Abstract: This paper presents a method of determining the relative abundance of the siphonophore species Apolemia uvaria in relation to depth in the Monterey Bay Canyon. This species of phynosect was found to exist intact only at depths below 250 meters. Further, density increased with depth below this level. In addition, it was found to be more abundant in habitats close to the canyon wall as compared to one mile off the wall. The unique resources of the Monterey Bay Aquarium Research Institute (MBARI) and its remote operated vehicle (ROV) were used to study and collect these organisms. A gut content analysis and description of the general feeding pattern of A. uvaria was also conducted with samples collected at depths of up to 500 meters. It was shown to have a novel feeding strategy in comparison to other siphonophores studied. In particular, its numerous, large gastrozooids allow for flexibility in prey size and percentage of gastrozooids containing prey. Introduction: Quantifving the depth distribution of deep-water siphonophores has proven difficult to accomplish. In particular, phynosects are extre mely fragile and can’t be collected by closing net systems (Pugh 1984). As a result, only depth ranges have been recorded for species of this suborder of siphonophores (Mackey 1987). This paper attempts to quantify the depth distribution of the phynosect A. uvaria. This species of giant apole mid has been observed at lengths of up to thirty meters. It has been previously sighted intact by the submersibles "Alvin" and Johnson-Sea-Link (Mackey 1987) and is often seen in fragments at shallow depths. The abundance relative to habitat was also examined and quantified. The habitats compared were defined by their proximity to the Monterey Bay Canyon wall. In addition to addressing questions of distribution, a gut content analysis was conducted by the collection and dissection of gastrozooids. The data collected on dietary composition was compared to similar work done on other siphonophore species to gain an understanding of general feeding patterns. (Biggs 1977: Purcell, 198 1a, 1981b, 1981c, 1982. 1983: Purcell and Kremer, 1983). It was also compared to limited work done on fragmented colonies of A. uvaria that drifted into surface waters (Purcell, 1981a). The resources of the Monterey Bay Aquarium Research Institute (MBARI) provide a unique opportunity to observe these organisms in situ at depths to four hundred and fifty meters. All dives by the remote operated vehicle (ROV) have been recorded by a Sony Betacam DXC 3000 and stored as time coded videocassettes in the archives. These tapes have been annotated for all sightings of various organisms including A. uvaria. These annotations include the date, time, video frame, depth, latitude, and longitude of the sightings. This infor mation can be recovered and printed from the database. Using the videotapes and database, the number of organisms sighted within certain depth ranges and the time spent by the ROV in each of these depth ranges was determined for a one year period. The density in a specific depth range was defined as the ratio of these two numbers. Hence, the units of density as calculated here are colonies sighted per unit time. This technique assumes that the time spent by the ROV at certain depths is related to the area sampled at that depth relative to other depths. Density by depth was also calculated for two separate habitat types defined by their closeness to the canyon wall. One set of dives is classified as being in the vicinity of the canyon wall; the other set of dives takes place approximately one mile off the canyon wall. By use of the methods above, the relative abundance of A. uvaria was determined by depth and habitat type. In addition, the general feeding pattern and gut contents are also described. Materials and Methods: Distribution The depth distribution of A. uvaria was examined using the archival tapes and database of annotations for ROV dives in the year 3/20/89- 3/20/90. All noted citations for Apolemia in the archives were reviewed in order to deter mine the number of colonies observed and the depth at which observations were made. In order to determine the relative abundance of A. uvaria over various ranges of depth, it was necessary to formulate an accurate measure of the area sampled by the ROV in each depth range. The amount of time spent in each depth range was used to measure the approximate area sampled by the ROV. In order to determine the time spent in each depth range, all depth reports by the ROV's computer with a value falling within the given range were counted over the year examined in this study. On average, these depth reports are recorded at two second intervals (Bruce Gritton, MBARI database manager). Therefore, the following equation was used to determine the time spent in a particular depth range: (of depth reports within depth range) X (2 sec/depth reportl- time in depth 60 sec./min. range (min.) This value was calculated over the following intervals: 0-25 meters, 26-100 m. 101-200 m. 201-250 m. 251-300 m. 301-350 m. 351-400 m and 401-450 m. Above twenty-five meters, most of the time is spent at or near the surface launching or recovering the ROV. The visibility during these operations is low and the sampling time is greatly reduced. Thus, this range has been separated to give a more accurate picture of the top one hundred meters. It is important to note that during the year studied, the ROV did not dive deeper than 450 meters. From the above infor mation, relative density was defined by the ratio of colonies observed in each depth range to time spent in that depth range, as follows: of colonies observed in depth range density (colonies/minute) time spent in each depth range (min.) In addition to examining the general depth distribution of A. uvaria the distribution within two different habitat types in the Monterey Bay Canyon was examined. The two habitats sampled were described as being either in the vicinity of the canyon wall or approximately one mile off the wall. Twelve dives taking place in the year examined were classified as being canyon wall dives; twenty-four dives were water column dives taking place about one mile off the wall. The latitude and longitude of these dive sites are marked in figure 1. For each of these dives, the time spent in the depth ranges of 0-25 meters, 25-100 m. 101-200 m. 201-300 m. 301-400 m, and 401-450 meters was determined in the manner described above. Again, the frequency of Apolemia occurrence in each of the two habitat types was deter mined with the units of colonies observed per unit time. This allows a comparison of relative abundance and depth distribution of Apolemia close to the canyon wall and in a habitat further off the wall. Collection of Samples The phynosect A. uvaria was collected using two different techniques. Three individuals were collected using the ROV suction sampler. This consists of a tube mounted on a poseable arm that uses suction to place samples into separate, sealed plastic cylinders. Using this technique, gastrozooids can be collected intact and preserved later at the surface to stop digestion. The colony itsesf is broken apart by this procedure. Individuals were collected on two separate daytime dives. Those ROV samples collected on 5/1/90 were preserved in 802 ethanol; those collected on 5/7790 were preserved in 102 solution of saturated formalin in seawater. These samples were preserved from three to five hours after collection. All ROV samples were collected at the C-4, C-5 canyon wall dive site as shown in figure 1. The second technique used to obtain gastrozooids involved the use of a sediment trap to fix live colonies in situ Cynthia Pilskaln of MBARI provided these samples. The detailed construction and operation of this apparatus is described by Honjo and Doherty (1987) and the discussion below is intended as an overview. A diagram of the trap is included as figure 2. The trap is suspended in the water column and collects samples passively in its conical portion for a period of two weeks. As the samples are funnelled into the collecting cylinders they are im mediately preserved with a buffered for maldehyde solution. Thus, the colonies swim into the trap and are immediately fixed, halting digestive processes. The exact number of individuals captured by this technique is unknown because only the gastrozooids are recovered intact. The three sets of trap samples were each collected in two week intervals beginning 9/6/89 and ending 10/4/89. The collections took place 100 meters above the canyon floor at a depth of 500 meters. The latitude and longitude of the collection site is shown in figure 1. It is important to note that this is a site near the canyon wall. Gut Content Analysis The contents of the 189 gastrozooids total collected from both techniques were isolated by dissection under 1.7-30 times magnification. The data recorded included the length of the longest side of each gastrozooid, prey description, prey size, and percentage of gastrozooids containing prey. This data is similar to that gathered by others on several other siphonophore species, making comparisons of general feeding patterns convenient (Purcell, 198 1a, 1981b, 198 1c, 1982, 1983; Purcell and Kremer, 1983; Biggs 1977). Results: Distribution The data used to calculate the depth distribution of A. uvaria are shown in Table 1. Abundance appears to increase with depth below 250 meters. Apolemia was not sighted above 250 meters during the year studied. A graphical representation of the relationship of density to depth range is included (Figure 3). Similar data for the canyon wall dives (Table 2) and the dives one mile off the wall (Table 3) are also recorded. The abundance of A. uvaria was found to be significantly higher in the canyon wall habitat. Distribution by depth in both of these two habitats followed the pattern established above, with abundance increasing with depth and no sightings above 250 meters. A comparison between the two habitats of the density by depth range is shown graphically (Figure 4). Gut Content Analysis The data collected on gastrozooid contents are summarized in Table 4. The samples gathered by the sediment trap showed a higher percentage of gastrozooids with prey compared to those collected using the ROV's suction sampler. This may be a consequence of the fact that the ROV samples were preserved hours after being collected. Digestive processes and egestion of contents may have occurred as a result of the delay. The trap specimens all had sergestids as a significant part of their diets. It is important to note that the unidentified digested material was most likely not sergestid soft body parts since examination of the material under 30 times magnification revealed no exoskeletal parts. Gastrozooid length along the longest side varied from 1.95-16.25 millimeters. Both the size and variation of size of the gastrozooids are large in comparison to other siphonophore species (Purcell, 198 1a). Further, the number of gastrozooids per individual was determined by videotape observations to exceed two hundred gastrozooids per colony in some organisms. Among siphonophores, this is considered to be a numerous amount (Purcell, 198 1a). Discussion: The unique resources of MBARI offer an opportunity to study organisms below the epipelagic zone with greater ease than previously available. The videotape archive provides an accessible body of infor mation and observation for many organisms. Combined with the ROV's sampling capabilities, it provides the tools to observe and collect otherwise elusive organisms. In this paper, I have presented a technique by which these tools can be used to quantify distributional questions. Using this technique, it was found that A. uvaria has a depth range starting below 250 meters, with density increasing with depth. The deep end of the depth range is currently unknown because the ROV tether only allows it to reach maximum depths of about 450 meters. This capability will soon be increased significantly to 1000 meters. At that time, it would be useful to examine the relative ab undance at these deeper depths. Twelve other siphonophores have been shown to have distribution tails beginning at about the same depth by Margulis (1980) and extending to deep waters approaching 4000 meters. No estimate of abundance by depth range was calculated in that study. Further, taxonomical and morphological data are not available for comparison given the previously stated difficulties of studying deep water siphonophores. The distribution by habitat type indicated a greater abundance close to the canyon wall as opposed to one mile off the wall. In fact, several intact individuals can often be seen in close proximity to one another in areas near the canyon wall. This correlates with the high percentage of sergestids found during gut content analysis. Visual observation seemed to indicate that sergestids were more abundant in areas closer to the canyon wall. Unfortunately, plankton tows haven't yet been collected and analyzed to confirm this observation. It is important to note that colonies can be observed in very close interaction with the canyon wall. Some colonies were recor ded on videotape being caught up on the rocky substrate and fragmenting. The gut content analysis carried out by Purcell (1981a) was based on similar tail fragments that apparently drifted to surface waters once separated from their nectophores. The study of digestion patterns in A. uvaria sets it apart from other species previously examined. The combination of large gastrozooids in great numbers implies a novel feeding strategy among siphonophores studied. In general, Purcell (198 1a) classified siphonophores into one of two categories: those having small gastrozooids («1 mm) in large number ( 20) and those having large gastrozooids O 1 mm) in smaller numbers («20). Phynosects were generally found to have large gastrozooids in smaller numbers. Further, those species having large numbers of small gastrozooids had a lower percentage of gastrozooids with prey («202) and consumed smaller prey items. A. uvaria had a highly variable percentage of gastrozooids with prey (11-502) and consumed prey of variable size (1.5-14.5 mm). The variation in the percentage of gastrozooids containing prey may simply be a result of the fact that the ROV samples digested and egested some of their gut contents in the time between capture and preservation. The sediment trap samples had a variation from 412 to 502 while the ROV collected samples had only from 112 to 172 of gastrozooids containing prey. Biggs (1977) found that egestion rates in siphonophores could be as rapid as 3-4 hours. However, in the two fragments analyzed by Purcell (1981a), the percentages of gastrozooids containing prey were 152 and 82.62. This variation existed even though both samples were fixed immediately after capture, halting all digestive processes. Unfortunately, the representativeness of those samples is suspect because they were fragments found well above the normal depth range of intact individuals. Another possibility is that the sediment trap samples were preserved at night, while the ROV samples were collected during the day. Several siphonophores have been shown to show diel differences in feeding (Purcell, 1981a). In any case, the physiological characteristics of A. uvaria imply à synthesis of the two feeding strategies described by Purcell. Apolemia combines large gastrozooids with great numbers, allowing more flexibility in its feeding pattern. In particular, large gastrozooids of variable size allow a wider variety of prey to be consumed and large numbers allow many more prey to be grazed. A variability in percentages of gastrozooids containing prey would also imply the ability to be flexible to variations in prey availability. It would again be interesting to take plankton samples at depths where Apolemia is found and determine if there is a high degree of variability in prey abundance. In addition, it would be useful to analyze gut contents from more samples collected at depth and fixed immediately upon capture. If variation in percentages of gastrozooids containing prey was found, this would further support the presence of a flexible feeding strategy. A study of prey size in relation to gastrozooid size would also be helpful in deter mining the role of gastrozooid size in allowing consumption of variable prey. In summary, the resources of MBARI have allowed a more quantitative estimate of the depth distribution of a phynosect than previously possible. Further, it was possible to collect and conduct a gut content analysis of gastrozooids collected from intact organisms at depth instead of fragments that have drifted to the surface. This provides a better estimate of the feeding behavior of the intact colony in its deep-water habitat. 8 — 6: Z 28 . . E 2 6 — X 6o oo 11 — Figure 2. Diagram of Sediment Trap Am 3 9. . ation of Depth Distribution Figure 3. Graphical Rep Colonies Observed per Unit Time at Each Depth Range 0.083 401-450 0.037 351-400 0.0365 301-350 0.0033 251-300 201-250 101-200 26-100 0-25 0.02 004 - 006 0.08 0.10 000 Colonies/Min. 13 Figure 4. Comparison of depth distribution near canyon wall versus one mile off the wall. Density Near Canyon Wall vs. One Mile Off the Wall 05 Tota 401-450 0.217 301-400 03 9.0005 201-300 00 Canyon Wall Dives 101-200 Water Column Dives (1 mile off wall) 26-100 0-25 0.1 92 0.3 0.0 Colonies/Hinute Table 1. Depth distribution data - 3/20/89-3/20/90. Depth Range (m) Depth Reports Time (minutes) Colonies Colonies/Min. 65688 0-25 m 2190 min. 26-100 m 52381 1746 min. 87384 101-200 m 2913 min. 201-250 m 90302 3010 min. 251-300 m 3856 1795 min. 0033 9—- 301-350 m 2// min. 0365 351-400 m 123086 4103 min. 152 037 401-450 m 43893 1463 min. 083 122 Table 2. Canyon wall depth distribution data - 12 dives total. Depth Range (m) Depth Reports Time (minutes) Colonies Colonies/Min. 0-25 m 7387 246 min. 26-100 m 2928 94 min. 101-200 m 239 min. 7161 5799 201-300 m 193 min. 01 301-400 m 42076 1436 min. 038 54 12739 401-450 m 425 min. 217 92 78090 Totals: 2603 min. 148 057 Table 3. Depth distribution data one mile off canyon wall - 24 dives total. Depth Range (m) Depth Reports Time (minutes) Colonies Colonies/Min. 0-25 m 8845 295 min. 6285 26-100 m 210 min. 34479 101-200 m 1149 min. 201-300 m 54443 00055 1815 min. 301-400 m 39263 1309 min. 0084 401-450 m 6092 0049 203 min. Totals: 149407 4980 min. 00261 13 15 Table 4. Dietary information for siphonophore Apole mia sp. Percentages of gastrozooids with prey calculated by (No. of prey/No. of gastrozooids) X 100. ROV-samples collected by suction sampler; Trap- samples collected with sediment trap; n- data unknown. of prey consumed a . . Trap 9/06/89 n 78 6.3 47 1.5-14.5 27 3 70 Trap 9/20/90 n 16 8.0 50 3.0-13.0 25 0 100 Trap 10/4/90 n 34 6.0 41 2.0-10.0 14 0 86 ROV 5/01/90 2 29 6.2 17 1.5-3.9 0 0 100 ROV 5/07/90 1 28 7.9 11 4.6-9.1 0 0 100 References Biggs, D. C. (1977). Field studies of fishing, feeding, and digestion in siphonophores. Marine Behavior and Physiology 4,261-274. Honjo, Susumo and Kenneth W. Doherty (1988). Large aperture time-series sediment traps; design objectives, construction, and application. Deep-Sea Research 35, 133-149. Mackie, G. O., P. R. Pugh, and J. E. Purcell (1987). Siphonophore Biology. Advances in Marine Biology 24, 97-262. Margulis, R. Ya. (1980). On the vertical distribution of siphonophores in the World's Oceans. In" The Theoretical and Practical Importance of the Ceolenterates:, pp. 60-65, Zoological Institude AN SSSR (in Russian). Pugh, P. R. (1984). The diel migrations and distributions within a mesopelagic community in the Northeast Atlantic. 7. Siphonophores. Progress in Oceanography 13, 461-489. Purcell. J. E. (1980). Influence of siphonophore behavior upon their natural diets: evidence for aggressive mimicry. Science N. F. 209, 1045-1047. Purcell, J. E. (198 1a). Dietary composition and diel feeding patterns of epipelagic siphonophores. Marine Biology 65. 83-90. Purcell. J. E. (198 1b). Feeding ecology of Rhizophysa eysenhardti a siphonophore predator of fish larvae. Limnology and Oceanography 26 424-432. Purcell, J. E. (198 1c). Selective predation and caloric consumption by the siphonophore Rosacea cymbiformis in nature. Marine Biology 62, 283-294. Purcell. J. E. (1982). Feeding and growth of the siphonophore Muggiaea atlantica (Cunningham 1893). Journal of Experimental Marine Biology and Ecology 61,39-54. Purcell, J. E. (1983). Digestion rates and assimilation efficiencies of siphonophores fed zooplankton prey. Marine Biology 73, 257-261. Purcell, J. E. and P. Kremer (1983). Feeding and metabolism of the siphonophore Sphaeronectes gracilis Journal of Plankton Research 19 95-106.