Abstract The jumbo squid, Dosidicus gigas, exhibits chromatophore-effected flashing behavior at frequencies of about 3 Hz. In order to characterize this behavior quantitatively and to make it possible to speculate as to its function, videos of D. gigas in its natural environment were taken and analyzed. There is great variability in the consistency of flashing rates among individuals: some stay relatively stable while others fluctuate widely in short spans of time. Videos of two squid encountering one another suggested that squid adjust to their conspecifics' flashing patterns in complex ways. A greater corpus of data will be necessary before specific hypotheses about the function of D. gigas’ flashing behavior can be formulated. Introduction Relatively little is known about Dosidicus gigas, commonly called the Humboldt squid, red devil or jumbo squid. Three intraspecific groups have recently been distinguished by their difference in adult size, with the largest squid reaching mantle lengths of 2 1 m and weights of 50 kg (Nesis, 1983). Adult males in the small-sized group reach 260 mm, and the medium-sized adult males reach 420 mm. Dosidicus gigas is fished extensively off Baja California and the California coast (37-40 deg. N) as well as the west coast of Central and South America (45-47 deg. S) (Nesis, 1970). It is epipelagic to several hundred meters, and rises to the surface both night and day to feed. Its diet consists of fish (most commonly epipelagic lanternfish) and squid (especially D. gigas), while it is preyed on by sperm whales, billfishes, tunas and humans (Norman, 2000). The life cycle of D. gigas is normally one year, and female fecundity is the highest known among cephalopods (Nigmatullin et al., 2001). In Peruvian populations, spawning takes place off the coast of Peru throughout the year, with a mean between October and January and a secondary peak between July and August (Tafur et al., 2001), and planktonic larvae are carried south to southern Peru and Chile (Fernández and Vásquez, 1995). Beyond this broadly drawn life history, little work has been published on the behavior of Dosidicus gigas, and what has been published is anecdotal. In particular, its flashing behavior has never been quantified or described in any detail. In Dosidicus as in other cephalopods, flashing is effected by millions of chromatophores, neuromuscular organs in the dermis that are innervated by the brain. Chromatophores are tiny cytoelastic sacs filled with pigment granules and encircled by excitatory radial muscle fibers Introduction Relatively little is known about Dosidicus gigas, commonly called the Humboldt squid, red devil or jumbo squid. Three intraspecific groups have recently been distinguished by their difference in adult size, with the largest squid reaching mantle lengths of 2 1 m and weights of 50 kg (Nesis, 1983). Adult males in the small-sized group reach 260 mm, and the medium-sized adult males reach 420 mm. Dosidicus gigas is fished extensively off Baja California and the California coast (37-40 deg. N) as well as the west coast of Central and South America (45-47 deg. S) (Nesis, 1970). It is epipelagic to several hundred meters, and rises to the surface both night and day to feed. Its diet consists of fish (most commonly epipelagic lanternfish) and squid (especially D. gigas), while it is preyed on by sperm whales, billfishes, tunas and humans (Norman, 2000). The life cycle of D. gigas is normally one year, and female fecundity is the highest known among cephalopods (Nigmatullin et al., 2001). In Peruvian populations, spawning takes place off the coast of Peru throughout the year, with a mean between October and January and a secondary peak between July and August (Tafur et al., 2001), and planktonic larvae are carried south to southern Peru and Chile (Fernández and Vásquez, 1995). Beyond this broadly drawn life history, little work has been published on the behavior of Dosidicus gigas, and what has been published is anecdotal. In particular, its flashing behavior has never been quantified or described in any detail. In Dosidicus as in other cephalopods, flashing is effected by millions of chromatophores, neuromuscular organs in the dermis that are innervated by the brain. Chromatophores are tiny cytoelastic sacs filled with pigment granules and encircled by excitatory radial muscle fibers (Hanlon and Messenger, 1988). When the fibers are stimulated, they contract and the chromatophore expands. The pigment granules create what looks like a tiny spot on the dermis, a state we will refer to as ON. In D. gigas, when a group of chromatophores are simultaneously switched ON, the entire area appears reddish-brown because of the red-brown chromatophore pigment. To colorblind conspecifics (and to us, in blue-shifted light), this state appears simply as dark. Somewhat counterintuitively, then, the bright flashes occur when the chromatophores are in the OFF state. Öther animals exhibit similar flashing behavior, both within the class Cephalopoda and outside it, notably in the class Insecta. Within Cephalopoda, the binary nature of D. gigas flashing, as well as its speed distinguishes it from the flashing of other species. Sepia species, for example, often use their ornate patterning as camouflage and countershading (Messenger, 1988). The best-studied group of flashing insects is the fireflies (genus Pteroptyx and Photinus) (Buck, 1988). Buck reviews many of the competing theories of the function of firefly flashing, all of which relate flashing behavior to courtship and mating. We used image analysis software to examine video clips of D. gigas swimming freely and on fishing lures. Pairs of individuals appear to synchronize briefly and sporadically with one another, and to maintain a more constant rhythm when they are isolated. Materials and Methods All of the videos used in analysis were filmed off Santa Rosita, BCS in October 2001. They were made by Bob Cranston, an independent filmmaker, using SCUBA at depths of 2 to 50 0 meters. The videos were filmed with a Sony broadcast HDTV camera and custom housing, and directly copied onto Hi-8 master videotapes. Selected sequences from these tapes were converted to .avi files using Adobe Premier and analyzed using ImagePro Plus. In ImagePro Plus, we used the AOI tool and the Intensity histogram to determine the grayscale value of selected areas of the image. At all times the original sampling rate of the video camera (30 Hz) was conserved. After culling data from the .avi clips, we employed Microsoft Excel 5.0 and IgorPro to generate the graphs and figures. Excel was used to create the original spreadsheets, which were imported to Igor to create the flashing graphs. Further observations on chromatophore density and arrangement in D. gigas were made using the Hi-8 tapes. Results The flashing behavior Dosidicus gigas is unique among the cephalopods. Instead of a large array of patterns, Dosidicus has two states, which we have called ON and OFF. ON indicates chromatophore expansion; the squid appears dark red. ÖFF indicates chromatophore contraction, in which the squid has a whitish appearance. When not flashing, the squids remain in either the ON or the ÖFF state; there is no clear "resting state“ despite the fact that the ON state requires ongoing excitation. Moreover, squid in the video documentation (free-swimming, at least several meters from the photographer) never exhibited partial activation: each flash seemed to be an all-or-none phenomenon. The minimum number of flashes in a train is just one. In abnormal conditions (usually while it was being tagged out of water), the squid sometimes flashed only its fins or arms. The maximal rate of flashing observed was 6 flashes/sec, though most of the squid in the clips examined flashed at a rate between 3/sec and 4/sec. A typical flashing cycle is shown in Figure 1, a series of twelve frames spanning 333 ms. To determine more quantitatively the flashing pattern of an individual squid, we examined several clips of single squid that remained relatively still and stayed close to the camera for at least five seconds. We then took intensity measurements on three or four areas of the squid’s body and plotted these as a function of time (see Figure 2). Along the squid body, there is no completely consistent pattern in peak ON states, but the chromatophores on the head and arms are generally activated first. They are followed within 100 ms by mantle and fin chromatophores. However, the mantle and fins do not follow a strict anterior-posterior waveform. Sometimes the fins are fully activated before the entire mantle reaches its peak brightness. Figures 3 through 5 examine the behavior of a single squid. These clips were chosen with the criterion that there be only one individual pictured for the duration of the clip. Of course, we cannot be sure that there were not squid nearby but outside the camera’s scope, but given the data this seemed the most reasonable method. Figures 3a, 4a, and 5a plot the light intensity of the squid mantle against time, while figures 3b, 4b and 5b plot the peak ON signals against time. These three sets of figures show the variability of D. gigas flashing behavior. The squid analyzed in Figure 3 shows a relatively stable rate of flashing, between 11 and 12 frames per cycle (or about 2.9 Hz). The following figure, Figure 4, exemplifies an intermittent flashing pattern, in which bursts of flashing are punctuated by periods of weaker flashing or a cessation of chromatophore activation. Finally, the graphs in Figure 5 are of a squid with a rapidly fluctuating flashing rate (between 7 and 15 frames per cycle, or 4.7 and 2.2 Hz). The markers at the very top of each figures 3a, 4a and 5a, are placed at each peak ON frame to indicate the regularity and relative frequency of flashing periods. We were also interested in answering the question of whether squids alter their behavior on the basis of other flashing squid in their proximity. Photic waves were plotted against time for multiple squid and their position relative to one another was noted. Graphs of this data for five groups of squid are included in Figures 6-9. In Figures 6-9, the first graph is a simple plot of grayscale intensity against time. Higher grayscale values correspond to a lighter appearance, which occurs when the chromatophores are in the ÖFF state. Figures 6b, 7b, 8b, and 9b are plots of the number of frames between peak ON states, which correspond to the darkest frame in the squid’s cycle, or low grayscale values. They present changes in the individual organisms’ flashing periodicity. Figure 6 is taken from a clip of two squid swimming parallel to one another. For the first two seconds they are nearly perfectly synchronized, and then within 1.5 periods they switch, becoming out of phase with respect to one another. Figure 7 is from a clip of two squid swimming parallel to one another. Remarkably, their flashing frequency shifts in similar ways so that they remain synchronized for eight cycles. Here, as in figure 8, it is difficult to tell whether and to what degree the two individuals adjust to the flashing behavior of their conspecifics. In the clip for Figure 8, one squid is swimming underneath the second, who is on a fishing lure. Drawings along the top of Figure 8 indicate the relative positions of the two squid during the course of the flashing sequence. The dashes below these drawings mark the peak ON periods of the two individuals. The two are flashing synchronously for the first three seconds. After a transition period they begin flashing exactly out of phase, and then after another transition they move back into phase, and then out of phase. Figure 8b includes a plot of the relative timing of each squid’s peak ON state (black line and markers). During the first segment, when the two are synchronized, the difference in peak ON states remains close to zero. During the transition segment the difference jumps rapidly to around half of a period (180 ms). The final “synchronized" segment shows the difference in peak ON states grow larger and then approach zero at the very end. Figure 9 is drawn from a clip of two squid intermittently near and far from each other. As the two squid approach, one is flashing steadily and the other begins to flash. They are not perfectly synchronized, but note the crossing lines plotting periodicity, in Figure 9b. This suggests that the two individuals are adjusting to one another’s tempo. The second squid remains quite steady in its flashing frequency as they swim apart, while the first one ceases to flash. As they approach one another again, the first squid resumes flashing and the second squid begins to modulate its frequency, possibly to synchronize with the first. We estimated the number of chromatophores on the skin of an adult Dosidicus gigas to be 3.9 x 10°. This estimate was taken from a video of a recently sacrificed, medium-sized adult with a skin surface area of about 0.5 m’ including fins, head and arms. Using video shots of squid calibrated with ruler that was placed on its skin, we found that there were approximately 70 chromatophores on an area equivalent to 9 mm’, or 7.8/mm’. An independent estimate based on the same video put chromatophore density at 10/mm’ (Packard, personal communication), yielding a total of 5 x 10° chromatophores. Whereas many cephalopods have several types of chromatophores that differ in pigmentation, D. gigas appears to have just one pigment (reddish brown). Although D. gigas has only one type of chromatophore pigment, two types of chromatophores were distinguished in this video on the basis of size rather than pigment. The large and small chromatophores appear to be grouped into separately controlled chromatomotor fields. The same video segment captured distinct groups of chromatophores that expanded synchronously upon being lightly touched. These patches varied in size between 10 and about 60 chromatophores, with a mean size of 23 chromatophores. These groups of chromatophores probably correspond to what Messenger calls chromatomotor fields (Messenger 2001), an amalgamation of Packard’s terms, motor field’ and ’chromatophore field’ (Packard 1974). Packard induced local expansion with a stimulating electrode, but his chromatophore fields share with the morphological units in the video the quality of having irregular edges that are complementary to the edges of adjacent fields. The touch-induced excitation also seemed to spread somewhat spontaneously to surrounding fields, with flickering lasting up to ten seconds after the initial stimulus. This suggests that chromatomotor field activation may be spread by muscular contraction after it has been initiated by excitation from the CNS. Discussion The graphs of between-peak-ON intervals show that there can be great variability when two squid are flashing in clear view of one another (see Figure 8b), or there can be relative stability (see Figure 7b). The same is true when one squid is flashing on its own (compare Figures 3b and 5b). However, more data must be gathered in order to refine this hypothesis. The study of this informationally rich behavior would be greatly aided by documentation at a higher sampling rate. Specifically, this would help disambiguate variability in periodicity of flashing. Another goal would be to obtain longer videotaped segments of small groups (2-4) of squids swimming or hovering near each other. This is a challenge because the animals are free- swimming; observation in captivity is one possible way of eliminating that problem. Such data sampling would also allow the study of variability in flashing frequency within one individual over longer periods of time, as well as giving us insight into possible correlations between signaling behavior and the individual’s sex. This might give us clues about the adaptive function of Dosidicus flashing behavior. As noted earlier, all the theories put forth to explain firefly flashing behavior are built on the assumption that it promotes mating in some way that makes it adaptive, either for the individual or the group. The reasons behind this assumption illuminate several differences between firefly and Dosidicus flashing. In fireflies, there are clear differences in the flashed signals of the two sexes, and those signals are stereotyped. For a given species, there will be an unchanging flash frequency or patter of flashing. In the videos of Dosidicus used in this study included some squid that never flashed, some that flashed for a few seconds and then stopped, and others that flashed continuously. However, no sex-correlated differences in flashing could be determined because there is little sexual dimorphism in D. gigas, and none that is distinguishable at the distances from which the video was taken. Another important difference between the two types of flashing is that the majority of firefly flashing behaviors involve synchronization of many individuals into either single pulses or wavelike flashes of light. Dosidicus gigas does not appear to have such rigid behaviors, perhaps due to the fact that they are usually in motion rather than resting stationery in trees, as the fireflies are. In short, the flashing patterns of D. gigas are complex, involving intermittent synchronizations between pairs as well as many intermittent individual patterns. These observations are certainly open to revision in light of a larger corpus of data, but at present they suggest that D. gigas uses its binary chromatophore-based signaling in multiple ways and probably for multiple purposes. These could include intraspecific signaling relating to navigation, aggression and mating, as well as defensive patterning directed at potential predators. Acknowledgements Many thanks to Dr. Gilly for your instruction, patience and contagious fascination with Cephalopoda. Thanks also to S. H. Thompson for the use of your computer. Literature Cited Buck, John. 1988. Synchronous Rhythmic Flashing of Fireflies. The Quarterly Review of Biology, September 1988. 265 - 286. Fernández, F., Vásquez, J.A., 1995. La jibia gigante Dosidicus gigas (D’Orbigny, 1835) en Chile: análisis de una pesqueria efimera. Estud. Oceanol. 14, 17-21. Hanlon, R. T. and J. B. Messenger. 1988. Adaptive coloration in young cuttlefish (Sepia officinalis L.): the morphology and development of body patterns and their relation to behaviour. Phil. Trans. of the Royal Society of London. B. Biological Sciences. 437-487 Hanlon, Roger T. and John B. Messenger. 1996. Cephalopod Behavior, Cambridge University Press, Cambridge, U.K. Moynihan, Martin. 1985. Communication and Noncommunication by Cephalopods, Indiana University Press, Bloomington, IN. Nesis, K.N. 1970. Biology of the Peru-Chilean giant squid, Dosidicus gigas. Oceanology 10, 140-152. Nesis, K. N. 1983. Dosidicus gigas. In: Boyle, P.R. (Ed.), Cephalopod Life Cycles. Academic Press, London, pp. 375-383. Nigmatullin, Ch. M., K.N. Nesis, A.I. Arkhipkin. 2001. A review of the biology of the jumbo squid Dosidicus gigas (Cephalopoda: Ommastrephidae). Fisheries Research 54, 9-19 Norman, Mark. 2000. Cephalopods: A World Guide, ConchBooks, Hackenheim, Germany. Packard, A. 1974. Chromatophore fields in the skin of the octopus. J. Physiol, Lond. 238, 38-40P. Tafur, Ricardo, Piero Villegas, Miguel Rabi, Carmen Yamashiro. 2001. Dynamics of maturation, seasonality of reproduction and spawning grounds of the jumbo squid Dosidicus gigas (Cephalopoda: Ommastrephidae) in Peruvian waters. Fisheries Research 54, 33-50. Fig. E Fig. 2 mantle anterior posterior 100- 80 60 - 40— - head 70 3 60 50- £ 40- 44 § 30- fin 90- 80- 70- 60- 50 - V 4 40 — time (s) 9 4 time (s) — — time (s) Fig. 3a 250 - 200- 5 150- 100- Fig. 3b 12- 2 - 0- N time (sec) time (sec) m 3 Fig. 4a 70- 5 60- 50 - 5 40- 30- Fig. 45 10- 2- O- EBE Wnd 10 8 time (sec) 8 time (sec) 10 Fig. 5a 140— 120 - 10- 80 - 60- 40 - Fig. 55 14- 12- 8- 6 - 2- o + E 1 4 2 8 10 12 time (sec) 8 10 time (sec) 12 Fig. 6a 70- 5 60 - 50 - 5 40- 30 Fig. 6b 12 - 10- 8 - 6 - 4 - 2- 0- time (sec) in phase St time (sec) Vvy 3 out of phase 4 Fig. 7a 54 - 52 - 50 - 48- 46- 44 - 42 - 40— 38— Fig. 75 12 - 10- 2- o S time (sec) time (sec) vt 3 3 4 4 Interval between peak ONs: Period (ms) prolines erel 8 O N- — 9. 5 80 N N eeedesses — — ta S 1. (su) eseud :z pinbs sA NO 1 pinbs 10 Sulun 7 Fig. 9a 100- 5 80- 5 60- 40 - Fig. 9b 14 - 12: 10- 8 — 6 - § 4- 2- o A ySWve time (sec) 2 time (sec) 0 VV 8