- C — ‘— “WV—v \ i W ‘ I \ t l H 'H i WM WWII « H H H 1 $35 iIHWHIHI THS THE ECOLOGY OF WiLD DUCKS IN TWO CENTRAL MICHIGAN IMPOUNDMENT MARSHES Thesis {at the Degree of M. S. MICHIGAN STATE UNWERSITY Frederick James lgnafoski 19566 3 L— LIBRA R Y Michigaz‘. State University THE ECOLOGY 0F WILD DUCKS IN TWO CENTRAL MICHIGAN IMPOUNDMENT MARSHES By Frederick James Ignatoski A THESIS Submitted to Michigan State University in partia] fuifiiiment of the requirements for the degree of MASTER OF SCIENCE Departnent of Fisheries and Wildiife 1966 ACKNOWLEDGMENTS I particularly want to thank Dr. Ralph Blouch, Game Research Supervisor, Michigan Department of Conservation, who made this study possible. This research project was supported by Pittman-Robertson Project, Michigan w-95-R. I would also like to express my thanks to Dr. Jerry Duvendeck, Biologist in Charge of the Houghton Lake Wildlife Research Station, and the remaining staff, who assisted in the project. To Dr. Miles D. Pirnie, Professor at Michigan State University, who gave unselfishly of his time and knowledge to assist in all aspects of the study, I am deeply indebted. I also wish to thank Dr. John Cantlon and Dr. Peter Tack, whose advice and consultation at the start of the project was most helpful; and Dr. Gordon Guyer and Dr. William Drew for their critical reading of the manuscript. I thank Carl Bennett, Jr., whose statistical assistance was most appreciated. Last, but not least, I wish to thank my wife, Janice, who typed and re—typed the manuscript. ii TABLE OF CONTENTS Chapter Page INTRODUCTION ................................................ 1 The Dead Stream Area ................................... 1 The Pike Marsh Area .................................... 3 Objectives ............................................. 5 MATERIALS AND METHODS ....................................... 6 Vegetation Sampling .................................... 6 Aquatic Insect Samples ................................. 6 Brood Census Method .................................... 7 Duck Gullet and Gizzard Analyses ....................... 9 Caloric Analysis of Duck Foods ......................... 11 RESULTS ..................................................... 12 Vegetative Samples ..................................... 12 Insect Samples ......................................... 18 Duck Brood Censuses .................................... 18 Duck "Stomach" Analyses ................................ 20 Caloric Food Values .................................... 25 DISCUSSION .................................................. 27 Management ............................................. 3O SUlr-lMARY ..................................................... 33 BIBLIOGRAPHY ................................................ 36 Table LIST OF TABLES Page Statistical data on the plant species found in Houghton Lake Flats Flooding in July of 1965 ....... 13 Transect data of vegetation from the Dead Stream in August, 1965, expressed as percent of total vegetation and the percent each plant provided as emergent, floating, or submergent growths .......... 17 Caloric value of seeds and adult insects available to waterfowl in the Dead Stream and Houghton Lake Flats .............................................. 26 iv LIST OF FIGURES Figure Page 1 The central area of the Dead Stream Impoundment ...... 2 2 The Pike Marsh ....................................... 4 3 Blue—winged Teal brood using the type of canal which also facilitated the travel for the Northern Pike fry and the author ................................... 8 4 Frequency expressed as a percentage of plants found per milacre plot in Houghton Lake Flats Flooding ..... 12 5 The establishment of cattail and Spike rush on an area scalped by the dragline ......................... 15 6 Total number of aquatic insects per ten Ekman dredge samples in relation to season in the Pike Marsh ................................................ 19 7 Importance values for seeds and animal foods found in ducks taken in the Pike Marsh ..................... 22 8 Preference values for seeds and animal foods consumed by waterfowl taken in the Pike Marsh ................. 23 9 The Dead Horse Flooding, developed in 1962, in Missaukee Co. Michigan, illustrating the loss of willow and emergent vegetation due to prolonged high water, and resulting in poor waterfowl pro- duction and use ......... 3 ............................ 32 INTRODUCTION The two areas used during this study are located near Houghton Lake in central Michigan. The Dead Stream Impoundment, also known as the Muskegon River flooding or Reedsburg backwater, Michigan's first combination waterfowl-fishing-recreation impoundment, was de- veloped in 1938. The other area, known as the Houghton Lake Flats Flooding, or "Pike Marsh", was developed in 1964 to serve wild ducks and to increase spawning of the Northern Pike (Esox lucius). The Dead Stream Area The Dead Stream-Muskegon Impoundment floods about six miles of meandering river valley, and has long been known for its excellent fishing and hunting; and the flooding has attracted nesting Ospreys, Bald Eagles and Great Blue Herons. A number of studies have been done at this area, starting back in 1939 when cover mapping was done by Ben Jenkins (1939). Later, studies of waterfowl and plant suc- cession were done by Norcross (1952) and Di Angelo (1953). However, active management was delayed until 1964, when a "draw- down“ (the controlled lowering of water levels) was tried to "reju- venate" the marsh, hopefully to create more suitable habitat for wild ducks. Also, at that time, 42,720 feet of channel and 273 “potholes" were dug by dragline to create openings across the denser marsh to provide nesting spots and loafing islands (Figure 1). The new islands were seeded with brome grass, rye, and Chewing's fescue. Also, 13 one-chain sample strips were sprayed with "Dalapon" at the rate of ten pounds per acre to provide openings across dense stands 45.»in mfimum Hmcmu mmfioguoa cmcmz cmwco cmcoz Loupmz cm>wm commxmsx awcpm macaw ucm_mH of sedges. The blasting of 83 potholes (12' by 3' by 60') with amonium nitrate was also part of the new management program. Trees and brush were Clear-cut from islands to improve nesting conditions (Figure 1). Figure 1 also shows the meandering Muskegon River and adjacent to it the wetter part of the marsh (darker area on photo- graph). The drier area (lighter area on photograph), where most management work was done, lacks water during extensive dry periods, except for ditches and "potholes". The Houghton Lake Flats Flooding, or Pike Marsh Houghton Lake, Michigan's largest inland lake, has undergone drastic changes in the last fifteen to twenty years, as have many other Michigan lakes. Most shallow marsh edges, which formerly provided spawning habitat for Northern Pike, are now all but gone, due to the growing devel0pment of resorts. Therefore, in the spring of 1964, the Michigan Department of Conservation developed the Houghton Lake Flats Flooding (also called the Pike Marsh area) pri- marily to create new spawning habitat for Northern Pike and also to attract nesting waterfowl. This area of land, located at the west side of Houghton Lake, is bordered by two major highways (Figure 2). Draglines were used to dike the north and south ends, and to dredge the series of canals feeding into the main ditch. This was done primarily to supply travel lanes for the pike fry, when water levels are lowered. Also, along these canals, "ponds“ were dug to give more open area for wa— terfowl (along the two southernmost canals, Figure 2). Later, the draglines were used to dig more small ponds along the other canals. O . \-b> u . ‘ . .. r . . \ ‘0‘. s . vi? . Al. I Q . o‘ I ~ . ure- .4. b Struc .y Diton ~r'. ...;‘)1;H ”Id otholes Canal 1 . “”34“; a... 1: Control sla T .4. T3 T - D v; .5 cu um i...“ H. "n av +131 C (October 13, 1965) 8.33811. L, I‘ e K .1 P e h T Figure 2. These lanes greatly facilitated the author's duckboat travel for marsh observations and surveys. The aims of this study were to provide infonnation on the ecology of waterfowl in terms of duck food and cover at the two nearby im— poundments, based on: ~(1) an analysis of the plant communities of each area, and comparing duck brood use; (2) finding the most-used and preferred foods of ducks in the Pike Marsh - an area lacking many of the "quality“ foods; (3) taking aquatic insect samples in the Pike Marsh to determine availability to wild ducks; and (4) detennining whether the selection of seeds or invertebrates was related to caloric value. Waterfowl ecology studies carried out by Bennett (1938), Munro (1943), Mendall (1949), Swank (1949), and Keith (1960) have all pro- vided valuable information. However, little has been done to deter- mine the actual role which foods or vegetation play in determining waterfowl productivity - a major objective of my study. MATERIALS AND METHODS Vegetation Sampling Vegetation was sampled, in July of 1965 and 1966, in the Pike Marsh, by establishing random milacre plots throughout the marsh every a to 4% chains. These figures were arrived at by substituting from the table of random numbers 8 chain for each number, ex: 1 = a chain. At each plot, all woody vegetation (Salix) was counted as stems at water level, but herbaceous vegetation (Cargx) was counted by taking two square-foot samples from the upper right hand corner of each plot; and again, all stems were counted at water level. Some herbaceous vegetation was present in such small quantities as to per- mit an entire count. From these plots, data on density, frequency, and coverage could be determined and used in analyzing the composition of the community. In the Dead Stream, 3 one-hundred foot transects were set out to obtain figures on the types of vegetation present and the percent of species composition. This was done by counting the stems touching the transect line. These data were used mainly to have some idea of the vegetation in relation to brood abundance, not to determine a de- tailed makeup of the marsh. Aquatic Insect Samples Aquatic insect samples were taken as described by Gerking (1962). Using this method, ten randomly chosen sites were sampled every 2% weeks, from June 15 through October 30, with an Ekman dredge (225 cmz). Each sample was put in a waterproof plastic bag and taken into the lab to be sorted through three U. S. Standard sieves, 3360, 1160 and 500 microns. The use of sucrose flotation was also applied in sort- ing. For further information on this sampling method, refer to Gerking (1962). Identification of insects was made according to Usinger (1963). The sample size was the minimum in which to receive some statis- tical information, while the time between samples was considered to be a maximum in which one could observe population changes in insect life. Time was a factor in sampling, and a more comprehensive study would have to be carried out to secure data of greater statistical validity. Brood Census Method The determination of the number of broods using areas of such size is usually a very difficult and time-consuming project. I thought the best technique at this time was one proposed by Bennett (1965). Using his method, one or two individuals, observing for one-half hour at each of six predetermined sites located in and around the marsh, can get a fairly accurate brood estimate, with only three days observation. At these sites, data on the species of duck, age class and number per brood are recorded, whenever possible. The canals provided good observations, since the ducks used these openings in travel (Figure 3). Then, using the following equations, I estimated the number of broods. .- . 1M;¢~~9 P-u‘5'quam 5“,, _ . Figure 3. Blue-winged Teal brood using the type of canal which also facilitated the travel for the Northern Pike fry and the author. x = b B = estimated total broods log 10 (n+1) b = number of different broods seen n = x log 10e dB/dN n = total broods sighted N = .43429 x N = theoretical estimated total .008 brood sighting necessary to Total Broods (B) = x loglON see all the broods For further infonnation on the develOpment of the equations and the limitation of this method, refer to Bennett (1965). Duck Gullet and Gizzard Analyses In analyzing waterfowl for foods taken in the Pike Marsh, I used the gizzard, proventriculus and gullet of each specimen. Of the 53 ducks analyzed during the summer and fall, 15 were from duck trapping casualties, 28 were donated by hunters, and only ten were taken under a United States Fish and Wildlife pennit. Twenty-seven of the ducks used were Mallards and Black Ducks, and 26 were Blue-winged Teal; and no duck under a 2b class (mostly feathered ducklings) were taken. Most ducks were collected in early morning, although some were taken in midday and evening to observe any change in diet. The contents of each "stomach“ was first washed into a 250- milliliter beaker and any floating seeds or insects were taken out. The remaining materials were then filtered and all soft-bodied in- sects were removed, so as not to destroy them by heat. The residue was placed in an oven at 50 c. for one-half hour to dry. The mate- rial was then screened and the seeds sorted. This was done first by eye and then by use of a disecting microscope. Samples of organic lO material were also placed under a high-powered microscope for pos- sible identification of algae (Prescott 1950). I collected seed ‘plants from the marsh and classified them according to Fassett (1940). This collection was then used as a reference in classi- fying all seeds found in the "stomachs". The Munscher seed collec- tion (Cornell University) was also very helpful. In analyzing the stomach contents, I followed the method out- lined by Keith (1960), in which percentage volume, percentage occur- rence, importance value, and preference rating were obtained for each food species. Percentage volume is the percentage of seed (in milliliters) compared to the total amount of organic material. Per- centage occurrence is the percentage of ducks which had the material. The use of these figures individually would not give a clear account of waterfowl use, for percentage volume could be misleading if one duck had a large volume of seed, or if in the case of percent- age occurrence one could have small amounts of seed in many ducks. Therefore, the products of the two previous figures (called “import- ance value") is used to help minimize the error of either figure. Preference rating is then obtained by dividing the percentage avail- ability of each seed into the importance value. Percentage availa— bility of seeds was taken into account in the milacre plots and aquatic insect samples. 11 Caloric Analysis of Duck Foods In selecting a method for obtaining the caloric values of duck foods, I took into account the limited size of many samples and the time it would take for analysis. The bomb calorimeter, which is the most widely used instrument, proved to have two weaknesses: (I) only a few samples can be analyzed in one day, and (2) a relatively large volume of seed is needed for each analysis. Therefore, I used the organic analysis by dichromate oxidation, as proposed by Maciolek (1962). Maciolek stated that data taken by this method varies only by :5 percent, when compared to the bomb calorimeter. ' By use of this method, one can obtain results from 20 to 30 sam- ples a day, with only 20 to 30 milligrams of material. This proved very beneficial, since large quantities of some seeds were not avail- able. Caloric values are obtained by determining the amount of oxygen consumed from the potassium dichromate. The material being tested is placed in a 30-milliliter flask, with five milliliters of potassium dichromate and ten milliliters sulfuric acid, and placed in a boiling water bath for three hours, after which the solution is titrated with ferrous sulfate. Then, by use of the following formulas, one can ob- tain the caloric value for each material. Oxygen consumed (0. C.) 0.0. = (reagent blank titer-sample titer) x N ferrous solution x 8 Amount of material Gram calories gm. cal. = 0. C. in mg. x 3.4 3.4 = Kcal/g. 02 in crude protein, lipid, carbohydrate For further details on the derivation and procedures of this method were reported by Maciolek (1962). RESULTS Vegetative Samples The vegetation data, taken in July, 1965, from 70 milacre plots were analyzed in terms of frequency, density, and cover. The frequency is plotted (Figure 4) according to the Raunkier classification, in which the vegetation is divided into five groups: A - those plants found in zero to 20 percent of the plots; B - 21 to 40; C - 41 to 60; D - 61 to 80; and E - 81 to 100 percent. The category for each plant species can be seen in Table 1. A B C D E 70 I/////// ' ///////// 60 ///////// ///////// so ///////// ///////// 4o ///////// ///////// 30 ///////// ///////// 20 ///////// Percent of Frequency 10 ///////// ///////// .[[//l[/ ////////A (//////// (//////// ’//////// Ill/Ill! ///////// ///////// ///////// llll/lll Figure 4. Frequency expressed as a percentage of plants found per milacre plot in Houghton Lake Flats Flooding. Frequency indicates a very homogeneous community by having few individuals in columns B, C, and 0, although using frequency only, as a criterion for determining structure, is not very useful (Oosting 1940). "Nearest neighbor“ results, according to Evans, gt_a1, (1954), along with the milacre plots, indicated a non-random or a clumped com- 12 0 000.0 00 000.00 000.000 000.000 00.0 00 000000 000000 0 . 000.0 000.000 000.000 000.00 00.0 00 0000000000 000000000000 0 000.0 000.0 000.000.0 000.000 000.00 00.0 00 0000 000000 0 000.00 000.0 000.000.0 000.000 000.0 00.0 00 .00 00000 00003 0 00 000.00 000.000 000.000 00.0 00 0000000000000 00000 0 00 000.00 000.000 000.000 00.0 00 000000000 000000000 0 00 000.00 000.000 000.000 00.0 00 000000000 00000 0 00 000 000.00 000.000 000.000 00.0 00 0000000> 00000000000 0 000 000.000 000.000 000.000 00.0 00 0000000 00000 0 000 000.000 000.000 000.000 00.0 00 000000000> 0000 0 000.00 000.000.00 000.000.00 000.00 00.0 000 000000000 0000000000 0 000 000.00 000.000.00 000.000.00 000.00 00.0 000 0000000000 0000000000000 0 000 000.000 000.000.00 000.000.00 000.00 00.0 0000 .00 0000000 0 x0000 0 000.000 000.000.000 000.000.00 000.00 00.00 000 00000000000 X0000 0b maomumngm: hucmscwgm Lm>ou mo 0.0.50 .50 9:95 000000. msg#00300 0.0me mg» 0.0 9:5: :00: mp0: wwwumam ucmugmm mEmum 303 P0008. mocmvicou xmm x c 0.0 Mmm .0000 00 00:0 :0 00000000 00000 0000 0000000: :0 00:00 0000000 00000 000 :0 0000 0000000000m H 0000» 14 munity, with a homogeneous structure. Table 1 shows the species com- position and statistical methods used in obtaining results. Valid statistical information was achieved in the sedges (Carex sp.), wil- low (Salix sp.), and to some extent bluejoint (Calamagrostic canadensis), spike rush (Eleocharis palustris) and spirea (Spirea alba). The wide range of error with the remaining plant species is due to the small numbers in which they were present. Some of these plants represent new species, which are just invading the area, such as blad- derwort (Utricularia vulgaris), smartweed (Polygonum amphibium), dock (Rumex verticillatus), spike rush, cattail (Typha latifolia). and duck potato (Sagittaria latifolia). These final three plants appeared on the newly disturbed sandy or barren soil area exposed by ditching (Figure 5). In August, 1965, six months after flooding, samples from an area, adjacent to the west side of the Pike Marsh, showed 70.83 percent more sedges and 46.84 percent more bluejoint, as compared to the flooded area. This initial loss was not as severe as it may appear, since the original vegetation probably was much too thick for good duck use; and a mean of 12 stems per square foot of sedge (Car§x_sp.) provided excel- lent cover. However, the July,1966 studies revealed no statistical differences in vegetation in the second year of flooding. The square- foot samples taken in July,1966 revealed a mean of 14.2 stems of sedge (Carex_sp.) per square foot, but with 95 percent confidence limit no significant change took place from the previous mean of 12 stems per square foot. The same was true for bluejoint, 3.1 stems per square foot, and other sedges (Scirpus sp.) were 2.9 stems per square foot at 15 Figure 5. The establishment of cattail and Spike rush on an area scalped by the dragline. 16 this time; but, again, no significant difference was observed. The remaining plants could not be compared, due to their presence in small numbers. From the author's observations, there was no die—off of willow, spirea, leatherleaf (Chamaedaphne calyculata), low birch (Betula pu— mila), and blue flag (Iris versicolor) during the second year of flooding. Tag alder (Alnus rugosa), which was present but not found in sampling, suffered a die-off of 63 percent due to flooding. Al- though no statistical difference could be observed during the second year of flooding, some changes were taking place in the disturbed ditch area. Bladderwort and algae were the only plants found during the first year of flooding; but, in the second year large amounts of duck potato and pondweeds (Potamogeton gramineus and Potamogeton confer- ygjges) had invaded the area, allowing for additional food selection by waterfowl. The results of the 3 one-hundred-foot transects at the Dead Stream Flooding revealed a very different community. Each species is represented in terms of percent of total vegetation and percent of emergent, floating, or submergent vegetation (Table 2). One can see from these data that the area differs considerably from the Pike Marsh, and that it contains some excellent waterfowl plants, such as burreed (Sparganium spp.), wildrice (Zizania aquatica), soft rush (Juncus effusus), and pondweeds (Potamogeton spp.), also observed by Norcross (1952). Since the sampling was not extensive, some plant species, which occurred in small or scattered areas, were not found. 17 Table 2 Transect data of vegetation from the Dead Stream in August, 1965, expressed as percent of total vegetation and the percent each plant provided as emergent, floating or submergent growths. Total percent Percent of Species Common Name of vegetation Emergent Emergent Sparganium spp. Burreed 37.66 61.75 Eleocharis spp. Spike Rush 5.66 9.29 Juncus effusus Soft Rush 6.00 9.84 Zizania agpatica Wildrice 4.66 7.65 Pontederia cordata Pickeral Weed 1.33 2.18 Iypha latifolia Cattail 3.66 6.01 Sagittaria latifolia Duck Potato 0.66 1.09 Carex & Scirpus spp, Sedge 1.33 2.19 Floating Nuphar variegatum & Nymphea odorata Water Lily 24.00 48.00 Potamogeton spp. Pondweed 26.00 52.00 Submergent Myriophyllum spp, Water Milfoil 49.00 79.89 Utricularia vulgaris Bladderwort 12.33 20.11 18 Insect Samples Ekman dredge samples from June 15 through October 30, 1965 in the Pike Marsh, are graphed in relation to the number of insects and time (Figure 6). A peak occurred between July 15 and August I, mostly due to hatching eggs, and then steadily declined, possibly due to pre- dation by Giant Water-Bugs (Belostomatidae), Predacious Diving-Beetles (Dytiscidae), Water-Scorpions (Nepidae), Back-Swimmers (Notonectidae), and Dragonfly larva (Odonata). The loss was mainly of Water-Boatmen (Corixidae), which made up over one-half of the insect life through August. In October, more Damselfly (Odonata) and Midge (Chironomidae) larvae were found, but no significant differences were found from September 1 through October 15. The water-boatmen, damselflies and midges made up 80 percent of the insect life; and eight other insect families made up the remaining population. Insects in terms of biomass or calories might give a more ideal picture of the actual role they play in the aquatic environment (Odum 1959). However, since the ducks sampled made only a limited use of aquatic insects, these were not emphasized. No invnetory of insects was made in 1966, because of limited use by ducks and their great abundance. Duck Brood Censuses With the data obtained from the six observation sites and using the equation previously described, a minimum estimate of 44 broods of duck used the Pike Marsh flooding during the first week of July, 1965. NUMBER 6)! UNSECIS / ' I V V V 1 J! J‘ O ‘3 (9 <3 TE. c; 4* 4‘ <%\ ‘Tb ‘1 <1 Gt, G' ‘SA ‘1? ‘0’ (z, 6 ’ 1,. ‘ f, z 2‘,‘ °‘ Figure 6. Total number of aquatic insects per ten Ekman dredge samples in relation to season, in the Pike Marsh. 20 x = 26 N = 824.6109 Log 10 (51+1) x = 15.19 B = 15.19 x 2.92 N = .43429 x 15.19 .008 B = 44 I believe this figure to be reasonably accurate because from personal observations throughout the marsh in June and July, I knew of 38 different duck broods using the area. I am also sure I did not see all the ducks using the 390-acre flooding. Broods identified were: 16 Mallard, 12 Blue-winged Teal, 7 Black Duck, and one each of Greenwing, Woodduck, and Goldeneye. The 1966 duck brood census, done in a similar manner, again indicated 44 broods of waterfowl. Brood surveys in the Dead Stream Flooding were done in a like manner, and covered about 350 acres. Most of the 24 broods of water- fowl found there in 1965 were Blue-winged Teal (eight) and Mallard (six). Done in a similar manner, the 1966 calculated brood estimate was 16 broods of ducks using the area. There is no way to distinguish between local and migrant ducks, unless marked. Therefore, the first week of July was considered to be the Optimum time to take the brood census, since the majority of the ducks would be hatched and in a flightless condition. Duck "Stomach" Analyses It is very difficult to obtain infonnation on feeding habits of wild ducks. Much of the data reported in the past was from gizzard analyses only; and, due to the persistance of hard seeds for long periods of time, and the speedy breakdown of plants and soft-bodied 21 insects, an accurate picture is seldom received from a gizzard sample. A gullet sample shows chiefly what the duck was eating at the time of collection or death. Using "gullets“ only would be impossible, due to the large number of ducks needed to obtain adequate data, since many gullets are empty. But by using both the gullet and gizzard, some reliable information can be obtained. The data were analyzed (Figures 7 and 8) according to food spe- cies importance and preference values. These data attempt only to show waterfowl use of the Pike Marsh, and to find just what waterfowl were selecting from an area lacking species which usually are rated as "good" duck foods. My Mallard and Black Duck "importance values" agree with those of McAtee (1918) and Mendall (1949), in which sedges and smartweeds are the most used seeds; and that animal life, like snails (Helisoma), play an important role in the waterfowl diet (Figure 7). Blue-winged Teal data by Mabbott (1920) and Bennett (1938) also pointed out that sedges were the most commonly used food. The Mallard and Black Duck used more animal food than the Blue-winged Teal; but the Blue-winged Teal made greater use of Sphagnum capsules. No other statistical dif- ferences in use could be detected. Lee gt_al, (1964) stated that waterfowl are opportunists and feed upon food readily available. I think this would be true to some extent, but the preference ratings do point out that they are also selective in feeding. Preference was also reported by Pirnie, when three Lesser Scaups, shot while feeding on a wild celery bed, had eaten only snails (Pirnie 1935). £22 1000' -MAI.I.ARD mulllmmus-wmoeo mu Aw ==__=_=_________==___=__=______====_=_ 1%. on _==_.A%w Car V0 . 6+ ___==___====_==_==_=__=__======,_=_,__.=.__==_.=======_.=_===_======____==_=______ ova 0 0 5 wUZ .— ”— ~— -——o ——-—— —-—— —-—o~ —— _-— n-nu—u-r —— _. ——-v m —.——- —— »..— , h—— F."- _—~— -.-- w '_—- Figure 0. Preference veluee for eeede end enieel foods «new by «terfowl ' . teheh in the "he Mel-eh. 24 Other individuals, such as Munro (1943), Martin and Uhler (1939), Bellrose and Anderson (1943), and Korschgen (1955) have also indicated food preferences, which depend on area and season of year. Data ga- thered by these workers have proved that smartweeds, cutgrass, corn, millet, pondweeds, and bulrushes are very important foods for waterfowl. The Mallard and Black Duck differed statistically from Blue- winged Teal in their preferences for smartweed, raspberry (Bubu§_sp.), and snails (Figure 8). Smartweed and snails have long been known as excellent foods for waterfowl, but not the raspberry. However, Men- dall (1948) showed "Rubus" as a food used by Black Ducks in Maine. This suggests that different areas provide different preference fobds, depending on the plant and animal life available. No other differences were observed, either because both species of duck used the food in nearly equal amounts; or both made little use of a food, so as not to provide adequate data. Some seed plants, such as blue flag and water plantain (Alflflli plantago), which were obviously available but not found in a duck, suggests that there are certain foods which waterfowl obviously omit or make little use of. Microscope studies revealed algae present in almost every "stomach", with Microspora sp. making up 95 percent of the algae found. Exact volume of this alga could not be measured and, since it occurred in large blooms, it could not be determined whether this item was selected by the ducks or taken in accidentally while feeding. Regardless of means of ingestion, algae may add to the nu- tritional value of the diet. Due to the scarcity of ducks in the Dead Stream area, no speci- 25 mens were taken there during the Spring or summer; and only six were available from the hunting season. Examination of stomach contents showed burreed, pondweed, and bulrushes the most commonly eaten seeds. This was not surprising, since these seeds were readily available, and are known to be used widely by waterfowl in northern Michigan (Pirnie 1935). In the spring of 1965, so as not to disturb potential breeding birds or reduce productivity, only eight waterfowl were collected in the Pike Marsh. From examination of these gullets and gizzards, snails proved to be the most important food, both in volume and num- bers taken. Coulter (1955), in a spring food study in Maine, found the seeds of sedges and burreed as the most used food, and reported no heavy use or selection of animal foods. Caloric Food Values The caloric value of waterfowl foods as determined by the dichro- mate analysis can be seen in Table 3. These data pointed out that the caloric values of most seeds range relatively close, from 4.5 to 5.7 K-cal. Animal food, such as snails and aquatic insects, range slightly higher than the seeds, from 5.5 to 6.6 K-cal, with duckweed (Lgmna_sp.) the lowest, 2.90 K-cal. Since the actual "efficiency” of waterfowl on each of these seeds is unknown, it would be impossible to say with any certainty that preferences can be based on caloric value alone. Animal foods have a higher value than seeds, yet only snails were heavily used by the Mallards and Black Ducks. Hater-boatmen, which had the highest value, and duckweed the lowest value, both had low use and poor preference ratings in the few examples studied. 26 _mu-¥ nfi.¢ mesz .aam mopcuocvgum quux m.¢ Agmcwu comm mmumcwucsgm mvcmpmca Pmoux No.¢ cmmzucmsm Esvawcqsm Escomeoa quux m.N ummzxoso .Qam mesa; Fmoux m.¢ opmuom xozo mwpowwpmr mwcmupwmmm Pmuux m.m mmcoam Ezcmmsam .amm Eanmmzmm Pmuux Ho.muwm.¢ mowcvaz muwpmmcm mwcmNPN Pmuux No.¢ coagucmsm EzwawsaEm Encomxpoa _mu-x Ho.m mmumm ”mam xmgou Foolx Ho.m mmvmm nmmmlmwumm Pmuux mo.m ummccsm Esmmexczm Eswcmmcmam quux mm.m xgcmnammm mUmew; mznzm :83. Z :2; 28$ 6: :83. 2 :25 Ma? Pmuux N.m swam mxwam mwcpszma mwcmgooum Faunx n.m swam wxwam mwcum:_ma,mwgmguom~m qunx 0.0 cmgamomucmpmz mmvwxvgou quux m.o cmspmomucmpmz mwmmeflmMM m:_m> owcopmo msmz :oEEou mmwumqm maFw> uwcopmu mEmz coEEoo mmvumam Emmcpm umma mpg—m mxme copgmso: .Emwcpm name as“ vcw mum—m mxmg coycmzo: cw Facmcmumz op mpamrwm>m mpumMCW “Pane new mnmwm Lo mm3~m> quOqu m mpnmh DISCUSSION Comparing two communities in relation to waterfowl ecology in terms of food and cover is a very complex proposition, since it in- cludes the changes in vegetation and other environmental influences, as well as the behavior of ducks themselves. However, the author has tried to obtain the results from the methods known to him, and has tried to evaluate them objectively. Sampling communities calls for a decision as to the size and time for samples which are needed to achieve reasonably accurate and meaningful data. The Pike Marsh habitat was sampled in July of both years, when broods were making greatest use of the area, and when most fruits could be found on the plant life. These data showed good statisti- cal validity on plants present in large quantities. However, it would be impossible to obtain sufficient data on plants newly invad- ing the area, or those present in small quantities. Data obtained from the 70 milacre plots in the Pike Marsh, during the first and second year of flooding, revealed a monotype community of sedges and willow, with scattered bluejoint, blue flag, spirea, and leatherleaf, along with newly invading aquatic plants such as water smartweed, cattail, duck potato, bladderwort, pondweeds and spike rush, with no significant differences in two years of flood- ing, except in the disturbed ditches. The Dead Stream Flooding consisted of many different kinds of emergent, floating and submergent vegetation such as burreed, wildrice, pondweeds and bulrushes, all of which are known to contribute to wa- terfowl diets (McAtee 1918, Pirnie 1935, Bennett 1938). ly observa- 27 tions revealed large stands of sedges with scattered bluejoint, smart- weed and duck potato. These also were observed by Norcross (1952). Duck brood inventories estimated a minimum of 44 broods of duck in 1965, and the same in 1966, in the Pike Marsh flooding, as com- pared to 24 and 16 in the Dead Stream area for these same years. The accuracy of the calculated brood estimates on the Pike Marsh is con- finned by many personal observations of flightless ducks during June, and July of both years. The Dead Stream area poses a greater problem, since much of the area is dense, emergent vegetation; and observations are very diffi- cult to make. Even the duck banding program tried in 1965 was of little help in estimating the number of broods. Therefore, if this area of the Dead Stream had significantly fewer broods of duck, it could be concluded that a single factor, or a combination of them, must influence the brood use in each water- fowl area. These factors being: (1) the lack of woody vegetation might have been restricting waterfowl use, since the Dead Stream had the same aquatic plants as in the Pike Marsh, plus many others, but lacked the woody shrubs of willows and spirea; and (2) since predation can have a great effect on a nesting population, raccoon and other predators in the Dead Stream may have affected the duck brood popula- tion there. The Pike Marsh was given a complete drawdown at the end of the 1965 duck season, in November; but even with the reflooding in early March, no shortage of animal foods has been observed in June, July and August, when most duck broods use the area. This could be ac- counted for since many insects overwinter as adults and can fly into the area in early spring, to mate and lay their eggs. Also, insects may be pumped into the area from Houghton Lake while flooding, and since all ditches maintain water throughout the year, insects may disperse into the shallow water after flooding. The stomach analyses of 53 ducks from the Pike Marsh revealed that food plays no determining role in attracting breeding waterfowl. The ducks used mostly sedge seeds and snails, but seeningly preferred such foods as smartweed and dewberry seeds. This is nothing out of the ordinary, since previous food studies by McAtee (1918), Pirnie (1935), Martin (1938), and Mendall (1948) have all indicated the im— portance of various seeds in the wild duck's diet. The Dead Stream provided these same aquatic plants plus wildrice, burreed, pondweed, and others, which are excellent duck foods. There- fore, if food supplied detennine waterfowl production and brood use, more or at least as many broods should be using the Dead Stream area. Another test was conducted in the spring of 1966, when corn was added at the Pike Marsh, with the hope of attracting breeding pairs. Four ducks (two Mallard, two Blue-winged Teal) were collected near the baited area, but only snails and sedge seeds were found in their gullets. It is possible they didn't find the corn, didn't have time to eat it, or did not care for it. Therefore, even though some ducks probably used the corn, it failed to attract large concentrations, since no more than four ducks were ever seen in the general area at one observation. Pirnie observed a situation on Houghton Lake where an area baited in May, was passed over by Lesser Scaups; but on 30 Wintergreen Lake, near Battle Creek, Michigan, Redheads found and used an offering of wheat in less than one hour! My results indicated no selection of seeds or animal life accord- ing to caloric value; and the caloric values of the seeds most used or preferred differ little from those that seemingly were overlooked. However, it may be possible that certain foods provide an essential nutrient or vitamin, no matter how low or high the calories. A more detailed study on the percentage of fats, protein, carbohydrates and possible vitamins of each seed should be conducted along with isotope feeding, so as to determine the actual roles of plant and animal foods in diet. This type of study may explain why some ducks chose one food over another. Management In looking at these area changes from a management standpoint, .each plays an important role for waterfowl. The Pike Marsh, evidently, serves as a duck rearing marsh, but in order to maintain this for waterfowl and Northern Pike, the following management procedures may be necessary: (1) Regulate water level by pumping in 18 inches of water in early spring, March lst, for pike spawning and waterfowl use. (2) Drawdown 12 inches just after the pike Spawn to return adults into the lake and prevent loss of ducklings; and bring water back to original level. (3) Drawdown around June lst, to release pike fry into the lake; then return water to 12 inch-level. (4) Drawdown around August 15, to prevent loss of grasses needed for pike spawning; this will also prevent the loss of shrubby woody vegetation such as willow and spirea. This type of management should retard the open pond succession, as ob- served in the Dead Horse Impoundment of once similar vegeta- tion (Figure 9). However, by fluctuating water levels, this type of habitat can be avoided. (5) If hunting is desired, water can be returned into the marsh prior to hunting season. It has been observed that this type of management will pennit good brood production and pike spawning. However, it will be conceded that some aspects will be subject to alteration, depending on future ecological changes. i k * . .‘lll .‘ V . .1 _‘ f’ I ‘ J ;¢#'Inma. Figure 9. 32 The Dead Horse Flooding, developed in 1962, in Missaukee 60., Michigan, illustrating the loss of willow and emergent vege- tation due to prolonged high water, and resulting in poor waterfowl production and use. SUMMARY The objectives of this study were to provide infonnation on the ecology of wild ducks, relating to food and vegetative type in two central Michigan impoundnents. Houghton Lake Flats, or Pike Marsh, is a 390-acre area, which serves as a Northern Pike spawning habitat and a waterfowl nesting area in the early spring. The Pike Marsh community is a monotype, consisting mainly of sedge and willow. The Dead Stream, however, is a diverse community of emergent, floating and submergent aquatic vegetation. Brood surveys in two successive years revealed a minimum of 44 broods of duck using the Pike Marsh, while the Dead Stream data showed 24 broods in 1965 and 16 in 1966. However, brood counts on the Dead Stream were very difficult, due to the dense stands of emergent vegetation; but the figure for this area is not too greatly underestimated, since a duck banding program in the Dead Stream in 1965 was not nearly as successful as in the Pike Marsh area. The greater population on the Pike Marsh, as compared to the Dead Stream, could be explained in two ways: (1) The shrubby, woody vegetation, which was so much more abundant in the Pike Marsh, was more condusive to nesting waterfowl. This could be possible, since, in early spring, most herbaceous cover is gone and only the woody shrubs are erect to give cover. (2) Predation by raccoon and other predators in the Dead Stream could have caused a reduction in nesting birds and broods. 33 34 The invasion of aquatic insects in the Pike Marsh is large enough each spring to produce a population of substantial numbers, so as not to cause a shortage of animal food to the newly hatched ducklings. Food studies of 53 wild duck gizzards and gullets during the summer and fall in the Pike Marsh established the fact that sedge seeds and snails were the most used foods, and smartweed and dewberry the most preferred, all of which are known to be used by waterfowl. Since these foods were also available at the Dead Stream, along with wildrice, pondweeds and burreed, all being excellent duck foods, it was concluded that food was not a major factor in holding breeding waterfowl. This was also observed in a spring food study, when an area baited with corn was little used, and no increase in ducks could be observed. It is possible the ducks did not find the corn; or they just did not use it. The caloric values of seeds and invertebrates, known to be used by wild ducks, differ little from those seeds which were seemingly overlooked. Therefore, no selection of seeds or animal life was made according to caloric value. But further studies must be carried out to detennine the actual role of each food in the diet, since each food could provide an essential nutrient, regardless of caloric value. In conclusion, each area plays an important role to waterfowl. An area such as the Pike Harsh, can be used to benefit both Northern Pike and waterfowl; and the Dead Stream, to provide a feeding and loafing site for local and migrating waterfowl, plus serve as a brood production area. 35 A management program can be set up for the Pike Marsh, to main- tain habitat for Northern Pike spawning and wild duck rearing, as was observed from the results of the second year of flooding. By fluc- tuating the water level at the right time, both pike and ducks can benefit. It is thought that this controlled lowering of water levels will retard the open pond succession, which has been detrimental to most flooded areas. BIBLIOGRAPHY Bellrose, F. C. and H. G. Anderson. 1943. Preferential ratings of duck food plants. Ill. Nat. Hist. Surv. Bull. 22 (5):417-433. Bennett, Jr., Carl. 1965. Some methods of waterfowl and marsh in- vestigations. Unpub. M. S. Thesis. Mich. State Univ., 42 pp. Bennett, L. J. 1938. The blue-winged teal. Collegiate Press Inc., Ames, Iowa, 44 pp. Coulter, M. N. 1955. Spring food habits of surface feeding ducks in Maine. J. Wildl. Mgmt., 19 (2):263-267. Di Angelo, Stephen. 1953. Aquatic plant succession at certain waterfowl flooding projects in Michigan. Unpub. M. Nildl. Mgmt. Thesis. Univ. of Mich., 112 pp. Evans, F. C., P. J. Clark, and R. H. Brand. 1955. Estimation of the number of species present on a given area. Ecology, 36 (2):342-343. Fassett, N. C. 1940. A manual of aquatic plants. McGraw-Hill Book Co., Inc., New York, vii + 381 pp. Gerking, S. P. 1962. Production and food utilization of Bluegill and Sunfish. Ecol. Mono. Vol. 32, No. 1, 50-62. Jenkins, B. C. and Cash H. Nonser. 1939. List of cover and soil types of Dead Stream area. Interoffice memo. Mich. Dept. Cons., 2 pp. Kadlec, J. A. 1960. The effect of a drawdown on the ecology of a waterfowl impoundment. Game Div. Rep. No. 2276, Mich. Dept. Cons., 181 pp. Keith, L. B. 1961. A study of waterfowl ecology on small impound- ments in southeastern Alberta. Nildl. Mono. No. 6, 88 pp. Korschgen, L. J. 1955. Fall foods of waterfowl in Missouri. Mo. Fish & Game Div., P-R Series No. 14, 41 pp. Lee, F. B., R. L. Jessen, N. J. Drdal, R. I. Benson, J. P. Lindmeier, and L. L. Johnson. 1964. Minn. Dept. Cons. Tech. Bull. No. 7, 210 pp. Mabbott, D. C. 1920. Food habitats of seven species of American shoal-water ducks. U. S. Dept. Agr. Bull. No. 862, 68 pp. 36 37 Maciolek, J. A. 1962. Limnological organic analyses of quantitative dichromate oxidation. U. S. Dept. Int. Res. Rep. No. 60, Pub. by Bureau of Sport Fish. 8 Nildl., Nash., D. C., 61 pp. Martin, A. C. and T. M. Uhler. 1939. Food of game ducks in the United States and Canada. U. S. Dept. Agr. Tech. Bull. No. 634, 151 pp. Mc Atee, N. L. 1939. Wildfowl food plants. Collegiate Press Inc., Ames, Iowa, 141 pp., illus. Mendall, H. L. 1949. Food habitats in relation to black duck manage- ment in Maine. J. Nildl. Mgmt., 13 (1):64-101. Moyle, J. B. and M. Hotchkins. 1945. The aquatic and marsh vegeta- tion of Minnesota and its value to waterfowl. Minn. Fish Res. Lab. Tech. Bull. No. 3, 122 pp. Munro, J. A. 1943. Studies of waterfowl in British Columbia. Mallard, Cana. J. Res. 21:223-260. Norcross, J. J. 1952. An ecological study of the Dead Stream Flood- ing project, Roscommon County, Michigan, with special reference to waterfowl. Unpub. M. S. Thesis. Mich. State Univ., 52 pp. Odum, E. P. 1959. Fundamentals of ecology. 2nd Edition. N. B. Saunders Co., Philadelphia and London, 546 pp. Dosting, H. J. 1956. The study of plant communities. 2nd Edition. N. H. Freeman and Co., San Francisco and London, vii + 440 pp. Pirnie, M. D. 1935. Michigan waterfowl management. Mich. Dept. Cons., Game Div., 328 pp. Prescott, G. N. 1951. Algae of the western Great Lakes area. Cran- brook Inst. of Sci. Bull. No. 31, vii + 946 pp. Reid, G. K. 1961. Ecology of inland waters and estuaries. Reinhold Publ. Corp., N. Y., 375 pp. Swank, N. G. 1949. Beaver ecology and management in Nest Virginia. Div. of Game Mgmt., Cons. Comm. of West Va., Bull. No. 1, 65 pp. Usinger, R. L. 1963. Aquatic insects of California. Univ. of Calif. Press., Berkley and Los Angeles, 507 pp. "111111111 1111111111111