than one puncture width between adjacent punctures; ventral surface of thorax only shallowly and sparsely punctate; mesothoracic carina well developed, platelike, highest between anterior coxae, where it extends ventrad well beyond base of trochanters; legs proportionately longer than in subgenus Rhopalimorpha (body only 1 ½ times as long as hind femur and tibia together); abdomen markedly concave, both above and below, at about level of scutellum apex; anterior abdominal spine moderately strongly developed, nearly straight, reaching midcoxae; the female with a pair of conspicuous dark ovoid patches on both 6th and 7th abdominal sterna, the 7th sternum broadly rounded and non-carinate, with its posterior margin very broadly excavated, and the anterior genital valves flat, with posterior margins nearly straight.

Type species: Rhopalimorpha (Lentimorpha) alpina sp. nov.

Lentimorpha resembles the typical subgenus Rhopalimorpha Dallas in the general form of the pronotum (not strongly declivous and not strongly arched between or behind posterior angles, so that the “hunch-backed” appearance of many Acanthosomatinae is lacking; sides straight or nearly straight; anterior margin broadly excavated; posterior angles broadly rounded, not prominent, not produced appreciably beyond abdominal margins); in the body being more or less ovoid, not definitely shield-shaped; tylus longer than juga, with apex broadly rounded; hemelytra barely surpassing apex of abdomen; spout of metathoracic scent-gland orifice developed as a large plate raised above general ventral surface and with outer and posterior margins free; the posterior connexival angles of the 7th abdominal segment in the female not produced back as acute spines; the anterior abdominal spine not strongly curved towards apex; and in the abdominal venter not being definitely carinate throughout its length.

These characters are shared with the allied Australian genus Eupolemus Distant, but in the genus Rhopalimorpha the pronotum is even flatter, with the sides not strongly reflexed, the sides of the head are much less strongly arcuate and concave, and the plate of the metathoracic scent-gland orifice is much broader (antero-posteriorly).

[The section below cannot be correctly rendered as it contains complex formatting. See the image of the page for a more accurate rendering.]

Lentimorpha differs from Rhopalimorpha in the more robust and broader ovoid form of the body; in the proportionately shorter head [width: length = 1.35 (1.1–1.25 in Rhopalimorpha)]; in having the first antennal segment well surpassing head and second segment shorter than third; the longer bucculae; the rostral segments extending proportionately further back, as described above (in Rhopalimorpha first segment extends only about 5/8 of way to base of head, second and third segments do not reach anterior and middle coxae respectively, and fourth segment reaches only to middle coxae); the pronotum with anterior angles not rounded, but acut and produced, and the base concave; the punctation of pronotal disc, scutellum, and thorax beneath finer, sparser, and more obscure, the much larger and expanded mesothoracic carina, the larger abdominal spine (not reaching middle coxae in Rhopalimorpha); the proportionately longer odoriferous plates and legs; and, in the female, the presence of conspicuous dark patches on both 6th and 7th abdominal sterna, the 7th sternum being without carina and very broadly excavated behind, and the anterior genital valves flat, not sulcate, their inner margins not raised.

Rhopalimorpha (Lentimorpha) alpina sp. nov. Figs. 5, 29, 30, 42.

♀. Length: 10·7 mm. Width: of abdomen, 5·8 mm.; of pronotum (between posterior angles), 5·2 mm. Colour: Body and appendages more or less uniformly ochreous, except for the darker brown dorsal punctures, tarsi, and apices of tibiae and claws, the pink ocelli, the reddish-brown eyes and abdominal venter, and the brownish-black apex of rostrum; no median pale line on pronotum or scutellum. Head: Tylus with only a few obscure, very shallow punctures; rounded apex projecting beyond juga. Rest of dorsal surface rugose, rather shallowly and sparsely punctate. Apices of juga well rounded, not forming acute angle at junction with tylus. Width across eyes 1.3 times length (47:36). Interocular space 3. 5 times as wide as eye (30:8 5). Ventral surface impunctate. Maxillary plate process a rather narrow shelf, reaching antennal base, not toothed nor much expanded. Bucculae long, reaching almost to base of head, behind level of posterior margins of eyes. Rostrum reaching hind coxae, first segment reaching to posterior ¼ of head, second to fore coxae, third to middle coxae; proportionate length of segments I-IV, 20:22:24:19 Antennal segments I-IV, 18:18:22:30; first segment well surpassing head. Thorax: Pronotum 2.6 times as wide across posterior angles as long (95:36), proportionate anterior width 50; moderately coarsely punctate, except on posterior part of disc, where punctures are shallower and sparser; calli confluent; postero-lateral margins very shallowly sinuate; base widely and rather shallowly concave. Scutellum with apex acute; length : anterior width : 58 : 50; rather finely and sparsely punctate. Mesothoracic carina reaching to anterior end of fore coxae and surpassing apex of second rostral segment, anteriorly expanded, extending ventrad well beyond base of trochanters. Ventral surface of prothorax very inconspicuously punctate, with punctures fine, shallow and sparse, and almost concolorus with general surface. Plate of metathoracic scent-gland orifice rather long, length (from outer end to mesial apex of orifice) over three times greatest width (20:6); ochreous, with margins pale and reflexed and a more or less well defined submarginal brownish-black line; broadest near the outer end, which is widely rounded; posterior margin nearly straight, only slightly and widely convex. Both anterior and posterior free lips of orifice prominent, the former the larger; apex of orifice continued mesially beyond their junction as a well-defined groove. Metathoracic evaporating area large, extending well beyond outer margin of plate and reaching close to lateral margin of segment, pale ochreous, very finely granular, with irregular folds; rest of metathoracic venter only shallowly, sparsely, and inconspicuously punctate. Mesothoracic evaporating area large, colour and texture as for metathoracic area, and with irregular folds; its posterior division occupying nearly half length of segment and with a deep transverse sulcus margined by a strong, rounded ridge; its lateral division reaching almost to anterior margin of segment; mesothoracic venter in front of posterior division impunctate. Legs impunctate, moderately long; body 1.5 times length of posterior femur and tibia together (198.133). Corium and cuneus rather finely and shallowly punctate; hemelytral membrane markedly declivous. Abdomen: Venter strongly convex, shining, reddish-brown, impunctate, finely and reticulately rugulose. Dorsal surface, so far as seen, dark brown. Anterior spine rather stout at base, elongate conical, reaching to anterior margin of mid-coxae and just in front of posterior end of thoracic carina. Connexivum moderately well developed, projecting slightly beyond sides of closed hemelytra. Postero-lateral

apex of seventh segment of ♀ a rounded right angle. A pair of conspicuous, dark, ovoid patches on each of sixth and seventh sterna of ♀, both close to posterior margin of former, the anterior pair the larger. ♀ with seventh sternum broadly convex, not at all carinate, its posterior margin very broadly excavated; anterior genital valves flat, obliquely rugulose, with inner margins not elevated, disc not longitudinally sulcate, and posterior margins nearly straight.

Locality: McKinnon Pass, Southland, S. Island; 28.xii.1919; 1 ♀; from C. E. Clarke coll.

Holotype: In collection of Auckland Museum.

Discussion: The erection of a new subgenus on the basis of a single specimen is done with a certain amount of reluctance. However, the set of diagnostic characters distinguishing it from the other known species of Rhopalimorpha is so marked as in my view to justify the separation. On the other hand, it seemed desirable to emphasize the undoubtedly close relationships of Rhopalimorpha and Lentimorpha, and their probable common origin within New Zealand, by treating the latter as of subgeneric status, rather than at this stage to split off a monotypic genus. This is especially so since in several features alpina more closely resembles lineolaris Pendergrast than the latter does obscura White. Such characters are: the plate of the metathoracic scent-gland orifice, the mesoand metathoracic “evaporating areas”, and the punctation of the posterior division of the metathoracic venter. In the rather broad ovoid form of the body, in the broad, low carina of the seventh abdominal sternum of the ♀, the shape of its posterior excavation and of the genital valves, and to a lesser extent in the development of the abdominal spine and mesothoracic carina, lineolaris is intermediate between the other two species.

Alpina is undoubtedly more closely allied to lineolaris than to obscura But in a consideration of affinities, these resemblances have to be weighed against the set of characters in which alpina differs from both these other species. (See p. 303.) Some of these characters are not evident at all in either lineolaris or obscura; others (the intermediate features listed above) represent tendencies shown to a minor degree in lineolaris but in alpina developed to an extent that gives this species a highly distinctive facres. In trying to decide what has been the main line of evolution, i.e., whether in most respects alpina has diverged from a lineolaris-like form or whether lineolaris and obscura are to be derived from a species showing many of the characters of alpina, it is necessary to decide, as far as possible, which features, apart from those common to all three species, may be regarded as ancestral for the genus. For data there are available the characteristics of allied genera like Eupolemus, due allowance being made for divergence within such a genus, and those of other more “typical” Acanthosomatine genera. On this basis the weight of evidence supports the view that, of the three species of Rhopalimorpha, alpina has retained ancestral characters to the greatest degree. Subject to the qualifications shortly to be made, these might include the form of the seventh sternum and genital valves and the presence of two pairs of sternal patches in the female, the large abdominal spine and mesothoracic carina, and, possibly, the acute anterior pronotal angles. At the same time, many of the characters distinctive of the subgenus Lentimorpha must be regarded as secondary acquisitions. The discovery of further species of either group would no doubt throw more light on the evolutionary trends involved.

The occurrence of such species, if they were annectent, might invalidate division of the genus or necessitate re-estimation of the subgenera; if they were not, the subgeneric distinctions might be accentuated. A study of the male genitalia of alpina is most desirable. It would appear, however, that the effect of the topographical and climatic isolation to which alpina has been subjected has been in two directions: the conservation of certain relic features and the development of divergent characters not shared either by related genera or by the other species of Rhopalimorpha.

The case against generic separation of Lentimorpha seems to me to be as follows: (a) The presence in all three species of Rhopalimorpha of common features not found in the closely allied genus Eupolemus. (b) The affinities already discussed between alpina and lineolaris. (c) The considerable intrageneric variation which may occur among Pentatomidae in such characters as size of anterior abdominal spine and mesothoracic carina, and development of the ventral thoracic punctation and of the sternal patches of the female. Until more is known of the detailed anatomy of members of the subfamily and of probable phylogenetic lines within it, it is difficult, in some cases, to judge which features should be regarded as relatively archaic, and to eliminate the possibility of convergence. For example, in “typical” Acanthosomatinae a large abdominal spine and mesothoracic carina and the presence of conspicnous (? sensory) patches on both sixth and seventh abdominal sterna of the female seem characteristic, but the variability of these structures within otherwise apparently compact genera has already been indicated. The genus Eupolemus, in which all these characters vary, is a case in point A large thoracic keel, e.g., might be regarded as a common ancestral characteristic of Eupolemus and Rhopalimorpha, but with a tendency to reduction already established before separation of the genera, the extent of reduction varying in different species of both. The specific appearance of many of these features is the expression of a stage in a trend, which may often be older than the species involved, rather than a clear-cut positive or negative character. Similarly, reduction of either or both pairs of sternal patches of the female has occurred apparently independently in different species of Eupolema and Rhopalimorpha In E. picturatus, e.g., only the pair on the seventh sternum is apparent, while in E. insularis neither pair is obvious. Pendergrast (unpublished; see p. 316) has shown that in R. lineolaris and R. obscura a specialized cuticular region exists on the seventh sternum probably comparable to that on the sixth, but unlike the latter not sharply differentiated from the general surface. Examination of intact specimens indicates that a similar condition probably holds in Eupolemus spp. The relative conspicuousness of these areas, though it may in some instances be useful in helping to evaluate evolutionary trends, can in itself scarcely be considered a character of major systematic importance.

It is the writer's opinion that, if alpina is to be separated generically, then so should obscura and lineolaris; that is, that depending on the taxonomic value accorded to the difference between them, either all three species should be relegated to the one genus or each should be placed in a different, at present monotypic, genus.

By analogy, the apparent importance of isolation in the differentiation of alpina suggests a means whereby obscura and lineolaris might have diverged by