191S.I F. H. (Jravkly: Passalidae of llw World. 13 



'^ No secondary tubercles or angular processes present en the 

 anterior margin of the licad (or behind the clypeus when tliis 

 is exposed) between the outer tubercles and anterior angles , 

 the posterior plate of the prosternuni almost always much 



narrowed, often pointed, behind (American forms) . . . . . 4. 



Vt least one pair of secondary tubercles or angular processes 

 present on the anterior margin of the head between the outer 

 tubercles and the anterior angles ; the posterior plate of the 

 prosternum more or less parallel sided, broadly truncate 

 behind (African forms) . . . . . . . Solenocychnae, p. 68. 



I The clypeus exposed and separated from the frons by a distinct 

 suture . . . . . . . , Pseudacalhinae, ]>. Ji'. 



The clypeus fused witli tlie frons- and usually hidden ., .. .. 5. 



( The outer tubercles obsolete or absent, the clypeus almogr. 

 .") J always exposed' . . . . . . . . Pfoodinae, p. 32. 



( The outer tubercles distinct, the clypeus rarely exposed . . Passalinae, p. 43. 



(The anterior margin of the head rarely with more than one pair 

 of tubercles,' though these may be compound in structure ; 

 the antennae rarely with less than four well developed 

 uv lamellae ; never with both these characters . . . . Macrolininae, p. 7G. 



I The anterior margin of the head iilwa\s with two pairs of 

 I snnple tubercles ; the antennae always with three well 

 \^ developed lamellae only .. .. .. .. Leplaxdadme, t^ \\\. 



Siiblaniily U'LACOCYCLLWIE. 

 Three of the genera dolinocl in my i)ieviou8 accMuint of this subfamily (1914c, pp. 192-3) 

 appear, in the light of further material, to be unnecessary. Two of these — Canlifer and 

 Aiiritidus — are monospecific, and may advantageously be merged in AuUicocijclus and 

 Ci/lindrocaulus, respectively. The third, Tristortlnis. may also be merged in Aidacoci/clus. 

 The best known species of Tristorthus is T. (ricuspis, Kaup, from New Caledonia, a species 

 shown by its short antennal lamellae to be most nearly allied to the Australian species 

 of Auldcocijcbis. With this are associated firstly, two other Xew Caledonian species, 

 apparently distinguislied largely' by differences in size, and in my opinion doubtftiUy 



* The processes especially characteristic of the Solenocyclinae are situated iiuniediately above the ventral tubercles. 

 Others may also, however, be present, and in all .Malagasy forms a pair is more or less distinctly (levcloi)ed imnio<liatclv 

 on the inner .^iile of the fronto-vertical suture. The most i)rimitive .Malagasy genus, Malmja.vihi.i. laeks the former pair of 

 processes, and the above definition is a])plicable only by reason of its jio^srssion of the latter. The African and 

 Malagasy groups are composed, broadly speaking, of parallel series of genera separatinl largely by the i>resenee or absence of 

 the latter pair of prma-sses. 1 do not know of any .Vfriean genus paralleling .\Inltiij<toilii.i .• but if one exist.s the above 

 definition can hardly be expected to apply to it. See also above, pp. 10-11. 



* Except in one sjxjoies of limtcx (below, pp. 4.1 & 47) ; imperfectly fused in one Sju>oiesof Chondmrfphnlus (Iwl.m 

 pp. 43-45, lig. vi, 1). Sice also above p. 11. 



» The clypeus is completely hidden only in tli.- genus Plfili/irms. though in the transitional six>eies Vcrrt* cnrlkoln 

 it is hardly apparent. In both these species the iinier tulwrclcs are situated on the anterior margin of the head and muv 

 readily be mistaken for outer tubercles, though a comparison with other species of Vtms, and es|>.'cially uifb 1'. 

 ■cavieoUis, at once settles their true homology. The heads of the sjx>cies in <piestion ar- shown in (ig v, p. .14. 



* Only in the genus Tarquiniiig, which hn? six well d-velopixl autonnal lamellae. 



