DOG. gators of the law of variation. One of the most riK-ent and snooessful ot these was a comparison by Windle of the skulls of some (50 varieties, rep- resentins most of the types, the results of which were extensively tahulaled and discussed in the Frocecdiiuis of the Zoohyical Society of London for 1S!*0. lie says that the most noteworthy fact learned from his comparative measurements of dog-crania was that the averages of the ditVereiit breeds, especially in points relating to some of the teeth, dilTer very little from one another; in other words, that, speaking generally, the teeth in one dog are relatively to the skull very similar in size to those of any other. It is also to be noticed that the range of variation in any breed is much greater, in almost every case, than that existing between any two breeds. Jlr. Windle says : "The extreme variations in any breed are probably due to the fact that, strictly speaking, so few animals of the same group are really in any sense of the same breed. The various mem- bers of a carefully selected strain of terriers, for example, bred by one breeder, might be com- parable with one another, and yet quite different in descent from another and perhaps equally good strain belonging to another breeder and to another part of the country. With dogs bred for show purposes, as so many of the pure strains are, and with constantly varying requirements of fashion, all sorts of crosses, as any manual of dog-breeding will show, have been tried with a view of attaining the ideal, whether of sym- metry, ])ace, or carriage. That such crosses should, at times at least, leave their marks upon the skulls, and cause differences in breeds which cannot be accounted for, is. of course, to be ex- pected. The presence of the disturbing factor can be appreciated, though its exact nature can- not always or even frequently be ascertained with any correctness. Tluis amongst the eleven bulldogs' skulls which we have examined, there was one which differed in measurements con- siderablj- from the rest. It was nearly one cen- timeter longer than any other, and. what is much more significant, it was seven centimeters hmger than it was broad, the average for the others being about three or four centimeters, iloreover, its palate was 1.90 centimeters longer than it was broad, whilst in every other case but one the breadth exceeded the length. In the second case, the length was .00 greater than the breadth of the palate, and the length of the skull nearly five centimeters greater than the zygomatic width. We cannot doubt that both of the skulls above mentioned were those of dog's in whom, to a greater or lesser degree,' there was an admix- ture of strain, of what kind it is impossible to say. And what is true of these is doubtless true also in les.ser degree of the greater number of specimens coming under examination. It thus becomes apparently a hopeless task to look for evidence as to the proximate or ultimate deriva- tion of the breeds of domestic dogs in their skulls or teetli." Having arranged his skills with reference to relative length and breadth, Mr. Windle found th.it the distinctly broad-headed dogs form a well-marked groiip by themselves, including the Chinese pug-nosed spaniel, the pug. bulldog, blaek-andtan toy terrier, and King Charles span- iel, a considerable interval existin<r between these and the next. All these are highly artificial 855 DOG. breeds, which require great care and attention in order to prevent deterioration, with its conse- quent elongation of the skull. Next to this group comes one, largely consisting of terriers, with heads inclining to be broad. A miscellaneous i;rc)u|) next foUows, gradually decreasing to the ilistinctly narrow-licadcd dogs, such as the Irish wolf-dog and the greyhound. Tiiese investigations demonstrated the fact that in the highly artificial broad-headed dogs elonga- tion of skull and palate is a sign of impure brwding, an evidence of admixture with the broad-headed strain of that of some other and narrower-headed dog. Examples of this may be seen almost any day in the streets in the shape of the half-bred pugs, in which the elongiited muz- zles present so great a contrast with the short, square faces of their pure-bred cousins. We have no facts before us to prove whether the long- headed dogs, such as greyhounds, tend to become broader when impurel.v bred, but it is highly probable that they would do so, and consequently that the dogs at both ends of the scale would, under the influence of promiscuous interbreed- ing, tend to approximate to the average head. Classification. Breeds of dogs have lieen variously classified. One Roman grouping men- tioned by old writers was into fighters (pug- naees), wise dogs (sagaces), and swift-footed ones (eeleres) ; the sagaces were said to have come from Greece, and the pugnaees from Asia. Another ancient grouping was into house-dogs, sheep-dogs, and sporting-dogs, the last embracing fighters, hounds hunting by scent and hounds hunting by sight. ^lodern naturalists have sub- stantially agreed upon six groups, with consid- erable differences in composition, however. Thus Col. Hamilton Smith, about 1830, arranged the list as follows: (I) The wolf-dogs, including the Siberian dog, Eskimo dog, Iceland dog. New- foundland dog. Nootka dog, sheep-dog, great wolf- dog, great Saint Bernard dog, Pomeranian dog, etc. (2) The watch and cattle dogs, including the German boarhound, Danish dog, matin, dog of the Xorth American Indians, etc. (3) The greyhoimds, including the Brinjarce dog, differ- ent kinds of greyhound, Irish hound, lurcher, Egvntian street dog, etc. (4) The hounds, in- cluding the bloodhound, old southern hound, stag- hound, foxhound, harrier, beagle, pointer, setter, spaniel, springer, cocker, Blenheim dog, water-dog or poodle, etc. (.5) The cur dogs, including the terrier and its allies. (G) The mastiffs, includ- ing different kinds of mastiff, the bulldog, pug dog, etc. The latest arrangement is that by Windle, based upon the shape of the skull and other features, rather than upon form and function, yet not greatl.v different from its predecessors. It is as follows: I. Wolf-liUe Dogs. — Arctic sledge-dogs; shep- herd dogs (collies) ; Newfoundland dog: Saint Bernard, rough and smooth, and Pomeranian or Spitz dog. IT. areyhounds. — Old Irish wolf-dog; modem Irish wolf-dog; greyhound: Italian greyhound; West Indian naked dog ('presumably'). TTI. flpanieU. — . varieties. TV. Uoutids. — Bloodhound: foxhound: har- rier: otter-hound, beagle, pointer, setter, etc. (hunting-dogs).
