The pendulum of interpretation may now be swinging back somewhat in favour of capital. Further estimates of capital formation have been laboriously assembled and those recently published by Feinstein for Britain are not in accordance with the Deane and Cole interpretation which had received growing acceptance during the 1960s. Feinstein's data, which will be discussed later, point to the investment ratio rising sharply during the eighteenth century from about 8% in the 1760s to 14% in the 1790s, almost doubling in a Rostow-Lewis manner. However, thereafter, with the exception of a dip during the Napoleonic Wars, investment's absorption of income in proportionate terms did not change and was not lifted to a new higher level by the coming of the railway. Recent estimates constructed by Lévy-Leboyer show a somewhat similar picture with the share of French national product invested increasing from about 7% in the 1810s to 12% by the 1850s and then remaining at this level for the next four or five decades. However although the Rostow-Lewis view may now be finding new foundations, the stress placed on the investment ratio has to be viewed in terms of a wider perspective. One reason for the appreciable rise in the ratio during the onset of industrialization is that at this stage in an economy's modernization, capital goods are relatively expensive compared with consumption goods. Initially there is no machine-producing industry and once producer goods have evolved beyond the point of being cheap, crude and self built, there may occur a period, albeit brief, when they are relatively costly. Subsequently the growing demand for capital goods will lead to the emergence of specialist builders. In addition to the play of specific factors, it may be important to turn to the movement of the consumption ratio. A rapid doubling or tripling of the investment ratio, partly caused by supply inelasticies, appears dramatic, an event consistent with a revolutionary interpretation of industrialization. Even though it may occur, it is accommodated by a small decline in the share of consumption, of the order of not much more than 10% during a period when income is continuously rising in the medium term relatively to population growth. Perhaps what is more important is not the comparison of the investment ratio at the beginning and end of the stage of the onset of industrialization but rather changes in its very long term behaviour.