O ecolog ia (B eri.) 31, 1 7 7 - 1 9 0 (1977) OecologiaiQ by S p ringer-V erlag 1977

A M odel of Renewable Resources and Limitation of Deposit-Feeding Benthic Populations

J.S. L ev in to n 1,2 and G .R. L o p ez2*1 D e p a rtm e n t o f E cology a n d E vo lu tio n , S ta te U n iv ers ity o f N ew Y ork ,S to n y B rook, N ew Y o rk , N Y 11794, U SA2 In s titu te o f E co logy a n d G enetics , U n iv e rs ity o f À rhus, Â rhus, D en m a rk

Sum m ary. 1. T he renewal rate of resources exploited by a population influences carrying capacity and com petitive interaction. A m odel of resource renewal is proposed where P is the fraction of the resource exploited a t any given time, p the fraction exploited per day, and a the fraction rem aining per day. A t equilibrium , P = p / ( l—a), and resource is always available if p e l — a. A logistic m odel of recovery is also proposed for living resources tha t are themselves lim ited by nutrients o r space.

2. T he models are used to predict carrying capacities of populations o f the m ud snail Hydrobia. The snail does no t reingest its ow n fecal pellets until they have broken down. Pellet b reakdow n rate may therefore be a lim iting factor to popu lation size. M easured pelletization rates and fecal pellet b reakdow n rates predict densities w ithin the range of natu ral Hydrobia populations. Pellet breakdow n in this case is the renew able resource. As m any natu ral sedim ents with deposit-feeders are com pletely pelletized we conclude tha t pellet breakdow n rate is an im portan t lim iting factor to deposit-feeder density and tha t coprophagy m ay be avoided in deposit-feeding mollusks.

1. Introduction

M any populations exploit resources which are renewable a t periodic intervals or are re turned a t a relatively continuous rate. The rate of renewal in p a rt m ay determ ine the pa tte rn of exploitation of the resource, and the favoring of intraspecific te rrito ria l defense m echanism s. If the resource m ust be m onopolized for the lifetime of the individual, then interference adap ta tions m ay be selected to discourage the usurping of the resource by a neighbor. T hus epifaunal benthic anim als have evolved a variety of defense m echanism s to avoid overgrow th o f their needed space for a ttachm en t to the substratum by conspecifics (e.g. Connell, 1961 ; Francis, 1973). In som e reported cases aggression am ong individuals alters the

* P resen t address: In s titu te o f B io logy, U n iv ers ity o f O dense , O dense , D e n m a rk

178 J.S. L ev in to n a n d G .R . L opez

outcom e of interspecific com petition (Lang, 1973 ; Kinzie, 1968). But a case can be m ade th a t the evolution of interspecific aggression as a result of interspecific com petition for resources should be rare (M acA rthur, 1972).

W here resources are renewed frequently relative to the exploiter’s lifespan selection for intraspecific interference or territo ria lity is lessened as the probability of a resource appearing w ithin one’s own foraging area is approxim ately as g reat as w ithin a neighbor’s foraging zone (Levinton, 1972a). In this case we w ould predict selection for efficiency at exploiting the renew able resource, ra ther than for discouragem ent of others from attem pts a t u tilization of the resource.

M obile deposit-feeding (sediment-ingesting) benthic anim als are exploiters of renew able resources. Sedim ent is ingested and the a ttendan t m icroorganism s are assim ilated with varying degrees of efficiency. Typically, the organic debris upon which m any of the m icrobial organism s live rem ains undigested, although fragm entation may occur (H argrave, 1970; Fenchel, 1970; Lopez et al., 1977). After a period o f time, the egested m aterial is recolonized by m icroorganism s, depending upon nutrien ts derived from interstitial w ater and organic debris. In the case of bacteria, recolonization m ay reach preingestion levels in only a few days (H argrave, 1976). T hus, after a period, the sedim ent is once again a food source for deposit- feeders.

W hen a resource is slowly renewed relative to the lifespan of the individual, carrying capacity is merely the to ta l am ount of resource divided by the am ount required per individual. This calculation is som ew hat com plicated by density- dependent resource exploitation and changes in resource requirem ent per in­dividual w ith age. N eglecting these com plications, the carrying capacity o f a barnacle popu la tion can be ascertained w ith a know ledge of the available space and the basal a ttachm en t area of an average individual. W here resources are renewed at a rapid rate relative to the individual’s lifespan, carrying capacity of the exploiter popu lation is differently determ ined for tw o reasons. F irst, ra ther than dividing the sta tic resource am ong the population , exp lo ita tion m ust be visualized as a balance o f explo ita tion and renewal. The m ore rapidly the resource is renewed the greater the exploiter population th a t can be supported . Second, the exploiter popu la tion can enhance the productivity of a renew able resource, and hence the ra te of renewal. A t low densities epibenthic algal and bacte ria l p roduction is stim ulated by anim al grazing, relative to ungrazed sedim ents (H argrave, 1970; Fenchel and K ofoed, 1976). G razing by the am phipod Orchestia grillus on S partina litter increases the colonization rate o f the m icrobial com m unity, relative to ungrazed contro ls (Lopez et al., 1977). U sing the grazing fish N otropis spilopterus. C ooper (1973) show ed tha t low level grazing enhanced the productivity o f au to troph ic m icroorganism s. R eduction o f standing crop and increased turnover o f the exploited populations was a possible explanation . Selective exploitation, relative digestibility o r ability to recover can shift the species com position of the exploited resource populations (e.g. Porter, 1973).

G iven these two factors, there m ust be a popu la tion size above which there is no available resource, and below which the regeneration rate of the resource, a t least on the sh o rt term , always allows som e resource to be available. W hen the resource is food, th is availability m ay be insufficient to m ain tain the needs o f the individual as the species com position o f the resource m ay have shifted to a less valuable p a rt of

A M o d el of R enew ab le R esources 179

the spectrum . But if the renewed resource is sedim ent available for ingestion, or shelters tem porarily used by foragers, then there is always some resource available neglecting the problem of finding the resource when it is rare.

In this study we construct a m odel show ing the lim itation of a deposit-feeding population by the renew able resource o f ingestable sedim entary particles. The model perm its a calculation of the population size th a t can be m aintained by the renewable resource and a pred iction of steady state resource levels at different exploitation rates. We em ploy laboratory microcosm s to investigate the balance of sedim ent ingestion and fecal pellet breakdow n when sedim ent is consum ed by the deposit-feeder Hydrobia. Hydrobia is show n to avoid ingestion of their own fecal pellets until they have been broken dow n to their constituent particles. The breakdow n rate of pellets may therefore lim it feeding. W e use Hydrobia only as a model system and em phasize the laboratory evidence th a t diatom recovery can lim it Hydrobia grow th (Fenchel and K ofoed, 1976). However, sedim ents a t some H ydrobiid localities and in m any shallow subtidal silt-clay sedim ents are pelletized (this study, R hoads, 1973) suggesting th a t the m echanism of pellet breakdow n may be im portan t in regulating deposit-feeder population size.

2. M ateria ls and M ethods

W e have w o rk ed u p o n la b o ra to ry p o p u la tio n s o f H ydrob ia m inuta co llec ted from a h igh in te rtid a l pool a t F lax P o n d , O ld F ie ld , N ew Y o rk , U S A , a n d H . ventrosa, co llected fro m a p o n d n ea r K alo S lo tsru in (H y lleberg a n d L assen , unpu b lish ed ), a n d from a la g o o n a t L e n d ru p (F enchel a n d K o fo e d ,'1976), b o th la tte r localities in Ju tla n d , D e n m a rk . E x p e rim en ts w ith H . m inuta w ere d o n e a t 21° C, w hereas those w ith H . ventrosa w ere k e p t a t ro o m te m p e ra tu re (ca. 21° C).

S ed im en t Finer th a n 62 p w as sieved fro m sed im en t co llected in th e sn a ils ’ n a tive h a b ita t . E qua l frac tio n s w ere p laced in 1 0 m l b ea k e rs o r sc in tilla tio n vials (w ith 1 0 m l o f seaw ater), w ith d ifferent n u m b e rs o f snails. A fte r specified p e rio d s o f tim e pelle ts w ere co llec ted on an 80 p sieve, tra n sfe rre d to a p rew eighed g lass fib re filter, a n d w eighed on an an a ly tica l ba lance . A neg lig ib le n u m b e r o f pe lle ts passed th ro u g h th e sieve. T h e p er c e n t p e lle tiza tio n w as th e n ca lcu la ted as th e w eight o f pe lle ts d iv id ed by th e w eigh t o f th e sed im en t th a t w as in tro d u c e d in to th e co n ta in er. T o e s tim a te pelle t b reak d o w n , we co llec ted fresh ly egested H . m inu ta p e lle ts a n d d ire c tly c o u n ted th e n u m b e r o f in ta c t pe lle ts rem ain ing after specified tim e in tervals . F o r H . ventrosa, e q u a l a liq u o ts o f pe lle ts less th a n 12 h o ld w ere p laced in 5 m l beakers w ith se aw a te r a n d sam p les w ere sieved a t in tervals on a n 80 p sieve to co llec t su rv iv ing in ta c t pellets.

R ec o lo n iz a tio n o f d ia to m s a n d b ac te r ia w as follow ed on fecal m a te r ia l th ro u g h use o f epi- f luorescence m icroscopy , a fte r th e m e th o d s d escrib ed in L opez an d L ev in to n (1978); co u n ts are expressed in n u m b e rs o f cells p e r m g sed im en t.

3. The D eposit-Feeding System and the M ud Snail Hydrobia

Deposit-feeding m arine benthic anim als are com m on in m arine bo ttom s and dom inate fine-grained sedim ents (Sanders, 1958; Levinton, 1972 b). T hey ingest sedim entary particles and organic debris, and egest fecal m aterial consisting of particles bound in to elongate pellets, in the case of Hydrobia. Pellets o f the am phipod Orchestia grillus are encased in a chitinous m em brane and rem ain intact for m onths unless m echanically disaggregated (Lopez e t ah, 1977). As the ingested m aterial passes through the gut, m icroalgae, bacteria, p ro tozoa and possibly small

Nu

mb

er

of

Pe

lle

t»

180 J.S. L ev in to n a n d G .R . L opez

5 0 0 C

0

5 0 0 0

0

5 0 0 0

8 sna ils

H *-

4 snails

2 snails

2 5 5 0

TIME (h)

100 '

b

Fig . 1. a P elle t a c cu m u la tio n by H ydrob ia m inu ta in c h a m b ers (1 4 m g sed im en t c h a m b e r - 1 ) w ith d iffe ring snail densities, b B reakdow n o f pe lle ts o v e r tim e

m eiofauna are digested w ith varying efficiencies. Egested m aterial is the site of recolonization by the m icrobial com m unity.

T he po ten tia l im portance o f coprophagy am ong deposit-feeders has been em phasized in several previous studies (Newell, 1965; F rankenburg and Smith, 1967; Johannes and Satom i, 1966). If dense populations o f deposit-feeders b ind all of the available sedim entary m aterial in to fecal pellets, it is possible th a t no sedim entary m aterial would be available for ingestion. Therefore, tritu ra tion of pellets would perm it access to this otherw ise unavailable resource. H owever, if fecal pellets were reingested, sedim entary grains m ight be devoid of digestable m icroor­ganisms. W hen sedim ent is fed to H ydrobia ventrosa, digestion of ingested diatom s may;be as high as 70 %. H owever, if fecal m ateria l is disaggregated and im m ediately refe'd to the snails, digestion is negligible (Lopez and Levinton, 1978). T hus there is selective value in avoiding reingestion o f recently egested fecal pellets, as they have n o t yet been fully recolonized by an ap p rop ria te food source.

D espite previous observations (Newell, 1965; F rankenburg and Sm ith, 1967), we have n o t observed H. ventrosa to reingest w hole fecal pellets until they have been broken dow n into com ponent sedim entary particles and organic debris. F igure 1 a shows the accum ulation of fecal pellets in populations of H. minuta. A t the highest snail density, ca. 70% of the sedim ent is pelletized after abou t 50 h and no m ore pellets accum ulate. A t this point, snails ceased feeding and either m oved abou t or

A M o d e l o f R enew ab le R esources 181

w ithdrew into their shells. I f pellets were disaggregated and presented to the snails, feeding resumed im mediately. The sam e experim ent was perform ed on Hydrobia ventrosa w ith sim ilar results. A lthough we have occasionally observed an individual to tear apart a fecal pellet, cessation o f feeding is the rule when m ost of the sedim ent is pelletized. The avoidance of reingestion of in tact fecal pellets seems adaptive in tha t it focuses feeding on sedim entary particles tha t have not recently passed through ano ther individual’s gut. L im itation of deposit-feeders, in the context of renew able resources may therefore be due to (1) recolonization of diatom s, bacteria and o ther m icroorganism s, and (2) the rate of pelletization as balanced by the rate of fecal pellet breakdow n. We concentrate on the latter lim iting factor below.

4. A M odel of Resource Level with Renewable Resources

A simple model of sedim ent pelletization and fecal pellet breakdow n m ust take into account the pellet egestion rate per snail, population density, and pellet breakdow n rate. A m ore elaborate m odel would incorporate density-related feeding rates, and the effect o f density upon fecal pellet breakdow n. Let P be the fraction of resource th a t is exploited; O fS P rg l. In the present context, P is the fraction o f the sedim ent th a t is pelletized at any given time. I f p is the fraction o f the sedim ent th a t is pelletized by a given popu la tion density after one day, and a is the fraction of pellets rem aining after one day, then after three days:

P = p + a -p + a -a -p .

T he first term of the series is the th ird day’s accum ulation, the second term is the 2 day’s accum ulation, m inus the fraction th a t has broken dow n in 1 day, while the last term is the 1 day’s pellet production , m inus 2 days’ disintegration. This generalizes to:

P = p t a \i = 0

w here n is the num ber of days. T his series converges (as a< 1) as:

P = p / ( l - a ) . (1)

T hus if pellet b reakdow n (or any resource return) is a constan t fraction per unit tim e, E quation (1) perm its the calculation of the equilibrium fraction pelletized as a balance betw een pellet p roduction and pellet disintegration. If p > ( l - a), then this equilibrium value is g reater than one and the sedim ent will be com pletely pelletized. Ifp <(1 - a), then there is always some available for ingestion. Therefore, a population density corresponding to carrying capacity can be calculated where p = ( l - a ) . C arrying capacity is b¡e, where e is the egestion ra te per snail, and b — { 1 —a). F igure 2a and b shows the effect of varying a and p, when a popu la tion o f deposit-feeders are added to an unpelletized sediment.

If a popu la tion is added to an environm ent w ith a renew able resource, an

182J.S. L ev in to n a n d G .R . L opez

4 3293 0 0 0 0

100

10000

505000

3 02 0

IME s

F ig . 2. a M o d el o f pellet a c cu m u la tio n w hen p e lle t fo rm a tio n is b a lan ced by p e lle t b reak d o w n . Dashed line ind ica tes w here P = 1.0 a t a ca rry in g ca p ac ity o f 14,329 snails m -2 (a = 0.9). b S tea d y state? p e lle tiz a tio n w here p = 0.1, an d a o b ta in s d iffering values. D ashed line show s P = 1

equilibrium value o f resource will be reached if p < (1 - a). Such a situation can be visualized when swarm s of larvae settle on a m ud bottom , a m igrating popu lation o f crustacea enter a feeding a rea w here shelters are required against pred iction betw een feeding periods, o r when a popu la tion o f fish or birds m igrate to a feeding ground and comm ence feeding on a renew able population. H owever, the tim e to reach equilibrium is dependent upon a, and n o t p. If

q — p/P = l + a + a2,

then

<?„= t a‘>i = 0

and a t equilibrium :

' e“o

K 100

/ . 95.9

5

a. .8

° 50

¿i- f .

3z 5

f ------------- .1

IO 2 0 30

T IM E ( D a y s )

A M odel o f R enew ab le R esources 183

q*= £ a' -1 /(1 - a ) .i = 0

If we wish to calculate the tim e to reach 98% of the equilibrium value:

(<?*-«„)/<?* = 0.02

¿ a‘ - ¿ d = an+1 + an+2 + an+3

= an+1(l + a + a 2)

= a n + 1 (1/(1 —a)).Thus

an+ ‘(1/(1 — a))= 0.02

(1/(1 - a ) )

an+l = 0 .02 .

Solving for n at 98% o f equilibrium : log 0.02

lo g a 1 . (2)

Because the relationship is logarithm ic, as a approaches one, the tim e to reach equilibrium greatly increases. Thus w ith a given population density, slow resource renewal results in a d isproportionately slow tim e to reach equilibrium . W ith a population a t carrying capacity, given a sm all value of a (high renewal), equilibrium will be rapidly atta ined and the effects of resource shortage im m ediately felt by the population. However, a popu la tion theoretically at carrying capacity when a is near unity (slow renewal) will bring the resource level to equilibrium very slowly (after a cessation of feeding as in a w inter slowdown), enhancing the possibility tha t o ther factors will contro l popu lation size. Population size above carrying capacity might decrease by em igration, m ortality , o r lowered reproduction.

The hyperbolic relation betw een P and (1 —a) indicates tha t when a is near unity, sm all changes in the la tte r will cause strongly non-linear changes in the steady state value of P. Thus, when a is near unity its value m ust be accurately determ ined. By contrast, P is linearly related to p, m aking sm all errors in estim ating p less im portant.

In the case of recovery by algae o r bacteria, the situation is quite different because of a presum ed upper lim it to m icrobial grow th. The logistic m odel of population grow th best describes such a situation. In integral form, m icrobial population size N , after tim e i, is:

N = K/{ l + e c~ rt),

where K is the upper lim it to grow th, r is the rate of increase of an unlim ited population, and c is a constan t of in tegration defining the position o f the curve relative to the origin (Pearl, 1930; A ndraw artha and Birch, 1954). This m ay be arranged to

In ( ( K - N ) /K ) = c - r t .

184J.S. L e v in to n a n d G .R . L opez

T he param eters r and c are the slope and intercepts, respectively, of the linear regression of ln((K — N)/K) on t . D ata is fitted to this equation by selecting values of K by tria l and e rro r until the best stra igh t line is obtained (Pearl, 1930).

If the m inim um ration required by an individual is know n, then a particle of sedim ent m ust be reingested a t a frequency no faster than it takes d ia tom recovery to reach the m inim um level. The deposit-feeder is essentially particle-lim ited as particles w ith attached diatom s are ingested (or bacteria). T herefore the m inim um ration is best expressed as num ber o f cells per particle, or per unit surface area as ingested particles vary in size.

5. Results

In o rder to calculate predicted carrying capacity for Hydrobia w ith respect to fecal pellet breakdow n, we m easured pelletization rate and pellet b reakdow n rate in H. minuta (3-5 mm long). Pellet accum ulation was m easured a t th ree snail densities: 8,4 and 2 per beaker. L inear regressions were calculated for pellet accum ulation in first 24 h (Fig. 1 a). If there is an effect on pellet b reakdow n due to snail grazing and m ovem ent, or a density effect on feeding ra te (estim ated by egestion rate), then the ratios of the slope of the high and m edium density treatm ents, and the m edium and low density treatm ents should differ from 2.0. T hey are how ever 2.1 and 1.9, respectively, so no such effects were noted in the first 24 h.

A fter ab o u t 50 h, pellet accum ulation in the high density bow l ceased a t abou t 68% pelletization (Fig. la). This m ay be due to the difficulty of finding particles which at th is po in t were diffusely spread th ro u g h o u t th e sedim ent, adhering to fecal pellets. It is also possible th a t these rem aining particles were close to 62 p in diam eter, and were difficult to ingest (Lopez and Levinton, 1978). In e ither case, this incom plete utilization was incorporated in to o u r calculations of carrying capacity.

Pellet breakdow n is show n in F igure lb . F o r the first four days b reakdow n consists o f m echanical fragm entation. H owever, pellets then began to become indistinct and breakdow n greatly increased. This change m ay be related to disin tegration o f the m ucous binding the pellets. T he change of breakdow n necessitates a sim ulation, adjusting the m odel so tha t a changes from 0.9 to 0.5 after the 5 th day.

In o rder to estim ate carrying capacity we m ake the following assum ptions. If H ydrobia exploits the sedim ent to a d ep th o f 2 m m there are 2000 cc m -2 of sedim ent available for grazing. If the w ater conten t of the sedim ent is 60 % (R hoads and Y oung, 1970), the specific gravity o f the sedim entary grains is 2.7, and a typical sedim ent consists of abou t 10% o f particles sm all enough to be ingested by H ydrobia (as is approxim ately the case in L im nfjord localities —Fenchel and K ofoed, 1976), then there are 216 g m - 2 available for exploitation. W e also assume th a t feeding ceases a t 68 % pelletization. F igure 3 shows the effect o f adding different popu la tion densities, using a pelle tization ra te o f 41 jig h - 1 snail- 1 (Levinton et al., 1977), and the breakdow n curve of F igure lb . These data predict a p opu la tion density o f 27,000 m - 2 . T ab le 1 show s H ydrobia densities a t variouslocalities for com parison.

A sim ilar study was done for H. ventrosa, b u t pelletization was followed for a

A M odel o f R enew ab le R esou rces185

' 6 0 0 0 0

150 i2 7 5 0 0

E

O2 0 0 0 0

100

lOOOO50|

e

5 0 0 0

1000

IO 20 3 0

Fig. 3. M odified pellet b reakdow n m ode l, w here a = 0.9 from d ay 1-5, th e a = 0.5. Dashed line ind ica tes w here P = 1.0

T I ME ( D a y s )

T able 1. T yp ica l n a tu ra l d ensities o f H ydrob ia p o p u la tio n s

Species L oca le S nails m 2 S ource

H. ventrosa H . ulvae H . ventrosa H . ulvae H . m inuta

L im n fjo rd , D e n m a rk L im nfjo rd , D e n m a rk K a lo P o n d , D e n m a rk W h its tab le , E ng lan d B a rn s ta b le H a rb o r,

M assach u se tts , U SA

0-37,0000-44,000

500-30,000400-12,000

75-24,700

T . F enche l (unpub lished) T . F enche l (unpub lished) J. H y lleb erg (unpub lished) N ew ell (1965)S an d ers e t a l. (1962)

longer period to see if a balance betw een pellet b reakdow n and pelletization resulted in a stabilization of the available resource, as predicted by the model. In this case, pelletization ra te was estim ated to be 38.0 ± 1.6 pg h -1 snail-1 (n=10). Figure 4 a show s pelletization over tim e a t tw o densities, corresponding to 8 and80,000 snails m - 2 , given the above assum ptions. A t the low density, pelletization increased to ab o u t 15 % b u t d id no t increase any further. As snails were observed to be actively feeding th roughou t the whole period o f the experim ent, we in terpre t this to be the result o f a balance between pelletization and pellet breakdow n. A t the higher density pelletization increased to ab o u t 70 %, feeding activity ceasing a t this point. This slow dow n represents the exhaustion of available resources. F igure 4b_ shows the m easured pellet breakdow n, show ing a near-linear decay of abou t 8 % per day. We do not know why the breakdow n curve was closer to linear in th is case as opposed to increasing, as in F igure 1 b. In H. ventrosa, pellet breakdow n seemed exclusively due to fragm entation. I t is unlikely th a t the species difference is im portant, bu t in the case o f H. minuta d a ta was collected in sum m er, while da ta was collected in w inter for H . ventrosa. Feeding, digestive activity and m ode of pellet com paction m ay have differed betw een seasons. U sing the assum ptions

186J.S. L ev in ton a n d G .R . L opez

80 i 2 0 Snai ls

o

I -UJ

Snai ls20-j

3 9 12 15 18 2 4 2 76

21a

T I M E ( D a y i )

100

« 50-1

b 2 0 3 0 4 0IOT I M E ( D a y s )

Fig . 4. a A c c u m u la tio n o f pe lle ts in ch am b ers (59 m g se d im e n t c h a m b e r" ‘) w ith tw o a n d 20 snails. V ertica l b a r is s ta n d a rd e r ro r (2 sn a ils: iV = 10, 20 sn a ils : N = 5). b B reak d o w n o f pe lle ts o ver tim e

discussed above and m easured pelletization and breakdow n rates, we arrive a t a carry ing capacity of ca. 13,000 snails m " 2 for H. ventrosa.

G iven the determ ined value of a at 0.92, and the m easured pelletization ra te we can p red ict an expected steady state pelletization for the low er snail density with E quation (1). In this case p = 0.033, giving a value of 0.41 for P. As the m easured value was only abou t 0.15, pellet b reakdow n m ust have been greater than estim ated in bowls w ithout snails. Therefore, over several days, snail activity o r excretions stim ulating m icrobial a ttack of fecal pellet m ucous may increase breakdow n rate,

A M odel o f R enew ab le R esources 187

80

_iUJQ.

4 0co

o .

201

O 5 20IO 15

S E D I M E N T / S N A I L ( m g )

F ig . 5. P e r cen t p e lle tiz a tio n a fte r 40 d ay s w ith d iffe ren t am o u n ts o f sed im en t p e r snail

as opposed to our d a ta for the first 24 h in H. minuta. It is also possible that conditions in the pellet breakdow n experim ent did not favor m icrobial grow th but da ta is lacking. Taking P as 0.15 and p as 0.033, a is calculated to actually be 0.78 in the experim ent. W ith this value of a, a carrying capacity o f ca. 35,800 snails m " 2 is predicted.

The pellet breakdow n m odel predicts th a t the steady state per cent pelletization of the sedim ent will increase w ith decreasing sedim ent per snail up to a m axim um pelletization rate. F igure 5 shows the per cent pelletization of experim ental H. ventrosa populations of differing am ounts of sedim ent per snail, kept for 40 days. Pelletization would be 100 % even in the lowest density population , were there no t a balance between breakdow n and pelletization rates. As can be seen, per cent pelletization increases to a m axim um of about 80% , som ew hat higher than reported above. Pelletization increases below about 5 mg sedim ent snail" l. Given the assum ptions outlined above o f sedim ent availability for a sedim ent o f 10 % silt- clay, this corresponds to a density of ab o u t 30,000 snails m " 2, a num ber close to th a t predicted from the m odel as revised by the anim al-related value of a a t 0.78.

F igure 6 shows the recovery o f d iatom s on particles less th an 10 p in size egested as H ydrobia pellets. The dashed line indicates the standing stock on naturally col­lected sedim ent a t L endrup (M ay 1977). R ecovery under room light to this level occurs in 6 days, and a p lateau is reached after about 10 days. As can be seen the logistic m odel adequately fits the data , w ith a best fit a t K = 1.3 x IO6 cells m g " 1 sedim ent. Independent evidence shows th a t the diatom s are not cropped below a density o f 5 x IO5 cells m g-1 (Lopez and Levinton, 1978). T he recovery tim e to a tta in the m inim um ra tion m ight be calculated by m easuring H ydrobia grow th at different densities and calculating the sedim ent ingested per snail per day above which no grow th occurs. G iven the independently m easured recovery of diatom s,

188J.S. L ev in to n a n d G .R . L opez

EmO

E

O

2 4 6 8 IO

T I M E ( D a y s )

Fig . 6. R eco v ery o f d ia to m s a tta c h e d to partic les in new ly egested H . ventrosa feces. D o tte d line rep resen ts d en s ity o f d ia to m s fo und in th e sed im en t a t L e n d ru p . S ta r t in g d ia to m d en s ity is th a t o f new ly

egested feces

we m ay infer w hether steady sta te field popu la tions are lim ited by th is particu lar renew able resource, and predict steady s ta te resource levels in laboratory m icrocosm experiments. T he approach is com plicated by the in teraction of recovery functions o f different resources, and by effects of anim al grazing on the recovery function (e.g. Fenchel and Kofoed, 1976).

6. DiscussionBecause deposit-feeders exploit a renew able resource, models o f the type proposed above predict the grazing level a t w hich no resource is available and the popu la tion is thus lim ited. A t popu la tion sizes below the carry ing capacity, particles are always available, perhaps for consum ption by an o th e r species. P opulation densities pred icted by the pellet breakdow n m odel are in the general range of Hydrobia popu la tions m easured in the field (Table 1). A t Lendrup, however, the sedim ent available for ingestion is not pelletized. A b undan t species such as the foram iniferan Elphidium williamsoni and the am phipod Corophium volutator can be show n to b reak dow n pellets in the labora to ry and p robab ly do so as well in th e field. H owever, o ther localities have sedim ents th a t are com pletely pelletized. A subtidal locality in the Lim nfjord (station 3 o f Jorgensen, 1977) was com pletely pelletized by Abra and Hydrobia. A t a protected flat n ear B egtrup (M ols, Ju tland), the sedim ent was dom inated by populations o f H ydrobia ventrosa and H. neglecta, a t densities of 25-30 x IO3 snails m - 2. Here, the ingestable sedim ent was 50-100 % pelletized in 5 cores exam ined (a visual estim ate as the preponderance of larger sand grains m ade sieving im practical). A t Sebbersund (locality 10 o f Fenchel, 1975) the silt-clay fraction of the sedim ent was com pletely pelletized. H. ulvae density was52,000 m - 2 . O bservations o f subtidal sedim ents dom inated by deposit-feeders often show com plete or nearly com plete pelletization (R hoads, 1967, 1973).

O ur observations are not in conform ance w ith m any reports in th e lite ra tu re of coprophagy in deposit-feeding benthic organism s. A lthough we have observed H ydrobia to dism em ber and ingest its ow n pellets, the general p a tte rn is fecal pellet

A M o d el o f R enew ab le R esou rces 189

avoidance. As freshly egested pellets are im m ediately ingested when dismem bered by the investigator, we presum e th a t the m echanism is merely cued to the size of the pellet (ca. 100 by 300 pi). Indeed, the p roduction of com pact pellets partitions that m aterial recently eaten from particles richer in microorganism s. Before the pellet breaks dow n and is available for reingestion, recolonization by the m icrobial com m unity will have occurred. T hus there is adaptive sense to avoidance o f coprophagy.

Recent experim ents suggest th a t digestion efficiency w ith regard to d iatom s on particles is influenced by the standing stock and digestibility o f the diatom s. This com plicates resource renew al models. W hen sedim ent o f increased diatom standing stock was fed to Hydrobia ventrosa, greater digestion efficiencies were m easured (Lopez and Levinton, 1978). This can probably be related to a minim um num ber of d iatom s m g- 1 sedim ent tha t could not be digested. W hen newly egested sedim ent was refed to Hydrobia, no digestion was m easured. Thus a period of recolonization is essential to renew the resource as fecal m aterial in this case has no nutritive value to the snails. W e m ight, however, imagine cases sim ilar to rabbits (M adsen, 1939) where bacteria-rich and nutritive poor feces are egested, the former being reserved for reingestion. If this can occur am ong deposit-feeding m arine invertebrates then coprophagy should be favored. O ur present evidence cannot exclude this possibility for Hydrobia.

T he rarity o f coprophagy in H ydrobia does not exclude the possibility th a t other cooccurring deposit-feeding species can obtain a nutritive rew ard from freshly egested Hydrobia pellets. F rac tions of the m icrobial com m unity not available to Hydrobia might be assim ilated by o ther species (see Lopez and Levinton, 1978).

In the case o f deposit-feeders where all renewable food resources depend upon particle ingestion, the one w ith the lowest recovery rate will lim it population size. If the m inim um diatom ra tion required by a Hydrobia renews less rapidly than pellet breakdow n, we m ust conclude th a t the la tte r will not be a lim iting renewable resource. But d iatom s m ay n o t be limiting if fecal pellets b reak dow n slowly. T hough the models p roposed above neglect several factors (e.g. density-dependent egestion), we feei th a t such a com parative approach to renew able resources is justified and perm its an analysis of population contro l by resources in deposit- feeding benthic populations.

A cknow ledgem ents. S. K irk p a tr ic k ass is ted w ith s tud ie s o f i f . m inuta. P . P e tra itu s p o in ted o u t to us the ca lc u la tio n fo r tim e to reac h eq u ilib riu m . W e a lso th a n k T . F enche l, J. H y lleberg , H . L assen a n d D. S ch n e id e r fo r help fu l d iscuss ion . T . F e n ch e l read a n d critic ized th e m a n u sc r ip t. T h is is c o n tr ib u tio n n u m b e r 239 to th e P ro g ra m in E co logy a n d E vo lu tio n , S ta te U n i\ e rsity o f N ew Y o rk , S to n y B rook. T h is p a p e r is d e d ic a te d to th e la te P ro fesso r R a lp h G o rd o n Joh n so n .

References

A n d ra w a rth a , H .G ., B irch , L .C .: T h e d is tr ib u tio n a n d ab u n d a n ce o f an im als. C h ic ag o : U n iv ers ity o f C h ic ag o P ress 1954

C o n n e ll, J .H .: T h e in fluence o f in terspecific co m p e titio n a n d o th e r fac to rs on th e d is tr ib u tio n o f the b a rn a c le C htham alus s te lla tus. E co lo g y 42, 710-723 (1961)

C o o p e r , D .C . : E n h a n cem e n t o f n e t p r im a ry p ro d u c tiv ity by h erb iv o re g raz in g in a q u a tic la b o ra to ry m icrocosm s. L im no l. O cean o g r. 18, 3 1 -37 (1973)

F en ch e l, T .: F a c to rs d e te rm in in g th e d is tr ib u tio n p a t te rn s o f m ud sn a ils (H y d ro b iid ae). O eco log ia (Beri.) 20, 1 -17 (1975)

190J.S. L ev in to n a n d G .R . L opez

F enche l, T .: S tu d ie s on th e d ec o m p o s itio n o f o rg a n ic d e tr itu s derived from th e tu rtle g rass Thalassia testudinum . L im n o l. O cean o g r. 15, 1 4 -2 0 (1970)

F enche l, T ., K o foed , L .H .: E v idence for ex p lo ita tiv e in terspecific co m p e titio n in m u d snails (H ydrob iidae). O ikos 27, 367-376 (1976)

F enche l, T ., K ofoed , L .H ., L a p p a la in en , A . : P a rtic le size se lec tion o f tw o d ep o s it feeders : the a m p h ip o d Corophium volu ta tor a n d th e p ro so b ra n c h H ydrob ia ulvae. M ar. Biol. 30, 119-128 (1975)

F ran c is , L .: In tra sp ec ific aggression a n d its effect o n th e d is tr ib u tio n o f A n thop leura elegantissim a. Biol.

Bull. W o o d s H ole 144~ 73 -9 2 (1973)F ra n k e n b u rg , D ., S m ith , K .L .: C o p ro p h a g y in m a rin e an im als. L im nol. O cean o g r. 12, 443 -4 5 0 (1967) H arg rav e , B .T .: T h e u tiliz a tio n o f b en th ic m icro flo ra by H ya le lla azteca . J . A n im . E col. 39, 427-437

(1970)H arg rav e , B .T .: T h e ce n tra l ro le o f in v e rte b ra te faeces in sed im en t d ec o m p o s itio n . In : T h e ro le o f

te rre s tr ia l a n d aq u a tic o rgan ism s in d ec o m p o s itio n p rocesses (J.M . A n d e rso n , A . M acF ad y en , eds.), p p .301-321 . O xfo rd : B lackw ell 1976

J o h a n n e s , R .E ., S atom i, M .: C o m p o s itio n a n d n u tr i tiv e v a lu e o f fecal pe lle ts o f a m a rin e cru stacean .L im nol. O ceanog r. 11, 191-197 (1966)

Jo rg en sen , B .B .: D is tr ib u t io n o f co lo rless su lfu r b a c te r ia (B egg ia toa spp .) in a co a sta l m a rin e sed im en t.

M ar. B iol. 41, 19-28 (1977)K inz ie , R .A .: R ep lacem en t o f Pseudosquilla ciliata by G onodacty lus fa lc a tu s (C ru s tac ea : S to m a to p o d a )

in H aw aii. P acific Sei. 22, 465 -4 7 5 (1968)K ofoed , L .H .: T h e feeding b io logy o f H ydrob ia ventrosa (M on tagu ). I. T h e a ss im ila tio n o f d iffe ren t

c o m p o n e n ts o f th e food. J. Exp. M ar. B iol. E col. 19, 233-241 (1975)L ang , J .: In te rspec ific aggression by sc le rac tin ian co ra ls . 2. W hy the race is n o t o n ly to th e sw ift. Bull.

M ar. Sei. 23, 260-279 (1973)L e v in to n , J .S . : S p a tia l d is tr ib u tio n o f N u cu la p ro x im a ( P ro to b ra n c h ia ) : a n e x p e rim en ta l a p p ro a c h . B iol.

Bull. W o o d s H o le 143, 175-183 (1972a)L e v in to n , J.S .: S tab ility a n d tro p h ic s tru c tu re in dep o sit-feed in g a n d su spension -feed ing com m un ities .

A m . N a t. 106, 472-486 (1972b)L e v in to n , J.S .: T h e eco logy o f deposit-feed ing c o m m u n itie s : Q uisse t H a rb o r M assach u se tts . In :

E c o lo g y o f m a rin e b en th o s (B.C. C ou li, ed.), p p . 19 1 -228 . C o lu m b ia : U n iv e rs ity o f S o u th C a ro l in a

P ress 1977L e v in to n , J.S., L opez , G .R ., L assen , H .H ., R a h n , U .: F e e d b a c k a n d s tru c tu re in d ep o sit-feed ing m a rin e

b en th ic co m m u n ities . In : P ro c . E leven th M a rin e B io logy S ym posium , G alw ay , pp . 409—416 (19^7) L opez , G .R ., L ev in to n , J.S .: T h e av a ilab ility o f m ic ro o rg an ism s a tta c h e d to sed im en t partic les as food

fo r H ydrob ia ventrosa. I n p re p a ra tio n (1978)L opez, G .R ., L ev in to n , J.S., S lo b o d k in , L.B. : T h e effect o f g raz in g by th e d e tritiv o re O rchestia grillus on

Spartina li tte r a n d its a sso c ia ted m ic ro b ia l co m m u n ity . O eco log ia (Beri.) 30, 111-127 (1977) M a c A rth u r , R .H .: G eo g rap h ic a l ecology. N ew Y o rk : H a rp e r a n d R o w 1972 M ad sen , H .: D oes th e ra b b it chew th e cu d ? N a tu re 143, 981 (1939)N ew ell, R . : T h e ro le o f d e tr i tu s in th e n u tr i t io n o f tw o m a r in e dep o s it feeders, th e p ro so b ra n c h H ydrob ia

ulvae a n d th e b ivalve M acom a balthica. P ro c . Z o o l. Soc. L o n d o n 144, 25 -4 5 (1965)P earl, R .: In tro d u c tio n to m edical b io m e try a n d s ta tis tic s . P h ila d e lp h ia : S au n d e rs 1930 P o r te r , K .G .: Selective g raz in g a n d d iffe ren tia l d ig e stio n o f a lg ae b y Z o o p lan k to n . N a tu re 2 4 4 ,1 7 9 -1 8 0

(1973)R h o a d s , D .C .: B iogenic rew ork ing o f in te r tid a l a n d su b tid a l sed im en ts in B a rn s ta b le H a rb o r a n d

B uzzards B ay, M assachusetts . J. G eo l. 75, 4 6 1 -4 7 4 (1967)R h d a d s , D .C .: O rgan ism -sed im en t re la tio n sh ip s o n th e m u d d y sea flo o r. O ceanog r. M ar. B iol. A nn.

'R e v . 12, 263 -3 0 0 (1973)R h o ad s , D .C ., Y oung , D .K .: T h e in fluence o f d ep o s it-feed in g o rgan ism s on sed im en t s tab ility and

co m m u n ity tro p h ic s tru c tu re . J. M ar. R es. 28, 150-178 (1970)S an d ers , H .L .: B en th ic s tud ie s in B uzzards Bay. 1. A n im a l-se d im e n t re la tio n sh ip s. L im n o l. O cean o g r. 3,

245 -2 5 8 (1958)S an d e rs , H .L ., G o u d sm it, E .M ., M ills, E .L ., H a m p s o n , G .E . : A s tu d y o f th e in te r tid a l fau n a o f B arn stab le

H a rb o r , M assachusetts . L im n o l. O ceanog r. 7, 6 3 -7 9 (1962)

R eceived Septem ber 15, 1977

O ecolog ia (B eri.) 31, 1 9 1 - 1 9 9 (1977) OecologiaiC by S p rin g er-V erlag 1977

Effects o f Salinity and Illumination on Photosynthesis and W ater Balance o f Spartina alterniflora Loisel

D.J. Longstreth and B.R. S train

D ep a rtm e n t o f B o tany , D u k e U n ivers ity , D u rh a m . N C 27706, U SA

Sum m ary. P lan ts o f th e sa lt m arsh grass Spartina alterniflora Loisel were collected from N orth C arolina and grow n under contro lled nutrient, tem pera­ture, and photoperiod conditions. P lants were grown a t two different illumi­nation levels ; substrate salinity was varied, and leaf photosynthesis, transp ira­tion, to tal chlorophyll, leaf xylem pressure, and specific leaf weight were m easured. C onditions were con tro lled so th a t gaseous and liquid phase resis­tances to C 0 2 diffusion could be calculated. G row th a t low illum ination and high salinity (30 pp t) resulted in a 50% reduction in photosynthesis. T he reduction in photosynthesis o f plants grow n a t low illum ination was correlated w ith an increase in gaseous resistance. Photosynthetic rates of plants grow n a t high salinity and high illum ination were reduced only slightly com pared to rates o f plants grow n in 10 pp t and H oag land’s solution. Both high salinity and high illum ination were correlated w ith increases in specific leaf weight. C hlorophyll da ta indicate th a t specific leaf weight differences were the result o f increases in leaf thickness. It is therefore hypothesized th a t photosynthetic response can be strongly influenced by salinity-induced changes in leaf structure. Sim ilarities in photosynthetic rate on an area basis a t high illum ination were apparen tly th e result o f increases in leaf thickness a t high salinity. Photosynthetic rates were generally quite high, even a t salinities close to open ocean w ater, and it is concluded th a t salinity rarely limits photosynthesis in S. alterniflora.

Introduction

Spartina alterniflora Loisel flourishes in an environm ent th a t is characterized by a num ber o f extrem e physical factors including high salinity and high irra ­diance. This p lan t species dom inates the intertidal sa lt m arshes o f the E astern and G u lf C oasts o f the U n ited States, often form ing a distinct vegetational zone (A dam s, 1963; H inde, 1954).

H igh salinity has been show n to reduce biom ass p roduction of S. alterniflora seedlings in labo ra to ry experim ents (M ooring e t al., 1971) and to be negatively correlated w ith dry m a tte r p roduction in field studies (Broom e e ta l . , 1975;