The flotation procedure for the detection of Cryptosporidium parvum and Cry
ptosporidium muris oocysts in feces was adapted for use on soil samples. So
il samples were seeded with known amounts of purified C. parvum or C. muris
oocysts and Cryptosporidium spp.-free bovine feces. The limit of detection
for this procedure was determined at different levels of inoculation for e
ach species. At each level of inoculation, 30 control samples were processe
d and the observer was blind to the status of the sample. All samples were
examined for the presence of Cryptosporidium spp. oocysts using phase-contr
ast microscopy. The samples were seeded with the following estimated counts
of C. parvum oocysts: 1000/g, 1250/g, and 1500/g. These levels had sensiti
vities of 88%, 90%, and 93%, respectively. All inoculation levels had a spe
cificity of 100%. Thirty additional samples were inoculated with C. muris a
nd the limit of detection was found to be 76 oocysts/g sample, with a sensi
tivity of 97% and a specificity of 100%.