28 
Fishery Bulletin 120(1) 
external and internal reproductive organs following Pratt 
(1979). Reproductive terminology follows Pratt (1979), 
Hamlett (1999), and Hamlett and Koob (1999). Only the 
right ovary in the blue shark is functional (Pratt, 1979); 
thus, for all specimens, reproductive organs from the right 
side of the body were measured to the nearest millime- 
ter to ensure consistency. Each fish was examined fresh 
to assign a mature or immature (juvenile) status, and 
detailed conditions, such as foreign objects in the body or 
excessive thinness, etc., were noted. Mature females were 
further classified as gravid or postpartum when appropri- 
ate. Gravid females contained embryos, and postpartum 
females had evidence of recent past pregnancy, such as 
flaccid uteri or ovaries in the process of oogenesis. Juve- 
niles with lengths between the smallest mature and largest 
immature sizes (transitional range) were further examined 
given that these sizes were in the range of median size at 
maturity, and because of the gradual nature of maturation, 
it was difficult to ascertain their condition. 
Female blue sharks were visually examined for the 
presence or absence of mating scars or injuries. Internal 
measurements included the width (measured at the wid- 
est point) of the upper oviduct and oviducal gland, uterus 
width and length, and ovary width and length. The diam- 
eter of the largest oocyte was measured, and the presence 
or absence of the vaginal membrane (hymen) was deter- 
mined by insertion of a probe through the posterior end 
of the uterus into the cloaca; the latter was used as an 
indicator of prior mating activity. 
For males, maturity was externally assessed by exam- 
ination of clasper condition on the basis of degree of 
clasper calcification (fully calcified, partially calcified, or 
uncalcified), ability of the clasper to easily rotate around 
the base, and the ability of the rhipidion to open (Clark 
and von Schmidt, 1965). Clasper length was measured 
on the external side from the insertion of the pelvic fin 
to the tip of the clasper. Siphon sacs, which le between 
the skin and the abdominal musculature, were measured 
as per Natanson and Gervelis (2013). Internal measure- 
ments for males included epididymis width, ampulla epi- 
didymis width, and testis length and width. Presence or 
absence of spermatophores was noted, as was coiling of 
the epididymis. 
For sharks that did not undergo a full workup, maturity 
was assessed on the basis of visual examination following 
criteria from Pratt (1979). In females, the condition of 
the epigonal tissue encasing the ovaries, presence and 
size of oocytes, and the general appearance of the upper 
oviduct, oviducal gland, and uterus were examined. In 
males, the clasper condition (as described in the previous 
paragraph), the amount of the epigonal tissue surround- 
ing the testis, and coiling of the epididymis were exam- 
ined. The specimens were then classified on the basis of 
development of these characteristics. 
Measurements of the reproductive organs of both sexes 
were plotted against FL to examine the growth of the 
organs throughout ontogeny. Those specimens not assigned 
a maturity status at the time of dissection were later clas- 
sified on the basis of comparisons of organ measurements 
in relation to FL, comparisons with organ measurements of 
staged individuals, and detailed notes on condition taken 
at dissection. Non-staged specimens within the transitional 
size range could not be assigned a status by using measure- 
ments alone because of the overlap of immature and mature 
stages in this size range. For some samples, there was not 
enough information to confirm maturity stage; thus, none of 
these samples were used in the ogive analyses. 
Pratt (1979) assigned 3 maturity stages for female blue 
sharks: immature (46.0—145.0 cm FL), subadult (145.0— 
185.0 cm FL), and mature (185.0—300.0 cm FL). In the 
subadult phase, the organs necessary for copulation were 
developed, while those required for generation (such as 
oviducal gland and ovaries) were still developing. There- 
fore, for ogive analysis in our study, we considered imma- 
ture any samples that would be assigned to the subadult 
stage in his classification. 
Median maturity analysis 
Median FL and weight at maturity were calculated for 
both sexes by using maturity ogives fit to binomial data on 
reproductive maturity status. The probability that a given 
individual 7 was mature was modeled as the outcome of 
a Bernoulli random variable, where y, is 0 for immature 
individuals and is 1 for mature individuals: 
y; ~ Bernoulli(p;), (2) 
where p; = the probability that shark 7 is mature. 
To examine life history changes through time, samples 
were divided into 2 discrete time periods (TPs). Time period 
1 (TP1) corresponded to the data collected during 1971-— 
1977, data that were originally analyzed by Pratt (1979). 
Time period 2 (TP2) included specimens collected during 
2003-2016, primarily by L. Natanson. The gap between 
time periods is approximately 5 generations, allowing 
time for density-dependent changes in life history. 
We modeled p; as a function of size (separately in terms 
of FL and weight) and time period as follows, with the logit 
link function constraining p; to values between 0 and 1: 
logit(p;) = Bo + Bi perioa S1ZEi; (3) 
where f, = an intercept term representing the mean prob- 
ability that a shark is mature; and 
B, = the time-period-specific effect of size in terms 
of either FL or weight. 
Models were fit to data on each sex separately by using 
maximum likelihood methods with functions available in R, 
vers. 3.6.3 (R Core Team, 2017). Models were run separately 
by sex because it is well-documented that life history char- 
acteristics (e.g., size at maturity) differ between male and 
female elasmobranchs (Cortés, 2000). Model fit was eval- 
uated by using the Akaike information criterion corrected 
for small sample sizes (AIC,) (Akaike, 1973; Burnham and 
Anderson, 2002). The best fitting model was the model with 
the lowest AIC, value. The difference in AIC, between each 
model (A;) was calculated as A;=AIC, ;—AIC, in, Where AIC, ; 
is the AIC, for time period 7 and AJC. is the lowest AIC, 
c,min 
