1. A study of meiosis in male Chorthippus elegans, Acrididæ (2n = 16 + X) shows the chromosome behaviour to be similar to that already described in Stenobothrus parallelus. Thus the chiasma frequency is an indirect function of length and has an interference curve of variation. Failure of pairing occurs in one chromosome type with a frequency in keeping with the curve. Terminalization depends entirely on the generalized repulsion in bivalents with closed loops and the localized spindle attachment repulsions are of the minimum degree found in Fritilaria. 2. A more extensive quantitative study makes it possible to show in the long chromosome type, that the number of terminal chiasmata at each stage between diplotene and metaphase is proportional to the number of interstitial chiasmata and increases from one stage to the next pari passu with the decrease in the number of the interstitial chiasmata. This confirms the earlier arguments that all interstitial chiasmata become terminal by movement without breakage while, on the other hand, terminal chiasmata always arise from earlier interstitial ones and in no other way. It is now clear that chiasmata always change their position after their formation at diplotene so that the configurations observed later are merely positions of changing equilibrium. 3. The opening of the diplotene loops has been followed in detail and shows that the chromosomes are single and undivided until this stage. The opening is therefore derived solely from the so-called "reductional" split, not from the "equational" one which only begins to appear at this time. 4. Syndiploidy occurs frequently just before metaphase.