L. Agassiz on Animals in Geological Times. 321 
looked by later writers. Who does not see what amount of error 
may accrue from the indiscriminate use of materials which are 
not first submitted to a very critical revision in these different re- 
spects, not to speak of the general difficulty of agreeing upon the 
limits of specific differences. With regard to this last point, 
however, I would say that whosoever would only use in discuss- 
ing general questions materials revised candidly with the same 
principles, could not fail to obtain at least uniform results. And 
when the results of investigations made upon materials corrected 
in different ways by different authors are compared with one an- 
other, if these differences are kept in view, the disagreement in 
the results would not be found so great as it might otherwise 
2m. The astronomers and physicists have long learned to cor- 
rect their observations before using them, and to take into consid- 
eration what they call the personal equation of different observ- 
ers; are we never to learn from them a lesson in the estimation 
of our respective investigations, and shall our facts for ever be 
used without being first ‘corrected for all the possible canses of 
error and disagreement? As long as there are differences of opin- 
ion respecting the natural limits “of genera and species, is it not 
absolutely necessary to reduce or expand the scale applied to the 
investigations of different authors, when using them for the same 
purposes, exactly in the same manner as thermometric observa- 
tions made with the scales of Reaumur or Celsius or Fahrenheit 
are reduced to the same standard, before being compared. 
In the second place, species must be referred to genera circum 
scribed within the same limits, before they can fairly be compared 
or at least lead to trustworthy general results. As long as certain 
bivalve shells of the carboniferous and oolitic series were referred 
oe e genus Unio, it could appear that the family of Natades 
an its existence at a very early period ; but since the oolitic 
Species of this kind have been na neblaneibis o differ essentially 
fom our freshwater shells, and to constitute a themseles asnat- 
ural genus more closely allied to Crassatella than to Unio, nobody 
thinks any longer of looking for Unios in marine deposits. ge 
these genera are the types could be supposed to have extended 
their tange far beyond the tertiary formations, a which how- 
ever no one of their representatives is to be found. Before the 
Sfiatangoids < divided into natural genera, ine genus Spatan- 
WS was mentioned among the fossils “al the oolitic as well as 
the “it Sai ‘a tertiary formations; now it is restricted to - 
among the fossils and found also among the living. I 
ieve that a single genuine species of Gorgonia i is found enemy 
the fossil Polypi, and yet that genus appears in the lists of ete 
ftom the paleozoic period to the present time. 
