Its

Geometrica is the form found in the extreme south-western part of Cape Colony, in the Cape and Malmesbury Districts. extension requires further investigation, but cannot be very wide.

Trimeni is found along the coasts to the north of geometrica, along each side of the mouth of the Orange River, that is, in Namaqualand and German South-West Africa.

Verreauxii has a very wide inland distribution, extending almost across the southern part of South Africa, from Basutoland on the east to German South-West Africa on the west. Within this area

have been obtained smithii, fiskii, seimundi, and boettgeri.

Oculifera occurs to the north of the zone of verreauxii, having been obtained from Barkly West, Griqualand, Bechuanaland, and Damaraland.

Thus each well-defined type of the geometrica-group occupies a very distinct geographical area in South Africa, the elevation and other physiographic conditions of which vary considerably. If one were able to study the peculiarities of the environment closely there is little question that the variations would be found to be largely adaptive, but the study of this aspect of the subject has not yet been possible. Even if the adaptive nature of the variations were established, it would not follow that the variations have been made by the environment; there seems much more evidence for the view that the transformations are primarily a manifestation of qualities inherent in the constitution of the organisms themselves-endonomic selection.

In a short paper on the South African tortoises of the genus Homopus (Records of the Albany Museum, Vol. I., p. 410), I have shown that the six species of this genus have likewise a very restricted distribution, coinciding in some ways with that of the members of the geometrica-group.

In contrast with the restricted distribution of the representatives of these two groups is the wide extent covered by two other species of South African tortoises, Testudo pardalis and Testudo angulata. In their general appearance these two are very different from one another, and also from the species of Homopus and the allies of T. geometrica. Any common origin must be far removed. Both species are plentiful, and apparently extend all over South Africa. Pardalis reaches to East Central Africa, and, along with angulata, is found in Natal; no local variations of any importance have been observed. It is very significant of the constitutional differences among organisms when one reflects that these two forms are to be found all over South Africa, without any important variations, while species of Homopus and the geometrica allies are highly distinctive. of special divisions of this area. Evidently the members of the Homopus and geometrica-groups have been, and probably still are, in a more variable or plastic state than pardalis and angulata; the first two groups are in process of transformation under South African conditions, while the others seem fixed. Of the two, Homopus is at present much less variable than the geometrica series, as no intergrading forms among its recognized species have yet been found.

C. ORIGINAL TYPE OF THE GEOMETRICA-GROUP.

It is impossible to say with certainty which of the described forms of the geometrica-group is nearest the original type from which the others may be supposed to have originated. In any ancient group we are not likely to obtain the primary type still living for, on the theory of evolution, it may have been lost in giving rise to new types. In modern groups it is more probable that the primary type is still living on, perhaps unchanged, in the ancestral home, while others, radiating, have been subjected to new environmental forces, and transformed along peculiar directions.

If we think of any type of the geometrica-group having the most generalized characters from which the others may have evolved in different directions, tentoria most nearly fulfils these requirements. In this sub-group are gathered together, in more or less incipient stages, all the characteristics which become elaborated to make up the many types of the geometrica-group as we know them. The knobbed form of the dorsal shields is the only character which presents an extreme variation in this sub-group, but in practically all its other features the sub-group has been taken as the starting point for characters elsewhere more pronounced.

If we desire an original type, around which all the divergent forms centre, then the verreauxii sub-group most nearly conforms to such. Taking the stages of the various characters, it is easy to see how by variation in one direction we may get such an extreme as the oculifera type, and, by variation in an opposite direction, we may reach the tentoria type, or the trimeni, and, much more remotely, the geometrica type.

The most logical course is unquestionably that which regards the tentoria type as the original, for from the variational tendencies represented within the sub-group itself, we can readily conceive how all the different combinations have arisen.

On the understanding that all these types have originated from some common ancestor, it may reasonably be asked why have not all the forms been transformed; why have some retained a simple condition in certain of their characters while others have gone on developing? Why have we oculifera, tentoria, and geometrica, representing very different stages of evolution, almost equally abundant at the same time? If the variations of the group were due to some simple internal force within the organism itself, we should have expected that practically the same level would have been reached by all at the same time, and that the original type would have disappeared in the latter transformations. But this is not the case. Germinal selection as a complex force may possibly have sufficed to produce such varieties, one series of determinants obtaining the upper-hand in one place, and others elsewhere.

Though the internal processes in the germ plasm might well have run a different course in different, separated colonies, yet it is much more probable that the explanation of the condition of affairs as we

find them to-day is to be sought in germinal variation under the selective influence of environmental changes. The study of their geographical distribution shows that different types occupy different areas, and there is some evidence that certain of the characters are selective as regards these different environments. I am therefore of opinion that the explanation of the present conditions is to be found in germinal variations along determinate lines, influenced by environmental forces. Given variation along definite lines, then adjustment to environment is to my mind the key to the occurrence of different forms in different districts.

III. SUMMARY.

I. Ten species of the geometrica-group of South African tortoises have been described by different systematists. In the present paper attention has been directed, first, to an analysis of the variations of the separate characters which are relied upon in describing the species, and, second, to the associations and combinations of the variations in the individual, involving the production of specific types.

2. With certain exceptions, all the characters discussed are present at one varietal stage or another in all the individuals. The general uniformity of the characters is taken as proof of the homogeneity and recent differentiation of the geometrica-group, and evolution consists in the gradual elaboration of characters traceable throughout the group.

3. Certain characters are wanting or incipient at one extreme, e.g., the femoral tubercle; certain new characters are restricted to what may be considered as intermediate members of the group, e.g., supplementary femoral shields, rows of caudal shields. These new characters are to be regarded as but further expressions of tendencies already indicated by other characters.

4. All the characters present a gradual change in complexity or elaboration, passing uninterruptedly from what we may consider as one extreme to the other. They afford proof of Continuous or Determinate Variation (Orthogenesis). There is no evidence of any rapid or sudden change from one condition to another, nor the sudden introduction of new characters: Discontinuous or Indeterminate Variation (Mutations). The transformation of any one character may be in only one direction or in more than one.

5. Certain characters are very fluctuating in some sub-groups, and constant in others. The former condition may be taken as indicative of the activity of the processes of germinal variation, perhaps associated with environmental changes, and is conducive to the establishment of new types; the latter state as indicative of a cessation of germinal variation and constancy of environment, and not conducive to the production of new forms.

6. In the production of the specific type, it is found that the stages reached in the transformation of different characters either differ in the same individual (non-correlated variation) or vary pari

passu with one another (correlated variation). In the former the individuals possess extreme and transitional stages or transitional stages of different degrees, in the latter they are either all extreme or all transitional stages.

7. Each varietal type has a restricted geographical distribution, without overlapping, and the more nearly related types occupy adjacent areas. Compared with other South African tortoises, it is found that each species of the genus Homopus has a restricted distribution, with characters showing little variation, while T. pardalis and T. angulata are widely distributed, also with slight variability.

8. Among the ten described species of the geometrica-group, only three distinct types of combination are recognizable: oculifera, geometrica, and tentoria. These may for the present be considered as sufficiently separated from any transitional forms to be given specific rank, and tentoria presents at least three well marked varieties or sub-species tentoria (sensu strictu), verre auxii, and trimeni. Four of the other described species (smithii, fiskii, seimundi, boettgeri) are not sufficiently distinct from the fluctuations of tentoria to warrant even varietal recognition, while strauchi may be a variety of geometrica.

BY J. E. DUERDEN, PH.D., A.R.C.S.

Professor of Zoology, Rhodes University College, Grahamstown.

Justly or unjustly the ostrich has become proverbial for stupidity. Both in literature and conversation the bird is frequently referred to as representing the extreme of foolishness from its supposed habit of hiding its head when alarmed, and leaving the rest of its body exposed to observation and danger. It must be understood that in applying such terms to an animal's behaviour we are, without warrant, presupposing consciousness and passing the same judgments upon it that we would upon the actions of an intelligent human being. Perhaps an enquiry into the actual habits of the ostrich and their underlying significance may serve to shed some light upon the justification or otherwise for the doubtful mentality usually accorded the bird. One generally discovers there is some foundation of fact in the habits of creatures or persons which have become bywords, though it is often found that in the first and superficial impressions the true significance of the phenomenon has been missed.

All ostrich breeders are aware that if ostrich chicks, while only a few days or weeks old, are suddenly startled they will at once crouch down, as if dead, and remain thus for some time. The action is much more likely to occur if the chicks have been hatched and reared out on the veld, away from human influence. An actual personal experience will best illustrate what happens. Driving over an ostrich farm one day a pair of ostriches were come upon, surrounded by a dozen or so chicks about a week old; the chicks had been hatched out on the open, and no personal attention given them. Dismounting from the cart, a sudden, clamorous rush was made towards the group, when the parent birds at once ran away and the chicks scattered and disappeared so suddenly and so completely that for the moment one knew not which way to turn in order to follow them. After remaining quiet for a few moments, and getting the eye accustomed to the surroundings, it was found that several chicks were lying quite near, so that one might almost have stepped upon them. Their bodies were prone upon the ground, with the neck and head stretched out resting upon the surface. The resemblance of the mottled black and brown bodies of the chicks to the rocky ground and scant vegetation was so perfect that it was with much difficulty the various individuals were discovered.

The recumbent chicks allowed themselves to be approached and picked up, when they were found to be quite limp and motionless, appearing altogether lifeless, and only recovered after being conveyed for some distance.

This behaviour on the part of the ostrich chick is a clear and striking instance of death-feigning or death-shamming. In this state. the animal appears and acts as if dead, and, thinking of it in terms of human actions, we explain it as if consciously assumed for the purpose of deceiving its enemies. The same instinct is also well developed in many birds (rails, bustards), mammals (opossum, jackals, foxes), and especially in insects.