In this paper, the general principles of lung growth already worked out in the rabbit (Short 1950) have been examined in the human lung. Using the same methods, a similar series of measurements have been made in embryonic and post-natal human lungs. Structural complexity has been measured in terms of internal surface area per unit volume of lung. Further support has been found for believing that the structural complexity of the lung is the result of tension in elastic fibres. At all stages of development a linear relationship similar to that found in the rabbit has been demonstrated between structural complexity of the lung on the one hand and a factor representing tension in elastic fibres (lung volume/interstitial volume) on the other. During the period of embryonic development, the structural complexity of the lungs of man and the rabbit is in very close agreement. It has also been found that the structural complexity per unit volume of the lungs at term in the mouse, rat, rabbit and man is the same. This suggests that the range of growth, between lung bud and term (responsible for the physical forces acting on elastic fibres and thus increasing the structural complexity), is the same in these four species. In man and the rabbit there are grounds for believing that the range of growth is indeed the same. Direct measurement of the volume of the lung rudiment at a stage when the tracheal bud has divided into two lumina in these two species shows them to be of the same order of size. Since the unit volumes at term are equal, the range of growth is equal in both species; a circumstance which accounts satisfactorily for the fact that the structural complexity at term is also the same in man and the rabbit. Calculation has also been made of the age at which formation of septa stops. In the two small species, the mouse and rat, the complexity of lung architecture probably increases throughout the post-natal period of growth. In the rabbit, differentiation is certainly complete by the third month of life, and in man is probably completed during the second year of life. Thereafter, continued growth of lung volume is accompanied by simple distension of the existing architecture. Further evidence bearing on growth and differentiation has been obtained from grossly hypoplastic lungs. In three cases of diaphragmatic hernia involving reduction of lung volume to one-sixth or less of normal size, structural complexity was found to be normal. Moreover, lobular size was also normal. Had the normal numbers of lobules been formed, their size must necessarily have been reduced. Therefore, reduction in total number of lobules and in total lung volume is not accompanied by structural hypoplasia of those lobules which are formed. The estimate of internal surface area per unit volume of lung in the adult mouse, rat, rabbit and man has been shown to increase as the smaller species are approached. On the other hand, the ratio of total internal surface area of the lung to body weight of these four species is probably to be regarded as a constant.