280 



SCIENCE. 



[N. S. Vol. IX. No. 217 . 



times. By this means we hoped to have 

 removed the greatest part of the neon, 

 krypton, metargon and xenon. Then we 

 liquefied the argon a fourth time, and as it 

 boiled away collected six samples, each after 

 one-fifth of the whole quantity had evapo- 

 rated. These samples were carefully puri- 

 fied and weighed. The density referred to 

 = 16 and the refractivity to air = 1 are as 

 follows : 



Density. Refractivity. 

 First fraction 19.65 0.962 



Second " 19.95 0.969 



Third " 19.95 



Fourth " 19.91* 



Fifth " 19.97 0.968 



Sixth " 19.95 0.966 



The first fraction possesses, as appears 

 from the table, a lower density and also a 

 lower refractivity. The other fractions 

 vary very little from each other. Since 

 these determinations were made by using 

 only 30 cubic centimeters, we have weighed 

 160 cubic centimeters of the fifth and sixth 

 fractions. The first determined density of 

 the fifth fraction was 19.935, but at a pres- 

 sure of 5 millimeters the spectrum of nitro- 

 gen was easily recognizable in a Pliicker 

 tube. After the gas had been again purified 

 by sparking, until all the nitrogen had been 

 removed, the densitj' was 19.957. In two 

 experiments the fourth fraction of gas gave 

 19.952 and 19.961. We must then accept the 

 true density of argon as not far from 19.96. 

 Independently Lord Eayleigh and I found 

 the density of argon to be 19.94 ; so it is clear 

 that the impurities of neon and the heavier 

 gases have little influence. The somewhat 

 greater density of pure argon arises from 

 the fact that the neon, which is the chief 

 impui-ity present, has been removed ; the in- 

 fluence of the other gases cannot be recog- 

 nized, owing to the insignificant quantities 

 present. In fact, in 15 liters of argon we 

 found no appreciable trace of xenon ; it can 



* Contained nitrogen. 



be prepared only out of large quantities of 

 liquid air. 



I must take this opportunity of thanking 

 you most sincerely for the honor you have 

 done me in inviting me to deliver this ad- 

 dress. It has been said by some scientist 

 that the gi-eatest joy of life lies in dis- 

 covering something which is new. There 

 is, however, another joy almost equally 

 great, that of making known the results of 

 an investigation to one's fellow scientists. 

 This joy, my friends, you have given me to 

 an extreme degree, and for this I express to 

 you my warmest thanks. 



A CASE OF CONVERGENCE* 

 In 1859 Girard (Proc. Acad. Nat. Sc. 

 Phila., p. 62) described a small blind fish, 

 Tt/phlichthys suhterraneus, from Bowling 

 Green, Ky. This species has since been 

 found to be abundant in the subterranean 

 waters east of the Mississippi and south of 

 the Ohio. 



In 1889 Garman (Bull. Mus. Comp. Zool. 

 XVII., No. 6) gave an account of a blind 

 fish from some caves in Missouri. Mr. 

 Garman says: " Compared with specimens 

 from Kentucky and Tennessee, they agree 

 so exactly as to raise the question whether 

 the species was not originated in one of the 

 localities and thence distributed to the 

 others. * * * There is no doubt that the 

 representatives of Typhlichthys suhterraneus 

 in the various caves were derived from 

 a single common ancestral species. The 

 doubts concern only the probability of the 

 existence of three or more lines of develop- 

 ment in as many difierent locations, start- 

 from the same species and leading to such 

 practical identity of result." 



Ably arguing the case from the data on 

 hand Garman came to the conclusion " that 

 these blind fishes originated in a particular 

 locality, and have been and are being dis- 



* Contributions from the Zoological Department of 

 the Indiana University, No. 27. 



