Content uploaded by Martin A. Hall
Author content
All content in this area was uploaded by Martin A. Hall
Content may be subject to copyright.
An ecological view of the tuna±dolphin
problem: impacts and trade-offs
MARTI
ÂNA.HALL
Inter-American Tropical Tuna Commission, 8604 La Jolla Shores Dr., La Jolla, CA 92037, USA. E-mail:
mhall@iattc.ucsd.edu
Contents
Abstract page 2
Introduction 3
Brief history of the ®shery 3
Types of purse seine sets 4
Stocks of dolphins 5
The association of tunas and dolphins 6
The tuna±dolphin problem 8
Estimation of incidental dolphin mortality
The early years
The Marine Mammal Protection Act of 1972 and its consequences
The internationalization of the ®shery
The Agreement for the Conservation of Dolphins and the International
Dolphin Conservation Program
The Declaration of Panama
Databases available
Mortality component variables
Estimation of dolphin abundance
Evolution of the ¯eet: technology and training
The ecological issues 12
Impact of the dolphin mortality on their populations
Ecological impacts of the ®shing operations
Factors to assess the ecological impacts of different ways of ®shing for
yellow®n tuna
Maximization of yield per recruit
Maximization of reproductive rate
Minimization of discards
Factors to assess the impact of ®shing on the rest of the ecosystem
Bycatches
Impacts of ®shing operations on the habitat
Impacts of lost or discarded gear
Generation of pollution and marine debris
`Subsidies' to some species
Alternative ways of purse seining and other ®shing methods 23
0960±3166 #1998 Chapman & Hall
Reviews in Fish Biology and Fisheries 8, 1±34 (1998)
Remain in the area and switch to purse seining on school®sh or logs
Move to other oceans and ®sh on school®sh or logs
Change gear and remain in the region
Switch to pole-and-line ®shing
Switch to longlines
Switch to gill nets
Remain in the area and develop new technologies
Conclusions 28
Acknowledgements 29
References 29
Abstract
After a brief description of purse seining and the other methods used to catch yellow®n
tuna in the eastern Paci®c Ocean, some considerations are made on the tuna±dolphin
association and the solution of the problem of dolphin mortality in the eastern Paci®c.
The association has been observed in other ocenas, but the frequency of setting in the
eastern Paci®c is much greater.
The mortalities of dolphins through ®shing have declined from about 133 000 in 1986
to around 2600 in 1996. The impact of recent levels of mortality on the dolphin
populations is not signi®cant from the population point of view. The mortality levels for
all the stocks are less than 0.1%, much lower than the 2% value used as a conservative
(low) estimate of net recruitment. All dolphin stocks have population sizes between
400 000 and 2 200 000, and most have remained stable for a decade or more.
Fishing operations can cause ecological impacts of different types: bycatches, damage
to the habitat, mortality caused by lost or discarded gear, pollution, generation of
marine debris, etc. A brief discussion follows, with a more detailed look at the
bycatches. For convenience we can separate the effects of the ®shery on the target
species, and on other species. Of the different ways of purse seining for tunas, sets on
dolphins catch tuna close to the optimum size to maximize yields and to allow for
reproduction, and result in discards of tunas of less than 1% of the catch. Sets on logs
catch small tunas, and result in the highest tuna discards (20± 25%). School sets fall in
the middle from the point of view of the sizes caught. Obviously, from the ecological
point of view, sets on dolphins are the best way to harvest yellow®n tuna.
After a discussion of the different ecological impacts a ®shery can cause to other
species on the habitat, a comparison is made of the bycatches generated by the different
types of purse seine sets. Bill®shes, sharks, mahi-mahi, wahoo and sea turtles are taken
as incidental catches by purse seiners. Log sets produce, by far, the largest bycatches,
followed by school sets and dolphin sets in that order. The bycatch levels in log sets are
usually tens to hundreds of times those in dolphin sets. The difference can be attributed
to the selection caused by the speed of movement of the tuna± dolphin group (slow-
moving species or individuals cannot keep up with the group), an effect that may be
magni®ed by the chase that precedes the dolphin sets. Log sets, on the other hand, are
made on a drifting community. The alternatives left to the ®shers if they were forced to
switch from the current ®shing methods to others are brie¯y discussed, considering their
feasibility, and comparing their ecological costs.
From the ecological point of view, and considering that the dolphin mortality is
2Hall
clearly sustainable, the impacts caused by the other types of sets, especially log sets,
could be more signi®cant than those caused by the dolphin sets. Some of the species
taken in log sets are endangered (e.g. sea turtles), others have unknown status and
potential vulnerability because of their low reproductive and juvenile survival rates (e.g.
sharks). Overall, the biodiversity of the eastern Paci®c appears to be better preserved by
a ®shery directed to dolphin sets than the other alternatives proposed for the purse seine
and for other gears.
Keywords: bycatch, dolphin, tuna
Introduction
The American purse seine ®shery for tunas in the eastern Paci®c Ocean started in the late
1950s and largely replaced the pole-and-line ®shery that had been operating for decades.
The new ®shery had much higher catch rates, a broader range of operations, and other
characteristics that made it very successful from the point of view of increasing tuna
catches (Francis et al., 1992). However, this new ®shery had an unwanted consequence:
often the schools of tunas were detected, and eventually encircled together with large
herds of dolphins. As the ®shers did not have the sophisticated gear or techniques needed
to release the dolphins, many of them were incidentally killed in the operations. When
the public became aware of the magnitude of this mortality (Perrin, 1968, 1969), the
outcry was one of the driving forces behind the passage of the Marine Mammal
Protection Act of 1972 by the US Congress. The level of dolphin mortality during the
1960s was estimated to be several hundreds of thousands of animals per year (Smith,
1983; Lo and Smith, 1986; Wade, 1995), but the data available were far too limited to
provide precise estimates (Smith and Lo, 1983). The mortality was not sustainable, and
most dolphin populations declined until the late 1970s (Anganuzzi and Buckland, 1994).
By 1996, however, incidental dolphin mortality had been reduced to close to 2500, and
the population decline had been stopped. In recent years, some sectors of the
environmental community had pushed the US and other governments to ban all ®shing
for tunas associated with dolphins. If the ®shery were to switch to alternative ways of
®shing, dolphin mortality in the eastern Paci®c might decline even further, but other
unwanted consequences of ®shing, such as discards and reduced yields per recruit of
tunas, and bycatches of other species, would increase considerably. The objective of this
review is to bring together the known impacts and trade-offs that would accompany this
change, and to try to compare, from the ecological point of view, the different
alternatives available to the ®shers and to the managers of the ®shery in their pursuit of a
solution that allows a rational use of the tuna resources while providing for adequate
protection for the dolphins and other components of the ecosystem.
Brief history of the ®shery
Fishing for tunas has taken place in the eastern Paci®c Ocean since early in the 20th
century. The main targets are the yellow®n tuna, Thunnus albacares, the skipjack tuna,
Katsuwonus pelamis, and the bigeye tuna, Thunnus obesus (all from the Fam.
Scombridae). Previous to about 1960, most of the vessels were pole-and-line vessels,
operating from California ports. They used live bait, and stayed mostly in coastal waters
Tuna±dolphin problem 3
because of limitations imposed by the need to renew the bait when exhausted. A new
technology resulting from several technical developments, the purse seine, began to
replace the bait-boats in the late 1950s (Alverson, 1960; Broadhead, 1962; Green et al.,
1971; Cole, 1980). Purse seine vessels also ®shed principally off Baja California and
California in the early years of this ®shery (Shimada and Schaefer, 1956). The success of
the purse seiners (higher catch rates, independence from the use of bait) led to the
construction of more and bigger vessels, that expanded rapidly the range of the ®shery to
the southern and offshore areas. They could at sea for longer periods and they were
capable of operating far offshore. These vessels could surround the school of tunas with a
wall of netting about 1.6 km long and 200 m deep, and when the circle is completed, a
cable passing through rings at the bottom of the net is pulled aboard the vessel, closing the
lower part of the net, forming a `purse'. Each of these operations is called a set. Purse
seining, as it is currently practised, is described by Ben-Yami (1994) and Sainsbury (1996).
Most of the tunas caught in the purse seine ®sheries are used by the canning industry.
Tunas are also caught in the eastern Paci®c Ocean by longline vessels (Nakano and
Bayliff, 1992). Lines of up to 120 km long with baited hooks are deployed at different
depths, depending on the preferred targets. These vessels, most of which are based in
Japan or Hawaii, catch several species of tunas, including yellow®n, and also bill®shes
(Fam. Istiophoridae and Xiphiidae) and sharks. The yellow®n caught are large (modal
length 130±140 cm, modal weight 46± 58 kg in 1981 ±1987; Nakano and Bayliff, 1992),
but the catch rates of longliners are much less than those of purse seiners. There are
several reasons for the difference in ef®ciency: (i) purse seines are directed at, and take,
whole schools of tunas while longlines are passive and take the ®sh individually; and
(ii) longliners' main target is bigeye tuna, not yellow®n, so the gear is deployed at
greater depths, where bigeye concentrations are found. Most of the tunas caught in the
longline ®sheries are used by the sashimi industry or by the fresh-®sh market.
Finally, there are still some pole-and-line boats operating in the eastern Paci®c. This
method of ®shing was described by Godsil (1938). They carry live bait in tanks, and when
they encounter a school of tunas, they chum the waters with the bait. When the tuna begin
taking the bait, the ®shers use unbaited hooks to catch them. This technique is limited to
some coastal areas because of the dif®culties in keeping the bait alive, and the results in
catches of mostly small ®shes (Tomlinson et al., 1992). In terms of catch per day at sea, it
is not very productive. In the past, it has had problems with depletion of bait.
Types of purse seine sets
Purse seining is conducted in three different ways that correspond to three ways of
detecting the tuna schools: (a) on free-swimming schools of tuna; (b) on tunas associated
with ¯oating objects; and (c) on dolphins.
On free-swimming (un-associated) schools of tuna. A tuna school is detected by
evidence of its presence on the surface of the ocean, i.e. the water appears to be
`boiling' or its surface is disturbed by what appears to be a local breeze, etc.
Frequently, birds associated with the tuna school are detected from the vessel with
radar. This operation is called school ®shing, and the sets are called school sets. This
technique usually produces small yellow®n (modal size of 50 cm, or 2.5 kg, for the
period 1976±1995, Fig. 2(b), below) and skipjack tuna.
On tunas associated with ¯oating objects. Tuna schools tend to associate with
4Hall
¯oating objects during the night, and then leave them early in the morning. When the
®shers ®nd a ¯oating object with tuna around it they surround it with the net shortly
after sunrise, capturing the ®sh associated with the object. Because the most common
objects are tree trunks and branches, this method of ®shing is called log-®shing, and the
sets are called log sets. This technique catches very small yellow®n (modal size of
40 cm, or 1.2 kg, for the period 1976±1995, Fig. 2(a), below) and skipjack tuna.
On dolphins. In the eastern Paci®c, yellow®n tuna are frequently found associated with
groups of dolphins. It is not known why they associate, but most researchers believe,
based in part on the way in which the ®shery operates, that the tunas follow the dolphins.
Most of the hypotheses proposed to explain the association are based on trophic reasons
or predator protection (Perrin et al., 1973; Stuntz, 1981; Anon., 1995: 31), but energetic
reasons (Edwards, 1992) have also been proposed. When the ®shers detect a group of
dolphins, of one or more of the species known to be associated with tunas (from the Fam.
Delphinidae: spotted dolphin, Stenella attenuata, spinner dolphin, S. longirostris,
common dolphin, Delphinus delphis, or, less frequently, striped dolphin, S. coeruleoal-
ba), they attempt to con®rm the presence of tuna either with the aid of the helicopter, or
from the vessel. When ®sh are present, they launch four or ®ve speedboats that chase the
dolphin herd, making a wide arc typically at a distance of 100±200 m to the side and
behind the herd. The chase usually lasts about 20 to 30 minutes, and when it ®nishes, the
dolphin herd has slowed down, or stopped. During this process, part or the whole dolphin
herd may evade the chase and=or encirclement, or, if it is not carrying tuna, may be
deliberately excluded from the encirclement area through the actions of the speedboats.
At this point, the seiner begins to surround them with the net, while the speedboats
manoeuvre in such a way as to keep them inside the encircled area.
Then, the net is `pursed', and both the dolphins and the tunas that were associated
with them are captured. The technique is called dolphin ®shing, and the sets are called
dolphin sets. At this point, the ®shers wish to release the dolphins and then bring the
tunas aboard the vessel. The average size of the dolphin group captured is about 400 to
500 dolphins, but it is common to see groups of more than 1000 dolphins in the net.
This technique produces almost exclusively yellow®n, and these are larger (modal size
of 70±80 cm, or 6.9±10.4 kg, for the period 1976 ± 1995, Fig. 2(c), below) than those
caught by other methods of purse seine ®shing. Unfortunately, because of natural
factors (currents, etc.), equipment malfunctions, or lack of expertise or motivation of
skippers and crews, many dolphins have died during these operations. The incidental
mortality of dolphins caused by this technique generated considerable controversy
around its legal, economic, political and ecological aspects, which are discussed by
Joseph (1994), and Scott (1996).
Figure 1 shows the spatial distribution of the different set types accumulated during the
1979±1996 period. Figure 2 (below) shows length frequencies of yellow®n tuna caught
in the different types of sets; the most recent available data, for 1976±1995, are depicted.
Stocks of dolphins
The dolphin species involved in this ®shery do not constitute a single group distributed
throughout the area; most authors believe that there are geographical subunits, that can
be identi®ed by morphological or other characteristics. These subunits with a limited
degree of mixing are called stocks, and are used as the units of management on the
Tuna±dolphin problem 5
grounds that there is genetic diversity in the units that must be conserved, and that their
population dynamics could differ. The classi®cation used for the stocks is the one
proposed by Perrin et al. (1985) and Dizon et al. (1994). There are two major stocks of
spotted dolphins, the north-eastern and the south-western, two of spinner dolphins, the
eastern and the whitebelly, and three of common dolphins, the northern, the southern and
the central. A subdivision of the northern stock of common dolphins proposed by
Heyning and Perrin (1994) has not been implemented because of dif®culties and
inaccuracies in the discrimination of the different groups by the observers.
The association of tunas and dolphins
The association of yellow®n tunas with dolphins has been observed in other oceans of the
world. An annotated bibliography is available (Donahue and Edwards, 1996). For the eastern
Atlantic, there are descriptions by Bane (1961), Simmons (1968), Mitchell (1975), Levenez
et al. (1980), Maigret (1981a, b, c), Coan and Sakagawa (1982), Pereira (1985), Stretta and
Slepoukha (1986), Cayre et al. (1988) and Santana et al. (1991). For the Indian Ocean,
Potier and Marsac (1984), Montaudouin et al. (1990), De Silva and Boniface (1991), De
Silva and Dayaratne (1991), and Leatherwood and Reeves (1991) mention the association.
For the central and western Paci®c, Paci®c Tuna Development Foundation (1977), Stuntz
(1981) and Dolar (1994) report on sightings or sets. There are also reports for other areas:
Fig. 1. Maps of the eastern Paci®c, showing the density of sets (for years 1979±1996) of the different
types in 1 degree 31 degree squares: (a) log sets; (b) school sets; and (c) dolphin sets. Notice that
intervals used are not equal.
6Hall
Fig. 1. (Continued).
Tuna±dolphin problem 7
Caldwell and Caldwell (1971) for the western Atlantic, Living Marine Resources (1982) for
the Gulf of Mexico, and some mentions in global reviews (Northridge, 1984, 1991). In some
cases (e.g. Di Natale, 1990), description of dolphin captures in tuna purse seines in the
Mediterranean are presented, but without stating whether the tunas and dolphins were
associated or the capture of the dolphins was simply by chance. One reference (Vilicic,
1985, cited in AlegrõÂa Hernandez, 1990) seems to suggest that blue®n tuna (Thunnus
thynnus) in the Mediterranean is found associated with dolphins; however, except for a small
proportion of skipjack, and occasionally bigeye, the association seems to be centred on the
yellow®n tuna. From the point of view of the dolphins, the genera or species found
associated with tunas in other ocean areas are the same or close to the species of dolphins
with which tuna associates in the eastern Paci®c Ocean.
When sets are described, their frequency seems to be much lower than in the eastern
Paci®c, (e.g. 0.0± 4.7% of the sets in the eastern Atlantic (Cayre et al., 1988; Santana et
al., 1991) versus 45± 70% in recent years in the eastern Paci®c. The low level of
observation in most oceans of the world makes it dif®cult to reach ®rm conclusions
concerning the signi®cance of this way of ®shing or its impact on the dolphin populations.
In a few cases there is information on mortality rates provided by the ®shers.
Levenez et al. (1980) report an average mortality of 15 dolphins per set, and a total
number of sets on dolphins per vessel per year of less than 10, based on interviews
with ®shing captains. Combining these values results in estimates of less than 150
dolphins killed per vessel per year, but we cannot establish a lower bound or average
with those values. Most captains interviewed, however, reported never setting on
dolphins. For the Philippines, Dolar (1994) presents a mortality rate of one dolphin for
every two tons of tunas produced by one of the commercial purse seine ¯eets operating
in the area, or a total of 300± 450 dolphins per vessel per year. In both cases, the
sample size was small, and the ®gures were obtained from captains or crew of the
®shing vessels. For comparison, in the eastern Paci®c the current mortality per set is
close to 0.33 dolphins, which translates to one dolphin per close to 70 tonnes of tuna
caught, and an annual average per vessel per year of 50 to 60 dolphins.
Estimates of dolphin mortality for purse seining ¯eets in other areas are practically non-
existent. There is an estimate for the eastern Atlantic of 3300 dolphins for 1977±1978,
produced by Maigret (1981b, cited in International Whaling Commission, 1982, p. 120).
Information on the characteristics of these associations (i.e. species, sizes) is very
scarce. They seem to have similar composition off Sri Lanka and the eastern Paci®c
(De Silva and Boniface, 1991, yellow®n tuna 100±120 cm in length associated with
dolphins). Coan and Sakagawa (1982) describe mostly sets on common dolphins in the
eastern Atlantic; that species of dolphin is much less frequent in the eastern Paci®c sets
(5% of the sets or less).
The tuna±dolphin problem
ESTIMATION OF INCIDENTAL DOLPHIN MORTALITY
The early years
The tuna±dolphin problem was ®rst brought to the attention of the public in the late
1960s (Perrin, 1968, 1969). The mortalities of dolphins in this ®shery were heavy during
the 1960s, but we'll never have a reliable estimate of their magnitude because of the very
scanty and biased database available (Smith and Lo, 1983: Table 4; Lo and Smith, 1986;
8Hall
Wade, 1993: Table 2; Wade, 1995). For the 1959±1970 period, there are data for only
four ®shing trips (out of a total of more than 3500); two from a biologist that was
allowed to participate in them (Perrin, 1968, 1969), and two from unsolicited letters fron
crew members (Smith and Lo, 1983). Of Perrin's two trips, the data for one could not be
used, because he had recorded mortality data only for the high-mortality sets; the
representativeness of the crew members' letters is questionable. A National Academy of
Sciences Committee addressing the tuna±dolphin programme (Francis et al., 1992)
concluded: ``In summary, the mortality estimates for the period before 1973 (peak values
of up to 350 000±653 751 in a year ... have little or no statistical value, and the only
conclusion that can be based on the data available is that mortality was very high.''
During this period, the vast majority of the vessels operating on dolphins were US-¯ag
vessels, the catches were processed by US canneries and sold in the US market.
The Marine Mammal Protection Act of 1972 and its consequences
In response to the US Marine Mammal Protection Act (MMPA) of 1972, a scienti®cally
designed and mandatory observer programme was begun by the US government in 1974.
The US National Marine Fisheries Service (NMFS) was put in charge of this programme.
Observers were placed aboard a sample of ®shing vessels to count the numbers of
dolphins killed, to make observations which would be used to calculate indices of
abundance of the various stocks of dolphins, and to gather information on the causes of
mortalities. Observer coverage in this period increased from about 10% in 1974± 1976 to
about 33% of all trips of US boats in 1977±1978. Besides the observer programme, other
actions were undertaken by the NMFS to develop methods and devices that could reduce
dolphin mortality (Coe et al., 1984), and many regulations based on these studies were
implemented. As a result of these steps, dolphin mortalities declined in the early 1970s,
and levelled-off in the early to mid-1980s.
The internationalization of the ®shery
In the 1970s, the participation of vessels from other nations began to increase, and the
tuna±dolphin problem became an international one. Besides expanding to include more
nations, the ®shery expanded geographically toward the offshore areas of the eastern
Paci®c, and soon a very signi®cant part of the catches came from international waters.
The markets also expanded to include European and Latin American canneries.
Operating in the region since 1949, the Inter-American Tropical Tuna Commission
(IATTC) is an international research organization charged with the collection of data
needed to study the population dynamics of the tuna species, of other related species,
and of the environment of the region, to provide advice to its member nations on
management issues. In 1976, the member nations of the IATTC decided to implement a
tuna±dolphin programme with the objectives of: `` ... strive to maintain a high level of
tuna production and also to maintain dolphin stocks at or above levels that assure their
survival in perpetuity, with every reasonable effort being made to avoid needless or
careless killing of [dolphins].''
The Agreement for the Conservation of Dolphins and the International Dolphin
Conservation Program
In 1992 an agreement initiating an International Dolphin Conservation Program was
signed by representatives of the nations participating in the ®shery, setting overall annual
Tuna±dolphin problem 9
`Dolphin Mortality Limits' for the ¯eets that decline every year from 1993 to 1999 (Anon.,
1993a). Those limits have been divided by the numbers of participating vessels, and each
vessel has been allocated an annual individual `Dolphin Mortality Limit' which, if reached,
forces the vessel to abandon ®shing on dolphins for the rest of the year. A list of
infractions and sanctions has been prepared for this programme. Compliance with them is
veri®ed by an international review panel which includes representatives of the paticipating
governments, the industry, and the environmental community, who are granted access to
the information gathered by the observers that accompany every ®shing trip.
The Declaration of Panama
In 1996, at an international meeting held in Panama City, all ®shing nations from the
area, together with many coastal nations and several major environmental groups,
produced the Declaration of Panama (Anon., 1997) that, if adopted, would consolidate
the gains achieved by the Agreement for the Conservation of Dolphins, and extend its
in¯uence. An acceleration of the schedule to reduce dolphin mortality, the introduction of
stock-by-stock limits, and addressing bycatches of other species are among the additions
of the Declaration of Panama to the previous agreement. As a prerequisite for the
implementation of the Declaration of Panama, some changes must be made in the US
legislation (lifting of embargoes, re-de®ning `dolphin-safe', etc.) These changes are being
discussed at the time of the writing of this manuscript. Several issues have been raised in
recent months by those opposed to implementation of the Declaration. A brief summary
of those is presented in Scott (1996). Many of those issues are neither ecological, nor
speci®c to this problem (i.e. free trade, alleged drug traf®c on tuna vessels, etc.)
Databases available
In 1979 observers from IATTC began to depart on vessels from the US (half of the
sampled trips) and from other ¯eets. In 1992, a national observer programme was started
in Mexico, with conditions similar to the programme of the United States. In the early
1980s, coverage of the US ¯eet remained close to 33%, but that of the other ¯eets was
very low (,6%). In 1986, Mexico, with the largest ¯eet operating in the area, joined the
programme, and the coverage of the non-US ¯eet climbed to 35% by 1988, while the US
¯eet coverage was close to 90%. In recent years, the percentage of the trips carrying
observers has continued to increase, and since 1991 all trips of vessels larger than 363
tonnes of capacity have been sampled.
Summarizing, the observer coverage for the US ¯eet prior to 1977 has probably been
insuf®cient to provide reliable estimates of mortality. For the other ¯eets, the coverage
prior to 1984 is insuf®cient, and only since 1986 have all ¯ags participated in the
programme. Since 1991, the coverage has been 100%, and the database is complete. The
data gathered by the observers may be affected by `observer effects,' if the presence of
the observer affects the behaviour of the crew or their decisions (Wahlen and Smith,
1985), or by interferences with the observer duties resulting from intimidation,
corruption, obstruction, etc. From the point of view of the observer's ability to see the
mortalities, after encirclement is complete, the far end of the net is usually less than
200 m away from him=her, who is equipped with binoculars, and on a deck 7 to 10 m
over sea level. Unless visibility is impaired (i.e. sets that end in darkness), the observer's
view of the net is quite adequate for detecting the presence of dead animals.
Observers also collect data on sightings of dolphin herds that are used to produce
10 Hall
indices of relative abundance. Line-transect methods are the main statistical technique
utilized for this purpose. Because the vessels are ®shing, rather than doing a scienti®c
survey, the data obtained violate many of the assumptions required for the validity of
the models (Anganuzzi and Buckland, 1994). Their use requires special adaptations, and
they yield estimates of relative abundance only, that is indices that are correlated with
abundance, rather than abundance.
Since 1987, IATTC observers began collecting data on the communities associated
with ¯oating objects, as a way to understand what makes them attractive to tunas, in
order to explore their potential as a source of alternative ®shing that could help reduce
effort on dolphins. Seeing the diversity and the large numbers of individuals
incidentally caught and killed in the ®shery, it was thought necessary to start a larger
programme in 1992, to study the bycatches (catches of unwanted species; or of
undersized or unmarketable tunas) in all types of sets in this ®shery.
MORTALITY COMPONENT VARIABLES
There are two statistical components that determine the dolphin mortality level: (1) the
average mortality of dolphins per dolphin set; and (2) the number of sets made on
dolphins. The ®rst depends on the skill and motivation of the captain and crew, and
availability and condition of equipment, and external factors such as occurrence of strong
subsurface currents. The second depends on the size of the ¯eet, the availability of tunas
associated with dolphins, regulations promulgated to limit effort on dolphins, and market
demand for large and small tunas. In order to reduce dolphin mortality, one of the options
is to switch effort away from dolphins, into forms of ®shing that seldom or never kill them.
If a way of ®shing were found that resulted in sustained and high levels of tuna catches, of
the sizes necessary to maintain near-optimal yield per recruit, without much higher costs,
and with little or no dolphin mortality, it would clearly result in major reduction or
elimination of dolphin mortality. But different ways of ®shing have different ecological
costs, and the reduction of dolphin mortality is only one objective of management.
ESTIMATION OF DOLPHIN ABUNDANCE
Estimates of dolphin abundance have been produced with line-transect methods.
Sightings of the dolphin herds, and estimates of group size are combined to produce a
value for the number of groups in area, and the number of individuals in them. Three
platforms have been used to produce the basic data for this purpose.
1. Research vessel data: surveys are planned following an experimental design.
Absolute abundance values have been produced in this way (Wade and Gerrodette,
1993).
2. Tuna vessel data: sightings from the tuna vessel observers are used to produce
indices of relative abundance showing the trends in the numbers of dolphins (Anon.,
1997).
3. Aerial surveys.
Given the very large area of the ®shery, and the limited resources available to
produce population estimates, it is quite clear that the coef®cients of variation of the
estimates will be large, and affected by environmental changes. The uncertainty around
them must be taken into consideration while selecting management options.
Tuna±dolphin problem 11
EVOLUTION OF THE FLEET: TECHNOLOGY AND TRAINING
Purse seine ®shing for tunas was made possible by the development of strong synthetic
webbing, which is more resistant to rotting and to tearing during the intense strain
exerted during ®shing operations, of the Puretic power block which ef®ciently retrieves
the net from the water, and to better methods of freezing the catches (Cole, 1980). Most
seiners today are around 60 to 70 m long and can carry 900 to 1100 tonnes of tuna.
Many carry helicopters to aid in the search for ®sh and `bird radar' that can detect even a
single frigate bird at distances of over 10 miles (.18.5 km). Birds and dolphins are
some of the most common signs of the presence of tunas. Large purse seiners can spend
up to three months at sea, depending on fuel consumption etc., and some of their ®shing
grounds are far offshore, up to 6700 km from the coast.
A series of technological developments, most of which were originated by the ®shers,
have been crucial in reducing dolphin mortality. Among these are: (1) the `backdown'
procedure, which consists of putting the vessel in reverse, after encircling the dolphins,
which forces the corkline to sink and opens an escape route for the dolphins; (2) the
Medina Panel, a section of smaller-meshed webbing in the part of the net with which
dolphins most often come in contact, to keep them from entanglement; and (3) the use
of rescue rafts and other means of hand rescue of dolphins from the net.
When the countries with vessels participating in the ®shery stepped up their efforts to
reduce dolphin mortality, many actions were initiated by the IATTC staff, in
cooperation with national scientists and technicians, to make sure that all vessels had
the right technology and that the ®shers were trained in their use. Statistical studies had
identi®ed a series of factors that led to increased dolphin mortality. These include
environmental factors (e.g. strong subsurface currents), behavioural factors (e.g. some
species or stocks of dolphins `cooperate' with the rescue operations, but others do not),
gear factors (e.g. nets not aligned properly, with holes in the webbing, or lacking some
of the dolphin-saving equipment), and crew factors (e.g. new, unskilled, or poorly
motivated captains or crews). Seminars are frequently held for captains, other crew
members, and vessel managers, where these factors are analysed and solutions
proposed. The operation of the equipment is tested periodically by IATTC technicians.
Trip records are analysed statistically, and the summaries are provided to the industry to
facilitate the follow-up of the progress of the captains and crews. Standards of
equipment and performance levels are required and enforced by the nations. In 1986,
close to 40% of the sets on dolphins had zero mortality; by 1996 this proportion had
climbed to about 88%. The average mortality of dolphins per set decreased from over
12 to 0.33 during the same period. These improvements allowed the ®shery to continue
operating, producing record catches of yellow®n in the late 1980s and early 1990s,
while at the same time reducing the impact on the dolphin populations.
The ecological issues
IMPACT OF THE DOLPHIN MORTALITY ON THEIR POPULATIONS
There are several questions that need to be considered under this heading.
·What are the levels of the populations?
·Are any of the populations in danger of extinction?
·Is the mortality sustainable?
12 Hall
·Are the populations increasing, stable, or decreasing?
·Are there trends in the populations independent of the purse seine ®shery?
In order to estimate the impact of any level of incidental mortality on a stock, it is
necessary to have data on the size of the population, trends in its abundance, and rates
of recruitment and natural mortality. During the 1986±1990 period, the NMFS
conducted annual research vessel and aerial surveys, using line-transect methods, to
assess the condition of the dolphin stocks of the eastern Paci®c (Wade and Gerrodette,
1993). Table 1 shows the estimates obtained. All dolphin stocks involved in the ®shery
have population sizes in excess of 400 000, and there seems to be no danger of
extinction for any of them, at least from the impact of this ®shery. Even though most of
these stocks have experienced serious declines because of the ®shery-caused mortality
(Smith, 1983; Wade, 1995), studies of their trends in recent years show that most of
them have remained at the same level for the past decade (Anganuzzi and Buckland,
1994; Anon., 1997).
Another variable of considerable importance in estimating that impact is the net
recruitment rate, de®ned for delphinids in a simple way as ``reproduction in excess of
mortality for a population as a whole'' (Perrin and Reilly, 1984). Unfortunately, no reliable
studies of net recruitment rate are available for any of the stocks of eastern Paci®c
dolphins, so we use, as a default, the 2% ®gure of Smith (1983: 9), which is believed to be
a conservative estimate of this parameter. Table 1 shows the absolute abundance, and the
most recent mortality estimates. It also shows con®dence intervals placed around the point
estimates for the proportions of mortality (Hall and Lennert, 1994).
To be conservative in assessing the impact of the ®shery on the dolphin population,
i.e. to minimize the possibility of wrongly believing that the stocks of dolphins are in
better condition than they actually are, several sources of uncertainty must be
considered. Estimates of population size and rates of recruitment and mortality can be
Tabl e 1 . Estimates of population abundance (pooled for 1986±1990; Wade and Gerrodette, 1993), of
incidental mortality in 1996, and of relative mortality (with approximate 95% con®dence intervals)
Stock Population Incidental Relative mortality (%)
abundance mortality
Estimate 95% CI
Offshore spotted dolphin
North-eastern 730 900 818 0.11 (0.085, 0.140)
Western ±southern 1 298 400 545 0.04 (0.033, 0.059)
Spinner dolphin
Eastern 631 800 450 0.07 (0.044, 0.108)
Whitebelly 1 019 300 447 0.04 (0.028, 0.058)
Common dolphin
Northern 476 300 77 0.02 (0.009, 0.035)
Central 406 100 51 0.01 (0.007, 0.025)
Southern 2 210 900 30 ,0.01 (0.001, 0.002)
Other dolphins2 802 300 129 ,0.01 (0.004, 0.005)
All 9 576 000 2547 0.03 (0.023, 0.030)
Includes the following species and stocks: striped dolphins (Stenella coeruleoalba), bottlenose dolphins (Tursiops
truncatus), Central American spinner dolphins (Stenella longirostris centroamericana), and unidenti®ed dolphins.
Tuna±dolphin problem 13
inaccurate as a result of methodological or other errors. A group of scientists in the
United States has been working to develop formulae to determine safe levels of take ±
a value called ``potential biological removal' or PBR (Anon., 1994; Barlow et al., 1995;
U.S. Public Law 103± 238, Marine Mammal Protection Act, Amendments of 1994). It is
a cautious scheme to provide managers with information concerning the levels of
incidental mortality (or harvest) that can be extracted from a population with a very
low probability of negative impacts. The PBR provides a conservative limit to mortality
by multiplying a conservative estimate of abundance by an estimate of recruitment rate
and an additional safety factor.
An application of the PBR approach to the eastern Paci®c ®shery is presented in
Table 2 to illustrate the different levels of caution that can be considered. To ensure
that the incidental mortality is sustainable, it is necessary to keep the mortality less
than or equal to the additions to the population during the period in question. When a
population is at its carrying capacity, additions and losses balance out. When a
population has been reduced as a result of some impact below carrying capacity, it is
expected to have a net increase that will depend on the abundance and reproductive
rates of the stock. When we try to estimate these values, we can follow the traditional
statistical approach to produce the best estimate, but caution dictates that we err by
underestimating population size and growth rate, rather than the opposite. With regard
to abundance, the point estimate can be replaced by the lower limit of some con®dence
interval. The PBR formula uses the 20th percentile of the log-normal distribution of
abundance estimates (Wade, 1994b).
Estimates of dolphin net recruitment rates are very dif®cult to measure. Kasuya
(1976) computed a net recruitment estimate of 2.3% and a maximum recruitment rate
(Rmax) of 4.4% for the striped dolphin, Stenella coeruleoalba. Wade (1994a) used
simulation models and a Bayesian approach to estimate maximum rates of increase of
3.8% for the north-eastern stock of spotted dolphins and 2.2% for the eastern stock of
spinner dolphins. However, the lack of observer data for the early years of the ®shery
leaves some doubt about the usefulness of these results, which rely heavily on
extrapolation. The PBR equation uses 1
2Rmax as an estimate of recruitment, with 2% as a
default value when Rmax is unknown.
The third component of the equation is the safety factor or `recovery factor'. This
factor attempts to account for uncertainties in estimates of incidental mortality and to
provide additional margin of error for population whose status is unknown or at risk.
The mortality estimates in particular are subject to several uncertainties. Some
mortalities may not be observed or reported, e.g. observers may overlook mortalities,
observers may be intimidated or corrupted to underreport mortalities, predation on
dolphins may be facilitated by the ®shing operation, or dolphins may suffer injuries that
later result in mortality. Some impacts, such as stress or interference with reproduction,
may not be observable in the short term. To account for such potential impacts, a
recovery factor is set between 0.1 and 1.0. Recovery factors of 0.1 are usually chosen
for endangered species (none of which are target species of the tuna ®shery), 0.5 for
stocks of unknown status or determined to be depleted under the MMPA (north-eastern
spotted and eastern spinner dolphins are depleted stocks), and 1.0 for populations
known not to be at risk (the little-exploited southern common dolphins). Intermediate
values can be chosen as well (the other stocks listed in Table 2 are conservatively given
recovery factors of 0.75 because these stocks have been reduced, but are not at risk).
14 Hall
Tabl e 2 . Potential biological removal (PBR) and zero mortality rate goal (ZMRG) values compared with 1996 dolphin mortality
Stock N
(31000)
CVNminy
(31000)
1=2
Rmaxz
FR
}
1996
PBR}
1996
ZMRG
1996
Mortality
North-eastern spotted dolphins 730.9 0.142 648.9 0.019 0.50 6165 616 818
Western=southern spotted dolphins 1298.4 0.150 1145.1 0.012 0.75 17 177 1718 545
Eastern spinner dolphins 631.8 0.238 518.5 0.011 0.50 2852 285 450
Whitebelly spinner dolphins 1019.3 0.187 871.9 0.02 0.75 13 079 1308 447
Northern common dolphins 713.7 0.288 562.7 0.02 0.75 8441 844 77
Central common dolphins 239.4 0.172 207.3 0.02 0.75 3109 311 51
Southern common dolphins 2210.9 0.217 1845.6 0.02 1.0 36 911 3691 30
Abundance estimate (N) and coef®cient of variation (CV) from Wade and Gerrodette (1993; unpublished data for northern and central common dolphins).
yMinimum abundance estimate (Nmin )N=exp f0:842 3[ln (1 CV2)]1=2g.
{
Maximum population growth rate (Rmax) default is 0.04; values for north-eastern spotted (0.038) and eastern spinner (0.022) from Wade (1993).
}
Recovery factor (FR)0:5 for depleted stocks, 1.0 for unexploited stocks, and a conservative value of 0.75 for stocks that have been reduced, but are thought to be
above OSP.
}Potential biological removal (PBR)Nmin 31
2Rmax 3FR.
Zero mortality rate goal (ZMRG)PBR 0.1.
Tuna±dolphin problem 15
Under this very conservative PBR scheme, limits on ®shery mortality can be set that
would allow populations to recover. An even more restrictive standard can be
implemented, however. The MMPA sets a zero mortality rate goal (ZMRG) for ®sheries
to achieve. It has been operationally de®ned as one-tenth of PBR, a level that is thought
to be biologically insigni®cant (Anon., 1994). The result is that we could remove 1%
(2% 30:5) of a depleted population and still allow it to recover. If the removals are
below 0.1% (1% 30:1), the ®shery would achieve the zero mortality rate goal.
It is clear that, even under the most conservative scenario, the mortality levels for
1996 are well below the assumed recruitment ®gures (Table 2), and it appears safe to
say that the current mortality levels are at least sustainable and that the ®shery has
achieved the zero mortality rate goal of the MMPA for all but two of the dolphin
stocks. Unless one or more of the sources of uncertainty mentioned above proves to be
much worse than anticipated by our safety factors, and under the current ®shery
conditions (and if all other biotic and abiotic factors allow it), these populations should
increase at rates close to the maximum. Given the high variability of the estimates and
the long life span of the dolphins, however, it should take several years for these
increases to become statistically signi®cant (Gerrodette, 1987).
Reducing the mortality caused by the ®shery does not guarantee that the populations
will recover to their pre-exploitation levels, however. Changes in the environment, or in
the structure and function of the ecosytem caused by the previous impact, or by other
impacts may prevent the recovery of the populations. Changes in geographical
distribution following oceanographic changes can also affect our estimates of trends
(Fiedler and Reilly, 1994). An interesting example in the eastern Paci®c is the decrease
in the indices of abundance of the `northern stock' of common dolphins (Anon., 1997),
even when incidental mortality values were at levels of 0.01% of the stock or less.
Studies in central California (Barlow, 1995a, b), to the north of the boundaries of the
®shery for tropical tunas, showed a large increase in the population abundance of the
same species. These increases have persisted over several years, indicating a large-scale
movement of an important part of the population. The causes of that shift in
distribution are unknown, but are not dependent on the ®shery. A study limited to the
boundaries of the ®shery would have shown a decrease in abundance that never took
place, but it might have been interpreted as a trend.
ECOLOGICAL IMPACTS OF THE FISHING OPERATIONS
Virtually all human activities have some impact on the ecosytem in which they take
place, and ®shing is not exception. Given the global increase in the human population, it
seems unlikely that the utilization of many resources could be halted, so it becomes
necessary to ®nd `ecologically sound' ways to utilize them. The meaning of this
expression has to be spelled out clearly. In the context of this paper it means that:
1. the use of the resource is concentrated, as much as possible, on the sizes and ages
of the target population that allow the greatest yields possible on a sustainable basis ±
high yield per recruit ratio;
2. the harvest is managed in a way that avoids, or at least minimizes, the loss of
genetic diversity;
3. the waste of the resource is kept at a minimum ± low [bycatch of target
species=catch] ratio;
16 Hall
4. the use of energy by the vessels is minimized ± low [energy use=catch] ratio;
5. the level of effort is appropriate for the harvest proposed ± high [catch=effort
ratio];
6. pollution originated in the ®shery is minimized ± low [pollution=catch] ratio;
7. the gear used is the best to harvest the resource with the least impact on the habitat
± low [habitat damage=catch] ratio;
8. the negative impact of the exploitation on other species of the system (e.g.
bycatches, competition for prey species) is kept at a minimum or, if possible,
eliminated ± low [bycatch of non-target species=catch] ratio;
9. the `positive' impact of the exploitation on other species of the system (`subsidies')
is also kept at a minimum or, if possible, eliminated ± low [subsidy=catch] ratio;
10. the population is maintained at levels that assure survival even if there are
unexpected, and possibly catastrophic, events such as die-offs.
To approach the problem of ®nding which are the `ecologically most sound' ways to
harvest the tuna populations in a scienti®c manner, it is necessary to compare the
ecological costs of catching them using different gears and techniques. To facilitate this
comparison, we can separate the effects on the target population (the object of the
exploitation) from those on other components of the ecosystem. This separation doesn't
imply any prioritization of the importance of the two groups of effects.
Factors to assess the ecological impacts of different ways of ®shing for yellow®n tuna
From the ecological point of view, a ®shery should operate in such a way that it meets or
approaches the conditions stated before. This view does not include economic or social
considerations, which may also be important to humans. For example, yields less than the
maximum possible may be preferable if the value of the ®sh caught or the employment of
®shers is increased, but larger catches increase employment in the processing plants.
Maximization of yield per recruit. In the case of yellow®n tuna, the optimum size for
maximization of yield per recruit is around 110±120 cm (27±35 kg). Figure 2 shows the
length frequencies of yellow®n tuna captured by the different ways of purse seining. Sets
on dolphins (Fig. 2(c)) produce catches closest to the optimum size. Based on yield-per-
recruit considerations, if the ®shery were to switch from ®shing predominantly on
dolphins towards the other forms, the purse seine catch of yellow®n would decline by
about 25% (Punsly et al., 1994). The decline could be considerably greater, however, if
some or all the scenarios discussed in that study take place (lower effort, reduction in the
range of the ®shery, reduction in yellow®n recruitment). The impact might be somewhat
mitigated, however, by greater catches of skipjack and exacerbated by greater catches of
small bigeye tuna.
Maximization of reproductive rate. With regard to reproduction, the vast majority of the
tunas caught on logs and on free-swimming schools are less than 100 cm in length, and
therefore most are sexually immature (Anon., 1993b; Fig. 2(a) and (b)). If the ®shery
concentrated on these types of sets as an alternative to ®shing on dolphins, the number of
®sh reaching sexual maturity would decline. However, as tunas are extremely fecund, it is
not certain that this decline would impair future recruitment. The information available up
to now (Anon., 1993a: 69± 70, 78) has not shown any relationship between the level of the
Tuna±dolphin problem 17
20 40 60 80 100 120 140 160 180
Length (cm)
2.8 9.6 22.8 44.7 77.5 123.4 184.7 263.6
Wet Weight (kg)
(c)
(b)
(a)
0
0.05
0.1
0.15
0.2
0.25
0.3
0
0.05
0.1
0.15
0.2
0.25
0.3
0
0.05
0.1
0.15
0.2
0.25
0.3
0
0.2
0.4
0.6
0.8
1.0
0
0.2
0.4
0.6
0.8
1.0
0
0.2
0.4
0.6
0.8
1.0
18 Hall
parental stock and the level of the recruitment, but it is possible that futher reductions in
the parental stock, outside the range of the data available, may show some impact.
Minimization of discards. With regard to discards, only relatively large tunas can keep up
with the cruising speed of a group of dolphins (Edwards, 1992) and stay with them
during the chase. That generates catches which are almost totally of market-size ®sh, and
of the most sought-after tuna species (discards are less than 1%). In contrast, drifting
objects produce catches of the smallest tunas caught in the ®shery; there is no selection
of size, and almost 20% of the catch has to be discarded because it is below the market
minimum requirement for size or condition. Sets on free-swimming schools have discard
levels of about 2%, and most of the ®sh retained are well below the optimum size in
terms of maximizing yield per recruit.
It is clear, from the point of view of maximizing yellow®n production and
minimizing bycatch, that ®shing on dolphins is a much sounder way of ®shing than the
alternatives.
Factors to assess the impact of ®shing on the rest of the ecosystem
There are many ways in which a ®shery can affect an environment, and it would be very
dif®cult to analyse all of them. A brief list could include: (1) bycatch of different species;
(2) impact of the ®shing operations on the habitat; (3) impact of lost and discarded gear;
(4) generation of pollution and marine debris; and (5) `subsidies' to some species. The
following section will discuss some of those impacts for the purse seine ®shery.
Bycatches. Dolphins. In the eastern Paci®c ®shery, the trophic relationships between tunas
and dolphins are not well known. It is not even known to what extent, if any, they compete
with one another or help one another. When the ®shery started, it was a pole-and-line
®shery that extracted tunas from the ocean without any dolphin mortality. This ®shery
lasted for decades, until it was largely replaced by the purse seine ®shery. In the early years
of the puse seine ®shery, dolphin mortalities were extremely high, and for a decade or so
remained at a high level. Afterward, dolphin mortality declined for several years, went up
again during the late 1980s, and then declined again. The impact of these differential
mortalities on the interactions between tunas and dolphins and on the ecosystem as a whole
are not known, and it is not possible to gather enough information to recreate with
adequate precision the processes that took place during the 1960s and 1970s.
It is clear that the dolphin populations associated with tunas experienced signi®cant
declines, caused by the ®sheries-in¯icted mortality, until the late 1970s. It has been
suggested that there may be other impacts on the dolphin populations: (1) cryptic
mortality: the ®shing operations, by disrupting the social structure of the dolphin group,
may facilitate attacks on dolphins by sharks and other predators: (2) abortions: the
chase prior to encirclement may cause females to lose their foetuses; (3) injuries: even
though severe injuries seen by observers are recorded and can be accounted for (about
Fig. 2. Length-frequency distribution of fork lengths in centimetres (expressed as percentage of the
catch, y-axis at left), of yellow®n tuna caught in the different types of sets in 1976±1995: (a) log sets;
(b) school sets; and (c) dolphin sets. Shaded curve shows the proportion of ®sh that are sexually
mature at each length (values on right vertical axis). Broken vertical line is the length at which 50%
of the tunas are sexually mature.
Tuna±dolphin problem 19
2% of the total mortality), other injuries may not be detected or assessed properly: (4)
stress: the stress caused by the ®shing operations may have a cumulative impact on the
individuals, reducing their ability to survive, reproduce, or grow (Myrick and Perkins,
1995). Unfortunately, there are no reliable data on the occurrence or level of any of
these potential problems. Sharks are frequently seen close to the nets, and there is one
report of sharks preying on dolphins or, more commonly, feeding on dead dolphins
inside the net or when released. Abortions associated with ®shing have not been
documented. The long-term effects of injuries are hard to assess. Indications of stress
have proved dif®cult to de®ne and measure, and no conclusive evidence one way or the
other has been published.
Fishing on ¯oating objects or on unassociated schools results in lower bycatches of
dolphins. If all the effort directed towards dolphins were directed towards these ways of
®shing, dolphin bycatch would be a few dozen animals per year.
Bycatches of species other than dolphins. One of the ecological costs of ®shing is the
bycatch of species which are not the target of the ®shery. In purse seining operations
the following species are caught incidentally, and are not usually retained:
small tunas (Fam. Scombridae): undersized yellow®n, bigeye, and skipjack tunas,
bullet tunas (Auxis spp.), black skipjack ( Euthynnus lineatus), bonito (Sarda spp.);
bill®shes: Fam. Istiophoridae: striped marlin (Tetrapturus audax), shortbill spear®sh (T.
angustirostris), black marlin (Makaira indica), blue marlin (M. nigricans), sail®sh
(Istiophorus platypterus); Fam. Xiphiidae: sword®sh (Xiphias gladius);
rainbow runner: Elagatis bipinulatus (Fam. Carangidae);
yellowtail: Seriola spp. (Fam. Carangidae);
wahoo: Acanthocybium solandri (Fam. Scombridae);
sharks: Fam. Sphyrnidae: hammerhead shark (Sphyrna spp.); Fam. Carcharhinidae:
blacktip shark (Carcharhinus limbatus), whitetip shark (C. longimanus), silky shark
(C. falciformis), dusky shark (C. oscurus), other sharks (Carcharhinus spp.);
rays: Fam. Mobulidae: manta ray (Mobula spp., Manta hamiltoni); Fam. Dasyatidae:
pelagic sting ray (Dasyatis violacea);
sea turtles: Fam. Chelonidae: olive ridley (Lepidochelys olivacea), green=black
(Chelonia mydas, C. agasizzi), loggerhead (Caretta caretta), hawksbill (Eretmo-
chelys imbricata);
mahi-mahis (dolphin-®sh): Fam. Coryphaenidae: Coryphaena hippurus, C. equiselis;
trigger®shes: Fam. Balistidae;
other large ®sh: Fam. Serranidae (sea bass, cabrilla) and Carangidae (jacks).
The list is far from complete, but it gives an idea of the main species caught,
although it is heavily biased towards the larger species which are easier to see and
identify. Many individuals of small species are also caught; the ®shers refer to some of
them as `bait®sh' (forage for tunas) (anchovies, fam. Engraulidae; herrings and
sardines, fam. Clupeidae; grunts, fam. Haemulidae, etc.) but not others (¯ying ®sh, fam.
Exocoetidae, etc.). In order to compare the bycatches in the different ways of purse
seining for tunas, three measurements have been used: (1) the cost of producing
1000 tons (909 tonnes) of tunas (yellow®n, skipjack and bigeye), in Table 3; (2) the
cost of producing 1000 tons of yellow®n tuna, which is the main target of the ®shery, in
Table 4; (3) and the cost per 10 000 sets of each type, in Table 5. The choice of
20 Hall
1000 tons is arbitrary; the choice of 10 000 sets is based in the fact that in the last
decade there have been, on average, about 10 000 sets on dolphins each year. The
bycatches are estimated by using individual counts in each set, or total weights divided
by average weights.
Of the three types of sets, those on ¯oating objects have, by far, the greatest
bycatches. As the logs are drifting, ®sh of all sizes and body con®gurations, slow or
fast-moving, can aggregate under them. On the other hand, groups of tunas and
dolphins cruise at high speeds, and prior to setting there is a chase at even higher
speed, so that when the group is encircled almost no small or slow-moving species of
®shes or other animals are encircled. Aside from dolphins, the bycatches of dolphin sets
consist of a few sharks and, occasionally, bill®shes, mahi-mahi, wahoo, and=or sea
turtles. The bill®shes may have been travelling with the tuna±dolphin aggregation, but
others were probably by chance in the water encircled. In comparing the columns of
Tables 3, 4 or 5, it is clear that dolphin sets are by far the `cleanest' in this respect.
Sets on unassociated schools have moderately low bycatches (in descending order of
magnitude) of sharks, yellowtail, mahi-mahi, bill®shes, sea turtles and wahoo. Log sets
have large bycatches of mahi-mahi, wahoo, sharks, rainbow runner, yellowtail, bill®shes
and sea turtles. Not included in these considerations are the aggregate classes `other
small' and `other large ®sh' because of uncertainty about their identity. The
invertebrates taken incidentally are almost always jelly®shes, and the observers'
estimates of weights or numbers of individuals are not reliable.
Tabl e 3 . Bycatches in numbers of individuals and discards of tuna (in tons)per 1000 tons of tuna
loaded for the different types of sets, based on combined data for 1993±1996. The numbers in
parentheses are sample sizes
Log sets
(n10 607)
School sets
(n13 112)
Dolphin sets
(n19 570)
Dolphins 0.0 0.1 34.1
Marlins 10.2 4.1 1.5
Sail®sh 0.4 6.6 2.9
Other bill®shes 0.7 0.3 0.1
Blacktip sharks 145.2 89.2 15.8
Silky sharks 51.1 16.3 3.8
Whitetip sharks 34.8 3.6 2.4
Other sharks and rays 59.3 86.1 17.0
Mahi-mahis 4722.7 193.8 2.4
Wahoo 2034.6 26.7 0.6
Yellowtail 110.7 553.8 9.9
Rainbow runner 130.0 36.5 0.0
Other large bony ®shes 54.1 457.3 0.2
Trigger®shes 4774.6 75.6 7.4
Other small ®shes 7286.3 1091.5 358.3
Unidenti®ed bony ®shes 7.3 4.6 10.6
Sea turtles 0.6 0.6 0.3
All tuna discards (tons) 228.5 33.6 9.9
1 ton 0.909 tonne.
Tuna±dolphin problem 21
Impacts of ®shing operations on the habitat. In the case of the tuna purse-seine ®shery,
there appears to be virtually no impact of the ®shing operation on the habitat. As
opposed to bottom trawls that may have an impact on the bottom, the seine rarely, if ever,
has any contact with it.
Impacts of lost or discarded gear. Because of the nature of the seining operation, gear is
seldom lost. Occasionally, pieces of webbing that have been replaced, or that may have
become irretrievably entangled with cables or propellers, may be discarded. There have
been no reports of ®sh or other animals entangled in lost or discarded purse seine
webbing; without ¯oating elements, it should sink. The impact of these pieces on the
bottom communities is unknown.
Generation of pollution and marine debris. Air pollution is generated by burning fuel
while searching for tunas, and given the high consumption of fuel of a seiner compared
with other types of tuna ®shing vessels, it could be a considerable amount. Vessels
®shing for dolphin-associated ®sh use more fuel than vessels ®shing for log ®sh, but the
former ®sh farther offshore, on average, than the latter. The former also carry a
helicopter in most cases, while the latter frequently do not. Water pollution is generated
by the vessels when they dump fuel, oil or other substances to the water. Occasionally,
when large catches are made soon after a vessel leaves port, the fuel stored in a well may
be dumped overboard to make room for the catch. The amount of marine debris, such as
garbage in plastic bags, or discarded containers, generated by the different ¯eets is
Tabl e 4 . Bycatches in numbers of individuals and discards of yellow®n (in tons) per 1000 tons of
yellow®n loaded for the different types of sets, based on combined data for 1993± 1996. The numbers
in parentheses are sample sizes
Log sets
(n10 607)
School sets
(n13 112)
Dolphin sets
(n19 570)
Dolphins 0.1 0.2 34.7
Marlins 58.4 6.1 1.5
Sail®sh 2.2 9.9 2.9
Other bill®shes 4.2 0.5 0.1
Blacktip sharks 829.8 132.8 16.1
Silky sharks 292.1 24.3 3.9
Whitetip sharks 199.1 5.4 2.4
Other sharks and rays 339.1 128.1 17.2
Mahi-mahis 26 987.2 288.4 2.4
Wahoo 11 626.2 39.8 0.6
Yellowtail 632.6 824.4 10.1
Rainbow runner 743.0 54.4 0.0
Other large bony ®shes 309.0 680.6 0.2
Trigger®shes 27 283.4 112.5 7.5
Other small ®shes 41 636.4 1624.6 364.5
Unidenti®ed bony ®shes 41.9 6.9 10.8
Sea turtles 3.6 1.0 0.3
Yellow®n discards (tons) 189.1 17.3 8.7
22 Hall
unknown. To determine the signi®cance of the problem, and compare the different gears,
evaluations are required (Natural Resources Consultants, 1990).
`Subsidies' to some species. The issue of `subsidies' to some species is a dif®cult one.
Many species of marine organisms have learned to use ®shing activities to their
advantage. Fish, birds and mammals follow ®shing vessels to catch prey which are made
vulnerable by the ®shing operation (forced to abandon shelter, confused, separated from
their school, etc.), `steal' prey from deployed gear (Nitta and Henderson, 1993), or feed
on discards (Britton and Morton, 1994; Couperus, 1994; Garthe and Hu
Èppop, 1994). This
type of interaction may also be important, because the ®shery may be tilting the
competitive equilibrium among different species. At the same time, the impact of this
situation is dif®cult to perceive or quantify, because in most cases it is not clear which
species are at a disadvantage in this situation. Some studies, such as those of Mearns et
al. (1981) and Olson and Boggs (1986), have been made of trophic relations of the
upper-level predators of the offshore pelagic zone of the tropical eastern Paci®c Ocean,
but not enough is known to predict the effects of selective removal of various
components of the ecosystem. In the case of the tuna purse seine ®shery, it is likely that
some species of birds and ®sh take advantage of the species that are made more
vulnerable by the ®shing operation or to feed on the discards. Research is needed to
determine which species are bene®ting and which are being harmed by this. Even the
most basic questions, those referring to the tuna ± dolphin interaction itself, have not been
answered (why are tunas and dolphins together? which species bene®ts from the
association? which, it any, is harmed by it? or do both bene®t?). Over the years, the
®shery has extracted tunas and not dolphins (prior to 1959), then dolphins at a much
higher rate (in relative terms) than tunas, and more recently fewer and fewer dolphins and
more and more tunas. If tunas and dolphins are competitors, as the studies on diet
overlap suggest (Perrin et al., 1973), then the early stages of the ®shery favoured the
dolphins, then, during the years of high dolphin mortality, the tunas, and more recently
the dolphins again. If the relationship is bene®cial to both, then any reduction in one of
them would be harmful to both.
Alternative ways of purse seining and other ®shing methods
If ®shing on dolphin-associated schools were eliminated the ®shers have four options.
REMAIN IN THE AREA AND SWITCH TO PURSE SEINING ON SCHOOLFISH OR LOGS
Some idea of the effects of switches to school®sh or logs can be obtained from Table 5,
which shows the bycatches which might be obtained if there were 10 000 sets made on
unassociated schools or logs. Actually, the bycatches would probably be less than
indicated in the table for two reasons: (1) adding arti®cial logs to increase the number of
those sets may result in logs `competing' for the individuals, with lower densities per log;
and (2) the abundance of many of the species would be reduced, which would reduce the
subsequent bycatches. If sets on dolphins were replaced mostly by school sets, the
bycatches would be far less than if they were replaced mostly by log sets. It seems more
likely that dolphin sets would be replaced mostly by log sets, as it happened in the period
1978±1982 (Anon., 1997: Table 4). Advances in knowledge of behaviour of ®sh relative
to oceanographic conditions or improvements in methods for detecting ®sh might make it
Tuna±dolphin problem 23
possible for ®shers to catch more ®sh in unassociated schools. However, ®shers have
recently been deploying arti®cial logs, called ®sh-aggregating devices (FADs), to catch
bigeye and skipjack tuna with considerable success, and it seems likely that if they were
required to cease ®shing on dolphins most of them would switch to ®shing on logs or
FADs. In the eastern Paci®c, the number of log sets in 1996 was the highest of the
decade, and it is almost twice the ®gure from 1991, the ®rst complete year under the
`dolphin-safe' policy. As a result of that development, the increasing catches of small
bigeye, and perhaps the much higher discards of small ®sh, have created a con¯ict with
the longline ®shery which is experiencing a decline in their catch rates (Anon., 1997).
There are various factors which make it impossible to predict the outcome of a switch
from dolphin to log ®shing. It is possible, for example, that there are already enough logs
in the eastern Paci®c Ocean to be close to a `saturation level' to accommodate all the ®sh
that wish to associate with them, so the addition of FADs will increase the catches only
until the `saturation level' is reached, and then further additions would only decrease the
average number of ®sh per log or FAD. The average log set produces about 36 tons of
tuna, of which about 9 tons is yellow®n, the preferred target for the ®shers, whereas the
average dolphin set produces about 18 tons of tuna, all of which is yellow®n. Besides,
when a vessel is ®shing on logs it usually makes only one set per day, early in the
morning, whereas a vessel ®shing on dolphins usually makes more than one set per day.
Most of the catch of log sets is skipjack tuna, which brings a lower price than yellow®n.
At times canneries have refused to accept skipjack, so all skipjack caught in log (and
school) sets had to be discarded. Also, the abundance of skipjack is more variable than
that of yellow®n (Forsbergh, 1989; Anon., 1993a: 81), which makes ®shing on logs more
Tabl e 5 . Bycatches in numbers of individuals and discards of tuna (in tons) per 10 000 sets, based on
combined data for 1993± 1996. The numbers in parentheses are sample sizes
Log sets
(n10 607)
School sets
(n13 112)
Dolphin sets
(n19 570)
Dolphins 6 11 4521
Marlins 3717 631 298
Sail®sh 138 1030 571
Other bill®shes 266 47 25
Blacktip sharks 52 800 13 827 3145
Silky sharks 18 587 2532 764
Whitetip sharks 12 669 562 477
Other sharks and rays 21 576 13 337 3374
Mahi-mahis 1 717 107 30 026 474
Wahoo 739 738 4141 124
Yellowtail 40 252 85 816 1967
Rainbow runner 47 277 5661 6
Other large bony ®shes 19 661 70 854 31
Trigger®shes 1 735 960 11 714 1474
Other small ®shes 2 649 192 169 125 71 309
Unidenti®ed bony ®shes 2664 717 2105
Sea turtles 232 100 64
All tuna discards (tons) 83 091 5210 1964
24 Hall
risky, ®nancially, than ®shing on dolphins. Most log sets are made in the coastal zone of
northern South America and southern Central America, within the Exclusive Economic
Zones of several coastal nations, generating access problems for some ¯eets. On the plus
side, log ®shing requires less fuel than dolphin ®shing.
The tables are based on groupings such as `sharks', `sea turtles', `bill®shes',
aggregate classes that are not adequate as management or conservation units. The
observers record the species whenever possible, which it is in most cases; the groupings
used are only a device to reduce the size of the table. A more detailed study that
includes all the species, the interannual variability, and the seasonal and spatial changes
in bycatches is being prepared. Even when we can break them down to the species
level, we don't have a clear idea of the stock structure, abundance, or other sources of
mortality. As the ®shery on ¯oating objects is more localized geographically, as was
mentioned before (Fig. 1), than the ®shery on dolphins, its impact could be felt by only
one or a few stocks of the species involved.
At the same time, however, the tables illustrate some of the different costs of the
different types of sets. If we maintain this scheme of a set-by-set switch, we can
compare, for instance, the type of set with the greatest bycatches with that with the
least by: (1) subtracting the average mortality per set on dolphins from the average on
logs for all the common species; (2) dividing both sides by the average mortality per
set for the dolphins, to obtain a crude correspondence showing the differential impacts
of the two ways of ®shing. The following `equation' is the result of those operations:
Differential costs of ®shing
Dolphin sets Log sets
1 dolphin 15 620 small tunas
0.1 sail®sh 382 mahi-mahi
0.1 manta ray 190 wahoo
7.6 rainbow runners
11.0 blacktip sharks
4.0 silky sharks
2.4 whitetip sharks
0.4 hammerhead sharks
2.9 other sharks and rays
0.3 black marlin
0.3 blue marlin
0.1 striped marlin
0.1 other bill®shes
4.3 other large ®sh
428 trigger®shes
800 other small ®sh
0.04 sea turtles
Except for the dolphins, the sail®sh, and the manta rays, the bycatches of all other
species are much greater in log sets than in dolphin sets. The protection of the one
dolphin on the left side of the equation results in the mortality of many other
organisms. How can we compare these impacts from the ecological point of view? The
answer to the question: how many sharks is a dolphin worth? is not to be found in the
Tuna±dolphin problem 25
ecological literature. We do not know what is the impact (even for the dolphins!) of the
mortality on the right side of the equation. Are all these species plentiful? Is this level
of mortality sustainable, or negligible? In the case of the dolphins, the current level of
mortality is not only sustainable but so low that it should allow these species to recover
to levels approaching those which existed before the advent of the purse seine ®shery
for tunas.
This equation emphasizes that there are two problems, not one, competing for our
attention. Solving one at the expense of exacerbating the other is not the ecologically
sound way out of this situation. Even though many people have a stronger aversion to
mortalities of dolphins than mortalities of sharks or other species, this preference has
no scienti®c basis. The bycatch in log sets is an issue that can and should be addressed
with a combination of management and technological innovation. Reducing its
magnitude is another goal that should be pursued. Intensifying its impact to eliminate
the mortality of dolphins through ®shing does not have a sound ecological basis.
MOVE TO OTHER OCEANS AND FISH ON SCHOOLFISH OR LOGS
Moving to other oceans is not a good solution to the problem. In the ®rst place, tunas in
the original area would be underutilized and there would be massive unemployment in
many coastal communities. In the second place, most stocks of tunas in other oceans are
already fully exploited, and the entrance of more vessels into those ®sheries would cause
problems for the vessels already there. There is very little information on the
composition and level of bycatches in other ocean areas, but it is known that the
communities associated with ¯oating objects are similar in most oceans, so the bycatches
in this type of sets occur but in other oceans (Bailey et al., 1996). Transferring the
problem to another ocean is therefore not an ecologically sound solution.
CHANGE GEAR AND REMAIN IN THE REGION
Switch to pole-and-line ®shing
The need for live bait limits the geographical range of this ®shery to coastal waters, and
therefore the catches are limited mostly to small tunas (Hennemuth, 1961). It is quite
inef®cient from the point of view of catch per unit of fuel consumed or catch per unit of
time, because of the time and energy spent in a `double' ®shery for bait and then for
tunas. The bait should be considered a bycatch, and we should remember that this way of
®shing may have led to depletion of bait species in some areas of the ®shery (Longhurst,
1971). On the other hand, there are no dolphin bycatches, and the bycatches of most
other species are minimal. It is unlikely that pole-and-line ®shing, under present
conditions, could supply the need for tunas by the canning industry.
Switch to longlines
Kanasashi (1960), Yoshida (1966) and Sainsbury (1996) describe the gear used. This
method catches small numbers of large tunas, bill®shes and sharks. Most of the yellow®n
caught (Nakano and Bayliff, 1992, Figures 63±65) are larger than the critical size of
116 cm (Anon., 1995: 58). If large numbers of longline vessels were constructed for
®shing in the eastern Paci®c Ocean, and if these vessels all deployed their gear in area-
time-depth strata in which yellow®n were most likely to be caught, their total catches
would still be much less than those of the puse seine ®shery, as that ®shery, especially
when directed at dolphin-associated ®sh, takes ®sh which are closer to what yield-per-
26 Hall
recruit analysis indicates is the optimum size for harvesting than does the longline
®shery. In order to make longlining pro®table, is is necessary to have a market that pays
a high price for these catches, which is the case with the Japanese market for some fresh
tunas. Because this market is limited, and because the supply of large tunas is limited, the
catches of tunas by the longline ®shery are much less that those by the purse seine
®shery. Setting the longlines at different depths changes the selectivity of the gear
(Boggs, 1992; Nakano and Bayliff, 1992). There is not much information of longline
bycatches in the eastern Paci®c, but longlines are known to cause bycatches of sharks,
bill®shes, sea turtles, sea birds and other species in other areas (Witzell, 1984; Brothers,
1991; Rey and Mun
Äoz-Chapuli, 1991; Hoey, 1992; Stevens, 1992; Nitta and Henderson,
1993).
Switch to gill nets
Gill nets are used to catch tunas in the oceans of the world. There is no experience with
offshore gill nets in the eastern Paci®c ®shing grounds, but it seems unlikely that it will
be a more selective way of ®shing than purse seining with respect to tunas and to other
species. In any case, a United Nations resolution has banned the use of high-seas gill
nets. Coastal gill nets ®shing for tunas produce incidental mortality of dolphins in most
®sheries that have been studied (Shomura, 1963; Perrin et al., 1994). Francis et al. (1992:
101) recommend against gill nets for ®shing for tunas. Fishers are trying to increase the
selectivity of gill nets, but when we have data, the costs in dolphins, other marine
mammals, and other species can be high. The Sri Lanka gillnet ®shery, which targets
mainly tunas, results in an incidental mortality of dolphins of between 5200 (Dayaratne
and Joseph, 1993) and 9000±12 000 (Leatherwood, 1994). As the total catches of that
®shery are close to 20 000 tonnes of `®sh' (Dayaratne and Joseph, 1993), most of which
is tuna, the cost in dolphins of producing that tuna is close to one dolphin per 4 tonnes of
tuna (of dolphin-safe tuna!), if we use the lowest mortality estimate, and one dolphin per
1.7 tonnes if we use the highest. These ®gures compared with one dolphin per 70 tonnes
of tunas in the eastern Paci®c tuna ®shery.
REMAIN IN THE AREA AND DEVELOP NEW TECHNOLOGIES
Can we develop ways of ®shing that catch the right sizes of tunas, without involving
dolphins, and without impacts on other species of the system? Using the purse seine
technology, that boils down to: (1) ®nding other ways to locate the schools of large
yellow®n tuna that are not associated with dolphins (if there are enough of them), or (2)
®nding a way to attract the large yellow®n tuna, or (3) ®nding a way to separate tunas
from dolphins before capture. Several technologies have been proposed to detect tuna
schools, based on laser (Oliver et al., 1994), acoustics, radar, etc. Up to now, the issue is
at an early experimental stage, and we have seen no great advances in the recent past.
FADs. To attract large yellow®n, the hopes are placed on the possibility of deploying
FADs (®sh-attracting devices) in areas and times where some records show that sets
have occasionally been made on large tuna in schools or associated to ¯oating objects
(Anon., 1991: Fig. 25). If those ®sh could be attracted (Armstrong and Oliver, 1995),
without large amounts of small tunas and other species, then the solution could be
ecologically sound. However, in the eastern Atlantic, where FADs have been used
intensively, the majority of the tuna vessel owners operating there have implemented a
voluntary ban on the practice in a time-area stratum (A. Fonteneau, pers. comm.),
Tuna±dolphin problem 27
which suggests that they perceive the negative effects of the practice to be quite
signi®cant. Experiments are needed to answer this question.
Separating tunas from dolphins. Several methods have been proposed to separate
tunas from dolphins during or prior to encirclement, based on herding the dolphins or
using sounds, chemical substances, or other means to attract or repel tunas or dolphins
(Coe et al., 1984; Edwards, 1996). Again experiments are needed because there are no
clear data on such methods; the issue of herding dolphins has proved quite intractable
up to now. We need to know a lot more about the social structure and schooling
behaviour of both tunas and dolphins to devise an effective way of separating them.
Some recent experiments attempting to explore the behaviour of tunas and dolphins are
discussed by Anon., (1993a: 60± 63; 1995: 51±52).
Mid-water trawls. Prado (1988) describes a trawl ®shery for albacore tuna (Thunnus
alalunga) in the Bay of Biscay, and unsuccessful tests performed in the Gulf of Guinea
in 1977. There are no reports of trawling for tunas in the eastern Paci®c. One study
based on experimental ®shing in the western Atlantic shows higher dolphin bycatches
per tonne of ®sh caught (roughly one dolphin per 4± 5 tonnes of tunas, Gerrior et al.,
1994) than for the eastern Paci®c purse seine ®shery. The most recent studies (Goudey,
1995, 1996) show very variable bycatch rates: one small cetacean per 124 tonnes of
tunas in 1994, and one per 10±11 tonnes of tuna in 1995. The cetaceans included pilot
whales and two species of dolphins, and the tuna catches were composed of albacore,
yellow®n and bigeye tuna. For comparison, the bycatch rate observed in the eastern
Paci®c purse seine ®shery is one dolphin per 70 tonnes of tuna. Given the experimental
nature of the pair trawl ®shery, it is possible that with more information on the
distribution, seasonality and causes of the bycatches, and the development of auxiliary
technology, the bycatch rates could be lowered substantially. As recently as 1986, the
eastern Paci®c purse seine ®shery had bycatch rates of the order of one dolphin per
1.6 tons of tuna caught. Ten years after that, the rates have been reduced by 97%. This
example shows the dif®culty of making a fair comparison between a new ®shery and a
mature one.
Conclusions
In comparison with all the other gears and techniques mentioned above, ®shing on
dolphins produces yellow®n tuna closer to the optimum size required to maximize yield
per recruit, allows them to reproduce, and wastes very little of the resource. The
bycatches of turtles and other species of ®sh are very low, and the bycatches of dolphins
have been brought under control, to levels that permit the recovery of the dolphin
populations, and eliminate the conservation concerns that gave this issue its high pro®le.
A `perfect' solution for the dolphins, one that eliminates all mortality and disturbance,
and provides them with total protection would be very costly, under current ®shing
practices, for all other species in the ecosystem, and for the ®shers.
Considerable further study is needed to produce a full assessment of the ecological
characteristics (positive and negative) of each way of ®shing. If wise management
decisions are to be made, it will be necesasry to evaluate the alternatives (DeMaster,
1992), and to have a much better knowledge than available today of the ecological costs
of ®sh production. To eliminate the ecological impacts is impossible, as removal of ®sh,
28 Hall
particularly selective removal of one or a few species, causes ecological impacts. But
we can choose among the impacts, and work to mitigate them.
Some recommended areas for research are:
1. identi®cation of management units for all species (stocks);
2. estimation of mortality rates and abundances for all species (stocks);
3. ecological fate of bycatches discarded at sea;
4. ecosystem interactions, especially those involving the target species with species
taken incidentally and with species subsidized by the ®shery;
5. modelling studies to assess the impact of management actions concerning bycatches;
6. technological improvements to increase gear selectivity;
7. development of techniques for pre-sorting the catch before loading it (¯oating cages,
®sh chutes, etc);
8. techniques to increase the survival of unmarketable species and individuals;
9. utilization of bycatches through marketing, changes in operations, etc.
If we can solve the tuna±dolphin problem in a satisfactory manner for most of those
involved, it would set up a model that can be used for other ®sheries. The approach
should be an international one, based on science and on the education of the ®shers to
produce the needed changes. It should not create a false dichotomy between the use of
a resource and the conservation of the ecosystem, or between jobs and the environment.
The industry should be held accountable by the nations and by the rest of the
community represented by non-governmental organizations, which are given access to
the basic information needed for monitoring the progress and the compliance with the
programmes agreed to. The solution should be based on scienti®c facts and on a
complete ecological perspective of the ®shery.
Acknowledgements
Lic. Marco GarcõÂa processed most of the information used in this document, and has
made himself indispensable many times, and in many different ways.
The author wishes to express his gratitude to Dr William Bayliff for his review of
this manuscript, and for years of patience, in which he made major contributions to the
legibility and to the contents of many documents. His solid common sense, his critical
mind, and his knowledge have kept the author from going astray in many occasions.
Dr Michael Scott is also thanked for his review and for many discussions that helped
in the development and analysis of some of the concepts presented in this document.
References
AlegrõÂa Hernandez, V. (1990) The tuna ®shery in the eastern Adriatic. ICCAT, Coll. Vol. Sci. Pap. 33,
101± 107.
Alverson, F.G. (1960) Distribution of ®shing effort and resulting tuna catches from the eastern tropical
Paci®c by quarters of the year, 1951±1958. Inter-Am. Trop. Comm. Bull. 4, 321± 446.
Anganuzzi, A.A. and Buckland, S. (1994) Relative abundance of dolphins associated with tuna in the
eastern Paci®c Ocean: analysis of 1992 data. Rep. Int. Whal. Comm. 44, 361 ± 366.
Anonymous (1991) Inter-Am. Trop. Tuna Comm. Ann. Rpt for 1989. 270 pp.
Tuna±dolphin problem 29
Anonymous (1993a) Inter-Am. Trop. Tuna Comm. Ann. Rpt for 1992. 315 pp.
Anonymous (1993b) Inter-Am. Trop. Tuna Comm. Quart. Rpt for April± June 1993. 58 pp.
Anonymous (1994) Report of the P.B.R. (Potential Biological Removal) Workshop. June 27±29.
(Southwest Fisheries Science Center, La Jolla, California.). 22 pp.
Anonymous (1995) Inter-Am. Trop. Tuna Comm. Ann. Rpt for 1994. 296 pp.
Anonymous (1997) Inter-Am. Trop. Tuna Comm. Ann. Rpt for 1995. 334 pp.
Armstrong, W.A. and Oliver, C.W. (1995) Recent use of ®sh aggregating devices in the eastern tropical
Paci®c purse-seine ®shery. La Jolla, CA: Southwest Fisheries Science Center. Admin Rept. LJ-95-
14.
Bailey, K., Williams, P.G. and Itano, D. (1996) By-catch and discards in western Paci®c tuna ®sheries:
a review of SPC data holdings and literature. Noumea, New Caledonia: Oceanic Fisheries
Programme, South Paci®c Commission, Tech. Rept 34.
Bane, G.W., Jr (1961) The distribution and abundance of tunas and tuna bait ®shes in the Gulf of
Guinea. MS thesis, Cornell Univ. 119 pp.
Barlow, J. (1995a) The abundance of cetaceans in California waters. Part I: ship surveys in summer
and fall of 1991. Fish. Bull., US. 93, 1 ±14.
Barlow, J. (1995b) The abundance of cetaceans in California waters. Part II: aerial surveys in winter
and spring of 1991 and 1992. Fish. Bull., U.S. 93, 15± 26.
Barlow, J., Swartz, S.L., Eagle, T.C. and Wade, P.R. (1995) U.S. marine mammal stock assessments:
guidelines for preparation, background, and a summary of the 1995 assessments. U.S. Dept of
Commerce, NOAA Tech. Memo. NMFS-OPR-6, 73 pp.
Ben-Yami, M. (1994) Purse Seining Manual. Oxford: Fishing News Books. 406 pp.
Boggs, C.H. (1992) Depth, capture time, and hooked longevity of longline-caught pelagic ®sh: timing
bites of ®sh with chips. Fish. Bull., U.S. 90, 642± 658.
Britton, J.C. and Morton, B. (1994) Marine carrion and scavengers. Oceanogr. Mar. Biol. Ann. Rev. 32,
369± 434.
Broadhead, G.C. (1962) Recent changes in the ef®ciency of vessels ®shing for yellow®n tuna in the
eastern Paci®c Ocean. Inter-Am. Trop. Tuna Comm. Bull.6, 283± 332.
Brothers, N. (1991) Albatross mortality and associated bait loss in the Japanese longline ®shery in the
Southern Ocean. Biol. Conserv. 55, 255±268.
Caldwell, D.K. and Caldwell, M.C. (1971) Porpoise ®sheries in the southern Caribbean ± present
utilizations and future potentials. Rosenstiel Sch. Mar. and Atmostph. Sci., Proc. 23rd Ann.
Session, Gulf and Caribbean Fish. Inst.: 195±205.
Cayre , P., Kothias, J.B.A., Diouf, T. and Stretta, J.M. (1988) Biologie des thons. In Fonteneau, A. and
Marcille, J. eds. Ressources, peÃche et biologie des thonideÂs tropicaux de l'Atlantique Centre-Est.
FAO, Doc. Tech. sur les Pe
Ãches No. 292, 157± 268.
Coan, A.L. and Sakagawa, G.T. (1982) An examination of single set data for the US tropical tuna
purse seine ¯eet. ICCAT, Coll. Vol. Sci. Pap. 18, 83± 94.
Coe, J.M., Holts, D.B. and Butler, R.W. (1984) The ``tuna-porpoise'' problem: National Marine
Fisheries Service dolphin mortality reduction research, 1970± 1981. Mar. Fish. Rev. 46, 18±33.
Cole, J.S. (1980) Synopsis of biological data on the yellow®n tuna, Thunnus albacares (Bonnaterre,
1788) in the Paci®c Ocean. Inter-Am. Trop. Tuna Comm. Special Rep. 2, 71± 150.
Couperus, A.S. (1994) Killer whales (Orcinus orca) scavenging on discards of freezer trawlers north
east of the Sheltand islands. Aquatic Mammals 20, 47 ± 51.
Dayaratne, P. and Joseph, L. (1993) A study on dolphin catches in Sri Lanka. Madras, India: Bay of
Bengal Program, 47 pp.
DeMaster, D.P. (1992) Strategic plan to develop and evaluate ``dolphin-safe'' methods of ®shing for
yellow®n tuna in the eastern tropical Paci®c. La Jolla, CA: Southwest Fisheries Science Center,
Admin. Rept LJ-92-16, 21 pp.
De Silva, J. and Boniface, B. (1991) The study of the handline ®shery on the west coast of Sri Lanka
with special reference to the use of dolphin for locating yellow®n tuna (Thunnus albacares).
30 Hall
IPTP (FAO), Coll. Vol. Work. Doc., 4. TWS=90=18, pp. 314 ± 324.
De Silva, J. and Dayaratne, P. (1991) Observations on the recently developed offshore ®sheries for
skipjack and yellow®n tunas in Sri Lanka. IPTP (FAO), Coll. Vol. Work. Doc., 4. TWS=90=17,
pp. 304± 313.
Di Natale, A. (1990) Marine mammal interactions in Scombridae ®shery activities: the Mediterranean
case. ICCAT, Coll. Vol. Sci. Pap.33, 140± 142.
Dizon, A.E., Perrin, W.F. and Akin, P.A. (1994) Stocks of dolphins (Stenella spp. and Delphinus
delphis) in the eastern tropical Paci®c: a phylogeographic classi®cation. NOAA Tech. Rep. NMFS
No. 119. 20 pp.
Dolar, M.L. (1994) Incidental takes of small cetaceans in ®sheries in Palawan, Central Visayas and
Northern Mindanao in the Philippines. In Perrin, W.F., Donovan, G.P. and Barlow, J., eds. Gillnets
and Cetaceans. Rept. Int. Whal. Comm. Special Issue, No. 15, pp. 355±363. Cambridge, UK.
Donahue, M.A. and Edwards, E.F. (1996) An annotated bibliography of available literature regarding
cetacean interactions with tuna purse-seine ®sheries outside of the eastern tropical Paci®c Ocean.
La Jolla, CA: NMFS, Southwest Fisheries Science Center, Admin. Rept LJ-96-20. 46 pp.
Edwards, E.F. (1992) Energetics of associated tunas and dolphins in the eastern tropical Paci®c Ocean:
a basis for the bond. Fish. Bull. U.S. 90, 678±690.
Edwards, E.F. (1996) Separation=attraction research on the tuna± dolphin bond: review and criteria for
future proposals. La Jolla, CA: NMFS, Southwest Fisheries Science Center, Admin. Rept LJ-96-
17. 25 pp.
Fiedler, P.C. and Reilly, S.B. (1994) Interannual variability of dolphin habitats in the eastern tropical
Paci®c. II: Effects on abundances estimated from tuna vessel sightings, 1975±1990. Fish. Bull.
U.S. 92, 451±463.
Forsbergh, E.D. (1989) The in¯uence of some environmental variables on the apparent abundance of
skipjack tuna, Katsuwonus pelamis, in the eastern Paci®c Ocean. Inter-Am. Trop. Tuna Comm.
Bull. 19, 433± 569.
Francis, R.C., Awbrey, F.T., Goudey, C.A., Hall, M.A., King, D.M., Medina, H., Norris, K.S., Orbach,
M.K. Payne, R. and Pikitch, E. (1992) Dolphins and the Tuna Industry. Washington, DC:
National Academy Press, 176 pp.
Garthe, S. and Hu
Èppop, O. (1994) Distribution of ship-following seabirds and their utilization of
discards in the North Sea in summer. Mar. Ecol. Prog. Ser. 106,1±9.
Gerrior, P., Williams, A.S. and Christensen, D.J. (1994) Observations of the U.S. pelagic pair trawl
®shery in the Northwest Atlantic. Mar. Fish. Rev. 56(3), 24± 27.
Gerrodette, T. (1987) A power analysis for detecting trends. Ecology 68, 1364± 1372.
Godsil, H.C. (1938) The high seas tuna ®shery of California. Calif. Dept Fish Game, Fish Bull. 51,41
pp.
Goudey, C.A. (1995) The 1994 experimental pair trawl ®shery for tuna in the Northwest Atlantic.
Cambridge, MA: MIT-Sea Grant 95± 6.
Goudey, C.A. (1996) The 1995 experimental pair trawl ®shery for tuna in the Northwest Atlantic.
Final Rept. MIT-Sea Grant Coll. Prog. NOAA Grant NA46RG0434, 12 pp.
Green, R.E., Perrin, W.F. and Petrich, B.P. (1971) The American tuna purse-seine ®shery. In
Kristjonsson, H., ed. Modern Fishing Gear of the World: 3. London: Fishing News (Books) Ltd,
pp. 182± 194.
Hall, M.A. and Lennert, C. (1994) Incidental mortality of dolphins in the eastern Paci®c Ocean tuna
®shery in 1992. Rep. Int. Whal. Comm 44, 349 ± 351.
Hennemuth, R.C. (1961) Size and year class composition of catch, age and growth of yellow®n tuna in
the eastern tropical Paci®c Ocean for the years 1954±1958. Inter-Am. Trop. Tuna Comm., Bull. 5,
1± 112.
Heyning, J. and Perrin, W.F. (1994) Evidence for two species of common dolphins (Genus Delphinus)
from the eastern North Paci®c. Nat. Hist. Mus. Los Angeles County Contrib. Science No. 442. 1±
35.
Tuna±dolphin problem 31
Hoey, J.J. (1992) Bycatch in U.S. Atlantic longline ®sheries for sword®sh and tuna. In Schoning, R.W.,
Jacobson, R.W., Alverson, D.L., Gentle, T.H. and Auyong, J., eds. Proceedings of the National
Industry Bycatch Workshop held at Newport, Oregon, Feb. 4±6, 1992, pp. 61± 70. Natural
Resources Consultants, 4055 21st W., Seattle, WA, USA.
International Whaling Commission (1982) Report of the Scienti®c Committee. Annex H. Ann. Rept
Int. Whal. Comm.32.
Joseph, J. (1994) The tuna±dolphin controversy in the eastern Paci®c Ocean: biological, economic, and
political impacts. Ocean Develop. Inter. Law 25(1), 1±30.
Kanasashi, Y. (1960) Longline ®shing: deck design and equipment. In Traung, J.-O., ed. Fishing Boats
of the World: 2. London: Fishing News (Books) Ltd. pp. 73±83.
Kasuya, T. (1976) Reconsideration of life history parameters of the spotted and striped dolphin based
on cemental layers. Sci. Rep. Whales Res. Inst., Tokyo 28, 73± 106.
Leatherwood, S. (1994) Re-estimation of incidental cetacean catches in Sri Lanka. Rep. Int. Whal.
Comm. Special Issue 15, Annex D, 64± 65.
Leatherwood, S. and Reeves, R.R. (eds) (1991) Marine mammal research and conservation in Sri
Lanka, 1985± 1986. Nairobi, Kenya: United Nations Environment Programme, Marine Mammal
Tech. Rept No. 1, 2nd edn. 138 pp.
Levenez, J., Fonteneau, A. and Regalado, R. (1980) Resultats d'une enquete sur l'importance des
dauphins dans la pecherie thoniere FISM. ICCAT, Coll. Vol. Sci. Pap.9, 176± 179.
Living Marine Resources, Inc. (1982) Gulf Coast tuna resource survey, April 1 to August 12, 1982,
F=V Providence. Silva Fishing Co., Grant-in-aid award No. NA82-GA-D-005 15, 37 pp.
Lo, N.C.H. and Smith, T.D. (1986) Incidental mortality of dolphins in the eastern tropical Paci®c,
1959± 1972. Fish. Bull. U.S. 84, 27± 34.
Longhurst, A.R. (1971) The clupeoid resources of tropical seas. Oceanog. Mar. Biol., Ann. Rev. 9,
349± 385.
Maigret, J. (1981a) Rapports entre les ceÂtaceÂs et la peÃche thonieÁre dans l'Atlantique tropical oriental.
Notes Afr. 171, 77± 84.
Maigret, J. (1981b) Groupe d'e tude des mammifeÁres marins dans l'Atlantique tropical Africain.
Gemmata=N.I. No. 3, Jan. 1981. 8 pp.
Maigret, J. (1981c) Introduction aÁ l'eÂtude des rapports entre CeÂtaceÂs et la Peà che thonieÁre dans
l'Atlantique tropical. Bull. Centre Natn Recherches Oce
Âanogr. Pe
Ãches Moaudhibou 10, 89±101.
Mearns, A.J., Young, D.R., Olson, R.J. and Schafer, H.A. (1981) Trophic structure and the cesium±
potassium ratio in pelagic ecosystems. Calif. Coop. Oceanic Fish. Invest. Rep.22, 99±110.
Mitchell, E. (1975) Porpoise, Dolphin and Small Whale Fisheries of the World. Status and Problems
(IUCN Monograph No. 3). Gland, Switzerland and Cambridge, UK: International Union for the
Conservation of Nature and Natural Resources. 129 pp.
Montaudouin, X. de, Hallier, J.P. and Hassani, S. (1990) Analyse des donneÂes collecteÂes lors des
embarquements a bord des senneurs baseÂs aux Seychelles (1986±1989). Seychelles Fishing
Authority, Tech. Rep. 34 pp.
Myrick, A.C. and Perkins, P.C. (1995) Adrenocortical color darkness and correlates as indicators of
continuous acute premortem stress in chased and purse-seined captured male dolphins.
Pathophysiology 2, 191± 204.
Nakano, H. and Bayliff, W.H. (1992) A review of the Japanese longline ®shery for tunas and bill®shes
in the eastern Paci®c Ocean, 1981±1987. Inter-Am. Trop. Tuna Comm. Bull. 20, 185±355.
Natural Resources Consultants, Inc. (1990) Survey and evaluation of ®shing gear loss in marine and
Great Lakes ®sheries of the United States. Final Report for National Marine Fisheries Service,
Contract 50ABNF-9-00144, 141 pp.
Nitta, E.T. and Henderson, J.R. (1993) A review of interactions between Hawaii's ®sheries and
protected species. Mar. Fish. Rev. 55, 83± 92.
Northridge, S.P. (1984) World review of interactions between marine mammals and ®sheries. F.A.O.
Fish. Tech. Paper No. 251. 190 pp.
32 Hall
Northridge, S.P. (1991) An updated world review of interactions between marine mammals and
®sheries. F.A.O. Fish. Tech. Paper No. 251, Suppl. 1. 58 pp.
Oliver, C.W., Armstrong, W.A. and Young, J.A. (1994) Development of an airborne LIDAR system to
detect tunas in the eastern tropical Paci®c purse-seine ®shery. U.S. Dept Commerce, NOAA Tech.
Memo. NOAA-TM-NMFS-SWFSC-204, 67 pp.
Olson, R.J. and Boggs, C.H. (1986) Apex predation by yellow®n tuna (Thunnus albacares):
independent estimates from gastric evacuation and stomach contents, bioenergetics, and cesium
concentrations. Can. J. Fish. Aquat. Sci.43, 1760± 1775.
Paci®c Tuna Development Foundation (1977) Final report tuna purse-seine charter to the Western
Paci®c. July± November 1976. 14 pp.
Pereira, J. (1985) Composition speci®que des bancs des thonides peÃches aÁ la senne, aux Acores.
ICCAT Coll. Vol. Sci. Pap.25, 395±400.
Perrin, W.F. (1968) The porpoise and the tuna. Sea Frontiers 14, 166±174.
Perrin, W.F. (1969) Using porpoise to catch tuna. World Fishing 18, 42±45.
Perrin, W.F. and Reilly, S.B. (1984) Reproductive parameters of dolphins and small whales of the
family Delphinidae. Rep. Int. Whal. Comm., Special Issue 6, 97 ± 133.
Perrin, W.F., Warner, R.W., Fiscus, C.L. and Holts, D.B. (1973) Stomach contents of porpoise, Stenella
spp., and yellow®n tuna Thunnus albacares, in mixed species aggregation. Fish. Bull. 71, 1077±
1092.
Perrin, W.F., Scott, M.D., Walker, G.J. and Cass, V.L. (1985) Review of geographical stocks of tropical
dolphins (Stenella spp. and Delphinus delphis) in the eastern Paci®c. US Dept Commerce, NOAA
Tech. Rept NMFS 28, 68 pp.
Perrin, W.F., Donovan, G.P. and Barlow, J. (eds) (1994) Gillnets and Cetaceans. Cambridge, UK: Rep.
Int. Whal. Comm., Special Issue 15. 629 pp.
Potier, M. and Marsac, F. (1984) La peÃche thonieÁre dans l'Ocean Indien: campagne exploratoire d'une
¯ottille de senneurs (1982±1983). Mission ORSTOM Seychelles, Rapp. Sci. No. 4. 73 pp.
Prado, J. (1988) Trawling for albacore. INFOFISH International,4=88, 50.
Punsly, R.G., Tomlinson, P.K. and Mullen, A.J. (1994) Potential tuna catches in the eastern Paci®c
Ocean from schools not associated with dolphins. Fish. Bull. U.S. 92, 132± 143.
Rey, J.C. and Mun
Äoz-Chapuli, R. (1991) Relation between hook depth and ®shing ef®ciency in surface
longline gear. Fish. Bull. U.S. 89, 729±732.
Sainsbury, J.C. (1996) Commercial Fishing Methods: an Introduction to Vessels and Gears. Oxford,
UK: Fishing News Books. 359 pp.
Santana, J.C., Ariz, J., Pallares, P. and Delgado de Molina, A. (1991) Nota sobre la presencia de
mamõÂferos marinos en la pesquerõÂa de tu
Ânidos al cerco en el AtlaÂntico Este intertropical. ICCAT
Coll. Vol. Sci. Pap.35, 196± 198.
Scott, M.D. (1996) The tuna± dolphin controversy. Whalewatcher 30(1), 16±20.
Shimada, B.M. and Schaefer, M.B. (1956) A study of changes in ®shing effort, abundance, and yield
for yellow®n and skipjack tuna in the eastern tropical Paci®c Ocean. Inter-Am. Trop. Tuna
Comm. Bull. 1, 347±469.
Shomura, R.S. (1963) Mono®lament gillnet ®shing in skipjack tuna in Hawaiian waters, 1961±1962.
U.S. Fish. Wildl. Serv., Circ., 170. 21 pp.
Simmons, D.C. (1968) Purse seining off Africa's West coast. Comm. Fish. Rev. 30(3), 21±22.
Smith, T.D. (1983) Changes in the size of three dolphin (Stenella spp.) populations in the eastern
tropical Paci®c. Fish. Bull. U.S. 81, 1±13.
Smith, T.D. and Lo, N.C.H. (1983) Some data on dolphin mortality in the eastern tropical Paci®c tuna
purse seine ®shery prior to 1970. US Dept Commerce, NOAA Tech. Memo. NOAA-TM-NMFS-
SWFC-34.
Stevens, J.D. (1992) Blue and mako shark bycatch in the Japanese longline ®shery off South-eastern
Australia. Aust. J. Mar. Freshwater Res.43, 227± 236.
Stretta, J.M. and Slepoukha, M. (1986) Analysis of biotic and abiotic factors associated with schools
Tuna±dolphin problem 33
of tunas. In Symons, P.E.K., Miyake, P.M. and Sakagawa, G.T., eds. Proc. ICCAT Conf.
International Skipjack Year Program. Barcelona: Imprenta Juvenil, pp. 161± 165.
Stuntz, W.E. (1981) The tuna± dolphin bond: a discussion of current hypotheses. Southwest Fisheries
Center, National Marine Fisheries Service, Admin. Rept No. LJ-81-19, August, 1981.
Tomlinson, P.K., Tsuji, S. and Calkins, T.P. (1992) Length-frequency estimation for yellow®n tuna
(Thunnus albacares) caught by commercial ®shing gear in the eastern Paci®c Ocean. Inter-Am.
Trop. Tuna Comm. Bull.20, 359±398.
U.S. Public Law 103 ± 238 ± Apr. 30, (1994) Marine Mammal Protection Act, Amendments of 1994.
Vilicic, A. (1985) Povjesni Razvoj Suvremenog Tunolova. VI. nakl., Univ. of Split. 199 pp.
Wade, P.R. (1993) Estimates of historical population size of the eastern spinner dolphin (Stenella
longirostris orientalis). Fish Bull. U.S. 91, 775± 787.
Wade, P.R. (1994a) Abundance and population dynamics of two eastern Paci®c dolphins, Stenella
attenuata and Stenella longirostris orientalis. PhD dissertation. Univ. California, San Diego. 255
pp.
Wade, P.R. (1994b) Managing populations under the Marine Mammal Protection Act of 1994: a
strategy for selecting values for Nmin, the minimum abundance estimate, and FR, the recovery
factor. Background document. Report of the P.B.R. (potential biological removal) Workshop. June
27± 29, 1994. (Southwest Fisheries Science Center, La Jolla, California.)
Wade, P.R. (1995) Revised estimates of incidental kill of dolphins (Delphinidae) by the purse-seine
tuna ®shery in the eastern tropical Paci®c, 1959±1972. Fish. Bull. U.S. 93, 345±354.
Wade, P.R. and Gerrodette, T. (1993) Estimates of cetacean abundance and distribution in the eastern
tropical Paci®c. Rep. Int. Whal. Comm. 43, 477 ± 493.
Wahlen, B.E. and Smith, T.D. (1985) Observer effect on incidental dolphin mortality in the eastern
tropical Paci®c tuna ®shery. Fish. Bull. U.S. 83, 521± 530.
Witzell, W.N. (1984) The incidental capture of sea turtles in the Atlantic U.S. Fishery Conservation
Zone by the Japanese tuna longline ¯eet, 1978±1981. Mar. Fish. Rev. 46(3), 56± 58.
Yoshida, H. (1966) Tuna ®shing vessels, gear, and techniques in the Paci®c Ocean. In Manar, T.A., ed.
Proceedings, Hawaii Governor's Conference of Central Paci®c Fishery Resources. Honolulu: Bur.
Comm. Fish., Biol. Lab., pp. 67±89.
Accepted 30 September 1997
34 Hall
