476 
BULLETIN OF THE BUREAU OF FISHERIES 
larger number on the outside surfaces is probably due to roughness as well as to greater 
area. The two series average 34.6 percent inside and 65.4 percent outside. In 
calculating the number of spat caught in a bag the counts made on inside surfaces 
are considered as 35 percent of the total. This proportion would not be correct if 
the shells were spread directly upon the grounds, for most of them, because of curva- 
ture, fall with the outer surfaces down. 
Table 20. — Number of spat caught on inside and outside surfaces of shells in wire bags 
Records were kept on the basis of a standard-size bag of shells, and at the time 
of counting the shells were carefully measured in a box of definite dimensions. Although 
some bags were fuller than others, the remeasurement eliminated any error from 
this source. It is true, of course, that it is impossible to obtain exactly uniform shells, 
and that at times they might be relatively smaller or larger than the average. It was 
noted, however, that this had little to do with the results. Generally, a standard bag 
contained 125 to 150 shells, though some held as many as 200 or as few as 100. When 
the shells are small they provide more surface per unit of volume, but impede the free 
circulation of water. When extra large, the water flows freely among them, though 
the area of surface is not so great. These factors appear to offset one another, in the 
case of bags of small diameter such as were used, and the resultant averages are 
relatively consistent. 
To standardize the system of counting and calculating the total number of spat 
caught by this method, 100 unselected shells from each bag were examined carefully, 
on their inside surfaces, with a binocular miscroscope, and every spat counted. To 
avoid error, circles were drawn around the spat as they were counted, for in some 
cases as many as 600 or 700 spat were found on the inside surface of a shell. The 
number of spat on the inside surfaces of 100 shells was used as a basis for calculating 
the number on both surfaces of all shells as described above. The variation between 
the different shells, with respect to the number of spat caught on the inside surfaces, 
was tremendous. In one typical case the extremes were 0 and 730. This was obvi- 
ously due to differences in the angles at which the shells are held as well as to their size 
and their position in the bag. On the shells in this particular bag, which was selected for 
