on the other. When the seed-parent is in class "C" or the pollen parent 

 is in "c" or "d" as described above, failure is to be expected unless the 

 judgment for this classing is inadequate. In general the results of the 

 pollinations indicate that in most cases the varieties are properly 

 classed. In several cases pollen of the "c" class seemed to yield cap- 

 sules as Bridesmaid x Bluebeard (A x c), but as the flowers could not 

 be bagged and the pollination thus controlled the pods in such cases 

 may have resulted from open cross-pollination. 



In several instances pollen of excellent grade (Class a) failed to 

 yield capsules on plants known to be able to produce fruit. In such 

 cases it may be that 



(a) the particular pollen used had lost its viability and had so to 

 speak "spoiled," or 



(b) that the pollination was not made at the proper time, or 



(c) that unfavorable weather conditions may have prevented some 

 process in fertilization or in the development of the capsule, or 



(d) that there are inherent incompatibilities in fertilization. 



The first three conditions just named are merely errors of manipula- 

 tion such as are to be expected in any work in breeding plants. They 

 are revealed and overcome by making a larger number of pollinations. 



A word should be said regarding incompatibility in fertilization. It 

 is now to be recognized that there are two distinct sorts of such incom- 

 patibility. 



First, there is the lack of affinity in fertilization that so often exists 

 between two different species. Thus in irises the crosses cristata x flori- 

 bunda and halophylla x tectorum may continually fail because of too 

 wide a difference between the crested and bearded irises or between the 

 beardless and crested groups. This is an old and well-recognized type 

 of sterility recognized as the inability of two different species to hybri- 

 dize. The studies here reported do not determine the limitations and 

 possibilities of crossing between different species and between different 

 groups of varieties. 



But quite a different type of physiological incompatibility frequent- 

 ly exists in plants that are fully hermaphrodite so that the individual 

 seedling or the entire clonal variety developed from it is "self-fruit- 

 less". Also cross-incompatibility frequently exists between closely re- 

 lated seedlings or clonal varieties. It is this type of sterility that is 

 frequent in the true lilies (Lilium) and in the day-lilies (Hennerocallis), 

 in certain fruit crops as apples, cherries and plums and in .many other 

 groups of plants. 



It is not, however, certain that real incompatibilities of this latter 

 type exist in the irises. Few varieties of irises of the above list are 

 classed as A and a; that is excellent both as a seed and a pollen parent. 

 A glance through the list of table II shows that tn most cases such 

 varieties apparently set seed to proper self-pollination. There are a 

 few exceptions to this; in the tests recorded above cengialti purpurea 

 failed to give capsules to self-pollination of three flowers and for Sta- 

 tellae five flowers failed to selfing. These failures merely suggest self- 

 incompatibility but larger numbers of tests are needed to determine the 

 point. 



The results seem to indicate that of the various groups of irises the 

 "Dutch bulbs" (bulbous Dutch irises) were most fruitful both in selfing 

 and in crossing. While the apogons or beardless irises as a group gave 

 numerous pods to their own pollen their use in crosses with varieties of 



37 



