A monitoring study of vertebrate community ecology in the northern Sonoran Desert, Arizona
Item Type text; Dissertation-Reproduction (electronic)
Authors Rosen, Philip Clark
Publisher The University of Arizona.
Rights Copyright © is held by the author. Digital access to this material is made possible by the University Libraries, University of Arizona. Further transmission, reproduction or presentation (such as public display or performance) of protected items is prohibited except with permission of the author.
Download date 29/09/2021 06:48:25
Link to Item http://hdl.handle.net/10150/289117 INFORMATION TO USERS
This manuscript has been reproduced from the microfilm master. UMI films the text directly from the original or copy submitted. Thus, some thesis and dissertation copies are in typewriter face, while others may be from any type of computer printer.
The quality of this reproduction is dependent upon the quality of the copy submitted. Broken or indistinct print, colored or poor quality illustrations and photographs, print bleedthrough, substandard margins, and improper alignment can adversely affect reproduction.
In the unlikely event that the author did not send UMI a complete manuscript and there are missing pages, these will be noted. Also, if unauthorized copyright material had to be removed, a note will indicate the deletion.
Oversize materials (e.g., maps, drawings, charts) are reproduced by sectioning the original, beginning at the upper left-hand comer and continuing from left to right in equal sections with small overiaps.
Photographs included in the original manuscript have been reproduced xerographically in this copy. Higher quality 6" x 9" black and white photographic prints are available for any photographs or illustrations appearing in this copy for an additional charge. Contact UMI directly to order.
Bell & Howell Infomiation and Learning 300 North Zeeb Road, Ann Art)or, Ml 48106-1346 USA 800-521-0600
A MONITORING STUDY OF VERTEBRATE COMMUNITY ECOLOGY
IN THE NORTHERN SONORAN DESERT, ARIZONA
by
Philip Clark Rosen
Copyright ® Philip Clark Rosen 2000
A Dissertation Submitted to the Faculty of the
DEPARTMENT OF ECOLOGY AND EVOLUTIONARY BIOLOGY
In Partial Fulfillment of the Requirements For the Degree of
DOCTOR OF PHILOSOPHY
In the Graduate College
THE UNIVERSITY OF ARIZONA
2000 UMI Number: 9965915
Copyright 2000 by Rosen, Philip Clark
Ail rights reserved.
UMI'
UMI Microform 9965915 Copyright 2000 by Bell & Howell Information and Learning Company. All rights reserved. This microform edition is protected against unauthorized copying under Title 17. United States Code.
Bell & Howell Information and Learning Company 300 North Zeeb Road P.O. Box 1346 Ann Arbor, Ml 48106-1346 2
THE UNIVERSITY OF ARIZONA ® GRADUATE COLLEGE
As members of the Final Examination Committee, we certify that we have
read the dissertation prepared by Philip Clark Rosen entitled A Monitoring Study of Vertebrate Community Ecology
in the Northern Sonoran Desert, Arizona
and recommend that it be accepted as fulfilling the dissertation requirement for the Degree of Doctor of Philosophy
Date
Robert jj. Frypi^r /J A/ Date ' , m hmuMM 2 //S //t Lundberg J Date^'
^"1:0 ^ 0- o O nez^el Rio Date
. /O hd ''Michael L. Ros'^n^eig / Date I.U.M Cecil R. Schwalbe Date Final approval and acceptance of this dissertation is contingent upon the candidate's submission of the final copy of the dissertation to the Graduate College.
I hereby certify that I have read this dissertation prepared under my direction and recommend that it be accepted as fulfilling the dissertation requirement.
% t ¥- Dissertation Director ^ Date 3
STATEMENT BY AUTHOR
This dissertation has been submitted in partial fulfillment of requirements for an advanced degree at The University of Arizona and is deposited in the University Library to be made available to borrowers under rules of the Library.
Brief quotations from this dissertation are allowable without special permission, provided that accurate acknowledgement of source is made. Requests for permission for e.xtended quotation from or reproduction of this manuscript in whole or in part may be granted by the copyright holder.
SIGNED: 4
ACKNOWLEDGEMENTS
The National Park Service established and funded the monitoring study. Natural Resources Management monitoring staff at Organ Pipe Cactus National Monument (Jonathan Arnold, James Barnett. Charles Conner. Roy Irving. Ami Pate. Jim Petterson) organized and carried out much of the monitoring reported on here. In addition to carrying out much of the monitoring with great precision and timeliness. NPS provided housing, tleld assistance and information, and protected the study from effects of vandalism and economic encroachment. The impressive monitoring efforts of Charles Conner and Ami Pate are primarily responsible for the datasets on mammals, lizards, and climate. Yet others played an indispensable role in e.xtending the monitoring efforts, and in providing important natural history detail and continuity: Peter Holm. Michael Lee. David Parizek. Shawn Sartorius. Elizabeth Wirt. Many others played major roles assisting with tleld work, including Steve Booth. Roger Eagan. Cal Lowe. Margaret Marshall, Daniel Mello, Julia Rosen; and many more have participated more occasionally. This was important for sustaining the scope of the work. The Department of Ecology and Evolutionary Biology supported me during this long project, including with funding to purchase trap materials. Many outside my own department have a.ssisted during this project, especially through the Cooperative Park Studies Unit (now in the U.S. Geological Survey) and elsewhere in the School of Renewable Natural Resources. University of Arizona, in particular helping administratively on many projects contributory to the monitoring: Brenda Carbajal. Joan Ford. Cecily McCleave. Sandra Mosoff. Mary Greene. Carole Wakely. Projects funded by the Arizona Game and Fish Department's Heritage Grants Program for work at ORPI allowed my monitoring efforts to continue. My committee members have encouraged and sustained me through an unusually long graduate career, and have been a source of intellectual challenge and striving for e.xcellence: Robert Frye. Lee Graham, John Lundberg. Carlos Martinez del Rio. Michael Rosenzweig, Cecil Schwalbe. David Vleck. Yar Petrysyzn not only designed the mammal monitoring protocol, but served as a de facto committee member by providing insight into functioning of the mammal assemblage. Dr. Schwalbe deserves many more thanks for years of protection from the hierarchy, as well as for advisement and collaboration on many projects, including the monitoring program. Charles H. Lowe, my committee chairman and mentor made this work possible. His ability to create opportunities where obstacles stood; to drive, stimulate, encourage and support; his demand for ecological realism; and his love for and knowledge of the desert and its workings have made this period one of great enjoyment for me. 5
TABLE OF CONTENTS
LIST OF TABLES 9
LIST OF FIGURES 13
ABSTRACT 16
INTRODUCTION TO THE DISSERTATION 18 OUTLINE OF THE DISSERTATION STRUCTURE 18 THE STATUS OF COMMUNITY ECOLOGY 18 CONNECTION BETWEEN ECOLOGY AND MONITORING 21 THE ORPI ECOSYSTEM MONITORING PROGRAM 22
I. SMALL MAMMAL POPULATION TRENDS 25 CHAPTER SUMMARY 25 INTRODUCTION 27 METHODS 30 STUDY AREA 30 LANDSCAPE STRUCTURE OF THE SMALL MAMMAL FAUNA 34 MONITORING METHODS 37 Sherman Trap Grids 37 Drift-fence Traps 39 Road Transect Method 40 Time-Constrained Search (TCS) 41 RESULTS 42 NOCTURNAL RODENTS 42 Dipodomys nierriami 42 Chaetodipiis penicillatus 55 Perognatlius ampins 57 Neotoma albigida 58 Other Nocturnal Rodent Species 59 DIURNAL RODENTS 60 LAGOMORPHS 60 CORRELATION AMONG METHODS 62 DISCUSSION 64 METHODS EVALUATION AND RECOMMENDATIONS 64 Road-Cruising, Drift-Fence Trapping, and Visual Search 64 Live-Trapping Protocoi 66 CAUSES OF POPULATION CHANGES 71 Climatic Effects, and Literature Comparisons 71 Life Historical and Inter-specific Competitive Effects 76 Predation and Assemblage Dynamics 79 6
TABLE OF CONTENTS -- Continued
Predation and Community Structure 84 Summary of Predation Effects 88
II. POPULATION MONITORING OF LIZARDS 90 CH.\PTER SUM.VI.\RY 90 INTRODUCTION 91 METHODS 93 STUDY AREA DESCRIPTION 93 Setting. Climate. Landscape and Vegetation Structure 93 Ecological Factors and Interactions Important for Lizards 97 Climate Methodology and History 100 LIZARD MONITORING METHODS LOL Transect Methods 102 Intensive Study Area: Trapping Methods 104 Time-Constrained Search (TCS) 105 Handling. Measurement and Marking 105 COMPUTATIONS AND STATISTICS 106 Analysis of Transect Results 106 Data Reduction and Analysis of Trap Captures 107 RESULTS 108 LINE TRANSECTS 108 Landscape Structure of the Fauna 108 All Sites Combined 110 Intensive Study Area (OSSA) 114 DRIFT FENCE TRAPS 118 CORRELATIONS AMONG METHODS 123 DISCUSSION 123 MONITORING METHODS 123 POPULATION RESPONSES TO CLIMATE AND BIOTIC INTERACTIONS 129
III. ENDOTHERMIC PREDATOR TRENDS. WITH A LITERATURE-BASED ESTIMATE OF DIET 136 CHAPTER SUMMARY 136 INTRODUCTION 137 METHODS 140 STUDY AREA 140 MONITORING METHODS 141 ESTIMATION OF DIETS FROM THE LITERATURE 143 RESULTS 146 COMPOSITION AND STRUCTURE OF THE FAUNA 146 7
TABLE OF CONTENTS - Continued
TEMPORAL TRENDS OF PREDATOR ABUNDANCE 151 DIET CHARACTERISTICS OF PREDATORS 154 TEMPORAL CHANGE IN THE PREDATOR ASSEMBLAGE 160 DISCUSSION 163
IV. MONITORING TRENDS FOR SNAKES 166 CHAPTER SUMMARY 166 INTRODUCTION 167 METHODS 169 STUDY AREA DESCRIPTION 169 Habitat and Climatic Setting 169 Potential Human Effects on the Study System 170 SNAKE MONITORING METHODS 172 Drift-fence Trapping Technique 172 Data Reduction and Variance for the Trap Results 173 Time-Constrained Search (TCS) Method 176 Road Transect Method 177 SYNTHESIS OF MONITORING METHODS RESULTS 177 RESULTS 178 PRELIMINARY OBSERVATIONS ON ASSEMBLAGE STRUCTURE 178 MONITORING RESULTS 178 DISCUSSION 191 ASSEMBLAGE STRUCTURE 191 TEMPORAL TRENDS 191
V. CLIMATIC AND TROPHIC CORRELATES OF POPULATION FLUCTUATIONS 198 CHAPTER SUMMARY 198 INTRODUCTION 200 METHODS 206 ESTIMATION OF ANNUAL PREY TAXA ABUNDANCE TIME-SERIES 206 Nature of the Data 206 Data Reduction 208 ESTIMATION OF ANNUAL PREDATION INTENSITY 209 CLIMATIC DATA AS A PROXY FOR PRODUCTIVITY 211 STATISTICAL METHODS: CORRELATION AND PATH ANALYSIS 212 RESULTS 222 STANDARDIZED TIME-SERIES 222 SMALL MAMMALS 225 Bivariate Correlation Structure 225 8
TABLE OF CONTENTS -- Continued
Model I Path Analysis 234 Model II Path Analysis 235 Model III Path Analysis 235 LIZARDS 236 Bivariate Correlation Structure 237 Model I Path Analysis 245 Model II Path Analysis 245 Model III Path Analysis 246 DISCUSSION 247 CORRELATIONS: EFFECTS OF PRECIPITATION 251 SMALL MAMMAL POPULATION REGULATION 253 Perognathines 254 Dipodomys merriami 256 Other Small Mammals 257 POPULATION REGULATION IN THE LIZARDS 259 CONCLUSION 262
DISSERTATION SUMMARY 264
APPENDIX I. List of monitored taxa 270
APPENDIX II. Weighting used to combine monitoring datasets into best available estimates of relative annual abundance 272
APPENDIX III. Standardized timeseries used in the correlation and path analysis for vertebrates at Organ Pipe Cactus National Monument. 1987-1998 273
REFERENCES 275 9
LIST OF TABLES
TABLE LL Nocturnal rodent species composition by macrohabitat 35
TABLE 1.2A. Population density of nocturnal rodent species at the OSSA monitoring site based on the Petryszyn Method index 43
TABLE 1.2B. Petryszyn Method biomass estimates (g/ha) for the OSSA site 43
TABLE 1.3A. Population density of nocturnal rodent species for five Core I valley monitoring sites based on the Petryszyn Method index 44
TABLE 1.3B. Petryszyn Method biomass estimates for 5 core valley desertscrub sites, ORPI 44
TABLE 1.4A. Population density of the four predominant nocturnal rodent species at the OSSA site based on the modified Lincoln- Petersen Index 45
TABLE 1.4B. Biomass estimates for the OSSA site based on Lincoln- Petersen procedure 45
TABLE 1.5A. Population density of the predominant nocturnal rodent species. 5 Core I monitoring sites, based on the modified Lincoln-Petersen Index .... 46
TABLE 1.5B. Biomass estimates for the 5 Core I valley sites, based on Lincoln-Petersen procedure 46
TABLE 1.6A. Number of drift fence trap captures of rodents 52
TABLE 1.6B. Percentage of juvenile rodents in the drift fence captures 52
TABLE 1.7. Consistency index comparing different monitoring methodologies 63
TABLE 1.8. Life history parameters for small mammals found at ORPI 77
TABLE 2.1. Species for species matching in the desert lizard fauna at Organ Pipe Cactus National Monument 109
TABLE 2.2. Yearly summary of lizard line transect results Ill 10
LIST OF TABLES -- Continued
TABLE 2.3. Seasonal values for lizard line transect results 117
TABLE 2.4. Yearly summary of drift fence trap results for lizards 119
TABLE 2.5. Time-constrained search results Coleonyx variegaius 124
TABLE 2.6. Pearson product-moment correlation coefficients comparing monitoring methodologies for lizards 125
TABLE 2.7. Special anti-predator tactics in the two major desert lizard communities atORPI 134
TABLE 3.1. Endothermic predators of vertebrates observed at ORPl 147
TABLE 3.2. Rate of observation (number seen per day) of different ta.xa of endothermic predators 148
TABLE 3.3. Results of informal monitoring of owl sounds at night 150
TABLE 3.4A. Nocturnal endothermic predators in valley habitat at ORPI; literature data on the diet in Western North America 155
TABLE 3.4B. Diurnal endothermic predators in valley habitat at ORPI: literature data on the diet in Western North America 156
TABLE 3.5. Species diets for Western North America, arid and semi-arid habitats .. 157
TABLE 3.6. Computed diet expected for endothermic predators in the desert valley ecosystem 159
TABLE 3.7. Frequency of encounter of endothermic predators that eat primarily small versus large prey 161
TABLE 4.1. Snake observations made during monitoring work on the intensive snake study region monitoring area 178
TABLE 4.2. Observation types used during monitoring work on the intensive snake study region monitoring area 180 11
LIST OF TABLES — Continued
TABLE 4.3. Preliminary estimate of biomass distribution among snake species on the intensive snake study region 181
TABLE 4.4. Snake relative abundance among years based on yr-adjusted drift-fence trap capture rate 182
TABLE 4.5. Nocturnal time-constrained search results for snakes 186
TABLE 4.6. Nocturnal sand-road transect results for snakes 188
TABLE 4.7. Annual variation in the reproductive frequency of rattlesnakes 197
TABLE 5.1. Correlation coefficients between small mammal population change and precipitation 226
TABLE 5.2. Correlation coefficients between small mammal population change and predation measures 227
TABLE 5.3. Correlation coefficients between small mammal population changes and abundances 228
TABLE 5.4. Time-lagged correlations between small mammal population changes and abundances 229
TABLE 5.5. Model I (2 independent variables) path coefficients from precipitation, intra-specific abundance, inter-specific abundance of possible competitors, and predation to small mammal population change . 230
TABLE 5.6. Model II path coefficients from precipitation, intra- and inter specific abundance, and predation to small mammal population change 231
TABLE 5.7. Summary model path coefficients for rodents 232
TABLE 5.8. Correlation coefficients between lizard species population change and precipitation 238
TABLE 5.9. Correlation coefficients between lizard population change and predation measures 239 12
LIST OF TABLES ~ Continued
TABLE 5.10. Correlation coefficients between lizard population change and lizard abundances 240
TABLE 5.11. Time-lagged correlations between lizard population changes and abundances 241
TABLE 5.12. Model 1 (2 independent variables) path coefficients from precipitation, intra-specific abundance, inter-specific abundance of possible competitors, and predation to lizard population change 242
TABLE 5.13. Model II path coefficients from precipitation, intra- and inter specific abundance, and predation to lizard population change 243
TABLE 5.14. Summary model path coefficients for lizards 244 13
LIST OF FIGURES
FIGURE 1.1 A. Annual (calendar year) precipitation during, and in the climatic phase preceding, this study 32
FIGURE I.IB. Seasonal precipitation record, 1982-1998 33
FIGURE 1.2. Rodent abundance on the Organ Pipe Cactus National Monument Snake Study Area 47
FIGURE 1.3. Rodent abundance on the 5 Core I Ecological Monitoring Program study sites 48
FIGURE 1.4. Annual trends of population density at ORPI Snake Study Area for the 4 abundant nocturnal rodent species 49
FIGURE 1.5. Annual trends of population density at 5 ORPI EMP monitoring sites for the 4 abundant nocturnal rodents 50
FIGURE 1.6. Drift fence trap capture rates for small mammals 53
FIGURE 1.7. Night-time road transect results for rodents 54
FIGURE 1.8. Nocturnal time-constrained search results for abundant rodent taxa .... 56
FIGURE 1.9. Combined road transect and time-constrained search results for three species of rabbits, ORPI 61
FIGURE 2.1. .Annual lizard line transect results. Arizona 112
FIGURE 2.2. Correlated and contrasting results for lizard species abundances on the transects 113
FIGURE 2.3A. Seasonal results of lizard line transect work at ORPI 116
RGURE 2.3B. Seasonal results of lizard line transect work at ORPI 117
FIGURE 2.4. Annual drift-fence trap capture rates for various lizard species at the OSSA site 120
RGURE 2.5. Simplified summary and contrasts of lizard abundance revealed by drift-fence trapping 121 14
LIST OF FIGURES -- Continued
FIGURE 2.6. Temporal aspects of the trophic dynamics of the monitored lizard assemblage at ORPI 130
FIGURE 3.1. Rate of observation of endothermic predators of vertebrates at ORPI .. 152
FIGURE 3.2. Visual observations for endothermic prey taxa 152
FIGURE 3.3. Standardized abundance trajectories for intensively monitored prey groups at ORPI 162
FIGURE 3.4. Contrasting temporal patterns of food-type sub-guilds at ORPI 162
FIGURE 4.1. Preliminary estimate of biomass distribution for snakes at the intensive snake study region. Valley of the Ajo. Arizona. 1987-1998 182
FIGURE 4.2. Standardized annual drift-fence trap capture rates for snakes at the ORPI Snake Study Area 184
FIGURE 4.3. Annual variation in time-constrained search (TCS) results for snakes at the intensive snake study region. ORPI 187
FIGURE 4.4. Annual variation in road-cruising results on a sand road. intensive snake study region. ORPI 189
FIGURE 4.5. Weighted, standardized mean of annual values for snakes from drift-fence, TCS. and road-cruise methods of snake monitoring on the intensive study region. ORPI 190
FIGURE 4.6. Temporal patterns in the trophic dynamics of the snake assemblage on the intensive snake study region, ORPI 193
FIGURE 5.1. Path diagram (generalized) with two predictor variables influencing population change 214
FIGURE 5.2. Path diagram (generalized) with three predictor variables influencing population change 215
RGURE 5.3. Path diagram (generalized) for the summary model of variable inter-relationships 219 15
LIST OF FIGURES - Continued
FIGURE 5.4. Small mammal population trajectories at Organ Pipe Cactus National Monument, 1987-1998, shown with other variables used in the analyses 223
FIGURE 5.5. Lizard population trajectories at Organ Pipe Cactus National Monument. 1988-1998. shown with other variables used in the analyses .... 224
FIGURE 5.6. Path diagrams summarizing conclusions for four ta.xa with the strongest monitoring data 248
FIGURE 5.7. Seasonal time-series of precipitation, and population structure and abundance of whiptail lizards 261 16
ABSTRACT
I synthesized monitoring results for vertebrates at Organ Pipe Cactus National
Monument (ORPI), 1987-1998. Small mammals, lizards, and predators were studied
using ongoing Ecological Monitoring Program (EMP) protocols (trap grids; transects)
and other methods (drift-fences, visual encounter, road-cruising). EMP protocols for
rodents and lizards performed well, but some recalibration is needed.
Populations declined to observed minima during a 1989-1990 drought, and increased with strong rains during 1990-1995. Small rodents (pocket mice) increased
fastest, but declined first, after 1992. The medium-sized Merriam's kangaroo rat increased
to a 1994 peak and then also collapsed. The larger packrat increased most slowly, not declining until after 1995. These temporal differences are consistent with a tradeoff of capacity for increase with resistance to predation pressure.
.After post-drought increases, most lizard populations declined when predator
pressure became high, after 1992. and then increased during dry years after 1995. while
predators declined.
Endothermic predators were monitored by simple daily record-keeping. They increased 3-4 fold from 1989-1995, with subsequent declines. A literature review showed two subguilds: a small-prey group, which increased rapidly in 1991-1993, and a larger-
prey group, which increased more slowly.
For most snakes, population fluctuations during 1989-1998 did not appear dramatic. Western diamondback rattlesnakes and coachwhips were the most important
mammal- and reptile-eating snakes, respectively. Large cohorts of young rattlesnakes 17 were produced during 1992 and 1993. The western diamondback approximately doubled by 1995. and the coachwhip increased during 1991-1993.
I summarized results for prey taxa using bivariate correlation and path analysis. I used precipitation as a proxy for food productivity, and constructed predation pressure indices that combined snakes and endotherms. Conspecitlc density was the most consistent (negative) correlate of population growth. Predation (negative) and productivity (positive) also had relatively consistent associations with annual prey population growth. Lizard population growth was positively correlated with summer rain, whereas some rodents and endothermic predators had positive correlations with winter rain. The analysis supported a competitive effect of Merriam's kangaroo rat on pocket mice.
I recommend additing predator monitoring to the EMP. and propose that resource management and academic ecology may develop a beneficial collaboration in the context of monitoring programs. 18
INTRODUCTION TO THE DISSERTATION
OUTLINE OF THE DISSERTATION STRUCTURE
This dissertation is arranged in five chapters, two about prey (small mammals,
lizards), two about predators (endotherms, snakes), and one synthesizing the time-series
results from the previous four chapters. The first four chapters present population
monitoring results from 1987-1998. and discuss them in relation to methodology,
monitoring needs, and relationships of focal taxa to climate and species interactions. Each chapter is designed so that it can be modified to stand alone.
THE ST.vrus OF COMMUNITY ECOLOGY
Community ecology is a difficult and challenging discipline. Whereas a great deal
is known about succession, guilds, assemblage and geographic patterns, and ecosystem-
level trophic structure (e.g., Clements. 1916; Lindeman, 1942; MacArthurand Levins.
1967; Cody and Diamond. 1973; Whittaker, 1975; Ricklefs and Schluter. 1993;
Rosenzweig. 1995; Poiis and Winemiller. 1996), predictive theory and specific
understanding of real systems remain elusive. We can list the kinds of interactions that species may have (predation. parasitism, competition, various mutualisms, and indirect effects) and provide examples of them (see Begon et al., 1996; Ricklefs and Miller. 2000;
Smith, 1996), but we have not evaluated their relative importances in general or in more
than a handful of specific instances. 19
We are still trying to settle key debates, many of them quite old (Mcintosh, 1995).
Are co-occurring species functionally integrated in communities or individualistically assembled (Clements, 1916; Gleason, 1926; Whittaker, 1975)? Are assemblages random or highly structured (see Strong et al., 1984)? Does diversity and food-web complexity beget stability (see Pimm. 1991; Naeem et al.. 1995; Polis and Winemiller. 1996)? Is food-chain length limited by energy, stability considerations, or some other factor
(Sterner et al.. 1997)? Are a few keystone taxa (Paine, 1966). or many species with varied and diffuse linkages and effects (McCann et al.. 1995; Polis. 1995). truly important for community dynamics and conservation (see Meffe et al.. 1997)? Are populations regulated by predation, or are they chaotically lluctuating (Hanski et al.. 1993)? Few of these and other key questions are resolved, and even ideas that have been largely rejected are still widely held among non-ecologists. The crisis in ecology has led to scathing self- criticism (Peters. 1991).
Three major research programs dominate modem community ecology. Most prominent is (1) direct experimentation with populations and abundance (Wilbur. 1972;
Davidson et al.. 1980; Bender etal.. 1984; Hairston. 1989; Scheiner and Gurevitch. 1993;
Lawton. 1995; Wilbur. 1997). usually in some sort of enclosure isolating a small, pre determined set of interactions, or in more complex but very small microcosms (Lawton.
1995. 1996; Ives et al.. 1996). In nearly three decades of intensive practice, this paradigm has had great success demonstrating the existence of competition, and elucidating its possible complexities (Carpenter et al., 1995), although its application to undisturbed natural ecosystems remains to be demonstrated. Another developing approach (2) has 20 involved careful analysis of optimal behavioral characteristics (e.g., Stephens and Krebs,
1986), which yields predictions and experiments that have readily measurable responses
(J.S. Brown, 1988). which in turn depend on the nature and strength of competitive and predatory interactions (see Rosenzweig and Abramsky, 1997). and thence reveal forces that may be structuring assemblages (J.S. Brown et al.. 1994a). Physiological and biomechanical capabilities have yet to be formally included in this mechanistic, behavioral paradigm, even though they may underlie it decisively. (3) Food-web modelling has been another approach, which has been more purely mathematical (May.
1973; Pimm. 1982, 1991; Martinez and Lavvton. 1995; Law and Morton. 1996). with increasingly promising results that, nonetheless, suffer from the difficulty (e.g.. Morton et al.. 1996) or the complexity (Polis, 1991) of real ecosystems.
Ecologists are developing impressive experimental methods for exploring small parts of ecosystems, yet lack a large database describing real system behavior. Further, the physical requirements for experimentation at the community level (Wilbur, 1997) are so great that most vertebrate ecosystems cannot be studied exclusively or primarily by such methods. A broader consensus may be emerging among ecologists developing or analyzing long-term vertebrate population data. European ecologists working with microtine rodents are concluding now that predation contributes strongly to control of dramatic population cycles (Hanski et al.. 1991; Lindstrom et al.. 1994; Hanski and
Henttonen, 1996; Lambin et al., 2000). and this is being recognized in Japan (Saitoh et al.. 2000) and North America (Korpimaki and Krebs, 1996), as well. European ecologists are extending their observations to other vertebrate systems (Nielsen. 1999; Redpath and 21
Thirgood. 1999). The endlessly reworked fur-trappers' dataset for lynx and hare is now. or rather again, recognized as an example of the importance of predation (Keith. 1983;
Keith et al., 1984; Sinclair et al., 1988; Krebs et al., 1992; and see also Sinclair et al.,
1993). And the debate about population regulation of moose by wolf predation has been resolved in favor of predation as the key force by adding the simultaneous effects of bear predation (McLaren and Peterson, 1994; Messier. 1994; Van Ballenberghe and Ballard.
1994; Messier, 1995). While all of these conclusions have drawn on experimental study, they are also based strongly on in-depth empirical study of mechanism, and all are fundamentally ba.sed on long-term monitoring data.
CONNECTION BETWEEN ECOLOGY .\ND MONITORING
Society's interest in biodiversity and natural systems conservation leads to fundamental questions we still cannot answer, such as, "What are normal ecological rtuctuations?". "What are their most important causes'?". "What will happen if we cause this or that to change?". "What are the temporal and spatial scales we need to be concerned about?", and "How important are local adaptation and local processes, versus species-level adaptation and landscape processes?" The unresolved questions of ecology are now in the public sphere. "Ecosystem management" has become a mantra for public land management agencies in the United States, yet ecologists can scarcely link classical ecosystem dynamics theory to individual species conservation (Callicott et al., 1999). The
U.S. National Park Service's ecological monitoring effort, which is the most advanced of an expanding array of publicly funded monitoring efforts, is a response to increasing public demands for appropriate conservation and decision-making by resource managers.
This monitoring effort offers an opportunity to create empirical databases on population and community dynamics, based on the extensive public resources now, and increasingly, available. It is also an opportunity for ecologists to develop and apply theoretical and methodological advances in the design, guidance, and interpretation of the monitoring.
In this dissertation, I summarize and synthesize results of the first twelve years of ecological monitoring in the vertebrate ecosystem at Organ Pipe Cactus National
Monument (ORPI), Arizona. This will illustrate the breadth of data that may become available to ecologists through monitoring programs. I would argue, conversely, for the resource manager, that monitoring bereft of theoretical constructs and ecological logic will not capture an understanding of system dynamics, and will not permit the most accurate or timely interpretation of results. I hope this project will facilitate continuity of the ORPI monitoring, encourage the addition of new protocols that monitor important ecosystem dynamic links, and attract participation of academically-inclined ecologists who can work synergistically within the framework supplied by monitoring.
THE ORPI ECOSYSTEM MONITORING PROGRAM
This study began in August 1987, after 4 or more years of incubation, as the
Sensitive Ecosystems Program, and is carried forth now as the ORPI Ecological
Monitoring Program (EMP). My work is part of a collaborative study involving National
Park Service (NPS) resource managers and biotechnicians. and academic mammalogists, herpetologists, botanists, and climalologists. We have monitored population trends of 23
interacting trophic components of the vertebrate ecosystem in the Valley of the Ajo, on a
Sonoran Desert thomscrub-desertscrub ecotone. During the past 17 years, we have
worked toward an ecosystem-wide research program in the biotically rich subtropical
scrub of one of the best-protected and least-damaged ecosystems in North America. My dissertation is the first attempt to synthesize results from this program.
EMP activities are carried out by the ORPI Natural Resources Management division, with up to 9 stuff members working in part or full on the program. University.
U.S. Geological Survey, and regional NPS personnel constitute an advisory group to design and assist in carrying out the program. My goals vis a vis the advisory committee, on which I .serve, are (1) to support an ecosystem-, landscape-, and community ecology-
based paradigm as the appropriate structure for the EMP. (2) to support monitoring of additional taxonomic and trophic groups, and (3) to assist development of additional
monitoring protocols. Further. I hope (4) to demonstrate that the EMP has intrinsic
interest for the research ecologist.
My individual role as an observer and field worker focused firstly on study and monitoring of snake populations, which comprise a significant part of the vertebrate ecosystem's predator diversity. I also monitored lagomorphs, diurnal rodents, and the diverse assemblage of endothermic predators in the system. ORPI monitoring personnel
have provided long term, quantitative data on nocturnal rodents, diurnal lizards,
vegetation, and climate, using formal monitoring protocols developed by research
personnel (ORPI, 1995). I have also collected additional monitoring data on lizards and
nocturnal rodents, some of which is utilized here. 24
Briefly, tfie academic objectives of tiiis dissertation are:
(1) to identify key trophic components of the ORPI ecosystem that should
be monitored.
(2) to identify causative factors (climate, predation or other species
interactions) required to understand system behavior under natural
or anthropogenic changes.
(3) to establish a 10-year empirical baseline for major trophic groups in the
vertebrate ecosystem.
(4) to develop a conceptual view and working hypotheses for the ORPI
ecosystem, specifying or indicating the relative importance of
climate and species interactions in the system. 25
I. SMALL MAMMAL POPULATION TRENDS
CHAPTER SUMMARY
Small mammals, including lagomorphs. diurnal rodents, and nocturnal rodents, were studied at Organ Pipe Cactus National Monument (ORPI). Arizona, as part of a formal Ecological Monitoring Program (EMP). Several monitoring methods (live-trap grids using a simple capture index and a modified Lincoln-Petersen Method index, road transects, time-constrained search, drift-fence trapping) were used to produce time-series of abundance for each taxon. The objectives of this paper are to (1) evaluate the established EMP protocol for mammals, (2) present baseline time-series of rodent abundances during 1987-1998 for further analysis. (3) provide a working hypothesis for rodent assemblage structure and dynamics that integrates predation. competition, and climate, and (4) examine the potential usefulness of adding a predator monitoring protocol to the EMP. I focused on sites representing a well-defined bajada and valley floor rodent assemblage, in which I concurrently obtained time-series data for predators.
The simplest Sherman live-trap index method (the established EMP protocol for nocturnal rodent monitoring) provided the most information, but modified Lincoln-
Petersen Method calculations showed that the index should be re-calibrated to accurately represent population size. Lumping recapture data over 5 sites for Lincoln-Petersen calculations proved to be an efficient approach to population estimation. 26
Desiccating conditions during the late 1980's culminated in a 1989-1990 drought during which prey and predators declined to their lowest observed levels. A strong. timely summer rainy season, which occurred only twice during this study, broke the drought in 1990. and was followed by an exceptional string of El Nifio-caused wet winters, with attendant leaf and seed productivity. This produced a mid-l990's upswing in the abundance of all of the small mammal taxa being studied. Over the cour.se of this climatic phase, rodent species trajectories diverged, and the correlation with climate was apparently altered by biotic interactions.
The four predominant nocturnal rodent species showed a succession of population eruptions and collapses. The desert pocket mouse Chaetodipus penicillatus) peaked in
1992 and then declined, and the smaller Arizona pocket mouse Perognaihus ampins responded similarly. A larger heteromyid, Merriam's kangaroo rat Dipodomys merriami, peaked in 1994 and then declined rapidly; and in 1995 the white-throated packrat
Neotoma albigula peaked, with a subsequent rapid decline. The ground squirrels
Ammospennophiliis harrisii and Speniiophilus tereticaudus also had large increases during the mid-1990's, with late peaks like the packrat. Lagomorphs (Sylvilagus aiidiiboni, Lepiis califomiciis, L. alleni) increased rapidly, like the perognathines (pocket mice, Perognatliiis, Cfiaetodipus), but declined more slowly from 1993 to a low point in
1997, during another drought. Most ta.xa showed low abundance during the 1996-1997 drought, and most showed a tendency to increase again in late 1998 following two successive seasons of strong rainfall. 27
I describe the relationship of small mammal population flux to (1) changes in predation pressure. (2) life history, and (3) predator diets. I hypothesize that the successionai order of small mammal changes reflects the capacity for rapid increase and a countervailing ability to resist predation. It is possible that predation may directly affect rodent community structure and dynamics. A formal predator monitoring protocol would likely assist interpretation of rodent monitoring results from the EMP.
INTRODUCTION
Rodents are abundant, readily observable, trophically important members of the
Sonoran Desert ecosystem. They are important as resource exploiters, competitors, and prey. Although they are relatively well known ecologically, important questions remain about their relationship to climate and predators, and the range of competitive and indirect interactions they participate in. Small mammals, including lagomorphs. nocturnal rodents, and diurnal rodents, are competitively important as granviores (e.g.. Davidson et al.. 1980) and perhaps foliavores (Chew and Chew. 1970; McAuliffe. 1986; Ernest. 1994) in Southwestern desert ecosystems. They are also primary prey for predator species, including most of those found in the Sonoran Desert study region (see Chapters 3 & 4).
Their predators form a complex assemblage in which intra-trophic level predation is widespread (Chapter 3). so small mammals may also have strong indirect effects through their effects on populations of interacting predators.
Observation and monitoring of small mammals, especially rodents, has been practiced for many decades, and indeed, nocturnal rodent trapping is one of the most 28 efficient vertebrate sampling methods known (Wilson et al.. 1996). Special sampling schemes and computational methods have been devised for rodent populations (Anderson et al.. 1983). Study of lagomorph population density and dynamics has been less frequently attempted (Mohr. 1947; Wilson et al.. 1996). even though lagomorphs are primary prey for top predators in the desert, such as coyotes, great homed owls, and buteos (see Chapter 3). Recent advances in distance sampling theory (Buckland et al.,
1993) may hold promise for lagomorph population studies.
Ecological monitoring as a discipline is a relatively recent trend, although it is already widely practiced by public lands agencies (e.g.. Halvorson and Davis. 1996). Its stated goal is usually to "monitor ecosystem health", often with a goal of "providing guidance for management action". However, the U.S. National Park Service (NPS) policy on science and monitoring also includes as a scientific objective, use of parks as natural controls for contrast to more disturbed areas. The monitoring movement is strong on methodology (Thompson et al.. 1998). but less so in applying trophic-dynamic and theoretical concepts relevant to ecosystem function. There tends to be an artificial dichotomy between ecological monitoring and ecological research. I would argue that consideration of major interacting trophic groups, rather than a few indicator taxa, is crucial for interpretation of changes observed through monitoring. Such interpretation will determine what, if anything, should be done in response to monitored changes. An objective of this chapter is to illustrate the connections between the monitored rodents at
ORPI and adjacent trophic levels that affect their abundance.
At Organ Pipe Cactus National Monument (ORPI), 180 km west of Tucson, 29
Arizona, an ongoing ecological monitoring program has been established through collaboration of researchers and NPS management personnel. Several key trophic groups in the vertebrate ecosystem have been monitored, as of this writing, for 5-12 years.
Monitoring protocols (ORPI. 1995) have been developed and carefully adhered to, yielding a highly consistent dataset.
A key problem in a broad-based study like the ORPI Ecological Monitoring
Program (EMP) is feasibility and funding. Whereas rodent home ranges are small, predator home ranges are much larger and their population dynamics occur on a still larger scale. At over 130,000 hectares, and with habitats ranging from desert floors at 300 m elevation to rocky slopes at over 1450 m, the ORPI project must make difficult choices about where and how to monitor. The EMP is faced with the task of encompassing the landscape yet producing valid, repeatable. and meaningful results from small sampling areas.
We developed a gradient-based landscape model (Yang and Lowe, 1956;
Whittaker and Niering, 1965; see Chapter 2), with some variations (see McAuliffe, 1994), that encapsulates major aspects of landscape variation. East-west and north-south changes across the monument have been superimposed on this design, and particularly interesting unique sites, as well as purely representative ones, have been included. Key needs now are (I) evaluation, validation, and external review of existing protocols, and (2) identification, design and implementation of protocols for un-monitored trophic groups, particularly predators, invertebrates, and primary producers.
The mammal monitoring protocol (ORPI, 1995) is focused on nocturnal rodents. 30
It involves live-trapping for brief periods on numerous small grids, with methods and
analytical procedures derived from Petryszyn (1982) and Petryszyn and Russ (1996). It
has the advantage of consistency with an even longer record of monitoring initiated
during the International Biological Program (IBP) at Sonoran Desert validation sites near
Tucson (Petryszyn. 1982. and manuscript). It was developed before some recent
methodological advances in rodent population study, and requires evaluation, presented
here. EMP protocols for diurnal rodents, lagomorphs. and predators have not yet been
established, but my monitoring results for these groups will be discussed here.
In this paper. I compare the results of nocturnal rodent trapping grids to other
methods I used, including road transects, time-constrained searches, and rodent captures
in drift-fence traps set for reptiles. I also used these methods to track lagomorph and
diurnal rodent populations. The objectives of this paper are to (1) evaluate the EMP
protocol. (2) present time-series of rodent abundances. (3) describe the implications of the
empirically derived time-series in relation to climate, predation. and assemblage structure
and dynamics, and (4) evaluate the usefulness of adding a predator monitoring protocol to
the EMP.
METHODS
STUDY AREA
The study area is in the core of the northern Sonoran Desert (Shreve and Wiggins,
1964; D.E. Brown and Lowe. 1980; D.E. Brown. 1994) at Organ Pipe Cactus National
Monument (ORPI), in western Pima County. Arizona. It is on an ecotone between arid 31
thomscrub (the Arizona Upland subdivision of the Sonoran Desert) and true desertscrub
(the Lower Colorado Valley subdivision). Annual precipitation (Fig. I.IA) during 1988-
1998 averaged 257 mm; it was 237 mm in valley habitats and reached its maximum
average of 328 mm in the Ajo Mountains, which rise from about 730 m at the base of the
rockslope through steep canyons to 1466 m. On the intensive study area discussed here,
the precipitation average was 230 mm/yr. Average precipitation during 1988-1998 was
about 25 mm above the long term average (see Sellers and Hill. 1974). Monsoonal summer rains (see Adams and Comrie. 1997) were poor except in 1990 and 1998; in
1996, the summer rain quantity was also good, but the onset was extremely late. In contrast, wet winters associated with El Nifio events (see Andrade and Sellers, 1988)-
which yielded ephemeral wildflower blooms in springs of 1992. 1993. 1995. and 1998- occurred in 4 of 11 years (Fig. I.IB), compared to an average frequency (pre-1978) of
roughly 1 per 5 years in the historical record.
The Sonoran Desert landscape of this region can be viewed as a gradient of both slope and biotic composition. The gradient runs from steep rocky slopes, through gently sloping bajadas. to flats on the valley floor. Soil particle size grades from stones, rocks,
boulders, and outcrops on the rockslopes, through gradations of cobble, gravel, and sandy
loams on the bajadas, to very fine sandy, silty, and even clay loams on the flats.
Vegetation transitions (life form, flora, species) correspond closely to changes in slope and soil particle size at ORPI, as reported for the desert at Tucson (Yang and Lowe, 1956;
McAuliffe, 1994). Structurally complex paloverde-mixed cacti associations on the rockslopes and upper bajadas grade to simple creosotebush consociations on the valley Annual precipitation, all 21 available sites, ORPI
1 400
7i 300
SF 200
Annual precipitation, intensive study area, Valley of the Ajo, ORPI 600
^ 300 -
S5S5S9SSd°>
Fig. 1.1 A. Annual (calendar year) precipilalion during and in ihe climalic phase preceding this sludy. Seasonal precipitation, all 21 available sites, ORPI
I I I—I—I—1—I—I—I I ^^inintotDt^h^ooco o>M<»ai(oaia>inM » ^ S S
SEASON and YEAR
Seasonal precipitation, intensive study area, Valley of the Ajo 300
MBMaMcoMaoi/><»(/)a>(/i(nc/>a>(/)a>(/)
SEASON and YEAR
Fig. MB. Seasonal precipicalion record, 1982-1998. Summer |S|: May-Sepiember. Winter |W|: Oclober-April). 34 bottom. At ORPI the biotic aspect of slope gradient has a conspicuous inflection point at the rockslope-bajada interface, and a less well-defined transition between bajada and tlats. Superimposed on the slope gradient is a network of dense xeroriparian scrub
(thomscrub) along arroyos. which contrasts with the surrounding upland desertscrub.
Additional details about the study area are presented in Chapters 2-4. in Warren et al.
(1981). and in a regional treatment by D.E. Brown (1994).
LANDSCAPE STRUCTURE OF THE SMALL MAMMAL FAUNA
The largest structural change in the small mammal fauna occurs at the interface between rockslope and bajada, where dominance by murid rodents (particularly the white-throated packrat Neotoma albigula. and the cactus mouse Peromyscus eremiciis) replaces heteromyid rodent dominance (Petryszyn and Russ, 1996; Cockrum and
Petryszyn, 1986; see Table l.l). Kangaroo rats (Dipodoniys spp.) are largely absent from the rockslopes. Bailey's pocket mouse Chaetudipus baileyi and the rock pocket mouse C. intermedins are the most abundant heteromyids on rockslopes and upper bajadas, whereas the desert pocket mouse C. penicillaius and Arizona pocket mouse Perognathus ampins share dominance with Merriam's kangaroo rat D. merriami in valley habitats below the upper bajada (Petryszyn and Russ, 1996). On some valley floor and middle bajada areas, large kangaroo rats (D. spectabilis, D. deseni) are also prominent (personal observations;
Cockrum and Petryszyn, 1986). N. albigula is moderately abundant throughout middle bajadas and along drainages in lower bajada and valley floor habitat. The southern grasshopper mouse Onychomys torridiis showed increasing abundance moving down- 35
Table 1.1. Nocturnal rodent species composition by macrohabitat. based on Sherman live trap captures at Organ Pipe Cactus National Monument. Arizona. Blanks indicate that no individuals were captured. Data is from 1991-1998. Cotton rats were seen earlier in desert mountain habitat. HABITAT: middle- riparian rock upper lower valley TOTAL SPECIES: oasis slope bajada bajada floor C.APTURES
Peromyscits merriatm 0.05 1
Peromyscus eremicus 6.18 16.56 0.52 0.30 0.39 240
Cluietodipus baileyi 0.39 13.98 32.99 1.02 1.52 582
Chaeiodipus intermedins 11.41 11.03 0.46 0.05 263 —
Seotoma ulbi^iihi 30.12 39.14 4.06 4.82 2.09 726
Perognatlius ampins 0.77 0.08 11.03 19.50 6.58 111
Chaetodipus penicillatus 49.42 18.75 33.71 39.41 57.64 3846
Dipodomys merriami 11.97 0.08 6.56 33.82 30.06 2037
Onycliomys torridiis 1.16 O.IO 0.30 0.99 50
Dipodomys spectabilis 0.30 0.18 14 — — —
Sigmodon arizonae 0.51 22 — — — —
TOTAL CAPTURES 259 1280 961 1969 4348 8558 TRAP-NIGHTS 588 3130 2738 4502 9892 20850 CAPTURE RATE (no/trap-night) 0.44 0.41 0.35 0.44 0.44 0.41 36
gradient (Table l.l), although its actual abundance was probably strongly underestimated
by the predominantly warm-season trapping (J.H. Brown and Zeng, 1989; J.H. Brown
and Heske, 1990a). Other species of nocturnal rodents were not widespread or abundant
in the area.
Three diurnal rodents are present. The rock squirrel Spermupliilus variegciius
occurs in moderate abundance on rockslopes; Harris's ground squirrel
Ammuspennophiliis harrisii is abundant on rockslopes. hills, and bajadas. and the round-
tailed ground squirrel Spennophilus tereticaudiis is abundant on bajadas and valley
Hoors. Grounds squirrel abundances do not approach those seen in urban or suburban
Tucson or in sand dune habitat west of ORPl (pers. obs.). Of three lagomorphs. all
abundant, the desert cottontail Syivilcigiis aiuliibonii is found primarily in areas with
denser xeroriparian vegetation, while black-tailed jackrabbits Lepus califomiciis occur
widely and abundantly in upland de.sertscrub. The larger antelope jackrabbit L. alleni is
seen most frequently on lower bajadas and valley floor tlats in relatively productive
mesquite-creosotebush floodplains with open vegetation structure. .-X shrew. Notiosorex
crciwfordi, is rare and restricted to riparian and the richest xeroriparian sites (Cockrum
and Petryszyn, 1986).
The landscape boundary between the rodent assemblages of rockslope-upper
bajada versus middle bajada-valley floor extends out onto gravelly and stony portions of
the bajada that could be classified as the upper part of a broadly-defmed middle bajada.
Thus, Table 1.1 is organized with the bajada split roughly equally, to display the existing
faunal pattern as clearly as possible. The characteristic valley fauna, which is the focus of 37 this paper, is clearly defined in the table. However, rockslope and upper bajada facets are somewhat obfuscated because many grids were astride the rockslope-bajada interface, in some cases on hills small enough to permit commingling of major faunal groups. In core rock slope areas, D. merriami and P. onipliis are absent. C. penicillatus and often even C. intermediiis are uncommon, and C. bailey may be the only heteromyid present in substantial numbers (Petryszyn and Russ, 19%).
MO.NITORING METHODS
Sherman Trap Grids
The mammal monitoring system discussed here is the one used in the ORPI EMP
(ORPI, 1995; Petryszyn and Russ, 1996). and was first developed during the IBP
(Petryszyn. 1982). It consists of rodent trapping grids with 49 standard folding Shennan live-traps placed singly in 7 x 7 grids at I5-m intervals. Traps were set before twilight, baited with a pinch of rolled oats, and checked from first light through sunrise. Grids were censused during the warm season, and the computed values presented here are all from June-early September. The work was performed by research personnel during 1988
(Petryszyn and Russ. 1996). by me during 1989-1990, and by ORPI EMP biological technicians during 1991-1998.
Trapping was conducted at 20 sites, with 2 grids per site, although not all sites were trapped every year. The original work was conducted at 16 sites (Petryszyn and
Russ, 1996). During 1989-1990 I carried out the protocol at only two sites. Work during
1991-1998 included some of the original sites, plus 4 new ones, with seven "Core I" sites 38 worked every year. For this study. I used data from a single Core I study site that was censused in 1988-1998. and from four other Core I sites censused 1991-1998. All five sites were on the lower 2/3 of the bajada and the valley floor. This dataset allows consideration of a homogeneous fauna in habitat where other trophic groups were also monitored.
Grids were run for 2-5 days, once or twice per year. In two of the years at some grids we experimented with different trap spacing or multiple traps per station. The computed values presented here are based only on the first two days of trapping, only on the 7 X 7 grids, and only on grids that had one trap/station.
Each trapped rodent was identified to species (Hoffmeister. 1986; Petryszyn.
1995) and se.x, weighed on a Pesola ™ spring balance, marked, and released immediately, all on the spot. Marking was either by toe-clipping or by using a large permanent magic marker on the venter. There was no mark loss during the 2-day trapping periods. The
EMP protocol uses an index-based density estimate based on 19 yr of previous work
(Petryszyn, 1982, 1995 and personal communication). It is computed by dividing the number of individuals captured (not the number of captures) during 2 nights of trapping by the overall mean fraction of total rodents previously determined to be captured during the first two nights of extended trapping (0.72; Petryszyn. 1982, 1995). The grids were assumed to sample the trapped area plus an edge equal to half the home range radius
(thus, approximately the average recapture distance; see Dice. 1938; Tanaka. 1980;
Petryszyn. 1982) for all species (15 m, yielding an effective sampling area estimated at
1.44 ha). Biomass was computed for each species as (density) x (mean weight) of the 39
individuals captured during two nights of trapping. The computed result is quite close to
the actual number and weight of individuals captured.
I also calculated a population estimate and variance based on the modified
Lincoln-Petersen Method (L-P) Method (Chapman, 1951; Seber. 1970) as recommended
by Otis et al. (1978). Krebs (1989). Pollack et al. (1990). and Thompson et al. (1998). For
each species, the initial sample consisted of individuals marked during the first night of
trapping, and the census was based on the proportion of recaptures in the second night's
sample. Density and biomass were computed using the area consideration as above. 1
combined data from all grids within each year into a single computation of population
size and its variance for each species. These were adjusted to density by dividing through
by the total area effectively sampled.
Drift-fence Traps
Rodents were also captured using 22 drift-fence traps set to catch snakes on a 1
km' study area (ORPI Snake Study Area: OSSA) that included an EMP site. The drift-
fence traps and the OSSA site are described in detail in Chapter 4. Briefly, the drift fences
were 12-18 m lengths of hardware cloth leading to funnel traps that in turn led to mostly-
sunken buckets from which rodents could almost never escape. These permanent
installations were checked once or twice per day when they were in use. When not in use.
the entrances were sealed and exit ramps were installed in the buckets. The trapping
started with 5 traps in 1988, 8 in 1989. 14 in 1990-91. 22 during 1992-3. 17 during 1994-
5. and 10-12 during 1996-1998. Seven drift-fence traps were in upland desertscrub, and 40
15 were in or near xeroriparian desertscrub. Rodents captured in the traps were identified to species, and age class was recorded as "small juvenile", "juvenile or subadult", or
"adult". Habitat in this area is a fine sandy loam supporting a mosaic of (1) paloverde- ironwood-saguaro habitat that includes complex shrub and sub-shrub strata and (2) areas predominantly vegetated by creosotebush, white bursage, triangle-leaf bursage. and big galleta grass.
Capture rates varied moderately among traps and months. Annual means were therefore computed by adjusting each month's capture rate to eliminate month and individual trap effects, as also described for lizards and snakes in Chapters 2 and 4.
Means were based on these adjusted figures, and are expressed as captures per trap-day.
Variance terms (standard errors) for the trap capture rates were estimated as described in
Chapter 3.
Road Transect Method
Standardized road-driving methods were utilized to monitor snakes and small mammals on a sandy road in an intensively studied middle bajada-valley floor region of
ORPI. the "intensive study region".. This road was cruised at 9-14 mph (14.5-22.5 km/hr) at night, mostly in April-October. Further details are in Chapter 4. All small mammals observed on or between the roadbanks were recorded as LOR (live on road), and identified to the finest possible taxonomic class without leaving the automobile. Some were only recorded as "rodent", "heteromyid", or "lagomorph". but most were identified to species, or in the case of pocket mice, to subfamily ("perognathines") or genus 41
{Chaetodipiis, Perognathiis). Murid mice (Onycliomys, Peromysciis) were infrequently
seen, and not proportionately recorded to species: they were usually recorded only as
"rodent" in most cases because my search image was poorly attuned to them. Age class
(juvenile, adult [including subadults]) was recorded when possible. Each trip across the
full road transect was counted as one observation; partial trips were not utilized here.
Arithmetic means and standard errors were computed as normal statistics from this set of
observations.
Time-Constrained Searcli (TCS)
Much of my work at ORPI was directed at locating, capturing, and marking snakes. During all routine, desert herpetological work at ORPI. I recorded starting and stopping time, area worked, activity carried out, and all reptiles, amphibians, mammals, and raptorial birds visually observed, including those seen after first being detected by ear. Numerous rodent counts were made during nocturnal searches for snakes. The
procedure was to move through the habitat looking and listening for animals, using a standard headlamp powered by four 1.5 V D-cell batteries (6 V total). Snakes encountered were captured and processed on the spot, and this handling time was deducted from the search time for each TCS using a formula based on handling times for different species of snakes during different years. Travel was at approximate rates of 0.5-
2 km/hr. Identical headlamps were used throughout the study. My distance vision did not change measurably during this study.
TCS results for nocturnal rodents presented here pertain to two EMP sites within 42 the region of intensive ecosystem study. This area included the drift-fence trapping area, plus a .second site 4 km away with a second pair of Sherman live-trap grids. The included habitat was bajada upland and xeroriparian desert woodland (ironwood. foothill paloverde. saguaro). lower bajada-valley floor creosotebush-bursage (Ambrosia spp.) associations, and mesquite-creosotebush floodplain with dense herbaceous vegetation on the valley floor. Moving down the slope gradient in this intensive study region, soils graded from fine sandy loam with very sparse gravel to very fine sandy and silty loam.
No special efforts were ever made to find rodents or increase the counts of them.
When they were detected. I sought them out for identification purposes. Each rodent seen was identified as precisely as possible, generally to species. Tallies were enumerated by species and annotated at the end of each search period in an all-purpose field notebook.
Each TCS result was e.xpressed as observations/hr (for a single observer), and annualized standard errors were computed using normal statistics.
RESULTS
NOCTURNAL RODENTS
Dipodomys meiriami.
Live-trap grids for the OSSA site and for the 5 Core I valley sites (Figs. 1.2-1.5) showed a consistent temporal pattern. Dipodomys merriami had the highest average biomass of any rodent at valley sites (Tables 1.2-1.5), although both Neotoma albigula and Chaetodipiis penicillatus were important. D. merriami showed an increase during
1991-1994, with a high point in 1994, a sharp decline in 1995, followed by a 1996 43
Table 1.2A. Population density (no. / ha) of nocturnal rodents at the OSS A intensive monitoring site.
YEAR: mean Species 88 89 90 91 92 93 94 95 96 97 98 among yrs
Dipodomys merriami 8.0 13.5 2.0 9.5 17.0 16,0 28.0 6.0 13.5 8.5 4.0 11.5 Chaetodipus penicillcilus 16.0 9.0 11.5 13.0 27.5 19,5 11.0 15.0 5.5 13.0 10.0 13.7 Perognathus ampins 5.5 2.0 2.8 7.5 9.0 9.0 9.0 12.0 4.5 6.0 2.0 6.3 Neotoma albigula 2.0 3.0 0.8 4.0 0,0 2,0 2.5 6.5 5.0 2.0 2.0 2.7 subtotal 31.5 27.5 17.0 34.0 53.5 46,5 50.5 39.5 28,5 29.5 18.0 34.2
Dipodomys spuciabilis 0.0 0.0 0.0 0.0 0,0 0.0 0.0 0.0 0,0 0.5 0.0 0.0 Chaetodipus baileyi 0.5 0.0 0.3 0.0 0,0 0,5 1.0 0.0 0,0 0.0 0.0 0.2 Peromvscus eremicus 0.0 0.0 0.0 0.0 1,0 0.0 0.0 0.0 0,0 0,0 0.0 O.I total 32.0 27.5 17.3 34.0 54,5 47.0 51.5 39.5 28.5 29.5 18.0 34.5
Table 1.2B. Petryszyn Method biomass estimates (g/ha) forOSSA site: population density was multiplied by mean weight of the individuals captured. YEAR: mean Species 88 89 90 91 92 93 94 95 96 97 98 amone vrs
Dipodomys merriami 298 504 75 354 646 590 937 242 493 331 169 422 Chaetodipus penicillaius 242 136 174 185 405 314 173 216 74 204 175 209 Perognathus amplus 56 20 28 87 95 110 101 128 49 69 21 70 Neotoma albigula 307 404 101 565 0 216 385 824 577 150 237 342
subtotal 903 1064 377 1190 11147 1230 1597 1410 1194 754 602 1042
Dipodomys spectabilis 0 0 0 0 0 0 0 0 0 59 0 5 Chaetodipus baileyi 25 0 6 0 0 16 28 0 0 0 0 7 Peromvscus eremicus 0 0 0 0 20 0 0 0 0 0 0 2 total 928 1064 383 1190 1167 1246 1625 1410 1194 754 602 1051 Table 1.3A. Population density (no. / ha) of nocturnal rodents 5 Core I valley monitoring sites. using the Petryszyn Method index. Methods are described in text. YEAR: Species 91 92 93 94 95 96 97 98 overall
Dipodomys merriami 13.7 13.2 18.8 32.8 9.4 18.8 16.2 8.7 16.5 Chaetodipus penicillatus 35.6 54.2 40.6 19.2 24.1 12.3 18.9 20.2 28.2 Perognatfiia aniplus 6.0 6.3 7.3 6.2 5.9 3.5 3.4 2.6 5.2 Neoioma albigula 1.5 ... 0-6 1.5 _ 2.4 . Id.. 1.7 0.7 2.5 1.8 subtotal '"56"8""74.2" '"68"2' ~60.5' 417""ITF' 39.2 34.1 '"5T7"
Dipodomys speclabilis 0.0 0.0 0.0 0.0 0.0 0.0 0.1 0.0 0.0 Chaetodipus baileyi 1.0 0.0 0.4 0.6 0.0 0.4 0.3 l.I 0.5 Chaetodipiu intermedins 0.0 0.7 0.0 0.0 0,0 0.0 0,3 0.0 0.1 Onychomys toridus 0.0 0.1 0.4 0.6 I.O 0.3 0.6 0.6 0.4 Peromyscus erenucus 0.3 0.7 0.6 O.l 0.3 0.0 0.3 0.0 0.3 Sigmodon arizonae 0.3 0.1 1.0 0.0 0.0 0.0 0.0 0.5 0.2 Grand total 58.4 75.9 70.6 61.7 44.0 37.0 40.6 36.2 53.3
Table 1.3B. Petryszyn Method biomass estimates for 5 Core I valley desertscrub sites, ORPI; population density was multipied by mean weight of the individuals captured. YEAR; Species 91 92 93 94 95 96 97 98 overall
Dipodomys merriami 484 488 649 1136 348 653 583 335 594 Chaetodipus penicillatus 529 843 641 297 339 170 259 356 424 Perognathus ampins 76 73 82 68 65 36 35 26 57 Neoioma albigula 223 92 180 _ 377 430 212 47 339 234 subtotal "*1313"T497" iTsT""l879 1182 To?™ 924'" 1056 1309
Dipodomys spectabilis 0 0 0 0 0 0 10 0 0 Chaetodipus baileyi 24 0 10 14 0 10 7 27 11 Chaetodipus intermedins 0 8 0 0 0 0 3 0 1 Onychomys toridus 0 4 11 14 25 7 14 16 11 Peromyscus eremicus 5 13 10 3 5 0 5 0 5 Sigmodon arizonae 20 10 69 0 0 0 0 33 16 Grand total 1361 1531 1651 1910 1212 1089 964 1131 1355 45
Table 1.4A. Population density of the four predominant nocturnal rodent species at OSSA intensive monitonng site, based on two grids, two-night trapping, 1-2 times/summer, with moditled Lincoln-Petersen Index using the _Dice^a^ca^M_mediod_Thejwo_ends_were_combine^jnto^one_comgutation^n^^ YEAR: Species parameter 90 91 92 93 94 95 96 97 98 overall
Dipodomys No. / ha 2.0 21.2 34.3 28.0 51.6 S.5 16.2 9.6 4.0 18.2 inerrUtmi SE 0.0 7.9 9.3 6,5 10.1 2.1 1.9 0.0 1.6
Chaetodipus .No. / ha 16.3 15.3 32.6 24.0 12.3 20.2 17.5 17.9 12.3 18.3 penicillatus SE 2.3 1.7 2.6 2.5 1.1 3.1 10.4 3.2 0.9
Pero^natlttis No. / ha 3,2 8.5 10.5 14.5 12.5 21.6 8.5 7,2 3.0 9,8 aniptus SE 0.5 4,1 2.3 5.9 3.5 1.2 1.4 0,9
Seoloma No. / ha i,j 3,3 0,0 3.5 J.3 14.5 3.3 3.3 3.0 5,4 aibigulci SE 0,6 1,7 0.0 1.7 6.3 0.7 1.7 1.4 1.3 species lumped No. / ha 25.3 47.6 71.5 70.2 74.8 64.8 41,8 39.1 24.6 50.5 SE 3.3 5.5 5.9 8.0 7.9 9.3 5.7 4.3 5.3 "> I
species sum 22.8 50.4 77.5 70.0 79.9 64.8 47.7 38.2 22.3 516 pooled SE 2.4 8.4 9.7 8.3 10.6 9.5 11.1 4.0 3.4 ^
Table 1.4B. Biomass estimates for OSSA site based on Lincoln-Petersen procedure: population density statistics were multipied by mean weight of the captured individuals. YEAR: Species parameter 90 91 92 93 94 95 96 97 98 overall
Dipodomys g / ha 75 788 1305 1033 1729 342 593 374 169 685 merriami SE 0 296 353 241 339 85 71 41 0 62
Chaetodipus g/ha 246 217 481 386 192 291 235 281 216 276 penicillatus SE 35 24 38 41 17 45 140 50 49 14
Perognathus g / ha 33 98 111 177 141 231 93 83 109 amplus SE 6 12 14 50 29 64 38 14 15 10
Neotoma g/ha 168 776 0 379 539 1838 635 263 356 645 albigula SE 82 244 0 187 178 802 82 130 168 156
species sum 522 1880 1897 1974 2601 2701 1556 1001 772 1715 pooled SE 90 385 355 312 385 810 181 146 175 168 46
Table 1.5 A. Population density of the four predominant nocturnal rodent species at the 5 Core I valley monitoring sites, ORPI. Each site had Sherman live-trap grids (7 X 7, 15 m spacing) run for two nights in summer. Modified Lincoln-Petersen Inde.x and the Dice-Tanaka area method, were used to
YEAR: Species parameter 91 92 93 94 95 96 97 98 overall
Dipodomys No. / ha 17.8 19.2 23.0 33.4 11.9 18.7 13.7 8.7 17.8 merriami SE 2.7 3.3 T 7 T t 2.0 1.5 0.8 I.l 0.7
Chaeiodipus .Mo. / ha 39.4 58.7 43.5 20.9 25.9 17.1 20.7 20.0 31.0 penicillatus SE 2.9 3.4 2.8 2.0 T -I 2.9 2.1 1.7 0.9
Perognathus No. / ha 5.9 8.9 10.7 6.7 6.8 4.6 3.0 3.7 6.4 ampins SE 0.8 2.1 2.7 l.I 1.3 1.3 0.4 1.3 0.5
Neotoma No. / ha 1.6 0.4 1.5 2.9 4.2 1.4 1.1 3.4 t -> albigula SE 0.5 0.0 0.3 0.8 1.1 0.2 0.6 I 1 0.3 species lumped No. / ha 65.2 85.5 78.9 64.2 49.3 40.5 37.4 35.7 57.3 SE 4.1 4.7 4.5 3.4 3.6 3.0 1.9 2.6 1.3
species sum 64.7 87.1 78.7 63.8 48.8 41.7 38.6 35.9 57.3 pooled SE 4.1 5.2 4.8 3.3 3.5 3.5 2.3 2.7 1.3
Table 1.5B. Biomass estimates for the 5 Core I valley sites, ORPI, based on Lincoln-Petersen Inde.x calculation. Population tlcnsUO^imean^^i^hlof^captured individuals yielded estimated biomass. YEAR: Species parameter 91 92 93 94 95 96 97 98 overall
Dipodomys g/ha 628 710 796 1158 442 650 493 336 638 merriami SE 94 123 94 76 76 53 27 43 23
Chaetodipus g / ha 586 913 686 324 364 235 284 352 466 penicillatus SE 43 53 45 31 31 40 28 30 14
Perognathus g/ha 75 103 121 74 75 48 31 37 71 ampins SE 11 24 31 12 14 13 5 13 6
Neotoma g/ha 225 66 172 460 540 173 73 469 285 albigula SE 72 0 40 134 146 24 40 166 39
species sum 1514 1792 1775 2016 1421 1106 882 1194 1459 pooled SE 127 136 116 158 168 72 56 174 48 47
Live-trap Grid Results: ORPI Snalte Study Area Petryszyn Method population estimates
60 • Oipodomys merriami —o— Chaetodipus penicillatus —0— Peragnathus amplus • • c- • Neotoma albigula 50 Total /
40
30
20
10
0 00 eo eo CO at o> a> O) 0>
14
12 •••a-- Neotoma albigula —o— Peragnathus amplus 10
8
6
4
0 eo CM eo CO eo 01 01 O) 01 O) s> O)
1.4 -o— Chaetodipus baileyi 1.2 Peromyscus eremicus 1.0 0.8 0.6
0.4 0.2 0.0 eo eo eo 0> 0> at 8> oi
Fig. 1.2. Rodent abundance on the ORPI Snake Study Area monitoring site, computed according to the monitoring protocol handbook (see te.xt). Values are based on one or two trapping periods/yr during summer, for two grids. 48
Live Trap Grid Results: 5 ORPI Valiey Sites: Petryszyn lUetliod population estimates 80 • «- Dipodomys merriami —o- Chaetodipus penicillatus —®— Perognathus amplus • a-- Neotomaalbigula Total
91 92 93 94 95 96 97 98
10 .. a • • Neotoma albigula 8 —Perognathus amplus
6
4 ..a 2
0 91 92 93 94 95 96 97 98
1.5 Chaetodipus baileyi ' • Peromyscus eremicus Onychomys torridus •••»•• Sigmodon anzonae 1.2
<0 « 0.9 d ^ 0.6
91 92 93 94 95 96 97 98
Fig. 1.3. Rodent abundance on the 5 EMP study sites at ORPI in valley desertscrub habitat. Values were computed according to the monitoring protocol handbook (see text), based on one or two trapping periods/yr during summer, for two grids. 49
Lincoln-Petersen Method results for ORPI Snake Study Area site. DIpodomys merriami Chaetodlpus penlciUstus ± T SE
Perognathus amplus Neotoma alblgula 25 ti SE «1 SE 20
15 •
(8 •D3 10 • 5 • T I I i I - i L 6 9 0 6 92 93 94 95 9 96 97 98
Sum of species, pooled variances Fig. 1.4. Annual trends of population density at the ORPI Snake Study Area intensive ±1 SE intensive monitoring site for the four abundant nocturnal rodent species. Values are modified Lincoln-Peterson estimates based on two grids run once or twice per year during summer for two nights. Smoothed lines connecting points are intended to facilitate comparison among graphs, and do not imply that populations followed the trajectory shown between points. 50
Lincoln-Petersen Method results for 5 valley study sites, ORPI
Olpodomys merrlaml Chaetodlpus penicillatus 70 1 * 1 SE t1 SE 4) 60 - 50 -
40 - M <8 3 30 - •o > 20 - C 10 -
0 -1
Perognathus amplus Neotoma albiguia
£l SE t1 SE
Sum of species, pooled variances Fig. 1.5. Annual trends of population density at 100 5 Core I ORPI EMP monitoring sites in valley desertscrub habitat for the 4 abundant nocturnal rodents. Values are modified Lincoln- Petersen estimates based on two grids run once or twice per year during summer for two nights. > : Smoothed line connecting year points is intended I 20 - to facilitate comparison among graphs, and does 0 -1 not imply the trajectory shown between points. 51 rebound and a subsequent decline. L-P results for the OSSA site (Fig. 1.4) suggest a 1993 decline that is non-significant, a possible artifact of small sample size. Variances for Fig.
1.5 are small, and this appears to be the best single dataset for 1991-1998 for all species for which population estimates could be calculated. The Petryszyn Method appeared to underestimate population size by about 7-24 compared to the L-P method (Tables 1.2-
1.5).
Drift-fence traps were not very effective at capturing kangaroo rats (Table 1.6.
Fig. 1.6). although they gave results concordant with other methods (Figs. 1.7. 1.8). All methods agreed that the amplitude of the kangaroo rat population oscillation in the intensive study region was very large—with a factorial increase of 5 fold or more from low to high points. The factorial amplitude for this species at the 5 valley sites was around 3 fold (Figs. 1.2, 1.4). which is also in the range seen for the other abundant species, discussed below, regardless of what method or area is considered.
Road transect results for Merriam's kangaroo rat (Fig. 1.7) are concordant with other methods except that they show a decline in 1994. but with high associated variance.
There was limited road-driving during 1994. all on August nights when kangaroo rat activity was not high, perhaps due to high snake activity (Chapter 4). Other than this, kangaroo rats were effectively monitored by this method. The road transects showed a clear explosion of the Merriam's kangaroo rat population starting with numerous juveniles seen in late August of 1998.
TCS results for D. merriami (Fig. 1.8) also failed to capture the details of the
1994 peak due to the limited work that year. This method also missed the ongoing Table 1.6A. Number of drift fence trap captures of rodent species on the ORPI Snake Study Area monitoring site. YEAR: Species 88 89 90 91 92 93 94 95 96 97 98 Total
Dipodomys merriami 0 5 2 67 29 68 12 1 16 3 6 209 Dipodomys spectabilis 0 0 0 0 0 2 0 0 0 1 0 3 Chaetodipus penicillatus 24 no 307 231 301 263 6 0 86 52 283 1663 perognathine, undet. 0 0 15 29 65 59 0 36 7 5 3 219 Perognathiis ampins 2 16 38 36 37 41 0 0 10 18 82 280 Neotoma alhif>ula 1 4 19 12 21 28 2 4 5 7 9 112 Peromyscus eremicus 0 0 0 5 4 1 0 0 1 0 0 11 Sigmodon arizonae 0 0 0 4 5 1 0 0 0 0 0 10 Onychomys torridus 2 1 2 9 32 20 1 0 2 4 3 76
Total 29 136 383 393 494 483 21 41 127 90 386 2583
Table I.6B. Percentage of juveniles in1 the drift fence captures, ORPI Snake Study Area monitoring site. YEAR: Species 88 89 90 91 92 93 94 95 96 97 98 Total
Dipodomys merriami 20 50 24 34 51 58 44 0 50 38 Chaetodipus penicillatus 33 30 58 29 49 18 17 42 56 36 39 Perognathus amplus 0 0 45 14 41 49 40 28 43 37 Neotoma alhigula 0 0 58 42 33 39 50 0 0 29 11 35
L/i lO — all perognathines ±1SE O. merriami >• 0.25 < 1.0 < 0 o a. 0.20 1 0.8 < cc 0.6 0.15 CO UJ 0.4 lu 0.10 oc OC i. 3 t- 0.2 H 0.05 H O. a. < < .1 u 0.0 u 0.00 00 OI o CM co in CO N. CO OI —r— 00 CO 00 OI OI OI OI OI OI OI OI OI OI N. 00 OI 0 CM CO in CO CO OI OI OI OI OI OI OI OI OI OI OI OI OI OI CO CO CO 01 OI OI OI OI OI OI OI OI OI T- T- V- T~ OI OI OI OI OI OI OI OI OI OI OI OI 01
• • a • - ground s^guirrels -X- Neotoma 0.05 0.06
>- >• 0.05 < 0.04 o< aI cL a. < 0.04 < 0.03 oc oc 0.03 (A 0.02 Ui (0 •I oc UJ 0.02 oc 3 0.01 < o 0.01 aI- < 0.00 0.00 o 00 o> ^ CM CO ^ m CO CO OI 00 CO Oi OI OI OI O) OI Ol OI OI OI r*-ooo>o^CMeo««' 01 at Oi OI OI OI OI OI OI OI OI OI CO CO oo Ok OI a% OI at -0.01 Ok Ok Ok Ok Ok Ok Ok Ok
Fig. 1.6. Drift-fence trap capture rates for small mammals at ORPI Snake Study Area site. Data were adjusted to remove the effects of season and differing capture rates of traps, as described in test. Data for Onychomys wrrulus, and for the individual perognathine species, were not recorded during 199.5. tyi Night-time Road Transect Results, ORPI
Dipodomys merriami Dipodomys spectabiiis s 8 o S 7 ±1 SE o ±1 SE c c J % f 6 2 • o
t •• Q. o 1 i 'i. f z -• 0 00 o> O
Fig. 1.7. Nighl-lime road iransecl results for rodenls showing annual mean and standard enor among iransecl runs. Only individuals present on or between the banks of the unpaved transect road were included. population boom during late summer 1998.
Chaetodipus penicillatus.
The numbers and biomass of desert pocket mice for the OSSA site and for the
combined Core I valley sites showed similar patterns (Figs. 1.2, 1.3). Both sets of results
have a peak in 1992. and low population levels during 1994-1997, with only a modest
increase observed for 1998. An explosive increase during 1998 (see below) was not
observed on the grids because it occurred after the annual grid monitoring was conducted.
When the live-trap grid results were adjusted using the Lincoln-Petersen Method
(Figs. 1.4, 1.5), the results were almost identical to the Petryszyn Method, both in pattern
and in relative amplitude of the temporal oscillation. The drift-fence, road transect, and
TCS methods also gave similar amplitudes. However. Petryszyn Method population size
estimates were 3.3-9.0% lower than the L-P values (Tables 1.2-1.5).
Drift-fence trap results for C. penicillatus (Fig. 1.6). the most frequently captured
rodent (Table 1.6), showed a pattern identical to the live-trap grids, except that they
revealed the late-1998 population eruption. The drift-fence results for this species suggest
an oscillation amplitude similar to that shown by the live-trap grids.
Road transect results for pocket mice (Fig. 1.7) show a similar trend, but with
some notable differences. These results illustrate the deepest part of the decline during
1989, which occurred during late summer and early autumn as the drought intensified,
and they illustrate the rebound that occurred during the 1990 summer rains (see also
Table 1.6B). The road-cruising transects appear to have captured transitory but marked 56
- OSSA site •average ' floodplain site
Dipodomys mem'ami 14 12 10 ^ 8 - o 6-
4
2 • 0 -f- CO 9) in o> 00h. flO 00 0) 9)CO 9) 9) $ 9) 9)00 at 0) 0) 0) 0) 9) 0) 9) 9) 9) 9) 9) a»
Perognathine rodents 10 9 8 7 6 5 o z 4 3 2 1 0 -n (O CJ CO 9) 9)in r- 9)00 9) 00 9) 9) 9)9> 9) 9) 9> 9) 9> 9> r-
Neotoma albigufa
(O Si s0> sO) s;O) s CM
Fig. 1.8. Nocturnal time-constrained search results for abundant rodent ta.xa in the intensive study region, ORPI. 57 population changes that were missed by the grids, but which are consistent with drift- fence data broken down by month (not shown). The road transect also showed a high point during 1993 rather than the 1992 peak shown by other methods.
TCS results for pocket mice (Fig. 1.8) support a 1989 low point, as well as the
1987 to 1988 increase seen in the road transect results (Fig. 1.7). However, the 1998 explosion that was clear from the trap and road data was not reflected in the TCS data: apparently the late 1998 increase was weakest at the fioodplain site, where most late summer-early fall walking was done in 1998.
Perognathus ampins.
Live-trap grid results for P. ampins (Figs. 1.2-1.5) show a relatively lengthy period of elevated population size during the mid-l990's. A mode during 1993 is found in these results, but the L-P method for the OSSA site had a high point in 1995 not seen in other results. Otherwise. Petryszyn Method and L-P were similar, although L-P produced estimates 13-19% higher than the index method. For this species the L-P results suggest a greater oscillation amplitude than the Petryszyn Method.
Drift-fence traps showed a 1992 high point for this species (Fig. 1.6). As in C. penicillatus. P. ampins showed an explosive increase in the drift-fence traps during late
1998. 58
Neotoma albigula.
Live-trap grid results for M albigula at OSSA are based on small sample sizes, but show a population peak in 1995. All of the live-trap grid results show an increase from a 1992 low point to a distinctive peak in 1995 (Figs. 1.2-1.5). Fluctuating populations during the late 1980's and early 1990's are indicated, and the magnitude of these changes appears to involve roughly a 3-4 fold change (Figs. 1.2-1.8). A 1998 increase of packrats in the live-trap grid data (Figs. 1.2-1.5) was not evident in results from other methods. The L-P method gave population estimates 18-38 % higher than the
Petryszyn method, which is especially important considering the large size of this taxon and its strong contribution to total rodent biomass (Tables 1.2-1.5).
Drift-fence trap results for the packrat (Fig. 1.6) suggest a long-wave increase, but as for other species, data are sparse for 1994 and not plentiful for 1995. Sample sizes in the traps were small, and it often appeared that a single rat attempted to take over one of the drift-fence buckets and would appear repeatedly. Thus, the drift-fence results for this species are not very useful.
Road transect results for M albigula (Fig. 1.7) are also based on limited sample sizes, but are also roughly consistent with results showing an early 1990's decline.
TCS results (Fig. 1.8) are based on substantial sample sizes, and they confirm that packrats are abundant on valley floor and bajada sites even though they constitute only a small fraction of the total number of rodents seen in traps or on roads. Their mound nests
(middens) are conspicuous and abundant at these sites. The visual observation data confirm an elevated population for 1988. and also are consistent with the mid-1990's 59
increase.
Other Nocturnal Rodent Species.
There were insufficient recaptures for O. lorridiis, D. spectabilis. and Sigrnodon arizonae to perform L-P computations for any of them, or for Peromyscus eremiciis.
The Petryszyn Method showed a brief increase of P. eremiciis in valley sites during 1991-1993 (Figs. 1.2, 1.3; Tables 1.2, 1.3), which was also recorded by the drift- fence traps (Table 1.6). I also observed several of these woodmice at night in the intensive study region during 1991-1993, and one during 1987. confirming a small, widespread, population pulse in the desert valley at the on.set of the mid-1990's El Nino.
The Petryszyn Method and drift-fences (Figs. 1.1, 1.2, 1.5) also showed a population boom of Onychomys torridiis during the mid-1990's. This was a large increase, far above foregoing or ensuing population levels, as also detected by a large increase in audible grasshopper mouse cries. The calling levels during spring-summer
1993. though not quantified, suggested an order of magnitude increase compared to earlier years.
The drift-fence traps also showed an unexpected pulse of 5. arizonae at OSSA in
1991-1992 (Table 1.6), and Sigrnodon also showed an increase elsewhere at ORPI at this time (Figs. 1.2, 1.3; Tables 1.2, 1.3). It was unknown at or near ORPI until 1988 when it was found in the Ajo Mountains (Petryszyn and Russ, 1996), and was not expected in the arid valley habitat at all.
The large banner-tailed kangaroo rat D. spectabilis (D. deserti is reported for 60
ORPI but I did not see any large kangaroo rats that were clearly distinguishable from D.
spectabilis, and therefore refer all large-Dipodomys observations to D. spectabilis)
appeared to decline in the early 1990's and increase with D. merriami in the mid-1990's
(Fig. 1.7). However, data are too sparse to ascertain the similarity suggested for these two
kangaroo rats.
Pocket gophers Tliomomys bottae also appeared briefly in the intensive study
region during the height of rodent abundance during 1992-1993. although none were
observed during any of the standardized work.
DIURN.AL RODENTS
Drift-fence traps are the only method presented here for monitoring diurnal
rodents. For S. tereticciiidiis and A. hcirrisii combined, the trapping results indicate a long
wave increase during the mid-1990's that peaked during 1995 (Fig. 1.6). Visual sighting
records and daytime road-driving showed similar trends, but these results have not been
prepared yet. The drift-fence records were not very numerous, and when broken down by
species portray a more erratic picture that differs from that suggested by visual encounter and road-cruising (pers. obs.).
LAGOMORPHS
Road transect results for lagomorphs (Fig. 1.9) indicate a decline to low abundance during dry years beginning in 1989. followed by a strong upswing beginning
in 1991 or 1992, and a gradual decline through 1997. Reproduction and an incipient 61
Nocturnal road transects, lagomorphs
h-00a>OT-(MC0^u)(0N00 0000000)0)0)0)0)0>0)0)0> 0)0)0)0)0)0)0)0)0)00)0)
Nocturnal TCS, lagomorphs
average OSSA site floodplain site
0) 0.8
is.eoo>o^cMco^in(or*iOO c0c0w0>0)0>0>0>0>0>0>0) 0)0)0>0)0)0>0)0>0)0>0)0
Fig. 1.9 Combined road transect and time-constrained search results for three species of rabbits, ORPI. 62 rebound were apparent during summer 1998. The amplitude of the population oscillation indicated here involved a 6-fold or greater increase.
TCS results for the nocturnal period (Fig. 1.9) are consistent with the road tran.sect findings: they suggest an earlier high point (1992 vs. 1993). and fail to reflect a late-1998 rebound seen on the road. The steep 1994 decline in the TCS dataset is at least partially an artifact of lack of sampling in the mesquite-creosotebush fioodplain in 1994. where rabbits were often conspicuously abundant. These data are otherwise consistent with road transect and other observations indicating declining of rabbit populations in the late
1980's.
Breaking down the lagomorph data by species resulted in sample sizes too small to be reliable. However, visual encounters and the yet-unanalyzed daytime work suggest that all three species probably followed a similar pattern.
CORRELATION AMONG METHODS
I used the mean of Pearson product-moment and Spearman rank correlation coefficients to examine the consistency of results among methods (Table 1.7). Each method produced a few low values that were caused either by non-normality (producing unreliable Pearson correlations) or by sets of quantitatively close values (producing overly low Spearman correlations). Table 1.7 shows high correlations among methods for taxa where sample sizes were adequate (as detailed above). The best correspondence is between comparisons of different trap grids or trap grid computational methods, which is as expected from the spatial and temporal homogeneity in this sampling protocol. Table 1.7. Consistency index comparing different monitoring metluxlologies. The index is the mean of the Spearman rank correlation and Pearson product-moment correlation coefficients. All nocturnal Perogmithus Dipodomys Neotonia Chaelodipus method comparison Species ampins Lagomorphs merriami (ilbiffiila penicillalHS mean OSSA vs. 5 Valley Sites 0.95 0.85 0.95 0.85 0.92 0.91 Lincoln-Petersen vs. Petryszyn Methods 0.98 0.85 0.92 0.82 0.93 0.91 Drift Fence vs. Trap Grids 0.57 0.29 0.96 0.23 0.61 0.53 Drift Fence vs. Road & TCS 0.63 0.37 0.77 0.06 0.53 0.53 Road vs. Trap Grids 0.79 0.73 0.17 0.57 Road vs. Fence & TCS 0.72 0.72 0.73 -0.21 0.50 TCS vs. Trap Grids 0.92 0.28 0.85 0.45 0.85 0.71 TCS vs. Fence & Road 0.67 0.37 0.72 0.75 0.28 0.53 0.55 taxon mean 0.82 0.56 0.72 0.86 0.41 0.80 0.70
o\ OJ 64
Differences among methods most often reflect seasonal differences in sampling times.
TCS appeared to perform slightly better than other alternative methods, and all of the species might potentially be monitored efficiently by visual encounter.
DISCUSSION
METHODS EVALUATION AND RECOMMENDATIONS
Road-Cruising, Drift-Fence Trapping, and Visual Search
The divergent methods used for monitoring small mammals gave largely concordant results, within constraints of sample sizes and seasonal differences. This suggests that a variety of approaches could suffice for tracking temporal changes. An advantage of the drift-fence trapping, road transect, and time-constrained search is that they often require no extra effort beyond record-keeping during normal work activities.
They are valuable in detecting changes that occur between the annual trapping periods.
For example, while the grids showed little change between 1997 and 1998. despite the spring 1998 bloom, the alternate methods revealed that mammal populations contracted after the 1997 trapping, and expanded again just after the 1998 trapping. Without the temporally broadened view of the alternative methods, a key feature of rodent ecology- population response to strong rainfall—would have been statistically invisible.
I experimented with distance sampling of lagomorphs in the dry July of 1992, when during the heat of the day all jackrabbits were ensconced in predictable, shady micro-sites, and could be observed from a vehicle or on foot in useful numbers.
Convenient transect-based monitoring like this could be added to the EMP at little cost. 65
In as little as one work day per month, one person could perform useful road-cruising or on-foot transects to provide data on currently un-monitored elements of the vertebrate ecosystem. Survey of lagomorph and diurnal rodent abundance jumps to mind. The ease of observation of packrats on foot at night, and their conspicuous middens, suggests that area- or time-constrained search could be successful for them.
Road-cruising could be used to collect monitoring data for certain taxa with little cost, if it was conducted according to a protocol that recognizes its limitations.
Regardless of protocol, monitoring in transit is subject to distractions, schedule commitments, and a variety of external conditions (weather, traffic, and so on). The solution is simply for the protocol to mandate canceling data collection at such times. A
less obvious problem is that road conditions may change. Traffic may continue to
increase, as it has on Arizona State Route 85 at ORPI; populations may decrease along the road (which would be of interest); or conducting the work without personal hazard could be affected by increased traffic. In the latter case, a protocol calling for specific driving speeds and behaviors would still suffice to guard against bias: the protocol might have to be terminated, but until then useful data would be obtained. A final problem for road transect work is that the system hardware may change: highways may be widened, cars may become quieter, headlights may be improved, dirt roads may be widened or may become entrenched. I assume that reasonably standard vehicles and headlights can be maintained. The problem is to monitor stretches of road during periods of years when they are not subject to physical alterations. As long the existing EMP trap-grid protocol always anchors any alternative methods, other protocols could be re-calibrated as needed. 66
Live-Trapping Protocol
The Sherman live-trapping grids provided the best results in this study. Their
greatest benefits were (I) high consistency from observer to observer; (2) substantial
sample sizes for numerous taxa; (3) spatially intensive focus permitting mark-recapture
estimation, and (4) consistent application by NFS personnel. Their most serious
limitations were (I) lagomorphs cannot be sampled. (2) diurnal trapping for squirrels
would be costly, or subject the animals to severe thermal stress; (3) non-granivorous
species such as the grasshopper mouse are under-sampled; and (4) the trapping may be
too time- and labor-intensive to permit seasonal tracking. For simple indicator-species
monitoring, the limitations of this method are not very significant. From the standpoints
of understanding community dynamics or estimating true abundance and biomass values,
there are some limitations to be considered.
Use of mark-recapture methods appears to be justified. The simplest mark-
recapture procedure, the Petryszyn Method, proved efficient and sufficiently accurate for
documentafion of temporal trends. By counting individuals, rather than raw captures, it
reduces the bias inherent in trap saturation. Despite all of the obvious problems one might expect from such a simplified system, I was unable to find evidence that the factorial
range between high and low population was markedly compressed in the results, which is the primary problem expected from trap saturation effects. Nor was I able to find evidence that trap competition or differential trapability among species produced anomalous temporal patterns: the results were reasonably robust with respect to L-P calculations and other monitoring methods. 67
The only consistent problem with the Petryszyn Method results was an apparent underestimation of population size by about 15 % (3 -39 % in individual cases) compared to the results from modified Lincoln-Petersen Index computations. Although absolute population size measurement was not a required goal for the EMP, it would be a useful one. For example, if population densities increase appreciably, trap saturation must occur, and the index would have to be adjusted to account for it. Such adjustment would be simple enough, but a standard of comparison is required, and the best standard is actual population density.
The most probable reason for underestimation by the Petryszyn Method is that it is founded on a slightly different protocol, in which the grids were checked and rebaited during late evening, also run in the morning as done under the ORPI protocol, and pre- baited in some cases (Petryszyn. 1982). This would increase the capture rate, and hence the 72% index value for individuals captured during two nights is predictably too high for the ORPI protocol. Precise measurements of actual population density would allow re- calibration of the Petryszyn Method index for the exact protocol used at ORPI.
There are key benefits of the Petryszyn Method that support its retention as the
EMP protocol. First of all, it contains the results of 12 yr data at ORPI. as of this writing
(1999). and is compatible with a temporally adjacent two-decade record in comparable
Sonoran desertscrub near Tucson. It hence pertains to an unparalleled monitoring record.
It also stands on its merits: it allows monitoring of any species that can be captured in even modest numbers on even a few grids. Rarely-captured species in the ORPI valley setting, such as Peromysciis eremicus, Sigmodon arizonae, and Onycliomys torridus were 68 successfully monitored under this protocol. In contrast, few species could feasibly supply adequate sample sizes for modem mark-recapture computations on any EMP site. When even more stringent demands are made of the mark-recapture logic, data intensiveness soars (e.g.. Anderson et al.. 1983; Wilson and Anderson. 1985). spatially extensive sampling becomes impossible, and successful estimation would always be restricted to a few species at their highest abundances. Furthermore, it is of the first importance to recognize that the Petryszyn Method-the existing EMP protocol—can be carried out successfully even during times of waning funding and logistical support for the long term monitoring program. A reasonable approach is to continue the protocol as is. Results should continue to be presented using the protocol's computational scheme, which only reports biomass/ha. but individuals/ha should also be reported. After re-calibration of the protocol, recomputed results for all years can be presented in one of the ORPI Natural
Resources Management division annual reports.
There are uncertainties in the Petryszyn method that require validation with more advanced mark-recapture programs. (1) Species-specific recapture rates under the ORPI protocol need to be precisely quantified; the assumption that all species have the same probability of capture, while crudely correct, clearly cannot hold with precision. (2) True population size needs to be carefully evaluated using larger grids run for longer time periods. (3) The effective grid sample area (i.e., the "edge effect") needs to be quantified under varying climatic and population density conditions, and specified for all species being reported. Although some of this information will accumulate under the existing protocol, re-calibration work would be most useful. 69
There is strong interest within monitoring projects in applying the latest logic from recent developments in population estimation theory (see Thompson et al., 1998).
The most practical ways to achieve accuracy and precision are (I) high recapture rates
(i.e.. 90 % or more), which minimize error more strongly than increased sample size, and
(2) large grids, which minimize edge effects and permit confidence that movements outside the "normal" curve of individual activity are not a behavioral feature that might falsify the estimates.
I recommend that two protocol test sites be established at ORPI: one in a core rockslope rodent assemblage, and one in a diverse core valley assemblage. At these sites, protocol calibration and demographic data could be collected as a major enhancement to the EMP.
Each site should have a single 9 ha grid (21 x 21. l5-m trap interval) with one trap per station. These grids should be run for 3-5 nights during two (April-May. August-
September. as suggested by Petryszyn and Russ. 1996) or more annual sampling periods.
Permanent marks (toe clips or ear-tags) should be used, and reproductive data should be collected. The objective would be to obtain high recapture rates for estimation of population density, demographics, and behavior with respect to trapability. home range
(i.e.. edge effects), and movement (i.e.. immigration or emigration). The grids should otherwise be operated like the standard 7x7 grids, but body mass should be measured at every recapture. Analysis of these grid results would use both the Petryszyn Method and the robust capture-recapture design (see Thompson et al.. 1998). Results for these grids may be contrasted with MCAPTURE (White, 1994) results treating each 7x7 sub-grid 70 as an individual unit, and with results from the standard EMP grids.
Observed differences between the OSSA site and other valley sites may reflect localized or transitory events, as well as simply the limited sample sizes. For example,
TCS and road-cruising produced no evidence of the 1998 packrat increase suggested by the live-trap grid data, and it is probable that the increases of abundant rodents discovered by various methods in 1998 were inconsistent among methods because of spatial variation in the eruptions. The 1995 mode shown by the Arizona pocket mouse at one site
(Figs. 1.2, 1.4) may be another example of a local effect. It does not appear likely that robust conclusions will be available for single sites in one or two years using the existing protocol and analysis.
My analytical tactic of lumping data for a species from several grids in similar habitat proved very effective, and suggest that 4or 5 grid sites (actually, 8or 10 7x7 grids) within a reasonably homogeneous habitat type provides excellent information on abundant species. The requirement of 8-10 grids per habitat may be adjustable to 8-10 per abundant species if habitat gradients are cleverly worked into the scheme to permit detection of montane-valley and east-west differences that can be expected within the system. Such changes may be required anyway to meet the demands of the validation grids and other work. 71
CAUSES OF POPULATION CHANGES
Climatic Effects, and Literature Comparisons
Small mammal populations responded dramatically to drought and enhanced rainfall (Fig. 1.1; see statistical treatment in Chapters). The study began in August 1987 with moderately large populations and good summer rains that were, however, the first of a regional pattern of delayed-onset summer monsoonal rainy seasons during the study.
The 1989 to July 5. 1990 drought brought monitored minima for all taxa. Summer rains in 1990 initiated population expansion in every species except the packrat N. albigula, which didn't begin a sustained increase until 1992. Strong winter rains during 1991-2 and
1992-3 were associated with continued population growth, or at least sustained high population densities. However, further details of the population trajectories were divergent among taxa.
A positive association between precipitation and rodent population density has been ob.served in other studies that tracked rodent populations in the Southwest (Beatiey.
1969; Whitford, 1976; Petryszyn, 1982; J.H. Brown and Zeng. 1989; J.H. Brown and
Heske, 1990a; Heske et al.. 1994). In some respects, the studies show similar trends, but there are important differences as well, including in methodology. Other studies involved repeated trapping over the yearly cycle at localized study areas; the present study covered a single, major habitat type, at a landscape scale, in the center of a major biogeographic region, and trapping was once per year during a time of average, relatively high abundance for most species. Factors common to all these studies, in addition to (I) the generally apparent climatic response, include (2) evidence for widespread, generally 72
modest, but nonetheless eruptive population behavior following seasons of strong
precipitation, and (3) a delayed or reduced response by packrat populations.
Petryszyn (1982) studied a fairly similar, primarily Sonoran Desert fauna at the
northeastern desert margin near Tucson during 1972-1980. and continued monitoring for
more than a decade additional (pers. comm.). His results are most similar to the ones
reported here for ORPI. He observed several sequences, each showing eruptions of
heteromyid rodent species during and following El Nino-related winter-spring blooms.
Packrats responded to yearlong periods of elevated rainfall (Petryszyn. 1982; see also
Chapters). Petryszyn found that the smallest pocket mice responded first, with peaks for
P. ampins followed with little or no delay by C. penicillatus and C intennediiis, followed
with slight delay by the larger C. baileyi, with roughly a year spanning the perognathine
peaks. The D. merriami population responded least rapidly, although it reached peak only
6-12 months later than C. baileyi. .All of these species also showed population collapses after their peaks. During a second bloom series in the late 1970's (Petryszyn. 1982). a similar sequence occurred, although the heteromyid peaks were less pronounced, and the collapses less clear. The detailed similarities between Petryszyn's (1982) results and the
ORPI ones (Figs. 1.2-1.5) are remarkable. At ORPI in the 1990's, P. ampins may have
had a broader peak than those apparent in the I970's Tucson data, although considering all monitoring results together indicates that P. ampins and C. penicillatns at ORPI showed rather similar temporal sequences.
The succession of peak population densities in the 4 dominant species at ORPI
was as follows: the desert pocket mouse Chaetodipns penicillatus peaked in 1992, with 73
the smaller Arizona pocket mouse Perognaihiis ampins peaking at about the same time or
in 1993; Merriam's kangaroo rat Dipodomys merriami reached its high point during 1994;
and finally, the white-throated packrat Neotoma albigiila peaked in 1995. The assemblage
did not fundamentally change: species composition was stable, even though brief (and
unimpressive) appearances were made by the cactus mouse and cotton rat during the
generalized rodent explosions.
Whitford (1976) reported on a somewhat similar, Chihuahuan Desert rodent
assemblage at two sites in a varied landscape at the Jornada Site in south-central New
Mexico, 1970-1974. He observed eruptive population behavior of all species (except, or
delayed in, Neotoma spp.) after rainy seasons, apparently with substantial following
population collapses as well. In that study, new species appeared (including the non-
native house mouse Mux musculus) and assumed at least temporary importance during the
rodent blooms. Further, Whitford observed a second, more substantial wave of rainfall
and plant production that yielded a generalized eruption of formerly rare or absent murid
mice, apparently throughout the varied landscape he examined. It appeared that the
heteromyid population expansions were roughly synchronous.
The Portal study (see J.H. Brown and Zeng, 1989; J.H. Brown and Heske, 1990a;
Heske et al., 1994; and others) is the most intensive, longest running North American desert project reported in the literature. The 20-ha, experimental and observational study site is in the Chihuahuan Desert Grassland of southeastern Arizona, roughly 6 km from a
Madrean montane front. Species were similar to Whitford's (1976), and thus slightly different from the Sonoran Desert faunas. Numerically, however, at this site, kangaroo 74
rats (D. tnerriami, D. spectabilis, D. orclii) totally dominated the fauna in numbers and
biomass, and were shown to behaviorally suppress smaller granivorous rodents (Munger and Brown, 1981; Frye, 1983; J.H. Brown and Munger, 1985; Bowers et al., 1985; Heske et al., 1994). These experimental findings reflected behavioral responses (immigration,
habitat and patch selection), rather than changes in populatio size, by small species that
could move in or out of small fenced (50 m x 50 m) kangaroo rat e.xclusion plots through tiny portals. While a positive population response would seem probable if the experiment could be extended in spatial scale (Heske et al., 1994), the numerical response of small granivores did not approach compensation for the removed biomass of kangaroo rats.
Even with kangaroo rat exclusion, small granivores were far less abundant than in the
Sonoran Desert studies or Whitford's Chihuahuan Desert study—especially pocket mice,
which are quite diverse in the Chihuahuan biogeographic province.
The Portal study area is in an unstable desert-grassland system in the Chihuahuan
Desert Grassland, and has a history of past overgrazing effects (see Bahre, 1991). The site has revealed (1) major biome-level shifts, possibly in response to climatic flux-- anthropogenic or otherwise (J.H. Brown et al., 1997)-Ieading to a rapid change from arid grassland to diverse Chihuahuan desertscrub, (2) a decline of characteristic Chihuahuan
Desert Grassland rodents {D. spectabilis, Perognaihus flaviis: J.H. Brown and Heske,
1990a; Valone et al., 1995; J.H. Brown et al.. 1997). and (3) a tendency for kangaroo rat exclusion to reverse biome-level change from grassland to scrub (J.H. Brown and Heske,
1990b; Heske et al., 1993). The variety of population dynamic patterns suggested a
Gleasonian (species-specific, individualistic; Gleason, 1926) dynamic in this community. 75 despite extensive evidence for potent community-organizing forces (J.H. Brown and
Zeng, 1989; J.H. Brown and Heske, 1990a; Heske et al.. 1994).
The Portal site is on the periphery of the Chihuahuan province (regardless of whether the arid southern plains grassland is included as part of it or not). It is close to semi-desert and woodland slopes, and it is apparently under human impact that is producing marked biome-type conversion. The species responses fit into categories, two of which are approximately equivalent to those seen in the Sonoran Desert studies (this study; Petryszyn, 1982): (1) fast eruptive responses to strong precipitation (the heteromyids, the murid Reithrodoniomys niegalotis): (2) slow or limited positive responses to precipitation (Neototna albigula, some granivorous murid); (3) strong or massive eruptive responses of murid rodents to only exceptionally good, protracted rainfall sequences (especially Peromysciis spp.. in part the response of R. megaloiis. and
N. albigula: as also seen by Whitford [1976] for these taxa and the non-native MILS musculiisy, and (4) declines of Chihuahuan Desert grassland-associated species (D. spectabilis, P. flaviis) and increases of others characteristic of shrubbier habitats
{Chaetociipiis penicillatus, C. baileyi\ J.H. Brown et al., 1997), and possibly others (D. merriami; see J.H. Brown et al., 1997, Heske et al., 1994, and J.H. Brown and Heske,
1990a) that are geographically wide-ranging warm desert species. The range of responses and lack of stability are related, in part, to special features of the site. The authors are certainly correct that the local assemblages' behavior is rather Gleasonian, but the
Sonoran Desert results (Petryszyn, 1982, pers. comm.; this study) suggest that undisturbed rodent communities within the core of major biogeographic entities might be 76 more stably organized and follow more predictable dynamics.
The Portal study results may differ from the Sonoran Desert findings (Petryszyn.
1982; this study), and apparently from those for the Chihuahuan Desen system in New
Mexico (Whitford, 1976), in having less dramatic climatically-correlated eruptions and collapses of important heteromyid taxa (Dipodomys spp.. C penicillatus: J.H. Brown and
Heske. 1990a; J.H. Brown et al., 1997). Nevertheless, the broad positive effect of precipitation on the rodent assemblage and its biomass and component species was a strong conclusion of the Portal research (J.H. Brown and Zeng. 1989; J.H. Brown and
Heske. 1990a). and the qualitative differences among studies overall are not very large.
Life Historical and Inter-specific Competitive EtTects
Population eruptions seen in this study and in Petryszyn {1982), J.H. Brown and
Zeng (1989), and Whitford (1976) suggest unrestrained population growth, and thus that the innate capacity for increa.se might be involved in the speed of their responses.
Available life history data (Table 1.8) shows that, for the 4 predominant species in valley habitats at ORPI (C. penicillatus, P. amplus, D. merriami, N. albigula), the order of population peaks during 1992-1995 corresponds to reproductive potentials expressed as annual offspring production in Table 1.8. Conley et al. (1977) also ranked these species in the same order (on an "r-K continuum"). Earlier-peaking species tend to have earlier maturity, and their short gestation periods may indicate capability to produce litters in rapid succession during optimal times. The evolutionary interaction between mortality patterns (which may be morphologically- and behaviorally-determined) and the capacity Table 1.8. Life history parameters for small mammals found at OUPl. The vaUies are condensed from Moffmeister (1986), Smith and Jorgensen (1975), Conley ei al. (1977), Jones (1985), Brown and Zeng (1989), Asdcll (1964), Schmidiy (1977), Zeng and Brown (1987), Wasser and Jones (1991), Sowls (1957), Cockrum (1982), Burl and Grossenheider (1975), and Nowak (1991). A question mark (7) indicates doubt about accuracy or precision of tabled value; values in parentheses are from species closely related to the one in the table. Lifespan is in yr for wild animals; values in mo are for captives. The estimate for mean number of litters per year is my estimate bused on literature values and descriptions.
Age at No. No. Mean No. Computed Peak Maturity Young/ Litters Litters No. Young Reproductive Reproductive Gestation Taxon (mo) Liner per Yr per Yr per Yr Season Season Lifespan Period (d)
Lcpus allc'tii 10-11 7 2.0 3-4 3.0 6 continuous Dec. - Sept. 7 (43) Lepus califoniicus 4- 11 2.8 1 -4 2.5 7 continuous Dec. - Sept. 7 84-f mo 43 Sylvilagus auduimiii 3-9 3.2 2-5 3.2 10 continuous Jan. - Aug. 7 2+ 26-30 AmmosperniopUilus luirrisii 3-11 6.8 1 -2 1.2 8 Dec. - June Feb. - Mar. 21-24 Spermopliilux lereticatulux 4- 12 6.3 1 -2 1.4 9 Jan. - July Mar. - Apr. 23-35 Thomomys boitae 6- 12 4.8 1 - 5 2.6 12 continuous Apr.,, Nov. 4 19 Neoloma albigula 5-9 1.9 2-^ 2.3 4 continuous Apr. - Sepi 4; 92 mo (30 ) 38 Onydiomys torridus 2 3.2 2-6 3.5 11 Jan. • Sept. Feb. -Sept. 3; 55 mo 26-35 Peromyxcus eremicus 17-6 3.0 2-4+ 3.0 9 continuous Jan. - Sept. 7 1.5 21 (-28) Sigmodon arizonae (1 -2) 6.5 7 (6 - 9 7) 4.5 29 continuous June - Nov. (^1) 19 (-27) Dipodomys deserti 37 3.2 2 1.7 5 Dec. - June (Jan. - Apr.) 29-31 Dipodomys spec tabilix 3- 10 2.1 1 -4 2.0 4 Dec. - June Jan. - Apr. 7; 66 mo 27(+7) Dipodomyx merriami 2 - 6 2.3 1 - 3 1.7 4 Feb. - Oct. Apr-May , Aug-Scpt 4; 66 nio (27-30) Chaetodipux baileyi (2-47) 3.9 1 (7) 1.5 6 Jan. - July Mar. -July (22-26) Chaeiodipux intermediux (2-47) 4.2 1 (7) 1.7 7 1-eb . July Apr. -July (22-26) Chaetodipux penicillaiux 3-47 3.9 2 - 3 2.5 10 l-eb. - Sept. Apr. - July 2 23-26 Pe rognath ux amplux (2-47) 3.6 1 - 2 1.5 5 Mar. - June Mar. - June 27;(95 mo) (22-23)
-J 78
for increase during flush times, may involve bi-directional compromises, or tradeoffs.
More proximately, it appears that species' characteristic life history patterns may
influence their tendencies to benefit from, or possibly even preemptively gain competitive
advantage at. certain stages of the generalized rodent eruptions that occur in the Sonoran
Desert.
There are clear lines of argument and evidence for inter-specific competition in
desert rodents. Competition hypotheses vary, and may be grouped as follows: (I)
exploitative food competition (J.H. Brown. 1973; J.H. Brown and Lieberman, 1973;
Reichman and Oberstein. 1977; Munger and Brown. 1981; J.H. Brown and Munger.
1985; Thomp.son and Fox, 1993; Fox and Brown. 1993). (2) habitat partitioning
(Schroder and Ro.senzweig. 1975; Price. 1978a; Stamp and Ohmart. 1978; Wondolleck.
1978). (3) competition and microhabitat partitioning under predation risk (Thompson.
1982a&b; Harris. 1984) (4) interference food competition (Frye. 1983; Mitchell et al..
1990; Heske et al., 1994). (5) seasonal partitioning ba.sed on foraging efficiency (J.S.
Brown, 1989), and (6) behaviorial niche partitioning based on modes of predator
avoidance in a competitive context (Rosenzweig and Winakur. 1969; Kotler, 1984; J.S.
Brown etal., 1992; Kotler etal., 1988. 1991; Norrdahl and Korpimaki, 1993). A number
of studies combine mechanisms to create hypotheses for community structure (see review
by Kotler and Brown, 1988). Evidence of intra- and inter-specific aggression.
territoriality, and interference is abundant, and interference competition appears to be the
most important competitive effect. The literature supports an expectation that inter specific competition among rodents is likely to be significant at ORPI. 79
The sequence of species population peaks at ORPI during 1992-1995 may reflect a successional series involving replacement of species that have high capacity for increase by species with advantages under intensified species interactions. The Portal results support a hypothesis that D. merriami may competitively suppress pocket mice, once its population reaches a high level. This may partially explain why its population peak was accompanied by a decline in the perognathines. as tlrst described by Petryszyn
(1982). It is plausible that competition from pocket mice could contribute to delayed increase in the kangaroo rat population, although further analysis did not support this
(Chapter 5). Regardless, it seems unlikely that competition between the granivorous heteromyids and the herbivorous N. cilbigulti would be strong. The slower increase of N.
(ilbigiila may partially reflect its low litter size, delayed maturity, and long gestation
(Table 1.8), and perhaps its need for sustained periods of perennial plant productivity
(Petryszyn. 1982). It is therefore unlikely that competition is the only interaction involved in the species sequence at ORPI.
Predation and Assemblage Dynamics
Population dynamics of the predominant nocturnal rodents studied at ORPI had the following important features; (1) rapid population expansions (eruptions), (2) a replacement sequence of taxon-specific peak populations, and (3) rapid and dramatic species population declines after their peak years (collapses). The rapidity of the declines suggests the action of predation. since evidence for debilitating disease or parasitism is absent. Could the pattern of predation contribute to the sequence of peak rodent 80
abundances?
Although competition was the dominant paradigm during the early development of desert rodent community ecology, predation has become a persistent theme (Kotler.
1984. 1985. Kotler etal.. 1988. 1991. 1992. 1993a&b; J.S. Brown etal.. 1988; Korpimiiki and Norrdahl. 1991; Longland and Price. 1991; Pierce et al.. 1992; Daly et al.. 1992;
Diaz. 1992; J^drzejewski et al., 1993; Meserve et al.. 1993. 1996; Lagos et al.. 1995. and others). Niche partitioning according to open vs. vegetation is a key feature of
Southwestern rodent assemblages (e.g., Rosenzweig and Winakur. 1969; Rosenzweig.
1973; Lemen and Rosenzweig. 1978; Price and Brown. 1983; Kotler and Brown. 1988).
Different locomotory modalities (particularly bipedal [Dipodomys, Microdipodop.s,
Jciciilus, Allciciolciga\ vs. quadrupedal [perognathines. Peromyscus, Gerbillus, and others]), often associated with enhanced hearing specializations, may confer differing predator detection and escape competencies that permit some species to exploit the environment with greater freedom.
Experimental attempts to demonstrate the details of this hypothesis have had some
limited success. Longland and Price (1991) showed that, in a semi-naturai cage, bipedal
heteromyids were better than quadrupedal heteromyids or Peromyscus maniculatus at escaping great homed owl attacks: yet they were so much more likely to draw attack,
particularly during their much more extensive utilization of open vs. bush microhabitat,
that they had higher total risk of predation. Pierce et al. (1992) found that in small arenas
with rattlesnakes, kangaroo rats (D. merriami) and deer mice {Peromyscus maniculatus)
were far superior to smaller heteromyids (bipedal Microdipodops megacephalus and quadrupedal Perognathiis fallax) in survival rates; but the results were confounded by geographic distributional patterns suggesting the less competent species were naive rather than biomechanically limited. Kotler et al. (1988) showed that predation by large owls
(great horned and bam owls) fell with disproportionate severity on larger compared to smaller heteromyids, yet they could not show a difference attributable to gait. Kotler
(1985. 1988) showed that long-eared owls (another large owl) showed selectivity for capturing quadrupedal over bipedal heteromyids when the bipedal species were most abundant, but when dominance of the quadrupedal set shifted from Perogncithus lungimembris to the murid Peromysciis maniailatus, selectivity shifted to the bipedal category. The reason was unclear. These detailed observations provide limited confirmation that bipedal species, which are known to use the owl-exposed microhabitat most extensively, are protected from owl predation. Perhaps the naivete of the rodents in their experimental surroundings, the high owl densities used (2 owls confined in a tiny fraction of a hectare), or other experimental details affected some of the experimental outcomes. The morphology-behavior-predation hypothesis remains a strong one. but requires more rigorous support.
Kotler et al. (1991; I993a&b; and therein) showed that Negev Desert Gerbilliis species avoided open ground under owl predation risk, but avoided shrubs with snake predation risk (as also in the North American desert [Bouskila. 1995]). Thus, owls and snakes might facilitate each other, as well as partition food resources and compete exploitatively over them. I might add that owls also eat the snakes, implying that 4 (or more) trophic interactions may be characteristic of just this single taxonomically-defmed 82 interaction involving species on the "same" trophic level. Moreover, if we add ontogeny and species diversity to the picture, we also find that snakes may eat nestling raptors as large as red-tailed hawks (C. Schwalbe, P. Rosen, pers. obs.). and large gopher snakes
(which are prey, as juveniles and even adults, for many raptors) may eat owls at least as large as screech-owls (Bent. 1938). To return to the point here, rodent foraging behavior is strongly affected by behavioral responses to predation risk. Predation may thus help determine the nature and outcome of inter-specific competition.
Endothermic predators showed a 3-4 fold increase during the mid-1990's, from a
1989-1990 low point to a 1994-1995 high point (see Chapter 3). Smaller predators that feed disproportionately on the smaller rodents apparently increased faster than larger predators (Chapter 3). The dominant rodent-eating snake, the western diamondback rattlesnake Crotcilus atrox increased by 1994-1995, and rattlesnake birth rates increased rapidly from a low in 1990-1991 to maximal levels in 1992 and 1993 (Chapter 4). The abundance of young rattlesnakes is especially significant, for although adult rattlesnakes have remarkably low feeding rate requirements (Beaupre, 1993; Beck, 1995; Secor, 1995;
Secor and Nagy, 1994), the rapidly-growing young do not, and they feed primarily on perognathine rodents. The taxon-specific predation indices (Chapter 5), which incorporated the major points recounted here, showed different peaks in predation pressure that correspond to the different rodents population peaks.
Thus, the increases, succession, and declines of rodent species in 1990-1996 occurred while the total intensity and species-specific relative intensities of predation varied. If we assume that pocket mice are the most susceptible to predation, their early 83 population collapse is consistent with a hypothesis that, while they may erupt most rapidly because of their capacity for Increase, they should also be the first to decline under intensifying predator pressure. Pocket mice are generally short-lived and early- maturing. particularly C. penicillaiiis (Table 1.8). and they seem relatively easy to catch.
Further, data suggesting a slightly later peak in P. ampins, which is congeneric with other silky pocket mice (genus Perognatlius) that are remarkable for captive lifespan (see. e.g..
Hoffmeister. 1986) and surprising survivorship under field conditions (French et al..
1967; Kenagy and Bartholomew. 1985). may also be consistent with this predation hypothesis.
Kangaroo rat.s are generally thought to be more capable of avoiding predation than quadrupedal rodents because they are faster runners, can u.se a quick bounce to evade a predatory strike, and have enlarged auditory structures for predator detection. They are probably longer-lived than the smaller, quadrupedal heteromyids. Under the hypothesis of a generalized, non-specific increase of predation pressure, they could continue to increase even as perognathine populations coIlap.se. as was observed at ORPI.
Finally, the generalized-increase predation hypothesis demands that the packrat N. albigula be the most predation-resistent species in the nocturnal assemblage. The available evidence supports this. Table 1.8 suggests it is among the most-long-lived
North American desert rodent, with relatively late maturity and low litter size. They nest within spine (cholla)-laden fortress houses of their own (culturally speaking) construction, and move about cautiously along trails also guarded by spiny cactus joints.
Although wild (R. Henry, personal communication) and domestic (personal observations) 84
felicls are efficient predators of packrats. and coyotes do occasionally excavate even the
cholla-laden middens, packrats nonetheless seem to be the best-defended rodent on the
desert floor. During night-time work, they evince some semblance of confidence as they
watch the field worker from their fortresses. Further, packrats are resistent to western
diamondback rattlesnake venom (Perez et al.. 1978; Beck, 1991): indeed, I have twice
observed subadult N. albigida fully recover from feeding strikes in the lab from large
adult western diamondbacks. Kotler et al. (1988) noted that the packrat Neotomci lepida
was disfavored as prey by bam owls that were fed live rodents in enclosures.
Predation and Community Structure
Here, I consider whether community structure, in addition to population dynamics, may be affected by predation patterns. Microhabitat partitioning and differential seed patch utilization (Kotler, 1984; Kotler et al., 1994; J.S. Brown et al.,
1994a) have been identified as key features of Southwestern desert heteromyid community structure. As discussed above, rodent gait and hearing capability, important in
relation to foraging ecology and economy (Rosenzweig and Winakur, 1969; Rosenzweig and Sterner. 1970; Reichman and Oberstein. 1977; Price. 1978b; Mitchell et al.. 1990;
Ziv et al., 1995), may reflect predation-resistance. The tradeoff between the high foraging efficiency of the small, exploitatively superior pocket mice (see Rosenzweig and Sterner.
1970; Kotler et al., 1994). which are much-restricted to the relative safety of vegetation cover, versus the less efficient but more predation-resistent foraging of kangaroo rats,
which can forage on high quality food patches in open habitats too risky for smaller taxa. 85 has been proposed as the explanation for this key feature of assemblage structure.
Bipedal rodents have long legs, which confer speed (see Kotler et al.. 1994;
Djawdan and Garland, 1988) and possibly evasive dodging ability (and necessarily much shorter arms to permit power in digging), and their enhanced hearing further suits them to exploit open spaces with high predation risk. Speed is associated first of all with large size (Djawdan and Garland, 1988), so open ground exploiters (be they kangaroo rats or jerboas) have greater food requirements, and hence cannot efficiently utilize low quality patches that smaller species can use. The possibility that nimbleness of quadrupedal. often smaller, taxa in dense low scrub, on trunks, and on and among rocks has also been suggested (Kotler. 1984; Kotler et al.. 1994). and although supporting data are not yet strong, the biomechanics seem clear enough.
Taxa that are very similar, such as the intensively-studied pair of gerbils Gerbilliis allenbyi and G. pyrimidiim in the Negev Desert, Israel (see Rosenzweig and Abramsky,
1997), are observed to tradeoff between successful interference of the larger species versus exploitation of lower-quality food patches by the higher-efficiency smaller species
(J.S. Brown et al., 1994b). Even in this best-analyzed case of closely similar congeners, recent evidence showing a strong modulation of the competitive interaction by predation risk (Abramsky et al., 1998) suggests that predation has important effects that are not yet well known. Such behavioral effects may be key factors permitting coexistence and producing structure within narrowly-defined guilds (or subguilds) of desert rodents.
At ORPI, these interactions of predation, competition, and morphology appear to explain why bipedal species are absent from the rocks: they cannot stand the intensity of 86 predation entrained by taxa that are more nimble on the complex footing, notably the murids Peromyscus eremicus and Neoioma albigula. On somewhat less rocky substrata
Chaetodipiis baileyi and C. iniennedius may retain a similar advantage over Dipodomys.
In valleys, speed on flat ground is at a premium, and murid biomass is effectively replaced by the heteromyids, especially D. merriami, with Neotoma persisting at much- reduced abundance in its self-engineered predator-free space. In the heteromyid group.
Dipodomys merriami persists because of its superior ability to exploit open spaces between shrubs, which tend to have more high-value seed clumps because they are too exposed to be heavily foraged by Chcieiodipus or Perogiicitlius. These perognathines persist because they are small, and hence need less energy, and thus can profitably forage on already-reduced seed patches that D. merriami ignores as unprofitable. Because the perognathines cannot afford extensive open-ground foraging, their activities are concentrated under shrub cover, and this increases the resource depression further under shrubs, accentuating microhabitat partitioning with D. merriami. To this I would add the likelihood that perognathines enjoy an anti-predation advantage in their selected habitat over the larger bipedal species.
This line of reasoning may explain the protracted collapse of D. spectabilis at the
Portal site, which is similarly seen in other shrub-invaded regions of southern Arizona (R.
Frye, pers. comm.).
In this literature-based scenario, predation risk determines behavioral habitat partitioning, but predation does not determine abundance directly. The question about whether small size and quadrupedalism confers a microhabitat-specific advantage over 87
larger bipedal species is important because it determines whether a simple predator-
avoidance niche axis can be partitioned to determine major aspects of community
structure via the direct population dynamic impacts of predation. The ORPI monitoring
results suggest that predation may be potent enough to regulate rodent abundances, and
this supports consideration of relatively direct, or simple, hypotheses of partitioning of
predator-free space. Somewhat analogously, just as smaller species should be more agile against predation in cover, they may have their best refugia against aggressive
interference by larger species in association with certain plants.
An important stnictural feature of desert rodent communities (although see J.H.
Brown and Kurzius, 1987; and Kelt et al.. 1996) is the regular size ratio among most- similar species (J.H. Brown. 1973: J.H. Brown and Lieberman. 1973; J.H. Brown. 1975;
Davidson et al., 1980; Fox and Brown. 1993). Size ratio consistency has been ascribed to a number of things associated with seed size and foraging economics, but could also reflect factors related to predation and life history. Rodents that are small are likely to be early-maturing and have high mortality (Table 1.8), and be preyed upon by smaller taxa of predators (Chapter 3. 4) that are also likely to be earlier-maturing, shorter-lived, and thus should also increase more quickly than larger taxa of predators. Other things being equal, the road to maturing early is maturing small, and hence the logic of life history constraints could participate in producing the observed size ratio structure. This would imply partitioning of a survivorship niche axis set up by temporally fluctuating predation. which is exactly what I have described for the desert valley community at ORPI. Of course, by differing in size, rodents could also avoid predation pressure from species 88 using primarily prey of a differing size, and this might also act to produce size differences among similar species "competing" for a position in niche space at which they are superior at avoiding predation. The general concept of competition (or apparent competition) for enemy-free space (Holt, 1977. 1984; Jeffries and Lawton, 1984, 1985) should apply here.
Summary of Predation Effects
For the nocturnal rodent community in Sonoran Desert valley habitat, a sequence of mildly eruptive species peaks, triggered by high rainfall, and followed by abrupt declines, appears to be controlled (at least in part) by the interaction of rapid reproductive increases and increasing predation pressure. Both the amount and nature of predator pressure appeared to vary over the timespan of the rodent sequence. Small rodents dominated at first; larger rodents, which have larger predators and may better withstand higher overall predation pressure, dominated later. A similar importance of differential susceptibility to predation in the determination of species-specific population dynamic differences has also recently been identified in Palearctic microtine rodents (Hanski and
Henttonen. 1996; Saitoh et al., 2000). This community dynamic process involving predation may contribute to community structure (I) directly, by determining which species persist and in what abundances. (2) less directly, by creating a partitionable niche axis of predator-free space, and (3) less directly still, through risk-avoiding behavioral decisions that may determine the outcome of competitive interactions.
The literature draws a distinction between predation as a "limiting" factor on 89 population size (one which contributes to the value achieved) versus a "regulating" factor
(one which acts in a density dependant way to produce a tendency toward stability; see
Sinclair, 1989; Begon et al., 1996). However, in practice regulation may be defined by its failure: the restricted definition of regulation seems to be asking what controls (and dampens) the fluctuations we do see. rather than those we don't. If the fluctuations we are trying to explain, especially eruptive ones, are those that reflect an escape from predation. concluding that predation does not regulate populations is circular. I find the older terminology of density dependent versus density independent processes more to the point, and 1 have used "population regulation" in a broad sense, to mean "determination of abundance".
To return to the Sonoran Desert. I would tentatively suggest that predation may have functioned as a density dependant regulating factor. Rodents declined under high predation pressure despite renewed El Nino activity. Whether the predator increase was affected by increased rodent density, rather than solely by a parallel but slower direct effect of climate, cannot be conclusively stated here, but a response to prey density is most plausible. 90
II. POPULATION MONITORING OF LIZARDS
CHAPTER SUMMARY
Population trends, climate, and predators were monitored over 12 years in a
Sonoran Desert lizard assemblage (16 species) at Organ Pipe Cactus National Monument,
Arizona. Lizards were monitored with line transect, trapping, and visual encounter
methods. These methods gave comparable results; the transect method was time-efficient
and non-invasive, but was not suitable for all species.
Lizard numbers contracted during a marked drought early in the study, and
expanded rapidly with drought-breaking rains in 1990-1991. However, continuing strong
rains during the mid-1990's. which were El Nino years with above average winter rains
but poor summer rains, failed to sustain the population expansion. Instead, lizard
predators increased, and lizard numbers declined. As wet conditions abated during the
late 1990's. this situation was reversed: predators declined and lizard numbers increased.
This indicates that predation can at times have an overriding effect on lizard populations, and suggests that predation may have an important effect on lizard community dynamics. 91
INTRODUCTION
Lizards are prominent and diverse in arid and semi-arid environments (Pianka.
1986; Schieich et al.. 1996; Vitt. 1995). Abundance and biomass can be high (Turner.
1977), apparently in some cases high enough to produce e.xploitative food competition among insectivorous species (Dunham. 1980; Smith. 1981). The biomass required for exploitative competition in an ectotherm implies that lizards should be important prey in many of the warm ecosystems where they are abundant (although it is true that lizard populations could conceivably depress insect abundance on only a spatio-temporally local scale—if a small part of the habitat was occupied by the lizards). Nonetheless, if we are to understand the functioning of arid-lands vertebrate communities, we should begin under the hypothesis that lizards are one of the key trophic linkages within the system.
Most long-term lizard population studies have focu.sed on one or a few populations, usually of one or a few species (see Turner. 1977; Andrews and Wright.
1994). Intensive mark-recapture studies of life history, experimental studies of inter specific competition, and broad community-pattern studies have predominated in the field of lizard ecology (see Milstead. 1967; Huey et al.. 1983; Dunham et al., 1988a; Vitt and
Pianka, 1994). Lizards are so tractable for research—easy to observe, mark, and census— that they have been considered "model organisms", the highest praise a group can receive in "modem" ecology. However, landscape patterns in lizards (e.g., Conroy, 1999) are little-incorporated into lizard population dynamics work; long time-series treating interacting species are rare; and population regulatory forces are poorly understood, especially where predation is concerned (see Case, 1994). Such studies have not been 92
pursued because they appear to be too time- and labor-intensive.
This suggests a need for efficient, non-invasive methods for population study in
lizards. Recent work suggests that line transects conducted within a distance sampling
conte.xt (Buckland et al., 1993) may efficiently produce a wealth of unbiased information,
and it appears that lizards have key characteristics such as abundance and observability
that suit them to this method. Landscape perspectives (Naveh and Lieberman, 1994) may
be relevant both to autecological and community studies. The tendency to study lizards
by mark-recapture results in selection of high-density populations for study, rather than
average or typical situations; and the range of population dynamic settings experienced
by regional lizard populations has been practically ignored. Land.scape patterns are
important in variation of habitat quality (structure, productivity, etc.), and because
predators (snakes, carnivorous birds and mammals) exploit lizard populations at scales
that are orders of magnitude larger than typical study areas.
In this context, an ideal ecological monitoring program for lizards would include
(1) long term sustainabilty, in (2) a multi-species, landscape context, with (3) tracking of
relative abundance at several or many representative sites, and (4) intensive study of
abundance and demographics on a more localized basis. This program should be linked to
(5) simultaneous monitoring of key environmental parameters, both abiotic (especially climate) and biotic (disease, parasitism, and especially, predation). A number of
monitoring methods have been used for lizards (e.g.. Campbell and Christman, 1982;
Degenhardt. 1966; Dunham et al., 1988b; Kams. 1986; Vogt and Hine, 1982), although
Turner (1977) criticized many of them. In this paper, I present results from a line transect 93 method applied to a desert lizard assemblage, and compare them to results from time- constrained search, pitfall trapping, drift-fence trapping, and mark-recapture methods that were simultaneously utilized.
METHODS
STUDY AREA DESCRIPTION
This study is part of a larger ecological monitoring program (EMP) at Organ Pipe
Cactus National Monument (ORPI) based on a series of representative study sites (see
ORPI. 1995). Sites were chosen to encompass known variation in the ORPI ecosystem, in accordance with a desert landscape paradigm. Therefore. I offer a detailed habitat description here to clarify the study design.
Setting, Climate, Landscape and Vegetation Structure
The ORPI site, in western Pima County, Arizona, is in the Basin and Range
Physiographic Province, within the Sonoran Desert (Shreve, 1951; D.E. Brown. 1994). It lies on an ecotone between arid thomscrub (the Arizona Upland subdivision of the
Sonoran Desert) and true desertscrub (the Lower Colorado Valley subdivision).
Precipitation on the desert floor at ORPI drops sharply from east to west: based on long- term averages, it drops from about 230 mm to 175 mm. over the 30 km span of the monument (i.8 mm ppt/km) This is higher than the drop of 0.4 mm/km over 175 km from Tucson to ORPI, and also higher than the 0.9 mm/km drop rate continuing west from ORPI to Sonoran Desert minima around 75 mm at Yuma, Arizona. Thus, the 94 ecotone between "thomscrub" and "desertscrub" corresponds to a steep precipitation gradient at ORPI. The monument also includes the northernmost vegetational representation of the Central Gulf Coast subdivision of the Sonoran Desert, which includes senita cactus, in the southern edge of the monument. While precipitation on intensively studied bajada sites at ORPI averaged 230 mm during 1988-1998, canyons in the Ajo Mountains (along the eastern monument boundary), which are much less xeric. averaged 328 mm/yr.
The landscape of this region of the Sonoran Desert comprises three major features—(1) rockslopes. (2) bajadas. and (3) valley floor flats. Bajadas (including pediment on upper bajadas. and coalesced alluvial fans on middle and lower bajadas) are flanking detrital skirts that surround the rockslopes (mountains, buttes, and rocky hills) from which they are derived. Smaller rocky features may be mostly submerged in this valley fill, with small emergent former mountaintops juxtaposing normally distant parts of the landscape slope gradient. The flats on the valley floor at ORPI are of two types.
The north side floodplain of the now-entrenched Rio Sonoyta is a broad alluvium deposited perpendicularly to the bajada slope, and thus burying much of the lower bajada: here, middle bajada inter-fingers in immediate juxtaposition with valley floor flats. These flats, as in similar formations elsewhere in Arizona, support a mixed vegetation dominated by saltbush (Atriplex spp.). Elsewhere at ORPI, lower bajadas merge almost imperceptibly with the flats, which then support creosotebush {Larrea triclentata) associations.
The slope gradient has an obvious inflection point at the rockslope-bajada 95 interface, and a variously well-defined or fuzzy transition between bajada and flats.
Moving down the slope gradient, soil particle size grades from boulders and outcrops, through gradations of cobble, gravel, and sandy loams on the bajadas. to very fine sandy, silty, and even clay loams on the flats. Sand dunes are absent: the only loose sand occurs adjacent to thomscrub in the arroyos. Vegetation transitions (life form, flora, species) correspond closely to changes in slope and soil at ORPI, as elsewhere in the northern
Sonoran Desert (Yang and Lowe, 1956; see McAuliffe, 1994).
Superimposed on and orthogonal to the three primary topographic features is a pattern of xeroriparian scrub along drainage lines, which contrasts sharply with the surrounding upland desertscrub. This is a ubiquitous juxtaposition of denser .xeroriparian thomscrub with true desertscrub or with sparse upland thornscrub (the paloverde-saguaro association). Xeroriparian desertscrub has shrubtrees (3-6 m tall; locally accompanied in some mountain canyon bottoms by small trees), dense shrub, sub-shrub (suffrutescent). and grass. The lower stories usually, and the taller shrub stories sometimes, form continuous thickets. This xeroriparian vegetation occurs as parallel lines on arroyo banks separated by sandy or rocky-sandy wash beds. Wash beds may become complex and braided in places. Here, the parallel lines of vegetation dissolve into productive desert floodplain environments.
ORPI has a diverse flora of at least 557 native + 48 introduced species (Bowers.
1982; Felger, 1990; and Felger et al., 1997). In the following descriptions, plant taxa are listed in order of structural and areal dominance. Names correspond to those found in
D.E. Brown (1994) and Bowers (1982). Major canyons support extensive xeroriparian 96 thicket overstories of mesquite, whitethorn and catclaw acacias, hopbush. desert and netleaf hackberry. Ajo oak, and others. Major thickets also occur on bajadas along the largest arroyos, where they are dominated by blue paloverde, mesquite. ironwood. and catclaw acacia. Most bajada arroyos are bordered by leguminous shrubtrees: open stands of ironwood, foothill paloverde, acacias, and mesquite. Xeroriparian habitats typically have a variety of small shrubs and similar-sized grasses and are heavily infused with large individuals of plant species characteristic of the surrounding upland desertscrub.
Productive valley floor thickets, some of which are quite extensive on highly braided and delta-like floodplains. are dominated by mesquite. creosotebush, blue paloverde. catclaw acacia, condalias. and crucifixion thorn; they also support a variety of characteristic shrubs, vines and grasses.
Upland desert.scrub on rockslopes is dominated by brittlebush, triangle-leaf bursage. jojoba, a variety of other small shrubs, foothill paloverde. and mixed cactus including prickly pear, saguaro. chollas. and organ pipe. Upland desertscrub on bajadas is dominated by creosotebush and triangle-leaf bursage. foothill paloverde. mixed cactus including several cholla species, saguaro. and fishhook barrel, small shrubs including white bursage. white ratany. and the shrub-like big galleta grass, as well as ocotillo. In bajada areas with deep sandy loams, there are open desert woodlands dominated by ironwoods and foothill paloverde shrubtrees. which usually stand along the ubiquitous small drainages, or washlets. The flats are vegetated by creosotebush (often in nearly pure consociation with scattered fishhook barrel cactus), white bursage. triangle-leaf bursage. and big galleta; or by saltbush associations. 97
More specific and systematized descriptions of these vegetation types and tiieir component species are available in Shreve and Wiggins (1964) and D.E. Brown (1994) for the Sonoran Desert region, and Warren et al. (1981) for ORPI.
Ecological Factors and Interactions Important for Lizards
Xeroriparian thickets may be penetrated by a human with difficulty. More open xeroriparian formations can be traversed by winding through openings in the vegetation, but many vegetation clumps are impenetrable. Movement of predators such as canids and felids may be obstructed by lower vegetation layers (shrub and sub-shrub) in xeroriparian scrub, where small animals such as snakes, lizards, and smaller rodents would seem to have an advantage. Conversely, cats may use such dense cover for ambush sites.
Upland de.sertscrub at ORPI can usually be traversed by humans with little contact with the vegetation, whereas at Tuc.son. the sub-shrub stratum presents a more nearly continuous canopy network that offers camouflage and cover against avian and mammalian predators. At ORPI. open desert woodlands are found from rockslopes down through middle bajadas. They consist of a dispersed array of single (or occasionally two or three) individual shrubtrees (2-6 m tall: foothill paloverde, ironwood, mesquite). which form roughly hemispherical complexes with fringing, variably dense stands of smaller shrubs (including especially wolfberry or thombush [Lyciiim] species), chollas, young saguaros and grasses, all of which benefit from the shadow of the shrubtree. These shrubtree complexes form obstacle courses easily negotiated by fleeing lizards, and offer a diversity of thermoregulatory options. 98
The desert vegetation, with the exception of creosotebush-dominated areas, is
thorny, very spiny, and armed with stout twigs, and is thus often repellent to lizard
predators but not to lizards. Even in the upland desertscrub, cholla, shrub-tree complexes,
and multiple-stemmed creosotebushes and bursages offer difficult-to- penetrate obstacle
courses utilized by variously-sized lizards and rodents. On the rockslope, crevices,
boulders, and talus offer cover, but obstruct the fast-running flight that is practicable in
valley habitats.
Arroyos in the upper bajada (or pediment) are typically incised 1-4 m into consolidated alluvia (conglomerate and caliche), or unconsolidated coarse-packed alluvia.
The embankments are richly peppered with cavelets. holes, and crevices utilized extensively by reptiles, small mammals, and arthropods. Bajada and valley tloor surfaces are extensively undermined by rodent burrows and burrow comple.xes. Bajadas. xeroriparian habitats, and especially rockslopes have substantial densities of packrat nests, which have fortress-like superstructures of plant fragments that often include high
proportions of very spiny teddy bear and chain-fruit cholla segments. These structures are
used by many species of lizards, snakes, murid rodents, and arthropods. Canids, badgers, and presumably other predators frequently excavate small soil burrows of prey
vertebrates, and occasionally excavate packrat houses, even very spiny ones.
Thermal refugia are widespread underground in rock and soil. Further,
xeroriparian thomscrub vegetation, and the shrubtree complexes, offer significant
buffering against extremes of heat and cold (Lowe and Hines. 1971). Arroyo banks, and soil pockets in bedrock, boulder, and rock areas, have elevated soil moisture and relative 99
humidity (pers. obs.) in a dry environment with essentially no surface water. Stark contrasts between sun and shade may enhance the camouflaging effect of the dazzling sunlight (see Norris, 1967).
Raptorial birds (great-homed, western screech, elf, and bam owls; red-tailed,
Harris', and Cooper's hawks; American kestrels; roadrunners; loggerhead shrikes; and golden eagles) are abundant year-round residents. They utilize saguaros, as well as shrub-
trees, rock promontories, and utility poles as hunting perches. Redtailed hawks, the most abundant buteo, are especially frequently seen hunting around desert rockslopes. Other abundant birds that probably eat lizards include thrashers, woodpeckers, ravens, and cactus wrens. Canids (coyote, kit and gray foxes) are abundant throughout the bajadas and valley floors, hunting primarily at night during most of the lizard activity season.
Badgers are occasionally abundant at local sites. Spotted skunks, ringtails, and mountain lions are regularly present on and near rockslopes, and bobcats are especially abundant there. Other endothermic predators are rare or localized (see Chapter 3; Cockrum and
Petryszyn, 1986).
A diverse assemblage of mammal- and reptile-eating snakes occurs throughout the monument (see Chapter 4), including 2 whipsnake species (Masticopliis flagelluin, M. bilineatus), the western patch-nosed snake, 6 rattlesnakes, 4 constricting colubrids, rosy boa, lyre snake, night snake, western coralsnake, and 2 leaf-nosed snakes (which are lizard egg specialists found on bajadas and valley floors). The diurnal, highly active,
whipsnakes are abundant throughout the monument.
Certain arthropods are also significant lizard predators at ORPI. The desert hairy 100 scoqjion Hadriiriis arizonensis is a large, abundant, lizard-eater on bajadas and valley floors throughout the monument (Lutz. 1980; G.L. Bradley, personal communication); tarantulas {Aplionopelnia sp.) are abundant throughout, and presumably eat lizards occasionally; and large centipedes {Scolopendrci hems), known lizard predators, are present in rocky areas.
The most prominent insectivorous birds that might compete with lizards are black-tailed gnatcatchers. curve-billed thrashers, wrens, verdins. and ash-throated flycatchers (Hensley. 1954; ORPI, unpublished). A diversity of scorpions, centipedes, spiders, and predatory ground-walking beetles are also abundant at ORPI (Kingsley.
1998), and might compete with lizards.
Climate Methodology and History
Climatic data are available from a reporting weather station at monument headquarters (el. 512 m) starting in 1949 (daily maximum and minimum temperature and precipitation). Ancillary data were provided by 8 Forester rain gauges at sites throughout the monument starting in the early 1960's. but consistently only from 1982.
Beginning in 1987, an extensive climatic study was initiated in tandem with the
EMP. Data was obtained at 9 EMP study sites using automated weather stations equipped with data loggers. Omnidata Datapods at each site recorded 6 to 8 of the following parameters into taped format downloaded to Lotus 1-2-3 spreadsheets; air temperature at
1.5 and 12.2 cm aboveground, precipitation, relative humidity, wind speed and direction, solar radiation, and soil temperature at 10 cm depth. lOl
Two additional stations were installed at EMP sites in 1995, and the system was converted to Handar recorders downloaded by palmtop computer using WeaBase software. In 1997. soil temperature and moisture monitoring was initiated at several sites to obtain depth profiles. For most climate parameters, data are recorded hourly, with daily and monthly summaries produced automatically.
The automated weather stations are monitored monthly by a technician who validates equipment function, makes needed repairs or adjustments, and scans the data and output for evidence of gross errors or detectable inaccuracies. Erroneous data is deleted, and care has been taken to avoid automated computations that permit such necessary data deletions to bias means or totals for the parameters. Without this intensive oversight of the machines, the automated weather station data might be virtually useless.
Rainfall at ORPI was exceptionally high during the wet period of 1977-1984.
Lizard populations in an area under study at the end of that period were also high (Rosen and Lowe. 1996). That period was followed by a late-1980's drought that was accompanied by high temperatures. The early and mid 1990's saw continued summer drought with few exceptions (1990, 1998), but several very wet winters (1991-2, 1992-3.
1994-5, and 1997-8) associated with the El Nino Southern Oscillation phenomenon.
Drought conditions occurred in 1989-1990 and 1996-1997.
LIZARD MONITORING METHODS
We established 30 lizard line transects at 18 sites, plus 18 pitfall traps and 22 drift fence traps at one site, to monitor lizard population changes from 1988-9 to 1998. 102
Nocturnal time-constrained search was used at three sites in an intensive study area, which included the trapping site. In addition, we conducted a mark-recapture study at the trapping site as a validation for the line transect methodology (see Rosen and Lowe,
1995, 1996).
Transect Methods
This study includes results from 21 of the transects at 13 sites that were established early in the study. Transects are 100-300 m long, permanently marked by tagged rebar stakes at 50 m intervals, and have been mapped using global positioning system technology. They are run by an observer walking the centerline defined by the rebar stakes, leaving only to verify identity of brietly-perceived lizards. Lizards were only recorded if they were within 7.5 m of the transect line during a transect walk.
Transects were run only on warm, clear or mostly clear days. Seasonally, each transect was run twice per year during the periods of highest lizard activity (spring: late
April to early June; summer monsoon: onset of summer rains [July-early August] to early
September). They were walked 4-8 times/day beginning at or prior to morning emergence of diurnal lizards. The rate of travel on the transects was adjusted by the observer and included brief stops to scan for motionless lizards; generally, the rate of progress was 5-
10 min per 100 m of transect. The walks of one transect during one morning comprise a run. For each species, the result of the run was tallied as the maximum number of individuals in one walk. During the course of a morning, species' activity levels increased as the environment warmed, and decreased as excessive heating occurred; thus the 103
maxima are here termed "peak values".
The transect method used here was designed as an abundance index for whiptail
lizard {Cnemidophorns tigris, C. biirii) abundance within sites over time. Transect walks
began before any lizard species reached high activity levels, and terminated when
whiptail lizards became inactive underground or in shade. Whiptail peak values sometimes occurred when shade activity, punctuated by brief forays in sun, was strong;
this period was followed by decreased activity, and tran.sect walks were terminated at this
time. Therefore, the method was less well-suited for thermophilic species--the desert
iguana Dipsosauriis dorsalis and the chuckwalla Saiirumcilus obesiis—ixud yielded no
information on the nocturnal western banded gecko Coleonyx variegcitiis.
For each lizard observed on-tran.sect, we recorded species, time, distance along
transect, distance from midline, size class, and sex (if possible). Age-classes were categorized in the field as large adult, adult, small adult, subadult. large juvenile, juvenile, small juvenile, hatchling, based on data from trapped animals. For analysis, the age-class data were collapsed to the following categories to reduce error and observer biases: adult, juvenile/subadult. and small juvenile/hatchling. Prior to starting each walk, we recorded cloud cover and wind conditions (qualitatively), air at 1.5 m and substratum temperatures
in sun (quick-reading thermometer, bulb shaded), starting time, and ending time.
A lizard monitoring protocol handbook was developed with very detailed guidelines for walking speed, walk timing and frequency, and observer behavior on
transect (Rosen and Lowe. 1995). Results discussed here were obtained by the author
(1989-1991) and a field technician (1991-1998) trained by the author according to the 104
protocol. Consistency of methodology was verified periodically by joint work, and by
examination of histograms produced from the distance from midline data (Rosen and
Lowe. 1995).
Intensive Study Area: Trapping Methods
Lizards were trapped in 22 drift fence traps set to catch snakes, and in 18 simple
pitfall traps. These traps were set on a 1 km' study area (ORPI Snake Study Area: OSSA)
that included one of the EMP sites. The pitfall traps were set in a grid pattern around two
of the snake traps near the center of OSSA. There were 4 lizard transects within the
OSSA boundary, including two that went through the pitfall grid.
Drift fences and habitat conditions at OSSA are described in detail in Chapter 4.
Brietly, the drift fences were 12-18 m lengths of hardware cloth leading to funnel traps
that in turn led to mostly-sunken buckets from which lizards could not escape. These
were permanent installations which were checked once or twice per day when they were
in use. When they were not in use, the entrances were sealed, and exit ramps were
installed in the buckets in case the seals failed. There were 5 traps in 1988. 8 in 1989. 14
in 1990-91. 22 during 1992-3, 17 during 1994-5, and 10-12 during 1996-1998.
Pitfall traps consisted of either (1) commercial 19 1 (5 gal) white buckets (n=9; 30 cm diameter, 36 cm depth) set ground flush and covered with 50 cm square plywood covers raised 2-5 cm above the ground or (2) metal cans (n=9: 15.5 cm diameter, 50 cm depth) set similarly but with 25 cm covers. These traps were checked once per day, and
were utilized only during 1988-199 L 105
Time-Constrained Searcti (TCS)
During all routine, desert herpetological work at ORPI, I recorded starting and
stopping time, area worked, activity carried out, and all reptiles, amphibians, mammals,
and raptorial birds visually observed. Numerous lizard counts were made during
nocturnal searches for snakes, daytime trap checks, and off-transect during lizard transect
monitoring. The results presented here are for geckos on the three EMP sites near and
including OSSA, including the trapping area plus areas within 4 km of it.
Handling, Measurement and Marking
Lizards were removed from traps during daylight hours, processed on the spot,
and released immediately. Geckos were released into small burrows, packrat houses, or
other thermal refugia, especially little-used ones where they were unlikely to encounter
predators. Sex was determined by inspection or probing (see Stebbins. 1985). Females
were checked for signs of reproductive condition (egg-distended abdomens or eggs
visible through body wall). The following data were recorded for each lizard captured: date, trap, species, sex, size class or size, and identification mark (if any). For each lizard
involved in the mark recapture study, I also recorded snout-vent length (SVL, mm, dangling in a straight, relaxed position along a ruler), original tail length, length of regenerated tail portion (if any), body mass (g), and release time.
During 1988 and 1989, lizards were marked at all drift fence traps in use. and all lizards captured in pitfall traps (1988 - 1991) were marked. During 1990-1994, lizards
were marked only at the 4 central drift fences on the study area, in the area that also had 106 the pitfall grid and two line transects. Marks were individual-specific toe clips consisting of 2-3 toes per individual, with no more than one toe cut per foot. Occasionally. mismarks were corrected with a fourth toe clip. Toe clipping was done with scissors that were disinfected with betadine or alcohol, and fresh marks were swabbed with betadine.
COMPUTATIONS AND STATISTICS
Analysis of Transect Results
Only transects studied during >15% of the 20 monitoring seasons (1989-1998) were included in the time trend analyses reported here, to minimize site-specific biases. A total of 21 transects at 13 EMP monitoring sites (Aguajita Wash. Alamo Canyon.
Armenta Ranch. Bum, Creosotebush. Dos Lomitas. East Armenia. Growler Canyon.
Lizard Grid. Pozo Nuevo. Salsola. Senita Basin. Vulture) are considered here; data from
9 transects at 5 sites are e.xcluded (Lost Cabin Mine. Lower Colorado Larrea, Middle
Bajada. Quitobaquito. Valley Roor) although they are under continuing study.
The peak value for each species was read from the data sheets, and adjusted to a common currency of lizards/100m. Means and standard errors for each monitoring season or year were computed directly from these data. The implied confidence intervals are therefore conservative with respect to inter-annual contrasts, since a substantial portion of the total variance is site-site variance. Sampling periods were considered significantly different if their standard error bars did not overlap. 107
Data Reduction and Analysis of Trap Captures
The objective of this study was to study variation in abundance among years.
Therefore. I computed a standardized annual capture rate for each species. Raw capture
rates varied among traps and seasons. The original 5 traps were less efficient than
subsequent installations; each trap had a characteristic capture rate for each species.
Capture rates also varied among months, although variance was smaller among months
than among traps. Therefore, capture data were standardized by first adjusting monthly
capture rates to account for inter-trap differences, producing a vector of standardized
monthly capture rates for each species. Using this vector as input, I computed a vector of
the characteristic trap capture rates for each individual trap. Again, this was done
separately for each species. These two values were then used to adjust the capture rate of
each year, yielding the standardized capture rate per year, again for each species. Details
of this procedure are discussed further in Chapter 4. As for snakes, variance terms
(standard errors) for the lizard capture rates were estimated from the formula for variance
of a binomial parameter (SE p = [P x Q/ 11"^, P= per trap-day capture rate. Q=\ - P,
N = no. trap-days [adjusted value]; for values of P> I (i.e.. in Cnemidophoriis tigris) I computed the standard error from daily capture rates using normal statistics. 108
RESULTS
LINE TRANSECTS
Landscape Structure of the Fauna
The lizard fauna at ORPI shows considerable ecological equivalency between
members of rockslope and valley (bajada + valley floor) assemblages (Table 2.1). with
apparent species-for-species matching (see Schluter, 1990) within broad niches. This
differs from descriptions for some North American Desert lizard assemblages (Gonzales-
Romero and Alvarez-Cardenas. 1989; Gonzales-Romero et al.. 1989; Maury and
Barbault. 1981). At ORPI. two genera (Cnemidophorus, Sceloporus) have species
replacements at the rockslope-bajada inflection point; in both cases the replacement
involves species markedly similar in size, behavior, and habitat use. In two other ca.ses.
similar-sized lizards, with similar diets, are replaced at the inflection point. Gcinibelia
wislizenii in the valley is replaced by Crotciphytus nebrius (formerly a sub-taxon of C.
collaris: Gambelia and Crotciphytus were recently congeneric, the split might more
appropriately be subgeneric). These species are medium-large, insectivorous-saurivorous
predators, but they use substrata differently. Gambelia is a flat-ground runner, while
Crotaphytiis characteristically perches on rocks. The other replacement involves the two
large herbivores—the ground-running Dipsosaurus dorsalis and the rock-perching, crevice-dwelling Saiiromaliis obesus. Although the ground-running Callisaiirus draconoides has no rockslope equivalent at ORPI (in more mesic regions of southern
Arizona, closely-related species—Holbrookia sp. or Cophosauriis te.xanus--do replace 109
Table 2.1. Species for species matching in the desert lizard fauna at Organ Pipe Cactus National Monument. There is occasional syntopy in this region for species pairs indicated with an "or".
VAI.LFY HABITATS
ROCKSI.OPF ARIZONA (rPF AND f OWFR rOi nRADO VALLEY
r. terrestrial ant-eating species (homed liyards) (niche vacant) Phrynosoma solare Phrynosoma plaryrhinos
[I. small terrestrial insecrivnre .'side-hlotched lizard) Uta stansburiana Uta stansbiiriana Uui stansburiana
HI. small arboreal and scansnrial Insectivnre (tree / brush lizards) Urosaurtis ormitus (or Uia stansburiana) Urosaurus ornatus Urosaurus graciosits (or U. ornatus)
IV. medium-sized, terrestrial, wait-dash insectovore (zebra-tailed lizard) (niche vacant) "allisaurus draconoide: Callisaunis draconoides
V. large arboreal insectnvnre (spinv lizards) Sceloprous clarkii Sceloporus magister Secloporits magister
VI. large terrestrial lizard\anhrnpnd-enting species (crotaphytines) Crotapliytus nebrius Gambelia wislizenii Gambeiia wislizenii
VII. large terrestrial herbivore (chuckwalla / desert iguana) Sauronialus obseits Dipsosaurus dorsalis Dipsosaiirus dorsalis
VIII. small, nocturnal, terrestrial inseetivore (gecko) Coleonyx variegatus Coleonyx variegatus Coleonyx variegatus
IX. medium-sized active-foraging, terrestrial inseetivore (whiptails) Cnemidophorus burti or C ligris Cnemidophorus tigris Cnemidophonis tigris
X. large carnivorous nestling predator (Gila monster) Heloderma suspecturn Heloderma sitspectum Heloderma suspectum
TOTAL SPECIES 8 10 10 110
Callisaiiriis upslope. but these are almost entirely absent from Sonoran desertscrub in southern Arizona), and Phrynosoma is apparently absent from the rockslope, the parallel structure remains striking. For present purposes, this ecological equivalency justifies combining matched pairs into "ecospecies", simplifying the analysis. Similarly, species
that display replacement across the Arizona Upland-Lower Colorado Valley boundary
(Table 2.1) are treated as a single ecospecies in the line transect analysis.
All Sites Combined
The most frequently observed lizard ta.xa on the 21 transects were, in decreasing order of abundance. Cnemidophonis, Uta stcinsburiana, Urosaurus, Callisaurus clraconoides, Sceloporus, Dipsosciiiriis clorsalis, the crotaphytines . and Phrynosoma solare (Table 2.2). Too few Gila monsters were observed for analysis, and no chuckwallas were recorded on the transects.
Cnemidophonis, Uta stansbiiriana, Urosaurus. Callisaurus draconoides, and
Sceloporus were all abundant on the transects. With the exception of U. stansburiana. their monitoring results followed strikingly consistent trajectories over the 10 yr of study
(Figs. 2.1. 2.2). Peak values for most taxa declined in 1990. rose steadily to highs in
1992. followed by lows in 1995. with maxima reached again in 1998. Deviations from this pattern were surprisingly minor; Sceloporus reached a minimum in 1994. slightly
below the 1995 value; and Cnemidophorus declined in 1998. although it still maintained high abundance. Dipsosaunis showed the same general pattern, while the crotaphytines, with limited data, were apparently somewhat divergent. Ill
Table 2.2. Yearly summary of lizard line transect results for ORPI. 1989-1998. Values are means of activity-peak values per lOO m of transect. YEAR TAXON 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998
Cnemidophoriis mean 2.09 1.57 3.02 3.36 2.65 2.38 1.86 2.16 3.44 2.98 SE 0.24 0.18 0.33 0.42 0.36 0.28 0.19 0.19 0.24 0.27
Uta staiisburiaita mean 0.87 1.25 0.93 0.99 1.36 1.59 1.88 2.46 2.48 SE 0.27 0.24 0.27 0.37 0.30 0.28 0.32 0.41 0.37 0.36
Ccillisciiirus draconoicles mean 0.66 0.46 0.91 1.35 0.67 0.53 0.26 0.33 0.70 1.14 SE 0.18 O.Il 0.22 0.43 0.19 0.12 0.10 0.09 0.14 0.32
Sceloporus mean 0.21 0.20 0.32 0.44 0.33 0.11 0.14 0.14 0.25 0.38 SE 0.08 0.07 0.10 0.11 0.08 0.04 0.05 0.05 0.08 0.11
Urosatints mean 1.23 0.52 l.Ol 1.04 0.83 0.43 0.41 0.63 1.22 1.50 SE 0.27 0.14 0.28 0.35 0.25 0.10 0.12 0.12 0.25 0.28
Dipsosattrits dorsalis mean 0.05 0.05 0.08 0.05 0.07 0.05 0.01 0.05 0.02 0.14 SE 0.03 0.04 0.05 0.03 0.03 0.03 0.01 0.03 0.02 0.05 crotaphytines mean 0.01 0.01 0.01 0.17 0.08 0.05 0.07 0.01 0.02 0.04 SE O.OI 0.01 0.01 0.09 0.04 0.05 0.04 0.01 0.02 0.03
Phrynosoma solare mean 0.00 0.03 O.OO 0.00 0.00 0.00 0.02 0.00 0.00 0.00 SE 0.00 0.03 O.OO 0.00 0.00 0.00 0.02 0.00 0.00 0.00
Helodemta suspectitm mean 0.00 0.00 O.OO 0.00 0.02 0.03 0.01 0.00 0.02 0.00 SE 0.00 0.00 0.00 0.00 0.02 0.02 0.01 0.00 0.02 0.00
Sum 5.12 4.10 6.27 7.40 6.01 5.16 4.66 5.78 8.00 8.66 No. /100 transect-m No. /100 transect-m No. / 100 transect-m S No. / 100 transect-m ^ to lO M U U A 3a ouio'uiocnoul b
1994
No. /100 transecl-m No. /100 transecl-m No. /100 transecl-m No. /100 transecl-m
1994 113
correlated population trajectories, ORPI —— Cnemidophorus • Callisaurus 4.0 — —Sceloporus • Urosaurus
8 3.0 c(0 (0 ^ 2.0 o o 1.0 o z — " - — — 0.0 a> CM CO in (O 00 oo CD o> o> at at 0) O) at at at
8 0.16 M s 0.12
I 0.08 O 0.04
0.00 O) o CM m lA (O 00 00 0) O) a> at at O) at at a> O) at O) at at at O) at at 0)
contrasting trajectories
_ all species • Cnemidophorus Uta
u 8 0) c« <0 6 o o
0) o CM CO m Fig. 2.2. Correlated and contrasting results for species abundances on the transects. The middle graph shows results for the taxa with small sample sizes. 114 Ufa stansburiana showed a contrasting, partially reversed, pattern (Figs. 2.1. 2.2). Its peak values rose in 1990, were at minimum in 1991-1992, and rose to a high point through 1996. In common with the other species, it maintained high abundance during 1997 and 1998. Total lizard abundance on the transects varied systematically but not dramatically, with an amplitude of only twofold variation between high and low points, although individual taxa had a greater range of variation. Variances suggest that the trends indicated for abundant species are real, although slightly differing interspecific differences should be interpreted cautiously (Fig. 2.2). Results for less abundant ta.\a (Dipsosciurus. crotaphytines. Phrynosomci. Helodenna) carry large variances and were insufficient for accurate determination of population trajectories. Intensive Study .\rea (OSSA) Abundance patterns on the four OSSA transects are relevant for comparison to other monitoring methods, which were used in the OSSA area. The patterns from these transects were similar in detail to those from all 21 transects, except that the infrequently recorded species (Dipsosaiiriis, Garnbelia) failed to show a late 1990's rebound in this more limited dataset (Fig. 2.3; Table 2.3). The four-transect results show wider and sharper fluctuations than those averaged for all 21 scattered throughout the monument, probably primarily as an artifact of small sample size, but also through a real averaging effect across sites. For the abundant taxa, the only differences in trends between the two datasets are that Urosaunis, Callisaiirus, and Sceloponis at OSSA maintained low. Table 2.3. Seasonal results (No. / 100 m transect, peak values) ("or lizard line transect resull.s, QRPI. YEAOEASON 1989 1990 1991 1992 199:^ 1994 1995 199ft 1997 1998 TAXQN Spr Sum Spr Sum Spr Sum Spr Sum Spr Sum Spr Sum Spr Sum Spr Sum Spr Sum Spr Sum Cnemidophorus mean 2.65 1.45 1.60 1.55 3.15 2.84 3.81 2.90 1.86 3.44 2.12 2.65 1.33 2.38 1.53 2.80 3.16 3.72 2.34 3.62 SE 0.35 0.26 0.27 0.25 0.40 0.55 0.54 0.63 0.22 0.65 0.36 0.44 0.26 0.24 0.19 0.28 0.38 0.31 0.22 0.47 Ula sUmsburiana mean 0.49 1.29 0.81 1.64 0.60 1.35 0.67 1.30 0.54 2.18 0.83 2.35 0.85 2.91 1.41 3.51 1.75 2.89 1..34 3.61 SE 0.14 0.54 0.25 0.37 0.25 0.53 0.29 0.69 0.12 0.53 0.25 0.44 0.23 0.52 0.34 0.67 0.42 0.59 0.24 0.58 Cullisaurus draconoides mean 0.76 0.54 0.31 0.59 0.96 0.86 1.43 1.26 0.84 0.49 0.56 0.49 0.21 0.30 0.35 0.30 0.68 0.72 1.12 1.15 SE 0.30 0.16 0.11 0.18 0.33 0.29 0.64 0.60 0.35 0.16 0.21 0.13 0.15 0.14 0.14 0.13 0.18 0.22 0.49 0.41 Sceloporus mean 0.18 0.26 0.21 0.19 0.25 0.40 0.37 0.51 0.31 0.36 0.07 0.14 0.09 0.19 0.06 0.22 0.28 0.23 0.25 0.52 SE 0.09 0,14 0.11 0.10 0.13 0.17 0.18 0.13 0.10 0.12 0.05 0.07 0.05 0.09 0.04 0.10 0.13 0.09 0.12 0.19 Urosaurus mean 1.58 0.84 0.51 0.53 1.26 0.68 0.93 1.15 1.13 0.52 0.46 0.39 0.59 0.24 0.61 0.64 1.42 1.01 1.53 1.46 SE 0.34 0.43 0.19 0.20 0.43 0.31 0.57 0.41 0.47 0.16 0.17 0.12 0.20 0.12 0.17 0.18 0.40 0.30 0.42 0.37 Dipsostmrus dorsalis mean 0.06 0.03 0.03 0.06 0.14 0.00 0.04 0.06 0.05 0.10 0.00 0.10 0.00 0.02 0.02 0.07 0.00 0.04 0.12 0.16 SE 0.06 0.03 0.03 0.06 0.08 0.00 0.04 0.04 0.03 0.06 0.00 0.06 0.00 0.02 0.02 0.05 0.00 0.03 0.07 0.07 crolaphytines mean 0.00 0.02 0.02 0.00 0.02 0.00 0.17 0.18 0.04 0.12 0.00 0.10 0.05 0.10 0.02 0.00 0.00 0.05 0.07 0.02 SE 0.00 0.02 0.02 0.00 0.02 0.00 0.14 0.10 0.03 0.07 0.00 0.10 0.05 0.07 0.02 0.00 0.00 0.05 0.05 0.02 Phrynosomu solare mean 0.00 0.00 0.07 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.05 0.(K) 0.(X) 0.00 0.00 0.00 0.00 0.00 SE 0.00 0.00 0.07 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.05 0.00 0.00 0.00 0.00 0.00 0.00 0.00 Helodenna suspecium mean 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.05 0 00 0.06 0.(X) 0.02 0.(KJ 0.00 0.00 0.00 0.05 0.00 0.00 SE 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.05 0.00 0.05 0.00 0.02 0.00 0.00 0.00 0.00 0.05 0.00 0.00 All 5,72 4.43 3.58 4.56 6.38 6.14 7.41 7.38 4.82 7.20 4.10 6.22 3.19 6.13 4.02 7.55 7.29 8.71 6.78 10.54 116 Cnemidophorus all (solid), OSSA (dashed) s ±1 SE (Q(o(ov)orv) Uta stansburiana 5 9)9»QQ*i»f» Dfpososaurus dorsaffs .3 3SSSSSS!3i3333Sa{$$«a)SS a3 a3a3a3a3 a3a3 a3 a.3 (/)a. (03 Urosaurus . ±1 SE SSSSSSoiSiSSSSSaiSSAnSS Fig. 2.3A. Seasonal results of lizard line transect work at ORPI, showing all transects (with standard errors) and OSSA site only (without standard errors). 117 Calllsaurus draconoides ti SE 2.0 S 8 o z 0.5 0.0 3! 3 Si n 3 3 v)(ou309MWU]03U} Sceloporus 1.0 i 1 SE 0.8 0.4 0.2 0.0 S) a 3 3j 3 3 3 3 3 crotaphytines .4 :1 SE =o .1 .0 « « 3 3 3 3 S 3 3 3 3 ,9- .3 ,0. 3 a 3 Q. 3 O. ,9- ,2 ALL LIZARDS '2 10 8 6 4 2 0 3 3 3 3 3 3 9! 3 3 3 3 caca«]O]«i Fig. 2.3B. Seasonal results of lizard line transect work at ORPI. showing all transects (with standard errors) and OSSA site only (without standard errors). 118 constant levels during 1993-1996, with sharp increases in 1997-1998; while for the entire ORPI dataset these taxa declined steadily to 1994-1995 minima and, as in Cnemidoplwrus and Uia in both datasets, showed increases by 1996. It is noteworthy here that summer rains were especially poor in the OSSA region of the monument during the mid-1990's. DRIFT FENCE TRAPS Drift fence traps sampled Cnemidoplwrus more effectively than other lizards because the whiptails active foraging made them most susceptible to drift fence capture (Figs. 2.4, 2.5; Table 2.4). These means carry very low variance, although they are, in some years, more sea.sonally limited than the tran.sect results. In Callisauim\ a sit-and- wait forager, and in the arboreal Urosaiirus orncitiis, trapping was of limited success compared to the transects. Coleunyx variegaius was well sampled by trapping, as were diurnal taxa with less specialized behavior (Table 2.4), including Dipsosciiirus and Gambelici, for which these data may be superior to those from the transects. For Cnemidoplwrus ligris. trapping (Figs. 2.4, 2.5) and transect results (Fig. 2.1, 2) correspond closely. The early 1980's population maximum was recorded in 1991, rather than 1992. in the trap dataset. but this was apparently because the traps sampled recruiting juveniles heavily during September-November 1991. well after the transect work for 1991 had ended. Cnemidoplwrus trapping results showed a minimum in 1994. rather than 1995, but only 3 days of trapping were conducted in 1994. For 1996-1998, the u^ap result is again offset by one season from the transect results, again probably 119 Table 2.4. Yearly summary of drift-fence trap results for lizards. ORPI, 1988-1998. Values are per trap-day capture rates, adjusted to remove the effects of trap location and months trapped. Standard errors were computed based on the binomial distribution (see te.xt). YEAR TAXON 1988 1989 1990 1991 1992 1993 1994 1993 1996 1997 1998 Cnemidophorus mean 0.55 0.78 0.52 1.21 1.05 0.76 0.41 0.46 1.09 0.83 1.01 SE .0426 .0206 .0143 .0121 .0110 .0166 .0747 .0311 .0190 .0240 .0166 Uta stansbiiriana mean 0.11 0.10 0.11 0.19 0.16 0.13 0.15 0.16 0.38 0.24 0.25 SE .0344 .0165 .0096 .0150 .0153 .0120 .0501 .0229 .0263 .0239 .0210 Callisaurus draconoides mean 0.05 0.05 0.03 0.05 0.04 0.02 0.00 0.00 0.04 0.03 0.04 SE .0146 .0103 .0046 .0085 .0066 .0058 .0000 .0000 .0133 .0121 .0117 Sceloporus mean 0.09 0.12 0.12 0.23 0.19 0.12 0.07 0.07 0.06 0.15 0.11 SE ,0242 .0172 .0095 .0162 .0142 .0131 .0510 .0149 .0148 .0229 .0175 Urosaurus mean 0.03 0.03 0.04 0.06 0.04 0.00 0.04 0.00 0.02 0.02 0.03 SE .0136 .0077 .0055 .0080 .0090 .0028 .0358 .0000 .0099 .0080 .0098 Dipsosaurus dorsalis mean 0.02 0.04 0.03 0.04 0.03 0.02 0.00 0.01 0.02 0.03 0.04 SE ,0090 .0109 .0051 .0083 .0065 .0056 .0000 .0041 .0094 .0121 .0110 crotaphytines mean 0.03 0.03 0.02 0.02 0.02 0.01 0.03 0.00 0.00 0.01 0.00 SE .0146 .0092 .0036 .0057 .0049 .0041 .0313 .0000 .0047 .0067 .0037 Phrynosoma solare mean 0.00 0.03 0.03 0.03 0.04 0.02 0.00 0.00 0.00 0.01 0.01 SE .0000 .0142 .0043 .0073 .0046 .0091 .0000 .0000 .0000 .0084 .0071 Coleonyx variegatus mean 0.08 0.11 O.Il 0.09 0.06 0.03 0.12 0.03 0.L7 0.25 0.19 SE .0182 .0146 .0092 .0117 .0089 .0064 .0458 .0099 .0231 .0259 .0203 C. tigris -X- Dipsosaurus dorsalis 1.5 Uta stansburiana 0.06 >>m >» T3 CQ 0.05 6. 1.0 TJ 6. 0.04 s S g *w c 0.03 T) 0.5 •o 0.02 d z 0.01 1 0.0 T r . 0.00 OOOO-t-CNCO^-. 00 O) O CM CO o> ifi - Sceloporus magister Coleonyx variegatus -X- Phrynosoma solare 0.3 0.05 a >. •o -oa 0.04 ^Q 0.2 iL 0.03 *c£ c T) 0.1 •o 0.02 6 z 0.01 / 0.0 0.00 /, 00 Ol 0 ra CO in to N. 00Ol Ol00 01Ol 0>Ol 0> OIO) OIo> O)o> Ol O) ^ T- Fig. 2.4. Annual drifl-fence trap capture rates for various lizards at the OSSA site, ORPl, with standard error bars (+ 1 SE). The values have been adjusted for inter-year differences in season of sampling and individual trap use. -X—all prey —>!—lizards —Callisaurus draconoides — • C. tigris —+• -rodents —X— Dipsosaurus dorsalis 0 • • -o - - Gambelia wisHzenii 0.06 >.2 5 a ^ 0.05 I2 0 *0 s K 0.04 5 w C 0.03 'C •o 1.0 ' 0.02 d Z 0 5 Z 0.01 0 0 0.00 CO at 0 r- CM m rt in to CO CO CO —Utastansburiana —X— Sceloporus magister -X— Phrynosoma solare • <3-• Coleonyx varlegatus • Urosaurus ornatus OM 0.08 >> a a>» ? 0.3 •o a. a 0.06 c 0.2 £ 0.04 'C 'iZ •o -a 0.1 0.02 d o z Z 0.00 0.0 tto)OT-c>jco^m Fig. 2.5. Simplified summary and contrasts of lizard abundance revealed by drifl-fence (rapping (see Fig. 2.4 legend). 122 reflecting late summer and fall trapping, rather than methodological differences. There is no doubt that vvhiptail lizard populations at ORPI were again exploding during late- season 1998. as tremendous juvenile growth and recruitment were obvious everywhere. For Uta (Figs. 2.4. 2.5), trapping results showed a gradual rise to ma.ximal levels during the late 1990's. and were generally similar to results from the line transects. For Sceloponis niagister {Figs. 2.4, 2.5), the comparison between methods shows a contrast like that for C. rigris. The two methods yielded almost identical trends, but the trapping method showed some trends somewhat earlier. As for C. ligris, this results from (I) numerous captures of recruiting juveniles during fall trapping (representing recruitment after transect work was completed), along with (2) the effectiveness of traps for sampling small juvenile lizards, which are less readily observed than adults in transect work. For Callisaiiriis, trap returns were relatively meager, but are still in general agreement with findings from the line transects. Trapping results for Dipsosauriis dorsalis also agree with transect data showing a mid-l990's minimum, but overall show only a rough correspondence for the two methods. For Gambelici wislizenii, trapping results are consistent with OSSA transect results showing a mid-late I990's minimum and little rebound during the late 1990's; this differs from the 1998 rebound suggested by the transect data. For the latter two species, however, the trapping data are more definitive than the transect results except in 1994. due to sample size considerations. Homed lizards became scarce at ORPI during the I990's. and although the transects were insufficient to record this event, it is clearly evident for Phrynosoma solare 123 in the drift fence trap results (Figs. 2.4. 2.5). The traps showed a consistent decline of Coleonyx variegatus to minima in 1993 and 1995. with large increases during 1996-7, and a modest decline in 1998 (Figs. 2.4. 2.5). Time-constrained search results for this gecko (Table 2.5) are consistent with trap results, although sample sizes were small. CORRELATIONS AMONG METHODS Correlation coefficients for the different methods are in Table 2.6. Lizard line results from all 21 transects were strongly correlated with those for only the 4 OSSA tran.sects, in all species for which adequate .sample sizes were available. Correlations between transect results and those from the drift-fence traps were fairly high for taxa with good data from both methods. For Coleonyx variegatus, there was a remarkably good correlation between results from trapping and TCS. considering the limited numbers of observations in some years. Although not summarized here, TCS data on diurnal lizards also matched trapping results. DISCUSSION MONITORING METHODS Population monitoring by pitfall or drift-fence trapping is the most common method of lizard monitoring. Pitfall trapping is less investigator-dependent than the line transect method: one must only ensure that traps are carefully set so that surface irregularities do not vary from year to year or investigator to investigator. However, there are drawbacks that are not widely appreciated. Pitfall grids may be less cost-efficient than Table 2.5. Time-constrained search results for Coleony.\ varic'satus, intensive study region, ORPI, 1988-1998. Values are per hour encounter rate for one person using a standard 6 volt headlamp. YEAR 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 overall mean 0.060 0.055 0.156 O.(KM) 0.059 0.078 O.(MH) O.(KX) 0.281 0.233 0.165 0.112 SE 0.042 0.045 0.052 0.045 0.057 0.162 0.129 0.121 0.027 No. searches 11 27 42 16 36 20 28 3 0 27 17 29 256 to Table 2.6. Pearson correlation coefficieiils comparing monitoring inelhodologies for lizards, QRPl, 1987-1998. comparison: TAXON transects (OSSA vs. all) transects vs. drift fences drift fences vs. tcs CnemUlophorus 0.94 0.60 Ula sumshuruina 0.98 0.76 Callisaurus dmctmoides 0.84 0.49 Sceloporiis 0.74 0.70 Urosaums 0.69 0.17 Dipsosau rus dorsalis 0.02 0.44 crotaphytines 0.81 0.05 Phrynosonui solare 0.02 Coleuny.x varie^atus 126 the line transect method we used (Bounds. 1996). and may have significant ecological effects. During study planning, special provisions must be made so that all pitfall traps are removed at the end of study, and this is often not done. I am aware of 4 sets of un-tended pitfall traps that were killing reptiles in southern Arizona during 1985-1999. and there were undoubtedly many more that 1 was not aware of. During a study, flooding, animals, and wind may open pitfalls when they are not being used, again leading to mortality. Of great concern, and not generally acknowledged, is predation within the traps. I inadvertently discovered that Sceloporus nuigisier were trap-lining the pitfall grid at OSSA. and eating smaller lizards. In some situations, grasshopper mice (Onycliomys sp.) are similarly pitfall-trapped at remarkably high frequencies, and there is no way to tell what exactly was in the traps with them. They are usually the only live animal present, often leaving no trace of their victims. Since it is likely that trap-liners enter occupied pitfalls deliberately, one cannot readily remove the bias they create. An even less tractable problem, though currently of unknown magnitude, is that larger snakes like whipsnakes {Masticophis spp.) may also trap-line, but since they are not themselves entrapped, the effect is undetectable. For all of these reasons, a pitfall trapping protocol that involves checking traps less than once per day is not very good for accurate quantitative monitoring of live animals. Even with daily, or ideally, twice-daily (sunrise, sunset) trap-checking, there will be some undetected mortality in the traps. An alternative, but not for long-term studies, is to cause all animals that enter the traps to be killed immediately. 127 There are other faults with the way pitfall trapping has been deployed. Typically, the traps are 36 cm deep (i.e., 5 gal buckets), which certain lizards (e.g.. Cnemidophonis biirti ssp.) and rodents (e.g., Dipodomys spp.) can escape from. To make pitfall traps reasonably safe for long-term quantitative monitoring of animals of a variety of sizes, it is necessary to install mesh filters in the bottom of the traps to allow various taxa to avoid each other. Although this is labor-intensive, it can be done, as I have with the drift-fence trap buckets. However, with 5 gal trap buckets, these filters would effectively raise the floor enough to allow individuals of many species to get out. Deeper pitfall traps (50 cm) with mesh filters would be adequate for monitoring, and could also simultaneously yield indices of arthropod abundance. This method would be an important addendum to the EMP. The alternatives to pitfall trapping used at ORPI were drift-fence traps, time- constrained search, and the lizard line transect protocol. The only TCS results presented here are for the nocturnal Coleonyx variegatus. The primary drawbacks of time- constrained search are inter-investigator differences and lack of spatial and temporal control. The line transect method appears to be useful for long-term, landscape-level study, especially in protected areas. In this study, we carefully calibrated the two data collectors (PCR, Charles W. Conner) to ensure that comparable results would be obtained. We also have inspected the distance-from midline results to verify that our ability to see lizards was consistent from person to person and year to year. Lizard transect results might be biased by inter-year variation in maximal lizard activity. However, we did not see cases where this appeared to be a serious effect. We 128 selected our work seasons to coincide with annual activity maxima, and further, we ran the lines twice a year—tending to average out seasonal phenomena. It would take two unusual seasons in which lizard activity remained low to significantly alter the result for the year. Although some activity depression was evident during early 1990, the effect was not apparently strong, and was followed by optimal conditions for activity in the wet summer of 1990. Occasional transect runs under ideal climatic conditions did yield unusually high transect results. This occurred after significant rains that ended prior to sunset of the previous day. allowing the ground surface to dry by the following sunrise. Such days, although rare, combined high humidity, proximity to rainfall, and high air and ground temperatures. Undoubtedly, this source of error is compensated for by various climatically marginal days. Mark-recapture results, which have thus far only been analyzed for Cnemidophorus ligris in 1989 and 1990. support the transect results. They yielded an estimated decline of 30% from 1989 to 1990 at the OSS A site, which is close to the observed change seen in the transect results. A second objection to the transect method is that differences in habitat structure and species behavior make it impossible to accurately compare different environments and divergent species. However, the distance-from-midline data are appropriate within the context of distance sampling theory (Buckland et al.. 1995) to evaluate differences among habitats, assuming the monitored taxon has approximately the same overall aboveground activity time in both habitats. This assumption probably holds roughly for the various abundant diurnal lizards at ORPI. However, results could be enhanced by a study of total aboveground activity of lizards during the time when transect peak values 129 are recorded. This calibration effort should be repeated at appropriate intervals to validate the distance sampling approach. POPULATION RESPONSES TO CLIMATE .AND BIOTIC INTERACTIONS Literature reports suggest that long-term desert lizard population monitoring results may be correlated with rainfall-related arthropod abundance (Whitford and Creusere. 1977: Anderson. 1994). Rainfall-based arthropod blooms (perhaps along with simple moisture availability) produced enhanced reproductive output and recruitment in lizards at ORPI (personal observations). Based only on data through 1991 or 1992 (Rosen and Lowe. 1996). I would have reached conclusions similar to Whitford and Creusere (1977). Longer running long-term studies (Turner et al., 1982: this study, .see Chapter 5) indicate that predation is probably also a major determinant of lizard population fluctuations, and may override the more obvious, beneficent, effects of precipitation. Lizard ecologists have tended to view lizard assemblage structure in terms of food niche partitioning (e.g.. Milstead, 1957, 1977; Turner et al.. 1969: Pianka, 1986; Vitt, 1995; and others). My observations suggest that predation is an important determinant of lizard population size in the Sonoran Desert. The decline of lizard populations during the high rainfall years of the mid-1990's was concomitant with a period of high endothermic predator abundance (Fig. 2.6; see Chapter 3), a correlation that holds up well when snake predation is added and other variables are controlled (see Chapter 5). If predation can have overriding effects on population size, might it also weigh heavily in lizard community structure? Several existing lines of evidence suggest that it 130 ORPl 1.6 o ,-+•, ^ 3 # * « 1.2 •a o .+ .a 0.8 u "s--' (Q T3 (Q 0.4 - - + - - lizard predator abundance (/) —X—lizard abundance on transects 0.0 oo o> o *- CM rt in (O oo 00 00 O) O) O) 05 O) o> o> a> Oi o> o> a> O) o> O) o> 9) O) O) o> T- T- 1.6 3 « 1.2 T3 0) 0.8 CO •o c 0.4 S -X-winter + summer rainfall (/> -X- lizard abundance on transects 0.0 00 0) O CM n U) (O N. 00 00 00 O) o> O) O) a> o> o> O) o> Oi O) a> a> O) O) o> O) 9) 0) O) T- T- T" T- 1— Fig. 2.6. Temporal aspects of the trophic dynamics of the lizard assemblage at ORPl. Lizard values are the total of species means; lizard predator abundance is as defined in Chapter 5; winter + summer rainfall is for all of ORPl. 131 may. (I) Interspecific competition experiments suggest moderate and variable competition between Southwestern lizard species that are very similar (Dunham. 1980; Smith. 1981). but more limited competition in more divergent taxa (Tinkle. 1982; Price et al.. 1993). Interspecific competition may have a less predominant role in continental lizard community dynamics than in insular situations. (2) Intra-specific density dependence is well documented for lizards (Massot et al.. 1992: Tinkle et al., 1993; Turner et a!., 1982). and both predation and energetics appear to play important roles in it (Ferguson and Fox. 1984; Massot et al.. 1992; Turner et al.. 1982). (3) Vitt and Congdon (1978) resolved an important research problem in lizard life history by indicating that foraging energetics and. especially, predation risk (Vitt. 1981; Vitt and Price. 1982). may interact with foraging behavior to lower the clutch size of actively-foraging species. (4) Huey and Pianka (1981) illustrated the susceptibility of actively foraging lizard species to sit-and-wait predation by snakes. The mechanisms underlying these observations may impose divergent predation regimes on lizards that forage differently; and. quite similarly, an analogous mechanism should function more broadly for foraging\basking\social activity place (i.e.. for microhabitat niche in general). Huey and Pianka's (1981) discussion leads me to suggest that active foragers like Cnemidophorus are the kangaroo rats of the desert lizard world. Sensory and locomotor capabilities (speed and hearing in kangaroo rats [Djawdan and Garland. 1988; Vial. 1962]; enhanced chemosensory ability, agility, speed, and high aerobic scope in whiptails [Cooper. 1994; Garland, 1993]) permit them to actively search for, find, and exploit the best food resources, under predation risks that cannot be successfully tolerated by species that 132 instead use sit-and-wait foraging tactics. North American Desert lizard assemblages, including at ORPI, tend to be structured by (I) geographic and macrohabitat replacements among similar species, and (2) broadly divergent patterns of morphology, size, and behavior among the actually syntopic species within communities (note, for example, that lizard communities at ORPI have exactly 1 species/genus [Table 2.1]). We may allow Table 2.1 to define 10 lizard niches for the Sonoran Desert of south-central Arizona, most of which tend to be full at any one site (see also Pianka, 1986). Taxa competent at escape within one of the 10 defined niches would also be relatively efficient foragers within that niche, solidifying their occupancy of it. It would be useful to separate population dynamic and evolutionary causes into components involving (1) energetics and exploitative food competition. (2) aggression and interference, and (3) escape ability and predation risk. Although many lizard assemblages probably are as simple as that at ORPI (see Pianka. 1986; Vitt. 1991; Vitt and Carvalho. 1995). more complex communities have been found in arid Australia and southern Africa (Pianka, 1986) and in wet and (certain) semi-arid Neotropical biotypes (Vitt, 1991, 1995; Vitt and Zani, 1996. 1998a&b). Many lizard assemblages contain congeners and other sets of ecologically similar species that have similar behavior and habitat and food type utilization. Species in these sets, as in North American desert rodents, differ consistently or considerably in size, and consequently in prey size (Magnusson et al., 1985; Vitt, 1991, 1995; Vitt and Carvalho. 1995; Vitt and Zani, 1996), suggesting interspecific competition, and a history of competitive character displacement (Case, 1983) or. presumably, ecological sorting (see Vitt. 1995 and therein). Yet it is 133 important to ask what role predation may play in producing and maintaining such size- structured sets (Dickie et al., 1987). Food size differences are not necessarily the underlying cause of a size-structured assemblage: they might be a result of it. It is entirely possible that lizard size may interact with mortality risk (and its density-dependencies), albeit in more complex ways than it interacts with diet. Population and community effects of predation are less convenient to study than the effects of competition. Density dependence in predation is likely to be expressed as apparent competition for enemy-free space (Holt, 1977, 1984; Holt and Kotler, 1987; Holt and Lawton, 1993; Jeffries and Lawton, 1984, 1985; Kotler and Holt. 1989). and predator-free space may comprise a broad category of effects. For example, low-density populations may capture a niche by not attracting behavioral attention from predators. At higher densities, prey behavior may be forced to contract to truly "safe" (but numerically limited) physical refugia—micro-environments at the core of the niche. At high prey densities, predator population increases may become entrained, leading to classical predator-prey scenarios. Which species entrain such increased predation pressure, and which ones can survive it. is not going to become clear from the kinds of data we normally collect for lizards. Vitt (1995) was among the first to present a systematic survey of escape behavior for an entire lizard assemblage, although other important efforts have been made to quantify escape behavior (Schall and Pianka, 1980; Price, 1992; Paulissen, 1994, 1995). I present a broad qualitative survey focusing on anti-predator tactics at ORPI (Table 2.7). The table suggests that the 10 niche groups have distinctive or unique traits, many 134 Table 2.7. Special anti-predator tactics in the two major desert lizard communities at ORPI. Organization of this table parallels saurofaunal structure shown in Table 2.1. The body of the table assumes (1) crypsis is important for all ta.\a, and (2) all retreat into burrows, tfee-holes, crevices, or under rocks, according to availability: added details are given when this behavior is strongly marked and specific in some way. ROrK.SI.OPF VAI.l.RY I. terrestrial ant-eating species (homed lizards) (niche vacant) motionlessness, spiney armor, distasteful blood-squirting II. small terrestrial insectivnre (side-hlotched lizard') dodges among steins of sub-shrubs, dodges among stems of sub-shrub and among rock obstacles 111, small arboreal and scansorial insectivnre (tree / hmsh lizards) hides behiiui rocks\irunks. hides behind trunks, ascends into foliage ascends into foliage IV. medium-sized, terrestrial. wait-Jash insectivore (zehra-tailed lizard) (niche vacant) tail-signalling, indcating escape ability. very fast running, followed by motionlessness V. large arboreal insectovore (spinv lizards) hides behind rocks, trunks, hides behind trunks, immediate flight into tree- ascends into foliage holes or descent into woodrat nests VI. large terrestrial lizard\arthropod-eating species (crntaphvtines) special tactics not known runs fast and freezes at shrub-base; bites, hisses VII. large terrestrial herbivore (chuckwalla / desert iguana) high vigilance, active in extreme heat, activity during extreme heat, immediate flight into crevices, inflates in crevices immediate submersion in burrow VIIl. small, nocturnal, terrestrial insectivnre (gecko) same as for valley habitat quick flight + motionlessness. (habitat differences not known) scorpion mimicry IX. medium-sized active-foraging, terrestrial insectivore (whiptails) evasive action in shrub or rock complexes, evasive action in medium-sized shrubs while remaining active or thickets, while remaining active X. large camivnrous nestling predator (Ctila monster) aposematic, painful venomous bite, aposematic, painfid venomous bite, osteodermal armor, shadow-mimicry in crevices osteodermal armor 135 specially for avoiding predation. We might categorize the "survival niches" at ORPI as (I) horned lizard suite. (II) subshrub dodging among small stems, (III) trunk crypsis and flight into foliage, (IV) rapid running on open ground, with "dazzle effect" (Norris, 1967) at high solar intensity. (V) trunk hiding and rapid flight into hard refugia. (VI) generalized crevice or burrow escape, plus biting, (VII) high vigilance with rapid, and/or specialized shelter use, (VIII) running burst with mimicry, followed by immediate, cryptic motionlessness. (IX) dodging in and among medium-sized obstacles, and (X) defensive venom (i.e.. active defense; see Lowe et al.. 1986). Further, most species have strong, background-specific crypsis or crypsis combined with aposematism (Norris and Lowe. 1964; Norris. 1967). Schall and Pianka (1980) have suggested that there may be some connection between modes of predator escape and assemblage structure among sympatric whiptail lizards. Lizard ecology might benefit from development and application of methods for quantifying the survival niche. 136 III. ENDOTHERMIC PREDATOR TRENDS, WITH A LITERATURE-BASED ESTIMATE OF DIET CHAPTER SUMMARY Endothermic predators were monitored in a desert valley community at Organ Pipe Cactus National Monument (ORPI) in conjunction with monitoring of prey (lizards and small mammals) and other predators (snakes). The endothermic predator assemblage included 17 species observed during primarily herpetological fieldwork totalling 650 work-days, at an average of 1.002 visual observations/day. Owls were also monitored by the frequency of calling. Coyotes Canis hitrans and red-tailed hawks Biiteo jamaicensis were the most frequently seen species; great homed owls Bubo virginianits. Harris' hawks Parabiiieo iinicinciiis, kit foxes Vulpes macrotis, roadrunners Geococcyx califoniiciniis. kestrels Falco span'erius. and screech-owls Otiis kermicottii were also abundant. The overall predator observation rate decreased to a minimum during a drought early in the study, and increased steadily during a wet period dominated by winter rainfall beginning in 1991. It reached a high point in 1995. declined again during drought in 1996-7, and increased during the wet winter-wet summer year of 1998. Annual endothermic predator abundance (as represented by observation rate) varied by a factor of 3 to 4, and smaller taxa appeared to increase more rapidly post-drought than larger species. The literature was used to construct an estimate of predator diets at ORPI. Data 137 from arid and semi-arid sites in Western North America were collected, averaged, and converted to local conditions by generic-level deletion of prey categories not available in the study area. Smaller predator taxa represented a sub-guild feeding on arthropods, lizards, and perognathine rodents (pocket mice, Perognatlius, Chaetodipiis). Larger taxa appear to feed most heavily on kangaroo rats, larger rodents, and lagomorphs: some of them added birds, arthropods, and reptiles to their menus. The bam owl specializes on mammals, intermediate between the smaller and larger taxa, and feeds especially heavily on the smallest taxa of rodents. The small-prey sub-guild reached an abundance peak in 1992-3 in my observations, while the large-prey sub-guild increased less abruptly and reached a high point in 1994-5. The temporal flux in amount and character of predation pressure could be important elements of niche space for predators and prey. INTRODUCTION Predation is a sudden and fast interaction—the most immediately potent and intense ecological interaction of all. It is difficult to observe and study, especially with vertebrate predators, which avoid us and occur at low densities to begin with. Ecology has tended to focus on competition as an organizing and driving force, and experimental ecology has focused on the successful demonstration of inter-specific competition (Schoener, 1983). Reviews of this experimental work found that both intraspecific competition and predation were being under-represented (Sih et al., 1985; Connell, 1983). In some systems, the effect of predation was found to be of key significance for species diversity, community suoicture, and community dynamics (Paine, 1966; see 138 Connell. 1975). Modem ecology still lacks a thorough understanding of importance of predation. Information relevant to the effects of predation must accumulate slowly—through dietary studies, models, accumulation of time-series data, and ultimately, through field experiments. Predation systems are often too spatially extensive or trophically complex for the kinds of experiments usually done by ecologists, and this has been a hindrance, at paradigmatic odds with the flowering of ecology as an experimental science. In recent years, vertebrate predation has nevertheless received rapidly increasing attention from empiricists, modelers, and experimentalists with a firm grounding in optimal foraging theory (see e.g.. Kotler. 1984; Hansson. 1987; Hanski etal.. 1991; Jaksic et al.. 1993; Marti et al.. 1993; J.S. Brown et al.. 1994a; Kotler et al.. 1994; Meserve et al.. 1996; Abramsky et al., 1998). Yet. predation's roles in population regulation and community organization remain controversial (Erlinge et al.. 1984, 1988; Hanski et al., 1993; J^drzejewski et al., 1996). Time lags, "chaos" (or complex dynamics), intra-trophic level predation. and even the old question of whether predators do more than remove the "surplus" (Errington. 1946), complicate our thinking about the Importance and function of predation. Established altitudes persist that individuals killed by predators are those "least likely to survive and reproduce", especially "the young, the homeless, the sick and the decrepit", and thus, "the effects of predation on the prey population will be far less than might be expected" (Begon et al., 1990). As ecology matures, complexities of competition and behavior are being explored (e.g, Rosenzwelg and Abramsky, 1997), and the relevance of 139 predation. even in systems long-viewed under the competition paradigm, is being prominently recognized (e.g., Kotler et al., 1994; Abramsky et al., 1998). More practically, the devastating effects of introduced predators across the globe (see Mooney and Drake, 1986; Lever, 1994; LaRoe et al., 1995) have become clear. Recent work building on long-running population data supports the importance of predation in such familiar systems as the lynx-hare, the vole cycle, and the moose-wolf-bear (see Introduction to the Dissertation). My work in the Sonoran Desert includes monitoring of small mammals (Chapter 1), lizards (Chapter 2), and snakes (Chapter 4). All of these are prey for endothermic carnivores, and snakes are also potential competitors. My objectives are to construct time- series for climate, prey, and predators and to search for climatic and biotic factors necessary for interpretation of long-term population monitoring results. My observations on endothermic predators show the broad outlines of temporal flux through a series of wet-dry climatic sequences from 1983-1998. The primary objective of this chapter is to present the temporal trends of endothermic predators for use elsewhere in quantitative analysis of prey monitoring results. Although an emerging trend in ecology views predatory effects as strong, non- lethal effects modulated via determination of prey behavior (Lima. 1998), Kotler et al. (1994) hypothesized that mortality from predation might also directly structure assemblages. If predation varies over time in intensity and kind, and potentially competing prey taxa vary markedly in exposure to risks and ability to cope with different states of the time-variant predation risk, coexistence of competitors can be promoted 140 directly through the population dynamic effects of mortality. Although J.S. Brown (1989) and Kotler et al. (1994) focus on seasonal variability and, analogously, spatial- and habitat-variance in predation pressure, in this paper I consider variation over a period of years. The data I present here suggest that there was important variation at this time scale in the character of predation facing small vertebrate animals as effects of a climatic wave rippled through the Sonoran Desert ecosystem. METHODS STUDY ARE.A The study area was centered at ORPI, a 1300 km" re.serve in western Pima County. Arizona, and includes data collected up to 10 km north of the park. The area is in the northern Sonoran Desert, and is transitional between the Arizona Upland and Lower Colorado Valley subdivisions of this desert. It supports a diverse and lush desert vegetation with (xeroriparian) thomscrub formations arranged along linear arroyos that form a dense network in the matrix of (upland) desertscrub. Short-stature desertscrub often occurs in an open woodland of desert shrubtrees 3-5 m tall. The flora and much of the fauna show a sharp transition at the inflection point between rocky desert slopes and gravelly-sandy-silty bajada plains. Surface water is scarce but widely distributed. The Ajo Mountains, on the E edge of the study area, and the centrally located Puerto Blanco Mountains each contain one tiny perennial water source. Quitobaquito Springs, in the SW. is a substantial oasis, and is about 3 km from a major perennial stream in Mexico, Rio Sonoyta, which parallels the S 141 monument boundary at a distance of 2-5 km. There is water at the centrally-located headquarters-campground complex, at 2 semi-perennial cattle-watering ponds within 10 km of the north boundary of ORPI. and one just W of the W boundary. At present, then, water is available within 13 km or less over most of the area, presumably well within the range of many raptorial birds and larger mammalian predators. Under natural conditions the distance would be greater, and animals in the Valley of Ajo, where most of the work was conducted, would have had to cross rugged mountains to reach Rio Sonoyta, or risk the confined approaches to water in rocky sites in the Puerto Blanco or Ajo Mountains. Some predator populations may be enhanced by such increased water availability, which now exists throughout the desert (see Broyles, 1997). This study was conducted mostly on the bajadas and valley floors at 350 - 700 m elevation (annual precipitation average 18-23 cm). Work at or near riparian habitat was excluded from the monitoring dataset, to avoid adding variance and bias. More detailed descriptions of climate, vegetation, floristics, microhabitat, and macrohabitat are available in Chapters 1, 2, and 4. MONITORING METHODS I recorded every visually-observed predator during days of routine herpetological fieldwork in desert (non-riparian) work at ORPI. Routine activities included low-speed road-cruising (Rosen and Lowe, 1994) on paved and dirt roads, walking at night with a headlamp looking for snakes, walking a 3-5 km route checking snake traps during daylight (Chapter 4), running lizard line transects (Chapter 2). and occasionally radio- 142 tracking snakes or installing equipment. I made no concerted or special efforts to locate or observe endothermic predators during this work. When I saw or heard them. I attempted to observe them and confirm their identification. 1 did not alter my search behavior or the visual focus of my attention significantly during the years of study, and my distance vision did not change measurably during the study. Observation activity was adjusted to correspond to temperature conditions conducive to snake activity, generally 20-40"C air temperature, and tended to occur disproportionately around the time the new moon. Almost all of the work was done during the warm season (April-October), and August-October was the most consistently worked seasonal period. A day was defined as any 8-12 hr of tleld activity during one 24 hour period. The observation units were tleld trips that lasted from 1 to 30 days; fractional days were summed in quarters, by considering the morning and afternoon-evening activity periods each as one half day. and a partial period as one quarter. Daily fieldwork activity for full days started between 0600 and 0800 hr, and morning fieldwork generally extended to at least 1100 hr. Afternoon-evening activity commenced between 1600 and 1900 hr. and generally continued until 2330 to 0230 hr. Observation rates were calculated for each field trip after deducting all days and portions of days when routine work was not conducted. The mean and variance for each year was computed by treating each trip as a single datum.. Owls were abundant but rarely seen during this study. Therefore, I also kept records of aural observations of the two numerically dominant vertebrate-eating species, the great homed owl (Bubo virginianus) and the western screech-owl (Otus kennicottii). 143 During 1988-1993 the work schedule frequently led me to sleep near the Intensive study area from 2400 or 0300 to 0545 or 0730, and for about 30 min prior to retiring I counted the number of owls calling. Generally, nights were still enough that air movement did not vary the distance at which animal sounds would be audible. Automobiles moving across a highway cattle guard at 55-85 mph were unmistakably audible to a distance of 2.9 km. coyotes to a presumed distance of 3-4 km. and owls to perhaps 0.5-1 km. Some abundant endothermic predator species were probably significantly under- observed. First. I did relatively little work in the Ajo Mountains, where golden eagles and felids are most abundant. Second, little of my work involved cool-season surveys, so wintering species of hawks were under-represented. Third, species likely to be detected only at a distance, such as larger falcons, would largely remain unobserved with the close-in focus of most herpetological fieldwork. ESTIMATION OF DIETS FROM THE LITERATURE I collected quantitative dietary data for predators at ORPI from the literature on Western North America. I focused first on studies from the Southwest that were conducted in desert, thomscrub, and desert grassland. In this sampling, taxonomic coverage was incomplete and sample sizes were limited, even for species of great importance. Therefore, I also included published diets from arid and semi-arid habitats elsewhere in Western North America. I separated the resulting data table into two categories: (I) arid and semi-arid Southwest (including most of Arizona, arid southern California and southwestern New 144 Mexico, all of Baja California, and most of northwestern Mexico), and (2) all other arid and semi-arid areas of Western North America (including the Great Basin, the central valley and associated foothills of California, and western Texas and Oklahoma). I did not utilize studies that appeared (based on habitat and prey spectrum) to pertain in substantial part to wetland, riparian, or forest environments. I summarized and reduced the data (Tables 3.4-3.6) utilizing prey categories useful for comparison to possible diets in the ORPI valley ecosystem. In many cases, this required me to re-compute the results based on the data presented, and the only consistently possible way was by frequency of occurrence. This presents several complications that must be recognized as limitations of the resultant tables. First of all. reporting of plant material varied in two ways that bias the result: (1) some authors may have overlooked plants, and (2) other authors varied the ta.xonomic level of the plant categories presented. Problem 2 also occurs for invertebrates, and is least significant for vertebrates in the diet. Problem 2 is important because almost all studies reported the occurrence of an item (i.e.. "plant material") as its percentage occurrence in a sample of discrete pellets, fecal samples, or digestive tracts. These are composite samples: if an author reported "vegetation" as the category, the total maximum total possible would be 100% frequency, but when several classes of plant material were included, the summed frequency could theoretically range to 100% times the number of subcategories. Therefore I modified the second diet table I derived (Table 3.5) to include plants separately, with the remaining categories normalized to 100%. I ignored similar, but smaller, problems with arthropods 145 and rodents. Rules for construction and interpretation of Tables 3.5 and 3.6, the index for expected diets at ORPI, were as follows. For plant material. I attempted to utilize a frequency based on the occurrence of the overall category, rather than the sum of sub categories; this was not possible for the Smith and Danner (1990) report, although generally it appeared that the occurrence of multiple plant taxa per scat or regurgitation pellet was infrequent. The bias is limited enough that the plant material values remain informative. For arthropods, this was not the case: methods varied, and precise interpretation of any single study would require considering it separately. However, it appeared that in the data I combined for each individual predator taxon, the mean results for arthropod frequency did not incorporate compounding or otherwise serious biases. For rodents, it was often necessary to lump unidentified rodents into the class "other small mammals", which thus contains larger small rodents (i.e.. cotton rats) as well as the rarely represented larger kinds (e.g.. muskrats). In the most such cases, the reported frequency of rodents is therefore low. and thus the type of biases discussed above can be presumed, and the true frequency of rodents is slightly underestimated. The category "reptiles" in my tables does not include the lizards and snakes that are separately tabled. However, the category "other endothermic predators" includes some predatory birds that are also represented again in the category "birds and bird eggs": the number of these was so small that the tabled values for birds were unaffected. I made no attempt to distinguish carrion from large herbivores ("hoofed animals") because my perception is that some authors were inaccurately hedging against the conclusion of 146 "harmful" affects of predators, and because it was often simply impossible to distinguish what the authors meant exactly. For the raven, the distinction between carrion and the other tabled values is highly problematic. This species watches the highway like a vulture; Camp et al.'s (1993) work suggests that ravens are capturing small birds and lizards in numbers, whereas recent results suggest that far more of these animals are available as roadkills than previously suspected [Kline and Swann, 1998; D. Swann, personal communication]). The extent of predation by ravens is not well-known, and the values in the table should be viewed as tentative. To obtain the estimate of predator diets in the ORPI valley ecosystem (Table 3.6). I deleted those categories not likely to be significant there (i.e., other endothermic predators, small murids and microtines, geomyids. other small mammals). The remaining results were normalized to 100%, and this result was used as the estimate of diet. With respect to the vertebrate diet categories especially, this procedure should have largely removed the biases discussed above. RESULTS COMPOSITION AND STRUCTURE OF THE FAUNA I saw a total of 651 individual endothermic predators of 19 species during 650 included days of herpetological work in the desert environment at ORPI (Tables 3.1 and 3.2). Predators were more abundant and diverse at the Quitobaquito Springs riparian area than in desertscrub. Species absent or rarely seen elsewhere on the monument, including herons, egrets, kingfishers and bird-eating hawks, were abundant at this oasis, as were 147 Table 3.1. Enodothermic predators of vertebrates visually observed during tul (8-12 hr) days of routine herpetological field work in desertscrub habitats, ORPI, Arizona. Most work was in the warm season, at times suitable for reptile activity, in valley habitat. YEAR: TAXON 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 total coyote I 4 8 16 15 6 41 3 21 8 4 2 129 kit fox 5 4 7 2 4 6 13 0 1 6 0 9 57 gray fo.x I 0 3 2 3 0 4 0 4 0 0 0 17 "fox" 0 0 0 -> 1 0 2 0 0 0 0 0 5 "canid" 0 0 1 0 2 0 1 0 7 0 0 0 11 bobcat 0 1 0 1 0 0 0 0 0 0 0 I 3 mountain lion 0 1 0 0 0 0 0 0 0 0 0 0 1 striped skunk 0 0 1 1 0 0 0 0 0 0 0 1 3 spotted skunk 0 0 1 0 0 0 0 0 0 0 0 0 I badger 0 0 1 0 0 0 0 0 3 0 0 0 4 "carnivore" 0 1 1 1 1 0 I 0 T 1 0 0 8 American kestrel 2 4 3 4 9 1 14 0 1 5 0 0 43 Coopers hawk I 0 I 1 2 0 3 0 0 0 1 0 9 northern harrier 0 0 I 0 3 1 0 0 0 0 0 0 5 red-tailed hawk 8 3 17 16 13 15 33 2 13 3 t 4 129 Harris' hawk 0 4 6 3 13 3 8 0 7 1 1 1 47 "buteo" 1 2 0 2 2 5 3 0 5 0 0 0 20 "hawk" 0 0 0 1 I 0 1 0 0 0 0 0 3 golden eagle 0 0 0 0 I 0 0 0 0 0 0 0 1 greater roadrunner 4 0 5 5 7 20 25 0 4 2 1 0 73 great horned owl 1 5 4 4 5 4 7 0 1 2 0 0 33 western screech-owl 1 0 0 0 0 0 3 0 3 1 0 1 9 bam owl 2 2 I 2 1 1 3 0 0 0 0 1 13 "owl- I 0 2 0 2 1 1 0 1 0 0 0 8 loggerhead shrike 0 0 3 1 4 2 5 1 0 1 1 1 19 sum 28 31 66 64 89 65 168 6 73 30 10 21 651 TOTAL DAYS 20 62 104 128 87 53 100 3.5 38 20 17 20 650 mean per day 1.44 0.50 0.64 0.50 1.02 1.24 1.69 1.71 1.93 1.48 0.61 1.05 1.00 SE (among trips) 0.50 O.ll 0.07 0.08 0.07 0.26 0.21 2.30 0.22 0.30 O.ll 0.21 148 Table 3.2. Rate of observation (number seen per day) of different taxa of endothermic predators at ORPI, based on data in Table 3.1. YEAR: YR 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 overall coyote .05 .06 .08 .13 .17 .11 .41 .86 .56 .40 .24 .10 0.20 kit fox .26 .06 .07 .02 .05 .11 .13 .00 .03 .30 .00 .45 0.09 gray fox .05 .00 .03 .02 .03 .00 .04 .00 .11 .00 .00 .00 0.03 "fox" .00 .00 .00 .02 .01 .00 .02 .00 .00 .00 .00 .00 0.01 "canid" .00 .00 .01 .00 .02 .00 .01 .00 .19 .00 .00 .00 0.02 bobcat .00 .02 .00 .01 .00 .00 .00 .00 .00 .00 .00 .05 0.00 mountain lion .00 .02 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 0.00 striped skunk .00 .00 .01 .01 .00 .00 .00 .00 .00 .00 .00 .05 0.00 spotted skunk .00 .00 .01 .00 .00 .00 .00 .00 .00 .00 .00 .00 0.00 badger .00 .00 .01 .00 .00 .00 .00 .00 .08 .00 .00 .00 0.01 "carnivore" .00 .02 .01 .01 .01 .00 .01 .00 .05 .05 .00 .00 0.01 American kestrel .10 .06 .03 .03 .10 .02 .14 .00 .03 .25 .00 .00 0.07 Coopers hawk .05 .00 .01 .01 .02 .00 .03 .00 .00 .00 .06 .00 0.01 northern harrier .00 .00 .01 .00 .03 .02 .00 .00 .00 .00 .00 .00 0.01 red-tailed hawk .41 .05 .16 .13 .15 .29 .33 .57 .34 .15 .12 .20 0.20 Harris' hawk .00 .06 .06 .02 .15 .06 .08 .00 .19 .05 .06 .05 0.07 "buteo" .05 .03 .00 .02 .02 .10 .03 .00 .13 .00 .00 .00 0.03 "hawk" .00 .00 .00 .01 .01 .00 .01 .00 .00 .00 .00 .00 0.00 golden eagle .00 .00 .00 .00 .01 .00 .00 .00 .00 .00 .00 .00 0.00 greater roadrunner .21 .00 .05 .04 .08 .38 .25 .00 .11 .10 .06 .00 0.11 great homed owl .05 .08 .04 .03 .06 .08 .07 .00 .03 .10 .00 .00 0.05 w. screech-owl .05 .00 .00 .00 .00 .00 .03 .00 .08 .05 .00 .05 0.01 bam owl .10 .03 .01 .02 .01 .02 .03 .00 .00 .00 .00 .05 0.02 "owl" .05 .00 .02 .00 .02 .02 .01 .00 .03 .00 .00 .00 0.01 loggerhead shrike .00 .00 .03 .01 .05 .04 .05 .29 .00 .05 .06 .05 0.03 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 overall small-prey taxa .51 .10 .14 .11 .26 .46 .53 .29 .21 .44 .18 .20 .26 large-prey taxa .97 .40 .50 .40 .78 .78 1.19 1.43 1.72 1.04 .48 .85 .75 149 many of the taxa common elsewhere on the monument. The riparian species, and certain others recorded rarely and at mesic sites, including raccoons and striped and hognosed skunks, are presumably present as populations along Ri'o Sonoyta. Owls were far more abundant than indicated by the visual records in Table 3.1, and I noted individuals 313 times during the nocturnal auditory monitoring during 1988- 1993 alone (Table 3.3). In addition to the species monitored. I saw many common ravens, particularly as they flew along the pavement competing with me for roadkill; I heard ferruginous pygmy owls on three occasions in 1988; and saw ringtails as well as sharp- shinned hawks during work not included in the monitoring. Various other predatory species, especially raptors, appear briefly and sporadically at ORPI (see Groschupf et al.. 1988), but are probably not significant for ecosystem dynamics. Like other aspects of the ORPI biota, the endothermic predator assemblage differed between rockpile and valley habitats. Species usually thought of as "top predators", the golden eagle and mountain lion, occurred primarily in larger rockpile habitats, especially including the less .xeric canyons of the Ajo Mountains. Top predators in valley habitats were thus presumably the coyote and great homed owl. Also characteristic of rockpile environments here and in similar Sonoran Desert areas, but not seen by me in the valleys (see also Cockrum and Petryszyn, 1986), were ringtails and western spotted skunks. Bobcats were far more prominent in the rocks than in valley settings (B. Henry, pers. comm.; pers. obs.; see also Cockrum and Petryszyn, 1982). Landscape partitioning by the fauna is undoubtedly more dramatic than described here: most of the work~and approximately 99% of the records in Table 3.1—pertains to valley Table 3.3. Results of informal monitoring of owl sounds at night, in valley deserlscmb habitat at ORPI. No monitoring was done in 1994-5.1 did not know the sounds made by barn owls until 1991; thus, although they were not abundant until 1993, they may have been present in small numbers during 1988-1990. # heard/night when owls recorded: # heard/night; average (index): YR QUQ WS-O QUO WS-O GHO WS-O # barn owls heard 1988 0.86 1.05 0.59 0.71 0.73 0.88 no data 1989 1.23 0.48 0.69 0.24 0.96 0.36 no data 1990 0.97 0.25 0.55 0.05 0.76 0.15 no data 1991 1.53 0.12 1.00 0.09 1.27 0.11 0 1992 0.81 0.92 0.38 0.44 0.60 0.68 1 1993 0.67 0.98 0.64 0.64 0.66 0.81 9 151 habitat, or to low rocky hills traversed by the paved highway. TEMPORAL TRENDS OF PREDATOR ABUNDANCE My first work on-site was in May, 1983. Predators seemed to be abundant during 1983-1985, at the end of the wettest period in the 1949-present ORPI weather record. At that time, I saw predators frequently, including species such as the badger, bam owl. and roadrunner, which were not widespread and abundant again until the mid-1990's. Thus, the decline from 1987 to low levels during 1988-1991 (Table 3.1, Fig. 3.1) probably represents the terminus of a foregoing period of elevated predator pressure. Following the late-1980's drought, endothermic predators showed a factorial increase of 3-4 fold to a high point in 1993-1995 (Fig. 3.1). There was a subsequent decline during the drought years of 1996-1997, and apparently a rapid response to the good winter-summer rainfall sequence of 1998. Thus, there was a long increase of predation pressure starting in 1991. and reaching its high point no earlier than 1993, and more likely in 1995 (Fig. 3.1). The ta.xon-specific pattern of increases also suggests that major groups, and most individual species, increased during the high winter-rainfall years of the early-mid 1990's (Table 3.2. Fig. 3.2). Certain prominent predators appeared to fluctuate markedly in both presence and abundance. The greater roadrunner, which by 1988-9 was largely localized in the largest, richest xeroriparian habitats, was abundant throughout the Valley of the Ajo by 1992 (see Fig. 3.2). Similarly, bam owls were rarely detected during the dry years, but were frequently heard during auditory monitoring in 1993 (Table 3.3). The badger, which was rarely seen, established residency on the intensive study site during 1994-5 152 ENDOTHERMIC PREDATORS OF VERTEBRATES, ORPI UJ 1.0- I I I I I I I I I 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 Fig. 3.1. Observation rate of endothermic predators of vertebrates at ORPI, based on Table 3.1. Graph shows mean + 1 SE, with arbitrarily smoothed trend line. 1.0 •canid 0.8 •owl — - buteo > < kestrel & shrike Q 0.6 roadrunner lU UJ 0.4 (0 0.2 0.0 O ^ CM in (o CO O) O) O) o> o> O) O) O) 0> ^ o> o> Fig. 3.2. Visual observations of endothermic predators at ORPI, from Tables 3.1 and 3.2. Arbitrarily smoothed trend lines are shown for taxa with large sample sizes. 153 (first detected by scent, in 1993), digging extensively and thoroughly over many hectares, and was not seen there at any other time. Bobcats were elusive, and not often observed; the only one I saw on the intensive study area was in 1998. However, I saw 8 during 1993-1996 in the northern Sonoran Desert, but only one earlier (1988) and thereafter (1998). Bobcats are normally prominent in the rocks or along major, exceptionally productive arroyos. I found them in valley habitats in 1988, 1996, and 1998. The data also indicate (Table 3.2, Fig. 3.2) that dominant consumers reached peak abundances during 1993-5. Canids and buteos reached peak levels during 1993-1995. with highest values recorded in 1995. Observations of foxes were most frequent early and late in the study (Table 3.2), but the accuracy of this is questionable because an individual den near a road transect, and individuals seen repeatedly during a series of consecutive nights at favored hunting grounds, added a strong random error to the temporal sequence. Although owls were too rarely seen to provide an accurate picture of their population trend, the auditory data through 1993 are more useful. The great horned owl. along with the coyote, was the most consistently detectable endothermic predator in the intensive study region, and auditory results suggest that it's population was less variable than other predators. It was at a high level early in the rainy period, and its highest detection rate was during 1991. In contrast, the western screech-owl population was almost undetectable during the 1989-1991 drought-associated trough, but achieved high levels in 1992-3 (Table 3.3). Overall, owl predation appeared to increased slowly during the mid- 1990's to reach highest levels during 1992-1996 (Fig. 3.2). 154 DIET CHARACTERISTICS OF PREDATORS My literature search yielded 24 species reports from the arid Southwest region and 23 from elsewhere in the arid West, with quantitative data on diet for each of the 17 endothermic predator species I observed in the valley ecosystem (Tables 3.4. 3.5). The 47 diets are impressive in scope and sampling effort, yet the data are barely adequate for several species, particularly for specific prey categories, including in some key predator species. The dataset indicates that endothermic predators in the Southwestern deserts utilize more lagomorphs, sciurids, Neotomci, and birds, and less Dipodomys, perognathines, geomyids. and small murids and microtines than the same predators do elsewhere in the West (Table 3.5). Although the focus of the present paper is on camivory, it is notable that most of the mammals include a considerable fraction of plant material in their diets. In many cases the material is substantial and nutritious, and at ORPI coyotes utilized mesquite beans heavily in some seasons (pers. obs.). Similarly, the contribution of arthropods to the diets of the predator taxa is not insignificant, especially since many are a.s large as the giant hairy scorpion {Hadnirus sp.). In cases in Table 3.4 with arthropods recorded at 20% or more, they probably form an energetically significant portion of the diet, despite their low average weight per item. The estimated diet for the Sonoran Desert valley fauna (Table 3.6) indicates that larger taxa tend to eat larger prey (Rosenzweig, 1966). but with some notable exceptions. Canids, felids, and buteos were computed to use larger prey types, with limited representation of smaller prey types, although the gray fox, which is less common than Table 3.4A. Nociunial endulherinic predaiors in valley liahiial ai ORPI: literalure dala on llic diei in Wesicrn North America. Tabled values are % frequency given by or computed from listed sources. Further details arc in text. List of sources follows Table 3.4B. O lA gco-S o 3 c ib-E -a SPECIES locale N j t I ^ ^ j K j j EE 1 jlll ^lg-£ badger ID link 3.1 38 8 3,5 0,6 9,5 0.4 1,2 0,9 27,7 0,0 19 bobcat AZ 237 2.3 0.8 27.8 4 2 23.6 2,1 3.8 24.0 10,6 0,4 12 bobcat AZ 79 0.0 48.1 13.9 12.7 1,3 0,0 2,5 0.0 0,0 21,5 26 coyote AZ 1157 III 33.8 12.0 7,9 1,5 33,8 2H coyote AZ link 1.8 12.8 17.3 2,4 1,0 52.4 0.9 20 coyote AZ 418 5.3 16.8 5.5 9 6 4,3 1,2 2,2 0.0 4,8 50.5 26 coyote CA 2250 1.4 0.1 16.8 18.9 7.2 16.0 2,8 1,7 10.4 26 2,3 5,5 1,0 0.1 11,3 2.3 9 coyote ID link 20.7 12.8 9,6 1,7 12,8 19 I.I 0.6 18,8 19 gray fox AZ 660 4.4 31.8 25.6 4,7 32.8 20 gray fox AZ 561 0 5 7.1 1.8 8,9 0,0 0.4 2.7 0.0 15,7 58.8 26 gray fox UT 747 4.8 0.0 0.0 13 2.3 0.0 0,0 2.8 2.3 2.3 1,6 00 1.5 0.0 32,5 47.9 29 gray fox SW 32S 1.2 0.3 3.0 4.6 4.9 4.0 11,3 3,7 0.6 9,4 1.8 3.4 0.0 25.0 27.1 30 kit fox AZ 614 0.2 0.2 147 11.9 9.4 7.0 3,1 2.6 2.8 10,4 0.2 0.3 0.0 9.3 27.8 7 bam owl AZ JJ4 0.0 00 11.0 19.0 18,0 34.0 60 8.0 0.0 22 bani owl CA 5U 0.0 0.0 0.8 0 2 1.0 6.8 43,5 8.8 37.4 0,0 0,2 08 0.6 0,0 0.0 8 bam owl CO 4366 0.0 0.0 1.5 0.0 0,1 4.0 7,3 78.2 7.3 0.2 1,4 00 0.0 0,0 0,0 0.0 18 bam owl ID link 0.2 0.1 7.9 6,8 76.4 5.7 0,1 19 bam owl OK link 0.0 0.0 0.0 0.0 3 9 0.0 42,1 41.0 II 11.8 0,0 00 0.0 0,0 0,0 0.0 1 bam owl VVA 262a 0.0 0.0 0.0 00 0,1 0.5 63,8 27.9 5.4 0,1 0,7 0,0 0.0 0,0 1,6 0.0 13 elf owl AZ link 0 5 10 98,5 14 great horned owl AZ link 34.0 36,0 28,0 0.0 2.0 5 great horned owl CA 1471 0.0 0.4 13.9 1.0 16,9 13,7 14 3.2 7.8 05 3 1 2,6 0.4 1.6 34,2 00 8 great homed owl CO 228H 0.0 0.2 14.0 0.2 0,3 2,2 1,9 660 8.4 0 5 4,5 0,0 0.0 0,3 1,6 0.3 18 great horned owl ID link 7.0 0.5 20,2 4.1 44.9 3.6 5,9 19 great homed owl MEX 115 0.0 00 24.4 0.0 3,5 0,9 11.3 0,9 0.0 6,1 5,2 8,4 7.3 0,0 23,5 0.0 15 great homed owl NEX 4940 6.1 13 2 16,0 I.I 12 4 1,6 4,3 36.2 0.0 7.0 6 great horned owl tJT 173 0.0 0.0 38.7 2.3 1,2 39,3 00 1,2 00 0,0 2,9 0,0 0.0 00 6,9 0.0 27 great homed owl WA H72 0,0 10 1.3 0.4 0,3 1,4 43,1 37.1 0 1 00 2,8 0 3 00 0,1 5.4 0.0 13 w. screech-owl ID link 0 1 00 2,3 12,3 54 3 1.0 10 2 4,5 19 w, scrcech-owl W N.A 604 0.0 00 0.0 0.0 0,0 0,0 1,2 1.3 0.0 00 0,8 0,3 1.0 0,0 95,2 0.0 24 Table 3.4B. Diurnal endothermic predators in valley habitat al ORPI: literature data on ihe diet in Western Nonh America. (U (/: E (A Ui c T3 'O B- 2 ^ •§ r3 ^ tuj eg O U CS S E s s V D- O 5 2 5 •§ 00 E ID ^ tc u- « E Clj ^ E x: O 0) O -O o o o ^ E •o C3 x: C *3 b > CJj (J S E b '''•I C cj a. s* SPECIES ItKale N O Q. ^C3 c. wj E E 15 C eg O- a>w cto Harris' hawk AZ 641 0.0 4.1 33.9 13.6 11.8 1.0 0.2 0.0 0.3 0.3 27.9 0.0 11.0 0.0 0.0 0.0 31 Harris' hawk AZ 160 0.0 0.6 46.3 14.4 12.5 0.0 0.0 0.0 0.0 0.6 17.5 0.0 8.7 0.0 0.0 0.0 16 Harris' hawk NM 889 0.0 1.1 63.0 4.4 5.4 2.8 0.0 0.1 0.4 0.1 10.3 0.0 3.7 0.0 10.0 0.0 2 red-lailed hawk AZ 55 0.0 0.0 36.4 16.4 1.8 0.0 3.6 1.8 1.8 0.0 0.0 27.2 11.0 0.0 0.0 0.0 17 red-lailed hawk AZ 326 0.0 0.2 37.6 12.2 2.5 1.1 0.0 0.4 0.2 0.0 13.8 6.2 8.6 0.0 1.2 0.0 21 red-tailed hawk CA 3688 0.9 0.3 8.7 28.6 1.6 0.9 0.3 0.2 21.5 0.2 4.6 10.4 12.8 0.0 4.4 0.0 10 red-tailed hawk ID uiik 17.4 29.9 7.0 0.1 3.8 3.9 10.4 2.8 2.6 19 Cooper's hawk AZ 567 0.0 0.9 14.3 4.1 2.1 0.0 0.0 0.0 0.0 0.2 73.9 1.2 1.9 0.0 2.3 0.0 21 Cooper's hawk CA 41 0.0 0.0 4.9 2.4 0.0 0.0 0.0 0.0 0.0 0.0 29.3 0.0 63.4 0.0 0.0 0.0 10 American kestrel AZ 122 0.0 0.0 0.0 0.8 0.0 0.0 0.0 1.6 0.0 0.0 23.8 1.6 15.6 0.0 39.3 0.0 21 American kestrel ID unk 0.0 0.2 0.0 0.0 21.1 0.2 0.0 0.4 66.0 19 prairie falcon AZ !5() 0.0 0.0 5.3 47.3 2.0 0.0 0.0 0.0 0.0 0.0 39.3 0.0 0.7 0.0 0.0 0,0 21 common raven CA 901 3.0 0.7 2.1 0.6 0.9 1.8 0.4 1.0 0.0 3.8 8.3 4.4 19.5 0.1 21.5 23.1 4 greater roadrunnei AZ unk 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 6.0 0.0 85.6 0.0 6.0 11 greater roadrunnei CA 100 0.0 0.0 1.2 0.0 0.0 0.0 0.0 2.3 0.0 0.0 1.7 0.1 3.7 0.0 81.1 9.9 3 greater roadrunnei TX 69 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1.4 0.0 0.0 0.0 4.8 10.6 5.8 76.8 0.0 23 loggerhead shrike CA 11846 0.0 0.0 0.0 0.0 0.0 0.0 0.0 2.6 0.0 0.0 0.4 0.0 9.9 0.0 87.1 0.0 25 (1) Ault, 1971; (2) Bednarz, 1988; (3) Bryant, 1916; (4) Campet al., 1993; (5) Dawson, 1998; (6) Donazar el al., 1989; (7) Fisher, 1981; (8) Filch, 1947;(9) Fitch, 1948; {10) Fitch et al., 1946; (11) Gorsuch, 1932 in Hughes, 1996; (12) Jones and Smith, 1979; (13) Knight and Jackman, 1984;(14) Ligon, 1968; (15) Llinas-Gutierrez et a!., 1991;(I6) Mader, 1975; (17) Mader, 1978; (18) Marti, 1974; (19) Marti el al., 1993; (20) McClure, 1993; (21) Millsap, 1981; (22) Mill.sap, 1998; (23) Paimley, 1982; (24) Ro.ss, 1969; (25) Scoii and Morrison, 1995; (26) Small, 1971; (27) Smith, 1969; (28) Smith and Danner, 1990; (29) Trapp, 1978; (30) Turkowski, 1969; (31 Whaley, 1986. Table 3.5. Species diets Tor Western North America, arid and semi-arid habitats, derived from Table 3.4 by computing the species mean among sites. Data are normalized to IttO% frequency for the animal fmid (plants are presented as additional here), even though some studies did not report minor data categories, as for the badger report. Blanks indicate that no information (including negative information) was available. ea CQ *n e JS c .a « ec u 4) cs tiu E E 5 B ts E s J: •c •o CQ -s a £ *3 u •oo "a a Ics oC 5 CQ a. s e E '£ O' o •§ u ti X3 JS I -2L. u *3 u o E u 2 3 e V ea CQ a S t3JS tu » *0 .S. e E u SPECIES S e ^ iS C8 ca N coyote 5.9 0.1 22.7 12J 6.4 9.8 1.7 6.2 4.5 15.3 5.1 1.8 1.4 0.0 6.9 34.8 3825+ kit fox 0.3 0.3 20.4 16.5 13.0 9.7 4.3 3.6 3.9 14.4 0.3 0.4 0.0 12.9 27.8 614 gray fox 3.6 0.2 13.5 3.3 4.1 2.1 5.9 4.0 2.0 17.0 4.4 0.9 3.8 0.0 35.4 42 2296+ bobcat 1.2 0.7 41.6 3.9 19 J 2.0 3.6 18.8 5.7 0.2 3.0 0.0 0.0 21.5 316 badger 3.6 45J 4.1 0.7 11.1 0.5 1.4 1.1 32J 0.0 unk Mammals, avenijied 2,7 a.i 20.4 I6..1 10.7 6.4 .11 .^4 2.6 1.1.7 7.4 O.'J 1.9 0.0 17.5 25.1 red-tailed hawk 0.3 U.2 27.S 24.2 2.1 2.7 1.0 1.8 7.2 0.1 6.9 14.4 9.3 0.0 2.3 0.0 4069+ Harris hawk 0.0 1.9 46.9 10.6 9.7 1.2 0.1 0.0 0.2 0.3 18.1 0.0 7.6 0.0 3.3 0.0 1690 CtHtper's hawk 0.0 0.4 9.5 3.2 1.0 0.0 0.0 0.0 0.0 0.1 51J 0.6 32.6 0.0 1.1 0.0 608 prairie falcon 0.0 0.0 5.6 50.0 2.1 0.0 0.0 0.0 0.0 0.0 41.5 0.0 0.7 0.0 0.0 0.0 150 American kestrel 0.0 0.0 0.0 0.5 0.0 0.0 0.0 11.3 0.1 0.0 25.2 0.8 8.4 0.0 53.6 0.0 122+ Hawks, uveragi'd O! 0.5 17.'J 17.7 .10 O.H 0.2 2 6 1.5 0.1 2S.6 .1.2 II.H 0.0 12. J 0.0 great horned owl 0.0 0.3 15.4 0.7 5.1 12.1 9.4 23.0 2.6 6.8 5.9 2.0 1.6 0.9 14.3 trace 9859+ ham owl 0.0 0.0 O.S 0.1 3.3 6.4 29.9 44.2 10.4 4.0 0.5 0.0 0.2 0.1 0.3 0.0 7641+ western screech owl 0.0 0.0 0.1 0.0 0.0 1.4 7.8 32.5 0.6 0.0 0.8 0.3 6.5 0.0 50.1 0.0 604+ elf owl 0.0 U.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.5 1.0 0.0 98.5 0.0 unk Owls, averaged 0.0 0! 4.0 0.2 2.1 II.H 24.>J .<4 2.7 J.H 0.7 2..1 0..1 40. S 0.0 roadrunner 0.0 0.0 0.4 0.0 0.0 0.0 0.0 1.3 0.0 0.0 2.7 2.3 6.3 1.9 85.0 3.3 169+ loggerhead shrike 0.0 0.0 0.0 0.0 0.0 0.0 0.0 2.6 0.0 0.0 0.4 0.0 9.9 0.0 87.1 0.0 11846 common raven 4.4 1.0 3.1 0.9 1.3 2.6 0.6 1.5 0.0 5.6 12.2 6.5 28.6 0.1 31.6 23.1 901 Misc., averaged 1.5 a? 12 O.J 0.4 o.v 0.2 J.8 0.0 / 9 5.1 2.<) 14 9 0.7 67.'J H.H Southwest Region t.4 0.5 10.4 V.I 6 2 .l.i .i.1 .1 .i O'J SO J.1.2 .11 6.1 0..1 21.S 15.4 12.265 other VV NA 0.5 0 2 9.6 6.4 2.6 6.2 10.5 22.4 5 4 I..1 4.2 I'J 5.9 0.6 22..1 .1.0 .12,445 158 other canids in the desert, is an exception. All of these taxa utilize lagomorphs heavily, but they vary widely in secondary prey dominance: bobcats, kit foxes, and apparently Harris' hawks showed heavy usage of packrats. while the coyote, kit fox, and both buteos had heavy utilization of diurnal sciurids. Red-tailed hawks also feed heavily on reptiles, especially snakes and larger lizards. Surprisingly. Harris' hawks, kit foxes, and great horned owls, rather than bobcats, seem to be the most prominent bird-eaters. My observations of domestic cats at a desert ranch site near Tucson. Arizona, also suggest that cats frequently capture packrats. cottontails, heteromyids. and lizards, and generally capture a limited number of birds, primarily fledglings (Rosen and Schwalbe. in press). The great horned owl is the largest desert owl here, and the most powerful raptor in the valley ecosystem. Its consumption of lagomorphs and the larger rodents resembles that of the canids, felids. and buteos. However, while it takes more large prey than other owls, it has, like the fo.xes, a very broad diet with a remarkably high representation of smaller rodents. The barn owl has an even stronger affinity for small rodents, and normalization of Table 3.4 to produce Table 3.6 suggests it is probably panicularly focused on the smallest rodent category in the Southwest valleys, the perognathines (its predominant diet further north was Microtus and other small murids). The western screech-owl appears to be strongly focused on small prey, particularly perognathines among its vertebrate prey, and the elf owl appears relatively insectivorous. Most species have broad diets. The badger is relatively specialized on diurnal sciurids (Table 3.6), as has been widely reported in less quantitative accounts (e.g.. Grinnell et al.. 1937). The prairie falcon, which is probably uncommon at ORPI, is Table 3.6. ('umpulcd did cxpccled fur endolherniic prcdalurs in the dc.st;r1 valley ecosystem al ()r);an Pipe Cactus National Monument, Arizcma. Metliodii used to derive this table from literature data are described in the text. Blanks indicate that no data Ls available. Only animal food Ls considered here, and the index was normalized to 100% for each taxon. Abbrtvialions under "Diet ('ateB)»rization" correspond to other headins-s. ao •Cu ii ilsl a e in a a B e i. ucs u amphibians / arthropods hoofed animals i other endothern sdurids birds, bird egf^s lizards predators Neoioma snakes SPECIES Dipodomys a DIET CATKtiOKI/ATION /. Cimxiswnlly iibumUml spct ies: coyote K.6 0.1 34.3 15.9 5.6 12.8 2.2 6.2 2.8 1.9 0.0 9.5 lagom, sciu, k-raLs kit fox 0.3 0.3 22.0 17.8 14.1 10.5 4.6 15.6 0.3 0.4 0.0 13.9 lagom, sciu, birds, Neot, arthr, k-raLs great homed owl 0.0 0.5 24.6 0.9 5.4 18.1 16.0 10.3 2.3 1.6 1.3 18.9 lagom, arthr, k-raLs, perog western screech owl 0.0 0.0 0.2 0.0 0.0 3.9 21.4 0.8 0.3 17.8 0.0 55.7 arthr, perog, lizrds elf owl 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.5 1.0 0.0 98.5 arthr red-tailed hawk 0.4 0.2 29.1 27.6 2.3 3.0 1.0 7.4 15.7 10.5 0.0 2.7 iagcmi, sciu, snakes, lii Harris hawk 0.0 1.9 47.2 10.6 9.8 1.2 0.1 18.2 0.0 7.7 0.0 3.3 lagom, birds, sciu, Neot,liz American kestrel 0.0 0.0 0.0 0.6 0.0 0.0 0.0 25.6 0.9 8.6 0.0 64.3 arthr, birds, liz loggerhead shrike 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.4 0.0 10.2 0.0 89.4 arthr, liz II. Variably abuiuUmi spet ies: badger U.U 4.1 51.2 0.0 4.6 0.8 1.6 1.2 0.0 .36.5 sciu, arthr bobcat 1.7 I.I 53.5 5.7 26.0 8.1 0.3 3.6 0.0 0.0 lagom, Neot, heteroniyids, birds barn owl 0.0 0.0 2.6 0.2 6.6 21.6 65.0 2.6 0.1 0.4 0.3 0.6 perog, k-raLs, Neot greater roadrunner 0.0 0.0 0.4 0.0 0.0 0.0 0.0 2.7 2.4 7.3 1.9 85.2 arthr, liz, snakes ///. AbuiKltiiil spi't ifi, pert vntUKv prfiliitiim v. ii/jinn iint crlam: common raven 4.7 I.I 3.3 0.9 1.4 2.8 0.6 13.1 7.0 30.8 0.2 34.0 arthr, liz, birds IV. Ihu onwum .\pf< if.\: gray fox 5.1 0.2 21.7 2.6 5.0 2.3 6.6 5.8 I.I 4.8 0.0 44.7 arthr, lagom, various small animals Cooper's hawk 0.0 0.4 9.6 3.2 1.0 0.0 0.0 51.3 0.6 32.6 0.0 I.I birds, liz, lagimi prairie falcon 0.0 0.0 5.6 50.0 2.1 0.0 0.0 41.5 0.0 0.7 0.0 0.0 sciu, birds V. "Avfriiftf diet" (mean njspfcirs uiuler ^nmpi 1 iiiiil II): 0.8 0.3 16.5 9.9 5.3 6.2 9.1 8.0 2.1 5.4 0.3 .36.1 160 reported to eat many sciurids. rather than strictly birds; and similarly, the Cooper's hawk, which I saw foraging among the large populations of quail during the mid-l990's. may consume substantial fractions of lizards and rodents, in addition to birds (Tables 3.4-3.6). Smaller predators—roadrunner. American kestrel, loggerhead shrike, elf owl-all utilize arthropods extensively. Lizards are reported as an important secondary diet component for all of these smaller taxa except the elf owl. which is poorly known. Similarly, the western screech-owl utilizes arthropods heavily, and showed lizards as a major, tertiary prey category (Table 3.6). These species have diets distinctive from the larger generalists, with the bam owl being somewhat intermediate. TEMPORAL CHANGE IN THE PREDATOR ASSEMBLAGE Two major sub-guilds are apparent within the assemblage (Tables 3.5. 3.6). One group eating primarily small vertebrates, such as lizards and perognathine rodents, is here defined to include skunks, the bam owl. westem screech-owl. Cooper's hawk. American kestrel, greater roadrunner. and loggerhead shrike. The other group, eating primarily larger rodents and lagomorphs. includes canids. felids, the badger, buteos. northern harrier, and great homed owl. These two sub-guilds showed different rates of increase during the onset of the 1990-1995 wet conditions (Fig. 3.3). The small-prey group increased more rapidly to a 1992-1993 peak, while the large-prey group peaked during 1994-5. Observations for 1992 and 1993 (Table 3.7). together or individually, have a significantly greater proportion of the small-prey guild than all other years combined. 1988-1991, 1994-1996, 1987-1991. and 1994-1998 (X' tests. P = 0.04 - 0.004). The Table 3.7. Frequency of encounter ( No. per day) of endothermic predators that cat primarily small prey (lizards and perognathine rodents) versus those eating primarily large prey (Dipodomys, Neotoma, sciurids, lagoniorphs), at ORPl, based on Tables 3.1 and 3.6. YEAR: 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 overall small-prey taxa .51 .10 .14 .11 .26 .46 .53 .29 .21 .44 .18 .20 .26 large-prey taxa .97 .40 .50 .40 .78 .78 1.19 1.43 1.72 1.04 .48 .85 .75 N SMAI.I. 10 6 15 14 23 24 53 1 8 9 3 4 170 19 25 52 51 68 41 118 5 65 21 8 17 490 ratio small/large 0.53 0.24 0.29 0.27 0.34 0.59 0.45 0.20 0.12 0.43 0.38 0.24 0.35 Trends: 1987 1988-90 1991 1992-3 1994-6 1996-9 10 0.53 35 0.27 23 0.34 77 0.48 18 ().20 16 0.35 19 128 68 159 91 46 162 Progression of Peak Prey Populations —*- Neotoma albigula - -A - Dipodomys merriami . - o perognathines —I—lizards r«ooo>Of-CMro^u}(or^oo 0000c00>0)0>0>0>0)0>0)0> o>o>o>o)o>o>o>o>o)o>a)o> Fig. 3.3. Standardized abundance trajectories for intensively monitored prey groups at ORPI, based on data from Chapters 1 and 2. Endothermic Predator Sub-Guilds: SMALL: lizard-, perognathine-eaters LARGE: larger rodent-, lagomorph-eaters 2.1 —X— small-prey taxa 0 1.8 '•a - • o - • large-prey taxa 1 1.5 > 0) 1.2 (0 13 0.9 O a---ci • o 0.6 z 0.3 0.0 N>eoo>or-eMco^u)(or«>eo 0000000)0>0)0>0)0>0)0)0) 0>00>0)0>0>0)0)0)0>0)0> Fig. 3.4. Contrasting temporal patterns of food-type sub-guilds at ORPI. Taxonomic composition of the groups is defined in text. 163 contrast of owl species based on auditory censusing (Table 3.3) also shows more rapid increases in the smaller owls than in the great homed owl. DISCUSSION The herpetological activities during which endothermic predator monitoring was carried out were consistent and diverse enough—and the observed changes in predator observations dramatic enough—to permit confidence that the overall results demonstrate a clear correspondence between climatic conditions and predators. The late-1980's drought produced a low point in the observation rate for virtually all taxa. and the protracted sequence of wet. El Nifio-related. spring bloom years (1991, 1992. 1993, 1995) produced a 3-4 fold increase in the visual observation frequency (Fig. 3.1). The following drought, and finally the double rainfall-event year of 1998. also were associated with concordant changes in the predators. Although ectotherms with low maintenance metabolic rates (like snakes) may vary in observability due to changing activity levels, this is less e.xpected in endotherms. The observed changes in endothermic predators were probably largely the result of immigration-emigration and recruitment dynamics. From the prey's standpoint, the endothermic response implies increased predation pressure, regardless of whether the cause was heightened activity or population increase of the predators. For species like the coyote, kit fox and buteos, recruiting young were apparent during the mid-1990's. I also interpret the auditory monitoring results for great homed owls and western screech-owls (Table 3.3) to reflect an increase in owl predation pressure, rather than 164 merely in individual vocalization. The increase in visual sightings for owls (Fig. 3.2) supports this, as does the change suggested for the bam owl. Elf owl (Micrathene whiineyi) calling was generally variable or episodic, suggesting that changing vocalization or migratory patterns might have influenced its results more than those of other owl species. For great homed and westem screech-owls, auditory census results tended to be remarkably consistent within years and seasons. The dietary data and estimates for the endothermic predators at ORPI (Tables 3.4- 3.6) show that there is great overlap among taxa in utilization of primary prey. Lagomorphs were estimated to be the most important vertebrate prey for all seven of the largest predators at the ORPI study area (Table 3.6), and were the primary food for five of them based on the average for all available reports (Table 3.5). Sciurids were estimated as a primary or secondary prey category for six species, including four which shared lagomorphs as primary prey. Smaller endothermic predators had a distinctive diet: the five smallest (westem screech-owl. elf owl, American kestrel, loggerhead shrike, and greater roadrunner) all had arthropods as primary prey (by frequency of occurrence), and lizards as secondary or tertiary prey (Tables 3.5. 3.6). The bam owl. a variable-population species that was apparently not recorded at ORPI until sightings in 1947 and 1969 (Groschupf et al., 1988). may have a rodent-dominated diet that resembles that of trophically important snakes (personal observations), but not other endotherms. Even though the diets of badgers. Cooper's hawks, and prairie falcons are distinctive, their prey categories are also heavily used by the rest of this assemblage. Certainly, each species' diet has a distinctive fingerprint, and Marti et al., (1993) 165 showed that nocturnal-diumal partitioning adds structure to such an assemblage. Yet it would appear that—assuming predators do affect their prey populations (see Chapters 1. 2, 5)--exploitative food competition at this trophic level is likely to be widespread. Differences in the timing of increases of small-prey and large-prey sub-groups of endothermic predators meets a key criterion necessary (but not necessarily sufficient) for the prey assemblage to be structured directly by the mass effect predation. Moreover, the succeeding pattern of abundance peaks of the major prey groups (Fig. 3.3) shows that the smaller prey taxa declined first, when the small-prey sub-guild of predators Increased (Fig. 3.4). The protracted increase in total predation pressure during the early-mid 1990's (Fig. 3.1) might differentially affect the prey. Vulnerable species such as the perognathines declined earliest, while faster running or better defended ones like Dipodomys and Neotoma maintained high abundances longer (see Chapter 1). Perhaps coexistence in this diverse assemblage of 17 endothermic and 15 reptilian vertebrate predators is facilitated by this kind of temporal flux, which may function as a temporal niche axis defined by predation risk. Further consideration of this hypothesis will be improved by on-site dietary data, and by a longer time-series based on formalized monitoring protocols. 166 IV. MONITORING TRENDS FOR SNAKES CHAPTER SUMMARY Desert snake populations were monitored in an Arizona valley for 12 yr using drift-fence trapping, time-constrained search, and a sand-road driving transect. Intensive observation and mark-recapture were conducted on a 3 x 8 km intensive area, and prey and predator populations were simultaneously monitored. Western diamondback and Mohave rattlesnakes Croicilus atrox and C. sciituUitus. and the coachwhip Masticophis Jlagelluiih dominated biomass and prey consumption in the 17 species assemblage: other species contributed small or negligible quantities. Monitoring results suggested an early decline in several species, during 1987-1989. apparently representing the end of foregoing population maxima during the very wet period of 1977-1984. There was relative stability of snake populations during 1989-1998. despite strong winter rains that produced a surge of prey availability during the mid-1990's. The snake population increases that were apparent responses to prey increases were moderate, and not all species showed responses concordant with changing prey availability. Reptile-eating species (e.g.. M.flagelluni, C. cerastes, and Hypsiglena torquata) had moderate increases and declines that were correlated with lizard abundances. Snake predators increased during the mid-1990's, and the species which most clearly showed a prey-correlated increase was the largest and, presumably, best-defended against predators~C atrox. 167 INTRODUCTION Snake populations are among the most difficult vertebrate populations to study (see Turner, 1977), primarily because snakes are secretive and have low activity levels (see, e.g., Parizek et al., 1996), making them difficult to observe and capture. Most snake population studies have been conducted in mild or cold temperate areas (e.g., Andren, 1982; Andren and Nilson, 1983; Baron et al., 1996; W.S. Brown, 1993; Flatt et al., 1997; Forsman and Lindell, 1991; Fukada, 1969; Gregory, 1977; King, 1986; Larsen and Gregory, 1989; Lindell and Forsman, 1997; Macartney et al., 1990; Madsen, 1988; Madsen and Shine, 1992; Madsen et al., 1993; Madsen and Sille, 1988; Parker and Brown, 1980), where hibernation sites can be fenced to increase capture rates. Such northerly, winter-refuging behaviors (see Crews and Garstka, 1982, for an e.xtreme example) reflect cold-limitation (Rosen. 1991), whereas biotic factors like competition, predation, and parasitism are likely to be more important in warmer areas. Long-term studies of snake assemblages (Fitch, 1949, 1982, 1998 and therein; Fitch et al., 1984; Fukada, 1969; Hirth and King, 1968; Klimstra. 1958; Parker and Brown, 1973, 1980; Seigel et al., 1995; Woodbury, 1951) are few, and many of them have seen changes brought about by human snake-killing and habitat-altering proclivities. Studies of snake community dynamics are lacking for tropical and subtropical assemblages. Snakes have low metabolic rates (Andrews and Pough, 1985; Beaupre, 1993; Beck, 1995; Chappell and Ellis, 1987; Lillywhite, 1987; Secor and Diamond, 1994), and although some ecologists have begun to recognize the importance of snake and owl predation for the behavioral ecology of desert rodents (Abramsky et al., 1998; Kotler. et 168 al.. 1992. I993a&b. Pierce et al.. 1992). there is a prevailing attitude that snakes are unlikely to affect prey populations (Porter and Tracy. 1974; Reichenbach and Dalrymple. 1986; Bouskila, 1995). On the other hand, there is evidence that snake population densities can be high in warm and arid ecosystems (e.g.. Fitch, 1949, 1998 and therein). Fitch (1949, 1958, 1982), Barbault (1970, 1971), Seigel (1984), and Reichenbach and Dalrymple (1986) reached divergent conclusions on the trophic importance of snakes. Fitch (1960. 1975. 1982), Hailey and Davies (1987), and Pomianowska-Pilipiuk (1974) estimated individual food consumption rates that should be high enough to influence prey under naturally-occurring abundances. Diller and Johnson (1988) and Forsman and Lindell (1996) found that snakes are likely to be trophically important predators, even though they worked in mild temperate climates. Savidge (1987) described the devastation of the Guam bird assemblage by the introduced brown tree snake Boiga irregularis. The role of snakes in ecosystems remains ambiguous, although the emerging view is one of trophic importance in some areas. Likewise, the regulation of snake populations is poorly understood. Klimstra (1958) and Scott and Klimstra (1955) reported time-series data for snakes and their prey and predators, but were not able to reach conclusions on the mechanism of observed population fluctuations in snakes. Forsman and Lindell (1996), Lindell (1997), and Lindell and Forsman (1997) systematically monitored snake life history and prey abundance, and presented evidence that varying prey abundance affects the demography and population dynamics of Vipera berus. No study has previously reported quantitative monitoring results for a snake assemblage and its predators and prey. 169 In 1983,1 initiated study at Organ Pipe Cactus National Monument (ORPI). in the diverse subtropical biota of the Sonoran Desert. I wanted to observe an undisturbed snake community in the contexts of climatic fluctuation, snake population biology, and prey and predator populations. This became feasible in 1987 when the National Park Service established the Sensitive Ecosystems Program at ORPI (which became the Ecological Monitoring Program [EMP]). The EMP now includes ongoing, formal monitoring projects for the principal prey the snakes at ORPI--diumal lizards and nocturnal rodents. My work under the Sensitive Ecosystems Program permitted me to establish and conduct the snake study and to also monitor other vertebrate predators and prey taxa (see Chapters 1-3). I thus have monitoring results for key components of the vertebrate community at ORPI. The purpose of this paper is to summarize the snake results for use with other monitoring results to evaluate prey population trends. METHODS STUDY AREA DESCRIPTION Habitat and Climatic Setting The study area lies on the transition between the Arizona Upland (thomscrub) and Lower Colorado Valley (desertscrub) subdivisions of the Sonoran Desert, in western Pima County, Arizona. Precipitation averages about 212 mm/yr in the valley habitat where most snake monitoring was conducted, but was about 25 mm above normal during the study period, a time also of above-normal temperatures. A 3 x 8 km focal region for intensive study of snakes was located in valley habitat at 470-555 m elevation, in middle 170 and lower bajada habitat, and barely extending onto the flats of the valley floor. The area has fine to very fine sandy loams, sparse gravel and essentially no rock, and supports a mosaic of open desert woodland and creosotebush-dominated desertscrub. It is everywhere transected by arroyos that support denser thornscrub. and there are important local areas of mesquite-creosotebush floodplain where the arroyos braid apart and spread broadly to irrigate extensive and productive flats. Generally, outside the arroyos and the least-degraded parts of the floodplains (see below), vegetation structure is readily penetrable by humans and, presumably, by various animals that prey upon snakes. A fuller description of habitat conditions is given in Chapter 2, and additional details of climate and landscape structure are in Chapters 1 and 3. The ORPI Snake Study Area (OSSA) is the central 1 km" area of the intensive study site. The habitat at OSSA is a fine sandy loam supporting a mosaic of (1) paloverde-ironwood-saguaro habitat including complex shrub and sub-shrub strata and (2) more open areas predominantly vegetated by creosotebush, white bursage, triangle- leaf bursage, and big galleta grass. The OSSA includes a major wash with large ironwood trees and extensive catciaw acacia-canyon ragweed thickets, two richly- vegetated medium-sized washes with diverse flora and much dense thicket microhabitat, and about 10 small washes with open shrub-tree and dense low shrub cover. Potential Human Effects on the Study System Human impact on the study area is limited. Although ORPI, specifically including the intensive snake study site region, was grazed by cattle and other livestock until 1976- 171 1978, the degradation has been limited compared to the degradation of most other ecosystems on earth. By 1987, at the latest, perennial grasses were widely established, and they appeared to be stable and flourishing during this study. The primary residual effect of grazing is entrenchment of the arroyos, reducing the rich floodplain environment as well as the density of the lower vegetation layers (subshrub, 0.3-0.5 m; large shrub 1.5- 2 m) along many affected arroyos. Since snake abundances are highest in un-degraded e.xamples of these habitat types (unpublished data), this is presumably a residual effect of grazing. In addition to grazing, floodplain habitat in the study area suffers from upstream stock-watering impoundments, which have de-watered and thus killed downstream mesquite-creosotebush floodplain woodlands. This has facilitated severe erosion and helped produce barren landscapes in their stead. Additionally. ORPI is at the arid limit of traditional floodwater-irrigation agriculture attempts, which have failed, including on part of the intensive snake study site. Clearing of floodplain xeroriparian desertscrub for this purpose, 4-5 decades prior to the initiation of this study, appears to have had a role in the erosion and devegetation afflicting the valley floor floodplain habitat. The intensive snake study site region had significant pre-Columbian human utilization, and is currently lightly used by National Park Service personnel, police, and, most recently, people illegally migrating from Mexico to the United States. Highway mortality of snakes occurs on the site (Rosen and Lowe. 1996), but probably has litde effect in areas where most of the data were collected. The benefits of highway scavenging to coyote and raven populations may have unknown effects on ORPI ecosystems. 172 Stock-watering facilities north of the monument presumably benefit mobile taxa, notably predators such as coyotes (see Golightly and Ohmart, 1983, 1984) and raptors. The most likely significant effect of this is intensified predation within the predator trophic level. For example, coyotes might have a pronounced effect on larger snakes that would be affected less, or at least differently, by kit foxes, which are less reliant on free water than coyotes (Golightly. 1981). During dry times, animal trails converging from at least 4 km (and probably much further) on these artificial water sources indicate heavy utilization by canids. deer, and peccaries. Despite all of this, the study area remains close to original natural conditions for this snake assemblage. S.NAKE MONITORING METHODS Drift-fence Trapping Technique Snakes were trapped using 22 drift-fence traps set on the 1 km" OSSA site. Seven drift-fence traps were in upland desertscrub, and 15 were in or near xeroriparian desertscrub. The drift fences were 12-18 m lengths of 3 mm (1/8 inch) mesh hardware cloth standing 0.75 m tall and 0.15 m into the ground. They had a screen funnel trap at each end. which led through a second funnel to mostly-buried, enclosed plastic trap buckets (75 or 110 1 capacity; sunk 42-52 cm into the ground). Trap bucket entrances were aboveground, to prevent flooding, and the bucket lids were covered by plywood to reduce heating in the bucket and surrounding soil. Double false bottoms of screen (12 mm. 24.5 mm) were installed to minimize predation within the trap buckets. The traps were permanently installed, in all habitat types, in sites likely to catch snakes. They were 173 checked once or twice per day. The trapping started with 5 traps in 1988. 8 in 1989, 14 in 1990-91. 22 during 1992-3, 17 during 1994-5. and 10-12 during 1996-1998. Data Reduction and Variance for the Trap Results The objective of this paper is to quantify variation in snake relative abundance among years. Therefore. I developed a spreadsheet format to compute the standardized annual trap capture rate for each species. This was necessary because capture rates varied among traps and seasons. The original 5 traps were less efficient than subsequent installations, and indeed, each trap had a characteristic capture rate for each species. Capture rates also varied among months, although this variance was much smaller than that among traps. Therefore, capture data were standardized by first adjusting monthly capture rates to account for inter-trap differences, producing a vector of standardized monthly capture rates for each species. Using this vector as input to standardize the results with respect to seasonal differences, I computed a second standardization vector containing the characteristic trap capture rates for each individual trap. These two values were then used to adjust the capture rate of each year, yielding the standardized capture rates per year. The details of this procedure are as follows. Total captures were always taken as the direct sum of actual captures, whereas the values for trap-days were adjusted to remove biases caused by variation among traps and months. Real (un-adjusted) trap-days were determined on a per field-trip basis as the sum of 24 hr cycles of trapping, plus a fractional day equal to the proportion of the 24 hr activity period encompassed by any partial trap-day. This was done separately for diurnal 174 and nocturnal snakes, with consideration also for intra-specific seasonal changes in activity time. With the trip values in hand, actual trap-days were summed by trap and month, and these were used in the calculations to produce adjusted values. For this, monthly capture rate was corrected for inter-trap variance by computing adjusted number of trap-days per trap for each month. This was done by multiplying actual trap-days for each trap by the relative capture rate of that trap ((capture rate for trap [il)/(mean trap capture rate}}. For each month, the adjusted number of trap-days was then the sum of the adjusted trap days of the individual traps operating during that month. Standardized monthly capture rates were then simply the elements of the vector containing the individual months values of (actual captures)/(adju.steci trap-days). Similarly, individual traps characteristic capture rates were adjusted to account for inter-trap differences in the months they were operated in any given year. The standardized monthly capture rate vectors (one vector per species, as defined above) were used to adjust the actual trap-days for each trap. Again, standardized trap capture rate were the elements of the vector of values of (captures)/(adjusted trap-days), but here the elements were the individual traps relative capture rates. These were quite close to the raw values, since month-month variation in capture rates was not extremely large, and inter-trap variation in actual trap-days per month was not large. Standardized values for monthly and individual trap capture rates were each corrected to a mean of unity. Adjusting months and traps for differences among years was not necessary because months and traps were exposed to the full range of inter-year variation during this study. Therefore, the standardized values retain very little bias. 175 Finally, to compute the standardized annual capture rates. I used these unitized adjusted capture rates to adjust the actual annual trap-days, as follows. Two huge (3- dimensional) tables were constructed: 1. (actual captures x month x trap); and 2. (actual trap-days x month x trap). Actual trap-days (for each month for each trap) were adjusted by multiplying them by the corresponding two correction factors (in the standardized vectors, for the appropriate month and trap). For each year, these adjusted trap-days were summed to give an adjusted total annual trap-days. Again, total captures was summed for each year, and standardized annual capture rate was then simply (capturesjAadjiistecl trap-days), in vector form, with the elements being years. The procedures and results were inspected for each species to ensure that sample size artifacts were not compounded in the standardization process. Since most capture values per trap per day were zero, and since there was variation in which traps were operating during any one year, it was impossible to compute variances from the raw data using normal statistics. [ derived an appro,ximation of the variance by treating the adjusted capture rates as binomial parameters. Variance terms (standard errors) for the trap capture rates were thus taken from the formula for variance of a binomial parameter (SE p = [P y. Q i . P •= per trap-day capture rate, Q = \ - P. N = no. trap-days [adjusted value]). 176 Less work was done during 1994 and 1995 than in other years as a result of other pressing concerns. The 1994 results were taken during a short period of maximal activity for many species, yet were limited enough that many less abundant species were not recorded at all. In contrast, 1995 work was. of necessity, carried out during a dry July- August. and was suspended prior to late-August rains. Since these biases tended to offset one another, and the combined data 1994-5 has sample sizes more comparable to those for other years, I combined the data for those years and recomputed the standardized capture rates. Although these results were cleariy superior to the individual results for 1994 and 1995, both are presented for the sake of completeness. Time-Constrained Searcii (TCS) Method During all routine, desert herpetological work at ORPI. 1 recorded starting and stopping time, area worked, activity carried out. and all reptiles, amphibians, mammals, and raptorial birds I saw. The data reponed here pertain only to nocturnal searches conducted to look for snakes—results obtained while other activities were simultaneously being done are not included. Further, these data apply only to work in the 3 x 8 km intensive study site. The TCS procedure was to move through the habitat looking and listening for animals; at night a headlamp powered by four 1.5 V D- cell batteries (6 V total) was used. Snakes encountered were captured and studied on the spot, and this processing time was deducted from the search time for each time-constrained search, using a formula based on handling times for different species of snakes during different years. Travel was at approximate rates of 0.5-2 km/hr. Identical headlamps were used 177 throughout the study. TCS results were expressed as observations/hr, and standard errors for annual means were computed in the normal way. Road Transect Method A sand road passing through the intensive study site, open only for service management and law enforcement, was used as a road-cruising transect. This road (4 m from bank to bank) was cruised at 9-14 mph {14.5-22.5 km/hr) at night and 15-20 mph (24-32 km.hr) during daylight. Road-driving transects were conducted mostly in April- October, from just before sunset to 2-4 hours after dark, but all seasons and hours were sampled. Only transects run during full darkness, April-October inclusive, and only when the full length of the road was travelled, are included here. Road milages were recorded to an accuracy of nearly 0.01 mi (16 m) by interpolating between 0.1 mi intervals on automobile odometers. Odometers were calibrated against known 10-50 mi highway stretches, and corrections applied to all data. Snakes encountered on the road were measured and marked on the spot, and released immediately. SYNTHESIS OF MONITORING METHODS RESULTS In order to combine the results of the three methods into a single estimate of the yearly trend of snake populations, I weighted the methods according to the sample size for each method, for each year, separately for each species. In order to perform the weighting, each dataset was first standardized to a mean of one. with resulting variation only among years, and the weighting was performed on these values. 178 RESULTS PRELIMINARY OBSERVATIONS ON ASSEMBLAGE STRUCTURE Seventeen species of snakes were found on the intensive study site (Table 4.1), 15 of which maintained populations on the OSSA site. Most species were effectively sampled by all three methods utilized (Table 4.2). except the sidewinder Crotaliis cerastes, which was heavily over-sampled by road-cruising because individuals travelled for great distances on the road. Other species normally crossed the road directly, as indicated by tracks. The most abundant species during this study were the western diamondback Crotaliis atrox, coachwhip Masticuphis JIagelliun. Mohave rattlesnake Crotaliis sciitiilatus, and long-nosed snake Rhinocheihis lecontei. Recapture rates were substantial for all of the vertebrate-eating snakes (i.e.. all snakes except Chionactis, Chilonienisciis, and Leptoryphlops). and indicated that the number of captures per species does not grossly misrepresent the relative abundance of any of them (unpublished data), with the exception of the sidewinder. For present purposes, the relative abundance of the sidewinder is represented by its occurrence in the trap and on-foot samples. Table 4.3 and Fig. 4.1 are derived from Tables 4.1 and 4.2 by utilizing the relative abundance of the sidewinder indicated by trapping and on-foot search only. MONITORING RESULTS Drift-fence capture rates (Table 4.4, Fig. 4.2) suggest that many species had relatively stable populations, although variances were high. Some lizard-eating species (Salvadora, Hypsiglena, Phyllorliynchus deciirtatiis, Rhinocheihis) had low observed Table 4.1. Snake observations made during monitoring work on the intensive snake study region monitoring area, in the Valley of the Ajo, Organ Pipe Cactus National Monument, Arizona, during 1987-1998. SPECIES 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 total Arizona elegans 4 13 16 11 3 10 10 2 7 1 3 1 81 Chilomensifus ductus () 1 1 2 0 0 2 0 0 0 0 0 6 Chionactis palarostris 0 0 0 1 0 0 1 0 0 0 0 0 2 Crotalus airox 10 31 18 55 40 37 43 11 7 35 20 16 323 Crotalus cerastes 38 19 21 56 71 46 60 1 22 44 11 61 450 Crotalus scutulatus 14 24 16 28 22 29 33 7 5 7 6 10 201 Hypsiglena torquata 2 17 14 28 14 23 16 0 1 3 5 14 137 Lutmpropeltis getula 1 3 0 7 0 2 0 0 0 1 1 1 16 Leptotyphlops humilis 0 2 3 5 1 1 1 0 0 0 0 2 15 Masticophis flagellum 1 9 18 48 49 54 61 3 10 18 10 25 306 Micruroides euryxanthus 1 1 1 13 8 2 2 1 1 1 1 0 32 Phyllorhynchus hrowni 0 2 1 2 2 7 2 0 1 0 1 2 20 Phyllorhynchus decurtatus 1 11 7 7 2 5 15 0 3 0 0 1 52 Pituophis melanoleucus 3 3 6 5 8 6 10 2 0 1 3 3 50 Rhinocheilus lecontei 5 7 17 33 23 16 28 2 6 8 4 12 161 Salvadora hexalepis 0 1 1 7 12 7 6 0 1 0 0 2 37 Trimorphodon hiscuiatus 0 0 0 0 0 0 1 0 0 0 0 0 1 total 80 144 140 308 255 245 291 29 64 119 65 150 1890 180 Table 4.2. Observation types used during moniotring work on the intensive snake study region monitoring area, in the Valley of the Ajo, Organ Pipe Cactus National Monument, Arizona. 1987-1998. METHOD: road- on-foot drift-fence SPECIES driving survey trapping total Arizona elegans 22 34 25 81 Chilomensicus cinctus 0 0 6 6 Chionaclis pcilarostris 1 0 1 2 Croialus atrox 24 219 80 323 Crotalus cerastes 395 36 19 450 Crotaliis sciitulatiis 37 108 56 201 Hypsiglena torquata 13 19 105 137 Lampropeltis getiila 4 4 8 16 Leptoryphlops hiimilis 0 4 11 15 Masticophis flagellum 6 42 258 306 Micruroides eiiryxanthiis 2 10 20 32 Phyllorhynchiis browni 8 6 6 20 Phyllorhynchus decurtatus 13 19 20 52 Pituophis melanoleuciis 10 16 24 50 Rhinocheilus lecontei 24 51 86 161 Salvadora hexalepis I 5 31 37 Trimorphodon bisciitatus 1 0 0 1 total 561 573 756 1890 181 Table 4.3. Preliminar>' estimate of biomass distribution among snake species on intensive snake study region. Relative abundances derived from Tables I, 2 (see text). Sample sizes for body mass are in Table I. Under diet. m=mammals. r=reptiles. i=invertebrates. relative mean OSSA index of SPECIES abundance (%) body mass (g) biomass (%) diet Arizona elegans 5.48 84.7 2.48 m r Chilomensicus cinctiis 0.41 8.8 0.02 i Chionactis palarostris 0.14 5.9 0.00 i Crotalus atrox 21.84 354.5 41.43 m Cmtalus cerastes 2.65 95.1 1.35 m r Crotalus scutulatus 13.59 289.0 21.02 m Hypsiglena torqiiata 9.26 12.9 0.64 r Lampropeltis getula 1.08 192.8 1.12 r m Leptotyphlops humilis I.OI 3.6 0.02 i Masticophis flagellum 20.69 170.0 18.82 r Micruroides euryxanthus 2.16 24.3 0.28 r Phyllorhynchus browni 1.35 20.7 0.15 r Phyllorhynchus cleciirtatiis 3.52 22.7 0.43 r Pittiophis melanoleiiciis 3.38 426.1 7.71 m Rhinocheilus lecontei 10.88 63.1 3.68 r m Salvadora hexalepis 2.50 63.4 0.85 r m Trimorphodon bisciitatus 0.07 40.4 0.01 r m Percentage of Snake Biomass at ORPl Snake Study Area Fig. 4.1. Preliminary estimate of biomass distribution among all snake species on the ORPl Snake Study Area, intensive snake study region. Valley of the Ajo, Arizona, 1987-1998. Table 4.4. Snake relative abundance among years based on yr-adjusled drift-fence trap capture rate, OSSA site, ORPl, Arizona. The values in the table are number of captures per 24 hr trap-day for one drift-fcnce with a trap at each end; they are adjusted for differences among years in seasons of sampling and indivdiual trap use (sec text). Combined YEAR; dula; TAXON 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 94&5 overall M. flagelliun .036 .054 .040 .066 .063 .057 .019 .031 .058 .036 .064 .029 .052 Rhinocheilus .009 .033 .022 .020 .008 .023 .014 .015 .013 .012 .016 .015 .017 Salvadora .(XX) .(X)l .(X)7 .017 .(X)7 .(X)8 .(XX) .016 .(XX) .(XX) .(X)4 .014 .(X)7 C. cerastes .(X)6 .003 .{X)3 .(X)4 .(X)5 .(X)4 .(XX) .(X)6 .(X)7 .(XX) .(X)3 .(X)5 .(X)4 Hypsiglena .040 .018 .020 .016 .020 .026 .(XX) .(X)3 .(X)9 .018 .046 .(X)3 .020 P. decurtatus .011 .(X)l .(X)4 .(X)7 .(X)3 .(X)8 .(XX) .(X)5 .(XX) .(XX) .(X)4 .(X)4 .{X)4 C. atrox .(XX) .(X)7 .013 .010 .018 .020 .031 .032 .039 .015 .013 .032 .017 C. scutulatus .(X)8 .011 .011 .016 .014 .010 .052 .(X)3 .013 .011 .(X)8 .011 .011 Pituophis .(XX) .(X)8 .(X)2 .(X)9 .011 .(X)5 .018 .(XX) .(XXJ .(X)5 .004 .(X)4 .(X)5 Arizona .(X)5 .(X)9 .(X)3 .(X)6 .(X)4 .(X)4 .(XX) .012 .(X)3 .(X)6 .(X)2 .010 .(X)6 Micruroides .021 .(XX) .(X)6 .(X)7 .(X)l .(XM .(XX) .(X)3 .(X)4 .(X)4 .(XX) .(X)2 .(X)4 Leptoiyphlops .(XX) .(X)l .(X)5 .(X)2 .(X)l .(X)2 .(XX) .(XX) .(XX) .(XX) .011 .(XX) .(X)2 overall .141 .151 .141 .184 .154 .175 .134 .124 .145 .105 .188 .128 .153 SE .029 .017 .011 .017 .012 .014 .050 .019 .022 .018 .021 .018 .(X)5 Adjusted drift fence - M. flagellum Adjusted drift fence C. atrox Adjusted drift fence o— Hypslglena reiull*, ORPl Salvadon results, ORPl a' C. scutulatus results, ORPl * P. decurlatus 0.09 • 0.050 t I SE 11 SE 0.08 0.07 0.040 0.06 0.030 0.05 - 0.04 - 0.020 0.03 0.02 0.01 - 0.010 •l: .41. 0.01 •j—J- 0.00 a. § § § S g g i i g i aOiQ«-Mcou)^ -o— MicfUfoides Ad|usled drift fence RhinocheUus Adjusted drift fence Pituophis '' ' ^-^ofypf*fop*.ChkmacU»,Chlk>meni*cus retulta, ORPl Arizona results, ORPl - • C. cenates 0.05 0.015 0.025 ± 1 SE 11 SE 11 SE 0.020 0.010 0.015 0.010 0.005 •4.... 0.005 0.000 OtOv-MOlOWh-flO § § S S aAfi A A A A A 2 * AAAA A A AAAAAA^AAA Fig. 4.2. Siandardized annual drift-fence irap capture rates (sec text) for snakes at the ORPl Snake Study Area. 185 abundance in the mid- 1990's. and appeared to begin a rebound at the end of the study period, when lizard populations were high. The coachwhip pattern also corresponded with lizard abundance, although its increases were not large. The western diamondback showed an increase during the mid-1990's, when its mammalian prey were also most abundant (see Chapter 1). Drift fence data for the rattlesnakes in 1988 are not indicative because the tlrst set of 5 traps was ineffective at capturing them. TCS results for the C. ciirox also show an increase during the mid-1990s and a lack of similar response by C. scutulaiiis (Table 4.5, Fig. 4.3). These results suggest long declines in several species (C. scuiulaius, Arizona. Hypsiglenci. Pliyllorliynchus): these results have mid-l990's low points that are consistent with the drift-fence trap record. Road-cruise transect results (Table 4.6, Fig. 4.4) also indicate a mid-l990's population increase in C curox. setting aside the values for 1994-1995, which are ba.sed on limited sampling. Large road-cruising samples were only obtained for the sidewinder (Fig. 4.4). The weighted combination of results from the three methods (Fig. 4.5) indicates relatively stable observed abundances for most species from 1988 or 1989 to 1998. Several species showed early declines from high levels in 1987-1988, including all three rattlesnakes. Table 4.5. Nocturnal time-contrained .search re.sults for .snake.s at ORPl inten.sive stiidy region. Values are the mean (among .searches) of number of indivdiuals found per hour of on-foot .search by one person (the author) using a headlamp, April-November, under thermal conditions suitable for snake activity, but primarily before 0230 hr. Searches when other activities (such as trap-checking) were underway are not included. Thus, the total sample here is roughly half that shown for "on-foot" captures in Table 4.2. YEAR: TAXON 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 overall SK C atrox 0.13 0.43 0.15 0.89 0.39 0.12 0.23 0.74 0.66 0.48 0.41 0.377 0.049 C. scutulatHS 0.16 0.43 0.09 0.34 0.08 0.11 0.13 0.21 0.00 0.05 0.00 0.128 0.030 Rhinocheilus 0.06 0.03 0.07 0.10 0.17 0.00 0.18 0.00 0.04 0.05 0.25 0.104 0.027 Arizona 0.15 0.14 0.09 0.16 0.00 0.00 0.01 0.20 0.00 0.10 0.00 0.057 0.016 Hypsigleua 0.00 0.06 0.02 0.00 0.04 0.06 0.00 0.00 0.06 0.00 0.00 0.026 0.011 Phyllorhynchm 0.00 0.13 0.07 0.00 0.00 0.00 0.06 0.00 0.00 0.00 0.04 0.035 0.011 all .snakes 0.55 1.25 0.52 1.49 0.72 0.32 0.66 1.25 0.79 0.71 0.70 0.756 0.067 No. of searches 11 27 42 16 36 20 28 3 0 27 17 29 256 No. of hours 16 36 55 11 47 17 25 11 20 14 16 269 Time^onstrained Search, Nocturnal, All Species TIme-Constralned Search, Nocturnal • D C. atrox 1.4 1 • C. scutulatus 2.0 1.2 1.6 - 1.0 0.8 0.6 0.8 0.4 0.4 0.2 0.0 0.0 CO o» CM CO ^ U> 00 OI CM CO in 00 00 00 00 OI a> Oi OI OI OI OI 00 00 00 OI OI OI OI OI OI OI o> o> OI OI OI OI OI OI OI OI OI OI Time-Constrained Search, Nocturnal o • Rhinocheilus Time-Constrained Search, Nocturnal ^ - Hypsiglena -•— Arizona Phyllorhynchus 0.4 T 0.2 T SE 0.3 ' i 0.2- 0.1 0.0 0.0 OI (M CO U> 00 00 OI CO in 00 eo 00 OI (71 OI OI OI OI 00 00 00 OI OI OI OI OI OI OI o» OI OI OI OI OI OI OI OI OI OI OI OI OI OI O) OI OI Fig. 4.3. Annuul vuriulion in time-consiruined .seurch (TCS) results for snukes at the intensive snuke study region, ORPI. Only taxu with n>8 lotul observutions ure included. Meun + 1 SE are shown. Table 4.6. Nocturnal sand-road transect results for snakes at ORPI intensive study region. Values are the mean (among runs) for number of indivdiuals found per 4.78 mi (7.7 km), April-November, at 9 - 15 mph (15-24 kph), under thermal conditions suitable for snake activity, but primarily before 0230 hr. Partial transect runs were not included; thus the sample size here is about half that shown under "road-driving" in Table 2. YEAR: all TAXON 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 YR C. atrox 0.11 0.(K) 0.(K) 0.02 ().(K) 0.(M) 0.(H) 0.(K) 0.(H) 0.(K) 0.06 0.08 0.022 C. scululatus 0.28 0.04 0.(K) 0.02 0.06 0.10 0.12 0.(K) 0.(K) 0.08 0.13 0.05 0.063 C. cerastes 1.06 0.31 0.44 0.55 0.85 0.52 0.47 0.(K) 0.47 1.33 0.63 1.53 0.731 Rhinocheilus 0.06 0.04 0.(K) 0.05 0.06 0.(M) 0.18 0.(M) 0.(M) 0.08 0.(X) 0.(X) 0.041 Arizona 0.06 0.07 0.04 0.03 0.(K) 0.(K) 0.06 0.(K) 0.07 0.(M) 0.(K) 0.(K) 0.028 Phyllorhynchus spp. 0.06 0.04 0.(K) 0.03 0.(M) ().(M) ().(H) ().(H) 0.13 0.(M) 0.06 ().(K) 0.025 all snakes ** 0.67 0.27 0.08 0.19 0.12 0.10 0.35 0.(M) 0.20 0.17 0.25 0.13 0.206 No. of transect runs 18 45 25 58 34 21 17 3 15 24 16 38 314 ** except C. cerastes Nocturnal road transect results, ORPl: all species Nocturnal road transect results, C. atrox except sidewinder ORPl: large rattlesnakes C. scutulatus 0.4 ±1SE ±1SE 00 -rN h*-a)0>o^cMco^&n(or^oo r^AOi-cMco«in Nocturnal road transects, ORPl: • ^ • Rhinocheitus Nocturnal road transect results, ORPl: medium-sized constrictors -o— Arizona sidewinder rattlesnake, Crotalus cerastes 0.35 2.0 ± 1SE ±\SE 0.30 g 0.25 ? 0.20 ^ 0.15 d Z 0.10 0.05 ^oo0o^rgco*tmcot-voo ^oocno^cMco^iniOi^oo 0000fl000>0>0>0>0>0)0)0) w€0€OdO>CRO>0>^00>0) (7) O) O) O) O) O) O) O) O) O) O) O) 0>0>(7>0)O0>0)0)0)0>0>cn Fig. 4.4. Annual variation in road-crui.sing result.s on a sand road, intensive snake study region, ORPl. Only taxa with n>8 total observations arc included. Weighted mean of • C. alrox c. cerastes Weighted mean of • ••e, - Phyllorhynchus spp. monitoring methods • C. scutulalus monitoring methods —Q— Hypslglena 4.0 5.0 Weighted mean of ' Rhlnochellus Weighted mean of - Masllcophls • • -x • • Salvadora monitoring methods - Arizona monitoring methods - PItuophls at u a TJ Ji < TJ0) N ••5 CB tJ C CQ s> Fig. 4.5. Weighted, standardized mean of annual values from drift-fence, TCS, and road-cruise methods of snake monitoring on the intensive snake study region, ORPl. 191 DISCUSSION ASSEMBLAGE STRUCTURE Dietary data indicates that the western diamondback rattlesnake was the most important mammal-eating snake, followed by the smaller Mohave rattlesnake. The gopher snake Pimophis melanoleiicus played a small role in overall mammal biomass consumption. By virtue of its size, abundance, and high activity, the coachwhip was the predominant lizard-eater, whereas the smaller, secretive, and less active long-nosed snake Rliinocheiliis lecontei and the small night snake Hypsigleiui torquam played relatively minor roles. However, various other snakes eat lizards as a minority fraction of their diets, and a rough biomass sum of 10% approximately reflects lizard biomass consumption by snakes other than the coachwhip. Although metabolic rate of the coachwhip is higher than for many other snakes (e.g.. Secor. 1995; Secor and Diamond, 1994), this should be reflected in higher activity, and was seen as high recapture rates for this species. Thus, Table 4.3 and Fig. 4.1 are reasonable first approximations of the relative trophic importances among snake species. TEMPORAL TRENDS Limited annual sample sizes make it difficult to infer individual species population trends from Figs. 4.2-4.5. although recapture results (below) confirm some of the indicated fluctuations. To clarify the broad outline of events in the snake assemblage. I used the standardized species means (Fig. 4.5) to construct a time-series for snakes as a 192 whole and for the two most prominent trophic groups of snakes (Fig. 4.6). The general stability of snake species populations through large fluctuations of climatic and prey abundance in the mid-1990's is striking in contrast to larger observed increases in endothermic predators (Fig. 4.6). Lizard-eating snakes generally increased, following the trend of lizard abundance, in both the early and late 1990's, although the magnitude of the response was moderate (Fig. 4.6). The snake decline seen starting in 1992 is not simply an artifact of sample size. Although I have only incorporated TCS and road-cruising data from fully-standardized work, a great many additional observations were not included because I could not accurately determine time-of-work values for them. This work also showed low capture rates beginning in 1992. For many species, high recapture rates were obtained during and following the most intensive years of study. 1988-1993. In 1989-1993, almost all long- nosed snakes on the OSSA site were marked: the 1992 population collapse was distinct in the mark-recapture results (pers. obs.). For leaf-nosed snakes, which are lizard-egg dietary specialists (Rosen et al.. 1996), the graphed results may underestimate the severity of decline. Leaf-nosed snakes were already at the tail end of activity when 1987 sampling began, on August 17. The low value for 1987 is thus an artifact, and both monitoring and mark-recapture results showed a large decline for them. Similarly, the night snake trend was also present in the mark-recapture data, although recapture rates for this species never exceeded 50%. For the western patch-nosed snake Salvadora hexalepis. the early increase during 1989-1991 was a distinctive one, in which the species appeared on the OSSA site and flourished for several years, during the period of intensive sampling. SNAKES: • + Lizard-eaters - Mammal-eaters All (A -a N 00 O) o CO U) (O oo 00 00 oo 9 O) O) a> Oi O) o> ai 9> O) O) 9) o» O) a 0) r- r* r" best assignments. 2 categories lizards -a— small mammals (A -a N oo 9> CM cn lA (O oo 9) at 0)00 0)oo 0>oo 0> s o»a> O) O)0) O) 0>0) oo> 0) • PPT (intensive study region) •endothermic predators « 2.0 S 0.8 (A 0.0 NOQ 00eo «o> 9 O) w Fig. 4.6. Trophic dynamic patterns of the snalce assemblage, intensive study region, ORP 194 Thus, for lizard-eating snakes, responses congruent with prey population changes (Fig. 4.6), although real, masked a remarkable diversity of species population responses. However, a more consistent response to lizard population increases appeared to be underway during the late 1990's. The remarkable stability of the Mcisticophis jlcigelhim population, for which high recapture rates are available from 1989-1998. corresponds to field observations of successful reproduction in this species despite drought conditions that sharply attenuated reproduction in the other species e.xcept C. cerastes (unpublished data). These two species showed relatively clear, though not dramatic. increa.ses (Figs. 4.2. 4.5) during the 1990's lizard increa.ses (Fig. 4.6). The coachwhip and its congener the Sonoran whipsnake M. bilineatus are pronounced ophiovores. which I have observed to consume long-nosed snakes (in lab. n=4). night snakes (in lab. n=l). patch-nosed snakes (in lab. n=l; and field, n=2 [1 of which escaped a predatory strike]), and rattlesnakes of surprising size (in lab, n=3; in field. n=2; see also Ortenburger. 1928; Secor. 1994; and Shaw and Campbell. 1974), although the coachwhip may avoid garter snakes (Thanmopliis marcianus, T. cyrtopsis) and perhaps coralsnakes (pers. obs.). Coachwhips were conspicuous and abundant in the trapping study, and it was obvious~at least during wet years when they were highly active and grew rapidly—that they posed a major, if not ubiquitous, danger to smaller snakes (pers. obs.). The monitoring results consistently showed a mid-1990's increase for C. atrox. The late I980's decline of this species and the other two rattlesnakes was also seen at a the more arid Pozo Nuevo EMP site in western ORPI. which was studied using TCS 195 during 1983-5 and 1987-8. At that site, there appeared to be approximately a 60% decline from the early to late 1980's (unpublished data). Thus, the 1987-1989 declines seen among several species in the current monitoring record (Fig. 4.5) probably represent the termini of population booms in the very wet period of 1977-1984. Overall, mammal- eating snakes increased slightly during the mid-1990's climatic amelioration (Fig. 4.6), in synchrony with the small mammal increase. However, as in lizard-eating snakes the species' responses varied considerably. C. airox was the only mammal-eating species for which a prey-correlated increase was consistently seen in three different monitoring methods. The gopher snake trajectory could not be accurately determined due to small sample sizes in the monitoring. Results for the Mohave rattlesnake suggest a remarkable stability during the post-1988 years of this study (Figs. 4.2, 4.5), while the glossy snake seemed to be in a gradual decline during much of this study (Fig. 4.5; see also Mendelson and Jennings. 1992). Leaving aside the inadequate sample from 1994. results for C cerastes indicate (1) an early decline also seen in other rattlesnakes (see also Rosen and Lowe, 1996). (2) relative stability from 1988-1995, possibly with an eariy response to good rainfall years in the early 1990's, and (3) increased abundance in the late 1990's (Fig. 4.5). In dry autumns, C. cerastes activity during prime sampling time for it (late-September - late October) may be low, possibly contributing to the dip seen in 1997. The trajectory of the sidewinder population bears more similarity to other lizard-eaters than to the other rattlesnakes. Although this species feeds heavily on pocket mice as well as lizards (unpublished data; see Funk, 1965), young juveniles are apparently too small too small 196 for anything but lizards, which they eat avidly. Given the consistent reproductive output of C. cerastes (below), its population fluctuations may have been more strongly affected by changes in juvenile feeding success, growth, and survival, rather than by the effects of food availability on adults. It was clear that adult survival of C. cerastes was low for a rattlesnake, with 12 and 24 mo old cohorts overwhelmingly dominating the autumn samples. Despite the correspondences of snake population tluctuations to those of climate and prey, the most remarkable pattern remains the relative stability of snake populations during most of this study (Fig. 4.6; see Chapters I. 2, 5). Snake reproduction (except, as noted above, in M. flagellum and C. cerastes) was obviously higher In the good years of the early- to mid-1990's. as shown in Table 4.7 for the rattlesnakes (and see Andren. 1982). During 1989, virtually no recruits were produced, while after 1990 gravid females of several species were frequently recorded. Predation risk from larger, endothermic predators increased rapidly during 1991-1995 (Fig. 4.6), suggesting that predation within the trophic level, including predation by snakes on other snakes, may have significantly regulated snake populations. At ORFI, very few snakes, except coachwhips. bear injury scars (unpublished data), which probably reflects the inability of most snakes in the desert to escape from predators once they are attacked. Although some of the larger rattlesnakes survived the duration of this study, most species had low survival rates: the oldest marked R. lecontei was 4 yr of age, the oldest M. flagelliim confirmed was a 6 yr old. and the oldest C. cerastes confumed was an 8 yr old. Other colubrid species for which substantial data were available also appeared to be short-lived. Table 4.7. Annual variation in reproductive frequency of rattlesnakes in the Sonoran Desert of southern Arizona. Reproductive status was "gravid" if snakes would have given birth in the year indicated. This was determined by: (1) dissection, (2) palpation, (3) collapsed posterior-lateral body wall plus low body mass during August - September (cutoff date species-dependent) in years when all other females had high or normal body mass. All individuals May- July (inclusive) could be categorized; a small fraction during August-September could not, and were excluded. If snakes were less than 21-25 mo old or smaller than the smallest known adult, they were classified as juveniles, and also excluded from the tabulations. Reproductive Yr Species Status 87 88 89 90 91 92 93 94 95 96 97 98 total C. atrox gravid 2 1 3 1 2 7 7 2 1 1 1 1 29 not gravid 0 5 0 6 3 2 1 1 0 5 3 4 30 n 2 6 3 7 5 9 8 3 1 6 4 5 59 % gravid 100 17 100 14 40 78 88 67 100 17 25 20 49.2 C. scutulatus gravid 3 3 1 2 1 4 3 0 1 0 0 0 18 not gravid 1 3 3 2 0 1 1 1 0 2 0 0 14 n 4 6 4 4 1 5 4 1 1 2 0 0 32 % gravid 75 50 25 50 100 80 75 0 100 0 nd nd 56.3 C. cerastes gravid 1 3 5 1 1 5 6 1 1 2 0 3 29 not gravid 0 1 0 1 1 0 0 0 0 0 0 0 3 n 1 4 5 2 2 5 6 1 1 2 0 3 32 % gravid 100 75 100 ^0 50 100 100 100 100 100 nd 100 90.6 combined gravid 6 7 9 4 16 16 3 3 3 1 4 76 not gravid 1 9 3 4 3 2 2 0 7 3 4 47 n 7 16 12 i . 8 19 18 5 3 10 4 8 123 % gravid 86 44 75 •> 1 50 84 89 60 100 30 25 50 61.8 198 V. CLIMATIC AND TROPHIC CORRELATES OF POPULATION FLUCTUATIONS CHAPTER SUMMARY I explored the correlation structure among population trajectories in a vertebrate ecosystem monitored over 12 yr (1987-1998) at Organ Pipe Cactus National Monument, in southern Arizona. Prey taxa (small mammals, lizards), predators (endotherms. snakes), and productivity (using precipitation as a proxy) were included, and possible competitive interactions among prey were also considered. Using a multi-step path analysis approach. I attempted to identify cau.sal variables for population changes. This was made difficult by an ecosystem-wide upwelling of abundance associated with strong El Nifio-effect winter rains during the middle ponion of the study. Observed abundance of each prey taxon had a negative correlation with its population change, indicating widespread density dependent population regulation in the 12 yr frame of study. This effect held up in all multi-variate path analyses. Most (if not all) prey taxa exhibited enhanced reproductive output during and after seasons of strong rainfall, yet many bivariate correlations and path coefficients failed to reflect positive effects of rainfall. In rodents, bivariate coefficients variously showed strongest positive effects from winter rains (perognathines), summer rains {Dipodomys merriami), or annualized rainfall (woodrats). In lizards, summer rainfall showed nearly ubiquitous 199 positive coefficients with population change, while winter rain had negative coefficients (except in the winter-active, spring-breeding Uta stansburiana). An index of predation pressure had negative bivariate correlations with population change in almost all taxa. For perognathine rodents, and especially for most of the lizards, predation pressure carried large coefficients that were sustained throughout path analysis procedures holding other variables constant. For the larger small mammals, predation coefficients had moderate or weak negative correlations with population change, and the statistical effect of predation tended to disappear when abundance of conspecifics was a co-predictor variable. Inter-specific competition hypotheses performed less well in the multi-variable analyses than did precipitation, intra-specific density-dependance. or predation. The ecosystem appears to be driven by responses to rainfall-induced variations in carrying capacity, modulated by predation and competition. The results support a regulatory role for predation in Sonoran Desert lizard and perognathine rodent populations. Inter-specific competition (or possibly, predation) from other lizards was suggested for some lizard species. Competitive dominance of D. merriami over the smaller, perognathine heteromyids was also supported. For kangaroo rats, ground squirrels, and woodrats, the statistical analysis gave inconclusive results for predation: I could not convincingly separate competitive and predation effects that were probably subsumed in intra-specific density-dependance. For these larger species, bivariate correlations, as well as anecdotal observations, suggest that the role of predation deserves continued consideration. Monitoring of the predator assemblage is likely to be relevant 200 for interpretation of the results of prey monitoring, which is ongoing. INTRODUCTION Predation is the most immediately potent inter-individual ecological interaction, with instantaneous and absolute fitness consequences. Yet it is difficult to observe, and there remains much debate about its importance for animal population regulation and community organization. Although, vertebrate predation has received increased attention in recent years (see e.g., Kotler. 1984; Hansson, 1987; Hanski et al.. 1991; Jaksic et al.. 1993; Marti et al., 1993; Kotler et al., 1994; Meserve et al.. 1996). it remains remarkably controversial (Erlinge et al.. 1983. 1984. 1988; Hanski etal.. 1993; Hekkilii et al.. 1994; Henttonen et al.. 1987; J^drzejewski etal.. 1996; Korpimaki et al.. 1991. 1994). Population regulation stands at the center of real-world ecology, where distribution and abundance, and the factors that control them, are the key facts and processes determining what we see around us. Yet, even the most basic words are contested. In this paper, by the phrase "population regulation" I mean, "to determine or cause the observed abundance"; and I note that such regulation may be density dependant or density independent. Evaluating the relative importance, consequences, and dynamics of ecological interactions (predation, parasitism, competition, mutualism), although difficult, is not trivial. It perhaps remains a central task in ecology. In this paper I attempt to evaluate especially the potential importance of predation in a vertebrate ecosystem. As our interest in natural landscapes, biodiversity conservation, and ecosystem management grows, it becomes more clear that we need a functional understanding of 201 ecological dynamics. The recent upsurge of interest in nature conservation has resulted in increasing efforts to monitor animal and plant populations (e.g., Haivorson and Davis, 1996). The oft-stated objective is to detect trends that indicate problems amenable to management. The monitoring paradigm tends to focus on methodology (Thompson et al, 1998) rather than of considerations of ecosystem dynamics. This is probably because the science of ecology has not properly resolved, or even focused on, the matter of population regulation. Recognition of trends, and evaluation of their significance (if any), will be accomplished using synthetic ecological logic well before individual data trends can be deemed statistically significant. The requirements for successful synthesis of monitored trends with their causal and dynamic processes should be firmly designed into ecological monitoring programs. Today, most ecological monitoring focuses on single species; a few programs involve representative indicator or index taxa that partially represent ecosystem diversity; very few, if any, are firmly established within the full context of ecological interactions. As humanity creates ecological problems of increasing scope, dynamic understanding, rather than piecemeal approaches, becomes increasingly important. A mutualistic interplay between a theoretically-grounded monitoring program and developing ecological science would be beneficial to both. Over the past 10 to 15 yr, ecologists have begun to recognize the great importance of predation in vertebrate populations. Classically given, but subsequently and quite recently much revisited, exemplars of predator-prey systems have only recently become accepted as examples of predation regulating prey populations. For the moose, the wolf 202 and bear; for the arctic hare, the lynx; and for microtine rodents, in Fennoscandia the weasels, and in North America other predators, are now recognized as important factors in population regulation (see Introduction to the Dissertation). Although many key authors reiterate the mantra of field experimentation for solving ecological problems, each of the systems listed is at bottom studied by a monitoring program. There is not a substitute for hard observational data on populations, regardless of how hard ecologists wish for one. Predation and disease are the least adequately treated ecological interactions. Predators are often larger, more mobile, less numerous, and more difficult to handle, study, or monitor than competitors. I will examine the correlations among time-series representing the dynamics of a Sonoran Desert ecosystem in southern Arizona. .A key objective is to assess the need for monitoring of predation for the ongoing Ecological Monitoring Program (EMP) at Organ Pipe Cactus National Monument (ORPI). Currently established EMP protocols cover climate, nocturnal rodents, and abundant lizards and birds, and to these I add my data on endothermic predators and snakes. Rodents are the best known component of the vertebrate assemblage of the North American Desert. Small mammals, including lagomorphs and nocturnal and diurnal rodents, are competitively important as granviores (e.g., Davidson et al., 1980) and perhaps foliavores (Chew and Chew. 1970; McAuliffe, 1986; Ernest. 1994) in Southwestern desert ecosystems. Although competition has been a dominant paradigm in desert rodent ecology, predation has become a persistent theme more recently (Thompson, 1982b; KoUer, 1984, 1985. Kotler et al., 1988, 1991. 1992, 1993a«&b; J.S. 203 Brown et al., 1988; Korpimiiki and Norrdahl, 1991; Longland and Price. 1991; Pierce et al., 1992; Diaz. 1992; J?drzejewski et al., 1993; Meserve et al.. 1993; Lagos et al.. 1995; Abramsky et al., 1998; and others). This growing literature suggests that predation risk shapes the utilization of foraging habitat, contributing to microhabitat partitioning. Thus far, the literature has not shown that mortality from predation regulates desert rodent populations, or that it directly structures rodent communities—although J.S. Brown et al. (1994) describe how it could. The results presented here suggest it may do both. North American Desert lizards comprise one of the best-studied lizard assemblages. Research has focused on natural history and demography, while only a few experimental studies have addressed inter-specific competition (Dunham. 1980; Smith, 1981; Tinkle. 1982). Inter-specific competition appears to be variable and moderate. With high niche overlap, competitor removals produced changes in energetics and survivorship (Dunham. 1980); with intermediate niche overlap, a survivorship response was observed (Smith. 1981); with more distinctive niche partitioning (as seen in lizards in the present study) little experimental response was found (Tinkle. 1982). Our understanding of competition in lizard community ecology (Dunham. 1983) has not yet been matched by a clarified appreciation of predation's role. A small number of long-term monitoring projects have been carried out on desert lizards (Creusere and Whitford. 1982; Medica et al.. 1973; Tanner and Krogh. 1973; Tinkle etal.. 1993; Turner et al., 1969a&b. 1970, 1982; Whitford and Creusere, 1977; see Vitt, 1991), but less attention has been paid to population regulation or predation (Turner, 1977; see Andrews and Wright, 1994). Turner et al. (1982) empirically modeled 204 population regulation in the side-blotched lizard {Uta stansbiiriana) in the Mohave Desert of southern Nevada. They found that density dependence was expressed in changing growth (and consequently, future clutch size), clutch frequency, and adult survivorship. Ferguson and Fox (1984) reported a density dependent survivorship response in juvenile U. stansbiiriana. Turner et al. (1982) concluded that their model was moderately successful, but that its greatest failing was probably in modelling predation. In contrast, data from another intensive study of the same species (Tinkle, 1969) revealed density dependence only in fecundity (see Turner. 1977. pp. 199-201). Predation. parasitism, density-dependance. and intra- and inter-specific competition can be presumed to contribute to lizard population dynamics and community structure, but their roles and importance remain unresolved. The present study began In 1987. with monitoring protocols first carried out in 1987-1989. depending on taxon. Full implementation of all protocols (except for birds) began in 1991. Thus, events of the 1990's are the most precisely known. The 1987 beginning was early in a very hot drought that followed the wettest period at ORPI (1977- 1984) since at least 1949. During 1983-1985, high densities of nocturnal rodents (Cockrum and Petryszyn, 1986) and lizards (Rosen and Lowe. 1996), as well as snakes (personal observations), were observed at ORPI. Endothermic predators were also abundant at that time, and certain fluctuating taxa, including the bam owl and badger were apparent (personal observations). The drought of the late 1980's broke with strong rains in summer 1990, and although summer rains were meager or average until 1998 (they were good but very late, and locally variable, in 1996). winter rains were quite 205 strong (1990-1, 1991-2, 1992-3, 1994-5; 1997-8), providing large seed and insect crops for rodents and lizards. This study, then, follows the progression of ecological change associated with a climatic wave involving a consistent deviation of rainfall from its normal pattern and mean. Such deviations can, however, be considered normal in the variable climate of the desert. There has been continued debate about whether observational (i.e., "natural history"), monitoring, and correlational approaches are suitable for understanding ecosystem function. For example. Korpimiiki and Krebs (1996) have suggested that further understanding of small mammal cycles must come from manipulative e.xperiments. They further criticize the "accounting" method of tallying up predation as a method of studying population regulation. On the other hand. Power et al. (1996) argue that a multifaceted approach is appropriate to demonstrate key facets of ecosystem dynamics. Here 1 hope to establish some of the groundwork for (I) using monitoring to reveal long-term trajectories and their correlative relationships, which might then be combined with (2) accounting of actual abundances and physiological requirements for prey consumption (to reveal the magnitude of predation, competition, or other interactions) and (3) demographic study to specify population dynamic effects of the interactions. Within an expanded natural historical framework such as this, experimentation should serve as the probe to examine intractable questions, and demonstrate or reject hypotheses based on and about the observed functioning of the un perturbed ecosystem. 206 METHODS The objective of the analysis presented here is to explore the linear correlation structure among population trajectories of potentially interacting species and climate. The goal of this is to determine which bivariate correlations may reflect parallel causation, and which are most robust within the multivariate context of rainfall, competition, and predation. ESTIMATION OF ANNUAL PREY TAXA ABUNDANCE TIME-SERIES Nature of the Data Monitoring results for prey taxa (small mammals, lizards) from 1987 or 1988 through 1998 are available in Chapters 1 and 2, where it is argued that these data represent changes in population size. In addition to the EMP protocols, intensive data were obtained at one study area (the ORPI Snake Study Area [OSSAj) using time- constrained search (TCS). sand-road transect, and drift-fence trapping. For most taxa. one or more of these methods was used in each year, but the efficacy of each method varied among taxa. In 1994 and 1995. only limited work was performed using these methods, although the EMP protocols were carried out in full. Starting in 1988 (rodents) or 1989 (lizards), formal monitoring protocols based on trapping grids (rodents) and transects (lizards) were performed once/yr (rodents) or twice/yr (lizards) through 1998 at sites with permanently-fixed quadrats or transects. Some sites, "Core I" NPS (National Park Service) monitoring sites, were monitored in every year. 207 For rodents, valley habitat and desert mountain (rockpile)-associated habitat had very different faunas (Chapter I), and may have had somewhat different dynamics: thus, only Core I sites in valley habitat were analyzed here. During 1989 and 1990. the rodent grid monitoring protocol was carried out at only one site, a Core I EMP study area within the OSSA site. For most years, modified Lincoln-Petersen method (L-P) computations could be performed on the available rodent data. The quality of the grid data was rated as 5-sites L-P > 5-sites non-L-P> I site L-P > I site non-L-P, and for each year, only the best among these methods was included as the grid data for further analysis. For lizards, faunal structure and dynamics were similar between valley and rockpile habitat, so standardized lizard line transect results from all EMP sites are utilized here. Most transects were run (twice) in every year, and any that were not run during at least 7 of the 10 yr were excluded from the present analysis (Chapter 2). For both rodents and lizards. University of Arizona research personnel carried out the standardized protocols during 1988-1990. NPS Resource Management biological technicians, trained and assisted by the research personnel, carried out the field work and data entry subsequently. Snake data used here is restricted to the OSSA site and surrounding valley terrain, and most of the endothermic predator monitoring records also come from this region and roads leading to it. 208 Data Reduction Although the EMP monitoring protocols yielded the best single dataset on nocturnal rodents and diurnal lizards, the other methods' records added significant scope to the study, both in duration and taxonomic coverage (see Chapters I and 2). Since data from different methods differed in quality. I weighted results from the various methods according to my best judgement. Based on considerations presented in Chapters 1 and 2. the NPS monitoring dataset was weighted most heavily in all cases except where sample sizes were small, and much smaller than for alternate methods. In order to perform the weighting, each dataset was first standardized to a mean of one. with resulting variation only among years. The weighting protocol is outlined in Appendix II. For rodents. EMP grid results carried weighting equal to the sum of weighting of all the other methods: they were weighted twice as heavily as the best alternate method and four times more heavily than other alternate method(s) for which data were available. For lizards, line transect and drift fence data were closely congruent except where one or the other dataset had low sample size. The transect data were weighted twice as heavily as the drift fence data, which was supported by strong sample sizes except in two lizard species. In one case it was weighted four times as heavily as drift fence results, and in one other the drift fence results were excluded, based on sample size considerations. Exceptions are noted in Appendix II. No EMP protocols were available for lagomorphs, sciurids, and nocturnal lizards (the gecko Coleonyx varie gains); for these the available datasets from OSS A were weighted using the same considerations. 209 ESTIMATION OF ANNUAL PREDATION INTENSITY Annual variation in the intensity of endothermic predation pressure on various animal groups was estimated from data presented in Chapter 3 (Tables 3.2 and 3.6). The abundances of predators within years (columns in Table 3.2) were multiplied by the proportion of each prey type in the predator's diet (i.e.. columns in Table 3.6), and these products summed down the list of predator taxa to give the best index of predation intensity on that specific prey category for each year. 1987-1998. The summation was then repeated for each prey taxon. Values of predation intensity were standardized to a mean of one. The year-to-year variation of snake predation was set equal to the variation in abundance of snakes as measured by standard drift-fence trapping procedures, weighted by mean total body mass for each snake species in each year (Chapter 4). Specifically, for each snake species, trap-capture rate was multiplied by the mean body mass for each year, and standardized by dividing through by the grand mean mass of all snakes captured (163 g). Data for 1994 and 1995 had to be combined to produce an estimate for those years (see Chapter 4), and the resultant value was used for both years here. For each year, the values for the snake species were summed, and a standardized index (mean 1.0) was derived directly from these annual sums. The values for predation pressure of snakes on lizards was based only on the following snake species: coachwhip Masticophis flagelliim, long-nosed snake Rhinocheilus lecontei, patch-nosed snake Salvadora hexalepis, sidewinder Crotaliis 210 cerastes, night snake Hypsiglena torquaia. For predation pressure on mammals, only the following snake species were considered: western diamondbackand Mohave rattlesnakes C atrox and C. sciitulaiiis. gopher snake Pitiiophis melanoleucus, and glossy snake Arizona elegans. Although this oversimplifies the diet of smaller and less common species of snakes (unpublished data), the resulting error is very small (see Fig. 4.1). Lizard predation pressure on other lizards was constructed specifically from the abundance of crotaphytine lizards (collared and leopard lizards. Crotaphyius, Gambelia) based on the mean value from lizard lines and drift fence trapping (Chapter 2). These values were then standardized to a mean of one. Predation intensity on any prey taxon was finally derived by combining values for endothermic, ophidian, and saurian predators. For lagomorphs, all predation was assumed to come from endotherms: the contribution of nestling predation by gopher snakes and adult lagomorph consumption by very large western diamondback rattlesnakes was neglected. For larger rodents (sciurids, Neotoma, Dipoclomys). endothermic predation was assumed to be twice as important as snake predation, and the standardized value for each weighted accordingly. For smaller rodents {Chaetodipus, Perognatlius). as well as lizards, I assumed that predation by endotherms and snakes was equally important, and combined the un-weighted values for each into a single set of mean values expressing the variation of predation intensity among years. In perognathines, the value used for snake predation combined the trap-value for mammal-eating snakes with the fecundity of rattlesnakes, the trap value counting twice as heavily (thus, the fecundity of rattlesnakes in the preceding year counted as 1/6 of the index). For the side-blotched lizard {Uta 211 smnsbiiriana), the predation index was taken as the mean of standardized values for lizard predation (as described here), crotaphytine abundance, and whiptail lizard abundance. This scheme was also tested for the gecko Coleonyx. Although the predator monitoring results do not distinguish between increases in predator abundance or individual activity levels, they represent the product of those two values (i.e., total predator activity), which is close to the proximate effector in predator- prey dynamics. CLIMATIC DATA AS A PROXY FOR PRODUCTIVITY I assumed that precipitation was a suitable proxy for overall plant growth and seed production, and hence represents food availability for small herbivorous vertebrates (see e.g., Abst. 1987). Rodents and lagomorphs are dependent on seeds and foliage for food. During wet summers and winters, arthropod abundance also increases markedly (see, e.g., Ballinger. 1977; Dunham, 1978, 1980, 1981: personal observations), and each spate of moisture produced a marked and often dramatic increase in activity of ground-walking arthropods (unpublished data), and presumably of other arthropods. Therefore, I also use precipitation as a proxy for food availability for small insectivorous vertebrates. Data on predator and prey abundance pertains to desert valley habitat at ORPI. and is somewhat weighted toward the OSSA site (see above). I weighted the ORPI-wide precipitation totals twice as heavily as the data pertaining only to the OSSA site, for use in correlation and regression analysis. 212 STATISTICAL METHODS: CORRELATION AND PATH ANALYSIS Path analysis (Sokal and Rohlf, 1981, Mitchell, 1993; Loehlin. 1987) was used for prey taxa (small mammals, lizards) to explore the correlation structure of population change {Apop) with respect to potential causative factors -- precipitation ippt), predation iprecf), and intra- and inter-specific competition (intra-, inter-). The standardized time- series dataset was used to create a correlation matrix of all species abundances and population growth rates, measures of ppt. and the estimate of taxon-specific predation pressure. Species Apop'?, were expressed as both - N ,-) and (In [population YR / population YR j-J). The difference equation produced a less skewed and kurtotic distribution than the logarithmic, and gave higher coefficients of correlation and determination across the board. Hence. I used the difference equation result in all ca.ses except for one, described below. The ppt measures were winter, summer, annual (i.e., calendar), seasonal-annual (winter YR y + summer YR y), and seasonal-interannual (summer YR ^ + winter YR t^,). The ppt and pm/data were also time-lagged ahead (T+l) and behind (T-1) one year, and included in the data for correlation. Path coefficients were computed from the resultant correlation matrix using Microsoft Excel 5.0 and PROC CALIS (SAS Institute Inc., 1989). only the latter being used for more complex models. The correlation matrix revealed that intra-specific abundance (intra-) was the most consistent single correlate of Apop, and it was therefore incorporated into most analyses. Three sets of path models (Figs. 5.1-5.3) were explored. Models I and II were 213 computed following Sokal and Rohlf (1981) and Mitchell (1993) on Microsoft Excel 5.0 spreadsheets. Additional models involving more predictor variables were also examined, using the RAM method in PROC CALIS, as were the Model III computations. Model I was expressed as two-way path diagrams (Fig. 5.1) that are equivalent to two-variable multiple regressions standardized to yield path coefficients. (Path coefficients are standardized multiple regression coefficients - meaning they represent the amount of change produced directly in the independent variable [always Apop in this case] by the specified causal variable, expressed as a fraction of the total variance of the independent variable). Path coefficients (unlike correlation coefficients) can be greater than one. since the variance produced by one causal variable may be offset by effects of other causal variables. Model I comparisons were run with pred combined with intra-, inter, or with the various measures of ppt. I computed path coefficients for ppt and preci with their time- lags synchronized — both within YR j, both in YR t-.,, and finally both in YR and also with offset time-lags. However, the various combinations of time lags revealed little not shown by the synchronized lagging of pred and ppt. so I chose the synchronized time- lags for further study. For Models I and II, I re-computed everything for each measure of ppt (winter, summer, annual, etc.) for each time-lag, and the coefficients presented are the mean of the ppt measures within each time-lag. I did this to avoid arbitrarily presenting the best fit. Model II was expressed as sets of three-way relationships among ppt. intra-, inter- and pred (Fig. 5.2). The basic model involved deriving all path coefficients from 214 PPT A POP PRED Other variables used: INTRA- INTER- Fig. 5.1. Path diagram with two predictor variables influencing population change. Unknown influences (unexplained variance) on the predictors and criterion variable were modelled but are omitted from this diagram. This model is equivalent to a two-way multiple regression. PPT Also used: INTER- INTRA- PRED Fig. 5.2. Path diagram with three predictor variables intluencing population change. This model is equivalent to three-way multiple regression. As in Fig. 5.1, arrows ("or unexplained variance are omitted for clarity in the diagram. 216 seasonal-annual ppt (i.e.. winter YR T- + summer YR species population size (intra-, inter-), and pred, all in YR t. I also computed this model for summer and winter ppr alone. Based on findings in Model 1.1 examined results only for equivalently-lagged variables. I found that the basic correlation structures found in the bivariate correlation table and path analysis Models I-III were not markedly changed by shifting time lags for most variables. The changes that occurred were either erratic or were obvious from examination of the bivariate results. Cascading effects of time lags through the correlation structure were. thus, not observed: and YR ^ is of primary interest regardless, since it incorporates the immediate effects of the systems parameters, all at about the time frame when they would be expected to have direct effects. Therefore. I have focused the presentation primarily on the non-time-lagged models. While the path analyses presented here are exploratory, and a large number of coefficients are examined. 1 present critical values at P = 0.05 for the bivariate correlations and basic path analysis models. These are indicative of the relative strength of each individual model, on the assumption that it was the only one being considered. Although this assumption is obviously violated, this was still the most informative and intuitively useful metric of model strength. Statistical significance must accordingly be assessed with caution, in light of the multiple comparisons examined. The critical values I present are derived identically to those for testing the null hypothesis that R = Q with N - K -1 degrees of freedom in a multiple regression problem (see Sokal and Rohlf, 1981), where R is the coefficient of multiple determination (and R' 217 is the proportion of variance accounted for by the model), N is the sample size, and K is the number of independent (or predictor) variables. For bivariate correlations, I present critical values for I-tailed tests of effects in the direction expected (e.g., predation is expected to have a negative impact on population size, and so on). Obviously, with N = 10 or 11 yr of population change, there is little possibility for statistical significance to be demonstrated in adding predictor variables in a multiple regression study (the general rule of thumb is that there must be 10 - 20 cases for each predictor variable). This is particularly difficult here, where several variables are apparently involved in determining population change. Therefore, for the path analysis models, 1 present critical values of R' and evaluate the effect of adding additional or differing variables on the suite of results of taxon-specific models under Models 1 and II. In considering these results as a whole, caution must be exercised against pseudoreplication. Many taxa were responding more or less in parallel to the same underlying factors; each individual taxon does not represent an independent set of responses. Conflating the significance of bivariate statistics across species may strengthen statistical confidence in the correlations, but it does not entirely alleviate the problem of parallel causation. Further, the multiple regression methods that underlie path analysis may be sensitive to normality and linearity assumptions. Inspection of the sample distributions shows they were not highly skewed or kurtotic, but sample sizes were too small for precise normality testing, and the magnitudes involved in non-normality effects are not accurately known. Models I and II clearly indicated that some of the results of simple bivariate 218 correlation could potentially be misleading. During much of this study, precipitation was followed by enhanced reproduction (evidenced in the field by gravid females and numerous young) and was generally followed, as may be expected, by increasing populations of prey, and. generally with additional time lag. predators. The several good winter rain years of the mid-l990's exacerbated the tendency for colinearity to occur under this causal arrangement. In some cases. Models I and II suggested that negative correlations of precl and Apop might reflect parallel causation, rather than predation as a direct effect. This was particularly a problem when pred and intra- were considered together, because intra- cannot be assumed to represent primarily or only intra-specific competition. It is possible or probable that behavioral and populational responses of predators to intra- may contribute to negative correlations of intra- and Apop. I used Model III (Fig. 5.3) to attempt to ameliorate the problem of colinearity and the potential cohabitation of predation and competition in the intra- variable. I developed an analog of a latent variable model (see Loehlin, 1987) for the system. The primary objective was to attempt to examine the co-effects of intra- and pred on Apop simultaneously, but separately from the co-effects of intra-, inter- and ppt. on Apop. I created latent variables analogous to birth (b) and death (d), based on the idea that Apop = b - d. Since I didn't directly measure b or d, I simply used Apop and -Apop as surrogates. To avoid singular correlation matrices, I used the logarithmic version of Apop for b and the difference version for d (since population growth is more likely than contraction to be an exponential phenomenon). winter PPT summer PPT energetics INTER- INTRA- mortality PRED Fig, 5.3, Palh diagram for ihe summary model of variable inier-relationships. Arrows for unknown variance, and lor the bivariate correlations among predictor variables, are omitted for claiity. lO 220 I incorporated factors into the most realistic model possible with the available parameters. Model III incorporated 5 variables (2 seasonal rainfall parameters (summer ppt for YR J and winter ppt for YR [as surrogates for food productivity], and 3 biotic interactions, intra-, inter-, and pred). The ppt variables and the two competition- involving variables intra- and inter-) were combined in a 4-way causal scheme affecting population growth, while intra- and precl were involved in a separate 2-way post-diction of Apop. The two sets of pathways were allowed to be constrained via interaction through the bivariate correlations among the causal variables, as in a simple multiple regression. I computed Model III in taxa that had both strong monitoring data and plausible interspecific competition hypotheses. For Neotoma. I assumed that sciurids were the most plausible competitor based on diet. size, and the bivariate correlation coefficient; computations assuming competition with other taxa all produced no support for an inter specific competition hypothesis (i.e.. there were no negative path coefficients for inter-). Since none of these models for Neotoma are suggested by the literature. I simply present the sciurid version. For Dipodomys merriami, I used competition hypotheses based on the literature. I examined a model based on competition with the larger banner-tailed kangaroo rats (primarily D. spectabilis. although D. deserti is also reported at ORPI), but this produced a positive path coefficient and was ignored because the larger taxa are not common at ORPI. I also examined sciurids as potential competitors, based on J.S. Brown (1989). again with uninteresting results, and I examined the smaller heteromyids (the perognathines) as potential competitors of D. merriami, based on their very great 221 abundance in the Sonoran Desert and similar diet. For the perognathines. I examined models of competition with D. merriami, based on J.H. Brown and Munger (1985) and Heske et al. (1994). The pairwise lizard species interactions that could occur at ORPI have not been previously studied. Nonetheless, exploitative competition among insectivorous lizards with microhabitat overlap (i.e.. terrestrial versus arboreal; see e.g.. Viit et al.. 1981) can at least be reasonably postulated. The spiny lizards at ORPI {Sceluporiis magisier, S. clarkii) are terrestrial-arboreal, and thus may have many potential competitors. I assumed that crotaphytines are among the potential competitors of Sceloporus because both are large and eat small lizards and large arthropods; and I assumed that the ant-specialist Phrynowma may compete with Sceloporus, which also eats ants in substantial quantity (Parker and Pianka. 1973; Pianka. 1986; Vitt and Ohmart. 1974). I also explored a variety of other models, including 4- and 5- way standardized multiple regressions, but found that Model III produced most consistent and intuitively reasonable results. I discuss the implications of the coefficients derived from these the other models, but present Model III as the best available summary of the correlation structure in this dataset. The results of the path analysis are not intended as a final or definitive statement of the correlation structure of this data. Sample sizes were small, and I maintained a simple linear statistical approach throughout the analysis. Both non-linear and threshold effects are entirely plausible for the systems modelled here. Nonetheless, the path analysis results illuminate key aspects of the data structure. RESULTS STANDARDIZED TIME-SERIES All taxa decreased during the late 1980's drought, followed by a generalized increase during the wet years of the mid-l990's (Figs. 5.4, 5.5; Appendix III). There was a subsequent generalized decline, with variable onset during 1993-1996 depending on taxon. By 1998. at the end of the time-series, many taxa had begun to increa.se again. The data for the 1990's are particularly strong, and document an oscillatory response of the ecosystem to a period of enhanced winter precipitation. Within this broad pattern, there were several distinctive variations (Fig. 5.4). Small heteromyids. the perognathines C. penicillatus and P. ampins, increased rapidly in the early 1990's. and began to decrease after 1992. with a slight recovery in 1995. Lagomorphs began to decline after 1993. The medium-sized heteromyid D. merriarni increased more slowly than the perognathines, and decreased after 1994. Larger rodents (sciurids. D. spectabilis. N. albigiiki) decreased relatively late in the mid-l990's. During the early-mid 1990's. species that were rare in the Valley of the Ajo, including Peromysciis ereniicus, Sigmodon cf. arizonae. and apparently. Thomomys bottae (see Chapter 1). also increased, during 1991-1993. Lizards showed a markedly more uniform pattern than thesmall mammals (Fig. 5.5). After 1990-1991 increases, declines began quite early, in 1992-3. continued through 1995, and were followed by large increases starting in 1996 or 1997. The crotaphytines and C. variegatus seemed to differ from this pattern, although data are scanty for them. One species, U. stansburiana, was different, showing a relatively lengthy, fluctuating 223 Standardized Small Mammal' Population Trajectories, ORPI -Neot —o—mer —pen •••••amp »—lagom —<>—sciurids spectab — Onyc S 2.0 a 2.0 eo s> g ^Mco^in^Koo OO 00 O 9 O) 0) 0) 9 91 0) O) 9 9 O O) Predation pressure index for: Predation pressure index: Neotoma Olpodomys sclurlds lagomorphs perognathlnes all endotherms 5 1.5 a 1.0 c 0.5 i« 0.5 * in « CO 010)0)0)0)0)0)0) Precipitation: Precipitation: —YR: calendar winter •YR: winter + summer ••••winter 3 1-0 a 1.0 m (Q cn 0.5 c 0.5 N 00 O) cv « u7 r* OO 9 O) OI A 9 s s s S S S! SO) O) SO) 0) 0) Fig. 5.4. Small mammal population trajectories at Organ Pipe Cacms National Monument, 1987-1998, shown with other variables used in the analyses. Accuracy is highest for the 1991-1998 and for the taxa shown in the upper left. Derivation of the standardized values is detailed in the Methods section. 224 Standardized Lizard I Population Trajectories CnemI U. ornatus -<]-Calli ••><' Scalop Uta stansburiana Species potentially interacting with Uta - Coleon -Oipso ' P. solare Cnemi crotaphytlnes n 0.5 01 @Ol Ol01 01Ol Olw 01w Ol 01w 001 01Ol OI 01Oi 01 - PPT, year • • PPT, summer - PPT, year — - predatlon: Uta -predatlon: lizards PPT, winter I 0.5 CO ^ M O 9 tn s s SS01 Ol2 SOl 2Ol s s Fig. 5.5. Lizard population trajectories at Organ Pipe Cactus National Monument, 1988-1998, shown with other variables used in the analyses. Accuracy is highest for 1989-1998 and for the taxa shown in the two top graphs. Derivation of the standardized values is detailed in the Methods section. 225 increase during most of the study. Predator pressures (Figs. 5.4. 5.5) were at minima in 1989 or 1990. increased to maxima in 1993-1996, and had a subsequent low point in 1996-1997 before increasing again. Peak times of estimated predation pressure varied among prey taxa. SMALL MAMMALS Bivariate Correlation Structure Positive correlations between population change and precipitation were seen in most small mammals, although the effects differed markedly among taxa. and were significant at P = 0.05 in only 5 of 18 cases involving summer YRj and winter YR (Table 5.1). For most taxa. it is precipitation in these two seasons that should lead to reproduction that would manifest itself through increased reproduction in the monitoring results for YR (recall that Apop = N - N However, for perognathines, there was a rapid summer breeding response to summer rains, which was seen in the monitoring results (Table 5.1) especially because this ta.xon was so effectively sampled during late- summer drift-fence trapping. Although winter rains were important for several taxa, as stated in the literature (Beatley, 1969; Petryszyn, 1982), summer rains also showed high bivariate correlations for D. merriami and C penicillatus. A summer rainfall effect may be appropriate for D. merriami. which may breed in autumn (Hoffmeister. 1986; Kenagy and Bartholomew, 1985). For YR J (the base year; see Table 5.2), pred had a negative correlation with Apop in 7 of 9 cases involving the taxon-specific indices (Pred TAXON)' ^ cases using Table 5.1. Correlation coefficient.s between small mammal population change and precipitation. Data arc for 1987-1998 (11 annual population changes) at ORPI. Critical value (P = 0.05) for the hypothesis that r > 0 is r = 0.52. For the hypothesis thai correlation for summer is higher than that for winter, the critical value f»)r rs - r^ is ca. 0.80. Highest values are indicated in boldface. Neotoma l)ipothmiy\ Onychomys Dipodomys Cluu'lotlipus Perof-natliiis all perog- lagomorphs sciurids alhifiiila spcctahilis torridus mcrriami penicillatus ampins nathines previous yr; winter rain -0.34 0.51 0.40 0.11 0.41 -0.37 -0.56 -0.52 -0.56 summer rain 0.27 -0.50 O.ll 0.21 0.24 -0.13 0.33 0.37 0.21 base yr; winter rain 0.19 -0.05 -0.26 -0.08 -0.13 0.27 -0.67 -0.65 -0.68 summer rain (W7 0.66 0.28 -0.19 0.27 0.54 0.03 0.14 0.14 larfn'l yr: winter rain 0.61 0.16 (».45 0.56 0.28 0.07 0.40 0.59 0.29 summer rain 0.32 -0.45 -0.31 0.13 -0.22 -0.04 0.43 0.19 0.49 mean ofyrs: winter rain 0.15 0.21 0.19 0.20 0.19 -0.01 -0.28 -0.19 -0.32 summer rain 0.22 -0.10 0.03 0.05 0.10 0.12 0.26 0.23 0.28 lO to OS Table 5.2. Correlalion cocfficienls between small mammal population change and predation indices (defined in text). Data for 1987-1998 (11 uniuial population changes) at ORPI. Critical value (P = O.O.'i) lor the hypothesis that r < 0 is r = - 0.52. Largest negative values are indicated in boldlacc. Ncouma Dipodomys Onycliimiy.s Dipodomys Chtu'iodipiis Pcrofiiuilhus all perog- iagoniorphs sciurids albifiula spcctabUis lorrklus iiu-niami penUillaiiis (iiiiph(s nuthines previous yr; t.NDO. -0.51 -0.14 -0.09 -0..30 -0.32 -0.36 -0.26 -0.20 -0.26 Prc(i,^r,a -0.34 Pred.v(((;K •0.24 -0.06 -(K37 Prcd/j//i(> -0.31 -0.35 Pred;./,K(;,; -0.31 -0.29 -0.39 haw yr: hNIU) -0.12 -0.03 -0.18 -0.20 -0.11 -0.25 -0.46 -0.48 -0.43 Pred,^,io -0.25 Prcdi( •/(//, 0.05 -0.12 0.(M) Pred/j//.^, -0.10 -0.41 Pred,./.^,,,; -0.78 -0,72 -0.41 turfiet yr: Pred^/j 0.16 0.24 0,05 0.22 0.06 0.09 -0.26 -0.09 -0.33 Pred/.((;,j -0.07 Predyc/,/^ 0.21 ''red/v^Yj/ 0.01 -0.15 Prcd/j//Y/ 0.01 0.19 Pred/y/„j„ 0.12 0.27 0.09 Table 5.3. Correlation coetficienl-s between .small mammal population chanijes and abundances. Data are lor 1987-1998 (11 annual population changes) at ORPI. Critical value (P = 0.05) for the hypothesis that r < 0 is r = -0.52. Intra-specific correlations are indicated in boldface; correlation coefficients with putative interspecific competitt)rs (based on the heteromyid literature) are indicated hy italic bold. Population Change of: Nfoionia Diptxlomys ()n\rli<>iiiy.\ Dip lO lo 00 Table 5,4, Time-lagged correlations between small mammal population changes and abundances. Data are lor 1987-1998 (11 annual population changes) at QRPI. Population rhanjje of: Neoioma Dipodomys Onychomys Dipodomys Cliaetodipus I'enifiiiatlms all perog- lagomorphs sciurids albifiula spectabiiis lorridus merriami penicillatus amplus nathines Time Lag (yr.l 3 -0.36 -0.13 0.19 -0.45 -0.32 -0.23 -0.20 -0.14 -0.17 2 -0.37 -0.58 0.04 -0.30 -0.25 -0.37 -0.20 0.03 -0.30 1 -0.38 -0.t)8 -0.42 0.04 -0.38 -0.12 -0.55 -0.58 -0.60 0 -0.51 -0.56 -0.66 -0.63 -0.50 -0.60 -0.54 -0.58 -0.55 putative competitor neot sped lagom mer neol |>erogs mer mer mer 3 0,19 -0.11 -0.24 -0.02 0.02 -0.41 0.65 0..59 0.71 2 -0.28 -0.53 -0.60 -0.48 -0.11 -0.22 -0.23 -0.42 -0.24 1 -0.26 -O.ll -0.1)8 -0.16 -0.13 -0.22 -0.42 -0.43 -0.36 0 -0.27 -0.01 ().(M) 0.38 -0.64 0.48 -0.29 -0.24 -0.42 putative competitor sci neot siu sciu mer sjiectab spectab spectab s|>ectab 3 -0.19 -0.64 0.01 0.21 -0.27 -0.22 0.02 0.01 0.02 2 -0.67 -0.38 -0.11 -0.50 -0.22 0.07 0.09 0.17 0.21 1 -0.27 0.08 -0.02 -0.13 -0.68 -0.71 0.04 -0.08 -0.08 0 -0.12 -0.18 -0.36 -0.33 0.39 0.29 -0.38 -0.49 -0.47 putative competitor S[K'Cl lagoin S|>fCl neol perogs pen amp Onyc Onyc 3 -0.78 -0.49 -0.04 t).22 -0.54 -0.17 0.15 0.63 0.70 2 -0.10 -0.27 -0.24 -0.47 -0.34 -0.32 0.10 0.26 0.29 1 -0.40 -0.11 0.01 -0.21 0.52 -0.26 -0.72 -0.50 -0.37 0 -0.27 0.10 -0.46 -0.17 0.34 0.53 -0.68 -0.16 -0.21 Table 5.5. Model 1 (2 indpendani variablc.s) path coefficienls Irom precipitation, intra-.specific abundance, inter specific ahundancc of possible coin|)eiitors, and predation to small mammal population change. Population cliaiige is from YR | to YR j, abundances are for YR , Critical value for R" is ca. 0.40. Data are for 1987-1998 (11 annual population changes) at ORPI. Path TO: PopulatiiMi Change of: Ncoioma Diiunlomys Onychomys Dipodoiiiys Chaeiodipus I'vroi'iuiilms all perog- FROM: lagonu)rphs sciurids albi^ula spcctahilis torridus mcrriumi pcnkilUitus umplus nathines precipilalion 0.27 0.31 0.07 -0.02 0.12 0.45 -0.22 -0.14 -0.32 predation -0.30 0.13 -0.08 -0.12 0.02 -0.43 -0.70 -0.66 -0.28 R- O.Ki 0.26 0.10 0.06 0.06 0..iS 0.66 0.57 0.31 intra-specific abundance -0.66 -1.21 -0.72 -0.9-1 -0.78 -0.63 -0.24 -0.10 -0.50 predation 0.21 0.91 0.15 0.49 0.48 0.04 -0.67 -0.65 -0.33 •> R- 0.29 0.72 0.46 0.54 o..iy 0..?6 0.65 0.53 0.4J "competitor": unk unk sciurids unk unk .sciurids Dipodomy.i nwn iami "c()m|)etiU)r" abundance -0.51 0.11 0.17 0.20 -0.28 predation 0.21 -0.48 -0.87 -0.83 -0.28 R- 0.15 (1.17 0.6.i 0.56 a 2.? "conipetitor" abundance -0.32 -0.13 -0.24 -0.18 -0.36 precipitation 0.02 0.43 -0.39 -0.29 -0.36 «- U. IH 0.2.1 0.27 0.21 0.34 lO OJ o Table 5.6. Model II palh coefficienls (rom prccipilalion, inlra- and inler-specific abundance, and predalion lo small manunal populalion change. Population change is from YR j lo YR i,precipialion, inlra-specific abundance, and predation are foi YR Critical value for R" is ca. 0.44. Data arc for 1987-1998 (11 annual population ciianges) at ORPI. Palh TO: Population Change of; NeoUmui Dipoddmys Onychomys Dipodoniys Chuctodipus Pcroiiiuitlius all perog- la};oniorphs sciurids alhifiiild spccuihilis lorridus iiicrriaiiii pcnicUUuux amplus nathines PROM; precipitation 0.40 0.16 0.08 0.10 ().(» 0.51 -0.22 -0.1 I 0.13 inlra-specific abundance -0.84 -1.07 -0.71 -0.96 -0.79 -0.63 -0.07 0.05 -0.56 predalion 0.28 0.88 0.14 0.46 0.50 0.(K) -0.65 -0.69 -0.31 K- 0.49 0.79 0.50 0.59 0.43 0.62 0.69 0.57 0.44 "COMPH'irrOR": unk unk sciurids unk unk sciurids /iipi idoniys iiwrritinii "compelilor" abundance -0.29 -0.07 0.15 0.21 -0.07 inlra-specific abundance -0.64 -0.65 -0.23 -0.12 -0.21 predalion 0.35 0.12 -0.76 -0.75 -0.64 R2 0.44 0.35 0.67 0.56 0.64 Table 5.7. Summary model (.see Fig. 5.3) path coelTicienls for seasonal precipitation, intra- and inter-specific competition, and predalion in relation lo rodenl population change. Population change is I'roni YR i to YR abundances arc for YR p; seasonal precipitation is for the adjacent .seasons: summer YR j and winter YR Data are for 1987-1998 at ORPI. NeoKmia Dipodoniy.V Cluu'todipu.s Penifiiialhu.s all perog albifiulei nieniami [H'liirillalu.s am pius nathines PATH TO: "population increase" FROM: summer precipitation -0.02 0.52 0.82 -0.08 0.02 O.(M) winter precipitation 0.44 0.21 0.27 0.80 0.82 0.72 intra-specific abundance -0.56 -0.40 -0.37 -0.73 -0.62 -0.79 "competitor" abundance -0.08 0.23 0.43 -0.38 -0.18 -0.39 •> K- 0.64 0.6« 0.76 0.S6 0.91 0.H2 putative "competitor" sciurids perognathines .sciurids D. iiwrriaiiii D. nieniami D. merriami PATH TO: "resistance to increase" FROM: intra-specific abundance 0.70 0.65 0.63 0.24 0.08 0.50 predation -0.16 -0.07 -0.05 0.67 0.67 ()..34 K- (U1 0.37 0.35 0.66 0.53 0.42 to 233 just endothermic predators as the index of predation pressure. However, for some taxa (the smaller heteromyids). the taxon-specific index performed better (Table 5.2). For the time-lagged predation index (Pred t., - for the yr prior to baseline and 2 yr prior to monitoring target for computation of Apop), the correlation was negative in 18 of 18 cases in Table 5.2. For this set of cases, the taxon-specific index again performed "better" than the generalized endotherm measure for the abundant and well-documented smaller heteromyids; however, it was not as strongly negative as for un-lagged predation values (Pred y-. Table 5.2). For several larger taxa. which were, however, less accurately monitored (lagomorphs, sciurids, D. spectabilis). correlations were weak for YR ^ and markedly stronger for YR This suggests that increases in these prey taxa may have produced large increases in the predators, but that those increases did not rapidly translate into prey population declines. For YR t-^,. correlations with predators in YR j were inconsistent through the dataset, with generally low absolute values (Table 5.2). Intra- showed a statistically significant or nearly-significant negative correlation with Apop in each small mammal taxon studied (Table 5.3). Intra-specific density dependence was apparently more consistent than inter-specific competition, predation. or even precipitation. Negative correlations indicating inter-specific competition among the various heteromyids that differ in size were not as prominent as I expected. However, correlations consistent with competitive effects were found among perognathines; and between lagomorphs and perognathines. possibly illustrating confounding effects of parallel causation. Given this, time-lagged correlations of Apop with intra- and with possible 234 candidates for inter- (Table 5.4) must be treated as only indicative, rather than as strong evidence, of potential interactions. Nonetheless, Table 5.4 shows that intra- generally gives the best correlation at zero time-lag. (Note that at a one year forward lag it would simply give the zero-lag correlation with opposite sign.) Time-lagged correlations involving potential competitors in Table 5.4 generally revealed few promising values, although time-lagged correlations between abundance of Dipoclomys merriami and Apop in perognathines are consistent with competitive effects of D. merriami on its smaller confamilials. Similarly, there is a large negative value at time-lag 1 suggesting an effect of the large Dipoclomys spectabilis on D. merriami. However. I view this as unlikely because the large kangaroo rats were not common at ORPI. Model I Path Analysis Path coefficients for intra- and inter- with ppt in the Model I (2 variable) path analysis (Table 5.5) were similar to bivariate correlation coefficients (Tables 5.1. 5.3). Pred also retained its bivariate characteristics when standardized with respect to ppt, although coefficients of determination were quite low for these models. However, path coefficients for pred tended to become less negative, and in some cases became positive, when computed with intra- as the co-predictor variable, for the larger taxa (lagomorphs. sciurids. Neotoma, D. spectabilis), as well as Onychomys and D. merriami. In other 2- predictor calculations (not shown), predation remained negative for YR t.,, but generally became positive for YR j in the larger taxa. Inter- also caused pred to change signs for Neotoma in the 2-prediclor path model 235 (Table 5.5), but not in D. merricitni or the perognathines. Precl retained strong negative coefficients in ail models involving the perognathines. Model I results had fairly high R' for most taxa when intra- was included, and in most calculations for the perognathines. Model II Path Analysis Results of the 3-variable path analysis (Table 5.6) retain the major elements seen in the foregoing model (Table 5.5). For perognathines, intra- and inter- became less important, but in the other taxa they became more important and pred assumed non- negative values. As in the bivariate case for intra-, and the Model I path analysis of intra- with precL critical values are, across the board, barely met, although Model II performs just slightly better in this regard. Model III Path Analysis This model was only performed for well-studied taxa for which competition hypotheses were readily available (Table 5.7). Ppt broken down by season had relatively strongly positive path coefficients for winter precipitation for all taxa, as expected from the literature, with summer ppr still indicated as strong for D. merriami. Summer for YR also had a large path coefficient in models for C. penicillatus, as my observations of rapid reproductive response suggested it should, but not for P. ampins. For Neotoma and D. merriami, neither competitors nor predators had scores suggesting they contribute to population regulation; ppt and intra- carried the models. All five factors scored relatively strongly in the perognathine models—with summer ppt included for C. 236 penicillatus. For perognathines. the results remain consistent with a regulatory effect of competition from D. merriami, as well as strong ones from pred (note that in Model III a positive sign for pred is consistent with regulation). I also performed simple standardized multiple regression computations for these taxa with the 5 predictor variables used in Model III. and then performed stepwise deletions of intra-, inter-, and ppt. Results were similar en toto to Model III. except that (1) summer ppt tended not to assume positive values. (2) inter- assumed a strong negative value for Neotoma in the 5-way model. (3) pred assumed a population-regulatory value for D. merriami in models with sciurids as competitor, and—in the only model to produce this— (4) pred failed to show the expected regulatory sign and magnitude in P. ampins and for the lumped perognathines in the 5-way regression. Although these calculations produced somewhat variable results overall, the latter point (4) is of interest here, since it is the only case in which predation failed as a predictor for perognathines. Stepwise variable deletions indicated that intra-, inter-, and winter ppt all shared some predictive power with pred for the perognathines (as a group and for the individual species studies). However, predation lost its negative sign only in the models described above, in which intra- was included. LIZARDS The lizard results were generally less varied and unstable with respect to predation than were the small mammal results, although for U. stansbiiriana all models were generally poor predictors of population behavior. 237 Bivariate Correlation Structure Negative correlations between population ciiange and precipitation are conspicuous in the bivariate study of the lizard dataset (Table 5.8). especially for winter ppt. In contrast, several taxa, which showed markedly similar population trajectories {Cnemitlophorus, Uroscuirm, Ccillisaurus, Sceloponis, Dipsosaurus: see Fig. 5.4), showed positive correlations of Apop with summer ppt. This was especially prominent for YR J, for which summer coefficients were generally high. The side-blotched lizard Uta stansburitma, which was abundant and well-monitored, showed a distinctly different pattern—negative relationship to ppt in summer, and positive for winter (Table 5.8), which might be expected for this winter-active (pers. obs.; C. Conner, pers. comm.) species. As in the small mammal ta.xa. despite the evident reproductive response to good seasonal rainfall, few of the correlation coefficients forp^r reach statistical significance at P = 0.05. For YR T. pfecl showed a negative correlation with Apop in 6 of 9 cases (Table 5.9). For the western banded gecko Coleony.x variegatus and the crotaphytines {Gcimbelia wislizenii, Crotciphytus nebriiis), negative correlations with pred occurred only at forward and backward time lags, although the monitoring results for these taxa are not very precise. For Uta the only, weak, negative value for pred was for time lag - I (YR the target year for population change). The predation-pressure index computed specifically for lizards performed better (i.e., more negatively) than the index based on all endothermic predators, and the index for Uta performed better than the other indices. Intra- showed a consistent negative correlation with Apop in each lizard taxon Table 5.8. Correlation coefficients between li/ard population change and precipitation. Data are lor 1988-1998 (10 annual population change.s) at ORPI. Critical value (P = 0.05) for the hypothesis that r > 0 is r = 0.55. For the hypothesis that correlation for summer is higher than that for winter, the critical value for rs - r^ is ca. 0.85. Highest values are indicated in boldface. Ciu'inuioplionis Coleonyx Uia Unmiiinis Cullisaurus Sccloimnts Dipsosaurus crolaphytines Phiynosonui previous yr: winter rain -0.21 0.15 -0.12 -0.12 -0.26 -0..^1 -0.59 -0.27 -0.42 summer rain -0.48 -0.(B -0.24 -0,.15 -0.09 -0.19 -0.17 0.62 0.18 base yr: winter rain -0.44 -0..18 0.49 -0.57 -0.51 -0.85 -0.40 -0.27 -0.49 .summer rain 0.69 -0.19 -0.40 0.57 0.27 0.48 0.05 0.04 0.20 larffi't yr: winter rain -0.29 -0.09 -0.08 -0.03 -0.12 -0.14 0.11 0.17 -0.09 .summer rain -0.1.3 0.22 0.31 -0.41 0.22 0.12 0.26 0.16 -0.18 mean of yrs: winterrain -0.31 -0.11 0.10 -0.24 -0.30 -0.44 -0.29 -0,13 -0.33 sununerrain 0.03 O.W) -0.11 -0.07 0.13 0.13 0.05 0.27 0.07 lo LU 00 Table 5.9. Correlalion coelficients between li/ard population change and predaiion indices (dellned in text). Data are (or 1988-1998 (10 annual population changes) at ORPl. Critical value (P = 0.05) for the hypothesis that r < 0 is r = -0.55. Largest negative values are indicated in hnldfuce. Cnemidophorus Cofeonw Uui Urosaurus Callisdiirux Sceloporiis Dipsosaunis crotaphytines Pliiyno.wnui previous yr: t:siw 0.09 0.26 0.05 0.47 0.01 0.05 0.20 -0.37 0.18 -0.(B -0.20 0.15 0.25 -0.38 -0.23 0.03 -0.54 0.15 Prcd,y„ -0.47 0.23 base yr: Prcd^/j /-Tv/jo -0.17 0.11 0.38 -0.08 -0.06 -0.40 -0.40 -0.12 -0.39 - mean ofyrs: Hsoo -0.13 0.04 0.29 0.06 -0.08 -0.31 -0.13 -0.21 -0.26 P'"Cd/./XHK/j -0.24 -0.19 0.19 -0.13 -0.30 -0.48 -0.26 -0.20 -0.32 Prcd,yy>, -0.23 0.06 to \0 Table 5.10. Correlation coelTicient.s between lizard population change and lizard abundance.s. Data are for 1988-1998 (10 annual population changes) ul ORPI. Critical value (P = 0.05) lor the hypolhe.si.s that r < 0 i.s r = -0.55. Inlra-.spccillc conelation.s are indicated in boldface; if any other value for a taxon is the most negative, it is indicated by italic bold. Population Change of: Cin'iiiiili>i>lntni\ CoU'onyx Uui Uroxaitrus Callisaiiriis Sicliifxirus Dipsosaurus crolaphytines I'lirynosoiiui AbuinlaiKc uf: Cnemidophoms -0.6W 0.43 -0.35 -O.I« -0.05 -0.29 0.01 0.32 -0.28 Coleony.x 0.06 -0.44 0.02 0.12 0.06 -0.03 0.71 -0.11 0.11 Uui 0.18 0.69 -0.24 0.58 0.53 0.38 0.28 0.14 0.04 Urosaunts -0.59 0.11 -0.27 -0.46 -0.28 -0.27 0.22 0.09 -0.03 Callisaurus -0.43 -0.17 -0.37 -0.24 -0.61 -0.29 0.01 -0.12 0.27 Sci'loporus -0.48 -0.14 -0.14 -0.39 -0.25 -0.48 -0.06 0.17 -0.35 Dipsosaurus -0.31 0.05 -0.56 -0.39 -0.06 -0.15 -0.54 0.52 -0.23 crotaphytines -0.50 -0.29 0.18 -0.37 -0.85 -0.69 -0.10 -0.67 -0.20 Phrynosoma -0.30 -0.13 -0.05 -0.58 -0.17 -0.32 -0.24 0.11 -0.53 Table 5. II. Tinie-laggeii correlalions between lizard populalion changes and abundances. Daia are for 1988-1998 (10 annual population changes) al ORPI. Populalion Change of: Ciu-midophoni.t C(tU' FROM; CnemUlophoms Coleon v.v Utn Urosaurus Callisaurus Sccloporus liipsosaurus crolaphytines I'luynusoma precipitation 0.01 -0.42 0.11 -0.11 -0.28 -0.31 -0.11 -0.17 -0.12 predation -0.37 0.14 O.OI -0.30 -0.08 -0.48 -0.65 0.14 -0.59 R' (U-) (U6 O.l.i O.M 0.17 0.61 0.46 0.06 0.42 intra-specific abundance -0.51 -0.56 -0.23 -0.62 -0.69 -0.36 -0.51 -0.77 -0.44 predation -0.35 -0.26 0.06 -0.60 -0.38 -0.59 -0.65 0.32 -0.56 0.47 0.24 0.06 0..U 0.5! 0.57 0.70 0.54 0.5ti "competitor"; CallisuHrus unk Ciumi Sceli)/)orus Ciu-ini Cm-mi unk unk unk "competitor" abundance -0.55 -0.54 -0.29 0.02 -0.10 predation -0.59 0.39 -0.37 -0.23 -0.64 R' 0.52 0.24 0.27 0.05 0.46 "competitor" abundance -0.42 -0.38 -0.01 0.11 -0.06 precipitation 0.05 0.02 -0.32 -0.41 -0.56 K' 0.19 0.40 0.29 0.70 to to Table 5.13. Model II path coefficients from precipitation, intra- and inter-specific abundance, and predation to lizard population change. Population change is from YR j to YR precipiation, intra-specific abundance, and predation are for YR |. Critical value for R" is 0.50. Data are for 1988-1998 (10 annual population change.s) at ORI'l. Path TO: Population Change of: Cnemidophorus Cok'ony.x Uui Urosaurus Ctillisaiiru.s Scclopoms Dipsdsaums crotaphytines I'hrvnosoma FROM: precipitation 0.18 -0.28 0.03 t).08 0.05 -0.10 0.11 0.10 -0.02 inlra-speciric abundance -0.50 -0.46 -0.23 -0.59 -0.69 -0.30 -0.54 -0.82 -0.43 predation -0.30 -0.12 -0.04 -0.54 -0.33 -0.48 -0.68 0.32 -0.53 H2 0.56 0.31 0.19 0.59 0.54 0.67 0.71 0.55 0.59 "COMPinrrOR"; CalUsaurux unk Ciii'ini Siclopdrus Ciiciiii Ciwiiii unk unk unk "competitor" abundance -0.40 -0.53 O..V^ 0.57 0.34 intra-specific abundance -0.27 -0.03 -0.84 -0.99 -0.59 predation -0.47 0.37 -0.73 -0.59 -0.63 K2 0.56 0.24 0.58 0.71 0.61 -p.to OJ Table 5.14. Summary model (see Fig. 5.3) path coefficients lor seasonal precipitation, intra- and inter-specific competition, and predation in relation to lizard species population change. F\)pulation change is from YR | to YR |>|; abundances are for YR x; seasonal precipialion is for the adjacent seasons, summer YR j and winter YR |+ |. Data are for 198X-I998 at ORPI. Ula Callisattrus CneiiiUloplutrus slansbunana Urosaiims clraconoides Sceloporus PATH TO: "population increase" FROM: summer precipitation 0.71 -0.52 0.57 0.46 0.46 winter precipitation 0.07 -0.04 0.30 -0.19 0.24 intra-specific abundance -0.16 -0.41 -0.08 -0.64 -0.47 "competitor" abundance -0.45 -0.36 -0.58 0.41 -0.22 •> R- 0.76 0.44 a 59 0.60 0.50 putative "competitors" Callisaurux Cnemidophoms Sccloporus Cncmidophorus Cncmidophorus Sicloporus Urosaurm Sccloporus Urosaurus Sceloporus Catliscmrus crotaphytines Phrynosoma PATH TO: "resistance to increase" FROM: intra-specific abundance 0.50 0.23 0.60 0.68 0..34 predation -0.06 0.59 0.36 0.59 R- 0.45 0.06 0.5.i 0.50 0.56 245 studied (Tables 5.10. 5.11), although only 3 of 9 individual cases were at or above the critical value for P = 0.05. These results, as for the small mammals, indicate widespread density dependence in the lizards. However, there were several other strongly negative correlations between species' Apop's and the abundance of other lizard species—especially involving taxa that might function as both predators and competitors (e.g., Cnemidophoriis with Uta and Urosaurus: Sceloponis with Urosciiirus: crotaphytines with Cnemidophoriis and Callisaiinis). Thus, the coefficients in Tables 5.10 and 5.11 do not strongly suggest inter-specific competition. Model I Path Analysis In lizards, the 2-predictor path analysis did not markedly alter the values of the corresponding bivariate correlation coefficients (Table 5.12). Coefficients of determination for YR (not shown) were generally low. as were several of the YR j models that included ppi (only 3 of 14 were at critical value) or inter- (3 of 8). Models with intra- and pred exceeded critical value for 7 of 9 taxa, and this combination also improved the fit of pred to regulatory (i.e., more negative) values for Coieonyx and Uta. Model n Path Analysis For the 3-predictor variable models, the general picture indicated by the 2- predictor (Model I) results was maintained (Table 5.13). R' values for YR ^ rose 0 - 13% compared to the best Model I results in Table 5.12. In the models employing ppt. intra-, and pred. all pred path coefficients (except for the crotaphytines) assumed negative 246 values and ppt coefficients tended to switch signs to assume more intuitively appropriate positive values. However, employing intra- and pred with inter- tended make the response of inter- more erratic, often more positive. Model III Path Analysis The summary model (Table 5.14) for the five lizard ta.xa with the best data (which have plausible hypotheses for interspecific competition) yielded path coefficients indicating strong positive effects for summer ppt, and weakly negative or positive path coefficients for winter ppt. except in Uta. In these models, as in the small mammals. intra- always assumed values consistent with density dependent population regulation. Inter- assumed negative values e.xcept in the case of CalUsauriis. although both sign and value were strongly dependent upon the taxa included as putative competitors: including Uta as a competitor for any other taxon generally resulted in a positive path coefficient for inter-, for example. I have displayed the models with the most negative values for inter- simply because they may hold some potential interest. At least one ppt measure had a substantial positive value in each case except Uta. Except for the Uta model. R' values indicate that the models explain 45 - 76% of the variance in population change. For Uta. Model III was an improvement over other models, but still performed poorly: the improvement in prediction power was apparently associated with the addition of the "competitors", although adding only Cnemidopliorus to the model produced similar results. The simple standardized multiple regression computations for these lizard taxa 247 (using 5, 4, or 3 predictor variables from Model III) sustained characteristics very similar to Model III, except that (1) intra- became weakly positive in the 5-way regression for Sceloporiis. (2) winter ppt became negative in the 3-way models, and (3) in the CaUisaiinis regressions, winter and summer ppt tended to assume more evenly-matched, positive values than in Model III. The 5-way regressions produced R' ranging from 0.78 - 0.92 (similar to those for rodents), except for the Uta model, which remained poor. DISCUSSION A summary condensation of the analytical results for major, closely-monitored prey groups is shown in Fig. 5.6. Correlative data can be misleading if variables do not have a cause-effect relationship, but rather share parallel causation. The critical caveat is, "correlation does not imply causation". In this study, precipitation produced an array of parallel causes in the ecosystem: plants produced food for animals, animals' constraints were directly reduced by water availability, and prey populations of insects and small vertebrates increased. The caveat is most critical when the array of alternative causal pathways cannot be specified and evaluated, which is not entirely true in this study. I examined relationships among the parameters most likely to be relevant to system behavior: species abundance, predator abundance, competitor abundance, and food resource productivity. Inferring probable causality in this circumstance is logically sound. Although not all possible causes were examined in this study, the most probable ones were. Even though the proposition that,"predator .v regulates prey species y", is not a complex or original hypothesis, it is an important one, and it exemplifies the hypothesis winter PPT winler PPT sumnner PPT summer PPT energetics energetics sciurids? 0. mernami A Neotoma A perog. Neotoma perognalhines mortality mortality predators predators winter PPT winter PPT summer PPT summer PPT energetics energetics perognathines competitors A mernami D. mernami conspecifics mortality mortality predators predators Fig. 5.6. Path diugramii summuri/ing conclusiuns I'or four laxa wall the htrongcht monitoring dalu, hu.scd on hivariatc corrcluiions, path analysis, observations, and literature reports. Dashed arrows reprcseni negaiivc ciTects, solid arrows posiiive cffccis; lo the width of the arrows corresponds with the probable strength of ihe effects. 4i. 00 249 testing that underlies this paper. The primary statistical problem I had was to tease apart the widespread colinearity produced by the generalized upwelling of desert life associated with the powerful El Nino events of the mid-1990's. During the late I980's, essentially all small vertebrates, except some snakes, responded by decreasing to minimal abundance at the 1989-90 drought maximum, and then began a dramatic increase with renewed precipitation. By 1992, all but two species {Coleonyx variegatus. Dipodomys spectabilis, both with limited monitoring data) had posted substantial increases, often dramatic ones eventually involving a 3-fold or greater increase. But despite high productivity through early 1995, marked population declines began in some well-studied ta.xa as early as 1992, notably in lizards and perognathine rodents. During 1993-5, declines began in the larger rodent species and lagomorphs. Most of the best-documented population trajectories saw population collapses at least as dramatic as the earlier explosions (Figs. 5.4. 5.5). Predation pressure was also at a low point in 1990, and steadily increased to 3 or 4 times the low value by 1994-1995; it then declined to a secondary low in 1997. Prey populations responded more quickly than predators, as is often the case. Such eruptive increases may be expected among early-maturing or high-fecundity prey species where, as with variable desert rainfall, carrying capacity of the environment increases abruptly. The reproductive biology of some prey species enables them outpace behavioral and populational responses of predators. The equal rapidity of the later collapses, as rainfall moderated, is too abrupt to result from only diminution of reproductive rates under intra- and inter-specific competition ~ not if we assume only attrition from invariant mortality 250 levels. Rather, the collapses indicate the operation of intensified mortality processes. This is a key point supporting interpretation of the path analysis results. Alternative interpretations-incorporating ecological interactions other than those I considered—do not now seem probable. I have no evidence that disease, parasitism, starvation, or aggression among competitors were widespread effectors of prey mortality. (1)1 have not observed weak or emaciated individuals of these taxa, except during fall 1989, when hatchling western whiptails {Cnemiduphorus tigris) were weak and emaciated, apparently dying outright from lack of water or food. (2) At the end of the 1997 drought I found adult western whiptails that were thin and mildly dehydrated in appearance; none evidenced a condition from which captives fail to recover without delay; and they all looked fine within days of drought-breaking rain. (3) In late June of the 1994 foresummer drought. I found two subadult N. albigiila and 2 doves that had apparently succumbed to exceptionally hot conditions (the Neotoma were in shallow retreats that I supposed were the best available to them under some kind of intra-specific aggression). (4) Animals captured in traps appeared robust and healthy (except those injured by other animals in the traps). (5) I explored the desert searching for and counting small animals during hundreds of diurnal and nocturnal forays, observing predatory strikes by various taxa and aggressive inter-specific chases among potential competitors: yet I did not see injured, weak, or dying individuals that might have suggested non- predation-related mortality. Nor are such indications found in the literature. The evidence against population regulation by mortality from inter-specific aggression or epidemic disease—without competition or predation as primary modulating 251 co-factors—appears to be substantial for this ecosystem. Disease and parasitism may well be important, but if they are major contributors to regulation, they apparently must affect population dynamics by influencing the outcomes of predatory and competitive interactions, rather than by causing outright mortality. The ecosystem appears to be driven by responses to rainfall-induced variations in carrying capacity, modulated by predation and competition. CORRELATIONS: EFFECTS OF PRECIPITATION It has recently been suggested that Southwestern desert rodent population changes are not significantly correlated with precipitation (Swanson, 1998), despite the widespread foregoing appreciation of such a causal association (Beatley, 1969; Whitford, 1976: Petryszyn, 1982; J.H. Brown and Zeng, 1989; and elsewhere). My results show why such a conclusion might be reached—there are multiple contributing causes for population change, which can obfuscate trends subjected to simple correlation analysis. The ORPI data reject the notion of no significant climatic correlation. In the summary model for small mammals, the mean path coefficient for ppt was +0.38 (95%CI 0.14 - 0.62), whereas for winter ppt it was +0.61 (95%CI 0.40 - 0.82). The ORPI data support the established idea that winter rains play a significant role in Southwestern rodent population dynamics, and suggest that summer ppt may also play a role for some taxa. Lizards display a different pattern. Although the means of taxon path coefficients for ppt in the summary model (Table 5.14) were not significantly different from R = Q, there is a clear pattern in Table 5.8 of positive association of the winter-active Uta with 252 winter ppt (bivariate r = 0.49) and the summer or summer-spring breeding taxa with summer ppt {r = 0.26, 95%CI 0.05 - 0.47). Excluding Uta, the difference between winter and summer ppt in correlation to Apop was 0.75 + 0.148 SE, and even with Uta the statistical significance remains. Table 5.1 also shows some positive correlations for summer ppt in YR in small mammals, specifically in lagomorphs and perognathines. This seems to be consistent with field observations of the rapid appearance of young after the on.set of good summer rainy seasons. In Neototna, the strongest correlations of precipitation and population change were for annual rainfall, as previously described by Petryszyn (1982): for each of the two bi-seasonal periods centered on YR p. r was equal to +0.52. For most small mammals, reproduction can occur in response to both Sonoran Desert rainy seasons. Among lizards. Urosciiirus oniatiis and Heloclenna siispectiim are summer egg-layers in this region (Asplund and Lowe. 1964; Goldberg and Lowe. 1997; Holm. 1988; Van Loben Sels and Vitt, 1984; but see Parker. 1973). and Phrynosoma solare and Gcimbelia wislizenii may also be (Howard. 1974; Parker, 1971; Parker and Pianka, 1976; Tinkle and Hadley, 1975), while Uta stcmsburiana is primarily a late winter-early spring breeder (Asplund and Lowe, 1964; Parker and Pianka, 1975; pers. obs.). However, most lizard species at ORPI produce eggs in both spring and summer (Goldberg, 1987; Parker, 1972; Parker and Pianka, 1973; Pianka and Parker, 1972; Smith et al., 1987; Tinkle and Dunham. 1986; Vitt and Ohmart, 1974, 1977a&b), and should be able to show rapid population increases in response to rainfall in either of the two Sonoran Desert rainy seasons. The absence of such consistent responses implies that other biotic interactions are important. 253 SMALL MAMMAL POPULATION REGULATION All of the rodents and lagomorphs I studied participated in the niid-1990's climate-associated oscillation, but with striking taxonomic variations within the pattern. Among the numerically dominant rodents, smaller species (perognathines) increased fastest and declined soonest; N. albigula increased most slowly and declined last; D. merriami was intermediate. This may be a regular feature of the system: Petryszyn (1982, and pers. comm.) has observed the same sequence in previous monitoring near Tucson. The rate at which rodents increased may be related to rapid maturation and high fecundity (see Chapter 1). Competition or predation (or both) subsequently became involved in the rapid down-regulation of these populations after their maxima. As also in lizards, the most consistent correlate of small mammal population change was intra-specitlc density (Tables 5.3. 5.10). This correlation must ultimately occur because populations do not typically march to zero or infinity, but instead fluctuate (however irregularly) about some central range of values. Yet this does not ensure that such negative correlations will appear so uniformly or rapidly. It is notable that (1) over only 10-11 yr all taxa showed this correlation quite strongly, and (2) the correlation was sustained in path coefficients throughout the multivariate analysis. Whatever statistical anomalies may exist in the data structure, they did little to remove the intra- effect. Only pred. especially in conjunction with ppt, statistically preempted the predictive role of intra- in any variable combination in Models I-III, and this occurred only rarely, in the perognathine rodent analysis. 254 Perognathines The apparently strong correlational evidence that predation affects population regulation in perognathine rodents (Tables 5.2. 5.5-5.7) could be tempered by consideration of variation in quality of the predation data. An important part of the very high coefficients for both C. penicillatus and P. ampins is a decline and rebound of both their populations (Fig. 5.4) synchronously with the pattern of the predation-pressure index, during 1994 (Fig. 5.4). Predator data for 1994 are quite scarce, and it happened that the endothermic predators observed were not ones modelled as important perognathine eaters. Nonetheless, there are other specifics that do support a hypothesis of predation-regulated populations in perognathines. Perognathine rodents seem vulnerable to predation: they are slower runners and have less auditory acuity than Dipodomys. and they do not occupy fortified nests like Neotoma. They occur disproportionately in the diets of all mammal-eating snakes with the possible exception of the gopher snake Pituophis melanoleucus (a constrictor that frequendy eats Neotoma and young lagomorphs). and are especially prominent in the diet of neonatal and yearling western diamondback and Mohave rattlesnakes. These young rattlesnakes are rapidly growing, frequently feeding, and were sometimes very numerous, and may play a greater role in pocket mouse ecology than I attributed to them in the predation index. Reproduction by these two large, abundant rattlesnakes had almost ceased by 1990. Despite good conditions, the slow pace of rattlesnake reproductive cycles (e.g.. Tinkle, 1962; Goldberg and Rosen, submitted-, although, see Fitch and Pisani, 1993) 255 allowed only moderate reproduction in 1991. In 1992 and 1993, most (ca. 80%) adult females gave birth. By late 1993. neonates and yearlings were remarkably abundant on the desert floor, along with a more moderate number of 2 yr-olds bom in 1991. By 1994- 5. monitored rattlesnake abundance had increased substantially (Chapter 4). The only other year large numbers of juvenile rattlesnakes were found was 1988. al.so coincident with a perognathine decline. In addition to snakes, small and medium-sized endotherms like the western screech-owl and bam owl prey on perognathines. and these endotherms appeared to increase more rapidly than larger predators during the early 1990's (Chapter 3). This may also have strongly and rapidly intensified predation on perognathines to an extent not fully represented by the predation pressure index, since these smaller predators, especially the noctumal ones, are harder to see than larger predators like hawks and coyotes. Considering these details, the path analysis results for predation on perognathines are not likely to be methodological artifacts. Competition with D. merricimi is also suggested by the path coefficients and bivariate correlations for perognathines (see Tables 5.3. 5.7). although these coefficients performed variably when predation was added as a co-predictor (Tables 5.5. 5.6). The remarkable abundance of D. merriami in 1994 (Fig. 5.4) must have accentuated the collapse of perognathine populations, quite possibly by increasingly restricting the smaller heteromyids to shrub-base microhabitat (Rosenzweig and Winakur. 1969; Rosenzweig, 1973; Price, 1978a) where they were intensely exposed to the risk of predation by snakes (Kotleret al., 1992, I993a&b; Bouskila, 1995). Competitive effects 256 of Dipodomys on perognathines involving interference (Frye, 1981; Heske et al.. 1994) have been experimentally demonstrated in the Chihuahuan Desert Grassland of southeastern Arizona. But, paradoxically in southeastern Arizona, the numerical response of perognathine populations to experimental exclusion of Dipodomys, though significant, was slight, and perognathines did not approach the abundance they normally achieve in the Sonoran Desert (J.H. Brown and Heske, 1990: Heske et al., 1994; J.H. Brown and Zeng, 1989; see e.g., Petryszyn, 1982, and Chapter 1 herein). Higher or less fluctuant predation pressure in the wetter environment of the southeastern Arizona desert grassland (Millsap, 1981), which may affect perognathines more than the larger heteromyids. could explain this experimental anomaly as well as the differing heteromyid assemblage structure between that site and the Sonoran Desert sites in southern Arizona. Dipodomys merriami Although in this species the highest bivariate correlations (Tables 5.1-5.3) were for imrci- (/• = -0.60), summer ppt (4-0.54). and pred (-0.41). the coefficient for pred was minimized or negated in models with intra- as a co-predictor (Tables 5.5-5.7). Similarly, the largest bivariate relationship for inter- (-0.26, with the sciurids) was reversed with pred as a covariate and minimized with ppt (Table 5.5), and was also was reversed in Model III, regardless of the taxon used as the putative competitor (Table 5.7). I considered this a possible keystone species in its assemblage, due to its size and abundance. I therefore analyzed a large number of variations on the path models, searching for a suitable model; none were a clear improvement over a simple one using 257 only intra- and ppi. The bivariate correlations suggest inter-specific competition and predation effects on D. merriami. However, the details don't seem to confirm a competitive effect of perognathines on D. merriami (Fig. 5.4): D. merriami increased steadily from 1992-1994 over a range of perognathine densities, and collapsed during 1994-5 when perognathine densities were low. Sciurids were not very accurately monitored, so it is not useful to contrast the details of their trajectory with D. merriami, beyond the already-noted coefficients. During the field work. I observed remarkably low densities of D. merriami in summer 1990. as shown in Fig. 5.4, and also during mid-late summer of 1997, a fact lost through the averaging process and in the timing of monitoring. It appeared to me. at these times, that as predator populations approaches their minima from previously high levels, certain prey populations decreased even more rapidly, and were driven to e.xtremely low levels by the remaining predators, presumably due to a scarcity of alternate prey. Demonstrating such reverse density dependance will require more refined information. Other Small Mammals The path analysis does not offer strong support for predation or competition as regulators of N. albigiila populations. Pred fared poorly as a co-predictor with any of the other variables. Although interspecific competition with D. spectabilis (or with sciurids) did slightly better, neither held up well when both ppt and intra- were used with it as co- predictors. Further, D. spectabilis never appeared abundant enough to be an important 258 competitor, and it and the sciurids differ from Neutumci in diel activity, diet, and/or microhabitat. As in D. merriami, the massive collapse in this species (a year later than the D. merriami decline) suggests a potentially strong role for predation. Anecdotal information also points toward predation in the mid-late 1990's declines of N. albigula. sciurids, and lagomorphs. Bobcats (R. Henry, pers. comm.; see also Chapter 3) and domestic cats (Felis sylvestri^-, pers. obs.) appear to be efficient N. albigula predators. Although I only saw bobcats in the Valley of the .Ajo in 1988, 1996, and 1998. in the desert between Tucson and ORPI 1 saw several during 1995-1996, suggesting they become more abundant late in the mid-1990's. Such delayed predation phenomena, which may be hard to document, may explain the late collapse in N. albigula. which occurred after it had sustained population growth for three years under apparently high predator pressure. Similarly, badgers, which are noted sciurid predators (see Chapter 3), were generally uncommon in the Sonoran Desert during this study. However, they appeared on the OSSA area in 1995, establishing residency, and by 1996 entire areas hectares in extent had nearly ubiquitous damage to their rodent burrow systems. During the lagomorph decline in 1994-1996,1 occasionally found collections of lagomorph intestinal remains surrounded by coyote tracks, again providing conspicuous, though anecdotal, evidence of intensified predation. Although for these taxa, as for D. merriami, the existing data do not consistently indicate a population dynamic role for predation, I suspect that more detailed monitoring data would. 259 POPULATION REGULATION IN THE LIZARDS The predation pressure index for lizards was correlated positively with winter ppt (r =0.54), and negatively with summer ppt (r = -0.30). Endothermic predator populations had a strong positive response to winter ppt (r = +0.50), and this beneficial effect on predators apparently outweighed the increased lizard reproduction associated with winter ppt. The inadequacy of the path models for Uta stcuisbiiriana is surprising, since it was clear early in the monitoring that this taxon was not responding as strongly as expected to winter rainfall. It experienced a broadly consistent increase that is perhaps not surprising considering the outstanding winter rains (Fig. 5.1) and amelioration of winter low temperatures during this time (P. Rowlands, pers. comm.). Previous intensive studies in the Mohave Desert of Nevada had suggested that this species was probably regulated in part by predation by one of the crotaphytines, Gambelia wislizenii (Turner et al.. 1982), but in the present study the available data for crotaphytines indicated no negative correlation with Apop in Uta. I attempted to use a uniform set of predictor variables arrived at a priori. I suspect that the Uta trajectory can be successfully modelled as a positive interaction between winter climate and one or more negative causal influences from abundant lizards that compete with and eat it. such as Cnemidophorus tigris. Several lizard taxa had sizable negative correlations between Apop and the abundance of other lizard species, particularly between the relatively large (and saurivorous) crotaphytines and medium-sized taxa the crotaphytines might affect via predation as well as competition. Similarly, predation could contribute to a negative 260 effect on Uta from the abundant medium-sized lizards that can catch them—Callimurus and Cnemidophonis. During this study, predation was observed by Gambelia on an adult Cnemiclophorus tigris, and by an adult Cnemidophonis tigris on an adult Uta. Similarly, predation by Sceloporus magister and 5. clarkii on syntopic Urosaurus oniatiis is at least a threat, although it has not been observed in the wild. Finally, primarily insectivorous passerine birds such as the curve-billed thrasher, which increased from 1990 to 1991 (ORPI, unpublished data), may also affect the population dynamics of lizards via predation (see Holm, 1988; and pers. obs.). Inter-specific competition (inter-) in lizards cannot be strongly inferred from this analysis, although it is suggested by the bivariate correlations (Tables 5.10. 5.11) and the summary path analysis model (Table 5.14). In attempting to locate variable combinations for predation that might reject the indications of its importance, coefficients for inter- sometimes changed erratically, suggesting they involved colinearities with other causal variables. This correlation study appears to be insufficient to evaluate inter-specific competition in lizards in a meaningful way. If competition (either inter- or intra-specific) was a primary determinant of lizard population decreases after 1992,1 would expect decreased reproductive output associated with declining populations during the mid-1990's. The proportion of juveniles instead tended to be higher in years which led to population declines (Apop = 0.47 - 0.03 x (%juveniles). r = -0.59,0.10 < P < 0.05; see Fig. 5.7). High reproductive output and good body condition of lizards were generally associated with good rainfall, while population growth was sometimes not (Fig. 5.7). Experimental studies indicate that strong Cnemidophorus, OR PI •HI adults EZD subadults juveniles PRECIP (MM) Q ^2.0 100 o 50 CO o> O) o o 5^ CM CM CO CO in in (O (D h- CO CO CO 00 o> o> O) a> a> o> o> o> o> a> O) O) OJ O) o> a> c» o> Q. 3 Q. 3 Q. 3 Q. 3 Q. 3 o. 3 a. 3 Q. 3 o. 3 a 3 0) 0) Ui (A 0> (A (0 if) Ui tn (0 (0 (0 (/> (/) u> (A Season and Year Fig. 5.7. Seasonal time-series of precipitation, and population structure and abundance of whiptail lizards. lo 262 competition between North American Desert lizard species may be unlikely during wet years (Dunham, 1980, 1981; Smith, 1981). During the 1989-1990 and 1996-1997 droughts, there were large numbers of large, old adult whiptail lizards (unpublished data), which appeared to be associated with low snake activity during these dry periods. The largest individuals disappeared rapidly in following rainy periods with high snake activity. Recapture data showed that mortality rates for C. tigrLs. and apparently also S. incigister, at ORPI were very high—few individuals reached age 3 yr (unpublished data). Finally, the decline of lizard populations during the mid-1990's El Nifio rainfall events was most apparent in the rich habitat at Alamo Canyon at ORPI, where endothermic predators were also most apparent. Sharply reduced lizard densities in similarly productive desert canyons appeared to be widespread in southern Arizona during 1993- 1995 (pers. obs.). The monitoring data and ancillary observations suggest that predation is of great importance to lizard populations, and that during 1987-1998 precipitation and predation combined to regulate lizard populations in the northern Sonoran Desert of southern Arizona. CONCLUSION The monitoring program for prey taxa would benefit from formalization of predator monitoring protocols, especially if ta,\onomic detail and breadth can be incorporated. In tandem with this, detailed accounting of actual abundances and prey consumption rates should prove highly informative. Experimental study might simultaneously proceed using fenced (Bock and Bock, 1994) or filtered (J.H. Brown and 263 Munger. 1985; Heske et al., 1994) enclosures. The scale of such enclosures should be larger than usually envisioned, to allow adequate numbers of each dominant rodent taxon are to be enclosed, and to preserve some semblance of their natural spatial ecology. Smaller enclosures would probably suffice for the more numerous lizards. While predator exclusion methods may definitively establish a population-limiting role for predators, they will not necessarily reveal how the real natural fluctuations of prey populations (observed by monitoring) have been caused. 264 DISSERTATION SUMMARY This dissertation summarizes and synthesizes population monitoring results for vertebrates in a desert valley community in the Sonoran Desert, at Organ Pipe Cactus National Monument (ORPl). Arizona. 1987-1998. Monitoring of small mammals, lizards, endothermic predators, and snakes was carried out by National Park Service personnel, in conformity with Ecological Monitoring Program (EMP) protocols, and by me using additional methods and during development of the EMP protocols. The synthesis objectives are to (1) provide material for review and validation of the protocols, (2) evaluate climatic and biotic correlations, and (3) identify additional protocols for inclusion in the EMP. Nocturnal rodents were monitored at two live-trapping grids at each of the five Core I EMP sites in desert valley habitat, using mark-recapture procedures. All rodents and lagomorphs were also monitored in an intensive study region using road transect, time-constrained search, and drift-fence trapping methods. The EMP live-trapping protocol was evaluated using modified Lincoln-Petersen Index calculations. It performed well except for a 12% underestimation abundance. It had the important advantage of being able to yield monitoring results for all nocturnal species, rather than only abundant ones. I recommend its continued use in the EMP. with re-calibration to actual abundance. All small mammals declined to minimum abundance during a 1989-1990 drought, and then increased with strong rains, especially winter rains, that occurred 1990-1995. 265 Individual species had divergent sub-patterns within this context. Small species (perognathine heteromyids Chaetodipus penicillatus, and Perognathiis ampins) increased most rapidly but then declined markedly after 1992. The medium-sized heteromyid, Merriam's kangaroo rat Dipodomys merriami increased more slowly to a peak in 1994 that was also followed by a rapid collapse. The largest species, the packrat Neotoma albigula. increased most slowly, reaching a 1995 peak, and also had a subsequent sharp decline. The rapidity with which these taxa reached maximal abundance corresponded to their apparent capacities for rapid reproductive increase. Their declines occurred during a time of increasing predator pressure, which apparently intensified first for the smaller taxa. 1 argue that population dynamics and community structure may both be regulated, in part, by predation. Diurnal lizards were monitored on 21 permanently-established EMP line transects at 13 sites, and also by drift-fence trapping and time-constrained search in the intensive study region. The transects yielded results consistent with limited mark-recapture efforts, as well as with other methods' results. Like small mammals, lizards reached population minima in the 1989-1990 drought and increased rapidly following drought-breaking rains that began July 5, 1990. However, all of the abundant, closely monitored taxa (Cnemidophorus, Sceloporiis, Urosaiiriis, Callisaiirus draconoides) except one {Uta stansbiiriana) began declines to low levels after 1992, even though strong rains continued through 1995. These declines, and sustained low abundances, occurred under elevated predator pressure. Lizard populations made dramatic expansions during the dry years after 1995, when predator abundances declined. I suggest that anti-predator behavior and 266 capability may be important determinants of community structure in these lizards. Endothermic predators (canids, felids, mustelids, hawks, owls, shrike, and roadrunner) were recorded whenever they were seen, and a monitoring dataset was constructed from field notes for full workdays consisting of 8-12 field-hr of routine herpetological work. These predators were observed at an average rate of 1.002 per day during 651 included days. 1987-1998. Endothermic predators were found at minimum in 1990. increased 3- or 4-fold by 1995. had subsequent declines through 1997. and a resurgence in 1998. Using the literature on Western North America. I extracted dietary information for these predators, and standardized the it with respect to prey available on the ORPI study area. Two sub-guilds were apparent, one eating smaller animals (especially lizards, perognathine rodents, and arthropods) and one eating snakes and larger small mammals (kangaroo rats, packrats. squirrels, and rabbits). Dietary overlap appears to be widespread. The small-prey sub-guild had a significantly earlier peak of observed abundance during the mid-1990's than did the large-prey sub-guild. Snakes were monitored in valley habitat on the intensive study region. This monitoring was done using the non-EMP monitoring methods, whose primary goal was to capture snakes for mark-recapture study: drift-fence trapping, time-constrained search, and road-cruising. Nearly 1900 observations of 17 species of snakes were made in this part of ORPI. Western diamondback rattlesnakes {Crotaliis atrox) and coach whips {Masticophis Jlagellum) were by far the most trophically important mammal- and reptile- eaters. respectively. For mammals, the Mohave rattlesnake (C. scutiilatiis) played an important secondary role, and the gopher snake (Pituophis melanoleuciis) played a 267 smaller role. Lizard consumption by all other snake species had a role secondary to that of the coachwhip. Snake reproduction clearly covaried with annual rainfall, and large cohorts of young rattlesnakes were produced during 1992 and 1993. For most snakes, population fluctuations during 1989-1998 did not appear to be dramatic. The western diamondback showed a consistent increase, approximately doubling by 1995, and the coachwhip increased during the early- to mid-l990's. Rattlesnake abundance apparently declined markedly during 1985-1988, after a period of greater abundance in an earlier wet cycle with strong summer rains. During 1998 (and in 1999), lizard-eating snakes showed substantial increases. In order to summarize results for this ecosystem, I constructed a set of best estimates for the population trajectories of rodent and lizard ta.xa by weighting the results from the various methods according to sample sizes, consistency, and efficacy for each species. I used precipitation as a proxy for food productivity for both small mammals and lizards. Predation pressure indices were constructed from the endotherm and snake monitoring results. Endothermic predators were weighted according the numbers observed per year, and snakes were weighted by the product of their drift-fence trap capture rate and mean body mass. Snake diets were determined by dissection of road kills. Values of monitored abundance and diet from endotherms and snakes were combined into predation-pressure indices for each major vertebrate prey taxon. Bivariate correlations and path analysis were used to evaluate relationships among predator, prey, and climate time-series. I looked for possible population-regulatory effects of predation, inter- and intra-specific competition, and changing carrying capacities 268 associated with the highly variable rainfall of the desert. Most taxa had some positive bivariate associations between population growth and precipitation, but this effect was not as overwhelming as might have been anticipated, and varied among taxa. Lizard population growth was generally correlated positively with summer rain but negatively with winter rain. Perognathine population growth was correlated with winter rain; in D. merricimi it was more strongly correlated with summer rain; and in N. cdbigiila with yearlong rainfall. Density of conspecifics was the most consistent (and negative) correlate of population growth, although the correlation results did not permit separation of intra- specific competition (in the usual sense) from density-dependent predation. Inter-specific competition was not consistently apparent in the correlations, although the analysis did support a competitive effect of D. merricimi on the perognathines. Predation had consistent negative correlations with lizard and perognathine population growth throughout the path analysis. In larger small mammals, negative bivariate correlations between predation and population growth were eliminated in path models that had density of conspecifics as one of the predictor variables. There was colinearity among some time-series, reflecting many parallel and interacting ramifications of multi-year wet or dry sequences. Thus, the data structure could not confirm the importance of predation for regulation of all taxa, although it was suggested for most, excluded for none, and strongly supported for some. The ecosystem appears to be driven by responses to rainfall-induced variations in carrying capacity, modulated by predation and competition. 269 The concepts of (apparent) competition for (or in) enemy-free space may be suitable for interpreting the structure of the assemblages studied here. I recommend addition of one or more predator monitoring protocols to the EMP, and I propose that natural resource management and academic ecology may develop a mutually beneficial collaboration in the context of monitoring programs. 270 Appendix I. List of monitored taxa. A. Small Vfammals Scientific Name Common Name lagomorphs rabbits and hares Lepiis cilleni antelope jackrabbit Lepiis califomicus black-tailed jackrabbit Sylvilagits auduboni desert cottontail sciurids squirrels A mmospennoph ilus liarrisii Harris' antelope ground squirrel Spennophilus lereticauclus round-tailed ground squirrel Dipodomys merriami Merriam's kangaroo rat Dipodomys spectabilis banner-tailed kangaroo rat Neotoma albigulci white-throated packrat (woodrat) Onychoniys torridiis southern grasshopper mouse Peromysciis eremicits cactus mouse Siginodon arizonae Arizona cotton rat perognathines pocket mice Chaetodipits baileyi Bailey's pocket mouse Chaetodipus iniermedius rock pocket mouse Chaetodipits penicillatus desert pocket mouse Perognathus ampins Arizona pocket mouse B. Lizards Cnemidoptionis whiptail lizards C. biirti red-backed whiptail C. tigris western whiptail crotaphytines leopard and collared lizards Crotaphytiis (collaris) nebrius collared lizard Gambelia wislizenii long-nosed leopard lizard Sceloponts spiny lizards S. clarkii Clark's spiny lizard S. rnagisier desert spiny lizard Coleonyx variegatiis western banded gecko Callisaunis draconoides zebra-tailed lizard Dipsosaunis dorsalis desert iguana Phrynosoma solare regal homed lizard Urosaunis omatiis tree lizard Urosatinis graciosiis long-tailed brush lizard Uta stansburiana side-bltoched lizard 271 Appendix I -- Continued C. Endothermic predators Scientific Name Common Name Can is latrans coyote Vulpes macrotis kit fox Urocyon cinereoargenteus gray fox Felis {Lyihx) riifiis bobcat Felis concolor mountain lion Mephitis mephitis striped skunk Spilogale piitorius spotted skunk TcLxiclea ta.xus badger Falco sparvariiis American kestrel Accipiter cooperii Coopers hawk Circus cyaneus northern harrier Buteo januiicensis red-tailed hawk Parabuteo unicintus Harris' hawk Aguila chrysaetos golden eagle Geococcy.x califomianus greater roadrunner Bubo virginiana great homed owl Otus kennicottii western screech-owl Tyto alba bam owl Lanius lucloviciamis loggerhead shrike D. Snakes Arizona elegans glossy snake Chilomensicus cinctus banded sand snake Chionactis palarostris Sonoran shovel-nosed snake Crotaltis atro.x western dimaondback rattlesnake Crotalus cerastes sidewinder rattlesnake Crotalus scutulatus Mohave rattlesnake Hypsiglena torquata night snake Lampropeltis getiila common kingsnake Leptotyphlops humilis westem blind snake Masticophis flagellum coachwhip (red racer) Micniroides euryxanthus westem coralsnake Phyllorhynchus browni saddled leaf-nosed snake Phyllorhynchiis decurtatus spotted leaf-nosed snake Pituophis melanoleucus gopher snake (bullsnake) Rhinocheilus lecontei long-nosed snake Salvadora hexalepis westem patch-nosed snake Trimorphodon biscutatus lyre snake 272 Appendix II. Weigiiting Used to Combine Monitoring Datasets into Best Available Estimates of Relative Annual Abundance. Zeros indicate dataset not utilized; "nt" means not yet tabulated. Method Time- Live-Trap Line Drift Road Constrained Taxon Grids Transects Fences Transect Search lagomorphs 3 1 sciurids I nt nt Neotoma albigula 4 I 1 2 Dipodornys rnerriami 3 1 I I D. speciabilis 0 0 3 1 Chaetodipiis penicillatus 3 2 1 Perognathiis ampins 3 2 I all perognathines 8 4 I 1 Onychomys 1 1 Cnemidophonis 2 I nt Uta stansbiiriana 2 I nt Urosaiiriis omatiis 1 0 nt Callisaiinis draconoides 4 I nt Sceloponis 2 I nt crotaphytines 2 1 nt Dipsosaunis dorsalis 2 1 nt Plirynosoma solare 1 0 nt Coleonlyx variegatiis I I Appendix III. Stundurdized timeseries used in tiie correlation and path analysis for vertebrates at Organ Pipe Cactus Nalional Monument. 1987-1998. A. Abundance of Smull Mammals YEAR TAXON 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 lagomorphs 0.45 1.04 0.25 0.24 0.40 1.17 2.78 1.74 1.52 1.43 0.50 0.61 sciurids 0.62 0.57 0.30 1.25 0.86 1.20 1.29 1.76 1.25 1.28 0.62 Neotoma alhigula 0.58 1.05 0.97 0.68 0.87 0.77 1.08 1.42 1.74 0.64 0.65 0.97 Dipodomys spectubilis 1.04 0.69 0.(K) 1.05 0.77 0.84 2.31 1.18 3.14 1.05 0.28 0.23 Onychomys torridus 0.69 0.10 0.10 0.50 0.99 1.19 1.93 2.20 0.61 1.17 1.06 Dipodomys merriami 1.26 0.76 0.65 0.13 1.15 1.17 1.62 2.05 0.60 l.(K) 0.51 0.56 Chaetodipus penicillatus 1.05 0.84 0.97 1.21 1.76 1.50 0.60 0.84 0.63 0.57 1.09 Perognathus ampins 0.82 0.66 0.42 0.91 1.72 1.44 0.79 1.08 0.70 0.55 1.68 ail perognathines 0.60 0.94 0.71 0.82 1.25 1.77 1.51 0.57 0.68 0.67 0.59 1.23 B, Abundance ofLl^urUs YEAR TAXON 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 Cnemidophorus 0.70 0.87 0.63 1.30 1.32 1.01 0.80 0.68 1.03 1.25 1.21 Coleonyx variegatus 1.45 0.56 1.04 0.50 0.64 1.01 0.13 0.53 0.12 1.93 2.50 Ula stanshuriami 0.60 0.56 0.70 0.76 0.71 0.80 0.94 1.07 1.72 1.40 1.48 Urosaurus oniatus 1.06 1.38 0.60 1.11 1.19 0.95 0.49 0.47 0.72 1.39 1.71 Callisaurus draconoides 1.75 1.06 0.71 1.37 1.78 0.88 0.60 0.29 0.61 1.02 1.53 Sceloporus 0.78 0.89 0.87 1.49 1.68 1.21 0.48 0.54 0.54 1.08 1.30 Dipsosaurus dorsalis 0.70 1.11 0.97 1.34 0.98 1.15 0.58 0.22 0.88 0.67 2.20 crotaphytines 1.60 0.70 0.53 0.61 2.86 1.34 1.31 0.99 0.26 0.53 0.68 Phrynosoma solare 0.(K) 2.22 1.94 1.76 2.45 1.55 0.(M) 0.(M) 0.(M) 0.58 0.49 Appendix 111. {conlinued) C. Precipitation YEAR INTERVAL 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 Annual ppt: calendar year 0.86 0.85 0.58 1.20 0.96 1.46 1.12 I.OI 0.65 0.66 0.72 1.07 summer i .i + winter 1 0.76 0.85 0.95 0.43 1.35 1.29 1.34 0.77 1.05 0.52 0.61 1.07 winter -|.| , + summer i 0.83 0.90 0.66 1.07 0.95 1.28 1.19 0.66 1.02 0.76 0.45 1.28 Seasonal ppt: winter 0.63 0.70 0.83 0.39 1.02 1.60 1.62 0.76 1.42 0.44 0.24 1.38 summer 1.03 1.10 0.49 1.75 0.88 0.97 0.77 0.55 0.61 1.08 0.66 1.18 n. Predalion Pressure Indices YEAR TAXON 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 endothermic predators (lumped) 1.44 0.50 0.64 0.50 1.02 1.24 1.69 1.71 1.93 1.48 0.61 1.05 index for lagomorphs 0.94 0.46 0.56 0.46 0.86 0.81 1.39 1.91 1.74 1.21 0.64 1.01 index for sciurids 1.09 0.46 0.56 0.53 0.84 0.99 1.34 1.85 1.62 1.15 0.67 0.99 index for packrats 1.10 0.66 0.54 0.51 0.92 0.94 1.30 1.29 1.34 1.40 0.59 1.49 index for kangaroo rats 1.07 0.63 0.49 0.54 0.85 0.93 1.40 1.70 1.41 1.41 0.63 1.01 index for pocket mice 1.83 0.83 0.61 0.58 0.87 1.09 1.42 0.94 1.32 1.03 0.45 1.12 index for lizards (except Uui) 1.29 0.64 0.86 0.73 1.08 1.13 1.39 1.15 1.05 0.99 0.71 0.90 index for side-blotched lizard 1.29 0.90 0.82 0.66 1.02 1.61 1.28 1.10 0.94 0.82 0.80 0.92 lO -J 4i- 275 REFERENCES Abramsky, Z., M.L. Rosenzweig, and A. Subach. 1998. Do gerbils care more about competition or predation? OIKOS 83:75-84. Abst, M.L. 1987. Environment and variation in life history traits of the chuckwalla. Saiiromaliis obesiis. Ecological Monographs 57:215-232. Adams, D.K. and A.C. Comrie. 1997. The North American monsoon. Bulletin of the American Meteorological Society 78:2197-2213. Anderson, D.R., K.P. Bumham. G.C. White, and D.L. Otis. 1983. Density estimation of small mammal populations using a trapping web and distance sampling methods. Ecology 64:674-680. Anderson, R.A. 1994. Functional and population responses of the lizard Cnernidophoriis ligris to environmental fluctuations. American Zoologist 34:409-421. .Andrade, E.R. and W.D. Sellers. 1988. El Niiio and its effects on precipitation in Arizona and western New Mexico. Journal of Climatology 8:403-410. Andren, C. 1982. Effect of prey density on reproduction, foraging, and other activities in the adder, Vipera hems. Amphibia-Reptilia 3:81-96. Andren, C. and G. Nilson. 1983. Reproductive tactics in an island population of adders with a fluctuating resource. Amphibia-Reptilia 4:63-79. Andrews, R.M. and F.H. Rough. 1985. Metabolism of squamate reptiles: allometric and ecological relationships. Physiological Zoology 58:214-231. Andrews, R.M. and S.J. Wright. 1994. Long-term population fluctuations of a tropical lizard: a test of causality. Pp. 267-285 in L.J. Vitt and E.R. Pianka (eds.). Lizard Ecology; Historical and Experimental Perspectives. Princeton University Press, Princeton, New Jersey, xii + 403 pp. Asdell, S.A. 1964. Patterns of Mammalian Reproduction. Cornell University Press, Ithaca, New York. 670 pp. Asplund, K.K. and C.H. Lowe. 1964. Reproductive cycles of the iguanid lizards Urosauriis oniatiis and Uta stansburiana in southeastern Arizona. Journal of Morphology 115:27-34. Ault, J.W. 197 L A quantitative esdmate of bam owl nesting habitat quality. M.S. thesis. 276 Oklahoma State University, Stillwater, cited in I. Taylor. 1994. Bam Owls: Predator-Prey Relationships and Conservation. Cambridge University Press. Cambridge. Great Britain. Bahre. C.J. 1991. A Legacy of Change: Historic Human Impact on Vegetation of the Arizona Borderlands. University of Arizona Press. Tucson. Ballinger, R.E. 1977. Reproductive strategies: food availability as a source of proximal variation in a lizard. Ecology 58:628-635. Barbault, R. 1970. Researches ecologiques dans la savane de Lamto (Cote d'lvoire): Les traits quantitatif du peuplement des ophidiens. Terre et Vie 24:94-107. Barbault, R. 1971. Les peuplements d'Ophidiens des savanes de Lamto (Cote d'lvoire). Annals of University Abidjan, ser. E. 4:133-193. Baron, J.P., R. Ferriere. J. Clobert, and H. Saint Girons. 1996. Life history of Vipera iirsinii ursinii at Mont Ventoux (France). Comptes Rendus de I'Academie des Sciences, Serie IIL Sciences de la Vie 319:1:57-69. Beaupre. S.J. 1993. An ecological study of oxygen consumption in the mottled rock rattlesnake. Crotaliis lepidiis lepidits. and the black-tailed rattlesnake, Crotcdiis inolossus molossiis. from two populations. Physiological Zoology 66:437-454. Beatley, J.M. 1969. Dependence of desert rodents on winter annuals and precipitation. Ecology 50:721-724. Beck. D.D. 1991. Physiological and behavioral consequences of reptilian life in the slow lane: ecology of beaded lizards and rattlesnakes. Ph.D Dissertation, University of Arizona, Tucson. Beck, D.D. 1995. Ecology and energetics of three sympatric rattlesnake species in the Sonoran Desert. Journal of Herpetology 29:2II-223. Bednarz, J.C. 1988. A comparative study of the breeding ecology of Harris' and Swainson's hawks in southeastern New Mexico. The Condor 90:311-323. Begon. M, J.L. Harper, and C.R. Townsend. 1990. Ecology: Individuals, Populations and Communities. 2nd Edition. Blackwell Scientific Publications, Boston. Massachusetts. Begon, M, J.L. Harper, and C.R. Townsend. 1996. Ecology: Individuals, Populations and Communities. 3rd Edition. Blackwell Science Ltd, Oxford, UK. xii-f-1068 pp. Ill Begon, M, M. Mortimer, and D.J. Thompson. 1996. Population Ecology; A Unified Study of Animals and Plants. 3rd edition. Blackwell Science. Great Britain. Bent, A.C. 1938. Life Histories of North American Birds of Prey. Vol. 2. U.S. National Museum Bulletin 170:482 pp. Bock, C.E. and J.H. Bock. 1994. Effects of predator exclusion on rodent abundance in an Arizona semidesert grassland. Southwestern Naturalist 39:208-210. Bounds, D.L. 1996. Comparison of pitfall traps and belt transects to e.xamine lizard populations in different vegetative communities. Unpublished doctoral dissertation. University of Arizona, Tucson, Arizona. 194 pp. Bouskila. A. 1995. Interactions between predation risk and competition: a field study of kangaroo rats and snakes. Ecology 76:165-178. Bowers. J. 1982. Checklist of vascular plants of Organ Pipe Cactus National Monument. Arizona. Southwest Parks and Monuments Association. Tucson, Arizona. Pp. 1032. Bowers, M.A., D.B. Thompson and J.H. Brown. 1987. Foraging and microhabitat use in desert rodents. Oecologia 72:77-82. Brown. D.E. 1994. Biotic Communities, Southwestern United States and Northwestern Mexico. 2nd Ed. University of Utah Press. Logan. Brown, D.E. and C.H. Lowe. 1980. Biotic Communities of the Southwest. USDA Forest Service General Technical Report RM-78. Map (one page). Rocky Mountain Forest and Range Experiment Station. Fort Collins. Colorado. Brown, J.H. 1973. Species diversity of seed-eating desert rodents in sand dune habitats. Ecology 54(4):755-787. Brown. J.H. 1975. Geographical ecology of desert rodents. Pp. 315-341 in M.L. Cody and J.M. Diamond {eds.). Ecology and Evolution of Communities. Belknap Press, Harvard University. Cambridge. Massachusetts. Brown. J.H. and E.J. Heske. 1990a. Temporal change in a Chihuahuan Desert rodent community. OIKOS 59:290-302. Brown, J.H. and E.J. Heske. 1990b. Control of a desert-grassland transition by a keystone rodent guild. Science 250:1705-1707. 278 Brown, J.H. and M.A. Kurzius. 1987. Composition of desert rodent faunas: combinations of coexisting species. Ann. Zool. Fennici 24:227-237. Brown, J.H. and G.A. Lieberman. 1973. Resource utilization and coexistence of seed- eating rodents in sand dune habitats. Ecology 54(4):788-797. Brown, J.H. and J.C. Munger. 1985. Experimental manipulation of a desert rodent community: food addition and species removal. Ecology 66:1545-1563. Brown, J.H., T.J. Valone. and C.G. Curtin. 1997. Reorganization of an arid ecosystem in response to recent climatic change. Proceedings of the National Academy of Science (USA) 94:9279-9733. Brown, J.H. and Z.Zeng. 1989. Comparative population ecology of eleven species of rodents in the Chihuahuan Desert. Ecology 70:1507-1525. Brown, J.S. 1988. Patch u.se as an indicator of habitat preference, predation risk, and competition. Behavioral Ecology and Sociobiology 22:37-47. Brown. J.S. 1989. Desert rodent community structure: a test of four mechanisms of co existence. Ecological Monographs 59(1): 1-20. Brown, J.S., B.P. Kotler, and W.A. Mitchell. 1994b. Foraging theory, patch use. and the structure of a Negev Desert granivore community. Ecology 75:2286-2300. Brown, J.S., B.P. Kotler, R.J. Smith, and W.O. Wirtz II. 1988. The effects of owl predation on the foraging behavior of heteromyid rodents. Oecologia 76:408-415. Brown, J.S., B.P. Kotler, and T.J. Valone. 1994a. Foraging under predation: a comparison of energetic and predation costs in rodent communities of the Negev and Sonoran Deserts. Australian Journal of Zoology 42:435-448. Brown, W.S. 1993. Biology, status, and management of the timber rattlesnake {Crotalus liorridiis): a guide for conservation. Society for the Study of Amphibians and Reptiles Herpetological Circulars 22;vi-h78 pp. Broyles, B. 1997. Wildlife water developments in southwestern Arizona. Journal of the Arizona-Nevada Academy of Science 30:30-42 Bryant, H.C. 1916. Habits and food of the roadrunner in California. University of California Publications in Zoology 17:21-58. Buckland, S.T., D.R. Anderson, K.P. Bumham, and J.L. Laake. 1995. Distance 279 Sampling: Estimating Abundance of Biological Populations. Chapman and Hall, London. Burt, W.H. and R.P. Grossenheider. 1964. A Field Guide to the Mammals. Houghton Mifflin Co., Boston. Massachusetts, xxv+284 pp. Callicott, J.B., L.B. Crowder, and K. Mumford. 1999. Current normative concepts in conservation. Conservation Biology 13:22-35. Camp, R.J., R.L. Knight, H.A.L. Knight. M.W. Sherman, and J.Y. Kawashima. 1993. Food habits of nesting common ravens in the eastern Mojave Desert. Southwestern Naturalist 38:163-165. Campbell. H.W. and S.P. Christman. 1982. Field techniques for herpetofaunal community analysis. Pp. 193-200 in N.J. Scott Jr. (ecL), Herpetological Communities: A Symposium. U.S. Fish & Wildlife Service Research Report 13, Washington, D.C. Carpenter, S.R.. S.W. Chisholm, C.J. Krebs. D.W. Schindler. and R.F. Wright. 1995. Ecosystem experiments. Science 269:324-327. Case, T.J. 1983. Sympatry and size-similarity in Cneniidophonis. Pp. 297-325 in R.B. Huey, E.R. Pianka, and T.W. Schoener {eels.). Lizard Ecology: Studies of a Model Organism. Harvard University Press. Cambridge, Massachusetts, vii+501 pp. Case, T.J. 1994. Introduction to Part IV. Population and Community Ecology. Pp. 261- 265 in L.J. Vitt and E.R. Pianka {eels.). Lizard Ecology: Historical and Experimental Perspectives. Princeton University Press. Princeton, New Jersey, xii + 403 pp. Chapman, D.G. 1951. Some properties of the hypergeometric distribution with applications to zoological censuses. Univ. Calif. Publ. Stat. 1:131-160. Chappell, M.A. and T.M. Ellis. 1987. Resting metabolic rates in boid snakes: allomeunc relationships and temperature effects. Journal of Comparative Physiology B 157:227-235. Chew, R.M. and A.E. Chew. 1970. Energy relationships of the mammals of the desert scrub {Lxirrea tridentata) community. Ecological Monographs 40:1-21. Clements, F.E. 1916. Plant succession: analysis of the development of vegetation. Carnegie Institution of Washington Publication 242:1-512. 280 Cockrum, E.L. 1982. Mammals of the Southwest. University of Arizona Press, Tucson. Cockrum. E.L. and Y. Petryszyn. 1986. Mammals of the Organ Pipe Cactus National Monument. USGS Biological Resources Division. Cooperative Park Studies Unit, University of Arizona, Tucson, Arizona, 85721. Special Report No. 5:1-89. Cody. M.L. and J.M. Diamond {eds.). 1975. Ecology and Evolution of Communities. Belknap Press of Harvard University. Cambridge. Massachusetts. Conley. W.. J.D. Nichols, and A.R. Tipton. 1977. Reproductive strategies in desert rodents. Pp. 193-215 in R.H. Wauerand D.H. Riskind (eds.). Transactions of the Symposium on the Biological Resources of the Chihuahuan Desert Region, United States and Mexico. U.S. National Parks Service Transactions and Proceedings Series No. 3. Connell. J.H. 1975. Some mechanisms producing structure in ecological communities. Pp. 460-490 in M.L. Cody and J.M. Diamond {eds.). Ecology and Evolution of Communities. Belknap Press of Harvard University, Cambridge. Massachusetts. Connell. J.H. 1983. On the prevalence and importance of competition in ecological communities: evidence from field experiments. American Naturalist 122:661-696. Conroy. S. 1999. Lizard assemblage response to a forest ecotone in northeastern Australia: a synecological approach. Journal of Herpetology 33:409-419. Cooper. W.E. Jr. 1994. Prey chemical discrimination, foraging mode, and phylogeny. Pp. 95-116 in L.J. Vitt and E.R. Pianka {eds.). Lizard Ecology: Historical and Experimental Perspectives. Princeton University Press, Princeton, New Jersey, xii + 403 pp. Creusere. M.F. and W.G. Whitford. 1982. Temporal and spatial resource partitioning in a Chihuahuan Desert lizard community. Pp. 121-128 in N.J. Scott Jr. {ed.), Herpetological Communities: A Symposium. U.S. Fish & Wildlife Service Research Report 13, Washington, D.C. Crews, D. and W.R, Garstka. 1982. The ecological physiology of a garter snake. Scientific American 247:5:159-168. Daly, M., P.R. Behrends, M.L Wilson, and L.F. Jacobs. 1992. Behavioural modulation of predation risk: moonlight avoidance and crepuscular compensation in a nocturnal desert rodent, Dipodomys merriami. Animal Behaviour 44:1-9. 281 Davidson, D.W., J.H. Brown, and R.S. Inouye. 1980. Competition and the structure of granivore communities. Bioscience 30:233-238. Dawson. J.W. 1998. Great homed owl. Bubo virginianus. Pp. 152-155 in R.L. Glinski (ed.). The Raptors of Arizona. University of Arizona Press, Tucson. Degenhardt, W.G. 1966. A method of counting some diurnal ground lizards of the genera Holbrookia and Cnemidophonis with results from the Big Bend National Park. American Midland Naturalist 75:61-100. Diaz, M. 1992. Rodent seed predation in cereal crop areas of central Spain: effects of physiognomy, food availability, and predation risk. Ecography 15:77-85. Dice, L.R. 1938. Some census methods for mammals. Journal of Wildlife Management 2:119-120. Dickie, L.M., S.R. Kerr, and B.P. Boudreau. 1987. Size-dependent processes underlying regularities in ecosystem structure. Ecological Monographs 57:233-250. Diller, L.V. and D.R. Johnson. 1988. Food habits, consumption rates, and predation rates of western rattlesnakes and gopher snakes In southwestern Idaho. Herpetologica 44:228-233. Djawdan, M. and T. Garland, Jr. 1988. Maximal running speeds of bipedal and quadrupedal rodents. Journal of Mammalogy 69:765-772. Donazar, J.A., F. Hiraldo. M. Delibes. and R.R. Estrella. 1989. Comparative food habits of the eagle owl Bubo bubo and the great homed owl Bubo virginianus in six Palearctic and Nearctic biomes. Omis Scandinavica 20:298-306. Dunham, A.E. 1978. Food availability as a proximate factor influencing individual growth rates in the iguanid lizard Sceloporus merriami. Ecology 59:770-778. Dunham, A.E. 1980. An experimental study of interspecific competition between the iguanid lizards Sceloporus merriami and Urosaurus omatus. Ecological Monographs 50:309-330. Dunham. A.E. 1981. Populations in a fluctuating environment: the comparative population ecology of the iguanid lizards Sceloporus merriami and Urosaurus omatus. Miscellaneous Publications of the Museum of Zoology, University of Michigan 158:1-62. Dunham. A.E. 1983. Realized niche overlap, resource abundance, and intensity of 282 interspecific competition. Pp. 261-280 in R.B. Huey. E.R. Pianka, and T.W. Schoener {eds.). Lizard Ecology: Studies of a Model Organism. Harvard University Press. Cambridge, Massachusetts, viii + 501 pp. Dunham, A.E.. D.B. Miles, and D.N. Reznick. 1988a. Life history patterns in squamate reptiles. Pp. 441-452 in C. Cans and R.B. Huey {eds.). Biology of the Reptilia, Vol. 16. Ecology B. Defense and Life History. A.R. Liss, New York. Dunham, A.E., P.J. Morin. and H.M. Wilbur. 1988b. Methods for the study of reptile populations. Pp. 331-386 in C. Gans and R.B. Huey (eds.). Biology of the Reptilia. Vol. 16. Ecology B. Defense and Life History. A.R. Liss, New York. Erlinge, S.. G. Gdransson, L. Hansson, G. Hogstedt, O. Liberg, LN. Nilsson. T. Nillson, T. Schantz, and M. Sylven. 1983. Predation as a regulating factor on small rodent populations in southern Sweden. OIKOS 40:36-52. Erlinge. S.. G. Goransson, L. Hansson, G. Hogstedt, O. Liberg. LN. Nilsson, T. Nillson. T. Schantz. and M. Sylven. 1984. Can vertebrate predators regulate their prey? American Naturalist 123:125-133. Erlinge, S., G. Goransson, L. Hansson, G. Hogstedt, O. Liberg, LN. Nilsson, T. Nillson, T. Schantz, and M. Sylven. 1988. More thoughts on vertebrate predator regulation of prey. American Naturalist 132:148-154. Ernest. K.A. 1994. Resistance of creosotebush to mammalian herbivory: temporal consistency and browsing-induced changes. Ecology 75:1684-1692. Errington, P.L. 1946. Predation and vertebrate populations. Quarterly Review of " Biology 21:144-177. Felger, R.S. 1990. Non-native plants of Organ Pipe Cactus National Monument. Arizona. USGS Biological Resources Division. Cooperative Park Studies Unit, University of Arizona. Tucson, Arizona. 85721. Technical Report 31:vii+92 pp. Felger, R.S., B. Royles, M. Wilson, and G.P. Nabham. 1997. The Binational Sonoran Desert Network and its plant life. Pp. 411-560 in R.S. Felger and B. Broyles (eds.). Dry Borders: Binational Sonoran Desert Reserves. Journal of the Southwest 39:3 and 4:297-860. The Southwest Center. University of Arizona. Tucson. Ferguson, G.W. and S.F. Fox. 1984. Annual variation of survival advantage of large juvenile side-blotched lizards, C/ta stansbiiriana: its causes and evolutionary consequences. Evolution 38:342-349. 283 Fisher, J.L. 1981. Kit fox diet in south-central Arizona. M.S. thesis. University of Arizona, Tucson. Fitch, H.S. 1947. Predation by owls in the Sierra Nevadan foothills of California. The Condor 49:137-151. Fitch. H.S. 1948. A study of coyote relationships on cattle range. Journal of Wildlife Management 12:73-78. Fitch, H.S. 1949. Study of snake populations in central California. American Midland Naturalist 41:513-579. Fitch, H.S. 1958. Home ranges, territories, and seasonal movements of vertebrates of the Natural History Reservation. University of Kansas Museum of Natural History Publications 11:63-326. Fitch, H.S. I960. Autecology of the copperhead. University of Kansas Museum of Natural History Publications 8:417-476. Fitch, H.S. 1975. A demographic study of the ringneck snake (Diadopliis punctcitus) in Kansas. University of Kansas Museum of Natural History Miscellaneous Publications 62:1-53. Fitch, H.S. 1982. Resources of a snake community in prairie-woodland habitat of northeastern Kansas. Pp. 83-97 in N.J. Scott Jr. (ed.), Herpetological Communities. U.S. Fish and Wildlife Service Research Report 13. Washington, DC. iv-f-239 pp. Fitch, H.S., V.R. Fitch, and W.D. Kettle. 1984. Reproduction, population changes, and interactions of small mammals on a natural area in northeastern Kansas. Occasional Papers of the University of Kansas Museum of Natural History 109:1- 37. Fitch. H.S., B. Glading. and V. House. 1946. Cooper's hawk nesting and predation. California Fish and Game 32:144-154. Fitch. H.S. and G.R. Pisani. 1993. Life history traits of the western diamondback rattlesnake (Crotalus atrox) studied from roundup samples in Oklahoma. Occasional Papers of the University of Kansas Museum of Natural History 156:1- 24. Fitch, H.S., F. Swenson, and D.F. Tillotson. 1946. Behavior and food habits of the red- tailed hawk. Condor 48:205-237. 284 Flatt, T., S. Dummermuth, and B.R. Anholt. 1997. Mark-recapture estimates of survival in populations of the asp viper, Vipera aspis aspis. Journal of Herpetology 31:558-564. Forsman, A. and L.E. Lindell. 1991. Trade-off between growth and energy storage in male Vipera bents (L.) under different prey densities. Functional Ecology 5:717- 723. Forsman, A. and L.E. Lindell. 1997. Responses of a predator to variation in prey abundance: survival and emigration of adders in relation to vole density. Canadian Journal of Zoology 75:1099-1108. Fox, B.J. and J.H. Brown. 1993. Assembly rules for functional groups in Nonh American desert rodent communities. OIKOS 67:358-370. French, N.R., B.C. Maza, and A.P. Aschwanden. 1967. Life spans of Dipodomys and Perognathus in the Mojave Desert. Journal of Mammalogy 48:537-546. Frye, R.J. 1983. Experimental evidence of interspecific aggression between two species of kangaroo rat {Dipodomys). Oecologia 59:74-78. Fukada, H. 1969. Biological studies on the snakes. XIIL Preliminary estimate of population size in Tambabashi study area. Bulletin Kyoto University Educ. Ser. B Math. Nat. Sci. 36:3-9. Funk, R.S. 1965. Food of Crotaiits cerastes lateropens in Yuma County, Arizona. Herpetologica 21:15-17. Garland, T. Jr. 1993. Locomotor performance and activity metabolism of Cnemidophoriis tigris in relation to natural behaviors. Pp. 163-210 in J.W. Wright and L.J. Vitt {eds.). Biology of Whiptail Lizards {Cnemidophoriis). Oklahoma Museum of Natural History, Norman, Oklahoma, xiv+417 pp. Gleason, H.A. 1926. The individualistic concept of the plant association. Torrey Botanical Club Bulletin 53:7-26. Goldberg, S.R. 1987. Reproductive cycle of the giant spotted whiptail, Cnemidophoriis biirti slictogrammiis. in Arizona. Southwestern Naturalist. 32:510-511. Goldberg, S.R. and C.H. Lowe. 1997. Reproductive cycle of the Gila monster, Heloderma suspectum, in southern Arizona. Journal of Herpetology 31:161-166. Golightly, R.T. 1981. Comparative energetics of two desert canids: the coyote Canis 285 latrans) and the kit fox (Vulpes macroiis). Unpublisiied doctoral dissertation. Arizona State University. Tempe. Arizona. 174 pp. Golightly. R.T. and R.D. Ohmart. 1983. Metabolism and body temperature of two desen canids: coyotes and kit foxes. Journal of Mammalogy 64:624-635. Golightly, R.T. and R.D. Ohmart. 1984. Water economy of two desert canids: coyote and kit fox. Journal of Mammalogy 65:51-58. Gonzales-Romero. A., A. Ortega, and R. Barbault. 1989. Habitat partitioning and spatial organization in a lizard community of the Sonoran Desert. Mexico. Amphibia- Reptilia 10:1-11. Gonzales-Romero, A., and S. Alvarez-Cardenas. 1989. Herpetofauna de la region del Pinacate. Sonora, Mexico: un inventario. Southwestern Naturalist 34:519-526. Gorsuch, D.M. 1932. The roadrunner. Arizona Wild Life, October 1932:1-11. Gregory, P.T. 1977. Life history parameters of the red-sided ganer snake (Tlianmopliis sirtalis parieuiUs) in an extreme environment, the interlake region of Manitoba. National Museums of Canada Publications in Zoology 13: 1-44. Grinnell. J.. J.S. Dixon, and J.M. Linsdale. 1937. Fur-Bearing Mammals of California. Vols. I and IL University of California Press, Berkeley. Groschupf, K.D.. B.T. Brown, and R.R. Johnson. 1988. An annotated checklist of the birds of Organ Pipe Cactus National Monument. Southwest Parks and Monuments Association, and USGS Cooperative Park Studies Unit, University of Arizona. Tucson. Arizona. 45 pp. Hailey, A. and P.M.C. Davies. 1987. Digestion, specific dynamic action, and ecological energetics of Matrix maura. Herpetological Journal 1:159-166. Hairston, N.G. Sr. 1989. Ecological Experiments: Purpose, Design, and Execution. Cambridge University Press, Cambridge, Great Britain. Halvorson, W.L. and G.E. Davis {eds.). 1996. Science and Ecosystem Management in the National Parks. University of Arizona Press, Tucson. Hanski, L L. Hansson. and H. Hentlonen. 1991. Specialist predators, generalist predators, and the microtine rodent cycle. Journal of Animal Ecology 60:353-367. Hanski, I, P. Turchin, E. Korpimiiki, and H. Henttonen. 1993. Population oscillations of 286 boreal rodents: regulation by mustelid predators leads to chaos. Nature 364:232- 235. Hanski, I. and H. Henttonen. 1996. Predation on competing rodent species: a simple explanation of complex patterns. Journal of Animal Ecology 65:220-232. Hanski, I., L. Hansson, and H. Henttonen. 1991. Specialist predators, generalist predators, and the microtine rodent cycle. Journal of Animal Ecology 60:353-367. Hans.son, L. 1987. An interpretation of rodent dynamics as due to trophic interactions. OIKOS 50:308-318. Harris. J.N. 1984. An experimental analysis of desert rodent foraging ecology. Ecology 65(5): 1579-1584. Hekkilii, J., A. Below, and I. Hanski. 1994. Synchronous dynamics of microtine rodent populations on islands in Lake Inari in northern Fennoscandia: evidence for regulation by mustelid predators. OIKOS 70:245-252. Hensley. M.M. 1954. Ecological relations of the breeding bird population of the desert biome in Arizona. Ecological Monographs 24:185-207. Heske, E.J.. J.H. Brown, and Q. Gao. 1993. Effects of kangaroo rat e.xclusion on vegetation structure and plant species diversity in the Chihuahuan Desert. Oecologia 95:520-524. Heske, E.J., J.H. Brown, and S. Mistry. 1994. Long-term experimental study of a Chihuahuan Desert rodent community: 13 years of competition. Ecology 75:438- 445. Hirth, H.F. and A.C. King. 1966. Biomass densities of snakes in the cold desert of Utah. Herpetologica 22:8-12. Hoffmeister, D.F. 1986. Mammals of Arizona. University of Arizona Press. Tucson. Holm, P.H. 1988. Two populations of the tree lizard {Urosauriis omatiis) in southern Arizona. Unpublished Masters Thesis, University of Arizona, Tucson, xi + 107 pp. Holt, R.D. 1977. Predation, apparent competition, and the structure of prey communities. Theoretical Population Biology 12:197-229. Holt, R.D. 1984. Spatial heterogeneity, indirect interactions, and the coexistence of prey 287 species. American Naturalist 124:377-406. Holt. R.D. and B.P. Kotler. 1987. Short-term apparent competition. American Naturalist 130:412-430. Holt. R.D. and J.H. Lawton. 1993. Apparent competition and enemy-free space in in.sect host-parasitoid communities. American Naturalist 142:623-645. Howard, C.W. 1974. Comparative reproductive ecology of homed lizards (Genus Phrynosomci) in southwestern United States and northern Mexico. Journal of the .Arizona Academy of Science 9:108-116. Huey, R.B. and E.R. Pianka. 1981. Ecological consequences of foraging mode. Ecology 62:9919-999. Hughes. J.M. 1996. Greater roadrunner. The Birds of North America no. 244. American Ornithologists' Union and Academy of Natural Sciences, Philadelphia. Pennsylvania. Ives. A.R., J. Foufopoulos, E.D. Klopfer. J.L. Klug. and T.M. Palmer. 1996. Bottle or big-scale studies: how do we do ecology? Ecology 77:681-685. Jaksic. P.M. P. Feinsinger, and J.E. Jimenez. 1993. A long-term study of the dynamics of guild structure among predatory vertebrates at a semi-arid Neotropical site. OikoS 67:87-96. J^drzejewski, W., L. Rychlik, and B. J(?drzejewski. 1993. Responses of bank voles to odours of seven species of predators: experimental data and their relevance to natural predator-vole relationships. OIKOS 68:251-257. J?drzejewski, W.. L. Rychlik, and B. J?drzejewski. 1996. Rodent cycles in relation to biomass and productivity of ground vegetation and predation in the Palearctic. Acta Theriologica 41:1-34. Jeffries, M.J. and J.H. Lawton. 1984. Enemy-free space and the structure of ecological communities. Biological Journal of the Linnean Society 23:269-286. Jeffries. M.J. and J.H. Lawton. 1985. Predator-prey ratios in communities of freshwater invertebrates: the role of enemy-free space. Freshwater Biology 15:105-112. Jones, J.H. and N.S. Smith. 1979. Bobcat density and prey selection in central Arizona. Journal of Wildlife Management 43:666-672. 288 Jones, T.J. 1985. Body size and life history variables in heteromyids. Journal of Mammalogy 66:128-132. Kams, D.R. 1986. Field herpetology: methods for the study of amphibians and reptiles in Minnesota. Occasional Papers of the James Ford Bell Museum of Natural History, University of Minnesota 18:1-86. Keith. L.B. 1983. Role of food in hare population cycles. OIKOS 40:385-395. Keith, L.B.. J.R. Gary, R.J. Rongstad, and M.C. Brittingham. 1984. Demography and ecology of declining snowshoe hare populations. Journal of Wildlife Management 90:1-43. Kelt. D.A.. J.H. Brown. E.J. Heske, P.A. Marquet, S.R. Morton, J.R.W. Reid, K.A. Rogovin, and G. Shenbrot. 1996. Community structure of desert small mammals: comparisons across four continents. Ecology 77:746-761. Kenagy, G.J. and G.A. Bartholomew. 1985. Seasonal reproductive patterns in five coe.xisting California desert rodent species. Ecological Monographs 55:371-397. King. R.B. 1986. Population ecology of the Lake Erie water snake, Nerodia sipeclon insiilarum. Copeia 1986:757-772. Kingsley, K.J. 1998. Invertebrates of Organ Pipe Cactus National Monument. U.S. Geological Survey, Cooperative Park Studies Unit, University of Arizona, Technical Report 60:viii-i-l87pp. Klimstra. W.D. 1958. Some observations on snakes activities and populations. Ecology 39:232-239. Kline. N.C. and D.E. Swann. 1998. Quantifying wildlife road mortality in Saguaro National Park. Pp 23-27 in G.L. Evink. P.Garrett. D. Zeigler. and J. Berry {eels.). Proceedings of the International Conference on Wildlife Ecology and Transportation. FL-ER-69-98. Florida Department of Transportation. Tallahassee. Rorida. Knight. R.L. and R.E. Jackman. 1984. Food niche relationships between great homed owls and common bam owls in eastern Washington. Auk 101:175-179. Korpimaki. E. and C.J. Krebs. 1996. Predation and population cycles of small mammals: a reassessment of the predation hypothesis. Bioscience 46:754-764. Korpimaki, E. and K. Norrdahl. 1991. Do breeding nomadic avian predators dampen 289 population fluctuations of small mammals'? OIKOS 62:195-208. Korpimaki, E., K. Norrdahl, and T. Rinta-Jaskari. 1994. Responses of stoats and least weasels to fluctuating food abundances: is the low phase of the vole cycle due to mustelid predation.Oecologia 88:552-561. Kotler, B.P, J.S. Brown. R.H. Slotow, W.L. Goodfriend and M. Strauss. 1993b. The influence of snakes on the foraging behavior of gerbils. OIKOS 67:309-316. Kotler, B.P. 1984. Risk of predation and the structure of desert rodent communities. Ecology 65:689-701. Kotler, B.P. 1985. Owl predation on desert rodents which differ in morphology and behavior. Journal of Mammalogy 66(4):824-828. Kotler. B.P. 1988. Temporal variation in the structure of a desert rodent community. Pp. 127-139 in D.W. Morrise, Z. Abramsky. B.J. Fox. and M.R. Willig (eels.). Patterns in the Structure of Mammalian Communities. Texas Tech University Press. Lubbock, Texas. Kotler, B.P., L. Blaustein and J.S. Brown. 1992. Predator facilitation: the combined effect of snakes and owls on the foraging behavior of gerbils. Ann. Zool. Fenn. 29:199-206. Kotler, B.P.. L. Blaustein and H. Dednam. 1993a. The specter of predation: the effects of vipers on the foraging behavior of two gerbilline rodents. Israel Journal of Zoology 39:11-21. Kotler, B.P. and J.S. Brown. 1988. Environmental heterogeneity and the coexistence of desert rodents. Annual Review of Ecology and Systematics 19:281-307. Kotler. B.P., J.S. Brown, and O. Hasson. 1991. Factors affecting gerbil foraging behavior and rates of owl predation. Ecology 72(6):2249-2260. Kotler, B.P.. J.S. Brown, and W.A. Mitchell. 1994. The role of predation in shaping the behavior, morphology and community organization of desert rodents. Australian Journal of Zoology 42:449-466. Kotler, B.P.. J.S. Brown, R.J. Smith, and W.O. Wirtz II. 1988. The effects of morphology and body size on rates of owl predation on desert rodents. OIKOS 53:145-152. Kotler, B.P. and R.D. Holt. 1989. Predation and competition: the interaction of two 290 types of species interactions. OIKOS 54:256-260. Krebs. C.J. 1989. Ecological Methodology. Harper & Roe, New York. Krebs. C.J., R. Boonstra, S. Boutin, S. Dale, S. Hannon, K. Martin. A.R.E. Sinclair, J.N.M. Smith, and R. Turkington. 1992. What drives the snowshoe hare cycle in Canada's Yukon? Pp. 886-896 in D.M. McCullough and R. Barrett (eds.). Wildlife 2001: Populations. Elsevier Press, London, UK. Lagos, V.O., L.S. Contreras, P.L. Meserve, J.R. Gutierrez, and P.M. Jaksic. 1995. Effects of predation risk on space use by small mammals: a field e.xperiment with a Neotropical rodent. OIKOS 74:259-264. Lambin, X.. S.J. Petty, and J.L Mackinnon. 2000. Cyclic dynamics in field vole populations and generalist predation. Journal of Animal Ecology 69:106-118. LaRoe. E.T., G.S. Farris, C.E. Puckett. P.D. Doran and M.J. Mac. {editors). 1995. Our Living Resources: a report to the nation on the distribution, abundance, and health of U.S. plants, animals, and ecosystems. U.S. Dept. Int.. Natl. Biological Serv. Wash.. D.C. Larsen. K.W. and P.T. Gregory. 1989. Population size and survivorship of the common garter snake. Tliamnophis sirtalis. near the northern limit of its distribution. Holarctic Ecology 12:81-86. Law. R. and R.D. Morton. 1996. Permanence and the assembly of ecological communities. Ecology 77:762-775. Lawton. J.H. 1995. Ecological experiments with model systems. Science 269:328-331. Lawton. J.H. 1996. The ecotron facility at Silwood Park: the value of big bottle experiments. Ecology 77:665-669. Lemen. C.A. and M.L. Rosenzweig. 1978. Microhabitat selection in two sympatric species of heteromyid rodents. Oecologia 33:127-135. Lever. C. 1994. Naturalized Animals: the Ecology of Successfully Introduced Animals. T & A D Poyser Ltd., London, Great Britain. Ligon, J.D. 1968. The biology of the elf owl. Miscellaneous Publications of the Museum of Zoology, University of Michigan 136:1-70. Lillywhite, H.B. 1987. Temperature, energetics, and ecological physiology. Pp. 422-477 291 in R.A. Seigel, J.T. Collins, and S.S. Novak (eds.). Snakes: Ecology and Evolutionary Biology. Macmillan, New York, xiv+529 pp. Lima. S.L. 1998. Nonlethal effects in the ecology of predator-prey interactions. Bioscience 48:25-34. Lindell, L.E. 1997. Annual variation in growth rate and body condition of adders. Vipera berus: effects of food availability and weather. Canadian Journal of Zoology 75:261-270. Lindell, L.E. and A. Forsman. 1996. Density effects and snake predation: prey limitation and reduced growth rate of adders at high density of conspecifics. Canadian Journal of Zoology 74:1000-10007. Lindeman, R. 1942. The trophic-dynamic aspect of ecology. Ecology 23:399-418. Lindstrom, E.R.. H. Andren, P. Angelstam, G. Cederlund, B. Homfeldt, L Jiiderberg. P. Lemnell, B. Martinsson, K. Skold, and J.E. Swenson. 1994. Disease reveals the predator: sarcoptic mange, red fox predation, and prey populations. Ecology 75:1042-1049. Llinas-Gutierrez, J.. G. Amaud, and M.Acevado. 1991. Food habits of the great horned owl {Bubo virginianiis) in the Cape region of Lower California, Mexico. Journal of Raptor Research 25:140-141. Loehlin. J.C. 1987. Latent Variable Models: An Introduction to Factor. Path, and Structural Analysis. Lawrence Erlbaum Associates, Publishers. Hillsdale, New Jersey, x + 273 pp. Longland, W.S. and M.V. Price. 1991. Direct observations of owls and heteromyid rodents: can predation risk explain microhabitat use? Ecology 72(6);2261-2273. Lowe, C.H. and D.S. Hinds. 1971. Effect of paloverde (Cercidiian) trees on the radiation tlux at ground level in the Sonoran Desert in winter. Ecology 52:916-922. Lowe, C.H., C.R. Schwalbe, and T.B. Johnson. 1986. The Venomous Reptiles of Arizona. Arizona Game and Fish Department, Phoenix, Arizona, ix+115 pp. Lutz, T.V. 1980. Taxonomy and distribution of the scorpions of Organ Pipe Cactus National Monument. Unpublished Masters thesis, Arizona State University, Tempe, Arizona, viii-i-38 pp. MacArthur, R.H. and R. Levins. 1967. The limiting similarity, convergence and 292 divergence of coexisting species. American Naturalist 101:377-385. Macartney, J.M., P.T. Gregory, and M.B. Charland. 1990. Growth and sexual maturity of the western rattlesnake, Crotcilus viridis, in British Columbia. Copeia 1990:528-542. Mader, W.J. 1975. Biology of the Harris' hawk in southern Arizona. The Living Bird 14:59-87. Mader. W.J. 1978. A comparative nesting study of red-tailed hawks and Harris' hawks in southern Arizona. Auk 95:327-337. Madsen, T. 1988. Reproductive success, mortality and sexual size dimorphism in the adder, Vipera bents. Holarctic Ecology 11:77-80. Madsen, T. and R. Shine. 1992. Determinants of reproductive success in female adders. Vipera beriis. Oecologia 92:40-47. Madsen. T.. R. Shine, R. Loman. and T. Hakansson. 1993. Determinants of mating success in male adders. Animal Behavior 45:491-499. Madsen, T. and B. Stille. 1988. The effect of size dependent mortality on colour morphs in male adders. Vipera berus. OIKOS 52:73-78. Magnusson, W.E., L.J. Paiva. R.M. Rocha. C.R. Franke, L.A. Kasper. and A.P. Lima. 1985. The correlates of foraging model in a community of Brazilian lizards. Herpetologica 41:324-332. Marti. C.D. 1974. Feeding ecology of four sympatric owls. The Condor 76:45-61. Marti. C.D., K. Steenhof. M.N. Kochert. and J.S. Marks. 1993. Community trophic structure: the roles of diet, body size, and activity times in vertebrate predators. OIKOS 67:6-18. Martinez. N.D. and J.H. Lawton. 1995. Scale and food-web structure ~ from local to global. OIKOS 73:148-154. Massot. M., J. Clobert, and T. Pilorge. 1992. Density dependence in the common lizard: demographic consequences of a density manipulation. Ecology 73:1742-1756. Maury. M.E. and R. Barbault. 1981. The spatial organization of the lizard community of the Bolson de Mapimi (Mexico). Pp. 79-87 in R. Barbault and G. Halffter {eds.). Ecology of the Chihuahuan Desert. Instituto de Ecologia. Distrito Federal, 293 Mexico. May. R.M. 1973. Stability and Complexity in Model Ecosystems. Princeton University Press, Princeton, New Jersey. McAuliffe, J.R. 1986. Herbivore-limited establishment of a Sonoran Desert tree, Cerciclium microphyUiim. Ecology 67:276-280. McAuliffe, J.R. 1994. Landscape evolution, soil formation, and ecological pattems and processes in Sonoran Desert bajadas. Ecological Monographs 64:111-148. McCann, K., A. Hastings, and G.R. Huxel. 1998. Weak trophic interactions and the balance of nature. Nature 395:794-798. McClure, M.F. 1993. Densities and diets of coyotes near Saguaro National Monument. M.S. thesis. University of Arizona, Tucson. Mcintosh, R.P. 1995. H. A. Gleason's 'individualistic concept' and theory of animal communities: a continuing controversy. Biological Reviews (Cambridge) 70:317- 357. McLaren. B.E. and R.O. Peterson. 1994. Wolves, moose, and tree rings on Isle Royale. Science 266:1555-1558. Medica, P.A., F.B. Turner, and D.D. Smith. 1973. Effects of radiation on a fenced population of homed lizards {Phrynosoma platyrhinos) in southern Nevada. Journal of Herpetology 7:79-85. Meffe, O.K., C.R. Carroll {eds.) and contributors. 1997. Principles of Conservation Biology. 2nd edition. Sinaur Associates. Sunderland, Massachusetts. Mendelson, J.R. Ill, and W.B. Jennings. 1992. Shifts in relative abundance of snakes in a desert grassland. Journal of Herpetology 26:38-45. Meserve, P.L., J.R. Gutierrez, and P.M. Jaksic. 1993. Effects of vertebrate predation on a caviomorph rodent, the degu {Octodon degus), in a semiarid thom scrub community in Chile. Oecologia 94:153-158. Meserve. P.L.. J.R. Gutierrez. J.A. Yunger. L.C. Contreras, and P.M. Jaksic. 1996. Role of biotic interactions in a small mammal assemblage in semiarid Chile. Ecology 77:133-148. Messier, P. 1994. Ungulate population models with predation: a case study with the 294 North American moose. Ecology 75:478-488. Messier, F. 1995. Trophic interactions in two northern wolf-ungulate systems. Wildlife Research 22:131-146. Millsap, B.A. 1981. Distributional status of falconiformes in westcentral Arizona ~ with notes on ecology, reproductive success and management. U.S.D.I. Bureau of Land Management, Technical Note 355:102 pp. Millsap. B.A. 1998. Bam owl, Tyto alba. Pp. 136-139 in R.L. Glinski (eJ.). The Raptors of Arizona. University of Arizona Press, Tucson. Milstead, W.W. {ecL). 1967. Lizard Ecology: .A Symposium. University of Missouri Press. Columbia. Missouri, .xi+300 pp. Milstead, W.W. 1957. Some aspects of competition in natural populations of whiptail lizards (genus Cnemiclophonis). Texas Journal of Science 9:410-445. Milstead. W.W. 1977. The Black Gap whiptail lizards after twenty years. Pp. 523-532 in R.H. Waur and D.H. Riskind {eels.). Transactions of the Symposium on the Biological Resources of the Chihuahuan Desert Region. National Park Service Transactions and Proceedings Series no. 3. Washington. DC. xxii-f-658 pp. Mitchell. R.J. 1993. Path analysis: pollination. Pp. 211-231 in S.M. Scheiner and J. Gurevitch {eds.). Design and Analysis of Ecological Experiments. Chapman and Hall, New York, xiv + 445 pp. Mitchell. W.A., Z. Abramsky, B.P. Kotler. B.P. Pinshow, and J.S. Brown. 1990. The effect of intra- and interspecific competition on foraging effort: theoretical development and a test with granivorous desert rodents in Israel. Ecology 71:844- 854. Mohr. C.O. 1947. Table of equivalent populations of North American small mammals. American Midland Naturalist 37:223-249. Mooney, H.G., and J.A. Drake {editors). 1986. Ecology of Biological Invasions of North America and Hawaii. Springer-Verlag, New York. Morton, R.D., R. Law, S.L. Pimm, and J.A. Drake. 1996. On models for assembling ecological communities. OIKOS 75:493-499. Munger, J.C. and J.H. Brown. 1981. Competition in desert rodents: an experiment with semipermeable exclosures. Science 211:510-512. 295 Naeem. S., L.J. Thompson, S.P. Lawler. J.H. Lawton. and R.M. Woodtln. 1995. Empirical evidence that declining species diversity may alter the performance of terrestrial ecosystems. Philosophical Transactions of the Royal Society of London, B. Biological Science 347:249-262. Naveh, Z. and A.S. Lieberman. 1994. Landscape Ecology: Theory and Application. Second Edition. Springer-Verlag, New York. xxvii4-360 pp. Nielsen, O.K. 1999. Gyrfalcon predation on ptarmigan: numerical and functional responses. Journal of Animal Ecology 68:1034-1050. Norris, K.S. 1967. Color adaptation in desert reptiles and its thermal relationships. Pp. 162-229 in W.W. Milstead {ed.). Lizard Ecology: A Symposium. University of Missouri Press, Columbia, Missouri, v+300 pp. Norris, K.S. and C.H. Lowe. 1964. An analysis of background color-matching in amphibians and reptiles. Ecology 45:565-580. Nowak. R.M. 1991. Walker's Mammals of the World. 5th edition. Johns Hopkins University Press. ORPI (Resources Management personnel). 1995. Climate monitoring protocol for the ecological monitoring program in Organ Pipe Cactus National Monument, Arizona. Section 9: 5 pp. in ORPI Resources Management division {coordinator). Organ Pipe Cactus National Monument Ecological Monitoring Program Monitoring Protocol Manual. Cooperative Park Studies Unit, National Biological Service (U.S.G.S.), University of Arizona, Special Report 11:12 Sections. ORPI. 1995. Organ Pipe Cactus National Monument Ecological Monitoring Program: Monitoring Protocol Manual. USGS Biological Resources Division, Cooperative Park Studies Unit, University of Arizona, Tucson, Arizona, 85721. Special Report No. 11. 307 pages in 12 sections. Ortenburger, A.I. 1928. The Whipsnake and Racers, Genera Masticophis and Coluber. Memoirs of the University of Michigan Museum of Zoology l:,\vii+247 pp. Otis, D.L., K.P. Bumham, G.C. White, and D.R. Anderson. 1978. Statistical inference from capture data on closed animal populations. Wildlife Monographs 62:1-135. Paine, R.T. 1966. Food web complexity and species diversity. American Naturalist 100:65-75. Parizek, D.A., P.C. Rosen, C.R. Schwalbe, and C.H. Lowe. 1996. Ecology of the 296 Mexican rosy boa and the Ajo Mountain whipsnake. Final Report for IIPAM 92051, Arizona Game and Fish Department Heritage Program, Phoenix, AZ. 108 pp. Parker, W.S. 1971. Ecological observations on the regal homed lizard {Phtynosoma solare) in Arizona. Herpetologica 27:333-338. Parker, W.S. 1972. Aspects of the ecology of a Sonoran Desert population of the western banded gecko, Coleonyx variegatiis (Sauria, Eublepharinae). American Midland Naturalist 88:209-224. Parker, W.S. 1973. Natural history notes on the iguanid lizard Urusmirus orncitus. Journal of Herpetology 7:21-26. Parker, W.S. and W.S. Brown. 1973. Species composition and population changes in two complexes of snake hibemacula in northern Utah. Herpetologica 29:319-326. Parker, W.S. and W.S. Brown. 1980. Comparative ecology of two colubrid snakes, Masticopliis t. taeniatus and Pitiiophis melanoleuciis deserticula in northern Utah. Milwaukee Public Museum Publications in Biology and Geology 7:1-104. Parker, W.S. and E.R. Pianka. 1973. Notes on the ecology of the iguanid lizard. Sceloporus inagister. Herpetologica 29:143-152. 1 Parker. W.S. and E.R. Pianka. 1976. Ecological observations on the leopard lizard {Crotaphytus wislizenii) in different parts of its range. Herpetologica 32:95-114. Parmley. D. 1982. Food items of roadrunners from Palo Pinto County, north central Texas. Texas Journal of Science 34:94-95. Paulissen, M.A. 1994. Microhabitat use and escape behaviors of syntopic clonal complexes of the parthenogenetic whiptail lizard Cnemidophorus laredoensis. American Midland Naturalist 132:10-18. Paulissen, M.A. 1995. Comparative escape behavior of parthenogenetic and gonochoristic Cnemidophorus in southern Texas. Copeia 1995:223-226. Perez, J.C., W.C. Haws, and C.H. Haws. 1978. Resistance of woodrats (Neotoma micropiis) to Crotaliis airox venom. Toxicon 16:198-200. Peters, R.H. 1991. A Critique for Ecology. Cambridge University Press, Cambridge. UK. xiv -I- 366 pp. 297 Petryszyn, Y. 1982. Population dynamics of nocturnal desert rodents: a nine year study. Ph.D Dissertation, University of Arizona. Tucson. Petryszyn, Y. 1995. Small nocturnal mammals monitoring protocol for the Ecological Monitoring Program in Organ Pipe Cactus National Monument. Section 5: iii + 45 pp. in ORPI Resource Management Division (coordinator). Organ Pipe Cactus National Monument Ecological Monitoring Program Monitoring Protocol Manual. Cooperative Park Studies Unit, National Biological Service (U.S.G.S.), University of Arizona, Special Report 11:12 Sections.. Petryszyn, Y. and S. Russ. 1996. Nocturnal rodent population densities and distribution at Organ Pipe Cactus National Monument, Arizona. Cooperative Park Studies Unit, National Biological Service (U.S.G.S.), University of Arizona, Technical Report No. 52:1-43. Pianka, E.R. 1986. Ecology and Natural History of Desert Lizards. Princeton University Press, Princeton, New Jersey, xi + 208 pp. Pianka, E.R. and W.S. Parker. 1972. Ecology of the iguanid lizard Callisaurus (Iraconoides. Copeia 1972:493-508. Pierce, B.M., W.S. Longland, and S.H. Jenkins. 1992. Rattlesnake predation on desert rodents: microhabitat and species-specific effects on risk. Journal of Mammalogy 73:859-865. Pimm, S.L. 1982. Food Webs. Chapman and Hall, London. Pimm, S.L. 1991. The Balance of Nature? University of Chicago Press, Chicago. Polis, G.A. 1991. Complex trophic interactions in deserts: an empirical critique of food- web theory. American Naturalist 138:123-155. Polis, G.A. 1998. Stability is woven by complex webs. Nature 395:744-745. Polis, G.A., and K.O. Winemiller. (eds.) 1996. Food Webs: Integration of Patterns and Dynamics. Chapman and Hall, New York. Pollack, K.H., J.D. Nichols, C. Browne, and J.E. Hines. 1990. Statistical inference for capture-recapture experiments. Wildlife Monographs 107:1-97. Pomianowska-Pilipiuk, I. 1974. Energy balance and food requirements of adult vipers Vipera bents (L.). Ekologia Polska 22:195-211. 298 Porter, W.P. and C.R. Tracy. 1974. Modelling the effects of temperature changes on the ecology of the garter snake and leopard frog. Pp. 595-608 in J.W. Gibbons and R.R. Sharitz {eds.). Thermal Ecology AEC Conference 730505. Oak Ridge, Tennessee. Power, M.E., D.Tilman, J.A. Estes, B.A. Menge, W.J. Bond, L.S. Mills, G. Daily, J.C. Castilla, J. Lubchenko, and R.T. Paine. 1996. Challenges in the quest for keystones. Bioscience 46:609-620. Price, A.H., J.L. LaPointe, and J.W. Atmar. 1993. The ecology and evolutionary implications of competition and parthenogenesis in Cnemidophorus. Pp. 371-410 in J.W. Wright and J.J. Vitt (eds.). Biology of the Whiptail Lizards (Gensu Cnemidophorus). Oklahoma Museum of Natural History, Norman, Oklahoma. xiv-)417 pp. Price, M.V. 1978a. The role of microhabitat in structuring desert rodent communities. Ecology 59:910-921. Price. M.V. 1978b. Seed dispersion preferences of coe.xisting desert rodent species. Journal of Mammalogy 59:624-626. Redpath, S.M. and S.J. Thirgood. 1999. Numerical and functional responses in generalist predators: hen harriers and peregrines on Scottish grouse moors. Journal of Animals Ecology 68:879-892. Reichenbach, N.G. and G.H. Dalrymple. 1980. On the criteria and evidence for interspecific competition in snakes. Journal of Herpetology 14:409-412. Reichman, O.J. 1975. Relation of desert rodent diets to available resources. Journal of Mammalogy 56:731-751. Reichman. O.J. 1979. Desert granivore foraging and its impact on seed densities and distributions. Ecology 60:1085-1092. Reichman. O.J. and D. Oberstein. 1977. Selection of seed distribution types by Dipodomys merriami and Perognat/iits ampins. Ecology 58:636-643. Ricklefs. R.E. and G.L. Miller. 2000. Ecology. 4th edition. W.H. Freeman and Co.. New York, xxxvii+822 pp. Ricklefs, R.E. and D. Schluter (,eds.). 1993. Species Diversity in Ecological Communities: Historical and Geographical Perspectives. University of Chicago Press, Chicago, Illinois. 299 Rosen, P.C. 1991. Comparative ecology and life history of the racer. Coluber constrictor, in Michigan. Copeia 1991:897-909. Rosen, P.C. P.A. Holm, and C.H. Lowe. 1995. Ecology and status of shovelnosed snakes {Chionactis) and leafnosed snakes (Phyllorhynchus) at and near Organ Pipe Cactus National Monument, Arizona. Final Report for IIPAM 92045, Arizona Heritage Program, Arizona Game and Fish Dept., Phoeni.x. 65 pp. Rosen, P.C. and C.H. Lowe. 1994. Highway mortality of snakes in the Sonoran Desert of southern Arizona. Biological Conservation 68:143-148. Rosen. P.C. and C.H. Lowe. 1995. Lizard monitoring protocol for the ecological monitoring program in Organ Pipe Cactus National Monument, Arizona. Section 4: V + 49 pp. in ORPI Resources Management division (coordinator). Organ Pipe Cactus National Monument Ecological Monitoring Program Monitoring Protocol Manual. Cooperative Park Studies Unit, National Biological Service (U.S.G.S.), University of Arizona, Special Report 11:12 Sections. Rosen. P.C. and C.H. Lowe. 1996. Ecology of the amphibians and reptiles at Organ Pipe Cactus National Monument. Cooperative Park Studies Unit, National Biological Service (U.S.G.S.), University of Arizona. Technical Report 53:.xii + 136 pp. Rosen. P.C. and C.R. Schwalbe. in press. Widespread effects of introduced species on aquatic reptiles and amphibians in the Sonoran Desert region. Pp. 000-000 in B. Tellman {ed.}. Introduced Species in the Sonoran Desert. University . Arizona Press, Tucson. ooo+OOO pp. Rosenzweig, M.L. 1966. Community structure in sympatric Camivora. Journal of Mammalogy 47:602-612. Rosenzweig, M.L. 1973. Habitat selection e.xperiments with a pair of coexisting heteromyid rodent species. Ecology 54:111-117. Rosenzweig, M.L. 1995. Species Diversity in Space and Time. Cambridge University Press, Cambridge. Great Britain. Rosenzweig, M.L. and Z. Abramsky. 1997. Two gerbils of the Negev: a long-term investigation of optimal habitat selection and its consequences. Evolutionary Ecology 11:733-756. Rosenzweig, M.L. and P.W. Sterner. 1970. Population ecology of desert rodent communities: body size and seed-husking as bases for heteromyid coexistence. Ecology 51:217-224. 300 Rosenzweig, M.L. and J. Winakur. 1969. Population ecology of desert rodent communities: habitats and environmental complexity. Ecology 50(3):558-572. Ross, A. 1969. Ecological aspects of the food habits of insectivorous screech-owls. Proceedings of the Western Foundation of Vertebrate Zoology 1:301-344. Saitoh, T., O.N Bjomstad. and N.C. Stenseth. 2000. Density dependence in voles and mice: a comparative study. Ecology 80:638-650. SAS Institute Inc. 1989. SAS/STAT User's Guide, Version 6. 4th edition, volume 1. SAS Institute Inc.. Cary, North Carolina. 943 pp. Savidge, J.A. 1987. E.xtinction of an island forest avifauna by an introduced snake. Ecology 68:660-668. Schall. J.J. and E.R. Pianka. 1980. Evolution of escape behavior diversity. .American Naturalist 115:551-556. Scheiner, S.M and J. Gurevitch {eds.). 1993. Design and .Analysis of Ecological E.xperiments. Chapman and Hall. New York. Schleich. H.H., W. Kastle, and K. Kabisch. 1996. Amphibians and Reptiles of North Africa. Koeltz Scientific Books. Koenigstein, Germany, ii+629 pp. Schluter, D. 1990. Species-for-species matching. American Naturalist 136:560-568. Schmidly, D.J. 1977. The Mammals of Trans-Pecos Texas. Texas A & M University Press. College Station. Texas. Schoener. T.W. 1983. Field experiments on interspecific competition. American Naturalist 122:240-285. Schroder. G.D. and M.L. Rosenzweig. 1975. Perturbation analysis of competition and overlap in habitat utilization between Dipodomys ordii and Dipodomys merriami. Oecologia 19:9-28. Scott, T.A. and M.L. Morrison. 1995. Opportunistic foraging behavior of loggerhead shrikes. Pp. 188-193 in Y. Reuven and F.E. Lohrer (e-rfi.). Shrikes (Laniidae) of the World: Biology and Conservation. Proceedings of the Western Foundation of Vertebrate Zoology 6:1-343. Scott, T.G. and W.D. Klimstra. 1955. Red foxes and a declining prey population. Southern Illinois University Monograph Series 1:123 pp. 301 Seber. G.A.F. 1970. The effects of trap-response on tag-recapture estimates. Biometrika 26:13-22. Secor. S.M. 1994. Natural history of the sidewinder, Crotcilus cerastes. Pp. 281-301 in P.R. Brown and J.W. Wright (eds.), Herpetology of the North American Deserts: Proceedings of a Symposium. Special Publication 5, Southwestern Herpetologists Society, Van Nuys, California, iv-f-311 pp. Secor. S.M. 1995. Ecological aspects of foraging mode for the snakes Crotalus cerastes and Masticophis flagelliim. Herpetological Monographs 9:169-186. Secor. S.M. and J.M. Diamond. 1994. Energetic correlates of foraging mode for the snakes Crotalus cerastes and Masticophis flagelliim. Ecology 75:1600-1614. Seigel, R.A. 1984. The foraging ecology and re.source partitioning pattern of two species of garter snakes. Unpublished Ph.D. di.ssertation. Univ. Kansas. Lawrence. Seigel. R.A.. J.W. Gibbons, and T.K. Lynch. 1995. Temporal changes in reptile populations: effects of a severe drought on aquatic snakes. Herpetologica 51:424- 434. Sellers, W.D. and R.H. Hill (editors). 1974. Arizona Climate, 1931-1972. University of Arizona Press, Tucson. Shaw. C.E. and S. Campbell. 1974. Snakes of the American West. Alfred A. Knopf, New York, xii-i-332 pp. Shreve. F. 1951. Vegetation and flora of the Sonoran Desert. Vol. 1. Vegetation. Carnegie Inst. Washington, Publ. 591:1-192. Shreve, F. and I.L. Wiggins. 1964. Vegetation and Flora of the Sonoran Desert, Volumes 1 and 2. Stanford University Press, Stanford, California. Sih, A., P. Crowley. M. Mcpeek, J. Petranka, and K. Strohmeier. 1985. Predation, competition, and prey communities: a review of field experiments. Annual Review of Ecology and Systematics 16:269-3 H. Sinclair, A.R.E. 1989. Population regulation in animals. Pp. 197-241 in J.M. Cherritt (ed.). Ecological Concepts. Symposium of the British Ecological Society No. 28, Blackwell Scientific Publications, Oxford, Great Britain. Sinclair, A.R.E., J.M. Gisline, G. Holdsworth, C.J. Krebs, S. Boutin, J.N.M. Smith, R. Boonstra, and M. Dale. 1993. Can the solar cycle and climate synchronize the 302 snovvshoe hare cycle in Canada? Evidence of tree rings and ice cores. American Naturalist 141:173-198. Sinclair, A.R.E., C.J. Krebs, J.N.M. Smith, and S. Boutin. 1988. Population biology of snowshoe hares. III. Nutrition, secondary compounds, and food limitation. Journal of Animal Ecology 57:787-806. Small. R.L. 1971. Interspecific competition among three species of camivora on the Spider Ranch. Yavapai Co.. Arizona. M.S. thesis. University of Arizona. Tucson. Smith, D.C. 1981. Competitive interactions of the striped plateau lizard {Sceloporiis virgatus) and the tree lizard {Urosauriis onicmis). Ecology 62:679-687. Smith. D.D., P.A. Medica, and S.R. Sanbom. 1987. Ecological comparison of sympatric populations of sand lizards (Copliosaurus texamis) and (Caltisciuriis draconoides). Great Basin Naturalist 47:175-185. Smith. D.G. 1969. Nesting ecology of the great homed owl. Bubo virginianus. Brigham Young University Science Bulletin Biological Series 10:16-25. Smith. H.D. and C.D. Jorgensen. 1975. Reproductive biology of North American desert rodents. Pp. 305-330 in I. Prakash and P.K. Ghosh (eds.). Rodents in Desert Environments. Monographie Biologicae. Vol. 28. Dr. W. Junk b.v.. Publishers. The Hague, Netherlands. Smith, N.S. and D.A. Danner. 1990. Gray fox diets relative to presence of coyotes and cattle grazing. Pp. 144-150 in P.R. Krausman and N.S. Smith (eds.). Managing Wildlife in the Southwest: Proceedings of the Symposium. Arizona Chapter of the Wildlife Society. P.O. Box 41337, Phoenix. AZ 85080-1337. Smith. R.L. 1996. Ecology and Field Biology. 5"* Edition. Addison-Wesley Educational Publishers, Menlo Park, California, xix+740 pp. +appendices. Sokal, R.R., and F.J. Rohlf. 1981. Biometry. The Principles and Practice of Statistics in Biological Research. W.H. Freeman, New York. Sowls, L.K. 1957. Reproduction in the Audubon cottontail in Arizona. Journal of Mammalogy 38:234-243. Stamp, N.E. and R.D. Ohmart. 1978. Resource utilization by desert rodents in the Lower Sonoran Desert. Ecology 59(4):700-707. Stephens, D.W. and C.J. Krebs. 1986. Foraging Theory. Princeton University Press, 303 Princeton. New Jersey. Sterner, R.W., A. Bajpai, and T. Adams. 1997. The enigma of food chain length: absence of theoretical evidence for dynamic constraints. Ecology 78:2258-2262. Strong, D.R. Jr., D. Simberloff, L.G. Abele, and A.B. Thistle {eds.). 1984. Ecological Communities: Conceptual Issues and the Evidence. Princeton University Press, Princeton, New Jersey. Swanson, B.J. 1998. Autocorrelated rates of change in animal populations and their relationship to precipitation. Conservation Biology 12:801-808. Tanaka, R. 1980. Controversial problems in advanced research on estimating population densities of small rodents. Researches on Population Ecology (Japan) Suppl. 2:1- 67. Tanner, W.W. and J.E. Krogh. 1973. Ecology of Phrynosoma platyrhinos at the Nevada Test site, Nye County. Nevada. Herpetologica 29:327-342. Thompson. P. and B.J. Fox. 1993. Asymmetric competition in Australian heathland rodents: a reciprocal removal experiment demonstrating the influence of size-class structure. OIKOS 67:264-278. Thompson, S.D. 1982a. Microhabitat utilization and foraging behavior of bipedal and quadrupedal heteromyid rodents. Ecology 63(5): 1303-1312. Thompson, S.D. 1982b. Structure and .species composition of desert heteromyid rodent species assemblages: effects of a simple habitat manipulation. Ecology 63(5):1313-1321. Thompson, W.L.. G.C. White., and C. Gowan. 1998. Monitoring Vertebrate Populations. Academic Press. San Diego, California. Tinkle, D.W. 1962. Reproductive potential and cycles in female Crutalus citrox from northwestern Texas. Copeia 1962:306-313. Tinkle. D.W. 1982. Results of experimental density manipulation in an Arizona lizard community. Ecology 63:57-65. Tinkle. D.W. and A.E. Dunham. 1986. Comparative life histories of two syntopic sceloporine lizards. Copeia 1986:1-18. Tinkle, D.W., A.E. Dunham, and J.D. Congdon. 1993. Life history and demographic 304 variation in the lizard Sceloporus graciosus: a long-term study. Ecology 74:2413- 2429. Tinkle. D.W. and N.F. Hadley. 1975. Lizard reproductive effort: caloric estimates and comments on its evolution. Ecology 43:223-229. Trapp. G.R. 1978. Comparative behavioral ecology of the ringtail and gray fox in southwestern Utah. Carnivore 1:3-32. Turkovvski, F.J. 1969. Food habits and behavior of the gray fo.x. Urocyon cinereoargemeiis the Lower and Upper Sonoran life zones of the southwestern United States. Ph.D. thesis, Arizona State University, Tempe. Turner, F.B. 1977. The dynamics of populations of squamates, crocodilians. and rhynchocephalians. Pp. 157-264 in C. Cans and D.W. Tinkle (eds.). Biology of the Reptilia. Vol. 7. Academic Press, New York. Turner, F.B., G.A. Hoddenbach, P.A. Medica, and J.R. Lannom Jr. 1970. The demography of the lizard, Uta stansbiirianci Baird and Girard, in southern Nevada. Journal of Animal Ecology 39:505-519. Turner. F.B., J.R. Lannom Jr.. P.A. Medica, and G.A. Hoddenbach. 1969b. Density and composition of fenced populations of leopard lizards (Crotapliytus wisUzenii) in southern Nevada. Herpetologica 25:247-257. Turner. F.B., P.A. Medica. K.W. Bridges, and R.L Jennrich. 1982. A population model of the lizard Uta stansburiana in Southern Nevada. Ecological Monographs 52:243-259. Turner, F.B.. P.A. Medica, J.R. Lannom Jr.. and G.A. Hoddenbach. 1969. A demographic analysis of fenced populations of the whiptail lizard. Cnemidophoriis tigris, in southern Nevada. Southwestern Naturalist 14:189-202. Valone, T.J. 1995. Catastrophic decline of a desert rodent, Dipodumys spectabilb: insights from a long-term study. Journal of Mammalogy 76:428-436. Van Ballenberghe, V. and W.B. Ballard. 1994. Limitation and regulation of moose populations: the role of predation. Canadian Journal of Zoology 72:2071-2077. Van Loben Sels, R.C. and L.J. Vitt. 1984. Desert lizard reproduction: annual and seasonal variation in Urosaurus omatiis (Iguanidae). Canadian Journal of Zoology 62:1779-1787. 305 Vial, J.L. 1962. The auditory bulla of Dipodomys deserti (Rodentia) and evidence of its adaptive significance. Revistade Biologia Tropical 10:11-18. Vitt, L.J. 1981. Lizard reproduction: habitat specificity and constraints on relative clutch mass. American Naturalist 117:506-514. Vitt, L.J. 1987. Communities. Pp. 335-395 in R.A. Seigel, J.T. Collins, and S.S. Novak {eds.). Snakes: Ecology and Evolutionary Biology. Macmillan. New York. .\iv-i-529 pp. Vitt, L.J. 1991. An introduction to the ecology of Cerrado lizards. Journal of Herpetology 25:79-90. Vitt. L.J. 1991. Desert reptile communities. Pp. 249-277 in G.A. Polls {ed.). The Ecology of De.sert Communities. University of Arizona Press. Tucson, Arizona, ix + 457 pp. Vitt, L.J. 1995. The ecology of tropical lizards in the Caatinga of northeast Brazil. Occasional Papers of the Oklahoma Museum of Natural History 1:1-29. Vitt, L.J. and C.M. Carvalho. 1995. Niche partitioning in a tropical wet season: lizards in the Lavrado area of northern Brazil. Copeia 1995:305-329. Vitt, L.J. and J.D. Congdon. 1978. Body shape, reproductive effort, and relative clutch mass in lizards: resolution of a paradox. American Naturalist 112:595-608. Vitt. L.J. and R.D. Ohmart. 1974. Reproduction and ecology of a Colorado River population of Sceloporus magisier (Sauria: Iguanidae). Herpetologica 30:410-417. Vitt. L.J. and R.D. Ohmart. 1977a. Ecology and reproduction of lower Colorado River lizards: 1. Callisaiinis draconoides (Iguanidae). Herpetologica 33:214-222. Vitt, L.J. and R.D. Ohmart. 1977b. Ecology and reproduction of lower Colorado River lizards: I. Cnemidophonis tigris (Teiidae), with comparisons. Herpetologica 33:223-234. Vitt. L.J. and H.J. Price. 1982. Ecological and evolutionary determinants of relative clutch mass in lizards. Heqjetologica 38:237-255. Vitt.. L.J., R.C. Van Loben Sels. and R.D. Ohmart. 1981. Ecological relationships among arboreal desert lizards. Ecology 62:398-410. Vitt, L.J. and P.A. Zani. 1996. Organization of a taxonomically diverse lizard 306 assemblage in Amazonian Ecuador. Canadian Journal of Zoology 74:1313-1335. Vitt. L.J. and P.A. Zani. 1998a. Ecological relationships among sympatric lizards in a transitional forest in the northern Amazon of Brazil. Journal of Tropical Ecology 14:63-86. Vitt, L.J. and P.A. Zani. 1998b. Prey use among sympatric lizard species in lowland rain forest of Nicaragua. Journal of Tropical Ecology 14:537-559. Vogt. R.C. and R.L. Hine. 1982. Evaluation of techniques for assessment of amphibian and reptiles populations in Wisconsin. Pp. 201-217 in N.J. Scott Jr. (ff/.), Herpetological Communities: A Symposium. U.S. Fish & Wildlife Service Research Report 13, Washington, D.C. Warren, P.L.. B.K. Mortenson, B.D. Treadwell, J.E. Bowers, and K.L. Reichardt. 1981. Vegetation of Organ Pipe Cactus National Monument. Cooperative Park Studies Unit (USGS), University of Arizona, Technical Report 8:1-79. Wasser. P.M. and W.T. Jones. 1991. Survival and reproductive effort in banner-tailed kangaroo rats. Ecology:72:771-777. Whaley, W.H. 1986. Population ecology of the Harris' hawk in .Arizona. (Journal of) Raptor Research 20:1-15. White. G.C. 1994. User's manual for program MCAPTURE. Unpublished report to USDA Forest Service. Whitford. W.G. 1976. Temporal fluctuations in density and diversity of desen rodent populations. Journal of Mammalogy 57:351-369. Whitford W.G., and F.M. Creusere. 1977. Seasonal and yearly fluctuations in Chihuahuan Desert lizard communities. Herpetologica 33:54-65. Whittaker, R.H. 1975. Communities and Ecosystems. 2nd edition. Macmillan. New York. Whittaker, R.H. and W.A. Niering. 1965. Vegetation of the Santa Catalina Mountains. Arizona: a gradient analysis of the south slope. Ecology 46:429-452. Wilbur, H.M. 1972. Competition, predation, and the structure of the Ambystoma-Rana sylvatica community. Ecology 53:3-21. Wilbur, H.M. 1997. Experimental ecology of food webs: complex systems in temporary 307 ponds. Ecology 78:2279-2302. Wilson, D.E., F.R. Cole, J.D. Nicholds, R. Rudran, and M.S. Foster. 1996. Measuring and Monitoring Biological Diversity: Standard Methods for Mammals. Smithsonian Institution Press, Washington, DC. Wilson. K.R. and D.R. Anderson. 1985. Evaluation of a density estimator based on a trapping web and distance sampling theory. Ecology 65:1185-1194. Wondolleck, J.T. 1978. Forage-area separation and overlap in heteromyid rodents. Journal of Mammalogy 59:510-518. Woodbury, A.M. 1951. A snake den in Toole County, Utah: introduction—a ten year study. Herpetologica 7:4-14. Yang, T.W. and C.H. Lowe. 1956. Correlation of major vegetation climaxes with soil characteristics in the Sonoran Desert. Science 123:542. Zeng, Z. and J.H. Brown. 1987. Population ecology of a desert rodent: Dipodomys merriami in the Chihuahuan Desert. Ecology 1328-1340. Ziv, Y., B.P. Kotler, Z. Abramsky, and M.L. Rosenzweig. 1995. Foraging efficiencies of competing rodents: why do gerbils exhibit shared-preference habitat selection? OIKOS 73:260-268.