In response to THIIL'S (1975) hypothesis that the food-limited deep sea is a small organism habitat, further data on average size of individnals representing various deep-sea taxa are presented. Our data were gathered with trawls and box corers between 200 and 5000 meters, in the western Noi.th Atlantic. For echinoderms, decapods and macrofauna there appears to be no steady, logarithmic decline i n size with increasing depth, bnt fishes are biggerdeeper.
1 . Introduction
It is of considerable ecological interest to know what effect the logarithmic decline i n biomass with depth has on animal contiiiiinities in the deep ocean. The decline in hionlass is generally thought to be reflected in the size of individual animals, and sniall size in the deep sea i s often referred to (SANDERS, HESSLER 6i HAMPSON 1965, FOUR-KIRR 1972, ROWE & MENZEL 1971, REX 1973). Konetheless. THIEL (19751, npon reviewing the literature, found that very few data concerning size of deep ocean animals were available. In part to stimulate the gathering of such data, he proposed the hypothesis that "which increasing depth and decreasing food concentration sniall organisrns gain importance in total coniniunitg nietabolisni". We offer here our own ohservations on the sizes of animals in a number of deep benthic groups in t h e hope of contributing to the debate opened by THIEL'S interesting ideas.
. Methods
0111.
data come from two sour(vs. Those for the megafauna rome from material collected with 16-ft and 41-ft bottom trawls on the continental slope and rise south of New England. We have madt, ninety-five successful hauls at approximately every 100 m of depth increase troni 100 t o 5000 in depth. For three dominant megafaunal taxaechinoderms, decapods, and fisheswe calciilate averagc weight of an individual in each collection by dividing the total \vet weight of tlie taxoii in question by the total number of individuals in that taxon. Macrofauna were collected from DSliV "Alvin" iising a modified Birge-Ekman box corer, REBC, (KOWE & ('LIBFOIW 1973) \ ith the sample sieved through 0.42 mm mesh. Average weight was determined by dividmg the u e t weight for the whole sample by the total number of individuals in the sample. We feel justified in iising Hct weight as suitable for comparative measurements on the basis of ROWE & n/IEnzE~'s (1971) determination for a series of infaunal samples that wet melght, dry weight, and organic carbon are all linear functions of one another. The validity of this assumption is supported b y ( ' A R E Y et ctl. (1974). Dominant taxa were polyrhaetes, other vermes (nemerteans & sipirncnlans), molliisrs, and crustaceans, but these were not considered separately. The size data for each collecil contribution from tlie \'ood~ Hole Oceanographic. Institution, \Voods Hole, Massachusetts