319 



tions were made from these forms and a furtlier study will be made of 

 them both for hardiness and as to fitness for medicinal use. 



Preliminary work has been done on seed selection with several forms 

 of Digitalis. Seed tests of light and hea\'y seeds obtained by means of 

 tlie api>;u";itus previously mentioned have given striking results in the 

 early stages of seedling growth as shown by the accompanying photo- 

 graphs. The following data will indicate the accuracy of this method of 

 seed selection and the uniformity in the seeds separated. It also demon- 

 sti-ates the practical value of the method if applied to the commercial 

 production of digitalis leaves. Seeds collected from the foregoing plants 

 of Digitalis purpurea L. were separated into light, medium and heavy. 

 The extreme smalluess of these seeds made it necessary to use tive hun- 

 dred seeds from each separation for weighings. Seeds of Digitalis grandi- 

 tiora Lam. obtained from Henry A. Dreer of I'hiladelphia wei'e also sep- 

 arated, weighed and tested. These were hetivy enough to be weighed in 

 one hundred lots and were of such uniformity that they were only sep- 

 arated into light and heavy portions. 



The following tal)le includes the results of these separations and 

 weighings : 



DiyitaUs purpurea. 



Light. Medium. Heavy. 



500 seeds 500 seeds 500 seeds 



0.0217 gm. 0.0253 gm. 0.0341 gm. 



Diyifalisi grdiuVipora. 



Light. Heavy. 



100 seeds 0.0168 gm. 100 seeds 0.0215 gm. 



100 seeds 0.0167 gm. 100 seeds 0.0223 gm. 



100 seeds 0.0161 gm. 100 seeds 0.0223 gm. 



100 seeds 0.0161 gm 100 seeds 0.0215 gm. 



100 seeds 0.0164 gm. 100 seeds 0.0217 gm. 



Total 0. 0S21 gm. Total 0. 1003 gm. 



In conclusion it is only necessary to say that the application of these 

 methods of breeding and the possibilities in drug plant improvement herein 

 suggested should be extended until they include such valuable forms £^s 



