BOTTOM SAMPLES, MURRAY ISLAND, THE BAHAMAS, AND FLORIDA. 253 



weight, however, it was necessary to know the fraction of the original grain 

 that the portion preserved represented. This was estimated and recorded. 



The slides of crushed grains or portions of grains were in the end prepared 

 directly with a rather thin solution of Canada balsam in xylol. In this way 

 slides that were interesting or difficult could be kept for future study. 



With the fine and very fine sand the procedure developed was to lay out, 

 in more or less accurate alinement, an indefinite number of grains taken from 

 all parts of the sample. From one end of this row a suitable number of 

 grains was then taken in the order in which they lay. This was done because 

 it was found difficult to keep track of a definite number of these fine grains 

 taken in advance, because the method was quicker, and because the original 

 weight of the grains taken was too small to be worth determining. 



The ratios of numbers of grains of different groups had to be converted 

 into ratios of weights of the groups because of apparent difference in weight 

 per unit volume; but to arrive at any accurate figures for these relative 

 weights proved very difficult. In one portion the unit weight of mollusk 

 grains would greatly exceed that of madrepores, in others the reverse. The 

 difficulty lay mainly in the fact that the grains were very variable in size and 

 that those identified by their external appearance or by removing a fraction 

 for microscopic study tended to be the larger ones. Consequently their aver- 

 age weight could not be taken as the average weight of all the grains of that 

 group originally present. Under these circumstances, with the small amount 

 of data available, any attempt at a strictly mathematical determination of 

 probable unit weights seemed inapplicable. It seemed best to adjust the 

 weights of the coarser portion on the basis of general considerations derived 

 partly from the portion itself and partly from portions previously studied. 

 In one case the first assumptions led to a total weight for the grains studied 

 which differed only 8 mg. in 383 mg. from the weight originally determined. 

 In another case 4 trials had to be made, adjusting the assumptions in each 

 case, before a sufficient correspondence was obtained. 



For the portions finer than the coarse sand ratios of weight units were 

 calculated from the results arrived at in the previous calculations. The 

 important factors were the relative weights of algae, corals, and mollusks. 

 Since there were no observed unit weights to go by in these finer portions, 

 much greater generalization of the results was necessary. The unit weights 

 of corals and mollusks were therefore taken as equal. The actual adjustment 

 was made in the "fine gravel" of No. 27337. Since large quantities were 

 weighed in that portion, observations were careful and full, and the unit 

 weights obtained seemed better balanced than in any other portion. The 

 unit weight of coral and mollusk was taken at a round figure about the mean 

 of the weights of each group in this portion and all other figures were kept 

 in their original ratio. From these relative unit weights the relative weights 

 of the total number of grains of each group present in a portion were calcu- 

 lated and from these the percentage by weight of each group was obtained. 



