464 G. K. Gilbert—Special Processes of Research. 
equal temperatures on the two thermometers: He would 
thence legitimately infer that the wet bulb temperature is 
never higher than the coincident dry-bulb temperature. 
Diagram No.3 is copied from one prepared by Professor 
Call to illustrate the relations of size exhibited by certain col- 
lections of the shell Helisoma trivolvis.* The ordinates repre- 
sent lengths in millimeters, the abscissas, which have the same 
scale, represent breadths; each circular dot shows the length 
and breadth, and thus 
Cec a stands for, an adult in- 
a co dividual collected from 
coeeree) a small lake in Utah; 
om each square dot  repre- 
Tortr} sents an adult individual 
5 “eel! collected from a late Qua- 
CECCeEEeerererr ternary deposit in Ne- 
treortr| vada. Now itis evident 
Ts ee ‘i -| by inspection that these 
° “7s RS dots do not form even 
Fie. 3.—Length and breadth of Helisoma approximately a straight 
trivolvis. 
line, and we therefore in- 
fer that in this species the length does not bear a constant ratio 
te the breadth. Itis further evident that the area covered by 
the circular dots is quite distinct from that covered by the 
square, and that it is farther from the origin. The modern 
specimens are therefore distinctly larger than the fossil, and 
this difference is, for the two localities in question, of a constant 
nature. Thirdly, each group of dots is horizontally elongate, 
the distal end of the group being slightly—and only slhightly— 
higher than the proximal, and from this we learn that at each 
locality the length is more constant than the breadth, variation 
of size being chiefly, though not entirely, confined to the lateral 
dimension. Now there can be no question that all of these 
results might have been obtained by mathematical methods, 
provided they had been suspected in advance, but without an- 
tecedent hypothesis they would never have been discovered, 
while the graphic method furnishes the information without 
the aid of leading questions. 
This example illustrates also a second and higner office of 
the graphic method, for it essentially invoives more than two 
variable phenomena. It not only shows the relation between 
the length and the breadth of this shell, but the variation of 
this relation with geologic horizon. In the black dots are com- 
pared the length and breadth of the round Helisoma trivolvis, 
in the square dots the corresponding dimensions of the same 
tee 
| 
* On the Quaternary and recent mollusca of the Great Basin; by R. Ellsworth 
Call. Bul. No. 11, U. S. Geol. Survey, Pl. III, Diagram II. 
