1432 PI.AXT BREEDING 



mycorhiza while on a hungry arid soil the number may rise to 98 per cent. 

 And the condition of the roots is always reflected in the number of flowers 

 with abortive ovaries. Further, where the nitrogen supply is so limited, 

 the whole vegetative growth of the olive tree suffers. Sandy soils poor in 

 lime and organic matter and liable to long periods of drought are particu- 

 larly favourable to the development of mycorhiza, whereas rich calca- 

 reous soils stimulate the growth of normal roots. 



Experiments were also carried out to determine whether the endoradi- 

 cal mycelium intercepted phosphoric acid. No interference in absorption 

 and transport was observed. 



Olive trees whose flowers are almost all sterile do not regain their 

 fertility even when generously manured and this fact would indicate that 

 the flowering branches, after being submitted to a prolonged course of mal- 

 nutrition, undergo profound physiological modifications which are irre- 

 versible. Further investigations are required to estabUsh whether the 

 change is really complete,, or whether the trees would eventually recover 

 with time and treatment. The nitrogenous manures which are specially 

 useful to stimulate the formation of a normal root system are those directly 

 available to the plant and owners of olive groves should be encouraged 

 to dress their gro^■es periodically with such fertilizers. 



PLANT 1069 - Correlated Characters in Maize Breeding.— Collins, o. x. in journal of A ■licultur- 



BREEDING al Research, Vol. VI., No. 12, pn. 435-153 + tallies XIy\' - L,XIII. Washiiutou, D. C, 



Tixne 19th, 191 6. 



Two principal methods of breeding can be distinguished, depending 

 on the manner in which selection is carried out : 



(i) vSelection may aim at the isolation and propagation of desirable 

 types of individuals. 



(2) vSelection may be -directed towards the variation of individual 

 characters, regarding which improvement is required. 



With the majority of crop plants the method of selection of types 

 has been b}^ far the most productive, but this method has been very little 

 used in the improvement of maize. Selection has been by characters 

 instead of types. 



It has not been clear why the isolation of t3-pes of plants has not been 

 a factor in the improvement of maize. Although the differences in the 

 characters are very clear and striking, few breeders have been able to dis- 

 tinguish well defined types of plants within the commercial varieties. If 

 recognizable types exist, it must mean that groups of characters tend to 

 appear together : in other words, the characters are correlated. The extent 

 to which obvious characters are correlated is therefore proposed as a mea- 

 sure of this tendency towards the persistence of types. The experimental 

 results here reported show that in the progeny of a hybrid between two 

 very different varieties of maize, the characters studied, instead of forming 

 coherent groups, are almost entirely independent in inheritance. 



In attempting to measure the extent to which types persist by means 

 of correlation coefficients, it is necessary' to distinguish different kinds of 

 correlations. For this purpose correlations are here classified as physical, 



