212 Fretp Musreum or Naturat History — Geotoey, Vot. III. 
The results agree very nearly with those obtained by Merrill* by 
the addition of 99 analyses, the principal difference being a larger 
percentage of Ca O in the present writer’s result. The present 
writer’s method of determining the minor constituents differed from 
that of Merrill in that the present writer divided the totals of these 
constituents by the total number of analyses instead of by the num- 
ber of analyses in which each constituent was reported. It is ev- 
ident that the writer’s method will produce too low a result, but 
the other method may give one too high, since the minor constituents 
may have been lacking in analyses in which they were not reported. 
It may further be suggested by way of discussion of the interesting 
comparison made by Merrill between stony meteorites and the earth’s 
crust, that only the lighter and more siliceous meteorites should be used 
for such a comparison. Stony meteorites having large percentages of 
free metal have too high a specific gravity to be strictly comparable 
with the earth’s crust. Again it should be recognized that the greater 
abundance of certain elements at the surface of the earth may be on 
account of their greater solubility. Thus limestones have grown 
successively more calcic and less magnesian since early times and an 
increase in the amount of soda and potash at the surface might take 
place in the same way. It does not appear that such a process would 
explain the discrepancy in the amount of alumina but it might act to 
increase the amount of silica. That the earth’s crust of earlier times 
was more nearly meteoritic in composition than the present seems to 
be indicated by the great deposits of iron oxide of earlier ages and the 
fact that the early limestones are more magnesian than the modern. 
Adding the analyses of iron meteorites p. 229 to those previously pub- 
lished, and omitting about 60 obviously imperfect ones, 318 analyses are 
obtained from which the average composition of iron meteorites can be 
calculated by summation. This sum is as follows: 
AVERAGE COMPOSITION OF IRON METEORITES 

i vate SME RRA RE A 90.85 
ING ia aerate ap paced Steed 5 2 AMY dese 208 emia Siaies Reel aan nen ee 8.52 
COT aie nied 64S Ds ee tia nen 5 eae J .59 
Pak ooug ed one GAIah Say Jae 2 ie ae ee a17 
iiss sare u ech hacia ae cynsc ds Beery tse ate] ame gee ame eee eo .04 
Go i inne bie s GOB pun de dD A 09 Gee > re a .03 
Ost Pr are CPr  ene neee huey as ace ater oh .02 
CP. ecviand so hue PLAT laos, are tea ee ne or ce OI 
100.23 
* Am. Jour. Sci. 1909, 4. 27; 471. 
