California. Dept. of Fish and Game.

California fish and game (Volume 54, no. 4) online

. (page 6 of 10)
Online LibraryCalifornia. Dept. of Fish and GameCalifornia fish and game (Volume 54, no. 4) → online text (page 6 of 10)
Font size
QR-code for this ebook

purpose of this paper is to present the fecundity data now available,
and to discuss their application to the problems of estimating adult


The data on anchovy fecundity were obtained from 19 females col-
lected off California or Baja California. Two were taken near the
entrance to San Diego Bay, March 29, 1951 ; 11 were from bait samples

1 Accepted for publication April 1968.

( 281 )



collected near San Diego, March 29, 1960; 3 were collected at CalCOFI
station 90.30. in tlio soutliorn Ranta Catalina Channel, Jnly 20, 1953;
and 3 at station 113.35, 55 nantit-al miles due north of San Benitos
Islands, Baja California, September 21. 1952. The fish were preserved
in formalin at the time of collection. Tlu^ data, recorded after preserva-
tion, include standard length, Aveight, ovary weight, and number and
size of eggs (Table 1). The methods used were similar to those of
]\IacGregor (1957). Comparative data for the anchoveta, Cetcngroulis
mysticctus, are from Peterson (1961) and for Anchoa jiaso from seven
females collected ot^' Baja California.


Measurements and Counts Related to Fecundity of Engraulis mordax

and Anchoa naso.





Advanced eggs


Standard length (mm)


of fish


Engraulis mordax

97 ... -















. 054







. 030

























.00 -.84

106 .

.52-. 66


108 ..

.66-. 76
.68-. 80


.58-. 70


.68-. 86


.54-. 62


.68-. 82


.68-. 80


.64-. 78


121 ...

.56-. 78
.52-. 66



.64-. 82
.70-. 80


.72-. 86




Anchoa naso


.50-. 66
.50-. 58
.62-. 78

.30-. 42

52 _

.34-. 46

58 ..

.32-. 40


.30-. 40


..34-. 42


.34-. 44


.34-. 44


The gonads of E. mordax and A. naso are elongate, tapered pos-
teriorly, and not fused. Both the shape and lack of fusion seem to be
typical of clupeoids. Since the gonads are the last organs to develop,
and because no special space is provided for them, they tend to crowd
into whatever space is available. They assume the general shape of the
body cavity and are contoured on one side by the body wall and on
the other side by the other internal organs.


The gonads of C. mysticetus differ from those of the other two species
in being tapered anteriorly (this species has a deeper body and a larger
body cavity posteriorly) and joined posteriorly.

The gonads of A. naso and E. mordax resemble those of most fish
species in that the left gonad is usually larger than the right. In
E. mordax, the left gonad composed 49 to 62% of the total gonad
weight for 29 females and 46 to 62% for 13 males; in A. naso, the
percentages were 48 to 65 for 7 females and 50 to 60 for 6 males. The
average for each of the four groups was between 55 and 56%). No data
are available on the comparative size of gonads of C. mysticetus.

While most fish eggs are spherical, engraulid eggs are more or less
oval. The length of the planktonic eggs of 10 species of engraulids from
tlie Gulf of Panama (Simpson, 1959) ranged from about 1^ to 2f times
their width. Bolin (1936) described the planktonic egg of E. mordax
as ellipsoidal, 1.23 to 1.55 mm long by 0.65 to 0.82 mm wide, with a
segmented yolk, no oil globules, and no sculpturing on the membranes.
The latter three characteristics were present in all engraulid eggs
sampled by Simpson.

The formalin-preserved ovarian eggs of E. mordax are transparent
and spherical up to a diameter of 0.14 mm, but larger eggs are elongate.
The eggs are translucent and darker at 0.20 mm because of yolk forma-
tion, and become opaque as size and yolk content increase. The largest
ovarian eggs I found were 0.86 mm long.

The eggs of A. naso are spherical throughout their ovarian develop-
ment. Yolk begins to form in eggs of 0.10-mm diameter. The largest
ovarian eggs measured 0.46 mm. Among the 14 species of engraulids
examined by Peterson (1956), A. naso was one of three that had spheri-
cal ovarian eggs ; the others had oval eggs. The egg developing within
the ovary has no perivitelline space and has the same shape as the yolk.
After spawning, the yolk, enclosed in the vitelline membrane, appar-
ently retains the same shape — either spherical or more or less ovate —
that it had within the ovary, but tlie shell of the egg becomes ovate.
Among the engraulid eggs in plankton figured by Simpson (1959),
some oval eggs had oval yolks and others had round yolks.

The ovarian eggs of the anchoveta are oval. Peterson (1961) reported
ovarian eggs of this species up to 0.66 mm long, while Simpson found
that planktonic eggs average 1.17 mm long and 0.56 mm wide.


The number of eggs per spawning is assumed to equal the number
of eggs in the most advanced modal group in the ovaries. As in other
fishes, the number of advanced eggs tends to increase as fish size in-
creases. Generally the number of eggs is proportional to the weight of
the fish and to the cube of the standard length. However, in many
species egg production occurs over such a relatively short range of fish
lengths, and variation in tlie number of eggs produced at each length
is so great, that the relationships of length and weight to egg production
are masked. As a rule of tlumib, the curvilinear relationship between
standard length and number of eggs produced is not apparent unless
the largest fish in the sample is more than twice as long as the smallest


Peterson (1961) fitted a least squares straiglit line to his length-
fecundity data for the anchoveta, E = 101,200 + 1.0006L, in which
E — the number of ova in the spawning mode and L =z standard
length in mm. His largest fish (1G2 mm) was only 1.4 times the length
of the smallest (115 mm), and he found no indication of a significant
departure from a linear relationship.

In my sample of 19 anchovies, the length of the largest fish (138
mm) also was only ].4 times the length of tlie smallest fish (97 mm) ;
the straiglit-line relationship is E = 2,175 + 7.700L. The parabolic
curve based on the cube of length is ^ = 0.00646L^, and that based on
the log-log transformation is E = 0.000574L2-^. The respective cor-
relation coefficients are 0.8] C. 0.820, and 0.830, which, though highly
significant, are not significantly different among tliemselves. If the
range of lengths of the fishes could be increased, the curvilinear rela-
tionships should become progressively better than the straight-line

Although in the present samples the correlation coefficient from the
log-log transformation is slightly better than that based on the cubic
relation, the latter will often yield a lower variance and higher cor-
relation coefficient. This relation also holds for the length-weiglit curve.
The log-log line heavily weights lower numbers in its transformation,
resulting in a parabolic curve that is only a rough and often mislead-
ing approximation of the best fitting curve. If spawning occurs over
a restricted range of fish lengths, the straight-line relationship is prac-
tical, whereas if the length range is greater, the cubic relationship is
generally satisfactory.

For some purposes, weight-fecundity relationships are more useful
than length-fecundity. An understanding of weight-fecundity relation-
ships may most easily be obtained by determining tlie number of ad-
vanced eggs produced per unit of fish weight. The mean numbers of
advanced eggs per gram of fish were 574 for 19 E. mordax, 836 for
86 C. mysficctus, and 326 for 7 A. nnso.

The number of eggs per g of fish does not appear to change witli size
of E. mordax; therefore, the weight-fecundity relationship may be
described hy E ^= 574 W in which E = the estimated number of ad-
vanced eggs in the ovary and IF =r fish weight in g.

l*eterson (1961) fitted a least squares straight line to his anchoveta
weight-fecundity data and obtained E = 3,304 + 927 W. The large
negative y-intercept indicates tliat the number of eggs is not propor-
tional to fish weight but gradually increases witli increasing fish size.
A comparison of eggs per g of fish for the various sizes of anchovetas
does not bear out this indication. The 13 smallest fish, 115 to 120 mm
SL, have a mean of 533 eggs per g of fish (standard error of the
mean = 35) and the remaining 73 fish, 121 to 162 mm sii. have a mean
of 890 eggs per g of fish (standard error of the mean = 26). The
number of eggs per g of fish does not change significantly among these
remaining 73 fish; the mean is 927 eggs for the 13 smallest (121 to 128
mm) in this group and 874 for the 13 largest (153-162 mm). If the
fecundities of these two groups are representative of the size groups
in the population, it might be more practical to use the formulae
E = 533 W for fish up to 120 mm sl and E = 890 W for larger fish.


The smaller specimens in a number of fish species have a lower
fecundity, but the increase in fecundity with size tends to be abrupt
rather than gradual. In A. naso, the five smaller specimens (50 to 76
mm) averaged 209 (180 to 254) eggs per g of fish, whereas the two
larger specimens (78 and 81 mm) averaged almost three times as many,
616 (546 and 686).

Although only 19 anchovies were used to determine fecundity, the
resulting data should be satisfactory for census purposes if the sample
is representative of the population. Eggs per g of fish were almost
identical for different areas (Santa Catalina, San Diego, San Benitos)
and different seasons (March, July, September) as well as for different
sizes of fish. Therefore, in spite of the small sample, I believe that it
is representative of the northern anchovy population.

The mean number of eggs per g of fish, 574, has a standard error of
30. The 95% fiducial limits are ± 60 eggs per g or ± 10%. For the
much larger sample of anchovetas, the 95% fiducial limits are ± 54
eggs or about ±


To determine the biomass of the adult anchovy population from egg
census data, we must know the sex ratio of the population by weight,
the average number of times the population spawns during the egg
census period (1 year for the present study), and the number of eggs
per spawning.

Of 3.500 anchovies sampled during the 1952-53 and 1953-54 seasons
(for which length and sex are available), 1,545 were males and 1,955
were females (Miller et al., 1955). The number of males equaled 79%
of the number of females. The males averaged 138.5 mm sl and the
females 142.2 mm sl. If we assume that there is no great difference in
form between male and female anchovies then, mean sl^ of females is
to mean sl^ of males as mean weight of females is to mean weight of
males. From this we can determine that the mean weight of males was
about 92% of the mean weight of females. Multiplying percentage
weight by percentage numbers gives a male biomass equal to 73% of
the female biomass.

Weight and sex data are available for a sample of 336 anchovies
taken from the San Francisco area between September 1952 and
February 1953. The 181 females averaged 149.5 mm sl and weighed a
total of 6,652 g; the 155 males averaged 145.3 mm sl and weighed a
total of 5,172 g, or about 78% of the total weight of the females.

From these data we may assume that the biomass of male anchovies
available to the fisliery equals about 75% of the female biomass. If for
one spawning a female produced 574 eggs per g of body weight.
5.2 X 10^ eggs would be produced per short ton of female biomass ;
therefore, we should expect the same number of eggs per 1.75 short tons
of adult anchovies or about 3.0 X 10^ egt?s per ton of adult anchovies.

The number of times an anchovy spawns in one season is not known.
There has been some speculation that the prespawning ratio of small
yolked eggs to large yolked eggs in the ovary is an indication of the
number of spawnings that will take place in that spawning season. That


is, if tlie ratio is approximately (nic lo one at the beginniiifj of tlie
spawiiiii<r season, tliore will be two spawiiin<i:s ; if it is two to one, tliere
will be three spawnings, etc. As the spawning season progresses, the
ratios would decrease by one after each spawning. This sjx'culation is
not supported, however, by my data for anchovies (nor b}^ iiiudi more
extensive unpnblished data for sardines). Ratios ajiparently are not
related to size of fish, area, or season. Eatios for individual fish ranged
from 1.3:1.0 to 2.4:1.0. The mean ratios for fish in the pi-esent collec-
tions were 2.2: 1.0 (San Diego, March 1951). 1.9: 1.0 (San Diego, March
l!)(i0), 1.7:1.0 (Santa Catalina Channel. July 1953), and 1.8:1.0, (San
Benitos Islands, November 1952).

The anchovy with the 2.4:1.0 ratio was most nearly ri])e, and had
eggs u]) to 0.86 mm long. It was also the only fish that exhibited a dis-
tinct trimodal distribution of yolked egg sizes. The ratios of the nund)ers
of eggs in each mode, from smallest eggs to largest, were 1.4:1.0:1.0.
If the largest eggs were omitted, the remaining disti-ibiition would
closely approximate the lower ratios in anchovies in wliich eggs in the
most advanced group were smaller. In addition, a number of the female
anehovies examined could not be used for fecundity estimates beeaus(>
they contained only small, yolked eggs with a miimodal size distribution.
The eggs in a few of these fish appeared to be undergoing resorption.

The ovaries offer no evidence of a fixed iiuhiImm- of spawnings. The
o^<X ratios suggest that the anchovy may spawn one, two, three, or more
times per year; the eggs remaining after a given si)awning might either
be resorbed or be replenished from the reserve of non-yolked eggs. If
replenishment does occur, the ratios are even less meaiiiii'_;ful.

In 19 anchovetas, the eggs in the advanced mode avt'raged more than
three times as many as the smaller yolked eggs (Peterson, 19()1). How-
ard and Landa (1958) noted the paucity of smaller eggs in this species
and concluded that only one batch of eggs was spawned, following
which the smaller eggs were resorbed. The short spawning stMson of
the anchoveta also supports the idea of a single spawning. Almost all
anchoveta eggs taken in plankton tows in the Gulf of Panama were
collected in November and December, and neither eggs nor ripe fish
could be found from mid-January to Oetober (Simpson. 1959).

In A. nnso, ratios of small yolked eggs to large yolked eggs
ranged from 0.0: 1.0 to 4.2: 1.0. Fish with develojiinfl: ovaries have been
taken in Oosta Pica in August, Septembei', and -lauuary (Peterson,
1956), and my two samples from Baja California were taken in July
and November. A protracted spawning season and multiple s]iawning
are indicated.

The spawning season of the uorthern anchovy also is proli'acted;
some spawning occurs in every month, and fish with developing: ova may
be found at any time of year. Nevertheless, there is a d(>finite annual
l^eak in spawning, as shown by the numb(M- of anchovy larvae tak(Mi
in ])lankton tows off the California and l>a.ia California coasts during
the 7 years 1951 throuirh 1957 (Ahlstrom 1953, 1954, 1958, 1959;
Ahlstrom and Kramer, 1955, 1956, 1957). Peak spawning occurred in
March (4 years). February (2 years), and January (1 j^ear),
and spawning during the peak month included 20 to 33% of the total
spawning for each year.


During the early part of the year, almost all of the mature females
in the population contain well-developed eggs or are recently spent ;
consequently, later spawnings must represent repeat spawnings by at
least part of the population, although younger, late-maturing females
may contribute. If the average time required to mature a second batch
of eggs after spawning of the first batch was one month, if all adult
females spawned in the peak month, and if 20 to 33% of the year's
spawning took place in the peak month, an average of three to five
spawnings per year would be indicated. The percentage of annual
spawning ranged from 34 to 49 for the 2 consecutive highest months
over the 7-year period. If the spawning interval were 2 months, an
average of two or three spawnings would be indicated.

The hake, Merluccius productus, which occurs in the same area and
has a peak spawning period very similar to that of the anchovy, spawns
only once a year (MacGregor, 1966). Larval occurrence indicates that
18% of the hake spawning takes place in January, 36% in February,
35% in March and 10% in April, with the remaining 1% scattered
through the next 8 months. The percentage of hake spawning for the
four peak months is about double that of the anchovy. If egg develop-
ment and spawning are otlierwise similar, but the hake spawns only once
a year, we might postulate that the anchovy spawns an average of
twice a year.

If one spawning produces 3 X 10^ eggs per short ton of spawning
fish, two spawnings would produce 6 X 10^, three spawnings 9 X 10^,
and four spawnings 12 X lO'* eggs per ton. If we arbitrarily estimate
the average number of spawnings a year to be two and a half, and
the actual number was only one, then the actual biomass would be
250% of our estimated biomass; if the actual number of spawnings
was two, actual biomass would be 125% of that estimated; if three,
83% ; and if four, 63%. Not knowing the average number of spawn-
ings per year can introduce considerable error into the calculations.

The number of eggs spawned per unit of fisli weight for one spawn-
ing and biomass ratio by sex probably change so little from yenv to
year that data obtained in one year could be used for any other year.
The number of spawnings, however, could change from year to year,
depending on the size composition of the spawning stock or on hydro-
graphic conditions. Anotlier factor affecting calculations would be the
change in biomass that results from the maturation of juvenile fish
and mortality of adults during the year. Biomass estimates would be an
annual mean A'alue weiglited to heaviest periods of spawning.

Probably a better way to estimate the adult population on the basis
of egg counts would be to sample intensively during a period of peak
spawning. If we could determine, by sampling, that portion of the adult
population which had spawned once during one or two of the peak
spawning months of a given year, we could calculate the average
biomass existing at that time from an egg census covering the same
period. This procedure largely would eliminate problems arising from
weighted annual biomass estimates, multiple spawning, and changes
in number of spawnings from year to year.



Tlio number of advanced eg^s produced per unit of fish weifrlit was
574 eggs per g of female for E. mordax, 886 for C. ninsticctns. and
320 for A. naso. Altliougli the number of eggs per gram of fisli did
not cliange witli size of fisli for E. mordax, it was considerably liigher
for larger specimens of eaeli of llic other two species.

The biomass of male anchovies in California commercial landings
equaled about 75% of the biomass of females. Therefore, on the basis
of one spawning, 3.0 X 10^ eggs would be produced per short ton of
spawning adult anchovies. The number of times a female anchovy
spawns in a season cannot be determined from presentlx- axailablr


Ahlstrom, Elhcrt H. 19r)8. l'ilcli;u(l <'gfrs and larvno .iiul hIIht f'lsli larvae. Pacific
Coast— 19.")!. U.S. Fish Wildl. Serv.. Spec. Sci. Kept- F'^^l'- (102) : 1-."..

. 1954. Pacific sardine Cpilchnrd) ejrs« <Tid larvae and other fish larvae.

Pacific Coast— 19.12. U.S. Fish Wildl. Serv., Spec. Sci. Kept.., (123) : l-7(i.

. 1958. Sardine eggs and larvae and other fish larvae. Pacific Coast. 19."i(i.

U.S. Fish Wildl. Serv.. Spec. Sci. Kept., Fish., (251) : 1-S4.

1959. Sardine eggs and larvae and other fish larvae. Pacific Coast, 1957.

U.S. Fish Wildl. Serv., Spec. Sci. Kept., Fish., (32S) : 1-99.
Ahlstrom. Elhert H., and David Kramer. 1955. Pacific sardine (pilchard) eggs

and other fish larvae, Pacific Coast, 1953. U.S. Fish Wildl. Serv.. Spec. Sci.

Rept., Fish., (155) : 1-74.
. 195G. Sardine eggs and larvae and other fish larvae. Pacific Coast. 19.54.

U.S. Fish Wildl. Serv., Spec. Sci. Kept., Fish.. (ISd) : 1-79.

1957. Sardine eggs and larvae and other fish larvae, Pacific Coast, 1955.

U.S. Fish AYildl. Serv.. Spec. Sci. Kept., Fish., (224) : 1-90.
Bolin, Rolf L. 19.".(i. Embryonic and early hnrval stages of the Califoniiii .incliovy

Engraulis mordaw (Jirard. Calif. Fish and Game, 22 (4) : 314-321.
Howard, Gerald V., and Antonio I.anda. 1958. A study of the age, growth, sexual

maturity, and spawning of the anchovetta (Ceiein/ratilis }in/sticctus) in the Gulf

of Panama. Inter-Amer. Trop. Tuna Comm.. Bull., 2 (9) : 389-437.
MacGregor, .Tohn S. 1957. Fecundity of the Pacific sardine ( i^ardiiiops carnilra) .

U.S. Fish Wildl. Serv., Fish. Bull.. 57 (121) : 427-449.
. 19(56. Fecundity of the Pacific hake, .l/cr/f/cr/i/.v produrlii^ (Ayres). Calif. and Game, 52 (2) : 111-1 IG.
Miller, Daniel ,T., Anita E. Daiigherty, Frances E. Fclin. ;iiid .lolm .MacCJregor.

1955. Age and length composition of the northern anclio\y catch off the coast of

California in 19.52-.53 and 19.53-.54. Calif. Dept. Fish and Game. Fish Bull..

(101) : 37-66.
I'eterson, Clifford Ij. 1956. ( thscM-vations on the taxonomy, biology, and ecology

of the engraulid and clupeid fishes in the Gulf of Xicoya, Costa Rica. Inter-Amer.

Trop. Tuna Comm., Bull., 1 (5) : 139-211.

1961. P'ecundity of the anchovetta (Ccfoiarnulis wysficrtus) in the Gulf

of I'anama. Inter-Amer. Trop. Tuna Comm., Bull., 6 (2) : 5.5-62.

Simpson, .John G. 1959. Identification of the egg, early life history and spawning
areas of the anchovetta. ('rlcnurdiilis iiii/slicrius (Gunther). in the (Juif of
Panama. Inter-Amer. Trop. Tuna Comm., Bull., 3 (10) : 441-538.

Calif. Fish and Game, 54(4) : 2,S!)-liD(;. I'.JGS.





Wildlife Management Branch

California Department of Fish and Game

A survey was made in 1965 to obtain a better understanding of des-
ert bighorn (Ovis canadensis) dependence upon water in the Santa Rosa
Mountains during the summer. Twenty bighorn were mariced at one
spring by using Hansen's dye spraying device. The nontoxic dye could
be seen on the horns of the sheep for 1 Vz months. Many of the ewes,
lambs, and young bighorn stayed within three-quarters of a mile of
waterholes during July and August. Rams in their prime or older trav-
eled as far as 3 miles from water and made fewer trips to water during
July and the first half of August. Optimum water distribution is at inter-
vals of 2 miles. Bighorn waterhole count data are evaluated.


A primary factor limiting bighorn sheep populations in the Santa
Rosa Mountains of California is lack of water. Jones, Flittner, and
Gard (1957) found that bighorn were restricted to within 1 mile of
water during the summer. The present study was conducted to obtain
more information on summer movements, frequency of trips to water,
and distance traveled to water. It was also hoped to get a better evalua-
tion of the waterhole counts regularly conducted by the Department
of Fisli and Game.

The desert slopes of the Santa Rosa Mountains above Palm Desert,
Riverside County, are ideal for such a study because of the convenient
location of roads, which made it possible to cover much of the study
area on foot in 1 day. We had a good knowledge of the summer big-

1 2 3 4 6 8 9 10

Online LibraryCalifornia. Dept. of Fish and GameCalifornia fish and game (Volume 54, no. 4) → online text (page 6 of 10)