Journal ofChemicalEcoiogy. Vol. 12, No. 11, 1986

SEX-IDENTIFYING URINE AND MOLT SIGNALS IN LOBSTER (Homarus americanus)

JELLE ATEMA

and D I A N E

F. C O W A N

Boston University Marine Program Marine Biological Laboratory Woods Hole, Massachusetts 02543 (Received June 27, 1985; accepted December 18, 1985)

Abstract--During courtship, premolt female lobsters, Homarus americanus, choose a male and initiate a pair bond by repeated approaches to his shelter. The male allows such a female to share his shelter for about one week. This knowledge formed the basis to search for quantitative evidence for lobster sex pheromone(s) used in courtship: male cues to allow premolt females to identify a preferred male, and female cues to allow males to identify a premolt mature female. In each of four 1500-liter naturalistic aquaria, the behavioral responses of one female and two male lobsters to male and female lobster urine (0.5 ml) and body odor (20 nal) stimuli were observed. These stimuli were injected once or twice per day into a continuously flowing delivery tube attached to lobster shelters. Habituation to stimulus introduction-a serious problem in earlier experiments--was apparently avoided in the more natural social and physical environment we employed in these experiments. We demonstrated that male and female molt body odors contain different chemical substances: females responded to male molt body odor and males responded to female molt body odor but not vice versa. In general, male and female intermolt urine caused strong responses; however, females responded only weakly to male urine. This suggests that male and female urine are chemically different. Female urine and molt body odor caused a typically male "high-on-legs" response. These results show that molt body odors and intermolt urine contain sex-specific substances, which may be used in lobster courtship as well as other social interactions. Key Words--Lobster, Homarus americanus, sex pheromone, urine odor, body odor, courtship, chemical communication. INTRODUCTION B e h a v i o r a l e v i d e n c e f o r f e m a l e s e x p h e r o m o n e s h a s b e e n o b t a i n e d in v a r i o u s C r u s t a c e a ( R y a n , 1966; D a h l et al., 1970; M c L e e s e , 1970; A t e m a a n d E n g s 2065 0098-0331/86/1100-2065505.00/0

9 1986 Plenum Publishing Corporation

2066

ATEMA AND COWAN

trom, 1971; Teytaud, 1971; Kamiguchi, 1972; Katona, 1973; Kittredge et al., 1971; Eales, 1974; Ameyaw-Akumfi and Hazlett, 1975). In a review of crustacean sex pheromone research, Dunham (1978) argued convincingly that important aspects of experimental design had often been lacking in these studies. However, in many of these studies the possibility of a sex pheromone acting other than as an attractant must be considered, as well as the role of behavioral reinforcement by social interactions, habituation to repeated presentation of unreinforced stimuli, natural biological rhythms, etc. (Atema et al., 1979; Atema and Cobb, 1980; Dunham, 1985). Despite a rather large number of studies by several authors on sex recognition in crayfish, Dunham's recent review (1985) concludes that it remains debatable whether these animals are capable of sex recognition or even species recognition by ehemieal stimuli, since results are inconsistent and inconclusive. Early attempts to identify and isolate a female sex pheromone in Homarus americanus failed (Atema and Gagosian, 1973; Gagosian and Atema, 1973; McLeese et al., 1977; Atema, unpublished) probably due to the lack of understanding of its biological function and context. It was formerly believed that female lobsters--like moths--were sedentary dispersers of pheromones. Through further behavioral studies in the laboratory (Atema et al., 1979; Cowan and Atema, 1985) and in the field (Karnofsky and Atema, 1979) we have learned that females, not males, are the active partners in sexual selection, seeking mature male lobsters in their shelters. This search and selection takes place in the premolt period and involves courtship and pheromones. Therefore, a bioassay requiring a male to locate the source of release is unnatural and, as experience showed, is doomed to failure. Under conditions where single males were presented with a social odor in the absence of other stimuli, responses often habituated or did not allow discrimination between odors (Atema, unpublished). Based on new understanding of the role of the female lobster sex pheromone, we designed a naturalistic odor testing situation in which natural context was preserved and habituation was avoided. Similarly, Gleeson (1980) and Gleeson et al. (1984) designed a successful bioassay for blue crab sex pheromone(s) based on the male's natural response to premolt females. We use the term "pheromone" to describe a body odor produced in the context of communication; it may be a mixture of compounds. "Body odor" refers to the entire chemical output which any organism produces as a result of metabolism; body odor may include urine, urine pheromones, and/or other pheromones. Many body odors may be detected and used in predation or social behavior without there being evidence for communication (Atema et al., 1980; Atema and Stenzler, 1977; Bryant and Atema, 1986). We chose three potential sources of chemical signals that may have social significance: intermolt urine, molt body odor, and intermolt body odor. Urine was shown to be the source of female sex pheromones in crabs (Ryan, 1966; Gleeson, 1980). Crustacean urine, including in lobsters, is stored in bladders,

2067

URINE AND MOLT SIGNALS IN LOBSTERS

an indication of the need for controlled release. Control of urine release was implied in a study of female crab pheromones (Christofferson, 1978). Urine is released from a bilateral pair of papillae (nephropores) at the base of the large (second) antennae into the powerful gill current (McPhie and Atema, 1984; Atema 1985), suggesting that the gill current is used--among other functions-as a broadcast system for chemical signals. Molt body odor is the mixture of metabolites in the water collected from the tank of a molted animal. Mature females molt about one half hour before mating. Molting and mating take place about midway during a cohabitation period of about one week in the male shelter (Atema et al., 1979). Mature female "molt water" contains the female lobster sex pheromone (McLeese, 1970; Atema and Engstrom, 1971) to which males show a typical posture variously described as "mating dance" or "high-on-legs" (Atema and Engstrom, 1971; Atema and Cobb, 1980). Similarly, intermolt body odor is the mixture of metabolites in the water collected from intermolt lobsters. There is no evidence for specific pheromones in intermolt water, but field and laboratory observations (Kamofsky and Atema, 1979; Cowan and Atema, 1985) suggest that lobsters can distinguish between intermolt individuals. One other aspect of odor signals is important to consider: dispersal and dilution. In mature lobsters, the gill current projects the animal's body odor some seven body lengths forward (McPhie and Atema, 1984; Atema, 1985) with or without the addition of urine. After about seven body lengths, the gill current has lost most of its momentum due to viscous drag (vortices and turbulence) and "dissolves" in a large dilute odor cloud. Ambient currents pick up this cloud and disperse and dilute it further. Reasonable estimates of odor dilution are not known in this or most other cases. Rough estimates can be obtained from measurements of dye diluting away from a stationary source, where dilution of 3-4 orders of magnitude were regularly observed at distances of 3-10 cm away from the source. These figures refer to the relevant parameter, odor peaks, not the commonly used but less significant parameter, average odor dilution. For a discussion of aquatic odor dispersal see Atema (1985). With our naturalistic bioassay (see Methods and Materials), we provide evidence for special molt substances in male and female molt body odor and for chemical differences between male and female urine. Our focus was on males responding to female cues, since the best evidence thus far was for female pheromones. This focus is reflected in the experimental design using intermolt males in shelter as the primary group of receivers. METHODS

AND MATERIALS

Four groups of three mature lobsters, two males and one female each, were kept in four 1500-liter aquaria under seasonally adjusted artificial light-dark cycles and ambient flow-through seawater conditions (Figure 1). The three lob-

2068

ATEMA AND COWAN

INTRO ,, CONSTANTFLOW "~~[~_~IR LIFT

~

3M

FIG. 1. Diagram of one of four similar naturalistic observation tanks used in these experiments. Observation shelters were placed in front of each window. A constantly flowing air-lift water circulation system with funnel interruption delivered a slow irregular flow of home tank water into these shelters at a mean flow rate of about 320 ml/min. Test and control stimuli were injected into the funnel. Ambient, unfiltered seawater flowed through each of the four tanks at a rate of about 25 liters/min (not drawn in figure). Substrate consisted of sand, gravel, shells, and rocks. The tank sketched here was one half of a 6-m-long tank. The two halves were separated by a solid barrier; each half had its own inflows and drains. Lobsters are drawn in to scale. The lobster in the left shelter is shown fanning his pleopods (FAN) at one of the shelter entrances; the lobster in the right shelter is shown checking one of the shelter entrances (CE). See Table 1 for description of behaviors.

sters per tank interacted normally, allowing each male lobster to recognize the presence of a male and female lobster. Thus, male and female odor stimuli would not be presented in a social vacuum. This feature was designed to prevent habituation to repeated odor introductions. Normal social interactions, including courtship and mating behavior, occurred in the tanks. Odor tests were done only when a lobster was alone in one o f the observation shelters. These handmolded concrete shelters had two entrances and dimensions comparable to those of lobster-made shelters in the field (personal observation). Such shelters were placed against the two observation windows in each tank to record behavior inside the shelter. Cinderblocks provided less preferred shelter space to the rear of each aquarium (Figure 1). The aquaria were supplied with fresh dead fish or squid daily in addition to the constant availability of a variety of live mussels, worms, crabs, fish, and squid. Uneaten dead food was removed the next day. Ample availability of food should keep the lobsters food-satiated to avoid the possibility that hunger motivated responses to social odors. Food-satiated lobsters responded to social odors but not to food odors (McLeese, 1973). Experiments were begun on May 24, 1984, when the water temperature reached 13 ~ and terminated on November 13, 1984, at the same temperature. Temperature was recorded daily and reached its summer peak of 23 ~ in mid-

URINE AND MOLT SIGNALS

IN LOBSTERS

2069

August. However, no experiments were done at temperatures above 21.5~ (August 2-27). Each observation shelter was provided with a constant-flow stimulus introduction system (Figure 1). Urine (0.5 ml) or body odor (20 ml) were introduced up to twice a day. We allowed at least 4 hr between tests for an individual lobster to minimize possible interaction between responses. The introduction system caused dilution of the stimuli; dye experiments showed urine to be diluted at least 10,000-fold and body odor at least 1000-fold before reaching the receiving animal. Most likely further dilution due to tank circulation and lobster-generated currents would result in negligible signal-to-noise ratios within minutes. Body odor stimuli were prepared by keeping a donor lobster in a 10liter aquarium with aeration and flow-through ambient seawater. After an hour of acclimation, the flow was shut off for 4 hr with the lobster in place. Aliquots were taken to serve as the body odor stimulus within 1 hr after collection. Control stimuli for body odor tests were either 20 ml home tank water (i.e., taken fresh from the tank of the recipient), or 20 ml fresh ambient seawater stored for 4 hr in a 10-liter aquarium with aeration. Home tank water was used as a control for experimental manipulations since, due to habituation, animals should show no novelty response to the water they were inhabiting. Fresh ambient seawater prepared in a manner identical to the test stimulus was the other necessary solvent control. Urine was collected by taking a lobster from a communal holding tank and holding two small vials up to the nephropores. An individual lobster could produce urine from a few drops to a powerful fine jet from both nephropores, occasionally providing as much as 5.0 ml from one nephropore. Often no urine was collected with this method. Urine samples (0.5 ml) were used within 1 hr after collection, usually within 20 min. For urine tests, 0.5-ml samples of home tank water served as controls. Durations of seven behaviors were recorded to the nearest second (Table 1). These behaviors were chosen from a longer list (Atema and Cobb, 1980; Cowan and Atema, unpublished). Nine different stimuli were tested for both males and females; 0.5 ml male or female intermolt urine; 20 ml male or female intermolt body odor; 20 ml male or female molt body odor (collected in a fiveday postmolt period); 0.5 ml and 20 ml home tank water controls; 20 ml fresh seawater control. All experimental lobsters, donors and recipients, were individually marked to follow their molt cycles. All animals were sexually mature (75-90 m m carapace length). Recipient animals never received any one stimulus more than 10 times; molted animals were replaced. The total number of all tests (including controls) done with any individual lobster ranged from 1 to 33, with a median of seven and a mean of eight. A test consisted of two continuous 5-min observation periods: (1) no stimulus and (2) control or test stimulus. A condition for doing a test was that in the no-stimulus observations the lobster was quiet in a shelter and not performing any of the behaviors listed in Table 1. The observer was blind to the type of stimulus being introduced.

2070

ATEMA AND COWAN

TABLE 1. DESCRIPTION OF LOBSTER BEHAVIORS IN STUDYa

Behavior

Description

Comments

Locate source (LS)

Reach up the glass at the front of the shelter and probe the stimulus inflow tube with chelae or clasp the tube with the dactyls of the pereiopods

Check glass (CG)

Turn to face the front of the shelter and place tips of the chelae at the window below the stimulus inflow tube.

Searching for the source of release; often the basis of experimental designs in other studies of crustacean chemical communication; follows from '~ glass"; always accompanied by antennule flicking (an odor sniffing behavior) Always precedes "locate source"; sometimes accompanied by antennule flicking and by scanning the shelter from entrance to entrance.

Check entrance (CE)

Walk over to and stand still at one of the two entrances (see Figure 1)

Generally accompanied by antennule flicking; common response to an approaching animal both in nature and in large aquaria (personal observation)

Pleopod fan (FAN)

Slow to rapid beating of the pleopods, almost always with tail extended (see Figure 1)

Creates (strong) current from front to rear of lobster; when standing with tail in shelter entrance, water is drawn into one entrance and blown out the other

Seizer open (SO)

Seizer claw fully opened (half open is the relaxed position)

Aggression (both offense and defense)

Seizer closed (SC)

Seizer claw fully closed

Seen in a variety of social contexts

High-on-legs (HOL)

Stand with all pereiopods (fully) extended

Described as a male response to female molt water (Atema and Engstrom, 1971); resembles male blue crab behavior in response to female sex pheromone (Gleeson, 1980); often accompanied by FAN

~For more complete behavioral context see Atema et al. (1979), Atema and Cobb (1980).

2071

U R I N E A N D M O L T SIGNALS IN L O B S T E R S

After normalizing the raw scores (x/~x + 0.5), data were analyzed with a three-way ANOVA for unequal replicates, comparing two sexes, seven behaviors, and nine stimuli simultaneously. This analysis was followed by Duncan's multiple-range test (c~ = 0.05) to determine differences in duration of individual poststimulus behaviors by comparison with the appropriate control tests and to establish differences between the six test stimuli, and between male and female responses.

RESULTS

The mean durations of all behavioral responses recorded are presented as raw duration scores in Figure 2. Statistical analysis of the results showed a strong three-way interaction between sex, behavior, and stimuli (ANOVA, P < 0.0005). Responses to tank water control and fresh seawater control were not different (test for least significant differences, P > 0.05). Subsequently all social odor tests were analyzed by comparing the behavioral responses to test stimuli with responses to ambient seawater controls (Duncan's multiple-range test, c~ = 0.05). Based on these comparisons (Table 2), it is evident that some stimuli, such as female urine, elicited the appearance of a number of different behaviors, while other stimuli, such as female molt body odor presented to females, did not. "Locate source" (LS) by itself was rarely a significant response. Since LS is an escalation of "check glass" (CG) (Table 1), we combined the two units into LS/CG for further analysis. The unit "seizer open" (SO) by itself showed no significant differences from control and is not listed in Tables 2-4. Overall, males responded more than females to each of the stimuli except to male molt body odor (Tables 2 and 3). Males responded to five of the six odors tested, whereas females showed no response to half the stimuli: i.e., both male and female intermolt body odor, and female molt body odor (Table 2). Males responded more strongly than females to female molt body odor and male urine (Table 3). The strongest response--as indicated by the largest number of different behaviors being significant--was elicited from males responding to female urine (Table 2). This response included the "high-on-legs" (HOL) behavior that turned out to be a unique male response displayed only to this stimulus and perhaps to female molt body odor (Figure 2); females never showed this behavior to a significant degree (Table 2). When compared with controls, female molt body odor elicited the greatest contrast between male and female responses: males showed a number of different behavioral responses to it, whereas females did not respond to it at all (Table 2). Conversely, male molt body odor caused stronger responses in females than in males (Tables 2 and 3). Male urine

+

168

CE

+

70

LS\CG

180

3

60

H

60

~

4o

E

ao

40

RECIP

M

DONOR

0 0 PRE

FMFMFM TTMFFM URINE

IDD [

FMFMFM

FMFM

RECIP

M

FMFMFM

FMFMFM

FMFM

TTMFFM

MFFM

DONOR

0

TTMFFM

TTMFFM

MFFM

PRE

URINE

INTERMOLT MOLTED BODY ODOR

FRN

INTERMOLT MOLTED BODY ODOR

HOL

7e §

20

18 --

RECIP DONOR

FM 0 0 PRE

i

,

i

M

FMFMFM

FMFMFM

FMFM

8

TTMFFM

TTMFFM

MFFM

PRE

URINE

a

F MF MF M T T MF F M URINE

F MF MF M T T MF F M

F MF M MF F M

RECIP DONOR

INTERMOLT MOLTED

INTERMOLT MOLTED

BODY ODOR

BODY ODOR

+ 120

+

+ SC

~

4e

20 ~0

10

RECIP

r RM

F M F M F M

F M

DONOR

8 0

T T MF F M

T T MF F M

PRE

URINE

SO

70

.

RECIP MF F M

INTERMOLT MOLTEI~ BODY ODOR

DONOR

j ,,ii,iii, i M

FMFMFM

FMFMFM

0

TTMFFM

TTMFFM

MFFM

INTERMOLT

MOLTEB

PRE

URINE

FMFM

BODY ODOR

F1G. 2. Mean durations _+ SEM of six behaviors (abbreviations and descriptions in Table 1) observed in 5 min following stimulus introduction. Histograms represent raw behavior duration scores approximated to the nearest second. There are two groups of recipients (RECIP): males (M) and females (F); three categories of donors (DONOR): males and females and control tank water (T); and eight stimuli: male and female urine and its 0.5-ml home tank water control, male and female intermolt body odor, male and female molt body odor, and 20-ml ambient seawater control for body odors. Also presented is behavior observed immediately preceding stimulus introduction (PRE). Subsequent data analysis showed behaviors marked ( + ) to be significantly different from their controls (see Table 2).

URINE AND MOLT SIGNALS IN LOBSTERS

2073

TABLE 2 . M A L E AND FEMALE RESPONSES TO SIX STIMULI a

B e h a v i o r units Response

Stimulus

M

N

LS/CG

CE

FAN

SC

HOL

* *

* *

* *

* --

* ---

----

*

--

-*

---

Urine Mobo Ibo

F F F

(60) (28) (16)

* -.

Urine Mobo Ibo

M M M

(21) (13) (15)

----

Urine Mobo Ibo

F F F

(25) (10) (8)

. .

Urine Mobo Ibo

M M M

(6) (6) (8)

.

.

.

* ---

*

* * *

* . .

-* .

* . .

. .

* * .

.

. . ---

.

.

.

a T h e asterisks indicate, f o r five different b e h a v i o r units, that the test stimulus c a u s e d a s t r o n g e r response ( D u n c a n ' s m u l t i p l e - r a n g e test c~ = 0 . 0 5 ) than s e a w a t e r control stimuli. F o r b e h a v i o r units see Table 1. N, s a m p l e size; M, male; F, female; M o b o , molt b o d y odor; Ibo, intermolt b o d y odor.

was

a strong

responded body

stimulus

weakly

odors Because

to the same

for males,

to male

molt

did not elicit other males stimuli,

responded

but not for females and

intermolt

significant more

comparisons

body

responses

and generally

between

(Tables odor (Table

more

the six stimuli

2 and

(Table

3). Males

2). Intermolt

2).

strongly

than

can be made

females only

by

TABLE 3. RESPONSE DIFFERENCE BETWEEN M A L E S AND FEMALES a

B e h a v i o r units Stimulus

LS/CG

Urine Mobo Ibo

F F F

. -.

Urine Mobo Ibo

M M M

--.

CE .

.

.

.

.

-. --.

FAN

SC

--

* m

* -.

HOL

-(*)

m

.

" T h e asterisks in this table indicate that males r e s p o n d e d significantly m o r e t h a n females ( D u n c a n ' s test a = 0 . 0 5 ) , except asterisk in p a r e n t h e s e s w h e r e f e m a l e s r e s p o n d e d significantly m o r e than males. See T a b l e 2 legend for abbreviations.

2074

ATEMA

TABLE

4.

STIMULUS

COMPARISON

AND

COWAN

a

Behavior units Stimulus Urine

F

Urine

M

Receiver

LS/CG

M

--

F

Urine

Urine

Urine

F

F

M

Mobo

F

Ibo F

Mobo

M

M

Ibo M

F

Mobo

Mobo

F

Ibo F

Mobo

Ibo F

M

Ibo M

Ibo M

-.

.

HOL

--

--

.

--

*

--

--

--

--

*

*

*

--

M

*

*

*

--

*

F

--

*

--

*

--

M

--

*

--

*

--

*

--

*

--

M

M

* .

SC

F

F

Mobo

FAN

M

F

Urine

.

CE

.

-.

-.

-.

-.

.

.

-.

--

.

.

M

--

--

M

.

.

.

.

.

F

.

.

.

.

.

M

.

.

.

.

F

--

. --

--

M

.

.

.

.

.

F

.

.

.

.

.

aThe asterisks in this table indicate that the stimulus listed on the left caused longer durations of the particular behavior than the stimulus on the right (Duncan's test ~ = 0 . 0 5 ) , except asterisk i n parentheses where this is reversed. For example, M o b o F caused significantly more S C i n male receivers than M o b o M , but female receivers showed significantly more SC to M o b o M than to M o b o F . S e e T a b l e 2 legend for abbreviations.

considering male and female responses separately. Stimulus comparison (Table 4) shows that female urine was a stronger stimulus than male urine for males. In addition, for both sexes, female urine was a stronger stimulus than both female molt and intermolt body odor (Table 4). Male urine caused stronger responses than both male molt and intermolt body odor in males. Males and females responded more strongly to the molt body odor of the opposite sex than to molt body odor of their own sex. Male molt body odor was a stronger stimulus than male intermolt body odor for females; males made no distinction between these male stimuli (Table 4) and showed only weak responses to both (Table 2). While SO did not occur with sufficient consistency to be significantly different from controls, the raw mean scores (Figure 2) show that SO and "seizer closed" (SC), the two independent changes from the relaxed half-open claw

URINE AND MOLT SIGNALS IN LOBSTERS

2075

TABLE 5. CONTINGENCY TABLE FOR TESTING DISTRIBUTION OF SO AND SC IN RESPONSE TO MOLT BODY ODORS (MOBO) IN TWO SITUATIONS: INTERSEXUAL AND INTRASEXUAL RESPONSES

Mobo

Intersexual

Intrasexual

SO SC

7 179

57 33

X2 = 118 P < 0.001

TABLE 6. CONTINGENCY TABLE FOR TESTING DISTRIBUTION OF SO AND SC IN RESPONSE TO INTERMOLT URINE IN TWO SITUATIONS; INTERSEXUAL AND INTRASEXUAL RESPONSE

Urine

Intersexual

Intrasexual

SO SC

11 151

42 158

X2 = 13 P < 0.001

TABLE 7. CONTINGENCY TABLE FOR TESTING DISTRIBUTION OF SO AND SC IN RESPONSE TO INTERMOLT BODY ODOR (IBO) IN T w o SITUATIONS: MALES RESPONDING TO FEMALE IBO (M TO F) AND FEMALES RESPONDING TO MALE IBO

(F TO M) Ibo

M to F

F to M

SO SC

2 42

22 8

X2 = 35 P < 0.001

position (Table 1), are not randomly distributed. In particular, intrasexual responses to molt body odor (i.e., males responding to male molt body odor, and females to female molt body odor) show longer SO than SC, while intersexual responses show the reverse effect. From X2 analysis of contingency tables (Tables 5 and 6), we conclude that for both sexes SO is a behavior associated more than expected with introductions of molt body odor and intermolt urine of animals of the same sex, while SC is associated with molt body odor and urine of the opposite sex (P < 0.001). This result is partially reflected also in Table 3, showing that males give stronger SC responses to female molt body odor, and vice versa. In contrast, intermolt body odors cause sex differences in claw responses: analysis of Table 7 shows that males are more likely to respond with SC to female intermolt body odor, while females are likely to respond with SO to male intermolt body odor (P < 0.001).

2076

ATEMA AND COWAN DISCUSSION

In interpreting these results, we must consider both the natural biological context of social interactions in lobsters and the experimental design of this study. The study was focused on male responses to female odors and employed two males and one female per tank as odor recipients. In addition, females consistently yielded more urine than males despite our attempts to collect urine from comparable numbers of males and females each day. This resulted in unequal sample sizes. Finally, our experimental design was based on observing responses of males in shelters receiving male and female urine and body odors. This is the natural context in which a male discriminates between lobsters approaching his shelter entrance, and by which he allows premolt females to enter and pair bond (Atema et al., 1979; Karnofsky and Atema, t979; Atema and Cobb, 1980; Atema, 1986). Courting females, however, do not sit in shelters; they walk around and approach male shelters selecting a mate. For females, the situation of receiving social odors while inside a shelter may not be conducive to producing behavioral responses meaningful to courtship. A bioassay focused on female behavior should be based on her natural courtship behavior; this was not the aim of this study. Nevertheless females showed interesting responses and differentiated well between various stimuli presented. Our interpretation of the results is based on a statistical comparison between responses to test and control stimuli to avoid being misled by the greatly unequal sample size of the different stimuli presented to males and females. In discriminating between responses, we are not placing great weight on the meaning of the different behaviors measured (see also Rose, 1984). Initially, we focus on the number of behavioral changes. In the interpretation of results, we will not specifically consider the possibility of masking "pheromones" that might suppress responses, although their existence is of theoretical interest. Overall, males responded more strongly than females when presented with the same stimuli. The six test stimuli caused greatly different responses in both sexes, both qualitatively and quantitatively. But can we conclude that there are qualitative chemical differences in the six stimuli? Body odors may contain various amounts of urine; males and females may produce different amounts of urine; recently molted animals may produce different amounts of urine than intermolt animals. Moreover, the response difference between males and females may be merely a result of a sex difference in response threshold, not the recognition of qualitatively different compounds in the six stimuli. Thus, one extreme hypothesis is that all stimuli are various dilutions of lobster urine, regardless of sex or molt state. On the other extreme one might postulate different chemical compositions of male and female urine and special molt signals for both sexes. We argue that our results contradict the urine dilution hypothesis, and we provide evidence for special male and female molt substances.

U R I N E A ND M O L T SIGNALS IN L O B S T E R S

2077

Females showed consistently strong responses to male molt body odor (Tables 2, 3, and 4). Whereas in general females were less responsive than males to the same stimuli, in the case of male molt body odor female response was greater than male response (Table 3). Additionally, in contrast to males, females responded more strongly to male molt body odor than to either female molt body odor, or male intermolt body odor or urine (Tables 2 and 4). These results support the hypothesis of a qualitatively different (set of) substance(s) in male molt body odor to which females but not males respond. There may or may not be other compounds in male molt body odor that caused the weak male response. Female intermolt urine caused similarly strong responses in males and females (Tables 2 and 3), although males showed HOL (Table 2). However, males responded more strongly than females to female molt body odor (Tables 2 and 3). If female molt body odor were merely a dilute female urine, one would expect it to cause weaker but still similar responses in both males and females. However, in reality males responded strongly to female molt body odor and females not at all (Table 2). This result is also evident in the direct stimulus comparisons of Table 4, where female urine and female molt body odor cause greater response differences in females than in males: females responded strongly to female urine but not to female molt body odor. These results support the hypothesis that females produce a (set of) special molt substance(s) to which males but not females respond. In addition to evidence for special molt substances in male and female molt body odor, the results point to differences between male and female intermolt urine. Female urine caused stronger responses than male urine in both sexes. Thus, according to the dilution hypothesis, female urine may contain a higher concentration of the same active fraction than male urine. However, based on male and female response similarity to female urine (Tables 2 and 3 and Figure 2), the dramatic difference between males and females in their responses to male urine (Tables 2 and 3 and Figure 2) would not be expected, if male urine were merely a dilute form of female urine. While these latter results cannot exclude the dilution hypothesis, it seems more likely to suspect some qualitative rather than merely quantitative sex differences in intermolt urine. The demonstration of a female-specific molt odor was expected based on previous work (Atema and Engstrom, 1971; Atema et al., 1979). The present experiments provide an indication of how a quantified bioassay might be designed to pursue chemical isolation and identification of the active compound(s). These and the female blue crab sex pheromone(s) (Gleeson et al., 1984) would become the first marine animal sex pheromones to be identified chemically. In contrast, we did not expect to demonstrate the presence of a male-specific molt odor causing particular excitement for females (Table 2: locate the source of release, check the shelter entrances, and close seizer claw), nor can

2078

ATEMA AND COWAN

we at present provide a natural context for such a response. Females residing in shelters are not likely to ever receive male molt odor and, if they did, we would not know the biological significance of responding to it. The possibility exists that molting in all lobsters is accompanied by a molt-specific odor which might contain, for instance, the arthropod molting hormone, crustecdysone (Atema and Gagosian, 1973) or a metabolite of it. However, the present experiments showed clearly that males and females responded differently to male and female molt odors. One might argue that if intrasexual competition were severe, then both sexes should respond to intrasexual molt odors. This would allow each sex to eliminate a freshly molted and thus vulnerable competitor. Similarly, one might argue that intermolt urine and body odor could act as warning signals announcing the presence of a potentially dangerous competitor, particularly for shelter possession. (It should be remembered that all recipients were tested in their shelters.) These expectations based on sexual selection theory may find some confirmation in the results of the analysis of SC and SO. Both sexes showed higher than expected durations of wide-open seizer claw (SO) in response to intrasexual molt body odor and intermolt urine. A wide-open seizer claw is seen in defensive and offensive agonistic interactions. Similarly, a higher than expected incidence of closed seizer claw (SC) was observed in response to intersexual molt body odor and intermolt urine. (It should be emphasized that SC and SO are, of course, mutually exclusive, but not dependent, because the relaxed seizer claw position is half open.) Finally, intermolt body odor also caused a sex difference in responses: males tended to close their seizer claw when presented with female body odor, whereas females more often opened their seizer claw toward male body odor. These results seem to indicate that intermolt females are wary of intermolt males, but that males are appeasing to females. Such an interpretation would fit data collected in the field showing that smaller males could successfully defend their shelters against larger females, but not the reverse (E.B. Kamofsky, J. Atema and R.H. Elgin, unpublished observation). To interpret these results properly will require further behavioral-ecological investigation in the field and in naturalistic environments. The pronounced responses to intermolt urine (0.5 ml diluted several orders of magnitude before reaching the lobster's chemoreceptors) indicate that urine may be used for social communication in general. The fact that male and female inte~rnolt body odors were weak stimuli for males, and caused no response in females, may indicate that body odor contained little urine. Perhaps lobsters do not release urine constantly or under the conditions of confinement in our body odor collecting tank, as argued for crabs (Christofferson, 1978). Together with the presence of a urinary bladder and the positioning of the nephropores just above the gill current, these results argue for controlled urine release and the use of urine as a distinct chemical signal. The urine signal could be injected under the animal's control into the gill current, which would carry it up to seven

URINE AND MOLT SIGNALS IN LOBSTERS

2079

animal lengths forward (McPhie and Atema, 1984; Atema, 1985) providing identification of species, sex, and perhaps dominance status, and individual identity. The gill current can also be abruptly reversed by fanning the exopodites of the maxillipeds (McPhie and Atema, 1984; Atema, 1985), thus allowing the animal to not project its body odor with or without urine and instead to sample water chemistry ahead of itself. Such control is required for a communication function. Identifying potential mating partners and competitors appears to be an important aspect of the seemingly loose social structure of a resident lobster population (E.B. Karnofsky, J. Atema and R.H. Elgin, unpublished observation). Under the normally dark, nocturnal conditions of their circadian activity period, chemical identification may well be the best social cue available, and considering the different control mechanisms available to the lobster, urine would be a convenient vehicle for chemical signals. Acknowledgments--Partial financial support for this study was provided by NSF grant BNS 8413661. We thank Donna McPhie and Mary Wisgirda for help in urine collection and in keeping the stimuli unknown to the observer, and Dr. Rainer Voigt and Gary Banta for help in computer programming and statistical analysis.

REFERENCES

AMEYAW-AKUMFI, C., and HAZLETT, B.A. 1975. Sex recognition in the crayfish Procambarus clarkii. Science 190:1225-1226. ATEMA, J. 1985. Chemoreception in the sea: Adaptations of chemoreceptors and behavior to aquatic stimulus conditions. Soc. Exp. Biol. Symp. 39:387-423. ATEMA, J. 1986. A review of sexual selection and chemical communication in the lobster, Homarus americanus. Can. J. Fish. Aquat. Sci. In press. ATEMA, J., and COBB, J.S. 1980. Social behavior, pp. 409-450, in J.S. Cobb and B.F. Phillips (eds.). The Biology and Management of Lobsters, Vol. 1, Academic Press, New York. ATEMA, J., and ENGSTROM, D.G. 1971. Sex pheromone in the lobster, Homarus americanus. Nature 232:261-263. ATEMA, J., and GAGOSIAN, R.B. 1973. Behavioral responses of male lobsters to ecdysones. Mar. Behav. Physiol. 2:15-20. ATEMA, J., and STENZLER, D. 1977. Alarm response of the marine mud snail Nassarius obsoletus: Specificity and behavioral priority. J. Chem. Ecol. 3:159-171. ATEMA, J., JACOBSON,S., KARNOFSKY,E., OLESZKO-SZUTS,S., and STEIN, L. 1979. Pair formation in the lobster, Homarus americanus: Behavioral development, pheromones and mating. Mar. Behav. Physiol. 6:277-296. ATEMA, J., HOLLAND,K., and IKEHARA,W. 1980. Olfactory responses of yellow-fin tuna (Thunnus albacares) to prey odors: Chemical search image. J. Chem. Ecol. 6:457-465. BRYANT, B., and ATEMA, J. 1986. Diet manipulation affects the social behavior of catfish; the importance of body odor. In preparation. CHRISTOFFERSON,J.P. 1978. Evidence for the controlled release of a crustacean sex pheromone. J. Chem. Ecol. 4:633-639. COWAN, D., and ATEMA, J. 1985. Serial monogamy, mate choice and pre- and postcopulatory guarding in the lobster, Homarus americanus. Biol. Bull. 169:550.

2080

ATEMA AND COWAN

DAHL, E., EMANUELSON,H., and VON MECKLENBURG,C. 1970. Pheromone transport and reception in an amphipod. Science 170:739-740. DUNHAM, P.J. 1978. Sex pheromones in Crustacea. Biol. Rev. Cambridge Philos. Soc. 53:555583. DUNHAM, P.J. 1985. Pheromones and Behavior in Cmstacea, in H. Laufer and R. Downer (eds.). Comparative Endocrinology, Vot. 2. In press. EALES, A.J. 1974. Sex pheromone in the shore crab Carcinus maenas, and the site of its release from females. Mar. Behav. Physiol. 2:345-355. GAGOSIAN, R.B., and ATEMA, J. 1973. Behavioral responses of male lobsters to ecdysone metabolites. Mar. Behav. Physiol. 2:115-120. GLEESON, R.A. 1980. Pheromone communication in the reproductive behavior of the blue crab, Callinectes sapidus. Mar. Behav. Physiol. 7:119-134. GLEESON, R.A., ADAMS, M.A., and SMITH,A.B., III 1984. Characterization of a sex pheromone in the blue crab, Callinectes sapidus: Crustecdysone studies. J. Chem. Ecol. 10(6):913-921. KAMIGUCHI, Y. 1972. Mating behavior in the freshwater prawn, Palaemon paucidens. A study of the sex pheromone and its effect on males. J. Fac. Sci. Hokkaido Univ. Ser. VI. Zool. 18:347355. KARNOFSKY, E.B., and ATEMA, J. 1979. Field and laboratory observations of lobster mating behavior. Biol. Bull. 157:374. KATONA, S.K. 1973. Evidence for sex pheromones in planktonic copepods. Limnol. Oceanogr. 81:374-383. KITTREDGE, J.S., TERRY, M., and TAKAHASHI,F.T. 1971. Sex pheromone activity of the molting hormone, crustecdysone, on male crabs. Fish. Bull. 69(2): 337-343. MCLEESE, D.W. 1970. Detection of dissolved substances by the American lobster (Homarus americanus) and olfactory attraction between lobsters. J. Fish. Res. Board Can. 27:1371-1378. MCLEESE, D.W. 1973. Chemical communication among lobsters (Homarus americanus, J. Fish. Res. Board Can. 30:775-778. MCLEZSE, D.W., SPRAGG~NS,R.L., BOSE, A.K., and PRIMANIK, B.N. 1977. Chemical and behavioral studies of the sex attractant of the lobster (Homarus americanus). Mar. Behav. Physiol. 4:219-232. MCPHIE, D., and ATEMA, J. 1984. Chemical communication in lobsters: Information currents. Biol. Bull. 167:510-511. ROSE, R.D. 1984. Chemical communication in crayfish: Physiological ecology, realism and experimental design. J. Chem. Ecol. 10:1289-1292. RYAN, E.P. 1966. Pheromone: Evidence in a decapod crustacean. Science 151:340-341. TEYTAUD, A.R. 1971. The laboratory studies of sex recognition in the blue crab, Callinectes sapidus (Rathbum). Univ. Miami Sea Grant Program, Sea Grant Tech. Bull. 15, Miami, Florida. 63 pp.