesiomorphic characters. Therefore the positions of races F and K are equivocal. In two other cladograms of Searle (1984), constructed on the basis of supposing a hybrid origin of some of the races, uncertainties are decreased. These two schemes differ only in the position of sister groups ED and UK and clarify the position of race G. But they do not decrease the uncertainty in the positions of F and K. Searle (1984) did not show the taxon- character matrix, but the method of Camin and Sokal (1965) requires that the primitive characters should be defined a priori. In this case it is easy to see that the characters n, p, q, and r (F race) are plesiomorphic with respect to np, qr (A race) and to nq, pr (B race) but it is hardly possible to polarize characters np, nq, qr, and pr. Besides, the DEGCH clade (Searle, 1984, fig. 3b) represents a group of races with continuous geographic ranges (Table 1). The inclusion of races C and H in the same sister group can be explained by insufficient data from the European part of Russia and from western Siberia. Currently 24 karyotypic races of S. araneus have been described (Table 1). Shrews from western Europe and Poland are the best studied: karyotypes of S. araneus from more then 50 populations outside of Russia have been studied using G- bands. Therefore, I have reconsidered the cladograms of Searle (1984), taking into account these new data. In addition, features of karyotypic races in the common shrew are convenient for formalization and thus allow various methods of dendrogram construction. In phenetic and cladistic analyses of relationships among karyotypic races of S. araneus, I employed a taxon-character matrix (Table 2) using UPGMA and WISS algorithms, respectively. Two alternative phenograms (Fig. In, b) obtained using the UPGMA algorithm (Sneath and Sokal, 1973) are considered reflections of levels of phenetic similarity of the karyotypic races. At higher levels of these trees, relationships among races agree with those revealed by cladistic analysis (see below). Comparing these two phenograms, there is no disagreement in the composition of UVGS and KBRFANJIMQ groups. In addition, a set of unique characters determines the same position of the L race in both trees. The positions of T, P, O, W races, as well as the ED and HC sister groups, vary between the phenograms. In general, these phenograms identify two groups of karyotypic races in S. araneus. One of them (group UVGS) is represented by races known only from Finland, whereas the other one (group KBRFANJIMQ) is geographically less compact, as races forming some of its subgroups (e.g., AN and IMQ) are widely distributed. Cladistic analysis using the WISS algorithm (after Farris et al., 1970) has evident advantages over the phenetic algorithm. First, it identifies sister groups on the basis of synapomorphies. Second, the method provides a possibility of determining the position of the ancestral karyotype at the base of the cladogram. And lastly, the distance between each pair of nodes in the cladogram reflects the number of Robertsonian fusions, whereas the length of each branch reflects the level of that race’s relative karyotypic advancement. Thus, the cladogram obtained using the WISS algorithm reveals both monophyletic groups and levels of relative change from the ancestor (Fig. 2). The WISS cladogram and phenograms (Fig. la, b) are similar only in recognizing the UV, GS, ED, and IMQ groups and that race L diverges from all other taxa at the very base of the trees. Race T (south Finland) is included with the PW group (east Poland) in one phenogram and is joined with the KBRFA- NJIMQ group in the second, thus sharing features in common with other races from Finland. In contrast, race T in the WISS cladogram is the sister group of the other Finnish races (V, U, S, G), although this race has a great number of autapomorphic characters separating it from other members of the group. Race C (Bialowieza) is placed with race H (Novosibirsk) in both of the phenograms and in the cladograms obtained by Searle (1984), whereas the WISS cladogram indicates it has a common ancestor with other races of east and north Poland (W, O, P). Race F (south England) forms a group with race B (south Sweden) and R (north Czechoslovakia) on both phenograms, but the WISS cladogram treats it as a sister form of the K, J, I, Q, and M races. Race K (west Germany) is a member of the BRFANJIMQ group according to the phenograms, whereas after cladistic analysis it is identified as the sister group of the Polish (W, O, P) and Finnish (T, V, U, S, G) races. The above-mentioned uncertainty of position of the F and K races in Searle’s (1984) paper is eliminated by the WISS method, as race K becomes ancestral to the A, N, F, R, B, J, Q, I, M races and race F becomes ancestral to the A, N, R, B races. The Novosibirsk race (H) clusters with race C (Bialowieza) in all phenograms and in Searle’s cladograms, but forms an independent branch in the WISS cladogram, which makes more sense geographically. The separation of branch H at the base of the cladogram reflects its uncertain position. However, this can be explained by an absence of karyotype data from localities in Siberia and European Russia. The independent origin of some biarmed autosomes has been previously suggested when distributional patterns of karyotypic races were evaluated (Searle, 1984). Analysis of the compo- sition of monophyletic groups of karyotypic races and their characters reveals with more certainty homoplasies at different levels of evolutionary diversification of chromosomal races of S. araneus. For instance, the qr combination, which is an apo- morphy uniting races A, I, H, D, probably appeared independ- ently in European (A, I) and Siberian (H, D) subgroups. Other cases of probable homoplasies are characters pr (races B and E), rk (races Q and S), kq (races M and G). Character ok, which is a synapomorphy for the clade ANFRB, appears independently in race U, thus also demonstrating homoplasy. The character hn appears independently at least three times, in the geographically and cladistically separated groups H, C, and VUSQ. Another case of homoplasy is demonstrated by character mn appearing in clades DE and WOP. The cladistic analysis outlined above allows identification of apomorphic characters in the races under consideration. These characters seem to be indicators of the historical development 1994 IVANITSKAYA— Cytogenetics and Systematics of Sorex 315 of the entire complex. At the present time, four groups of races are recognized in Europe, each group consisting of races of common origin (Fig. 3). In addition, race L, which is known only from Vallais, Switzerland, represents an independent line of karyotypic evolution. This conclusion differs markedly from those of Wojcik and Zima (1987), who recognized seven karyotypic groups in Europe. However, the above authors failed to identify synapomorphic and symplesiomorphic groups among karyotypic races. Further studies of chromosomal evolution in the common shrew may change the above-suggested phylogeny. However, such changes would most likely affect only those parts of the cladogram that contain Siberian races and the position of races Ms and Mn from eastern Europe. Phylogenetic Relations Within the Genus Sorex In S. araneus, groups of karyotypic races are easily recognized as monophyletic assemblages of subspecies- semispecies taxonomic rank. However, in other representatives of this genus identification of such groups is not so simple. First, chromosome numbers are more or less fully known only for Palearctic species. Second, data on differentially-stained chromosomes of Sorex are fragmentary, and serve as a source of data for phylogenetic reconstructions only in some groups of species. At present, chromosome numbers are known for 30 Sorex species, but only ten of them have been differentially- stained. Such fragmentary data allow limited speculation on directions of chromosome rearrangement resulting in the present diversity of karyotypes in the genus Sorex. Nevertheless, even scant information on G-, C-, and N-banded chromosomes per- mits a hypothesis of chromosomal evolution in the genus Sorex. Olert and Schmid (1978) explained differences between S. araneus and S. coronatus by three pericentric and two paracentric inversions and one reciprocal translocation. Later, comparing results obtained from G -banded chromosomes of Siberian populations of S. araneus and S. tundrensis with the data of Olert and Schmid (1978), Aniskin (1987) proposed that Recent karyotypes are derived from transformations of the hypothetical ancestral karyotype of 46 autosomes, rather than one Recent karyotype being derived directly from another. Various sequences of fusions of the ancestral chromosomes led to the origin of the three species considered here. Additional data on C- and N-banded chromosomes of S. araneus and S. tundrensis led to the proposal that the ancestral karyotype of these species had not 46, but 50 autosomes (48 acrocentrics and two small metacentrics) (Ivanitskaya, 1985, 1989). Comparison of G-banded chromosomes of four species of shrews with 2N = 42 showed the identical arrangement of G -bands in the autosomes (Ivanitskaya, 1985). In previous studies, some differences in autosome morphology of these species were explained by pericentric inversions (Halkka et al., 1970; Kozlovsky and Orlov, 1971). However, my data do not confirm that the existence of a chromosome rearrangement of this type in the morphology between separate pairs is due to the different position of centromeres without change of the arrangement of G -bands. Supposedly these chromosomes were formed from the same ancestral elements by centromere-telomere fusion with subsequent differentiation of active centromeres. The karyotypes of S. minutus, S. isodon, S. caecutiens, S. unguiculatus, and S. roboratus consist of 42 chromosomes but differ in autosomal arms. Sorex minutus and S. isodon possess a similar pattern of G -bands in the first five pairs of biarmed autosomes; G-bands of the large meta- and submetacentric autosomes of S. isodon and of the large acrocentric autosomes in S. minutus are similar as well. This cannot be explained by pericentric inversions. It seems more probable that chromo- somes in S. isodon and S. minutus, which are morphologically different, have been formed by centromere-telomere fusions of acrocentric chromosomes of the ancestral karyotype and by conserving as active different ancestral centromeres. Detailed comparison of G-banded chromosomes of shrews characterized by multiple sex chromosomes (S. araneus and S. tundrensis), with shrews having diploid numbers of 42 (S. caecutiens, S. isodon, S. unguiculatus, and S. roboratus) has yielded only two similarities: Y chromosomes; and the “e” arm of X chromosomes of S. araneus and S. tundrensis and X chromosomes of shrews having 2N = 42 (Fedyk and Michalak, 1982; Ivanitskaya, 1985). No autosomes or autosome sections of significant length with identical bands are found in these species. This can be explained by the complex structure of shrew chromosomes formed as a result of various types of tandem fusions of a great number of relatively small acrocentric elements. This hypothesis is indirectly confirmed by an investigation of Sorex chromosome structure by Schmid et al. (1982), which demonstrated unusual phenomena for mammals, namely diffuse distribution of AT-rich sites of DNA within the length of the chromosomes of S. araneus and S. coronatus, and the complete absence of structural heterochromatin in the autosomes. AT-rich (or GC-rich) sites of DNA in chromosomes of most mammals are located in centromeric regions. It is possible that the diffusely distributed AT-rich sites in the chromosomes of Sorex species with low 2N are regions of the ancestral centromeres. Thus, these comparative data allow identification of prevalent types of chromosome changes in the evolution of this group. Apparently, no significant changes in heterochromatin quantity have taken place in the process of karyotype divergence, nor have pericentric inversions played an essential role in chromosome rearrangement. The ancestral karyotype of Sorex supposedly consisted of a large number of acrocentric chromosomes. As the greatest number of autosome arms found in shrews is 94 {S. fumeus), presumably the ancestral karyotype of shrews contained a minimum of 94 small acrocentric autosomes. Karyotype structure (chromosome number and morphology) can be used for cladistic analysis. As G-band data are not available for all shrew species, use of that character for cladistic analysis is not now possible. However, the hypothesis concerning the direction of chromosome evolution discussed above, the pattern of chromosome rearrangement, and the structure of the proposed ancestral karyotype of shrews makes it possible to use chromosome morphology and chromosome number as characters for cladistic analysis. For this character system the WISS algorithm (Farris et al., 1970) was used. 316 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 The initial data include the autosome number of the proposed ancestral karyotype (2N^ = 94 with NF^ = 94), the number of autosomal arms (NFa), and the number of biarmed autosomes (M) in the karyotypes of Recent species. The height of species / (H,) is the distance between the ancestral karyotype and the karyotype of the /th species and was calculated according to the formula H, = 2N- - NFa,. + M-. The distance (D) between species i and j was accordingly D, = NFa,. - NFa,. + M,. - M,.. The height of species k, which is ancestral to the species / and j, was calculated as H, = 0.5 (H, + H. - D,^.). The karyotype containing the greater number of chromosomes and greater number of arms was considered more primitive. Information on karyotypes of 30 species of shrews (Table 3) was included in the analysis. In defining the karyotype ancestral to the araneus-coronatus-tundrensis group, data on G-banded chromosomes of these species were taken into consideration. However this correction did not change the tree topology considerably from that based on only 2N and NFa parameters. Based on the cladogram obtained (Fig. 4), shrews form three polytypic groups. One group includes representatives of the subgenus Ot isorex: American species of the “^vagrans” group and, related to them, S. bendirii, the Amphiberingian species S. ugyunak, and two related American species, S. ornatus and S. Vagrans bairdi. The ancestral karyotype for the subgenus Otisorex appears to be related to karyotypes of Recent S. leucogaster and S. cinereus. Possibly, the karyotype related to S. cinereus was also ancestral to a large group of Palearctic species which were initially attributed to the nominate subgenus. Along the line pa-l-ci-3 (Fig. 4), there are five proposed centromere-centromere and 14 centromere-telomere fusions which define a second polytypic group. This species group is divided into two subgroups. In one of these subgroups (with the terminal species S. raddei and S. gracillimus), centromere- centromere fusions are the main type of chromosome rearrangements. In the other (S. alpinus, S. minutus, S. volnuchini, S. bucharensis), centromere-centromere and centromere-telomere chromosome fusions are substantially equal. Representatives of these subgroups— 5. isodon, S. unguiculatus, S. roboratus, S. caecutiens (point cc on the cladogram) and S. minutus (m/)— are similar in their patterns of G-bands, i.e., the same ancestral chromosomes formed biarmed autosomes in one group (cc) and acrocentric autosomes in the other one (mi). It is difficult to imagine that all these events were independent. Moreover, the cladistic algorithm does not allow for the process of differentiation of active centromeres. Thus, considering the G-band data linking these two subgroups, node 3 must be placed at least five evolutionary steps higher. As a consequence, the part of the cladogram situated beyond node 3 represents a more compact group. The third group in the cladogram is represented by species with multiple sex chromosomes in males. Along the branch of cladogram 19, centromere-telomere fusion should occur. Between nodes 2 and 5, an X autosome translocation and another Robertsonian fusion of ancestral autosomes occurred. Thus, an ancestral karyotype in this group could have included 48 acrocentric and two metacentric autosomes. Three separate evolutionary lines can be identified within this group. The first includes S. araneus, S. coronal us, S. asper, S. granarius, and S. tundrensis, the last species having an Amphiberingian distribution, whereas the others are Palearctic endemics. Derivatives of the second branch are the Asian species S. daphaenodon and the European S. satunini. The third subgroup of shrews with multiple sex chromosomes includes the North American S. arcticus. In this scheme two pairs of species, S. araneus-S. satunini and S. arcticus-S. tundrensis, which were formerly considered closest relatives, belong to different phyletic branches. The American species S. funieus and S. trowbridgii represent independent phyletic branches on the cladogram. The position of the European S. samniticus is also isolated and does not indicate relationships of this species. From this karyologically-based cladogram, the following three groups of subgeneric rank are recognized: Sorex s. str. with type species S. araneus L., Otisorex de Kay with type species S. cinereus, and Homalurus Schultze with type species S. alpinus. Thus, the subgenus Sorex of earlier authors is divided by karyological data into two unrelated taxa. Sorex trowbridgii and S. funieus, which were previously included in this subgenus, are placed distantly on the cladogram which makes it impossible to include them in Sorex s. str. Vorontsov and Krai (1986) included each of these species in a monotypic subgenus. However, the question of formal taxonomic treatment of their relationships requires more investigation. The European species S. samniticus, which was considered until recently a subspecies of S. araneus, appears on the cladogram close to S. trowbridgii, thus forming with the latter a sister group relative to Sorex s. str. As they are dissimilar in macromorphology and geographically distant, this probably results from parallel evolution of similar chromosome structures. The karyotype of S. samniticus seems to be similar to that interpreted as ancestral for the subgenus Sorex s. str. Analysis of G-banded chromosomes is necessary for more accurate determination of the relationships of these two species. The cladogram constructed herein shows interesting similarity to the phylogenetic tree constructed from analysis of biochemical data by George (1988). Thus, S. trowbridgii (together with the karyologically unstudied S. merriami and S. arizonae) and S. fumeus (together with S. dispar) are placed distant from each other and from other groups in both schemes. George’s (1988) data also indicated monophyly of the subgenus Otisorex, which agrees with the cladogram constructed from karyotypic data. But in the cladogram based on allozyme analysis, all Palearctic species represented a compact group. Acknowledgments I thank G. I. Shenbrot of the Institute of Animal Evolution, Morphology and Ecology, Russian Academy of Sciences, for writing the special computer programs and valuable comments 1994 IVANITSKAYA — Cytogenetics and Systematics of Sorex 317 on the manuscript, and I. Ya. Pavlinov of the Zoological Museum of Moscow University for discussion of the results. I am very grateful to Dr. S. B. George and Dr. J. B. Searle for comments that helped improve the manuscript. Literature Cited Aniskin, V. M. 1987. Karyological characteristics and phylogenetic relationships of some species of shrews of the genus Sorex within the araneus-arcticus complex (Insectivora, Soricidae). Zoologicheskii Zhumal, 66:1 19-123 (in Russian). Aniskin, V. M., and I. V. Lukyanova. 1989. A new chromosomal race and analysis of hybrid zone of two Sorex araneus karyomorphs (Insectivora, Soricidae). Doklady Akademii Nauk SSSR, 309:1260-1262 (in Russian). Aniskin, V. M., and V. T. Volobuev. 1980. Chromosomal polymorphism in Siberian populations of the shrews of the araneus- arcticus complex (Insectivora, Soricidae). II. Sayan population of the arctic shrew, Sorex arcticus Kerr (1972). Genetica (USSR), 16:2171-2175 (in Russian). . 1981. Chromosomal polymorphism in Siberian populations of the shrews of the araneus-arcticus complex (Insectivora, Soricidae). III. Three chromosomal forms of the common shrew Sorex araneus L. Genetica (USSR), 17:1784-1791 (in Russian). Brown, R. J. 1974. A comparative study of the chromosomes of three species of shrews, Sorex bendirii, Sorex trowbridgii, and Sorex vagrans. The Wasmann Journal of Biology, 32:303-326. Brown, R. J., and R. L. Rudd. 1981 . Chromosomal comparisons within the Sorex ornatus-S. vagrans complex. The Wasmann Journal of Biology, 39:30-35. Camin, j. H., and R. R. SokAL. 1965. A method for deducing branching sequences in phylogeny. Evolution, 19:31 1-326. Dolgov, V. A., and I. V. Lukyanova. 1966. On the male genital structure of Palearctic shrews (Sorex, Insectivora) as a taxonomic character. Zoologicheskii Zhumal, 45:1852-1861 (in Russian). Dulic, B. 1978. Chromosomen morphologie bei Waldspitzmausen Sorex araneus Linne 1758, aus einigen Gegenden Jugoslawiens. Saugetierkundliche Mitteilungen, 26:184-190. Farris, J. S., F. G. Kluge, and M. J. Eckardt. 1970. A numerical approach to phylogenetic systematics. Systematic Zoology, 19:172-189. Fedyk, S. 1986. Genetic differentiation of Polish populations of Sorex araneus L. II. Possibilities of gene flow between chromosome races. Bulletin of the Polish Academy of Sciences, 34: 161-171 . Fedyk, S., and E. Yu. Ivanitskaya. 1972. Chromosomes of Siberian shrews. Acta Theriologica, 17:475-492. Fedyk, S., and I. Michalak. 1982. Banding patterns on chromosomes of the lesser shrew. Acta Theriologica, 27:61-70. Findley, J. S. 1955. Speciation of the wandering shrew. University of Kansas Publications, Museum of Natural History, 9:1-86. Fredga, K. 1968. Chromosomes of the masked shrew (Sorex caecutiens Laxm.). Hereditas, 60:269-271 . Fredga, K., and J. Nawrin. 1977. Karyotype variability in Sorex araneus L. (Insectivora, Mammalia). Chromosomes Today, 6:153-161. George, S. B. 1988. Systematics, historical biogeography and evolution of the genus Sorex. Journal of Mammalogy, 64:443-461. Graf, J.-D., J. Hausser, and P. Vogel. 1979. Confirmation du statut specifique de Sorex samniticus Altobcllo, 1926 (Mammalia, Insectivora). Bonn zoologischcr Beitrage, 30:14-21. Gureev, A. A. 1971. Shrews (Soricidae) of the World Fauna. Science Press, Leningrad, 252 pp. (in Russian). Halkka, O., U. S KAREN, AND L. Halkka. 1970. The karyotypes of Sorex isodon Turov and S. minutissimus L. Annales Academiae Scientiarum Fennicae. Series A, IV Biologica, 161:1-5. Halkka, L., O. Halkka, U. Skaren, and V. Soderlund. 1974. Chromosome banding pattern in a polymorphic population of Sorex araneus from northeastern Finland. Hereditas, 76:305-314. Halkka, L., V. Soderlund, U. Skaren, and J. Heikkila. 1987. Chromosomal polymorphism and racial evolution of Sorex araneus L. in Finland. Hereditas, 106:257-275. Hausser, J., F. Catzeflis, A. Meylan, and P. Vogel. 1985. Speciation in the Sorex araneus complex (Mammalia, Insectivora). Acta Zoologica Fennica, 170: 125-130. Hausser, J., E. D annelid, and F. Catzeflis. 1986. Distribution of two karyotypic races of Sorex araneus (Insectivora, Soricidae) in Switzerland and the postglacial reconstruction of the Valais: First results. Zeitschrift fiir zoologischer Systematik und Evolutionsforschung, 24:307-314. H EPTNER , V. G., AND V. A. DOLGOV. 1967. On the systematic status of Sorex mirabilis Ognev, 1937 (Mammalia, Soricidae). Zoologicheskii Zhumal, 56:1419-1422 (in Russian). Hoffmann, R. S. 1971 . Relationships of certain Holarctic shrews, genus Sorex. Zeitschrift fiir Saugetierkunde, 36:193-200. Ivanitskaya, E. Yu. 1985. Taxonomical and cytogenetical analysis of transberingian relations of shrews (Sorex: Insectivora) and pikas (Ochotona: Lagomorpha). Unpublished Ph.D. dissert.. Institute of Animal Evolutionary Morphology and Ecology, USSR Academy of Sciences, Moscow, 257 pp. (in Russian). 1986. A new chromosome race of the common shrew (Sorex araneus). Pp. 63-64, in IV Meeting of the All-Union Theriological Society, Volume 1, Moscow (in Russian). 1989. Constitutive heterochromatin and nucleolar organizer regions in karyotypes of some shrews (Soricidae, Insectivora). Genetica (USSR), 25:1188-1198 (in Russian). Ivanitskaya, E. Yu., and A. I. Kozlovsky. 1983. The karyological evidence of absence of the arctic shrew (Sorex arcticus) in the Palearctic. Zoologichesky Zhumal, 62:399-408 (in Russian). 1985. Karyotypes of Palearctic shrews of the subgenus Otisorex with comments on taxonomy and phylogeny of the group “cinereus." Zoologichesky Zhumal, 64:950-952 (in Russian). Ivanitskaya, E. Yu., and V. M. Malygin. 1985. Chromosome complements of insectivorous mammals from Mongolia. Bulleten moskovskogo Obshestva Ispytateley Prirody, 90:15-24 (in Russian). Ivanitskaya, E. Yu., S. I. Isakov, and K. V. Korobitzyna. 1977. Chromosome sets of two species of shrews from Tadjikistan, Sorex buchariensis Ognev, 1921 and Crocidura suaveolens Pallas, 1811 (Soricidae, Insectivora). Zoologichesky Zhumal, 56:1896-1900 (in Russian). Ivanitskaya, E. Yu., A. I. Kozlovsky, V. N. Orlov, Yu. M. Kovalskaya, and M. I. Baskevich. 1986. New data on the karyotypes of red-toothed shrews of the USSR fauna (Sorex, Soricidae, Insectivora). Zoologichesky Zhumal, 65:1228-1236 (in Russian). Kozlovsky, A. 1. 1971. Karyotypes and systematics of some populations of shrews usually classed with Sorex arcticus (Insectivora, Soricidae). Zoologichesky Zhumal, 50:756-762 (in Russian). 1973a. Results of karyological investigations of allopatric forms of the lesser shrew (Sorex minutus). Zoologichesky Zhumal, 52:390-398 (in Russian). 1913b. Somatic chromosomes of two species of shrews from Caucasus. Zoologichesky Zhumal, 52:571-576 (in Russian). Kozlovsky, A. I., and V. N. Orlov. 1971. Karyological 318 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 confirmation of species independent of Sorex isodon Turov (Soricidae, Insectivora). Zoologichesky Zhumal, 50:1056-1061 (in Russian). Kral, B., and S. 1. Radjably. 1976. Karyotypes and G-bands of western Siberian shrews Sorex arcticus and S. araneus (Soricidae, Insectivora). Zoologicke Listy, 23:327-334. MatthEY, R., and A. Meylan. 1961. Le polymorphismique de Sorex araneus L. (Mammalia, Insectivora). Etude de deux portees de cinq et neuf petits. Revue suisse de Zoologie, 68:223-227. Meylan, A. 1965a. La formule chromosomique de Sorex minulus L. (Mammalia, Insectivora). Experientia, 21, 268:1-4. Meylan, A. 1965fc. Repartition geographique des races chromosomiques de Sorex araneus L. en Europe (Mammalia- Insectivora). Revue suisse de Zoologie, 72:636-646. Meylan, A. 1967. Chromosomes of four species of shrews (Soricidae-Insectivora). Mammalian Chromosomes Newsletter, 8(3):187-190. Meylan, A., and J. Hausser. 1973. Les chromosomes des Sorex du groupe araneus-arcticus (Mammalia, Insectivora). Zeitschrift fiir Saugetierkunde, 38:143-158. Olert, J., and M. Schmid. 1978. Comparative analysis of karyotypes in European shrew species. I. The sibling species Sorex araneus and S. gemellus: Q-bands, G-bands and position of NORs. Cytogenetics and Cell Genetics, 20:308-322. Schmid, M., W. Schempp, and J. Olert. 1982. Comparative analysis of karyotypes in European shrew species. !!. Constitutive heterochromatin, replication patterns and sister chromatid exchanges in Sorex araneus and S. gemellus. Cytogenetics and Cell Genetics, 34:124-135. Searle, j. B. 1984. Three new karyotypic races of the common shrew Sorex araneus (Mammalia: Insectivora) and a phylogeny. Systematic Zoology, 33:184-194. Sneath, P. H. A., AND R. R. SOKAL. 1973. Numerical Taxonomy: Principles and Practiee of Numerical Classification. W. H. Freeman Co., San Francisco, 573 pp. Stroganov, S. U. 1957. Mammals of Siberia. Insectivora. Academy of the USSR Science Press, Moscow, 267 pp. (in Russian). Takagi, N., and Y. Fujimaki. 1966. Chromosomes of Sorex shinto saevus Thomas and Sorex unguiculatus Dobson. Japan Journal of Genetics, 41:109-113. Tsuchiya, K. 1979. A contribution to the ehromosome study in Japanese mammals. Proceedings of the Japan Academy, Series B, 55:191-195. Vorontsov, N. N., and B. Kral. 1986. Karyological differentiation and a system of the genus Sorex. Pp. 48-50, in IV Meeting of the All-Union Theriological Society, Volume 1, Moscow (in Russian). WoJClK, J. M. 1986. Karyotypic races of the common shrew (Sorex araneus L.) from northern Poland. Experientia, 42:960-962. WOJCIK, J. M., AND J. ZiMA. 1987. Cytogenetika ryjowski aksamitnej, Sorex araneus Linnaeus, 1758. Przegl^d Zoologiczny, 31:439-456. ZiMA, J., AND B. Kral. 1985. Karyotype variability in Sorex araneus in central Europe (Soricidae, Insectivora). Folia Zoologica, 34:235-243. Table 1. — The karyotypic races of the common shrew and their characteristics. The designation ofS. araneus races used in this paper follows Searle (1984) for ease of comparison with his cladograms. Race Combinations of Autosomal Arms (g-r) Range of Distribution Reference A hi, gm, jl, ko, np, qr Northeastern Scotland Searle, 1984 B hi, gm, jl, ko, nq, pr Southern Sweden Fredga and Nawrin, 1977 C hn, gr, jl, ik, mp, q, o Eastern Poland Fredga and Nawrin, 1977 D hi, gk, jl, mn, op, qr Russia, Tomsk region Aniskin and Volobuev, 1981 E ho, gk, jl, mn, iq, pr Russia, south of Krasnoyarsk region Aniskin and Volobuev, 1981 F hi, gm, jl, ko, n, p, q, r Southern England, northern Poland Searle, 1984; Fedyk, 1986 G hn, gm, jl, kq, ip, or Russia, Novosibirsk region Aniskin and Volobuev, 1981 H hn, go, jl, ik, mp, qr Northern Finland, northern Sweden Halkka et al., 1974 I hi, gm, jl, kq, no, pr Central Britain Searle, 1984 J hi, gm, jl, kp, nr, oq Central Sweden Fredga and Nawrin, 1977 K hi, gm, jl, k, n, o, p, q, r West Germany, Czechoslovakia, Olert and Schmid, 1978; Dulic, Hungary, Yugoslavia 1978; Zima and Kral, 1985 L gi, hj, lo, kn, m, p, q, r Switzerland (Valais) Hausser et al., 1985 Ms hi, gm, jl, kp, no, qr Russia, south of Moscow region Ivanitskaya, 1986 Mn hi, gm, jl, kr, no, pq Russia, north of Moscow region Aniskin and Lukyanova, 1989 N hi, gm, jl, ko, np, q, r Western and northeastern Poland Fedyk, 1986; Wojcik, 1986 0 hk, gr, jl, io, mn, p, q Northern Poland Fedyk, 1986; Wojcik, 1986 P ik, gr, jl, hq, mn, o, p Northeastern Poland Fedyk, 1986; Wojcik, 1986 Q hi, gm, jl, kr, no, p, q Switzerland (Vaud) Hausser et al., 1986 R hi, gm, jl, ko, nr, p, q Northern Czechoslovakia, Moravia Zima and Kral, 1985 S hn, gm, jl, kr, ip, oq Northeastern Finland Halkka et al., 1987 T hk, gq, jl, mo, ip, nr Southeastern Finland Halkka et al., 1987 U hn, gq, jl, ko, ip, mr Central and southern Finland Halkka et al., 1987 V hn, jl, oq, ip, mr, g, k Southwestern Finland Halkka et al., 1987 W hi, gr, jl, ko, mn, p, q Northeastern Poland Fedyk, 1986 1994 IVANITSKAYA— Cytogenetics and Systematics of Sorex 319 TdhX&l.—Taxon-charactermatrix of the superspecies Sorex araneus, based on the presence (1) or absence (0) of arm combinations in the karyotypic races. Characters Karyotypic races A B c D E F G H I i K L M N 0 p Q R s T u V w hi 1 1 0 1 0 I 0 0 1 1 1 0 1 1 0 0 1 1 0 0 0 0 1 hn 0 0 1 0 0 0 1 1 0 0 0 0 0 0 0 0 0 0 1 0 1 1 0 ho 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 jh 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 gm I 1 0 0 0 1 1 0 1 1 1 0 1 1 0 0 1 1 1 0 0 0 0 gr 0 0 1 0 0 0 0 0 0 0 0 0 0 0 1 1 0 0 0 0 0 0 1 gk 0 0 0 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 go 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 gi 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 Jl 1 I 1 1 1 1 1 1 1 1 1 0 1 1 1 1 1 1 1 1 1 1 1 lo 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 ko 1 1 0 0 0 1 0 0 0 0 0 0 0 1 0 0 0 1 0 0 1 0 1 mn 0 0 0 1 1 0 0 0 0 0 0 0 0 0 1 1 0 0 0 0 0 0 1 mp 0 0 1 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 kq 0 0 0 0 0 0 1 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 ik 0 0 1 0 0 0 0 1 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 kp 0 0 0 0 0 0 0 0 0 1 0 0 1 0 0 0 0 0 0 0 0 0 0 kn 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 kr 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 1 0 0 0 0 hk 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 1 0 0 0 hq 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 gq 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 0 0 np 1 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 nq 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 op 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 iq 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 ip 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 0 nr 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 1 0 1 0 0 0 qr 1 0 0 1 0 0 0 1 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 Pr 0 1 0 0 1 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 or 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 oq 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 1 0 0 1 0 io 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 mo 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 mr 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 0 320 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 Table 3. — The karyotypic characters ofSote.\ species. 2N is the number of chromosomes in males; 2Na, the number of autosomes; NFa, the number of autosomal arms. * for additional information, see Table 1. Species 2N 2Na NFa Reference S. cinereus 66 64 66 S. ugyunak 60 58 62 S. leucogaster 66 64 70 S. ornatus 54 52 76 S. bendirii 54 52 66 S. vagrans 54 52 58-62 S. V. bairdi 53-54 51-52 56,60 S. fumeus 66 64 94 S. trowbridgii 34 32 38 S. samniticus 52 50 50 S. alpinus 58 56 66 S. minutus 42 40 54 S. volnuchini 40 38 56 S. bucharensis 40 38 56 S. unguiculatus 42 40 70 S. isodon 42 40 70 S. caecutiens 42 40 70 S. roboratus 42 40 66 S. minutissimus 38 36 68 S. raddei 36 34 60 S. mirabilis 38 36 60 S. gracillimus 36 34 60 S. arcticus 29 26 34 S. granarius 37 34 34-36 S. daphaenodon 27 24 42 29 26 42 S. satunini 25 22 42 S. coronatus 23 20 40 S. araneus 21-29 18-26 36 S. asper 33 30 52 S. tundrensis 31-37 28-34 52 39-41 36-38 54 Meylan, 1967 Ivanitskaya and Kozlovsky, 1985 Ivanitskaya and Kozlovsky, 1985 Brown and Rudd, 1981 Brown, 1974 Brown, 1974 Brown, 1974 Meylan, 1967 Brown, 1974 Grafetal., 1979 Zima, in litt. Meylan, 1965n Kozlovsky, 1973n Ivanitskaya et al., 1977 Takagi and Fujimaki, 1966 Kozlovsky and Orlov, 1971 Fredga, 1968 Ivanitskaya and Malygin, 1985 Halkkaetal., 1970 Kozlovsky, 1913b Ivanitskaya et al., 1986 Tsuchiya, 1979 Meylan and Hausser, 1973 Hausser et al., 1985 Fedyk and Ivanitskaya, 1972 Ivanitskaya et al., 1986 Kozlovsky, \913b Meylan, 1965ft Matthey and Meylan, 1961* Ivanitskaya et al., 1986 Kozlovsky, 1971; Krai and Radjably, 1976; Aniskin and Volobouev, 1980; Ivanitskaya and Kozlovsky, 1983 Ivanitskaya et al., 1986 1994 IVANITSKAYA — Cytogenetics and Systematics of Sorex 321 L P 14 0 r> E T Q M 1 M A r R B K C H S O *J U P u T U u a. s H C O 0 rC 8 R E A M J 1 M O? Fig. 1. — Two variants (a, b) of phenograms reflecting the levels of similarity between karyotypic races (A-W) of Sorex araneus. Fig. 2. — Cladogram WISS reflecting phylogenetic relationships of karyotypic races (A-W) of S. araneus. 322 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 Fig. 3. — Distribution of karyotypic races of S. araneus in Europe. Karyotypic races are combined in four groups due to their apomorphies: 1 (triangle) - hj, gi, lo; 2 (open square) = gm, hi; 3 (closed circle) = gr; 4 (open circle) = ip. 1994 IVANITSKAYA — Cytogenetics and Systematics of Sorex 323 s a Fig. 4. — Cladogram WISS reflecting phylogenetic relationships of the Sorex species: pa, a proposed ancestral Sorex karyotype with 2Na = 94; at, S. arcticus (2Na = 26); ar, S. araneus (2Na = 18-26); as, S. asper (2Na = 30); al, S. alpinus (2Na = 54); bu, S. bucharensis (2Na = 38); be, S. bendiri (2Na = 52); ba, S. bairdi (2Na — 51-52); cc, S. caecutiens, S. unguiculatus, S. isodon, S. roboratus (2Na = 40); co, S. coronatus (2Na = 20); ci, S. cinereus, S. leucogaster (2Na = 64); da, S. daphaenodon (2Na == 24, 26); fu, S. fumeus (2Na = 64); gr, S. gracillimus (2Na = 34); gn, S. granarius (2Na = 32); mi, S. minutus (2Na = 40); ms, S. minutissimus (2Na = 36); mr, S. mirabilis (2Na = 36); or, S. ornatus (2Na = 52); ra, S. rciddei (2Na = 34); sm, S. samniticus (2Na = 50); sa, S. satunini (2Na == 22); tr, S. trowbridgii (2Na = 32); tu, S. tufidrensis (2Na = 28-37); ug, S. ugyuiiak (2Na = 58); va, S. vagrans (2Na = 52); vo, S. volnuchini (2Na = 38). THE EVOLUTION OF THE SORICIDAE AS SHOWN BY THE VARIABILITY OF CRANIAL MORPHOLOGY V. A. Dolgov Department of Vertebrate Zoology and General Ecology, Moscow State University, 119899 Moscow, Russia Abstract Three skull variables were examined for their levels of variation within populations of species and of genera of shrews in the Family Soricidae, and among related families of primitive placental mammals. Characters which exhibit highly correlated variability at the population level and over time also manifest the same correlations throughout the geographic range. The amount of variability increases significantly at the generic level compared to the population and species levels, and soricid shrews have slightly more variability in the proportions of the skull than other families of primitive placental mammals. Introduction The pattern of character variability at the population level reflects the early adaptive peculiarities of the populations and expresses the morphological reorganization of organs in phylogenesis. I examined this variability using concrete data. In choosing a methodology, I tried to maintain simplicity and universality for all cases studied in order to reveal the most genera! principles. The Order Insectivora and its close relatives were chosen because they comprise an old, diversified, and widely distributed group. I studied the families traditionally grouped in the Order Insectivora, although in the view of many contemporary authors some of these families, e.g., Macroscelididae, should be placed in separate orders. This approach permits broader comparisons and more universal conclusions. Although there is controversy on the number of genera in the Family Soricidae, I recognized 21 genera, which is closest to the position of Sokolov (1973). Materials and Methods Material came from the author’s collection, and the collections of several Russian museums and the British Museum (Natural History). The choice of characters for study of taxonomic variability presents substantial difficulties (Simpson, 1948), the most fundamental of which are: (1) the choice of individual characters for measurement, (2) the ability to express measurements in comparable units, and (3) the assignment of weights to the characters in order to obtain reliable general averages. In addition, the coefficient of variation of volumetric values calculated using linear measurements is not comparable to that of linear and surface measurements (Schmalhausen, 1935). Three basic craniometric characters were selected for analysis: the condylobasal length of the skull, the length of the upper toothrow, and the width of the rostrum. These characters can be evaluated quantitatively and are more comparable at all levels of the taxonomic hierarchy. Functionally, they are interpreted as indicators of the distribution of strength for seizing, holding, extracting, and dismembering prey in various ecological situations. Among the various statistical measurements, special attention is given to the coefficient of variation (CV): ^ a ♦ 100 X The CV is abstracted from the absolute value of the characteristic and is an indicator of the strength of the controlling factor (Jablakov, 1966). In this paper, the CV of the measurements of individuals is used to characterize the variability of populations (geographic selection). The average CV of populations is used to characterize the variability of species, the average CV for species is used to characterize the variability of genera, and the average CV for genera is used to characterize the variability of families. In all cases, the averages were weighted according to importance. X = E w. (Zaks, 1976) For a more complete description of forms, taxa, and their interactions, I used orthographic regression. The straight line is derived by the method of least squares, according to the formula: tg20 = E (X.-X) iY,-Y) i=l E (X.-X)^ - E (F.-IO^ 1=1 1=1 This line is called the line of orthographic regression of the gradient line (Lirmick, 1962). Generally, the points (Pj) are located in an elongate strip and the angle (X, the angle of the gradient line to the X axis) is derived simultaneously. The equation for linear regression is as follows: Y = Bo + B, * X The coefficient Bq characterizes the significance of Y when Bj 325 326 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 = 0. This is a theoretical situation. The angle X depends on the coefficient Bj and is used in this paper since it has significance which can be interpolated from the biology of the taxa. To some degree, X characterizes the changes in skull proportions with allometric growth of the parts of the skull. The data presented establish general tendencies in a large number of observations. As a rule, the null hypothesis was not checked. Only illustrative examples are provided, not the complete body of data. Genera such as Suncus, Sorex, and Crocidura are characterized by an abundance of species and wide geographic distributions, permitting a multifaceted comparison of their variability. For this reason, they have been used most frequently for analysis. Individual characteristics change at the population level, as do combinations of characters, in species of the same and different genera. Results In Crocidura suaveolens from Tadjikistan (“Tiger Ravine” collection), the CV of individual characters (Fig. 1) increased in the sequence: condylobasal length of the skull (2.7), length of the upper toothrow (3.2), and width of the rostrum (3.7). Sometimes the upper toothrow measurements showed the greatest variation. For example, in C. hirta from Lake Tanganyika the corresponding CVs were 3.6, 5.1, and 3.1, respectively. The amount of population variability in individual linear characters of the species in the genus Crocidura was similar, although it fluctuated within a broad range. The variability of the upper toothrow ranged from 1.7 (C. fumosa from Tanzania) to 7.6 (C. flavescens from Ethiopia). The mean values of CV for populations of Crocidura species (16 populations, 7 species) were 3.1, 3.4, and 3.4, respectively. In Crocidura populations, pairs of characters tend to vary together. This correlation is more pronounced than differences in the variability of characters taken individually. The correlation coefficient between condylobasal length of the skull and the length of the upper toothrow in C. suaveolens from “Tiger Ravine” was 0.7. For the condylobasal length and the rostrum width, the correlation coefficient was 0.6, and it was 0.5 for the upper toothrow length and the rostrum width. These differences are not great, but the pattern held for all populations. The average correlation coefficients of these character pairs were 0.80, 0.61, and 0.54, respectively. These characters are even more consistent within populations of species in the genus Sorex. The variability increases in the same sequence as with the Crocidura: condylobasal length of skull, length of upper toothrow, and rostrum width. In S. sinalis from the Perm region, for example, the CVs of these characters were 1.4, 1.8, and 2.6, respectively. On the average (11 populations, 7 species) the CVs of these characters in Sorex were 1.54, 1.90, and 2.30, respectively. The amount of character variability within species of Sorex was not significant (range 1.05 to 2.98), indicating that the potential evolutionary capabilities of the lowest evolving units in a genus are similar. In populations of the genus Sorex, the correlation coefficient of the condylobasal length of the skull and the upper toothrow length is greatest, and that between the upper toothrow length and rostrum width is the least. For example, the correlation coefficients of these pairs were 0.70 and 0.20, respectively, in S. sinalis from Kamchatka. On average, the variability of craniometric features of populations of the genus Sorex was less than that of species of Crocidura. The levels of craniometric variability within populations of Neomys fodiens were similar to those of the populations of Sorex. The CV of the condylobasal length of the skull in several populations ranged from 1.6 to 2.2, that of the upper toothrow length ranged from 2.1 to 2.4, and the CV of the rostrum width ranged from 1.9 to 2.5. Usually the condylobasal length of the skull was the least variable character, but the variability of all characters was not significantly different. Populations of species of Suncus also show a broad range of character variability, and the amount of variability of each character within populations of a single species is also large. For instance, the CV of the condylobasal length for populations of Suncus murinus from Madagascar was 5.2, that of the upper toothrow was 10.2, and that of the width of the rostrum was 4.7. In populations of S. murinus from Sri Lanka, the CVs for these characters were 11.5, 10.2, and 12.7, respectively. Thus, cranial characters of shrews are seen to vary in ways characteristic for the taxa. The absolute values of the characters, the amount of variability of single characters, and the relative variability of several characters within a taxon are specific to that taxon. There are also deviations within taxa and similarities among taxa, which suggests that the variations are influenced not only by the relationships within a taxon but also by the relationships among taxa. This does not reject the theory that closely related species populations show more similar variability in character complexes than do populations that are less closely related. Populations display character variability through time in much the same manner. In Crocidura suaveolens from southwestern Tajikistan, the average CV of the condylobasal length of specimens collected over an eight-year period was 0.8. The CV of the upper toothrow was 1.2 in these specimens and that of the rostrum width was 1.6. There was a definite increase in variability from one character to another. There was a statistically significant difference in the CV of the condylobasal length of specimens taken in 1970 and 1972, and a similar difference in the CVs of the rostrum widths of specimens taken in 1960 and in 1972. In fact, the average indicators for these years were significantly different from the average of all specimens taken in all years, although each character varied differently. This shows some degree of variability in these features over time in a single population. The differences in character variability over time in this group also correlate with the degree of differences in variability of these characters within a single season. The correlation coefficient between the variability of the upper toothrow length and the rostrum width was the least (0.5) of the three pairs. Fewer differences were found between the condylobasal length of the skull and the rostrum width (correlation coefficient of 0.61). On the other hand, the direction of change of the correlation coefficient agrees with the change in the direction of the gradient line by year. The CV of the angle L (defining the 1994 DOLGOV — Variation in Skull Morphology in the Soricidae 327 slope of the line) for condylobasal length/rostrum width was 4.0, whereas for the upper toothrow length/rostrum width pair it was 15.1 (Fig. 1). The lower the correlation coefficient for two indicators, the greater the multiyear variability in their proportions and vice versa. The wide variability of the size of the skull from year to year is illustrated by comparing subadults and overwintering adults of Sorex and Neotnys. In some instances, subadults were larger, in others they were smaller. Obviously, this pattern is connected with variations in favorability of living conditions from year to year (Dehnel, 1952; Dolgov, personal observation). Geographic variability of species is also related to population and chronological variability. Forms living in temperate zone plains with smooth gradations of environmental conditions typically exhibit clinal variability in metric indicators. The genus Sorex is dominant in these parts of the geographic range of the Family Soricidae. Usually the variability is not smooth, but is polyclinal. For example, the variability of the condylobasal length of the skull of Sorex caecutiens in the eastern part of its range decreased to the west (Dolgov, 1966, 1972). In the central part of the range, it decreased from south to north and simultaneously to the west. In S. tundrensis from the Altai, variability of the condylobasal length decreased to the west, north, and northeast. The variability of this character in S. sinalis decreased in the central part of its range from west to east (Dolgov, 1966, 1972). The direction and amount of change in variability may be similar between characters (homoclinal), or may be different (heteroclinal). For example, the variability of both the condylobasal length and the upper toothrow length of S. sinalis increased in the central part of the range, whereas the variability of rostrum width remained the same. The same relationship in the variability of these characters was observed in other widely distributed species, including S. caecutiens and S. tundrensis (Dolgov, 1966, 1972). Overall, characters which exhibit highly correlated variability at the population level and over time also manifest the same correlations throughout the geographic range. For example, the condylobasal length of skull and the length of the upper toothrow are homoclinal with respect to each other. Characters which are less correlated with each other at the population level and over time also are heteroclinal over the geographic range. Forms belonging to families with southern and tropical distributions exhibit a regional-mosaic type of geographic variation. For example, neighboring populations of Crocidura suaveolens from the ridges of Tien-Shan, Pamir, and Pamiro- Alaya Mountains and nearby regions of Afghanistan formed a mosaic conglomerate of more or less differing forms without showing vectorized trends (Dolgov, personal observation). A similar example of regional-mosaic variability was provided by Suncus murinus (Dolgov, personal observation). The degree of difference in the genic variability was less in the genera Suncus and Crocidura than in the genus Sorex. This corresponds well with the smaller differences in population variability of individual characters in these two genera, and with the smaller differences in correlation coefficients for character pairs for these genera as compared to Sorex. Thus, for these genera a pattern is seen between population and geographic variability. The degree of variability of linear characters of species of the widely distributed genus Sorex is greater at the species level than at the population level. For instance, the CV of the condylobasal length of the skull varied from 2.0 (S. araneus) to 3.7 (5. tundrensis), whereas in the more narrowly distributed species it ranged from 0.6 (5. gracillimus) to 1.7 {S. bedfordae). But there were instances in which a species with a wide distribution (5. sinalis, 2.2) had less variability than one with a smaller distribution (S. tundrensis, 3.7). The indices of regression for Sorex at the species level demonstrate a tendency toward correlation of features at the population level related to the geographic variability of pairs of characters. The minimal variability of the angle of the gradient line for the condylobasal length/toothrow length pair is typical. It varied for widely distributed species from 7.3 (S. caecutiens) to 13.8 (5. minutus). This agrees with the maximum coefficient of correlation in their populations and with their geographic homoclinality. The CV of the linear characters (condylobasal length, toothrow length, and rostrum width) of the genus Crocidura was less at the species level in some cases than it was at the population level. For instance, the CVs of C. flavescens were 5.8, 6.0, and 5.7, respectively, and in C. hirta the CVs of the same characters were 2.5, 2.6, and 2.6. In other instances (Fig. I), the CVs of the features at the population level were smaller than they were at the species level (C. lasiura: 4.6, 5.9, 6.3) and in still other instances they did not vary significantly (C. suaveolens: 2.7, 3.2, 3.0). The variance of individual characters differs between species to some extent. In some instances the pattern of variability of all characters was of the same sequence (C. hirta, C. flavescens). In others, the upper toothrow length showed the most variability (C. leucodon: 2.7, 41.0, 2.8). Rostrum width was often the most variable character, but the difference between its variability and that of other characters was not great. On the whole, the genus Crocidura typically exhibited diversity in the variability of individual characters. Unlike the linear features, there are species differences in the angle (X) of the regression line. In individual species the differences are similar in their sequences, but are quantitatively different. The X of the regression of toothrow length on rostrum width was most variable: 14.8 in C. fumosa, 29.9 in C. suaveolens, and 37.2 in C. russula, whereas the condylobasal length regressed on rostrum width was least variable: 5.9 in C. lasiura and 11.8 in C. suaveolens (Fig. 1). The variability of species of other genera was similar qualitatively and quantitatively. On the whole, variability of X at the species level increased, reflecting increases in the versions of the cranial proportions in different geographical populations of a species as compared with the variation in one population over many years. At the generic level, the amount of variability of characters increases significantly. The CV of the condylobasal length of the skull in the genus Crocidura was 23.2, the upper toothrow length was 25.6, and the rostrum width 24.2 (Fig. 1). In the genus Sorex, the CVs of the same characters were 12.0, 13.0, 328 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 and 15.7, respectively, whereas the species maxima for the characters were 3.7, 4.5, and 5.6, respectively. This shows that variability of characters relative to each other preserves the same sequence at generic and specific levels. The rostrum width is the most variable character, followed by the upper toothrow length and the condylobasal length. By contrast, the variability of X (indicative of variability in skull proportions) does not increase significantly at the generic level. In the genus Crocidura (22 species examined), the variability of the condylobasal length regressed on upper toothrow length was 8.5, the variability of the condylobasal length regressed on upper rostrum width was 12.9, and that of upper toothrow length regressed on rostrum width was 28.4 (Fig. 1). In 19 Sorex species, the variabilities for these regressions were 15.8, 14.4, and 20.8, respectively. This indicates definite stabilization in divergence of proportions of the crania at the generic level. Evolution of individual characters relative to each other preserves at generic level both variations in populations through time and geographic variation at the species level . Condylobasal length and toothrow length are most strongly correlated geographically and chronologically; the proportions of these characters are similar in both genera and species. Since these characters have minimum impact on the variations of genera and species over time, it is more important that genera and species exhibit independent geographic modification. This was demonstrated by comparing the average generic and specific values. Diplomesodon had the widest rostrum relative to the condylobasal length of the skull; Sorex and Micro sorex (now also called Sorex) had the narrowest. All other species in the family, including Suncus and Crocidura, occupied intermediate positions. The same general relationships among genera were observed in the relationship of the rostrum width relative to the upper toothrow length, although there was less difference among species and Blarinella had the narrowest rostrum. The genera of the Family Soricidae have undergone a similar evolution in the relationship of the upper toothrow length relative to the condylobasal length of the skull; no divergence was observed (Fig. 2). The distribution of cranial proportion by average species character is valid for all species of these genera (Fig. 3). An ecological interpretation of differences in variability of these characters at the generic level is possible. Species of Sorex (including the former Microsorex) are adapted to the extraction of comparatively slow-moving and soft-bodied prey which are mined from the litter of the boreal forest floor. These species have developed a dolichognathic skull. Species of the genus Diplomesodon have a brachyognathic skull, which supports holding and cracking of more mobile, harder-bodied prey found on the surface in arid zones. The species of the genera Crocidura and Suncus occupy intermediate positions. In general, the degree of dolichognathicity or brachyognathicity of the skull is associated with aridity of the ranges occupied by the genera: Diplomesodon occupies deserts, Crocidura and Suncus occupy plains, and Sorex (including Microsorex) live in forests. Thus, the tendencies of variability of characters at the population and species levels are realized in the phylogenetic specificity of form. Increase or decrease in the size of any character causes preconditions for changes in both the qualitative and quantitative functionality of the other characters. A disproportionate change in a single character relative to the others increases even more the possibility of a qualitative change in function, permitting divergence of many kinds (Dolgov, 1976). The factors which determine the correlation of character variability at the species and subspecies levels were discussed by Schmalhausen (1940). At the species level, correlations of character variability indicate physiological interrelationships of the individual. Above the species level, they indicate phylogenetic relationships. This is the border between micro- and macroevolution. The continuity of similarities and differences in variability is preserved in this instance. There was slightly less variability (CV) in the linear measurements in the Family Soricidae (Table 1) than in the same characters at the generic level, especially in the genus Crocidura (Fig. 1). Single character variability reflects the peculiarities of genera; for instance, the length of the skull is less variable than the width of the rostrum. The stability of the proportions of the skull noted at the generic level (Table 2) becomes even more pronounced at the family level (Fig. 1). The stabilization of variation of these characters and their relationships reflects stabilization of the cranial morphotype of shrews at the family level. The similarity of this cranial morphotype in various biologically prosperous genera demonstrates its universality and high biomechanical utility. Variants are most often found in the specializations of individual species. The Soricidae has more genera than other insectivore families, and soricid shrews are the most widely distributed. They are established in a wide spectrum of landscape zones and biotypes. Regardless of this ability, the variability of individual cranial characters at the family level is the lowest of any family considered here (Table 1). Soricid shrews, however, exhibit the same or slightly more variability in the proportions of the skull than is found in the other families of primitive placental mammals (Table 2). Variability in individual cranial characters increases above the family level (Fig. 1, Table 1), which reflects a well-defined divergence of families. At the same time, the variability of cranial proportions is not significant (Fig. 1, Table 2). The relation between skull length and rostrum width was most variable in all groups. Discussion The families considered here have developed common cranial proportions. According to Schmalhausen (1945), “...this stability is not the absence of variability, but rather it is based on that variability always being led in a specific direction. ” The stability of the cranial morphotype of shrews at the family level may be the result of variation of stabilizing selection (Schmalhausen, 1946). This is canalizing selection, in which feedback mechanisms support development of a standard phenotype despite changes in genes, genotypes, and fluctuations in external conditions (Waddington, 1957, 1960). A number of 1994 DOLGOV — Variation in Skull Morphology in the Soricidae 329 the morphological characters of the skulls of these families are primitive, leading Wilson et al. (1975) to postulate a low rate of evolution for the families listed in Table 1. Thus, the pattern of character variation, the individual and comparative qualitative characteristics of this variation, and the correlation between the variability of some characters and the independence of others can be traced from the population— the primary evolutionary unit — through the taxonomic hierarchy to species, genus, family, and order. This pattern is considered sequential, the most likely chain of events in the phylogenetic development of characters and their complexes. The morphometric variability of characters and the corresponding variability in their functions emerge as population adaptations and simultaneously produce the prerequisites for evolutionary transformation of organs. In Chapter 14 of The Origin of the Species, Darwin (1859) noted, “The larger and more dominant groups within each class thus tend to go on increasing in size; and they consequently supplant many smaller and feebler groups.” These and other indicators of biological progress were also developed by Schmalhausen (1940, 1946) and Heptner (1965), based on taxonomic material. According to these theories, Sorex, Crocidura, and Suncus are biologically progressive genera, since there are many species in each genus, both the genera and their constituent species inhabit geographically wide areas, their karyotypes exhibit polymorphism, and they evolve new forms. In these theories, biological progress is tied to morphological change. The morphology of shrews corresponds to a high degree with the demands of their environment and “does not require” radical improvements. However, sometimes the environment changes radically, from warmer to cooler or from wetter to drier. In order to survive, it is necessary to constantly conform to new conditions and shrews have “solved” this problem. They possess the capacity for high rates of evolution within their already developed morphotypes and the maximum adaptive variability at the specific and generic levels. This supports morphological adjustment to the environment. The morphological changes in shrews are not vectorized, but reflect established changes in morphology at lower taxonomic levels. The evolutionary pattern in shrews may be said to be an “indefinitely mobile cladogenesis,” as demonstrated by rapid changes in a limited number of morphological variants. If evaluated as directional (vectoral) changes in morphology, they seem to evolve slowly. In actuality, the rate of their evolution may be high. Perhaps it is precisely such an evolutionary mechanism which allowed shrews and their ancestors to pass successfully from a role as the forefathers of placental mammals in the Cretaceous era to a position as one of the most flourishing groups existing today. It can be assumed that even in the early stages of evolution, their morphotypes were highly adapted to their environments. The possibilities of successful indefinitely mobile cladogenesis are realized by the less specialized genera, such as Sorex, Crocidura, and Suncus. This provides the basis for the assumption that it is precisely these groups which will continue to be the central branches on the evolutionary tree of the Soricidae in the future. Literature Cited Darwin, C. 1859. On the Origin of Species by Means of Natural Selection, or the Preservation of Favoured Races in the Struggle for Life. John Murray, London. DEHNEL, a. 1952. The biology of breeding of the common shrew in laboratory conditions. Annals of the University of Marie Curie- Sklodowska, C, 6, 11:359-376. Dolgov, V. A. 1966. About some regularities in geographical variability of mammals. Reports of Academy of Sciences, USSR, 171(5):1230-1233. 1972. The craniometry and the regularities in geographical variability of craniometrical characters of the Palearctic Sorex species (Mammalia, Sorex). Series Proceedings of the Zoological Museum of Moscow State University, Moscow, 13:150-186. 1976. The intraspecific variability and the early stages of shape formation. Reports of the Academy of Sciences, USSR, 231(4):1021-1024. Heptner, V. G. 1965. The structure of the systematic groups and the biological progress. Zoological Journal, 4(9): 1291-1308. Jablakov, a. V. 1966. The Variability of Mammals. Publishing House Science, Moscow. Linnick, Ju. V. 1962. The Method of the Least Squares and the Bases of the Mathematical Statistical Treatment of the Observation Processing Theory. Physic-Mathematical Literature Publishing House, Moscow. Schmalhausen, I. I. 1935. The determination of fundamental concepts and the method of growth investigation. Pp. 8-60, in The Growth of Animals, Biomed Publishing House, Moscow- Leningrad. 1940. The Ways and the Regularities of the Evolutionary Process. Publishing House of the Academy of Science, USSR, Moscow-Leningrad. 1945. The stabilizing selection and the problem of transmission of the sexual characters from one sex to another. Journal of General Biology (Moscow), 6:363-380. 1946. The Factors of Evolution. Publishing House of the Academy of Science, USSR, Moscow-Leningrad. Simpson, G. G. 1948. The Rates and the Forms of Evolution. Foreign Literature Publishing House, Moscow. Sokolov, V. E. 1973. The Systematics of Mammals. Highest School Publishing House, Moscow. Waddington, C. H. 1957. The Nature of the Genes. Allen and Unwin, London. Waddington, C. H. 1960. Experiments on canalizing selection. Genetic Researches, Cambridge, 1:140-150. Wilson, A. C., G. L. Bush, S. M. Case, and M. C. King. 1975. Social structuring of mammalian populations and of chromosomal evolution. Proceedings of the National Academy of Science, 72:5061-5065. Zaks, L. 1976. The Statistical Estimated Value. Publishing House Statistics, Moscow. 330 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 Table 1.— Variation in the craniometric characters of members of the Soricidae and related families. Condylobasal Length Upper Toothrow of the Skull (mm) Length (mm) Rostrum Width (mm) MIN MEAN MAX MIN MEAN MAX MIN MEAN MAX Family o CV N 0 CV N 0 CV N Soricidae 11.8 21.95 39.4 5.1 10.02 18.3 2.90 6.57 13.2 4.53 20.66 20 2.15 21.47 20 1.62 24.64 20 Erinaceidae 29.5 47.67 90.5 15.4 24.67 49.9 9.2 15.83 27.4 16.39 34.39 6 9.23 37.42 6 5.46 34.49 6 Talpidae 19.5 30.85 56.0 8.90 14.20 30.7 5.80 8.7 17.8 8.4 27.24 13 5.02 35.37 13 2.82 32.43 13 Potamogalidae 34.7 49.0 67.0 15.1 23.0 32.3 10.3 14.1 18.8 2 2 2 Tenrecidae 16.2 42.86 112.4 7.60 20.14 59.3 4.90 11.57 28.7 16.51 38.52 7 9.88 49.14 7 4.10 35.44 7 Solenodontidae 74.3 80.04 83.3 38.7 40.22 41.0 23.8 24.32 24.9 1 1 1 Chrysochloridae 15.80 23.45 40.3 7.0 10.7 18.7 6.50 8.70 14.1 6.26 26.70 4 2.67 24.95 4 1.73 19.91 4 Macroscelididae 28.7 12.32 43.58 28.27 64.1 7.8 22.45 30.5 3.80 6.95 14.4 4 5.08 22.63 4 3.27 47.04 4 All Families 11.80 42.62 112.4 5.10 20.75 59.30 2.90 12.25 28.7 17.23 40.43 8 9.05 43.62 8 5.38 43.95 8 Table 2. —Relative proportions of cranial measurements of families of the Soricidae and related families. Ratio of Condylobasal Length Ratio of Condylobasal Length Ratio of Uppi er Toothrow Length to Uppe r Toothrow Length to Rostrum Width to Rostrum Width MIN MEAN MAX MIN MEAN MAX MIN MEAN MAX Family 0 CV N 0 CV N 0 CV N Soricidae 55.00 64.10 68.0 63.00 71.70 82 52.00 57.30 72 4.05 6.32 20 5.24 7.30 20 6.25 10.91 20 Erinaceidae 55.00 63.00 67 69.00 77.83 90 49.00 65.00 90 3.92 6.22 6 5.27 6.78 6 13.45 20.70 6 Talpidae 40.00 64.07 83 64.00 77.46 91 47.00 64.15 90 10.72 16.88 12 8.61 11.11 12 15.64 24.39 12 Potamogalidae 50.0 55.0 60 63.00 68.50 74 59.00 60.50 62 2 2 — 2 Tenrecidae 45.00 67.00 75 30.00 69.90 85 30.00 51.14 78 9.89 14.75 7 16.86 24.14 7 13.43 26.26 7 Chrysochloridae 68.00 71.50 78 72.00 79.00 86 44.00 63.00 85 4.09 5.72 4 6.52 8.25 4 16.69 26.49 4 Macroscelididae 57.00 62.50 69 82.00 87.00 96 79.00 85.25 100 4 4 4 1994 DOLGOV— Variation in Skull Morphology in the Soricidae 331 (by years) (species) suaveolens suaveolens (species) (genus) (family) (order) (by years) Tadjikistan Fig. 1. — The variability of linear characters and proportions (the gradient line angle) of the primitive placental mammals at different levels of taxonomic hierarchy. A, condylobasal length of skull; B, upper toothrow length; C, rostrum width. 332 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 n n Fig. 2. — The distribution of soricid genera according to craniometric characters, using the letters as defined in Fig. 1. The genera are: 1, Diplomesodon; 2, Crocidura\ 3, Suncus; 4, Microsorex; 5, Sorex; 6, Anourosorex; 1, Feroculus; 8, Blarinella; 9, Myosorex; 10, Surdisorex; 11, Scutisorex; 12, Solisorex; 13, Blarina; 14, Cryptotus; 15, Neomys; 16, Soriculus; 17, Notiosorex; 18, Chodsigoa; 19, Chimarrogale; 20, Nectogale. n = number of genera. 1994 DOLGOV— Variation in Skull Morphology in the Soricidae 333 So rex n. // 9 ■ 7- 5- £l C A c. ^ ^,3 b Fig. 3. — The variability of four soricid genera according to craniometric characters, using the letters as defined in Fig. 1. i.^. Sif -V ^ ■■ j' %■ M »ai»*' ■- f ^ 4 ? •- t V- ii'.r " ■%r ^ '!;;ir ; ''i'ir- r "‘,wn w 'Jir-- - * y*’* ■ A M>« y u.. ■ - ' ' »• H ' .M..1 , ., c y%*# i* [ j J, -, ’■- i — ■-HW J\- i“i 1-1 *■ ’ h 'Mi 1 i — iij- — u •“t ‘V! ij, .ti,. .' >f\' • - ' > !•> l!‘l.-' ’ll I'M 1^, I Tnlftt’^4Cprf I* '1, '£Hi,rU'» ffii' ■ • ' <1^ , ■ (JriiJ*j(i,rii r.T-ff.iy .>j-vf.ffr, t ‘iVi, CHROMOSOMAL EVOLUTION IN THE GENUS CROCIDURA aNSECTIVORA: SORICIDAE) T. Maddalena' and M. Ruedi' ’ Institut de zoologie et d’ecologie animale, Universite de Lausanne, CH-1015 Lausanne, Switzerland Abstract Available Crocidura karyotypes (38 taxa) are used to investigate chromosomal evolution in the genus. Intraspecific variability is relatively rare, whereas interspecific differentiation is common. Diploid chromosome numbers range from 2N = 22 to 2N == 60, and fundamental numbers from FN = 34 to FN = 86. Based on comparisons of differentially-stained chromosomes of Palearctic, Afrotropical, and Oriental species, a hypothetical ancestral karyotype is proposed; it has 38 elements consisting mostly of acrocentric or subtelocentric chromosomes, and four metacentrics. Three main chromosomal evolutionary tendencies are recognized. The first retains a karyotype close to the ancestral state with around 40 chromosomes and is found, as expected from its plesiomorphic state, in all zoogeographic regions inhabited by Crocidura. The second tendency is to increase both diploid and fundamental numbers, as is characteristic of most African Crocidura. The third tendency, which reduces the diploid number mainly through Robertsonian fusions, is apparently restricted to Eurasian species. The karyological data corroborate a phylogenetic hypothesis derived from biochemical data that the genus Crocidura contains two monophyletic clades: an Afrotropical clade and a Palearctic-Oriental clade. Introduction According to Repenning (1967), the two extant subfamilies of Soricidae (Crocidurinae and Soricinae) diverged from a common ancestor during the Oligocene. The genus Crocidura appeared later, probably during the Miocene, somewhere in Africa where most of the species now occur (Heim de Balsac and Meester, 1977). Since the Miocene, the genus Crocidura has radiated considerably. More than 140 species occur in the Old World from the African continent to the Palearctic and Oriental regions (Hutterer et al., 1982). Recently, methods such as allozyme electrophoresis have added considerably to knowledge of the evolution of the Soricidae (Catalan, 1984; Catzeflis, 1984), the Soricinae (George, 1986), and the Crocidurinae (Maddalena, 1990n). The last author compared Palearctic and Afrotropical Crocidura cladistically (Fig. 1). The results suggested the existence of two monophyletic groups, each corresponding approximately to a different zoogeographic region. Crocidura bottegi and C. luna do not group with either clade, probably because they retain many ancestral isozymes. Based on morphological characters, the two species were also regarded as primitive (Heim de Balsac and Lamotte, 1956, 1957; Butler and Greenwood, 1979). The first cytogenetical studies of Crocidura were done about 40 years ago by Bovey (1949) in Switzerland, and later by Meylan and coworkers (see Reumer and Meylan, 1986, for a review). These early studies showed relatively great interspecific chromosomal variation and a very low level of intraspecific polymorphism. Undifferentially-stained karyotypic preparations were useful for recognizing and characterizing species, but of limited utility for establishing chromosomal homologies. Only recently have differentially-stained karyotypes of Crocidura been published (Harada et al., 1985; Tada and Obara, 1986; Hutterer et al., 1987a; Grafodatsky et al., 1988; Maddalena, 1990/j; Maddalena and Vogel, 1990), thus enabling comparative studies. First we present a synopsis of known Crocidura karyotypes, and analyze intra- and interspecific variation. Second, on the basis of comparison of G-banding patterns of European, African, and Asian species, we propose a hypothetical ancestral karyotype for the genus Crocidura and discuss how the different extant species evolved from the primitive state. Finally, we compare findings based on chromosomal data with the hypotheses of relationships based on biochemical data (Fig. 1). Synopsis of Crocidura Karyotypes In checklist form, Reumer and Meylan (1986) listed the chromosome formulas of 21 Crocidura species. Since then (1991), information for 17 additional species has accumulated. Table 1 includes all karyotypes of Crocidura species available, covering about one-fourth of the known species. Unless noted, the scientific names are as in the original papers. In a few cases, the chromosome formula was adapted to standardize chromosomal characteristics, which are the diploid number (2N), the fundamental number of chromosomal arms including the two female sex chromosomes (FN), and the shape of both male (Y) and female (X) sex chromosomes. Species are listed according to geographic origin and by increasing diploid number. The genus Suncus, sometimes considered a subgenus of Crocidura (Ellerman and Morrison-Scott, 1966; Heaney and Timm, 1983) is not included. Results and Discussion Karyotypic Variability in the Genus Crocidura Intraspecific Variation. — Intraspecific chromosomal polymorphism has been observed in only six of the 38 known karyotypes (Table 1). These are found in all of the three zoogeographical regions occupied by the genus Crocidura. Five cases are non-Robertsonian variations, but in one C. suaveolens from Japan, a Robertsonian fusion of two acrocentric chromosomes was described, and the diploid number consequently decreased from 2N = 40 to 2N = 39 (Tsuchiya, 1987). A similar case was recently discovered in a Crocidura species from the Mediterranean island of Pantelleria (Vogel et 335 336 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 al., 1992). Apart from these rare events, non-Robertsonian variations are more frequent. Variation of the diploid number involving B-chromosomes have been reported in C. suaveolens, C. cf. malayana, C. poensis, and C. crossei (Meylan and Hausser, 1974; Maddalena, 1990ft; Ruedi et al., 1990). According to the classification of Volobuev (1981), these are group II B-chromosomes. Another kind of intraspecific non- Robertsonian polymorphism involves extensive FN variations in the Oriental species C. cf. malayana and C. cf. fuliginosa and in the African C. theresae (Table 1). This is probably due to a varying amount of heterochromatin, which leads to the formation of additional short chromosomal arms. The female X chromosome is usually a metacentric of large size, although it is submetacentric or acrocentric in six species (Table 1). Intraspecific variation of the X chromosome has been reported only in C. suaveolens (Tembotova, 1983; Ivanitskaya, 1989; Maddalena, 1990ft). The Y chromosome is generally acrocentric or subtelocentric and of small size. However, in C. bottegi, C. horsfieldi, C. cf. malayana, and C. cf. fuliginosa (Meylan, 1971; Krishna Rao and Aswathanarayana, 1978; Ruedi et al., 1990), it is a relatively large submetacentric element. In C. dsinezumi, C. horsfieldi watasei, C. suaveolens, C. sibirica, C. leucodon, C. poensis, and C. theresae, the Y chromosome appears to be entirely heterochromatic based on C- banding (Harada et al., 1985; Tada and Obara, 1986; Grafodatsky et al., 1988; Maddalena, 1990ft; Maddalena and Vogel, 1990), which is probably the general case in Crocidura. The male sex chromosome of C. suaveolens provides the only real example of intraspecific polymorphism: it has geographically variable size and shape (Tembotova, 1983; Grafodatsky et al., 1988; Ivanitskaya, 1989; Maddalena, 1990ft). This is similar to the variation found in Suncus murinus (Yong, 1971; Yosida, 1985). Interspecific Variation. — Chromosomal variation is extensive among the species of Crocidura (Table 1). The diploid number ranges from 2N = 22 in C. pergrisea to 2N = 60 in C. cf. bicolor, whereas fundamental numbers vary from FN = 34 to FN = 86. Although only about one-fourth of Crocidura species have been karyotyped, some interesting tendencies appear when 2N and FN are plotted together (Fig. 2). Most African species have a high diploid number, generally with around 50 chromosomes and more than 60 chromosomal arms. In contrast, all Palearctic and Oriental species have about 40 chromosomes or less and a relatively low fundamental number. The Thai species C. attenuata is a notable exception among Oriental species because it has a chromosome formula similar to African species, with 2N = 50 and FN -- 66 (Tsuchiya et al., 1979). But if the karyotype of C. attenuata is compared to the conventionally-stained karyotypes of the African species C. Olivieri and C. viaria, which have the same chromosome formula (Maddalena, 1990ft), it is obvious that they are not homologous. The Oriental species has most biarmed elements larger than the acrocentric chromosomes, whereas the opposite situation characterizes the African species. Unfortunately, a banded karyotype is not available for C. attenuata, therefore it is not possible to explain the process responsible for this unusual augmentation of chromosome number. The African species C. luna and C. cf. bottegi and the Palearctic C. canariensis and C. sicula share the same chromosome formulas (Table 1), which is probably due to the retention of ancestral characters, even though some differences appear on G-banding (Fig. 3, and Maddalena, 1990ft). In most Crocidura species, the X chromosome is large, meta- or more rarely submetacentric, representing about 6-8% of the haploid genome (Meylan, 1971; Krishna Rao and Aswathanarayana, 1978; Grafodatsky et al., 1988). However, C. pergrisea is an apparent exception in having an acrocentric X chromosome (Grafodatsky et al., 1988). Another peculiarity is found in C. poensis, wherein the X chromosome is very large and metacentric, representing about 12% of the total genome. Such unusually large sex chromosomes are known in other mammal groups and have generally been attributed either to the insertion of large blocks of heterochromatin or to the translocation of an autosome onto the sex chromosome (Ohno et al., 1964; Hood and Baker, 1986). In C. poensis, most of the X chromosome is euchromatic in C -banding (Maddalena, unpublished data). The heterochromatin portions are restricted to the centromere and to a distal band, both of which are also present in the normal-sized X chromosomes of C. dsinezumi, C. horsfieldi watasei, C. suaveolens, and C. leucodon (Tada and Obara, 1986; Grafodatsky et al. 1988). Thus, translocation involving an autosome is the more probable origin of this very long chromosome in C. poensis. Hypothetical Ancestral Karyotype Among the techniques developed to differentiate chromosomes, G-banding (Seabright, 1971) has become the most popular because of its high resolution power. Thus it is possible to compare the banding pattern of chromosomes of different species and to recognize homologies. We propose here a hypothetical ancestral karyotype on the basis of homologous chromosomes shared between three Palearctic (C. russula, C. suaveolens , and C. canariensis) and three Afrotropical (C. luna, C. olivieri, and C. poensis) species. Hie results are presented in Fig. 3. Crocidura cf. malayana (Fig. 4) was the Oriental representative for comparison. According to the parsimony principle, homologous chromosomes shared between Palearctic and Afrotropical species were considered primitive characters, and were retained to reconstruct the ancestral karyotype. Unfortunately, no true outgroup could be used to root a cladistic analysis, either because the G -banded karyotype of sister taxa is unknown (e.g., Sylvisorex) or the karyotypes are too divergent to permit recognition of homologies (e.g., Soricinae, Grafodatsky et al., 1989). The comparative approach recognized 19 pairs of homologous chromosomes (Fig. 3), giving a formula of 2N = 38 and FN = 54 to FN - 58. Only four pairs of chromosomes would be metacentric, whereas most of the complement would be acrocentric with a variable number of short chromosomal arms (Fig. 5). This model remains approximate because homologies of the smallest chromosomes are difficult to establish, and because few representatives could be compared. However, that the morphologically and biochemically primitive C. luna (2N = 36), C. cf. bottegi (2N = 36), and C. bottegi 1994 MADDALENA AND RUEDI— Chromosomal Evolution in the Genus Crocidura 337 (2N = 40) possess karyotypes close to the model reinforces our conclusion (Heim de Balsac and Lamotte, 1957; Butler and Greenwood, 1979; Maddalena, 1990a; Fig. 1). Chromosomal Evolution in the Genus Crocidura Although only 38 of the 149 species of Crocidura have been karyotyped (Table 1), several tendencies of chromosome evolution are evident. First, at the intraspecific level, karyological polymorphism is rare, and if present, it is mainly due to non-Robertsonian processes such as the presence of supernumerary chromosomes (B-chromosomes, Fig. 4) or variation in the number of additional heterochromatic arms (Ruedi et al., 1990). At the supraspecific level, however, there is much more variation. Studies have shown that Palearctic species such as C. horsfieldi watasei, C. leucodon, and C. pergrisea have reduced diploid numbers due to Robertsonian translocations (Harada et al., 1985; Grafodatsky et al., 1988) and perhaps also to tandem fusions (Grafodatsky et al., 1988). In other Palearctic species and a few Oriental species there is karyotypic stability (Table 1), with a chromosome formula close to the hypothetical ancestral type (Fig. 5), even though some differences appear in G-band patterns (Harada et al., 1985; Grafodatsky et al., 1988; Grafodatsky et al., 1989; Maddalena, 1990/j). As several Eurasian species share a similar karyotype with 2N = 40, Grafodatsky et al. (1988) regarded that formula as the primitive karyotype for Crocidura. Those species are C. suaveolens (including gueldenstaedtii, cypria, and monacha), C. sibirica, C. dsinezumi, C. lasiura, C. dracula, C. caspica, and C. cf. fuliginosa. Other methods should be used to test whether these cytologically similar species represent a monophyletic unit within the Palearctic-Oriental species-group. Crocidura cf. malayana, which could also be included in this group, differs by the presence of B-chromosomes, which raise the original diploid number of 38 to 40 (Ruedi et al., 1990). The Oriental C. attenuata has a chromosome formula identical to some African species (Tsuchiya et al., 1979; Table 1). However, the morphology of its chromosomes appears different suggesting convergence. However, the process responsible for such an increase of the chromosome number is unknown. Whereas the chromosome numbers of Palearctic and Oriental species often decrease or remain stable, Afrotropical Crocidura show an increase of both diploid and fundamental numbers. This tendency may lead either to karyotypes with mostly acrocentric chromosomes (e.g., giant shrews with 2N = 50 and FN = 66, Maddalena et al., 1989), or to mostly biarmed elements (e.g., C. wimmeri with 2N = 50 and FN = 84, Meylan and Vogel, 1982). Exceptions to this general pattern are found in C. luna, C. cf. bottegi, and C. bottegi which retain a chromosomal formula close to the hypothesized ancestral type. In C. lusitania, the karyotype (2N = 38) consists mainly of metacentric chromosomes (Maddalena, \99Qb), and is thus interpreted as the result of secondary reduction in the diploid number by a Robertsonian process, whereas the fundamental number remains high (FN = 74) as in other African species. The divergent karyotypic evolution of Palearctic-Oriental and Afrotropical species is illustrated in Fig. 2 and Fig. 6. These figures show a major difference between species in these zoogeographic regions, and suggest the existence of two distinct lineages among Crocidura. This is in full agreement with the previous phylogenetic hypothesis derived from biochemical data (Maddalena, 1990a; Fig. 1). Also reinforcing this conclusion is the intermediate condition of C. russula from both karyological and biochemical points of view. With 2N = 42 and FN = 60, this species has the highest chromosomal formula among Palearctic species and thus is intermediate between those and the Afrotropical species (Fig. 6). The zoogeographic origin of C. russula is probably North Africa (Richter, 1970), from which it entered Europe by crossing the Strait of Gibraltar (Catzeflis, 1984; Vogel and Maddalena, 1987), unlike the four other European species which entered Europe from the east (Catzeflis 1984; Poitevin et al., 1986; Vogel et al., 1986; Maddalena and Vogel, 1990). This zoogeographic pattern is confirmed by the biochemical results, which set C. russula well apart from the other European taxa. Other North African species (e.g., C. whitakeri or C. tarfayaensis) should be investigated to see if they also possess a karyotype intermediate between Afrotropical and Palearctic species. Another contact zone between these zoogeographic regions, which lies in Arabia, should also be more intensively studied. Acknowledgments We are grateful to A.-M. Mehmeti for technical assistance and to M. Jotterand-Bellomo (Lausanne), L. R. Heaney (Chicago), R. Hutterer (Bonn), and P. Vogel (Lausanne) for critical comments on the manuscript. Literature Cited Bovey, R. 1949. Les chromosomes des Chiropteres et des Insectivores. Revue suisse de Zoologie, 56:371-460. Butler, P. M., and M. Greenwood. 1979. Soricidae (Mammalia) from the Early Pleistocene of Olduvai Gorge, Tanzania. Zoological Journal of the Linnean Society, 67:329-379. Catalan, J. 1984. Methodes genetiques appliquees a la systematique des Soricidae du midi de la France. Diplome EPHE, Montpellier, 96 pp. Catzeflis, F. 1984. Systematique biochimique, taxonomie et phylogenie des musaraignes d’ Europe (Soricidae, Mammalia). Unpublished Ph.D. dissert.. University of Lausanne, 164 pp. Catzeflis, F., T. Maddalena, S. Hellwing, and P. Vogel. 1985. Unexpected findings on the taxonomic status of east Mediterranean Crocidura russula auct. (Mammalia, Insectivora). Zeitschrift fiir Saugetierkunde, 50:185-201. De Hondt, H. A. 1974. Karyological studies of two insectivores of Egypt. Proceedings of the Egyptian Academy of Sciences, 25:171-174. Dippenaar, N. J. 1980. A taxonomic revision of the monax- dolichura, luna-fumosa, glassi and zaodon-poensis complexes of Afrotropical Crocidura (Mammalia: Soricidae). Unpublished Ph.D. dissert.. University of Natal, Pietermaritzburg, 814 pp. Ellerman, j. R., and T. C. S. Morrison-Scott. 1966. Checklist of Paiaearctic and Indian mammals 1758 to 1946. British Museum (Natural History) Publications, London, 810 pp. George, S. B. 1986. Evolution and historical biogeography of Soricine shrews. Systematic Zoology, 35: 153-162. 338 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 Grafodatsky, A. S., S. I. Radzhabli, A. V. Sharshov, and M. V. Zaitsev. 1988. Karyotypes of five Crocidura species of the USSR Fauna. Tsytologya, 30:1247-1255 (in Russian, English summary). Grafodatsky, A. S., S. I. Radzhabli, and M. V. Zaitsev. 1989. Chromosome relations in Insectivores. Pp. 661-662, in Abstracts of the Fifth International Theriological Congress (E. Capanna, and L. Boitani, eds.), Rome, 1061 pp. Harada, M., T. H. Yosida, S. Hattori, and S. Takada. 1985. Cytogenetical studies on Insectivora. III. Karyotype comparison of two Crocidura species in Japan. Proceedings of the Japanese Academy, Series B, 61:371-374. Heaney, L. R., and R. M. Timm. 1983. Systematics and distribution of shrews of the genus Crocidura (Mammalia, Insectivora) in Vietnam. Proceedings of the Biological Society of Washington, 96:115-120. Heim DE Balsac, H., and M. Lamotte. 1956. Evolution et phylogenie des Soricides africains. I) La lignee Myosorex- Surdisorex. Mammalia, 20:140-167. 1957. Evolution et phylogenie des Soricides africains. II) La lignee Sylvisorex-Suncus-Crocidura. Mammalia, 21:15-49. Heim de Balsac, H., and J. Meester. 1977. Order Insectivora. Pp. 1-29, in The Mammals of Africa: An Identification Manual (J. Meester and H. W. Setzer, eds.), Smithsonian Institution Press, Washington, D.C. Hood, C. S., and R. J. Baker. 1986. G- and C-banding chromosomal studies of bats of the family Emballonuridae. Journal of Mammalogy, 67:705-711. Hutterer, R. 1983. Crocidura grandiceps, eine neue Spitzmaus aus Westafrika. Revue suisse de Zoologie, 83:699-707. 1984. Status of some African Crocidura described by Isidore Geoffroy Saint-Hilaire, Carl J. Sundevall and Theodor von Heuglin. Annales du Musee Royal d’Afrique Centrale, Sciences Zoologiques, 237:207-217. Hutterer, R., G. S. Jones, O. L. Rossolimo, R. G. van G elder, AND S. Wang. 1982. Insectivora: Soricidae. Pp. 67-102, in Mammal Species of the World, A Taxonomic and Geographic Reference (J . H. Honacki, K. E. Kinman, and J. W. Koeppl, eds.), Allen Press, Inc. and The Association of Systematics Collections, Lawrence, Kansas, 694 pp. Hutterer, R., L. F. Lopez-Jurado, and P. Vogel. 1987a. The shrews of the eastern Canary Islands: A new species (Mammalia: Soricidae). Journal of Natural History, 21:1347-1357. Hutterer, R., E. Van der Straeten, and W. N. Verheyen. I9ilb. A checklist of the shrews of Rwanda and biogeographical considerations on African Soricidae. Bonner zoologische Beitrage, 38:155-172. Ivanitskaya, E. Yu. 1989. Constitutive heterochromatin and nucleolar organizer regions in karyotypes of some shrews (Soricidae, Insectivora). Genet ika, 25:1188-1198 (in Russian, English summary). Jenkins, P. D. 1976. Variation in Eurasian shrews of the genus Crocidura (Insectivora: Soricidae). Bulletin of the British Museum (Natural History) Zoology, 30:269-309. Krishna Rao, S., and N. V. Aswathanarayana. 1978. Karyological analysis of the Horsfield’s shrew from peninsular India. The Journal of Heredity, 69:202-204. Maddalena, T. 1990a. Systematics and biogeography of Afrotropical and Palaearctic shrews of the genus Crocidura (Insectivora: Soricidae): An electrophoretic approach. Pp. 297-308, in Vertebrates in the Tropics (G. Peters and R. Hutterer, eds.). Museum Alexander Koenig, Bonn, 424 pp. \990h. Systematique, evolution et biogeographie des musaraignes Afrotropicales et Palearctiques de la sous-famille des Crocidurinae: Une approchegenetique. Unpublished Ph.D. dissert.. University of Lausanne, 172 pp. Maddalena, T., and P. Vogel. 1990. Relations genetiques entre Crocidures mediterraneennes: Le cas des musaraignes de Gozo (Make). Vie Milieu, 40:119-123. Maddalena, T., A.-M. Mehmeti, G. Bronner, and P. Vogel. 1987. The karyotype of Crocidura flavescens (Mammalia, Insectivora) in South Africa. Zeitschrift fur Saugetierkunde, 52:129-132. Maddalena, T., P. Vogel, and W. N. Verheyen. 1989. The African giant shrews: Taxonomy and relationships. P. 674, in Abstracts of the Fifth International Theriological Congress (E. Capanna, and L. Boitani, eds.), Rome, 1061 pp. Meylan, a. 1966. Donnees nouvelles sur les chromosomes des Insectivores europeens (Mamm.). Revue suisse de Zoologie, 72:636-646. 1971. Chromosomes de Soricides de Cote d’Ivoire (Mammalia, Insectivora). Revue suisse de Zoologie, 78:603-613. Meylan, a., and J. Hausser. 1974. Position cytotaxonomique de quelques musaraignes du genre Crocidura au Tessin (Mammalia, Insectivora). Revue suisse de Zoologie, 81:701-710. Meylan, A., and P. Vogel. 1982. Contribution a la cytotaxonomie des Soricides (Mammalia, Insectivora) de I’Afrique occidentale. Cytogenetics and Cell Genetics, 34:83-92. Ohno, S., W. Becak, and M. L. Becak. 1964. X autosome ratio and the behavior pattern of individual X chromosomes in placental mammals. Chromosoma, 15:14-30. Orlov, V. N., and N. S. Bulatova. 1983. Mammalian comparative cytogenetics and karyosystematics. Doklady Akademii Nauk SSSR, Moscow, 405 pp. (in Russian, English summary). Orlov, V. N., N. S. Bulatova, and A. N. Milishnikov. 1989. Karyotypes of some Mammalian species (Insectivora, Rodentia) in Ethiopia. Pp. 95-109, in Ecological and Faunistic Studies in South- Western Ethiopia (V. E. Sokolov, ed.). The USSR Committee for the UNESCO Programme of Man and the Biosphere (MAB) and The Institute of Evolutionary Morphology and Animal Ecology of the USSR Academy of Sciences, Moscow, 326 pp. (in Russian, English summary). PoiTEViN, F., J. Catalan, R. Fons, and H. Croset. 1986. Criteres d ’identification et repartition biogeographique de Crocidura russula (Hermann, IISO) et Crocidura suaveolens (Palla.s 1811). Terre Vie, 41:299-314. Repenning, C. A. 1967. Subfamilies and genera of the Soricidae. United States Geological Survey Professional Paper, 565:1-74. Reumer, j., and A. Meylan. 1986. New developments in vertebrate cytotaxonomy. IX. Chromosome numbers in the order Insectivora (Mammalia). Genetica, 70:119-151. Richter, H. 1970. Zur Taxonomic und Verbreitung der palaearktischen Crociduren (Mammalia, Insectivora, Soricidae). Zoologische Abhandlungen aus dem staatlische Museum fiir Tierkunde in Dresden, 31:293-304. Ruedi, M., T. Maddalena, H.-S. Yong, and P. Vogel. 1990. The Crocidura fuliginosa species complex (Mammalia: Insectivora) in peninsular Malaysia: Biological, karyological and genetical evidence. Biochemical Systematics and Ecology, 18:573-581. Seabright, M. 1971. A rapid banding technique for human chromosomes. Lancet, 2:971-972. Tada, T., and Y. Obara. 1986. Karyological relationship between the Japanese house shrew, Suncus murinus riukiuanus and the Japanese white-toothed shrew, Crocidura dsinezumi chisai. Proceedings of the Japanese Academy, Series B, 62:125-128. Tembotova, F. a. 1983. Taxonomic weight of some features of 1994 MADDALENA AND RUEDI — Chromosomal Evolution in the Genus Crocidura 339 Crocidura species from Caucasus. Doklady Akademii Nauk, SSSR, Moscow, 184-185 (in Russian). Tsuchiya, K. 1987. Cytological and biochemical studies of Insectivora in Tsushima Island. Pp. 111-124 in Nature of Tsushima, Tsushima Natural Resource Research Report, Nagasaki Prefecture (in Japanese, English summary). Tsuchiya, K., T. H. Yosida, K. Moriwaki, S. Ohtani, S. Kulta- Uthai, and P. Sudto. 1979. Karyotypes of twelve species of small mammals from Thailand. Report of the Hokkaido Institute of Public Health, 29:26-29. Vogel, P. 1986. Der Karyotyp der Kretaspitzmaus, Crocidura zimmermanni Wettstein, 1953 (Mammalia, Insectivora). Bonner zoologische Beitrage, 37:35-38. 1988. Taxonomical and biogeographical problems in Mediterranean shrews of the genus Crocidura (Mammalia, Insectivora) with reference to a new karyotype from Sicily (Italy). Bulletin de la Societe Vaudoise des Sciences Naturelles, 79:139-148. Vogel, P., and T. MaddalENA. 1987. Note sur la repartition altitud inale et la frequence de la musaraigne musette {Crocidura russula yebalensis) au Maroc. Mammalia, 51:465-467. Vogel, P., T. Maddalena, and F. Catzeflis. 1986. A contribution to the taxonomy and ecology of shrews (Crocidura zimmermanni and C. suaveolens) from Crete and Turkey. Acta Theriologica, 31:537-545. Vogel, P., T. Maddalena, and S. Aulagnier. 1988. Lecaryotype de Crocidura bolivari Morales Agacino, 1934 (Mammalia, Soricidae). Revue suisse de Zoologie, 95:779-783. Vogel, P., R. Hutterer, and M. Sara. 1989. The correct name, species diagnosis and distribution of the Sicilian shrew. Bonner zoologische Beitrage, 40:243-248. Vogel, P., T. Maddalena, and M. Sara. 1992. The taxonomic status of Crocidura cossyrensis Contoli, 1989 and its relationship to African and European Crocidura russula (Mammalia: Insectivora). Israel Journal of Zoology, 38:424. VOLOBUEV, V. T. 1981. B-Chromosomes system of the mammals. Caryologia, 34:1-23. Yong, H.-S. 1971 . Chromosome polymorphism in the Malayan house shrew, Suncus murinus (Insectivora, Soricidae). Experientia, 27:589-591. Yosida, T. H. 1985. The evolution and geographic differentiation of the house shrew karyotypes. Acta Zoologica Fennica, 170: 31-34. Table 1. — Diploid number (2 N) , fundamental number (FN) and shape of the female (X) and male (Y) sex chromosome of all the known karyotypes q/" Crocidura species. The abbreviations used are M = metacentric, SM = submetacentric, ST = subtelocentric, and A = acrocentric. Species 2N FN X Y References Palearctic origin C. pergrisea 22 34 1 C. horsfieldi watasei^ 26 48 SM A 2 C. leucodon 28 56 M A 1,3 C. zimmermanni 34 44 M A 4 C. canariensis 36 56 M ST 5 C. siculct 36 56 M ST 6,16 C. suaveolens'^ 40 50 M A, SM 3,9,10 40,41,42 50,52,54 M A 7,16 39,40 50 M A 8 C. sibirica 40 50 M ST 1 C. lasiura 40 54 M A 11 C. dsinezumi 40 56 SM ST 2,12 C. russula 42 60 M A 5,7,16 Oriental origin C. horsfieldi^ 38 48 M SM 13 C. cf. malayana^ 38,39,40 62-68 SM M, SM 14 C. cf. f uliginose^ 40 54-58 SM SM 14 C. attenuata 50 66 15 Afrotropical origin C luna 36 56 M ST 16 C. cf. bottegf 36 56 16 C. lusitania 38 74 M A 16 C. bottegi 40 60 SM SM 17 C. nanilla 42 74 M 16 C. ebriensis^ 44 66,72 M A 16,17 C. crossei^ 44,45 72,73 M A 16,18 340 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 Table 1 (cont.) C. grandicep^ 46 68 C. nigrojuscd 48 78 C. olivier? 50 66 C. viaric^ 50 66 C. hirta 50 66 C. flavescens 50 74 C. nigeriae 50 76 C. wimmeri 50 84 C. theresae 50 82,86 C. luna macmillan^ 50 C. lamottei 52 68 C. poensis 52,53 70,72 C. hildegardeae 52 76 C. cf. gracilipes 52 86 C. fuscomurina 56 86 C. bicolor^ 60 M A 18 M A 16 M A 16,17,18,19 M A 16,20 M A 21 M A 22 M A 18 M ST 18 M A 16,17 23 M A 17 M A 16,17 M A 16 M 18 M A 16 23 References: 1. Grafodatsky et al. (1988) 13. Krishna Rao and Aswathanarayana (1978) 2. Harada et al. (1985) 14. Ruedi et al. (1990) 3. Meylan (1966) 15. Tsuchiya et al. (1979) 4. Vogel (1986) 16. Maddalena (1990b) 5. Hutterer et al. (1987a) 17. Meylan (1971) 6. Vogel (1988) 18. Meylan and Vogel (1982) 7. Meylan and Hausser (1974) 19. De Hondt (1974) 8. Tsuchiya (1987) 20. Vogel et al. (1988) 9. Tembotova (1983) 21. Maddalena et al. (1989) 10. Ivanitskaya (1989) 22. Maddalena et al. (1987) 11. Orlov and Bulatova (1983) 12. Tada and Obara (1986) 23. Orlov et al. (1989) Notes: a. Reumer and Meylan (1986) considered C. horsfieldi watasei from Japan and C. horsfieldi from India as different species. b. Vogel (1988) previously used the name C. caudata, now replaced by C. sicula (Vogel et al., 1989). c. Crocidura russula monacha, C. cypria, and C. gueldenstaedtii are conspecific with C. suaveolens (Catzeflis et al., 1985). d. Ruedi et al. (1990) applied these names for two species from peninsular Malaysia which were previously confused in a single taxon (Jenkins, 1976). e. This female shows a karyotype different from C. bottegi (Maddalena, 1990^), and represents a distinct species (R. Hutterer, in prep.). f. This species is mentioned in Reumer and Meylan (1986) as C. crossei, but after R. Hutterer (in litt.) the specimens from the Ivory Coast differ from C. crossei and probably represent a distinct species for which the name C. ebriensis is available. Variations of FN are probably caused by methodological differences in the preparation of karyotypes. g. Meylan and Vogel (1982) provisionally used the name C. cf. planiceps, but their voucher specimen probably belongs to C. crossei (R. Hutterer, in litt.). h. Named C. cf. nimbae by Meylan and Vogel (1982) but Hutterer (1983) referred this specimen to C. grandiceps. i. According to Hutterer et al. (1987/)) this name antedates C. zaodon in the sense of Heim de Balsac and Meester (1977) and Dippenaar (1980). j. The name C. olivieri includes the forms giffardi, kivu, rnanni, occidentalis, and spurrelli (Maddalena, 19906). k. The name C. bolivar i is a junior synonym of C. viaria (Hutterer, 1984). l. This subspecies may be referable to C. thalia (R. Hutterer, in litt.) m. Certainly not C. bicolor, but impossible to identify from text (R. Hutterer, in litt.). 1994 MADDALENA AND RUEDI — Chromosomal Evolution in the Genus Crocidura 341 Fig. 1 . — Phylogenetic relationships of African and Palearctic shrews derived from a cladistic analysis of electrophoretic characters; the outgroup (not shown) was Sylvisorex megalura (after Maddalena, 1990n). Fig. 2. — Scattergram of diploid and fundamental numbers of Crocidura species in Table 1. The geographic origin of the species correlates generally with their chromosome formulas. 342 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY Q. T- « •* O 1^ Q_ h- O CM O. ’” * o 2^ _J o ■ _J o ■ -I ts. O 5- s*> f o «<» o > » *r> Z> CO _j f , * => -> 00 3 ■ 3 ^ -J j«rs*4 < CM O ^ 5 ■ ^ 5> csrr? =5 (O CO ♦}*se; g ■ 3 DC ■ ^t* ^ {£) im ^ C75 DC T- .* g- g> ♦ f #4# O r,N Q_ O) i» ♦ 9 * Ql T- » « - O .- 0. •MMT O Q. ^*#1 < =# 2 X _J O CO * - O ^ ' C# o ^ m r d'' m o X '■ ' »4ff 3 T- _J T- ** 3 '■ ^:-lf 3 lO _J ’- 3 ■ < T- O T- «« 1 2 ^ '5^11 < (O O ^ •f^: 2 cn 2 X ■ 3 in CO 3 ■ CO s ■ 3 >< # CO 3 _ DC T- *fii g ^ 3 DC ■ 2® * 2 X #1' LU 1 5 */ *,« 2 ® O CL ■ ♦ * o Dl ^ < O r- ^ O 05 *■' * '"1? 3 ^ _l o ■ 3 _i > v)»A _l O ’■ 3 . -1 ci< 3 ^ CO ^ A 2 ^ < O . 2 ■ 3 . CO W ” =5 . CO /If* .u'^si RU 3;2( 3 DC ■ CC T" =5 - DC CO ■"«' .'4!^ ' —1 O in s m S>^ O in O 0. ■ _l o . f i OL PO 19 18 « O Q. CM « O W 3 _J CM 3 . _i *> • ■> »t O i« 3 CO i? V s ^ Q CD K«T5. ^ CO o 3 CO c c-1 < o ■ 3 ID CO ^ 3 DC ■ 4 f O’" • ^ D. CM «, f CO ” O CM 2- v 2 ? 2 ■ 2 ■ « ^ O O * Ry Q_ CM » i ■ , o° O ” _i o . _i O ■ o o 0 ::= < S U 1 1 II .2 ^ , o gO S I |2 -9 ^ G O Q CS G & ? s ^ O r ! O ^ < abO « ^ o 9 !- ON X o\ o ^ S2 - 5 t>5 o a 6C « .2 *3 0 TJ 2 1 S P-. NO. 18 1994 MADDALENA AND RUEDI — Chromosomal Evolution in the Genus Crocidura 343 Fig. 4.-—G-banded karyotype of C. cf. malayana from Southeast Asia. This karyotype has nearly the same banding pattern as the Paiearctic species in Fig. 3. Supernumerary B-chromosome is indicated by an arrow. Fig. 5. — Model of the hypothetical ancestral karyotype of Crocidura species. The precise diploid and fundamental numbers are not known but should be around 38 and 54-58 respectively. 344 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 Palearctic and Oriental clade with 2n = 22-40; FN = 34-68 \w Stability or decreasing chromosomal number African clade with 2n =42-60; FN = 66-86 and fundamental number Fig. 6. — Principal lines of chromosomal evolution in the genus Crocidura. According to this scheme, karyotypes with 36-40 chromosomes (e.g., C. bottegi, C. luna) represent primitive characters. The two main clades are defined by opposing chromosomal evolution, one leading to the Afrotropical species with high diploid and fundamental numbers and the other to the Palearctic-Oriental species with stable or low diploid number. Exceptions to this pattern are discussed in the text. PHYLOGENY AND DISTRIBUTION OF THE CROCIDOSORICINAE (MAMMALIA: SORICIDAE) Jelle W. F. Reumer Natuurmuseum Rotterdam, P. O. Box 23452, NL-3001 KL Rotterdam, The Netherlands Abstract Typical features distinguishing the subfamily Crocidosoricinae Reumer, 1987 from other groups of Soricidae are found in the lower incisor, in the lower fourth premolar, in the number of lower antemolars, in the mandibular condyle, and in the upper incisor. Crocidosoricinae are the most plesiomorphic shrews and they gave rise, during the Miocene, to the other subfamilies. A cladogram is presented showing the relationships to some other insectivore families. The Nyctitheriidae are placed as the sister taxon of Heterosoricidae, Pies io soricidae, and Soricidae; the Heterosoricidae as the sister taxon of Plesiosoricidae and Soricidae; and the Crocidosoricinae as the sister taxon of all other Soricidae. The Crocidosoricinae originated in Asia, immigrated to Europe at the beginning of the middle Oligocene, and peaked in diversity during the early Miocene. Climatic deteriorations around the early/middle Miocene boundary and the middle/late Miocene boundary, in combination with increased competition by more apomorphic forms, caused an impoverishment, a southward retreat and, finally, the extinction of the Crocidosoricinae. Introduction The subfamily Crocidosoricinae Reumer, 1987 was established for housing a number of morphologically related shrews from Oligocene and Miocene deposits, most of them located in Europe. In the original descriptive paper (Reumer, 1987) the taxonomic relationships were briefly discussed and a number of taxa were mentioned as possible members of the Crocidosoricinae. Ziegler (1989) greatly expanded our knowledge of the subfamily. Here the Crocidosoricinae will be placed on a firmer foundation, initially with a discussion about the morphological characteristics of the subfamily and the phylogenetic relationships between the Crocidosoricinae and other insectivore taxa. Next, the biostratigraphic and biogeographic distribution of the subfamily will be investigated, using published literature. The aim of this study is to summarize what is known of the early history of the Soricidae, thereby stimulating further research for the origins and phytogeny of the family. Materials and Methods This is primarily a literature study. Terminology follows Reumer, 1984, unless otherwise noted. Dental elements from the upper jaw are indicated by superscripted numbers (for example, P'* or M’); dental elements from the lower jaw (mandible) by subscripted numbers (for example, or Mj). For the compilation of Tables 2, 3, 4, and 5 and for the design of Fig. 2, 3, 4, and 5, I used the following literature sources (numbers in parentheses refer to source indications in Tables 2-5): Aguilar et al., 1986 (1); Agusti et al., 1984 (2); Baudelot, 1972 (3); Bohlin, 1942 (4); de Bruijn and Riimke, 1974 (5); Brunet et al., 1981 (6); Bulot, 1986 (7); Crochet, 1975 (8); Daams and Freudenthal, 1981 (9); Doben-Florin, 1964(10); Doukas, 1986(11); Engesser, 1972(12), 1980(13); Engesser et al., 1981 (14); Fahlbusch and Wu, 1981 (15); Gaillard, 1899 (16); Gibert, 1975« (17), \915b (18); de Giuli et al., 1987 (19); Heizmarm et al., 1989 (20); Heizmann and Fahlbusch, 1983 (21); Hugueney, 1974 (22), 1976 (23); de Jong, 1988 (24); Lavocat, 1951 (25), 1961 (26); Li et al., 1983 (27); Rabeder, 1978 (28); Remy et al., 1987 (29); Repenning, 1967 (30); Russell and Zhai, 1987 (31); Stehlin, 1940 (32); Storch, 1988 (33); Sulimski, 1969 (34); Tobien, 1939 (35); Viret and Zapfe, 1951 (36); Yanovskaya et al., 1977 (37); Zapfe, 1951 (38); Ziegler, 1989 (39); Ziegler and Fahlbusch, 1986 (40). The stratigraphic framework used in the analysis is mostly based on Schmidt-Kittler (1987) for the Oligocene and Mein (1990) for the Neogene (Mio-Pliocene). The Oligocene is subdivided into MP (Mammalian Paleogene) biozones, and the Miocene and Pliocene into MN (Mammalian Neogene) biozones. The early Oligocene is considered to span MP 18 (La Debruge) through MP 22 (Villebramar), the middle Oligocene spans MP 23 (Itardies) through MP 27 (Boningen), and the late Oligocene MP 28 (Pech du Fraysse) through MP 30 (Coderet). The Oligocene/Miocene boundary is placed at the MP 30-MN 1 boundary. The early Miocene as used in this article is defined as spanning MN 1 (Paulhiac) through MN 4 (La Romieu). The middle Miocene spans MN 5 (Pont Levoy) through MN 8 (Anwil), the late Miocene spans MN 9 (Can Llobateres) through MN 13 (El Arquillo), and MN 14-16 is the Pliocene. See Lindsay and Tedford (1990) for a correlation of names of North American, central European and southwest European mammal ages and stages. Morphological Characteristics Crocidosoricinae are part of the family Soricidae, which family is here considered sensu stricto, excluding the Heterosoricidae Viret and Zapfe, 1951. The Soricidae are characterized by being generally small-sized Insectivora, by possessing a deeply pocketed internal temporal fossa, by the lack of a zygomatic arch, and by having a dorsoventrally separated mandibular condyle. The subfamily Crocidosoricinae, established by Reumer (1987) has the following diagnosis: “The lower incisor is cuspulate and relatively small; the 345 346 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 posterior extension of its buccal face is not, or only slightly developed. The P4 is tetrahedron shaped; its wear surface is V-shaped; there is a (sometimes only weakly defined) posterior groove or sulcus. Sometimes 2, but usually 3-5 lower antemolars. Upper incisor not fissident. Condyle Crocidura-\i\ie, small, with articular facets that are only slightly separated.” A differential diagnosis was also given, in which the Crocidosoricinae were distinguished from the other subfamilies then recognized: “The Crocidosoricinae... can be distinguished from the other subfamilies of Soricidae by usually having 3-5 lower antemolars, and by the small lower incisor which lacks a strong buccal extension. They differ from the Soricinae and the Limnoecinae in having a tetrahedron- shaped, more or less symmetric P4, with a wear-surface that is V-shaped rather than surrounding a posterolingual basin, or comma-shaped. They differ from the Soricinae in having less strongly separated condylar facets. ” The recent resurrection (Reumer, 1992) of a fifth soricid subfamily, the Allosoricinae Fejfar, 1966 in addition to the Crocidosoricinae, the Crocidurinae, the Limnoecinae, and the Soricinae, requires an adaptation of this differential diagnosis, discussed later. Several morphological traits of the Crocidosoricinae need a short discussion. The lower incisor (Ijnf ) is generally short. In the normal crocidosoricine situation, the Ij„f is bicuspulate and the axes of crown and root are not parallel, but placed under a weak angle, e.g., Ziegler, 1989:table 2, fig. 2 (Ulmensia ehrensteinensis Ziegler, 1989); table 4, fig. 4 (Soricella discrepans Doben-Florin, 1964); and table 6, fig. 2 {Florinia stehlini [Doben-Florin, 1964]). The short crown results in only a slight posterior extension of Ij„f at the buccal side of the mandible. In the derived state, the Ij„f develops the tendency to elongate buccally and to proceed posteriad below the subsequent dental elements, in some cases reaching below the middle of Mj (e.g., in Blarinoides mariae Sulimski, 1959, a Ruscinian Soricinae: see Reumer, 1984:plate 26). The tetrahedron-shaped P4 is included in the diagnosis. The archetypical crocidosoricine P4 has a V-shaped or Y-shaped wear facet. From the tip, two ridges (posterior arms) run to the posterolingual and the posterobuccal comers of the tooth, enclosing a posterior sulcus. On both posterior arms a small terminal cuspule may be present. This is what Jammot (1983) calls a “type Myosorex" P4, named after the living crocidurine genus Myosorex. The posterolingual arm may become elongated (seen for example in the crocidosoricine Clapasorex sigei Crochet, 1975), or the posterobuccal (posterolabial) branch may become elongated (seen for example in the crocidosoricine Carposorex sylviae Crochet, 1975, and in an advanced form in the subfamily Soricinae; see Crochet, 1975; Reumer, 1987:fig. 2). If both posterior arms reduce, the wear facet becomes triangular, a situation seen in most species of the subfamily Crocidurinae. There are exceptions to this rule. The simple fact that Jammot (1983) called the typical crocidosoricine P4 after a living Crocidurinae {Myosorex) indicates that in this genus apparently the archetypical P4 remained. It seems better not to speak of a “type Myosorex" P4, but of a crocidosoricine type of P4, when referring to the type of P4 with two posterior arms provided with accessory cuspules. A character not mentioned before is the double-rooted nature of the P4, although this too is not strictly diagnostic. Most Crocidosoricinae have a double-rooted P4, the only exception being Florinia stehlini (Doben-Florin, 1964) (Ziegler, 1989). On the other hand, members of other subfamilies also may have a double-rooted P4, such as the living genus Surdisorex (Crocidurinae); the fossil Paenelimnoecus micromorphus (Doben-Florin, 1964) and P. crouzeli Baudelot, 1972 (Allosoricinae); and Antesorex compressus (Wilson, 1960) (an Arikareean Soricinae; see Repenning, 1967; Reumer, 1992). Another character of the Crocidosoricinae is the possession of generally more than two lower antemolars. The number may vary from three to five, with only one exception — the early Miocene Miosorex pusilliformis (Doben-Florin, 1964) had only two lower antemolars. This is apparently an early “Reduktionsstadium” (Ziegler, 1989). On the other hand, several shrews are known that possess three lower antemolars and that do not belong to the Crocidosoricinae: e.g., Alluvisorex arcadentes Hutchison, 1966, a Barstovian Soricinae from Oregon; Angustidens vireti (Wilson, 1960), an early Miocene Limnoecinae from Colorado; Limnoecus niobrarensis Macdonald, 1947, a Barstovian Limnoecinae from Nebraska; and the genus Myosorex, a living Crocidurinae from Africa (Repenning, 1967). As this latter genus was also found to have retained the crocidosoricine type of P4, it may be considered a plesiomorphic Crocidurinae or a living relic of the Crocidosoricinae (see also Reumer, 1987). It would be worthwhile to study its isozyme and chromosome constitution in order to reveal its relation to other Crocidurinae. The mandibular condyle of the Crocidosoricinae is relatively small, it has upper and lower articular facets, and the interarticular area is not very large. This type of condyle is retained in most Crocidurinae and in the American Limnoecinae, albeit sometimes larger. The Soricinae are marked by a development of the condyle in which the interarticular area becomes larger dorsoventrally, the separation of the facets thus becoming more pronounced, and in which the interarticular area shows a lingual (medial) emargination. In the Allosoricinae, the upper articular facet acquires a triangular form. The upper incisor is invariably single tipped. The development of fissident upper incisors, which can be found in many Soricinae (tribes Soriculini, Beremendini, some Soricini), in Allosorex (Allosoricinae), but also in the Heterosoricidae, is apparently an apomorphic feature that has a polyphyletic origin. In the framework of the present study the presence or absence of pigmentation in the teeth is ignored. Most authors, when describing shrews, mention the presence or absence of pigment as a diagnostic feature. I think that in fossils this character must be treated with utmost care. As a result of the fossilization processes the original absence or presence of a pigment of iron-containing compounds in the enamel may be impossible to reconstruct. 1994 REUMER — Crocidosoricinae Phylogeny and Distribution 347 Relationships Because the Crocidosoricinae are by far the earliest Soricidae, they are critical in discussing the relationship of the Soricidae to other insectivore families and the relationships within the Soricidae. Reumer (1987) hypothesized that the Soricidae (Crocidosoricinae) sprung off from Eocene or early Oligocene Nyctitheriidae, independently from the Heterosoricidae. Nyctitheriidae and Soricidae share several derived characters: elongate and cuspulate lower incisors and a complex condylar facet. The Nyctitheriidae furthermore have a P"* much like the soricid P'*, and the alveolar pattern of the lower antemolars is identical to that of the Crocidosoricinae; all elements are single rooted, except for the P4, which is double rooted (Sige, 1976; Butler, 1988). On the other hand, there appear many differences: Nyctitheriidae have a zalambdodont dentition, Soricidae are dilambdodont; Nyctitheriidae have strongly molarized premolars, which Soricidae lack; there is a hypoconulid in the nyctitheriid lower molars, which is absent in shrews. Butler (1988) places the Soricidae (sensu lato) far from the Nyctitheriidae in his postulated phylogenetic tree, and considers the Plesiosoricidae a sister taxon of the Soricidae. These two families share many derived characters: dilambdodont dentition, large first incisors, reduction of the rest of the antemolar dentition, reduction or loss of the hypoconulid, the emphasis on P^-Mj shear, and, this is important, the loss of the zygomatic arch (Butler, 1988; Sig^, 1976). Butler (1988:132, fig. 5.6) places the Plesiosoricidae and the Soricidae as a sister group of a group comprising Dimylidae, Talpidae, and Proscalopidae. Sige (1976:68), however, considers the common characters of Plesiosoricidae and Soricidae as parallelisms, and states: “ ...Plesiosorex realise pendant I’Oligocene et le Miocene une adaptation parallelisant plus ou moins celle des soricides.” The Heterosoricidae share many characters with the Soricidae (to which family they were always considered to belong), but there is one important difference. Heterosoricidae do have a zygomatic arch (Gaillard, 1915; Viret and Zapfe, 1951), and lack the specialized masticatory apparatus of the Soricidae (see Repenning, 1967; Reumer, 1987). Some characters shared by the Heterosoricidae and the Soricidae may prove to be points of convergence. For example, the fissident upper incisor could well have different origins: the emergence of a median tine in the Soricidae versus a spadelike widening of the apex in Heterosoricidae. The condyle, which is of a complex shape in both families, shows dorsoventrally separated facets in the Soricidae, while in the Heterosoricidae the facets show also a conspicuous mediolateral separation. The fact that the molar morphology of the Heterosoricidae (weak posterior emargination, continuous endolophs) strongly resembles that of a highly advanced group of Soricinae (viz., the group with the genera Petenyia and Blarinella) can only be explained as a convergence. The presence of many possible convergences hinders the construction of a clear and unambiguous phylogenetic tree or cladogram. It is often impossible to decide whether a certain character (for example, the loss of the zygomatic arch or the formation of a complex condyle) is either a shared derived character (synapomorphy) or a case of parallelism (convergency). Thus, as a working hypothesis only, the following cladogram is proposed (Fig. 1). The nodes are characterized by the following hypothesized synapomorphies: A: somewhat enlarged first incisors, often serrate (cuspulate); complex condyle; lower antemolar alveolar pattern: all single rooted except for double- rooted P4. B: demolarization of antemolars except P**; reduction of antemolars; strongly enlarged first incisors; dilambdodont upper molars; loss or reduction of hypoconulid; P'^-Mj shear. C: loss of zygomatic arch. D: internal temporal fossa; dorsoventral separation of condylar facets. Reumer (1987, 1989, 1992) hypothesized that the Crocidosoricinae were ancestral to the other soricids. The emergence of other subfamilies began in the Miocene. The oldest record of a Limnoecinae is Angustidens vireti (Wilson, 1960) from the early Miocene (Arikareean) of Colorado (see Repenning, 1967). The oldest record of an Allosoricinae is also from the early Miocene (Orleanian): Paenelimnoecus rnicromorphus (Doben-Florin, 1964) from several localities in Bavaria, Germany (Ziegler, 1989; Reumer, 1992). The earliest record of a Soricinae is not, as mentioned before (Reumer, 1989), Paenelimnoecus crouzeli Baudelot, 1972, because this is an Allosoricinae (Reumer, 1992). The oldest mention of a possible Soricinae in Europe is ?Hemisorex sp. from the early Miocene (Orleanian, MN3-4) of Stubersheim 3 in Germany (Ziegler, 1989). An unambiguous Soricinae is Hemisorex robustus Baudelot, 1967 from the middle Miocene (MN6) of Sansan in France (Baudelot, 1972); this taxon is already highly developed and seems to stand at the base of the lineage leading to Blarinella and Petenyia. In fact, a comparison of the type material of H. robustus and Blarinella {B. dubia [Bachmayer and Wilson, 1970] and B. europaea Reumer, 1984) could reveal that they are congeneric. Antesorex compressus (Wilson, 1960) seems the earliest American record of a Soricinae; it is from the late Arikareean (late early Miocene) of Colorado (see Repenning, 1967). However, the specimens need to be examined to decide their subfamiliar assignment. The possession of a double-rooted P4 in A. compressus could indicate an affinity to the Crocidosoricinae. The oldest recorded Crocidurinae was supposedly a late Miocene Crocidura sp. from Kenya (see Repenning, 1967). It was described by Butler and Hopwood (1957), but Butler (1985) subsequently suggested that it was probably of Pleistocene age. No other Miocene Crocidurinae have been described; the origin of this subfamily is still unknown. Morphologically, the Crocidosoricinae can be distinguished from the Limnoecinae by the shape of the P4, which possesses a comma-shaped wear facet in the Limnoecinae; better separating criteria do not seem to be present, as the majority of limnoecine species have three or four lower antemolars and possess a crocidosoricine type of condyle, albeit rather large. The Crocidosoricinae are distinguished from the Allosoricinae most readily by the absence or extreme reduction of the entoconids and entoconid crests (entocristids in the terms of Ziegler, 1989) in the latter subfamily, and by the different 348 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 condyles; allosoricines have a triangular upper condylar facet. The oldest Allosoricinae retain some crocidosoricine features, such as the double-rooted P4. The difference between Crocidosoricinae and Soricinae is found in the P4 (see above and Reumer, 1987), in the condyle (soricines having well -separated articular facets), and in the lower number of lower antemolars (only Alluvisorex arcadentes Hutchison, 1966 retains a minute P3, see Repenning, 1967). Also, soricines have a more elongate lower incisor, and many have a fissident upper incisor. The difference between Crocidosoricinae and Crocidurinae is less pronounced. In general, the crocidurines have a simpler P4, lacking accessory cuspules and a posterior sulcus. Also the number of lower antemolars is lower, but, as already noted above, the living crocidurine Myosorex is an exception to the rules. Biostratigraphy and Biogeography General Remarks For a (literature) study of the biostratigraphic and paleobiogeographic distribution of the Crocidosoricinae, a complete listing of assigned taxa is necessary. Table 1 gives a compilation of the genera and species used in the present study, including some taxa that are still of uncertain generic assignment. I do not want to claim the list as being complete and exhaustive. Considerable changes in the list will no doubt occur or be necessary, for the following reasons. Many articles give faunal lists of a given locality in which the Soricidae are mentioned as “Soricidae indet.,” “Soricidae species A and B,” or the like. Such records have been excluded from this study, except when authors explicitly claim close relationship to known crocidosoricine taxa. Therefore, many records may have been “overlooked.” Furthermore, the identification of species often needs to be viewed with some caution. Certainly, researchers of the last few decades, such as Baudelot, Hugueney, Doben-Florin, Crochet, and Ziegler, who all described Crocidosoricinae, provided us with excellent descriptions and identifications, but many of the earlier records, whatever their morphology, were identified as “'Sorex.'' Likewise, Miocene shrews tended to be identified as either grivensis antiquus or “dehmi." Because I was not able to check all such identifications, I have for the present compilation adhered to the original designations. The appropriate species have, however, been assigned to the genera as currently used (see Ziegler, 1989, who revised a great part of crocidosoricine taxonomy). TTius, records of “Sorex” dehmi are treated in the lists as Lartetium dehmi. Finally, maps such as the ones reproduced here (Fig. 2-5) more often show the current geographical distribution of vertebrate paleontologists rather than a trustworthy distribution of the taxa under consideration. This produces a bias in the distribution maps. Oligocene Figure 2 gives the distribution of Soricidae (Crocidosoricinae) in Oligocene deposits. Table 2 summarizes the records. The earliest published report (not seen) of a Soricidae is in the early Oligocene of Mongolia: an indeterminate shrew from the Ergilin-Dzo Svita (Yanovskaya, 1977, in Russell and Zhai, 1987). If the identification is correct, this find appears to underline the hypothesis of the Asiatic origin of the family, as stated by Butler (1985, 1988) and Reumer (1987). Sulimski (1969) described the middle Oligocene Gobisorex kingae from the Shand-Gol Svita in Mongolia (see also Russell and Zhai, 1987); its biostratigraphic correlation to the three recorded French middle Oligocene finds is unknown. The remains resemble Heterosoricidae, but are best compared to “primitive Miocene species of the genus Sorex...” (Sulimski, 1969:68). The sizes of the teeth are relatively large, but compare well to those of, for example, Soricella discrepans (see Ziegler, 1989). G. kingae has either five single-rooted antemolars, or four such elements including a double-rooted P4. In the latter case, the posterior root of P4 is larger than the anterior one, a situation also known in Ulmensia ehrensteinensis (Ziegler, 1989). As Sulimski (1969) put it, Gobisorex represents an earlier stage of soricine evolution showing some resemblances to heterosoricines. Until I can examine the material (which lacks a mandibular ascending ramus), I can make no definitive choice between Soricidae and Heterosoricidae, and hence accept the opinion of Sulimski that Gobisorex is an early soricid (“soricine”). The presence of at least four lower antemolars and the anterior position of the buccal side of Ij„f furthermore seem to justify inclusion into the Crocidosoricinae. The late Oligocene marks the onset of the blooming of the subfamily. An indeterminate find from Yindirte (Taben Buluk) in China (Bohlin, 1942; Russell and Zhai, 1987) and eight reports from Europe (France and Germany) indicate a proliferation of taxa. Two new genera (Crocidosorex and Ulmensia) are reported, indicating morphological diversification of the subfamily. A possible ninth record, of Mysarachne picteti Pomel, 1853 from Chauffours (Puy-de-D6me, France) was neglected due to the unavailability of the material and of the name (see Lavocat, 1951). Miocene Figures 3, 4, and 5 show the geographical distribution of the Crocidosoricinae in, respectively, the early Miocene, middle Miocene, and late Miocene. Tables 3, 4, and 5 list the records concerned. The early Miocene (MN zones 1-4) shows an extensive proliferation of Crocidosoricinae, both in terms of geographical distribution and in terms of taxonomical diversity. Figure 3 and Table 3 show the presence of eight genera with at least 16 species (plus indeterminate material or material of uncertain affinities), a total of over 60 records. The distributional emphasis (see Fig. 3) is situated in central Europe and France. Figure 4 and Table 4 show the situation in the middle Miocene (MN zones 5-8). There is one report from Asia (Xiacaowan in China; Li et al., 1983), but in Europe the diversity is considerably less than that of the early Miocene, although the geographical distribution seems somewhat enlarged with records from Turkey (three localities; Engesser, 1980) and from 1994 REUMER — Crocidosoricinae Phylogeny and Distribution 349 Morocco in northern Africa (Beni-Mellal; Lavocat, 1961). The distributional emphasis has shifted from central Europe to Spain. Miosorex grivensis is the most successful species, while, apart from the mentioned record from Xiacaowan, the entire genus Crocidosorex appears to have vanished. We have less than 40 records from the middle Miocene, including only two genera and four species (plus some indeterminate material). With the onset of the late Miocene, the subfamily Crocidosoricinae has disappeared almost completely (Fig. 5, Table 5). TTiere are only three late Miocene records, all of Miosorex grivensis from Spanish localities (Gibert, 1975/?; de Jong, 1988). All three are stratigraphically dated to the earliest late Miocene (MN 9); MN zones 10-13 are devoid of Crocidosoricinae. Pliocene Finally, there are three closely-linked records that can be attributed to the Pliocene: an island relic from three sites at Apricena in the Italian peninsula of Gargano, that once was an island (Fig. 5, Table 5). It is taxonomically indeterminate (de Giuli et al., 1987). L. Maul (personal communication, January 30, 1989) mentioned the report of Crocidosorex" from Akkulaevo in Bashkiria (Russia) by Sukhov, 1970. Apparently the taxon was only mentioned in one table on page 18 of Sukhov’s article without any description in the text. Because I have not seen this record, I ignore it beyond this brief mention. Another Pliocene record concerns Myosorex meini described by Jammot, 1977, as a Crocidurinae from three Pliocene localities: Seynes and Balaruc 2 in southern France, and lies Medas in northern Spain. Crochet (1986), who mentioned the taxon, cited Jammot (1977) who stated that M. meini and '^Sorex" dehmi have important affinities. This could imply that M. meini is a Crocidosoricinae. I have not seen the original material and Jammot ’s thesis is unobtainable. Apart from the fact that M. meini has never been published as a new taxon according to the rules of the International Code of Zoological Nomenclature and is therefore a nomen nudum, I prefer to adhere to Jammot’s original notion of the taxon as a Crocidurinae. It is therefore excluded from the present study. Discussion I stated earlier (Reumer, 1989) that the Soricidae probably came to Europe as part of the faunal immigration wave often called “Grande Coupure,” about 33 million years ago. In terms of MP (Mammalian Paleogene) biozonation the Grande Coupure is situated between MP 20 and 21 (Tobien, 1987). The oldest shrews from Europe date from the middle Oligocene of France. Srinitium marteli Hugueney, 1976 from Saint-Martin-de- Castillon is dated to biozone MP 23 (Schmidt-Kittler, 1987). An indeterminate shrew from Garouillas is correlated to the biozone d’Antoingt, biozone 10 according to Remy et al. (1987), which corresponds to MP zone 25. A find of Srinitium sp. from biozone 13 of the Pech du Fraysse, also in France, is still slightly younger (Remy et al., 1987); it is dated to either biozone MP 27 or MP 28. Engesser and Mayo (1987) mention a “soricid indet.” from the Swiss locality of Balm, dated to the assemblage zone of Balm, which is MP 22 (Schmidt-Kittler, 1987). A recent reinvestigation has shown the absence of Soricidae, however, in this locality (Reumer, in press). The results of the literature survey of the distribution of the Crocidosoricinae (Fig. 2-5; Tables 2-5) show the peak of crocidosoricine diversity in the early Miocene. After the early Miocene, a decline started, leading to a virtual extermination of the subfamily by the onset of the late Miocene (Vallesian, MN zone 9). This finding strongly contradicts my earlier hypothesis (Reumer, 1989) that the climatic deterioration at the Miocene-Pliocene boundary was responsible for the extinction of the Crocidosoricinae on the European continent. I drew a parallel with the climatic event at the Plio-Pleistocene boundary of 2.4 million years ago and its impact on soricid diversity, and with the Pleistocene glacial history. Because the Crocidosoricinae were almost entirely absent during the late Miocene, the climatic events at the Miocene-Pliocene boundary could, consequently, not have caused the extinction of the subfamily. Van der Meulen and Daams (1992) studied the paleoecological conditions in the early and middle Miocene, based primarily on the composition of rodent associations from Spanish localities. Their studies resulted in a Relative Humidity Curve and a Relative Temperature Curve for the interval studied (MN zones 3 through 9). A drop in relative humidity is noted in MN 4, followed by a considerable cooling event starting at the onset of MN 5. Both events, but especially the cooling event, coincide with the boundary between early and middle Miocene as used in the present study. I postulate that the considerable drop in crocidosoricine diversity, as noted above, is at least partly caused by the climatic deterioration of MN 4/5 (see Reumer, 1989, for the impact of climatic events on fossil Soricidae). A second cooling event is noted at the onset of MN 9, leading to temperatures lower than those recorded before. It coincides with an increase in relative humidity. As mentioned above, the Crocidosoricinae disappear almost entirely in MN 9, which marks the beginning of the late Miocene. It is here postulated that this extinction is at least partly caused by the MN 9 cooling event. Another development paralleled the climatic events of the Miocene: the gradual evolution of advanced types of shrews. As noted above, Limnoecinae and Soricinae are present in America from the early Miocene onwards; in Europe the Allosoricinae are present from the early Miocene and unambiguous Soricinae from the middle Miocene onwards. Such advanced forms must have lived in competition with the Crocidosoricinae; it is therefore likely that the Crocidosoricinae disappeared through the combined influence of climatic deteriorations and increased competition. Figures 3, 4, and 5 furthermore show a gradual shift of the major distributional areas into a southward direction. In the early Miocene, most Crocidosoricinae lived in central Europe; in the middle Miocene most records are from Spain and other Mediterranean regions; the few late Miocene and Pliocene 350 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 records are all from south European localities. This pattern suggests a gradual southward retreat. It parallels the phenomenon observed by Reumer (1984) for the Plio-Pleistocene genus Episoriculus and provides further evidence for the impact of climatic cooling trends on the distribution of fossil Soricidae. Conclusions The Soricidae most probably originated in Asia and came to Europe right after the early Oligocene “Grande Coupure.” Only a few records are known from Oligocene deposits, but then the shrews (subfamily Crocidosoricinae) bloomed during early Miocene times with the main distributional emphasis in central Europe. Climatic deteriorations around MN4/5 coincide with a decline in diversity, and the biogeographic focus moved southward to Mediterranean regions. A second climatic deterioration around MN 9 (earliest late Miocene) coincides with the virtual extinction of the Crocidosoricinae; only an island relic in Italy survived into the Pliocene. Increased competition by more apomorphic groups of shrews (in Europe mainly Allosoricinae and Soricinae) probably contributed to this extinction process. Acknowledgments The article has greatly improved as a result of discussions with A. J. van der Meulen (Utrecht), whom I gratefully acknowledge. J. de Vos (Leyden) critically read the manuscript and provided many helpful suggestions. R. K. Rose, B. Rzebik- Kowalska, and an anonymous colleague acted as reviewers, thereby improving the quality of the paper. Its contents remain my own responsibility. Literature Cited Aguilar, J.-P., M. Calvet, J.-Y. Crochet, S. Legendre, J. Michaux, and B. SlGE. 1986. Premiere occurrence d’un Megachiroptere Pteropodide dans le Miocene moyen d’ Europe (gisement de Lo Foumas-II, Pyrenees-Orientales, France). Palaeovertebrata, 16:173-184. Agusti, j., a. Moya-Sola, and j. Gibert. 1984. Mammal distribution dynamics in the eastern margin of the Iberian Peninsula during the Miocene. Paleobiologie continentale, 14:33-46. Baudelot, S. 1972. Rude des Chiropteres, Insectivores et Rongeurs du Miocene de Sansan (Gers). Unpublished Ph.D. dissert., Universite de Toulouse, Toulouse, France, 364 pp. Bohlin, B. 1942. The fossil mammals from the Tertiary deposit of Taben-buluk, western Kansu. Part I: Insectivora and Lagomorpha. Palaeontologica Sinica, New Series, 8a: 1-1 13. Brunet, M., M. Hugueney, and Y. Jehenne. 1981. Cournon-Les Soumeroux: un nouveau site a vertebres d’Auvergne; sa place parmi les faunes de I’Oligocene superieur d’Europe. Geobios, 14:323-359. Bulot, C. 1986. Distinction de deux niveaux fossiliferes dans le gisement traditionnel de La Romieu (Gers). Bulletin du Museum national d’Histoire naturelle, 4e serie, section C, 4:483-497. Butler, P. M. 1985. The history of African insectivores. Acta Zoologica Fennica 173:215-217. 1988. Phylogeny of the insectivores. Pp. 117-141, in The Phylogeny and Classification of the Tetrapods (M. J. Benton, ed.). The Systematics Association Special Volume, 35B, volume 2, x + 329 pp. Butler, P. M., and A. T. Hopwood. 1957. Insectivora and Chiroptera from the Miocene rocks of Kenya Colony. Fossil Mammals of Africa, 13:1-35. Crochet, J.-Y. 1975. Diversite des Insectivores Soricides du Miocene inferieur de France. Colloque international C.N.R.S., 218:631-652. 1986. Insectivores Pliocenes du Sud de la France (Languedoc- Rousillon) et du Nord-Est de I’Espagne. Palaeovertebrata, 16:145-171. Daams, R., and M. Freudenthal. 1981. Aragonian: The SUge concept versus Neogene Mammal Zones. Scripta Geologica, 62:1-17. DE Bruun, H., and C. G. Rumke. 1974. On a peculiar mammalian association from the Miocene of Oschiri (Sardinia). I and II. Proceedings Koninklijke Nederlandse Akademie van Wetenschappen, Series B, 77:46-79. DE Giuli, C., F. Masini, D. Torre, and V. Boddi. 1987. Evolution of endemic mammal faunas in the Gargano Neogene (Italy): The problem of endemic variation as a chronological tool. Annales Instituti Geologic! Public! Hungarici, 70:137-140. DE Jong, F. 1988. Insectivora from the Upper Aragonian and the Lower Vallesian of the Daroca-Villafeliche area in the Calatayud- Teruel Basin (Spain). Scripta Geologica, Special Issue, 1:253-285. Doben-Florin, U. 1964. Die Spitzmause aus dem Alt-Burdigalium von Wintershof-Westbei Eichstatt in Bayern. Bayerische Akademie der Wissenschaften, Mathemathisch-naturwissenschaftlicheKlasse, Abhandlungen, Neue Folge, 117:1-82. Doukas, C. S. 1986. The mammals from the Lower Miocene of Aliveri (Island of Evia, Greece). Proceedings Koninklijke Nederlandse Akademie van Wetenschappen, Series B, 89:15-38. Engesser, B. 1972. Die obermiozane Saugetierfauna von AnwiI (Baseband). Tatigkeitsberichte der Naturforschenden Gesellschaft Baseband, 28:37-363. 1980. Insectivora und Chiroptera (Mammalia) aus dem Neogen der Tiirkei. Schweizerischen Palaontologischen Abhandlungen, 102:47-149. Engesser, B., and N. A. Mayo. 1987. A biozonation of the Lower Freshwater Molasse (Oligocene and Agenian) of Switzerland and Savoy on the basis of fossil mammals. Miinchner Geowissenschaftliche Abhandlungen, A 10:67-84. Engesser, B., A. Matter, and M. Weidmann. 1981. Stratigraphic und Saugetierfaunen des mittleren Miozans von Vermes (Kt. Jura). Eclogae Geologicae Helvetiae, 74:893-952. Fahlbusch, V., AND W.-Y. Wu. 1981. Puttenhausen: Eine neue Kleinsauger-Fauna aus der Oberen Susswasser-Molasse Niederbayems. Mitteilungen der Bayerischen Staatssammlung fiir Palaontologie und historische Geologic, 21:115-119. Gaillard, C. 1899. Mammiferes miocenes nouveauxou peu connus de La Grive-Saint-Alban (Isere). Archives du Museum d’Histoire Naturelle de Lyon, 7 b:l-78. 1915. Nouveau genre de musaraignes dans les depots miocenes de la Grive-Saint-Alban (Isere). Annales de la Societe linneenne de Lyon, 62:83-98. Gibert, J. 1975a. New insectivores from the Miocene of Spain. I and II. Proceedings Koninklijke Nederlandse Akademie van Wetenschappen, B, 78:108-133. 1975fc. Distribucion bioestratigrafica de los Insectivoros del Mioceno en el NE. de Espana. Biotopos, comparacion de cuencas y localidades. Relaciones faunisticas con America del Norte. Acta Geologica Hispanica, 10:167-169. Heizmann, E. P. j., and V. Fahlbusch. 1983. Die mittelmiozane 1994 REUMER — Crocidosoricinae Phylogeny and Distribution 351 Wirbeltierfauna vom Steinberg (Nordlinger Ries). Eine Uebersicht. Mittcilungen der Bayerischen Staatssammlung fiir Palaonto logic und historische Geologic, 23:83-93. Heizmann, E. P. J., G. Bloos, R. Bottcher, J. Werner, and R. Ziegler. 1989. Ulm-Westtangente und Ulm-Uniklinik: Zwei neue Wirbeltier-Faunen aus dem Untercn Susswasser-Molasse (Untermiozan) von Ulm (Baden-Wurttemberg). Stuttgarter Beitrage zur Naturkunde, Serie B (Geologic und Palaonto ligie), 153:1-14. HUGUEiNEY, M. 1974. Gisements de petits mammileres dans la region de Saint-Gcrand-le-Puy (Allier): Stratigraphic relative. Revue scientifique Bourbonnais, Moulins, 1974:52-68. 1976. Un stade primitif dans revolution des Soricinae (Mammalia, Insectivora): Srinitium marleli nov. gen., nov. sp. de rOligocene moyen de Saint-Martin-de-Castiilon (Vaucluse). Comptes rendus. Academic des Sciences, Paris, serie D, 282:981-984. Hutchison, J. H. 1966. Notes on some Upper Miocene shrews from Oregon. Bulletin, Museum of Natural History, University of Oregon, 2:1-23. Jammot, D. 1977. Les musaraignes (Soricidae, Insectivora) du Plio-Pleistocene d’Europe. Unpublished Ph.D. dissert., Universite de Dijon, Dijon, France, 314 pp. 1983. Evolution des Soricidae. Symbioses, 15(4):253-273. Lavocat, R. 1951. Revision de la faune des mammileres oligocenes d’Auvergneet du Velay. Ph.D. dissert., Universite de Paris, Paris, France, 153 pp. 1961. Le gisement de vertebres miocenes de Beni Melial (Maroc). Etude systematique de la faune de mammileres et conclusions generales. Notes et Memoires du Service Geologique du Maroc, 155:1-117. Li, C., Y. Lin, Y. Gu, L. Hou, W. Wu, and Z. Qiu. 1983. The Aragonian vertebrate fauna of Xiacaowan, Jiangsu. 1. A brief introduction to the fossil localities and preliminary report on the new material. Vertebrata PalAsiatica, 21:313-327. Lindsay, E. H., and R. H. Tedford. 1990. Development and application of land mammal ages in North America and Europe, a comparison. Pp. 601-624, in European Neogene Mammal Chronology (E. H. Lindsay, V. Fahlbusch, and P. Mein, eds.), NATO AS I Series, A 180, Plenum Press, New York, 658 pp. Mein, P. 1990. Updating of MN zones. Pp. 73-90, in European Neogene Mammal Chronology (E. H. Lindsay, V. Fahlbusch, and P. Mein, eds.), NATO ASl Series, A 180, Plenum Press, New York, 658 pp. Rabeder, G. 1978. Die Saugetiere des Badenien. Pji. 467-480, in Chronostratigraphie und Neostratotypen, Miozan M4 Badenien (A. Papp, 1. Cicha, J. Senes, and F. Steininger, eds.), Verlag Slowakische Akademie der Wissenschaften, Bratislava. Remy, j. a., J.-Y. Crochet, B. Sige, J. Sudre, L. de Bonis, M. Vianey-Liaud, M. Godinot, J.-L. Hartenberger, B. Lange- Badre, and B. Comte. 1987. Biochronologie des phosphorites de Quercy: Mise a jour des listes fauniques et nouveaux gisements de mammiferes fossiles. Munchner Geowissenschaftliche Abhandlungen, A 10:169-188. Repenning, C. A. 1967. Subfamilies and Genera of the Soricidae. U. S. Geological Survey Professional Paper, 565, iv + 74 pp. Reumer, j. W. F. 1984. Ruscinian and early Pleistocene Soricidae (Insectivora, Mammalia) from Tegelen (The Netherlands) and Hungary. Scripta Geologica, 73:1-173. 1987. Redefinition of the Soricidae and the Heterosoricidae (Insectivora, Mammalia), with the description of the Crocidosoricinae, a new subfamily of Soricidae. Revue de Paleobiologie, 6:189-192. 1989. Speciation and evolution in the Soricidae (Mammalia: Insectivora) in relation with the paleoclimate. Revue suisse de Zoologie, 96:81-90. 1992. The taxonomical position of the genus Paenelimnoecus Baudelot, 1972 (Mammalia, Soricidae): A resurrection of the subfamily Allosoricinae. Journal of Vertebrate Paleontology, 12:103-106. In press. The insectivorous mammals (Marsupialia: Didelphidae; Insectivora: Nyctitheriidae; Chiroptera) from the early Oligocene of Balm, Switzerland. Eclogae Geologicae Helvetiae, accepted. Russell, D. E., and R.-J. ZHAI. 1987. The Paleogene of Asia: Mammals and stratigraphy. Memoires du Museum national d’Histoire naturelle, Serie C, 52, 488 pp. Schmidt-Kittler, N. (ED.). 1987. International Symposium on Mammalian Biostratigraphy and Paleo ecology of the European Paleogene. Munchner Geowissenschaftliche Abhandlungen, A 10:1-312. SiGE, B. 1976. Insectivores primitifs de I’Eocene superieur et Oligocene in ferieurd ’Europe occidentale. Nyctitheriides. Memoires du Museum national d’Histoire naturelle, Nouvelle Serie, Serie C, 34, 140 pp. Stehlin, H. G. 1940. Zur Stammesgeschichte der Soriciden. Eclogae Geologicae Helvetiae, 33:298-306. Storch, G. 1988. Insectivora (Mammalia) aus dem Kalktertiar (Oberoligozan-Untermiozan) des Mainzer Beckens. Geologisches Jahrbuch, A 110:337-343. Sukhov, V. P. 1970. Poznepliosenovye melkie mlekopitaiushchie Akkulaevskogo mestonakhozhdeniia v Bashkirii [Late Pliocene Small Mammals from Akkulaevo faunal locality in Bashkiria], Nauka Press, Moscow, 94 pp. (in Russian). SULIMSKI, A. 1969. On some Oligocene insectivore remains from Mongolia. Pp. 53-70, in Results of the Polish-Mongolian Palaeontological Expeditions, Part II (Z. Kielan-J aworowska , ed.), Palaeontologica Polonica, Volume 21. Tobien, H. 1939. Die Insektenfresser und Nagetiere aus der aquitanen Spaltenfullung bei Tomerd ingen (Ulmer Alb). Naturforschende Gesellsehaft Freiburg, Berichte, 36:159-180. 1987. The position of the “Grande Coupure” in the Paleogene of the Upper Rhine Graben and the Mainz Basin. Munchner Geowissenschaftliche Abhandlungen, A 10: 197-202. van der Meulen, a. j., and R. Daams. 1992. Evolution of early to middle Miocene rodent faunas in relation to long-term paleoenvironmental changes. Paleogeography, Paleoclimatology, Paleoecology, 93:227-253. ViRET, J., AND H. Zapfe. 1951 . Sur quelques Soricides miocenes. Eclogae Geologicae Helvetiae, 44:41 1-426. Yanovskaya, N. M., Ye. N. Kurochkin, and Ye. V. Devyatkin. 1977. The find loeality of Ergilin-Dzo; the stratotype for the lower Oligocene in southwestern Mongolia. Pp. 14-33, 163, 169, in Fauna, Flora i biostratigrafiya Mezozoya i Kainozoya Mongolii (R. Barsbold, et al., eds.), Sovmestnaya Soviet-Mongolian Paleontological Expedition, Tr. no. 4 (in Russian). Zapfe, H. 1951. Die Fauna der miozanen Spaltenfullung von Neudorf a.d. March (CSR). Oesterreichische Akademie der Wissenschaften , Sitzungsberichte Mathemathisch-naturwissenschaftliche Klasse, Serie 1, 160:449-480. Ziegler, R. 1989. Heterosoricidae und Soricidae (Insectivora, Mammalia) aus dem Oberoligozan und Untermiozan Siiddeutsehlands. Stuttgarter Beitrage zur Naturkunde, Serie B (Geologie und Palaontologie), 154: 1-73. Ziegler, R., and V. Fahlbusch. 1986. Kleinsauger-Faunen aus der basalen Oberen Susswasser-Molasse Niederbayerns. Zitteliana, 14:3-80. 352 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 Table 1.— Alphabetic list of the taxa included in the subfamily Crocidosoricinae Reumer, 1987. Carposorex sylviae Crochet, 1975 Clapasorex bonisi Crochet, 1975 C. sigei Crochet, 1975 Crocidosorex antiquus (Pomel, 1853) C piveteaui Lavocat, 1951 C. thauensis Crochet, 1975 Florinia stehlini (Doben-Florin, 1964) Gobisorex kingae Sulimski, 1970 Lartetium dehmi (Viret and Zapfe, 1951) L. petersbuchense Ziegler, 1989 L. prevostianum (Lartet, 1851) ‘‘"Limnoecus" truyolsi Gibert, 1975 Miosorex desnoyersianus (Lartet, 1851) M. grivensis (Deperet, 1892) M. pusilliformis (Doben-Florin, 1964) '"Oligosorex” bruijni Gibert, 1975 "'Sorex” collongensis Mein, 1958 “Sorex” gracilidens Viret and Zapfe, 1951 Soricella discrepans Doben-Florin, 1964 Srinitium marteli Hugueney, 1976 Ulmensia ehrensteinensis Ziegler, 1989 Table 2. — List of names and localities of Oligocene Crocidosoricinae used in the analysis. Numbers refer to the articles mentioned in the Materials and Methods section. + , type locality. Late Oligocene Crocidosorex cf. thauensis Eggingen Erdbeerhecke (Bavaria, Germany) 39 id. Eggingen Mittelhart (id.) 39 id. Ehrenstein 4 (id.) 39 id. Hochheim (Germany) 33 Crocidosorex sp. Coumon-les-Soumeroux (Puy-de-D6me, France) 6 Ulmensia ehrensteinensis Ehrenstein 4 (Bavaria, Germany) (-(-) 39 Ulmensia sp. Hochheim (Germany) 33, 39 indet. Yindirte (Taban Buluk, China) 4, 31 indet. Hochheim (Germany) 33 Middle Oligocene Gobisorex kingae Shand-Gol Svita (Mongolia) ( + ) 31, 34 Srinitium marteli St-Martin-de-Castillon (Vaucluse, France) (-I-) 23 Srinitium sp. Pech du Fraysse (Quercy, France) 29 indet. Garouillas (Quercy, France) 29 Early Oligocene indet. Ergilin-Dzo Svita (Mongolia) 31, 37 1994 REUMER— Crocidosoricinae Phylogeny and Distribution 353 Table 3. — List of names and localities of Early Miocene (MN1-MN4) Crocidosoricinae used in the analysis. Numbers refer to the articles mentioned in the Materials and Methods section. + , type locality; *, uncertain taxonomic affiliation. Crocidosorex antiquus id. id. id. id. id. id. id. Crocidosorex cf. antiquus Crocidosorex thauensis id. Crocidosorex piveteaui Crocidosorex sp. id. Montaigu-le-Blin (Allier, France) ( + ) Langy (id.) Saulcet (id.) Chaveroches (id.) St-Gerand-le-Puy (id.) Ulm Westtangente (Bavaria, Germany) Tomerdingen (Germany) Oschiri (Sardinia, Italy) Budenheim (Germany) Bouzigues (Herault, France) ( + ) Paulhiac (Lot-et-Garonne, France) Marcoin-pres-Volvic (Puy-de-D6me, France) ( + ) Ulm Westtangente (Bavaria, Germany) Weisenau (Germany) 8, 20, 35, 33, 20, 32 32 32 32 6 39 39 5 39 8 8 8 39 33 Miosorex grivensis id. id. id. id. id. id. Miosorex aff. grivensis id. Miosorex pusilliformis id. id. id. id. Miosorex cf. desnoyersianus Valdemoros lA (Calatayud, Spain) Munebrega 1 (id.) Torralba 1 (id.) Villafeliche 4 (id.) Valdemoros 3B (id.) La Romieu (Gers, France) Rembach (Bavaria, Germany) Forsthart (id.) Vieux-Collonges (Rhone, France) Wintershof West (Bavaria, Germany) ( + ) Stubersheim 3 (id.) Erkertshofen 2 (id.) Petersbuch 2 (id.) Navarrete del Rio (Teruel, Spain) Petersbuch 2 (Bavaria, Germany) 18 18 18 18 18 7, 9 40 40 9 10 39 39 39 9 39 Lartetiiim dehmi id. id. id. id. Lartetium petersbuchense id. Erkertshofen 2 (Bavaria, Germany) Rauscherod lb (id.) Rembach (id.) Forsthart (id.) Vieux-Collonges (Rhone, France) Petersbuch 2 (Bavaria, Germany) (-I-) Erkertshofen 2 (id.) 40 40 40 40 9 39 39 Florinia stehlini id. id. id. Wintershof West (Bavaria, Germany) (-!-) Petersbuch 2 (id.) Erkertshofen 2 (id.) Rauscherod Ib/lc (id.) 10 39 39 40 Carposorex sylviae Carposorex sp. id. id. Laugnac (France) (-I-) Paulhiac (id.) La Brete (Aquitaine, France) Stubersheim 3 (Bavaria, Germany) 8 8 8 39 Clapasorex sigei Bouzigues (Herault, France) ( + ) 8 Clapasorex bonisi Paulhiac (Lot-et-Garorme, France) (4-) 8 354 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 Table 3 (cont.) Soricella discrepans Wintershof West (Bavaria, Germany) ( + ) 10 id. Stubersheim 3 (id.) 39 id. Erkertshofen 2 (id.) 39 id. Petersbuch 2 (id.) 39 id. Budenheim Hessler (Hessen, Germany) 39 id. Navarrete del Rio (Teruel, Spain) 9 id. Chaveroches (Allier, France) 22, 39 id. Dolnice (Czechoslovakia) 39 Soricella cf. discrepans Ulm Westtangente (Bavaria, Germany) 20, 39 Soricella n. sp. Petersbuch 2 (id.) 39 indet. Stubersheim 3 (id.) 39 indet. Aliveri (Greece) 11 indet. Weisenau (Germany) 33 ‘‘‘‘ Oligosorex" bruijni* Villafeliche 2A (Calatayud, Spain) 17 id.* Ateca 1 ( + ) and 3 (id.) 17 '"Limnoecus" truyolsi* Villafeliche 4 (Calatayud, Spain) ( + ) 17 id.* Valdemoros 3B (id.) 17 “Limnoecus" sp.* Vieux Collonges (Rhone, France) 9 ""Sorex” collongensis* id. (id.) ( + ) 3 Table 4.— List of names and localities of Middle Miocene (MN5-MN8) Crocidosori cinae used in the analysis. Numbers refer to the articles mentioned in the Materials and Methods section. + , type locality; *, uncertain taxonomic affiliation. Crocidosorex sp. Xiacaowan (China) 27 Miosorex grivensis id. id. id. id. id. id. Miosorex aff. grivensis id. id. id. id. id. id. Miosorex desnoyersianus id. Armantes 4 and 7 (Calatayud, Spain) Las Planas 4A and 4B (id.) Manchones (id.) Arroyo del Val 6 (id.) Hostalets de Pierola (Valles Penedes, Spain) La Grive St. Alban (France) ( + ) Puttenhausen (Bavaria, Germany) Las Planas 5B, 5H, and 5K (id.) Valalto 2c (id.) Boijas (id.) Villafeliche 9 (id.) Alcocer 2 (id.) Toril (id.) Solera (id.) Lo Foumas II (Pyrenes Orientales, France) Sansan (Gers, France) ( + ) 18 18 18 18 2, 18 16, 36 15 24 24 24 24 24 24 24 1 3 Lartetium dehmi id. id. id. Lartetium dehmi africanus La Grive St. Alban (France) ( + ) 36 Puttenhausen (Bavaria, Germany) 40 Steinberg (id.) 21 Neudorf a.d. March ( = Devinska Nova Ves, Czechoslovakia) 28, 38 Beni Mellal (Morocco) 3, 26 1994 REUMER— Crocidosoricinae Phyiogeny and Distribution 355 Table 4 (cont.) Lartetium prevostianum Sansan (Gers, France) ( + ) 3 id. Lo Foumas II (Pyrenees Orientales, France) 1 id. Cases de Pene (id.) 1 indet. Vermes (Jura, Switzerland) 14 indet. Anwil (Baselland, Switzerland) 12 indet. (comparable to M. desnoyersianus) Steinberg (Bavaria, Germany) 21 indet. (comparable to M. grivensis) Sari Qay (Turkey) 13 id. Pa§alar (Turkey) 13 indet. (comparable to L. dehmi) (^andir (Turkey) 13 “’Sorex" gracilidens (*) ""Limnoecus” truyolsi (*) Neudorf a.d. March ( = Devinska Nova Ves, Czechoslovakia) ( + ) Las Planas 4 A (Calatayud, Spain) 28, 36, 38 17 Table 5. — List of names and localities of Late Miocene and Pliocene Crocidosoricinae used in articles mentioned in the Materials and Methods section. the analysis. Numbers refer to the Pliocene indet. Apricena F15, F21b, F21c (Gargano, Italy) 19 Late Miocene (MN9-MN13) Miosorex grivensis Can Ponsich (Valles Penedes, Spain) 18 Miosorex aff. grivensis Carrilanga 1 (Calatayud, Spain) 24 id. Pedregueras 2A (id.) 24 NYCTITHERIIDAE HETEROSORICIDAE PLESIOSORICIDAE CROCIDOSORICINAE other SORICIDAE Fig. 1.— Cladogram showing {postulated interrelationships of some insectivore higher taxa. The nodes are characterized by hypothesized synapxpmorphies: A — enlarged, often serrate first incisors; complex condyle; lower antemolars single rooted except for double-rooted P4; B — demolarization of antemolars except for F*; reduction of antemolars; strongly enlarged first incisors; dilambdodont upper molars; loss or reduction of hypoconulid; P'^-Mj shear; C— loss of zygomatic arch; D— internal temporal fossa; dorsoventral separation of condylar facets. 356 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 "O TJ T3 g go o 12 « § M ^ o O T3 oT 1) C M 4) ij 8 -2 60 "O O o >. 8 t: ‘-' S § l: o (U > ii o TJ O (0 .2 g S 8 ja 0) § ^ C4) C4) •S .s ^ & o o ^ JS Vi w 4> U a O4 o o V- ti 3 3 W U) : o n T3 so 0) 60 SO IS IS A PRELIMINARY ANALYSIS OF BIOGEOGRAPHY AND PHYLOGENY OF CROCIDURA FROM THE PHILIPPINES Lawrence R. Heaney* and Manuel Ruedi^ *Field Museum of Natural History, Roosevelt Road at Lake Shore Drive, Chicago, Illinois 60605; ^Institut de Zoologie et d’Ecologie animale, Universite de Lausanne, 1015 Lausanne, Switzerland Abstract White-toothed shrews of the genus Crocidura from the Philippine Islands were compared with species from mainland Southeast Asia and the continental shelf islands of the Sunda Shelf. External and cranial variation indicate that three groups of taxa are present in the Philippines: a population of C. attenuata from the Batanes Islands; a group of medium-sized species including the named taxa C. beatiis, C. grayi, C. halconus\ and a group of large species, including C. grandis, C. mindorus, C. negrina, and C. palawanensis . One external and three dental synapomorphies are shared by C. heatus, C. grayi, and C. halconus, and at least one cranial synapomorphy is shared by this group and C. mindorus, C. negrina, and C. grandis, indicating that all six may form a monophyletic group restricted to the Philippines. Four synapomorphies unite all of these species with the C. Juliginosa group of continental Southeast Asia, rather than with the C. attenuata group. Analysis of allozyme variation in three Philippine populations (C. beatus, C. grayi, and C. mindorus from Sibuyan) and three reference populations from the Malay Peninsula revealed close relationships between Philippine populations and one member of the widespread Asian C. fuiiginosa group (provisionally referred to C. cf. malayana). Sister-taxon relationship between C. beatus and C. grayi was supported by genetic distance data and cladistic analysis of allelic characters. Results of genetic analyses were strongly concordant with those of morphological analyses. Crocidura occur on virtually every Philippine island, including geologically Recent oceanic islands, indicating that they colonize well. The distribution of individual species corresponds strongly to the extent of land areas during the late Pleistocene when sea level was 120 m lower than at present. Current patterns of diversity of shrews in the Philippines have probably resulted from three or four separate colonizations from Asia and subsequent speciation which involved both vicariance and colonization Introduction The Philippine Islands support a mammal fauna that is remarkable for its diversity and level of endemism; of the roughly 170 species, over 100 species (59%) are endemic (Heaney et al., 1987). These figures exceed even Madagascar, where about 80 mammal species are endemic (Jenkins, 1987:63). This exceptional degree of diversity is the result of both colonization from outside sources and phylogenetic diversification within the archipelago (Heaney, 1986, 1991; Heaney and Rickart, 1990). However, the number of individual cases is small in which there is documentation of the origin of the Philippine fauna and of its subsequent history within the archipelago. Moreover, statements about distributional patterns and levels of endemism depend on accurate information concerning taxonomic and geographic limits of species. When taxonomic status is arbitrary (that is, when actual species are not recognized or minor geographic variants are mistakenly recognized as species), patterns of endemism may easily be misinterpreted. In the Philippines, many species and genera have been reviewed recently (see references in Heaney et al., 1987), so that general patterns are discemable. However, several genera have never been reviewed in a systematic fashion, and these may influence current interpretations of the origin and evolution of the Philippine mammalian fauna. Equally importantly, phylogenies have been proposed for only a few groups, limiting our ability to interpret patterns of speciation and diversification. Foremost among these diverse but unstudied groups are the shrews of the genus Crocidura. The purposes of this paper are: 1) to determine if any Philippine Crocidura are conspecific with those on the adjacent Asian continent, 2) to define taxonomic events. Sympatry is restricted to members of different species groups. and geographic limits of species in the archipelago, 3) to analyze information concerning phylogenetic relations of the species, and 4) to discuss the evolutionary biogeography of the Philippine shrew fauna. Unfortunately, only small samples of conventional specimens (i.e., skins and skulls) are available from most taxa considered in this study. Compounding this problem is a paucity of postcranial skeletons and fluid-preserved specimens and of frozen tissue for biochemical studies, making it especially difficult to define species limits and determine the phylogenetic relationships of these species. In this paper, we summarize all available information as a means of clarifying some past points of uncertainty, and of developing hypotheses for further study. Methods All twelve cranial measurements were taken by Heaney with dial calipers graduated to 0.05 mm or digital calipers graduated to 0.01 mm. Cranial measurements were defined in Heaney and Timm (1983). External measurements were taken from specimen labels, and so are likely to include substantial interobserver variation. Terminology for morphology of the R* follows Jenkins (1984), and for foot pads follows Brown and Yalden (1973). Principal components analyses were conducted based on both correlation and variance-covariance matrices of base- 10 logarithm-transformed cranial measurements, using SAS Version 6.03 (SAS Institute Inc., 1988). Results of the two analyses were nearly identical, and so only the results of the correlation matrix analysis are reported here. Cladistic analysis of morphological characters was conducted using PAUP (ver- sion 3.0; Swofford, 1990), using methods described in the text. Photographs were taken in an Amray Model 1810 scanning 357 358 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 electron microscope using untreated specimens. Micrographs were prepared of all Philippine taxa except C. grandis, which is known only from the holotype. Tissues of 11 freshly-trapped individuals from three populations were frozen in liquid nitrogen and later transferred to a freezer for storage at — 80°C. Biochemical analyses of allo- zyme variation using starch-gel electrophoresis was conducted by Ruedi for 32 loci using procedures described by Pasteur et al. (1987). These were compared to two Malaysian species of the C. fuliginosa complex recently analyzed by Ruedi et al. (1991). The first (C. cf. fuliginosa) is characterized by a relatively low fundamental number of chromosome arms, and is represented by one population from the Cameron Highlands of the Malay Peninsula (designated “CH” in figures; n = 8). The second (C. cf. malayana), characterized by a higher funda- mental number of chromosomes, is represented by a population from the mainland of the Malay Peninsula (designated “UG” in figures; n = 4) and by one from nearby Tioman Island (desig- nated “TI” in figures; n = 6). We used Crocidura russula (n — 10) from Switzerland (Maddalena, 1990) as an outgroup. Nei’s genetic distance (Nei, 1978) was calculated from allele frequencies in each population. Nei’s genetic distances were used to estimate a phylogenetic tree using the Fitch-Margoliash algorithm (Fitch and Margoliash, 1967) as implemented on the FITCH program of PHYLIP (Felsenstein, 1989), and tested for robustness to choice of taxa using the jackknife approach of Lanyon (1985). Phylogenetic relationships also were assessed by parsimony analysis of allozyme distribution, considering each allele as a binary character (Buth, 1984; Dowling and Brown, 1989). The most parsimonious tree was found by the branch-and-bound algorithm using Wagner parsimony (PENNY program of the PHYLIP 3.2 system; Felsenstein, 1989). Each branching point was tested for significance by bootstrapping characters (Felsenstein, 1985), giving the proportion of 100 replicated trees based on resampled data sets possessing the node in question. Results We approached the goals of this study through a series of steps that allowed us to test successive hypotheses regarding Philippine populations of Crocidura. As the first step, we used cranial morphometric variation to test the null hypothesis that Philippine shrews are not distinguishable from mainland or continental shelf populations of shrews, i.e., that they are conspecific. As the second step, we reexamined populations for which we could not reject the null hypothesis, investigating nonmetric variation in cranial, dental, and external characters. In the third step, we developed a hypothesis of phylogenetic relationships using morphological characters; and in a final step, we examined genetic variation in several populations as an independent test of our hypothesis of relationships. Cranial Morphometric Variation Sixty individuals from the Philippines with complete crania, along with 75 individuals of six Asian species from 1 1 localities, were included in a principal components analysis to assess overall mensural similarity and geographic variation. Cranial measurements and sample sizes for Philippine taxa are given in Table 1. The results (Fig. 1, Table 2) indicate that most variation (89%) is explained by a single axis. All measurements contribute approximately equally on this axis, indicating that it represents size variation. The second axis is weighted heavily by the breadth of the palate between M^s; this is the most direct measure of palatal width. Rostral length, rostral width, and condyle-to-glenoid length also contribute to this axis, although less strongly. Examination of the third and fourth axes yielded no interpretable pattern and explained little variation, and so they were not considered further. Inspection of Fig. 1 shows that most of the species included in this analysis fall into three groups readily defined by size. Crocidura horsfieldi, C. maxi, C. monticola, and C. suaveolens (G-K in Fig. 1) form a group of small species (low loadings on the first axis) that are unlike any populations in the Philippines. Sample sizes are small, but the magnitude of the differences suggests a high level of distinctiveness. We therefore reject conspecificity between these species and those in the Philippines. A second group (A-B, M-S, X in Fig. 1) includes populations of intermediate size. These are C. attenuata from Taiwan and Vietnam and shrews from the main body of the Philippines (falling within current definitions of C. beat us, C. grayi, and C. halconus), plus a population from the Batanes Islands, which lie midway between Luzon and Taiwan. Within this group, two clusters are apparent. Crocidura attenuata (A, B) scores higher on axis II than do the other species; in other words, C. attenuata has a broad palate and short rostrum. Only one specimen of C. beatus overlaps C. attenuata. The population of unknowns from Batan Island (X) lies closest to the cluster of C. attenuata, with some overlap. The second cluster (M-S) appears to be internally identical on the second axis but shows variation along the first axis, with individuals from Mindoro tending to be smallest, from Mindanao larger, from Luzon still larger, and from Bohol, Leyte, and Maripipi the largest. On this basis, we reject conspecificity of C. beat us /grayi /halconus (hereafter referred to as the “C. grayi group”) with any reference taxa, but fail to reject the hypothesis that the Batan population is conspecific with C. attenuata. The third group (C-F, T-W, Y-Z in Fig. 1) is made up of individuals of largest size. It consists of the C. fuliginosa group from Borneo and mainland Asia (C-F), and the named Philippine taxa C. grandis, C. mindorus, C. negrina, C. palawanensis , and an individual of unknown species from Sibuyan Island (T-W, Y-Z). Sample sizes are small for all Philippine taxa in this group, restricting the strength of the interpretations. Within the group, C. negrina (U) tends to be smaller than C. fuliginosa (C-F), whereas C. palawanensis (V) tends to score higher on the second axis than do most specimens of C. fuliginosa (F). Crocidura mindorus (T) and the unknown from Sibuyan (Y) fall near each other, but are indistinguishable from C. fuliginosa on these axes; C. gratuiis (Z) is also in the midst of this cluster. On this basis, we are unable to reject the null hypothesis of conspecificity of these large shrews with the C. fuliginosa group, although some differences are present in 1994 HEANEY AND RUEDI— Evolution of Philippine Shrews 359 two cases (C. negrina and C. palawanensis). Qualitative Variation Examination of specimens indicated that four suites of features varied substantially between the populations under study; interorbital breadth, posterior palatal breadth, shape of P^, and size and position of the pads of the hind feet. Interorbital Breadth . — Crocidura grandis , C. mindorus, and the unknown from Sibuyan all have exceptionally broad interorbital regions (Table 1); C. grandis is the most extreme. Crocidura negrina and members of the C. grayi group have moderately broad interorbital regions, and C. palawanensis , C. fuliginosa, and C. attenuata progressively less so. Posterior Palatal Breadth. — As noted in the cranial morphometric analysis, there is substantial variation in the shape of the posterior portion of the palate. In Crocidura attenuata from the mainland and Taiwan and the unknowns from the Batanes Islands, the portion of the palate that supports the molars is much broader, and progressively increases in breadth posteriorly, in comparison to specimens of similar size from the main body of the Philippines (the C. grayi group). This is readily visible in Fig. 2, and is apparent in measurements of the labial width across M“s and palatal width between M^s (Table 1). In C. fuliginosa, C. palawanensis, and the C. mindorus group the molar-bearing portion of the palate is most similar to that of the C. grayi group (Fig. 2); the palate is narrow relative to the size of the molars, as in the C. grayi group, rather than broad as in C. attenuata. Moreover, the line defined by the lingual margins of the upper molars is at a small angle to the midline, rather than substantially diverging posteriorly as in C. attenuata. Shape of P^. — As noted by prior authors (e.g., Jenkins, 1976, 1982, 1984; Heaney and Timm, 1983), the shape of P^ often varies substantially between species. Taxa included in this study vary in size and shape of the talonid heel, the degree of concavity of the posterior margin of the tooth, prominence of the parastyle, and in the degree of development of the lingual cingulum (Fig. 3, 4). Crocidura attenuata has the most distinctive P'^ of the taxa studied (Fig. 3, 4). The posterior margin of the tooth is highly concave and the talonid heel well-developed, so that a wide space is left between the P'* and M* except where they touch at the labial edge. The lingual cingulum is typically low and moderately narrow. The population from Batan Island has a P“* similar in shape, except that the posterior edge is slightly less concave. Additionally, the lingual cingulum is slightly broader but does not project as far anteriorly. Thus the tooth has low but conspicuous points at both the protocone and the anterior tip of the cingulum, rather than being nearly smoothly curved as is typical. The parastyle forms a moderately high and prominent cusp. Crocidura attenuata and the Batanes specimens have narrower molars (in the anterior-posterior axis) than all other populations in the study. At the opposite extreme in these respects are the shrews of the C. grayi group. The P"^ of C. grayi (Fig. 3, 4) has a very broad, smoothly curving talonid heel that leaves a narrow gap between it and the M*. Most of this expansive talonid is rimmed by the lingual cingulum. The angle formed by the talonid and the shearing edge of the paracone and distostyle (defined by Jenkins, 1984:66) is fairly abrupt (Fig. 3). The parastyle is low, forming only a small, rounded projection. Eight specimens from Mount Isarog in southern Luzon and four from Haights-in-the-Oaks in northern Luzon are very similar to the specimen shown in Fig. 3. One specimen shows slightly more expansion of the talonid and two slightly less. In one, the cingulum terminates just posterior to the protocone so that the protocone forms a more sharply defined angle in the edge of the tooth. Little if any of this variation is due to age, since the basal outline of the tooth is modified only in very old animals. Crocidura beatus and C. halconus are very similar to C. grayi. TTie latter differs only in being slightly smaller; the former has a slightly broader tooth in the labiolingual plane, and in particular has a slightly broader talonid region. The P“* of C. fuliginosa (Fig. 3, 4) is most similar to that of the C. grayi group. It differs in being larger, in having the entire lingual portion of the tooth broader, and in having the talonid slightly broader and more strongly expanded posterolingually. However, the parastyle is moderately large and distinct, much like that of C. attenuata. A series of 15 specimens of C. fuliginosa foetida from Mount Kinabalu, Sabah, exhibits little variation from the example in Fig. 3. A few have less of an anterolingual expansion, there is slight variation in the degree of concavity of the posterior edge, and the lingual edge is slightly convex in one specimen, rather than flattened or slightly concave. In comparison with C. fuliginosa, C. palawanensis has slightly less concavity to both the anterior and posterior edges of P^, and the protocone is placed slightly more anterolingually. The specimen of C. palawanensis shown in Fig. 3 has a narrower talonid than other specimens from Palawan and Balabac; in the others, the talonid is greater in area and more rounded, and less elongate on the labiolingual axis. The P* of the single specimen of C. grandis is indistinguishable from the series of C. fuliginosa foetida. None of the C. fuliginosa, C. palawanensis, or other members of the C. mindorus group has a P'* as small as that of C. negrina, and it appears that the parastyle is smaller in C. negrina than in other populations with the exception of the C. grayi group. The lingual edge of the P* of C. mindorus is convex in both known specimens (unlike the typical state in C. fuliginosa). Additionally, the talonid of the holotype of C. mindorus is greatly enlarged (Fig. 3) and is virtually identical to that of the unknown from Sibuyan. The paratype of C. mindorus has the talonid less enlarged, roughly equal to that of typical C. fuliginosa foetida. The P* of C. grandis is indistinguishable from that of C. fuliginosa foetida. On this basis, we are again unable to reject the hypothesis that the animals from the Batanes Islands are conspecific with C. attenuata. The members of the C. grayi group are distinct from both C. attenuata and the C. mindorus group. Within the C. grayi group, C. halconus does not differ from C. grayi, but C. beatus differs slightly. Within the C. mindorus group, although there are few specimens with which to assess variation, C. negrina is apparently distinct on the basis of small differences. Crocidura mindorus and the unknown from Sibuyan 360 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 are very similar to each other, but slightly different from all others. Crocidura fuliginosa and C. palawanensis differ subtly, and are most similar to the C. mindorus group. Hind Foot Pads. — Comparisons were made using all taxa for which fluid-preserved specimens are available: C. attenuata (from Taiwan), C. beat us, C. fuliginosa foetida (from Borneo), C. grayi, C. negrina, and the unknowns from Batan and Sibuyan. Several morphotypes are present. The first includes only C. attenuata from Taiwan and the unknowns from the Batanes. These have very short, broad feet with small, rounded plantar pads (Fig. 5). Pigmentation is relatively light, and the “granulae” (the small, often pigmented fleshy bumps) on the ventral surface of the foot are relatively inconspicuous. Hairs on the dorsal surface of the feet are short and comparatively small in diameter (i.e., fine rather than coarse). Animals from the Batanes tend to have plantar pads slightly smaller and closer together, but are otherwise identical to those from Taiwan. The second morphotype is represented by C. beatus and C. grayi (Fig. 5). Their hind feet are longer but of the same width as those of C. attenuata, so that they are proportionately narrower. They have heavily pigmented ventral surfaces, and the granulae are prominent. The plantar pads are of moderate size, with a tendency to be slightly elongate and flattened proximally. Specimens from Leyte, Maripipi, and Luzon are nearly identical. The distinctiveness of the foot morphology of this species group again causes us to reject its conspecificity with any other taxa. The third type is shown by C. fuliginosa foetida, C. negrina, and the Sibuyan shrew (Fig. 5). They are generally similar in having hind feet that are long but varying in width. The soles are moderately pigmented, and granulae are prominent. The plantar pads are more elongate than in the C. beatus type, but proximal flattening is equivalent. Within this group, C. fuliginosa has broad, heavy hind feet, and the unknown from Sibuyan has the narrowest, least heavy feet, with C. negrina intermediate. The thenar and hypothenar pads of the Sibuyan shrew are closest to the interdigital pads, so that the central area of the palm is smallest, and the thenar tends to be more elongate than in the other two species. We tentatively conclude that C. negrina and the Sibuyan animal differ from Bornean C. fuliginosa. Phylogenetic Analysis of Morphological Data Analysis of the above characters is limited by lack of a phylogenetic framework for Southeast Asian Crocidura as a whole. In particular, determination of the polarity of characters is complicated by uncertainty about the most appropriate outgroup. For consistency with genetic analyses (for which tissues samples are scarce), we selected C. russula from Switzerland as an outgroup for determining polarity of the characters listed in Table 3. Jenkins (1976) considered C. russula to be morphometrically “central” in an analysis of Eurasian shrews, unlike C. fuliginosa which was found to be distinctive. Phylogenetic analyses by Maddalena (1990) and Maddalena and Ruedi (1994), based on allozymes and karyotypes, respectively, have shown C. russula to be an early derivative of the clade of Palearctic shrews that also includes the C. fuliginosa group. The cladistic analysis of data in Table 3 found the shortest tree to be of 14 steps; there were 14 trees of this length. The large number of trees is associated with missing data for several characters and with a few unstable nodes, and the consistency index for the 14 trees is high (0.86). Fig. 6A shows a “semi- strict” consensus tree for the 14 shortest trees. That is, it includes those nodes that are not contradicted by alternative arrangements, and indicates the percentage of the 14 trees that supported a given node. The Batanes shrews are the sister group to Taiwan C. attenuata, with no characters that distinguish them. These two species are distantly related to all other taxa. Crocidura palawanensis, C. fuliginosa, and C. negrina form progressive sister taxa to the clade including other Philippine taxa. This clade is made up of an unresolved quadrichotomy, one branch of which is made up of the C. grayi group. A second perspective on the cladistic analysis is presented in Fig. 6B. This is a 50% majority-rule consensus of the 14 shortest trees. This analysis accepts nodes that are supported by 50% or more of the 14 trees, and indicates the percent of the trees that support a given node. It is very similar to the prior tree (Fig. 6A), except that C. palawanensis and C. fuliginosa are now part of an unresolved trichotomy. The quadrichotomy involving the C. grandis and C. grayi groups has been partially resolved, with the C. grayi group as the sister taxon to C. mindorus and the Sibuyan shrew, with C. grandis and C. negrina as successive outgroups. Table 3 shows that there are two synapomorphies that unite C. attenuata and the population from the Batanes Islands, but no derived characters shared by these two and any of the other taxa. Four synapomorphies (one dental and three from the hind foot pads) unite C. fuliginosa and C. palawanensis with the shrews from the Philippines. Within the remaining Philippine shrews, three synapomorphies (3b, 4b, and 7c) are shared by the C. grayi group, but there are no synapomorphies within the group. One character (5c; low degree of concavity of the posterior edge of P^) is either a synapomorphy for all Philippine shrews but reversed to 5b in C. negrina and C. grandis, or was developed independently in the C. grayi group and in C. mindorus and the unknown from Sibuyan. Crocidura grandis, C. mindorus, and the shrew from Sibuyan share one synapomorphy (exceptionally broad interorbital region), but there are no synapomorphies that unite these with C. negrina unless 5c is a true synapomorphy. Genetic Analyses The frequencies of alleles at 32 loci are presented in Table 4. The matrix of genetic distances and associated standard errors (Nei, 1978) derived from these data is presented in Table 5. Sample sizes are small, and all of the genetic analyses should be regarded as provisional, to be refined and expanded when more specimens are available. A Fitch-Margoliash analysis of the genetic distance matrix (Fig. 7A; see Methods) indicates that C. beatus and C. grayi are sister taxa, the Sibuyan shrew is their sister taxon, and the two populations of C. cf. malayana are their collective sister 1994 HEANEY AND RUEDI— Evolution of Philippine Shrews 361 taxon. Crocidura cf. fuliginosa was the sister taxon to all of these. Coding of these data for the presence or absence of each allele in Table 3 produced 63 binary characters, of which 21 are informative about ingroup relationships. A cladistic analysis of these data using boot strapping for 100 repetitions produced nearly the same relationships (Fig. 7B). The only difference was a shift of the Sibuyan shrew to a position as sister taxon to the two populations of C. cf. malayana rather than the C. grayi group. However, in both cases the genetic distance is small and the nodes are not strongly supported, so that we view these results as indicating an unresolved trichotomy between the Sibuyan shrew, the two populations Of C. cf. malayana, and the C. grayi group. These results suggest that the Philippine shrews and C. cf. malayana share a common ancestor which invaded the Philippines relatively recently. The other Malayan shrew (C. cf. fuliginosa) is apparently more distantly related to this clade, and apparently did not colonize the Philippines. The results of these analyses of allozyme characters (Fig. 7) are strongly concordant with those of morphological characters (Fig. 6). One important difference regards definition of members of the C. fuliginosa species group. In the genetic analysis (Fig. 7), it is apparent that one particular member of this group, C. cf. malayana, is most closely related to the Philippine shrews, whereas C. cf. fuliginosa is more distant. There is currently no documented morphological difference between C. malayana and C. fuliginosa, and they have been considered synonyms (Jenkins, 1984). Until species limits and relationships within the species group are resolved, we can make no further comment on this problem. In both genetically-based analyses, C. beatus and C. grayi are sister taxa, differing by several characters from the Sibuyan shrew and representatives of the C. fuliginosa group. The placement of the Sibuyan shrew is unresolved in the genetic analysis, but weakly defined as the sister taxon to the C. grayi group in morphological analyses. More specimens of nearly all taxa are needed to further resolve relationships. Species Limits The lack of characters that distinguish the population in the Batanes Islands from C. attenuata lead us to recommend that these be considered conspecific. Similarly, the number of synapomorphies and the absence of more than very slight differences between C. mindorus and the shrew from Sibuyan lead us to consider them to be conspecific, but we urge continued study of these shrews. Of the previously-named taxa in the C. mindorus group, we suggest no changes. The three species appear to be closely related, but differences in size, degree of pilosity of the tail, and shape of the F* (see species accounts) indicate separate evolutionary histories. We recommend that C. grayi and C. halconus be considered conspecific, as noted previously by Heaney et al. (1987). Although there are slight differences in size and shape of P'*, these are trivial and seem to represent only minor geographic variation. We do not suggest use of a trinomial because this would give a false impression of the precision of current knowledge. In spite of the close relationship between C. grayi and C. beatus, we do not recommend treating these as conspecific at this time. The differences in external, dental, and cranial morphology and in gene frequencies make it apparent that the sometimes subtle variation is consistent in indicating evolutionary independence. It is possible that more extensive data will show that they would best be recognized as subspecies, but we currently lack the information to make such fine distinctions. Synopsis of Relationships As a means of formalizing interpretation of the above results, we present Fig. 8 as our current hypothesis of relations for the Philippine shrews. All data indicate sister-group relationship for C. beatus and C. grayi. The morphological data show strong support (Fig. 6A, B) for the C. mindorus group as the sister taxon to the C. grayi group. One of the genetic analyses (Fitch-Margoliash; Fig. 7 A) supports this, and the other is uninformative in this respect. Relationships within the C. mindorus group are generally unresolved by the analyses. We thus place the C. mindorus group plus the C. grayi group as an unresolved quadrichotomy in Fig. 8 as a means of indicating both general relationships and the node that is most in need of further investigation. Crocidura palawanensis is so similar to members of the C. fuliginosa group that we can distinguish it only with difficulty (see species accounts), and so we place it as the sister taxon of C. fuliginosa (sensu lato). These two taxa are consistently identified in morphological and genetic analyses as the outgroup to the shrews from the main body of the Philippines (Fig. 6, 7), and we include them in that position in Fig. 8. Crocidura attenuata is unambiguously identified by morphological analyses as the most distant branch in phylogenetic analyses, and we show it so in Fig. 8. Discussion and Conclusions Biogeography The five endemic and two widespread species here recognized from the Philippine Islands occur throughout the archipelago (Fig. 9), on landbridge (Palawan and Balabac), old oceanic (Luzon, Mindanao, and associated smaller islands), and young oceanic islands (Batan, Negros, and Sibuyan). They have been found on every island on which they have been sought. Three endemic species are each confined to single large Pleistocene islands. Crocidura beatus, known from Biliran, Bohol, Leyte, Maripipi, and Mindanao, and C. grandis, known only from Mindanao, are restricted to the Pleistocene island of Greater Mindanao; C. negrina occurs on the Pleistocene island of Greater Negros-Panay; and C. palawanensis is known only from Palawan and Balabac, which were united as part of Greater Palawan during the late Pleistocene. Additionally, C. attenuata is known in the Philippines only from one island in the Batanes group; these were isolated from other islands during the Pleistocene. The Batanes Islands lie midway between Luzon and Taiwan. This represents the first 362 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 evidence of colonization of the northern Philippines from Taiwan (Heaney, 1986). Other species are more widespread. Crocidura mhidorus is now known from Mindoro, which remained separate from other islands during the Pleistocene, and the nearby small oceanic island of Sibuyan. Crocidura grayi occurs on Luzon and Catanduanes, which were joined as Greater Luzon, and on the adjacent island of Mindoro. Diversification Within the Philippines With seven species, Crocidura is the second most speciose genus of nonvolant mammal in the Philippines; only the endemic forest mice of the genus Apomys are more speciose (Musser, 1982; Heaney, 1986). This large number of endemic species is atypical of nonendemic genera in the Philippines. Of 15 nonendemic genera, only two {Crocidura and Rattus; 13%) have more than one species in the Philippines, whereas 53% of the endemic genera have more than one species. This difference implies that Crocidura and Rattus have unusually good colonizing ability, and this is supported by the presence of both genera on isolated oceanic islands within the Philippines and in the oceanic portions of the Indo-Australian archipelago to the south. Our conclusions regarding phylogenetic relationships and species limits (Fig. 8) carry implications regarding the likely historical processes of diversification within the Philippines. It appears that there have been three or four separate colonizations of the Philippines from the Asian mainland. Perhaps the most recent of these was the arrival of C. attenuata in the Batanes Islands, presumably from the nearby island of Taiwan. The deep water surrounding the Batanes and lack of evidence of dryland connection to Asia implies overwater colonization. Another recent colonization was that which produced C. pa~ lawanensis , a species only slightly different from C. fuliginosa. This occurrence is strongly concordant with the Pleistocene history of Palawan. The fauna of the Palawan island group shares very little with the main body of the Philippines. Instead, 96% of the genera and about 60% of the species of nonvoiant mammals are shared with Borneo and the rest of the continental shelf islands. This pattern is due to the probable presence of a land bridge between Palawan and Borneo during the middle Pleistocene (ca. 160,000 years ago), when sea level dropped to about 160 m below the present level. Water between Palawan and Borneo is about 140 m deep (Heaney, 1986). Sea level dropped only to 120 m below current sea level during the late Pleistocene, thereby keeping Palawan isolated, so that the two species probably have been disjunct for about 160,000 years. The three species of the C. tnindorus group probably represent a third colonization. Although they might have originated from a common ancestor with C. palawanensis , it is more likely that they are derived from the common ancestor of C. fuliginosa and C. palawanensis, since they share no characters with C. palawanensis that are not also shared with C. fuliginosa. This group has apparently diversified along the western rim of the Philippines, with each species now on a separate island. There is no evidence that these islands were ever connected as a single land mass (Heaney, 1986, 1991), so overwater colonization on three occasions is implied. Finally, the C. grayi group might have originated from the C. mindorus group within the Philippines, or from the common ancestor of C. fuliginosa, C. palawanensis, and the C. mindorus group. Portrayal of the relevant nodes on Fig. 8 indicates that the latter possibility is more likely, but the weakness of the relevant node does not allow firm conclusions. Once in the Philippines, they speciated between Greater Luzon and Greater Mindanao. Overwater colonization across the narrow San Bernardino Straits is the most likely mechanism, although the possibility of a land bridge across the straits cannot be entirely discounted (Heaney, 1986). Colonization from Luzon to Mindoro by C. grayi is likely, since current evidence does not indicate past dryland connection between these two islands (Heaney, 1986, 1991). We conclude that shrews of the genus Crocidura have entered the Philippines on three or four occasions, once from the north reaching only the Batanes Islands, and two or three times from the south. Having entered the Philippines from the south, they have repeatedly crossed narrow sea channels. In one case (C. palawanensis) speciation followed vicariance due to a rise in sea level. In three cases (C. negrina, C. grandis, and C. mindorus) it seems certain and in one case (C. beatus and C. grayi) likely that speciation followed overwater colonization. In two cases (C. attenuata in the Batanes and C. grayi on Mindoro), overwater colonization has not resulted in recognized speciation. Sympatry of Crocidura in the Philippines involves two cases in which members of the C. mitulorus group and the C. grayi group occur sympatrically, but never members of a single species group. The biogeographic history of the genus Crocidura in the Philippines is one of both colonization and diversification, with the latter taking place through both vicariance and dispersal, and with current patterns of species richness the result of both speciation and direct colonization. Taxonomic Accounts of Philippine Species Crocidura attenuata 1872. Crocidura attenuata Milne- Edwards, Rech. Mamm., p. 263. Type locality, Moupin, Szechwan, China. Remarks.— In comparison with C. attenuata from Taiwan and with C. grayi from Luzon, specimens from Batan Island have a short tail (ca. 42 mm) like shrews from Taiwan, rather than moderately long as on Luzon (ca. 60 mm). Bristle hairs are present on the proximal three-fourths of the tail of all three taxa, but are longer on shrews from Taiwan and Batan than from Luzon (10 vs. 6 mm, respectively). The shorter hairs that are closely appressed to the tail are short in Batan and Taiwan shrews, giving the tail the appearance of being nearly naked, whereas shrews from Luzon have longer hairs so that their tails look moderately hairy and have inconspicuous scales. The feet are pale brown on specimens from Batan and Taiwan, and dark brown on those from Luzon. The pelage of shrews from Batan and Taiwan is moderately dense but short, whereas shrews from Luzon have denser and longer pelage. Specimens from Batan are unlike all others in having P^ with a more strongly developed protocone which is placed more anterolingually than in other species, giving the tooth a 1994 HEANEY AND RUEDI — Evolution of Philippine Shrews 363 distinctive shape (Fig. 3). Specimens examined. — Batan Island. Batanes Prov.: Basco Airstrip (2 USNM); Itbud (1 USNM); no specific locality (2 USNM). Crocidura beatus 1910. Crocidura beatus Miller, Proc. U. S. Nall. Mus., 38:392. Type locality, Mt. Bliss, Mindanao, elev. 5750 ft. Holotype, USNM 144647. 1934. Crocidura parvacauda Taylor, Monogr. Bur. Sci. (Manila), 30:83. Type locality, Saub, Cotabato Prov., Mindanao, sea level. Holotype, UIMNH 33390. Remarks. — A small shrew, condyloincisive length 20.0-21.5 mm. Relative to C. grayi, there are usually only a few vibrissae-like hairs on the basal one-half of the tail, and these tend to be fine and short (rather than many coarse hairs over most of the length of the tail). The claws on fore and hind feet are slightly shorter and less robust. The cranium is slightly more elongate and narrow, and the braincase is slightly more inflated. In side view of the cranium, the anterior tip of the nasal aperture has walls that come to an acute anterodorsal point, nearly forming a right angle between anterior edge and dorsal edge, rather than being lower and more rounded. The upper P'* and molars are slightly broader and the toothrow is slightly longer (Fig. 2). Relative to C. attenuata, the braincase is more inflated, the interorbital region is broader, the molars are proportionately larger, and the posterior margins of F* and upper molars are very slightly concave, rather than strongly so (Fig. 3). The tail is longer, and all hairs are longer giving the tail a hairier appearance. The pelage is longer and denser, and the feet more darkly pigmented. Relative to C. negrina, the cranium is smaller and has a markedly shorter postpalatal region, and the ascending ramus of the mandible is lower; dental proportions are similar. Other taxa differ substantially in morphometric traits (Fig. 1, discussion above). Crocidura parvacauda is known only from the holotype. The skin label now has the note, “skin made up from alcohol specimen, skull not found Feb. 1956, carcass in alcohol.” A search of the UIMNH collection by Heaney in 1983 failed to produce the skull. The skin does not differ in any conspicuous way from C. beatus. Only the tail differs in being shorter (35 mm vs. 52-60 mm for those we have examined), which was noted by Taylor as the prime distinguishing feature. All of the measurements of the skull given by Taylor (1934:84) fall within the range of those for a series from Mindanao. Given the absence of any difference other than tail length, and the frequency of tail damage in small mammals, we recommend that C. parvacauda be considered a synonym of C. beatus until additional short-tailed specimens may be found. Specimens examined. — Biliran. 5 km N, 10 km E Naval, elev. 850 m (1 USNM). Bohol. 1 km S, 1 km E Bilar, 320 m (1 USNM). Leyte. Leyte Prov., Mt. Lobi Range: Tampas, Burauen, 2300 ft (2 DMNH); Mt. Pangasugan, 8.5 km N, 2.5 km E Baybay, 500 m (2 USNM); Mt. Pangasugan, 10 km N, 4.5 km E Baybay, 950 m (6 USNM). Maripipi. Leyte Prov.: 1 km N, 1.5 km W Maripipi town, 400 m (2 UMMZ). Mindanao. Agusan Prov.: Mt. Hilong-Hilong, Siwod, 3500 ft (3 DMNH). Bukidnon Prov.: Mt. Katanglad, Malaybalay, 5200 ft (2 DMNH); Mt. Katanglad (1 SMF). Cotabato Prov.: Saub, sea level (1 UIMNH). Zamboanga Prov.: Dabiak, Labao (1 FMNH); summit of Mt. Bliss, 5750 ft (1 USNM); Mt. Malindang, Duminigat, 5200 ft (1 FMNH). Crocidura grandis 1910. Crocidura grandis Miller, Proc. U. S. Natl. Mus., 38:393. Type locality. Grand Malindang Mt., Mindanao, 6100 ft. Holotype, USNM 144648. Remarks. — A large shrew, condyloincisive length 23.6 mm. Relative to C. mindorus, the tail is thicker with fewer long vibrissae-like hairs, hind feet longer and apparently less heavily pigmented; cranium more elongate, braincase slightly less globose, interorbital region broader, molars slightly larger. Relative to C. fuliginosa, quite similar, interorbital region broader, braincase slightly more inflated. Relative to C. negrina, generally larger; cranium proportionately more elongate, interorbital region broader, postpalatal region longer, unicuspids slightly broader. All other Philippine taxa are substantially smaller. Specimens examined. — Mindanao. Grand Malindang Mt., 6KX) ft (1 USNM). Crocidura grayi 1890. Crocidura grayi Dobson, Ann. and Mag. Nat. Hist., ser. 6, 6:494. Type locality, Philippine Islands; precise locality unknown. Holotype, BMNH 55.12.24.421. 1910. Crocidura halconus Miller, Proc. U. S. Natl. Mus. 38:391 . Type locality, spur of main ridge of Mount Halcon, Mindoro, 6300 ft. Holotype, USNM 144652. Remarks. — A small shrew (condyloincisive length 18.8-20.7 mm) with a narrow rostrum and proportionately narrow molar teeth. Differs from C. beatus and C. attenuata as discussed above. Other Philippine taxa are appreciably larger (Table 1). Specimens examined. — Catanduanes. 4 km E Summit, 250 m (1 USNM). Luzon. Benguet Prov.: Haights-in-the-Oaks (near Barrio Sayangan, Atok, ca. 2400 m) (4 USNM). Camarines Sur Prov.: Mt. Isarog, 475 m (1 PNM), 900 m (2 USNM), 1 125 m (25 USNM), 1350 m (3 USNM), 1550 m (2 USNM), 1750 m (6 USNM). Mountain Prov.: Mt. Data(l BMNH). Rizal Prov.: “probably Manila” (1 AMNH). No specific locality (2 BMNH). Mindoro. Bulalacao (1 USNM). Main ridge of Mt. Halcon, 4500 ft (3 USNM); spur of main ridge of Mt. Halcon, 6300 ft (1 USNM). Occidental Mindoro Prov.: Mt. Iglit Station (2 MMNH); 1 mi W Mt. Iglit Station (1 MMNH); 0.5 mi NE Mt. Iglit Station (1 MMNH); 3.5 mi NE Mt. Iglit Station (12 MMNH); 4 mi NE Mt. Iglit Station (6 MMNH); 5 mi NE Mt. Iglit Station (2 MMNH). Other records.— \Mzon. Abra Prov.: Lagangilang (Lawrence, 1939). Crocidura mindorus 1910. Crocidura mindorus Miller, Proc. U. S. Natl. Mus., 38:392. Type locality, summit of main ridge of Mt. Halcon, 6300 ft. Holotype, USNM 144654. 364 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 Remarks. — A moderately large Crocidura with a well- expanded, relatively globose braincase. Relative to C. fuUginosa, cranium generally less elongate; braincase proportionately shorter, more globose; molars laterally narrower; toothrow shorter. Posterior margins of upper molariform teeth less concave. Lower incisors less robust, lower molars with lower crowns, coronoid process lower, less robust. Relative to C. grandis, the tail is thiimer with fewer long hairs on proximal half, hind feet shorter and heavily pigmented; cranium less elongate, interorbital region slightly narrower. Relative to C. negrina, braincase more globose, interorbital region broader, slightly narrower upper molars, lower coronoid process. Other taxa appreciably smaller. Specimens examined. — Mindoro. Main ridge of Mt. Halcon, 6300 ft (2 USNM). Sibuyan Island. Romblon Prov.: NW slope Mt. Guitinguitin, 4.5 km S, 4 km E Magdiwang, 325 m (1 FMNH). Crocidura negrina 1952. Crocidura negrina Rabor, Nat. Hist. Misc., Chicago Acad. Sci., 96:6. Type locality, Dayungan, elev. 4500 ft, Cuemos de Negros (= Mt. Talinis), Negros Island. Holotype, FMNH 78445. Remarks. — A medium-sized shrew, condyloincisive length 22-23 mm. Relative to C. beatus, larger, proportionately longer rostrum, shorter postpalatal region, lower coronoid process on ascending ramus. Relative to C. grayi, larger, wider rostrum, broader molars, and longer toothrow. Relative to C. fuUginosa, shorter postpalatal region, broader interorbital region, posterior margin of upper molariform teeth much less concave. Relative to C. mindorus braincase less globose, narrower interorbital region, slightly broader upper molars, higher coronoid process on ascending ramus. Specimens examined. — Negros. Negros Oriental Prov.: Dayungan, Cuemos de Negros, 4500 ft (1 FMNH); Camp Lookout, Valencia, 500 m (1 SU); Naguro, Siaton, 700 m (1 DMNH); Lake Balinsasayao, 3 km N, 14 km W Dumaguete, 830 m (1 UMMZ), 1000 m (1 UMMZ), 1200 m (1 UMMZ). Crocidura palawanensis 1934. Crocidura palawanensis Taylor, Monogr. Bur. Sci., Manila. Type locality, Brooke’s Point, Palawan Island, Philippines. Holotype, AMNH 241679. Remarks. — A relatively large Crocidura, condyloincisive length 21.7-24.7 mm, cranium relatively elongate, interorbital region narrow, posterior margin of upper molariform teeth moderately concave. Relative to C. gratulis, interorbital region narrower and braincase slightly broader and shorter, posterior margin of molariform teeth less concave. Relative to C. mindorus cranium more elongate; braincase less globose; interorbital region narrower; molar teeth laterally broader, posterior margins more highly concave. Lower incisors more robust; lower molars with higher crowns; coronoid process higher, more robust. Relative to C. negrina, longer postpalatal region, narrower unicuspids, posterior margins of molariform teeth more concave. Other Philippine taxa substantially smaller, differing in quantitative traits (Fig. 1, discussion above). Relative to C. fuUginosa, specimens from Palawan and Balabac are slightly larger than those from Borneo, have narrower palates, and have P'^s that have less concave posterior edges. External differences are not consistent. Specimens from Palawan have longer tails and longer pelage than specimens from Borneo and Balabac. Balabac specimens have tails of equal length but shorter pelage than those from Borneo. The differences between C. palawanensis and the Sunda Shelf shrews are weak, and the Palawan shrews may eventually be synonymized with a species from the Sunda Shelf. We do not take such action here both because of the uncertainty regarding species limits of shrews on the Sunda Shelf, and in order to keep attention focused on the differences that are present. Specimens examined.— EdXah&c,. Palawan Prov.: Dalawan Bay, Minagas Point (2 USNM). Palawan. Palawan Prov.: Mt. Mantalingajan, 3600-4350 ft (1 USNM); Babuyan, Puerto Princesa, sea level (1 FMNH); Brooke’s Point (1 AMNH). Acknowledgments For boundless hospitality and unending cooperation in the course of field studies, we thank our many friends and colleagues at the Philippine National Museum (especially P. Gonzales), the Protected Areas and Wildlife Bureau (especially C. Custodio, W. Dee, and S. Penafiel), Silliman University (especially A. Alcala, L. Dolar, and R. Utzurrum), and the Visayas State College of Agriculture (especially L. Raros and P. Milan). Access to specimens was granted by the curators in charge of the following collections: American Museum of Natural History (AMNH); Bell Museum of Natural History, University of Minnesota (MMNH); The Natural History Museum, London (BMNH); Delaware Museum of Natural History (DMNH); Field Museum of Natural History (FMNH); Silliman University Museum of Natural History (SU); Senckenberg Museum, Frankfurt (SMF); University of Illinois Museum of Natural History (UIMNH); and the United States National Museum (USNM). We thank R. Hutterer and F. Petter for helping to solve the mystery of Crocidura edwardsiana. Earlier drafts of this manuscript benefitted greatly from comments and suggestions from P. D. Heideman, S. M. G. Hoffman, R. Hutterer, G. G. Musser, A. T. Peterson, E. A. Rickart, and R. W. Thorington, Jr. B. Strack gave patient instruction in the use of the SEM, J. Sedlock drew Fig. 5, and T. B. Griswold prepared the other illustrations. Our studies have been supported by the Field Museum of Natural History (Ellen Thome Smith and Marshall Field Funds), the Smithsonian Institution Office of Fellowships and Grants, and the National Science Foundation (BSR-8514223). Literature Cited Brown, J. C., and D. W. YalDEN. 1973. The description of mammals — 2. Limbs and locomotion of terrestrial mammals. Mammal Review, 3:107-134. Buth, D. G. 1984. The application of electrophoretic data to systematic studies. Annual Review of Ecology and Systematics, 15:501-522. Dobson, G. E. 1890. Descriptions of new species of Crocidura. Annals and Magazine of Natural History, series 6, 6:494-497. Dowling, T. E., and W. M. Brown. 1989. Allozymes, mitochondrial DNA, and levels of phylogenetic resolution among four minnow species (Notropis: Cyprinidae). Systematic Zoology, 38:126-143. Felsenstein, j. 1985. Confidence limits on phylogenies: An approach utilizing the bootstrap. Evolution, 39:783-791. 1989. PHYLIP (Phylogeny Inference Package), Version 3.2 1994 HEANEY AND RUEDI — Evolution of Philippine Shrews 365 Manual. University of Washington, Seattle (on disk). Fitch, W. M., and E. Margoliash. 1967. Construction of phylogenetic trees. Science, 155:279-284. Heaney, L. R. 1986. Biogeography of mammals in Southeast Asia: Estimates of rates of colonization, extinction and speciation. Biological Journal of the Linnean Society, 28:127-165. 1991. An analysis of patterns of distribution and species richness among Philippine fruit bats (Pteropodidae). Bulletin of the American Museum of Natural History, 206:145-167. Heaney, L. R., and E. A. Rickart. 1990. Correlations of clades and dines: Geographical, elevational, and phylogenetic distribution patterns among Philippine mammals. Pp. 321-332, in Vertebrates in the Tropics (G. Peters and R. Hutterer, eds.). Museum Alexander Koenig, Bonn, 424 pp. Heaney, L. R., and R. M. Timm. 1983. Systematics and distribution of shrews of the genus Crocidura (Mammalia: Insectivora) in Vietnam. Proceedings of the Biological Society of Washington, 96:115-120. Heaney, L. R., P. C. Gonzales, and A. C. Alcala. 1987. An annotated checklist of the taxonomic and conservation status of land mammals in the Philippines. Silliman Journal, 34:32-66. Hollister, N. 1912. A list of the mammals of the Philippine Islands, exclusive of the Cetacea. Philippine Journal of Science, 7:1-64. 1913. A review of the Philippine land mammals in the United States National Museum. Proceedings of the United States National Museum, 46:299-341 . HONACKI, J. H., K. E. K INMAN, AND J. W. KOEPPL (EDS.). 1982. Mammal Species of the World: A Taxonomic and Geographic Reference. Allen Press, Inc. and The Association of Systematics Collections, Lawrence, Kansas, ix + 694 pp. Jenkins, M. D. (ED.). 1987. Madagascar, an Environmental Profile. International Union for the Conservation of Nature and Natural Resources (lUCN), Gland, Switzerland, 374 pp. Jenkins, P. D. 1976. Variation in Eurasian shrews of the genus Crocidura (Insectivora: Soricidae). Bulletin of the British Museum (Natural History), Zoology Series, 30:271-309. 1982. A discussion of Malayan and Indonesian shrews of the genus Crocidura (Insectivora: Soricidae). Zoologische Mededelingen Leiden, 56:267-279. 1984. Description of a new species of Sylvisorex (Insectivora: Soricidae) from Tanzania. Bulletin of the British Museum (Natural History), Zoology Series, 47:65-76. Lanyon, S. 1985. Detecting internal inconsistencies in distance data. Systematic Zoology, 34:397-403. Lawrence, B. 1939. Mammals. Collections from the Philippine Islands. Bulletin of the Museum of Comparative Zoology, 86:28-128. Maddalena, T. 1990. Systematique, evolution et biogeographie des musaraignes Afrotropicales et Palaearctiques de la sous-famille des Crocid u rinae: U ne approche genetique . Unpublished Ph.D. dissert. , Lausanne University, 172 pp. Maddalena, T., and M. Ruedi. 1994. Chromosomal evolution in the genus Crocidura (Insectivora: Soricidae). Pp. 335-344, in Advances in the Biology of Shrews (J. F. Merritt, G. L. Kirkland, Jr., and R. K. Rose, eds.), Carnegie Museum of Natural History Special Publication no. 18, x + 458 pp. Miller, G. S., Jr. 1910. Descriptions of two new genera and sixteen new species of mammals from the Philippine Islands. Proceedings of the United States National Museum, 38:391-404. Musser, G. G. 1982. Results of the Archbold Expeditions, No. 108. The definition of Apomys, a native rat of the Philippine Islands. American Museum Novitates, 2746:1-43. Musser, G. G., and L. R. Heaney. 1985. Philippine Rattus: A new species from the Sulu Archipelago. American Museum Novitates, 2818:1-32. Nei, M. 1978. Estimation of average heterozygosity and genetic distance from small numbers of individuals. Genetics, 89:583-590. Pasteur, N., G. Pasteur, F. Bonhomme, J. Catalan, and J. Britton-Da vidian. 1987. Manuel technique de genetique par electrophorese des proteines. Tec and Doc, Paris, 217 pp. Ruedi, M., T. Maddalena, H. S. Yong, and P. Vogel. 1991. The Crocidura fuliginosa species complex (Mammalia, Insectivora) in peninsular Malaysia: Biological, karyological, and genetical evidence. Biochemical Systematics and Ecology, 18:573-581. SAS Institute Inc. 1988. S AS/ST AT User’s Guide. SAS Institute, Cary, North Carolina, 1028 pp. SWOFFORD, D. 1990. PAUP, Phylogenetic Analysis Using Parsimony, Version 3.0. Illinois Natural History Survey, Champaign, Illinois (on disk). Taylor, E. H. 1934. Philippine Land Mammals. Monographs of the Bureau of Science, Manila, 30: 1-548. Trouessart, E. L. 1880. Description d’une Espece nouvelle de Musaraigne de la Collection de Musee de Paris. Le Naturaliste, 1(42):330. 1897. Catalogus Mammalium Tam Viventium Quam Fossilium. Pt. 1. Frielander and Sohn, Berolini, 664 pp. Appendix Crocidura edwardsiana from “lie Soulou” (= Jolo Island, Sulu Archipelago), is eurrently misidentified in systematic literature. Trouessart (1880) named this species at a time when Pachyura ( = Suncus) was considered a subgenus of Crocidura. Although the original description made no subgeneric designation, it stated that there are 28 teeth, and only four teeth are mentioned anterior to the molariform teeth, making it fit the current definition of Crocidura. However, Trouessart (1897) listed edwardsiana as a variety of Crocidura (Pachyura) caerulea, which is now considered a synonym of Suncus murinus. Hollister (1912) listed edwardsiana as a species of Pachyura on the basis of Trouessart’s (1897) description of four upper unicuspids, the crucial diagnostic character of Suncus. However, Taylor (1934) listed edwardsiana as a species of Crocidura, apparently on the basis of the original description of the unicuspids. Subsequent authors have followed Taylor uncritically (e.g., Honacki et al., 1982; Musser and Heaney, 1985). The original description was based on two specimens; no others have been obtained subsequently. The paratype was not found by F. Petter in 1986 (Petter, personal communication) and is presumed lost. The holotype of C. edwardsiana (MNHNP 1881-3895) was examined by R. Hutterer on 29 January 1987 (Hutterer, personal communi- cation). He found it possessed all of the characteristics of a juvenile of one to two weeks age: pelage hairs short and gray, skull not fully calcified, basisphenoid suture open, and teeth unworn. The right side of the palate was still covered by flesh, but the left side had been cleaned and the fourth upper unicuspid was visible. Cranial measurements (as defined by Heaney and Timm, 1983) were: CIL 25.0; UTR 11.8; MB 8.9; GW 10.5; PL 11.6; lO 5.5; RL 8.2; PPD 5.1; RB 3.6; PPL 10.4; CGL 8.8; P4M3 6.3; M2M2 8.7; PWM3 4.1; COR 6.4; LTR 11.2. He concluded that the holotype is a juvenile Suncus murinus of “caravanning age.” The presence of four unicuspids, with the measurements given here, and identification of the holotype as a juvenile, make identification of this taxon as Suncus murinus unambiguous, in keeping with Trouessart’s (1897) assessment. 366 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 5 § "I 5 •I -c eS 3 'o o l-l U t; Xi nj H On q q q 1 o d' d d d 3" q do do o <5 NO +1 o +1 •Cl *T o +i d +1 ON d 1 i£ o Um NO d •C) 00 cl d NO q »o PC PC q PC d d d 1 id •Cl r< q q PC ■5 ON *0 d' ci d d d Cl d PC d q do q PC *5 OX) so 0\ +1 r- NO -H NO PC d •Cl On +1 NO o »n +1 NO 00 d r- 1 •Cl H o »T) •o •Cl o PC PC d d •ci On •ci On d •5 NO *o •Cl 00 PC d PC NO 3 OX) f<) +1 -H o +1 +1 d d d o e Tt* 1 ON 1 00 1 •Cl 1 d 1 q 1 1 q 1 •o NO 1 od 1 1 c* o 00 d PC PC 00 o •Cl •C) J f<) d PC d NO ci PC od d NO •C) *T) o d PC Cl oo q q q •Cl q q q q q do q 2 JS *2 •o f<) o -H ci 1 d cl 1 d +1 Cl 1 q q d H-1 Cl 1 d +l ci i d +1 ci 1 d -H ci 1 q •Cl q ci 1 o q cl 1 d +1 ci 1 d r4 ri o PC ci q 00 Cl ci d o q PC 00 Cl ci ci q ci q o q ci cu »n d d PC d q d d d q d d d q ri ci Cl ci ci ci ci NO NO PC •Cl ^'k o q d d PC PC Cl •cT d o d q CQ o d d d d On d d d d NO d NO d NO O' NO NO d d d NO o CQ vd 'd +! 00 1 PC d 1 +1 1 q d 1 d +1 o 1 q +1 1 q +1 •Cl 1 q +1 NO 1 q NO NO 1 NO q NO 1 C' +1 PC 1 q d i hJ fo d d o d d q d O q q d d d •ri Nd d d NO NO •ci NO NO d 00 d 2 cl q d o q d rt q d? PC f<) o d 1 d d d 1 q d 1 d •ci 1 d •ci d •ci d •c C' •Cl •Cl o •ci d •ci 1 O +1 PC 1 +1 +1 +1 1 +1 1 +1 1 •ri •Cl ! •Cl 1 +! d w «n »d r- d d On q d q PC q d o q d •ci q •ci PC •Cl q Oh cl d d q d q d Cl •ci q q q •ci •ci q d •Cl •c •ri •ci rc PC d On u D CU o ■5 t Cl o +1 On J PC PC On 1 PC d +1 PC d 1 q d d cl d +1 d 1 Cl d +1 q d ) Cl d +1 PC d 1 Cl d +1 d d i d od i d od 1 •Cl od o C'' d od C'' c od 00 +1 Cl +1 q +! +1 +1 +1 00 1 d +1 s od ■g o r- od q PC q d q d cl q NO q Cl q d od q od od PC q o 5 fC od d q PC d q d ON d 00 d go q d U od d d od d d d 00 C“ •Cl cl NO d 1 J3 do d o PC d PC PC d Cl d Cl C d q d “S & •3 OX) c 00 d +1 On 1 d 1 d +1 d 1 •Cl q d o +1 1 d +1 d 1 d +1 d 1 d +1 d f q d •Cl q d o q d d +! d i q d 1) ON ON o PC On q o q d •c q On q o q cl q d •/^ d •A PC q o ON d Cl d d q od q od q od d q gd CL, On d o od od od od fC d NO PC •C) Cl O q d <*H d? d do do q q 1 •5 *o I ON d +1 PC 1 yn q Cl 1 d +1 cl 1 q PC 1 o q q d +1 ci 1 d +1 ci { d +1 ci 1 q •Cl q ci 1 o q ci t d +! ci I q o > c4 *o *0 Cl q o PC cl q ci d On PC o q ci ci q ci q PC q ci oc r- d d q d d q d q d q d d ci q ci cl ci Cl ci ci ci 1 rc •C) q o cl - q 00 o d? d d — 00 o *« o- •B Cl 'U Q 3.7 3.8 d +1 cl 1 d 3.7 P PC d +1 NO PC j q 4.0 1 q d +1 o 1 q d +1 On 1 q d +1 d PC 1 q d +1 o PC i 4.2 (4.1) o q d 1 o PC 1 q 4.0 o On PC d PC PC Cl q PC q PC q PC PC PC CL. fC PC PC PC p Cl P NO q q 00 q Tt Cl cl d •Cl q •Cl u 3 OX) NO d od q q d d •o d d d d od d od d od q o d d d od 00 e 1 +1 1 +1 1 cl +I 1 +1 1 +1 i +1 1 +1 1 +1 1 o o d d 00 d ■ 2 d q q q q o •Cl On PC Cl q q NO q o •5 21.2 d .-H «C) d • 2 •C) ci d d d d d d q PC d •Cl ci •C) ci d d ■_0 OX) c 1) 1 +1 o d 1 NO q Cl 1 NO +1 00 cl 1 O 3 Cl Cl Cl 1 q +1 d Cl 1 q +1 00 d i +1 PC cl j PC +1 o cl 1 q cl 1 PC Cl ci d 1 q -H Cl Tf 23.-: >% "S J fC d d d q d c d q PC q q d 3 •Cl od 3 f"l ci 3 d ci q d 3 Q d d d d d PC d d o c d ti C d d o O cl cl d '1 Cl cl 3 3 1 3 ■§ C OC c -Ci - - 'O d On *3 Cl d PC L. 00 o 00 *3 •C PC C _ QC I c •Cl Oh 01 a o 3 3 Q 3 3 O 3 3 3 >> L. 3 !§* 0Q c L. 3 o a k. 3 c o o N L. 3 o 3 c 3 2 L. 3 •h, 3 e C Leyte eQ CQ eO "3 N >3 o O ■3 "3 CQ *« o o J ‘C a CQ w s Bohol 'C OQ S 3 e s ’C a X) .2 CQ cri -2 CQ cu ‘C a 3 i-J d J Z G •? 2 1 o 2 C ^ •- o S £ 1 'C a Negr( ‘C a Batan Sibuy 1994 HEANEY AND RUEDI— Evolution of Philippine Shrews 367 Table 2.— Results of a principal components analysis of log-transformed cranial measurements <^Crocidura from the Philippines and adjacent areas. Axes I and II are graphed in Fig. 1. Axis Variable I II Condyloincisive length 0.30 -0.05 Braincase width 0.29 -0.02 Interorbital width 0.28 -0.12 Rostral length 0.30 -0.15 Postpalatal depth 0.29 0.04 Rostral width 0.27 -0.27 Postpalatal length 0.29 -0.10 Condyle to glenoid 0.29 -0.23 I' to 0.30 -0.03 P'^ to 0.29 -0.04 to (labial) 0.30 0.12 Palatal width at 0.26 0.90 Cumulative % variance 89.5 92.0 Eigenvalue 10.7 0.29 Table 3. — List of qualitative characters discussed in the text, giving the character state for each of the populations from the Philippines and from three reference populations. Character numbers and state codes are those used in Fig. 6. Species 1 Interorbit 2 Posterior Palate 3 Parastyle of P^ 4 Lingual Exposure of P^ 5 Concavity of P^ 6 Thenar and Hypothenar 7 Pigmentation of Feet 8 Plantar Granulae Crocidura attenuata narrow broad prominent low great small and rounded slight inconspicuous Crocidura sp. (Batanes) narrow broad prominent low great small and rounded slight inconspicuous C. beatus moderate narrow low moderate low elongate and flattened heavy prominent C. grayi moderate narrow low moderate low elongate and flattened heavy prominent C. halconus moderate narrow low moderate low elongate and flattened heavy prominent C. fuliginosa narrow narrow prominent high moderate elongate and flattened moderate prominent C. grandis broad narrow prominent high moderate elongate and flattened — — C. mindorus broad narrow prominent high low elongate and flattened — — C. negrina moderate narrow prominent high moderate elongate and flattened moderate prominent C. palawanensis narrow narrow prominent high moderate elongate and flattened — — Crocidura sp. (Sibuyan) \ broad narrow prominent high low elongate and flattened moderate prominent C. russula narrow narrow prominent high great small and slight inconspicuous rounded 368 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 Table A.— Frequency of allozymes in seven populations of Crocidura. Dashes indicate values of zero. Locus Relative Speed C. grayi n = 9 C. beams n = 1 Crocidura sp. (Sibuyan) n = 1 C. fidiginosa (CH) n = % C. malayana (UG) n = 4 C. malayana (TI) n = 6 C. russula n = \0 ADA + 155 __ 0.08 + 140 — — — 0.25 — 0.84 — + 120 0.75 — — 0.75 0.50 0.08 — + 100 0.25 1.0 1.0 — 0.50 — 1.0 ADH -100 1.0 1.0 1.0 1.0 1.0 1.0 1.0 AK-1 + 112 — — — 0.13 — — — + 100 1.0 1.0 1.0 0.87 1.0 1.0 1.0 ALB + 95 0.05 — 1.0 1.0 0.50 1.0 — + 91 0.28 — — — 0.50 — 1.0 + 84 0.67 1.0 — — — — — CK-1 + 100 1.0 1.0 1.0 1.0 1.0 1.0 1.0 CK-2 + 100 — — — 1.0 1.0 — 1.0 + 71 1.0 1.0 1.0 — — 1.0 — EST-1 + 135 — — — — — — 1.0 + 106 1.0 1.0 1.0 — 1.0 1.0 — + 87 — — — 1.0 — — — GOT-1 + 200 — — — 1.0 — — — + 100 1.0 1.0 1.0 — 1.0 1.0 1.0 GOT-2 -100 1.0 1.0 1.0 1.0 1.0 1.0 1.0 a-GPD + 153 — — — 0.13 — — — + 100 1.0 1.0 1.0 0.87 — — 1.0 + 67 — — — — 1.0 1.0 — G-6-PD + 120 — — — 1.0 — — ■— + 115 1.0 1.0 1.0 — 1.0 1.0 — + 107 — — — — — — 1.0 HBB -100 1.0 1.0 1.0 — 1.0 1.0 1.0 -82 — — — 1.0 — — — HK-1 + 163 — — — 0.13 — — — + 110 0.11 — — — — 0.33 — + 100 0.89 1.0 1.0 0.87 1.0 0.67 1.0 HK-2 -100 1.0 1.0 1.0 — 0.50 0.83 1.0 -81 — — — — 0.25 — ~ -76 — — — 1.0 0.25 0.17 — GDH + 118 1.0 1.0 1.0 — 1.0 1.0 — + 100 — — — 1.0 — — — + 0 — — — — — — 1.0 IDH-1 + 155 1.0 1.0 1.0 1.0 1.0 1.0 1.0 IDH-2 -100 1.0 — 1.0 1.0 1.0 1.0 — -75 — — — — — — 1.0 -45 — 1.0 — — — — — SOD-1 + 102 — — — 0.06 — — — + 100 — — — 0.88 0.25 — — + 94 — — — 0.06 0.38 1.0 — + 84 0.06 — I.O — 0.25 — 1.0 + 79 0.55 1.0 — — 0.12 — — + 65 0.39 — — — — — — SOD-2 -100 1.0 1.0 1.0 1.0 1.0 1.0 — -60 — — — — — — 1.0 SOD-3 -100 1.0 1.0 1.0 1.0 1.0 1.0 — -85 — — — — — — 1.0 LAP + 100 1.0 1.0 1.0 1.0 1.0 1.0 1.0 LDH-1 + 100 1.0 1.0 1.0 1.0 1.0 1.0 1.0 1994 HEANEY AND RUEDI— Evolution of Philippine Shrews 369 Table 4 (cont.) LDH-2 -110 1.0 1.0 1.0 1.0 1.0 1.0 1.0 MDH-1 + 100 1.0 1.0 1.0 1.0 1.0 1.0 1.0 MDH-2 -100 1.0 1.0 1.0 1.0 1.0 1.0 1.0 MOD -125 — — 1.0 -115 — 1 .0 — -105 1.0 1.0 1 .0 1 .0 1.0 — MPI + 137 — 0.50 0.87 0.88 0.42 0.80 + 100 1.0 1.0 0.50 0.13 0.12 0.58 0.20 PA + 80 1.0 1.0 1.0 1.0 1.0 1.0 1.0 6-PGD + 180 — 1.0 — 0.50 0.48 + 140 1.0 1.0 — 0.50 0.52 1.0 + 100 — 1 .0 — — — PGI -200 — — 1.0 -100 1.0 1.0 1 ,0 i .0 1.0 1.0 — PGM + 150 — 0.19 — — — + 120 0.28 1 .0 1 .0 0.81 1.0 1.0 + 100 — — — 1 .0 + 75 0.72 — PROT-A + 100 1.0 1.0 1 .0 1.0 1 .0 1 .0 1 .0 Table 5.— Matrix of Net’s genetic distances (and standard errors) for Southeast Asian Crocidura, calculated from data in Table 4. Abbreviations for taxa as in Table 4. C. juliginosa (CH) C. malayana (UG) C. malayana (TI) C. grayi C beatus Crocidura sp. Sibuyan C. mal. 0.34689 (UG) (0.11285) C. mal. 0.49254 0.11089 (TI) (0.13896) (0.05430) C. grayi 0.53944 0.19448 0.21593 (0.15011) (0.07679) (0.08019) C. beatus 0.51422 0.15998 0.18357 0.07695 (0.14610) (0.06685) (0.07675) (0.04265) C. sp. 0.42709 0.10561 0.13397 0.15663 0.09996 Sibuyan (0.13264) (0.05382) (0.06788) (0.06724) (0.05872) C. russula 0.54250 0.44351 0.56945 0.50845 0.46134 0.36350 (0.15099) (0.12827) (0.15528) (0.14121) (0.13624) (0.11989) 370 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 CD (M CXJ U CD CO II Dd PQ S . a I I H a e § 0 .2? a ^ 1 5 U § S oS . > V 5 o « ^ = 2 o 2 "O .00 •T3 S . « (j C/5 S “ .& « rz; -o _c os - c/2 a, (U B - i S a o S cj 2 o -=: ;j <4-1 V? O ^ C • • |Q 2| C« 5 0) u 6C9 cn s c o as ti; u CJ C3 <4- 03 o ’c: • S2 <5 cn ^ >, « CC ^ C c cs -C: M .2? c !a s ^ go §0 i§ •2" £ o C kT^ •c > E o ^ c/5 .i: Q - 1 (U 3 S 'E, S '5. . 'C Cj 0* S E <0 o 3 CQ 3 E -Q o . 5 E 5 P 2 o a o ^ ii s s 2 o <*2 3 O -s-S I § .. 03 E- § 6 S u O ^ ■ S § ^ CO § 6 i = S y .j -2 N-J c« c *3 g « cs e 3 o O J- c/5 S -~ o 3 e E o o to - P C 3 Ml i- a lu cO *5 Z <0 3 S I <*: „ Co 3 I U *«• ^ ^ N U ^ I CQ £55 1994 HEANEY AND RUEDI— Evolution of Philippine Shrews 371 0 1 2mm Fig. 2. — Scanning electron micrographs of the upper toothrows and palates of A: C. attenuata from Vietnam (FMNH 46641); B: Crocidura sp. from Batan Island (USNM 463794); C: C. grayi from Luzon (USNM 573365); D: C. beatus from Maripipi Island (UMMZ 160372); E: C. hatconus from Mindoro (FMNH 87388); F: C. negrina from Negros Island (UMMZ 158881); G: C. fuliginosa from Borneo (FMNH 33055); H: Crocidura sp. from Sibuyan Island (FMNH 137022); I: C. mindorus from Mindoro (USNM 144653); J: C. palawanensis from Palawan Island (FMNH 63022). 372 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 Fig. 3. — Scanning electron micrographs of the P^s of Crocidura] all labels as in Fig. 2. 1994 HEANEY AND RUEDI — Evolution of Philippine Shrews 373 0 1mm 1 1 I Fig. 4.— Scanning electron micrographs of the upper left toothrows of Crocidura-, all labels as in Fig. 2. 374 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 Fig. 5.— Hind feet of C. attenuata from Batan Island (left); C. grayi from Mt. Isarog, Luzon (center); and C. mindorus from Sibuyan Island (right). 1994 HEANEY AND RUEDI— Evolution of Philippine Shrews 375 A B 50% 1 00% 1 00% 1 00% ■ 100% 71% C. attenuata Batan I. C. beat us C. grayi C. halconus C. grand is C. mindorus Sibuyan I. C. negrina C. fuliginosa C. palawanensis C. russula 100% 50% 100" 1 00% 100% 57% 57% C. attenuata Batan I. C, beat us C. grayi C. halconus C. mindorus Sibuyan I. C. grand is C. negrina C. fuliginosa C. palawanensis C. russula Fig. 6. — Results of cladistic analysis of morphological characters in Table 3. A: “Semistrict” consensus of 14 shortest trees. B: “Majority-rule” consensus of 14 shortest trees (see text for definitions). Index of consistency = 0.86. 376 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 B Fig. 7.— A: Results of Fitch-Margoliash analysis of Nei’s genetic distances based on allozyme frequencies presented in Table 2. Estimated branch lengths are proportional to those shown in the figure. Black circles indicate nodes that are robust to a jackknife manipulation of taxa. B: Results of cladistic parsimony analysis of data in Table 2 (see Methods and text). C. russula (Europe) C. attenuata C. fuliginosa C. palawanensis (Asian mainland and Balan I.) (Asian mainland and continental shell) (Peripheral continental shelf) C. negrina (Negros i.) C. grandis (Mindanao) C. mindorus (Mindoro & Sibuyan) C. baatUS (Greater Mindanao) C. gray! (Greater Luzon & Mindoro) Fig. 8.— Final hypothesis of phylogenetic relationships; distributions indicated in parentheses. 1994 HEANEY AND RUEDI— Evolution of Philippine Shrews 377 Fig. 9. — Map of the Philippine Islands showing the extent of islands during the late Pleistocene period of low sea level when water dropped 120 m, and the localities from which Crocidura have been taken. 1 = C. beatus, 2 = C. palawanensis , 3 = C. grandis, 4 — C. grayi, 5 = C. negrina, 6 = C. mindorus, 1 — C. attenuata. :b f ' Hl« .. . ^ A*. . . *Lri ~4‘ 1 ‘8i ■I ’ .••; ' •i- ; i/*/j ' k . •■. U: • ^ ■ ■■'•^ r'V ■ ■ ''k |! t»} V. ‘.^ ! ' i ‘ > - iy ■., tvui _4;‘ f-A*-' QaiiROa ri^'’’’ t>* ' -' ' ■'. ;iTi K . i-'4U»i'.j ■ 'h'*' ^T:*^** ^'* ' M'in’ .;.(/■ "S • > „(\i>v • .•' >i)|j,*iii vt .up . •;■. ‘ EVOLUTION AND PHYLOGENETIC AFFINITIES OF THE AFRICAN SPECIES OF CROCIDURA, SUNCUS, and SYLVISOREX aNSECTIVORA: SORICIDAE) Laura J. McLellan Division of Mammals, National Museum of Natural History, Smithsonian Institution, Washington D.C. 20560; Current address: Biology Department, Central Missouri State University, Warrensburg, Missouri 64093-5053 Abstract Members of the shrew genus Crocidura are Old World in distribution, with 96 of the 151 described species occurring in Africa. The fossil record is meager and consequently it is necessary to rely on comparative studies of Recent species for reconstruction of phylogeny. Attempts to elucidate the systematics of African Crocidura have resulted in the recognition of species groups that share a few general characteristics without rigorous consideration of their phylogenetic affinities, A set of 42 discrete morphological characters was used to evaluate phylogenetic relationships within the genus Crocidura, and among Crocidura, Suncus, and Sylvisorex. Myosorex was used as the out-group. Parsimony and cluster analyses of external and cranial features of African crocidurine shrews were used to examine phenetic affinities, determine phylogenetic relationships, and to infer possible speciation events based on biogeographic patterns. African species of Crocidura, Suncus, and Sylvisorex do not represent monophyletic lineages. The generic level distinctions among Crocidura, Suncus, and Sylvisorex based on tooth number are not maintained when additional characteristics are included in parsimony or cluster analyses. Further, the Crocidura species groups that have previously been defined vary according to which characters are included in the analysis. A high degree of homoplasy in cranial and external morphology warrants the use of additional characters such as soft tissue anatomy, anatomy karyotypes, and biochemical data to derive testable phylogenetic hypotheses. Most of the crocidurine shrews examined in this study occur throughout the lowland forests and savannas of east and west Africa, and in the highlands of Ethiopia, Kenya, Tanzania, Uganda, and Zaire in presumed refugia of arid phases in east and west Africa. Past climatic changes have undoubtedly been a major mechanism in the speciation of crocidurine shrews in Africa. Introduction The soricid genus Crocidura is the largest and one of the least understood genera of mammals. Species range in weight from 2.5 to 100 g, and are characterized by lack of pigment on the incisors, relatively long tails that are usually covered with a mixture of long and short bristly hairs, paired lateral musk glands between the front and hind legs, and 28 teeth. Members of this Old World genus occur in Africa, southern Europe, and Asia, as well as Indonesia and the Philippine Islands. Of the 151 recognized species, 96 occur in Africa, 22 in southern Europe and Asia, and 31 in Indonesia and the Philippine Islands (Honacki et al., 1982). Throughout their geographic range, members of Crocidura inhabit damp and dry forests, grasslands, deserts, and areas of human habitation (Smithers, 1983; Nowak, 1991). Phylogenetic affinities among African species of Crocidura, Suncus, and Sylvisorex, all members of the subfamily Crocidurinae, are unclear. Crocidurinae is characterized by the retention of primitive characters. Modem forms differ from those of the late Miocene and from one another by the loss of one, two, or three upper and lower antemolars; reduction in the talonid of the lower third molar; and a greater emargination of the posterior basal outline of the upper premolar and upper first molar. The generic boundaries are, however, based on very few characters. Suncus differs from Crocidura by the retention of a fourth upper antemolar. Sylvisorex differs from both Suncus and Crocidura by a lack of tail bristles, but further differs from Crocidura by the retention of a fourth upper antemolar. Heim de Balsac and Lamotte (1956, 1957) examined phylogenetic affinities among the African shrew genera within the subfamily Crocidurinae, and associated evolutionary advancement with increases in body size, pilosity of the tail, size of the rostral portion of the skull, a flattening of the braincase, and a decrease in the complexity of the dentition. Myosorex is considered the most ancient of the living genera because it has the largest number of teeth {n = 32), and Suncus, Sylvisorex, and Scutisorex are viewed as descendants of this stock because these genera have one less lower antemolar (Fig. 1). Crocidura and Paracrocidura have the most reduced dentitions (n =28), having lost one upper and one lower antemolar. Paracrocidura is described as a distinct radiation derived from an ancestor common to Sylvisorex and Crocidura. Crocidura is considered the most advanced of the genera. However, the relationship between Crocidura and its two closest sister groups, Suncus and Sylvisorex, was unresolved. Heim de Balsac and Lamotte (1956) considered that Crocidura might be diphyletic, originating from both Sylvisorex and Suncus by loss of an upper antemolar. Dippenaar and Meester (1989) examined the cladistic relationships among species of the C. luna-fumosa complex using morphological data. Sylvisorex and Myosorex were used as out-groups. Their conclusions about plesiomorphic and apomorphic character-states in Crocidura are consistent with the findings of Heim de Balsac and Lamotte (1956, 1957). The resulting cladogram was unresolved in that no apomorphy defined the C. luna-fumosa complex. The species complex was established on phenetic grounds, so one or more of the species may be more closely related to members of other species complexes. The extent to which these phenetically defined species complexes reflect phylogenetic affinities remains to be tested. Butler et al. (1989) examined relationships among African crocidurine shrews based on multivariate analysis of mandibular 379 380 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 data (Fig. 2). Two groups of Sylvisorex are noted: the lunaris group (5. morio, S. lunaris, S. ollula) and the granti group {S. grant i, S. megalura, S. howelli, S. johnstoni, S. olduvaiensis). The lunaris group is considered primitive in most characters, whereas the granti group is judged more advanced because of a decrease in dental complexity. Suncus forms a single group that is similar to the Sylvisorex granti group, with the exception of S. murinus. Crocidura suaveolens shares characters with Crocidura fuscomurina. The funiosa group of Crocidura (maurisca, dolichura,fumosa, mariquensis , pitmani) is assumed to have a common origin with the Sylvisorex granti and lunaris groups. Given the prevalence of parallelism, monophyly is viewed as unlikely. Because five groups of Crocidura are confined to Africa, they probably originated there (Butler et al., 1989). Most species of Suncus are from the Oriental region; thus, they may have arisen from a Sylvisorex-like form from Asia, such as Feroculus, that subsequently extended its range to Africa (Butler et al., 1989). Finally, a diphyletic origin for Crocidura was again suggested (Butler et al., 1989). Maddalena (1989) examined systematic relationships among 21 European and African species of Crocidura using allozyme data. Sylvisorex spp. were used as out-groups. Palearctic and Afrotropical species were separated into two major groups using phenetic and cladistic analyses of 25 loci (Fig. 3). Within the Palearctic group, C. russula forms the sister taxon to the remaining species. Branch 2 of the cladogram contains Afrotropical species of Crocidura with the exception of C. luna and C. bottegi. Crocidura luna is in a clade by itself and is the primitive sister-group to the other species of Crocidura examined. The large-sized shrews C. flavescens, C. lamottei, C. olivieri, and C. viaria form a monophyletic group. The remaining species of branch 2 form a sister-group to the giant shrews. These Crocidura species form nine monophyletic groups, but the relationships among these groups remain unresolved. Past studies of the systematics of African Crocidura, without rigorous consideration of the phylogenetic affinities with Sylvisorex and Suncus, have resulted in the recognition of species groups that share some general characteristics. The objectives of this study were to identify additional morphological characters useful in evaluating phylogenetic affinities, provide a detailed description and analysis of each potential character, examine relationships among Afrotropical Crocidura and out-group taxa using phenetic and parsimony methods, compare results derived from different methods, and consider the evolution and biogeography of African crocidurine shrews. Materials and Methods 1 examined 18 African species and subspecies of Crocidura, five species of Myosorex, three species of Suncus, and four species of Sylvisorex. Only those Crocidura species that possess primitive dental characteristics were included in this study, under the assumption that evolutionary advancement is associated with a reduction in the complexity of the dentition. The species of Crocidura included in this study are medium- sized, with the second upper antemolar equal in size to the third (rather than smaller), and having a hypocone on the fourth upper premolar, or a complex talonid on the third lower molar, or both. All specimens of Myosorex, Sylvisorex, Crocidura, and Suncus from Africa available at the National Museum of Natural History were examined. A set of 42 characteristics was selected for parsimony and cluster analyses of Sylvisorex, Suncus, and Crocidura. These characters are discrete (present or absent), or form a series of states that represent modifications or alternative forms of a homologous structure. A detailed description of each character follows. Character states were coded from primitive to advanced indicating likely directional trends (i.e. , polarity) over evolutionary time using the out-group method (Watrous and Wheeler, 1981). Suncus and Sylvisorex are sister-groups of Crocidura. Myosorex, the most primitive member of the subfamily Crocidurinae, was used in out-group analysis of Crocidura, Suncus, and Sylvisorex. Survey of the Characters The 42 characters selected for cladistic analysis are grouped by anatomical region. Characters of the teeth (Tl-15), naso- maxillary region (Nl-6), palatal region (Pl-3), brain case (Cl-4), orbito-temporal region (01-3), basicrania (Bl-6), and external features (El-5) were examined (Table 1). The letters A through D refer to character states as listed in Table 1. Character Tl. Posterior Cusp off Bicuspid. — A, absent; B, present. This character is unique to Crocidura, and thus is treated as the derived condition. Character T2. Proximal Width (Lateral View) of f . — A, weak {<'/i rostral depth); B, robust (>'A rostral depth). The character state T2 is a measurement of the width of I* relative to the rostral depth (rostral depth is measured from the superior border of the nasals to the superior I* alveolus). The robust condition is treated as the derived condition. It is weak in Sylvisorex, but robust or weak in Suncus and Crocidura. The robust condition is noted most often in the savatma species. Enlargement of the incisors may allow consumption of larger and tougher prey items, and may be adaptive in areas where a large proportion of hard-bodied prey items is available (Hutterer, 1986). Character T3. Cusp-Like Posterior Lingual Cingulum of f . — A, absent; B, present. This trait appears to be produced by the crowding of the first upper antemolar and l'. The occurrence of a cusp-like cingulum on I* is a character whose polarity is difficult to determine from the occurrence within out- groups. Crocidura as well as the out-groups show both conditions. Character T4. Relative Size of the Second and Third Upper Antemolars. — A, second upper antemolar smaller than third (length or width); B, second upper antemolar equal in size to third. The second upper antemolar may be smaller than or equal in size to the third. Heim de Balsac and Lamotte (1956) concluded that a decrease in the size of the third upper antemolar represented the derived condition. The distribution of this character within out-groups generally supports this assumption. Character T5. P"^, Hypocone. — A, absent; B, present. 1994 MCLELLAN— Evolution and Phylogeny of African Crocidurine Shrews 381 Retention of the hypocone is considered primitive by Heim de Balsac and Lamotte (1956). The polarity of this character is difficult to determine based on the occurrence within out- groups. The presence and absence of a P‘* hypocone are seen in Crocidura, Suncus, and Sylvisorex, but the presence is far less frequent in Crocidura than in the other species, suggesting that parallel losses have occurred in Suncus and Sylvisorex. Tooth wear may obscure reliable observation of this character. Character T6. , Parastyle Height (Lateral View).— A, less than third antemolar; B, equal to or greater than third antemolar. A tall P^ parastyle relative to the height of the third antemolar appears to be the primitive condition based on the pattern seen within out-groups. This feature may be produced in one of several ways: the P^ parastyle may be proportionately large, the third antemolar may be proportionately small, or differences in the size or shape of the maxilla may alter the positional relationship between the teeth. Character T7. Size Relative to M~ Size (Anterior- Posterior).— A, small (length < '/i that of M“); B, large (length > '/i that of M^). The of Crocidura ranges in size and shape from broad and square to long and narrow. Heim de Balsac and Lamotte (1956) concluded that the decreases in size with evolutionary advancement. Out-group analysis confirmed these conclusions. Character T8. Fourth Upper Antemolar.— A, absent; B, present. This tooth is present in some Myosorex, and all Suncus and Sylvisorex species, but absent in Crocidura. It has traditionally formed the basis for the distinction between Crocidura and its two closest sister genera Suncus and Sylvisorex. A reduction in the number of teeth has been associated with evolutionary advancement in crocidurine shrews (Heim de Balsac and Lamotte, 1957). Character T9. Ij, Cutting Surface with Accessory Cusps or Denticulations . — A, absent; B, single; C, double; D, triple. The presence of accessory cusps on the Ij is considered primitive by a number of authors (Heim de Balsac and Lamotte, 1957; Butler and Greenwood, 1979). This character is well-developed in several species of Sylvisorex; most have two denticulations. Sylvisorex lunaris has three denticulations, and the third may be an autapomorph of that species (Butler et al., 1989). The cutting surface of the I[ is slightly serrated (one or two accessory cusps) to smooth in species of Suncus and Crocidura. It is not clear whether the presence of a single accessory cusp on the Ij is due to a reduction in the number of cusps or an independently derived character. Character TIO. Ij, Upwardly Curved (Distally). — A, absent; B, present. The distal end of the Ij is straight in certain species of Myosorex, Sylvisorex, and Crocidura. Suncus and most Crocidura species have an upwardly curved Ij. The upturned condition is considered the derived condition by Heim de Balsac and Lamotte (1956). A straight lower incisor is the derived condition, based on the occurrence within out-groups. Character Til. f Lingual Groove.— A, low; B, elevated. The medial side of the lower incisor has a groove that curves above the notch of the basal border in Myosorex, Suncus, and some Sylvisorex (lunaris and ollula), and below the notch in other Sylvisorex (granti and megalura) and Crocidura. An elevated groove is considered the plesiomorphic condition. Character T12. Mj, Complex Talonid with a Fovea and Entoconid.—A, absent; B, present. The loss of the M3 entoconid and reduction of the talonid to a single cusp is seen in species of Suncus, Sylvisorex, Scutisorex, Crocidura, and in some soricines (e.g.. Crypt ot is). Members of Crocidurinae with a well -developed talonid basin of the M3 often retain the distinctiveness of the entoconid and hypoconid. Loss of the M3 entoconid may have a history of frequent parallelism within Soricidae. Heim de Balsac and Lamotte (1957) concluded that evolutionary advancement in crocidurines is associated with simplification of the M3 talonid. A fovea occurs only in species with a complex M3 talonid. The presence of an M3 fovea is most likely primitive, based on the distribution pattern of this character within out-groups. This trait occurs in Myosorex, but has been lost in some species of Suncus, Sylvisorex, and Crocidura. Character T13. P^ Metaconid. — A, absent; B, present. The P4 metaconid is common in Sylvisorex and Myosorex, but often absent in Suncus and Crocidura. The metaconid is lost with evolutionary advancement. Character T14. First Lower Antemolar. — A, tricuspid; B, bicuspid; C, unicuspid. The most primitive living crocidurines (Myosorex spp.) and extinct crocidurines have a tricuspid first lower antemolar, whereas Suncus and Crocidura have a unicuspid first lower antemolar. Therefore, the loss of cusps on the first lower antemolar is the advanced condition. Character T15. Mental Foramen Position Relative to the P^.—A, under; B, posterior. The mental foramen occurred below the P4 in Miocene crocidurines, while in Pliocene to Recent forms, it occurs below the anterior root of the Mj (Butler and Greenwood, 1979). The anterior position is considered the primitive condition. Character Nl. Rostrum Length Relative to Width.— A, short broad (>50%); B, intermediate (<50%, >45%); C, long narrow (<45%). The rostral width just anterior to the interorbital foramen multiplied by 100 and divided by the distance between the posterior margin of the lateral wall of the interorbital foramen and the anterior edge of the l' alveolus was used to describe rostrum shape. Heim de Balsac and Lamotte (1957) concluded that evolutionary advancement is associated with an increase in the size of the rostral portion of the skull. Rostral shape is a character whose polarity is difficult to determine based on the occurrence of this trait within out- groups. Crocidura as well as Suncus and Sylvisorex have rostra that vary from short and broad to long and narrow. Character N2. Maxilla, Zygomatic Process. — A, absent; B, present. The zygomatic process of the maxilla is the site of origin for the external pterygoid muscle. A prominent zygomatic process may occur in species that eat food requiring lateral grinding motions. A prominent zygomatic process is present in Myosorex and Sylvisorex, absent in Suncus, and either present or absent in Crocidura. The presence of the zygomatic process is the primitive condition. Character N3. Maxilla, Lateral Wall of Infraorbital Foramen.— A, broad (width > length); B, narrow (width < length). The lateral wall of the maxilla may be broad or 382 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 narrow, independent of specimen size. A broad lateral wall of the infraorbital foramina is the primitive condition seen in Myosorex. Crocidura as well as the out-groups show both conditions. Character N4. Maxilla, Bimaxillary Width Relative to Breadth Between Infraorbital Foramina. — A, narrow (>60%); B, intermediate (51-59); C, wide (<50%). The breadth between infraorbital foramina multiplied by 100 is divided by the greatest width across the maxillary region to measure bimaxillary width. Bimaxillary width is either intermediate or wide in Myosorex, Sylvisorex, and Suncus, suggesting that a narrow bimaxillary width is the derived condition in Crocidura. Character N5. Lacrimal, Tubercle. — A, absent; B, present. This trait is a projection on the lacrimal just exterior to the lacrimal foramen on the anterior border of the lateral wall of the infraorbital foramen. A lacrimal tubercle is lacking in Myosorex and either present or absent in Sylvisorex, Suncus, and Crocidura. The presence of a lacrimal tubercle is the derived condition. Character N6. Nasals, Lateral Margins. — A, straight, with a gradual medial curve posteriorly; B, convex, with an anterior and a posterior constriction. A convex lateral margin on the nasal bone, seen only in Crocidura, is considered the derived condition. Character PI. Posterior Accessory Incisive Foramen. — A, absent; B, 1; C, 2; D, 3. One to three accessory foramina may be present posterior to the incisive foramina. The incisive foramina transmit the palatine branch of the trigeminal nerve from the maxillary branch (nasopalatine nerves). The function of the accessory foramina, however, is unknown. The plesiomorphic state is the lack of accessory incisive foramina or one accessory incisive foramen, while the occurrence of two or three posterior incisive foramina is the apomorphic state. Character P2. Diastema Between Third or Fourth Upper Antemolar and F*. — A, absent (teeth touching); B, weak (obscured by parastyle in ventral view of skull); C, well- developed (not obscured by F* parastyle in ventral view of skull). The absence of a diastema between the third upper antemolar and P'* is the apomorphic state in Crocidura. Suncus and Sylvisorex have a weak to well-developed diastema between the fourth upper antemolar and P**. An even larger diastema can be obtained with the loss of the fourth upper antemolar. The loss of this diastema may be accomplished either by a shortening of the rostrum or an increase in the robustness of the dentition. Character P3. Diastema, Between M^ and M~ (Lingual). — A, absent (teeth touching); B, weak (obscured by protocone of M“ in ventral view of skull); C, well-developed (not obscured by protocone of M^ in ventral view of skull). The absence of a diastema between M* and M“ is the apomorphic state in Crocidura. Suncus and Sylvisorex have a well-developed diastema between M' and M“. The loss of this diastema may result from an increase in the robustness of the dentition. Character Cl. Cranial Shape (Dorsal View).— A, short, broad (width > length); B, long, narrow (width < length). Cranial shape is calculated by dividing cranial length (distance between the glenoid fossa and the occipital condyle) by the greatest width of braincase. Heim de Balsac and Lamotte (1957) concluded that evolutionary advancement in crocidurine shrews is associated with elongation of the cranium. Myosorex and Sylvisorex have short, broad crania whereas Suncus and some Crocidura species have elongated braincases; thus, the short, broad braincase is most likely the plesiomorphic state. ' Character C2. Cranial Depth (Lateral View).— A, bulbous (braincase depth >55% width); B, flattened (braincase depth <55% width). Cranial depth is measured as the distance between the basioccipital and the highest point on the dorsal surface of the braincase divided by braincase width. Heim de j Balsac and Lamotte (1957) concluded that evolutionary advancement is associated with a flattening of the cranium. Out- group analysis does not generally support this hypothesis. j Sylvisorex, one Suncus, and two Myosorex species have bulbous braincases. Two Suncus species have flattened crania, but both the inflated and the flattened braincases occur in species of » Crocidura. Thus, the inflated state is most likely the ! plesiomorphic state. Character C3. Temporalis Muscle Origin.— A, small (meets sagittal crest at union with lambdoidal crest); B, large (meets sagittal crest anterior to union with lambdoidal crest). This muscle scar is produced by the origin of the temporalis muscle. Myosorex has a small muscle scar, as does the lunaris group of : Sylvisorex. Suncus, some Crocidura, and the granti group of Sylvisorex have large temporalis muscle scars. The polarity is I difficult to determine based on the pattern of occurrence within j[ out-groups. Although this character may reflect dietary ■ differences among species, a lack of data prevents confirmation ; of this idea. Character C4. Interparietal. — A, absent; B, present. An interparietal is present in Myosorex and Scutisorex, but lacking in Suncus and most Sylvisorex and Crocidura. The presence of ‘ an interparietal is the plesiomorphic state for this character. Character 01. Incomplete Fusion Between Temporalis and Squamosal. — A, absent; B, present. When present, incomplete fusion between the temporalis and squamosal persists in adults. This character, absent in Sylvisorex but present in some Suncus, j| Crocidura, and Myosorex, suggests that incomplete fusion between temporalis and squamosal may have a history of parallelism within Crocidurinae. Character 02. Squamosal, Shape of Glenoid Fossa, Dorsal View. — A, rounded; B, angular. The shape of the glenoid fossa jj may be influenced by the size of the external pterygoid muscle. One of its heads originates on the anterior edge of the superior I articular facet of the mandibular fossa. The polarity of this character is difficult to determine based on the study of out- groups. Myosorex has an angular glenoid fossa, whereas Suncus, Sylvisorex, and Crocidura show both conditions. The shape of the glenoid fossa may reflect dietary differences among taxa. II Character 03. Infraorbital Constriction. — A, wide (>55% f of braincase width); B, intermediate (51-54% of base width); C, narrow (<50% of braincase width). This character is a ratio between the least infraorbital width to the greatest width of the | braincase. The polarity of this character is difficult to determine j' based on the occurrence within out-groups. A wide to | 1994 MCLELLAN— Evolution and Phylogeny of African Crocidurine Shrews 383 intermediate infraorbital constriction is seen in Myosorex species, whereas all three conditions are seen in Sylvisorex, Suncus, and Crocidura. However, because a narrow infraorbital constriction is rare in all three genera, a narrow infraorbital constriction may be the apomorphic condition. Character Bl. Basioccipital, Lateral Grooves ExteiuUng Anteriorly to the Basisphenoid. — A, absent; B, present. The rectus capitus muscles, used to flex the head, insert on the basioccipital producing lateral grooves. Polarity of this character is difficult to determine. Lateral grooves are present in Sylvisorex, but both conditions are present in Myosorex, Suncus, and Crocidura. The absence of the basioccipital lateral grooves may be due to a reduction in the size of the rectus capitus muscle or a shift in the insertion. Character B2. Basioccipital and Basisphenoid , Medial Groove. — A, absent; B, present. The longus capitus muscle inserts on the basioccipital and basisphenoid producing the medial groove. This muscle is also used to flex and support the head. The absence of the medial groove may be due to a reduction in the size of the longus capitus. This medial groove is lacking in Myosorex, Sylvisorex, and most Crocidura species, but is common in Suncus. Therefore, this character is a synapomorphy of Suncus and Crocidura. Character B3. Basioccipital and Basisphenoid , Shape of Raised Portion. — A, V-shaped; B, hourglass-shaped; C, Y- shaped. The basioccipital and basisphenoid have raised areas that may appear V-shaped, hourglass-shaped , or Y-shaped. The polarity of this character cannot be determined based on the occurrence within out-groups. Character B4. Petrosal, Medial Process.— A, runs lateral to the basioccipital grooves; B, interrupts the basioccipital grooves. The medial process of the petrosal either runs lateral to the basioccipital grooves or interrupts them. There is a positive correlation between the medial process interrupting the basioccipital grooves and a reduction of the basioccipital grooves. The medial process of the petrosal running lateral to the basioccipital grooves is the plesiomorphic state, seen in Myosorex and Sylvisorex. Both character states are seen in Suncus and Crocidura species. Character B5. Alisphenoid Bone, Position of Vidian Foramen. — A, medial to pyriform fenestra; B, anterior to or parallel with the anterior border of the pyriform fenestra. The vidian foramen carries the vidian nerve. Evolutionary advancement in this character is associated with the vidian foramen which occurs posterior to the anterior edge of the pyriform fenestra. The vidian foramen occurs anterior to or parallel with the pyriform fenestra in Myosorex and most Sylvisorex species, whereas both conditions occur in Suncus and Crocidura. The anterior position is the primitive condition. Character B6. Alisphenoid Bone, Vascular Foramen Size. — A, absent; B, small (similar in size to vidian foramen); C, large (much larger than vidian foramen). The function of this foramen is not known. There are no blood vessels passing through this foramen in C. russula and it begins to ossify with age in some species. Evolutionary advancement is associated with a decrease in size of this foramen. The vascular foramina are large in Myosorex, medium to large in Suncus, and absent to large in Sylvisorex and Crocidura. Character El. Tail, Proportion Covered by Bristles. — A, none; B, proximal Va; C, proximal %; D, all. A bristled tail is a synapomorphy of Crocidura and Suncus, not seen in Sylvisorex and Myosorex. Heim de Balsac and Lamotte (1956) concluded that an increase in pilosity of the tail is associated with evolutionary advancement. A transformation series for this character is difficult to determine. Species of Suncus either have a tail that is fully bristled, or one that is two-thirds bristled. Pilosity of tail in Crocidura varies from species with no tail bristles to those with fully bristled tails. This suggests that tail bristles have been gained and secondarily lost in Crocidura, or that Crocidura is paraphyletic arising from Suncus and Sylvisorex. Character E2. Tail Length Relative to Head and Body Length.— A, short {<^A HB); B, medium ( > '/i HB, < HB); C, long (>HB). Myosorex species have short tails except for M. longicaudatus, Suncus have medium tails, Sylvisorex have medium to long tails, and Crocidura species have a full range of tail lengths. A long tail is the derived condition. However, the presence of a long tail relative to the body length is a character that may have a history of parallelism. Climbing species of Sylvisorex, Suncus, and Crocidura have long tails. The adaptive advantage of having a long tail for balance in climbing species of mammals is well-documented (Hutterer, 1985). Character E3. Body Size (Head-Body Length). — A, small (<80 mm); B, medium (80-110 mm); C, large (>110 mm). An increase in body size has been associated with evolutionary advancement by Heim de Balsac and Lamotte (1957). Myosorex and Sylvisorex are small to medium in body size, where as Suncus and Crocidura range in size from small to large. A large body size is the derived condition. However, the plesiomorphic condition may be either small or medium. Character E4. Foot Length Relative to Head and Body Length. — A, small (<15%); B, medium (16-19%); C, large ( >20%). Suncus species have medium-sized hind feet, but species of Myosorex, Sylvisorex, and Crocidura may have small, medium, or large hind feet. The plesiomorphic state for this character is a medium-sized foot relative to body length. A large hind foot may be associated with climbing ability; Crocidura dolichura and Sylvisorex megalura are known climbers having both long tails and large hind feet. Character E5. Pelage Color, Ventral. — A, same as dorsal pelage; B, slightly lighter than dorsal pelage; C, distinctly lighter. The plesiomorphic state for this character is a ventral pelage slightly lighter than dorsal pelage. Sylvisorex, Suncus, and nonburrowing Myosorex species have the plesiomorphic state. Burrowing species of Myosorex and one species of Crocidura examined do not have a lighter ventral pelage. A distinctly lighter ventral pelage is an apomorphy of Crocidura. Statistical Methods PAUP (Phylogenetic Analysis Using Parsimony, version 2.4; Swofford, 1985) was used to examine phylogenetic affinities among Crocidura (n = 19), Suncus (n — 3), and Sylvisorex (n = 4) taxa. The Outgroup method of PAUP was used to root the 384 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 tree. Four species of Myosorex were examined and a hypothetical ancestor was used as the out-group. The polarity of ten characters could not be determined, so they were treated as unordered. These include: T3, TIO, Nl, N2, N3, Ol, 02, Bl, B3, and El. The FARRIS method with alternate branch swapping, MULPARS, and CONTREE of PAUP were used. Character states assigned to each species are listed in Table 1. CLUSTER, a clustering program of SYSTAT (The System for Statistics, Evanston, Illinois: SYSTAT, Inc., 1986), was used to examine phenetic affinities among species using the same character-state data subjected to parsimony analysis. The Euclidean distance and average linkage methods were used. Results and Discussion Phylogenetic and Phenetic Relationships The maximum of 50 trees was produced by PAUP with a length of 2 1 8 and a consistency index of 25 % . A consensus tree was produced using CONTREE, a consensus tree program of PAUP (Fig. 4). Beginning at the top of the consensus tree, Suncus murinus from northeast Africa, S. varilla from southern Africa, and C. lamottei from west Africa form a monophyletic group. The relationships within this first group are unresolved. Both species of Suncus had to gain an upper antemolar for this set of relationships to be valid. Teeth are generally lost rather than acquired in mammalian evolution. The next monophyletic group consists of C. luna luna from the highlands of Kenya, C. 1. selina from the highlands of Uganda, C. 1. schistacea from the highlands of Kenya and Tanzania, C. fumosa from the highlands of east Africa, and C. batsei from west and central Africa. Crocidura 1. luna and C. /. selina share a close affinity to one another, while C. /. schistacea shows a closer affinity to C. fumosa and C. batsei. This grouping corresponds to the C. luna-fumosa group described by Dippenaar and Meester (1989), with the exception of C. batsei. The third monophyletic group includes C. glassi from the Ethiopian highlands, C. z. zaodon and C. z. tarella both from central and east Africa, and C. littoralis from central Africa. Relationships within this third group are unresolved. Crocidura nigeriae from west Africa and C. tephra from Sudan form the fourth monophyletic group. The fifth group includes C. mariquensis subspecies from southern Africa and C. poensis from west Africa. Crocidura lanosa and C. congobelgica, both from central Africa, form the sixth monophyletic group. The seventh monophyletic group consists of Suncus lixus from southern Africa; Sylvisorex granti from central Africa; Sylvisorex megalura from east, central, and west Africa; and C. dolichura from central Africa. The relationship between Sylvisorex species and C. dolichura within this group remains unresolved. Finally, C. nilotica from central Africa, Sylvisorex lunaris from central Africa, and Sylvisorex ollula from west central Africa are monophyletic species. The results of cluster analysis show a close phenetic affinity among the hypothetical ancestor, Sylvisorex lunaris and Sylvisorex ollula (Fig. 5). In the second cluster, Suncus murinus, C. lamottei, and Suncus varilla form a group. Crocidura lamottei has a close phenetic affinity to Suncus and belongs to the subgenus Afrosorex (Hutterer, 1986). Members of the subgenus Afrosorex are savanna species that share several derived characteristics (Hutterer, 1986). Crocidura congobelgica and C. dolichura show a close phenetic affinity to the members of the Sylvisorex granti group (Sylvisorex megalura and Sylvisorex granti) in the third cluster. Crocidura batsei, C. fumosa, and C. 1. schistacea form the fourth cluster. The fifth cluster includes C. 1. selina, C. 1. luna, C. tephra, C. nigeriae, and C. nilotica. The sixth cluster includes C. m. shortridgei, C. m. mariauensis, C. poensis, and C. littoralis. Crocidura glassi, C. z- tarella, and C. z. zaodon form the seventh cluster. The last group, at the bottom of the phenogram, includes Suncus lixus and C. lanosa. Parsimony analyses suggest a close relationship between the three C. luna subspecies and C. fumosa, whereas cluster analysis does not show a close phenetic affinity. This association noted in parsimony analysis corresponds to the luna- fumosa group of Dippenaar and Meester (1989), which is based on cranial and dental features. However, Butler et al. (1989) placed C. fumosa in the fumosa group (along with C. maurisca, C. dolichura, and mariquensis), and C. luna in the C. turba group (along with C. monax and C. foxi). Suncus does not form a monophyletic group. Suncus murinus and S. varilla form a monophyletic group along with C. lamottei, whereas Suncus lixus shows a close affinity to the Sylvisorex granti group. Crocidura dolichura is more closely related to the Sylvisorex granti group than to the other species of Crocidura examined. Based on these associations, monophyly in Crocidura, Suncus, and Sylvisorex seems unlikely. The two Sylvisorex species groups, granti and lunaris, are distinctive enough to be considered separate subgenera based on the study of only four of seven species. Thirteen characteristics separate Sylvisorex granti and Sylvisorex megalura from Sylvisorex lunaris and Sylvisorex ollula (Table 2). The remaining species of Sylvisorex must be studied before it can be determined whether subgeneric boundaries should be defined. The relationships of Sylvisorex, Suncus, and Crocidura are unclear. The results of this study suggest that Suncus varilla, S. lixus, S. murinus, and the Sylvisorex granti group have evolved from Crocidura through the acquisition of an upper antemolar. However, teeth are lost more often than gained in mammals over evolutionary time. An alternative explanation is that a long history of co-occurrence in Africa (12 million years) may have led to many parallel and convergent evolutionary trends in morphology, and through extinction, Suncus and Sylvisorex may have lost many primitive members, leaving behind a few primitive and some relatively advanced species. Historical Biogeography and Speciation The concept that animal speciation in Africa has been strongly influenced by cycles of arid and moist climatic phases is well-established (Moreau, 1952). These fluctuations in climate apparently isolated some animal species and led to the distribution expansions of others. During the Pleistocene, montane biomes extended from the highlands of Cameroon to the Ethiopian highlands and to southern Africa (Grubb, 1978; Moreau, 1952, 1962). Following the Pleistocene, the forest areas decreased in size, resulting in the fragmentation of west 1994 MCLELLAN— Evolution and Phylogeny of African Crocidurine Shrews 385 and central African lowland forests, and the isolation of numerous islands of montane forest at higher elevations which have become refugia (Fig. 6; Grubb, 1978; Robbins, 1978). The pattern of fragmentation, however, is not clear. Few African mammal taxa have a sufficiently large geographic distribution and adequate geographic variation to recognize centers of dispersal or refugia. Crocidura, however, is an excellent model for studies of African and Old World biogeography: it is speciose; has low vagility; and occurs throughout Europe, Africa, Asia, Indonesia, and the Philippine Islands. The African montane species of Crocidura possessing primitive dental characteristics might represent fragmented relics of formerly widespread species that have subsequently diverged within these isolated areas. Examples of this pattern are seen in C. luna and C. fumosa from the highlands of Kenya, Uganda, and Tanzania, and in C. glassi from the highlands of Ethiopia. Species adapted to drier and more open habitats, such as C. lamottei, appear to be phylogenetically advanced and probably evolved in Recent times with the expansion of the Sahara Desert and the retreat of the central forest region, as suggested by Hutterer (1986). Conclusions Crocidura, Suncus, and Sylvisorex do not represent monophyletic lineages. The generic distinctions among Crocidura, Suncus, and Sylvisorex, based on tooth number, are not maintained when additional characteristics are included in parsimony or cluster analyses. Similarly, the species groups that have been defined vary according to which characters are included in the analysis. A lack of fossil data and a high degree of homoplasy in cranial and external morphology warrants the use of additional characters such as soft tissue anatomy, karyotypes, and biochemical data to derive testable hypotheses. More evolutionary data clearly are needed both for testing these hypotheses and for developing more rigorous models of speciation. Literature Cited Butler, P. M., and M. Greenwood. 1979. Soricidae (Mammalia) from the Early Pleistocene of Olduvai Gorge, Tanzania. Zoological Journal of the Linnean Society, 67:329-379. Butler, P. M., R. S. Thorpe, and M. Greenwood. 1989. Interspecific relations of African crocidurine shrews (Mammalia: Soricidae) based on multivariate analyses of mandibular data. Zoological Journal of the Linnean Society, 96:373-412. Dippenaar, N. J., and J. A. J. MEESTER. 1989. Revision of the luna-fumosa complex of Afrotropical Crocidura Wagler, 1832 (Mammalia: Soricidae). Annals of the Transvaal Museum, 35:1-47. Grubb, P. 1978. Patterns of speciation in African mammals. Pp. 152-167, in Ecology and taxonomy of African small mammals (D. A. Schlitter, ed.). Bulletin of the Carnegie Museum of Natural History no. 6, 214 pp. Heim de Balsac, H., and M. Lamotte. 1956. Evolution et phylogenie des Soricides Africains. I. La lignee Myosorex- Surdisorex. Mammalia, 20:140-167. 1957. Evolution et phylogenie des Soricides Africains. II. La lignee Sylvisorex-Suncus-Crocidura. Mammalia, 21:15-49. Honacki, j. H., K. E. Kin man, and J. W. Koeppl. 1982. Mammal Species of the World: A Taxonomic and Geographic Reference. Allen Press and the Association of Systematics Collections, Lawrence, Kansas. 694 pp. Hutterer, R. 1985. Anatomical adaptations of shrews. Mammal Review, 15:43-55 1986. African shrews allied to Crocidura fischerc. Taxonomy, distribution and relationship. Cimbebasia, (A)4:23-35. Maddalena, T. 1989. Systematics and biogeography of Afrotropical and Palaearctic shrews of the genus Crocidura (Insectivora: Soricidae): An electrophoretic approach. Pp. 297-308, in Vertebrates in the Tropics (G. Peters and R. Hutterer, eds.). Museum Alexander Koenig, Bonn, 424 pp. Moreau, R. E. 1952. Africa since the Mesozoic: With particular reference to certain biological problems. Proceedings of the Zoological Society of London, 121 :869-913. .. 1962. Vicissitudes of the African biomes in the late Pleistocene. Proceedings of the Zoological Society of London, 141 :395-419. Nowak, R. M. 1991. Walkers’s Mammals of the World. 5th ed. Johns Hopkins University Press, Baltimore, 1:1-642 pp. Robbins, C. B. 1978. The Dahomey Gap — A reevaluation of its significance as a faunal barrier to the west African high forest mammals. Pp. 168-174, in Ecology and Taxonomy of African Small Mammals (D. A. Schlitter, ed.). Bulletin of Carnegie Museum of Natural History no. 6, 214 pp. SWOFFORD, D. L. 1985. PAUP (Phylogenetic Analysis Using Parsimony). Version 2.4. Illinois Natural History Survey, Champaign, Illinois (on disk). Smithers, R. H. N. 1983. The mammals of the southern African subregion. University of Pretoria, Republic of South Africa, 736 PP- Watrous, L. E., and Q. D. Wheeler. 1981. Out-group comparison method of character analysis. Systematic Zoology, 30(1):1-11. 386 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 Table \ .— Character states for species included in parsimony and cluster analysis. Character abbreviations refer to anatomical regions as follows: tooth characters (Tl-15), naso-maxillary region (Nl-6), palatal region (PI -3), braincase (Cl-4), orbito- temporal region (01-3), basicrania (Bl-6), external features (El-5). For a more detailed description of each character, see survey of the characters in Materials and Methods. Character State TTTTTTTTTTTTTTTNNNNNNPPPCCCCOOOBBBBBBEEEEE Species 12 3456789012345123456123123412312345612345 Ancestor S. murinus S. lixus S. varilla Sy. lunaris Sy. granti Sy. ollula Sy. megalura C. 1. luna C. 1. selina C. 1. schistacea C. fumosa C. m. mariquensis C. m. shortridgei C. glassi C. lamottei C. poensis C. batsei C. nigeriae C. z. zaodon C. z. tarella C. littoralis C. lamosa C. dolichura C. congobelgica C. tephra C. nilotica AABABBBBBABBBAACBABAAAACABAABBBAACABCAABBB ABBABABBABBAACABABCAABCCBBBABBCBBBBBBDBBBA AAAAABBBABAABCAAAABAA7CCBBBAABABBBABBCBA?? ABBBAAABABAAACBAABCBAABCBBBABBAAAABACDAABB AABABABBDABBBBBBAABBAABCAAAAABCBAAABAABBBB AAAAABBBCAABBBBAAAABAACCBABAAAABABAABACACB AABBBABBBABABAACBBBBABBCAAAAABBBBBABAABBBB AAAAAABBBBABBCAABAAAAACCBABBAAABACABCACACB ABBABABAABBBBCBAAABAABCCAABABBABACBACCBBBB ABBAAABAABBBBCBAAABAABCCAABABBABACBABBBBBB ABAAAABAABBABCBABABAABBBAABABBABAAAABCBBBB AAAAAABABABABCAABAAAABBBAAAABAAAAABACCBBBB AABBBABABBBBACBCAACAABCCBABAAAABBBAAABBACB AABBAABABBBBACBCABBAABCCBABABAABACAABBBBCB AAABABBAAABAACACABBBBBBBBABAAAABABBAABBBBB BBBBBAAAABBAACBBAACBABBBBBBABBBBAABABDABAA AAABAABABBBAACBBAABBABCCBBBABAABABABBCBBCB ABAAAABABBBBBCBBAABBABABABBABBABAABBBBBBBB AAAAAAAAAABABCBBABBBABCCBBBABBABACAACBBBBB AAABABBABABBACBBAABBBBBBBABABBBBACAABDBBBB AAABABBAAABBACBBAABBBBCBBABABBABACAAACBBBB AAABAABABBBBACBBABBBABBCBAAABBABACAAAABBBB AABBAABABBBABCACABBAA7CCAABBBBAAAABAAABBCB AAAAABBABBBBACAAAAABA2CCBABBAABAABBAAACACB AAABAABABBBAACAAAABBABCBBABBBBBAABBACABABB ABABAABAABBAACBBABBBABCCAABABBBBACBABCBBBB ABABABBABABABCBCAABBABCCAABAABBBABABCCBBBB Table 2. — Diagnostic features observed in the Sylvisorex granti and lunaris groups. granti Group lunaris Group Character (megalura) (lunaris) Cusp-like posterior lingual cingula on the upper incisor absent present hypocone absent present Height of lingual groove on the lower incisor low high Rostrum length short long Bimaxillary width narrow wide Size of diastema between the third upper antemolar and R* large small Braincase shape long short Size of temporalis muscle scar large small Interorbital constriction width wide narrow Vascular foramen present absent Tail length relative to head and body length equal to or longer shorter Body size small large Hind foot size large small to medium 1994 MCLELLAN — Evolution and Phylogeny of African Crocidurine Shrews 387 Cr oc I d ur a Fig. 1. — Phylogenetic affinities among African crocidurine shrews based on cranial and dental morphology from Heim de Balsac and Lamotte (1957). Fig. 2. — Phylogenetic affinities among species of Sylvisorex, with possible branching points of S uncus, and Crocidura based on multivariate analysis of mandibular morphology following Butler et al. (1989). 388 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 C Sy! visa rex megalura Sylvisorex lu nan's Crocidura luna Crocidura bottegi Crocidura russula Crocidura zimmermanni Crocidura suaveolens Crocidura leucodon Crocidura canariensis Crocidura sicuia Crocidura hidegardeae Crocidura nigrofusca Crocidura poensis Crocidura theresae Crocidura lusitania Crocidura crossei Crocidura nanilla Crocidura jouvenetae Crocidura fuscomurina Crocidura ollvieri Crocidura viaria Crocidura lamottei Crocidura flavescens Fig. 3. — Cladistic relationships among 21 Crocidura species based on 25 polymorphic loci from Maddalena (1989). 1994 MCLELLAN— Evolution and Phylogeny of African Crocidurine Shrews 389 {Z Ancestor S. murinus north east Africa S. variUa southern Africa C. lamottei west Africa ' C. /. luna highlands of Kenya and Tanzania ’ C. 1. se/ina highlands of Uganda ' C. /. schistacea highlands of Kenya and Tanzania ' C. fumosa highlands of east Africa ' C. batsei west and centra! Africa ‘ C. g las si Highlands of Ethiopia ' C. z. zaodon east and centra! Africa ■ C. z. tareila east and centra! Africa ■ C. iittoralis cendal Africa C. nigenae west Africa ' C. tephra Sudan ■ C. m. mariquensis southern Africa ■ C. m. shortridgei southern Africa ■ C. poensis west Africa ■ C. lanosa centra! Africa • C. congobelgica centra! Africa ■ S. southern Africa 5/. grand centra! Africa ■ Sy. megatura east, central, and west Africa ■ C. doUchura centra! Africa ■ C. nilodca centra! Africa • S/. oHuHa west centra! Africa “ S/. lunaris centra! Africa Fig. 4.— Phylogenetic relationships among Suncus, Sylvisorex, and Crocidura based on parsimony analysis of 42 morphological characters. The geographic distribution of each species is listed to the right. 390 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 Ancestor Sy! visor ex oUula Sylvisorex lunaris SunciJS murinus * Crocidura lamottei Suncus varilla ■ Crocidura congobelgica , Crocidura doiichura Syivisorex megaiura Syivisorex grand Crocidura batsei Crocidura fumosa . Crocidura i. schistacea Crocidura i. seiina Crocidura /. luna Crocidura tephra Crocidura nigeriae Crocidura niiotica Crocidura m. shortridgei Crocidura m. mariquensis Crocidura poensis Crocidura iittoraiis Crocidura giassi Crocidura zaodon tareiia Crocidura zaodon zaodon Suncus iixus Crocidura ianosa 1.00 0.70 0.60 i. 0.37^ 0.66 0.58 0.81 0.56 " 0.64 u 11 0.51 : 0.75 0.56 : 0.49 w . -I w II 0.69 ' 0.27 ii 0.48 !! 0.54 i 0.60 0.63 0.44 I ^ - II 0.51 0.55 li II 0.61 0.59 0.35 II 0.70 !! 0.72 Fig. 5.— Phenetic affinities among Suncus, Syivisorex, and Crocidura based on cluster analysis of 42 morphological features using the Euclidean distance and the average linkage method of SYSTAT (System for Statistics, 1986). Distance values between groups are listed on the right side. 7.50 1994 MCLELLAN — Evolution and Phytogeny of African Crocidurine Shrews 391 200 "T 200 Fig. 6.— Faunistic divisions of the forest biome in Africa, with presumed major refugia of extreme arid phases stippled and more Recent refugia of less extreme arid phases in intervening zones. W, western region, Dahomey Gap on the eastern boundary; WC, west-central region, bounded by the Oubangui-Congo nvers, Sanaga River (S), and Ogoue River (O); EC, east-central region, bounded by the Congo River; SC, south-central region; E, eastern region, the forest is restricted to small mountains and coastal forest Islands. Taken from Grubb (1978). 097. ■ i|S . ■ f ■ ••w# H'.-MiEtMSP ;:^Mt«;'.s3t-jar*:w«f5iS^^W53s^ ■■■. - . .■-, ■: •• ■ ■■ .-.-Ji-,;,..; •— . „, .,... - . C>. \ .•' j h- jTf-~ -rii'' , 4l • ,' Itc-. ' ,X /I ti/-, ,1.1 (?^'/q)'yni'iax«oiK!tyi"h i<- ♦vi, r^' <• ■, > fV»,'. 'M UH^ t:.-. l/^ IftlijitO UflSR' I IDENTIFICATION OF THE CAROLINIAN SHREWS OF BACHMAN 1837 Charles O. Handley, Jr.* and Merrill Yarn' 'Division of Mammals, National Museum of Natural History, Smithsonian Institution, Washington, D.C. 20560 Abstract In his 1837 monograph of North American shrews, Bachman named seven new species and redescribed six of other authors. These 13 names represent seven currently recognized taxa. Of particular interest are the species with which Bachman was most familiar, Sorex carolinensis ( = Blarina carolinensis), S. cinereus ( = Cryptotis parva), and S. longirostris , which Bachman described from the vicinity of Charleston, South Carolina. Topotypical material had been too meager to be useful, but by means of formalin-charged pitfalls, we obtained near-topotypical series of all three species. We have used these specimens to redefine Bachman’s taxa. The names Sorex carolinensis and S. longirostris are correctly applied in current nomenclature, but S. cinereus is not a synonym of Cryptotis parva parva Say as previously thought. Introduction Three species of shrews {Sorex longirostris, Blarina carolinensis, and Cryptotis parva) are common and widespread in the lowlands of the southeastern United States. All three were discovered by John Bachman in the vicinity of Charleston, South Carolina. He described and named them in 1837 in a monograph of the shrews of North America. Bachman’s monograph contained descriptions of 13 species of shrews in North America north of Mexico. These represent seven species in today’s nomenclature. Bachman stated that shrews were the least studied and most poorly collected group of North American mammals. He believed that more species would be found. The most recent checklist for the same area (Jones et al., 1986) listed 31 species of shrews. As a taxonomist Bachman was ahead of his time. His descriptions are easily interpreted, and they are thorough without being cluttered with minutia. His species accounts consist of a brief diagnosis (“Characters”), dental formula, dentition, form, color, measurements, comparisons, distribution, and habits. He recognized that species are variable and perceived the need for collecting series of specimens. Some of Bachman’s specimens still exist in the Academy of Natural Sciences of Philadelphia (ANSP) and in the National Museum of Natural History, Smithsonian Institution (USNM). However, they are in poor condition and difficult to interpret in the absence of good supporting material. Until recently, topotypical specimens have not been available, and it has been necessary to use specimens from northeastern Georgia and central North Carolina to represent Bachman’s species (Miller, 1895:40; Jackson, 1928:86). Consequently, application of Bachman’s names at the level of subspecies has been clouded with doubt. In current nomenclature Bachman’s Sorex longirostris and Sorex carolinensis are the nominate subspecies S. 1. longirostris and Blarina c. carolinensis. Sorex cinereus is thought to be a synonym of Cryptotis parva parva. A major objective of this project was to collect series of shrews in coastal South Carolina, near Charleston, where Bachman obtained his specimens. We were spurred by questions on several fronts to obtain fresh series of Bachman’s shrews. For example, we wondered whether the federally threatened Sorex longirostris fisheri Merriam might be widespread in coastal Virginia and the Carolinas, rather than restricted to the Dismal Swamp. Could it be the same as Bachman’s Sorex longirostris! Also we were suspicious, after studying his description, that Bachman did not have a specimen of Cryptotis parva parva in hand when he described Sorex cinereus. Bachman’S Shrews From literature Bachman knew six species of North American shrews, described from the northern United States and Canada prior to 1837. We list them here with page numbers in Bachman (1837), in order of their current nomenclatorial status. Note that in Bachman’s day all shrews were included in the Linnaean genus Sorex. Sorex personatus I. Geoffroy Saint-Hilaire 1827 Page 398; a synonym of Sorex cinereus Kerr (see Jackson, 1925:55). Sorex forsteri Richardson 1828 Page 386; a synonym of Sorex cinereus Kerr (see Miller, 1895:40). Sorex palustris Richardson 1828 Page 396; Sorex palustris Richardson. Sorex brevicaudus Say 1823 Page 381; Blarina brevicauda Say. Sorex talpoides Gapper 1830 Page 397; Blarina brevicauda talpoides Gapper. Sorex parvus Say 1823 Page 394; Cryptotis parva Say. In addition to these six established taxa of North American shrews, which Bachman treated briefly in his 1837 monograph, he described seven more as new species. Three of the new species, the Carolinian shrews which were based on his own material, Bachman described in considerable detail. They are the subject of this paper. The other four new species, based on references in literature and on specimens supplied by correspondents, received more cursory treatment by Bachman. We list the new species here, with page numbers in Bachman (1837), in order of their current nomenclatorial status; the Carolinian shrews are indicated with an asterisk. Sorex richardsoni Bachman Page 383; a synonym of Sorex arcticus Kerr (see Jackson, 1925:55 and Miller, 1895:38). Sorex cooperi Bachman Page 388; a synonym of Sorex cinereus Kerr (see Miller, 393 394 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 1895:41). *Sorex longirostris Bachman Page 370; Sorex longirostris Bachman (see Miller, 1895:40, 52-53, 56). Sorex dekayi Bachman Page 377; a synonym of Blarina brevicauda talpoides Gapper (see Bangs, 1902:75 and Merriam, 1895:10). *Sorex carolinensis Bachman Page 366; Blarina carolinensis Bachman. Regarded as a subspecies of Blarina brevicauda Say by Merriam (1895:14) and subsequent authors until restored to the status of a species by Handley (1971:300). See also Tate et al. (1980). *Sorex cinereus Bachman Page 373; Cryptotis parva floridana Merriam. A homonym of Sorex cinereus Kerr. Long thought by all authors to be a synonym of Cryptotis parva parva Say; shown in this paper to be a synonym of Cryptotis parva floridana Merriam. Sorex fimbripes Bachman Page 391. Enigmatic; unidentified. Most authors (e.g.. Hall, 1981:28) have disregarded the details of Bachman’s description and have regarded Sorex fimbripes as a synonym of Sorex cinereus cinereus Kerr. However, there is no reason to believe that Bachman did not describe Sorex fimbripes exactly as he saw it in his hand. Hollister (1911:381) was correct in rejecting the assignment of Sorex fimbripes to the synonymy of Sorex cinereus and in his assessment, “The description of Sorex fimbripes differs so widely from any known American shrew that the name is probably unidentifiable.” We venture that it is only unidentifiable because the holotype is lost (Hollister, 1911) and the animal has never been collected again. In form and pelage, but not in size, coloration, or dentition, this shrew is reminiscent of Nectogale elegans of the eastern Himalayas. Otherwise it is unlike any known shrew. Sorex fimbripes Bachman is the type species of Hydrogale Pomel, 1848, a name preoccupied by Hydrogale Kaup, 1829. Study Area and Methods We collected within a few kilometers of the coast, in and near Francis Marion National Forest, 40-60 km northeast of Charleston, Charleston County, South Carolina. Collecting localities are listed in the species accounts. For Sorex longirostris we sought early to midsuccession mixed forest with herbaceous ground cover on damp organic soil, such as is occasionally found on swamp edges (Rose and Padgett, 1991). For Cryptotis parva we initially sought grassy ditch and dike banks in Spartina salt marsh. Eventually we settled for dense, not recently cut, grass hayfields. We expected to catch Blarina carolinensis wherever we set traps, except in salt marsh. We used Museum Special snap traps to catch specimens to be prepared as dry study skins. We used pitfalls for specimens to be stored in alcohol after extracting skulls. We used six topless 2-liter plastic bottles (11 cm diameter X 20 cm deep) and a 1-gal metal can (15 cm diameter X 17.5 cm deep) in arrays of seven pitfalls each. The pitfalls were arranged in a three-leaf clover pattern (120° between arms), with the gallon can at the center and 2-liter bottles on either side, near the distal end of each arm (drift fence). The drift fences, made of aluminum siding, 1.2 m long by 30 cm high, met at the central can. An array fit into a triangle a little less than 2.5 m from comer to comer. From July 1989 to September 1990 we had 154 pitfall traps distributed in 22 arrays of seven pitfalls each in three transects on swamp edges. In addition, from January to September 1990 we had 45 pitfalls, without drift fences, at 10-m intervals in two transects on shmb-bordered ditch banks in sandy weed fields. Because of ants and crabs, snap trapping proved infeasible during warm seasons. In January and Febmary 1990 we used Museum Special snap traps baited with rolled oats or mouse parts (attached to trap treadles with twist-ties) in 100-trap transects in hayfields and weed fields to capture Cryptotis parva and Blarina carolinensis. On 21 September 1989 Hurricane Hugo devastated coastal South Carolina with a tidal surge up to 5.9 m (19.5 ft) above normal high tide that inundated the salt marshes and adjacent coasts, and winds of 217 km/h (135 mph) that leveled the forests. In our study area more than 80% of the trees were uprooted or broken off. The storm eliminated any chance of catching Cryptotis parva in marshes, but it dramatically improved habitat for Sorex longirostris on the swamp margins. What had been dark, closed canopy forest with patchy herbaceous ground cover became open and surmy, with scattered trees and dense, continuous herbaceous ground cover. Within six months we were catching more Sorex longirostris, as well as open country mammals such as Cryptotis parva and Reithrodontomys humulis, where there had been forest. Measurements. — Only specimens judged to be adult by tooth wear— molariform teeth showing little to moderate wear (subadult and adult of Choate, 1970) — were measured. Individuals with unworn teeth (young) or excessively worn teeth (old adult) were not measured. External measurements, assumed to have been taken in the conventional manner, are from specimen labels. Shrews immersed in liquid in pitfalls shrink or stiffen and yield an erroneously small total length. Thus, total length of shrews from fluid-charged pitfalls is not comparable to total length of shrews measured fresh from snap traps or live traps, so it is not included in the tables. Lengths of tail, hind foot, and ear do not change appreciably in liquid and were used in our tables of measurements. Cranial measurements were made with dial calipers and a binocular microscope as described by Jackson (1928), Choate (1970), and Junge and Hoffmann (1981). Abbreviations of measurements in the tables are as follows: condylobasal length (CBL), palatal length (PAL), maxillary breadth (MAB), interorbital breadth (lOB), maxillary toothrow length (MTR), cranial breadth (CRB), total length (TL), tail vertebrae (TA), hind foot (HF), and ear from notch (EA). 1994 HANDLEY AND VARN-Shrews of Bachman 1837 395 Results and Discussion Sorex longirostris Bachman 1837 ISorex personatus \. Geoffrey Saint-Hilaire, 1827:122. Sorex longirostris Bachman, 1837:370. Oftisorex] longirostris: DeKay, 1842:23. Corsira forsteri: Lesson, 1842:89 = 7 Sorex longirostris Bachman. Musar[aneus] (CroefiduraJ) bachmani Pomel, 1848:249. Sorex personatus: Baird, 1857:30. Sorex wagneri Fitzinger, 1868:512. Type locality. — Swamps of the Santee, Hume Plantation, Cat Island, Georgetown County, South Carolina. Cat Island is on the north side of the mouth of the Santee River. Holotype. — Academy of Natural Sciences of Philadelphia ANSP 479, “Mounted skin, skull inside except for rostrum which has been extracted” (Koopman, 1976). Collected by Alexander Hume. Description. — Sorex longirostris of coastal South Carolina is a small long-tailed shrew; total length usually less than 90 mm; tail a little more than one-third of the total length (mean 31 mm); tail, forefeet and hindfeet small and delicate; tail naked in reproducing old adults of both sexes; ears hairy and relatively large, protruding conspicuously through fur; eyes minute but visible; vibrissae long, conspicuous, reaching ear when laid back. Coloration of upper parts uniformly brown, except flanks slightly paler, and vibrissal area black (“masked”); underparts pale orange-brown, not sharply defined from color of flanks; tail fuscous on dorsal surface and on distal third of ventral surface, buff on proximal two-thirds; forefeet and hindfeet pale buffy brown. Rostrum unusually short and broad for a Sorex (palate averages 39.9% of the condylobasal length and maxillary breadth 29.7 % of the condylobasal length); postmandibular foramen absent (n = 22); tooth formula I 1/1, U 5/2, P 1/0, M 3/3 = 10/6 X 2 = 32; medial tine of upper incisor (where upper incisors touch one another) pigmented and within the pigmented area of the main cusp (9 of 16 cases), pigmented and contiguous with pigmented area of main cusp (5 of 16 cases), and unpigmented and separated from the pigmented area of the main cusp (2 of 16 cases); upper unicuspid row not particularly crowded; usually smaller than U'* (16 of 22 cases), subequal in six cases (27%); fully visible in lateral view in 21 of 22 cases, partly hidden in one; only the tips of the upper unicuspids pigmented and all (// = 21) lack a pigmented ridge extending down to the cingulum on the lingual faces of the teeth. Measurements. —See Table 1. Comparisons. — Sorex longirostris from Charleston County, South Carolina, is medium-sized for the species and has a relatively long tail, short rostrum, and broad skull. Shrews with these characteristics are found in the Coastal Plain and Piedmont south and southwestward from coastal South Carolina to Georgia, Florida, Alabama, and Mississippi. Shrews are slightly smaller in the Ohio and Mississippi valleys, north Georgia, and in the interior of the Carolinas, Virginia, and Maryland (Table 1). Sorex longirostris ftsheri Merriam from the Dismal Swamp, southeastern Virginia and adjacent North Carolina is larger and has a relatively longer tail and longer. narrower rostrum. South Carolinian shrews do not resemble S. 1. ftsheri. They do, however, resemble S. 1. eionis Davis of the Gulf Coast of Florida more than do specimens of S. 1. longirostris from the interior. Compared with coastal S. longirostris, S. 1. eionis has a longer head and body, relatively shorter tail, and cranial measurements averaging slightly larger (Davis, 1957). Nomenclature.— Sorex personatus I. Geoffroy Saint-Hilaire (1827) may be the oldest name for the southeastern shrew. Beginning with Baird (1857), S. personatus was used as the name for the southeastern shrew. In naming S. personatus Geoffroy Saint-Hilaire cited only the United States as its type locality. However, Baird (1857:31) believed that “The original specimen [of S. personatus^ was collected in some one of the Atlantic States by Milbert, probably somewhere in the south...,” i.e., within the range of S. longirostris. Baird went on to say, “It is with much pleasure that I am enabled to identify the hitherto obscure Sorex personatus of Geoffroy. A comparison of the specimen before me [USNM 637/1788 from Washington, D.C.] not only with the description, but with the beautiful figure given in Guerin’s Magazin de Zoologie, 1833, shows a much more than usual agreement between the two in color, shape, dimensions, etc.” On the basis of this comparison, Baird lodged S. longirostris in the synonymy of S. personatus. Incidentally he was correct in his identification of USNM 637/1788 as a southeastern shrew. Baird described the skin and teeth of that specimen in such detail that together with his table of measurements it is certain that, in spite of a fragmentary skull, the individual now labeled USNM 637/1788 is the specimen Baird described. Miller (1895:40) called attention to Baird (1857:31) and raised the possibility that the reference pertained to Sorex longirostris Bachman. Miller examined USNM 637/1788, then without a skull, and pronounced it “wholly unidentifiable.” He included the Baird reference with question in the synonymy of S. personatus I. Geoffroy Saint-Hilaire. Of S. personatus he said “...the original specimen was collected by Milbert in the United States, possibly in New York (Milbert collected the type of Rhinichthys cataractae Cuv. and Val. at Niagara Falls, N.Y.). The description is sufficiently accurate to show that the animal was the smaller common Long-tailed Shrew of the eastern United States [5. personatus = S. cinereus]." Another perspective on S. personatus came from Merriam (1895:62): “Respecting the pertinence of the name personatus for this Shrew [the common small Sorex of eastern North America], Dr. G. E. Dobson wrote me from Netley, England, under date of October 5, 1885, as follows: T have lately returned from Paris, where I have been studying the Soricidae in the Museum of the Jardin des Plantes. I have found there the type of Sorex personatus Geoff, which is certainly = S. cooperi, the latter name becoming therefore, a synonym.’” Of S. cooperi. Miller (1895:41) said: “The Sorex cooperi which Bachman named in 1837 is without doubt the present species {S. personatus).” Thus, authority for identification of Sorex personatus I. Geoffroy Saint-Hilaire as the common eastern North American masked shrew rests not with Baird, Miller, or Merriam, but 396 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 with Dobson as quoted by Merriam (1895). Dobson was the only one of the four who actually examined the holotype of S. personatus. We do not know whether the holotype still exists. Miller did not mention it in his manuscript notebooks describing types in European museums. Neither did Rode (1943) in his catalog of types of Insectivora in the Museum National d’Histoire Naturelle, Paris. We do not believe that any of these determinations (Baird, Dobson, Miller, or Merriam) can be taken seriously, since Dobson did not know what all the possibilities were and could not have made pertinent comparisons, and neither Baird, Miller, nor Merriam had more than nondefmitive descriptions to go by. Without a holotype or a definite type locality we believe Sorex personatus I. Geoffroy Saint-Hilaire is unidentifiable. Miller (1895) associated the name S. personatus with the northern masked shrew, and Jackson (1925) synonymyzed it with S. cinereus Kerr. Sorex longirostris Bachman (1837) is the earliest name that pertains unequivocally to the long-tailed shrew of the southeastern United States. Recognition of this fact dates from Miller’s (1895:52) redescription of the species. Musaraneus bachmani Pomel (1848) and Sorex wagneri Fitzinger (1868) are renamings of Sorex longirostris Bachman. Specimens examined. — All in the National Museum of Natural History unless otherwise noted. S. 1. eionis: Florida: Wakulla County: St. Marks National Wildlife Refuge, St. Marks Unit, 1 skin and skull. S. 1. fisheri: All localities are in the Great Dismal Swamp National Wildlife Refuge. North Carolina: Gates County: Cross Canal, 3 mi SW Corapeake, 4 skulls; Weyerhauser Ditch, 1.6 mi N North Carolina Highway 158, 5 skulls. Virginia: City of Chesapeake (= Norfolk County): East Ditch, 6 skulls; Feeder Ditch, 3 skulls; Portsmouth Ditch, 1 skull; 4.7 mi NNE Wallaceton, 1 skin and skull. City of Suffolk (= Nansemond County): Badger Ditch, 2 skulls; Jericho Ditch, 6 skulls; Lake Drummond, 7 skins and skulls (including holotype of S. 1. fisheri), 2 alcoholics and skulls; Railroad Ditch, 1 mi E Desert Road, 5 skulls; West Ditch, 1 skull. S. 1. longirostris: Alabama: Chambers County: 2 mi N Gold Hill, 12 skins and skulls. Arkansas: Benton County: 3 mi N War Eagle, 1 alcoholic and skull (University of Arkansas, Department of Zoology). Indiana: Tippecanoe County: 10 mi W Lafayette, 6 skins and skulls. Mississippi: Noxubee County: Macon, 1 skull. North Carolina: Wake County: Raleigh, 3 skins and skulls. South Carolina: Charleston County: Iron Swamp, 4.0 km SW Awendaw, 7 alcoholics and skulls, 2 alcoholics; Head of Mill Branch Swamp, 3.3 km NW McClellanville, 15 alcoholics and skulls, 27 alcoholics Virginia: Amelia County: Amelia Court House, 3 skins and skulls, 1 skin. Brunswick County: Seward Forest, near Triplett, 2 skins and skulls. Chesterfield County-Powhatan County line: Keswick Farm, 4 mi N Midlothian, 1 skin and skull. Hanover County: 1.2 mi S Montpelier, 1 skull. Montgomery County: Blacksburg, 2000 ft, 5 alcoholics and skulls (Virginia Commonwealth University). Sorex carolinensis Bachman Sorex carolinensis Bachman, 1837:366. Blarina brevicaudata: Lesson, 1842:89 = ISorex carolinensis Bachman. Sorex (Anotus) carolinensis: Wagner, 1855:550 (part.). Blarina carolinensis: Baird, 1857:45. Blarina brevicauda: Allen, 1869:221. Blarina brevicauda carolinensis: Merriam, 1895:13. Type locality. — Bachman (1837:368) specified “South Carolina”, “...both in the upper and maritime districts.” The only specific area Bachman mentioned was Abbeville District in western South Carolina. Merriam (1895:13) restricted the type locality to “Eastern South Carolina”. We further restrict the type locality to Charleston County, South Carolina, because Bachman lived in Charleston and was most familiar with the local fauna. He specified that S. carolinensis had been known to him for “nearly twenty years. ” He also mentioned seeing burrows of this shrew in plowed ground and ditch banks. Furthermore, the species that Bachman (1837:366) described, total length 3% inches (108 mm), was unusually large for Blarina carolinensis. We believe the largest shrews are found near the coast. Neotype. — None of Bachman’s type material is known to exist. Since Blarina carolinensis is geographically variable and additional subspecies are likely to be named, we designate as neotype USNM 574157, adult male with moderately worn teeth, skin and skull, collected 27 July 1989, by Charles Handley and Merrill Vam, beside Awendaw Creek, 3.2 km E Awendaw Post Office, Charleston County, South Carolina, in a thicket at the edge of a salt marsh. Field number COH 15236. Description. — This species can be recognized as a Blarina by its medium size (for a soricid), stout body, short tail (about as long as the head), blackish coloration, slit-like ear hidden in fur, tiny eye, and 32 teeth tipped with reddish pigment. Dorsal coloration in fresh pelage blackish with a slight brownish tint flecked with gray where hair bases show through; underparts paler, washed with brown, not sharply defined from dorsum, pallor not invading flanks; summer pelage slightly paler; vibrissae whitish, extending beyond the eye but not as far as the ear when laid back; forefeet buffy; hindfeet pale fuscous; tail blackish above, slightly paler beneath. Pelage becoming progressively browner in museum storage. Cranium moderately angular; tooth formula I 1/1, U 5/2, P 1/0, M 3/3 = 10/6 x 2 -- 32; U* and U“ large and subequal, U^ and U'* small and subequal, U^ tiny but usually visible in lateral view; toothrows relatively uncrowded; maxillary process extends behind M“; ascending ramus of mandible in lateral view bends up well behind posterior molar; posterior edge of P‘*, M^ and M“ concave. Measurements.— The. neotype, in mm (additional measurements in Table 2): Total length 105, tail vertebrae 21, hind foot 12, condylobasal length 19.4, palatal length 8.7, maxillary breadth 6.8, interorbital breadth 5.2, maxillary toothrow length 7.3, cranial breadth 10.7. Comparisons .—Shrews of coastal populations in South Carolina and North Carolina have a slightly longer rostrum than shrews of inland (Piedmont) populations in South Carolina and Virginia (Table 2). This variation is seen in means of condylobasal length (19.0 and 18.9 vs 18.6 and 18.6), palatal length (8.5 and 8.5 vs 8.1 and 8.2), and maxillary toothrow 1994 HANDLEY AND VARN-Shrews of Bachman 1837 397 length (7.3 and 7.3 vs 6.8 and 7.1). Other cranial and external measurements differ little, if any, between coastal and Piedmont populations. Nomenclature. — Sorex carolinensis DeKay (1842:21), based on specimens from New York, is not the same as S. carolinensis Bachman. It is a synonym of Blarina brevicauda Say. On the other hand, Sorex (Anotus) carolinensis Wagner (1855:550) is a composite, based partly on S. carolinensis Bachman and partly on S. carolinensis DeKay. However, S. carolinensis DeKay alone is the type species of Anotus, which Wagner used tentatively as a subgeneric name for a species he mistakenly believed had 36 teeth and no external ear. Talposorex Pomel (1848) was also based on S. carolinensis DeKay. Baird (1855:47) noted the anomalous description of S. carolinensis DeKay and did not include DeKay ’s name in the synonymy of S. carolinensis Bachman. Merriam (1895) indecisively referred to S. carolinensis Bachman as a species, Blarina carolinensis, or a subspecies, Blarina brevicauda carolinensis. Merriam thought it remarkable that S. carolinensis Bachman had remained free of synonyms. For 75 years following Merriam (1895), ail authors used the name combination Blarina brevicauda carolinensis. Handley (1971)> on geographic grounds, and Genoways and Choate (1972) and many subsequent authors on genetic grounds, have regarded Blarina carolinensis as a distinct species. Specimens examined. — All in the National Museum of Natural History unless otherwise noted: North Carolina: Currituck County: Knotts Island, 5 skins and skulls (Norfolk [Virginia] Museum). South Carolina: Charleston County: Awendaw Creek, 3.2 km E Awendaw Post Office, 1 skin and skull; Iron Swamp, 4.0 km SW Awendaw Post Office, 4 alcoholics and skulls, 9 alcoholics; 3.4 km NE McClellanville, 2 skins and skulls; Head of Mill Branch Swamp, 3.3 km NW McClellanville, 3 alcoholics and skulls, 14 alcoholics. Darlington County: Society Hill, 1 skull. Dorchester County: 12.9 km SE St. George, 1 skin and skull. Georgetown County: Georgetown, 1 skin and skull; Plantersville, 2 skins and skulls. Richland County: Columbia, 4 skins and skulls, 1 skin. Williamsburg County: Lanes, 1 skin and skull. Virginia: Amelia County: Amelia, 25 skins and skulls. Sorex cinereus Bachman Sorex cinereus Bachman, 1837:373 (preoccupied by Sorex arcticus cinereus Kerr, 1792). Corsira (Blarina) cinerea: Gray, 1838:124. Blarina brevicaudata: Lesson, 1842:89 = 1 Sorex cinereus Bachman. M[usaraneus] (Cryptotis) cinereus: Pomel, 1848:249. S(orex) carolinensis: Bachman, in Audubon and Bachman, 1854:344 (part.) = Sorex cinereus Bachman. Blarina cinerea: Baird, 1857:48. Blarina brevicauda: Allen, 1869:221 (part.). [Blarina] (Soriciscus) [cinereus]: Coues, 1877:649. Blarina (Cryptotis) parva: Merriam, 1895:17. Cryptotis parva: Miller, 1912:24. Blarina (Cryptotis) fhridana Merriam, 1895:19. Cryptotis parva floridana: Harper, 1927:270. Type locality. — “Goose Creek about twenty-two miles from Charleston...” (Bachman, 1837:374). This must be road mileage because the town of Goose Creek, in Berkeley County, South Carolina, is 16 mi NNW of Charleston, measured from the Battery at the tip of the Charleston Peninsula (the point in Charleston most distant from Goose Creek). “They [the type specimens] were ploughed up from time to time from an old field which had laid in an uncultivated state for some years and was partially overgrown with weeds and bushes” (Bachman, 1837). Syntypes. — Bachman (1837) based his description of Sorex cinereus on six specimens from Goose Creek that had been given to him by Mr. W. Wesner. Bachman wrote that he had “...received about twenty other specimens from various parts of the low country of Carolina — all of the size and colour of the above.” At least three of Bachman’s specimens of Sorex cinereus still exist. Two are in the Academy of Natural Sciences of Philadelphia, ANSP 477 and ANSP 478. They are labeled, “Bachman, Blarina cinerea, N.A.” They are badly discolored skins with skulls inside. The third specimen is USNM 94 [skin]/1771 [skull] in the National Museum of Natural History, received as an exchange from the Academy of Natural Sciences, and cataloged on 26 April 1852. It is labeled “Dr. Bachman, Carolina, Sorex cinereus.” All that remains of this specimen in the National Museum is a fragmentary rostrum of the skull (1771). The skin was transferred to the University of Michigan 1 April 1859 and was cataloged there. However, Handley could not find it at the University of Michigan on 31 May 1954. Presumably these three specimens were among the six syntypes from Goose Creek rather than among the 20 other specimens mentioned by Bachman. Probably there is no way to be certain. In any case, we do not select one of them as lectotype. Description.— Sorex cinereus Bachman is a small, short- tailed shrew with a slender tail about one-third of the head and body length (one-fourth of total length), delicate forefeet and hindfeet, slit-like ear wholly concealed in fur, acute snout, and tiny eyes. Coloration clear blackish gray above and grayish white below, fairly sharply divided on flanks; dorsal hairs tricolored, with wide blue-gray basal band, narrow whitish subterminal band, and short blackish tip imparting a “salt and pepper” appearance to dorsum; hairs of underparts bicolored, with wide blue-gray basal band and short whitish tip; vibrissae white, short, reaching back about halfway between eye and ear; forefeet and hindfeet whitish to grayish; tail bicolored, whitish below, blackish above. Skull narrow and slender; braincase rounded, not angular; auditory ring relatively large, much wider than basisphenoid between rings. Tooth formula I 1/1, U 4/2, P 1/0, M 3/3 = 9/6 X 2 = 30 teeth; maxillary teeth relatively small and posterior concavities on these teeth relatively shallow; U'^ usually visible in lateral view; distribution and sometimes intensity of chestnut pigmentation on teeth usually reduced. Measurements. —See Table 3. Comparisons .—To evaluate geographic variation in Cryptotis parva floridana, we selected four samples from the National Museum of Natural History to study in detail: southern Florida (Brevard County, representing typical C. p. floridana), northern 398 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 Florida (St. Marks area, Franklin and Wakulla counties), coastal South Carolina (Charleston County, representing typical S. cinereus), and coastal North Carolina (Dare, Hyde, and Cartaret counties). For comparison we used a series from Raleigh, North Carolina, to represent the local C. p. parva (typical C. p. parva from Blair, Nebraska, averages slightly larger). The sample from southern Florida averages slightly, but not significantly, larger than the series from northern Florida in cranial dimensions (Table 3). External measurements of the southern series are significantly larger in total length with little overlap, but the specimens from northern Florida have a relatively longer tail (26% of total length vs 24%). There is broad overlap between the northern Florida and South Carolina samples, with the South Carolina specimens averaging slightly smaller. Between South Carolina and coastal North Carolina samples there is little difference in skull measurements (the same or slightly larger in South Carolina), but coastal North Carolina specimens average a little larger in external dimensions. The tail, however, averages slightly longer in South Carolina specimens (25% of total length vs 24%). In summary, with few exceptions there is a gradual change from larger in the South to smaller in the North in coastal Cryptotis. In the North there is an abrupt change to much smaller size in the interior at Raleigh, with little overlap in most measurements. This is shown dramatically in total length (declines only 10.2% between southern Florida and coastal North Carolina, but 12.3% between coastal North Carolina and Raleigh), and condylobasal length (declines 7.8% between southern Florida and coastal North Carolina, and 7.0% between the coast and Raleigh). Tail length, a characteristic that distinguishes C. p. floridana from C. p. parva, averages about 25% of total length along the coast from Florida to North Carolina. It averages 21% at Raleigh (also 21% at Blair, Nebraska). Clinal diminution in size of Cryptotis along the coast from Florida to North Carolina is illustrated very well in the ratio of condylobasal length to maxillary breadth (Fig. 1, 2). The location of each score in the plot represents the position of each specimen along the axis defined by the discriminant function for that population. Both plots indicate that specimens from coastal South Carolina are similar to both C. p. floridana from Florida and specimens from coastal North Carolina previously reported as C. p. parva. All of these populations are well differentiated from C. p. parva of inland North Carolina. In fresh pelage C. p. floridana is clear gray without a hint of brown, whereas C. p. parva is clear brown without a hint of gray. Summer and winter pelages vary in shade from pale to dark, but not in hue. However, in some instances old worn pelage of C. p. floridana may take on a brownish tinge in the field. Postmortem changes in coloration can be dramatic in Cryptotis. Take for example a specimen (USNM 314998) that Handley noted in about 1960 to be “...in fresh, clean, winter pelage... almost clear gray dorsally, with hardly a hint of brown... indistinguishable from Florida C. floridana, but... strikingly distinct from topotypes of C. parva in similar pelage.” In 1990 this doesn’t look like the same specimen. Now it is brown dorsally, with hardly a hint of gray. It is similar to topotypes of C. parva. In 30 years this specimen changed from resembling one subspecies, C. p. floridana, to resembling another, C. p. parval Fortunately, not all postmortem changes in the museum are so dramatic. For example, consider the series from St. Marks National Wildlife Refuge, Florida, and that from McClellanville, South Carolina, used in this study. Sample size is the same in each. The skins are well-made, collected at exactly the same season (January-February), 12 years apart (1978 and 1990). Each series varies similarly from pale to dark, and when arranged by shade the two series are almost indistinguishable dorsally. Eleven months after they were collected the new specimens were still gray or gray with the slightest hint of brown; the old specimens were still gray but now slightly brownish dorsally. On the other hand, ventrally, the 1990 McClellanville specimens are easily recognized. Their whitish underparts contrast with the yellowish of the St. Marks specimens. We suspect coloration of the McClellanville and St. Marks series would be identical if the collections had been made in the same year. We interpret the yellow-brown cast in the St. Marks series to be postmortem foxing. Such changes may be related to peculiarities in specimen preparation as well as conditions of storage. Foxing in Cryptotis was noted also by Choate (1970:204). This phenomenon must be taken into account in any color comparisons involving museum specimens of Cryptotis. The dorsal coloration of specimens from coastal North Carolina, collected in 1939, is brown, not gray. These specimens are dark, rich brown; in some cases blackish-brown. They are darker than Raleigh specimens and differ from them also in the “salt and pepper” effect that characterizes C. p. floridana. In addition to size of body and coloration of pelage, Merriam (1895) used several dental characters to distinguish C. floridana from C. parva. Measurements show that some of Merriam’s characters (Table 4) are useful, some not. Shallowness of the posterior concavity on and M* and reduction in crowding of unicuspids (U“* visible in lateral view) group specimens from Florida and coastal South Carolina together and distinguish them from Raleigh, North Carolina, specimens. Specimens from coastal North Carolina are intermediate. The amount of color on the teeth and the depth of the anterolingual notch on F* are variable and not diagnostic. Tooth color is often, but not always, paler in coastal and southern specimens. Nomenclature. — During much of the 19th century Sorex cinereus Bachman had a curious history of cyclically hiding in the synonymy of Blarina brevicauda and being elevated to the position of type species of new subgenera. Lesson (1842:89) included Bachman’s Sorex cinereus as a probable synonym of Blarina brevicauda. Similarly, Allen (1869) regarded all of the American short-tailed shrews as subspecies or synonyms of Blarina brevicauda. He thought the smallest short-tailed shrews, such as Cryptotis cinerea and C. parva, were merely young of Blarina brevicauda. 1994 HANDLEY AND YARN— Shrews of Bachman 1837 399 On the other hand, Pomel (1848:249), contrary to Lesson (1842), withdrew Sorex cinereus Bachman (= Sorex parvus Say) from synonymy and set it apart as the type species of a new subgenus, Cryptotis. Three decades later, Coues (1877) refuted the opinions of Allen (1869), restored the distinction between large and small short-tailed shrews, and described a new subgenus, Soriciscus, once again using "'Sorex parvus Say or S. cinereus Bachm.” as the type species. Oddly, Coues placed Cryptotes [sic] Pomel in the synonymy of Blarina brevicauda. Baird (1857) had a third opinion. He did not recognize subgenera in Blarina, but he did recognize several species within the 30-toothed section of Blarina (= Cryptotis). Although he noticed that the Florida shrew was generally larger and longer-tailed, Baird was the first author to perceive a relationship between Cryptotis of Florida and coastal South Carolina. He referred a specimen (USNM 2155/3110) from Indian River, Florida, (near the type locality of the later named Blarina floridana) to Blarina cinerea. Furthermore, Baird ( 1 857 : 62) distinguished between Cryptotis floridana ( = Blarina cinerea) and C. parva (= Blarina exilipes). Baird’s specimens of C. floridana were from Florida and coastal Georgia and South Carolina (also one from Pennsylvania on the basis of tail length). His specimens of C. parva were from Mississippi, Missouri, Tennessee, Illinois, and Virginia. Thus, the geographic ranges of Baird’s two forms were very similar to the ranges we ascribe to C. p. floridana and C. p. parva in this paper. Like Baird, Merriam (1895) recognized 32-tooth (subgenus Blarina) and 30-tooth (subgenus Cryptotis) divisions of Blarina. Also like Baird, Merriam recognized two forms of Cryptotis in the southeastern United States: the widespread Cryptotis parva and a new species, C. floridana, in Florida. Merriam (1895:18) noted that specimens from the coastal region of South Carolina and Georgia (the same specimens that Baird had referred to Sorex cinereus Bachman) are “somewhat larger than the typical form [C. /Jflrva]... appreciably [larger] than those from Raleigh, N.C.” However, in spite of their larger size and gray rather than brown coloration, to Merriam’s eye these specimens showed no approach to C. floridana in dental characteristics. Thus, he synonymyzed Sorex cinereus Bachman with Cryptotis parva Say, and he has been followed by all authors to the present day. As we have shown in our comparisons, Merriam’s (1895) synonymy is incorrect. Actually, Sorex cinereus Bachman is a prior name for Blarina floridana Merriam. However, since Sorex cinereus Bachman is a primary homonym of Sorex cinereus Kerr, it is unavailable and falls into the synonymy of Blarina floridana Merriam. Thus, Merriam’s B. floridana is the first available name for the homonym Sorex cinereus Bachman. Its holotype is USNM 16510/23937, Chester Shoals, Florida. Specimens examined.— AW in the National Museum of Natural History unless otherwise noted. Cryptotis parva floridana: Florida: Brevard County: Chester Shoals, 1 1 mi N Cape Canaveral , 2 alcoholics and skulls (including holotype of Blarina floridana)-, Micco, 1 alcoholic and skull; Oak Lodge (East Peninsula, opposite Micco), 13 skins and skulls (Museum of Comparative Zoology, Harvard), 1 skin (Museum of Comparative Zoology, Harvard), 1 skull (Museum of Comparative Zoology, Harvard). Franklin County: Alligator Point, 8 mi SSE Panacea, 2 skins and skulls. Wakulla County: St. Marks National Wildlife Refuge, Panacea Unit, 2 skins and skulls, 1 skin, 3 skulls, 1 alcoholic; St. Marks National Wildlife Refuge, St. Marks Unit, 8 skins and skulls, 2 skulls, 12 alcoholics. North Carolina: Carteret County: 6 mi NE Beaufort, 2 skins and skulls; Bogue Island, near Morehead City, 2 skins and skulls. Dare County: Buxton, 1 skin and skull (North Carolina State Museum); Hatteras, 1 skin (American Museum of Natural History); 8-10 mi SW Stumpy Point, 5 skins and skulls, 1 skull. Hyde County: 10 mi N Englehard, 2 skins and skulls; 3 mi W Lake Landing, 7 skins and skulls, 1 alcoholic. South Carolina: Berkeley County: “Carolina” [= Goose Creek?], 1 skull; “North America” [= Goose Creek?], 2 skins with skull inside (Academy of Natural Sciences, Philadelphia). Charleston County: McClellanville, 4 skulls, 2 alcoholics and skull, 2 alcoholics; 3.4 km NE McClellanville, 8 skins and skulls; Mt. Pleasant, I skin and skull (Museum of Comparative Zoology, Harvard), 1 skin and skull. Georgetown County: Georgetown, 1 skin and skull. Cryptotis parva parva: Nebraska: Washington County: Blair, 1 1 skins and skulls (topotypes of C. p. parva). North Carolina: Wake County: Raleigh, 5 skins and skulls (Museum of Comparative Zoology, Harvard), 22 skins and skulls. Acknowledgments We are grateful to many persons who helped us in a variety of ways. J. Jacobs, United States Fish and Wildlife Service, Annapolis, Maryland, made initial contacts in South Carolina for us. O. Stewart, Head Wildlife Biologist, South Carolina National Forests, advised us on collecting areas. J. Cely, South Carolina Wildlife and Marine Resources Department, provided collecting permits. G. Stapleton, District Ranger, made us feel welcome, gave us permission to collect in Francis Marion National Forest, and provided a place for us to prepare specimens. D. Carlson, Forest Wildlife Biologist, patiently guided us around the forest to locate suitable study sites, secured permission for access to private lands for us, and proved to be a real friend. Mayor R. Leland and Mrs. J. L. “Maggie” Yergin of McClellanville allowed us to trap in their fields. G. Kirkland, Jr., Shippensburg University, showed us the basic techniques of pitfall trapping. A. Wilson, T. and B. Handley, and M. Leo helped check the pitfalls. Dr. B. Wanamaker, Orangeburg, South Carolina, arranged for Roper Memorial Hospital in Charleston to supply us with formaldehyde when we ran short in the field. L. Gordon and J. Jacobs, National Museum of Natural History, expedited cleaning the skulls of the South Carolina shrews. F. C. Thompson, Systematic Entomology Laboratory, United States Department of Agriculture, advised us on questions of nomenclature. D. Handley assembled the data for the tables, helped with manuscript preparation, and was patient throughout. Funds from the Smithsonian Office of Product Development and Licensing, L. Stevenson, Director, made the many trips to South Carolina possible. 400 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 Literature Cited Allen, J. a. 1869. Catalogue of the mammals of Massachusetts: With a critical revision of the species. Bulletin of the Museum of Comparative Zoology, l(8):143-252. Audubon, J. J., and J. Bachman. 1854. The Quadrupeds of North America. Volume 3. V. G. Audubon, New York, viii + 349 pp. Bachman, J. 1837. Some remarks on the genus Sorex, with a monograph of the North American species. Journal of the Academy of Natural Sciences of Philadelphia, 7(2):362-402, pi. xxiii. Baird, S. F. 1857. Reports of explorations and surveys, to ascertain the most practicable and economical route for a railroad from the Mississippi River to the Pacific Ocean. 8(1, Mammals):xix-xlviii, 1-757. Bangs, O. 1902. Descriptions of two new, insular Blarinas from eastern Massachusetts. Proceedings of the New England Zoological Club, 3:75-78. Choate, J. R. 1970. Systematics and zoogeography of Middle American shrews of the genus Cryptotis. University of Kansas Publications, Museum of Natural History, 19(3):195-317. COUES, E. 1877. Precursory notes on American insectivorous mammals, with descriptions of new species. Bulletin of the United States Geological and Geographical Survey, 3(3):63 1-653. Davis, J. A., Jr. 1957. A new shrew (Sorex) from Florida. American Museum Novitates, 1844:1-9. DeKay, j. E. 1842. Zoology of New-York, or the New-York fauna; etc. Part I. Mammalia. State of New York, Albany, xvi + 146 pp., 33 pis. Fitzinger, L. j. 1868. Kritische Untersuchungen iiber die der natiirlichen Familie der Spitzmause (Sorices) angehorigen Arten. II. Abtheilung. Sitzungsberichte des Kaiserlichen Akademie der Wissenschaften., Mathematisch-naturwissenschaftlichen Classe, Wien, 57(1):425-514. G ENDWAYS, H. H., AND J. R. CHOATE. 1972. A multivariate analysis of systematic relationships among populations of the short-tailed shrews (genus Blarina) in Nebraska. Systematic Zoology, 21:106-116. Geoffroy SaiNT-Hilaire, I. 1827. Memoire sur quelques espies nouvelles ou peu connues du genre Musaraigne. Memoires du Museum d’Histoire Naturelle, 15:117-144. Gray, J. E. 1838. [Revision of the genus Sorex, Linn.]. Proceedings of the Zoological Society of London, part 5 (1837): 123-126. Hall, E. R. 1981. The Mammals of North America. Volume 1 . John Wiley and Sons, New York, xviii -f 600 + 90 pp. Handley, C. O., Jr. 1971. Appalachian mammalian geography— Recent Epoch. Pp. 263-303, in The Distributional History of the Biota of the Southern Appalachians, Part III: Vertebrates (P. C. Holt, R. A. Paterson, and J. P. Hubbard, eds.). Research Division Monograph 4, Virginia Polytechnic Institute and State University, Blacksburg, Virginia, 306 pp. Harper, F. 1927. The mammals of the Okefinokee Swamp region of Georgia. Proceedings of the Boston Society of Natural History, 38(7):191-396. Hollister, N. 1911. Remarks on the long-tailed shrews of the eastern United States, with description of a new species. Proceedings of the United States National Museum, 40(1825):377-381. Jackson, H. H. T. 1925. The Sorex arcticus and Sorex arcticus cinereus of Kerr. Journal of Mammalogy, 6:55-56. 1928. A taxonomic review of the American long-tailed shrews (Genera Sorex and Microsorex). North American Fauna, 51:i-vi + 1-238. Jones, J. K., Jr., D. C. Carter, H. H. Genoways, R. S. Hoffmann, D. W. Rice, and C. Jones. 1986. Revised checklist of North American mammals north of Mexico, 1986. Occasional Papers, The Museum, Texas Tech University, 107:1-22. JUNGE, J. A., AND R. S. Hoffmann. 1981. An annotated key to the long-tailed shrews (genus Sorex) of the United States and Canada, with notes on middle American Sorex. Occasional Papers of the Museum of Natural History, The University of Kansas, 94:1-48. Kerr, R. 1792. The animal kingdom, or zoological system, of the celebrated Sir Charles Linnaeus. Class I. Mammalia... being a translation of that part of the Systema Naturae, as lately published, with great improvements, by Professor Gmelin of Goettingen..., etc. Edinburgh. 644 pp. Koopman, K. F. 1976. Catalog of type specimens of Recent mammals in the Academy of Natural Sciences at Philadelphia. Proceedings of the Academy of Natural Sciences of Philadelphia, 128(1): 1-24. Lesson, R.-P. 1842. Nouveau Tableau du Regne Animal. Mammiferes. Arthus-Bertrand, Paris. 204 pp. Merriam, C. H. 1895. Revision of the shrews of the American genera Blarina and Notiosorex. North American Fauna, 10:5-34, 102-107 (pis. 1-3). Miller, G. S., Jr. 1895. The long-tailed shrews of the eastern United States. North American Fauna, 10:35-56, 108-113 (pis. 4-6). 1912. List of North American land mammals in the United States National Museum, 1911. United States National Museum Bulletin 79, i-xiv -f 1-455 pp. POMEL, A. 1848. Etudes sur les camassiers insectivores. Seconde partie. — Classification des insectivores. Archives des Sciences Physiques et Naturelles, Geneve, 9:244-251. Rode, P. 1943. Catalogue des Types de Mammiferes du Museum National d’Histoire Naturalle. III. Ordre des Insectivores. Bulletin du Museum National d’Histoire Naturelle, (2)14:307-314, 382-387. Rose, R. K., and T. Padgett. 1991. Sorex longirostris fisheri Merriam. Pp. 562-564, in Virginia’s Endangered Species (K. A. Terwilliger, coordinator). McDonald and Woodward Publishing Co., Blacksburg, Virginia. Tate, C. M., J. F. Pagels, and C. O. Handley, Jr. 1980. Distribution and systematic relationship of two kinds of short-tailed shrews (Soricidae: Blarina) in south-central Virginia. Proceedings of the Biological Society of Washington, 93(l):50-60. Wagner, J. A. 1855. Die Saugthiere, in Abbildungen nach der Natur mit Beschreibungen, von Dr. Johann Christian Daniel von Schreber. Leipzig. Supplementband, xxvi + 810 pp. Table 1.— Cranial and external measurements of Sorex longirostris from localities in the southeastern United States. Historic and recent samples of Sorex longirostris fisheri are from areas now included in the Great Dismal Swamp National Wildlife Refuge. Abbreviations are defined in the section on methods. 1994 HANDLEY AND VARN-Shrews of Bachman 1837 < u ^ o <=> a; +1 c!) 00 \Q +1 I q q q d d u. X m q O o +1 7 —I ^ 6 q q d cN +'3 ra ^ p d +'3 d O m o d (N 7 d \o o 7 o d rn +'3 < H H CQ O U n u q q d ^ m 00- o\ (N 00 in o rj ^ On S CM q CO CC •i: 00 O d 'C +1 fO I C' § q (N •b 04 00 o d +1 00 O -H d +1 fi. q ir! 'A 04 7 q d +1 04 7 ^ Tj- '9 so q 75 2 o q q 8 q 00 q o d fl d as 04 d cn 3 en I'. T}- +1 0- 21 o o +1 1 r-* q +1 q - +1 1 o 00 Ui d d d m m O m m fO c8 a cS u OO o 00 O 00 c ° § +' s c d +1 d 1 04 U q 3 O q O'" 04 O' % X +i 2 VO lU m iC oo u 04 -CJ u q O as I O' o d S +i 00 q 04 rn 2 So g s: q 'O ^ o +i J. 7 Tf 00 o d 7 so C4 +1 ^ O' o\ >n >0) 04 q d >oi .2 'S ’5) ha > .‘O C v_ I s: to +1 04 OO oo as I >05 d C P CO B O' 0 >05 U PQ o >05 ci o q 75 q q 2 e 75 Q q u d +1 d ^ q 43 d +1 d 1 q 22 cS d +i VO 1 r- d +1 7f ■'t d u d B vr! q vr! d •S2 d < VO *5^ d -c 'q >05 o q N, O o q CN 04 1 O' q d 7f CO O +1 i CN +i 1 04 m 7j* rt Ti- Tf O o\ q q d VT) d d +1 1 00 r-- +1 1 00 q VT5 q VO d d o CN q q d d d +1 1 v£) +1 1 q q m Tf d Os ^ q ^ § +1 7 O' ^ q >05 00 2 o as ^ Os o 00 q VO On q d o\ d ^ 05 ° d +1 ^ 04 d d NO ro m d d +1 7 7 d V05 W OO 04 +1 w W W W (O coi 2-> c/> COI 2r* c/J y oc C4 04 W 04 M 04 2« S s: S s: S K C c a CQ +'(S +|(S +|j2 +1^ ■K 'K 'K 401 Table 2. — Cranial and external measurements o/’Blarina carolinensis from coastal (Charleston and Currituck counties) and Piedmont (Richland atuJ Amelia counties) localities in the Carolinas and Virginia. Abbreviations are defined in the section on methods. 402 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 00 o +1 (i O iri r- o O o q 9 q >n q d d d d d d d ♦-H +1 1 00 -H 1 m +1 1 cn +1 1 o q o o Tt cs CN d 9 d CN CN CN CN —< d 1—1 m o 8 o o (N q 9 q d d d d d d +1 +1 CN CN cs 1 00 1 m + 1 cn +1 1 o q o q q Tf d q ^ 9 00 d 00* .-H CN On CN ’-H CN q vd +1 q rN o \o d +! c O 3 ^ O U 3 O t/3 ° ^ +1 3 9 § 2 u ca o q no’ 00 q o \o q 43 o 00 c 9 d d 5 d d 5 d vd c < d d +l 1 00 q 2 +1 1 u-j q +1 1 cn r- +1 1 9 cn d 3 u m d 1 00 d 3 NO «<} r* u CO o 03 1 o d 9 d •5 d d .d d q d sZ 2 q d d o 1 +1 1 q r- £ +1 1 On 7j- C +! 1 00 2 +1 1 00 2 q d 2 q rf •S Os 'd 'VL q ■d d a d ?> "Ct d •S •5 'C C (n 00 o Qq q 00 5 CN q >s q q q 03 o NO d d d d d < 2 +1 1 7 \o +! 1 Tj- -S 03 +1 ! CN m +1 1 m 9 NO r-- d 9 d 9 d NO NO’ NO NO CN m CN • t" CN 00 CN CN 00 j O oo’ d 00 d 00* d 00 < 0, +1 1 CN 00 +1 I (N +1 1 00 m +1 1 ri d d d d ON +1 o 00 +1 CM 1 q d d q d rn ;® a +1 d ! q NO 4) q 4) -a — ' !> o <=> d Z oo d O NO d 3 o 00 ^ od O U od 3 U t/2 a a o 05 cd 4) jL«i « s: CM q in a K q in to 4) q m s q in >T) ^ 9 q :§ d d (N -§ ° d d r- a o d 1 CN , o ,o d 1 m c o +1 1 00 C +1 o 1 r- +1 U 1 »r; 1 +' 1 q ^ +1 1 q CN CM in d - 9 d •C 9 d CS * d ! NO ?S LO CS s: ,2 qcj Q qs C! _ Cl, Cl os i: o q q r o c • r- ^ ITi C o q d ■•C o o d 'Tf d r- Cl, o d CN §,o d C^ +1 i 1 m ! +' 1 •S +1 1 q SS +1 r* 1 q CN 00 d b q d C3 SO m 2 LO rN i d u cn L) m ^ d Cs B Ci, d O CM S'!, o in 00 oo 00 q 1.9 d d CN d d 1 (N d r- d d CN bo +1 d 1 C^J +1 1 q +1 1 q +1 -fl q 1 CN d d cn d q d q d lO d d CM o o 00 q q q q q q d CN d d 1 d d 1 r- d d 1 CN d d 1 m +1 1 r- +1 1 q +1 1 +1 1 q +1 1 00 CN d q NO NO »r; d q d NO NO d d Tj- Os lO q q q q q d t-’ d C'" d NO d d d d +1 3 00 +1 1 +1 1 q in +! 1 m o +1 1 22 d q NO q lO q d oo d r-' no' NO d d w w w , w w CO o CO o CO y. CO -l g. u ■S o' 02 _ 3 ^ e i.* .... '2 S ■s ^ "" 3 ■?! ^ 3 Q n- 0) 3 S3 H Co 3 s: ^ VO fn VO -L 0 o o •S H U E 2 t c g 2 - o ^ 73 C o U •S — bn 2 J 00 ^ ^ ^ ^ ^ IT) (N VO in Tf On 00 VO m ^ ^ ^ ^ 0\ VO — ' ov m VO Tt in 00 06 00’ 00 00 r- OV 00 00 in OV 00 vr> 1-H in m m fn CN F— I 00 00 7^- in 7l* 0 cn T^- W' 'W'' ,— 1 7t r" (N 0 VO 7t 00 , figs. 9-11). Differences with C. maurisca can be seen in a comparison with the figures given in Heim de Balsac and Mein (1971). Crocidura kivuana has a broader maxillary region and very different teeth (Heim de Balsac, 1968). The same applies to C. congobelgica, which may belong to this species group but has a very wide maxillary region (Hollister, 1916). 1994 HUTTERER — Shrews of Ancient Egypt 409 There exists some fossil evidence which indicates a rather old age for this species group. Wesselman (1984) referred a mandibular fragment from upper Pliocene sediments (dated as 3.03 + 0.07 my B.P.) of the Omo valley, Ethiopia, to C. aff. dolichura. Two teeth from Plio-Pleistocene sediments east of Lake Turkana, Kenya, were also identified with this species by Black and Krishtalka (1986). Both localities are outside the present range of the species group. The limits of the C. dolichura species group are not clear. Dieterlen and Heim de Balsac (1979), Dippenaar (1980), and Hutterer (1981, 1982, 1986) discussed various aspects of this group and their conclusions are not fully congruent. One problem is the possible inclusion in this group of species such as C. littoralis, C. monax, C. oritis, C. ultima, C. lanosa, C. stenocephala, C. usambarae, C. manengubae, C. tansaniana, and C. telfordi. These are medium to large blackish shrews which have a more or less naked tail of medium length. The skull is generally less inflated, the braincase hexagonal rather than rounded, the maxillary region broad and the dentition often heavy. Although there are certainly relationships between this assemblage and the C. dolichura species group, I prefer to treat them as separate groups. Crocidura balsamifera resembles only one of these species, C. manengubae. They are simitar in size, but the proportions and shapes of the upper and lower teeth are different and C. manengubae has a hexagonal braincase. After evaluation of all morphological characters, I assign C. balsamifera to the C. dolichura group, in the narrow sense defined above. The extant species of this group all occur in the high forest zone of Africa (Fig. 4). They are not abundant in numbers in a population, but up to three species of this group may occur together at one locality (Brosset, 1988). Only one species, C. roosevelti, occurs not within, but along the outer edge of the forest zone (Hutterer, 1981). From the known habitat preferences and morphology of these related species it may be deduced that C. balsamifera was a “forest shrew” with similar ecological requirements. Changing Environments and Faunas The presence of an extinct forest species in the embalmed fauna of ancient Egypt raises questions about the landscape in Egypt 2400 years ago. Of the 68 shrews unwrapped by Heim de Balsac and Mein (1971), 27 were Crocidura religiosa, 26 C. Olivieri, 1 Suncus etruscus, 3 C. floweri, 3 C. balsamifera (including the lost specimen), and 2 C. fulvastra. The most abundant species, the small C. religiosa, is rarely found now in the upper Nile valley, as is C. floweri (Osborn and Helmy, 1980), suggesting that conditions for shrews are now less favorable. Both of these species have been identified in fossil material from middle Paleolithic lake deposits of southwestern Egypt (Kowalski et al., 1989). The fauna of Bir Tarfawi, which is about 135,000 yr B.P. in age, shows that there were lakes with dense vegetation including trees in southwestern Egypt at that time. Farther south at Oyo, northwestern Sudan, Ritchie et al. (1985) found pollen evidence for a humid phase supporting savanna and grassland between 8490 and 4920 yr B.P. Climatic changes in the Nile valley have been studied by Adamson et al. (1980) and Paulissen and Vermeersch (1987a, 1987Z?) Their results indicate alternating humid and arid periods since the Pleistocene, including severe flooding of the Nile in Egypt about 13,000-12,000 yr B.P. due to an overflow from Lake Victoria. The last major wet period in upper Egypt occurred during the early Holocene between 11,000 and 6,000 yr B.P. We may therefore hypothesize that during certain periods, swamps or gallery forests along the Nile were of sufficient size to support tropical shrews and allow them to extend their ranges. It is also possible that during heavy floods, small mammals were shifted northward on floating mats of vegetation and tree trunks. The ancestor of C. balsamifera could have reached Egypt that way. Although the aridity of the climate increased after 6,000 yr B.P. in upper Egypt, remnants of gallery forest could have remained until destroyed, either from natural causes or as a result of human exploitation. The present shrew fauna of Africa provides examples which indicate that some “ forest shrews” inhabit marshes or gallery forests adjacent to forest borders. For example, Heim de Balsac and Verschuren (1968) found C. littoralis to occur in considerable numbers in marshes in the Guinea savanna of the Garamba National Park, northeastern Zaire. This could possibly serve as a model for the habitat of C. balsamifera in ancient Egypt. Acknowledgments This study has been undertaken on the background of almost 15 years of working through museum collections in Europe, Africa, and the United States. I express my gratitude to all curators and staff members who provided facilities and necessary support. My travels to Lyon, Pennsylvania, and Chicago were funded by the Museum Alexander Koenig and the German Research Association DFG. I am particularly grateful to Professor P. Mein (Universite Lyon) and Mr. R. Mourer (Musee Guimet de Lyon) for their kind permission to study the irreplaceable specimens described herein. Dr. D. Yalden (Manchester) supplied me with interesting literature, and Professor E. Brunner-Traut (Tubingen) allowed me to reproduce figures from her work. I am also grateful to the two referees who improved this text. Literature Cited Adamson, D. A., F. Gasse, F. A. Street, and M. A. J. Williams. 1980. Late Quaternary history of the Nile. Nature, 288:50-55. BI.ACK, C. C., AND L. Krishtalka. 1986. Rodents, bats, and insectivores from the Plio-PIeistocene sediments to the east of Lake Turkana, Kenya. Contributions in Science, 372:1-15. Boessneck, J. 1988. Die Tierwelt des Alten Agypten. C. H. Beck, Miinchen, 197 pp. Brosset, A. 1988. Le peuplement de mammiferes insectivores des forets du nord-est du Gabon. Revue de Ecologie (Terre Vie), 43:23-46. Brosset, A., G. Dubost, and H. Heim de Balsac. 1965a. Une nouvelle esp^e de Crocidura du Gabon. Mammalia, 29:268-274. 1965fc. Mammiferes inedits recoltes au Gabon. Biologia Gabonica, 1:147-174. Brunner-Traut, E. 1965. Spitzmaus und Ichneumon als Tiere des Sonnengottes. Nachrichten der Akademie der Wissenschaften in 410 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 Gottingen I. Philologisch-Historische Klasse, 1965, 7:123-163. Corbet, G. B. 1978. The Mammals of the Palaearetic Region: A Taxonomie Review. British Museum (Natural History) and Cornell University Press, London and Ithaca, 314 pp. Dieterlen, F., and H. Heim de Balsac. 1979. Zur Okologie und Taxonomie der Spitzmause (Soricidae) des Kivu-Gebietes. Saugetierkundliche Mitteilungen, 27:241-287. Dippenaar, N. J. 1980. New species of Crocidura from Ethiopia and northern Tanzania (Mammalia: Soricidae). Annals of the Transvaal Museum, 32:125-154. Dollman, G. 1916. On the African shrews belonging to the genus Crocidura — VII. Annals and Magazine of Natural History, (8)17:188-209. Engesser, B. 1975. Revision der europaischen Heterosoricinae (Insectivora, Mammalia). Eclogae Geologicae Helvetiae, 68:649-671. Geoffroy Saint-Hilaire, I. 1827. Memoires sur quelques espies nouvelles ou peu connues du Genre Musaraigne. Memoires du Museum d’Histoire naturelle, Paris, 15:117-145, pi. 4. Goodman, S. M. 1988. The food habits of the little owl inhabiting Wadi el Natrun, Egypt. Sandgrouse, 10:102-104. Heim de Balsac, H. 1968. Considerations preliminaires sur le peuplement des montagnes africaines par les Soricidae. Biologia Gabonica, 4:299-323. Heim de Balsac, H., and J. J. Barloy. 1966. Revision des Crocidures du groupe flavescens-occidentalis-manni. Mammalia, 30:601-633. Heim de Balsac, H., and P. Mein. 1971. Les musaraignes momifi&s des hypogees des Thebes. Existence d’un metalophe chezles Crocidurinae (sensu Repenning). Mammalia, 35:220-244. Heim de Balsac, H., and J. Verschuren. 1968. Exploration du Parc National de la Garamba. Mission H. de Saeger, 54:1-50. Heller, E. 1910. New species of insectivores from British East Africa, Uganda, and the Sudan. Smithsonian Miscellaneous Collections, 56(15): 1-8. Hollister, N. 1916. Shrews collected by the Congo expedition of the American Museum. Bulletin of the American Museum of Natural History, 35:663-680. HuguENEY, M. 1976. Un stade primitif dans 1’evolution des Soricinae (Mammalia, Insectivora): Srinitium marteli nov. gen., nov. sp. de rOligocene moyen de Saint-Martin-de-Castillon (Vaucluse). Compte Rendu Hebdomadaire des Seances de I’Academie des Sciences, Paris, D, 282:981-984. Hutterer, R. 1981. Nachweis der Spitzmaus Crocidura roosevelti fiir Tanzania. Stuttgarter Beitrage zur Naturkunde, A, 342:1-9. 1982. Crocidura manengubae n. sp. (Mammalia: Soricidae), eine neue Spitzmaus aus Kamerun. Bonner Zoologische Beitrage, 32:241-248. 1984. Status of some African Crocidura described by Isidore Geoffroy Saint-Hilaire, Carl J. Sundevall and Theodor von Heuglin. Annales du Mus^ Royal de L’Afrique Centrale, Sciences Zoologiques, 237:207-217. 1986. Diagnosen neuer Spitzmause aus Tansania (Mammalia: Soricidae). Bonner Zoologische Beitrage, 37:23-33. Hutterer, R., and D. L. Harrison. 1988. A new look at the shrews (Soricidae) of Arabia. Bonner Zoologische Beitrage, 39:59-72. Hutterer, R., and P. D. Jenkins. 1980. A new species of Crocidura from Nigeria (Mammalia: Insectivora). Bulletin of the British Museum of Natural History (Zoology), 39:305-310. Hutterer, R., and M. Tranier. 1990. The immigration of the Asian house shrew Suncus murinus into Africa and Madagascar. Pp. 309-319, in Vertebrates in the Tropics (G. Peters and R. Hutterer, eds.). Museum Alexander Koenig, Bonn, 419 pp. Hutterer, R., and D. W. Yalden. 1990. Two new species of shrews from a relict forest in the Bale Mountains, Ethiopia. Pp. 63-72, in Vertebrates in the Tropics (G. Peters and R. Hutterer, eds.). Museum Alexander Koenig, Bonn, 419 pp. Jenkins, P. D. 1984. Description of a new species of Sylvisorex (Insectivora: Soricidae) from Tanzania. Bulletin of the British Museum of Natural History (Zoology), 47:65-76. Kessler, D. 1989. Die heiligen Tiere und der Konig. Teil I. Beitrage zu Organisation, Kult und Theologie der spatzeitlichen Tierfriedhofe. Agypten und Altes Testament, O. Harrassowitz, Wiesbaden, 16:1-303, 10 pis. Kowalski, K., W. Van Neer, Z. Bochenski, M. Mlynarski, B. Rzebk-Kowalska, Z. Szyndlar, a. Gautier, R. Schild, A. E. Close, and F. Wendorf. 1989. A last interglacial fauna from the eastern Sahara. Quaternary Research, 32:335-341. Lesson, R. P. 1827. Manuel de Mammalogie ou Histoire naturelle des Mammiferes. Roret, Paris. Lortet, L., and C. Gaillard. 1903. La faune momifiee de I’ancienne Egypte. I. Archives du Museum d’Histoire naturelle de Lyon, 8:1-205, pi. 1-5. LURKER, M. 1974. The Gods and Symbols of Ancient Egypt. Thames and Hudson, London, 142 pp. Maddalena, T., A.-M. Mehmeti, G. Bonner, and P. Vogel. 1987. The karyotype of Crocidura flavescens (Mammalia, Insectivora) in South Africa. Zeitschrift fiir Saugetierkunde, 52:129-132. Osborn, D. J., and I. Helmy. 1980. The contemporary land mammals of Egypt (including Sinai). Fieldiana Zoology, New Series, 5:xix, 1-579. Paulissen, E., and P. M. Vermeersch. 1987a. Earth, man and climate in the Egyptian Nile valley during the Pleistocene. Pp. 29-67, in Prehistory of Arid North Africa, Essays in Honor of Fred Wendorf (A. Close, ed.). Southern Methodist University, Dallas. 1987fc. Contributions to Pleistocene environments of the Egyptian Nile valley. P. 82, in Evolution Passee et Future des Deserts (N. Petit-Maire and C. Vanbesien, eds.). PICG 252, Rapports CNRS, Paris. Ritchie, J. C., C. H. Eyles, and C. V. Hayres. 1985. Sediment and pollen evidence for an early to mid-Holocene humid phase in the eastern Sahara. Nature, 314:352-355. Setzer, H. W. 1960. Two new mammals from Egypt. Journal of the Egyptian Public Health Association, 35:1-5. Wesselman, H. B. 1984. The Omo micromammals, systematics and paleo ecology of early man sites from Ethiopia. Contributions to Vertebrate Evolution, 7:x, 1-219. 1994 HUTTERER— Shrews of Ancient Egypt 411 Table 1. — Synopsis of the past and present shrew fauna of Egypt. “The name C. religiosa is used in accordance with Corbet (1978) who selected a neotype. The application of the name C. nana Dobson, 1890, (type locality: Dollo, Somaliland) for this species by Heim de Balsac and Mein (1971) and Osborn and Helmy (1980) is not followed, based on an examination of the holotype of nana and the neotype q/" religiosa and comparisons with material from Egypt, Ethiopia, and Somalia. C. religiosa does not seem to occur outside Egypt. ^Ihe use of C. olivieri instead of C. flavescens deltae Heim de Balsac and Barloy, 1966, follows Corbet (1978) and Maddalena et al. (1987). ‘'C. fulvastra has priority over C. sericea (Hutterer, 1984). ‘^A recent examination of the holotype of C. suaveolens matruhensis Setter, 1960, in the Field Museum of Natural History, Chicago, has shown that it bears all characters of C. whitakeri. The specimen constitutes the first and only record of this species from Egypt. "The species has not been collected in Egypt since 1832 (Hutterer and Tranier, 1990). Species Known from Egypt Conspecifics or Close Relatives in Ancient Present Africa Palearctic Orient Crocidura religiosa^ (I. Geoffroy Saint-Hilaire, 1827) x Crocidura floweri Dollman, 1915 x Crocidura olivieri^ (Lesson, 1827) x Crocidura fulvastra"" (Sundevall, 1843) x Crocidura balsamifera n. sp. x Crocidura whitakeri^ de Winton, 1897 — Suncus etruscus (Savi, 1822) x Suncus murinus" (Linnaeus, 1766) — X X X X X (X) X X X X X X X Table 2.— Cranial measurements of the type specimens of species assigned to the C. dolichura species group. Abbreviations as explained in Materials and Methods. Holotypes marked with an asterisk were measured by the author, other measurements were taken from original description. Sources: 1 , American Museum of Natural History, New York (Hollister, 1916); 2, National Museum of Natural History, Washington, D.C.; 3, Museum national d ’His Wire naturelle, Paris; 4, Zoologisches Museum der Humboldt- Uni versitdt, Berlin; 5, The Natural History Museum, London; 6, Naturhistorisches Museum Basel (holotype skull lost; topotype skull measured from the Stuttgart Museum, SMNS 13413); 7, Musee Guimet d ’Histoire naturelle, Lyon; 8, Centre de Paleontologie, Universite Claude Bernard, Lyon; 9, unlabeled skull from Heim de Balsac 's collection, now in the Museum national d ’Histoire naturelle, Paris (MNHNP 1981-1090). Species CIL C. polia 18.2 C. ludia 18.2 C. muricauda 18.4 C. crenata — C. dolichura 19.2 C. niobe 19.6 C. latona 19.8 C. maurisca 20.4 C. roosevelti — C. kivuana 20.8 C. balsamifera 22.1 C. balsamifera — C. grassei 23.1 UTR PAL GW HCC PGL 7.8 8.2 7.8 — 8.2 — — 8.1 7.8 8.1 4.9 4.8 8.1 7.8 — — — 8.1 7.4 8.2 4.8 5.7 8.2 7.8 9.1 5.3 6.1 8.7 — 8.9 — — 8.9 — 9.3 5.5 6.0 8.6 8.5 8.4 — 5.7 9.2 — 9.6 5.9 6.2 9.6 9.1 9.8 5.6 6.4 9.9 9.8 — — — 10.1 — 8.9 5.2 — MB lO Notes, Source 5.2 3.8 holotype 1 5.4 5.0 holotype 1 5.0 3.8 holotype* 2 5.0 paratype* 3 5.5 4.2 holotype* 4 6.1 4.5 holotype* 5 6.1 — holotype 1 5.9 4.1 holotype* 5 5.9 4.6 holotype* 2 6.6 4.8 topotype* 6 6.4 4.5 holotype* 7 6.4 4.6 paratype* 8 6.7 4.8 holotype?* 9 412 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 Fig. 1.— Examples of the role of shrews in ancient Egypt: Left, a small bronze dedicated to the winged god Homs; Right, transcription of hieroglyphics from the Kahun papyms (both figures reproduced from Bmnner-Traut, 1965, with permission). Fig. 2.— Skulls and mandibles of Crocidura balsamifera n. sp., holotype on left, paratype on right side of figure. Scale is 5 mm. 1994 HUTTERER — Shrews of Ancient Egypt 413 1 Fig. 3. — Occlusal view of M^-M^ of the 1 mm. holotype (upper figure) and paratype (below) of Crocidura balsamifera n. sp. Scale is Fig. 4. A comparison of the distribution of the Crocidura doUchura species group (11 species as listed in Table 2) and the new species from Egypt. Records are based on literature and author’s unpublished data. i. 1 ,’n ■'V I: K ..... -0'"i' - •■■/i':. ' vr’ J-!-' I' . •«i}. ■ .■.f, ' : ^ = >■• p^. ". II-.. M^i'AA. r ■ ;•? ■• ■••/••. -JiH' .J. ■- I ■ * -.iAr .W&s* ■ :;V ftiW- .J, V4^^» ' '_•_.' ':i‘ ’•' ■A.: .>1.' - . jjy r-'!’.;\ :■ ,- ...1' ,t;. t* : ... . ?*•> vi-^’ ‘'Si; ^ . 5lr' •' ■■ 4'V' ■r, . ' ? '■<'> 1 '■( ■ >r 5 PROGESTERONE (P4) AND ESTRADIOL (E2) SECRETION BY SUNCUS MURINUS OVARIES AND ADRENALS IN VITRO Stasia Stoklosowa', Janice Bahr^, and Gil Dryden^ 'institute of Zoology, Jagiellonian University, Krakow, Poland; ^Department of Animal Scienee, University of Illinois, Urbana, Illinois 61801; ^9100 South Mill Site Road, Ashland, Missouri 65010 Abstract Early studies of eaptive musk shrews {Suncus murinus) here and in Japan questioned the role of ovarian steroids in reproductive behavior and function and suggested adrenal involvement. Later studies documented “normal” plasma estradiol (E^) concentrations and populations of uterine E2 receptors but steroid production by explanted shrew ovaries and adrenals have not been reported. We determined relative in vitro secretions of P4 and E2 by these organs in both hCG-pretreated shrews and in organs supplemented with hCG in vitro. Paired whole ovaries from pretreated shrews released five times more P4 than control ovaries but less than twice that of ovaries in medium supplemented with hCG. Ovarian E2 output was variably elevated by hCG pretreatment. P4 secretion by sliced adrenals was depressed by hCG in vivo and in vitro whereas E2 output was unaffected by hCG. These results suggest that the ovary is an important source of P4 and E2 in this species. Perhaps the ovaries of this shrew produce other active steroids; alternatively, this shrew has poor E, utilization, or it uses P4 “unconventionally.” Introduction The hormonal functions of musk shrew (Suncus murinus) ovaries and adrenals are poorly understood. Especially enigmatic are concentrations of £3 in blood and the role of this hormone in reproduction. Hasler (1973) was unable to measure serum E^ by radioimmunoassay but more sensitive techniques indicated plasma £3 concentrations of approximately 0.03 ng/ml in intact adult shrews (Rissman and Crews, 1988). Uteri bind tritiated £3 (Dryden and Anderson, 1977; Keefer and Dryden, 1982) but the hormone fails to elicit the classic uterotropic response in this species (Dryden and Anderson, 1977; Rissman and Bronson, 1987). Females exhibit no behavioral estrus cyclicity (Dryden, 1969; Rissman et al., 1988). Some ovariectomized shrews copulate (Dryden and Anderson, 1977) but most do not (Rissman and Bronson, 1987). Early suggestions that the adrenal gland might supply steroids in support of sexual behavior (Dryden and Anderson, 1977) have been confirmed by adrenalectomy and ovariectomy experiments. Ovariectomy abolishes sexual behavior only slightly more than does adrenalectomy (Rissman and Bronson, 1987). Furthermore, studies by Furumura et al. (personal communication) strongly suggest adrenal and/or placental contributions of steroids toward implantation as well as pregnancy maintenance. Likewise, Rissman and Bronson (1987) suggested that adrenal-ovarian interaction drives sexual behavior and uterine development, perhaps separately or synergistically. They moreover raised the possibility that steroids other than £3 are important modulators of reproductive functions in this species (Rissman and Bronson, 1987). Ovaries of intact musk shrews respond to injected human chorionic gonadotropin (hCG) or luteinizing hormone (LH) by ovulating and producing morphologically normal corpora lutea (Dryden, 1985). It was therefore of interest to determine directly if secretion of sex steroids by ovaries and adrenals is under gonadotropin control. There are no published data on the secretion of £3 or P4 by Suncus ovaries and adrenals maintained in culture. This study was therefore undertaken to determine in vitro secretion of these hormones by preovulatory ovaries and by adrenals from animals treated with hCG. Materials and Methods Treatment Groups Sixteen multiparous shrews three to five months old were divided into three groups: 1) ovaries and adrenals of five shrews were controls; 2) five animals were injected with 50 lU of hCG (Sigma, Stock No. CG-2), and ten hours after injection (about five hours prior to expected time of ovulation), ovaries and adrenals were removed; 3) ovaries and adrenals were removed from six untreated animals and hCG (10 lU/culture well) was added to the culture medium in which the tissue was maintained. Culture of Ovaries and Adrenals All ovaries and adrenals were aseptically removed, cleaned of adjoining tissue under the stereomicroscope, and placed in culture wells, each containing 1 ml of culture medium. Culture was performed using multi-well plates (Coming) according to Fainstatt (1972). Tissue was distributed in the following way; both ovaries and one adrenal from each shrew were placed in culture plate wells but the other adrenal was saved for peri fusion. Since the diameter of shrew ovaries did not exceed 1 mm, they were cultured in toto. Adrenals were incised in the middle to facilitate penetration of culture medium in the perifusion apparatus. Culture medium was a MEM (GIBCO) supplemented with 10% calf serum (GIBCO) and with penicillin, streptomycin, and mystatin in doses routinely used in tissue culture. Cultures were kept in a humid incubator at 37°C for 24 h in 95 % O3 atmosphere. After 24 hr in culture, media were collected and frozen for steroid analysis, and ovaries were frozen for protein assay. Because only one adrenal from each animal was used in the tissue culture, the other adrenal from control and hCG-injected shrews was placed in a perifusion system, one per column. Tissue was perifused with a MEM for 6 hr (flow rate = 1 ml/90 min) at 37°C. Perifusion media were frozen for steroid analysis. 415 416 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 Radioimmunoassays were run using very specific antibodies obtained from Drs. O. D. Sherwood and G. D. Niswender. Intra-assay and interassay variations, both between 5 % and 8%, were used as genera! averages in this study. Steroids were measured according to the procedure previously reported by Bahr et al. (1980). Tissues were extracted with petroleum ether for progesterone (P^) assay. Samples were extracted first for E2 with anesthesia-grade ether. After evaporation of the ether, the residue was extracted again with a mixture of 1 ml hexane and 1 ml 75% methanol. The hexane phase was removed and the methanol evaporated. Samples were then assayed as previously described. Double extraction and antibody specificity obviated chromatographic separation for £3 measurements. Culture media and peri fusion effluents were assayed for £3 and P4 content according to the method of Bahr et al. (1980). Several glands were fixed in Bouin’s solution for histology. Peri fusion data were analyzed by ANOVA for repeated measures using SAS program 5. 16. Results Ovarian Morphology Histologic appearances of follicles and interstitia of ovaries in control and treated groups conformed to previous descriptions (Dry den, 1969). Ovaries of untreated shrews contained many atretic and unenlarged follicles whereas ovaries of shrews treated with hCG underwent preluteinization hypertrophy and vascular/lymphatic distention similar to that observed in mated shrews. One ovary exposed in vitro to hCG contained an old corpus luteum unquestionably from a previous pregnancy. Interstitial hypertrophy was most extensive in ovaries of shrews pretreated with hCG and less so in those from control (uninjected) animals. Ovarian Secretion Control ovaries secreted 0.28 + 0.07 ng P^I\>aXxl2A hr of culture. The amount of P4 released by ovaries from hCG- injected shrews was approximately five times greater than the amount of P4 produced by ovaries of control shrews (1.38 ± 0. 10 ng/pair/24 hr; P < 0.001; Fig. 1). Conversely, levels of P4 in medium containing ovaries treated with hCG in vitro were only 0.45 ± 0.05 ng. This concentration is statistically similar {P — 0.09) to that released from ovaries of shrews pretreated with hCG and with values for control ovaries. Ovarian E2 Secretion The estradiol concentrations secreted into the culture medium by ovaries of different shrews were highly variable. Mean control E2 secretion was 44.05 ± 20.98 pg/pair/24 hr. Ovaries from hCG-injected animals were steroidogenically more active, releasing 266.31 ± 128.12 pg/pair/24 hr. Although this output was five times greater than that of control ovaries, the difference was not significant {P = 0.09), presumably a function of group size and variability {n = 5, coefficient of variation = 107.28%, Fig. 1). Ovaries cultured in hCG- supplemented medium secreted 92.64 ± 19.99 pg E2/pair/24 hr, which was not significantly different from control concentrations. Adrenal Morphology Tissue arrangements and appearances were similar between groups. Adrenals from all treatment groups were similarly compact. Cortical zonation or cellular hypertrophy did not differ. The deep zona fasciculata of all adrenals appeared steroidogenically active. Adrenal P^ Secretion Each control adrenal released 57.19 ± 15.66 ng P4/24 hr, whereas the adrenals of hCG-pretreated shrews and adrenals of shrews treated with hCG in vitro produced 38.99 ± 7.73 ng/24 hr and 45.23 + 8.05 ng/24 hr, respectively. All values were statistically identical (Fig. 2). A similar but more pronounced “suppressing” effect of hCG on P4 secretion was observed in peri fused adrenals (control = 5.92 + 2.40 ng/gland/24 hr; hCG-injected shrews were statistically significant by 4.5 hr (P < 0.05) with increased significance at 6 hr (from 5.59 + 1.05 to 2.84 ± 0.62 ng/gland/24 hr; P < 0.01, Fig. 3). Adrenal E2 Secretion Control adrenals released 4.35 + 0.69 pg E2/gland/24 hr. There was no significant difference in Ej amounts secreted by adrenals obtained from hCG-injected shrews or by adrenals treated with hCG in vitro (Fig. 2). This lack of response by adrenals to hCG was also observed for adrenals in the peri fusion system. Discussion Previous studies of S uncus murinus have shown that animals injected with hCG ovulate the same number of oocytes (Dryden and Pucek, 1976) at the same time (Singh and Dominic, 1982) as mated shrews. Ovaries of hCG-injected shrews also form corpora lutea of pseudopregnancy which are anatomically similar to those following sterile mating (Dryden, 1985). The present results extend these parallel responses to the ability of cultured ovaries to respond to hCG administered systemically or in vitro. This study also reports ovarian and adrenal P4 and E2 secretory responses under these two conditions. Estradiol secretion by cultured, whole Suncus ovaries was extremely variable and low but could be enhanced five-fold by hCG-pretreatment of the donor shrew. Presumably more convincing enhancement of E2 production by ovaries following pretreatment with hCG could be obtained by varying the gonadotropin dosage, time of assay after treatment, and by employing larger numbers of animals. The weak response of shrew ovaries to hCG in vitro is puzzling. Cultured ovaries obtained from immature and mature mice secrete greater amounts of P4 and Ej under similar treatment (Neal and Baker, 1974; Ryle et al.,1975). In rats, the timing of hormone secretory response to ovarian receptor proliferation seems precisely timed (Szoltys, 1976, 1981; Szoltys et al, 1982; Madej , 1986). We have no such insight into the ovarian response of any shrew. Intervals of hormonal 1994 STOKLOSOWA ET AL — Suncus Steroid Secretion 417 sensitivity to stimulation may be transitory. Also, the shrew ovary in vitro may have a longer latent period than that of the mouse before becoming responsive to hCG. Clearly, more rigorous testing of S. murinus ovarian responses to varied hCG and treatment times needs to be done now that this species is firmly established in several research laboratories. The role of the adrenal in the regulation of reproduction in Suncus is gaining support. Plasma and adrenal P4 concentrations are significantly higher in receptive vs nonreceptive females (Hasler, 1973). Highest levels of adrenal P4 occur in females 15 hr after copulation (Hasler and Nalbandov, 1980). Some ovariectomized shrews sustain pregnancy, mammary development, and high plasma P4 concentrations (Furumura et al., 1983). Administered E2 fails to enhance sexual behavior of ovariectomized females but socially interacting, intact females do experience elevated plasma P4 levels (Rissman and Crews, 1988). These observations, along with the apparent dissociation of central and peripheral effects of steroids (Rissman and Bronson, 1987), strongly suggest adrenal involvement in the reproductive coordination of this species. This interpretation is supported by the ovari ec tomy / adrenalectomy manipulations of Furumura et al. (personal communication), who hypothesize the participation of both adrenals and ovaries in pregnancy maintenance. Ovariectomy followed by steroid replacement has also demonstrated that female scent-marking behavior of musk shrews, unlike that of the male musk shrew, is not gonadally controlled (Tennant et al., 1987). That study indicated the necessity of adrenal steroid support for some but not all of a female’s reproductive behavioral repertoire. The brains as well as uteri of female musk shrews concentrate radiolabelled testosterone and estradiol in similar compartments; moreover, their vaginal epithelia bind higher concentrations of testosterone than of estradiol (Keefer and Dry den, 1985). We cannot currently make much of hCG-induced “suppression” of P4 production by peri fused adrenals. They may not be responsive to hCG and physiologically deteriorate in culture, thus producing an illusory secretory depression over the time course monitored in this study. Could the adrenals of female Suncus secrete testosterone under gonadotropic stimulation, as do those of marsupials (Vinson and Renfree, 1975)? Rissman et al. (1990) have shown that testosterone administered to ovariectomized S. murinus restores sexual behavior, also raising the issue of a possible role of androgen aromatization in female musk shrews. The present results from unweighed cultured organs provide only hints about relative ovary and adrenal functions under endogenous gonadotropic influence in intact musk shrews. But, along with results of other studies discussed here, they certainly suggest the presence of intriguing endocrine mechanisms controlling reproduction which remain to be elucidated. Acknowledgments We thank A. Jablonka for expert technical assistance, K. Furumura, E. Gregoraszuzk, and K. Ota for sharing unpublished data, as well as J. Anderson, P. Johnson, E. Rissman, and A. van Tienhoven for helpful criticism of an early draft of this paper. Literature Cited Bahr, J., R. Gardner, P. Schenck, and N. Shabhi. 1980. Follicular steroidogenesis: Effect of reproductive condition. Biology of Reproduction, 22:817-826. Dryden, G. L. 1969. Reproduction in Suncus murinus. Journal of Reproduction and Fertility Supplement, 6:377-396. 1985. Development and regression of the corpus luteum of pregnancy in Suncus murinus. Acta Zoologica Fennica, 173:263-264. Dryden, G. L., and J. N. Anderson. 1977. Ovarian hormone: Lack of effect on reproductive structure of female Asian musk shrews. Science, 197:782-784. Dryden, G. L., and Z. Pucek. 1976. Insectivores in reproduction studies, with emphasis on ovulation in American, Asian and European shrews. Pp. 39-50, in The Laboratory Animal in the Study of Reproduction (Th. Antikatzides, S. Ericksen, and A. Spiegel, eds.), Gustav Fischer Verlag, Stuttgart, 185 pp. Fainstatt, T. H. 1972. Submerged organ culture: An improved method. In Vitro, 7:300-303. Furumura, K., K. Ota, A. Yokoyama, and S. Oda. 1983. Mammary growth and plasma progesterone level during pregnancy in the house musk shrew, Suncus murinus Linnaeus. Endocrinologica Japonica, 30:621-630. Hasler, M. J. 1973. Progesterone levels in non-receptive and receptive female musk shrews (Suncus murinus). Biology of Reproduction, 9:94. Hasler, M. J., and A. V. Nalbandov. 1980. Ovulation, ovum maturation, and changes in plasma and adrenal progesterone concentration in the musk shrew (Suncus murinus). Biology of Reproduction, 22:377-381. Keefer, D. A., and G. L. Dryden. 1982. Nuclear uptake of radioactivity by cells of pituitary, brain, uterus, and vagina of the Asian musk shrew (Suncus murinus) following (^H) estradiol administration. General and Comparative Endocrinology, 47:125-130. 1985. Oestradiol and testosterone concentration in female Suncus murinus. Acta Zoologica Fennica, 1973:265-267. Madej, E. 1986. Effect of exogenous hormones on estrogen and progesterone release from cultured rat granulosa cells from various stages of estrous cycle. Endocrinologica Experimentia, 20:293-400. Neal, P., and T. G. Baker. 1974. Response of mouse ovaries in vivo and in organ culture to pregnant mare’s serum gonadotropin and human chorionic gonadotropin. III: Effect of age. Journal of Reproduction and Fertility, 39:41 1-414. Rissman, E. F., and F. H. Bronson. 1987. Role of the ovary and adrenal gland in the sexual behavior of the musk shrew, Suncus murinus. Journal of Reproduction and Fertility, 36:664-668. Rissman, E. F., and D. Crews. 1988. Hormonal correlates of sexual behavior in the female musk shrew: The role of estradiol. Physiology and Behavior, 44:1-7. Rissman, E. F., J. Silveria, and F. H. Bronson. 1988. Patterns of sexual receptivity in the female musk shrew (Suncus murinus). Hormones and Behavior, 22: 186-193. Rissman, E. F., A. L. Clendenon, and R. W. Krohmer. 1990. Role of androgens in the regulation of sexual behavior in the female musk shrew. Neuroendocrinology, 51:468-473. Ryle, M., J. Court, T. Smith, and R. Morris. 1975. Gonadotropic stimulation of oestrogen synthesis by cultured immature mouse ovaries. Journal of Endocrinology, 66:225-232. Singh, J. S., and C. J. Dominic. 1982. Induction of ovulation in the musk shrew Suncus murinus L. by hCG. Indian Journal of 418 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 Experimental Biology, 20:10-12. SZOLTYS, M. 1976. Progestagen dynamics in preovulatory follicles of rats. Journal of Reproduction and Fertility, 48:397-398. 1981. Oestrogens and progesterone in rat ovarian follicles during the oestrous cycle. Journal of Reproduction and Fertility, 63:221-226. SzOLTYS, M., S. Stoklosowa, AND F. H. Kasten. 1982. Hormonal secretion of cultured rat ovarian follicles isolated at various hours of proestrus. In Vitro, 18:463-467. Tennant, L. E., E. Rissman, and F. H. Bronson. 1987. Scent marking in the musk shrew {Suncus murinus). Physiology and Behavior, 39:677-680. Vinson, G. P., and M. B. Renfree. 1975. Biosynthesis and secretion of testosterone by adrenal tissue from the North American opossum, Didelphis virginiana, and the effects of tropic hormone stimulation. General and Comparative Endocrinology, 27:214-222. Fig. 1. — Progesterone and estradiol secretion by cultured shrew ovaries (2 ovaries/24 hr). C, control cultures: in vivo, ovaries from shrews injected with 50 lU hCG 10 hr before removal from the shrew; in vitro, 10 lU hCG/ml culture medium; ***, P < 0.001. 1994 STOKLOSOWA ET AL — Suncus Steroid Secretion 419 Fig. 2. — Progesterone (P4) and estradiol (E2) secretion by cultured shrew adrenals (1 adrenal/24 hr). Treatment groups as defined in Fig. 1. No significant differences among groups. < LU GC Q < LU U) + 1 O LU Z o cc LU H C/) LU O O QC Q. 1-5 3 4.5 6 HOURS Fig. 3.— Progesterone secretion by perifused shrew adrenals; *, P < 0.05; **, P < 0.01. See text for experimental conditions. PG ± SEM/ADRENAL/DAY tu>sftp»y - 4i'i’ ■ <"< h. I; ^lif ‘’’^' ti:D' 1 : .1 -J. %: * '• ■■•- ^ ' ' ' ■‘’' ' I . -’’■Sulill’ :^•f;;>. 'ATut -j...i S;i *Hi‘ r;y I '-«M^ -saawif : .« ' 7^^, ■S;m06lL ";«5# .aiT. <'1^14?, Ii"^ -''fV V . '’ 1.^1;-'^. ; .’ , '''.! V>i 'I'i ■ >tr vt, jW iV,' ,ri I I ■>,. f'/lll*''',; |- ii'i V'-a' '■ ' 't'iw I fFvit'l .. : iS't... J ; ' s!i ■«’'• • '5'- . O. S 1 rnftt ! <• t 7™ 1 in. S' W +’ D ’5? :« '* I l*^‘ * ‘■ r-’ ■ j ^ 1:1 M *^v.; © -3; Ui' t'A- r ■•4v i I ^ .•?U’f? c,.r • TXi ■ i. Hi J, :i - ' tu ■ O * .0 *.'ft *>v,'.’; I . -«■* ' ; f ' »f "i-' T J*.. ^ « ‘^■V*'", ' .. ;: V,'. f METABOLIC RATES AND REGULATION OF CARDIAC AND RESPIRATORY FUNCTION IN EUROPEAN SHREWS Alfred Nagel Hans-Thoma-Strafie 5, 61440 Oberursel/Taunus, Germany Abstract As measures for metabolic rate, the eardiac and respiratory activity, oxygen consumption, heart rate, and respiratory rate were measured at different ambient temperatures in the white-toothed shrew Crocidura russula and the red-toothed shrews Sorex araneus and Sorex minutus to investigate systematie differences and thermoregulatory adaptations to speeific habitats. The thermal neutral zone (TNZ) in sueh small mammals is very narrow and appears around 30°C in C. russula, but around 20°C and 25®C in S. araneus and S. minutus, respectively. Basal oxygen consumption in C. russula is 12% above the expected value predieted by the Kleiber curve and increases by 420% at 0°C. Basal respir- atory rate increases 3.8 times, and basal heart rate to 1.7 times basal value at Q°C. Oxygen eonsumption per breath at 0°C is only 15% above the value in TNZ, whereas basal oxygen pulse inereases to 255% at 0°C. Minimal oxygen eonsumption in S. araneus and S. minutus is 267% and 322% higher than predieted and increases only to 1.7 times basal value at 0°C. In both species, basal respiratory rate amounts to 435 min“’. Consumed oxygen per breath amounts to 2.7 /il and 1.5 pi with an increase at 0°C of only 30% and 40% in S. araneus and S. minutus, respectively. Basal heart rates are 528 min“* (S. araneus) and 786 min“’ (S. minutus)', the values at 0°C are 20% higher. Oxygen pulse, which depends on heart mass, is 2.3 pi and 0.85 pi with an inerease of up to 120% and 150% at 0°C in S. araneus and S. minutus, respectively. The two main possibilities to deliver higher amounts of oxygen in the organism are to inerease frequencies of respiration and heart aetivity, or to increase oxygen intake per breath and oxygen transport per heart beat. While in both Sorex species, the contribution of frequency regulation of respiration equals that of the volume regulation (50% each), C. russula shows a pure frequency regulation (95%), which is an exception among mammals. Regarding heart activity, frequency regulation accounts for only 33% and 25 % of increased cardiac output in C. russula and S. minutus, respectively, whereas the volume and frequency regulation Introduction All small homeothermic animals show high mass-specific metabolic rates because of a disadvantageous relationship between body mass and body surface (Kleiber, 1967). In red- toothed shrews (Soricinae) especially, these metabolic rates are even higher than in other small mammals of the same body size (Pearson, 1947; Morrison, 1948; Tomasi, 1984; Nagel, 1985). In white-toothed shrews (Crocidurinae), mean body temperature is comparatively lower than in other mammals of the same size, and their metabolic rates are almost as low as in “standard” mammals (Vogel, 1976; Nagel, 1985; Mover et al., 1986). Little is known about adaptations concerning respiration or regulation of cardiac and respiratory function in shrews, information which would be required to support the elevated metabolic rates. Therefore oxygen consumption, respiratory rate, heart rate, and body temperature were investigated in the common shrew (Sorex araneus), the lesser shrew (S. minutus), and the common European white-toothed shrew Crocidura russula in order to investigate the physiological aspects of thermoregulation in small endotherms in general and to investigate systematic differences combined with geographic distribution and thermoregulatory adaptations to specific habitats. Materials and Methods Seven specimens of Crocidura russula (x = 11.4 g), 11 Sorex araneus (x = 8.0 g), and six Sorex minutus (x = 3.4 g) were kept outdoors under semi natural conditions in large cages and fed alternately with commercial dog food and meal worms. Oxygen consumption (Vq2) was measured in an open system (Servomex, type OA 184). A measurement chamber, described portions are almost equal (50%) in S. araneus. in Nagel (1986), was used to determine Vq2, heart rate (HR), and respiratory rate (RR) in undisturbed shrews. The air flow through the measurement chamber was 15 L X h“’ (STPD). Oxygen consumption measurements lasted from one hour, at extreme ambient temperatures (Ta), to 24 hours at Ta to which the shrews were normally accustomed. The animals fasted for at least two hours before and during the experiments. The measurements were taken in a temperature-controlled cabinet regulated to +1°C. Body temperature (Tb) was measured rectally at the end of each experiment by a calibrated electronic thermometer (Ahlbom MeB- und Regeltechnik, Therm 2245); the thermistor was inserted about 18 mm into the rectum in C. russula and 10 mm in S. araneus and S. minutus. Respiratory rate was measured by means of the whole body plethysmography (Drorbaugh and Perm, 1955; Malan, 1973; Epstein and Epstein, 1978) with a micropressure transducer (Furness Controls Limited; type PC 011, measuring range ±2 millibar). To get sufficiently high signals, the air flow had to be adjusted to a pressure difference of — 0. 1 millibar between the inside and outside of the measurement chamber. To diminish the influences of the membrane pump, a thin capillary was inserted between measurement chamber and membrane pump. The electronic signals were registered either with a recorder (Goerz Metrawatt, type SE 120, Hewlett Packard, type 7702 B) or were analyzed with a storage oscilloscope (Tectronics, type 654). Heart rate was registered by foot electrodes (Nagel, 1986). Signals were recorded and analyzed parallel to measurements of RR with the same equipment. Oxygen consumption, RR, and HR were analyzed in normothermic shrews only, during the last 15-30 min of each experiment. Data are presented as the mean +SD. Statistical comparisons were made using a 2-sample Student’s t-test (Bortz, 1979). 421 422 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 Differences were considered significant for P < 0.05. The theorem of Bienayme-Tschebychev (Schaich, 1977) was used to compare basal metabolic rate with the mass-predicted value (Kleiber, 1967). The regression analysis was based on the least squares method. Results Body Temperature The first parameter to determine the accuracy of temperature regulation is to measure Tb at different Ta. The range of Ta in which Tb is rather constant is the animal’s state of normothermia. At lower Ta, Tb drops and the animals become hypothermic. At the upper end of normothermic Ta, Tb increases and the animals become hyperthermic. Tb at different Ta is shown in Fig. 1. The highest mean Tb at all Ta, shown by S. minutus, ranged from 37.6 ± 0.9°C to 37.9 ± 0.3°C at Ta of 20°C and below, but increased to 39.4 ± 0.3°C and 39.2 ± 0.2°C at 25 and 30°C, respectively. In S. araneus, Tb was always slightly lower than in S. minutus (Ta < 20°C, mean Tb - 37.2 + 0.6-37.7 + 0.9°C) but was also relatively constant. At Ta above 20°C in S. minutus, and 25 °C in 5. araneus, Tb increased to 39.2 + 0.2°C and 38.3 ± 0.9°C, respectively. At higher Ta, the shrews were not able to relax and died within half an hour due to heat stress. The response of Tb to changing Ta was remarkably different in C. russula. In the range of normothermia (Ta < 30°C), this species showed a wide variation of Tb with low mean values between 35.7 + LUC (Ta = 0°C) and 34.2 + 0.8°C (Ta = 25 °C). At Ta above 30°C, the animals lost normothermia and their Tb increased up to 37.9 + 1.1 °C (Ta = 37°C). The experiments had to be interrupted above 37 °C, because C. russula did not relax and no values during rest could be measured. In no single case could Tb be regulated below Ta. Nevertheless, C. russula tolerated higher Ta than S. araneus and S. minutus. Oxygen Consumption The energetic expenditure of temperature regulation in normothermic animals is expressed by oxygen consumption (Fig. 2) at different Ta compared to the basal values. Because of the small size of S. minutus, its Vq2 was very high. The range of Ta in which Vq2 was lowest is denoted as the thermal neutral zone (TNZ) which appeared around 20 °C in S. araneus and 25 °C in S. minutus. The thermal neutral zone could not be determined more precisely because all measurements were made in 5°C Ta intervals. Ambient temperatures below and above TNZ at which Vq2 is increased is the lower critical temperature (Tic) or the upper critical temperature (Tuc), respectively. The basal Vq2 in S. minutus, 12.0 ± 2.2 ml O2 X g“’ X h“' was 3.2 times higher than predicted (Kleiber, 1967); (log M — 1.87 + 0.739 log W ± 0.03, M = basal metabolism [kcal X day“*l, W = body mass [kg], 1 ml O2 = 4.8 cal). Oxygen consumption increased at Ta below 25 °C. This increase can be described by the calculated regression equation: Vq2 [ml X g~’ X h-’j - -0.288 X Ta + 19.6, n = 42, r = -0.635. The slope of this regression line is handled as the minimal thermal conductance. It was —0.288 ml O2 X g“* X h~* X °C“*. The actual value lies 48% below the mass-predicted value (Herreid and Kessel, 1967). Minimal Vq2 in S. araneus was 8.3 ± 1.6 ml O2 X g~‘ X h“* at 20°C. As in S. minutus, this value is 267% higher than the mass-predicted value. Oxygen consumption increased with decreasing Ta following the equation: Vqj [ml x g“* x h“*] = —0.302 X Ta -t- 14.7, n = 128, r = —0.677. Minimal thermal conductance in S. araneus, —0.302 ml X g“* X h“* X °C~*, was 15% below the predicted value. At 0°C the temperature-regulating response of Vq2 was 1.7 times the basal value in both Sorex species. In normothermic C. russula, mean Vq2 at different Ta was lowest at 30°C. The mean value of basal Vq2, 2.32 + 0.42 ml X g“' X h“*, was only 12% above the expected value and did not differ significantly from it. In spite of the relatively low Tb in the TNZ (Tb = 35.4 ± 0.7°C, Fig. 1), Vq2 was still higher than in other mammals of the same size. Below Ta of 30°C, Vq2 increased steadily to 9.76 ± 1.55 ml X g“^ X h“’ at the Ta of 0°C, which is 4.2 times higher than in TNZ. The temperature-dependent increase of described by the linear equation: Vq2 [ml X g“* x h“‘] = —0.253 X Ta + 9.49, n = 108, r = —0.893. Thermal conductance (0.298 ml X g“’ X h“* X °C“’ was about 15% lower than the expected value. In normothermic and torpid C. russula, Vq2 strongly depends on the level of regulated Tb which is shown in Fig. 3. At a Tb of 18°C (Ta = 15°C), Vq2 is even lower than 1 ml X g“’ X h“*. At higher Tb Vq2 increases steadily with increasing Tb to a value of 7.5 ml x g“* X h”* at a Tb of 37 °C. At high Tb, Vq2 in C. russula is almost as high as in S. araneus with the same Tb. Hence, the main difference in Vq2 between the species was the different level of regulated Tb. The remaining differences in Vq2 can be explained by the smaller body mass of S. araneus. Respiratory Rate Mean respiratory rate (RR) in relation to Ta was similar in S. minutus and S. araneus (Fig. 4). Basal RR was also almost identical (437 ± 86 min“* in S. minutus and 433 ±81 min~* in S. araneus) and were 86% and 130% above the mass- predicted value (Stahl, 1967). At 0°C, RR increased due to the increased Vq2 to 598 ± 113 min”* and to 557 ± 103 min”* which were 1.4 and 1.3 times the basal values, in S. araneus and S. minutus, respectively. Respiratory rate remained nearly constant at Ta above the upper critical Ta. The mean basal value of RR in C. russula was much lower (40% below the mass-predicted value) than for both Sorex species and amounted to only 103 ± 19.8 min”* within the TNZ. By Tb of 0°C, RR had increased to 393 ± 80.1 min”*, which was a 3.8-fold increase over the basal value. This increase corresponds approximately to the 4.2-fold increase in Vq2- In contrast to S. araneus and S. minutus, RR in C. russula increased at Ta above the upper critical temperature. Highest respiratory rates could be found during sniffing. For short periods values of 800 min~* could be reached. 1994 NAGEL— Metabolic Rates in European Shrews 423 Oxygen Consumed per Breath The calculated amount of oxygen per breath, treated as the amount of oxygen absorption per breath, is the main parameter of the depth of breathing. Because of the different size of the shrews, which limits respiratory volume, the calculated oxygen consumed per breath was different among the species (Fig. 5). The basal value was 1.5 ± 0. 1 ^rl in S. minutus, and 2.7 ± 0.5 p\ in S. araneus. At 0°C, oxygen consumed per breath was elevated to 2.0 + 0.5 pi and 3.4 ± 0.7 p\, which is a 1.3-fold increase in basal value for both species. Due to the constant RR at Ta above the upper critical Ta (Tuc), alterations in oxygen consumed per breath were negligible. In C. russula, volumes of oxygen consumed per breath increased from a mean of 4.3 + 0.7 pi (Ta = 30°C) to 4.9 ± 0.5 pi (Ta = 0°C) which was only 15% above the value in TNZ. Above Tuc the volume of oxygen consumed per breath decreased slightly to a mean of 4. 1 + 1.1 (Ta = 35°C) and 3.8 ± 1.0 pi (Ta = 37°C). The two main possibilities to satisfy higher oxygen demands in the body are: 1) to increase frequency of respiration, or (2) to increase oxygen intake per breath. In both Sorex species, frequency regulation accounted for 52 % and volume regulation for 48% of the increase in respiratory performance calculated according to Bartholomew and Tucker (1963). In contrast, C. russula showed a pure frequency regulation (95%). Heart Rate High metabolic rates in shrews should be combined with high heart rate (HR) to guarantee the supply of oxygen to the body. In fact, HR was lower than expected (Fig. 6). Basal HR in S. minutus was lowest at Ta of 15°C whereas the lower critical Ta (Tic) was near 25 °C; it amounted to 774 ± 22 min”* and was only 29% above the expected value (600 min”*, Wang and Hudson, 1971). Heart rate was raised to 922 ± 105 min”* at 0°C. In S. araneus, minimal heart rate (528 + 55 min”*) occurred in the TNZ and was only 9% above the expected value. It increased to only 638 ±119 min”* at 0°C. In both Sorex species, HR increased 1.2 times the basal value at Ta of 0°C. In C. russula, mean basal HR was 444 ± 37 min”* at Ta of 30 °C. This value corresponds exactly to the expected value. Lowering Ta to 0°C led to an increase of HR to 779 ± 108 min”*. The relative increase of HR in C. russula was 75%. Corresponding to Vq2> heart rate of C. russula also depends on the Tb (Fig. 7) of the animal. At a Ta of 15°C, the lowest measured heart rate during torpor (Tb = 18°C) was 81 min”*. At higher Tb, heart rate increased to about 700 min”* at a Tb of 36°C. Oxygen Pulse The calculated Vq2 per heart beat, the oxygen pulse, is given in Fig. 8. In S. minutus, basal oxygen pulse was 0.85 ± 0. 1 /xl. It increased to 1.3 ± 0.3 /xl at 0°C, which is 50% higher than the basal value. In S. araneus, basal oxygen pulse amounts to 2.3 ± 0.5 pi. The increase at 0°C was to 2.8 ± 0.5 pi, or 1.2 times the basal value. In C. russula, the basal value was 0.98 ± 0.2 pi', oxygen pulse increased with increasing Vq2 (Ta = 0°C) about 255% to a mean value of 2.5 ± 0.4 ^1. Above Tuc oxygen pulse increased to 1.3 ± 0.2 pi (Ta = 37°C) in C. russula, yet it remained almost constant in S. minutus and S. araneus. In contrast to Vq2 and HR, oxygen pulse does not depend on Tb during torpor or normothermia in C. russula (Fig. 9); most of the values lie between 1-2 pi at Tb of 18-37°C. In all cases, oxygen pulse of S. araneus was higher than the corresponding values of C. russula although heart mass and HR were lower and Vq2 higher than in the white-toothed shrew. This effect can be explained by a higher mass-specific oxygen- transporting capacity of the heart of S. araneus (Table 1), which generally was higher in the red-toothed shrews. Only at Ta of 0°C did C. russula have a higher value than S. minutus, which is three times smaller and has only half the heart mass. The frequency regulation portion accounted for 33% and 25% of heart activity in C. russula and S. minutus, respectively, using the formula of Bartholomew and Tucker (1963). In iS. araneus, the volume and frequency regulation portions were almost identical (51% and 49%, respectively). Discussion Body Temperature The present investigation has shown that C. russula had a lower and less constant mean Tb than S. araneus and S. minutus (Fig. 1). Low Tb, also known in Crocidura leucodon, Crocidura suaveolens, Suncus etruscus (Nagel, 1977, 1985; Frey, 1979; Mover et al., 1986), and Suncus murinus (Balakrishnan et al., 1974; Hasler and Nalbandov, 1974), thus seems typical for white-toothed shrews in general. In contrast, S. minutus and S. araneus had high and constant Tb, also confirmed by Gebczynski (1977) and Sparti and Genoud (1989), in Neomys fodiens (Nagel, 1985), in Sorex coronatus (Sparti and Genoud, 1989), in Blarina brevicauda (Doremus, 1965; Platt, 1974; Merritt, 1986), in Cryptotis parva (Layne and Redmond, 1959), and in Sorex cinereus (Morrison et al., 1959). Above Tuc no real regulation of Tb took place; Tb fluctuated only with Ta. Comparing the species, the minimal difference between Ta and Tb depends on basal metabolic rate and is a sign for heat tolerance. The smaller this difference, the greater is the heat tolerance of the animal. Thus red-toothed shrews (Soricinae) show symptoms of severe heat stress even at a Ta of 30°C, a Ta at which C. russula is in thermoneutrality. The Soricinae are, like other small mammals with high Tb, typically adapted to a cold environment, a result also confirmed by Irving (1972) but not by Scholander et al. (1950), Precht et al. (1955), and Cossin and Bowler (1987). Thus, the regulation of high Tb obviously is no adaptation to high Ta, but the tolerance of a high Tb without regulatory response on it is favorable to endure high Ta. There is no doubt that white-toothed shrews (Crocidurinae) have a more flexible temperature regulation, particularly due to their ability to enter torpor either spontaneously (Frey and Vogel, 1979; Genoud, 1985) or during periods of starvation (Wahlstrom, 1929; Kusnetzov, 1972; Vogel, 1974; Nagel, 424 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 1977, 1985; Frey, 1979). During torpor, Tb can drop passively to 18°C (Nagel, 1985). This low Tb in torpor always was higher than Ta (range 26-18‘’C), but does not depend otherwise on Ta; the shrews do not react as poikilotherms do. There are no obvious explanations about the level of TT? in torpor, but shrews regulate the certain in torpor very well by alterations of Vq2- The arousal from torpor, which takes place spontaneously or after a gentle disturbance of the animal, is followed by an increase in Tb at the rate of 0.5-0.9°C X min“* (Nagel, 1985). Even during severe food deprivation (Martinsen, 1969; Gebczynski, 1971a), red-toothed shrews are not capable of torpor. Lindstedt (1980a), however, observed shallow hypothermia in the red-toothed shrew Notiosorex crawfordi, a desert-dwelling shrew from the southwestern USA. White-toothed shrews develop torpor by the fourth day of life (Nagel, 1989), when the young are still blind and naked. In young shrews torpor is only expressed if the animals are able to form groups of at least two animals. In field studies Vogel et al. (1984) observed groups of nesting C. russula only during winter, when food supply was reduced. Perhaps these animals formed “torpor groups” as described in the marsupial Sminthopis crassicaudata and the house mouse Mus musculus (Morton, 1978). Oxygen Consumption Both Sorex species in this study had high basal metabolic rates which were approximately 100% above the expected value (Kleiber, 1967). In white-toothed shrews, however, the relationship of metabolic rate to body mass corresponds approximately to the Kleiber curve. The exceedingly high basal metabolic rates in several species of Soricinae have been reported by many investigators (Morrison and Pearson, 1946; Pearson, 1947, 1948; Morrison, 1948; Morrison et al., 1952, 1953, 1959; Hawkins et al., 1960; Pfeiffer and Gass, 1962; Buckner, 1964; Doremus, 1965; Gebczynska and Gebzynski, 1965; Gebczynski, 1965, 1971a, \91\b\ Martinsen, 1969; Neal and Lustick, 1974; Platt, 1974; Vogel, 1976; Lindstedt, 1980(>; Tomasi, 1984). In contrast, the basal metabolic rate of C. russula was only a few percent higher than the expected Kleiber value, which is typical for all European Crocidurinae (Fons and Sicart, 1976; Vogel, 1976; Nagel, 1977, 1985; Frey, 1979) as well as for Suncus murinus (Dryden et al., 1974), Crocidura occidentalis (Hildwein, 1972) and Crocidura russula monacha (Mover et al., 1986). Although mean Tb was comparatively low in white-toothed shrews, basal metabolic rate was higher than the predicted value in all cases. The influence of regulated Tb on metabolic rate (Fig. 3) is expressed by a very strong correlation between Vq2 and Tb in C. russula. The values are compared with the corresponding values in S. araneus at a Ta below Tic of both species (Ta -- 15°C). The main conclusion is: An individual of C. russula at a given high Tb has the same Vq2 as an individual of S. araneus with the same Tb. Consequently, below Tic in both species, differences in mean Tb are the principle explanation for the differences between both species. The role of lowered Tb and of actually lowered metabolic rate on the differences between C. russula and S. araneus is explained by Fig. 10, which shows that the role of different Tb is greater than that of actual different basal metabolic rates. Comparing the actual values of the Crocidurinae and Soricinae (Fig. 11) with those expected from the Kleiber curve, basal metabolic rates of the Soricinae seem to lose the dependence on body mass, whereas the Crocidurinae have reconfirmed Kleiber’s correlation of basal metabolic rate and body mass by the loss of a constant and high Tb. At low ambient temperatures (Ta = 0°C), Vq2 in C. russula increased to a value 4.2 times the value at TNZ (Fig. 2). Because a similar increase in Vq2 can be found in all other European Crocidurinae (Nagel, 1985), these white-toothed shrews must be regarded as adapted to high Ta. In contrast, Vq2 in S. minutus and S. araneus increased only to 1.5 times the basal value. This relatively small response is a result of the high basal metabolic rate which has been interpreted as an adaptation to cold climatic conditions (Scholander et al., 1950; Irving, 1972; Weathers, 1979). The slope of increasing Vq2 at decreasing Ta (thermal conductance; Fig. 2) shows the supplementary Vq2^ needed to maintain constant Tb at Ta below Tic by an increase in heat production. This temperature-regulatory reaction of metabolism shows the total effect of Ta on Vq2 of a homeothermic animal, neglecting the relationship to several parameters on which it depends. Thermal conductance, which depends on insulation, heat production, heat loss, and body temperature, shows a very similar pattern in all three species, varying between 0.288 ml O2 X g“' X h“* X °C~* {S. minutus), 0.302 ml O2 X g“* X h~' X °C~* (5. araneus), and 0.298 ml O2 X g~* X h“* X (C. russula). These values are about 15% in 5. araneus and C. russula and in S. minutus about 48 % below their mass- predicted values (Herreid and Kessel, 1967). There is no real difference in the reaction to cold between S. araneus and C. russula in spite of different levels of TT? and of different basal metabolic rates. In contrast, in other shrew species, thermal conductance depends indirectly on body mass (Nagel, 1977, 1985). The Tb correction of thermal conductance (Bradley and Deavers, 1980; McNab, 1980) leads to a more differentiated result of 0.21 ml O2 X g“’ X h“* Xx in C. russula and 0. 17 ml O2 X g~* X h“* X in S. araneus, respectively. The lower thermal conductance in S. araneus can be explained by a better insulation, due to longer hair and a thicker fur. Nevertheless, the real effect of low Ta was the same in the two almost equal-sized species S. araneus and C. russula. Respiratory Rate Corresponding to their high metabolic rates, mean basal respiratory rate (RR) in S. minutus and S. araneus are also very high, at 86% and 130% respectively above the mass-predicted value (Stahl, 1967). In contrast, mean basal RR is 40% below the predicted value in C. russula. The two main possibilities to satisfy higher oxygen demands in the body are to increase frequency of respiration or to increase oxygen intake per breath. At low Ta in both Soricinae, RR and consumed oxygen per breath increased in the same way and the regulation of oxygen absorption was achieved not only by alteration of RR but also by changing tidal volume or oxygen extraction rate, or both. In both Sorex species, frequency regulation accounted for 52% and 1994 NAGEL — Metabolic Rates in European Shrews 425 volume regulation for 44% of increase in respiratory performance (Bartholomew and Tucker, 1963). This form of regulation is typical for most mammals and birds (Casey et al., 1979; Withers et al., 1979; Muller and Rost, 1983; Blake and Banchero, 1985; Miiller, 1985; Prinzinger, 1988). In C. russula, Vq2 is frequency regulated. Because of the small changes in volume of oxygen consumed per breath and the almost identical increase of Vq2 (4.2-fold) and RR (3.8-fold) in C. russula, it can be assumed that respiratory depth and oxygen extraction rate are nearly constant. This statement must be preliminary, because no suitable methods are available to do these measurements in such small mammals. A similar situation was also found in the North American white-footed mice (Withers, 1977; Chappel, 1985), in hummingbirds (Prinzinger and Schuchmarm, 1985; Prinzinger and Jackel, 1986) and even in the bat Leptonycteris sanborni (Carpenter and Graham, 1967), and seems to be typical for small animals with high metabolic rates and high RR. At Ta above Tuc, high Tb leads to a decrease of consumed oxygen per breath. It is assumed that respiration becomes more and more shallow, possibly due to increasing rates of evaporative heat loss. Real panting with very low volumes of consumed oxygen per breath was not observed, either in C. russula or in the Sorex species, although it is developed in birds of similar size (Prinzinger and Schuchmarm, 1985; Prinzinger and Jackel, 1986). Heat dissipation by evaporation of water would probably be insufficient in this species, because normally no free water is available in the habitat of C. russula. Heart Rate Basal heart rate (HR) in both Sorex species is only 10-30% higher than predicted, whereas it corresponds exactly to the mass specific value in C. russula (Wang and Hudson, 1971), (HR = 816 X W“®'^^, W = body mass [g]), (Fig. 12). Compared with the formula reported by Stahl and Gummerson (1967), (HR = 241 x W“°'^^, W = body mass [kg]), basal HR is 26-62% lower. All data of basal HR in nonshrew mammals, for example of Mus minutoides (Goethel and Nagel, 1990) studied in my laboratory are much lower than the predicted value calculated after Stahl and Gummerson (1967) but fit quite well to the formula of Wang and Hudson (1971). In my opinion, the formula of Stahl and Gummerson (1967) is not suitable for mammalian allometry, especially at small body mass, and therefore is not considered further here. Dryden et al. (1971) reported a HR of 563-604 min~’ and Balakrishnan et al. (1974) a mean HR of 550 min“’ in the relatively large Suncus murinus. Heart rate is about 750 min~' in the North American species Blarina brevicauda (Doremus, 1965), and it ranges from 600 min“* in the masked shrew Sorex cinereus (Morrison et al., 1959). Weibel et al. (1971) and Bartels et al. (1979) investigated the smallest known mammal, Suncus etruscus, and found HR of 1000 min~* and 700-1400 min“' respectively. A comparison of the data from these investigations with the mass-expected and the reported values leads to the conclusion that the reported results are much too high and artificially distorted due to inappropriate methods. Measurement of HR by foot electrodes (Nagel, 1986) is an efficient and. therefore, favorable method for measuring heart frequency during rest, because the animals are not disturbed by handling and the results are not influenced by narcosis or other factors. The relatively low HR compared both to metabolic rates and to mass-predicted values in shrews can be explained by a particularly large heart (Stahl, 1965), which involves greater stroke volumes combined with a greater oxygen transport per heart beat. In European shrews, relative heart mass is 70-110% higher than in other small mammals (Bartels et al., 1979; Nagel, 1980, 1985), and must be regarded as a general adaptation to their high metabolic rates. Due to this adaptation, heart rate values that were 25-33% below the expected value (Nagel, 1980, 1985, 1991) are capable of delivering the oxygen to sustain their high metabolic rates. Another adaptation to high metabolic rates is the reduction in body mass, called Dehnel’s phenomenon (Buchalczyk, 1961; Mezhzerin, 1964; Pucek, 1965, 1970; Hyvarinen, 1969). Only heart mass remains constant and leads to a higher relative heart mass, probably to avoid too high HR during the strong metabolic efforts during winter. S. araneus reduces its body size during winter, but Dehnel’s phenomenon could not be found in any white-toothed shrew. The relative increase of HR in S. minutus and S. araneus was only 20% whereas that of C. russula was 75%. Compared with Vq2> hr increased only moderately in all three species; the increased oxygen demands in tissues at lower Ta can be satisfied mainly by an increase in the transported oxygen amount per heart beat or an increase of the oxygen utilization in the blood. The amount of transported oxygen per heart beat carmot be estimated, but the effectively used amount of oxygen per heart beat can be calculated as Vq2 per heart beat, i.e., as the oxygen pulse. In comparison with the pure frequency-dependent regulation of ventilation in C. russula for regulation of heart function, frequency regulation as well as volume regulation were pronounced in all three species. In heart function regulation, the regulatory role of oxygen pulse was two times higher than heart frequency in S. minutus and C. russula (Bartholomew and Tucker, 1963). Only in S. araneus were frequency regulation and volume regulation equivalent. Heart rate of the Crocidurinae during torpor depends very strongly on the Tb of the animal (Fig. 7). The lowest measured value was 60 min“* in Crocidura leucodon (Tb = 18°C, Nagel, 1980). In spite of strong decreases in heart rate and Vq2, oxygen pulse remains constant at a certain ambient temperature (Fig. 9; Nagel, 1980), which points out that torpor is a well-regulated state, because heart excitability depends on Tb (Nagel, 1986) and probably also on the contraction of the heart, but not Vq2 per heart beat. Oxygen pulse itself can be influenced only by alterations in ambient temperature. In spite of this strong increase of oxygen pulse, in no single case did mean values reach those in S. araneus, a species which is considerably smaller. This is astonishing because C. russula had a higher HR than S. araneus (Ta = 0°C) in spite of its greater body mass, lower Tb, lower metabolic rate, and lower basal HR. The mass-specific effect of oxygen transport per heart beat was considerably higher in S. araneus than for the 426 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 other two species (Table 2). No explanation for this fact can be given because no differences in the blood parameters are known between the species (Bartels et al., 1969; Wolk, 1974). Perhaps differences in arteriovenous oxygen levels, in the structure of the heart or heart activity regulation must be taken into account. Conclusions Comparing the distribution of the Soricinae and the Crocidurinae in the Old World, the red-toothed shrews are found from the temperate zone, to the subarctic and arctic regions. In contrast, the Crocidurinae are found in the tropical and subtropical regions, with few species in the cooler temperate zones. Their comparable low Vq2 and low Tb, combined with a great tolerance to high Ta, point to their tropical origin (Table 2). As a group, the Crocidurinae have a good tolerance against heat, their TNZ is high (approximately 30°C) and their average basal metabolic rate is much lower than in the Soricinae. The stable climatic conditions of the warm regions, especially in the tropics, favor this economical energy budget. Several small and middle-sized subtropical and tropical mammals, such as rodents, insectivores, and primates, show the same adaptations in temperature regulation (Hildwein, 1972; Muller, 1975, 1979; Muller and Kulzer, 1977). The Crocidurinae have adopted another strategy, entering torpor, which also serves to save energy, especially in response to lack of food. However, torpor occurs at Ta between 10-25°C and cannot be equated with hibernation. The northern boundary of their distribution shows that, in spite of the ability to enter torpor, the Crocidurinae are not adapted to the climatic conditions of subarctic and arctic regions. Red-toothed shrews have adapted a completely different strategy. Their well-regulated TT?, low TNZ and high basal metabolic rates enable them to endure the lowest temperature conditions, comparable to lemmings (Hart and Heroux, 1955), snow voles (Bienkowski and Marszalek, 1974), and weasels (Iversen, 1972; Casey and Casey, 1979). Acknowledgments I thank A. Pressmar for animal care and assistance during the experiments. I further thank for statistical advice my friend P. Jackel and for extensive discussion my colleague H. Burda. This study was supported in part by DFG Na 178/1-1. Literature Cited Balakrishnan, N., G. N. Ambdcatmajan Nair, and K. M. Alexander. 1974. A study of some aspects of the physiology of the Indian musk shrew Suncus murinus viridescens. Journal of Animal Morphology and Physiology, 21:98-106. Bartels, H., R. Schmelzle, and S. Ulrich. 1969. Comparative studies of the respiratory function of mammalian blood. V. Insectivora: Shrew, mole and nonhibemating and hibernating hedgehog. Respiration Physiology, 7:278-286. Bartels, H., R. Bartels, R. Baumann, R. Fons, R. Jurgens, and K. D. Wright. 1979. Blood oxygen transport and organ weights of two shrew species. American Journal of Physiology. 236:221-224. Bartholomew, G. A., and V. A. Tucker. 1963. Control of changes in body temperature, metabolism, and circulation by the agamid lizard, Amphiholurus barbatus. Physiological Zoology, 36:199-218. Bienkowski, P., and U. Marszalek. 1974. Metabolism and energy budget in the snow vole Microtus nivalis. Acta Theriologica, 19:55-67. Blake, C. I., and N. Banchero. 1985. Effects of cold and hypoxia on ventilation and oxygen consumption in awake guinea pigs. Respiration Physiology, 61:357-368. Bradley, S. R., and D. R. Deavers. 1980. A re-examination of the relationship between thermal conductance and body weight in mammals. Comparative Biochemistry and Physiology, 65A:465-476. Bortz, J. 1979. Lehrbuch der Statistik fiir Sozialwissenschaftler. Springer-Verlag, Berlin, 871 pp. Buchalczyk, A. 1961. Variation in the weight of the internal organs of Sorex araneus. Acta Theriologica, 5:229-252. Buckner, C. H. 1964. Metabolism, food capacity and feeding behaviour in four species of shrews. Canadian Journal of Zoology, 42:259-279. Carpenter, R. E., and J. B. Graham. 1967. Physiological responses to temperature in the long-nosed bat, Leptonycteris sanborni. Comparative Biochemistry and Physiology, 22:709-722. Casey, T. M., and K. K. Casey. 1979. Thermoregulation of arctic weasels. Physiological Zoology, 52:153-164. Casey, T. M., P. C. Withers, and K. K. Casey. 1979. Metabolic and respiratory responses of arctic mammals to ambient temperature during the summer. Comparative Biochemistry and Physiology, 64A:33 1-341. Chappell, N. A. 1985. Effects of ambient temperature and altitude on ventilation and gas exchange in deer mice (Peromyscus inaniculatus). Journal of Comparative Physiology B, 155:751-758. CossiN, R., AND K. Bowler. 1987. Temperature Biology of Animals. Chapman and Hill, London, 339 pp. Doremus, H. 1965. Heart rate, temperature and respiration rate of the short-tailed shrew in captivity. Journal of Mammalogy, 46:424-425. DroRBAUGH, j. E., AND W. 0. FENN. 1955. A barometric method for measuring ventilation in newborn infants. Pediatrics, 16:81-87. Dryden, G. L., R. Gossrau, AND H. E. Dale. 1971. The electrocardiogram and impulse conduction system of the Asian musk shrew. Pfliiers Archiv, 323:105-120. Dryden, G. L., M. Gebczynski, and E. L. Douglas. 1974. Oxygen consumption by nursling and adult musk shrews. Acta Theriologica, 19:453-461. Epstein, M. A. F., and R. A. Epstein. 1978. A theoretical analysis of the barometric method for measurement of tidal volume. Respiration Physiology, 32:105-120. Fons, R., and R. Sicart. 1976. Contribution a la conaissance du metabolisme energetique chez deux Crocidurinae. Mammalia, 40:299-311. Frey, H. 1979. La temperature corporelle de au cours de I’activite, du repos normothermique et de la torpeur. Revue Suisse de la Zoologie, 86:653-662. Frey, H., and P. Vogel. 1979. &ude de la torpeur chez Suncus etruscus en captivite. Revue Suisse de la Zoologie, 86:23-36. Gebczynska, Z., and M. Gebczynski. 1965. Oxygen consumption in two species of water shrews. Acta Theriologica, 10:203-214. Gebczynski, M. 1965. Seasonal and age changes in the metabolism and activity of Sorex araneus. Acta Theriologica, 10:303-331. 1971a. Oxygen consumption in starving shrews. Acta Theriologica, 16:288-292. \91\b. The rate of metabolism of the lesser shrew. Acta 1994 NAGEL — Metabolic Rates in European Shrews 427 Theriologica, 16:329-339. 1977. Body temperature in five species of shrews. Acta Theriologica, 22:521-530. Genoud, M. 1985. Ecological energetics of two European shrews: Crocidura russula and Sorex coronatus (Soricidae: Mammalia). Journal of Zoology, (London), 207:63-85. Goethel, H., and a. Nagel. 1990. Sauerstoffverbrauch, Atem- und Herztatigkeit bei der afrikanischen Zwergmaus Mus minutoides (Rodentia, Muridae). Zeitschrift fiir Saugetierkunde, Sonderheft, 55:19. Hart, J. S., and O. Heroux. 1955. Exercise and temperature regulation in lemmings and rabbits. Canadian Journal of Biochemistry and Physiology, 33:428-435. Hasler, M. J., and a. Nalbandov. 1974. Body and peritesticular temperatures in the musk shrew. Journal of Reproduction and Fertility, 36: 397-399. Hawkins, A. E., P. A. Jewell, and G. Tomlison. 1960. The metabolism of some British shrews. Proceedings of the Zoological Society of London, 135:99-103. Herreid, C. F., and B. Kessel. 1967. The thermal conductance in birds and mammals. Comparative Biochemistry and Physiology, 21:405-414. Hildwein, G. 1972. Metabolisme energetique de quelques mammiferes et oiseaux de la foret equatoriale. Archive des Sciences physiologiques, 26:397-400. Hyvarinen, H. 1969. On the seasonal changes in the skeleton of the common shrew (Sorex araneus) and their physiological background. Aquilo Serie Zoologica, 7:1-32. Irving, L. 1972. Arctic Life of Birds and Mammals. Springer- Verlag, Berlin, 192 pp. Iverson, G. 1972. Basal energy metabolism of mustelids. Journal of Comparative Physiology, 81:341-344. Kleiber, M. 1967. Der Energiehaushalt von Mensch und Haustier. Parey, Hamburg, 325 pp. Kusnetzov, V. J. 1972. On ecology of Crocidura suaveolens and Diplomesodon pulchellum in the Karakum Desert. Theriology, 1:266-276. Layne, j. N., and j. R. Redmond. 1959. Body temperature of the least shrew. Saugetierkundliche Mitteilungen, 7:169-172. Lindstedt, S. L. 1980a. Regulated hypothermia in the desert shrew. Journal of Comparative Physiology, 137:173-176. 1980fe. Energetics and water economy of the smallest desert mammal. Physiological Zoology, 53:82-97. Malan, a. 1973. Ventilation measured by body plethysmography in hibernating mammals and in poikilotherms. Respiration Physiology, 17:32-44. Martinsen, D. L. 1969. Energetics and activity patterns of short- tailed shrews (Blarina) on restricted diets. Ecology, 50:505-510. McNab, B. K. 1980. On estimating thermal conductance in endotherms. Physiological Zoology, 53:145-156. Merritt, J. F. 1986. Winter survival adaptations of the short-tailed shrew (Blarina brevicauda) in an Appalachian montane forest. Journal of Mammalogy, 67:450-464. MEZHZERIN, V. A. 1964. Dehnel’s phenomenon and its possible explanation. Acta Theriologica, 8:95-1 14. Morrison, P. R. 1948. Oxygen consumption in several small wild mammals. Journal of Cellular Physiology, 31:69-96. Morrison, P. R., and O. P. Pearson. 1946. The metabolism of a very small mammal. Science, 104:287. Morrison, P. R., and F. A. Ryser. 1952. Weight and body temperature in mammals. Science, 1 16:231-232. Morrison, P. R., F. A. Ryser, and A. R. Dawe. 1953. Physiological observations on a small shrew. Federation Proceedings, 12:100-101. 1959. Studies on the physiology of the masked shrew Sorex cinereus. Physiological Zoology, 32:256-271 . Morton, S. R. 1978. Torpor and nest-sharing in free-living Sminthopsis crassicaudata (Marsupialia) and Mus musculus (Rodentia). Journal of Mammalogy, 59:569-575. Mover, H., A. Ar, and S. Hellwing. 1986. Non-steady-state O2 consumption and body temperature in the shrew Crocidura russula monacha (Soricidae, Insectivora). Physiological Zoology, 59:369-375. Muller, E. F. 1975. Temperature regulation in the slow loris (Nycticebus coucang). Naturwissenschaften, 62:140-141. 1979. Energy metabolism, thermoregulation and water budget in the slow loris (Nycticebus coucang). Comparative Biochemistry and Physiology, 64A: 109-1 19. 1985. Untersuchungen zur Temperaturregulation bei der Wiistenrennmaus Gerbillus perpallidus Setzer, 1958. Zeitschrift fiir Saugetierkunde, 50:337-347. Muller, E. F., and E. Kulzer. 1977. Body temperature and oxygen uptake in the kinkajou (Polos Jlavus, Schreber) a nocturnal tropical carnivore. Archives Internationales de Physiologic et de Biochimie, 86:153-163. MOller, E. F., and H. Rost. 1983. Respiratory frequency, total evaporative water loss and heart rate in the kinkajou (Polos Jlavus Schreber). Zeitschrift fiir Saugetierkunde, 48:217-226. Nagel, A. 1977. Torpor in the European white-toothed shrews. Experientia, 33:1455-1456. 1980. Sauerstoffverbrauch, Temperaturregulation und Herzfrequenz der europaischen Spitzmause (Soricidae, Mammalia). U npublished Ph . D . dissert., Eberhard-Karls-Universitat, Tubingen, 86 pp. 1985. Sauerstoffverbrauch, Temperaturregulation und Herzfrequenz bei europaischen Spitzmausen (Soricidae). Zeitschrift fiir Saugetierkunde, 50:249-266. 1986. The electrocardiogram of European shrews. Comparative Biochemistry and Physiology, 83A:791-794. 1989a. Development of temperature regulation in the common white-toothed shrew, Crocidura russula. Comparative Biochemistry and Physiology, 92A:409-413. 1991 . Metabolic, respiratory and cardiac activity in the shrew Crocidura russula. Respiration Physiology, 85:139-149. Neal, C. M., and S. J. Lustick. 1974. Energetics and evaporative water loss in the short-tailed shrew. Physiological Zoology, 47:180-185. Pearson, O. P. 1947. The rate of metabolism of some small mammals. Ecology, 28:127-145. 1948. Metabolism of small mammals. Science, 108:44. Pfeiffer, C. J., and G. H. Gass. 1962. Oxygen consumption in the small short-tailed shrew Cryptotis parva. Transactions of the Illinois State Academy of Science, 55: 130-132. Platt, W. J. 1974. Metabolic rates of short-tailed shrews. Physiological Zoology, 47:75-90. Precht, H., j. Christophersen, and H. Hensel. 1955. Temperatur und Leben. Springer-Verlag, Berlin, 514 pp. Prinzinger, R. 1988. Energy metabolism, body-temperature and breathing parameters in nontorjiid blue-naped mousebirds Urocolius macrourus. Journal of Comparative Physiology B, 157:801-806. Prinzinger, R., and S. Jackel. 1986. Energy metabolism, respiration frequency and 02-consumption per breathing act in 1 1 different sunbird species during day and night. Experientia, 42:1002-1003. Prinzinger, R., and K. L. Schuchmann. 1985. Respiration frequency and tidal volume in resting hummingbirds during 428 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 daytime. Journal fiir Omitologie, 126:107-108. PUCEK, Z. 1965. Seasonal and age changes in the weight of internal organs of shrews. Acta Theriologica, 10:369-438. 1970. Seasonal and age changes in shrews as an adaptive process. Symposium of the Zoological Society of London, 26:189-207. SCHAICH, E. 1977. Schatz- und Testmethoden fur Sozialwissenschaftler. Franz Vahlen, Miinchen, 435 pp. SCHOLANDER, P. F. R. HOCK, V. WALTERS, AND L. iRVING. 1950. Adaptation to cold in arctic and tropical mammals and birds in relation to body temperature, insulation, and basal metabolic rate. The Biological Bulletin, 39:259-271. Sparti, A., AND M. Genoud. 1989. Basal rate of metabolism and temperature regulation in Sorex coronatus and Sorex minutus (Soricidae: Mammalia). Comparative Biochemistry and Physiology, 92A:359-363. Stahl, W. R. 1965. Organ weights in primates and other mammals. Science, 150:1039-1042. 1967. Scaling of respiratory variables in mammals. Journal of Applied Physiology, 29:453-460. Stahl, W. R., and J. R. Gummerson. 1967. Systematic allometry in five species of adult primates. Growth, 31:21-34. Tomasi, T. 1984. Shrew metabolic rates and thyroxine utilization. Comparative Biochemistry and Physiology, 78A:43 1-435. Vogel, P. 1974. Kalteresistenz und reversible Hypothermic der Etruskerspitzmaus S«ncM5 etruscus. Zeitschrift fiir Saugetierkunde, 39:78-88. 1976. Energy consumption of European and African shrews. Acta Theriologica, 21:195-206. Vogel, P., J. C. Ricci, and D. Cantoni. 1984. Raumliche und zeitliche Beziehung zwischen markierten Tieren einer Wildpopulation der Hausspitzmaus Crocidiira russula. Zeitschrift fiir Saugetierkunde Sonderheft, 49:63-64. Wahlstrom, a. 1929. Beitrage zur Biologic von Crocidura leucodon. Zeitschrift fiir Saugetierkunde, 4:157-185. Wang, L. C., and J. W. Hudson. 1971 . Temperature regulation in normothermic and hibernating eastern chipmunk Tamias s trial us. Comparative Biochemistry and Physiology, 38:59-90. Weathers, W. W. 1979. Climatic adaptation in avian standard metabolic rate. Oecologia, 42:81-89. Weibel, E. R., P. H. Buri, and H. Claasen. 1971. The gas exchange apparatus of the smallest mammal Suncus etruscus. Experientia, 27:724. Withers, P. C. 1977. Metabolic, respiratory and haematological adjustments of the little pocket mouse to circadian torpor cycles. Respiration Physiology, 31:295-307. Withers, P. C., A. K. Lee, and R. W. Martin. 1979. Metabolism, respiration and evaporative water loss in the Australian hopping- mouse Notomys alexis (Rodentia: Muridae). Australian Journal of Zoology, 27: 195-204. WOLK, E. 1974. Variations in the hematological parameters of shrews. Acta Theriologica, 19:315-346. Table 1. — Oxygen pulse (pi) per 100 mg heart mass in S. minutus, S. araneus, and C. russula, determined by actual values of oxygen pulse and of mean values of heart mass in the different species. S. minutus (46.6 ±4.4 mg, n = 3); S. araneus (72.2 ± 14.3 mg, n = 17);) C. russula (89.0 + 19.6 mg, n = 20). Ambient Temperature (°C) Species 0 5 10 15 20 25 30 35 37 S. minutus 2.55 2.75 2.34 2.45 2.49 2.27 2.14 S. araneus 3.88 3.72 3.54 3.32 3.13 3.09 3.05 C. russula 2.82 2.47 1.90 1.74 1.59 1.29 1.10 1.42 1.46 Table 2.— Comparison of physiological features in S. araneus and in Crocidura russula as typical representatives of the subfamily Soricinae atui Crocidurinae, respectively, in comparison to “standard” mammals. Species Sorex araneus Crocidura russula Body temperature Basal oxygen consumption Metabolic adaptation Respiratory rate Heart rate Ecological adaptation Distribution standard very high heat production very high standard to cold temperate, subarctic, and arctic zone low standard torpor low standard to heat temperate, subtropical, and tropical zone 1994 NAGEL— Metabolic Rates in European Shrews 429 Fig. 1. — Mean body temperature (±SD) at different ambient temperatures in Sorex minutus (triangle), Sorex araneus (square), and Crocidura russula (circle). 5 0 5 10 15 20 25 30 35 37 amoient lemoerature fX 1 Fig. 2.— Mean oxygen consumption ( + SD) at different ambient temperatures in Sorex minutus (triangle), Sorex araneus (square), and Crocidura russula (circle). NO. 18 430 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY r — I I JZ X O) c o Q. E ZJ (/) c 0 u c 01 cn >> X o 18 20 22 24 26 28 30 32 34 36 38 40 body temperature [°C] Fig. 3. — Oxygen consumption (single measurements) of torpid (Tb < 30°C) and normothermic (Tb > 30 °C) Crocidura russula (circles) and Sorex araneus (squares) in relation to body temperature at ambient temperature of 15 °C. Values from Nagel (1985) are included. £ 500 ■ 10 15 20 25 ambient temperature [“C ] J L_J 30 35 37 Fig. 4. — Mean respiratory rate (±SD) at different ambient temperatures in Sorex minutus (triangle), Sorex araneus (square), and Crocidura russula (circle). 1994 NAGEL — Metabolic Rates in European Shrews 431 ®r 0*- I — 1 > 1 I I 1 I I I "5 0 5 10 15 20 25 30 35 37 ambient temperature TCl Fig. 5. — Mean oxygen consumption per breath ( + SD) at different ambient temperatures in Sorex minutus (triangle), Sorex araneus (square), and Crocidura russula (circle). C 600 E I I I -5 0 S 10 15 20 25 30 35 37 ambient temperature (’Cl Fig. 6. — Mean heart rate (±SD) at different ambient temperatures in Sorex minutus (triangle), Sorex araneus (square), and Crocidura russula (circle). 432 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 800 - ^ 600 - C Qj 400 -f- ' 0 -4— » 1 200 - 100 18 20 22 24 26 28 30 32 34 36 body temperature [°C] Fig. 7. — Heart rate (single measurements) of torpid (Tb < 30°C) and normothermic (Tb > 30°C) Crocidura russula (circles) in relation to body temperature at ambient temperature of 15°C. Values from Nagel (1985) are included. 3,0 - 0.5 - oL - 5 0 5 10 15 20 25 30 35 37 ambient temperature [X] Fig. 8. — Mean oxygen pulse (±SD) at different ambient temperatures in Sorex minutus (triangle), Sorex araneus (square), and Crocidura russula (circle). 1994 NAGEL — Metabolic Rates in European Shrews 433 3r I I O) ZD Q- C O) cn X o («) (S) ® (5) 0 0 0 0 0 I I 1 I L_l_ 18 20 22 I I I I I I i i \ ! i 1 i i ! 24 26 28 30 32 34 36 38 body temperature [°C] Fig. 9.— Oxygen pulse (single measurements) of torpid (Tb < 30°C) and normothermic (Tb > 30°C) Crocidura russula in relation to body temperature at ambient temperature of 15 °C. Values from Nagel (1985) are included. M 'O) X c5^ c o CL E U) cz o o c O) cn >> X o 0 5 10 15 20 25 30 35 40 ambient temperature [°C] A oxygen consumption is caused by. A body temperature A basal metabolic rate Fig. 10. — Influence of body temperature and different basal metabolic rates in Crocidura russula (circles) and Sorex araneus (squares) on oxygen consumption. 434 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 Fig. II. — Metabolism of red-toothed shrews (Soricinae) and white-toothed shrews (Crocidurinae) compared with other mammals. body mass [g] Fig. 12. — Basal heart rates in relation to body mass (double log sclaes) in S. niinutus (closed triangles), S. araneus (closed squares), and C. russula (bull’s eye) compared with the mass-predicted values (Wang and Hudson, 1971) and with data from the literature: Neomys fodiens (closed circle), Crocidura suaveolens (open circle), Crocidura leucodon (open square) (Nagel, 1980); Suncus murinus (open diamond) (Dryden et al., 1971; Balakrishnan et al., 1974); Blarina brevicauda (asterisk) (Doremus, 1965); Sorex cinereus (closed diamond) (Morrison et al., 1959); Suncus etruscus (open triangle) (Weibel et al., 1971; Bartels et al., 1979). THE WHITE-TOOTHED SHREW CROCIDURA RUSSULA MONACHA: HOW DOES ITS PHYSIOLOGY FIT MAMMALIAN ALLOMETRY? Hay A Mover’, Amos Ar’, and Salo Hellwing’ ’Department of Zoology, Tel Aviv University, Ramat Aviv, Tel Aviv, 69978, Israel Abstract By eomparing physiologieal and reproduetive parameters of the eroeidurine shrew Crocidura russula monacha to interspecific eutherian mammalian allometric equations, deviations from common mammalian patterns are revealed. These include higher-than-predicted basal metabolic rate, heat transfer constant, heart mass, ventilation and gestation time, and lower-than-predicted lung mass and heart rate. Physiological limits, ecological adaptations, and inherited trends are thought to explain these deviations. Introduction Allometry provides a practical tool for analyzing the physiological consequences of body mass (Gould, 1966). Although body mass contributes a major fraction to the variability of many physiological parameters among mammals (Schmidt-Nielsen, 1984), significant deviations from a general allometric prediction occur and may reflect special cases, such as ecological adaptations (McNab, 1986), life history and reproductive patterns (Read and Harvey, 1989), or phylogenetic inherited characteristics (Gittleman, 1986; Calder, 1987; Elgar and Harvey, 1987). When plotted allometrically, values of shrew variables occupy the extreme left side of mammalian allometric curves, raising the questions of a possible limit to mammalian size on one hand, and of mammalian physiological potential on the other. A high basal metabolic rate (BMR) (at rest, postabsorptive state, and in thermoneutral i ty ) is characteristic of shrews. The white-toothed shrew Crocidura russula monacha Thomas, 1906, cf. Crocidura suaveolens Pallas, 1811 (Catzeflis et al., 1985) has a body mass of 7-9 g. Its specific oxygen consumption rate at BMR is 26% higher than expected from body mass allometry (Mover et al., 1986). Thus, high cardiac output, respiratory ventilation, and food intake are also expected. Reproductive parameters such as gestation period and postnatal growth rate are also known to be related to maternal rate of energy expenditure (McNab, 1986). The purpose of this study is to examine if circulatory, respiratory, and reproductive variables measured in C. r. monacha fit the expected values obtained from interspecific allometric relationships, i.e. , that they may be explained by body mass considerations alone, and thus if possible deviations may be related to physiological limitations, special ecological adaptations, or inherited characters. Material and Methods Animals. — Nonreproductive female shrews were kept long enough in separate cages (40 X 30 X 5 cm) at 29 °C and 14L: lOD conditions to be considered su mmer-acc 1 i mated . Live fly larvae and minced meat mixed with milk powder were supplied once a day. Water was supplied ad libitum. Resting oxygen consumption rates, respiratory frequencies, tidal volumes, and heart rates were measured simultaneously. Six females were used in this experiment. The female gender was chosen to provide a base line for comparisons with reproductive parameters. All measurements were performed in a constant temperature cabinet at 31°C, which is within the thermoneutral zone of these shrews (28-36°C; Mover et al., 1986). Heart and Lung Masses. — Hearts and lungs were removed from 15 nonreproductive females. The lungs including tracheae were weighed immediately after removal. Blood in the ventricles and atria was rinsed with saline, the fluids were absorbed, and the empty hearts were immediately weighed (Mettler BA28; +0.01 mg). Ventilation and Oxygen Consumption.— A measuring chamber consisting of two small cylindrical, horizontal plastic tubes (internal diameter [ID] = 4 cm) separated into head (3 cm long) and body (7 cm long) compartments by an elastic latex membrane was constructed from the finger of a surgeon’s glove. The membrane had a hole in its center which was lubricated with silicon grease and fitted to be gas-tight around the shrew’s neck, separating head from body compartment, without apparent interference of the normal respiratory rate. This was confirmed by comparing the respiration rate to that of an unrestrained animal in the same chamber. The body compartment tube was perforated to allow free air movement due to volume changes during respiration. In order to measure accurately the rate of oxygen consumption, dry, C02-free air was drawn through the head compartment which served as an “open” metabolic chamber, at a constant rate (75-90 ml /min) by an air pump. The same air was also used to calibrate the O2 analyzer ( ± 0.005 % ) . Oxygen consumption was calculated using the redried C02-free outgoing flow rate ( + 0.5 ml) and the 0-, concentration in it as: O2 consumption rate = outgoing gas flow rate times O2 fraction difference between incoming and outgoing gas divided by (1 — O2 fraction of incoming gas) (Depocas and Hart, 1957) where the O2 uptake and the flow rate are given in ml dry gas at 0°C, pressure of 760 Torr, per hour, STPD. An Applied Electrochemistry O2 Analyzer, model S-3A, was used to measure oxygen concentrations, and 16 X 0.7 cm ID columns of 20-mesh “Drierite” and “Ascarite” were inserted in the flow paths in series to absorb water and CO2 from the gas before entering and after leaving the head compartment. The changes in air-flow rates caused by inspiration and expiration in the head compartment were 435 436 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 measured across a fixed resistance portion of the outflow tubing as pressure differences (range: + 1 mm H2O) by a Valedyne DP15TL pressure transducer, and used for calculating respiratory rate. Calibration of tidal volume was performed by producing artificial tidal volumes (10-70 /:tl) at frequencies of 150-450 min“*. This was done by moving mercury in a vertical graduated glass tube like a piston, using a variable speed motor, and inspecting the mercury deflections. The glass tube was connected to the head compartment for calibration through a metal tube of the same dimensions as the trachea of a shrew (10 mm long X 0.9 mm ID). Respiratory flow oscillations produced both artificially and by animals were recorded on a Gould recorder (model 2660), digitized off-line (Hipad Tablet Digitizer), and fed to a computer (PDP 11/23) where time integrations of the respiratory flow cycles around the mean flow were made to produce water vapor saturated tidal volume values expressed in fil at experimental temperature and pressure. Heart Rate. — The floor of the body compartment contained two rows of three electrically separated, copper-conducting square sheets (each 2x1 cm). A lead from each copper sheet was connected to oscilloscope (Tetronix 502) inputs through a selection switchbox. For ECG inspection, pairs of leads from these sheets, which were spread with electrically conductive paste in contact with the legs, were selected. The amplified ECG was recorded simultaneously with ventilatory parameters. Animals were allowed to reach steady-state values before the experiments began. Each female was measured once. Each measurement included O2 consumption and 5-15 sequences of at least ten respiratory cycles and the simultaneously-measured heart rate. The reproduction and thermoregulation parameters were taken from Mover et al. (1986, 1988). The expected values were calculated from relevant allometric equations using the mean body mass of 7.9 g. Results Table 1 summarizes a comparison of mass specific oxygen consumption rate in BMR conditions between C. r. monacha (O2 consumption of 2.56 ml [STPD]/[g-h] ±0.56 SD) and allometric predictions, and some values from the literature. Table 2 describes the BMR and heat transfer constant of C. r. monacha recalculated in energy units assuming a respiratory coefficient of 0.88 (Mover et al., 1986), and its ventilatory and circulatory parameters at BMR conditions. The results are compared with available interspecific allometric predictions from the literature. In Table 3 we have compared measured values of gestation time, litter mass, neonatal brain mass, and averaged embryonic growth rate of C. r. monacha (newborn mass divided by gestation time) taken from Mover et al. (1988) with values predicted using allometric equations. Discussion Nonreproductive State. — Unlike the soricine shrews, the values of the oxygen consumption rate at rest of crocidurine shrews tend to approach those that are expected from their body masses (Vogel, 1976). However, in the white-toothed shrew Crocidura russula monacha, the values obtained were generally higher than those predicted for a typical eutherian mammal (Table 1). Our measurements were made on summer-acclimated animals during the resting phase of the circadian cycle (daytime), at complete rest, and only mean minimal values were taken into account (see Mover et al., 1986, for details). Summer-acclimated C. russula are also known to have an elevated oxygen consumption in low ambient temperatures (Lardet, 1989). Nevertheless, this oxygen consumption is much lower than expected within its systematic group due to the much elevated BMR of the Soricinae (Table 1; Nagel, 1994). The elevated value of oxygen consumption of C. r. monacha compared with mammals in general could result in a relatively high body temperature as was found in the soricine shrews (Sparti and Genoud, 1989). However, because it is accompanied by a proportionately elevated heat-transfer constant (Table 2), body temperature (32-37°C: Mover et al., 1986) varies on the low side of the normal mammalian range (Schmidt-Nielsen, 1984). Why such a strategy of high energy expenditure and heat loss has evolved in this species is not clear. The values given by Nagel (1985) and Sparti (1990) are 31 and 37 % lower for the European C. russula, but Sparti (1990) reports a value similar to ours for C. suaveolens. Our summer-acclimated animals may be adapted primarily to withstand heat-stress conditions where a high thermal conductance is advantageous in preventing overheating. Lardet (1989) found for the European C. russula a heat transfer constant in winter which was 41 % lower than in our shrews, and in summer, a constant which was only 32% lower. The potential for high heat production helps to overcome the low heat capacity, which is a function of small size. The nocturnal mode of activity, when ambient temperatures in the field are relatively low, can now be associated with increased energy expenditure and heat production during this activity. Conflicting demands due to fluctuations in ambient temperatures between day and night, especially in winter, at the ground level, and in the near-arid zones where C. r. monacha are found in Israel may also explain the presence of a significant capacity for nonshivering thermogenesis (NST) found in our summer-acclimated shrews (4.4 ml [STPD]/(g-h); unpublished data). How are the higher-than-expected oxygen demands of C. r. monacha met by the circulatory and respiratory systems? Table 2 shows that the changes from the expected, which occur in the heart and lungs of this shrew, deviate in opposite directions: the heart is relatively large and slow in rate, while the lungs are relatively small and respiratory frequency is high. This leads to an unusual ratio of heart rate to respiration frequency of 2. 1 in C. r. monacha, compared with the ratio of 4.3 as predicted from mammalian allometry (Stahl, 1967). A relatively large heart may lead to efficient heart work since it enables a high stroke volume resulting in only a minor increase in the heart’s oxygen consumption, while an increase in heart rate notably increases the heart’s metabolic needs (Roth, 1976). A smaller heart with a higher beat frequency could reach the limit of contraction-relaxation time of the heart muscle, especially 1994 MOVER ET AL. — Shrew Physiology and Mammalian Allometry 437 during activity (Schmidt-Nielsen, 1984). The large density of mitochondria needed to activate a small, fast-beating heart may impair its mechanical activity (Weibel, 1984) and thus a larger heart may have a favorable mitochondrial fraction. In addition, as in athletes, a relatively large heart may increase the scope of its performance during activity. Ventilation is more than proportionately elevated compared with the oxygen demands. The higher-than-expected ventilation is achieved by increasing respiration frequency while tidal volume is kept near expected values (Table 2). Because the small size of the shrew makes its CO2 capacity low, in the face of a very high specific CO2 production rate, the high ventilation rate may be necessary to regulate both CO2 stores and blood pH more precisely. It will be interesting to test this idea in other small mammals. From the measured data and assuming a dead space of one-third tidal volume (Dejours, 1981), we calculated an alveolar O2 pressure of about 1 10 Torr for C. r. monacha, which is very high compared with the 88 Torr predicted according to Weibel (1984). This “hyperventilation” maintains a high oxygen head pressure in the alveoli which facilitates oxygen diffusion to the blood. Since both lung mass (Table 2) and lung diffusion capacity of the shrew correspond to body mass and not to its higher-than-expected energy metabolism (Weibel et al., 1981), relative hyperventilation may provide a mechanism for overcoming O2 diffusion difficulties through the alveolar-blood barrier. It may also provide an important avenue for evaporative heat loss since the specific ventilation of about 1.5 ml/(g-min) is very high (in man it is about 0.1 ml/[gTnin]), but in order to evaluate it quantitatively, expired air temperature has to be measured. Reproductive State. — In eutherian mammals, high metabolic rates will tend to correlate with a short gestation period for a given maternal size (McNab, 1980, 1986). Martin (1981) contends that maternal metabolic rate during gestation determines the neonatal brain mass, and Hofman (1983) found a strong correlation between maternal metabolism, litter birth mass, and neonatal brain mass. For C. r. monacha these relationships do not hold; although BMR exceeds the expected value, the gestation period is far longer than expected (Table 3). According to the above correlations, the combination of such a long gestation period and high BMR should have resulted in a rapid intrauterine growth and a large neonate brain mass. However, neonatal brain mass, embryonic growth rate, and litter body mass in C. r. monacha fit rather precisely the allometrically expected values (Table 3). In fact, crocidurines are similar to soricine shrews in their growth rate (Vogel, 1972) despite their longer gestation period. This may indicate that the newborn of C. r. monacha are relatively more advanced in their development at birth as indicated for Crocidurinae in general (Vogel, 1981). While it seems that maternal body mass explains some of the reproductive variables in C. r. monacha, the long gestation, relative precociality, and short postnatal development period (Hellwing, 1971) are better explained as phylogenetically inherited characters. As in the metatherian pattern (Lillegraven et al., 1987), no correlation between body mass, metabolic rate, and gestation period can be detected. The long gestation period in C. r. monacha may serve to partition in time the maternal energy investment during simultaneous lactation and pregnancy (Mover et al., 1989). Females conceive on the day of delivery. During the first seven days of gestation, the embryonic litter mass is almost nil. It reaches a total mass of 0.08 g only at the end of the tenth day (Mover et al., 1988), while during these ten days energy intake for milk production is rising to a value which is 4.8 times that needed to sustain the nonpregnant, nonlactating female shrew (Mover et al., 1989). Omitting these first ten days brings the gestation time of 28 days closer to the predicted value (Table 3). In conclusion, many characteristics of the white-toothed shrew C. r. monacha fit the general predictions for mammalian values expected for such a body mass. However, there are deviations due to three main constraints: 1) physiological performance limitations, such as a lower-than-expected heart rate apparently to avoid approaching the upper frequency and efficiency limits; 2) ecological adaptations where the low body capacity limitations challenge homeostasis, e.g., thermoregulation, as demonstrated by the high intensity of metabolism and the high heat-transfer coefficient that nevertheless result in a “normal” body temperature, and the relative hyperventilation which enables both higher-than- expected oxygen delivery and fast regulation of CO2 and pH levels; and 3) an inherited reproductive strategy in which body mass and the high BMR have no visible correlation with the neonatal development rate. Instead, the longer-than-predicted gestation period helps to partition the energy intake demands of the simultaneously lactating and pregnant female. Literature Cited Bartels, H. 1980. Aspects of respiratory gas transport in mammals with high weight specific metabolic rates. Verhandlungen der Deutschen Zoologischen Gesellshaft, 73:108-201. Bradley, S. R., and D. R. Deavers. 1980. A re-examination of the relationship between thermal conductance and body weight in mammals. Comparative Biochemistry and Physiology, 65A:465-476. Brody, S. 1945. Bioenergetics and Growth. Reinhold, New York, 1023 pp. Calder, W. A. 1982. The pace of growth: An allometric approach to comparative embryonic and post-embryonic growth. Journal of Zoology, London, 198:215-225. 1987. Scaling energetics of homeothermic vertebrates: An operational allometry. Annual Review of Physiology, 49:107-120. Catzeflis, F., T. Maddalena, S. Hellwing, and P. Vogel. 1985. Unexpected finding on the taxonomic status of east Mediterranean Crocidura russula auct., (Mammalia, Insectivora). Zeitschrift fiir Saugetierkunde, 50:185-201. Dejours, P. 1981. Principles of Comparative Respiratory Physiology. Elsevier/North-Holland Biomedical Press, Amsterdam, 265 pp. Depocas, F., and S. J. Hart. 1957. Use of the Pauling oxygen analyzer for measurement of oxygen consumption of animals in open-circuit systems and in a short-lag, closed-circuit system apparatus. Journal of Applied Physiology, 10:388-392. Elgar, M. A., and P. H. Harvey. 1987. Basal metabolic rate in mammals; allometry, phylogeny and ecology. Functional Ecology, 1:25-36. Gittleman, J. L. 1986. Carnivore life history patterns: Allometric, 438 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 phylogenetic, and ecological associations, American Naturalist, 127:744-771. Gould, S. J. 1966. Allometry and size in ontogeny and phylogeny. Biological Review, 41:587-640. Hayssen, V., and R. C. Lacy. 1985. Basal metabolic rates in mammals: Taxonomic differences in the allometry of BMR and body mass. Comparative Biochemistry and Physiology, 81A:741-754. Hellwing, S. 1971. Maintenance and reproduction in the white-toothed shrew, Crocidura russula monacha Thomas, in captivity. Zeitschrift fiir Saugetierkunde, 36:103-113. Hofman, M. A. 1983. Evolution of brain size in neonatal and adult placental mammals: A theoretical approach. Journal of Theoretical Biology, 105:317-331. Kleiber, M. 1961. The Fire of Life. An Introduction to Animal Energetics. John Wiley, New York, 454 pp. Lardet, J.-P. 1989. Variations saisonnieres de la thermoregulation de la musaraigne musette, Crocidura russula (insectivores, soricides). Revue Suisse de Zoologie, 96:90-108. Lillegraven, J. A., S. D. Thompson, B. K. McNab, and J. L. Patton. 1987. The origin of eutherian mammals. Biological Journal of the Linnean Society, 32:281-336. Martin, R. D. 1981. Relative brain size and basal metabolic rate in terrestrial vertebrates. Nature, 293:57-60. McNab, B. K. 1980. Food habit, energetics, and the population biology of mammals. American Naturalist, 116:106-124. 1986. The influence of food habits on the energetics of eutherian mammals. Ecological Monographs, 56:1-19. 1988. Complications inherent in scaling the basal rate of metabolism in mammals. Quarterly Review of Biology, 63:25-54. Mover, H., A. Ar, and S. Hellwing. 1986. Non-steady-state O2 consumption in the shrew Crocidura russula monacha (Soricidae, Insectivora). Physiological Zoology, 59:369-375. Mover, H., S. Hellwing, and A. Ar. 1988. Energetic cost of gestation in the white-toothed shrew Crocidura russula monacha (Soricidae, Insectivora). Physiological Zoology, 61:17-25. Mover, H., A. Ar, and S. Hellwing. 1989. Energetic costs of lactation with and without simultaneous pregnancy in the white-toothed shrew Crocidura russula monacha. Physiological Zoology, 62:919-936. Nagel, A. von. 1985. Sauerstoffverbrauch, Temperaturregulation und Herzfrequenz bei europaischen Spitzmause (Soricidae). Zeitschrift fiir Saugetierkunde, 50:249-266 1994. Metabolic rates and regulation of cardiac and respiratory function in European shrews. Pp. 421-434, in Advances in the Biology of Shrews (J. F. Merritt, G. L. Kirkland, Jr., and R. K. Rose, eds.), Carnegie Museum of Natural History Special Publication no. 18, x + 458 pp. Rahn, H. 1980. Comparison of embryonic development in birds and mammals: Birth weight, time and cost. Pp. 124-137, in A Companion to Animal Physiology (C. R. Taylor, K. Johansen, and L. Bolis, eds.), Cambridge University Press, Cambridge, 365 pp. Read, A. F., and P. H. Harvey. 1989. Life history differences among eutherian radiation. Journal of Zoology, London, 219:329-353. Roth, C. F. 1976. Cardiodynamics. Pp. 337-361, in Physiology (Selkurt, E. E., ed.). Little, Brown and Co., Boston. Schmidt-Nielsen, K. 1984. Scaling: Why is Animal Size So Important? Cambridge University Press, Cambridge, 241 pp. Sparti, A. 1990. Comparative temperature regulation of African and European shrews. Comparative Biochemistry and Physiology, 97A:391-397. Sparti, A., and M. Genoud. 1989. Basal rate of temperature regulation in Sorex coronatus and S. minutus (Soricidae: Mammalia). Comparative Biochemistry and Physiology, 92A:359-363. Stahl, W. R. 1967. Scaling of respiratory variables in mammals. Journal of Applied Physiology, 22:453-460. Vogel, P. 1972. Vergleichende Untersuchung zum Ontogenesermodus einheimischer Soriciden (Crocidura russula, Sorex araneus und Neomys fodiens). Ph.D. dissert., Universitat Basel. 1976. Energy consumption of European and African shrews. Acta Theriologica, 21:195-206. 1981. Occurrence and interpretation of delayed implantation in insectivores. Journal of Reproduction and Fertility, Supplement, 29:51-60. Weibel, E. R. 1984. The Pathway for Oxygen. Structure and Function in the Mammalian Respiratory System. Harvard University Press, Cambridge, 425 pp. Weibel, E. R., C. R. Taylor, P. Gehr, H. Hoppler, O. Mathieu, and G. M. O. Maloiy. 1981. Design of the mammalian respiratory system. Functional and structural limits for oxygen flow. Respiratory Physiology, 44: 151-164. 1994 MOVER ET AL.— Shrew Physiology and Mammalian Allometry 439 Table 1. — A comparison of mass-specific oxygen consumption rate at rest, thermoneutrality , and post-absorptive state, between Crocidura russula monacha, selected allometrically predicted values, and observed values. The upper part compares C. r. monacha with general mammalian pattern; the middle part, within a limited systematic group which includes soricid values; the lower part, with other measurements of populations of the same species. Positive values represent an increase of C. r. monacha above predicted or observed; negative values, below predicted or observed. Type of Comparison Deviation (%) Authority Eutherian mammals, full range + 16 Brody, 1945 tt ff " " + 26 Kleiber, 1961 tt ft " " + 8 Hayssen & Lacey, « t) up to 450 g + 22 to + 35 McNab, 1988 Rodents and insectivores up to 100 g + 11 Bartels, 1980 Insectivores -83 Hayssen & Lacey, Soricomorpha -100 McNab, 1988 C. russula -4 Nagel, 1985 C. russula + 11 Nagel, 1994 C. russula -14 Sparti, 1990 C. suaveolens + 10 Nagel, 1985 C. suaveolens -13 Sparti, 1990 Table 2. — Measured and expected heat transfer coefficients, ventilatory, and circulatory parameters of the white-toothed shrew Crocidura russula monacha. ^Kleiber, 1961; ^Bradley and Deavers, 1980; ^Stahl, 1967. Variable Oxygen Consumption (niW/g) Heat Transfer Coefficient (mW/g °C) Respiration Frequency (1/min) Tidal Volume (mO Ventilation (BTPS) (ml/min) Lung Mass (mg) Heart Mass (mg) Heart Rate (1/min) Measured 14.28 2.29 273 47 11.7 81.4 68.2 566 ±SD 3.1 0.01 68 12 2.3 9.51 8.15 71.7 Predicted 11.38' 1.75“ 188^ 50^ 7.88^ 93. 7^ 50.5^ 808^ Measured/predicted 1.26 1.31 1.45 0.94 1.48 0.87 1.36 0.70 Table 3. — Measured and expected reproduction parameters of the white-toothed shrew Crocidura russula monacha. ^ Mover et al. , 1988; ^Hofman, 1983; ^Rahn, 1 980; ^Calder, 1982. Variable Gestation Time (d) Litter Body Mass (g) Neonatal Brain Mass (mg) Embryonic Growth Rate (g/d) Measured' 28 2.79 61.91 0.032 + SD 1 0.14 6.95 0.002 Predicted 17“ 2.61^ 63.03“ 0.034^ Measured/predicted 1.65 1.06 0.98 0.94 ..'- -n^ .1 ’^■ ' \'"' i- /-J.- ... V ^ k - ■'■'?»!■ • . , *«» ■ ' . . .. ,4^..*.i. -T -A «aKakaitt^.»«.^^ ;'-rfi<'.^ ‘ .■'•i^%* "% ■ifi;, ' '■ j_ \ "' -/4)WVi .. , ' f‘'‘ ■ • -^-nM .. ... , . ■ _> ., .* '■‘ ” " ''“ ■ » . ■4>.»''y!)ft'.li3ll^' i>iB: ■AUrrf.m .‘> ^ .1 1.4 i j.iiV',’ .' )■’. V,.. ' _ ••A*«i'> "I'. I ' ■. 'V‘'> ■’•J^v'i|pil^3«<«r''- .. .-■ 5 ,•" •;< fw rs ■■■’•-•■ j fja '. , ■ 1 ' ' • .< ■> ' H *p*->i 5 i ’■'■i ^ '•' >„ H ■' >in' ,,.. :.^W44.a'= ^ ',.iWi*V' i .. ' THE STRUCTURE AND ADAPTIVE PECULIARITIES OF PELAGE IN SORICINE SHREWS Ernest V. Ivanter Department of Zoology, University of Petrozavodsk, Petrozavodsk 185640, Russia Abstract The fur structure of four soricine shrews, Sorex araneus, S. caecutiens, S. minutus, and Neotnys fodiens, was investigated in animals collected at different seasons in extreme western Russia. Hairs of shrews can be classified into four categories based on structure and function: monotrich. Type I guard. Type II guard, and woolly. All except monotrichs are tapered with longitudinal bands, have three to seven segments, and grow singly and perpendicularly from the skin surface. Guard and especially woolly hairs are short and thin with well-marked distal segments, whereas monotrichs are long and thick. Winter hairs are thinner and longer than summer hairs, due mainly to an increase in the number of segments from three to four to six to seven. Also, the density of the pelage increases by 20-30 percent as winter approaches. Three distinct concentric regions (core, cortical, and cuticular) were revealed in the microstructure of the hair shaft. A shrew hair is characterized by the presence of two longitudinal furrows near the tips of guard hairs, smooth margins of cuticular scales, an interrupted pattern of the cortical layer of the shaft, and zonal coloration of the shaft. The insulative property (coefficient of heat conductivity) of shrew pelage also was examined. Physical thermoregulation is important in the maintenance of temperature homeostasis during the molt period. Introduction Soricine shrews have extremely high levels of metabolism (McNab, 1991; Nagel, 1994) yet must survive the low winter temperatures of the temperate and subarctic locations where they occur. Their small size precludes use of long, dense hairs or fatty layers for insulation, and consequently they lose much heat by thermal conductance. To minimize heat loss and maintain a favorable energy balance, these shrews have acquired a complex of ecological and morphological adaptations for surviving the energy costs of thermoregulation . The objective of this study was to investigate the role of the fur, including the microstructure of different hairs and peculiarities of the molt, as adaptations for physical thermoregulation. Materials and Methods The skins of 157 shrews of four species {Sorex araneus, S. caecutiens, S. minutus, and Neomys fodiens) were examined. For comparison, representatives of other families of Insectivora and Rodentia were studied as well, including Talpa europaea {n — 37), Sicista betulina {n = 18), Clethrionomys glareolus {n 42). and Microtus oeconomus {n = 14). Investigations of hair distribution on the skin, structure of a single hair shaft, density and length of hair, and topography of the hair cover were conducted according to traditional methods (Williams, 1938; Borowski, 1952; Sokolov, 1973). The microstructure of hair shafts was observed using a microscope at 600-1 350 X magnification (immersion lens). Imprints of the cuticular layer were made on slides coated with clear nail polish. The insulative properties of skins were determined (by the coefficient of thermoconducti vi ty method) with the use of a special apparatus, IT-3, from the Institute of Technical Thermophysics of the Ukraine Academy of Sciences in Kiev. This instrument measures heat loss from a constant source through a dried, flat skin. Results General Characteristics of the Fur The pelage of soricine shrews is distinctive. The fur is short, velvety, soft, perpendicular to the skin surface, and does not form the so-called “hair flow” which is common in most mammalian pelage. Because of their segmental structure, the hairs lie easily in any direction, do not rumple, and retain air. Shrew hairs are thin and consist of slightly thickened or widened segments interrupted by occasional constrictions where the shaft bends at an obtuse angle (Fig. 1, 2). This feature makes the fur elastic and gives a layered effect to the surface regardless of the direction the hair tips point. All four species of shrews examined had dark, predominantly brown hairs on the dorsal side, and light, dirty-white hairs on the venter. Most had a well -pronounced transitional color zone on each flank, but this feature was absent in one specimen of Neomys fodiens. The color of fur is determined by the presence and concentration of melanin and lipochrome, the black and brown pigments. The major tone of coloration is determined by the concentration of lipochrome in the distal segments of the guard hairs. Other segments of all hairs contain melanin. Pigments are absent in the upper parts of the distal segments, which have no coloration. However, the light-reflecting scales add a specific silver luster to the fur. Newly molted fur looks more vivid, shiny, and lustrous compared to old, dull fur with frayed ends. Differentiation of the Hairs Shrew hairs are classified into four categories: monotrich. Type I guard. Type II guard, and woolly (Fig. 1, 2). The monotrich is thick, long, and elastic. The shaft is nearly straight and spindle-shaped but round in cross section. Its base is not twisted, and segments and constrictions are absent. The shaft gradually narrows at the tip, becoming filiform and nearly colorless. The thickness of the shaft also narrows at the root. 441 442 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 Monotrichs are outnumbered 150-200: 1 by the other hair types. Although not adding significantly to the fur cover, monotrichs fulfill an important sensory function. Compared to other hair types, the sacs of monotrichs lie in deeper layers of the dermis where the nerve ganglia are more developed. The guard hair is shorter and thicker than the monotrich, and is distinctly segmented. The distal lancet-shaped segment of the guard hair is most pronounced, being longer and wider than other segments and comprising 25-50 percent of the length of the shaft. The longitudinal bends of the shaft allow the guard hair to lie in any direction. There are two types of guard hairs. The Type I guard hair is widest and longest in segments among the segmented hairs, and is longer in winter having six segments than in summer with 3-4 segments present. Whether in summer or winter pelage, the distal segment constitutes up to 50 percent of the length of the shaft. The distinctive feature of the guard hair is the presence of longitudinal grooves on the sides of the distal segments, creating an H-shaped profile in cross section (Fig. 2). These grooves are closed in the constriction and tip of the guard hair, a feature that is distinctive and diagnostic for the Soricinae. Species-specific features are less pronounced. A distinctive feature in N. fodiens but in other soricines is the “internal bar,” which is formed by a diagonal slat on the bottom of the groove (Fig. 2). The Type II guard hair is slightly shorter and thinner than Type I. Its terminal segment comprises up to 25 percent of the shaft length. In both winter and summer. Type II hairs have one segment more than Type I guard hairs; i.e.. Type II hairs have seven segments in winter and 4-5 segments in summer (Tables 1-3). The guard hairs have zonal coloration, a character which varies in different parts of the pelage. On the abdomen, the end segments of guard hairs are white, but on the back and flanks they are dark brown but colorless at the very tips. Other segments of the guard hairs of all parts of the body are intensely black. The highest concentration of melanin occurs in the core of widened parts of segments, but in narrow areas the melanin is diffusely distributed, producing lighter coloration. Large inter- and intracellular air cavities refract the light, producing a lustrous sheen to the hair in appropriate lighting. Together with the zonation of coloration, the combination of dark and light segments determines the variety of fur shades in different parts of the pelage. The main functions of guard hairs are to protect the more delicate woolly hairs from mechanical damage and rumpling, and provide the structure or matrix for the other hairs of the pelage. Besides bending to cover the woolly layer, the thickened terminal segments of guard hairs promote the formation of the insulative air layer and enhance thermoregulation. Woolly hairs, although thinnest and shortest, have the same number of segments as Type II guard hairs. The terminal lancet-shaped segment is less developed than the other seg- ments. The density of woolly hairs is greater than that of other types, and can make up 40-60% of all hairs. Their main func- tion is to provide insulation and reduce thermal conductance. These characteristics of the four main types of hair are extremely heterogeneous within each type. Yet, in comparison with other mammals, hair differentiation in shrews is much less variable. This is explained by their ancient origin, their fossorial adaptations, and the presumed adaptive value for each hair type and the specific combination of types (Ivanter et al., 1985). Each hair type participates equally in mechanical and thermal protection of the body, but the presence of multiple types of hair permits the specialization of some types to enhance thermoregulation and others to improve the mechanical and protective qualities without an increase in length, density, or mass of the fur. Such an increase would impede the active movements of small mammals. Hair Shaft Structure Two distinctly delimited concentric layers (cortical and cuticular) surrounding a core region were revealed in the microstructure of the hair shaft (Fig. 3). The cuticula, forming the external covering of the shaft, consists of one layer of homed, nonconcentric scale cells with the open side directed toward the tip of the shaft. TTiese unpigmented, semitransparent cuticular scales differ in form and size in different regions of the shaft. In both the preroot zone and constricted sections, cuticular scales are elongated with smooth, round borders and do not fit tightly with one another. (In rodents, the scale borders in such places are toothed.) This loose fit in shrews provides the necessary flexibility of the shaft. Where the shaft widens, particularly in the distal segment, the cuticular scales become short, wide, and toothed on the borders (Fig. 4, Sa, Sc, Sm, Sb). This design keeps the hair shaft rigid. In rodents, by contrast, scale cells of cuticula look like narrow, twisting ribbons that cling to the shaft, giving each a characteristic appearance (Fig. 4, Cg and Mo). The characteristics of winter and summer hair cuticula varied only slightly in each species of shrew. Age variation was also minor. There were few differences in form and size of scales, and the general morphology was conservative. Compared to hair cuticula from other systematic groups inhabiting similar environments, no common features were noted. However, closely related species from different habitats had similar cuticular structures. For example, the cuticula of shrews in the genus Sorex differed more from Clethrionomys from the same habitats than from the cuticula of other insectivores, such as Talpa, a fossorial mammal, or Neomys, a semiaquatic mammal. Overall, hair characteristics are determined more by taxonomic affiliation than by ecological factors. The moderately thick cortical layer, which lies between the cuticular layer and the core, consists of the spindle-shaped and highly keratinized cells which provide the tensile strength of the shaft. The cortical layer is well-formed along the entire length of the shaft, and forms the cover of the central channel. The cortical layer has no pigment, but in the Rodentia it contains melanin granules and, together with the core, determines the coloration of the hair. In rodents, the cortical layer is thinner and more evenly distributed along the shaft. The coloration and thickness of the hair depend on the development and structure of the core, especially the size and position of the lens-shaped cells, the presence of pigment, and 1994 I VANTER— Structure and Function of Soricine Pelage 443 intra- and intercellular air-filled cavities. The diameter of the core changes in the growing hair, and varies in different parts of the hair shaft. The end segment of the core is most highly developed, where numerous pigment granules, lying like loosely stacked coins, determine the hair coloration. In the widest part of the segment, core cells are arranged in three rows (4-6 rows in rodents), whereas at the base and tip of the shaft and near constrictions, the core becomes two-rowed. In thin, bent areas of the hair, such as between segments and in the preroot section of the growing hair, the core is very narrow, thread-like, sometimes interrupted, and has a diffuse distribution of pigment. In these places, the hair is covered with the longest scales. In the preroot section and tip of the fully grown hair, both core and pigment are practically absent. In contrast, the core region is well-developed and pigmented in the root of the growing hair shaft, especially in the initial period of growth when hairs are located inside the skin. As a hair grows, the most recently produced section at the base of hair shaft is filiform. The gradually widening channel initially does not have a pronounced cellular structure and is characterized by a scattering of pigment. Later this changes to a strict alternation of air cavities and pigmented cells, reaching maximum development in the terminal segment of the hair shaft. Within the general microstructural features of the hair shaft are found numerous variations and deviations, depending on the type of hair, grade of maturation, location on the skin, season of the year, and species of mammal. The core region is most variable. For example, the core of each monotrich and guard hair has 1-2 rows of core cells in its thinnest parts but many rows in the segments of the thicker part. In contrast, in woolly hairs, the core is either almost nonexistent or interrupted, has only one row of core cells along the entire shaft, and contains numerous air chambers. The relationship between the core and cortical regions differs among hair types. The cortical layer is best developed in monotrichs, less in guard hairs, and least in woolly hairs, whereas the core is best developed in the shafts of woolly hairs. The seasonal changes are more apparent. As a rule, the shaft of a winter hair has a thicker core but a thinner cortical layer. The winter increase in core thickness is due mainly to the enlargement of the air cavities, thus improving the insulative qualities of the fur. Fur Density Hair density is one of the most variable indices of the fur structure. Both intra- and interspecific differences were found (Table 4). The ftir of shrews is denser than that of rodents in both winter and summer. Among the Insectivora, the densest fur belongs to Talpa europaea with hair density 1.5 times greater than that of Sorex minutus, the smallest shrew examined. Between these extremes of density lie, in order, Neomys fodiens, Sorex araneus, and S. caecutiens. Clethrionomys glareolus has the least dense hairs of all, but Sicista betulina, with 100-150 hairs per mm^, has the densest fur in its summer pelage. Although the body appears fully furred, hair density varies at different locations (Table 4). In most of the eight species examined, the thickest fur occurred on the back. The topography of fur density depends on the animal’s ecology as well as its habitat. The hair density of Talpa europaea is thinnest on the sides, while hair density in shrews is thinnest on the abdomen. However, the hair density of Neomys fodiens is uniform across the sides and abdomen, but is still thickest on the back. Seasonal changes in hair density are more obvious. The winter pelage is 1.2-1. 4 times thicker than in summer, due to increased numbers of woolly hairs rather than guard or directing hairs, whose numbers remain constant. The highest density was observed in the fur of molting animals, when “old” hairs remain among intensively growing “new” hairs. The hair density also depends on the age of the animal, confirmed by the negative relationship between hair density and body size. Thickness of the Hair Shaft Shafts of different types of hair vary in thickness (Table 5). Woolly hairs not only are thinner than guard hairs and monotrichs, but are more uniform in thickness. The width of the lancet-shaped terminal segment of a woolly shaft from Sorex araneus varies from 4. 5-5.0 fi, with a coefficient of variation (CV) of 3.9%. In the Type II guard hair, the diameter varies from 26.0-28.7 /x, with a CV of 13.6%. The shaft thickness also changes with season. The winter hairs of all types are thinner, softer, and lighter in color than summer hairs. The relative thickness of the core channel increases from summer to winter, from comprising 50-65% of the total diameter in summer to 60-85 % in winter (Table 5). Therefore, although the total thickness of the hair shaft increases in summer, the core narrows. In winter, the shaft narrows and the thickness of the cortical region increases. Among the majority of species investigated, in cross section the hairs are thickest on the back and thinnest on the abdomen. Neomys fodiens and Talpa europaea are exceptions, with abdominal hair slightly thicker than that of the sides and back. This is due to the animals’ specific life habits. As previously mentioned, the hair shafts with greatest diameter are those of monotrichs; guard hairs are slightly thinner, and woolly hairs the thirmest. However, an opposite relationship holds for the proportion of the core in cross section. The core is large in woolly hairs, intermediate in guard hairs, and small in monotrichs. Therefore, the thinner the shaft, the greater the contribution of the core and the smaller the contribution of the cortical layer. Length of the Hairs Three or four zones based on the length of hairs were identified in the pelage of eight species of small mammals (Fig. 5). Among Sorex, the sacrum (stippled pattern. Fig. 5) had the longest and densest fur. A second zone, with hairs 0.5-0. 7 mm shorter, covered nearly all of the remaining back and sides. A third zone surrounded the second like a narrow ribbon; and a fourth zone covered the venter, where the hairs are shortest. According to the nomenclature of Tserevitnov (1958), such pelage topography is called the “sacral-equal type.” The “equilateral type” of fur, characteristic of Neomys fodiens, had 444 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 hair of moderate length on the back, long hair on the sacrum and flanks, and short hair on the venter. The “equal type” of fur on Talpa europaea consists of a relatively constant length of fur over the entire body. The pelage of shrews caught in summer was much shorter than that of winter shrews (Table 6). The summer fur was formed by three-segmented Type I guard and four-segmented woolly and Type II guard hairs, whereas the winter fur had six- segmented (Type I) and seven-segmented (Type II) guard hairs and woolly hairs. The summer guard hairs of Sorex araneus averaged 4.2 mm, with the terminal segments comprising 2.6 mm of this length. In the winter, the guard hairs averaged 7.5 mm and the terminal segments were 2.8 mm. Thus, the winter guard hairs were 1. 6- 1. 8 times longer than in summer, an increase seen also for other categories of hair (Table 6). Overall, all categories of hairs from winter-caught S. araneus were longer than in summer. Among the monotrichs, winter fiir covering different parts of the four species of soricine shrews was 0.9-3. 8 mm (X = 2.3 mm), or 30% greater than summer values. Among the guard hairs, this range was 1.6-3. 7 mm (X = 2.7 mm, 42%), and among woolly hairs, the range was 0.7-3. 8 mm (X = 2.00 mm, 37%). Thus, from summer to winter the hair length increased an average of 30-40%. Insulative Properties of the Pelage The insulative properties of the pelage depend in part on the “inert” air in the core region but more importantly on the thickness of the “stable” air that reduces the convection current which moves heat from the skin surface through the pelage to the surface of the fur. These features combine to provide an effective thermal insulation (Scholander et al., 1950; Hammel, 1955; Irving and Krog, 1955; Sokolov, 1973; Schmidt-Nielsen, 1975; Ivanter et al., 1985). A pelage of densely packed thick, long hairs intermixed with thin hairs provides a stable air layer, and consequently heat is better conserved compared to other pelage. Thermoconductivity also depends on the amount of air in the core of hair shafts (and thus is greater when the core has little air) and on the pores of the skin surface and, in turn, on the density and thickness of the skin. Hair structure also affects thermoconductivity based on the presence of longitudinal grooves on the shaft, the number of segments, and the manner in which the hairs lay in the pelage. The pattern of heat conductivity from dry prepared skins of animals collected at different seasons confirmed that the combination of changes in hair density and quality combined to affect the insulative properties of the pelage (Table 7). The thermoconductivity of the fur showed well-pronounced taxonomic and seasonal variations, as determined by the structural peculiarities of the pelage. The longer, thinner and more numerous the hair shafts and the thicker the skin, the lower the thermoconductive coefficient and, consequently, the better the insulation. Each species demonstrated a pronounced seasonal variability, and this suggested a correlation between the thermoconductive coefficient and hair length, thickness, and numbers. The coefficient of heat conductivity (CHC) of summer skins was greater than that of winter skins (Table 7); in Sorex araneus this difference was 27.3%; in S. caecutiens, 34.6%; in S. minutus, 26.8%; in Talpa europaea, 30.2%; in Clethrionomys glareolus, 21.5%; in Microtus oeconomus, 30.9%; and in Neomys fodiens, 7.3%. These values corresponded to the seasonal changes in the length and thickness of the fur. The rank correlation index of Spearman (r,) between hair length and specific thermoconductivity of the skin was statistically significant (r, = —0.41, t = 2.3, P < 0.05). The correlation between CHC and density of the fiir was higher and more significant (r^ — —0.66, t = 3.6, P < 0.01). When hair length and hair density were combined and correlated with CHC, the correlation was still greater (r, = 0.90, P < 0.001). In brief, the rate of heat loss was inversely proportional to the density of the pelage. The fiir of shrews was shorter and denser than that of voles. In addition, the segmentary structure of hair prevents rumpling and allows the fiir to lie easily in any direction. This keeps the “stable” or trapped air within the fur, and improves its insulative properties. Neomys fodiens and Talpa europaea, which have the longest and densest pelage of the insectivores examined, have insulative properties nearly twice that of any of the three Sorex species. The insulative properties of mammal skins at the peak of molting were better those of winter skins (Table 7). As mentioned, thermoconductivity is reduced by the thickening of the skin and increase in fur density due to the combination of old hairs and newly emerging ones. Thus, the efficiency of thermoregulation in shrews during the molting period is increased during this transition. Energy demands are increased by 20-30% during the molting period. Because molt may occur during seasons when the abundance and quality of food is changing, energy savings resulting from pelage would be highly adaptive during the molt. Discussion Among the four insectivores in this study are representatives of three adaptive types, i.e., semiterrestrial {Sorex sp.), fossorial {Talpa), and semiaquatic {Neomys). This division allowed the identification of the most typical and obviously adaptive features of hair structure for each species and assessment of these features from an ecological perspective. By life habits, environment, and fur structure, shrews combine many features that are characteristic of terrestrial and fossorial mammals. They inhabit temperate and cold climates where conditions of constant thermal deficit prevail. Thus, these insectivores must have fiir which is light and suitable for easy movement, but with high insulative properties. The fur is characterized by moderate density and length of the hair, an uneven density and length of cover over different parts of the body, the mace-shaped form of the terminal segment, the slight twisting of the shaft at the base of the hair, and a relatively well-developed core in the center of the hair shaft. It is well-known that solid hair shafts are not good insulators because of the relatively high thermoconductivity of the keratin. The principal role in thermoprotection , besides the “inert” air in air chambers of the shaft, belongs to the so-called “stable” air formed in the pelage as a result of maximum reduction of 1994 IVANTER — Structure and Function of Soricine Pelage 445 convection currents. In shrews, the structure of the fur promotes the maintenance of this insulative layer of air. Hair shafts are located singly and perpendicular to the skin surface, are segmented and twisted, and differentiated into categories. Guard and monotrichs grow among and protect the woolly hairs, not only protecting the fur from rumpling but also reducing the escape of air and its heat. By having a relatively less developed hair core as compared with rodents, but with much higher hair durability by the thickened cortical layer, shrews compensate for the deficit of “inert” air in the core of individual hairs by an increased content of “stable” air in the fur. As a result, the thermoconductivity coefficient of shrew fur is less than that of rodents (Table 7). However, the lower insulative properties of the fur of voles is not maladaptive. Voles compensate by obtaining other mechanisms to maintain a favorable energy balance; e.g., they have better chemical thermoregulation, occupy microhabitats possessing burrows with constant temperatures, and construct nests to aid in reducing thermal conductance. The uneven character of the hair cover on different zones of the body, a characteristic of semi terrestrial species, is an important additional component of physical thermoregulation . The zone in the middle of the back, which undergoes the most severe cooling, is covered by the longest and densest fur. The venter is least protected from the cold, because the hair cover is sparser and shorter. The usefulness of such topography is apparent when the shrew rolls into a ball during sleep. In this configuration, the ventral part of the body is covered, exposing a minimal area of surface for thermoradiation. Also, the exposed parts of the sleeping shrew are mostly well-furred. The pelage also protects the shrew from mechanical damage by reducing the pressure of the surrounding substrate as shrews move in narrow burrows, leaf litter, bushes, and grasses. To some extent, the mechanical protection function is opposed to the insulative function. The most durable and protective hair shafts would have poorly developed cores, whereas thinning leads to a decrease in insulative properties. This contradiction is solved in several ways. One is the numerical increase in thin- cored shafts, a compromise that promotes both good mechanical and insulative properties. A second way is the differentiation of hair types, a sort of division-of-labor function where the guard and monotrichs offer mechanical protection and the woolly hairs provide good insulation. A third mechanism involves the irregularity of the structure of the hair shaft along its length. For protection, coreless ends of hair shafts are especially valuable, with their high mechanical properties provided by the well -developed cortical layer. Other narrowed parts of the hair shaft also contribute to high durability with their slightly developed or absent cores and thick tougher cortical layers. In such places the hair shaft is easily bent to any side without damage, providing elasticity and thereby preventing the shaft from being fractured or crushed. In Talpa europaea, the categories of hair types are similar, producing fur that is relatively short, smooth, and dense with a slightly pronounced pile and little taper to the tail. Such specific structure of the fur allows the animal to move freely forward and backward in narrow burrows. Microstructural peculiarities of mole hair have pronounced adaptive value. For example, the slight development of the core layer of the hair shaft and consequent greater cortical layer improve the mechanical qualities. That this applies only to directing and guard hairs and not to woolly hair, which has thicker cores than other hairs, improves the thermoresistance of the hair. The woolly hair, covered by the guard and directing hairs, experiences minimal mechanical disturbance, thereby reducing the need for mechanical properties and permitting a specialization toward greater insulative properties. As it has denser, longer fur and thicker skin than the shrew, the mole’s external cover offers greater thermoresistant properties. The structure of the hair cover of Neomys fodiens shares features common to both terrestrial shrews and moles. Its hairs, unlike those of other insectivorans, are longest on the flanks, a feature which may improve the hydrodynamics as it swims and dives. Other morphological peculiarities include the presence of longitudinal grooves of the flattened guard hairs of the flanks; the relatively slight development, compared with other Insectivora, of the core in grana of covering hairs; the thinning and lengthening of all types of hair, the latter by increase in segment number; and the higher density and thickness of the skin. These characters combine to provide an improved ability to retain heat in an air layer within the fur. Thus, the fur stays dry, making a more perfect insulative covering. The character of molting is also distinctive, for the changes occur gradually and molting is prolonged (Ivanter et al., 1984, 1985). Finally, N. fodiens has the countershading of the body that is typical of aquatic animals, with a dark back and light underside and no transitional zone at the flanks. According to Cott (1940), this is due to the optical effect, desirable for inhabitants of the upper layers of water, which obscures the body by masking the shadow. Other characteristic features of the aquatic N. fodiens include the more pronounced differentiation of hairs into categories, and the uniform type of pelage topography. The monotrichs and guard hairs are typically wider and have more flattened terminal segments. These produce an overlapping, almost tile-shaped cover which, because of the surface tension of the water, retains the insulative air layer in the woolly hair (Gudkova- Aksenova, 1951; Sokolov, 1973). Entrapment of the air layer is enhanced by the density of woolly hairs, the twisting character of their shafts, and the comparatively better-developed core region (72-73% versus 49-50% in covering hairs). Typically, the guard hairs are straighter than in other shrews. Contacts between cuticular scales and the hair shaft, and between the scales themselves produce a sleek surface for the hair. In addition, close spacing of the rows of scales provides an effective protection from moistening. This research also supports suggestions by a number of authors (Appelt, 1973; Hutterer and Hurter, 1981; Vogel and Kopchen, 1978) about the typical profile of the cross section of guard hairs from N. fodiens. Because of the numerous diagonal cuticular slats that cover the bottom of longitudinal grooves on the guard hairs on the flanks, the H-profile has 1. 5-2.0 times more teeth than shrews in the genus Sorex. This peculiarity is also an adaptation to swimming, because additional teeth 446 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 contribute to trapping air within the fur. In conclusion, shrews have four types of hair that differ in their contributions to protection and thermoregulation. The long, thick, and elastic monotrichs comprise only 0.5% of hairs, but serve an important sensory function. Wooly hairs are thin and short, but constitute 40% (summer) to 60% (winter) of the pelage density and are important in thermoregulation . The two types of guard hairs protect as well as insulate. Winter pelage is 20-40 % thicker and hairs 30-40% longer than in summer. In winter, all hairs are thinner, softer, and lighter in color than in summer, and the hollow cores of the hairs increase substantially. Despite being thinner in diameter, winter hairs have larger hollow cores, comprising 60-85% of hair cross section. Insulation is due to the dead air space of the pelage that reduces heat convection from the body, and, to a lesser extent, to the hollow nature of hairs. Tests with dried skins showed that winter pelage was about 30 % more effective at retaining heat than summer pelage. Overall, shrews are able to minimize thermal conductance during the winter months through a combination of changes in hair quality and hair density. Literature Cited Appelt, F. 1973. Fellstrukturuntersuchungen an Wasserspitzmausen. Abhandlungen und Berichte aus dem Museum “Mauritianum” Altenburg, 8:81-87. Borowski, S. 1952. Sezonowe smiany uwlosienia u Sorididae. Annals of the University of Marie Cu rie-Sklodowska , Section C, 7:65-117. COTT, H. B. 1940. Adaptive Coloration in Animals. Methuen, London. Gudkova- Aksenova, I. S. 1951 . Environment and its influence on organization of some water Insectivora and Rodentia. Annals of the University of Gorky, 19: 135-174 (in Russian). Hammel, T. H. 1955. Thermal properties of fur. American Journal of Physiology, 182:369-371. Hutterer, R., and T. Hurter. 1981. Haarstrukturen bei Wasserspitzmausen (Insectivora, Soricinae). Zeitschrift fur Saugertierkunde, 46: 1-11. Irving, L., and J. Krog. 1955. Temperature of skin in the Arctic as a regulator of heat. Journal of Applied Physiology, 7:355-364. IVANTER, E. V., T. V. IVANTER, AND R. V. Levina. 1984. The adaptive peculiarities of the structure of hair cover and of the molt in semiaquatic mammals, Neomys fodiens taken as an example. Soviet Journal of Zoology, 63:245-255 (in Russian). IVANTER, E. V., T. V. IVANTER, AND I. L. TUMANOV. 1985. Adaptive Peculiarities of Small Mammals. Nauka, Leningrad (in Russian). McNab, B. K. 1991. The energy expenditure of shrews. Pp. 35-45, in The Biology of the Soricidae (J. S. Findley and T. L. Yates, eds.). Special Publication, The Museum of Southwestern Biology, 1:1-91. Nagel, A. 1994. Metabolic rates and regulation of cardiac and respiratory function in European shrews. Pp. 421-434, in Advances in the Biology of Shrews (J. F. Merritt, G. L. Kirkland, Jr., and R. K. Rose, eds.), Carnegie Museum of Natural History Special Publication no. 18, x + 458 pp. Schmidt-N IELSEN , K. 1975. Animal Physiology: Adaptation and Environment. Cambridge University Press, England. Scholander, P. F., R. Hock, V. Walters, F. Johnson, and L. Irving. 1950. Heat regulation in some arctic and tropical mammals and birds. Biological Bulletin, 99:237-258. Sokolov, V. E. 1973. The skin cover of mammals. Nauka, Moscow (in Russian). Tserevitinov, B. F. 1958. The topography peculiarities of hair cover of fur-bearing animals. Annals of the Institute of Furs, Moscow, 17:256-267 (in Russian). Vogel, P., and B. KopcheN. 1978. Resondere haarstrukturen der Soricidae (Mammalia, Insectivora) und ihre taxonomische deutung. Zoomorphologie, 89:47-56. Williams, C. S. 1938. Aids to the identification of mole and shrew hairs with general comments on hair structure and hair determination. The Journal of Wildlife Management, 2:239-250. 1994 IVANTER — Structure and Function of Soricine Pelage 447 Table 1. — Characteristics of summer fur from three body regions in Sorex araneus. Type of Hair n Number of Hairs Per 4 mm^ Length of Hair (mm) Thickness of Hair (ji) Number of Segments Back Monotrich 17 3.6 ± 0.2 5.1 ± 0.06 50.0 ± 0.3 1 Type I guard 25 114.0 ± 4.0 4.2 ± 0.01 50.0 ± 0.4 3 Type II guard 25 89.8 ± 5.7 4.0 ± 0.05 27.2 ± 0.4 4 Woolly 25 235.3 ± 7.1 3.6 ± 0.02 12.6 ± 0.1 4 Side Monotrich 16 2.3 ± 0.1 4.7 ± 0.03 49.9 ± 0.5 1 Type I guard 24 94.6 ± 5.7 4.3 + 0.04 49.7 ± 0.6 3 Type II guard 25 66.3 ± 6.5 4.1 ± 0.09 26.1 ± 0.4 4 Woolly 26 232.4 ± 5.7 3.4 ± 0.01 9.8 ± 0.2 4 Abdomen Monotrich 18 2.0 ± 0.1 4.4 ± 0.04 49.8 ± 0.6 1 Type I guard 25 169.2 ± 6.7 4.1 ± 0.04 49.7 ± 0.4 3 Type II guard 24 83.6 ± 4.9 3.6 ± 0.03 28.0 ± 0.3 4 Woolly 23 224.6 ± 7.1 3.3 ± 0.01 10.2 ± 0.1 4 Table 2. — Characteristics of winter fur from three body regions in Sorex araneus. Type of Hair n Number of Hairs Per 4 mm^ Length of Hair (mm) Thickness of Hair (pi) Number of Segments Back Monotrich 16 4.1 ± 0.3 8.5 i 0.09 30.2 ± 0.2 1 Type I guard 20 119.2 ± 3.1 7.5 + 0.08 30.1 ± 0.6 6 Type II guard 20 135.6 ± 5.0 7.0 ± 0.08 25.4 ± 0.4 7 Woolly 24 264.2 + 5.1 6.2 i 0.03 7.2 + 0.3 7 Side Monotrich 15 3.0 ± 0.1 8.5 ± 0.06 30.1 ± 0.3 1 Type I guard 16 117.0 + 4.2 7.1 ± 0.05 30.0 + 0.5 6 Type II guard 18 104.2 + 3.6 6.9 ± 0.03 25.2 ± 0.3 7 Woolly 21 205.4 ± 3.9 5.9 i 0.02 7.2 ± 0.4 7 Abdomen Monotrich 17 2.3 ± 0.1 7.9 + 0.09 31.0 ± 0.4 1 Type I guard 24 122.3 ± 3.3 6.1 ± 0.07 30.2 ± 0.1 6 Type II guard 26 141.7 + 5.1 5.6 ± 0.05 25.1 ± 0.8 7 Woolly 22 299.1 ± 5.0 4.6 ± 0.02 7.4 ± 0.6 7 448 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 Table 3. — Characteristics of summer fur from three body regions in Neomys fodiens. Type of Hair n Number of Hairs Per 4 mm^ Length of Hair (mm) Thickness of Hair (p.) Number of Segments Back Monotrich 17 3.5 ± 0.2 8.0 ± 0.02 50.0 + 0.4 1 Type I guard 25 137.2 ± 5.1 6.5 + 0.11 27.6 ± 0.5 4 Type II guard 16 108.1 + 5.3 6.2 ± 0.01 23.0 ± 0.3 5 Woolly 18 218.6 ± 5.0 6.0 ± 0.02 9.8 + 0.1 5 Side Monotrich 16 2.9 ± 0.1 8.5 ± 0.08 50.0 + 0.6 1 Type I guard 25 120.2 ± 3.4 7.0 + 0.06 28.1 ± 0.5 4 Type II guard 21 106.9 ± 3.9 6.5 ± 0.09 21.3 + 0.3 5 Woolly 24 120.9 + 4.1 6.2 + 0.03 10.0 + 0.1 5 Belly Monotrich 14 2.1 ± 0.1 7.5 ± 0.08 50.0 ± 0.7 1 Type I guard 18 128.3 + 4.1 5.9 + 0.07 37.1 ± 0.6 4 Type II guard 19 107.1 + 4.0 5.6 + 0.05 27.2 + 0.2 5 Woolly 25 130.6 ± 3.9 5.3 ± 0.02 10.2 ± 0.1 5 Table 4. — Seasonal difference in the density of hairs (number of hairs/4 mrn^) in five species of insectivores and three species of rodents from northwestern Russia. Species Season n Back Side Abdomen Sorex araneus summer 25 442.7 ± 7.3 395.6 ± 7.9 479.8 ± 4.6 winter 25 523.1 ± 7.2 429.6 + 4.4 565.4 + 9.0 Sorex caecutiens summer 25 389.9 ± 6.8 376.3 ± 7.7 332.8 ± 7.6 winter 25 538.5 + 8.9 432.7 + 8.7 460.9 ± 9.1 Sorex minutus summer 14 364.0 ± 6.7 328.3 ± 6.0 266.4 ± 4.3 winter 15 463.4 ± 5.3 413.1 ± 5.1 395.9 ± 4.7 Neomys fodiens summer 18 467.4 + 6.1 350.9 + 5.3 368.1 + 5.0 Talpa europaea summer 16 521.1 + 9.6 368.9 ± 9.2 457.9 ± 5.9 winter 15 728.7 ± 9.8 538.1 + 9.1 572.0 ± 9.2 Sicista betulina summer 18 576.3 ± 8.3 495.7 ± 3.1 381.9 ± 7.0 Clethrionomys glareolus summer 15 299.9 ± 3.0 288.8 ± 3.1 225.4 ± 1.8 winter 13 422.5 + 4.6 399.2 ± 4.1 284.7 ± 2.1 Microtus oeconomus summer 15 334.0 ± 2.9 288.1 ± 1.9 317.6 ± 2.1 winter 15 478.7 ± 6.1 453.6 ± 5.6 460.2 ± 5.8 1994 IVANTER — Structure and Function of Soricine Pelage 449 Table 5. — Thickness of hairs on the backs of five insectivores and three rodents from northwestern Russia. “Percent ” refers to the thickness of the cortical layer as a percentage of the total thickness of the hair. Species Season Guard Hairs Woolly Hairs n Thickness of Hair {pi) Percent n Thickness of Hair (pi) Percent Sorex araneus summer 25 40.0 ± 0.4 58.0 25 12.6 + 0.1 63.0 winter 20 30.1 ± 0.6 63.4 24 7.2 ± 0.3 77.0 Sorex caecutiens summer 50 31.1 + 0.6 64.3 28 11.0 ± 0.8 61.8 winter 28 27.0 ± 0.4 61.5 26 7.9 + 0.3 78.5 Sorex minutus summer 25 27.6 ± 0.4 54.7 29 10.0 ± 0.3 61.0 winter 24 23.8 ± 0.02 64.7 33 6.1 ± 0.3 80.3 Neomys fodiens summer 16 27.6 ± 0.05 50.0 18 9.8 + 0.1 72.1 Talpa europaea summer 25 32.4 + 0.3 50.9 19 10.9 ± 0.2 54.2 winter 17 27.8 ± 0.5 66.1 25 9.7 ± 0.2 70.1 Sicista betulina summer 20 17.4 + 00.2 79.9 18 15.8 ± 0.2 81.0 Clethrionomys glareolus summer 22 49.1 + 0.3 79.5 20 28.1 + 0.2 78.2 winter 18 20.1 + 0.2 71.3 18 11.9 ± 0.2 79.9 Microtus oeconomus summer 10 49.1 ± 0.6 87.0 10 22.8 + 0.7 54.3 winter 9 48.6 ± 0.6 87.8 9 16.4 + 0.3 81.6 Table 6. — Length of hairs (mm) in five species of insectivores and three species of rodents. Sample sizes in parentheses. Species Season Back Side Abdomen Sorex araneus summer 4.2 ± 0.01 (25) Guard Hairs 4.3 ± 0.04 (24) 4.1 ± 0.04 (25) winter 7.5 ± 0.08 (20) 7.1 ± 0.05 (16) 6.1 + 0.07 (24) Sorex caecutiens summer 4.0 + 0.05 (50) 4.0 ± 0.04 (60) 3.7 ± 0.3 (50) winter 7.5 + 0.07 (28) 7.7 ± 0.04 (29) 5.2 ± 0.5 (26) Sorex minutus summer 3.3 ± 0.01 (25) 3.2 ± 0.01 (25) 2.9 ± 0.01 (24) winter 5.2 ± 0.02 (24) 5.1 ± 0.01 (25) 4.7 ± 0.01 (25) Neomys fodiens summer 6.5 + 0.11 (25) 7.0 + 0.06 (25) 5.9 ± 0.07 (18) Talpa europaea summer 7.9 ± 0.02 (25) 7.7 ± 0.01 (18) 6.4 ± 0.04 (18) winter 11.2 ± 0.05 (19) 10.9 ± 0.04 (18) 10.0 + 0.06 (19) Sicista betulina summer 8.1 ± 0.09 (25) 7.1 ± 0.09 (25) 6.7 ± 0.04 (25) Clethrionomys glareolus summer 8.4 ± 0.03 (22) 8.0 ± 0.03 (26) 7.5 ± 0.02 (28) winter 11.2 ± 0.03 (18) 9.9 ± 0.03 (22) 9.1 ± 0.03 (26) Microtus oeconomus summer 12.2 ± 0.05 (14) 11.9 ± 0.05 (11) 10.1 ± 0.06 (9) winter 15.1 ± 0.05 (9) 14.7 ± 0.05 (10) 13.8 ± 0.06 (16) Sorex araneus summer 3.6 + 0.02 (25) Woolly Hairs 3.4 ± 0.01 (26) 3.3 ± 0.01 (23) winter 6.2 ± 0.03 (24) 5.9 ± 0.02 (21) 4.6 ± 0.02 (22) Sorex caecutiens summer 3.5 ± 0.02 (25) 3.3 ± 0.01 (25) 3.1 ± 0.01 (40) winter 5.2 + 0.03 (26) 5.0 + 0.01 (25) 3.8 + 0.02 (23) Sorex minutus summer 2.8 ± 0.04 (29) 2.8 ± 0.04 (27) 2.4 ± 0.01 (25) winter 4.2 ± 0.02 (33) 4.0 ± 0.04 (25) 3.6 ± 0.05 (27) Neomys fodiens summer 6.0 ± 0.02 (18) 6.2 ± 0.03 (24) 5.3 + 0.02 (25) Talpa europaea summer 6.5 ± 0.02 (29) 6.4 + 0.03 (25) 5.1 + 0.01 (25) winter 10.1 ± 0.02 (25) 9.4 ± 0.01 (20) 8.9 ± 0.02 (25) Sicista betulina summer 7.5 ± 0.11 (27) 6.5 ± 0.08 (25) 5.7 ± 0.04 (24) Clethrionomys glareolus summer 7.3 + 0.02 (20) 7.0 + 0.04 (22) 6.6 ± 0.02 (32) winter 9.4 ± 0.03 (18) 9.1 ± 0.03 (28) 7.9 ± 0.02 (24) Microtus oeconomus summer 10.9 ± 0.04 (16) 10.0 ± 0.03 (15) 8.6 ± 0.06 (10) winter 12.3 + 0.07 (9) 11.9 ± 0.05 (10) 11.0 ± 0.02 (14) 450 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 Table 7. — Coefficient of heat conductivity (CMC = 10 ^ w/m °K) of dry skins in seasonal samples of five insectivores and three rodents from northwestern Russia. Note: autumn specimens were molting. Species Season n CHC X ± SE Sorex araneus summer 16 38.2-50.0 47.6 + 0.8 winter 15 35.0-45.4 37.4 ± 0.6 autumn 12 33.1-41.2 34.7 ± 0.5 Sorex caecutiens summer 16 50.0-60.0 54.0 ± 1.2 winter 15 38.3-42.6 40.1 ± 0.5 autumn 13 37.1-44.8 39.4 ± 0.3 Sorex minutus summer 15 49.0-59.2 53.5 + 0.9 winter 14 38.0-49.0 42.2 ± 0.5 autumn 9 36.0-45.0 39.0 ± 0.4 Neomys fodiens summer 18 30.0-42.1 32.5 + 0.9 autumn 14 28.0-39.0 30.3 + 0.8 Talpa europaea summer 9 24.2-30.0 26.7 ± 0.3 winter 8 19.0-22.4 20.5 ± 0.3 autumn 8 18.9-21.0 19.1 ± 0.3 Sicista betulina summer 22 48.0-50.3 48.6 + 0.3 autumn 4 37.9-38.2 38.1 ± 0.4 Clethrionomys glareolus summer 20 45.2-53.1 49.8 + 0.9 winter 16 37.0-50.0 41.0 ± 0.8 autumn 14 36.1-42.1 38.7 ± 0.8 Microtus oeconomus summer 11 46.7-53.2 47.8 ± 0.4 winter 5 34.1-38.4 36.5 + 0.4 autumn 8 34.1-40.9 36.1 ± 0.6 1994 451 IVANTER — Structure and Function of Soricine Pelage Fig. 1. — Summer hairs in Sorex araneus (Sa), Neomys fodiens (Nf), and Talpa europaea (Te). Numbers refer to: 1, monotrichs; 2, Type I guard hairs; 3, Type II guard hairs; 4, woolly hairs. Fig. 2. — Structure of winter hairs in Sorex araneus. I, monotrichs; 2, Type I guard hairs; 3, Type II guard hairs; 4, woolly hairs; 5-7, texture of the cuticular layer: 5, in lower zone of hair; 6, in zone of bending; 7, wide part of hair segment; 8, H-shaped profiles in cross section of guard hairs of S. araneus (upper) and Neomys fodiens (lower). 452 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 Sa Nf Fig. 3. Microstructure between segments (1) and in the distal segments (2) of the guard hair shaft in Sorex araneus (Sa), and Neomys fodiens (Nf, 3). 1994 IVANTER — Structure and Function of Soricine Pelage 453 Fig. 4.— Structure of the cuticular layer in Sorex araneus (Sa), S. cinereus (Sc), S. minutus (Sm), Sicista betulina (Sb), Clethrionomys glareolus (Cg), and Microtus oeconomus (Mo). 454 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 Fig. 5. Topography of hair length (mm) in Sorex araneus (Sa), 5. cinereus (Sc), S. minutus (Sm), Neomys fodiens (Nf), Talpa europaea (Te), Sicista betulina (Sb), Clethrionomys glareolus (Cg), and Microtus oeconomus (Mo). Each figure represents a ^ed Skin and the patterns represent the areas of uniform hair lengths of the dimensions given next to the symbols in the boxes. The stippled pattern is found on the sacral region (rump). SORICID BIOLOGY: A SUMMARY AND LOOK AHEAD Robert K. Rose Department of Biological Sciences, Old Dominion University, Norfolk, Virginia 23529-0266 The International Colloquium on the Biology of the Soricidae, which was held from 8 to 14 October 1990 at the Carnegie Museum of Natural History’s Powdermill Biological Station, represents the first comprehensive international meeting devoted to the study of shrews. During the first half of this century, cricetid and murid rodents occupied center stage in research to understand the biology of mammals, and more recently research by bat biologists has yielded important insights into mammalian physiology, systematics, community structure, neural activity, and ecomorphology . Papers presented at the Powdermill colloquium and included in this volume revealed that shrews also have the potential for advancing our understanding in many areas of mammalian biology. In this chapter, I summarize the information and ideas presented by participants in the colloquium, while trying to identify areas of significant advances as well as potential areas for future research. The family Soricidae (shrews) is the largest family in the order Insectivora, with 20 genera and 290 species. Among families of Recent mammals, only the Muridae (murid rodents, 1,122 species) and Vespertilionidae (common bats, 313 species) are larger. Shrews are widespread in the Nearctic, Palearctic, Oriental, and Ethiopian biogeographic regions, but barely extend into the northern Neotropical and are absent from the Australian region. Two suborders of shrews are recognized: Soricinae (red-toothed shrews) and Crocidurinae (white-toothed shrews). Soricine shrews, including such representative genera as Sorex, Blarina, Crypt ot is, and Neomys, are largely Holarctic in distribution, with a few species in Central America and northwestern South America. They are most abundant and diverse in cool, moist boreal forest regions. In contrast, crocidurines are abundant and widespread in the Old World tropics and subtropics, especially Africa, the Indian and Indochinese subcontinents, and Malaysia. Limited geographic overlap between the distributions of soricine and crocidurine shrews occurs mainly in southern Europe and along the Palearctic-Oriental boundary. With 185 species in such genera as Crocidura, Suncus, and Myosorex, crocidurine shrews are more diverse than soricines. They are also more diverse in body size and thermoregulatory ability, and are potentially more complex in patterns of distribution and ecology. We know considerably less about the biology of crocidurines than about soricine shrews because the distribution of the majority of mammalogists, who reside principally in Eurasia and North America, coincides with the distribution of soricines. Many crocidurine species, including some of the nearly 150 species of Crocidura, are known from relatively few specimens or localities. To illustrate the magnitude of the problems of studying crocidurines, Zaire, with 52 species of crocidurine shrews, has more shrew species than are found in the Americas. Methodologies Most information on the distribution, populations, and community structure of shrews has come from studies using kill traps, either snap or pitfall traps. In North America, knowledge of shrew distribution and abundance has been greatly advanced in the past two decades with the increasingly widespread use of pitfall traps; some species, such as Sorex hoyi and S. longirostris , can only be studied effectively with this method. However, methods or procedures for pitfall trapping differ widely for a variety of reasons, including the objectives of the study, the nature of the substrate, and the availability of materials and manpower. Results are most comparable to other studies when standardized methods for the shape of pitfall arrays or the spacing and length of pitfall transects (such as those presented in this volume) are used. For some investigations, live trapping is clearly superior to kill trapping because the dynamic features of movement, lifespan, territory shifts, changes in body mass, and many others can be studied only by repeatedly observing an animal. Several papers in this volume report some of these details from mark-and-release studies. However, because many shrews do not survive well in live traps, it is difficult to secure long-term trapping records of individual shrews. Some shrews are so tiny that trap sensitivity can be a compounding problem. In addition, trapping throughout the winter poses further problems. Hawes’ (1977) year-round study of two Sorex species in Canada remains a model for others to emulate. Although radiotelemetry continues to hold promise in the study of shrews, the reductions in the size of radiotransmitters of the 1970s were not continued in the 1980s, consequently transmitters of the present generation are still too large for use with most species of shrews. Other electronic methods, such as tiny implants with holographic bar codes, are needed for the study of free-ranging shrews in the field or in seminatural enclosures. In sum, substantial challenges remain to secure shrews for study. Nevertheless, in many parts of the world great advances in understanding can be achieved by using the present methods. Origin and Paleohistory As with mammals in general, living shrews are a subset of past experiments in soricid evolution. Although shrews date from the mid-Tertiary, neither their small size nor their 455 456 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 distribution has changed much during their evolutionary history, although soricines apparently moved westward from Asia and later diversified in Europe. Soricines still dominate in the Holarctic, whereas crocidurines generally have tropical distributions, with relatively few species in temperate regions. Although many details are known about the fossil history of shrews in some regions, little or nothing is known elsewhere. With the increasing knowledge of the differences in physiologies and environmental tolerance between soricines and crocidurines, shrews are very important in understanding past environments. Systematics and Evolution The systematics and evolutionary relationships of many shrews remain uncertain. Because the cranial features important for identification have little phylogenetic value in some groups of shrews, it was not surprising to learn that cranial characters alone sometimes are not useful in phylogenetic studies. Some shrews show promise for yielding insight into how chromosomal evolution occurs, whether at the micro- or regional scale, because of their great variation in diploid numbers and chromosomal banding patterns. Furthermore, island populations of shrews are models to study the rate of evolutionary divergence between island and mainland populations, and how small mammals traverse water barriers and colonize offshore islands. The diversity, abundance, small size, and variability of shrews should contribute to their increased study and to substantial advances by systematic and evolutionary biologists during the coming decade. Growth and Development Because only a few species of shrews have been bred and raised successfully in the laboratory, the opportunities to study embryos of known age to determine developmental rates and to evaluate developmental processes have been limited. Soricine shrews have among the shortest gestation times of placental mammals. Young are bom at an early stage of development (almost like that of newborn marsupials), and have a relatively long period of postnatal development. Whether altricial young affirms the primitive evolutionary status of shrews or represents a derived feature, such as in passerine birds and many rodents, is moot. Extremely small maternal body size and large litter size may be part of the explanation for the tiny size of altricial newborn shrews, but the high metabolic rates of shrews should hasten development, both before and after parturition (McNab, 1980). Of course, the efficiency of transfer in the placenta, enzyme kinetics, and the quality of milk are other factors that could account for the seemingly slow rate of development in at least some soricine shrews. For whatever reasons, there are notable differences in the developmental rates of shrews and mice (in which development has been thoroughly studied). Postnatal development is also puzzling because shrews apparently differ from most small mammals in that young stay in the nest until virtually fully grown. By contrast, the young of most small mammals become semi -independent of the nest (and often are weaned) when only half or two-thirds grown. The most plausible explanation for the failure to catch juvenile or subadult shrews has been that young shrews are not exposed to traps because they remain nestbound during the 4-5 weeks of postnatal development. This is merely an assumption because almost nothing is known about rates or stages of postnatal growth in shrews, and even less about the behaviors associated with this period. Many such details could be learned in the laboratory using electronic devices and video cameras. Studies of development could also be important in understanding the evolutionary relationships of metatherian and eutherian mammals. Anatomy and Morphology Studies of shrew anatomy are potentially rewarding for several reasons. The shrew brain differs in appearance and proportions from the brains of most mammals, especially in the presence of exceptionally large olfactory bulbs. The brain of some soricid shrews physically shrinks more than the body during the winter months, a surprising revelation reported by Dehnel (1952). Dehnel’s phenomenon, first reported for Sorex araneus in Poland and confirmed in other species of Sorex, remains in need of a plausible explanation. Other unusual anatomical modifications include the teeth, which in some soricines are highly modified to include red enamel (produced by iron pigments) and large procumbent incisors, which sometimes possess tines. In contrast, crocidurines are called “white- toothed shrews.” The cheek teeth of shrews also are modified (“unicuspids”) compared to those of most mammals, making the entire dentition, the jaw musculature, and masticatory mechanics a challenge for functional morphologists to study and understand. In some soricines, the hairs change in size and shape between seasons, as do the relative proportions of white and brown adipose tissues. Although the eye of shrews has the basic eutherian components, the eyeball is small and many features are reduced. The ear likewise has the basic eutherian design but the small size of the head limits the auditory discrimation capacity of shrews. Although shrews were reported to navigate by echolocation more than 25 years ago, the link between hearing and behavior remains to be clarified. Much could be achieved by research on this topic. The olfactory system is well- developed and smell may surpass hearing as the most important sense in shrews. Species of Blarina, Neomys, and the solenodons have modified submaxillary (salivary) glands which produce toxins; these shrews and the duck-billed platypus, males of which have a spur associated with poison glands on the hind leg, are the only venomous mammals. The nature of the salivary toxin, its adaptive value, and whether the poison enables shrews to overpower and kill prey larger than themselves or to immobilize prey for later consumption are important issues in need of study. Finally, histological studies of gonads could be important in understanding the annual cycle of breeding in shrews. Soricine shrews are believed to be among the most reliable seasonal breeders of all small mammals, but the details of the annual cycle of the testis and ovary remain largely unknown. 1994 ROSE— Summary 457 Reproduction and Life-History Traits Crocidurines differ from soricines in their life-history traits by being physically larger and having fewer and larger neonates after a longer gestation period. Weaning occurs sooner in Crocidura than in Sorex, and their young are weaned at a smaller mass measured as a percentage of adult body mass. If grams of young produced across pregnancy and lactation are considered in relation to maternal mass, Sorex allocates more than twice as much energy to reproduction as does Crocidura. The average Sorex female (6.5 g) produces 34.2 g of weaned young, or a mass of young equal to 526% of maternal mass, in 45.1 days. This value conforms to Pearson’s (1944) observation that an 1 1-g Blarina brevicauda female supports a litter mass of 55 g near the end of lactation. By contrast, the average Crocidura female produces weaned young equal to 255% of maternal mass over 49. 1 days, and for Suncus murinus, the largest crocidurine shrew, the value is 181 % over 48. 1 days. Expressed as a percentage of maternal mass per day, the litter mass in these three genera accrued ten times faster during lactation than during pregnancy. For example, embryo mass of Sorex grew at a rate of 2.34% of maternal mass per day during pregnancy, but litter mass of neonates grew at a rate of 22.56 % of maternal mass per day. For Crocidura these values are 1.06% per day during pregnancy and 11.19 % per day during lactation, and for Suncus murinus the values are 0.64% and 6.26% per day, respectively. Thus, although Sorex females seemingly allocate twice as much energy per gram of maternal mass to the production of a weaned litter compared to Crocidura or to Suncus murinus, the three genera are similar in the greatly increased cost of lactation over pregnancy. I agree with Bronson (1989) that the energetic costs of lactation in shrews can be enormous. Physiology and Parasites/Disease As the smallest mammals with the highest metabolic rates, shrews therefore have the highest energy demands on a mass- specific basis. Yet many soricine shrews live in extremely cold environments where they apparently must pay huge energy costs to survive. Because their small size precludes storing much fat or gaining much insulation value from winter pelage, shrews in winter would seem to need plentiful, energy-rich prey. Alternatively, they would need to modify dietary selection, build highly insulative subterranean nests, change solitary behavior to permit communal nesting, hibernate or enter torpor, or employ combinations of these to survive in cold environments. Soricine shrews probably avoid extremely low air temperatures by living underground and in the subnivean space between the ground and snow cover; nevertheless, they must confront temperatures near 0°C for many months of the winter. Although some soricines accumulate brown fat in autumn which is mobilized to produce heat by nonshivering thermogenesis as needed in the coldest months (Merritt, 1986), other strategies of shrews to survive winter conditions are poorly understood. Some shrews eat more vegetable matter and a few may aggregate during the winter months. Surprisingly Crocidura (not Sorex) has a relatively low body temperature and metabolic rate, and the ability to enter torpor, features that seemingly would be more adaptive in arctic rather than in tropical environments. Much can be learned in the coming years by relating physiology to ecology, such as by electronic monitoring for location and movements and changes in body temperature and heart rate. Dietary studies are badly needed, too, particularly throughout the year at locations with cold and snowy winters. Such studies could provide information on where shrews obtain the calories to maintain high body temperatures during cold months. Shrews have not been much studied from the standpoint of diseases or parasites, in part because shrews generally are not economic pests although they sometimes are serious seed predators. Furthermore, shrews do not live as commensals to humans, thereby minimizing their potential role in disease transmission. However, a knowledge of the histopathological changes in populations may be important in understanding the dramatic year-to-year density changes that typify some populations of shrews. Populations and Communities Several papers in this volume substantially advance understanding of the ecology of soricine shrew populations and communities. Population density often fluctuates greatly from year to year in some species, but a long-term study of a Blarina population revealed mostly low-amplitude annual cycles and no relationship to the high-amplitude multiarmual cycles of two syntopic microtine rodents. The causes of year-to-year differences in shrew population density are unknown but may be influenced by fluctuating amount and availability of food or by predator densities. Shrews apparently move modest distances, and the size and stability of territories are related to age and breeding activity. Shrews have fairly predictable breeding seasons, starting later in northern than in more southerly locations. Mortality rates are fairly constant throughout the year in some populations; mortality from autumn to spring may be as high as 75%, and the chance of surviving a second winter nil. Whether these patterns found in soricines also apply to crocidurines remains to be learned. In the Holarctic biogeographic region, small mammal communities in most mesic, midlatitude sites are characterized by the presence of two or more soricine shrews. Shrew species tend to become proportionately more abundant in small mammal communities at higher latitudes. In North America the greatest diversities of shrews are found in the Oregon-Washington- British Columbia and northern Appalachian Mountain regions, whereas the greatest number of coexisting soricine shrew species apparently is in Siberia, where as many as nine species may be found in an area. Little is known about the diversity and role of crocidurine shrews in small mammal communities of Africa and Asia, but some localities surely have several species. Soricine shrews eat a variety of foods; however, each species usually specializes on a few species of common and abundant invertebrate prey. Prey size is not always related to the size of the shrew. If food of a particular size is abundant, the shrew species that optimally selects that food may become dominant in the shrew community. It is postulated in this volume that 458 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 18 large shrew species, which have twice the per capita food requirements of small shrews, exploit more productive habitats and have more stable local populations. By contrast, small species may exploit smaller and less productive habitat patches, and have transitory populations. It is puzzling that the smallest soricine species seem to be the rarest and often have patchy distributions. Their low body masses and high metabolic rates, which result in short starvation times, tend to produce high extinction rates for small local populations. Their continued presence can be sustained, it seems, only by great fecundity and equally great dispersal capability. Sorex minutissimus in Eurasia and Sorex hoyi in North America, the smallest shrews, also seem to be the rarest. Perhaps this is due to a sampling problem, because in theory one of the best defenses against extinction of populations is large population size. Most studies of shrews in North America, and to an extent elsewhere, have been conducted as adjuncts to studies of other species of small mammals, frequently microtine rodents. Although such studies are valuable, they rarely include experimentation or hypothesis testing for shrews, and mostly describe the dynamics and components of fitness of the population or composition of the small mammal community. Nevertheless, such information is badly needed for many shrew species, especially crocidurines, as a basis for formulating hypotheses and experiments for future studies. Relatively few studies have used live-trapping methods to mark and release shrews, while documenting the events of their lives. Fewer still have attempted to follow the daily lives of individual shrews via radiotelemetry or other electronic methods to evaluate physiological and behavioral changes within and among days and seasons. The studies reported here give a foretaste of what future investigations might reveal. In conclusion, much remains to be learned about these fascinating small mammals, the smallest body masses of which are less than the theoretically smallest mass that can sustain metabolic rates (they haven’t read the manual!), and which somehow survive the world’s harshest winter environments despite the high energy costs associated with rapid heat loss, apparently low availability and quality of food, and inability to hibernate or enter torpor or even to form communal groups. Truly, the study of shrews will continue to be well repaid and to produce more results that will amaze us or continue to defy explanation. Acknowledgments Special thanks to G. L. Kirkland, Jr., for comments that substantially improved this manuscript; any errors, of course, remain mine. Literature Cited Bronson, F. H. 1989. Mammalian Reproductive Biology. University of Chicago Press, Chicago. Dehnel, a. 1952. The biology of breeding of the common shrew in laboratory conditions. Annales of University M . Curie-Sklodowska, Section C, 6, 11:359-376. Hawes, M. L. 1977. Home range, territoriality, and ecological separation in sympatric shrews, Sorex vagrans and Sorex obscurus. Journal of Mammalogy, 58:354-367. Merritt, J. F. 1986. Winter survival adaptations of the short-tailed shrew (Blarina brevicaiida) in an Appalachian montane forest. Journal of Mammalogy, 67:450-464. McNab, B. K. 1980. Food habits, energetics, and the population biology of mammals. American Naturalist, 116:106-124. Pearson, O. P. 1944. Reproduction in the shrew (Blarina brevicauda Say). American Journal of Anatomy, 75:39-93. 1^'' ■V \ }■ I c5S" t) -31 I AN IJH57 Tujj' 3 9088 00966 524 THE CARNEGIE MUSEUM OF NATURAL HISTORY