This file was created by scanning the printed publication. Errors identified by the software have been corrected; however, some errors may remain.

Herpetofauna in Riparian Habitats Along the River in Grand Canyon 1

2 Peter L. Warren and Cecil R. Schwalbe

Abstract.-- population densities and composition were sampled in riparian and non-riparian habitats along the Colorado River. The highest densities were found in shoreline habitats, moderate densities in riparian habitats and lowest densities in non-riparian habitats. Rapidly fluctuating river flow levels may have a deleterious effect on lizard populations by trapping populations on alluvial bars and inundating nest sites.

For years riparian habitats have been recognized as making a contribution to the densities in upland vegetation may actually be structural diversity and species richness of higher. natural communities that exceeds the relative areal extent of those habitats. The availability One group that has received relatively little of additional water permits growth of plant attention with respect to the importance of species and growth forms that are lacking in the riparian habitats to their density and diversity surrounding upland vegetation. Their occurrence is the . It is common to find comments in in turn provides food and habitat resources the literature about the higher density of some without which some populations may not species in riparian sites (Lowe and Johnson, 1977; otherwise persist in the upland community. To Vitt and Ohmart, 1977; Tinkle, 1982) and some most biologists these patterns are obvious, but in studies of lizard demography have been performed many cases they are surprisingly poorly in riparian areas (Tinkle, 1976; Tinkle and documented. Dunham, 1983; Vitt and Van Lohen Sels, 1976). However, quantitative studies comparing Some of the best studied examples of the density and diversity in riparian and adjacent contribution of riparian habitat to local species non-riparian habitats are few. Only recently has density and diversity are for birds and mammals. emphasis on riparian ecosystems has begun to Gallery forests of cottonwood and willow along address effects of management practices and exotic some Southwestern rivers have been shown to have riparian vegetation on riparian reptile some of the highest densities of nesting birds in communities (Szaro et al., 1985; Jakle and Getz, North America, much higher than in surrounding 1985; Jones and Glinski, 1985). semiarid upland sites (Johnson et al., 1977; Anderson, Higgins and Ohmart, 1977). Riparian The present study was designed to examine the habitats contribute breeding sites, feeding areas patterns of distribution of reptile species and migratory routes for birds. Hammal species relative to riparian habitats along the Colorado diversity is also higher along watercourses, where River in Grand Canyon National Park. This work is some species find necessary cover that is lacking part of a larger study to determine the effects of in more open adjacent arid vegetation (Anderson, fluctuating flows from Glen Canyon Dam on plant Drake and Ohmart, 1977), although small mammal and animal populations in and along the Colorado River. Data presented here were gathered during constant flow levels of approximately 40,000 cubic 1 feet per second (cfs) in June and 25,000 cfs in Paper presented at the First North American August, 1984. Additional censuses will he made Riparian Conference [University of , during lower, fluctuating flow levels. The Tucson, April 16-18, 1985]. results presented here are from the first year of 2 a multi-year project, and are restricted to only Peter L. Warren is Research Assistant at the those species for which the most data were Arizona Remote Sensing Center, Office of Arid gathered, the diurnal . Lands Studies, University of Arizona, Tucson, Arizona. Cecil R. Schwalbe is Herpetologist with Nongame Branch, Arizona Game and Fish Department, 2222 W. Greenway Road, Phoenix, Arizona.

347 STUDY AREA Habitats Sampled

We censused lizard populations at a series of Sampling was performed in ten different sites along the Colorado River above and in Grand habitats that are distributed in three zones Canyon National Park beginning near Lees Ferry and relative to the river. The first zone comprised extending downstream 220 miles almost to Diamond shoreline habitats within 5 meters of the river Creek. The elevation at river level is shore. The second zone included all riverine approximately 945 meters (3,100 feet) at Lees riparian vegetation greater than five meters from Ferry and drops to 427 meters (1,400 feet) at the the river shore. The third zone included non­ last census locality at mile 220. The vegetation river habitats, both upland and riparian (Table through which the river flows is generally Mohave 1). desertscrub. However, there is a gradual transition in species composition from more cold­ Three distinct habitats were sampled in the tolerant species at the upper end of the study river shoreline zone. These were cobble shore, area to a flora composed of many frost-sensitive rocky shore, and cliff faces at the water's edge. species at the lower end (Warren et al., 1982). In all cases shoreline samples were characterized by low vegetation cover, usually less than ten The riparian corridor along the river is percent. Cobble shores generally were charac­ characterized by two vegetation zones that are terized by num~rous rocks less than 0.5 meters in more or less distinct in species composition and diameter and rounded by erosion. Larger, uneroded distribution. Previous to the construction of boulders were absent and large patches of bare Glen Canyon Dam in 1963 the river channel was sand were occasional. Cobble shores generally were scoured by floods on a regular basis, and the only found at the mou~hs of tributary canyons where the riparian vegetation occurred as a belt along the coarse alluvium that forms level cobble bars was high water line where flood disturbance was at a washed into the river. m1n1mum. Since dam construction lack of large­ volume flooding has permitted plants, many of them exotics, to grow along the water's edge (Turner and Karpiscak, 1980). The resulting pattern is Table 1.--Location of study sites at which lizard one in which the original riparian vegetation, transect sampling was performed. The number consisting largely of mesquite (Prosopis of habitats sampled in each vegetation zone glandulosa) and cat-claw acacia (Acacia greggii), is indicated for each site. is perched on talus slopes and alluvial terraces several meters above the current normal water Site River Shore­ River Non­ level. The new riparian vegetation, dominated by Name Mile line Riparian River tamarisk (Tamarix chinensis) and arrowweed Lee's Ferry -1R 1 1 (Tessaria sericea), occupies sand and cobble bars along the water's edge. Badger 8R 1 none 16L 1 none 20R 1 North Canyon 20.5R 1 METHODS none 43.5L 1 Saddle Canyon 47R 1 3 1 Visual belt transects, modified from the Nankoweep 53R 2 3 2 Emlen (1971) bird census technique, were used to Kwagunt 56R 1 census the common diurnal species (Lowe and Cardenas 71L 1 4 Johnson, 1977). This method involves walking a Cremation 86L 1 transect through representative areas of the none 94L 1 target habitats and recording all individuals Crystal 98R 1 observed within a belt of predetermined width of Bass 108.5R 1 2 1 four meters. Transect length varies with size of Elves Chasm 116.5L 1 1 the habitat patch, but usually varied from 100 to Forster 123L 2 300 meters in length. Transect sites were Tapeats 134R 1 1 selected to sample a range of variation within none 140L 1 1 old- and new-riparian habitats and in adjacent Kanab 143.5R 1 4 non-riparian desertscrub. The time of day at the National 166L 2 1 beginning and end of each transect was recorded as 171R 1 well as a temperature profile consisting of soil Stairway 2 none surface temperature, air temperature at 5 mm and 185R 1 3 1 air temperature at 1.5 m. Weather conditions such Whitmore 188R 3 Parashant as cloudiness and wind speed were also noted. 198R 1 3 1 Granite Park 209L 1 1 Three Springs As each individual lizard was sighted, the 216L 1 220 mi. Canyon 220R 1 distance along the transect and the substrate upon which it was first observed were recorded, as well Total Transects 24 36 8 as its sex and age, when possible. The substrate categories used were bare soil, litter, rock (less Total Transect than one meter diameter), boulder (greater than Length (meters) 2665 5522 2420 one meter diameter), cliff face, or tree. When individuals were in a tree, the tree species and height above ground were also recorded.

348 In contrast, rocky shores were composed of can be considered "new-zone" or post-dam habitats. rock fragments of varying sizes ranging from These were open tamarisk with 15 to 40 percent cobbles up to boulders several meters in diameter. cover, dense tamarisk with 60 to 100 percent These shores were generally composed of uneroded cover, and arrowweed with cover similar to the talus and rockfall debris and may include pockets open tamarisk. of bare sand of varying sizes that were trapped among the boulders. In contrast to the level Finally, two habitats were sampled in the cobble shores, rocky shores usually fell steeply non-river zone. These were desertscrub on canyon to the water's edge and were commonly very rugged slopes generally ranging from 15 to 30 percent and irregular. slope with 15 to 30 percent vegetation cover, and non-river riparian habitats along perennial Sandy shores and heavily vegetated shores tributary streams. were examined but not sampled systematically for several reasons. Heavy vegetation immediately at RESULTS AND DISCUSSION the water's edge was relatively uncommon. In most locations where dense cover was present near the Sampling was performed during 18 days in June shore it occured on sandy soil. Frequently and five days in August, 1984. A total of 68 erosion of sandy soil along the river's edge kept the immediate shoreline free of dense cover even transects were sampled at 27 localities, with though adjacent sandy bars were thickly vegetated. between one and five habitats sampled per locality Open sandy shorelines that lacked vegetation or (Table 1). Preliminary habitat assessments were rock cover were found to be almost completely free made during September, 1983 and April, 1984. of reptile and amphibian activity, and although such sandy shores were spot-checked repeatedly, no Five common diurnal lizard species were systematic transects were sampled. successfully sampled using the belt transect method. One lizard species ( maculata), two toad species (Bufo punctatus and B. Within the riverine riparian zone five woodhousei), one frog species (Hyla arenicolor~ habitats were sampled. These included two that and three snake species (Crotalus viridis, can be considered "old-zone" or pre-dam habitats, Masticophis taeniatus, and ~ flagellum)were mesquite/acacia alluvial terraces and encountered in numbers too small for adequate mesquite/acacia talus slopes. The remaining three conclusions to be drawn concerning distribution patterns.

Table 2.--Distribution of lizard occurrence on different substrates along the Colorado River in Grand Canyon, June and August, 1984. Numbers in parentheses indicate the percent of individuals of each species that were observed on each substrate.

Species Substrate Litter Bare Rock Boulder Cliff Tree Total Soil 1m

Uta 2 70 71 2 1 4 150 stansburiana ( 1. 3) (46.7) (47.3) ( 1. 3) (0.7) (2.7)

Cnemido,ehorus 9 78 4 3 0 1 95 tigris (9.5) (82.1) (4.2) (3.2) ( 1.1)

Sceloporus 11 11 7 34 3 22 88 magister (12.5) (12.5) (8.0) (38.6) (3.4) (25.0)

Urosaurus 3 1 9 16 27 5 61 ornatus (4.9) ( 1.6) (14.7) (26.2) (44.3) (8.2)

Crotaphytus 0 1 4 2 0 0 7 insularis (14.3) (57.1) (28.6)

Sauromalus 0 0 0 1 0 0 1 obesus (100)

Holbrookia 0 1 0 0 0 0 1 maculata (100)

Total 25 162 95 58 40 32 403

349 Substrate Preference Patterns of Density and Diversity

The lizards showed strong species-specific The most striking observation was the large patterns of substrate preference (Table 2). All differences in lizard densities among the habi­ species were observed most frequently on a tats sampled (Table 3). Total lizard densities substrate different from any other species, were highest in shoreline and open, "new zone" although four species were commonly observed along riparian habitats and lowest in desertscrub, with a single transect in one habitat. intermediate densities in "old zone" sites. Most of the individual species followed the same pat­ Side-blotched lizards (Uta stansburiana) were tern with highest densities in shoreline and "new the most common species observed as well as the zone" habitats and lowest density in desertscrub. smallest. Utas were found predominately in open The only exception to this pattern were collared sites and the substrates upon which they were most lizards which, although relatively rare, were seen frequently observed were rocks less than one meter more commonly in desertscrub than any other in diameter and bare soil. They were almost never habitat. seen at a distance greater than one meter away from cover in the form of rocks or small shrubs. The pattern of differences in lizard densities among habitats was stable through time Western whiptail lizards (Cnemidophorus as shown by comparison of June and August data tigris), the second most abundant species, were (Fig. 1). Regression analysis of density data found most frequently on bare soil or litter. gathered in the same habitats during two months They frequently occurred in the same habitats with indicates that although overall observed densities Uta, but were rarely seen perched on small rocks declined from June to August, the ranking of habi­ ag- Uta does. Cnemidophorus was the only species tats based on density remained the same. This observed commonly roaming up to several meters was possibly the result of cooler, cloudier across bare sand away from cover. weather encountered during the August census and consequent lower activity levels of some species. Desert spiny lizards (Sceloporus magister) Whiptails and desert spiny lizards both declined were approximately equal in abundance to in observed densities by approximately one-half Cnemidophorus, although they were less noticeable between the two census periods. due to more sedentary habits and preference for cryptic substrates with a strong vertical Comparison of density values derived from component, such as large boulders and trees. visual transects in this study with density data Desert spiny lizards were seen most commonly on available in the literature is difficult for boulders larger than one meter in diameter, and several reasons. First and most important is that usually on those with fractures and crevices. At our visual census does not attempt to account for sites that lacked boulders but had trees, such as every lizard in the study site as a mark/recapture tamarisk stands on sand bars, this species was study on a permanent grid does. Visual transect also found on larger tree trunks. On those estimates will therefore generally be lower than a occasions when they were observed on the ground, comparable mark/recapture estimate. Second, they were almost invariably at the immediate base lizard densities vary to large degree between of a large tree or boulder. sites, between years, and even seasons or days, at a single site. Thus any comparison of densities, Tree lizards, Urosaurus ornatus, were also regardless of the sample technique, is fraught found on substrates with a strong vertical compo­ with problems unless the sampling is performed nent. However, they showed a clear preference for simultaneously at all sites to be compared. With sheer, vertical rock faces on cliffs or large these problems in mind, it is still useful to boulders. The highest densities of tree lizards compare our results with those density data that were found on cliff faces that dropped vertically are available in the literature. into the river, usually along eddies or quiet stretches. They often sat less than one meter In general the lizard densities observed above water level, just above the splash zone, on along the Colorado River fell within the range of faces that had no fractures or other protection values that have been observed for these species and that were up to 20 to 40 meters away from the in other areas (Table 4). That species which we nearest water-level alluvial soil. observed to occur in the highest density, Urosaurus ornatus, was also reported by several Black collared lizards, Crotaphytus authors to have the highest density of lizard insularis, and chuckwallas, Sauromalus ohesus, species studied. Similarly, of the four most were observed much less frequently than the four common species, we generally found Sceloporus preceding species. These two species also magister to have the lowest density. This species differed from the others insofar as both species was reported by several authors usually to have were seen more often in desertscruh than in the lower densities than the other three common riparian corridor. Collared lizards generally species. These results indicate that visual were observed perched on rocks or small boulders transect data are roughly comparable with that were approximately one meter in diameter or mark/recapture data. slightly smaller. Chuckwallas rarely were seen on transects, hut additional observations indicated The observed average June densities of 858 that they prefered deeply fractured boulders and lizards per hectare on shoreline cliff-faces and rock outcrops. 300 lizards per hectare in non-river riparian habitats equal or exceed lizard densities reported

350 Table 3.--Lizard densities in habitats along the Colorado River in Grand Canyon, Arizona during June and August, 1984. Values indicate number of individuals encountered per hectare.

Habitat Lizard S2ecies All Month Uta Cnemi- Scelo2- Uro- Crota- Lizards do,ehorus ~ saurus 2hytus

Shoreline (<5m2: Rocky Shore June 48 23 60 20 0 150 Aug. 20 0 0 100 0 120

Cobble Bar June 68 40 15 0 3 125 Aug. 60 18 13 0 0 90

Cliff Face June 0 0 0 858 0 858 Aug. 0 0 0 223 0 223

River RiEarian (>Sm~: New Zone

Open Tamarisk June 53 78 55 13 0 195 Aug. 53 60 60 0 0 173

Arrowweed June 35 35 5 0 0 73 Aug. 33 18 18 0 0 68

Dense June 0 13 40 0 0 53 Tamarisk Aug. no sample

-----Old Zone Terrace June 30 15 15 3 1 65 Aug. 0 0 13 25 0 38

Talus June 28 10 15 0 0 53 Aug. no sample

Non-River:

Desertscrub June 18 8 5 0 2 30 Aug. 5 5 0 0 5 15

Riparian June 25 0 125 150 0 300 Aug. 208 0 0 0 0 208.

Grand Mean June 35 25 23 10 0.7 93 (All habitats) Aug. 30 13 13 23 1 80

in the literature for any habitat. This resources. Many shoreline sites appear to have observation is of particular interest considering much greater numbers of insects than non-riparian the expected under-estimate of visual census areas for two major reasons. First, debris washed compared to mark/recapture methods discussed up along the water's edge in eddies and backwaters above. The lizard densities we observed in is frequented by many insects. Second, many riparian habitats along the Colorado River were riparian plant species support a larger insect higher than those in most habitats thusfar fauna than non-riparian species (Stevens, 1976). documented in the Southwest. They were up to an The highest local lizard densities observed order of magnitude higher than densities in anywhere along the river were both at sites along desertscrub immediately adjacent to the river the shoreline where lizards were feeding upon corridor. insects. The highest density was observed at Cardenas where a total of eight Cnemido,ehorus The most likely explanation for these high tigris and five Scelo2orus magister were observed densities is an increased abundance of food feeding along the shoreline in an area of

351 Table 4.--Comparison of average lizard densities in Grand Canyon with those from other localities. Ranges are shown in parentheses. In some cases the range of values are from replicate sampling in adjacent sites, and in some cases from sampling in different years. The range of values is not published in all cases.

Species Average Density Location Source (Number/Ha)

Jta 140 (62-238) Tinkle, 1967 stansburiana 22 Ariz. desertscrub Vitt & Van Loben Sels 1976 7 Ariz. mesquite II II 7 Ariz. riparian II II 33 (0-208) All habitats This study

Cnemidoehorus 12 (8-18) Nevada Turner, et al, 1969 tigris 8 (3-15) Texas Degenhardt, 1966 17 Colorado McCoy, 1965 30 Nevada Tanner & Jorgensen, 1963 114 (45-184) Texas Milstead, 1965 .... 3 Ariz. grassland Lowe & Johnson, 1977 12 Ariz. desertscrub Vitt & Van Loben Sels, 1976 32 Ariz. mesquite II II 32 Ariz. riparian II II 7 Ariz. dry wash II II 19 (0-78) All habitats This study

Sceloeorus 15 , riparian Tinkle, 1976 magister 10 Ariz. desertscrub Vitt & Van Loben Sels, 1976 25 Ariz. mesquite II II 25 Ariz. riparian II " 18 (0-125) All habitats This study

Urosaurus 158 (131-188) Ariz., spring Tinkle & Dunham, 1983 ornatus 101 (42-161) Ariz., summer II II 370 Ariz. mesquite Vitt & Van Lob en Sels, 1976 185 Ariz. riparian II II 16 (0-858) All habitats This study

Total 6 (2-12) Southwest deserts Pianka, 1967 Lizards 55 Ariz. riparian Lowe & Johnson, 1977 66 Ariz. grassland II " 8 Ariz: Chihuahuan desert II II 593 Ariz. mesquite Vitt & Van Loben Sels, 1976 277 Ariz. riparian II II 89 Ariz. Sonoran desert II II 12 Ariz. dry wash " II 86 (15-858) All habitats This study '·/ approximately three by seven meters, or a density habitats that contained a mosaic of bare sand and equivalent to 6,500 lizards per hectare!! In cover such as cobbles and small shrubs. Uta spite of their close proximity to one another, no juveniles were the most common and were often seen antagonistic interactions were observed between on cobble bars and shoreline. Tinkle's (1967) individuals of either species, all of which were observations that average first-year dispersal of active in the area for an hour. The second juvenile Utas is less than six meters suggests highest density was observed on a vertical rock that these--habitats are the location of higher face at the waterline on which eight Urosaurus reproductive activity than non-riparian sites. ornatus were observed in an area of two by twenty­ Future study of nest site selection will clarify five meters, or 1,600 per hectare. Again, they the level of reproductive activity in the were feeding on insects at the waters edge. different habitats.

Reproductive activity of lizards along the The distributions of several of the lizard Colorado River was not evaluated directly, but species studied were consistent with the concept indirect evidence of reproduction was inferred of "preferential" riparian species as used by from the distribution of first year immature Johnson et al. (1984) in their discussion of plant individuals. The greatest number of immature species distributions. Urosaurus, Cnemidoehorus, lizards were observed in shoreline and riparian Sceloeorus and Uta could be considered

352 second, rising water during the breeding season / from May to July may inundate nest sites in shore­ • / line and riparian-zone sand. / ACKNOWLEDGEMENTS This research was supported by the Bureau of / Reclamation and the Arizona Game and Fish Department as part of the Glen Canyon Environmen­ / tal Studies through cooperative agreement 4-A6-40- 01810. The National Park Service provided collec­ tion permits. Park Service personnel, especially >. /. ·c;; John Thomas, Steven Carothers, James Gaddy, and -c: Jon Dick, assisted with logistics. Many Arizona (1) "C Game and Fish personnel contributed to various (1) c: aspects of project initiation and field work :::::1 J including James C. DeVos, Jr., Terry Johnson, Ray Lee, Henry Maddux, William Persons, and Bruce Taubert. Terry B. Johnson reviewed the manu­ script. Humphrey Summit Associates provided cheerful, professional logistic support.

LITERATURE CITED

Anderson, B.W., A.E Higgins and R.D. Ohmart. 1977. Avian use of saltcedar communities in the August density lower Colorado River valley. In Importance, Preservation and Management -of Riparian Figure 1.--Comparison of average total lizard Habitats: a Symposium. U.S.D.A. Forest densities for two dates in seven habitats Service Gen. Tech. Rep. RM-43 p. 128-136. along the Colorado River in Grand Canyon. Circles indicate shoreline habitats (less Anderson, B.W., J.F. Drake and R.D. Ohmart. 1977. than 5 meters from shoreline), triangles Population fluctuations in nocturnal rodents indicate riparian habitats (greater than 5 of the in the lower Colorado River valley. meters from shoreline), and square indicates In Importance, Preservation and Management of desertscrub. Regression equation is y=1.28x Riparian Habitats: a Symposium. U.S.D.A. + 3.75 (r=0.98, n=7). Forest Service Gen. Tech. Rep. RM-43 p.183- 192.

"preferential" riparian species by virtue of their Carothers, S.W. and S.W. Aitchison, eds. 1976. higher densities in riparian habitats compared to An ecological survey of the riparian zone of non-riparian. As with the original application of the Colorado River between Lees Ferry and the these terms to plant distributions, it is Grand Wash Cliffs, Arizona. Tech. Rept. No. important to note that these classifications refer 10, Grand Canyon National Park, Colorado only to local distribution and do not apply River Research Series. 251 pp. throughout the species' ranges. Emlen, J.T. 1971. Population densities of birds derived from transect counts. Auk. 88:323- CONCLUSIONS 342.

Shoreline lizard densities along the Colorado Degenhardt, W.G. 1966. A method of counting some River were found to be higher than densities in diurnal ground lizards of the genera riverine riparian vegetation, which in turn were Holbrookia and Cnemido:ehorus with results higher than non-riparian desertscrub densities. from the Big Bend National Park. Amer. Shoreline densities for the four most common Midland Natur. 74:61-100. species were higher than densities previously reported for those species anywhere else in the Jakle, M.D. and T.A. Gatz. 1985. Herpetofauna use southwest. The reason for the high densities of four habitat types of the middle Gila observed is probably high food availability on River drainage, Arizona. This conference. riparian plants and on debris along the water's edge. Johnson, R.R., L.T. Haight and J.M. Simpson. 1977. Endangered species vs. endangered It is possible that rapidly fluctuating river habitats: a concept. In Importance, flow levels will have deleterious effects on Management and Preservation of Riparian shoreline lizard populations for two reasons. Habitats: a Symposium. U.S.D.A. Forest First, rapidly rising water could trap and destroy Service Gen. Tech. Rep. RM-43 p. 68-79. large numbers of individuals on alluvial bars, and

353 Johnson, R.R., P.L. Warren, L.S. Anderson and C.H. Tanner, W.W. and C.D. Jorgensen. 1963. Reptiles Lowe. 1984. Stream order in ephemeral of the Nevada Test Site. Brigham Young watercourses: a preliminary analysis from the University Sci. Bull., Biol. Series III(3). Sonoran Desert. Proc. Arner. Water Resources 31 p. Assoc.--Southwestern Div. 14:89-100. Tinkle, D.W. 1967. Horne range, density, dynamics, Johnson, T.B. and C.H. Lowe. 1978. A survey of and structure of a Texas population of the the amphibians and reptiles of Coronado lizard Uta stansburiana. In: W.W. Milstead, National Memorial. Tech. Rept. No. 5, Coop. ed. Lizard ecology-a symposium. Univ. Miss. Nat. Park Resources Studies Unit, University Press. p. 5-29. of Arizona, Tucson. 1976. Comparative data on the popula­ Jones, K.B. and P. Glinski. 1985. Microhabitats tion ecology of the desert spiny lizard, of lizards in a Southwestern riparian Sceloporus magister. Herpetologica. 32:1-6. community. This conference. 1982. Results of experimental density Lowe, C.H. and T.B. Johnson. 1977. Fishes, manipulation in an Arizona lizard community. amphibians and reptiles of of the Rosemont Ecology. 63:57-65. site. In R. Davis and J.R. Callahan, eds. An environmental inventory of the Rosemont Tinkle, D.W. and A.E. Dunham. 1983. Demography area in southern Arizona. University of of the tree lizard, Urosaurus ornatus, in Arizona, Dept. of Biological Sciences. central Arizona. Copeia. 1983:585-598. Unpublished report. Turner, F.B., P.A. Medica, J.R. Lannom Jr. and McCoy, C.J., Jr. 1965. Life history of and G.A. Hoddenbach. 1969. A demographic ecology of Cnernidophorus tigris septentri­ analysis of fenced populations of the onalis. Ph.D. diss. Univ. of Colorado, whiptail lizard, Cnernidophorus tigris, in Boulder. southern Nevada. Southwestern Naturalist. 14:189-202. Milstead, w.w. 1967. Lizard ecology - a symposium. University of Missouri Press, Turner, R.M. and M.M. Karpiscak. 1980. Recent Columbia. 300 pp. vegetation changes along the Colorado River between Glen Canyon Darn and Lake Mead, Pianka, E.R. 1967. On lizard species diversity: Arizona. U.S.G.S. Prof. Paper No. 1132. North American flatland deserts. Ecology. 48:333-351. Vitt, L.J. and R.D. Ohrnart. 1977. Ecology and reproduction of lower Colorado River lizards: Stevens, L.E. 1976. Insect production of native II. Cnernidophorus tigris (Teiidae), with and introduced dominant plant species. In: comparisons. Herpetologica. 33:223-234. S.W. Carothers and S.W. Aitchison, eds. An ecological survey of the riparian zone of the Vitt, L.J. and R.C. Van Loben Sels. 1976. The Colorado River between Lee's Ferry and the herpetofauna of the Orme Darn site. Arizona Grand Wash Cliffs, Arizona. Grand Canyon State University, Dept. of Zoology. National Park Colorado River Research Series Unpublished report. 57 pp. Contr. No. 38. Warren, P.L., K.L. Reichhardt, D.A. Mouat, B.T. Szaro, R.C., S.C. Bellfit, J.K. Aitkin and Brown and R.R. Johnson. 1982. Vegetation of J.N.Rinne. 1985. A preliminary study on the Grand Canyon National Park. Coop. National impact of grazing on riparian garter snakes. Park Resources Studies Unit Tech. Rep. No. 9, This conference. University of Arizona. 140 pp.

354