ered that produce shipments from Florida peak in the late May-
early June period. At the time of the March survey, Florida
produce and ornamentals movements were approximately 3,000
truckload-equivalents per week, a rate which had been main-
tained since early January and would continue virtually un-
changed for several more weeks (USDA, 1980-1983). The USDA
Agricultural Marketing Service (AMS) characterized truck avail-
ability as "adequate" to "slight surplus" (USDA, 1983). At the
time of the June survey, the shipment rate had risen to around
7,000 truckload-equivalents per week, and availability was
classified by the AMS as "slight shortage" or "shortage."
Therefore, in June the reservation price of a carrier would
have been greater than in March, even if the direct variable or
accounting costs associated with acquiring a load had not
changed.

 The positive parameter estimate associated with DAYLOSS
(50.83) has an absolute magnitude about six times that of its
standard error. This result lends strong support to DeVany and
Saving's contention that value of service pricing schemes can
evolve in competitively structured, unregulated markets. Again
however, it cannot be determined to what extent this is due to
costs associated with expedited service or to insurance. Em-
ploying the parameter associated with DAYLOSS, it is estimated
that for every 1,000 dollar increase in the average daily loss
per truckload PT rises by 118 dollars.7

 Not surprisingly, PT is strongly and positively associated
with distance. The estimated parameter associated with D
(1.856) is over six times the magnitude of its standard error.
The estimated parameter associated with DSQ (-.00003456) is
also large in absolute terms relative to its standard error.
The peak in the rate function across distance is at 2,685
miles, well beyond the 1,705 mile maximum length of haul in the
data. These results indicate a tapering rate/distance schedule
similar to that shown in Figure 1.