The hypothesis I put forward provisionally with respect to thecomposite bow, and which Mr. Balfour appears to have adopted, was, that the Esquimaux bow, consisting of separate pieces put together with rivets, strengthened and rendered elastic either by means of numerous strands of sinews tied on at the back, or of sinews formed into a stout cord and bound on upon the convex side, represented the survival of the earliest form of the composite bow, which primeval man, in the absence of a better class of weapon, was compelled to form in order to serve his purpose as a bow. That the bows in which, like those of the Californian Indians and other tribes of the north-west coast of America, the sinew instead of being formed into a strong cord or numerous small strands at the back, is spread over the back in thin layers and glued on to it, represent an improved form which all must have gone through before they developed into the more advanced form of Tatar, Indian, Persian, and Chinese bows, in which the sinews or other elastic materials besides being spread over and glued to the back, are bound up and covered over with bark or some other suitable substance, so as to give it the appearance of a single piece like the plain bow. I also showed that the connection between the Chinese bow and the bow of the Western Esquimaux is rendered certain by the adoption in both, of the curved back straight pieces at the ends, which Mr. Balfour terms "ears," united to the body of the bow, at an angle or elbow, the particular use of which is not very clear, though its influence on the flight of the arrow may be conjectured. The probability of its being a form of some special use is made more likely by the fact of its having been adopted in India with the steel bow, made entirely of one piece of that metal, and consequently not a necessary adjunct of any composite construction of the weapon; unless indeed it was adopted in the steel bow through sheer unreasoning conservatism, like so many survivals in the material arts. I think, however, that this form may have a tendency to draw the bow-string taut in the direction of its length during the release, and thereby possibly to increase the initial velocity of the arrow.

But there is another point connected with the origin of this class of bow into which Mr. Balfour has not entered. He has given detailed descriptions of some varieties in the construction of the composite bow, but he has not said anything about the advantages which the different changes and additions were destined to achieve; yet each variety must either have been intended as an improvement, or must have been introduced through some unknown conditions affecting the craft of the bowyer. If we could get at these we should be in a better position to appreciate the causes of the variations and the spread of the different varieties. In my catalogue I endeavoured to collect a few facts relating to the performances of these two classes of bow in respect of range and accuracy. It is not an easy matter to obtain reliable information on the subject, for the users of the long bow have never been proverbial for the accuracy of their statements concerning it. In the early part of my professional career as Chief Instructor of Musketry, I had considerable experience in the methods of testing the range and accuracy of missile weapons, and I am well aware how much care would be required for such an investigation. Yet the information is not altogether inaccessible, and from what I was able to gather, the composite bow does not appear to be a superior, but, if anything, an inferior weapon to the plain bow, when made of the proper wood and in skilful hands. We know how tenaciously the soldiers of our own country clung to the long bow for sorne time after the first introduction of fire-arms, and how many works were published in praise of it at that time. But this has an important bearing on the origin of the composite bow, which, being of more complex structure, must certainly be of later introduction.