The Coccidae of South Africa. 
81 
but slight difference from the most common pinkish or purpHsh tints 
Pseudococcus stelli sp. n., Pseudococcus craioi and solani are pale yellow ; 
Pseudococcus capensis always appears slightly brownish, claviger reddish- 
brown ; formicarii yellowish-brown ; aphyllonis and olivaceous olive- 
brown ; quercus greenish-brown ; saliniis grey. The filamentosus group 
appears purplish-black, and solitarius sp. n. shiny black, etc. 
When recording the colour of specimens in boiling KOH it is important 
to note the shade at once, as an excess of alkali changes the colour rapidly. 
(d) Segjiientation. — It is usual, for some reason or other, to state 
whether the segmentation of the body is conspicuous or not. As a 
character for determination this is quite useless. In the majority of cases 
it merely amounts to whether the mealy secretion is abundant or scant. 
The segmentation of the body is alw^ays more distinct in the second 
nymphal stage than in the adult. The distinctness gradually diminishes 
as the adult becomes replete. It varies, too, in the same species, as it 
depends somewhat upon whether the insects are in sheltered or exposed 
positions. 
(e) Lateral and Caudal Filaments. — The presence or absence of lateral 
or caudal filaments, and their nature when present, is a character which 
may assist, in a preliminary way, in the determination of species. In a 
few cases (e.g. P. transvaalensis sp. n.) caudal filaments are present and 
lateral ones absent. In P. adonidum all the filaments are more slender 
than are those of P. citri. 
There is naturally the relation between these filaments and the dermal 
spine-gland areas which gives the point significance. 
(/*) Antennae. — The number of segments in the antennae is a character 
which is generally given generic rank, and in the majority of cases seems 
quite satisfactory" as such. In the genera which I now include in the 
tribe Pseudococcini, however, this particular character is undoubtedly 
unstable, and is therefore unsuitable to retain this significance. I shall 
deal with this point briefly under the discussion of the genus Pseudococcus. 
I wish to draw attention here to the practice of giving the antennal 
formula as a specific character. 
This is done in a large number of the descriptions of the species, and it 
is only after one has really endeavoured to determine specimens with its 
aid that the impossibility of the task is realized. A formula is composed 
of numbers indicating the joints, the largest, i.e. longest, being given first, 
the shortest last, with the others in order of their length. When several 
segments vary so much that the sequence may be altered, the numbers are 
placed in brackets. 
I am inclined to think that many, if not most, of the antennal formulae 
have been made on simple comparisons of the segments under the 
microscope, and not by actual measurements. 
7 
