wants increase. A single best variety is an impossibility, the very qualities that make a sort best for one set of conditions, make it one of the poorest of all under other circumstances. Very slight differences between two sorts, that can not be recognized except by an expert, may make it possible to grow one to a good profit under conditions in which the other could only be grown at a loss, though the latter kind would be by far the more satisfactory under other conditions. The maturing of the entire crop of peas so that they can all be picked at once is a most valuable quality where they are grown for canning or for shipping to a distant market, but is a real disadvantage in the home garden. Success in vegetable culture depends as much upon the selection of a variety suited to the prevailing conditions as upon any other factor. The wise planter, therefore, is eager to discover and secure the sort best suited to his needs. And the work of the seedsmen is the supplying of his customer with that sort. He must first know what the customer's conditions are and then know what sort will fit him best. He should not waste time in search of varieties which are simply new, merely novelties, but be ever on the alert to secure and develop sorts which will suit certain conditions better than those now in use. To do this intelligently there must be more accurate knowledge of sorts-one source of such knowledge is the study of different varieties as grown in trial grounds -please note that I say the study of varieties. A trial ground, on matter how extensive and well conducted, which is visited only now and then, and where there is no systematic and careful study and record of every characteristic of the varieties planted, is of little value. I know of trial grounds where the plants are admirably grown, but where there is but very little study and practically no records of the characteristics of the different sorts, the records kept consisting of little more than a brief statement of the relative purity of the different stocks. It may be argued that the record of whether a certain variety of bean has a smooth or rough leaf, the pod a long or short point, is of no practical importance, and that many of the important qualities such as hardiness and flavor can not be determined by inspection. While this is true, it is also true that in all vegetables there is a wonderful correlation of qualities, so that the presence or absence of an easily seen characteristic is often a reliable guide as to the development of some other less discernible but more important qualities. Thus the smooth leaved beans will commonly stand more heat than those where the leaf is of a coarser and more open texture. The character of the point of the pod is an almost certain indication of its stringiness. If we select from a field of Red Valentine beans, one plant bearing full round, fleshy pods of the desired type and another where the pods are flatter and less fleshy, and save the seed of the two, it will be found that the seed from the first plant is more

or less contorted and not nearly so symmetrical and handsome a sample as that from the other and inferior plant. But the closest study of varieties as they appear in any one trial ground is an imperfect guide. The most that can be learned is how the different sorts behave under one set of conditions. We must see varieties grown under different conditions of climate, soil and culture, if we would fully know them. So we should learn either by inspection or report, of their behavior in other trial grounds than our own, and this knowledge should be supplemented by that of results obtained from fields planted and cared for, and the crop marketed by men who do it for their bread and butter, as well as for their cake and pie. In my opinion, this lack of definite knowledge of varieties and of their characteristics is the weak point of American seedsmen, and yet to the production of seed of the highest type, as well as to a wise selection of sorts, such knowledge is essential. But for our ideal, it is not enough to know what is wanted, we must know how to produce it and this can only be done in accordance with natural laws. My understanding of the law of heredity is that the character of every living organism is made up of the sum and combination of the different influences inherited from each of its ancestors back for an indefinite number of generations and the relative power of each of these distinct influences varies, or may vary in each individual and so individuals of the same ancestry may vary. Of six peas from the same pod, the first may develop into a plant like that which produced it: the second into one most like a plant, possibly very different in character, which produced the pollen with which the flower was fertilized: the third, into a plant exactly intermediate in character between these two; the fourth into a plant much like a great grand parent and entirely distinct from the first three; the influence of the producing plant seeming to be entirely overcome by that of the remote ancestor; in the fifth all of the influences may be equally balanced; while the sixth may be precisely like the first. The case is analogous to the indicating of numbers by the use of the digits. The proportionate differences between two numbers, each of ten places, and alike, except that one has 5 and the other 6 in the unit's place, is not great, but if the two digits mentioned occupy the billion's column, there is a vast difference in the resulting numbers. So if the influence of a great great grandparent, which was very different and interior to the plant which produced the seed, is relatively slight, as it usually is, (if it occupies the unit's place) it is not an important factor in determining the characteristics of the plant the seed will develop into, but if, as is sometimes the case, it is the dominating influence, outweighing that of the intermediate parents, (if it occupies the billion's place) it is of the utmost importance. We can only predict with certainty the character of plant a seed will develop into, (the number

1

the digits will express) when each and every one of its ancestors were precisely alike (the using of only one digit), or when the relative power of all the different influences (the order in which the digits are arranged) is precisely the same. Neither of these conditions is ever fully possible with plants, for there never were two plants precisely alike, but we may come as near to it as possible by first forming a conception, to the minutest detail, of the kind of a plant we want and then seeing to it that every ancestor of our seed comes as near to this conception, and hence all are as nearly alike, as possible, always remembering that no matter how slight the variation may be in any single ancestor, it may affect the result, just as the use of a digit even one larger or smaller than the other, may most materially change the resulting number, when they are thrown together. I think that seed growers often fail in this respect. If they find an ear of Crosby corn much larger than the type they have been breeding to, but perfect in every other respect, they put it in stock seed, forgetting that this increased size is almost certain to be correlated with other and less evident differences, and these differences will be likely to disturb that balance of ancestral influences, which, undisturbed, would give us an ideal Crosby. The second method of securing uniformity is to have the influences of the different individuals, no matter how different, always have the same relative powers in the combination, (the placing of the digits in the same order.) This is even more difficult to accomplish fully than the first condition, but here too, we may work towards it, for among the other differences in plants, there is a difference in prepotency-tendency to reproduce themselves. In some plants the influence over the character of the seed is dominant, overcoming that of the other individual ancestors. (They always crowd into the billion's column.) We can, by experiment, ascertain which plants possess this power and only breed from them, thus increasing the probability of the seed being like this parent in spite of tendencies received from other ancestors. This quality is often very marked. Frequently, of two plants, each conforming equally well to a certain type, the seed of one will give plants 75% of which to disturb that balance of ancestral influences, which, undisturbed, would give us an ideal Crosby. The second method of securing uniwill be with difficulty distinguished from the parent while the seed of the other will give plants 75% of which differ materially from the parent and from each other. This selection of prepotent strains of any variety is little practiced but a most important factor in successful seed breeding.

But for an American, it is not enough to know what he wants and how to secure it, he must be able to produce it in abundance and at the least possible cost. Under present commercial conditions it is quite impracticable to raise seed for sale, even in limited quantities

and at high prices, from plants selected one by one. We can not even afford to reject any considerable part of a seed crop. We often hear of seed that has been carefully rogued. Except where there has been a slight mechanical mixture of some very distinctly different sort, rogueing really accomplishes but very little. The most that can be accomplished by it is to prevent the stocks deteriorating; there can be no real improvement. Assuming that a field of peas is sown broadcast and that there are on the average five plants to the square foot, the field would contain 217,800 plants to the acre, the removal of even one per cent of the poorest plants would take out 2,100 plants to the acre, where, in practice, what is called thorough rogueing rarely takes as many as 200 plants to the acre. The removal of onetenth of one per cent of the inferior plants can not materially change the real value of the seed product. Rogueing, the rejecting of the poorest is fighting against the law of heredity, which is merely kicking against the pricks. The better way is not to fight against the law but to use it in selecting and breeding only the best, and it is entirely practicable to do this, even with commercial seed. I can spend half a day in studying the character of a variety of tomato, forming an exact ideal of just what, in every particular, the sort should be, then spend another half day in selecting five plants which come as near to the ideal as possible. Save the seed of each of these separately and the next season, plant say, an acre with these five lots, and when in fruit, spend another half day in ascertaining which one of the lots is most uniformly like the ideal, and from that lot select enough of the best plants to give 5 pounds of seed. This seed will plant 40 acres the next year, which should give three to five thousand pounds of seed, very much more prepotent and sure to grow superior crops than it would be possible to produce by the most careful and thorough rogueing, both of the stock seed and of the seed crop, which rogueing would cost in labor and loss of crop, not less than $5 per acre, or $200 for the 40 acres, while the actual cost of such selecting and breeding, as I have suggested, even if it required the services of an expert costing $10 per day, would not amount to more than $50, or less than 2 cents a pound for the three to five thousand pounds, of the product. I think the saving and growing of seed from individual plants separately is of great importance. Among the hundreds of lots of such plants selected by myself with the greatest care and with such rigid adherence to the type wanted that it was extremely difficult to distinguish one plant from another and often quite impossible to do so, yet in 90% of the cases there was a material difference in the degree to which the product of such plants adhere to the type selected, often the product of one plant which to all appearances was fully equal to the others would prove so variable and inferior as to not only be without value itself but if the seed of the five

plants had been thrown together, this would have ruined the whole lot. The growing of such stock seed must be under the immediate eye of the seedsmen and must needs be grown under his immediate control, but the growing of the seed crop, the production of the commercial seed, can generally be much more profitably done by contract with individual farmers so located as to be most certain of securing a full crop and at the least cost, and if we use such stock seed as I have described and secure farmers of integrity, distance is not a serious objection. We need men who can be relied upon, not only to use land which in quality and preparation will be most likely to give a good return, but so located that our seed crop will be in the least possible danger of contamination from other crops in the vicinity. We do not think it important that the field be rogued, but we do think that it is of the utmost importance that these farmer growers, no matter how far away, should be visited at least once each season, not for the sake of watching them, but for the sake of advice and counsel as to methods of culture and the handling of the crop, so that we may know our own stocks, and how they behave under field conditions and last but not least, that we may come into friendly relations personally with the grower. To my mind there are very few business relations where mutual confidence and respect is of greater importance than between seedsmen and their farmer growers. I do not think it possible for a seedsman to be permanently successful who does not strive to cultivate such relations and I do know that often such relations have been foundations of the most successful seed growing, not only as to the quality of the product, but the obtaining of it at the least possible cost. Perhaps I can make my meaning clearer by illustrating. A seed farmer once said, "I don't know as I make any more money out of these seed beans that I grow for these people than I would out of field beans, but so long as I don't actually lose money, I can afford to put in ten acres each year merely for the sake of the inspector's visits, he never came here yet but what I learned from him something about my beans or some other crop, that was worth more to me than the care of the ground I had in for them."

To sum up my ideal in seed growing:

First. In a thorough knowledge of the sorts and their fitness for certain conditions.

Second. In understanding and using the natural laws by which we can produce seed which is sure to develop into plants of the distinct type desired.

Third. Having the seed grown in such locations and circumstances that it can be produced at the least possible cost.