386 



SCIENCE 



[N. S. Vol. XLII. No. 1081 



ignation of sucli a mating is 58. This will be 

 made clear by repeating the pedigree diagram 

 of bird Nj Diagram II., and adding to it the 

 proper letters and mating designations from 

 Table I. 



The simplicity of the scheme is obvious. No 

 argument appears necessary as to its useful- 

 ness in experimental breeding operations. The 

 writer has found it extremely helpful and 

 clarifying. 



A word should be added in regard to the 

 system by which the numbers have been as- 

 signed to the matings. It might at first sight 

 appear as though the arrangement were an 

 entirely haphazard one. It is not. On the 

 contrary the numbers will be found to con- 

 form to the following general principles, 

 which seem likely to be of aid in practical 

 work, as tending to make it easy to recall from 

 a number just what its particular pedigree 

 looks like. 



1. All even numbers refer to back-cross mat- 

 ings. 



2. All odd numbers refer to co-fraternal or 

 intra-generation matings (not back-crosses). 



3. Matings below 2 are of parental genera- 

 tion individuals: between 2 and 8 inclusive 

 are of F^ individuals ; matings over 10 are of 

 F, individuals. 



4. Even numbers from 10 to 36 inclusive 

 designate back-crosses of F„ individuals with 

 their grandparents, or individuals of the 

 grandparental generation. 



5. Even numbers from 40 up designate 

 back-crosses of F. individuals on Fj individ- 

 uals. 



6. In the case of the odd numbers from 11 

 up it is, in a general way, true that the smaller 

 the designating number of a mating the more 

 closely related to each other are the two indi- 

 viduals entering that mating likely to be. 

 This principle of assigning the numbers could 

 not be so precisely followed as the preceding 

 five, but still is perhaps worth a little. 



In using this system in one's notes or writ- 

 ing it is of course essential to have the basic 

 table always at hand. If the plan should ap- 

 peal to any number of experimental workers 

 it would be a simple matter to have copies of 

 Table I. printed on heavy cardboard to be used 



in breeding houses and pens, in the field and 

 at the desk. 



Raymond Pearl 

 Agricultukal Experiment Station, 

 Orono, Me. 



the chemical composition of bohnite 



Since the analyses of crystallized material 

 from Cornwall by Plattner,^ bornite has gen- 

 erally been considered to be a cuprous sulfo- 

 ferrite, Cu3FeS3(3Cu,S-Fe,S3). In 1903 Har- 

 rington- made a critical study of the published 

 analyses, added several new analyses, and con- 

 cluded that the chemical formula of bornite is 

 Cu,FeS,(5Cu,S-Fe.S3). Recently Kraus and 

 Goldsberry^ made an analysis of crystallized 

 bornite from Bristol, Conn., which gave the 

 formula CujjFe.SjCeCu.S-FejSj), and also con- 

 firmed Harrington's formula Cu^FeS^, of crys- 

 tallized bornite from the same locality. They 

 conclude that bornite is of variable chemical 

 composition, and in order to explain the facts 

 they assume a morphotropic series of minerals 

 ranging from chalcopyrite, CuFeS,, through 

 barnhardtite, Cu^FcSj, and various bornites 

 Cu,Fe,S„, Cu^Fe.S,, Cu,„Fe,S„ Cu, Je,S„ up to 

 CUjijFe^S^, finally ending with chalcocite Cu,S, 

 each member of the series differing from the 

 one below it by the addition of one molecule 

 of Cu,S. 



As a metallographic examination of the tVo 

 analyzed bornites showed no foreign admix- 

 ture, the work of Kraus and Goldsberry fur- 

 nishes, for the first time, proof that bornite is 

 variable in composition. It is believed, how- 

 ever, that there is a more rational explanation 

 of the variability in composition of bornite 

 than the one advanced by Kraus and Golds- 

 berry. 



The recorded analyses of bornite show a 

 copper content varying from 77 to 55 per cent., 

 and an iron content varying from 18 to 6 per 

 cent. In Fig. 1 I have plotted on the triangu- 

 lar coordinate diagram of J. Willard Gibbs the 

 available bornite analyses (59 in number) 

 given in Hintze's " Handbuch " and in the 



1 Fogg. Ann., 47, 351, 1839. 



2 Amer. Jour. Sci., 16, 151, 1903. 

 s Amer. Jour. Sci., 37, 539, 1914. 



