324 JOURNAL OF THE ROYAL HORTICULTURAL SOCIETY. 



This change, which I could follow through all its stages, was abso- 

 lutely a sudden one. There were no preliminary foreshadowings, no 

 intermediate types. One could easily imagine that the peloric variety 

 would be produced by a gradual increase, either of the number of the 

 spurs, or of the number of five-spurred flowers on the spikes. Such a slow 

 improvement, however, was not observed. The new variety arose suddenly, 

 without any previous indication. But it arose in more than one specimen, 

 and these at once reproduced it from seed. 



Oenothera biennis (fig. 116) and Oenothera biennis cruciata. 

 (fig. 117). — Sudden and repeated appearances of striking varieties have 

 been very often recorded in horticultural literature ; they constitute a 

 clear and simple type, which the florist usually calls " sports." 



In nature the same thing occurs. I shall adduce only one instance, 

 that of the common evening primrose, which, from time to time, pro- 

 duces specimens with very narrow, almost linear petals. Instances of 

 the unexpected appearance of this most curious and constant variety have 

 been observed in Germany and in Holland. They arise from the parent 

 species without previous steps and without intermediate links. 



Dahlia fistulosa (fig. 118). — In the same way, some years ago, a 

 tubular Dahlia arose in my cultures. It has its florets rolled into tubes in 

 the opposite way from that of the Cactus Dahlia, the upper surface being 

 within the tube. It has no horticultural value, but it was interesting to 

 observe its sudden appearance in guarded cultures. 



Chrysanthemum) segetum plenum (fig. 119). — Double flowers and 

 double flower-heads of Composites are among the most common instances 

 of ever-sporting varieties. I have produced such a form from the ordinary 

 corn-marigold under the strict conditions of experimental cultures. It 

 has been possible thereby to follow the process in all its steps from the 

 very beginning. I started from the large-flowered variety of the gardens, 

 which has, on an average, a row of twenty- one florets on each head. There 

 is, of course, a wide range of fluctuating variability in this number, which 

 may go up to twenty-four and down to nine. 



My plan was to select the plants which had the greatest number of 

 rays on the their flower-heads. The effect of this plan could be expected 

 to be twofold. In the first place, the average number of the rays might be 

 gradually increased, and the brightness of the variety might thereby be 

 improved. But, on the other hand, there was a chance of catching the 

 double variety, provided this lay hidden among the plants of my culture. 

 If this were the case, this variety would probably be brought to light, 

 and even in a short time. On the other hand, however, the progress 

 brought about by selection would be but slight and, from a horticultural 

 point of view, an insignificant one. 



Chance favoured me in this case. After the lapse of four years, I 

 succeeded in increasing the number of rays considerably, by selecting 

 specimens with thirty-four, forty-eight, sixty-six and ultimately with 

 about a hundred rays. From these, the next year, the full double variety 

 arose, with over two hundred rays^on many flower-heads. In the course of 

 five years the double type had been fixed. There was, of course, no need 

 for purification from the influence of crosses, since such had been excluded 

 from the very first. But the novelty had betrayed itself through a slight 



