The bacterial composition of bulk tank milk from 13 farms was examined over
a 2-wk period to characterize sudden elevations in the total bacterial cou
nt referred to as "spikes." Bulk tank milk samples collected at each pick-u
p were analyzed for standard plate count, Petrifilm aerobic count, somatic
cell count, Gram-negative organisms, and streptococci. Twenty standard plat
e count spikes were observed: 12 associated with streptococci, 4 associated
with Gram-negative organisms, 2 associated with streptococci and Gram-nega
tive organisms, and 2 that were not definitively characterized. Spikes rang
ed from 14,000 to 600,000 cfu/ml. Streptococcus uberis was isolated as the
predominant organism from 11 spikes, and Escherichia coli was isolated from
4 spikes. Statistical analysis of total bacterial counts indicated a high
correlation (r = 0.94) between standard plate counts and Petrifilm aerobic
count. Regression analysis of standard plate counts and Petrifilm aerobic c
ounts yielded the equation log(10) (standard plate count) = 0.73 + 0.85log(
10) (Petrifilm aerobic count), indicating that the correlation, although st
rong, is not one to one. In a related pilot study, triplicate bulk tank mil
k samples were collected and analyzed for total bacterial count and presump
tive streptococcus, Gram-negative, and staphylococcus counts. Two-way ANOVA
of these triplicate data indicated a lack of significant variation among t
he triplicate samples, suggesting that one sample can reliably gauge the mi
crobial status of the entire bulk tank.