SAN) HARVARD UNIVERSITY Library of the Museum of Comparative Zoology ', J MUS. COMP. ZOOU ^/^,f\\ LIBRARY cirp o 107/1 HARVARD A MULTIVARIATE university EVOLUTIONARY ANALYSIS OF THE ANDEAN IGUANID LIZARDS OF THE GENUS STENOCERCUS THOMAS H. FRITTS SAN DIEGO SOCIETY OF NATURAL HISTORY MEMOIR 7 1974 A MULTIVARIATE EVOLUTIONARY ANALYSIS OF THE ANDEAN IGUANID LIZARDS OF THE GENUS STENOCERCUS THOMAS H. FRITTS SAN DIEGO NATURAL HISTORY MUSEUM SAN DIEGO SOCIETY OF NATURAL HISTORY MEMOIR 7 1974 SAN DIEGO SOCIETY OF NATURAL HISTORY MEMOIRS MEMOIR 7, pages 1 - 89 /ssued June 21, 1974 Frontispiece. Open areas and exposed rocks on the steep valley walls of the Ri'o Urubamba near Chilca, 1760m, Departamento Cuzco, Peru are inhabited by Stenocercus ochoai. At lower elevations of the Ri'o Urubamba Valley, S. ochoai occurs in sympatry with S. crassicaudatus. Working on the bromeliad-covered slopes is William E. Duellman, of the University of Kansas who assisted with field work in southern Peru. PUBLISHED WITH FINANCIAL AID FROM THE W.W. WHITNEY PUBLICATIONS ENDOWMENT CONTENTS INTRODUCTION 9 Acknowledgments 10 MATERIALS AND METHODS 11 Materials 11 Definition of Characters 11 Definition of Ecological Variables 14 Univariate Methods 14 Quantitative Phyletic Methods 15 Multivariate Methods 16 Taxonomic Treatment 17 RESULTS OF ANALYSES 18 DISCUSSION 22 Morphology and Cladistics 22 Morphology and Ecology 28 Niche Theory and Evolution 29 EVOLUTIONARY CONCLUSIONS 33 TAXONOMIC ACCOUNT 34 Ophryoessoides 35 Stenocercus 35 Key to species of Stenocercus 36 Stenocercus apurimacus 39 Stenocercus boettgeri 39 Stenocercus carrioni 41 Stenocercus chrysopygus 42 Stenocercus crassicaudatus 45 Stenocercus cupreus 47 Stenocercus empetrus 47 Stenocercus festae 49 Stenocercus formosus 51 Stenocercus guentheri 54 Stenocercus haenschi 55 Stenocercus humeralis 56 Stenocercus ivitus 56 Stenocercus marmoratus 57 Stenocercus melanopygus 57 Stenocercus moestus 58 Stenocercus nigromaculatus 59 Stenocercus nubicola 59 Stenocercus ochoai 60 Stenocercus orientalis 60 Stenocercus ornatissimus 61 Stenocercus ornatus 62 Stenocercus praeornatus 62 Stenocercus rhodomelas 63 Stenocercus roseiventris 63 Stenocercus simonsii 64 Stenocercus trachycephalus 65 Stenocercus variabilis 65 Stenocercus varius 66 TAXONOMIC SUMMARY 68 LITERATURE CITED 71 APPENDIX A 84 APPENDIX B 88 INTRODUCTION Many approaches to the study of organic evolution have been explored since the dawning of Darwinian theory in 1859. A majority of the evolutionary studies prior to 1940 were based on empirical surveys of biological diversity and similarity. As a result of increased demands for scientific accuracy and greater understanding of some aspects of evolution, later studies empha- sized experimental techniques and were more limited in scope than the more ambitious but less definitive earlier studies. However, the use of experimental methods often limits the number of parameters that can be studied and results in informative but disjunct analyses of evolutionary mecha- nisms. Results from diverse sources subse- quently must be integrated and placed in a conceptual framework. Although the inte- gration of information from diverse sources often places the evolutionary biologist on the margins of his personal competence, it is one of the most important functions he will perform in pursuit of increased understand- ing of organic evolution. In assuming the task of analysis and integration he must be "willing to err moderately often in order that inadequate evidence shall more often suggest the right answer" (Tukey, 1962). Atchley (1971) used multivariate statis- tics to analyze simultaneously morphological and ecological data. Such techniques facili- tate understanding of causal factors of morphological variation and are potentially useful in broader evolutionary studies (Atchley, 1970; 1971). Multivariate tech- niques can be applied in testing hypotheses as well as in heuristic studies. This multivariate study of the lizard genus Stenocercus is the result of my interest in the evolutionary interaction between eco- logical and morphological components in vertebrate populations. It is an attempt to integrate morphological, ecological, and behavioral data in a quantitative analysis of evolutionary trends in the Andean lizards of the genus Stenocercus. Stenocercus Dumeril and Bibron (1859) is a member of the tropidurine group of the family Iguanidae, as defined by Etheridge (1964). The genus includes 29 nominate species distributed throughout the central Andes of South America at elevations of 200-3800 m. Some species of Stenocercus are comparable in their ecology to North American fence lizards of the genus Sceloporus, whereas others are more nearly equivalent to species in the North American genera Sator and Petrosaurus. The purpose of this study is threefold: 1) to infer evolutionary events in the radiation of the genus Stenocercus) 2) to estimate ecological influence on past evolu- tionary trends and present distributional patterns; and 3) to integrate ecological and morphological data in defining adaptive abilities and in considering the evolutionary history in terms of niche theory (MacArthur and Levins, 1964). A heuristic approach is employed in the analysis in order to facilitate generation of testable hypotheses of evolutionary mechanisms. Conceptual aspects of niche theory and adaptive radiation are examined in the context of evolution within Stenocercus. A taxonomic revision of the genus is presented to facilitate discussion of the species. A montane group was studied not only for physiographic and distributional diver- sity but also for faunal simplicity. At most Andean localities the fauna is small, and most elements are conspicuous and more easily studied than in more diverse lowland tropical regions. In addition, the close geographic proximity of ecologically diver- gent areas allows comparison of ecological factors with minimal geographic distortion. Lizards were studied because they have many morphological attributes that can be quantitatively measured, and because they can be observed and collected in reasonable series. 10 THOMAS H. FRITTS ACKNOWLEDGEMENTS This study was made possible through the provision of travel funds and three one-year traineeships in Systematics and Evolutionary Biology by the National Science Foundation to the University of Kansas (GB-4446X, George W. Byers principal investigator and GB-8785, J. Knox Jones, Jr., principal investigator). Addi- tional support was received from the F. William Saul Fund of the Museum of Natural History, University of Kansas. The University of Kansas Computational Center provided funds for computer analyses. William E. Duellman gave generously of his time in both field and laboratory phases of this study. Richard Etheridge provided notes on type material in Europe and shared his knowledge of iguanid lizards. Numerous persons in South America assisted field work by providing information and facilities. Among these are Oscar Ochoa M., Nellie Carrillo de Espinosa, Ismael Ceballos B., Filemon Lujan M., Luis Inope, Manuel Moro, Hugo Paz, James Muir, Fernando Ortiz, Hernando de Macedo, Elizabeth Gambini, Delfin Goitia, Victor Melendez Sandoval, Efrain de la Vega, Efrain Gonzales, Miguel Galegos L., Sisinio Morales Z., Nicanor Condorena and Evila Arias Davila. Mario Canedo, Vicente Lauro Gomez and Francisco Castilla Z. provided meteorological data for Bolivia, Ecuador and Peru, respectively. Several scientists al- lowed free access to material in their respective institutions as follows: William E. Duellman, Museum of Natural History, University of Kansas (KU), Ernest E. Williams, Museum of Comparative Zoology (MCZ); Charles F. Walker, University of Michigan, Museum of Zoology (UMMZ); Richard G. Zweifel and Charles W. Myers, American Museum of Natural History (AMNH); James A. Peters, United States National Museum (USNM) and James A. Peters private collection (JAP); Nellie Carillo de Espinosa, Museo Javier Prado, Lima (MJP); Hymen Marx, Field Museum of Natural History (FMNH); Gustav Orcez v., private collection (GOV); Wade C. Sherbrooke private collection (WCS); James R. Dixon, Texas A & M Museum (TAM); Robert Mertens, Senckenbergische Natur- Museum (SM); David B. Wake, Museum of Vertebrate Zoology (MVZ); and Gaston F. de Witte, Institut Royal des Sciences Naturelles de Belgique (BIN). Type material of Stenocercus also exists or existed in the following institutions: British Museum (Natural History) (BMNH); Koenglisches Zoologisches und Anthropologisch-Ethano- g^aphisches Museum, Dresden (D); Musee d'Histoire Naturelle de Neuchatel (MHNN); Museum National d'Histoire Naturelle, Paris (MHNP); Museo Zoologico di Turin (MZT); Naturhistorischen Museum in Wies- baden (WNM); Zoologische Museum der Humboldt Universitat, Berlin (ZMB); Zoo- logisches Staatsinstitut und Zoologisches Museum, Hamburg (ZMH). Janalee P. Caldwell, Robert R. Beatson and Richard R. Montanucci commented on various aspects of this report; Craig Black, William E. Duellman and Peter R. Ashlock read the manuscript. I am grateful to all of these people for their aid. ANDEAN IGUANID LIZARDS 11 MATERIALS AND METHODS MATERIALS More than 1000 specimens of the genus Stenocercus were examined during this study. Of these, over 800 were collected and examined in life during 14 months of field work in the Andes of Bolivia, Ecuador, and Peru. Preserved material of every species recognized herein was examined; notes on coloration of living animals, colored photo- graphs, and karyotypic materials are avail- able for most species. Karyotypic studies will be the basis for a future report. A list of those specimens examined is given in Appendix A, whereas those specimens used in statistical analyses are indicated in Appendix B. In most analyses emphasis is placed on the local sample in order to allow for consideration of climatic and ecological parameters measured at a specific locality. Except in certain cases a local sample is composed of four or more adult specimens of the same species collected within a five kilometer radius of a point. Stenocercus roseiventris , which is known largely from material collected before 1910 with general or imprecise locality data, is herein repre- sented by two grouped samples from the Bolivian departments of Cochabamba and Santa Cruz. Stenocercus moestus is repre- sented by a series of specimens with imprecise data. Other species for which local samples of four or more adult specimens are lacking are omitted from quantitative analyses. Grouping of specimens from widely separated localities or different biogeo- graphic or ecological environments was not done for several reasons. If the species does exhibit geographic variation in morphology, the statistical computations resultant from pooled samples will merely be reflective of the proportions of individuals from the various localities and habitats; the source of the variation — selection, drift, environ- mental influence — will not be detectable. Although examples of pooling of samples are still common in zoological literature, the change in emphasis from consideration of geographic isolation and gene flow as primary evolutionary factors, to definition of selective pressures on local and micro- ecological units will cause the decline of those practices that mask the causes of variation. DEFINITION OF CHARACTERS The characters used in this study were chosen so as to measure as many attributes of the lizards as possible. Specific names for individual scales and characters based on specialized nomenclature have not been used for two reasons: 1) a generalized char- acter, such as the number of scales between temporals (see definition below), can be compared among many groups having considerable variation in cephalic scutellation, whereas a character based on a particular type of scale, such as number of postoccipitals, depends upon the definition of homologous epidermal units and necessi- tates comparable nomenclature and cephalic patterns in all specimens studied; 2) the study of variation is facilitated by use of characters that involve counting or measur- ing morphological units quantitatively in- stead of grouping them into subjective classes. Although meristic characters (scale counts) were preferred for quantitative study, some qualitative characters have been included. Since qualitative measure- ments involve grouping data into classes, considerable information can be lost if the grouping is not done carefully. Several of the characters treated qualitatively in the present study could be measured in a more quantitative manner with increased invest- ment of time. The character based on the size of the lateral scales included two grouped classes: lateral scales which were greater than one-half the size of dorsals, and 12 THOMAS H. FRITTS lateral scales which were one-half the size of dorsals or smaller; the attribute could have been expressed as a measurement of the lateral scales or a ratio of laterals to dorsals. Characters measured qualitatively herein were of less value than they might have been if treated in a more quantitative manner. Twenty-nine characters were meas- ured on each specimen used in statistical analyses, except in cases where specimens were damaged, thereby preventing observa- tion of one or more attributes. Aspects of coloration were compared for all populations for which notes on coloration in life were available. Of particu- lar interest are sexually dimorphic charac- teristics. The ventral coloration of males is often markedly different from adult females, as well as from males of other species occur- ring in sympatry or in close geographic proximity. The ventral surfaces of the thighs and pelvic region are yellow or black in adult males of some species. The entire venter or that part lateral to a midventral stripe is brightly colored in some males. In the females of some species, the chin, side of head, and side of neck is black and in contrast to the ground color. Females of many species are similar in coloration, although differences do exist between them. The present analysis has not included any quantitative analysis of coloration due to the inherent problems of ontogenetic and sexual differences. A brief description of the method used to measure each character employed in quantitative analyses is provided below, in order to facilitate understanding of the results and the design of comparative studies of related or similar groups of lizards. Scales Between Nasal and Subocular (NSO).— The number of scales in contact with the supralabials between the subocular and the nasal is noted for both the right and left sides; the number used is the sum of the two sides. The elongate subocular and the nasal are not included. The subocular scale of Stenocercus is an elongate scale between the ventral margin of the eye and the supralabials. Supraoculars (SO).— The number of supraocular scales in a transverse line at the greatest width of the supraocular semicircle is given. The supraoculars are approxi- mately in longitudinal rows; therefore the character state is the maximum number of rows in one semicircle. Internasal Scales (IN).— The number of scales between the nasals and in contact with the rostral varies from two to four. Scales Between Temporals (SBT). — The number of dorsal head scales is counted along a transverse line between the temporal scales of the right and left sides and posterior to the interparietal. The counts originate and end with the row of angulate temporal scales posterior to the ultimate superciliary scale. Parietal Eye (PE).— The presence or absence of a small unpigmented spot associated with the parietal eye is noted. Posterior Head Scales (PHS).— The scales on the posterior part of the head are compared with the adjacent scales on the dorsal part of the neck. Scales Along Midline of Head (SMH). — The scales along the midline of the head from the occipital region to the rostral are counted; the rostral is included. Head Scales (HS).— The scales of the interorbital and frontal regions are noted to be smooth or keeled. Scales of the temporal region of nearly all species are keeled and are not recorded. Size of Scales Posterior to Dorsal One -half of Tympanum (SDT).— The size of the scales posterior to the external auditory meatus is compared to the size of the adjacent scales of the dorsal neck. The scales posterior to the tympanum are considered to be small if they are one-half (or less) of the size of the dorsal neck scales. Type of Scales Posterior to Dorsal One-half of Tympanum (TDT).— The scales in this region are recorded as granular, imbricate smooth, or imbricate keeled. Antehumeral Fold (AF).— A fold of epidermis and dermis immediately anterior to the insertion of the forelimb is considered to be an antehumeral fold. ANDEAN IGUANID LIZARDS 13 Neck Fold (NF).— The folds anterior to the antehumeral fold on the side of the neck are termed neck folds. In some species the neck fold and antehumeral fold are con- nected by a longitudinal fold. Auricular Lobules (AL).— Some of the scales on the anterior margin of the tympanic opening are denticulate and project into the ear opening; they are termed auricular lobules. Repeated counts of the number of auricular lobules are variable and the character is considered to be of little utility. Posthumeral Pocket (PHP). — A dis- tinct depression or pocket of skin in the axillary region is termed a posthumeral pocket. Postfemoral Pocket (PFP).— A distinct pocket or involution of skin behind the insertion of the hindlimb is termed a postfemoral pocket. Those species having a slight fold which is present only when the hindlimb projects posteriorly are considered to lack a postfemoral pocket. Condition of Dorsal Scales (DS).— The dorsal scales are granular, imbricate smooth, or imbricate keeled. In species where two or more types of scales occur, the condition of the scales near the vertebral row is recorded. Condition of Lateral Scales (LS).— The lateral scales are granular, imbricate smooth, or imbricate keeled. Size of Lateral Scales (SLS). — The size of the scales on the sides of the body is compared with that of adjacent dorsal scales. Laterals one-half the size of the dorsals are noted as small. Vertebral Scale Row (VR).— The row of scales on the middorsal line from the head to the base of the tail is termed the vertebral row. A partial or diverging row is recorded as a discontinuous vertebral row, whereas a row comprised of scales in a single unbroken Hne is considered as a continuous row. Dorsal Crest (DCT).— A dorsal crest is recorded as present if the scales of the vertebral row are morphologically distinct from adjacent dorsals. The crest of some species is not raised above the adjacent dorsals, whereas the crest of other species consists of high denticulate scales. Scales on Ventral Surface of Arm (VA). — The scales of the ventral surface of the proximal forelimb are granular, imbricate smooth, or imbricate keeled. Scales on Posterior Thigh (PT).— The scales on the posterior surrface of the thigh are granular, imbricate smooth, or imbri- cate keeled. Subdigital Lamellae on Fourth Finger (LF). — The subdigital lamellae are counted from the point of attachment of the third and fourth fingers to the terminus of the fourth finger. Subdigital Lamellae on Fourth Toe (LT). — The subdigital lamellae are counted from the point of attachment of the third and fourth toes to the terminus of the fourth toe. Caudal Scales (CD).— The size of the caudal scales and degree of keeHng varies. Mucronate and conspicuously keeled caudals tend to be one and one-half times the size of ventral scales and are termed spinose. Smaller scales in relation to ventral scales have less prominent keels and are termed normal. Caudal Whorls per Autonomic Segment (AS).— The position of autonomic planes of the caudal vertebrae is indicated by caudal whorls composed of larger scales than in other whorls. The number of caudal whorls or transverse rows of caudals between consecutive autonomic points is noted. Scales Around Middle of Body (SMB).— The number of longitudinal rows of scales is determined by counting in a transverse line around the middle of the body halfway between the insertion of the limbs. Dorsal Scales in Transverse Rows (DTR).— The dorsals are counted in a longitudinal paravertebral row from the occipital scales on the posterior margin of the head to the base of the tail at the posterior margin of the thigh. Scales Across Gnlar Region (SAG).— The gular scales are counted in a transverse arch across the gular region from the ventral margin of one tympanum to the other. The scales of the chin are arranged in 14 THOMAS H. FRITTS longitudinal rows radiating from the labial region allowing precise counts of the number of rows transversed from one tympanum to another. The granular scales in the interior part of the tympanic depression are not included in the count. DEFINITION OF ECOLOGICAL VARIABLES An exhaustic analysis of the ecology of the lizards of the genus Stenocercus is not possible at present. All ecological data are examined on the level of the local sample in order to allow incorporation of variables pertaining to local demes instead of individual specimens. Meteorological data are taken from published records for Bolivia (Mendez, 1965), Ecuador (Garcia, 1959- 1968), and Peru (Anonymous 1960-1963; 1964; 1965a; 1965b). Observations of local populations for periods of several hours to several days during field work are the basis for data on habitat preferences, associated fauna and flora, and behavior. Eight variables are employed in the quantitative analyses. A brief description of these variables is presented below in order to facilitate interpretation of the analyses. All meteorological variables are recorded as means of observations from two to ten years if data were available. Observations for one year are used for localities for which additional data are lacking. Maximum Precipitation in Any Month (MXP).— The amount of precipitation for the one month of each year having the maximum rainfall was taken from meteoro- logical records. Mean Monthly Precipitation (MOP). — The mean monthly precipitation for a locality for each year was taken from meteorological records. Minimum Precipitation in Any Month (MNP).— The amount of precipitation for the one month of each year having the minimum rainfall was taken from meteoro- logical records. Mean Annual Temperature (MAT). — In most cases the mean annual temperature was computed by averaging the mean monthly temperatures. For the few locali- ties for which no temperature data are available, the mean annual temperature was computed using the relationship of 6°C per 1000 m (Eidt, 1969). Instead of using 30°C for the mean annual temperature at sea level for the basis of computations as proposed by Eidt, the actual mean annual temperature of a nearby locality in the same drainage system was used and corrected according to the difference in elevation between the two sites. Latitude (LAT).— The latitude of each locality from which local samples were obtained was converted from degrees and minutes to degrees accurate to two decimal places. Altitude (ALT).— The elevation at which the sample was collected is recorded in meters. Drainage (DNG).— The location of each locality is noted as within the Atlantic or Pacific drainage system. Habitat (HAB).— Lizards of a local sample are considered to be 1) terrestrial if individuals are observed on the ground as well as on exposed roots of shrubs and rock piles, 2) rock-dwelling if individuals of both sexes are confined to surfaces of rock piles and rock cliffs, or 3) arboreal if individuals of both sexes are frequently seen on vertical trunks or elevated branches of shrubs and trees. UNIVARIATE METHODS Basic statistics of all variables for each local sample were computed using a program written by Dennis M. Power. Intraspecific variation was examined by the Sum of Square-Simultaneous Test Proce- dure (SS-STP). The SS-STP is an a posteriori test of significance of means (Sokal and Rohlf, 1969) and defines nonsig- nificant subsets of samples after significant variance is demonstrated by an overall analysis of variance. The test is used for samples of only two means realizing that the probability of 0.05 is that of an a priori test, not an a posteriori test. Significance in analyses of only two samples is therefore tentative. ANDEAN IGUANID LIZARDS 15 QUANTITATIVE PHYLETIC METHODS The quantitative phyletic algorithm (Kluge and Farris, 1969) was employed for evolutionary analysis of morphological rela- tionships. The computer program adapted for use on the Honeywell-635 computer at the University of Kansas attempts to compute a most parsimonious rooted den- drogram from morphological data. The den- drogram is rooted by a set of characters considered to have been present in a common ancestor of the group. Selection of the characters to be used for the common ancestor is made on the basis of criteria outlined by Wagner (1961) and Kluge (1969). Although considerable discussion has centered around the subjectivity and circu- larity of these criteria for selecting charac- ters present in ancestral forms, few alternatives exist (see Inger (1967), and Kluge and Farris (1969) for contrasting viewpoints of this problem.) The criteria for selection of primitive (ancestral) character states (Wagner, 1961) likely to have been present in a common ancestor of all species in the genus Stenocercus are: 1) a primitive state is present in numerous representatives of closely related groups; 2) a primitive state is more widespread within the group than any one derived state; 3) a primitive state is one associated with states of other charac- ters known to be primitive by other evidence. Although tentative ancestral char- acter states can be assigned for most characters on the basis of criteria one and two, ancestral character states for those characters having a large range of states {e.g. scales around the middle of the body and scales across the gular region) must be inferred on the basis of criterion three. The concordance of criteria one and two in the selection of ancestral states for many characters of Stenocercus is possibly a result of the large size of the group being studied, as well as the similarity of the logic used in the formulation of the two criteria. On the basis of both criteria one and two, the inferred ancestor possesses a low number of scales between the nasal and subocular (6.5); a moderate number of supraoculars (5.0); a low number of internasals (2.0); a moderate number of scales across the occipital region between the temporals (10.0); a parietal eye in nearly all individuals (1.8); posterior head scales distinct from adjacent scales of dorsal neck; a moderately low number of scales between occiput and rostral (15.0); smooth head scales, small, imbricate, keeled scales on side of neck; antehumeral and neck folds; a few auricular lobules projecting into ear opening (2.0); no posthumeral pocket; imbricate, keeled scales on dorsum; imbri- cate keeled scales on lateral body which are nearly equal in size to dorsals; a continuous vertebral row; granular scales on posterior thigh; a low number of subdigital lamellae on fourth finger (18.0); a low number of subdigital lamellae on fourth toe (24.0); normal caudal scutellation; and three caudal whorls per autonomic segment. On the basis of criterion one in concordannce with criterion three the inferred ancestor has no postfemoral pocket; no dorsal crest; imbricate, keeled scales on ventral surface of proximal arm; a moder- ately low number of scales around the body (50.0); a moderately low number of dorsals between the head and the base of the tail (53.0); and a low number of scales across the gular region (24.0). All of the above character states are tentative in being based on indirect evidence. Some of the character states are likely to be erroneous; however, if most of the inferred ancestral character states approximate those of an ancestral form, reevaluation of others with additional evidence will be possible. The use of diverse kinds of morphologi- cal data in the quantitative phyletic program prompts consideration of weight- ing and scaling procedures. Kluge (1969) and Kluge and Farris (1969) discussed methods of weighting data for evolutionary analyses according to relative conservatism of a particular character in relation to others. The weighting function I use is one proposed by Kluge (pers. comm.) involving the formation of a ratio of interclass 16 THOMAS H. FRITTS variation/intraclass variation, or specifi- cally, the standard deviation of the popula- tional means/the mean standard deviation of the individual populations. The weighting function, a real number, is multiplied by the value of the character state for each OTU (operational taxonomic unit). Each charac- ter has an independently computed weight- ing function. The purpose of the weighting function is to increase the relative number of estimated evolutionary steps between the states of a character in proportion to the conservatism of that character. The weight- ing function allows the conservative charac- ters to have more influence in the evolutionary construct. Kluge and Farris (1969) included diverse kinds of morphological characters in a quantitative phyletics analysis of the anurans, but all characters were coded and the ranges of character states were small and approximately equal. Most characters used by Kluge (1969) in a similar study of Hemidactylus were based on scutellational counts; only one character, "type of digit," included coded character states. The ranges of the populational means of the meristic characters varied; however, none of the characters included more than 15 evolution- ary units before being weighted. Thus, the ranges of the characters within both of these data matrices analysed using the quantitative phyletic algorithm had com- parable ranges. The morphological data matrix used in this study includes several diverse kinds of characters. Both meristic characters, counts of the number of scales on a particular area of the body, and coded qualitative charactes describing differences in scale pattern or morphology are utilized. The minimal number of states for a character is two (e.g. ventral scales smooth (1), ventral scales keeled (2)). The maximal number of states is 83 for the number of transverse rows of dorsal scales between head and base of tail (DTR). The number of character states observed in a character directly influences the respective weighting function com- puted for it. Thus, it is necessary to avoid duplication of this effect by rescaling all observed character states to an approxi- mate common range. This is accomphshed by dividing all meristic characters by the minimal character state for that character. As a result most characters have a range of two. The characters are then weighted by multiplying the scaled values by the weight- ing functions. Other scaling techniques are possible but remain to be considered for use in similar analyses. MULTIVARIATE METHODS Multivariate studies were performed using programs in the BMD (Biomedical Computer Programs) and SFA Program libraries at the University of Kansas Computation Center. Multivariate analyses using mean values for local samples were performed on 29 samples of 19 species, for which morphological, climatic, and ecologi- cal data are available. Matrices of product- moment correlation coefficients were com- puted for 28 morphological variables and eight ecological variables using the SFA OlC computer program. The two separate correlation matrices were factored independ- ently through the use of the SFA 03D program of incomplete-components analysis. The vectors thus defined were subsequently rotated orthogonally according to the Varimax Criterion utilizing SFA 13B. Finally, factor scores were computed for both the ecological and morphological principal components analyses utilizing SFA 22C. The principal components analysis summarizes trends in the correlation matrix through the definition of vectors delineating maximal variance of the original matrix. Orthogonal rotation of the multivariate vectors insures independence of the trends defined. The computation of factor scores allows viewing of the individual OTUs on the vectors. Thus, each vector defined in the analysis can be described by its constituent OTUs in relative order and position. The eigenvalue of each vector is a direct measuremennt of the amount of variance included in the vector; the eigenvector is a description of the vector and indicates the relative amount each ANDEAN IGUANID LIZARDS 17 character has influenced the delimitation of the vector. Factor scores of the independently- factored morphological and ecological data sets for the 29 localities were analysed using a canonical correlation program, BMD X75. Canonical variables were computed for each canonical correlation. Whereas principal components analysis attempts to explain maximal variance through transformation of original data into linear equations, canonical correlation at- tempts to find maximal covariance between two sets of data through delimitation of a series of linear equations. Canonical correla- tion analyses can compute as many correlations as there are variables in the smaller of the two data sets. All canonical correlations are statistically independent from others. TAXONOMIC TREATMENT The taxa assigned to the genera Stenocercus and Ophryoessoides were re- viewed. Several species are included in Stenocercus that were previously assigned to Ophryoessoides. Generic assignment was made on the basis of morphological and ecological similarity within and between genera. The delimitation of species was made on the basis of morphological concord- ance and ecological similarity between populations. The taxonomic treatment is necessarily a conservative one due to the paucity of material and rudimentary knowl- edge of the distribution and ecology of many species. Some populations are morphologi- cally distinguishable from others of the same species but which are similar in most other aspects of distribution, ecology, and morphology. Sympatry between closely related species is rare, although better icnowledge of the distributions of all species will likely provide examples of parapatric populations which do not interbreed and which warrant specific recognition. Geo- graphic variation is discussed when samples are sufficiently large to allow statistical analysis. The ranges used for meristic characters in the diagnoses are absolute ranges for all populations of the species. Individual populations are discussed when the popula- tion means are significant at p < 0.05 level in SS-STP analyses. All means are com- puted for local samples and not for species. A mean character state for a species composed of more than one sample is of no statistical and little descriptive value. Taxa are listed in alphabetical order. 18 THOMAS H. FRITTS RESULTS OF ANALYSES Three major cladistic events are prominent in the evolutionary construct, resulting from the Wagner-tree analysis (Fig. 1). The right evolutionary branch includes nineteen populations of ten spe- cies. The species on the right branch in approximate order from basal to distal positions are: Stenocercus omatissimus, S. variabilis, S. humeralis, S. praeomatus, S. varius, S. boettgeri, S. simonsii, S. cras- sicaudatus, S. carrioni and S. empetrus. The remaining OTUs are positioned on two major branches diverging near the ancestral condition. The central branch includes eight local samples of three species. In approximate order from basal to distal positions are OTUs of Stenocercus mela- nopygus, S. moestus, and S. chrysopygus. The left branch includes nine OTUs of six species. Stenocercus orientalis, S. formosus, S. ochoai, S. roseiventris, S. marmoratus, TOTO HUAR V CHAV RICH CHIL YAUP LIMA HUAM CAJA \ UNIO and S. cupreus are represented by OTUs on the left evolutionary branch with the latter most distal from the ancestor. The left branch includes 273 patristic units, whereas the right branch has 473. The shorter and less diverse central branch includes 200 patristic units. The 8x8 correlation matrix of ecologi- cal data for 29 local samples reflects several relationships among the variables. As elevation increases the minimum monthly precipitation in the year decreases. Mean annual temperature is inversely correlated with altitude; maximum and minimum monthly rainfall values are highly correlated with annual precipitation. Annual precipi- tation and maximum precipitation per month are correlated with mean annual temperature. Occupation of rock habitats by lizards is correlated with maximum monthly rainfall. In southern latitudes the lizards CELE SACS \ ALAM MACHv^ PORC COMA PAUC TAND I GIRO , SARG PALM LOJA MEJO AZUL AYACs]^'^ OTUZ 75 PATRISTIC UNITS Figure 1. Wagner dendrogram of evolutionary relationships of 36 OTUs. See Appendix B for localities represented by OTUs. ANDEAN IGUANID LIZARDS 19 occupy Amazonian slopes and interandean valleys, whereas in northern latitudes lizards occupy interandean valleys and Pacific slopes. The principal components analysis of the ecological correlation matrix (Table 1) delineates three orthogonal factors "explain- ing" 81.4 percent of the original variance. The first factor "explains" 34.5 percent of the original variance and includes precipita- tion and habitat utilization. The second ecological factor "explains" 21.3 percent of the original variance and emphasizes relationships between latitude and drain- age. The third ecological factor "explains" 25.5 percent of the variance and reflects the inverse relationship of altitude with mean annual temperature. The 28 X 28 correlation matrix of the morphological variables indicates that me- ristic characters involving number of scales of a particular area or type are correlated; not all scale counts are positively correlated. The presence of a posthumeral pocket, postfemoral pocket, keeling on dorsal scales, and keeling on head scales correlate Table 1. Varimax Rotated Factor Matrix for 8 Ecological Variables. Factor loadings less than 0.25 are recorded as zero. Columns are factors; rows are variables. Columns two and three are transposed to facilitate visualization of the variables loading on the factors. 1 3 2 MXP 0.923 0. 0. MOP 0.843 0.339 0.279 MNP 0.758 0.469 0. HAB 0.718 0. -0.360 ALT 0. -0.901 0. TEM 0. 0.893 0.263 LAT 0. 0. 0.832 DNG 0. 0. 0.825 Cumulative Percent Variation "Explained" 34.55 60.08 81.40 with few other variables. Six factors are defined from the 28 x 28 matrix of correlation coefficients of morpho- logical variation through principal compo- nent analyses (Table 2). The six factors "explain" 82.9 percent of the original variance after orthogonal rotation. The first morphological factor "explains" 27.9 percent of the variance and emphasizes several characters measuring scale number and size, including aspects of the body and limbs. The variables measuring scale size are inversely correlated with those measur- ing the number of scales. The second morphological factor emphasizes character- istics of the tail and head and "explains" 17.8 percent of the original variance. The third factor is one reflecting trends in the dorsal crest and snout morphology; it "explains" 13.9 percent of the original variance. The percentage of the population with a parietal eye is inversely correlated with other characters emphasized in this factor. The variation in thigh and head morphology is described by the fourth factor; 7.1 percent of the variance is accounted for by this factor. Characteris- tics of neck morphology are emphasized in the fifth factor; 11.2 percent of the total variance is "explained". The sixth factor accounts for only 5.1 percent of the total variance; aspects of the posthumeral pocket, posterior thigh, and dorsal scutellation are reflected in the factor. All other factors "explain" less than 5 percent of the original variance and are less interpretable. Three relationships between the inde- pendently derived ecological and morpho- logical factors are noted (Table 3). Scutella- tion (first morphological factor) is negatively correlated with latitude-drainage (second ecological factor), rij = .5476, p < 0.01 (Fig. 2). Thigh and head morphology (fourth morphological factor) are correlated with precipitation and habitat (first ecological factor), rij = .4263, p < 0.05. Caudal and posterior head morphology (second morpho- logical factor) are correlated with precipita- tion and habitat, rij = .3655, p < 0.05. Canonical correlations between morpho- 20 THOMAS H. FRITTS Table 2. Varimax Rotated Factor Matrix for 28 Morphological Variables. Factor loadings less than 0.25 are recorded as zero. Columns are factors; rows are variables. Columns four and five are transposed to facilitate visualization of the variables loading on the factors. 1 2 3 5 4 6 LF 0.904 0. 0. 0. 0. 0. LT 0.853 0. 0. 0. 0. 0. LS -0.847 -0.359 0. 0. 0. 0. VA -0.839 -0.271 0. 0. 0. 0. SAG 0.830 0.371 0.319 0. 0. 0. SMB 0.794 0. 0.285 0.276 0. 0. DTR 0.760 0. 0.330 0.313 0. 0. SBT 0.653 0.561 0. 0. 0. 0. AL 0.598 0. 0. 0. -0.533 0. PT -0.568 0. 0. 0. 0.504 0.302 SO 0.532 0.308 0. 0.476 0. 0. PHS 0.515 0.727 0. 0. 0. 0. AS 0. -0.919 0. 0. 0. 0. CD 0. 0.905 0. 0. 0. 0. SLS 0.403 0.684 0.345 0. 0. 0. SMH 0.483 0.655 0.286 0.335 0. 0. DCT 0.325 0. 0.765 0. 0. 0.351 IN 0.471 0. 0.754 0. 0. 0. NSO 0. 0.409 0.740 0. 0. 0. VR 0.324 0. 0.703 0. 0. 0.317 PE 0. -0.362 -0.688 -0.445 0. 0. PHP 0. -0.299 0.588 0. -0.461 -0.304 SDT 0. 0. 0. -0.909 0. 0. NF 0. 0. 0. 0.902 0. 0. TDT -0.436 -0.477 0. -0.557 0. 0. PFP 0.302 0. 0. 0. 0.765 0. HS 0. -0.420 0. 0. 0.645 0. DS 0. -0.252 0. 0. 0. 0.840 Cumu lative Percent Variation "Explained" 27.86 45.63 59.60 70.73 77.84 82.89 ANDEAN IGUANID LIZARDS 21 Table 3. Matrix of Correlation Coefficients of Three Eco- logical Factors with Six Morphological Factors. Critical value at p = 0.05 level for correlation coefficients equals 0.367. Ecological Factors 1 2 3 Morphologica 1 Factors 1 0.25 -0.55 -0.29 2 0.37 -0.05 0.10 3 0.26 0.19 0.18 4 0.44 0.16 0.21 5 0.10 0.06 -0.12 6 0.05 -0.06 0.30 logical and ecological factors (Table 4) resemble the simple correlation described above. The signs after the factor names indicate relationship of individual factors to the canonical correlation. Latitude-drainage (-) and altitude-temperature (-) show a positive relationship with thigh-head (-), scutellation (^-), and crest-snout (-) mor- phology, Re = .7434, p < 0.05. Figure 3 illustrates the relative position of the OTUs in this canonical correlation. The second canonical correlation indicates that precip- itation-habitat {+) and latitude-drainage (-) are correlated with scutellation (+) and tail- head (+) morphology, Re = .6793, p < 0.1 (Fig. 4). The third canonical correlation is uninterpretable and does not approach significance. Redundancy values for the first canoni- cal correlation indicate that 33.4 percent of the ecological variance is "explained" by the morphological data; similarly 16.7 percent of the morphological variance is "explained" by the ecological data. Redundancy values for the second canonical correlation are similar, 33.2 and 16.7 percent, respectively. 0Chap 0Tolo 0Pich ^Tira Ayac0 Mejo 0Yaup Chiq 00Pal<: Huan^ ^ Huam ^Unio #Caia ^Coma 0Chu 0Mach ^Pauc 9chac ^Cele 0Chav 0Huar ^Bamb 0Otu ^Lima #Loja )Sarg 0Al3 r . -.5476 P < .01 MORPHOLOGICAL COMPONENT NO. 1 (SCUTELLATION) Figure 2. Correlation of first morphological component with second ecological component, based on 29 OTUs. 22 THOMAS H. FRITTS DISCUSSION The primary purposes of this study are the estimation of ecological and morpholog- ical variation, the definition of evolutionary trends, and the integration of evolutionary information and niche theory. Studies by Atchley (1970; 1971) and by Smith and Koehn (1971) are examples of recent attempts at integration of ecological and morphological data sets in a quantitative manner. Atchley (1971) studied intraspecific variation through multiple regression of morphological characters on climatic data. Smith and Koehn (1971) used species as the basic unit of studies and compared catasto- mid fishes through cladistic annd multivari- ate analyses. The use of local samples as the basic unit in the present study is an attempt to incorporate intra- and interspecific Table 4. Matrix of Canonical Correlations and Canonical Coefficients for Three Ecological Factors and Six Morphological Factors. Canonical Correlation 1 2 3 0.743 0.679 0.303 Coeffi icients for Cai of Ecologies nonical Variables )l Factors Factor 1 -0.515 0.811 -0.278 Factor 2 -0.592 -0.571 -0.569 Factor 3 -0.620 -0.129 0.774 Coefficients for Canonical Variables of Morphological Factors Factor 1 Factor 2 Factor 3 Factor 4 Factor 5 Factor 6 0.501 -0.299 -0.484 -0.604 -0.179 -0.244 0.811 0.458 0.114 0.330 0.087 0.055 0.073 0.012 -0.142 -0.152 -0.509 0.832 variation in the same analysis. Since selective pressures resultant from ecological and climatic factors can differ within the range of a species just as they differ from those of other species, simultaneous exami- nation of intra- and interspecific variation can possibly contribute to our knowledge of factors causing speciation, as well as geographic variation below the species level. MORPHOLOGY AND CLADISTICS The Wagner-tree (Fig. 1) computed for 36 local populations included in the present analysis depicts three major evolutionary lines. These lines represent computed cladistic lines and shall be treated sepa- rately in order to facilitate discussion of the various parameters of interest. The right branch is composed of ten species of Stenocercus. All species assigned to the S. humeralis species group by Etheridge (1970) are positioned on the basal portion of the right branch. The most terminal OTUs of the right branch are members of the S. crassicaudatus species group of Etheridge. The most basal OTU on the right branch (OTUZ) represents S. omatissimus, a member of the S. chrysop- ygus group. The populations of the right branch are mainly from high interandean and Pacific slopes of northern Peru and Ecuador with only the more basal ones from areas below 3000 m on Amazonian slopes of central Peru. Stenocercus crassicaudatus is the only species on the right branch that reaches southern Peru. Stenocercus omatis- simus occurs at intermediate elevations on the coastal slopes in northern Peru. The central evolutionary branch in- cludes three species, all members of the S. chrysopygus group as defined by Etheridge (1970). These species are distributed in coastal and interandean valleys of northern and central Peru. Stenocercus melanopygus (one of the species on the central branch) ANDEAN IGUANID LIZARDS 23 and S. ivitus (a closely related species not included in the analysis) are known to occur in sympatry with congeneric species of the right and left branches. The left branch is composed of six species, all of which inhabit Amazonian drainages. The most basally positioned OTU (CHAC) represents S. orientalis, which occurs in northern Peru. The remaining OTUs on the left branch are from central and southern Peru and Bolivia. Stenocercus orientalis most closely agrees with the S. chrysopygus species group and is patristi- cally close to the central branch. Intermedi- ately positioned OTUs representing the species now known as S. formosus (Tschudi, 1845, non Boulenger, 1880) and S. ochoai have been placed in Ophryoessoides by previous workers. Populations of S. rosei- ventris and S. marmoratus are terminally positioned on the left branch; both of the latter species were placed in the S. crassi- caudatus species group by Etheridge. Steno- cercus cupreus, also represented by a terminally positioned OTU, was placed in the S. chrysopygus species group. The evolutionary construct agrees in several ways with the species groups proposed by Etheridge. Basic differences exist concerning the affinities of S. rosei- ventris, S. marmoratus, and S. cupreus. Whether the specialized caudal morphology shared by the latter three species is an example of parallel evolution with the species of the crassicaudatus group as indicated in the dendrogram or indicative of close relationships between them as indi- cated by Etheridge's classification will be discussed below. In biogeographic terms the phyletic 3 o o I- < 3 O Cele I Giro Huan Pale Yaup kUnio 'Chuc kComa •Chap Pich— <^ 0Chac Tolo • ^^ T ^^ AHuam ^\ •Chiq Caja9 Chav • huar RdX ' -332 RpY = -167 Mejo( Otuz I Ayac RC2 = .679 P • .1 (PRECIPITATION — HABITAT, LATITUDE-DRAINAGE) Figure 4. Second canonical correlation between morphological data (abcissa) and ecological data (ordinate). Redun- dancy values (Rjx and ^dy^ ^""^ computed according to the method of Cooley and Lohnes (1972). ANDEAN IGUANID LIZARDS 25 upon correlations and concordant trends of groups of characters in order to estimate affinities. Because principal components analysis attempts to explain maximal variance by defining trends or factors, I hypothesized that the factors would define the major modes of adaptation. The radiation of a group into a different environment from the one for which it is adapted must necessarily involve considera- ble genotypic and phenotypic change. Similarly, the evolution of faunal complexity is known to necessitate genotypic and phenotypic modification, even when the local fauna is made up of competing taxa that are not closely related (MacArthur and Levins, 1964). It can be hypothesized that these adaptive genotypic shifts would account for major variational trends, where- as local variation, mosaic patterns, and minor adaptive changes would not. In a study involving extensive faunal complexity with sympatry of ecologically similar but distantly related groups, the factors derived from principal components would be of more value in explaining ecological relationships than phylogenetic ones. An example might be a multivariate study of the trends in which seed-eating birds of an area differed. The trends which explain maximal variance would be those facilitating the avoidance of competition and successful occupation of the area while exploiting a common resource type. However, in a study like the present one, in which the sympatric and parapatric species are evolutionarily closely related, the principal components would be of both evolutionary and ecological utility to the investigator. The hypothesis can be tested by comparing the grouping and ordering of factor scores for the OTUs on the individual vectors with patristic and cladistic events indicated on the evolutionary construct. Acceptance of the hypothesis should depend on two criteria: 1) the concordance of trends with cladistic and patristic events; 2) demonstrable association of distributional or ecological trends producing causal selec- tive pressures. The latter is attempted through correlation of the morphological factors with ecological ones. It should be remembered that principal components are computed along axes of greatest variation of the data set, not any one character. Thus, OTUs positioned at opposite ends of the vector are most likely to differ in the particular set of characters emphasized by the vector. On the scutellational vector (Fig. 5) the three OTUs having the highest factor scores are terminally located on the left evolution- ary branch of the dendrogram; these OTUs have low scutellational values. The immedi- ately adjacent OTUs on the vector include the remaining OTUs of the left branch in addition to two of the most basally positioned OTUs of the right evolutionary branch. The other end of the scutellational vector includes all terminally positioned members of the right evolutionary branch; these populations have high actual scutella- tional values. The OTUs positioned interme- diately on this vector are members of the central evolutionary branch and the most basal members of the right and left evolutionary branches. Thus, the scutella- tional vector is concordant with the large patristic distance between terminal right and left branches and approximates the cladistic event separating these two branches. The second multivariate vector (Fig. 5) is one reflecting caudal morphology and cephalic scutellation. The three OTUs having the highest factor scores represent terminal members of the left branch except for Stenocercus cupreus (HUAN) which has a lower factor score. Stenocercus cupreus differs in several respects from the other species on the terminal left branch and has a large patristic distance from them. Adjacent on the second vector but distinctly sepa- rated are the most terminal right branch OTUs. Other OTUs are positioned on the lower end of the vector; these are patristically closer to the ancestral condi- tion. This vector indicates a marked change from the ancestral condition in caudal and head morphology in terminal OTUs of the evolutionary construct; two groups are approximated on the positive 26 THOMAS H. FRITTS (upper) end of the factor. The positioning of the OTUs on this factor indicates a closer relationship between terminal members of the phyletic construct than the dendrogram illustrates or convergent (or parallel) evolu- tion with regards to this subset of the data. The variance "explained" possibly reflects the change of head and tail morphology from the primitive condition to one or more derived sets of character states. The third morphological factor de- scribes dorsal crest and snout morphology (Fig. 6) and has OTUs less markedly grouped at extremes of the vector than previous factors. However, examination of the OTUs at the extremes of the vector is informative. The populations having the highest factor scores are those basally positioned on the right and left branches of the dendrogram; those having the lowest factor scores include all members of the central evolutionary branch, except for CAJA which has a more intermediate factor score. The OTUs of the basal portions of the right and left branches are patristically similar to the OTUs of the central evolutionary branch but are those most diametrically positioned from OTUs of the central branch on this factor. Terminal OTUs of the right and left branches differ from the OTUs of the central branch on the basis of other factors, but are intermedi- COMPONENT NO. 1 SCUTELLATION COMPONENT NO 2 TAIL and HEAD _1 8_ Chap _1.6_ _1.0_ Ayac _ .8_ Huan Chac _ .6 _ _ .4 _ —- Caja ^ -S- Chav _0.0_ _-.2_ Coma _-.4_ Cele _..6_ Giro _-.8_ _-1.0_ land 1 _-1.2_ _-1.4_ Mach 1 nla _-1.6_ _-1 .e_ _-2.0_ Tira Toto Otuz Rich Mejo Yaup Lima Chiq Huar Pale Unio Sarg Chuc- Mach ' Sarg- Huan - Palc- Pauc - Loja ^ Chav- Huar ' Ayac - _1.6_ _1.4_ _1.0_ _ .8_ _.6_ _A_ _.2_ _0.0_ -.2 _-.4_ -.6 _-.8_ _.1.0_ _-1.2_ _-1.4_ _-1.6_ Chap Tira Cele Giro Alam Huam Caja Coma Otuz Rich Chiq Unio Mejo Yaup Figure 5. Diagrammatic representation of first and second orthogonal components of morphological data. Refer to Fig. 1 for position of OTUs on the dendrogram. ANDEAN IGUANID LIZARDS 27 ately positioned on this factor. Thus, this factor is concordant with the cladistic event separating the right and left branches from the central branch. The fourth morphological factor empha- sizes thigh and head morphology (Fig. 6) and is of interest from two different viewpoints. The highest factor scores are those of basal and intermediate members of the left branch. The lowest factor scores are those of OTUs on the basal right branch. Terminal OTUs of the left branch also have lower factor scores than basal and interme- diate OTUs of the left branch. On this factor, basal and intermediate OTUs of the left branch differ from the OTUs which are patristically closest to them on the dendro- gram. The most terminal OTUs of the left branch are from intermediate elevations of interandean and Amazonian slopes, whereas basal and intermediate members of the left branch are from lower elevations of Amazonian slopes. The OTU most closely positioned with basal left branch OTUs (those having high factor scores) is MACH which also occurs on moist Amazonian slopes. MACH, representing S. crassi- caudatus and occupying an evolutionary position in the distal right branch, is the only OTU of that branch inhabiting southern Amazonian slopes. In this component it most closely resembles geographically and COMPONENT NO. 3 CREST and SNOUT COMPONENT NO 4 LIMBS and HEAD _1 6_ Coma _1.6_ 1.4 Yaup _1.2_ _1.0_ _ 8 _ _ ,6_ PIch Pauc _ .4 _ Alam _ -2 _ Mach _0.0_ Loia Chuc _■ 2_ Tolo L--"- Cele _-,6_ Lima _-.8_ _-1.0_ Ayac Mejo Tand Sarg Giro Huan Bamb Caja Chac Tira Chap- Chiq Unio- Alam- Loja- Pauc- Tira^ Sarg- Giro- Chav- Palc- Huan- Mejo - - Yaup - Mach - Huam - Chuc - Tand — -] Chiq — Coma — Toto — Caja — Lima — Bamb '^ Huar - Ayac Figure 6. Diagrammatic representation of third and fourth orthogonal components of morphological data. Refer to Fig. 1 for position of OTUs on the dendrogram. 28 THOMAS H. FRITTS ecologically similar, but patristically distant, OTUs. On the basis of neck morphology, which influences the fifth vector (Fig. 7), the most basal members of the central evolutionary branch are diametrically positioned from the most patristically similar OTUs on the terminal central branch. The former have the lowest factor scores for this vector; other central branch OTUs and basal OTUs of right and left branches have higher scores. This vector emphasizes morphologi- cal differences within the central evolution- ary branch. Those OTUs with low factor COMPONENT NO. 5 NECK MORPHOLOGY Chiq- Toto- Mejo- Ayac- Pich- Giro- Chap- Bamb- Palc' .of - Tira -Loja - Unio - Yaup -Sarg - Mach -Otuz -Huan I KM Caja- -2.9_ _-3.1_ Figure 7. Diagrammatic representation of fifth orthog- onal component of morphological data. Refer to Fig. 1 for position of OTUs on the dendrogram. scores are the only members of the central branch known to occur in sympatry with other iguanid lizards. Although this factor serves to differentiate only a few taxa in the present sample, several species from Ecua- dor having neck morphology similar to the basal central branch OTUs remain to be added to the analysis. Several of the Ecua- dorian taxa occur in sympatry with other iguanids, including other members of the genus Stenocercus. The results of the principal components analysis tend to describe patterns similar to major cladistic and patristic events of the Wagner dendrogram. Each factor, in being independent of others, defines a different trend. Factors tend to differentiate between patristically close, but cladistically distinct, groups. This supports the hypothesis that the major factors will be ones describing evolutionary trends. The OTUs not involved in a particular divergence being defined, because of other aspects of their biology, would be expected to show little sensitivity to the variational trends accounting for major portions of the variance. The intermediate positions of distantly related species in some of the trends support this interpretation. MORPHOLOGY AND ECOLOGY In order to examine further the relationships of the multivariate factors to evolutionary events and to estimate any adaptive shifts, ecological data are incorpo- rated into the analysis. Factor scores for the multivariate factors derived independently from ecological and morphological data are examined for correlations. The significant correlation of trends in scutellation with latitude-drainage (Fig. 2) can be interpreted through examination of the individual variables. Lizards from southern Amazonian regions tend to have fewer scales than those from northern interandean and Pacific drainages. As noted above, the scutella- tional factor shows morphological differ- ences between southern and northern species. Geographically intermediate regions are intermediate on the factor. Since this ANDEAN IGUANID LIZARDS 29 factor explains 28 percent of the morpholog- ical variance and is markedly concordant with the primary branching pattern of the dendrogram, further investigation of selec- tive trends in scutellation is warranted. Rock-dwelling species on the right branch occur largely in sympatry with congeneric species from the central or left branches. Rock-dwelling species on the left branch occur in sympatry with terrestrial or arboreal species of other iguanid genera. Comparison of sympatric pairs of congeneric species indicates in each case the two species are quite widely separated from each other patristically. Field observations indicate that each of the pairs includes one species that is basically terrestrial or confined to small solitary rocks in its activity. Both sexes of the other species feed and bask in large rock piles or on rock walls or rocky cliffs. In each case the terrestrial species occupies a basal position on the dendrogram, whereas the rock-dwelling species is more distally positioned. NICHE THEORY AND EVOLUTION The canonical correlation analysis dem- onstrates a relationship between precip- itation-habitat, latitude-drainage and mor- phological factors of tail, head, and scutella- tion. These same morphological factors are ones in which sympatric congeneric species differ abruptly. Those inhabiting rock habitats usually have more numerous head scales, larger mandibular musculature, spiny tails with a reduced number of caudal whorls per autonomic segment of the tail and more numerous scales on limbs and body. The occurrence of high factor loadings for precipitation and habitat variables in the first ecological factor does not mean that rock or elevated habitats are more common in moist areas, but that they are used more exclusively by species in moist areas. The occurrence of rock cliffs and outcrops depends largely on the geological character- istics of a locality, although precipitation and other climatic variables do influence weathering and soil formation. Although the more highly evolved species (distally posi- tioned on dendrogram) are seldom seen feeding or basking in other than rock habitats, even when terrestrial species are rare or absent, the converse is observed. In localities where no specialized rock-dwelling species occur, the terrestrial, or less highly evolved species utilizes terrestrial as well as rock habitats for feeding and basking. Males show a preference for rock or elevated habitats, whereas females are predominant on the ground. For example, Stenocercus ochotti, in the Rio Urubamba Valley of Peru, is found on the ground, on solitary rocks, and on large rock piles at localities where no other iguanid lizards occur. But where S. ochoai occurs in sympatry with the rock- dwelling species, S. crassicaudatus, S. ochoai is observed only on the ground and on low solitary rocks. Other species pairs have similar ecological relationships. In the terminology of Hutchinson (1957) the funda- mental and realized niches of the rock- dwelling species are approximately equal (with reference to habitat), whereas the realized niche of terrestrial species in areas of sympatry with other iguanid species is a subset of the fundamental niche. Although the terms generalized and specialized are often sources of confusion, I think that they describe the terrestrial and rock-dwelling species, respectively, with reference to both morphological attributes and ecological relationships. Utilization of terrestrial habitats can be observed in species at nearly every locality. Terrestrial species utilize terrestrial and rock habitats in allopatry or only terrestrial surfaces in allopatry or sympatry. Exclusive utilization of the rock habitat is confined to specialized species in more mesic areas. Examination of these observations in the light of MacArthur and Pianka's (1966) comments on the optimal use of a patchy environment allows a testable hypothesis to be formulated with respect to distribution of sympatric species pairs and relative abund- ance of individual species. Among the factors they list for increased specialization are:l) greater food density and 2) increased mobility of the animal in feeding activity. Studies by Terborgh (1971) in the Cordillera 30 THOMAS H. FRITTS Vilcabamba of Peru have shown increased species diversity and density of insectivor- ous birds on an altitudinal and moisture gradient. Other studies measuring general productivity along moisture gradients (Whittaker, 1970) have recorded increased productivity and energy flow with increased precipitation, especially at the lower ex- tremes of such gradients. Since insects are important primary and secondary con- sumers, a similar increase in insect produc- tivity and biomass might be expected in more mesic environments. Such an increase in insect abundannce along a moisture gradient would result in greater density of food for insectivorous lizards. Increased morphological specialization {e.g. increased number of toe lamellae) for mobility and efficiency of prey capture in a subset of habitat types would tend to increase the ability to obtain food. Although documentation is scant for lizards, studies by CoUete (1961) and others have suggested selective pressures on such morphological attributes in nonterrestrial lizards. MacArthur and Pianka considered relative search and pursuit times for a species and reasoned that if search time is long, pursuit time must be relatively short; under such conditions the species loses little search time while pursuing prey. In an area of low food density, such a relationship would be advantageous. However, if pursuit time is long due to difficulty in reaching prey, or the percent of prey captured is low, a high food density resulting in low search time would be necessary for exploitation by a species as a coarse-grained resource. Such a relationship exists on the irregular and vertical surfaces of rock cliffs and rock piles. The positive correlation between thigh- head morphology and precipitation-habitat indicates that lizards occurring in rock or arboreal situations and in more mesic localities possess smaller postfemoral pock- ets, fewer denticulate scales on the anterior margin of the ear and larger scales on the posterior thigh than terrestrial and more xeric adapted lizards. Another morphological trend, caudal and cephalic morphology, is correlated with precipitation gradients and habitat exploita- tion. All species with spinose caudal scutellation, except Stenocercus cupreus, frequent rock habitats and largely avoid activity on the ground surface. Species in which both sexes are known to frequent rock habitats in a coarse-grained manner (sensu MacArthur and Pianka, 1966) tend to lack parietal eyes, have more scales on the posterior margin of the dorsal head surface, smaller head scales, and smaller scales on the lateral body. Spinose tail scutellation is known in rock-dwelling lizards of several families (Bellairs, 1970). The reduction of lateral scale size may be adaptive in allowing the lizards to escape predators in rocks. Small scales allow the body to be inflated, wedging it within rock crevices and thus resisting extraction by predators. Similar behavior is known in Cordylus and Uromastix (Bellairs, 1970). The increase in number of head scales and decrease in size of the individual scales is accompanied by enlargement of the masseter musculature and head size in rock-dwelling species. In taking shelter in irregular crevices, biting may be a major defense of rock-dwelling lizards. Increased jaw musculature and head size would facilitate biting of a predator that could enter the confined spaces where these lizards hide. Little is known concerning the differences in physiology and basking behavior between rock-dwelling and more terrestrial lizards. Norris (1958; 1967) discussed thermoregulation and compara- tive aspects of behavior, habitat, and color pattern in relation to temperature. The loss of the unpigmented spot of the interparietal evidencing parietal position may be related to thermoregulatory differences between rock and ground dwelling lizards. Since reduction of all original variables to a few robust multivariate variables would tend to limit flexibility of relationships defined between ecological and morphologi- cal data sets, canonical correlation is also employed. The first canonical correlation is defined between thigh-head, scutellation, crest-snout morphology, and latitude- drainage, altitude-temperature (Table 4), and emphasizes differences between left ANDEAN IGUANID LIZARDS 31 branch OTUs and all others. Only the sample of Stenocercus cupreus (HUAN) approaches the cluster of right and central branch OTUs in position in Figure 3. OTUs from more southern Amazonian latitudes and in progressively lower and warmer environments exhibit aspects of scutellation, thigh-head and crest morphology that covary -with environmental and biogeo- graphic parameters. Recall that two of the morphological factors, scutellation and thigh-head morphology, emphasize differ- ences between OTUs on the left branch and OTUs on other branches; the multiple correlation of these factors with distinctive environmental combinations and the group- ing of OTUs on the left branch apart from central and right branches possibly indicate that these vectors are measuring variation resultant from adaptation to environmental and ecological parameters of southern Amazonian localities. The second canonical correlation of scutellation and caudal-head morphology with precipitation-habitat and latitude- drainage (Table 4) is statistically independ- ent of the previous canonical correlation. In this correlation (Fig. 4) OTUs in progres- sively more mesic and northern localities and those utilizing rock habitats display more marked specialization in scutellation and tail morphology than others. The terminal OTUs of the right evolutionary branch have the highest canonical scores for both the ecological and morphological data sets. OTUs on the basal right branch and those on the left and central branches have lower canonical scores for both data sets. Those OTUs on the left branch with specialized caudal and head morphologies are distinctly separated from, but are closest to, the terminal OTUs on the right branch. This relationship can be viewed in terms of sympatric relationships of the various OTUs with other iguanid lizards. Those OTUs in sympatry with other congeneric species occur in more mesic areas than those in allopatry. All of the pairs of sympatric specis are composed of one species from the lowermost positions of Figure 4 (having generalized tail, head and scutellational morphology and terrestrial in habitat preference in sympatry) and one other species from the upper right of Figure 4 (having speciaUzed caudal, head and scutellational morphology, and occurring in rock or arboreal habitats). Pursuit of prey on rock surfaces would involve more time and energy than on terrestrial substrates and consequently one would expect fewer successful captures per prey item sighted. Fine-grained utilization of terrestrial and rock habitats (in the frequency in which they are encountered) would be feasible in areas of low food density, such as xeric environments. Coarse-grained utilization of one or the other habitats through morphological and behavioral specialization would be adaptive only if food resources were sufficient or patch size large enough to reduce the necessity of migration from one patch to another (MacArthur and Levins, 1964; Levins, 1968). Species of Stenocercus that are termed terrestrial species are generalized "search- ers" exploiting open areas, bases of shrubs, and the smallest rock patches in sympatric situations, all areas where insects are low in density. In allopatry the terrestrial species treat habitat types as fine-grained re- sources, although males are more common on rock piles, and females are usually on the ground. The rock-dwelling lizards are found in moist areas where food density is sufficient to allow exploitation of rock habitats as coarse-grained resources. They are probably more efficient than terrestrial lizards in the pursuit of prey in such habitats. Terminal members of the right and left branches are most patristically distant from the inferred ancestor and treat the patchy habit as a coarse-grainned resource in utilizing only rock habitats. Most basal members of right and left branches as well as most members of the central branch occur in xeric habitats and utilize the patchy habitat as a fine-grained resource. Whereas sexual dimorphism in size of adults of the rock-dwelling species is minimal, terrestrial species show marked 32 THOMAS H. FRITTS size differences between sexes. Sexual dimorphism in size has been shown by Schoener (1967) to be greater in areas where intraspecific competition is greater than interspecific competition. This possibly results in greater niche breadth. Lizards exploiting limited resources or having little interspecific competition for similar re- sources would tend to be sexually dimorphic in size, allowing each sex to prey on different sizes of food. The occurrence of sexual dimorphism in size in only those forms occurring in allopatric populations and those occupying terrestrial habitats indirectly supports the previous discussion of habitat exploitation in relation to productivity and food abundance. Food is abundant in rock habitats in mesic environ- ments and rock-dwelling lizards are not sexually dimorphic. Food is scarce in terrestrial habitats causing intense intraspe- cific competition and favoring increased variation in size. Allopatric populations are largely confined to xeric environments which similarly have low food levels causing intraspecific competition and favoring in- creased variation in size, in this case sexual dimorphism. The analysis of feeding habits of various species of Stenocercus will serve as a test of the hypothesis of fine-grained and coarse-grained use of habitat types. Quanti- tative measurement of food density in terrestrial and rock habitats at mesic and xeric localities should indicate greater food density in rock habitats and at mesic locali- ties. Analysis of prey items of sympatric species will allow differences resulting from habitat selection to be substantiated. Simi- larly, differences in prey item size between males and females of sexually dimorphic species should exist. The food habits of Stenocercus are currently being studied in preserved material. Another test of the hypothesis will be the addition of other species in the multivariate analysis. Similar ecological relationships should exist in other species not previously analysed. The convergent evolution of two groups of rock-dwelling lizards indicated in the evolutionary construct is understandable in terms of niche theory and resource utiliza- tion. The amount of precipitation on the Pacific slopes of northern Peru is positively correlated with altitude. The higher eleva- tions are those with the most mesic conditions, and supporting sympatric spe- cies of Stenocercus. In the Amazonian slopes of southern Peru and Bolivia the amount of precipitation is more independent of altitude but tends to increase at lower elevations. The area around Cochabamba, Bolivia is an exception in having high rainfall at high elevations (Mendez, 1965). Terminal OTUs of the right branch of the dendrogram occur at higher elevations in northern Peru and southern Ecuador, and are sympatric with other species of Steno- cercus. Terminal OTUs of the left branch are distributed at lower elevations on the Amazonian slopes of central and southern Peru and Bolivia. The OTUs representing S. marmoratus are from high altitudes of Bolivia in the extremely mesic region in the Departamento Cochabamba. Thus converg- ent evolution of the right and left branches toward a specialized morphology adaptive in different mesic envvironments is under- standable. Both lines have evolved lizard species adapted for coarse-grained utiliza- tion of a patchy environment, the right branch species at high altitudes in northern Peru and southern Ecuador, and the left branch species at low elevations in southern Peru and Bolivia. Terminal members of the right and left branches are similar only in those aspects of morphology related to rock dwelling. Both lines have resulted in some lizards exploiting rock habitats in sympatry with other iguanids. The evolution of faunal com- plexity, through evolution of sympatric species pairs includes much of the evolution- ary change within the genus. ANDEAN IGUANID LIZARDS 33 EVOLUTIONARY CONCLUSIONS Three evolutionary lines are delimited within Stenocercus. The right branch of the dendrogram contains species from central and northern Peru and from Ecuador. Basal members on the right branch are terrestrial species which occupy xeric environments and occur allopatrically in relation to other species of Stenocercus. Terminal members of the right branch are saxicolous (rock- dwelling) species inhabiting mesic environ- ments at higher elevations and occurring in sympatry with other species of Stenocercus or other tropidurine genera. The central evolutionary branch includes species known to occur in sympatry with species on the terminal right branch. Some members of the central branch occur on the coastal slopes of central and northern Peru. Members of the central branch are terres- trial and patristically similar to the inferred ancestor. The left evolutionary branch includes species inhabiting Amazonian slopes of Peru and Bolivia. Terminal members of the left branch are similar to those of the terminal right branch in occurring in mesic environ- ments in sympatry with other iguanid lizards, in occupying rock habitats, and in having morphological adaptations for saxi- colous habits. Species included in terminal positions on the right and left branches differ in most aspects of scutellation and are convergent only in those characteristics associated with saxicolous habits. Basal members of the left branch are terrestrial and occur in allopatry or in sympatry with species on the terminal right branch. The evolution of rock-dwelling species occurring in sympatry with other species is likely in response to increased productivity in mesic environments. The presence of sexual dimorphism in size in terrestrial species suggests increased intra- specific competition for food in terrestrial habitats and indirectly supports the hypoth- esized relationship between productivity and grain size in habitat utilization. The environment inhabited by Stenocercus is a patchy one in relation to rock outcrops and insect density. Evidence available suggests that allopatric species utiHze the patchy habitat as a fine-grained resource in xeric areas where food density is low. Sympatric species are limited to mesic areas with greater food density and utilize patches in a coarse-grained manner. The evolution of two groups of rock- dwelling lizards is likely a result of high productivity and food density in rock habitats in two different biogeographic zones, one in subtropical Amazonian low- lands and the other in temperate montane areas with high precipitation. More detailed investigations are needed to measure the effects of variation in patch size and patch density. Some species, S. guentheri and S. chrysopygus, occur in allopatry at mesic localities where few rocks are exposed. Whether the size of rock patches is too small or rock patches are too infrequent to allow sympatric species to treat them as coarse-grained resources remains to be determined in a future study. The evolution of the diverse genus Stenocercus has occurred in an ecologically mosaic region. The basic trend of evolution is from allopatrically distributed species in xeric environments to more specialized species occurring in sympatry in areas of high productivity and in allopatry in uniform or xeric environments. The restric- tion of Stenocercus to Andean regions indicates that most of the evolution of the genus likely occurred as a result of the Pliocene-Pleistocene orogenies of the central Andes. Ecological succession during early stages of mountain building likely pro- gressed from xeric environments to geo- graphically variable extremes of precipita- tion. The occurrence of alternating climatic shifts in tropical regions during the Pleisto- cene simultaneous with the last major uplift of the Andes results in a complex and still poorly understood history for the central 34 THOMAS H. FRITTS Andes (Vuilleumier, 1971). Present evidence indicates that climatic and ecological factors were of paramount importance in the evolu- tion of Stenocercus. Future studies with emphasis on past and present ecology of the Andes should further understanding of the evolution of montane tropical communities. Quantitative measurement of morphological and ecological variation and associated selective regimes with less emphasis on gene flow and geographic isolation is a short step toward such an understanding. TAXONOMIC ACCOUNT Peters and Donoso Barros (1970) and Etheridge (1970) followed previous workers in considering the species of Ophryoessoides to be generically distinct from those of Stenocercus on the basis of the antehum- eral, caudal and dorsal crest morphology. Some Stenocercus have large spinose caudal scales, whereas others have small imbricate caudal scales similar to those of all species presently included in the Ophryoessoides. Most species of Stenocercus have a conspic- uous antehumeral fold, although one species is known to lack this fold. The majority of species of the genus Ophryoessoides have no antehumeral fold, but four do have a fold. Granular scales are present on the inner surfaces of antehumeral fold in all species of Stenocercus, except the single species lacking an antehumeral fold. Imbri- cate scales are found in the antehumeral region of most Ophryoessoides, although granular scales are known in three species. The dorsal crest is absent in some Stenocercus; the crest is present in all Ophryoessoides. The above characters have been empha- sized by previous authors (Etheridge, 1970, 1966; Boulenger, 1885) but no one character separates the two genera. Ether- idge (1966) pointed out that several species previously considered to belong in Ophry- oessoides or Liocephalus might be assign- able to the genus Stenocercus. Within Ophryoessoides (sensu Etheridge, 1966) are several species occurring in lowland for- ested regions below elevations of 1000 m. These species have large supraoculars, large dorsal scales, keeled ventral scales and relatively long and thin or laterally compressed tails. They also have inscrip- tional ribs which meet on the ventral midline; such an inscriptional rib pattern was noted by Etheridge (1965) as distinct from the pattern of other tropidurine genera and other species of Ophryoessoides. Most are arboreal. Included in this group is the type species of the genus Ophryoes- soides, 0. tricristatus Dumeril (1851), as well as O. aculeatus (O'Shaughnessy, 1879), 0. caducus (Cope, 1862), 0. erythrogaster (Hallowell, 1856), 0. iridescens (Giinther, 1859b) and 0. scapularis (Boulenger, 1901). The other species of Ophryoessoides resem- ble species of Stenocercus in having small supraoculars, smaller dorsal scales, smooth ventrals and in most species more robust tails. These species have inscriptional ribs which do not meet on the midventral line. These species of Ophryoessoides are largely confined to montane forests, grasslands and dry interandean valleys, as are most species of Stenocercus. They occupy the ground substrate or rock piles and seldom frequent tree trunks or tall shrubs. Therefore, it is clear that some members of the genus Ophryoessoides, including the type species, differ in scutella- tion, rib morphology and ecology from other species included in the genus. These other species resemble species of Stenocercus in all of the above aspects of morphology and ecology. Thus, the present generic arrange- ment does not reflect the evolutionary affinity between montane Ophryoessoides and Stenocercus. The morphological diverg- ence of lowland species of Ophryoessoides ANDEAN IGUANID LIZARDS 35 from other tropidurines and especially from Stenocercus and montane Ophryoessoides casts doubt on their relationships within the tropidurine group of the Iguanidae. In order to present a taxonomic arrangement con- sistent with evolutionary relationships, as I understand them, redefinition of the genus Ophryoessoides is necessary. A more restricted genus including the lowland species and any other having the diagnostic characters presented herein is proposed. Ophryoessoides Dumeril, 1851 Ophryoessodoides Dumeril, 1851:66 [Type species. — Ophryoessoides tricristatus Dumeril, 1851, by mono- t.yp.vl- Brachysaurus Hallowell, 1856:232 [Type species.— BrachysauTus erythrogaster Hallowell, 1856, by monotypy]. Srnrtisciis Cope, 1862:182 [Type species. — Scartiscus caducus Cope, 1862, by monotypy]. Diagnosis.— Small to moderate sized lizards having the following combination of character states: 1) interparietal small or absent; 2) supraoculars enlarged; 3) poste- rior head scales large; 4) antehumeral fold absent; 5) gular fold absent; 6) ventral scales keeled; 7) dorsal scales large, imbri- cate; 8) subdigital lamellae keeled; 9) femoral and preanal pores absent; 10) tail long, thin, or laterally compressed; 11) paired inscriptional ribs joining one another on midventral line; 12) sternal fontanelle large. Ophryoessoides is the only tropidurine genus having pairs of inscriptional ribs joining at the midventral line (Etheridge, 1966). Ophryoessoides also differs from Stenocercus in having large posterior head scales, large supraoculars, and keeled ventrals. Content. — As defined here, the genus contains six nominate species: 0. tricris- tatus Dumeril, 0. aculeatus (O'Shaughn- essy), 0. caducus (Cope), 0. erythrogaster (Hallowel), 0. iridescens (Giinther) and 0. scapularis (Boulenger). Two undescribed species from Amazonian Peru possibly are members of the genus. Distribution. — The species of the genus are known in peripheral Andean regions from the Sierra Nevada de Santa Marta, Colombia (Ruthven, 1922), south- ward to Paraguay (Etheridge, 1970). Most species occur at elevations below 1000 m east of the Andes; however, 0. iridescens inhabits the Pacific lowlands of Ecuador. Stenocercus Dumeril and Bibron, 1837 Stenocercus Dumeril and Bibron, 1837:349 [Type species.— Stenocercus rosei-ventris D'Orbigny, 1837 in Dumeril and Bibron, 1837, by monotypy]. Trachycyclus Dumeril and Bibron, 1837:355 [Type species.— Trachycyclus marmoratus D'Orbigny, 1837 in Dumeril and Bibron, 1837, by monotypy]. Steironotus (Stenocercus) Fitzinger, 1843:71 [Type species. — Stenocercus rosei-ventris D'Orbigny, 1837 in Dumeril and Bibron, 1837, by monotypy]. Heterotropis {Trachycyclus) Fitzinger, 1843:71 [Type species. — Trachycyclus marmoratus D'Orbigny, 1837 in Dumeril and Bibron, 1837, by monotypy]. Scelotrema Tschudi, 1845:154 [Type species.— Scelo- trema crassicaudatum Tschudi, 1845, by subsequent designation Etheridge, 1970]. Saccodeira Girard, 1857:197 [Type species. — Saccodeira omatissima Girard, 1857, by monotypy], new synon- ymy- Microphractus Giinther, 1859a:90 [Type species. — Microphractus humeralis Gunther, 1859a, by mono- type]. Diagnosis. — Small to moderate sized lizards having the following combination of character states: 1) interparietal small or absent; 2) supraoculars not enlarged; 3) pos- terior head scales usually small; 4) ante- humeral fold present or absent; 5) gular fold absent or if present caudal scales large and spinose; 6) ventral scales smooth or only weakly keeled; 7) dorsal scales small or moderate in size, those near middorsal region imbricate; 8) subdigital lamellae keeled; 9) femoral and preanal pores absent; 10) tail normal or spinose, not elongate or distinctly compressed laterally; 11) paired inscriptional ribs not joining one another on midventral line; 12) sternal fontanelle large; 14) free xiphisternal rods rarely present. Description.— Lizards of the genus Stenocercus are small to moderate in size. The head scales are not well differentiated, and the interparietal is small or absent. A parietal eye is evidenced by an unpigmented spot on the interparietal of some species, but it is absent or only sometimes present in others. A continuous vertebral scale row is 36 THOMAS H. FRITTS present or absent and in some species is modified so as to form a distinct dorsal crest. The dorsal scales are imbricate over the entire dorsum or only in the paraverte- bral region. The lateral body scales are nearly equal in size to the dorsal scales or are much smaller. An antehumeral fold lined with granular or imbricate scales is present only in some species. The scales on the posterior surfaces of the thighs are imbricate or granular. The caudal scales are imbricate, keeled, enlarged, and spinose. The species with spinose tails usually have only two whorls of scales between autonomic points of the tail. Femoral and preanal pores are absent. Males of many species have yellow or black coloration on the ventral surfaces of the thighs and pelvic region. Sexual dimorphism in dorsal and ventral coloration is present in some species. Consent.— Twenty-nine nominate spe- cies are assignable to the genus as presently defined. Two additional species are known but remain unnamed due to the lack of ade- quate material and distributional data. Whereas some species of Stenocercus have been included in the genus since their original descriptions, others have been considered to be in separate genera. In addition to the genera considered synony- mous with Stenocercus, species have been cited most often in combination with the generic names Liocephalus, Ophryoessoides, and Proctotretus. Key to Species of Stenocercus 1) Scales on posterior surface of thigh 7) Antehumeral fold conspicuous, lateral imbricate 2 venter of males not red or pink, no Scales on posterior surface of thigh posthumeral pocket granular 12 S. nigromaculatus 2) Caudal scales large, nearly twice the Antehumeral fold weak, lateral venter of size of ventrals, two caudal whorls per males bronze-red, deep posthumeral autonomic segment 3 pocket S. omatus Caudal scales small, approximately equal 8) Occipital scales larger than other head in size to ventrals, three or four cau- scales, wider than long, posterior dal whorls per autonomic segment margin bluntly curved S. apurimacus 4 Occipital scales subequal to other head 3) Vertebral row discontinuous, inconspic- scales, not bluntly curved on poste- uous; scales around midbody fewer rior margin 9 than 55 S. marmoratus 9) Scales around midbody 60 or Vertebral row continuous, forming more S. guentheri prominent dorsal crest; scales around Scales around midbody 60 or fewer ... 10 midbody more than 53 10) Gular region of males black, chin usually S. roseiventris not black; venter of males not pink; 4) Antehumeral fold present 5 posthumeral pocket weak or moder- Antehumeral fold lacking 8 ately developed 11 5) Neck folds present anterior to ante- Gular region of males pink; venter of humeral fold 6 males pink with black midventral Neck folds lacking 7 stripe from pectoral region to ventral 6) Three caudal whorls per autonomic seg- thighs; posthumeral pocket deep ment; scales around midbody fewer S. rhodomelas than 71 S. ochoai 11) Ventrals weakly keeled; pectoral region Four caudal whorls per autonomic seg- of males pink; posthumeral pocket ment; scales around midbody more lacking S. trachycephalus than 75 S. formosus Ventrals smooth in adults, weakly ANDEAN IGUANID LIZARDS 37 keeled in juveniles, pectoral region light pink, yellow or black; posthum- eral pocket present S. festae 12) Caudal scales large, nearly twice the size of ventrals, two caudal whorls per autonomic segment 13 Caudal scales small approximately equal in size to ventrals, three caudal whorls per autonomic segment 17 13) Vertebral row discontinuous 14 Vertebral row continuous from neck to base of tail 15 14) Caudal whorls subequal in size S. empetrus Caudal whorls alternating in size, smaller whorls one-half size of adja- cent whorls S. carrioni 15) Scales of vertebral row similar to adja- cent dorsals not forming distinct dorsal crest, scales around midbody fewer than 65 S. cupreus Scales of vertebral row distinct from adjacent dorsals, forming low dorsal crest; scales around midbody more than 70 16 16) All dorsal scales imbricate; dorsals meeting smaller lateral scales abrupt- ly in dorsolateral region . . . S. simonsii Dorsals near vertebral row imbricate, dorsals near dorsolateral region smaller, granular; dorsals gradually reduced laterally, grading into granu- lar lateral scales . . . .S. crassicaudatus 17) A neck fold anterior to antehumeral fold present 20 Neck fold lacking 18 18) Scales around midbody fewer than 40 S. moestus (part) Scales around midbody more than 40 19 19) Antehumeral fold lacking S. ivitus Antehumeral fold present S. melanopygus 20) Scales of vertebral row morphologically distinct from adjacent dorsals, form- ing a dorsal crest 25 Scales of vertebral row not morpholog- ically distinct from adjacent dorsals, no dorsal crest 21 21) Scales posterior to tympanum granular, head scales smooth S. chrysopygus Scales posterior to tympanum imbricate; head scales keeled or smooth 22 22) Head scales keeled S. orientalis Head scales smooth 23 23) Scales across gular region between tym- pana fewer than 30 24 Scales across gular region between tym- pana more than 35 S. haenschi 24) Scales around midbody fewer than 40 S. moestus (part) Scales around midbody more than 45 S. omatissimus 25) Scales around midbody fewer than 75 S. variabilis Scales around midbody more than 75 26 26) Black bar on ventral neck of males S. praeomatus No black bar on ventral neck of males 27 27) Black midventral stripe in males S. nubicola No midventral stripe in males 28 28) Subditial lamellae of fourth toe more than 35; scales around midbody more than 99 S. humeralis Subdigital lamellae of fourth toe 35 or fewer; scales around midbody fewer than 99 29 29) Posterior head scales morphologically distinct from adjacent scales of neck; one row of supraoculars between larg- est supraocular row and posteromedial border of semicircle S. boettgeri Posterior head scales similar to adjacent scales of neck; two rows of supraocu- lars between largest supraocular row and posteromedial margin of semi- circle S. varius 38 THOMAS H. FRITTS 66 '© ■ '. / .^ ^'^^ 12 fO) 16 200 400 600 Kilometers AREA ABOVE 3000 METERS 82 T8 Figure 8. Distribution of Stenocercus apurimacus (circles), S. praeomatus (squares), S. variabilis (triangles) and S. boettgeri (stars) in the Peruvian Andes. ANDEAN IGUANID LIZARDS 39 Stenocercus apurimacus Fig. 38 Stenocercus apurimacus Fritts, 1972:2. Type Material.— Holotype KU 134273 and paratypes KU 134270-134272, 134274-134309 from Puente Pacha- chaca, 15 km W Abancay, 1800m, Departamento Apurimac, Peru. Diagnosis.— 1) Posthumeral pocket deep; 2) postfemoral pocket absent; 4) scales posterior to tympanum keeled, imbri- cate, equal in size to dorsal neck scales; 5) scales on posterior thigh keeled, imbri- cate; 6) vertebral row continuous, forming high dorsal crest extending onto tail; 7) caudal scales normal, three caudal whorls per autonomic segment; 8) antehumeral fold absent; 9) neck fold absent; 10) scales around middle of body 48-61; 11) scales across gular region between tympana 18-26; 13) venter of males gray-beige to gray with a blue hue; 17) posterior head scales large, wider than long, rounded posteriorly, a pair of large parietals in broad contact poste- rior to small interparietal. Stenocercus apurimacus is the only species of Stenocer- cus occurring in southern Peru which lacks antehumeral and neck folds. Distribution. — The species is known in Peru from elevations of 1800-2700 m in the interandean valleys of the Rio Apurimac and Rio Pachachaca in the departments of Cuzco and Apurimac (Fig. 8). Ecological Observations.— The species occurs on the ground and on rocks in xeric habitats where Acacin, thorn shrubs and cactus form the predominant plant cover. Stenocercus apurimacus occurs in sympatry with S. ochoai at Curahuasi, 2700 m, Departamento Apurimac. Remarks.— Variation in the number of scales around the body and chin coloration was described by Fritts (1972). Stenocercus boettgeri Boulenger Fig. 35 Stenocercus boettgen Boulenger, 1911:22. Type Material.— Syntypes BMNH 1911.12.13.25-32 (RR 1946.8.11.92-99) from Huancabamba (Departa- mento Piura), Peru. Diagnosis.— 1) Posthumeral pocket absent; 2) postfemoral pocket present; 4) scales posterior to tympanum granular, less than one-half size of dorsal neck scales; 5) scales on posterior thigh granular; 6) ver- tebral row continuous, forming indistinct dorsal crest; 7) caudal scales normal, three caudal whorls per autonomic segment; 8) antehumeral fold present; 9) neck fold present; 10) scales around middle of body 78-91; 11) scales across gular region between tympana 43-55; 16) one row of supraoculars between largest row of supra- oculars and posteromedial margin of semi- circle; 17) scales across occipital region similar to other head scales, distinct from dorsal neck scales; 18) scales of nasal region large. Stenocercus boettgeri and S. varius have common ranges for several scutella- tional characters. They can be distinguished by the following characters: the scales between the nostrils of S. boettgeri are large; those of S. varius are smaller and more irregularly positioned. The largest supraocular row is separated from the interorbital scales along the posteromedial margin of the semicircle by one small supra- ocular row in S. boettgeri: two small supra- ocular rows He between the largest supra- ocular row and the posteromedial interor- bital scales in S. varius. The scales of the posterior margin of the dorsal surface of the head are notably larger in S. boettgeri than in S. varius (Fig. 10). The posterior head scales of S. varius are small and nearly granular, thereby resembUng the dorsal neck scales. Distribution. — Stenocercus boettgeri is known from the Amazonian drainage of central Peru (Paucartambo, 2900-3250 m, Departamento Pasco) and from extreme northern Peru (Huancabamba, Departa- mento Piura, at approximately 3000 m). Several species of Stenocercus are known to occur in the area between the extremes of the range of Stenocercus boettgeri, but no specimens of this species are known from the intermediate localities (Fig. 8). Due to inaccessibility, the eastern margins of the Rio Maranon remain poorly known and may eventually prove to support geographically 40 THOMAS H. FRITTS 76 100 200 300 Kilometers REA ABOVE 3000 METERS ELEVATION — 6 80 78 76 Figure 9. Distribution of Stenocercus guentheri (closed circles), S. haenschi (open circles), S. carrioni (closed squares), S. varius (open squares), S. simonsii (triangles) and S. humeralis (stars) in the Andes of Ecuador. intermediate populations of the species. The large distance between the populations from Piura and Pasco departments extends from the northern tributaries of the Rio Maranon valley across the Rio Huallaga drainage and into the upper regions of the Rio Perene. Ecological Observations. — Individuals of both sexes are active on large rock piles ANDEAN IGUANID LIZARDS 41 5. varius S. boet tg er i Figure 10. Dorsal cephalic scutellation of Stenocercus boettgeri, KU 134014, male, x 5. and S. varius, KU 132494, x 7. Arrows indicate anterior margin of tympanum. and rock walls. The dorsal coloration of yellow flecks on a black ground color closely approximates the color of the dark rocks partially covered with yellow lichens on which the lizards occur near Paucartambo, Departamento Pasco. The lizards seek refuge in crevices within rock piles and are capable of inflating the lungs so as to wedge the body tightly in the irregular spaces between rocks. Stenocercus boettgeri and S. nigromaculatus likely occur in sympatry near Huancabamba, Departamento Piura, Peru. The latter species is largely confined to the ground and rows of Agave. Remarks.— Little can be noted con- cerning the coloration of S. boettgeri in the northern populations. Boulenger (1911) presented the following notes: "Olive above with more or less indistinct darker cross bars". Boulenger did not see the specimens in Hfe and probably was using preserved specimens for the basis of this description. Specimens from Paucartambo, 2900-3250 m, in central Peru, are black dorsally with numerous yellow flecks and lines on the dorsal and lateral scales. Specimens from both central and northern Peru have a pinkish tint on ventral surfaces. Etheridge (1970) considered Stenocer- cus juninensis Shreve (1941) to be synony- mous with S. boettgeri. I have examined the type of S. juninensis (MCZ 45820) and find it to differ from S. boettgeri. in having fewer dorsals in paravertebral row from head to base of tail 87, (90-113 in S. boettgeri); fewer scales across gular region between tympana 35, (43-55 in S. boettgeri) and fewer subdigital lamellae on fourth toe 21, (25-28 in S. boettgeri). Stenocercus carrioni Parker Stenocercus carrioni Parker, 1934:268. Type Material.- BMNH 1933.6.24.75 (RR 1946.8.11.83) from Zamora, 3250 ft. Provineia Loja, Ecuador. The city of Zamora is presently included in Provineia Zamora, Ecuador. Diagnosis. — 4) Scales posterior to tym- panum granular, less than one-half size of dorsal neck scales; 5) scales of posterior thigh granular; 6) vertebral row discontinu- 42 THOMAS H. FRITTS ous in some specimens, dorsal crest visible on neck only; 7) caudal scales large, extremely spinose, alternating large, small, large, small, two caudal whorls per auton- omic segment; 8) antehumeral fold present; 9) neck folds present; 10) scales around middle of body 77-86. Of the species of Stenocercus with 2 caudal whorls per autonomic segment only S. carrioni, S. empetrus, and S. marmoratus lack complete vertebral rows. Stenocercus marmoratus differs from the S. carrioni in possessing imbricate scales on the posterior thigh and fewer subdigital lamellae. Steno- cercus empetrus differs in having caudal whorls which are subequal in size and in having a larger and more robust body. Stenocercus carrioni differs from S. simonsii in having caudal whorls markedly alternat- ing in size; the smaller caudal scales are approximately one-half the size of the larger ones. The absence of a complete vertebral row is also distinctive in S. carrioni in comparison to S. simonsii. Distribution. — Although the type spec- imen was reported from Zamora, Ecuador, on the Amazonian slopes, the only additional specimens are from the vicinity of Alamor, Provincia Loja, Ecuador, a locality in the Pacific drainage (Fig. 9) and adjacent regions of Provincia El Oro. The type was part of a collection given to the British Museum by Clodovo Carrion. Parker (1934) reported on many species in the Carrion collection; the report contains reference to several other species from "Zamora, 3250 ft.," which on the basis of present knowledge are doubted or known not to occur at the locality. Parker listed Atelopus ignescens from Zamora, although it is confined to higher elevations throughout the Andes of Ecuador; despite Parker's record from Zamora it is extremely doubtful that the Andean Pholidobolus montium occurs at elevations below 2000 m in southern Ecuador. I view the possibility of Stenocer- cus carrioni occurring on Amazonian slopes as slight. However communication between Atlantic and Pacific drainages through low passes in the Andes has been postulated for other species in southern Ecuador and extreme northern Peru (Schmidt and Walker, 1943). Ecological Observations. — Stenocercus ornatus and a species of Ophryoessoides are known from the environs of Alamor and are likely to be in sympatry with S. carrioni. On the basis of the phenetic similarity to many of the rock-dwelling species of the genus, S. carrioni is likely to confine its activity to rock outcrops and associated habitats. Stenocercus chrysopygus Boulenger Fig. 23-24 Stenocercus chrysopygus Boulenger, 1900:183. Type Material.— Syntypes BMNH 1900.6.20.8-17 (RR 1946.8.9.33-46) from Carao (Caraz), 8000 ft., BMNH 1900.6.20.18 (RR 1946.8.11.84) from Huaraz, 10,000 ft. and BMNH 1900.6.20.19 (RR 1946.8.5.98) from Recuay, 11,000 ft., Departamento Ancash. Peru. The type locality is herein restricted to Huaraz, 10,000 ft., Departamento Ancash, Peru. Diagnosis.— 1) Posthumeral pocket absent; 2) postfemoral pocket weakly devel- oped; 4) scales posterior to tympanum granular, less than one-half size of dorsal neck scales; 5) scales of posterior thighs granular; 6) vertebral row continuous or broken, not forming dorsal crest; 7) caudal scales normal, three caudal whorls per autonomic segment; 8) antehumeral fold present; 9) neck fold present; 10) scales around middle of body 54-82; 15) head scales smooth. Of the species having granular scales on the lateral surface of the neck and lacking a distinct dorsal crest, only Stenocercus chrysopygus and S. haenschi lack a spinose tail. Stenocercus chrysopygus differs from S. haenschi in having fewer than 32 scales across the gular region between tympana, fewer than 23 subdigital lamellae on the fourth finger and in lacking green colora- tion. Stenocercus chrysopygus resembles S. omatissimus in aspects of scutellation and body form, but differs from it in having a distinct postfemoral pocket, and granular scales on the lateral neck. It differs from S. orientalis in lacking keeled head scales and in having granular scales on the lateral neck. Distribution.— This species is wide- spread on both sides of the Cordillera ANDEAN IGUANID LIZARDS 43 78 66 16 12 — J 200 '& .'^^ 400 Kilometers 600 AREA ABOVE 3000 METERS '\ / ^ dp — 0 78 Figure 11. Distribution of Stenocercus ivitus (open circles), S. chrysopygus (closed circles), S. orientalis (open squares), S. moestus (closed squares), S. omatissimus (triangles), S. melanopygus (stars) in the Andes of central and northern Peru. 44 THOMAS H. FRITTS Blanca of central Peru (Fig. 11). Populations are known in the upper reaches of the Rio Chiquian and Rio Santa on the Pacific slopes of the Departamento Ancash at elevations of 2265 to 3500 m. Populations are known in the Atlantic drainage from La Union, 3100 m, Departamento Huanuco, and Chavin de Huantar, 3230-3500 m, Departamento An- cash. The Cordillera Blanca, the highest mountain range in Peru attains elevations of 4500-5000 m between the Atlantic and Pacific localities in this area of Peru. The populations from Caraz, Huaraz, and Recuay are from different elevations of the Rio Santa in the Departamento Ancash, and can be considered as a single panmictic unit. Chiquian is situated in the valley of the Rio Chiquian to the south and is separated from the geographically close valley of the Rio Santa by a ridge at 4300 m. Chavin de Huantar and La Union are localities in different subvalleys of the Rio Marahon. Ecological Observations. — Males are most often observed on small rocks or mud walls. Females are found under rocks and at the bases of Agave and small shrubs. No other lizards are known to occur in sympatry with this species. A species of Liolaemus occurs at higher elevations between the valleys of the Rio Santa and the Rio Chiquian. Several species of Tropidurus are known from lower elevations on the Pacific slopes of this region. Remarks.— The males of this species exhibit considerable variation in coloration within and between populations. Males from the valley of the Rio Santa (Recuay, Huaraz, Caraz) have gray-white venters with dull yellow coloration on the ventral thighs and ventral pelvic region. An inverted yellow triangle on the ventral pelvic region is present in some males. The chin and gular region are light with a variable blue suffusion and prominent gray-black blotches. The ventrolateral body of some specimens is light to medium brown with a Hght blue suffusion. The gray-brown dorsolateral stripes are indistinctly defined on head and neck or are absent entirely. Males from other localities (Chavin de Huantar, La Union, and Chiquian) have black coloration on the ventral pelvic region and thighs. The ventral base of the tail is yellow with black flecks or black in males from Chavin. The black ventral coloration extends anteriorly to the pectoral region. Some males from Chavin have a gray-black suffusion over a lighter ventral ground color. The chin and gular regions are light to medium blue with prominent black reticulations. Some specimens have a yellow or green suffusion on the apex of chin. The ventrolateral body is gray-blue to medium blue with black blotches. A prominent white or yellow stripe extends from the subocular region, passing dorsal to the tympanum, onto the shoulder. A short yellow or white stripe passes ventral to the tympanum from the subocular region to the base of the fore- limb. These stripes are most distinct in males from La Union and Chiquian. Females from all localities are similar in coloration. Most individuals have gray dorsolateral stripes and two longitudinal series of irregular dark blotches on the dorsum. Females have a gray middorsal stripe without black flecks or blotches. The venter is beige-white with scattered gray flecks. The populations studied are uniform in most aspects of the scutellation. The number of scales around the middle of the body and the number of transverse rows of dorsals from the head to the base of the tail vary geographically. These two characters are highly correlated and show similar patterns of variation. The Pacific popula- tions (Chiquian, Huaraz, and Recuay) possess similar numbers of dorsal scales whereas the La Union population possesses higher numbers and Chavin has the highest of all (Table 5). Individuals from La Union and Chiquian are markedly melanistic; lizards from Chavin are slightly melanistic but resemble lizards from Huaraz and Recuay in having the more typical dorsal coloration of terrestrial lizards of this genus. Lizards from La Union have a continuous vertebral row of scales on the body whereas lizards from other localities seldom have a continu- ous vertebral row. ANDEAN IGUANID LIZARDS 45 Table 5. Basic statistics of five samples of Stenocercus chrysopygus. Vertical lines connect non- significant (p < 0.05) subsets of SS-STP analysis. Sample N X Range SD 2SE^ Scales Around Middle of Body (SMB) Chavin 17 74.06 67-82 4.10 2.00 1 La Union 10 65.80 59-70 3.16 2.00 1 Recuay 37 60.59 55-68 3.20 1.05 Huaraz 22 58.23 53-65 3.28 1.40 Chiquian 15 58.07 54-62 2.46 1.27 Scales Across Gular Regi on Between Ty mpana (SAG) La Union 10 28.20 25-31 2.00 1.26 Chavin 17 27.65 25-30 1.62 0.78 Recuay 37 25.54 23-29 1.63 0.53 Chiquian 15 25.40 24-27 1.35 0.70 Huaraz 22 24.64 21-28 1.47 0.62 Because of the small number of samples available, few trends can be delimited with confidence. The two darkest color patterns occur in the most southern localities (Chiquian and La Union); lizards from Pacific drainages have fewer scales around the middle of the body than do those from Atlantic populations. Stenocercus crassicaudatus (Tschudi) Fig. 28 Scelotrema crassicaudatum Tschudi, 1845:155. Type Material.— Holotype MHNN unnumbered, from Peru; Tschudi (1846) further specified the type locality as Urubamba (Departamento Cuzco), Peru. Stenocercus torquatus Boulenger, 1885:133. Type Material.— BMNH 61.5.22.4 from Peru. Urocentrum meyeri Werner, 1904:4. Type Material.— The holotype (number unknown) was destroyed in the Dresden Museum during World War II. The type locality (cited as] Lima, Peru is probably erroneous. Stenocercus ervingi Stejneger, 1913:545. Type Material.— USNM 49550 from Huadquinia, 5000 ft. (Departamento Cuzco), Peru. Diagnosis.— 3) Dorsal scales near vertebral row imbricate and keeled, those near dorsolateral region granular, lateral scales granular; 4) scales posterior to tympanum granular; 5) scales on posterior thigh granular; 6) vertebral row continuous, forming low dorsal crest; 7) caudal scales large and spinose, two caudal whorls per autonomic segment of tail; 8) antehumeral fold present; 9) neck fold present; 10) scales around middle of body 85-119; 11) scales across gular region between tympana 50-59; 12) subdigital lamellae of fourth finger 30-37. Stenocercus crassicaudatus differs from all other species with spinose caudal scutellation in having more subdigital lamellae on the fourth finger, a dorsal coloration of dark gray with black spots, and usually having more scales around the body. Distribution.— Stenocercus crassicau- datus is known from elevations of 1000- 2500 m on Amazonian slopes of central and southern Peru (Fig. 12). On the basis of two specimens (UMMZ 68115) from Yungas de Cochabamba, Bolivia, the species is consid- ered to extend into Amazonian Bolivia. Whether this species has a continuous distribution from north-central Peru south to Amazonian Bolivia or not cannot be ascertained at this time. Invasion of the 46 THOMAS H. FRITTS 12 — 16 200 400 600 Kilometers AREA ABOVE 3000 METERS V ._..' 82 78 Figure 12. Distribution of Stenocercus crassicaudatus (circles), S. empetrus (squares) and S. nigromaculatus (triangles). ANDEAN IGUANID LIZARDS 47 interandean valleys of Rio Urubamba and Rio Apurimac has been recorded to eleva- tions of 2500 and 1400 m respectively (Mertens, 1952). Ecological Observations.— Stenocercus crassicaudatus occurs on large rock piles and rocky cliffs at elevations of 1000 to 2500 m. Stenocercus ochoai occurs in sympatry with S. crassicaudatus between 2000 and 2500 m in the Rio Urubamba valley. Stenocercus ochoai is confined to the ground, solitary rocks and small rock piles in areas of sympatry. Mertens (1952) recorded Stenocercus crassicaudatus and Liocephalus arenarius (Stenocercus ochoai?) in sympatry in the valley of the Ri'o Apuri- mac: Aina, 1400 m, Departamento Ayacu- cho, Peru. Remarks.— Specimens from Departa- mento Cuzco, Peru have numerous black spots on gray dorsal ground color on neck and body; those from Departamento Junin, Peru have a conspicuous black bar in the antehumeral region. The bar is occasionally complete across the dorsum forming a dorsal and lateral collar. One to three transverse black bars can be present on the neck. Specimens from central and southern Peru are similar in all scutellational attributes observed. Two specimens from Yungas de Cochabamba, Bolivia (UMMZ 68115) have fewer scales around the middle of the body (85 and 92) than any of the Peruvian specimens (more than 98 scales around body) of Stenocercus crassicaudatus. Stenocercus cupreus Boulenger Fig. 33 Stenocercus cupreus Boulenger, 1885:135. Type Material.— BMNH 76.7.4.4 from Huanuco, (De- partamento Huanuco), Perii. Diagnosis.— 1) Posthumeral pocket absent; 2) postfemoral pocket present; 3) dorsal scales keeled, twice as large as smallest laterals; 4) scales posterior to tympanum imbricate, one-half size of dorsal neck scales; 5) scales on posterior thigh granular; 6) vertebral row continuous, not forming distinct dorsal crest; 7) caudal scales large, not spinose, two caudal whorls per autonomic segment; 8) antehumeral fold present; 9) neck fold present; 10) scales around middle of body 52-64. This is the only species having two whorls per autonomic segment and lacking spinose caudal scales. Of the species with two caudal whorls per autonomic segment, only S. roseiventris and S. marmoratus have as few scales around the body as S. cupreus. Stenocercus cupreus differs from S. roseiventris in lacking pink ventral coloration of males and granular scales on the posterior thigh, and differs from S. marmoratus in having a complete vertebral row and granular scales on the posterior thigh. Distribution.— The species is known from the valley of the Rio Huallaga, Departamento Huanuco, Peru between elevations of 1900 and 2300 m (Fig. 13). Stenocercus chrysopygus occurs in the adjacent upper Maranon Valley; one pass between the two valleys is at an elevation of 3900 m. Liolaemus walkeri is known to occur at elevations above 3400 m in the Nudo de Pasco, where the source waters of the Rio Huallaga form. Ecological Observations. — The vegeta- tion of the xeric valley of the Rio Huallaga consists of scattered cacti, grasses, and low shrubs. Males are observed basking on rocks, at bases of cacti and shrubs and rarely in the lowest branches of shrubs. Females are frequently at bases of shrubs and under rocks. There is little vegetation cover on the slopes of the valley; no other iguanids have been observed in sympatry. Stenocercus empetrus Fritts Fig. 39-41 Stenocercus empetrus Fritts, 1972:7. Type Material.— Holotype KU 134394 and paratypes KU 134379-134393, 134395-134400 from Huamachuco, 3350m, Departamento La Libertad, Peru. Diagnosis.— 1) Posthumeral pocket absent; 2) postfemoral pocket large; 3) dorsals imbricate, smooth or weakly keeled; 4) scales posterior to tympanum granular, smaller than dorsal neck scales; 5) scales on posterior thigh granular; 6) vertebral row 48 THOMAS H. FRITTS Figure 13. Distribution of Stenocercus marmoratus (open circles), S. ochoai (closed circles), S. roseiventris (open squares), S. formosus (closed squares), S. cupreus (triangles) and S. nubicola (stars). ANDEAN IGUANID LIZARDS 49 discontinuous, only a few scales on neck and anterior body morphologically distinct from adjacent dorsal scales; 7) caudal scales spinose, enlarged, subequal, two caudal whorls per autonomic segment; 8) antehum- eral fold present; 9) neck fold well- developed; 10) scales around middle of body 87-115; 11) scales across gular region between tympana 39-49; 13) dorsal body dark brown or black, numerous yellow flecks and spots on dorsum in some populations. Three species of Stenocercus have spinose caudal scutellation and lack a continuous vertebral row. Stenocercus em- petrus differs from S. marmoratus in having more than 86 scales around the middle of the body, more than 38 scales across the gular region between the tympana, and granular scales on the posterior thigh. Stenocercus empetrus differs from iS. carrioni in having caudal whorls which are subequal in size and in having a more robust body. It differs from S. simonsii in lacking a vertebral row forming a distinct dorsal crest, in having smooth or weakly keeled dorsals, and in having a less spinose caudal scutellation. Distribution. — The species is known from the valley of the Rio Crisneja, adjacent tributaries of the Rio Maranon, and a single locality in the Pacific drainage in the upper valley of the Rio Moche (Fig. 12). It is distributed in the Peruvian departments of La Libertad and Cajamarca at elevations between 2650 m and 3200 m. Ecological Observations.— Individuals of both sexes are saxicolous and occur in sympatry with the terrestrial species Stenocercus melanopygus in Atlantic locali- ties. Stenocercus empetrus is likely sympat- ric with S. omatissimus on Pacific slopes of north central Peru. Remarks.— Considerable variation in dorsal coloration is evident (Fritts, 1972). Additional material is necessary for evalua- tion of morphological variation and relation- ships within the populations assigned to this taxon. Stenocercus festae (Peracca) n. comb. Fig. 37 Liocephalus festae Peracca, 1897:6. Type Material.— Three syntypes MZT (numbers un- known) from Cuenca, (Provincia Azuay), Ecuador; none of the syntypes could be located in Turin by Richard Etheridge in 1967. Diagnosis. — 1) Posthumeral pocket present; 2) postfemoral pocket well- developed; 3) lateral scales not greatly reduced in size in relation to dorsals; 4) scales posterior to tympanum nearly imbricate, keeled, equal in size to dorsal neck scales; 5) scales on posterior thigh imbricate, keeled; 6) vertebral row contin- uous, forming raised dorsal crest; 7) caudals normal, three caudal whorls per autonomic segment; 8) antehumeral fold absent; 9) neck fold absent; 10) scales around middle of body 47-60; 12) subdigital lamellae of fourth finger 16-22; 13) venter of males black; 14) gular region of males black. Stenocercus festae differs from all other species lacking an antehumeral fold, in having a black blotch on the gular region of males, a black venter in males, smooth ventrals in adults, and a posthumeral pocket. Distribution. — Stenocercus festae is known in Ecuador from throughout the Cuenca Basin (upper valley of Rio Paute) and the adjacent Pacific slopes at elevations from 2300 to 2900 m (Fig. 14). The valleys of the Rio Paute (Atlantic drainage) and Rio Jubones (Pacific drainage) are separated by a ridge less than 3000 m in elevation. All localities are in the provinces of Canar, Azuay, Loja and El Oro. Ecological Observations. — Stenocercus festae occurs in sympatry with S. simonsii in the upper Rio Jubones drainage. In this area S. festae utilizes the leaves of agave plants and surrounding ground surface for basking and feeding activity, whereas S. simonsii is restricted to large rocks or rock piles. Stenocercus rhodomelas and S. ornatus are known from geographically proximal regions. Remarks. — The juveniles of this species have weakly keeled ventrals, al- 50 THOMAS H. FRITTS 76 100 200 300 C Kilometers \REA ABOVE 3000 METERS ELEVATION 80 78 76 Figure 14. Distribution oi Stenocercus festae (circles), S. rhodomelas (squares) and S. omatus (triangles) in the southern Andes of Ecuador. though no trace of keeling remains in sub- adults or adults. The ventral coloration is variable between geographically proximal localities. Male individuals from the valley floor at 2500 m near Cuenca have a light yellow chin and lateral venter; some of those from 8 km NW Cuenca, 2620 m are similar whereas others have a solid black ventral surface from the snout to the vent. Males from SW of Cuenca near Cutchil, ANDEAN IGUANID LIZARDS 51 2500-2900 m, Provincia Azuay, have a light orange coloration on the chin and pectoral region. No notes on the color of live specimens are available for adults from the Pacific slopes, although juveniles have light yellow coloration on the labial scales. Preserved specimens from the Pacific slopes retain the black coloration on the gular and ventral regions. Specimens from Azogues, 2500 m, Provincia Canar, have more scales around the middle of body 59.3 and more scales across the gular region between tympana 28.5 than those from Cuenca, 2540 m, Provincia Azuay 50.9 and 24.7 respectively (p < 0.05; SS-STP). Stenocercus formosus (Tschudi) n. comb. Fig. 31 Scelotrema formosus Tschudi, 1845:155. Type Material.— Two unnumbered syntypes in MHNN from Peru; locality further specified by Tschudi (1846) as Tuliumayo, Montanas des Mittlern Peru. Liocephalus rhodogaster Boulenger, 1901:547. new syn- onymy. Type Material.- BMNH 1900.11.27.24-25 (RR 1946.8. 29.81-2) from La Merced, 3250 ft., Ri'o Perene, (Departamento Juni'n), Peru. Liocephalus lineogularis Werner, 1901:3. new synon- ymy- Type Material.— D 1781 from Chanchamayo, (Departa- mento Junin), Peru; type destroyed during World War IL Stenocercus seydi Andersson, 1908:301. new synon- ymy- Type Material. — Single specimen WNM (number un- known) from La Merced, 1000m, (Departamento Junin), Peru. Diagnosis.— 1) Posthumeral pocket absent; 2) postfemoral pocket deep; 3) lateral scales nearly equal in size to dorsals; 4) scales posterior to tympanum imbricate, less than one-half size of dorsal neck scales; 5) scales on posterior thigh imbricate, keeled; 6) vertebral row continuous, form- ing raised dorsal crest; 7) caudal scales normal, three or four caudal whorls per autonomic segment; 8) antehumeral fold present; 9) neck fold present; 10) scales around body 78-81; 11) scales across gular region between tympana 29-31; 13) venter of adult males pink with black midventral stripe. Of the species having imbricate, keeled scales on the posterior surface of the thigh, only S. formosus and S. ochoai have ante- humeral folds, neck folds and normal caudal scutellation. Stenocercus formosus differs from S. ochoai in having pink ventral coloration of adult males, in having more than 75 scales around the body and in having four caudal whorls per autonomic segment near the base of the tail. Distribution. — Stenocercus formosus is known from the "Chanchamayo" region of Amazonian Peru (Fig. 13). The distribution extends from 1000 to 1600 m in the departments of Junm and Pasco. Previous workers (Burt and Burt, 1930; Mertens, 1952) have confused this species with S. ochoai from the valleys of the Rio Apurimac and Rio Urubamba in southern Peru. Stenocercus formosus is confined to the drainage of the headwaters of the Rio Perene. Stenocercus ochoai occurs in the more southern valleys of the Rio Urubamba and Rio Apurimac and is not known to reach the Rio Perene drainage. Stenocercus formo- sus occupies several small valleys on the eastern slopes of the Andes. Ecological Observations.— Individuals are observed on the ground, in small shrubs and on small rock piles. They seek refuge in rock piles and holes around the base of shrubs. Montane tropical forest covers most of the geographic distribution of this species. No other Stenocercus species occurs in sympatry. Stenocercus praeor- natus. S. boettgeri, and S. variabilis are known from higher elevations in the same region. Remarks.— The trivial name of this species has been confused by all previous workers; Steironotus arenarius Tschudi (1845) was considered a senior synonym of Liocephalus rhodogaster by Roux (1907) and by subsequent workers. I have found Steironotus arenarius Tschudi to be a species of Tropidurus synonomous with Tropidurus tschudi Roux (1907). Evidence exists which indicates that Roux inadvert- antly used the type specimen of Steironotus arenarius Tschudi as the type specimen of Tropidurus tschudii. In the same bottle as 52 THOMAS H. FRITTS the specimen of Tropidurus named by Roux were two specimens which Roux considered to be the types of Steironotus arenarius and found to be conspecific with Liocephalus rhodogaster Boulenger (1901). The name Scelotrema formosum Tschudi (1845) has never been allocated by subsequent workers (Boulenger, 1885; Burt and Burt, 1933). Roux reviewed some of Tschudi's types but did not comment on S. formosum presum- ably because he did not find the type in the Neuchatel collection. Most authors have considered S. formosum likely to be allied to the "South American Liocephalus" Ophry- oessoides, or Stenocercus. Boulenger (1885) listed Scelotrema in the synonymy of Liocephalus although he failed to allocate S. formosum, perhaps because of uncer- tainty and the homonomy that would result with Liocephalus formosus Boulenger (1880). In studying Tschudi's original descrip- tion (1845) and his redescription (1846) of Scelotrema formosum, I find that the characters discussed closely fit those of Liocephalus rhodogaster (Boulenger, 1901). Most characteristics given by Tschudi are comparable with ones observed in a recently collected series of lizards from central Peru (Table 6). The illustration of Scelotrema formosum in Tschudi (1846, pi. 1, fig. 1), the description and the measurements indicate that Tschudi was in possession of a subadult or adult male, although the measurements do not ag^ee with those of Roux (1907) and cannot be accepted with confidence. The illustration of the type, although of poor quality, is recognizable as a Stenocercus or Ophryoessoides on the basis of the colora- tion. Scelotrema formosum as described by Tschudi differs only in minor points from Liocephalus rhodogaster. Scelotrema for- mosum is said to lack pterygoid teeth; L. rhodogaster has 1 or 2 minute pterygoid teeth per side which are sometimes covered by the mucosa of the mouth. If Tschudi merely looked at the pterygoid region without probing in the mucosa, he could have overlooked the inconspicuous pterygoid teeth of this species. The white spots on the body of S. formosum are greenish white in Table 6. Comparison of Steironotus arenarius Tschudi, Scelotrema formosum Tschudi and Liocephalus rhodo- gaster Boulenger. Data from Tschudi (1846) and Boulenger (1901). The description of the last species is supplemented by a recent series of lizards (KU 134109-1341 13). Characters Steironotus arenarium Scelotrema formosum Liocephalus rhodogaster Interparietal large small small Gularfold present absent absent Parietals "large" ? small Lateral venter green with black spots violet violet Chin of males light pink pink Shoulder 5 transverse drak blotches light stripe light stripe Postfemoral pocket weak deep deep Scales of ventral pectoral region in relation to ventrals larger equal equal Suborbital stripe ? prominent prominent Pterygoid teeth 4-6 0 1-2 ANDEAN IGUANID LIZARDS 53 life in L. rhodogaster. Tschudi referred to the cinammon-brown dorsum with red on the posterior margin of each scale; speci- mens of L. rhodogaster are various shades of brown with some males having a reddish suffusion which is most conspicuous on the lateral body. The dark dorsal blotches of L. rhodogaster are often outlined with white or yellow whereas those of the type of S. formosum are said to be outlined with yellow. The lateral body of males of L. rhod- ogaster is brown, slightly darker than the dorsal head whereas it is dull black in females. Tschudi's reference to a velvet black lateral head may be a result of discoloration of the specimen in alcohol, or the use of both sexes in compiling the description. Tschudi's statements referring to a violet ventral coloration and a pink chin indicate that he had a male. Roux (1907) indicated the presence of three specimens in a jar labeled Steironotus arenarius in the Neuchatel museum. Roux concluded errone- ously that the two specimens similar to L. rhodogaster represented the types of Steironotus arenarium and that the third specimen was an undescribed Tropidurus; in fact the two specimens similar to L. rhodogaster were the types of Scelotrema formosum and the specimen of the "new" species of Tropidurus was the type of Steironotus arenarius. Roux's work despite his erroneous conclusions does indicate that more than one specimen of Scelotrema formosum was available to Tschudi and faclitates understanding of the reference to characteristics of both sexes in the descrip- tion (Tschudi, 1846). The midventral stripe of males of L. rhodogaster is absent in juveniles, gray in subadults and black in adults. Tschudi referred to a yellow-green midventral stripe. Liocephalus rhodogaster has conspic- uous dark transverse markings on the tail; S. formosum was described as having dark transverse markings with yellow centers. Except for the differences noted, the morphology described for Scelotrema for- mosum Tschudi agrees with that of Lio- cephalus rhodogaster Boulenger. Dr. Rich- ard Etheridge (pers. comm.) examined the specimens assumed by Roux to be types of Steironotus arenarius but here shown to be the types of Scelotrema formosum; Ether- idge found the cotypes to ag^ee with the type of Liocephalus rhodogaster which he had examined previously in the British Museum (Natural History). On the basis of the original descriptions and subsequent discussions of the two taxa, I consider Liocephalus rhodogaster Boulen- ger to be a junior synonym of Scelotrema formosum. Scelotrema formosum is assigned to the genus Stenocercus on the basis of the morphology which compares well with other species and the present diagnosis of the genus. The type locality of Stenocercus for- mosus given by Tschudi (1847) of "Montanas des mittlern Peru: am Tullumayo" probably refers to the Rio Tulumayo, Departamento Junin, a tributary of the Rio Perene. Most of the Rio Tulumayo is within the range of specimens previously assigned to L. rhodo- gaster. It is a moot point to argue further whether the specimens examined by Roux and Etheridge and found to be idenical with Liocephalus rhodogaster were the types of Scelotrema formosum. The recognition of Scelotrema for- mosum Tschudi as a Stenocercus creates secondary homonymy with Liocephalus formosus Boulenger (1880), which is also a member of the genus Stenocercus. Replace- ment of the name Liocephalus formosus Boulenger (1880) with an available junior synonym is necessary. See the species account of Stenocercus haenschi Werner (1901). Among the characteristics listed by Andersson (1908) for Stenocercus seydi are several which indicate that a specimen of Stenocercus formosus was being described. Andersson described a repeating sequence of 4 caudal whorls (per autonomic segment); S. formosus is the only species in the genus with more than 3 caudal whorls per autonomic segment. The aspects of scutella- tion and coloration mentioned by Andersson agree with those known in S. formosus and prompt synonymy of the two taxa. 54 THOMAS H. FRITTS Stenocercus guentheri (Boulenger) n. comb. Fig. 19-21 Liocephalus guentAen Boulenger, 1885:169. Type Material.- Syntypes BMNH 58.7.25.16, 58.7. 25.18, 60.6.16.18-21, 80.12.8.53, 59.9.20.b, 71.4.16. •53(9), 71.2.7.7-10 from Guayaquil (Provincia Guayas), Ecuador; Sarayacu (Provincia Past' ia), Ecuador; and an unspecified locality in Colombia. Stenocercus guen- theri occurs at none of the localities given for the type material. The presence of a black gular patch on at least one of the male syntypes indicates that it was collected in the Pichincha Basin of Ecuador; the type locality is herein restricted to San Antonio de Pichin- cha, 2500m, Provincia Pichincha, Ecuador. Diagnosis.— 1) Posthumeral pocket absent or only weakly developed; 2) post- femoral pocket present, variable in size; 3) lateral scales not greatly reduced in size in relation to dorsal scales; 4) scales posterior to tympanum imbricate, keeled, nearly equal in size to dorsal neck scales; 5) scales on posterior thigh imbricate, keeled; 6) vertebral row continuous, form- ing raised dorsal crest; 7) caudal scales normal, three caudal whorls per autonomic segment; 8) antehumeral fold absent; 9) neck fold absent; 10) scales around middle of body 58-89; 11) scales across gular region between tympana 25-38; 13) venter of males variable in coloration; 14) chin of males not black. Stenocercus guentheri differs from all species lacking antehumeral and neck folds except S. apurimacus in usually having more than 60 scales around the midbody, and more than 24 scales across the gular region between tympana. It differs from S. apurimacus in lacking a deep posthumeral pocket, large scales which are wider than long in the occipital region, and in having a postfemoral pocket. Distribution. — Stenocercus guentheri is the most widely distributed species of the genus in Ecuador, where it extends from Imbabura Province southward to Azuay Province (Fig. 9). It occurs altitudinally from 2100 m to 3890 m. Most localities are from the slopes of the interandean valleys between the two principal Andean Cor- dilleras; both Pacific and Atlantic drainages are inhabited. Ecological Observations. — Stenocercus guentheri is not known in sympatry with other Stenocercus. Males tend to bask and search for food on rocks and other elevated positions, whereas females are most com- monly noted feeding on the ground. At localities where exposed rocks and rock walls are uncommon, individuals of both sexes seek refuge in holes in the ground at the bases of plants (Agave, Stipa, and others). One female (KU 134474) was observed covering a small hole containing two eggs near Guamote, Chimborazo Province, on 11 June 1970. Eighteen females collected near Mulalo, Provincia Cotopaxi deposited two eggs each in the laboratory between 16 May and 18 June 1972. Remarks. — Stenocercus guentheri re- mains taxonomically bewildering in exhibit- ing several divergent morphological types including considerable geographic variation. A detailed study of variation within the species is underway. However, present knowledge precludes all but a superficial discussion of variation. Males from the Imbabura and Pichincha basins have a narrow yellow midventral stripe with lateral areas of venter gray-blue to greenish gray; males from the Central Valley (Cotopaxi and Chimborazo basins) have yellow venters with numerous dark spots. The population from Alausi, Provincia Chimborazo, south of the southern margin of the central valley includes males with red-orange venters, others with light yellow venters and additional individuals with blue-green ventral fields separated by a black midventral stripe. Males from the Pichincha Basin have black patches on the gular region similar to those of Stenocercus festae; males from other valleys lack such black markings. The number of scales around the middle of the body is higher in some populations in the Central Valley than in those in the Imbabura and Pichincha Basins. This possibly reflects a greater range of variation within the Central Valley rather than a distinct geographic trend. Three local populations from the Imbabura Basin have means of 64.0 to 68.1; three populations ANDEAN IGUANID LIZARDS 55 from the Pichincha Basin have means of 64.2 to 68.0. Nine populations from the Central Valley have means of 66.1 to 77.2. Means for the two populations south of the Central Valley, Alausi', Provincia Chimborazo, and Sevilla de Oro, Provincia Azuay, are 67.2 and 62.1 respectively. A similar pattern of greater variability of populational means in the Central Valley can be seen in the number of scales from occiput to rostral. The ranges of the means for the Imbabura Basin, Pichincha Basin and Central Valley are 15.0 to 16.8, 15.1 to 15.7 and 16.0 to 18.4 respectively. The samples from Alausi' and Sevilla de Oro have means of 17.9 and 17.1 respectively. The extent of sexual dimorphism in size and color pattern remains to be fully investigated. Field observations suggest that increased struc- tural diversity in habitat (i.e. large patches of rock similar to the lava fields at Lago Cuicocha, Provincia Imbabura, and rock- shrub habitat near Alausi', Provincia Chim- borazo) is accompanied by greater sexual dimorphism and increased variability. The absence of patchy resources or low density and small size of such patches can limit sympatric occurrence of ecologically similar species (Levins, 1968). The absence of structural diversity in habitat [i.e. the scarcity of exposed rocks in the northern Ecuadorian Andes) is a possible explanation of the allopatric distribution of S. guentheri. The distribution of Pleistocene volcanism in Ecuador, as mapped by Saur (1957), closely approximates that of Stenocercus guentheri. The rapid weathering of certain volcanic products possibly explains the absence of extensive areas of exposed rock and the absence of a completely saxicolous lizard in sympatry with S. guentheri. Where hetero- geneous habitat patches do exist locally, morphological adaptation by one or both sexes within the limits of genetic plasticity available is likely. Such an hypothesis would explain the extreme inter- and intrapopula- tional variation known in S. guentheri. Stenocercus haenschi (Werner) n. comb. Liocephalus formosus Boulenger, 1880:43. A junior sec- ondary homonym of Scelotrema formosum Tschudi (1845) = Stenocercus formosus (Tschudi). Type Material.- BIN 2007 from Andes del' Equateur; the type locality is herein restricted to Balzapampa, Provincia Bolivar, Ecuador. Liocephalus haenschi Werner, 1901:595. Type Material.— ZMB 16595 from Balzapampa, 750m, (Provincia Bolivar), Ecuador. Diagnosis.— 1) Posthumeral pocket absent; 2) postfemoral pocket absent; 3) dorsal scales imbricate, keeled; 4) scales posterior to tympanum imbricate and keeled, one-half size of dorsal neck scales; 5) scales on posterior thigh granular; 6) ver- tebral row continuous, not modified into dorsal crest; 7) caudal scales normal, three caudal whorls per autonomic segment; 8) antehumeral fold present; 9) neck fold present; 10) scales around body 57-62; 11) scales across gular region small, approx- imately 40; 15) head scales smooth. The blue-green coloration noted by Boulenger and Werner in the two known specimens of S. haenschi is approximated in only two other species— S. varius and S. humeralis. Both differ from S. haenschi in having more scales around the body, more cephalic scales between rostral and occiput, and a distinct vertebral crest. Distribution. — The species is known from two specimens which are the types of Leiocephalus haenschi Werner and Leio- cephalus formosus Boulenger, respectively. The former was reported from Balzapampa [Provincia Bolivar], Ecuador; the latter from the Andes of Ecuador (Fig. 9). The town of Balzapampa at 800 m elevation lies at the base of the major range of the Cor- dillera Occidental. The locality "Balza- pampa" possibly includes a great range of elevations, above 800 m, within a close geographic proximity to the town. Ecological Observations. — Little is known concerning this species. The green dorsal coloration and high number of sub- digital lamellae of the fourth finger (26-28) indicate that Stenocercus haenschi is likely an arboreal species. Both S. humeralis and 56 THOMAS H. FRITTS S. varius have green dorsal coloration and are arboreal. All other species of the genus having a high number of subdigital lamellae of the fourth toe occupy arboreal or rock habitats. Remarks.— The status of the two names, presently considered synonyms of Stenocercus haenschi, remains doubtful since their publication over 70 years ago. Examination of the type specimens of both species has revealed close similarity in morphology necessitating the present syn- onymy. Both specimens are clearly members of the genus Stenocercus. 2700 m in altitude. Ecological Observations. — Near the city of Loja at 2150 m Stenocercus humeralis is abundant only in specific habitats. Individuals occur in small shrubs and on the trunks of eucalyptus trees in mesic quebradas as well as in agave plants on top of earth fences. Stenocercus ornatus occurs in sympatry in the upper Rio Zamora Valley; the latter species is confined to more open habitats at bases of shrubs and in low agave rows. Stenocercus festae possibly occurs in sympatry with S. humeralis on the Pacific slopes. Stenocercus humeralis (Giinther) Microphractus humeralis Giinther, 1859a:90. Type Material.- BMNH (RR 1946.8.11.76-77) from Andes of Ecuador; the type locality is herein re- stricted to Loja, 2150m, Provincia Loja, Ecuador. Diagnosis. — 1) Posthumeral pocket absent; 2) postfemoral pocket present; 3) dorsal scales near vertebral row imbri- cate, keeled; dorsals more laterally posi- tioned granular, small, dorsals grading into small, granular lateral scales; 4) scales pos- terior to tympanum granular, less than one-half the size of dorsal neck scales; 5) scales on posterior thigh granular; 6) ver- tebral row continuous, forming low dorsal crest; 7) caudal scales normal, three caudal whorls per autonomic segment; 8) antehum- eral fold present; 9) neck fold present; 10) scales around body more than 99; 11) scales across gular region between tympana more than 36; 13) dorsal colora- tion green. Stenocercus humeralis differs from the other species with green coloration, S. haenschi and S. varius, in having more scales around the body and more subdigital lamellae on the fourth finger. Distribution. — Stenocercus humeralis is known from southern regions of Ecuador in Loja Province in the upper Rio Catamayo (Pacific drainage) and Rib Zamora (Atlantic drainage) at elevations of 2000-3000 m (Fig. 9). In its lowest areas the continental divide in Loja Province is approximately Stenocercus ivitus Fritts Stenocercus ivitus Fritts, 1972:10. Type Material.— Holotype KU 134654 from summit of Cordillera Huancabamba between Canchaque and Huancabamba, 3100m, Departamento Piura, Peru. Diagnosis.— 1) Posthumeral pocket absent; 2) postfemoral pocket present; 4) scales posterior to tympanum imbricate, keeled, nearly equal in size to dorsal neck scales; 5) scales of posterior thigh granular; 6) vertebral row discontinuous, vertebral scales not forming dorsal crest; 7) caudal scales normal, three caudal whorls per autonomic segment; 8) antehumeral fold absent; 9) neck fold absent; 10) scales around middle of body 44. Stenocercus ivitus differs from all other species lacking antehumeral folds and neck folds in lacking a continuous vertebral row which forms a dorsal crest and in having granular scales on the posterior surface of the thigh. Distribution. — Stenocercus ivitus is known only from the type locality at the summit of the Cordillera Huancabamba (Fig. 11). This Cordillera forms the continen- tal divide between the Rio Piura (Pacific drainage) and the Rio Huancabamba (Atlan- tic drainage). Ecological Observations. — The adult female was observed basking on a pile of dead branches at the margin of a low shrub forest, and in sympatry with S. nubicola. ANDEAN IGUANID LIZARDS 57 Stenocercus marmoratus (D'Orbigny) Trachycyclus marmoratus D'Orbigny, 1837, in, Dumeril andBibron, 1837:356. Type Material.— MHNP 2513 from Rio Grande, Brazil; in error. Stenocercus difficilis Werner, 1910:23. Type Material.— Holotype ZMH (number unknown) from Cochabamba (Departamento Cochabamba), Bolivia. Diagnosis.— 3) Dorsals imbricate, keeled; 4) scales posterior to tympanum granular, less than one-half size of dorsal neck scales; 5) scales on posterior thigh imbricate, keeled or smooth; 6) vertebral row rarely complete, vertebral scales not distinctly modified from other dorsal scales; 7) caudal scales large, with prominent keels, two caudal whorls per autonomic segment; 8) antehumeral fold present; 9) neck fold present; 10) scales around middle of body 43-54. This is the only spiny-tailed Stenocer- cus with a discontinuous vertebral row and imbricate scales on posterior thighs. Distribution. — Stenocercus marmora- tus is known from Andean slopes between 2700 and 3350 m in the Departamento Cochabamba, Bolivia (Fig. 13). Stenocercus mxirmoratus occupies the highest elevations of any of the southern species of Stenocercus. Ecological Observations. — A species of Liolaemus similar to L. alticolor is known from similar elevations within the range of Stenocercus marmoratus; the two species possibly occur in sympatry. The Liolaemus is small and exhibits a morphology similar to ground-dwelling species of Stenocercus and other Liolaemus, whereas Stenocercus marmoratus possesses a spinose tail and lacks a high dorsal crest similar to rock- dwelling species of Stenocercus. The Liol- aemus occurs on the ground and at the bases of clump grasses. No field observa- tions are available for Stenocercus marmoratus. Remarks.— Dumeril and Bibron (1837) cite D'Orbigny "Voy Amer. mer. Zoolog. Rept. tab. 4, fig. 2. (non encore publie.)" as the author of the name. D'Orbigny (1847) later published his work. Dumeril and Bibron (1837) were cited as the authors of the name Trachycyclus marmoratus by Burt and Burt (1933) and by Etheridge (1970). The interpretations of authorship described in the International Code of Zoological Nomenclature (1964: art. 50-51) clearly proide for authorship by an author in the work of another author or other authors if such responsibility "is clear from the contents of the publication. . ." The citation of plates in D'Orbigny's then unpubUshed work and the failure to place "Nobis" behind the name both imply that Dumeril and Bibron considered D'Orbigny responsible for the proposal of the name Trachycyclus marmoratus. Stenocercus melanopygus Boulenger Fig. 26 Stenocercus melanopygus Boulenger, 1900:182. Type Material.- Syntypes BMNH 1900.3.30.6-9 (RR 1946.8.11.85-88), 1900.3.30.10-13 (RR 1946.8.11.78-81) from Banos, 9000 ft. (Departamento) Cajamarca, Peru. Diagnosis.— 4) Scales posterior to tympanum imbricate, keeled, nearly equal in size to those of dorsal neck; 5) scales on posterior thigh granular; 6) vertebral row discontinuous, no dorsal crest; 7) caudal scales normal, three caudal whorls per autonomic segment; 8) antehumeral fold present; 9) neck fold absent; 10) scales around middle of body 44-58. Stenocercus melanopygus and S. nigro- maculatus, are the only species in the genus which have antehumeral folds, but lack neck folds and posthumeral pockets. Stenocercus melanopygus differs from S. nigromaculatus in lacking a continuous vertebral row and high dorsal crest. Some individuals of Stenocercus moestus seemingly lack neck folds and have only small patches of granular or small imbricate scales in the approximate position of the neck folds. Stenocercus moestus differs from S. melan- opygus in having less than 40 scales around the middle of the body. Distribution.— The species is known at elevations of 2700-3250 m in the interandean valleys in the Peruvian departments of 58 THOMAS H. FRITTS Cajamarca and La Libertad (Fig. 11). All localities except Celendin are within the drainage of the Rio Crisnejas. Celendin lies in an adjacent valley separated from the Rio Crisnejas drainage by a ridge approximately 3300 m in elevation. The species is likely widespread in the western tributaries of the Rio Maranon. The species is not known from Pacific drainages, although the distribution occu- pies the most western parts of the Amazonian drainage in north-central Peru. The western slopes of the Rio Maranon Valley are broken only by minor ridges into east-west rivers feeding into the Rio Maraiion. The major physiographic features which correlate with the limits of the distribution of this species are the low elevations and dry conditions prevalent in the area of the Abra Porculla in the north, the Sierra Occidental on the west, and the low elevation of the Rio Marafion Valley on the east. Ecological Observations. — Stenocercus melanopygus occurs in sympatry with Stenocercus empetrus. The former species utilizes the ground, blades of agave plants, solitary rocks, and the extreme edges of large rock piles for basking and feeding activity, whereas the latter species inhabits large rock piles, rock cliffs and rock exposures. Localities where S. melanopygus occurs are moist and support moderate vegetation for this region. On a local basis, specimens are more common on dry hillsides where rocks are scattered, and vegetation is sparse; in moist areas on the valley floor or on gentle slopes, both species are found together in proportions seemingly corre- lated with their respective habitat preferences. Near Cajamarca, 2800 m, Departa- mento Cajamarca, Peru, S. empetrus is rare on extremely dry hillsides. Individual males of S. melanopygus occupy large rock piles and rock exposures which are usually utilized by S. empetrus in areas of sympatry. The males of S. melanopygus in this area are larger :SVL 54-76 mm) than males from near Huamachuco (SVL 50-64 mm). The males from the vicinity of Cajamarca have a color pattern differing from those from other localities and resembling the sympat- ric congener S. empetrus. The increase in size of males and their possession of a color pattern similar to rock-dwelling species probably facilitates exploitation of rock habitats in the local area where S. empetrus is rare. Possibly, populations considered to be Stenocercus melanopygus from other localities are specifically distinct from the lizards from Cajamarca; further knowledge of geographic variation, ecology and behav- ior is needed for understanding the varia- tion within this species. Remarks.— The most conspicuous vari- ation in Stenocercus melanopygus is in size and coloration of males. Specimens from Cajamarca show more marked sexual dimorphism in size than those from other populations. The males from Cajamarca differ from those of other areas in having a black ventral pelvic region and ventral thighs instead of the yellow coloration present in other populations. Stenocercus moestus Boulenger Stenocercus moestus Boulenger, 1885:136. Type Material.— BMNH 75.2.13.3 from Lima (Departa- mento Lima), Peru. Diagnosis.— 1) Posthumeral pocket absent; 2) postfemoral pocket weak; 4) scales posterior to tympanum imbricate, usually keeled, nearly equal in size to dorsal neck scales; 5) scales on posterior thigh granular; 6) vertebral row discontinuous, no dorsal crest; 7) caudal scales normal, three caudal whorls per autonomic segment; 8) antehumeral fold present; 9) neck fold weak or absent; 10) scales around middle of body 33-38. Stenocercus moestus resembles S. mel- anopygus and S. ivitus in having extremely large scales on the lateral surface of the neck, but differs in usually having a weak neck fold. Stenocercus moestus also differs from S. melanopygus in having less than 40 scales around the middle of the body and from S. ivitus in having an antehumeral fold. ANDEAN IGUANID LIZARDS 59 Distribution. — The species is Itnown only from the valley of the Rio Rimac near Lima, Departamento Lima, Peru (Fig. 11), from sea level to 762 m (Etheridge, 1970). The Rio Rimac is one of the major rivers of the central Peruvian coast. The Abra Teclio, 4815 m in the upper Rio Rimac is one of the lowest passes to Atlantic drainages at this latitude. Ecological Observations. — I have not observed this species in the field. The valley of the Rio Rimac is xeric with little natural vegetation. The extensive irrigation of the valley for agricultural practices has left only the rock and sand slopes and cliffs unmolested. Species of Tropidurus occur throughout the range of Stenocercus moes- tus. Stenocercus omatissimus is known to occur at higher elevations, 3340 m of the Rfo Rimac, and possibly replaces Stenocercus moestus at intermediate elevations. The presence of a Stenocercus at low elevations of coastal Peru only in the valley of the Rio Rimac is puzzling. Whereas other Stenocer- cus are known from Pacific drainages of Peru, all occupy elevations of 1500 m or higher. Stenocercus nigromaculatus Noble Fig. 18 Stenocercus nigromaculatus Noble, 1924:112. Type Material.— Holotype MCZ 17975 and paratypes MCZ 18766-18769 from Huancabamba, Province of (Departamento) Piura, Peru. Diagnosis.— 1) Weak depression in axillary region, posthumeral pocket absent; 2) postfemoral pocket large; 4) scales posterior to tympanum imbricate, keeled, nearly equal in size to dorsal neck scales; 5) scales on posterior thigh imbricate, keeled; 6) vertebral row continuous, form- ing high dorsal crest; 7) caudal scales normal, three caudal whorls per autonomic segment; 8) antehumeral fold distinct; 9) neck fold absent; 10) scales around middle of body 48-56; 11) scales across gular region between tympana 21-25; 13) venter of males with midventral black stripe bifurcating anteriorly at pectoral girdle, bifurcating posteriorly onto ventral surfaces of thighs, lateral venter light bronze; 14) chin and gular region white with faint gray reticulation. Stenocercus nigromaculatus most close- ly resembles S. ornatus and S. rhodomelas , from which it differs in having a strong antehumeral fold, and in lacking bright pink or red on ventral surfaces of males. Distribution. — The species is known from the vicinities of Huancabamba and Chumaya, Departamento Piura, Peru (Fig. 12). I have observed individuals from 1900 to 2300 m in the valley of the Ri'o Huancabamba. The specimens that Burt and Burt (1931) reported from Wavepongo [= Llavepongo], Ecuador are identifiable as Stenocercus guentheri; Stenocercus nigro- maculatus is not known from Ecuador. The Ri'o Huancabamba, a northern tributary of the Rfo Marandn, is separated from coastal regions by the Cordillera Huancabamba with passes as low as 2790 m and 3100 m. The Ri'o Maranon Valley at this latitude has an elevation of 500 m and possibly limits the species on the east. Ecological Observations.— Individuals of Stenocercus nigromaculatus bask and forage in rows of Agave and around bases of shrubs. Juveniles enter holes at the bases of small shrubs to avoid capture. The slopes of the valley have small shrubs, and extensive grass and shrub cover; few rocks are exposed. Stenocercus boettgeri was re- ported from Huancabamba by Boulenger (1911); additional specimens were collected near Huancabamba by Noble. I did not observe S. boettgeri in this region. Steno- cercus nubicola and S. ivitus are known to occur at 3100 m in the Cordillera Huanca- bamba. Tropidurus stolzmanni and an unnamed species of Ophryoessoides are known from lower elevations of the Rio Huancabamba. Stenocercus nubicola Fritts Fig. 16 Stenocercus nubicola Fritts, 1972:11. Type Material.— Holotype KU 134107 and paratype KU 134108 from summit of Cordillera Huancabamba be- tween Canchaque and Huancabamba, 3100m, Departa- mento Piura, Peru. 60 THOMAS H. FRITTS Diagnosis. — 1) Posthumeral pocket absent; 2) postfemoral pocket present; 4) scales posterior to tympanum imbricate and smooth or granular, less than one-half size of dorsal neck scales; 5) scales on posterior thigh granular; 6) vertebral row complete, forming dorsal crest; 7) caudal scales normal, three caudal whorls per autonomic segment; 8) antehumeral fold dis- tinct; 9) neck fold present; 10) scales around middle of body 93-94; 11) scales across gular region between tympana 38-39; 13) black midventral stripe in males, bifurcate ante- riorly, extending onto ventral thighs posteriorly. Of the species of Stenocercus having black pigment extending from the pelvic to pectoral girdle on the venter of males, only S. nubicola has more than 85 scales around the middle of the body. Stenocercus nubicola closely resembles the geographically adjacent species, S. nigromaculatus in coloration, but differs from it in having a neck fold, granular scales on posterior thigh and more scales around the body. Distribution. — Stenocercus nubicola is known only from the type locality (Fig. 13). The Cordillera Huancabamba forms the continental divide in extreme northern Peru and separates the Rio Huancabamba Valley (Atlantic drainage) and the Rio Piura Valley (Pacific drainage). Ecological Observations. — Stenocercus nubicola occurs in sympatry with Stenocer- cus ivitus, in montane moist forest. Steno- cercus nigromaculatus is known at lower elevations in the same region. Stenocercus ochoai Fritts Fig. 32 Stenocercus ochoai Fritts, 1972:13. Type Material.— Holotype KU 133888 and paratypes KU 133874-133887, 133889, 139263 from Chilca. 10 km NW Ollantaytambo, 2700m, Departamento Cuzco, Peru. Diagnosis. — 1) Posthumeral pocket absent; 2) postfemoral pocket deep; 3) dor- sal scales imbricate, keeled, but slightly larger than lateral scales; 4) scales posterior to tympanum imbricate, usually keeled, one-half size of dorsal neck scales; 5) scales on posterior thigh imbricate, keeled; 6) ver- tebral row continuous, forming raised dorsal crest; 7) caudal scales normal, three caudal whorls per autonomic segment; 8) antehum- eral fold present; 9) neck fold present; 10) scales around middle of body 57-67; 11) scales across gular region between tympana 19-26; 13) venter of adult males black. Stenocercus ochoai is most similar to Stenocercus formosus (Tschudi) from which it differs in having fewer scales around the middle of the body, the entire ventral coloration of males black and three caudal whorls per autonomic segment. Distribution. — Stenocercus ochoai (Fig. 13) is known between 2000 and 3000 m in the valleys of the Rio Urubamba (Burt and Burt, 1930) and Rio Apurimac (Mertens, 1952). Ecological Observations.— Stenocercus ochoai occurs in sympatry with Stenocercus crassicaudatus between 2000 m and 2600 m in the Ri'o Urubamba Valley, and with Stenocercus apurimacus at Curahuasi, 2700 m, Departamento Apurimac, in the Rio Apurimac Valley. The species uses both ground and rock substrates. Stenocercus orientalis Fritts Fig. 15 Stenocercus orientalis Fritts, 1972:14. Type Material.— Holotype KU 134466 and paratypes 134447-134465, 134467-134473 from Chachapoyas, 2340m, Departamento Amazonas, Peru. Diagnosis.— 1) Posthumeral pocket absent; 2) postfemoral pocket present; 4) scales posterior to tympanum imbricate, keeled, one-half size of dorsal neck scales; 5) scales on posterior thigh granular; 6) ver- tebral row continuous, not forming dorsal crest; 7) caudal scales normal, three caudal whorls per autonomic segment; 8) antehum- eral fold present; 9) neck fold present; 10) scales around the middle of the body 47-58; 15) head scales prominently keeled. Stenocercus orientalis, S. omatissimus, S. moestus, S. haenschi and S. chrysopygus differ from other species of Stenocercus in ANDEAN IGUANID LIZARDS 61 lacking a dorsal crest and spinose caudal scutellation, and in having antehumeral and neck folds. Stenocercus orientalis is the only one of these having prominently keeled scales on the snout and interorbital region. It also differs from S. omatissimus in having a well-developed postfemoral pocket, from S. moestus in having more than 46 scales around the middle of the body, from S. haenschi in lacking a green dorsal coloration and in having fewer than 30 scales across the gular region and from S. chrysopygus in having imbricate scales posterior to the tympanum and in having keels of dorsals in longitudinal rows. Distribution.— The species is known in Peru only from the type locality in the valley of the Ri'o Utcubamba (Fig. 11). Ecological Observations. — In the vicin- ity of Chachapoyas patches of open chapar- ral are intermixed with dry subtropical forest. The species is observed on the ground at the bases of shrubs in open areas. No other iguanid species are known in the area. Stenocercus omatissimus (Girard) Fig. 25 Saccodeira OTnatissima Girard, 1857:198. Type Material.— Cotype USNM 5655 from Yanga, Peru and cotype (location unknown) from Obrajillo, Peru. Diagnosis.— 1) Posthumeral pocket absent; 2) postfemoral pocket absent, a weak fold behind base of hindlimb; 4) scales posterior to tympanum imbricate, keeled, usually one-half size of dorsal neck scales; 5) scales on posterior thigh granular; 6) ver- tebral row usually discontinuous or diverg- ing on body, not forming dorsal crest; 7) caudal scales normal, three caudal whorls per autonomic segment; 8) antehumeral fold present; 9) neck fold present; 10) scales around the middle of the body 43-57; 15) head scales smooth. Of the species with a discontinuous vertebral row and lacking a dorsal crest, only S. omatissimus and S. moestus have imbricate scales on the lateral neck, an antehumeral fold and neck folds. Stenocer- cus omatissimus differs from S. moestus in having more than 42 scales around the middle of the body and larger neck folds. Stenocercus omatissimus resembles S. chry- sopygus in aspects of general scutellation and body form, but differs from it in lacking a distinct postfemoral pocket, and in having imbricate scales on the lateral surface of the neck. Stenocercus orientalis differs in having a continuous vertebral row, in having keeled head scales and in having a prominent postfemoral pocket. Distribution. — Stenocercus omatissi- mus is known from few specimens, and its range remains to be defined accurately. The two localities from which the type material was obtained could not be located on any of the maps available to me. Girard referred to the "lower cordillera," probably indicating the coastal slopes of the Andes. Specimens from near Otusco, 2700-3000 m, Departa- mento La Libertad, Peru, closely approxi- mate the original description and the more complete redescription given by Girard (1857; 1858). A single subadult female from Cacray, near Infernillo, 3340 m, Departa- mento Lima, Peru, is tentatively assigned to S. omatissimus. The species is probably limited to intermediate elevations (2000 to 3400 m) of Pacific slopes in the departments of Lima, Ancash and La Libertad, Peru (Fig. 11). The coastal ranges of north- central Peru are quite irregular in height and continuity. A continuous more central range exceeding 4000 m altitude extends from Departamento Cajamarca, Peru, south to Chile (Schmidt and Walker, 1943). Ecological Observations. — Males bask on rocks and Agave, whereas females fre- quently are at bases of shrubs and on the ground between rock patches. Stenocercus empetrus is known to occur near Otuzco, Departamento La Libertad, and is probably sympatric with S. omatissimus. The one specimen of S. omatissimus from 3340 m in Departamento Lima was collected with a species of Liolaemus by Dr. J. Blancas (pers. comm.). Stenocercus moestus is known to occur at lower elevations in the Rimac Valley as do several species of Tropidurus. Tropidurus occipitalis occurs in the more xeric regions west of Otuzco at 62 THOMAS H. FRITTS elevations from sea level to 1500 m. Stenocercus omatus (Gray) n. comb. Fig. 17 Liocephalus omatus Gray, 1845:219. Type Material.- BMNH xxii.lOle (RR 1946.8.29.72) from Guayaquil (Provincia Guayas), Ecuador, in error. The type locality is herein restricted to Loja, 2150m, Provincia Loja, Ecuador. Diagnosis. — 1) Posthumeral pocket deep; 2) postfemoral pocket present; 3) lateral scales not greatly reduced in size in relation to dorsal scales; 4) scales posterior to tympanum imbricate, keeled, equal in size to dorsal neck scales; 6) vertebral row continuous, forming moderately high dorsal crest; 7) caudals normal, three caudal whorls per autonomic segment; 8) antehum- eral fold weakly formed; 9) neck fold absent; 10) scales around middle of body 47-58; 13) venter of males bronze-red with black midventral stripe, chin light red or pink with dark suffusion in some specimens; 14) gular region of males light red or pink. Stenocercus omatus differs from other species with a black midventral stripe in males, in having a bronze-red venter. It differs from Stenocercus rhodomelas in having a weak antehumeral fold, a deeper posthumeral pocket and a higher dorsal crest. It differs from S. formosus and S. ochoai in lacking a neck fold and from S. nigromaculatus in having a deep posthum- eral pocket and a higher dorsal crest. Stenocercus omatus differs from S. festae in lacking a black gular patch in males. Distribution. — Stenocercus omatus oc- curs in Ecuador in the valley of the Rio Zamora (Atlantic drainage), Provincia Loja, at elevations of 2065-2700 m (Fig. 14). Stenocercus omatus is known from the crest of the range dividing the Rio Zamora drainage from the Rio Catamayo drainage. The complex biogeography of southern Ecuador and the relatively poor knowledge of the ecology of the area prevents more precise definition of the distribution of this species. Ecological Observations. — Stenocercus ornatus and S. humeralis occur in sympatry in the Loja Valley above 2100 m. The former species occupies open areas on the ground and rows of Agave, whereas the latter species is confined to tree trunks, shrubs and specimens of Agave on top of earth fences in more mesic areas. The Loja Basin is largely mesic, and probably supported patches of montane forest prior to clearing by man. Remarks.— The specimens from the Loja basin agree in most aspects of morphology with the type of S. omatus (Gray, 1845). The type (BMNH 1946.8.29.72) has 46 scales around the body, whereas specimens from the Loja Basin have 47-58. The type has a solid black midventral stripe, whereas most specimens from the Loja Basin have a yellow stripe dividing the broader black stripe. The chin is black on the type, whereas it is light red or pink in the other specimens. The type agrees with specimens from the Loja Basin in having a moderate black blotch anterior to insertion of forelimb, a deep posthumeral pocket, a weak dorsolateral crest on neck produced by enlarged scales and a high dorsal crest. Burt and Burt (1931:271) misquoted Boulen- ger (1885) by exchanging the number of scales around the body of S. guentheri with the number of S. omatus. Stenocercus praeomatus Fritts Fig. 30 Stenocercus praeomatus Fritts, 1972:16. Type Material.— Holotype KU 134231 and paratypes KU 134224-134230, 134232 from Comas, 3220m, De- partamento Junin, Peru. Diagnosis. — 1) Posthumeral pocket consisting of weak fold; 2) postfemoral pocket large; 4) scales posterior to tym- panum granular, less than one-half size of dorsal neck scales; 5) scales on posterior thigh granular; 6) vertebral row continuous, forming high dorsal crest; 7) caudal scales normal, three caudal whorls per autonomic segment; 8) antehumeral fold present; 9) neck fold present; 10) scales around middle of body 89-121; 11) scales across gular region between tympana 32-57; 13) ventral coloration of males rose pink; ANDEAN IGUANID LIZARDS 63 14) gular region of males black. Of the species having a black ventral collar in males, Stenocercus praeomatus is the only one having more than 88 scales around the middle of the body and having pink ventral coloration in males. Distribution. — The species is known in central Peru from the higher elevations of the Rio Tulumayo in Departamento Junin (Fig. 8) and from northern Peru, Abra Porculla, 2144 m, Departamento Piura. Presence of the species in the area between these two regions remains to be docu- mented; a similar distribution is exhibited by Stenocercus boettgeri. Ecological Observations. — Stenocercus praeomatus occurs on large rock piles and in terraces made of rocks. Populations of S. variabilis and S. boettgeri are known from adjacent areas of Departamento Junin. Remarks .— The population in northern Peru is known to differ from that in central Peru in scutellation (Fritts, 1972). Stenocercus rhodomelas (Boulenger) n. comb. Fig. 22 Liocephalus rhodomelas Boulenger, 1899:455. Type Material.— Syntypes BMNH 99.10.30.11-14 (RR 1946.8.29.77-80) from Oiia, 6500 ft. (Provincia Azuay), Ecuador. Diagnosis.— 1) Posthumeral pocket deep; 2) postfemoral pocket present; 3) lat- eral scales not greatly reduced in relation to dorsals; 4) scales posterior to tympanum imbricate, keeled, equal in size to dorsal neck scales; 5) scales on posterior thigh imbricate, keeled; 6) vertebral row continu- ous, forming high dorsal crest; 7) caudals normal, three caudal whorls per autonomic segment; 8) antehumeral fold absent; 9) neck fold absent; 10) scales around middle of body 46-56; 11) scales across gular region 18-23; 13) venter of males bright pink with mid ventral black stripe, ventral surfaces of thighs black; 14) chin black, gular region pink in males. Stenocercus rhodomelas differs from S. ornatus in having a smaller blotch anterior in insertion of forelimb, a shallower posthumeral pocket, and a pink ventral coloration in males, and in lacking any trace of an antehumeral fold. Distribution. — Stenocercus rhodomelas is known from the valley of the Rio Jubones below 2100 m in elevation on the Pacific slopes in the Ecuadorian provinces of Azuay, Loja and El Oro (Fig. 14). The species does not reach elevations of 2240 and 2500 m at Giron, Provincia Azuay, and Saraguro, Provincia Loja, respectively. Whether it occupies other smaller valleys on the Pacific slopes remains to be determined. Ecological Observations. — Ophryoes- soides iridescens and Tropidurus occipitalis possibly occur in sympatry with Stenocercus rhodomelas at the lower elevations of its distribution. Stenocercus simonsii and S. festae are known allopatrically at elevations above 2200 m and 2500 m, respectively, in the valley of the Rio Jubones. The area occupied by Stenocercus rhodomelas is xeric with sparse vegetation of cactus and Agave. Individuals were observed on large rocks and on the ground at the bases of cactus. Stenocercus roseiventris D'Orbigny Fig. 34 Stenocercus rosei-ventris D'Orbigny, 1837 in, Dumeril and Bibron, 1837:350. Type Material.— MHNP 6879 from Bolivia. Stenocercus atrigularis Werner, 1913:11. Type Material.— Holotype ZMH (number unknown) from Provinz (Departamento) Beni, Bolivia, Sudliches Quellgabiet des Amazonas; destroyed in Hamburg during World War II. Diagnosis. — 1) Postfemoral pocket present; 4) scales posterior to tympanum granular, less than one-half size of dorsal neck scales; 5) scales on posterior thigh large, imbricate, keeled; 6) vertebral row continuous, forming distinct dorsal crest; 7) caudals large, extremely spinose, two caudal whorls per autonomic segment; 8) antehumeral fold present; 9) neck fold present; 10) scales around middle of body 54-76; 13) ventral coloration of males rose-pink. Stenocercus roseiventris and S. mar- moratus are the only species of the genus having spinose caudal scutellation and 64 THOMAS H. FRITTS imbricate scales on the posterior thigh. Stenocercus roseiventris differs from S. marmoratus in having a continuous verte- bral row which forms a distinct dorsal crest, caudal scales which are more spinose, and more scales around the middle of the body. Distribution. — Stenocercus roseiven- tris occurs in lowland and peripheral Andean regions of Amazonian Peru, Bolivia, northern Argentina and western Brazil (Fig. 13). The species is known in Argentina from the provinces of Jujuy and Salta, and in Brazil from Acre. It is known as far north as the lower Rio Maranon and Rio Ucayali in north-central Amazonian Peru. All speci- mens are from the southern Amazon Basin or adjacent regions of the Rio Bermejo and Rio Salado. Ecological Relationships.— The single specimen of this species observed in the field was in a coffee grove near some large rocks. On the basis of the localities from which the species is known, it is probably confined to forests of lower elevations in the peripheral Amazon Basin. Since the tail is extremely spiny, similar to the tails of other Stenocercus confined to rock habitats, S. roseiventris is expected to be found in rocky habitats near rivers and in broken areas of the forest canopy. Remarks.— Dumeril and Bibron (1837) cite D'Orbigny as the author of Stenocercus roseiventris. Since D'Orbigny 's work re- mained unpublished at that time, the date of publication is that of Dumeril and Bibron's work (1837). My consideration of D'Orbigny as the author of the name is based on the indication by Dumeril and Bibron; Etheridge (1970) considered Dumeril and Bibron to be the proper authors. Refer to the discussion given under S. marmoratus. Stenocercus simonsii Boulenger Fig. 27 Stenocercus simonsii Boulenger, 1899:454. Type Material.— Syntypes BMNH 99.10.30.5-6 (RR 1946.8.11.73-74) from Oiia, 6500 ft. (Provincia Azuay), Ecuador. Diagnosis.— 1) Posthumeral pocket absent; 2) postfemoral pocket moderately large; 3) dorsals imbricate, keeled; 4) scales posterior to tympanum granular; 5) scales on posterior thigh granular; 6) vertebral row continuous, forming low dorsal crest; 7) caudal scales large, strongly keeled, two caudal whorls per autonomic segment; 8) antehumeral fold present; 9) neck fold present; 10) scales around middle of body 73-102. Stenocercus simonsii differs from S. carrioni in having caudal whorls that are nearly equal in size. It differs from S. crassicaudatus in having fewer scales around midbody (less than 103), and in having all dorsal scales imbricate and larger than granular lateral scales. Distribution. — The species is known from Ecuador at elevations of 2240-2500 m in the drainage of the Rio Jubones in the provinces of Azuay and Loja (Fig. 10). The distribution possibly includes adjacent val- leys of the Pacific drainage of Ecuador. A pass at 3100 m exists between the Rio Jubones and Rio Zamora drainages. The mesic habitats inhabited by S. simonsii do not extend into the lower elevations of the Rio Jubones. Ecological Observations. — Individuals of both sexes are confined to rock walls and rock piles. Stenocercus festae occurs in sympatry with S. simonsii near Saraguro, Provincia Loja, Ecuador. Stenocercus festae basks on agave plants, on solitary rocks and on the ground near small shrubs. Remarks.— Populations from Giron, Provincia Azuay, and Saraguro, Provincia Loja, differ in dorsal coloration and the size of dorsal bloches. Lizards from the former locality have a light gray-green dorsal coloration with heavy black transverse blotches; the lizards of the latter population have a light gray dorsal coloration with smaller, less distinct black blotches. The sample from Saraguro differs from that from Giron in having more scales around the body 86-102 (96.7 ± 2.07) vs. 73-99 (87.5 ± 3.34), p < 0.05, more subdigital lamellae on fourth toe 30-34 (33.0 ±0.58) vs. 27-32 (29.3 ± 0.59), p < 0.001, and more scales across the gular region 49-56 )51.1 ± 0.91) ANDEAN IGUANID LIZARDS 65 VS. 39-52 (45 ± 1.35), p < 0.005, when using SS-STP analysis. Stenocercus trachycephalus (Dumeril) n. comb. Holotropis trachycephalus Dumeril, 1851:70. Type Material.- Syntypes MHNP 1787 (2), 2393 (2, 'from Nile. Grenada and MHNP 2394 (2) from Bogota, (Colombia). Dumeril (1851) listed the types as being from Santa Fe de Bogota (Departamento Cundina- marca), Colombia. Diagnosis.— 1) Posthumeral pocket absent; 2) postfemoral pocket well devel- oped; 3) lateral scales not greatly reduced in size in relation to dorsals; 4) scales posterior to tympanum imbricate, keeled, nearly equal in size to dorsal neck scales; 5) scales on posterior thigh imbricate, keeled; 6) ver- tebral scale row continuous, forming raised dorsal crest; 7) caudal scales normal, three caudal whorls per autonomic segment; 8) antehumeral fold absent; 9) neck folds absent; 10) scales around middle of body 42-62; 13) venter of males black, pectoral region bright orange; 14) gular region of males black. Of the species lacking antehumeral folds, only S. trachycephalus, S. festae, and S. guentheri have a black gular patch. Stenocercus trachycephalus differs from S. guentheri in usually having fewer than 60 scales around the middle of the body, a black venter in males and weakly keeled ventrals. Stenocercus trachycephalus differs from S. festae in lacking a posthumeral pocket, in usually having weakly keeled ventrals and in having a bright orange chest. Some males of S. festae have light orange or yellow coloration on chin and ventrolateral neck. Distribution.— The species is known in Colombia from throughout the Sabana de Bogota at elevations of 1909-3200 m (Dunn, 1944) in the departments of Boyaca, Cundinamarca, and Santander. Two series of Stenocercus trachycephalus, USNM 153987-153991 and AMNH 27611, 27613 from Girardot, Departamento Cundina- marca, and La Dorada, Departamento Caldas, respectively, are considered to have erroneous locality data. Both localities are below 300 m in elevation in the Rio Magdalenas Valley. The Magdalena Valley has an extremely hot and xeric climate, which would prevent habitation by a montane lizard. Ecological Observations.— Dunn (1944) commented on the habitat of S. trachyce- phalus near Bogota, Colombia. Individuals frequent rocks and other types of cover on the ground but are rarely found in shrubs. Two eggs are laid in a single clutch (Osorno, 1938). Remarks. — Stenocercus trachycephalus is the only species of the genus occurring in Colombia. The absence of this and other species from regions of the Colombian Andes other than the Sabana de Bogota re- mains to be documented by thorough field- work. The hiatus in the range of the genus between the Sabana de Bogota and the nor- thern Ecuadorian border is inexplicable. Stenocercus variabilis Boulenger Fig. 36 Stenocercus variabilis Boulenger, 1901=546. Type Material.- Syntypes BMNH 1900.11.27.12-14 (RR 1946.8.11.89-91) from Palca, 1000 ft., Bolivia (in error); the type locality is herein restricted to Palca, 1000 ft., Departamento Junin, Peru. Stenocercus juninensis Shreve, 1941:75. new synonymy. Type Material.— Holotype MCZ 45820 and paratype 45821 from Huasqui, near Tarma, Departamento Juni'n, Peru. Diagnosis.— 1) Posthumeral pocket indicated by shallow depression in axilla; 2) postfemoral pocket present or absent; 4) scales posterior to tympanum small, granular or smooth-imbricate; 5) scales on posterior thigh granular; 6) vertebral row continuous, forming distinct raised dorsal crest; 7) caudal scales normal, three caudal whorls per autonomic segment; 8) antehum- eral fold present; 9) neck fold well- developed; 10) scales around middle of body 59-74; 15) scales on dorsal head smooth. Of the species with granular scales on the posterior thigh, a distinct dorsal crest, and neck folds, only S. variabilis has fewer 66 THOMAS H. FRITTS than 75 scales around the middle of the body. Stenocercus variabilis differs from S. ochoai and S. formosus in having granular scales on the posterior thigh; it differs from S. cupreus in lacking enlarged caudals and in having three caudal whorls per autonomic segment and a raised dorsal crest. Distribution.— This species is known in Peru in the valleys of the Rio Mantaro and Rio Tarma at elevations of 1200-3000 m (Fig. 8). The failure of subsequent workers to record additional specimens of this species from Bolivia and the discovery of morphologically similar lizards in central Peru cast doubt on the type locality as given by Boulenger (1901). The type series was collected by P. 0. Simons and is a part of a large collection of reptiles from Peru and Bolivia. Among the other species named by Boulenger from the Simons collection in the same paper as S. variabilis, are three from the vicinity of the Rio Perene, Departa- mento Junin, Peru. One of the only land routes to the Rio Perene region from the central Andes in the nineteenth century, as well as in modern times, passes from Tarma at 3100 m down the valley of the Rio Tarma to the Rio Perene or "Chanchamayo region". On this route at 3000 m (- 9843 ft.) eleva- tion is a small village named Palca. I have collected specimens of S. variabilis at Palca and several other localities in the Rio Tarma Valley. Simons probably passed through Palca, and collected the series of lizards on his way to the Chanchamayo region; the country of origin was probably not noted on the tags of the specimens, and was thus confused upon the arrival of the collection at the British Museum. The upper river valleys of the Rio Tarma and Rio Mantaro are separated by a ridge at least 3500 m in elevation. Liolaemus alticolor is known from higher elevations between the two drainage systems. Ecological Observations.— Individuals were collected in rock piles, on solitary rocks and on the ground at the bases of shrubs. The region around Palca in the upper Rio Tarma is moderately mesic with Agave, grass, and shrub cover. The Mantaro Valley is more closed and xeric in appearance. The lower elevations at which the species is known have extremely hot and dry climates with Acacia and several cactus species as prevalent plants. Although sympatry with other iguanids remains to be documented, Stenocercus praeomatus and S. boettgeri are known from similar elevations near Palca. A species of Ameiva, similar to A. bifrontata, is sympatric with Stenocercus variabilis near Villa Azul, = 1200 m, Departamento Huancavelica. Remarks.— Specimens from the valley of the Rio Tarma have a distinct postfemoral pocket; those from localities near the Rio Mantaro lack a postfemoral pocket. Varia- tion in four scutellational characters is notable (Table 7). Samples from Palca and Villa Azul have higher means than those from La Mejorada and Ayacucho for the number of scales around the middle of the body. Specimens from Palca and La Mejorada have more scales across the gular region than ones from the southern localities of Ayacucho and Villa Azul. The chin coloration of males is darker in more southern localities. From north to south, specimens from Palca have gray- brown chins with blue suffusion, those from Villa Azul and La Mejorada have extensive black areas on lateral margins of chin, and those from Ayacucho have solid black chin and gular regions. Females show a different pattern of variation; those from Villa Azul, the lowest elevation, have a solid black chin and lateral head. Females from localities at higher elevations possess gray to gray- brown chin coloration. Since considerable variation in coloration and some scutella- tional aspects exists between populations of this species, further study of the taxon with larger samples for morphological analyses and more detailed data on behavior and ecology may show the existence of more than one species. Stenocercus varius Boulenger Fig. 29 Stenocercus varius Boulenger, 1885:134. Type Material.— BMNH 71.4.16.53 from an unknown ANDEAN IGUANID LIZARDS 67 Table 7. Basic statistics of four samples of Stenocercus variabilis. Vertical lines connect nonsignificant (p < 0.05) subsets of SS-STP analysis. Note small sample sizes. N Range SD 2SEx Paica Ayacucho Villa Azul La Mejorada Villa Azul PaIca Ayacucho La Mejorada La Mejorada Pa lea Villa Azul Ayacucho PaIca Villa Azul Ayacucho La Mejorada Scales Across Occipital Region Between Temporals (SBT) 5 12.4 12-13 0.55 0.49 14 11.21 10-13 1.05 0.56 4 11.00 11-11 — — 9 10.4 9-11 0.88 0.59 Scales Around Vlidbody (SMB) 4 71.00 69-74 2.16 2.16 5 70.00 63-74 4.36 3.90 15 62.53 57-67 0.80 1.60 9 62.44 60-64 1.51 1.01 Scales Across Gu lar Region (SAG) 9 32.00 29-36 2.29 1.53 5 32.00 30-34 1.58 1.41 4 27.50 27-29 1.00 1.00 15 27.40 25-31 1.76 0.91 Scales Along Midline of Head Between Occiput and Rostral (SMH) 5 19.80 19-21 0.84 0.75 4 17.75 16-21 2.22 2.22 14 17.07 16-19 1.07 0.57 9 15.89 14-17 0.93 0.62 locality. The type locality is herein restricted to Tan- dapi, 1460m, Provincia Pichincha, Ecuador. Diagnosis. — 1) Posthumeral pocket absent; 2) postfemoral pocket present; 4) scales posterior to tympanum granular, less than one-half the size of dorsal neck scales; 5) scales on posterior thigh granular; 6) vertebral row continuous, distinct from adjacent dorsal scales, forming low verte- bral crest; 7) caudal scales normal, three caudal whorls per autonomic segment; 8) antehumeral fold present; 9) neck fold pres- ent; 10) scales around the middle of body 79-87; 13) dorsum light green with varying amounts of brown mottling, venter yellow; 16) supraoculars between largest row of supraoculars and posteromedial margin of semicircular in two rows; 17) scales across occipital region granular, similar to dorsal neck scales; 18) scales of nasal region small. Stenocercus varius, S. humeralis, and S. haenschi are light green dorsally; S. varius differs from S. humeralis in having fewer scales around the middle of the body; it differs from S. haenschi in having more scales around the body and lateral scales that are markedly smaller than the dorsal scales. Stenocercus boettgeri resembles S. 68 THOMAS H. FRITTS varius in several aspects of scutellation; they differ in characteristics of the occipital region of the head, snout, and supraocular scales (see diagnosis of S. boettgeri). Distribution. — Stenocercus varius is known from the Pacific slopes of Pichincha Province, Ecuador, at elevations of 1500- 1800 m (Fig. 9). The specimens recorded by Burt and Burt (1931) as S. varius from southern Ecuador are referable to Stenocer- cus carrioni. Ecological Observations. — Individuals are observed on tree trunks, fallen logs and rocks in partially cleared areas of montane forest. The forest is quite dense and lush, except where broken by steep hillsides, rock outcrops, or clearing by man. TAXONOMIC SUMMARY Twenty-nine species are recognized in the genus Stenocercus. Steironotus arenar- ius Tschudi 1845 is synonymous with Tropidurus tschudii Roux, 1907 and is a member of the genus Tropidurus. The populations previously known as Ophryoes- soides arenarium (Tschudi, 1845) are identi- fiable as Stenocercus formosus (Tschudi, 1845). Liocephalus formosus Boulenger, 1880 and Liocephalus haenschi Werner, 1901 are synonymous and are assigned to the genus Stenocercus. The former species is a secondary junior homonym of Stenocercus formosus (Tschudi) and is replaced by an available junior synonym, Stenocercus haenschi (Werner, 1901). Ophryoessoides trachycephalus (Dum- eril, 1851), 0. rhodomelas (Boulenger, 1899), O. ornatus (Gray, 1845), 0. guentheri (Boulenger, 1885), 0. festae (Peracca, 1897) and 0. haenschi (Werner, 1901) are transferred to the genus Stenocercus. Stenocercus nigromaculatus (Noble, 1924) is a valid species of the genus Stenocercus despite its omission from Etheridge (1970) and Peters and Donoso Barros (1970). D'Orbigny is considered as the author of the names Stenocercus marmoratus (D'Orbigny) and S. roseiventris D'Orbigny on the basis of authorship noted by Dumeril and Bibron (1837). Stenocercus juninensis Shreve, 1941 is a junior synonym of S. variabilis Boulenger, 1901 and not identifi- able as S. boettgeri Boulenger, 1911 as reported by Etheridge (1970 The genus Ophryoessoides includes six nominate species: 0. triscristatus Dumeril, 1851, 0. aculeatus (O'Shaughnessy, 1879), 0. caducus (Cope, 1862), O. erythrogaster (Hallowell, 1856), 0. iridescens (Giinther, 1859) and 0. scapularis (Boulenger, 1901). The status of all trivial names associated with the genera Ophryoessoides and Steno- cercus is given in Table 8. ANDEAN IGUANID LIZARDS 69 Table 8. Current status of trivial names associated with the genera Ophryoessoides Dumeril, 1851 and Stenocercus Dumeril and Bibron, 1837; in part after Etheridge (1970). Original Name Leiocephalus aculeatus O'Shaughnessy, 1879 Liocephalus angulifer Werner, 1901 Stenocercus apurimacus Fritts, 1972 Steironotus arenarius Tschudi, 1845 Stenocercus atrigularis Werner, 1913 Stenocercus boettgeri Boulenger, 1911 Liocephalus bolivianus Boulenger, 1890 Scartiscus caducus Cope, 1 862 Stenocercus carrioni Parker, 1 934 Stenocercus chrysopygus Boulenger, 1900 Scelotrema crassicaudatum Tschudi, 1845 Stenocercus cupreus Boulenger, 1885 Stenocercus difficilis Werner, 1910 Ophryoessoides dumerilii Steindachner, 1869 Stenocercus empetrus Fritts, 1972 Stenocercus ervingi Stejneger, 1913 Brachysaurus erythrogaster Hallowell, 1856 Leiocephalus festae Peracca, 1 897 Scelotrema formosum Tschudi, 1845 Liocephalus formosus Boulenger, 1880 Liocephalus guentheri Boulenger, 1885 Liocephalus haenschi Werner, 1901 Microphractus humeralis Gunther, 1859 Stenocercus ivitus Fritts, 1972 Liocephalus iridescens Gunther, 1859 Stenocercus juninensis Shreve, 1941 Liocephalus lineogularis Werner, 1901 Trachycyclus marmoratus D'Orbigny, 1837 Stenocercus melanopygus Boulenger, 1900 Stenocercus moestus Boulenger, 1885 Urocentron meyeri Werner, 1900 Stenocercus nigromaculatus Noble, 1924 Stenocercus nubicola Fritts, 1972 Stenocercus ochoai Fritts, 1972 Stenocercus oriental's Fritts, 1972 Saccodeira ornatissima Girard, 1857 Leiocephalus ornatus Gray, 1845 Stenocercus praeornatus Fritts, 1972 Liocephalus rhodogaster Boulenger, 1901 Liocephalus rhodomelas Boulenger, 1899 Stenocercus rosei-ventris D'Orbigny, 1837 Liocephalus scapularis Boulenger, 1901 Stenocercus seydi Andersson, 1908 Stenocercus simonsii Boulenger, 1899 Stenocercus tacnae Shreve, 1941 Stenocercus torquatus Boulenger, 1885 Holotropis trachycephalus Dumeril, 1851 Ophryoessoides tricristatus Dumeril, 1851 Stenocercus variabilis Boulenger, 1901 Stenocercus varius Boulenger, 1885 Current Name Ophryoessoides aculeatus Ophryoessoides aculeatus Stenocercus apurimacus Tropidurus arenarius Stenocercus roseiventris Stenocercus boettgeri Ophryoessoides caducus Ophryoessoides caducus Stenocercus carrioni Stenocercus chrysopygus Stenocercus crassicaudatus Stenocercus cupreus Stenocercus marmoratus Ophryoessoides tricristatus Stenocercus empetrus Stenocercus crassicaudatus Ophryoessoides erythrogaster Stenocercus festae Stenocercus formosus Stenocercus haenschi Stenocercus guentheri Stenocercus haenschi Stenocercus humeralis Stenocercus ivitus Ophryoessoides iridescens Stenocercus variabilis Stenocercus formosus Stenocercus marmoratus Stenocercus melanopygus Stenocercus moestus Stenocercus crassicaudatus Stenocercus nigromaculatus Stenocercus nubicola Stenocercus ochoai Stenocercus oriental is Stenocercus ornatissimus Stenocercus ornatus Stenocercus praeornatus Stenocercus formosus Stenocercus rhodomelas Stenocercus roseiventris Ophryoessoides scapularis Stenocercus formosus Stenocercus simonsii Liolaemus alticolor Stenocercus crassicaudatus Stenocercus trachycepahlus Ophryoessoides tricristatus Stenocercus variabilis Stenocercus varius ANDEAN IGUANID LIZARDS 71 LITERATURE CITED Andersson, L. G. 1908. A remarkable new gecko from South-Africa and a new Stenocer- cits-species from South-America. Jahrb. Nassau. Ver. f. Naturk. Wiesbaden 61: 299-306. Anonymous 1960- BoletinClimatologico Anual. Direc- 1963. 'cion General de Meteorologia del Peru, Lima, Peru. 1964. Anuario Climatologico. Direc- cion General de Meteorologia! del Peru, Lima, Peru. 1965a. Anuario Meteorologico. Direccion General de Meteorologia 1965, Lima, Peru. 1965b. Anuario Meteologico Complemen- tario. Direccion General de Mete- orologia 1965, Lima, Peru. Atchley, W. R. 1970. A biosystematic study of the sub- genus Selfia of Culicoides (Diptera: Ceratopogonidae). Univ. Kansas Sci. Bull. 49: 181-336. 1971. A comparative study of the cause and significance of morphological variation in adults and pupae of Culicoides: a factor analysis and multiple regression study. Evolu- tion 25: 563-583. Barbour, T. and G. K. Noble 1920. Amphibians and reptiles from southern Peru collected by the Peruvian expedition of 1914-1915 under the auspices of Yale Univ- ersity and the National Geographic Society. Proc. U.S. Natl. Mus. 58: 609-620. Bellairs, A. 1970. The life of reptiles. Universe Books, New York. 590 p. Boulenger, G. A. 1880. Reptiles et batraciens dans les Andes de I'Equateur. Bull. Soc. Zool. France 1880: 41-48. 1885. Catalogue of the lizards in the British Museum (Natural History). 2nd Ed. Printed by order of the 1890, 1899. 1900. 1901. 1911. Burt, C 1930. 1931. 1933. CoUette 1961. Cooley, 1971. Cope, E 1862. Trustees, London. Vol. 2. xiii 497 p., 23 pi. First report on additions to the lizard collection in the British Museum (Natural History). Proc. Zool. Soc. London 1890 (I): 77-86. Descriptions of new reptiles and batrachians collected by Mr. P.O. Simons in the Andes of Ecuador. Ann. Mag. Nat. Hist. (7) 4: 454-457. Descriptions of new batrachians and reptiles collected by Mr. P.O. Simons in Peru. Ann. Mag. Nat. Hist. (7) 6: 181-186. Further descriptions of new rep- tiles collected by Mr. P.O. Simons in Peru and Bolivia. Ann. Mag. Nat. Hist. (7) 7: 546-549. Descriptions of new reptiles from the Andes of South America, pre- served in the British Museum. Ann. Mag. Nat. Hist. (8) 7: 19-25. , E. and M. D. Burt The South American lizards in the collection of the United States National Museum. Proc. U.S. Natl. Mus. 78 (6): 1-52. South American lizards in the col- lection of the American Museum of Natural History. Bull. Amer. Mus. Nat. Hist. 61 (7): 227-395. A preliminary check list of the lizards of South America. Trans. Acad. Sci. St. Louis 25 (1): 1-104. . B. Correlations between ecology and morphology in anoline lizards from Havana, Cuba, and southern Flor- ida. Bull. Mus. Comp. Zool. 125 (5): 137-162. W. W. and P. R. Lohnes Multivariate data analysis. John Wiley and Sons, Inc., New York. 364 p. . D. Contributions to neotropical saur- ology. Proc. Acad. Nat. Sci. Phil- 72 THOMAS H. FRITTS adelphia 1862: 176-188. D'Orbigny, A. D. 1837. 350-353, 356-359 p. In. Dumeril, A.M.C. and G. Bibron, 1837. Erpe- tologie generale. Vol. 4, 572 p. 1847. Reptiles. 12 p. Voyage dans 1' Amerique meridionale. Tom. 5, part 1. Dumeril, M. A. 1851. Catalogue methodique de la col- lection des reptiles. Mus. D'Hist. Nat. Paris. 128 p. Dumeril, A. M. C. and G. Bibron 1837. Erpetologie generale. Vol. 4, 572 p. Dunn, E. R. 1944. Herpetology of the Bogota area. Rev. Acad. Colombiana Cien. Ex. Fis. Nat. 6: 68-81. Edit, R. C. 1969. The climatology of South Amer- ica, p. 54-81. In, Fittkau, E. J. et al., ed., Biogeography and ecology in South America Vol. I. Dr. W. Junk N. V. Publishers, The Hague. 442 p. Etheridge, R. 1964. Comparative osteology and sys- tematic relationships of sceloporine lizards. Copeia 1964 (4): 610-631. 1965. The abdominal skeleton of lizards in the family Iguanidae. Herpeto- logicia 21 (3): 161-168. 1966. The systematic relationships of West Indian and South American lizards referred to the iguanid genus Leiocephalus. Copeia 1966: 79-91. 1970. Stenocercus. p. 254-257. In, Peters, J. A. and R. Donoso-Barros, Cata- logue of neotropical Squamata: Part II. Lizards and amphis- baenians. Bull. Smithsonian Inst., 297: 1-293. Fitzinger, Leopoldo J. F. J. 1843. Systema reptilium. Vol. 1. Vienna, 106 p. Fritts, T. H. 1972. New Species of Lizards of the Genus Stenocercus from Peru (Sauria:Iguanidae). Occas. Pap. Mus. Nat. Hist. Univ. Kansas (10): 1-21. Garcia, A. 1959- Anuario Meteorologico No. 1-8. 1968. Servicio Nacional de Meteorologia e Hidrologi'a, Quito, Ecuador. Girard, C. 1857. Descriptions of some new reptiles collected by the United States exploring expedition, under the command of Capt. Charles Wilkes, U.S.M. Fourth part including the species of saurians exotic to North America. Proc. Acad. Nat. Sci. Philadelphia 1857: 195-198. 1858. Herpetology. hi. United States exploring expedition 1838-42. Under the command of C. Wilkes. Vol. 20, xvii, 496, 10 p.: 32 pis. Gray, J. E. 1845. Catalogue of the specimens of liz- ards in the collection of the British Museum. Printed by order of the Trustees, London. 289 p. Gunther, A. 1859a. List of the cold-blooded vertebrata collected by Mr. Eraser in the Andes of western Ecuador. Proc. Zool. Soc. London 1859: 89-93. 1859b. Second list of cold-blooded verte- brata collected by Mr. Eraser in the Andes of western Ecuador. Proc. Zool. Soc. London 1859: 402-420. Hallowell, E. 1856. Notes on the reptiles in the col- lection of the Academy of Natural Sciences of Philadelphia. Proc. Acad. Nat. Sci. Philadelphia. 8: 221-238. Hutchinson, G. E. 1957. Concluding remarks. Cold Springs Harbor Symp. Quant. Biol. 22: 415-427. Inger, R. F. 1967. The development of a phylogeny of frogs. Evolution 21: 369-384. Kluge, A. G. 1969. The evolution and geographical origin of the new world Hemi- dactylus mabouia-brooki complex (Gekkonidae, Sauria). Misc. Publ. ANDEAN IGUANID LIZARDS 73 Mus. Zool. Univ. Michigan 138: 1-78. Kluge, A. G. and J. S. Farris 1969. Quantitative phyletics and the evolution of anurans. Syst. Zool. 18: 1-32. Levins, R. 1968. Evolution in changing environ- ments. Princeton Univ. Press, Princeton. 120 p. MacArthur, R. and R. Levins 1964. Competition, habitat selection, and character displacement in a patchy environment. Proc. Natl. Acad. Sci. 54: 1207-1210. MacArthur, R. and E. Pianka 1966. On the optimal use of a patchy environment. Amer. Nat. 100: 603-609. Mendez-Torrico, G. 1965. Anuario Meteorologico Servicio Nacional de Meteorologia e Hid- rologia. Ministerio de Obras Pub- licas, La Paz, Bolivia. Mertens, R. 1952. Amphibien und Reptilien I. p. 257-266. In, Titschack, E. Beitrage zur fauna Perus. Bd. III. Verlag Gustav Fischer, Jena. Noble, G. K. 1924. New lizards from northwestern Peru. Occas. Papers Boston Soc. Nat. Hist. 5: 107-113. Norris, K. S. 1958. The evolution and systematics of the iguanid genus Uma and its relation to the evolution of other North American desert reptiles. Bull. Amer. Mus. Nat. Hist. 114 (3): 251-326. 1967. Color adaptation in desert rep- tiles and its thermal relationships, p. 162-229. In, Milstead, W., ed.. Lizard ecology: a symposium. Univ. Missouri Press, Colum- bia. 300 p. O'Shaughnessy, A. W. E. 1879. Descriptions of new species of lizards in the collection of the British Museum. Ann. Mag. Nat. Hist. (5) 4: 295-303. Osorno, E. 1938. On the habits of Leiocephalus omatus trachycephalus. Copeia 1938: 200. Parker, H. W. 1934. Reptiles and amphibians from southern Ecuador. Ann. Mag. Nat. Hist. (10) 14: 264-273. Peracca, M. G. 1897. Viaggio del Dr. Enrico Festa nell 'Ecuador e regioni vicine. Boll. Mus. Zool. Comp. Anat. Univ. Torino 12 (300): 1-20. Peters, J. A. and R. Donoso-Barros 1970. Catalogue of the neotropical Squa- mata Part H. Lizards and amphis- baenians. Bull. Smithsonian Inst. 297: 1-293. Roux, J. 1907. Revision de quelques especes de reptiles et amphibians du Perou. Rev. Suisse Zool 15: 293-303. Ruthven, A. G. 1922. The amphibians and reptiles of the Sierra Nevada de Santa Marta, Columbia. Univ. Michigan Mus. Zool. Misc. Publ. 8: 1-69, pi. 1-12, 1 map. Schmidt, K. P. and W. F. Walker, Jr. 1943. Snakes of the Peruvian coastal region. Field Mus. Nat. Hist., Zool. Ser. 24: 297-324. Schoener, T. W. 1967. The ecological significance of sex- ual dimorphism in size in the lizard Anolis conspersus. Science 155: 474-477. Shreve, B. 1941. Notes on Ecuadorian and Peru- vian reptiles and amphibians with description of new forms. Proc. New Englnad Zool. Club 18: 71-83. Smith, G. R. and R. K. Koehn 1971. Phenetic and cladistic studies of biochemical and morphological characteristics of Catostomus. Syst. Zool. 20: 282-297. Sokal, R. R. and F. J. Rohlf 1969. Biometry. W. H. Freeman and Company, San Francisco. 776 p. Steindachner, F. 74 THOMAS H. FRITTS 1869. Reptilien. 98 p. +3 fig. In, Reise der Osterreichischen Fregatte Novara um die Erde. Zoologischer Theil. Stejneger, L. 1913. Results of the Yale Peruvian expe- dition of 1911. Batracians and reptiles. Proc. U.S. Nat. Mus. 45: 541-547. Stoll, N. R., Chm. 1964. International code of zoological nomenclature. International Com- mission on Zoological Nomencla- ture, London. 176 p. Terborgh, J. 1971. Distribution on environmental gradients: theory and a prelim- inary interpretation of distribu- tional patterns in the avifauna of the Cordillera Vilcabamba, Peru. Ecology 52: 23-40. Tschudi, J. J. 1845. Reptilium conspectum quae in in republica Peruana reperiuntur et pleraque observata vel collecta sunt in itinere. Arch, fur Naturg. 11 (1): 150-170. 1846. Herpetologie. p. 1-80. In, Unter- suchungen uber die fauna Peruana. St. Gallen. Tukey. J. W. 1962. The future of data analysis. Ann. Math. Stat. 33: 1;67. Vuilleumier, B. S. 1971. Pleistocene changes in the fauna and flora of South America. Sci- ence 173 (3999): 771-780. Wagner, W. H., Jr. 1961. Problems in the classification of ferns, p. 841-844. In, Recent ad- vances in biology. Univ. Toronto Press, Toronto. Werner, F. 1900. Reptilien und batrachier aus Peru und Bolivien. Abh. Ber. K. Zool. Anthro.-Ethn. Mus. Dresden 9 (2): 1-14. 1901. Uber reptilien und batrachier aus Ecuador und Neu-Guinea. Verh. Zool.-Bot. Ges. Wien 51: 593-614. 1910. Uber neue oder seltene reptilien des naturhistorischen museums in Hamburg. Mitt. Naturhist. Mus. Hamburg 27 (2): 1-46. 1913. Neue oder seltene reptilien und forsche des naturhistorischen mu- seums in Hamburg. Mitt. Natur- hist. Mus. Hamburg 30 (2): 1-39. Whittaker, R. H. 1970. Communities and ecosystems. Macmillan Co., London. 162 p. 76 THOMAS H. FRITTS FIGURES 15-17 Figure 15. Stenocercus orientalis, KU 134456, male, SVL 65 mm. '►=",*•• ■ ♦ '••';t r Figure 16. Stenocercus nubicola, KU 134107, male, SVL 72 mm. Figure 17. Stenocercus omatus, KU 134140, male, SVL 79 mm. ANDEAN IGUANID LIZARDS 77 FIGURES 18-21 Figure 18. Stenocercus nigromacu- latus, KU 134098, male, SVL 68 mm. I'igure 19. Stenocercus guentheri, cf. KU 134610 134617, male, SVL approximately 74 mm, 4 km W San Antonio de Pichincha, Provincia Pichincha, Ecuador. Figure 20. Stenocercus guentheri, KU 134432, male, SVL 81 mm, 5 km NE Alausi, Provincia Chimborazo, Ecuador. Figure 21. Stenocercus guentheri, KU 134503, male, SVL 86 mm, 10 km N Riobamba, Provincia Chimborazo, Ecuador. 78 THOMAS H. FRITTS FIGURES 22-25 > ^ > > Figure 22. Stenocercus rhodomel- as, cf. KU 141164-141166, male, SVL approximately 85 mm. Figure 23. Stenocercus ckrysopy- gus, KU 133912, male, SVL 70 mm, Caraz, Departamento Ancash, Peru. Figure 24. Stenocercus chrysopy- gus, KU 134333, male, SVL 66 mm, Chiquian, Departamento Ancash, Peru. Figure 25. Stenocercus omatissi- mus. KU 134351, male, SVL 76 mm. ANDEAN IGUANID LIZARDS 79 FIGURES 26-29 Figure 26. Stenocercus melanopy- gus, KU 134038, male, SVL 69 mm. Figure 27. Stenocercus simonsii, KU 134165, female, SVL 71 mm. Figure 28. Stenocercus crassicau- datus, KU 139264, male, SVL 78 mm. Figure 29. Stenocercus varius, KU 142704, female, SVL 74 mm. 80 THOMAS H. FRITTS FIGURES 30-33 Figure 30. Stenocercus praeoma- tus, cf. KU 134224-134232, male, SVL approximately 100 mm. Figure 31. Stenocercus formosus, KU 134112, male, SVL 84 mm. Figure 32. Stenocercus ochoai, KU 133888, male, SVL 95 mm. Figure 33. Stenocercus cupreus, KU 133979, female, SVL 56 mm. ANDEAN IGUANID LIZARDS 81 FIGURES 34-37 jp^igfr* Figure 34. Stenocercus roseiventris, KU 134156, male, SVL 83 mm. Figure 35. Stenocercus boettgeri, KU 134011, male, SVL 109 mm. Figure 36. Stenocercus variabilis, KU 134138, male, SVL approxi- mately 65 mm. Figure 37. Stenocercus festae, KU 134608, male, SVL 83 mm. 82 THOMAS H. FRITTS FIGURES 38-41 Figure 38. Stenocernis apurimac- us, KU 134285, male, SVL 84 mm. Figure 39. Stenocercus empetrus, KU 134390, male, SVL 92 mm, Huamachuco, Departamento La Libertad, Peru. Figure 40. Stenocercus empetrus, KU 134413, male, SVL 89 mm. Cajamarca, Departamento Caja- marca, Peru. Figure 41. Stenocercus empetrus, KU 134424, male, SVL 88 mm, Celendin, Departamento Caja- marca, Peru. 84 THOMAS H. FRITTS APPENDIX A A list of specimens examined is given for each species of Stenocercus. The species are arranged alphabetically. The abbreviations of museum names are explained in the materials and methods section. All altitudes have been expressed in metric units. Stenocercus apurimacus PERU: Apurimac: Curahuasi, 2700 m, KU 134240, 134242-134249, 134251-134252; Hacienda Matara, Dist. Matara, MCZ 62253; Pte. Pachachaca, 1800 m, KU 134270- 134309; Rio Apurimac, Pte. Cunyac, 1830 m, KU 134235- 134239, 134253-134257. Cuzco: Limatambo, 2700 m, KU 134263-134269; 3 km E Pte. Cunyac, 1830m, KU 134258; 8 km E Pte. Cunyac, 2300 m, KU 134259-134262. Stenocercus boettgeri PERU: Junin: Chanchamayo, MCZ 45881. Pasco: Hua- chon. MCZ 45842-45843; Oxapampa, AMNH 13625-13628; Paucartambo, 3000 m, KU 134011-134017; 1-3 km NE Paucartambo, 2920-2980 m, KU 139476-139477. Piura: Huancabamba, AMNH 5279, 13502-13509, UMMZ 51277. Stenocercus carrioni ECUADOR: El Oro: Cordillera de Chilla, Llano de Guavos, AMNH 18308; RioLunama, AMNH 22185. Loja: Alamor, MCZ 34866, AMNH 21847, 22112, 22133, 22136-22154, 22156-22157, 22164, 22168-22177, 22187, 22191-22210; 10 km N Celica, 1900 m. MCZ 93589; Guainche, AMNH 22120; Siboyal, 5 mi. NW Alamor, AMNH 21848-21851. Stenocercus chrysopygus PERU: Ancash: Caraz, 2265 m,'KU 133915-133917, MCZ 8073, UMMZ 58920(12), 58915; Carhuas, UMMZ, 58916; Chavin de Huantar, 3230 m, KU 134334-134350; Chiquian, 3200-3500 m, KU 134320-134323, 134325-134333, 134509; Huaraz, 3200-3400 m, KU 133891-133914, MCZ 45832- 45836, AMNH 63475-63476, UMMZ 89487; Recuay, Tica- pampa, FMNH 3943-3944; 5 km N Recuay, 3450 m, KU 133918-133954. Huanuco: 5 km NE La Union, 3100 m, KU 134310-134319. Stenocercus crassicaudatus BOLIVIA: Cochabamba: Yungas de Cochabamba, UMMZ 68115(2). PERU: Cuzco: Huadquinia, USNM 49550; Machu Picchu, 2000-2450 m, KU 133955-133971, 139264- 139266; Rio Cosireni, USNM 60736, 60731-60732, MCZ 12401-12403; San Fernando, Rio Cosireni, USNM 60710, 60712-60713. Junin: Chanchamayo, MCZ 8081, 45882, AMNH 57171. 57177, 56415, 23152-23153, FMNH 40619- 40624. 45481-45483; La Merced, 1525 m, AMNH 57172; Perene. MCZ 29303-29304, AMNH 23124-23133, 23137- 23139, 23143-23144, 23146, 23148, 23151, 23188. Loreto: Uchpayacu, Ri'o Cushabatay, AMNH 57176. San Martin: Pachisa, AMNH 57173. Stenocercus cupreus PERU: Huanuco: Ambo, FMNH 5606-5614; Huanuco, 1900 m. KU 133972-133993, 133995-133999, AMNH 63474, FMNH 3547-3552, 16168-16183, MCZ 43788-43793; 5 km N La Esperanza, 1900 m, KU 133994; Pachachupan, above Acomayo, 2300 m, W.C.S. 2995; Turuca, MCZ 43787. Stenocercus empetrus PERU: Cajamarca: Banos, MCZ 8084; Cajamarca, 2800 m, FMNH 3941-3942, 5710-5711, KU 134411-134413, 3 km E Celendin, 2650 m, KU 134414-134427; Hacienda Huagac, UMMZ 60094-60095. La Libertad: Huamachuco, 3350 m. KU 134379-134400; Laguna Sacsacocha, 12 km E Huama- chuco, 3200 m, KU 134401-134410; Otuzco, FMNH 5708. Stenocercus festae ECUADOR: Azuay: Bastion, AMNH 18315; 2.4 km S Chordeleg on road to Sigsig, 2300 m, KU 141146-141147; Contrayerbas, AMNH 23439-23444; 4 km E Cuenca, 2540 m, KU 134574-134595; 6 km N Cuenca, AMNH 91819-91826; 8.8 km NW Cuenca, 2620 m, KU 141155- 141161; S edge Cuenca near Rio Tomebamba. 2545 m, JAP 2365-2367; 1.2 km S Cuenca, 2565 m, KU 141154. JAP 2369-2375; 1-4 km S Cutchil, 2570-2855 m, KU 141144, 141145, 141149-141153; 8.5 km S Cutchil, 2905 m, KU 141148; Laguna de Zurucuchu, 3200 m, KU 121094; 4 km W San Cristobol, 2500 m, KU 121095; Saraguro, 2500 m, KU 134120-134126; Sinchay. AMNH 23416, 23418- 23421, 23436. 23447-23454. Canar: 3 km S Azogues. 2500 m, KU 134596- 134609; 0.5 km S jet. Pan American Hwy. and Rio Canar. 2835 m, KU 141143. El Oro: Salvias, AMNH 18313-18314. Stenocercus formosus PERU: Junin: Chanchamayo, MCZ 11295; Perene, AMNH 23196. Pasco: Rio Paucartambo, Yaupi, 1600 m, KU 134109-134115. Stenocercus guentkeri ECUADOR: Azuay: Sevilla de Oro, 2286 m, JAP 6522- 6525, 6527-6528, 6530. Chimborazo: Alausi, 2400 m, KU 134428-134429; Alausi Valley, USNM 33859; 5 km NE Alausi, 2600-2750 m, KU 134430-134446; 10 km NW Alausi, 2895 m, JAP 6390, 6529, 6526, 6531; 8.4 km N Cubijies. between Riobamba and Bafios. 2520 m. KU 141251-141255; Guamote, 3050 m, KU 134474-134486, 138827; 9.5 m S first bridge S Guamote, 3190 m, KU 141229-141232; 20.1 km S first bridge S. Guamote, 3190 m, KU 141233141235; Mt. Chimborazo, 4000 m, USNM 103275-103286; 5.2 km S Palmira, 3170 m, KU 141244- 141248; 8.2 km SE Pungala, on road to Aloa, 3085 m, KU 141256; Riobamba, 2780 m, KU 121096-121102; 12.1 km E Riobamba, on road to Licto, 2850 m, KU 141257-141258; 10 km N Riobamba, 2730 m, KU 134487-134508; 7 km W Riobamba, Rio Chibunga, 3291 m, JAP 1726; 4.1 km E San Juan, 3220 m, KU 141259; 6.2 km N Tixan, 3240 m, KU 141242141243; 0.3-3.0 km S Tixan, 2885-2900 m, KU 141236-141239; 4-5 km S Tixan, 2785-2820 m, KU 141240- 141241, 141249. Cotopaxi: 32 km S Aloag, Hacienda Ra- faelito, 3000 m, KU 141174-141178; 1 km W Latacunga, 2750 m, KU 127096-127111; 1 km NE Mulalo, 3050 m. KU 127112-127128. 127095. 128062. 141180-141184, 141189-141194, 142710-142713; 1 km S Mulalo, 3050 m. ANDEAN IGUANID LIZARDS 85 KU- 141 179; 0.3 km W Mulald, 2980 m, KU 141185-141188, 141288; NudodeTiopulla, El Chasqui, 3490 m, KU 132490; Paramd de Tegua, 3700 m, KU 134570-134573; 66.4 km W Pujili, 3585 m. KU 141250. Esmeraldes: Cachabi, AMNH 1349213493. Imbabura: Ibarra, AMNH 13490-13491; Laguna Cuicocha, 10 km W Quiroga, 3000 m, KU 134564- 134569; Otovalo, 2550 m, KU 134535-134555; Quebrada San Miguel, 1 km N Otavalo. 2560 m, KU 118057-118069; San Antonio de Ibarra, 2290 m, KU 134556-134562. Manabi: Santo Domingo, AMNH 27150-27152. Pastaza: Chichirota, Ri'o Bobonaza ?, GOV 8848; Ri'o Solis, Caber- ceras del Ri'o Bobonaza ?, GOV 6493. Pichincha: near Cayambe, 2800 m, UMMZ 105899-105906(8); Guachala, near Cayambe, UMMZ 92174; Laguna de Limpios, N base Volcan Cotapaxi, 3840 m, KU 122120122121; Machachi, UMMZ 55762, 84109; below Facto, GOV 6499; Quito, AMNH 28737-28741, 28743-28744, 28746-28747, 28749- 28767, JAP 3721-3722, GOV 6735-6737, 6966-6967, 7007; 1 mi. ENE Quito, 2895 m, JAP 1675-1677; 1.5 mi. SSE Quito, 2743 m, JAP 1658-1659; San Antonio, 2500 m, KU 134618-134652; 4 km W San Antonio, 2550 m, KU 134610- 134617; 4 km W San Antonio de Pichincha, JAP 4141; Sloa, UMMZ 55512-55513(4). Tungurahua:Ambato, 2700 m KU 121103-121125, GOV 7062-7066, UMMZ 90807(6); 10 km E Ambato, 2560 m, KU 134510-134534; 17.8 km W Banos, 2345 m, KU141204-141225; Chambo Grande, 7.6 km SE Pelileo, 2340 m, KU 142706-142709; Hacienda Leito (Cerro Llanganauti), 2438-2743 m, JAP 5996-6000, 5972; Hacienda Leito, 8.9 km SSE Patate, 2480 m, KU 141203; road SE Patate, 2320 m, KU 141198; 5.6 km SSE Patate, 2415 m, KU 141199-141202; Rio Patate, 2 km S Los Andes, 2133 m, JAP 5919-5923; 3 km SSW San Mi- guelito, 2620 m, KU 132491; 6-7 km S Tungurahua- Riobamba road, 2450-2510 m, KU 141226-141228; Vinge, jet. Rib Chambo and Rio Patate, 2110 m, KU 141195- 141197. Stenocercus haenschi ECUADOR: No locality given, BIN 2007. Stenocercus humeralis ECUADOR: Loja: Loja, 2150 m. KU 121136, 121138, 134000-134004, JAP 2385-2386, 2388-2390, MCZ 126136; 2-3 km E Loja, 2135-2200 m, KU 121137, 141162; 5 km N Loja, 2150 m, KU 134003-134006; 4 km S Loja, MCZ 89795; 12.2 km S. Loja, Rio Malacatos Valley on road to Vilcabamba, 2275 m, KU 141163; 1 km W Loja, 2072 m, JAP 2385-2390; no locality given, AMNH 18307; Valle de Casanga, El Empalme, GOV 6164-6166, 6168- 6171. Stenocercus ivitus PERU: Piura: Summit Cordillera between Canchaque and Huancabamba, 3100 m, KU 134653. Stenocercus marmoratus BOLIVIA: Cochabamba: Arani, 2572 m, UMMZ 68121(2); Pocona, 2700 m, UMMZ 68122(4); Tiraque, UMMZ 68120(20); Totora, 2500 m, USNM 94093, UMMZ 68123- 68124(8). Stenocercus melanopygus PERU: Cajamarca: Barios, 2800 m, MCZ 126133; Caja- bamba, 2700 m, KU 134035-134036; Cajamarca, 2800 m. KU 134037-134088; 3 km E Celendin, 2650 m, KU 138825- 138826; no locality given, FMNH 5712. La Libertad: Huamachuco, 3350 m, KU 134018-134027, 134030-134034; Laguna Sacsacocha, 12 km E Huamachuco, 3200 m, KU 134028-134029. Stenocercus moestus PERU: Lima: Chosica, 762-914 m, FMNH 152204-152207, 39361; vicinity of Lima, MJP 5 unnumbered specimens. Stenocercus nigromaculatus PERU: Piura: Chumaya, AMNH 28531-28532; Huanca- bamba, 1900-2300 m, KU 134089-134106, MCZ 17975, 18767-18768, AMNH 28553-28557, 28559, 28588-28598. Stenocercus nubicola PERU: Piura: Summit of Cordillera between Canchaque and Huancabamba, 3100 m, KU 134107134108. Stenocercus ochoai PERU: Apurimac: Curahuasi, 2700 m, KU 134241, 134250. Cuzco: Chilca, 10 km N Ollantaytambo, 2760 m, KU 133874-133889, 139263; Chospyoe, Rib Huaracondo, USNM 60705-60706; Ollantaytambo, USNM 60810, MCZ 41984; Rio Huaracondo, 3048 m, MCZ 12410; Torontoy, USNM 60811; Ruinas de Machu Picchu, 2400-2450 m, KU 117108, 134233-134234, 139267-139268. Stenocercus orientalis PERU: Amazonas: Chachapoyas, 2340 m, KU 134447- 134473. Stenocercus omatissimus PERU: La Libertad: Otuzco, 2730 m, KU 134351-134378. Lima: Verrugas, USNM 75398. No department given: Yanga, USNM 5655. Stenocercus omatus ECUADOR: Loja: Loja, 2150 m, KU 121126, USNM 98925, 98938; 1-3 km E Loja, 2050-2200 m, JAP 2379- 2384, 2451-2454, KU 121127-121134; 4-5 km N Loja, 2065- 2150 m, KU 134150-134155, 141168-141170; 3-4 km W Loja, 2150-2280 m, KU 134127-134139, 141167; 15 km W Loja, 2150-2700 m, KU 134140-134149. Stenocercus praeomatus PERU: Junin: Comas, 3220 m, KU 134224-134232. Piura: Abra Porculla, 2144 m, SM 65252, 65254, 65257-65259, MVZ 82370. Stenocercus rhodomelas ECUADOR: Azuay: 4 km SW Catavina (2 km NE Abdon Calderon), 1600 m, JAP 3540; 7-8 km SW Giron, 2100 m, JAP 3518-3522; 12 km SW Giron, 2000 m, JAP 3524; Pan American Hwy at Rio Leon, N Oiia, 1885 m, KU 141164- 141166; Rib Leon, 12.5 km N Ofia, 1920 m, KU 142699- 142703; 4 km E San Francisco, 1250 m, JAP 3636-3654; Santa Isabel, GOV 6487-6490, 6982, 6491, 6978, 6988, 7028, 7033, 7039-7040; 10 km W Santa Isabel, JAP 3709- 3710. Stenocercus roseiventris BOLIVIA: Cochabamba: No locality, AMNH 6766, UMMZ 68119; Todos Santos Trail, AMNH 9037; Tucachaca, 86 THOMAS H. FRITTS UMMZ 68118(5);Yungas de Cochabamba, UMMZ 68132, USNM 94094. Santa Cruz: Buena Vista, FMNH 21523, 16164, AMNH 37096, UMMZ 60582-60587, 60601, 60612, 60811; Provincia Sara, MCZ 126140; no locality given, UMMZ 68116-68117. No department given: Rio Sirutu, UMMZ 63783(2). Sucre: No locality given, MCZ 11298, FMNH 3947. PERU: Ayacucho: Ocros, Hacienda Pajonal, 2000 m, FMNH 81424. Cuzco: Cordillera Vilcabamba, 870 m, AMNH 101384-101385; Marcapata, Hacienda Cadena, FMNH 62937-62938; Mission Coribeni, 24 km ENE La Rosalina, 680 m, KU 134156; Quincemil, 780 m, FMNH 168233-168234, 168242, 168246, 168256, 168266, 168351, 168353-168355. Huanuco: Rio Pachitea, Monte Alegre, AMNH 57200. Juni'n: Chanchamayo, AMNH 56309, 57167-57170. Loreto: Bajo Amazonas, Iquitos, 1200 m, FMNH 45478; Ucayali, Yarinacocha, 160 m, FMNH 45479-45780. Madre de Dios: Avispas, 480 m, FMNH 168293. Stenocercus simonsii ECUADOR: Azuay: Giron, 2240-2500 m, KU 134157- 134164. Loja: Saraguro, 2500 m. KU 134165-134174. Stenocercus trachycephalus COLOMBIA: Boyaca: Duitama, MCZ 67978. Caldes: La Dorada ?, AMNH 27611, 27613. Cundinamarca: vicinity of Bogota, MCZ 14159-14160, 17136-17144, 22011-22013, AMNH 7644-7646, 24216, 24224-24225, 24260, 24282- 24283, 32682-32697, 91751, USNM 75958, 92493. 95178, 90064, 153972-153973. UMMZ 56758, 78293, 100192; 16 km N Bogota, AMNH 91745-91750; Bosque Popular, NE Bogota, MCZ 67977; Cajica, Sabana de Bogota, MCZ 67979; Embalsa La Regadera, AMNH 97365; Facatativa. AMNH 27625; Girardot ?, USNM 153987-153991; Gutier- rez, SE Bogota, AMNH 32678-32679, 32760-32761; La- guna La Granja, N Bogota, MCZ 67981; Madrid, Sabana de Bogota, MCZ 67980; Pantanos Martos. Finca Vista Hermosa, AMNH 97364; Paramo de Cruz Verde, Bogota, MCZ 110743; Rio Chico, N Bogota, MCZ 67975; Sabana de Bogota. MCZ 67974; Suba, Sabana de Bogota. MCA 67976. No department given: Guasin. MCZ 17092. San- Under: Mountains near Muso. AMNH 27528, 27532, 27534, 27539, 27546-27547, 27552-27553, 27556-27557, 27559, 27566. ^ Stenocercus variabilis PERU: Ayacucho: Ayacucho, 2760 m, KU 134198-134207, 134214-134219. 134221-134223; 4 km N Ayacucho, 2750 m, KU 134211-134213; 20 km N Ayacucho, 2700 m, KU 134208-134210. Huancavelica: La Mejorada (Mariscal Caceres), 2790 m, KU 134180-134192; Villa Azul, 17 km by road ESE Colcabamba, approx. 1200 m. KU 134193- 134197. Junin: Huacapistana, 2500 m, FMNH 40617; Huasqui, Tarma region, MCZ 45820-45821; 4-5 km W Palca, 3000 m, KU 134175-134179. Stenocercus varius ECUADOR: Pichincha: 2 km W Campemento Silante. 2100 m, KU 132492-132494; 5-6 km E Chiriboga, 1830 m, JAP 3715, UMMZ 128830; Finca Santa Lucia, 7.7 km E Chiriboga. 2120 m, KU 142704; Tandapi, 1460-1500 m, KU 121184-121186, 121135, 134563. No province given: Las Maguinas, AMNH 27135. Stenocercus sp. ECUADOR: Carchi: Valle de Chota, GOV 9043-9067, 9070, 9072; no specific locality, GOV 6159-6163. Imbabura: El Juncal, Valle de Chota, GOV 7008-7018; nr. Salinas, 1600 m, UMMZ 105912; Pimpanpiro, San Nicolas, 2000 m, UMMZ 105913-105916. Pichincha: Palma Real, Gov 6975-6976. Stenocercus sp. ECUADOR: Loja: 10 km ENE Catamayo, 1500 m. KU 134116-134119. THOMAS H. FRITTS APPENDIX B The samples used in the quantitative analyses are listed in alphabetical order. Samples indicated with asterisks are used only in the quantitative phyletic analysis. ALAM. — Stenocercus carrioni, Ecuador: Provincia El Ore: Alamor, =600 m, AMNH 21847. 22136, 22168- 22169, 22172, 22187, 22196, 22201. AY AC. — Stenocercus variabilis, Peru: Departamento Ayacucho: Ayacucho, 2760 m, KU 1.34198-134199, 134204-134206. 134209-134214, 134217. 1,34219, 134221- 134223. *AZUL. — Stenocercus variabilis. Peril: Departamento Huancavelica: Villa Azul, =1200 m, KU 134193, 134195- 134197. BAMB. — Stenocercus boettgeri, Peru: Departamento Piura: Huancabamba, =1900 m, AMNH 13502-13503, 13505, 13508-13509. CAJA. — Stenocercus melanopygus, Peru: Departa- mento Cajamarca: Cajamarca, 2800 m, KU 134037- 134088. CELE. — Stenocercus empetrus, Peru: Departamento Cajamarca: Celendin, 2650 m, KU 134414. 134417- 134420, 134422-134427. CHAC. — Stenocercus orientalis, Peru: Departamento Amazonas: Chachapoyas, 2340 m, KU 134447-134473. CHAP. — Stenocercus roseiventris, Bolivia: Departa- mento Cochabamba: vicinity of Yungas de Chapare, =300 m, UMMZ 68118, 68119(4), 68132(5). CHAV. — Stenocercus chrysopygus, Peru: Departa- mento Ancash: Chavin de Huantar, 3230 m, KU 134334- 134350. *CHIL.— Stenocercus ochoai, Peru: Departamento Cuzco: Chiica, 2760 m. KU 133874, 133878-133886, 133888-133889. CHIQ. — Stenocercus chrysopygus, Peru: Departamento Ancash: Chiquian, 3200-3400 m, KU 134320-134333, 134509. CHUC. — Stenocercus empetrus, Perii: Departamento LaLibertad: Huamachuco, 3350 m, KU 134382-134388, 134390-134400. *CRUZ. — Stenocercus roseiventris, Bolivia: Departa- mento Santa Cruz: vicinity of Buena Vista, UMMZ 60582(3), 60583, 60585, 60586, 68116(2), 68117(4). COMA. — Stenocercus praeomatus, Peru: Departamento Juni'n: Comas, 3220 m, KU 134224-134232. GIRO. — Stenocercus simonsii, Ecuador: Provincia Azuay: Giron, 2240-2.500 m, KU 134157-134164. HUAM. — Stenocercus melanopygus, Peru: Departa- mento La Libertad: Huamachuco, 3350 m, KU 134018- 134027, 134030-134034. HUAN. — Stenocercus cupreus, Peru: Departamento Huanuco: Huanuco, 1900 m, KU 133972-133993, 133997- 133999. HUAR. — Stenocercus chrysopygus, Peru: Departa- mento Ancash: Huaraz, 3250 m, KU 133892-133909, 133911-133914. LIMA. — Stenocercus moestus, Peru: Departamento Lima: vicinity of Lima (5 specimens unnumbered MJP). LOJA. — Stenocercus humeralis, Ecuador: Provincia Loja: Loja, 2150 m, KU 134000-134004, JAP 2385-2386, 2388-2390. MACH. — Stenocercus crassicaudatus, Peru: Departa- mento Cuzco: Machu Picchu, 2000-2400 m, KU 133957- 133971, 139264-139266. MEJO. — Stenocercus variabilis, Peru: Departamento Huancavelica: La Mejorada, 2790 m, KU 134180-134182, 134184-134187, 134191-134192. PALC. — Stenocercus variabilis, Peru: Departamento Junin: vicinity of Palca, 3000 m, KU 134175-134179. *PALM. — Stenocercus humeralis, Ecuador: Provincia Loja: Valle de Casanga, El Empalme, GOV 6164-6166, 6168-6171. PAUC. — Stenocercus boettgeri, Peru: Departamento Pasco: Paucartambo, 3000 m, KU 134011, 134013- 134017. PICH. — Stenocercus ochoai, Peru: Departamento Cuzco: Machu Picchu, 2000 m, KU 117108, 134233, 139267- 139268. *PORC. — Stenocercus praeomatus, Peru: Departa- mento Piura: Abra PorcuUa, 2144 m, SM 65252, 65254, 6.52.57-65259, MVZ 82370. OTUZ. — Stenocercus omatissimus, Peru: Departamento La Libertad: Otuzco, 2730 m, KU 134351-134378. *RECY. — Stenocercus chrysopygus, Peru: Departa- mento Ancash: 5 km N Recuay, 3450 m, KU 133918- 133954. *SACS. — Stenocercus empetrus, Peru: Departamento La Libertad: Laguna Sacsacocha (12 km E Huama- chuco), 3200 m, KU 134401-134408, 134411-134413. SARG. — Stenocercus simonsii, Ecuador: Provincia Loja: Saraguro, 2500 m, KU 134165-134168, 134170. 134172, 134174. TAND. — Stenocercus varius, Ecuador: Provincia Pi- chincha: Tandapi, 1500 m, KU 132492-132494, 134563. TIRA. — Stenocercus marmoratus, Bolivia: Departa- mento Cochabamba: Tiraque,=3000 m, UMMZ 68120- (13). TOTO. — Stenocercus marmoratus, Bolivia: Departa- mento Cochabamba: Totora, =2902 m, UMMZ 68123- 68124(8). UNIO. — Stenocercus chrysopygus, Peru: Departamento Huanuco: 5 km NE La Union, 3100 m, KU 134310- 134318. YAUP. — Stenocercus formosus, Peru: Departamento Pasco: Yaupi, 1600 m, KU 134109-134110, 134112, 134114. ANDEAN IGUANID LIZARDS 89 Figure 42. Each dot represents one of the 27 localities from which the 29 samples used in the quantitative analyses were taken. Samples of two sympatric species were used from Huamachuco, La Libertad and Machu Picchu, Cuzco, Peru. Date Due ACME BOOK"l^^'riG CO., inc. NOV 28 1984 100 CA. c^iREET a*.RLESi\jWN, MASS. 3 2044 093 SeiTg