list very little good is obtained, and then only by accident. This occurs very seldom in Burbank's cultures.
When he wishes to experiment with wild flowers Burbank goes out himself in search of specimens. He carefully compares the different places of growth and investigates the variation in individuals. Many days are thus employed in gathering together one kind in order to find out existing dissimilarities or to see whether they promise anything for future cultivation. Such specimens are then transferred to his experimental grounds, and when established are subjected to crossing.
With crossing or hybridization we usually understand the sexual union of two individuals belonging to different species or varieties. In practical plant breeding, however, it is not sufficient to combine two types, but three, four, and even five or six kinds are thus united, so as to bring out as many desirable qualities as possible in one single variety. It is, of course, impossible to predict what result will be obtained, and it must be left to chance and the future to decide what combinations are the most desirable. Often crossings are made only with the object in view that among all the combinations something good may turn up. In this case the breeder wants to destroy the equilibrium of existing characters, to make the constant forms unstable, and then to select the best out of the many balancing properties. When the parents themselves are variable their offspring will naturally be more so, and the number of differences increases with the number of hybrids experimented upon.
There is also a chance that latent or sleeping characters may be brought to light. From a scientific point of view we know, as yet, nothing about this, but Burbank holds the opinion that in many cases one character prevents another from becoming visible. For instance, in crossing, the first one meets an opponent which has kept it back—as is often the case in the crossing of varieties—and this latent character gets an opportunity of becoming active. We can naturally not detect what dormant qualities are hidden in a plant, and may, therefore, expect all kinds of surprises. The combinations may be desirable, and the hybrids can be propagated immediately, or they may be the reverse and need further crossing before the unfavorable traits are eliminated. Unknown atavistic properties may in this way become evident and may play an important part in the development of future generations.
In other cases the crossings are made with a certain purpose in view. These are the instances from which we learn the most, and which at the same time give the best chances for quick and favorable, results. A certain number is selected of species or varieties, which together contain those characters we want combined in one type; the undesirable properties we try to eliminate. As the crossings result in all kinds of combinations, it is necessary to produce them in
