PROPAGATION BY SEED 81 



root and grow. Therefore seed of Algerian Deglet 

 Ntirs can not be expected to produce particularly 

 favorable results. 



The Majhlil date of Tafilalet, on the other hand, 

 seems to give a better average of results, and this is 

 probably because it and varieties similar to it pre- 

 ponderate in certain oases, so that the race has been 

 unknowingly kept purer than usual. The same thing 

 has taken place in the case of Deglet Nur in some of 

 the oases of Southern Tunisia. 



Fardh has given good results as a seedling. It is 

 grown only in one valley, where two-thirds of the 

 trees are of that variety; therefore there is an excellent 

 chance that the male may have been a Fardh seedling. 



Asharasi should give fairly good results, since the 

 most valued male in the districts where it is grown 

 is itself a seedling Asharasi. 



But better still, one may secure from the Bureau 

 of Plant Industry seeds of Deglet Niir which have been 

 inbred, so the probability of their coming true to seed 

 is notably better than those of the ordinary Deglet 

 Nur of commerce. As the breeding process goes on, 

 better seeds will be obtainable every few years, 

 although it will be a quarter of a century before any 

 quantity of seeds of an approximately pure race can 

 be obtained. 



Having secured his seed the grower will sort it, 

 throwing out small ones, and then soak it in water for 

 a week, in order to hasten germination. The easiest 

 way to do this is to place the seeds in a sack or per- 

 forated can and leave it in an irrigating ditch. They 

 may then be planted in any ordinary soil. Spring is, 



