l"»-'0) FIELD CROPS. 529 



ceiitly broufjlit to li^'ht In New South Wales, Is Kiven. Tliis plant Is said to be 

 closely allied to strawberry clover (T. frai/ifcntm), but sutlificiitly individual to 

 be classilied as a (listiiict variety. 



Selection in sell-fertilized lines as the basis for corn improvement, D. P. 

 JoNKs (./OH/-. .l//((7-. Soc. Anion., 12 il'J20), No. S, pp. 77-100. pl.s. 5, fif/s. .'}). — 

 In tliis iiajier. a contribution from the Connecticut Experiment Station, tiie 

 author levieus the projcre.ss of uiaize breedinjr, specially discussing the pedigree 

 record .system, tlu- ear-to-row method, significance of inbreeding, results of self- 

 fertilization, and the value of inbred strains. He urges the application of 

 selecti<in in self-fertilized lines to corn breedijig and outlint's a ti'iitative phin 

 of procedure. 



A teosinte-niaize hybrid, G. N. Cor.i.i.xs and .7. II. Kemi'To.\ (Jour. Ayr. lie- 

 search [U. .s'.], /.'' [I!i,i0), Xo. 1, pp. 1-S8, pis. 7, fiys. .«).— This paper, a con- 

 tribution from tln' lUiri'au of Plant Industry, U. S. Department of Agriculture, 

 de.scribes the behavior of a number of the more sharply contrasted characters 

 in the second generation of a hybrid between Florida teosinte and Tom Thumb 

 pop corn in an attempt to determine more definitely the relation of teosinte to 

 the origin of maize. 



Six Fi plants were grown from hybrids secured in a greenhouse by u.'^ing 

 teosinte plants as the female jjarents. From the self-fertilized seed of one 

 of these, 127 Vi plants were produced. Field measurement of characters, in- 

 cluding dates of dowering and size and number of the several organs, were 

 transferred to punched cards, each card representing an individual plant. 

 Practically all calculMtions were made by the use of electric sorting and tabulat- 

 ing machines, the distribution and means being obtained by sorting with respect 

 to each character and counting the cards in each cla.ss with the tabulator. 

 The use of mechanical tabulating machines in calculation of correlations is 

 explained in detail. The data on the characters are presented in tabular form 

 and the fretiuency distribution by numei'ous figures. 



Many of the 33 characters recorded and correlations considered fell into 

 groups, the members of which .seemed to be mutually related, either physically 

 or physiologically. Eight such groups, including height, nodes above, tassel, 

 male branch, alicole, nodes silking, prophyllary, and number of rows are 

 recognized, comprising in all 26 characters. Among the 7 remaining characters 

 considered as indepi'iident, physiological relations, if existing, were more ob- 

 scure. 



The observations made can be summarized as follows: The F: plants, like 

 the Fi plants, showed characters largely intermediate between tho.se of the 

 parents. The distribution of 33 of the recorded differentiating characters with 

 one or two exceptions showed little or no evidence of alternative or Mendelian 

 Inheritance. With re.spect to individual characters the extreme variants ap- 

 proached, and sometimes exceedetl, those of the parents, but none possessed 

 any large number of characters of either maize or teo.sinte. 



The greatest freedom of recombination was shown, all combinations of char- 

 acters appearing that might lie expected from so few individuals. The many 

 In.stances of coherence or partial couiding were nearly erpuiled by instances 

 where characters derived from ditTerent parents showed a tendency to combine 

 more frequently than would be expected as the result of chance. However, 

 it was found impossible to distinguish primary from secondary correlations. 



Although no incompatible combinations appeared, there were no completely 

 Independent characters. Every character I'ecorded showed significant correla- 

 tion with one or two others, which in turn were correlated with still others, 

 resulting in the interrelation of all characters and the formation of a single 



