were reported as early as 1861 by Steenstrup (p. 688) while Hertwig (1874, p. 358) also noted that new dermal denticles are formed between an existing generation of denticles.

Within the Squaloidea, Jordan and Hubbs (1925, p. 107) briefly describe denticle changes in Centroscymnus owstonii, but their findings appear to have been overlooked by later writers. Bigelow and Schroeder (1954, p. 47, fig. 2) amplify Tortonese's account of similar changes in Centroscymnus coelolepis, and suggest (p. 50) that the same situation occurs in C. owstonii. More recently (1957, p. 14) they note that denticle changes take place in Oxynotus spp. and describe changes (p. 48, fig. 3) in the denticles of Etmopterus princeps.

In the present revision of the New Zealand squaloids, denticle changes involving differences in shape, sculpture or proportions between juveniles and adults were observed in 11 of the 14 species examined. For the remaining three species (Etmopterus lucifer, Dalatias licha and Scymnodalatias sherwoodi), an insufficient range of material was available. In many of the species, the changes involved are not great, but they are included here to show the universality of the phenomena, and also to give some indication of the trends of the changes. It should be noted that change in size was not regarded as a criterion, insofar as adults of all species have larger denticles than juveniles, though there is no direct proportion between total length and denticle size.

Species showing the least changes in their denticles are those with spine-like or thorn-like denticles—e.g., Etmopterus baxteri, Echinorhinus brucus and E. cookei. In Etmopterus baxteri (and also in E. princeps as noted by Bigelow and Schroeder [1957, p. 48]) juvenile form denticles are slender and bristle-like, but are replaced by stouter, thorn-like denticles in adults. In Echinorhinus brucus, strongly ridged bucklers with deeply indented margins appear to be the juvenile form, and are succeeded by finely ridged bucklers with more nearly entire margins; similar but less contrasting forms appear in juveniles and adults of E. cookei (Garrick, Part X).

Species with blade-like dermal denticles show a greater variety of changes, though with one or two exceptions they follow trends similar to those observed by Radcliffe (1916) in the carcharhinids. These are that the blades of adult form denticles have additional longitudinal ridges on their external surfaces, and reduction of the teeth on the posterior margins. There might be added for the Squaloidea the trend for the blades of adult denticles to lie more nearly horizontal compared with the near vertical or steeply directed blades of juvenile denticles. Good examples of these trends are seen in Centrophorus squamosus, Centroscymnus crepidater and Scymnodon plunketi (in Garrick, Parts VII, VIII and IX respectively). Exceptions are in Centroscymnus owstonii (Garrick, Part VIII) where longitudinal ridging is reduced and lost in adult denticles, in Oxynotus bruniensis and Deania calcea (Garrick, Part XI) where additional teeth are acquired posteriorly or posterolaterally on the blades, and also in O. bruniensis where adult form denticles are directed more steeply from the skin than juvenile denticles. Another trend common to the genera with blade-like denticles mentioned here, but not including Squalus spp., is for the anterior margins of the denticle bases to become multi-angled. This is accompanied by the development of additional ridges extending from the angles on to the pedicle. In comparison the denticle bases of Squalus spp. remain rhomboid even in adults, and have only the four major ridges extending on the pedicle.

Despite the above trends, considerable variation is shown in the nature of the changes between different species. It is of particular interest (and value also to the systematist) that the variation can occur even in species of the same genus. A striking example is seen in Centroscymnus spp. where C. coelolepis and C. owstonii show a complete elimination of the longitudinal ridges and the posterior marginal teeth in adult form denticles, while in C. crepidater these features remain strongly developed (Garrick, Part VIII). In all three species the juvenile denticles are

very similar. Likewise in three Squalus spp. examined (Text-fig. 3), two of them, S. acanthias and S. megalops, agree in their juvenile form denticles but differ widely in their adult denticles; the third species, S. blainvillii, contrasts in its juvenile denticles, but shows agreement with S. acanthias in its adult denticles.

Squalus spp. are also unusual in that although their adult form denticles are horizontal blades with a very high median longitudinal ridge, their juvenile form denticles show little development of the horizontal blade (except in S. blainvillii) but strong development of the median ridge (Text-fig. 3). In species of the other genera examined, the horizontal blade is well developed in juvenile denticles, and in Scymnodon plunketi there is initially no median ridge though it becomes progressively stronger as replacement denticles appear.

The above observations were mostly made on the denticles from the side of the trunk below the 1st dorsal fin, an area long regarded as standard. However, during the study it became apparent that examination of only the standard area could be misleading in specimens of critical size—i.e., those about one-third grown. In such specimens the denticles are still mostly of juvenile form, except on the side of the caudal peduncle. Replacement of denticles on the side of the trunk appears to begin posteriorly, and to progressively advance forwards. Jordan and Hubbs (1925, p. 107) came to the same conclusion after examining Japanese specimens of Centroscymnus owstonii, but I was not aware of their account when describing (1959 b; 1959 c) this phenomenon in New Zealand specimens of C. owstonii and also Scymnodon plunketi. Actually earlier replacement begins on the dorsal midline at the anterior ends of the bases of the dorsal fins and of the epiural lobe of the caudal, and also between these fins in a longitudinal area parallel to and alongside the mid-dorsum. In this respect there is fairly close correspondence between the sites of origin of the denticles of the embryo and the sites of the first replacing denticles of each later generation. However, as the mid-dorsal denticles often have different facies from those on the side of the trunk as a whole, it is perhaps better that they be disregarded in denticle change studies (in Squalus acanthias the first mid-dorsal denticles to erupt in the embryo are tridentate, whereas on the side of the trunk there are one or more generations of spear-head shaped denticles before tridentate forms appear).

There can be little doubt that with further study of denticle changes, the value of denticles as systematic criteria will show a considerable increase. This will require, as pointed out by Radcliffe (1917, p. 26) the assembly and examination of skin samples as a normal routine, but compared to many procedures this requires little time or effort. The result should well exceed Radcliffe's opinion (p. 27) “that these (dermal denticles) are not an open sesame by which all sharks may be readily identified, but that in a troublesome group they are an aid”.

In addition to the problems of relationships which stem from changes in the dermal denticles, other problems are also apparent. Present knowledge of the structure and development of dermal denticles does not recognise further growth of the denticles once they are erupted. This means that any change in the shape or structure of the denticles throughout the life of an individual must be interpreted in terms of denticle replacement. Such changes are numerous in some species, and examinations of a series of specimens, even to near maximum size, gives a picture of continual increase in denticle size or continual change in shape. Accordingly denticle replacement must occur at a matching pace, even in the larger specimens. With some species, support can be given to this view, for in Squalus spp. and Dalatias licha newly erupted denticles, not yet pigmented, are evident even on near-maximum size specimens. On the other hand, Bigelow and Schroeder (1954, p. 50) in their study of Centroscymnus coelolepis observe that “the uniformity in size and shape of the denticles on adults, and the regularity with which they are arranged, suggest that the ovoid ones that first develop on specimens a little more