136 BOARD OF AGRICULTURE. 



progeny ; for we have not yet made the experiments in breed- 

 ing that will enable us to ascertain that with any degree of 

 certainty, because, in all the experiments of which I have been 

 able to find any record, the breeding individuals have been taken 

 as if they had no ancestry — as if no qualifications could be 

 transmitted to the progeny besides those of the mother and of 

 the father. And yet, if we look to this law of ancestral trans- 

 mission, we know that any progeny may show characteristics 

 which are neither those of the mother nor of the father, but 

 those of a remote ancestor, three generations back. 



Now, therefore, we must begin our experiments with refer- 

 ence to the transmission of qualifications from the male or the 

 female, if we would have at all a trustworthy basis. And how 

 shall we proceed ? Here I propose one problem for solution. I 

 have no results to give, gentlemen, and you will at once see 

 how difficult it will be to obtain a result at all ; what extraordi- 

 nary, costly and difficult conditions must be met in order to 

 obtain a result that shall have any value whatever. But I think 

 the time has come when we must stop arguing on a loose basis, 

 when we must begin to riiake experiments that shall have all 

 that scientific accuracy on which we can rely. I am snre that 

 Massachusetts farmers are the men to do this work for the 

 progress of agriculture, for I see from their discussions that 

 whatever tliey do, they do thoroughly ; that whatever operations 

 they enter into they analyze to their satisfaction. Now, if they 

 would ascertain what are the laws of inheritance, or in reference 

 to breeding, let them first secure individuals from which they 

 have eliminated the elements of ancestral transmission. That 

 is the first thing to do ; just as when astronomers compute their 

 observations ; they begin by looking over the observations, in 

 order to know which they are to take into account, and which 

 not. There are observations made by unskilful hands, and if 

 they were taken into consideration, in a computation in which 

 a thousandth part of a minute is an element of great impor- 

 tance, you sfce at once that a single incorrect observation would 

 vitiate all the results of the good ones. 



Now, the first thing an astronomer does when he goes to work 

 on his observations, is, to see how the observations were made, 

 and on looking over the books, he sees at once that here are 

 observations tiuit he must leave out, page after page ; and here 



