again produce three black to one white. This may be shown diagrammatically, the squares representing males, the circles females.

About 1836, black and white Spanish chickens were introduced into England. The cross of these two was a finely mottled gray known as the blue Andalusian, much esteemed by fanciers. To their disgust, this "blue" never bred true, about half of the next generation was "blue," the other half black and white. We now see that they were dealing with a hybrid which of necessity produced the original types in a certain percentage of the

cases.

We may now briefly illustrate what happens when two sets of characters are crossed, and again we will take Mendel's peas, recalling that round is dominant to wrinkled, yellow to green.

This gives, as will be noted, a ratio of 9:3:3:1, or 9 yellow-round, 3 yellow-wrinkled, 3 green-round and 1 green-wrinkled.

Mendel was keen enough to see that the explanation of

these results lay in the nature of the germ-cells not in any outer conditions. He could not give a complete answer to the questions raised, but he did bring out the fact that the factors causing these phenomena were separately heritable. The rediscovery of his work, in connection with the results of the botanists just named gave a great stimulus to the investigation of heredity and such phenomena are now called "Mendelian " in honor of their first discoverer. Curiously enough, we now know that as early as 1820 John Gross of Devonshire, seeking a new variety of pea, had observed the result of crossing different types but had never followed it up nor discovered any law underlying it.

It so happened that this newer evidence supplemented some of the claims made a decade before by one of the greatest of the successors of Darwin, August Weismann. After 1867 Weismann had undertaken to develop a suggestion made by Virchow in his "Cellular Pathology" in 1858. Inasmuch as every individual starts life as a single cell, and inasmuch as these cells all come from those of earlier generations by a process of division, Weismann became convinced that it was to the germ cells, not to the body as a whole, that we must look for the facts of heredity and variation. He therefore came to think of the germ plasm as something independent of the body in which it was housed and which was passed along generation after generation practically unchanged. He thought that within the germ cells there must be "determiners" of some sort for the various parts of the bodies and saw that if this were true that some of them would have to be discarded or else the union of the male and female cells would double the number. This, as we have seen, was soon found to be true.

If we assume then that there is in the chromosomes a determiner for the different characters and do not forget that each individual gets the chromosomes of his germ cells from two sources and that these are segregated in the first "reduction division" we may hazard an explanation of the above phenomena. Take the case of the guinea pigs. The black pig came of pure black ancestry and we say he was duplex or homozygous in so far as that character was concerned. The white pig had a white ancestry and was likewise duplex. Now when these are mated every ovum of the mother carries a determiner for white and every sperm of the father one for black. Hence all the offspring are simplex or heterozygous in so far as color is concerned, and are all black inasmuch as black is dominant over white. When these hybrids are mated however the chances are that out of four possible combinations black will unite with black once, black with white twice, and white with white once. Such a result is based of course on many matings for one case is not decisive. A cent thrown in the air may come down heads up ten times in succession, but out of a large number of throws heads and tails will be about equal. We may represent what takes place by a simple diagram:

SUGGESTED EXPLANATION OF COLOR INHERITANCE IN GUINEA PIGS Inasmuch as the one at the left in the third generation is duplex black it evidently cannot produce white if mated

with another pure black; while the one at the right is duplex white it cannot produce black again if mated with white. Moreover experience has shown us that white is a recessive and will not show if there is a single determiner for black. The two in the middle are simplex and will therefore produce both white and black if mated.

Suppose now a simplex black be mated with a duplex white, what will be the result? The law of averages will give two simplex black and two duplex white. If duplex black be crossed with simplex black there will be two duplex black and two simplex black.

Some five years ago the writer was given a pair of kittens, brother and sister. Both were short-haired, but the donor claimed that some of their ancestors were longhaired. It is known that short-hair is the dominant, long hair the recessive. For two years only short-haired offspring appeared. Then there was a litter containing one long-haired kitten. This was mated with its father and about half of the kittens since born have been long-haired equal so far as hair length is concerned to Angoras of aristocratic lineage. Inasmuch as these short-haired cats are duplex recessives, they can produce only long-haired kittens if bred with each other.

The difficulty in these cases is that the dominant-recessive cross is of the same color as the pure dominant. In the case of the "four o'clocks" bred by Correns the color is different and the nature of the germ cells or "gametes is revealed thereby.

Interesting and significant as these facts are it is impossible at the present time to tell how far-reaching the Mendelian phenomena are or what practical use may be made of them. On these points the biologists are divided, some holding that many of the phenomena of inheritance

are of another order, while others believe that as soon as we know the facts we shall be able to interpret on the Mendelian basis. On this subject Darbishire concludes as follows:

"In the opinion of those who accept Mendel's theory as foreshadowing, if not as, in its present state, actually constituting a valid theory of heredity in general, the number of characters concerned every time a fertilization takes place is certainly very large; it is nothing less than the sum total of the characters of the organism in question. According to this generalized Mendelian theory, the organism is made up of a number of characters which are called unit-characters, because they are transmitted as independent units in inheritance. These unit-characters were, in the early days of Mendelian speculation, considered to be associated together in pairs, but . . . the pair is now regarded as consisting in the presence of a particular character as its dominant member, and the absence of this character as the recessive member. But this is a secondary feature of the theory. The essence of it is that the organism is built up of an obviously immense number of separately transmissible unit-characters, the number, limits, and nature of which can be determined by experimental breeding. With regard to the soundness of the theory, all we know at present is that it applies to the relatively small number of characters which have been dealt with in Mendelian studies. This knowledge is sufficient to justify its application to practical problems, if there is reason to believe that the inheritance of the hereditary characters under consideration is of, or approximates the Mendelian type. But this knowledge is not as yet by any means sufficient to warrant even the hope that the