IIAVNKS. I'ANDAMDAl-:. llll'l'OLVTIDAK. AN1)CKAN(;()NIIIAK I.AKVAK 



where females were releasing larvae (for these 4 

 species, females were releasing larvae at depths of 

 about 85, 35, 50, and 100 m, respectively, based on a 

 trawling survey). 



Time of release of pandalid larvae varies with 

 species. In Kachemak Bay in 1972, Stage I larvae of 

 Panda Ins boreal i-x were not caught until the first half 

 of April; Stage I larvae of P. goniurus and P. hyp- 

 sinotus were caught later, in the latter half of April. 

 In British Columbia waters, P. borealis larvae are 

 also released earlier than larvae of either P. goni- 

 urus or P. hypsinotus (Berkeley 1930; Butler 

 1964). 



Time of lar\'al release is also related to water 

 temperature. F^or example, a residual layer of 

 relatively cold (sometimes subzero) water remains on 

 the central West Kamchatkan shelf at a depth of 

 50-150 m throughout the summer. Decapods living in 

 this layer of cold water release their larvae later than 

 decapods living in warmer waters to the north and 

 south. In the western North Atlantic Ocean, pan- 

 dalid shrimp also release their larvae later in colder 

 waters than in warmer waters (Haynes and Wigley 

 1969). 



Depth distributions of larvae of P. borealis and P. 

 goniurus in Kachemak Bay, 1972, were usually 

 similar. Few larvae were in the 0-10 m stratum; most 

 were between about 10 and 40 m. The abundance of 

 larvae remained relatively constant below about 50 

 m. Numbers of Stage I P. borealis larvae, however, 

 increased below about 70 m, possibly reflecting their 

 recent release. These depth distributions differ from 

 the depth distribution of P. jordani off the Oregon 

 coast. Younger (Stages V-X) P. jordani larvae were 

 found closer to the surface (0-10 m stratum) than 

 older (Stages XI-XV) larvae (to 160 m). 



Water temperature has profound effects on larval 

 survival, growth, and size at metamorphosis. For ex- 

 ample, survival of P. jordani larvae (Stages I-III) is 

 markedly less at 17° than at 5°C. For the oldest 

 stage (Stages IX-XIII), the relation between survival 

 and temperature is reversed, and survival is lowest 

 at 5°C (Rothlisberg 1979). For larvae of P. platy- 

 ceros, survival is reduced by sudden changes in 

 temperature, particularly about 20 °C and below 9°C 

 (Wickins 1972). At a given temperature (range 

 5°-14°C), growth increments for all larval stages of 

 P. jordani decrease with increasing size; however, 

 the higher the temperature, the more rapid the 

 molting frequency (Rothlisberg 1979). 



Shrimp larvae can probably influence the direction 

 and extent of their dispersal. For instance, in Kache- 

 mak Bay, 1972 and 1976, pandalid shrimp larvae 

 were released in the central portion of the outer bay. 



Some of these larvae were carried northward out of 

 the bay in the direction of the current, but others 

 were dispersed southwestward in a direction op- 

 posite the current. In the southern area of the 

 western Kamchatka shelf, Crangon larvae released 

 close to shore with larvae of other species, such as 

 king crab, Paralithodes camtschatica, remained close 

 to shore. Larvae of the other species, however, were 

 carried seaward. In the northern area of the western 

 Kamchatka shelf, where currents are faster than in 

 the southern area, Crangon larvae were carried 

 seaward (Makarov 1967). The causes for dispersal of 

 larvae against known water currents are unknown, 

 but dispersal may be dependent, at least in part, on 

 the swimming capability of the larvae. 



Some pandalid shrimp larvae migrate vertically in 

 a diel cycle. In Kachemak Bay in 1972, Stages I and 

 II larvae of P. borealis and P. goniurus were most 

 abundant between the surface and 15 m during low 

 light levels (1800-0800 h); however, during high light 

 levels (1000 and 1600 h), they were most abundant 

 between 30 and 60 m. Although present, a pro- 

 nounced thermocline did not prevent larvae from 

 moving vertically. Whether later stages of P. 

 borealis and P. goniurus migrate similarly is 

 unknown; however, in waters off Oregon, only 

 Stages XII-XVI larvae of P. jordani migrate ver- 

 tically in a diel cycle. During the day, these P. jor- 

 dani larvae are distributed from the surface to 1 50 m 

 by age: the deeper the water, the older the lar\'ae. At 

 night, P. jordani larvae migrate upwards in the 

 water column, and the stages remain somewhat 

 uniformly distributed with depth. 



Foods of pandalid larvae have been determined 

 during attempts to rear the larvae in the laboratory 

 and from examination of shrimp stomachs. Larvae of 

 P. jordani and P. platyceros have been reared on 

 brine shrimp, A rtemia salina nauplii (Modin and Cox 

 1967; Lee 1969; Price and Chew 1972), P. hypsino- 

 tus larvae have been reared on brine shrimp nauplii 

 and algae (Haynes 1976), and P. kessleri larvae have 

 been reared on small pieces of crab, shrimp, and 

 mussel tissue (Kurata 1955). In 1976, I made a 

 preliminary study (unpublished) on foods eaten by 

 pandalid shrimp larvae in Kachemak Bay by examin- 

 ing their gut contents. The larvae mostly ate 

 diatoms, especially Coscinodisais types, and larval 

 crustaceans. Many of the guts also contained black 

 pigment and ommatidia. The assumption that pan- 

 dalid larvae feed on eyes of other decapod larvae was 

 subsequently confirmed when I observed a P. 

 borealis zoea ingesting the eye of a live king crab 

 zoea. Calcareous fragments (probably molluscs), 

 coccolithophores, spines of larval echinoderms, and 



285 



