190 The American Geologist. March, 1897 
died away here, but of that we are never positively informed. 
Wlien a species disappears at one horizon it may return. 
The migration of species in this area during the Galena 
(Trenton) and Maqiioketa (Hudson) periods was gradual 
when the fauna is taken as a whole into consideration, and 
there are many minor changes as illustrated in the groups of 
species already described, which make the division of the 
series into beds easy, and the delimitation of formations 
practicable. There are morever local peculiarities of all de- 
grees besides those that are regional. A priori, one is not 
able to tell which species is of the greatest value for correla- 
tion but investigation soon proves some to be better than 
others. Very abundant species and large series of very 
closely related ones are of great importance because their re- 
lation to the strata is most easily proved. Isolated species 
like Bellerophon wisconsensi.'i Whitf. (bed No. 2) which char- 
acterize one zone, and other common ones that are absent for 
a short time over the whole area like JJ. hilohatus Sow. are im- 
portant guides that are not to be neglected in correlation. Of 
least value in the Galena and Maquoketa series is the common 
method of correlation by computing the percentage of like 
fossils since the identity of fossils does not prove strictly con- 
temporeneous deposits, but only related faunas. In north- 
eastern Iowa, as I have incidentally shown, paleontologists 
and geologists have recently mistaken the true Maquoketa 
(beds No. 11, 12), in numerous large exposures, for "Trenton" 
(beds 5, 4, etc.), and the next above the Maquoketa formation 
(13)for Galena (9, 8, etc.), and in fact a percentage computa- 
tion of faunas might lead to the same mistake. Collections 
from different beds in two diiferent localities may have a higher 
per cent, of like species than collections from equivalent de- 
posits of the same bed at the two localities. The paleontol- 
ogic evidence, therefore, that I have used in correlating in the 
Galena and Maquoketa series has not simply consisted of 
gross percentage computations, but upon the inductively de- 
termined limits, range and distribution of species, and in 
fact of the commonest species. 
