The Castleton area contains an extensive and complex karst drainage sy
stem. Recharge is provided by allogenic stream sinks and by infiltrati
on into a soil covered autogenic catchment. Concentrations of the sani
tary indicator bacteria faecal coliform (FC) were measured weekly over
a 84-week period at three stream-sinks (P6, P7 and P8) and at two con
trasting springs (Russet Well and Peak Cavern Rising). Russet Well dra
ins the allogenic catchment, but also receives some autogenic recharge
whereas Peak Cavern Rising receives only autogenic recharge except at
high stage when it functions as an overflow spring for the Russet Wel
l system. Over the year as a whole and during each three-month season,
median FC concentrations at P6 were significantly higher than at Russ
et Well. The difference was greatest during summer/autumn and was lowe
st in winter/spring and it appears that FC concentrations at the risin
g are a complex function of faecal inputs and flow-through time. The r
elationship between FC concentrations at Russet Well and at Peak Caver
n Rising proved to be complex. Over the sampling period as a whole and
during the spring and autumn there was no significant difference betw
een median FC concentrations at the two risings; during the winter, wh
en discharge was highest, median FC concentrations at Russet Well were
significantly higher than at Peak Cavern Rising; and during the summe
r FC concentrations at Peak Cavern Rising were significantly higher th
an at Russet Well. The high FC concentrations in the sinking streams a
nd at the risings suggest that there could be a health risk to cavers,
especially during storm events and the summer. Furthermore, the fact
that the waters from both springs contain significant concentrations o
f FC bacteria indicates that even soil covered karat systems are unabl
e to filter out potentially harmful micro-organisms.