48 Nr. 9 EN HI S STITUT BOLIVIA GS MONOGR de ileal taht ered a me german a nn N ee eet tn ee ee LP RETNA AN FOI ARMIN ERATOR AM Yat FING NOIRE CAE RR ty yt re BONNER ZOOLOGISCHE MONOGRAPHIEN Die Serie wird vom Zoologischen Forschungsinstitut und Museum Alexander Koenig herausgegeben und bringt Originalarbeiten, die für eine Unterbringung in den „Bonner zoologischen Beiträgen“ zu lang sind und eine Veröffentlichung als Monographie rechtfertigen. Anfragen bezüglich der Vorlage von Manuskripten sind an die Schriftleitung zu richten; Bestellungen und Tauschangebote bitte an die Bibliothek des Instituts. This series of monographs, published by the Alexander Koenig Research Institute and Museum of Zoology, has been established for original contributions too long for inclu- sion in „Bonner zoologische Beiträge“. Correspondence concerning manuscripts for publication should be addressed to the editor. Purchase orders and requests for exchange please address to the library of the institute. L’Institut de Recherches Zoologiques et Muséum Alexander Koenig a établi cette série de monographies pour pouvoir publier des travaux zoologiques trop longs pour étre inclus dans les „Bonner zoologische Beiträge“. Toute correspondance concernante des manuscrits pour cette série doit étre adressée a l'éditeur. Commandes et demandes pour échanges adresser a la bibliotheque de l'insti- tut, s. v. p. BONNER ZOOLOGISCHE MONOGRAPHIEN, Nr. 48, 2000 Preis 61,- DM Schriftleitung/Editor: G. Rheinwald Zoologisches Forschungsinstitut und Museum Alexander Koenig Adenauerallee 150-164, D-53113 Bonn, Germany Druck: JFeCARTHAUS, Bonn ISBN 3-925382-52-6 ISSN 0302-671 X AMPHIBIAN DIVERSITY IN BOLIVIA: A STUDY WITH SPECIAL REFERENCE TO MONTANE FOREST REGIONS by JORN KOHLER BONNER ZOOLOGISCHE MONOGRAPHIEN, Nr. 48 2000 Herausgeber: ZOOLOGISCHES FORSCHUNGSINSTITUT UND MUSEUM ALEXANDER KOENIG BONN Die Deutsche Bibliothek — CIP-Einheitsaufnahme Köhler, Jörn: Amphibian diversity in Bolivia: a study with special reference to montane forest regions / Jörn Köhler. Hrsg.: Zoologisches Forschungsinstitut und Museum Alexander Koenig. - Bonn : Zoologisches Forschungsinst. und Museum Alexander Koenig, 2000 (Bonner zoologische Monographien ; Bd. 48) Zugl.: Bonn, Univ., Diss., 2000 ISBN 3-925382-52-6 CONTENTS Page Kinoducbiom ne een ws ine ta cr arian tame kh Rat 5 Biodnyensiyae ee a ee a N) Bolwaar ra mesadiversityicountiy 0.0.0 neuen if Ampliolans' a ee i ee ae Na ae RN 8 Investigations on Bolivian ph bias = AMI StOnICAl VIEW 2.2.3. . 2er... 12 Wine cine stole Ee study 2 Se el ae tlh ee ale ak 15 eweseou similar studies in the Neotropics 3... ye ee os es 17 Sy TER 45 yc aed TEE RR er oe RE EESEN 18 BOING, ee elt oO Rema ea Re N DOC ea ree 18 Gemstell NL nie ree ONE ent en RE tak, ce, 18 SORTIERE DAL EEE IDEE. RES ER 20 Alla. de 3 arias oe RR 20 We CEPAULOME ze CORCRIONSY ur elle. 22 FEINE hc ET RE EA ie Rr Cae ea oO 24 IN BOE RCOMSCIV al ons pres an ee en ee 29 Mine SUI ALE CES IES en un en se ri izle 29 Diatenalsandemeinodse., sus. ee Nr 38 Fas) Gy (Ot eee een ee ER hE 38 Sy Ma OMIT MIAN ENO OS elem susanne IE 38 Tear AeA Ain eet cent Ashore ot SS Re clea eG Bekoveenty aagliita D4 40 ASSOCIRUSG! CEE VETIN as rare Abe een 40 IeparavlonvOl VOUCKET SPECIMENS). ahs an. een ee 4] EXONOmV re aa ar Oe ee Red fr 42 We Cle SMG MUI CAUIONS + ssh ae penn ee 42 Nlomenelaiiiea ee RR 44 Bioacoukliet Er re ane iey sell d analy ENR eal ahd Oe etc Dats 44 REGIE Se Ne NE EUR RE eh Sees 4 Samplınes analysis, and presentation =. a. 2.2 ec. eae eee G nee a 45 @algdesenip ons an are PMA Peet end near tgs ee oes Noam SUNeH ER 45 Imelussonseleiiterature-dalay er. 0... ice oo eM Ca ee alone sh 46 Barsimonyzanalysisofendemism "2.2 Yi cae eee ee en oe te a 47 NeiehboBjolmıme analysis ol endemism 2.2.2... wessen 48 |Liimittafntom ob N tele ie Bee ute ca ucmaye cue aroeManay opens 48 RISSE IS a east ee Huee 49 Brelimmaryzchecklist and distribution: .............. nennen. 49 7siinotatfonsitorthe checklistn. ee) can sieges ees Eves ete wis 59 ddistonsstoschezlist re ee else 39 Delcuonsstrom.cherlistner ee ae 60 Spceiestcomplexeser a nn exe i 61 Resurrections from Synonymy ....... = u. 232: 22. 0 = eee 62 Unnamed species 25 2.0. 2. ua kann ee Ses Ge 65 Miscellaneous notes - taxonomic problems - ... „u... 2.2 eee 65 Species predicted to oceur ın Bolivia 2.......2.... ae 69 Species diversity and endemism in Bolivia ..../......:27) wer 69 Taxonomic diversity 3.8.0. 002 00.22 bao nn ae 69 Spatial patterns of species diversity and endemism =. . +25. 222 ere 70 Diversity and distribution in montane forest regions ...... 22a 9 Species accounts 2.40040 ea eee es 2 Eee 79 A-transect model 2.0... sind. 0.00. 300 sense ee oe 146 Ecological comparisons 22.2.4 645. 202.222 222 2 er 154 Comparisons of diversity and endemism using PAE and NJAE ...... 162 Large scale distribution patterns of montane forest species .......... Il Discussion. „inneren ee een Re ON e 1707 The.degree of amphibian diversity in Bolivia .......22 2. 9) Segoe 177 Comparison to other’studies „u... ..00.0: 020.5496 es A 177 Ecological determinism . 2... sa¢.2.ce.0esc00 ee eee 181 Recent climate : 2.2.2.2... 0 ne ee A 181 Habitat diversity «. 2.2.0. add en i A 184 Historical perspectives ....20..00..000 00002 u ee 2 ee 187 Recommended conservation priorities .........:...... se ae 193 Future research u... ea ae eee 195 Acknowledgments ..........u. 0240450 see eee 195 SUMMAIY ans ee ee th oe a 197 RESUMED 4 ic odes see an a ae oe ee 200 Zusammenfassung ..... Ged ede seule ve nels ow eels ean ee 202 Literature cited) 2... 800000 00000 bees oe ewes oo oon ee eee 205 INTRODUCTION Biodiversity Life and its extraordinary diversity is the unique wealth which distinguishes the earth from all other planets. Biological diversity in all its aspects represents the foundation of human existence and mankind 1s a critical element of this diversity. We have just become aware of the real dimension of the earth’s diversity in recent years and at.the same time it is increasingly evident that due to the rapid growth of the world’s human population, this diversity is undergoing a dramatic change. The growing recognition and knowledge of the importance of biodiversity has become part of the public awareness of the dual role of biodiversity: as an eco- nomic resource, and as an essential condition for the survival of individuals and biotic communities. It is becoming evident that the loss of biodiversity has seri- ous ecological and economical consequences. As a result, biodiversity is now seen as a critical component of global environmental change. Biodiversity is not equally distributed on the earth’s surface. The most diverse ecosystems are found in tropical countries and certain subtropical areas while the industrialized countries harbor comparatively low biological diversity. This enor- mous contrast between megadeveloped countries and megadiversity countries reinforces us to devote greater intention to establish efficient projects in the fields of research, conservation, and development. One of the most significant interna- tional agreements about conservation, exploration, and sustained use of biodiver- sity is the Convention on Biological Diversity (CBD) of Rio de Janeiro, 1992. This convention regulates the use of biological resources through a fair and well- balanced procedure of benefit sharing and was signed by almost all nations of the world. Terminology Often the term biological diversity or biodiversity is confused with species diver- sity. Biodiversity is far more inclusive and describes diversity in all aspects of biology; that is the morphological, physiological, ethological, ecological, and genetic diversity in populations, species, or higher taxonomic categories. Usually, biodiversity is defined considering three different hierarchic levels (e.g.. Solbrig 1991, 1994): (1) ecosystem diversity is the result of the diversity of abi- otic factors which are available as different combinations of potential resources. Living organisms are related to different kinds and combinations of resources to use them for their reproduction. There are unalterable interactions between these abiotic factors and living organisms. These interactions as well as all ecological processes are elements of ecosystem diversity; (2) species diversity is a substan- tial part of ecosystem diversity. Species diversity is the result of the evolutionary trend to develop an increasing diversity of combinations of genes on the one hand 6 and to ‘freeze’ this different combinations occasionally in distinct units, the species, on the other hand. These units are more or less limited in their ability to exchange genes with other units. Every unit (species) is characterized by the use of a particular combination of resources (ecological niche); (3) intraspecific diver- sity reflects the tendency of life to diversify. The spectrum of intraspecific diver- sity includes minimal differences in the genome to differences resulting in the development of races and subspecies. If these differences result in speciation processes, often depends on external (and mostly random) influences like e.g. geographical isolation. Measuring diversity is discussed controversially. Many authors only consider species numbers (which are easy to determine) within a particular space as index for diversity. In addition, a variety of diversity indices has been developed, con- necting species numbers with abundance (e.g., Pianka 1977, Spellerberg 1991). However, the combination of species numbers and abundance as only two quali- ties out of numerous others connected with species appears somewhat arbitrary to describe the complex patterns of biodiversity. At least, these indices do not answer the basic questions of biodiversity research: “Why are there so many organisms and how do they manage to coexist?” and “Why are there differences in numbers of coexisting species at different places?” For researchers investigating biogeog- raphy and biodiversity, species which are small, rare or less abundant may have the same importance than abundant larger species (see Brown 1988). Moreover, comparison and description of biodiversity will not become more exact because of indices, since not all species contributing to the diversity of a particular region or area are discovered and described. However, this work mainly focuses on species diversity as a value for describing diversity patterns. It is used here in the sense of numbers of coexisting species in a certain space. An important concept differentiates species diversity into three categories (Cody 1986): alpha, beta, and gamma diversity. Alpha diversity equals the number of species existing at one place, beta diversity describes the species- turnover along an ecological gradient, and gamma diversity describes the rate of species substitutions within ecologically similar habitats which are separated by a certain geographical distance. Alpha diversity reflects the maximum density of species, wheras beta diversity gives a value for habitat specialization and ecolog- ical plasticity of species. Gamma diversity is strongly depending on the ability of taxa to evolve and their tendency to endemism. Significance of Research on Species Diversity Research on systematics and taxonomy, largely ignored in an era of genetics and biochemistry (e.g., Butler et al. 1998), is experiencing a comeback under the mod- ern designation ‘biodiversity research’. This answers to an urgent need, for a few years ago it became apparent that only a small percentage of the earth’s diversity 7 in species is scientifically known. Since 1758 - the official beginning of scientif- ic nomenclature - some 1.8 million species have been described. Today, the actu- al species number inhabiting the earth has been estimated to include anywhere from 10 to 100 million species (e.g., May 1992). At the same time, it becomes increasingly evident that the near future will bring species extinction on a scale such as it has occurred only a handful of times in the earth’s history. In other words, innumerable species will become extinct before we will ever have had the opportunity to study and know them. The species that remain hidden from scientific and general knowledge are by no means only inconspicuous insects or worms. Major groups of vertebrates are still unknown. For example, the recent discoveries of new bovid species in the forests of Vietnam (Dung et al. 1993, Peter & Feiler 1994) demonstrate the defectiveness of our knowledge in a spectacular way. Bolivia — a megadiversity country The term megadiversity country was first developed by Mittermeier (1988) in order to stress the importance of those few countries which harbor a major portion of the world’s biological diversity. Of course, animal and plant species are not aware of geopolitical borders but the megadiversity approach acknowledges that conservation is managed at country level (Groombridge 1990). However, Bolivia is still neglected in recent publications concerning megadiversity countries. Moraes & Beck (1992) were the first to propose Bolivia to be included in the list. Bolivia is still one of the least investigated countries of South America and scien- tists unfamiliar with it tend to underestimate its diversity. This diversity is mani- fested at all levels, from its abiotic conditions, called ‘geodiversity’ by Barthlott et al. (1996), to the hierarchic biological systems which exist within its borders. The fact that Bolivia actually is a megadiversity country was stressed by several authors in a book edited by Barthlott & Winiger (1998) which resulted from a con- gress on biodiversity with main focus on Bolivia. Ecosystem Diversity The ecosystem constitutes an important hierarchic level of biological systems. Only few countries in the world can match Bolivia’s ecosystem diversity (see Fig.4). There is probably no other tropical country which has access to as many biogeographical regions and biomes as Bolivia. In his scheme of tropical vegeta- tion, Lauer (1986) classifies zonal vegetation according to hydrothermical units. Almost all of them, from desert to rainforest, from hot lowlands to glaciers, can be found in Bolivia. Furthermore, there is a complex differentiation between zonal and azonal ecosystems. For example, Ribera (1992) lists more than 40 different ecoregions for Bolivia. 8 Species Diversity Species are a second important level of biological systems and although it is not possible to estimate species numbers for many groups of Bolivian organisms, Bolivia undoubtedly contains extraordinary high levels of species diversity. This high degree of species diversity is due to Bolivia’s abiotic diversity in space and time. All the historical, geological, orographical and climatic processes are responsible for the richness at species level we can find today on Bolivian territo- ry. All factors and mechanisms which stimulate and accelerate speciation or guar- antee the maintenance of high species diversity, such as habitat heterogeneity, extinction-buffering long-term stability, isolation and local medium disturbances are active in Bolivia (Ibisch 1998). Moreover, its geographical location facilitates the immigration of very different biogeographical elements (see also discussion). Genetic Resource Diversity All organisms which are actual or potential providers of resources for human life are genetic resources. The diversity of Bolivia’s genetic resources can be assumed to be tremendous. Genetic resources include stable foods and medicines, con- struction material and clothing. Bolivia has dozens of cultivated and wild plant species which are of enormous importance for global food security (Cardenas 1989). As an example, one of the most important stable foods worldwide, the pota- to, originated in the high Andes of Bolivia and Peru. Currently, 38 species of pota- to with hundreds of local varieties are cultivated by local farmers. In the Departamento Santa Cruz alone, Vazquez & Coimbra (1996) identified 130 edi- ble fruits of wild plants, at least ten of them with high export potential. Ethno-Cultural Diversity Beside these kinds of diversity, Bolivia is especially rich in ethno-cultural diver- sity which in a broader sense is also part of the biological diversity. Today, vast portions of the Andean region are occupied by Aymara and Quechua people. Ethnodiversity in the Bolivian lowlands is much greater, especially in humid areas which are biologically more diverse. About 30 different cultures that developed special adaptations to their different natural resources can still be found in the Bolivian lowlands (Libermann 1995). Amphibians Amphibians inhabit a variety of life zones. With the exception of oceans they can be found from deserts to the subpolar region, from sea level to snow line, every imaginable type of freshwater, from the ground up to the highest treetop. Amphibians seem to have once again reached a level of diversity comparable to their first “golden age”, the Carboniferus and the Permian. Equipped with lungs 9 and limbs, they were the first vertebrates to leave the water in the Devonian some 350 to 360 million years ago to conquer the land masses previously uncolonized by vertebrates. During this process they developed an enormous diversity in forms and species, including representatives of several meters in length and with very lit- tle resemblance to present-day amphibians. The fossil evidence of Paleozoic amphibians ends with the Trirassic, but amphibians did not become entirely extinct. The first frog-like creature (Triadobatrachus massinoti) appeared in the Triassic in Madagascar, still sporting a short tail, but already displaying first signs of a saltatory mode of life. In the Jurassic, diverse recent frog families already existed such as tongueless frogs (Pipidae) and disk-tongued frogs (Discoglossidae). Salamanders and caecilians followed in the Cretaceous. From that time on, amphibians experienced a second golden age that lasts into the pres- ent days. Species Diversity Amphibians need freshwater. Thus, it comes as no surprise that the moist envi- ronment of the tropics is home to their greatest diversity. High temperatures and the constant access to water in form of precipitation create ideal conditions, and the number of amphibian species increases the nearer one gets to the equator. But actually the situation is far more complex and should not be seen just against the background of current climate conditions. Instead, past climate oscillations and orographic changes must also be considered, since they created environments that were hostile to amphibian survival or, on the other hand, caused the extinction of many species. Today, more than 5100 extant amphibian species have been described (Glaw et al. 1998a) and the number increases at a yearly rate of approximately 70 to 100 newly discovered species (Glaw & Kohler 1998). The rate in discovering new amphib- ian species has never been as great as it is now. Since 1994, the number of known recent amphibians is greater than that of mammals and the biggest burst in new species descriptions is not yet reached (Glaw & Köhler 1998). With approximate- ly 4500 valid species, frogs (Anura) are by far the most species-rich amphibian group, only one tenth (450 species) are salamanders (Urodela) and a mere 165 species count among caecilians (Gymnophiona). By far, most of the new species described in the recent years originate from the Neotropical region (68%; Glaw & Kohler 1998, Glaw et al. 1998b). Of course, cur- rent species descriptions do not necessarily reflect the true status of existing species diversity, because different levels of research intensity have to be considered. Research may be more intensely focused on Latin America than on Africa and Asia. Nevertheless, current studies show that there is still no end to be seen in the discovery of new amphibians in tropical South America. Therefore it is probable that amphibian species diversity actually is concentrated in the Neotropics (Köhler et al. 1998a). Nearctic 2% Ethiopian 4% 2 IN BOoEHB 2,73 Australo-Papuan 9% Fig.1: Percentages of am- phibian species described from the beginning of eee 1986 to the end of 1995 69% in different biogeographi- cal regions (after Glaw & Köhler 1998). Despite the recognizable increase of investigation efforts during the last decades of this century, recent surveys indicate that actual species diversity in amphibians is still underestimated in nearly every tropical forest. Due to the use of modern techniques such as biochemical and genetic analysis as well as the almost obliga- tory analysis of species-specific advertisement calls in anurans, the real degree of Species numbers becomes more and more evident (see also Hanken 1999). Still another important factor is the research in previously not or only poorly investi- gated areas. For example, Pethiyagoda & Manamendra-Arachchi (1998) suggest- ed the actual number of frog species inhabiting Sri Lanka to be more than 250 instead of the 54 species recognized in the current literature. Similar cases can be found in other regions such as for example Madagascar (Glaw & Vences 1994, Glaw 1999), Vietnam (e.g., Inger et al. 1999), or Bolivia (De la Riva et al. 2000). Natural History Few vertebrates, with the exception of fishes, are as dependent on environmental moisture — usually in the form of precipitation — as amphibians. The geographic range, ecology, behavior, and natural history of amphibians is strongly influenced by the distribution and abundance of freshwater. As a result, the spontaneous and often synchronized breeding of several species of frogs with the first rainfalls is a well-known phenomenon, especially in areas where rainfall is strongly seasonal. On the other hand, the multitude of other ecological factors which interact to affect amphibian activity and life history is poorly understood (McDiarmid 1994a). Amphibians may occur in terrestrial, aquatic, arboricol, or fossorial habitats, or in a combination of those. In most cases, only little is known about the species’ nat- 1] ural history. Concerning the reproduction, a major part of the anuran species has external fertilization of the eggs and an aquatic development of tadpoles which represent a completely different life form. After a period of growth larvae under- go metamorphosis and move back to a terrestrial environment where they devel- op into mature adults. Others undergo direct development, that means, they lack an independent larval stage. The eggs hatch into nonfeeding larvae or small froglets. A few forms are ovoviviparous or viviparous combined with internal fer- tilization. However, within this generalized modes of reproduction several vari- eties in reproductive efforts evolved which can be interpreted as adaptations to different environmental conditions. These include different modes of parental care like for example attendance to egg clutches or juveniles, feeding of tadpoles, and back pack carrying of eggs, larvae, or froglets, as well as production of foam nests or development of the larvae inside the stomach or vocal sac of one of the parents. An overview about reproductive modes and parental care in anurans was given by several authors (e.g., Crump 1974, McDiarmid 1978, Duellman & Trueb 1986, odI1990)._- Caecilians (Gymnophiona) are aquatic or fossorial and thus difficult to sample. Due to their secretive habits, very little is known about their life history and ecol- ogy. Male caecilians have a protrusible copulatory organ, the phallodeum, and presumably fertilization is internal in all species. Most caecilians seem to be vivip- arous, although some are ovoviviparous. Salamanders (Caudata) are mainly distributed in the Holarctic region, but a major radiation of plethodontid salamanders with direct development has evolved in the Neotropics. Salamanders display a variety of courtship patterns and reproductive modes. Most groups have internal fertilization without copulation, but few large species have external fertilization. Eggs of aquatic species are laid singly, in strings, or in clumps in ponds or streams, sometimes beneath stones or attached to vegetation. These species have aquatic larvae which usually metamorphose and move back to a terrestrial environment. As adults they return to aquatic environ- ments for reproduction. Most plethodontid salamanders are terrestrial or arboricol and deposit egg clumps in moist sites in leaf litter, bromeliads, beneath rocks and logs, and have direct development of the young. Visual and chemical signals appear to be more important for communication than in anurans. In anurans, the most important medium to communicate seems to be acoustic. It has been demonstrated that frog calls have different kinds of functions such as advertis- ing, territorial, or distressing (e.g., Blair 1958, Duellman & Trueb 1986, Hédl & Gollman 1986). Today, the analysis of mating or advertisement calls is almost oblig- atory in some groups as a character to distinguish species which might be morpho- logically very similar. Advertisement calls work as a very effective pre-zygotic iso- lating mechanism. However, it became evident that other forms of communication of visual or even seismic character might play important roles as well (e.g., Harding 1982, Lewis & Narins 1985, Narins 1990, Cardoso & Heyer 1995). 12 The amount of different life forms, behaviors, and reproductive modes evolved in amphibians as well as the variety of habitats used by them is hardly to be exceed- ed by any other vertebrate group. This makes it a challenge for every researcher studying the biology of amphibians. Global Amphibian Decline? The phenomenon of a worldwide decline of amphibian populations has been the subject of several articles published in scientific journals as well as in commercial newspapers. Since it became evident that several populations in different parts of the earth were declining, a discussion about possible reasons began, whether this decline mirrors natural fluctuations in population size or might be caused by human impact (e.g., Pechmann et al. 1991, Blaustein et al. 1994). If they were due to human impact, the question if local or global factors are responsible often remained open (Blaustein & Wake 1990). The possibility of a global phenomenon was seriously discussed, because several populations, especially in Central America and Andean South America, obviously declined although they inhabit apparently undisturbed habitats (e.g., Crump et al. 1992, La Marca & Lötters 1997). Many theories appeared to explain this phenomenon, including the influ- ence of increased ultraviolet radiation, acid precipitation, fragmentation of habi- tats, overcollecting, chemical pollution as a result of volcanic activity, and pathogens (e.g., Morell 1999). Recent findings in southern Central America are alarming. Lips (1997) observed declines of anuran populations in Panama in previously very diverse communities, including findings of dying individuals. Her observations strongly argue for a dis- ease probably caused by a virus. Possibly, amphibians react more sensitive in response to environmental changes than other vertebrates because of their permeable skin and an aquatic stage in their life cycle. This would make them important indicator organisms, but well-man- aged monitoring projects are needed to throw more light on the factors influenc- ing fluctuations in population size. Investigation of Bolivian amphibians — a historical view The very first reference referring to Bolivian amphibians is the description of Hylaplesia picta (= Epipedobates pictus) by Bibron (in Tschudi) in the year 1838, with the type locality Santa Cruz de la Sierra. The second reference is the descrip- tion of Leiuperus marmoratus (= Pleurodema marmoratum) from the Departamento Potosi by Dumeril & Bibron (1841). Some years later, in 1847, the French explorer and naturalist Alcides d’Orbigny published some herpetological results in his volume V of “Voyage dans l’Amerique Meridionale”. In his work, which at this time contained very important botanical, zoological, and anthropo- logical information, d’Orbigny presented some data and illustrations of amphib- ans collected on Bolivian territory. 3 As a matter of fact, the first known regions of Bolivia were the early settled val- leys of the highlands, namely the vicinities of the cities of La Paz and Cochabamba. Early investigations in these areas resulted in several species descriptions late in the century (Boettger 1891, Boulenger 1882, 1887, 1891, 1898, 1902, Steindachner 1892, Werner 1899, 1901). All these publications were of basic taxonomic contents and mainly included taxa from the mentioned valleys and the adjacent Altiplano. A majority of the species described in these publica- tions was collected by P. O. Simons. Only when the vast oriental areas of the Bolivian lowlands became accessible to some explorers, people got an approximate imagination of Bolivia’s fauna. During the first half of this century, various papers contributed to the knowledge of Bolivian amphibians (e.g., Andersson 1906, 1932, Barbour & Noble 1920, De Grys 1938, Dunn 1942, 1949, Eisentraut 1932, Gaige 1929, Mertens 1929, Miiller 1924, Müller & Hellmich 1936, Nieden 1923, Parker 1927b, 1928, 1934, 1940, Procter 1921), mainly publishing results of larger expeditions. Among the most important expeditions covering Bolivian lowlands were the “Swedish Chaco- Cordillera Expedition” (1901-1902) under direction of Earland Nordenskiöld (see Andersson 1906), the American “Mulford Exploration of the Amazon Basin” (1921-1922), accompanied by the herpetologist Everet N. Pearson, and the “Deutsche Gran Chaco-Expedition” by Franz Krieg at the end of the 1920’s (results published by Miiller & Hellmich 1936). However, at these times investi- gations were mostly restricted to regions around religious missions founded by Jesuits in the eighteenth century. Between 1910 and 1950 the German family Steinbach collected many amphibians and other animals at different Bolivian localities, but mainly at Buenavista, Departamento Santa Cruz (type locality of Hamptophryne boliviana, Pseudopaludicola boliviana, and Scinax parkeri). The collected specimens are deposited in various collections (see Ergueta 1991b). In the 1950s and 60s, research activity concerning amphibians was relatively low and several of the publications dealt only in part with Bolivian populations (e.g., Barrio 1965, Bokermann 1964, Cochran 1955, Cochran & Goin 1970, Duellman 1956, Funkhouser 1957, Gallardo 1961a, b, 1965, Lutz 1973, Rivero 1961, Vellard 1951, 1957, 1960). Apart from these works with a different geographic emphasis, some publications were exclusively on Bolivian anurans (Cei 1968, Donoso- Barros 1969a, b, 1970, Gans 1960, Shreve 1959). Among these papers are sever- al descriptions of new species and subspecies which today are treated as junior synonyms of previously described taxa. The 1970s were somehow more fruitful in contributing to the knowledge of Bolivia’s amphibian fauna. Bolivian specimens have been included in revisions of taxonomic groups or particular regions, mainly published by North American her- petologists (e.g., Duellman 1971, 1972a, 1973, 1974a, b, Duellman & Fritts 1972, Edwards. 1974, lynch 1975, 1976, Heyer 1970, 1973, 1977, 1978.) 1979, 14 Silverstone 1976, Trueb & Duellman 1971, Vellard 1970). Subsequently, Charles M. Fugler published some results of his investigations in the northern Departamento Beni (Fugler 1983, 1984, 1985, 1986, 1988) and few other authors provided new taxonomic information (Cannatella 1980, 1983, Lynch & McDiarmid 1987, Wake 1984). In the same period, Gorham (1974) published his list of world amphibians, including data on Bolivian species. However, the first compiled list of amphibian species known to occur in Bolivia was provided by Harding (1983). Although at this time, Harding’s (1983) list rep- resented an important contribution, it contained mistakes and omissions. The next account of Bolivian amphibians was included in “Amphibian species of the world” edited by Frost (1985), and in the additions and corrections to this work by Duellman (1993). The first comprehensive checklist was provided by Ignacio De la Riva (1990a). His list contained 112 amphibian species, distribution data, com- ments on the status of several taxa, first records for the country, a list of species he predicted to occur in Bolivia, as well as for the first time color pictures of many of the species. De la Riva’s (1990a) work was an useful basis for subsequent stud- ies and somehow the starting point of an “investigation boom” concerning Bolivian amphibians. In the 1990s, many more publications appeared dealing with Bolivia’s amphibian fauna. The main part of these articles was contributed by De la Riva (1990b, 1992a, b, 1993a, b, c, 1994a, b, 1995a, b, c, d, 1996, 1998, 199926222 D000: 70 250 Oo oO present work © 200 & = © = 2 = a 5 50 : = = ® = D c 8 De la Riva (1990a) x 20 Ö 5 zo) 5 100 3 o © a 220 Harding (1983) = ay 4 xe) 2 50 5 a 10 = S © © 0 0 © © © © © © © OD © OD © OO W OO) imo comc NO st 0 oO FF © oo oo = NAN © St) oo © © © © D0 © © 9 © 9 DOD © © © Fogo > Tat Sar ze N A TE N = = Na So ae N Nom Saas Soni N > b Fig.2: Figure demonstrating the increase of investigation efforts in Bolivia: Numbers of amphibian species described from Bolivian territory since 1828 (species currently consid- ered as synonyms included), and (€) total number of valid amphibian species known from Bolivia (Harding 1983: 83 species; De la Riva 1990a: 112 species; present work: 200 species). 15 De la Riva & Gonzales 1998, De la Riva & Köhler 1998, De la Riva & Lynch 1997) who focused his Ph.D. thesis on an amphibian community in the northern part of the Departamento Santa Cruz (De la Riva 1993d). His publications includ- ed the descriptions of nine new frog species as well as several first records for the country. De la Riva also was the first who, together with Rafael Marquez and Jaime Bosch, published data of the advertisement calls of many Bolivian frogs (Bosch et al. 1996, De la Riva et al. 1994, 1995, 1996a, b, c, 1997, Marquez et al. 199321995, 1996). Other publications in the recent years include the descriptions of new species (Harvey 1996, Harvey & Ergueta 1998, Harvey & Smith, 1993, 1994, Harvey & Keck 1995, Köhler 2000a, b, Köhler & Jungfer 1995, Kohler & Lötters 1999a, 2000, Köhler et al. 1998c, Lötters & Köhler 2000a, Lavilla & Ergueta 1995a, b, 1999, Reichle & Köhler 1997, Reynolds & Foster 1992), new distribution data (e.g., Köhler 1995b, 1997a, b, Köhler & Lötters 1999b, Köhler & Reichle 1998, Lötters & Köhler 2000b, Reichle & Köhler 1996a, b, Reichle et al. 1997), contri- butions on ecology and/or community structure (Ergueta 199la, 1993, Harvey 1998, Hoogmoed 1993, Ibisch & Böhme 1993, Köhler et al. 1995a, b, Köhler & Böhme 1996, Reichle 1997a, b, c, Reichle & Köhler 1998), as well as an overview about Bolivia’s amphibian species diversity (Köhler et al. 1998b). The populations of two Bolivian species, Telmatobius culeus and T. albiventris (the latter name was placed as a junior synonym of the former by Vellard 1992), were considered to be of vulnerable or endangered status and therefore are listed in the “Libro rojo de los vertebratos de Bolivia” (Ergueta & Harvey 1996). As a result of these most recent publications, the taxonomic status of many nom- inal species was clarified, many taxa were added on Bolivia’s list, and a lot more is known now about distribution and biology than few years before. The increase of studies on Bolivian amphibians is illustrated by the remarkable fact that 31 out of 55 (= 56%) valid amphibian species described from Bolivian territory since 1838 were described in the last ten years (see Fig.2). However, many more species are still to be discovered and their ecology and distribution has to be studied. Objectives of the study Seeing the introducing words about the present state of knowledge of Bolivian amphibians, it is obvious that a study on their diversity, distribution, and biology can be nothing else than preliminary. This is especially true when there are well defined limits in research time, funding, and personnel resources like in this study. New amphibian species are continuously discovered in Bolivia and the checklist could be updated monthly. However, a comprehensive revision of the Bolivian amphibian fauna is not the purpose of this work. Due to the limitations mentioned above, this would need much more financial and personnel efforts than available herein. For example, examining all the amphibian specimens harbored by collec- 16 tions distributed all over the world was beyond the possibilities of this thesis. As a result, this study includes erroneous and insufficient information, and many omissions. Nevertheless, it appears interesting enough to provide an analysis of diversity and distribution patterns at the present state. Although or because new data arise permanently, it seems to be important to draw an integrative and sum- marizing picture early to identify tendencies and relationships as well as deficits. This will at least draw attention to unanswered questions and stimulate further research. The main objective is to provide a preliminary documentation of the amphibian diversity of Bolivia, not only at the regional level but also at a local level. The intention is also to figure the quantitative distribution of diversity and to discuss factors implied. In this study, it is tried to describe amphibian diversity and distri- bution from a more or less synthetic point of view leading to a more entirely understanding of patterns. Generally, amphibian diversity and distribution 1s relat- ed to altitude and latitude and dependent from the amount of precipitation and the degree of temperatures. Therefore, ecoregions were defined as one possible scale to analyze spatial patterns, leading to the questions “How is amphibian diversity and distribution linked to ecoregions?” and “Where can we find the highest degrees in diversity and endemism?” However, ecoregions represent a rather rough scale not adequate to answer the question “How does the degree of diversi- ty change within short distances?” Elevational gradients within the diverse mon- tane forests of Bolivia were chosen as a principal study area to receive insights to the different levels of species diversity. Another purpose is to characterize com- munity structures and to provide new biological data of the investigated species, like for example advertisement calls and habitat use, since they are also part of biodiversity and necessary to understand ecological relationships. Eventually, possible reasons explaining the identified patterns are discussed with biogeo- graphical and historical background. Why a Diversity Study in Bolivia? Bolivia still is the least explored Neotropical country with respect to amphibians (and most other groups of organisms). Despite the remarkable increase of investi- gations in the past ten years, the picture to be drawn is only fragmentary. A further important reason is that almost all of the relevant South American eco-geograph- ical regions are unified on Bolivian territory. Bolivia is an ideal region to investi- gate the change of diversity patterns along ecological gradients. Finally, it appears significant to conduct biodiversity and biogeographical studies at the level of political countries, although they mostly represent artificial and randomly limited areas. Biodiversity research always has also a political dimension. Nowadays, genetic diversity is regarded as a resource of the country and also conservation policy occurs at the country level. Ibe Review of similar studies in the Neotropics Summarizing, there is no similar study to that presented herein hitherto, focusing on general biogeographical patterns within the political borders of one country as well as on diversity patterns within different montane forest areas. By far, most of the studies published on Neotropical amphibians deal with alpha taxonomy or var- ious aspects of a particular species (distribution, physiology, behavior, etc.). Others are about phylogenetic relationships within different taxonomic categories (e.g., Hillis & de SA 1988, Graybeal 1997, Vences et al. 2000). As far as I know, similar studies on diversity patterns, including investigations along altitudinal transects, are now taking place in central Peru, carried out independently by E. Lehr and L. O. Rodriguez (pers. comm.), but the results are not published yet. However, some other categories of publications include at least aspects similar to the contents of this work. The most similar study is probably the one on the distribution of frogs of the genus Eleutherodactylus in the Cordillera Occidental, western Colombia (Ruiz-Carranza et al. 1997, Lynch 1998). In two separate publications the authors briefly described ten sampled transects and Lynch (1998) summarized and discussed the findings of distributions of species and diversity of communities. Although only a single genus was considered, the study provides data for 76 species exclusively distributed in montane forests (Lynch 1998) which represents a greater number of species than involved in the analysis herein. Lynch & Duellman (1997) summa- rized the distributions of Eleutherodactylus species on the Andean slopes of Ecuador. In addition, there are few studies dealing with amphibian distribution along a par- ticular altitudinal transect. Heyer (1967) investigated sites at different elevations in the Cordillera de Tilaran, Costa Rica. Cadle & Patton (1988) published results for vertebrate distributions at the eastern versant of the Andes in southern Peru, also including valuable data for amphibian species. Johnson (1989) focused on biogeographic patterns in southern Mexico, providing information on altitudinal ranges for certain groups. In an unpublished thesis, Franzen (1994) investigated the herpetofauna in the Guanacaste National Park, Costa Rica, including amphib- ian distribution on the slopes of the volcanoes Orosi and Cacao. All these papers provide at least some data usable for superficial comparisons with findings in the present study. Another category of publications is the one dealing with general herpetofaunal distribution patterns in South American. In a book edited by Duellman (1979a), several authors discussed the origin and history of patterns known at that time (e.g., Gallardo 1979, Hoogmoed 1979, Lynch 1979). Subsequently, the same was subject in publications for example by Duellman (1982), Heyer & Maxson (1982a, b), Heyer (1988), and Kress et al. (1998). All of them reflect upon patterns on a large geographical scale, comprising distributions almost all over the sub- 18 continent as a basis for discussing general mechanisms of speciation and disper- sal (see discussion). Other studies mainly focused on amphibian (or herpetofaunal) communities of particular areas with limited expanse. Among these are also long term ones pro- viding valuable data and insight to Neotropical amphibian community structures and distribution patterns. In the following, only some of the most important ones are listed: Martin (1955) — Mexican cloud forest; Stebbins & Hendrickson (1959) — Colombia; Crump (1971) — Belém, Brazil; Duellman (1978c) — Santa Caecilia, Ecuador; Toft & Duellman (1979) — Rio Llullapichis, Peru; Schlüter (1984, 1987a, b) — Panguana, Peru; Heyer et al. (1990) — Boracéia, Brazil; Rodriguez (1992) — Cocha Cashu, Peru; Duellman & Mendelson (1995) — northern Loreto, Peru. Analogous studies were published for Asian (e.g., Brown & Alcala 1961, Lloyd et al. 1968, Inger 1969, Inger & Colwell 1977) and African communities (e.g., Barbault 1974, 1976, Rödel 1996). However, comparisons of the data resulting from research at single sites revealed interesting patterns with respect to distribu- tion, community composition, habitat use, and reproductive modes (e.g., Duellman 1988, 1989, 1990). STUDY AREA Bolivia General Information The state of Bolivia reached its independence from Spain on August 6, 1825. In the following, Bolivia lost more than half of its territory as a consequence of wars (1879-1935) with all its neighboring countries (Argentina, Brazil, Chile, Paraguay, and Peru). At present days, Bolivia’s surface is 1 098 581 km? and therefore it represents the fifth largest country on the South American continent (Montes de Oca 1989). Politically, it is divided into nine departments and more than hundred provinces. Capital is the town Sucre in the Departamento Chuquisaca, but La Paz is the governmental seat as well as the largest city in the country, with more than 1.2 million inhabitants, followed by Santa Cruz de la Sierra and Cochabamba. Bolivia is inhabited by more than 7 million people, with a mean population density of approximately 6 persons/km?. The majority of the human population (70-80%) inhabits the Andean regions, an area constituting 38% of Bolivia’s surface (Montes de Oca 1989). Official languages are Spanish, Aymara, and Quechua. Bolivia contains the highest portion of indigenous people of all South American countries and it is considered to represent the second poor- est country on the continent. Human settlement on the territory of present-day Bolivia started ten to twelve thousand years ago after the last glacial period of the Pleistocene. Approximately 100 years BC, the culture of Tiwanaku erected its center in the Andean highlands near the lake Titicaca. In the thirteenth century, the Tiwanaku culture was fol- Wy) lowed by smaller groups of Aymara tribes existing parallel; subsequently the Quechua speaking Inca overcame the whole territory (15th and 16th century). In some regions, the Inca reached their power shortly before the Spanish conquerors arrived, but they did not manage to include most of the Bolivian lowland tribes in their empire. Later in the 16th century, with the colonization by the Spanish, Bolivia became an important factor in the worldwide growth of economy. Especially the silver mines of the Cerro Rico in Potosi contributed essentially to the richness. of the Spanish empire. As a consequence of the silver exploitation, Potosi became the largest city of the world, larger than Paris or London at that Beni ® Trinidad BIS Aa 7 \ Cochabamba Santa Cruz Cochabamba na \ ® us \_ Santa Cruz de la Sierra Potosi Sn a saca Potosi \, CCIE: nn raya Tarıja PARAGUAY CHILE ARGENTINA 65° Fig.3: Bolivia’s position to its neighboring countries and its political division into nine Departamentos with their corresponding capital cities. 20 time. Since the day of independence, Bolivia was ruled by nearly 70 presidents. There were many armed risings and terror regimes, but since 1982 Bolivia is developing in a relatively stable and democratic way. Today, Bolivia’s economy is mainly based on mining, oil and gas production, cul- tivation of industrial crops (e.g., soy beans, rice, cotton), cattle, and timber extrac- tion (e.g., Ibisch 1998). Another important factor not to be depreciate is the pro- duction of coca and/or cocaine. At a rough estimation, 30% of Bolivia’s gross domestic product comes from the production of drugs (see Müller 1999). Main developmental problems are the impoverishment of the rural population, the migration pressure on cities and unsettled tropical rainforest regions, as well as the destruction of natural environments (for further information see Ibisch 1998). Geography The country is situated between 09°38’ and 22°53’ southern latitude and 57°25’ and 69°38’ western longitude. Highest mountains are in the western Cordillera the Sajama (6542 m a.s.l.) and the Pomerape (6222 m a.s.l.) and in the eastern Cordillera the Illampu (6412 m a.s.l.) and Illımanı (6402 m a.s.l.). The Chiquitania mountains reach 1.300 m a.s.l. Some pre-Cambrian outcrops (inselbergs) of the Brazilian shield might reach 500 m altitude. Bolivia also is the place to find the most important watershed on the continent. Approximately 66% of the country’s surface belong to the Amazon river system, with the large rivers Beni, Guaporeé, Madre de Dios, and Mamore. Other 21% of Bolivia’s surface are part of the La Plata river system (rivers Pilcomayo, Bermejo, Paraguay, Parana), and the rest is part of the Altiplano water system. Generally, Bolivia can be divided into the following physiogeographic regions which are characterized by different geomorphological and historical conditions: (1) the Altiplano which is limited by the (2) western Cordillera, and the (3) east- ern Cordillera. The (4) sub-Andean regions including the inter-Andean valleys are a transition zone to the (5) eastern lowlands. In the east, the lowlands meet the (6) Brazilian shield. This contact zone is interrupted by the (7) Chiquitania mountain chains (Montes de Oca 1989). The final uplift of the Andes took place five to three million years ago. This event was accompanied by drastic changes in climatic conditions. In the quaternary, geomorphological processes were strongly influenced by the cycles of glacial and inter-glacial periods, resulting in changes of temperatures and humidity which accounted for different amounts of glacier covering of the Andean region. For an overview of Bolivia’s geomorphological history and its geoecology see Hanagarth (1993) and Hanagarth & Szwagrzak (1998). Climate According to the definition of the tropics by Lauer (1975), Bolivia is a tropical country without thermal seasons. As a result of the Andean uplift, Bolivia contains 2] warm and hot lowland tropics as well as cool and cold highland tropics. Additionally, the hygric differentiation is very complex and results in a high diver- sity of tropical ecosystems from very humid to arid. Due to its location in the cen- ter of the South American continent, Bolivia is the only country that has equiva- lent portions of Amazonian rainforest vegetation, Cerrado formations, Chaco dry- forest, as well as the climatic highly diverse Andean region. It is the unique loca- tion within an area of different climatic and biogeographic transition and contact zones which accounts for Bolivia’s diversity (Solomon 1989). With the help of pollen analysis, it was shown that the vegetation of the Andean highlands experi- enced drastic vertical dislocations during the Pleistocene (e.g., Graf 1994). At the climax of the last glacial period 18—19 000 years ago, puna vegetation was locat- ed 1000 m lower than today. Mean annual temperatures at that time were approx- imately 7°C lower, but the amount of annual precipitation was about 50% above the values of present days. Vast areas of the Bolivian Andes were covered by gla- ciers, the snow line was situated at 4600 m a.s.l. (today 5200 m a.s.l.), and the for- est line at 2000-2500 m a.s.l. At the maximum of the last Pleistocene inter-glacial period, mean temperatures were approximately 2°C higher than today. Since 5000 years, the phenomenon of “EI Nifio” is existent. “EI Nino” periodically causes extreme climatic conditions resulting in less precipitation in the Andes of Bolivia during the rainy season (“El NiAo-Southern-Oscillation”). During the last ice-age, temperatures in the Bolivian lowlands were 3-4°C lower than today and the amount of precipitation was reduced. Humid rainforests had a more restricted distribution, but probably were not replaced by completely forest- free formations like postulated in the theory of Pleistocene refugia (e.g., Haffer 1969, Brown 1982, Bush 1994, Vanzolini & Williams 1981). In a more recent the- ory, the main presumption is that regions with extremely stable ecological condi- tions remained in times of drastic climatic changes (“Ecologically Extremely Stable Areas - EESAs”; Fjeldsa 1995, Fjeldsa et al. 1999). These regions do not have to be forests, they only have to guarantee the survival of pretentious species (Fjeldsä 1995). At present days, Bolivia’s climate is very diverse and depending on different degrees of altitude and humidity (Lauer 1986). The mean temperature decreases with increasing altitude (0.5—0.6°C/100 m). The temperature dependent altitudinal zones have been classified into Tierra caliente, T. templada, T. fria, T. helada and T. nevada (e.g., Lauer & Erlenbach 1987). Within the Bolivian Chaco close to the Argentinean border the hottest spot of the continent is located, with temperatures reaching 48°C (see Spichiger & Ramella 1989). Periodically, cold southern winds from Antarctic regions (“surazos”) have important climatic influences. They are most common in the dry season in the middle of the year and might result in a drop of temperature below 3°C. These temperature droppings reach the northern savan- nas of the Beni (Hanagarth 1993). 22 = The convective tropical climate results ın a decrease of steam contents in the air with increasing altitude. Due to cool downs, the steam content increases stepwise. In the eastern versants of the Andes, two important condensation levels can be observed. The first is below 2000 m a.s.l. and the second above 2700 m a.s.l. At the second level, a broad bank of fog (or clouds) is usually present. A maxımum of precipitation can be found in lower montane rainforests of the Yungas de Cochabamba region at approximately 1500 m a.s.l. Annual precipitation in that region can be expected to be more than 6000 mm. The Yungas of La Paz are some- what less humid (ca. 3000 mm estimated). The western part of Bolivia is dry due to the influence of the cold Pacific Humbold stream (like western Chile and Peru). Generally, the situation concerning amounts of precipitation is very complex with- in the Andean region, mainly influenced by high mountain chains forming water- sheds. Precipitation in the northern Bolivian lowlands (1700-2000 mm) increases from the northeast to the southwest, parallel to the Andean slopes (Killeen 1998). In contrast to the dryness in the Andean highlands, the above mentioned EI Nifo- Southern-Oscillation causes unusual high amounts of rainfall in the northeastern lowlands of Bolivia (Hanagarth 1993). The western lowlands are remarkably drier, with minima in precipitation in the central Chaco (< 400 mm). The El Nino phenomenon probably did also affect the present study. The rainy season 1997/98 was strongly influenced by the presence of an El Nifo effect. As a result, the first heavy rains in western Bolivia started late (middle of December) and the absolute amount of rainfall was lower than in non-El Nifio years. At the same time, precipitation in the Yungas region probably increased. Vegetation — ecoregions Supposedly, in past times more than 600 000 km? of Bolivia were covered by forests. After data provided by the Worldbank (1994) Bolivia had 556 000 km2 of forest in 1980 and 493 000 km? in 1990. This is about half of the countries’ sur- face, placing it in the ranks of the ten most forest rich countries of the world (rank five or six among tropical countries; Ibisch 1998). Annual deforestation is about 6200 km? which equals 1-2% of the remaining forests. The estimation of forest extent in historical times in the Andes is difficult. According to Kessler & Driesch (1994), 90% of the Andean forests (mainly Polylepis spp.) have been destroyed. The floristic diversity of Bolivia is high. Eighteen to twenty-thousand plant species might occur on Bolivian territory (Moraes & Beck 1992, Ibisch 1996, Beck 1998), among them about 2700 species of trees (Killeen et al. 1993). Generally, the Bolivian flora is still insufficiently known which is illustrated by the large number of new species described in recent times. For example, today more than 1300 species in the most species-rich family, the Orchidaceae, are known from Bolivia (Vasquez 1996) and still more are discovered every year. The ne is a Ecoregions of Bolivia i 000% (ome) 09 00” fo) (o) 69° Amazonian rainforests Campos within Amazonian forests wet savannas humid transition forests humid forests of the pre-Cambrian shield semi-deciduous Chiquitania forests Campos Cerrados Chaco dry-forests (0) (e) Oo {e) Oo (e) Po! 9 (a or ‘ 0.64 0 ons) Polo 30-0 0009, Er Ko) 2 (e) | | ape fo) J \ OnE fo) OÖ He He MT 4 ; Inne: &. 612 59° Chaco montane forests Tucumanian-Bolivian forests inter-Andean dry-valleys high-Andean forests dry Puna “ ~ salt lakes (Salares) cloud forests ("Ceja") humid montane rainforests (Yungas) Fig.4: Schematic map of Bolivia showing its defined ecoregions. 24 number of known Bolivian orchids was only 322 in 1922 (Schlechter 1922) and about 500 were listed by Foster (1958). Several plant groups have their center of diversity in Bolivia (e.g., Cactaceae, Amaranthaceae, Cleistocactus, Puya, Fosterella; see Ibisch 1998). Bolivia’s high diversity of different ecosystems is due to its geographical location. Bolivia is an Amazonian, Andean, Chaco, and Cerrado country. An useful overview of the vegetation of Bolivia and its ecoregions was provided by Beck et al. (1993). More recently, Ibisch (1996) characterized the ecoregions of Bolivia in detail, compiling own and literature data. The following brief characterization of Bolivian ecoregions is mainly based on the data given by Ibisch (1996). Information about conservation areas was taken from Ergueta & Gomez (1997). Chaco dry-forest Located in the Departamentos Santa Cruz, Chuquisaca, and Tarija; also in western Paraguay and northern Argentina; 300-600 m a.s.l.; mean annual temperature 25-26°C; maximum temperature 48°C at the Argentinean border; minimum tem- perature 1°C; mean annual precipitation 400-900 mm; about 1000 mm precipita- tion at the Andean foothills and in the northern transition zone to the Pantanal: 6-8 arid months; low dry-forest of 10-15 m height with various succulent plants; 50-100 tree species; important plant genera Ziziphus, Geoffrea, Ruprechtia, Stetsonia, Cereus; biogeographical relationships to the inter-Andean dry-valleys; land use: timber extraction, cattle; conservation areas: recently funded Parque Nacional y Area Natural de Manejo Integrado Kaa-lya (see Taber et al. 1997). Chaco montane forest Located in the Departamentos Santa Cruz, Chuquisaca, and Tarija; also in north- ern Argentina; 600-1500 m a.s.l.; mean annual temperatures 18-22°C; mean annual precipitation 1000-2000 mm; 6-7 arid months; deciduous forest of medi- um height (< 25 m); important tree species: Schinopsis haenkeana, Astronium urundeuva, Lithraea ternifolia, Zanthoxylum coco; 100-200 tree species; rela- tionships to Caatinga formations; land use: cattle, oil hauling; no areas with con- servation status. Inter-Andean dry-valleys Located in the Departamentos La Paz, Cochabamba, Chuquisaca, Santa Cruz, Tarıja; similar dry-valleys in Argentina and Peru; 1500-3000 m a.s.l.; mean annu- al temperature 12—16°C; maxima above 30°C, minima below 0°C; mean annual precipitation 500-700 mm; 6-8 arid months; (semi-)deciduous dry-forests of medium height (10-20 m); important plant species: Prosopsis spp., Schinus molle, Acacia spp., Tipuana tipu, Schinopsis haenkeana, Erythrina falcata, Kageneckia 2 lanceolata; 100-200 tree species; almost all natural forests destroyed; area of high human population density; problems with soil erosion; conservation areas: only parts of the Parque Nacional Carrasco include small areas. High-Andean forests Located in the Departamentos La Paz, Cochabamba, Oruro, Chuquisaca, Potosi, Tarija; forest type continues in Argentina, parts of northern Chile, and Peru; 2500-4600 m a.s.l. (Polvlepis growth up to 5200 m a.s.l. around the Sajama); con- sidered the highest forests of the world; mean annual temperature below 10°C; temperatures below 0°C relatively common; mean annual precipitation 500-700 mm; 6-8 arıd months; low to medium high evergreen montane forests (5-15 m); most important tree species: Polylepis spp., Baccharis spp., Berberis spp., Escallonia spp., Senna spp.; 10-50 tree species; large parts destroyed; land use: extraction of fire-wood, grazing; conservation areas: Parque Nacional Sajama, Parque Nacional Llica, Reserva Nacional Eduardo Avaroa, Reserva Nacional Ulla Ulla. Semi-deciduous Chiquitania forests Located in the Departamento Santa Cruz in the Provincias Velasco, Nuflo de Chavez, Sandoval, and Chiquitos; unique ecoregion in South America; transition zone between Amazonian rainforests and Chaco dry-forest; relationships to the Brazilian Cerrados; 300-1200 m a.s.l.; mean annual temperatures 18-23°C; mean annual precipitation 1000-1500 mm; 3-5 arıd months; forest of medium height (15-25 m), large parts evergreen; important trees: Cordia alliodora, Terminalia argentea, Astronium urundeuva, Schinopsis brasiliensis; 200-400 tree species; land use: farming, cattle, timber extraction, slash and burn culture; nearly no con- servation status, only the small Parque Nacional Historico Santa Cruz la Vieja. Humid forests of the pre-Cambrian shield (and Campos Cerrados) Located in the Departamento Santa Cruz in the Provincias Nuflo de Chavez and Velasco, and parts of the Departamento Beni; also present in Brazil; 200-1000 m a.s.l.; mean annual temperatures 18-25°C; mean annual precipitation 1500-1800 mm; 2-4 arid months; evergreen forest of 15-30 m height; azonal vegetation on inselbergs and sandstone ridges; important trees: Swietenia macrophylla, Terminalia oblonga, Schizolobium amazonicum, Gallesia integrifolia, Ocotea guianensis; 400-650 tree species; few floristic relationships to the Chaco; land use: timber extraction, slash and burn cultures, gold mining, rubber collection (historically); conservation areas: Reserva Nacional Rios Blancos y Negros, Parque Nacional Noel Kempff Mercado (protects humid forests as well as savan- nas on the Huanchaca plateau). 26 Wet savannas Located in the Departamentos Beni, Santa Cruz, and northern La Paz; 130-250 m a.s.l.; mean annual temperature around 26°C; mean annual precipitation 1000-2000 mm; 2-6 arid months; swamps and grass savannas with few small groups of trees; important tree species: Guazuma ulmifolia, Genipa americana, Rheedia achachairu, Scheela priceps; 200-400 tree species; the southern Beni savannas are closely related with the Pantanal, the northern part is more closely related to Campo Cerrado formations (Hanagarth & Beck 1996); land use: cattle; conservation area: Reserva Biosfera Estacion Biolögıca del Bent. Humid lowland transition forests Located in the Departamentos Beni, Santa Cruz, and Cochabamba; forest type unique to Bolivia; 150-250 m a.s.l.; mean annual temperature around 25°C; mean annual precipitation 1200-1800 mm; 2-4 arid months; evergreen rainforests of 25-30 m height; azonal gallery forests along rivers and wet savannas; important tree species: Hura crepitans, Swietenia macrophylla, Terminalia oblonga, Irartea deltoidea, Bactris gasipaes; 650-800 tree species; close relationships to the moist forests of the pre-Cambrian shield; land use: timber extraction, slash and burn cul- tures, coca plantation; no conservation areas. Tucumanian-Bolivian montane forests Located in the Departamentos Santa Cruz (Prov. Florida, Caballero, Vallegrande), Chuquisaca, and Tarija; eastern Andean slopes south of Santa Cruz de la Sierra; continue south to the subtropical montane forests of Argentina; 800-3000 m a.s.1.; mean annual temperature 13-23°C; mean annual precipitation 1000-2000 mm; 3-5 arid months; montane forest of medium height (< 20 m); important tree species: Blepharocalyx salicifolius, Myrcianthes pseudomato, Cinnamomum por- phyria, Cedrela lilloi, Juglans australis, Sambucus australis, Podocarpus parla- torei, Alnus acuminata; 200-400 tree species; many endemic species for the ecoregion (Argentina and Bolivia); land use: timber extraction, agriculture, oil hauling; conservation areas: Reserva Nacional de Fauna y Flora Tariquia. Amazonian rainforests Located in the Departamentos Pando, Beni, La Paz, as well as forests of the Andean foot in the Departamentos Cochabamba and Santa Cruz (reaching Provincia Ichilo); continue in Peru and Brazil; 100-500 m a.s.l.; mean annual tem- perature 25—27°C; mean annual precipitation 1800-2200 mm; 0-3 arid months; high evergreen rainforest (30-45 m); important tree species of the terra firme for- est: Bertholletia excelsa, Hevea brasiliensis, Couratari guianensis, Manilkara bidentata, Enterolobium contortisiliqum, Mezilauris itauba, Phenakospermim 27 guianensis; trees of the varzea forest: Calophyllum brasiliense, Ceiba pentandra, Ficus spp.; more than 800 tree species; typical Amazonian species are lacking in the forests of the Andean foot; land use: timber extraction, rubber, paranut col- lecting, coca plantation, oil hauling; conservation areas: Reserva Nacional Amazonica Manuripi-Heath, Parque Nacional Madidi, Parque Nacional Isiboro- Secure, Parque Nacional Pilon Lajas, Parque Nacional Carrasco, Parque Nacional Amboro. Humid montane rainforests — Yungas Located in the Departamentos La Paz, Cochabamba, and Santa Cruz; continue to Peru; 500-2500 m a.s.l.; divided in upper montane rainforests (1500-2500 m) and lower montane rainforests (500-1500 m); mean annual temperature 15—24°C; minima below 0°C above 2300 m a.s.l.; mean annual precipitation 2500 to more than 6000 mm; Kessler (1999) suggested a yearly precipitation of 8000 mm in some parts of the Yungas de Cochabamba; maximum rainfall between 1500 and 1800 m a.s.l.; 0-2 arid months; characterized by steep slopes and deep valleys; evergreen montane rainforest of medium height (15-30 m); important plant gen- era: Guatteria, Cyathea, Acalypha, Aniba, Nectandra, Persea, Inga, Trichila, Ficus, Solanum, Oreopanax, Brunellia, Hedyosmum, Clethra, Weinmannia, Clusia, Ocotea; extremely rich in epiphytic plants; 400—650 tree species; land use: coca and Locoto plantations; conservation areas: Parque Nacional Amboro, Parque Nacional Carrasco, Parque Nacional Cotapata, Parque Nacional Pilon Lajas, Parque Nacional Isiboro-Sécure. Cloud forests — “Ceja” Also called “ceja de la montana” which means eyebrow of the mountains; locat- ed in the Departamentos La Paz, Cochabamba, and Santa Cruz; continue in Peru; 2500-3500 m a.s.l.; at the perhumid northeastern versants of the Bolivian Andes, above 3200-3500 location of tree line; mean annual temperature 10—14°C; mean annual precipitation 2500-3500 mm; 0-2 arid months; low evergreen cloud forest (5-15 m); important tree species: Thibaudia crenulata, Gaiadendron punctatum, Persea ruizii, Oreopanax pentalandianus, Freziera spp., Weinmannia spp., Polylepis spp., Escallonia spp.; 50-100 tree species; land use: fire wood extrac- tion, potato and Locoto plantations; conservation areas: Parque Nacional Amboro, Parque Nacional Carrasco, Parque Nacional Cotapata. Fauna The fauna of Bolivia comprises Amazonian, Andean, Chacoan, as well as Cerrado elements. Every of Bolivia’s ecological life zones is inhabited by a typical fauna. There are many transition zones where elements of different origin meet to form special and unique communities. 28 In respect to vertebrates, Bolivia seems to be especially rich in fish and bird species (Ergueta & de Morales 1996). According to Armonia (1995), 1385 bird species are known from Bolivian territory which represent 43% of all South American avifauna (Rocha & Quiroga 1996). Remsen & Parker (1995) assumed that as many as 1088 species of birds could potentially exist within the conserva- tion area of Parque Nacional Madidi (10 000 km?). If this assumption is correct, the area has the potential to become the planet’s richest park for birds and proba- bly for other terrestrial biota as well (Remsen & Parker 1995). Particularly, the eastern slopes of the tropical Andean region is rich in endemic bird species (Fjeldsä & Rahbek 1998). Detailed data on the number of fish species are lacking, but the species number was estimated to be around 500 occurring in Bolivia (Sarmiento & Barrera 1996). Table 1: Knowledge of Bolivian species diversity in selected groups (* = estimates). Number of Known Species Reference Mammals 327 Anderson (1997) Birds 1385 Rocha & Quiroga (1996) Reptiles 229 Dirksen (1995) Fish >500* Sarmiento & Barrera (1996) Vascular Plants 18 000-19 000* Moraes & Beck (1992) Orchidaceae 1330 Vasquez (1996) Today, 327 species of mammals are known to occur in Bolivia (Anderson 1997) which represent about one third of all South American mammal fauna (Hutterer 1998). Several taxa are endemic to Bolivia, including two primates (Callicebus modestus and Callicebus ollalae), two marsupials (Marmosops dorothea and Monodelphis kunsi), and several rodents. The Andean highlands have a very spe- cial mammal fauna including rare species like the vicuna (Vicugna vicugna) and the Andean cat (Felis jacobita). The humid Yunga forests are still home to the Andean bear (Jucumari, Tremarctos ornatus), Mazama chunyi, and many endem- ic rodent species (Tarifa 1996). The Amazonian lowland regions of Bolivia harbor a typical fauna including pygmy anteaters, sloth, primates, cats, tapirs, deer, giant otter, and opossums. The pink river dolphin (Boutu, /nia geoffrensis) occurs in the Madre de Dios, Beni, and Mamoré river systems. The Beni savannas are an impor- tant habitat for the swamp deer (Odocoileus dichotomus) and the rare maned wolf (Chrysocyon brachyurus). Mammal diversity at some sites in the dry Chaco forests is comparable with that at Amazonian sites. The Chaco mammal fauna includes important and endangered species as for example the giant amardillo (Priodontes maximus), giant anteater (Myrmecophaga tridactyla), Chacoan pec- cary (Catagonus wagneri), and the Chacoan Tuco-Tuco (Cfenomys conoveri). About 220-230 species of reptiles were recorded from Bolivian territory (Dirksen 1995, Pacheco & Aparicio 1996). Undoubtedly, this number is far from complete 29 because only very few inventory studies took place concerning reptiles. Four species of crocodiles (Caiman latirostris, Caiman yacare, Melanosuchus niger, Palaeosuchus trigonatus) and 13 turtle species are known from Bolivia. The largest group is represented by snakes with approximately 125 species (Fugler & Cabot 1995). Recently, Dirksen & De la Riva (1999) reported 102 species of lizards from the country. Until today, no estimates on the species number of the little known group of inver- tebrates can, be given. Data and collections are far from complete and well man- aged investigation projects are necessary to seize Bolivia’s invertebrate fauna. Nature conservation Bolivia’s natural richness is protected by 29 conservation areas (listed by Ergueta & Gomez 1997) covering approximately 14% of the countries surface. These areas have different categories of conservation status, for example “Parques Nacionales’, “Reservas”, “Reservas de la Biosfera”, “Refugios de Vida Silvestre”, and “Areas Naturales de Manejo Integrado”. Despite of few private organized reserves, the Direcci6n General de la Biodiversidad (DGB), La Paz, is the respon- sible governmental institution for conservation matters. Only part of these areas really enjoy a managed and controlled protection. The other part only exists on the paper and there are no fundings to fulfill conservation managements. Additionally, in several areas the boundaries are not properly defined. For these reasons, many of the protected regions suffer from human population pressure, slash and burn cultivation, illegal hunting, timber extraction, or gold mining. In recent times, more and more funding for conservation efforts were received from external, non-Bolivian sources and large international organizations like Conservation International and the World Wildlife Fund for Nature (WWF) began to recognize the value of Bolivia’s diverse biota and started initial projects. Until today, vast areas of almost undisturbed ecosystems are still present, and unlike many other countries, Bolivia still has the opportunity to decide how to use and manage its natural resources. Investigated Sites Own investigations on Bolivia’s amphibian diversity and distribution were con- ducted in the years 1994, and 1997-1999, in total comprising eleven months of presence in the country. The itinerar (Fig.5) shows the areas covered by investi- gation efforts. As obvious from this figure, the largest part of the investigations was focused on the humid montane forests in the Yungas de Cochabamba and Santa Cruz regions. Nevertheless, data on amphibians were obtained whenever travelling through the country and repeatedly interesting findings were even made when having stopped the car for a break to relax. Although the main focus of the study was on the Systematic Sampling Survey (SSS), all these collected data are pieces of the present work. Fig.5: Itinerar. Spots indicate sites which were investigated at least four person days. Open squares indicate sites investigated by other herpetologists. Data from these studies were included in the distribution analysis. The regions covered by own studies are roughly the following: moist forest of the pre-Cambrian shield in the northeren Departamento Santa Cruz including granitic rock outcrops (inselbergs) and floating meadows (October 1994); semi-humid Chiquitania forests in northern Departamento Santa Cruz (October 1994); south- ern Beni savannas and humid transition forests west of Trinidad (November 1994); Chiquitania and Chaco formations in the vicinity of Santa Cruz de la Sierra (November 1997/98 — February 1998/99); Chaco montane forests around Camiri (December 1997); inter-Andean temperate-valleys in eastern Departamento Chuquisaca (December 1997); inter-Andean dry-valleys between Samaipata and i Comarapa, Departamento Santa Cruz (November 1994, December 1997, January and November—December 1998); inter-Andean temperate-valley of Vallegrande (January 1998); high-Andean areas around Tiahuanacu and Lake Titicaca (December 1994): dry-puna in the Departamentos Oruro and Potosi (December 1994, January 1999); the inter-Andean valley of Cochabamba and adjacent high- Andean zones (December 1994, January-February 1999); seasonal Amazonian lowland forests around Cobija, Departamento Pando (January 1998); Amazonian lowland forests at the Andean foot in the Departamentos Cochabamba and Santa Cruz (December 1994, November 1997, February 1998, January 1999). For mon- tane forest sites considered more detailed in this study see below. Moreover, data obtained by colleagues in almost all regions of Bolivia, as well as data from museum specimens, mainly deposited in Bolivian collections, became part of this study (indicated by open squares in Fig.5). The sites within montane forest regions investigated more thoroughly during this study are listed and briefly characterized below. Unfortunately, for almost all the sites detailed.data on climate are lacking. So, the annual precipitation given for a site is only an estimation. “Old” Chapare road. — Departamento Cochabamba, Provincia Chapare; this term comprises several sites within the Parque Nacional Carrasco, all located along the “old” road connecting lowland Paractito with Andean Cochabamba. The road lies within a region which is among those with the highest amount of rainfall in Bolivia. Kessler (1999) suggests the yearly amount to be around 8000 mm in some parts of the Parque Nacional Carrasco. It runs on the slopes of the Rio San Mateo valley close to the border of Provincia Tiraque and is in reasonable good condition but not passable above approximately 2250 m a.s.l. due to a large land- slide. Although the road is generally out of use for regular traffic (a new road run- ning more or less parallel west of the “old” road was constructed in the 1970s), it is the only access to a still managed bauxit mine and it is probably also used for the transportation of coca leaves. Only the villages Paracti and El Palmar are to pass when travelling the road. The region is characterized by extremely steep slopes, with slope inclination frequently ranging between 40° and 85° (Ibisch 1996). The sites to be characterized in the following have no available local name and therefore their elevation is used as specification. (1) 500 m a.s.l. — 7 km on road S from Paractito; 17°04’ S, 65°29’ W; investigat- ed 3-4 February 1998; 12-14 December 1998, 2-3 January 1999; Amazonian rainforest at the Andean foot; slightly disturbed (coca plantations); many small to large streams, roadside ditches, no ponds; annual precipitation supposedly 2500-3500 mm. (2) 700 m a.s.l. — 14 km on road S from Paractito, close to the valley of El Palmar; 17°06’ S, 65°30’ W; investigated 6-7 February 1998 and 20-21 December 1998; 32 3000 elevation wet Chapare transect ee (Rio San Mateo valley) 2500 cloud forest 2000 upper montane rainforest 1500 lower montane rainforest o 1000 £ a rainforest of [m Andean foothills 2 = je} ec © =) a F 500 a a © << >= lowland rainforest approximate road km 52 44 37 30 24 132 2620 Fig.6: Schematic profile of the Chapare transect, roughly showing the study sites on an ele- vatıonal gradient. lower montane rainforest; disturbed by coca plantations; many small to large streams, roadside ditches, few artificial ponds; annual precipitation supposedly 2500-3500 mm. (3) 950 m a.s.l. (Fig.7) — 24 km on road S from Paractito; 17°06’ S, 65°34’ W; investigated 4 February 1998 and 19 December 1998; lower montane rainforest; almost undisturbed; many small to large streams, roadside ditches, no ponds; annual precipitation supposedly 3000-4000 mm. (4) 1250 m a.s.l. — 30 km on road S from Paractito; 17°07 * S, 65°34’ W; investi- gated 5 February 1998, 18 December 1998, and 3 January 1999; montane rain Fig.7: “Old "Chapare road at’ SS0Emeasık flooded after heavy down-pour; 4 Feb. 1998. 33 Fig.8: “Old” Chapare road, montane rainforest at 1650 m acs. forests; undisturbed; steep slopes; many small to large streams, roadside ditches, no ponds; annual precipitation supposedly 3500-4500 mm. (5) 1650 m a.s.l. (Fig.8) — 37 km on road S from Paractito; 17°07’ S, 65°35’ W; investigated 16-17 December 1998 and 3 January 1999; montane rainforests; undisturbed; steep slopes; many small to medium-sized streams, roadside ditches, no ponds; annual precipitation supposedly 4500-5000 mm. (6) 1850 m a.s.l. — 44 km on road S from Paractito, 17°08’ S, 65°36’ W, investi- gated 15 December 1998 and 28 January 1999; upper montane rainforests; undis- turbed; steep slopes; many small to medium-sized streams, roadside ditches, no ponds; annual precipitation supposedly 3500-4500 mm. (7) 2150 m a.s.l. (Fig.9) — 52 km on road S from Paractito, 17°09’ S, 65°37’ W, investigated 14 December 1998 and 28-30 January 1999; upper montane rain Fig.9: “Old "@hapare road, border from upper montane rainforest to cloud forest, 2150 m a.s.l. forests; slightly disturbed through mining activities; steep slopes; many small to medium-sized streams, roadside ditches, no ponds; annual precipitation suppos- edly 3500-4500 mm. S of Cuevas (Fig.10). — Departamento Santa Cruz, Provincia Florida; few km by road S of Cuevas; 1300-1400 m a.s.l.; 18°14’ S, 63°41’ W; investigated 30-31 December 1997; semi-deciduous forest; partly disturbed; relationship to the Tucumanian-Bolivian montane forests; few small streams, ephemeral puddles present; 800-1000 mm annual precipitation; area within the influence of “sura- 99 ZOS . El Fuerte (Samaipata). — Departamento Santa Cruz, Provincia Florida; 5 km by road E of Samaipata; 1650-1950 m a.s.l.; 18°10’ S, 63°50’ W; investigated November 1994, 21 December 1997, 27 January 1998, and 8 February 1998; semi-deciduous forest, dry-valley vegetation, as well as elements from humid montane forests; large parts disturbed; small and medium-sized streams present in the area, many ephemeral water bodies of different sizes; 700-1000 mm annual precipitation; for a detailed description of the area see Kohler et al. (1995b) and Ibisch et al. (1996). Fig.10: Bellavista, south of Cuevas, 1300-1400 m a.s.l. >» Empalme (La Siberia). — Departamento Santa Cruz, Provincia Caballero; 31 km by road W of Comarapa; close to the Santa Cruz-Cochabamba border; 2450-2650 mas; 17°51’ S, 64°42’ W; investigated 20 December 1997 and 23-25 November 1998; cloud forest (“Ceja”), parly disturbed; many small streams, numerous roadside ditches; annual precipitation expected to be around 2500-3800 mm; temperatures presumably not dropping below 0°C; usually strong wind from north-east. SE of Guadalupe. — Departamento Santa Cruz, Provincia Vallegrande; 29 km by road SE of Guadalupe; 1650 m a.s.l.,;, 18°39’ S, 63°59’ W; investigated 7-8 January 1998; situated within Tucumanian-Bolivian montane forest, partly dis- turbed; many ephemeral ponds and puddles, one large stream; annual precipitation expected to be around 1500 mm; stronger relationships to the montane rainforests of the Yungas than to the Chaco montane forests; area within the influence of “surazos’’. Fig.11: La Hoyada, 1800 m a.s.l.; montane rainfor- est with Locoto planta- tions in the foreground 36 Incachaca. — Departamento Cochabamba, Provincia Chapare; 2250-2350 m a.s.l.; 17°15’ S, 65°49’ W; investigated 7-9 February 1998; upper montane rain- forest partly influenced by “Ceja” climate; large parts disturbed; some parts cov- ered by artificial conifer forest; many large and medium-sized streams, artificial ponds; annual precipitation expected to be around 2500-3500 mm. Karahuasi. — Departamento Cochabamba, Provincia Carrasco; 1800-2200 m a.s.l.; 17°44 S, 64°44’ W; north of Empalme; at the western limits of the Parque Nacional Amboro and eastern limits of Parque Nacional Carrasco; investigated 3-4 January 1998, and 22-26 November 1998; upper montane rainforest; at lower elevations disturbed through Locoto plantations; relatively steep slopes; all kinds of water bodies available; annual precipitation expected to be 2500-4000 mm. La Hoyada (Fig.11). — Departamento Santa Cruz, Provincia Florida; 1650-1900 m a.s.l.; 17°54 S, 63°08’ W; north of Aguaclara, at the southern limits of the Parque Nacional Amborö; investigated 16-18 November 1998; disturbed humid montane forest, partially logged and cultivated areas, Locoto plantations; many small rivers and creeks, some artificial ponds; valley partly situated on the south- western flanks of the Cordillera Oriental; area within the influence of “surazos’”; annual precipitation expected to be 2000-2500 mm. La Yunga (Fig.12). — Departamento Santa Cruz, Provincia Florida; 2250-2350 m a.s.l.; 18°04’ S, 63°55’ W; north of Mairana, at the southern limits of Parque Nacional Ambor6; investigated 31 December 1997 to | January 1998; upper mon- tane rainforest and adjacent cloud forest; disturbed in the upper parts; small streams but nearly no lentic water; annual precipitation expected to be 2500-3500 mm; influenced by “surazos”. Macufiucu. — Departamento Santa Cruz, Provincia Ichilo; 500 m a.s.l.; 17°44’ S, 63°36’ W; campsite within the Parque Nacional Amboro; investigated 1-3 December 1998; semi-humid lowland rainforests; many streams of all sizes; few swampy areas; relationships to Amazonia as well as to transition forests; annual precipitation expected to be around 1000-1500 mm; influenced by “surazos”. Fig.12: “Va Yungaa(de Mairana), 2300 m a.s.l.; understory of cloud for- est. 31 Mataracu. — Departamento Santa Cruz, Provincia Ichilo; 500 m a.s.l.; 17°33’ S, 63°52’ W; campsite within the Parque Nacional Amborö; investigated 15-19 November 1997 and 16-17 January 1999; humid Amazonian forest at the Andean foot; undisturbed; many small and medium-sized streams, swamps, small ephemeral ponds; annual precipitation around 1800-2300 mm; seasonal climate. Paracti, Rio Roncito. — Departamento Cochabamba, Provincia Chapare; these localities are close to each other on the road connecting Villa Tunari and Cochabamba’ (see Reynolds & Foster 1992 for details); 1600-1950 m a.s.l.; 17°11’ S, 65°47’ W; investigated 9-10 February 1998 and 13-14 January 1999; montane rainforest; partly disturbed along the roads; influenced by heavy traffic; small and medium-sized streams, roadside ditches; annual precipitation expected to be around 2500-4000 mm. Fig.13: Approximately 40 km west of Rio Seco, view from 1000 m a.s.l. to the eastern Chacoan lowlands. 38 Remates. — Departamento Santa Cruz, Provincia Caballero; 2000-2300 m a.s.l.: 17°53’ S, 64°21’ W; at the southern limits of the Parque Nacional Ambor6o, north of San Juan del Potrero; investigated 2-3 January 1998; upper montane rainforest and adjacent cloud forest; small and medium-sized streams, ponds and ephemer- al puddles present; annual precipitation expected to be around 2500-3500 mm; influenced by “surazos”. W of Rid Seco (Fig.13). — Departamento Santa Cruz, Provincia Cordillera; approximately 30 km (airline) west of Rio Seco; 950-1200 m a.s.l.; 18°35’ S, 63°32’ W; investigated 6-10 December 1997; Chaco montane forest; largely dis- turbed through cattle; sandstone formations; islands of forest remnants; small and medium-sized seasonal rivers, few ponds; annual precipitation expected to be 1000-1200 mm; influenced by “surazos”. Sehuencas. — Departamento Cochabamba, Provincia Carrasco; 2100-2300 m a.s.l.; 17°29’ S, 65°17’ W; within the Parque Nacional Carrasco; north of Montepunco; investigated 29 November to 6 December 1994 and 19-20 December 1997; upper montane rainforests; steep slopes; many small and larger streams, many puddles and roadside ditches; annual precipitation expected to be 3000-5000 mm; a detailed description of the area was given by Köhler et al. (1995a) and Ibisch (1996). W of Vaca Guzman. — Departamento Chuquisaca, Provincia Luis Calvo; 13 km by road W of Vaca Guzman; 1340 m a.s.l.; 19°50’ S, 63°49’ W; investigated 15-17 December 1997; large artificial lagoon, surrounded by disturbed Tucumanian- Bolivian montane forest formations; annual precipitation expected to be 1200-1500 mm; influenced by “surazos”. MATERIAL AND METHODS Field Work Sampling Methods Several techniques are available for compiling species lists or information on species richness for a site. The common field techniques are methods of general collecting, as historically practiced by herpetologists. Usually, they involve searching and collecting of specimens in all appropriate microhabitats during both, day and night and result in moderate habitat modification, such as turning rocks and fallen logs or removal of epiphytes. These general collecting techniques have been used for both long-term and short-term sampling projects, and accord- ing to Scott (1994), they are probably the most efficient way to estimate the species richness in an area within constrained time. No other collecting method is as productive in amassing species for a list and in obtaining series of specimens (Scott 1994). 3) The purpose in this study was to obtain as many species of amphibians as possi- ble from a certain site to compare its relative species richness with those from other investigated sites. Because various sites had to be sampled, the available time for research was strongly limited at each site. To approach the necessary Species inventories of the chosen sites, short-term, number-constrained sampling method called Systematic Sampling Survey (SSS) were used (see Scott 1994). This method has been used with birds (Terborgh 1989) and was suggested to be appropriate for tropical amphibian faunas inhabiting forest litter by Scott (1976, 1994). The Systematic Sampling Survey can be used to compare and rank habitats and sites according to relative species richness. This SSS sampling method depends on the validity of the following assumption: more species are present in a limited sample of a species-rich fauna than are present in a similarly sized sam- ple from a less rich fauna (see Hurlbert 1971). The SSS method requires equivalent preselected numbers of specimens (number- constrained) sampled at different sites. According to Scott (1994), samples of approximately 100 specimens may be adequate to rank a series of diverse faunas with respect to species richness. If the site has not been adequately sampled in the investigators’ view, efforts can be concentrated on the collection of additional species (not specimens). During this study, amphibians were searched and collected at the chosen sites by at least two experienced investigators (up to five). Usually, the major habitat types at one site were identified and briefly surveyed during the day. Thereby, it was most important to find possible amphibian breeding sites (e.g., water bodies with egg masses or tadpoles) which were investigated more thoroughly at night. The specimens were mainly encountered by visual sightings and recognition of calling males. During the day, rocks, fallen logs, and bolsters of moss were turned to dis- cover hidden specimens. Small fishing nets were used to obtain tadpoles from their aquatic environment. It was tried to search all suitable habitat types present at one site. Each site was sampled until approximately 100 specimens (a supposedly suffi- cient number according to Scott 1994) were collected or encountered. If it was not possible to obtain the number of specimens required after a prolonged time of searching (e.g., due to dry weather conditions), the site was investigated for a sec- ond period. Therefore, the sampling time at each site varied considerably, from few hours to several days. During searching, some individuals were identified by their advertisement call and were not seen and collected. Usually, only a repre- sentative part of the collected specimens was prepared as vouchers. Limitations: The results from short-term sampling are highly depending on collecting and environmental variables. Possibly, one of the most important vari- ables in sampling amphibian species is the weather during sampling. Using the SSS technique, unfortunate conditions lead to a prolonged sampling time to receive comparable data. Furthermore, SSS enables the investigator to reduce bias 40 in collecting efforts (in contrast to passive techniques such as pitfall arrays which depend on trap location and species susceptibility). On the other hand secretive, fossorial, canopy-dwelling, and deep-water species are more difficult to invento- ry and may require specialized searching methods. In practice, there is usually no time (and/or money) to use specialized techniques to investigate a special habitat type. Therefore, a principally different habitat distribution of frog species among different sites will reduce the comparability of the received data. Finally, with SSS the actual number of species occurring in a defined area will not be estimated accurately. It only enables an investigator to rank sites and habitats according to their relative species richness. For this reason, own results were com- bined with literature data to conduct the Parsimony Analysis of Endemism (see below). Moreover, sampled sites apparently were not always of the same size. Therefore, the comparisons according to species richness have to be taken with some pre- caution, because they are not reflecting absolute values. Biological Data Although the primary goal of the investigation was to obtain data on the amphib- ian diversity in Bolivian montane forests, many observations were made on the species’ biology in the field. These included for example the kind of microhabi- tats used by a species, the calling activity of males, the time of reproduction, the character and size of egg clutches, the kind off egg laying sites, observations on escape and/or defense behavior, predation, and miscellaneous other things. Most of these observations provide valuable new information to the knowledge of cer- tain Bolivian species and it would be somehow irresponsible to withhold such data from any herpetologist. Moreover, these biological data can give at least some insight to the complex ecological relationships within an amphibian community. As a standard, at least following biological observations were noted for every specimen collected: kind of microhabitat used, substrate, general activity, calling activity in males, minimum distance to other calling males of the same species, reproductive state of females, as well as any other remarkable behavior observed. Additionally, it was tried to record the vocalization of every species encountered calling (for methods see below). The advertisement call often significantly helped to identify the species. Moreover, analysis of the recordings subsequently to the field trip may reveal the presence of additional anuran species which were unable to be collected. Associated Data Besides the collecting of specimens and the observation on the species’ biology, certain abiotic parameters were measured. Geographic position was obtained using a Magellan 3000 XL GPS receiver. Elevation above sea level was measured Al with a Thommen altimeter. Air and water temperatures were obtained with a Greisinger GTH 215 digital thermometer (precision 0.1°C). Climatic conditions (e.g., dry, light rain, heavy rain, fog, rained before sampling, wind, etc.) during sampling time were noted, although the absolute amount of precipitation was not measured due to the limited time at each site. Additionally, color slides were taken from the sampled habitats to better remind the general conditions. All data obtained, whether biological or abiotic, were noted in a field book or on a field catalogue sheet similar to that figured by Inger (1994:62). Preparation of voucher specimens After collecting, amphibian specimens were carried in transparent plastic bags. Color slides and notes on coloration in life were taken from living specimens before preparing them as voucher specimens. The specimens were killed in Chloretone solution that was prepared in dissolving a small amount of hydrous chlorobutanol crystals in 0.5 liter of water. This solution was freshly prepared every three weeks when gradually loosing its strength (compare McDiarmid 1994b). Species showed noticeable differences responding to the Chloretone solu- tion. Some died rather quickly (within 2 minutes), others took longer (up to 10 minutes). After death, the completely relaxed specimens were fixed with 96% ethanol in a plastic tray with white paper towels on its bottom soaked with ethanol. Each spec- imen was positioned in the tray in a way that facilitate measurements and exami- nation of key characters. In frogs, the limbs were drawn in next to the body and flexed into a natural position; fingers and toes were straightened and spread to dis- play tubercles and webbing. Since the ethanol does not penetrate the body of larg- er specimens sufficiently within a short time, additional 96% ethanol was inject- ed into their body cavity through the anus. The fixation time depended on the specimens’ size and on the species (e.g., individuals of the Ayla pulchella species group needed significantly more time for fixation). Usually, small specimens were fixed within a few minutes and then had to be removed from the 96% ethanol quickly to avoid desiccation. Large specimens remained in the covered fixation tray for several hours. When fixed, a tag with a field number (JKSL) was attached to each specimen. Each number corresponds to the data for the respective specimen noted in a field book. Among these data are information about the exact locality, date, time, col- lector, habitat, weather conditions, calling activity, coloration in life, and miscel- laneous observations. The fixed and tagged specimens were transferred to jars with 70% ethanol for final preservation. Usually, smaller, more fragile specimens were put into separate small jars to avoid damages, whereas more robust specimens of one locality were preserved and transported together in a larger jar. 42 Tadpoles were killed, fixed, and preserved putting them into 5-8% formalin immediately after collecting (see McDiarmid 1994b). Due to a larger content of water, larvae require a stronger fixative than adults and ethanol is seemingly not adequate. Each tadpole sample from one locality received a JKSL field number that was put together with the sample into the jar. Taxonomy Species identifications Comparison with museum specimens and literature data Correct and accurate species identification is the required basis for every study on amphibian diversity, distribution, and biology. The identification of specimens included in this study was partly based on the comparison of morphological key characters with the data provided in the literature, mainly original species descrip- tions or subsequent revisions of species groups. Additionally, collected specimens were compared with material already deposited in scientific collections, especial- ly with type specimens. Measurements of specimens (for comparisons) were taken to the nearest 0.1 mm using dial calipers. The following account contains all insti- tutions from which material was examined for the purpose of proper species iden- tifications as well as institutions where specimens collected during this study were deposited subsequently. BMNH Natural History Museum, London CBF Coleccion Boliviana de Fauna, La Paz CM Carnegie Museum, Pittsburgh KM Musaeı Zoologici Uniwersytetu Jagiellonskiego, Krakow KU Kansas University, Natural History Museum, Lawrence MNCN Museo Nacional de Ciencias Naturales, Madrid MNHN Muséum national d’Histoire naturelle, Paris MZUSP Museu de Zoologia da Universidade de Sao Paulo NHMG Naturhistoriska Museet Göteborg NKA Museo de Historia Natural “Noel Kempff Mercado” (amphibian collection), Santa Cruz de la Sierra NMW_ Naturhistorisches Museum Wien NRM Naturhistoriska Riksmuseet, Stockholm SMF _ Forschungsinstitut und Naturmuseum Senckenberg, Frankfurt/Main SMNS Staatliches Museum für Naturkunde, Stuttgart USNM National Museum of Natural History, Smithsonian Institution, Washington ZFMK Zoologisches Forschungsinstitut und Museum Alexander Koenig, Bonn ZMB Zoologisches Museum Berlin ZSM Zoologische Staatssammlung München 43 Following type specimens have been examined during the present study: Atelopus tricolor (ZFMK 28103, lectotype), Bufo acutirostris (ZSM 1147/0, holotype), Bufo amboroensis (NKA 953, holotype), Bufo echinodes (USNM 257799, holo- type), Bufo fissipes (BMNH 1947.2.20.64, holotype), Bufo inca (USNM 49557, holotype), Bufo justinianoi (NKA 950, holotype), Bufo pleuropterus (KM 1030, holotype), Bufo simus (NMW 16521, syntype), Bufo stanlaii (CBF 3346, ZFMK 60464, 67096-97, USNM 257796-98, ZSM 144/1999, holotype and paratypes), Centrolenella bejaranoi (KU 182369-71, holotype and paratypes), Centrolenella bergeri (KU 182363-68, holotype and paratypes), Centrolenella phenax (KU 162263-64, holotype and paratype), Centrolenella pluvialis (KU 173224-27, holotype and paratypes), Dendrobates eucnemis (NMW 19190 [1,2,4], syntypes), Protherapis bolivianus (BMNH 1947.2.13.89, —91, lectotype and paralectotype), Hyla aperomea (KU 181812, holotype), Ayla armata (BMNH 1947.2.13.60, syn- type), Hyla callipleura (BMNH 1947.2.13.64-74, lectotype and paralectotypes), Hyla carinata (NRM 1874, syntypes), Hyla leali (KU 92058-59, paratypes), Hyla minima (NMW 19436, holotype), Hyla ocapia (NRM 1873, syntypes), Hyla prasina (ZMB 4675, holotype), Hyla riveroi (CM 37433, holotype), Hyla zebra (MNHN 4817, syntypes), Nototrema bolivianum (NMW 16490, holotype), Adenomera griseigularis (ZFMK 31800, holotype), Eleutherodactylus ashkapara (CBF 3344, ZFMK 70318, holotype and paratype), Eleutherodactylus danae (KU 162307, holotype), Eleutherodactylus dundeei (USNM 50789799, paratypes), Eleutherodactylus lindae (KU 162305, holotype), Eleutherodactylus llojsintuta (CBF 3300-01, NKA 3475-76, ZFMK 66387-89, holotype and paratypes), Eleutherodactylus mendax (KU 173234-35, holotype and _ paratype), Eleutherodactylus olivaceus (CBF 3329-30, ZFMK 67132-33, holotype and paratypes), Eleutherodactylus pluvicanorus (NKA 1100-04, ZFMK 60186-91, holotype and paratypes), Eleutherodactylus rhabdolaemus (KU 175082-83, paratypes), Eleutherodactylus samaipatae (ZFMK 59600, holotype), Eleutherodactylus zongoensis (CBF 2503, holotype), Eleutherodactylus species A (CBF 3341, ZFMK 60402, holotype and paratype), Hylodes cruralis (BMNH 1947.2.15.70, holotype), Hylodes fenestratus (NMW 19940 [1,2], syntypes), Hylodes gollmeri bisignata (NMW 16502, holotype), Hylodes granulosus (BMNH 1947.2.15.72, holotype), Hylodes peruvianus (NHMG 490, holotype), Hylodes platydactylus (BMNH 1947.2.15.91-92, —94, lectotype and paralecto- types), Ischnocnema sanctaecrucis (NKA 1198, holotype), Phrynopus kempffi (NKA 480, paratype), Phrynopus pinguis (CBF 1906-08, 1911-12, holotype and paratypes), Phyllonastes carrascoicola (ZFMK 59569-73, holotype and paratypes), Phyllonastes ritarasquinae (CBF 3350, holotype), Telmatobius huayra (CBF 1223, holotype), Telmatobius jahuira (CBF 1675-76, holotype and paratype), Telmatobius verrucosus (NMW 22922, holotype), Telmatobius yuracare (NKA 511-13, paratypes), Caecilia marcusi (ZSM 79/1982, 83/1982, holotype and paratypes). 44 Additionally, topotypic material of following species has been examined: Bufo castaneoticus, Bufo quechua, Epipedobates pictus, Hyla charazani, Scinax cas- troviejoi, Telmatobius edaphonastes, and Hamptophryne boliviana. Comparison of Advertisement Calls Another important resource for distinguishing species was the analysis of adver- tisement calls. It has been demonstrated that mating calls are an effective pre- zygotic isolation mechanism in anurans (e.g., Blair 1958, 1962, Duellman 1967, Fouquette 1960, Littlejohn 1965, Penna 1997). Every species has its own distinct call which differs from calls of other species. This 1s especially true for co-exist- ing species where barriers in time and space are only insufficiently developed (Hödl & Schaller 1978). In the present study, recorded advertisement calls were compared with already published data and/or with own recordings from other localities and species. Nomenclature Generally, terminology and taxonomic classification follows Frost (1985) and Duellman (1993). In cases where the taxonomic status of a specimen or a popula- tion deserves comments, these are given in the taxonomic account. Some scien- tific species names used herein are incomplete, containing only the generic name with the addition “species A, B, ...”. This refers to populations which have already been identified as distinct species but no species name is available. Most of these new species will be named in the near future and in some cases the description is already in press. If so, this is mentioned in the taxonomic account. A “cf.” in front of the species name means that the specific identification is only preliminary and possibly the populations actually correspond to a closely related but different species. Bioacoustics Recording Unless otherwise mentioned, frog calls were recorded in the field using a Sony WM-D6C professional walkman or an Aiwa HS-F150 cassette recorder, respec- tively, a Sennheiser Me-80 directional microphone, and TDK-MA60 cassettes. No filters or noise reduction systems were used during recording. Gain settings were adjusted manually to ensure that calls were recorded at optimal levels, avoiding clipping or distortion. The distance between microphone and recorded individual varied from few centimeters to several meters and depended on the accessibility of the habitat and call motivation of the individual. Disturbance of calling males caused by the procedure of recording was tried to reduce to a minimum. During recording the air and/or water temperature were measured as close as possible to 45 the calling specimens. The following associated data were spoken on the tape pre- vious or subsequent to the call recording: date, time of recording, locality, specif- ic identification of recorded individual (if possible), calling site habitat, distance to calling individual to be recorded, general weather conditions, other species call- ing in background, and air and/or water temperature. The greatest part of these data as well as the corresponding field number of the voucher specimen were later noted in a field book and in part also on the tape box. Sampling, analysis, and presentation Recordings were sampled with a rate of 22.05 kHz and 16-bit resolution using IBM compatible computers. Analysis of the calls was conducted with the sound analysis software Cool Edit 96 (Syntrillium Software Corporation). The choices of the recordings selected for analysis were based on the certainty of the identifi- cation of the recorded individual as well as on criteria of sound quality. Frequency information was obtained through fast Fourier transformation (FFT, width 1024 points). Temporal information was measured in oscillograms. In some recordings, frequency sections not containing call structures were filtered to remove back- ground noise. Settings for frequency and time ranges and resolutions were chosen according to the essential structural parameters to be measured. A representative audiospectrogram (FFT width 256 points) and oscillogram of a ‘typical’ call is presented in the species account chapter. The figured time segment was chosen to provide as much information as possible on the principal structure of the call. Call descriptions Terminology in call descriptions generally follows Heyer et al. (1990). However, note and pulse repetition rates were calculated within calls or within notes, respec- tively, following method “B” of Scoville & Gottlieb (1978). They were not reck- oned up with call repetition rates like done by other authors (e.g., Marquez et al. 1993, 1995). The experience showed that different call repetition rates may be the result of differences in individual calling motivation, whereas calculation of rep- etition rates within calls (or notes) resulted in a character which is very species specific and independent from individual motivation. In the call descriptions, the range of numerical parameters is followed by the mean and one standard devia- tion in parentheses. To facilitate the understanding, brief definitions of the used terms follow. Audiospectrogram: A visual representation of a call displaying the frequen- cy of the sound over time. Call: An acoustic unit of frog vocalization, may be composed of either identical or different notes; separated from other calls by a period longer than the call; can function alone as an independent vocalization. 46 Call group: Calls may be organized into groups which are separated by long periods of silence; spacing of calls in groups is regular or changing in a pre- dictable pattern. Call (repetition) rate: Number of calls repeated in a defined period of time. The value is provided as calls per minute. Dominant frequency peak: The frequency of the call (or note) at which most sound energy is concentrated. Frequency modulation: Changes in frequency of a sound over time. Frequency range: The frequencies of the call at which at least some sound energy is recognizable. Often, the actual frequency range is difficult to measure, because its representation in an audiospectrogram is dependent from the spectral settings of the used software program, or the call’s frequency range is overlapped by background noise in the frequency analysis. Harmonic: Many sounds have their energy concentrated in several separated, evenly spaced frequencies called harmonics. These frequencies are multiples of the lowest or first harmonic. In pulsed calls, some frequency bands might reflect amplitude modulation generated by the laryngial glottis and is not to be confused with the carrier frequency and its harmonics generated by the vocal cords (see Bradbury & Vehrencamp 1998). Note: Calls are often broken into smaller subunits by 100% amplitude modula- tion with only short intervals between them relative to length of note. A call which is amplitude modulated to 100% is said to be made up of notes; one which is mod- ulated at less than 100% is said to be pulsed. Note (repetition) rate: Number of notes repeated in a defined period of time within a call. The value is provided as notes per second. Oscillogram: A visual representation of a call displaying the amplitude of the sound as it changes over time. Pulse: The smallest named subunit of a call (or a note), produced by amplitude modulation of less than 100%. A note which is modulated to whatever depth is said to be pulsed. A call in which the primary modulation is not 100% is said to have only one note which is pulsed. Pulse (repetition) rate: Number of pulses repeated in a defined period of time within a note. The value is provided as pulses per second. Inclusion of literature data When investigating a certain area within limited time, it is mostly not possible to find all the species that actually occur there. Especially the inhomogenous climat- ic conditions during sampling time may be responsible for obviously incomplete sampling results. Moreover, as mentioned above the used SSS technique is not adequate to estimate the actual number of species occurring in a defined area 47 accurately. For this reason, it was necessary to combine own findings with distri- bution data already published in papers concerning Bolivian amphibians to con- duct the Parsimony Analysis of Endemism (PAE) described below. In some cases, species have not been recollected since their description, but there is no reason to believe that they got extinct. In many cases, there are vouchered records in the lit- erature from sites or elevations which were not accessible during the present study. The information was included in the analysis of general distribution pat- terns, if the records were regarded to be reliable. Parsimony analysis of endemism Although identifying areas of endemism is widely recognized as critical in all methods of biogeographic analysis (e.g., Harold & Mooi 1994), few methods of determining patterns of endemism exist. Phenetic clustering methods have been used to analyze species similarities between sites, but this technique is plagued with problems, with different similarity indices and clustering methods producing different dendrograms. Recently, parsimony analysis, developed for phylogenetic studies, has been used to determine hierarchical patterns of endemism. This method, Parsimony Analysis of Endemism (PAE), was first described by Rosen (1988) and Rosen & Smith (1988), and was later adopted for herpetofaunal analy- sis (e.g., Raxworthy & Nussbaum 1996, 1997, Harvey 1998). Under ideal condi- tions (i.e., when faunas are known completely) this technique produces dendro- grams that link sites on the basis of shared species. Species endemic to one or more areas are treated as apomorphies. PAE resembles cladistic phylogenetic analysis, except that the operational taxo- nomic units are geographic areas rather than taxa, and the characters used in PAE are species distributions. The character state for each species distribution is either present or absent. Shared presence of species provides evidence of biogeographic affinity between different sites, and is used to produce a hierarchical pattern of endemism. PAE reversals either represent species that have gone extinct or were missed during surveys. Dendrograms resulting from PAE might demonstrate historical relationships among the faunas. However, linkage between sites might simply reflect shared environmental conditions that result in colonization by similar faunas. Parsimony analysis of endemism was done using PAUP* (Phylogenetic Analysis Using Parsimony) version 4.0 (Swofford 1998). Heuristic searches were per- formed using the TBR (tree bisection reconnection) branch swapping algorithm. When more than one most parsimonious tree were found, a strict consensus of all trees was calculated. Trees were rooted using a hypothetical outgroup area devoid of all species (see Rosen & Smith 1988). All characters were analyzed unordered, without differential character weighting. No upper limit was imposed on the max- imum number of trees saved. To get an indication of the robustness of the pro- 48 duced topologies, bootstrap analyses were performed (Felsenstein 1985) as imple- mented in PAUP*. This method builds trees based on the same number of charac- ters as the maximum parsimony analysis, but the characters are chosen randomly, and characters are not eliminated from the pool of characters. Thus, some charac- ters will be used more than once for tree calculation, while others will not be used. A total of 500 of these tree pseudoreplicates was calculated. The percentage in which a certain clade is present in these pseudoreplicates is the bootstrap value. Data for the PAE (and NJAE) analysis of sites were taken from the following sources: Balta, Peru (Duellman & Thomas 1996); Cocha Cashu, Peru (Rodriguez & Cadle 1990, Rodriguez 1992); Cuzco Amazonico, Peru (Duellman & Salas 1991); Pakitza, Peru (Morales & McDiarmid 1996); Panguana, Peru (Schlüter 1984, Aichinger 1985); southeastern Peru (Cadle & Patton 1988 and misc. publ.); Puerto Almacén, Bolivia (De la Riva 1993d); Manaus, Brazil (Zimmerman & Rodriguez 1990); Los Colorados, Argentina (Lavilla et al. 1995). Used data sets may be obtained from the author. Neighbor joining analysis of endemism Neighbor Joining (NJ) analysis is a second method used in phylogenetic studies to produce dendrograms of relationships. NJ analysis first calculates a distance matrix between data sets and then searches for the tree which connects all sets with the minimum amount of branch length. Starting from the initial distance matrix, the program produces a further matrix which contains the distance between nodes. The two nodes with the closest distance are connected in the tree and are replaced by a new node which corresponds to their putative last common ancestor. This cluster is thus considered as one unit only in the further analysis. The program now searches again for the nodes with closest distance, etc., until all data sets in the tree are connected. Here, the method was adopted to analyze patterns of endemism. Analogue to the PAE, this method is here called Neighbor Joining Analysis of Endemism (NJAE). The NJAE was conducted using exactly the same data sets and options as in the PAE. NJAE was based on total character difference. The NJAE method produced exactly the same dendrograms as PAE, but partly resulted in different bootstrap values. In the results chapter, usually only one dendrogram is presented giving both bootstrap values, that for PAE and NJAE. Limitation of data Like already stated, data resulting from field surveys in general, and especially from short term ones, are usually not complete. It has been demonstrated in long term stud- ies that abundance of amphibian species in tropical forests show rather chaotic patterns (Duellman 1995). According to Pearman et al. (1995), accumulation of species num- 49 bers is reached best by the usage of a combination of different sampling methods. Even then, the species diversity at a site can not be discovered completely. As is obvious from the sampling methods described above, there was not the opportunity to use (or test) different and time consuming methods for sampling. All obtained data and the conclu- sions drawn from it have therefore to be regarded as preliminary. Including literature data, as done in the analysis (PAE, NJAE), provides a slightly more realistic picture of actual patterns. However, comparably few publications deal with Bolivian amphibians and the available data to fill up existing gaps are less than sufficient. Therefore it is almost sure that future findings will restrict the results and conclusions presented here- in. In practice, and like any other method for discovering patterns among regional faunas, PAE is influenced by sampling errors. The degree to which incomplete data affect the results of PAE are not yet quantitatively assessed. Nevertheless, PAE provides testable biogeographic hypotheses of faunal relationships. RESULTS Preliminary checklist and distributions In the following, an updated checklist of the amphibians of Bolivia is provided. Many of the included data resulted from own studies presented herein. Although most of the species are also listed in De la Riva et al. (2000), some differences exist which are due to my personal point of view and/or data that were not included in the mentioned pub- lication. In addition to the account of species, data on the distribution within the politi- cal borders of Bolivia’s Departamentos as well as within the ecoregions suitable for the species are provided. Moreover, it is stated, if the species is considered to be endemic for Bolivia or not. An asterisk (*) following the year of description indicates that the type locality of the species is in Bolivia. The checklist is followed by some annotations, because certain records deserve comments. Abbreviations used in the table are as follows. (1) Abbreviations of Departamentos: LP — La Paz; CB — Cochabamba; SC — Santa Cruz; BE — Beni; PA — Pando; PO — Potosi; OR — Oruro; CH — Chuquisaca; TA — Tarija. (2) Abbreviations of suitable ecoregions: AM — Amazonian rainforests; Cej — “Ceja” (cloud forest); Chi — Chiquitania forests (includes the Cerrado formations); CL — Chaco lowland forests; CM — Chaco montane forests; HiA — high-Andean vegetation; HTf — humid transition forests; LAV — inter- Andean dry-valleys; PCS — forests of the pre-Cambrian shield; TB — Tucumanian- Bolivian forests; WSa — wet savannas; YU — Yungas — montane rainforests. (3) Others: e — occurrence expectable; X — occurrence documented by voucher specimens and/or published data. ee. pales Sad EDS ESM NA SId ‘AVI JLH UDO ‘WV AVI VIH gl ‘AVI pew ote NA WY esMm ‘GL ‘IO ‘19 UI SOd LH ‘WV SOd JLH WV ‘ ‘ ESM SId HLH IND II WV gl ‘AVI WO 19 ESESFSEIEIESESTOEN LR es) eae od as = ss] Ca (z86\ ‘UeWUIIONG 9 ejlsjeuuedg) SI/eIAN/d eyauesyIOD „9661 AanıeH ejou eyjauesyoo9 „(0861 ‘eljeyeuued) ouesefagq eyjauesyoo09 sepiuajosjuay (/r6L ‘pPsel|aA) SuJUaAUgn SnosiusAuydouREjay (Lr6L ‘UllaIN]) smynumw SNosiu/UYydoJpUuEeGg 2S8L YPIWYIS sısuanbeuan oyng x91dwoo snıuoydA} oyng 0002 121404 3 S12997 IIejueJs oyng reg, ‘uuewBbalj, snsojnuids oyng ZEB] ‘OZWeD ıogwni oyng „196, ‘Opseyj|ey enyaanb ojng gGrgl ‘Ipnyos| ıBıddaod oyng GZ6L ‘ZIN7 siwauseued oyng (85/7 | 'snaeuul) snuuew oyng „7661 ‘UWS 9 AanıeH joueiuysn! oyng 66/1 Jepleuyos snjeynb oyng rzs\ Xıds snsoynueiß oyng €06| sebuajnog sedissy oyng L66L IIBMPIEO SNdijJOaUe}sed oyng 2981 „E66, ‘UWS 9 AanıeH Ssısuaosoqwe ojng ‘ asus} Wnieuase oyng Z06| ‘Jabuajnog sooo sndojs}y sepiuo/ng Dil NA ‘AL ‘AVI 'VIH [89 SOd HLH ‘14D WV SOd JLH WV SOd LH WV Gr6L ‘UOSsepuYy eaNyig e/AH „8681 ‘Jebuajnog uezjeg ejAH „2.06, JeBbuajnog gJeune ejAH vz6l ‘JOIN eulpue eJAH yZ8l xıds ejejoundogye e/AH „EZ6L ‘USPaIN esBiuog/e e/AH 796| ‘uUUeWIAaY0g euealoe ej/AH V ds esayjoyses (LZ8, 'epeds3 ej ap Zauawir) eauipnjsea] EI9YJONSEH (981 ‘pluyos) suapuajds eDayjoNse (LpgL ‘uOuqig $ JupWung) ejyeidnsuew edaYyjOUJSeH „L66L ‘CAIN P| 8G BeOUNZNe/ eCOaYjJOIJSeEH aepl|AH (rzgı 'xıds) smyeyum sajeqopadidg A8E8L Ipnyos] u UOIgig) smaıd seJegopadıd (E88, ‘abuajnog) wauyey saJegopadıdg „(Z06, sebusjnog) snuemimog sajeqopedidy \ ‘ds snyJ9JsoJo9 788| ‘JeBusinog snJeaum SnyJ3JS0J09 „2661 21504 °F SpIoUAay /pIUUeIpoW SNYJ97S0J09 (1881 ‘edoD) snauunig snyJ9}S0J09 8g ‘JeBuainog) sijesowa sEJegoJ]W sepieqoipueq (O86) ‘eyjayeuued) Vabiaqg wnıysenegouneÄH PSM a Wo) ate 6881 ‘’ebuajnog eueu ejAY CLEL ‘S1B}8q ejnumw e/AH 1881 ‘dod eaiABseuejaw ejAY GEBL ONSUUS SISU9NDEÄEIES e/AH ZOBL UENINT 9 }PIeEYUIaY Einpunaignı e/AY 0/61 ‘UIOd 9 UeIYDOD IOJBAL EIÄH 658, ‘JeujJUNd ejdedopoys ejAY (7981 ‘edod) sdaaıueı ejAY (66/1 ‘Japleuyos) eJejound ejAH (897 | ‘queineq) eyesouuew e/AH ‘ ESM SId 19 ESM SId ‘LH “19 UI WV SOd JLH WV NA ‘Al ‘AVI site) 2661, ‘OZIWED sejruenuew e/AY PSM SId 4LH WV (EBL ‘slouleg) eJejaydoangı e/AH 796) ‘UUeWUAaYOg //e9a/ EIÄH a a AVI [89 See SOd JLH WV SOd ‘$LH NV jo] (s)uo1ß341093 o}uswe 6861 Gens] 9 UeWI]aNG /uly990Y BJÄH 788, Jebueinog esoueub ejAH pzg\ xıds ealydesboab e/AH 6581 Ja}so-+ 9 spjouAay eajysouojyo EIÄH „0/61 ‘PIEIaA JueZzeseYyo e/AH „2.061 ‘Jabuajnog eunajdijjeo ‘jo eIÄH Jayjuns ejyeiose; ejAH ‘ «C661 gr8Ll jeyosoi, eyesegjeo ejAH (85/ L ‘sneeuuly) sueog ejAY ne ea eel = En Kl Bu BR ®| 8] ESM ‘SOd ‘LH ‘WY fae al m eS a el ae ae a] BE ae BESENEBBBBBEBRBEIFREISE =) alse] | | [six ePsI<[el<] [ee lees Biel Bs) bel | best fc] [|> Reben ie ee Biel |<] =Pebe] Te beix<[e [| Tel psles aa leis] = jel