number, and resembling each other in texture. And in the arrangement of the feather-tracts the humming-birds approach more nearly to the swifts than they do to any other birds; and altogether differ from the sun-birds, which, in this respect as in so many others, resemble the honey-suckers of Australia and other true passerine birds.

Having this clue to their affinities, we shall find other peculiarities common to these two groups, the swifts and the humming-birds. They have both ten tail-feathers, while the sun-birds have twelve. They have both only sixteen true quill-feathers, and they are the only birds which have so small a number. The humming-birds are remarkable for having, in almost all the species, the first quill the longest of all, the only other birds resembling them in this respect being a few species of swifts; and, lastly, in both groups the plumage is remarkably compact and closely pressed to the body. Yet, with all these points of agreement, we find an extreme diversity in the bills and tongues of the two groups. The swifts have a short, broad, flat bill, with a flat horny-tipped tongue of the usual character; while the humming-birds have a very long, narrow, almost cylindrical bill, containing a tubular and highly extensible tongue. The essential point however is, that whereas hardly any of the other characters we have adduced are adaptive, or strictly correlated with habits and economy, this character is pre-eminently so; for the swifts are pure aërial insect-hunters, and their short, broad bills, and wide gape, are essential to their mode of life. The humming-birds, on the other hand, are floral insect-hunters, and for this purpose their peculiarly long bills and extensile tongues are especially adapted; while they are at the same time honey-suckers, and for this purpose have acquired the tubular tongue. The formation of such a tubular tongue out of one of the ordinary kind is easily conceivable, as it only requires to be lengthened, and the two laminae of which it is composed curled in at the sides; and these changes it probably goes through in the young birds. When on the Amazon I once had a nest brought me containing two little unfledged humming-birds, apparently not long hatched. Their beaks were not at all like those of their parents, but short, triangular, and broad at the base, just the form of the beak of a swallow or swift slightly lengthened. Thinking (erroneously) that the young birds were fed by their parents on honey, I tried to feed them with a syrup made of honey and water, but though they kept their mouths constantly open as if ravenously hungry, they would not swallow the liquid, but threw it out again and sometimes nearly choked themselves in the effort. At length I caught some minute flies, and on dropping one of these into the open mouth it instantly closed, the fly was gulped down, and the mouth opened again for more; and each took in this way fifteen

or twenty little flies in succession before it was satisfied. They lived thus three or four days, but required more constant care than I could give them. These little birds were in the "swift" stage; they were pure insect-eaters, with a bill and mouth adapted for insect-eating only. At that time I was not aware of the importance of the observation of the tongue, but as the bill was so short and the tubular tongue not required, there can be little doubt that the organ was, at that early stage of growth, short and flat, as it is in the birds most nearly allied to them.

In respect of all the essential and deep-seated points of structure, which have been shown to offer such remarkable similarities between the swifts and the humming-birds, the sun-birds of the Eastern hemisphere differ totally from the latter, while they agree with the passerine birds generally, or more particularly with the creepers and honeysuckers. They have a deeply-notched sternum; they have twelve tail-feathers in place of ten; they have nineteen quills in place of sixteen; and the first quill, instead of being the longest, is the very shortest of all; while the wings are short and round, instead of being excessively long and pointed. Their plumage is arranged differently; and their feet are long and strong, instead of being excessively short and weak. There remain only the superficial characters of small size and brilliant metallic colours to assimilate them with the humming-birds, and one structural feature-a tubular and somewhat extensile tongue. This however is a strictly adaptive character, the sun-birds feeding on small insects and the nectar of flowers, just as do the humming-birds; and it is a remarkable instance of a highly peculiar modification of an organ occurring independently in two widely-separated groups. In the sun-birds the hyoid or tonguemuscles do not extend so completely over the head as they do in the humming-birds, so that the tongue is less extensible; but it is constructed in exactly the same way by the inrolling of the two lamina of which it is composed. The tubular tongue of the sun-birds is a special adaptive modification acquired within the family itself, and not inherited from a remote ancestral form. This is shown by the amount of variation this organ exhibits in different members of what is undoubtedly one family. It is most highly developed in the Arachnotheræ, or spider-hunters, of Asia, which are sun-birds without any metallic or other brilliant colouring. These have the longest bills and tongues, and the most developed hyoid muscles; they hunt much about the blossoms of palm-trees, and may frequently be seen probing the flowers while fluttering clumsily in the air, just as if they had seen and attempted to imitate the aerial gambols of the American humming-birds. The true metallic sunbirds generally cling about the flowers with their strong feet; and they feed chiefly on minute hard insects, as do many humming-birds.

There is, however, one species (Chalcoparia phoenicotis) always classed as a sun-bird, which differs entirely from the rest of the species in having the tongue flat, horny, and forked at the tip; and its food seems to differ correspondingly, for small caterpillars were found in its stomach. More remotely allied, but yet belonging to the same family, are the little flower-peckers of the genus Diceum, which have a short bill and a tongue twice split at the end; and these feed on small fruits, and perhaps on buds and on the pollen of flowers. The little white-eyes (Zosterops), which are probably allied to the last, eat soft fruits and minute insects. We have here a whole group of birds, considerably varied in external form, yet undoubtedly closely allied to each other, one division of which is specially adapted to feed on the juices secreted by flowers and the minute insects that harbour in them; and these alone have a lengthened bill and double tubular tongue, just as in the humming-birds. We can hardly have a more striking example of the necessity of discriminating between adaptive and purely structural characters. The same adaptive character may coexist in two groups which have a similar mode of life, without indicating any affinity between them, because it may have been acquired by each independently to enable it to fill a similar place in nature. In such cases it is found to be an almost isolated character, connecting apparently two groups which otherwise differ radically. Non-adaptive, or purely structural characters, on the other hand, are such as have, probably, been transmitted from a remote ancestor, and thus indicate fundamental peculiarities of growth and development. The changes of structure rendered necessary by modifications of the habits or instincts of the different species have been made, to a great extent, independently of such characters, and as several of these may always be found in the same animal, their value becomes cumulative. We thus arrive at the seeming paradox, that the less of direct use is apparent in any peculiarity of structure, the greater is its value in indicating true, though perhaps remote, affinities; while any peculiarity of an organ which seems essential to its possessor's well-being is often of very little value in indicating affinity for other creatures.

This somewhat technical discussion will, it is hoped, enable the general reader to understand some of the more important principles of the modern or natural classification of animals, as distinguished from the artificial system which long prevailed. It will also afford him an easily remembered example of those principles, in the radical distinctness of two families of birds often confounded together,—the sun-birds of the Eastern Hemisphere and the humming-birds of America; and in the interesting fact that the latter are essentially swifts-profoundly modified, it is true, for an aërial and flowerhaunting existence, but still bearing in many important peculiarities of structure the unmistakable evidences of a common origin. ALFRED R. WALLACE.

DR. NEWMAN'S THEORY OF BELIEF.1

(Conclusion.)

THE historical method, as understood by Dr. Newman, would test the value of a creed by its fruitfulness, coherence, persistence, and power of assimilating congenial and rejecting alien matter, or, in a word, by its vitality. Such a method has two remarkable consequences. In the first place, it tends to set aside the direct and obvious tests of the old-fashioned apologists. We need not ask with the philosophers whether the creed gives a worthy or intelligible conception of the universe; for such inquiries only lead into the endless labyrinth of metaphysical argumentations. We need not inquire with the critics into the evidence for particular historical statements, for the facts are intelligible only as part of a vast and complex evolution, which must be appreciated as a whole before it can be understood in detail. And, in the second place, the method lays particular stress upon the process by which ideas "percolate" (as Dr. Newman somewhere says) by other than directly logical means. The dogmas of the creed are not revealed in full scholastic precision and nicety of definition. They are not reasoned out like mathematical propositions by direct demonstration. The germs are

planted by revelation; they grow spontaneously in the minds of believers, obeying a law which is not consciously apprehended, but which may be afterwards elicited, and which becomes more manifest as the process is developed. Once seized it may be stated as a logical formula; but during the earlier period it is in the state of implicit logic-an informing and animating principle, not a recognised and avowed law of belief.

Some kind of logical organon is required, as I tried to point out in my previous article, in order to extract from this theory an available logical test. The truth of a theory must be the ultimate reason for believing it; and the question is, briefly, how from the vitality of a creed are we to infer its truth? An answer is attempted in the Grammar of Assent; and the theory expounded in that book harmonizes throughout with that which is implied in the doctrine of development. The method of classification adopted is the same in both cases. Creeds, according to the historical theory, are measurable according to their degrees of vitality; and so the Grammar of Assent opens with an elaborate scale of assents or beliefs, varying from the faintest to the most vivid, and from the most abstract to the most concrete. Beliefs, that

(1) Concluded from the Fortnightly Review for November.

is, are classified by their fitness to form part of a vigorous creed. The faculty, whose existence is postulated in the doctrine of development, that by which the mind draws remote inferences without a conscious syllogistic process, is now carefully analyzed, and receives the name of the Illative Sense. And, finally, we are again struck by the absence of the direct logical method. A Grammar of Assent, one would say, ought to correspond to a treatise on logic. We ought to assent to true propositions, and therefore should begin by inquiring what is the test of truth. But the very name of the treatise seems designedly calculated to set aside such inquiries, and contemplates at least the possibility of a divorce between the faculty of believing and the faculty of perceiving the truth. The method, as we shall see, is calculated—whether designedly or not-to evade the purely logical question. Indeed, Dr. Newman lays it down as a principle that "in no class of concrete reasonings is there any ultimate test of truth and error in our reasonings besides the trustworthiness of the illative sense that gives them its sanction." 1 Our duty is to cultivate that faculty, and then trust implicitly to its decisions.

· ...

The meaning of this will appear as we proceed; but it is important to notice at once the precise nature of Dr. Newman's problem. He is investigating, one may say, the physiology of belief in the individual as he before considered the physiology of religious faith in a society. He looks upon belief from outside, as a phenomenon which is to be examined, and whose laws are to be discovered by observation. The problem is in truth this, What are the general conditions of belief? How do men, as a fact, reach the state of mind called "certitude"? If an exhaustive answer could be given, we should know the laws of belief. But it must be distinctly observed that "law" is here used in its scientific not in its narrower

and more proper sense. The code investigated is not that imposed by logic, but that which is necessarily and always obeyed by the working of the human mind. We are seeking the laws of all belief, not the laws of right belief; and our theory would explain the growth of error just as much as the growth of sound knowledge. Every opinion, true or false, must necessarily obey the laws of thought, when the phrase is used in this sense; and it is a further and different question which of the opinions generated are true, or, in other words, correspond to the facts. Logic may be regarded from this point of view as a particular province of the wider science of belief in general, and it is with that wider science that Dr. Newman is primarily concerned. It will require a distinct step to reach the purely logical problem. Before that step is made, his conclusions may be useful in discriminating between real and sham beliefs, but do not touch the (1) Grammar of Assent, p. 352.