--

Published October 13

San Francisco State University, Bomberg nburon Center. 3150 Paradbe Drive. Tiburon. Califoraia 94920. USA 'Caliioraia Department of Fish and Came. 4001 N. W k o n Way. Stockton. Calltorma 95209. USA

ABSTRACT The clam Porsmocorbda amurensls was lnuoduced into the fan Franasco Bay esm- Icaliforma. USA1 in 1986 and became abundant in late 1987. With a year. chlorophyll concentration and the abundance of adults of 3 common estuarine copepoa speaes had dedined by 53 to 91 Sb, pro- vidrng an opponunitf to examine mechrnlrms by wbch b e n h c grazing might control the abundance of pelagic populanons. Decbes in chlorophyll and aoundance of the 3 speaes of copepod coinaded appmxamatelg with the geoqrapluc range of the dam populaaon. The decline m abundance of the copepod Ewytemora a f k was accompanied by a aeaease ra the ratio of nauplii to adults. but nor in the ratio of eggs to females. Therefore. the decline in abundance may be due to elevated monality of neuplii rather than food lirmtation of reproductive rare. We argue that direct predaeon by P. amweasis is the cause of the reduced survival of nauplk. and therefore of the depressed abundance of adults. Expenmentally derermined clearance rates of P. amrrrePrir on E. ~ ~ ~ L Z U S nauplii averaged 0.11 1 clam-' d-'. If that clearance rate applied in the field. the clams could remove 8.2% of the nauphi d-I. T ~ I S removal rate is suffiaent to explain the observed rate of population decline. P. amwepsis appears to have become well established, and copepod populations of the bay so far have failed to rebound. Thus ths invasion may have permanent effecrs. In a broader sense. predanon on zooplankton by soft-bonom benthos may be an zmponant, and heretofore overlooked, source of monalrty in shallow waters. . f elecriviry occurring tbrough differences in escape r.esponse and vemcal position could make bivalve predation an imponant rrrfluence on biomass and spedes composition of inshore zooplankton.

-

KEY WORDS: Potamocorbula amurensis . E v e m c r a afdnir. Predation . San Francisco Bay . Cope-

WraODUCLlON mference from reduced se t thg rates of meroplankton near filtering adults (Cowden et al. 1984, Sebens &

Benthic grazing is believed to control phytoplankton Koehl 1984. Hunt et al. 1987, Ertmar~ & Jumars 1988, biomass in some shallow waters (Cloern 1982. Officer et Young & Gotelli 1988, but see Young 1989). al. 1982. Nichols 1985. Alpine & Cloern 19921, A large Predation on copepods has b~ described for body of literature indicates that certain benthic organ- benthic species including brittle stars and cnidarians isms can graze substantially on phytoplankton and (Ferrari & Dearborn 1909, Sullivan & Banzon 1990. mall zooplankton (Wiiams 1980, Buss &Jackson 1981, Sullivan et al. 1991), but not bivalves (Homed et al. Wright et al. 1982. b a n & Juman 1988, Bingham & 1988). Since bivalves are an important component of Walters 1989. Madsaac et al. 1991). Bivalves can influ- shallow soft-bottom benthos (Nichols 1985), a signrfi- ence benthic recruitment or reduce the abundance of cant predation rate by bivalves could be a key lim~ta- certain zooplankton by consuming small. relatively tion to zooplankton abundance and speaes composi-

C Inter-Research 1994 Rere;* of fuU article nor p e m n e d

W-593

w

82 Mar. Ect;l. Proa. Srr. 113: 31-93, 1994

The clam Potamocorbula amurensis arrived in San Franasco Bay. California. USA. in 1986. probably in s h p ballast water (Carlton et al. 1990, Nichols et al. 1990). P. amurgnas is a small (5 to 30 mm) dam with an epibenhc habitat and apparently euryhahe distribu- tion: it has been found at freshwater and seawater s h t i e s . Its type locality is the &nur River in eastern Siberia. Lttle is known of its dismbution. abundance. - or ecology in the source reqon (Carlton et al. 1990).

Potamocorbula kmurensir spread very rapidly. be- coniing abundant throughout the northern reach of the bay by mid-summer 1987, a period of drought (Carlton et ai. 1990). Ths spread apparently occuged at the e?pense of other salt-tolerant benthicspecies. which, under these drought conditions, would ordi- narily have invaded the northern reach by 1987-88 (Nichols et al. 1990). Ths invasion, potentially disas- trous to the food web of the upper estuary, offered a unique oppormnity to observe the effects of benthic filter feeding on the plankton of the escuay;. Alpine & Cloern (1992) have demonstrated the influence of r h s dam on the phytoplankton of the estuary, while Werner & Hollibaugh (1993) have shown it to be capable of consuming bacteria. although at a lower ,rate than phytoplankton.

C o n m e n t with the spread of Potamocorbula amu- .-. . rensis came a decrease in chlorophyll of about 5-fold

and elimination of the spring bloom in 1988 to 1990 i (Alpine & Cloern 1992). This period has included the 4

longest drought since the 1930s. and the dimmution of chlorophyll was much greater than seen in a shorter but more severe drought in 1976-77 (Arthur & Ball 1979, Cloern et al. 1983. Alpine & Cloern 1992). At the same time the abundance of several common estuarine copepod speaes declined markedly, leading to the belief that the dam was qfluendng copepod abun- dance either indirectly through redudon of food supply or directly through predation. In t h . ~ ~ paper we describe circu~tant ial evidence

suppordng the theory that the rapld dedine in cope- pod abundance was caused by predation by the dams on copepod n a u p g - a g e r tJ~-t&ough reductton in

----.-- .__- _..- food supply. We focus on abundance patterns of the copepods Eupremora af f in~~. Sinocafaaus doerrii, and Acaro'a spp.. wbch were historically the most abun- - dant speaes in the upper estuary (Ambler et al. 1985. Orsi d :Mecum 1986). In particular, E. a f h ~ was of interest because it occupies the saliruty range 1 to 6 psu (Oni & .Mecum 1986), where a large part of the monitoring effort has been concentrated and where the dam was known to be abundant IAlpiue 3r Cloern 1992). E. aff- is a wdeiv distributed estuarine pelaqc copepod with a broad range of s a h t y toler- ance (ca 0 to 15 psu: Roddie et al. 1984. Nagaraj 19921. Unlike .4cartia speaes, wtuch are broadcast spawnen.

or S. doerrii. wbei. releases eggs m dumps 01 up to 4 , E. a f f i ~ ~ . ~ carries its eggs in a slngle egg sac. S. doeri; was inaoduced ;o the estuary around 1979 and hac since been abundant at a position slightly upseeam 0: the center of po~ulatmn of &. a f b !Orsi et al. 1983).

San Frandsco Bay (Fig. 1) is one of the most urban- ized estuaries, with one of the most closely managed freshwater flow regunes, in the world (Nichols e! aL 1986). Most of the freshwater flow enters the estuary from the Sacamento and Sari Joaquin Riven. which discharge tbrough an edensive delta. We concentrate on the reach of the estuary including ~uis& Bay and the western delta (Fig. 11, where Pota- mocorbufa amurensis is abundant (Hymanson 1991). This reach indudes broad shoal areas and several deep (20 m) channeis. Tidal currents in f2u.s region can exceed 2 m s - '.

Freshwater flow data were obtained from the C&- fornia Depamnent or Water Resources (DWR) DayRow program. This proqram calculates a hypothetical net Freshwater outflow through the westein delta into the estuary at the nominal location of Chipps Island (river lan 75. Fig. 1). Net delta outflow is Lhe difference between inputs to the delta. consisting of .gaged and ungaged river flows and preapitation less evaporation. and withdrawals. including consumption by farms w i t b the delta and export flows at several major pumping facilities iFig. 1). Monthly net delta outflow estimates were corrected using recently revised esti- mates of consumption in the delta from DWR ('I. Kadk DWR. pen. comm.!.

Zooplankton data. Zooplankton abundance data were obtamed from *e California D e p m e n t of Ti and Game zooplanicton monitoring program. for,which methods are described more fully in Orsi & Mecum

'

(1986). Samples were taken at ca 35 to 10 StatioS throughout the upper estuary (Fig. I ) , monthly a March and November and twice monthly in April , to October. 1971 to 1992. Samples were taken with a 154 p m mesh, 10 cm diameter Clarke-Bumpus zet towed obLiquely from the bottom to the surface. and J with a pump whose mtake was slowly raised through , .

the water column. i he net sample and a 1.7 1 corn- posite subsample of the pump discharge were Pr* serfed in 5 % formaldehyde. f

In the laboratory the zooplankton net sample concenmted on a 134 pm mesh saeen and sub- pled to obtain at ieast 200 amm& for, coundng. me pump sample wqs sassed through a 154 Irm mSk ,

saeen and collected on a 43 pm mesh screen. and the entire sample was counted. ~ d u l t copepods were iden-

F. bunerer er al.: Clam preaauon orqcopepoas 83

.... .... .,..- ..... -.... ... ._-..v*-.L ..C -*.

...... :::

-. _-.. ......... . ._. ; .'..-- ...-. "" -. . . . ... . .

. i . . - .-.-... ... <, .'. .' .-. I

..... "-. . . . . . . . . . . . . . 100 R;l ,ypl l . . . I

. . . I ... .-,. .... ." ., .. _... .*: ': a v ---. .*

- . . . ....._ ;'

...... .............

--., ,,;-

........ ...

.... . . . . .... --. ....

..... __ .." . .......... ........"."

-.. :..? ,:...*%"."A: ..,. .... - .,. : , , : :*'

'.^" -i-.., .. ..,.-. - . ...- .h. .... 3:: .:: :.A" ..... - ...'"-%,. .-.. : . . . .

i:".."' I.. .... , ... I

Fig. 1. Upper San Francisco Bay estuary, showing location ofstations used in sampling and river kilometers mentioned in the text. Arrows at the boaom of the figure indicate the sites of the 2 major expon pumping facilities

I tified to species or genus. Nauplii. were identified'to . ,

species for Euryremora affinis only for 1985 to 1989. 'Abundance was calculated as the sum of abundance I values trom the net and pump samples after fractions-

- tion. Ratios of nauplii to adulb were calculated for 1985 to 1989; if the egg production and egg mortality rates are constant this ratio is inversely related to the ,mar- : I t d t y rate of nauplii (Kimmerer 1987).

Egg ratios of Eurytemora affiais were determined using 62 samples from spring (March to May) 1984 to 1990. Samples from other seasons were not chosen

_ because there were usually too few E. affinis in sum- mer and fall after 1987. Females and eggs in entire samples were counted if available: in some cases only subsamples had been archived. Eggs were identified as either attached. in detached egg sacs, or as loose eggs. The latter could not be unequivocally identified

to speaes, and were most abundant when Acartia spp. were abundant, so only eggs in egg sacs were included in the counts. Females were counted whether or not they had eggs, so that the mean population egg ratio was being estimated. Egg ratios are proportional to egg production rate at a constant egg development time, which depends on temperature (McLaren 1965, Edmondson 1968); temperature had no systematic variation before versus after 1987. At a constant tem- perature, egg ratios and egg produdon rate can be M t e d by food supply, either because of reduced number of eggs per dutch or an increased interval between clutches.

A vertically stratified series of 0.5 rn3 pump samples was taken on July 21 to 22. 1986 at about 1.2 psu . salinity, near the maximum in abundance of Eury- femora affinis. Samples were taken hourly at 1 m, mid-

>tar. Ecol. Prog. Ser. 113. 81-93. 1094

depth, and 1 m off the bottom. and all E. &finis were idennhed to Life stage and counted. For the present purposes. the abundance of nauplu at each depth was used to determine theu relatlve abundance rn the near-bottom samples, and thereby therr relanve vul- nerabhty to benthlc predanon.

.At each station surface speczfic conductance and temperature were measured with a conductiv~ty meter, and samples were taken for chloroptiyll a analysls begmmng in 1976. Salmty was detennrned from specific conductance according to the Practical Salinity Scale (UNESCO 1981) and is thus q v e n in pradcal s&ty units (psul. .Additional sahmty data were rob- tamed from dady means of surface speafic conduc- tance from a contmuous monitoring station at river km 75 (DWR station near Chipps Island. Fig. 1).

Data analysis. Chlorophyll a concentrations and abundance of copepods are :losely related to salinity, with peaks at intermediate sah i t i es and lower values at hgher and lower salimties (see Fig. 2). The salt field moves considerably with changes in freshwater flow (Peterson et al. 1975), so that the abundance peaks move as well. Temporal variation in abundance can therefore be confounded with variation due to. move- ment of the salt field; s ~ d a r l y . seasonal effecrs can obscure longer-term trends. Since we were mterested in temporal patterns other than those due to season and to movement of the salt field, we elirmnated these effects from the data as follows. Data from 1972 to 1987 were partitioned into 20 s a h t y dasses. each having about 500 obsenrations (Table 11. These sa lh ty classes have mean sahuties at roughly equal logarithrmc

intervals. Fint we ploned the mean chlorophyil a and log abundance (with 10 added to allow ior zerovaluesj of each copepod speaes in each salinity class before and aiter ihe end of 1987 to determine at what salm- ties the declines occurred.

The next step was to determine the geographic range and therefore the saliruty range likely to be affected by the dam. In 1991 the abundance of Poraqocorbula

1 i, amurensis was over 1000 rne2 up to about nver km 75,

and between 1 and 1000 m-2 to about k n 90 (Hymanson . 1991). The tidal excursion is about 10 krn in thu regon: i thus the effect of the dam population should be de- .$r

tectable up'to a nominal position of nver !cm 100. k corresponds on average to a surface saiirzlty of about 0.45 psu (Fig. 3), or s b t y class 10 (Table 1).

Abundance anomalies were then caiculated using a l l the data in salinity dasses 1 10 (2 16 for ..\carria speaes because they are not abundant at lower sahities). The mean of each combinadon of s&ty class and month was subtracted from the raw data to obtain anomalies virtually independent of saLin~ty and persistent sea- sonal variation. These anomalies were then combined into monthly means for determining the timing of step changes in abundance unng m e series analysls. Miss- ing data, parcicul.uly for the months of December to February, were filled m using linear interpolation to satisfy the requrrements of the data for time series analysis. We then applied intervention analysis to determine the magnitude and tirmng of the largest step changes in the data. Intervention analvsis is useiul in detemng step changes m data with sigmficant auto- correlation I Shumwav 1988).

, , - = 10 a f 6 2 a

.. .,.. 2 1 . 5 ......... Fig. 2. San Franuso Bay esnrary.

0 . . . , , . . 1.0 , . , , , . , . . , Chlorophyll and log abundance 0 2 4 6 8 10 12 14 16 18 20 l2 l6 [number m-' 10) of adults of

common copepod specles in 1-0 salinity classes dunng 1972 to

3.0 1987 (solid Line with 95 46 confi- dence limits of the me- sene5 began m 1976 for cbiorophyL

I 1979 for Sinocalanus doerrii) and 1988 to 1992 !dotted line).

9 Dark bars along the x - a m- g 1.5 dicate s U t y classes in wbch A differences were s~gnrficant It-

. . 1 .O . , . , test on annual means. p < 0.05).

I 2 4 6 3 10 12 14 16 18 20 0 2 A 5 8 10 12 .14 la 1 8 20 la) Chlorophyll: (bl Eqemora clifias: I C I Siaocaldnus deed:

Salin~ty Class Salinity Class (dl Acama spp.

Kmunerer et d: Clam preaaaon on copepoa. : 8 2

10: ,. a srnaie senes of amfiaal predation experiments i i . ' I $ 1 ,

' i-- I - - a ? . . 7s ISinparaJah 19691. A Pastex pipet attached to a length I ' I of tublng was used to siphon 150 ml of water out oi

250 mi oi water contammg E. aftrrus nauph of mixed stages, in 225 rmn. The mouth opemg of the pipet was.about 1 mrn. or abou: lox the width of the iarger nauplii. The nauplii in eacn fraction were counted and compared to the expected vaiue, based on the propor- uon of water siphoned. Sote that the flow veloaty was

. considerably tugher than expected m a dam siphoc, so these results mhcate oniy qualitadvely the possibihty

0.1 87 88 89 90 91 92 93 that escape response explains differences in frequency

YEAR - of capture i~? ciams. A senes of predation emerirnents was conducted to

Fig. 3. San Franaso Bay estuary. Surface s b t y from the zooplankton s w e y s at nver b 75. 90. and 100 for 1987 to 1991. The honzonta b e IS the geomemc mean s b r y at

100 km averaged f~om rmd-1987 to the end oi 1991

Experimental work. Experiments were conducted to ,determme whether dams could consume copepod naupb. and some aspects of the mechanisms of capture. Clams for experslents were gathered f ~ o m hallow locations m the uppe: estuary and maintained in aquaria with slowly flouing whole surface water for a few days before experiments. Copepod nauph were collected by concentratiug surface water from about 5 psu salan~ty through a 35 pm mesh saeen using upward flow to minimize damage. On one occasion nauplii from about 30 psu were used, and dams were gradually acclimated to the higher salinity. Nauplii were maintained for a few hours before experiments in insulated buckets of unscreened surface water at the experimental s b t y and temperature.

We observed ingestion of copepod nauplii by Pota- mocorbula amurensis in perri dishes under a d i s s e d g microscope. Nauplii of Eunrtemora affinis, Acarh'a spp., Oithona spp.. and ~ u t e ~ i n a acutifrons were starned with neutral red to aid in observation and, in some of the observations, the clams were glued to the peni &shes. To detect escape responses, we conducted

Table 1. SaLarty classes used m the analyses with upper l h t s of s a h t y bractical s a h h . units, psu) in each class

Saliruty Upper S a h r ) . Upper Saliruty Upper class salicllry dass saliruty class salinity

1 0.08 8 0.22 15 3.91 2 0.09 9 0.31 16 5.32 3 0.10 10 0.45 17 6.98 4 0.11 11 0.69 18 8.95 5 0.13 12 1.08 19 11.97 6 0.14 '13 i.79 20 20.17 7 0.18 14 2.71

estlmate the clearance rate of Potamocorbula amu- rensis on naupk. Nauph were bvided among 8 insu- lated 40 1 tanks containing 30 1 of water from the same samphg site at temperatures slightly above ambient (20 to 35OC). Mixing m the tanks was provided by a .slowly fca 1 rpml rotafins paddle. Clams were col- lected several days earlier and maintained in an aquar- ium with flowing whole suriace water and an a~ stone. To begin the expenmen: 250 to 300 dams (mean shell length 8.5 mmj were placed on plastic mays whch were lowered into 4 of the tanks. The experiment was terminated by removing the dams, then draining and rinsing the tanks through a 35 p mesh saeen an& preserving the sample. Subsamples were taken with a stempel pipet. 200 to 500 naupb were counted. and the number per Liter in the tank was calculated from the fraction sampled and tank volume. Grazing rates were calculated as:

v (No\ 9 = hi- h l

where _q is the grazing rate (1 darn-' h-I). V is vol- ume of the tank (I), t is duration of the experiment in hours. C is the number of dams, No is the initial abundance, and N the final abundance of nauplii in the tanks (I-'). In a series of prehmary experiments, consumption

rates of nauplii of several copepod species (Euryternora affinis. Paracalanus quasimodo. Acartia spp., Oitbona fpp.. and Euterpina acutifrons) were measured. Twelve clams were placed in each of four 1 1 beakers, and no dams in 4 controls. Nauplii obtained by rearing from field-caught females were added to each beaker in 700 rnl of water from the salinity range at which the copepods were collected. After 2 h incubation near ambient temperatures laround 20 "C), during which beakers were stirred manually eveq 10 min and dams were observed filtering, the contents of the beakers were strained through a 35 pm mesh saeen, stained with neutral red, preserved with formaldehyde, and . . counted.

Table 2. Log chlorophyll and log abundance (number m-' +

10) oi 3 common copepods m the San Franasco Bay esruary before (1971 to 1987) and alter (1989 to 19921 the dam Poramocoroula a m w e m became abundant. Values are means and 95?& coddence h t s of annual mean values for sakruty classes m wtuch s~gruf~cant dedines occurred (dark bars 'in Fig. 1), and percentage dedines m h e geo-

memc means

Before 1988 1988-1997- Percent change

Chlorophyll 0.74t0.10 0.23=0.18 89% Euryremoraaffinrs 2.4~0.1 1.3 = 0.2 93 % Sinocaldnus doerrii 2.3 a 0.2 1.q = 0.2 76 % Acdnid Spp. 3.3 a 0.2 2.1 2 0.1 95 %

L

Table 3. Results of intervennon analvsls on variables in the San Franaxo Bav estuary to determrne t m g of change tn anomahes of log chloropnvU and log abundance (number m- I - 10) of 3 common copepods ior sahmt)r dasses 10 to 20 (16 :o 20 for Acaraa sup.1. Values gven are the t value for tesmg the signrhcance of the largest step change identified m the data. the date (month and year! at which b a t change oc- curred. and the percent change detennaned as (1 to loc).

where cis the coeffiaent of the step change

1

t-value Date Percent change

Chlorophyll -6.0 Jun 1987 -53 ?/o E t , u p e m o r a a f . -10.0 Jan1988 -88 ' 6

,;- Sinocaianus doemi -7.0 Sep 1988 -78 96 Acama spp. -6.3 Jul 1987 -91 06

- ::> - . - 4 -.

RESULTS to 1992. Intervention analysis showed several statuti- I . . c d y s i w c a n t step changes in each of the data , - Changes in abundance series, the largest of wtuch occurred during 1987-88

for chlorophyll a and all 3 zooplankton species Chlorophyll a and the abundance of the 3 copepod (Table 3). The decline in chlorophyll occurred in

species all dedined substantially and iigmficantly at mid-1987, and the dedines in copepod abundance bgher salimties in their ranges (Fig. 2. Table 2). followed in order Acartia spp.. Euryremora a f k , Except for Acartia spp.. which is rare at low s-ties, and ~inocalan& doerrii. These time periods are ap- si@cant dedines occurred in saliilties down to proximate since the dedines occurred over a, period about 0.17 psu. of several months. Annual mean anomalies (Fig. 4)

The dedines in copepod abundance were the most show considerable vanability, but were lower after substantial changes in the data record from 1972 the dates determined by intervention analysis than

before ixi every year =cept 1989 for A m . a spp. Some of the variability be-

> 0.6 - 1.0- fore the dedines idennfied above can be a a

i esplained by the effects of drought or 2 ..- 2 0.5- ! 0 - .. - - .. flood [Alpine & C!oern 19921. 5 0.0- 0 ;

0 5 0.0- - - - - - ' ..

The effect oi grazmg by dams may 2

- 0 . : . - >

4 42. . W -. U , , i - - ' -

be coniounded by the effects of variability 5 z 4,s- . . .

.-I

.- a in river flow (Alpine & Cloern 1992).

. .. - = 4.4- 2 a . -. - A I f .LO. We plotted anomalies of chlorophyll a q 4 0 . 0 .

' 1 - I I concentration and copepod abundance - . : a 1 - 9 . -. - . t 4 8 .I.s2

I I ! against the log of monthly mean net delta - . . n 7 1 m - m i a ~ 1 ~ 1 ~ n 7 1 m : a m m ~ m m m ~ ..- - 0.6. 1.5- outflow ior s w e v s before and after mid- - -. - P = *

a 0.2. 1.0. d 1987. Chlorophyll anomalies beiore mid-

a I I A . 2 .- r

: ! 1987 were generally bgher at mtermedi- - 3 a o - 3 C 0." ' , -

z - z : I ate flows and lower at high and low flows -. a 4 2 - -. 0

a *. - - U1 u ' - 0 . ' I !

(fig. jal. .After md-1987, chlorophyll was - 2 4.4. 0 0 , : - -

' _ T O J . . ' - : ! i nearly unifordy low in spite of variation

-I ;i a , ' - . - a a#. , a o .I.* ! , A l g in outflow. The copepods had no apparent . -- z z . Z 3

m 41- - ; .a,,. ; relationships with outflow (Fig. jb to dl.. u a : u

.1.0* except for some reduced values at high

, , - n : * m n m n ~ m ~ a ~ ~

' d

' 72 m * a * - m 4 flows, probably due to madequate Sam- . :- - YEAR YEAR . I . - piing of the downstream regon. The 4- - - Fig. 4. San Franciso Bay estuary. .Annual means and 95% confidence h t s U? after md-1987 were generally low*

' I of anomahes m log cnlorophyll concenuaaon (pg l"1 and log abundance ' pareicularfy for Ememora affiais and. . - : ,3 (number m-.' - 10) oi common cooepoa speaes m s d h t y classes 10 to 20 116 &&'a spp. Seasonal paaerns for both E

to 10 for .4cart7a spp.~. la1 ChlorophvU: I ~ I Euryremora dlfuz: {CI Sinocalanus d o e m ldl Acarna sup. affirus and . - \ ca~ ia changed after 1987;

with nearly normal values o c m g dur-

Kunmerer er al.: Clam preaanon on copewas 6;

mg spnng. when flows are bgh. and low values m the iow-flow periods oi summer and fd.

Log rauos of nauph to adults of E u p temora affrnis were signii~cantly iower in

I;? , .. - 1987to!989thanir.1985and1986.Ifthe , r , break IS made aiter 1987, the difference IS .%

still sianificant (Fig. 6: p c 0.000:. nested ANOVA. Tabie 41. Egg ratlos were vari- abie among years. but nor between groups of "ears elther 19& ro '1986 vs

f 1987 to. 1990 or 1984 to 1987 vs 1988 to

I@? 1990 (Fig. 6: p > 0.1. nested ANOVA.

:a; Table 41. The lowest annual mean egg -.- % rafio was m 1986, when naupb and % adults were moderately abundant. and '5 1s: there was no sigmficant relationsbp : berween egg ratios and chlorophyll a (by -. .- inspection and b e a r reg-resslon. r' = 0.04,

P 34m.p>0.11. : Vemca! &stributions of ~ ~ n e r n a r a

' - a{- afhPls nauplii showed relatively little ten- denq of the nauplii to avoid deep water.

LOG DELTA OUTFLOW. cts LOG DELTA OUTFLOW. cts

Fig. 5. San Franciso Bay em- . Means by month of log chlorophvli anomahes and log abundance anomabes of common copepod speaes m s a h y dasses as m Fig. 4, vsjog of net delta outflow rnto esruap [cubic feet per second Idsl. 1 ft' -0.028 mJ]. lo1 1971 to June 1987: 1-1 July 1987 to 1991. (a) Chiorophvll:

(b) Eq-remora a f h ; (c) Sinocaianus doerrii: (dl Acama spp.

The mean depth of the nauplii was 0.8 t 0.5 m {mean with 95 ?O coniiaence hits) above mid-depth. and the

- abundance at 1 m off the bottom averaged 77 = 13 % (mean with 95% confidence limits) of the geometric

r;' tion the nauplii are vulnerable to predation by the

s dams. g s-.

&i. s.. .- . Experimental resuits I:: ..

I L ,-. IF- Euyternora affinis nauplh had a moderately strong 1 =: -. escape response to the shear caused by the incunent f . . z flow of the dam siphons. This response often enabled

1 them to evade the incurrent flow and sometimes to 2.t.

. j:; become entrained in the strong flow from the excur- r; . rent siphon. Nauplh that entered the incurrent siphon -* - .. were not observed to emerge, and some were later

found entangled in pseudofeces, apparently dead. We ' observed quatative differences among copepod _li- ' species in the ablimty to evade the siphon. with E. afdnv

and Acarb'a spp. having a moderate escape response. the harpacticoid Euterpina acutifrons having a strong response, and the neritic calanoid Paracalanus quasi- modo having a weak or noneristent response. Using a pipet as an aruficial clam siphon, we found a sigmfi- cant escape response in the nauplii of E. affinis (135 of 263 nauplu captured by the siphon, f = 78, p < 0.0001. 1 df. 3 experiments pooled).

Clams consumed Eurytemora affinis nauplii and chlorophyll a at sigmficant rates in all 8 tank experi- ments (Fig. 7). Filtration rates averaged 0.005 1 clam-'

0.0 d3 84 85 86 8? 88 8k 90 $1

YEAR

Fig. 6. Eurytemora afikus in the San Frandso Bay estuary. Ratios of nauplii to adults for salinity dasses 10 to 19, between 0.3 and 10 psu and ratios of eggs to females for March to May in sahity dasses 10 to 19. Means and 95 q.6 confidence limits

by year

h- ' on nauplii: and 0.038.1 dam" h'l for chlorophyll. The apparent downward trend in Fig. 3 for E. afiinis may have been related to ambient temperatuie, which increased during this period, although experimental temperature did not show a trend (Eg. 7).

In a single preliminary experiment using beaken, filtration rates on Eurytemora affinis nauplii averaged 0.03 1 dam-' h-'. Ln preliminary experiments using nauplii from higher salinity (ca 30 psu), clearance rates were bghest on Paracalanus quasirnodo, lower on Acartia spp., and not sigmficantly different from 0 on Euterpina acutifrons.

38 .MU. Ccoi. Prog. Ser. 113: 81-93. 1094

Tabie 4. Xnaivm of vanance resuits for ratios of naupb to &h.~b (top) and eggs to females (bottom). Resuits compare pre-dam wth post-dam groups of years. with 1987 bemg asugned to the former (left) or the latter (right]. 3esuits m&cated as elther

s i w c a n t at p c 0.01 I ") or p < 0.001 ("' ) or not nyruficant

Source 1985-1987 vs 1988-1989 1985-1986 vs 1987-1989 Sum of df Mean F Sum of di .Mean squares square squares square

- I Nauplii to adult ratio Year group 1.49 1 1.49 16.09"' Year wvlttun group 1.64 3 0.55 5.91" Error 3.24 35 0.09

Source 1984-1987 vs 1988-1990 Surnof - df Mean F P squares square ?

2.61 1 2.61 28.3"' 0.37 3 0.17 1.32 3.32 35 0.09

1984-1986 vs 1987-1990 Sum of df . Mean - - squares square

Egg to iemale ratio Year group 0.14 1 0.14 1.41 0.27 1 0.27 1.66 Year within group 3.06 5 0.61 6.02 "' 2.90 5 0.58 j.71"' Error 5.69 56 0.10 5.69 56 0.10

DISCUSSION

Tlus may be the first demonstration of a major impact by bivalves on populations of e s t u m e copepods. This result was a fortuitous outcome of an ecoiogical dis- aster. the introduction of a competent and prolific exotic speaes to San Francisco Bay. The finding that bivalves can have such an impact on copepod popula- tions may have strong implications for all shailow estu- anes with abundant bivalves. Thereiore, we examine below the steps leading to the conclusion that preda- tion by the clams caused the dedine in copepods.

The pvldence developed here is circumstantial in that direct measurements of in situ consumption rates were not made. To date, field coasumption;rates of phytoplankton by blvalves have been measured only a few ames, usually under conditions oi low flow veioa- ties. shallow water, and high bivalve densities. such as over mussel beds (Wright et al. 1982. Frechette & BOU-

get 1985a. ~Muschenheun & Newell 1992). To our knowiedge. all stuhes reporting broad-scale effects on phytoplankton have been based on 1aborato'r)r values of grazing rates rather than field measurements of grazlng (e.g. Cloern 1982. Officer et al. 1982. Cohen et al. !984. Loo & Rosenberg 1989. Alpme & Cloern 19921. The pnnapal difficulties seem to be patduness m grazing rates and Logstical problems in w o r b g in deep water wth saong tidal currents. The fonner diffi- culty would be magmfied for zooplankton. since che consumpaon rates and densities oi nauplii are lower than for ?hytoplankton. In addition. knowledge of pop- ulation dynamics of the copepods would be required to place the rnortaiity in context. Suffiaendy detaded measurements oi populanon dynarmcs of estuarine

zooplankton nave been made only once I Landry 1978)

for a population'in a small, enclosed body of water; doing a slmilar study on a population in a dynamic estuary would be extremely difficult. Thus. practical impediments prevent a thorough test of the theory that

---- 1 2 1 4 5 6 7 8 Mean .

1 2 1 4 5 6 7 8 Mean fiwriment Number

Fig. 7. San Francvo Bay esruary. Expenmentally determined grazrng rate Imean t range. I clam-' h-') of the mtroduced dam Potamocorbula amurensu on (a) chlorophyll a d fbl E w e r n o f a &ms nauplii. Numbers m fbl are tempera-

tures l°C1 aunng e.rpenrnents

KUnmerer et a;.: Clam

/ . , I a

predanon by dams exerts sqIuflcanr mondty on b e copepod popuiauons. and we are torced to rely on

i mference.

i . Dedines in abundance

I The declvle m copepod abundance nearly concur- rent with the spread of Potamocorbula amurensis IS the fin: plece or evlaence that the two are reiated. These deches are tne iaraes; and sharpest long-tenr, decreases h abundance we are aware of for esruannr zoopiankton speaes. Tine colnadence of the deches in copepod abundance wth the increase in dam abun- ; dance can be seen by comparing Fig. 4 with Fig. 3 in

Carlton et al. (1990i and Fig. 3 in Alpine & Cloern (1992). whch show the abundance of P. arnurensis to have increased by summer 1967 to values ofrer.

' exceeding 1000 m-' m Susun and San Pablo Bays. On e a l y statistical grounds it could be argued that t h ~ 1 offers oni!; 2 polnu lbeiore and after], providrng no

, power ior tesmg the strength of the interacnon. How- ever. the decimes m these copepod speaes were the

' I largest on record. exceedmg all previous deches , and no other potentiai cause has been identified. Further-

- more, a 5- to 10-fold de&e in abundance of 3 speaes witlun a year would appear to rule out such causes as diseases and parasites, whith would presumably be I ! species-spetific. Other potential predators such 'as small fish have. if anythmg, declined during the same penod (Herbold et al. 19921.

' 1 Based on the intervennon analysis, the dedine in ' chlorophyll occurred nearly simultaneously with the

outbreak of the dam population in summer 1987. The abundance of the 3 copepod species declined in order of the degree of exposure to predation, with Acartia spp, dechmg first foillowed by Eurytemora affinis and. later. Sinocalaous doerrii. Of these 3 copepod taxa,

. i s.' doerrii occupies the most freshwater habitat, with the buik.of the population at a salinity range below 1 psu or saiiniry class 12, generally upstream of river

, km 90 (Figs. 2 & 3). Thus the population of S. d o e m / was upstream of the bulk of the Pofamocorbu4a amu- ;-. rensis population for most of the period, resulting in

reduced exposure. Considering the degree of expo- sure of these populations, the timing of dedines of copepods coinades with the spread of the P. amurensis - * .. I population.

I The declines in copepod abundance were almost

certainly not related to the drought. Freshwater flow .- since the invasion has been low but within the range of

previously observed values. wlule copepod abundance values have been extraordinarily low. There was no apparent relationship between flow and copepod abundance before or after h e dam invasion (Fig. 5).

preaatlon on copepoas 8 C

Prevlous drouahts have been charanenzed by inva- '

slons of other blvaives. notabiy Mva arenana (Nichois 1985i. that may have reduced phytopiankton biomass m 1977 (Alpine & Cloern 19921. There are 3 possible explanations why no effect, of that grazing was gb- served on zoop~ankron. First. chlorophvll concenna- tion dunng that penod was about an order of magmi- tude higher than in the more recent drought Eig. 3 in Alpme 8: Cloem 19921. Li t b implies an order of magnitude lower grarmg rate. the resulting effect on zooplankton would not be detectable over the limited penod of that drought. Second,' s a t ) ' in the area affected by M. arenana was'too iugh for most of the copepod~opulaaons except Acartia spp. %a. M. aren&a is an iniaunal speaes, in contrast to the epifaunal habitat of Potamocorbula amurenesis. The difference in habitat may imply differences in grazing effect on nauplii.

Causal mechanisms

If the dedine in copepod populations was caused by Potamocorbrtla amurensis, there are 2 plausible mech- anisms: (11 consumption of nauplii land eggs, for .

copepods that release their eggs into the water) by P. amurensis, and (2) reduction in food availability due to grazing by the dams. Evidence for a decline in chlorophyll has already been presented (Alpine & Cloern 1992), and is supported by the data in Figs. 2. 4 & 5, although Eurytemora affiais is also capable of feeding on detritus (Heinie & Remer 1975). We argue here that E. affinis was probably more affeaed by pre- dation than competition, although we cannot make the same argument for the other copepod species.

Reduction in food could lower reproductive rates of the copepods, reduce development rates of nauplii. or increase their mortality rates. Only the mortality rates . of the nauplii would be affected by direct predation on the nauplii. However, all of these would .manifest themselves as reduced abundance of nauplii relative to - adults, as was observed.

. Growth rates of juvenile stages of copepods are generally believed to continue into the adult stages, at least for the female, where they are expressed as egg production (Sekiguchi et al. 1980. Berggreen et al. 1988). Reproductive rates appear to be at least as sen- sitive as g~owth rates of copepodites to reduction in food availability (Berggreen et al. 1988). Data on food limitation of nauplii are scarce. Feeding rates of Calanuspacifiwnauplii and stage I copepodites satu- rate at similar food concentrations (Fernandez 1979), and the concentrations of food needed to saturate feed- ing rates increase with inaeasing copepodite stage (Vidal 1980). This suggests that in C. pacificus at least.

food iimtation should occur in the reproductive rates of the adults before it shows up in the feeding rates of nauplii. If that is m e of copepods in general. then one should seek evidence of food b t a d o n in the egg pro- duction rate of the adults. There is some evidence from the field to suggest that early life stages of copepods are less food -h t ed than later, particularly adult. stages (Runge et al. 1985. Peterson et al. 1991. Walker & Peterson 1991). TO. our knowledge there is no example of a copepod population in which nauplii or early copepodites were food-lirmted w w e adults were satiated.

The egg ratio of Eurytemora affinis was presented as ,i a surrogate for egg production rate. Since temperahe was not Uferent among the years of the study, repro- ductive rates should be closely correlated with egg ratios. No dedine in egg ratios was observed in 1987, and the lowest annual value was seen in 1986. before the spread of Potamocorbula amurensis. Furthermore. egg rauos during t h ~ time were uncorrelated with chloropnyU (we do not know the reasons for the inter- annual vanation, such as the low value in 19861. Thereiore, egg producrion after 1987 was not more food Lunited than before. in spite of the reduction in ph.(roplankton that attended the spread of the dam. Bv the assumptions in the previous paragraph. the growth rates of the nauplii (and tbereiore their mor- tality rates in the absence of predation) would not have changed. 'RI.IS leaves predation as the likely cause oi the reduced abundance of nauplii. although the possi- bility oi. iood lirmtadon could never be completely elimnated.

Acaraa speaes apparently are food limited 91 south San Francisco Bay spring iKimrnerer unpubi. data), so the poss~bhty oi competition for food is bgher with th speces. On the other hand, -4cartia spp. release the^ eggs, whch are negatively buoyant and would be vulnerable to consumption by dams. Thus either pre- daaon or competition could be the dominant Whence on the abundance of .-\carria. . .

T-ie behavioral observationi demonstrated that Para- mocorbula amurensis 1s capable of ingesting copepod nauplii, and the expenrnents in tanks demonstrated sigmiicant consumpaon rates on Eurytemora atfirus. T i u is not too surpnsrng given the saength oi the ieed- ing currents observed. However, ingestion oi copepod naupLii by bivalves has been only inirequenrly demon- strated before iHonted et al. 1988).

The vemcal dismbudon of nauplii in *he water column suggests that they should be vulnerabie to lngesaon by the iield population of dams unless there is a depiedon oi nauuk tn the benrbtc boundary laver. Such a depletion has been observed for phytopiaakton and bacrena over dense mussel beds (Frechette & Bourget 198.5b. Frechette et al. 1989. Muschennem h

Newell 19921. Recent esperiments wth beds oi am- fiaal dams scaled to represent PoramocorblL'd dmu-

rensis (11 000 siphons m-'1 have shown a maumn~ deplehon in concenrration of about 50% for a sub- stance that s filtered with 100°6 efficiency (0 '30rdan et al. in press; see also > l o m ~ t h et al. 1990). Tne effi- ciency of pumping ii.e. effective filtration rate divided bv pumpmg rate) for chlorophyll is probably much Less than 100°6 (see Doering & Oviatt 19861, so that [or

, .nauplii should be less than 13 %, based on the ratio of consumption of nauplii to chlorophyll in our experi- ments. A s s U g a filtrh,tion effiaency of 13 96 , and a mean density of 3500 dams rn-2 (Hynanson 19911, the depletion oi nauplii would amount to about 2 3e in the concentradon boundary laver. Thus. dam grazing would not lead to a measurable depletion in concentra- tion of nauph in the benthic boundary layer. On the other hand. a behamoral avoidance of the bottom by the nauplii. not addressed in th.s study. could result in a reduced affect of consumpaon by clams.

Field clearance rates

Other workers on chis speaes have obtained dear- ance rates on chlorophyll of 0.9 to 4.7 1 dam-' d-' (Cole et al. 1992. Werner 81 Hollibaugh 1993) higher than those we observed. ?heir rates could have been biased upward by their use ox cultured rather than natural phytoplankton iDoenng & OviaR 1986). If our experimental design resulted in iow estimates of the clearance rates for ddorophvfl. Then our dearance rates for nauplii may aiso be underestimates.

Measurement oi grazug rates or bivalves is compli- cated by their responses to a variety of factors indud- ing seston concenuadon, flow. 'and substrate type. (Morton 1983. Frechene & Bourget 1985a. Edunan et al. 1989. Levlnton 1991. Cole et al. 19921. Thus the observed dearance rates are unlikely to match closely the rates that would p e m in the field. -4 better method of estimating g r w g rates is through the use of reorculating Llurnes !Cole et ai. 19921, although scaling these properly can be difficult (Jurnars & Nowell 1984). In any event. it is .unlikely that any experimental setup would suffice to determine field qr-g rates. Since direct measurements of these rates is also ememely difficult qven iield conditions (Sf32

above), we are left with indirect demonstrations of the isfluence of predation by Potamocorbula amurensis.

Ln view of the problems extrapolamg laboratory rates to the field, we must take 'Lhese rates as only order-of-magnitude estimates of what the fieid rates could be. It is still instructive to calculate the predatory impact on the copepod opulaaon to determine whether redation could be responsible for the

hmmerer e: a!.: Clam preaauon on conepoa . 0:

Table 5 . Calculauon of unpac. of Poramocorbuia amruens~s oranno on Eun~emora d h s naupli. Tne upnnopai assumption 1s

mar filtrauon races m b e expenmenrs are the same as mose ine i~ela

I kaue Source

G r m p rate (1 dm-: n ' l l

Mean welght of dams (gi Corresponding shell length 1cm1 Me&= shell length m field I ~ I Faao: to conea rate tor s u e Field gaung rare I1 dam-: d - ' I Flelo apundance oi dams rxr, "I

Mean depth Iml ,

G r m o lmpact rd-ll

Proporuon of tune as nauphus Effecnve rnonahn. rp population I d - ' 1 Davs to reach 12 90 of populanon

- ~

From Fig. 7 ! Mean of all erpenment. HoUibaugh & Werner 11993 I I Hymanson ( 199 1 I Hollibauon & Werner 11993 1 I ~ulap ly gravna rare by corremon factor anti 24 n d - I Hynanson (1991 I . DWR rnomtomg data Nauacal cham. area betwee'n nver km 55 and 75 Multlply graung rate by abundance. &nde by depth and b!. 1000 Hernle & Fiemer 119751

I Muhply graung unpad by relatlve auraaon i From effecave mona+ty. dssumrng that ~t IS m excess over monaty I

of steadyrate popUaaon 1 i

observed decline. Calculations are summari2ed in Tabie 5. We calculated consumption rates in percent of the water column deared per day from the measured filtration rates, the abundance.& dams m the upper estuar).. and the mean dep&. We then convened ths to a daiiy mortality rate of the nauplii and calculated the rate of dedine of the population, assurmng that the mortabty is excess over that e a t i n g in the population before dams arrived, and that no compensatory response of the population occurred d u n g this time period. We estimate that the dams were consuming on average 8.2% d-' of Eurytemora affinis nauplii and 62 % d" of chlorophyll in the northern reach of the estuary (Table 5). The calculated time to declme to 12% of the E. affinis population is 73 d. This is some- what shorter than the observed dedine, which took several months: if nauplii are less thari 100 YO ~ulnera- ble to predation the calculated time scale tvould be longer.

O u i data show that the abundance of Eur-vternora affinis and the other 2 species of copepod aeclmed sharply at about the m e of the spread of Potamocor- bula arnurends. Our estimates of clearance rate are - consistent with the explanation of the decline in copepods as due to predation by dams. There is no evidence of increased food limitation in the E. a f f i ~ ~ s population. Thus we have strong, if circumstantial, evi- dence that predation by dams caused the d e c h e in the copepod population.

Potamocorbula amurensis has become well estab- lished in this estuary since the begmning of the re- cent drought (Carlton et al. 1990. Nichols et al. 1990). Although it would be premature to forecast a perma- nent change in the zooplankton, there is cause for concern: several species of fish that pass through their larval stages in the upper estuary are also in a Serious state of d e h e (Stevens et al. 1985. Moyle et al. 1992).

The differences in predation rates on difierent spe- a e s in the p r e l u n m q experiments are provocative. Differences in predation by fish or other predators or. zooplankton prey are often descrfoed by terms such as 'selemviry' or 'preference'. The mfluence of prey escape mechanisms on predator 'choice' IS often over- looked. It is W e l y that a dam could select one cope- pod nauplius over another.

Cenain copepod speaes are numerically dominant in many estuaries and marine bays at least partly because of the low vuinerabiliry of adults and copepodites to pre- dation (Kimmerer 1 99 1. Ueda 1991). Similarly. differ- ences in vulnerability to predation by bivalves. if borne out by further experimental work, could have implica- tions for the controi of copepod species composition. Bivalves are highly abundant in many shallow estuaries and bays ( ~ i c h o h 1985). Potamocorbula amurensis does not appear particularly unusud in size or strength of siphon currents, so one might expect a substantial graz- ing impact on zoopiankton populations wherever ' lqh abundance of bivalves occurs in shallow water. l b predation should be selemve if there are differences in escape responses of the nauplii. Thus it is likely that benthic gradng could exert considerable control on the abundance and species composition of zooplankton in these locations.

acknowledgements. Fundrng was provided to J.O. by,the California lnteragenq Ecologcal Studies Program and Department of Fish and Game and to WX. and E.G. by the California Depamnent of Water Resources. We thank Doug Batl for helpful discussloas. Zach H ~ a n s o n for data on abundance of clams. Randy Brow for encouragement and tunding, Lee Mecum. Greg Schrmdt and Sally Skelton for help with fieldwork and sample analysis, Ai-Ling Chai for help with data analysis. Cathy O'hordan and Jeff Koseff for helpful drscuss~ons. Barbara Sullivan for helpful reviews and encouragement, and Doug Ball. BiJl Bemen. Pat Brandes. AM Durbm. Zach H!manson. Peggy Lehman. Bill Peterson. and Don Stevens for helpful comments on earlier dratts.

d p m e . .A. 5. Cloern. J. E. (19921. Trophc mteracnonj and direc: pnysicai effects convol phytoplankton blomass and proaucuon m an estuary. Lmnol. Oceanogr. 37: 946-955

-bier. J. LL'.. Cloern. J. E.. Hutdunson. A. (1985). Seasonal q d e s or z o o p l ~ t o n from San Franasco Bay. Hydro- bioloqa !29: 177- 197

Arthur. J. . A . Ball. 41. D. 11979). Factors Lniluencing the enrapmerit ot suspended matenal in the San Frannsco bay-delta estuary. In: Conomos. T. J. (ed.1 San Franasco Bay: the urbanued estuary. .Amencan AssoclaQon for the Advancement of Saence. San Frandsco, p. 143-174

Berggreen. Li.. Hansen. B.. Kiarboe. T. (19881. Food sqe specua, ingesnon and growth of the copeood dcania tonsa during development: irnplicanons for determination + of cooepod production. Mar. Biol. 99: 341-352

Bingham. B. L.. Walten, L. J. (1989). Solitary ascidians as predators oi ~nvenebrare larvae - ev-idence from gut analyses and plank~on .samples. J. e.xp. mar. Biol. Ecol. 131; 147-139

Buss. L. W.. Jackson, J. B. C. (1981). Planktonic food avail- abi!ity and suspension-feeder abundance: evidence of in sltu depieuon. J. exp. mar. Biol. Ecol. 49: I5 1-161

Carlton, J. T.. Thompson.. J. K.. Schemel. L. E.. Nichols. F. H. l1990). Remarkable invasion of San Francisco Bay ICalrfor- ma. USA) by the Asian dam Potamocorbula amurensu. 1. Invoducaon and dispersal. Mar. EcoL Prog. Ser. 66: 81-94

Cloern. J. E. r 19821. Does the benthos conuol phytoplankton biomass in south San Franaco Bay7 Mar. Ecol. Prog. Ser. 9: 191-202

Cloem. J. E., Xlpme. X.. Cole. 5.. Wong. R., Arthur. J.. Ball. .M. (1983). River discharge controls phytoplankton dyadrmcs in the northern San Francisco Bay astuary. Estuar. coast Shelf Sa . 16: 415-429

Cohen. R. .R. H.. Dresler. P. V.. ? U p s . E. J. P.. Cory. R. L. (19&1. The effect of the Asiatic dam. C o r b i d a flrrmmea. on phytoplankton of the Potomac River. Maryland. Ltmnol. Oceanogr. 29: !i0-180

Cole. B. 5.. Thompson. J. K., Cloem. J. E. (19921. bIeasure- ment of fiitrauon rates by uuaunal blvalves m a rearculat- ing flume. Mar. Sinl- 113: 219-225

Cowden. C.. Young. C. Ll.. Chia. F. S. (1984). Differenaal predaaon on manne invenebrate larvae by two beriUuc predaton. Mar. Ecol. Prog. Ser. 14: 145-149

Doenng, P. H.. Ovlatt. C. .-\. (19861. Application of filtradon rate models to held populations of blvalves: an assessment usmg expenmental mesocosms. Mar. Ecol. Prog. Ser. 31. 265-375

Eclrman. J . E.. Peterson. C. H.. Cahalan, J. A. (19891. Effects oi flow speed. turbulence. and onentaaon on growth of Iuve- rule bav scallops .Ugopecten irradiaar concenmcus ISavl. J. exp. mar. Biol. Ecol. 132: 123-140

Edmondson. W. T. I 19681. .A graphical method for.evaluanag the use of the egg raao tor measuring brnh and death rates. Oecoloqa 1: 1-37

Emnan. S. C.. Jumars. P. A. I 19881. Effecrs of blvalve s~phonal nurents on !he settlement of inen pamdes and larvae. J. mar. Res. 46: 797-a13

Fernandez. F. i1979). Nuuluon studies m the nauplius Larva of Calmus pacrficus (Copepoda: Calano~da~. .Mar. Biol. $3: 131-117

Ferran. F. D.. Dearborn. J. 3. I 19891. X second examuration oi predanon on pelaqc copepods by the bnttle star .&uo- Coma sgasazu. J. Plankton Res. 11: 1315-1320

Frechette. 41.. 3ourget. E. I 1985a1. E n e r p flow between rhe peiaqc dnd henth~c zones: [actors connolling paradare

organic matter avaiiabie to an mtertidal mussel bed. Can. J. Fish. .Aquat. 5c. 42: 1158- 1165

. Frechene. ,M.. Bourget. E. (l985b). F o o d - b t e d growvr.'. of M.yliius eduhs L. m relation to the benthrc boundary laver. Can. J. Fish. Aauar. 3n. 42: 1166-1 130

Fr-ette. .M.. Bucman. C. .A.. Cever. W. R. 119891. The Impor- tance of bound--laver flows in supplymg phytoplankton to the bentbc suspension feeder. iWWus edulir L. Lunnol. Oceanogr. 34: 19-36

Herrrle. D. R. Flemer. D. .A. (1975). Carbon requvements of a population of the estuarine copepod Eurytemora a i f ~ . Mar. BioL 31: 235-247

Herbold. B.. 3assbg. A. D.. Moyle. P. M. (1992). Status and Bends repon on aquatic resources m the San Francisco Esruary. San Francisco Estuary -Project. U.S. Enwon- '' mental Protecaon Agency. San Frannsco

Homed. S. J.. Nielsen. T. G.. Riemann. B.. Pock-Steen. J.. Bjonrson. P. K. (1988). Regulation of zooplankton by suspension-feeding b~vaives and fish in e s r u m e enclo- sures. &Mar. Ecol. Prog. Ser. 48: 217-224

Hunt. J. H.. . a b r o s e . ' ~ . G. Jr, Petenon. C. H. (1987). Effecrs of the gasnopod. Ilvanassa obsoleta (Say). and the bivalve. 'Mercenaria mercenana (L.), on lanral settlement and juve- nrle recnriment of iniauna. J. erp. mar. Biol. Ecol. 108: 729-240

Hymanson. Z. P. (1991). Results of a spatially intensive s w e y of Poramocorbula amwends in the upper San Francisco Bay estuary. Califorma fnteraqency Ecolopcal Studies Rogram Techolcal Report 30. Sacramento

Jumars. P. A. Nowell. A. R M. (1984). Fluid and sediment dynamic eifeas on marine benthic c o m m u t y structure. .A.m. Zool. 24: 45-55

Kimmerer. W. J. (1987). The theorp of secondary 'production calculations for conunuouslv reproduang populations. LimnoL Oceanogr. 32: 1-13

Kimmerer. W. J. [ 1991). Predatory rnfluences on copepod dis- tributions in coastal waters. in: Gye. 5 - 1 . . Nisbda. S.. Ho. J.-S. (eds.) Proceedings of the iourth international conierence on Copepoda. Bull. Plankton Soc. Japan. Spec. Vol., p. 161-174

Landry. :M. R. (1918). Population dynamics and produdon of a planktonic marine copepod. dcarda dau i i . in a small temperate lagoon on S a . Juan Islana. Washgton. h t . Revue ges. Hydro biol. 63: 37- 119

Levinton. J. S. (1991 ). Vanable feeding behanor in 3 species of .Macoma (Bivalvla. Teiljnacea~ as a response to water flow and sedment manspon. .Mar. Biol. 110: 375-383

Loo. L 0.; Rosenberg. R. 11989). Bivalve suspenuon-feeding dynamics and benthic pelaqc coupling m an eutroph- cated marine bay. J. esp. mar. Biol. Ecol. 130: 153-276

Mactsaac. H. 1.. Spnrles. W. G.. Leach. J. H. 11991). Ingestion of small-bodied zooplankton by zebra mussels (Dreirrerra ,aolymorphar - can cannibalism on !arvae rnfluence pop- ulanon dynarmcs? Can. J. Fih. Aquat. Sci. 48: 2051-2060

Mdaren. I. A. (19651. Relaaonsbp between temperature and egg sue. bodv sue. development rate. and fecundity of the copepod Pseudocalauus h o L Oceanogr. 10: 528-530

~M0msnut.h. S. G.. Koseif. J. 8. Thompson. J. K. O'Riord- C. .A.. Nepi. H. M. (1990). A srudy of model bivalve nphonai currents. Limnol. Oceanoqr. 35: 680-696

Mortou. B. (1983I. Feebag and cbqesaon in Bivalvib Sanuidin. X.. Wilbure. Z 1eds.1 The Mollusca VoL 51

Phynology. Part 2. Academic Press. Yew York. p. 65-14? Moyle. P. 5.. Herbold. 5.. Stevens. 3. t. .LliLler. L W. (1992).

Life history and status oi delta smelt In the Sacrament* San-Joaquin Estuary. Califorma. Trans. .a. Fib. Sot 121: 67-77

Kuxmerer e: ai.: Clam preaauor. on co~epoa i 0:

- . . MwcnennellI;. C. K.. Newell. C. i;. 11991.. Udluanon or.sestor.

flux over a mussel bet. Mar. Ecoi. Proa. Ser. 85: :3i-136 Nagara). M. 119921. Combmec eiiects of temperature anc

sari on me aeveiopmen: o: me copepod Eury7emors affuw. Aquacuinrre 105: 65-7:

~lchols. F. H. 119851. lnaeased benmc pazing: an alternanve expiananon for low ph~opiankton blomass m northern San Franasco Bay dunno the ?976-!977 arouqnr. Esruar. coast. SheU Su. 21: 376-386

~ichols . F H.. Cloern. J . E.. Luomk. 5 . . Peterson. D. X. 119861. The modrficauon of an estuary. Sclence 231: 567-573

Nichols. F. H.. Thompson. J. K.. Scheme:. i. E. (19901. Remark- abie rnvaslon of San Franascc. bay lC&orma. USA1 by the Asian dam Potamocorbuii amurenus. 2. Displace- men[ of a iormer commumn ,Mar Ecoi. Prog. Ser. 66: 95-10;

O'koraan. C. A.. Morusmi&.. S. G.. tiosef!. J. R. fin press). The effea oi b~valve excurrent )e! d~n-narmcs on mass transfer m a b e n b c boundary lave:. Lmnoi. Oceanogr.

Officer. C. B.. Smavda. T. J.. Manx. :. R . (19821. B e n h c filter feeding: a natural euuoptucanon control. Mar. Ecol. Proc;. Ser. t: 203-310

Orsi. J . J.. Bo-an. T. E.. MarreL. 2. 'c.. Hutchson . A. (19831. Recent lnrroducnon o! rbe planktomc calanoid copepod Sinocalanus doemi fCentropa@dael f ~ o m m u - land Chmc ro the Sacramento-Sar Joaqurr. Estuap of Cakfoma. J. Plankton Res. 5: 257-575

Orsi. J . J . . mec cum. W. L. 119861. Zoopidnicton dismbunon and abundance m the Sacramento-Sari joaqujn Delta rn rela- non to c e n u enwonmental factors. Estuaries S: 326-339

Peterson. D. I-!.... Conomos. T. J.. Broenirou.. W. J.. Doherty. P. C. (19751. Locaaon of the non-naai current null zone in northern San Franasco Bay. War. coast. mar. Sa. 3: 1-11

Peterson. W. T.. Tiselius. P.. Kierboe. T. 11991). Copepod egg production. moulmg and growth rates, and second-

, producnon. m the Skagerrak m Augur 1988.3. Plankton Res. 13: 131-154

Rodde. B.. Leakey. R.. Berry. A. 119841. SWty-tempera.ture tolerance and osmoreNanon m Eurytemora affinrs 1 Poppel (Copepoda: Calanoidaj in relation to its distribu- t~on In the zooplankton of the upper reaches of the Forth Estuary. J. exp. m u . Biol. Ecol. 79: 191-31 1

Runge. J. A. (1985). Relauonshp of egg producticn of Calanus pacficus to seasonal changeiVm phytoplankton avail- an&? m Puget Sound. Washington. h o l . Oceanogr. 30: 362-396

Sebens. K. P.. Koehl. M. A. R. (19841. Predation on zoo- plankton by the benthc anthozoans Acyoniurr: siden'um IAlcyonaceai and Memdium senile [Acdniarial in the Kew England subtidal. Mar. Biol. 81: 255-271

S e i u q u a . E.. McLaren. :. .A,. -Coriten. C. i 119801. ileiauon- siun oeween arowitl rate ana ego proauction m the cope- poc .-\carua ciausi huasomca. Mar. 510;. 56. :3;-136

Slnuaraian. K. \'. 119691. Escape reacaons oi zooplankton: tne avolaance oi a pursumg slpnon tuoe, exp. mar. Bloi. Eco~. 2. l?i-178

Shumwav. R. H. 119881. Appired s t a m c a l nme senes anaivsls. Prenuce Hall. Enolewood Cbfis. ii-!

Stevens. D. E.. Kohlhorsi. E. h'.. miller. L. h'.. KeLley. D. W. 11 985 I . The d e c h e of smped bass m tne Sacramento-Sari Joaaurn Estuary. Caliiorrua. Trans. .-.. Fish. Soc. ! 1 4 . 19-3C

SUvar.. B. K. . Banzor,. P. V. 119901. fooc lrrmtatron and benrhc regulauon oi populanons of me copepoc Acarua butisomccr P~nhey I= , nument-hmtec and nurnent-

. ennchec svstems. h o : . Oceanogr. 25. 1616-1631 Sulhvc. B. fi. Doering. P. H.. Ovlan. C. X.. Keller. A. A..

Fnthsez. J iB. (1901i. Lnteracnons wi& the bentnos alter p e i a d food web suucrwe m coastai waters. Can. J. Fuh. Amat. So. 48: 2216-2284

Ueaa. E. 119911. Honzontai cismbunons oi plankromc cope- pods m ralet waters. Ir: Uye. S.4.. !+shdh. S.. Ho. j . 4 . leas.; Proceedihgs oi the 10- mternanonai conier- ence on Copepoda. B d . Plankton Soc. japan. Spec. Vol., D. 143-160

UNZSCO f19811. Tenth repon of the Joint Panel on Oceano- g a p k c Tables and Srandards. Sidney. BC

Vidd. J. i19801. Physioecoiog) of zooplaakton. I . Efiects of pny'toplankton concenaahon, temperature. and body sue on the growth rates oi Calanus padficus and Pseudo- calanus sp. Mar. Biol. 56: 111-134

W a c . D. R.. Peterson. K. T. 11991). Relanonsbps berween hvdroqaphy, phpoplarrkton produclron. b~omass, cell sue and sneaes composition. and copepod producnon m the southern Beawela upwelhg system m April 1988. S. Att. J. mar. Sd. 11: 289-305

Werner. 1.. Hollibaugh J . T. (19931. Potamocorbuia am wensis (Molluscs. Pelecyopodal: comparison of clearance rates and asnmilation effiaenaes tor phyloplankton and bac- tenopiankton. Lmnoi. Oceanogr. 38: 949-964

Wiams. J. G. (19801. The d u e n c e of adults on the settle- ment of spat of the dm. Tapes japoaica. J. mar. Res. 38: 729-741

Wnght. R. T.. Coffin. R. 9.. Ersing, C. P.. Pearson. D. (1963). Field and laboratory measurements of bivalve filtration of natural bactenop1ank:on. Lmnol. Oceanogr. 27: 91-96

Young. C. M. (1989). Larr-z! depletion by asa&ans has little effeet on settlement of epifauna. Mar. Biol. 102: 481-489

Young. C. M.. Got&. S. (1988). Larval predation by bar- nacles: eflects on parch colonization in a shallow subtidal comrnmty. Ecolog?. 65: 624-634

1 Ti m'cle was presented by M. R. Landn (Senior Edrtorial Manwmp: first rece1rpec: .%ovember 30. !993 .. Adrisorl. Honolulu. Hawaii. USA Revrred version accepred: :May 16. 1994 p

1 . . .

February 20, 1995

recent Delta invader, eritiger bifasciatus)

na 0. Swensonl

omia-Nevada Fisheries Society

i

ve Biology, University of mia, Berkeley, CA 94720 and Fisheries Biology, U ty of California, Davis, CA 95616

u The tidewater goby (Eucyclogobius newberry

estuaries (Moyle 1976, Swift et al. the females are quite aggressive and active provide parental care for the eggs (Swift e populations currently exist, which promp endangered species in 1994 (Brewer et al. 19 development of coastal wetlands and alte and exotic species such as centrarchids

The shimofuri goby (~n'dent i~e; b California from Asia via ship ballast water tidewater goby and it closely resemble another Asian goby found in the San

. separate species (Akihito & Sakamot goby (Tomiyama 1936). The two sp shimofuri goby lives in nearly fresh conditions ( k h i t o & Sakamoto 1989.

The shimofuri goby was first recorded 5 and by 1989 was the most abundant fish sampled in Suisun Marsh (U ia at Davis trawls). In 1990 shimofuri gobies traveled south via the CaI yramid Reservoir. By 1992 shimofuri gobies were downstream of Pyl-a Creek, a tributary of the Santa Clara River, which harbors tidewater gobi nof'uli gobies lnay also gain

W-606