Habitat use by fishes of the San Bernardino National Wildlife Refuge

Item Type text; Thesis-Reproduction (electronic)

Authors Maes, Ronnie Andrew

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 23/09/2021 19:46:53

Link to Item http://hdl.handle.net/10150/291514 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 corner and continuing from left to right in equal sections with small overlaps. Each original is also photographed in one exposure and is included in reduced form at the back of the book.

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. UMI A Bell & Howell Information Company 300 North Zeeb Road, Ann Arbor MI 48106-1346 USA 313/761-4700 800/521-0600

HABITAT USE BY FISHES OF THE

SAN BERNARDINO NATIONAL WILDLIFE REFUGE

by

Ronnie Andrew Maes

A Thesis Submitted to the Faculty of the

SCHOOL OF RENEWABLE NATURAL RESOURCES

In Partial fulfillment of the Requirements For the Degree of

MASTER OF SCIENCE WITH A MAJOR IN WILDLIFE AND FISHERIES SCIENCE

In the Graduate College

THE UNIVERSITY OF ARIZONA

19 9 5 UMI Number'. 1378304

UMI Microform 1378304 Copyright 1996, by UMI Company. All rights reserved.

This microform edition is protected against unauthorized copying under Title 17, United States Code.

UMI 300 North Zeeb Road Ann Arbor, MI 48103 2

STATEMENT BY THE AUTHOR

This thesis 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 thesis are allowable without special permission, provided that accurate acknowledgment of source is made. Requests for permission for extended quotation from or reproduction of this manuscript in whole or in part may be granted by the head of the major department or the Dean of the Graduate College when in his or her judgment the proposed use of the material is in the interests of scholarship. In all other instances, however, permission must be obtained from the author.

SIGNED K- VA,

APPROVAL BY THESIS DIRECTOR

This thesis has been approved on the date shown below:

S O. Eugene Maughan, Thesis Dir&ctor Date Professor of Wildlife and Fisheries Science , ii.y William J. Matter Date Associate Professor of Wildlife and Fisheries Science

C - Carole Mclvor Date Associate Professor of Wildlife and Fisheries Science 3

ACKNOWLEDGEMENTS

This study was made possible by the help of many people. Funding was provided

by the San Bernardino National Wildlife Refuge, USFWS, Douglas Arizona. I would

like to thank Kevin Cobble, refuge manager, for his cooperation and information on the

refuge.

Special thanks go to Dr. O. Eugene Maughan for his guidance and support

throughout this project and allowing me the opportunity to pursue this Master's degree. I

would also like to thank Dr. William Matter and Dr. Carole Mclvor for their helpful

comments and suggestions while serving on my committee.

The time 1 spent at the University of Arizona was an enjoyable and memorable

one. I owe the good memories to Cliff Schleusner, Rob Dudley, Cynthia Martinez,

Leslie Ruiz, Trisha Roller, and Craig Springer.

Many people assisted in the field: Cliff Schleusner, Armando Ortiz, Leslie Ruiz,

David Kitcheyan, Trisha Roller, Rob Dudley, and my wife Janalyn. Without your help the work would have been impossible and unbearable. I want to thank Sarah Fritz, Mary

Soltero, Marge O'Connell, and Carole Wakely for their help with the administrative hurdles and Hatsumi for the many cups of hot coffee.

Lastly, I would like to thank my wife for her support, help in the field, patience, and especially her forbearance during stressful times, and my parents for their love and support during my lengthy college career. DEDICATION

This thesis is dedicated to the memory of my loving sister, Michelle Annette

Maes. I only wish that you had been here to see me through the completion of this endeavor. 5

TABLE OF CONTENTS

LIST OF TABLES 7

LIST OF FIGURES 8

ABSTRACT 10

INTRODUCTION 11

Species Present 14

Objectives 19

Study Area 22

METHODS 24

Habitat Use 24

Habitat Availability 29

Operational Definitions 30

Data Analysis 31

RESULTS 34

Habitat Use .. ., 34

Topminnow 37

Beautiful Shiner 46

Yaqui Chub 51

DISCUSSION 64

Biotic Interactions 74

Applications 76

APPENDIX A 80 TABLE OF CONTEmS-Continued

LIST OF REFERENCES 7

LIST OF TABLES

Table 1. Fishes of the San Bernardino Valley in southeastern Arizona 15

Table 2. Description of sites studied on the San Bernardino National Wildlife

Refuge 24

Table 3. Modified Wentworth particle size scale 27

Table 4. Cover types used to define cover used by fish 28

Table 5. Size categories of schools 33

Table 6. Sample sizes of species by lifestage in three ponds on the San

Bernardino National Wildlife Refuge, Arizona 35 8

LIST OF FIGURES

Figure 1. Main portion of the San Bernardino National Wildlife Refuge,

southeastern Arizona 12

Figure 2. Leslie Creek Canyon in the southern extension of the Swisshelm Mtns.,

San Bernardino National Wildlife Refuge, southeastern Arizona 26

Figure 3. Distribution of available substrates for North Pond versus Oasis Pond,

SBNWR, Arizona 36

Figure 4. Distribution of relative depths used by Yaqui topminnows in Tule

Pond 41

Figure 5. Substrates used versus available to topminnows in Tule, North and

Oasis Ponds 42

Figure 6. Cover types used by topminnows in Tule, North, and Oasis Ponds 43

Figure 7. Relative depths used by Yaqui topminnows in Oasis and North Ponds. ... 44

Figure 8. Depths used by topminnows in North and Oasis Ponds 45

Figure 9. Relative depths used by beautiful shiners in Oasis Pond 47

Figure 10. Depths used by beautiful shiners in Oasis Pond 48

Figure 11. Substrates used by beautiful shiners in Oasis Pond 49

Figure 12. Cover types used by beautiful shiners in Oasis Pond 50

Figure 13. Relative depths used by young-of-year chubs (YOY) at all sites 54

Figure 14. Relative depths used by Yaqui chubs (excluding YOY) at all sites 55

Figure 15. Depths used by Yaqui chubs at all sites 56

Figure 16. Substrates used by Yaqui chubs in North and Oasis Ponds 57 9

LIST OF FIGURES-Conrtwwerf

Figure 17. Substrates used by Yaqui chubs in Leslie Creek 58

Figure 18. Cover types used by different life stages of Yaqui chubs in North

Pond 59

Figure 19. Cover types used by Yaqui chubs in Oasis Pond 60

Figure 20. Cover types used by different life stages of Yaqui chubs in Leslie

Creek 61

Figure 21. Mean water column velocity for Yaqui chub observations in Leslie

Creek 62

Figure 22. Nose velocity for Yaqui chubs in Leslie Creek 63

Figure 23. Distribution of Yaqui topminnow, beautiful shiner, and Yaqui chub

in a cross section of a pool 73 10

ABSTRACT

I quantified microhabitat conditions used by Yaqui chub (Gila purpurea). Yaqui topminnow (Poeciliopsis occidentalis sonoriensis). and beautiful shiner (Cyprinella lutrensis) on the San Bernardino National Wildlife Refuge, Arizona. Different species and different lifestages used different microhabitats. Smaller fish selected shallower water than adults. Yaqui topminnow and Yaqui chub showed seasonal variation in microhabitats used.

Yaqui topminnows were found closer to cover when in the presence of beautiful shiners. Close proximity to cover may indicate a negative interaction. Yaqui chubs did not use microhabitats differently when in the presence of the other two species.

Microhabitats used by Yaqui chubs in the ponds and Leslie Creek did not differ. Yaqui chub preferred pools with little or no flow.

Management of aquatic environments on the refuge should focus on vegetative thinning. Stocking of beautiful shiner with Yaqui topminnow should be postponed until further research is conducted on the interactions between the two species. 11

INTRODUCTION

San Bernardino National Wildlife Refuge (SBNWR) is located in the San

Bernardino Valley of southeastern Arizona. The San Bernardino Valley is a part of

the Rio Yaqui Basin of Sonora and Chihuahua, Mexico (Figure 1.) It contains a

unique assemblage of desert aquatic habitats which, in the past, has been home to

nearly one quarter of the native fish fauna of Arizona (Williams 1991). The refuge

consists of two parts. The main portion is located along the Mexican border 25 km

east of Douglas, Arizona. It contains the intermittent stream, Black Draw, artesian

ponds, cienegas, and natural springs. A smaller, more recently acquired portion is in the southern extension of the Swisshelm Mountains, 25 km north of Douglas, and includes the perennial section of Leslie Creek.

Prior to the 19th Century, a perennial stream flowed through the San

Bernardino Valley and numerous springs emerged in the area (USFWS 1987). These areas supported a unique fauna. By the mid 1960's most of these aquatic environments were destroyed or altered and only remnant fish populations remained

(Williams 1991). Habitat changes resulted in the extirpation of most of the Rio

Yaqui fishes from the United States (Minckley 1973, McNatt 1974, USDI 1984,

USFWS 1987, Williams 1991). Intensive pumping of the underground aquifer and diversion of the streams and springs to irrigate agricultural fields along Black Draw may have modified aquatic habitats and reduced spring flows (USDI 1984, USFWS

1987). Vegetation removal caused by farming, cattle grazing and trampling, Figure 1. Main portion of the San Bernardino National Wildlife Refuge, southeastern Arizona.

I———I f

North Pond

Oasis Pond Tule Pond

House Pond —i

Mexico 13

and altered water flows probably resulted in extensive erosion throughout the region

(Hastings 1959, Hastings and Turner 1966, McNatt 1974).

Non-native fish such as mosquito fish (Gambusia affinis), largemouth bass

(Micropterus salmoides). black crappie (Pomoxis nii>romaculatus'). black bullhead

Clctalurus melas). and bluegill (Lepomis macrochirus) may have once posed a threat to

the Rio Yaqui fishes in the United States. Some may still be a problem since they

exist on or near the refuge. For example, the mosquito fish is present in Black Draw

and could be inadvertently introduced to other locations on the refuge. As a result, the refuge is currently closed to the public in an attempt to prevent transport of fish from one location to another. Additionally, the non-native fish species present in the waters of Mexico may constitute a threat to native fish populations (Hendrickson et al. 1980,

USDI 1984).

The refuge contains a variety of terrestrial environments dominated by

Chihuahuan desert scrub. However, the area along Leslie Creek supports a lush riparian zone. A mesquite bosque, abandoned croplands, and riparian vegetation occur along Black Draw on the main part of the refuge (USFWS 1987). Marshy areas caused by spring and pond overflows are being managed and expanded on the refuge.

The majority of aquatic habitats that now exist on the refuge are the result of man- made ponds maintained by clear water springs and artesian wells. The ponds are maintained (e.g., vegetation thinning) to offer refugia for Rio Yaqui fishes. The degree of artificiality of aquatic habitats on the refuge is of concern, but such refugia 14

are necessary until more secure natural habitats are available (Minckley and Deacon

1991).

The SBNWR was established because the area provided the best opportunity in

the United States for protecting and recovering the fishes and habitat of the Rio Yaqui

(USFWS 1987). There was a dependable, inexpensive source (little or no pumping

required) of high quality water in subterranean artesian reservoirs able to maintain

populations during drought. Land and water rights were purchased by the Nature

Conservancy in 1979, and were acquired by the U.S. Fish and Wildlife Service

(USFWS) in 1982 (USFWS 1987).

Species Present

The refuge was established to provide a refugium for six species of Rio Yaqui fishes that are listed as endangered or threatened, or proposed for listing as endangered or threatened (Table 1). These fishes are the Yaqui chub (Gila purpurea). Yaqui topminnow (Poecilionsis occidentalis sonoriensis), beautiful shiner (Cvorinella formosa), Yaqui catfish (Ictalunis pricei), Yaqui sucker (Catostomus bernardini). and the Mexican stoneroller (Camnostoma ornatum). The last three species do not currently occur on the refuge, but there are plans to reintroduce them (USFWS 1987).

The Yaqui form of the longfin dace (Agosia chrvsogaster) is also present on the refuge. Roundtail chub (Gila robusta) also may have been present historically in refuge waters. They are present just south of the international boundary of the United

States (USFWS 1987). These two species will be managed on the refuge on an incidental basis (USFWS 1987). In addition to the fishes of the Rio Yaqui, the refuge

will also be managed to maintain the rare San Bernardino snail (Fontelicella sp.).

Table 1. Fishes of the San Bernardino Valley in southeastern Arizona.

Common Scientific Present on Federal Name Name Refuge Listing

Yaqui topminnow Poecilioosis occidentalis Yes Endangered sonoriensis

Yaqui chub Gila purpurea Yes Endangered

Yaqui beautiful CvDrinella formosa Yes Threatened shiner

Longfin dace Aeosia chrvsoeaster Yes (Leslie Creek) Not Listed (Yaqui form)

Mexican stoneroller CamDostoma ornatum No (Rucker Canyon) Candidate

Yaqui catfish Ictalurus pricei No Threatened

Yaqui sucker Catostomus bernardini No Candidate

Roundtail chub Gila robusta No Not Listed (Yaqui form)

The Yaqui chub is listed as endangered by the USFWS (USDI 1984). This species is of special concern because populations (Rios Sonora, Matape, and Yaqui basins) in Sonora, Mexico (Williams et al. 1989) recently were described as a new species, Gila eremica (Demarias 1991). This redefinition of species makes Gila purpurea even more rare. Historically Yaqui chub were extirpated from the main 16

refuge, but 200 individuals were moved from Astin Spring to Leslie Creek prior to

their extirpation (Minckley 1973). Fish from Leslie Creek were later removed to

Dexter National Fish Hatchery (NFH) for future reintroduction into historic ranges

(USFWS 1987). Fish reintroduced from this source now inhabit Leslie Creek, Black

Draw, North Pond, Twin Pond, House Pond, Cottonwood Pond, and Robertson Pond

(Personal Communication, Kevin Cobble, Refuge Manager, SBNWR).

Yaqui chub occupy deep pools of small streams near undercut banks or debris,

and are often found in association with higher aquatic plants (Minckley 1973, 1980,

McNatt 1974). They also occupy permanent, deep, undercut zones adjacent to cliffs,

boulders, or roots (Hendrickson et al. 1980) and pools of creeks, marshes, and other

quiet water often associated with dense aquatic vegetation (Rinne and Minckley 1991).

Their diets consist of algae, terrestrial insects, and arachnids near stream sources

(Minckley 1973); and aquatic insects and small fishes in small streams (Minckley

1980).

The Yaqui topminnow is a small livebearer listed as endangered by the

USFWS (Minckley 1973, USFWS 1983). It is abundant and widely distributed

throughout the Rio Yaqui basin in Mexico (Hendrickson et al. 1980, Williams et al.

1989), but the expanding range of the exotic mosquito fish may threaten existing

populations (Hendrickson et al. 1980). The topminnow currently occupies all permanent waters on the refuge (Personal Communication, Kevin Cobble, Refuge «

Manager, SBNWR). The biology of the Yaqui topminnow may be similar to that of the Gila topminnow. Gila topminnow can be found in shallow, warm, quiet waters of 17

springs, cienegas, marshes, streams (Meffe 1983), and the edges of large rivers (Meffe

and Snelson 1989). They also have been reported to occur in intermediate-sized riffle

and pool habitats (Silvey 1975); small, intermittent, hot, low-desert streams

(Hendrickson et al. 1980); and near the water surface in shallow water near aquatic

vegetation (Rinne and Minckley 1991). Foods include detritus, algae, and aquatic

invertebrates (Minckley 1973, Rinne and Minckley 1991).

The beautiful shiner is federally listed as threatened (USDI 1984). It was

extirpated from the United States (Minckley 1973, USDI 1984), but it has been

reintroduced to the refuge from stock maintained at Dexter NFH. The Dexter

population originated in Mexico (USFWS 1990). Beautiful shiner currently occupy

Twin and Robertson ponds on the refuge. They are also widespread but not generally

abundant over most of the Rio Yaqui in Mexico (Hendrickson et al. 1980). Rinne and

Minckley (1991) reported some locations in Mexico with abundant populations. The

exotic red shiner (Cvprinella lutrensis) may threaten populations in Mexico by genetic swamping or through interspecific competition (USDI 1984). The beautiful shiner occupies riffles of small streams, midwater areas of pools and runs, shorelines in large rivers (Hendrickson et al. 1980, USDI 1984, Rinne and Minckley 1991), and ponds

(McNatt 1974, Hendrickson et al. 1980). They feed on drifting aquatic and terrestrial insects (Rinne and Minckley 1991).

The longfin dace is widespread throughout the Gila, Bill Williams, and Rio

Yaqui Basins (Hendrickson et al. 1980, Sublette et al. 1990, Rinne and Minckley

1991). This species currently inhabits Black Draw and Leslie Creek on the refuge. Its 18

habitat ranges from clear, cool mountain brooks to small, intermittent desert streams

with sand or gravel substrates (Sublette et al. 1990). During periods of low water,

individuals may take refuge under moist algal mats and organic debris until suitable

conditions return (Minckley and Barber 1971, Sublette et al. 1990). The adults of this

species feed principally on detritus but also take zooplankton, aquatic insects, and

filamentous algae (Sublette et al. 1990).

The threatened Yaqui catfish currently does not occur on the refuge (USFWS

1987). However, it was originally described from a location near the current Mexico-

United States boundary in San Bernardino Creek (Rutter 1896) and is believed to have

once occurred on the refuge (USFWS 1987). This species is common in the Rio

Yaqui basin in Mexico (Hendrickson et al. 1980), but genetic swamping and

competition from exotic channel catfish (Ictalurus punctatus). may constitute a threat

(USDI 1984, USFWS 1987, Kelsh and Baca 1991).

The Yaqui sucker (a species being considered for listing) was extirpated from

the United States (Minckley 1973), but the species is still widespread throughout the

Rio Yaqui basin in Mexico (Hendrickson et al. 1980). There are plans to reintroduce

both the Yaqui sucker and Yaqui catfish to the refuge (USFWS 1987).

The Mexican stoneroller is a candidate for the endangered species list

(Williams et al. 1989). The species was extirpated from Leslie Creek and from other

streams (Whitewater Draw west of Douglas) before the refuge was formed (McNatt

1974). Mexican stonerollers are believed to have once occurred on the main portion of the refuge because they are still abundant just south of the border in San 19

Bernardino Creek (Hendrickson et al. 1980). A small population also exists in Rucker

Creek in the Chiricahua Mountains north of Leslie Creek (Minckley 1973, McNatt

1974). There are plans to reintroduce the stoneroller to the refuge, but the lack of

appropriate stream habitat could limit success (USFWS 1987).

Objectives

Currently, little is known of the life histories and habitat requirements of the

six Rio Yaqui species that occur or may soon occur on the refuge. Most management

decisions are based on extrapolations of data on closely related species (USFWS

1987). Today, the majority of aquatic environments on the refuge consist of artificial

ponds and short stretches of stream. Ponds on the refuge are isolated from one

another and are closed to immigration and emigration. There is little or no protection

for these fishes in Mexico. Thus, these ponds and streams may become the last refiigia for the species. The limited numbers of these fishes and the need to maintain

healthy, stable populations make it necessary to understand habitat requirements.

Fishes require certain conditions and resources in order to successfully complete their life cycle. Shortages, or superabundance of these requisites often lead to changes in physiology, behavior, foraging strategy or habitat use. Fishes show

"preferences" for certain conditions within environments they occupy. These preferences probably are linked to survival and reproduction and are often referred to as limiting factors in fish management. 20

We have historically managed game fish species by identifying limiting habitat

factors and then manipulating those factors to increase the production of that species.

However, we have not generally used the same approach to manage native non-game

fish. The reason for the difference in management approach is that in native non-

game fish management, we are generally seeking to maintain or restore a complex

interactive system, rather than increase production of a single species. Therefore, we

have emphasized the protection of intact ecosystems in managing native non-game

fish, rather than the habitat manipulation approach we have used to manage game fish.

However, there is no option to preserve intact systems for Rio Yaqui fishes in the

United States. In addition, we are precluded from reintroducing the species into typical habitat; desert streams in the Rio Yaqui basin in the United States typically contain no water for extended periods. We are thus forced by circumstances to learn how to manage Rio Yaqui fishes in artesian ponds, where the species probably never historically occurred.

To ensure we take the right management actions with Rio Yaqui fishes, we must understand the life requirements of each species and how those requirements are modified by changes (including biotic changes) in the ecosystem. Clearly, understanding the life requirements of these species would be best accomplished by studying them in optimal habitats, followed by subsequent observations in less optimal habitats. However, such an approach was not possible with Rio Yaqui fishes.

Historical habitats in the United States have ceased to exist and habitats in Mexico are often remote, located on private property, and are not generally accessible. 21

Water bodies on the refuge are not entirely typical of historical habitats.

However, they might not be as different as they first appear. Native fishes in the Rio

Yaqui basin have evolved in response to a monsoon-drought cycle of water flow.

Surface flow often ceases during periods of drought and fish are restricted to isolated

pools. In addition, several species select for relatively quite backwaters and pools

even under mesic conditions. These pools may not be entirely different in character from the small ponds to which most of the fish are constrained on the refuge. Also, refuge ponds and Leslie Creek represent the only areas that have been dedicated to the preservation and recovery of these species in this country. To accomplish even the preservation of these species in the United States, we will need to learn how to manage aquatic habitats on SBNWR for the greatest benefit of Rio Yaqui fishes.

Therefore, my first objective was to evaluate what conditions were used by native Rio

Yaqui fishes in ponds and in Leslie Creek on SBNWR. My second objective was to use this information to determine how refuge ponds might be modified or managed in order to make them better for Rio Yaqui fishes.

The fishes I studied on SBNWR are stream fishes that have been able to survive and even thrive in artesian ponds on the refuge. However, because they are typically stream fishes, physical factors such as depth, velocity (in Leslie Creek) and substrate would seem likely to be important habitat factors. Therefore, 1 studied use of these factors in North, Oasis and Tule Ponds and Leslie Creek on SBNWR.

Fishes occasionally occur in monotypic assemblages, but generally co-occur with other species. Although, the refuge currently does not support exotic fish species, 22

several species of Rio Yaqui fishes do co-occur in the ponds. However, ability to co-

occur does not indicate lack of interactions. Therefore, my third objective was to

evaluate physical habitat utilization of these fishes where they occurred alone versus

where they occurred in combination with other native fish species. Changes in habitat

use patterns in areas of co-occurrence might give insight into how to better manage

ponds to benefit Rio Yaqui fishes.

Study Area

Tule, North and Oasis Ponds all occur on the main refuge. They differ in area

and maximum depth, but are all characterized by clear water with extensive vegetation

(Typha sp., Naias marina. Potamogeton spp., and Chara sp.). All the wells and springs draw water from the same aquifer (USFWS 1987).

Tule Pond is about 0.03 ha and is fed by a natural seep within the pond basin

(see also Table 2). The maximum depth in the pond is 1.5 m. Without regular maintenance, this pond quickly becomes covered with emergent aquatic vegetation

(USFWS 1987). The dominant aquatic vegetation within the pond basin are sedges, cattails, and pondweed. This pond is small and undergoes rapid development. During my study, this site continually experienced low dissolved oxygen (DO) levels caused by the decomposition of organic matter within the basin. Tule Pond contains only

Yaqui topminnows.

North Pond is a relatively old pond of 0.10 ha that receives water from an artesian well near the pond. The maximum depth is 2.0 m. During summer, it 23

becomes overgrown with emergent (mostly Tvpha sp.) and submergent vegetation

(primarily Naias marinus and Potamotieton spp ). Periodically, refuge personnel thin vegetation (cattails). Yaqui topminnows co-occur with Yaqui chubs in North Pond.

Oasis Pond is about 0.12 ha and has a maximum depth of 2.1 in. The pond receives water from an artesian well. It was constructed relatively recently and does not contain much vegetative during the summer. Oasis Pond also lacks the woody riparian vegetation present at the other two sites. Yaqui topminnows co-occur with the

Yaqui chubs and beautiful shiners in this pond.

Leslie Creek contains a short reach of permanent water which is characterized by very shallow riffles and runs interspersed with shallow to deep pools. Over 3 years in the early 1970's the permanent water averaged 843 meters in length (Silvey 1975).

On an April 23, 1994 fish survey, the stream measured approximately 1032 m. The stream had a mean annual flow rate of 0.0074 mVsec in the 1989 water year (USGS

1989). 24

Table 2. Description of sites studied 011 the San Bernardino National Wildlife Refuge.

Site Area Maximum Source Mean Mean Mean Species Depth (m) Secchi Conductivity Dissolved Present (m) (mmho) Oxygen

Tule Pond 0.03 1.49 Natural Seep 2.17 411 2.45 Topminnow

North Pond 0.10 2.02 Artesian Well 2.98 358 8.81 Topminnow Chub

Oasis Pond 0.12 2.14 Artesian Well 2.41 361 12.63 Topminnow Chub Shiner

Leslie Creek - 1.98 Spring 3.81 464 5.80 Chub Topminnow? Longfin Dace

METHODS

Habitat Use

The three species studied occur in four different combinations on the refuge.

Therefore, habitat used by topminnows was observed in the absence of the other two species in Tule Pond. Habitat used by chubs was observed in the presence of topminnow in North Pond. Habitat used by beautiful shiner was observed in the presence of chubs and topminnows in Oasis Pond. Habitat used by Yaqui chubs was also observed in the absence of the other two species in Leslie Creek. Longfin dace were present in Leslie Creek, but very few were actually seen in the section of stream studied (Figure 2).

Microhabitat use was evaluated by snorkeling a predetermined set of transects in each pond at a constant rate and marking the locations of individual fish with a 25

labeled, weighted buoy. General types of environments (e.g., pools, riffles) occupied

by species are referred to as "macrohabitats". Within macrohabitats a species may

restrict their activity to a localized subset of the available environment that best satisfy

their immediate requirements (Ortli 1985, Shirvell and Dungey 1983) and can be related to physical characteristics such as water depth, substrate, and velocity (Lewis

1969, Devore and White 1978). The occurrence of each species, approximate lengths, frequency of individuals, substrate, cover type, distance to cover, and activity were noted along each transect. In addition, the position of fish in the water column above the substrate or focal point elevation (Moyle and Baltz 1985) was estimated in reference to graduated tick marks on the buoy string. Bank observations were used to supplement snorkeling observations in shallow areas of ponds. Physical measurements of the area occupied by fish were made after all transects were observed. Water temperature at the fish's position was also noted. The size of fish was estimated

(Griffith and Fuller 1979, Griffith 1981, Platts et al. 1983) to identify size specific differences in habitat use (Orth et al. 1982). Objects underwater appear to be 33% larger than actual size to the human eye (Helfman 1985). I validated the estimated size of the fish by comparing it to a known-sized object (Springer 1993), for example, my hand or a thermometer. Habitat use information for fish which were visually disturbed, either by approaching or avoiding (fleeing) the diver or observer, were not used in the analysis (Fausch and White 1981, Heggens et al. 1990). Water depth and depth to fish from the substrate were measured using a surveying rod marked in

English measurements and later converted to metric. 26

Figure 2. Leslie Creek Canyon in the southern extension of the Swisshelm Mtns., San Bernardino National Wildlife Refuge, southeastern Arizona (25 km north of Douglas).

Spring Origin

ARIZONA

Existing Rock Dam

Thin lines indicate study transects.

Study site between rock and gage station dams.

Z^USGS Gage Station #09537200

Leslie Canyon Road

Stream Terminus April 23, 1994 27

Table 3. Modified Wentworth particle size scale.

Category Code Size (mm)

Silt 01 <0.0625 Sand 02 0.0625-2 Gravel 03 2-16 Pebble 04 16-64 Cobble 05 64-256 Boulder 06 >256 Bedrock 07 Parent Material* Organic Detritus 08 ** Vascular Plants 09 ** Attached Algae 10 *•

* Inorganic substrate in which only the surface could be seen (i.e., three sides were not visible). ** Organic detritus with a diameter in excess of 2 mm or > 75% coverage of plants.

Substrate was coded according to a modified Wentworth particle size index

(Table 3). Additions to the particle size index include categories for organic detritus,

vascular plants, and attached algae. Substrate type was determined by identifying the

dominant substrate category located within a 0.25-m2 area around each fish. If the area was covered by greater than 75% detritus, vegetation, or attached algae, the substrate category was assigned a value of 8, 9, or 10, respectively. Proximity to cover (Table 4) was defined as the distance to the nearest cover that would fully conceal the fish (Fausch and White 1981, Heggens et al. 1990). 28

Table 4. Cover types used to define cover used by fish.

Category Code

Cattail 01 Sedge 02 Watercress 03 Flooded Terrestrial Vegetation 04 Pondweed 05 Chara 06 Spiny Naiad 07 Woody Material 08 Floating Organics 09 Undercut Bank 10 Substrate 11 Root Wads (exposed roots) 12

Additional depth measurements were taken at 25-cm intervals along an imaginary 1-nr x-y axis originating at the point of observation. These values were averaged to obtain a mean depth at the observation point. The range of these values was also used to identify habitat use along vertical structure such as undercut banks.

Habitat used by Yaqui chubs was also observed in Leslie Creek (Figure 2).

Bank observations were used to supplement observations made by snorkeling in shallow habitats. Snorkeling transects were established using a stratified random selection approach (Bain and Finn 1991). The area of each available macrohabitat was estimated for a representative segment of the permanent reach of the stream; Leslie

Creek between the old U.S. Geological Survey (USGS) weir and the USGS gage station (#09537200). This portion was selected for study because of water clarity and 29

permanency. The percentage of each macrohabitat present in the study area was

determined and a proportional number of randomly selected transects were established

in each macrohabitat type. The same physical variables were measured in the stream

as were measured in the ponds. In addition, mean water column velocity and nose

velocity were measured in the stream. Mean water column velocity was measured at

0.6 of the water column for depths < 1 m. For depths > 1 m, velocity was measured at 0.2 and 0.8 of the water column and then averaged to obtain the mean column velocity. Nose

velocity was measured at the fish's vertical position within the water column. Nose velocity is used to describe the conditions at the precise point occupied by an individual fish and is a better descriptor of microhabitat (Shirvell and Dungey 1983). All water velocities were measured with a Marsh-McBirney electronic flow meter.

Habitat Availability

Habitat availability was estimated by measuring physical variables at 1-m intervals along permanent transects in each pond; the first measurement was taken 10 cm from the water's edge. The physical variables measured included depth and substrate. Point measurements represent average values for the segment 1 m wide and extending half way to adjacent transects (Jones et al. 1984, Bovee 1986).

Bias associated with differences in visibility is one potential problem in using direct observations to measure habitat use (Keenylside 1962). To standardize for this bias, observations were discontinued if visibility was reduced so that the substrate at the bottom 30

of the transect line could not be seen. During brief periods in the summer, visibility was

also limited by the presence of aquatic vegetation.

Habitat availability in Leslie Creek was determined by measuring physical variables

at 0.25-m intervals on permanent transects stratified randomly across the macrohabitat

types. Mean water column velocity was also measured along these transects.

Habitat use and availability were measured about every 8 weeks to identify seasonal

changes. Habitat used by different species and size classes was examined.

Operational Definitions

Life stages of Yaqui chub were identified based on estimated total length and size

at sexual maturity. Young-of-year (YOY) were defined as fish < 15 mm, total length (TL).

Juveniles were defined as fish with total lengths > 15 mm and < 40 mm. Adults were defined

as fish > 40 mm, TL. The definitions of juvenile and adult life stages were based on

Demarais and Minckley (1993) who determined that Yaqui chub reach sexual maturity at as

little as 40 mm, TL.

Life stages of Yaqui topminnow were also defined based on the estimated total length

and size at sexual maturity. Juvenile topminnows were defined as fish < 25 mm TL. Adult

topminnows were defined as fish > 25 mm, TL. These categories were based on the minimum

length at which breeding males were observed and data from Minckley (1973). Breeding males can be readily identified because they are dark brown to black. Breeding males were generally > 25 mm, TL. Minckley (1973) reported that during the peak of the breeding season

98% of females longer than 24 mm standard length (SL) were pregnant. 31

I did not classify beautiful shiner by size because of small sample size.

Therefore, analysis of habitat used by beautiful shiners included all size-classes combined.

Substrate in the ponds were partitioned into open substrates (inorganic and detrital substrates) versus vascular plant substrates (> 75% coverage by plants) for analysis.

Substrates in Leslie Creek were silt, sand, large substrates (gravel through boulder and bedrock), detritus, and plants.

Data Analysis

Total depth of the water column at fish observation points is commonly the primary descriptive variable for microhabitat use by fish amongst species (Bain and Finn 1981,

Cunjack and Green 1983, Faush and White 1981, Gorman 1987, Gorman 1988,

Heggenes 1990, Heggenes et al. 1990, Keenleyside 1962, Rinne 1992) as well as within species (Carpenter and Maughan 1993). However, vertical position, or relative depth, also serves as an equally descriptive variable (Moyle and Baltz 1985). Relative depth (Gorman 1987, Gorman 1988) was calculated by dividing the distance the fish was above the substrate by the total water column depth. I recorded relative depth as

0 at the bottom of the water column and 1 at the surface. Gorman (1988) stated that vertical position provided the most useful information on habitat segregation in the community of minnows studied.

Vertical position and water depth distributions between sites were compared using the Kolmogorov-Smirnov two-sample (K-S) test with an oc level of 0.001. I 32 chose an oc level of 0.001 to minimize biases associated with observability (i.e., abundance of aquatic vegetation and/or water clarity) between sites. I used the Mann-

Whitney-Wilcoxon (M-W) test with an oc level of 0.05 to test for differences within species and sites. Spearman's Rank Correlation Coefficients were used to determine if significant (p < 0.001) linear relationships existed between relative depth, total water column depth, species size and frequency of occurrence. The effect of school size (# offish per observation) on habitat used was investigated by grouping the frequency of occurrence into categories (Table 5). I used the M-W pair-wise test to compare depth and velocity with the Bonferroni correction factor (Miller 1981) to determine the appropriate significance level (e.g., p < 0.0083 for 6 pairwise comparisons). Seasons were defined as spring,summer, fall, and winter. Spring included the months of

March, April, and May. Summer included the months of June, July, and August. Fall included the months of September, October, and November. Winter included the months of December, January, and February. Seasonal use was compared using a

Kruskal-Wallis one-way analysis of variance (K-W one-way ANOVA) test (p < 0.05).

Dunn's (1964) multiple comparisons test was used to determine which pairs differed (p

< 0.05).

Use of substrate was expressed in relative frequency histograms. A chi-square test-of-independence was used to determine if substrate and cover type used differed between species, sites, and life stages (p < 0.05). Cover type used by each species was expressed in a relative frequency histogram. Cover was treated as a discrete 33 binary variable (association with cover or not) and analyzed using a chi-square test

(Cunjack and Green 1983). Spearman Rank Correlation Coefficients were used to determine if the distance to cover was related to the size of fish, total depth of the water column, relative depth, or frequency of fish observed.

Table 5. Size categories of schools.

Category Number of Individuals

Small 2-5 Medium 6-25 Large >25

A chi-square test of homogeneity was used to compare availability of depth, substrate, velocity, and relative depth with use (Ordway and Krausman 1986, Carson and Peak 1987, Thomas and Taylor 1990). Whether or not there was selection for individual habitat categories was determined using Bonferroni Z-tests with simultaneous confidence intervals (Marcum and Loftsgaarden 1980). Substrate categories were grouped as required to validate the Chi-square analysis. Validation required an expected frequency of at least 5 for each cell. Substrates in the ponds were partitioned into open substrates (inorganic and detrital substrates) versus plant substrates (>75% coverage of plants). Some substrates in Leslie Creek were also consoljdated for the chi-square analysis. Substrates in Leslie Creek were silt, sand, large substrates (gravel through boulder and bedrock), detritus, and plants. 34

Availability of relative depth was treated as a known constant with equal

proportions. Therefore, a Chi-square goodness-of-fit test was used to compare use

versus availability (Thomas and Taylor 1990). If a significant difference was detected

at the 0.05 oc level, the Bonferroni Z-test with simultaneous confidence intervals

(Bonferroni SCI)was employed to determine preference of vertical position. The range

of values from 0.00 to 1.00 were divided into vertical zones. 1 used the definitions of

preference and selectivity described by Johnson (1980). A preference is indicated

when the components are offered to the consumer on an equal basis and a higher

proportion is used than would be expected. Selection takes place when a species

selects one resource disproportionate to its availability when the resources are not

available on an equal basis. The two terms are currently used interchangeably. But

since a separate analysis is used for each scenario, 1 chose to distinguish between the

two by referring to them with individual terms.

RESULTS

Habitat Use

A total of 385 observations on individuals and schools of fish of the three

species were collected over 18 months (Table 6). The majority of observations made

were on chubs (56%), followed by topminnows (39%). Because beautiful shiners were only observed in Oasis Pond, the probability of observing them was smaller than that for chubs or topminnows. Small sample size was further complicated by low 35

visibilities in Oasis Pond. A phytoplankton bloom occurred midway through the study

and lasted until the conclusion of data collection in November 1994. Based on

relative abundance (number of fish observed during survey), the community in Oasis

Pond consisted of 85% topminnows, 11% beautiful shiners, and 4% chubs. The

community in North pond consisted of 83% topminnows and 13% chubs. The only

resource component in which Oasis and North Ponds differed significantly was in the

composition of substrates (x = 51.1, df = 2, p < 0.001, Figure 3). Independent

analysis of substrates used by site showed little difference. Therefore, data from these

two ponds were combined in the final analysis.

Table 6. Sample sizes of species by lifestage in three ponds on the San Bernardino National Wildlife Refuge, Arizona.

Species Life Stage Frequency Freq %

Yaqui Topminnow Juvenile 83 22 Adult 66 17 Total All 149 39

Beautiful Shiner All 22 6

Yaqui Chub Young-of-Year 26 7 Juvenile 82 21 Adult 106 28 Total All 214

Overall Total 385 56

Tule Pond, which contains only topminnows, had low dissolved oxygen levels

(as low as 0.5 mg/1, 15 cm below the water's surface) throughout most of the year. 36

Figure 3. Distribution of available substrates for North Pond versus Oasis Pond, SBNWR, Arizona.

Distribution of Available Substrates North vs. Oasis Pond

100 -|

detritus vegetation Category

oasis 37

Topminnows were able to survive in the oxygen-rich surface layer of the water column, but the necessity to maintain contact with the surface layer may have affected their normal distribution in the water column. Therefore, depth use by topminnows in

Tule pond was analyzed separately.

Topminnow A total of 149 observations were made on 4,827 topminnows.

Mean size of topminnow observed was 26 ±11 mm, TL. The average school size was 32 ± 84 fish. Topminnows in Tule Pond ranged in size from 5 to 50 mm, TL.

Relative depths used between adult and juvenile topminnows in Tule Pond differed significantly (M-W test, p < 0.05, corrected for ties). Juvenile topminnows in Tule

Pond preferred the top portion of the water column {yj = 114.00, df = 4, p < 0.05;

Bonferroni SCI, p < 0.05; Figure 4). Adults showed no preference for vertical position in the water column (X = 4.30, df = 4, p > 0.05; Figure 4). Although the bottom portion of the water column was used less than expected, an avoidance was not detected by the Bonferroni SCI (p > 0.05). The mean relative depth used by juveniles and adults was 0.92 ± 0.21 and 0.58 ± 0.30, respectively. Depths used did not differ between juveniles and adults (M-W test, p > 0.05, corrected for ties). Depths used by topminnows in Tule Pond were proportional to availability (K-S test, p > 0.05).

Depths used did not differ by season for juveniles or adults (juveniles: X = 7.35, p >

0.05, corrected for ties; adults: x2 = 3.22, p > 0.05, corrected for ties). Vertical position did not differ by season for adults (K-W one-way ANOVA, y} = 0.35, p >

0.05, corrected for ties), but it did for juveniles (K-W one-way ANOVA, x2 = 13.74, p 38

< 0.05, corrected for ties). Juveniles used areas higher in the water column in the

winter than during other seasons (Dunn's multiple comparisons test, p < 0.05).

Substrate and cover used by juveniles and adults in Tule Pond did not differ

(juveniles: yS = 0.43, df = 2, p > 0.05; adults: X = 7.26, df = 5, p > 0.05). Adults selected open substrates more frequently than they selected vegetated substrates

(Figure 5). Open substrates consisted of inorganic as well as detrital substrates (i.e., non-vegetated substrates). Juveniles showed a similar distribution, but selection was not verified; Bonferroni SCI were not significant (p > 0.05). Topminnows were not found in close association with cover in Tule Pond (p > 0.05). However, when cover was used, it most often was cattails or floating cattail leaves (Figure 6). Distance to cover averaged 22 ±48 cm.

Topminnows in Oasis and North Ponds ranged in size from 5 to 55 mm.

Relative depths used differed significantly between adult and juvenile topminnows in

Oasis and North Ponds (M-W test, p < 0.05, corrected for ties). Juvenile topminnows preferred the surface of the water column (x,2 =36.65, df = 4, p < 0.05; Bonferroni

SCI, p < 0.05; Figure 7). However, there also was some use of the midwater column.

Adults selected a wider distribution for relative depth than would be expected and there was a slight bimodal distribution (Figure 7). Adult topminnows in Oasis and

North Ponds did not show a preference for relative depth (x2 = 4.30, df = 4, p > 0.05).

The mean relative depth used by juveniles and adults in Oasis and North Ponds was

0.72 ± 0.32 and 0.52 ± 0.30, respectively. There was also a significant difference in mean relative depth used between juvenile topminnows in Tule Pond and those in 39

North/Oasis Ponds (K-S test, p < 0.001). Juvenile topminnows in Tule Pond were

observed almost exclusively at the surface (Figure 4), while juvenile topminnows in

North and Oasis Ponds were observed more frequently at positions further down in the

water column (Figure 7). Nonetheless, juvenile topminnows at both sites preferred the

surface of the water column (x2 = 36.65, df = 4, p < 0.05; Bonferroni SCI, p < 0.05).

There was no difference in the depths used by juveniles and adults in North/Oasis

Ponds (M-W, p > 0.05, corrected for ties). Depths used were disproportionate to what

was available (x2 =16.93, df =4, p < 0.05). Topminnows were generally observed in

water < 80 cm deep (Figure 8).

Substrate and cover used in North and Oasis Ponds did not differ among life stages (substrate: x2 = 0.43, df = 2, p > 0.05; cover: x2 = 9.13, df = 5, p > 0.05).

Both juveniles and adults, like topminnows in Tule Pond, selected open substrates more frequently than would be expected based on availability (juveniles: %2 =23.82, df = 2, p < 0.05; adults: x2 = J3.76, df = 2, p < 0.05; Bonferroni SCI p < 0.05;

Figure 5). Topminnows in North Pond were not found in close association with cover

(p > 0.05). When topminnow were near cover in North Pond they used a wide variety of plants. They often were seen near cattails and flooded terrestrial vegetation such as

Bermuda grass (Cvnodon dactvlon"). In deeper water they were found near spiny naiad, Chara, and pondweed (Figure 6). Topminnows in Oasis Pond were found in close association with cover (p < 0.05). They were often near Chara (Figure 6).

Distance to cover averaged 46 ± 77 cm in North Pond, but only 14 ± 17 cm in Oasis

Pond. 40

In general (Spearman's Rank Correlation Coefficients, p < 0.05), as topminnow size increased the position in the water column decreased (-0.5134, Spearman coefficient). As school size increased the distance to cover increased (0.4076) and total depth used decreased (-0.2702). Topminnows were observed in water that ranged from 13 to 30'°C. 41

Figure 4. Distribution of relative depths used by Yaqui topminnows in Tule Pond. Asterisks denote a significant difference using the Bonferroni Z-test simultaneous confidence intervals. If the horizontal bar extends beyond the expected proportion and contains an asterisk it indicates a preference. If the bar does not extend beyond the expected proportion and an asterisk is present, an avoidance is indicated.

Relative Depth Used by Topminnow Tule Pond Surface

0.8 a. a

0.2

Bottom 0.2 0.4 0.6 0.8 Proportion

juvenile i' I adult 42

Figure 5. Substrates used versus available to topminnows in Tule, North and Oasis Ponds. Asterisks denote a significant difference between use and availability using the Bonferroni Z-test simultaneous confidence intervals.

Substrate Used by Topminnow

0.8

n= 149 ;

0.6

c o c o 0.4 o

0.2 -f

IOpen Detritus Plant Substrate Type

| observed [ | expected 43

Figure 6. Cover types used by topminnows in Tule, North, and Oasis Ponds. Cover types are as follows: 1) cattail, 2) sedge, 3) watercress, 4)flooded terrestrial vegetation, 5) pondweed, 6) Chara. 7)spiny naiad, 8) woody material, 9) floating organics, 10) undercut banks, 11) substrate, 12) root wads.

Cover Used by Yaqui Topminnow

0.7

0.6

0.5

c .2 0.4 - o S" r £ 0.3 O. : |:

0.2 -

0.1 -i nn 0 - 5 6 7 10 11 12 Cover Type

I Tule HH North | j Oasis I 44

Figure 7. Relative depths used by Yaqui topminnows in Oasis and North Ponds. Asterisks denote a significant difference between use and expected proportions.

Relative Depth Used by Topminnow Oasis and North Pond Surface i I 1

r

expected = 0.2

Bottom 0 0.1 0.2 0.3 0.4 0.5 0.6 j Proportion ]

| H juvenile adult J J 45

Figure 8. Depths used by topminnows in North and Oasis Ponds.

Total Depth Used Topminnow 0.35

0.25 r

2 0.15

0-40 41-80 81-120 121-160 >161 Total Depth (cm)

observed [ j expected 46

Beautiful Shiner Only 22 observations on a total of 105 beautiful shiner were

collected over the length of the study. Therefore, all seasons and life stages were combined for analysis. The average school size for beautiful shiners was 4.8 ± 6.5 fish.

Shiners ranged in size from 8 to 55 mm, TL. Relative depth used by beautiful shiners in

Oasis Pond was just above the mid-point of the water column (Figure 9). Beautiful shiners were observed most often between 0.6 and 0.8 relative depth in water 40 to 100 cm deep

(Figure 10). Beautiful shiners were observed in the top half-meter of the water column regardless of the depth of the water column. Depths < 40 cm seemed to be avoided by shiners (X = 10.27, df = 4, p < 0.05; Bonferroni SCI, p < 0.05). They were found over both open (inorganic and detrital substrates) and vascular plant substrates. They used substrate in proportion to what was available (x2 = 4.21, df = 1, p > 0.05,

Fisher's Exact Test; Figure 11). Shiners were not found in close association with cover (p > 0.05). Nearest available cover consisted primarily of pondweed and Chara

(Figure 12) and average distance from cover was 23 + 16 cm. Beautiful shiners occupied water with temperatures ranging from 18 to 28°C. Temperatures in Oasis

Pond ranged from 17 to 31°C. 47

Figure 9. Relative depths used by beautiful shiners in Oasis Pond. Asterisks denote a significant difference between observed and expected value for individual depth zones.

Relative Depth Used Beautiful Shiner Surface in = 22

0.8 Q.

—

0.2 1*

Bottom

0.1 0.2 0.3 0.4 0.5 Proportion 48

Figure 10. Depths used by beautiful shiners in Oasis Pond. Asterisks denote a significant difference between observed and available proportions for individual depth categories.

Total Depth Used Beautiful Shiner 0.5

n = 22 0.4

§ 0.3 - '€ OQ. r 2 & 0.2 -

0.1

0-40 41-80 81-120 121-160 >161 Total Depth (cm)

observed j . j expected 49

Figure 11. Substrates used by beautiful shiners in Oasis Pond.

Substrate Used by Beautiful Shiner

n = 22

Open Plant Substrate Type

| observed j i expected 50

Figure 12. Cover types used by beautiful shiners in Oasis Pond. Cover types are 1) cattail, 2) sedge, 3) watercress, 4) flooded terrestrial vegetation, 5) pondweed, 6) Chara. 7) spiny naiad, 8) woody material, 9) floating organics, 10) undercut bank, 11) substrate, and 12) root wad.

Cover Used by Beautiful Shiner

0.5

n = 19 i I 0.4 -!

§0.3 , t i

(XI 0.2 ^!

01 1H .1.1 5 6 7 8 9 10 11 12 Cover Type Yaaui Chub A total of 214 observations were made on Yaqui chub (Table

4). The 214 observations included 2,322 fisli with an average school size of 10.9 ±

26.7 fish. Relative depth used did not differ among sites (K-S test, p > 0.001),

therefore data from all sites were combined for analysis. Relative depth differed

between juveniles and adults (M-W test, p < 0.05, corrected for ties) but not between

YOY and juveniles (p > 0.05). YOY chub (Figure 13) did not show a preference for

vertical position (j; = 2.00, d = 3, p > 0.05), but there was a bimodal distribution for

relative depth used by juveniles (Figure 14). Adults and juveniles were observed most

often at the bottom of the water column (Figure 14). Typically, juveniles and YOY

were found in water 1 m deep (Figure 15), but adults were usually at depths > 1 m

(Figure 15). Total depth used by season did not differ for juveniles (K-W one-way

ANOVA, x2 = 7.50, p > 0.05, corrected for ties) and adults (K-W one-way ANOVA,

X2 = 4.49, p > 0.05, corrected for ties). However, relative depth did vary by season

for both groups (K-W one-way ANOVA, corrected for ties; juvenile: X =10.86, p >

0.05; adults: X ~ 19.85, p < 0.05). The mean relative depth used by juveniles during

summer differed from the remaining seasons (Dunn's multiple comparisons test, p <

0.05). Mean relative depth for juveniles in summer was 0.59 ± 0.32 versus 0.33 ±

0.28 in the remaining seasons. The mean relative depth used by adults during winter differed from the remaining seasons (Dunn's multiple comparisons test, p < 0.05).

Mean relative depth for adult chubs in winter was 0.02 ± 0.05 versus 0.28 ± 0.23 for the remaining seasons. 52

Substrates used in the ponds differed among life stages (x* = 11.89, df = I, p <

0.05) but did not differ among life stages in Leslie Creek (x2 = 4.40, df = 2, p > 0.05).

Juveniles in the ponds were found over silt substrates more often than expected (x2 =

5.02, df = 1, p < 0.05; Bonferroni SCI, p < 0.05; Figure 16). Adults were found almost exclusively over plant substrates (97.3%) usually at the top or within the plant substrate (x2 = 6.05, df = 1, p < 0.05; Bonferroni SCI, p < 0.05; Figure 16). Both juveniles (%2 = 42.38, df = 4, p < 0.05; Bonferroni SCI, p < 0.05; Figure 17) and adults (x2 = 44.95, df = 4, p < 0.05; Bonferroni SCI, p < 0.05; Figure 17) in Leslie

Creek selected silt substrates more often than would be expected based on availability.

Cover types used by YOY chubs in North and Oasis Ponds consisted mostly of cattail and flooded terrestrial vegetation. Juveniles were seen most frequently near spiny naiad and adults were most frequently observed near Chara and pondweed

(Figures 18 and 19). Chubs in the ponds were found in close association with cover

(p < 0.05). Distance to cover did not differ among life stages (M-W test, p > 0.05, corrected for ties). However, mean distance to cover in the ponds and in Leslie Creek did differ (M-W, p < 0.05, corrected for ties). Distance to cover in Leslie Creek was nearly twice that in the ponds (41 ± 52 vs 24 ± 27 cm). Chubs in Leslie Creek were not found in close association with cover (p > 0.05). Cover used in Leslie Creek consisted primarily of root wads (Figure 20). In general (Spearman Ranks), as the size of chub increased total depth used also increased, but relative depth and fish density decreased. Chubs were observed in water ranging from 10 to 30°C. 53

Adult chubs in Leslie Creek were observed in areas where mean water column velocities were disproportionate to what was available (%2 = 4.84, df = 2, p < 0.05;

Figure 21). Adults avoided velocities > 0.04 m/sec (Bonferroni SCI, p < 0.05).

Juveniles used velocity in proportion to availability (%2 = 1.82, df = 2, p > 0.05;

Figure 21). However, Bonferroni SCI (p < 0.05) detected an avoidance of velocities >

0.04 m/sec. Nose velocity for both life stages was commonly at or near 0 m/sec

(Figure 22). 54

Figure 13. Relative depths used by young-of-year chubs (YOY) at all sites.

Relative Depth Used by YOY Chub All Sites

n = 26

0.2

0.05 0.1 0.15 0.2 0.25 0.3 0.35 Proportion 55

Figure 14. Relative depths used by Yaqui chubs (excluding YOY) at all sites.

Relative Depth Used by Yaqui Chub Surfacc -

expected = 0.2

Bottom

0 0.1 0.2 0.3 0.4 0.5 0.6 { Proportion j

i HI iuvenile£^adult^| j 56

Figure 15. Depths used by Yaqui chubs at all sites. Asterisks denote a significant difference between use and expected proportions based on availability.

Total Depth Used by Yaqui Chub

0-40 41-80 81-120 121-160 >161 Total Depth (cm)

| juvenile [ i adult expected 57

Figure 16. Substrates used by Yaqui chubs in North and Oasis Ponds. Asterisks denote a significant difference between use and availability.

Substrate Used by Yaqui Chub North/Oasis Pond

0.8|

Open Plant Substrate Type

] juvenile f j adult expected 58

Figure 17. Substrates used by Yaqui chubs in Leslie Creek. Asterisks denote a significant difference between use and availability.

Substrate Used by Yaqui Chub Leslie Creek 0.7 -r I

0.6 -

0.5 -

• 2 0.4 • t: oQ. S 0.3 - o.

0.2 -

0.1 -

0 - H il Silt Sand Large DetritusI Plant Substrate Type 59

Figure 18. Cover types used by different life stages of Yaqui chubs in North Pond. Cover types are 1) cattail, 2) sedge, 3) watercress, 4) flooded terrestrial vegetation, 5) pondweed, 6) Chara. 7) spiny naiad, 8) woody material, 9) floating organics, 10) undercut bank, 11) substrate, and 12) root wads.

Cover Types Used by Yaqui Chub North Pond

1 2 3 4 5 6 7 8 9 10 11 12 Cover Type I • yoy BUI juvenile[ j adult 60

Figure 19. Cover types used by Yaqui chubs in Oasis Pond. Cover types are 1) cattail, 2) sedge, 3) watercress, 4) flooded terrestrial vegetation, 5) pondweed, 6) Chara. 7) spiny naiad, 8) woody material, 9) floating organics, 10) undercut bank, 11) substrate, and 12) root wads.

Cover Types Used by Yaqui Chub Oasis Pond

Cover Type 61

Figure 20. Cover types used by different life stages of Yaqui chubs in Leslie Creek. Cover types are 1) cattail, 2) sedge, 3) watercress, 4) flooded terrestrial vegetation, 5) pondweed, 6) Chara. 7) spiny naiad, 8) woody material, 9) floating organics, 10) undercut bank, 11) substrate, and 12) root wads.

Cover Types Used by Chub Leslie Creek 0.8

0.6

c o t: o 0.4 O- o

0.2

12 3 4 5 6 7 9 10 11 Cover Type

! ! y°y |juvenile j I adult 62

Figure 21. Mean water column velocity for Yaqui chub observations in Leslie Creek. Asterisks denote a significant difference between observed and expected proportions for that individual category.

Mean Water Column Velocity Yaqui Chub

0 0.01-0.04 >0.04 Velocity (m/sec) Figure 22. Nose velocity for Yaqui chubs in Leslie Creek.

Nose Velocity Yaqui Chub Leslie Creek

0.01-0.04 >0.04 Velocity (m/sec)

|juvenile^__J^ 64

DISCUSSION

Yaqui topminnows used different depth and cover profiles in certain ponds and in the presence of other fish species. In Tule pond, where Yaqui topminnows occurred alone, they occupied the surface layers of the water and showed no selection for cover. However, occupation of the surface layer may not represent a real preference because low oxygen levels were persistent in this pond throughout the summer. Juvenile topminnows in North and Oasis Ponds, where dissolved oxygen levels were not low, showed the same preference for the surface zone. However, in these two ponds juveniles also used the mid portion of the water column. Oxygen levels in Tule pond appeared to be low because of decomposition of organic matter.

Fish may have remained at the water surface to facilitate contact with the oxygen rich layer; oxygen content at the surface was consistently at or above 2 mg/1 versus 0.2 mg/1 at the bottom. Topminnows have evolved in habitats that are subjected to periodic low oxygen conditions as a consequence of droughts or flow cessation. They have evolved behaviors such as air gulping to allow periodic survival of these conditions. However, since ponds and streams on San Bernardino National Wildlife

Refuge represent the last refugia for this species, it would seem prudent to develop techniques and protocols for augmenting oxygen supplies during critical periods, or eliminate low oxygen episodes by managing ponds to limit the amount of organic matter subject to summer decomposition.

Adult topminnows (> 25 mm) were distributed throughout all depths even in

Tule Pond, but were not observed during periods of low oxygen levels. It is possible 65 that the largest topminnow congregated near the surface among thick cover to avoid aerial predation. Very large adults (> 45 mm), usually females, were often located at mid depth amongst or near submerged vegetation. It was at these depths that most breeding activity was observed. Large adults were seldom seen except during breeding.

Minckley et al. (1977) state that Gila topminnows prefer water depths < 25 cm.

However, Stefferud (1984) reported that Sonoran topminnow (Poeciliopsis occidentalis;

Gila and Yaqui topminnow) were observed in water up to 1 m in depth. The results reported by Stefferud (1984) and Forrest (1992) are similar to the results I obtained for

Yaqui topminnow. Forrest (1992) reported that Gila topminnow were observed in water ranging from 1 cm to over 1 m in depth. Yaqui topminnow occurred in water <

80 cm about 63 % of the time, but were often observed in water up to 2 m deep, usually over submerged vegetation. Several factors may contribute to the use of shallow areas by poeciliid fishes. One may be the avoidance of predation by larger fish (Meffe and Snelson 1989).

Substrate could be important for one or more life history functions or feeding.

However, for livebearing fish such as the topminnow, it appears probable that substrate might play a secondary role in the selection of habitat. Substrates at the depths used were dominated by silt and detritus. However, topminnows seldom fed over silt substrates, but were often seen picking epiphytic algae from submerged plants. The apparent selection for silt substrates could be a by-product of selection for shallow water. 66

Minckley (1973) states that detritus is a major food source for topminnows.

Brooks (1985) suggested that detrital substrates might benefit Gila topminnow, but found that the presence of silt did not significantly improve the success of Gila topminnow re-introductions. Forrest (1992) reported that Gila topminnow had no clear substrate preference. 1 seldom observed Yaqui topminnow consuming detrital items off the substrate. Most observations on feeding topminnows involved consumption of epiphytic algae and particulates suspended in the water column. Therefore, plant substrates appear to be important to topminnow.

During winter, juvenile topminnows were frequently found in shallow areas.

The higher temperature of shallow areas may be important in the selection for these habitats during winter (Meffe and Snelson 1989). Forrest (1992) also reported that

Gila topminnow (Poeciliopsis occidentalis occidentalis) selected shallow areas because of higher water temperatures. Water temperatures were also warmer where ground­ water entered ponds. Topminnows, especially in North Pond, congregated in large schools at the PVC pipe inlet. They remained at this location during daylight hours, even though the temperatures around the pipe occasionally were lower than those in other parts of the pond. This preference for the inlet area may be the result of stability in diurnal temperatures or possibly a preference for positions at or near flowing water. Both of the these hypotheses are in disagreement with what Forrest

(1992) reported for Gila topminnow. He noted that Gila topminnow exhibited diel movements in relation to temperature (movement into warmer areas) and avoidance of moving water. Sonoran topminnow (Poeciliopsis occidentalis) routinely use 67 springheads with high C02 and low pH (Meffe and Snelson 1989). Longsworth (1991)

reported the pH as 7.82 ± 0.39 and CO: concentrations as 83 ± 60 mg/1 from well water on SBNWR. Kepner (1988) reported the average pH from well discharge to be

7.78 ± 0.72.

Schooling was most prevalent among juvenile topminnows. The use of shallow areas by topminnows may be the result of avoidance of predation by larger fish (Meffe and Snelson 1989). Although the risk of predation in shallow water is lessened, the threat still remains. Schooling may provide better protection from predators as well as increased swimming efficiency (Moyle and Cech 1988).

Of the three species studied, the least is probably known of the beautiful shiner. Beautiful shiner are closely related to red shiner (Cyprinella lutrensis) of the

Mississippi River drainage. Matthews and Hill (1979) reported that temperature, current speed, and water depth had the greatest influence on the selection of habitat by red shiner. Dissolved oxygen, turbidity, shelter, shade and substrate are probably of lesser importance in habitat selection by red shiner (Matthews and Hill 1979).

Beautiful shiner in Oasis Pond (SBNWR) were most often observed in the subsurface zone in water between 41 and 80 cm. Matthews and Hill (1979) stated that red shiner avoid shallow water, possibly due to thermal instability or exposure to predation.

Starrett (1951) suggested that red shiner might avoid shallow water to avoid becoming stranded in shallow pools during drought. Beautiful shiner in the Rio Yaqui drainage of Mexico are primarily a lotic system inhabitant occupying riffles of smaller streams and intermittent pools of creeks that have a high percentage of riffles in wetter periods 68

(Hendrickson et al. 1980). The avoidance of water < 40 cm may be in response to the

absence of moving water and an effort to occupy areas less likely to become dry

during drought.

Both topminnovvs and beautiful shiners in Oasis Pond were concentrated in

warmer water areas near the artesian inflow during winter (January and February).

Water on the north shore of the pond was as much as 3°C warmer than at other

locations. Although fish survived without access to these thermally moderated areas,

it would seem prudent to define the thermal preferences and tolerances of these

species. Such a study was beyond the scope of the current work and would have to be

conducted under experimental conditions in a laboratory. Filamentous algae also was

present on the north shore. Both species were observed consuming algae.

Although primarily found in small streams, beautiful shiner were relatively

abundant in an earthen tank in Mexico (Hendrickson et al. 1980). They also appear to

be maintaining viable populations in the artesian fed ponds of the SBNWR. Since

beautiful shiner only occur in still water habitats on the refuge, investigation of use of

velocity was not possible.

Beautiful shiner used the upper mid-portion of the water column and formed aggregations in open water in areas with minimal cover. Matthews and Hill (1979)

reported that red shiner exhibited similar behavior. Beautiful shiner did not select areas of dense aquatic vegetation. However, they did stay around the margins of the

ponds. These pond margins are most susceptible to becoming overgrown with aquatic vegetation. Therefore, in order to provide acceptable habitat for beautiful shiner, it 69

may be necessary to periodically reduce the amount of vegetation around part of the

pond margin.

The nearest available cover for beautiful shiner consisted mostly of mid water

plants such as pondweed and Chara. These plants were common at intermediate depths. However, beautiful shiner seldom used these plants for cover. Instead, when disturbed they swam rapidly and laterally remaining in water of similar depth. This

behavior suggests that cover is of lesser importance to beautiful shiner than water depth in selecting habitat. Schooling by beautiful shiner may reduce the need for physical cover (Moyle and Cech 1988).

Juvenile Yaqui chub used similar conditions in both flowing and standing waters. Adults also used the same conditions in both flowing and standing waters, but consistently occupied deeper water and a narrower spectrum of depths than juveniles.

Therefore, it appears that the depth, velocity, substrate, and cover profiles identified for adult and juvenile Yaqui chub in my study represent real preferences. These profiles should be used as design criteria to construct or renovate habitats for this species on the refuge. Of particular importance to chub is the presence of deep (> 80 cm) pools in or adjacent to areas of thick vegetative cover (North and Oasis Ponds) or root wads (Leslie Creek).

Adult Yaqui chub were generally closer to the substrate in Leslie Creek than in the ponds. Substrate in the ponds was most often covered with vegetation whereas it was not in Leslie Creek. In the ponds, adult chub were generally at the top fringes or within submerged vegetation. Fish may have been keying on the "false bottom" 70 created by the surface of the plants. It is also possible that I failed to see fish adjacent to the substrate due to visual obstruction by plants. The association between adult chub and vegetated substrates in the ponds and with silt in Leslie Creek may be an artifact of this preference for water > 1 m deep. Juvenile chub were generally observed over silt substrates. Silt substrates were found in water < 1 m deep in the ponds. However, in Leslie Creek where silt substrates dominated, chub of both life stages selected silt substrates. This selection in Leslie Creek appears to be based on a

"preference" for pool habitats > 40 cm deep rather than for any particular substrate.

There was zonal stratification of different size classes of chub within pools in

Leslie Creek. Stratification based on size may explain the bimodal distribution in relative depth used by juvenile chub (Figure 15). The "ecological species" concept described by Gorman (1987,1988) is a complimentary use of vertical positions in areas of varying water depths by YOY and older life stages of fish of the same species.

That is, although YOY fish utilized shallower water, they used vertical zones similar to those used by juveniles and adults in deeper water. Yaqui chub differed in that individual size classes separated vertically. Chub were seldom seen within 25 cm of the water's surface regardless of size, except on two occasions. Low oxygen also occasionally appeared to be a factor in profiles of depths used by Yaqui chub. During periods of zero flow in Leslie creek, chub were sometimes seen piping at the surface.

Oxygen measurements indicated that levels were 5.5 mg/1 at the surface and < than 1 mg/1 at the bottom. Chub avoid water in which DO levels drop below 2.5 mg/1. The pool was also thermally stratified. The surface temperature was 1.5°C wanner than 71

the bottom. However, chub frequently occupied water with lower temperatures.

Therefore, I conclude that dissolved oxygen concentrations were responsible for

aberrant vertical positions in these two pools.

Yaqui chub were generally found in waters with little or 110 velocity.

However, it is unclear whether they were avoiding velocity or selecting deep water.

Water in which velocities were greater than 0.05 m/sec were under 20 cm deep.

Juvenile chub occupied pools with velocities that approached 0.1 m/sec, but the

highest focal point velocity for any chub was 0.07 m/sec. Water depth seems to have

the most influence on distribution of chub. Roundtail chub (Gila robusta) also occupy

pool environments characterized by slower to zero velocity (Rinne 1992).

Minckley (1980) noted that Yaqui chub occupy areas "in association with

higher aquatic plants", implying a high affinity for cover. Areas occupied by chub in

the ponds on SBNWR and Leslie Creek had abundant aquatic plants and chub sought

cover in these plants when disturbed. In addition, chub in ponds had a close

association with cover. However, the same was not true in Leslie Creek. Several

factors may be responsible for this difference. Leslie Creek has a dense overstory of

riparian vegetation. The shade provided by the overstory in conjunction with water >

40 cm deep may minimize the need to seek protection in instream cover. Smaller size classes of chub reacted differently when an observer approached the bank of the stream than when the observer was within the water column. In the first case, fish responded by swimming lower in the water column. If approached by an observer in 72 the water column, fish immediately darted into exposed root wads on the vertical banks.

Vertical position was found to provide the most information on habitat segregation of the six habitat variables considered in a niche analysis by Gorman

(1987, 1988). Moyle and Baltz (1985) stated that vertical position was important in describing the microhabitat use by species. Moyle and Baltz (1985) and Gorman

(1987, 1988) also found that although juveniles and adults used different depths, they used similar vertical positions. For both Yaqui topminnow and Yaqui chub 1 found that vertical positions as well as depths used differed between juveniles and adults.

Yaqui topminnow juveniles were observed in the shallowest water near the edge of the pool, while the adults were observed in deeper water and further down in the water column (Figure 23). Both variables were inconsistent for the two lifestages of Yaqui topminnow. The same can be said for vertical position and depths used by Yaqui chub. As mentioned earlier, Yaqui chub exhibited a vertical stratification of size classes in both the ponds and Leslie Creek (Figure 23). It was more prominent in

Leslie Creek. Juvenile chubs in the ponds were most frequently observed in shallower areas and at positions higher in the water column (Figure 23). Different lifestages for beautiful shiners were not examined, but most observations occurred between the adult

Yaqui topminnows and the juvenile Yaqui chubs. Beautiful shiner were observed higher in the water column tlian either the adult topminnow and the juvenile chub

(Figure 23). Figure 23. Distribution of Yaqui topminnow, beautiful shiner, and Yaqui chub in a cross section of a pool.

POOL CROSS SECTION

SURFACE

TOPMINNOW

SHINER

CHUB BOTTOM 74

Biotic Interactions

Depth, velocity, substrate and cover profiles utilized by Yaqui chub were

unaffected by the presence of Yaqui topminnows or beautiful shiners. There was also no indication in my data that the presence of Yaqui chub affected habitat selection by

Yaqui topminnows. Minckley (1980) reported that adult Yaqui chub feed on Yaqui topminnow. 1 saw no indications of predation on Yaqui topminnows by Yaqui chub during underwater observation in my study, but such predation is certainly possible.

Yaqui topminnows selected cover in Oasis pond where they co-occurred with Yaqui chub. However, they failed to select cover in North pond where they also co-occurred with Yaqui chub. Tf predation by Yaqui chub were involved in cover selection by

Yaqui topminnows, one might expect cover selection to have occurred in both ponds.

These data coupled with the absence of any direct observation of interactions would seem to indicate that Yaqui chub could continue to be placed together with Yaqui topminnows. However, an effort should be made to evaluate predation by Yaqui chub on Yaqui topminnows.

It is impossible to determine whether habitat utilization by beautiful shiners was modified by the presence of Yaqui chub or Yaqui topminnow because they co- occur in Oasis Pond and one other pond on the refuge. There are no base line for habitat use by beautiful shiner alone. Development of habitat use profiles for beautiful shiner in the absence of other species might allow evaluation of how representative my profiles from Oasis pond were. 75

Yaqui topminnows selected cover when in the presence of beautiful shiners and

Yaqui chub (Oasis Pond) but did not select cover when in the presence of Yaqui chub

(North Pond) or where they occurred alone (Tule Pond). These data may indicate

negative interactions between beautiful shiner and Yaqui topminnow. However, they might also be indicative of differences in predation pressure from bullfrogs (Rana catesbiana) 011 Yaqui topminnow in the three ponds. Bullfrogs occur in all three study ponds and are known to feed upon Yaqui topminnow, Chiricahua leopard frogs and garter snakes (C. Schwalbe personal communication). Since the early 1990's intensive bullfrog control and removal programs have occurred on the refuge to protect leopard frogs and garter snakes. However, bullfrog control efforts have not been uniform over the three study ponds. The most extensive bullfrog removal efforts have been on

North Pond and the least extensive efforts 011 Tule Pond. Control efforts have been intermediate on Oasis Pond (C. Schwalbe personal communication). If bullfrog predation were the causative factor for cover selection by Yaqui topminnow, one might expect Yaqui topminnows to also select areas of cover in Tule pond. However, no such selection was observed. Despite the complexities of interpreting my observations, I would suggest that beautiful shiners not be stocked in combination with

Yaqui topminnows until there has been a more exhaustive study of the causative factor or factors for cover selection by Yaqui topminnows in Oasis pond. 76

Applications

Management for the three species I studied has focused primarily of ensuring the survival of the species within the United States. Much of the historic range of these fishes within the United States no longer can support the native aquatic fauna.

Therefore, recovery and subsequent downlisting and delisting will depend on the status of these species in Mexico. Within the SBNWR, populations of these fish occur in closed systems (i.e., ponds). Within these systems, habitat must insure perpetual existence of each species.

Over time, lentic environments experience successional ontogeny (Wetzel

1975). In shallow systems, such as the ponds on SBNWR in which littoral dominance occurs, this ontogeny is accelerated. Left untreated, these systems move from a rheotrophic (lake-like conditions) state to that of a marsh, saturated with sediments and having little standing water (Wetzel 1975). Marsh conditions would likely be unsuitable for both Yaqui chub and beautiful shiner. Therefore, maintenance of these ponds in an early successional stage is beneficial to the survival of Yaqui chub and beautiful shiner. Specifically, water greater than 1 m deep would be needed for Yaqui chub. Also, areas free of dense aquatic vegetation and > 40 cm deep may be needed by beautiful shiner.

Tule Pond became choked with sedges, cattails, and Phratimites and the accumulation of organic matter and subsequent decomposition resulted in reduced dissolved oxygen concentrations. With the absence of oxygen-producing submergent vegetation, dissolved oxygen concentrations remained low and continued to drop. 77

Although topminnow are capable of surviving such conditions, they may not be beneficial to the population. The limited available habitat resulting from low oxygen levels could be detrimental to larger fish, primarily females, which appear to prefer positions in the water column below the surface.

The historic composition of fishes in Leslie Creek on the SBNWR is uncertain.

It is believed that Yaqui chub, Yaqui topminnow, Mexican stoneroller and longfin dace co-occurred in the stream. The stream is highly variable in length and the area of surface flow averages 850 to 900 m in length. However, during drought years surface flows cease and only deep pools contain water. During drought longfin dace can persist by taking refuge in moist detritus and algal mats (Sublette et ai. 1990), but

Yaqui chub and topminnow maintain populations in the deeper pools. It is under drought conditions that Yaqui cluib may prey heavily upon topminnow. During one drought in the mid 1980's, Leslie Creek was reduced to one very deep pool (Kevin

Cobble, SBNWR Manager, Personnel communication) and the topminnow population in Leslie Creek was reduced severely. Following cessation of drought conditions, the topminnow population did not recover in Leslie Creek. It is not known whether the current conditions in Leslie Creek are suitable for longterm survival of topminnow.

Topminnows have consistently been collected from Leslie Creek during annual surveys by SBNWR and Arizona Game and Fish Department personnel, but numbers have always been low and appear to be diminishing. During recent surveys topminnow have not been collected. The absence of topminnow has prompted the refuge manager to consider restocking topminnows in Leslie Creek. I suggest that any restocking be 78 experimental and be followed by monitoring of persistence, habitat selection, and survival.

Ponds and Leslie Creek on the refuge are isolated, closed environments which inhibit emigration and immigration. There is concern that captive breeding of endangered fishes will result in changes and may preclude there successful reintroduction into nature (Meffe 1987, Meffe and Vrijenhoek 1988) or may not be beneficial to wild populations (Allendorf and Ryman 1987, Echelle 1991, Hindar et al.

1991). Although the populations are not considered captive breeding populations and do exist in their native range, they are no different from the captive populations that were maintained at the Dexter National Fish Hatchery and Technology Center.

Maintenance of large stocks can be considered as a remedial measure to prevent deleterious effects and can be applied in the management of isolated populations (Demarais and Micnkley 1993). Although comparisons with different populations was not possible for Yaqui chub, Demarais and Minckley (1993) found no evidence for reduced genetic variability. However, reduced variability was reported for similarity isolated topminnows (Vrijenhoek et al. 1985) and other arid-land poeciliids and cyprinodontids (Echelle 1991). Certain life history characteristics, such as high fecundity and the ability to establish large populations quickly, may be conducive to maintenance of high levels of genetic variation (Demarais and Minckley

1993). Yaqui chub exhibit these characteristics. The life history characteristics exhibited by Yaqui chubs may account for the low amount of differentiation in the 79

natural populations as well as in the managed stocks in ponds on the refuge (Demarais

and Minckley 1993).

Interconnectedness of aquatic environments is currently not a priority on the

refuge. Although Yaqui chubs may be somewhat immune to the problems of

isolation, other species such as the Yaqui topminnow or the beautiful shiner as well as

other species proposed for reintroduction may be adversely affected by such conditions. One solution to this potential problem could be the allowance of outflow from certain ponds into Black Draw. However consideration should be given to the

possible colonization of ponds on the refuge by non-native species present in the Rio

Yaqui basin within Mexico. Another solution may be the implementation of pedigree development programs. Periodic genetic monitoring and human-assisted exchange of genetic material between populations may prevent genetic problems faced by many other endangered species.

Suitable habitat for Yaqui chubs is present in Leslie Creek to support a fairly large population. However, during drought, populations may be reduced to only 100's or even 10's of individuals resulting in possible bottlenecking (Demarais and Minckley

1993). The problem in Leslie Creek may be alleviated by supplemental flows by well water similar to that in Bathouse Spring. APPENDIX A

Bonferroni confidence intervals for preferences and selectivities. 81

RELATIVE DEPTH PREFERENCES

JUVENILE TOPMINNOW--TULE POND

Category Expected Observed Bonferroni Intervals Significant Avoidance Proportion Proportion /Preference

0 - 0.2 0.2 0.025 -0.03859 0.0885899 Yes Avoidance

0.2 - 0 4 0.2 0.025 -0.03859 0 0885899 Yes Avoidance 1 0 © ON 0.2 0.025 -0.03859 0 0885899 Yes Avoidance

0.6 - 0.8 0.2 0.05 -0.038769 0.13876693 Yes Avoidance

0.8 - 1.0 0.2 0.875 0.7402977 1.0097023 Yes Preference

JUVENILE TOPMINNOW—NORTH/OASIS PONDS

Category Expected Observed Bonferroni Intervals Significant Avoidance Proportion Proportion /Preference

0 - 0.2 0.2 0.1162791 -0.009648 0.2422064 No O © ts) 0.2 0.0232558 -0.03595 0.0824621 Yes Avoidance

0.4 - 0.6 0.2 0.255814 0.0844125 0.4272154 No

0.6 - 0.8 0.2 0.0697674 -0.030309 0.1698443 Yes Avoidance 0 © © 0 1 0.2 0.5348837 0.3389442 0.7308233 Yes Preference

JUVENILE CHUB--ALL SITES

Category Expected Observed Bonferroni Intervals Significant Avoidance Proportion Proportion /Preference

0 - 0.2 0.2 0.3292683 0.1955816 0.462955 No

0.2 - 0.4 0.2 0.1463415 0.04 57955 0.2468874 No 1 ON © © 0.2 0.1219512 0.0288637 0.2150387 No

0.6 - 0.8 0.2 0.2560976 0.1319324 0.3802627 No 0 0 © © 1 0.2 0 1463415 0.0457955 0.2468874 No 82

ADULT CHUB-ALL SITES

Category Expected Observed Bonferroni Intervals Significant Avoidance Proportion Proportion /Preference

0 - 0.2 0.2 0.5754717 0.4518034 0.69914 Yes Preference

0.2 - 0.4 0.2 0.1698113 0.0758683 0.2637544 No

0.4 - 0.6 0.2 0.1698113 0.0758683 0.2637544 No

0.6 - 0.8 0.2 0.0754717 0.0093803 0.1415631 Yes Avoidance

0.8 - 1.0 0.2 0.009434 -0.014753 0.0336209 Yes Avoidance

BEAUTIFUL SHINER—OASIS POND

Category Expected Observed Bonferroni Intervals Significant Avoidance Proportion Proportion /Preference

0 - 0.2 0.2 0.0454545 -0.068944 0.1598533 Yes Avoidance

0.2 - 0.4 0.2 0.1363636 -0.052109 0.3248367 No

0.4 - 0.6 0.2 0.1818182 -0.030007 0.3936436 No

0.6 - 0.8 0.2 0.4545455 0.18108 0.7280109 No 0 o o 0 1 0.2 0.1818182 -0.030007 0.3936436 No

TOTAL DEPTH SELECTIVITIES

TOPMINNOW DEPTH SELECTION IN NORTH/OASIS PONDS

Category Expected Observed Bonferroni Intervals Significant Avoidance Proportion Proportion /Preference

0-40 cm 0.2183174 0.3085106 0.1857922 0.4312291 No

41-80 cm 0.2193823 0.3297872 0.2048749 0.4546995 No

81-120 cm 0.1469649 0 0851064 0.0109671 0.1592457 No

120- 160cm 0.1576145 0.0957447 0.0175666 0.1739227 No

> 160 cm 0.257721 0.1808511 0.0785868 0.281153 No 83

JUVENILE CHUB DEPTH SELECTION-ALL SITES

Category Expected Observed Bonferroni Intervals Significant Avoidance Proportion Proportion /Preference

0-40 cm 0.4722222 0.2962963 0.1831105 0.4094821 Yes Avoidance

41-80 cm 0.1518519 0.2685185 0.1586628 0.3783743 Yes Preference

81-120 cm 0.1175926 0.1944444 0.0963421 0.2925468 No

120-160cm 0.0944444 0.1574074 0.0671349 0.2476799 No

> 160 cm 0.1638889 0.0833333 0.0148241 0.1518426 Yes Avoidance

ADULT CHUB DEPTH SELECTION-ALL SITES

Category Expected Observed Bonferroni Intervals Significant Avoidance Proportion Proportion /Preference

0-40 cm 0.460886 0.0471698 -0.005874 0.1002134 Yes Avoidance

41-80 cm 0.1489161 0.2075472 0.1060771 0.3090173 No

81-120 cm 0.1244109 0.3396226 0.2211311 0.4581142 Yes Preference

120-160cm 0.0961357 0.1981132 0.0983877 0.2978387 Yes Preference

> 160 cm 0.1696513 0.2075472 0.1060771 0.3090173 No

BEAUTIFUL SHINER IN OASIS POND

Category Expected Observed Bonferroni Intervals Significant Avoidance Proportion Proportion /Preference

0-40 cm 0.1954545 0.0454545 -0.068944 0.1598533 Yes Avoidance

41-80 cm 0.2409091 0.4545455 0.18108 0.7280109 No

81-120 cm 0.1909091 0.1818182 -0.030007 0.3936436 No

120-160cm 0.1454545 0.0909091 -0.066976 0.2487944 No

> 160 cm 0.2272727 0.2272727 -0.002883 0.4574282 No 84

SUBSTRATE SELECTIVITIES

TOPMINNOW—ALL SITES

Category Expected Observed Bonferroni Intervals Significant Avoidance Proportion Proportion /Preference

Open 0.1276024 0.2214765 0.1400378 0.3029152 Yes Preference

Detritus 0.1484218 0.2550336 0.169547 0.3405202 Yes Preference

Plant 0.7239758 0.5234899 0.4255361 0.6214437 Yes Avoidance

BEAUTIFUL SHINER IN OASIS POND

Category Expected Observed Bonferroni Intervals Significant Avoidance Proportion Proportion /Preference

Open 0.1 0.2272727 0.0270482 0.4274973 No

Plant 0.9 0.7727273 0.5725027 0.9729518 No

JUVENILE CHUB IN NORTH AND OASIS PONDS

Category Expected Observed Bonferroni Intervals Significant Avoidance Proportion Proportion /Preference

Open 0.2391304 0.4347826 0.2031382 0.666427 No

Plant 0.7608696 0.5652174 0.333573 0.7968618 No

ADULT CHUB IN NORTH AND OASIS PONDS

Category Expected Observed Bonferroni Intervals Significant Avoidance Proportion Proportion /Preference

Open 0.2285714 0.0357143 -0.042879 0.1143078 Yes Avoidance

Plant 0.7714286 0.9642857 0.8856922 1.0428792 Yes Preference 85

JUVENILE CHUB IN LESLIE CREEK

Category Expected Observed Bonferroni Intervals Significant Avoidance Proportion Proportion /Preference

Silt 0.3367174 0.6949153 0.5404979 0.8493326 Yes Preference

Sand 0.1201354 0.1016949 0.0003315 0.2030584 No

Large 0.0964467 0.0338983 -0.026792 0.0945887 Yes Avoidance Substrates

Detritus 0.1489002 0.1355932 0.0207784 0.250408 No

Plant 0.2978003 0.0338983 -0.026792 0.0945887 Yes Avoidance

ADULT CHUB IN LESLIE CREEK

Category Expected Observed Bonferroni Intervals Significant Avoidance Proportion Proportion /Preference

Silt 0.3307692 0.525641 0.3799956 0.6712864 Yes Preference

Sand 0 1153846 0.0512821 -0.013053 0.1156174 No

Large 0.1012821 0.1153846 0.0221988 0.2085704 No Substrates

Detritus 0.1602564 0.2948718 0.1618725 0.4278711 Yes Preference

Plant 0.2923077 0.0128205 -0.019993 0.0456338 Yes Avoidance

VELOCITY SELECTIVITIES FOR LESLIE CREEK

JUVENILE CHUB

Category Expected Observed Bonferroni Intervals Significant Avoidance (m/sec) Proportion Proportion /Preference

0 0.830508 0.830508 0.75566 0.90536 No

0.01-0.04 0.11168 0.135593 0.06729 0.20389 No

> 0.04 0.07953 0.033898 -0.0022 0.07 Yes Avoidance 86

ADULT CHUB

Category Expected Observed Bonferroni Intervals Significant Avoidance (m/sec) Proportion Proportion /Preference

0 0.8153846 0.8846154 0.7980134 0.9712174 No

0.01-0.04 0.1102564 0.1025641 0.0203254 0.1848028 No

> 0.04 0.074359 0.0128205 -0.017674 0.0433154 Yes Avoidance 87

LIST OF REFERENCES

Allendorf, F. W., and N. Ryman. 1987. Genetic management of hatchery stocks. Pages 141-159. In: N. Ryman and F Utter, Eds. Population Genetics and Fishery Management. University of Washington Press, Seattle, Washington.

Bain, M. B,. and J. T. Finn. 1991. Analysis of microhabitat use by fish: investigator effect and investigator bias. Rivers 2:57-65.

Boner, F. C., C. F. Smith, W. B. Garrett, and A. Komeczki. 1989. Water resources data. Arizona. United States Geological Survey, Water Data Report AZ-89-1.

Bovee, K. D. 1986. Development and evaluation of habitat suitability criteria for use in the instream flow incremental methodology. Instream Flow Information Paper No. 21. United States Fish and Wildlife Service. Biological Report 86(7).

Byers, C. R., R. K. Steinhorst, and P. R. Krausman. 1984. Clarification of a technique for analysis of utilization-availability data. Journal of Wildlife Management 48:1050-1053.

Carpenter, J,, and O. E. Maughan. 1993. Macrohabitat of Sonora Chub (Gila ditaenja) in Sycamore Creek, Santa Cruz County, Arizona. Journal of Freshwater Ecology 8:265-277.

Carson, R. G., and J. M. Peek. 1987. Mule deer habitat selection patterns in northcentral Washington. Journal of Wildlife Management 51:46-51.

Collins, J.P., C. Young, J. Howell, and W. L. Minckley. 1981. Impact of flooding in a Sonoran Desert stream, including elimination of an endangered fish population (Poeciliopsis o. occidentalis. Poeciliidae). The Southwestern Naturalist 26:415-423.

Conover, W. J. 1971. Practical nonparametric statistics. John Wiley and Sons, New York, New York.

Cunjack, R. A., and J. M. Green. 1983. Habitat utilization by brook char (Salvelinus fontinalis) and rainbow trout (Salmo gairdneri) in Newfoundland streams. Canadian Journal of Zoology 61:1214-1219.

Demarais, B. D. 1991. Gila eremica. a new cyprinid fish from northwestern Sonora, Mexico. Copeia 1:178-189. 88

Demarais, B. D., and W. L. Minckley. 1993. Genetics and morphology of Yaqui chub Gila purpurea, an endangered Cyprinid fish subject to recovery efforts. Biological Conservation 6:195-205.

Devore, P. W., and R. J. White. 1978. Daytime response of brown trout Salmo trutta to cover stimuli in stream channels. Transactions of the American Fisheries Society 107:763-771.

Dunn, O. J. 1964. Multiple comparisons using rank sums. Technometrics 6:241- 252.

Echelle, A. A. 1991. Conservation genetics and genie diversity in threatened fishes of western North America. Pages 141-153. Itr. W. L. Minckley and J. E. Deacon, Eds. Battle Against Extinction: Native Fish Management in the American West. University of Arizona Press, Tucson, Arizona.

Fausch, K. D., and J. R. White. 1981 Competition between brook trout (Salvelinus fontinalis) and brown trout (Salmo trutta) for positions in a Michigan stream. Canadian Journal of Fisheries and Aquatic Sciences 38:1220-1227.

Forrest, R. E. 1992. Habitat use and preference of Gila topminnow. Master's Thesis. The University of Arizona, Tucson, Arizona. 84 pp.

Galat, D. L., and B. Robertson. 1992. Response of endangered Poeciliopsis occidentalis sonoriensis in the Rio Yaqui drainage, Arizona, to introduced Gambusia affinis. Environmental Biology of Fishes 33:249-264.

Gorman, O. T. 1987. Habitat segregation in an assemblage of minnows in an Ozark stream. Iir. W. J. Matthews, and D. C. Heirs, Eds. Community and Evolutionary Ecology of North American Stream Fishes. University of Oklahoma Press, Norman, Oklahoma. 310 pp.

Gorman, O. T. 1988. An experimental study of habitat use in an assemblage of Ozark minnows. Ecology 69:1239-1250.

Griffith, J. S. 1981. Estimation of the age-frequency distribution of stream dwelling trout by underwater observation. Progressive Fish-Culturist 43:51-53.

Griffith, J. S., and R. K. Fuller. 1979. Ability of underwater observers to estimate size of stream-dwelling salmonid fishes. Pages 69-70. In: O. D. Markham, and W. J. Arthur, Eds. Proceedings on a Symposium of the Idaho National Engineering Laboratory and Ecology Program. United States Department of Energy, IDO-12088. Jackson, Wyoming. 89

Hastings, J. R. 1959. Vegetation change and arroyo cutting in southeastern Arizona. Journal of the Arizona Academy of Science 1:60-67.

Hastings, J. R. and R. M. Turner. 1966. The changing mile. University of Arizona Press, Tucson, Arizona.

Heggenes, J. 1990. Habitat utilization and preferences in juvenile Atlantic salmon (Salmo salar) in streams. Regulated Rivers Research and Management 5:341-354.

Heggenes, J., A. Brabrand, and S. J. Saltveit. 1990. Comparison of three methods for studies of stream habitat use by young brown trout and Atlantic salmon. Transactions of the American Fisheries Society 119:101-111.

Helfman, G. S. 1985. Underwater methods. Pages 349-370. In: L. A. Neilson and D. L Johnson, Eds. Fisheries techniques. American Fisheries Society. Bethesda, Maryland.

Hendrickson, D. A., W. L. Minckley, R. R. Miller, D. J. Siebert, and P. H. Minckley. 1980. Fishes of the Rio Yaqui, Mexico and United States. Journal of the Arizona-Nevada Academy of Sciences 15:65-106.

Hindar, K., N. Ryman, and F. Utter. 1991. Genetic effects of cultured fish on natural fish populations. Canadian Journal of Fisheries and Aquatic Sciences 48:945- 957.

James, P. W. 1989. Reproductive ecology and habitat preference of the leopard darter, Percina pantherina. Ph.D. dissertation. Oklahoma State University. Stillwater, Oklahoma. 169 pp.

James, P. W., and O. E. Maughan. 1989. Spawning and habitat use of the threatened leopard darter, Percina pantherina. Southwestern Naturalist 34:298-301.

Jones, R. N., D. J. Orth, and O. E. Maughan. 1984. Abundance and preferred habitat of the leopard darter, Percina pantherina (Moore and Reeves), in Glover Creek, Oklahoma. Copeia 2:378-384.

Johnson, D. H. 1980. The comparison of usage and availability measurements for evaluating resource preference. Ecology 61:65-71.

Keenleyside, M. H. A. 1962. Skin diving observations of Atlantic salmon and brook trout in the Miramichi River, New Brunswick. Journal of the Fisheries Research Board of Canada 19:625-634. 90

Kelsh, S. W., and M. J. Baca. 1991. Biochemical analysis of gene products in the Yaqui catfish, lctalurus pricei. United States Fish and Wildlife Service, Dexter National Fish Hatchery.

Kepner, W. G. 1988. San Bernardino National Wildlife Refuge contaminant study. U. S. Fish and Wildlife Service, Ecological Services, Phoenix Field Office. Environmental Contaminants Program. 18 pp.

King, D. L., and J. W. Eibler. 1989. Applicability of the instream flow incremental methodology to a Michigan trout stream. Study No. 615. Project No. F-35-R- 14. Michigan Department of Natural Resources, Lansing, Michigan.

Lewis, S. L. 1969. Physical factors influencing fish populations in pools of a trout stream. Transactions of the American Fisheries Society 98:14-19.

Longsworth, S. A. 1991. Geohydrology and chemical quality of ground water, San Bernardino National Wildlife Refuge, Arizona. U. S. Geological Survey, Water-Resources Investigations Report 90-4190. 28pp.

Marcum, C. L., D. O. Loftsgaarden. 1980. A nonmapping technique for studying habitat preferences. Journal of Wildlife Management 44:963-968.

Matthews, W. J., and L. G. Hill. 1979. Influence of physico-chemical factors on habitat selection by red shiners, Notropis lutrensis (Pisces: Cyprinidae). Copeia 1979:70-81.

McNatt, R. M. 1974. Re-evaluation of the native fishes of the Rio Yaqui in the United States. Proceedings of the 54th Annual Western Association of State Game and Fish Commissioners 1974:273-279.

Meffe, G. K. 1983. Ecology of species replacement in the Sonoran topminnow (Poeciliopsis occidentalism and the mosquitofish (Gambusia affinis) in Arizona. Ph.D.Dissertation. Arizona State University, Tempe, Arizona. 143 pp.

Meffe, G. K. 1987. Conserving fish genomes: Philosophies and practices. Environmental Biology of Fishes 18:3-9.

Meffe, G. K., and F. F. Snelson, Jr. 1989. An ecological overview of Poeciliid fishes. Pages 13-31. In: G. K. Meffe, and F. F. Snelson, Jr. Ecology and Evolution of Livebearing Fishes (Poeciliidae). Prentice Hall. Englewood Cliffs, New Jersey. 453 pp. 91

Meffe, G. K., and R. C. Vrijenhoek. 1988. Conservation genetics in the management of desert fishes. Conservation Biology 2:157-169.

Minckley, W. L. 1973. Fishes of Arizona. Arizona Game and Fish Department. Phoenix, Arizona. 293 pp.

Minckley, W. L. 1980. Gila purpurea (Girard), Yaqui chub. Page 171. In: P. S. Lee, C. R. Gilbert, C. H. Hocutt, R. E. Jenkins, D. E. McAllister, and J. R. Stauffer, Eds. Fishes of North America, North Carolina State Museum of Natural History. Raleigh, North Carolina.

Minckley, W. L., and W. E. Barber. 1971. Some aspects of biology of the longfin dace, a cyprinid fish characteristic of streams in the Sonoran desert. The Southwestern Naturalist 15:459-464.

Minckley, W. L., and J. E. Deacon. 1991. Battle Against Extinction: Native Fish Management in the American West. The University of Arizona Press. Tucson, Arizona. 517 pp.

Moyle, P. B., and D. M. Baltz. 1985. Microhabitat use by an assemblage of California stream fishes: developing criteria for instream flow determinations. Transactions of the American Fisheries Society 114:695-704.

Moyle, P. B., and J. J. Cecil. 1988. Fishes: An Introduction to Ichthyology. Prentice-Hall, Inc., Englewood Cliffs, New Jersey. 559 pp.

Ordway, L. L., and P. R. Krausman. 1986. Habitat use by desert mule deer. Journal of Wildlife Management 50:677-683.

Orth, D. L. 1985. Aquatic habitat measurements. Pages 61-84. In: L. A. Neilson and D. L Johnson, Eds. Fisheries Techniques. American Fisheries Society. Bethesda, Maryland.

Platts, W. S., W. F. Megahan, and G. W. Minshall. 1983. Methods for evaluating stream, riparian, and biotic conditions. General Technical Report INT-138. United States Department of Agriculture, Forest Service Intermoutain Forest and Range Experiment Station, Ogden, Utah. 70 pp.

Rinne, J. N. 1992. Physical habitat utilization of fish in a Sonoran Desert stream, Arizona, southwestern United States. Ecology of Freshwater Fish 1992:35- 41.

Rinne, J. N., and W. L. Minckley. 1991. Native fishes of arid lands, a dwindling 92

resource of the desert Southwest. General Technical Report INT-206. U.S. Department of Agriculture, Forest Service, Rocky Mountain Forest and Range Experiment Station, Ft. Collins, Colorado. 45 pp.

Rutter, C. 1896. Notes on the fresh water fishes of the Pacific slope of North America. Proceedings of the California Academy of Sciences 6:245-267.

Shirvell, C. S., and R. G. Dungey. 1983. Microhabitat chosen by brown trout for feeding and spawning in rivers. Transactions of the American Fisheries Society 1 12:355-367.

Silvey, W. 1975. Fishes of Leslie Creek, Cochise County, Arizona. Statewide Fisheries Investigations, Project F-7-R-17, Job No. VI, Seg. 2, Arizona Game and Fish Department, Phoenix, Arizona. 11 pp.

Springer, C. L. 1993. Temporal variation in habitat use by smallmouth bass in the Black River, Arizona. Master's Thesis. The University of Arizona, Tucson, Arizona. 102 pp.

Stefferud, S. E. 1984. Recovery Plan for Gila and Yaqui topminnow Poeciliopsis occidentalis (Baird and Girard). United States Fish and Wildlife Service, Albuquerque, New Mexico. 67 pp.

Sublette, J. E., M. D. Hatch, and M. Sublette. 1990. The Fishes of New Mexico. University of New Mexico Press, Albuquerque, New Mexico. 393 pp.

Thomas, D. L., and E. J. Taylor. 1990. Study designs and tests for comparing resource use and availability. Journal of Wildlife Management 54(2):322-330.

United States Department of the Interior (USDI). 1967. Endangered and threatened wildlife and plants. Federal Register 32(48):4001.

United States Department of the Interior (USDI). 1984. Endangered and threatened wildlife and plants. Federal Register 49(171):34490-34497.

United States Fish and Wildlife Service (USFWS). 1983. Gila and Yaqui topminnow recovery plan. U.S. Fish and Wildlife Service, Albuquerque, New Mexico. 56 PP

United States Fish and Wildlife Service (USFWS). 1987. San Bernardino National Wildlife Refuge management plan. U.S. Fish and Wildlife Service, Albuquerque, New Mexico. 112 pp. 93

United States Fish and Wildlife Service (USFWS). 1990. Report to Congress: endangered and threatened species recovery program. U.S. Department of the Interior, Fish and Wildlife Service. Washington, D.C. 406 pp.

Vrijenhoek, R. C., M. E. Douglas, and G. K. Meffe. 1985. Conservation genetics of endangered fish populations in Arizona. Science 229:400-402.

Wetzel, R. G. 1975. Limnology: Second Edition. Saunders College Publishing, Philadelphia, Pennsylvania. 767 pp.

Williams, J. E., J. E. Johnson, D. A. Hendrickson, S. Contreras-Balderas, J. D. Williams, M. Navarro-Mendoza, D. E. McAllister, and J. E. Deacon. 1989. Fishes of North America, endangered, threatened, and of special concern. Fisheries 14:2-20.

Williams, J. E. 1991. Preserves and refuges for native Western fishes: history and management. Pages 171-189. In: W. L. Minckley, and J. E. Deacon, Eds. Battle Against Extinction. The University of Arizona Press. Tucson, Arizona. 517 pp.