STATION FOR I'LANT EKEEDING AT SVALOF, SWEDEN. 817 



iiiid in n. II, tiy. 2, is shown :i siiuill pUit of [u'lis whicli is the pi-oduet 

 of the .sint;lc best plant thrown in a similar plat the year l)efore. 



A» soon a,s one of the plats seems to warrant it, i. c, wlien its 

 nniforniity of type and (general excellence of yield have been demon- 

 strated, a nuni])er of the best ^'ieklinj^- plants, exceptino- of course the 

 sint>le best one, are chosen to plant an increase phit. From the seed 

 obtained from this increase ])lat, ])i-o\'idin<i- the unifoi'inity of type 

 and j;eneral excellence still hold «;ood, the variety test plats are planted. 

 Here is applied the third and tinal test, and all but the one or two 

 very best are eliminated. The variety or varieties that hold out 

 throu^^h all this process of selection are increased for distribution. 

 The method of selection outlined has been in use at Svalof since 1892. 



Althouo'h directly productive of no novelties, the early years of 

 activity were not lost, for the}^ resulted in the discover}- of luuncrous 

 valuable methods and instruments for correctl}- appraising the prac- 

 tical value of certain characters of cereal and fodder plants. They 

 made it ]50ssiblc for the investigator to express, often in statistical 

 form, the value of a variety of grain. 



The amateur plant breeder is often blinded to the real problem at 

 issue b}- the ease with which a host of new forms can be produced l\y 

 cross fertilization. He learns later, as Dr. Nilsson says, that it is 

 easy enough to produce new varieties, ])ut far more difficult to deter- 

 mine which of their characteristics are worth reproducing, or can be 

 reproduced. When Dr. Nilsson hit upon the idea of conducting all 

 experiments from the standpoint of a single plant, he was able with 

 the aid of the experience of former j^ears to correctly and (pnckly 

 decide the value of such new varieties as he observed and selected from 

 his fields. He was able hy the same means to select from each fol- 

 lowing generation the desirable tN'pe for propagation. 



The discover}^ of certain easily observable seed characters on the 

 grains of barley made it possible to group the many varieties of each 

 subspecies into 4 tolerably distinct strains, and to analyze quickly 

 large quantities of seed of a given sort in order to determine its 

 purity as a variety {Sorfenreinhelt). The botanical an-angement of 

 the oat inflorescence, i. e., the number of flowers per spikelet, etc., 

 have made a tolerable classification of the varieties of this plant possi- 

 l)le, and the breeder can detect at a glance in his experiment plats the 

 pi'csence of a valuable hottiiilail variation. 



These purely botanical characters have correlated with them ^•alua- 

 l>le economic ones, and it has been one of the services of the Svalof 

 station to point out some of these correlations and to emphasize th(> 

 importance of this principle. For example, it has been found that 

 the close headed wheats have the strongest straw, and the three-flowered 

 spikelcts of the oat bear the largest grains. 



Too much stress can not l)e laid, according to Dr. Nilsson, upon the 



