CLARKE: DIEL FEEDING PATTERNS OF MESOPELAGIC FISHES 



the ranges and standard deviations of the data are 

 broad relative to the differences in means, the diel 

 trend in the latter is similar to that observed here 

 for H. proximum, i.e., peak value was reached 

 early in the evening and then dropped to low val- 

 ues and probably zero before the dawn descent. 



Most previous studies have used visual esti- 

 mates of fullness with a scale of 3-5 ranks. Because 

 of the lack of "intercalibration" between inves- 

 tigators, only the rank for "empty" can be com- 

 pared unequivocally, and it is not certain in what 

 manner the ranks might correlate with percent- 

 ages of the fishes' dry bodily weight. Finally the 

 validity or absence of trends and details thereof 

 are questionable because scales of only 0-3 or 0-4 

 are rather insensitive. (Had only visual estimates 

 of fullness been used for the present study, only in 

 a few cases, e.g.,//. proximum or Valenciennellus 

 tripunctulatus, would the diel trends have been 

 obvious.) 



Anderson's ( 1967) data on T. mexicanus indicate 

 a peak in stomach fullness just before sunrise, but 

 his data on degree of digestion indicate that fresh 

 food items were most frequent between sunset and 

 midnight. His data for Bathylagus stilbius (car- 

 diac portion of the stomach only) indicate two 

 separate periods of increasing fullness at night 

 and the sharpest decrease prior to ascent at dusk. 

 This pattern correlates with frequency of less- 

 digested prey items and is very similar to that 

 observed for several myctophids in this study. 



Similar indices were used in the studies of four 

 species of high latitude myctophids: Benthosema 

 glaciale (Gjosaeter 1973) and Stenobrachius 

 leucopsarus, Diaphus theta, and Tarletonbeania 

 crenularis (Tyler and Pearcy 1975). Both studies 

 examined large numbers of specimens from each 

 of a few, very broad time periods. Their data indi- 

 cated highest percentages of full or nearly full 

 stomachs at night and highest percentages of low 

 values during the day. The occurrence of some full 

 stomachs during the day led Gjosaeter to conclude 

 that diel variation in feeding was not great and 

 Tyler and Pearcy to conclude that there was no 

 evidence against diurnal feeding. Both studies 

 noted a higher degree of digestion during the day. 

 These results are, however, consistent with the 

 possibility that like many of the myctophids in the 

 present study, their species descended at dawn 

 with full stomachs and did not evacuate them 

 completely until the dusk ascent. The latter may 

 well have not been detected in these studies due to 

 the broad time periods used. 



Data on myctophids from recent studies by Mer- 

 rett and Roe (1974) and Baird et al. (1975) are 

 consistent with nocturnal feeding but are 

 equivocal to varying degrees due to low numbers 

 of specimens, incomplete diel coverage, or 

 methodology. Both studies based stomach fullness 

 estimates on counts of identifiable prey items. Ap- 

 parently, the presence of a single resistant part, 

 e.g., a Pleuromamma button, was counted the 

 same as an intact, whole individual of the same 

 taxon. Because of this and the likelihood that some 

 prey taxa or parts of prey are digested — and con- 

 comitantly rendered unrecognizable — at different 

 rates (e.g., Pandian 1967; Gannon 1976), such 

 counts seem to be insensitive or possibly biased 

 estimates of gut fullness — especially so when the 

 counts are used to back-calculate dry weight as 

 done by Baird et al. Furthermore, neither study 

 corrected the fullness estimate for fish weight, 

 which (using standard length ranges given by 

 these authors and assuming that weight is 

 roughly porportional to the cube of the length) 

 varied by factors of ca. 7-15 in the myctophids 

 covered by Merrett and Roe and ca. 2.75 in D. 

 taaningi, the species studied by Baird et al. 



Merrett and Roe's data for L. cuprarious indi- 

 cated peak fullness in the middle of the night and a 

 decrease before the dawn descent — a pattern simi- 

 lar to that of//, proximum. Their data (or Lobian- 

 chia dolfleini and A^. valdiviae include no samples 

 between dusk and near dawn, but show fuller 

 stomachs at dawn. Data of Baird et al. for D. 

 taaningi are also similar to that for H . proximuni . 

 The rise in fullness from empty or nearly empty 

 stomachs in the afternoon to fairly high values in 

 early evening is evident and based on 39 and 9 

 specimens, respectively; however, the subsequent 

 decline is based on a single specimen from late 

 night and 4 from just after dawn ( 1 which con- 

 tained a fair amount of food). 



Fewer stomiatoids have been examined 

 elsewhere, but much of the data available is con- 

 sistent with diurnal feeding. Perhaps the most 

 convincing data (because of good diel coverage and 

 numerous specimens) presented by Merrett and 

 Roe (1974) is that for Valenciennellus 

 tripunctulatus, which does not migrate in their 

 study area. The pattern is clearly similar to that 

 observed for the Hawaiian specimens. Hopkins 

 and Baird ( 1977) cited their own unpublished data 

 also indicating diurnal feeding for the same 

 species. Merrett and Roe (1974) interpreted dusk 

 peaks of numbers of items/nonempty stomach as 



509 



