64 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY. 
tate was formed, which was filtered off, and the amount of phosphate 
estimated separately as above. 
In most cases no precipitate, or only the slightest turbidity, was 
noticed, and after the solution was heated, 50 cubic centimeters of the 
standard solution of ammonium oxalate were run in from a calibrated 
burette. This precipitated all the calcium as oxalate, and left an excess 
of oxalate in the solution. The beaker of solution was then cooled to 
the temperature of the air and poured into a calibrated flask of 250 
cubic centimeters’ capacity, and filled with distilled water up to the 
mark in the neck. The solution was next filtered through a dry filter 
paper, and 100 cubic centimeters of the clear filtrate measured out from 
a calibrated burette. To this were added 25 cubic centimeters of dilute 
sulphuric acid; the mixture was then warmed to about 50°C., and 
titrated with standard permanganate solution from a calibrated burette. 
The value of the permanganate had been previously obtained by means 
of pure ferrous ammonium sulphate and its equivalent in the standard 
ammonium oxalate solution, determined by titration. With these data 
the amount of oxalate which had combined with the calcium could be 
estimated, and from that the percentage of calcium carbonate present in 
the rock was obtained. When phosphate or insoluble residue was pres- 
ent in determinable quantity, the amounts were ascertained separately 
by the methods already described, and the percentage quantity added to 
that of the calcium carbonate. The amount of magnesium carbonate in 
the rock was then represented by the difference between this figure and 
one hundred. 
This method was used by Mr. Hart-Smith in making determinations 
of specimens from the Funafuti core. Later, I had the opportunity of 
making a number of analyses from different depths of the same boring. 
With both of us the method has yielded consistent and fairly accurate 
results, and it is this means of analysis which I have mainly employed 
— confirmed here and there by gravimetric determinations — in work- 
ing on the limestones from the islands about to be considered. In order 
to test the method more fully, and with a view to demonstrating to what 
extent the results yielded by it may be relied on, I made a series of de- 
terminations in duplicate, and another series in which gravimetric and 
volumetric analyses of similar specimens were compared. An examina- 
tion of the results showed that the maximum difference in 27 determina- 
tions of calcium carbonate in duplicate from different limestones was .8 
per cent, while the average difference was rather less than .4 per cent. 
Similarly a comparison of gravimetric and volumetric results of calcium 
