XXIV.] RUNNING OUT OF POTATOES. 381 



throughout the experiment, but they were constantly 

 shifted over an area of from five to eight acres, so 

 that potatoes were not grown two seasons upon 

 exactly the same ground ; and during the time when 

 these potatoes were decreasing in yield, the garden 

 was each year producing better crops of other kinds, 

 and the newer varieties of potatoes did well. In this 

 case it may be argued that the plants showed signs 

 of wearing out rather than of running out by varia- 

 tion, but there is no evidence to show that the plants 

 were in any way weaker or less able to perpetuate 

 themselves after they had run out than before, for 

 it is probable that seed -production increased as tuber - 

 production decreased ; at all events, we cannot deter- 

 mine if the varieties wore out so long as we have no 

 record of their seed -production. It seems, rather, 

 that the plants returned to a comparatively tuberless 

 condition. Large potato tubers are abnormal, to 

 begin with, and it is not strange if their characters 

 are transitory. (See page 28.) 



At present I see no reason for supposing that fruits 

 propagated by buds run out, to any extent, so long as 

 equal conditions of cultivation and soil fertility exist ; 

 but if the buds are taken from parts which are ab- 

 normally or unusually developed, as they are in the 

 case of the potato, I should expect that we could not 

 long hold the offspring up to their assumed character. 



The conclusion of the whole matter is simply this : 

 Varieties grown from seeds tend to vary or run out, 

 while varieties grown from buds tend to remain per- 

 manent or nearly so, unless the parts which are 

 propagated possess abnormal, or what we might call 



