CALIFORNIA
FISH™ GAME

"CONSERVATION OF WILD LIFE THROUGH EDUCATION"

California Fish and Game is published quarterly by the California Department of
Fish and Game. It is a journal devoted to the conservation and understanding of fish
and wildlife. If its contents are reproduced elsewhere, the authors and the California
Department of Fish and Game would appreciate being acknowledged.

Subscriptions may be obtained at the rate of $10 per year by placing an order with
the Editor, California Department of Fish and Game, 1416 Ninth Street, Sacramento.
CA 95814. Checks or money orders in U.S. dollars should be made out to: California
Fish and Game. Inquiries regarding paid subscriptions should be directed to the
Editor. Complimentary subscriptions are granted on an exchange basis.

California Department of Fish and Game employees may request complimentary
subscriptions to the journal.

Please direct correspondence to:

Dr. Eric R. Loft, Editor-in-Chief
California Fish and Game
1416 Ninth Street
Sacramento, California 95814

RESOURCES AGENCY

CALIFORNIA

DEPARTMENT

u

1

]

VOLUME 79

SUMMER 1993

NUMBER 3

Published Quarterly by

STATE OF CALIFORNIA

THE RESOURCES AGENCY

DEPARTMENT OF FISH AND GAME

-LDA--

STATE OF CALIFORNIA
PETE WILSON, Governor

THE RESOURCES AGENCY
DOUGLAS P. WHEELER, Secretary for Resources

FISH AND GAME COMMISSION

Benjamin F. Biaggini, President San Francisco

Albert C. Taucher, Member Long Beach

Frank D. Boren, Member Carpinteria

Gus Owen, Member Dana Point

Robert R. Treanor, Executive Director

DEPARTMENT OF FISH AND GAME

BOYD GIBBONS, Director

John H. Sullivan, Chief Deputy Director

Al Petrovich Jr., Deputy Director

Banky E. Curtis, Deputy Director

H.K. (Pete) Chadwick, Chief Bay-Delta Division

Rolf Mall, Chief Marine Resources Division

Tim Farley, Chief Inland Fisheries Division

Terry Mansfield, Chief Wildlife Management Division

John Turner, Chief Environmental Services Division

Susan A. Cochrane, Chief Natural Heritage Division

DeWayne Johnston, Chief Wildlife Protection Division

Richard Elliott, Regional Manager Redding

Ryan Broddrick, Regional Manager Rancho Cordova

Brian F. Hunter, Regional Manager Yountville

George D. Nokes, Regional Manager Fresno

Fred Worthley, Regional Manager Long Beach

CALIFORNIA FISH AND GAME
1993 EDITORIAL STAFF

Eric R. Loft, Editor-in-Chief Wildlife Management

Jack Hanson, Betsy C. Bolster, Ralph Carpenter,

Arthur C. Knutson, Jr Inland Fisheries

Dan Yparraguirre, Douglas R Updike Wildlife Management

Steve Crooke, Doyle Hanan, Jerome D. Spratt Marine Resources

Donald E. Stevens Bay-Delta

Peter T. Phillips, Richard L. Callas Environmental Services

CONTENTS

Food Habits of the Brown Smoothhound Shark {Mustelus Henlei)

From Two Sites in Tomales Bay Steven L Haeseker

and Joseph J. Cech, Jr. 89

Distribution, Ecology, and Status of the Fishes of the San Joaquin

River Drainage, California Larry R. Brown and

Peter B. Moyle 96

Fire Effects on a Montane Sierra Nevada Meadow Robert S.

Boyd, Roy A. Woodward, and Gary Walter 115

A Sampler for Quantifying the Vertical Distribution of

Macroinvertebrates in Shallow Wetlands Jeffrey Mackay

and Ned H. Euliss, Jr. 126

Tule Elk Relocated to Brushy Mountain, Mendocino County,
California Kent B. Livezey 131

BACK ISSUES WANTED! If you have been collecting California Fish and Game for
a long time, or have acquired old copies of issues, and don't necessarily want to keep
them, please consider offering them back to us! There are several avid collectors who
are trying to put together a complete set of the journal dating back to 1 9 1 4. If you would
like to contribute, please contact the editor.

CALIFORNIA FISH AND GAME
Calif. Fish and Game 79(3):89-95 1 993

FOOD HABITS OF THE BROWN SMOOTHHOUND
SHARK (MUSTELUS HENLEI) FROM TWO SITES IN

TOMALES BAY

STEVEN L HAESEKER1 and JOSEPH J. CECH, JR.

Department of Wildlife and Fisheries Biology

University of California

Davis, California 95616

Brown smoothhound sharks were non-destructively sampled for
stomach contents at two locations in Tomales Bay. Sharks near Hog
Island consumed more slender crabs and fish, but fewer yellow shore
crabs and polychaete worms than those near Indian Beach. Thus, brown
smoothhound shark food habits are not homogeneous throughout the
Bay, as was implied from previous studies using shark derby-caught fish.
Survival of sharks from stomach eversion was > 85%.

INTRODUCTION

The brown smoothhound {Mustelus henlei) is a common, bottom-oriented shark
in shallow waters off California. Although it has been found from the Gulf of
California to Humboldt Bay (Miller and Lea 1972), it is most common inside the bays
and estuaries of Northern California such as San Francisco and Tomales bays (Love
1991). This shark has a subterminal mouth and small, coarse teeth, and its main food
items are crabs, shrimp, polychaete worms, and fish (Karl 1979, Russo 1975).
Previous diet studies in Tomales Bay used specimens obtained from shark fishing
derbies, which were not site-specific within the bay.

Numerous methods have been developed for removal of stomach contents without
harming fish. The simplest of these is merely removing the contents with a forceps
(Wales 1962). Another method uses water to flush contents out of the stomach
(Seaburg 1 957). The stomach flushing technique is efficient with high survival offish
after treatment, but efficiency is negatively correlated with size (Meehan and Miller
1978). The ability of sharks to evert their stomachs under stress and survive is well
known (Russo 1975, Springer and Gold 1989). Researchers have used this ability to
develop a stomach eversion technique for examining the diet and feeding habits of
lemon sharks (Negaphon brevirostris) (Cortes and Gruber 1 990) and grey smoothhound
sharks (Mustelus californicus) (SanFilippo 1992). With this technique, the stomach
and its contents are manually pulled out through the esophagus and mouth.

The purpose of this study was to examine the food habits of the brown smoothhound
sharks captured with short duration gill-net sets at two different locations in Tomales
Bay, using a stomach eversion technique. This method is non-destructive to the sharks
but allows for complete sampling of the stomach contents, as has been shown for gray

1 Present Address: North Carolina Cooperative Fish and Wildlife Research Unit, Box 7617, Department
of Zoology, North Carolina State University, Raleigh, NC 27695-7617.

89

90 CALIFORNIA FISH AND GAME

smoothhound sharks (SanFilippo 1992). In addition, gill-net placement allows
investigation of site-specific food habits.

METHODS
Fish Collection

Fish were collected during 7-16 July 1992 at two locations in Tomales Bay. The
first was in the outer bay region (Smith et al. 1991) near Hog Island at a depth of 4-
6 m, and the second was in the inner bay region near Indian Beach at a depth of 3-6
m (Fig. 1). Two gill-nets, 3 m high x 100 m long with 6 cm mesh were deployed at
each site by boat and allowed to set for 30 min. Captured fish were removed from the
net and placed into a seawater-filled, insulated cooler where they were immediately
anesthetized using dissolved CO, (Summerfelt and Smith 1990).

Stomach Eversion

After approximately 10 min in the CO,-seawater bath, the fish were removed,
sexed, measured to the nearest cm, and held caudal end up. Stomach tongs (20 or 45
cm length) were gently inserted through the esophagus and closed onto a portion of
the stomach near the duodenum. The tongs were then slowly withdrawn, everting the
stomach and releasing the stomach contents onto a 1-mm stiff plastic mesh situated
over a bucket. After collection of the stomach contents, the shark was then held snout
end up which allowed the stomach to slide down the esophagus and be swallowed
back to normal position. The food items were sealed in a plastic bag. The shark was
then marked by punching a small hole in the posterior region of the dorsal fin, revived
with a fresh seawater flow over the gills, and released. Sharks were marked to prevent
multiple sampling of the same fish.

Prey Analysis

Stomach contents were analyzed at Bodega Marine Laboratory (BML) using a
dissecting microscope and taxonomic keys (Morris et al. 1980, Smith and Carlton
1 975, Miller and Lea 1972). Frequency of the items within the stomach (Hyslop 1 980)
and the carapace width (mm) for crabs were recorded. Shark length and prey data were
compared using regression and Mest analyses (Sokal and Rohlf 1981).

Survivorship

For survivorship studies, 7 sharks were transported to BML in oxygenated,
seawater-filled coolers. After 12-30 h acclimation in seawater tanks, shark stomachs
were everted. Contents of this eversion with those from field-sampled sharks were
used in the study. Following a 36 h recovery, stomachs were everted again. After 24-
48 h subsequent recovery, surviving sharks were released into the wild. No control
group was used because of time and space constraints.

FOOD HABITS OF BROWN SMOOTHHOUND SHARK

91

^.Dillon
t;. Beach

V

■; : •:. \

Pacific

Ocean

Site#2

4 N 38° 6'

W 122° 54'

San Francisco
40 km I

O

7"

Figure 1. Map of Tomales Bay, on the northern California coast, showing sampling sites #1 (Hog Island)
and #2 (Indian Beach).

92 CALIFORNIA FISH AND GAME

RESULTS

Fifty-four brown smoothhound sharks were captured and stomach contents
collected by stomach eversion. The sharks ranged from 67 to 92 cm total length, and
there was no difference in mean fish length between sampling sites (Mest, P > 0.05).
Only three sharks had empty stomachs. Stomach contents analysis indicated some
dietary differences between the collection sites. At the Hog Island site, 75% of the
sharks contained slender crabs (Cancer gracilis) while 25% contained yellow shore
crabs (Hemigrapsus oregonensis). However at the Indian Beach site, 91% contained
H. oregonensis and only 12% contained Cancer gracilis (Table 1). Sharks sampled
from the Hog Island site contained a higher (/-test, P < 0.05) number of C. gracilis
while the number of ingested H. oregonensis was higher (/-test, P < 0.05) at Indian
Beach. Pea crabs (Scleroplax granulata) and kelp crabs (Pugettia producta) were
only rarely found in the stomach contents.

Bay shrimp (Crangon stylirostris) also constituted an important part of the
smoothhound diet. No significant difference (Mest, P > 0.05) in the mean number of
C. stylirostris existed between the two sampling sites. However, the mean was
somewhat greater at the Indian Beach location. Other shrimp such as the bay ghost
shrimp (Callianassa californiensis) and the blue mud shrimp (Upogebia pugettensis)
were rarely found and did not show significant differences (Mest, P > 0.05) between
sites.

Of the sharks captured, 27% of the stomachs contained fish. Highly digested
remains made species determination difficult, and thus a large proportion offish are
recorded under the "Unidentified" category. However the brown smoothhounds
captured at Hog Island contained a significantly higher mean number of total fish than
the Indian Beach individuals. In addition, the mean number of staghorn sculpins
(Leptocottus armatus) was higher (Mest, P < 0.05) at Hog Island.

Marine worms were also highly digested and could only be classified under the
title "Polychaeta." The mean number of polychaetes differed between the two
sampling locations with Indian Beach showing a higher (/-test, P < 0.05) mean. The
mean number of eelgrass {Zostera marina) blades ingested was higher (/-test, P <
0.05) at Hog Island than at the Indian Beach site.

Although the prey item proportions differed between sampled sites, the sharks
maintained a similar diversity of prey and feeding level. The mean number of
identified species between locations as well as the mean number of items between
locations showed no significant (/-test, P > 0.05) differences. Regressions comparing
shark length with food item size, number of food items, and the number of different
species present in the stomach indicated no significant (P > 0.05) relationships.

The stomach eversion technique for sampling stomach contents resulted in little
mortality, with 6 of 7 brown smoothhound sharks surviving two stomach eversions
within 36 hours. Although reviving the sharks after the procedure typically took 2-5
min, the sharks demonstrated no ill-effects afterward. The one mortality that occurred
took place within the first 24 hours after arrival at the laboratory, quite possibly from
transportation-related stress.

FOOD HABITS OF BROWN SMOOTHHOUND SHARK

93

O"
u

Os o co —

CO —

00 — CM

CM CO — 00

SO CM — CO

73

—

O

f-

^ ^ ^ ^

^ ^ ^

#

£ £ *£ £

£ ^ 5R ^

N IT) vO d

CO CM ^fr

^c

rt SO CM I/O

— rf CM SO

r— co

CO

ri

— ^

~-

E
u

00 CO
— 00

©

n

O

CO CM

SO o

t

c

CO
CO

©

©

CI

O

r-

CO Tt CM

— o o

<

cm O

©

d

d d

d

d

d

d

d

d

odd

z

-f

ri

3*

a-

— i Tf CM

co

— ^ o —

ri

CM O O co

O CM O O

73

E

73

u

OQ

EA

-
E

c
o

■ —

c

—

o

E

c

y.

-«:
u

73

x:

—
c

a

o

c
o
E

v-

=

1

—
c
■/-.

E
u

—

I

u

73
U

PQ

c

73

c

o

u.

c

3

x:

CO

E

—

la

OJ

CO

E

CD

^ ^ ^ ^

^ £ ^

S

^ ^ ^ t£

^ ^ ^ ^

— CM SO CO

— O co

«/-,

so O O On

O SO o o

On — '

«*

CO

00 -h >0
— < CM |p

CO

o

3C

O CO

o o

r-

-r

so o

o o

O 0">

o o

Q
O

o

©

o

<

O 00

q —

co d d

d

d

d d

d

d d

d d

d

d

d

z

U0 CM

i/-> i/i — O

CO

73
1—

u

*
=:

?

5 S

b- ft.

vj to

ft,
a

TT — —

r\

O CO — W0

~ -S3

6a

2 ■

3

t0

=:

*-.
*—

8

ft,

a

2

O 92

a, -5 U O 13 i £

S

s
a

SO O — CO

eP

^ ^ ^

^ ^ ^

#

#

c^ ^

5^

^ ^ ^ ^

>n

m in o

O >n >n

©

o

>n m

i/",

O O v> >/"}

rj

r^

CM

—

—

CM

co —

O

o

m

o

o

O

CO

in

O

©

©

©

IT)

c

©

CO

CO

o
©

in

o

<

d

—

d

d

d

d

©

d

©

d

d

d

©

d

d

Z

•n m

00 CO

co cm'

t»3 C

a ^

a § S

rs *«• •*— »

2 a c

o s: <u v)

r* ST- ^5 C bl

.sf 5^

>.

*

a

,c

C

g

S
a

u.

U

-^

E

u

u

u

CL

^~-

wa

4-

•*-

0

0

u

u.

U

u

X

X

E

E

3

3

z

z

=

B

73

4>

U

S S

-o

C

—

73

u

_c

c
u
u

4-1

u

-C

u

E

3
C

u

E

c

u
u

c

c

73
'J

e

c/5

94 CALIFORNIA FISH AND GAME

DISCUSSION

Past food habit studies on Tomales Bay brown smoothhound sharks relied upon
shark derbies for specimens for stomach content analysis (Russo 1975, Karl 1979).
Although many individuals can be captured per day, collecting sites are frequently not
well documented. Dietary sampling biases may also be expected from angled
specimens which may be hungry and willing to consume bait or artificial lures. Our
fish were sampled at two known sites by gill-nets and should better represent the food
habits of the population at these locations. We found that food habits of the brown
smoothhound shark were not homogeneous within Tomales Bay, contrary to previous
studies (Russo 1975, Karl 1979).

We found similar diet items, but some significant differences in brown smoothhound
sharks' species selection between sampling locations in Tomales Bay (Table 1).
Because of the similar temperature conditions throughout the bay during summer, the
sharks would be expected to have similar metabolic demands. The differences in prey
species composition between sites may result from different salinities and/or habitats
found in Tomales Bay.

The outer bay containing the Hog Island site is basically oceanic habitat with
oceanic salinities, while the inner bay containing the Indian Beach site is a bay habitat
demonstrating bay salinities (Smith et al. 1991). Cancer gracilis is intolerant of
brackish waters and its distribution would be restricted to the oceanic conditions of
the Hog Island site (Morris et al. 1 980). Hemigrapsus oregonensis is more euryhaline,
tolerating both hypersaline and brackish waters (Morris et al. 1980) and would
expectedly be abundant at the Indian Beach site. The observed differences in brown
smoothhound shark diet may result from a differential abundance in prey species
between locations. Our regression analysis showed that the sampled sharks
demonstrated no prey size selectivity with regard to shark length (P > 0.05), possibly
a function of the relatively small size range sampled with the gill-nets.

ACKNOWLEDGMENTS

We thank Richard SanFilippo for shark stomach eversion instruction and Scott
Matern and Jose Setka for their help in collecting and everting the sharks. Paul Siri,
laboratory manager, and his staff facilitated our efforts at the BML. Laura Brink drew
the figure, and Peter Moyle kindly criticized the manuscript.

LITERATURE CITED

Cortes, E., and S. Gruber. 1990. Diet, feeding habits and estimates of daily ration of young
lemon sharks, Negaprion brevirostris (Poey). Copeia 1990:204-218.

Hyslop, E.J. 1 980. Stomach contents analysis- a review of methods and their application. J. Fish
Biol. 17:411-429.

Karl, S.R. 1979. Fish feeding-habit studies from Tomales Bay, California. M.S. thesis, Univ.
Pacific. 44 p.

FOOD HABITS OF BROWN SMOOTHHOUND SHARK 95

Love, R.M. 1991. Probably more than you want to know about the fishes of the Pacific Coast.

Really Big Press. Santa Barbara, CA. 215 p.
Meehan, W.R., and R.A. Miller. 1978. Stomach flushing: effectiveness and influence on

survival and condition of juvenile salmonids. J. Fish. Res. Bd. Can. 35:1359-1363.
Miller, D.J., and R.N. Lea. 1972. Guide to coastal marine fishes of California. Calif. Dept. Fish

and Game, Calif. Fish Bull. 157:1-249.
Morris, R.H., D.P. Abbott, and E.C. Haderlie. 1980. Intertidal invertebrates of California.

Stanford Univ. Press. Stanford, CA. 690 p.
Russo, R.A. 1 975. Observations on the food habits of leopard sharks (Triakis semifasciata) and

brown smoothhounds (Mustelus henlei). Calif. Fish and Game 61:95-103.
SanFilippo, R. 1992. Diet and gastric evacuation of gray smoothhound sharks, Mustelus

californicus, from Elkhorn Slough National Research Reserve, Monterey Bay, California.

Proc. Eighth Ann. Meeting Amer. Elasmobranch Soc. 161. Abstract.
Seaburg, K.G. 1957. A stomach sampler for live fish. Prog. Fish-Cult. 19:137-139.
Smith, R.I., and J.T. Carlton. 1975. Light's Manual, 3rd ed. University of California Press.

Berkeley, CA. 721 p.
Smith, S.V., J.T. Hollibaugh, S.J. Dollar, and S. Vink. 1991. Tomales Bay metabolism: C-N-

P stoichiometry and ecosystem heterotrophy at the land-sea interface. Est. Coast. Shelf Sci.

33:223-257.
Sokal, R.R. and F.J. Rohlf. 1 98 1 . Biometry , The principles and practice of statistics in biological

research, 2nd ed. W.H. Freeman and Co. NY.
Springer, V.G. and J. P. Gold. 1989. Sharks in question. Smithsonian Institution Press.

Washington, D.C.
Summerfelt, R.C. and L.S. Smith. 1990. Anesthesia, surgery, and related techniques. Pages

213-272 in C.B. Schreck and P.B. Moyle, eds. Methods for fish biology. American

Fisheries Society, Bethesda, Maryland.
Wales, J.H. 1962. Forceps for removal of trout stomach content. Prog. Fish-Cult. 24:171.

Received: 18 February 1993
Accepted: 26 July 1993

CALIFORNIA FISH AND GAME
Calif. Fish and Game 79(3):96-1 14 1 993

DISTRIBUTION, ECOLOGY, AND STATUS OF THE FISHES
OF THE SAN JOAQUIN RIVER DRAINAGE, CALIFORNIA

LARRY R. BROWN1 AND PETER B. MOYLE

Department of Wildlife and Fisheries Biology

University of California

Davis, CA95616

In 1985 and 1986 we sampled streams of the San Joaquin River
drainage in south-central California. The purposes of the survey were: (i)
to see if further declines in native fish populations had occurred since
1 970 when they were last surveyed; (ii) to verify the species distributions
and species-habitat relationships observed in previous studies; (iii) to
verify the species assemblages observed in previous studies; and (iv) to
determine the status of the recently described Kern brook lamprey
(Lampetra hubbsi). We also reviewed the status of the native fish fauna
as compared to pre-European times. Only 1 1 species of the original fauna
of 19 species were found and only 6 of the 11 were common. Hardhead
{Mylopharodon conocephalus) and hitch {Lavinia exilicauda) were found
in fewer localities than in 1970. The decline in hardhead was associated
with an expansion of smallmouth bass (Micropterus dolomieu)
populations. Three assemblages of native species were identified, in
agreement with earlier studies. A fourth assemblage identified in earlier
studies, composed largely of introduced fishes, was divided into two
subgroups on the basis of our analysis. Each assemblage of species was
associated with a distinct set of habitat characteristics. Populations of
Kern brook lamprey were found in the Kaweah, Kings, San Joaquin, and
Merced rivers. Lampreys were absent from the lower reaches of the rivers
and, except in the Kings River, were only found below major dams. In the
Kings River lampreys were captured above and below Pine Flat Reservoir.
Because the populations are restricted in range, effectively isolated from
one another, and all but one can be affected by reservoir operations,
special protection for them is warranted.

INTRODUCTION

About half of California's water flows through the Sacramento-San Joaquin
drainage basin, which includes the Central Valley (Karhl et al. 1978). Despite the
large size of the drainage, only 34 species of freshwater or anadromous fish are native
to it, 17 of them endemic (Moyle 1976, Moyle and Williams 1990). The San Joaquin
basin is the most southern and most arid portion of the Sacramento-San Joaquin
drainage and historically contained 19 of the 34 species (Table 1), including 12 of the
1 7 endemic forms (Moyle 1 976). The Kern brook lamprey (Lampetra hubbsi) is found
only in the San Joaquin River drainage (Vladykov and Kott 1976), as are three
subspecies of rainbow trout (Oncorhynchus mykiss whitei, O.m. aquabonita, and 0.

'Present adress: U.S. Geological Survey, 2800 Cottage Way, Rm. W-2234, Sacramento, CA 95825

96

SAN JOAQUIN FISHES 97

m. gilberti) (Moyle 1976, Berg 1987). In addition, the California roach (Lavinia
symmetricus) appears to have a number of distinctive populations in the drainage,
although their taxonomic status has not been fully determined (Brown et al. 1992).

Because the San Joaquin Valley is intensively farmed, most of the water that once
flowed into the San Joaquin River or into large lakes on the valley floor, has been
diverted for irrigation (Karhl et al. 1978). The valley lakes have been drained and
converted to farmland. All major streams entering the San Joaquin Valley have been
dammed. Additional development of the limited water remaining is taking place in
response to the rapid growth of human populations in the region, especially in the
vicinity of the cities of Modesto, Fresno, and Bakersfield.

Not surprisingly, the native fish fauna of the region is in decline (Brown and
Moylel992), as is the fauna of the entire state (Moyle and Williams 1990). In 1970,
Moyle and Nichols ( 1 973, 1 974) surveyed the fish fauna of the foothill streams on the
east side of the San Joaquin Valley between elevations of 90 and 1 100 m and found
evidence of considerable decline in the distribution and abundance of the native
fishes. They only surveyed a relatively limited portion of the entire drainage because
of time constraints and because preliminary surveys indicated that native fishes were

Table 1. Native fishes of San Joaquin River Drainage, California.

Percentagea

Species 1986 1970 Status"

Pacific lamprey, Lampetra tridentata 6 0 D?

Kern brook lamprey, Lampetra hubbsi 5 0 D

White sturgeon, Acipenser transmontarws 0 0 R

Delta smelt, Hypomesus transpacificus 0 0 R

Chinook salmon, Oncorhynchus tshawytscha 0 0 D

Rainbow trout, Oncorhynchus mykiss 32 20 C

Thicktail chub, Gila crassicauda 0 0 E

Splittail, Pogonichthys macrolepidotus 0 0 R

Sacramento blackfish, Orthodon microlepidotus 0 0 C

Hitch, Lavinia exilicauda 5 10 D

California roach, Lavinia symmetricus 27 32 D

Sacramento squawfish, Ptychocheilus grandis 31 38 C

Hardhead, Mylopharodon conocephalus 7 9 D

Sacramento sucker, Catostomus occidentalis 48 42 C

Prickly sculpin, Cottus asper 7 2 C

Riffle sculpin, Cottus gulosus 4 2 D

Threespine stickleback, Gasterosteus aculeatus 3 1 D

Sacramento perch, Archoplites interruptus 0 0 E

Tule perch, Hysterocarpus traski 0 0 R

' Percentage refers to the percentage of samples in which the species was collected during this study in 1 986

(n = 186) and by Moyle and Nichols (1973. 1974) in 1970 (n = 130).
1 Status within the drainage is abbreviated as follows: E = extinct; R = rare, probably no longer resident;

D = depleted and declining, range and numbers substantially reduced; C = common, widely distributed.

98 CALIFORNIA FISH AND GAME

largely absent from lower elevation sites, and rainbow trout, mainly from introductions,
were the principal inhabitants of sites at higher elevations.

In 1985 and 1986, we resurveyed the fish fauna of the San Joaquin River drainage
but expanded the survey to more sites, including sites at lower elevations, sites on the
west side of the valley, and sites in the Kern River drainage (Fig. 1). The Kern River
is the southernmost of the major San Joaquin Valley rivers and was not sampled by
Moyle and Nichols ( 1 974). However, we did not sample the intensively farmed valley
floor because sampling by Saiki ( 1 984) and Jennings and Saiki ( 1 990) demonstrated
the scarcity of native fishes there. The purpose of the survey was to answer the
following questions:

1 . Have further declines in the distribution and abundance of native fishes taken
place since the 1970 surveys of Moyle and Nichols (1973, 1974)?

2. Would the species distributions and species-habitat relationships observed by
Moyle and Nichols (1973, 1974) hold up under more extensive sampling and
use of additional sampling gear (electrofishers)?

3. Would the species assemblages described by Moyle and Nichols (1973, 1974)
hold up under more extensive sampling, use of additional sampling gear and
application of more sophisticated statistical analyses?

4. What is the status of the Kern brook lamprey, a species described subsequent to
the 1970 surveys and known from only two localities?

STUDY AREA

The San Joaquin River drainage consists of all the streams that flow into the San
Joaquin Valley of south-central California, a drainage area of about 83,000 km2. Most
of the region receives an average of less than 25 cm of rain per year, and the main
streams depend on run-off from the Sierra Nevada on the east side of the valley. The
Coast Range on the west side of the valley is comparatively low and arid and supports
only a few small streams. During this study, only three westside streams inspected
contained water and fish: Warthan Creek, Los Gatos Creek, and Puerto Creek. The
streams flow into two main basins, Tulare Basin and San Joaquin River Basin.

Historically, the Tulare Basin was dominated by four huge, interconnected
terminal lakes occupying the low center of the southern half of the valley. The largest
was Tulare Lake, with a surface area of over 2000 km2. The lakes were created by
water flowing in from the Kern, Kaweah, Tule, and Kings River drainages, as well as
several smaller drainages. During wet years, these lakes overflowed into the San
Joaquin River, which also received water from the upper San Joaquin drainage and
from the Fresno, Chowchilla, Merced, Tuolumne, and Stanislaus rivers. Natural
flows in these streams were highly seasonal, with high flows occurring in spring
following snow-melt in the Sierra Nevada. By late summer, the flows in the main
streams were very low and small tributaries were often intermittent. Presently,
virtually all streams of any size are dammed and stream flows on the valley floor are
almost completely controlled, except during the largest floods. As a consequence,
Tulare Lake is dry (and farmed) and during the summer the San Joaquin River on the

SAN JOAQUIN FISHES

99

'A

STOCKTON

3M
\MOI

\9

19 \ nJL^MERCED

40 KM

Figure l . Drainages sampled during this study: ( l ) Stanislaus River, (2) Tuolumne River, (3) Merced River,
(4) Bear Creek, (5) Miles Creek, (6) Mariposa Creek. (7) Chowchilla River, (8) Fresno River, (9) San
Joaquin River, ( 1 0) Dry Creek, (11) Fancher Creek, (12) Kings River, (13) Kaweah River, ( 14) Tule River,
(15) Deer Creek, (16) White River, (17) Poso Creek, (18) Kern River, (19) Puerto Creek, (20) Los Gatos
Creek, and (2 1 ) Warthan Creek. Also shown are Fresno Slough (22) and the following majorcanals (dashed
lines): (23) Madera Canal, (24) Friant-Kem Canal, and (25) California Aqueduct. Circles indicate locations
where Kem brook lamprey were collected.

1 00 CALIFORNIA FISH AND GAME

valley floor flows mainly with polluted irrigation return water. Stream flows in the
Stanislaus, Tuolumne, and Merced Rivers are increased in the fall to attract and
provide spawning habitat for chinook salmon {Oncorhynchus tshawytscha).

METHODS

A preliminary survey of 33 sites was conducted in September 1985. This survey
was used for distributional studies but not for statistical analyses. The main survey
was conducted from July through September 1 986. Highest priority for sampling was
given to the 1 30 sites sampled by Moyle and Nichols ( 1 973, 1 974). However, we were
denied access to some sites on private land, others were inundated by new reservoirs,
and others were dry so we were able to sample only 84 of the original sites.

Because we were interested in comparing the abundances of the fishes among sites
and wanted to make sure that we recorded all species present at a site, three sampling
methods were used: electrofishing, seining, and snorkeling. The methods chosen for
each site were those that would sample it most thoroughly. At each site at least 50 m
of stream was sampled, except for extremely small streams or for intermittent streams
where only isolated pools were present. Sampling was halted if no additional species
or habitat types were encountered after approximately 15 minutes.

Electrofishing was used primarily in shallow, rocky streams. A single pass was
made through each reach using a Smith-Root Type VII or XI backpack electrofisher
(battery powered). In most cases, there was one person shocking the fish and one
person dipnetting them. Seines were used in habitats too large for effective electrofishing
or in water with high conductivity. Depending on the habitat, seines used were 6.6 x
1 .3 m or 1 0 x 1 .3 m with 6-mm mesh. Snorkeling was used in deep pools and runs of
the larger streams. One or two researchers swam in an upstream direction and counted
fish of all species and estimated their standard lengths (SL). All fish captured by
seining or electrofishing were measured (SL), unless more than 50 individuals of a
species were caught, in which case a representative sample of 25-50 fish was
measured.

At each site, the following environmental variables were measured: water and air
temperature (°C); maximum depth (cm); pH (measured with an electronic pH meter);
and conductivity (measured with a YSI S-C-T meter). We estimated: flow (estimated,
mVmin); water clarity (1-5 scale, where 1 = crystal clear and 5 = extremely muddy);
percentage of bottom covered with rooted aquatic plants; percentage of water surface
covered with floating aquatic plants or algae mats; percentage of habitat in pool, run,
and riffle; percentage of water surface likely to be shaded most of the day; extent of
human modification ( 1 -5 scale, where 1 = unmodified and 5 = extremely modified,
e.g. a cement-lined ditch); and percentage substrate as mud, sand, gravel, rubble,
boulder, and bedrock (according to the Wentworth particle scale, Bovee and Milhous
1978). Mean depth (cm) was estimated by measuring the depth at a point determined
by eye to represent the average depth in the study reach. Mean stream width (m) was
estimated by measuring the width of the stream at a point determined by eye to
represent the average width of the study reach. In addition, elevation, gradient, and

SAN JOAQUIN FISHES 101

stream order (Strahler 1957) were determined from USGS 7.5' or 15' topographical
maps.

All fish except lamprey were identified using keys in Moyle ( 1 976). Because there
is no taxonomic key to the ammocoetes of California lampreys, they could not be
identified to species with certainty. Pacific lamprey {Lampetra tridentata) and Kern
brook lamprey have different numbers of trunk myomeres (Vladykov and Kott 1976,
Richards, Beamish, and Beamish 1982). All ammocoetes with trunk myomere counts
of 63-69 were assumed to be Pacific lamprey and those with myomere counts of 5 1 -
57 were assumed to be Kern brook lamprey. This means it was possible we
misidentified river lamprey (L. ayersi) ammocoetes as Pacific lamprey and Pacific
brook lamprey (L. richardsoni) as Kern brook lamprey. However, river lamprey
appear to be most abundant in the lower Sacramento-San Joaquin River system and
have not been reported from the areas we sampled (Moyle 1976). Also, to verify our
tentative identification of low myomere count ammocoetes as Kern brook lamprey,
we returned in March and May 1987 to two of the sites from which we collected
ammocoetes with low myomere counts and collected transforming individuals. Those
with well developed tooth plates were identified as Kern brook lamprey.

We sampled 1 86 sites in 1986 (Table 2). After reviewing the data, we decided that
data from 156 of the sites were appropriate for quantitative statistical analyses. The
remaining 30 sites were big river sites where the edges were sampled for lampreys.
Most other fishes were not adequately sampled because of great depth, high flows, or
low visibility. Because of the variety of sampling methods used, rank abundance of
each species in each sample was used for analyses rather than actual numbers
collected or observed. Only species occurring in at least 5% of the samples were
included in quantitative analyses.

For each common species, we calculated Pearson product moment correlations
between species rank abundances and each of the environmental variables measured.
Correlations were also calculated between species rank abundances and percentage
of native fish at each site and total number of species at each site. Strictly speaking,
these latter correlations were not statistically valid because the variables were not
independent. However, as shown in Moyle and Nichols (1973), the comparisons do
have descriptive value. Correlations were considered significant at P < 0.05. For each
species, the mean and standard deviation of each of the environmental variables was
calculated based on data from the stations at which the species was found.

A principal components analysis was conducted on the rank abundance data to
determine patterns of co-occurrence among species. Principal components with
eigenvalues greater than or equal to one were rotated using a varimax rotation (SAS
1982). Principal components analysis is a multivariate method for reducing a large
number of intercorrelated variables to a reduced number of orthogonal (independent)
variables. The loadings of the original variables on a principal component indicate the
correlations of the variables to the principal component. The varimax rotation makes
a final adjustment to maximize the amount of variation explained by the first few
principal components. Pearson product moment correlations were also calculated
among species, using rank abundances.

102

CALIFORNIA FISH AND GAME

Table 2. Total number of sites sampled and number of sites included in quantitative analyses
for each drainage sampled during this study.

Total

Quantitative

Drainage

Sites

Sites

Stanislaus River

19

12

Tuolumne River

21

16

Merced River

7

3

Bear Creek

2

2

Miles Creek

2

2

Mariposa Creek

2

2

Chowchilla River

13

13

Fresno River

9

9

San Joaquin River

21

19

Dry Creek

1

1

Fancher Creek

1

1

Kings River

27

20

Kaweah River

18

IS

Tule River

16

15

Deer Creek

6

6

White River

1

1

Poso Creek

1

1

Kern River

16

12

Puerto Creek

1

1

Los Gatos Creek

1

1

Warthan Creek

!

1

Total

186

156

RESULTS
General Distributional Patterns

Eleven of the 19 original native species were collected in this study (Table 1), in
addition to 19 introduced species (Table 3). Only six species of native fish (rainbow
trout, hitch, hardhead, Sacramento squawfish, California roach, Sacramento sucker)
were collected frequently enough to use in the quantitative analyses, as were six
introduced species (brown trout, mosquitofish, green sunfish, bluegill, largemouth
bass, smallmouth bass). Distributions of native species were similar to those shown
in Moyle and Nichols (1974), with exceptions noted in the species accounts below.
Native fishes were largely confined to a narrow band of habitat in the foothills, usually
above the major Sierra Nevada foothill dams or in tributary streams below them. We
review the results of other studies in the following species accounts to give the most
accurate depiction possible of the distribution and abundance of each species.

SAN JOAQUIN FISHES 103

Table 3. Introduced fishes collected from streams of the San Joaquin River Drainage during this
study in 1986 and by Moyle and Nichols (1973, 1974) in 1970.

Species

Largemouth bass, Micropterus salmoides
Smallmouth bass, Micropterus dolomieu
Redeye bass, Micropterus coosae
Green sunfish, Lepomis cyanellus
Bluegill, Lepomis macrochirus
Redear sunfish, Lepomis microlophus
Black crappie, Pomoxis nigromaculatus
White crappie, Pomoxis annularis
Bigscale logperch, Percina macrolepida
Western mosquitofish, Gambusia affinis
Carp, Cyprinus carpio
Goldfish, Carassius auratus
Golden shiner, Notemigonus chrysoleucas
Fathead minnow, Pimephales promelus
White catfish, Ameiurus catus
Black bullhead, Ameiurus melas
Brown bullhead, Ameiurus nebulosus
Brown trout, Salmo trutta
Threadfin shad, Dorosoma petenense

Percentage3

1986

1970

19

31

7

7

1

0

25

46

12

23

4

1

<1

0

2

0

<1

0

23

26

2

2

<1

1

3

1

<1

0

3

9

4

0

0

7

9

1

<1

0

a Percentage refers to the percentage of samples in which the species was collected in 1986 (n = 156) and
in 1970 (n= 130).

Native Fishes

Rainbow trout. Rainbow trout occurred at 59 (38%) of the 156 sites in 1986. Their
abundance was positively correlated with stream order, depth, gradient, elevation,
percentage riffle, and percentage boulders and was negatively correlated with
temperature, water clarity, pH, conductivity, percentage rooted and floating vegetation,
percentage pools, human modification, and total number of species (Table 4). These
correlations indicate that rainbow trout were characteristic of clear, cold high
elevation, high gradient streams, the same pattern found by Moyle and Nichols
(1973).

Hitch. These native cyprinids were found at only 8 (5%) of the 156 sites. Moyle
and Nichols (1973) found them at 13 (10%) sites, despite sampling a more limited
geographic area. In 1986, hitch were absent from the Tule River, Bear Creek and
Fresno River drainages where they were collected in 1970. We did capture hitch in
the Fresno River in 1985. Hitch abundance was correlated with percentage run and
percentage shade and negatively correlated with stream order, water temperature, and
percentage pool (Table 4). The typical hitch stream was a sandy bottomed, low
gradient stream at moderately low elevations, containing a mixture of native and
introduced species. The mean elevation of sites with hitch was 418 m, the lowest for
any native species (Table 4). The habitat where we found hitch was similar to that

104

CALIFORNIA FISH AND GAME

o
■J-.

'S.

o

'3
u

Q.
y,

U

>

eO

c

><

CO

c

£

c

|

>
c
u

U

■*-

u

u

u

Q

u

s

c

c u

2 y

U 3

co as

CO

CO

E

o

x:
o
es

c

ca

U

O JS

(8
: £

o cr

CO «s

T3
CO

u

1

u

o

X
C

"3

04

i

IT!

^S^go

§P$5*2

i

CN CN

r-- — . r<-> cn ~ £2, —

CN 2 -

C<1

ri

f~-^TtSSt1^r--<^vo(N. £ ch 2 >§ — — o

ON

IT1
CN

^ jv OV i^ f*l

IT)

ID

cn

^ OS
CN 00

F- - ""> ©

^ vo

On ^ ^

00

in — ""! 00

-:oot^r^rf(2''-i'r-

m * o\

in

vo cn cn ^ o — r~-

>/-) CN

PO — • CN CN

oc

cn r~

00

;_,

ri

•?

rn

oc

r-_

cn

cn

sC

— 1

sC

-r

d

m

d

cn

n-

cn

d

- -

*■"*'

^— '■

• — '

^-^

■ — '

' — '

■ — *

■ — ■

1-

o

<*

i/-,

■tf

c

«t

-T

t-

-r

r\

ri

cn

cn

CN

d

CN

1ft

CN

d

Tf ^ tT CN ^

cn >n cn <n tj-

Ov * rs - N m n
Tfr' oo

fl

cn cn

0C

^, C- C- o — m r

>/-,

m m

cn cn £ ^ — cn cn ;u ill cn ~

- S^<P

oo _;

cn

S o » *> 2 22 <=? - £

- cn - ^ cn ^ ^

oc

m o\
_: m

oo oo ^ i^ ^ [2

-. ^ 2> ^ fl

CN~ <^

. i/~) lI7 ^^ ^^ ^T r*^ ro ^^ \& ^"

cj^^d^Si^cJccgCc

IT)

4J

E

3

z

u ^

^ E

o 3

fc- —

3 x:

2 a.

E E

>r> ,3 ^

ill

u
•a

r e

42 >> «

II

.2 "-3

60 (U
<U >
> 60

•S "9 c

t#

CO
O

u. -»i

T3 O

1 "O

o a>

'3

u
a

x:

IS
u
>

CO
C

o
'3

y;

C

o

c
u

©
©
v
a.

T3
C
CO

m

©
d
v
ft.

>

so

M

o
C

C

CO

O

d
V
ft,

+

+

•o
a

CO

in

©

d
v
ft.

>

-1

o >

E S
.2P-3

IS °-

5 <u

.9 £

2 'I
a. >

•r <u

b g

<— c0

S" c

5 3
n CO

SAN JOAQUIN FISHES 105

described by Moyle and Nichols (1973).

Hardhead. Hardhead were found only at 10 (6%) sites. Moyle and Nichols (1973)
found them at 12 (9%) sites. Five of the sites in this study were on the Kern River,
which was not sampled by Moyle and Nichols (1973, 1974). We failed to collect
hardhead from three streams where they were found by Moyle and Nichols (1973,
1974), Horse Creek in the Kaweah River drainage and Big Creek and Dinky Creek
in the Kings River drainage. In 1 986, we did not find hardhead at three sites where we
found them in 1985. We also did not collect hardhead from the Tuolumne River,
where intensive sampling indicated they have been present in small numbers (EA
Engineering, Science, and Technology, 1990, unpublished data). The small number
of sites containing hardhead limited the correlations to a negative correlation of
abundance with temperature and a positive correlation with percentage of native fish
(Table 4). The positive correlation is consistent with the results from other studies that
indicate hardhead are found in the least disturbed sections of the larger streams
dominated by native species (Moyle 1976). The absence or scarcity of hardhead in
otherwise seemingly suitable habitats that contain smallmouth bass, especially in the
upper Kings River, suggests that predation by bass may be limiting hardhead
numbers. Overall, hardhead were uncommon in the San Joaquin River drainage and
appear to be declining.

Sacramento squawfish. These predatory cyprinids were abundant and widely
distributed in both the 1 970 and 1 986 studies. In 1 986, they were present at 46 (29%)
sites and were collected at most sites where they were found in 1970. Squawfish
abundance was positively correlated with stream order, maximum depth, pH, percentage
of rooted vegetation, percentage of shade, and percentage of native fish but negatively
correlated only with elevation (Table 4). This pattern indicates that squawfish were
most common in the larger streams at lower elevations at sites with well developed
riparian vegetation and native fish communities.

California roach. Roach were found at 47 (30%) of the sample sites in 1986 and
42 (32%) in 1 970. They were not collected from the Kern River drainage but were the
dominant species in the three streams sampled on the west side of the San Joaquin
Valley. Their abundance was positively correlated with conductivity, percentage of
rubble, and percentage of native fish. Roach were negatively correlated with stream
order, stream width, turbidity, percentage of sand, and number of species (Table 4).
These fish were most common in small, clear intermittent streams where they were
often the only species present, although they were found in habitats ranging from cool
trout streams to warm isolated pools. In the latter habitats, they were often found at
extremely high densities. Taxonomic analyses of the roach collected in this study
indicated that the population in each tributary drainage exhibits minor morphological
and meristic differences from populations in the other drainages (Brown et al. 1 992).

Sacramento sucker. Sacramento suckers were the most commonly collected
native fish in both the 1970 (42% of sites) and the 1986 (48% of sites) surveys. They
were found in a wide variety of habitats. Their abundance was positively correlated
with stream order, maximum depth, flow, percentage bedrock, and percentage native
fish (Table 4). Sucker abundance was negatively correlated with percentage of rooted

1 06 CALIFORNIA FISH AND GAME

vegetation, percentage of pool, percentage of shade, and total number of species.
Suckers were most abundant in the larger, clear, cool streams that also contained either
rainbow trout or California roach. Even though suckers were commonly collected
with introduced species in disturbed habitats, they were rarely abundant in such
situations.

Kern brook lamprey. Presumptive Kern brook lamprey were collected in 1985,
1986, and 1987 from lower Merced River, the San Joaquin River below Friant Dam,
the Kings River above and below Pine Flat Dam, and the lower Kaweah River (Fig.
1). Wang (1986) described ammocoetes that probably belonged to this species from
a site above Friant Dam (Millerton Reservoir) but below Kerckoff Reservoir. With the
exception of the upper Kings River site, all collection localities were below major
dams. Typical collection localities for the ammocoetes were sandy-bottomed
backwaters or shallow river edges. Brook lampreys were not collected from similar
habitats in the lower Kern, Tule, Tuolumne and Stanislaus rivers, although considerable
effort was made to find them. When encountered, ammocoetes were usually locally
abundant. Previous records of Kern brook lamprey were from the Merced River and
the Friant-Kem Canal (Vladykov and Kott 1976, 1984). This canal delivers water
from Millerton Reservoir to farmlands to the south. In 1988, ammocoetes and adults
of Kern brook lamprey were collected by California Department of Fish and Game
personnel from the silty-bottomed siphons of the Friant-Kern canal when the siphons
were treated with rotenone to rid the canal of white bass (Morone chrysops) (Moyle
et al. 1989).

Pacific lamprey. Presumptive Pacific lamprey ammocoetes were found in the
lower Stanislaus, Tuolumne, Merced, and Kings rivers, as well as the San Joaquin
River below Friant dam (in 1 985 only). Lampreys were not collected in 1 970 because
electroshockers were not used and ammocoetes are not vulnerable to seines, the main
method of collection in 1970. We expected to find Pacific lampreys in the Stanislaus,
Tuolumne, and Merced rivers because they are anadromous and these rivers are
accessible from the sea as indicated by small runs of chinook salmon {Oncorhynchus
tshawytscha). However, the San Joaquin and Kings river sites are likely to be
accessible only during wet years, when water spills from the dams. During wet years,
lampreys and salmon (Moyle 1970) may occasionally gain access to these rivers and
spawn. Because Pacific lampreys may persist as larvae for 5-7 years in streams (Moyle
1 976) the progeny of such spawnings will persist in the system for a number of years.
It is likely that the ammocoetes we captured were the result of such an event.

Uncommon native fishes. Prickly sculpins {Cottus asper) riffle sculpins (Cottus
gulosus) and threespine sticklebacks {Gasterosteus aculeatus) were collected at only
a small number of sites in the 1970 and 1986 studies. Prickly sculpins were found in
the lower Stanislaus, Merced, Fresno, Kings, and Kaweah rivers and in the San
Joaquin River below Friant Dam. This sculpin disperses readily because of its
planktonic larval stage and tolerance of a wide variety of environmental conditions.
They are common in many of the reservoirs in the drainage. The reservoirs probably
act as a source of larvae to downstream areas, so small numbers were expected below
the dams (Moyle 1976). Riffle sculpins, in contrast, require small, cold, permanent

SAN JOAQUIN FISHES 107

streams and disperse very slowly because they have benthic larvae (Smith 1982). We
found them mainly in a few isolated populations above dams in the upper San Joaquin
River, the Kings River, and the Kaweah River, although a substantial population also
exists in a 4-km reach of the Tuolumne River below LaGrange Dam, with small
numbers occurring as far as 36 km downstream from the dam (EA Engineering,
Science, and Technology, unpublished data, 1 990). Sculpins of undetermined species
have also been observed in the Middle Fork Stanislaus River, which we did not sample
(Brian Quelvog, Calif. Dept. Fish and Game, pers. comm.). Threespine sticklebacks
were found only in the San Joaquin River immediately below Friant Dam, in a small
tributary to the San Joaquin River above Kerckoff Reservoir and in a 30-km reach of
the Kings River below Pine Flat Dam.

Introduced Fishes

Brown trout and smallmouth bass. These two species have scattered but
complementary distribution patterns throughout the drainage that probably more
reflect their planting history by humans than their ecological requirements. Brown
trout were captured at 16 (10%) sites and their abundance was correlated with the
same environmental factors as rainbow trout (Table 5). Smallmouth bass were
captured at nine sites in both 1970 and 1986 and showed few correlations with
environmental variables (Table 5). The main difference between the 2 years is that
bass were more abundant and widely distributed in the upper Kings River drainage
in 1986 than they were in 1970, an expansion associated with decline of hardhead in
the same area. Sites from which smallmouth bass were collected in 1970 but not in
1986 were either dry in 1986 or inundated by new reservoirs.

Largemouth bass. Largemouth bass were captured at 28 sites (18%). Their
abundance was positively correlated with water temperature, turbidity, percentage of
rooted and floating vegetation, percentage of pool, percentage of sand, and human
modification. They were negatively correlated with gradient, elevation, percentage of
riffle, and percentage of rubble (Table 5). As found by Moyle and Nichols (1973), we
found largemouth bass in warm, pond-like habitats in highly disturbed areas. They
were often particularly abundant below impoundments, which not only reduce flows
downstream (resulting in more pools and warmer water) but act as sources of
immigrants. Where largemouth bass were common, native fishes were uncommon or
absent.

Green sunfish and mosquitofish. These two species were abundant, widely
distributed, and often co-occurring. Green sunfish were present at 43 (28%) sites, and
mosquitofish were present at 30 (19%) sites. The abundance of both species had
positive correlations with water temperature, turbidity, percentage of rooted vegetation,
percentage of pool, percentage of sand, and human modification (Table 5). They were
negatively correlated with stream order, maximum depth, gradient, elevation,
percentage of riffle, and percentage of rubble, boulder, and bedrock. These results
support Moyle and Nichols' (1973) characterization of their habitat as warm,
intermittent streams in areas highly disturbed by livestock grazing and human activity

108

CALIFORNIA FISH AND GAME

o
'3

O"

o

Oil

OO rt W ^O fi

(U

^_^ N— «• S» • ^— ^ ^— «*

3

r~ u-j cn no

CQ

o i

3

J3

■t— >

3
O co

C co

■a*

E

00

o 2
co

© L

n "i m

oriSBBd^-^8 2

8>

t ~>^

^.' 00 i/-> ■*

■<* — n

rr -^ CN "* j-j 5 —

• cn oo © r- r-

rtW CN

_: f-

— w-)

en

— CM On ^J

^ o^
©

q
en

00 ^° £-. ^-v

S £ £ 2 o

moo

D

CO

* — '

nC

1-

f^ r- cm ,— . — m
^ — cm on _; CM

-k >/■> m no

°^ — CN

00

en

CM

en -• ^ S 2

cn (©

oe i^vg^vqocn,.

oc

r<-) in r~ On

r- o^ >> °i "o

oo —J en

cm on

ITi (Nl » _' *

<n cm rn o<

(^ 00 "* 00 "^ NO On

oo t

°°. J-v 00 CN vO

oo r~", fn !fl "^ ^

ri

f- *£> CN rt — .

_• cm on _; <=*

— CN

• ON « CN — 2

O ^ ^ <N. S C

CN

en

O so 00 CN

en Z- r- —
en O

d cn -. w w - w w

cn

^^£3

O^

On ^~-

_; m

Q
-t-

CM

^^CS S

cn o — —

i

N 00 Is; tJ" O1 ^

CM TT — ^) cn £2

,-s NO

2 ©

nO o _
-. CN — .

cn — r-

r~ NO On cn cn —
— ^ ^; no —

CN

CN

ri

ri

oo Z.

^ r»

ri

i^,

— >n

-*• cn

"O "O (o *

^C

I ^^ 00

cs p: Tt

. r- o ^ ~ ! 2 o ;

CN — — '0Csjcnvpo''

oo on m oo 't cn

rj

CN O

O N-*

o — on r- cn fs, __

ON

m

cn — — cn —

CN

<U

E

3

z

u

^

o

E

g

o

3

J=

^

4— *

i-
o.

o-

E

E 5a

OJ

3 V,

T_

■t— »

E^

OJ

u

*-*

"3 ^

"O

«

C3 O

6

^SE

»o

c§3

•P ^ >

• — ■

C
^O
^-

*— '
OJ
00

u

>

-o

£ #

fi # R,

# # -

-O "O >

3t^5S^: = i2
HOwoSft-oiSonO

QoS

O

x> 3 -o

3 O m

a: co cq

S 2

co

4J
'G

s

o.

CO

i-

s:

u

'S

u

D.

J-,

—

O

s:

c

o

ca

u

t

o

o

^

f^

o

d

V

a.

i

^3

C

?3

ir,

O

d

V

a.

'

— .

u

>

K)

SXj

'J

C

TD

C

r3

^ — .

^—

O

d

V

ft,

+

+

TD

C

a

m

O

d

V

Q~

+

^— ^

<u

>

*■!

V

O

_o

ax

03

g

73

u

>

|g

3

'E

U

oo

'co

'co

>,

o

X

4^

Cu

u

OJ

X

•5

c

—

W3

■4— 1

a.

?

o

u.

o

UJ

T3

a.

c

3

3

oo

X

n

w

SAN JOAQUIN FISHES 109

where other species of fish are uncommon. Green sunfish are widely distributed
because of their natural dispersal abilities (Smith 1 982). They frequently are found in
low numbers in relatively undisturbed streams with native fishes. They are able to
invade these streams in low numbers but do not become abundant until physical
conditions favor them.

Bluegill and other species. Bluegill were found at 18(1 2%) sites but were usually
not abundant. It is likely that most of the fish represent escapees from farm ponds or
reservoirs. Their abundance was positively correlated with water temperature,
turbidity, and percentage of pools, indicating their presence in pond-like habitats with
other introduced warmwater fishes (Table 5). Twelve other introduced warmwater
fishes had only sporadic occurrences, mainly in river reaches below reservoirs (Table
3). These and at least five other introduced species are common in waters of the valley
floor (Saiki 1984, Jennings and Saiki 1990). The only species of note is redeye bass
{Micropterus coosae) from a single site on the South Fork of the Stanislaus River,
where they were common. Redeye bass were introduced there in 1962 (Moyle 1976)
and have recently been collected from New Melones Reservoir on the Stanislaus
River (Randy Kelly, Calif. Dept. Fish and Game, pers. comm.), indicating that they
are slowly dispersing downstream.

Species Assemblages

Moyle and Nichols (1973) used correlation analysis to identify four assemblages
of species: (i) the Rainbow Trout Association (mainly rainbow trout alone); (ii) the
California Roach Association (California roach plus juvenile Sacramento suckers);
(iii) the Native Cyprinid-Catostomid Association (mainly Sacramento squawfish,
hardhead, and Sacramento sucker); and (iv) the Introduced Fish Association (mainly
largemouth bass, bluegill, green sunfish, and mosquitofish). A similar pattern was
evident in our own correlation analyses but the principal components analysis (PCA)
indicated that the Introduced Fishes Association could be divided into two subgroups.
The PCA yielded five components with eigenvalues greater than one. The five rotated
principal components (varimax rotation) explained 69% of the variance in rank
abundance of the 12 most abundant species (Table 6).

The first principal component defined the Native Cyprinid-Catostomid Association,
but without hardhead. Squawfish and suckers showed high positive loadings on this
component indicating a high degree of association of the two species. Hitch and
brown trout showed low negative loadings indicating they were present at sites where
squawfish and suckers were absent. Green sunfish and mosquitofish loaded positively
on the second principal component 2; rainbow trout and brown trout load negatively
on it. The second principal component thus identified the Rainbow Trout Association
and a subset of the Introduced Species Association. It separated fishes associated with
undisturbed, clear, high gradient streams from those associated with highly disturbed,
turbid, low gradient, small streams. The third principal component represented the
remainder of the Introduced Species Association that occurred in larger streams and
was loaded heavily by largemouth bass and bluegill. The negative loading of hitch

1 1 0 CALIFORNIA FISH AND GAME

Table 6. Principal component loadings (after Varimax rotation) of species rank abundances
for common fishes of the San Joaquin River Drainage. Loadings of less than 0.4 are not
shown.

Principal Component

Species

1

2

3

4

5

Sacramento squawfish

0.80

Sacramento sucker

0.78

Brown trout

-0.56

-0.50

Rainbow trout

-0.75

Green sunfish

0.75

Mosquitofish

0.66

Hitch

-0.42

-0.41

Largemouth bass

0.79

Bluegill

0.81

Smallmoufh bass

0.86

Hardhead

-0.59

Calif, roach

0.85

Variance explained (%)

17

17

15

10

10

Cumulative variance explained (%)

17

34

49

59

69

showed that hitch were negatively correlated with the abundance of the other two
species. The fourth principal component separated smallmouth bass from hardhead.
Neither of these species showed many correlations with environmental factors
measured, in part due to the small sample sizes for both. However, both species were
commonly found in deep pools of large, cool streams, suggesting that the two species
were not found together because of interactions between them, presumably smallmouth
bass predation on juvenile hardhead. The fifth principal component identified the
California Roach Association that dominated the small, warm intermittent streams.

DISCUSSION

It is clear that native fishes have been greatly reduced in numbers and distribution
since the arrival of Europeans in California. Only 11 of the 19 native species were
collected in this study and only six of the eleven were found at more than a few sites.
Except for Pacific lamprey, the five uncommon species occurred in scattered, isolated
populations, where localized extinction was possible, with no possibility of natural
recolonization because of dams. This situation was also characteristic of the distributions
of two of the more widely distributed species, hitch and hardhead. Introduced species
now dominate many streams, especially those altered by human activity, and native
fishes are largely gone from the valley floor (Saiki 1984, Jennings and Saiki 1990).

Two of the eight species not collected in this study (delta smelt, federally listed
as threatened, and white sturgeon) are estuarine species that formerly ascended further
up the San Joaquin River than they do today. Neither was collected by Saiki (1984);

SAN JOAQUIN FISHES 1 1 1

however, white sturgeon were collected by Jennings and Saiki (1990). Four of the
eight species were once abundant on the valley floor but are now either extinct
(thicktail chub, Sacramento perch) or very rare there (tule perch, Sacramento splittail)
(Moyle 1976, Saiki 1984, Jennings and Saiki 1 990, Moyle and Williams 1990). Tule
perch and splittail still persist in the estuary (Moyle et al. 1982) and are occasionally
collected from valley floor rivers (Jennings and Saiki 1990). The only native species
that is still abundant on the valley floor is Sacramento blackfish (Saiki 1 984), a species
that is remarkably tolerant of poor water quality. Chinook salmon were not found at
our sampling sites because their presence is seasonal and they are not present in the
summer. Salmon spawn in the lower Stanislaus, Tuolumne, and Merced rivers but the
juveniles usually emigrate during March through May (Sasaki 1966). Prior to the
construction of the dams on the rivers we sampled, it is likely that chinook salmon
were present at many of our sample sites; runs of adult salmon were once 300,000-
500,000 or more per year in the drainage (Lufkin 1990). In 1990-91, less than 1,000
adult salmon were present in the drainage (Calif. Dept. Fish and Game 1992).

Have further declines of the native fishes taken place since 1970? Fifteen years
is a short time to detect differences in fish distribution over a wide area, especially
between two surveys that did not overlap completely in their sampling sites. However,
in 1986 hitch were not found in the Tule River, Bear Creek, or Fresno River where
they were found in 1970 and hardhead were absent from the Kaweah River drainage
and two streams in the Kings River drainage where they were found in 1970. The
decline of hardhead may be associated with the expansion of smallmouth bass
populations. Further declines may have taken place since this survey was completed
as the period from 1986 to 1992 was one of severe drought in the drainage, which
would exacerbate the already precarious position of many of the native fish populations.

Are the species distributions, species-habitat relationships and species
assemblages observed consistent with those of Moyle and Nichols (1973, 1974)?
The species distributions and species-habitat patterns observed in 1970 were very
similar to the results of our 1986 survey, except for the differences already noted for
hitch and hardhead. The species assemblages identified in 1 986 differed slightly from
those identified by Moyle and Nichols (1973, 1974). Similar to Moyle and Nichols
( 1 973, 1 974), three assemblages of native fishes were found, each occupying a narrow
elevational band in the Sierra foothills. They were the Rainbow Trout Association,
Native Cyprinid-Catostomid Association, and the California Roach Association. The
California Roach Association was absent, apparently naturally, from the Kern River
drainage, but was the only assemblage found in the few permanent streams on the arid
west side of the valley. The Introduced Fishes Association was not distinctly defined
in 1986 compared to 1970. Moyle and Nichols (1973) characterized the Introduced
Fishes Association by the presence of largemouth bass, green sunfish, bluegill, and
mosquitofish. The principal components analysis indicated that the Introduced Fishes
Association can be divided into two subgroups. The first group included green sunfish
and mosquitofish that characterized the smaller streams and the second group
included largemouth bass and bluegill that were generally found in larger streams.
The values of the Pearson product moment correlations among these species were

1 1 2 CALIFORNIA FISH AND GAME

similar in both studies; therefore, the division of the Introduced Fishes Association
into subgroups is most likely the result of the greater sensitivity of principal
components analysis to species relationships compared to simpler techniques.

What is the status of the Kern brook lamprey? The Kern brook lamprey was
of special interest in this study because it is a recently described taxon and the only
species identified as endemic to the San Joaquin Valley. The five or six known Kern
brook lamprey populations probably need special protection, because they are
isolated from one another and with one exception, occur below dams, so are subject
to the vagaries of water releases from the dams. Moyle et al. (1989) regard them as
a Species of Special Concern, a status which seems appropriate.

CONCLUSIONS

Based on our results and other studies (Saiki 1 984, Jennings and Saiki 1 990), only
five of the original 19 species of the San Joaquin drainage are reasonably abundant
and widely distributed (rainbow trout, Sacramento squawfish, California roach,
Sacramento blackfish, Sacramento sucker). Once abundant anadromous fishes are
now a tiny fraction of their original numbers. Most aquatic habitats on the valley floor
are probably degraded beyond hope of recovery to the point where they can support
native fish assemblages. Most foothill streams, even those containing native fishes,
are in a degraded condition (Brown and Moyle 1 992). It is clear that the fish fauna has
declined in distribution and abundance since Europeans came to California, a decline
that appears to be continuing and seems unlikely to be halted soon. The best hope for
protecting some remnants of the fauna and their aquatic habitats (and the attendant
native plants, invertebrates, and amphibians) is to establish a series of preserves or
Aquatic Diversity Management Areas, along the lines suggested by Moyle and
Yoshiyama(1992).

ACKNOWLEDGMENTS

Preliminary surveys in 1985 were funded by the Water Resources Center,
University of California (W-659). The 1986 survey was sponsored by the Endangered
Species Tax Checkoff Fund, California Department of Fish and Game (Contract C-
260607), supervised by D. McGriff . D. White, E. Wikramanayake, and G. Sato helped
to collect the field data. C. Chadwick, D. Christensen, W. Loudermilk, and N. Villa
of the California Department of Fish and Game suggested sampling sites, especially
for lamprey. B. Quelvog and an anonymous reviewer significantly improved the
paper.

LITERATURE CITED

Bovee, K. D., and R.T. Milhous. 1978. Hydraulic simulation in instream flow studies: theory
and technique. Instream flow information paper 5. United States Fish and Wildlife Service,
Biological Services Program FWS-OBS 78-33.

SAN JOAQUIN FISHES 113

Brown, L.R., and P.B. Moyle. 1992. Native fishes of the San Joaquin drainage: status of a

remnant fauna and its habitats. Pages 89-98 in D. F. Williams, T. A. Rado, and S. Byrde,

eds. Proceedings of the Conference on Endangered and Sensitive Species of the San Joaquin

Valley, California. California Energy Commission, Sacramento, CA.
, , W. A. Bennett, and B. Quelvog. 1992. Morphological variation among

populations of California roach (Cyprinidae: Lavinia symmetricus): implications for

conservation strategies. Biol. Cons. 62:1-10.
California Department of Fish and Game. 1 992. San Joaquin River chinook salmon enhancement

project annual report fiscal year 1 990- 1 99 1 . Calif. Dep. Fish and Game, Fresno, CA. 84 p.
Jennings, M.R., and M.K. Saiki. 1990. Establishment of red shiner, Notropis lutrensis, in the

San Joaquin Valley, California. Calif. Fish and Game 76:46-57.
Kahrl, W.L., W.A. Bowen, S. Brand, M.L. Shelton, D.L. Fuller, D.A. Ryan. 1978. The

California water atlas. The Governor's Office of Planning and Research, Sacramento, CA.

118 p.
Lufkin, A., ed. 1990. Salmon and steelhead in California. University of Calif. Press, Berkeley,

CA. 305 p.
Moyle, P.B. 1970. Occurrence of king (chinook) salmon in the Kings River, Fresno County.

Calif. Fish and Game 56:314-315.

. 1976. Inland fishes of California. University of Calif. Press, Berkeley, CA. 405 p.

, and R.D. Nichols. 1973. Ecology of some native and introduced fishes of the Sierra

Nevada foothills in Central California. Copeia 1973:473-490.
, . 1974. Decline of the native fish fauna of the Sierra Nevada foothills, Central

California. American Midi. Nat. 92:72-83.
, and J.E. Williams. 1990. Biodiversity loss in the temperate zone: decline of the native

fish fauna of California. Cons. Biol. 4:275-284.
, , and E.D. Wikramanayake. 1989. Fish species of special concern of California.

Calif. Dept. Fish and Game, Sacramento, CA. 222 p.
_, and R.M. Yoshiyama. 1 992. Fishes, aquatic diversity management areas, and endangered

species: a plan to protect California's native aquatic biota. The California Policy Seminar,

University of California, Berkeley, CA. 222 p.
Richards, J.E., R.J. Beamish, and F. W.H. Beamish. 1 982. Descriptions and keys for ammocoetes

of lampreys from British Columbia, Canada. Can. J. Fish. Aquat. Sci. 39:1484-1495.
Saiki, M.K. 1984. Environmental conditions and fish faunas in low elevation rivers on the

irrigated San Joaquin Valley floor. Calif. Fish and Game 70:145-157.
Sasaki, S. 1966. Distribution and food habits of King salmon, Oncorhynchus tshawytscha, and

steelhead rainbow trout, Salmo gairdneri, in the Sacramento-San Joaquin Delta. Pages 1 08-

114 m J.L. Turner and D.W. Kelley (compilers). Ecological Studies of the Sacramento-San

Joaquin Delta, Part II, Fishes of the Delta. Calif. Dept. Fish and Game, Fish Bull. 136:1-

168.
SAS Institute Inc. 1982. SAS user's guide: statistics. SAS Institute Inc., Cary, North Carolina.

584 p.
Smith, J.J. 1982. Distribution and ecology of stream fishes of the Sacramento-San Joaquin

drainage system, California. II. Fishes of the Pajaro River system. Univ. Calif. Publ. Zool.

115:83-170.
Strahler, A.N. 1957. Quantitative analysis of watershed geomorphology. Trans. Amer.

Geophysical Union 38:913-920.
Vladykov, V.D. and E. Kott. 1976. A new nonparasitic species of lamprey of the genus

Entosphenus Gill, 1862, (Petromyzonidae) from south central California. Bull. So. Calif.

Acad. Sci. 75:60-67.

1 1 4 CALIFORNIA FISH AND GAME

Vladykov, V.D. and E. Kott. 1984. A second record for California and additional information

on Entosphenus hubbsi Vladykov and Kott 1976 (Petromyzontidae). Calif. Fish and Game

70:121-127.
Wang, J.C.S. 1986. Fishes of the Sacramento-San Joaquin estuary and adjacent waters,

California. Interagency Ecol. Stud. Prog, for the Sacramento-San Joaquin Estuary, Tech.

Rept. 9, Sacramento, CA.

Received: 9 July 1991
Accepted: 3 September 1993

.

CALIFORNIA FISH AND GAME
Calif. Fish and Game (79)3 : 1 1 5- 1 25 1 993

FIRE EFFECTS ON A MONTANE
SIERRA NEVADA MEADOW

ROBERT S. BOYD

Department of Botany and Microbiology

and Alabama Agricultural Experiment Station

Auburn University, AL 36849-5407

ROY A. WOODWARD

Bechtel Corporation

50 Beale Street

San Francisco, CA 941 19-3965

GARY WALTER

California Department of Parks and Recreation

P.O. Drawer D

Tahoma, CA 95733

The effects of a late fall burn on a mountain meadow at Grover Hot
Springs State Park, California, were evaluated. Both wet (Carex sp.
dominated) and dry (Poa sp. dominated) meadow plots were burned by
a low to moderate intensity fire in mid-November 1987. Fire resulted in
few detectable changes in species composition 1 0.5 months later. Postf ire
decreases in Poa sp. and Juncus sp., and increases in Carex sp. and
Muhlenbergiasp. were observed. Burning increased bare ground area by
more than three fold in both wet and dry plots, but there was no significant
invasion of burned areas by exotic species after the fire. Under high soil
moisture conditions, burning resulted in relatively little change in meadow
vegetation.

INTRODUCTION

Fire is a natural process in most grassland ecosystems (Vogl 1974). Despite
widespread acceptance of the natural role of fire in grasslands and many other
vegetation types, little is known about fire effects in North American mountain
meadows (Ratliff 1985). Rundel et al. (1977) pointed out the lack of information
about fire effects on high elevation Sierra Nevada communities in general. There is
but one published account of a wildfire burning a meadow in the Sierra Nevada
(DeBenedetti and Parsons 1979).

Meadow fire may be a positive factor essential to maintaining meadows against
invasion by woody species (Gibbens and Heady 1964). Fire may also have a
destructive effect by consuming the organic matter forming the bulk of many wet
meadow soils (Bennett 1965, Vogl 1974). There are few data available from which
to draw any conclusions. DeBenedetti and Parsons (1979, 1984) documented rapid
recovery of the vegetation despite apparently extensive destruction of soil organic
matter and root biomass due to the occurrence of fire during drought conditions.

115

1 1 6 CALIFORNIA FISH AND GAME

Managers of most public wildlands are charged with maintaining them in a
'natural' condition (Parsons et al. 1986). The California Department of Parks and
Recreation (DPR) has been exploring the utility of fire in managing many vegetation
types. Fire has not been used in mountain meadows due to a lack of information on
its historical role in such communities and reports of negative impacts (Bennett 1965,
Vogl 1974). As with fire in other communities, timing of a fire relative to soil and
vegetation moisture conditions in a meadow can be the deciding factor in determining
a fire's overall effect (Wright and Bailey 1982). A properly timed prescribed burn can
prevent destruction of soil organic matter while decreasing litter buildup and
returning nutrients to the soil in plant-available form (Wright and Bailey 1982).
Resource ecologists at DPR had observed the effects of a 1985 wildfire occurring on
a small portion of the mountain meadow in Grover Hot Springs State Park, Alpine
County, California. They noted an apparent increase in abundance of the native
perennial grass Elymus triticoides in the drier areas of the meadow and increased vigor
of the vegetation in the burned area.

The meadow at Grover Hot Springs State Park is currently little-impacted by
human disturbances and is relatively free of weedy non-native species. Livestock
grazing and fire have been excluded for over 25 years and litter levels in some portions
of the meadow are high. Managers recognized that fire may improve vigor of meadow
species by increasing nutrient release from litter (Wright and Bailey 1982), but
desired to avoid drastic changes in species composition or invasion by alien species.
The objective of this study was to determine temperature extremes following burning
of wet and dry meadow types at a montane meadow at Grover Hot Springs State Park.
We also observed short-term changes in species composition of meadow vegetation
due to prescribed fire in order to determine the acceptability of prescribed burning as
a management tool.

METHODS

Grover Hot Springs State Park (lat 38.7°N, long 1 19.8° W) is located on the east
side of the Sierra Nevada mountains 28 km south of Lake Tahoe at an elevation of
1,830 m. Annual precipitation, based on a 30-year average from a United States
Geological Survey weather station at nearby Woodfords, California, is 50 cm. Most
precipitation (87%) falls between October and April, with more than 25 cm of snow
falling each month between December and February. The meadow area used for this
study is north of Hot Springs Creek and east of Buck Creek, covering approximately
21 ha (Figure 1 ). The southern boundary of the meadow is a thin band of riparian trees
(Salix sp. and Alnus sp.) along Hot Springs Creek, which is eroded as much as 2 m
below the level of the meadow. The eastern and northern edges are bounded by
coniferous forest, dominated by Pinusjeffreyi Grev. & Balf . in A. Murr. There are two
small 'islands' of forest within this meadow area growing on slightly higher, rocky
outcrops. Wet meadow areas had a dense cover of Carex sp. and J uncus sp., whereas
the dry areas (bordering the forest) were much more open and supported a more
diverse assemblage of species.

FIRE EFFECTS ON A MOUNTAIN MEADOW

117

N

LEGEND

Park Boundary

USFS special use boundary
Meadow
Burned area
Dirt service roads

Figure 1 . Map of Graver Hot Springs State Park, California. The meadow bordering Hot Springs
Creek is shown, along with the portion of the meadow burned in the escaped fire. Solid lines not
delimiting the meadow, or burned meadow portion, are paved roads.

We stratified our samples by hydrologic meadow type based on plant species
composition and cover. We employed a paired plot approach; pairs of circular plots,
20 m diameter, were placed in each meadow type, one member of each pair was to be
burned and one unburned. Plots were subjectively selected in pairs, and we tried to
match the vegetation of members of each pair in order to make detection of the effects
of fire easier. Members of pairs were usually within 20 m of each other in order to keep
other site characteristics constant. It was impossible to have the same number of plots

1 1 8 CALIFORNIA FISH AND GAME

in both wet and dry meadow types because most of the meadow was composed of the
wet type and many of the dry areas had such large percentages of bare soil that a
prescribed fire under high-moisture conditions would not carry through the plots.

Five pairs of dry area plots and 10 pairs of wet area plots were established in early
August 1987. The center of each plot was marked with a numbered metal stake and
the vegetation of each plot was characterized by five tosses of a 0.25 m2 sampling hoop
in which the species present were recorded and the cover of each species was visually
estimated. Species cover included all standing, live or dead, material. Litter was
defined as dead plant material lying on the soil surface. Voucher specimens of plant
species identified in this study were deposited with DPR. Plant species names follow
Munz and Keck (1968).

Maximum fire temperature was estimated with temperature-indicator crayons
(Omega Engineering, Stamford, CT). Crayons with 39 melting points were used,
ranging from 66-62 1°C. Crayons from 66-260°C had melting points separated by ca
7°C (28 crayons), from 260-343°C differed by ca 14°C (7 crayons), and from 343-
62 1 °C by 50°C (4 crayons). Temperature indicators were made by stapling a piece of
each crayon into an envelope of 1 mm mesh steel window screen. One indicator,
containing all 39 pieces of crayon, was placed near the undisturbed center of each plot
at the mineral soil surface. Litter removed during placement was replaced over the
indicator. After burning, the indicator was retrieved and the maximum temperature
determined by identifying which crayon pieces remained unmelted in the envelope.

Statistical comparisons of temperature data were made with the Mann- Whitney U
test (Zar 1984). This is a nonparametric rank-sum technique. Vegetation data was
compared using the Wilcoxon paired sample nonparametric procedure or the Mann-
Whitney £/-test (Zar 1984).

Fuel quantity, fuel moisture, and soil moisture were measured prior to burning
each plot by tossing a 0.25 m2 sampling hoop into the plot. All above-ground plant
material within the hoop was clipped and a surface soil sample (0-3 cm depth)
collected, weighed, dried at 60°C for 3 days and re-weighed to determine moisture
levels at the time of burning.

The burn prescription, based on recommendations from State Park Resource
Ecologists and complying with local fire and air pollution ordinances, required air
temperatures of 4-16°C, 20-40% relative humidity, wind speed 0-8 kph, and fuel
moisture (1 hr) ranging from 5-10%. Burning of the plots was begun 1 1 November
1987 after the meadow had been moistened by several storms. Conditions during
burning were: 2-14°C, 21% relative humidity, wind 2 kph from the west, and fuel
moisture (1 hr) 5-7%. Most of the plots were burned the first day and the fires were
of low intensity. On the second day, increased winds and lower humidity forced
cancellation of further burning. Conditions that day were: 4-20°C, relative humidity
16%, wind 8-16 kph from the west, and fuel moisture (1 hr) 4-6%. However, a
smoldering spot in one of the previous day's burned plots crossed its firebreak. The
resulting escaped fire, of low to moderate intensity, burned most of the meadow north
of Hot Springs Creek and east of Buck Creek (21 ha total), including a number of
control plots (Fig. 1); of the 10 dry plots 9 were burned, of the 20 wet plots 8 burned,
1 1 were unburned, and 1 was partially burned - the partially burned plot was not

FIRE EFFECTS ON A MOUNTAIN MEADOW 1 1 9

included in the analysis. The escaped fire was extinguished by fire fighters using fire
breaks and water.

The following summer, data on number of species and amount of live cover by
each species were collected from plots selected by tossing a 0.25 m2 sampling hoop
five times into each plot. The area of bare soil and amount of litter were also recorded,
with litter being defined as all dead plant material. Because litter was defined
differently in each of the two sampling years, species cover data were not directly
comparable. To circumvent this problem we subtracted 1988 live cover from 1987
cover for each species, classified each plot as burned or unburned, and tested for a fire
effect on species covers by the Mann- Whitney U test. In this way relative changes due
to fire could be detected by comparing the differences of burned and unburned areas.
A similar procedure was performed to detect fire effects on bare soil area and species
richness/plot.

After the escaped fire, we supplemented our original experimental design by
including another sampling area which contained both dry and wet meadow areas.
When it had been extinguished, the western edge of the escaped fire was oriented
north-south, along the main moisture gradient from dry to wet meadow types. This
border area was little disturbed by fire-fighting activities, the fire having stopped due
to a change in wind direction combined with water applied from a pumper truck. The
major disturbance were tracks left by vehicles at the border of the burned and
unburned areas. We marked this fire border in 1987 and returned in 1988 to sample
burned and unburned meadow areas. At each of 20 sites spaced every 5 m along the
border, we tossed a 0.25 m2 sampling hoop at least 5 m into the burned and unburned
parts of the meadow. The 5 m distance was sufficient to avoid the areas disturbed by
vehicle tracks and also decreased the influence of edge effects. We recorded the
species present and estimated their living covers, and clipped the unburned and
burned plots at every other site (for a total of 10 samples of each type). Dry biomass
was determined by weighing after drying at 60°C for 3 days.

One of the presumed benefits of fire is an increase in nutrient input to the soil. We
checked nutrient status of the plants in the biomass samples by measuring total
nitrogen in a ground sub-sample from each biomass sample (10 burned and 10
unburned samples) using a Leco CHN-600 nitrogen analyzer.

RESULTS

Prior to the fire, species richness (Table 1) was greater on the dry plots (24
species) than on the wet plots (9 species). Large differences existed in the cover of
species present in both areas. Carex species covered 64% of wet plots but only 9% of
dry plots. The native perennial grasses Poa and Elymus were more abundant on dry
plots, with a twelve-fold difference in Poa and no Elymus at all on wet plots. Dry
meadow plots had more litter and bare soil than wet meadow plots.

Most of the species listed in Table 1 are native. Only Senecio vulgaris, Rumex
acetosella, Taraxacum officinale, and Bromus tectorum are exotic, and of these, only
Bromus tectorum averaged more than 1 % cover (and this only on dry meadow plots).

120

CALIFORNIA FISH AND GAME

Table 1. Species covers (expressed as mean percent cover, standard error in parentheses) in
the wet and dry area experimental plots prior to burning.

Species

Wet plots (n=20)

Dry plots («=10)

Carex spp.

64.00

(5.80)

8.90

(2.40)

Juncus spp.

14.00

(3.40)

13.00

(1.90)

Achillea lanulosa Nutt.

0.31

(0.26)

0.80

(0.59)

Potentilla gracilis Doug, ex Hook.

0.52

(0.23)

1.50

(0.66)

Poa spp.

1.60

(0.68)

19.00

(3.40)

Senecio vulgaris L.

0.30

(0.15)

0.25

(0.25)

Muhlenbergia richardsonis (Trin.) Rydb.

1.20

(0.50)

1.40

(0.43)

Taraxacum officinale Wiggers.

0.01

(0.10)

0

Elymus glaucus Buckl.

0

7.70

(2.30)

Trifolium wormskioldii Lehm.

0

0.10

(0.10)

Equisetum laevigatum A.Br.

0

0.02

(0.02)

Polygonum douglasii Greene.

0

0.22

(0.13)

Stipa columbiana Macoun.

0

0.12

(0.12)

Penstemon oreocharis Greene.

0

0.17

(0.11)

Rumex acetosella L.

0

0.02

(0.02)

Vicia americana Muhl. ssp. oregana (Nutt.) Abrams

. 0

0.17

(0.14)

Bromus tectorum L.

0

1.30

(0.84)

Epilobium paniculatum Nutt. ex T. & G.

0

0.02

(0.02)

Sysyrinchium idahoense Bick.

0

0.01

(0.01)

Erigeron spp.

0

0.05

(0.04)

Others

0.16

(0.51)

0.30

(0.22)

Litter

17.00

(16.00)

34.00

(3.00)

Bare Ground

0.71

(2.30)

11.00

(2.80)

At time of fire:

Soil moisture (%)

76.0

(14.0)

23.0

(4.2)

Above-ground dry weight (g/m2)

1,260.0

(148.0)

440.0 i

; ioo.O)

Wet and dry meadow plots varied at the time of the burn. Almost three times as
much fuel was present on wet meadow plots (Table 1 , P < 0.001). Soil moisture was
also 3.3 times higher in wet meadow plots at the time of the burn (Table 1 , P < 0.005).
Temperatures at the soil surface during the fires ranged from 79 to 302°C. Wet and
dry meadow plots did not differ in mean maximum burn temperature: 153 (± 61 .2 )°C
and 185 (± 83.8)°C, respectively, P = 0.51 (note: ± value is one standard error).
Failure of fire temperatures to reflect the difference in fuel was probably due to drier
conditions in dry meadow plots at the time of burning. Two of the temperature
indicators were in wet meadow plots burned by the escaped fire. They had a higher
mean temperature when compared to the five prescribed-burned plots in wet meadow
areas, 222 (± 54)°C versus 1 69 (± 94)°C. Flame lengths during the prescribed fire were
under 0.5 m (considered a low intensity fire) but exceeded 1 m for the escaped fire
(moderate intensity).

FIRE EFFECTS ON A MOUNTAIN MEADOW 121

Burned wet meadow plots showed an increase in bare ground compared to
unburned plots 10.5 mo later. Percent bare ground/plot was 1 .7 (± 1 .2) prior to the fire
and 8.3 (± 2.0) after the fire, whereas unburned plots averaged 0.09 1 (± 0.09 1 ) in 1 987
and 0.018 (± 0.018) in 1988. There was no significant change in species richness
between burned and unburned wet meadow plots. Burned plots averaged 4. 1 (± 0.44)
species prior to the fire and 4.9 (± 0.58) after, whereas unburned plots contained 2.7
(± 0.27) species in 1987 and 3.1 (± 0.46) in 1988.

Individual species exhibited few differences in cover due to fire. Of five taxa
examined (Table 2), only Poa had less cover on burned plots one year after the fire
(Table 2). Poa species were not reliably distinguishable in a vegetative condition in
the field and hence several species are included in this result. Weedy species not
included in Table 2 were Senecio vulgaris and Taraxacum officinale. Senecio was not
observed in 1988 on any plots. Taraxacum was only recorded on one plot prior to the
fire and occurred on two others (both burned) after the fire. Cover by Taraxacum on
these plots in 1988 was low (<2%).

Examination of the fire border plots also showed few significant differences
between burned and unburned areas. Fire increased living Carex cover almost
threefold (Table 3), and increased Muhlenbergia richardsonis (Trin.) Rydb. while
decreasing cover by J uncus. Species richness was not affected by fire, but the amount
of bare ground increased almost threefold. Biomass was lower on burned plots, but
tissue moisture levels were significantly higher. We found no significant difference
in nitrogen content of plant material clipped from burned and unburned plots (Table

3).

Almost all control (unburned) dry meadow plots burned in the escaped fire,
leaving only one pair unburned. The remaining unburned pair of plots was not enough
to allow statistical analysis; however, the fire border plots cover the gradient from wet
to dry areas and include some plots equivalent to the dry meadow plots we selected
before the fire. These selected dry meadow plots averaged 9% cover by Carex (Table

Table 2. Comparison of 1987 prefire and 1988 post fire data for wet meadow plots. Species and
bare ground data are mean changes in estimated percent cover (1988 data subtracted from 1987
data, S.E. in parentheses). Statistical significance refers to results of the Mann-Whitney U -test.

Species/Characteristic
Compared

Carex spp.

Juncus spp.

Poa spp.

Muhlenbergia richardsonis

Achillea lanulosa

Bare ground
Species richness/plot

Burned

Unburned

plots (n = 8)

plots (

n= 11)

P

38.00 (8.40)

54.00

(6.80)

0.23

11.00 (4.60)

11.00

(4.50)

0.98

-2.90 (0.88)

0.35

(0.49)

0.003

0.28 (0.56)

1.00

(0.71)

0.44

-0.25 (0.17)

-0.09

(0.09)

0.38

-6.60 (2.70)

0.07

(0.09)

0.004

-0.75 (0.31)

-0.36

(0.47)

0.30

1 22 CALIFORNIA FISH AND GAME

Table 3. Cover by species for burned and unburned border plots. Statistical comparisons were
made by the Wilcoxon paired-sample test. Species, bare ground, and litter data are mean
estimated percent covers (standard errors in parentheses).

Species/trait examined

Burned Plots

Unburned Plots

P

Carex spp.

18.00

(3.70)

6.20

(1.50)

0.001

Juncus spp.

4.60

(0.61)

6.60

(1.00)

0.05

Poa spp.

2.10

(0.84)

5.50

(1.40)

0.05

Muhlenbergia richardsonis

3.80

(2.0)

0.25

(0.25)

0.05

Achillea lanulosa

0.25

(0.25)

3.50

(2.10)

0.20

All others

13.00

(1.20)

15.0 0

(1.60)

0.20

Bare ground

21.00

(4.50)

7.60

(2.10)

0.002

Litter

49.00

(4.30)

62.00

(5.30)

0.05

Species richness

3.10

(0.30)

3.80

(0.33)

0.30

Fresh biomass (gm/m2)

332.00

(81.00)

572.00

(137.00)

0.02

Dry biomass (gm/m2)

184.00

(58.00)

416.00

(114.00)

0.01

Percent water

48.00

(2.00)

31.00

(2.80)

0.01

Percent nitrogen

2.10

(0.05)

1.90

(0.12)

0.11

1). Therefore, those plots with 5% or less Carex along the fire border probably were
equivalent to our original dry meadow plots. Assuming these plots were properly
paired before the fire, these plots (five pairs) were analyzed separately to determine
the effect of fire on bare soil area of dry meadow plots. Bare soil area was almost three-
fold higher in burned plots than unburned plots: 34% (± 33) versus 13% (± 12) (P <
0.05).

Vehicles were driven across the meadow during efforts to control the escaped fire
and left ruts due to soil compaction. A year later these access routes were still visible
in wet meadow areas. In 1988, the amount of live Carex in vehicle-compacted areas
was noticeably less than in the surrounding site. We also noted that some areas had
large amounts of soil disturbance due to maneuvering of vehicles, smolders that had
destroyed plant material and litter down to bare mineral soil, and excavation of
smoldering areas by fire crews in an effort to fully extinguish the fire. These disturbed
areas were still relatively bare in 1988. During the summer of 1989, the vehicle-
compacted areas were still visible and were the only noticeable difference between
burned and unburned areas.

DISCUSSION

Neither the wet nor dry meadow sites at Grover Hot Springs fit easily into the
Sierra Nevada meadow classification system proposed by Ratliff (1982). Wet
meadow areas contained 4 identifiable Carex species (Carex aquatilis Wahl., Carex
bolanderi Olney, Carex douglasii Boott., and Carex prae gracilis W. Boott.), none of

FIRE EFFECTS ON A MOUNTAIN MEADOW 1 23

which were used in his system. His Nebraska sedge class (being less wet than the
beaked sedge class) seems the closest fit, although Juncus was more abundant on our
plots.

Dry meadow areas were dominated by Poa species, including Poa pratensis L.,
Poa nevadensis Vasey ex Scribn., and Poa scabrella (Thurb.) Benth. ex Vasey.
Dominance by Poa suggests the Kentucky bluegrass meadow class, which is found
on relatively well-drained soils and may have significant cover by Carex, Juncus, and
bare ground (Ratliff 1982). Elymus is an important component of these plots but was
not mentioned by Ratliff ( 1 982). Our search for dry meadow plots containing Elymus
undoubtedly inflated its importance.

Fire left the species composition of the meadow relatively unchanged. There was
no marked increase in exotic species or major shifts in native species composition.
Species richness in both the fire border and wet meadow plots did not change greatly
as a result of the fire. In the fire border plots, cover of some native species increased
while others decreased. Decreased biomass on the burned plots was primarily due to
reduction of litter volume, as reflected in the decreased cover by litter and the increase
in bare ground. This was also indicated by the higher moisture level of material
clipped from burned area plots, reflecting the greater amount of green plant material
present relative to litter. We should note that this study was conducted during a
drought. For example, data from a weather station maintained at the park from June
1987 until May 1988 totaled only 33% of the 30 year precipitation average from the
nearby weather station at Woodfords, California.

Fires in grasslands generally release nutrients from burned litter (Vogl 1974). In
our study, burning did not result in increased nitrogen levels in plant tissues even
though post fire growth appeared more vigorous. This may have been due to dilution
of the green plant portion of each plot by a larger amount of dead material (litter). In
a study of spring burning in Rocky Mountain meadows, Nimir and Payne (1978)
found fire increased soil pH, potassium, and nitrate levels but they did not monitor
plant nutrient levels. However, since fertilization has been shown to increase Sierra
Nevada meadow productivity (Evans and Neal 1982, Ratliff 1985), release of
nutrients by fire in our study should have had a positive effect. This positive effect
would only have occurred if released nutrients were not flushed from the soil during
the winter, because we burned when plants were dormant and therefore unable to
immediately take up any released nutrients. Because this fire was not hot enough to
destroy soil organic matter, retention of released nutrients is likely. The increase in
Carex cover (Table 3) may reflect improved growth in response to an increase in
nutrients.

The increase in bare soil after the fire appeared to be the only undesirable impact
of fire on the meadow. Dry meadow plots had such high amounts of bare soil in their
unburned state (see Table 1) that we believe they were not made more vulnerable to
invasion by exotic species due to creation of openings by fire. Some weedy exotic
species (e.g., Bromus tectorum) were present in both burned and unburned dry
meadow plots of the fire border study and showed no dramatic response to fire. In
contrast, unburned wet meadow plots had very little bare area and this was significantly

■) 24 CALIFORNIA FISH AND GAME

increased by fire. We did not detect substantial invasion of these bare areas by 1988,
but did note more Taraxacum in some areas outside of the burned plots. If open areas
are rapidly closed by the recovering native plants, increase in Taraxacum will be
limited. Overall, given the rapid recovery of the meadow one year after burning, it
does not appear that there will be an invasion problem in the wet meadow areas. The
limited evidence available (Ratliff 1985) points to natural fires occurring in meadows
at intervals of several hundred years. We therefore do not recommend a repeat of the
burn for a fairly long period of time to allow complete reinvasion of bare areas by
native species.

Wright and Bailey (1982) and Ratliff (1985) predicted that burning under
relatively wet conditions would probably not result in long-term vegetation changes
in mountain meadows. Our study supports their prediction. The drastic change in
meadow vegetation of a subalpine Sierran meadow (DeBenedetti and Parsons 1979,
DeBenedetti 1980, DeBenedetti and Parsons 1984) was primarily due to the fire
occurring during drought conditions. Our results show that burning when plants are
dormant and soil moisture is high enough to prevent organic matter in the soil from
burning prevents change in meadow vegetation and allows rapid post fire recovery.

Ratliff (1985) concluded that not enough was known about fire in mountain
meadows to prescribe its use as a management tool. We have shown that, under the
prescription used here, fire does not change meadow vegetation greatly or make
meadows significantly more susceptible to invasion by exotic species after one year.
Thus, it can be used without these effects on the vegetation. However, neither have
we demonstrated a benefit of prescribed burning sufficient to justify the costs of
conducting prescription burns. Prescribed fire may be beneficial in ways not detected
or not pertinent to the meadow at Grover Hot Springs: by accelerating nutrient
cycling, making meadows less likely to suffer from catastrophic fire during drought
years, or discouraging invasion by woody species. We hope that our work will
encourage further experimentation with fire in montane meadows.

ACKNOWLEDGMENTS

We thank Wayne Owen, Department of Geography, University of California,
Davis for many plant species identifications and D. Folkerts, A. Causey, M. Manning,
and J. Freeman for suggesting improvements to an earlier version of this paper. We
also thank the many members of the California Department of Parks and Recreation
who assisted in the planning, execution, and provision of logistical support for this
project. This work was supported by a grant from the California Department of Parks
and Recreation to M. G. Barbour, University of California, Davis. This is AAES
Journal No. 6-902695P.

LITERATURE CITED
Bennett, P. 1965. An investigation of the impact of grazing on ten meadows in Sequoia and

FIRE EFFECTS ON A MOUNTAIN MEADOW 1 25

Kings Canyon National Parks. M.S. Thesis, San Jose State University. 164 p.
DeBenedetti, S. H., and D. J. Parsons. 1979. Natural fire in subalpine meadows: a case

description from the Sierra Nevada. Journal of Forestry 77:477-479.
. 1984. Postfire succession in a Sierran subalpine meadow. American Midland Naturalist

111:118-125.
Evans, R. A., and D. L. Neal. 1982. Nutrient testing of soils to determine fertilizer needs on

Central Sierra Nevada deer-cattle ranges. Journal of Range Management 35:159-162.
Gibbens, R., and H. Heady. 1964. The influence of modern man on the vegetation of Yosemite

Valley. University of California Division of Agricultural Science Publication, 44 p.
Munz, P. A., and D. D. Keck. 1968. A California Flora with Supplement. University of

California Press. 1,905 p.
Nimir, M. B., and G. F. Payne. 1979. Effects of spring burning on a mountain range. Journal

of Range Management 3 1 :259-263.
Parsons, D. J., D. M. Graber, J. K. Agee, and J. W. van Wagtendonk. 1986. Natural fire

management in National Parks. Environmental Management 10:21-24.
Ratliff, R. D. 1982. A meadow site classification for the Sierra Nevada, California. USDA

General Technical Report PSW-60, Pacific Southwest Forest and Range Experimental

Station, Berkeley, California. 1 5 p.
. 1985. Meadows in the Sierra Nevada of California: a state of knowledge. USDA

General Technical Report PSW-84, Pacific Southwest Forest and Range Experimental

Station, Berkeley, California. 74 p.
Rundel, P. W., D. J. Parsons, and D. T. Gordon. 1977. Montane and subalpine vegetation of the

Sierra Nevada and Cascade Ranges. Pages 559-599 in M. G. Barbour and J. Major, eds.,

Terrestrial vegetation of California. John Wiley and Sons, New York.
Vogl, R. J. 1974. Effects of fire on grasslands. Pages 139-194 in T. T. Kozlowski and C. E.

Ahlgren, eds., Fire and ecosystems. Academic Press, New York.
Wright, H. A., and A. W. Bailey. 1 982. Fire Ecology, United States and Southern Canada. Wiley

and Sons, New York. 501 p.
Zar, J. 1984. Biostatistical analysis. Prentice-Hall, Englewood Cliffs, New Jersey. 718 p.

Received: 5 November 1992
Accepted: 16 April 1993

CALIFORNIA FISH AND GAME
Calif. Fish and Game 79(3) :1 26-1 30 1 993

A SAMPLER FOR QUANTIFYING THE VERTICAL

DISTRIBUTION OF MACROINVERTEBRATES IN

SHALLOW WETLANDS

JEFFREY MACKAY1

U. S. Fish and Wildlife Service

Northern Prairie Wildlife Research Center

6924 Tremont Road

Dixon, CA 95620

and

NEDH. EULISS, JR.

U.S. Fish and Wildlife Service

Northern Prairie Wildlife Research Center

Route 1 , Box 96C

Jamestown, ND 58401-9736

A sampler for quantifying the vertical distribution of aquatic
macroinvertebrates in wetlands is described. This device will facilitate
quantitative sampling of macroinvertebrates in waterfowl ecology and
related studies. Because it simultaneously collects benthic and pelagic
invertebrates, the sampler reduces bias associated with sampling
macroinvertebrates that occupy the benthic-pelagic interface of wetlands.
The sampling device also separates benthic and pelagic macro-
invertebrates into separate vertical profiles to facilitate studies of
distribution patterns or the influence of chemical and physical gradients
on invertebrate vertical distribution.

INTRODUCTION

Interest in quantifying aquatic macroinvertebrates in wetlands is increasing
because they are consumed by waterfowl (Drobney and Fredrickson 1979, Swanson
et al. 1979, Euliss and Harris 1987) and shorebirds (Hicklin and Smith 1979,
Baldassarre and Fischer 1984, Boates and Smith 1989). They are also important in the
epizootiology of avian botulism (Bell et al. 1955, Jensen and Allen 1960, Duncan and
Jensen 1976) and are valuable as biological indicators of pesticide residues in
wetlands (Grue et al. 1989). Consequently, there is a need to develop a variety of
quantitative sampling gear for aquatic invertebrates to address broad ecological
questions. In this paper, we describe a tube sampler that separates the water column
and benthos of shallow wetlands into distinct vertical profiles. We developed the
device to describe vertical distribution patterns of invertebrates in wetlands for
research on avian botulism epizootiology and microhabitat selection by wetland
macroinvertebrates in relation to vertical abiotic gradients.

The described sampler (Fig. 1) is similar to the multiple tube sampler described
by Euliss et al. (1992) in that benthic and pelagic invertebrates are quantified

' Present address: Ruby Lake National Wildlife Refuge, HC 60 Box 860, Ruby Valley, NV 89833

126

A VERTICAL SAMPLER FOR QUANTIFYING MACROINVERTEBRATES 1 27

/

143 cm

0.1 cm

In Open Position

In Closed Position

Figure 1 . Schematic view of sampler in open and closed position. Numbered parts are: 1 ) Radiator-type hose
clamps; 2) Closure cable handle storage hooks; 3) Closure cable handle; 4) Sampler handle; 5) Braided
nylon line; and 6) Sediment tube section door (in open position).

1 28 CALIFORNIA FISH AND GAME

simultaneously. Thus, bias from double sampling does not occur because separate
devices are not required to sample benthic and pelagic organisms (Euliss et al. 1992).
The portion of the tube that samples the water-column is partitioned into individual
compartments to facilitate sampling of specific profiles in the water-column.

The sampler is constructed by stacking several interconnected 10.1 -cm ID (1 1.3-
cm OD) clear acrylic-tube water-column samplers on a bottom tube that samples
benthos (Fig. 1). We used 15.2-cm long tubes, but tube length may be modified for
individual needs. Several acrylic doors isolate water-column samples into separate
vertical profiles. The leading edges of the doors are sharpened to cut minor vegetation
as they are closed. The bottom of the tube is also sharpened to facilitate penetration
of sediments. Acrylic doors (flat 6-cm-thick acrylic) are integrated into the base of
each compartment (Fig. 2) to physically block the movement of invertebrates among
profiles. A 9-cm-diameter 0.5-mm-mesh stainless steel screen is attached with clear
silicone adhesive to the center of each door. The doors are sanded to reduce their
thickness and to facilitate rapid closure. A laminated acrylic sheet is attached flush
with the top of each compartment to form a firm attachment plate with the adjacent
compartment. Acrylic cement was used to bond door housing units and joint plates.

The top tube section of the sampler must be long enough to facilitate placement
of sampler handles and brackets used to lock sampler doors. Sampler handles and the
door closure handle brackets are attached with standard hose clamps to the top tube.
Handles may be obtained from hardware stores and the two brackets were constructed
from 16 mm x 7 mm (0.25 inch) rods welded to a metal plate.

The water-column compartments and the sediment tube are stacked vertically and
coupled together at the corners with 7 mm bolts. Wing nuts are used to facilitate
breakdown and assembly of the sampler in the field.

Compartment doors are secured to the closure cable with a loop of small-diameter
braided nylon line. Nylon line is inserted through two holes in the rear of each door
housing and then into two holes drilled in the leading edge of the door. A knot in the
nylon line prevents the line from pulling out of the doors when they are closed. A snap
swivel attaches the loop of nylon line to the closure cable. The closure cable is
constructed with 1.5 mm plastic-coated braided wire cable, a 7 mm OD wooden
dowel, and a handle of 25 mm OD wooden dowel. The 90 cm wooden dowel is
attached to the wire cable to prevent tangling.

A sample is collected by lowering the sampling device into the water with doors
in the open position. The bottom tube section is forced into the sediment to the joint
plate of the sediment compartment. Immediately after the sampling device is seated
into the sediments, the closure cable is raised to simultaneously close all doors. The
sampler is then rocked back and forth to break the sediment seal and lifted to the
surface. Individual water-column compartments are disassembled and the samples
washed into specimen jars. Re-assembly of the sampler averages 2 minutes per
individual water-column sample.

Benthic samples are removed from the sampling device by hand. The core is
forced through the top of the tube by hand until the protruding core is the desired depth
from the benthic-pelagic interface. The door is then closed to sever the sample from

A VERTICAL SAMPLER FOR QUANTIFYING MACROINVERTEBRATES 1 29

Figure 2. Expanded view of water-column compartment and door closure unit. Numbered parts are: 1 ) Tube
compartment; 2) Screen; 3) Sliding door; 4) Cable line clamp;5) Barrel swivel;6) Nylon tie; 7) Snap; 8) Bolt;
9) Joint plate (top section); 10) Wood dowel; 1 1) Plastic coated wire cable; and 12) Wing nut.

the remaining core. Individual sediment cores are then placed in a modified self-
cleaning screen (Euliss and Swanson 1989) to concentrate sample residues and placed
in specimen jars.

We used the described sampling device for two field seasons to collect
macroinvertebrates in shallow wetlands. Our study sites contained minimal submergent
and emergent vegetation that would interfere with the operation of this device. The
sampler withstood rigorous use and required only occasional minor maintenance of
the cable used to close the doors. Materials for the sampling device cost about
$175.00. Total construction time was about 8 hours.

1 30 CALIFORNIA FISH AND GAME

We thank H. R. Murkin, F. A. Reid, and G. A. Swanson for critical review of the
manuscript and D. M. Mushet for preparing figures.

LITERATURE CITED

Baldassarre, G. A., and D. H. Fischer. 1 984. Food habits of fall migrant shorebirds on the Texas

USA high plains. J. Field Ornithol. 55:220-229.
Bell, J. F., G. W. Sciple, and A. A. Hubert. 1955. A microenvironment concept of the

epizootiology of avian botulism. J. Wildl. Manage. 19:352-357.
Boates, J. S., and P. C. Smith. 1989. Crawling behavior of the amphipod Corophium volutator

and foraging by semipalmated sandpipers Calidris pusilla. Can. J. Zool. 67:457-462.
Duncan. R. M., and W. I. Jensen. 1976. A relationship between avian carcasses and living

invertebrates in the epizootiology of avian botulism. J. Wildl. Dis. 12:1 16-126.
Drobney, R. D., and L. H. Fredrickson. 1979. Food selection by wood ducks in relation to

breeding status. J. Wildl. Manage. 43:109-120.
Euliss, N. H., Jr., and S. W. Harris. 1987. Feeding ecology of northern pintail and green-wing

teal wintering in California. J. Wildl. Manage. 51:724-732.
, and G. A. Swanson. 1989. Improved self-cleaning screen for processing benthic

samples. Calif. Fish and Game 75:124-128.
, , and J. Mackay. 1992. Multiple tube sampler for benthic and pelagic invertebrates

in shallow wetlands. J. Wildl. Manage. 56:186-191.
Grue, C. E., M. W. Tome, T. A. Messmer, D. B. Henry, G. A. Swanson, and L. R. DeWeese.

1989. Agricultural chemicals and prairie pothole wetlands: meeting the needs of the

resource and the farmer - U. S. perspective. Trans. N. Am. Wild. Nat. Resour. Conf. 54:43-

58.
Hicklin, P. W., and P. C. Smith. 1979. The diets of 5 species of migrant shorebirds in the Bay

of Fundy, Nova Scotia, Canada. Proc. Nova Scotia Inst. Sci. 29:483-488.
Jensen, W. I., and J. P. Allen. 1960. A possible relationship between aquatic invertebrates and

avian botulism. Trans. N. Am. Wildl. Nat. Resour. Conf. 25:171-181.
Swanson, G. A., G. L. Krapu, and J. R. Serie. 1979. Foods of laying female dabbling ducks on

the breeding grounds. Pages 47-57 in T.A. Bookhout, ed. Waterfowl and wetlands— an

integrated review. Proc. Symp. North Cent. Sect., The Wildl. Soc.

Received: 7 January 1992
Accepted: 31 March 1993

CALIFORNIA FISH AND GAME
Calif. Fish and Game 79(3):1 31 -1 32 1 993

TULE ELK RELOCATED TO BRUSHY MOUNTAIN,
MENDOCINO COUNTY, CALIFORNIA

KENT B. LIVEZEY

Department of the Navy

Engineering Field Activity, Northwest

Natural Resources Management Section

3505 N.W. Anderson Hill Road

Silverdale, WA 98383-9130

Nineteen tule elk (Cervus elaphus nannodes) from Grizzly Island Wildlife Area,
Solano County, California and 1 1 tule elk from Tupman Tule Elk Reserve, Kern
County, California were relocated to Brushy Mountain, Mendocino County, California
on 30 July and 1 8 September 1 986, respectively, as part of the California Department
of Fish and Game tule elk management program (Koch 1 987, Livezey et al. 1 988). The
elk from Grizzly Island were captured by a helicopter/corral system (Koch 1987), and
the elk from Tupman were immobilized with the drug carfentanil (Mueleman et al.
1984). Ten of the 30 elk were fitted with radio-telemetry collars and all 30 elk were
eartagged.

A total of 221 radio-locations (57%) or observations (43%) of radio-collared elk
and more than 130 observations of non-collared elk were collected from July 1986
through October 1988 (Livezey 1987, 1988, and 1991). The two groups of released
elk fragmented into at least 1 2 smaller groups or individuals that travelled as far as 28
km from the release site. There were eight confirmed elk mortalities during the study
period. Renarcotization of the carfentanil (Stanley et al. 1988) caused the death of two
cows and a male yearling. Four cows and one bull died of unknown causes. One of
these, a cow, was killed and eaten or scavenged by a black bear (Ursus americanus),
and another, a bull, was scavenged by coyotes (Canis latrans). Two cows examined
within 10 hours and two days of death, respectively, showed no apparent cause of
death and had not been scavenged.

The relocation efforts were a success; two small subherds of elk were established.
The main group, which settled in the Brushy Mountain/Eden Valley area 0-6 km from
the release site, was comprised of at least five cows, three bulls, two female yearlings
and three male yearlings. The other group, which established itself in the Laytonville
area 1 8-20 km west of the release site, was comprised of one cow and one bull that
joined with two cows, a female yearling, and a calf that were individuals or their
decendents from a 1979 or 1980 tule elk relocation (BLM 1983). No elk calves were
observed during the 1987 calving season, but at least five 1988 calves were born from
cows released in 1986.

ACKNOWLEDGMENTS

This project was financed through California State Contract C- 1 624. D. Koch and
J. Booth provided very helpful guidance, expertise, and review. J. Booth supplied

131

1 32 CALIFORNIA FISH AND GAME

locations of elk from June through October 1988.

LITERATURE CITED

BLM. 1983. The tule elk in California: 7th annual report to Congress. Bureau of Land

Management, Sacramento. 40 p.
Koch, D. B. 1987. Tule elk management: problems encountered with a successful wildlife

management program. Trans. West. Sect. Wildl. Soc. 23:1-3.
Livezey, K. B. 1987. Radio-telemetry study of tule elk released near Eden Valley, Mendocino

County, California during summer 1 986. Calif. Dept. Fish and Game Final Rep., Sacramento.

State Contract C-1624. 64 p.
. 1988. Protective frame for a 2-element, hand-held Yagi antenna. J. Wildl. Manage.

52:565-567.
, D. B. Koch, and J. W. Booth. 1988. Mendocino tule elk management plan. Calif. Dept.

Fish and Game Rep., Sacramento. 25 p.
. 1991. Home range, habitat use, disturbance, and mortality of Columbian black-tailed

deer in Mendocino National Forest. Calif. Fish Game 77:201-209.
Mueleman, T., J. D. Port, T. H. Stanley, and K. F. Williard. 1984. Immobilization of elk and

moose with carfentanil. J. Wildl. Manage. 48:258-262.
Stanley, T. H., S. McJames, J. Kimball, J. D. Port, and N. L. Pace. 1988. Immobilization of elk

with A-3080. J. Wildl. Manage. 52:577-581.

Received: 10 November 1988
Accepted: 22 June 1993

94 84723

INSTRUCTIONS FOR CONTRIBUTORS

EDITORIAL POLICY

California Fish and Game is a technical, professional, and educational journal
devoted to the conservation and understanding of fish, wildlife, and native
communities. Original manuscripts submitted for consideration should deal with
California flora or fauna, or provide information of direct interest and benefit to
California researchers and managers.

MANUSCRIPTS: Refer to the CBE Style Manual (5th Edition) and a recent issue of
California Fish and Game for general guidance in preparing manuscripts. Specific
guidelines are available from the Editor-in-Chief.

COPY: Use good quality 215 x 280 mm (8.5 x 11 in.) paper. Double-space throughout
with 3-cm margins. Do not hyphenate at the right margin, or right-justify text. Authors
should submit three good copies of their manuscript, including tables and figures to
the Editor-in-Chief. If written on a micro-computer, a 5.25 or 3.5 in. diskette of the
manuscript in word processor and ASCII file format will be desired with the final
accepted version of the manuscript.

CITATIONS: All citations should follow the name-and-year system. See a recent
issue of California Fish and Game for format of citations and Literature Cited. Use
initials for given names in Literature Cited.

ABSTRACTS: Every article, except notes, must be introduced by an abstract.
Abstracts should be about 1 typed line per typed page of text. In one paragraph
describe the problem studied, most important findings, and their implications.

TABLES: Start each table on a separate page and double-space throughout. Identify
footnotes with roman letters.

FIGURES: Consider proportions of figures in relation to the page size of California
Fish and Game. Figures and line-drawings should be of high-quality with clear, well-
defined lines and lettering. Lettering style should be the same throughout. The
original or copy of each figure submitted must be no larger than 215 x 280 mm (8.5
x 1 1 in.). Figures must be readable when reduced to finished size. The usable printed
page is 117 x 191 mm (4.6 x 7.5 in.). Figures, including captions cannot exceed these
limits. Photographs of high-quality with strong contrasts are accepted and should be
submitted on glossy paper. Type figure captions on a separate page, not on the figure
page. On the back and top of each figure or photograph, lightly write the figure number
and senior author's last name.

PAGE CHARGES AND REPRINTS: All authors will be charged $40 per printed page
and will be billed before publication of the manuscript. Reprints may be ordered
through the editor at the time the galley proof is submitted. Authors will receive a
reprint charge form along with the galley proof.

_. o

m

^. w

Q.

-*' =5;

^»

o> 0

0

-^

7 3

3' o>

0

s

3- 0

r-

22 -8

^

- o>

O

0) §■
w 5

33

J*

0 0

^ =•

r*

0) .-+

3 a

CD ^

33
(0

Fish
nto,

a:

O 03

5

> 3

0

Game
95814

2

T3 to

c

<D O

CO ro

11

■0

• c

>

0 *

CO 33

° O

D

d >

<o '

> H

& O

Q m

<D >

m