306 JouRNAL OF COMPARATIVE NEUROLOGY. 
intensive. Of still more importance are the facts that the ana- 
tomical data can be directly correlated by physiological experi- 
mentation, and the method is open to experimental control all 
along the line. Our degeneration methods open up _possibili- 
ties here which are incomparably more valuable than the most 
precise anatomical observation. 
And nature has performed for us a series of experiments 
which are in a sense the converse of our degeneration methods. 
The various sensori-motor systems are very unequally developed, 
some animals possessing one ina high state of elaboration, 
some another. If therefore we begin our studies on the visual 
system for instance, with animals such as most birds with very 
highly developed eyes, and then compare with animals with 
vestigeal eyes, it is evident that we have here a means of isolat- 
ing the system for scientific study which has some points of 
superiority over artificial experimental methods. Fortunately 
within the group of fishes, whose brains are all constructed on 
a plan fundamentally similar, we have the most remarkable di- 
versity in the degree of development of the several systems, so 
that this is a favorable starting point for this method, especially 
since the brain is composed almost wholly of the simpler reflex 
mechanisms without the complications which we find in mam- 
mals due to the enormous development of higher associational 
centers in the forebrain. Some fishes have huge eyes, some 
are blind; some have elaborate olfactory apparatus, some very 
slight ; some show a marvelous hypertrophy of the organs of 
taste, or touch, etc. These organs are all open to physiologi- 
cal study and so the functions can be accurately determined. 
Then, having found the cerebral pathways for each system 
where it reaches its maximum development, we can more easily 
trace out the system in other types, and thus arrive ultimately 
at a full knowledge of its evolutionary history. 
All scientific method is both analytic and synthetic. In 
the phyletic type of neurological method, these two processes 
are apt to be far separated and the observed facts may remain 
inert and relatively meaningless, because imperfectly under- 
stood, incoordinated. In our third type of method, on the other 
