a hold of a transshipment vessel. The transshipped 
catches of individual seiners are kept separate, but an 
entire seiner’s catch, which is usually caught in several 
areas and over several months, is generally loaded into a 
single hold. 
Length-Frequency Samples 
The sampling procedures were virtually the same as 
those recommended by Hennemuth (1957) for the 
eastern tropical Pacific fishery, i.e., a stratified, by area 
and month, two-stage subsampling procedure (Cochran 
1963) was used. The boundaries of the areas (Fig. 1) were 
drawn according to the distribution and concentration of 
fishing effort of the American fleet. 
The first stage of the sampling was to choose the well 
(or hold if a transshipment vessel‘) to sample. The se- 
cond stage was to draw from the selected well a random 
sample of each species (yellowfin, skipjack, and bigeye 
tunas). Ancillary information, such as well number and 
catcher vessel name, was recorded for each sample. The 
date, location (NMFS area — Fig. 1), and tonnage of the 
catch sampled were obtained from logbooks after the 
samples were drawn. 
Before 1972, samples were obtained on an opportunis- 
tic basis and the sample size varied (10 to 300 fish). Since 
1971, a goal of 6 skipjack samples and 12 yellowfin 
samples of 50 fish each from each NMFS area and each 
fishing month was established in an attempt to ensure a 
more complete area-month coverage. The larger number 
of samples for yellowfin tuna was required because of the 
greater variability in sizes of this species (Hennemuth 
1957). This goal, however, was not attained in any of the 
years. 
Total Catch by NMFS Area-Month 
Total catch, by species and month, of tunas caught in 
the Atlantic by American vessels is tabulated annually 
by the IATTC from landing receipts. Logbook infor- 
mation on estimated catch by species, 1° area, date, and 
well number in which the catch is stored for each net set 
is also collected by the IATTC from virtually the entire 
American fleet. This logbook information was used to 
identify those seine sets that contributed to the catch in 
wells that were sampled and also to prorate the total 
catch by species of the entire fleet into catch by NMFS 
area-month strata. 
ANALYTICAL PROCEDURES 
Different areas apparently contain different sizes of 
fish, at least for yellowfin tuna in the eastern tropical At- 
lantic (International Commission for the Conservation of 
Atlantic Tunas 1974a), and the stratified sampling 
procedure, by month and area, was designed to account 
partly for this difference. Sizes of fish in the total 
‘It is not uncommon to find several species stored in a well. The fish are 
partially thawed in the wells before unloading at the canneries or onto a 
transshipment vessel. Measurements were made on partially thawed fish. 
a 
~ § 
on 
2 o 
Figure 1.—Map of the eastern tropical Atlantic where American 
seiners fish for tropical tunas (shaded area). Numbered statistical 
areas used in this report are shown. 
American catch were estimated from the stratified 
length-frequency samples and catch. 
When a sample is drawn from a well, complete infor- 
mation on date, location, and tonnage of catch is not al- 
ways available; complete information is obtained later 
from the ship’s logbook. Fish that may have been caught 
in different strata were, consequently, sometimes drawn 
in a single sample. These samples were not used in our 
analysis unless 75% or more of the tonnage in the well 
was caught in a single NMFS area-month stratum. 
Because of this rule, virtually all samples from trans- 
shipment vessels were not used in this study. Analyses 
currently underway which contain a special stratum for 
transshipment samples may lead to new procedures for 
utilizing much of the rejected samples. 
Weighting Factor 
The sample size was fixed and not proportional to the 
numbers of fish in the well. Each sample was therefore 
weighted by a factor (number of fish) based on the 
species tonnage and average weight of fish in the well or, 
when this was unavailable, on the species tonnage and 
average weight of fish in seine sets that contributed to 
the catch in the sampled well. 
Substituting Samples 
Samples were unavailable for all area-month strata in 
which the fleet caught fish. For strata without samples, it 
was necessary to make assumptions about the catch and 
substitute samples from adjacent strata, within year to 
estimate the sizes of fish caught by the fleet. Sub- 
stitution was on the basis of the following rules: 1) use 
same month and adjacent areas; 2) use same area and 
adjacent months; 3) use adjacent months and adjacent 
