54 BULLETIN OF THE BUREAU OF FISHERIES. 



manifested in six of the eight pairs of contrasted figures, the June figures (both for the 

 Bay and Sound) being exceptions. 



(4) The average density for Buzzards Bay (1.02314°) is lower than that for Vine- 

 yard Sound (1.02372) and is particularly low at the head of the Bay. This condition 

 is readily understood by reference to the estuaries which discharge into it. 



(5) Certain seasonal differences appear which are, perhaps, of questionable sig- 

 nificance. In Vineyard Sound the density figures for the seasons may be arranged in 

 the following order : 



J un e 1-02339 



March i- 02358 



August 1. 02387 



November 1. 02403 



For Buzzards Bay the figures can not be given for the entire area, since in March 

 only six stations, nearly all of them in the upper half, were visited. Taking the figures 

 for these same six stations for the four months we find the following order to obtain : 



March r .... 1. 02266 



June 1. 02273 



November , 1. 02299 ■ 



August 1. 02327 



The figures for the different seasons were obtained at intervals of about three 

 months and by two different observers. Differences due to 'personal equation" 

 have thus perhaps played a part in the results. And even if that source of error were 

 eliminated, it is quite likely that the figures for the same month in different years would 

 not agree at all exactly. In November, 1908, eight of the determinations of the pre- 

 ceding November were repeated. The average difference between the earlier and later 

 figures was 4 in the fourth decimal place, i. e., a quantity in excess of some of the sea- 

 sonal differences appearing in the foregoing tables. 



In order to compare the results of hydrometer readings with those obtained by titra- 

 tion for chlorine, 17 water samples were subjected to both tests. 6 The chlorine deter- 

 mination in each case was compared with the value, computed from Pettersson's table, 

 for water of the specific gravity recorded. It was found that the actual and the expected 

 values differed on the average by 1.5 per cent. On the assumption that the figures for 

 the titrations were absolutely correct, which is scarcely allowable, this discrepancy 

 implies an average error in the salinometer readings amounting to a little over 3 in 

 the fourth decimal place. We have thus, in any case, some measure of the accuracy 

 of the specific gravity determinations here recorded. As already stated, the fourth 

 figure is not entirely trustworthy. It must be remembered, however, that local dif- 

 ferences have been pointed out within our region equal to about ten times the amount 

 of this average error. 



Seven water samples obtained by us in August, 1909, at points within Nantucket 

 Sound and beyond its eastern end, yielded specific gravities varying only from 1.0237 

 to 1.0239. These figures are close to, but slightly lower than, those found in Vineyard 

 Sound during the same month two years previously. 



a This figure is somewhat too low, since only the upper half of the Bay was represented ia the March series. Here, as 

 stated, the density is particularly low. 



t> These titrations were for the most part made by Dr. W. M. Clark, then a scientific assistant at the Woods Hole 

 laboratory. 



