|
|
Feeding patterns of juvenile Atlantic salmon (Salmo
salar)
Adam Storch and Theodore Smith
A Brief History of Atlantic salmon
The historic range of Atlantic salmon includes the North Atlantic Ocean
and its freshwater tributaries from Lake Ontario to Ungava Bay and southward
to the state of Connecticut. In Europe this species was found from Russia's
White Sea to coastal Portugal. A large portion of this range, to date,
has been drastically reduced, with remnant populations found in the rivers
of Ireland, the United Kingdom, the Faroe Islands, Iceland, Norway, Sweden,
Finland, Russia, France, Spain, Canada, and the United States (The Atlantic
Salmon Federation, 11 Aug. 2001).
Atlantic salmon, historically, were the only migratory salmonine native
to New York State. This species was abundant in many lakes and rivers
during the early nineteenth century. Various anthropogenic activities
such as urban development and the introduction of invasive
species resulted in the extirpation of Atlantic salmon by the latter portion
of the 1800's. Despite a greater public awareness of the problems facing
the existence these salmon, enormous tolls are continuously being exacted
on remaining populations
What are Atlantic salmon
Atlantic salmon are a member of the family Salmonidae.
Carolus Linnaeus first formally described Salmo salar, a Latin name meaning
"The Leaper," in 1758 (The Atlantic Salmon Federation, 11 Aug.
2001). When young, a primary identifying characteristic of these fish
is the large size of their pectoral fins. As adults,
many Atlantics display distinct x-shaped markings on their flank.
They are an anadromous species, spawning in freshwater habitats but spending
the majority of their life in marine environments. Unlike Coho a Chinook
salmon, Atlantics are non-semelparous, meaning that they do not die subsequent
to reproduction. The life history of these salmon is fairly complex, beginning
as a pea-sized orange egg 
buried
under several inches of streambed in autumn. Under this substrate, the
eggs incubate, and early the following spring thousands of tiny Atlantic
salmon alevin emerge with their nourishing yolk sacs still attached. From
this life-stage, the fish continues to develop until the yolk sac is completely
absorbed, at which point they are referred to as fry. Subsequently, the
young Atlantics move into open water to begin feeding. After a period
of time, the fish acquire vertical markings on the lateral portions of
their bodies, identifying them as parr. Proceeding a variable period of
stream life, usually two to six years, the salmon assume a distinct silver
color, and in early spring, the Atlantics, now described as smolts, move
into the sea. Before this migration occurs, body chemistry changes in
order to accommodate the move from freshwater to saltwater as well as
to stimulate a heightened capacity for growth. Salmon that remain in close
vicinity to the coast return to the rivers in which they were born after
one winter in the sea. These fish are referred to a grilse. Alternatively,
fish that travel to distant feeding grounds, remain at least two winters
in the sea before returning at a substantially larger size than their
counterpart grilse. Atlantic salmon have strongly developed homing instincts,
allowing for an accurate return to the same portion of the stream from
which they were born. At this point the cycle begins anew (Webster 1982).
There is an exception to this life history. Some populations of Atlantic
salmon never migrate to a marine environment. Instead, they inhabit lake
and river systems bordering the North Atlantic. These fish are termed
"landlocked" Atlantic salmon. The life of these salmon follows
a pattern similar to that of the previously described sea-run Atlantics,
however migration occurs between deep-lake feeding areas and spawning
grounds along shorelines or in tributaries (The Atlantic Salmon Federation,
11 Aug. 2001).
The Future of Atlantic salmon
Currently considerable research is being done in an effort to reestablish
sustainable populations in New York State, as well as many other regions
throughout their natural range. However, the past seems to have cast a
bleak shadow on the restoration potential of this charismatic species.
Between the years 1994 and 1999, the number of adult Atlantic salmon available
to return to North American rivers is estimated to have dropped from approximately
200,000 to 80,000 (The Atlantic Salmon Federation, 11 Aug. 2001).
Atlantic Salmon Restoration Project: Diet Analysis
Advancements in fishery biology and aquatic ecology has helped to create
a template for the restoration of Atlantic salmon in New York State, as
well as many other areas throughout their original range. In order to
properly evaluate the potential for this restoration it is imperative
to gain an understanding of virtually all associated aspects of the species.
Few studies have been conducted, and comparatively little information
has been accumulated, pertaining to Atlantic salmon feeding. Because diet,
and related interactions, may be seen as a major key to their survival,
studies involving the analysis of Atlantic salmon feeding may prove to
be integral portions of the much larger picture.
The primary objective of this study is to quantitatively evaluate feeding
patterns of juvenile Atlantic salmon in two New York State streams, Hemlock
Creek and the Owasco Lake Inlet. In addition it is our hope that a better
public understanding of the necessity of such endeavors, will result.
Methods and Materials
Preliminary
Work
Several thousand Atlantic salmon fry were stocked at various sites throughout
Cayuga and Oswego Counties, including Hemlock Creek and The Owasco Lake
Inlet. This was done to provide for later studies pertaining to the potential
for the restoration of this species in New York State.
Field Work
Age 0+ Atlantic salmon (less than one year of age) were collected from
Hemlock Creek and the Owasco Lake Inlet, both of which are found in the
town of Locke, Cayuga County, New York. Hemlock Creek flows in a northerly
direction for approximately 11.1 km befor
e
entering the Owasco inlet. The Inlet, also displaying a northbound flow,
originates approximately 14.7 km from the point at which the two bodies
of water converge, and subsequently continues for 14.0 km until reaching
its final destination, Owasco Lake. Because of their close geographical
association, both streams display similar topographical characteristics,
which have resulted from various glacial movements.
Hemlock Creek was sampled for Atlantic salmon at four times (7:00am, 10:45
am, 3:30 pm, and 8:00 pm respectively) throughout an approximate twelve-hour
period, on August 1, 2001. A total of 29 fish were collected from this
site using a back-mounted electroshocker set at a voltage of 300v and
a frequency of 90 Hz. These were found to be the most favorable settings,
through our previous fieldwork, to reduce inadvertent fish mortality.
Upon shocking, stunned fish were captured
in standard scap nets (29cm x 41cm) and immediately
placed in a food grade bucket, filled with
water at a temperature similar to that of the stream. Following
collection at each time interval, the length, weight, and gape width of
each fish was measured (an attempt was made to obtain a minimum of 10
0+ Atlantic salmon per sampling episode). Length
was recorded from the tip of the snout, to the end of the laterally compressed
caudal fin. Weight was measured on a digital
scale, and recorded to the nearest one-tenth of a gram. Width of gape
was determined to be the horizontal distance across the corners of the
mouth when opened, using a pair of calipers.
Immediately after these measurements were taken for each individual fish,
its stomach contents were extracted using a flushing
technique. The contents were then placed in a "whirl-pak"
bag filled with 70% alcohol and labeled accordingly. Upon processing
the fish, in the above manner, the invertebrate filled bags were organized
in a large cooler for storage. The salmon were then returned to the section
of stream from which they were removed, with care taken to revive them
when necessary.
The Owasco lake Inlet was also sampled for Atlantic salmon at four intervals
(7:00 am, 11:15 am, 3:00 pm, and 7:30 pm respectively) throughout an approximate
twelve-hour time period, on August 2, 2001. The above procedure
was also used for the processing of these fish. A total of 44 0+ Atlantic
salmon were collected from this site.
Lab Work
The contents of each "Whirl-pak"
bag were placed in a petri-dish and viewed under a stereomicroscope. Aquatic
invertebrates were identified to the level of family
whereas terrestrial insects were identified to order.
The relative number of each taxa were then tallied in reference to the
specific fish from which they were extracted. These data, and others,
were then organized into tabular form. Statistical analyses were performed
using a specialized computer software package.
Feeding Rates
The number of distinguishable invertebrates found in the stomach samples
of Hemlock Creek Atlantic salmon varied markedly from 0 to 111 individuals.
Alternatively, less variation was observed in the stomach content samples
of the Owasco Lake Inlet salmon, which ranged from 0 to 44. It was apparent,
that at any time period throughout the day, one could find individual
salmon containing large numbers of prey items and others containing nominal
numbers (Allan 1981).
The mean diversity index for invertebrates
found in the Hemlock Creek salmon began comparatively high during the
first sampling period and subsequently declined near mid-day. At approximately
3:00 pm the mean diversity index peaked at a value just exceeding
0.5. This value then once again subsided during the evening hours to a
similar value of that found for the early-morning sample.
The mean diversity index for the prey items extracted from the stomachs
of Owasco Inlet Atlantic salmon began with an initial peak at 7:00 am.
This figure then continued to decline through the mid-day and afternoon
time periods. During the 7:30 to 8:00 pm interval the mean diversity index
of removed prey items once again showed a relative increase reaching 0.4
(Figure 1.).
Body Length vs. Gape Width
The smallest Atlantic salmon, in reference to length, found in Hemlock
Creek was 57 mm, while the largest measured 80 mm. The minimum gape width
for this stream was found to be 7.3 mm, as compared to a maximum gape
of 9.8 mm. The minimum length for salmon collected from the Owasco Lake
Inlet was 60 mm. Alternatively, the longest Atlantic was 87 mm. The smallest
gape width of fish sampled in the Owasco Inlet 6.6 mm with a maximum of
10.4 mm. Comparisons between body length and gape width for the two streams
displayed a positive linear relationship (Figure 2.).
Using a significance value of P< 0.05 as a point of reference, the
P-value for this association was determined to be
0.006 with an r² (coefficient of determination)
of 0.154. Therefore, for both streams, as body length
increased, so to did gape width.
Gape Width vs. # of Insects
Through the analysis of data collected pertaining to gape width and the
number of extracted insects, a positive linear relationship was found
(Figure 3.). Comparison between these two measurements
generated a fairly significant P-value (P<0.05) of 0.001 and an r²
value of 0.2958, providing evidence that the number of invertebrates found
in the stomachs of Atlantic salmon for these two streams increased with
increasing gape width.
Gape Width vs. Diversity Index
There appeared to be no significant relationship between gape width and
diversity index for each individual Atlantic salmon (Figure
4.). For these data, the P-value was found to be 0.95, and the coefficient
of determination of 0.001.
Mean % Composition of Major Invertebrate Taxa Extracted During Four
Sampling Periods on Hemlock Creek
The mean invertebrate composition by percentage of Atlantic salmon stomach
contents for this stream during the early-morning sampling period (7:00
am) consisted primarily of Baetidae larvae and Chironomidae adults, comprising
40.9% and 40.4% of the salmonid's diet respectively. The remaining 18.7%
represented four different families, in addition to Chironomidae larvae
(Figure 5).
Both Baetidae larvae and adults, comprising 34% each, dominated the mean
composition by percentage of salmon stomach contents during the late-morning
period (10:45 am). Chironomid larvae and adults also contributed significantly,
representing 9.3% and 10.8% respectively. The remaining
11.9% was made up of three distinct families, two of which belonged to
the order Trichoptera (caddis flies), in addition to two more inclusive
groupings (Figure 6).
The afternoon sampling period (3:30 pm) displayed a mean percent composition
consisting primarily of Chironomidae adults (41.9%), Baetidae larvae (32.3%),
and Simuliidae adults(11.4%). The remaining 14.4% represented five families,
two belonging to the order Trichoptera and two of the order Ephemeroptera
(mayflies) (Figure 7).
During the evening (8:00 pm) sampling period, salmon diet seemed to be
predominantly composed of Baetidae larvae and Simuliidae adults, comprising
35.1% and 24.0% respectively. Significant contributions were also made
by Hydropsychidae larvae (15.9%) and Chironomidae adults (18.8%). The
remaining 6.2% was made up of two families and one unidentified Trichopteran
group (Figure 8).
Mean % Composition of Major Invertebrate Taxa Extracted During Four
Sampling Periods on the Owasco Lake Inlet
The mean percent composition of Atlantic salmon stomach contents for the
Owasco Inlet during the early-morning sampling period (7:00 am) consisted
primarily of Baetidae and Hydropsychidae larvae, comprising 39.5% and
38.1% respectively. Chironomid larvae also contributed significantly making
up 13.5% of the average diet. Eight families of invertebrates, one of
which was a terrestrial grouping, accounted for the remaining 8.9% (Figure
9).
Hydropsychidae larvae (32.3%), Baetidae larvae (23.2%), and Tipulidae
adults (17.4%) comprised the largest portions of the average salmon diet
during the late-morning sampling of the Owasco Lake Inlet. The remaining
27.1% consisted of three families and three, more inclusive groupings
(other, Hydrachnidia, and unidentified Trichopterans)(Figure
10).
During the afternoon (3:00 pm) sampling period, Atlantic salmon diet appeared
to be predominantly composed of Hydropsychidae larvae (53.9%) and Baeitidae
larvae (22.0%). Chironomidae pupae also contributed a significant proportion
(9.4%). The remaining 14.7% was comprised of five families (Figure
11).
The final evening (7:30 pm) sampling period for the Owasco Lake Inlet
displayed a mean percent composition of major invertebrate taxa which
was composed primarily of Baetidae, Hydropsychidae, and Hydroptilidae
larvae representing 40.0%, 18.2%, and 16.0% of the salmon diet respectively.
Seven families accounted for the remaining 25.8% (Figure
12).
Discussion
There appears to be little or no correspondence of mean diversity index
between Hemlock Creek and the Owasco Lake Inlet throughout the twelve-hour
periods in which they were sampled. The stomach contents of the fish collected
in Hemlock Creek shows a definite peak in diversity during the afternoon
sampling period, whereas the greatest mean diversity index for the Owasco
Lake Inlet reaches a maximum in the early-morning (Figure
1). Despite the geographic similarities between these two bodies of
water, they retain many other characteristics that differ (i.e. temperature,
vegetative cover, etc.). These may account for a portion of the variation.
However, without drift, benthic, and a more comprehensive survey of the
physical characteristics, only hypotheses may be made.
There are several factors to which the lulls in diversity of extracted
invertebrates may be attributed. At the times of day where these minimums
occur, it is possible that the collected Atlantic salmon where more selective
in their feeding, choosing certain individuals over others. In addition,
the decreases may have been caused by a lower rate of feeding during these
periods, thus lowering the probability that a larger array of insects
will be found.
A positive linear relationship was observed when comparing body length
and gape width of Atlantic salmon collected on both streams (Figure
2). This is substantiated by the small P-value (0.006), allowing for
the rejection of the hypothesis stating that there is no difference between
the length of collected salmon and gape width. Thus it may be assumed
that as body length increases so to does gape width. Unfortunately, the
minimal r² value may lend some question into the integrity of this
model. However, the general trend of this association may help to corroborate
evidence found in other analyses (i.e. gape width vs.
# of invertebrates and gape width vs. diversity). The relationship between
body length and gape width is understandable in that both are associated
with the overall growth of the Atlantic salmon.
The relationship between gape width and the number of invertebrates found
in the stomachs of the salmon collected from both streams also shows a
positive linear relationship (Figure 3). The small
calculated P-value (0.001) provides evidence of a similar trend, and as
in the above association, allows for the rejection of the hypothesis stating
that there is no difference between gape width and the number of extracted
insects. That is, as gape width increases so to does the number of extracted
invertebrates. In addition, the small r² value prompts some reservations
as to the accuracy of the model. Intuitively, it seems possible that the
number of extracted invertebrates could increase with
increasing gape width, however there may be many other factors contributing
to this relationship, which must be taken into account before a definite
association is established.
No correlation was found between gape width and the diversity of insects
found in the salmon of both studied streams (Figure
4). The large P-value (0.95) substantiates this finding. Therefore,
the hypothesis stating that there is no significant difference between
gape width and the diversity of extracted insects may be accepted. That
is, as gape width increases diversity tends to remain the same. The r²
value for this comparison was found to be an extremely small number, and
therefore the appropriateness of the model may be in question. As previously
explained, diversity of stomach contents could be attributed to a number
of factors, and therefore few conclusions can be drawn from such a weak
relationship, as was found in this analysis.
Hemlock Creek seemed to be dominated by three families of invertebrates;
Baetidae(mayflies), Chironomidae(midges), and Simuliidae(black flies).
From observations under a stereomicroscope it was found that both the
Chironomid and Simuliidae groups were comprised of very small individuals.
This may be a reason as to why the collected Atlantic salmon, in the juvenile
stage of their life-cycle, seem to have fed more on these invertebrates.
Atlantic salmon may have fed more intensley on all three of these families
due to their greater abundance in the habitat. However, this cannot be
deduced without proper drift and benthic analysis. (make "drift and
benthic analysis" a link bringing the reader to the "to be continued"
section) Overall the mean % composition of major invertebrate taxa extracted,
remained fairly consistent throughout the twelve-hour period, indicating
little change in feeding habit during different periods of the day. A
very small percentage of the average composition was made up of terrestrial
insects. This may indicate a greater preference for benthic as opposed
to drift feeding.
The Owasco Lake Inlet appeared to be dominated by only two families of
invertebrates; Baetidae and Hydropsychidae (caddisflies). The large presence
of Hydropsychids may indicate a preference for benthic feeding. It is
also possible that these individuals where the greatest source of prey
items. As with Hemlock Creek, feeding seemed to be fairly consistent throughout
the day, also indicating little variation in feeding. The Atlantic salmon
collected in this area seemed to consume a greater number of large prey.
Unlike the situation with Hemlock Creek, this consumption of larger sized
individuals, in all likelihood, cannot be attributed to their juvenile
status. Therefore greater numbers of large invertebrates in the stomach
contents may be concomitant with a greater abundance of large invertebrates
in the stream.
Conclusions
Through the analysis of various data it was found that the collected Atlantic
salmon, for whatever reason, were dependent on only a few families of
invertebrates. In addition a stable feeding pattern was displayed throughout
the day. Why is this study, and those similar to it important in the restoration
of sustainable Atlantic salmon populations? In addition
to causing detriment to salmon, compromising agents, such as habitat destruction
and the introduction of invasive species, can negatively impact the prey
items on which they feed. Therefore it is necessary to understand the
feeding habits of these fish in order to determine how trophic interactions
will be affected.
To Be Continued…
Significant amounts of data were collected, throughout the summer, in
addition to that displayed in this diet analysis study. While obtaining
stomach content sample, drift and benthic samples
were also collected. In addition, stomach contents were extracted from
two other species of fish, brown 
trout
(Salmo trutta) and rainbow trout (Oncorhynchus mykiss), in order to investigate
diet overlap between the three species. Fish were also sampled at other
various site in Cayuga and Oswego counties, in an effort to determine
population density, survival, and growth estimates of these stocked fish.
At all sites studied (a total of 19), physical stream characteristics
such as velocity, depth, width, temperature, and substrate composition
were recorded. Due to the immense amount of data that was collected, only
that presented in this particular study was analyzed. However, in the
very near future we will be observing the remainder of what was collected,
additions will certainly be made to this presentation, covering a much
wider range of Atlantic salmon biology.
Contact Information
Adam Storch: ephemerella4@hotmail.com
Theodore Smith: tsmith03@syr.edu
Acknowledgements
We would like to offer our utmost appreciation to the The Roosevelt
Wild Life Station for giving us this unique opportunity.
We would also like to thank Dr. Neil Ringler, Dr. Don Leopold, and future
Dr. Steve Coghlan.
About the Authors
Adam Storch and Theodore Smith are both seniors at the SUNY College
of Environmental Science and Forestry, each focusing their studies on
aquatic science. Their involvement with the ongoing SUNY-ESF Atlantic
salmon restoration project began during the summer of 2000 as field assistants
to PhD. Candidate Steven Coghlan. Through the generosity of the Roosevelt
Wildlife Station, the students have been able to conduct their own research
during the summer of 2001, under the guidance of Dr. Neil Ringler and
Steven Coghlan. Upon graduation, both Adam and Theodore plan to pursue
advanced degrees in aquatic science, hopefully leading to greater opportunities
in which to help restore Atlantic salmon to its natural range.
Literature Cited
Allan, J. David. “Determinants of Diet of Brook Trout (Salvelinus
fontinalis) in a
Mountain Stream.” Canadian Journal of Fisheries and Aquatic Science
38 (1981): 184-192.
The Atlantic Salmon Federation. “The Atlantic Salmon.” The Atlantic
Salmon’s Life. http://www.asf.ca/Overall/atlsalm.htm (11 Aug 2001).
Webster, Dwight A. “Early History of Atlantic Salmon in New York.”
New York Fish and Game Journal 29 (1982): 26-44.
|
|