TEXAS THE JOURNAL OF SCIENCE GENERAL INFORMATION MEMBERSHIP.— Any person or members of any group engaged in scientific work or interested in the promotion of science are eligible for membership in The Texas Academy of Science. Dues for regular members are $30.00 annually; supporting members, $60.00; sustaining members, $100.00; patron members, $150.00; associate (student) members, $15.00; family members, $35.00; affiliate members, $5.(X); emeritus members, $10.00; life members, 20 times annual dues; corporate members, $250.00 annually; corporate life members, $2000.00 in one payment. Library subscription rate is $50.00 annually. The Texas Journal of Science is a quarterly publication of The Texas Academy of Science and is sent to most members and all subscribers. Payment of dues, changes of address and inquiries regarding missing or back issues should be sent to Dr. Brad Henry, Department of Biology, The University of Texas-Pan American, Edinburg, Texas 78539. AFFILIATED ORGANIZATIONS American Association for the Advancement of Science, Texas Council of Elementary Science Texas Section, American Association of Physics Teachers Texas Section, Mathematical Association of America Texas Section, National Association of Geology Teachers Texas Society of Mammalogists The Texas Journal of Science (ISSN 0040-4403) is published quarterly at Lubbock, Texas U.S. A. Second class postage paid at Post Office, Lubbock, Texas 79402. Postmaster: Send address changes, and returned copies to The Texas Journal of Science, Box 43151, Texas Tech University, Lubbock, Texas 79409-3151, U.S. A. THE TEXAS JOURNAL OF SCIENCE Volume 47, No. 1 February 1995 CONTENTS Noteworthy records of mammals from the Edwards Plateau of central Texas. By Jim R. Goetze, Franklin D. Yancey, II, Clyde Jones and Burhan M. Gharaibeh . 3 A fossil specimen of the long-nosed snake Rhinoceilus from the pliocene of southern New Mexico. By Spencer G. Lucas, Andrew B. Heckert and Paul L. Sealey . 9 Saturn in late 1993. By Richard W. Schmude, Jr. and Danny Bruton . 13 A comparison of six test statistics for detecting multivariate non-normality which utilize the multivariate squared-radii statistic. By Dean M. Young, Samuel L. Seaman and John W. Seaman, Jr. . 21 Vomerolfactory exploration of novel environments by the parthenogenetic whiptail lizard Cnemidophorus laredoensis (Sauria: Teiidae). By Loraine R. Rybiski and Mark A. Paulissen . 39 Habitat use of introduced and native anoles (Iguanidae: Anolis) along the northern coast of Jamaica. By Allan J. Landwer, Gary W. Ferguson, Rick Herber and Mark Brower . 45 Kallikrein-like enzyme from the venom of Crotalus basiliscus basiliscus (Serpentes: Crotalidae). By Robert D. Gaffin, Jon B. Scales and Rodney L. Cate . 53 Abundance and diversity of aquatic birds on two south Texas oxbow lakes. By Diane Teter and David L. McNeely . 62 General Notes First report of the Acanthocephalan Macracanthorhynchus ingens from the domestic dog Canis familiaris in Kansas. By Omar M. Amin, Charles L. Kramer and Steve J. Upton . 69 Placobdella parasitica (Rhynchobdellida; Glossiphoniidae) from the eastern river cooter (Chelonia: Emydidae) in Oklahoma. By William E. Moser . 71 Book Review - Birds and other wildlife of south central Texas: A handbook. By Terry C. Maxwell, Robert C. Dowler and Raymond Stone, Jr . 75 Instructions to Authors . 76 THE TEXAS JOURNAL OF SCIENCE EDITORIAL STAFF Manuscript Editor: Jack D. McCullough, Stephen F. Austin State University Managing Editors: Michael J. Carlo, Angelo State University Ned E. Strenth, Angelo State University Associate Editor for Botany: Robert 1. Lonard, The University of Texas — Pan American Associate Editor for Chemistry: John R. Villarreal, The University of Texas — Pan American Associate Editor for Geology: M. John Kocurko, Midwestern State University Associate Editor for Mathematics and Statistics: E. Donice McCune, Stephen F. Austin State University Associate Editor for Physics: Charles W. Myles, Texas Tech University Manuscripts intended for publication in the Journal should be submitted in TRIPLICATE to: Jack D. McCullough TAS Manuscript Editor Department of Biology - Box 13003 Stephen F Austin State University Nacogdoches, Texas 75962 Scholarly papers in any field of science, technology, or science education will be considered for publication in The Texas Journal of Science. Instructions to authors are published one or more times each year in the Journal on a space-available basis, and also are available from the Manuscript Editor at the above address. The Texas Journal of Science is published quarterly in February, May, August, and November for $30 per year (regular membership) by The Texas Academy of Science. Second-class postage rates (ISSN 0040-4403) paid at Lubbock, Texas. Postmaster: Send address changes, and returned copies to The Texas Journal of Science, Box 43151, Texas Tech University, Lubbock, Texas 79409-3151, U.S.A. TEXAS J. SCI. 47(l):3-8 FEBRUARY, 1995 NOTEWORTHY RECORDS OF MAMMALS FROM THE EDWARDS PLATEAU OF CENTRAL TEXAS Jim R. Goetze, Franklin D. Yancey, 11, Clyde Jones and Burhan M. Gharaibeh The Museum and Department of Biological Sciences, Texas Tech University, Lubbock, Texas 79409-3191 Abstract.— Distributional notes based upon recent field collections are reported for nine species of small mammals from the Edwards Plateau of central Texas. These include one species of bat (Nycticeius), one armadillo (Dasypus), one kangaroo rat (Dipodomys), four mice {Reithrodontomys , Peromyscus and Baiomys), one woodrat {Neotoma) and one skunk (Conepatus). Field studies conducted from 1990 through 1993 have provided additional distributional data on several species of small mammals from the Edwards Plateau of central Texas. Additional noteworthy specimens from this same geographical area were also noted to be present in the Collection of Recent Mammals of The Museum, Texas Tech University. The following species accounts are the result of research efforts conducted during the course of this study. Voucher specimens are deposited with the holdings of The Museum at Texas Tech University (TTU). Nycticeius humeralis humeralis (Rafmesque) (Evening Bat) Distributional notes— Tht evening bat is found throughout the eastern half of Texas and reaches its western distributional limits on the Edwards Plateau (Schmidly 1991). While specimens are reported from several counties on the Edwards Plateau (Manning et al. 1987; Schmidly 1991; Dowler et al. 1992), overall representation from this geographical area of Texas is poorly known. This report represents, the first specimen record of the evening bat from Blanco County. Material examined,— \ mi N, 10 mi E of Johnson City, Blanco County, Texas, 13 May 1990, three specimens (TTU 57802-57804) netted over small stream. A single specimen (TTU 57804) was gravid with two fetuses measuring 20 mm in crown-rump length. 4 THE TEXAS JOURNAL OF SCIENCE-VOL. 47, NO. 1, 1995 Dasypus novemcinctus mexicanus Peters (Nine-banded Armadillo) Distributional notes.— Although the armadillo has been reported to range throughout eastern and southern Texas (Schmidly 1983) and is currently expanding its range onto the Llano Estacado (Jones et al. 1993), records of its occurrence from the northern Edwards Plateau are rare. The recent collection of a single specimen from near Buffalo Gap represents the first specimen record of this species from Taylor County. Material examined.— 2 mi S, 8 mi W of Buffalo Gap, Taylor County, Texas, 27 June 1993, one specimen (TTU 63410). Habitat.— Tht collection locality is in rough, broken lands of the Callahan Divide, which is a northern extension of the Edwards Plateau. Dipodomys merriami ambiguus Merriam (Merriam’s Kangaroo Rat) Distributional notes.— ^\h\Q Merriam’s kangaroo rat has been reported from several Texas counties east of the Pecos River (Jones & Jones 1992), the distribution of this species is poorly known for this region. It has been reported from adjacent Midland, Reagan (Jones & Jones 1992), and Crockett Counties (Hollander et al. 1987). This report represents the first record of this species from Upton County. Material examined. — 5 mi S, 4 mi E of Crane, Upton County, Texas, 19 March 1992, three specimens (TTU 63074-63076). Habitat.— Tht collection locality is adjacent to a fencerow which borders a mesquite pasture. Vegetation in this area consisted of mesquite (Prosopis glandulosa), creosote (Larrea tridentata), grama grass (Bouteloua sp.) and broomweed (Guitierrezia sp.). Soils at the locality were eroded from a nearby escarpment and were sandy with small gravel. Reithrodontomys fulvescens laceyi Allen (Fulvous Harvest Mouse) Distributional Schmidly (1983) reported this species from Bosque and Hays Counties of central Texas. This report represents additional county records for the Callahan Divide and Edwards Plateau. GOETZE ET AL. 5 Material examined.—^ mi S, 4.5 mi W of Clyde, Callahan County, Texas, 15 July 1991, one specimen (TTU 59852). 4.5 mi N of Oplin, Callahan County, Texas, 6 June 1993, one specimen (TTU 63418). 3.5 mi E of San Saba, San Saba County, Texas, 16 July 1992, one specimen (TTU 63419). Reithrodontomys megalotis megalotis (Baird) (Western Harvest Mouse) Distributional notes. — The western harvest mouse ranges from the Llano Estacado south into the Trans-Pecos area of west Texas. This species was previously believed to be restricted to the Llano Estacado at the extreme southeastern part of its range (Davis 1974). This is the first report of the western harvest mouse from Glasscock County. Choate et al. (1992) mentioned the possible occurrence of this species in this county at the southern border of the Kansan Biotic Province. The collection locality is, however, on the Edwards Plateau immediately east of the Llano Estacado. Material examined.— 5 mi N, 7 mi W of Garden City, Glasscock County, Texas, 1 April 1990, five specimens (TTU 59103-59107). Peromyscus attwateri (Allen) (Texas Mouse) Distributional notes .—Tht Texas mouse is distributed from central Texas north to the Red River and west to the escarpment of the Llano Estacado (Davis 1974). This report represents a new county record for this species within its known range. The collection site is within the Callahan Divide. Material examined.— \ mi S, 9 mi W of Buffalo Gap, Taylor County, Texas, 27 June 1993, one specimen (TTU 63425). Habitat. — The collection locality was a rocky slope dominated by juniper (Juniperus ashei) and sideoats grama {Bouteloua curtipendula) . Baiomys taylori taylori (Thomas) (Pygmy Mouse) Distributional notes.— TYvt geographic range of the pygmy mouse in Texas has been well documented (Davis 1974; Jones et al. 1987; Choate et al. 1990; Jones et al. 1993). This species is distributed over the eastern two- thirds of the state and onto the Llano Estacado and adjacent 6 THE TEXAS JOURNAL OF SCIENCE-VOL. 47, NO. 1, 1995 areas. Several specimens collected during the course of this study represent county records from within its range. Material examined.— 2 mi W of Oplin, 15 July 1991, five specimens (TTU 59834-59838); 1.5 mi N of Oplin, 6 June 1993, one specimen (TTU 63442); 4.5 mi N of Oplin, Callahan County, Texas, 6 June 1993, one specimen (TTU 63443). 4 mi W of Big Lake, Reagan County, Texas, 22 July 1991, two specimens (TTU 59841-59842). 3.5 mi E of San Saba, San Saba County, Texas, 16 July 1992, two specimens (TTU 63447-63448). 5 mi S, 4 mi E of Crane, Upton County, Texas, 19 March 1993, one specimen (TTU 63077). 4 mi S, 5 mi W of Eden, Concho County, Texas, 19 June 1993, one specimen (TTU 63444). 1 mi N, 4 mi W of San Marcos, Hays County, Texas, 16 July 1993, two specimens (TTU 63445-63446). Habitat.— K\\ specimens were collected from habitats characterized by dense vegetative cover. Neotoma micropus micropus Baird (Southern Plains Woodrat) Distributional notes southern plains woodrat ranges throughout the western two- thirds of Texas (Jones & Jones 1992). This record represents the first reported occurrence of this species from Callahan County. The collection locality lies within the Callahan Divide of the Edwards Plateau. The species has been reported from adjacent Eastland and Taylor Counties. Material examined.— mi S, 4.5 mi W of Clyde, Callahan County, Texas, 15 July 1991, one specimen (TTU 59844). Habitat. — The specimen was collected from a rocky slope dominated by juniper and grama grasses. Conepatus mesoleucus meamsi Merriam (Hog-nosed Skunk) Distributional notes.— Tht currently known range of the hog-nosed skunk in Texas was reported by Manning et al. (1986). While this report noted a sight record of C. mesoleucus meamsi from Nolan County, no voucher specimens were collected. This current study reports the first voucher of the hog-nosed skunk from Nolan County as well as the first specimen record of this species from Reagan County on the western Edwards Plateau. GOETZE ET AL. 7 Material examined. —I mi S of Big Lake, Reagan County, Texas, 1 September 1991, one specimen (TTU 60063). 6 mi N, 7 mi E of Blackwell, Nolan County, 17 August 1993, one female specimen (TTU 63450), partial salvaged skull, no reproductive activity noted, external measurements: total length, 660 mm; tail length, 257 mm; hind foot, 70 mm; length of ear from notch, 28 mm. Acknowledgments We are grateful to Richard W. Manning and an anonymous reviewer for their critical reviews of this manuscript. Literature Cited Choate, L. L., J. K. Jones, Jr., R. W. Manning, & C. Jones. 1990. Westward ho: Continued dispersal of the pygmy mouse, Baiomys taylori, on the Llano Estacado and in adjacent areas of Texas. Occas. Papers Mus., Texas Tech Univ., 134:1-8. _ , R. W. Manning, J. K. Jones, Jr., C. Jones, & S. E. Henke. Mammals from the southern border of the Kansan Biotic Province in western Texas. Occas. Papers Mus., Texas Tech Univ., 152:1-34. Davis, W. B. 1974. The mammals of Texas. Bull. Texas Parks Wildlife Dept., 41:1-294. Dowler, R. C., T. C. Maxwell, & D. S. Marsh. 1992. Noteworthy records of bats from Texas. Texas J. Sci., 44(1): 121-123. Hollander, R. R., C. Jones, R. W. Manning, & J. K. Jones, Jr. 1987. Distributional notes on some mammals from the Edwards Plateau and adjacent areas of south-central Texas. Occas. Papers Mus., Texas Tech Univ., 110:1-10. Jones, C., R. D. Sutkus, & M. A. Bogan. 1987. Notes on some mammals of north-central Texas. Occas. Papers Mus., Texas Tech Univ., 115:1-21. Jones, J. K., Jr., & C. Jones. 1992. Revised checklist of Recent land mammals of Texas, with annotations. Texas J. Sci., 44(l):53-74. _ , R. W. Manning, F. D. Yancey, II, & C. Jones. 1993. Records of five species of small mammals from western Texas. Texas J. Sci. , 45(1): 104-105. Manning, R. W., J. K. Jones, Jr., & R. R. Hollander. 1986. Northern limits of distribution of the hog- nosed skunk, Conepatus mesoleucus, in Texas. Texas J. Sci., 3 8(3): 289-291.^ 8 THE TEXAS JOURNAL OF SCIENCE— VOL. 47, NO. 1, 1995 _ , _ , _ , & C. Jones. 1987. Notes on distribution and natural history of some bats on the Edwards Plateau and in adjacent areas of Texas. Texas J. Sci., 39(3): 279-285. Schmidly, D. J. 1983. Texas mammals east of the Balcones Fault zone. Texas A&M Univ. Press, College Station, xviii + 400 pp. _ . 1991. The bats of Texas. Texas A&M Univ. Press, College Station, xv 4- 188 pp. TEXAS J. SCI. 47(1):9-12 FEBRUARY, 1995 A FOSSIL SPECIMEN OF THE LONG-NOSED SNAKE RHINOCHEILUS FROM THE PLIOCENE OF SOUTHERN NEW MEXICO Spencer G. Lucas, Andrew B. Heckert and Paul L. Sealey New Mexico Museum of Natural History and Science, 1801 Mountain Road N W. , Albuquerque, New Mexico 87104 Abstract.— This study reports a Blancan (late Pliocene) fossil snake from southern New Mexico. This is the first documentation of a late Cenozoic fossil snake from New Mexico and the first state fossil record of the genus Rhinocheilus . It also extends the western range of the known fossil distribution of Rhinocheilus and appears to represent the currently known most complete fossil specimen of the long-nosed snake. The Miocene-Pleistocene Santa Fe Group, which is the sedimentary fill of the Rio Grande rift basins of central New Mexico, is known to contain extensive assemblages of fossil mammals (Tedford 1981). Much less collected and studied are the fossil herpetofaunas of these strata (Kues 1993). The fossil specimen detailed in this study was collected at tlie New Mexico Museum of Natural History and Science (NMMNH) locality 02854 in the NEl/4 NEl/4 SWl/4 NW 1/4 sec. 23, T19S, R4W, Doha Ana County, New Mexico. This stratum is in the lower part of the Camp Rice Formation of the Santa Fe Group. Fossil mammals associated with this stratum indicate a Blancan age (Tedford 1981). The fossil specimen (NMMNH P-18816) represents a single individual, and consists of a braincase, 9 cervical vertebrae, 44 trunk vertebrae, 27 ribs (3 complete) and various fragments (Fig. 1). These bones were found slightly disarticulated and in close association. The trunk vertebrae lack hypapophyses, have relatively long and thin neural spines, are much longer than they are wide, and have thin and distinct hemal keels and vaulted neural arches. Based upon the work by Holman (1979), the above characteristics identify this fossil specimen as a colubrine. The trunk vertebrae exhibit features characteristic of Rhinocheilus lecontei as enumerated by Hill (1971), namely flat zygosphenes in anterior view; obovate to oval prezygapophysial faces; thick neural spines (within the colubrines) that are also flat dorsally, overhanging centra posteriorly, with indented anterior and posterior edges; short centra, no epizygapophysial spines; depressed neural arches; round to 10 THE TEXAS JOURNAL OF SCIENCE-VOL. 47, NO. 1, 1995 Figure 1. Selected bones of NMMNH P-18816, partial skeleton of Rhinocheilus lecontei from the Pliocene of southern New Mexico. A-C; dorsal (A), ventral (B) and lateral (C) views of braincase. D,F; lateral (D) and dorsal (F) views of trunk vertebra, E; ventral view of trunk vertebra, G-H; dorsal (G) and right lateral (H) views of two articulated trunk vertebrae. slightly compressed cotyla; postzygapophysial faces that are obovate to orbicular, strongly developed hemal keels and subcentral ridges, and swollen, very flat accessory processes. Holman (1979) recognized Rhinocheilus on the basis of trunk vertebrae that are almost equally long as wide, have strong subcentral ridges, and very long and thick accessory processes. These characters are evident in the fossil material. For these reasons this fossil specimen is assigned to the genus Rhinocheilus. Hill (1971) and Holman (1979) noted that fossil specimens of Rhinocheilus from west Texas are indistinguishable from specimens of R. lecontei of Recent origin. A comparison of NMMNH P-18816 with a juvenile extant specimen of R. lecontei from the University of New LUCAS, HECKERT 8l SEALEY 11 Mexico Museum of Southwestern Biology (UNM 47182) revealed that, except for size differences resulting from their differing ontogenetic stages, the two specimens were identical. Comparison of NMMNH P-18816 with a Recent specimen of R. antoni (UNM 41567), revealed major differences in the two specimens. The fossil specimen is not assigned to R. antoni due to the fact that the trunk vertebrae of the fossil specimen possess neural spines that are not strongly indented anteriorly and posteriorly, as well as articular facets that are broad, flat, and round in dorsal view. Fossil specimens of Rhinocheilus lecontei have been reported from strata of Blancan age in Scurry County of west Texas (Rogers 1976) and from strata of Rancholabrean age in Kendall County of central Texas (Holman 1969a; 1969b; Hill 1971; Holman 1979). The present-day range of Rhinocheilus extends from California north to Nevada and south to Texas and Chihuahua (Wright & Wright 1957). This report of the New Mexican fossil of Rhinocheilus (1) represents the first record of a fossil snake from the late Cenozoic of New Mexico as well as the first fossil Rhinocheilus from the state; (2) extends the known range of fossil Rhinocheilus westward from Texas into central New Mexico; and (3) provides a much more complete fossil specimen of Rhinocheilus than those reported by Holman (1969a), Hill (1971) and Rogers (1976). Acknowledgments We wish to thank Bill Lang and Pete Reser for expert preparation of the fossil specimen and the Herpetology Division of the Museum of Southwestern Biology with the University of New Mexico for access to skeletons of extant Rhinocheilus , and two anonymous reviewers for their helpful comments. Literature Cited Hill, W. H. 1971. Pleistocene snakes from a cave in Kendall County, Texas. Texas J. Sci., 22:209-216. Holman, J. A. 1969a. The Pleistocene herpetofauna from Kendall County, Texas. Publ. Mus. Mich. St. Univ., 4:163-192. _ . 1969b. Herpetofauna of the Pleistocene Slaton local fauna of Texas. Southwestern Nat. , 14:203-212. _ 1979. A review of North American Tertiary snakes. Publ. Mus. Mich. St. Univ. Paleont. Ser., 1:203-260. 12 THE TEXAS JOURNAL OF SCIENCE- VOL. 47, NO. 1, 1995 Kues, B. S. 1993. Bibliographic catalogue of New Mexico vertebrate fossils. New Mexico Mus. of Nat. His. and Sci., Bull., 2:199-279. Rogers, K. L. 1976. Herpetofauna of the Beck Ranch local fauna (upper Pliocene: Blancan) of Texas. Publ. Mus. Mich. St. Univ. Paleont. Ser., 1(5): 167-200. Tedford, R. H. 1981. Mammalian biochronology of the late Cenozoic basins of New Mexico. Geol. Soc. Amer. Bull., 92:1008-1022. Wright, A. H., & A. A. Wright 1957. Handbook of Snakes of the United States and Canada. Cornell Univ. Press, Ithaca, New York, 1105 pp. TEXAS J. SCI. 47(1): 13-20 FEBRUARY, 1995 SATURN IN LATE 1993 Richard W. Schmude, Jr. and Danny Bruton Division of Natural Sciences and Nursing, Gordon College, Barnesville, Georgia 30204 and Department of Physics, Texas A&M University, College Station, Texas 77843 Abstract.— A combination of photometric and visual/photographic studies of the planet Saturn were conducted during September and October of 1993. Solar phase coefficients in mag. /deg., of Saturn (globe+rings) measured through Johnson B, V, R and I filters were 0.055 ±0.(X)7; 0.054±0.005; 0.047 ±0.010 and 0.053 ±0.(X)7 respectively. The normalized magnitude of Saturn (excluding the rings) was V(1 ,0) = -8.84±0.04. Visual observations and photographs indicate that Saturn has returned to an appearance similar to that which was observed in 1989 prior to the occurrence of the large white storm of 1990. Saturn attained opposition on 19 August 1993 and was at a declination of -15°. The ring plane was tilted at an angle of 12-13° with respect to the Earth in late 1993; this tilt is considerably less than 17° during 1992 and 20° during 1991. The changing tilt produces a change in the brightness of Saturn which can be photometrically measured. The changing aspect of Saturn’s rings may also provide additional information relative to the colors of Saturn and its rings. One of the primary objectives of observing Saturn during 1993 was to detect any changes in color, appearance or brightness which may have resulted due to the large white storm which occurred during 1990 (Barnet et al. 1992; Beebe et al. 1992; Benton 1992; Heath & McKim 1992). This study provides data on the solar phase coefficients of Saturn (globe + rings) through the B, V, R and I filters along with normalized magnitudes extrapolated to a phase angle of zero degrees. The position, color and intensity of various atmospheric features of Saturn during late 1993 is also documented. Photometric Methods The 14 inch (36 cm) f/11 Schmidt-Cassegrain telescope at Texas A&M University Observatory was used during all measurements and observations. An Optec SSP-3 solid-state photometer was used with Johnson U, B, V, R and I filters during all photometric measurements. For additional information relative to the photometer and filters, refer to Optec (1988) and Schmude (1992). 14 THE TEXAS JOURNAL OF SCIENCE— VOL. 47, NO. 1, 1995 The comparison star used during all measurements was 7-Cap. This star has respective U, B, V, R and I magnitueds of +4.20, +3.99, + 3.67, +3.44 and +3.31 (Iriarte et al. 1965), and has coordinates (2000.0) of R.A.= 2P 40.09"^ and Dec. - -16^ 39.75*" (Hirshfeld & Sinnott 1985). The measured magnitudes of Saturn along with other relevant information is listed in Table 1 . All magnitudes have been corrected for atmospheric extinction. Both the solar phase angle of Saturn, a and the tilt of the ring plane, B, were changing during the time period covered in Table 1. The brightness of Saturn depends on a and B according to: V(1,0) - V^eas - 51og[gh] -c,Q? + 2.60sin[B] - 1.25sm"[B] where V(1 ,0) is the normalized magnitude of Saturn without rings, is the measured V-filter magnitude of Saturn (globe + rings), g and h are the Saturn- Earth and Saturn-Sun distances in astronomical units, c^ is the solar phase angle coefficient in the V filter, a is the solar phase angle of Saturn and B is the ring tilt. The "2.5 log k" term (where k is the fraction of the disc which is illuminated by the sun as seen from the Earth) is negligible for Saturn and is thus not included in the above equation. The primary goal of the photometric study was to measure the solar phase angle coefficient, c^, of Saturn (globe + rings). Therefore, any change in brightness created by a changing B must be eliminated. The magnitude change. Am, of Saturn as a function only of ring tilt is reported (Harris, 1961) to obey: Am - 2.60sin(B) - 1.25sin2(B) The value of Am varied from 0.511 up to 0.528 during the time period covered in Table 1; therefore, in order to correct for B (or in other words, hold B constant), a scaling factor of (0.522 - Am) was subtracted from all normalized V(l,a) values to obtain the V(l,c^)' values. The 0.522 is Am when B-13.0®; the average value of B in Table 1. The slope of the V(l,a)' versus a plot is 0.054 mag. /deg. The slope, c^, is the solar phase coefficient of Saturn (globe + rings) in the V-filter when B — 13°. A similar procedure was used in the V filter measurements made in 1991 and the corresponding solar phase coefficient is listed in Table 2. The difference between the 1991 and 1993 solar phase coefficients is SCHMUDE & BRUTON 15 Table 1. Summary of wideband photometric measurements made of Saturn in 1993. The solar phase angle is a and the ring tilt angle is 6. Date (U.T.) a B Filter Measured Magnitude X(l,a) Sept. 21.264 3.2" + 12.7° B + 1.54 -8.18 Sept. 21.266 3.2" + 12.7° V +0.53 -9.19 Sept. 23.178 3.4" + 12.8° B + 1.54 -8.18 Sept. 25.210 3.5" + 12.8" B + 1.54 -8.19 Sept. 25.212 3.5" + 12.8" V +0.56 -9.17 Sept. 25.215 3.5" + 12.8" R -0.05 -9.78 Sept. 25.219 3.5" + 12.8° I -0.10 -9.83 Sept. 26.095 3.6" + 12.8" U + 2.30 -7.43 Sept. 26.097 3.6" + 12.8" B + 1.57 -8.16 Sept. 26.099 3.6° + 12.8° V +0.58 -9.15 Sept. 26.102 3.6° + 12.8" R +0.01 -9.72 Sept. 26.106 3.6" + 12.8° I -0.09 -9.82 Sept. 30.160 4.0" + 12.9" B + 1.61 -8.13 Sept. 30.162 4.0° + 12.9° V +0.59 -9.15 Sept. 30.165 4.0" + 12.9" R +0.03 -9.71 Sept. 30.167 4.0" + 12.9° I -0.04 -9.78 Oct. 1.178 4.1" + 12.9° B + 1.60 -8.14 Oct. 1.180 4.1" + 12.9" V +0.60 -9.14 Oct. 1.181 4.1" + 12.9" R +0.05 -9.69 Oct. 1.182 4.1" + 12.9° I - 0.02 -9.76 Oct. 2.248 4.2" + 13.0" B + 1.57 -8.17 Oct. 2.250 4.2° + 13.0° V +0.61 -9.13 Oct. 2.252 4.2" + 13.0" R +0.03 -9.71 Oct. 2.255 4.2" + 13.0" I - 0.03 -9.77 Oct. 11.103 4.8" + 13.1" U +2.46 -7.31 Oct. 11.105 4.8" + 13.1" B + 1.66 -8.11 Oct. 11.106 4.8" + 13.1" V +0.67 -9.10 Oct. 11.108 4.8" + 13.1° R +0.09 -9.68 Oct. 11.111 4.8° + 13.1" I +0.01 -9.76 Oct. 11.134 4.8° + 13.1" U +2.47 -7.30 Oct. 11.136 4.8° + 13.1° B + 1.65 -8.12 Oct. 11.137 4.8" + 13.1° V +0.66 -9.11 Oct. 23.049 5.4" + 13.2" u +2.52 -7.29 Oct. 23.052 5.4" + 13.2" B + 1.75 -8.06 Oct. 23.053 5.4" + 13.2° V +0.74 -9.07 Oct. 23.055 5.4° + 13.2° R +0.14 -9.67 Oct. 23.057 5.4° + 13.2° I +0.06 -9.75 Oct. 26.042 5.5" + 13.2° B + 1.71 -8.11 Oct. 26.044 5.5° + 13.2° V +0.73 -9.09 Oct. 26.046 5.5" + 13.2° R +0.14 -9.68 Oct. 26.049 5.5" + 13.2° I +0.10 -9.72 Oct. 28.082 5.5° + 13.2° B + 1.76 -8.06 Oct. 28.083 5.5" + 13.2" V +0.76 -9.07 Oct. 28.083 5.5" + 13.2" R +0.18 -9.64 Oct. 28.087 5.5" + 13.2° I +0.12 -9.70 believed to be real and due to the change in ring tilt. Solar phase coefficients are reported in Harris (1961) for the years 1914-15, 1917, 1918 and 1920; these values are listed in Table 2 along with the ring tilt (Nautical Almanac 1911; 1915a; 1915b; 1917). It appears that the solar 16 THE TEXAS JOURNAL OF SCIENCE- VOL. 47, NO. 1, 1995 Table 2. Solar phase coefficients and normalized magnitudes measured for Saturn (globe + rings). Year B Qf-range Filter c (mag. /deg.) X(1,0)' 1993 + 13° 3.2°-5.5° B 0.055+0.007 -8.38 1993 + 13° 3.2°-5.5° V 0.054 +0.005 -9.37 1993 + 13° 3.2°-5.5° R 0.047+0.010 -9.92 1993 + 13° 3.2°-5.5° I 0.053+0.007 -10.01 1992 + 17° 5.2°-5.7° V 9.45^ 1991 +20° 0.9°-5.7° B 0.049 +0.010 ____ 1991 +20° 0.4°-5.7° V 0.038+0.007 -9.50 1920^ -6° — 0.049 ____ 1918^ -17° — 0.043 1917’^ H -22° — 0.033 1914-15 -27° — 0.033 — - Assuming a solar phase coefficient of 0.044 mag. /deg. ^ From (Harris 1961). phase coefficient increases with decreasing B (or increasing shadowing in the range 27°>B>13°. The increase in c^ with decreasing B is consistent with the behavior of the Moon and Mercury. Essentially, as the shadowing of the Moon (Harris 1961) and Mercury (Veverka 1988) increases, the value of c^ increases; likewise, as the shadowing of the rings increases, c^ increases. Determining the solar phase angle coefficients through the B, R, and I filters is more difficult because there is no equation available which relates Am to the ring tilt. For a close approximation, however, the same magnitude correction used for the V filter can be applied to the B, R and I filters. Such an approximation is reasonable since it is only the change in B, R and I magnitudes for ring tilt angles 12.7°- 13. 2° that is of interest. The resulting solar phase coefficients are listed in Table 2. The normalized magnitude of Saturn calculated using the 1993 solar phase coefficient and the above first equation is V(1,0) =-8.84+0. 03 which is close to the accepted value of -8.88 (Nautical Almanac 1992; Harris 1961). The color indexes of Saturn (globe + rings) in 1993 at B = 13° and a=0° are: B-V= +0.99±0.04; V-R= +0.55+0.04 and R-I= +0.09+0.04. The uncertainties are estimated based on a combination of possible errors arising from star magnitude, atmospheric extinction, the above second equation and in the V(1,0)' versus a plot. SCHMUDE & BRUTON 17 Table 3. Summary of Saturnicentric latitudes of various features on Saturn. Feature abbreviations follow those of Schmude (1990). Feature Reticle Photograph Video Visual Selected Value NPR® 56'’N 57°N - — - — 57°±4°N NTB^^ 44°N 38°N - — 35°N 39°±5°N neb" 26°N 27°N 21 °N 26°N 25°±2°N neb" 16°N 16°N 13°N 15°N \5° ±2°N - — - — - — 9°N 9°±4‘’N STB^ 34°S 36‘’S - — 35°S 35°±3°S Appearance of Saturn The objective of this visual/photographic study was to determine the positions/colors of Saturn’s atmospheric features. A combination of visual and photometric studies should yield a more accurate picture of the overall behavior of Saturn. Four different methods were used in measuring the Saturnicentric latitudes which are: visual, reticle, photograph and video measurements. Latitudes were computed from the formulae in Peek (1958) and the data in the Nautical Almanac (1992). The resulting latitudes are summarized in Table 3. In Table 3, NPR3= south edge of the north polar cap; NTBe= center of the north temperature belt, NEB„ and NEB^ are the north and south edges of the north equatorial belt respectively; STBc= center of the south temperate belt and Xc = center of belt which is either the equatorial band or a portion of the north equatorial belt which split. The selected values are the equally weighted average of the values from the four methods. Uncertainties are estimated from: standard deviations, consistency between the different methods and visibility of the feature. Between 1989 and 1991, the NEB appears to have moved northwards (Benton 1990; 1992; 1993; Heath 1992a; 1992b; 1993b; Schmude 1990), but this movement ceased during 1993. The 1993 position of the south temperate belt and the north polar region are consistent with those observed during the 1960s and 1970s (Hollis 1980). The colors and intensities of the zones and belts of Saturn are listed in Table 4. The colors were estimated visually through the eyepiece and from color slides of Saturn. Kodak 100 HC color slide film was used 18 THE TEXAS JOURNAL OF SCIENCE— VOL. 47, NO. 1, 1995 Table 4. Summary of intensity and color estimates of various atmospheric features on Saturn. Feature Intensity Color NPR 4.2 Gray NTZ 3.5 Yellow-gray NTB 4.0 Brownish-gray NTrZ 3.1 Yellow-gray NEB 5.1 Brownish-gray EZ 2.4 Yellow- white STrZ 2.8 Yellow- white STB 4.2 Gray STZ 3.3 for all slides. Exposure times of 5-10 seconds were used at f/106 for the slides. Intensities were estimated on a scale from 0-bright white to 10-black using the same methods as reported by Schmude (1990). The colors and intensities, with the exception of the C ring, are similar to those in 1991 (Heath 1993; Benton 1993), 1990 (Heath 1992; Benton 1992) and 1989 (Benton 1990; Schmude 1990; Heath 1992). The consistency in colors and intensities is in agreement with the relatively constant photometric magnitudes of Saturn measured during 1991-1993. The C ring was more distinct in 1993 than in the period 1987-1992. This may be due to the lower tilt of Saturn’s rings. According to McKim & Blaxall (1984), the contrast between the C ring and the black sky background increased as B decreased. Discussion The photometric results for Saturn in 1993 suggest that the solar phase coefficient depends on the tilt of the rings (angle B). The B-V color index for 1993 was 0.99+0.04 which is a little lower than previous measurements (Nautical Almanac 1992; Harris 1961); this discrepancy may be due to the changing tilt of Saturn’s rings. In future studies, photometric measurements need to be carried out at different values of the ring tilt so that variations due to the rings can be determined. Only in this way can changes due to Saturn’s globe be detected using photometry. Definite statements about the globe of Saturn by itself thus can not be made until photometry of the rings is better understood. The major contribution of this work is that it adds to the current knowledge of the photometric behavior of Saturn’s rings. SCHMUDE & BRUTON 19 Conclusions In summary, a detailed photometric/photographic and visual study of Saturn was conducted during late 1993. Respective solar phase coefficients in the B, V, R and I filters were measured as: 0.055+0.007; 0.054+0.005; 0.047+0.010 and 0.053+0.007 in late 1993. It is concluded that the solar phase coefficient of Saturn (globe + rings) in the V-filter changes with respect to the ring tilt, B. A normalized magnitude of Saturn calculated from equation 1 and a solar phase coefficient of 0.054 mag. /deg. is V(l,0)=-8.84. It is also concluded that the large white storm of 1990 did not create large changes in the color of Saturn in 1991-1993, but that a small shift in belt latitudes may have taken place. References Barnet, C. D., J. A. Westphal, R. F. Beebe, & L. F. Huber. 1992. Hubble Space Telescope Observations of the 1990 Equatorial Disturbance on Saturn: Zonal Winds and Central Meridian Albedos, Icarus, 100:499-511. Beebe, R. F., C. Barnet, P. V. Sada, & A. S. Murrell. 1992. The Onset and Growth of the 1990 Equatorial Disturbance on Saturn, Icarus, 95:163-172. Benton, J. L. 1990. The 1988-89 Apparition of Saturn: Visual and Photographic Observations, J. Assoc. Lunar & Planet. Obs., 34:160-169. Benton, J. L., Jr. 1992. The 1990-91 Apparition of Saturn: Visual and Photographic Observations, J. Assoc. Lunar & Planet. Obs., 36:49-62. _ . 1993. The 1991-92 Apparition of Saturn: Visual and Photographic Observations, J. Assoc. Lunar & Planet. Obs., 37:1-13. Harris, D. L. 1961. Photometry and Colorimetry of Planets and Satellites, in: Planets and Satellites, Kuiper, G. P. and Middlehurst, B. M.-eds. The University of Chicago Press, Chicago 601 pp. Heath, A. W. 1992a. Saturn 1989, J. Brit. Astron. Assoc., 102:85-92. _ . 1992b. Saturn 1990, J. Brit. Astron. Assoc., 102:205-209. _ . 1993. Saturn 1991, J. Brit. Astron. Assoc., 103:228-234. _ & R. J. McKim. 1992. Saturn 1990: The Great White Spot, J. Brit. Astron. Assoc., 102:210-219. Hirshfeld, A. & R. W. Sinnott- Editors. 1985. Sky Catalog 2000.0 Volume 1: Stars to Magnitude 8.0, vol. 1, Cambridge University Press, New York, 604 pp. 20 THE TEXAS JOURNAL OF SCIENCE— VOL. 47, NO. 1, 1995 Hollis, A. J. 1980. Saturn-The Latitudes of the Belts 1946-1976, J. Brit. Astron. Assoc. 91:41-53. Iriarte, B., et al. 1965. Five-Color Photometry of Bright Stars, Sky and Telesc., 30:21-31. McKim, R. J. & K. W. Blaxall. 1984. Saturn 1943-1981: A Visual Photometric Study-I, J. Brit. Astron. Assoc. 94:145-151. Nautical Almanac for the Year 1914. 1911. Govt. Printing Office, Washington D. C. 742 pp. Nautical Almanac for the Year 1917. 1915a. Govt. Printing Office, Washington D. C. 750 pp. Nautical Almanac for the Year 1918. 1915b. Govt. Printing Office, Washington D. C. 750 pp. Nautical Almanac for the Year 1920. 1917. Govt. Printing Office, Washington D. C. 786 pp. Nautical Almanac for the Year 1993. 1992. U. S. Govt. Printing Office, Washington D. C. 539 pp. Optec, Inc. 1988. Model SSP-3 Solid-State Photometer Technical Manual for Theory of Operation and Operating Procedures", Lowell, MI, 28 pp. Peek, B. M. 1958. The Planet Jupiter. Faber and Faber, London, 283 pp. Schmude, R. W., Jr. 1990. Observations of Saturn in the Spring and Summer of 1989, Texas J. Sci., 42:283-288. _ . 1992. The 1991 Apparition of Uranus, J. Assoc. Lunar & Planet. Obs., 36:20-22. _ . 1993. Wideband Photometry of Mercury, Venus, Jupiter, Callisto and Saturn, J. Roy. Astron. Soc. Can., 87:133-139. Veverka, J., et al. 1988. Photometry and Polarimetry of Mercury, Mercury, Vilas, F. et al.-eds.. The University of Arizona Press, Tucson, 794 pp. TEXAS J. SCI. 47(1)21-38 FEBRUARY, 1995 A COMPARISON OF SIX TEST STATISTICS FOR DETECTING MULTIVARIATE NON-NORMALITY WHICH UTILIZE THE MULTIVARIATE SQUARED-RADII STATISTIC Dean M. Young, Samuel L. Seaman and John W. Seaman, Jr. Hankamer School of Business, Baylor University, Waco, Texas 76798 Abstract.— This study presents tabulated emprically-derived critical values for Hawkins’ test for non-normality, and compares the power of this test to five other test statistics designed to detect multivariate non-normality, all of which are functions of the multivariate squared-radii statistic. The power comparison has been accomplished using a Monte Carlo simulation with two sample sizes, two observation dimensions, and ten multivariate non¬ normal distributions. Among the six test statistics considered in the present study, the one proposed by Hawkins (1981) has proven to be the best omnibus test statistic for detecting multivariate non-normality. Empirically calculated critical values for Hawkin’ s test statistic for detecting multivariate non-normality are given as an appendix. Strategies for testing the hypothesis of multivariate normality of a population from a set of sampled multivariate observations are numerous in the statistical literature. To date, over forty different test statistics have been recommended for this purpose. The interested reader is referred to thorough reviews by Gnanadesikan (1977), Mardia (1980), Koziol (1986), and Looney (1986). Attempts at detecting deviations from multivariate normality, using sample evidence from a set of multivariate observations, have typically employed one of the following strategies: (1) apply univariate techniques to detect marginal univariate non-normality for each dimension, (2) utilize multivariate techniques to detect joint non-normality, or (3) employ a univariate summary statistic to test for multivariate non-normality. Given a set of /^-dimensional random variables Xj, X2, ..., X^, the statistic most often utilized in testing for multivariate normality after the manner of strategy (3) is the squared sample radii statistic defined by where Dj =(X, -X)'S'' (X, -X) n-\ (1) Several of the techniques designed to test the multivariate normality hypothesis have employed some variation of the multivariate squared radii statistic defined in (1). For example, Healy (1968) suggested that 22 THE TEXAS JOURNAL OF SCIENCE-VOL. 47. NO. 1, 1995 the tables in Wilk et al. (1962) be used to construct a plot so that the multivariate normality hypothesis can be tested visually when p = 2. Malkovich & Afiffi (1973) proposed applying the Cramer-Von Mises and the Kolmogorov-Smirnov statistics to test the hypothesis that the values of D- have an approximate x^(p) distribution. For cases where /7 > 2, Small (1978) proposed plotting the order statistics of against the expected order statistics from a multiple of a beta distribution since, under the hypothesis of multivariate normality, the marginal distribution of D- is proportional to [(/?-l)-l]' n-l Beta p (/?-l)-p-l 2’ 2 where Beta {a, b) denotes a beta distribution with shape parameters a and b. Another test procedure utilizing the squared-radii statistic has been formulated by Hawkins (1981) for simultaneously testing the assumption of multivariate normality of two or more sets of multivariate observa¬ tions. He has proposed transforming the squared radii statistics into statistics with approximate /^-distributions, assuming that multivariate normality of all data sets holds. He has shown that, under the assump¬ tion of multivariate normality, the tail probabilities will be distributed uniformly on the open unit interval. Hawkins suggested using the Anderson-Darling test statistic to test the assumption of uniformity for the transformed tail probabilities. Moore & Stubblebine (1981) proposed a multivariate normality test statistic which also is based upon the squared- radii statistic. The test statistic is of the form nk which has an approximate distribution of x^(^), k — 1 < q < k, where O- is the number of /)•, i = 1,2,...,«, whose values are in cell j,j = 1,2,...,/:. One advantage of this statistic is that approximate critical values are easily obtained. Fattorini (1982) proposed two statistics based on D- that may be used to test for multivariate non- normality. The first statistic is the average relative discrepancy among the sample order statistics and the expected order statistics from a multiple of the beta distribution. The second statistic utilizes the Theil index to measure the goodness of fit between YOUNG, SEAMAN & SEAMAN 23 the sample order statistics and the expected order statistics from the beta distribution. Koziol (1982) derived the asymptotic distribution of the Cramer-Von Mises type test of Malkovich and Afifi and also derived critical values via a Monte Carlo simulation for various sample sizes, dimensions, and significance levels. However, these critical values are not reported in the paper. Royston (1983) formulated a test for multivariate normality based on the squared-radii statistic in which, assuming the hypothesis of multi¬ variate normality, the squared-radii are transformed to near normality and then summed to form an approximate random variable. Booker et al. (1984) noted that the y^(p) reference distribution used by Malkovich & Afifi (1973) in their Kolmogorov-Smirnov type test for multivariate normality could be improved by applying a multiple of the beta distribution as the reference distribution. However, the power of this test was examined for only the limited case of p = 2. Paulson et al. (1987) proposed two tests for multivariate normality utilizing the squared-radii statistic. They find empirical critical values for dimensions one through five for an Anderson-Darling type statistic and they also formulate a test based on the Kullback divergence statistic. Tsai & Koziol (1988) suggested using the Pearson correlation coefficient as a measure of the strength of the relationship between the order statistic for the squared-radii, D-, and the approximate expected order statistics of the D- when assuming multivariate normality of the underlying population. This paper compares the relative powers of six test statistics for detecting multivariate non- normality, all of which are functions of the squared-radii statistic. The power comparison is accomplished using a Monte Carlo simulation encompassing a variety of multivariate non¬ normal distributions. Additionally, a table of empirically-derived critical values for Hawkins’ test statistic for multivariate non-normality is constructed. Tabled critical values can be found in an appendix whilst a description of the simulation used to generate those critical values can be found in the next section along with a brief discussion of each of the six test statistics that are based on the squared-radii statistic and that have been compared in this study. Then, in Section 3, a brief description is given of the Monte Carlo simulation used for the power comparison and the results of that power comparison is presented. 24 THE TEXAS JOURNAL OF SCIENCE- VOL. 47. NO. 1, 1995 Finally, in Section 4, comments on the simulation results are given and recommendations are made regarding the choice of a test statistic. Six Test Statistics for Detection of Multivariate Non-Normality Based upon the Squared-Radii Statistic The goal of this study is the comparison of the powers of six test statistics designed to detect multivariate non- normality. Each of these statistics is a function of the squared- radii statistic defined in (1). For completeness a brief description of these test statistics is presented. Hawkins Test Statistic (HAW) Hawkins (1981) proposed a statistic for detecting multivariate non¬ normality which is a function of the squared-radii statistic, D-. His procedure, which may be applied to observations from one or more populations simultaneously, is based upon a transformation of the squared-radii into statistics which have exact F-distributions under the assumption that the underlying populations are multivariate normal. If this assumption is true, the tail probabilities of the proposed statistic are distributed uniformly on the interval (0,1). The Anderson-Darling methodology is then employed to assess the uniformity of the tail probabilities. Hawkins’ test statistic for detecting multivariate non¬ normality of a single population may be described as follows. Let D- be defined as in (1) and let p (/?-p-l)^?A p[{n -1)- - nD.y Let A- = P[F > FJ denote the tail area of a random variable with an F-distribution having p and {n — p — \) degrees of freedom. Hawkins’ test statistic for detecting multivariate non-normality is based on the n order statistics y4(i) < ^^2) ^ ^ ^(n) of the A^’s and may be written as HAW = -n - - ^ {2j - l)[log + log(l - )]. // j=\ Large values of HAW indicate a departure from the multivariate normal model. Note that this is nothing more than an application of the Anderson-Darling statistic to test uniformity of the /4- values. Empirical critical values for Hawkins test statistic have been obtained via a Monte Carlo simulation that is described in the next section. YOUNG, SEAMAN & SEAMAN 25 Empirically-derived Critical Values for Hawkins * Test Statistic For each combination of sample size n = 10, 20, 30, 40, 50, 75, 100 and dimension p = 2,3, 4, 5, 6, 8, 10, 12, and 15, four sets of 5,000 sample observations of the statistic HAW have been generated from the /7-dimensional standard normal distribution (i.e. a /?- variate normal distribution with mean vector 0 and covariance matrix E = I) . Notice that it will be sufficient to use the /?- variate standard normal since the V- are invariant under linear transformations. For each combination of n and p, each set of 5,000 observations has been ordered and the appropri¬ ate sample quantile selected to estimate the critical value. The critical values that have been tabulated in the appendix, are actually the averages of the four sample quantiles for significance levels of 0.1, 0.05, 0.025, 0.01 and 0.005. Computations have been performed on an IBM 4381 computer under the VM/CMS operating system in the Casey Computer Center at Baylor University. The code has been written in the SAS/IML software. One further observation about the empirically generated critical values is worth mentioning here. A comparison of the asymptotic, Anderson- Darling critical values (recommended by Hawkins) and the empirically- derived critical values found in the appendix suggests that the asymptotic critical values may be quite conservative when applied to Hawkins’ test statistic. Consider, for example, the case where n = AQ, p = 5, and a significance level of . 10 is adopted. Under these experimental condi¬ tions, the asymptotic critical value, which is independent of p, is 1.933, A glance at the tabled empirical values in the appendix, however, re¬ veals that the /7-value corresponding to 1.933 is actually less than 0.005. The Paulson-Roohan-Sullo Test Statistic (PRS) The PRS test statistic was formulated by Paulson et al. (1987). This test statistic for detecting multivariate non-normality, like Hawkins’ statistic, is based on the Anderson-Darling formulation. The PRS statistic may be expressed as PRS = -1 (2; - 1)( log G(D, ) + log[l - )]) where G(«) is the cumulative distribution of a x^(p) random variable and is the j order statistic of the squared-radii statistic defined in (1). Note that Hawkins’ test procedure differs from the PRS test procedure 26 THE TEXAS JOURNAL OF SCIENCE-VOL. 47. NO. 1, 1995 in that the latter statistic adopts a ^ -approximation for the D/s while the former test statistic utilizes a transformation of the D,’s, resulting in tail probabilities with exact uniform distributions. Empirical critical values for the PRS statistic can be found in Paulson et al. (1987). The Tsai-Koziol Test Statistic (TK) The TK test statistic may be described as follows. Let < Q2 < < |2n denote the expected order statistics in a sample of size n from a -distribution with p degrees of freedom. The Tsai-Koziol statistic, then, is of the form _ ] « _ 1 « where ^ ~ X A and Q =-Y Q . Note that this nt; statistic is nothing more than the Pearson correlation estimate for the correlation between the expected and empirical order statistics for the multivariate squared-radii statistic. The null hypothesis of multivariate normality is rejected for sufficiently small TK values. A selected group of sample critical values may be found in Tsai & Koziol (1988). The Extended Malkovich and Afifi Test Statistic (EM A) Malkovich & Afifi (1973) proposed another test statistic for detecting multivariate non-normality which is a function of the squared-radii statistic. Their statistic is essentially an extension of the Lilliefors statistic for testing univariate normality. The Malkovich and Afifi statistic is of the form EMA = sup|F„(z)-G(z)| where F„(z) is the sample cumulative distribution function of the squared-radii statistic and G{z) is the cumulative distribution function of a x^(p) random variable. Concerning the performance of their test statistic, Malkovich and Afifi state "... a better approximation than x^ (p) may be appropriate as the hypothetical distribution of Z)- ...." Jennings et al. (1990) applied a multiple of a beta random variable as an YOUNG, SEAMAN & SEAMAN 27 approximation to the distribution of Z)-. This statistic is of the form EMA* =sup|F„(z)-G*(z)| where G*(z) is taken to be a scaled Beta distribution. Note that this formulation is an extension of the statistic proposed by Booker et al. (1984) in the case where the dimensionality of the observation vectors is greater than two. Empirical critical values for this test statistic have been generated by Jennings et al. (1990). * The Cramer-Von Mises Test Statistic (CM) Koziol (1982) derived a test statistic for detecting multivariate non¬ normality which is based on the Cramer-Von Mises distance measure between two distribution functions. This distance measure is of the form J[F(z)-G(z)]-dG,(z) 0 where F and G are cumulative distribution functions. The Cramer-Von Mises test statistic formulated by Koziol (1982) is expressed as CM=— +y 12^? tr 27-lf 2n where G is the cumulative distribution function of a random variable with p degrees of freedom. Unfortunately, Koziol (1982) includes only a limited number of empirical critical values. The Percent Mean Difference Test Statistic (PME) Fattorini (1982) suggested using the percent mean difference of the estimated quantiles of from the approximated expected quantiles of the squared-radii statistic assuming multivariate normality. The approximate expected quantiles of D- are calculated as functions of approximate beta quantiles of order i-{a-\)l2a 2a 2b 28 THE TEXAS JOURNAL OF SCIENCE- VOL. 47. NO. 1, 1995 denoted by q-, where a = pH and b = (n — p — l)/2. The expected quantiles for D, may then be approximated by v, = [{n — The percent mean difference in the estimated quantiles, D( j ), and the approximate expected quantiles of D-, v^, is then expressed as PME=J-yliLjilil. Note that large values of PME indicate evidence of multivariate non¬ normality. Empirically derived critical values for selected sample sizes with dimensions 2 through 6 are given in Fattorini (1982). The Simulation for Power Comparisons To evaluate the relative powers of the six test statistics for detecting multivariate non- normality which are functions of the squared-radii statistic defined in (1), we conducted a Monte Carlo simulation using SAS/IML under the VM/CMS operating system on an IBM 4381. The simulation was performed in the following manner. Sets of ten thousand random vectors for sample sizes n = 20 and n = 50 from various nonnormal multivariate populations of dimensions p = 2 and p = 6 were generated. We evaluated each of the six test statistics using all possible configurations of sample size, dimension, and form of the nonnormal distribution. The power study simulation made extensive use of the r-normed exponential distribution family which consists of symmetric, multivariate distributions. The reader may consult Goodman & Kotz (1973) or Chhikara & Odell (1973) for a complete discussion of this family. This study used multivariate r-normed exponential distributions with r = 1 , 1.1, 1.2, 1.3, 1.4, 1.5, and 10. Other nonnormal distributions used in this study include four /^-dimensional distributions with marginal variables having 1, 2, 3, and 4 degrees of freedom; and a /7-dimension- al distribution with marginal uniform variates. The non-normality of these distributions was assessed using multivariate measures of skewness and kurtosis formulated by Mardia (1970). The multivariate skewness measure is P...= I i.j,k=l e,/.g=l YOUNG, SEAMAN & SEAMAN and the multivariate kurtosis measure is 29 ij=ik,i=\ where 1 - H/)"]. and = (a*-' ) for i,j = 1,...,/?. Mardia (1970) showed that = 0 and 132^ p = pip +2) for multivariate normal distributions. It is noted, here, that there are many types of non-normality and that Mardia’ s measures of multivariate skewness and kurtosis do not characterize all of them. Empirical powers were calculated as the proportion of rejections at both the a = 0.10 and the a = 0.05 levels of significance. Empirically-generated critical values (based upon 10,000 samples) are employed for all test statistics in the interest of equitable power comparison. Power Simulation Results The results of our power comparison (for a = 0. 10) is given in Figure 1 . Each pair of histograms in the figure shows the powers of the six test statistics for both sample sizes (n = 20 or n = 50) when p = 2 (graph on left) or 6 (graph on right). When a. = 0.05, the powers of all tests, naturally, are smaller compared to a. = 0.10. The relative performances of the tests, however, are unaffected by the choice of ot (0.10 or 0.05) for each experimental combination of n, /?, and type of non-normality examined. In the interest of brevity, therefore, we have not presented the results for a = 0.05. Careful examination of each of the graphs in Figure 1 reveals that no test statistic is uniformly most powerful over all of the configurations considered in the simulation study. Indeed, the results presented here give us reason to reiterate, albeit less vigorously, the recommendations of Andrews, Gnanadesikan, & Warner (1973:95) who advise that "...a variety of techniques with differing sensitivities to the different types of departures" should be used when testing for multivariate non-normality. In the subsections that follow, we comment on the "sensitivity" and 30 THE TEXAS JOURNAL OF SCIENCE-VOL. 47. NO. 1, 1995 relative performance of each of the test statistics examined over the various combinations oin, p, and type of non- normality. Hawkins Test Statistic (HAW) Hawkins’ statistic yields excellent power characterstics for many of the distributions considered in this study. This test statistic has, for example, excellent power against both symmetric, heavy-tailed and skewed distributions regardless of the sample size or dimension. From the graphs in Figure 1, it is clear that Hawkins’ test statistic very often enjoys increased statistical power when the dimension is expanded from 2 to 6, for both sample sizes. In contrast, all but one of the competing statistics tended to lose power as the dimension was increased, especially with small samples (n=20). The one exception is the PME statistic which we describe in more detail below. Hawkins’ statistic does not, however, exhibit high power against symmetric light-tailed distributions and, in fact, the comparatively poor power of Hawkins’ test for these types of distributions worsens with increasing dimension. Percent Mean Difference Test Statistic (PME) Fattorini’s PME test statistic also exhibits good power against skewed and symmetric, heavy-tailed distributions. For all of these types of non¬ normality, however, the power of the PME statistic declined markedly when the dimension is reduced from 6 to 2. This phenomenon is especi¬ ally noticeable with small samples (n=20). The practical implication, here, is that while the statistic’s relative performance is good for skewed and symmetric, heavy-tailed distributions, that performance depends in large part upon the ratio, nip, and suffers greatly when this ratio is large. In addition, the PME statistic, like Hawkins’ statistic, has relatively poor power against symmetric, light-tailed distributions. There is little reason, then, to recommend the PME statistic over Hawkins’ statistic unless the non- normality is likely to be in the form of very heavy-tailed distributions and the ratio, nip, is quite small. Paulson-Roohan-Sullo Test Statistic (PRS) The PRS test statistic provides superior power only on those occasions where the non-normality manifests itself in the form of symmetric, light¬ tailed distributions. Results presented in Figures 1, however, do demon¬ strate that on such occasions, the PRS test statistic has markedly better relative power than all but one other statistic (see CM below). The YOUNG, SEAMAN & SEAMAN 31 practical implication is that this statistic is most powerful in situations where Hawldns’ test and the PME statistic provide relatively poor per¬ formances. The reader is cautioned, however, that both the PRS and CM statistics have relatively poor power against skewed and symmetric, heavy- tailed distributions. Indeed, on some occasions the powers of these tests can be smaller than the actual level of the test (i.e. against symmetric, medium to heavy- tailed distributions where the ratio of the sample size to dimension is relatively small {nip < 4)). Thus, the PRS and CM test statistics biased tests for detecting multivariate non¬ normality. Tsai-Koziol Test Statistic (TK) In terms of statistical power, the TK test statistic proves to be markedly inferior to nearly all of the other test statistics except on two occasions where the ratio of sample size to dimension was large and the non-normality occurred in the form of moderately skewed distributions. There is little reason to consider this statistic in the data analytic setting or in any future research efforts. Cramer-Von Mises Test Statistic (CM) The CM statistic, like the PRS test statistic enjoys power advantages only on those occasions characterized by symmetric, light-tailed distributions. Even then, the power of the CM statistic is less than that of the PRS statistic. Both of these statistics are inferior to nearly all other tests examined when attempting to detect multivariate non¬ normality in the form of heavy- tailed or skewed distributions. Finally, there is little reason to prefer CM to PRS when choosing a test that will be sensitive to symmetric, light-tailed forms of non- normality. Extended Malkovich and Afifi Test Statistic (EM A) The EM A statistic enjoys adequate relative power against skewed and symmetric, heavy- tailed distributions, typically ranking third (after Hawkins and PME) in order of relative power for these experimental conditions. The EM A statistic does enjoy power advantages over all other tests, on those occasions where the nonnormal distribution is extremely skewed and the ratio of sample size to dimension is quite large. However, under these conditions all of the tests have reasonably large powers and any power differences are likely to be inconsequential. Incidently, the EM A statistic also gives a relatively poor performance 32 THE TEXAS JOURNAL OF SCIENCE-VOL. 47. NO. 1, 1995 against multivariate non-normality in the form of light-tailed distributions. Conclusions This study compares the relative powers of six test statistics (all of which are ftmctions of the squared-radii statistic) that can be used to detect multivariate non-normality. While none of the statistics considered here was most powerful against all of the alternative distributions simulated, Hawkins’ test statistic appears to have relatively good power against many of the types of multivariate non- normality considered in the present study. This is especially true of non- normality in the form of skewed or heavy-tailed distributions. Even on those occasions when Hawkins’ test statistic does not yield superior power (for symmetric, light-tailed distributions), the power is fairly good in that the power differences between Hawkins’ test and the "best" test statistic never exceeds about .10. It is also worth mentioning that Hawkins’ test statistic is one of a few tests examined here that benefits (enjoys increased power) from increases in the dimension with no associated increases in sample size. Additionally, Hawkins’ test statistic is not difficult to compute and is readily applied in the research setting. Two cautionary notes on applying Hawkins’ test statistic is in order. If the ratio of the sample size n to the dimension p is too small, the statistic H, can be negative and, therefore, useless. Simulation results indicate that this condition can usually be avoided if care is taken to insure that the ratio of sample size to dimension will be greater than or equal to 2. Also, in his paper Hawkins (1981) uses asymptotic critical values for an example application and states that the asymptotic critical values seem to be adequate. Unfortunately, this study reveals that the asymptotic critical values suggested by Hawkins (1981) yield test levels that differ considerably from the assumed test levels using asymptotic critical values. As noted in Section 2 critical values for Hawkins’ statistic have been tabulated and are given in the appendix. Literature Cited Andrews, D. F., R. Gnanadesikan & J. L. Warner. 1973. Methods for assessing multivariate normality. P. 95-116 in Multivariate Analysis- Ill. (P. Krishniah, ed.). Academic Press, New York. Booker, J. M., M. E. Johnson & R. J. Beckman. 1984. Investigation of an empirical probability measure based test for multivariate normality. ASA Proceedings of the Stat. Comp. Section 208-213. YOUNG, SEAMAN & SEAMAN 33 Chhikara, R. S. & P. L. Odell. 1973. Discriminant analysis using certain normal exponential densities with emphasis on remote sensing applications. Pattern Recognition 5:259-272. Fattorini, L. 1982. Assessing multivariate normality on beta plots. Statistica, 42:251-257. Goodman, I. R. & S. Kotz. 1973. Multivariate ^-generalized normal distributions. J. Multivariate Anal. 3:204-219. Hawkins, D. M. 1981. A new test for multivariate normality and homoscedasticity. Technometrics 23:105-110. Healy, M. J. R. 1968. Multivariate normal plotting. Applied Statistics 17:157-161. Jennings, L. W., D. M. Young & J. W. Seaman. 1990. An extension of the Malkovich and Afifi test for multivariate normality. Unpublished paper, Baylor University, Waco, 12 pp. Koziol, J. A. 1982. A class of invariant procedures for assessing multivariate normality. Biometrika 69:423-427. _ . 1986. Assessing multivariate normality: a compendium. Comm. Statist. Theor. and Meth. 15:2763-2783. Looney, S. W. 1986. A review of techniques for assessing multivariate normality. ASA Proceedings of the Stat. Comp. Section 280-285. Malkovich J. F. & A. A. Afifi. 1973. On tests for multivariate normality. Journ. Amer. Statist. Assn. 68:176-179. Mardia, K. V. 1970. Measures of multivariate skewness and kurtosis with applications. Biometrika 57:519-530. Moore, D. S. & J. B. Stubblebine. 1981. Chi-square tests for multivariate normality with applications to common stock prices. Commun. Stat.-Theor. and Meth. 10:713-738. Paulson, A. S., P. Roohan & P. Sullo. 1987. Some empirical distribution function tests for multivariate normality. J. Statist. Comput. Simul. 28:15-30. Royston, J. P. 1983. An extension of Shapiro and Wilk’s test for normality to large samples. Applied Statistics 3 1 : 1 15- 124. Small N. J. H. 1978. Plotting squared radii. Biometrika 65:657-658. Tsai, K. & J. A. Koziol. 1988. A correlation type procedure for assessing multivariate normality. Commun. Statist.- Simula. 17:637- 651. Wilk, M. B., R. Gnanadesikan & M. J. Huyett. 1962. Probability plots for the gamma distribution. Technometrics 4:1-20. 34 THE TEXAS JOURNAL OF SCIENCE— VOL. 47. NO. 1, 1995 Figure 1. Powers of the Six Test Statistics for I ^N=20 E:ZIII]N=50 a = 0.10 p=6 p=2 X\3): pu=5.33 P2^=16.00 PRS - 0.133 TK - 0.239 CM H 0.124 RMA X\3): p 1.6=16 p 2.6=72.00 IT 10.895 868 0.768 “1 X\4): P,.2=4 P2.2=14.00 X\4): P,.6=12 P2.6=66.00 0.890 0.852 0.907 0.801 Unifomi: Uniform: Pi,6=0 0.366 - HAW - 0.219 0.446 0.556 0.209 0.477 PME PRS TK CM RMA - 0.243 - 0.178 - 0.204 0.896 0.945 0.758 POWER YOUNG, SEAMAN & SEAMAN 35 Figure 1. Powers of the Six Test Statistics for 1 IN=20 ^^HN=50 a = 0.10 P=2 p=6 MEP(r=1.0): Pu=0 p„=14.00 MEP(r=1.0): p,.6=0 P2.6=66.00 MEP(r=l.l): pi^=0 p2j=12.55 MEP(r=l.l): p,.6=0 p2.6=61.65 MEP(r=U): p,j=0 P2;!=11.48 0.615 0.577 0.581 ) 1 0.272 1 0.263 0.277 0.183 0.244 0.312 MEP(r=U): p,.s=0 P2.6=58.46 - TK -0.125 "T . . I .. ; -|o.42: 0.231 0.719 POWER 36 THE TEXAS JOURNAL OF SCIENCE— VOL. 47. NO. 1, 1995 Figure 1. Powers of the Six Test Statistics for I I N=20 N=50 a = 0. 1 0 p=2 p=6 MEP(r=lJ): Pi^=0 P2^=10.67 MEP(r=lJ): p 1.5=0 P2.6=56.02 0.447 0.456 0.439 0.472 0.213 - PRS -0. - 0.279 - 0.274 -0.081.' ' ... |o.359 -0.125 ; 0.187 - 1 - : - - - 0.046 . 0.302 -0.192 0.440 0.576 0.574 MEP(r=1.4): p,.2=0 p2.2=10.30 MEP(r=1.4): p,.6=0 P2.6=54.10 0.376 0.228 . HAW - 0.221 0.458 0.336 0.183 . PME - 0.220 0.451 0.344 0.180 - PRS - - 1 - 0.078 j jo.249 0.179 0.137 - TK - 0.115 |o.l47 0.33C 1 0.164 - CM - 0 049 0.205 0.360 0.222 ■ RMA - 0.163 1 0.336 MEP(r=1.5): P1.2-O P2.2=9.52 0.308 MEP(r=1.5): Pi.6=0 P2.6=41.30 0.189 - HAW - 0.184 0.359 60 0.145 - PME - 0.184 0.350 70 0.142 - PRS - 0.079 0.135 0.152 s 0.120 - TK - 0.126 1 0.135 62 I 0.130 - CM - 1 . 0 051 1...'. . 0.148 0.175 - RMA - 0.145 . .\-,'jo.265 YOUNG, SEAMAN & SEAMAN 37 Figure 1. Powers of the Six Test Statistics for ]N=20 [=50 a = 0.10 p=2 p=6 MEP(r=10): pij=0 P2j=0.18 MEP(r=10): Pi.6=0 P2,6=5.76 Xi[l): Pu=16 P2;.=32.00 X^l): Pi.6=16 p2.6=120.(X) Xi:2): p,j=8 P2j=20.00 0.504 ■ 0.786 ^ 0.397 - i 0.439 - 0.485 - POWER X^2): P,.6=24 p2.6=84.00 HAW - 0.639 bsl4v:^^4lo.988 PME - 0.605 PRS - 0.254 TK - 0.417 ;>:^&|i%^Jo.896 CM - 0.240 1 RMA - 0.568 0.0 OJ 1.0 POWER 38 THE TEXAS JOURNAL OF SCIENCE— VOL. 47. NO. 1, 1995 o p ooor-f^-- CO ON 00 — ^ O ON 00 ON ON 00 rsi ^ ' O ^ On ^ ^ m CO 00 O nC On On 00 r*^ On On NO CO 00 CO On CO On 00 Tf »r) ON O 00 00 00 00 00 On NO NO »n in in CO in o 00 00 00 00 oo On 00 ON r' ^ vC NC 00 00 00 00 00 CO NO 00 cs »rj rr — in o O' NO O' TJ- -sj- (N Tt ON 00 Q — < 00 o p sc o^ r- NO NO r*' NO nc vo ON o r-* 00 m VC vo 'O 'n Nc o VO tN m oo m ^ lo lO NO in NO lo ^ CO CO P VC VO NC NO ^ CO m p 04 NO NO NC NC NO m (N 00 (N — CN vo gj ? 04 NO CN 3 S NO O O NO CS 5 5 - 2 5 2 S -2 (N P m Tf r^i m CO CO CO CO CO CO CO CO CO CO CO CO CO CO CO CO CO u > On JO ^ ^ ^ ^ iq — ^ ^ O 00 fs ^ d> <5 d> cS d> d d d d d o o d o o O O o o m o o o o o »o O o o o o o o o o in S o O O m 2 cN m 2 (N CO tt 2 0 04 04 00 04 0 Tf rf cn Tf ^ rj- ^ .05 04 ON 04 O 04 04 NO NO — cn cn 04 04 cn m m cn cn 00 0- 0- NO m NO 04 04 cn 04 m m 04 04 04 04 04 04 04 04 NO ^ in 00 ^ nc On ^ 0 04 ^ ^ 04 04 04 Tf ^ 00 NO 04 m 0 On 00 r- 00 00 On o O ^ cn — ' o^ m 04 m cn 14O «n in p p p p p p p cn m 0^ ON in in >0 00 00 On On — p p On On ON ON 0 0 0 d d d d 00 04 ^ 0^ ^ 0- On cn in NO NO 00 ON On O' On On On On On d d d d d d d ON On On NO 04 cn cn cn m m m NO On GS On O' O' On On d d d d d d d Sample Dimension Size 0 0 C 0 0 m 0 — < fN m in 0 II II II II II II li 0 0 0 0 0 m 0 — 04 m in 0 II II II II II II li s: 0 0 0 0 0 m 0 04 m Tj- in r~ 0 II II II II II II li s: 0 0 0 0 0 m 0 — 04 m Tf m 0 II II II II II II 7 s; O'! II Cl. m II =x ij- II Cl. m II ex Appendix. Approximate upper .10, .05, .025, .001 and .0005 level critical values for Hawkins’ Test for multivariate normality of a single distribution. TEXAS J. SCI. 47(l):39-44 FEBRUARY, 1995 VOMEROLFACTORY EXPLORATION OF NOVEL ENVIRONMENTS BY THE PARTHENOGENETIC WHIPTAIL LIZARD CNEMIDOPHORUS LAREDOENSIS (SAURIA: TEIIDAE) Loraine R. Rybiski and Mark A. Paulissen Department of Biological and Environmental Sciences, McNeese State University, Lake Charles, Louisiana 70609 Abstract. — The frequency rate of tongue touch behavior of the whiptail lizard Cnemidophorus laredoensis was observed in the laboratory to determine the nature of this activity when exposed to a variety of different environmental habitats. This study reveals that C. laredoensis increases vomerolfactory exploration equally in all novel environments but, unlike other species of lizards, may be unable to distinguish among different novel odors. Many studies have shown that lizards are capable of detecting ecologically relevant odors (see reviews by Simon 1983; Cooper 1994). In addition, many species are capable of detecting odors of conspecifics and/or heterospecifics (DeFazio et al. 1977; Duvall 1979; Duvall et al., 1980; Bissinger & Simon 1981 ; Cooper & Vitt 1987). Studies of tongue flick rates (Bissinger & Simon 1979) and responses to prey odors (Cooper 1990; Cruz-Neto & Andrade 1993) indicate that members of the lizard family Teiidae exhibit well developed vomerolfaction (sensu Cooper & Burghardt 1990). While vomerolfaction may also play a role in species identification and detection of conspecifics in teiids (Carpenter 1962; Simon 1983), this premise has never been tested. The Laredo striped whiptail, Cnemidophorus laredoensis, is a parthenogenetic teiid lizard which occurs in the sandy disturbed habitats in and near the Rio Grande Valley of Texas and Mexico (Walker 1987). It arose from hybridization between the bisexual species C gularis and C. sexlineatus (cf. McKinney et al. 1973; Bickham et al. 1976; Wright et al. 1983). Most habitat sites occupied by C. laredoensis are also occupied by C gularis (cf. Paulissen et al. 1992). Both species forage for arthropods by moving widely through the habitat while rapidly flicking the tongue; both also use small burrows for overnight retreats (Walker et al. 1986; Paulissen et al. 1988). When a specimen of C. laredoensis forages, it moves away from the vicinity of its burrow and will consequently encounter a variety of novel vomerolfactory stimuli such as odors left on the ground by conspecifics or C. gularis. Specimens may alter their exploratory behavior in novel environments to gain information relative to the presence of other lizards in the 40 THE TEXAS JOURNAL OF SCIENCE- VOL. 47, NO. 1, 1995 immediate area. This study analyzes tongue touch rates of captive C laredoensis to determine the extent that this species uses the tongue to chemically sample novel environments. Methods and Materials Nine adult specimens of Cnemidophorus laredoensis ( = clonal complex LAR-A; Walker 1986; 1987) and four adult specimens of C. gularis were collected in Cameron and Hidalgo counties in south Texas. Specimens were housed individually in 10-gallon (36 by 50 by 31 cm) terraria provided with 3 cm of clean sand, a water dish, and a 5 by 6 cm piece of cardboard for shelter. Each terrarium was heated with a 60 watt lamp placed against the side of the terrarium; overhead fluorescent lights connected to timers simulated mid- summer photoperiod (day length of 14 hr). Each specimen was fed 2-3 mealworms a day and water was provided as needed. Specimens were tested in a terrarium of above dimensions containing soiled sand and a cardboard shelter and heated by a 60 watt lamp. Each specimen was tested in six different odor environments: resident odor (simulating the test specimen’s home environment), conspecific odor (of another C. laredoensis), congeneric male odor (of male C. gularis), congeneric female odor (of female C. gularis), non-saurian odor (no lizard odor), and artifical odor (perfume - Avon’s Knowing). To create the resident odor environment, the test lizard was placed in the test terrarium and left for 24 hours. The specimen was then removed for a few seconds, returned to the test terrarium, and observed. To create the environments with odors of another C. laredoensis or a C. gularis, a lizard was placed into the test terrarium and left for 24 hours after which it was removed and replaced with the test lizard. The environment of non-saurian odor was created by using a terrarium that had not been previously occupied. The artifical odor environment was created by spraying clean sand with perfume just before testing began. Each specimen was exposed to the six odor environments in random order. The cardboard shelter was removed just prior to testing to prevent the test lizard from escaping from the observer’s view. Data were collected by placing a test lizard in a test terrarium and observing it for 20 minutes. The following were measured: (1) number of tongue touches (i.e. extrusions of the tongue onto the sand or glass of the terrarium); (2) time the specimen spent moving (as opposed to basking in front of the light) measured in minutes with a stopwatch; (3) RYBISKI & PAULISSEN 41 number of tongue touches per minute while the specimen was moving about in exploratory fashion (to account for the fact that the room temperature varied slightly among experiments causing lizards to spend varying amounts of time basking). Differences between odor environ¬ ments were statistically evaluated by analysis-of- variance and Duncan’s multiple range tests (SAS Institute 1985). Upon being placed in a test terrarium specimens initially remained stationary for several minutes; they would then begin moving about in the terrarium while rapidly touching the tongue to the sand. After a few minutes the specimen typically basked in front of the heat lamp for a short time; specimens were not observed to tongue touch while basking. When the specimen resumed moving, its behavior consisted of tongue touching, digging, and pushing its nose against the glass or trying to climb out. This alteration of moving and basking continued until the end of the 20 minute testing period. Results and Observations The mean number of tongue touches and time spent moving were significantly lower in the resident odor environment versus the other odor environments, and the number of tongue touches per minute was significantly higher than for the other odor treatments (Table 1). When specimens were tested in their resident odor environment, they moved about and rapidly tongue touched for short amounts of time primarily during the first half of the test period. The rest of the time was spent near the heat lamp. When placed in any of the novel odor environ¬ ments, specimens were observed to spend less time basking and more time moving. Moving involved not only tongue touches, but also attempts to escape (i.e. climbing up the glass, pushing nose against the glass, and digging in the sand). Specimens were much more active in the novel environments and rested only after being away from the heat of the lamp for several minutes. This behavior of alternating basking and trying to escape continued throughout the entire test period. The results of this experiment suggest that specimens of Cnemidophorus laredoensis increase vomerolfactory exploration in novel or unfamiliar environments. Presumably, this potentially allows them to detect novel or unfamiliar areas or burrows. Field observations appear to support this premise. When a specimen of C. laredoensis encounters several burrow entrances in the field, it rapidly tongue touches at first one, then another before finally entering a burrow. These observations suggest that C. laredoensis may be using vomerolfaction to locate their 42 THE TEXAS JOURNAL OF SCIENCE— VOL. 47, NO. 1, 1995 Table 1 . Mean (x) and standard deviation (SD) of vomerolfactory exploration measures for Cnemidophorus laredoensis exposed to six different odor environments for 20 minutes. Results that are significantly different from all other results are indicated by an asterisk (ANOVA Duncan’s multiple-range tests; P < 0.05). Sample size was nine lizards for each odor environment. Odor Number of Environment Tongue Touches Time Spent Tongue Moving (Min) Touches/Min While Moving Resident X = 114.7* X = 5.1* X = 23.4* Odor SD = 52.0 SD = 2.1 SD = 7.4 Conspecific X = 209.2 X = 12.1 X = 17.7 Odor SD = 40.4 SD = 2.6 SD = 3.8 Congeneric X = 203.3 X 11.1 X 18.8 Male Odor SD = 45.6 SD = 3.0 SD = 3.6 Congeneric X = 193.4 X = 11.4 X z= 17.1 Female Odor SD = 30.9 SD = 1.9 SD = 2.2 Non-saurian X = 201.1 X = 12.9 X 15.7 Odor SD = 30.5 SD = 2.2 SD = 1.9 Artificial X = 211.6 X 12.3 X 18.1 Odor SD = 29.7 SD = 3.2 SD = 4.3 own individual retreats and/or avoid burrows occupied by other lizards. This behavioral pattern may serve to minimize contact with conspecifics and so reduce intraspecific competition. Also, it may aid in avoiding male specimens of C. gularis which occasionally copulate with C. laredoensis to produce hybrids that are morphologically and genetically different from the parthenogens (Walker et al. 1989). Conclusions This study reveals that specimens of C. laredoensis exhibit the same tongue-touch rates in all five of the novel odor environments tested (Table 1). This contrasts with the studies of other lizards such as Sceloporus jarrovi which was reported to tongue touch significantly less frequently in a clean cage (i.e. no odor treatment) than in a cage previously occupied by a conspecific (DeFazio et al. 1977; Bissinger & Simon 1981). The fact that C. laredoensis exhibits the same tongue touch rate in different novel environments suggests that either it responds to any novel environment with the same elevation of tongue touch rate or is unable to distinguish among different novel odors. RYBISKI & PAULISSEN 43 Acknowledgments We wish to thank M. Wygoda, G. Haigh, and three reviewers for providing constructive comments on this manuscript. Laboratory space and equipment were provided by the Department of Biological and Environmental Sciences, McNeese State University. Specimens were collected under the authority of scientific collecting permit no. SPR- 0691-408 granted by the Texas Parks and Wildlife Department to M. A. Paulissen. This study is the result of a Senior Honor’s thesis conducted by L. R. Rybiski. Literature Cited Bickham, J. W., C. O. McKinney, & M. F. Mathews. 1976. Karyotypes of the parthenogenetic whiptail lizard Cnemidop horns laredoensis and its presumed parental species (Sauria: Teiidae). Herpetologica, 32:395-399. Bissinger, B. E. & C. A. Simon. 1981. The chemical detection of conspecifics by juvenile Yarrow’s Spiny Lizard, Sceloporus jarrovi. J. Herpetol., 15:77-81. Carpenter, C. C. 1962. Patterns of behavior in two Oklahoma lizards. Amer. Midi. Nat., 67:132-151. Cooper, W. E., Jr. 1990. Prey odor discrimination by teiid and lacertid lizards and the relationship of prey odor detection to foraging mode in lizard families. Copeia, 1990:237-242. . 1994. Chemical discrimination by tongue- flicking in lizards: a review with hypotheses on its origin and its ecological and phylogenetic relationships. J. Chem. Ecol., 20:439-487. ___ & G. M. Burghardt. 1990. Vomerolfaction and vomodor. J. Chem. Ecol., 16:103-105. _ _ & L. J. Vitt. 1987. Intraspecific and interspecific aggression in lizards of the scincid genus Eumeces: chemical detection of conspecific competitors. Herpetologica, 43:7-14. Cruz-Neto, A. P. & D. V. Andrade. 1993. The effect of recent diet on prey odor discrimination by juvenile tegu lizard, Tupinambis teguixin (Sauria: Teiidae). Zool. Anz., 230:123-129. DeFazio, A. C., C. A. Simon, G. A. Middendorf, & D. Romano. 1977. Iguanid substrate licking: a response to novel situations in Sceloporus jarrovi. Copeia, 1977:706-709. Duvall, D. 1979. Western fence lizard {Sceloporus occidentalis) chemical signals I: conspecific discriminations and release of a species-typical visual display. J. Exp. Zool., 210:321-326. 44 THE TEXAS JOURNAL OF SCIENCE-VOL. 47, NO. 1, 1995 _ , R. Herskowitz, & J. Trupiano-Duvall. 1980. Responses of five-lined skinks {Eumeces fasciatus) and ground skinks (Scincella lateralis) to conspecific and interspecific chemical cues. J. Herpetol., 14:121-127. McKinney, C. O., F. R. Kay, & R. A. Anderson. 1973. A new all¬ female species of the genus Cnemidop horns . Herpetologica, 29:361- 366. Paulissen, M. A., J. M. Walker, & J. E. Cordes. 1988. Ecology of syntopic clones of the parthenogenetic whiptail lizard, Cnemidophorus 'laredoensis'. J. Herpetol., 22:331-342. _ , _ , & _ . 1992. Can parthenogenetic Cnemidophorus laredoensis (Teiidae) coexist with its bisexual congeners? J. Herpetol., 26:153-158. SAS Institute. 1985. SAS User’s guide: statistics. SAS Institute, Cary, NC., xvi -f- 956 pp. Simon, C. A. 1983. A review of lizard chemoreception. Pp. 119-133, in Lizard ecology: studies of a model organism, (R. B. Huey, E. R. Pianka, & T. W. Schoener, eds.). Harvard Univ. Press, Cambridge, MA., viii + 501 pp. Walker, J. M. 1986. The taxonomy of parthenogenetic species of hybrid origin: cloned hybrid populations of Cnemidophorus (Sauria: Teiidae). Syst. Zool., 35:427-440. _ . 1987. Distribution and habitat of the parthenogenetic whiptail lizard, Cnemidophorus laredoensis (Sauria: Teiidae). Amer. Midi. Nat., 117:319-332. _ , S. E. Trauth, J. M. Britton, & J. E. Cordes. 1986. Burrows of the parthenogenetic whiptail lizard Cnemidophorus laredoensis (Teiidae) in Webb co., Texas. Southwest. Nat., 31:408-410. _ , J. E. Cordes, & M. A. Paulissen. 1989. Hybrids of two parthenogenetic clonal complexes and a gonochoristic species of the Cnemidophorus, and the relationship of hybridization to habitat characteristics. J. Herpetol., 23:119-130. Wright, J. W., C. Spolsky, & W. M. Brown. 1983. The origin of the parthenogenetic lizard Cnemidophorus laredoensis inferred from mitochondrial DNA analysis. Herpetologica, 39:410-416. TEXAS J. SCI. 47(l):45-52 FEBRUARY, 1995 HABITAT USE OF INTRODUCED AND NATIVE ANOLES (IGUANIDAE: ANOLIS) ALONG THE NORTHERN COAST OF JAMAICA Allan J. Landwer, *Gary W. Ferguson, *Rick Herber and *Mark Brewer Department of Biology, Drawer N, Hardin-Simmons University, Abilene, Texas 79698 and "^Department of Biology, Texas Christian University, Fort Worth, Texas 76129 Abstract. — A comparison of the spatial distribution of the introduced Anolis sagrei with the native Anolis grahami and Anolis lineatopus was conducted along the north coast of Jamaica during the spring of 1983, 1987 and 1994. Spatial distribution was determined as a product of perch height and perch diameter. Perch heights differed significantly among the three species. Due to the presence of considerable overlap in the two niche dimensions, these variables were natural log transformed to construct a two-dimensional niche space to view species overlap. Anolis sagrei is an invader that has apparently integrated into the native Anolis community at this locality. The results of this study are discussed regarding implications for coexistence and potential competition between these three species of anoles. Community structure may arise by both rapid processes occurring in ecological time as well as by more long term processes acting over evolutionary time (Roughgarden et al. 1983; Grant 1986). Communities may be structured through mutual co-adaptation of community members, or by invaders. For an invading species to be successful as a colonist, it must be pre-adapted to integrate in with other members of the community (Rummel & Roughgarden 1983; 1985). Caribbean anoles provide good model systems for analyzing community structure due to their relative simplicity. The extent to which such communities are the result of coevolution among species, however, remains ambiguous (Williams 1972; Roughgarden 1989; Losos 1992a; 1992b). Investigation into the effects of the invasion by the introduced Anolis sagrei may provide some degree of insight into the evolution of Anolis communities within the Caribbean area. The purpose of this study was to examine the spatial niche or structural habitat of three common species of anoles (Anolis grahami, A. sagrei and A. lineatopus) along the northern coast of Jamaica. For a complete description of the three species, refer to Underwood & Williams (1959). Four site localities were chosen for the ease of long term sampling and due to the fact that the flora of each is typical of the area. The "structural habitat" (Rand 1964) of an Anolis species refers to both the flora and other objects upon which these arboreal lizards are found. Rand (1964) originally used perch height and perch diameter to 46 THE TEXAS JOURNAL OF SCIENCE— VOL. 47, NO. 1, 1995 distinguish the spatial niche dimensions utilized by anole species. Other authors (Schoener 1967; Schoener & Schoener 1971; Losos et al. 1993) have also used these measures to quantify habitat use. Spatial habitat use has also been described by perch type and by the degree of insolation (Schoener & Schoener 1971; Loses et al., 1993). Perch height and perch diameter may be the most conveniently quantified measures of structural habitat. This is because these characteristics are correlated with other habitat features and if partitioned between or within species will allow a large number of lizard species to coexist in a relatively small area. For example, in studying five species of anoles at three lowland localities in Jamaica, Schoener & Schoener (1971) found that the anoles had partitioned their spatial niches to reduce spatial overlap and avoid potential intraspecific and interspecific competition. However, such measures may not accurately represent the total niche usage by the community fauna. Anolis sagrei is an invader that is expanding its range across the island of Jamaica (Underwood & Williams 1959; Williams 1969; Mayer 1989), and it is reasonable to expect changes in the anole community. This study was designed to quantify perch height and perch diameter, and compare these distributions among the species. It also examines spatial overlap of these dimensions with plots of habitat space. Materials and Methods Observations were made at the Hofstra University Marine Laboratory campus at Priory, Saint Ann’s Parish, Jamaica. The data were collected on March 10-13, 1983, March 10-13, 1987, and January 4-7, 1994 between 0830-1800 hours from four selected study sites on the campus. Study sites included a beach and three highly disturbed areas. The disturbed areas included laboratory buildings, concrete walls, telephone poles, a barb wire fence, lumber and brick piles, and a low wooden pier. Flora present in all of these areas included red mangrove, coconut, banana trees, and several varieties of cultivated garden plants. Potential habitat availability was similar in all of the selected study sites. Each study site was carefully examined by one or two observers several times during the day. When an anole specimen was observed, its perch height and perch diameter were recorded by tape measure to the nearest cm. Specimens observed on the ground or on walls were assigned a perch height of zero. Only large adult males were used in the study due to the fact that they are larger and more conspicuous due to their social dewlap displays. Several other studies on anoles have 25 20 15 10 5 . 25 20 15 10 5 0 25 20 15 10 5 0 L I six ET AL. 47 a A. grahami C A. lineatopiis 50 100 150 200 250 300 350 400 450 500 550 Perch Height (cm) requency histogram of recorded perch heights for: (a) Anolis grahami (N = points above 6 m were excluded), (b) Anolis sagrei (N = 164), (c) Anolis ? (N = 76). 48 THE TEXAS JOURNAL OF SCIENCE— VOL. 47, NO. 1, 1995 focused exclusively on adult males (Rand & Williams 1969; Williams 1972; 1983). Mean perch heights and perch diameters were calculated for all study sites for all three years. Perch heights were compared using Kol- mogorov-Smirnov Two-sample tests (two-tailed). When the same data were used in pairwise multiple comparisons, the sequential Bonferroni technique (alpha = 0.05) was used to judge statistical significance (Rice 1989). Probabilities reported remain significant with the Bonferroni correction. Also, habitat space plots were constructed from natural log transformed data of perch height and perch diameter to compare overlap in these niche dimensions. Results The frequency distributions of perch heights for each species during all years of the study are graphed in Figure 1. Komolgorov-Smirnov Tests indicate statistically significant differences in perch heights among all three species during all years of the study (A. grahami vs. A. sagrei D = 42.57, n = 197, 164, P< 0.001; A. grahami vs. A, lineatopus D = 22.23, n = 197,76, P<0.001;^. lineatopus A, sagri D = 14.15, n = 76,164, P<0.001). Highly statistically significant differences were found to exist among species in these niche dimensions. Yet, habitat space plots revealed that there is substantial overlap in both perch heights and perch diameters among all three species during the study (Figure 2). The largest overlap in these measures of habitat use was found between the invading species A, sagrei, and the resident species A. lineatopus. Anolis grahami tended to occupy the highest and largest diameter perches when compared to the other two species studied. Discussion The results of this study indicate that there are statistically significant differences among these three species in perch heights, even though our space plots reveal that there is still considerable overlap on these niche axes. Because of this result, it would be useful then to examine other niche axes to determine the coexistence of these species. Statistical differences in perch heights may not be biologically meaningful measures of habitat use due to the fact that they do not adequately or accurately represent the lizard’s structural niche. Other measures such as degree of insolation, and/or finer scale habitat measures may be Perch Diameter (cm) Perch Diameter (cm) Perch Diameter (cm) LANDWER ET AL. 49 Figure 2. Scatterplot of perch height vs. perch diameter for: (a) Anolis grahami (N = 197), (b) Anolis sagrei (N = 164), (c) Anolis lineatopus (N = 76). 50 THE TEXAS JOURNAL OF SCIENCE-VOL. 47, NO. 1, 1995 needed to more adequately quantify habitat use. Questions arise as to the nature of the factors which allow these species to coexist. There may be ecological differences among these species(i.e., A. sagrei may be pre-adapted to such habitats before it invaded). Pre-invasion body-size differences may affect the size of prey items these lizards take (Schoener 1967; Roughgarden 1974; but see Floyd & Jenssen 1983). Other possible unquantified niche axes upon which these species may differ in their overlap include differences in insolation and preferred body temperatures. Anolis lineatopus was observed more frequently in the shade than the other two species, and hence may not compete with the other species for shady perches. Cuban and Floridian specimens of A. sagrei are thermophilic and utilize sunny habitants (Ruibal 1961; Salzburg 1984). Preferred body temperatures may differ and prevent overlap among these three species. All of these pre-invasion factors may be involved in facilitating the coexistence that was observed in this Anolis community. Future studies should involve detailed microhabitat use quantification and account for other niche axes that may help to understand the important ecological and evolutionary factors that are responsible for these faunal assemblages. Such information may be useful in predicting the outcome of invasions on lizard community structure. Finally, it must be stressed that conclusions resulting from this study relative to competitive effects must remain tentative in the absence of data on the entire marked populations over a longer time period, data from con¬ trolled field manipulations, and data on the effects of climate, resource availability and the ongoing habitat disturbances upon these species. Such thorough investigations are necessary to reveal post-invasion responses to competitive interactions among these species. Acknowledgements We are grateful to Melissa Garretson for assistance with lizard observations and perch height measurements, and also the Hofstra Marine Station officials, including Eugene Kaplan and his staff for making this study possible. We thank George Stevens, Jim Brown and Jim Dixon for constructive comments on the manuscript, and George Stevens, Mark Taper and Lee Fitzgerald for their statistical advice. Literature Cited Floyd, H. G., & T. A. Jenssen. 1983. Food habitats of the Jamaican LANDWER ET AL. 51 lizard, Anolis opalinus resource partitioning and seasonal effects excimined. Copeia 1983:319-331. Grant, P. R. 1986. Ecology and Evolution of Darwin’s Finches. Princeton Univ. Press: Princeton. Losos, J. B. 1992a. A critical comparison of the taxon-cycle and character displacement models for the size evolution of Anolis lizards in the Lesser Antilles. Copeia 1992:279-288. _ . 1992b. The evolution of convergent structure in Caribbean Anolis communities. Syst. Biol. 41:403-420. _ , J. C. Marks & T. W. Schoener. 1993. Habitat use and ecological interactions of an introduced and a native species of Anolis lizard on Grand Cayman, with a review of the outcomes of anole introductions. Oecologia 95:525-532. Mayer, G. C. 1989. Deterministic patterns of community structure in West Indian reptiles and amphibians. Ph.D. Dissertation, Harvard University. Rand, A. S. 1964. Ecological distribution in anoline lizards of Puerto Rico. Ecology 45:745-752. _ & E. E. Williams. 1969. The anoles of La Palma: aspects of their ecological relationships. Breviora 327:1-19. Rice, W. R. 1989. Analyzing tables of statistical tests. Evolution 43:223-225. Roughgarden, J. 1974. Niche width: biogeographic patterns among Anolis lizard populations. Am. Nat. 108:429-442. Roughgarden, J., J. D. Rummel & S. W. Pacala. 1983. Experimental evidence of strong present-day competition between Anolis populations of the Anguilla Bank - a preliminary report. Pp. 499- 506, in Herpetology and Evolutionary Biology: Essays in honor of Ernest E. Williams (A. Rhodin and K. Miyata, eds.). Museum of Comp Zool, Cambridge. _ . 1989. The structure and assembly of communities. Pp. 203- 226, in Perspectives in Ecological Theory (J. Roughgarden, R. M. May and S. A. Levin, eds.), Princeton Univ. Press, Princeton. Ruibal, R. 1961. Thermal relations of five species of tropical lizards. Evolution 15:98-111. Rummel, J. D., & J. Roughgarden. 1983. Some differences between invasion- structured and coevolution-structured competitive communities: a preliminary theoretical analysis. Oikos 41:477-486. _ . 1985. A theory of faunal build-up for competition communities. Evolution 39:1009-1033. Salzburg, M. A. 1984. Anolis sagrei and Anolis cristatellus in southern 52 THE TEXAS JOURNAL OF SCIENCE-VOL. 47, NO. 1, 1995 Florida: a case study in interspecific competition. Ecology 65:14-19. Schoener, T. W. 1967. The ecological significance of sexual dimorphism in size of the lizard Anolis consperus. Science 155:474- 478. Schoener, T. W., & A. Schoener. 1971. Structural habitats of West Indian Anolis lizards. Jamaican lowlands. Breviora 368: 1-53. Underwood, G., & E. E. Williams. 1959. The anoline lizards of Jamaica. Bull. Inst. Jamaica, Science Series 9:1-48. Williams, E. E. 1969. The ecology of colonization as seen in the zoogeography of anoline lizards on small islands. Q. Rev. Biol. 44:345-389. _ . 1972. The origin of faunas. Evolution of lizard congeners in a complex island fauna: a trial analysis. Evolutionary Biology 6:47- 89. _ . 1983. Ecomorphs, fauna, island size, and diverse end points in island radiations of Anolis. pp. 326-270, in Lizard Ecology: Studies of a model organism (R. B. Huey, E. R. Pianka & T. Schoener, eds.). Harvard University Press: Cambridge. TEXAS J. SCI. 47(1):53-61 FEBRUARY, 1995 KALLIKREIN-LIKE ENZYME FROM THE VENOM OF CROTALUS BASILISCUS BASILISCUS (SERPENTES: CROTALIDAE) Robert D. Gaffin, Jon B. Scales and Rodney L. Cate Department of Biology and Department of Chemistry, Midwestern State University, Wichita Falls, Texas 76308 Abstract. —A kallikrein-like enzyme was isolated from the venom of Crotalus basiliscus basiliscus (Mexican west coast rattlesnake) and assayed for several physiological properties. Isolation of the enzyme was accomplished with the use of Sephadex G-75, DEAE (diethylaminoethyl) Sephadex A-50, QAE (diethyl [2-hydroxypropyl] aminoethyl) Sephadex A-50 and UPLC Aquapore CX-300 (carboxy methyl) chromatography columns. Fractions were assayed for arginine ester hydrolase activity with the synthetic substrate BAEE (a-N- benzoyl-L-arginine ethyl ester). The purified enzyme exhibited a molecular mass of 31.2 kDa on SDS-PAGE. When added to bovine plasma, the venom enzyme was able to liberate 0.70 moles of bradykinin per mole of enzyme in the isolated rat uterus assay. The venom enzyme, without an exogenous source of kininogen, did not cause an increase in capillary permeability as determined by the Evans blue stain method. The venom enzyme produced a drop of 7 mm Hg in systolic pressure when administered intravenously to a rabbit at 0.07 pgtg body weight. Recent work has yielded kallikrein-like enzymes from the venoms of Crotalus adamanteus (cf. Markland et al. 1982), Crotalus atrox (cf. Bjarnason et al. 1983), Crotalus ruber ruber (cf. Mori & Sugihara 1988; 1989), Crotalus viridis viridis (cf. Komori et al. 1988), Agkistrodon piscivorus piscivorus (cf. Nikai et al. 1988) and Vipera aspis aspis (cf. Komori & Sugihara 1988). Kallikrein is known to possess arginine ester hydrolase activity and thus is able to cleave the ethyl ester group from the synthetic substrate BAEE (a-N-benzoyl-L- arginine ethyl ester). Many of these enzymes have also been functionally characterized for their kallikrein-like activity (i.e., their ability to release kinins, which cause contraction of visceral smooth muscle, an increase in capillary permeability and hypotension). Physiological assays have been devel¬ oped to test each of these effects. This study presents findings on the isolation and functional characterization of a kallikrein-like enzyme from the venom of Crotalus basiliscus basiliscus. Materials and Methods Whole, lyophilized venom was obtained from Latoxan, lot number PA 346. Chromatography resins, a-N-benzoyl-L-arginine ethyl ester HCl (BAEE), bradykinin, kallikrein, bovine plasma and SDS-PAGE 54 THE TEXAS JOURNAL OF SCIENCE— VOL. 47. NO. 1, 1995 molecular weight standards were obtained from Sigma Chemical Company. SDS-PAGE pre-made gels were purchased from Bio-Rad. Solvents used for HPLC procedures were of HPLC grade. All other chemicals were of reagent grade quality. BAEE activity. — BAEE activity assays were performed by mixing 5-25 yiL of fractionated venom with 3 mL of buffered BAEE substrate (0. 1 mg/mL BAEE buffered in a solution of 0. 1 M Tris-HCl, pH 8.0 at ambient temperature). Activity was monitored with a Perkin- Elmer Lambda 7 UV/VIS spectrophotometer at 253 nm. BAEE hydrolase units are reported as the change in absorbance per minute per 25 /xL of enzyme solution. Determination of molecular wdg/zL— SDS-polyacrylamide gel electrophoresis was carried out on 4-20% gradient and 12% non-gradient polyacrylamide Mini- PROTEAN II ready gels. Electrophoresis buffers, voltages, and coomassie brilliant blue staining and destaining were carried out according to instructions from the manufacturer. Kinin-releasing activity assay method for kinin-rel easing activity in the isolated rat uterus was adapted from Erspamer & Erspamer (1962), Trautschold (1970), and Komori et al. (1988). Venom enzyme (5.1 /xg/mL) mixed with an equal volume of bovine plasma was tested for the initiation of contractal events for smooth muscle preparatons. Uterine horn contractions were monitored on a myograph (Model F-60, Narco Bio-Systems) connected to a pen recorder (Narcotrace 40 physiograph. Narco Bio-Systems) set at a sensitivity of 50 mV/cm. Capillary permeability-increasing activity .—Tho, assay method for capillary permeability- increasing activity was originally described by Miles & Wilhelm (1955) and later modified by Miller & Tu (1989), who used mice. Venom enzyme (2.0 and 2.5 pg in 50 /xL), kallikrein (1.2, 2.5 and 25.0 pg in 50 pL) or whole venom (5.2 pg in 50 jLxL) was injected intradermally using a Vi inch-27 gauge needle. Hypotensive activity .—Tht procedure for this assay was adapted from Presley (1984:5-8) using the following Locke’s solution (Lockwood 1961): NaCl, 9.0 g; KCl, 0.42 g; CaCl^, 0.24 g; NaHC03, 0.15 g; glucose, 1.0 g/L; pH 7.3 at ambient temperature. Direct carotid blood pressure was measured in the urethane anesthetized New Zealand white rabbit. The carotid catheter was connected to a linear core pressure transducer (Model P-IOOOB, Narco Bio-Systems) set at 20 mV/cm GAFFIN, SCALES & CATE 55 60 50 40 30 20 10 0 0 400 800 1200 1600 Elution volume (mL) Figure 1. Sephadex G-75 separation of whole venom. Protein absorption at 280 nm ( was monitored continuously and BAEE hydrolase activity as the change in absorbanc( 253 nm/min/mL of fraction ( o ) was determined as indicated. Elution volume (mL) Figure 2. DEAE Sephadex A-50 separation of the BAEE hydrolase fraction from the third Sephadex G-75 rechromatography column (not shown). Protein absorption at 280 nm (— ) and BAEE hydrolase activity ( o) were determined as noted in Fig. 1. KCl concentra¬ tion, M, (A) was determined by conductivity measurements as indicated. BAEE hydrolase units si BAEE hydrolase units 56 THE TEXAS JOURNAL OF SCIENCE— VOL. 47. NO. 1, 1995 sensitivity for direct measurement of blood pressure. Heart rate and respiration was monitored through thoracic needle electrodes coupled to a hi-gain coupler and an impedance pneumograph coupler, respectively. RESULTS Isolation and molecular weight.— 5. g of whole venom was dissolved in 45 mL of 50 mM Tris-HCl, 50 mM KCl buffer (pH 7.2 at 0-4° C), and applied to a Sephadex G-75 column (5.5 x 50 cm) equilibrated with the same buffer. A flow rate of 63 mL/hr at 0-4° C was maintained with the use of a peristaltic pump. Fractions were collected every 15 minutes. Fractions B1 (Fig. 1) were pooled, dialyzed, lyophilized and rechromatographed on a second G-75 column equilibrated with a buffer of greater ionic strength (50 mM Tris, 50 mM KCl, 100 mM NH4AC, pH 7.2 at 0-4 °C). The increase in ionic strength alleviated protein precipitation problems which were observed in the first G-75 column. Fractions containing 59% of the total BAEE hydrolase activity from the second G-75 column were pooled and rechromatographed on a third G-75 column (50 mM Tris-HCl, 50 mM KCl, pH 7.2 at 0-4°C). Fractions containing 92% of the total BAEE hydrolase activity were applied to a DEAE Sephadex A-50 ion exchange column (2.6 x 81 cm) equilibrated with Tris buffer (50 mM Tris-HCl, 20 mM KCl, pH 7.2 at 0-4° C). The sample was eluted with an isocratic wash of approximate¬ ly three bed volumes and then developed with a linear salt gradient to 1 M KCl. Fraction E4 (Fig. 2) was then applied to a QAE Sephadex A-50 ion exchange column (2.6 x 79 cm) equilibrated and developed with buffer and salt gradient identical to the DEAE column. Aliquots of fraction B5 (Fig. 3) were then applied to a an HPLC Aquapore CX-300 cation exchange column (220 x 4.6 mm) equilibrated with 5% buffer A (10 mM Tris-Ac, 0.2 M NH4AC, pH 7.0 at ambient temperature) and 95% buffer B (10 mM Tris-Ac, pH 7.0 at ambient temperature). The column was developed with a linear gradient from 5 % A to 60% A over 25-40 minutes with a flow rate of 1 .0 mL/min and then re-equilibrated with 5 % A and 95 % B for ten minutes prior to the next run. Fractions B6 (Fig. 4), the kali ikrein-1 ike enzyme, were then collected and weighed, SDS-PAGE of the purified enzyme revealed a single band correspond¬ ing to a molecular weight of 31.2 kDa. 57 GAFFIN, SCALES & CATE c D 0 (/) 0 2 >% LU LU < CO Figure 3. QAE Sephadex A-50 separation of the BAEE hydrolase containing fraction ’E4’ from the DEAE Sephadex A-50 separation (see Fig. 2). Protein absorption at 280 nm (— ), BAEE hydrolase activity ( o ) and KCl concentration (A) was determined as noted in Fig. 2. 40 30 20 10 10 15 20 25 Elution volume (mL) 30 35 w ‘c 3 0 C/) 0 O 1— "D LU LU < CO Figure 4. HPLC Aquapore CX-300 chromatographic separation of an aliquote of the BAEE hydrolase containing fraction, B5, from Fig. 3. One of eight runs where protein absorp¬ tion at 280 nm (— ) and BAEE hydrolase activity ( o ) were determined as noted in Fig. 1 and sodium acetate concentration, M, (A) was calculated from the gradient program. 58 THE TEXAS JOURNAL OF SCIENCE- VOL. 47. NO. 1, 1995 Table 1. Isolated rat uterus assay (kinin-releasing activity). Assay substances at 110-130 pmol were mixed with an equi volume of bovine plasma (150 fiL) unless noted otherwise and incubated at 37° C for five minutes. Assay Mixture Uterus Contraction (g) Bradykinine quivalents released (pmol) Moles of bradykinin releas^/mole of substance Enzyme -1- Plasma 0.75 89 0.70 Kallikrein* + Plasma 0.93 210 1.9 Enzyme only 0.00 0 0.00 Kallikrein only 0.00 0 0.00 Plasma only 0.00 0 0.00 * Porcine pancreatic kallikrein; molecular weight = 34 kDa (Fieldler, 1976) Kinin-releasing activity in isolated rat uterus.— Kinin-relesLsing activity assays showed that the venom enzyme (3.8 jxg, 130 picomoles, in 75 jxL), when added to an equal volume of plasma, induced contractions of the isolated rat uterus equal to 0.75 g force (Table 1). Comparison of the bradykinin dose-response curve with the venom enzyme-plasma mix¬ ture indicated that the enzyme produced a contraction force equal to the tissue response for 89 pmol of bradykinin. The enzyme, on a per mole basis, released kinin from the plasma corresponding to approximately 0.70 mol of bradykinin. Capillary-permeability increasing activity .—Inirndermsd injection of venom enzyme, without an exogenous source of kininogen, did not produce an increase in capillary permeability. Hypotensive <3cnV/ry.— Intravenous injections of venom enzyme resulted in a small decrease in mean arterial pressure followed by a rapid recovery (Figure 5). DISCUSSION The purified enzyme exhibited a molecular mass of 3 1 .2 kDa. Recent studies have shown that other kail ikrein-1 ike enzymes from snake venoms have similar masses (Bjarnason et al. 1983; Komori et al. 1988; Mori & Sugihara 1988; Nikai et al. 1988). The enzyme exhibited the ability to release kinin causing visceral smooth muscle contraction. The equivalent of 0.70 moles of bradykinin released per mole of enzyme indicates a moderate kinin-releasing ability. GAFFIN, SCALES & CATE 59 Figure 5. Mean arterial pressure of the rabbit receiving venom enzyme. Intravenous injections were given at time 0 (0.07 iigig body weight) and as indicated by the arrow (0.1 iig/g body weight). An increase in capillary permeability was not seen following direct intradermal injection of either the venom enzyme or plasma kallikrein. The lack of any capillary permeability response may be due to the lack of sufficient endogenous kininogens in the injection area. Intradermal injections of whole venom did produce a noticible blue spot of 0.4 cm^ The whole venom activity is most likely due to the presence of spreading factors such as hemorrhagic enzymes. In the preliminary studies, low concentrations of venom enzyme (0.07 fig enzyme per gram of body weight) produced a 7 mm Hg decrease in systolic pressure. Comparisons of this low response to the results obtained with equal concentrations of bradykinin and kallikrein show that the enzyme does not produce a hypotensive kail ikrein-1 ike response in the rabbit model. Other kail ikrein-1 ike enzymes have been shown to cause sizable decreases in blood pressure resulting in prolonged hypoten¬ sion in rats (Komori & Sugihara 1988; Komori et al. 1988; Mori & Sugihara 1988). These studies report the results from the administration of kallikrein-like enzymes at considerably higher concentrations than reported in this study. Further studies of the kallikrein-like enzyme concentration effects on blood pressure will be of considerable interest. 60 THE TEXAS JOURNAL OF SCIENCE- VOL. 47. NO. 1, 1995 ACKNOWLEDGEMENTS This study was supported in part by a Robert A. Welch Foundation Grant to the Department of Chemistry. A portion of this study was used in partial fulfillment of the Master of Science degree, Midwestern State University, Wichita Falls, Texas. LITERATURE CITED Bjarnason, J. B., A. Barish, G. S. Direnzo, R. Campbell & J. W. Fox. 1983. Kallikrein-like enzymes from Crotalus atrox venom. J. Biol. Chem., 258:12566-12573. Erspamer, V., & G. F. Erspamer. 1962. Pharmacological actions of eledoisin on extravascular smooth muscle. Brit. J. Pharmacol., 19:337-354. Fiddler, F. 1976. Pig pancreatic kallikreins a and b. Pp. 289-303, in Methods in enzymology (L. Lorand, ed.). Academic Press, New York, XLV Part B.XIX 4-1-939 pp. Komori, Y., & H. Sugihara. 1988. Physiological and biochemical properties of a kallikrein-like enzyme from the venom of Vipera aspis aspis (aspic viper). Toxicon, 26:1193-1204. _ , T. Nikai & H. Sugihara. 1988. Biochemical and physiological studies on a kallikrein-like enzyme from the venom of Crotalus viridis viridis (Prairie rattlesnake). Biochimica et Biophysica Acta. , 967:92- 102. Lockwood, A. P. M. 1961. "Ringer" solutions and some notes on the physiological basis of their ionic composition. Comp. Biochem. Physiol., 2:241-289. Markland, F. S., C. Kettner, S. Schiffman, E. Shaw, S. S. Bajwa, K. N. N. Reddy, H. Kirakossian, G. B. Patkos, 1. Theodor, and H. Pirkle. 1982. Kallikrein-like activity of crotalase, a snake venom enzyme that clots fibrinogen. Proc. Natl. Acad. Sci., USA, 79: 1688- 1692. Miles, A. A., & D. L. Wilhelm. 1955. Enzyme-like globulins from serum reproducing the vascular phenomena of inflammation. 1. An activable permeability factor and its inhibitor in guinea-pig serum. Br. J. Exp. Pathol., 36:71-81. Miller, R. A., & A. T. Tu. 1989. Factors in snake venoms that increase capillary permeability. J. Pharm. Pharmcol., 41:792-794. Mori, N., & H. Sugihara. 1988. Kallikrein-like enzyme from Crotalus ruber ruber (red rattlesnake) venom. Int. J. Biochem., 20:1425- 1433. GAFFIN, SCALES & CATE 61 _ & _ . 1989. Characterization of kallikrein-like enzyme from Crotalus ruber ruber (red rattlesnake) venom. Int. J. Biochem., 21:83-90. Nikai, T., K. Imai, M. Nagasaka & H. Sugihara. 1988. Kallikrein-like enzyme from the venom of Agkistrodon p. piscivorus. Int. J. Biochem., 20:1239-1245. Presley, L. N. 1984. Some physiological effects of Crotalus basiliscus venom on the mammalian system. Unpubl. M. S. thesis, Midwestern State University, Wichita Falls, Texas, 48 pp. Trautschold, I. 1970. Assay methods in the kinin system. Pp. 53-81, in Handbook of experimental pharmacology (E.G. Erdos, ed.), Springer-Verlag, Berlin, XXV:XIX-f- 1-768 pp. TEXAS J. SCI. 47(l):62-68 FEBRUARY, 1995 ABUNDANCE AND DIVERSITY OF AQUATIC BIRDS ON TWO SOUTH TEXAS OXBOW LAKES Diane Teter and David L. McNeely Department of Biology, University of Texas at Brownsville, 80 Ft. Brown, Brownsville, Texas 78520 Abstract.— Avian use of two oxbow lakes located in urban surroundings of south Texas was studied for one year. Both abundance and diversity of aquatic birds were notably reduced in the more intensely developed area. Possible factors resulting in these differences are reviewed. Recent studies have investigated the relationship between birds and the diversity of habitat in which they are found. Mac Arthur & Mac Arthur (1961) reported that foliage height diversity was important for the maintenance of bird species diversity. Rotenberry et al. (1979) noted a positive relationship between vegetational structure and avian diversity. Lancaster & Rees (1979) reported a positive linear relationship between avian diversity and foliage height diversity and an inverse relationship between bird species diversity and the man-made component of structural diversity. Bird species diversity and species abundance have also been shown to decrease with urbanization and to increase with foliage height diversity and total vegetation (Emlen 1974; Guthrie 1974; Lancaster & Rees 1979; Cicero 1980). In addition to the above studies, many investigators have specifically examined the resulting impact and relationship that man-made changes in the environment have had upon the avian community (Emlen 1974; Gavareski 1976; Rice et al. 1984; Naik & Parasharya 1987; Derleth et al. 1989). This study was undertaken in an effort to determine the nature of man’s influence as a factor in determining the use of urban lentic bodies of water by aquatic birds in south Texas. Study Area Two oxbow lakes, known locally as resacas, which are located adjacent to the University of Texas at Brownsville campus were selected for observation during the course of this study. The two bodies of water are similar with respect to several physical features. They are less than 200 m apart and separated by a paved roadway and artificially con¬ structed levee. Both are of the same approximate size in surface area (0. 1 km^) and lie within 1 km of the Rio Grande and the international TETER & McNEELY 63 border with Mexico. While similar in many ways, the two bodies of water exhibit differences of major importance. Fort Brown Resaca.—Thi^ horse-shoe shaped lake is located in a highly developed urban area. It adjoins the campus of the University of Texas at Brownsville and is surrounded by residential homes, hotels and offices. The grounds in this area are maintained and characterized by the presence of introduced cultivated plants. Lawns are regularly trimmed and there are extensive paved streets and parking areas. While there is occasional weedy vegetation along the shoreline, most of the lake’s edge consists of man-made vertical retaining walls. A nearly constant water level is maintained by the Brownsville Public Utilities Board with water pumped from the Rio Grande. The minimum water depth is maintained near 0.8 m and the bottom is deep silt. Lozano Banco. — This hook-shaped body of water is located in a former agricultural area which is currently in the early stages of succession to streamside woodland and Tamaulipian brushland. The shoreline is heavily vegetated with both native and introduced vegetation includeing trees and shrubs that form a dense thicket around much of the perimeter. A stand of giant reed grass occupies a small area of the lake margin. While the banks slope naturally around most of the lake, they are very steep adjacent to artificial levees. Lozano Banco receives most of its water from run-off and its level fluctuates. The lake is shallow with a maximum depth of approximately 1 m. The lake bottom slopes gently near the shore and is composed of deep silts. Methods A bird census of both resacas was conducted on a weekly basis. This consisted of 15 min. walking surveys conducted during midday (Mar. 92 - Feb. 93) and hour-long observational periods conducted at dawn (Aug. 92 - Jan. 93). Observations were also made with binoculars from a vantage point that afforded a view of most of the lake surface and perimeter. Estimates of both numbers of birds and numbers of species were based upon visual observations. Diversity was estimated using Shannon’s formula as given by Begon et al. (1990). While both adult and juvenile birds were included in the estimation of numbers of birds (Figure 1), juvenile birds of uncertain identification were not used in determining the numbers of species (Figure 2) nor the bird species diversity (Figure 3). 64 THE TEXAS JOURNAL OF SCIENCE-VOL. 47, NO. 1, 1995 Figure 1 . Total numbers of birds recorded at Fort Brown and Lozano Banco Resacas from February of 1992 through January of 1993. Results During the course of this study 22 species (Table 1) of aquatic birds were recorded present in the study area. Of these, only five species were observed on both bodies of water. These were the Double-crested Cormorant, Neotropic Cormorant, Great Egret, Snowy Egret and Black- bellied Whistling-Duck. The Laughing Gull was observed only on Ft. Brown Resaca. The remaining 16 species were reported only from Lozano Banco. The total numbers of birds observed during the course of this study (Figure 1) were consistently higher at Lozano Banco than at Fort Brown Resaca. These differences were observed to be greatest during the autumn and early winter months due to the presence of large numbers of migrants. The observed differences during October and November were primarily due to the presence of Cattle Egrets. Additionally, abundance was also affected during these months by the overnight roosting of several species of herons in dense thickets of trees overhanging the water. This was not observed at Fort Brown Resaca. The number of species, or species richness, was also noted to be consistently higher (Figure 2) at Lozano Banco in all months except May of 1992 when it was equal for both lakes. During that time, only two TETER & McNEELY 65 ■ FT. BROWN □ LOZANO BANCO Figure 2. Number of bird species recorded at Fort Brown and Lozano Banco Resacas from February of 1992 through January of 1993. 4.5 ^ 3.5 + I- w 3 oc ui i 2.5 O 2 z 1.5 nj li I li FT. BROWN □ LOZANO BANCO Figure 3. Bird species diversity recorded at Fort Brown and Lozano Banco Resacas from February of 1992 through January of 1993. species of aquatic birds were observed to be present on each of the two lakes. Diversity (Figure 3) was also equal to or higher at Lozano Banco than that observed at Fort Brown Resaca during most of the year. The most noteworthy exception to this trend was during the period of 66 THE TEXAS JOURNAL OF SCIENCE-VOL. 47, NO. 1, 1995 Table 1 . Aquatic birds censused on Ft. Brown Resaca and Lozano Banco from February of 1992 through January of 1993. Podilymbus podiceps Tachybaptus dominicus Phalacrocorax auritus** Phalacrocorax brasilianus** Ardea herodias Bubulcus ibis Butorides virescens Casmerodius alb us** Egretta caerulea Egretta thula** Egretta tricolor Nyctanassa violacea Nycticorax nycticorax Anas discors Dendrocygna autumnalis** Fulica americana Gallinula chloropus Earns atricilla* Sterna maxima Ceryle alcyon Ceryle torquata Chloroceryle americana Pied-billed Grebe Least Grebe Double-crested Cormorant Neotropic Cormorant Great Blue Heron Cattle Egret Green Heron Great Egret Little Blue Heron Snowy Egret Tricolored Heron Yellow-crowned Night-Heron Black-crowned Night-Heron Blue- winged Teal Black-bellied Whistling-Duck American Coot Common Moorhen Laughing Gull Royal Tern Belted Kingfisher Ringed Kingfisher Green Kingfisher *Recorded only on Ft. Brown Resaca **Recorded on both Ft. Brown Resaca and Lozano Banco Other species listed were observed only on Lozano Banco overnight heron roosting. The dominance of these birds appears responsible for this exception. Discussion There appears to be a number of man-made factors which could account for or contribute to the differences in numbers of birds, number of species and avian diversity between Fort Brown Resaca and Lozano Banco observed during the course of this study. Human Activity .—Ki Fort Brown Resaca, human traffic is continually high; disturbance from pedestrians, vehicular traffic and recreational use from adjacent facilities all appear disruptive of avian activities. This is not the case for Lozano Banco. Shoreline Vegetation. — The presence of an abundance of both native TETER & McNEELY 67 and introduced vegetation along and near the shoreline of Lozano Banco appears to represent a major contributing factor in the increased avian activity noted in this area. Likewise, the lack of similar vegetation and its replacement with lawns, paved areas and man-made structures in the Fort Brown area affords little usable vertical habitat adjacent to the shoreline. Physical Differences .—Tht nature of the differences in shoreline modification, slope of lake bottom and maintenance of a constant surface water level at Fort Brown Resaca are beyond the scope of this investigation. Likewise are possible differences in bottom sediments and non-avian biota between the two lakes. Additional studies would be required to resolve the effect of these variables upon avian use of the lakes. Conclusions This study was conducted upon only two lakes of noted dissimilarity for only one year’s duration. Considering this and the number of potential variables involved, the results of this study must be considered preliminary. While it appears likely that the man-made physical and biological changes at Fort Brown Resaca are responsible for the observed reduction in avian abundance and diversity, when compared to Lozano Banco, additional studies are required to substantiate this tentative conclusion. This study should be useful in planning future studies. Meanwhile, the results may provide useful information for those in positions of authority who are responsible for the planning and/or modification of existing bodies of water in urban areas with respect to aquatic bird use. Ackno wl edgements This study was supported in part by funding from the University of Texas at Brownsville. We wish to thank Dana Lee Bohne, Rose Farmer, Mike Farmer, Merriwood Ferguson and Rick Teter for their assistance and to Alfred Richardson, Rodney Sullivan and Alfredo Munoz for reviewing an earlier draft of this paper. Literature Cited Begon, Michael, John L. Harper & Colin R. Townsend. 1990. Ecology, Individuals, Populations and Communities, Second Edition. Blackwell Scientific Publications, Boston. 945 pp. 68 THE TEXAS JOURNAL OF SCIENCE- VOL. 47, NO. 1, 1995 Cicero, C. 1980. Avian community structure in a large urban park: controls of local richness and diversity. Landscape and Urban Planning 17:221-240. Derleth, E. L., D. G. McAuley, and T.J. Dwyer. 1989. Avian com¬ munity response to small-scale habitat disturbance in Maine. Canadian Journal of Zoology 67:385-390. Emlen, J. T. 1974. An urban bird community in Tucson, Arizona: derivation, structure, regulation. The Condor 76:184-197. Gavareski, C. 1976. Relation of park size and vegetation to urban bird populations in Seattle, Washington. The Condor 78:375-382. Guthrie, D. A. 1974. Suburban bird populations in Southern California. The American Midland Naturalist 92(2): 46 1-466. Lancaster, R. K. & W. E. Rees. 1979. Bird communities and the structure of urban habitats. Canadian Journal of Zoology 57:2358- 2368. Mac Arthur, Robert M. & John W. Mac Arthur. 1961. On bird species diversity. Ecology 43:594-598. Naik, R. M. & B. M. Parasharya. 1987. Impact of the food availability, nesting-habitat destruction and cultural variations of human settlement on the nesting distribution of a coastal bird Egretta gularis in Western India. Journal of the Bombay Natural History Society 84(2): 350-360. Rice, J., B. W. Anderson & R. D. Ohmart. 1984. Comparison of the importance of different habitat attributes to avian community organization. Journal Wildlife Management 48(3) : 895-9 1 1 . Rotenberry, J. T., R. E. Fitzner & W. H. Rickard. 1979. Seasonal variation in avian community structure: differences in mechanisms regulating diversity. The Auk 96:499-505. TEXAS J. SCI. 47(1) 69 GENERAL NOTES FIRST REPORT OF THE ACANTHOCEPHALAN MACRACANTHORHYNCHUS INGENS FROM THE DOMESTIC DOG CANIS FAMILIARIS IN KANSAS Omar M. Amin, ^Charles L. Kramer, & *Steve J. Upton Institute of Parasitic Diseases, P. O. Box 28372, Tempe, Arizona 85285-8372 and Dept, of Zoology, Arizona State University, Tempe, Arizona 85287-1501 , and *Div. of Biology, Ackert Hall, Kansas State Univ., Manhattan, Kansas 66506-4901 On 21 August 1993 a single adult specimen of the acanthocephalan Macracanthorhynchus ingens (Oligacanthorhychidae) was collected in Manhattan, Kansas. The 10 cm long female specimen was defecated by a two year old Portuguese water dog which weighed 20 kg and was boarded at a residence located in a wooded hillside area of the city. The specimen of M, ingens was stained in acid carmine, preserved in glycerol, and partially dissected to expose the proboscis for diagnostic purposes. It was distended with ovarian balls at various stages of development. The size and stage of maturity of the specimen is indicative of normal development in the host’s intestine. The specimen is deposited with the United States National Museum (USNM 84482). A follow up fecal examination of the canine host, which receives 136 /Ag/month ivermectin orally for heartworm prevention, produced only Giardia sp. cysts. The natural host of M. ingens is the common raccoon Procyon lotor. This animal is an opportunistic omnivore that has a diverse diet which includes insects, frogs, snakes and garbage (Whitney & Underwood 1952; Johnson 1970). Elkins & Nickol (1983) list additional definitive hosts of M. ingens. These include the striped skunk {Mephitis mephitis), mink {Mustela vison) and black bear (Ursus americanus). The domestic dog, Canis familiaris, has also been reported as an occasional definitive host (Fahnestock 1985; Georgi 1992). Beetles assigned to the genera Phyllophaga and Ligyrus (cf Moore 1946), the millipede Narceus americanus (cf. Crites 1964) and woodroaches of the genus Parcoblatta (cf. Elkins & Nickol 1983) are also reported to represent intermediate hosts. Visceral establishment of M. ingens has been reported in frogs of the genus Rana (cf. Moore 1946) and garter snakes of the genus Thamnophis (cf. Elkins & Nickol 1983). A review of acanthocephalan host systems was presented by Schmidt (1985). 70 THE TEXAS JOURNAL OF SCIENCE-VOL. 47, NO. 1, 1995 The area of Manhattan, Kansas where this case occurred affords a wide range of insect- amphibian-reptilian host systems with considerable habitat overlap with raccoons and skunks. Woodroaches, millipedes, beetles, and snakes of the genera Thamnophis and Diadophis are abundant. There are few frogs but the toads Bufo woodhousei and B, cognatus are common. Raccoons and skunks often enter the grounds of residences in Manhattan during hours of darkness and defecate. Clearly, an enzootic population of M. ingens is well established in and around these Manhattan woodlands. The prevalence of M. ingens in raccoons in rural Manhattan is 17.4% (Robel et al. 1989) The canine host is given daily freedom of the grounds and has a history of coprophagy of raccoon feces. This is not, however, regarded as the definitive cause of this reported case. The eggs of M. ingens would first have to develop to an infective cystacanth stage in an intermediate insect host. Although unlikely, it is possible that the specimen of M, ingens was defecated intact by a raccoon and ingested by the dog. The canine host may also have ingested intermediate or paratenic host(s) which are naturally infected with cystacanths. The means of transmission of M. ingens to the canine host remains unclear at this time. Should the intermediate/paratenic hosts prove to be the source of this dog’s infection, then Canis familiaris should be regarded as a natural host and should be added to the list of regular definitive hosts of M. ingens. It is possible that this association has not been commonly observed due to ecological segregation rather than physiological incompatibility. The documentation of an active infection of M. ingens in a 10-month-old child in Texas (Dingley & Beaver 1985) attests to the ability of this acanthocephalan species to infect a wider range of vertebrate hosts than previously reported. Literature Cited Dingley, D. & P. C. Beaver. 1985. Macracanthorhynchus ingens from a child in Texas (USA). Amer. J. Throp. Med. Hyg., 34:918-920. Crites, J. L. 1964. A millipede, Narceus americanus, as a natural intermediate host of an acanthocephala. J. ParasitoL, 50:293. Elkins, C. A, & B. B. Nickol. 1983. The epizootiology of Macra¬ canthorhynchus ingens in Louisiana. J. ParasitoL, 69:951-956. Fahnestock, G. R. 1985. Macracanthorhynchiasis in dogs. Mod. Vet. Pract. Jan., Feb., 31, 81. Georgi, J. R. 1992. Canine Clinical Parasitology. Lea and Febiger, Phil., 217 p. GENERAL NOTES 71 Johnson, A. S. 1970. Biology of the raccoon {Procyon lotor varius Nelson & Goodman) in Alabama. Bull. Auburn Univ. Agr. Exper. St., 402:1-148. Moore, D. V. 1946. Studies on the life history and development of Macracanthorhynchus ingens Meyer, 1933, with a redescription of the adult worm. J. Parasitol., 32:387-399. Robel, R. J., N. A. Barnes, & S. J. Upton. 1989. Gastrointestinal helminths and protozoa from two raccoon populations in Kansas. J. Parasitol . , 75:1 000- 1 003 . Schmidt, G. D. 1985. Development and life cycles. In Biology of Acanthocephala, D. W. T. Crompton & B. B. Nickol (eds.). Cambridge Univ. Press, London, pp. 273-305. Whitney, L. F. & A. B. Underwood. 1952. The Raccoon. Practical Science Publ, Co., Orange, Connecticut, 177 pp. PLACOBDELLA PARASITICA (RHYNCHOBDELLIDA: GLOSSIPHONIIDAE) FROM THE EASTERN RIVER COOTER (CHELONIA: EMYDIDAE) IN OKLAHOMA William E. Moser School of Biological Sciences, University of Nebraska-Lincoln, Lincoln, Nebraska 68588-0118 Present address: Department of Invertebrate Zoology, Division of Worms, National Museum of Natural History, Smithsonian Institution, Washington, D. C. 20560 While the large glossiphoniid leech Placobdella parasitica (Say 1824) has been commonly reported throughout much of North America, distributional data on this species from the southern and western states remains poorly known (Sawyer 1972; Klemm 1982). On 15 March 1994, three adult specimens of P. parasitica were collected from the plastron of a single male specimen of the eastern river cooter Pseudemys concinna concinna (LeConte 1830) from Lake Texoma, Marshall County, Oklahoma. The length of the turtle’s carapace was 21.5 cm and its width was 15.5 cm. All three leech specimens were dorsoventrally flattened with a greenish-brown dorsum, cream-colored mid-dorsal band of variable width, and irregular lateral patches. The ventrum of each exhibits 8 to 12 bluish-green longitudinal stripes. Specimens measured 2.0 cm, 3.2 cm and 3.5 cm in length. The crop ceca of all three specimens contained blood, indicating a natural association with the turtle. Two specimens (HWML 37846) are 72 THE TEXAS JOURNAL OF SCIENCE— VOL. 47, NO. 1, 1995 Table 1. Reported turtle hosts of Placobdella parasitica. TURTLE SPECIES REFERENCES Chelydra serpentina serpentina Chelydra serpentina osceola Macroclemys tenvninckii Sternotherus depressus Sternotherus odoratus Kinosternon subrubrum Clemmys guttata Clenvnys insculpta Graptemys geographica Graptemys pseudogeographica Trachemys scripta Trachemys scripta elegans Pseudemys concinna concinna Pseudemys nelsoni Chrysemys picta picta Chrysemys picta marginata Emydoidea blandingii Sawyer 1972, 1986 Ernst & Barbour 1972 Ernst & Barbour 1972 Dodd 1988 Ryerson 1915; Sawyer 1972 Sawyer & Shelley 1976 Ryerson 1915; Sawyer 1972 Koffler et al. 1978; Ricciardi & Lewis 1991 Say 1824; Sawyer 1972 Sawyer 1986 Martin 1972; Sawyer 1972 Hendricks et al. 1971 This study Ernst & Barbour 1972 Sawyer 1972, 1986 Ryerson 1915; Sawyer 1972; Amin 1981; Ricciardi & Lewis 1991 Sawyer 1972; Amin 1981 deposited with the holdings of the H. W. Manter Parasitology Laboratory, University of Nebraska State Museum. This report of Placobdella parasitica represents a new addition to the leech fauna of Oklahoma. It also represents the first record of this species from Pseudemys concinna concinna. This leech is commonly reported to blood-feed on Chelydra serpentina and Chrysemys picta turtles (Sawyer 1972, 1986; Amin 1981), which may represent the preferred hosts for this leech. Placobdella parasitica has now been reported from 17 different species and subspecies of turtles (Table 1) as well as Rana pipens (cf. Meyer & Moore 1954), indicating that P, GENERAL NOTES 73 parasitica blood-feeds upon a wide variety of turtles and possibly other hosts. Acknowledgements The assistance of Dr. John D. Lynch with the turtle identification, the turtle collection by Scott Snyder and the critical reviews of the manuscript by Dr. Duane Hope and Dennis J. Richardson are gratefully appreciated. Literature Cited Amin, O. M. 1981. Leeches (Hirudinea) from Wisconsin, and a description of the spermatophore of Placobdella omata. Trans. Am. Microsc. Soc., 100:42-51. Dodd, C. K. Jr. 1988. Patterns of distributional and seasonal use of the turtle Stemotherus depressus by the leech Placobdella parasitica J. Herpetol., 22:74-81. Ernst, C. H., & R. W. Barbour. 1972. Turtles of the United States. The University Press of Kentucky, Lexington, 347 pp. Hendricks, A. C., J. T. Wyatt & D. E. Henley. 1971. Infestation of a Texas red-eared turtle by leeches. Tex. J. Sci., 22:247. Klemm, D. J. 1982. Leeches (Annelida: Hirudinea) of North America. US Environmental Protection Agency, Cincinnati, 177 pp. Koffler, B. R., R. A. Seigel & M. T. Mendonca. 1978. The seasonal occurrence of leeches on the wood turtle, Clemmys insculpta (Reptilia, Testudines, Emydidae). J. Herpetol., 12:571-572. Martin, D. R. 1972. Distribution of helminth parasites in turtles native to southern Illinois. Trans. Ill. Acad. Sci., 65(3/4): 61 -67. Meyer, M. C., & J. P. Moore. 1954. Notes on Canadian leeches (Hirudinea), with the description of a new species. Wasmann J. Biol., 12:63-96. Ricciardi, A., & D. J. Lewis. 1991. New records of freshwater leeches (Annelida: Hirudinea) from Quebec. Can. Field-Nat., 105:368-371. Ryerson, C. G. S. 1915. Notes on the Hirudinea of Georgian Bay. Contributions to Canadian Biology, Sessional Paper No. 39b. 5 George V,pp. 165-175. Sawyer, R. T. 1972. North American freshwater leeches, exclusive of the Piscicolidae, with a key to all species. Ill. Biol. Monogr., 46:1- 154. 74 THE TEXAS JOURNAL OF SCIENCE™ VOL. 47, NO. 1, 1995 . 1986. Leech biology and behaviour. Vols. 1-3. Oxford University Press, Oxford, 1065 pp. _ & R. M. Shelley. 1976. New records and species of leeches (Annelida: Hirudinea) from North and South Carolina. J. Nat. Hist., 10:65-67. Say, T. 1824. Keating’s narrative of an expedition to the source of St. Peter’s River, Lake Winnepeek, Lake of the Woods, etc., in 1823, under S. H. Long, 2 Vols. Philadelphia. Vol. II Appendix, D. Class Vermes, pp. 14-16. TEXAS J. SCI. 47(1):75 FEBRUARY, 1995 BOOK REVIEW Birds and Other Wildlife of South Central Texas: A Handbook by Edward A. Kutac & S. Christopher Caran. University of Texas Press, 1994, First Edition, 203 pages, 28 maps, glossary, selected references, indices. ISBN 0-292-75550-3 HB, ISBN 0-292-74315-7 PBK. Although E. A. Kutac and S. C. Caran are listed as authors of this book, five of the eight chapters are actually authored in part or exclusively by others. This is an excellent guide to the natural history of the 19 county area covered in this handbook. The preliminary pages on geological history, climate, ecology and locations of interest are very well done and should serve as a source for professionals as well as for nonprofessionals. The chapter on birds is well done. The intent of the bird list is to inform about abundance, distribution, and timing of species’ occurrence. It does this well without elaborating on vouching details. Without details on dates of occurrence or other confirming evidence for the status of many species, the list cannot be used as a technical source. But this was not the intent of the authors. The chapter on mammals represents a useful compilation of mammal species occurring within the defined area of South Central Texas. The scientific names and sequence for species of mammals follows a recent checklist of North American mammals, but no explanation is given for the use of subspecific names for several bats and only species designations for the remainder of the mammals. Although it is unclear how the authors’ determined relative abundance for mammal species, most assessments appear accurate. These minor considerations notwith¬ standing, the authors have succeeded in elucidating the status of the mammalian fauna within this region of Texas. The annotated checklist of amphibians and reptiles is a compendium of useful information regarding species abundance, habitat and distribu¬ tion. There are no errors worth noting and the classification is up to date. The checklist should be useful to both professionals and non¬ professionals. Additionally, there are chapters on fishes, land snails and butterflies of this area of Texas. Terry C. Maxwell Robert C. Dowler Raymond Stone, Jr. 76 THE TEXAS JOURNAL OF SCIENCE-VOL. 47, NO. 1, 1995 INSTRUCTIONS TO AUTHORS Scholarly manuscripts in any field of science or technology will be considered for publication in The Texas Journal of Science. Prior to acceptance, each manuscript will be reviewed by both knowledgeable peers and the editorial staff. Authors are encouraged to suggest the names and addresses of two potential reviewers to the Manuscript Editor at the time of submission of their manuscript. No manuscript submitted to the Journal is to have been published or submitted elsewhere. Upon completion of the peer review process, the corresponding author is required to submit the final revised manuscript in both typed (two copies) and 3.5 inch computer diskette forms. The diskette must be compatible with IBM or Macintosh computers. Format Except for the corresponding author’s address, manuscripts must be double-spaced throughout (including legends and literature cited), pages numbered, and submitted in TRIPLICATE (typed or photocopied) on 8.5 by 11 inch bond paper, with margins of approximately one inch. Computer generated manuscripts must be reproduced as letter quality or laser prints, not dot matrix. Do not justify the right margin. Do not break words at the right margin. The text can be subdivided into sections as deemed appropriate by the author(s). Possible examples are: Abstract; Materials and Methods; Results; Discussion; Summary or Conclusions; Acknowledgments; Literature Cited. Major internal headings are centered and capitalized. References Cite all references in the text by author and date in chronological (not alphabetical) order; Jones (1971); Jones (1971; 1975); (Jones 1971); (Jones 1971; 1975); (Jones 1971; Smith 1973; Davis 1975); Jones (1971), Smith (1973), Davis (1975); Smith & Davis (1985); (Smith & Davis 1985). If more than two authors, use Jones et al. (1976) or (Jones et al. 1976). Citations to publications by the same author(s) in the same year should be designated alphabetically (1979a; 1979b). AUTHOR GUIDELINES 77 Literature Cited Journal abbreviations in the Literature Cited section should follow those listed in BIOSIS Previews ® Database (ISSN: 1044-4297). This volume is present in all libraries receiving Biological Abstracts. Ask your interlibrary loan officer or head librarian. If not available, then use standard recognized abbreviations in the field of study. Be certain that all citations in the text are included in the Literature Cited section and vice versa. Consecutively-paged journal volumes and other serials should be cited by volume, number, and pagination. Serials with more than one number and that are not consecutively paged should be cited by number as well (Smithson. Misc. Coll., 37(3): 1-30). The following are examples of a variety of citations: Journals & Serials. — Jones, T. L. 1971. Vegetational patterns in the Guadalupe Mountains, Texas. Am. J. Bot., 76(3):266-278. Smith, J. D. 1973. Geographic variation in the Seminole bat, Lasiurus seminolus J. Mammal., 54(l):25-38. _ & G. L. Davis. 1985. Bats of the Yucatan Peninsula. Occas. Pap. Mus., Texas Tech Univ., 97:1-36. Books. — Jones, T. L. 1975. An introduction to the study of plants. John Wiley & Sons, New York, xx -1-386 pp. _ , A. L. Bain, & E. C. Burns. 1976. Grasses of Texas. Pp. 205- 265, in Native grasses of North America (R. R. Dunn, ed.), Univ. Texas Studies, 205:xx+ 1-630. Unpublished. — Davis, G. L. 1975. The mammals of the Mexican state of Yucatan. Unpublished Ph.D. dissertation, Texas Tech Univ. , Lubbock, 396 pp. In the text of the manuscript, the above unpublished reference should be cited as Davis (1975) or (Davis 1975). Do not make citations to unpublished material that cannot be obtained nor reviewed by other investigators (such as unpub. or unpub. field notes). 78 THE TEXAS JOURNAL OF SCIENCE- VOL. 47, NO. 1, 1995 Graphics, Figures & Tables All tables must be included as a computer generated addendum or appendix of the manuscript. Computer generated figures and graphics must be laser quality and camera ready, reduced to 5.5 in. (14 cm) in width and not exceed 8.5 in. (20.5 cm) in height. Shading is unacceptable. Instead, use different and contrasting styles of crosshatching, grids, line tints, dot size, or other suitable matrix to denote differences in graphics or figures. Figures, maps and graphs should be reduced to the above graphic measurements by a photographic method. A high contrast black and white process known as a PMT or Camera Copy Print is recommended. Authors unable to provide reduced PMT’s should submit their originals. They will be charged $10 per reduction for this service. Figures and graphs which are too wide to be reduced to the above measurements may be positioned sideways. They should then be reduced to 9 in. (23 cm) wide and 5 in. (12.5 cm) in height. Black and white photographs of specimens, study sites, etc. should not exceed 8 in. in width and be mounted on 8.5 by 11 in. paper or backing. Color photographs cannot be processed at this time. Each figure should be marked on the back with the name of the author(s) and figure number. If confusion might result as to arrangement of a figure, label "top". All legends for figures and tables must be typed (double¬ spaced) on a sheet(s) of paper separate from the text. All figures must be referred to in text as "Figure 3" or "(Fig. 3)"; all tables as "Table 3" or "(Table 3)". Galley Proofs & Reprints The principal author will receive galley proofs along with edited typescript. Proofs must be corrected and returned to the Managing Editor within five days; failure to promptly return corrected galley proofs will result in delay of publication. The Academy will provide 100 reprints without charge for each feature article or note published in the Journal. These will be mailed to the senior author or other such designated contact person following the publishing of each issue of the Journal. The distribution of reprints among co-authors is the responsibility of the senior author. Authors will have the opportunity to purchase additional reprints (in lots of 100) at the time that the corrected galley proofs are returned to the Managing Editor. AUTHOR GUIDELINES 79 Page Charges Authors are required to pay $50 per printed page. While members of the Academy are allowed four published pages per year free of charge on publications with one or two authors, all authors with means or institutional support are requested to pay full page charges. Full payment is required for those pages in excess of four. In the case of the publication of a manuscript with two authors which is granted four free pages, both author and co-author of the article are considered to have expired their allotment of free pages for that calendar year. All publications authored by three or more persons are required to pay full page charges. Nonmembers of the Academy are required to pay full page charges for all pages. The Academy, upon written request, will subsidize a limited number of contributions per volume. These exceptions are, however, generally limited to students without financial support. Should a problem arise relative to page charges, please contact Dr. Michael J. Carlo at Angelo State University. Additional Guidelines An expanded version of the above author guidelines which includes instructions on style, title and abstract preparation, deposition of voucher specimens, and a listing of standardized abbreviations is available upon request from: Dr. Jack D. McCullough TAS Manuscript Editor Department of Biology - Box 13003 Stephen F. Austin State University Nacogdoches, Texas 75962 THE TEXAS ACADEMY OF SCIENCE Membership Information and Application MEMBERSHIP. — Any person or members of any group engaged in scientific work or interested in the promotion of science are eligible for membership in The Texas Academy of Science. Dues for regular members are $30.00 annually; supporting members, $60.00; sustaining members, $100.00; patron members, $150.00; associate (student) members, $15.00; family members, $35.00; affiliate members, $5.00; emeritus members, $10.00; life members, 20 times annual dues; corporate members, $250.00 annually; corporate life members, $2000.00 in one payment. Library subscription rate is $50.00 annually. Application for Membership (please print or type) Name _ Last First Middle Mailing Address _ City _ State _ Zip Type of Membership _ Send Application Form and Check or Money Order to: Dr. Brad C. Henry TAS Executive Secretary Department of Biology University of Texas-Pan American Edinburg, Texas 78539 Please photocopy this Application Form for additional copies. THE TEXAS ACADEMY OF SCIENCE, 1994-95 Officers President: President-Elect: Vice-President: Immediate Past President: Executive Secretary: Corresponding Secretary: Manuscript Editor: Managing Editors: Treasurer: A A AS Council Representative: Ned E. Strenth, Angelo State University Donald E. Harper, Texas A&M University at Galveston Kenneth L. Dickson, University of North Texas David Buzan, Texas Parks and Wildlife Department Brad C. Henry, University of Texas-Pan American Donald E. Harper, Texas A&M University at Galveston Jack D. McCullough, Stephen F. Austin State University Michael J. Carlo, Angelo State University Ned E. Strenth, Angelo State University Michael J. Carlo, Angelo State University Sandra S. West, Southwest Texas State University Directors 1992 Joe C. Yelderman, Jr., Baylor University David D. Diamond, Texas Parks and Wildlife Department 1993 Dovalee Dorsett, Baylor University Tom Conry, Brazos River Authority 1994 Neil Ford, University of Texas at Tyler Barbara ten Brink, Texas Education Agency Sectional Chairpersons Biological Science: Robert S. Baldridge, Baylor University Botany: Fisayo Jejelowo, Texas Southern University Chemistry: William B. Harrell, Texas Southern University Computer Science: James Lamiell, Brooke Army Medical Center Conservation: Warren Pulich, Jr., Texas Parks & Wildlife Environmental Science: Tom Vaughan, Texas A&M International University Freshwater and Marine Science: Earl Chilton, II, Texas Parks & Wildlife Geography: Dafrel McDonald, Stephen F. Austin State University Geology: James O. Jones, University of Texas at San Antonio Materials Science: Dennis W. Mueller, University of North Texas Mathematics: Ben Sultenfuss, Stephen F. Austin State University Science Education: Barbara ten Brink, Texas Education Agency Systematics and Evolutionary Biology: Jeff Pittman, Lamar University Terrestrial Ecology: Norman Horner, Midwestern State University Counselors Collegiate Academy: Jim Mills, St. Edward’s University Junior Academy: Kathy Mittag, Sam Houston State University 2nd CLASS POSTAGE PAID AT LUBBOCK TEXAS 79401 BOUND PRINTED MATERIAL RETURN POSTAGE GUARANTEED THE TEXAS JOURNAL OF SCIENCE Box 43151, Texas Tech University Lubbock, Texas 79409-3151, U.S.A. library acquisitions SMITHSONIAN INSTITUTE room 25 NHB WASHINGTON, DC 20560 GENERAL INFORMATION MEMBERSHIP.— Any person or members of any group engaged in scientific work or interested in the promotion of science are eligible for membership in The Texas Academy of Science. Dues for regular members are $30.00 annually; supporting members, $60.00; sustaining members, $100.00; patron members, $150.00; associate (student) members, $15.00; family members, $35.00; affiliate members, $5.00; emeritus members, $10.00; life members, 20 times annual dues; corporate members, $250.00 annually; corporate life members, $2000.00 in one payment. Library subscription rate is $50.00 annually. The Texas Journal of Science is a quarterly publication of The Texas Academy of Science and is sent to most members and all subscribers. Payment of dues, changes of address and inquiries regarding missing or back issues should be sent to Dr. Brad Henry, Department of Biology, The University of Texas-Pan American, Edinburg, Texas 78539. AFFILIATED ORGANIZATIONS American Association for the Advancement of Science, Texas Council of Elementary Science Texas Section, American Association of Physics Teachers Texas Section, Mathematical Association of America Texas Section, National Association of Geology Teachers Texas Society of Mammalogists The Texas Journal of Science (ISSN 0040-4403) is published quarterly at Lubbock, Texas U.S. A. Second class postage paid at Post Office, Lubbock, Texas 79402. Postmaster: Send address changes, and returned copies to The Texas Journal of Science, Box 43151, Texas Tech University, Lubbock, Texas 79409-3151, U.S. A. THE TEXAS JOURNAL OF SCIENCE Volume 47, No. 2 May 1995 CONTENTS Helminth parasites of unisexual and bisexual whiptail lizards (Teiidae) in North America. IX. The plateau spotted whiptail {Cnemidophorus gularis septemvittatus) . By Chris T. McAllister, James E. Cordes and James M. Walker . . 83 Cranial and dental variation in the nine-banded armadillo, Dasypus novemcinctus , from Texas and Oklahoma. By Frederick B. Stangl, Jr. , Stacie L. Beauchamp and N. Gayle Konermann . 89 Distributional records of small mammals from the southwestern Rolling Plains of Texas. By Franklin D. Yancey, II, Jim R. Goetze, Burhan M. Gharaibeh and Clyde Jones . 101 Hydrocarbon degrading bacteria at Oil Springs, Texas. By Thomas G. Benoit and Robert J. Wiggers . 106 A new subspecies of the polytypic lizard sp^ies Sceloporus undulatus (Sauria: Iguanidae) from northern Mexico. By Hobart M. Smith, David Chiszar and Julio A. Lemos-Espinal . 117 Sums of products of positive integers. By David R. Cecil . 144 Growth models of loblolly pine from east Texas and Lx)st Pines seed sources growing in the Post Oak Belt of Texas. By W. David Hacker and M. Victor Bilan . . 151 Hybridization among members of the genus Morone (Pisces: Percichthyidae) in Galveston Bay, Texas. By Rocky Ward, Ivonne R. Blandon and Britt W. Bumguardner . 155 General Note An additional record of the native American elk (Cervus elaphus) from north Texas. By Brian S. Shaffer, Bonnie C. Yates and Barry W. Baker . 159 THE TEXAS JOURNAL OF SCIENCE EDITORIAL STAFF Manuscript Editor: Jack D. McCullough, Stephen F. Austin State University Managing Editor: Ned E. Strenth, Angelo State University Associate General Editor: Michael J. Carlo, Angelo State University Associate Editor for Botany: Robert 1. Lonard, The University of Texas — Pan American Associate Editor for Chemistry: John R. Villarreal, The University of Texas — Pan American Associate Editor for Geology: M. John Kocurko, Midwestern State University Associate Editor for Mathematics and Statistics: E. Donice McCune, Stephen F. Austin State University Associate Editor for Physics: Charles W. Myles, Texas Tech University Manuscripts intended for publication in the Journal should be submitted in TRIPLICATE to: Jack D. McCullough TAS Manuscript Editor Department of Biology - Box 13003 Stephen F Austin State University Nacogdoches, Texas 75962 Scholarly papers in any field of science, technology, or science education will be considered for publication in The Texas Journal of Science. Instructions to authors are published one or more times each year in the Journal on a space-available basis, and also are available from the Manuscript Editor at the above address. The Texas Journal of Science is published quarterly in February, May, August, and November for $30 per year (regular membership) by The Texas Academy of Science. Second-class postage rates (ISSN 0040-4403) paid at Lubbock, Texas. Postmaster: Send address changes, and returned copies to The Texas Journal of Science, Box 43151, Texas Tech University, Lubbock, Texas 79409-3151, U.S.A. TEXAS J. SCI. 47(2): 83-88 MAY, 1995 HELMINTH PARASITES OF UNISEXUAL AND BISEXUAL WHIPTAIL LIZARDS (TEIIDAE) IN NORTH AMERICA. IX. THE PLATEAU SPOTTED WHIPTAIL (CNEMIDOPHORUS GULARIS SEPTEMVITTATUS) Chris T. McAllister, James E. Cordes and James M. Walker Division of Natural and Applied Sciences, Cedar Valley College, 3030 N. Dallas Avenue, Lancaster, Texas 75134-3799; Division of Sciences, Louisiana State University at Eunice, P. O. Box 1129, Eunice, Louisiana 70535; and Department of Biological Sciences, University of Arkansas, Fayetteville, Arkansas 72701 Abstract.— Sixteen percent of 83 specimens of the plateau spotted whiptail lizard, Cnemidophorus gularis septemvittatus Cope 1892, were found to be infected with one or more of six species of parasitic helminths. One percent were found with Oochoristica bivitellobata Loewen 1940, eight percent with tetrathyridia of Mesocestoides sp. Vaillant 1863, one percent with larval Physaloptera sp. Rudolph! 1819, five percent with Parathelandros texanus Specian & Ubelaker 1974, two percent with Pharyngodon warneri Harwood 1932, and six percent with oligacanthorhynchid acanthocephalan cystacanths. This report, the ninth in a series on helminths of the genus Cnemidophorus, reports new host records for Oochoristica bivitellobata, Physaloptera sp., Pharyngodon warneri, and oligacanthorhynchid acanthocephalan cystacanths. The plateau spotted whiptail, Cnemidophorus gularis (= septem¬ vittatus) septemvittatus Cope 1892, is a large bisexual teiid lizard that ranges from the Big Bend region of Texas south through much of Chihuahua and Coahuila, Mexico (Conant & Collins 1991). This lizard is known from only six counties of southwestern Texas (Dixon 1987; Cordes et al. 1990) where it frequents rocky outcroppings with sparse vegetation in rugged mountainous regions and desert foothills. Little is known about the natural history and ecology of this lizard (Wright & Vitt 1993). Likewise, little has been documented on its parasites. Specian & Ubelaker (1974a; 1974b) described two species of oxyurid nematodes from C. gularis septemvittatus (syn. C. scalaris) in Brewster County, Texas. In addition, McAllister et al. (1991c) reported Mesocestoides sp. tetrathyridia in C. gularis septemvittatus from Presidio County, Texas. To our knowledge, no additional information has been published on parasites of this lizard. The objective of this paper, the ninth in a series of reports on helminth parasites of Cnemidophorus (McAllister 1990a; 1990b; 1990c; 1990d; 1992; McAllister et al. 1991a; 1991b; 1991c), is to provide additional data on the identity, prevalence, and intensity of parasites infecting C. gularis septemvittatus from Texas. 84 THE TEXAS JOURNAL OF SCIENCE— VOL. 47, NO. 2, 1995 Materials and Methods Eighty- three juvenile and adult specimens of C. gularis septemvittatus, including 44 males and 39 females (mean ± 1 SD for snout-vent length (SVL) = 86.1 ± 16. 1, range 47-1 15 mm) were collected between June to September 1989 and again during May 1990. Most specimens (N = 78) came from "Campo Nuevo" in San Antonio Canyon, 45 km N of Presidio, Presidio Co., Texas (29°53’N, 104°29’W, elevation 900 to 1500 m). The remainder were from sites in Brewster (N = 2), Pecos (N = 1), and Terrell (N = 2) counties. Lizards were killed with .22- caliber rat shot or immobilized with rubber bands and euthanized by overdose with sodium pentobarbital. Lizards were field preserved in 10 percent formalin and stored in 70 percent ethanol until examination. Detailed methods for processing hosts and parasites have been previously described by McAllister (1990a). Representative helminths have been deposited in the United States National Museum (USNM) Parasite Collection (USDA) Belts ville, Maryland 20705, as follows: Oochoristica bivitellobata (USNM 83364), Mesocestoides sp. (USNM 83363), Physaloptera sp. (USNM 83366), Parathelandros texanus (USNM 83367), Pharyngodon wameri (USNM 83368), acanthocephalan cystacanths (USNM 83365). Voucher speci¬ mens of C. gularis septemvittatus are deposited in the University of Arkansas Department of Zoology Collection (UADZ). Results and Discussion Thirteen of 83 (16%) specimens of C. gularis septemvittatus were found to be infected with at least one kind of helminth (Table 1). Seven specimens (54%) harbored a single parasite whereas multiple infections were observed in six specimens (46%), including two helminths in five lizards and one lizard with three species. Only lizards from Presidio County harbored parasites. On the average, mature and larger lizards exhibited a higher prevalence of infection as the overall SVL’s of infected (102.2 + 9.4, 84-115 mm, N = 13) and uninfected lizards (83.1 ± 15.2, 47-115 mm, N = 70) were significantly different (Student’s unpaired t-test, t = 6.01, 81 df, P < 0.0005). If juveniles (60.1 ± 8.8, 47-69 mm, N = 13), all of which did not harbor parasites (Table 2), are excluded from the analysis, SVL’s of infected and uninfected adults (88.5 ± 10.6, 70- 114 mm, N = 57) remain significantly different (t = 4.63, 68 df, P < 0.0005). There was nearly a three-fold difference in prevalence of McAllister, cordes & walker 85 Table 1. Helminths found in CnemMophorus gularis septemvittatus from southwestern Texas. Helminth Site of infection ftevalence* Cestoidea Cyclophyllidea Mesocestoides sp.** coelom, heart, liver, lungs, intestines, mesenteries, ovaries, stomach 7/83 (8%) Oochoristica bivitellobata*** duodenum 1/83(1%) Nematoda Spimrida Physaloptera sp.*** stomach 1/83(1%) Oxyurida Parathelandros texanus rectum 4/83 (5%) Pharyngodon warneri*** colon, rectum 2/83 (2%) Acanthocephala Oligacanthorhynchida unidentified cystacanths*** muscle fascia 5/83 (6%) *Number infected/ number examined (percent). **Mesocestoides sp. tetrathyridia (Cestoidea: Cyclophyllidea) reported previously from 6/70 (9 percent) C gularis septemvittatus by McAllister et al. (1991c). ***New host record. infection among the sexes as 23 percent of all males and eight percent of all females harbored parasites (Table 2). When SVL’s of these 13 infected lizards are compared, adult male SVL’s (105.9 + 6.0, 98-115 mm, N = 10) were significantly different (t = 2.96, 11 df, P < 0.005) than SVL’s of adult females (90.0 ± 8.7, 84-100 mm, N = 3). As for within sexual comparisons, SVL’s of infected adult males were significantly different (t = 5.01, 37 df, P < 0.0005) than uninfected adult males (91.8 ± 11.2, 70-114 mm, N = 29) whereas SVL’s of infected adult females were not significantly different (t ^ 0.924, 29 df, P > 0.20) than SVL’s of uninfected adult females (85.1 ± 9.0, 70-106 mm, N = 28). Three third-stage larval spirurid nematodes, Physaloptera sp. Rudolphi 1819, were found in a single male whiptail (UADZ 3967, SVL = 111 mm) collected in May 1990. Although prevalence of infection and intensities may be low in some whiptails, as is the case here, these 86 THE TEXAS JOURNAL OF SCIENCE-VOL. 47, NO. 2, 1995 Table 2. Prevalence of helminths infecting different age, sex, and size classes of C. gularis septemvittatus from southwestern Texas. Age, sex, and size class* Prevalence** Juveniles (47-68 mm SVL) 0/13 (0%) Adult females (70-106 mm SVL) 3/31 (10%) Adult males (70-115 mm SVL) 10/39 (26%) All females (47-106 mm SVL) 3/38 (8%) All males (52-115 mm SVL) 10/45 (22%) All adults (70-115 mm SVL) 13/70 (19%) All lizards (47-115 mm SVL) 13/83 (16%) *Lizards that reached SVL’s > 70 mm were considered sexually mature. **Number infected/ number examined (percent). spirurids have been reported numerous times from Cnemidop horns spp. (cf. McAllister et al. 1992). Ten oxyurid nematodes, Pharyngodon wameri Harwood 1932, each were found in two adult male C. gularis septemvittatus (UADZ 3636, 3766, SVL = 99, 105 mm) collected in June 1989 and May 1990. This nematode is a common component of the helminth fauna of whiptail lizards (McAllister 1990d, 1991a). Two linstowiid tapeworms, Oochoristica bivitellobata Loewen 1940, were found in a male C. gularis septemvittatus (UADZ 3625, SVL = 100 mm) collected in June 1989. This cestode is another common parasite of Cnemidop horns spp., although prevalence of infection may be quite low (McAllister et al. 1991b). Six oxyurid nematodes, Parathelandros texanus Specian & Ubelaker 1974, were found in four (three males, one female) C gularis septemvittatus (UADZ 3625, 3634, 3646, 3658, SVL = 99.0 ± 10.4, range 84-107 mm) collected in June and July 1989; mean intensity was 1.5 ± 0.6 worms. The type locality of P. texanus is in nearby Brewster County (Specian & Ubelaker 1974a), approximately 150 km southeast of the present locale in Presidio County. Specian & Ubelaker (1974a) previously reported P. texanus from C. gularis septemvittatus (syn. C scalaris). It has also been reported from three of 58 (5%) C. dixoni at the Campo Nuevo site (McAllister et al. 1991a) as well as three of 27 (11%) C. tesselatus from Presidio County, Texas (McAllister 1990a), McAllister, cordes & walker 87 one of 289 (0.3%) C gularis from Jeff Davis County, Texas (McAllister 1990d), and one of 23 (4%) C. flagellicaudus from Catron County, New Mexico (McAllister 1992). Other hosts and localities include C. inomatus, C. tigris (=marmoratus) , Sceloporus merriami, S. undulatus, Cophosaurus texanus, and Urosaurus omatus from Presidio County, Texas (Specian & Ubelaker 1974a), and U. omatus from Arizona (Babero & Matthias 1967; Walker & Matthias 1973). Oligacanthorynchid acanthocephalan cystacanths were recovered from five male C. gularis septemvittatus (SVL = 107.8 + 7.7, range 99-115 mm) collected in June 1989 and May 1990. Cystacanths are apparently common helminths of whiptails as they have been reported previously from Cnemidop horns spp., including C dixoni from the same site included herein (McAllister et al. 1991a). In summary, four new host records are documented for parasites of C. gularis septemvittatus, and all of its parasites are shared with the unisexual species, C. dixoni A and C. tesselatus. This is not surprising given that C. gularis septemvittatus is considered to be one of the parental congeners of both species by Wright & Vitt (1993). Acknowledgments We wish to thank A. Real, general manager of the Mesquite Ranch, Presidio County, Texas, for lodging and field assistance for JEC. Lizard collections were made under the authority of Texas Parks and Wildlife Department Scientific Collecting Permit no. 61. Literature Cited Babero, B. B., & D. V. Matthis. 1967. Thubunaea cnemidophoms n. sp., and other helminths from lizards, Cnemidophoms tigris, in Nevada and Arizona. Trans. Amer. Micros. Soc., 86:173-177. Conant, R., & J. T. Collins. 1991. A field guide to reptiles and amphibians of eastern and central North America. Houghton- Mifflin Co., Boston, 3rd ed., 450 pp. Cordes, J. E., J. M. Walker, & R. M. Abuhteba. 1990. Cnemidophoms gularis septemvittatus (Teiidae) from Pecos County, Texas. Texas J. Sci., 42:209-210. Dixon, J. R. 1987. Amphibians and reptiles of Texas. Texas A&M University Press, College Station, Texas. 434 pp + xii. McAllister, C. T. 1990a. Helminth parasites of unisexual and bisexual whiptail lizards (Teiidae) in North America. 1. The Colorado 88 THE TEXAS JOURNAL OF SCIENCE- VOL. 47, NO. 2, 1995 checkered whiptail (Cnemidop horns tesselatus). J. Wildlife Dis., 26:139-142. _ . 1990b. Helminth parasites of unisexual and bisexual whiptail lizards (Teiidae) in North America. II. The New Mexico whiptail (Cnemidop horns neomexicanus) . J. Wildlife Dis., 26:403-406. _ . 1990c. Helminth parasites of unisexual and bisexual whiptail lizards (Teiidae) in North America. III. The Chihuahuan spotted whiptail (Cnemidophorus exsanguis). J. Wildlife Dis., 26:544-546. _ . 1990d. Helminth parasites of unisexual and bisexual whiptail lizards (Teiidae) in North America. IV. The Texas spotted whiptail (Cnemidophorus gularis). Texas J. Sci., 42:381-388. _ . 1992. Helminth parasites of unisexual and bisexual whiptail lizards (Teiidae) in North America. VIII. The Gila spotted whiptail (Cnemidophorus flagellicaudus) , Sonoran spotted whiptail (Cnemidophorus sonorae), and plateau striped whiptail (Cnemidophorus velox). Texas J. Sci., 44:233-239. _ , J. E. Cordes, & J. M. Walker. 1991a. Helminth parasites of unisexual and bisexual whiptail lizards (Teiidae) in North America. VI. The gray-checkered whiptail (Cnemidophorus dixoni). Texas J. Sci., 43:309-314. _ , S. E. Trauth, & D. B. Conn. 1991b. Helminth parasites of unisexual and bisexual whiptail lizards (Teiidae) in North America. VII. The six-lined racerunner, Cnemidophorus sexlineatus. Texas J. Sci., 43:391-397. _ , J. E. Cordes, D. B. Conn, J. Singleton, & J. M. Walker. 1991c. Helminth parasites of unisexual and bisexual whiptail lizards (Teiidae) in North America. V. Mesocestoides sp. tetrathyridia (Cestoidea: Cyclophyllidea) from four species of Cnemidophorus. J. Wildlife Dis., 27:494-497. Specian, R. D., & J. E. Ubelaker. 1974a. Parathelandros texanus n. sp. (Nematoda: Oxyuridae) from lizards in west Texas. Trans. Amer. Micros. Soc., 93:413-415. _ . 1974b. Two new species of Pharyngodon Diesing, 1861 (Nematoda: Oxyuridae) from lizards in west Texas. Proc. Helminthol. Soc. Wash., 41:46-51. Walker, K. A., & D. V. Matthias. 1973. Helminths of some northern Arizona lizards. Proc. Helminth. Soc. Washington, 40:168-169. Wright, J. W., & L. J. Vitt. 1993. Biology of whiptail lizards (genus Cnemidophorus). Oklahoma Mus. Nat. Hist., Norman, Oklahoma. 417 + xiv pp. TEXAS J. SCI. 47(2): 89-100 MAY, 1995 CRANIAL AND DENTAL VARIATION IN THE NINE-BANDED ARMADILLO, DASYPUS NOVEMCINCTUS, FROM TEXAS AND OKLAHOMA Frederick B. Stangl, Jr., Stacie L. Beauchamp and N. Gayle Konermann Department of Biology, Midwestern State University, Wichita Falls, Texas 76308 Abstract.— Variations of cranial and dental features were examined in the skulls of 60 specimens of Dasypus novemcinctus from north Texas and southern Oklahoma. Each specimen was assigned to one of three relative age groups (subadult, adult or old adult) based on the degree of fiision of ventral braincase elements. Measurements of cranial and mandibular characters were quantitatively analyzed. Evidence for sexual dimorphism appears lacking. The considerable size overlaps of morphometric measurements between relative age groups suggests a rapid and perhaps less-than-uniform rate of growth among individuals. Timing and sequence of tooth eruption, tooth loss, and dental anomalies are also described. The nine-banded armadillo {Dasypus novemcinctus) is a common and conspicuous mammal across much of its range. While much has been written about the general anatomy, physiology, and reproductive biology of this species (for summaries refer to Newman & Patterson 1910; Newman 1913; Talmage & Buchanan 1954; McBee & Baker 1982), few efforts have been directed towards the investigation of morphological variation within this species. No comprehensive assessment of geographic variation exists for this species, although the topic has received cursory treatment in studies of broader taxonomic scope (Wetzel & Mondolfi 1979; Wetzel 1985). This includes such aspects as cranial and dental variation, growth and development, and of tooth replacement. Post-partum growth and development have yet to be detailed, although details of embryonic and fetal stages of development have long been known (e.g. Newman 1913; Martin 1916). An understanding of these aspects of armadillo biology remains in need of clarification. Detailed studies of nongeographic variation do not exist, although regional comprehensive treatments of mammals often provide repre¬ sentative measurements such as those by Lowery (1974) for Louisiana, Bee et al. (1981) for Kansas, Schmidly (1983) for east Texas, Dalquest & Horner (1984) for north Texas and Sealander & Heidt (1990) for Arkansas. Perhaps a reason for this void in the literature on such a widespread, conspicuous, and regionally common mammal is due to its 90 THE TEXAS JOURNAL OF SCIENCE— VOL. 47, NO. 2, 1995 Figure 1 . Cranial and mandibular measurements of Dasypus novemcinctus utilized in this study and summarized in Table 1. Abbreviations and characters are as follows: sk, greatest length of skull; na, greatest width of nasals; ro, rostral width; zy, greatest zygomatic width; io, interorbital breadth; pm, length of premaxillary; mx, length of maxillary, pa, length of palatine; oc, occipital breadth; ut, length of maxillary (upper) tooth row; It, length of mandibular (lower) toothrow. poor representation in many systematic collections. This observation is supported by the surprisingly few specimens reported from regions such as Louisiana Lowery 1974) and east Texas (N=49; Schmidly 1983), where the armadillo is particularly abundant. This may be because most existing study specimens of D. novemcinctus are obtained by salvage of dead animals from along roadways and such collections are often considered an inconvenient or unpleasant task to perform. A series of skulls of D. novemcinctus from Texas and southern Oklahoma comprise the basis for this examination of variation by individual, sex, and age in this species. A method of determining relative age groups within the species is described, which may encourage investigations of geographic patterns of cranial variation in the nine-banded armadillo. Methods and Materials Eleven measurements were attempted from the skulls and mandibles of 60 specimens of Dasypus novemcinctus originating from Texas and STANGL, BEAUCHAMP & KONERMANN 91 PS Figure 2. Ventral view of braincase in a subadult Dasypus novemcinctus , as indicated by open sutures (arrows). Components are: ps, presphenoid; bs, basisphenoid; and bo, basioccipital. southern Oklahoma (Appendix 1). These included right alveolar toothrows of the mandible and maxillary, greatest skull length, greatest zygomatic breadth, least inter orbital breadth, greatest width of nasals, basal width of rostrum, greatest occipital breadth, and greatest lengths of the premaxillary, maxillary, and palatine bones from along the medial suture as viewed from the ventral perspective (Figure 1). Measurements were taken with a digital caliper, and recorded to the nearest 0.01 mm, although the fragmentary or damaged nature of some material prevented obtaining all measurements from many specimens. Dental observations on tooth eruption patterns and congenital anomalies were recorded for each specimen. Tooth losses considered as natural events were distinguished from accidental loss during the preparatory process by the partial or complete filling in of alveoli by spongy bone. When possible, specimens were classed by sex. Three age classes were assigned on the basis of ventral braincase suture closures between the basioccipital and basisphenoid, and between the basisphenoid and presphenoid (Figure 2). Subadults were defined by the persistence of both sutures. Young adults were characterized by fusion of the basioccipital with basisphenoid, and fully erupted permanent dentition. Specimens defined as old adults were those in which both braincase sutures were obliterated. Each month of the year was represented, except for the period of 92 THE TEXAS JOURNAL OF SCIENCE— VOL. 47, NO. 2, 1995 Table 1 . Variation by age class in cranial measurements (in mm) of Dasypus novemcinctus from the Rolling Plains of Texas and Oklahoma. Descriptive statistics are: sample size (N), mean, standard deviation (SD), range, and coefficient of variation (CV). Age Class (N) Mean-I-SD Range CV Duncan’s Maximum length of skull ** Subadult (19) 95.34 + 3.29 86.10 - 100.59 3.45 1 Adult (14) 101.01 + 1.80 97.07 - 104.17 1.79 Old adult (15) 102.37 + 2.27 99.09 - 107.26 2.21 Rostral width ** Subadult (21) 16.83 + 2.84 11.31 - 23.28 16.89 1 Adult (17) 18.63 2.12 14.92 - 23.03 11.39 Old adult (19) 19.12 + 1.49 16.94 - 21.88 7.80 Minimum interorbital breadth ** Subadult (23) 22.23 + 0.91 21.37 - 24.78 3.91 1 Adult (17) 24.34 0.95 22.98 - 26.10 3.89 Old adult (20) 24.56 + 0.84 22.64 - 25.91 3.43 Maximum zygomatic width ** Subadult (14) 38.02 + 1.79 35.18 - 40.41 4.70 1 Adult (12) 40.91 + 1.82 38.21 - 43.79 4.44 Old adult (12) 42.09 -h 1.86 38.83 - 45.10 4.42 Maximum width of nasals ** Subadult (16) 9.42 + 0.65 8.14 - 10.18 6.90 1 Adult (16) 10.15 + 0.48 9.31 - 11.12 4.69 Old adult (12) 10.29 0.54 9.32 - 11.04 5.27 Occipital breadth ** Subadult (19) 23.55 + 0.72 21.72 - 24.84 3.04 Adult (16) 23.92 -f 0.79 22.10 - 25.10 3.31 Old adult (21) 24.54 -h 0.92 22.76 - 26.60 3.75 1 Length of premaxillary Subadult (16) 8.56 + 0.97 7.21 - 10.57 11.36 Adult (16) 9.05 + 0.82 7.53 - 10.36 9.04 N.S. Old adult (13) 8.79 -h 0.50 7.92 - 9.92 5.71 Length of maxillary ** Subadult (23) 37.14 2.31 32.20 - 41.17 6.23 1 Adult (17) 39.52 1.47 36.17 - 41.63 3.73 Old adult (19) 39.45 + 1.70 39.45 - 43.17 4.31 Length of palatine ** Subadult (23) 16.98 -1- 1.32 14.61 - 19.17 7.79 1 Adult (17) 18.29 -f 2.03 11.95 - 20.70 11.10 1 Old Adult (20) 19.23 + 1.01 17.44 - 20.79 5.26 1 Length of right maxillary toothrow Subadult (22) 25.19 -h 1.39 20.80 - 27.43 5.53 Adult (17) 25.64 -h 1.29 22.74 - 28.16 5.03 N.S. Old adult (19) 25.85 + 1.69 21.57 - 28.28 6.54 Length of right mandibular toothrow ** Subadult (22) 25.54 + 1.56 22.34 - 28.89 6.12 1 Adult (17) 26.88 + 1.02 25.54 - 29.18 3.80 Old adult (20) 26.75 + 0.89 25.06 - 29.12 3.31 1. One-way analysis of variance (ANOVA): **=0.01 >P>0.001. 2. Duncan’s Multiple Range Test. Age classes grouped together are not significantly different (P<0.05). STANGL, BEAUCHAMP & KONERMANN 93 Figure 3. Lateral view of emerging permanent dentition in a subadult Dasypus novemcinctus (MWSU 17046). Stippling indicates position of remnant deciduous precursors. June-September, during which time a single animal was recorded from 5 July. All examined specimens are deposited in the Midwestern State University (MWSU) Collection of Recent Mammals. Statistical analyses were performed with the Number Cruncher Statistical Systems (NCSS; Hintze 1990). Two-way ANOVAs (by sex and age) were conducted to assess the interactive effects between sex and age for each measurement, and found to be negligible. Sexual dimorphism was then evaluated with one-way ANOVAs by age class for each character. Only subadults displayed any variation by sex, and these differences (palatine length, P =0.0435; interorbital breadth, P=0.0491) were marginally significant. Therefore, individuals of both sexes were pooled with unsexed individuals for one-way ANOVAs to evaluate age variation for each character. Significant subsets of age groupings for each set of measurements were determined by Duncan’s multiple range tests. Results and Discussion Relative age determinations A fairly accurate means of assigning individuals from natural populations to relative age categories is a necessary prerequisite to any meaningful study of morphological variation in Dasypus novemcinctus, for which no hard data exists in the literature on postpartum growth and development of the species. Skeletal and dental observations lend support to the validity of our relative age groupings based on crania, although considerable overlap exists among each age group for all measurements (Table 1). Four of the 23 subadult individuals (as defined by cranial fusions) still retained their deciduous dentition, 15 were marked by newly erupting permanent dentition (e.g. Figure 3), and the remaining seven had new to lightly 94 THE TEXAS JOURNAL OF SCIENCE-VOL. 47, NO. 2, 1995 worn permanent dentition. Each of the seven skeletons from animals of this age category were characterized by persistent epiphyseal cartilage. No skeletons were available from animals designated as young adults, although each skull from this age category possessed fully emerged and unworn or little worn permanent dentition. Dentition of old adults was typically worn, and tooth loss was prevalent. Of the six skeletons of older animals available for inspection, all epiphyseal sutures were fused beyond recognition. Additionally, while we do not have reproductive data for all specimens, each animal for which we have data inferring reproductive maturity (males with enlarged testes and lactating or pregnant females) had been classed as old adults. If Martin (1916) is correct in her assessment that animals achieve near-adult size within four to six months, then most specimens designated as subadults are less than one year old. The chronological distribution of subadults in our study was continuous October- April, with peaks in April (N=7) and November (N=4), doubtless reflecting the extended six-month breeding season and year-around occurrence of juvenile specimens (< 3 kg) as reported by McDonough (1990). The young adult stage probably represents individuals in their second year, and the range of ages for animals classed as old adults may span many years, for Crandall (1974) provides the longevity for a zoo animal at more than six years, and McDonough (1994) refers to a captive specimen exceeding 22 years old. Morphometric variation. —S\gn\f\c2C[ii age variation exists for all characters but premaxillary length. However, extensive overlaps in size occur for each of the 11 measurements. The arbitrary nature of age category assignments doubtless contributes to this observation, and individual rates of age-diagnostic cranial fusions cannot be discounted. Subadults were significantly smaller than older age categories for eight of the characters examined, although only two characters distinguished young adults from older individuals. Morphometric similarity of the two older relative ages suggests that the rapid growth rates postulated by Martin (1916) decelerate during the second year, followed by an imperceptible rate of growth for the duration of the life of an individual. Dentition—Mdirim' s (1916) examination of a large series of embryos and newborn specimens provided a detailed assessment of the early dental development and typical dental formula of D. novemcinctus . STANGL, BEAUCHAMP & KONERMANN 95 Based on her account, young animals are born with eight functional upper teeth which originate on the maxillary. Only the first seven are apparently replaced by permanent teeth. The last eight of 13-14 tooth germs in the lower jaw ultimately give rise to milk teeth, although only the first seven of these are replaced by permanent teeth. This study substantiates Martin’s (1916) description of the typical functional armadillo dentition as comprised of eight teeth in each of the upper and lower quadrants, and that the last of each tooth row lacks a deciduous precursor. She asserted that premolars, canines, and incisors were probably represented among the deciduous and replacement teeth. This cannot apply to upper teeth, which all originate in the maxillary, for when present in therian mammals, incisors occur on the premaxillary and the canine at the premaxillary-maxillary junction. While we agree that the poster iormost teeth of both upper and lower dental batteries are molars, precise homologies of the other teeth cannot be made with any certainty. As a matter of expedience in numbering the teeth, we refer to deciduous teeth and their replacements as premolars, and to the last single tooth in each tooth row as a molar. Deviation from the normal 8/8 dental formula in D. novemcinctus attributed to tooth loss appears to be a function of age, and presumably is not particularly detrimental for a species that feeds mostly on soft-bodies invertebrates. Loss of the simple, peg-like teeth of armadillos during the cleaning of cranial materials is common, and leaves a well-defined alveolus, and unless only recently lost, the alveoli of teeth lost during the life of an animal are initially filled in with a scaffolding similar in appearance to spongy, or perhaps woven, bone. Ten of 26 subadults were missing teeth, as well as seven of 16 young adults, and 14 of 20 old adults. The upper molars were most commonly missing, and contributed to most instances of tooth loss among the subadults. One cannot discount the possibility that these small molars sometimes never erupt, for there was seldom any trace of spongy or cancellous bone which fills empty the empty alveolus of teeth lost during life. Among specimens of the two younger age classes, only one individual (MWSU 19514) was missing a permanent upper premolar. Among old adults, eight specimens were missing one or more permanent premolars, and two particularly aged individuals (MWSU 1729 and 19215) were lacking a total of 6 and 12 premolars from their dentition. Tooth replacement of premolars occurs in subadults nearly the size of adults. However, specimens varying in size and exhibiting comparable stages of tooth replacement were taken in every month but March- 96 THE TEXAS JOURNAL OF SCIENCE— VOL. 47, NO. 2, 1995 September, suggesting either a prolonged replacement process or considerable variation in timing of the acquisition of adult dentition. The upper molars are typically erupted prior to full emergence of their lower counterparts. However, eruption sequences of upper and lower premolars appear to be almost simultaneous events, initiated by the appearance of the fourth, fifth, and sixth premolars (Figure 3). Martin (1916) first noted that the permanent teeth often wear through the deciduous teeth, leaving persistent remnants along the anterior and posterior margins of the alveoli. We found such traces to be common in armadillos of all ages, and even the largest and possibly oldest of examined specimens (MWSU 19542; as determined by greatest skull length and extent of tooth wear and loss) still retained traces of nine deciduous teeth along the alveolar margins. Dental variation attributed to developmental or genetic factors was noted in four specimens. A very small supernumerary tooth, little more than a slender spicule and likely derived from one of the usually atro¬ phied anterior tooth buds noted by Martin (1916), marked the lower left jaw of one animal (MWSU 19515), just anterior to the normal dentition. The maxillary tooth rows of another animal (MWSU 1681), although comparable in length and normal in tooth composition, were staggered, with the left row situated anteriorly by about 10 mm. A third specimen (MWSU 18214) possessed five sets of paired maxillary teeth and one set of paired mandibular teeth (Figure 4a), apparently the result of a splitting of the tooth buds. The fourth aberration (MWSU 19215) involved the staggered clustering of three premolars on the left maxillary toothrow and of two teeth on the right (Figure 4b). Additional Investigations Despite the fact that Dasypus novemcinctus is both common and conspicuous throughout much of its range, detailed studies of geographic variation in this species are lacking. Of particular interest would be comparisons between populations originating from animals introduced into Florida, which have since spread and made contact with those of the original range. This study suggests that both sexed and unsexed specimens can be pooled for such analyses, although efforts to collect or salvage additional specimens will probably be necessary to bolster sample sizes in many areas. Published information on early life history of the nine-banded armadillo from neonate through juvenile stages is both sketchy and STANGL, BEAUCHAMP & KONERMANN 97 Figure 4. Aberrant alveolar placement in upper dentition of two old adult Dasypus novemcinctus: a) "twinning" of teeth (MWSU 18214); and b) alveolar gathering (MWSU 18215). Stippling indicates alveolus filled with spongy bone, following tooth loss earlier in life. anecdotal. Postpartum development and growth rates with known- aged individuals have yet to be detailed in the literature, and such data may be useful in assigning absolute ages to museum specimens and to test the accuracy of our relative age groupings. Acknowledgments We wish to thank the two anonymous reviewers whose thoughtful comments and suggestions on an earlier version of this manuscript greatly improved the final product. Literature Cited Bee, J. W., G. E. Glass, R. S. Hoffmann & R. R. Patterson. 1981. Mammals in Kansas. Univ. Kansas Mus. Nat. Hist., Public Ed. Ser., 7:ix + 300 pp. Crandall, L. S. 1964. Management of wild mammals in captivity. Univ. Chicago Press, xv + 769 pp. Dalquest, W. W. & N. V. Horner. 1984. Mammals of north-central Texas. Midwestern State Univ. Press, Wichita Falls, Texas, 254 pp. 98 THE TEXAS JOURNAL OF SCIENCE— VOL. 47, NO. 2, 1995 Hall, E. R. 1981. The mammals of North America. John Wiley & Sons, New York, 2nd ed., l:xv + 1-600 + 90. Hintze, J. L. 1990. Number Cruncher Statistical Systems, Version 5.03. Pacific Ease Co., Santa Monica, California, 422 pp. Lowery, G. H., Jr. 1974. The mammals of Louisiana and its adjacent waters. Louisiana State Univ. Press, Baton Rouge, xxiii + 565 pp. Martin, B. E. 1916. Tooth development mDasypus novemcinctus . J. Morphology, 27:647-691. McBee, K. & R. J. Baker. 1982. Dasypus novemcinctus. Mammalian Species, 162:1-9. McDonough, C. M. 1994. Determinants of aggression in nine- banded armadillos. J. Mamm., 75:189-198. Newman, H. H. 1913. The natural history of the nine-banded armadillo in Texas. American Nat., 47:513-539. _ & J. T. Patterson. 1910. The development of the nine-banded armadillo from the primitive streak stage to birth: with especial reference to the question of specific polyembryony. J. Morphol., 21:359-423. Sealander, J. A. & G. A. Heidt. 1990. Arkansas mammals; their natural history, classification, and distribution. Univ. Arkansas Press, Fayetteville, xiv + 308 pp. Schmidly, D. J. 1983. Texas mammals east of the Balcones Escarp¬ ment. Texas A&M Univ. Press, College Station, xvii + 400 pp. Talmadge, R. V. & G. D. Buchanan. 1954. The armadillo {Dasypus novemcinctus). A review of its natural history, ecology, anatomy and reproductive physiology. Rice Inst. Pamphlet, Monogr. Biol., 41:1-135. Wetzel, R. M. 1985. Taxonomy and distribution of armadillos, Dasypodidae. Pp. 23-46, in The evolution and ecology of armadillos, sloths, and vermalinguas (G. G. Montgomery, ed.). Smithsonian Institution Press, Washington, D. C., 451 pp. _ & E. Mondolfi. 1979. The subgenera and species of long-nosed armadillos, genus Dasypus. Pp. 43-63, in Vertebrate ecology in the northern neotropics (J. F. Eisenberg, ed.) Smithsonian Institution Press, Washington, D. C., 271 pp. STANGL, BEAUCHAMP & KONERMANN 99 Appendix Following is a listing of localities and MWSU numbers for 60 specimens of Dasypus novemcinctus from the Midwestern State University Collection of Recent Mammals examined during this study. Counties are listed in alphabetic order by state. Oklahoma.— GvcQt County: 7.5 mi W of Quartz Mountains State Park, one specimen (16267). Jackson County: 8 mi E of Altus, one specimen (13342). Roger Mills County: 1 mi S of Cheyenne, one specimen (16630). Stephens County: 8 mi N, 5 mi W of Waurika, one specimen (16360). Texas. — Archer County: Lake Kickapoo, one specimen (9961); 2 mi N of Scotland, one specimen (18453); Archer City, two specimens (1683 & 11173). Baylor County: 8 mi NE of Seymour, one specimen (6422); 7.3 mi E of Seymour, one specimen (19516). Clay County: no specific locality, one specimen (1198); 3 mi N of Shannon, one specimen (17105); 15 mi N of Henrietta, one specimen (15442); 4.2 mi N of Henrietta, one specimen (19515); 7 mi W of Henrietta, one specimen (13378); 4 mi W of Henrietta, one specimen (17103). Cooke County: 5 mi E of Saint Jo, one specimen (7512). Grayson Co.: 5 mi N of Gordonville, one specimen (19213). Hopkins County: 10 mi N of Sulphur Springs, one specimen (16706). Jack County: 5 mi W of Crafton, four specimens (1759-1762); 36 mi S of Wichita Falls, one specimen (1682); 7.4 mi N of Graford, one specimen (18722); 15 mi N, 2 mi W of Jacksboro, one specimen (11525); 14 mi N, 1 mi W of Jacksboro, one specimen (13377). Montague County: 6 mi NNE of Nocona, one specimen (16705); 4.7 mi NW of Nocona, one specimen (18491); 5.6 mi W of Nocona, one specimen (19542); 3 mi E of Nocona, one specimen (19970); 3 mi E of Forestburg, one specimen (10983); 3 mi S of Vashti, one specimen (11204). Stephens County: 18 mi N of Breckenridge, one specimen (19513); 15 mi NE of Brecken- ridge, one specimen (3130). Wichita County: no specific locality, two specimens (1199 & 11175); 2.7 mi ENE of Kamay, one specimen (18137); 3 mi N of Iowa Park, one specimen (11174); 1 mi S of Iowa Park, one specimen (15441); near Iowa Park, one specimen (19517); 6 mi NE of Punkin Center, one specimen (19215); 10 mi S of Electra, one specimen (19512); 10 mi S of Burkburnett, one specimen (13379); 10 mi W of Wichita Falls, one specimen (18214); 2 mi W of Wichita Falls, one specimen (15440); Wichita Falls, 10 specimens (1201, 1274, 1465, 100 THE TEXAS JOURNAL OF SCIENCE-VOL. 47, NO. 2, 1995 1729, 7208, 10984, 17046, 19214, 19514 & 19518). Wise County: no specific locality, one specimen (1681). Young County: 1.3 mi S of Markley, one specimen (18721); 3 mi W, 3.8 mi S of Graham, one specimen (17104); 14 mi S of Windthorst, one specimen (16704). TEXAS J. SCI. 47(2): 101-105 MAY, 1995 DISTRIBUTIONAL RECORDS OF SMALL MAMMALS FROM THE SOUTHWESTERN ROLLING PLAINS OF TEXAS Franklin D. Yancey, 11, Jim R. Goetze, Burhan M. Gharaibeh and Clyde Jones The Museum and Department of Biological Sciences, Texas Tech University, Lubbock, Texas 79409-3191 Abstract. — Distributional notes based upon recent field collections are reported for nine species of small mammals from the southwestern Rolling Plains of Texas. These include one mole (Scalopus), one bat (Tadarida), one armadillo {Dasypus), one kangaroo rat (Dipodomys) , two woodrats (Neotoma) and three mice {Reithrodontomys and Baiomys). Recent field studies conducted on the southwestern Rolling Plains of Texas have resulted in the collection of specimens of nine mammalian species, which represent county records or additional distributional data for these species from this area of the state. The following species accounts are the result of this research. Voucher specimens are deposited with the holdings of The Museum at Texas Tech University (TTU). Scalopus aquaticus aereus (Bangs) (Eastern Mole) Distributional notes.— k single specimen of the eastern mole was collected south of White River Lake in Garza County. The nearest reported occurrence of S. aquaticus aereus is from Dickens County, approximately 40 km to the northeast (Choate 1990). This record of the eastern mole represents the first from the escarpment breaks of the Rolling Plains in Garza County. Material examined. — 12 mi S of White River Lake, Garza County, Texas, 6 December 1982, one specimen (TTU 40657). Habitat .—So\\s> at the collection locality are loamy, fine sands, which are suitable for habitation by moles. Tadarida brasiliensis mexicana (Saussure) (Brazilian Free-tailed Bat) Distributional Although this species is distributed statewide in Texas (Hall 1981), records are most common from the central and Trans-Pecos regions of Texas (Schmidly 1977; Manning et al. 1987). 102 THE TEXAS JOURNAL OF SCIENCE— VOL. 47, NO. 2, 1995 Schmidly (1991) lists this species from adjacent Garza and Stonewall counties. This is the first report of specimens from Kent County. Material examined.— Girard, Kent County, Texas, 9 October 1984, 24 specimens (TTU 41952-41965, 41967-41974, 49285-49286); 17 October 1985, 62 specimens (TTU 43205-43206, 45461-45520); 23 September 1986, six specimens (TTU 49279-49284). Specimens include fluid-preserved, skins and skulls, and skeletons. Dasypus novemcinctus mexicanus Peters (Nine-banded Armadillo) Distributional notes.— Pddiough. the armadillo ranges throughout eastern and southern Texas (Schmidly 1983) and is currently expanding its range onto the Llano Estacado (Jones et al. 1993), there are few records of occurrence for the southwestern Rolling Plains. A single specimen each for both Scurry and Garza counties represent new county records for this species. Material examined. — \2 mi N, 9 mi W of Snyder, Scurry County, Texas, 6 June 1993, one specimen (TTU 63409). 8 mi S, 9 mi E. of Post, Garza County, Texas, 22 July 1993, one specimen (TTU 63408). Dipodornys ordii medius Setzer (Ord’s Kangaroo Rat) Distributional notes.— species has been reported to occur throughout west-central Texas (Hollander et al. 1987) and has been taken in several counties of the Rolling Plains region (Jones et al. 1991). This report represents the first specimen record for Kent County. Material examined.— 1 mi S, 17 mi W of Clairemont, Kent County, Texas, 20 November 1993, three specimens (TTU 63393-63395). One female specimen (TTU 63394) was gravid with three fetuses measuring 19mm in crown- rump length. Habitat. —Sptcimtn^ were collected along the sandy terraces of the Double Mountain Fork of the Brazos River southwest of Clairemont. Reithrodontomys megalotis megalotis (Baird) (Western Harvest Mouse) Distributional notes. — The western harvest mouse exhibits a wide¬ spread distribution on the Llano Estacado and ranges southward into the YANCEY ET AL. 103 Trans-Pecos region of Texas (Davis & Schmidly 1994). This species was previously believed to be restricted to the Llano Estacado at the extreme southeastern part of its range (Choate et al. 1992). This report extends the range of this species onto the western Rolling Plains. Material examined.— 1 mi S, 17 mi W of Clairemont, Kent County, Texas, 3 April 1993, one specimen (TTU 63054). 3 mi N, 9 mi E of Justiceburg, Garza County, Texas, 1 May 1993, one specimen (TTU 63315). Habitat.— Tht collection locality southwest of Clairemont is adjacent to the Double Mountain Fork of the Brazos River in an area of sandy, alluvial soil. Vegetation was dense and dominated by tall grasses. The collection locality northeast of Justiceburg is in an area of upland mesquite rangeland. Reithrodontomys montanus griseus Bailey (Plains Harvest Mouse) Distributional notes .—Tht plains harvest mouse is found in western and central Texas (Davis & Schmidly 1994). This report represents the first record of this species from Kent County. This species appears to prefer more xeric, upland habitats than does R. megalotis (cf. Jones et al. 1983). Although these two species of harvest mice are sympatric in parts of their ranges in Texas, they are usually separated in their specific habitats and are rarely taken together at the same locality. Material examined.— 1 mi S, 17 mi W of Clairmont, Kent County, Texas, 3 April 1993, two specimens (TTU 63057-63058); 20 November 1993, one specimen (TTU 63396). All three specimens are males with testicular measurements of 5 by 3 mm for the April specimens and 2 by 1 mm for the November specimen. Baiomys taylori taylori (Thomas) (Pygmy Mouse) Distributional notes.— Tht expansion of the geographic range of the pygmy mouse to include the Llano Estacado and adjacent areas was summarized by Choate et al. (1990). This species is distributed over the eastern two- thirds of the state wherever suitable habitat may be found, but records from the southwestern Rolling Plains previously were lacking. Material examined.— 1 mi S, 17 mi W of Clairemont, Kent County, 104 THE TEXAS JOURNAL OF SCIENCE- VOL. 47, NO. 2, 1995 Texas, 3 April 1993, three specimens (TTU 63039-63041); 24 June 1993, one specimen (TTU 63331). A single specimen (TTU 63331) was gravid with three fetuses measuring 5 mm in crown-rump length. Habitat.— Tht habit at this locality consists of sandy alluvial terraces dominated by sand dropseed {Sporobolus cryptandrus) and other tallgrass species. Neotorna albigula albigula Hartley (White- throated Woodrat) Distributional notes.— Tht white-throated woodrat is distributed throughout much of northwestern, central, and Trans-Pecos Texas (Jones & Jones 1992), but records from the southwestern Rolling Plains are rare. This report represents the first record of this species from Kent County. Material examined.— 1 mi S, 17 mi W of Clairemont, Kent County, Texas, 12 May 1993 and 20 November 1993, four specimens (TTU 63340-63342, 63404). Habitat species prefers saxicolous habitats. The collection locality in Kent County is in the rocky, juniper woodlands of the escarpment breaks. Neotorna micropus canescens J. A. Allen (Southern Plains Woodrat) Distributional notes .—Tht southern plains woodrat is found in the western two-thirds of Texas (Jones & Jones 1992), though no previous records from Kent County exist. Material examined.— 1 mi S, 17 mi W of Clairemont, Kent County, Texas, 3 April 1993, one specimen (TTU 63059); 12 May 1993, one specimen (TTU 63345). One specimen (TTU 63059) was gravid with two fetuses measuring 10 mm in crown- rump length. Specimens were collected in an area of sandy, alluvial soils dominated by wild plum (Prunus sp.). Acknowledgments We wish to thank R. R. Hollander and an anonymous reviewer for their comments on this manuscript. YANCEY ET AL. 105 Literature Cited Choate, L, L. 1990. First record of the mole, Scalopus aquaticus, on the Llano Estacado of western Texas. Texas J. Sci., 42(2): 207. ___, J. K. Jones, Jr., R. W. Manning & C. Jones. 1990. Westward ho: Continued dispersal of the pygmy mouse, Baiomys taylori, on the Llano Estacado and in adjacent areas of Texas. Occas. Papers Mus., Texas Tech Univ., 134:1-8. _ , R. W. Manning, J. K. Jones, Jr., C. Jones & S. E. Henke. 1992. Mammals from the southern border of the Kansan Biotic Province. Occas. Papers Mus., Texas Tech Univ., 152:1-34. Davis, W. B. & D. J. Schmidly. 1994. The mammals of Texas. Texas Parks and Wildlife Dept., Nongame and Urban Program, x -f 338 pp. Hall, E. R. 1981. The mammals of North America. John Wiley & Sons, New York, 2nd ed., l:xv + 1-600 +90. Hollander, R. R., C. Jones, R. W. Manning & J. K. Jones, Jr. 1987. Distributional notes on some mammals from the Edwards Plateau and adjacent areas of south-central Texas. Occas. Papers Mus., Texas Tech Univ., 110:1-10. Jones, J. K., Jr., D. M. Armstrong, R. S. Hoffmann & C. Jones. 1983. Mammals of the northern Great Plains. Univ. of Nebraska Press, Lincoln, xii + 379 pp. _ & C. Jones. 1992. Revised checklist of Recent land mammals of Texas, with annotations. Texas J. Sci., 44(l):53-74. _ , R. W. Manning & J. R. Goetze. 1991. Noteworthy records of seven species of small mammals from west-central Texas. Occas. Papers Mus., Texas Tech Univ., 143:1-4. _ , _ , F. D. Yancey, II & C. Jones. 1993. Records of five species of small mammals from western Texas. Texas J. Sci., 45(1): 104-105. Manning, R. W., J. K. Jones, Jr., R. R. Hollander & C. Jones. 1987. Notes on distribution and natural history of some bats on the Edwards Plateau and in adjacent areas of Texas. Texas J. Sci. , 39(3): 279-285. Schmidly, D. J. 1977. The mammals of Trans-Pecos Texas. Texas A&M Univ. Press, College Station, xiii + 225 pp. _ _ . 1983. Texas mammals east of the Balcones Fault zone. Texas A&M Univ. Press, College Station, xviii + 400 pp. _ . 1991. The bats of Texas. Texas A&M Univ. Press, College Station, xv + 188 pp. TEXAS J. SCI. 47(2): 106-1 16 MAY, 1995 HYDROCARBON DEGRADING BACTERIA AT OIL SPRINGS, TEXAS Thomas G. Benoit and Robert J. Wiggers Department of Biology, Stephen F. Austin State University, Nacogdoches , Texas 75962 Abstract. — Hydrocarbon degrading bacteria capable of heterotrophic growth in air were isolated from the oily mousse (biodegraded oil) from two separate sites at Oil Springs in east Texas. This is an area of long-standing crude oil seepage into a freshwater system. Of the four species of bacteria isolated from one site, and the five from the other, only two species were common to the oil at both sites. The isolates of these common species, however, were biotypically different from each other. Samples from both sites contained strains capable of growth on crude oil as a sole source of carbon (unconditional strains) along with strains that could degrade oil if proteose peptone was available in the growth medium (conditional strains). Additionally, the groups of isolates as a whole from each sample displayed remarkable similarities in the sizes of extrachromosomal elements, adhesion to crude oil, and hydrocarbon substrate utilization. The results indicate that common selection pressures may have produced groups of bacteria of different genera with similar physiologies at both sites. Hydrocarbon degrading bacteria are widely distributed in nature (ZoBell 1946; Atlas 1981; Leahy & Colwell 1990). They may constitute up to 100% of the viable bacterial community in areas exposed to hydrocarbons (Atlas 1981). These communities develop at least partly from the autochthonous bacteria when the area is first exposed to hydrocarbons. The numbers and proportions of the hydrocarbonoclastic bacteria in the community will increase after exposure (Colwell & Walker 1977; Atlas 1981; Floodgate 1984; Cooney 1984). Eventually, only a few genera may dominate in the community (Llanos & Kjoller 1976); or, community diversity may remain unchanged (Pinholt et al. 1979; Olsen & Sizemore 1981) or even increase (Hood et al. 1975). These communities characteristically degrade hydrocarbons at higher rates than similar bacterial communities in unexposed areas (Leahy & Colwell 1990). In general they develop as the result of the availability of hydrocarbons for bacterial growth, apparently favoring strains that can grow on hydrocarbons over those that cannot. The predominance of hydrocarbon-utilizing bacteria may become permanent in areas subject to chronic exposure. Petroleum-polluted environments also contain a much higher number of plasmid-bearing bacterial species than similar unpolluted areas (Hada & Sizemore 1981; Leahy & Colwell 1990; Ogunseitan et al. 1987). In Pseudomonas spp., the genes responsible for hydrocarbon degradation BENOIT & WIGGINS 107 often reside on plasmids. In these species the genes responsible for degradation of toluates, camphor, salycilate, alkanes, and naphthalene are found on, respectively, the TOL, CAM, SAL, OCT, and NAH plasmids (Chakrabarty 1976). In other known hydrocarbon-degrading bacteria such as Acinetobacter sp. HOl-N and A. calcoaceticus the genes are located chromosomally (Singer & Finnerty 1984). Interest in the degradative activities of hydrocarbonoclastic bacteria in marine and terrestrial environments has produced many studies to date. Not as much information has been obtained, however, on similar bacteria found in freshwater systems. Furthermore, most studies have been carried out in areas following accidental spills where selection pressures exerted by the oil may be of relatively short duration. Knowledge of the bacteria present in areas subjected to long-term, chronic exposure is lacking and would be of scientific interest since bacteria in these areas likely have formed stable associations. Oil Springs, Texas is such an area of long-standing hydrocarbon exposure (Pate 1987). This report characterizes hydrocarbon degrading bacteria capable of heterotrophic growth in air from the oily mousse at two sites there. Materials and Methods Site description and sampling.— Iht oil seepage was located in a forested area about 25 miles southeast of the Stephen F. Austin State University campus. Historical records indicate that as early as 1790 pioneers used the oil on the surface in this area for axle grease while traveling westward (Pate 1987). The first producing well in Texas also was drilled in this area in 1866. Samples of oily mousse (biodegraded oil/water mixture) were collected from two locations: 1) a site where crude oil seeps into a small basin of freshwater and 2) the bank of a stream receiving crude oil seepage. This stream also receives the discharge from the basin and was sampled approximately 100 m upstream of that point. The two sites are separated by a distance of approximately 200 m. Isolation and identification of bacterial Samples were returned at ambient temperature (approximately 25 °C) to the lab. Approximately one gram was used to inoculate nutrient broth containing 10 gm/1 crude oil as an enrichment for hydrocarbon-utilizing strains. These 24 h cultures then were streaked onto nutrient agar (Difco) to produce bacterial colonies. Preliminary sorting of isolates was carried 108 THE TEXAS JOURNAL OF SCIENCE-VOL. 47, NO. 2, 1995 Table 1 . Composition of growth media. Medium Composition (gm/liter) 1. SS^ (NH4)2S04,(2); K2HPO4, (14); KH2PO4, (6); MgS04, (0.2) 2. OSS Crude Oil, (2); in SS 3. PPSS Proteose peptone (Difco), (10); in SS 4. OPPSS Crude Oil, (2); proteose peptone, (10); in SS 5. DSS Diesel fuel, (2); in SS 6. MOSS Mineral oil, (2); in SS 7. NSS Naphthalene, (2); in SS 8. TSS Toluene, (2); in SS 9. Luria agar Tryptone, (10); yeast extract, (5); NaCl, (5); agar, (15) ^ Spizizen, J. 1958. Transformation of biochemically deficient strains of Bacillus subtilis by deoxyribonucleate. Proc. Nat. Acad. Sci. USA. 78:2893-2897. out on the basis of colony and cellular morphologies, cellular arrangements, and Gram stain reaction. Each unique isolate then was identified by using a combination of the following tools: Bergey's Manual of Systematic Bacteriology, John G. Holt (ed.), Williams and Wilkins, Baltimore / London; 20E and Rapid NET Identification Systems (BioMerieaux- Vitek), and Microscan Rapid Identification System (Baxter). Media and growth conditions .—Tht components of each growth medium are listed in Table I. All incubations were carried out at 25 °C with vigorous shaking. Extrachromosomal DNA isolation and agarose electrophoresis . — An isolation procedure designed to recover both large and small plasmids was used. Bacteria were grown overnight on Luria agar at 25 °C. Approximately one square centimeter of cells was scraped from the plate and resuspended in 0.3 ml of lysis buffer containing 50 mM Tris-HCl and 3% (w:v) sodium dodecyl sulfate adjusted to pH 12.6 with 10 M NaOH. Cells were incubated at 60 °C for 30 minutes. The resulting lysate was extracted with an equal volume of Tris-buffered phenol: chloroform (1:1, pH 7.5). The aqueous phase was recovered and electrophoresed through 1 % (w:v) agarose gels in TBE buffer (89 BENOIT & WIGGINS 109 Table 2. Degradation of crude oil and microbial adhesion to hydrocarbon (MATH) assays of bacterial isolates. Bacterial isolate Amount of crude oil degraded^ (ppm) in each medium OSS OPPSS MATH assay'’ percentage decrease in optical density Basin site 1. P. aeruginosa^ OSBl 1500-2000 1-500 0 2. E. cloacae^ OS2A 0 1500-2000 0 3. A. faecalis OS2B 1500-2000 1500-2000 75 4. B. cereus OS3 0 0 0 Stream site 1. S. marcescens CR2B 0 1000-1500 52 2. E. americana GREW 1-500 1500-2000 0 3. P. aeruginosa^ CRFYl 1000-1500 1-500 0 4. E. cloacae"^ CRFY2 0 1500-2000 0 5. F. odoratum CRLY 1000-1500 1000-1500 35 ^ Starting concentration of crude oil was 2000 ppm. Values reported are in ranges consistently observed in several trials. ^ Measurements as described in Materials and Methods. Averages of three trials. ^ The two P. aeruginosa and the two E. cloacae isolates are different biotypes. mM Tris base, 89 mM boric acid, 2 mM EDTA). DNA was stained with ethidium bromide and visualized with 254 nm wavelength UV light. Hydrocarbon Individual isolates were tested for the ability to utilize crude oil as a sole source of carbon and energy for growth by inoculating 100 ml of OSS broth (Table 1) with 0.1 ml of an 18-24 hour-old culture growing in PPSS broth (Table 1). OSS broth cultures were incubated seven days after which the remaining hydrocarbon was measured by infrared spectrophotometry according to Standard Methods for the Examination of Water and Wastewater (American Public Health Association 1990). In some experiments identical tests were conducted in PPSS broth. The ability to utilize diesel fuel, mineral oil, naphthalene, and toluene as sole sources of carbon and energy for growth was determined after incubation for seven days in DSS, MOSS, NSS and TSS broths (Table 1), respectively. Growth was considered positive if the cultures became turbid and dense populations of cells were detected microscopically. Microbial Adhesion to Hydrocarbon (MATH) /455'<3y5'.— MATH assays were carried out by a modification of the method of Rosenberg et al. (1980). Cultures that were 18-24 hours old were washed once and 110 THE TEXAS JOURNAL OF SCIENCE— VOL. 47, NO. 2, 1995 resuspended in SS ( Table 1). The density of the washed cultures was adjusted with SS to approximately 0.1 absorbance units at 600 nm wavelength. Two- tenths of a ml of crude oil was added to a 1.0 ml portion of the resulting suspensions which were then vigorously mixed for 120 seconds after which the hydrocarbon phase was allowed to separate from the aqueous phase for 15 minutes. The absorbance of the aqueous phase was measured. The percentage decrease in absorbance is a direct measurement of the percentage of cells removed from the suspension by adhesion to the crude oil. Results The four species of bacteria isolated from the basin and the five species from the stream site samples are listed in Table 2. All of these, with the exception of Erwinia, are common in freshwater habitats. It is probable that other species were present in the samples and were not culturable by the enrichment and isolation procedure. The organisms that were isolated, therefore, probably do not constitute the entire bacterial community present in the oil. A typical electrophoretic gel is shown in Figure 1 . All of the strains tested, except Pseudomonas aeruginosa CRFYl, contained the smaller of the two size classes of extrachromosomal elements present. Additionally, Enterobacter cloacae OS2A and CRFY2, Serratia marcescens CR2B, and Erwinia americana CRFW all possessed the larger size class of element. No other size classes of elements were detected. Each isolate was tested for its ability to utilize crude oil as a sole source of carbon and energy for growth in OSS broth (Table 2). Only P. aeruginosa OSB 1 and Alcaligenes faecalis OS2B from the basin site and E. americana CRFW, P, aeruginosa CRFYl, and Flavobacterium odoratum CRLY from the stream site utilized the oil. The other species were unable to grow or degrade the oil in this medium. The inability of the species to grow in OSS apparently was due to inability to utilize the oil as a carbon source and not to a missing growth factor since all could grow in SS supplemented with 1% (w:v) glucose. When similar experiments were conducted in OPPSS broth, which contains proteose peptone as an additional carbon source, degradation patterns changed for most of the isolates (Table 2). Degradation of the oil by the two P. aeruginosa isolates was reduced in this broth. This BENOIT & WIGGINS 111 X CRLY CRFWOS3 OS2B OS2A OSBl CR2B CRFYl CRFY2 X 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 23 kb 9.4 kb 6.5 kb Figure 1 . Agarose gel indicating the extra-chromosomal DNA found in the bacterial isolates. Lanes 1 and 20 contain Hind III digested lambda phage DNA as a size standard. Sizes of the visible standard bands are indicated on the right of the figure. Lysate from each bacterial isolate was run in side by side lanes; even lanes contained 20 /4I of lysate and odd lanes 40 /il. The species name for each isolate is given in Table 2. Both extrachromosomal DNA elements trail the 23 kb band of the lambda standard and are indicated by arrows. Several lanes also exhibit undegraded chromosomal DNA which co¬ migrates with the 23 kb standard band. Although not clearly visible, the presence of the larger element in CR2B was confirmed on other gels. likely was due to repression of enzyme activity by constituents of the proteose peptone, a phenomenon which has been observed in this species during paraffin oxidation (van Eyk & Bartles 1968). Degradation by all of the other isolates increased markedly, except for F. odoratum CRLY and Bacillus cereus OS3 which remained unchanged. It is unknown whether the oil was utilized for growth or was degraded by cooxidation. All of the isolates grew vigorously in OPPSS and so failure to degrade the oil was not due to growth inhibition by components of the oil. The experiments in OSS and OPPSS broths provide a basis for separating each isolate into one of three categories: unconditional degrader, conditional degrader, or non-degrader. The unconditional degraders can utilize oil as a sole source of carbon and energy. Isolates in this category are P, aeruginosa OSB 1 and A. faecalis OS2B from the basin site and E, americana CRFW, P. aeruginosa CRFYl, and F. odoratum CRLY from the stream site. The conditional degraders can degrade oil when an additional carbon source is present, in this case proteose peptone. The isolates in this category are E. cloacae OS2A 112 THE TEXAS JOURNAL OF SCIENCE-VOL. 47, NO. 2, 1995 Table 3. Growth of unconditional degrader strains on diesel fuel, mineral oil, naphthalene, and toluene as sole sources of carbon and energy. Isolate Growth in each medium * DSS MOSS NSS TSS Basin site 1. P. aeruginosa + + - - 2. A. faecalis OS2B + + - - Stream site 1. E. americana CRFW + + - - 2. P. aeruginosa CRFYl + + - - 3 . F. odoratum CRLY + ■ * growth, ( + ); no growth, (-). Growth was measured as described in Materials and Methods. See Table 1 for composition of media. from the basin site and S. marcescens CR2B and E. cloacae CRFY2 from the stream site. The only non-degrader in either medium was B. cereus OS3 from the basin site. Several attempts were made to induce growth of the conditional degraders and the non-degraders by culturing them in OSS broth in various combinations with the unconditional degraders (data not shown). All attempts failed; therefore, these species apparently cannot utilize the metabolites produced by the unconditional degraders during growth on oil. All species, however, could grow together in OPPSS broth and so growth inhibition of one species by another was not a factor in these results. No correlation was found between utilization of crude oil by an isolate and the ability to adhere to crude oil (Table 2). Among the unconditional degraders, only A, faecalis OS2B and F. odoratum CRLY adhered. Neither of the P. aeruginosa isolates adhered, a finding which is consistent with previous reports on this organism (Rosenberg, 1991). However, during growth of these organisms the medium became foamy and the oil appeared to be emulsified. A bioemulsifier may therefore have been produced, allowing these strains to utilize the oil without adhering to it. The other unconditional degrader, E. americana CRFW, did not adhere. Of the conditional degrader strains, only S, marcescens CR2B was observed to adhere to crude oil; this finding is consistent as well with previous experiments with this species (Rosenberg 1991). The non-degrader strain (B. cereus OS3) also was non-adherent as has been observed previously (Rosenberg 1991). BENOIT & WIGGINS 113 The unconditional degrader strains were tested for the ability to utilize various components of crude oil. Mineral oil was chosen to represent a mixture of n-aliphatic compounds; diesel fuel was chosen to represent a mixture of straight and branched chain hydrocarbons; and naphthalene and toluene represented specific aromatics. Results are given in Table 3. There was remarkable homogeneity among the strains. All utilized diesel fuel; all but F, odoratum CRLY utilized mineral oil; and none was able to utilize naphthalene or toluene under the experimental conditions. The conditional degraders and the non-degraders were not tested. Discussion The oily mousse from the basin and stream sites contained different groups of bacteria. Although both samples contained a Pseudomonas aeruginosa and an Enterobacter cloacae strain, the strains from each were biotypically different from each other. No other cultured species were common to both samples. These assemblages of bacteria, therefore, appear to have formed independently. The unconditional degrader strains represent species which are known to utilize crude oil for growth (Bartha & Atlas 1977) except for Erwinia, which to our knowledge has not been reported to do so. This isolate is unusual, therefore, and its presence in the oily mousse is unexpected since this genus is associated with plants in nature. Notable similarities were found in the physiological categories of bacteria in the samples from both sites. Both contained unconditional and conditional degrader strains. Furthermore, among the unconditional degraders, there was almost complete homogeneity with respect to the substrates (diesel, mineral oil, naphthalene, or toluene) that can be utilized for growth. The unconditional degrader strains may be important hydrocarbon oxidizing strains at the sampled sites, while perhaps the conditional degrader strains also consume or cooxidize a portion while utilizing an additional carbon source for growth. The laboratory experiments do not indicate that the unconditional degrader strains can supply carbon for growth of the conditional strains in the form of hydrocarbon oxidation products. The presence of extrachromosomal elements in the isolates from both sites is consistent with previous observations in other systems (Hada & Sizemore 1981; Leahy & Colwell 1990; Ogunseitan et al. 1987). In those systems, extrachromosomal DNA genes may have contributed to 114 THE TEXAS JOURNAL OF SCIENCE-VOL. 47, NO. 2, 1995 the ability of the isolates to degrade oil. It is unknown at this point whether there is a similar function for the DNA in the Oil Springs isolates. The presence of adherent strains in the samples is readily understood as cells of these organisms likely posses hydrophobic cell surfaces or fimbriae, or produce adhesins (Rosenberg 1991), which would maintain physical contact between the cell and the crude oil substratum. For example, the hydrophobic pigment, prodigiosin, of Serratia marcescens has been shown to increase adhesion of the cells to hydrocarbons (Rosenberg 1984) as well as has the serraphobin outer surface protein of this organism (Bar-Ness & Rosenberg 1989). Serratia marcescens CR2B produces the red pigment prodigiosin characteristic of the species and probably adheres at least for that reason. The presence of hydrocarbon non-adherent strains may be explained by their adherence to organic particles within the oily mousse or co¬ adhesion to the adherent strains. Pseudomonas aeruginosa for example is well-known to produce exopolysachharides which allow it to form biofilms on a variety of surfaces. Cells of other species can become trapped in these biofilms as well (Wilderer & Characklis 1989). The physiological similarities between the bacterial populations from the two samples may be a result of selection pressures exerted by the availability of the same crude oil at both sites. Crude oil from this location is high in C-20 or larger compounds with cyclic hydrocarbons present in lesser amounts (Pike 1977). The preference of the uncondi¬ tional degrader strains for diesel and mineral oil likely reflects this content of the crude oil. Non-utilization of naphthalene and toluene by the isolates might be expected since cyclic hydrocarbons are not present in high amounts in this oil. This along with the hydraulic flow that moves bacteria and oil downstream away from the seep may inhibit strains capable of utilizing the cyclic portion from colonizing the oil at the seepage sites. Common selection pressures, therefore, appear to have produced the two different groups of hydrocarbonoclastic bacteria with similar physiologies at both sites, an expected outcome in stable, similar microenvironments. The unconditional strains do not appear to provide nutrients for the conditional strains and so the interaction between them in the oily mousse, if any, is not clear. Further investigations will reveal whether some other interaction exists and whether additional similarities can be found at the two seepages among the sediment and free aquatic bacteria not cultured in this study. BENOIT & WIGGINS 15 Ackno wl edgements This work was supported in part by special item appropriation from the State of Texas. The authors gratefully acknowledge the assistance of the following students: R. Elton Sloan, Thomas Vyles, Aaron Schultz, Garlene Kayanan, and Kalyn Sowell. Literature Cited Atlas, R. M. 1981. Microbial degradation of petroleum hydrocarbons: an environmental perspective. Microbiol. Rev. 45:180-209. Bar-Ness, R. & M. Rosenberg. 1989. Putative role of a 70 kda surface protein in mediating cell surface hydrophobicity of Serratia marcescens . J. gen. Microbiol. 135:277. Bartha, R. & R. M. Atlas. 1977. The microbiology of aquatic oil spills. Adv. Appl. Microbiol. 22:225-266. Chakrabarty, A. M. 1976. Plasmids in Pseudomonas. Annu. Rev. Genet. 10:7-30. Colwell, R. R. & J. D. Walker. 1977. Ecological aspects of microbial degradation of petroleum in the marine environment. Crit. Rev. Microbiol. 5:423-445. Cooney, J. J. 1984. The fate of petroleum pollutants in freshwater ecosystems. M R. M. Atlas (ed.). Petroleum Microbiology, pp. 399- 433. MacMillan Publishing Co., New York. Floodgate, G. 1984. The fate of petroleum in marine ecosystems. In R. M. Atlas (ed.). Petroleum Microbiology, pp. 355-397. MacMillan Publishing Co., New York. Hada, H. S. & R. K. Sizemore. 1981. Incidence of plasmids in marine Vibrio spp. isolated from an oil field in the northwestern Gulf of Mexico. Appl. Environ. Microbiol. 41:199-202. Hood, M. A., W. S. Bishop, S. P. Myers, & T. Whelan, III. 1975. Microbial indicators of oil-rich salt marsh sediment. Appl. Microbiol. 30:982-987. Leadbetter, E. R. & J. W. Foster. 1960. Bacterial oxidation of gaseous alkanes. Arch. Mikrobiol. 35:92-100. Leahy, J. G. & R. R. Colwell. 1990. Microbial degradation of hydrocarbons in the environment. Microbiol. Rev. 54:305-315. Llanos, C. & A. Kjoller. 1976. Changes in the flora of soil fungi following oil waste application. Oikos 27:377-382. Ogunseitan, O. A., E. T. Tedford, D. Pacia, K. M. Sirotkin, & G. S. Sayler. 1987. Distribution of plasmids in groundwater. J. Ind. Microbiol. 1:311-317. 116 THE TEXAS JOURNAL OF SCIENCE— VOL. 47, NO. 2, 1995 Olsen, K.D. & R. K. Sizemore. 1981. Effects of an established offshore oil platform on the autochthonous bacterial community. Dev. Ind. Microbiol. 22:685-694. Pate, C. S. 1987. A subsurface study of the queen city formation in Nacogdoches and Angelina counties. Unpublished thesis, Stephen F. Austin State University, Nacogdoches, 160 pp. Pike, R. W. 1977. A study of the biodegradation of an east Texas shallow well crude oil and a deep well crude oil using an autochthounous bacterium. Unpublished thesis, Stephen F, Austin State University, Nacogdoches, 46 pp. Pinholt, Y. S., S. Struwe, & A. Kjoller. 1979. Microbial changes during oil decomposition in soil. Holarct. Ecol. 2:195-200. Rosenberg, M., D. Gutnick, & E. Rosenberg. 1980. Adherence of bacteria to hydrocarbons: a simple method for measuring cell-surface hydrophobicity. FEMS Microbiol. Lett. 9:29-33. Rosenberg, M. 1984. Isolation of pigmented and non-pigmented mutants of Serratia marcescens with reduced cell surface hydrophobicity. J. Bacteriol. 160:480. Rosenberg, M. 1991. Basic and applied aspects of microbial adhesion at the hydrocarbon: water interface. Crit. Rev. Microbiol. 18: 159-173. van Eyk, J. & T. J. Bartles. 1968. Paraffin oxidation in Pseudomonas aeruginosa: I. Induction of paraffin oxidation. J. Bacteriol. 96:706- 712. Wilderer, P. A. & W. G. Characklis. 1989. Structure and function of biofilms. In Characklis and Wilderer (eds.). Structure and Function of Biofilms, pp. 5-17. John Wiley and Sons, New York. ZoBell, C. E. 1946. Action of microorganisms on hydrocarbons. Bacteriol. Rev. 10:1-49. TEXAS J. SCI. 47(2): 117-143 MAY, 1995 A NEW SUBSPECIES OF THE POLYTYPIC LIZARD SPECIES SCELOPORUS UNDULATUS (SAURIA: IGUANIDAE) FROM NORTHERN MEXICO Hobart M. Smith, David Chiszar and Julio A. Lemos-Espinal Department of EPO Biology , University of Colorado, Boulder, Colorado 80309-0334 Department of Psychology , University of Colorado, Boulder, Colorado 80309-0345 Laboratorio de Conservacion, Escuela Nacional de Estudios Profesionales Iztacala, UN AM, Apartado Postal 314, Tlalnepantla, Estado de Mexico, Mexico NhsiT2iQi.—Sceloporus undulatus belli is described from northern Mexico. This new taxon represents the southernmost known subspecies and ranges from northwestern Chihuahua to southern Coahuila, northwestern Zacatecas and northeastern Durango. In northern central Chihuahua it occurs within a few km of the known range of Sceloporus undulatus speari without evidence of intergradation. Sympatry is possible and at least close parapatry appears likely. These two subspecies are assigned to different exerges; the new taxon is assigned its own and is separate from the adjacent consobrinus exerge. The relationships of the four exerges and 1 1 subspecies currently assigned to this species are discussed, comparisons made, and a key proposed. Taxonomic intraspecific parapatry is exemplified in the widespread lizard species Sceloporus undulatus (Bose & Daudin) in both primary (in situ) and secondary (circular range approximation) contexts (Figure 1). Instances of the former are numerous, but only two probable examples of the latter have so far been noted. In one example, S. undulatus garmani Boulenger and S, undulatus erythrocheilus Maslin (of the consobrinus and tristichus exerges, respectively) assuredly have closely approximated ranges in Baca and perhaps Boulder counties, Colorado (Maslin 1964; Hammerson 1982). So markedly different are these two taxa, even at their closest range approximation (a few km) that they would qualify as distinct species were it not for the indirect linkage provided by the obvious intergradation of S. undulatus consobrinus Baird & Girard with not only S. undulatus garmani but also with S. undulatus tristichus Cope, with which S, undulatus erythrocheilus in turn clearly intergrades. Their habitat preferences differ so strongly that at least macrosympatry could occur. The second example of circular range approximation occurs between S. undulatus elongatus Stejneger and S, undulatus erythrocheilus in southeastern Wyoming, where the former subspecies, otherwise limited to western slopes, occurs in the upper reaches of the eastward-draining 118 THE TEXAS JOURNAL OF SCIENCE— VOL. 47, NO. 2, 1995 SMITH, CHISZAR & LEMOS-ESPINAL 119 North Platte River, which passes through the range nearby, in much the same habitat, of S. undulatus erythrocheilus . At present the ranges of the two subspecies are known to come no closer than about 40 km, but the distribution of this species in this area is poorly known; the two subspecies probably occur considerably nearer to one another than is now apparent. Since they belong to the same exerge, and have the same habitat preferences, interbreeding rather than sympatry would be a likely outcome were their ranges to contact one another. Both of these cases of circular range approximation involve populations whose ranges extend northward on either side of the natural barriers of the Rocky Mountains or the Great Plains. No such examples, however, have been expected in this species in the southern part of its range, in the absence of comparable barriers and of known subspecific differentiation. It is therefore of great interest that this study of Mexican populations has revealed a third case of close range approximation without intergradation. The recently described subspecies of the consobrinus exerge, S. undulatus speari Smith et al. (1994), is limited, so far as known, to the vast sand dunes of the northern central part of the state of Chihuahua. Only a short distance to the south and west of this area, and extending southeastward into Durango, Zacatecas and Coahuila, occurs a distinctly different subspecies bearing much the same geographic relationship to S. undulatus speari as S. undulatus erythrocheilus does to S, undulatus garmani. This new taxon is herein described. Specimens are deposited with the University of Colorado Museum at Boulder (UCM), the Museum of Comparative Zoology at Harvard (MCZ), the University of Michigan Museum of Zoology at Ann Arbor (UMMZ), the University of New Mexico Museum of Southwestern Biology at Albuquerque (MSB) and the University of Texas at El Paso (UTEP). Sceloporus undulatus belli, new subspecies (Figures 2 & 3) Holotype—kdnXi male (UCM 41539), 2 mi S of Leon Guzman, Durango, Mexico, 19 June 1966, collected by Richard L. Holland. Paratypes .—Om hundred and fifty-six specimens in the following museums. MCZ.— Six specimens, all from Chihuahua: 32.5 km ESE Cd. Cuauhtemoc, 1850 m (126874); 12.5 km SE Moctezuma (126854); 16 120 THE TEXAS JOURNAL OF SCIENCE- VOL. 47, NO. 2, 1995 Figure 2. Dorsal views of S. undulatus belli. Upper, male holotype (UCM 41539, 68 mm SVL), 2 mi S Leon Guzman, Durango. Lower, female paratype (MCZ 126856, 77 mm SVL), 16 km SSE Moctezuma, Chihuahua. The cross-barred pattern is typical of females, the nearly unicolor dorsum with prominent, broad lateral dark stripes typical of males. km SSE Moctezuma (126855-7); 2 km W Ricardo Flores Magon, 1485 m (126933). UMMZ.—Thxtt specimens, all from Chihuahua: 11 mi W Cd. Cuauhtemoc, 6850 ft (1 18971 A-B); 30 mi S El Sueco, 4700 ft (118972). MSB.— 20 specimens, all from Chihuahua: 16 mi E Aldama, Hy 16 (33222, 33411); 4.5 mi S Cd. Camargo, Hy 45 (7196-7); Cd. Chihuahua, International Airport runway (39945); jet Hy 16 and 45, SW Cd. Chihuahua (56335); 20.5 mi N Cd. Chihuahua, Hy 45 (8956-7); 17.1 mi W El Carmen (8968); 9. 1-9.3 mi S El Sueco, Hy 45 (8961, 8966); General Trias, centro (33223); Ojo de los Reyes, 4.7 mi SE Galeana, 1 mi N Angostura (39946-7); ruins of Paquime, Casas Grandes Viejo (35041-2); 9 mi SE Rio Florida in Jimenez (8922); 15 km E, 7 km N San Buenaventura, Hy 10 (35043-5). UCM.— 10 specimens, from Chihuahua and Durango: Chihuahua: 15.6 mi S Cd. Chihuahua (41504); 22 mi S Cd. Chihuahua (41505-13); 3 mi S Encinillas (24285); 7 mi NE Escalon (16812-4, 16816-8); 18 mi SMITH, CHISZAR & LEMOS-ESPINAL 121 Figure 3. Ventral views of the same specimens as in Figure 2, Note the typical, unmarked venter of this large female, and the typical, extensively fused gular semeions, with no anterior or medial edging with black, in the large male. S Gallegos (16801-11); S Jimenez (41518); 7 mi N Jimenez (49677); 7 mi S Jimenez (49516-7); 12.5 mi S Jimenez (41519-32); Meoqui, S side Rio San Pedro (41514); 8.6 mi S Moctezuma (41502-3); 2.5 mi S Saucillo (41515). Durango: 2 mi N Cuencame (41535-7); 20.4 mi S Cuencame (41538); 19 mi SW Gomez Palacio, Rio Nazas, 3800 ft (51414); Leon Guzman (41542); 1.1 mi N Leon Guzman (41534); 1.7 mi SW Leon Guzman (41540-1); 5 mi NE Pedricena (50001); 17 mi S Rodeo, 5500 ft (24286). UTEP.—51 specimens, fom Chihuahua, Coahuila, Durango and Zacatecas. Chihuahua: Aldama, 20.3 rd mi NE, Hy 16 (9238); Ascencion, 9.9 rd mi SW (4271); Camargo, 1 mi E (4513); Camargo, 15 mi N (4536); Camargo, 15 mi S, Hy 45 (9443); Camargo, 16 mi W, Presa Boquilla (9442); Cd. Chihuahua, 4 km NW, Lago Jacales (9441); Cd. Chihuahua, 5 km NW, Casa Salud (9437-40); El Sauz, 11 mi SSE turn-off to, Hy 45 (4518-20); El Sauz, 16 mi NNE turn-off to, Hy 45 (4537-8); El Sueco, Hy 10 3.2 km WNW jet Hy 45 and 10 (14514); 122 THE TEXAS JOURNAL OF SCIENCE- VOL. 47, NO. 2, 1995 Escalon, 13 rd mi NW, Hy 49 (9236); Gallego, 1.5 mi N (4529-30); Gallego, 6 mi S (4531); Jimenez, 20 mi SE (4515-7); La Perla, 15.2 rd mi N (9237); Parral, 7 mi E (4512); Ricardo Flores Magon, Rio del Carmen (4514). Coahuila: El Chiflon, 36 km W Saltillo, Hy 40 (4459, 6707-8, 6783-5); Francisco E. Madero (Chavez), 3 km E, Hy 30 (6779, 6811, 7567-8). Durango: Atotonilco (4532-3); Ceballos, 2 mi SE (4539); Ceballos, 2.4 mi E (6415); Mapimi, 17.7 mi W, Hy 30 (4525-6); Sombreretillos, 7.9 rd mi SE (9514); Zavalza, 6.4 mi SE, Hy 49 (4524). Zacatecas: An^uac, 2 km E (4460); Anahuac, 3 km E (6802-6); Anahuac, 6 km E (7465); Camacho, 27.8 mi NE, rd to Mazapil (9445); Cinco de Mayo, 5 km NE (4446); La Presa de Junco, 3.5 rd mi W, 24 17’45”N, 101 4r45”W (3612); Rio Grande, 13 rd mi SSE (4540-1). Range (Figure 1). — Extreme northwestern Chihuahua (only —50 km from Hidalgo County in the southwestern corner of New Mexico) southeastward east of the Sierra Madre Occidental through central Chihuahua to northeastern Durango, and eastward through northwestern Zacatecas and southern Coahuila. Diagnosis. — A member of the polytypic species Sceloporus undulatus, assigned to its own monotypic exerge (refer to Recommendation 6B of Article 6 of of the International Code of Zoological Nomenclature 1985, for use of this term in lieu of "subspecies group"), distinguished by the combination of usually poorly developed dorsolateral light lines; conspicuous dorsal crossbars in females, but usually a nearly unicolor, unmarked dorsum in males, with a broad, conspicuous lateral dark line (a sexual dimorphism unique to the subspecies); almost invariably, complete absence of ventral markings in females except for small, dim and diffuse gular semeions in about half of the specimens; blue component of gular semeions always fused, usually fully but at least in part in all males except juveniles (<45 mm SVL; unique to the subspecies), not or weakly black-bordered anteriorly or medially. Differences from other subspecies are summarized in the following sections on Relationships, Subspecific Key and Adjacent Subspecific Comparisons. Description of Holotype.— An adult male (Figures 2 & 3), 68 mm SVL. Four postrostrals; four internasals in a row between middle of nasals; three scales on either side of midline between median frontonasal and rostral; frontonasal s in contact side to side, not subdivided; prefrontal s in median contact, not subdivided; frontal in two parts. SMITH, CHISZAR & LEMOS-ESPINAL 123 anterior and posterior, not further subdivided; interparietal and frontal in contact; 1-1 frontoparietals. Supraoculars 5-5; 8-9 oculociliaries; 1-1 subnasals, separated by one row of lorilabials from supralabials; 2-2 canthals, in normal position; 1-1 preoculars; lorilabial rows reduced to 1-1 by at least one scale below subocular; loreals 1-1. Outer labiomental row ending anteriorly opposite 1-1 infralabials, inner row opposite 3-3 infralabials. Dorsals 39, minimum count interparietal to level of rear margin of thighs held at right angles to body axis; femoral pores 17-18; interfemoral pore scales six. Dorsal ground color a light tan, without a dorsolateral series of dark spots; a well-defined lateral line, dark brown, 1-3 scale rows wide, not sharp-edged, from orbit to base of tail, narrowest on neck, widest at midabdomen, bordered medially by a dim dorsolateral light line one and two half scale rows wide; a broad, grayer vertebral area four and two half scale rows wide. Sides of abdomen below lateral dark line light tan, merging ventrally with the light blue abdominal semeions (AS) that extend between levels near axilla to near groin; a narrow medial black border on AS, not reaching either axilla or groin; a minimum of four scales between AS black borders. No black flecks on chest, inter- AS abdomen, or ventral surfaces of limbs or tail. Gular semeions (GS) mostly light blue, narrowly bordered posteriorly by an extension from a black patch in each nuchal pocket, extending toward but not onto foreleg. No black pigment bordering GS anteriorly or on sides, and none anterior to GS. A short, narrow, black divider posteriorly between the right and left halves of the GS, and a dim, median whitish indentation anteriorly between the two blue halves. Variation. — The most conspicuous deviation from the preceding description occurs in female coloration. Females invariably have a dorsolateral series of dark spots (Figure 2); in a few specimens (e.g., UCM 16809, 41522) they are fused to form a more or less continuous dorsolateral dark line. Typically some or all of the spots form transverse bars that cross the dorsolateral light line, and where this occurs, frequently a bright light spot in the dorsolateral light line lies immediately posterior to the crossbar. A narrow lateral light line is often present, bounded ventrally in some specimens with a still narrower, irregular sublateral dark line, and dorsally by the area occupied in males by the lateral dark line, which is typically poorly defined in females. 124 THE TEXAS JOURNAL OF SCIENCE- VOL, 47, NO. 2, 1995 On the contrary, males are usually essentially unicolor light brown between the lateral dark stripes, with no dark markings whatever or, at most, a few very narrow, dark, longitudinal streaks. The dorsolateral light lines are typically obscure in fully mature males, distinct in juveniles. The lateral dark line is invariably broad, continuous and uninterrupted. A lateral light line is usually present ventral to the lateral dark line, and is bordered ventrally by a narrow, sublateral dark line. Some males, typically young (e.g. UNM 8961, 8966, 35043, 35045, 39947, 56335) are exceptional in having a dorsolateral series of dark marks; in most they are much less prominent than in females, but they do interrupt the dorsolateral light line. Sexual dichromatism is equally conspicuous ventrally (Figure 3). With one exception (UTEP 4524), no female has a well-defined GS, and less than a third (18 of 63) exhibit any evidence whatever of them; in the mentioned exception they are distinct but small and fused. In one other they are clearly evident, but weak, and in the others they are very dim, diffuse and widely separated. In only one of 63 (the same exception, UTEP 4524, noted previously) is there any evidence whatever of AS; in that exception they are poorly defined but certainly visible. Otherwise the only dark markings are a very narrow midventral streak, often interrupted or faint, or both, and in a few specimens a few similar streaks on throat or chest. All males, on the contrary, with a SVL of 43 or more (81), have conspicuous AS and GS, with the exception of one (of two) at 43 mm SVL; that exception, and four smaller ones, show no evidence of either GS or AS. The blue components of the GS are fused in every male, although their basically paired nature is frequently evidenced by a partial median division. In no other subspecies are the blue components of the GS fused; if the GS are fused at all, it is the black borders that are involved. The AS are invariably separated, by a minimum of 1-7 scales (one, 2; two, 4; three, 8; four, 18; five, 14; six, 7; seven, 1). Usually no dark ventral pigmentation, other than the AS, occurs on chest, midabdomen or legs, except for the fine midventral line and dim streaks on chest of a few. The AS extend onto the thigh in the largest specimens, and the GS barely reach the foreleg in a few. The GS are bordered posteriorly (but seldom anteriorly or medially, and narrowly in those exceptions) by black or dark blue (not readily distinguishable) in some, and some brown pigment is scattered over the throat anterior to the GS in most specimens. Thus, unlike other subspecies of S. undulatus, S. undulatus belli has SMITH, CHISZAR & LEMOS-ESPINAL 125 the two sexes trenchantly distinguishable in almost all fully mature adults in both dorsal and ventral color /pattern; this is a unique feature of the subspecies. Variation in scalation does not appear to be sexually dimorphic, hence the following data are not sorted by sex. Postrostrals usually (84) four, three in six, two in two. Internasals in a row between nasals usually (85) four, three in one, two in four. Scales between median frontonasal and rostral, on either side of median line, usually (64) 3-3, 2-2 in three, 2-3 in six, 3-4 in eight, 4-4 in six. Frontonasals usually (82) undivided and in serial contact, subdivided or separated on one or both sides in eight. Prefrontals usually (50) in contact medially, separated by an azygous scale in 24, by contact of frontal and median frontonasal in 17. Frontal usually (78) divided into anterior and posterior sections, without subdivision, abnormal (subdivided, undivided or with anomalous fusions) in 12. Frontal usually (79) in contact with interparietal, separated by an azygous scale in two, by contact of frontoparietal s in nine. Frontoparietals usually (76) 1-1, but 1-2 in 11, 2-2 in two, 2-3 in one. Supraoculars most frequently (25) 5-5, but 5-6 in 20, 5-7 in two, 6-6 in 23, 6-? in two, 6-7 in five, 7-7 in 12, 8-8 in one. Total (both sides) oculociliaries 14 (three), 15 (one), 16 (three), 17 (five) 18 (eight), 19 (five), 20 (five), 21 (seven), 22 (ten), 23 (seven), 24 (nine), 25 (three), 26 (six), 27 (five), 29 (three), 30 (six), 31 (one), 32 (one). In 6 of 83 specimens, one or more supraoculars are in at least narrow contact with one or more median head scales on one or both sides. The subnasal is abnormal in three of 88 (1-2 in one, fused with anterior canthal on one side of two). The canthals are normal in 71; the first is displaced dorsally on one side in three, is in contact with the lorilabials on one side in five, both sides of eleven, fused with the loreal on both sides in two, one side in two, and fused with subnasal on one side of two. The preoculars are 1-1 in 81, 1-2 in four, 2-2 in two, 2-3 in three. The two rows of lorilabials between subocular and supralabials are complete, or are reduced to one at some point on one or both sides as follows: 1-1, 32; 1-2, 20; 2-2, 36; 1-?, 1; 2-?, 2. The loreals are 1-1 in 74, 1-2 in 6, 2-2 in 4, 2-3 in one, 3-3 in one, and are fused with the first canthal on both sides in two, one side in two. The outer labiomental row extends anteriorly to the mental (M) or to the level of the 1st or 2nd infralabial as follows: 1-1, 77; 1-?, 2; 2-2, 1; M-1, 5; M-M, 5; M-?, 1. The inner labiomental row extends anteriorly to the levels of the 2nd, 3rd or 4th infralabials as follows: 2-2, 1; 2-3, 2; 3-3, 84; 3-4, 2. Dorsal scale count 35-44 (35, one; 36, one; 37, four; 38, eight; 39, 126 THE TEXAS JOURNAL OF SCIENCE-VOL. 47, NO. 2, 1995 thirteen; 40, twelve; 41, eighteen; 42, fifteen; 43, ten; 44, four). Femoral pores 14-20 (14, ten; 15, twenty-eight; 16, sixty-two; 17, fifty; 18, twenty-one; 19, eight; 20, two). Minimal scale count between femoral pore series, 3-8 (three, 1; four, 18; five, 34; six, 26; seven, 8; eight, 2). Lamellae under free part of 4th toe 20-24 (20, two; 21 , eight; 22, four; 23, three; 24, three). SVL parameters (in mm) in the 54 males (and parenthetically in the 37 females) are: range, 43-74 (52-77); 40-49, 1(0); 50-59, 17 (9); 60-69, 32 (21); 70-77, 4 (7). Etymology. — The subspecies is named for Dr. Edwin L. Bell, a retired Professor of Biology at Albright College, Reading, Pennsylvania, in recognition of his meticulous study of geographic variation in the western member, S. occidentalism of the undulatus exerge. He also first observed and brought attention to the distinctness of the present taxon. Habits and Habitat.— Tht authors have had no experience with this subspecies in the field. Tanner (1987: 396) stated that most specimens he there reported of "S. undulatus consobrinus" (all 46 now referable to S. undulatus belli except for one S. undulatus speari from 30 mi S Cd. Juarez) were "seen and collected ... along the valley roads, many of them sunning on rocks and mounds of soil which had been left as the roads were constructed." Dr. Robert G. Webb, who has observed the species over many years in Chihuahua, states (pers. comm.) that "The S. undulatus in eastern Chihuahua south of El Sueco into northeast Durango are generally ground-dwellers in areas of small bushes, shrubs and ground debris. The general area is grassland with scattered shrubs that may be mesquite, creosote, acacia and/or cholla, the terrain either flat with hard-packed clay soils or more hilly with a rocky substrate. During inclement weather (overcast, windy, rather cold with rain threatening) lizards have been found under rocks. In looking at my field notes, occasional specimens have been noticed on fence posts, associated with cement- supported road culverts, and a few inches above ground on low shrub branches." Field notes on UMMZ 118971(2) from 11 mi W Cuauhtemoc, Chihuahua, state that the specimens were taken at 6850 ft., in steep, rocky hills of a pass, where there were oaks and some pines and junipers, with bushes and weeds between rocks. UMMZ 118972 from 30 mi S El Sueco, Chihuahua, was taken at 4700 ft. , on bushy flats with scattered clumps of weeds and grass amongst tall mesquite trees, with bare earth between; the lizard was asleep in a bush. SMITH, CHISZAR & LEMOS-ESPINAL 127 Intensive searches for S, undulatus conducted as a part of this study from Sept. 28 - Oct. 4, 1993 (Lemos-Espinal et al. 1994) in the extensive sand dunes east and west of Hy 45 south of Cd. Juarez, as far south as El Sancho, revealed only S. undulatus speari, no S, undulatus belli. It appears certain that the latter does not occur in the dunes area, although it is here recorded from as near as 12.5 km SE Moctezuma, and farther west it occurs as far north as does S. undulatus speari. Whether either subspecies occurs in the intervening territory remains to be determined. Relationships This study retains the same taxa and exerges as proposed by Smith et al. (1992), but adds another exerge to accommodate S. undulatus belli, adds S. undulatus speari to the consobrinus exerge (as proposed by Smith et al. 1994), and indicates the proposed derivation of each taxon as indicated primarily by the existence of intergradation. None is known between S. undulatus speari and its presumed ancestor, S. undulatus consobrinus, but may well occur in areas between the known limits of their geographic ranges, whence no material is now available although no barrier to the occurrence of the species appears to exist. Sceloporus undulatus belli, with its relatively large size, cross-barred females, usually unspotted males, almost complete lack of ventral markings in females, the light gular semeions with their unique fusion of the blue components in mature males, and its terrestrially cursorial habits and habitat, fits with none of the other exerges. The undulatus exerge is limited to eastern North America, hence is not involved, although it is different in being scansorially arboreal, cross-barred dorsally and more heavily pigmented ventrally in both sexes. Members of the consobrinus exerge are relatively small, their dorsolateral light stripes tend to be bright, dorsal spotting is usually present in both sexes except in the bleached, arenicolous taxa, and all are terrestrially cursorial. Members of the tristichus exerge tend to have distinct, dark dorsal crossbars or spots in both sexes, are scansorially rupicolous, are heavily pigmented ventrally in both sexes, and are relatively large. Subspecific Key Sceloporus undulatus consobrinus appears twice in this key, in adjustment to the variation in the intensity of the dorsolateral light stripe, consistently bright in eastern populations, but subdued in some western 128 THE TEXAS JOURNAL OF SCIENCE— VOL. 47, NO. 2, 1995 intergrade populations, through the influence of S, undulatus tristichus. Sceloporus undulatus speari likewise appears twice to account for the difference between the bleached summer dorsal pattern and the darker winter one. lA. B. 2A. B. 3A. B. 4A. B. 5A. B. 6A. B. 7A. B. 8A. B. 9A. B. lOA B. IIA B. East of the Mississippi River and dorsal scale count, interparietal to level of rear margins of thighs held at right angles to base of tail, less than 38 (92%; Smith, 1938) . S. undulatus undulatus Not as above . 2 Dorsal count 46 or more (86%; Smith 1938) . S. undulatus elongatus Fewer (>85%; Smith, 1938, and present data) . 3 Dorsum bleached, whitish . 4 Not as above . 6 Gular semeions absent or vestigial in both sexes; maximum SVL 54 mm in males, 59 mm in females . S. undulatus tedbrowni Gular semeions well developed, at least in males; maximum SVL 60-64 mm in males, 69 mm in females . 5 Semeions not black-bordered . S. undulatus speari Semeions black-bordered . S. undulatus cowlesi Gular semeions absent in both sexes . S. undulatus garmani Gular semeions prominent at least in males . 7 Females without ventral markings, or with only dim, diffuse gular semeions, and boldly cross-barred dorsally; fully adult males usually without dark marks dorsally, blue components of gular semeions at least partially fused, usually not black-bordered medially or anteriorly . S. undulatus belli Not as above . 8 Dorsolateral light lines distinct, broken or not . 9 Dorsolateral light lines poorly defined, broken . 10 Semeions not black-bordered in either sex . S. undulatus speari Semeions black-bordered . S. undulatus consobrinus Black-bordered abdominal semeions present in both sexes . S. undulatus tristichus Abdominal semeions not present in females, or if present not black- bordered . 11 Dorsum prominently cross-barred; venter with scattered black pigment . . 12 Not as above . S. undulatus consobrinus 12A West of the Great Plains; males with lips and throat reddish or yellow¬ ish at least in breeding season . S. undulatus erythrocheilus B. East of the Great Plains; no red or yellow on head in either sex at any season . S. undulatus hyacinthinus Adjacent Subspecific Comparisons As indicated in the preceding key to subspecies, the standard criteria of dorsal pattern, ventral pattern, size and habits/habitat provide most SMITH, CHISZAR & LEMOS-ESPINAL 129 Table 1. Chi-square values and percent occurrences for statistically significant mor¬ phological differences between S. undulatus belli and S, undulatus tristichus Character- state N chi2 df P % 14 90(291) 72.21 1 <.001 71(23)* 2 33(159) 61.36 1 <.001 85(18)* 9 90(298) 56.74 1 <.001 90(45) 18 88(276) 52.79 1 <.001 72(28)* 6 86(286) 44.41 1 <.001 66(27) 16 89(301) 19.85 1 <.001 7(30) 10 91(300) 8.57 1 <.005 45(62) The number of S. undulatus belli is followed parenthetically by the number of S. undulatus tristichus. Sequence the same in percent column. Diagnostically significant differences are indicated by an asterisk. See text for character-states. character- states readily differentiating the eleven recognized subspecies from each other. Dorsal scale count is useful in distinguishing S, undulatus undulatus and S, undulatus elongatus from all other subspecies, and from one another, but scalation has been little utilized otherwise in subspecies characterization. Data taken on certain features of external morphology of the four subspecies of immediate concern relative to S. undulatus belli (S, u.consobrinus, S. undulatus speari, S, undulatus tristichus) reveals that a few diagnostically useful as well as significant but individually non-diagnostic differences between those taxa do exist among those features, as well as in the standard criteria. Hence these subspecies are demonstrably more than mere pattern classes. Chi-square values of all compared morphological characteristics are given in Tables 1-5, where the character-states are given the following numbers: tail/total length ratio (1) .58 or more, or (2) .59 or more, or (3) .60 or more, or (4) .62 or more; dorsal scale count (5) 39 or less, or (6) 41 or less, or (7) 41 or more; scales between femoral pore series (8) five or more; intemasals in a line between median frontonasal and rostral, on each side of median line, (9) 3-3 or more; prefrontals (10) separated medially; frontal (11) separated from interparietal, or (12) abnormal (fused, subdivided, tic.); frontoparietals (13) more than 1-1; supraoculars (14) 5-6 or more, or (15) 5-7 or 6-6 or more, or (16) one or more in contact with median head scales on one or both sides; total oculociliaries , on both sides, (17) 19 or more, or (18) 20 or more; canthals (19) with the 1st contacting lorilabials on one or both sides, or (20) abnormal (fused, subdivided, contacting lorilabials, or displaced); preoculars (21) 1-2 or more; lorilabials below subocular (22) in two 130 THE TEXAS JOURNAL OF SCIENCE— VOL. 47, NO. 2, 1995 Figure 4. Dorsal views of S. undulatus tristichus. Upper, male (UCM 7378, 65 mm SVL, 9 mi E Youngsville, Rio Arriba Co., New Mexico). Lower, female (UCM 7329, 58 mm SVL, 2 mi S Chromo, Archuleta Co., Colorado). Note the weak dorsolateral light lines (as compared with S. undulatus speari and S. undulatus consobrinus) , the dark markings in the male (as compared with S. undulatus belli, S. undulatus speari) . complete, uninterrupted rows, or (23) reduced to one row by one or more scales on one or both sides. Differing character- states in a single character in the preceding list, e.g. 1-4, 5-7, etc., were selected ad hoc for maximum discrimination between the particular compared taxa, and would be useless in compari¬ son of certain other taxa. Table 6 documents the variation of all 23 selected character- states in the four subspecies of present concern. Because many statistical tests were conducted, it was inappropriate to evaluate all of them at the 0.05 level of significance, as this could have elevated the rate of type I errors to an unacceptable level. To protect the over- all error rate, individual comparisons were judged to be significant at the 0.005 level (Kirk 1982:101-106). Sceloporus undulatus belli, for example, is diagnostically distinguished from S. undulatus tristichus (from whose range it is narrowly separated by the western arm of S. undulatus consobrinus; see Figure 1), the nearest subspecies of comparable size (maximum SVL 71 mm in males, 75 mm in females, vs 74 mm and 77 mm, respectively. SMITH, CHISZAR & LEMOS-ESPINAL 131 Figure 5. Ventral views of S. undulatus tristichus. Upper, male, same as in Figure 4. Lower, female (UCM 7635, 64.5 mm SVL, 5 mi E Gallina, Rio Arriba Co., New Mexico). Note the medial and anterior black borders of the gular semeions in the male, and the remarkably masculine markings of the female. in S. undulatus belli) and with reduction of the dorsolateral light stripes, in three morphological characters, and, in addition, statistically significant differences exist that are not individually diagnostically useful in four others (see Table 1). Five pattern differences are easier to evaluate (although less objective) and are the primary basis for distinguishing these two subspecies (Figures 2-5): (1) complete separation of gular semeions in males (0% in S, undulatus belli, 71 % in S, undulatus tristichus)', (2) gular semeions absent or faint in females (100% in S, undulatus belli, 5% of S. undulatus tristichus)', (3) absence of any indication of abdominal semeions in females (100% in S. undulatus belli, 0% in .S', undulatus tristichus)', (4) absence of black median and anterior borders of the gular semeions in males (100% of 5'. undulatus belli, 0% of .S', undulatus tristichus)', and (5) absence of scattered black pigment on venter (95% in S. undulatus belli, 0% in S. undulatus tristichus). In addition, S. undulatus belli is reputedly terrestrial and cursorial, S. undulatus tristichus petricolous and scansorial. 132 THE TEXAS JOURNAL OF SCIENCE-VOL. 47, NO. 2, 1995 Figure 6. Dorsal views of S. undulatus consobrinus from the western arm of the subspecies’ distribution, in southeastern Arizona. Left, female, UCM 56071, 55 mm SVL, Santa Cruz Co., 5 mi SE Elgin. Center, a female, UCM 56884, 62 mm SVL, Santa Cruz Co., Audubon Research Ranch, 6 mi SSE Elgin. Right, a male, UCM 56886, 64 mm SVL, same locality as the preceding. No scalation features distinguish S. undulatus belli from S. undulatus consobrinus, although the latter is smaller (74 mm SVL maximum and 67% 60 mm or more in males, and 77 mm, 76% 60 mm or more in females of S, undulatus belli; 72 mm and 27% 60 mm or more in 131 males, and 74 mm, 39% 60 mm or more in 122 females of S. undulatus consobrinus) . The primary distinctions are in pattern (Figures 2-3 , 6-7) , S. undulatus belli having (1), with rare (one) exception, gular semeions faint or absent in females (vs distinct); (2) usually no anterior or medial black borders on broadly fused, blue gular semeions in males (vs present, blue not fused); (3) dorsolateral light stripes usually absent or poorly defined in fully mature adults (vs well defined throughout life); and (4) a strong sexual dimorphism in dorsal pattern, females being brightly cross-barred, males unspotted and unicolor between lateral dark lines (vs no distinct pattern dimorphism, both sexes spotted or cross- SMITH, CHISZAR 8l LEMOS-ESPINAL 133 ^ 9 SS ''ui Figure 7. Ventral views of S. undulatus consobrinus, same specimens and arrangement as in Figure 6. barred dorsally). Four significant but individually non-diagnostic differences exist; see Table 2. Similarly, no scalation features distinguish S. undulatus belli from S. undulatus speari, although the latter is smaller (74 mm SVL maximum and 67% 60 mm or more in males, 77 mm, 76% 60 mm or more in females of S. undulatus belli', 64 mm and 14% 60 mm or more in males, 69 mm and 38% 60 mm or more in females of S. undulatus speari). Three statistically significant but independently non-diagnostic morphological differences between the two subspecies occur, however; see Table 3. Pattern differences are diagnostic (see figures in Smith et al. 1994), with S. undulatus belli (1) having the dorsolateral light lines usually absent or poorly defined, at least in fully mature adults (vs well developed); (2) a strong sexual dimorphism in dorsal pattern, with females strongly barred and fully adult males usually without dark marks 134 THE TEXAS JOURNAL OF SCIENCE— VOL. 47, NO. 2, 1995 Table 2. Chi-square values and percent occurrences for statistically significant mor¬ phological differences between S. undulatus belli and S. undulatus consobrinus Character- state N chi2 df P % 22 90(170) 19.34 1 <.001 64(36) 9 90(171) 19.08 1 <.001 90(65) 11 90(170) 8.08 1 <.005 12(28) 21 90(171) 6.93 1 <.005 10(23) The number of S. undulatus belli is followed parenthetically by the number of S. undulatus consobrinus. Sequence the same in percent column, where none are of individual diagnostic significance. See text for character-states. and with a unicolor dorsum between the lateral dark lines (vs the same striped pattern in both sexes); (3) 100% of the males with the gular semeions in contact (vs 21 %); and, significantly but non-diagnostically, 1% of females with distinct gular semeions (vs 35%). The most distinctive single feature of S. undulatus belli, distinguishing it not only from the three subspecies most closely associated geographically, dealt with in the three preceding paragraphs, but from all others of the species, is the fusion, partial or complete (usually the latter), of the blue components of the GS in adult and subadult males. In all other subspecies the GS are either completely separated or have their black borders fused. Comparisons of S. undulatus consobrinus with S. undulatus speari are given in the description of the latter (Smith et al. 1995). The latter was not compared in detail with S. undulatus tristichus, however. Three diagnostically significant morphological differences and five individually non-diagnostic morphological differences exist (see Table 4). Size differs somewhat; in S. undulatus speari the maximum SVL in males is 64 mm, and 14% are 60 mm or more in SVL (in females, 69mm and 38%), whereas in S. undulatus tristichus those figures are 71mm and 22% for males, 69mm and 55% for females. Three diagnostically significant differences exist in pattern; S. undulatus speari has (1) distinct dorsolateral light lines (vs faint); (2) no black borders on its semeions (vs present); and (3) abdominal semeions faint or absent in 100% of females (vs distinct, in no less than 89%, but probably close to 100% excluding possible intergrades). Non-diagnostically but statistically significantly, no more than 35% of S. undulatus speari females have distinct gular semeions, whereas at least 95% of S. undulatus tristichus females do. SMITH, CHISZAR & LEMOS-ESPINAL 135 Table 3. Chi-square values and percent occurrences for statistically significant mor¬ phological differences between S. undulatus belli and S. undulatus speari Character- state N chi2 df P % 5 86(65) 32.84 1 <.001 31(78) 8 89(65) 20.54 1 <.001 79(43) 9 90(68) 10.98 1 <.001 90(69) The number of S. undulatus belli is followed parenthetically by the number of S. undulatus speari. Sequence the same in percent column, none of individual diagnostic significance. See text for character-states. Finally, although Applegarth (1969) synonymized S. undulatus tristichus with S. undulatus consobrinus, this study provides ample evidence that they are distinct taxa (allosubspecific). The only structural difference at a diagnostic level that was found is in tail /total length ratio, but statistically significant differences of independently non-diagnostic magnitude exist in at least six features of scalation (see Table 5). The other diagnostically significant differences lie in pattern: in S. undulatus consobrinus, (1) 29% of males have the gular semeions separated (vs 71% in S. undulatus tristichus)’, (2) in 100% of females, abdominal semeions are absent or faint (vs at most 11%, most of which are probably intergrades); and (3) all except intergrades have distinct dorsolateral light lines (vs indistinct lines). A distinct difference between these two taxa also exists in habits and habitat, S. undulatus consobrinus being cursorial and terrestrial, S. undulatus tristichus scansorial and saxicolous (personal observation). Future Research The intricate populational relationships of the subspecies currently recognized in S. undulatus are likely to require study for many years to come. In particular, the wide-ranging subspecies S. undulatus consobrinus and S. undulatus hyacinthinus merit further study of geographic variation. The taxonomic rank of the four exerges, particularly of belli and undulatus, need attention. The taxonomic status of the relic population near Limon, Colorado (Smith et al. 1993), remains enigmatic (although here viewed as relic intergrades between S. undulatus garmani and S, undulatus consobrinus), and the possible sympatry of several terminal populations is intriguing. Areas of intergradation are in need of refinement in almost every case. The plethora of subspecies in New Mexico (certainly six, and possibly seven 136 THE TEXAS JOURNAL OF SCIENCE— VOL. 47, NO. 2, 1995 Table 4. Chi-square values and percent occurrences for statistically significant mor¬ phological differences between S. undulatus speari and S. undulatus tristichus Character- state N chi2 df P % 7 65(286) 177.52 1 <.001 6(88)* 3 48(159) 94.11 1 <.001 85(13)* 18 63(276) 61.45 1 <.001 75(28)* 8 65(295) 42.06 1 <.001 43(82) 14 65(291) 39.44 1 <.001 63(23) 9 68(298) 12.92 1 <.001 69(45) 16 68(301) 12.8 1 <.001 9(30) 23 68(296) 11.16 1 <.001 51(70) The number of S. undulatus speari is followed parenthetically by the number of S. undulatus tristichus. Sequence the same in percent column. Diagnostically significant differences are indicated by an asterisk. See text for character-states. or eight if S. undulatus speari and/or S, undulatus belli occur there) is an open invitation for study. Applegarth’s review (1969) of northeastern populations in the state did not distinguish S. undulatus consobrinus and S. undulatus tristichus, but did not consider sexual dimorphisn in ventral coloration. A further problem exists in the sexual dimorphism in dorsal pattern in S. undulatus belli', this is a character unprecedented in the S. undulatus complex. Its phylogenetic origin and adaptive value are completely unknown. Also, S, undulatus speari presents a unique, presumably hormonal adaptation that permits a bleached dorsal pattern in summer to be replaced in the fall with a darker pattern that enables even adults, which in other species have gone into hibernation (or brumation), to remain as active as in midsummer (Smith et al. 1995). No intergradation of S. undulatus speari with other subspecies has been confirmed and needs investigation. Morrison’s (1988) study concludes with allospecificity of S. undulatus hyacinthinus on one hand, and S, undulatus garmani plus S. undulatus consobrinus on the other. On the contrary, McCoy (1961) found all three subspecies intergrading in limited zones in Oklahoma. Further study is needed for a definitive conclusion. Although the undulatus and tristichus exerges appear to be parallel lines of evolution, separated by the consobrinus exerge (Figure 1), there appears to be no reason to question the conspecificity S, undulatus tristichus and S. undulatus consobrinus, because of the occurrence of intergradation between them in central and southern New Mexico and SMITH, CHISZAR & LEMOS-ESPINAL 137 Table 5. Chi-square values and percent occurrences for statistically significant mor¬ phological differences between S. undulatus consobrinus and S. undulatus tristichus Character- state N chi2 df P % 1 101(159) 51.8 1 <.001 73(28)* 17 160(276) 46.37 1 <.001 72(38) 15 166(300) 36.35 1 <.001 31(10) 9 171(298) 17.33 1 <.001 65(45) 20 176(299) 14.4 1 <.001 23(10) 12 171(301) 11.65 1 <.001 18( 7) 16 180(301) 8.74 1 <.005 18(30) The number of S. undulatus consobrinus is followed parenthetically by the number of S. undulatus tristichus. Sequence the same in percent column. Diagnostically significant difference indicated by an asterisk. See text for character-states. southern Arizona. Intergradation is indicated by the shift in females from absence of abdominal semeions {S. undulatus consobrinus) to presence of them (S, undulatus tristichus), by the shift in both sexes from bright dorsolateral light lines (the former) to their reduction or loss (the latter), and by the shift from terrestrial to petricolous habitats, respectively. Attention is called to the intergrades from Bernalillo (15 mi E Albuquerque; 2-4 mi N Isleta Pueblo), San Miguel (2.3 mi W Sands) and Santa Fe (3 mi N Pena Blanca; 5-9 mi E Santa Fe; 9 mi S Santa Fe) counties, all referred to S, undulatus consobrinus although with very weak AS in females. No records indicate whether any of these specimens were in a terrestrial or petricolous habitat. A series of seven specimens from Water Canyon, 1 mi NE and 1 mi NW Manzano, Torrance Co., is assigned to S. undulatus tristichus, but inasmuch as only one of the four adult females has prominent AS, this series too could be regarded as intergradient. Habitat records would be of considerable value in population evaluation where intergradation may be involved, inasmuch as habitat is sharply different among central and southern subspecies of S. undulatus. Intergradation of S. undulatus belli with S. undulatus consobrinus certainly occurs in northwestern Chihuahua. A specimen (UTEP 3568) from 24 air mi NNE Ascencion is an adult male (61 mm SVL) with a spotted dorsum, distinct dorsolateral light lines and separate GS, as in S. undulatus consobrinus, but the black borders on the GS are scarcely visible to the naked eye, as in S. undulatus belli. Yet a 72 mm SVL male from 9.9 rd mi SW Ascencion (UTEP 4271) has the typical unicolor dorsum, dim dorsolateral light stripes and broadly fused GS, as 138 THE TEXAS JOURNAL OF SCIENCE-VOL. 47, NO. 2, 1995 Table 6. Data on the Four Major Southwestern Subspecies of Sceloporus undulatus for all Character-States of Tables 1-5 Character- State S. undulatus belli S. undulatus consobrinus S. undulatus speari S. undulatus tristichus 1 33(91) 101(73) 48(92) 159(28) 2 33(85) 150(61) 48(88) 159(18) 3 33(70) 150(41) 48(85) 159(13) 4 33(33) 150(-12) 48(65) 159( 1) 5 86(31) 241(27) 65(78) 286( 4) 6 86(66) 159(52) 65(97) 286(27) 7 86(55) 241(56) 65( 6) 286(88) 8 89(79) 164(82) 65(43) 295(82) 9 90(90) 171(65) 68(69) 298(45) 10 91(45) 172(59) 68(63) 300(62) 11 90(12) 170(28) 67(18) 298(19) 12 90(13) 171(18) 68( 4) 301 ( 7) 13 90(16) 168(22) 68(28) 298(19) 14 90(71) 166(48) 65(63) 291(23) 15 88(49) 166(31) 65(43) 300(10) 16 89( 7) 180(18) 68( 9) 301(30) 17 88(77) 160(72) 63(87) 276(38) 18 88(72) 160(59) 63(75) 276(28) 19 90(18) 176(19) 69( 0) 299( 8) 20 90(26) 176(23) 69(12) 299(10) 21 90(10) 171(23) 69(19) 277(22) 22 90(64) 170(36) 68(37) 290(19) 23 89(60) 254(74) 68(51) 296(70) The figures in each column give first the total number of specimens examined for the given character-state, followed parenthetically by the percent of that number having that character-state. See text for the numbered character-states. in S, undulatus belli, to which it is clearly referable (it also has the maximum SVL we have recorded for male S. undulatus consobrinus, but within the range - to 74 mm - of male S, undulatus belli). Although the locality for S. undulatus belli south of Ascencion is no more than 50 km from the New Mexico line, that subspecies does not occur in the latter state, to judge from the 83 specimens of the species in the UNM collection from various localities in Hidalgo Co. , New Mexico; all are clearly S. undulatus consobrinus. Intergradation of the two subspecies may also occur in in southeastern Chihuahua and southern Coahuila, where their ranges are presumably in broad contact. However, the distinction is sharp between the easternmost sample of S. undulatus belli we have seen (UTEP 4459, 6707-8, 6783-5, El Chiflon, 36 km W Saltillo, Hy 40, Coahuila) and the nearest sample of S. undulatus consobrinus (UNM 8940- 1 , 1 mi W and 1 mi S Villa de Garcia, Nuevo Leon). SMITH, CHISZAR & LEMOS-ESPINAL 139 Figure 8. Proposed phylogenetic concept of the eleven subspecies of S. undulatus currently recognized. The flared base of an arrow indicates apparent existence of intergradation between the connected taxa. No evidence of intergradation of S. undulatus speari with other taxa is known but may exist in areas of potential contact as yet not sampled. The names of the taxa and outlines of the ranges of the exerges are so displayed as to reflect approximate geographic position, but length of arrows is not of any significance phylogenetically. The symbol x indicates an isolated population of relictual intergrades between S. undulatus garmani and S, undulatus consobrinus (cf. Smith et al., 1993). Rationale Our continued preoccupation with subspecies is justified in part by the most fundamental purpose of taxonomy: to recognize and provide names for all populational, genetically distinctive "kinds" of organisms. Geographically consistent, genetically distinctive segments of species are also "kinds," just as real as species, and in some cases are more readily distinguished than some species, inasmuch as reproductive isolation, a necessary attribute of species, may be effected without discernible phenotypic effect. On the contrary, the absence of reproductive isolation, an attribute of subspecies, commonly results in readily observed phenotypic differences. 140 THE TEXAS JOURNAL OF SCIENCE— VOL. 47, NO. 2, 1995 Although the working diagnoses of some of the subspecies of S. undulatus have been based on the 70% level of differentiation (Simpson, 1961: 173-176; Mayr, 1969: 188-193), an admittedly arbitrary criterion, S. undulatus belli is categorically different from the other subspecies of the S. undulatus complex. Hence this is not dealing with a category of convenience (sensu Frost & Hillis 1992), but with a taxon that has an evolutionary history, potentially traceable by phylogenetic- cl adi Stic methods. Figure 8 represents our hypotheses regarding the outcome of such analyses when sufficient data become available to permit their application to the S. undulatus complex. Acknowledgments We are much indebted to the personnel of the various museums from which material has been loaned for our study: Dr Shi-Kuei Wu of UCM; Dr. Arnold Kluge and Greg Schneider of UMMZ; Dr. Howard L. Snell, Dr. William G. Degenhardt (also for zoogeographic counsel) and Allan Landwer, MSB; Ernest A Liner, private collection; and Dr. Robert G. Webb, UTEP (also for habitat information from his field notes on S, undulatus belli). Logistic support for field work and facilities for laboratory study were kindly provided by Drs. William M. Lewis, Michael C. Grant and Shi-Kuei Wu of the University of Colorado, and by Dr. Fermm Rivera Agiiero of CyMA-UNAM Iztacala; and our collecting permit A00702. -05929 by Dr. Ezequiel Ezcurra of the Instituto de Ecologia-SEDESOL. We are especially grateful for the indulgence by Joan, Adam and Laine Chiszar that has been vital for our work on this and related projects, in the field and in the laboratory. Literature Cited Applegarth, John S. 1969. The variation, distribution, and taxonomy of the Eastern fence lizard, Sceloporus undulatus Bose in Latreille, in northeastern New Mexico. Albuquerque, Univ. New Mexico. M.S. Thesis. 124 pp., 21 figs. Bell, Edwin L. 1954. A taxonomic and evolutionary study of the Western fence lizard, Sceloporus occidentalism and its relationships to the Eastern fence lizard, Sceloporus undulatus. Urbana, Univ. Illinois, Ph.D. Diss. v, 162 pp., 7 figs. Conant, Roger & Joseph T. Collins. 1991. A field guide to reptiles and amphibians: eastern and central North America. Boston, Houghton Mifflin, xx, 450 pp., 48 pis., 155 figs., 334 maps. Ferguson, George M. 1982. Distribution, variation and genetic SMITH, CHISZAR & LEMOS-ESPINAL 141 relationships of the lizard Sceloporus cautus Smith in northeastern Mexico. El Paso, Univ. Texas M.S. Diss. xiii, 195 pp., 30 figs. Hammerson, Geoffrey. A. 1982. Amphibians and reptiles in Colorado. Denver, Colorado Div. Wildlife, vii, 131 pp., ill. (col.). International Code of Zoological Nomenclature, 1985. H. Charlesworth & Co. Ltd, Huddersfield, England. 338 pp. Kirk, R. E. 1982. Experimental design. Pacific Grove, California, Brooks/Cole. 911 pp. Lemos-Espinal, Julio A., David Chiszar & Hobart M. Smith. 1994. Results, and their biological significance, of a fall herpetological survey of the transmontane sand dunes of northern Chihuahua, Mexico. Bull. Maryland Herp. Soc., 30(4): 157-176, Fig. 1. Maslin, T. Paul. 1964. Amphibians and reptiles of the Boulder area. Univ. Colorado Mus. Leaflet, (13): 75-80. Mayr, Ernst. 1969. Principles of systematic zoology. New York, McGraw-Hill, xiii, 428 pp., ill. _ & Peter D. Ashlock. 1991. Principles of systematic zoology. Second edition. New York, McGraw-Hill, xx, 475 pp., ill. McCoy, Clarence J., Jr. 1961. Distribution of the subspecies of Sceloporus undulatus (Reptilia:Iguanidae) in Oklahoma. SW Nat., 6(2): 79-85, map 1, figs. 1-2. Morrison, Randall L. 1988. A re-examination of the taxonomic status of Sceloporus undulatus sensu lato in Oklahoma and Texas. Lincoln, Univ. Nebraska, M.S. Diss. vi, 109 pp., 17 figs. Simpson, George G. 1961. Principles of animal taxonomy. New York, Columbia Univ. xiii, 247 pp., 30 figs. Smith, Hobart M. 1938. Remarks on the status of the subspecies of Sceloporus undulatus, with descriptions of new species and subspecies of the undulatus group. Occ. Pap. Mus. Zool. Univ. Michigan, (387): 1-17, map. _ , Edwin L. Bell, John S. Applegarth & David Chiszar. 1992. Adaptive convergence in the lizard super species Sceloporus undulatus. Bull. Maryland Herp. Soc., 28(4): 123-149, figs. 1-9. _ _ _ & David Chiszar. 1989. The subspecific identity of the population of Sceloporus undulatus sympatric with S. occidentalis . Bull. Maryland Herp. Soc., 25(4): 143-150, figs. 1-2. , _ , Emmett Evanoff & Jeffrey B. Mitton. 1993. The range of the so-called relictual intergrades between the lizards Sceloporus undulatus garmani and S. undulatus erythrocheilus . Bull. Maryland Herp. Soc., 29(1): 30-36, fig. 1. _ , Julio A. Lemos-Espinal & Edwin L. Bell. 1995. The 142 THE TEXAS JOURNAL OF SCIENCE— VOL. 47, NO. 2, 1995 Cabeza de Vaca Basin subspecies of the lizard Sceloporus undulatus. Trans. Kansas Acad. Sci. (in press). _ & Charles R. Nadvornik. 1990. Another record of occurrence of the Northern plateau lizard {Sceloporus undulatus elongatus) in southern central Wyoming. Northwestern Naturalist, 71: 11-12. Stebbins, Robert C. 1985. A field guide to western reptiles and amphibians. Boston, Houghton Mifflin, xvi, 336 pp., 48 pis., 202 maps, 69 figs. Tanner, Wilmer W. 1987. Lizards and turtles of western Chihuahua. Gr. Basin Nat., 47(3): 383-421, figs 1-8. Appendix A total of 801 specimens of Sceloporus undulatus was examined during this study. The 69 specimens of S. undulatus speari and all but 85 of the 255 S, undulatus consobrinus are listed in Smith et al. (1994), and are not repeated here; two exceptions (both UTEP) are from 24 air mi NNE Ascencion, Chihuahua (3568), and 30 mi S Monel ova, Coahuila (5626), and the remaining 83 (all MSB) are from various localities in Hidalgo County, New Mexico. The 156 S. undulatus belli are cited as types in preceding paragraphs, and the 302 S. undulatus tristichus (all UCM) are as follows. ARIZONA. Maricopa Co. : Sycamore Creek Campground, nr Sycamore (44030). Pinal Co.: Oak Flat Campground, 4000 ft, 4 mi E Superior (46054-5). Yavapai Co.: 1 mi NW Yarnell (13302, 13305-11); 1 mi W Yarnell (13303-4). COLORADO. Archuleta Co.: 2 mi N Chromo (7313, 7317, 7320); 2 miS Chromo (7311-2, 7314-6, 7318-9, 7321, 7323-33, 7335-6); 2 mi S, 0.5 mi W Chimney Rock Peak (51952-3, 52046); 0.5 mi SE Pagosa Springs (5995-6004); 5 mi E Pagosa Springs (52119). NEW MEXICO. Catron Co.: 2 mi W Aragon (6254); 2 mi E Beaverhead (6314-21); 1 mi E Beaverhead Ranger Station (6304-13); 3 mi NE Horse Springs (6255); 12 mi S Horse Springs, Bat Cave (6286-6303); 1.5 mi SE Horse Springs (6280-3); 1 mi W Horse Springs (2 specimens, unnumbered); 7 mi W Horse Springs (6256); 6 mi WSW Horse Springs (6257-68, 6270-1, 6273-9); nr Old Horse Springs (34140-6). Colfax Co. : 5 mi W Hoxie, Koehler mine (7054, 7062). Harding Co. : 19 mi E Mosquero (13844-7). Mora Co. : Hy 120, 3 mi W Canadian River (13848-60). Rio Arriba Co.: 5 mi E Gallina (7360-8); 2 mi N Regina (7356-9); 9 mi E Youngsville SMITH, CHISZAR & LEMOS-ESPINAL 143 (7377-84). Sandoval Co.: Bandolier National Monument Headquarters, 7500 ft (43951-3); 2 mi W Bernalillo (23449-55); 8 mi S Cuba, Hy 44 (10627); 4 mi SW Cuba, Hy 197, 6800 ft (43849-50, 43869-70); 15 mi NW San Ysidro, Hy 44 (10626); Rio Grande 3 mi N Pena Blanca (16820-37, 23219-22, 43896-900, 43905, 43907-9); Rio Grande 3,5 mi N Pena Blanca (43945-7). Sierra Co. : Rito de los Frijoles (73a-c); 18 mi W Winston (6322-36). Taos Co.: 3 mi W Arroyo Hondo, Rio Grande (7011-31); Hy 285, 16 mi S Colorado state line (14692-5); 3.3 mi E Ojo Caliente, 7700 ft (43948-50); Questa (6227); 2.5 mi E Taos (7032-49). Torrance Co.: Water Canyon, 1 mi NE Manzano (23456-7); Water Canyon, 1 mi NW Manzano (23458-62). UTAH (intergrades, S. undulatus tristichus x S. undulatus elongams; Smith & Chiszar, 1989). Washington Co.: 4. 2-7. 2 mi NW Leeds (56061); 25 mi NE St. George (13863-5). TEXAS J. SCI. 47(2): 144-150 MAY, 1995 SUMS OF PRODUCTS OF POSITIVE INTEGERS David R. Cecil Department of Mathematics , Texas A&M University-Kingsville, Kingsville, Texas 78363 Abstract.— Although formulas for, and numerical values of, sums of powers of positive integers appear in any number of handbooks and reference volumes devoted to mathematics, few relationships can be found involving sums of products of positive integers. Included in this paper are a number of formulas for sums of products of two and of three variables. Since sums of products occur frequently, in such diverse fields as combinatorial analysis, numerical approximation of double integrals and complexity of algorithms, these formulas should be useful in a number of applications. Two examples are given illustrating the usefulness of the formulas. Sums of powers and sums of products of positive integers occur in many combinatorial problems. Although formulas for (Beyer 1970; Selby 1974), and numerical values of (Abramowitz & Segun 1968), sums of powers of positive integers appear in any number of handbooks and reference volumes devoted to mathematics, few relationships can be found involving sums of products of positive integers. Certain of these sums can easily be obtained from the well known formulas for sums of powers of positive integers. For example, to obtain the sum of products of two consecutive positive integers, one has N-1 (l*2)+(2*3)+ - ■f((V-l)*iV) = Y. i=l N-1 N-1 = E * E '■ = . i=l i=l ^ As another example, consider the sum of products of two consecutive odd positive integers, followed by two even ones, then two odds, and so on, yielding the following: N-2 (l*3)+(2*4)+(3*5)+ +((V-2)*A) = Y i=l N-2 i=i N-2 2Yi- 1=1 iN-2) *(iV-l)*(2V+3) 6 CECIL 145 Sums of Products of Two Variables This paper presents formulas for certain sums of products of two or of three positive integer variables. Theorem 1 ^ z= A[( ^ 2 lN * * E' = "at + ~*IN^+3N^->-3NKN] . 1=1 2 As in Theorem 3 , assume that a^ is a fifth degree polynomial in N and solve the resulting difference equation for the unknown coefficients. Theorem 4 Proof: Add the results of Theorems 2 and 3. Theorem 5 ^ N j-1 N-1 N j=2 1=1 1=1 j=i+l Proof: Letting Sn be the value of the summation, by induction it is seen that the equation for recursion, at stage M, is given by: M Vl - * E'" ^2 = . i=l Iterating for S3, and so on, produces the first of the two double summations. The second double summation is obtained from expanding the first one and then regrouping like terms in p. A Product of Three Variables Theorem 6 Proof: Note that (1.2. .iv)’ = - ^£v*^‘** • 148 THE TEXAS JOURNAL OF SCIENCE-VOL. 47, NO. 2, 1995 The conclusion follows by using the results of Theorems 2 and 3 along with the formulas for the sums of i and of ? ( these formulas appeared in the proofs of Corollaries 2 and 5 of Theorem 1). Linear Combinations of Two and Three Variables Let a and b be any two fixed real numbers. Theorem 7 Proof: Letting denote the summation of the ai + bj terms, the recursion relationship is given by: N-1 - Vi = (E ^ Nb*iN-l) i=l Assuming that the solution for S^, has the form Ci*N + C2*N^ + C3*N^ and using the method of undetermined coefficients, the result follows. Corollary N*(N+1) 6 *[(a^2b)^(N-l) + 3(fl+^>)] Proof: Add N N + b)i = (a + b)*Y,i i=l i=l to the result of the theorem. N-1 N-1 N-1 N=2 N=3 N-1 ^ Lemma CECIL 149 Proof: By expressing iV-1 N-l N-l + j: ,• + .. - E = N=2 N=3 N-l (iV-1) +... + 5+ 4+ 3+ 2 + (iV-1) + ... + 5+ 4+ 3 + (A^-1) + - + 5 + 4 + (N-l) = (iV-2)*(A^-l) + + 5*(4) + 4*(3) + 3*(2) + 2*(1) it is clear that this is the same as (except for one term less) the first relationship given in the introduction, thus giving the desired result. Theorem 8 = ^♦(a+26+3c)*(JV-2)*(Ar-l)t(JV+l) the recursion relationship is given by: iV-1 N-l N-l N=2 N=^ N-l a[l(N-2) + 2iN-3) + + (A^-2)(l)] + Ncl(N-2) + (iV-3) + + 1] which by the lemma becomes N-2 N-2 ^[— *(^-l)*(A^-2)] + i*(N-i-l) + cN'^ i ^ i=l i=l and this simplifies down to ia^lb^c) ^jv*(Af-l)*(Ar-2) . 6 Assuming that = ( Cj + C2 Nf + C3 N + C4 ) * N, substituting into the recursion equation and solving for the Cj terms gives the desired result. 150 THE TEXAS JOURNAL OF SCIENCE-VOL. 47, NO. 2, 1995 Conclusions Two examples illustrating the usefulness of the formulas are now presented, one from combinatorics and the other from the field of numerical approximations. Given a fixed positive integer N greater than 1; on each day i, from day 1 through day N- 1 , choose 1 man from a group of i men and choose 1 woman from a group of at least i+1, but no more than N, women. Then X;ij, where 1 < i < j < N, gives the total number of possible selections made during the N-1 days. This value can then be found using the formula given in Corollary 2 to Theorem 1 . A method for approximating the value of the double integral fo fo is to divide [0,1], along both the x- and the y-axes, into subintervals of length Ax = 1/N = Ay. Then, using the (i/N,j/N) endpoint in each square area [(i-l)/N,i/N] by [(j'l)/N,j/N] to calculate f(x,y) = x^ y. (Ax)(Ay) E l\xv'mg winter months it was noted to be present in relatively high numbers in the low shrub areas of the reserve. Aratinga acuticaudata (Vieillot) (Blue-crowned Parakeet) Occurrence .—Th\s> species is found within the reserve from spring to fall; it was not observed to be present during the winter months. It nests within the boundaries of the reserve. Habitat.— observed transitory in other areas of the reserve, it exhibits a preference for the uppermost strata of the upper forest areas. Aratinga leucophtalmus (Muller) (White-eyed Parakeet) Occurrence .—This species is most commonly observed during the spring and summer months. It is not present during the winter and is only transitory during the autumn. This species does nest within the boundaries of the reserve. Habitat .—^h\\t this species may occur from the low shrub to the upper forest areas, it is most often observed in the upper forest. Forpus xanthopterygius (Spix) (Blue-winged Parrotlet) Occurrence .—T>\xv'mg the course of the two year study, this species was observed on only two occasions during summer months. The reserve borders upon the southwestern limit of the range of this species. It is considered rare at El Bagual. Habitat. — When observed, it was in the upper forest. 178 THE TEXAS JOURNAL OF SCIENCE— VOL. 47, NO. 3, 1995 Acknowledgements Special thanks go to J. G. Fernandez, A. Balabusic, F. Moschione, J. R. Dixon and K. A. Arnold. Alparamis S. A. of Buenos Aires provided logistic and financial support for this project. This is Scientific Contribution Number 37 for El Bagual Ecological Reserve. Literature Cited Balabusic, A. M., R. A. Banchs, & F. N. Moschione. 1990. Proyecto Amazona aestiva. Direccion Nacional de Fauna Silvestre, Buenos Aires, 9 pp. Contreras, J. R. 1987. Lista preliminar de la avifauna de la provincia de Formosa, Repilblica Argentina. Hist. Nat. 7:33-52. Dabbene, R. 1935. Los loros deben ser considerados plaga nacional? Hornero 5:39-59. Emlen, J. T. 1971. Population densities of birds derived from transect counts. Auk 88:323-342. Morello, J. & J. Adamoli. 1974. La vegetacion de la Republica Argentina. Las grandes unidades de vegetacion y ambiente del Chaco Argentino. Segunda parte: vegetacion y ambiente de la provincia del Chaco. INTA Serie Fitogeografica Number 18. 130 pp. Narosky, T. & D. Izurieta. 1987, Guia para la identificacion de las aves de Argentina y Uruguay. Asoc. Ornitol. del Plata., Buenos Aires. 345 pp. Perez, J. J. & L. E. Eguiarte. 1989. Situacion actual de tres especies del genero Amazona {A, ochrocephala, A. viridigenalis y A, autumnalis) en el noreste de Mexico. Fida Silvestre Neotropical 2:63-67. Waterton, C. 1825. Wanderings in South America. Reprint 1983, Century Publ., London. 520 pp. TEXAS J. SCI. 47(3): 179-190 AUGUST, 1995 HABITAT USE BY WINTERING SHOREBIRDS ALONG THE LOWER LAGUNA MADRE OF SOUTH TEXAS Timothy Brush Department of Biology, The University of Texas - Pan American, Edinburg, Texas 78539 Abstract.— Censuses of wintering shorebirds were conducted in mudflat habitats along the lower Laguna Madre during November-February of 1992-1993 and 1993-1994. Widespread species such as Western Sandpipers (Calidris mauri) , Willet {Catoptrophorus semipalmatus). Black-bellied Plovers {Pluvialis squatarola), and dowitchers {Limnodromus spp.) used both high and low mudflats, peaking during low water conditions. Long-billed Curlews (Numenius americanus) and Marbled Godwits (Limosa fedoa) used mainly low mudflats, while American Avocets (Recurvirostra americana) and Stilt Sandpipers {Calidris himantopus) were restricted to pools on the high mudflats. Piping Plovers {Charadrius melodus). Snowy Plovers {Charadrius alexandrinus) and other small Charadrius plovers were uncommon, occurring mainly on high mudflats and algal flats before these were covered with glasswort {Salicornia bigelovii). The lower Laguna Madre, which is located in subtropical southern Texas, is characterized by extensive unvegetated mudflats and sandflats near mean sea level as well as large areas covered by seagrasses below mean sea level (Britton & Morton 1989). Mild winters and extensive intertidal flats support the existence of many shorebirds in the Laguna Madre region of both Texas and Tamaulipas (Mitchell & Boyd 1992; Morrison et al. 1993; Withers & Chapman 1993). Some of the largest known wintering populations of the threatened Piping Plover and the declining Snowy Plover also occur in the Laguna Madre of south Texas (Haig & Plissner 1993). This study was undertaken to determine patterns of habitat use by shorebirds and their responses to changing habitat conditions during winter months along the lower Laguna Madre of south Texas. Study Areas All study areas were located along the western shore of the lower Laguna Madre, Cameron Co., Texas, 26°04'>26°22'N and 97° 10'- 97°20' W. Habitats were classified based on frequency of exposure and presence of pools, algal flats, or seagrass. Large areas of intertidal habitat are alternately exposed and covered by wind shifts associated with fronts during the winter (Mitchell 1992), whereas the effects of astronomic tides are restricted to passes (Breuer 1962). 180 THE TEXAS JOURNAL OF SCIENCE-VOL. 47, NO. 3, 1995 High Mudflat/Pool. — This area is comprised of a large complex of enclosed shallow water and exposed, firm mudflat just west of Horse Island in Unit 5, Laguna Atascosa National Wildlife Refuge (LANWR). This is the only plot to have a relatively permanent, shallow pool, which covered the deepest portion of the study area. Strong south and southeast winds exposed extensive mudflats > 2000m wide, whereas north and northwest winds exposed more limited areas of mudflats. Blue-green algae formed isolated mats over the mud in areas periodically exposed, but such algal mats were poorly developed and temporary. Although nearly devoid of vegetation in November 1992, some of the mudflats within 50 m of the usual water’s edge became partially covered with glass wort {Salicomia bigelovii) during December-February 1992- 1993. By November 1993, glasswort covered large areas of secondary coves and other regularly- inundated sites throughout the study area. High Algal Flat.— Th\s area is located immediately south and east of Horse Island and is characterized by wind-tidal flats covered with a well- developed algal mat. The algal mat provided a firm, tough surface in much of this study area. High Algal Flat was covered with water during strong southeast winds or seasonal high water, and was exposed by strong north and northwest winds. Limited Salicomia covered this flat by February 1993, but coverage never became dense in this area. There was no increase in glasswort density during the second winter, but dead glasswort stalks remained on the flats during the second winter. New germination was not noticed until late February 1994. Low Mudflat.— This area (the shoreline section of Bayside Tour Loop, Unit 7, LANWR) includes the shoreline and associated mudflat and seagrass areas visible from the shore. These mudflats are frequently covered with water and are exposed mainly after cold fronts with associated north and northwest winds (Mitchell & Boyd 1992). Seagrasses (primarily Halodule beaudettei, with some Thalassia testudinum, Syringodium flliforme and Halophila engelmannii) covered a limited area exposed offshore during the period of lowest water levels. Low Seagrass Flat.— This area includes the intertidal area between Port Isabel and Laguna Heights, north of Texas Route 100. Turtle-grass {Thalassia testudinum) covered about half of these low mudflats, which were inundated and exposed on a daily basis by tides throughout most of the winter months. BRUSH 181 Methods Each area was censused every two weeks, as accessibility allowed. Usually, all areas were censused during the same day, although occasionally two days were needed. During some rainy periods, not all areas were censused, due to road conditions. Each study area was censused 15-17 times. The distance to the water’s edge was estimated at reference points during each census. Species were identified based on plumage, silhouette, behavior, and calls (Hayman et al. 1986). Because of frequent identification difficulties, only 4% of all dowitchers were identified to the species level. Therefore, Short-billed and Long-billed Dowitchers were treated as "dowitchers" in this paper. At times, mainly on the low mudflat plot, small sandpipers were lumped as Calidris spp. ("peeps") due to distance. This category does not include the Stilt Sandpiper (C. himantopus) , a much larger species. 10 X 40 binoculars were used to scan for shorebirds, and a 15-45 X zoom telescope was used to identify most individuals. During each census, microhabitat use was recorded for each species, as shallow water only, shallow water and exposed mud, or exposed mud only. Results General patterns.— K total of 23 shorebird species (Table 1) were observed on the study plots, with 19 species recorded in numbers averaging at least one individual per census on at least one study plot (Table 1). The highest number of shorebirds present at a single time on a study area was 20,978 birds on High Mudflat/Pool, on 11 January 1994. That plot averaged the highest total numbers of shorebirds, and contained the largest single species total: 16,930 Western Sandpipers, also on 11 January 1994. Eleven species were recorded on all four plots, but three species were restricted to <2 study plots (American Avocet, Marbled Godwit, and Stilt Sandpiper). Peak total numbers of shorebirds occurred at different times on each of the four plots (Fig. 1). High Mudflat/Pool. — The largest numbers of many shorebird species were recorded here. Several species, including American Avocet (peak 1200 on 23 January 1993) and Stilt Sandpiper (peak 480 on 17 December 1992), were essentially restricted to this habitat. Total monthly shorebird numbers peaked in December 1993 and January 1994 (Figure 1), and Western Sandpipers averaged 48% of the shorebird community. Total shorebird numbers peaked during low water (Fig. 2). Least Sandpipers reached peak abundance here (Table 1) and foraged 182 THE TEXAS JOURNAL OF SCIENCE— VOL. 47, NO. 3, 1995 Table 1 . Mean and maximum (in parenthesis) numbers of shorebirds for species averaging at least 1 /census on at least one study plot Lower Laguna Madre, November-February 1992-1993 and 1993-1994. Asterisk (*) indicates highest mean or maximum for a given species. Habitat Type Avian Species High Mudflat Low Mudflat Algal Mudflat Seagrass Flat Black-bellied Plover, (Pluvialis squatarola) 77.4 (320) *246.9 (1450) 1.1 (5) 17.5 (75) Snowy Plover *17.9 (Charadrius alexandrinus) (51) 1.2 (21) 8.0 (36) 0.0 (0) Semipalmated Plover 13.1 (Charadrius semipalmatus) (91) 2.5 (23) *14.2 (95) 0.6 (3) Piping Plover (Charadrius melodus) 4.9 (24) 2.1 (13) *5.1 (69) 0.1 (1) Killdeer (Charadrius vociferus) *5.8 (23) 0.9 (5) 0.8 (4) 0.4 (6) American Avocet *412.9 (Recurvirostra americand) (1200) 0.7 (12) 0.0 (0) 0.0 (0) Greater Yellowlegs (Tringa melanoleuca) *52.4 (198) 16.2 (110) 3.6 (12) 1.7 (10) Lesser Yellowlegs (Tringa flavipes) *69.8 (360) 22.9 (270) 3.2 (15) 0.2 (3) Willet 127.7 (Catoptrophorus semipalmatus) (590) *261.5 (1100) 1.5 (6) 99,7 (400) Long-billed Curlew (Numenius americanus) 0,1 (1) *71.2 (205) 0.0 (0) 4.8 (25) Marbled Godwit (Limosa fedoa) 0.0 (0) *37.7 (55) 0.0 (0) 0.0 (0) Ruddy Turnstone (Arenaria interpres) 0.1 (2) *3.8 (12) 0.3 (4) 0.9 (6) Sanderling (Calidris alba) *1.3 (7) 1.0 (8) 0.5 (6) 0.0 (0) Western Sandpiper (Calidris mauri) *2259. 1 (16930) 390.9 (4370) 36.4 (208) 58.5 (800) Least Sandpiper (Calidris minutilla) *281.7 (1320) 4.1 (40) 61.0 (500) 0.0 (0) Peep spp. (Calidris spp. ) 330.0 (2000) *1156.5 (7700) 33.3 (500) 10.0 (150) Dunlin (Calidris alpina) *367.3 (1960) 85.9 (640) 3.5 (30) 11,9 (124) Stilt Sandpiper (Calidris himantopus) *126.0 (480) 0.0 (1) 2.4 (35) 0.0 (0) Dowitcher spp. (Limnodromus spp. ) ^ 556.0 (2000) *660.4 (2290) 191.6 (1340) 69.7 (600) Total ^ (Maximum '*) *4704.8 (20978) 2966.5(11980) 366.6 (1932) 276.8 (1528) ' Other species, detected in lower numbers, include Wilson’s Plover (Charadrius wilsonia), American Oystercatcher {Haematopus palliatus), Red Knot (Calidris canutus), and Semipalmated Sandpiper (Calidris pusilla). ^ Of the 4% of all dowitchers identified by call, 77 % were Long-billed Dowitchers, Limnodromus scolopaceus, while 23 % were Short-billed Dowitcher, Limnodromus griseus. ^ Average number of shorebirds, all species combined, per census. ^ Maximum number of shorebirds, all species combined, present during the same census. BRUSH 183 Table 2. Winter foraging locations of shorebirds along the Lower Laguna Madre. Species in shallow water and exposed mudflat categories were observed foraging in those habitats on >95% of observations. Asterisk (*) indicates species that were exclusively foraging in shallow water or exposed mudflat. Shallow water Exposed mudflat Water/mudflat American Avocet * Greater Yellowlegs Lesser Yellowlegs Stilt Sandpiper * Black-bellied Plover Snowy Plover * Semipalmated Plover Piping Plover * Sanderling Killdeer * Ruddy Turnstone * Least Sandpiper Willet Long-billed Curlew Marbled Godwit Western Sandpiper Dunlin Dowitchers both on wet and dry mudflats and among Salicomia. Species in the water- foraging group regularly occurred in large numbers here (Table 2). During each winter, such pool-users peaked in December and January, but monthly averages fluctuated relatively little, compared to numbers of the shorebird community as a whole (Figures 1 & 3). Large pool-using species such as American Avocets often foraged in or near ardeids, such as Reddish Egret (Egretta rufescens) and Snowy Egret {Egretta thula), which often occurred in large numbers. Small plovers, including Piping Plover, Snowy Plover, and Semipalmated Plover, occurred mainly on High Mudflat/Pool and on High Algal Mudflat. They were observed in small, loose flocks away from or at the edge of larger groups of other mudflat foragers. Plover numbers varied considerably, with lowest numbers in January 1993 and February 1994 (Figure 4). Piping Plovers averaged > 10/census only in November 1992 and December 1993. Snowy Plovers peaked during November-December of each winter, but they remained fairly common until February 1994. In contrast, Semipalmated Plovers were uncommon during most of the first winter but became fairly common during December 1993 and January 1994. Snowy Plovers foraged both near and far from the water’s edge, while Piping and Semipalmated Plovers foraged mainly on the wet flats near the water’s edge (below the glass wort zone, once it became established). High Algal Mudflat.— This area supported lower numbers of shorebirds, and was used sporadically (Figure 1). Peak numbers occurred during intermediate water levels, with more limited use during peak low or high water (Figure 2). Least Sandpipers were the most common Calidris sandpiper identified here and often foraged near small 184 THE TEXAS JOURNAL OF SCIENCE— VOL. 47, NO. 3, 1995 HIGH MUDFLAT/POOL HIGH ALGAL FLAT MONTH MONTH LOW MUDFLAT LOW SEAGRASS FLAT MONTH MONTH Figure 1 . Monthly mean number of shorebirds wintering on study areas along the Lower Laguna Madre. See text for description of habitats. Please note that vertical scales differ in all graphs. plovers. Flocks of dowitchers occurred erratically. Mixed flocks of small plovers consistently occurred here during only November- December 1992. Peak plover numbers were 95 Semipalmated Plovers on 27 November 1992, and 69 Piping Plovers on 14 November 1992. However, small plovers were nearly absent during the second winter and were never seen foraging among the glasswort. Low Mudflat. — This area, along Bayside Drive, was used heavily by shorebirds during extreme low water conditions in November 1993 and January 1994 (Figures 1 & 2). Peak numbers for the following wide¬ spread shorebirds were recorded on this plot: Calidris spp.: 7700 on BRUSH 185 HIGH MUDFLAT/POOL HIGH ALGAL FLAT LOW MUDFLAT LOW SEAGRASS FLAT 100000 10000 1000 10000 1000 100 100 100 DISTANCE TO WATER (M) 10 100 DISTANCE TO WATER (M) Figure 2. Relationship of water level and shorebird numbers on study areas along the Lower Laguna Madre. Both axes are scaled logarithmically (base 10). Please note different vertical scales on different graphs. 5 January 1994; dowitchers: 2290 on 5 January 1994; Black-bellied Plover: 1450 birds, on 27 November 1993; and Willet: 1100 on 11 January 1994. Western Sandpipers were common to abundant, while Dunlin were occasionally common. Least Sandpipers were uncommon (max. 40 individuals). Small plover flocks were limited to extreme low water conditions, during which they foraged on moist offshore bars with Calidris sandpipers and Black-bellied Plovers. The largest shorebird species. Long-billed Curlews and Marbled Godwits were essentially or entirely restricted to Low Mudflat. Long- 186 THE TEXAS JOURNAL OF SCIENCE-VOL. 47, NO. 3, 1995 1500 □ STILT □ WILLET □ LESS. YELL. □ GR YELL. □ AVOCET MONTH Figure 3, Monthly mean numbers of pool-using birds wintering on High Mudflat/Pool study area. Species are Stilt Sandpiper, Willet, Lesser Yellowlegs, Greater Yellowlegs and American Avocet. billed Curlews peaked at 205 on 26 November 1993, while Marbled Godwits peaked at 150 on 17 December 1992. Both these species were often seen in relatively deep water, near Reddish Egrets, Great Egrets (Casmerodius albus) and White Ibis (Eudocimus albus). Curlews foraged in loose flocks of 10-50 or in small groups of 2-3 birds, and were often observed flying between the intertidal flats and inland areas of LANWR. They foraged by probing into the exposed mud and into mud or seagrass below shallow water. Godwits usually occurred in more compact, larger flocks, when present. They foraged on exposed wet mud or in shallow water, and were never observed flying inland. Low Seagrass This long, linear study area was utilized by a limited number of shorebirds, mainly during low water (Figures 1 & 2). Only Willets occurred here in large numbers (peak 400 on 16 December 1993), but dowitchers. Black-bellied Plovers, and Western Sandpipers occurred in modest numbers during low water. Willets foraged frequently in turtle-grass when its tops were exposed, along with some Long-billed Curlews, Reddish Egrets and Tricolored Herons {Egretta tricolor). The other shorebirds foraged on the bare, moist mudflats exposed during lowest water conditions. Peak monthly averages of all shorebirds exceeded 1500 only during January 1994, while shorebirds were absent during the two November 1993 censuses (Figure 1). BRUSH 187 HIGH ALGAL FLAT HIGH MUDFLAT/POOL MONTH MONTH Figure 4. Monthly mean numbers of small Charadrius plovers wintering on high mudflat study areas along the Lower Laguna Madre. Species are Semipalmated Plover, Snowy Plover and Piping Plover. Please note the difference in vertical scales between the two graphs. Discussion and Conclusions Many shorebirds of several species wintered commonly on the mud¬ flats of the lower Laguna Madre. Although densities could not be determined in an unbiased fashion, specific patterns of habitat use were primarily based on the foraging requirements of each species. Pool¬ using birds preferred the large enclosed pool on one study area (High Mudflat/Pool) and were relatively consistent in numbers compared to mudflat foragers. The substantial wintering Stilt Sandpiper population on High Mudflat/Pool during the first winter was unexpected, given its winter status in the United States as irregularly uncommon (Oberholser & Kincaid 1974) or casual (AOU 1983). The large, enclosed pool which the Stilt Sandpipers used is typical Stilt Sandpiper habitat (Hayman et al. 1986) but probably represents a rare habitat in the Lower Laguna Madre. Mudflat and mudflat/pool foragers, dominated by Western Sandpipers and other Calidris spp. , fluctuated in response to changing water levels in all study areas. Highest numbers were recorded when extensive flats in the High Mudflat/ Pool and Low Mudflat study areas were newly exposed and still moist. When mudflats dried out due to prolonged exposure, such as in High Algal Mudflat, shorebird use declined. Since ideal conditions are partially dependent on wind shifts due to passage of periodic fronts, such use by shorebirds appears to be very opportunistic. 188 THE TEXAS JOURNAL OF SCIENCE-VOL. 47, NO. 3, 1995 Relative abundances of particular species in tlie Low Mudflat area were generally similar to those found by Mitchell & Boyd (1992) in the same area (Bayside Drive). Assuming that most of the Calidris spp. sandpipers in Low Mudflat were Western Sandpipers, then only Black- bellied Plovers, both yellowlegs, and Marbled Godwits were dispropor¬ tionately more common during my censuses than Mitchell & Boyd’s (1992) in Low Mudflat habitat. The peak of >1400 Black-bellied Plovers appears high, as it represents more than three times the total number seen along the Laguna Madre of Tamaulipas (Morrison et al. 1993). Marbled Godwits, uncommon during this study, may winter in large flocks in the Laguna Madre of Tamaulipas , a similar hypersaline lagoon in northeastern Mexico (Morrison et al. 1993). The nearness of coastal saline prairie to extensive intertidal feeding areas may explain the higher numbers of Long-billed Curlews on Low Mudflat than in my other areas or in Oso Bay (Withers & Chapman 1993). Snowy Plovers and Dunlin were somewhat less common during my censuses than in Mitchell and Boyd (1992), particularly given the generally higher numbers of shorebirds found in this study. The total absence of Black¬ necked Stilts during this study is inexplicable, given their winter occurrence in the Corpus Christi area (Withers & Chapman 1993). The modest peak numbers of small Charadrius plovers were expected, but the failure of many small plovers to winter in any plots was somewhat unexpected. Open algal flats normally support Snowy and Piping Plovers and Least Sandpipers in the Laguna Madre, but the spread of glasswort onto both algal flats and high mudflats may have made those habitats unsuitable for small plovers. This would be particularly true for Snowy Plovers and Piping Plovers, which forage on somewhat higher flats - the area covered by Salicomia. None of the three small plovers was recorded foraging in Salicomia-cowered areas, even in very open stands. Relatively large numbers of plovers only occurred when damp or wet flats were available above or below the Salicomia-cowcT&d areas. Shorebird use of the lower seagrass meadows and lower, bay-margin mudflats by large numbers of Calidris sandpipers, Willets, and Black- bellied Plovers was linked with cold fronts (northers) and seasonal low water in winter (Breuer 1962). Such low mudflats probably supply abundant food resources for shorebirds when they are exposed. These flats remain moist during short-duration exposures during the cool temperatures of winter. In contrast, the frequently exposed higher mudflats often dry out and are apparently unsuitable for most shorebirds BRUSH 189 in the absence of rainfall (Skagen & Knopf 1993). Shorebirds have seldom been observed previously foraging in seagrass, except in intertidal areas (Dann 1987). There is variability in all wetland/mudflat ecosystems, but the climatic variability, large size, shallow water, and flatness of the lower Laguna Madre ecosystem appear to make it particularly variable. Areas heavily used one day, month, or season may be little used at another time due to slight changes in water level. During times when neither the High Mudflat/Pool area nor the Low Mudflat area would be suitable for large numbers of mudflat- foraging shorebirds, other nearby areas may be suitable. For example, on the 1993 Laguna Atascosa Christmas Bird Count, > 20,000 shorebirds (mainly Western Sandpipers and dowitchers) were recorded on the Buena Vista Ranch, located adjacent to LANWR. Mitchell (1992) noted similar shifts in Redheads (Ay thy a americana) in the Lower Laguna Madre, due to changing water levels. Evidence to date suggests that (1) extensive areas along an elevational gradient are necessary to support large numbers of wintering shorebirds, and (2) that high mudflats free of Salicomia are needed to support wintering flocks of small plovers, including the declining Snowy and Piping Plovers. Ackno wl edgments Special thanks go to Steve Thompson and the staff of the Laguna Atascosa National Wildlife Refuge for their crucial logistic support during this study and to the Faculty Research Council of the University of Texas-Pan American for funding this study. Thanks also go to Curt Zonick for information on Piping and Snowy Plovers in South Texas. Appreciation is extended to Chris Onuf for providing information relative to seagrass cover along Bay side Drive, and to Brian R. Chapman and Terry C. Maxwell for their constructive comments on an earlier draft. Literature Cited American Ornithologists’ Union. 1983. Check-list of North American Birds. 6th ed. American Ornithologist’s Union, Lawrence, Kansas. 877 pp. Breuer, J. P. 1962. An ecological survey of the Lower Laguna Madre of Texas. Publ. Inst. Mar. Sci. 8:153-183. Britton, J. C. & B. Morton. 1989. Shore ecology of the Gulf of 190 THE TEXAS JOURNAL OF SCIENCE-VOL. 47, NO. 3, 1995 Mexico. Univ. Texas Press, Austin. 387 pp. Dann, P. 1987. The feeding behaviour and ecology of shorebirds. pp. 10-20 in Lane, B. A. Shorebirds in Australia. Nelson, Melbourne, Australia. 187 pp. Haig, S. M. & J. H. Plissner. 1993. Distribution and abundance of Piping Plovers: Results and implications of the 1991 international census. The Condor 95: 145-156. Hayman, P., J. Marchant & T. Prater. 1986. Shorebirds. An identifica-tion guide to the waders of the world. Houghton Mifflin, Boston. 412 pp. Mitchell, C. A. 1992. Water depth predicts Redhead distribution in the Lower Laguna Madre, Texas. Wildl. Soc. Bull. 20:420-424. Mitchell, K. & R. L. Boyd. 1992. Saltwater lagoon [Winter bird population study]. J. Field Ornithol. 63:24-25, Suppl. Morrison, R. 1. G., R. K. Ross, J. Guzman P. & A. Estrada. 1993. Aerial surveys of Nearctic shorebirds wintering in Mexico: preliminary results of surveys on the Gulf of Mexico and Caribbean coasts. Canad. Wildl. Serv. Progr. Notes 206:1-14. Oberholser, H. C. & E. B. Kincaid, Jr. 1974. The bird life of Texas, Volume 1. University of Texas Press, Austin. 530 pp. Skagen, S. K. & F. L. Knopf. 1993. Towards conservation of midcontinental shorebird migrations. Conserv. Biol. 7:533-541. Withers, K. & B. R. Chapman. 1993. Seasonal abundance and habitat use of shorebirds on an Oso Bay mudflat. Corpus Christi, Texas. J. Field Ornithol. 64:383-392. TEXAS J. SCI. 47(3): 191-202 AUGUST, 1995 LATE QUATERNARY SEDIMENTATION, LOWER NUECES RIVER, SOUTH TEXAS Frank G. Cornish and Jon A. Baskin Suemaur Exploration, Inc., Texas Commerce Plaza, Corpus Christi, Texas 78470 and Department of Geosciences, Texas A&I University, Kingsville, Texas 78363 Abstract.— Three valley fill and four alluvial terrace units are recognized from Holocene and late Pleistocene sediments in the lower Nueces River valley where this river is entrenched in the late Pleistocene Beaumont Formation. The three valley fill units are included in the Cayamon Creek Alloformation and are designated (from youngest to oldest) Cayamon Creek allomembers 3, 2, and 1. Allomember 3 consists of late Holocene sands. This is unconformably underlain by the muddy sand of allomember 2. Gastropod shells from this unit were radiocarbon dated at 965 ± 95 YBP. The unconformably underlying allomember 1 consists mainly of coarse sands and gravels. A log buried in this unit has been dated at 13,230 ±110 YBP. Four late Pleistocene terraces occur at elevations of (1) 0-5.7 m (0-19 ft), (2) 8.4-9 m (28-30 ft), (3) 10.8-13.5 m (36-45 ft), and (4) 14.4-16.5 m (48-55 ft) above the present day flood plain. The four terraces (from youngest to oldest) are the Angelita, Fort Lipantitlan, Bluntzer, and Corpus Christi. River valleys and their alluvial sediments of the lower Gulf Coastal Plain are the products of changing climate and sea level during the Quaternary. The best studied late Pleistocene and Holocene river deposits of the Texas Coastal Plain are associated with the Colorado River (for summary, see Blum & Valastro 1994). The Nueces River drains an area of 43,900 km^, a little more than 40% of the 103,000 km^ drained by the Colorado River of Texas (US Bureau of Reclamation 1983). The present study encompasses the lower Nueces River valley, from the mouth of the river in Nueces Bay to Wesley Seale Dam, near Mathis, Texas. The lower Nueces River valley and its terrace and valley fill deposits are a testament to the effects of changing sea level on the lower Texas coast. At the western edge of the city of Corpus Christi, where IH 37 crosses the Nueces River, the valley floor is more than 25 m below the top of the valley wall of the Beaumont Formation and there is at least another 10 m of latest Pleistocene valley fill below that. This approximately 35 m of erosion is quite remarkable in coastal South Texas, where the land surface slopes gently gulfward at a rate of 0.5 m per km (2.5 ft per mile) (Weeks 1945). This valley contains a complex variety of sedimentary fill. 192 THE TEXAS JOURNAL OF SCIENCE-VOL. 47, NO. 3, 1995 Nueces River Terrace Stratigraphy Terraces are geomorphic features that are recognized on the basis of landscape position. Deussen (1924) was the first to map the terraces of the Nueces River. In the vicinity of Calallen (south of Odem), he recognized two terrace levels. Both are present in the region between San Patricio/Bluntzer and Odem. Price (1933) described these two terraces in greater detail. He named the upper unit the Corpus Christi Terrace and designated its type area as occurring approximately 1.6 km southeast of Odem, at an elevation of 15 m (50 ft). He showed only this upper terrace continuing to the east of Odem to Corpus Christi Bay. This terrace is shown (Price 1933: Fig. 17) offset (down to the coast) by the Clarkwood fault. 3.3 km east of Bluntzer, the terrace is shown occurring at an elevation of 21-24 m (70-80 ft), approximately 15-18 m (50-60 ft) above the floodplain. West of Odem, there is a lower terrace, which he named the Angel ita Terrace. The type area for the Angel ita Terrace is the surface 5 km to the southwest of Odem at an elevation of 4. 5-7. 5 m (15-25 ft). Near Bluntzer, it is shown at an elevation of 9-10.5 m (30-35 ft), approximately 4.5 m (15 ft) above the floodplain. Weeks (1945: 1707) apparently considered the gravels of the Angelita Terrace near Calallen (probably the Fordyce pit at San Patricio) equivalent to the Uvalde gravels. He (1945: Fig. 4) showed four terraces above the Nueces River west of Dinero (outside of our study area), where the river is entrenched in the Goliad and/or Fleming Formations, but did not discuss them. The highest terrace near Dinero is approximately 30 m above the river bed; the lowest 9 m. Doering (1956: Fig. 6) showed three terraces in the area between Odem and San Patricio at 3-4.5 m (10-15 ft), 12 m (40 ft), and 15 m (50 ft) above stream level. The uppermost terrace of Weeks (1945) near Dinero is shown as a continuation of the Beaumont surface northwest of Mathis; the lower three correspond to the terraces near Odem. Doering (1956) interpreted that the two upper terraces between Odem and Bluntzer were equivalent to the Corpus Christi Terrace as defined by Price (1939). Doering (1956) concluded that the middle terrace correlated with the single Corpus Christi Terrace east of the Clarkwood fault and restricted the name to this unit. He suggested that the upper terrace was equivalent to the Eunice (Sixth Street) Terrace of the Colorado River. Conkin et al. (1962) studied gastropod and other fossils from the Fordyce Quarry near San Patricio. This quarry is within the Angelita Terrace of this report. They determined that the terraces were late Wisconsinan based on fossil evidence. The Geologic Atlas of Texas CORNISH & BASKIN 193 (Barnes 1975) recognized three sets of exposed (Dewey ville) terrace deposits at 1.5-6 m (5-20 ft), 6-9 m (20-30 ft), and 12-16.5 m (40-55 ft) above the Recent Nueces floodplain. Baskin (1991) reported reworked early Pliocene horse remains along with late Pleistocene fossils in the Angelita Terrace near Odem and in the valley fill near Bluntzer. The present study has the advantage of more detailed USGS topographic maps (7.5 minute series, 1" = 2000’) available in 1969 and 1979, as compared to 15 minute series maps (1" = 5280’) prior to those years. This study recognizes four terraces (Figs. 1 & 2) at the following elevations above the floodplain: (1) 0-5.7 m (0-19 ft), (2) 8.4-9 m (28- 30 ft), (3) 10.8-13.5 m (36-45 ft), and (4) 14.4-16.5 m (48-55 ft). The lowest terrace corresponds with Price’s Angelita Terrace. The Angelita is distinctive in possessing large meander loops with a mean radius of curvature of 900 m. It dips below the floodplain in the lowest reaches of the floodplain, as illustrated by Price (1933). The second terrace probably is equivalent to the middle terrace of the Geologic Atlas and is here given the name Fort Lipantitlan Terrace. The Fort Lipantitlan Terrace is named for the type area at the Fort Lipantitlan historical site on the south side of the Nueces River valley, 4.8 km due west of San Patricio at an elevation of 19.5 m (65 ft). The third terrace is equivalent to Doering’s middle terrace, which occurs at a lower elevation than the type of the Corpus Christi Terrace and herein is called the Bluntzer Terrace. The type area is located 300 m north of Bluntzer along state highway 666 at an elevation of 16.8 m (56 ft). The uppermost terrace correlates with Price’s Corpus Christi Terrace west of the Clarkwood fault. This terrace is equivalent to the uppermost of Doering’s (1956) three terraces. Nueces River Allostratigraphic Units An additional unit, the Cayamon Creek Alloformation, is recognized for the valley fill beneath the present day floodplain between Odem and Sandia (Figs. 1 & 2). Allostratigraphic units are three-dimensional sedimentary units that are delimited by their bounding unconformities (North American Commission on Stratigraphic Nomenclature 1983). This unit is defined from two sections measured at the Wright Materials, Inc. quarries in Nueces County on the south side of the Nueces River, 4 and 4.5 km respectively south southwest of San Patricio. The surface elevation of the quarries is approximately 10 meters above sea level, the level of the flood plain. A pump is used to lower the water level and the pits are quarried by a drag line. The drag- lines quarry a total of 13 194 THE TEXAS JOURNAL OF SCIENCE- VOL. 47, NO. 3, 1995 36ni 27 18 9 0 • 9 Figure 1. Cross sections of the Nueces River valley. A. cross section from south of Fort Lipantitlan to near Willow Lake. B. cross section from Bluntzer through the gravel pits to San Patricio. Locations of sections are shown in Figure 2. Abbreviations: AT, Angelita Terrace; BF, Beaumont Formation; B/LF, Beaumont and/or Lissie Formation; BT, Bluntzer Terrace; c, colluvium; CAla, Cayamon Creek Allomember 1, sandy gravel unit; CCA, Cayamon Creek Alloformation; CAlb, Cayamon Creek allomember 1, sand and muddy sand units; CA2, Cayamon Creek allomember 2; CA3, Cayamon Creek allomember 3; CCT, Corpus Christ! Terrace; LF, Lissie Formation; LT, Fort Lipantitlan Terrace; NP, north pit, Wright Materials, Inc.; NR, Nueces River; and SP, south pit, Wright Materials, Inc. to 15 meters from the surface. The lower one to two meters are under water even during pumping. Quarrying is halted on encountering a yellow-green clay, presumably representing the Beaumont Formation, or a calcareous-cemented sandstone. This sandstone could possibly represent the Goliad Fm. , but it is also encountered at the Odem locality (Baskin 1991) which is too far downdip to encounter the Goliad at such a shallow depth. A 14 meter section (Baskin & Cornish 1989) was measured in the north wall of a pit 1 km northeast of FM 3088 (Figs. 1 & 2: NP). A 13.5 meter section was measured in the northwest wall of a pit 0.7 km southwest of FM 3088 (Figs. 1 & 2: SP). The Cayamon Creek Alloformation can be subdivided into three allomembers: allomember 1, the lowermost sand and gravel dominated unit; allomember 2, a mud dominated unit; and allomember 3, a fine sand unit. Allomember 1 is by far the thickest of the three. The lower 3 to 4 meters of allomember 1 are dominated by gravel. The coarsest material is from the base of the section (mainly underwater) . The long diameter of the 20 largest cobbles easily visible at the base of the gravel pit section had an average diameter of 13.5 cm and a maximum diameter of 18 cm. The lower 1.5 to 2.5 m are mainly a sandy pebble gravel, with crudely horizontal and tabular-planar bedding (up to 60 cm thick), interbedded with cross-bedded sand. CORNISH & BASKIN 195 Figure 2. Map of the Nueces River valley showing stratigraphic units in the area of study. Contours are taken from the San Patricio quadrangle, USGS 7.5 minute series. The lines passing through A” and through B-B’ show the locations of the cross sections shown in Figure 1 . Abbreviations are those of Figure 1 . Pebble imbrication is present, but rare. The sandy gravel consists of 55-65% gravel, 35-45% sand, and 1-2% mud, and has a bimodal distri¬ bution, with modes at approximately 5.75 and 0.30 mm. The cross- bedded sand has a modal value of 0.21 mm and contains about 90% of sand and 5% of mud and gravel respectively. The bottom part of the unit includes two or three major cut and fill events. Gravel units are separated by 20 to 50 cm beds of slightly gravelly muddy sands. 196 THE TEXAS JOURNAL OF SCIENCE- VOL. 47, NO. 3, 1995 The next 1.5 m are mainly sand, arranged in lateral accretion units dipping 4-5 ° . The predominant sedimentary structure in this subunit of the section is tabular-planar cross-bedding (up to 1 m thick), as well as trough cross-bedding. Within and at the top of the unit there are thin gravel lenses. The sands are moderately well to very well sorted and have mean size ranges of 0.25 to 0.18 mm. The top of this unit is marked by a locally thin sandy gravel capped by a sandy mud. The next 4 m consist of alternating muddy sands and sandy muds. The beds are 5 to 30 cm thick, with a tendency for thinner beds higher in the section. A possible soil horizon occurs about 1 m above the base of this unit. Near the top of the unit are scattered, small caliche nodules. Allomember 2 unconformably overlies the lower unit and varies in thickness from 2 to 3 m. It is a sandy mud that ranges from a light orange below to dark-gray or black at the top. The dark sandy clay has a rich terrestrial and fresh- water gastropod fauna. The upper unit (allomember 3) is a poorly-exposed, fine sand that is channeled into the underlying units. This unit is not present at the south pit. Fluvial Architecture of the Cayamon Creek Alloformation Cayamon Creek allomember 1 has an overall fining-upward trend, related to decreasing flow strength and depth, but the lower part of the section, in particular, records several cut and fill episodes. Thickness of individual units can vary considerably over distances of tens of meters, presumably related to variation in distance from the axis of the paleochannel. The basal gravel, cut and fill, vertical decrease in grain size, lateral accretion bedding, and the suite of sedimentary structures described above all indicate that allomember 1 was deposited by a coarse-grained meandering fluvial system. Coarse-grained meandering fluvial systems are well described by Bluck (1971), McGowan & Garner (1970), and Jackson (1978), among others. The coarsest material from the base of the section represents channel lag deposits. The gravel cobbles consist mainly of brown chert derived from Cretaceous carbonates of the Edwards Plateau, black chert from the Maravillas Formation of West Texas, caliche from the nearby Goliad Formation, and silicified wood from Tertiary Coastal Plain sediments (Russell 1981). There are also clay galls, derived from the underlying Beaumont Formation. Long, intermediate, and short diameters were measured for a sample of the 13 largest chert cobbles easily visible at CORNISH & BASKIN 197 the base of the section. The mean intermediate diameter is 7.8 cm. Five large cobbles from the sorted material pile had a mean intermediate diameter of 10.3 cm. Gustavson (1978) recorded mean intermediate clast length for the 10 largest cobbles from 13 sample sites along the upper Nueces River south of the Balcones Escarpment as ranging from 3.5 to 10.8 cm. Velocities of about 3-6 m/sec at 1-10 m above the stream bed are required to transport particles this size (Sundborb 1956). Floods this size occur at an average interval of 8 years on the upper Nueces (Gustavson 1978). Similar velocities must have been in force in the lower Nueces in the late Pleistocene. Runoff from major tropical storms and hurricanes is probably sufficient to account for this stream competence. The maximum discharge for the Nueces River at Three Rivers (72 km upstream from the study area) since 1875 was 3,950 m^/sec following Hurricane Beulah on 23 September 1967. The previous maximum flow was 2,380 m^/sec on 18 September 1919 after a major hurricane. The maximum recorded flow for the Nueces River at Mathis before the construction of the Wesley Seale Dam was 1,650 mVsec on 20 September 1919. The sand-dominated, lateral accretion beds represent point bar deposits. Gravel lenses in this unit could represent some minor abandoned channel fill units or chute bars. The uppermost alternating sequence of sands and muds result from overbank crevasse splay deposition or near-channel levee deposits. These are similar to "rhythmites" in crevasse splay sequences of the Mississippi River (Farrell 1987) that developed by overbank sheet floods and were deposited by waning flow. They represent vertical accretion of the flood plain. Subaerial exposure is indicated by numerous caliche zones, similar to flood plain deposits on the Brazos River (Bernard & Major 1963). Allomembers 2 and 3 represent early to late Holocene floodplain deposits. They may be contemporaneous modern deposits of the Nueces River and represent flood plain muds and channel /point bar sands, respectively, of a muddy meandering fluvial system. Chronostratigraphy Formation of the Nueces River terraces and valley fill is related to changing climate and sea level during the late Quaternary. Possible age relationships of the alluvium to oxygen isotope stages are discussed in Baskin (1991). The Angelita Terrace predates the late Wisconsinan 198 THE TEXAS JOURNAL OF SCIENCE— VOL. 47, NO. 3, 1995 maximum drop of sea level and is probably equivalent in age to the Eagle Lake Alloformation of the lower Colorado River (Blum & Valastro 1994), where the river is cut into the alluvial plain of the Beaumont Formation. Blum & Valastro (1994) correlated the Eagle Lake Alloformation with the Sixth Street Terrace in the Austin area and stated that deposition of this unit took place from 20,000 to 14,000 YBP. The Nueces River valley was excavated at least 35 m into the Beaumont Formation during the maximum drop in sea level during isotopic stage 2, approximately 15,000 YBP (Baskin 1991). Blum et al. (1994) stated that excavation of bedrock valleys in the Edwards Plateau by the Colorado River took place from about 14,000 to 11,000 YBP. Following the late Wisconsinan low stand, sea level began to rise (oxygen isotope stage 1), continuing on into the Holocene, reaching its present level about 5000 YBP. Broecker et al. (1988) stated that large meltwater influxes into the Gulf of Mexico began around 14,000 YBP, peaked rapidly around 12,500 YBP, and then decreased rapidly. This time of rapid melting and concomitant sea level rise corresponds to the time of onset of deposition of the valley fill deposits. Toomey et al. (1993) stated that cave faunas from the Edwards Plateau indicate that average summer temperatures increased rapidly from 15,000-13,000 YBP to near present values and that effective moisture decreased and then increased again from 14,000-10,500 YBP. This time of dry conditions on the Edwards Plateau corresponds to meltwater discharge peaks in the Gulf of Mexico. Blum et al. (1994) stated that deposition of complex valley fills took place in the last 11,000 years. Cayamon Creek allomember 1 of the Nueces River valley corresponds with Columbus Bend Allomember 1 of the lower Colorado River, which was deposited from approximately 12,000 to 5,000 YBP (Blum & Valastro 1994). The oldest date they reported for this unit is 12,970 ± 640 YBP. However Lundelius (1992) reported a radiocarbon date of approximately 15,000 YBP for the equivalent First Street Terrace in Austin. This and the radiocarbon date for the Cayamon Creek allomember 1 may indicate that deposition of the valley fill may have begun somewhat earlier than 12,000 YBP. Pleistocene fossils and a piece of wood that was radiocarbon dated were collected in place in the cross-bedded gravelly sands and sandy gravels of Cayamon Creek allomember 1. The wood was from the north pit, approximately 4 m above the base of the section. The sample was dated by the Southern Methodist University Radiocarbon laboratory following a partial cellulose extraction pretreatment. The fractionation CORNISH & BASKIN 199 corrected age for sample SMU 2306 is 13,230 ± 110 YBP with a d of -26.3% (Haas, pers. comm.). The wood sample had absorbed large amounts of sulfur, which may indicate deposition in a swampy environment (Haas pers. comm.) In the gravels in the lower part of the section are scattered lenses of pebbles coated with manganese dioxide. Cayamon Creek allomember 2 corresponds to Blum et al.’s (1994) Columbus Bend Allomember 2, which was deposited from approximately 5,000 to 1,000 YBP. The snail-bearing upper unit of allomember 2 has a C-13 corrected date on snail shells of 965 ± 95 YBP with a d ^^C of -8.5% (Krueger Enterprises, Geochron Laboratories sample GX- 19928). This unit may be the same as Conkin et al.’s (1962) snail bearing unit from the top of the Fordyce Quarry, which they concluded was late Pleistocene. Toomey et al. (1993) suggested that the Edwards Plateau was experiencing relatively more mesic conditions from 2500-1000 YBP. This may explain the cooler and more humid conditions indicated by the Fordyce snails that Conkin et al. (1962) thought were indicative of a late Wisconsinan age. Conclusions The formation of Gulf Coast terraces was first addressed in detail by Fisk (1944). He related Mississippi Valley terrace levels to glacio-eustatic sea level changes: sea level highstands caused floodplain formation, while lowstands caused downcutting through the flood plains, leaving behind an elevated terrace. It is implicit in this theory that all rivers debouching into the Gulf of Mexico should have similar terrace sequences. However, several recent studies suggest that climate as well as glacio-eustatic changes affect terrace formation. Blum & Valastro (1994) suggested that changes in climate, coupled with degradation of upland soil mantles, affected the rate of runoff and sedimentation rates. Chronostratigraphic correlation of valley fill/terrace sequences between the Nueces River and the Colorado River support an argument against strict glacio-eustatic controls for terrace formation. In the Colorado River valley, the Columbus Bend Allomember 2 and Eagle Lake Allomember are exposed as terraces in the upper reaches of the valley (Blum and Valastro, 1994). However, these terraces are buried below the modern flood plain in its lower reaches. This effect is apparent for the Angelita Terrace of the Nueces River valley. Within the study area, it dips from 9 m above the valley floor, just below the Wesley Seele Dam to floodplain level immediately east of IH 37. 200 THE TEXAS JOURNAL OF SCIENCE-VOL. 47, NO. 3, 1995 The chronostratigraphic equivalent of the Columbus Bend Allomember 2 is the Cayamon Creek allomember 2, which is not exposed as a terrace, but as floodplain valley fill. This downdip change in relative terrace height indicates that terrace height above the valley floor is not useful for between- valley correlations. Four terraces in the Nueces valley would not necessarily correlate to four terraces in any other valley. Different dates for similar late Pleistocene terrace levels in different valleys caused controversy among workers using the Fisk concept. Some researchers found it easy to dismiss other workers’ radiocarbon dates that did not fit their model. Now, with more chronostratigraphic studies and a different concept of terrace formation, researchers should be able to piece together the complex history of external controls on late Quaternary Gulf Coast fluvial processes. Acknowledgments This study was funded in part by Organized Faculty Research, Texas A&I University. We are very grateful to the management and staff of Wright Materials, Inc., particularly Milus Wright, Ruth Wright, and Mark Truesdale, for access to their sand and gravel pits. George Sorensen granted permission to investigate the sand and gravel pit on his property. M. Guzler of Texas A&I performed the grain-size analyses. R. Thomas of Texas A&I assisted with the field work. The figures were prepared thanks to the invaluable assistance and unlimited patience of C. Tipton of Texas A&I. Literature Cited Barnes, V. E. 1975. Geologic Atlas of Texas; Scale 1:250,000; Corpus Christ! Sheet. Bur. Econ. Geol., Univ. Texas. Baskin, J. A. 1991 . Early Pliocene horses from late Pleistocene fluvial deposits. Gulf Coastal Plain, South Texas. J. Paleontol., 65:995-1006. Baskin, J. A. & F. G. Cornish. 1989. Late Quaternary fluvial deposits and vertebrate paleontology, Nueces River valley. Gulf Coastal Plain, South Texas, p. 23-30. In South Texas clastic depositional systems. (J. A. Baskin and J. S. Prouty, eds.) Gulf Coast Assoc. Geol. Soc. 1989 Convention Field Trip, Corpus Christ! Geol. Soc. Bernard, H. A. & C. F. Major, Jr. 1963. Recent meander belt deposits of the Brazos River: an alluvial ’sand’ model. Am. Assoc. Petrol. Geol. Bull., 47:350. Bluck, B. J. 1971. Sedimentation in the meandering River Endrick. Scottish J. Geol., 7:93-138. CORNISH & BASKIN 201 Blum, M. D., R. S. Toomey, III, S. Valastro, Jr. 1994. Fluvial response to Late Quaternary climatic and environmental change, Edwards Plateau, Texas. Palaeogeogr. , Palaeoclimatol. , Palaeoecol. , 108:1-21. Blum, M. D. & S. Valastro, Jr. 1994. Late Quaternary sedimentation, lower Colorado River, Gulf Coastal Plain of Texas. Geol. Soc. Am. Bull., 106:1002-1016. Broecker, W. S., M. Andree, W. Wolfli, H. Oeschger, G. Bonani, J. Kennett & D. Peteet. 1988. The chronology of the last deglaciation: implications to cause of the younger Dryas event. Paleoceanography, 3:1-19. Conkin, J. E., B. M. Conkin & W. T. Mason, Jr. 1962. Pleistocene snails from San Patricio County, Texas. Trans. Kentucky Acad. Sci., 23:25-50. Deussen, A. 1924. Geology of the coastal plain of Texas west of the Brazos River. U.S. Geol. Surv. Professional Paper, 126:1-145. Doering, J. A. 1956. Review of Quaternary surface formations of Gulf Coast region. Am. Assoc. Petrol. Geol. Bull., 40:1816-1862. Parrel, K. M. 1987. Sedimentology and facies architecture of overbank deposits of the Mississippi River, False River Region, Louisiana. Pp. 111-120, In Recent developments in fluvial sedimentology (F. G. Ethridge, R. M. Flores, and M. D. Harvey, eds.), Soc. Econ. Paleontol. Mineral. Spec. Publ., 39:1-389. Fisk, H. N. 1944. Geological investigation of the alluvial valley of the lower Mississippi River. Mississippi River Commission, Vicksburg, 78 pp. Gustavson, T. C. 1978. Bedforms and stratification types of modern gravel meander lobes, Nueces River, Texas. Sedimentology, 25: 401-426. Jackson, R. G., 11. 1978. Preliminary evaluation of lithofacies models for meandering alluvial streams. Pp. 543-576, In Fluvial sedi¬ mentology (A. D. Miall, ed.), Canadian Soc. Petrol. Geol., Mem. 5. Lundelius, E. L. 1992. The Avenue Local Fauna, late Pleistocene vertebrates from terrace deposits at Austin, Texas. Ann. Zool. Fennici, 28:291-299. McGowan, J. H. & L. H. Garner. 1970. Physiographic features and stratification types of coarse-grained point bars: modern and ancient examples. Sedimentology, 14:77-111. North American Commission on Stratigraphic Nomenclature. 1983. The North American Stratigraphic Code. Am. Assoc. Petrol. Geol. Bull., 67:841-875. 202 THE TEXAS JOURNAL OF SCIENCE- VOL. 47, NO. 3, 1995 Price, W. A. 1933. Role of diastrophism in topography of Corpus Christi area, South Texas. Am. Assoc. Petrol. Geol. Bull., 17:907- 962. Russell, J. L. 1981. Fluvial, eolian, and estuarine-lagoonal depositional environments of South Texas. Pp. 35-41, In Modern depositional environments of sands in South Texas. Corpus Christi Geol. Soc., Field Trip for 1981 Meeting of Gulf Coast Assoc. Geol. Soc. Sundborg, A. 1956. The River KlarMven; a study of fluvial processes. Geografiska Annaler, 38:125-316. Toomey, R. S., Ill, M. D. Blum & S. Valastro, Jr. 1993. Late Quaternary climates and environments of the Edwards Plateau, Texas. Global and Planetary Change, 7:299-320. U.S. Bureau of Reclamation. 1983. Nueces River basin: a special report of the Texas basins project. 345 pp. Weeks, A. W. 1945. Quaternary deposits of Texas Coastal Plain between Brazos River and Rio Grande. Am. Assoc. Petrol. Geol. Bull., 29:1693-1720. TEXAS J. SCI. 47(3):203-222 AUGUST, 1995 ENVIRONMENTAL ASSESSMENT OF LA QUINTA CHANNEL, CORPUS CHRISTI BAY, TEXAS Christopher M. Martin and Paul A. Montagna University of Texas at Austin, Marine Science Institute, P. O. Box 1267, Port Aransas, Texas 78373 Abstract.— Benthic invertebrate communities sampled during 1992 and 1993 from each of three stations in La Quinta Channel and Corpus Christ! Bay were analyzed. Higher numbers of individuals were found in the top 3 cm at all stations during each sampling period. Biomass, however, was greater in the bottom (3-10 cm) sediment section. There were higher numbers of species, greater species richness, and more evenness in La Quinta Channel than in Corpus Christ! Bay. Both communities were dominated by deeper-dwelling equilibrium species and the low abundance of pioneering species. The results of this study indicate that the sediments of La Quinta Channel differ little from those of Corpus Christ! Bay. Results are consistent with findings that contamination concentrations are low in the La Quinta Channel when compared to those of Corpus Christ! Bay. Man has caused negative impacts on estuaries in a variety of ways, including chemical pollution, reduced inflow, channelization and suspended wastes. Corpus Christi Bay is a large bay near Corpus Christi, Texas, the sixth largest port city in the United States. Corpus Christi has a population of over 250,000 people. Its metropolitan center is located along the west side of the bay. La Quinta Channel traverses the north shore of Corpus Christi Bay and is a spur of the main Corpus Christi Ship Channel which serves the Port of Corpus Christi (Fig. 1). La Quinta Channel serves several industrial plants, including DuPont Chemical, Reynolds Aluminum, and Occidental Chemical Company. This concentration of industry has led to public concerns about the health of this environment. It is necessary to determine whether or not anthropogenic environmental stresses have affected La Quinta Channel. Benthic organisms have been especially useful to detect pollution and to estimate overall effects of pollution on a community (Oglesby 1967; Houston et al. 1983; Ferraro et al. 1991). Benthic species and com¬ munities, rather than fish or planktonic fauna, have often been regarded as being the best indicators of organic pollution because of their limited mobility (Wass 1967; Flint et al. 1980). For example, an estuary under severe environmental stress will be depauperate of benthic species, while a similar estuary that is relatively unaffected by human activities will generally have a much higher number of benthic species (Marques et al. 204 THE TEXAS JOURNAL OF SCIENCE-VOL. 47, NO. 3, 1995 Figure 1. Map of Corpus Christi Bay reference stations. Site NCC is located near the mouth of Nueces Bay, NCD is locat^ north of the Laguna Madre, and NCE is just south of Point Mustang. Station LQA is on the north end of La Quinta Channel, LQB is located in Ingleside Cove, and LQC is near the juncture of the La Quinta Ship Channel with the Corpus Christi Ship Channel. All stations are away from any obvious anthropogenic influences. 1993). There are several reasons why benthic organisms are good indi¬ cators of environmental stress (Soule 1988). (1) Because of gravity, pollutants end up in bottom sediments. Even pollutants in freshwater will be transported to coastal sea bottoms. (2) Living organisms die and end up in the detrital food chain, which is utilized by the benthos. Pollutants are usually tightly coupled to organic matrices, therefore benthos have great exposure through their niche (food) and habitat (living spaces) to pollutants. (3) Benthos are relatively long-lived and sessile and thus integrate the effect of pollutants over long temporal and spatial scales. (4) Benthic invertebrates are sensitive to pollutants. (5) Bioturbation and irrigation of sediments by benthos affect the mobilization and burial of xenobiotic materials. The objective of this study was to measure the environmental condi¬ tions of La Quinta Channel to assess if the Channel has been affected by anthropogenic input. The approach used was to compare the benthic community of La Quinta Channel shoal areas with reference stations in Corpus Christi Bay. This was accomplished by measuring benthic abun¬ dance, biomass, and community diversity for one year. The relative MARTIN «& MONTAGNA 205 characteristics of the benthos in the two study areas is assumed to be indicative of the environmental conditions present. Methods Study Area Description.— Tht Corpus Christi Bay system is one of seven major estuarine systems along the Texas Gulf Coast. The open bay bottom surface area totals 432.98 km^ (Flint et al. 1983). It is separated from the Gulf of Mexico by a barrier island, Mustang Island. There is one main tidal inlet, Aransas Pass, and two forms of fluvial flow, the main being Nueces River and the other Oso Creek. The eco¬ system includes the following bodies of water: Nueces Bay, Corpus Christi Bay, Oso Bay, the northern portion of the upper Laguna Madre, and the southern portion of Redfish Bay (Fig. 1). Flint et al. (1981) noted that, due to the small volume of riverine input, salinities reflect a more oceanic condition. They also stated that the Corpus Christi Bay system is relatively sensitive to changes from factors such as infrequent surges of freshwater usually occurring from large storms in the watershed. Bottom sediments within Corpus Christi Bay are primarily mud in the interior portions, with muddy and shelly sands around the shoals (Flint et al. 1980). The composition of the suspended sediment load is primarily inorganic silt and clay detritus, with a subordinate organic skeletal fraction dominated by diatoms (Shideler 1980). Sampling Sites.— A total of six sampling stations (Fig. 1) were examined; three stations in the shoal areas of La Quinta Channel and three reference stations in Corpus Christi Bay away from any obvious anthropogenic influence. The stations were named for their location and station (i.e., Nueces-Corpus Christi/Station C was NCC and La Quinta Channel shoal area/Station A was LQA). Stations NCC (3.4 m) and NCD (2.7 m) are part of a series of stations that have been sampled previously (Montagna & Kalke 1992), and NCE (3.4 m) was added in April, 1991. Samples were collected in October, 1992, and January, April, and July, 1993. The stations in La Quinta Channel are LQA (2.5 m), LQB (1.8 m), and LQC (2.5 m) (Fig. 1). Station LQA was behind Occidental Chemical Company on the south side of the channel, station LQB was in Ingleside Cove, and station LQC was 225 m west southwest of Naval Station Ingleside. The sampling area at each station was 2 m^. Hydrography.— Hydrographic measurements were taken just below 206 THE TEXAS JOURNAL OF SCIENCE— VOL. 47, NO. 3, 1995 the water surface and at the bottom of the water column using a Hydrolab Surveyor II. Data (units and accuracy) were collected on: depth (±1 m), water temperature (±0.15 C), pH (±1 unit), dissolved oxygen (±0.2 mg I'O, specific conductivity (±0.015-1.5 mmhos cm'^ depending on range), redox potential (±0.05 mV), and salinity (±0.7 ppt. Measurements were taken at all stations during each collection. Sediment —Stdimtni samples were taken during the first collection period. Each sample was placed in a jar and filled with distilled water and hydrogen peroxide. The samples were allowed to sit for one week to digest organic material. The samples were filtered, dried, and weighed (mg) (Folk 1964). The percent of rubble, sand, silt, and clay was then calculated. Benthic Abundance and Biomass.— abundance and community structure were measured using the standard techniques Montagna & Kalke (1992) have been using since 1984. Samples were collected using core tubes that were 6.7 cm in diameter yielding a sample area of 35.4 cm^. The cores were sectioned to sediment depths of 0-3 cm and 3-10 cm to examine the vertical distribution of macrofauna. Each sediment section was stored in 4% Formalin made with filtered sea- water. Animals were extracted with a 0.5 mm mesh sieve. After the species had been identified and enumerated, biomass was measured. Mollusk shells were removed by an acidic vaporization technique (Hedges & Stern 1984). Animals were separated into their major groups: annelids, crustaceans, mollusks, nemerteans, ophiuroids, sipunculids, and others and put on aluminum pans. The pans were dried at 50 °C for a minimum of 48 h to obtain dry weight biomass. Diversity Analyses.— DivQxsiiy was calculated using Hill’s diversity number one (Nl) (Hill 1973). It is a measure of the effective number of species in a sample, and indicates the number of abundant species. It is calculated as the exponentiated form of the Shannon & Weaver (1949) diversity index {H"). Richness is an index of the number of species present. The obvious richness index is simply the total number of all species found in a sample regardless of their abundances. Hill (1973) named this index NO. Another well known index of species richness is the Margalef (1958) index (Rl). R1 is based on the relationship between the number of species and the total number of individuals observed. Evenness is an index that expresses that all species in a sample are MARTIN & MONTAGNA 207 equally abundant. Evenness is a component of diversity. El is probably the more common, it is the familiar J' of Pielou (1975). It expresses H* relative to the maximum value of H\ El is sensitive to species richness. The variable El was chosen because it is calculated with NO and Nl. When El is equal to 1.0 then there is complete evenness. Statistical Analyses.—Si'di\\s\AC2i\ analyses were performed on sediment, biomass, abundance, and diversity data using general linear model procedures to reveal differences among sampling periods, stations, and sediment depths (SAS 1985). Three-way analysis of variance (ANOVA) models were used where sampling dates, stations, and sediment sections were the three main effects. Two-way ANOVA models were used where sampling dates and stations were the two main effects on total cores. The residuals were tested for conformance to normality using the Shapiro-Wilk statistic (SAS 1985). Orthogonal linear contrasts were used to test the a priori null hypothesis that La Quinta Channel stations were different from Corpus Christi Bay stations. Results Hydrography .—Mtdin salinity readings ranged from 23.5 - 30.5 ppt in La Quinta Channel from October 1992 to July 1993, due to rainfall that occurred during the year, especially in May 1993. The trend was similar in Corpus Christi Bay, although the salinity was slightly lower. There was a large seasonal drop in temperature, from 24 C to 13 C, between October and January, but the temperature increased to 40 C by July. The pH remained constant (7.9 units) and was similar among each of the six stations all year. Dissolved oxygen (DO) concentration in the water had a mean of 7.0 mg f’ for all stations except NCC and NCD, which were lower (5. 5-6. 5 mg I'O, during October. In January and April DO was higher (8. 5-9. 5 mg f') for all stations. By July it had decreased to 5. 0-6. 5 mg and was similar at all stations except for the bottom at NCD, where it was only 1.7 mg 1’^ The redox potential remained constant (0.23 mV) during the entire year at all stations. Overall, parameters recorded at each of the stations were very similar to one another. Sediment .SdiVid was predominant at most stations, except for NCC (Fig. 2). The sand content was usually greater in the top 3 cm of the samples. This measurement was taken only once in October, 1992 (NCD was taken October, 1991). All of the La Quinta Channel stations 208 THE TEXAS JOURNAL OF SCIENCE-VOL. 47, NO. 3, 1995 Stations ■ rubble M sand M silt ^ clay Figure 2. Sediment composition of La Quinta Channel and Corpus Christi Bay stations. Percent dry weight of sediment components in each station (station NCD sample taken October, 1991; all other stations taken October, 1992). had sand comprising greater than 75% of the sediment. Both layers at NCC consisted mostly of silt and clay, making up 79.7% in the upper section and 89.2% in the lower section, while there was more sand in NCD (82.4%) and NCE (56.7%). Benthic Abundance and Biomass. — There were more individuals in the top 3 cm of sediment during each study period, except for July 1993 (Fig. 3). The average number of individuals found in the top section for all stations and periods was 13,024 m’^, which is 61% of the total average of both sections. The average of the top section at LQA was 12,126 individuals m‘^, which is lower than in the bottom section (13,307 individuals m'^). In October 1992, and January and April 1993, the average density in the upper layer was 67 % , while the density was only 41% in the upper layer for July 1993. There were significant interactions (3-way ANOVA, P=0.0003) between the sections, dates, and stations for the density. Average macrofauna density, to 10 cm depth, in La Quinta Channel was not significantly different from Corpus Christi Bay (linear contrast, P=0.0873). La Quinta Channel had an average of 25,796 individuals m'^ to 10 cm, while Corpus Christi Bay exhibited 20,123 individuals m'^ MARTIN & MONTAGNA 209 60,000 50,000 P ^ 40,000 I 30,000 CD "D i 20,000 < 10,000 0 Oct 1 992 Jan 1 993 Apr 1 993 Jul 1 993 ■ 3-10 cm mO-3 cm Figure 3. Vertical distribution of macrofaunal density (mean n m'^) in La Quinta Channel and Corpus Christ! Bay for each station and sampling period (sediment cores were vertically sectioned at 0-3 cm and 3-10 cm intervals). to 10 cm. The three stations in La Quinta Channel and NCE in Corpus Christi Bay had similar numbers of species throughout the study periods, while NCD was consistently lower (Fig. 3). La Quinta Channel and Corpus Christi Bay both had a greater bio¬ mass (g m'^) in the bottom (3-10 cm) sediment section (Fig. 4). La Quinta Channel had an average of 72% and Corpus Christi Bay aver¬ aged 69% of the biomass in the bottom sections. Overall, the percent of biomass found in the top 3 cm was 29 % , but composition was differ¬ ent among stations. LQA had 17% in the top section, LQB had 35%, there was 39% at LQC, NCC had 23%, NCD was found to contain 50%, and NCE had 33% (Table 1). The high percentage of biomass at NCD occurred in January and April 1993, when NCD had 64% of the biomass in the top 3 cm while the other 2 sampling periods had less than 40% each in the top 3 cm at NCD (Fig. 4). There were significant interactions between the sections, dates, and stations for the biomass (3- way ANOVA, /^=:0.0126). Biomass in La Quinta Channel to a depth of 10 cm was significantly different from Corpus Christi Bay (linear contrast, P=0.0112). La 210 THE TEXAS JOURNAL OF SCIENCE-VOL. 47, NO. 3, 1995 ■ 3-10 cm 11 0-3 cm Figure 4. Vertical distribution of macro faunal biomass (mean g m'^) in La Quinta Channel and Corpus Christ! Bay for each station and sampling period (sediment cores were vertically sectioned at 0-3 cm and 3-10 cm intervals). Quinta Channel had an average biomass of 13.8 g m'^ and Corpus Christi Bay had an average biomass of only 10.0 g m‘^. Without station NCD, the Corpus Christi Bay average biomass would increase to 13.8 g ml Polychaetes were the dominant organisms at all stations in both the top and bottom sections of all samples throughout the study period (Table 1). Polychaetes made up 77 % of the mean total abundance found in the top 3 cm and 90% for the bottom 7 cm of La Quinta Channel and 84% of the total species in the upper section and 90% in the bottom section of Corpus Christi Bay samples (Table 1). Polychaetes had the greatest biomass at all stations and sections ranging from 0.8 - 9.3 g m'^ (Table 1). The biomass of all taxa, except polychaetes, was greater in the top 3 cm than in the bottom 7 cm (Table 1). Polychaete biomass in the bottom section made the overall average biomass of the bottom section greater than that of the top section. Benthic Diversity —Tht mean NO (total number of species in a sample) for all stations combined was 18 species. There was a difference in the mean total species between La Quinta Channel (13 MARTIN & MONTAGNA 211 Table 1 . Vertical distribution of macrofaunal taxa for entire study year, Oct 1992 - July 1993. Mean biomass (g m'^) and abundance (n m'^) of taxonomic categories. SD = Standard Deviation. Station Taxa Section 0-3 cm 3-10 cm n m'^ g m'^ n m'^ Mean SD Mean SD Mean SD Mean SD LQA Crustacea 449 372 0.024 0.021 236 398 0.016 0.033 Hemichordata 0 0 0 0 24 32 0.020 0.070 Mollusca 544 573 0.169 0.331 544 901 0.138 0.271 Nemertea 284 270 0.229 0.513 236 266 0.155 0.326 Other 142 226 0.024 0.045 165 283 0.602 1.476 Ophiuroidea 95 140 0.008 0.019 0 0 0 0 Polychaeta 10,613 4,794 3.049 2.095 12,078 4,846 15.692 9.694 Sipunculida 0 0 0 0 24 82 0.010 0.033 TOTAL 12,127 6,375 3.503 3.024 13,307 6,808 16.633 11.903 LQB Crustacea 473 425 0.027 0.022 142 284 0.012 0.028 Mollusca 473 677 0.149 0.348 165 352 0.022 0.043 Nemertea 165 225 0.019 0.028 142 284 0.023 0.042 Other 2,600 5,259 0.622 1.297 189 425 0.009 0.021 Ophiuroidea 71 176 0.049 0.162 24 82 0.214 0.743 Polychaeta 8,367 3,556 3.003 4.595 8,202 3,896 6.764 7.155 Sipunculida 24 82 0.001 0.004 24 82 0.005 0.016 TOTAL 12,173 10,400 3.870 6.456 8,888 5,405 7.049 8.048 LQC Crustacea 780 438 0.064 0.100 118 190 0.027 0.074 Hemichordata 24 82 0.005 0.018 0 0 0 0 Mollusca 1,064 757 0.572 1.013 615 896 0.101 0.165 Nemertea 449 255 0.482 1.228 355 299 0.096 0.085 Other 567 1,128 0.054 0.114 71 128 0.015 0.032 Ophiuroidea 165 190 0.190 0.470 24 82 0.005 0.018 Polychaeta 10,069 3,262 2.613 3.100 6,547 3,456 6.045 4.822 Sipunculida 47 no 0.072 0.221 0 0 0 0 TOTAL 13,165 6,222 4.052 6.264 7,730 5,051 6.289 5.196 NCC Crustacea 449 534 0.014 0.018 118 225 0.015 0.039 Mollusca 449 534 0.664 1.070 24 82 0.035 0.120 Nemertea 165 190 0.024 0.036 189 221 0.261 0.584 Other 47 164 0.002 0.007 24 82 0.002 0.007 Ophiuroidea 236 237 0.598 1.686 165 190 3.670 6.732 Polychaeta 12,457 14,698 1.332 0.831 3,876 1,970 5.092 3.592 Sipunculida 95 185 0.067 0.164 0 0 0 0 TOTAL 13,898 16,542 2.701 3.812 4,396 2,770 9.075 11.074 NCD Crustacea 520 754 0.030 0.039 95 221 0.004 0.011 Mollusca 780 486 0.071 0.118 24 82 0.001 0.004 Nemertea 260 307 0.046 0.066 0 0 0 0 Other 71 128 0.243 0.719 0 0 0 0 Ophiuroidea 24 82 0.014 0.050 71 176 0.256 0.850 Polychaeta 13,898 14,467 0.806 0.626 5,815 10,316 0.951 0.976 Sipunculida 165 330 0.010 0.018 0 0 0 0 TOTAL 15,718 16,554 1.220 1.636 6,005 10,795 1.212 1.841 NCE Crustacea 1,702 2,739 0.092 0.156 544 534 0.073 0.088 Mollusca 638 470 0.249 0.427 142 191 0.022 0.039 Nemertea 307 255 0.071 0.077 189 252 0.294 0.690 Other 331 433 0.058 0.098 165 190 0.048 0.066 Ophiuroidea 165 225 0.050 0.101 142 191 0.724 1.357 Polychaeta 7,847 3,237 4.730 5.817 8,107 5,462 9.316 6.057 Sipuneulida 71 176 0.013 0.041 0 0 0 0 TOTAL 11,061 7,535 5.263 6.717 9,289 6,820 10.477 8.297 212 THE TEXAS JOURNAL OF SCIENCE- VOL. 47, NO. 3, 1995 OQ (J O Q O O O U -I -I Z Z Oct 1992 < CD s> {Sorghastrum nutans) series” (Diamond & Smeins 1985). Occurring with these dominant species are several other native grasses: switchgrass (Panicum virgatum), eastern gamagrass (Tripsacum dactyloides) , big bluestem (Andropogon gerardii), tall dropseed {Sporobolus asper), sideoats grama (Bouteloua curtipendula) , Florida paspalum (Paspalum floridanum), Scribner panicum {Dicanthelium oligosanthes), Texas wintergrass {Stipa leucotricha) , and fimbry (Fimbristylis puberula). A diverse array of forbs is present. Native woodland characterizes the drainages of Clymer Meadow. The principal species of trees are cedar elm {Ulmus crassifolia), bois d’arc {Madura pomifera), and Texas sugarberry {Celtis laevigata). The primary woody species in the under story are coralberry {Symphoricarpos orbiculatus) , roughleaf dogwood {Comus drurnmondii), and poison ivy {Toxicodendron radicans); very little of the ground surface is covered by forbs or grasses. Water flow in these draws is intermittent, with timing and amount of water related to pattern and amount of precipita¬ tion. Mottes of woody vegetation, surrounded by grassland, occur sporadically on the hillsides and hilltops. Generally, such a motte consists of a central bois d’arc or sugarberry tree beneath which is a thicket of poison ivy and coralberry. Methods and Materials Grid studies. — Population densities of small mammals were deter¬ mined by analyzing mark- recapture data obtained from live- trapping on grids and transects (Blair 1941; Hayne 1949). The method used to estimate population sizes was based on the minimum number of different individuals handled during a sampling session; this method is almost always an underestimate of population size (Overton 1972). 246 THE TEXAS JOURNAL OF SCIENCE-VOL. 47, NO. 4, 1995 Two trapping grids were established on the Parkhill tract of Clymer Meadow to sample rodents in the two primary terrestrial habitats representative of tallgrass prairies. One grid, the grassland grid, sampled unperturbed prairie. The other, the wooded-draw grid, sampled the wooded habitat characteristic of lower elevations and streambeds located between the higher, grassy hillsides and hilltops. These grids have been undisturbed by burning, grazing, or haying since before the Nature Conservacy of Texas acquired the property in 1986; prior to this, the only use of this tract was as a hay meadow. As such, the data from these grids represent a database useful in assessing effects of various perturbations of habitat. The grassland trapping grid consisted of 50 trap stations arranged in a rectangular 5 by 10 array. The grid in the wooded draw consisted of three columns of 10 stations each for a total of 30 traps; the columns followed the contours of the draw. Distance between adjacent trap stations in the same and adjacent rows was 15 m in the grassland grid and 10 m in the wooded-draw grid. The effective sampling areas of each grid consisted of the area within each grid plus the area of a boundary strip one-half the distance between traps (7.5 m wide for grassland grid; 5 m wide for wooded-draw grid; as per Blair 1942). Hence, the grassland and wooded-draw grids sampled 1. 12 ha and 0.30 ha, respectively. One Sherman livetrap (23 by 9 by 8 cm) was placed at each trap station on each grid. Trapping was conducted seasonally for 3 years, from April 1989 through January 1992. Spring, summer, fall, and winter sampling was conducted in April, July, October, and January, respectively. Each sampling session consisted of four consecutive days. Traps were placed, set, and baited with rolled oats on the afternoon of the first day, and were reset and rebaited on the afternoons of the second and third days. Traps were checked for presence of rodents and then closed soon after daybreak on the second, third, and fourth days. Closure of traps during daytime precluded mammals from entering traps during daytime. For every rodent live-trapped, species identity, sex, age, reproductive condition, body weight, and trap station were recorded in the field prior to release at the site of capture. On occasion of first capture, each rodent was marked to enable future identification upon recapture. Marking was by toe-clipping (Wilkins 1977), a traditional method sanctioned by the American Society of Mammal ogists (1987). Longevity for a particular species on a particular grid was determined by averaging WILKINS 247 the lengths of time from first to last capture for each individual that was recaptured at least once. Vegetation analysis. —Seasonal analysis of vegetative cover on each grid was conducted to assess relationships of quantity and type of vegetation to the rodent species present. A convenient method of canopy-cover analysis, modified from Daubenmire (1959), has worked effectively in other studies of small mammals (Wilkins 1977; Wilkins & Schmidly 1981) and was used as follows in this study: Cover provided by 15 categories of plants was estimated at selected heights above the soil surface: litter at 5 cm; standing dead material at 10, 25, and 50 cm; big bluestem at 25 and 50 cm; switchgrass at 25 and 50 cm; other grasses at 25 and 50 cm; forbs at 5, 10, and 25 cm; and woody species at 50 cm and 1 m. Included in the "other grasses" category were many of the important species, such as little bluestem, indiangrass, and eastern gamagrass. Amount of cover provided was encoded as follows: 0 indicates 0%; 1, 1-5%; 2, 6-25%; 3, 26-50%; 4, 51-75%; 5, 76-95%, 6, 96-100%. The values used in computation are the midpoints of the respective ranges: code 1, 3%; 2, 15.5%; 3, 38%, 4, 63%; 5, 85.5%; 6, 98%. Vegetative cover within a 0.2 by 0.5 m quadrat (area=0.1 m^) was assessed seasonally at the same 10 randomly selected trap stations in each grid during the 4-day interval when rodents were sampled. Statistical analysis of vegetation data was conducted by use of the MEANS and ANOVA procedures of the Statistical Analysis System (SAS Institute Inc. 1985). Sampling of small mammals away from grids .—Additional sampling was conducted at Clymer Meadow beyond the locations and habitats represented on the grids. Sampling the other subhabitats present (disturbed grassland, grassland inhabited by fire ants, woodland, streamside, fencelines) allowed detection of species perhaps not occurring in the habitats represented by the grids and, thereby, facilitated a more-complete inventory for the preserve. Comparison of species compositions of rodent communities in different habitats allowed asssessment of impact of various habitat perturbations. Sherman live traps (baited with rolled oats) were placed in various subhabitats in transects of 30-40 stations (approximately 10 m apart) for one night. Information recorded was number of individuals of each species in the transect and, later in the study, the number of traps in which imported fire ants {Solenopsis invicta) were found. The vast majority of these rodents were released unharmed; the few animals pre- 248 THE TEXAS JOURNAL OF SCIENCE— VOL. 47, NO. 4, 1995 Table 1. Summary statistics for rodent population densities (computed from minimum number known alive) on the grassland and wooded-draw sampling grids at Clymer Meadow, Hunt Co., Texas, 1989-1992. Listed are the means (and standard deviations) of number of rodents per hectare for each of six species averaged by season for the 3- year study period. Species Grid and Season Sigmodon hispidus Reithrodontomys Reithrodontomys Peromyscus fulvescens humulis leucopus Peromyscus maniculatus Neotoma floridana Grassland Spring 5.9 (2.8) 3.0 (3.7) 1.2 (1.0) 1.2 (1.4) 0.9 (0.9) Summer 16.7 (5.2) 2.7 (3.2) 0.9 (0.9) 1.2 (1.4) 0.6 (0.5) Fall 16.7 (4.1) 2.1 (1.9) 0.0 (0.0) 0.0 (0.0) 0.0 (0.0) Winter 12.8 (5.9) 4.5 (4.5) 0.0 (0.0) 0.3 (0.5) 0.3 (0.5) Wooded- draw spring Summer Fall Winter 0.0 (0.0) 0.0 (0.0) 1.1 (1.9) 0.0 (0.0) 32.2 (10.7) 24.4 (18.4) 16.6 (11.5) 16.7 (12.0) 23.3 (17.3) 0.0 (0.0) 4.4 (5.1) 32.2 (18.4) 3.3 (3.3) 14.4 (6.9) 4.4(1.9) 3.3 (3.3) pared as voucher specimens (deposited in the collection of vertebrates at Baylor University) generally were taken from these transects rather than from the grids. Results Population dynamics of grassland grid.—Vwt species of rodents, listed in order of decreasing abundance, were captured on the Parkhill grassland grid during 1,800 trapnights of sampling: Sigmodon hispidus (cotton rat; 153 different individuals), Reithrodontomys fulvescens (fulvous harvest mouse; 32 individuals), Reithrodontomys humulis (eastern harvest mouse, 7 individuals), Peromyscus leucopus (white¬ footed mouse; 7 individuals), and Peromyscus maniculatus (deer mouse; 7 individuals). Sigmodon hispidus was the only species caught every sampling session; densities based on minimum number of different individuals present ranged from 2.7/ha to 21.4/ha (Table 1). For each year of the study, springtime densities of Sigmodon hispidus were lower than for any other season. Though none were handled during the first year of the study, Reithrodontomys fulvescens was trapped during each of the remaining eight sampling sessions; maximum density was 8.9/ha. Individuals of the other three species were captured only sporadically during the study; maximum densities for Peromyscus leucopus. WILKINS 249 Peromyscus maniculatus, and Reithrodontornys humulis were estimated to be less than 3 /ha throughout the study (Table 1). Stick houses of eastern wood rats (Neotoma floridana) were con¬ structed at the bases of woody mottes scattered over the grassland portions of the prairie. One such motte and nest was located at the northeastern corner of the grassland grid, yet no wood rats were caught there. Longevity could not be assessed for Reithrodontornys humulis as none was ever recaptured. This species heretofore was unknown in Hunt County, Texas, thus two of the seven individuals were prepared as voucher specimens (Wilkins 1991). Three Peromyscus leucopus were recaptured, but all recaptures were during the same sampling session of their initial capture. Only one Peromyscus maniculatus was recaptured during a session later than its initial capture; the longevity of this male was at least 6 months. Nine different Reithrodontornys fulvescens were recaptured. Two were captured twice during the same session, four others were caught again during the next sampling session, one was recaptured 6 months after its first capture, another was recaptured 9 months after being marked, and one male was caught four times, the last occasion being 12 months after first capture. Mean minimum longevity computed from these nine mvXh^Xy-cdi^iuvtd Reithrodontornys fulvescens was 4.3 months, nearly twice the mean lifespan reported by Cameron (1977). Sixty of 153 Sigmodon hispidus handled (39.2%) were recaptured at least once. The greatest longevity evident among these was 6 months {n=7). The last recapture for 21 cotton rats was 3 months after initial capture. For 32 individuals, all recaptures were during the same session that they were marked. On the basis of these 60 animals, mean mini¬ mum longevity for Sigmodon hispidus on this grid was 1.75 months, slightly shorter than values (2.0 - 2.6 months) summarized by Schmidly (1983). Approximately 650 trapnights of effort were required for the initial capture of all five species of rodents that were caught on the grassland grid. Three species {Sigmodon hispidus, Peromyscus leucopus, Peromyscus maniculatus) were handled during the first sampling session of the first year of the project (i.e. , during the first 150 trapnights). The first specimens of the fourth {Reithrodontornys fulvescens) and fifth species {Reithrodontornys humulis) were obtained during the first sampling session of the second year (i.e., after 650 trapnights). 250 THE TEXAS JOURNAL OF SCIENCE— VOL. 47, NO. 4, 1995 Population dynamics of wooded-draw grid.— Sampling (1,080 trap- nights) indicated the presence of four species of rodents in the wooded- draw habitat. One of these species, Sigmodon hispidus, was represented by only one individual (density = 3.3/ha), a subadult caught one time near the ecotone of woods and grassland. Twenty different wood rats (Neotoma floridana) were trapped on this grid. Two other species were present in similar numbers: Peromyscus leucopus (n=52 individuals) and Peromyscus rnaniculatus {n=46). Densities of Peromyscus leucopus ranged from 3. 3 /ha to 40.0/ha (Table 1); those of Peromyscus rnaniculatus varied from 0/ha during four sessions to 33.3/ha (Table 1). A general trend evident for Peromyscus rnaniculatus is lower densities in summer and fall, and higher densities in winter and spring. No Neotorna floridana were caught during the winter of year two or spring of year three, but they were present at densities of 3. 3 /ha to 20.0/ha during the rest of the study. For each year, their peak densities were during summer (Table 1). Fifteen of the 20 Neotorna floridana handled were never recaptured. Three of the remaining five were recaptured only within the session during which they were marked. The initial capture and recapture of one male spanned three seasons for a minimum longevity of six months. Another male was handled during spring of years one and two. Mean minimum longevity for Neotorna floridana as based on these five animals was 3.6 months. Of the 52 Peromyscus leucopus captured, 21 (40.4%) were recaptured at least once. Four were recaptured only during their session of initial capture. The last record for seven more individuals was in the session immediately after session of first capture. Four others were captured two sessions after their original capture. Three were last captured in the third session following initial capture. Longevity in three others spanned four, six, and eight trapping sessions, respectively. Mean minimal longevity as based on these 52 mice was six months, more than three times the 1.7 month longevity determined by Waggoner (1975). On this grid, mean minimal longevity for Peromyscus rnaniculatus was 2. 1 months. Over one- third (37 % , /2 = 17) of the deer mice handled were recaptured one or more times. One male survived at least one year, being captured in consecutive winters. The trapping records for two other Peromyscus rnaniculatus spanned six months. Four others were last captured during the session immediately following the session of initial capture. Ten of the 17 recaptured individuals were recaptured only during the session of first capture. WILKINS A: Grassland grid 251 Figure 1. Seasonal trends in cover provided by 15 categories of vegetation from grassland (A) and wooded-draw (B) grids at Clymer Meadow, Hunt Co., Texas, 1989-1992. The vegetation categories are described in Methods and Materials, The three species (Perornyscus leucopus, Perotnyscus maniculatus, Neotoma floridana) constituting the vast majority of the rodent community of the wooded-draw habitat were initially sampled during the first sampling session of the first year of the study (i.e., within 90 trapnights). A fourth species {Sigmodon hispidus) was first trapped in this habitat after approximately 990 trapnights of effort, during the third session of the third year of the study. Vegetation analyses of grassland Seasonal summaries of cover provided by 15 categories of vegetation are presented in Figure 1 and Table 2. The predominant grasses of the Blackland Prairie at Clymer Meadow are little bluestem and indiangrass. Indeed, these species in combination with the less-abundant switchgrass and eastern gamagrass offered the vast majority of vegetative cover on the grassland grid. Big Woody Im Table 2. Summary statistics for plant cover on the grassland and wooded-draw sampling grids at Clymer Meadow, Hunt Co., Texas, 1989-1992. Listed are the means and standard deviations (located below the mean) of percentage cover provided by each of 15 categories of vegetation averaged by season for the 3-year study period. Each mean was computed from 30 sampling quadrats each with an area of 0.1 m^. 252 THE TEXAS JOURNAL OF SCIENCE-VOL. 47, NO. 4, 1995 ^ VO O' Im) during winter when deciduous species had dropped their leaves to maxima for cover at 50 cm during spring (21.5%) and for cover at >lm during summer (70.8% ; Table 2). During any season for any year, no category of dead herbaceous material contributed more than 14% cover. Litter in the woods most often was composed of fallen leaves of shrubs and trees rather than of stems and leaves of grasses as in the grassland. Big bluestem, little bluestem, and switchgrass were absent from the wooded draw. Most of the grassy cover was provided by grasses in the "other grasses" category; the dominant grass was river oats 254 THE TEXAS JOURNAL OF SCIENCE-VOL. 47, NO. 4, 1995 {Chasmanthium latifolium) , occurring in areas devoid or nearly so of woody cover. As in the grassland grid, forbs were insignificant contributors to total vegetative cover. Maximum coverage by any category of forbs during any year of the study was 11.2% at 5 cm (during spring of year three) and 1 1.0% at 10 cm (during spring of year one). Herbaceous ground cover varied from virtually absent beneath a dense canopy to rather dense in light gaps. Vegetation comparison of grassland and wooded-draw grids. — Cover on the two grids was provided by different groups of plants, with grasses and litter derived therefrom dominating on the grassland grid, and woody shrubs and trees dominating for the wooded-draw grid (Fig. 1). Analyses of variance were conducted to compare grids on the basis of amount of cover provided by each vegetation category for each season. For all four categories of dead herbaceous vegetation, the grids differed significantly (P<0.054; d.f =59). Even more pronounced were the differences in cover provided by woody vegetation (F< 0.0001 ; d.f. =59), Big bluestem and switchgrass, both absent from the wooded draw, provided only minor amounts of cover in the grassland. The amount of cover provided by “other grasses” was significantly greater in the grassland at 25 cm and 50 cm during fall (P< 0.05 15; d.f =59) and at 25 cm during summer (P< 0.0001; d.f =59); during other seasons at both 25 cm and 50 cm, amount of cover provided by “other grasses” was similar. Values for cover provided by the various categories of forbs for both grids hovered at or below 10% for all seasons. Forbs at 25 cm offered significantly greater amounts of cover in the grassland (8.3% vs. 0.4% in wooded draw) during summer (P< 0.008; d.f. =59). During fall, cover from forbs at 10 cm and 25 cm was significantly greater in the grassland (P< 0.0030; d.f. =59); probably these values are not biologically meaningful as the range in these cover values was 0.1% to 3. 1 %. Care must be taken in compar¬ ing the forbs and the “other grasses” categories between grids because these categories comprised different species in the two habitats being compared. Additional small mammal sampling.— Etyond sampling on the trap¬ ping grids, 1,653 additional trapnights of effort were invested in other areas of the preserve having different land uses (Table 3). During the 3 -year study, 167 rodents of five species were livetrapped away from the grids for a trapping efficiency of 10.1%. Sampling in the disturbed (primarily by haying) native grassland, the habitat most like the native grassland grid, yielded five species: Sigmodon hispidus, Peromyscus WILKINS 255 Table 3. Summary of numbers of different individuals of each rodent species caught in six habitats at Clymer Meadow, Hunt County, Texas, 1989-1992. Habitat Undisturbed Hayed Bermuda- Undisturbed Cedar elm- Wooded grassland native grass wooded-draw ash hilltops Species (grid) grassland pasture (grid) woodland and mottes Sigmodon hispidus 153 94 3 1 — 1 Peromyscus leucopus 7 23 5 52 3 1 Peromyscus maniculatus 1 15 4 46 — — Neotoma floridana — — — 20 — 2 Reithrodontomys fulvescens 32 14 — — — Reithrodontomys humulis 1 — — — — — Mus musculus — 3 — — — — Total rodents 206 149 12 119 3 4 Total traps set 1,800 1,344 150 1,080 90 69 Trapping efficiency 11.4% 11.1% 8.0% 11.1% 3.3% 5.8% leucopus, Peromyscus maniculatus, Reithrodontomys fulvescens, and Mus mus cuius. Presence of Mus mus cuius and absence of Reithrodontomys humulis distinguished the disturbed grassland from the undisturbed grassland. A pasture dominated by coastal bermuda grass {Cynodon dactylon) and being lightly grazed by cattle during our trapping had a less-diverse rodent fauna (only Sigmodon hispidus, Peromyscus leucopus, and Peromyscus maniculatus) than native grassland. Trapping in wooded hilltop mottes produced a different set of three species {Sigmodon hispidus, Peromyscus leucopus and Neotoma floridana). The habitat with the most-depauperate rodent community was a low-lying woodland dominated by cedar elm and ash (Fraxinus); only Peromyscus leucopus was trapped there. Another factor possibly germane to differences in the rodent communities in these habitats is the presence or absence of imported fire ants. These ants were absent from the tract containing the two grid sites during the entire study, but were present in all hayed, grazed and other disturbed tracts. Discussion Habitat associations .—SdxrvpXmg revealed the presence of six native species of ground-dwelling rodents {Sigmodon hispidus, Reithrodontomys fulvescens, Reithrodontomys humulis, Peromyscus leucopus, Peromyscus maniculatus, Neotoma floridana) and one exotic species {Mus musculus) in the various prairie habitats. The microhabitat selectivities of these species at Clymer Meadow generally corresponded with those reported for these species elsewhere in their ranges. 256 THE TEXAS JOURNAL OF SCIENCE-VOL. 47, NO. 4, 1995 Though Sigmodon hispidus may be found in a variety of habitats, their greatest densities obtain in monocot-dominated situations where dense stands of grasses with a thick mulch layer and intermixed with forbs offer suitable cover (Kaufman & Fleharty 1974; Kincaid & Cameron 1985; McMurry et al. 1994). Cotton rats use less-favored habitats when population densities are high (Fleharty & Mares 1973); this tendency probably explains the isolated occurrence of a subadult cotton rat in the wooded draw of Clymer Meadow. Peromyscus leucopus and Peromyscus maniculatus occurred in the grassland and wooded habitats at Clymer, with Peromyscus leucopus predominating in the woodland and achieving much lower densities in the grassland where there were similar densities of Peromyscus maniculatus. On Konza Prairie, white- footed mice achieved dramatically greater densities in wooded habitats than in habitats lacking trees (McMillan & Kaufman 1994), a trend also documented by Cummings & Vessey (1994). Peromyscus maniculatus in the tallgrass prairies of Kansas avoids sites with well-developed litter layers, selectively foraging instead in patches with sparse litter ( < 1 cm high— Clark and Kaufman 1991). The low densities of deer mice in the Clymer grassland may be explained by the presence of thick, dense litter. Despite its widely documented preference for grassy areas (Kaufman & Fleharty 1974; Schmidly 1983; Schwartz et al. 1994), Peromyscus maniculatus also was caught frequently in the wooded draws of Clymer Meadow, where its greater-than-expected presence perhaps was related to microhabitat patchiness (McMillan & Kaufman 1994). This pattern probably was effected by the narrowness of the belts of woodland penetrating the more-extensive grasslands of the preserve. In eastern Texas, Reithrodontomys fiilvescens occurs in a variety of habitats, all of which include a significant component of grasses (Schmidly 1983). At Clymer Meadow, fulvous harvest mice were trapped only in undisturbed native grassland and in hayed native grass¬ land, two situations in which little bluestem and switchgrass were the dominant grasses. Also caught in this undisturbed native grassland at Clymer Meadow was R. humulis, a species that appears to be uncom¬ mon in all eastern Texas localities from which it is known (Schmidly 1983; Wilkins 1991). There is evidence, however, that eastern harvest mice might be more abundant than most reports indicate (Dunaway 1968); Cawthorn & Rose (1989), using a modified trapping method that excluded larger species of rodents, estimated densities averaging about 20/ha, with peaks as high as 44.4/ha, in old fields in southeastern Virginia. WILKINS 257 The habitats in which Neotoma floridana occurred at Clymer Meadow approximate those described for eastern woodrats in Kansas, where they preferred living in tangled underbrush in low-lying areas along timbered streams (Kellogg 1915). With the subsequent destruction of this habitat in Kansas, woodrats shifted to hedgerows where the dominant tree was osage orange (bois d’arc), and coralberry constituted a significant proportion of the understory (Rainey 1956). Wooded habitats, both in drainages and on hilltops, throughout the Clymer preserve boast numerous stick houses built by woodrats; preferred locations seem to be at the bases of multistemmed bois d’arc trees. Biogeographic considerations .—Oi the seven species of rodents sampled in the prairie habitats, one {Reithrodontomys humulis) was previously unknown from Hunt County; this westward extension of geographic range has been reported elsewhere (Wilkins 1991). Three additional native species of rodents, Taylor’s pygmy mouse {Baiomys taylori), hispid pocket mouse (Chaetodipus [Perognathus] hispidus), and thirteen-lined ground squirrel (Spermophilus tridecemlineatus) , were anticipated but not found at Clymer preserve despite extensive sampling during an array of weather conditions during all seasons over 3 years. Baiomys taylori, a euryoecious species that frequently inhabits grassy areas, is known to the west in nearby Dallas and Denton counties (Schmidly 1983). Only in the 1960’s did this species enter north-central Texas as part of a continuing northward expansion from subtropical south Texas (Schmidly 1983; Pitts & Smolen 1989). Possibly, Baiomys taylori simply has not yet reached Clymer Meadow; regular monitoring might reveal its entry into the area in the next decade or two. Having a well-documented statewide range (Davis & Schmidly 1994), Chaetodipus hispidus almost certainly is present at or in the vicinity of Clymer Meadow. My observations from work in eastern and central Texas are in agreement with McCarley’s (1959) report that C. hispidus is not abundant in habitats with clayey soils (e.g., blackland prairies), but is more common in habitats with well-drained, sandy soils (e.g., sandy alluvium along Brazos River in McLennan County, pers. obs.; Wilkins 1977; Wilkins & Schmidly 1980). Though Bailey’s distribution map (1905: plate XIII) for S. tridecemlineatus included the western half of Hunt County, no evidence of this species was found at Clymer Meadow. Continued sampling might yet reveal these species on this preserve. The assemblage of ground-dwelling rodent species at Clymer preserve apparently is unique, different by at least one or two species from faunas 258 THE TEXAS JOURNAL OF SCIENCE-VOL. 47, NO. 4, 1995 described in the few other studies of small-mammal demography in various habitats in eastern Texas. These studies include an inventory of mammalian species in the Big Thicket region (Schmidly et al. 1979, 1980); assessment of effects of mowing of highway rights-of-way vegetation on rodent species in Brazos and Madison counties (Wilkins 1977; Wilkins & Schmidly 1980); comparison of rodent communities on natural, reclaimed, and strip-mined habitats near Fairfield, Freestone County (Waggoner 1975); and studies of rodent population dynamics in coastal prairies near Houston (Joule & Cameron 1974; Cameron 1977; Cameron et al. 1979), and in old fields in the postoak savanna of Brazos County (Grant et al. 1985). Little other work on the population dynamics of small mammals on native tallgrass Blackland Prairies in Texas has been conducted, and apparently none has been published. However, one such project was conducted in Falls County, north of Marlin, on the Dorothea Leonhardt Preserve, owned by the Nature Conservancy of Texas. The Leonhardt Prairie, a tallgrass prairie dominated by little bluestem and switchgrass, is located approximately 270 km SSW of Clymer Meadow. R. S. Baldridge and other faculty members and students of the Department of Biology, Baylor University (unpublished data), conducted mark-release- recapture studies of small mammals from 1984 until about 1987. They found Sigmodon hispidus and Baiomys taylori to be co-dominant species of the Leonhardt Prairie rodent community. Other ground-dwelling rodents at this site were Peromyscus leucopus, Peromyscus maniculatus, Reithrodontomys fulvescens, Chaetodipus hispidus, Neotoma floridana, and Mus mus cuius. Imported fire ants invaded the Leonhardt preserve near the end of the study. Ackno wl edgments The efforts of many people contributed to the successful completion of this project. Jeff Weigel, of the Nature Conservancy of Texas, was involved significantly and enthusiastically in all administrative phases of the project. Numerous students from my classes at Baylor University provided much of the personpower needed in the field segments of the project; most notable among these were graduate students Russ Fraze and Rick Wiedenmann. W. Keith Hartberg, chairman of the Baylor Biology Department, facilitated this research by allowing use of departmental vehicles and other equipment. Prof. Walter Holmes, Baylor University, graciously identified several species of plants found on the preserve. Several Nature Conservancy interns, including Ben WILKINS 259 Bernard, Scott Grimm, and Jamie Ingold, supplied invaluable field assistance. The Clymer houses were welcome refuges from the heat, cold, rain, insects, and other natural elements necessarily endured by field workers; the generosity of the Nature Conservancy of Texas in allowing us to use these facilities is greatly appreciated. The Small Grants program of the Nature Conservancy of Texas funded this research. Literature Cited American Society of Mammal ogists. 1987. Acceptable field methods in mammalogy: preliminary guidelines approved by the American Society of Mammal ogists. Supplement, J. Mammal., 68, 18 pp. Bailey, V. 1905. Biological survey of Texas. North Am. Fauna, 25:1- 222. Blair, W. F. 1941. Techniques for the study of mammal populations. J. Mammal., 22:148-157. Blair, W. F. 1942. Size of home range and notes on the life history of the woodland deer-mouse and eastern chipmunk in northern Michigan. J. Mammal., 23:27-36. Blair, W. F. 1950. The biotic provinces of Texas. Texas J. Sci., 2:93-117. Cameron, G. N. 1977. Experimental species removal: demographic responses by Sigmodon hispidus and Reithrodontornys fulvescens. J. Mammal., 58:488-506. Cameron, G. N., W. B. Kincaid, C. A. Way, & J. O. Woodrow, Jr. 1979. Daily movement patterns of Sigmodon hispidus. Southwestern Nat., 24:63-70. Cawthorn, J. M. & R. K. Rose. 1989. The population ecology of the eastern harvest mouse {Reithrodontornys humulis) in southeastern Virginia. Amer. Midi. Nat., 122:1-10. Clark, B. K. & D. W. Kaufman. 1991. Effects of plant litter on foraging and nesting behavior of prairie rodents. J. Mammal., 72:502-512, Cummings, J. R. & S. H. Vessey. 1994. Agricultural influences on movement patterns of white- footed mice (Peromyscus leucopus). Amer. Midi. Nat., 132:209-218. Daubenmire, R. 1959. A canopy coverage method of vegetational analysis. Northwest Sci., 33:43-64. Davis, W. B. & D. J. Schmidly. 1994. The mammals of Texas. Texas Parks and Wildlife Dept., Austin, x + 338 pp. Diamond, D. D. & F. E. Smeins. 1985. Composition, classification 260 THE TEXAS JOURNAL OF SCIENCE— VOL. 47, NO. 4, 1995 and species response patterns of remnant tallgrass prairies in Texas. Am. Midi. Nat., 112:294-308. Diamond, D. D. & F. E. Smeins. 1993. The native plant communities of the Blackland Prairie, Pp. 66-81, in The Texas Blackland Prairie: Land, History, and Culture (M. R. Sharpless &J. C. Yelderman, Jr., eds.), Baylor Univ. Program for Regional Studies, Waco, Texas, xvi + 369 pp. Dunaway, P. B. 1968. Life history and population aspects of the eastern harvest mouse. Amer. Midi. Nat., 79:48-67. Fleharty, E. D. & M. A. Mares. 1973. Habitat preferences and spatial relations of Sigmodon hispidus on a remnant prairie in west-central Kansas. Southwestern Nat., 18:21-29. Gehlbach, F. R. 1991. The east- west transition zone of terrestrial vertebrates in central Texas: a biogeographical analysis. Texas J. Sci., 43:415-427. Gould, F. W. 1975. The grasses of Texas. Texas A&M Univ. Press, College Station, viii + 653 pp. Grant, W. E., P. E. Carothers, & L. A. Gidley. 1985. Small mammal community structure in the postoak savanna of east-central Texas. J. Mammal., 66:589-594. Hayne, D. W. 1949. Two methods for estimating population from trapping records. J. Mammal., 30:399-411. Joule, J. & G. N. Cameron. 1974. Field estimation of demographic parameters: influence of Sigmodon hispidus population structure. J. Mammal., 55:309-318. Kaufman, D. W. & E. D. Fleharty. 1974. Habitat selection by nine species of rodents in north-central Kansas. Southwestern Nat., 18:443-452. Kellogg, R. 1915. The mammals of Kansas with notes on their distribution, life histories and economic importance. Master’s thesis, Univ. Kansas, Lawrence. Kincaid, W. B. & G. N. Cameron. 1985. Interactions of cotton rats with a patchy environment: dietary responses and habitat selection. Ecology, 66:1769-1783. McCarley, H. S. 1959. The mammals of eastern Texas. Texas J. Sci., 11:385-426. McMillan, B. R. & D. W. Kaufman. 1994. Differences in use of interspersed woodland and grassland by small mammals in northeastern Kansas. Prairie Nat., 26:107-116. McMurry, S. T., R. L. Lochmiller, J. F. Boggs, D. M. Leslie, Jr., & D. M. Engle. 1994. Demographic profiles of populations of cotton WILKINS 261 rats in a continuum of habitat types. J. Mammal., 75:50-59. Overton, W. S. 1972. Estimating numbers of animals in wildlife populations, Pp. 403-456, in Wildlife management techniques (R. H. Giles, ed.). Wildlife Society, Washington, D.C., vii -f 633 pp. Pitts, R. M. & M. J. Smolen. 1989. Status of Baiomys taylori in Texas, with new localities of record in the southern part of the state. Texas J. Sci., 41:85-88. Rainey, D. G. 1956. Eastern woodrat, Neotoma floridana: life history and ecology. Univ. Kansas Publ., Mus. Nat. Hist., 8:535-646. SAS Institute Inc. 1985. SAS user’s guide: statistics. Fifth ed. SAS Institute Inc., Cary, North Carolina, 958 pp. Schmidly, D. J. 1983. Texas mammals east of the Balcones fault zone. Texas A&M Univ. Press, College Station, xviii + 400 pp. Schmidly, D. J., B. R. Barnett & J. A. Reed. 1979. The mammals of Big Thicket National Preserve and east Texas. Report prepared for the Office of Natural Resources, Southwest Region, National Park Service, Santa Fe, New Mexico, Contract no. CX700050442, 345 pp. Schmidly, D. J., W. G. Norton & G. A. Barber. 1980. The game and furbearing mammals of Big Thicket National Preserve with comments on the small mammal fauna of selected units. Report prepared for the Office of Natural Resources, Southwest Region, National Park Service, Santa Fe, New Mexico, Contract no. CX70290019, 141 pp. Schmidly, D. J., D. L. Scarbrough & M. A. Horner. 1993. Wildlife diversity in the Blackland Prairies, Pp. 82-95 in The Texas Blackland Prairie: Land, History, and Culture (M. R. Sharpless & J. C. Yelderman, Jr., eds.), Baylor Univ. Program for Regional Studies, Waco, Texas, xvi + 369 pp. Schmidly, D. J., K. T. Wilkins, R. L. Honeycutt & B. C. Weynand. 1977. The bats of east Texas. Texas J. Sci., 28:127-143. Schwartz, O. A., A. M. Vivas, A. Orris & C. J. Miller. 1994. Small mammals species associations in three types of roadside habitats in Iowa. Prairie Nat., 26:45-52. Smeins, F. E. & D. D. Diamond. 1983. Remnant grasslands of the Fayette Prairie, Texas. Amer. Midi. Nat., 110:1-13. Strecker, J. K. 1924. The mammals of McLennan County, Texas. Baylor Bull., 27:3-20. Strecker, J. K.. 1926. The mammals of McLennan County, Texas: second paper, supplementary notes. Contrib. Baylor Univ. Mus., 9:1-15. Tharp, B. C. 1939. The vegetation of Texas. Texas Acad. Sci., Publ. Nat. Hist., Non-Tech. Ser., 1:1-74. 262 THE TEXAS JOURNAL OF SCIENCE- VOL. 47, NO. 4, 1995 Waggoner, K. V. 1975. The effects of strip-mining and reclamation on small mammal communities. M.S. thesis, Texas A&M Univ., College Station, 79 pp. Wilkins, K. T. 1977. The effects of highways on small mammals and other wildlife. M.S. thesis, Texas A&M Univ., College Station, 103 pp. Wilkins, K. T. 1991. Distributional records and ecological notes for two species of rodents in Texas. Texas J. Sci., 43:337-339. Wilkins, K. T. & D. J. Schmidly. 1980. Highway mortality of vertebrates in southeastern Texas. Texas J. Sci., 32:343-350. Wilkins, K. T. & D. J. Schmidly. 1981. The effects of mowing of highway rights-of-way on small mammals, Pp. 55-1—55-13, in Proc. 2nd Natl. Symp., Environmental Concerns in Rights-of-Way Management (R. E. Tillman, ed.). Electric Power Reserach Institute, Project WS 78-141. TEXAS J. SCI. 47(4):263-268 NOVEMBER, 1995 NOTES ON HARVEST MICE {REITHRODONTOMYS) OF THE BIG BEND REGION OF TEXAS Franklin D. Yancey, II, Clyde Jones and Jim R. Goetze Department of Biological Sciences and the Museum of Texas Tech University, Lubbock, Texas 79409 Abstract. — Extensive live-trapping for small mammals on the Big Bend Ranch State Park, Texas, yielded several specimens of Reithrodontomys fulvescens and Reithrodontomys megalotis from a variety of habitat types. On two occasions, these two species of harvest mice were taken in sympatry. The ecological relationships and reproductive biology of these two species of harvest mice are reviewed. In addition, the distributions of R. fulvescens, R. megalotis, as well as their congener, Reithrodontomys montanus, in the Big Bend region of Texas are discussed. The presence of three species of harvest mice (Reithrodontomys fulvescens, R. megalotis and R, montanus) in the Big Bend region of Texas has been previously documented (Schmidly 1977; Hall 1981; Jones & Jones 1992; Davis & Schmidly 1994). However, information is not available with regard to the details of geographic distributions and ecological relationships of these rodents in the area. In addition, little is known about the life histories of these species in the Trans-Pecos area of Texas (Schmidly 1977). This report is a summary of information obtained recently on the harvest mice of the Big Bend region of Trans- Pecos Texas. Methods and Materials Between January 1994 and April 1995, small mammals were sampled on the Big Bend Ranch State Park, Texas. This area was formerly known as the Big Bend Ranch State Natural Area. This area encompasses approximately 1087 square kilometers of Presidio and Brewster counties. Sherman live traps were set in the four major habitat types that exist in the area; Chihuahuan Desert scrub, Chihuahuan Desert grassland. Juniper roughland, and riparian woodland. During the course of this study, 7,439 traps were set. Traplines consisted of 40-50 traps baited with rolled oats, and set at 10 meter intervals. They were set approximately one hour before sundown and retrieved approximately one hour after sunrise the following morning. Animals acquired were identified and voucher specimens (standard museum skin and skull) were prepared. From selected specimens, tissues (muscle, liver, heart, and kidney) were collected and immediately frozen in liquid nitrogen. 264 THE TEXAS JOURNAL OF SCIENCE— VOL. 47, NO. 4, 1995 In the following accounts, all measurements provided are in milli¬ meters. All localities are based on Universe Transverse Mercator (UTM) coordinates taken from a hand held global positioning system. As a reference point for these UTM coordinates, Big Bend Ranch State Park Headquarters (Sauceda Ranch), which has UTM coordinates 13 601203E, 3260 197N, is situated approximately 9 km S, 41 km E of Presidio. Voucher materials (TTU numbers) and frozen tissues (TK numbers) are deposited in the Collection of Recent Mammals in the Natural Science Research Laboratory of the Museum of Texas Tech University. Results and Discussion Reithrodontornys fulvescens (fulvous harvest mouse). — Although this species has an extensive geographic range in North and Central America (Hall 1981; Spencer & Cameron 1982), it occurs mostly east of the 100th meridian in Texas, except for parts of the Panhandle and the Trans-Pecos region (Schmidly 1977; Jones & Jones 1992; Davis & Schmidly 1994). This species was recorded previously from single localities in Presidio and Reeves counties, from four localities in Brewster County, and from five localities in Jeff Davis County, Texas (Schmidly 1977). During this study, seven specimens of R. fulvescens were obtained from three locations in Presidio County, and a single specimen from one site in western Brewster County, all within the boundaries of the Big Bend Ranch State Park. According to Schmidly (1977), in the Trans-Pecos, R. fulvescens favors rough grasslands. However, all seven individuals acquired from Presidio County during this study were taken in riparian habitat. Each was captured along small, permanent streams where cottonwood {Populus sp.), willow (Salix sp.), false willow (Baccharis sp.), and various grasses were the dominant plant species. Interestingly, R. megalotis was taken along with R. fulvescens at two of the three localities in Presidio County. Such close ecological associations of these two species of harvest mice were not reported previously for the Big Bend region of West Texas (Schmidly 1977), Chihuahua, Mexico (Anderson 1972), or Coahuila, Mexico (Baker 1956). Other rodents associated with these riparian areas included the white-footed mouse (Perotnyscus leucopus), the deer mouse (P. maniculatus) , the white- ankled mouse (P. pectoralis), and the hispid cotton rat (Sigmodon hispidus) . YANCEY, JONES & GOETZE 265 The single specimen of R. Julvescens taken from western Brewster County was captured in a dry, shallow arroyo in the basin of the unique geological formation known as the Solitario. The habitat at this site was Chihuahuan Desert scrub dominated by mesquite (Prosopis sp.), catclaw {Acacia sp.), and medium grass. Other small mammals taken in conjunction with this specimen included Nelson’s pocket mouse {Chaetodipus nelsoni), the desert pocket mouse (C. penicillatus) , and Merriam’s kangaroo rat {Dipodomys merriami). In Texas in general, fulvous harvest mice are known to breed from February to October (Davis & Schmidly 1994). However, a male in reproductive condition (testes 10 by 5 mm) was captured on 10 November, and a gravid female (3 embryos with crown-rump length of 14 mm) was taken on 17 November. Given a gestation period of about 21 days (Davis & Schmidly 1994), these data suggest that the breeding season in the Big Bend region may extend well into November. Non- gravid females were obtained on 10 January and 13 and 14 February. In addition, three non-reproductively active males (testes 4 by 2 mm) were taken on 13 February. Material examined.— County: Big Bend Ranch State Park, UTM coordinates 13 617281E, 3261484N, one specimen (TTU 67413). Presidio County: Big Bend Ranch State Park, UTM coordinates 13 587250E 32629 18N, five specimens (TTU 67407, TK 41751; TTU 67408-67410; TTU 67414, TK 46437); Big Bend Ranch State Park, UTM coordinates 13 601619E 3260741N, one specimen (TTU 67411); Big Bend Ranch State Park, UTM coordinates 13 586937E 3262923 N, one specimen (TTU 67412). Reithrodontomys megalotis (western harvest mouse).— This species has a large geographic range in North America (Hall 1981; Webster & Jones 1982), but in Texas is known only from the western Panhandle, the northwestern Edwards Plateau, and the Trans-Pecos region (Schmidly 1977; Jones & Jones 1992; Davis & Schmidly 1994; Goetze et al. 1995). This mouse was considered uncommon in Coahuila, Mexico (Baker 1956), but was reported as widespread both geographically and ecologically in Chihuahua, Mexico (Anderson 1972). According to Schmidly (1977), "This is the most widely distributed harvest mouse in the Trans-Pecos..." In addition to four localities reported previously from Presidio County (Schmidly 1977), 10 specimens of the western harvest mouse were collected during this study from seven localities within Big Bend Ranch State Park, Presidio County. 266 THE TEXAS JOURNAL OF SCIENCE— VOL. 47, NO. 4, 1995 In the Trans- Pecos, R. megalotis is reported to prefer grasslands with dense ground cover (Schmidly 1977). However, of the seven sites from which collections of this mouse were made, only two fit this description. At one of these sites, two western harvest mice were taken in tall grass associated with yucca {Yucca sp.) and mesquite. Other mammals taken from this site included the deer mouse and Mearn’s grasshopper mouse {Onychomys arenicola). At the other grassland site, a single /?. megalotis was taken in heavily grazed short grass with thick desert- willow (Chilopsis sp.). Associated small mammals included the deer mouse and the white- footed mouse. As mentioned previously, western harvest mice also were found at two riparian localities. See the above account on R. fulvescens for specifics. The three remaining sites from which specimens of western harvest mice were acquired are best described as desert scrub. Associated vegetation included creosotebush {Larrea sp.), mesquite, mariola (Parthenium sp.), catclaw, and various grasses. The desert pocket mouse, Merriam’s kangaroo rat, the cactus mouse {Peromyscus eremicus), the deer mouse, and the southern plains woodrat {Neotoma micropus) also were taken at these localities. Western harvest mice are considered to breed throughout the year (Schmidly 1977). Reproductive data regarding specimens from this study are as follows. A male collected in January had testes measuring 5 by 1 mm. Specimens obtained in February included five males with testes measuring 3 by 2 mm (two specimens), 6 by 4 mm, 5 by 3 mm, and 2 by 2 mm, as well as two non-gravid females. A female bearing three embryos (crown- rump length 8 mm) was taken in April. One non- gravid female was collected in November. Material examined.— County: Big Bend Ranch State Park, UTM coordinates 13 600694E, 325975 IN, one specimen (TTU 67415); Big Bend Ranch State Park, UTM coordinates 13 587250E, 32629 18N, two specimens (TTU 67416-67417, TK 41752); Big Bend Ranch State Park, UTM coordinates 13 601619E, 326074 IN, two specimens (TTU 67418-67419); Big Bend Ranch State Park, UTM coordinates 13 576836E, 329625 IN, one specimen (TTU 67420, TK 41778); Big Bend Ranch State Park, UTM coordinates 13 576970E, 3296222N, one specimen (TTU 67421); Big Bend Ranch State Park, UTM coordinates 13 601973E, 3253265N, one specimen (TTU 67605, TK 46450); Big Bend Ranch State Park, UTM coordinates 13 605746E, 3261330N, two specimens (TTU 67606, TK 46458; TTU 67607, TK 46460). YANCEY, JONES & GOETZE 267 Reithrodontomys montanus (plains harvest mouse). — This species is distributed widely in central and southwestern North America (Hall 1981; Wilkins 1986), and it ranges throughout much of Texas (Jones & Jones 1992; Davis & Schmidly 1994). However, Schmidly (1977) considered it to be the rarest of the species of Reithrodontomys occurring in the Trans- Pecos region. Other than the records of R. montanus summarized by Jones et al. (1993), no recent information is known on this species from the Big Bend region of Texas. Reithrodontomys montanus was reported from five localities in Chihuahua, Mexico, the southern limit of the geographic range of the species (Anderson 1972). The three species of harvest mice {R. fulvescens, R. rnegalotis and R, montanus) in the Big Bend Region of Texas should be subjected to thorough review with regard to systematic affinities. In addition, given the overlaps of the geographic ranges of R, fulvescens and R. rnegalotis in restricted habitats on the Big Bend Ranch State Park, it would appear that unique opportunities exist there for detailed studies of the ecological relationships of these two sympatric species. Acknowledgments Mammals were collected on the Big Bend Ranch State Park in accordance with scientific collecting permits issued by the Texas Parks and Wildlife Department (permit numbers SPR-0790-189, 4-94, 25-95). Financial assistance was provided by the Natural Resources Program (David H. Riskind, Director) of the Texas Parks and Wildlife Department. Logistic support was provided by personnel of the Big Bend Ranch State Park (Luis Armendariz, Superintendent). Assistance in the field was provided by Mary Ann Abbey, Jennifer Brice, Charlene Cunningham, Craig Jones, Kinzea Grace Jones, Maryann Lynch, and Richard Manning. We thank James G. Dickson and an anonymous reviewer for their comments on this manuscript. Literature Cited Anderson, S. 1972. Mammals of Chihuahua taxonomy and distribu¬ tion. Bull. American Mus. Nat. Hist., 148:149-410. Baker, R. H. 1956. Mammals of Coahuila, Mexico. Univ. Kansas Pubis., Mus. Nat. Hist., 9:125-335. Davis, W. B., & D. J. Schmidly. 1994. The mammals of Texas. Texas Parks and Wildlife Dept., Austin, x + 338 pp. Goetze, J. R., F. D. Yancey, II, C. Jones, & B. M. Gharaibeh. 1995. Noteworthy records of mammals from the Edwards Plateau of central 268 THE TEXAS JOURNAL OF SCIENCE— VOL. 47, NO. 4, 1995 Texas. Texas J. Sci., 47(l):3-8. Hall, E. R. 1981. The mammals of North America. John Wiley & Sons, New York, 2nd ed., l:xv + 1-600 + 90 and 2:vi + 601-1181 + 90. Jones, J. K., Jr., & C. Jones. 1992. Revised checklist of Recent land mammals of Texas, with annotations. Texas J. Sci., 44(l):53-74. Jones, J. K., Jr., R. W. Manning, F. D. Yancey, II, & C. Jones. 1993. Records of five species of small mammals from western Texas. Texas J. Sci., 45(1): 104-105. Schmidly, D. J. 1977. The mammals of Trans-Pecos Texas including Big Bend National Park and Guadalupe Mountains National Park. Texas A&M Univ. Press, College Station, xiii + 225 pp. Spencer, S. R., & G. N. Cameron. 1982 Reithrodontomys fulvescens. Mammalian Species, 174:1-7. Webster, Wm. D., & J. K. Jones, Jr. 1982. Reithrodontomys megalotis. Mammalian Species, 167:1-5. Wilkins, K. T. 1986. Reithrodontomys montanus. Mammalian Species, 257: 1-5. TEXAS J. SCI. 47(4):269-276 NOVEMBER, 1995 KARYOTYPES OF SEVEN SPECIES OF NORTH AMERICAN WRENS (PASSERIFORMES: TROGLODYTIDAE) *Shamone M. Minzenmayer, Terry C. Maxwell and Robert C. Dowler Department of Biology, Angelo State University, San Angelo, Texas 76909 ^Present address: Department of Wildlife and Fisheries Sciences, Texas A&M University, College Station, Texas 77843 Abstract.— Karyotypes, new to cytology, are described for seven North American species of wrens (Troglodytidae); Campylorhynchus brunneicapillus (2n = 74), Salpinctes obsoletus (2N = 80), Catherpes mexicanus (2N = 80), Thryothorus ludovicianus (2n = 76), Thryomanes bewickii (2N = 76), Cistothorus platensis (2N = 76), and Cistothorus palustris (2N = 76). All species examined share the first three macrochromosomes. Compared to other vertebrate classes (e.g. mammals and reptiles), avian karyology is poorly known. Less than 10% of the world’s species of birds have been karyotyped, of which very few have been passerines (Takagi & Sasak 1974; Tegelstrom & Ryttman 1981; Shields 1983; 1985; De Boer 1984). The family Troglodytidae includes some 45 species in 15 different genera (Morony et al. 1975; Sibley & Monroe 1990). Of these, only two species have been karyotyped. Udagawa (1956) described the chromosomes of the Nearctic species. Troglodytes troglodytes (2n = 86) and De Lucca & Chamma (1977) reported those of the neotropical wren, Thryothorus leucotis (2n=78). Nine species of wrens occur in North America north of Mexico (AOU 1983). This study provides the first description of the chromosomes of seven of these species: Cactus Wren {Campylorhynchus brunneicaplT lus), Rock Wren {Salpinctes obsoletus), Canyon Wren {Catherpes mexicanus), Carolina Wren {Thryothorus ludovicianus), Bewick’s Wren {Thryomanes bewickii). Sedge Wren {Cistothorus platensis), and Marsh Wren {Cistothorus palustris). Nomenclature follows that of the AOU Check-list of North American Birds (1983). This study represents the only cytogenetic information reported for these species. Materials and Methods Cytological techniques followed those of Christidis (1985) and Hafner & Sandquist (1989). A chromosome was designated as a macrochromo¬ some if either the centromere or separate arms could be differentiated. 270 THE TEXAS JOURNAL OF SCIENCE-VOL. 47, NO. 4, 1995 All Other chromosomes were designated as microchromosomes. At least 10 metaphase cells of each species were examined in order to determine the diploid number. Both sexes were examined for each species. A total of 26 specimens were examined during the course of this study. Specimens were collected in accordance with scientific collecting permit numbers PRT-674149 (U.S. Fish and Wildlife Service) and SPR-0290- 021 (Texas Parks and Wildlife Department) issued to Terry C. Maxwell. Voucher study skins of 24 salvageable specimens are deposited with the holdings of the Angelo State Natural History Collections (ASNHC) at Angelo State University. Material examined.— Thryomanes bewickii.—5.5 mi N, 4.5 mi W of San Angelo, Tom Green County, Texas, one male specimen (ASNHC 1058) and two female specimens (ASNHC 1054, 1056); 5 mi S of Christoval, Tom Green County, Texas, one male specimen (ASNHC 1057). Thryothorus ludovicianus .—5 mi S of Christoval, Tom Green County, Texas, two male specimens (ASNHC 1050, 1051) and one female specimen (ASNHC 1053); 4 mi S of Christoval, Tom Green County, Texas, one male specimen (ASNHC 1048). Campy lorhynchus brunneicapillus .—\0 mi S 4.5 mi W of San Angelo, Tom Green County, Texas, one male specimen (ASNHC 1073); 10 mi SW of San Angelo, Tom Green County, Texas, one male specimen (ASNHC 694); 6.3 mi N, 13.2 mi W of Mertzon, Irion County, Texas, one female specimen (ASNHC 1074). Salpinctes obsoletus.—\Q.5 mi S, 2.3 mi W of San Angelo, Tom Green County, Texas, three female specimens (ASNHC 1061, 1062, 1063). Catherpes mexicanus.— 6.3 mi N, 13.2 mi W of Mertzon, Irion County, Texas, one female specimen (ASNHC 1063); 2 mi. S, 1 mi W of Leakey, Real County, Texas, one male specimen (ASNHC 1065) and one female specimen (ASNHC 1064). Cistothorus platensis . — W .?> mi N of Freeport, Brazoria County, Texas, two male specimens (ASNHC 1066, 1067) and three female specimens (ASNHC 1068, 1069 1070). Cistothorus palustris .—4 mi S of San Angelo, Tom Green County, Texas, one male specimen (ASNHC 1072) and one female specimen (ASNHC 1071). Results and Discussion The diploid number of chromosomes of wren species examined during this study ranges from 74 in Campy lorhynchus to 80 in both Salpinctes and Catherpes and is similar to diploid numbers reported in most birds. Shields (1980) reported that 70% of all birds karyotyped have diploid numbers between 76 and 84. Although no intraspecific polymorphisms were detected in karyotypes examined during this study, larger sample MINZENMAYER, MAXWELL & DOWLER 271 Figure 1. Karyotype of Campy lorhynchus brunneicapillus (ASNHC 694 cJ). Scale is 10 ^m. Figure 2. Karyotype of Salpinctes obsoletus (ASNHC 1061 9), Scale is 10 sizes with wider geographic representation would be required to accurately assess intraspecific variation. Campy lorhynchus brunneicapillus (Cactus Wren) exhibits a diploid number of 74 (Fig. 1). The eight largest pairs of macrochromosomes (including the sex chromosomes) are distinguishable by size from the remaining 29 pairs of smaller microchromosomes. The two largest pairs of macrochromosomes are submetacentric, the third largest pair is subtelocentric, the next two pairs are submetacentric and the sixth and seventh pair are telocentric. The eighth pair of macrochromosomes are the sex chromosomes. In this species, the W chromosome is telocentric, 272 THE TEXAS JOURNAL OF SCIENCE— VOL. 47, NO. 4, 1995 Figure 3. Karyotype of Catherpes mexicanus (ASNHC 1064 9). Scale is 10 pm. whereas the Z chromosomes are subtelocentric. Cistothorus palustris (Marsh Wren) (Fig. 2) and C. platensis (Sedge Wren) (Fig. 3) appear to have identical karyotypes, each with a diploid number of 76. They have eight pairs of macrochromosomes (including the sex chromosomes) and 30 pairs of microchromosomes. In both species, the largest pair is submetacentric, the next four pairs are subtelocentric, and the two smallest pairs are telocentric. The W chromosome appears to be telocentric, whereas the Z chromosome is subtelocentric. Thryomanes bewickii (Bewick’s Wren) (Fig. 4) and Thryothorus ludovicianus (Carolina Wren) (Fig. 5) also appear to exhibit identical karyotypes with a 2n of 76. They have nine pairs of macrochromo¬ somes (including the sex chromosomes) and 29 pairs of microchromo¬ somes. All other species in this study exhibited eight pairs of macrochromosomes. The largest pair for both species are submeta¬ centric, the next two pairs are subtelocentric, the fourth pair is submetacentric and the four smallest pairs are telocentric. Again, the W chromosome is telocentric, whereas the Z chromosomes are subtelo¬ centric. The diploid number of 76 observed in this study for T. ludovicianus differs from that reported for the only other karyotyped MINZENMAYER, MAXWELL & DOWLER 273 Figure 4. Karyotype of Thryothorus ludovicianus (ASNHC 1051 6). Scale is 10 fim. Figure 5. Karyotype of Thryomanes bemckii (ASNHC 1058 S). Scale is 10 fim. species within the genus. De Lucca & Chamma (1977) reported a 2n of 78 for T, leucotis. Salpinctes obsoletus (Rock Wren) (Fig. 6) and Catherpes mexicanus (Canyon Wren) (Fig. 7) have very similar karyotypes, both with a diploid number of 80. Both have eight pairs of macrochromosomes (including the sex chromosomes) and 32 pairs of microchromosomes. In both species, the largest pair is submetacentric, the next four pairs are subtelocentric, and the two smallest pairs are telocentric. The Z chromosome appears to be a large telocentric. The only difference in 274 THE TEXAS JOURNAL OF SCIENCE— VOL. 47, NO. 4, 1995 Figure 6. Karyotype of Cistothorus platensis (ASNHC 1069 9). Scale is 10 pm. Figure 7. Karyotype of Cistothorus palustris (ASNHC 1072 (5). Scale is 10 ^m. the karyotypes appears to be the W chromosome. In the Rock Wren, the W chromosome is metacentric, whereas that of the Canyon Wren appears to be telocentric. Some authors have considered these two species to be congeneric (Mayr & Short 1970; Morony et al. 1975). MINZENMAYER, MAXWELL & DOWLER 275 All genera of wrens that have been karyotyped share the first three largest pairs of macrochromosomes, with one exception; the second pair in Campy lor hynchus is submetacentric instead of subtelocentric as in all the others. The tendency for the largest three or four pairs of chromosomes to be identical among different groups of birds is fairly common (Stock et al. 1974; Ryttman et al. 1979; Ryttman & Tegelstrom 1981). Cistothorus, Catherpes, and Salpinctes all share pairs four and five whereas Thryothorus, Thryomanes, and Campy lorhynchus share pair four. With regard to pair five, Campy lorhynchus is submetacentric, whereas Thryothorus and Thryomanes are both telocentric. All species share pairs six and seven. Only Thryothorus and Thryomanes have the eighth pair of macrochromosomes. Thryomanes , Thryothorus , Cisto¬ thorus, and Campy lorhynchus all appear to share the Z and W chromosomes. Catherpes, and Salpinctes share the same Z chromosome but appear to have different W chromosomes. This study represents the only cytogenetic information reported for these species of wrens. Further karyological investigation, with chromosomal banding, should provide clarification of divisions within this phylogenetically challenging assemblage. Acknowledgments We acknowledge the field and laboratory assistance of the following individuals: A. Boyd, D. Roeder, D. Marsh and S. Parrish. Thanks to the Boulware family for use of the Head-of-the-River Ranch and for their scholarship which partially funded this research. We also thank the staff of the U.S. Fish and Wildlife Service, San Bernard National Wildlife Refuge for their assistance in collecting Sedge Wrens. Literature Cited American Ornithologists’ Union. 1983. Check-list of North American Birds, 6th Edition. Allen Press, Inc., Lawrence, Kansas, xxix -I- 877 PP- Christidis, L. 1985. A rapid procedure for obtaining chromosome preparations in birds. Auk, 102:892-893. De Boer, L. E. M. 1984. New developments in vertebrate cytotaxonomy VIII . A current list of references on avian karyology. Genetica, 65:3-37. De Lucca, E. J. & L. Chamma. 1977. Estudo do complemento cromosomico de 11 especies de aves das Ordens Columbiformes, 276 THE TEXAS JOURNAL OF SCIENCE-VOL. 47, NO. 4, 1995 Passeriformes e Tinamiformes. Rev. Bras, de Pesquisas Med. e Biol., 10:97-105. Hafner, J. C. & D. R. Sandquist. 1989. Post mortem field preparation of bird and mammal chromosomes: an evaluation involving the pocket gopher, Thomomys bottae. Southw. Nat., 34(3):330-337. Mayr, E. & L.L. Short. 1970. Species Taxa of North American Birds. Publ. No. 9, Nuttall Ornith. Club., Cambridge, Mass. Morony, J. J., Q. J. Bock & J. Farrand. 1975. Reference list of the birds of the world. American Museum of Natural History, New York. Ryttman, H., H. Tegelstrom & H. Jansson. 1979. G- and C-banding in four related Lams species (Aves). Hereditas, 91:143-148. Ryttman, H. & H. Tegelstrom. 1981. G-banded karyotypes of three Galliformes species. Domestic Fowl (Callus domesticus). Quail (Cotumix cotumix japonica), and Turkey (Meleagris gallopavo). Hereditas, 94:165-170. Shields, G. F. 1980. Avian cytogenetics: New methodology and comparative results. Proc. XVIII Intern. Ornith. Cong., 1978:1226-1231. Shields, G. F. 1983. Bird chromosomes. Pp. 189-209, in Current Ornithology Vol. 1 (Richard F. Johnston, ed.). New York, Plenum Press. Shields, G. F. 1985. Organization of the avian genome. Pp. 271-290, in Perspectives in Ornithology (A. H. Brush and G. A. Clark, Jr., ed.), Cambridge University, New York. Stock, A. D., F. E. Arrighi & K. Stefos. 1974. Chromosome homology in birds; banding patterns of the chromosomes of domestic chicken, ring-necked dove, and domestic pigeon. Cytogenet. Cell Genet., 13:410-418. Sibley, C.G. & B.L. Monroe, Jr. 1990. Distribution and taxonomy of birds of the world. Yale Univ. Press, New Haven, Conn, xxiv 4- 1 1 1 1 pp. Takagi, N. & M. Sasaki. 1974. A phylogenetic study of bird karyotypes. Chromosoma, 46:91-120. Tegelstrom, H. & H. Ryttman. 1981. Chromosomes in birds (Aves): evolutionary implications of macro- and microchromosome numbers and lengths. Hereditas, 94:225-233. Udagawa, T. 1956. Karyogram studies in birds VIII. The chromosomes of some species of the Turdidae and Troglodytidae. Jap. Jour, of Zool. 12:105-111. TEXAS J. SCI. 47(4):277-286 NOVEMBER, 1995 CHANGES IN THE GEOGRAPHIC CENTERS OF THE POPULATION OF TEXAS FROM 1850 TO 1990 Christopher S. Davies and Robert H. Brinkman Department of Geography and Bureau of Economic Geology, The University of Texas, Austin, Texas 78712 Abstract.— This study presents a spatial and tabular analysis of the movement of the center of population of Texas for each of the Decennial Censuses from 1850 to 1990. The population center’s geographic location has remained remarkably constant over the past 140 years, demonstrating that even with dramatic spurts of population growth at various times and in various locations throughout the state, the relative distribution of Texas’s city sizes by population has remained surprisingly constant. This suggests a regularity and stability in the Texas settlement pattern over time. Officials of the US Bureau of the Census state that a nation or state’s gravity center of population is the “preeminent measure for analyzing population distribution and change” (US Bureau of the Census 1974). Perturbations in the location of such a center reflects ongoing economic and settlement changes. This study presents a spatial and tabular analysis of the movement of the center of population of Texas for each of the Decennial Censuses from 1850 to 1990. One could envision the center of the population of Texas as the demographic heart or balance point of the state’s population distribution. On an imaginary, flat, weightless and rigid map of Texas, where each Texan is assumed to have equal weight, and where each Texan exerts an influence on the center of population proportional to his or her distance from the center, the state is demographically balanced at this population center of gravity. Viewed another way, if the state’s total population aspired to be spatially efficient by assembling at a point whose location minimized the aggregate sum of all straight-line travel distances from their residences to this point, then this pivotal point or group centroid would represent the state’s population center or demographic heart. The present population settlement pattern of Texas can be character¬ ized as clusters of densely populated metropolitan areas, their inter¬ metropolitan peripheries, and thinly populated rural sectors (Davies 1986: Figure 1). The coalescence of Dallas-Fort Worth, Waco, Austin, San Antonio and the Houston-Bay area into a triangle of dense, con¬ joining urban corridors, is a regional feature that contrasts with the state’s more isolated cities, such as Amarillo, Lubbock, McAllen, Brownsville, El Paso and Wichita Falls. 278 THE TEXAS JOURNAL OF SCIENCE-VOL. 47, NO. 4, 1995 Center Computation In calculating the location of the center of population of Texas for the 15 censuses between 1850 and 1990, along with its movement from decade to decade, the county was used as the geographic unit of measurement. The coordinates of a county’s center of population must be calculated for the state’s 254 counties, for each of the 15 censuses from 1850 to 1990. A county’s population center generally lies around its most populous city. If a county’s population does not coalesce in this manner, a center is determined that reflects the proportional influences on its location of the county’s three or four largest cities. A typical population pattern found in most Texas counties is a dispersion around a radial transportation network centered on a county seat, with smaller settlements strung out along the highway, and scattered populations filling the interstices. The population center of each county north and south of the parallel line is multiplied by the distance of this center from the parallel to provide the north and south moments for that county. The county’s east and west moments from the meridian line are similarly derived. In calculating the moments the distances are measured in minutes of arc. These moments are the directional corrections each county population exerts on the location of the state’s center of population. Under a weighted-mean system these exertions are expressed in a way that standardizes correctional differences. The county center of population moves north or south, or east or west of the parallel and meridian lines respectively, depending upon which directional product (population times distance) or moment is in excess. The difference between the north and south product sum or moments is divided by the total population of the county to give the number of degrees, minutes, or seconds of arc that the county’s center of population is north or south of this parallel line. Similarly, the difference between the east and the west product sums is divided by the total county population to give the measure of arc the population center is east or west of the meridian line. For example, the center of population of Anderson County in 1850 is calculated to lie near Palestine, at 95°38’ West longitude, and 31°46’ North latitude. To calculate the center of population of Texas, an assumption is first made as to its location in 1850. From archival records on the distri¬ bution of 212,592 people in 1850 in Texas, the state’s population center DAVIES & BRINKMAN 279 is assumed to intersect at a meridian of longitude, 97° West, and a parallel of latitude, 31°North, in northern Milam County. One can perceive of this meridian line and parallel line as the axes of moments crossing the state. The relationship of the coordinates of each of the 254 county centers of population to the coordinates of the center of population of Texas for each of the 15 Decennial Censuses is now estimated. For example, the relationship or location of Anderson County’s center of population (95°38’ West, 31°46’ North) to that of the population center of Texas located in Milam County in 1850 (97°West, il’North) is recorded as 82 min East, and 46 min North of this Milam County location. This difference is Anderson County’s distance or correctional factor. The populations of the 254 counties for each of the individual 15 Decennial Censuses are then multiplied by their own correctional distances to provide the population influence each county has on the center of population of Texas for that census date. For example, in 1850 Anderson County’s population was 2,884. Therefore, in 1850 Anderson County’s population exerted an influence, or a refinement on the coordinates of the population center of Texas of 236,488 people-min East (82 min East x 2,884 people), and 132,664 people-min North (46 min North x 2,884 people). The sum of all 254 county population influences is the total influence or pull of all county populations on the state’s center of population for each census between 1850 and 1990. The total westward influence is subtracted from the total eastward influence, and the total southern influence is subtracted from the total northern influence. In 1850, this statewide population influence on the center of population of Texas is 15,495,632 people-min to the East, and 673,006 people-min to the North. These above East and North correctional influences are then divided by the 1850 population of Texas of 212,592 to give a correctional factor in longitude and latitude to the 1850 population center of Texas of 72.8891 people-min to the East (15,495,632/212,592), and 3.165 people-min to the North (673,006/212,592). This is the total moment-arm correction on the center of population of Texas for the 1850 census date. The assumed center of population of Texas for 1850 of 97°West, and 31°North, is now adjusted by this statewide correctional factor of 72.8891 min East, and 3.1657 min North, to establish the final location of the center of population of Texas of 95 °47’ 280 THE TEXAS JOURNAL OF SCIENCE-VOL. 47, NO. 4, 1995 Table 1. Geographical position of Texas’ centers of population from 1850 through 1990, followed by the resident county of each site and the directional movement in miles since the last census. Year Longitude Latitude County (*) Directional Movement (in miles) WENS 1850 95047. 7.. 31° 3’10" Madison (Midway) 1860 96°14’32" 31°13’25" Leon (Marquez) 23.5 10.2 1870 96°21’55" 31° 9’11" Robertson (Easterly) 6.3 4.2 1880 96°39’27" 31°22’48" Limestone (Thornton) 15.0 13.6 1890 96°51’56" 31°28’28" Falls (Perry) 10.2 5.7 1900 96°54’12" 31°28’41" McLennan (Riesel) 1.9 0.2 1910 97°15’19" 31°31’46" McLennan (Waco) 18.0 3.1 1920 97019,33.. 31°29’26" McLennan (McGregor) 3.6 2.3 1930 97034, 9.. 31°22’41" Coryell (Oglesby) 12.4 6.7 1940 97°28’15" 31°15’32" Bell (Pendleton) 5.0 7.1 1950 97°35’59" 31° 7’46" Bell (Nolanville) 6.6 7.8 1960 97°4r48" 31° 5’47" Bell (Killeen) 5.0 2.0 1970 97032,12" 31° 5’ 0" Bell (Nolanville) 8.2 0.8 1980 97028’ 1" 30°59’24" Bell (Salado) 3.6 5.6 1990 97°26’54" 30°58’28" Bell (Little River) 1.0 0.9 (*) Nearest town to the geographic center within the county. West, and 31 °3’ North for 1850. This procedure is repeated to establish the coordinates of the center of population of Texas for the 7th through the 21st United States Censuses, 1850-1990 (Table 1, Figure 1). Center Movement The location and movement of the state’s center of population reflects the direction the population of Texas has grown and, to a certain extent, the direction the economic life of the state has taken. What Texans do for a living has a profound effect on the growth and shape of the state’s urban pattern and the shifts in the fortunes of its various regions. As to directional movement what has been the directional movement of this center over this 140 year interlude, and how extensive has this movement been? Are there regularities, cycles, and underlying mechanisms associated with its location over this 140 period? Figure 1 shows that from 1850 to 1910 Texas’s center of population moved due west. Then from 1910 to 1990 it veered almost directly south, with a slight easterly gait. The decided southward direction shows the pulling DAVIES & BRINKMAN 281 Figure 1. Geographic centers of population of Texas from 1850 through 1990. effect or influence on the center’s location of large population gains in Houston, Austin, San Antonio and the Gulf city complex. This southward drift would be even more emphatic if it were not counter-balanced by the population growth of the Dallas-Fort Worth-Denton metropolitan area. Although Texas has the greatest area of any of the continental 48 states, the movement of its center of population is not large, since “the population development in nearly all parts of the state has been nearly uniform” (US Bureau of Census 1974). The accumulated straight-line movement of the center for this 140 year period is a mere 145 miles, or a shift of 10.3 miles every 10 years. The center has shown little movement over the last 40 years being firmly ensconced in Bell County. Table 1 shows in straight-line distances the direction West, East, North or South this population center dislocated for every census decade. For example, the population center traveled 18 miles West, and 3.1 miles North, between 1900 and 1910. The center’s shortest movements occur from 1970 to 1980, and from 1980 to 1990, when Texas gained its largest ever absolute population increases, some 3.03 million people, and 2.76 million people, respectively. This reemphasizes the almost uniform constancy of this absorption of people into the settlement pattern of Texas. 282 THE TEXAS JOURNAL OF SCIENCE-VOL. 47, NO. 4, 1995 In comparison with the movement and the direction of the center of population of the nation and the remaining states, what has been the fate of the center of Texas? The US Bureau of the Census has calculated the center of population for the United States since the first census of 1790. Then the U.S. center of population was located in Maryland, some 23 miles east of Baltimore. By 1990 this population center had migrated southwestward to a site in southern Missouri as a reflection of the influence of the nation’s population growth in the west and the south. At intermittent times since 1880, the Bureau has also calculated the center of population for each state, principally, from 1880 to 1920, and from 1950 to 1970. Using the last period some 8, 17, 7, and 9 states exhibit North, South, East or West movements respectively, with 9 states and the District of Columbia showing no directional trend. For 1950 to 1970 Texas joins the 17 states with a southern drift and is also similar to 16 other states in that its center of population lies close to its state capital. Nevada’s center of population exhibits the greatest motion with 41.2 miles, and 32.2 miles, for 1950 to 1960, and 1960 to 1970 respectively. Rhode Island had the smallest movement of 0.5 mile and 0.3 miles, with the center of population of Texas moving five miles and six miles for these same periods. Sources of Error/Comparisons Some potential sources of error in the authors calculation of the center relate to the accuracy of the US Bureau of the Census’s county population counts, the assumption that the grid of longitudes and latitudes over Texas is perfectly flat, when the surface is very slightly curved, and the authors use of the county as the geographic unit of analysis. This study and the US Bureau of the Census used different geographical units in the calculation of the population center of Texas for the decades when such comparisons were permissible. The geographic unit used by the Bureau groups the population of Texas by square degrees (that is, by areas included between consecutive parallels and meridians). This study used the county as the geographical unit of analysis. Differences in the siting of the center of population of Texas is almost negligible when either geographic unit of analysis is used (Table 2). This suggests that the county is as effective a geographic unit of analysis as that of square degrees. Furthermore, the county population is more easily obtainable and understandable as a unit of analysis than that of square degrees. DAVIES & BRINKMAN 283 Table 2. Comparison of the geographical centers of Texas as determined by the U.S. Bureau of Census versus this study followed by the differences in miles between the two studies. The U.S. Census Bureau did not calculate the geographical center of Texas for 1930, 1940, 1980 or 1990. Year U.S. Bureau Census Longitude U.S. Bureau Census Latitude This Study Longitude This Study Latitude Diffenence in Miles 1880 96°38’30" 3P20’50" 96°39’27" 31°22’48" 2.1 1890 96°50’52" 3r26’ll" 96°5r56" 31°28’28" 2.5 1900 96°52’56" 31°28’35" 96°54’12" 31°28’41" 1.5 1910 97°15’14" 31°3r23" 97°15’19" 31°3r46" 0.4 1920 97°19’12" 31°28’34" 97°19’33" 31°29’26" 0.9 1950 97°35’10" 31° 7’30" 97°35’59" 31° 7’46" 0.8 1960 97°40’59" 31° 5’56" 97°41’48" 31° 5’47" 0.7 1970 97°31’30" 31° 4’52" 97°32’12" 31° 5’ 0" 0.6 Results and Discussion In 1850 when Texas’s population totaled 212,529, the state’s population center was located at longitude 95°47’7”, and latitude 31°3’10”, around the town of Midway in Madison County, (sited in Walker County in 1850). One hundred and forty years later, with Texas’s population registering 17 million in 1990, the state’s population center had shifted to longitude 97°26’54”, and latitude 30°58’28”, near the town of Little River in Bell County. The directional movement of the center of population of Texas has moved West and then from 1910 to the present, almost vertically South (Figure 1). In the course of this 140 year interlude, the population center had transported little more than 145 miles, a rate of a mere mile a year. Why is the overall location of the center of population of Texas so constant over these 140 years, when various parts of the state have experienced dissimilar spurts of growth at different times, which trans¬ lates into increases in settlement size where that growth occurs? The maturation of the Texas urban system has resulted from spatial instabili¬ ties in which random discoveries of the state’s residue of available energy that evoked the rise of such cities as Amarillo or Odessa, became amplified, triggered growth, and in the process remolded the state’s settlement structure. As Texas went through a series of growth cycles, different parts of its present urban system emerged. Distinct differences in the pace of urbanization occurred for different regions of Texas during the period 1890-1980. For expediency, the state has been divided into six broad regions, within whose arbitrary 284 THE TEXAS JOURNAL OF SCIENCE- VOL. 47, NO. 4, 1995 boundaries estimates were made of the proportional increase in the pace of urbanization for 1890 to 1980 (Davies 1986: Figure 9). Generally, it is assumed that the Panhandle and West Texas (PWT) is the rural bastion of Texas. Yet from 1890 onward, following an attack on the state’s oil and gas resources, it underwent a proportional increase in the rate of urbanization far in excess of other Texas regions. This is obviously not true for absolute numerical increases, since the Dallas and North Texas Region (DNT), the Central Texas Region (CT), and the Houston and Gulf Coast Region (HGC) started off with much larger initial population bases. What is much more significant than the mere numerical increase in the urban population of each region is the proportional rate of growth experienced by these diverse areas. The East Texas region experienced the slowest rate of urban growth. The other regional profiles are similar, with the Houston Gulf Coast Region the front runner, although lagging well behind the PWT region in the proportional rate of growth. The logarithmic curves show almost parallel slopes since 1960 for all regions, suggesting that the pace or rate of urbanization in Texas is now fairly uniform throughout the state (Davies 1986). While pressure exerted by each individual Texan on the state’s population center increases in proportion to the distance traveled away from it, making population growth in such peripheral cities as Tyler- Longview-Marshall, Midland-Odessa, Laredo- El Paso and so forth, a more powerful relative influence on the center’s location than similar population growth in cities close to the center, such as those found in the state’s urban triangle (Dallas-Fort Worth, Austin, San Antonio and Houston) the larger population increases in this triangle, counters or nullifies the stronger relative influence of population growth in cities more distant from the center. The circular and cumulative causation behind the population growth of Texas since its founding as a state creates a multiplier effect, in that the regional dominance of the state’s heartland or triangle has hardly weakened over this 140 year interlude. Whatever population increase has occurred in the state’s peripheral locations over the past 140 years, such increases failed to substantially affect the location of the state’s population center. In fact the influence is almost negligible when the locations of the center in Figure 1 is examined. This affirms the fact that for the past 140 years, the pace or rate of population absorption into the settlement pattern of Texas has been fairly uniform throughout the state. The present population of Texas has grown, as Malthus observed, exponentially, that is by DAVIES & BRINKMAN 285 constant proportions and has been absorbed into the state’s settlement pattern according to the law of the urban hierarchy of Texas, which is the law of regularity (Davies 1986). The constancy of the center’s location demonstrates that even with these spurts of population growth at various times and in various locations, the distribution of the city sizes of Texas by population, relative to each other, has remained remarkably constant, suggesting a regularity and stability in the Texas settlement pattern over time. If Texas cities are ranked in size from the largest downward, then the population of a given city tends to be equal to the population of the largest city divided by the rank of the given city. The J-shaped curve of Figure 10 of Davies (1986) (which is a plot of the city sizes of Texas and rank for selected years from 1860 to 1980) suggests that, as the overall population of Texas increased, Texas cities remained distributed in simple rank conformance. The straightening of population curves (Davies 1986: Figure 10) over time for Texas is evidence of increasing regularity (like that of the United States as a whole) and suggests that in the development of the urban system of Texas two conditions prevail: equal increases in the number of towns of the lower size order and uniform distribution of the population increment through most city sizes. The rank-size distribution of Texas is a logical outcome or a by-product of its urban system and is sufficiently stable to allow one to project future patterns. It appears that whatever the general growth of the Texas economy, this growth is distributed throughout the urban system, so that no one city class size monopolizes all the growth at any one time. It also appears that in the history of the urban development of Texas, a city begins with a random size and thereafter grows in an exponential manner, that is, proportional to its size. The Texas urban hierarchy grows under the law of allometric growth which is to say that the rate of growth of an individual city within the system is proportional to that of the system as a whole. The whole process of the civilization of Texas has been largely one of the aggregation of greater and greater numbers of people into limited areas; an apparently irresistible pressure to work toward a climax state, a provisionally stable pattern. In a spatial context, since distance is an inconvenience one should minimize it. The site and situation of the location of the population center of Texas suggests that it is an ideal location for hotel investment, if minimizing travel for all Texan’s is a criteria for site selection. The 286 THE TEXAS JOURNAL OF SCIENCE— VOL. 47, NO. 4, 1995 selection by Apple Computer of Round Rock in Williamson County as the site for their new factory has merit beyond such vicissitudes as abatement incentives, in that, being close to the state’s population center, it minimizes travel distances for all those visiting the plant and for deliveries from the plant. The Texas economy is moving from a material-intensive to an information-intensive base; from a mechanical model, with its roots in mechanical energy, to the biological model, whose antecedents lie in information. How will this transformation affect the urban cadastre of Texas? From evidence on the location of the center of population of Texas, especially its location for the past forty years, the import is clear. With the population of Texas predicted to be 22.3 million by the year 2026, which maybe an underestimation, given Texas’s large 1990 population of 17 million, this population will be absorbed into the state’s present settlement pattern in a manner that will not affect it substantially, and neither consequently that of the location of the center of population of Texas. Acknowledgements Appreciation is extended to Mr. Frank Brazile for the preparation of Figure 1. Literature Cited Davies, C. S. 1986. Life at the edge. Urban and industrial evolution of Texas, frontier wilderness-frontier space, 1836-1986. South¬ western Historical Quarterly 89(4): 443-554. Texas Comptroller of Public Accounts. 1992. The changing face of Texas, Texas through the year 2026: Economic growth, cultural diversity. Austin, Texas, 35pp. US Bureau of the Census. 1925. Statistical Atlas of the United States. US Government Printing Office, Washington, D.C., 476pp. US Bureau of the Census. 1974. Centers of population for states and counties. US Government Printing Office, Washington, D.C., 96pp. TEXAS J. SCI. 47(4):287-294 NOVEMBER, 1995 WEIGHT ESTIMATION FOR AXIS, FALLOW, SIKA AND WHITE-TAILED DEER IN TEXAS David A. Osborn, Stephen Demarais and R. Terry Ervin School of Forest Resources, University of Georga, Athens, Georgia 30602 and Department of Range and Wildlife Management and Department of Agricultural Economics, Texas Tech University, Lubbock, Texas 79409 Abstract.— Predictive equations provide wildlife managers and sportsmen with a practical and reliable estimate of deer weight. These equations have not been reported for axis {Axis axis), fallow {Dama dama), sika {Cervus nippon), or white-tailed deer {Odocoileus virginianus) in Texas. To address this need, the relationships between heart girth, live weight, dressed weight, and carcass weight were assessed for these cervid species in central Texas. Separate predictive equations were required for each species, age class, and season. General models using heart girth to provide an estimate of weight had values of 0.76, 0.75, and 0.75 for live weight, dressed weight, and carcass weight, respectively. General models using dressed weight to predict live weight and using live weight to predict carcass weight had R^ values of 0.89 and 0.83, respectively. Predictive equations based on heart girth or partial body weight provide a simple and reliable estimate of deer weight. The low cost and convenience of weight equations makes them practical for use by wildlife managers and sportsmen (Smart et al. 1973). Viable populations of non- native deer species are present in Texas and offer a popular sporting alternative to indigenous big game. Axis, fallow, and sika deer are the most common species of exotic deer present on Texas rangelands (Traweek 1989). However, weight estimation equations have not been reported for these species nor for native white-tailed deer in the Edwards Plateau Region of Texas. The objectives of this study were to test for effects of season, age, and species on regression models for body weights of axis, fallow, sika, and white-tailed deer and to develop predictive equations of body weight for each species. Study Area and Methods Deer were harvested on four privately owned ranches located in Kerr and Real Counties, Texas during two winter (15 December 1987-15 January 1988 and 15 December 1988-15 January 1989) and two summer (15 July 1988-15 August 1988 and 15 July 1989-15 August 1989) sampling periods. These ranches were predominantly rangeland used for 288 THE TEXAS JOURNAL OF SCIENCE-VOL. 47, NO. 4, 1995 Table 1 . Linear regression equations developed to estimate live weight from heart girth for axis, fallow, sika, and white-tailed deer in Texas, 1987-89. Live Weight (kg) Species Age Season n X SE® Equation Axis Adult & Subadult Summer 55 44.1 3.6 Y = -14.38 -H 0.08(HG) Adult Winter 46 46.8 3.0 Y = - 2.67 + 0.07(HG) Subadult Winter 10 39.5 3.7 Y = - 2.67 -h 0.06(HG) Fallow Adult Summer 38 37.1 4.1 Y = -22.02 -f 0.08(HG) Adult Winter 42 38.0 2.6 Y = -10.31 -}- 0.07(HG) Subadult Summer 16 29.5 2.9 Y = -20.47 4- 0.08(HG) Subadult Winter 4 33.8 2.8 Y = - 8.75 + 0.06(HG) Sika Adult & Subadult Summer 44 37.1 3.7 Y = -20.46 -f 0.08(HG) Adult Winter 40 39.1 3.3 Y = - 8.75 + 0.07(HG) Subadult Winter 6 33.4 2.7 Y = - 8.75 + 0.06(HG) Whitetail Adult Summer 47 34.5 3.2 Y = -22.88 -f 0.09(HG) Adult Winter 50 34.3 2.7 Y = -11.17 -f 0.07(HG) Subadult Summer “ 2 24.5 Y = -20.46 4- 0.08(HG) Subadult Winter 3 27.8 0.2 Y = - 8.75 -h 0.06(HG) ^ Standard error of the estimate. ^ Based on the general model Y = - 8.745 -I- 0.058(HG) -I- 6.078 (Axis) - 11.714 (Summer) - 1.564 (Fallow*Adult) - 2.421 (Whitetail* Adult) + 0.007 (Adult*Girth) + 0.018 (Summer*Girth) - 0.003 ( Axis*Summer* Adult*Girth) 4- 0.003 (Whitetail*Summer*Adult*Girth). = 0.7615, Adj. R^ = 0.7561, N = 403, F = 139.5. ^ Y = live weight in kg, HG = heart girth in mm. ^ Too few observations to calculate SE. the production of domestic livestock and wildlife and were characteristic of the Edwards Plateau Region (Butts et al. 1982; Landers 1987). Body weights and heart girth measurements were recorded for 111 axis, 100 fallow, 90 sika, and 102 white-tailed deer. All deer were female and classified as either subadult (1 .0-1 .5 years-old) or adult (>1.5 years-old) based on tooth eruption and wear criteria (Severinghaus 1949; Graf & Nichols 1966; Chaplin & White 1969; Duff 1969). A linear measurement of heart girth (Smart et al. 1973) using a flexible steel tape and recorded to the nearest 0.5 cm was taken with the deer lying on its side prior to evisceration. Live weight, defined as the weight of a recently harvested deer minus blood loss from the gunshot wound to the head, was measured using a spring scale. An electronic scale was used to determine carcass weight, defined as the weight of the eviscerated deer minus the head, hide, and feet. The combined weight of the head, hide, and feet was then measured using a spring scale and added to the carcass weight to arrive at the field dressed weight. Spring OSBORN, DEMARAIS & ERVIN 289 Table 2. Linear regression equations developed to estimate dressed weight from heart girth for axis, fallow, sika, and white-tailed deer in Texas, 1987-89. Dressed Weight (kg) Species Age Season n X SE ^ Equation Axis Adult Summer 30.1 2.2 Y = -15.11 + 0.06(HG) Adult Winter 56^ 28.5 1.7 Y = -13.95 + 0.06(HG) Subadult Winter 10 24.6 2.9 Y = -13.95 + 0.05(HG) Fallow Adult Summer 38 21.7 2.4 Y = -17.33 + 0.06(HG) Adult Winter 42 22.7 2.2 Y = -16.18 + 0.05(HG) Subadult Combined 20 19.4 1.6 Y = -13.95 + 0.05(HG) Sika Adult Adult & Summer 37 23.6 2.6 Y = -15.11 + 0.06(HG) Subadult Winter 53 23.8 2.7 Y = - 8.75 + 0.07(HG) Whitetail Adult Adult & Summer 47 21.9 2.0 Y = -12.44 + 0.05(HG) Subadult Winter 55 22.3 2.1 Y = -11.28 + 0.05(HG) ^ Standard error of the estimate. ^ Based on the general model Y = - 13.953 + 0.048(HG) + 2.669 (Whitetail) - 2.223 (Fallow* Adult) - 1.156 (Summer* Adult) + 0.004 (Adult*Girth) + 0.003 (Summer*Girth) + 0.005 (Axis*Girth) - 0.004 (Whitetail*Adult*Girth). R^ = 0.7543, Adj. R' = 0.7493, N = 403, F = 151.2. ^ Y = dressed weight in kg, HG = heart girth in mm. ^ This equation also represents axis subadults during the summer. scales were calibrated with the electronic scale, prior to each weighing session, and all weights were recorded to the nearest 0.45 kg. Statistical analyses were performed with the PC version of the computer program SHAZAM (White et al. 1988) on an IBM PC-AT. Linear regression equations were developed describing the weight-girth and weight- weight relationships. Differences among age, season, species and the various interaction terms in the regression equations were tested with the creation of dummy variables (Leistritz 1973). Results Coefficient of determination (R^) values suggested most of the variability in our data were accounted for in the various regression models. General models using heart girth to provide an estimate of weight had R^ values of 0.76, 0.75, and 0.75 for live weight, dressed weight, and carcass weight, respectively. General models using dressed weight to predict live weight and using live weight to predict carcass weight had R^ values of 0.89 and 0.83, respectively. Equations developed to estimate live weight from heart girth (Table 1) varied by age and season. For each species, the live weight to girth 290 THE TEXAS JOURNAL OF SCIENCE— VOL. 47, NO. 4, 1995 Table 3. Linear regression equations developed to estimate carcass weight from heart girth for axis, fallow, sika, and white-tailed deer in Texas, 1987-89. Species Age Season n Carcass Weight (kg) X SE “ Equation Axis Adult Summer 45 24.2 1.9 Y = -13.24 + 0.05(HG) Adult Winter 46 23.4 1.6 Y = -14.19 + 0.05(HG) Subadult Summer 10 19.8 1.8 Y = -14.95 -F 0.05(HG) Subadult Winter 10 19.6 2.6 Y = -15.91 -}- 0.05(HG) Fallow Adult & Subadult Summer 54 15.9 2.1 Y = -10.22 -H 0.04(HG) Adult & Subadult Winter 46 16.8 2.1 Y = -11.18 -h 0.04(HG) Sika Adult & Subadult Summer 44 17.7 2.1 Y = -17.07 -H 0.05(HG) Adult & Subadult Winter 46 18.8 2.3 Y = -18.03 -h 0.05(HG) Whitetail Adult & Subadult Summer 49 17.2 1.7 Y = -15.98 0.05(HG) Adult & Subadult Winter 53 17.6 1.8 Y = -16.93 + 0.05(HG) ^ Standard error of the estimate. ^ Based on the general model Y = - 18.025 4- 0.049 (HG) + 2.119 (Axis) + 6.850 (Fallow) -I- 1.094 (Whitetail) + 0.953 (Summer) + 1.718 (Axis*Adult) - 0.012 (Fallow*Girth). = 0.7505, Adj. = 0.7460, N = 403, F = 169.7. ^ Y = carcass weight in kg, HG = heart girth in mm. relationship was different between age-classes during winter requiring separate equations for adults and subadults. Age-specific equations were needed for fallow and white-tailed deer in the summer but not for axis or sika deer. The dressed weight of a deer can be estimated from a linear heart girth measurement (Table 2). Separate equations were required for axis, fallow, sika, and whitetail adults harvested during summer. However, sika and whitetails each showed little variation in the dressed weight to girth relationship among adults harvested during the winter or subadults in either season. Age-related variation was apparent for axis and fallow deer. The relationship of carcass weight to heart girth (Table 3) observed for axis, sika, and white-tailed deer differed from that of fallow deer; however, none of the species showed an age or season effect on this relationship. Seasonal variation in intercept existed; therefore, separate equations in summer and winter were needed for fallow, sika, and white- tailed deer. Four equations were necessary for axis deer because OSBORN, DEMARAIS & ERVIN 291 Table 4. Linear regression equations developed to estimate live weight from dressed weight for axis, fallow, sika, and white-tailed deer in Texas, 1987-89. Species Age Season n Live Weight (kg) X SE ^ Equation Axis Adult Summer 45 45.7 2.3 Y = 7.20 + 1.28(DW) Adult Winter 46 46.8 2.1 Y =13.98 + L12(DW) Subadult Summer 10 37.1 1.0 Y = 7.20 -f 1.18(DW) Subadult Winter 10 39.5 2.3 Y =13.98 + 1.02(DW) Fallow Adult Summer 38 37.1 3.3 Y = 8.54 + 1.31(DW) Adult Winter 42 38.0 2.5 Y =15.31 + l.Ol(DW) Subadult Summer 16 29.5 1.9 Y = 7.20 + 1.21(DW) Subadult Winter 4 33.8 1.8 Y =13.98 + 0.91(DW) Sika Adult Summer 37 38.2 2.3 Y = 7.20 + 1.31(DW) Adult Winter 40 39.1 2.0 Y =13.98 + l.Ol(DW) Subadult Summer 7 31.5 2.3 Y = 7.20 -f 1.21(DW) Subadult Winter 6 33.4 1.9 Y =13.98 -h 0.91 (DW) Whitetail Adult Summer 47 34.5 1.9 Y = 0.50 + 1.55(DW) Adult Winter 50 34.3 1.8 Y = 7.27 -h 1.19(DW) Subadult Summer ^ 2 24.5 Y = 0.50 + 1.39(DW) Subadult Winter 3 27.8 0.6 Y = 7.27 + 1.09(DW) ^ Standard error of the estimate. Based on the general model Y = 13.975 + 0.906 (DW) - 6.705 (Whitetail) - 6.773 (Summer) + 1.337 (Fallow*Adult) -1- 0.103 (Adult*DW) + 0.115 (Axis*DW) -f 0.184 (Whitetail *DW) + 0.300 (Summer*DW) - 0.141 (Axis*Summer*DW) -I- 0.059 (Whitetail*Adult*Summer*DW). R^ = 0.8855, Adj. R^ = 0.8825, N = 403, F = 303.0. ^ Y = live weight in kg, DW = dressed weight in kg. ^ Too few observations to calculate SE. of age- and season-related variation in intercept terms. Equations designed to predict live weight from dressed weight (Table 4) and carcass weight from live weight (Table 5) reflect age-related differences for axis, fallow, sika, and white-tailed deer. Variation in the live weight to dressed weight relationship also occurred by season for all species.. Only axis and fallow deer required season-specific equations to estimate carcass weight from a live weight measurement. Discussion Strong correlation among body weights and heart girth has been reported for white-tailed deer in the southeastern United States (Smart et al. 1973; Urbston et al. 1976; Weckerly et al. 1987). Results of this study verify this relationship for axis, fallow, sika, and white-tailed deer in Texas with standard errors for specific estimates similar to those reported for white-tailed deer in Illinois (Roseberry & Klimstra 1975) 292 THE TEXAS JOURNAL OF SCIENCE— VOL. 47, NO. 4, 1995 Table 5. Linear regression equations developed to estimate carcass weight from live weight for axis, fallow, sika, and white-tailed deer in Texas, 1987-89. Species Age Season n Carcass Weight (kg) X SE ^ Equation Axis Adult Summer 45 24.2 1.3 Y = 0.57 -f 0.52(LW) Adult Winter 46 23.4 1.6 Y = 0.57 -f 0.49(LW) Subadult Summer 10 19.8 1.1 Y = -1.46 + 0.57(LW) Subadult Winter 10 19.6 1.5 Y = -1.46 + 0.54(LW) Fallow Adult Summer 38 16.3 1.8 Y = 2.97 + 0.36(LW) Adult Winter 42 16.9 2.2 Y = -1.46 + 0.36(LW) Subadult Summer 16 14.9 1.5 Y = 2.97 -f 0.53(LW) Subadult Winter 4 15.3 2.0 Y = -1.46 + 0.53(LW) Sika Adult Combined 77 18.5 1.8 Y = -1.46 -f 0.52(LW) Subadult Combined 13 17.0 2.4 Y = -1.46 -f 0.57(LW) Whitetail Adult Combined 97 17.5 1.3 Y = -0.27 + 0.52(LW) Subadult Combined 5 14.4 0.9 Y = -1.46 + 0.57(LW) ^ Standard error of the estimate. Based on the general model Y = - 1.459 + 0.567(LW) + 4.432 (Fallow*Summer) + 2.025 (Axis*Adult) + 1.194 (Whitetail*Adult) - 0.050 (Adult*LW) - 0.031 (Axis*LW) - 0.037 (Fallow*LW) + 0.031 (Axis*Summer*LW) - 0.123 (Fallow*Summer*LW) . = 0.8256, Adj. R2 = 0.8216, N = 403, F = 206.7. ^ Y = carcass weight in kg, LW = live weight in kg. and Tennessee (Weckerly et al. 1987). Analyses revealed the need for separate weight estimation equations among cervid species, age classes, and seasons in the Edwards Plateau Region of Texas. Although separate regression equations presented in this study have been determined to be statistically different (Tables 1-5), biologically significant differences may not exist in cases where both intercepts and slopes are very similar. Models were developed using data collected exclusively from female deer and, therefore, may or may not be applicable for males of the respective species (Smart et al. 1973; Weckerly et al. 1987). Also, the Edwards Plateau Region of central Texas supports a very dense deer population. Because deer weights may be population specific, caution should be used when applying these predictive equations to other deer populations. Acknowledgments We thank the owners and managers of the Bowman, Johnson, Two- Dot, and Y.O. ranches for their hospitality and assistance. We would also like to thank the Texas Wild Game Cooperative, All That’s Deer, OSBORN, DEMARAIS & ERVIN 293 Inc., J. J. Jackley, M. Bailey, and several students at Texas Tech University for their assistance in data collection and R. S. Lutz, L. M. Smith, and J. J. Jackley for reviewing the manuscript. This research was supported by the Exotic Wildlife Association, Texas Tech University, the Houston Livestock Show and Rodeo, and various private landowners. This manuscript is Texas Tech University, College of Agricultural Sciences Contribution T-4-590. Literature Cited Butts, G. L., M. J. Anderegg, W. E. Armstrong, D. E. Harmel, C. W. Ramsey & S. H. Sorola. 1982. Food habits of five exotic ungulates on Kerr Wildlife Management Area, Texas. Texas Parks and Wildlife Dept. Technical Series, No. 30. Chaplin, R. E. & R. W. G. White. 1969. The use of tooth eruption and wear, body weight and antler characteristics in the age estimation of male wild and park Fallow deer {Darna darna). J. Zool., London, 157:125-132. Duff, K. R. 1969. Tooth eruption as a guide to aging Japanese sika deer (Cervus nippon) in Dorset. Deer, 2:566-567. Graf, W. & L. Nichols, Jr. 1966. The axis deer in Hawaii. J. Bombay Nat. Hist. Soc., 63(3):629-734. Landers, R. Q. 1987. Native vegetation of Texas. Rangelands, 9:203-207. Leistritz, F. L. 1973. The use of dummy variables in regression analysis. Dep. Agric. Econ., N. Dakota State Univ., Misc. Report 13, Fargo, N. Dakota. Roseberry, J. L. & W. D. Klimstra. 1975. Productivity of white-tailed deer on Crab Orchard National Wildlife Refuge. J. Wildlife. Manag., 34(l):23-28. Severinghaus, C. W. 1949. Tooth development and wear as criteria of age in white-tailed deer. J. Wildlife Manag., 13(2): 195-216. Smart, C. W., R. H. Giles, Jr. & D. C. Guynn, Jr. 1973. Weight tape for white-tailed deer in Virginia. J. Wildlife Manag. , 37(4) :553- 555. Traweek, M. S. 1989. State- wide census of exotic big game animals. Job Report, Federal Aid Project W-109-R-12, Texas Parks and Wildlife Dept., Austin, Texas. Urbston, D. F., C. W. Smart & P. F. Scanlon. 1976. Relationship between body weight and heart girth in white-tailed deer from South Carolina. Proc. Annu. Conf. Southeastern Assoc. Game and Fish Comm., 30:471-473. 294 THE TEXAS JOURNAL OF SCIENCE— VOL. 47, NO. 4, 1995 Weckerly, F. W., P. L. Leberg & R. A. Van Den Bussche. 1987. Variation of weight and chest girth in white- tailed deer. J. Wildlife Manag., 51(2):334-337. White, K. J., S. A. Haun, N. G. Horsman & S. D. Wong. 1988. SHAZAM econometrics computer program. McGraw-Hill Book Co., New York. TEXAS J. SCI. 47(4):295-307 NOVEMBER, 1995 LATE PREHISTORIC SNAKES OF E. V. SPENCE AND O. H. IVIE RESERVOIR BASINS OF COKE, COLEMAN, CONCHO, AND RUNNELS COUNTIES, TEXAS Okla W. Thornton, Jr. and J. R. Smith Colorado River Municipal Water District, I vie Reservoir Field Office, HCR 82, Box 4B Lead ay, Texas 76888 Abstract.— Eighteen fossilized snake vertebrae representing at least four different genera were recovered from the archaeological surveys of four prehistoric campsites along the Colorado River of west-central Texas. The sites are now inundated by E.V. Spence and O.H. Ivie reservoirs. All fossils examined represent extant forms that are currently present in this region of Texas. The occurrence of these species indicates that the prehistoric climate in this area was very similar to that of today. The incompleteness of the fossils and their proveniences suggests that they were part of the aboriginals diet. Four snake genera are identified from four different prehistoric campsites along the Colorado River, Texas. Although few in number, the vertebrae have zooarchaeological significance. Because of their present-day habitat preferences, zoogeographic distribution, and the convergence of three physiographic regions in or near the fossil sites, the vertebrae found at these prehistoric campsites demonstrate a dietary use of snakes by early native Americans. Many archaeological investigations (Ruecking 1953; Sjoberg 1953; Newcomb 1961; Williams- Dean 1978; Steele & Mokry 1985; Shafer 1986; Steele & Hunter 1986; Steele 1986a; 1986b; Hellier et al. in press) have reported snake consumption but there appears to be little interest in the identification of those species which were being consumed. There are probably numerous snake bones stored in archeology holdings that are never examined in a critical manner (Parmely pers. comm.). It would appear that to fully understand the dietary habits of early native Americans, archaeologists would need to determine what types of snakes were being consumed by native Americans. Study Areas E. V. Spence Reservoir.— T\\t Sand Creek archaeological site (41CK79) is within E. V. Spence Reservoir basin. Coke County, Texas (Fig. 1). The Robert Lee Dam, which impounds the reservoir, is 48 km NNW of San Angelo. The prehistoric campsite where the vertebrae were recovered is 21 km NW of the dam on a 9 m terrace of Sand 296 THE TEXAS JOURNAL OF SCIENCE- VOL. 47, NO. 4, 1995 Figure 1 . E. V. Spence and O. H. Ivie reservoir locations along the upper Colorado River basin of west-central Texas. Creek about 305 m from the bank of the Colorado River (Shafer 1971). The elevation of the Sand Creek site is between 576-579 m above mean sea level (msl). The site is within the Mesquite Plains, a subregion of the Rolling Plains (Smeins & Slack 1982). O. H. Ivie Reservoir.— -ThvtQ prehistoric sites, now inundated by O. H. Ivie Reservoir, are located within a 6.1 km radius of the confluence of the Colorado and Concho Rivers (Fig. 1). The S. W. Freese Dam, which impounds the reservoir, is 26 km downstream from this con¬ fluence. The reservoir is approximately 74 km ENE of San Angelo. Site 41CN19 is 6.5 km north of the Freese Dam service spillway (Coleman County) on a left-bank terrace above the Colorado River in proximity to an unnamed ravine through which a perennial spring flowed. Site 4 ICC 131 is situated 7.8 km NW of Freese Dam (Concho County) on a right-bank terrace above the Concho River in proximity to an unnamed intermittent stream traversing a ravine. Site 41RN169 is 15.5 km NW of Freese Dam (Runnels County) on the left-bank of the Colorado River, about 228 m upstream on Rocky Branch, an intermit¬ tent stream. The elevation of the sites ranges from 457-469 m above msf. A detailed description of each site is presented by Lintz et al. (1993). THORNTON & SMITH 297 Methods and Materials Identifications were made by comparing the fossil vertebrae with modern reference skeletons in the collection of the senior author. Drawings of fossilized vertebrae from published literature were also reviewed to aid in comparisons and identifications. Catalog or lot and site numbers (in parentheses) refer to the field specimen and site collections (trinomial code), respectively, of the Texas Archeological Research Laboratory (TARL) and Mariah Associates, Inc. Radiocarbon (Carbon- 14) dates are based upon communitive charcoal recovered at each site. The fossil specimens are maintained at the Texas Archeological Research Laboratory. The classification system follows Dowling & Duellman (1976) with standard scientific names from Collins et al. (1978). Results The following species accounts include specimens of at least four different genera assigned to the families Colubridae and Viperidae. Drawings of fossil specimens examined during the course of this study are presented in Figure 2. Elaphe obsoleta (Say) Material examined.— Ont large trunk vertebra (FS 114, 41CC131; Fig. 2A-C) was found 130 cm below the surface north of the deep gully, but insufficient diagnostic artifacts were found with the vertebra to allow a relative age assignment. The large size of the vertebra indicates a mature individual. The neural arch is not highly vaulted and the subcentral ridges are relatively deep. Elaphe obsoleta is presently widespread in the eastern two-thirds of the state while the study area is near the western range limit (Dixon 1987). Elaphe obsoleta lives in a variety of habitats, but within the basin it is typically found in wooded stream valleys and rocky canyons (Conant & Collins 1991). Masticophis sp. or Coluber sp. (indeterminable) Material examined.— Eight trunk vertebrae (one FS 566.1, four FS 685.1, 41CN19; two FS 1218.1 (Fig. 2D-F), one FS 114, 41CC131) and two cervical vertebrae (one FS 1224.1, 41CC131; one FS 685.1, 41CN19) are identified as those of either Masticophis or Coluber. Vertebrae of Masticophis are inseparable from those of Coluber (cf. Holman 298 THE TEXAS JOURNAL OF SCIENCE- VOL. 47, NO. 4, 1995 Figure 2. Fossil vertebrae of Elaphe obsoleta (A-C), Masticophis/ Coluber sp. indet. (D-F), Thamnophis sp. (G-I), Crotalus sp. (J-O), each shown in dorsal, posterior and lateral views respectively. Scale is 1.0 cm. THORNTON & SMITH 299 1979; 1981; Parmley 1986; 1988a; 1988b; 1990). Though non- distinguishable, these vertebrae (FS 114, FS 1224.1, FS 566.1, FS 685.1, and FS 1218.1) exhibit the following characters typical of these genera: general shape long and constricted medially; well developed epizygapophyseal spines; thin, long neural spine; thin, relatively uniform in width hemal keel; strong posterior neural spine overhang; and high, domed neural arch. Masticophis I Coluber vertebrae are the most common elements recovered at the sites. Masticophis flagellum, Masticophis taeniatus, and Coluber constrictor currently occur sympatrically within the reservoir area (Dixon 1987). Coluber constrictor inhabits fields, grasslands, brushlands, and open woodlands; M. flagellum frequents grasslands, mesquite savannahs, arid brushlands, and many other more or less open habitats; and M. taeniatus prefers rocky breaks and stream valleys (Conant & Collins 1991). Of the ten vertebrae, all post- and prezygapophseal processes were damaged. One trunk vertebra (FS 566.1) of an adult snake was found 160-170 cm below the surface and shows evidence of being burned and slightly weathered. Four trunk and one cervical vertebrae (FS 685.1) were found 190-200 cm below the datum, and they appear to be from a single individual and show moderate weathering and signs of burning. The cervical vertebra was badly damaged and burned. Two trunk vertebrae (FS 1218.1) were found 40-50 cm below the surface. These specimens indicate two individuals with minute signs of weathering, however, both clearly show signs of being burned. One cervical vertebra (FS 1224.1) was found 190-200 cm below the datum. This specimen was slightly weathered with signs of being burned. One trunk vertebra (FS 114) was found 130 cm below the surface. It appears to be from a moderately sized individual, with some weathering and signs of being burned. Carbon- 14 dates associated with these vertebrae are: FS 566.1 and FS 685.1, associated materials and stratigraphic data suggests site utilization between A.D. 600 to 1050; FS 1218.1 and FS 1224.1, material remains suggest an age span of A.D. 1000 to 1300; and FS 114, no dates are associated because insufficient diagnostic artifacts were found to provide a relative age for this specimen. Thamnophis sp. Material examined.— Ont trunk vertebra (Lot 133, 41CK79; Fig. 2G- I) was found 22.9-30.5 cm below the surface in association with rich cultural deposits including freshwater mussel shell, flint, and hearthstones. Artifacts recovered suggest a time span of about A.D. 800 300 THE TEXAS JOURNAL OF SCIENCE- VOL. 47, NO. 4, 1995 to 1600 (1 100 to 200 YBP). This specimen appears to be Thamnophis, because of its elongated shape. However, it could be Nerodia harteri paucimaculata, although the neural arch is not quite the same. Nonetheless, the hypapophysis of Thamnophis is similar to N. harteri (Parmley pers. comm.), therefore, the possibility that this specimen could be N. harteri paucimaculata cannot be ruled out. At any rate, lack of adequate N. harteri reference skeletons make a positive identification difficult to impossible. Both Thamnophis marcianus marcianus and Thamnophis proximus rubrilineatus presently occur in or near the site. Thamnophis marcianus marcianus is widely distributed in the arid Southwest and T proximus rubrilineatus occurs in central to west-central Texas. Both seldom stray far from streambeds, springs, or other places where water may be present (Conant & Collins 1991), however, this site occurs at the western extremes of N. harteri paucimaculata range (Dixon 1987). Crotalus sp. Material examined.— Five trunk vertebrae (FS 141.1, 41CC13 1 ; Lot 115 (n=2). Lot 77, and Lot 231, 41CK79) are identified as Crotalus. Vertebra FS 141. l(Fig.2J-L) was found 170 cm below the surface during the testing phase in TP 4, north of the deep gully. No dates are associated and insufficient diagnostic artifacts were found to provide a relative age for this specimen. The vertebra appeared to be that of a moderately sized individual with signs of burning. FS 141.1 was identified as Crotalus sp. indet., in that it is rather short and wide, the zygosphene is thick, and the hypapophysis base is thick like a viperid. The remaining four trunk vertebra (Lot 115, Lot 77, and Lot 231) are from moderately sized rattlesnakes and more associated with artifacts suggesting a time span of about A.D. 800 to 1600. Two large vertebrae (Lot 115) were found 22.9-30.5 cm below the surface during the recovery phase in Level 4. All were charred and extremely fragmented. One vertebra (Lot 77) was found 15.2- 22.9 cm below the datum during the recovery phase in Level 3, and one vertebra (Lot 231) was found 30.5-38.1 cm below the surface during the recovery phase in Level 5. The overall shape (squareness), robustness, and low neural arch of the vertebra are characteristic of the genus Crotalus. Because of the poor preservation and fragmentation of the specimens, allocation to species cannot be ascertained. One trunk vertebra (FS 692. 1, 41RN169; Fig. 2M-0) was found 70- 80 cm below the surface within the central pit of a burned rock oven. THORNTON & SMITH 301 This vertebra is relatively well preserved and appears to be either C atrox or C. viridis. Both species presently occur at or near the site, but the single vertebra is indistinguishable from either species. In fact, Crotalus molossus may occur there as well. This vertebra is rather large indicating a large, adult snake with signs of slight weathering and being burned. The artifacts and C-14 dates support use of the oven between A.D. 700 and 1200. The specific identification of Crotalus vertebrae presents special problems that are yet to be worked out (Holman 1981; Parmley 1986; 1988a; 1990). Parmley (1988a) states that fossil vertebrae of C. atrox and C viridis of comparable size cannot be distinguished. He further states that specific identification of isolated Crotalus vertebrae has never been satisfactorily analyzed, and it is possible that Crotalus vertebrae are not diagnostic at the species level. The same problem exists for C. molossus. Both C. viridis and C molossus are medium-sized rattlesnakes (range 89-114 cm, record 144.8 cm; 76-106.7 cm, record 125.7 cm, respectively) compared to C. atrox (range 76-183 cm, record 213 cm) (Conant & Collins 1991). All three species presently occur at or near the fossil sites (Dixon 1987). Discussion The purpose of analyzing the few herpetofaunal remains recovered from 41CN19, 41CC131, 41RN169 and 41CK79 was twofold. First, the prehistoric occupation sites where the fossils were collected are in proximity to the mainstem stream riffles of the Colorado River. With the knowledge that the aboriginal inhabitants utilized riverine based resources (Carlson et al. 1982), the question was asked whether natracines, specifically, the Concho water snake {N. harteri paucimaculata) , occurred in the region prior to European occupation. Second, the opportunity to report the dietary use of prehistoric snakes by early native Americans. More than 50,000 bone specimens were recovered during the archaeological testing and data recovery within the I vie Reservoir basin. A majority of the bone specimens recovered represent small- to large¬ sized animals, many of which persist in or near the area today. Only a few of these bones were attributed to avifaunal, ichthyofaunal, and herpetofaunal assemblages (Lintz et al. 1993). The 13 snake vertebrae recovered represents only 0.03% of the total bones recovered. A total of 147 vertebrate bone specimens were recovered from the Spence Reservoir site, and snake vertebrae represented only 3.4% of 302 THE TEXAS JOURNAL OF SCIENCE-VOL. 47, NO. 4, 1995 that total. A review of earlier studies (Carlson et al. 1982; Steele & DeMarcay 1985; Steele & Mokxy 1985; Steele 1986a; 1986b; Steele & Hunter 1986; Hellier et al. in press) indicates that numerically the herpetofaunal portions of the assemblages are poorly represented. Steele (1986b) explains this apparent disparity by suggesting that mammals constitute the largest faunal remains because of a greater utilization by early man. His reasons were: (1) the bones of the other classes are smaller and more fragile, therefore more subject to destruction and loss; and (2) when the amount of meat contributed by each class of vertebrates is considered, the mammals are the most important, since most species of mammals are considerably larger than other vertebrates. Driver (1969) indicated that when large game animals were scarce, the hunters and gatherers relied more on rodents, reptiles, and insects. However, Steele & Mokry (1985) state that faunal utilization involved all classes and sizes of vertebrates, and furthermore, there is no evidence indicating any fauna was favored to the exclusion of others. Another question to be addressed is whether the fossils recovered at human habitation sites were intrusively or consumptively deposited. Steele & DeMarcay (1985) lists two criteria with which one can, with assurance, make these decisions. First, the disarticulation of the skeleton and its overall incompleteness represented by single bones or bone fragments indicates probable consumption. Second, a skeleton found articulated and generally complete, indicates an intrusive element. However, they admit one should be judicious when drawing conclusions based upon faunal remains. Nonetheless, the archaeological evidence suggests that when faunal remains are recovered disarticulated, few in number, and within the tight stratigraphic context directly associated with an archaeological habitation, then the taxon represented was most likely consumed. Bones recovered from hearth areas and showing signs of burning (charring) and disarticulation clearly point to human involvement, particularly dietary consumption (Shafer 1986). Furthermore, through coprolite examination Williams-Dean (1978) confirmed that snakes were consumed. A complete review of the available literature, concerning the dietary use of natracines by prehistoric peoples was not attempted, however, one example was found. Steele (1986b) examined the vertebrate remains recovered from a Late Prehistoric site (ca. A.D. 1250-1500) and identified four snake genera, one being a Nerodia (sp. indet.). He noted that five aquatic taxa, including the water snake, indicated the presence of a nearby stable aquatic habitat. THORNTON & SMITH 303 If colubrines and crotalines have consistently been reported as probable, if not actual food items, why then has Nerodia been noticeably absent. The close proximity of Prehistoric campsites to intermittent and perennial streams draining this xeric region, in conjunction with an abundance of mussel shell accumulations, clearly indicates the utilization of riverine resources. If natracines occurred here at this time, as the paleoherpetofaunal evidence attests (Holman 1981), and if the Prehistoric peoples were in somewhat of a continuous contact with this riparian habitat, why is one of the most numerous water snakes, N, harteri paucimaculata, (or any other water snake species) not represented in the fossil collections. To answer this question, an analogy between two predator-prey relationships is necessary. First, Parmley (1986) examined a herpetofaunal assemblage from a sinkhole located in the karst terrain characteristic of south-central Texas. The faunal remains were primarily attributed to feeding activities of the Common Barn Owl {Tyto alba), although some may have entered voluntarily or simply fallen into the hole (Dal quest et al. 1969). Parmley (1986) explains the apparent lack of natracine fossils compared to the overwhelming abundance of colubrines fossils by indicating a bias in raptor selection of prey. He suggested that the woodland and grassland forms (especially colubrines) were an easier prey for the owls simply because they often forage in the open. However, aquatic forms such as natracines were usually more restricted to their habitat, hence, more protected. Second, the vast majority of the faunal remains recovered and attributed to food procurement, were diurnal creatures. Following Parmley ’s (1986) suggestion, the analogy between predators (humans or owls) becomes obvious. Furthermore, the herpetofaunal remains recovered in the cultural context strongly suggests warm weather harvesting (Steele & Hunter 1986). As diurnal temperatures increase in the summer months, natracines are known to become more crepuscular or nocturnal. Another important consideration is the medium in which the various snakes escape predation. Although the terrestrial refuges sought by escaping snakes makes it somewhat difficult for hunters to capture their prey, it can be done. Contrary to this, field experience with water snakes reveals that once they enter the water, the chance of capturing them diminishes greatly. It then becomes apparent why colubrines dominate the fossil snake assemblages in ancient campsites. First, terrestrial snakes are more numerous; second, most of the larger forms are diurnal; and third, the hunting area and it’s 304 THE TEXAS JOURNAL OF SCIENCE-VOL. 47, NO. 4, 1995 resources are dominated by the arid landscape. Today in this region water is still a limited resource surrounded by the arid uplands. One could reliably conclude that the faunal assemblage, whether intrusively or culturally deposited, would reflect this ecological reality as well. The actual dietary use of natracines in this region cannot be documented, but can be confirmed for some of the crotalid and colubrid taxa recovered. Ruecking (1953) and Sjoberg (1953) have documented the consumption of rattlesnakes by the Historic Coahuiltecan and Tonkawa Indians. Crotalids have also been documented as food items for Archaic, Prehistoric, and Late Prehistoric Indians (Newcomb 1961; Steele & Mokry 1985; Shafer 1986; Steele & Hunter 1986; Steele 1986a; 1986b; Hellier et al. in press;). Colubrids were consumed as well (Newcomb 1961; Steele & DeMarcay 1985; Steele & Mokry 1985; Shafer 1986; Steele 1986b), however, Steele & Hunter (1986) admit that the ranges in size of the snake vertebrae recovered clearly indicates that a variety of these reptiles were being harvested. Interestingly, from Post- Pleistocene hearths dated ca. 8,000 YBP, 16 species of snake were recovered ranging from large rattlesnakes to very small nonpoisonous snakes (Shafer 1986). The identified genera include Crotalus (prob. C. atrox), Agkistrodon (prob. A. contortrix), Elaphe (sp. indet.), Nerodia (sp. indet.), and Larnpropeltis (sp. indet.). Excluding the unidentified snake species, the percentages of the remaining snakes are: crotalids, 57.1%; terrestrial colubrids, 35.8% ; and aquatic natracines, 7.1%. The percentages of snake taxa recovered indicates a strong preference for crotalids, however, if indeterminate species are included the percentages are: crotalids, 44.4%; colubrids, 27.8%; indeterminate snakes, 22.2%; and natracines, 5.5 % . Presently, 36 taxa of snakes occurs in or near the fossil sites (Dixon 1987). Of these, colubrids represent 83.3% and crotalids 16.7%. Upon examination of the snake remains recovered from 41CN19, 41CC131, and 41RN169, it is apparent that these Prehistoric people were consuming E. obsoleta, Crotalus (prob. C. atrox), Masticophis (prob. M, flagellum), and possibly Coluber {^voh. C. constrictor). This is based upon the observation that E. obsoleta, C. atrox, and M. flagellum occur in the area today, and they are the largest members (length and mass) of the locally occurring terrestrial or aquatic taxa. These snakes are also more commonly found than their con-specifics {E. guttata, C. viridis, C. molossus, and M. taeniatus). THORNTON & SMITH 305 Acknowledgments We sincerely thank Dr. Chris Lintz and Mariah Associates, Inc., for the opportunity and the privilege of studying the snake fossils recovered during the course of the Stacy Dam Archeological Project. Dr. Lintz also reviewed this paper and provided valuable inputs. Gratitude is also extended to Dr. Darrell Creel of the Texas Archeological Research Laboratory for loaning us the snake vertebrae recovered during the Lake Spence archeological investigations. Sincere gratitude is extended to Dr. Dennis Parmley for taking the time to confirm our tentative identifications and for making the final taxonomic determinations of the fossils. Dr. Parmely’s critical review greatly improved this manuscript. The archeological investigations of the Texas Archeological Salvage Project (Lake Spence) were in fulfillment of the terms of Memorandum of Agreement 14-10- 7:931-27 submitted to the National Park Service. The cultural resource studies (Stacy Dam Project) of Mariah Associates, Inc., were permitted under United States Army Corps of Engineers (USAGE) Permit W-N-443-41 -Permit 225 and Texas Antiquities Committee Permit Numbers 609 and 809. Funding and facilities were provided by the Colorado River Municipal Water District, Big Spring, Texas, during the course of this study. Drawings are by the junior author. Literature Cited Carlson, D. L., D. G. Steele & H. L. Bruno. 1982. Archaeological Investigations at the Allison Site (41NU185), Nueces County, Texas. Archaeological Research Laboratory. Texas A&M University. Reports of Investigations No. 1, iv + 80 pp. Collins, J. T., J. E. Huheey, J. L. Knight & H. M. Smith. 1978. Standard common and current scientific names for North American amphibians and reptiles. Committee on Common and Scientific Names. Society for the Study of Amphibians and Reptiles. Lawrence, Kansas. Herpetological Circular No. 7. iv 36 pp. Conant, R. & J. T. Collins. 1991. A field guide to reptiles and amphibians of eastern and central North America. Houghton Mifflin Company, Boston, xx 4- 450 pp. Dalquest, W.W., E. Roth & F. Judd. 1969. The mammal fauna of Schulze Cave, Edwards County, Texas. Bull. Fla. Mus. 13:205-276. Dixon, J. R. 1987. Amphibians and reptiles of Texas. Texas A&M University Press. College Station, xii H- 434 pp. Dowling, H. G. & W. E. Duellman. 1976. Systematic herpetology: a synopsis of families and higher categories. Hiss Publ., New York. 306 THE TEXAS JOURNAL OF SCIENCE— VOL. 47, NO. 4, 1995 118 pp. Driver, H. E. 1969. Indians of North America. The University of Chicago Press. Chicago. Hellier, J, R., D. G. Steele & C. A. Hunter. Analysis of Vertebrate Faunal Remains. In Archaeological Investigations At The Loma Sandia Site (41LK28): A Prehistoric Cemetery And Campsite In Live Oak County, Texas, by A. J. Taylor and C. L. Highley. Texas State Department of Highways and Public Transportation, Highway Design Division, Contract Reports in Archaeology, p. 1223-1285: In press. Holman, J. A. 1979. A review of North American Tertiary snakes. Publ. Mus. Mich. State Uni v. Paleontol. Ser. 1:200-260. Holman, J. A. 1981. A review of North American Pleistocene snakes. Publ. Mus. Mich. St. Univ. Paleont. Ser. 1:261-306. Lintz, C. &W. N. Trierweiler, A. C. Earls, F. M. Ogelsby, M. Blum, P. L. O’Neill, J. Kuhl, R. Holloway, L. Scott-Cummings & D. Scurlock. 1993. Cultural resource investigations in the O. H. I vie Reservoir, Concho, Coleman, and Runnels counties, Texas. Vol. I: Project introduction, setting and methods. Mariah Associates, Inc., Austin, Texas. Technical Report 346-1. Texas Antiquities Committee Permit Number 609. xiii -t- 392 pp. Newcomb, W. W., Jr. 1961. The Indians of Texas: From Prehistoric to Modern Times. Austin: University of Texas Press, xviii + 404 pp. Parmley, D. 1986. Herpetofauna of the Rancholabrean Schulze Cave local fauna of Texas. J. Herpetol., 20(1): 1-10. Parmley, D. 1988a. Additional Pleistocene amphibians and reptiles from the Seymour Formation, Texas. J. Herpetol. 22(l):82-87. Parmley, D. 1988b. Early Hemphillian (late Miocene) snakes from the Higgins local fauna of Lipscomb County, Texas. J. Vert. Paleontol., 23:322-327. Parmley, D. 1990. Late Pleistocene snakes from Fowlkes Cave, Culberson County, Texas. J. Herpetol., 24(3): 274-279. Ruecking, F., Jr. 1953. The economic system of the Coahuiltecan Indians of southern Texas and northeastern Mexico. Texas J. Sci., 5: 470-489. Shafer, H. J. 1971. Investigations into south plains prehistory, west central Texas. Survey Report No. 20. Texas Archaeological Salvage Project. University of Texas, Austin, x -f 174 pp. Shafer. H. J. 1986. Ancient Texans: Rock art and lifeways along the Lower Pecos. Texas Monthly Press. Austin, xiv + 247 pp. THORNTON & SMITH 307 Sjoberg, A. F. 1953. The culture of the Tonkawa, a Texas Indian tribe. Texas J. Sci., 5:280-304. Smeins, F. E. & R. D. Slack. 1982. Fundamentals of ecology laboratory manual. Kendall/Hunt Publishing Co., Dubuque, Iowa. V 140 pp. Steele, D. G. & G. B. DeMarcay. 1985. Analysis of faunal remains recovered during the 1984 excavations at Rancho de las Cabras. Appendix C, In Archaeological Survey and Testing at Rancho De Las Cabras, 41WN30, Wilson County, Texas, Fifth Season, by A. J. Taylor and A. A. Fox. Center for Archaeological Research, The University of Texas at San Antonio, Survey Report 144:62-75. Steele, D. G. & E. R. Mokry, Jr. 1985. Archaeological investigations of seven prehistoric sites along Oso Creek, Nueces County, Texas. Bull, of the Texas Archaeological Soc., 54:288-308. Steele, D. G. & C. A. Hunter. 1986. Analysis of vertebrate faunal remains from 41MC222 and 41MC296, McMullen County, Texas. Appendix III, In The Prehistoric Sites at Choke Canyon, Southern Texas: Results Phase II Archaeological Investigations, by G. D. Hall, T. R. Hester, and S. L. Black. Center for Archaeological Research, The University of Texas at San Antonio, Choke Canyon Series. Vol. 10:452-502. Steele, D. G. 1986a. Analysis of vertebrate faunal remains from 41LK201, Live Oak County, Texas. Appendix V, In Archaeological Investigations at 41LK201, Coke Canyon Reservoir, Southern Texas, by C. L. Highley. Center for Archaeological Research, The University of Texas at San Antonio, Choke Canyon Series. Vol. 11:200-249. Steele, D. G. 1986b. Analysis of vertebrate faunal remains from 41JW8, Jim Wells County, Texas. Appendix VII, In The Clemente and Herminia Hinojosa Site, 41JW8: A Toyah Horizon Campsite in Southern Texas, by S. L. Black. Center for Archaeological Research, The University of Texas at San Antonio, Special Report 18:108-136. 302 pp. Williams-Dean, G. W. 1978. Ethnobotany and cultural ecology of Prehistoric man in southwest Texas. Unpublished Ph.D. dissertation, Texas A&M Univ., College Station. TEXAS J. SCI. 47(4):308-314 NOVEMBER, 1995 OBSERVATIONS ON COMPARATIVE GROWTH RATES AND EARLY DEVELOPMENT IN TWO LITTERS OF THE MEXICAN GROUND SQUIRREL, SPERMOPHILUS MEXICANUS (RODENTIA: SCIURIDAE) Frederick B. Stangl, Jr., Billie J. Brooks and Carla B. Carr Department of Biology, Midwestern State University, Wichita Falls, Texas 76308 Abstract.— Two captive-bom litters of the Mexican ground squirrel {Spermophilus mexicanus) were maintained for a period of 75 days. Body weights, and lengths of hind foot and tail of young animals were recorded at regular intervals. The first litter was characterized by individuals which grew rapidly and at very similar rates. By contrast, the second litter was characterized by an initially depressed growth rate, a partially compensatory terminal acceleration in growth, and considerable variation in individual growth rates. Observations on the timing sequences of eye-opening and juvenile and subadult molt sequences are also presented. Levels of variability in patterns and rates of growth and development suggest that these data are too preliminary to permit a full characterization for this species in regard to these aspects of growth and development. Even for many common and widespread mammalian species, summaries in the literature of reproductive biology, growth rates, and development patterns are often sketchy and anecdotal. Further, if the data are few, the possibility exists of presenting (or perpetuating) an oversimplified view of uniformity of behavior, rates, or patterns. This was found to be true for the fox squirrel, Sciurus niger (cf. Stangl 1993). Little is known of the reproduction, growth, and development of the Mexican ground squirrel, Spermophilus mexicanus (cf. Young & Jones 1982). The basis for general references in the literature detailing these aspects of the biology of this species (e.g. Davis & Schmidly 1994); Schmidly 1977; Young & Jones 1982) appears to be Edwards’ (1946) description of a newborn litter and immature animals of unspecified age. This study details growth rates for two captive-born litters of S, mexicanus, as measured by body weight and lengths of the tail and hind foot. Rates of development are described, and the extent of inter- and intralitter size variation is discussed. Methods Adult females of Spermophilus mexicanus from Haskell Cemetery, Haskell County, Texas, were housed in large glass aquariums lined with STANGL, BROOKS «fe CARR 309 cedar shavings and maintained on a diet of alfalfa pellets and water. Animals were kept captive for 35 days to allow sufficient time for visible manifestation of any pregnancy. Five of six adult females taken on 15 May 1993 had not been impregnated, but a single animal gave birth on 23 May to six young. Neonates were toe-clipped, weighed, and measured for greatest lengths of tail and hind foot. Weights and measurements were taken of neonates when first observed, and at five- day intervals thereafter. Observations on growth and development were recorded on a daily basis. The mother was removed from the cage after 50 days, and released at the pi ace- of- capture. Young animals were sacrificed and prepared as study skins with skeletons at 75 days-of-age. The same protocol outlined above was followed for seven adult females collected on 25 March 1994. The single pregnant female gave birth on 23 April to seven young. One neonate was missing on the second day and presumed to have been cannibalized. Weights and measurements were recorded for neonates, and on days 10, 15, 20, 30, 40, 50, and at 75 days-of-age, when they were sacrificed and prepared as study specimens. All voucher specimens were deposited in the Collection of Recent Mammals, Midwestern State University. Regression analyses, ANOVAs to detect any sexual dimorphism, and t-tests for interlitter comparisons, were computed with NCSS, version 5.03 (Hintze 1990). Results and Discussion Growth rates for litter This litter was comprised of three males and three females. The neonates averaged 6.9 g, somewhat larger than the mean weight of 4.3 g for four animals reported by Edwards (1946), although tail and hind foot measurements were similar to his respective findings of 13.9 mm and 8.6 mm (Table 1). Individuals remained docile and were easily managed during data-recording sessions. Growth rates for each character of each age category were comparable among litter- mates, providing uniform patterns of growth (Fig. 1). The most striking example of individual variation was noted in weights of neonates ranging from 5. 2-8. 3 g, although coefficients of variation (CV) steadily de¬ creased with age (from 17.41 to 2.76; Table 1). Increase in tail lengths and body weights were almost imperceptible after day 60, although hind foot measurements achieved adult dimensions by day 50. While other workers (Cothran 1983; Yancey et al. 1993) have described sexually dimorphic aspects of the cranium, no variation by sex was noted in this study for any character at any age category (one-way ANOVAs, P > 0.05). 310 THE TEXAS JOURNAL OF SCIENCE— VOL. 47, NO. 4, 1995 LITTER 1 LITTER 2 AGE (days) (days) Figure 1. Regressions for comparative growth rates of weight (g), and lengths of tail and hind foot (mm) from two captive-bom litters of Spermophilus mexicanus from Haskell County, Texas. Growth rates for litter Five males and one female comprised this litter. Compared to the first litter, these animals became progressively wild and aggressive during handling sessions, and no attempts were made to weigh or measure living specimens after day 50. This litter showed extensive individual variation, and individuals were smaller and slower-growing than their litter one counterparts (Table 1). Initially depressed rates of weight gain and tail growth were followed by a period of rapid recovery of weight and tail dimensions between days 50 and 75. This was especially noticeable for two males, whose stunted growth rates were largely responsible for the litter’s consistently lower Two-tailed T-test: * 0.05 > P >0.01; ** 0.01 > P >0.001 STANGL, BROOKS & CARR 3i: 00 VO VO p\ io ^ O Ui O U) K) 'P VO b ^ to to to ON to 4^ Ui 00 4i^ .■^ l/i b ^ OV Os 'O -o Ul >— On -Fi- ^ U> OJ ON U> b ^ 4^ CZ) ^ to in °® rs. to . O to O . U» -O L/\ )D y* u) b S^l+ n b ON to / — ^ "O 4^ )0 to 1/1 00 00 I_f- b to VO 4i. to L/1 b ^ to * * VO O .0° b'^ --I £'+ 1/1 2; ^ tOl Oi 00 to n b ?° 'T'H- b ON ^ to VO Ov Ui -o ON b H- ^ b 1+ --J ^ l-t- ^ tJ Q o VO 9 1/1 b2:i+ oj ’ll; to o b tJ n ^ -o to to 4^ U) ■ ^1+ sft- b ^6 to ^1+ o) r U) 00 to ^ 1/1 OJ 0^1+ 'jfep 00 4i^ OJ -V 'O u. ^ H- 00 7 § O i. 1+ U) OO L 1+ U) -N ^ 4^°1+ 00 isv ' ' rv ' ' 1 to;^^ o£;to 'Oil^^to Ov ON U) ^ to ^1+ 00 VO to V/1 Ov Stj to o oj n b P'/’l-f- VO Ul H- ON O to bv b ^ 00 z 00 to d o b ^ .n o toi ON 00 V^I+ 4S>■ 00 H- ^ b I— VO ^1+ ov >— b I— ^ oo 00 4^ to 03 ^ -O 4^ b l-f 00 bo b b'^ .3 + 2. 24-30 8.86 b b 1+ 4^ Ob p b 4^ ^ 4:- bo c/1 Ui to to b /— s ^ b — M u) o 1+ feooP 00 Sp iiS o 8t ns ON '' — ^ 00 ^p\+ VO OJ bl-h V/J ^ b tJ ^ 'O 4^ N b b Ov to b S'+ VO 00 ^ -O -o ;i: b b ' 1+ 4^ r to ON O OJ S b VO y^ ooi+ )SbP ovs^y v/i to ■o ^ b 9^1+ 00 o bo to > 00 cv ol ft: o; I ^ n?i