r,2r. w- cnM'KitKNci: i.mi-;ii\\ti(in\i.I': w. cknkthh i;. 



[ilanl |irov('cl il In lie m li\liii(l lirlwccii n si\-n>\v('il iiaricy willi liliwk t-.w-. .-nul 

 llie l.yiH\ 



Many analogoiis cases wcri' round in liic culUiros. ami llic occurrcmc ol' 

 sucli naliiral liybrids lietwocn clemenlarv .spccies, ami cvcn belween Linn(i?an 

 species ol' cnltivalcd barley, sbows how ncarly liiry arc relaled. AU tlio 

 possible coml)inalions of Ihc rospeclivc characlers arc lound. and seifrof^alion 

 occurs acçordini^ lo Mcndcl's law . No case of an inliorilcd innialion lias bccn 

 obs(Mned. 



'■2. Vnri'ihh- rhamrlrr^. 



Nunun'ous expcrimcnls bave bccn carried ont since \Wi . al tlic Kxjicri- 

 nienlal Slalion at Tunis, witb llie objecl of ascerlaining (bc vahie of sclcction 

 of large jifrains and long ears, with regard bolb lo Ibe ininiediale yield. and 

 aiso lo Ibe effccl on Ibe lype. 



The opinion Ibal tbere is an advanlage in sowing lai'ge grains is generally 

 b.eld among European eullivalors, whereas among tiie Arabian cullivalors. a 

 préférence is shown for sowing small grains, for wliicb a liigber pnce is sonic- 

 limcs paid. 



I. (".omprehensi\ e experiinenls wilb bard wlical sbow Ibal Ibe wcigbl ol' 

 Ibe individual grain sown bas no influence on liie weighl of Ibe grains liar- 

 vesled from Ibe same planl. Planls arising from large grain tiller more tban 

 tbose from small grain and il may be supposed Ibal Ibis is a charaelcr depen- 

 ding on nulrilion in Ibe early vegelalive period. Tbere is an inerease in Ibe 

 yield from large grains in proporlion lo Ibe inerease in Ibe lillering. 



Tbese experimenis also sbow Ibal a given weigbl of small grain gives rise 

 lo a larger nnmber of lillers Iban Ibe same \\eigbl of large grain, liie increased 

 lillering in Ibe case of the large grain is more Iban compensaled for by Ibe 

 jaiger num])er of small grains sown. 



II. Influence of Ihe posilion of Ibe grains in Ibe car. 



Experimenis were carried oui wilb eleven lypes of bard wheal, and one 

 lype of winlcr barley, in wbicb Ibe grains from Ibc Iwo exiremilies of the ear, 

 and tbose from Ibe middle of the ear were sown separalely. Tiie resulls show 

 Ibal there is no advanlage in selecling the heavier grains from Ibe middle of 

 Ibc ear. 



III. Iniluencc of liie lenglb of llie ear. Grains from long aiul sborl ears 

 were sown separalely. The ears haivesled were practically idenlicai. 



IV. Influence of cullural conditions. Experimenis wilb wheats, iii wliicii 

 the influence on succeeding générations of cullural conditions was tesicd. 

 show Ibal no marked advanlage is obtained by providing specially favourable 

 culture. 



GENEnvi. CONCLUSIONS. — A pure variety of Cereal must be regarded as a 

 natural lype whicb il is impossible to modify by any dii'ect action. Tbe bere- 

 diliiry ebaracters of such a lype dépend on Ibe anceslry; the variable charac- 

 lers dépend on the conditions of life: sélection of individuals is wilboul effecl. 

 By the sélection of large grains an inerease in the yield may be obtained under 

 certain conditions and wilbin limits which our experimenis bave nol yet llxed. 



The only means by wbich improvemenls can be effected are, l)y bringing 

 lo perfection cullural melhods, in Ibe production of new lypes liy bybridiza- 

 lion, and in Iaking advanlage of forluitious variations, of wbich many owe 

 llicir origin lo nain rai livbfiili/.alion. 



