BIOLOGICAL SURVEY OF WOODS HOLE AND VICINITY. 31 
Referring to the first three heads, it must be stated that the ordinary glacial drift 
of the region, like that which is distributed so widely elsewhere, consists of a mixture, 
in varying proportions, of sand, gravel, and stones. These three terms are not employed 
in the same definite sense as they are, for example, by E. J. Allen (1899). This writer 
restricts the word “‘sand”’ to mixtures of particles the coarsest of which will pass through 
a 1% mm. sieve, the finest passing through a 14 mm. sieve, but not remaining in suspen- 
sion for more than one minute in sea water. Under this main class he recognizes three 
subdivisions. ‘‘Gravel’’ (also subdivided into ‘‘fine,’’ ‘‘medium,’’ and ‘‘coarse’’) com- 
prises aggregations of particles ranging from 1.5 mm. to 15 mm. in diameter. Any 
inorganic material coarser than this was listed by him as ‘‘stones.’’ Our use of these 
terms, though far less precise than Allen’s, we believe to correspond more nearly with 
common usage. In many cases our ‘“‘sand’’ would probably comprise Allen’s finer 
grade of ‘“‘gravel,’’ and our “‘gravel’’ would comprise much which he would term “‘stones.”’ 
Thus stones which were frequently as large as an inch or more in diameter were con- 
sidered as belonging to the “‘gravel.”’ 
The truth is that any such classification is arbitrary, and, unless actual measure- 
ment is employed, as has been done by Petersen and by Allen and Worth, these designa- 
tions must be extremely ambiguous. Moreover, it is very doubtful whether an exact 
classification, such as the foregoing, would be of any service in the case of our local sea 
bottoms, which vary so much, even within the limits of a single dredge haul.¢ As will 
be pointed out later, the nature of the methods employed renders it possible to state 
with only a rough degree of approximation the extent of the correlation between the 
distribution of a given species and the character of the sea floor. 
Another source of difficulty relates to the character of dredge employed at a given 
station. A canvas bag (p. 17) would retain all of the ingredients, and this could be 
washed and sifted and properly described. Such a small bag would frequently fill almost 
immediately, however, and thus fail to represent the entire course of the haul. During the 
earlier portion of the work the sample was commonly collected by an ordinary dredge 
net having a very close mesh at the bottom. It is obvious that if the mixture consisted 
of sand and gravel, much of the former might be lost during the reeling in of the dredge 
line, and that the sample might be listed as merely “‘gravel,’’ whereas sand predominated 
at the outset. On the other hand, a sample in which sand predominated was doubtless 
at first often listed as ‘‘sand”’ in cases where careful washing would have revealed the 
presence of small proportions of gravel or shells. The beam trawl, having no cutting 
edge, and having a net with a wide-meshed bottom, would bring up merely the loose 
stones lying freely upon the surface. Thus the ‘“‘stony’’ bottoms of the earlier records 
may in some cases have included a certain proportion of sand and fine gravel, though 
such cases are probably infrequent, since the beam trawl was commonly not used upon 
bottoms known to be stony. Where no stones appeared in the trawl net it was usually 
assumed, in the absence of data to the contrary, that the bottom was sandy. However, 
as already stated, a small dredge was generally used along with the beam trawl. 
a An idea of the variability in the character of the bottom within comparatively narrow limits will be gained from considering 
the results of some of our supplementary dredgings, in the course of which over 100 of the original stations were repeated with a 
rather rough approach to accuracy. On comparing in each instance the earlier and later record for the same station it was 
found that in only 14 per cent of the cases were identical types of bottom recorded, while in only 33 per cent of the others were 
they substantially identical. In 47 per cent of the cases the ingredients recorded were partly the same, while in 6 per cent they 
were totally different. The later entries were as a rule fuller than the earlier ones, and this fact doubtless accounts for some of 
the differences, but they are likewise largely the result of real differences in the bottom passed over. 
