Fish Populations, Following a Drought,
In the Neosho and Marais des Cygnes Rivers
of Kansas

BY

JAMES EVERETT DEACON

CONTENTS

TABLES

INTRODUCTION

This report concerns the ability of fish-populations in the Neosho and Marais des Cygnes rivers in Kansas to readjust to continuous stream-flow following intermittent conditions resulting from the severest drought in the history of the State.

The variable weather in Kansas (and in other areas of the Great Plains) markedly affects its flora and fauna. Weaver and Albertson (1936) reported as much as 91 per cent loss in the basal prairie vegetative cover in Kansas near the close of the drought of the 1930's. The average annual cost (in 1951 prices) of floods in Kansas from 1926 to 1953 was $35,000,000. In the same period the average annual loss from the droughts of the 1930's and 1950's was $75,000,000 (in 1951 prices), excluding losses from wind- and soil-erosion. Thus, over a period of 28 years, the average annual flood-losses were less than one-half the average annual drought-losses (Foley, Smrha, and Metzler, 1955:9; Anonymous, 1958:15).

Weather conditions in Kansas from 1951 to 1957 were especially noteworthy: 1951 produced a bumper crop of climatological events significant to the economy of the State. Notable among these were: Wettest year since beginning of the state-wide weather records in 1887; highest river stages since settlement of the State on the Kansas River and on most of its tributaries, as well as on the Marais des Cygnes and on the Neosho and Cottonwood. The upper Arkansas and a number of smaller streams in western Kansas also experienced unprecedented flooding (Garrett, 1951:147). This period of damaging floods was immediately followed by the driest five-year period on record, culminating in the driest year in 1956 (Garrett, 1958:56). Water shortage became serious for many communities. The Neosho River usually furnishes adequate quantities of water for present demands, but in some years of drought all flow ceases for several consecutive months. In 1956-'57, the city of Chanute, on an emergency basis, recirculated treated sewage for potable supply (Metzler et al., 1958). The water shortage in many communities along the Neosho River became so serious that a joint project to pump water from the Smoky Hill River into the upper Neosho was considered, and preliminary investigations were made. If the drought had continued through 1957, this program might have been vigorously promoted. Data on stream-flow in the Neosho and Marais des Cygnes (1951-'59) are presented in Tables 1-4.

These severe conditions provided a unique opportunity to gain insight into the ability of several species of fish to adjust to marked changes in their environment. For this reason, and because of a paucity of information concerning stream-fish populations in Kansas, the study here reported on was undertaken.

Table 1. Stream-flow in Cubic Feet per Second, Neosho River near Council Grove, Kansas. Drainage Area: 250 Square Miles

Table 2. Stream-flow in Cubic Feet per Second, Neosho River near Parsons, Kansas. Drainage Area: 4905 Square Miles.

Table 3. Stream-flow in Cubic Feet per Second, Marais des Cygnes River Near Ottawa, Kansas. Drainage Area: 1,250 Square Miles.

Table 4. Stream-flow in Cubic Feet per Second, Marais des Cygnes River at Trading Post, Kansas. Drainage Area: 2,880 Square Miles.

DESCRIPTION OF NEOSHO RIVER

The Neosho River, a tributary of Arkansas River, rises in the Flint Hills of Morris and southwestern Wabaunsee counties and flows southeast for 281 miles in Kansas, leaving the state in the extreme southeast corner (Fig. 1). With its tributaries (including Cottonwood and Spring rivers) the Neosho drains 6,285 square miles in Kansas and enters the Arkansas River near Muskogee, Oklahoma (Schoewe, 1951:299). Upstream from its confluence with Cottonwood River, the Neosho River has an average gradient of 15 feet per mile. The gradient lessens rapidly below the mouth of the Cottonwood, averaging 1.35 feet per mile downstream to the State line (Anonymous, 1947:12). The banks of the meandering, well-defined channel vary from 15 to 50 feet in height and support a deciduous fringe-forest. The spelling of the name originally was "Neozho," an Osage Indian word signifying "clear water" (Mead, 1903:216).

Neosho River, Upper Station.—Two miles north and two miles west of Council Grove, Morris County, Kansas (Sec. 32 and 33, T. 15 S., R. 8 E.) (Pl. 28, Fig. 2, and Pl. 29, Fig. 1). Width 20 to 40 feet, depth to six feet, length of study-area one-half mile (one large pool plus many small pools connected by riffles), bottom of mud, gravel, and rubble. Muddy banks 20 to 30 feet high.

According to H. E. Bosch (landowner) this section of the river dried completely in 1956, except for the large pool mentioned above. This section was intermittent in 1954 and 1955; it again became intermittent in the late summer of 1957 but not in 1958 or 1959.

A second section two miles downstream (on land owned by Herbert White) was studied in the summer of 1959 (Sec. 3 and 10, T. 16 S., R. 8 E.) (Pl. 29, Fig. 2 and Pl. 30, Figs. 1 and 2). This section is 20 to 60 feet in width, to five feet in depth, one-half mile in length (six small pools with intervening riffles bounded upstream by a low-head dam and downstream by a long pool), having a bottom of gravel, rubble, bedrock, and mud, and banks of mud and rock, five to 20 feet in height.

Neosho River, Middle Station.—One mile east and one and one-half miles south of Neosho Falls, Woodson County, Kansas (Sec. 3 and 4, T. 24 S., R. 17 E.) (Pl. 26, Fig. 1). Width 60 to 70 feet, depth to eleven feet, length of study-area two miles (four large pools with connecting riffles), bottom of mud, gravel and rock. Mud and rock banks 30 to 40 feet high.

According to Floyd Meats (landowner) this section of the river was intermittent for part of the drought.

Neosho River, Lower Station.—Two and one-half miles west, one-half mile north of Saint Paul, Neosho County, Kansas (Sec. 16, T. 29 S., R. 20 E.). Width 100 to 125 feet, depth to ten feet, length of study-area one mile (two large pools connected by a long rubble-gravel riffle), bottom of mud, gravel, and rock. Banks, of mud and rock, 30 to 40 feet high (Pl. 26, Fig. 2).

This station was established after one collection of fishes was made approximately ten miles upstream (Sec. 35, T. 28 S., R. 19 E.). The second site, suggested by Ernest Craig, Game Protector, provided greater accessibility and a more representative section of stream than the original locality.

DESCRIPTION OF MARAIS DES CYGNES RIVER

The Marais des Cygnes River, a tributary of Missouri River, rises in the Flint Hills of Wabaunsee County, Kansas, and flows generally eastward through the southern part of Osage County and the middle of Franklin County. The river then takes a southeasterly course through Miami County and the northeastern part of Linn County, leaving the state northeast of Pleasanton. With its tributaries (Dragoon, Salt, Pottawatomie, Bull and Big Sugar creeks) the river drains 4,360 square miles in Kansas (Anonymous, 1945:23), comprising the major part of the area between the watersheds of the Kansas and Neosho rivers. The gradient from the headwaters to Quenemo is more than five feet per mile, from Quenemo to Osawatomie 1.53 feet per mile, and from Osawatomie to the State line 1.10 feet per mile (Anonymous, 1945:24). The total length is approximately 475 miles (150 miles in Kansas). The river flows in a highly-meandering, well-defined channel that has been entrenched from 50 to 250 feet (Schoewe, 1951:294). "Marais des Cygnes" is of French origin, signifying "the marsh of the swans."

Marais des Cygnes River, Upper Station.—One mile south and one mile west of Pomona, Franklin County, Kansas (Sec. 12, T. 17 S., R. 17 E.) (Pl. 27, Fig. 1). Width 30 to 40 feet, depth to six feet, length of study-area one-half mile (three large pools with short connecting riffles), bottom of mud and bedrock. Mud banks 30 to 40 feet high.

According to P. Lindsey (landowner) this section of the river was intermittent for most of the drought. Flow was continuous in 1957, 1958 and 1959.

There are four low-head dams between the upper and middle Marais des Cygnes stations.

Marais des Cygnes River, Middle Station.—One mile east of Ottawa, Franklin County, Kansas (Sec. 6, T. 17 S., R. 20 E.) (Pl. 27, Fig. 2). Width 50 to 60 feet, depth to eight feet, length of study-area one-half mile (one large pool plus a long riffle interrupted by several small pools), bottom of mud, gravel, and rock. Mud and sand banks 30 to 40 feet high.

This section of the river was intermittent for much of the drought. In the winter of 1957-'58 a bridge was constructed over this station as a part of Interstate Highway 35. Because of this construction many trees were removed from the stream-banks, the channel was straightened, a gravel-bottomed riffle was rerouted, and silt was deposited in a gravel-bottom pool.

Marais des Cygnes River, Lower Station.—At eastern edge of Marais des Cygnes Wildlife Refuge, Linn County, Kansas (Sec. 9, T. 21 S., R. 25 E.). Width 80 to 100 feet, depth to eight feet, length of study-area one-half mile (one large pool plus a long riffle interrupted by several small pools), bottom of mud, gravel, and rock. Mud banks 40 to 50 feet high.

This section of the river ceased to flow only briefly in 1956.

METHODS

Electrical Fishing Gear

The principal collecting-device used was a portable (600-watt, 110-volt, A. C.) electric shocker carried in a 12-foot aluminum boat. Two 2 × 2-inch wooden booms, each ten feet long, were attached to the front of the boat in a "V" position so they normally were two feet above the surface of the water. A nylon rope attached to the tips of the booms held them ten feet apart. Electrodes, six feet long, were suspended from the tip and center of each boom, and two electrodes were suspended from the nylon rope. The electrodes extended approximately four feet into the water. Of various materials used for electrodes, the most satisfactory was a neoprene-core, shielded hydraulic hose in sections two feet long. These lengths could be screwed together, permitting adjustment of the length of the electrodes with minimum effort. At night, a sealed-beam automobile headlight was plugged into a six-volt D. C. outlet in the generating unit and a Coleman lantern was mounted on each gunwale to illuminate the area around the bow and along the sides of the boat (Pl. 3a). In late summer, 1959, a 230-volt, 1500-watt generating unit, composed of a 115-volt, 1500-watt Homelite generator was used. It was attached to a step-up transformer that converted the current to 230 volts. The same booms described above were used with the 230-volt unit, with single electrodes at the tip of each boom.

A 5.5-horsepower motor propelled the boat, and the stunned fish were collected by means of scap nets. Fishes seen and identified but not captured also were recorded. On several occasions fishes were collected by placing a 25-foot seine in the current and shocking toward the seine from upstream.

The shocker was used in daylight at all six stations in the three years, 1957-'59. Collections were made at night in 1958 and 1959 at the middle Neosho station and in 1959 at the lower Neosho station.

Seines

Seines of various lengths (4, 6, 12, 15, 25 and 60 feet), with mesh-sizes varying from bobbinet to one-half inch, were used. The 4-, 12-, and 25-foot seines were used in the estimation of relative abundance by taking ten hauls with each seine, recording all species captured in each haul, and making a total count of all fish captured in two of the ten hauls. The two hauls to be counted were chosen prior to each collection from a table of random numbers. Additional selective seining was done to ascertain the habitats occupied by different species.

Trap, Hoop, and Fyke Nets.—Limited use was made of unbaited trapping devices: wire traps 2.5 feet in diameter, six feet long, covered with one-inch-mesh chicken wire; hoop nets 1.5 feet to three feet in diameter at the first hoop with a pot-mesh of one inch; and a fyke net three feet in diameter at the first hoop, pot-mesh of one inch with wings three feet in length. All of these were set parallel to the current with the mouths downstream. The use of trapping devices was abated because data obtained were not sufficient to justify the effort expended.

Gill Nets

Gill-netting was done mostly in 1959 at the lower Neosho station. Use of gill nets was limited because frequent slight rises in the river caused nets to collect excessive debris, with damage to the nets.

Gill nets used were 125 feet long, six feet deep, with mesh sizes of ¾ inch to 2½ inches. Nets, weighted to sink, were placed at right angles to the current and attached at the banks with rope.

Sodium Cyanide

Pellets of sodium cyanide were used infrequently to collect fish from a moderately fast riffle over gravel bottom that was overgrown with willows, making seining impossible. The pellets were dissolved in a small amount of water, a seine was held in place, and the cyanide solution was introduced into the water a short distance upstream from the seine, causing incapacitated fish to drift into the seine. Most of these fish that were placed in uncontaminated water revived.

Rotenone

Rotenone was used in a few small pools in efforts to capture complete populations. This method was used to check the validity of other methods, and to reduce the possibility that rare species would go undetected. Rotenone was applied by hand, and applications were occasionally supplemented by placing rotenone in a container that was punctured with a small hole and suspended over the water at the head of a riffle draining into the area being poisoned. This maintained a toxic concentration in the pool for sufficient time to obtain the desired kill. Rotenone acts more slowly than cyanide, allowing more of the distressed fish to rise to the surface.

Dyes

Bismark Brown Y was used primarily at the upper Neosho station to stain large numbers of small fish. The dye was used at a dilution of 1:20,000. Fishes were placed in the dye-solution for three hours, then transferred to a live-box in midstream for variable periods (ten minutes to twelve hours) before release.

Determination of Abundance

In the accounts of species that follow, the relative terms "abundant," "common," and "rare" are used. Assignment of one of these terms to each species was based on analysis of data that are presented in Tables 9-16, (pages 402, 403, 404, 405, 408, 410, 411, 414-415, and 416). The number of fish caught per unit of effort with the shocker (Table 10) and with seines (Table 11) constitute the main basis for statements about the abundance of each species at all stations except the upper Neosho station. Species listed in each Table (10 and 11) are those that were taken consistently by the method specified in the caption of the table; erratically, but in large numbers at least once, by that method; and those taken by the method specified but not the other method.

For the species listed in Table 10, the following usually applies: abundant=more than three fish caught per hour; common=one to three fish caught per hour; rare=less than one fish caught per hour.

Tables 12-16 list all fish obtained at the upper Neosho station by means of the shocker, seines, and rotenone.

Names of Fishes

Technical names of fishes are those that seem to qualify under the International Rules of Zoological Nomenclature. Vernacular names are those in Special Publication No. 2 (1960) of the American Fisheries Society, with grammatical modifications required for use in the University of Kansas Publications, Museum of Natural History.

ANNOTATED LIST OF SPECIES

Lepisosteus osseus (Linnaeus)
Long-nosed Gar

The long-nosed gar was abundant at the lower and middle Neosho stations and the lower Marais des Cygnes station. Numbers increased slightly in the period of study, probably because of increased, continuous flow. The long-nosed gar was not taken at the upper Neosho station. At lower stations the fish occurred in many habitats, but most commonly in pools where gar often were seen with their snouts protruding above the water in midstream. Gar commonly lie quietly near the surface, both by day and by night, and are therefore readily collected by means of the shocker. Twice, at night, gar jumped into the boat after being shocked.

Young-of-the-year were taken at the middle and lower stations on both the Neosho and Marais des Cygnes rivers, and all were near shore in quiet water. Many young-of-the-year were seined at the lower Neosho station on 18 June 1959, near the lower end of a gravel-bar in a small backwater-area having a depth of one to three inches, a muddy bottom, and a higher temperature than the mainstream. Forty-three of these young gar averaged 2.1 inches in total length (T.L.).

Comparison of sizes of long-nosed gar taken by means of the shocker and gill nets at the lower and middle Neosho stations revealed that: the average size at each station remained constant from 1957 to 1959; the average size was greater at the lower than at the middle station; and, with the exception of young-of-the-year, no individual shorter than 13 inches was found at the middle station and only one shorter than 16 inches was taken at the lower station (Table 5).

Table 5. Numbers and Sizes of Long-nosed Gar Captured by Shocker and Gill Nets at the Middle and Lower Neosho Stations in 1957, 1958 and 1959.

Because collecting was intensive and several methods were used, I think that the population of gars was sampled adequately. Wallen (Fishes of the Verdigris River in Oklahoma, 1958:29 [mimeographed copy of dissertation, Oklahoma State University]) took large individuals in the mainstream of the Verdigris River in Oklahoma and small specimens from the headwaters of some tributaries. Because I took young-of-the-year at the lower Neosho station, it is possible that long-nosed gar move upstream when small and then slowly downstream to the larger parts of rivers as the fish increase in size. This pattern of size-segregation, according to size of river, merits further investigation.

Ripe, spent, and immature long-nosed gar (38 males and 10 females) were taken in three gill nets, set across the channel, 150 to 500 yards below a riffle, at the lower Neosho station on June 16, 17, and 18, 1959. On 23 June, 1959, 12 males and two females were taken in gill nets set 50, 150, and 400 yards above the same riffle. Operations with the shocker between 24 June and 10 July, 1959, yielded 29 males and three females. The fish were taken from many kinds of habitat in a three-mile section of the river.

Direction of movement as recorded from gill nets shows that of 67 gar taken, 45 had moved downstream and 22 upstream into the nets. Only ten of the above gar were taken from the nets set above the riffle; six of the ten were captured as they moved downstream into the nets.

On one occasion I watched minnows swimming frantically about, jumping out of the water, and crowding against the shore, presumably to avoid a long-nosed gar that swam slowly in and out of view. I have observed similar activity when gar fed in aquaria. Stomachs of a few gar from the Neosho River were examined and found to contain minnows and some channel catfish.

Long-nosed gar have a relatively long life span (Breder, 1936). This longevity and their ability to gulp air probably insure excellent survival through periods of adverse conditions. The population of long-nosed gar probably would not be drastically affected even in the event of a nearly complete failure of one or two successive hatches. Maturity is attained at approximately 20 inches, total length.

Collections at the middle Neosho station in 1958 indicate that the long-nosed gar is more susceptible to capture at night than in daytime (Table 9, p. 402).

Lepisosteus platostomus Rafinesque
Short-nosed Gar

Only one short-nosed gar was taken in 1957, at the lower station on the Neosho River. In 1958 this species was taken at the lower station on the Marais des Cygnes and in 1958 and 1959 at the lower and middle stations on the Neosho. More common in the Neosho than the Marais des Cygnes, L. platostomus occurs mainly in large streams and never was taken in the upper portions of either river. Although short-nosed gar were about equally abundant at the middle and lower stations on the Neosho, the average size was greater at the lower station (Table 6). This kind of segregation by size is shared with long-nosed gar, and was considered in the discussion of that species. Short-nosed gar were taken only in quiet water. Both species were collected most efficiently by means of gill nets and shocker. While shocking, I saw many gar only momentarily, as they appeared at the surface, and specific identification was impossible. The total of all gar seen while shocking indicated that gar increased in abundance from 1957 to 1959 (see Tables 5 and 6). Judging from the gar that were identified, the increase was more pronounced in short-nosed gar than in long-nosed gar.

At the lower Neosho station in 1959, two ripe females and one spent female were taken in gill nets (16, 23 and 17 June, respectively) and were moving downstream when caught. No males were taken in the nets. Subsequently, by means of the shocker (26 June-8 July), two spent and two ripe males were captured in quiet water of the mainstream that closely resembled areas in which the gill nets were set. No females were taken by means of the shocker.

Table 6. Numbers and Sizes of Short-nosed Gar Captured by Shocker and Gill Nets at the Middle and Lower Neosho Stations in 1958 and 1959.