591,97
I 878

Yomi2, R. T,

'■E>9

The life of Devils lake,
Ilorth Liakota.

North Dakota Biolcs^ical Station
1924

QH
105
Y 79

/ '

THE LIFE OF DEVILS LAKE

NORTH DAKOTA

BY

R. T. YOUNG

■jgjyge*.

••BA^*

PUBLICATION OF THE NORTH DAKOTA
BIOLOGICAL STATION

1924

THE LIFE OF DEVILS LAKE

NORTH DAKOTA

BY

R. T. YOUNG

mo fWt^fn

'^*S5&*»'''

PUBLICATION OF THE NORTH DAKOTA

BIOLOGICAL STATION

1924

^^rav> •^- ^

^- ^gvc.

ijp™*

MATE'S 1928 /

TABLE OF CONTENTS

Page

Introduction 5

Location and History 7

Climate — Rainfa'll and run-off, temperature, humidity and evaporation,

wind 13

Main Lake — Physiography 17

Physics 18

Chemistry 28

Biology 32

Life Zones 33

Plancton Methods and Results 34

East Lake 48

Mission Lake 49

Lamoreau Lake 50

Stump Lake 50

Lake A 54

Lake C 55

Spring Lake 55

Lake N 56

Lake 0 56

Ivake P 56

Annotated list of species — Bacteria 57

Algae 60

Diatomaeeae 68

Spermatophyta 74

Protozoa 75

Planaria 81

Nematoda 81

Rotifera 81

Gastrotricha 87

Crustacea 87

Insecta 94

Hydracarina 98

Pisces 98

Amphibia 99

Aves 100

Mammalia 100

Parasites 100

Discussion 100

Conclusions Ill

Bibliography Ill

Index 115

3 4 5 i) C

INTRODUCTION

In 1909 a biological laboraton'- Avas established by the State of
North Dakota, Tinder the auspices of the State I^'nivcrsity, on the
shore of Devils Lake. Since that time the writer, acting at first as
assistant and later as director, has been engaged in studying the
life of the lake, and this report is the result of such studies. His
thanks are due principally to the former director, Dr. INI. A. Bran-
non, now Chancellor of the University of Montana, to whose initiative
the laboratory owed its establishment, for his encouragement and
friendly assistance during the first five years of this work. He has
also received much valuable assistance from his present colleagues,
Professors H. E. Simpson, G. A. Abbott and Karl Fussier and
E. D. Coon, and from Drs. F. H. Heath, C. E. King and Messrs.
Robert Hulberd, Eric G. Moberg and R. H. Johnstone, formerly of
the University of North Dakota. Professor Simpson's valuable
work on the Physiography of Devils Lake (1912) is the basis of
that part of the present report descriptive of the history of Devils
Lake. Drs. Heath and King have furnished the major part of the
chemical data. Mr. Moberg, as my former assistant, has aided in
the collection of material and has made most of the plancton counts,
while the other gentlemen have kindly assisted me in various ways.

Several specialists have aided in the identification of organ-
isms, without which assistance the work could not have been carried
on. Messrs. C. F. Rousselet and David Bryce have identified the
rotifers, Professors C. J. Needham and R. Mattheson and Messrs.
C. R. Plunlcett and C. K. Sibley the insects. President E. A. Birge
and Professors Chancey Juday, A. Willey and C. Dwight Marsh
the Crustacea, Dr. N. A. Cobb, the nematodes. Dr. Karl Viets and
Professors R. H. Wolcott and Ruth Marshall, the mites, and Drs.
Geo. T. ]\Ioore and Nellie Carter, the algae. Professor C. H. Edmond-
son has spent several weeks at the station in 1914 and 1917, working
on the Protozoa, and has worked over several collections sent him at
other times. Professor C. J. Elmore has identified the diatoms, in part
from collections sent to him, and in part from material collected
by himself during a visit to the lake in 1915. To all of these gentle-
men the writer's thanks are due for their invaluable aid. Professor
Jacob Reighard and President E. A. Birge have given very helpful
advice regarding methods and apparatus, the former loaning the
proof of a chapter in Ward and Whipple's "Fresh Water Biology"
which deals with these subjects.

The charts accompanying the paper are the work of Mr. Wilfred
P. Lowe.

The first two years were spent chiefly in collection of apparatus,
development of methods and in a general preliminary survey of the
lake. Since that time collections, both qualitative and quantitative,

of the organisms inhabiting the lake, together with physical and
chemical studies of its water, have been made at freqvient intervals
from 1911 to 1914 with the exception of certain periods in 1912 and
1913, during which time the absence of the writer interfered with
the conduct of the work, and during the winter when few collections
were made. A few collections were made in 1909 and 1910, but
these were preliminary to the main series. An additional series was
also made between December, 1922, and July, 1923. The collections
and determinations, aside from the purely preliminary work, have
thus covered a period of twelve years (1911-1923).

The major part of the work has been devoted to a study of the
plancton of the lake in its ecological relations, this forming the most
important part of the life of the lake. Several general chemical anal-
yses (both organic and inorganic) have been made during the study,
and analyses of certain constituents, especially the ammonias, nitrates,
nitrites, carbon dioxide and oxygen, have been made during part
of the time at much more frequent intervals. This will be explained
in detail in the proper place. In connection with the collections and
analyses temperature records have been kept, and a study of the
distribution of light at various depths has been made. The data of
the U. S. Weather Bureau Station in the city of Devils Lake, eight
km. distant, have been drawn upon extensively, and these have been
amplified to a certain extent b}^ our own more or less fragmentary
records of air temperatures and precipitation taken at the station.

There are several reports on various phases of the biology of
the Devils Lake region beginning with that of Pope (1908), but
nothing as yet of a comprehensive nature. There are brief notes
on the physiography of the region in Upham (1896) and Willard
(1921), while Simpson (1912) gives a thorough treatment of this
subject.

HISTORY AND LOCATION

The uortliern United States and Canada are dotted with numer-
ous picturesque lakes bounded by rolling- hills — legacies of the great
ice sheet, which thousands of years ago held this region in its
frigid grasp. During its retreat from the upper Mississippi Valley
its waters were in part collected in a great lake (Lake Souris) in
northern central North Dakota and southwestern Manitoba, occupy-
ing the watershed of the present IMouse River, and in part that of
the Assiniboine, while further to the north in the valley of the
Saskatchewan, was glacial Lake Saskatchewan.

Before the ice had retreated sufficiently to uncover the northern
outlets of these lakes, both of the latter at first drained southward
into the Missouri River. Later new outlets were opened, at first
thru the James River and later the Sheyenne into Lake Agassiz.
"With the retreat of the ice north of the Shej'enne River its drain-
age was collected in a broad, shallow valley running from north-
west to southeast between longitude 98 and 99 and latitude 47 and
4S, while sout?i of this depression a tumbled mass of moraines
served as a dam to the glacial waters, thru which different channels
at various times conducted them southward into the Sheyenne. Thus
was formed the glacial lake Minnewaukan* the forerunner of tho
present Devils-Stump Lake complex in North Dakota.

''We can little comprehend the vast flood of water which passed
this way from the southern and western front of the great ice sheet.
From the far northwest, including even the basin of the great Assi-
niboine River and glacial Lake Saskatchewan, 300 miles to the
northward, came the flood of glacial waters through this great chahi
of lakes and their connecting rivers, which must have somewhat
resembled straits, to the Mississippi River and Gulf of Mexico. This
was flood time in the Devils-Stump Lake region, w^hen Lakes Minne-
waukan and Wamduska stood at their highest level."* The chief
of these outlets led from the western side of the present Stump
Lake into the Sheyenne.

With the development of this outlet the waters of Lake Souris
and the extensive drainage area to the north found an easier path of
discharge thru Lake Minnewaukan, which at its prime must have
carried a \ery large volume of water from the melting ice. Another

'■'Mluuewaukan or Spirit Water was the Indian name translated bj- the wliite
man into Devils Lake. This translation is probably traceable to an Indian legend
associating the spirit of the lake with the wild winds which so often transform it
into a foaming mass of fury, and which, according to the legend, at one time over-
whelmed a party of .Sioux warriors returning from a successful foray against their
Chippewa neighbors to the north. Wamduska is the Sioux name of Stump Lake
from its fancied resemblance to a serpent, and applied by Simpson (1912) to that
lake at the time of its earlier connection with Devils Lake. Because of the fact
that botli lakes were at first one body a single name is preferable ; lience I have
cliose.n the name of the larger, iMinnewaukan, to apply to both.

♦Simpson, 1. c. p, 142.

8 THE LIFE OF DEVILS LAKE

change iu drainage channels gave Lake Souris an eastern outlet
into Lake Agassiz through the Pembina Elver, and Lake ]\Iinne-
waukan, its chief source of supply cut off, began to drop and soon
was divided into two parts— Stump Lake and Devils Lake, the
later history of both of which has been one of stead}^ recession.

The Devils Lake region is in the midst of what Simpson (I.e.)
has termed the "Drift Prairie Plain". It forms an intermediate
area between the old Lake Agassiz floor, which forms what is now
known as the Red River Valley, on the east, and the Great Plains
Plateau, which sweeps westward thru Montana to the Rocky Moun-
tains.

"The Drift Prairie Plain varies in width from about 200 miles
at the north to 100 miles at the south and has a general elevation of
from 1,500 to 1,800 feet above sea level. This plain has a gradual
but gentle slope eastward from the Coteau du Missouri to the Pem-
bina escarpment and Coteau des Prairies and southward from the
international boundary line to the South Dakota line. This double
slope determines the direction of the drainage, causing the several
main streams to take a general southeasterly course. The topography
of this plain is that of the young drift type characteristic of
all that portion of the prairie plains which lies within the limits of
the latest ice invasion, and varies from gently undulating through
rolling to hilly, the form being due almost entirely to the original
disposition of the unmodified glacial drift upon a nearly level plain.
The soft, shaly character of the underlying rocks is such that they
do not influence the surface topography to any marked extent, ex-
cept where occasional groups of low well-rounded hills or full bodied
ridges rise above the plain, and these are so well veneered with drift
that only their form reveals their origin.

More important because more numerous, though less conspicu-
ous, are the groups of hills and knobby, irregular ridges which
stretch across the prairie in a northwest to southeast direction. At
times the ridge effect is pronounced and they lengthen out into long
looped curves, and again there seems to be simply a confusion of low
rounded hills, both types being so characteristic as to suggest at
cnce to the student of physiography their origin in glacial moraines.
This prairie is otherwise a gently rolling drift plain cut by a few
abnormally deep and well defined valleys such as those of the
James, the Sheyenne, and the Maple rivers, trending southward and
eastward, marked by many shallow, irregularly winding coulees and
dotted by thousands of small lakes and marshy areas, occupying
numerous sags and swales, which testify to the undrained condition
of llie land.'"*

*Sinipsou, 1. c. pp. 107-8.

Map 5ho\A//ng Topographic reafures of /Vorth Dakota

A/tamonf Moraine

Soi/ndary of
G/aciatet!f crreo

f/ecf ff/'ver- l/a//ey

Bed of Lake ■Sour/s

Dr-ifr P/a/'n

Missouri' ^Jjereau

Plate I. From LeDnard (1919)

clia
int
wa
wa;

lat<

has
are

km

Pic

tai

at

frc

bii

bii

int

slo

ma

of

all

the

rol

dis

Th

do

cei

rid

thf

OU!

str
tin
lor
roi
on
Tl]
ab
Ja
ea!
do
nn
of

Plate 2. Above, Devils Lake from North Shore, with Sully's Hill in background
Below, East Lake, summit of "A" cliff in middle foreground.

Si-

■t^M

M

'^k-^-

^^^^m

M

W"

-.-.?;'■

w

w^

■7

• 't' •/: ■ ■/„ ,:;.:■''" ■

::'^''^''^pi

:--^i-

\

1^

'W^'^>.

^HBHftt^

.^^^^

Hi

• 4. "" ■ ^^r -**^"^^^^^I^Ss<3^

Plate 3. Upper, shore of Main Lake, showina; C, D and E benches.
Lower, A and B cliffs of Lake Minnewaukan, with the large tree midway between them.

From Simpson (1912).

THE LIFE OF DEVTT;S T.AKE 9

The Devils-Stump Lake complex occupies the southern part of
a drainage basin extending from the Turtle Mountains on the north
to the morainal mass of hills just south of Devils Lake, and with
poorly defined eastern and western boundaries. Its estimated area
is 9000 sq. km.*

"There is a gradual slope throughout the basin southward to
these two lakes; the fall is so slight, however, and the surface so
irregular that the drainage is but very imperfectly developed. Small
lakes and ponds abound, especially in the southern portion. Coulees
are few and very shallow, rarely containing running water except
in wet seasons. Formerly these coulees and the chains of lakes
connected by them emptied considerable water into Devils Lake
through Mauvaise Coulee and by several converging coulees into
both the eastern and western arms of Stump Lake. Mauvaise
Coulee was the most important drainage line of the entire basin.
Its headwaters w^ere gathered beyond the international boundary
line and in its course southward it drained the Sweetwater chain of
lakes bj' Lake Irvine through which it passed, and entered Mauvais«=
(or Minnewaukan) Bay of Devils Lake as a large and permanent
stieam. Today no surface streams flow into either Devils Lake or
Stump Lake except very minor flows during spring thaws and after
excessive falls of rain. Both lakes, however, undoubtedly receive
the extensive underground seepage from the drift cover of the large
drainage basin, the waters of which move slowly down the slope
from the north over the impervious floor of Pierre shale and through
the lower sandy portions of the drift. Little of the surface drainage
of this inland basin ever reaches either of these lakes. In fact, but
a very small fraction, almost negligible, of the rain falling in the
basin reaches the lakes by running otf over the surface. The amount
that reaches the lakes and the Sheyenue River by underground seep-
age is no doubt greater, but this amount cannot satisfactorily be
estimated."*

The history of Devils and Stump Lakes may be read in a
series of more or less well marked beaches parallelling their present
shores. There are two of these old shore lines that are well marked,
indicating the earlier levels, and several later ones which are only
occasionally distinct. The two former have been designated by
Simpson (1. c.) the A and B and the latter the C. D. and E.
beaches respectively. There is some evidence that after the ice had
left the basins of what are now Devils and Stmnp Lakes it ex-
tended an arm between them forming a dam to the waters of the
former. In this event the outlet must have been thru one of the
channels leading directly south from Devils Lake to the Sheyenne,

•Chandler (1911)

^Simijson, 1. c, pp. 10910.

10 THE LIFE OF DEVILS LAKE

two of which have been described by Simpson, one from what is
now Ft. Totten Bay and the other from Minnewaukan Bay. Indi-
cations of such a barrier are a small moraine extending north and
south across the divide between Devils and Stump Lakes and the
existence of "small cliffs and wave-cut terraces at elevations of 7
to 20 feet above the level of the A beach . . . The evidence of
earlier stages . . . are, however, fragmentary and inconclusive
. . . The life history of such must necessarily have been brief
and unimportant. ' '

The first definite shore line, or A beach, marks the lake level
when the outlet from what is now Stump Lake to the Sheyenne
River existed, and marks a level of 445 m. above the sea, and 12 m.
above the present surface. During the existence of this outlet the
water surface was evidently held approximately level for many
3-ears, as is attested by the distinct bluff which was formed by
wave action at various points, (plates 3 and 4). At other points
well marked beaches were formetl by the deposition of sand and
gravel by waves, with the occasional addition of boulders, due un-
doubtedly to ice shove. That this latter has played an important
part in developing the shores of the lake is evidenced further by the
boulder piles in the main section of Devils Lake known as Bird
Island and the Rock Pile, (plate 5) and similar islands in Stump
Lake, which show evidence of ice work, and by the frequent de-
struction or displacement of piers and boat houses in spring. At
this time the melting ice, driven before the fierce winds which
characteristically sweep these prairies, is frequently piled to heights
of 2 to 3 m., and must exert a tremendous pressure against the
shores (plate 4).

With the loss of the supply from Lakes Saskatchewan and
Souris, the outlet of Lake Minnewaukan into the Sheyenne River
was cut off and the lake dropped from the level of the A beach
to that of the B, 2 m. lower, marking the period previous to its
separation into Stump and Devils Lakes. That the level of the
lake remained approximately constant for many j^ears at this period
is shown by the well marked cliffs and beaches developed at this
time. In some places indeed the B cliff was so far developed as
to undermine and destroy the preceding A cliff (plate 4). Tem-
porary periods of rise and fall between the A and B stages are in-
dicated by minor bluffs and beaches between these.

According to Upham (1895) the elevation of Devils Lake was
441 m. about 1830, while traditions telling of the existence of certain
islands as late as 1867 indicate that this level was maintained up to
about the latter date. The later levels (C, D, & E) are poorly
marked, only rarely developing small bluffs, the most recent tem-
porary levels being indicated chiefly by zones of plants along the

m ^

« to

9 -

c't

< c « E

a 'u

'THE LIFE or DKVirvS LAKK

11

Figure I. Showing levels of Devils Lake, depths in meters as ordinates, dates as abscissas.

Prior to 1867 the latter are approximate only. The "A" stage is

indicated at 1727 on the chart.

12 THE LIFE OF DEVILS LAKE

shore which have found foothold in small depressions formed above
the small sand beaches thrown up by the waves, where a little
moisture gathers (plate 3).

Between 1867 and 1901 more or less well authenticated data are
available on the levels of Devils Lake, while since the latter date
a gauge has been maintained and numerous yearly readings are on
record. Figure 1 shows graphically the lake levels from the date
of the A stage to the present, the former being unknown, and the
rest of the earlier ones being only approximate.

In spite of our record of the steady fall of the Devils-Stump
Lake complex, there is considerable evidence of an earlier, still
lower, stage than the present. Stump Lake derives its name from
the remains of an old forest which once occupied its floor. Many
of these stumps are still standing, while many others exist as fallen
logs (plate 5). A few of these logs occur in Devils Lake, but they
are most numerous in Stump Lake. This forest must necessarily
postdate the glacial epoch, the termination of which in this region
is marked by the A beach. The period of lower level, perhaps of
absolute dryness of the lake must then have succeeded that of the
A stage. On the other hand, history and tradition testify to the
existence of a stage at least approximating that of the B stage with-
in recorded time. How high the lakes rose, and at what time, after
their lowest period; whether the beaches from the B downward were
developed before or after such periods, or partly both, we have no
certain knowledge. It seems most likely however that the prominent
B beach was developed before such a low period ensued, rather than
later, as the supply of the lake was more likely to have been held
constant for a longer period at that time, while recent changes have
been relatively rapid.

Other evidence of the existence of a former lower lake level are
the rocky islands in both Devils and Stump Lakes. The fact that
these islands until recently were below the lake surface and yet are
largely the work of ice shows that they must have been formed dur-
ing a period of lower level.

This evidence corresponds moreover with that from other
sources regarding climatic changes in the past.

Thus Huntington (1914) and others by means of measurements
of the annual rings of growth of the Sequoias, have alread.y demon-
strated alternate periods of moisture and aridity in California dur-
ing three thousand years. Similar evidence has been adduced from
the former shore lines of lakes in different parts of the world
(Huntington 1915 et. al.) and from the rise and fall of ancient
civilizations, as indicated by both historic and archaeologic evidence
(Huntington, 1. c).

•rriK LIFE OF DKVrLS LAKE 13

CLIMATE

The cause of the decline of the Devils-Stump Lake complex is
excess of evaporation over intake. The sources of such intake are
two, with a possible third, rainfall, run-off and possibly under-
ground seepage.

The weather bureau record of rainfall at the city of Devils Lake
for the twenty-four year period from 1897 to 1920 inclusive shows
that the mean annual rainfall is 44.6 cm., the greatest fall occurring
in spring and early summer, during the growing season, while late
summer and autumn, the harvest time, are apt to be dry.

Records taken at the Biological Station on the north shore of
Devils Lake about five miles distant from the town, from June 29
to August 27, 1914, show a precipitation of 10.1 cm. w^ith occasional
additional traces, making a total of about 10.2 cm. compared with a
total of 10.5 cm. for the weather bureau station in Devils Lake.
During summer the rainfall is apt to be local and inconstant, so
that while one area maj^ receive a good supply, another, a short
distance removed, may receive but little.

The snowfall is usually light, overaging 82.8 cm. for twenty
years. Under the high wind the lake surface is usually swept bare,
and the snow piled in heavy drifts about its shores.

The amount of run-off reaching the lake cannot be estimated.
^\''ith the loss of the supply from the ]\[auvaise Coulee about 1885*,
probably due to the extensive settlement and consequent cultivation
of the territory contiguous thereto in the early eighties, coupled
with the drouth in the latter half of this decade, the drainage basin
was reduced to the area immediately bordering the lake, in which
there are no streams and only a few feeble springs. Most of this
land moreover is either cultivated or wooded, so that there is very
little run-off.

The extensive cultivation of the soil is a well known cause of
lowering the w^ater table in the ground thru increased evaporation
from the surface and especially thru the transpiration of plants,
but it has an even more important effect in preventing run-off.

The major part of the run-off occurs during the spring thaw,
when the ground is frozen to a depth of several feet. The amount
of this run-off depends not alone on amount of snowfall, but even
more on the character of the thaw. When the latter is sudden,
practically all of the melting snow reaches the lake, causing a mark-
ed rise ; when gradual, most of the water percolates into the ground
and the lake level is but little changed. The average amount of
this rise for ten vears has been about 0.15 m.

*As a result of e.xccptionally hoavy snowfaU in the winter of 1915-16 this
coule'e was flowing' approximately 1.3249 cu. m. per day on Aug. 1.5, 1916.

14 THE LIFE OF DEVILS LAKE

The amount of seepage water reaching the lake is likewise an
unknown factor. The porous mantle of glacial till overlying the
impervious Pierre clay undoubtedly furnishes a certain supply in
this way J especially in spring when the melting snow surcharges the
ground with moisture, as may be seen in the little pockets in the
fields which are too wet for plowing. The slight slope of most of
the drainage basin, however, coupled with the fact that during
the summer the rainfall is seldom more than enough for the grow-
ing crops, must keep this supply at a minimum. Conversely the
loss thru seepage must be a negligible factor owing to the impervious
clay underlying the lake.

The daily amount of evaporation at the lake shore determined
by readings on a Livingstone porous cup atmometer from Aug. 4
to Sept. 5, 1914, shows, as would be expected, a great range of vari-
ation. These readings, however, have a comparative value only, and
are of no significance in themselves.

Factors determining evaporation are three — temperature, wind
and humidity. Like other interior regions at this latitude and alti-
tude the climate of the Devils Lake region is marked by great ex-
tremes of temperature, maxima and minima for 24 years from 1897
to 1920 at Devils Lake being 41.1°C on July 28, "^1917 and 42.2°
on January 12, 1916.

Daily extremes are considerable. The average daily range of
temperature at Devils Lake city compiled monthly for 19 years is
11.98°C. The greatest average variation occurs in August, but the
maximum was in January (30.84°).

Maximum and minimum temperatures at the lake shore were
taken from June 13 to Sept. 5, 1914, with a few omissions. These
showed a maximum daily range of 23.3° as compared with one of
22.8° at Devils Lake city. The former, however, occurred in June,
and the latter in August, so there is evidently no very close cor-
respondence in the temperature range of the two localities.

The annual and monthly mean maxima and minima from 1900
to 1920 are shown in table 1. The annual average number of
sunny days* is 125 while at the summer solstice the number of
hoiirs of sunshine is sixteen.

Twilight lasts till about 11 :00 p. m. and the first flush of day
appears about 3 :00 a. m. This great amount of sunshine ensures
rapid growth and early maturity to the crops of grain for which
the northwest is famous, and doubtless determines in large measure
the development of organisms in Devils Lake in early summer.
Midday in summer is rarely oppressive and the nights are almost in-
variably delightful, days of excessive heat being rare, and usually
enduring for but brief intervals. Devils Lake usually freezes be-

•Ovar 70% sunshine.

THE LIFE OF DEVILS LAKE 15

tween the 15th and 25th of November, while it is usually free from
ice between April 20 and 25. The ice commonly reaches a thickness
of one metre, a few centimetres of which are frozen snow.

The comparatively low relative humidity (table 2) and the
characteristically high winds of the region, are mainly responsible
for the rapid evaporation, rather than the occasional brief extremes
of temperature.

TABLE) 1
Showing mean maximum (A) and minimum (B) monthly and annual tempera-
tures (C°) at Devils Lake City from 1900 to 1920.

Jan. Feb. Mar. Apr. May June July

Aug. Sept. Oct. Nov. Dec. An-
nual

A
B

— 10.2 —9.6 — O.S 10.9 18.1 22.8 26.3
— 22.1 —21.3 — 12.7 — 1.9 4.1 12.4 12.7

25.2 20.3 12.5 2.8 — 7.2 9.3
11.1 6.3 — 0.2 — 8.5 —16.7 — 3.3

a

TABLE 2
Showing monthly and annual mean relative
period of 13 years.

humidities at Devils Lake city for

Jan. Feb. Mar. Apr. May June

July Aug. Sept. Oct. Nov. Dec. An-
nual

7

a. m 88 88 85 80 77 79

83 85 85 84 86 88 84

7_

p. m 84 83 75 57 51 57

57 55 59 65 78 84 67

Wind is one of the most important factors, not only in lowering
the level of Devils Lake ; but in moulding its shores, thru ice and
wave action, and in determining its currents, the distribution of its
inhabitants, its dissolved gases, and other chemical constituents, its
temperatures and turbidity.

The average total annual wind movement at Devils Lake for
13 years was 168531 km., maximum velocities of 97 km. per hour
being occasionally reached. The prevailing directions are N W. and
S. E.

The average annual precipitation at Devils Lake is 44.6 cm.,
while during a period of 14 years (1907-1920) the level of the main
part of Devils Lake fell from 435.1 to 433.1 m, or 2.0 m, an average
annual fall of 0.15 m.

The amount of water represented by this drop can only be
estimated since no accurate survey of the lake has been made since
1883, and the shore lines are constantly changing. According to
the report of the state engineer (Atkinson, 1912) on the survey
for a proposed diversion of the Mouse River into Devils Lake, the
amount of water needed to make good the annual excess of evapora-
tion is 24,635, 790 eu. m. Basing an estimate on the above data
and allowing for the total rainfall reaching the lake, plus 50%
additional in run-off and seepage, the average amount of evapora-
tion for the past 15 years may be estimated at 0.8 m. or 131,390,880
cu. m. This figure is of course the merest estimate, based on the
average rise from melted snow in spring, when most of the run-off
occurs and amounting to about 35% of the annual rainfall, with an
additional 15% for run-off during the rest of the year and for

16

THE LIFE OF DEVILS LAKE

<s^Ai n

THE LIFE OF DEVILS F.AKE 17

seepay:e. During the sumraor of 1916 in spite of an estimated
daily flow of 13,249 cu. ni. ■ from the Mauvaise Coulee, the lake
fell about 0.24 m. from May 1 to November 5.

The future of Devils Lake can only be read in the record of
the past. With present climatic conditions continuing it will prob-
ably disappear during the next forty or fifty years. The annual
drop is apt to increase with decrease in depth. "From him that
hath not shall l)e taken away even that which he hath" is ap-
plicable to lakes as well as to men. The eastern section of Devils
Lake, which was separated from the main body in 1903 by a high-
way at the Narrows, and which in 1913 had an average depth of
1.8 to 2.4 m. was more than 0.9 m. lower than the latter in 1922,
and is now (1923) little more than a mud flat, covered in places
with a few centimetres of water, while Minnewaukan Bay, which
in 1916 contained 0.9 to 1.2 m. of Avater (plate 10) virtually dis-
appeared in 1917, while the main body fell only 0.4 m.

Whether climatic cycles may ensue to raise the lake to its for-
mer high estate must be left to the future to decide. Such changes,
however, are reckoned in centuries, not years. The immediate
prospect is one of continued and increasing fall and ultimate dry-
ness, the fate of so many other glacial lakes in North America. The
youth and maturity of Devils Lake are past, the period of old age
has arrived.

MAIN LAKE, PHYSIOGEAPHY

As may be seen from the map (fig. 2) Devils Lake has the
form of a long, narrow body with several arms at right angles to
the latter. These arms or bays probably mark the course of tribu-
taries of the old pre-glacial river in whose valley the lake now lies
The area of Devils Lake in 1883 was approximately 401.5 square
km. with a shore line over 290 km. long. Today the area probably
does not exceed 180-200 km. and the shore line has also been consid-
erably reduced, tho not to the same extent as the former.

The recession of Devils Lake, together with the construction of
highway embankments at various points across the lake, has resulted
in the development of several smaller lakes, besides the separation of
the glacial lake Minnewaukan into the present Devils and Stump
Lakes. Some of these smaller parts, which are spring fed, have re-
tained their freshness; while others have a high concentration of
salts, the proportions of which vary of course in each individual
ease. At present Devils Lake proper is divided into four chief
parts — Minnewaukan Bay*, Main, East and Lamoreau Lakes ; the
principal of which, and that on which most of the data in this
treatise are based, is Main Lake. This is an irregular sheet of

•Aug. 15.

♦Virtually dry — 1924.

18 THE LIFE OF DEVILS LAKE

water approximately 130 sq. km. in extent and at present** of about
5 m. maximum depth. It has two main arms extending north —
Creel Bay, about 5 km., and Six Mile Bay about 6 km. long, each
averaging from 0.8 to 1.2 km. in width. At the eastern end Main
Lake is cut off from East Lake by a highway across the Narrows,
while to the southeast Mission Lake was cut off by a dam in 1908,
The floor of Main Lake is mostly covered with a very fine, impalpa-
ble, black, foul smelling ooze, but in places sandy or gravelly bottom
is found, with occasional glacial boulders scattered over the lake
floor. The greater part of the lake bottom is nearly level, with
a gradual rise about the shores, which in places are covered with
glacial boulders, and in others are low, muddy flats.

PHYSICS

The physics and chemistry of the water naturally varies with
the changing level of the lake, and with the seasons of the year.

The specific gravity in 1907, as given by Pope (1908) was
1.006, while in 1912 this* had increased to 1.0099.

The color of the water, determined by the platinum-cobalt
standard of the IT. S. Geological Survey, usually ranges from 15 to
70, the lower figure representing conditions in winter, when chem-
ical and physical changes are at a minimum, while the higher rep-
resents the condition in spring, accompanying the run-off from the
melting snow, with its large amount of dissolved material from the
humus in the soil. There is no appreciable difference in color be-
tween the surface and the bottom of the lake.

On March 11, 1923, the water from a hole cut in the ice near
the shore was a deep wine color, which faded very markedly after
standing a few days. Not having any color standard solutions
available at the time a color reading was impossible. The color
moreover was not comparable with the platinum-cobalt standard,
being distinctly pinkish, so much so, in fact that an alkalinity read-
ing with phenolthalein was impossible. The ice in the neighbor-
hood of the hole was thickly filled with Ruppia, and the color was
apparently due to an extract of the latter. That this should have
occurred to such an extent at a temperature of about 0° C. seems
remarkable.

Comparative analysis of the water from the shore hole and from
one some distance out in the lake, made on March 22, when the
color had nearly disappeared, showed, as was to be expected, a
much higher amount of both free and albuminoid ammonia in the
shore sample.*

**1924

•Mid lake — free NHa 0.52, albuminoid Nils 2.71 ppin.

Shore — free NH^ 3.56, albuminoid Nil,, 7.24 ppm.

THE LIFE OF DKVILlS LAKE 19

Turbidity has been determined by means of the Jackson turbid-
imeter described by Whipple and Jackson (1900). This ranges
from 70 to below 25 p p m.** of SiOg and averages sligiitly higher
at the surface than the bottom, the reason for which is not clear,
but is probably a slight excess of phytoplankton at the former level.
The difference, however, is slight and inconstant. The turbidity is
largely due to bacteria and algae (chiefly blue-greens), but consid-
erable quantities of amorphous matter were present in all the
plancton collections except in winter when only traces of it occurred.
This amorphous matter is probably in part the result of deteriora-
tion of the plancton in samples which have stood for several years,
and in part composed of air borne dust and more largely of the
flocculent ooze from the lake bottom. While I have no positive
evidence on this point, there is little doubt that the lake is frequent-
ly stirred to the bottom b}^ high winds and that the ooze thus dis-
turbed takes a long time to settle.

To the action of the wind is also probably due the rather
uniform distribution of the plancton between surface and bottom,
at least at the depth of 3-4 m., where most of my collections have
been made, and the occasional massing of the algae along the shore
at certain planes, especially in the case of Nodularia, the principal
filamentous form in the lake.

Horizontal currents are due to the wind, and hence are variable
in strength and direction. No exact measurement of these currents
are available but approximate determinations have been made by
timing the drift of submerged bottles over measured courses.

Exact measurement of these currents would have little sig-
nificance unless made over a period of years to include the maxi-
mum, minimum and average effects of wind. To show in detail
wind influence it would also be necessary to make current determi-
nations both as to velocity and direction for all parts of the lake,
illustrated with chai'ts. Time and equipment have rendered such
detailed study impracticable, and it is further doubtful if the re-
sults of such a stud}^ would justify the necessary time and effort.
The approximate results obtained are sufficient to illustrate the im-
portant role which wind may play in the horizontal distribution of
the organisms (especially filamentous plants) in the lake. The
determinations were made on August 27, 1917, during a high wind,
and showed a current of about 13 cm. per second, or approximately
1.4 km. in three hours, at which rate a mass of plancton might read-
ily drift from one side of the lake to the other in a day.

The determination of the penetration of light in water is a
matter of some difficulty. The problem has been attacked by many
investigators, employing several types of apparatus. The simplest

•*In winter less than 10 ppra.. compai-ed with a silica standard.

20 THE LIFE OF DEVILS LAKE

of these consists of a white disk, Avhich is lowered in the water and
the depths of disappearance and of reappearance upon raising not-
ed, the mean of which gives a reading which maj'' be compared with
similar readings on other waters, or on the same water at other
times, or under different conditions. A modification of this type
of apparatus is the turbidity rod of the U. S. Geological Survey, in
which a platinum wire is substituted for the disk of the Secchi ap-
paratus. In another modification an electric light replaces the disk.
Another type depends on the exposure of a sensitive plate or film at
various depths for various intervals, the relative degrees of penetra-
tion varying inversely as the times of exposure. This method has
been employed by several investigators in both marine and fresh
water work, using many different kinds of apparatus for this
purpose. A third plan is that of Regnard who used the resistance of
a selenium cell as registered by a galvanometer to determine the
strength of light at different depths in the harbor of Monaco.*
Birge and Juday, in their work on Wisconsin lakes and elsewhere,
have employed an instrument, the pj^rlimnometer, which measures
the radiation effect of the sun at various depths, by means of
electrical thermo-couples and a galvanometer. Finally Shelford and
Gail (1922) have employed a photo-electric cell consisting of potas-
sium mounted in an atmosphere of hydrogen contained in a glass
cell, which is connected in circuit with a battery of 80 to 160 volts,
and a galvanometer, by means of insulated wire. The amount of
current passed thru the cell is proportional to the intensity of the
light striking it, and the cell is most sensitive to the blue rays.

A detailed discussion of the relative merits of these various
devices need not be given here. Suffice it to say briefly that all
give values which are relative only.

The first, since it involves a personal factor, is valuable only
when used by the same observor at all times. It furthermore tells
nothing as to amount or quality of light penetrating to various
depths.

The second method involves the elimination of certain rays by
absorption in passing thru the glass window covering the sensitive
film. The amount and quality of the rays absorbed by the glass
can be tested spectroscopicall}^, however, and in any event is proba-
bly small.

A more serious ob.jectiou to this method is the difference iu
relative density with different times of exposure, and development.*

Another serious objection is that, in comparing the density of
plates or films of different exposures with any standard, or Avith
each other, the personal error enters and it is difficult, or impossible.

♦Fide' SliPlford nud Gail (1922). I Iiave not had access to Regnard's original
work.

*Sec Nutting (1912)

THE LIFE OF DEVJUS LAKE 21

to obtain closely accurate results. An essential condition of this
method is, of course, uniformity in the plate or film employed. By
using the same kind in all experiments, practical uniformity of re-
sults is probably assured.

The objections to the third method have been detailed by Shel-
ford and Gail (1. c) and may be summarized by the statement that
the sensitivity varies in different cells and under different condi-
tions (temperature, time of illumination and quality of light).

The pyrlimuometer of Birge and Juday measures the radiatior
(heat) energy of light, rather than its chemical effect, which latter
is the more important factor to be determined, — since the heat effect
can be found by thormometric methods.

The objections to the use of a photo-electric cell would appear
to be its expense, difficulty of construction and operation, especially
on rough waters and the absorption of some rays by the glass cell.

In all of these methods probably the greatest difficulty in-
volved is in the variable condition of the water surface and of the
air at different times. "When the surface is rippled there is ob-
viously much greater reflection than when it is smooth. The differ-
ence in penetration of the light thru the surface layer may vary
momentarily more than 100%. Thus Shelf ord and Gail (1. c. pp.
160-1) have shown that on Puget Sound the light penetrating the
surface varied in three minutes from 67.3% of the total (in the
air) to 26.0% with changes in the surface, due to tide ripples. Even
marked changes in intensity of light, as recorded by photo-electric
cells, may occur within brief intervals in bright sunlight. "Marked
difference in apparent brightness of the sunlight and in the reading
may take place in five or ten minutes when no clouds are visible."
(Shelf ord and Gail, 1. c. p. 155).

The results obtained by these various investigators naturally
vary widely according to the various methods and conditions of the
experiments. No two waters are exactly alike in character, nor does
any one body remain constant at all times. In general, the penetra-
tion is greater in winter than in summer, partly because of the
direct influence of temperature on the absorptive power of water,
partly because of the amount of plancton and amorphous matter
present at these seasons ; tho individual waters will vary greatly
at different times and under different conditions.

Helland-Hansen*, using photographic plates, found in the At-
lantic Ocean a penetration of from 900 to 1500 m., while Grein
(1913)**, employing a similar method, found a penetration of about
1500 m. Andersen and Walker (1920), in their work on the Sand-
hill lakes of Nebraska, found an elimination of light at 4 dm. vary-

*See Murray & Hjort (1912, p. 257).
**Fide' Shelford & Gail (1. c).

22

THE LIFE OF DEVILS LAKE

THE LIFE OF DEVILS LAKE 23

ing from 89 to 95%, while at 8 diu. it ranged from 96 to 99%
depending on the time of exposure.

In the (Trannsee) Linsbauer (1905)"*** found a reduction in
light between the air and a 10 m. depth of 98.6% ; while Whipple
(1914) found that the Secchi disk disappeared at a depth of 1.8
m. in the highly colored (92 on the Pt.-Co. U. S. G. S. scale) water
of the Chestnut Hill Reservoir near Boston, as compared with a
depth of 30 m. in the comparatively clear waters of Lake Tahoe in
California.

In my own work on Devils Lake I have employed the apparatus,
which is illustrated in (fig. 3). It consists of two brass plates a
and b, separated by a metal strip c, around three sides, so spaced
as to admit a photographic plate and cover between them. The
opening on the fourth side is closed by a plate carrying a tongue,
which closely fits the opening in the box so as to render it light
tight. Two springs bring the plate into position against a. In a
are four holes placed 90 degrees apart. Upon a is a circular disk
(1, with a tongue on its lower surface fitting into a groove in a. d
also lias four holes so placed as to exactly match those in a, when
d is in the proper position, d is revolved by means of a spring coil,
s, attached to a pin, p, on d and a plug, e, which passes thru d and
is screwed into a. Around the edge of d are eight equally space 1
teeth, t,t'. Attached to a point on a is a bar, f, with a tooth at g.
When s is wound up, f is pressed against d by a spring, h, in such
a manner that g engages one of t and d is held in position covering
the holes in a. By means of a vertical arm, i, which swings on a
pivot^ j, and carrying a cog which engages one end of f, g may
be released and d revolved thru 45 degrees by the action of the
spring coil, s, thereby exposing one of the holes in a. i is con-
trolled by a cord in the hand of the operator. Releasing i allows g
to engage the next tooth, t. d is automatically prevented from re-
volving too far, and thereby closing the hole in a too soon, by the
end of the bar f opposite g which engages the tooth t' when t is
released. The closure of the hole is affected in the same manner
as its exposure by simply pulling the string attached to i. The ap-
paratus is supported by four cords attached at one end to each
corner and at the other end to a small plate, which is low-
ered to any desired depth by means of a marked line. The film
is protected from water by a glass cover sealed to the plate by
celloidin. The plates employed were Cramer's "isochromatic."

For making the experiments clear, still days were chosen so far
as possible, a combination which is very unusual at Devils Lake,
occurring perhaps half a dozen days all told in any year. Some of
the experiments, however, were made thru the ice, when the wind
factor was, of course, eliminated.

***Fide' Steuer (1910).

24 THE LIFE OF DEVILS LAKE

In each experiment the time and weather conditions were noted
and one exposure made in aii- for comparison with those under
water.

The results obtained with this apparatus, some of which are
illustrated in plate 6, show marked differences on different dates
dependent on the altitude of the sun and the turbidity of the water.
Many other factors enter in, such as condition of the air and of the
surface, but the first two are most important.

Regarding the influence of the condition of the surface on the
readings, both Shelf ord and Gail (1. c.) and Anderson and Walker
(I.e.) agree that a large proportion of the light is cut off by a
rough surface and that this proportion may change very greatly
from time to time. I have not made a careful study of this point,
as most of my readings have been made in comparatively quiet
water. Two series of exposures to test it were, however, made on
October 10, 1923, between 2 :13 and 2 :24 p. m., one where the water
was almost calm, the other where the surface was rippled. They
show only a slight difference in surface penetration in nearly calm
and rippled water.

The results show^ a much greater diminution of light in Devils
Lake than in Puget Sound (fide Shelf ord and Gail), about the
same as that in Seneca, Cayuga, and Canandaigua Lakes, N. Y.
(fide Birge and Juday 1921), but considerably less than that found
by Anderson and Walker in some of the Sandhill lakes in Nebraska.
Neither the first nor the last of these, however, have recorded the
turbidities at the time of observation, which makes impossible any
direct comparison of our results. Birge and Juday have recorded
the transparency in terms of the Secchi disk, their results showing
very clearly the influence of turbidity on transmission of solar ener-
gy-

My own observation show fnrtlier a marked difference in per-
cent of transmission between July, August and November. This
agrees with the results of Anderson and Walker and is to be ex-
pected from the difference in altitude of the sun on those dates.

A brief examination has also been made of the penetration
of red and blue, as compared with white light; using for this pur-
pose filters of red and blue glass, whose transmission has not been
tested speetroscopically, together with Cramer's "spectrum process"
and "iso" plates. The results indicate a lower penetration of blue,
and red, than of white light.

Temperature records have been taken with all of the plancton
collections. In some eases the Negretti-Zambra thermometer ha.s
been used, but generally they have been taken on a chemical ther-
mometer. These have not been standardized and varied somewhat.

Plate 6. A. Comparative light intens-
ities in air and at 0.6. 2.1 and 2.7 m.
depths on October 6, 1923.

1 — exposed 1 sec. in ;iir.

2 — exposed 2 1-.5 soc. ;it O.ll ni. dejitli.

3 — exi)osed 3 4-.5 sec. :\t 2.1 m. depth.

4 — exposed o sec. at 2.7 m. depth.

Comparative light intensities in air
and just below the surface of the
lake on October 7, 1923.

1 — exposed 1 1-5 sec. in air.

2 — exposed 1 1-5 sec. at surface.

3 — exposed 2 1-10 sec. at surface.

4 — exposed 2 9-10 sec. at surface.

Comparative light intensity in air
and at 91/2 m- and I2V2 m-
on November 4, 1923.

1 — exposed 4-5 sec. in air.

■_' — exi)osfd 29* sec. at 2.9 m. depth.

:'. — cxiinsed 60* sec. at 2.9 m. depth.

1 — exposed 120 sec. at 3.8 m. depth.

*Possible error of 1 sec.

THE LIFE OF DEVILS LAKE

25

the maximum difference in no case exceeding 1^' C, and usually
being less than that. Some additional readings have been made,
especially along shore in summer time to determine temperature ex-
tremes in various parts of the lake.

Beside these readings a series of records were made with a
thermograph during the summers of 1914-15, to obtain a comparative
record cf temperatures in air, at the surface and near the bottom
for all hours of day and night, for a period of several weeks, one
of which is shown in figure 4. These show, as might be expected,
the greater changes occurring at the surface, as compared with
lower levels of the lake.

ze.hf

21.11

Figure 4. Thermograph records on Creel Bay for air and surface on August 13-14 and

25-26, 1914.

In a shallow lake such as Devils Lake no thermocline develops
and there is usually but little temperature difference between top
and bottom. In hot summer days the surface temperature rises a
few degrees above the bottom, but there is no distinct break between
the two. Similarh' the shallow water near shore, especially on
quiet sunny days, reaches a higher temperature than the surface
water in mid lake. Vice versa the surface and shore water are
cooled at night more than the bottom, so that more or less mixing
of the water occurs daily.

As the lake warms in the spring the shore water naturally
warms faster than the main body, a condition similar to that de-
veloping on warm summer days, while reverse conditions occur as
the lake cools in autumn. Figure 5 well illustrates these differences.

26

THE LIFE OF DEVILS LAKE

I
I

20-

Au&.i<i,m^

15-

AuG. 25,111^

Figure 5. Graphs illustrating surface temperatures across Creel Bay on August 19 and

25, 1914.

JunriS.IIII

"\"IXV..

Figure

6. Temperature and oxygen curves for Main Lake. Ordinates represent depths
metres, abscissas, temperatures in degrees centigrade and cc. of 0:> pr. I.

In fall, as the lake freezes, the water at the surface, next to
the ice, is cooled slightly faster than that at the bottom, and its ex-
pansion near the freezing point causes the cooler water to remain at
the surface for a time. Gradually, however, conduction equalizes
the temperature of the two regions, which thereafter remains con-
stant until spring, when, the warming of the surface layer, and its
consequent condensation, again produces thoro mixing of the entire
body of water, the surface and bottom temperatures thereafter re
maining nearly constant until summer conditions ensue. Temporary
and slight variations occur, dependent on the changing conditions
produced by melting ice and snow, alternate freezing and thawing
during the early spring and the varying temperature of the air.

Some tyjiical temperature and oxygen curves for different sea-
Bons are shown in figures 6, 7 and plate 7.

THE LIFE OF DEVILS LAKE

27

0.

d'cS'o

II M II

'■"^^■i a g s 's °al/ ?, a 2 tf, o to
r~T I I r^ r I I I I r

"'■'(TCO C-\i> tft ■*• lO <VI T- O -^

's/ii

■s^A

<« o

S "

j; E

28

THE LIFE OF DEVILS LAKE

CHEMISTRY

The chemistry of Devils Lake Avater naturally varies with the
variation in the lake level. A large number of analyses have been
made in different years by different chemists. Many of them arc-
partial, but there is a good series, which is complete for all the
most important constituents of the water. The methods employed
have been, in general, those recommended bj- the American Public
Health Association. The results show an increase in salt concentra-
tion, accompanying the decrease in lake level, from 8471 ppm. in
1899 to 15210 in 1923. This chemical change is, as will be seen
later, probably the most important factor in determining the
changes in the inhabitants of the lake. The analysis (table 3)
shows the character of Devils Lake water to be typical of other
brackish water lakes of the interior.

Table 3, showing
except as noted.

chemical analy

ses of lakes iu

Devils

Lake

lomplex

all in

1919,

Locality

Jfain

a

a
o

i

3

a>
u

o

a

C4

P.

a

<:

O

c

CQ

&4

Normal carbonates as
CaCOa

305

1 571

1 320

290

350

290

374

84

Bicarboiiates as

as Ca(HC0.i)3 ....

1 1
458 ' 902 1 525

489

525

564

448

56

Total solids

1 11755 1 '218

1(1907)1(1907)

13462 1 37S01 32640

I 62929 48179

(1920)1(19201

14932

19000

25450
(1923)

19003

9143

15755

16165

Calcium

70 1 203 1 82

107

239

25

1 91

317

844 1 931 1 1532
(1913)1 1

727

870

883

445

691

735

Sodium

2548 1 8232 | 4012

4012

33 02

3946

2146

4201

3660

Potassium

204 1 1

133

764

273

AljOi and Fe>0- .

95 1 88 i 134

63

42

24

45

32

Chlorine

1310 1 3250 1 2750

1 1442

1803

1950

;f95

1 1610

1283

Sulphates (SO,)

7187 1 19960 1 16129

1 8070

10084

10488

4589

1 8320

8988

Iron

1 1

1

4.0

SiO:j

1 168 1

29.4

Osmotic pressure in
atmospheres

1
6.3 12.8 9.7
(1920) (1920)1(1920)
1

6.75
(1920)

1
8.9
(1920)1

I

The salt content of surface and bottom layers is alike, but some
differences e.\ist between different parts of the lake, dependent on
the scanty supply of fresh water drainage at one or two points, in-
flux of sewage, and probably to some extent also on other less
obvious factors, such as temperature, depth, kind and amount of
organic matter in the ooze and kind and number of organisms pres
ent in the water.

THE LIFE or DEVILS LAKE

29

The orjijaiiic material, represented b}' free and combined am-
monia, naturalh^ varies markedly, not only with season, l)ut also
with locality and depth (fig'. 8).

PrtSTS
ptr

Plilliofl-

I

A

SwnrAcf

20-

\\

■• ^'i' (VH,

QOTTO^

i

\\
\\

^V„, MM

15-

1
1

\\

1

\\

1-

\\

10-

/\

\\

/ 1,

\\

/ f\

\ \

5-

/ ji|\

•S \

'-

/ A

\\ '\

\

■

J/j

\

v^--..

. V-.. .,j^

'am Feb. riwtAm. Hw Junt July flut

Sre'Otr'Nov'Dec'

-rto,

• no.

P»<tr5
5 -,

j/w'r£e.'n,.B'ii«'niT'ji-«'Xj«'fl>,o'Sfp'£XT.'iv«'D«'

Figure 8. Seasonal occurrence of organic material in Main Lake, Jan. -Mar.", 1914,

Apr.-Dec, 1913.

Dissolved gases also show marked variations with time and place,
which, to a certain exient, when plotted, take the form of cj'clical
curves dependent on the seasons.

Especially is this true of dissolved oxygen. The amount of this
present in the lake depends on temperature, which in part deter-
mines the absorptive capacity of the water for gases and the
photosynthetic activity of plants; on amount and velocity of the
wind, determining amount of aeration ; on the number and activity
of the organisms present, and on the extent of decomposition taking
place in the water.

The apparatus used for collecting oxygen samples is that recom-
mended in "Standard Methods" of the American Public Health
Association.

Exact determination of oxygen in the lower layers is a matter
of considerable difficulty, due to currents generated by the raising
and lowering of the apparatus. When an oxygen free stratum

30 THE LIFE OF DEVILS LAKE

several inches thick occurs at the bottom, it is usually possible to
definitely establish this fact, but if this stratum is very thin, or if
oxygen is present at the bottom, but in quantities much smaller
than at higher levels, it is practically impossible to obtain duplicate
readings checking each other within less than 25 or 30%, Espe-
cially is this true in wind.y weather, when the motion of the boat
increases the currents generated in raising and lowering the appa-
ratus.

There is some question regarding the accuracy of the Winkler
method, which was emploj'ed in this study, for the determination
of dissolved oxygen. Birge and Juday (1911) found a fairly close
agreement in the results of the Winkler and boiling out methods.
They were working, however, on waters, none of which approached
Devils Lake in salt concentration. In waters of such high concen-
tration as the latter, the iodiraetric method employed by Winkler
is liable to be affected by the various salts present. Relative
thereto, Dr. Heath, chemist at the station in 1914-15, says in ms.
notes on his results :

"For a variety of reasons this method appears to be inaccurate
when applied to such a salt solution as Devils Lake water . . .
The presence of nitrates in the water causes the Winkler method
to give too high results for oxygen, and nitrates may interfere by
a variety of chemical changes. When the amount of nitrate is
known a correction may be applied, but even then there are so
many complicating factors that the method is unsatisfactory. It is
probable that the other substances present in the water also inter-
fere with this determination to a certain extent."

The results obtained by the Winkler method, however, agree
well with those obtained by other workers elsewhere and with
what might be expected in Devils Lake. The boiling method, more-
over, is tedious, and difficult to use in the field, for which reasons
the Winkler method has been used and probably gives results which
are sufficiently accurate for comparative purposes.

During the winter, when the ice sheet reaches an average
thickness of one metre, there is little opportunity for aeration, and
by spring the oxygen in the lower levels is greatly diminished if
not entirely absent. With the melting of the ice and influx of a
considerable body of fresh water from air and melting snow, the
oxygen content rises sharply, gradually diminishing again during
the warm days of summer. After a few hot still days, which but
rarely occur, the bottom water may be almost or entirely free from
oxygen, but a fresh breeze quickly changes this condition and restores
oxygen to the lower layers. With the advent of cool fall weather
the oxygen content again rises, remaining at a maximum until the
lake is once more sealed by ice.

THE LIFE OF DEVILS LAKE

31

The lowering of the oxj-gen in the lower levels is due, not to
the absence or scarcity of chlorophyl-bearing organisms at these
depths, where they are about as numerous as they are near the
surface, but rather to the decomposition of the organic matter con-
tained in the ooze covering the lake floor. This decomposition
evidently occurs even at zero temperature, judging by the diminution
of oxygen in the bottom level during the winter.

Similar, though much less marked differences, occur between the
shore water and that further out in the lake, while local differences
may occur, due to aggregations of plants at different points.

No free carbon dioxide appears in any of the analyses and is
probably never present at any point in the lake, though it is pos-
sible that at times of high water it might temporarily appear at
the mouth of the sewer from Devils Lake city, which discharges
into the lake. In the form of bound (mono-) and half bound (bi-)
carbonates it usually shows but little difference between surface and
bottom, and the shore and middle parts of the lake. At the bottom
and near shore the bicarbonates are higher than elsewhere. When
the ice is melting the surface water is temporarily much lower in
carbonate alkalinity than the lower water. With the disappearance
of the ice and stirring of the lake by the wind, this difference
promptly disappears. (Figure 9.)

too-

^

~.^

\

^~~— -• —

\

^ - ^*^^

J

4O0 J

\

"^

300.

_-— —

CO, ^^"-^^^^^'--^ — ""

\ ~ /-/

^r^^-

^

tx-

\

\\ / /

ioo-

viry

■J- :i^

^ jf- S-

-c -^ Jl

;>

'I

Figure 9.

Distribution of CO.- in Main Lake at 4.0 m. from 9,24, 13 to 6, 20/14.
OrdinatES represent parts pr. million.

Numerous tests have been made for hydrogen sulphide, but
the results are not very conclusive.

Birge and Juday (I.e.) have suggested the inexactness of the
standard methods for the determination of HoS and recently Heath
and Lee (1923), as the result of studies based primarily on Devils
Lake water, have pointed out the difficulties in its determination,
due to the presence of various salts and their reaction with iodine.
These authors suggest the substitution of a colorimetric method for
the iodimetric one.

32 THE LIFE OF DEVTT.S LAKE

Dr. Heath in ms. notes thus summarizes the results of his work
on Devils Lake :

"The waters of the Lake seem to carry no dissolved hydrog:en
sulphide except in a few places where contaminated by sewage. The
mud on the bottom and the water in immediate contact with it
carry the gas in small amounts."

Comparatively little is known regarding the relation between
dissolved nitrogen and aquatic life. It probably is always present
in water, forming a large percentage of the dissolved gas. Judging
by its inert character, it seems improbable that it plays any import-
ant part in the life of aquatic animals, although there is some
evidence to show that in excess it may be injurious to fishes (Marsh
and Gorham, 1905). It may possibly be an important part of the
food of aquatic plants, though this is not definitely known.

But two tests have been made for nitrogen in Devils Lake, both
in the summer of 1914. Two 1200 cc samples gave by the boiling
method 84.9% and 89.8% respectively of nitrogen in the dissolved
gases, the remainder being oxygen. In these two tests no methane,
carbon monoxide or free hydrogen were detected. These are the
only tests which have been made for these gases.

BIOLOaY

The biological investigation of any area comprises an enumer-
ation of its inhabitants, and a determination of their distribution
and abundance in time and space, in reference to their physical
environment and their reactions upon one another. Without the
cooperation of many specialists the first phase of the problem is
impossible, while the second phase is of course dependent on the
first. Due acknowledgment of the aid received in the prosecution of
this work has already been made and need not be repeated here.
The list of species is admittedly incomplete since some of the iden-
tifications have not been finished, and part of the work was pre-
liminary in character. This is notably true of the nematodes,
insects, and bacteria.

In spite of these deficiencies, however, it has not seemed advis-
able to delay publication further, since to obtain complete lists of
species would probably require several years delay, and by that
time the fauna and flora of the lake would probably differ markedly
from their composition when the work Avas begun. In fact at the
present time distinct changes have taken place, and in the case of
some parts of the lake (East and Mission Lakes) the present fauna
and flora have little resemblance to those of a few years previous.
Furthermore, the data at present available are sufficient to give a
fairly adetiuate idea of the life of the lake and its responses to its
changing environment.

THE LIFE OF DEVIDS LAKE 33

The biological problems may be grouped under three main
heads ; first, the life zones of the lake and their characteristic inhabi-
tants; second, the seasonal distribution of the plancton, and third,
an annotated list of species.

LIFE ZONES

The life zones of the lake are not well defined regions because
of its shallowness, the gradual slope of the bottom, and especially
because of the influence of the wind in mixing its waters. They may
in general, however, be defined as four, namely (1) littoral, the
shore zone to a depth of 0.6 ra., (2) Ruppia, the region between
0.6 and 2.0 ra. in depth, where Ruppia grows abundantly, (3)
pelagic, the area over 2.0 m. deep, and (4) bottom, the ooze cover-
ing the floor of the lake. Zone 1 is characterized in places by Enter-
omorpha which grows abundantlj' on the rocks at some points along
shore. It is not common, however, occurring only where the shores
are covered with rocks, chiefly on the islands or "rock piles" in the
lake. Its presence indeed is so exceptional that it should perhaps not
be considered a "characteristic" of this zone.

Apart from Enteromorpha, zone 1 has no specific character-
istics. Cladophora al.so occurs here to a considerable extent,
attached to the rocks or to submerged logs, but this is more char-
acteristic of the following zone. It is a region where, at times,
certain species occur in large numbers, only to disappear again as
quickly. This erratic occurrence of organisms in zone 1 is due
chiefly to two factors ; first, temperature, and second, wind. The
former is one of the most important factors controlling the multi-
plication, movement and consequent distribution of organisms in
the lake. Consequently it is not surprising that wide variations in
the abundance of organisms in the littoral zone should accompany
the temperature differences in this zone, to which reference has
already been made.

Wind is also an important factor in determining the distribu-
tion of organisms in this zone, especially the massing of filamentous
algae which occasional]}' occurs here, as already noted.

Further reference to these features will be made in the discus-
sion of plancton distribution. In general it may be said that
zone 1 is marked rather bj' the absence than the presence of char-
acteristic forms.

The Ruppia zone is the most interesting, as well as important
one in the lake. It is interesting because of the variety of its
inhabitants, and important because of the influence of the great
mass of Ruppia, which develops here in summer, upon the chemistry
of the water, as an important source of the organic matter in the
ooze accumulating on the lake bed and as a breeding place for
numerous organisms, both animal and plant.

34 THE LIFE OF DEVILS LAKE

This zone is characterized primarily, as its name implies, by
the Ruppia, which grows here abundantly, and secondarily by the
organisms which live attached to the Ruppia or to other organisms
which grow upon it. Chief of these is Cladophora, which forms
extensive masses and which, perhaps no less than Ruppia, is respon-
sible for the evolution of large quantities of oxygen in sunny
weather, and contributes an important amount of the bottom ooze
thru the decay of its filaments. The exact quantitative role which any
one of several organisms plays in a situation of this sort is a diffi-
cult one to determine and no attempt has been made to do so in
this case. It is, moreover, of minor importance since both organisms
exert a similar influence, the effect produced being a collective one.

Growing on the Ruppia and Cladophora are numerous micro-
scopic forms, chief of which are sessile diatoms, peritrichous Pro-
tozoa and attached rotifers. There are many free-living forms also
which are numerous here and rare, or absent, elsewhere in the
lake. Chief among these are the Protozoa, rotifers, nematodes, and
mites, while a few others occur.

The extent of zones 1 and 2 is naturally a variable one, depend-
ent on the slope of the lake bottom at different points.

The pelagic zone, which comprises the major part of the lake,
has no forms peculiar to it, since all of its organisms occur in the
other two zones as well. It is characterized by the greater uni-
formity in the distribution of its inhabitants, tho great differences
occur here also, as will be seen later, and by the absence or rarity
of those forms which characterize zones 1 and 2.

The bottom zone, which comprises the ooze on the lake floor, is
characterized positively by the presence of the larvae of Chirono-
mus, by nematodes and the rhizopod Arcella, and negatively by
the absence of pelagic species, tho some of the algae, especially
Coelosphaerium may occur there. This zone is frequently lacking
in oxygen, as has been mentioned previously, and will be discussed
later in connection with its inhabitants.

PLANCTON METHODS AND RESULTS

The determination of the distribution in time and space of its
inhabitants is the central problem in the investigation of any area,
for this problem involves the reaction between organisms and environ-
ment, and the adaptation of the former to the latter, which are the
essential questions of ecology. This determination is a matter of
considerable difficulty, for both land and water animals. Many
attempts have been made to devise satisfactory apparatus for
plancton collection, but thus far none has been invented which is
uniformly satisfactory for all kinds of organisms. Full discus-
sions of the advantages and disadvantages of various types ot

THE LIFE OF DEVILS LAKE 35

apparatus are found in the works of many authors,* and need
not be repeated here, where it may suffice to summarize the diffi-
culties involved in their use.

1. Nets, even of the finest bolting cloth available, allow the
escape of a large part of the nanno-plancton.

2. The amount of water filtered by the net is a function of its
form, the rate at which it is handled, the age of the silk and the
consequent extent of clogging, and probably also of the temperature
and the amount of plancton present.

3. Plancton traps and pumps which depend upon the use ol
nets for the concentration of the catch are subject to the first diffi-
culty above mentioned.

4. Pumps are open to the further objection that their current
may serve to drive away some of the more active negatively rheo-
tropic organisms from the intake.

5. Water bottles collect only small amounts of water, and fur-
ther cannot be filled without the creation of some current, sub-
jecting them to the same criticism as applies to pumps.

6. The centrifuge in general handles only small amounts of wa-
ter, and if constructed of larger capacity is bulky, and consequently
impractical for general field work. Further, unless run at a high rate
of speed, some of the organisms, such as filamentous algae, especially
diatoms with flotation spines, may not be fully thrown down. The
large centrifuges require electricity, while the small hand instru-
ments require considerable strength and endurance on the part ot
the operator to run at a speed of 3,000 revolutions per minute,
for periods of one minute, which is hardly adequate for the pre-
cipitation of the lighter forms.

7. The Sedgwick-Rafter method, involving the use of a funnel
tube closed by a piece of bolting cloth and a column of sand, is
open to objection 1 above, since the sand in combination with the
bolting cloth does not retain all of the finer organisms. When
there is sufficient amorphous material and zooglea in the water to
retain the nanno-plancton it retards filtration, and the apparatu&
works very slowly, if at all, six hours or more being required with
some collections, even tho the scum on the surface of the sand b(»
broken occasionally with a glass rod. Many of the organisms are
retained by the sand when filtration is ended and some may be
retained by the walls of the tube itself. A careful discussion of its
accuracy has been given by Whipple (1914) who considers it accui-
rate to 10%, In obtaining samples for filtration, moreover, the
difficulties mentioned above under 4 and 5 apply.

*See especially Steuer (1910), Whipple (1914), Reighard in Ward and Whipple
(1918, Chapter III), and Juday (1916).

36 THE LIFE OF DEVILS LAKE

8. Concentration of the planctou by filtration thru paper, in
addition to the errors involved in collection of the samples ana
enumerated above, involves also one due to the adherence of a
considerable amount to the paper and its consequent loss in the
counting cell.

Were the errors involved in any method constant, they would
not cause any serious difficulty in plancton investigations, but i.a
every case they are the function of several variables and hence are
themselves variable. Thus in the use of a pump the current at
the intake depends on the rate of pumping, which is necessarily
inconstant for hand driven pumps, while machinery for the purpose
is expensive, bulky and generally impractical for field work.

This variation in current rate at the intake may influence the
number of organisms which escape due to their rheotropism. Further-
more the movement of the end of the hose due to the rocking
of the boat, which is obviously a very variable quantity, may influ-
ence their reaction, and the same objections apply in the case of
the water bottle, where the rate of flow is a function of the pressure
and this in turn of the depth.

Furthermore the rheotropic reaction on the part of plancton
organisms is very probably a function of temperature and other
less obvious factors in the life of the organism.

Satisfactory preservation of material is one of the most serious
difficulties in any plancton investigation. If the material is to be
counted fresh, it must be done promptly, especially in warm weather,
or it will deteriorate, and this is impractical in the case of collec-
tions made in the field at some distance from a laboratory. If the
material be living when counted the movement of larger forms such
as crustaceans and rotifers disturb the contents of the counting
cell so that accurate counting is impossible. Furthermore at the
time collections are made, press of more immediate duties may
render counting impractical. On the other hand preserved material,
especially those forms such as the alga Dictyosphaerium and bac-
terial zooglea, which are held together by delicate gelatinous sheaths,
is almost certain to deteriorate in course of time, so that if counting
be delayed for months or years a very serious error is introduced.
The enumeration of the plancton has to be made with low
powers (about 100 diameters) which renders the specific determina-
tion of some of the smaller forms difficult or impossible.

Still another difficulty in obtaining accurate plancton records,
and one to which sufficient consideration has not been generally
given, is the irregularity in the distribution of the plancton organ-
isms themselves. This irregularity has been noted by occasional
investigators in the past, but it has nevertheless been generally
assumed that one sample taken at any point in a given body of

THE LTFE OF DEVILS EAKE 37

water is representative of any other sample collected at a different
point, provided environmental conditions at the two points are
identical. This question has recently been carefully investigated
at Devils Lake by Moberg (1918), who found an average variation
ranging from 70% for some of the algae Coelosphaerium, Chro-
(icoccus, etc.) to 185% for the crustecean Moina.*

Maximum variations for the entire series are of course much
larger. Thus in one series Diaptomus showed a range of 400% of
variation from the mean and Monia 297%, and in another series
Brachionus satanicus showed one of 876%.

While Birge (1897) maintains the general uniformity in hori-
zontal distribution of plancton, he finds nevertheless in a series of
collections made at ditl'erent points on the same day* a range of
variation from the mean from 49%. for Cyclops** to 180% for
Daphnia pulicaria and Ergasilus. These variations, while not
necessarily proving the existence of swarms, as Birge points out,
do nevertheless show an irregularity of horizontal distribution so
great as to invalidate any general conclusions regarding plankton
distribution unless based on a large number of counts.

"My observations show so much variation in catches made at
the same place and in succession that I have little confidence in
the differential method of determining vertical distribution ; unless
a very large number of observations is made and averaged, so as
to eliminate the chance of variation in the single observation.'"
(Birge 1897, p. 376.)

Marsh (1898) in a series of collections made at the same place
on Green Lake, Wisconsin on two successive days, found that Diap-
tomus ranged from 291 to 2966, and similarly large variations were
found by him in other genera.

Birge himself (I.e.) and other writers have described the
occurrence of swarms, chiefly among the Cladocera and I have
observed them in both Moina and Diaptomus.***

Similar results have been obtained by other investigators
(Reighard, 1894, Apstein, 1896, Gantl()lti-lIorn3-old and Aimeroth,
1915, et al.), altho, in general, with smaller variations. Moberg's
work was done, chiefly on 500 cc samples, but in two series 19 litre
collections were taken. His results warranted the following con-
clusions. "1) The zooplankton in Devils Lake shows a great
irregularity in horizontal distribution, and this irregularity'' can-
not be correlated with any variations in amount of phytoplankton
or in the chemical and physical environment. It is more likely due

*Asplanchna showed a range of 219%. The number here is too bTuall, however,
to have much significance.

♦Time required for the series not stated. Presumably they were made under
as nearly as possible similar conditions.

•'Including two species.

•••See Moberg (1918. p. 266, f. n. 8).

38 THE LIFE OF DEVILS LAKE

to the habit of swarming among plankton animals, due perhaps
to a social instinct, similar to that found in many other groups of
the animal kingdom. Plankton swarms are at times visible, even at
considerable distances, to the naked eye, 2) With larger samples
(19 litres) the variations tend to be reduced, but even here they
are at times greater than in the smaller ones (% litre). 3) These
variations invalidate the usual assumption that a given sample of
water is representative of a large area, at least in respect to its
animal inhabitants, and necessitate the collection of large numbers
of samples before definite conclusions regarding their distribution
or movement can be drawn."

These difficulties, however, while serious in themselves, are
usuallj^ insufficient to obscure the comparative results of an ample
series of collections, and it is these comparative results in which
the ecologist is chiefly interested. It may, of course, be desirable
to know the actual amount of plancton per hectare which a given body
of water will produce, in order to know the number of fish which
that body will support. But who knows the number of cubic centi-
metres of plancton per cubic metre of water which will furnish
adequate sustenance for a fish of a given weight and given species?
And in general the suitability of any water for the cultivation of fish,
or other economic animals, depends more on temperature, chemical
character of water, breeding grounds, and oxygen supply, than on
the available amount of food; for in any water where the physico-
chemical environment is suitable for fish life there is pretty sure
to be adequate plancton for the support of as many fish as the
water is suited to contain.

On the other hand the comparative variation in amount of
plancton in both time and space, correlated with changes in the
environment, answers the ecologist 's questions regarding the reac-
tion between organism and environment and adaptation of the former
to the latter.

While no one method is adequate for the purpose, it has not
seemed practical for various reasons to employ more than one, and
I have accordingly employed that one which seemed best adapted
to the general purpose from every standpoint. This method has
been checked and supplemented from time to time by others, but
the results which follow are based almost entirely upon it.

In collecting samples I have used a water bottle closed with
a two hole rubber stopper, thru one hole of which extended a glass
tube with a rubber hose connection reaching to the surface, the
other hole being closed with a wooden plug attached to a string.
The bottle was contained in a metal box, weighted and open on
the sides, so as to insure easy sinking, attached to a marked chain.
When the box was lowered to the desired depth, as indicated on

THE LITE OF DEVILS LAKE 39

the chain, the plug was jerked from the cork and the water allowed
to enter in a steady stream as the air escaped thru the rubber hose.
"When oxygen samples were taken at the same time as the plancton,
the sample of the latter was taken from the larger bottle of the
Winkler apparatus used in collecting the former.

Prior to September 4, 1913, surface and shore collections were
taken by immersing a bottle below the surface. Thereafter they were
made by scooping it up and transferring to a jar. Comparative
tests of the methods of sampling by means of the scoop, the bottle,
the Winkler apparatus and the pump have been made for the
surface, and in the case of the two latter methods, for the bottom
as well. In these tests from 48 to 100 cc. of the samples collected by
each method was carefully measured and the animals contained in
it counted by eye in a white porcelain dish. The method is not a
very accurate one, but is sufficiently so to demonstrate any con-
siderable or constant variation in the different methods. With the
scoop water was taken both from the surface and from seven to
ten centimetres below. With the Winkler apparatus and pump
the intake was necessarily held a few centimetres below the surface,
and as nearly as possible at the same level in all collections. The
samples were taken as nearly as possible in the same spot and with
brief intervals between each collection. The results of these tests
show no constant difference between the various methods, and such
differences as occur are smaller in general than those found by
Moberg (r.c.) in the horizontal distribution of the plancton. In
the shore samples, however, where the number of Crustacea was
large, there is a distinct advantage for the scoop method.

A comparison of the results obtained by the plancton trap,
described by Juday (1916), the pump, Sedgwick-Rafter, centrifuge
and filter paper methods indicates no constant difference for the
Crustacea and rotifers in favor of either the Sedgwick-Rafter, pump
or trap methods, such differences as exist being inconstant and
doubtles due to variable distribution already referred to.*

With the bacteria and algae the results vary. For the fila-
mentous forms (Nodularia and Lyngbya) the Sedgwick-Rafter
method gives distinctly larger results. For Coelosphaerium,
Oocystis, Merismopedium and Micrococcus there is a distinct
advantage in favor of the centrifuge ; while in the case of Dictyos-
phaerium, Chroococcus and Chaetoeeros, there is but little difference
one way or the other.

In these experiments a small hand centrifuge, holding 15 cc.
of water in each tube and run for one minute at approximately
3,000 revolutions was employed and it was very difficult to throw
down all of the Nodularia. With a larger machine, operated at

♦See pp. 36-8.

40 THE LIFE OF DEVILS LAKE

a higher speed and for longer periods the results would doubtless
have been different.

Examinations have been made of the filtrate from the Sedg-
wiek-Rafter tubes which show a considerable loss of such forms as
Coelosphaerium, while, vice versa, filter paper, even if carefully
washed, retains a considerable quantity of material, so that in
either case the results are too low.

The usual planeton series includes readings at the shore, the
surface, some distance from shore, and at depths of 0.6, 2.1 and 3.0
to 4.3 m., the latter reading depending on the depth of the lake at
the point and time of collection. There was some variation from
this general plan, especially in the earlier years of the work, when
the collections were preliminary to the main series. Some of the
collections were lost so that the series is not quite complete. It is
sufficiently so, however, to give a good idea of the planeton abund-
ance and distribution thruout the periods of observation.

The counts have been made mostly by my former assistant,
Mr. E. G. Moberg. Some of them are my own, however, but these
have been cheeked with his bj^ comparative counts on the same
material. Whenever possible it is desirable that readings should
be made by the same observer, in order to remove the personal
equation, which cannot otherwise be eliminated entirely from work
of this character.

The series prior to August, 1913, are more or less fragmentary.
Beginning at this time, however, readings were taken at intervals
of usually a week to ten days until September, 1914. In some cases
readings were taken at intervals of one to two days, while there
are one or two intervals of two weeks duration. During winter
and early spring the interval was lengthened to from three to six
weeks. An additional series of readings with two to three week
intervals* was taken from Dec. 3, 1922, to July 23, 1923.

Readings were made at several levels instead of a single one
to ascertain the distribution of organisms at various depths. In
several instances these readings were taken in parallel series for
different hours of the day and night in an effort to determine the
vertical migration of the zooplancton. The results of these readings,
which were very indefinite, will be discussed later. The series of
readings from top to bottom gives a much better record of the
planeton abundance than would readings at only one level.

The accompanying charts give a graphic representation of the
occurrence of the various selected species and groups, during the
periods of observation fi'om 1911 to 1923.** The chart of one year
is not strictly comparable with that of another, in respect to total
numbers, at least for some of the algae and bacteria ; since the period

•In one caHe six weeks.

"♦I'.asi'Cl I'll oiii" liti'c samples.

THE LIFK OF DKVILS LAKE 41

elapsing between the time of collection and tliat of counting was
different for different years, with the result that some of the collec-
tions deteriorated more than others. Especially is this true of the
1914 series, which for the most part was not counted until 1919
and in which considerable disintegration of some of the algae and
bacteria occurred. No record has been made of Dictyosphaerium for
this year, since it appears in the counts only sporadically and has
evidently deteriorated. Deterioration of the closer colonies, such
as Coelosphaerium, occurred to u much less degree and probably
did not materially affect the records, while the diatoms were unaf-
fected. In general the counts of any one season are closely com-
parable with one another since they were made for the most part
at nearly the same time.

A study of the charts in general reveals several interesting
facts.

1. There is, in general, no evident relation between depth and
plancton abundance, conditions at the surface and at lower levels
not usually differing greatly, except in respect to light, which factor
is considered at greater length later.* The absence of a thermo-
cline in Devils Lake renders conditions here very different from
those in deeper lakes elsewhere. Oxygen may occasionally be
greatly reduced or absent at the bottom, but the layer of water
in which this occurs is thin, and it is quite possible that oxygen,
in the dissolved condition at least, is of less importance than has
generally been believed. This ([uestion is discussed at greater
length elsewhere, in connection with the organisms inhabiting the
ooze.**

2. There are great variations in the collections at all points,
tho most conspicuous in the case of those taken near shore. They
may be due in part, in the case of some of the colonial algae, to
imperfect preservation of material, altho this is unlikely, since the
various samples must in most cases at least, have been influenced
equally by this factor. In the case of the animals, swarms no doubt
are very important, tho the extent of their influence is still a
problem.*** Even among the algae, Moberg, in his study of hori-
zontal distribution, found great variations, too great to be attributed
to errors in collecting or in counting the samples, and apparently
due to chance. In any event the plancton, both plants and ani-
mals are 7wt distributed uniformly, as is generally assumed. In
the shore collections these variations are probably due largely to
temperature. "Wind may also play an important part, especially
with the filamentous algae, but other, less evident factors, appear
to enter.

♦See p. 46.

**See p. 106.

••♦In this connection, see Moberg (1918).

42 THE LIFE OF DEVILS LAKE

3. There is no close correspondence between temperature and
plancton abundance. It is true that in winter most species are
either absent from, or very rare in the collections; but from spring
to autumn there may occur several maxima and minima of plancton
numbers, which are apparently wholly unrelated to temperature.
These naturally differ for different species and different years, but
in general, occur, one in early summer, and the other in early
autumn. The causes are obscure and are probably manifold. In
spring there is usually a considerable influx of drainage water,
carrying with it dissolved materials, including much organic matter,
from the soil. Following this influx, and with increasing light and
rising temperature, there is a great development of both algae and
bacteria, especially the latter, accompanied in turn by a great
increase in animals, especially the rotifers. The maximum of early
summer is succeeded by a more or less well marked period of
depression during July, which may be succeeded by other periods
of maxima and minima in August. In the fall there generally
occurs a high point in the distribution curves accompanying a falling
temperature and decrease of light. These waves of production are
irregular and inconstant and their explanation is obscure. It is
not impossible that in midsummer the temperature is too high for
the successful development of many of the organisms in the lake;
or that, with their development in early summer the dissolved mate-
rials in the lake necessary for plant growth are largely consumed,
with consequent inhibition of development; it is further possible
that the great growth of bacteria may be responsible for the produc-
tion of some toxins inimical to the growth of other forms. The flrst
of these hypotheses seems the more likely, but none of them are
entirely satisfactory and the reason remains obscure. Similar
periods of maxima and minima in plancton production have been
described by other writers, but it is difiicult to compare the sea-
sonal distribution of plancton in different lakes because of 1) dif-
ferences in its environment and 2) differences in its component
species. A detailed comparison of my own work with that of others
would require far too long a discussion. It must suffice to indicate
some of the more important results.

1.) In studying the distribution curves for any lake, one Js
at once struck with the numerous minor irregularities which they
display. Take for example, the curves illustrating the vertical
distribution of the Crustacea and nauplii in Lake Mendota on
September 8, 1896 (pi. XLII, Birge, 1897). According to these
there were more adult Crustacea by about 35% at the 6 than at
the 4 m. level, and slightly more at 3 than at 2 m. ; in spite of
the fact that they were decreasing rapidly from above downward,
except in the first half metre. The nauplii vice versa, while

THE LIFE OF DEVILS LAKE 43

increasing from above downward, were nearly 50% fewer at the
4 than at the 3 m. level.

Again, consider the Diaptomus charts of Marsh (1898, pi VII).
In August, 1896, there occurred, according to the chart, two well
marked maxima and minima, with numbers ranging from 1563 to
3803*, a difference of nearly 150%. Similar, tho less marked
irregularities are shown in the curves of Birge (I.e.)-

Irregularities in vertical distribution, similar to those just
mentioned, have been discussed by Birge and Juday (1911, p. 116)
who consider them evidence of stratification of the organisms con-
cerned. They deny the likelihood of their causation thru errors in
collection or counting, but apparently overlook the possibility of
irregularity in horizontal distribution being responsible for an
apparent irregularity in vertical distribution, nor do these authors
attempt an explanation, merely contenting themselves with the
statement that "such results . . . should be expected . . ."

These differences may, of course, be accentuated, or the reverse,
by the scale and time interval chosen, and the number of collections
averaged. Taken in the aggregate they are of little significance and
do not obscure the main features of the curves. Considered sepa-
rately, however, they do have significance, in all probability indi-
cating the irregularity of horizontal distribution already referred
to.**

2. There appears to be no constant relation between the
environment and the periods of maximum and minimum production.
Reproduction is, in most cases at least, greater in summer than
winter, so that, obviously the factors of light and temperature play
the all important roles in determining abundance. But even this
rule has exceptions. Diatoms frequently develop in enormous num-
bers in winter, as noted by Marsh (1900, p. 176) in Green Lake,
Wisconsin, and others. The above statement obviously does not
mean that plankton production is independent of environment. It
is closely dependent thereon. But the factors are so many, and
so closely interwoven with one another, that it is difficult, if not
impossible, to determine any general laws governing their inter-
dependence.

There is often a plancton maximum in spring associated with
increase of temperature and light, with overturn and consequent
mixing of the water and influx of food materials in the run-off from
rain and melting snow. In addition to this maximum there may
occur other less constant maxima in mid-summer and autumn.
This is the general type of plancton production presented by Devils
Lake, Lake Mendota*** and the lake of Zurich,**** etc.

•Total catch.
**See pp. 36-S.
***Fide' Birge (1. c.)
****Fide' Lozeron (1902).

44 THE LIFE OF DEVILS LAKE

In no two lakes, however, are the environmental conditions,
or the fauna and flora exactly identical; it would therefore be
too much to expect that the behavior of the plancton should ba
the same in both. The behavior of any organism at any time, and
in any situation, constitutes a problem in itaelf, and a very complex
problem at that.

3. "While, in a general way, the forms of the curves of any
species are similar in two successive years, nevertheless their size
(i. e. the number of individuals present) varies greatly. This is
at once obvious on inspection of the plancton curves for two or
more successive years for any lake. Compare, for example, those
given by Birge (1. c. pi. XXV) for Diaptomus in Lake Mendota
in 1894-95 and '96.

"The feature of the annual distribution of the Crustacea which
surprised me most in the progress of my work is the great differ-
ence between the numbers of the same species of Crustacea present
in successive years. I do not refer so much to the larger or smaller
numbers of forms like Cyclops, for whose variations causes can be
assigned, at least in part, but rather to such facts as those shown
by Daphnia retrocurva and by Diaphanosoma, which are either
absent, or present in very small numbers in one season and appear
in great numbers in another year. For such variations it is very
difficult to assign even conjectural causes.

"A similar fact has appeared in the succession of the algae.
It is not true for lake Mendota that the forms of algae succeed
one another in a definite order in successive seasons, so that one
can be sure of finding certain forms at certain times of year, as
would be the case with plants of woodland or prairie." (Birge,
1. c, p. 317.)

Kobert (1919-20), after reviewing the work of several authoriii,
together with his own, reaches the conclusion (p. 41) that "Nous
ne pensons pas qu'un seul des facteurs generalement invoques:
temperature, circulation ou stratifications des eaux, puisse a lui
seul expliquer la date d' apparition des maxima et minima du
plancton. Ceux-ci dependent sans doute de facteurs fort complexes
et diflficiles a isoler, parmi lesquels ceux que nous avons etudies
jouent probablemciit un certain role."

Only the principal plancton species are represented in the
charts. Many otliers are of sporadic* occurrence, probably being
carried from the Ruppia into the pelagic zone by the action of
the wind. A few forms, such as Characium, which grows epiphy-
tically on the Crustacea and elsewhere, are of frequent but incon-
stant occurrence in the collections; while others, such as Cothumia
and Pediastrum occur in too small numbers to have much signi-

THE LIFE OF DEVIDS LAKE 45

ficance in the study of the plancton distribution. They will be
considered individually in the annotated list.

Only two phyla of animals have any considerable representa-
tion ill the plancton. These are the Crustacea and the rotifers. Of
the former there are three important genera of one species each,
and of the latter two genera and three species.

Of the Crustacea Diaptomus sicilis is of most importance, both
because of its numbers and its presence at all seasons of the year.
At times it is out-numbered by Cyclops and Moina, but these occur
in considerable numbers for comparatively brief periods, while
Diaptomus is always more or less common.

The nauplii are mainly those of Diaptomus, but no attempt has
been made to differentiate between them and those of Cyclops.
The latter are probably too few in general to affect the distribution
curve to any extent. In general the maxima of the nauplius curve
alternate with the minima of that of Diaptomus as might be
expected. The number of nauplii is, in general, much greater than
those of Diaptomus and Cyclops combined, which undoubtedly indi-
cates a high mortality in the larval stage. In Lake Mendota also
Birge (1897) reaches a similar conclusion regarding their mortality.

Contrary to his experience, on the other hand, I find the
nauplii apparently absent in Devils Lake in winter, tho gravid
females and eggs of Diaptomus occasionally occur at this time.

Birge (1. c.) finds the nauplii more abundant in sununer near
the thermocline, while, with the disappearance of this layer, their
distribution becomes more uniform. In Devils Lake nauplii are
somewhat more numerous near the bottom than the surface in
summer, the difference not being evident in spring and fall. The
reasons for this difference in distribution is not clear. Nor does
Birge offer any satisfactory explanation.

The curves for the various species will be considered in con-
nection with each individually. In the case of the zooplanctou,
especially the rotifers, the numbers and abruptness of the waves
indicate a brief life cycle and rapid development, but we have as
yet few data regarding the life span of any planctont. Needham
and Lloyd (1916) state that "the rotifer, Hydatina is said to have
a length of life of some thirteen days," but give no authority for
their statement. Steuer (1910, p. 269) gives the ages of a few
copepods based on the investigations of Burckhardt (1900) and
Ekman (1904), the average being 13 months.

This figure appears high for the average life span of fresh
water copepods. Judging from the curves for Devils Lake and
for Lake Mendota and Green Lake, Wis., as given by Birge (1897)
and Marsh (1898), the average life span is very much less than

46 THE LIFE OF DEVILS LAKE

this. But the facts, so far as known at present, do not warrant
any final conclusion.

Of the three species of rotifers which appear in the plancton
in any considerable numbers, two (Pedalion and Brachionus satani-
cus) far outnumber the third (Brachionus plicatilis). The
two former, in respect to numbers, and probably also in respect to
mass, are the most important constituents of the zooplancton.

The greater part of the phytoplancton is composed of the blue-
green algae and bacteria, altho diatoms and a few other green algae
are present in considerable numbers. It is difficult to draw any gen-
eral conclusions from the distribution curves, as they vary greatly,
not only for different species, but for different seasons in the same
species. These variations may be due in part, especially in the
case of the shore collections, to the effect of wind in driving the
filamentous forms especiallj^ onto a lee shore; in part to the influ-
ence of temperature, rainfall and other factors. In general there
is a maximum both for algae and bacteria in the spring or early
summer and in fall, with occasional smaller maxima between. Some
possible causes of these fluctuations have already been briefly out-
lined, but no final explanation can yet be given.

The light reaction of zooplancton is generally believed to be a
very definite one, the animals retiring from the surface by day and
distributing themselves more evenly at night. That the zooplancton
of Devils Lake is positively phototropic can be clearly shown
by keeping them in a blackened jar, with a small opening for
admission of light. After a few hours they will be found
swarming about the window. Vice versa toward sunset and after
the sun has been clouded for sometime they become nearly equally
distributed. Further, the effect is less striking in indirect than in
direct sunlight. If there are two openings in the jar, one (A) in
the direct rays of the sun and another (B) at 90 degrees thereto,
the majority of the Crustacea and rotifers gather at A, while a sec-
ondary gathering occurs at the edge of B nearest to A.

In the lake, however, they are under the influence of a com-
plex of variable factors, such as temperature, currents, dissolved
salts and gases, food, etc., and their resulting behavior cannot be
so easily analyzed. Ritter (1912) and others have well emphasized
the need for correlation between laboratorj^ and field studies of
animal behavior. In the study of a simple response the laboratory
method is undoubtedly better, because of the ease of controlling the
conditions of experiment, but in order to know how the animal will
respond to its natural environment, it is necessary to go to nature
to find out.

In Devils Lake, the light influence is obscured by so many
others that it is impossible to determine its effect, a careful series
of observations, extending over a period of years, having failed to

THE LIFE OF DEVILS LAKE

47

'^

/ °-

■a E

■S^

j^ ^

m a
u o

s *^ '-'

3 .S c

— 03

O *'

^■fr/s^

^

■3 H

N W -

48 THE LIFE OF DEVILS LAKE

show any influence of light on the distribution of the plancton
animals. In most of this work 500 c. c. samples were used, but in
two series 19 litres were pumped thru a Kofoid net, the amount
being measured with a water meter. The tests usually covered a
period of twenty-four hours in quiet weather, and were taken on a
line across Creel Bay. at both shores, and at the 0, 0.6, 0.9 and
3.0 m. depth near the middle of the bay.

The collections were made at approximately the following
hours: 12 N. 5. and 9 p. m., 12 M., and 9 a. m.

One reason for the failure of the tests to give definite results
is possibly the irregular distribution of the plancton, as Moberg
(I.e.) has already suggested.

EAST LAKE (pi. 2)

East Lake is a long, narrow body of water extending southeast
from the "Narrows" for some 17 km. It formerly was connected
with Main Lake on the one hand and with Lamoreau Lake on the
other, forming one continuous body of water, and is so represented
on the present maps. It was separated from Main and Lamoreau
Lakes about 1903. At the present time it has retreated about 2
km. from Lamoreau Lake, its former bed being occupied by a low,
grassy swale in which are two nearly fresh water ponds, evidently
spring fed.

With its separation from the other parts of Devils Lake the
fall in East Lake has been rapid. When work -was begun in 1911
it had a maximum depth of about 3 m., while at present it can
hardly be more than a metre. The reasons for this are possibly
threefold. In the first place the shores of East Lake are in gen-
eral very low and flat and the drainage per unit of lake area is
probably much less than that of the other parts. In the second
place the supply from springs may be less, altho this source of
supply is probably very small in all three bodies. Thirdly the
shallow waters of East Lake are more disturbed by the wind than
are the deeper waters of the other two parts, and hence the evapora-
tion is greater.

The decrease in depth has naturally been accompanied by pro-
found changes in its physics, chemistry and biology. I have but
two color records for East Lake, both of which were made on sam-
ples which had been collected several days previously. One of
these gave a reading of 86 for the surface and 64 for the bottom,
and the other of 68 for the surface and 40 for the bottom.

The few turbidity readings taken show a range of from 25 to
46.

Some curves of temperature and dissolved gases are given in
figures 10 and 11. They show differences between top and

Plate 8. Above, Lamoreau Lake.
Below, Mission Lake in left foreground. Mission Bay of Main Lake in right background.

THE LIFE OF DEVIDS LAKE 49

bottom similar to those of Main Lake. The shallowness of the
lake renders it more quickly responsive to the air temperature than
the deeper waters of the other lakes. In the earlier years of the
work (1911-14) there were the same differences, altho not so mark-
ed, between the shore and mid lake temperatures as noted for
Devils Lake, but at present such differences are undoubtedly negli-
gible, if existent. Its shallowness renders the temperature of the
entire body higher in summer, while its greater salt content renders
the freezing point, and hence the water temperature, somewhat
lower in winter than the others. At present it freezes solid in win-
ter.

The chemical character of East Lake is naturally similar to
that of the main body, but has gradually become more concentrated
with decreasing depth (table 3). Formerly East Lake showed the
same zonation as Main Lake, but at present no such distinction can
be made. Plancton records cover the period from 9/2/13 to 9/1/14,
with a few prior thereto. They were taken at the shore and at
depths of 0.6 and 2.0 m. The plancton species, at the time these
records were made were similar to those of i\Iain Lake.

Probably owing to poor preservation of material the data for
Micrococcus, Sarcina, Merisniopedia and Dictyosphaerium are inade-
quate and have accordingly been omitted. Brachionus plicatilis
which appears in the collections of June 21 and July 12 only, is
also omitted.

MISSION LAKE (pl. 8).

This is a small body of water resulting from the isolation of a
bay of the main lake in 1909 or 1910. In spite of the fact that the
lake probably receives a small amount of seepage from a fresh
water lake situated on a terrace about a half kilometer distant, its
history has been similar to that of East Lake, so that today it is
only a shallow pool, nearly filled with foul smelling ooze. In win-
ter it freezes solid and in summer is heated to 30° so that its life is
subjected to a strenuous existence.

The chemical analysis and osmotic pressure are given in table 3.
The life of Mission Lake in kind, distribution and abundance was
formerly the same as that of the parent body, but with its separa-
tion, and consequent changes in its physico-chemical character, great
changes have naturally also taken place in its fauna and flora.
Only a few quantitative collections have been made, so that there
are no records available of the seasonal distribution. Compared with
collections from Main Lake these collections show a considerably
greater number of zooplanctonts and much fewer phytoplanctonts
than collections taken at the same, or nearly the same time from the
former water. Whether this difference is constant or not the
collections are too few to determine, and if constant, the reason is

50 THE LIFE OF DEVILS LAKE

not obvious. It would appear that the Crustacea and rotifer*?,
which constitute the great bulk of the zooplancton, finding in Mis-
sion Lake a favorable environment, had multiplied rapidly; and,
using the plants as food, had reduced the latter accordingly. At
present, (1924) the Crustacea (except Marshia) and the rotifers have
apparently disappeared.

LAMOREAU LAKE (pl. 8).

This was formerly the eastern end of Devils Lake, but about
1903 was separated from East Lake by the construction of a high-
way across the Odessa Narrows. Its size, depth, drainage area and
possibly some supply from springs have sufficed to keep it in about
the same condition as Main Lake, so that its life, in general, is
similar to that of the latter. It covers an area of about 15 km. and
in 1914 had a maximum depth of 7 m., slightly greater than Main
Lake at the same time. The chemical analysis together with osmotic
pressure is given in table 3, while Og, and temperature are shown
in figure 12. Quantitative plancton collections are too few to havs
much significance. They indicate a smaller number of plancton ani-
mals (especially rotifers) than were present about the same time
(8/6/14) in Main Lake, and a somewhat greater abundance of
Crustacea than were present in the adjoining East Lake.

In the algae the most conspicuous differences between La-
moreau and Main Lakes at this time were the large numbers of
the diatom Cyclotella in the former and its comparative rarity in
the latter.

Nodularia, which is so characteristic of these lakes, was also much
more abundant in Lamoreau than in Main Lake (Creel Bay) at this
time, but much less so than in East Lake. These differences, how-
ever, may have been due to some local conditions and of no import
in reference to the life of the lakes as a whole.

STUMP LAKE (pl. 5)

"With the possible exception of Main Lake this is the largest
body in the whole complex and its depth is greater than any.

As with Devils Lake, the shores are in places rocky, in others
muddy and flat. On a flat at the southeastern end are the stumps,
from which the lake derives its name. In places the old shore Lines
can be traced, marking the levels of glacial lake Minnewaukan,

The lake is mainly supplied by run-off, but in places, especially
at the southern end, are some considerable springs, while near the
northwestern end are a series of small ponds, seepage from which
may reach the lake underground. There are no records of the
levels of Stump Lake, but the old shore lines, and the character of
the water teU a story similar to that of Devils Lake.

THE LIFE OF DEVILS LAKE

On

51

2 -

4-

5-

fe-

7-

8.

Temp
0,

O" he g.,
^0° 21' 22°

T 1 1 — ■■» -1 —

3^< Vrc 5"cc 6^c 7 -a
2 3° 2*° XS" 2.6° ' Tkmp

Figure 12. Temperature and oxygen curves for Lamoreau Lake, 8/6/14. Ordinates represent
depths in metres, abscissas degrees of temperature (centigrade) and c. c. of O2 pr. 1.

52

THE LIFE OF DEVILS LAKE

Srump LnnE ij^j'ie

Figure 13. Curves of dissolved oxygen in Stump Lake (near west end), A. M., N. and
P. M. 7/9/18. Ordinates represent depths in metres.

Studies of Stump Lake, as with all others of the complex, ex-
cept Main and East Lake, have been confined mainly to the chemical

1-

c

/

/'

/

■t-

I I " I I 1 I

*/• 3Z' 23* 2V' 73' 26" X7' r<fr,0.
A.. Z.^ 3-. Vr. See 6<t 7.. 0,

Srur^P Lfi>H£ l/lijjQ

Figure 14. Curves of temperature and dissolved gases in Stump Lake (west end) 7/1 1/18.
Ordinates represent depths in metres.

THE LIFE OF DEVILS LAKE

1-

53

z -

3-

<t -

5-

b-

1 '

a

CO:,, M CO^

7~Z^

I I 11 I I

!O0 300 W50 5l» «<» TOOf.p t^

i<r 20'

23' Z"*' Tfy^f.

Figure 15. Temperature and CO2 graph, for Stump Lake (middle) 7/10/18. Ordinates

represent depths in metres, abscissas temperatures in degrees centigrade,

c. c. of Oj pr. 1. and ppm. of carbonates.

54 THE LIFE OF DEVILS LAKE

analysis of the water and the determination of its inhabitants. A
few records of temperature and dissolved gases for 1918, indicate
very similar conditions to those of Main Lake (figs. 13-15).

The deepest part of the lake (7.6 m. in 1918) is near the
southeastern end of the main body. Thence the lake gradually
shallows in both directions, while the western arm in 1918 did not
exceed a depth of 3 m. at any point. The bottom is covered with
an ooze, similar to that on the floor of Devils Lake.

The chemical anah^sis and osmotic pressures are shown in table
3. There is a much greater amount of dissolved substances due to
the higher chlorine and sulphate content, with a correspondingly
greater osmotic pressure. A considerable part of the chlorine is
present as NaCl, as evidenced by the distinctly salty taste of the
water, which is lacking in Main Lake, or concealed by other in-
gredients.

The same life zones occur as in Main Lake, and the constituent
organisms are similar. No extensive plancton studies have been
made and there are no coincident records from Main Lake for com-
parison. The few collections made in 1918 indicate, however, a
comparative scarcity of rotifers, a fairly average number of crust-
aceans and a considerable abundance of Nodularia and Chaetoceros.
The distribution of the latter is evidently local however, as it ap-
peared in large numbers (1000-1500 cells per cc.) at the eastern
end of the lake on July 10, 1918, while only from 20 to 40 cells
per cc. were present at the southwest end, about eight miles distant
the following day.

The considerable abundance of Nodularia is of interest, as
this alga appears to increase with increasing salt content, at least
up to a certain limit. In East Lake also there was an enormous
amount of Nodularia present in 1913-14, correlated possibly with
the greater concentration of its water. But, on the other hand, no
increase of this alga has been detected in Main Lake since 1911,
while the concentration of its water has materially increased since
that time,* so that the reason for its abundance in East and Stump
Lakes is uncertain.

LAKE A (pl. 9).
This small lake was cut off from the main body about 1885.
It is approximately 1.3 sq. km. in extent, with a maximum depth of
about 2.5 m. While there has been a marked drop in the level of
the lake since its isolation, it is, at present, fairly constant in level,
due undoubtedly to a small supply from springs. At one point a
very slight stream trickles into the lake, and there is probably some
underground seepage. Kun-off and rainfall can hardly be ade-

*Cp. the charts of distribution of Nodularia in Main and East Lakes (pl. 15 and
fig. 20) with the chemical anals'ses (table 3.)

Plate 9. Above, Spring Lake.
Below, Lake A.

THE LIFE OF DEVILS LAKE 55

quate to maintain a constant level. Like other shallow lakes its
temperature in summer probably reaches a maximum of 30° C.

It has a high salt concentration, as shown in the analysis
(table 3), with a correspondingly high osmotic pressure.

Its inhabitants, so far as studied, are similar to those of Main
Lake.

LAKE C

Lake C is about 0.3 sq. km. in extent with a maximum depth
of 1.8 m. It was separated from what is now East Lake, about
1885. It lies in a shallow draw between two morainic ridges, and
is probably spring fed, having apparently maintained its present
level for some time.

In spite of its constancy of level, it shows a considerable con-
centration of dissolved solids as indicated by the analysis (table 3).

The study of Lake C has not been thoro enough to admit of
any very definite comparison between its life and that of other
lakes in the complex. In general this appears to be much less
abundant both in species and individuals than that of neighboring
ponds. A striking feature is the apparent absence of Nodularia,
which is so characteristic of the more alkaline lakes of the complex.
The same absence is noticeable in Spring Lake and Lakes N, 0 and
P. The reason thereof is not obvious. Where comparatively few
collections have been examined, as is the case with these lakes, it
would not be safe to be too emphatic in making negative conclusions.
A sufficient examination has, however, been made to reveal Nodu-
laria were it common, as it is in the principal lakes of the complex,
so that its absence is probably significant, tho of what, cannot at
present be said. The salt content of lake C is rather low, about the
same in amount and proportion as that of Main Lake in 1907. In
1909 Nodularia was observed as one of the most characteristic species
in Main Lake. The other lakes, with the possible exception of Lake
P, are all shallow, N and 0 being but temporary pools, and rang-
ing in different seasons from nearly fresh to exceedingly alkaline
ponds or mud flats. This may explain the absence of Nodularia
in them, but no satisfactory explanation is at present evident for
Lake C.

SPRING LAKE (pl. 9).

This is little more than a mud flat covered by a third to a
half a metre of water. Were it not for a supply from springs it
would undoubtedly have long since been dry. It has been separated
from the main body for about 40 years.

In winter it freezes solid, while in summer it reaches a temper-
ature of about 30°.

56 THE LIFE OF UEVILH LAKE

The chemical character of the water is indicated iii table 3.

Spring Lake possesses several species of rotifers, not as yet re-
ported elsewhere in complex, but in other respects, its characteristics
are chiefly negative.

LAKE P (pi. 10).

This small pond was separated from Devils Lake about 1906.
It is probably spring fed, as it maintains a fairly constant level from
year to year, in spite of the lack of supply, apart from run-off and
possible springs. Its maximum depth is between one third and half
a metre.

The chemical analysis and osmotic pressure are given in table
3.

Lake P has several species of algae not reported elsewhere in
the complex, which are of interest, coupled with the absence of
Nodularia noted above.

One of the most interesting inhabitants of this lake is the larva
of the dragon fl.y Sympetrum. This has not been found in any
other of the brackish lakes of the complex, and is probably to be
considered as a typically fresh water animal.

Triaenodes flavescens is another species occurring only in Lakes
C and P and in neighboring fresh water ponds; all of these facts
seem to relate Lake P rather more closely to fresh than brackish
waters, and j'et the analysis (table 3) shows the water to be highly
brackish. I cannot at present reconcile these apparently irrecon-
cilable facts.

LAKES N AND O

Both of these are shallow pools in spring and continue as such
during wet seasons. They occupy parts of the old lake basin, from
which they were separated about 1886-87. Like other pools they
freeze solid in winter, when water is present, w^hile in summer their
temperatures must occasionally run to 30° or over.

Conditions in these lakes are too variable, and their study is
as ,yet too incomplete, to justify any conclusions regarding their
characteristics.

Besides the lakes already enumerated there are several fresh
water ponds which at one time were a part of Devils or Stump
Lake. These I have omitted from the discussion, because, being
fresh water, they are very different, both chemically and biological-
ly, from most of the cut-off lakes forming part of the Devils-Stump
Lake complex.

Another lake has only recently (1923) been separated from
Devils Lake by the construction of a highway, but its recent separa-
tion renders its consideration unnecessarv here.

Plate 10. Above, Minnewaukan Bay in 1916.
Below. Lake P.

THE LIFE OF DEVILS LAKE

57

ANNOTATED LIST OF SPECIES

BACTERIA

The bacteria of Devils Lake are abundant. Apart from the
usual sources of infection tliru wind and drainage, considerable
sewage from the city of Devils Lake 5 km. distant, flows into Creel
Bay thru a drainage ditch.

In the plancton collections the bacteria include one of the most
common types. Naturally no Mttciiiiit has Ix-oii made to distinguish

Figure 16. Water bottle for bacteriological samples, see p. 58 lor description.

58 THE LIFE OF DEVILS LAKE

species in the enumeration, and the form specified probably includes
several. It is designated by form simply as Micrococcus. Besides
these a less frequent type, Sarcina, is of common occurrence.

The bacteriology of Devils Lake was studied during the sum-
mer of 1916 by Mr. (now Dr.) Fi. "W. Stevenson, a pre-medical
student of the University, under the direction of Dr. L. D. Bristol,
then director of the Public Health Laboratory of the State. The
methods followed by Mr. Stevenson are those recommended by the
American Public Health Association in their "Standard Methods of
Water Analysis."

Samples for quantitative analysis were collected in the morn-
ing, shipped to the University in an iced container, and plated in
the evening of the same day. With those intended for qualitative
analysis only, especially with those from outlying parts of the lake
complex, several days elapsed in some instances between time of
collection and plating, and no attention was paid to keeping the
sample at a low temperature. In collecting samples from points
below the surface of the lake, the apparatus illustrated in figure 16
was employed. It consists of a metal frame f, weighted at the
bottom with a block of lead. w. The frame carries a collar c, one
side of which is hinged and can be swung open to admit a bottle,
the latter being held in place by screwing the two parts of the coi-
lar together at s. The stopper of the bottle is held by a screw clamp
k, at the base of a rod, r, which latter is attached to a cord, 1, and
surrounded by a spring, p. The rod, r, passes thru a collar, d, which
is bolted to the top of the frame by means of a nut, n. The frame
is suspended by two cords, m, which are attached to a small plate,
h, thru a collar, e, at the center of which the cord, 1, passes, and
which is in turn suspended by a line, t. The apparatus is lowered
into the water to any desired depth by the line t, when a pull on
the cord, 1, removes the cork and allows the bottle to fill. When
the latter is full, the cork is replaced by repleasing the tension on
the spring, p.

The seasonal distribution of bacteria is indicated by the curves
for Micrococcus platted from the planeton counts (pi. 11). They
show, in general, an enormous development in summer, with a
marked decrease or total disappearance in winter. Further than
this, it is difficult to draw any general conclusions from the dis-
tribution curves, as they differ considerably in different years.

In 1912 and 1913 definite maxima occur about June 1 with
indications of a mid-summer minimum and a small maximum near
the end of August, at least in 1913. In 1911 a minimum is indicat-
ed about August 15, tho the records here are too few to warrant
definite conclusions. In 1914 there is a June maximum without,
however, any sharply marked peak, and a very distinct minimum
al)ont Jnlv 20. ffdlnwed bv a great increase in August. The 1914

THE LIFE OF DEVILS LAKE 59

material was in poor shape, however, when counted, and the form
of the curve may be somewhat obscured as a result. In the records
from December 3, 1922, to July 23, 1923, Micrococcus was disap-
pearing from the plancton, when the records began. During the
winter it was absent from most of the collections, tho present in
every series in the shore samples. It increased rapidly in May and
June to a large and well marked maximum about the middle of the
month, falling off rapidly until the end of the collection period.

Much of the material included under Micrococcus may be
Clathrocj^stis, which occurs commonly in the Devils Lake plancton,
and which in preserved material "might closely resemble the coccus
form of the bacterium and I doubt if it would be possible to separate
the two".*

Micrococcus, while present in East Lake, had for the most part
deteriorated when the counts were made.

Sarcina is much less abundant than either of the preceding,
and the records in general are too scattering to admit of anj^ con-
clusions reg'arding seasonal distribution other than that a marked
decrease occurs in winter, with a corresponding increase in summer
and autumn. Its distribution is approximately similar to that of
the two preceding, altho it is conspicuously less abundant in spring.
It is possible that some of it may have been included with Micro-
coccus.

The quantitative analyses covering the period from June 27 to
August 9, 1916, and including 27 samples were designed to show
rather the spatial than the temporal distribution of the bacteria in the
Devils Lake complex. They showed quite clearly, as might be expected,
the much greater abundance of bacteria in situations where there was
a greater amount of decaying matter, such as the shore and bottom,
compared with the surface, or the lake some distance out from
shore, and the heads of shallow bays as compared with the middle
of the lake. Especially numerous were the bacteria near the mouth
of the Devils Lake sewer at the head of Creel Bay. The B coli group
were not limited to the former situations, but were widely distribut-
ed thruout the lake.

These results are of interest when compared with the organic
analyses, in a majority of which (about 2 to 1) the organic con-
tent is higher at the bottom of the lake than at the surface.

The distribution of the bacteria in the various lakes of the com-
plex has not been studied thoroly enough to warrant any definite
conclusions.

During the summer of 1916 the Mauvaise Coulee was flowing into
Devils Lake carrying with it large numbers of bacteria. This un-

•From a letter to the writer from Dr. Geo. T. Moore, who Identified the
Devils Lake algae.

60 THE LIFE OF DEA^LR LAKE

questionably influenced the numbers of bacteria in Minnewaukan
Bay and possibly to some extent in the main lake at this time.

None of the bacteria in Devils Lake occur uniformly in all of
the eollectious, altho most of them occur frequently. In most cases
their presence is apparently accidental and not characteristic.

The list of species identified is probably incomplete ; which,
coupled with the irregularity in their occurrence and the lack of
adequate data regarding- each, prevents a detailed discussion of them
here.

Bacillus subtilis appears to be the commonest bacterium in the
lake, but it is difficult to make any selection among the others.

The list follows:

Micrococcus candicans, citreus, descidena. orbicularis, plumosus.
radiatus, rosettaceus, simplex. Streptococcus albus, vermiformis.
Sarcina alba, lutea, subflava. Bacillus aerogenes, albus, aquatilis,
coli communis, convolutum, fluorescens, formosus, proteus vulgaris,
reticularis, subtilis, vermiculosura, Bacterium aerogenes, desidiosum,
flexuosum, nubilum, tiogense.

ALGAE

Of the algae the blue-greens and diatoms furnish by far the
greatest number of species, the three most characteristic genera
being Nodularia, Coelosphaerium and Chaetoceros. Many of the
species described do not occur, or cannot be recognized, in the
planeton samples, due in part to their rarity, in part possibly to
poor preservation, in part to the difficulty of recognition under the
low magnifications used in counting and in part to their presence
chiefly among attached algae, where the samples were not taken.

Dr. George T. Moore who has made a preliminary report on
the Myxophyceae and Chlorophyceae of Devils Lake (1917) says:
"The algal flora of Devils Lake can hardly be regarded as a
rich one. The Myxophyceae are, it is true, fairly well represented,
and the fact that they constitute practically 50 per cent of the
genera and species present* may be regarded as one of the effects
of the increasing salinity of the water. The almost entire elimina-
tion of the Conjugales, which so frequently constitute the greatest
number of species in fre.sh-water lakes, is likewise to be attributed
to the high content of salts, and perhaps the absence of this order
is sufficient alone to account for the small total number of species.
With the exception of Spirulina nordstedtii,* all the species listed
are frequently, if not invariably, found in fresh water, so that the
algal flora of the lake is to be regarded as typically a fresh-water
one, showing no effect of the gradual concentration of the water.

♦Not inrludlng thp diatoms.

•Subtllissima is also present. Moore describes it as "a salt water and sulphur
spring form."

THE LIFE OF DEVILS LAKE 61

"Furthermore, of the species identified, it is worthy of note
that with minor differences of measurement, they were all abso-
lutely typical, with no indication of any effect of the, in many
cases unusual, if not unfavorable, environment."

The following list is compiled mainly from Dr. Moore's report,
the notations being partly from his, and partly from my own data.

Myxophyceae — Most of the species recorded below are free
floating and of general distribution thruout the lake, which may
be assumed to be the case for each species, unless it is specifically
stated to the contrary.

Aphanothece castagnei — This species occurs commonly in tow-
ings. In the quantitative samples numerous ovoid cells were found
which probabi}' belonged here. The colonies soon disintegrate how-
ever so that it is difficult of identification. It occurs from the end
of May to December, but the records are too scattering to justify
any conclusions regarding seasonal distribution.

Bhahdoderma sigmoid ea has been reported from Main, East
and Mission Lakes, and var. minor from Lake P.

Chroococcus limneticus and minutus. Both of these species
are common and occur frequently in the plankton samples. In
counting no attempt has been made to differentiate between them.

It is difficult to draw any conclusions regarding the cycle
of Chroococcus from the distribution curves. It occurs in rather
large colonies, and one count may show none or few, while another
shows a large number. To this fact may be due, in part at least,
the more or less zigzag form of the cui-ve, which is especially noticea-
ble in 1914. It is rarely very numerous in the collections, which,
together with its colonial habit, and the liability of the colonies to
disintegrate in preserved material, thus rendering it difficult to
identify, tends to obscure the form of the curve. In its seasonal
occurrence it is similar to other algae, appearing early in April
and disappearing again toward the end of December. Its maximum
falls between August 10 and September 15. Its occurrence in the
East Lake collections is too scattering to warrant any conclusions
regarding its seasonal distribution there (pi. 11 and fig. 18).

A similar irregularity is evident in Lake Mendota (Birge and
Juday, 1922). Chroococcus is present here for about four weeks from
the end of May to the end of June, then apparently disappearing
until the middle of September, after which it occurs until winter.
In 1916 it occurs from the first of May till mid-October, with the
exception of a brief interval about August 1, but is not present later
in the autumn.

Two other species of Chroococcus, dispersus and turgidus, occur
in Lake P, and the latter has been reported from Main Lake.*

♦Brannon (1910).

62

THE LIFE OF DEVILS LA.KE

-r/i

1^

0^ S

"

-'V

J^

I \

m \i\>

I'm

i- ^'

t

/

THE LIFE OF DEVILS LAKE 63

Clathrocystis aeruginosa. Reported by Dr. Moore as "abundant
in towings and associated with filamentous algae." In the quan-
titative plancton samples it has not been possible to distinguish this
species. The cells are very small and the indefinite form of the
colony renders it difficult, in preserved material and with low
magnifications, to differentiate it from masses of zooglea, so that
some of it may be included under "Micrococcus" in the counts.

Microcystis incerta. Taken in Lake P only.

Coelosphaerium kuetzingianum. This is one of the most abun-
dant forms in the lake. With low magnifications it is very difficult
to distinguish it from the following species, so that both have been
listed together in the counts. Reported from Main Lake and Lake
C.

Gomphosphaeria aponina. A common and widely distributed
species. Var. cordiformis has been found in Lake P. The seasona\
distribution of this species is shown, together with that of the pre-
ceding, in the chart (pi. 12), which shows that they first appear in
March or early April, and are present until after the freeze-up in
December. The record for 1913-14 shows their presence thru the
winter until March 1, followed by a brief absence and re-appearance
about April 1. In 1911 a marked drop occurred in August parallel
to that of Nodularia, Merismopedia and Chaetoceros, but not parallel-
ed by Chroococeus or Oocystis. This behavior does not correspond
to the record of subsequent years. In 1912-14 there is apparently
but a single maximum for each year, but this occurs at different
times in 1912 (Aut. 25-Sept. 1), 1913 (Oct. 20) and 1914 (May 20-
30). In the latter year a poorly marked minimum occurred July 25
to August 1, followed by a slight increase in August and September.

Coelosphaerium and Gomphosphaeria are much less common in
East than in Main Lake (fig. 19).

The occurrence of Coelosphaerium in Lake Mendota (Birge and
Juday, 1922) is less regular than in Devils Lake, for in the former,
in 1915 it was absent or rare until June, developing in considerable
numbers from July onward.

Merismopedia convuluta, elegans, glauca and tenuissima. Of
these the latter "is by far the most abundant." No attempt has
been made to differentiate between them in the counts, the results
of which are shown in plate 13.

The curve corresponds rather closely to that of Chroococeus,
with a distinct drop in mid-August in 1911 and a maximum at the
same time in 1912 and 1913. In the 1914 collections probably most
of the material had disintegrated before the counting was done, so
no definite conclusions can be drawn for that year. In 1912 a small
maximum is apparent about the end of May, and this is also indi-
cated in the 1913 curve, altho the collections here are rather too
scanty to admit of any definite conclusions. Merismopedia sp. oc-

64 THE LIFE OF DEVILS LAKE

curs occasionally iu the East Lake collections, but the records are
very few. M. elegans has been found in Mission Lake only.

Tetrapedia gothica. "Sparingly found in two collections, one
from towings, the other associated with Cladophora near the shore
in front of the laboratory. This, so far as is known, is the first
recorded collection of this genus in the United States, and it is like-
wise the first time the species has been found since its original dis-
covery by Reinsch in Germany."

Tetrapedia sp. has been found in Spring Lake.

Arthrospira jenneri. occurs occasionally in Main, East and Mis-
sion Lakes.

Oscillatoria. This genus is represented by ten species, one of
them indeterminate. Of these amphibia is 'the most abundant and
widely distributed." Specimens of Oscillatoria occur occasionally
in the plancton collections, but not often enough to warrant any
counclusious relative to its seasonal distribution. Beside amphibia
the following species occur, hrevis, chalyhea, chlorina, geminata,
janthiphora, Umosa, suhtilissima, tenuis and sp-, "a small form
. . . resembling in a general way the descriptions of O. sub-
tilissima. ' '

Plectonema tenue. "Numerous filaments of this species were
obtained from surface tow in one of the smaller bays of the lake.
While it is probable that the material had become detached from
stones, a search along the shore in shallow water failed to reveal
any indication of from where the specimens had come."

Spirulina. This genus is also of sporadic occurrence in the
plancton samples, and hence is not represented in the charts. Of
the four species present, major, nordstedtii, suhtilissima and tener-
rima, the first is the most common. Nordstedtii and suhtilissima
are of interest as being the only salt water representatives of either
Myxophyceae or Chlorophyceae present in the lake.

Anahasna, represented by two species, flos-aquae and spiroides.
occurs rarely. It is found occasionally in the plancton collections,
but the data do not enable me to draw any conclusions regarding its
seasonal distribution.

Lyngbya conlorta was formerly one of the most characteristic
forms of the plancton (pi. 14).

In 1911, 12 and 14 it was abundant from September on to the
season's close*, while in 1913 it showed a great decrease in Septem-
ber and October. In both 1912 and 13 a distinct maximum occurred
at the end of May and early June, while in 1914 this maximum,
tho present, was very small. In 1913 a very large and well marked
maximum occurred about the middle of August, and apparently a
similar maximum occurred about two weeks later in 1912. In 1914

♦In 1914 this' \v;is on Septt-niljcr 14.

XI

;i , ,-^> . , (-y^F-rr.

^'d

i ■ 'i

THE LIFE OF DEVIDS LAKE 65

there was a marked increase about September 1, which apparently
corresponds to the mid-August maximum of 1913.

In 1911 the records are too few to warrant any definite con-
clusions.

In 1923 there are only a few scattering records of Lyngbya.
Apparently it is much less numerous now than formerly, correlated
possibly with the increasing concentration of the lake. It appears
to be absent from Stump and Lamoreau Lakes and very scarce in
East Lake. It was abundant in Main Lake in 1916 and present in
some quantity in 1918. altho the records for this year are too few
to warrant any conclusions as to its relative abundance at this time.
It is of interest that in 1916, when still abundant in most of Main
Lake it was apparently absent from Minnewaukan Bay which was
nearly fresh at this time*. Its absence or scarcity in Stump and
East Lakes supports the hypothesis of its dependence on a certain
degree of concentration (10000-15000 ppm.) ; but its apparent ab-
sence from Lamoreau Lake, which has about the same chemical
character as Main Lake, negatives this conclusion and renders the
cause of its erratic distribution in the complex uncertain. It was,
moreover, common in two collections from Mission Lake in 1916,
altho the concentration of the latter was 32640 ppm. at this time.

Lyngbya martensiana has been reported from Lake 0.

Nodularia spumigena. Excepting Coelosphaerium and Lyngbya
this is the most common alga in the lake. It is frequently driven
by the wind into extensive masses along the lee shore, while other
regions may be comparatively free from it. It develops in consider-
able quantity rather later than most of the other algae, appearing
sparingly during April and May and showing a rapid increase in
Jvine, with a maximum about the end of July and a more or less
distinct minimum about the middle of August. This was particu-
larly striking in 1913 in which it fell from an average for the sev-
eral depths of collection of 1720 on August 1 and 2 to an average
of 285 on August 20-26. But it occurred also in 1911 (August 8-
22) and 1914 (August 17). In 1912 the data are too few at this
season to warrant any conclusions.

In 1914 a marked minimum occurred July 20-22, but the sud-
denness of the previous drop and the succeeding rise renders this
minimum open to question. Prior to July 20 Nodularia developed
mainly along shore and at the surface, so it is possible that a shift in
the wind, driving it into other parts of the lake, may have caused
this apparent drop (pi. 15).

Nodularia appeared to be much more abundant in East than
in Main Lake, in 1913 and 1914 (fig. 20), but the records are too
few to permit of any definite conclusions*. It was abundant in

*See table .1.
♦See p. 54.

66

THE LIFE OF DEVILS LAKE

Mission Lake in 1923, tho the total solid concentration of the latter
at this time was 32600 ppm.

Var. major has been reported from Lamoreau Lake.

-

,0

— „ 6/,

-

SI,ore

375-

250-

V

125-

\

v

\

<>'^.

■

'^ '^ ^^^^<^"

-

— i — "— r

Figure 20. Seasonal distribution of Nodularia in millimeters in East Lake.

Tolypothrix lanata. "Found in two collections growing with
other algae attached to twigs near shore."

Calothrix hraunii. "Collected but once, growing on a single
large boulder."

Calothrix fusca. Reported from Main Lake.

Calothrix parietina. "Slightly incrusted, forming brownish
patches on small stones near shore."

Gloetrichia natans. "Several large masses in one towing."

CHLOROPHYCEAE

Closterium occurs occasionally in the plancton samples from
Main Lake from July 1 to October 24,* but the species has not been
determined. Both C. dianae and its variety (arcuatum) have been
found in Lake P.

Cosniarium. Three species of this genus (formulosum, scop-
ularum and sub-costatum have been taken in Lake P.

Staurastrum sp. occurs occasionally in the plancton of Main
Lake from May 31 to Dec. 13. The dates for both Closterium and
Staurastrum are significant only in so far as indicating an extensive
seasonal distribution for both.

Mougeotia sp. "Found near shore in three collections . . .
with Tolypothrix and Anabaena. "

Pandorina morum. Reported from Main Lake.

Pandorina sp. An indeterminate Pandorina has also been re-
ported from East Lake.

*See below, under Staurastrum.

THE LIFE OF DEVILS LAKE 67

Botryococcus braunii. "Abundantly present in one towing.
Doubtfully present in two other towings. " Also in neighboring
fresh waters.

Ankistrodesmus falcatus. Found infrequently in plancton col-
lections from Main Lake.

Characium hookeri. Occurs abundantly as an epiphyte on
Crustacea and filamentous algae. Found frequently but irregularly
in the plancton samples at all seasons. Its irregularity renders it
difficult to draw any definite conclusions as to its seasonal distribu-
tion. Numerous detached cells are found in the collections but
normally it is probably always epiphytic, mainly on Crustacea.

"Apparently the first record of this species in America."

Coelastrum microporum. "Rare in towings."

DictyosphaeriiDii. This, in common with Characium and
Oocystis, is the most characteristic of the green algae in the plancton.
It readily disintegrates in preserved material so that the records are
incomplete. Especially is this true for most of the 1914 collections,
which could not be counted until 1919. No attempt has been made
therefore to plot a distribution curve for this year. The records for
1913, which are complete except for the period from June 24 to
August 1, show practically none in winter, with a rather uniform
rise from none on May 1 to a maximum about October 10, and a
return to zero by the end of December.

In 1923 Dictyosphaerium appeared on May 12 and gradually
but slowly increased up to July 23 when the collecting ended
(pl. 15).

Both ehrenbergianum & pulchellum occur in the complex.

Nephrocytium naegelii. "Occurring rather frequently in tow-
ings. ' '

Eudorina elegans. A rare species reported only from Main
Lake.

Oocystis is one of the characteristic genera, but very irregular in
its presence in the plancton samples. In 1913 a small maximum
occurred about September 1, while in 1914 there Vvas a very marked
rise on September 14, when the collections were ended (pl. 13).

Judging from the diagrams of Birge and Juday (1922) Oocystis
is rather more regular in its occurrence in Lake ]\Iendota than in
Devils Lake, but here too marked irregularities occur.

It is represented in the complex by five species, crassa, lacustris,
naeglii, pusilla and solitaria, of which the latter is the most numer-
ous.

Scencdesnms hijugatus and quadricauda typiciis. "Both species
rather common in towings, also associated with filamentous algae
along shore," and of occasional occurrence in the plancton samples.

Zoochorella conductrix. "In Stentor from towings."

68 THE LIFE OF DEVILS LAKE

Actinastrum hantzschii. Reported from Main Lake and fresh
waters.

Pediastrum angulosum and horyanum. "Both species rather
frequent in towings. " No attempt has been made to differentiate
them in the counting of the plancton collections, and their occur-
rence here is too irregular to warrant the construction of a distribu-
tion curve. Pediastrum apj)ears sporadically from April to Decem-
ber.

The following attached forms are characteristic of the shore
and Ruppia zones. Of these Cladophora is the most characteristic,
altho Entermorpha is very common in several scattered localities.

Microspora loefgrenii. The cell wall of this species may be
occasionally .006 mm. thick, which is perhaps an adaptation to its
saline environment, altho without a change in permeability mere thick-
ness would hardly be adequate to prevent plasmolytic action of the
water on the cell contents. The increase in thickness moreover only
occurs in some specimens, which is a further argument against its
adaptive character.

Ulothrix zonaia. "With Cladophora along shore."

Enteromorpha. "Although Enteromorpha is ordinarily re-
garded as a marine form, E. prolifera has been reported from sev-
eral fresh water lakes in the west." Intestinalis( ?), prolifera and
an indeterminate species occur in the complex.

Protoderma viride. "On stones along shores."

Stigeoclonium nanum. "Attached to Cladophora and associat-
ed with Enteromorpha."

Cladophora. The most characteristic littoral genus in all the
lakes of the complex. Both kuetzingiana and an indeterminate
species occur.

Rhizoclonium hieroglyph icum has been found in Lakes P and N.

Vaucheria sp. One record from Main Lake (6/24/15).

Oedogonium sp. Taken in Lake P. only.
DIATOMACEAE

The diatoms of Devils Lake have been studied by Profesor C. J
Elmore of William Jewell College, both at the lake and in collec-
tions sent to him from time to time. They are the most numerous
in species, altho not greatest in numbers of any class of plants in
the lake. Serving as food for many species of animals they play a
very important part in the biology of the lake. They include a
total of 80 species,* 46 of which are fresh water types with two
doubtful forms additional ;18 inhabit both fresh and brackish water;
1 fresh, brackish or marine ; 4 marine or brackish, while four are
brackish and four marine types only. There is one new species
(Navicula minnewaukonensis).

•Including one siibspecios (Navicula orptocephala venpta).

xj

-^^^^^

W'l.'^x

'Mh.

i /v"N\

THE IJFE OF DEVILS LAKE 69

Of the 46 species, which are typically fresh water forms, "there
is nothing in their appearance to indicate that they have been in
any way modified I)y their changed environment." That many
species of diatoms are very insensitive to changes in their environ-
ment is show^n by a comparison of the flora of Minnewankan Bay
in 1916 w^ith that of the main hike. In that year by reason of heavy
snow fall the preceding w'inter, and of heavy rains in the spring the
Mauvaise Coulee, draining a wide extent of conntry north and
west of Devils Lake, was flowing for the first time in many years.
The construction of a highway grade across the mouth of the bay,
leaving it connected wnth the main lake by a culvert, thru which
water was flow-ing at the rate of about 7750 cu. m. daily, converted
the former into a lake with about 1/3 of the salt content of the
main lake ^4464 as compared with 12920 ])pm.). Of the fourteen
species found in the bay in this year thirteen were identical with
those in the main lake, while only one (Stephanodiscus niagrae) was
new% evidently brought down by the coulee. It is not unlikely that
a more thoro study would have revealed a larger number of species
common to both places, but the data suffice to show that many
species at least are very tolerant of great changes in their chemical
environment.

The Ruppia zone is the chief source of the diatoms in tlie lake,
partly because it furnishes suitable attachment for the sessile species,
and probably because of the greater amount of dissolved food-stuffs
in this zone. They occur in considerable numbers in the pelagic
zone, however, and especially in the shore zone at certain times.
Thej^ may be carried into these zones from the Ruppia by currents,
or they may develop there independently of the latter. ProTably
both factors are involved in their presence in these zones.

Their occurrence in the plancton .samples is very erratic. That
this is due to improper preservation, or to errors in sampling and
concentrating the material is unlikely. It is possibly to be explained
by the action of currents, just mentioned, especially in the case of
the shore samples, where the diatoms may be present in large num-
bers at one time, and soon after absent or verj' rare in the same
place.

The follown'ng genera occur more or less frequently in the planc-
ton collections : Amphora, Amphiprora. Cyclotella, Cymbella, Cys-
topleura, (ryrosigma, Xavicula, Rhoicosphenia, S u r i r e 1 1 a and
Synedra, of which Cyclotella, Navicula, Surirella and Cymbella are
the commonest.

The individual distribution of all of these has been plotted
as well as that of the diatoms collectively, but only the latter chart
and that of Chaetoceros are reproduced here.* Apart from the

•Plato 16 and figs. IT, 21.

70 THE LIFE OF DEVILS LAKE

irregularity of their distribution, and their frequently greater
abundance in the shore collections, to which attention has already
been called, the charts show little except that, like other algae, they
are much more abundant in summer than in winter. That this
is due to temperature is unlikety, because in some instances at least
there is a marked increase in early spring or late autumn. Similar

Figure 21. Seasonal distribution of diatoms in individuals in East Lake.

conditions are described by Needham and Lloyd (1916, pp. 302-3)
and Marsh (1900, p. 176). In r. letter to the writer from Grand
Island, Nebraska, (Dec. 31, 1920) Professor Elmore says, "Here,
diatoms are much more abundant in the winter, when the water
is partly frozen over." Their infrequent occurrence then in the
winter collections from Devils Lake is probably due to other factors,
such as reduction in amount of light or dissolved food-stuffs.

In shore collections, especially in bays of the lake, the numbers
of diatoms, especially Navicula and Synedra, may run very high,
the maximum number recorded being found at the head of Creel
Bay, near the mouth of the sewer ditch from Devils Lake City, on
August 13, 1912, as follows: total 1292, Navicula 748, Synedra 376,
Cymbella 128, scattering 40, per cc. Other high records are from
the same place: total 292, Navicula 144, Synedra 140, scattering 8,
date?, and total 168, Navicula 152, Synedra 16; Aug. 9, 1916, and
from the mouth of Minnewaukan Bay on July 13, 1916, Navicula
116, Synedra 4. These are the only records of more than 100
diatoms per cc, but there are a few others of 90 or more.

Other species of occasional occurrence in the plancton are
0 d o n t i d i u m elongatum, Mastigloia elliptica and Carapylodiscus

THE LIFE OF DEVILS LAKE 71

clypeus, tho it is likely that all of the species in the lake are occa-
sionally driven from their natural habitat in the Ruppia zone
into the open water.

The list of species follows.

Chaetoceros elmorei. One of the characteristic phytoplanc-
tonts of the lake. It also frequently serves for the attach-
ment of other plauctonts such as Characium, Vorticella and
Cothurnia. It is of interest to note that this representative of a
marine genus should find in this brackish lake so favorable a
habitat that it now forms one of its characteristic species.

The curve of seasonal distribution is given in pi. 16 which
shows that it is more limited in its occurrence than most of the
algae, seldom appearing before June, and disappearing about
December first. It has a distinct minimum about mid-August, with
maxima about the end of July and August, while in 1914 a second
sharp minimum occurred July 20. The few records from East Lake
show a much smaller number in Main Lake (fig. 17).

Cyclotella. Three species of this genus, meneghiniana, striata
and comta occur in the Devils Lake complex. No attempt has been
made to differentiate them in the plancton counts. Cyclotella
occurs occasionally, tho rarely in winter, and rather frequently
at other seasons. In only one year (1913) is there evidence of a
rather definite maximum, which occurred in early September.
Meneghiniana has been found in Main, East and Spring Lakes
and in fresh water. Striata, which is typically a marine and
brackish water species, is reported from Main and East Lakes
and Lake P. Professor Elmore writes me that Devils Lake is
the only place where he has taken the latter species, altho it is
reported by Boyer from blue clay near Philadelphia.

Stephanodiscus carconensis is widely distributed in the complex,
being reported by Elmore from Main, Mission, Stump and East
Lakes and Lakes C and P, altho he tells me that this is the only
region in which he has found it. He further states that it is not
given by Van Heurck (1896), and I have not found it in Schon-
feldt (i913), so it is evidently a rare species. Professor Elmore
reports it as common in Devils Lake in the plancton samples.

Oyrosigma. Three species of Gyrosigma occur in Devils Lake,
of which kutzingii is common and the others rare. G. kutzingii Is
reported by Elmore from East and Mission as well as Main Lake,
and G. acuminatum from Main and Spring Lakes; while G. spen-
cerii is reported from Main Lake only in three collections. No
attempt has been made to differentiate the three species in the
plankton counts. The genus appears sporadically from April to
November, but the data are inadequate for drawing any conclu«
sions as to its seasonal distribution.

72 THE LIFE OF DEVILS LAKE

Mastigloio. Two species of Mastigloia, elliptica and smithii,
occur in the complex, of which the former is common and widely
distributed; while of the latter there is but one record (Main Lake
12/29/15), altho Professor Elmore reports it from a fresh water
lake in the vicinity. The genus appears sporadically in the plancton
collections.

Scolioplcura peisonis. In his "Diatoms of Nebraska" Elmore
(1921) gives Devils Lake as his only record for this species. It is
not mentioned by Schonfeldt (1913), while Professor Elmore
informs me that Van Heurck (1896) records it as marine. Elmore's
only record from Devils Lake is Minnewaukan Bay, 7/20/15.

Navicula. No attempt has been made to ditferentiate the
various species of Navicula in the plancton samples. The genus
occurs irregularly in the collections thruout the year. It has a
fairly well marked maximum in May-June and two other rather
indefinite maxima in July-August and October-November. Tl
frequently appears in large numbers in the shore collections*
while absent or rare in the others. This irregularity renders it diffi-
cult to draw any very definite conclusion regarding its seasonal
distribution.

Notwithstanding the irregularity of its occurrence it is, as
might be expected, the richest in species, and one of the richest
in numbers of the genera inhabiting the lake complex. The follow-
ing species have been identified by Professor Elmore : ambigua,
anglica, brebissonii, cincta, crucicula, cryptocephala, hinigarica,
iridis, lacustris, lanceolata, fulva, gastrum, major, minnewan-
konensis, oblonga, parva, pygmaea, rostrata, scandinavica, sculpta,
sphaerophora and sub-capitata.

AinplLora. Of the two species of Amphora present in the com-
plex ovalis is common and widely distributed, occurring in fresh
and alkaline waters alike; while proteus is much rarer, being
reported by Elmore only in Main and Mission Lakes.

Amphora occurs occasionally in tlie plancton collections, but
no attempt has been made to chart its seasonal distribution.

CjjniheUa. Four species of Cymbella occur in the complex
(aequalis, cistula, parva and pusilla), of which the latter is the
most common and widely distributed. Parva occurs in fresh waters
as well as in Mission Lake, while ae(|ualis is reported only from
Lake P.

The genus occurs sporadically in the plancton collections from
the first of May to the middle of November, but the data are insuf-
ficient for drawing any conclusions regarding its seasonal distri-
bution. '

— — —

*Soo p. 70.

THE LIFE OF J)EVll7S LAKE 73

Goitipkunciiia is rei>resentcd l»y Ihrce species, graeile, moiitamim
and parvulum, of which tlie two latter are the commonest, graeile
being reported only twice, 8/21/13 and 8/28/13, both from Main
Lake. Gomphonema appears but rarely in the plancton samples,
hence no attempt has been made to determine its seasonal distri-
bution.

Ehoicosphenia curvata is common in Main Lake, tho not often
taken in the plancton collections. Reported also from Lamoreau
Lake.

Amphiprora alata. Taken occasionally in the plancton collec-
tions from Main Lake, and reported once from Lamoreau Lake.

Coeconeis. Two species of Cocconeis, pedicnlus and placen-
tula, are common and widely distributed in the complex and in
neighboring- fresh waters.

AchnantJies is represented in the complex by three species,
delicatula, lanceolata and microcephala, none of which are common,
and there is a doubtful record of a fourth, hungarica.

Hoinoeocladki is represented in the complex by eight species,
augustata, amphioxys, eommutata, hungarica, lanceolata, palea, try-
blionella and vermicularis, with one uncertain record (Main Lake,
8/28/13) of another species, stagnorum. Of these hungarica is the
most common, followed, in the order named, by palea and am-
phioxys.

C amp ylodi sens clypeus is fairly common in Main Lake, tho
of infrequent occurrence in the plancton samples. Reported also
from Mission and Spring Lakes and Lakes A and P.

Surirella occurs frequently in the plancton catches, appearing
about the first of April and disappearing about the middle of
November, somewdiat earlier than most of the algae. The irregu-
larity of its occurrence renders it ditificult to draw any conclusions
relative to its seasonal distribution. Of the two species in Devils
Lake, striatula and ovalis ovata, the former is the more common.

Denticula elegans. A rare species, reported only from Main
Lake.

Fragilaria construois. Reported in one collection 8/19/20 from
Lake P.

OdoiitidiiDH elongaium. Reported from Main Lake only. A
fairly common species, but too irregular in the plancton collec-
tions to w^arrant any conclusions regarding its seasonal distribu-
tion.

Synedra. Four species of Synedra occur in the complex,
pulchella, tabulata, tenuissima and ulna, with a doubtful record
(Main Lake 8/28/13) of a fifth, amphicephala. Of these pulchella
is common, while the others occur rather rarely. No attempt has
been made to differentiate the several species in the plancton

74 THE LIFE OF DEVILS LAKE

coiiuts, which shf)W that t]ie distribution of the gcnns follows rather
closely that of Navienla. It occurs rarely in winter and has
one or two rather definite maxima in May and June. Its irregular
occurrence, especially in the shore collections, renders it impossible
to make any statements regarding maxima at other times. It some-
times appears in the shore collections in large numbers.*

Cystopleura. There are six species of Cystopleura in the com-
plex, argus, gibba, ocellata, sorex, ventricosa and zebra, of which
gibba is common and the others rare. It appears erratically in the
plancton samples and is usuall.y too infrequent to justify any
conclusions regarding its seasonal distribution. Occasionally, how-
ever, it appears in large numbers (80 per c. c. surface Creel Bay,
8/10/14).

Eunotia lunaris. There is but one record of this species for
the complex (Main Lake, 6/27/14).

Sphinctocystis Uhrilis. There is but a single record of this
species in the complex (Main Lake, 8/28/13).

FUNGI

Spores of a mold occasionally occur in the plancton collections,
but are probabl}' accidental.

SPERMATOPHYTA

The only flowering plant of importance in Devils Lake is
the ditch grass (Ruppia maritima) but in periods of high water
sedges (Cyperus sp.) may occasionally be found in shallow places
close to shore.

The Ruppia probabl.y plays a more important role in the life
of the lake than any other organism. It has already (p. 33) been
mentioned as characterizing one of the life zones of the lake,
which derives its name from this plant. It not only functions in
the interchange of chemical substances in the water, but con-
tributes a considerable part of the annual deposit of ooze on the
lake floor. It also furnishes shelter and attachment for many
smaller animals and plants. It is widely distributed, occurring in
most of the lakes of the complex, both brackish and fresh water.

It grows in from 0.6 to 2.4 m. of water, forming a zone, which
naturally varies in width at different points, depending on the con-
figuration of the lake bottom. During winter much of the leafy
parts of the plant die. Enough of the stems survive, however,
to renew the growth the following year. During the break up of
the ice in spring considerable masses of Ruppia are carried by
floating ice cakes from point to point, or thrown on shore to die

(pl. 4).

It flowers chiefly from July to September.

* See p. 70.

THE LIFJ'] OF J^EVILS LAKE 75

The sedges, which occasionally occur in the edge of the lake,
play an insignificant part in its life, and may accordingly be dis-
missed with brief mention. In 1916, however, as a result of the
abundant supply of fresh water thru the Mauvaise Coulee, Minne-
waukan Bay extended over a wide expanse of old lake floor, and
numerous sedges and rushes were found growing in its waters
(pl. 10).

PROTOZOA

The data on the Protozoa in Devils Lake are largely the result
of studies by Professor C. H. Edmondson, in 1914 and 1917. Some
data have, however, been taken from my own notes.

As a result of his studies Edmondson (1920) found that:
"The protozoan fauna of the Devils Lake complex ... in many
respects, was such as one might expect in a fresh water lake of simi-
lar depth, j^et some very pronounced differences M^ere disclosed
. . . A most noticeable feature is the apparent total absence of
numerous forms universally found in fresh water. . . .

"Experiments of a preliminarj^ character indicate that certain
protozoa having adjusted themselves to fresh water conditions are
not in all cases at least, readily adaptable to the waters of Devils
Lake."

Just how large a part the Protozoa plaj' in the bionomics of
Devils Lake cannot be estimated. They are of relatively infrequent
occurrence in the plancton catches, and then mostly attached peri-
trichs, such as Cothurnia and Vorticella. Any collection of Ruppia,
however, after standing in the laboratory for a few days, contains
large numbers of Protozoa of several species, and it is this mate-
rial which is the chief source of Protozoa in the lake. No attempt
has been made to estimate the number of the Protozoa as a whole,
nor the relative abundance of the different species, except in ix
general way.

Many of the forms are so minute as to escape thru the sand
filter in the Sedgwick-Rafter tube, and even if retained in the
preserved sample, they are so distorted as to prevent identification.
They are comparatively rare, so that the only way to obtain them
in considerable numbers for study is to allow a culture of ditch
grass and algae, or of ooze to stand in a culture .jar for several
days, when many of them develop in enormous numbers. In this
way it is possible to obtain them at any season, even in winter,
when they doubtless occur encysted in the ooze.

The vitality of some of the Protozoa is quite remarkable.
Apart from the results of Professor Edmondson 's experiments, men-
tioned above, which show the ability of some species to withstand
sudden transfer from lake water to fresh, I have observed an

76 THE LIFE OF DEVIL8 LAKE

Oxytricha which was still living:, althn inactive, after a 25-26 hour
confinement beneath a cover glass in a sealed Petri dish.

Many Protozoa occur in the lower levels of the lake and in
the ooze at the bottom, where dissolved oxygen (as shown by the
Winkler test) is absent. This is discussed elsewhere in this report.*

One hundred and six species are recorded from the lake, a list
of which follows.

Amcha proteus. Found occasionally at several points iti the
main lake. Also reported from East Lake.

Ameha radiosn. Eare. Taken in Euppia in I\Iinncwa\dvan Bay.
Also reported from Lake A.

Ameha Umax. "Associated with Euppia and algae at the
head of Creel Bay . . .and the east side of the main lake (numer-
ous) ", also taken ("numerous") in Lake A and Mission Lake
(1917).

Amelia rerrucosa. "Observed but once."

A7neba gnitula. Of fairly common occurrence in the lake.

Amoeba striata. One specimen noted in a plant infusion from
Stump Lake.

Auieha vitraca. Eare.

Difflugia pyriformis. Taken in Lake A near seepage from a
fresh water spring.

Difflugia constricta. Found near the ou.tlet of the l^evils Lake
sewer in Creel Bay, and in Lake A near seepage from a spring.

Arcella vulgaris. Found in ooze in Creel Bay, and abundant
in Lake A near seepage from a spring.

Cyphoderia ainpidla. "One specimen only has been observed."
Taken in Euppia.

Euglypha alveolata. '"Observed but once" in the overflow of
a tank of lake water. It is jiossible that this species does not occur
in the lake, its presence in the lake water overflow may have been
accidental.

Actinoplirys sol. Eare, in Eui)pia from ^linnewaukan Bay.

Cercomonas sp. "Probably Cercomonas longicauda. . . .
Observed in infusions from Stump Lake only."

Manas sp). Three difit'erent species have been found in the oo/e
from Creel Bay.

Ileteromita glohosa. " Li dredged material from Creel Bay."

Heteromita sp. "Probably . . . ovata." Taken from ooze
on rocks near the Station. Also taken in Lake A.

TrepoiHonas agilis. Common in ooze and Ruppia cultures fr"om
all parts of the lake.

Euglena viridis. Occurs occasionally at various points in the
main lake. Also taken in Lake A.

"See pp. 100-7.

THE LIFE OF DEVILS LAKE 77

Euglena deses. Common and of widespread occurrence in ooze
and Ruppia, in the main lake. Also taken in Mission and East
Lakes and Lake A.

Phacus purum. Common, and widely distributed in the main
lake. Also numerous in Lake A.

Eutreptia viridis. "From the surface among Ruppia" in
Lake A.

Astasia tricophora. Common in Ruppia and algae. Wide-
spread in the main lake. Reported also in Lake A.

Petalomonas mediocanellata. Occasional in ooze in the main
lake, and the surface of Lake A.

Petalomonas sp. "From the ooze of Creel Bay."
Heteronema acus. "From Six-mile Bay and from the ooze of
Creel Bay."

Anisonema grande. "Among Ruppia and algae at the head
of Creel Bay."

Notosolenus sp. Occasional. Taken in Main and Stump Lakes.
Tetraselmis cordiformis. "Taken from Stump Lake only."
Polytoma uvella. Occasional in the main lake.
CMamydomonas pulviscidus. "Taken from the head of Creel
Bay."

Glenodinium pidv is cuius. "Taken from the surface and from
the ooze at the bottom of Creel Bay."

Holophrya ovum. "Among algae from Creel Bay."
Urotricha lahiata. Common in the main lake.
Prorodon teres. A fairly common and widely distributed spe-
cies in the main lake in Ruppia and algae. Also in Lake A.

Prorodon edentatus. Occasional in Ruppia. Taken in Minne-
waukan Bay and Lake A.

Enchelys sp. Noted by Edmondson in ooze of the main lake
and in overflow from a tank of lake water. What was probably
the same species was observed by me in the shells of rotifers in a
shore collection, July 20, 1914.

Spathidium spatula. "Among algae from the head of Creel
Bay."

Two indeterminate species of Spathidium were taken by Prof.
Edmondson. one in a Ruppia infusion from the head of Creel Bay
and one from "ooze of the main lake."

Chaenia teres. "Among algae from the head of Creel Bay."
Mesodinium pulex. One of the most common Protozoa in the
lake, found in infusions of Ruppia, in surface tows and in the
ooze of the main lake. Undoubtedly occurs in other lakes of the
complex, but has not yet been observed other than in the main lake..
Didinium nasutum. Occasional and of wide distribution in
Main Lake.

78 THE LIFE OF DEVILS LAKE

Lacrymaria olor. "Among Ruppia in Creel Bay."

Lacrymaria truncata. "Among Ruppia from the north end ol
the main lake."

Lacrymaria lagenula. "In ooze from the main lake."

Lacrymaria cohnii. "In an infusion from Stump Lake."

Lionotus fasciola. Common and widely distributed in main
lake. "Also taken from Stump Lake", and Lake A.

Lionotus sp. "Among algae from Creel Bay."

Amphileptus meleagris. "From algae at the head of Creel
Bay." Also in Stump Lake.

Nassula rub ens. In overflow of lake water from tank.

Nassula ornata. Taken from Lake N only.

Chilodon cucullulus. "In infusions of algae from Creel Bay"
and Lake A.

Chilodon caudatus. "Among Ruppia from Minnewaukan Bay."

Aegyria pusilla. "Among algae near the Station."

Glaucoma scintillans. In an infusion of algae from the west
end of the lake.

Glaucoma margaritaceum,. "Very abundant" and widely dis-
tributed in Main Lake both at the surface and in the ooze. Reported
also from Stump Lake.

Leucophrys patula. "One specimen only obseryed from . . .
the Main Lake."

Frontonia leucas. Common and widely distributed in Main
Lake, and reported from East Lake.

Loxocephalus gramdosus. "Taken only in the ooze of Lake A
near the inseepage of fresh water."

Uronema marinum. "One of the most common species in the
lake. Abundant everywhere both at the surface and in the ooze."

Uronema is also reported by Edmondson among algae in Stump
Lake. Presumably this species.

Colpidium putrinum. "F^rom algae at the east side of Creel
Bay."

Tillina saprophila. "Taken only in the overflow of lake water
from fish-tank."

Parainaecium trichinum. Occasional.

Paramaecium caudatum. "Taken from Lake A near the inseep-
age of fresh water."

Cyclidium glaucoma. Abundant everywhere in Main Lake at
the surface and in the ooze.

Cyclidium litomesum. "Numerous in infusions from the head
of Creel Bay and in the ooze from the main lake."

Pleuronema chrysalis. "Observed in infusions from Stump
Lake only."

Metopus sigmoides. "Common in dredged material from Min-

THE LIFE OF DEVILS LAKE 79

newaukan Ba.y, Creel Bay and the main lake. Abundant in East
Lake."

Spirostomum amhiguum. "Observed in dredged material from
Creel Bay."

Steyitor sp. Reported by Moore (1917) from Main Lake.

Halteria grandinella. Common and widely distributed in Main
Lake, both in ooze and infusions of Ruppia and algae.

Uroleptus agilis. Occasional in ooze in ]\Iain Lake.

Uroleptus rattulus. Occasional in Ruppia in Main Lake.

Oxytricha. This genus is probably the most characteristic
genus in the lake, both in respect to variety of species and number
of individuals, almost every infusion of Ruppia containing them.
Four species have been determined, but it is Professor Edmondson's
opinion, that further study would reveal others. These four are
fallax, pellionella, parvistyla, and bifaria, of which pellionella ia
the commonest. The latter is reported from Lake A also. The
others from Main Lake onl}-. Bifaria? has been seen with con-
tained diatoms (Navicula?).

Histrio erythysticus. In Ruppia in Creel Bay.

Stylonichia notophora. "With algae from Creel Bay."

Holosticha vernaUs. "Among Ruppia from the main lake."

Pleurotricha lanceolata. "Taken at the head of Creel Bay."

T achy soma parvistyla. "Observed in infusions from Stump
Lake only."

Euplotes charon. This species and E. patella are among the
most abundant Protozoa in the lake. Charon is reported from East
Lake also, while patella occurs in Stump, East Lake and Lake A.

Aspidisca costata. Abundant in Ruppia infusions thruout the
main lake.

The genera Vorticella and Cothurnia are the only Protozoa
which occur with any frequency in the plancton. This may be due
to the fact that they are carried out of their natural habitat in
the Ruppia by animals or plants to which they are attached. No
attempt at specific determination has been made in the plancton
counts. Vorticella occurs frequently at all seasons while Cothurnia
appears more rarely, but it is probably perennial also. My records
run from June 22* to December 30.

Six species of Vorticella, two of them indeterminate, have been
observed by Professor Edmondson as follows : telescopica, convallaria,
octavo, microstoma and two sp. Of these convallaria is reported from
Lake A. also.

Gerda annulata. "Among algae and Ruppia from the north
end of the main lake."

EpistyUs plicatilis. "From . . . Creel Bay."

•June 21 in East Lake.

80 THE LIFE OF DEVILS LAKE

Epistylis branchiophila. "Among algae near the head of
Creel Bay."

Carchesium epistylidis. "Among algae from Creel Bay."

Zoothamnium alterans. "Among Ruppia and algae from Stump
Lake."

Zoothamnion sp. Fairly common and widely distributed.
Among Ruppia in Main, Stump and East Lakes.

Vaginicola crystalUna. "Numerous among algae from East
Lake, also taken from Stump Lake and from the north end of
the main lake."

Cothurnia imherhis. "Commonly attached to floating diatoms,
from dredged material and also among Ruppia in Creel Bay. Also
taken from Stump Lake."

Cothurnia curva. "Among Ruppia at the north end of the
main lake."

Podophyrya Uhei-a. "Numerous at east side of the main lake."

Podophyrya sp. "Attached to algae from the main lake."

Sphaerophrya ynagna. "From Stump Lake and the east side
of the main lake."

Acineta sp. "From floating material in the main lake and
also among algae from Stump Lake."

Acineta sp. "Attached to algae from Stump Lake. Commonly
feeding on XJronema. ' '

Species incertae sedis

Triiuastigidae gen. "Numerous among Ruppia from Creel
Bay."

Fam. and gen. An undetermined ciliate was found by Pro-
fessor Edmondson at the "surface of the main lake and among
Ruppia and algae. ' '

Genus sp. A small green flagellate sometimes occurs in large
numbers attached to Pedaliou, and probably to other rotifers. It
is flexible but not ameboid ; in form ovoid, stipitate. It bears a
single anterior flagellum, which is vibratile thruout its length, and
which about equals the body jn length, and is inserted near the
bottom of a depression, one side of which projects forward as an
overhanging lip, and which in depth about equals one third the
length of the body.

There is a single vacuole situated posteriorly. Staining with
magenta brings out a number of dark granules, but 1 have not been
able to differentiate any definite nucleus. Nor have I determined
the form of the chromatophore, the entire cell appearing filled with
chlorophyl. There is no eye spot. Measurements of three speci-
mens (exclusive of the stalk) follow: .0049x.0033, .0057x.0024, .0057x
.0024 mm.

---^f

THE IJFK OF DKA'IlvS I.AKK 81

PLAN ARIA

But a single species of this group (Gyratrix hermaphroditus)
has been found in the Devils Lake complex, an occasional specimen
being taken in collections from Ruppia in the main lake.

XEI\IATODA

The nematodes of the complex are numerous in individuals
and undoubtedly in species, but as yet complete identifications are
not available. Dr. N. A. Cobb of the U. S. Department of Agri-
culture has kindly given me the following list, based on fresh mate-
rial which reached him, for the most part, in poor condition.

Monohystera sp. "Very commou."

Monohystera sp. "Not common."

Diplogaster sp. "Common."

Achromadora sp. "Rare."

Ironus sp. "Not common."

Plectus sp. "Not common."

Dorylaimus sp. "Rare."

Dorylaimus sp. "Rare."

Cephalobus sp.

Chromadora americaiia n. sp. "Rather common."

The nematodes occur mainl}^ in the Ruppia zone and in the
ooze, but are occasionally present in the plancton catches. They
occur at all seasons of the year, both in winter and summer, but
no attempt has been made to determine their relative abundance
at different times.

ROTIFERA

Both in respect to numbers and variety the rotifers are the
most important animals in the Devils Lake plancton. Thirty-one
species and one sub-species occur, two of them, Brachionus satanicus
and Pedalia fennica, at certain times in great numbers.

The attached forms naturally are found in the Ruppia zone,
which also shelters a number of the rarer, free swimming forms.
All of the latter are probably of widespread distribution thruout
the lake, but the rarer forms occur so infrequently that but little
can be said regarding their distribution, either spatial or temporal.

Reproduction is almost entirely parthenogenetic, males of any
species having been seen in very few instances. The difficulty of
recognizing males in preserved material may, however, partly
account for this.

The seasonal distribution of the rotifers as a whole is shown
ii) pi. 17 and figure 22, anent which the same remarks apply as
those made anent that of the Crustacea.* The distribution of those
species which occur in the collections in numbers large enough to

♦See p. 87.

82 THE LIFE OF DEVILS LAKE

Figure 22. Seasonal distribution of rotifers in East Lake.

be significant is shown in plates 18 to 21 and will be discussed
briefly in the annotated list.

Absent, or practically so, in winter, the rotifers develop rapidly
in early summer and may reach their maximum for the entire year
by the first of June. There are usually two maxima, one in June
or July and another in August, but in 1913 there are three sueh
periods, the last one appearing about October 6, which is unusual.
The curves show very sharp breaks, indicating rapid development
and equally rapid disappearance.

While the rotifers of Devils Lake are mostly fresh M^ater types,
like rotifers in general elsewhere, there are a few which are marine
as well as fresh water. These are Colurella adriatica, colura, Brach-
ionus plicatilis, Notholca striata thalassica and Pedalia fennica.
Collotheca cornuta, Brachionus calyciflorus pala, Polyarthra trigla,
Keratella cochlearis, and possibly Asplanchna silvestrii* occur in
both fresh and brackish water, while Brachionus satanicus and B.
pterodinoides, both of them new species, belong in the latter class,
judging from their distribution in the complex and its adjoining
waters.

The widespread distribution of so many species in both the
brackish and fresh water lakes of the complex indicates the
adaptability of the group to widely varying environments; the
apparent restriction of some to one or a few lakes, being due
possibly to the incompleteness of the investigation. The rotifers
show marked light reactions, a discussion of which has already
been given.**

Rotifers occur occasionally in oxygen free water.***

The annotated list follows:

Cephalodella catellina. Occurs occasionally in the main lake,
where it is of widespread distribution. Reported also from East
Lake.

Diaschiza sterea. Widely distributed in main lake. Reported
also from East and Mission Lakes.

♦The character of the Chilean lake from which this species was first taljen Is
not known.

**See p, 46.
•♦♦See p. 107.

82

:/

be si
briefl

in ea
by tl:
or Jt
perio
The .
and

like 1

as w€

ionus

CoUo

Kera

both

ptero

judgi

watei

brack
adapl
appai
possil
show
been
]

(

when

Lake.

J

also :

Dot ki

THE LIFE OF DEVIDS LAKE 83

D. megalocephala. One record only, a single specimen in a
collection from Creel Bay (8/31/14).

Brachionns plicatilis. A very common and widely distributed
species but not usually as numerous as either Pcdalia or Brachionns
satanicus; occurring in Main, East, Mission, Lamoreau, Spring and
Stump Lakes, Lakes A, C, and 0, and in certain fresh water
lakes of the complex as well. Owing to the close similarity between
this species and its variety spatiosus no attempt has been made tr>
distinguish them in the plancton counts. Typical spatiosus is
sufficiently distinct, but intergradations render it impossible to
determine where every specimen belongs. Accordingly I have?
included both the species and variety in the chart (pi. 18), which
illustrates the seasonal distribution. This shows its presence thru
a comparatively short period. In 1911 a few were present up to the
end of September. In 1912 the records show their practical absence
prior to June 26, one or two records being all that I have before
this date. A few were present later in the season, but the records
are too scanty to represent as a curve. In 1913 they were absent
prior to June 24, when the spring collections ended, occurring in
small numbers from August 1 to September 6. The 1914 records
cover a longer period (May 15 to September 14) but even here
they were present in considerable numbers during July only. In
1918 on August 7 a rough comparison of the numbers of various
species of rotifers taken in a tow showed 63 Brachionns p. spa-
tiosus and 1 plicatilis? to 36 Pedalia and 3 Brachionns satanicus.
August 12, 1918, a similar comparison showed 103 B. p. spatiosus,
47 Pedalion, 2 B. satanicus and 1 Colurella? With remarkable
suddenness, however, the plicatilis disappeared, for on August 21,
I found but one specimen, as compared with 86 Pedalia and 9 B.
satanicus.* Usually two maxima with an intervening minimum
occur in a season. In 1913 there is a very definite minimum on
August 11 folloAved by a definite maximum on the 20th and pre-
ceded by a maximum sometime in July or early August. In 1914
there was a distinct maximum on July 20, and another, less well
defined, between August 17 and 26, with a verv^ distinct minimum
on August 2. In 1911 the records are not adequate for inter-
pretation. The records for East Lake are too scattering to have
much value.

Brachioniis plicatilis spatiosus. The spatial distribution of the
variety described by Rousselet (1912) is similar to that of the
parent species, altho there is no record from Lake P. from whicfi
the parent species is recorded. In the main lake, where the most
extensive collecting has been done, spatiosus is much commoner than
plicatilis. In his description of this form Rousselet says (p. 374)

*See p. 45.

84 THE LIFE OF DEVILS LAKE

"I was ill doubt at first whether this form should not be called a
variet}^ of B. :\Iulleri with which structurally it is undoubtedly more
nearly allied than with any other species; but after comparing
it with the mounted specimens in my collection from various locali-
ties I have decided to give it specific rank on account of its strik-
ing shape." My reason for classing it here as a variety of B. plica-
tills are the occasional intermediate forms which occur in my
material as already noted.*

Brachionus capsuliflorm quddridodatns. Of occasional occur-
rence and widespread in the main lake, and reported in one collec-
tion each for East Lake (1918) Stump Lake, Spring Lake, and
Lakes A., C, 0., and P. It occurs also in several fresh water
lakes of the complex.

Brachionus urceolaris. Taken near the head of Minnewaukan
Bay in 1916. Also in fresh water lakes in the vicinity. As already
noted the former was nearly fresh at this time.

Brachionus pterodinoides. This species, described by Rous-
selet (1918), is fairly common and widely distributed in the com-
plex and in fresh waters. While common, it appears only sporadic-
ally in the planeton collections, and hence there are not sufficient
data for plotting its seasonal distribution. It occurs first in early
May, disappearing again early in November.

Brachionus calyciflorus pala. Reported from Minnewaukan
Bay.

Brachionus dolahratus. Reported from Minnewaukan Bay
(June 6, 1921) and Spring Lake (July, 1920).

Brachionus sataniciis (pi. 19, 20). This species, described by
Rousselet (1911), is one of the most abundant and characteristic
species in the lake. The curve of its seasonal distribution in j\Iain
Lake shows that two cycles, one major and one minor, occurred prior
to June 26 in 1912. Three of them are evident in 1913, with
maxima about June 11, August 1, and September 1, and in 1914
there are also three, with maxima between June 10 and 24 and on
July 2 and September 2.

It is not only one of the commonest, but naturally also, one
of the most widely distributed species in the complex, occurring in
all the lakes examined except N and in fresh water as well. Lake
N is merely a temporary pool which possibly explains the apparent
absence of B. satanicus there.

In a letter to the writer (July 17, 1912) Mr. Rousselet says
"I have found two distinct seasonal varieties of this species, evi-
dently winter forms, which during the warmer weather of July will
produce the normal type . . . These ... are not immature
. . . forms, for they carry eggs. These two seasonal varieties

♦See p. S3.

i

V;c

^^i^"=^^.,

''^-

-^^r^.

A I r

A p

Plate 19. Seasonal distribution of Bracliic

THP] JJFE OP l^EVTTvR LAKE 85

of B. satanicus present some remarkable features for not only are
the posterior spines reduced and curved inwards, but the foot open-
ing is situated on the dorsal side, which does not occur in any
other known Braehionus." Further investij?ation, however, has
shown the presence of both short and lonj;-spined forms in summer
(Aujjust 22, 1913), so that it is doubtful if these varieties are sea-
sonal ones.

Platyias quadricornis. Reported from Main. Stump, and Spring
Lakes and fresh waters. It occurs in all of the zones of the lake
except the ooze.

Keratella cochleari.';. Present in Spring Lake — also in fresh
waters.

Keratella qnadrnta. Reported from Stump Lake and fresh water
lakes.

Notholca striata thalasslca. There are a number of records
from Main Lake, where it is widely distributed but not common.
Reported also from Mission Lake (1914), Lamoreau Lake, and Lakes
C and P and in fresh waters.

Mytilina ventralis brevispiita. One record only for Main Lake.
Recorded also in Spring Lake and in fresh water lakes of the
complex.

Lecane inermis. Of occasional occurrence and widely dis-
tributed in Main Lake, and reported from East Lake and Mission
Lake and Lake C.

Lecane tuna. Reported in two collections from Main Lake.
Also in Stump, East and Spring Lakes and fresh waters.

Monostyla cornuta. Recorded only from Spring Lake and fresh
waters.

Monostyla bulla. Present in a surface tow from Creel Bay
(8/13/12). Also in Stump and Spring Lakes and fresh waters.

Monostyla lunaris. Two specimens in a surface tow from Ft.
Totten Bay (10/2/17) is the only record for Main Lake. Also in
Lake N and fresh water.

Monostyla quadridentata. Reported from Minnewaukan Bay
(Main Lake) Spring Lake and fresh waters.

Lepadella patella. Present in a few towings from Main Lake.
Also in fresh waters.

Squatinella mutica. One record, from Ruppia, Creel Bay,
(8/26/14).

Cohtrella adriatica. Occurs occasionally and is widely dis-
tributed in Main Lake. Present also in Mission Lake (1914-15) and
in Lakes A and P.

Colurdla colura. Occurs occasionally and is widely distributed
in Main Lake. Reported also for Mission Lake (1914).

86 THE LIFE OF DEVILS LAKE

Filinia longiseta. In September and early October 1909 this
species was of occasional occurrence in the planeton collections from
the main lake. Since with increase in salt concentration of the
water above 1% it appears to be absent there, but was present in
two collections from Minnewankan Bay in 1916, when this Bay was
reduced about 2/3 in salinity and greatly increased in area, due
to the influx from the Mauvaise Coulee, it appears to be unable to
adapt itself to as great concentration of salts as several other species
in the complex.

It has also been reported in Spring Lake. It is in general very
rare in the complex but in the Minnewaukan Bay samples of 1916
Mr. Harry K. Harring of the United States National Museum re-
ports it as "abundant" and "common" in two collections. My last;
record from the Main Lake is May 17, 1912.

Pedalia fennica (pi. 20, 21). Abundant and widely distributed
in the complex and in fresh waters. It is evidently readily adapta-
ble to a great variety of habitats.

Its distribution curve for Main Ijake shows that in 1911 three
rather poorly defined maxima occurred, one about mid-July, one in
early and one in late August. In 1912 a distinct maximum oc-
curred about May 25. The records for this year, altho incom-
plete, suggest also another smaller one about the end of August.
In 1913 a small maximum occurred the last of May, and three
well marked maxima from August to October (8/12/20, 9/3/10,
and 9/29-10/6), altho the latter consists chiefly of a great develop-
ment at shore on the last date, which may have been due to a
swarm and not indicate a time maximum at all.

Occurs rarely in winter, but is in active reproduction very early
in spring, gravid females having been taken as early as April 8 (9)
when the lake was still frozen.

In a letter to the writer of July 17, 1912, Mr. Kousselet says,
"It is remarkable that Pedalia fennica should be found in ac-
tive reproductive stage - - so early in the year (May 17)
. . . in Europe Pedalion mirum is a summer form and is hardly
ever seen before June."

Asplanchna silvestrii. During summer Asplanchna occurs
rather frequently in the planeton samples and occasionally in con-
siderable numbers (100 pr. 1. 0.6 m. depth, 6/24/12). It is too
erratic however to justify any attempt at an analysis of its seasonal
distribution. It usually occurs in numbers of 50 pr. 1. or less. It
appears rather suddenly in considerable numbers, about the end
of June, disappearing again in September or early October.
My earliest record for Main Lake is 4/20/12 and the latest
10/15/10. Present also in Mission Lake (1914), in Spring Lake, in
Lakes A, and C. and in fresh water lakes of the complex. Numerous
in a collection from Creel Bay (6/24/10) in which several males were

750

500

250

Seasonal distribution of Brachi

86

spe(
the
wat
two
red
to 1
ada
in 1

ran
Mr.

por
reci

in
ble

rat'
ear

GUI

pie
In
wel
an(
me
sw

in
wh

''I

tiv

evi

ra
sic
er:
dii
ap
of
M
10
L{
in

THE LIFE OF DEVILS LAKE 87

present. Prior to its discovery in the Devils Lake complex it had
only been reported from Chile, which is doubtless due to incomplete
study of our waters.

Testudinella patina. One record only for Main Lake (10/18/14),
Reported also from Spring Lake, and there are several records from
fresh water lakes of the region.

Ptygura sp. Two specimens in a collection from Mission Lake
(8/18/14).

Collotheca cornuta. Infrequent. Reported only from Main
Lake.

GASTROTRICHA

Chaetonotus maximns. Occasionally noted in collections from
the Ruppia zone in Main Lake.

CRUSTACEA
In plate 17 is shoAvn the annual records of the Crustacea for
Main Lake, from 2/6. 1911 to 9/14, 1914 and 12/3, 1922 to
7/23, 1923, and in figure 23 those for East Lake for 1913-14.
There are gaps in the records for 1912 (6/25-8/6. 8/28-10/7), 1913
(6/24-8/12) and 1923 (1/27-3/11). The curves represent the aver-
age of all collections from different depths on a given day or for a
period of successive days. The shore collections have not, however,

Figure 23. Seasonal distribution of Crustacea in East Lake.

been included until 9/3/13 because of their scarcity prior thereto.
The number of collections have varied from time to time, so that
the probable accuracy of the record is not the same thruout. Es-
pecially is this true of 1911, in which year comparatively few col-
lections were made. In spite of these defects the records are fairly
consistent thruout, except for 7/13/23 when great numbers of
nauplii appeared, averaging over 1000 per 1., to disappear again as
quickly as they came. Furthermore, the 1923 records in general
are very much higher than for former years. What the explana-
tion for this is I cannot say. There was a marked development of
Coelosphaerium about the end of June in this year, but not con-
spicuously more than in 1914, while the diatoms and other algae
showed no unusual development. Lyngbya, however, was greatly
reduced in 1923 as compared with former years. Birge (1897) con-
siders an abundance of this genus prejudicial to the Crustacea
since the latter do not feed on it. I doubt very much whether this
hypothesis holds good for Devils Lake because the filaments of the

88 THE LIFE OF DEV1D8 LAKE

former are so extremely delicate that they eonld scarcely be shunned
by the Crustacea because of their inability to feed upon them, as
Birge believes; altho, as he suggests, there may be factors other
than size which cause them to avoid it. He undoubtedly refers to
another species (L. birgei?). with filaments about 20 ce wide,
from that occurring in Devils Lake ( L. contorta), which has fila-
ments about 1.5 ec in width.

As to whether or not the Crustacea do feed on Lyngbya, I
cannot sa.y, as I have paid little attention to their feeding habits.

The only other obvious difference between conditions in 1911-14
and 1923 is the increase in concentration of the lake water, which,
a priori, one would expect to be inimical to fresh water animals,
rather than the reverse. Differences in plankton abundance in dif-
ferent years are of frequent occurrence elsewhere and have been
previously discussed in this paper.* The curves in general, show
two maxima, one in May-June and another larger one in August-
September.

The sex ratio in the Crustacea presents some interesting fea-
tures. In Diaptomus, during the summer,** when reproduction is
most active, the females outnumber the males in the ratio of 662:118,
while the ratio for the remainder of the year is 905 :942. In Cyclops
the summer ratio is 167:112, and for the balance of the year 32:10.
While the Tuimbers of the latter ratio are too small to have much
significance, there is evidently here not the same difference for dif-
ferent seasons as occurs in Diaptomus. When Moina first appears
it is represented almost exclusively by females, males usually ap-
pearing about the end of August. In 1911 and 12 the numbers of
Moina are two small to have much significance. In 1913 males
first appeared on August 27, from which time until the last ap-
pearance of the genus on October 26 the ratio was 131 females to 66
males. In 1914, out of a total of 59 individuals taken from August
2 to September 14, no males were observed.

The above data are based on collections from Main Lake only.
and, of course, on sexually mature specimens.

Kofoid (1908) finds similarly large variations in the sex ratio
of various species of Entomostraca of the Illinois River.

Diaptomus. There are several species of this genus in the
complex, but most of them are of only occasional occurrence, D.
sicilis being by far the most common. No attempt has been made
to differentiate the various species in the plankton counts
(pi. 22 and fig 24).

Diaptomus is one of the most characteristic forms of the plane-
ton. It occurs in considerable numbers thruout the year, and is,

*Soe page 44.

**The data for the "summer" poriods are as follows: 1911, 7/7-fl/ll ; 1913,
8/21-9/.'if>; 1914. 6/8-9'/14 (end of collecting period).

.,-1

^

f'^afeyi^gfe,.^

D/;:i-3/OV^ Hrx-», ^r^-

'h v,l'

1

^ I

'/^A

THE LIFE OF DEVILS LAKE 89

in fact, the only species, either animal or plant, which is not wholly
absent from, or very rare in the plancton catches in winter time.
The other important genera, Cyclops and Moina, may, at times
outnumber it, but are on the whole of secondary importance.

Figure 24. Seasonal distribution of Diaptomus in East Lake.

Its distribution curve shows rather poorly marked maxima and
minima. In general, there is a minimum in April-May, succeeded
by a maximum in May- June, a minimum in July-August, a maximum
in August-September and a minimum in September, with one or
more less well defined waves in October-November. Its abundance
in fall and winter is in general greater than its minimum number
in summer, therefore temperature is not the controlling factor in
its distribution. Its w'aves alternate in a general way with those of
the nauplii (pi. 20 and 23), as might be expected, tho in midsum-
mer, when both Diaptomus and Cyclops show maxima, this relation
is not very clear.

Diaptomus is interesting as the 'only genus, either plant or
animal, which is common in winter. Whil? its reproduction de-
creases materially at this season, immature individuals, nauplii
(probably of this species) pregnant females and eggs occur occasion-
ally beneath the ice.

From the occurrence of a winter maximum of D. graciloides in
lake Plon and a summer maximum of the same species in Dobers-
dorfer lake Apstein (1896) concludes that temperature is ineffective
in influencing the development of Diaptomus. and Marsh (1896-97)
reaches a similar conclusion. Birge, (1897) on the contrary, con-
siders temperature the chief factor determining the seasonal dis-
tribution of Diaptomus (oregonensis )in Lake Mendota. Its maxima
here occur in the warmer months (May-September) with a great
reduction in winter. He states that "it is the first of the perennial
Crustacea to slacken its reproductive activity in the autumn and
this occurs when food is at its maximum. I can attribute this check
only to the fall in temjierature. Indeed, my observations show thai
the reproductive activity of D. oregonensis is more promptly check-
ed by the decline of temperature than is that of any other of the
perennial species." (p. 326).

]\Iy own observations on Devils Lake indicate that the develop-
ment of Diaptomus is in inverse ratio to the temperature because of

90 THE LIFE OF DEVILS LAKE

its greater abundance in winter and early spring than in mid-
summer. What the factor is which determines the relatively large
number present at the former season is impossible to say. After
the lake becomes ice-bound there is almost a total absence of algae
so that food is very scarce. In Lake Mendota, on the contrary,
Birge (1. c. page 320) states that "the food supply is ample." In
Devils Lake I believe that the winter population consists of indi-
viduals which have developed in the preceding summer or fall ; that,
due to decrease of temperature and food, there is no active repro-
duction in winter, but this, on the other hand, reduces metabolism
and lowers the death rate and that consequently the numbers remain
fairly stationary. The diminution in spring, which in 1923 occur-
red before the lake opened, is probably due to the death of many
of the individuals which have lived thru the winter, possibly due to
lack of food. Rome of these survive, however, and give rise to a
large number of nauplii and adults in May or June.

This genus is frequently colored a deep red, due, apparently,
to the presence of a highly colored oil derived from its food, prob-
ably diatoms. This was especially noticeable on one occasion
(7/13/16) when the water above the grade across Minnewaukan Bay
was literally colored red by great masses of Diaptomus driven on
shore by the wind. This material may be present in the nauplii
even before hatching.

Diaptomus sicilis elsewhere is a cold water species, which might
explain its greater abundance in winter than in mid-summer in
Devils Lake, were it not for the August-September maxima of 1913
and 1914. In 1913 this occurred between 9/1 and 9/8 a period of
high and rising temperature, culminating in a surface temperature
of 23 C° at 1 p. m. on 9/6. In 1914, on the contrary, it occurred
a week earlier (8/24-8/31) when the temperature was falling*
(17° at the surface at 6 p. m. 8/26). It is impossible therefore,
from my experience to find any close correlation between tempera-
ture and the seasonal distribution of this species. There is probably
no one factor which determines it, as Robert (1. c.) has pointed out.

Diaptomus sicilis is a widely distributed form, adapting
itself readily to a great variety of environments. Hence it is
not strange that it should occur even in so brackish an environ-
ment as Devils Lake. It occurs more or less commonly in all of the
bodies of the Devils-Stump Lake complex and is ubiquitous in its
distribution, being present in all parts of the lake, with the excep-
tion of the ooze.

Dioptomvs shoshove has been reported in one collection
(7/27/20) from Main Lake and from Lake P and Spring Lake, and
a neighboring pond, not included in the complex.

*See pi. 22.

THE LIFE OF DEVIDS LAKE 91

DiaptouiKs leptopus piscinae. Reported in one collection from
Lake A (9/23/17).

Diaptomns siciloides. This species has been recorded from
Main and Stump Lakes and Lake A ; in Stump Lake since 1922
only. Prior to this time D. sicilis was common as elsewhere in the
complex, but appears to have given place to siciloides at this time.
That this is a case of seasonal succession is unlikely, since the
1922 and '23 collections were made at times (8/25/22 and
7/21/23) when sicilis was formerly common. In 1922-23 siciloides
was apparently" as common as sicilis had been formerly.

Cyclops viridis americanus, (pi. 20 and 22). Cyclops is gener-
ally present in much smaller numbers than Diaptomns, and, with
the exception of the shore records, which are erratic, for reasons
already noted, shows little evidence of waves of abundance thru
the season. It is almost wholh'- absent in winter, appears first in the
collections in April or May, the time of its appearance varying in dif-
ferent years, and disappears again in November. In 1914 it
appeared much later than usual, the earliest record being May 30.
The reason for this is not evident. There was no corresponding
scarcity of other zooplancton this spring and the temperature was
similar to that of preceding years. Cyclops is seldom abundant
and the number present in the spring months is always so few as
to have little significance. Their total absence in the early spring of
1914 is nevertheless notable.

It is difficult to locate the time of maximum production of
Cyclops, because of its general scarcity and the influence of the
erratic shore collections. The curves are fairly regular from May
to November, with the crest about August 1. In Lake Mendota
Birge (1. c.) found Cyclops brevispinosus reproducing under the ice.
This I have not observed in C. v. americanus in Devils Lake. He
traces the decline of the former from its spring maximum to lack
of food. In Devils Lake there is no evident relation between food
supply and abundance of Crustacea, and I suggest as one possible
factor in determining the summer maxima and minima, a brief life
cycle accompanying rapid metabolism, which in turn is depcn(!'ent
on temperature.*

It is similar in its spatial distribution to Diaptomns.

Cyclops leuckarti. I have one record of this species from Lake
A, 9/23/17.

Cyclops serrulatus. There is only one record for this .species
in the complex — Main Lake, 10/2/17. It has also been taken in a
neighboring fresh water coulee.

Moina macrocopa (pi. 20 and fig. 25). Appears in the plancton
samples about August 1, rises to a maximum about four weeks later,

*See pp. 45 and 90.

92

THE LIFE OB^ DEVILS LAKE

disappearing again in October ; altho I have a single record for
November 4, 1909, and one from a towing on June 24, 1914. This
was a female with three nearly developed young in her brood pouch.
Moina was also numerous in Minnewaukan Bay on July 20, 1916.
It forms a relatively small part of the plancton, tho occurring at
times in dense swarms in the warm water close to shore.** Gen-
erally distributed throughout the lake with the exception of the
ooze, where however the ephippia occur in winter. Reported also
for Stump, East, Lamoreau and Mission Lakes. In Lake A an
indeterminate Moina has been reported which is probably this
species.

f

30Q _

0 metres

200 _

i

£■1 "
Shore

fOO _

=^

/tx

.'^

5^:^

V

^ , — „_

r 1 1

Till

T

t 1

"n 1 "\

r^^^P^

Figure 25. Seasonal distribution of Moina in Main Lake.

Daphnia })iar/mi. Eeported from ]\Iinnewaukan Bay in 1916,
probably due to the fresh water supply which the latter received
in that year. It occurs also in Spring Lake and fresh waters.

Daphnia psittaeea. Occurs infrequently in widely separated
lakes. One record each from Main, Stump and Spring Lakes.

Daphnia longispina varf This species has been found occasional-
ly in Main Lake, but the variety has not been determined.

Daphnia pidex? An immature specimen was taken in Main
Lake (10/2/17). It occurs also in fresh waters in the vicinity.

Daphnia sp. Daphnia occurs in Lamoreau Lake and Lake 0,
but the species are indeterminate, as only ephippia have been found.
One male, sp. ?, has also been reported for East Lake (1917).

Ceriodaph'nia pulchella, l^i-ynocephalu^ vetulus, Alona rectan-
(jttla. Chijdurus sphaericits. There is one record only for each of
the above species — Main Lake, 10/2/17. The last species, however,
has been taken in neighboring fresh waters.

Bosniina longirostris. One record from Lamoreau Lak(\ and
one from a fresh water lake in the neighborhood.

**As many as 70,000 p<n- 1. See Moliprg (191S).

THE TJFE OF DEVILS LAKE 93

Diaphanosoma hrachiuruin. I'rcseiil, in East and Mission Lakes
in 1911, and one record from Main Lake 10/2/17.

Marshia albuquerquensis. Of occasional occurrence. !My rec-
ords show its presence in Main, East, Mission, A and 0 Lakes,
as well as in fresh water. Present in Main Lake occasionally froni
July 20 to Sept. 6.

There are but three previous records of this species, that of
Herrick (1895), Dodds (1920) and Willey (1923a).* Doubtless
the examination of other brackish lakes in the western United States
would reveal its presence elsewhere. In the Devils Lake complex
it occupies a wide variety of habitats, showing its adaptability to
widely different conditions. It occurs but rarely in Devils Lake
proper, but recenth' has developed in considerable number.s in
East and Mission Lakes.

Marshia appears occasionally in the plancton catches from
East Lake, but too infrequently to warrant very definite conclu-
sions regarding its seasonal distribution. IM}^ earliest record is
May 5 and my latest August 2.

Laophonte calamorum. I have but one record of this species
for the Devils Lake complex, from Lake C, Sept. 1917. It has re-
cently been reported by Willey {1923a) as a new species from the
Quill Lakes, Saskatchewan, which are quite similar in their chemical
character to those of the Devils Lake complex, and in Lake St. John,
Quebec. According to Willey the other species of the Laophontidae
are strictly marine, but this species has adapted itself to both
brackish and fresh water.

Hyallela azteca. A common form, but not of sufficient num
bers to appear, other than occasionally, in the plancton. It occurs
in all zones and at all depths of the Main Lake, and is widely dis-
tributed thruout the complex, and neighboring fresh waters. It is
notably phototropic in its reactions, but, as with other zooplankton,
does not appear to show any relation to light in its distribution.
Stereotropism appears to be an important factor in its distribution,
for it is frequently found crawling over submerged objects, such
as timbers. It is more or less of a scavenger, and is found in large
numbers on the bodies of dead birds. In their moist feathers it
will live for several days after removal from the water. I have
observed it copulating as early as April 15, while the lake was still
frozen.

Cyprls pellucida. This species has been definitely determined
from Main Lake onh', but doubtless occurs in other parts of the
complex also.

Ostraeods occur but rarely- in the plancton catches; consequent-
ly no attempt has been made to follow their seasonal distribution.

*From Devils Lake complex.

94 THE LIFE OF DEVILS LAKE

They are seldom found in the open water, being restricted mostly
to the Ruppia zone near shore.

Crustacea may occur in oxygen-free water, together with other
animals.*

INSECTA

The aquatic insects of Devils Lake and adjoining waters have
been studied by Mr. C. K. Sibley of the Department of Entomology
of Cornell University, assisted by Miss Helen Stegenga formerly of
the University of North Dakota. A few data have been supplied
by Professors C. J. Needham, E. Matheson, Mr. C. R. Plunkett and
C H. Kennedy. Many of the forms, however, are still awaiting
specific determination.

With the exception of the corixids and occasional chironomids,
the insects of Devils Lake are found in the shore and Ruppia zones
and in the ooze. While they occasionally occur in the plancton
catches, they do so too infrequentlj- to warrant any conclusions re-
garding their seasonal distribution. Much remains to be done, not
only in their classification, but also in a study of their food, repro-
duction and reactions. The present report can be considered only
as a beginning.

Most important, in respect both to variety and number, are the
chironomids. Their larvae are present at all times in the year in
the ooze, while in summer the most conspicuous insects about the
shores are members of this family, Avhich fill the air with their
swarms in the evening, and during the day lie hidden in the grass
and brush.

Second in importance to the chironomids are the corixids, of
which at least two species inhabit the lakes, and which are common
everj'W'here, especially in the Ruppia zone.

On the whole, however, the insect fauna of Devils Lake cannot
be said to be rich in species, tho plentiful in number of individuals.

Its constituent species are characteristically those of fresh
water, tho a species of Ephydra, which is a salt and alkaline water
form, is common in Main and Mission Lakes.

Several species of dragon and damsel flies naturally occur in the
Devils Lake region, but only one, so far as known, breeds in the
main lake. This is an undetermined (probably new) species of
Enallagma with a two-segmented gill, which according to Professor
Needham is "unlike anything in our eastern fauna."

Damsel fly nymphs are very common in the shore and Ruppia
zones of Main and Stump Lakes, and have been taken in Lake 0.

*See p. 107.

THE LIFE OF DEVILS LAKE 95

They probably occur in most of the lakes of the complex, but these
are the only records.

Sympetrum corruptum breeds in Lake P, but apparently not
in other parts of the complex.

Arctecorixa sp. Common and widely distributed in all lakes
of the complex and in fresh water.

Corixa sp. A common species. Reported from Main, Stump,
East and Spring Lakes and Lakes A and P.

Corixids are both thermo — and phototropic. I have found them
numerous in the open water along shore in spring, when most of
the lake was frozen and the shore water was several degrees warmer
than that beneath the ice. They are attracted by the light of a
lantern, but their manner of response appears to be that of trial and
error, not direct.

They are present at all seasons of the year, but while occa-
sional in the plancton samples are not sufficiently numerous to ad-
mit of any definite conclusions regarding their seasonal distribution.
While occurring in all zones and at all levels, they appear to be more
numerous near shore and in the Ruppia zone. I have observ^ed them
copulating in August and September and have found immature
specimens in July and August. In a laboratory they do not live
long. Whether this is due to a natural bre\aty of life, or to the
artificial conditions of their environment I cannot say.

Notonecta sp. Prior to 1914 the back-swimmer was occasionally
taken in Main Lake. It was present thru the summer and beneath
the ice in March.

Notonecta undulata. Reported in Lake 0, also in fresh water.

Biienoa niargaritacea. A single specimen from Lake P about
September 5, 1919, is the only record of this species.

Triaenodes flavescens. Larvae were found in Lakes C, 0 and
P and in fresh waters.

Limnophilus rhomhicus L. This widely distributed species is
the commonest caddis fly in the region. Present in Minnewaukan
Bay and in fresh waters (1921).

Phryganea sp. A few caddis fly larvae were found among
algae near shore in Main Lake in 1922 by my assistant, Miss
Stegenga. This record is of considerable interest because of the
very unusual occurrence of caddis larvae in other than fresh water.*
Apparently these larvae are not of regular occurrence in Devils
Lake, for this is the only record of their presence here. Larvae of
P. interrupta occur rather rarely in fresh water lakes in the vicinity
of Devils Lake.

♦See p. 102.

96 THE LIFE OF DEVILS LAKE

Chirononius. At least three species of this genus occur in
Devils Lake, the specific determination of which has not yet been
possible.

Midges, including both C'hironomus and Protenthes, are the
most characteristic insects in all part of the Devils Lake complex.
Of these P. pnnctipennis is probably the most common. They may
be seen swarming about the lake shore any still evening in summer
or early autumn until after the middle of September. During the
daj' they seek the shelter of the herbage on the shore. Some species
as larvae inhabit the Ruppia zone, while others live in the ooze,
but determination of the exact distrib\ition of the various species,
naturally awaits the determination of the latter. Those which in-
habit the ooze in the deeper parts of the lakes are frequently in an
environment free from oxygen." Occasionally the ooze inhabitants
are found in the upper levels of the lake, probably as the result of
wind action. Tn the ooze the larvae live in tubes.

Midge larvae of probably more than one species occur rarely
in the plancton samples, not being normally inhabitants of the open
water. The east pupal skins frequently occur in large numbers on
the surface of the lake, and especially on shore where they are driven
by the wind.

Thienemann (1918-20) has classified lakes faunistically as
1) Chironomus, 2) Chironomus plus Corethra — and 3) Tanytarsus
lakes ; the first being characterized by the presence of the red
Chironomus larvae at the bottom, the second by the presence of
Corethra larvae in addition to those of Chirononms and the 3) by
the absence of the two former and the presence of the larvae of the
Tanytarsus group of chironomids (i. e. Lauterbornia). In the first
of these the oxygen content of the bottom water in summer ranges
from 35 to 50% of saturation; in the second it lies below 37% and
in the third above 50%.

Thienemann asks w^hether his classification is of universal ap-
plication or not. It appears to apply nicely to Devils Lake, which
belongs in the first group above mentioned (i. e. Chironomus lakes).
Neither Corethra nor Tanytarsus occur here, and the oxygen content
is probably often zero in the ooze where these midge larvae live.
This ooze is probably frequently aerated by Avind action in summer,
hut the oxygen content quickly drops upon subsidence of the wind,
in consecpience of the oxydation of the large amount of organic
material which it contains. In early spring also, before the melting
of the ice, the bottom water may be oxygen-free as already noted.*

Tmiypvs sp. Recorded only from Lake P and a fresh water
pond in the vicinity.

*See pp. 106-7.
*Seo p. ;50.

THE TJFE OF DEVILS LAKE 97

Procladius sp. Several specimens, examined by Prof. Needham,
were taken in a surface tow in Creel Bay on July 6, 1911. "Ap-
parently a new species," but impossible of determination because of
poor condition.

Sayomyia sp. One specimen was taken in a surface tow near
Bird Island, July 31, 1911.

Odontmnyia sp.

Stratiomyia sp.

Soldier flies include two indeterminate species, both of which
are common as larvae in shallow water near shore among decaying
vegetation, and as pupae in decaying vegetation and under logs and
stones along shore. They have been taken in Main, Lamoreau and
Stump Lakes and Lake P, as well as in fresh waters.

Nemotehis sp. Taken in Main Lake, Lake P and a fresh water
pond near Stump Lake.

Tahanus sp. Adult horse flies are common in the Devils Lake
territory. "Larvae and pupae were quite common in the sand
at the waters edge" in Main and East Lakes and a fresh water pond
near Stump Lake.

Chrysops sp. Deer flies are exceedingly common in all of the
Devils Lake territory. The larvae occur in "sand at the water
line." Thej^ are not common in the alkaline waters of the complex,
and probably develop mainly in fresh waters.

Eristalis tenax Rat-tailed maggots ars common in decaying
vegetation along the shores of the main lake and Lake P.

There are several species of beetles in the complex, but it has
not been possible to obtain identifications of most of them.

Octhehius sp. is common and widely distributed thru the com-
plex.

Bidessus lacustris. This and the following genus has been
determined by Dr. R. Matheson of Cornell. I have one record from
a surface tow near the station on July 14, 1911,

Coelamhus. Two undetermined species were taken in the same
collection as the above. C. nubilus has been taken in Lake N.

Deronectes sp. Common thruout the Devils Lake complex. Oc-
curs also in fresh waters.

Berosus sp. "Larvae were numerous" at the grade across
Minnewaukan Bay, "and in a fresh water pond north of East
Lake." (1921).

Hydrous triangularis. "Larvae were common at the Minne-
waukan grade" and in fresh waters (1921).

Tropisiernus sp "Larvae were found at the Minnewaukan
grade," and in Six-Mile Bay, also in fresh waters. (1921).

Agalus sp. There is one record only (Creel Bay 4/21/23).

98 THE LIFE OF DEVILS LAKE

Philhydrus sp. This is the common hydrophilid of Devils Lake.
All stages were found. The species was also present in Lake A,
Lake P, Stump Lake and a fresh water pond.

HYDRACARINA

There are several species of mites in the Devils Lake complex, as
yet only partly identified. Dr. Karl Viets has determined Hy-
drachna schneideri and H. valida, both from Main Lake, the former
from Spring and C Lakes also. One specimen of Diplodontus despiciens
from Spring Lake has been identified by Prof. Ruth Marshall, while
Prof. R. H. Wolcott has found a third species of Hydrachna, as yet
undetermined, in Main Lake. Eylais sp. occurs rather commonly
in Main Lake and E. extendens has been taken in Lakes Spring and
P. A single specimen of Hydryphantes was found by Prof. Wolcott
in a collection from a shore pool close to the lake edge. The latter
was probably formed in part by rain and partly by wash from the
lake, from which the Hydryphantes undoubtedly came.

Mites are fairly common in Main Lake but do not occur in
sufficient numbers in the plancton samples to give definite records
of their seasonal distribution.

Other than Main Lake I have recorded mites from Spring Lake
and Lakes C and P only, but they doubtless occur elsewhere in the
complex.

FISH

At present there is but a single species of fish inhabiting Devils
Lake, the stickleback (Eucalia inconstans), altho previously it is re-
ported to have swarmed with pickerel (Esox lucius?), and Lord
(1884) reports the "shiner," and Pope (1908) the minnow (Pime-
phales phomelas) as occurring here.

The reasons for the disappearance of fish from Devils Lake are
probably primarily the increase in concentration of the lake water,
and secondarily the loss of suitable breeding grounds for the pickerel,
thru the cutting off of the coulee which formerly flowed into Min-
newaukan Bay, but which in recent years has only done so excep-
tionally. The ruthless destruction of the pickerel by the fishermen
was probably a contributing cause, and there is some evidence (but
very uncertain), from the accounts of early observers, of an epidemic
among the fish.

Since 1908 numerous experiments have been made in the intro-
duction of various species of fish (chiefly yellow perch — Perca
flavescens) into Devils Lake, but without any permanent success,
and on the tolerance of fish for several different salts in varying
concentrations.*

In another part of this report** I am discussing briefly the

•See Popp (1908), Brannon (1911, 1913) and Young (1923).
•♦See p. 106.

THE LIFE OF DEVILS LAKE 99

great difference in tolerance for salt concentration shown by various
species of both plants and animals in the Devils Lake complex, and
need not consider it further here.

The stickleback occurs in both Main and Stump Lakes and
probably in other lakes of the complex, as well as in fresh waters.
It was formerly common in Main Lake, but appears to be much less
so at present. In the summer of 1914 several were found dying
in the lake, for a reason which could not be ascertained. The tem-
perature of the water at the time was not higher than probably oc-
curs frequently in lakes where the stickleback abounds, and bacteri-
ological examinations made a few hours after death gave no evi-
dence of pathogenic bacteria in the fish.

In 1912 also the sticklebacks were found dead in large numbers
on shore in spring about the time of opening of the lake. The cause
of this fatality also is obscure. It is possible, altho it seems un-
likely, that the fish may have been trapped in the ice and frozen
to death. Lack of oxygen will hardly account for it, since it al-
ways appears to be ample, thru the greater part, at least, of the
water stratum.

The freezing of a layer of water a metre thick in a lake only
6 or 7 metres deep necessarily causes a great concentration of salts
in the remaining water, and it is possible that the weaker stickle-
backs were unable to withstand this increased concentration; and
further that their death in summer may have been due to salt con-
centration, coupled with high temperature.

In recent years their death has not been noted, altho they ap-
pear to be present in considerably diminished numbers.

They breed in the Ruppia, nests with eggs having been found
during late June and early July.

AMPHIBIA

Two amphibians occur in Devils Lake, the pickerel frog (Rana
pipiens) and the salamander (Amblystoma tigrinum).

The former of these is occasionally common about the edges of
the lake and has been taken in Stump Lake, while the latter is a
regular inhabitant of its waters. Neither species, so far as known,
breeds in the lake.

The salamander was reported by Pope (1908) as Crytobranchus
allegheniensis, and later reported by me (Young 1912) as Ambly-
stoma. They occur both as larvae and adults. Some of the adults
are unusually large, measuring 25 to 28 cm in length.

There are numerous small fresh water ponds, ditches and
swamps in the Devils Lake complex, which in spring are usually
full of water, and which probably serve as breeding grounds for
both frogs and salamanders. In one of these, near the south end
of Stump Lake, the salamanders breed abundantly.

100 THE LIFE OF DEVILS LAKE

The ability of frogs and salamanders to tolerate the highly
concentrated water of Devils Lake, is of great interest, in view of
the inability of most species of fish to live therein. Further con-
sideration of this question is given below.*

AVES

In his recent report on the birds of North Dakota Wood (1923)
lists 67 species of water birds at Devils Lake, divided among the
following orders : Colymbiformes, 6 ; Ciconiiformes, 5 ; Anseri-
formes, 23; Charadriiformes, 33.

While birds are not, properly speaking, a part of the life of a
lake, they nevertheless play an important part in its bionomics by
feeding extensively on the insects and other organisms which it con-
tains.

MAMMALS
The only mammal present in Devils Lake, other than accident-
ally, is the muskrat, (Fiber zibethicus cinamominus). This is
frequently seen swimming in the lake, and what are undoubtedly
its burrows may be found occasionally on the shore, but I have
never seen any of its "houses" in the lake. It is not numerous,
and plays a minor role in the biology of the lake.

PARASITES
There are of course several parasites infecting various mem-
bers of the Devils Lake fauna, but no study has been made of them.
While, in a broad sense, they might be considered members of the
fauna, in a stricter sense, they are not, since their habitat is deter-
mined for them by their hosts, and is not, at least so far as known,
influenced by the physico-chemical environment of the water in
which the hosts may happen to live.

DISCUSSION OF RESULTS

Devils Lake is interesting as well for what it does not, as for
what it does contain. Derived from an originally fresh body of
water, it probably at one time contained representatives of all the
larger groups of fresh water animals. At present there are no mem-
bers of the sponges, celenterates, nemerteans, annelids, bryozoans,
molluscs or reptiles in the alkaline waters of the complex, altho
sponges, leeches and molluscs and turtles occur in some of the fresh
water ponds, which at one time were part of Devils Lake. Thus far
no hydroids, nemerteans or bryozoans have been found in any part
of the complex, either alkaline or fresh. Large numbers of snail
shells, including several species occur along the shores of Devils
Lake, but apparently no living species at present inhabit its waters.
At one point, in a gravel bed, many mussel shells occur, evidently re-
mains of a former fresh water fauna.

♦See p. lOG.

THE LIFE OF DEVILS LAKE 101

Among the plants all of the larger groups are represented, but
the flowering plants by practically only a single species. While
both fauna and flora have more fresh than salt water affinities,
there is one genus (Chaetoceros) and several species, which are
typically marine or brackish water forms.

But while the fauna and flora are characteristically fresh wa-
ter, a comparison with those of fresh waters elsewhere reveals sev-
eral striking differences. Here, as in brackish water elsewhere, the
blue-green algae play a predominant role, while the diatoms and
desmids, which often form so important a part of the plancton, are
here, with the exception of the marine genus Chaetoceros, very ir-
regular in occurrence and comparatively small in number, altho of
the diatoms a large variety of species occur. Nodularia, which at
certain seasons is one of the most abundant plants in Devils Lake
is seldom met with in fresh water, while forms such as Aphanizome-
non, Aphanocapsa, Clathrocystis (Microcystis), Oseillatoria, and
Anabaena are either absent or of wholly minor importance.

Of the Protozoa there are none of importance in the plancton,
and some forms, such as Ceratium, Synura and Dinobryon which
in other waters are often so abundant, appear to be wholly absent
here ; while the scarcity of shell-bearing rhizopods is noteworthy.

The number of important species of rotifers and Crustacea is
comparatively few, there being but three of each. Among the for-
mer Brachionus satanicus and B. plicatilis spatiosus are new, while
Pedalia fennica has hitherto been reported from only a few locali-
ties. The rotifers which are usually most important in fresh water
are here either absent or of secondary importance. In their ex-
amination of a large number of lakes in the northwestern United
States Kemmerer et al. (1923), report Pedalia* from only one —
Medical Lake, Washington, which is so "distinctly alkaline" that
"no fish will live there for any period of time."

The Crustacea are likewise signalized by the absence or rarity
of several animals, chiefly Cladocera (Daphnia, Bosmina, Leptodora,
etc.) which are characteristic of the plancton of fresh water lakes
in general.

Of the nine orders of insects which Needham and Lloyd
(1916) give as common in inland waters four (Plecoptera,
Ephemerida, Neuroptera and Lepidoptera) have not been found in
the Devils Lake complex, while the Odonata and the Trichoptera are
represented by a single species each.

Comparatively little work has yet been done on the alkaline
lakes of North America, so far as I am aware. Recently Huntsman
(1922), Willey (1923a) and Bailey (1922) have made a brief survey
of the Quill Lakes in central Saskatchewan, which shows a striking

•Species not given.

102 THE LIFE OF DEVILS LAKE

similarity of fauna and flora in these lakes, as compared with
Devils Lake, in correspondence with their similarity in physical and
chemical characteristics.

Like Devils Lake, the Quill Lakes are shallow and of high
salinity (about 11000 ppm. of total solids in Little Quill and
16500 in Big Quill). The study of the Quill Lakes has been too
brief to permit of any comparison between their life and that of
Devils Lake, either as to quality or quantity, but a few points may be
noted. In both Diaptomus sicilis appears to be the principal crus-
tacean with Hyalella azteca a common species. Cyclops viridis is
represented in Devils Lake by var. americanus and in the Quill
Lakes by parens. Laophonte calamorum, a new species from the
Quill Lakes, occurs in the Devils Lake complex, altho it has not yet
been found in the main body of Devils Lake. The Odonata are
represented in Devils Lake by Enallagmia sp. and in the Quill Lakes
by E. calverti ? It is interesting to note the presence of caddis fly
larvae (Phryganea interrupta in the Quill Lakes and what is prob-
ably the same species in Devils Lake) as these are characteristically
fresh water inhabitants.

The presence of numerous gastropod shells along the lake shores,
but the absence of their occupants in the water is noticeable in both
regions.

Sticklebacks are the principal species of fish in both waters,
but the presence of other species in the Quill Lakes and their ab-
sence from Devils Lake is a noticeable difference.* In both regions
the sticklebacks are frequently infected with large tapeworms
in the eelome, the identity of which has not been determined.

The algae reported from the Quill Lakes are, in most cases,
species occurring in Devils Lake. Of particular interest is Nodularia
spumigena which is abundant in the Quill Lakes and is one of the
characteristic forms in Devils Lake, but is not common elsewhere.
Of much interest also are the diatoms. Seven of these, which are
characteristically marine types, are reported by Bailey (1922) from
the Quill Lakes, of which three, either identical or closely related
forms (Chaetoceros, Surirella striatula and ovalis) are common to
both regions.

Among the sandhill lakes of Nebraska are found waters o4
similar character to Devils Lake, the algae of which have been
studied by Andersen and Walker (1920).

Detailed analyses of these lakes are not available, but they
range from comparatively fresh waters (111 ppm. of "alkali") to
highly alkaline ones (1129 ppm.). While the depth of only one is
given, they are apparently all shallow (under 3 m.) Since the
periods and localities of collection were not the same in all, an exact

•For a further discussion of this question s'ee Young (1923, pp. 387-8).

THE LIFE OF DEVIDS LAKE 103

comparison of their algae is impossible, but, in general, the alkalinity
appears to be a limiting factor in the distribution of these plants,
the higher the amount of alkalinity the smaller the number of
species. Big Alkali lake with 622 ppm. of "alkalinity'' is probably
most like Devils Lake, but as this was visited only once, only a very
superficial comparison of their algae is possible. Of 26 species re-
corded from this lake 16 occur in Devils Lake and every genus is
represented except Chara.

Recently Moore and Carter (1924) have extended the survey
of the former (Moore, 1917) to include about 70 lakes in central
North Dakota, including both fresh and alkaline types. As a re-
sult of this study these authors conclude; "that as far as the Myxo-
phyceae are concerned the 'alkaline' lakes are considerably richer
in species than the 'freshwater' lakes. On the other hand a def-
inite 'bloom' was rarely developed in strongly alkaline waters,
whereas in several of the 'freshwater' lakes there was frequently
a copius 'bloom' consisting of Clathrocystis aeruginosa, or some-
times of Aphanizomenon flos-aquae, one or other of these species
being dominant, and often the only representative of the Myxophy-
ceae in the sample. In the alkaline lakes there was never such a
pronounced development of a single species of Oscillatoria, Nodu-
laria spumigena, Arthrospira Jenneri and Spinilina spp. The tiny
species Rhabdoderma sigmoidea, described here for the first time,
also occurred in some of the alkaline lakes as a dominant constituent.

With regard to the Chlorophyceae, the condition is reversed.
Many of the species seem to be very intolerant of increasing con-
centration of salts, and are therefore eliminated from the flora of
the strongly alkaline lakes. On the whole the Conjugatae are very
poorly represented, as pointed out by Moore (I.e., p. 302), although
a few species are able to thrive under the unusual conditions. The
genus Oocystis seems to maintain itself satisfactorily in the alkaline
waters, as do also some species of Scenedesmus, Pediastrum, and
Dictyosphaerium spp. Many of the Protococcales commonly found
in freshwater plancton are not represented at all, however, in lakes
of the 'alkaline' group.

The 'alkaline' lakes were conspicuous by reason of their paucity
of species, the richest lakes of this type containing only 5-9 species,
whilst some of the 'freshwater' lakes had as many as 12-17 species."

While the number of species inhabiting Devils Lake is rather
restricted, especially in certain groups; the number of individuals
of a few species is large. This is especially noticeable in the case of
the rotifers Brachionus and Pedalia and in that of the Myxophy-
ceae, Nodularia and Coelosphaerium. It is quite possible that the
paucity of certain species, thereby restricting competition, has per-
mitted a greater development of others than would otherwise have
occurred. More important factors, however, are probably the favor-

104 THE LIFE OF DEVILS LAKE

able physical environment, the shallowness, and the lack of inlet
or outlet to the lake. Shallow lakes, in general, are better plancton
producers than deeper ones, and Devils Lake is no exception to
the rule. It is well known, also, that certain temporary pools, may,
at times, be swarming with life, while in bodies of changing water,
especially rivers, the number of organisms is frequently much less.

A comparison of the amount of plancton in lakes and rivers is
difficult to make because of the great individual differences between
different bodies of water, due to various factors. Thus Birge &
Juday (1922) in their study of three Wisconsin lakes* found that
the total amounts of dry net plancton in milligrams per cm. ranged
from 491 for Mendota to 2182 for Waubesa, while the corresponding
amounts for centrifuged material ranged from 3090 to 5665. Apstein
(1896 p. 92) found variations running from 4.5 cc. pr. cm. to 454.3
cc. and similar variations have been found by other authors.

A comparison between the amount of plancton in lakes and rivers
made by Whipple (1914, p. 20), iu spite of great individual varia-
tion, shows clearly enough the relative paucity, of plancton in many
streams as compared with lakes.

Kofoid (1908), however, records amounts of plancton for the
Illinois River,* which compare favorably with the figures for most
lakes. So much depends on the individual character of the water in
question that it is unsafe to generalize too broadly.

A comparison between the plancton production of Devils Lake
and that of other waters is difficult to make because different
workers have used different methods of expressing their results, and
in many cases, moreover, the period of observation is too brief to
enable one to draw any conclusions concerning maximum, minimum
or average productivity of the water in question. Some data art,
however, available and maj^ be of interest. The average number of
Crustacea (incl. nauplii) in Devils Lake for the periods of collection
from 1911 to 1923 inclusive is 228,000 pr. cm. From table B in
Birge (1897) I have computed the average for Lake Mendota in
1895-96 as 37,350 pr. cm. In each case the average was obtained
"by adding all the collections for a given date and dividing by their
number. From the data given by Birge & Juday (1914, '21) I have
computed the averages for Canandaigua, Cayuga, Hemlock, Otisco
and Seneca Lakes, N. Y. and Green Lake, Wis., the maximum of
which is 67,800 pr. em. in Otisco Lake on August 16, 1910. It
should be borne in mind that these averages are for very few col-
lections; for Hemlock, Green and Otisco Lakes, only one; and for
the others three each.

The rotifer average for Devils Lake for the same period is
350,000 pr. cm., while, for the reproductive season only, the average

•Mendota, Monona, Waubesa.
♦See Kofoid, p. II, p. 16.

THE LIFE OF DEVILS LAKE 105

is 560,000. The maximum for the lakes studied by Birge & Juday
was about 40,000 pr. em. for Cayuga Lake, while the others were
distinctly lower.

In a recent paper by Kemmerer et al. (1923) there are data for
51 lakes in the northwestern United States. I have computed the
averages for a fe-sv of these, chosen more or less at random, but in-
cluding both deep and shallow lakes. The maximum number of
Crustacea was 212,400 for Calvert Lake, Washington, a small lake
about 12 m. deep. The other lakes ran very much lower than this.
The rotifer maximum for these lakes was 34,000 pr. cm. in Cottage
Lake, Washington on August 13, 1913. Here, too, it should be re-
membered that these averages are, in most cases, for only a single
series of collections.

The data for the nannoplancton have, in most cases, been based
on net collections, and hence are much below their true values.
Birge & Juday (1923), however, give some data from centrifuged
collections. Neither method, however is closely comparable with the
Sedgwick-R after method, which stands between the other two in
respect to the amount of nannoplancton recorded by it.

The average number of diatoms in Devils Lake for the years
1911- '14 inclusive is 50,000 pr. cm. which is higher than most of
the figures recorded by Kemmerer et al. (I.e.),* altho in a few
instances their figures are very much higher, ^t is in every
instance far below the figures given by Birge & Juday (1921) for
their centrifuged samples, and in several cases lower than their
results for the net plancton.

I have attempted no comparison for the other algae for the
additional reason that many of my results are in "standard
units,"** or mm., while other authors state their results in colonies
or filaments.

Devils Lake may be compared to a large pool, where, finding
conditions favorable to existence, and little competition, and there
being no flow to carry them away as they multiply, certain types
of life have developed extensively; while others have developed
slightly, or not at all, in an environment, which is unfavorable to
them. Rate of reproduction enters here, of course, as an important
factor, and the whole question is complicated and obscure.

The study of the lake has revealed new species as follows:
Protozoa — Urotricha labiata, Gerda annulata, (Edmondson 1920) ;
Diatoms — Chaetoceros elmorei, Navicula minnewaukonensis, (Elmore
1921) : Rotifiers — Brachionus satanicus, B. spatiosus and B.
pterodiuoides, (Rousselet 1911, 12 & 13) ; while several hitherto known

♦In Tipper Klamath Lake, Orejron, on July 20, 1913 there waa the large average
of more than 20.000,000 pr. cm. This is a shallow lake in most of its area, with
a maximum depth of about 11 tn.

•*Seo Whipple (1914).

106 THE LIFE OF DEVILS LAKE

only from other continents have been found to occur here. These
include: Tetrapedia gothica, previously known only from Germany,
Characium hookeri from Europe and Asplanchna sylvestrii, previ-
ously recorded only from Chile, so far as I know.

The difference in tolerance of different types of life to waters
of various salt concentration opens up a wide field of interesting
experiment and speculation. I have already (Young 1922) dis-
cussed this question at some length in connection with experiments
on fish. These, and other experiments indicate, not only considera-
ble specific, but also individual differences in resistance to unfavora-
ble environment.

In his paper on the African Lakes Cunuington (1920) has pointed
out the absence or rarity of certain species of animals and plants
in Tanganyika and Kivu, which occur in other lakes of Central
Africa, or even in tributary streams.

Thus, with the exception of an occasional isolated specimen
there are no Cladocera in the two former lakes, altho they occur in
the Lofu River, one of the tributaries of Tanganyika, and in Lakes
Nyasa, Victoria, Albert and Edward Nyanza. Rousselet* records
23 species of rotifers from the river and only 10 from the
lake, only one of which was common to both lake and river; while
West* found 30 species of phytoplancton in the river, but
one of which was present in the lake.

Cunuington seeks an explanation for these facts in the saline
character of the water of both Tanganyika and Kivu. He gives
no analyses of either, but states that the amount of magnesium
salts is high, altho Tanganyika water is "fresh."

In the absence of adequate data no satisfactory comparison can
be drawn with the water of Devils Lake. The character of the
three waters is probably somewhat similar, however, since in all
of them magnesium salts are present in considerable quantity. In
Devils Lake, however, these salts have not excluded the Cladocera,
nor apparently restricted the rotifers. Some species are, however,
so restricted. The scarcity of the Conjugales** and the shell-bear-
ing rhizopods is noteworthy,** and there are several groups present
in Tanganyika which are Avholly unrepresented in Devils Lake.***
It is doubtful then if salinity alone will explain these deficiencies in
the fauna of the African Lakes and what the explanation is re-
mains obscure.

The presence of animals in the oxygen — free ooze at the bottom
of lakes is a discovery of great physiological interest, since it has
generally been regarded that oxygen is a sine qua non for the exist-

*Fide Cunniuprton.

♦♦See Moore (1017) and Edmondson (1920).

•"•Gastropoda. MamelUbraiuhiata. I'ol.yzoa. Macrura, Bmchyura, Branchiura,
Olipochaeta, Iliiudinea, Ilydrozoa and Porifera. Of the above those italicized occur
also in Kivu.

THE LIFE OF DEVILS LAKE 107

enee of animal life. Recently, however, Juday (1908) has found
many species of animals living in the bottom of Lake Mendota, in-
cluding several Protozoa, rotifers, annelids, crustaceans, insect lar-
vae, a nematode, and a mollusc. Juday has shown very clearly the
absence of dissolved oxygen in the bottom water of Lake Mendota
at certain seasons, both by boiling and titration tests.

In my own work on Devils Lake I have used exclusively th«
latter method, which, as already noted,* is probably not very ac
curate in alkaline water. The decaying character of the ooze,
which has a very foul odor, is strong presumptive evidence, however,
which, together with the titrations and the results of Juday 's work,
renders it practically certain that the bottom ooze of Devils Lake is
frequently oxygen free.

In this ooze I have found several species of Protozoa and a
red Chironomus larva, while in the water immediately overlying the
ooze, which showed no oxygen by titration, I have found Diaptomus,
Moina, Cyclops and nauplii, in addition to the Protozoa. These
animals may, however, have been temporary invaders of this layer,
which is never very thick.

As yet we have no satisfactory explanation of the way in which
these animals live in an oxygen-free environment, tho various ex-
planatory theories have been advanced, which have been discussed
at some length by Cole (1921).

In any study of distribution of either animals or plants, one
finds some wide-spread, adaptive species, and others of restricted
distribution. The latter may not necessarily be non-adaptive, how-
ever, since the question of ease of dispersal enters into any prob-
lem of distribution.

The questions of adaptability and dispersal lead up to that of the
source of the life of Devils Lake. While it is possible to trace with
a fair degree of probability the origin of land floras and faunas,
from the distribution of both living and extinct forms ; this is
much more difficult in the case of aquatic organisms, especially
those of the plancton ; because of their ease of dispersal and conse
quent wide spread distribution; their great adaptability, in many
cases, to widely different environments, and especially to our lack
of knowledge concerning the distribution of many of them.

Thus, in at least three cases, in the present stud}'** organisms
have been recorded for the first time in North America, and in a
recent paper on the algae of several other North Dakota lakes by
Moore and Carter several other species are recorded as new to
America.

Diaptomus sicilis has hitherto been considered as a cold water
species. Thus, according to Marsh in Ward & Whipple (1918) it oc-

*See p. 30.
**See p. 106.

108 THE LIFE OF DEVILS LAKE

curs commonly in winter in Green Lake, Wis. but is rare in sum-
mer ; but is found in the cold waters of the Great Lakes thruout the
summer. In the Devils Lake complex, however, it is common at
certain times in summer at high temperatures.** Brachionus
plicatilis is given as a marine and brackish water form in Brauer
(1912), but here it occurs in both fresh and alkaline waters. These
instances might be multiplied, but are sufficient to illustrate how
much we have yet to learn regarding the distribution and adapta-
tion of aquatic life.

It is unsafe therefore to attempt at present anj' delimitation of
life zones, or to draw conclusions regarding the origin of any fauna
or flora, based upon their plancton components.

This opinion agrees with that of other workers in this field.
Thus Jennings, in Ward & Whipple (1918), emphasizes the essen-
tially cosmopolitan character of rotifers, whose distribution is deter-
mined by the character of their environment, rather than its geo-
graphical position; while the former factor is ineffective in the case
of many species, which can inhabit fresh and brackish, or even sea
water alike. And Edmondson (1912) in his account of the Protozoa
from some high mountain lakes in Colorado, says (p. 67) "By com-
paring the following list with the protozoa reported from sea-level
in tropical Tahiti (Science, September 9, 1910), there will be seen
not only the latitudinal but also the great altitudinal range which
many of our species have." In a letter to the writer (Augu.st 21,
1923) Prof. C. J. Elmore who has worked on the diatoms of Devils
Lake says:

"I have never been able to find that certain species of diatoms
are northern and others southern. A good many species that are
common in this region (Nebraska) were first described as Arctic
diatoms . . . As to locality, I have found just the same species
in collections from India that I find at home."

Birge, in Ward & Whipple (I.e.), says (p. 687) of the Clado-
cera, "The geographical distribution . . . offers little of inter-
est that can be stated in a brief sketch, chiefl.y because the species
are so widely distributed. Some are . . . cosmopolitan. A
majority (of those) found in this country are found also in Europe."
The study of our fo7'nis has not gone far enough to enable us to
speak of the local distribution of each species within the area which
it covers, but it is known that the rare species are very irregularly
distributed. On the whole the fauna of the various regions of the
country is strikingly similar, but with some forms peculiar to each
region.

Sharp in the same place says (p. 800) "Owing to the varia-
tions in habitat, and the vicissitudes to which most fresh water

•♦Over 20° C.

THE LIFE OF DEVILS LAKE 109

Ostracoda arc subject and because of the variable and inconstant
nature of their surroundings, it is almost impossible to work out
their exact distribution."

So far as I am aware the only attempts which have been made
thus far to compare the zonal distribution of plaucton with that
of higher animals have been those of Dodds (1920, 1924) in his papers
on Entomostraca and life zones in Colorado, who found (1920, pp.
104, 5) that ''The zonation is definite, even though the present collec-
tions do not enable as to discriminate beween all of the colder zones.
It is equally difficult, on the basis of our present knowledge to
sharpl}^ differentiate these same zones in their continental extent
in Canada and northern United States . . . Though these
zones have been established and defined on the basis of organisms
other than Entomostraca, j^et the agreement is striking and real,
and extends even to common species inhabiting the same zones
in localities hundreds of miles apart and separated by many
degrees of latitude."

Marsh also in his chapter on the copepods in Ward & Whipple
(1918) points out the influence of temperature in limiting the dis-
tribution of certain species, but not of others.

If we attempt to determine the life zone of Devils Lake or the
origin of its fauna on the basis of its entomostracan population we
find that the majority of its species are wide spread or cosmopoli-
tan in their distribution. These include the three species of Cy-
clops (viridis, leuckarti and serrulatus) and the Cladocera, Bos-
mina longirostris, Moiua macrocopa, Daphnia longispina, psittacea,
pulex and magna, Diaphanosoma brachyurum, Ceriodaphnia pul-
chella, Alona rectangula, Chj'dorus sphaericus and Simoeephalus
vetulus and the ostracod Cypris pellucida. Diaptomus sicilis has
hitherto been considered a cold water type restricted to the Great
Lakes and adjoining waters. Recent studies, however, have ex-
tended its area westward to Devils Lake and northward into
central Saskatchewan (Huntsman, 1922).

D. leptopus piscinae occurs thruout the northwestern United
States, Alberta and Manitoba. D. shoshone occurs in mountain lakes
of the western United States and southern Canada. "The occur-
rence in Devils Lake is interesting in that it is out of the moun-
tains and is the most easterly location that has been reported."*
Its presence here brings it into the transition zone.

D. siciloides. Its occurrence in the Devils Lake complex is
the most northerly yet recorded bringing this species into the
transition zone from the south.

♦Letter from Dr. C. Dwight Marsh to the writer.

110 THE LIFE OF DEVILS LAKE

Marshia albuqerqensis has hitherto been reported only from
New Mexico** (Herrick and Turner 1895) and Colorado (Dodds,
1920). Our present knowledge regarding its distribution is too
limited to consider it in attempting to determine the zonal position
of the lake or the origin of its fauna.

Laophonte calamorum. This species, recently described by
Willej^ (1923 a), is evidently widely distributed in both fresh
and alkaline waters, having been found thus far in the Quill
Lakes, Saskatchewan, De^dls Lake, N. D. and Lake St. John,
Quebec.

The origin of the fauna and flora of Devils Lake must be
traced back to the days of the old glacial lake Minnewaukan when
the latter was a part of the Mississippi drainage basin. At this
time it was possible for both northern and southern forms to
have entered directly, and without any carriage by birds or any
air blown dust, etc. Later, with the loss of this connection and
the establishment of one into Lake Agassiz, only northern ele-
ments could enter directly, and this condition remained after the
retreat of Lake Agassiz and while Devils Lake was draining
through the Sheyenne River into the Red River. The loss of this
outlet has been comparatively recent. Since then additions to the
fauna and flora must have been thru outside agencies.

The foregoing facts indicate then that the fauna and flora of
Devils Lake are in general widespread in their distribution and un-
doubtedly have come from manj" sources.

At present the life of the lake is distinctly fresh water in
character with, however, a considerable admixture of brackish water
forms. The species are for the most part cosmopolitan in their
distribution. Of those types with restricted distribution it is dif-
ficult to say whether northern or southern ones predominate, until
we have fuller information regarding the distribution of aquatic
animals and plants.

The fauna and flora, moreover, are not static, but are changing
from year to year with the changing character of the various lakes
in the complex. Thus, as already pointed out,* Filinia lougiseta
has apparently disappeared from Main Lake with the gradual in-
crease in concentration of its water. Similarly the Crustacea, with
the exception of Marshia, have apparently disappeared from East
and Mission Lakes in recent years, while the latter, on the contrary,
has increased in numbers (1922).

No careful study has been made of these changes, but the above
data are sufficient to demonstrate clearly the great changes that are
taking place.

**In liis puMieation Ilprrick apparently does not mention the locality. Marsh,
howpvpr, in Warrl &. Wliipplc (1918. p. 780) kivos it as New Mexico.
*See p. 86.

THE LIFE OF DEVIDS LAKE 111

CONCLUSIONS

1. The Devils Lake complex comprises a series of large shallow
pools of alkaline water ; which were originally parts of a large, deep,
fresh water lake; in which certain organisms, finding conditions
favorable for their existence, develop at certain times in enormous
numbers; while other organisms, though present in adjacent waters,
are rare or ab.sent.

2. Conditions in the different parts of the complex vary, es-
pecially in respect to depth and chemical character. In response
to these changing conditions marked changes in their faunas and
floras have occurred and are progressing.

3. The character of these latter is that of fresh water, being
distinguished therefrom rather by the absence of many fresh water
species, than by the presence of many brackish or marine types, tho
a few of these occur.

4. The source of fauna and flora is manifold. They are not
characteristic of any particular region, but are rather cosmopolitan
in aspect.

BIBLIOGRAPHY

1920 American Public Health Association, Standard Methods for
the Examination of Water and Sewage — Ith Ed.

1920 Anderson, E. M. & Walker, E. R. An ecological study of the
algae of some sandhill lakes. Trans. Am. Mic. Soc. Vol.
XXXIX, pp. 51-85.

1896 Apstein, C. Das Siisswasserplankton. Lipsius and Tischer.
Kiel and Leipzig.

1912 Atkinson, T. R. Proposed diversion of ]\Iouse River to Devils
Lake. Fifth Bien. Rep., State Engineer, 1911-12, pp. 20-32.

1922 Bailey, L. W. Diatoms from the Quill Lakes, Saskat-
chewan, and from Airdrie, Alberta. Contrib. to Can. Biol.,
No. XL (1921).

1897 Birge, E. A. Plancton studies on Lake IMendota II. The
Crustacea of the plancton, July 1894-Dec. 1896. Trans. Wis.
Acad. Sci. Arts & Letters, Vol. XI, pp. 274-448.

1911 Birge, E. A. & Juday, C. Inland lakes of Wisconsin. Dis-
solved gases and their biological significance. Madison.

1914 A limnological study of the finger lakes of New York. Bull.
U. S. Bur. Fish. Vol. XXXII, pp. 527-609.

1921 Further limnological observations on the finger lakes of New
York. Bull. U. S. Bur. Fish. Vol. XXXVII, pp. 211-52.

1911 Brannon, M. A. First Biennial Report of the North Dakota
Biological Station (1909-10), State Pub.

1913 Second Biennial Report of the North Dakota Biological Sta-
tion (1911-12), State Pub.

112 THE LIFE OF DEVILS LAKE

1912 Brauer, A. Siisswasserfaima Deutsclilands. Heft 14. Gustav
Fischer, Jeiia.

1900 Burekhardt, G. Quantitative Studien iiber das Zooplankton
des Vierwaldstattersees. Mitt, iiatnrf. Gesell. Luzerne.

1911 Chandler, E. F. The Red River of the North. Quart. Jour.
Univ. of North Dakota, Vol. I, pp. 227-55.

1921 Cole, A. E. Oxygen supply of certain animals living in water
containing- no dissolved oxvgen. Jour. Exp. Zool. Vol.
XXXIII, pp. 293-320.

1920 Cunnington, W. A. The fauna of the African Lakes. Pro-
ceedings of the Zoological Society of London, pp. 507-622.

1920 Dodds, G. S. Entomostraca and life zones. Biol. Bull., Vol.
XXXIX, pp. 89-107.

1924 Notes on Entomostraca from Colorado. Proc.

U. S. Nat. Mus., Vol. LXV, pp. 1-7.

1912 Edmondson, C. H. Protozoa of high mountain lakes in Colora-
do. Univ. Colo. Studies, Vol. IX, pp. 65-74.

1920 Protozoa of Devils Lake Complex, North Da-
kota. Tran. Am. Mic. Soc. Vol. XXXIX, pp. 167-98.

1904 Ekman, S. Die Phyllopoden, Cladoceren u. freileb. Copepoden
der nordschwed. Hochgebirge. Zool. Jahrb. Syst. Bd. 21.

1915 Gandolfi-Hornyold & Almeroth. Mitteilungen iiber die Ver-
teilung von Daphnia hyalina im Genfersee. Int. Rev. Hydro-
biol. u. Hydrgr. Vol. VII, pp. 426-32.

1913 Grein, K. Untersuchungen iiber die Absorption des Lichts
in Seewasser. Ann. de 1' Institut Oceanographique, Vol. I, pp.
1-24.

1923 Heath, F. H. & Lee, F. A. Errors in the determination of

hydrogen sulfide. Jour. Am. Chem. Soc. Vol. XLV, pp.

1643-7.
1895 Herrick. C. L. & Turner, C. H. Entomostraca of Minnesota.

St. Paul.
1914. Huntington, E. et al. The climatic factor as illustrated in

arid America. Carn. Inst., Washington.

1915 Huntington, E. Civilization and climate. Yale Univ. Press.,
New Haven.

1922 Huntsman, A. G. The Quill Lakes of Saskatchewan and their
fisherj^ possibilities. Contrib. to Can. Biol. No. 9.

1908 Juday, C. Some aquatic invertebrates that live under an-
aerobic conditions. Trans. Wis. Acad. Sci. Arts & Letters, Vol.
XVI, pp. 10-22.

1916 Limnologieal apparatus. Trans. Wis. Acad.

Sci. Arts & Letters, Vol. XVIII, pp. 566-92.

1923 Kemmerer, G., Bovard, J. F. & Boorman, W. R. Northwest-
ern Lakes of the United States. Bull. U. S. Bur. Fish., Vol.
XXXIX, pp. 51-140.

THE LIFE OF DEVILS LAKE 113

1919 Leonard, A. G. The surface features of N. Dakota and their
origin. Quart. Jour. Univ. N. D., Vol. 9, pp. 209-295.

1905 Linsbaner, L. Photometrisehe Untersuchungen iiber die Be-
lenchtung-s-verhiiltnisse im Wasser. Sitzb. K. Akad. Wiss.
Wien. Math. Nat. Kl. Bd. 114.

1884 Lord, H. ^Y. The fish of Devils Lake, Dakota. Bull. U. S.
F. C, Vol. IV, pp. 351.

1902 Lozeron, H. La reparition vertical dn plancton dans le Lac
de Zurich. Vierteljahrschrift d. naturf. Ges. -Zurich. Jhrg. 47.

1898 Marsh, C. D. On the limnetic Crustacea of Green Lake. Trans,
Wis. Acad. Sci. Arts & Letters. Vol. XI, pp. 179-224.

1900 The plancton of fresh water lakes. Ibid.

Vol. XIII, pp. 163-87.

1905 Marsh, j\I. C. & Gorhaui, F. P. The gas disease in fishes.
Rep. U. S. Bur. Fish (1904) pp. 345-76.

1918 Moberg, E. G. Variation in horizontal ditsribution of planc-
ton in Devils Lake, North Dakota. Trans. Am. Mic. Soc. Vol.
XXXVII, pp. 239-67.

1917 ^loore, G. T. Preliminary List of Algae in Devils Lake, N.
Dakota. Ann. Mo. Bot. Gar. Vol. 4, pp. 293-303.

1924 Moore, G. T. & Carter, N. Algae from lakes in the north-
eastern part of North Dakota. Ann. Mo. Bot. Gard. Vol. 10,
pp. 393-492.

1912 Murray, J. & Hjort, J. Depths of the Ocean. MacMillan,
London.

1916 Needham, J. G. & Lloyd, J. T. Life of inland waters. Corn-
stock Pub. Co., Ithaca, N. Y.

1912 Nutting, P. G. Outlines of applied optics. Blakiston, Phila-
delphia.

1908 Pope, T. E. B. Devils Lake, N. D. U. S. Bur. Fish. Pub. 634.

1894 Reighard, J. E. A biological examination of Lake St. Clair.
Bull. Mich. Fish Com., No. 4.

1912 Ritter, W. E. The jNIarine Biological Station of San Diego
Univ. Cal. Pub. Zool. Vol. IX, pp. 137-248.

1911 Rousselet, C. F. On three new species of Rotifera. Jour.
Quekett Mic. Club. Vol. XI, pp. 161-64.

1912 On Notholca triarthroides, Cathypna brachy-

dactyla, and on a new Brachionus from Devils Lake, N. D.
Ibid!^ pp. 371-74.

1913 The Rotifera of Devils Lake, with description

of a new Brachionus. Ibid. Vol. XII, pp. 57-64.

1919 Robert, H. Contribution a I'etude du zooplancton du lac
de Neuchatel.

1920 Reprint from Bull. Soc. Neuch. Sc. Nat. Vol. XLV, pp. 75.
1907 Schonfeldt. H. von. Diatomaceae Germaniae. Junk, Berlin.
1922 Shelford, V. E. & Gail, F. W. A study of light penetration

into sea water. Pub. Puget Sound Biol. Station, pp 141-76.

114 THE LIFE OF DEVILS LAKE

1912 Simpson, H. E. Physiography of the Devils-Stump Lake Re-
gion, North Dakota. N. D. Geol. Survey, 6th Bien. Rep. pp.

105-57.
1910 Steuer, A. Planktonkunde. Teubner, Leipzig and Berlin.
1918-1920 Thienemann, A. Untersuchungeu iiber die Beziehungen

zwischeii dem Sauerstoffgehalt des Wassers und der Zusam-

mensetzung der Fauna in norddeutschen Seen. Archiv fUr

Hydrobiol., Vol. XII, pp. 1-65.
1896 Upham, W. Glacial Lake Agassiz. Monograph of the U. S.

Geol. Survey, Vol. XXV.
1896 Van Heurck, H. A treatise on the Diatomaceae. (tr. Baxter)

London.
1918 Ward, H. B. & Whipple, G. C. F^esh Water Biology. Wiley,

New York.
1904 & 1908 Wesenberg, Lund C. Studier over de danske Soers

Plankton. Dansk ferskvands-Biologisk Laboratorium Kjoben-

havn, Nordisk Forlag (2 Vol.)
1914 Whipple, G. C. Microscopy of drinking water. Wiley, New

York.
1900 Whipple, G. C. & Jackson, D. D. A comparative study of the

methods used for the measurement of the turbidity of water.

Technology Quarterly, Vol. XIII, pp. 274-94.
1921 Willard, D. E. Story of the Prairies. 9th Ed. Kalamazoo,

Mich.
1923 (a) Willey, A. Notes on the distribution of free-living Cope-

poda in Canadian waters. Contrib. to Can. Biol. No. 16.
1923 (b) Ecology and the partition of biology. Trans.

Roy. Soc. Can. Vol. XVII, pp. 1-9.
1923 Wood, N. A. A preliminary survey of the bird life of North

Dakota. Univ. of Mich. Mus. Zool., Misc. Pub., No. 10, pp. 96.
1912 Young, R. T. Cryptobranchus alleghaniensis, ... in

North Dakota. Science, N. S. Vol. XXXV, pp. 308-309.
1923 Resistance of fish to salts and alkalinity. Am.

Jour. Phys., Vol. LXIII, pp. 373-88.

INDEX

The following index includes only the more im{)ortant topics covered in
the paper. Eeferences to authors, many of the genera and, for the most part,
species are omitted. The two latter may be found iu their proper group by
refering to the table of contents on paige 3.

Page Page

A. F.

Adaptation to salt concentration 106

Algae, characteristic genera 60

relation to alkalinity 60-1

Amphibia, breeding 9&

Asplanchna silvestrii, seasonal distri-
bution 86

Bacteria, apparatus and methods 57, 8

distribution 59-60

seasonal occurrence 58-9

Brachionus plicatilis, distribution,

spatial °^

temporal 83

var. spatiosus 83-4

satanicus', distribution, spatial 84

temporal 84

variation 84-5

Caddis fly larvae, brackish watei' 95

Carbon dioxide 31

Chaetoceros, distribution, seasonal.— 71

Chironomus, habitat 96

relation to oxygen..- 96. 107

Chlorophvceae, relation to alkalin-
ity . .■- 60, 103

Chroococcus. distribution, seasonal.... 61

Cladocpra, distribution, unrestricted.. 109

swarms 37

Cladophora, distribution 34

Coelosphaerium, distribution, &• e a-

sonal 63

Color of water 18

Copepoda, life cycle - 45

Corisids, reactions 95

Cosmopolitan species 109

Crustacea, amount, comparison of

lakes 104-5

comparison with fresh water 101, 106

distribution, seasonal 87

vertical 42-3

zonal 109-10

see Cladocera, Copepoda

Currents 19

Cyclops, distribution, seasonal 91

relation to oxygen 107

D.

Devils Lake, comparison with life in

other lakes 100-4

conditions in 105

Devils-Stump Lake complex history.. 9-12
parts 17-8

Diaptomus, color 90

distribution, seasonal 89

oxygen 107

sicilis. distribution 107-8

siciloides, succession 91

Enteromorpha, habitat 33, 6S

Bntomostraca, see Crustacea.

Eucalia, death 99

distribution 99

Evaporation 14

Fauna, character 100-4, 110

changes ilO

comparison with fresh water 100-1

source 110

Fish, experiments 98, 106

disappearance 98

Filinia longiseta, relation to environ-
ment 86

Flora, character 101-4

changes 110

comparison with fresh water 101

source 110

G

Gomphosphaeria, see Coelosphaerium.

H.

Hyallela azteca, reactions 93

Hydrogen sulphide 31-2

Ice, thickness

work

Insects, characteristic forms

comparison with fresh water.

15
10
94
94

L.

Lake C, life, comparison with other

lakes 55

Lake P, life, comparison with other

lakes 56

Lamoreau Lake, life, comparison

with other lakes 50

Light apparatus 19-24

penetration 21-4

reaction, see Plancton.

Lyngbya, distribution in complex 65

seasonal 64-5

relation to Crustacea 87-8

M.

Merismopedia, seasonal distribution.. 63
Micrococcus, seasonal distribution.... 58-9
Mission Lake, life compared with

Main Lake 49-50

Mites, distribution 98

Moina, distribution, seasonal 91-2

relation to oxygen 107

Mvxophyceae, relation to alkalin-
ity 60, 103

Nauplii, cycles 45

vertical distribution 45

mortality 45

relation to oxygen 107

Navicula, seasonal distribution 72

Nematodes, habitat 81

Nitrogen 32

Nodularia, distribution, seasonal 65

relation to salt content 54-6, 65-6

Page
O.

Oozo, animals in 106-7

Oxygen, cycle 30

methods 29-30

ooze 31. 106-7

production, factors in 29-30

P.

I'edalia feunica, seasonal occurrence SG
Plancton, amount, comparison with.

other lakes 104-5

comparison of lakes and rivers 104

distribution, horizontal, irregular-
ity in 36-8

imrestricted 107-9

vertical 41, 43

production, factors in 42-4

variation in 43-4

reaction to light 46, 48

species, rt'presentative 45

Precipitation 13

Protozoa, distribution, unrestricted.... 108

environment, relation to 75

habitat 75

relation to, oxygen 107

species, comparison vrith fresh

water 75, 101

vitality 75-6

Page
R.

Rotifers, adaptability 82

amount, comparison with other

lakes 104-5

distribution, seasonal 81-2

habitat „ 81

life cycle _ 45

reproduction 81

species, comparison with other

waters 82

Run-off 13

Ruppia 74

S.

Sarcina, seasonal occurrence 59

Seepage 14

Species, new 105

Specific gravity 18

Stump Lake, life, comparison with

Main Lake 54

T.

Temperature 24-6

Thermocline, see Temperature.

Turbidity 19

W.
Wini 15