'Pitted but not pitied' or, does smallpox make you small? -super-1

This article examines three propositions put by Leunig and Voth: that smallpox reduced stature irrespective of location, that stunting was most apparent among adolescents, and that these relationships were obscured in my earlier work by small sample size. It tests these claims by re-examining the original data-including the neglected Wandsworth data set-and questioning the meaning of the chosen method of graphical representation. Furthermore, and most fundamentally, the relationship between smallpox and stunting is advanced by adding new data on a further 34,310 prisoners. Using considerably larger data sets with many more juveniles, and refined definitions of rural and urban locations, this article confirms that the 'smallpox effect' varied by location, age, gender, and time period. That the relationship between smallpox and stunting was mediated through place and time suggests the role played by evolving urban conditions. The article offers a warning on the dangers of aggregating data without paying heed to important composition effects, and it argues that size does matter: the size of the smallpox effect, population size, sample size, and the size of the p-statistic. The reply concludes by again questioning the likely causes of stunting in the world's first great metropolis, London, arguing for the importance of examining chronic illness as a source of ongoing nutritional insult. Copyright Economic History Society 2006.