Pacific Science (1973), Vol. 27, No.1, p. 1-7 Printed in Great Britain

Information Feedback from Photophores and Ventral

Countershading in Mid-Water SquidI

RICHARD EDWARD YOUNG2

ABSTRACT: The arrangement ofphotosensitive vesicles and photophores in two species of mid-water squid suggests that the vesicles function in detecting the in­ tensity ofdownward-directed surface light and the intensity oflight from their own photophores. This information is precisely what is required for an animal to eliminate its ventral shadow by the production ofa ventral bioluminescent glow. This arrangement, therefore, offers strong support for the theory of ventral countershading in mid-water animals.

CEPHALOPODS have photoreceptive structures The function of the photosensitive vesicles other than the eyes. In octopods these organs beyond their photosensitive capacity is un­ have been called epistellar bodies (Young, known. One probable function, however, has 1929); in squid and cuttlefish they have been recently emerged during the course of a study labeled the parolfactory vesicles (Boycott and which is attempting to correlate modifications Young, 1956). In both instances the names are of the photosensitive vesicles with certain as­ associated with the location of the organs: in pects of the ecology of mid-water cephalopods octopods, on the stellate ganglia; and in squid off Hawaii. and cuttlefish, near the olfactory lobe on the op­ I would like to thank J. Z. Young, University tic stalk of the brain. In spite of the different College London; N. B. Marshall, British Mu­ locations of these organs, they are probably seum (Natural History); C. F. E. Roper, Smith­ homologous structures (Nishioka, Hagadorn, sonian Institution; J. M. Arnold, Pacific Bio­ and Bern, 1962; Nishioka et aI., 1966, and per­ medical Research Center; T. Okutani, Tokai sonal observations). It is, therefore, no longer Regional Fisheries Research Laboratory; and desirable to maintain a separate terminology in J. Walters and S. Amesbury, University of the different groups. Since neither ofthe present Hawaii, for reading and commenting on the names is appropriate to all cephalopod groups, manuscript. I also thank T. Clarke, University it becomes necessary to rename the organs. ofHawaii, for providing some of the specimens Since these organs are photoreceptors (Nishioka examined. et aI., 1966; Mauro and Baumann, 1968; Mauro and Sten-Knudsen, 1972), it is suggested that RESULTS the epistellar bodies andtheparolfactoryvesicles be called the photosensitive vesicles. 3 Pelagic cephalopods living below the epi­ pelagic zone during the daytime (i.e., in the mid-water environment) have photosensitive I. This work was supported by grant GB 20993 from the National Science Foundation. Manuscript received vesicles which exhibit many variations; these 10 May 1972. often involve hypertrophy of the organs. Two 2 DepartmentofOceanography, University ofHawaii, such variations have already been reported Honolulu, Hawaii 96822. (Messenger, 1967; Baumann et aI., 1970). The 3 An abstract, treating the modifications ofthe photo­ sensitive vesicles in mid-water squid, was published in most striking modifications that I have observ­ the Proceedings of the Joint Oceanographic Assembly, ed occur in some members of the squid family Tokyo, 1971. In this abstract I used the term "photic Enoploteuthidae. Although the arrangement vesicles." J. Z. Young has since informed me that this varies somewhat between the species examined, term had already been used for certain light-sensitive most basic elements of the general pattern re­ structures in gastropods. Following his suggestion, I now adopt the term "photosensitive vesicles." main constant. The animals studied in most 1 1-2 2 PACIFIC SCIENCE, Volume 27, January 1973

.OP. 5T. E5.

CE. CART...... ~"lt-- ...... (T

P. PHOTo VE5.--t-ihH!r~

NER.---+l:-\

FIG. 1. Lateral dissection of the head of Abraliopsis Sp. SYMBOLS: A. V. PHOTo VES., anterior ventral photosensitive vesicles; CEo CART., cephalic cartilage; D. FU. PHOT., dorsal funnel photophores; D. PHOTo VES., dorsal photosensitive vesicles; L. FU. PHOT., lateral funnel photophores; M. V. PHOTo VES., mid-ventral photosensitive vesicles; NER., nerve; OP. ST., optic stalk; P. PHOTo VES., posterior photosensitive vesicles; V. FU. PHOT., ventral funnel photophores.

detail belong to two undescribed species of stalks of the brain. All of the vesicles lie within Abraliopsis (Burgess, in manuscript). Both of the head adjacent to the inner wall of the ce­ these species are commonly captured in mid­ phalic cartilage. One set is situated dorsally; water trawls off Hawaii. Although the data are two are ventrally located; and the fourth, which rather sparse, these species seem to occupy is absent in one species, occupies a posterior depths during the daytime around 600 meters position within the head (Fig. 1). Each member, and at night they migrate into the upper hun­ or lobe, ofa set consists of a number ofindivi­ dred meters. dual vesicles, each with a lining of cell bodies The structure and arrangement of the photo­ that have photosensitive processes extending sensitive vesicles in these species of Abraliopsis into a central lumen. differ greatly from those found in neritic or Each lobe of the dorsal set of vesicles lies epipelagic squid. The unusual arrangement of close to the dorsal surface of the head and pro­ vesicles is described in the following paragraphs trudes into a concavity in the cephalic cartilage. and provides the basis for the subsequent As a result, the cartilage above each lobe is very discussion. thin. Muscle tissue, although present on all sur­ Squid occupying near-surface waters gener­ rounding areas, is absent from most of the area ally have a single, small set ofvesicles. In con­ immediately above the dorsal vesicles. There­ trast, these deeper living species of Abraliopsis fore, between the vesicles and the surface lie have three to four greatly enlarged sets, all well only the thin cephalic cartilage and the integu­ removed from the usual position on the optic ment. The integument over most ofthe surface Photophores and Ventral Countershading in Squid-YoUNG 3

FIG. 2. A, dorsal view ofhead ofAbraliopsis sp. B, ventral view ofhead ofAbraliopsis sp. with the funnel adductor muscles cut and the funnel reflected posteriorly. Shaded portion indicates areas covered by chromatophores. C, lateral view of Abraliopsis sp. showing ventral predominance of photophores. SYMBOLS: CHR., chromatophore; CUT. FU ADD. MUSe., cut funnel adductor muscle; D. FU. PHOT., dorsal funnel photophores; D. WIN., dorsal window; FU., funnel; NOR. POS. FU., normal position of the funnel; PHOT., photophores; POS. A. V. PHOTo VES., position of the anterior ventral photosensitive vesicles; POS. M. V. PHOTo VES., position of the mid-ventral photosensitive vesicles; V. WIN., ventral window. of the head contains two layers of chromato­ extend ventrally along the inner surface of the phores. The innermost layer is absent from the cephalic cartilage and enter the brain in the re­ integument immediately above the vesicles and gion of the peduncle lobe. the other layer is usually observed with all The ventral sets of vesicles are shaped quite chromatophores contracted. This arrangement differently from the dorsal one. The most an­ provides a distinctive"window" on the surface terior of the two ventral sets consists of some­ of the head through which light has easy access what flattened, sausage-shaped structures with to the dorsal vesicles (Fig. 2A). Viewed dor­ thickened and laterally curved anterior ends sally, each lobe has a nearly circular shape except (Fig. 1). Each lobe of this anteroventral set lies for a V-shaped indentation on its lateral margin. in a depression in the floor of the cartilage ex­ The dorsal surface of the lobe is smooth and cept for a small medial projection which par­ convex while the ventral surface is approxi­ tially lies in a foramen in the cartilage. Each lobe mately flat but rather irregular. seems to consist ofa single layer ofvesicles. At Each dorsal lobe appears to consist of a the thickened anterior end, the vesicles are single layer of elongate vesicles oriented per­ elongate and have their long axes approximately pendicular to the dorsal surface ofthe lobe. The perpendicular to the convex ventral surface of cross sections ofindividual vesicles are approxi­ the organ. The flatter, posterior portions of mately circular. Cell bodies of the photosensi­ each lobe have somewhat irregular elongate tive cells are found primarily along the sides and vesicles which lie parallel to the ventral surface dorsal ends of the individual vesicles, and the of the organ. Cell bodies of the photosensitive photosensitive processes, in general, are aligned cells occur in a compact layer on all surfaces of with the long axis of each vesicle. Large flat­ the individual vesicles but predominate on the tened bundles of nerves from each dorsal lobe side and ventral margins of the vesicles. Each 4 PACIFIC SCIENCE, Volume 27, January 1973 anteroventral lobe is in contact with another In order to interpret the probable functions lobe, the mid-ventral lobe, via a narrow strand of the dorsal and ventral sets of vesicles, it is of vesicles. necessary to examine the arrangements of In contrast to the anteroventral lobes, the photophores in these squid. These species of mid-ventral lobes are very irregular in outline Abraliopsispossess numerous photophores over and are extraordinarily thin (Fig. 1). Individual the surfaces of the body, head, and arms. As vesicles are short and irregular in shape and lack with many mid-water animals, the photophores apparent orientation. Cell bodies of the photo­ are most abundant on the ventral surfaces and sensitive cells predominate along the sides of only a few scattered organs are present dorsally the vesicles and are often absent from the (Fig. 2C). The photophores are very complex, flattened dorsal and ventral surfaces. containing pigment screens, reflecting devices, A large nerve passes from the ventral sets of and lenses. The photophores on the ventral vesicles along the inner wall of the cephalic half of the animal are oriented with lenses cartilage to the region of the peduncle lobe of directed ventrally. Some of the photophores the brain. on the funnel provide a clear exception to this The surface of the head immediately below general arrangement. the vesicles has a distinctive appearance. This The funnel bears six longitudinal series of area, which also overlies the funnel, lacks photophores (Fig. 1). Two series are located on chromatophores and photophores. Head re­ the ventral surface and each has 19 to 20 vent­ tractor muscles which cover much of this area rally directed photophores. Another series is are transparent in the living animal. The ventral located on each ventrolateral portion of the portion of the head between the ventral sets of funnel and has about 10 to 12 ventrally directed vesicles and the funnel, therefore, forms an ex­ photophores. The final two series are located on tensive ventral window through which light the dorsal surface of the funnel, one on either may reach the photosensitive vesicles (Fig. 2B). side of the anterior funnel adductor muscles. The final pair of lobes, the posterior photo­ Each of these series has 22.to 23 photophores sensitive vesicles, lies between the optic lobes which point in a general dorsal direction to a of the brain and the posterior wall of the ce­ centrally located area on the ventral surface of phalic cartilage (Fig. 1). Each is elliptical in out­ the head. The most posterior of these photo­ line except in the extreme ventral portion where phores slant anteriorly and medially, whereas it tapers toward the mid-ventral vesicles with the most anterior ones slant medially only which it mayor may not connect. The posterior (Fig. 1). All of these photophores, therefore, lobes, in contrast to the colorless condition are directed precisely toward the ventral sets of (when preserved) of the other lobes, have a dis­ photosensitive vesicles. It appears that light tinctive yellow coloration in preserved (alcohol) passes from the dorsal funnel photophores specimens. Often several vesicles of the mid­ through the "ventral window" to these vesic­ ventral lobes which are nearest the posterior les. These squid appear to be monitoring light lobes exhibit this same yellow pigmentation. from their own photophores. Each posterior organ is thin but thicker and more turgid than the mid-ventral lobes. The elongate vesicles of the posterior lobes are regularly aligned and oriented parallel to the DISCUSSION ventral surface of the animal, i.e., they extend Information feedback from these photo­ from the median margin toward the lateral mar­ phores in Abraliopsis could be useful in regu­ gin of each organ; some vesicles traverse the lating the spectral composition of the light entire width of a lobe. Cell bodies of photo­ emitted or in regulating some type of flash sensitive cells can be found, equally concentra­ pattern. It seems more likely, however, that the ted, around the entire circumference ofthe indi­ information is used in regulating the intensity vidual vesicles. Nerves appear to run in several of the bioluminescent light. Although there is flattened bundles from the posterior vesicles to no direct evidence that these species have this a point in the brain near the peduncle lobe. capacity, such an ability would be ofgreat value. Photophores and Ventral Countershading in Squid-YoUNG 5

These two species of Abraliopsis live in meso­ in this position around a night light near the pelagic waters during the day where the pene­ Bahama Islands. In this position, the dorsal tration of surface light is probably of sufficient photosensitive vesicles would be well situated intensity to be detected by the highly sensitive for the detection ofdownward-directed surface eyes of many mid-water animals (Clarke and light. The ventral vesicles are ideally suited for Denton, 1962). Clarke (1963) and Fraser (1962) detecting the intensity of light from the have pointed out that animals living in this animal's own photophores. The necessary sen­ zone would be silhouetted against the down­ sory mechanism which would allow ventral ward-directed light and thereby become visible countershading, therefore, is present. to predators below them. They have suggested This mechanism may have an analogous that the predominantly ventral arrangement of counterpart in certain mid-water fishes. Sto­ photophores found in many animals living in miatoid fishes have photophores that are di­ this zone allows their possessor to produce a rected into the eyes (Marshall, 1954). Many beam of downward-directed light which explanations have been proposed for this matches that from the surface and obliterates peculiar arrangement. Piitter (1902) thought the silhouette. Nicol (1967) has discussed this that this light would produce a subliminal illu­ hypothesis of ventral countershading at some mination of the retina, thereby lowering the length, and Denton (1970) has pointed out that effective visual threshold. Brauer (1908) con­ the structure ofthe photophores and the pattern sidered that this light might make it easier for a ofemitted light in the hatchet fish, Argyropelecus, species to recognize the distinctive quality ofits as well as the coloring of this fish, fit well with own light, thereby facilitating recognition of this hypothesis. Foxton (1970) has found that conspecific individuals. Marshall (1954, 1966) the distributional patterns of decapod crusta­ suggested that the light might sensitize the ceans are compatible with this hypothesis. For retina prior to turning on the strong body countershading to be most effective it is prob­ photophores. Nicol (1967) wondered if these ably necessary for bioluminescent light to be photophores might permit the fish to compare produced throughout the daylight period. It is luminous output with the light of the environ­ possible, however, that intermittent counter­ ment or that they might enable a fish to signal shading may be part of an escape mechanism by clocking its emission against a response from which utilizes retreat rather than concealment another fish. Both Brauer and Nicol, therefore, through countershading. Unfortunately, there have suggested that these fish have a system of is no direct observational evidence that con­ bioluminescent feedback. The analogous (and cerns this problem. more clear-cut) situation in Abraliopsis supports In order to utilize ventral countershading, a this general interpretation. species might distribute along a specific isolume. If ventral countershading does occur, it is Such a situation, however, would restrict a almost certainly not the sole function of these species to a narrow depth range due to the rapid photophores, as each species exhibits a distinc­ change of light intensity with depth and would, tive photophore arrangement which suggests a presumably, necessitate frequent movement due species-recognition function. Nor can ventral to short-term fluctuations in light intensity. countershading be the only reason for the Alternatively, ventral countershading would ventral predominance of photophores. Foxton be feasible if the animals could determine the (1970) pointed out that some species ofSergestes intensity of downward-penetrating surface (Sergia) possess ventral photophores but occupy light and could regulate the luminous output of levels (below 700 m off the Canary Islands) at their photophores to match this light. which the low light intensity would probably Abraliopsisprobably orients in the water with make countershading ineffective. This is also its longitudinal body axis in a horizontal posi­ true for some deep living cephalopods (e.g., tion. This is the typical attitude for squid in the Vamp)lToteuthis i'!fernalis, Mastigoteuthis spp.). families Loliginidae, Ommastrephidae, and In both sergestids and cephalopods the photo­ probably others. I have observed the closely re­ phores found in the deepwater species are not lated enoploteuthid. Abralia lJeraf!J'i, swimming as complex as those found in the shallower- 6 PACIFIC SCIENCE, Volume 27, January 1973 living species. Complex photophores (i.e., provides a mechanism which allows adjust­ photophores with reflectors and lenses) are ment of photophore light intensity relative necessary for ventral countershading in order to the intensity of downward-penetrating that most ofthe emitted light be collimated and light from the surface. A ventrally produced directed vertically downward (Denton, 1970). beam of bioluminescent light of the proper Ifthe light were not collimated to some degree, intensity would eliminate the animal's the portion of the expanding light beam be­ silhouette and thereby aid in concealment. neath the animal would decrease in intensity much faster than the downward-penetrating surface light. LITERATURE CITED In conclusion, the ventral photosensitive vesicles in these species ofAbraliopsis appear to BAUMANN, F., A. MAURO, R. MILECCHIA, S. detect bioluminescent light from the animal's NIGHTINGALE, and J. Z. YOUNG. 1970. The own photophores while the dorsal vesicles are extra-ocular light receptors of the squids well situated to detect downward-penetrating Todarodes and IIlex. Brain Res. 21: 275-279. surface light. This arrangement strongly sug­ BOYCOTT, B. B., and J. Z. YOUNG. 1956. The gests that a countershading mechanism is subpedunculate body and nerve and other operating and lends considerable support to the organs associated with the optic tract of countershading hypothesis. cephalopods, p. 76-105. In: Bertil Hanstrom: Zoological papers in honour of his sixty­ fifth birthday. Zool. Inst., Lund. BRAUER, A. 1908. Die Teifsee-Fische. 2 Anat. SUMMARY Teil. Wiss. Ergebn. 'Valdivia' 15: 1-266. 1. The arrangement and structure of the CLARKE, G. L., and E. J. DENTON. 1962. Light photosensitive vesicles of two species of and animal life, p. 456-468. In: M. N. Hill Abraliopsis are described. [ed.] The sea. Vol. 1. John Wiley and Sons, 2. One set of vesicles is located near the dorsal Interscience, New York. surface ofthe head, a second near the ventral CLARKE, W. D. 1963. Function of biolumines­ surface, and the third near the posterior sur­ cence in mesopelagic organisms. Nature, face. The surface of the head adjacent to the Lond. 198: 1244--1246. dorsal and ventral organs either lacks or has DENTON, E. J. 1970. On the organization of a reduced number of chromatophores and, reflecting surfaces in some marine animals. thereby, forms"windows" for the passage Phil. Trans., ser. B, 258: 286-313. of light. FOXTON, P. 1970. The vertical distribution of 3. The dorsal vesicles are well positioned to pelagic decapods (Crustacea: Natantia) col­ detect downward-penetrating surface light. lected on the SOND cruise 1965. J. Mar. 4. The ventral vesicles lie within the head im­ BioI. Ass. u.K. 50: 939-960. mediately above the funnel and oppose a FRASER, J. 1962. Nature adrift. DuFour series of photophores located on the funnel Editions, Chester Springs, Pa. 178 p. and directed at the vesicles. It is suggested MARSHALL, N. B. 1954. Aspects of deep sea that the ventral vesicles provide a feedback biology. Hutchinson, London. 380 p. mechanism for determining the intensity of --. 1966. The life of fishes. World Pub­ photophore light. lishing Co., Cleveland. 402 p. 5. These squid live in a region of low light in­ MAURO, A., and F. BAUMANN. 1968. Electro­ tensity and possess numerous photophores, physiological evidence of photoreceptors in located on the ventral half of the animal, the epistellar body of Eledone lJJoschata. which are oriented in a ventral direction Nature, Lond. 220: 1332-1334. (with the exception of the photophores on MAURO, A., and O. STEN-KNUDSEN. 1972. the dorsal surface of the funnel). Light-evoked impulses from extraocular 6. It is suggested that the arrangement of photoreceptors in the squid Todarodes. photophores and photosensitive vesicles Nature, Lond. 237: 342-343. Photophores and Ventral Countershading in Squid-YOUNG 7

MESSENGER, J. B. 1967. Parolfactory vesicles as NISHIOKA, R. S., 1. YASUMASU, A. PACKARD, photoreceptors in a deep-sea squid. Nature, H. A. BERN, and J. Z. YOUNG. 1966. Nature Land. 213: 836-838. ofvesicles associated with the nervous system NICOL, J. A. C. 1967. The luminescence of of cephalopods. Z. Zellforsch. 75: 301-316. fishes. Symp. Zool. Soc. Lond. 19: 27-55. PUTTER, A. 1902. Die Augen der Wassersauge­ NISHIOKA, R. S., 1. R. HAGADORN and H. A. thiere. Zool. Jber. 2, 17: 99. BERN. 1962. The ultrastructure of the epi­ YOUNG, J. Z. 1929. Sopra un nuovo organo dei stellar body ofthe octopus. Z. Zellforsch. 27 : cephalopodi. Boll. Soc. Ital. BioI. Spero 4: 406-421. 1,022-1,024.