Journal of Chemical Ecology, Vol. 10, No. 8, 1984
L e t t e r to t h e E d i t o r
CHEMICAL COMMUNI'CATION IN CRAYFISH: Physiological Ecology, Realism and Experimental Design 1 In writing my response (Rose, 1982) to Itagaki and Thorp (198 l) concerning sex pheromones and Procambarus clarkii, I compiled a list of potential criticisms and reinterpretations of their methods, data, interpretations, and conclusions. I then compiled the evidence for and against each item. Some items subsequently were included while others were not. Hazlett (1984) has recently resurrected one item (not included) and examined this item in the broader context of the role of realism in ecological experimental design. The present contribution addresses both the specific and general questions raised by Hazlett. The specific question brought forth by Hazlett is whether or not lentic animals should be tested behaviorally under lotic conditions. He suggests they should not; I believe experimental design depends on the specific question being examined and that there are no hard and fast rules. There are two issues relevant to Itagaki and Thorp's experiments: (1) the effectiveness of the stimuli tested, and (2) the appropriateness of the overall design. 1. In the crustacean pheromone literature, one finds cause to doubt the effectiveness of Itagaki and Thorp's (1981) stimuli. In brief, Itagaki and Thorp drew stimulus waters from "head" tanks that contained animals for a 90-min acclimation period. Test duration was 20 rain. Therefore the total time that the "stimulus" animals were in the head tanks was 1 l0 min. Ryan (1966) reports that there was "no regularity of response" when "stimulus" animals were in "head tanks" for less than 2 hr in his tests on sex pheromones in portunid crabs (similar data for crayfish does not exist to my knowledge). Furthermore, Christofferson (1978) has shown restraint to inhibit release of pheromone in this same crab for up to several hours. It is not inconceivable that animals respond similarly (at least with respect to micturition) to situations of"restraint" or handling and introduction into a new environment. (There is substantial evidence that some crustacean pheromones are borne in the urine produced by the antennal glands.) It is therefore possible that Itagaki and Thorp's apparatus was frequently either (1) devoid of sex pheromone or (2) subjected to extremely low levels (but see Rose, 1982) especially if urine is the vehicle for putative sex pheromone in P. clarkii. ~This work supported by NIMH 08323 and NY-NPA 454154. 1289 0098-0331/84/0800-1289503.50/0 9 Plenum Publishing Corporation
1290
LETTER TO THE EDITOR
2. Procambarus elarkii is the red swamp crayfish. Current flow rates of swamps vary tremendously depending on microlocation. In some areas water is virtually stagnant, while in others there is appreciable flow. It is therefore difficult to evaluate the flow rate in Itagaki and Thorp's apparatus (1.5 cm/sec) with respect to P. clarkii and chemoreception. Commercially raised crayfish from Louisiana are raised under sprinkler systems (Waubun Laboratories, Schriever, La., personal communication; these animals are commercially available through Carolina Biological Supply). Itagaki and Thorp used commercially available animals. One may question whether or not a signal perceived by such animals will be appropriately acknowledged behaviorally under such experimental conditions as described (see also below). Furthermore, Gleeson (1980) has found some behaviors attenuated or partially lacking in the blue crab when observed under laboratory conditions. Hazlett also raises a more general question concerning the role of ecological realism in experimental design. This role varies. In some instances one is interested in the "normal situation" (e.g., some crayfish communieation studies), and one would like to mimic, as nearly as possible, the natural situation. In such studies one changes only those parameters necessary for standardization of results and quantification of effect, if any effect exists. Other studies, however, require the examination of some aspect of an organism at or near its physiological limits, e.g., Schneider's (1957) studies of pheromone detection by moths. Thus, in order to assure the best experimental design, one must use an integrative approach and draw upon the findings of all physical and biological sciences. Frequently there are physiological tests for the demonstration of reception of specific stimuli (e.g., electroretinograms, single unit electrophysiological recordings, etc.) or the components necessary for reception (e.g., microspeetrophotometry of visual pigments), although there may be difficulty in extracting signal from noise. There are also anatomical criteria for the demonstration of certain receptor types, e.g., cones in the cat retina (see Blough and Yager, 1972). Anatomy and physiology should be studied in conjunction with ecology (see Ameyaw-Akumfi and Hazlett, 1975). Positive results from anatomical and physiological tests may suggest perception (see below), but behavioral results are necessary for its demonstration. Clear statement of appropriate hypotheses eases seleetion of experimental design and interpretation of results. Factors such as motivation of subject animals may further complicate the choice of the most appropriate experimental design. Three examples are drawn from the literature on vision in which motivation and testing conditions have altered results, and it is likely that as the pheromone literature increases, examples from this field will be available also: (I) Rose and Menzel (1981) have behavioral evidence suggesting that foraging honeybees exhibit differences in their ability to discriminate colored disks depending on location
LETTER TO THE EDITOR
1291
(hive entrance vs. food source) and independent of illumination of those disks. It seems doubtful that the receptive cells or their physiology are different under the two sets of circumstances. (2) Food-deprived cats (80% free-feeding weight, 22 hr hungry) can discriminate colors while satiated cats will not Mello and Peterson, 1963; Daw and Pearlman, 1970). (3) Weiskrantz et al. (1974), studying "blindsight" in man have shown that striate cortex lesion results in functional blindness with respect to verbal tests for sightedness, but "sightedness" with respect to nonverbal responses to visual stimuli. These examples serve to illustrate that animals are often aware of stimuli even though they may not be responsive. Furthermore, they may be aware of, and responsive to, additional cues or motivational states, as in (1) above, which are unforseen by the experimenters yet are important to the test subjects. Thus one must be extremely cautious in the design and choice of behavioral experiments, In many instances one measures performance, not motivation nor perception. One must bear this in mind always, especially in the interpretation of negative results. It is possible to infer perception (and therefore the existence of something perceived) from differential performance in different test situations, but one cannot conclude nonperception (and nonexistence of perceivable stimuli) from nondifferential performance. Realism may be manifest in terms of experimental habitat, stimulus frequencies and intensities, and overall experimental design. Presumably one is more likely to obtain accurate results when the experimental situation closely approximates the natural one (see Hazlett, 1984). Efforts at ecological realism, however, should not encroach upon the interpretation of differential results. Furthermore, behaviors observed under laboratory conditions may not be indicative of those occurring in the field; behaviors may be lacking, attenuated, inappropriate, or different. Therefore it is frequently beneficial for an investigator to consider (and where possible incorporate) all that is known about the animal, or system, under study in terms of habitat, behavior, physiology, biochemistry, evolution, and genetics when designing realistic ecological experiments aimed at answering particular questions. R.D. Rose Department of Neurobiology and Behavior State University of New York at Stony Brook Stony Brook, New York 11794
REFERENCES AMEYAW-AKUMFI,C., and HAZLETT,B.A. 1975. Sex recognition in the crayfish P r o c a m b a r u s clark ii. Science 190:1225-1226. BLOUGtt, D.S., and YAGER, D. 1972. Visual psychophysics in animals, pp. 732-763, in D.
1292
LETTER TO THE EDITOR
Jameson and L.M. Hurvich (eds.). Handbook of Sensory Physiology. Springer-Verlag, Berlin. CHRISTOFFERSON,J.P. 1978. Evidence for the controlled release of a crustacean sex pheromone. J. Chem. Ecol. 6:633-639. DAw, M.W., and PEARLMAN, A.L. 1970. Cat colour vision: Evidence for more than one cone process. J. Physiol. (London) 211:125-137. GLEESON, R.A. 1980. Pheromone communication in the reproductive behavior of the blue crab, Callinectes sapidus. Mar. Behav. Physiol. 7:119-134. HAZLETT, B.A. 1984. Experimental design and ecological realism. J. Chem. Ecol. 10:1281-1282. ITAGAKI, H., and THORP, J.H. 1981. Laboratory experiments to determine if crayfish can communicate chemically in a flow-through system. J. Chem. Ecol. 7:115-126. MELLO, N.K., and PETERSON, N.J. 1964. Behavioral evidence for color discrimination in cat. J. Neurophysiol. t7:289-294. RosE, R.D. 1982. On the nature of chemical communication by crayfish in a laboratory controlled flow-through system. J. Chem. Ecol. 8:1065-1071. RosE, R.D., and MENZEL, R. 1981. Luminance dependence of pigment color discrimination in bees, J. Comp. Physiol. 14l:379-388. RYAN, E.P. 1966. Pheromone: Evidence in a decapod crustacean. Science 170:739-740. SCnNEmER, D. 1957. Elektrophysiologische Untersuchungen yon Chemo- und Mechanorezcptoren er Antenne des Seidenspinners Bombyx rnori L. Z. Vergl. Physiol. 40:8-41. WEIS~:RANTZ, L., WARR1NGTON,E.K., SANDERS, M.D., and MARSHALL,J. 1974. Visual capacity in the hemianpoic field following a restricted occipital ablation. Brain 97:709-728.
L e t t e r to t h e E d i t o r
CHEMICAL COMMUNI'CATION IN CRAYFISH: Physiological Ecology, Realism and Experimental Design 1 In writing my response (Rose, 1982) to Itagaki and Thorp (198 l) concerning sex pheromones and Procambarus clarkii, I compiled a list of potential criticisms and reinterpretations of their methods, data, interpretations, and conclusions. I then compiled the evidence for and against each item. Some items subsequently were included while others were not. Hazlett (1984) has recently resurrected one item (not included) and examined this item in the broader context of the role of realism in ecological experimental design. The present contribution addresses both the specific and general questions raised by Hazlett. The specific question brought forth by Hazlett is whether or not lentic animals should be tested behaviorally under lotic conditions. He suggests they should not; I believe experimental design depends on the specific question being examined and that there are no hard and fast rules. There are two issues relevant to Itagaki and Thorp's experiments: (1) the effectiveness of the stimuli tested, and (2) the appropriateness of the overall design. 1. In the crustacean pheromone literature, one finds cause to doubt the effectiveness of Itagaki and Thorp's (1981) stimuli. In brief, Itagaki and Thorp drew stimulus waters from "head" tanks that contained animals for a 90-min acclimation period. Test duration was 20 rain. Therefore the total time that the "stimulus" animals were in the head tanks was 1 l0 min. Ryan (1966) reports that there was "no regularity of response" when "stimulus" animals were in "head tanks" for less than 2 hr in his tests on sex pheromones in portunid crabs (similar data for crayfish does not exist to my knowledge). Furthermore, Christofferson (1978) has shown restraint to inhibit release of pheromone in this same crab for up to several hours. It is not inconceivable that animals respond similarly (at least with respect to micturition) to situations of"restraint" or handling and introduction into a new environment. (There is substantial evidence that some crustacean pheromones are borne in the urine produced by the antennal glands.) It is therefore possible that Itagaki and Thorp's apparatus was frequently either (1) devoid of sex pheromone or (2) subjected to extremely low levels (but see Rose, 1982) especially if urine is the vehicle for putative sex pheromone in P. clarkii. ~This work supported by NIMH 08323 and NY-NPA 454154. 1289 0098-0331/84/0800-1289503.50/0 9 Plenum Publishing Corporation
1290
LETTER TO THE EDITOR
2. Procambarus elarkii is the red swamp crayfish. Current flow rates of swamps vary tremendously depending on microlocation. In some areas water is virtually stagnant, while in others there is appreciable flow. It is therefore difficult to evaluate the flow rate in Itagaki and Thorp's apparatus (1.5 cm/sec) with respect to P. clarkii and chemoreception. Commercially raised crayfish from Louisiana are raised under sprinkler systems (Waubun Laboratories, Schriever, La., personal communication; these animals are commercially available through Carolina Biological Supply). Itagaki and Thorp used commercially available animals. One may question whether or not a signal perceived by such animals will be appropriately acknowledged behaviorally under such experimental conditions as described (see also below). Furthermore, Gleeson (1980) has found some behaviors attenuated or partially lacking in the blue crab when observed under laboratory conditions. Hazlett also raises a more general question concerning the role of ecological realism in experimental design. This role varies. In some instances one is interested in the "normal situation" (e.g., some crayfish communieation studies), and one would like to mimic, as nearly as possible, the natural situation. In such studies one changes only those parameters necessary for standardization of results and quantification of effect, if any effect exists. Other studies, however, require the examination of some aspect of an organism at or near its physiological limits, e.g., Schneider's (1957) studies of pheromone detection by moths. Thus, in order to assure the best experimental design, one must use an integrative approach and draw upon the findings of all physical and biological sciences. Frequently there are physiological tests for the demonstration of reception of specific stimuli (e.g., electroretinograms, single unit electrophysiological recordings, etc.) or the components necessary for reception (e.g., microspeetrophotometry of visual pigments), although there may be difficulty in extracting signal from noise. There are also anatomical criteria for the demonstration of certain receptor types, e.g., cones in the cat retina (see Blough and Yager, 1972). Anatomy and physiology should be studied in conjunction with ecology (see Ameyaw-Akumfi and Hazlett, 1975). Positive results from anatomical and physiological tests may suggest perception (see below), but behavioral results are necessary for its demonstration. Clear statement of appropriate hypotheses eases seleetion of experimental design and interpretation of results. Factors such as motivation of subject animals may further complicate the choice of the most appropriate experimental design. Three examples are drawn from the literature on vision in which motivation and testing conditions have altered results, and it is likely that as the pheromone literature increases, examples from this field will be available also: (I) Rose and Menzel (1981) have behavioral evidence suggesting that foraging honeybees exhibit differences in their ability to discriminate colored disks depending on location
LETTER TO THE EDITOR
1291
(hive entrance vs. food source) and independent of illumination of those disks. It seems doubtful that the receptive cells or their physiology are different under the two sets of circumstances. (2) Food-deprived cats (80% free-feeding weight, 22 hr hungry) can discriminate colors while satiated cats will not Mello and Peterson, 1963; Daw and Pearlman, 1970). (3) Weiskrantz et al. (1974), studying "blindsight" in man have shown that striate cortex lesion results in functional blindness with respect to verbal tests for sightedness, but "sightedness" with respect to nonverbal responses to visual stimuli. These examples serve to illustrate that animals are often aware of stimuli even though they may not be responsive. Furthermore, they may be aware of, and responsive to, additional cues or motivational states, as in (1) above, which are unforseen by the experimenters yet are important to the test subjects. Thus one must be extremely cautious in the design and choice of behavioral experiments, In many instances one measures performance, not motivation nor perception. One must bear this in mind always, especially in the interpretation of negative results. It is possible to infer perception (and therefore the existence of something perceived) from differential performance in different test situations, but one cannot conclude nonperception (and nonexistence of perceivable stimuli) from nondifferential performance. Realism may be manifest in terms of experimental habitat, stimulus frequencies and intensities, and overall experimental design. Presumably one is more likely to obtain accurate results when the experimental situation closely approximates the natural one (see Hazlett, 1984). Efforts at ecological realism, however, should not encroach upon the interpretation of differential results. Furthermore, behaviors observed under laboratory conditions may not be indicative of those occurring in the field; behaviors may be lacking, attenuated, inappropriate, or different. Therefore it is frequently beneficial for an investigator to consider (and where possible incorporate) all that is known about the animal, or system, under study in terms of habitat, behavior, physiology, biochemistry, evolution, and genetics when designing realistic ecological experiments aimed at answering particular questions. R.D. Rose Department of Neurobiology and Behavior State University of New York at Stony Brook Stony Brook, New York 11794
REFERENCES AMEYAW-AKUMFI,C., and HAZLETT,B.A. 1975. Sex recognition in the crayfish P r o c a m b a r u s clark ii. Science 190:1225-1226. BLOUGtt, D.S., and YAGER, D. 1972. Visual psychophysics in animals, pp. 732-763, in D.
1292
LETTER TO THE EDITOR
Jameson and L.M. Hurvich (eds.). Handbook of Sensory Physiology. Springer-Verlag, Berlin. CHRISTOFFERSON,J.P. 1978. Evidence for the controlled release of a crustacean sex pheromone. J. Chem. Ecol. 6:633-639. DAw, M.W., and PEARLMAN, A.L. 1970. Cat colour vision: Evidence for more than one cone process. J. Physiol. (London) 211:125-137. GLEESON, R.A. 1980. Pheromone communication in the reproductive behavior of the blue crab, Callinectes sapidus. Mar. Behav. Physiol. 7:119-134. HAZLETT, B.A. 1984. Experimental design and ecological realism. J. Chem. Ecol. 10:1281-1282. ITAGAKI, H., and THORP, J.H. 1981. Laboratory experiments to determine if crayfish can communicate chemically in a flow-through system. J. Chem. Ecol. 7:115-126. MELLO, N.K., and PETERSON, N.J. 1964. Behavioral evidence for color discrimination in cat. J. Neurophysiol. t7:289-294. RosE, R.D. 1982. On the nature of chemical communication by crayfish in a laboratory controlled flow-through system. J. Chem. Ecol. 8:1065-1071. RosE, R.D., and MENZEL, R. 1981. Luminance dependence of pigment color discrimination in bees, J. Comp. Physiol. 14l:379-388. RYAN, E.P. 1966. Pheromone: Evidence in a decapod crustacean. Science 170:739-740. SCnNEmER, D. 1957. Elektrophysiologische Untersuchungen yon Chemo- und Mechanorezcptoren er Antenne des Seidenspinners Bombyx rnori L. Z. Vergl. Physiol. 40:8-41. WEIS~:RANTZ, L., WARR1NGTON,E.K., SANDERS, M.D., and MARSHALL,J. 1974. Visual capacity in the hemianpoic field following a restricted occipital ablation. Brain 97:709-728.
