I have set forth the complete tragedy, as it is fairly often enacted, but not always. The insect is not a machine, unvarying in the effect of its mechanism; it is allowed a certain latitude, enabling it to cope with the eventualities of the moment. Any one expecting to see the incidents of the struggle unfolding themselves exactly as I have described will risk disappointment. Special instances occur--they are even numerous--which are more or less at variance with the general rule. It will be well to mention the more important, in order to put future observers on their guard. Not infrequently the first act, that of paralysing the thorax, is restricted to two thrusts of the sting instead of three, or even to one, which is then delivered in the foremost segment. This, it would seem, from the persistency with which the Ammophila inflicts it, is the most important prick of all. Is it unreasonable to suppose that the operator, when she begins by pricking the thorax, intends to subdue her capture and to make it incapable of injuring her, or even of disturbing her when the moment comes for the delicate and protracted surgery of the second act? This idea seems to me highly admissible; and then, instead of three dagger-thrusts, why not two only, why not merely one, if this would suffice for the time being? The amount of vigour displayed by the caterpillar must be taken into consideration. Be this as it may, the segments spared in the first act are stabbed in the second. I have sometimes even seen the three thoracic segments stung twice over: at the beginning of the attack and again when the Wasp returned to her vanquished prey. The Ammophila's triumphant transports beside her wounded and writhing victim are also subject to exceptions. Sometimes, without releasing its prey for a moment, the insect proceeds from the thorax to the next segments and completes its operation in a single spell. The joyous entr'acte does not take place; the convulsive movements of the wings and the acrobatic postures are suppressed. The rule is paralysis of all the segments, however many, in regular order from front to back, including even the anal segment if this boast of legs. By a fairly frequent exception the last two or three segments are spared. Another exception, but a very rare one, of which I have observed only a single instance, consists in the inversion of the dagger-thrusts of the second act, the thrusts being delivered from back to front. The caterpillar is then seized by its hinder extremity; and the Ammophila, progressing towards the head, stings in reverse order, passing from the succeeding to the preceding segment, including the thorax already stabbed. This reversal of the usual tactics I am inclined to attribute to negligence on the insect's part. Negligence or not, the inverted method has the same final result as the direct method: the paralysis of all the segments.