STERILITIES OF WILD AND CULTIVATED POTATOES. 7 



media (PI. IV, figs. 1 and 2). Thus, in the best pollen bearers of the cultivated 

 varieties there is much impotence of pollen, and the relative number of grains that 

 germinate is low. Varieties with such pollen, however, have a record of being able to 

 function best as pollen parents. 



Class 2. — In general appearance and dehiscence anthers of class 2 are like those in 

 class 1; the percentage of plump pollen is low (5 per cent to less than 30 per cent), 

 with only a few of these grains germinating. Usually the tubes are of feeble growth, 

 but for some varieties listed here tubes 250 ^ long were observed for a very few grains. 



Green Mountain is fairly typical of this class. The pollen chamber is less spacious 

 than in McCormick (PL I, fig. 6). From the examinations of 30 mounts of pollen, 

 mostly on 15 + 1 naedia it appeared that from 5 to 10 per cent of all grains have granular 

 contents but that less than 1 per cent of all pollen could germinate, and the tubes for 

 all these were short (PI. V, fig. 1). In some cultures of pollen of the Green Mountain 

 no germination was seen. 



The record for crosses with pollen of this class quite uniformly shows poor results. 

 Green Mountain selfed onll5 flowers gave only 2 seed balls; on 51 varieties and seedlings 

 its pollen failed in 44 combinations, and of the 806 flowers involved only 57 set fruit. 

 Yet pollen of Green Mountain on McCormick in one set of crosses gave 27 fruits for 44 

 flowers pollinated, an indi^-idual record that is exceedingly high for pollen of any 

 plant in this class. 



Class 3. — -Anthers usually well developed, occasionally remaining somewhat 

 greenish at apex; dehiscence of nearly all stamens good, but in some varieties poor for 

 some stamens; pollen usually scant, but sometimes fairly abundant; percentage of 

 plump grains low (usually less than 10 per cent and often less than 1 per cent), rarely 

 giving germination, and failing in numerous crosses on varieties known to set fruit 

 readily when viable pollen is used. 



As a class, the plants here listed have poorer pollen than those of class 2. Possibly 

 the making of a large number of germination tests would show some germination of 

 pollen for plants here listed in class 3. The Australian Blue variety is in this class. 

 The anthers are large and well developed and dehisce fully, but the pollen chambers 

 are narrow (PI. I, fig. 5), pollen is scant, and there is seldom a plump grain. The 

 shriveled pollen is aggregated in irregular and rather compact masses, as shown in the 

 illustration. The empty and in some cases shrunken pollen grains of the Irish Cobbler 

 and Triumph, which also belong in this class, are shown in Plate VI, Figures 1 and 2, 

 respectively. 



The records of crosses made by the Bureau of Plant Industry show that only a few 

 of the varieties of this class have been used as pollen parents in crosses with varieties 

 known to be good seed parents. Irish Cobbler pollen occasionally showed a very few 

 pollen tubes in the germination tests, but the use of pollen of this variety in crosses on 

 12 varieties with a total of 156 flowers failed in every case. Early Rose pollen failed 

 completely on 8 varieties involving 83 flowers. Such failures are to be expected from 

 the appearance of pollen of these varieties and its lack of germination on media. 



Class 4- — Anthers mostly greenish; dehiscence irregular, with some or even all an- 

 thers failing to open; pollen very scant, with scarcely a plump grain, or pollen grains 

 lacking. No germination. 



In Rose No. 4, which belongs in this class, the anthers are greenish yellow; the de- 

 hiscence is irregular, with some anthers failing to open; the pollen is scant, with very 

 rarely a plump grain. The Russet Burbank variety has some orange-colored anthers, 

 but many are greenish, tapering at apex, indehiscent, and no good pollen was observed . 

 Berrick and Sensation illustrate well the condition in which anthers are decidedly 

 undersized, greenish, and almost, if not completely, indehiscent. It is obvious that 

 varieties belonging in this class an rarely ever function as pollen parents. 



These observations and the tests for germination give results that 

 fully support the very general view that pollen sterility is a condition 

 which accounts for much of the failure of fruit production in potatoes. 

 In the cultivated varieties and also in certain wild species there is a 

 very general impotence of pollen. 



The varieties grouped in class 1 are the only ones of those studied 

 which can with reason be expected to function at all well as male 

 parents, and even in the best of these a large percentage of the pollen 

 IS impotent. Possibly a few of the plants in class 2 can function oc- 

 casionally as pollen parents, and if pollen germinates more readily on 

 pistils than on artificial media perhaps varieties in class 3 may some- 

 times function as pollen parents under specially favorable conditions. 



