The egg, the five larval instars, cocoon, prepupa and pupa of Bracon tachardiae, Cameron, are described.

Two series of larvae, one male and one female, occur, these series overlap very considerably and it is only towards the maximum and minimum of the range that larvae are respectively entirely female or male.

It is shown that the five larval instars of B. tachardiae may be grouped by the widths of the head-capsule, and that an instar can be characterised by the head-width, particularly if only averages are considered. The actual range of head-width in each instar is rather large, and in some cases the extremes overlap. In actual practice head-width is a safe indication of the instar of a larva as the extremes rarely occur; the widths calculated on Dyar's principle also approximate sufficiently closely to the observed widths to preclude the chance of an ecdysis having been overlooked.

The length of the mandible is for practical purposes identical in a larva of any given instar and in its exuviae. The larval instars and their exuviae can be grouped by the lengths of the mandibles, particularly if averages are considered; the range for a given instar is wide but the extremes do not overlap. The lengths calculated on Dyar's head-width principle approximate closely with the observed lengths, sufficiently so to preclude the possibility of overlooking an ecdysis, and making possible the identification of the instar of fixed and mounted specimens of larvae or exuviae, for which purpose head-width is useless.

The factors of increase of head-width and mandibular length are fairly similar, falling near to 3 √2 =1.26, and for practical purposes one may be substituted for the other, the calculated widths and lengths so obtained still approximating to the observed widths and lengths.

Various uses of this fact are indicated. Preliminary investigations point to the fact that similar observations apply to Aphrastobracon flavipennis, Ashm., an ectoparasite of Eublemma scitula, Ramb. It is suggested that similar observations may be true for other ectoparasitic Braconids.

The growth of the body of the larva from instar to instar is independent of the growth of the head-capsule, larvae increasing in weight and in volume by a figure lying between 3 and 4 times from instar to instar. The actual figure obtained theoretically for volume increase was 3–6. The head-size increases by the usual double at each moult.