TAXONOMY OF TROMBICULIDS 
125 
typical of the whole congeneric group., having in fact drawn attention to itself by some excessively 
developed character or other. The genus Doloisia (p. 157) and subgenus Trombiculindus 
(p. 1 41) are examples. It has often happened in the Trombiculidae that the characters chosen 
for generic separation are recurrent polyphyletic characters, which appear independently in 
various unrelated groups. Such characters and their significance are discussed below. In the 
case of such genera raised on outstanding characters or groups of characters, it may later be 
found that a series of the other species puts the genotype in its proper perspective and the genus 
may then be identified through that series with some other genus, or it may appear to be at the 
most of no more that subgeneric rank. The selection of some exceptional species as the type of 
a new genus may give rise to a long delay before genuinely congeneric species are accommodated : 
the result in any case is frequently a genus of which the genotype is actually not typical of the 
natural group. This may not be very important, but the delay in recognising the congeneric 
species may be important and it is simply necessary to bear this in mind constantly. It follows, 
therefore, that monotypic genera should be raised cautiously and for exceptionally good reasons, 
especially in the earlier stages of classification. 
The third way in which a genus or subgenus comes to be recognised is as a distinct species- 
group which is in time considered to have a higher status. An example is the tsutsugamushi- 
group of Nagayo and his colleagues. This group was recognised early because a number of 
similar species were found together in Japan, and the genus Leptotrombidium was proposed to 
accommodate them. This genus is temporarily considered as a subgenus of Trombicula sensu 
lato. Brennan (1952) has drawn attention to the T. cynos group as a potential genus. In the 
case of such genera, a good representative genotype can be selected from the group. This is 
important because the only question one may ask of a new species is whether or not it is 
congeneric with any known genotype. 
A review of the progress of trombiculid taxonomy reveals four important stumbling blocks, 
viz. (a) incomplete early descriptions, making many comparisons impossible without reference 
to the genotype itself — this should be remedied by the redescription of type material (vide 
Fuller 1952) and by the publication of detailed and illustrated comparisons with related species 
at every opportunity. ( b ) The interpretation of recurrent polyphyletic characters, (c) The 
delayed study of the postlarval stages and the interpretation of post-larval characters, (d) The 
inadequate genotype of Trombicula — it is proposed in this paper to adopt, as a provisional 
substitute for the genotype, a species known in both larva and nymph which appears for very 
good reasons to be very close indeed to it, if not identical. It will thus at least be possible to 
build up a consistent subgenus. Points (b) and (c) are discussed below. 
Interpretation of recurrent polyphyletic characters 
A number of characters are now well recognised as being of polyphyletic origin and 
occurring as the same deviations from the standard in different groups. There should be a 
tendency to acknowledge and erect subgenera rather than genera when such characters are 
themselves made the chief points of difference. Care should also be taken to avoid grouping 
unrelated species because they share such a recurrent character. Furthermore, only a long 
study of adequate numbers of species may show whether a recurrent character should be 
considered as of more than specific importance. It may be wise in each case to start by 
ignoring the recurrent character completely and attempting to discover the relationships of the 
species concerned on the remaining characters. Audy (1952) has commented on the poor 
chitinisation and the posterolateral peninsulas of the scutum of Trombicula dimolinae : this 
group might include members in which the occurrence of PL setae off the scutum is no more 
than a specific character. In a group related to Doloisia (p. 157) this same character appears 
perhaps to be of subgeneric importance. The more important recurrent polyphyletic 
MALAYA , No. 26, 1953 
