It is in the search after the true boundaries and characteristics of orders that we may expect the greatest advance by the naturalists of the present day; and yet there is now much discrepancy among them, some mistaking orders for classes, others raising families to the dignity of orders. This want of agreement in their results is not strange, however; for the recognition of orders is indeed exceedingly difficult. If they are, as I have defined them, groups in Nature founded upon a greater or less complication of structure, they must of course form a regular gradation within the limits of their class, since comparative perfection implies comparative rank, and a correct estimate of these degrees of complication requires an intimate and extensive knowledge of structure throughout the class. There would seem to be an arbitrary element here,--that of our individual appreciation of structural character. If one man holds a certain kind of structural characters superior to another, he will establish the rank of the order upon that feature, while some other naturalist, appreciating a different point of the structure more highly, will make that the test character of the group. Let us see whether we can eliminate this arbitrary element in our estimate of these groups, and find any mode of determining orders that shall be unquestionable, and give us results as positive as a chemical analysis according to quantitative elements. I believe that there are such absolute tests of structural relations. It is my conviction, that orders, like all the other groups of the Animal Kingdom, have a positive existence in Nature with definite limits, that no arbitrary element should enter into any part of our classifications, and that we have already the key by which to solve this question about orders.