The structure of the shell of eggs of Pieris brassicae (L.), together with changes in it and associated membranes during embryonic development, have been investigated in relation to the penetration and toxicity of simple chemicals. The rigid outer shell consists of two proteinaceous layers, covered externally by a relatively hydrofuge cement, by which the egg is attached to the leaf surface. The egg has respiratory pores over its surface, and a single apical micropyle penetrating these layers. The inside of the rigid shell is lined with a layer of unsaturated oil—an unusual feature for an insect egg. When the egg is first laid, the vitelline membrane is directly applied to the inner surface of the solid shell over the region immediately around the micropyle, but within four hours this contact is broken, and the oil layer flows into this region also, and becomes complete. As development proceeds, the vitelline layer is replaced by membranes of embryonic origin, but before eclosion both these epembryonic layers, and also the oil, are resorbed.

The egg is remarkably resistant to water-soluble poisons which have no oil-solubility, except during the first four hours of development. This resistance is attributed almost entirely to the oil layer, and the early susceptibility to its absence over the micropylar region. These changes are not reflected in the effect of oil-soluble poisons or fumigants. The solid portions of the shell do not seem to be of great importance in restricting the entry of liquid poisons, even though the cement is comparatively hydrofuge; from experiments with wetting agents and with eggs immersed in poisons under vacuum, it does not appear that the respiratory air spaces in the shell are preferential channels of access; rather, the poisons penetrate through the solid portions of the shell. This penetration, even of oil-soluble materials, is slow, for they can be effectively washed out of the shell again, some considerable time after dipping. On the other hand, non-volatile oily materials can interfere with the respiration of the egg by blocking the air spaces in the shell.

The secretion of epembryonic layers does not appear to change the resistance of the egg to water-soluble materials; this is to be expected, for they do not contain lipoid. On the other hand they do add appreciably to the resistance to oil-soluble materials. There is no evidence that poisons are accumulated in these epembryonic membranes, and released during the pre-eclosion period. Experiments with covalent compounds, such as mercuric chloride, suggest that their oil-solubility accounts for their toxicity, whereas electrovalent compounds containing similar heavy metals are only effective while the direct micropylar path of entry is available to them.