In the first of the two explosions the three towns were almost equally attacked per head of their populations; in the second explosion, in mid-winter, Darlington had relatively only half as many cases as each of the other two, which had about the same number of cases as in the former six-weeks’ period. In both periods, of six weeks each, the three towns had together 1334 cases of typhoid, while the country districts near them had a mere sprinkling. A flooded state of the Tees appeared to be a relevant antecedent to each of the explosions. The Tees is a broad shallow river flowing rapidly, subject to frequent inundations, tortuous in its lower course, forming at its mouth, where Middlesborough stands, a wide estuary bordered by low flat grounds. The rainfall at Middlesborough was 6·3 inches in August, of which 2·2 inches fell on the 12th of the month, the river being high in flood thereafter. There were again high floods in November, chiefly caused by the melting of snow in the upper basin (5 inches fell at Barnard Castle in November, 3·1 inches at Middlesborough, while the December fall was 1·2 inches at the former and 1·4 inches at the latter). To apply correctly the ground-water doctrine of enteric fever to these explosions, other particulars would have to be known, more especially the extent of the previous dryness of the subsoil (the rainfall at Middlesborough was 9·3 inches in the first half of 1890, 15·6 in the second half, and below average for the whole year). But the flooded state of the Tees valley in August and November must have changed abruptly the state of the ground-ferments within the areas of the respective towns and so afforded, according to the general law, the conditions for an abrupt increase of enteric fever in these its endemic or perennial soils[408].