1920] AGRICULTURAL BOTANY. 31 



Tlie prodiU'tiim of slhiic is coiiiuh-UhI with the amount of nitrogen lixed, and 

 is inlliu-nccd by tiii' nu'diuni. Tlio naturo of the cell sap also influenct's the 

 capabilities of the bacteria. The form of the nodule depends primarily on the 

 nature of the environment of the host acting through the cell sap. 



Observations on the anatomy of ash-wood with reference to water- 

 conductivity, M. (J. HoLifKs (Ann. Hot. [London], 33 (1019), No. 130, pp. 255- 

 26.'f, fis/s. 7). — The results obtained from a quantitative investigation of the 

 constitution of ash wood In young shoots, with special reference to vessel con- 

 tent, have bet'u discussed in comparison with similar data obtained for hazel 

 wood. It is stated that on the whole there is a fall in absolute water con- 

 ductivity and a rise in specific conductivity from the base of a shoot to its 

 apex in both kinds of wood examined. The figures for specific conductivity 

 are in general lower in ash than in hazel, but the water conducting elements are 

 more numerous in hazel. 



The floras of tFie outlying; islands of New Zealand and their distribution, 

 J. C. Willis (Ann. Hot. [Loiiilon], 33 (IHI!)), No. 131, pp. 267-293, fhjs. 2).— 

 In the present pajter the author deals chiefly with the smaller islands that lie 

 at some distance from New Zealand, but on the same submarine plateau, fol- 

 lowing up the work on the taxonomic distribution of the New Zealand flora 

 given in preceding i)aper.s (E. S. II., 42, p. 724). As previously, a number of 

 suggested predictions are put to the test of actual count and tabulation, and 

 the bearing of the results on the particular and general aspects of the problem 

 are indicated with discussion. 



The method of prediction and verification has, it is claimed, been employed 

 as many as 82 times. It is shown that the floras of these islands must be 

 very old, fnmi their outlying positions. The.se islands have proportionately 

 more families in conmion with Steward Island, whose flora is also old. than 

 with New Zealand proper. 



It is claimed that the age and area hypothesis has now been used successfully 

 in case of 67 predictions. The facts are considered to show that the increase 

 of area witli age is the chief positive factor in determining the distribution 

 of plants, barriers constituting the chief negative factor. 



A'ariation in Hevea brasiliensis, S. Wiiithy (Ann. Bat. [Tjondon], 33 (1919), 

 No. 131, pp. 313-321, fif/. /). — The data here presented with discussion refer to 

 a population of 1,011 trees in their third year of tapping on a normal planta- 

 tion area of seven-year-old Para rubber. They were intended to show the 

 extent to which variation occurs in the amount of rubber yielded by individual 

 trees of H. braniliensis of like age and growing under like conditions, also the 

 extent of possible correlation between yield of rubber and girth of trunk. 

 Ob.servations more or less incidental furnished information on the extent of 

 variation in the rubber content of latex from individual trees and on some 

 other points. 



Strength of latex appeared to be characteristic ami constant for indivichial 

 Irees. Yield, while affected by various cause.s, was c<tmparatively constant 

 when taken to cover a .series of days (6 to 12). Correlation between yield and 

 girth, while definite and positive, is not sufficient in its extent to warrant plac- 

 ing much emphasis on girth in thinning out. 



Figures are given indicative of considerable possibilities in connection with 

 seed selection to improve rubber yields. 



Peculiarities of certain individual trees include a rapid discoloration of the 

 latex, a tendency to rapid coagulation, and a marked cream-straw color in 

 case of a few trees not marked by a high rubber content. 



On double stocks, P. .7. .Taramillo and F. .T. CriiTTKNDKN (Jour. Roi/. Hort. 

 Noc, .'/// (1919), pp. 7^1-82, pis. 2).— The authors have followed up the sugges- 



