BOTTOM SAMPLES, MURRAY ISLAND, THE BAHAMAS, AND FLORIDA. l6l 



shore area Halimeda was relatively more abundant than farther from the 

 shore. The following calculations show the proportion in which the two 

 types of alga? would have to be present in order to make the calculated and 

 directly observed results for MgC0 3 agree. The results are not at all in 

 conflict with what would be expected from field experience with coral reefs. 

 For sample No. 27353a, if the proportions of the two types of alga? were 

 Coralhnaceae 69.5 per cent, Halimeda 30.5 per cent, then there would be 

 in the whole sample, Coralhnaceae 22.8 per cent, Halimeda, 9.8 per cent; 

 and their respective contributions to the whole sample would be: 



Correcting the calculated results by these amounts, we have, for sample 



No. 27353^.- 



The fact that now the calculated amount of CaC0 3 has fallen below the 

 observed amount proves that these revised calculations are not quite ade- 

 quate to account for the discrepancies, though they bring about a closer 

 agreement. 



If in sample No. 27337 the Corallinaceae were 57.05 per cent and Hali- 

 meda 42.95 per cent of the total algae, then there would be in the whole sample 

 Corallinaceae 24.25 per cent, Halimeda 18.25 per cent, and their respective 

 contributions to the whole sample would be: 



For the total of sample No. 27337, we then have: 



Here the error in the calculated amount of CaC0 3 is increased, indicat- 

 ing that there is no consistent error running through the counts and calcula- 

 tions to explain all the discrepancies in chemical composition. 



