Full text of "Smithsonian miscellaneous collections"

See other formats

Spe TATE

a a Ne

iy vf

'
Ui
=i
ta
‘ t
Tibs Z
‘8
»
Al i pe \)
i RET NS ales
i in 3 i : pity ype Mi ; eh
OP AMAA ING Uh Dia) UR EATSY, cei ant Mg) A a uh
Pia aie
+. HG AM '
' re :
i ’ 1
1 i
Riya fy
Re ( prey y
i ie ; fil
hy tet
il aa 6 va iy
~ i ,/ Le a?
rn Oe 6 Sk ay hal

SMITHSONIAN
MISCELLANEOUS COLLECTIONS

VOE. «122

“EVERY MAN IS A VALUABLE MEMBER OF SOCIETY WHO, BY HIS OBSERVATIONS, RESEARCHES,

AND EXPERIMENTS, PROCURES KNOWLEDGE FOR MEN’’—JAMES SMITHSON

(Pustication 4219)

CITY OF WASHINGTON
PUBLISHED BY THE SMITHSONIAN INSTITUTION
1955

The Lord Baltimore Press

BALTIMORE, MD., U. & A.

HSOWN
AN 14 Vy

LIBRARY

The Smithsonian Miscellaneous Collections series contains, since the
suspension in 1916 of the Smithsonian Contributions to Knowledge,
all the publications issued directly by the Institution except the An-
nual Report and occasional publications of a special nature. As the
name of the series implies, its scope is not limited, and the volumes
thus far issued relate to nearly every branch of science. Papers in
the fields of biology, geology, anthropology, and astrophysics have
predominated.

LEONARD CARMICHAEL,
Secretary, Smithsonian Institution.

(iii)

\ ‘ f Ba yrs ‘estan ne
‘irae ela

10.

HOES

12.

13:

1A.

CONTENTS

Assot, C. G. Long-range effects of the sun’s variation on the
temperature of Washington, D. C. 14 pp., 5 figs. May 12,
1953. (Publ. 4131.)

Witson, Mitprep Stratton. New and inadequately known
North American species of the copepod genus Diaptomus. 30
pp., 58 figs. Aug. 4, 1953. (Publ. 4132.)

Snopcrass, R. E. The metamorphosis of a fly’s head. 25 pp.,
7 figs. June 25, 1953. (Publ. 4133.)

Assgot, C. G. Solar variation, a leading weather element. 35 pp.,
22 fies: Aus.\4, 1053. (Publ. 4153.)

Hoover, W. H., and Frortanp, A. G. Silver-disk pyrheliometry.
TO pp: 1 ng. Aug. 4, 1953: (Publ. 4136.)

Cuao, Hstu-Fu. The external morphology of the dragonfly
Onychogomphus ardens Needham. 56 pp., 50 figs. Sept. 15,
1953. (Publ. 4137.)

Bryan, Kirx. The geology of Chaco Canyon, New Mexico, in
relation to the life and remains of the prehistoric peoples of
Pueblo Bonito. 65 pp., 11 pls., 3 figs. Feb. 2, 1954. (Publ.
4140.)

WeErTMorE, ALEXANDER. Further additions to the birds of Panama
and Colombia. 12 pp. Dec. 17, 1953. (Publ. 4142.)

Snoperass, R. E, Insect metamorphosis. 124 pp., 17 figs. Apr.
I, 1954. (Publ. 4144.)

WHITTINGTON, Harry B. Two silicified carboniferous trilobites
from West Texas. 16 pp., 3 pls., 1 fig. Apr. 22, 1954. (Publ.
4146.)

Ciark, AiLsa M., and Crarx, Austin H. A revision of the
sea-stars of the genus Tethyaster. 27 pp., 12 pls., 2 figs. Apr.
8, 1954. (Publ. 4147.)

Roru, Louis M., and Wi11s, Epwin R. The reproduction of
cockroaches. 49 pp., 12 pls. June 9, 1954. (Publ. 4148.)
Aszot, C. G. Washington, D. C., precipitation of 1953 and 1954.

4 pp., I fig. Apr. 20, 1954. (Publ. 4170.)

Concer, PauL S. A new genus and species of plankton diatom
from the Florida Straits. 8 pp., 4 pls. July 15, 1954. (Publ.
4171.)

(v)

ih oe

yy ,
+a b

al Roe

ica @
Loli ]

SMITHSONIAN MISCELLANEOUS COLLECTIONS
VOLUME 122, NUMBER 1

Roebling Fund

EONG-RANGE EFFECTS OF THE SUN'S
VARIATION ON THE TEMPERATURE
OF WASHINGTON, D. C.

BY.
Cc. G. ABBOT

Research Associate, Smithsonian Institution

(PusicaTion 4131)

CITY OF WASHINGTON
PUBLISHED BY THE SMITHSONIAN INSTITUTION
MAY 12, 1953

The Lord Baltimore Hress

BALTIMORE, MD., U. 8 Ac

Roebling Fund

LONG-RANGE EFFECTS OF THE SUN’S VARIATION ON
Pie TEMPERATURE OF WASHINGTON; D.C.

By C. G. ABBOT
Research Associate, Smithsonian Institution

In a closely knit series of four recent papers ' I have shown (1) that
the sun’s output of radiation varies regularly in 23 periods, all in-
tegrally submultiples of 222 years; (2) that customary methods of
tabulating weather records, giving normal values therewith, are faulty
for computations of periodic terms because the normals are taken as
a whole, without segregation of times of high and of low sunspot
frequency ; (3) that with proper normal values and attention paid to
phase changes, depending on the seasons of the year and on the sun-
spot frequency, the precipitation at Peoria, Ill., shows plainly control
by the regular periodic variations of the sun; (4) that similar control
by solar variation is to be found in the precipitation at Albany, N. Y.

Since the variation of the sun operates primarily and directly on
the temperature of the atmosphere, and only indirectly on precipita-
tion, it seemed probable that a study of temperature might show even
more perfect control by solar variation than does precipitation. I
therefore take up in the present paper the temperature of Washington
in relation to the 23 known regular periodic variations of the sun’s
output of radiation. As in the Peoria and Albany papers, I employ,
for the most part, the monthly mean values published in the three
volumes of ‘World Weather Records,” but supplement these by U. S.
Weather Bureau publications since 1940.

As I have shown, in Smithsonian Publication No. 4090, that the
normals customarily published are misleading for my purpose, I com-
puted new normals as follows, suited to high and low sunspot activity.
I chose as the dividing line a Wolf sunspot number of 20. The tem-
peratures which follow are in degrees Fahrenheit. .

Jan, Feb. Mar, Apr. May June July Aug. Sept. Oct. Nov. Dec:

3. P.<20.... ery MEY Eol sles, 53-9 64.0 72.5 se 74.1 68.1 56.9 46.0 36.4
PE 2Oswcis 3350 . 35:60 43.7 Bata 03:0) yee ny OSes OLS S6sa00) 45-5) 0 35-3

1 Smithsonian Misc. Coll., vol. 117, No. 10 (Publ. 4088) ; No. 11 (Publ. 4090) ;
No. 16 (Publ. 4095), 1952; vol. 121, No. 5 (Publ. 4103), 1953.

SMITHSONIAN MISCELLANEOUS COLLECTIONS, VOL. 122, NO. 1

2 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. I22

From these normal temperature values I computed departures, ex-
pressed in tenths of degrees, for all months available from 1854 to
1939. There is a gap in “World Weather Records” of Washington
temperatures through 1860 and 1861. To avoid embarrassment by
large jumps of temperature from month to month, I computed 5-
month running means of the departures. That is, for March use
Jan.-Reb.- Mare Apr-May , and similarly for all months.

From these smoothed temperature departures I computed the effects
of the 23 regular periodic variations of the sun’s output of radiation,
employing only the interval from 1854 to 1939. For I wished to use
these results to forecast the behavior of Washington temperature from
1940 to 1951, and to compare such forecast with the actual event.
Obviously it is not to be hoped to find in such a manner very close
agreement between forecast and event, because of the complexity of
the earth’s surface and the turbulence of the atmosphere. But if it
can be shown that a general forecast of seasons, whether they are to
be on the whole warm or cold, wet or dry, can be made with reason-
able success for 10 years in advance, it would be of inestimable value
to people in many walks of life.

As was shown in the studies of precipitation at Peoria and Albany,
changed atmospheric conditions at different seasons of the year and
at different activity of sunspots displace the phases of the terrestrial
responses to solar variations. The same holds true for the temperature
of Washington. In short, the amplitudes and forms of the marches
of terrestrial responses to the regular periodic solar variations do not
alter greatly, though of course affected by interference of all other
periodicities. But the phases of the terrestrial curves shift from
season to season and alter with sunspot activity. It is not possible to
subdivide the data sufficiently to follow all these phase changes ac-
curately. I have contented myself with separate tabulations for three
seasons, viz: January to April—May to August—September to De-
cember ; and with two states of sunspot activity, viz: S. P.<20,
S. P.>20 Wolf numbers.

The method of tabulation follows closely that used in the study of
precipitation at Albany. Readers are referred to Smithsonian Publi-
cations No. 4095 and No. 4103 for information as to this method. I
have gone still farther in the direction of the modifications of Peoria
procedure as used at Albany, so as to strengthen the mean values in
the Washington temperature tabulations. For, before taking means,
I have shifted to a common phase the phases of all six mean tabula-

NO. I TEMPERATURE OF WASHINGTON, D. C.—ABBOT 3

tions for the three seasons, and for the two intervals, 1854-1899, and
1900-1939, with all 13 periods up to 153 months. At Albany only seven
periods were thus treated. 1 have also cleared every long period from
224 months to 91 months of overriding shorter periods, which are
integral submultiples of these long periods. In this way it was found
unnecessary to use periods longer than 454 months, for all the ampli-
tudes of still longer periods were produced by overriding shorter ones.
The 20 periods actually used for Washington temperatures were as
follows, expressed in months:

4%, 5%, 6-1/15, 7, 8%, 96, 93, 10-1/10, 10-6/10, 113, 13-1/10, 13-6/10, 15%, 224,
243, 274, 305, 345, 383, 453.

To illustrate the points brought out above I give several figures.

Figures I and 2 relate to the period of 13.6 months, as tabulated in
tables 1 and 2. Figure 1 and table 1 cover the times when Wolf sun-
spot numbers exceeded 20, and figure 2 and table 2, the times when
these were below 20. As usual, for periods of less than 224 months
tables 1 and 2 each comprise six independent subordinate tables, which
I am accustomed to designate as Ai, Az; Bi, B2; Ci, Cs. Subscripts 1
and 2 relate, respectively, to times before and after 1900 in the span
of years 1854 to 1939. Letters A, B, C, relate, respectively, to the
months January to April, May to August and September to December.
Symbols ok, +, | indicate whether curves were unchanged, moved
earlier, or moved later in their phases before taking means marked M.

In the 13.6-month tabulation for sunspots>20, the subordinate
tables have the following numbers of columns:

Designation : Ai Aa Bi Ba Ci (Oo
No. of columns: 6 5 7 8 5 6

Without giving dates of beginnings of columns or the temperatures
found in the individual columns, and recalling to the reader that, in
order to keep average lengths exactly 13.6 months, certain tempera-
tures are duplicated so that the columns as tabulated are 14 months
long, I now give in table 1 the mean values for A,, A2; Bi, Bz; Ci Cz
and their departures from the averages of these mean columns. The
means and the departures are stated in hundredths of a degree Fahren-
heit.

The columns of departures from table 1 are plotted in figure 1 with
the appropriate letters. Along with their letters are given symbols
ok, 4, or |, to show what shifts of phases were required to bring the
six curves to a common phase. In table 3 these changes of phase are
made, and the mean of the departures is taken as thus arranged. This
mean of departures is always employed, but reduced back to its proper

Means

Fic. 1—Sunspots>2o0

scissae, months.

numbers.

Six determinations of the periodicity of 13.6 months and their mean at
uniform phases in each figure. Ordinates, hundredths degree Fahrenheit. Ab-
Other symbols explained in the text.

Fic. 2.—Sunspots<20
numbers.

TABLE 1.—Illustrating tabulation for 13.6 months. S.P.>20

Ay
Mean Dep.
AS) tye
—I123 —123
Of 07
ma 43 nd
+ 32 + 32
+135 +135
“+-135 -+135
+125 +125
ele lay,
Boe, ie Ge
cates) Ls,
Se AD AS
aa Dona BOO
— 78 — 78

A
Mean

Dep.
ATi (5
17 + 29
Slo4-0 05
49° 4, 3
2I + 25
BN oi
59 =113
Slo at ed
44+ 2
93, =" 47
36 + I0
Sh rie us
61 — 15
81 — 35
46

Cy
Mean
—16 +17
—28 + 5
34) ak
—44 —II
= 5017
== S421
—Io +23
42: >=) 9
7-28 + 5
+20 +53
GOL ats
—72 —39
+20 Ae,
ar)

Dep.

As
Mean Dep.
+ 38 + 57
+ 8+ 25
Se ey
— 84 — 65
— 6— 77
— 88 — 60
— 32 — 13
— 20 — I
— 94 — 75
— 64 — 45
+ 74 + 93
+ 78 + 97
+ 30 + 49
— 2+ 17
—— LO)

Bs
Mean
=30"=—89
—2I —30
—II —20
Slee
—I7 —26
+9 O
+11 + 2
PIO ee 7
+25 +16

Dep.

Co
Mean Dep.
— 20 —87

o —67
70 = s
--, 00 4-23
+100 +33
+120 +53
“157 50
+105 +38
+118 +51
+ 62—5
ces ae
=~ Sal =e
+ 42 —25
Ses) eae
FOF.

NO. I TEMPERATURE OF WASHINGTON, D. C.—ABBOT 5
phase status in the syntheses to be described below. It is used instead
of the individual columns of departures given in table 1, because it
rests on 37 columns of temperatures, instead of on 5, 6, 7, or 8
columns, like the individual sets of departures in table 1. The reader
should recall that nearly 20 other periodicities have their effects upon
the columns of temperatures used to determine the periodicity of 13.6
months. Hence it is highly desirable to screen out these interferences
by numerous repetitions of the temperature columns.

The final mean of departures is graphed in the heavy line, M, of
figure I.

With this explanation of figure 1 and tables 1 and 3, it will not be
necessary to explain in detail figure 2 and tables 2 and 4. But it is
interesting to point out that the two heavy curves, M, of final columns
of mean departures, plotted in figures 1 and 2, are very similar in
form and amplitude but differ in phase, and that they are derived from
wholly independent groups of temperatures, one group coming solely
from times when Wolf sunspot numbers exceed 20, and the other
when these were below 20.

In the 13.6-month tabulation, table 2, for sunspots<z2o the sub-
ordinate tables have the following numbers of columns:

Designation zi Ai Az Bi B: Ci C,
No. of columns: 8 6 4 5 6 6

The mean of departures shown in table 4 therefore rests on the
temperatures contained in 35 columns summarized in table 2.

TABLE 2.—I/lustrating tabulation for 13.6 months. S.P.<20

Ay By Cy As Bs Cy
Mean Dep. Mean Dep. Mean Dep. Mean Dep. Mean Dep. Mean Dep.
—4o0+ 3 + 7+10t +2-+2 ¢163 +109 —14 —37. +123 + 6
—B8r +12 — 57 + 37 +40 -++40 + 80+ 26 +80+57 +148 +31
—26+19 —87+ 7 +20+20 + 12—42 472+49 +155 +38
—i1r1+ 32 —r182— 8 +15 +15 —13— 67 +48 +25 +132 +15
—I12+ 31 —225 —I3I +23 +23 —13— 67 +84 +61 +110 — 7
—24+ 19 —202—108 +78478 + 7—47 +80+57 +105 —12
— 36+ 7 —18— 86 +50+50 —22—76 +42+19 + 78 —39
— 56— 13 —I90— 96 +50-+50 + 28— 26 +428+5 + 67 —50
== 60 — 17") —157 — 63). +23 '-1-23 O— 54 +78 +55 + 48 —69
—15+ 28 —11I0—16 —47—47 +105+51 +18—5 + 78 —39
—51— 8 —50+54 —87—87 +105 +51 —44—67 +103 —14
—74— 31. +15+109 —72—72 +100+ 46 —72—95 . +165 +48
—95— 52 +25 +119 —60 —60 + 94+ 40 —16 —39 +138 +21
—74— 3 -+- 70 +164 —40 —40 +1io + 56 —62—85 +187 +70

Means — 43 — 94 00 + 54 +23 +117

6 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122
In tables 5 and 6, and figure 3 I give the evidence which shows that
it is unnecessary to employ the periodicity of 544 months in Washing-

ton temperature. As usual, I employ the symbols A, and A, to indi-

TABLE 3.—Combined table for 13.6 months. S.P.>20

As cok Biok Book Cyok Cos Mean
45 +57 +5 a0 aek7, —I7 —5
53 +25 +29 30 ro —25 — 8
— 78 +7 +15 —20 — I —32 —18
igs —65 —3 —45 —It —87 —59
—I123 —77 +25 —26 — 8 —67 —46
— 8&7 —69 +15 0 —I7 + 3 —26
— 43 —I3 —I3 + 2 —2I +23 —IlI
+ 32 ++ 2 aps 7 +23 +33 +17
+135 —75 +2 +16 —9 +53 +20
+135 —45 —47 +15 +5 +50 +18
+125 +93 +10 +63 +53 +38 +64
+127 +97 +15 +50 —3 +51 +64
+ 52 +47 —25 +18 —39 = 5 8
— 13 +17 —35 —14 +7 —I2 — 8

Ave Asok Byok Bate onerpenn (ats Mean
—17 +109 +101 +19 +60 +38 +50
+28 + 26 137 +5 +50 +15 +27
=e — 42 7 +55 +23 aay +5
—3I — 67 — 88 —5 —47 —I2 —42
—52 — 67 —I31 —67 —87 —39 —74
—3l — 47 —108 —95 —72 —50 —67
+3 — 76 — 86 —39 —60 —69 —55
+12 — 26 — 96 —85 —4o —39 —46
+19 — 54 — 63 —37 +2 —I4 —25
+32 + 51 —= 06 +57 +40 +48 +35
+31 + 51 + 54 +49 +20 +21 +38
+19 + 46 ++ 109 +25 +HI5 +70 +47
+7 + 40 + 49 +61 +23 + 6 +31
—I3 + 56 +164 +57 +78 +31 +62

cate results from temperatures recorded before and after 1900, during
the interval of years 1854 to 1939. In table 5 the column A, is the
mean of eight columns and the column A, of nine columns. The
departures shown in these tabulations having been plotted in figure 3,
the tabulation of A, discloses the presence of the overriding periodicity
of 9¢ months, approximately one-sixth of the 544-month period.

To eliminate it, the departure values in column Ag, table 5, were
arranged in six columns and their mean taken as shown in table 6.

NO. I TEMPERATURE OF WASHINGTON, D. C.—ABBOT 7.

These mean departures, repeated six times, are given in table 5, and,
being subtracted from column Az, give the departure column A*,. The
values A1, are plotted in figure 3, and show great similarity in form
and phase relations to the departures A,. So the mean of A, and A,
is taken in table 5, and plotted in figure 3. It is now obvious that the

Fic. 3—The 544-month periodicity, cleared of superriding periodicities, as
explained in the text.

curve has an overriding periodicity of half its length. Hence the mean
of departures of columns A, and A}, is analyzed for a periodicity of
(544-2) months, yielding the results shown in table 6 and repeated
end to end in table 5. Subtracting from the values given in the next
preceding column, and plotting the remainder in figure 3, it is now
obvious that only the effect of the overriding periodicity of 114, or
approximately one-fifth of 544 months, remains. Hence it proves
unnecessary to employ the periodicity of 543 months at all in the
synthesis of Washington temperatures. Similar steps eliminate the
periodicities of 684 months and 91 months from consideration.
Weare now prepared to test the usefulness of the 20 periodic terms
which have been worked out in the Washington departures from

Means...

TABLE 5.—Clearing the periodicity of 544 months of overrides

Mean

Dep.
Ay

“ts

|
to
ON

|
w&
Ny

Mean

Dep.

— 49
— 4
mae
— 50

ite

ia et pe

edleetes|
oR Go

The
93M.

A
Aly
oo
—26
—40
= 55
ss 58
—86

Mean
A; & Al,

—I2
—I3
—II
==45
—45
=95
—38
ae.
tar
+40
+62
+37
— 2
+ 6
— 8
—I4
—29
Seis
+48
+64

The

27iM.

—I9
—24
—I19
—40
—44
—63
=o
me)
TE

NO. I TEMPERATURE OF WASHINGTON, D. C.—ABBOT 9

TABLE 6.—Periodicities 94 and 274M

The 93-month periodicity from departures A»

Mean
— 49 + 2 — 32 +11 + 6 oO —I0
anes + 47 ao —32 + 32 —I4 Stas
re! + 89 +101 +1 + 29 +15 +36
as +118 + 46 == ae 9 =H +22
— 50 + 59 + 29 —43 + 55 +30 +8
a 7o ame + 41 —48 +106 +25 +8
—="03 — 44 + 17 —35 + 45 —2 —19
—102 — 23 — iI —53 + 58 —26 —24
— 53 — 19 + 18 —26 — 14 —40 —22
The 27}-month periodicity from mean A; and Al,

Mean

—I2 —26 —I19

—I13 —36 —24

—II —28 —I19

—45 55 —40

—45 —43 —44

—98 —29 SOs

0D —I2 53

—84 —28 —56

O80 ae hs

ah fe} == 3

+21 —2I fe)

+40 7109 O

+62 +15 +38

=1-37 a4, +42

= 3 37 +18

aa +62 +34

= +38 +5

—14 +45 +15

—29 +32 +1

— 4 she. +2

+48 +19 +33

+64 +63 +63

+64 +21 +42

+54 2 +28

+43 aie +16

+17 —29 — 6

1-16 —45 a9

normal temperatures for the interval from 1854 to 1939. It is pro-
posed to synthesize the results in such a manner as to forecast the
march of temperatures from 1940 to 1951, 12 years, and to compare
this synthetic forecast with the event. As the departures in monthly
records used for the interval 1854 to 1939 were smoothed by 5-month
running means, it is proper to compute the monthly departures from

Io SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

the same normals over the period 1940 to 1951, and to smooth these
departures also by 5-month running means.

As it is common knowledge that the temperature of eastern United
States has been gradually rising for the past century, it is highly prob-
able that we shall find that the departures from our normals, which
I computed from records of 1854 to 1939, will be prevailingly plus
during the interval 1940 to 1951. On another account it is also unlikely
that the scale of the synthetic summation of the effects of 20 periodic-
ities will be exactly the scale of our normal values. For the accumu-
lation of such inaccuracies as have resulted from computing depar-
tures from averages of 20 means, such as are shown in tables 1 and
2, must almost infallibly result in a plus or a minus departure in the
synthesis. Hence, on both accounts, just mentioned, we can expect
that there may be a systematic difference in level between the synthe-
sis and the event for the years 1940 to 1951.

Furthermore, as appeared in the study of the precipitation at Peoria
and Albany and, indeed, in the tabulation of Washington temperatures,
in comparing results before and after 1900, there are encountered
brief, as yet unpredictable, shifts of phase between synthesis and event
in the study of the control of weather by periodic solar variations.
Therefore we are to expect not only some systematic difference in
scale level between the synthesized forecast and the event in Washing-
ton temperatures from 1940 to 1951, but we may also expect occasional
brief unpredictable shifts of phase between the predicted and observed
results. With these remarks we preface the results obtained.

In table 7 I give a sample of the synthesis covering only part of
the year 1940. Figure 4 shows in the thin full line the synthesis, and
in the thick full line the event, for the years 1940 to 1951. The system-
atic difference in scale referred to above amounts to 3.0 degrees
Fahrenheit, the synthesis being lower than the event. It has been
removed in the thick dotted line by a flat addition of 3.0 degrees to
the synthesis, in order that attention might not be diverted from the
comparative marches of the two curves. That is the real test of the
method. In figure 4 the lighter line represents the synthetic forecast,
as computed after the manner of table 7. It is apparent that the princi-
pal features are found in the curves both of forecast and event. But
throughout the 12 years the event runs behind the forecast by several
months. From 1940 to May 1941 the lag is 4 months. Thence, in
the long interval to July 1948 the lag holds steadily at 3 months.
Thence to October 1951 it is only 2 months. In the dotted line I have
made these indicated shifts of phase, retreating the features of the

wonN — we et
aentass
++I141/i
mo eonotet
wm Ss = SS Oe
+ ++4+++4++

AN},
ia,

HIN
Na
FAP

20 periodicities deter-
ed; dotted curve, pre-

Se a |
gil al

aa a ee shot
° aes Fey
t+HEQaca Aa S
SI eae a

7. a
ahi ; atte
eS NALA i
WAV EAC TE

Aaa Ld; A PY

Sat itiey ‘Mie M3 @Ae VI SAN AS
NN | \ NW
eS
, 9) Ae ee — ies

—+—
MO{) 71 7
Fic. landed and observed temperature departures for Washington, D. C., years 1940 to I95I.
"Heavy curve, observed ; full taht curve, pio ep dotted curve, pre-

mined from records, 1854 to 1939. Ordinates, degrees Fahrenheit. Abscissae, years.
dicted, altered in phases and scale as described in the text. All temperatures smoothed by 5-month running means before used.

gZ —
foe

‘saanjpapgap aanqosaguay uozburysp 44 ryporsag fo Sisayjuts fo ajguvs p—Z alav

f9—
vr—
of—
61—
bz—

gi—
¥lz

Ll—
Ig—
Leta
|e
6
1 --

fre

¢zz

gI—
ZS

41+

ox €r

(0)
Si+
gh
eas
qIl

eb—
Iz—
4 —
z+
1€+
19+

oro he

br—
1Z—
SI—
go-F
ZI+
p+

9To I

€z—
ei.
€z+
Lp
S9-+
ge+
6

Sz—
Ocam
vi—
So-\-
9S++
ze-+
8

“4yDqJ 2a4bap D fo syIpapuny ui Sadnpavgap aanyosaguiay, “syjuou ur spor4sag

“* oun
oe ARN

‘aidy

"Tey

‘Pa
‘uel
oré1

I2 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

heavy dotted curve of the forecast, as just indicated. Thus the fore-
cast in the dotted line can readily be compared with the heavy full
curve of the event.

To gather more data on the sporadic changes of phase, as yet un-
predictable, I synthesized the periodicities from 1934 to 1939 and
compared the synthesis with the event. I was surprised to find that
in this interval, when, as one might say, the synthesis should be tailored
to fit the event, there was less satisfactory accordance than in the fore-
casted interval, 1940 to 1951.

From September 1936 to September 1938 synthesis and event are
exactly in the same phase. From September 1938 to January 1940
immediately preceding my forecast, the synthesis goes ahead of the
event by 3 months, as it does in most of my forecasted interval, but is
not yet 4 months, as immediately followed in the interval January
1940 to May 1941.

The scale level of the synthesis from 1934 to 1939 lies about 3
degrees below that of the event, as it did later, through most of the
interval from 1934 to 1939, but less in the months nearer 1934. If
the causes of the changes of level and of phase in these comparisons
could be unraveled and such changes predicted, a very great advance
in meteorology would ensue.

I think it can hardly be denied that there is a similarity between the
main features of the 12-year forecast and of the event. This simi-
larity is especially strongly marked in the rise of temperature from
1940 through 1941, though marred by the excessive rise of forecasted
temperature at the end of 1940. The similarity is even more striking
from May 1948 to December 1950, 8 to 11 years after the forecast
began. But here an additional systematic difference of about 1 degree
in level is seen.

There are many who are so impressed by the elegance of the meth-
od of correlation coefficients as an index of the worth of a forecast,
that they are contemptuous of curve comparisons as a test. To me
this seems unfair and misleading. For instance, old water mills used
to employ tooth and pin gears, irregularly made by ordinary carpenters
and having large and variable amounts of backlash. There was really
100 percent correlation in the running of a pair of these gears. But
they were often out of step, owing to the combined effects of imperfect
spacing and wide backlash. Computed coefficients of correlation would
fall far short of 100 percent.

In the control of weather by solar variation, obvious and certain
though it is, the complexity of the earth’s surface and atmosphere

NO. I TEMPERATURE OF WASHINGTON, D. C.—ABBOT LS

causes variations in the lag of response to regular periodic variations.
Consequently, when it is quite obvious that a pair of curves of forecast
and event are related, a rapid rise or decline may be found in one curve
slightly in advance of the other. This causes large departures between
the two curves and may bring down the computed correlation coeffi-
cient to apparent meaninglessness. Mere obstacles to the free opera-

TABLE 8.—Forecast of Washington 5-month running mean Fahrenheit
temperatures, 1952 to 1959

1952 Jan. 39°1 1954 Jan. 34°4 TOSOMJian. 3727 1958 Jan. 38°8
Feb. 40.4 Feb. 35.7 Feb: 37.3 Feb. 41.0
Mar. 47.3 Mar. 43.4 Mar. 43.7 Mar. 47.1
Apr. 56.9 Apr. 54.5 Apr. 55.2 Apr. 55.8
May 66.2 May 64.8 May 64.1 May 64.7
June 72.6 June 73.1 June 70.6 June 71.2
July 75.7 July 78.4 July: 75).7 July 76.0
Aug. 73.9 Aug. 76.8 Aug. 74.2 Aug. 75.3
Sept. 66.2 Sept. 72.6 Sept. 68.8 Sept. 70.1
Och 15558 Oct6s53 Oct soa7 Oct. 58.8
Nov. 44.2 Nov. 51.7 Nov. 49.1 Nov. 48.6
Dec. 35.2 Dec. 40.8 Dec. 40.5 Dec. 37.8

1953 Jan. 35.3 1955 Jan. 37.4 1957 Jan. 37.6 1959 Jan. 35.6
Feb. 35.9 Feb. 37.4 Feb. 38.4 Feb. 36.6
Mar. 43.1 Mar. 45.1 Mar. 45.0 Mar. 43.3
Apr. 54.2 Apr. 54.3 Apr. 54.5 Apr. 54.6
May 63.7 May 61.8 May 63.7 May 65.6
June 72.4 June 70.3 June 70.8 June 74.3
July 78.9 July 75.9 July 75.5 July 77.8
Aug. 76.9 Aug. 74.3 Aug. 76.0 Aug. 76.2
Sept. 71.6 Sept. 71.4 Sept. 68.1 Sept. 70.4
Oct ores Oct. 26283 Oct 57.1 Oct) 50.2
Nov. 49.8 Nov.ic51.7 Nov. 47.7 Nov. 48.1
Dec. 38.90 Dec. 40.4 Dec. 38.2 Dec. 40.0

tion of a cause may, in the correlation method, so far obscure the cause
that it fails altogether of recognition as the cause. Yet, for practical
purposes, the forecast may tell the interested agriculturalist quite
nearly enough, in time and amount, the change which he wishes to
know in advance.

I regard the results of this test of forecasting Washington tempera-
ture as so promising that I have ventured to synthesize the expected
Washington temperatures from 1952 to 1959. This forecast is given
in table 8. These forecasted 8 years of Washington temperatures I
have reduced from the status of departures from normal to actual
temperatures Fahrenheit. In making the forecast I have assumed that
the lag between synthesis and event will be reduced to zero, and that

14 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. I22

the scale of mean temperatures will remain 2 degrees above synthesis,
as now prevailing. The comparison of forecast is to be with Weather
Bureau Records, means between averages of monthly maxima and
monthly minima, at the main Weather Bureau Office, 26th and M
Streets, NW., Washington, D. C. Obviously, to check the accuracy
of the forecast, the observed temperatures of future years must first
be smoothed by 5-month running means.

<2?) 1 ee!
FERRE Ay

_ pens sees! et 4
BEERSy USES! SESE s SERPs es eee Pee
aaee Prt fee Pe
ae ee

Fic. 5.—Washington temperature departures, 1950 to 1952, predicted (light
curve) and observed (heavy curve). Correlation, 50.4 = 9.7 percent. Tempera-
tures, degrees Fahrenheit. All temperatures smoothed by 5-month running
means before used.

To fix upon the probable scale difference and lag, I prepared figure 5,
in which departures from normal in the synthesis are plotted from the
upper zero line and the right-hand scale of ordinates. The departures
observed are plotted from the lower zero line and the left-hand scale
of ordinates. The plot begins with 1950 and extends through 1952.
A lag of one to two months is seen, as stated above, in the years 1950
and 1951, but seems to vanish in 1952. As for the scale, the synthetic
values seem to run about 2° Fahrenheit below the observed values in
these three years. So I have assumed that the same scale difference
and zero lag will continue till 1959, as stated above.

In view of unpredictable changes of scale and lag heretofore noted,
one hardly hopes that such changes will not occur before the end of
this forecast. I can hardly hope to live to see it verified to the end. It
is really a forecast for 20 years in advance, beginning with the year
1940. Considering that the basis of my forecast lies in records of 1854
to 1939, centering about 1900, one may even justly say that the fore-
cast, 1952 to 1959, is over a half century in advance.

For those who prefer correlation coefficients to graphs, figure 5 gives
a correlation coefficient of 50.4+9.7 percent with the scale difference
of 2° Fahrenheit removed.

SMITHSONIAN MISCELLANEOUS COLLECTIONS
VOLUME 122, NUMBER 2

NEW AND INADEQUATELY KNOWN
NORTH AMERICAN SPECIES OF
THE COPEPOD GENUS ‘DIAPTOMUS

MILDRED STRATTON WILSON

Arctic Health Research Center
U. S. Public Health Service
Anchorage, Alaska

(PusLicaTIon 4132)

CITY OF WASHINGTON
PUBLISHED BY THE SMITHSONIAN INSTITUTION
AUGUST 4, 1953

The Lord Baltimore Press

BALTIMORE, MD., U. 8 Ae

NEW AND INADEQUATELY KNOWN NORTH
AMERICAN SPECIES, OF. THE COPEPOD
GENUS DIAPTOMUS

By MILDRED STRATTON WILSON

Arctic Health Research Center
U. S. Public Health Service
Anchorage, Alaska

INTRODUCTION

In the preparation of a new key to the calanoid Copepoda for the
revised edition of Ward and Whipple’s ‘““Fresh-Water Biology,” some
new species of Diaptomus have been recognized and the status and
distribution of other species have been clarified. In order that these
new forms may be included in the key, the following diagnostic de-
scriptions and notes are presented. More detailed treatment is reserved
for the future monographic review of the North American species.

A considerable part of the present report deals with the species that
have in one way or another been confused with Diaptomus shoshone
Forbes. It became apparent early in the study of the subgenus Hes-
perodiaptomus that it would be necessary to establish the typical form
of D. shoshone before it and several closely related species could be
correctly separated from one another. All that remains of the orig-
inal collection, which is in the Illinois State Natural History Survey,
are slides consisting mostly of dissected appendages. These have
been found adequate to determine both the important and unknown
diagnostic characters of the type. Study of literature and other col-
lections, particularly the Marsh and Light accessions in the U. S.
National Museum, has shown that several definable forms can be
unqualifiedly separated from the typical. Two of these (D. caducus
and D. nevadensis) have already been distinguished by Light (1938),
though he was unaware that they had been included in published rec-
ords of D. shoshone. The others are herein described as new species.

The confusion of these species with Diaptomus shoshone has been
largely due to the fact that certain fundamental characters of the genus
have been neglected in the descriptions of North American diapto-
mids. Two of the most important of these are the setation of the
antennules of the female, and that of the left side of the male, and
the exact form of the left exopod of the male fifth leg. Both of these

SMITHSONIAN MISCELLANEOUS COLLECTIONS, VOL. 122, NO. 2

2 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. I22

characters are significant in the taxonomy of the subgenus Hesperodi-
aptomus, and particularly so in the case of D. shoshone and its allies.

The setation of the antennule was recognized as a fundamental
specific character by Schmeil (1896) in his comprehensive analysis
of the genus, and its invariability has been emphasized by Gurney
(1931, p. 114). The exception mentioned by Gurney has been clari-
fied by Kiefer (1932, p. 512). I have noted, in examination of nu-
merous American specimens, that anomalies sometimes occur in the
setation, but these are very rare and are recognizable as such because
they occur in isolated individuals of a sample and on only one anten-
nule of a pair in the female. Since the subgenus Hesperodiaptomus

TABLE 1.—Antennule setation in the subgenus Hesperodiaptomus
2 and left side d

Segment

Species and sex 2 6 10 1 LFol4e els b- 26) «. el eD,
COMUCHS. eos a Shes alas O Ol v4: 2. A one. Be 2 SB ee ee
hirsutus ...0.000.. fs 3 Ey aa aa a ae eh Oe

sere Bee fie enters ol ey as) Gates a ie a ee
shoshone ......0+. a ee Me ce Oe a Ls oa
novemdecimus ..... Orgies Lier Twenty and @i Pian Derren re
Va Te Dish [oad dey pod Peuitine tn iw Big oee
SCREG CTA waitin nese EF a Oe Oe ae ee On AOE ENE Rk oS
Ie 7 aS HO eee SR UR QR RN OS ala NE Bie or
NeVAdENSIS .....06. wal 3. kL a a te he
CVSOMU ABS ISES Po arb Met Wine tips Cy) pr eee
aycticusis. ty. BEA OSes ive, ele seo tay Bea ee

* schefferi—d left may occasionally have 2 on 6, as it always has on the right.

belongs to what has been termed the “multisetaceous” group, in which
the number of setae on segment I1 is 2, and on segments 13-Ig is
either I or 2, it is highly desirable that the setation of all the species
be known. It is therefore presented in table 1. The subgenus differs
from any group that has been recorded in literature in having a species
in which 4 setae are present on segment 2. In addition, some of the
proximal segments may have 2 setae rather than the customary single
seta.

The preliminary consideration that has been given in this study
to the structure and distribution of Diaptomus shoshone and the forms
closely allied to it suggests that the group as a whole may be valuable
material for studies in variation and distribution. Such studies not
only might contribute to the zoogeography of this group but also might
have wider application in the much-needed evaluation of structural
characters. Since the knowledge of variation and distribution is so

NO. 2 COPEPODS OF THE GENUS DIAPTOMUS—WILSON 3

incomplete, no reliable analysis can be made as to the systematic status
of these forms. It is therefore not superfluous to emphasize that it
is of much importance that published records of Diaptomus shoshone
and related forms be based upon accurate identification.

SYSTEMATIC DESCRIPTIONS
DIAPTOMUS (HESPERODIAPTOMUS) SHOSHONE Forbes
Figures 1-8

Diaptomus shoshone Forzes, 1893, p. 251, pl. 42, figs. 23-25—-ScHACHT, 1897
(in part), p. 141, pl. 26, fig. 3—MarsH, 1907, p. 431, pl. 28, figs. 2-5; 1920
(in part), p. 8j; 1929 (in part), p. 17—Dopps, 1915a, p. 102, fig. 9; 1915,
p. 290, fig. 65; 1917, p. 76; 1924, p. 4.

Diaptomus (Hesperodiaptomus) shoshone, Licut, 1938, p. 67.

Specimens examined.—The material studied consisted of slides in
the type (Forbes) and Schacht collections in the Illinois Natural His-
tory Survey; specimens from all the localities reported by Marsh for
which slides are available in the National Museum; unpublished rec-
ords in the S. F. Light accession in the National Museum; and a recent
alcoholic collection, consisting of 57 females and 31 males, collected
by J. S. Stanford, Dry Lake, Cache County, Utah. This latter collec-
tion agrees with the type material in the basic characters, which have
been checked on all the specimens.

The descriptive diagnosis and illustrations given here are based
upon the Forbes slides, except for the habitus, which has been made
from the Utah specimens inasmuch as it is desirable that study of
whole specimens be made whenever possible from those that have
been undistorted by cover-glass pressure. In all cases the term typical
form refers to the type lot (Forbes did not designate an individual
type specimen), or to individuals from other samples that agree with
the type in the basic characters. The common variations that have
been found are given in parentheses.

Diagnosis (emended).—Length (after Forbes) : Female, 3.1 mm. ;
male 2.59 mm. (Utah specimens, female, 3.7-4.4 mm.; male, 3.0-3.5
mm.). Greatest width in both sexes in the mandibular-maxillary re-
gion, the metasome tapering sharply so that the posterior portion is
noticeably narrower than the anterior. Metasomal wings of female
not produced outwardly, directed posteriorly, reaching to near the
end of the swollen portion of the genital segment (not bifid as stated
by Forbes). Urosome of female 3-segmented. Genital segment asym-
metrical, the lateral areas bearing the minute sensilla produced into
prominent lobes, that of the right side larger and directed backward.
Caudal rami subequal to segment 3, both margins ciliate. (The illus-

VOL. 122

ONS

ONIAN MISCELLANEOUS COLLECTI

SMITHS

opposite page.)

1-14.—(See legend on

Fics.

NO. 2 COPEPODS OF THE GENUS DIAPTOMUS—wWILSON 5

trations of the female urosome given by Marsh, 1907, and by Dodds,
1915a and 1915b, agree with Forbes’s description and slides and are
correct. That of Schacht, 1897, pl. 26, fig. 1, is not of D. shoshone.
It was found in checking slides in the Schacht collection that those
labeled shoshone female were of a leptodiaptomid.)

Antennules reaching to near end of metasome, setation of female
and left side of male identical, having 3 setae on segment 2, I seta on
6 and 10, Segments 11 and 13-19:

II 13 14 15 16 17 18 19
2 I 2 I 2 I 2 I

Right antennule of male with spines of segments 10, 11, and 13 not
grossly developed ; that of 10 less than width of segment, those of 11
and 13 hardly longer ; that of 13 a little longer than that of 11. (These
spines show considerable individual variation.) Segment 15 without
spinous process. Segment 16 with a long, distally placed process
(usually varying from about 30-42 percent of the length of the margin
of the segment ; I specimen examined has the extreme of 61 percent).
The process of segment 23 reaching to the middle of the last segment
(rarely beyond), its apex pointed (frequently rounded).

Maxilliped of both sexes grossly developed, with greatly enlarged
clawlike setae on the inner side of the endopod, the terminal and outer
setae much reduced in size. Leg 1 with the spine of exopod 1 long and
setiform, reaching to near the end of segment 2. Leg 2 lacking
Schmeil’s organ.

Leg 5 of female slender. Relative lengths of the exopod and endo-
pod of Forbes’s slide 507 (fig. 2):

Exopod 1 Exopod 2 (outer) Endopod Endopod setae
40 36 35 19:22

The endopod indistinctly segmented (or distinctly so), armed apically
with large, flat spinules; the setae unornamented (only a few speci-
mens show, at high magnification, scattered hairs on these setae;
Schacht’s figure is undoubtedly of shoshone, but the dense plumosity

Fics. 1-8.—Diaptomus (Hesperodiaptomus) shoshone Forbes: 1, Female,
metasome segments 5-6 and urosome, dorsal; 2, female, leg 5; 3, male, right
antennule, apical segments; 4, male, right antennule, segments 15-16; 5, male,
leg 5, posterior view; 6, male, leg 5, left exopod, segment 2, anterior view; 7,
ies antennule, setae of segments 1-6; 8, female, antennule, setae of segments
13-19.

Fics. 9-11.—Diaptomus (H.) novemdecimus, new species: 9, Female, leg 5;
10, male, leg 5, anterior view; 11, male, leg 5, processes of left exopod, anterior
view.

Fics. 12-14.—Diaptomus (H.) kenat, new species: 12, Male, leg 5, left exopod
and endopod, anterior view; 13, male, right antennule, apical segments; 14,
male, leg 5, posterior view, with detail of lateral spine.

6 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

of the endopod setae is an exaggeration). Length of the second exo-
pod segment three times its greatest width, the whole of the claw very
slender, evenly tapered from the area of the third exopod segment.
Setae of the third exopod short, the inner the longer.

Leg 5 of male, right: Claw swollen at its base (not divided as
shown by Forbes) ; shorter than the rest of the leg, about 25: 33.
Basipod I without inner lamella, sensillum on well-developed cutic-
ular prominence. Basipod 2 without prominent raised ridge or pro-
trusion. Exopod 2 nearly parallel-sided, with small spinule on inner
posterior face; lateral spine straight, shorter (or a little longer) than
the width of the segment. Endoped a little longer than the inner mar-
gin of exopod I (I- or 2-segmented).

Left leg reaching to about the middle of the right second exopod
segment. Basipod 2 with the proximal inner half of the anterior face
hardly protuberant (individually variable). Segment 1 of the exopod
considerably longer than segment 2 (about 3:2). Inner process of
distal segment a long, slender, tapering, distally directed spine whose
basal portion is hardly widened and which reaches to the end of the
outer process (or farther) ; its length more than half that of the outer
margin of exopod 2 (measured to the base of distal process). The
medial spinules of the distal pad very gross, those of the posterior
and anterior faces very small, arranged in groups. Endopod 1-(or 2-)
segmented.

Distribution —The type locality is Lake Shoshone, Yellowstone Na-
tional Park, Wyo. Other Yellowstone Park records given by Forbes
are: Lewis Lake, Yellowstone Lake, Swan Lake, and an alkaline pond.

It has been determined from examination of the Marsh slide col-
lection in the National Museum that only the following records pub-
lished by Marsh (1920, 1929) are of typical shoshone: Yellowstone
Lake, Wyo.; Corona, Irwin, and Pikes Peak, Colo.; Nioche Valley
and Salinas, Wasatch Mountains, Utah. The Toronto, Canada, record
is questionable ; the only slide available is of a cyclopoid. All the other
Marsh records and also those of Carl (1940) are referable to one or
another of the species discussed below.

The Light accession contains three unpublished records of typical
shoshone. One is an additional Rocky Mountain locality: A pond 28
miles east of Cooke, Mont., 9,000 feet, A. G. Rempel, collector. The
others are from the Sierra Nevadas of California: Iceberg Lake,
Madera County, 10,100 feet, P. R. Needham; Helen Lake, Fresno
County, 10,896 feet, H. J. Rayner.

Dodds, whose illustrations agree with the typical form, has pointed
out that in regions of the Rocky Mountains studied by him, Diaptomus

NO. 2 COPEPODS OF THE GENUS DIAPTOMUS—WILSON 7

shoshone was found only in lakes at very high altitudes (around 9,000-
12,000 feet). The elevation given for the type locality, Lake Shoshone,
was 7,740 feet (Forbes, 1893, p. 214).

On the basis of present knowledge the distribution of typical sho-
shone appears to be restricted. All the authentic records are from high
altitudes (6,000-12,000 feet) in the Rocky Mountains or Sierra Ne-
vadas. It is not intended to suggest here that this is proof of the
altitudinal distribution of the species. It should be pointed out, how-
ever, that a trend is apparent which is worthy of investigation and
which may have bearing on the zoogeography of this and some of the
species discussed below.

DIAPTOMUS (HESPERODIAPTOMUS) NOVEMDECIMUS, new species
Figures 9-11

Type lot.—Slides from the Light collection consisting of mounted
appendages of both sexes. Temporary pond, 2 miles south of Charlo,
Flathead Reservation, Mont., elevation about 3,000 feet, Gordon B.
Castle, April 28, 1940. Occurring with D. wardi. Type slide, U.S.N.M.
No. 94624.

Since only mounted appendages are available, no measurements or
description of the habitus can be given. The size of the appendages
indicates that the body size of both sexes is similar to that of D.
shoshone.

Diagnosis.—Antennule setation of female and left side of male: 3
setae on segment 2, I on segments 6 and 10. Segments 11 and 13-19:
II 13 14 15 16 17 18 19
2 I 2 I 2 I 2 2

Right antennule of male: Spines of segments 10, 11, and 13 not long,
that of 13 of irregular shape, longer than that of 11. Segment 15
lacking a process; 16 with a distally placed process, its length about
20 percent of that of the segment. Process of segment 23 stout,
straight, the apex pointed, reaching to near the end of segment 25.
Leg 5 of female: Exopod 1 (outer margin) a little longer than
exopod 2, 47:45. Exopod 2 wider at base than in typical shoshone,
the width to total length, 20: 45 ; the claw not so slender as in shoshone,
gradually tapered beyond the middle. Endopod 2-segmented (or
indistinctly so), longer than the inner margin of exopod 1. The inner
seta of exopod 3 subequal to, or not as stout or long as, the outer.
Leg 5 of male: Claw not swollen at base, nearly as long as the rest
of the leg, 37:35. Right basipod without modification or armature.
Right exopod 2 somewhat enlarged, with a blunt spinule on inner pos-
terior face. Right endopod 2-segmented, longer than exopod 1. Left

8 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

basipod 2 with prominent inner proximal protrusion; serrate cuticle
particularly conspicuous on inner anterior face (such a serration is
a rudimentary structure which may be present or absent in hespero-
diaptomids). Exopod 1 longer than exopod 2 (about 37:25). Inner
process a stout spine with a slightly widened base, reaching to near
the end of the distal process, its length a little less than one-half that
of the outer margin of exopod 1 (to base of distal process).

The trivial name of this species refers to segment 19 of the antennule,
which differs from that of shoshone in the presence of 2 setae rather
than 1 seta. The question of whether D. novemdecimus is a subspecies
of D. shoshone should be considered in future studies. The status of
species has been given here because of the antennular setation, which
has long been considered by competent authorities to be a stable spe-
cific character. In the several samples of typical shoshone that have
been examined no individuals of either sex have been found to have
2 setae on segment 19. In addition, the two can be separated by two
definable characters of the male fifth leg which also differ in other
species of Hesperodiaptomus—that is, the length and shape of the
claw and the size and shape of the inner process of the left exopod.
This combination of a pattern of close structural similarity and de-
finable differences in seemingly basic characters appears to make these
two species valuable for studies in the interrelationships of the hes-
perodiaptomid group and the problem of evaluation of characters.
Until adequate knowledge of variation and distribution is available,
it is my opinion that any attempt at subspeciation is both arbitrary
and premature.

DIAPTOMUS (HESPERODIAPTOMUS) KENAI, new species
Figures 12-17

Diaptomus shoshone, MarsH, 1920 (in part), p. 8j; 1929, p. 17—CaRL, 1940,
p. 81; ? 1944, p. 30.
? Diaptomus shoshone, THACKER, 1923, p. 88.

Type lot.—t1oo specimens of both sexes. Shallow mountain pond
on Palmer Creek Road, about 12.6 miles southeast of Hope, Kenai
Peninsula, Alaska, Charles S. Wilson, August 24, 1949. Occurring
with D. tyrrelli. Holotype female, U.S.N.M. No. 94632; allotype
male, U.S.N.M. No. 94633.

Diagnosis.—Length (dorsal view): Female, 2.03-2.08 mm.; male,
1.87-2.04 mm.

The wings of the last metasomal segment of female a little asym-
metrical, the lateral tip of each side drawn out, that of the right side
larger than the left. Urosome of female 3-segmented. Genital seg-

NO. 2 COPEPODS OF THE GENUS DIAPTOMUS—WILSON 9

ment symmetrical, without lateral protrusions. Third segment and
caudal rami subequal in length; the greatest width of the rami a little
more than one-half their length (21: 35), ciliate on inner margin.

Antennules of female reaching to near the middle of the genital
segment. Numerical setation: 3 on 2, I on 6 and Io, 2 on II, and 1
on 13-19. The seta of segment 1 short, not reaching to the end of
segment 2; all setae comparatively short, none reaching beyond the
middle of the succeeding three segments. Left antennule of male
with same setation as female.

Right antennule of male with the spines of 10, 11, and 13 thick,
none longer than the width of their segments, proportions to one an-
other, 11:16:21. Segment 15 without a process; segment 16 with a
distally placed process reaching beyond the end of the segment, its
length about 28 percent of the length of the segment. The process of
segment 23 reaching to about the middle of the last segment, straight
or outcurved.

Maxilliped not so grossly developed as in shoshone. Setiform spine
of exopod segment 1 of leg 1 only about half the length of segment 2.
Leg 2 lacking Schmeil’s organ.

Leg 5 of female: Exopod 1 (outer margin) a little longer than
exopod 2 (35: 33). Proportion of greatest width to length of exopod
2, about 15:33; this great width gradually decreased throughout the
length of the “claw” to near its apex where it may be abruptly nar-
rowed. The outer seta of exopod 3 always stouter and usually much
longer than the inner. Endopod 2-segmented, as long as, or longer
than, the inner margin of exopod 1. The apex more or less prolonged
on the inner side, armed with a few short spinules, the length of the
subterminally placed setae about half that of the endopod.

Leg 5 of male, right: Claw short, only a little longer than the exo-
pod ; its base hardly swollen. Exopod 2 with spinule on the posterior
inner face; the lateral spine short, characteristically incurved on the
inner side. Endopod about as long as exopod 1, I- or 2-segmented.
Left leg reaching to a little beyond the right exopod 1. First segment
of exopod a little longer than the second. The inner process a broad-
based spine, not reaching to the end of the distal process.

This species is the only one of the shoshone group that has the seta-
tion of the antennule reduced to one on segments 13-19. This char-
acter distinguishes it from all the others, and particularly from D.
caducus, to which it would appear to be most closely allied. The female
fifth leg has a distinctive shape given to it by the widening of the
second exopod segment. The inner process of the left exopod of the
male fifth leg is distinguished by the broadened base from which it
tapers to a slender spine.

SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

Lif,

S. 15-28.—(See legend on

opposite page.)

Fic

NO. 2 COPEPODS OF THE GENUS DIAPTOMUS—WILSON II

Distribution —tThe specimens reported as shoshone by Marsh from
the Pribilof Islands, Alaska, and from Wheat Meadows, Calif., are
referred to this species. Marsh is supposed to have identified the speci-
mens reported by the Thackers, but slides from their British Columbia
localities have not been found in the National Museum collections.
They have here been questionably referred to D. kenai because the
record falls within the distribution pattern of this species. Slides
labeled D. shoshone by Carl and reported in the British Columbia rec-
ords of his 1940 paper are in the Light accession and have been iden-
tified by me with D. kenai. In addition to these collections and the
type lot from Alaska, a large number of collections from Oregon and
California have also been examined and referred to this species. These
records are extensive enough to show that the species is not altitudin-
ally restricted. In Oregon and California it is rare on the coast but of
frequent occurrence in the Cascade and Sierra Nevada mountain ranges,
where it has been collected to an elevation of 9,000 feet. The species
also occurs in the Cascades of Washington, having been found in a
collection from Lake George, Mount Rainier National Park, referred
to me by C. C. Davis.

DIAPTOMUS (HESPERODIAPTOMUS) CADUCUS Light

Diaptomus caducus Licut, 1938, p. 67, figs. 1-5, 23.
Diaptomus shoshone, MarsH, 1929 (in part), p. 17.
Diaptomus sicilis, CARL, 1940 (in part), p. 81.

Specimens reported by Marsh from Vancouver Island, British
Columbia, are referable to this species. A slide labeled D. sicilis by
Carl is present in the Light collection; the locality given is: Pond,
Victoria, British Columbia. This specimen is clearly identifiable from
the antennular setation as caducus.

Diaptomus caducus has been adequately described by Light, and if
proper attention is given to the highly important setation of the anten-

Fics. 15-17.—Diaptomus (Hesperodiaptomus) kenai, new species: 15, Female,
metasome segments 5-6 and urosome, dorsal, with detail of right “wing”; 16,
female, leg 5, with detail of lateral setae of exopod; 17, female, antennule, setae
of segments 13-10.

Figs. 18-25.—Diaptomus (H.) hirsutus, new species: 18, Female, metasomal
wings and distal portion of urosome, dorsal; 19, female, leg 5, with detail of
lateral setae of exopod (type lot); 20, female, leg 5, endopod with elongate
setae (lot from Eldorado County, Calif.) ; 21, female, antennule, segments I-10,
showing setae of segments I, 2, 3, 6, and 10; 22, male, leg 5, posterior view; 23,
male, leg 5, profile of protrusions of right basipod segments; 24, male, right
antennule, apical segments; 25, male, leg 5, detail left exopod and endopod,
anterior view.

Fics. 26-28.—Diaptomus (H.) nevadensis Light: 26, Female, metasome seg-
ments 5-6 and urosome, dorsal; 27, female, leg 5, with detail of apex of endopod,
2 different views; 28, female, leg 5, detail of lateral setae of exopod.

I2 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

nules (table 1) there should be no confusion of this species with any
other known hesperodiaptomid. The species is unique among known
diaptomids in having 4 setae rather than the usual 3 on the second
segment in both sexes, including the right antennule of the male. Oc-
casional specimens have been found in which 2 setae are also present
on segments 4-7 on one antennule of a pair, apparently an anomaly
rather than a variation of the species.

Present knowledge of distribution confines this species to the Pacific
coast areas from central California to British Columbia, where it
characteristically occurs in temporary ponds and roadside ditches.
The one mountain record given by Light is referable to the new species
described below.

DIAPTOMUS (HESPERODIAPTOMUS) HIRSUTUS, new species
Figures 18-25
Diaptomus caducus LicHtT, 1938 (in part), p. 69.

A single female from Granite Lake, Amador County, Calif., was
incorrectly assigned to caducus by Light. Another collection made
at a later date from the same locality yielded numerous specimens and
has been made the type lot of the new species.

Type lot—100 specimens of both sexes. Granite Lake, Amador
County, Calif., 6,800 feet, June 22, 1937, R. E. Smith. From Light
collection in the U. S. National Museum. Holotype female, U.S.N.M.
No. 94628; allotype male, U.S.N.M. No. 94629.

Other California mountain collections in the Light accession refer-
able to this species are: Several ponds in Lassen National Park; pond
at Columbia, Sierra County ; pond near Summit Lake, Shasta County ;
Silver Ford, Eldorado County.

Diagnosis —Length : Female, about 1.88 mm. ; male, about 1.79 mm.
Greatest width in both sexes in the middle of the cephalic segment,
that of the female about 28 percent of the length. Posterior margin
of metasomal wings of female slightly bifid, the outer portion pro-
duced laterally. Urosome of female 3-segmented. Genital segment
symmetrical, not swollen laterally. Caudal rami shorter than segment
3 (about 29: 35) ; their width about 71-76 percent of the length ; both
margins and entire dorsal surface hairy. Urosome of male symmet-
rical, length of caudal rami subequal to segment 5, with hairs on the
inner margins only.

Antennules reaching nearly to end of metasome. Those of female
having 3 setae on segment 2, I on 6, 2 on 10, 11, and 13-19; seta of
3 unusually long, reaching about to end of segment 10. Left antennule

NO. 2 COPEPODS OF THE GENUS DIAPTOMUS—WILSON 13

of male differing from that of female in having 1 seta on segments 10
and 13 (2 on 11 and 14-19), seta of 3 reaching about to segment 8.
(This sexual difference found in all the lots of specimens examined.)
Right antennule of male: Length of spines of segments 10 and 11 less
than the width of their segments and shorter than that of 13, which is
exceedingly slender and reaches about to the middle of segment 14;
segment 16 with a short, distally placed process not reaching beyond
the end of the segment. Process of segment 23 spatulate, its tip always
rounded, reaching to the end of 24 or to the middle of 25. Leg 2
lacking Schmeil’s organ.

Leg 5 of female: Endopod 2-segmented, as long as, or longer than,
exopod I ; apex truncate, with few spinules; setae subequal in length
to endopod or longer.

Leg 5 of male, right: Claw subequal to (or a little longer than)
the rest of the leg. Basipod 1 without prominent inner lamella. Basi-
pod 2 with raised ridge on posterior surface, produced proximally
into a rounded lobe (fig. 23 shows profile of ridge, without pressure) ;
this structure reduced to indefinite shape by cover-glass pressure (fig.
22). Exopod 2 with small spinule on inner posterior face, lateral
spine a little shorter (or longer) than width of segment. Left exopod:
Segment 1 a little longer than segment 2, 19:15. Inner process a
tapered spine with a narrowly expanded base. Distal pad with minute
spinules arranged in groups on the anterior side ; those of the posterior
side larger and thickly set, extending far up on the face of the segment.

The trivial name of this species refers to the presence of hairs on
the dorsal surface of the caudal rami of the female, a condition un-
usual in Diaptomus. D. caducus has been found to have hairs on both
surfaces of the caudal rami, but they are few in number and scattered,
in contrast to the numerous thickly set hairs of hirsutus. The two
species appear to be related. They are the only ones known in which
the segments of the female antennule proximal to segment I1 have
some of the setae multiplied. D. hirsutus is clearly defined in its
characters and the male is strikingly different from that of caducus
not only in the setation of the left antennule (table 1) but in several
of the characters of the fifth leg (greater length of the claw, modifica-
tion of the right basipod 2, and the elongate form of the inner process
of the left exopod). These characters, as well as those of the female,
have been found in all the several collections examined. Present knowl-
edge of distribution confines caducus to the Pacific coast area and
hirsutus to the mountains of northeastern and north-central California.

14 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

DIAPTOMUS (HESPERODIAPTOMUS) NEVADENSIS Light
Figures 26-28

Diaptomus nevadensis LicHT, 1938, p. 60, figs. 6-7.
Diaptomus shoshone, MarsuH, 1929 (in part), p. 17.

The specimens reported by Marsh as shoshone from Devils Lake,
N. D., are referable to this species.*

Only males were present in the type lot from Nevada. Since publi-
cation, Light collected and identified the species from another Nevada
and several California localities, as listed: An alkaline lake, Washoe
County, Nev., 5,000 feet; Honey Lake and Horse Lake, Lassen
County, Calif.; Middle Lake, Cedarville, Modoc County, Calif.

The above collections contained female’, from which an allotype
specimen has been selected for description.

The male in the new Nevada and California collections shows no
significant differences from that described by Light from the type lot,
except that the lateral spine of the right second exopod of the fifth
leg is noticeably longer, equaling at least the width of the segment (in
the type specimen, which has been examined, it is considerably less).
The typical male is characterized by:

Large size (about 3.5 mm.).

Left antennule: 2 setae on II, I on 13-19.

Right antennule: Spines of 10 and 11 exceptionally slender, that
of 11 longer than that of 13. Short spinous processes on both segments
15 and 16. Process of segment 23 long and curving.

Leg 5: Claw comparatively short, only a little longer than the
exopod ; its base swollen. Right basipod 1 with prominent inner lamel-
lar expansion. Second exopod segment of right leg greatly enlarged,
with a very small spinule on inner posterior face. Left exopod: Seg-
ment I a little longer than segment 2; the inner process a very short,
but broad-based, toothlike spine, not reaching beyond the base of the
distal process. Distal pad having the spinules closely set, not arranged
in groups.

The North Dakota specimens differ from the Nevada and California
males in having the second exopod segment of the right leg not con-
spicuously enlarged, and in the absence of the short spinous process
on the 15th segment of the right antennule. The female shows no
differences.

Diagnosis of female—Allotype female: U.S.N.M. No. 94627.

1 Specimens recently reported as D. shoshone by J. E. Moore (Can. Journ.
Zool., vol. 30, p. 422, 1952) from saline lakes in Saskatchewan have also been
examined and found to be D. nevadensis.

NO. 2 COPEPODS OF THE GENUS DIAPTOMUS—WILSON 15

Honey Lake, Lassen County, Calif., June 1938, collected and identi-
fied by S. F. Light.

The female is large but comparatively slender. Length 3.85-4.05 mm.
Greatest width only 23.6 percent of length. Metasomal wings not
expanded or produced laterally, symmetrical. Urosome 30 percent of
total length, 3-segmented. Genital segment widened proximally with
slight lateral protrusions, the sensillum of the right side borne on a
larger protrusion than that of the left. Caudal rami subequal in
length to segment 3, proportions of length to width about 3: 2; ciliate
on both margins.

Antennules reaching to the middle of the genital segment. Numeri-
cal setation: 3 on segment 2, 1 on 6 and 10, 2 on II, and I on 13-19.
The seta of segment 1 not elongate, hardly reaching to the middle of
segment 2; that of 3 reaching to end of segment 6. All the segments
extremely slender, the length of segments 17-19 are 4 to 5 times their
width. Setae of segments 7, 9, and 14 exceptionally long, that of 7
the longest, reaching to middle of segment 13. Relative proportions
of these setae: Segment 7: 265; segment 9: 210; segment 14: 175.

Maxilliped very gross, as in shoshone, the endopod shorter than
the preceding basipod segment, and armed with very stout clawlike
setae. Leg 2 lacking Schmeil’s organ.

Leg 5 elongate and slender, the total exopod almost twice the length
of the basipod. Relative lengths:

Basipod 1+ 2 Exopod I Exopod 2
90 90 85
The inner spine of exopod 3 much stouter and longer than the outer
spine, armed with stout marginal spinules. Endopod shorter than
inner margin of exopod (61:85), indistinctly 2-segmented, the inner
apex produced into a sharp prolongation which is armed with coarse
hairs ; the terminal setae about two-thirds the length of the endopod.

Although it lacks any striking modification of the second basipod
segment in the male right fifth leg, D. nevadensis appears to be refer-
able to the eiseni rather than to the shoshone group of Hesperodiapto-
mus. This is evident in the male fifth leg in the regular arrangement
of the spinules of the distal pad of the left exopod and the prominent
inner lamellar expansion of the first basipod segment ; and in the right
antennule by the presence of a spinule on the fifteenth segment, which
though not always present in members of the eiseni group, has not
yet been found in those of the shoshone group. The fifth leg of the
female is strikingly similar to that of typical etseni from which it can
be distinguished by the prominent prolongation of the apex of the
endopod ; in eiseni the endopod has only a minute production of the
apex.

16 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

DIAPTOMUS (HESPERODIAPTOMUS) SCHEFFERI, new species
Figures 33-42
Diaptomus shoshone var. wardi, JuDAY AND MutTTKOWSKI, I915, p. 23, fig. I,
A-E.
Diaptomus wardi, MArsH, 1920, p. 8j, pl. 3, fig. 10; 1929 (in part), p. 23.

This interesting Pribilof Island species was erroneously identified
by Juday and Muttkowski as wardi Pearse, which they considered to
be a variety of shoshone. Marsh (1920, 1929) accepted this incorrect
identification but did not refer the species to shoshone. Study of
Montana specimens which are referable to typical wardi (see below)
show several distinctive differences between the two forms. The most
striking and the one that has hitherto been misinterpreted is the
structure of the protrusion of the second basipod segment of the male
right fifth leg.

There has been available for study some of the original material
examined by Juday and Muttkowski, now in the Marsh collection in
the National Museum, and additional specimens also from the Pribilof
Islands, referred to Dr. Light by Dr. Victor B. Scheffer, chief of
Pribilof Fur Seal Investigations, U. S. Fish and Wildlife Service.
The species is named for Dr. Scheffer.

Type locality—Upper Ice House Lake, St. Paul Island, Pribilof
Islands, Alaska. Holotype female, U.S.N.M. No. 94625; allotype
male, U.S.N.M. No. 94626.

Diagnosis.—Length: Female, about 2.66 mm.; male, about 2.5 mm.
Urosome of female 3-segmented, symmetrical, the sensilla borne on
very slight lateral protrusions. The caudal rami longer than the third
segment (15:11) with hairs only on the inner margin.

Antennules of female reaching about to the middle of the genital
segment. Numerical setation: 3 on 2, I on 6 and 10, 2 on II, I on
13-19. The seta of segment I reaching to the middle of segment 4;
that of 3 subequal in length to that of 1, reaching to near the end of
segment 6. Left antennule of male usually armed as in the female,
the seta of segment 1 not so long, reaching to the middle of 2, that of
3 to the middle of 7. The right antennule differing from the left in
having 2 setae on segment 6 (occasional specimens have 2 setae on the
left, but this is not usual) ; spines of segments 10, 11, and 13 only
moderately developed, the length of all less than the width of their
segments, that of 11 longer than that of 13. The midportion of the
antennule only moderately swollen, segment 15 lacking a process, that
on 16 short, distally placed, its length about 15 percent of the length
of the segment. The process of segment 23 usually long, slim, always
pointed, reaching to near the end of, or beyond, the last segment, the

NO. 2 COPEPODS OF THE GENUS DIAPTOMUS—WILSON 7,

inner edge usually smooth, but it may have one to several rounded
notches.

Leg 5 of female slender throughout. Exopod segments subequal
to one another, claw of exopod 2 very slim, Lateral seta of exopod
2 shorter than third exopod segment. Outer seta of exopod 3 very
short and narrow, about one-half the length of the inner which is usu-
ally stout (slim in some specimens), armed with spinules. Endopod
a little shorter than the inner margin of exopod 1, indistinctly 2-seg-
mented, or distinctly so ; the apex truncated, without apical production,
armed with a few stout spinules and hairs. The setae set terminally,
the outer the longer; proportions of endopod to outer seta to inner
seta, 38:28:17. The setae always armed with short stout hairs, often
plainly visible at low power.

Leg 5 of male, right: Claw about as long as the rest of the leg,
slender, curving, symmetrical throughout. Basipod 1 with moderately
expanded inner protrusion. Basipod 2 with long heavy ridge on
posterior face, and a rectangular lamella placed just above the middle
of the inner margin; this lamella clearly not a mere continuous pro-
trusion of the segmental body but a cuticular outgrowth consisting of
a heavy medial portion and an outer membrane. Exopod 2 lacking
the usual spine of the inner posterior face. Left leg: Basipod 2 with
the proximal inner portion protruding. Second exopod segment a
little longer than the first. Both pads large; the distal with its spine-
lets thickly set and not arranged in groups. The inner process a short
slender spine swollen at its base, reaching a little beyond the edge of
the pad.

DIAPTOMUS (HESPERODIAPTOMUS) WARDI Pearse
Figures 29-32
Diaptomus wardi PEARSE, 1905, p. 148, pl. 13, figs. 1-4.

The type locality of D. wardi is Spokane, Wash. So far as is known,
types do not exist in any available collection, although Juday and
Muttkowski (1915) mentioned that they examined specimens referred
to them by Pearse. Marsh’s (1920) figure of D. wardi from Pribilof
Island material is D. schefferi.

The confusion of these two species would be difficult to clarify
without specimens of D. wardi. Fortunately the Light collection con-
tains slides on which are appendages of two unidentified females
(leg 5) and two males (leg 5 and antennules) which are so like
Pearse’s illustrations of D. wardi that there can be no doubt of their
identity. These specimens occurred in the Montana collection with

Fics. 29-42—(See legend on opposite page.)

NO. 2 COPEPODS OF THE GENUS DIAPTOMUS—WILSON 19

Diaptomus novemdecimus, described above. This apparently consti-
tutes the first valid record of the species since its description.

The protrusion of the second basipod segment of the male right
fifth leg is not at all like the lamella on the medial margin of D. schef-
feri, which is definitely of cuticular origin. That of wardi appears
instead to be an outwardly projecting lobed protrusion of the proximal
inner portion of the segment itself. Until unmounted appendages not
distorted by cover-glass pressure can be examined, its exact structure
may not be determinable, but one of the slides contained a profile view
which appears to be quite natural (fig. 29). In the other mount, the
outline of the protrusion is clearly visible, though flattened (fig. 30).
It appears to be of the nature of that described above for D. hirsutus.

The Montana specimens agree with Pearse’s description in other
characters of the male fifth leg: The elongate cuticular prominence
of the right basipod 1 which bears the minute sensillum ; the very long
endopod of the right leg; the extremely slender claw subequal to the
rest of the leg; and the structure of the left exopod and the inner
process, similar to D. schefferi. These characters preclude possibility
of identity with D. shoshone. Pearse did not indicate a minute spinule
on the posterior face of the second exopod, which is present in the
Montana specimens. The process of segment 23 of the right antennule
agrees exactly with that shown by Pearse; it reaches beyond the apex
and is rounded at the tip. Segment 6 has only 1 seta, and the spines
of segments 10, 11, and 13 are all longer than the width of their seg-
ments, that of segment 11 longer than that of 13; both segments
15 and 16 have minute processes, that of segment 16 is at the middle
of the segment, thus differing from the usual distally placed lamelli-
form process of other hesperodiaptomids. The left antennule has 2
setae on segment 11 and 1 on 13-19. Unfortunately, the antennule of
the female had not been dissected, and it can only be assumed for the
present that its setation is like that of the male.’

2 Whole specimens of D. wardi have been examined since this report was first
written. The number of setae on these segments of the female antennule is the

Fics. 29-32.—Diaptomus (Hesperodiaptomus) wardi Pearse (Montana): 20,
Male, leg 5, right basipod (profile view of inner protrusion of basipod 2) ; 30,
male, leg 5, anterior view; 31, female, leg 5, with details of endopod apex and
exopod setae; 32, male, right antennule, apical segments.

Fics. 33-42—Diaptomus (H.) schefferi, new species: 33, Female, rostral fila-
ments and segments 1-3 of antennule (with detail of setae of segments I and 3) ;
34, male, leg 5, posterior view; 35, male, leg 5, right basipod 2 and endopod;
36, female, leg 5; 37, female, leg 5, detail of lateral setae of exopod; 38, female,
leg 5, detail of endopod apex and setae; 39, male, left antennule, setae of seg-
ments 1-6; 40, male, right antennule, apical segments, with variation of process
of segment 23; 41, male, leg 5, detail of left exopod segment 2, anterior view;
42, male, right antennule, segments 6-16, showing setae of 6, and spines and
processes of other segments.

20 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

The female fifth leg is exactly like that illustrated by Pearse, though
it appears to be more slender. The endopod has two well-developed,
equal, nonplumose setae, about half the length of the endopod which
is very slightly produced between them. The setae of exopod 3 are
very short and subequal, that of exopod 2 is very minute.

DIAPTOMUS (ARCTODIAPTOMUS) ARAPAHOENSIS Dodds

Diaptomus arapahoensis Dopps, 1915a, p. 99, figs. 3-6.
Diaptomus bacillifer, MARSH, 1924 (in part), p. 485; 10920, p. 8.

Marsh (1920) reported the occurrence of the Eurasian species
Diaptomus bacillifer on the Arctic coast of Canada and on St. Paul,
Pribilof Islands. In 1924 he supposedly extended its range in North
America by placing in synonymy with it the species arapahoensis, de-
scribed by Dodds from the Rocky Mountains of Colorado. I have
examined Marsh’s specimens of bacillifer and find his identification
to be correct. Further examination of cotypes of arapahoensis, which
are in the U. S. National Museum, and a new collection in the Light
accession, from the Rocky Mountains of Montana (Hidden Lake,
G. B. Castle collector), indicates that Dodds’s species is not referable
to D. bacillifer as Marsh had supposed.

The fifth leg of the male in most groups of Arctodiaptomus shows,
as in many species of Hesperodiaptomus, very close structural simi-
larity, and it is necessary to take into consideration all the characters
of the copepod when making identifications. The male fifth leg of
arapahoensis is built on the same general plan as that of bacillifer.
The most noticeable difference is the presence of a large, cuticular,
spinelike structure on the midposterior face of the right second exo-
pod segment. This is absent in bacillifer but is similar to that found
in other species (salinus, acutilobatus). This process is much larger
than depicted by Dodds.

The setation of the female antennules and the left male antennule
of the Canadian and Alaskan specimens of bacillifer agrees with that
given by several authors for Eurasian specimens. There are 2 setae
on segments II and 13 and I on segments 14-19, and the seta of seg-
ment 1 in the female is very long. Diaptomus arapahoensis has been
found to have the following setation in the female:

II 13 14 15 16 17 18 19
2 2 I 2 I 2 I I

The seta of segment I reaches to near the end of segment 5 and is
sparsely plumose, being similar in this to bacillifer. The male left

same as in the male. The seta of segment I is very long, reaching about to
segment 12,

NO. 2 COPEPODS OF THE GENUS DIAPTOMUS—WILSON 21

antennule differs from the female in having 1 seta on all segments
13-19. This difference exists in both the type (Colorado) and Montana
collections.

D. arapahoensis bears unmistakable resemblance to the Asiatic
species D. acutilobatus Sars (1903). The antennule setation of the
female agrees with that given for this species by Gurney (1931, table,
p. 115). In order that its exact identity may be known, and particu-
larly since there is a difference in the antennule setation of the two
sexes, a fact not known for acutilobatus, it appears best to await com-
parison of actual specimens of the two forms, before a decision as to
their conspecificity is made.

DIAPTOMUS (LEPTODIAPTOMUS) PRIBILOFENSIS Juday and
Muttkowski

Diaptomus pribilofensis JupAY AND MutTTKowSKI, I015, p. 25, figs. 1-6.
Diaptomus tyrelli, MARSH, 1915 (in part), p. 459; 1929, p. 23.—Hooprr, 1947,
p. 80.

This is a form widely spread in Alaska and western Canada and
has for years been considered synonymous with D. tyrrelli (corrected
spelling). It is closely allied to D. coloradensis from the Rocky Moun-
tains and forms with it and the Asiatic species Diaptomus angustilobus
Sars (1898) a group of seemingly allopatric species. Its supposed
synonymy with tyrrelli has been unfortunate in obscuring the pattern
of its distribution and its closer relationship to the other species of
the group. Specimens reported by Hooper as tyrrelli from western
Canada have been examined and identified as pribilofensis.

DIAPTOMUS (EUDIAPTOMUS) GRACILIS Sars

Diaptomus gracilis is a well-known Eurasian species new to North
American fauna. It has been found recently in several of my Alaskan
collections. It appears to be common in the Arctic regions of Alaska,
having been found on the western coast (lakes of the lower Yukon
River and Bristol Bay areas) and on the Arctic slope at Umiat. In
south-central Alaska it occurred in collections of the Kuskokwim River
area at McGrath and in Wonder Lake, Mount McKinley National
Park.

DIAPTOMUS (AGLAODIAPTOMUS) MARSHIANUS, new species
Figures 43-51

Type lot—13 females, 45 males. Lake Jackson, Leon County, Fa.,
April 3, 1950, Murray H. Voth collector. Holotype female, U.S.N.M.

SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

Fics. 43-58.—(See legend on opposite page.)

NO. 2 COPEPODS OF THE GENUS DIAPTOMUS—WILSON 23

No. 94634; allotype male, U.S.N.M. No. 94635. Occurring with D.
(Arctodiaptomus) floridanus Marsh and D. (Skistodiaptomus) mis-
sissippiensis Marsh.

Diagnosis—Length, dorsal view : Female, 1.55-1.89 mm. ; male, 1.3-
1.58 mm. Greatest width of female in segments 2 and 3, equaling
about 28 percent of the length. Metasomal segment 5 with an unusual
dorsal protuberance consisting of an erect cuticular frill, placed mostly
on the right side. The wings of the last segment not produced laterally,
the inner portion rounded so as to form a lobe in dorsal view.

Urosome of female 3-segmented. Genital segment noticeably asym-
metrical, the right side being very tumid. Caudal rami longer than
segment 3, 25:20, their width 64 percent of their length, ciliate on
inner margin only; the inner dorsal seta as long as the inner caudal
seta.

Antennules reaching beyond the caudal rami by the last 2-3 seg-
ments. Numerical setation: 2 on segment II, I on segments 13-19.
The seta of segment 1 short, reaching to the middle of segment 2, that
on 3 reaching to segment 6. Setae of segments 17, 19, 20, and 22 stiff
and uncinate, their length less than, or equal to, that of the segment.
Setation of left antennule of male like that of the female, including
the uncinate setae. Right antennule of male having spines of 10 and
11 longer than the width of their segments, that on segment 13 longer
than that of 11, strongly outcurved ; segments 15 and 16 with spinous
processes. The process of segment 23 reaching to about the middle of
24, outcurved, accompanied by a narrow membrane.

Maxilliped not grossly developed, the distal lobe of the basal seg-
ment with 3 setae in both sexes. Schmeil’s organ present on the
endopod of leg 2.

Leg 5 of female: Basal segment with large sensillum on broad base.
Exopod 1 and 2 subequal in length, exopod 3 not separated. Seta of
exopod 2 present, set closely with the setae of exopod 3; the inner
seta with marginal serrations. Endopod longer than inner margin of

Fics. 43-51.—Diaptomus (Aglaodiaptomus) marshianus, new species: 43,
Female, metasome segments 5-6 and urosome, dorsal; 44, female, metasome
segments 5-6, lateral view; 45, female, leg 5, with detail of lateral setae of
exopod; 46, female, leg 5, detail of endopod setae; 47, male, right antennule,
spines and processes of segments 10-16; 48, male, leg 5, right basipod (with
profile of inner protrusion) and outline of endopod; 49, female, antennule, de-
tail of uncinate seta of segment 19; 50, male, leg 5, posterior view; 51, male,
right antennule, apical segments.

Fics. 52-58.—Diaptomus (Mastigodiaptomus) texensis, new species: 52, Fe-
male, metasome segments 5-6 and urosome, dorsal; 53, female, metasome seg-
ments 5-6, lateral view; 54, male, leg 5, left exopod, apical segment, posterior
view; 55, male, leg 5, posterior view; 56, female, leg 5, with details of exopod
setae and apex of endopod; 57, male, right antennule, spines and processes of
segments 10-16; 58, male, distal segments of urosome, dorsal.

24 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

exopod I, its setae not longer than half its length, both with enlarged
bases and thickly plumose.

Leg 5 of male, right: Basipod 1 with a long spinelike sensillum
reaching to near the middle of the next segment. The inner proximal
portion of basipod 2 with a prominent lobed protrusion accompanied
distally by a small narrow hyaline lamella (fig. 48) ; the distal portion
of the segment lacking the cuticular process present in many species
of Aglaodiaptomus. The first exopod segment about as long as basipod
2, the outer distal portion produced. Proportions of exopod 1 to exo-
pod 2, 25:40. Inner portion of exopod 2 deeply grooved, the anterior
side with a protruding flange ; relative length of lateral spine to exopod
2, 26: 40. Claw subequal to exopod 2, very stout and curving. Endopod
a little less than one-half the length of exopod 1.

Left leg: Sensillum of basipod 1 a stout spine. Exopod 1 noticeably
longer than exopod 2, 24:16. Exopod 2 broadened distally, the distal
process digitiform, nearly one-third the length of the outer margin of
the segment ; the inner process a long, curving, setiform spine, nearly
3 times the length of the distal (14:5), spinulose on its inner margin ;
the sclerotized marginal area of the segment produced to a point at its
base. The proximal pad consisting of a hairy region on the upper
inner margin; the distal pad of spinulose areas on the posterior face.
The endopod very large, reaching to near the end of exopod 2, the
inner portion grooved, the entire surface thickly spinulose from above
the middle to the end.

Taxonomic position—The subgenus Aglaodiaptomus was proposed
by Light in 1938. The original list of included species should be
revised as follows:

D. piscinae Forbes (1893) should be recognized as a synonym of
D. leptopus Forbes (1882). The type collections of Forbes (Illinois
Natural History Survey) as well as those of Schacht and Marsh have
been examined, and no definable structural difference has been found.
(The details of this study are reserved for future publication.) In
the synonymy of leptopus should also be placed D. manitobensis Ar-
nason (1950). I wish to acknowledge Dr. Arnason’s courtesy in
permitting me to examine type material of his proposed species.

Diaptomus spatulocrenatus Pearse (1906) was omitted from Light’s
list.

The species Diaptomus pseudosanguineus Turner (1921), which
was omitted by Marsh (1929), should be recognized, although there
are certain inadequacies in the description. The species was described
from the St. Louis, Mo., area, and on the basis of the description it
is not referable to any of the known species of Aglaodiaptomus. The

NO, 2 COPEPODS OF THE GENUS DIAPTOMUS—-WILSON 25

female of pseudosanguineus is described as having a pair of long,
curved spines on the ventral portion of the genital segment, and the
photographic illustrations show a process distad to the genital pro-
tuberance in 2 lateral views of what appear to be two separate indi-
viduals (Turner, 1921, pl. 1, fig. 3; pl. 2, figs. 1 and 2). Such a
process does not occur on the genital segment of any of the species
of Aglaodiaptomus, all of which have been examined. The male fifth
leg is most comparable to that of spatulocrenatus, resembling it in the
proportions of the segments of the left exopod which are subequal,
and the endopod which is described as having a crenate inner margin.
The right first exopod segment differs from spatulocrenatus in having
the distal outer portion produced as in conipedatus and marshianus.
No detail can be made out, from the illustration, of the right second
basipod segment, adequate knowledge of which is extremely important
in the taxonomy of this group. No type material of Turner’s species
is known to be in existence.

Diaptomus marshianus is distinguished in the female by the peculiar
dorsal protuberance of the metasome. There is no evidence in any of
the other species of Aglaodiaptomus of such cuticular development.
There can be no question, however, of the reference of this female to
Aglaodiaptomus and hence to the male described from this collection,
because the female shows unmistakable aglaodiaptomid characters in
the uncinate setae of the antennule, the presence of three rather than
four setae on the distal lobe of the basal segment of the maxilliped,
and the dense plumosity of the endopod setae of the fifth leg. The
male fifth leg most closely resembles that of spatulocrenatus and
conipedatus, from which it can be distinguished by the lack of a distal
cuticular process of the right second basipod segment and by the
grosser development of the left endopod. It appears to differ from
pseudosanguineus in having the first leit exopod segment considerably
longer than the second.

It is a personal pleasure to give the name of Dr. C. Dwight Marsh
to a distinctive American species of Diaptomus. In this connection,
attention should be drawn to the fact that Kiefer (1936, p. 309) has
shown that the species named D, marshi by Juday (1914) should be
known as D. colombiensis Thiebaud. Kiefer has stated that Thiebaud’s
paper was actually published as a separate in 1912 instead of 1914.

Acknowledgment is due Murray Voth and Dr, Irene Boliek, of
Florida State University, for specimens and information of this in-
teresting species.

26 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

DIAPTOMUS (MASTIGODIAPTOMUS) TEXENSIS, new species
Figures 52-58

Type lot.—200 specimens of both sexes. Temporary roadside pond,
county road to Bayside, about 1.5 miles west of Rockport, Aransas
County, Tex., “spring” of 1945, Joel W. Hedgpeth collector. Holo-
type female, U.S.N.M. No. 94630; allotype male, U.S.N.M. No.
94631.

Diagnosis ——Length, dorsal view: Female, 1.5-1.6 mm. Greatest
width of female in segment 3, 26-28 percent of length. Distal part of
fifth metasomal segment of female usually with a small, medially
placed, rounded dorsal protuberance (not always present). Wings
not expanded, the left larger than that of the right side, both with
spinelike sensilla, that on the inner portion of the left side usually
directed outward. Urosome of female 3-segmented. Genital segment
symmetrical, without lateral protrusions, lateral sensilla stout. Caudal
rami longer than segment 3 (26:20) ; their width about 61 percent of
their length; both margins ciliate; the dorsal seta about one-half to
three-fourths the length of the inner caudal seta. Urosome of male
asymmetrical ; segment 4 produced backward on the right side ; caudal
rami asymmetrical, the right longer than the left, with a cuticular
process on its ventral side near the base of the inner setae.

Antennules of female reaching beyond caudal rami by last two
segments. Setation: 2 on segment II, I on segments 13-19; seta of
segment I short. Left antennule of male like that of female. Right
antennule of male: Spine on segment 8 not enlarged, that of segment
10 hardly larger than that of 8, that of 11 nearly as long as width of
its segment, that of 13 not much longer than the width of its segment,
but exceedingly stout. Segments 14, 15, and 16 with stout spinous
processes. Proportions of spines and processes to one another :

8 10 II 13 14 15 16
3 5 14 20 13 14 5

Segment 23 with a short, outcurved process, reaching about to the
middle of segment 24.

Maxilliped slender, setation of basal segment normal ; the inner setae
of the endopod not clawlike, all shorter than the endopod; the outer
and terminal setae longer than the endopod (40: 37). Schmeil’s organ
present on endopod of leg 2.

Leg 5 of female stout, width of exopod 1 about half its length.
Sensillum of basipod 1 a long, stout, flat spine. Exopod 1 a little
longer than exopod 2 (27: 25). Exopod 3 separated, its outer seta short
and spinelike, closely set with and usually overlying the inner seta;

NO. 2 COPEPODS OF THE GENUS DIAPTOMUS—WILSON 27

lateral seta of exopod 2 present, shorter than exopod 3. Claw with
spinules on both margins. Endopod nearly as long as inner margin
of exopod 1, 2-segmented, bearing 2 setae, the outer longer than the
inner, which is set considerably above the tip of the endopod; tip of
endopod with double row of stout hairs.

Leg 5 of male, right: Sensillum of basipod 1 a stout spine, in pos-
terior view overlying the second basipod segment and directed toward
a protrusion of the segment whose central portion consists of a cres-
cent-shaped sclerotized lamella. The inner proximal portion with an
inwardly and sometimes distally directed small marginal lamella. Exo-
pod 1 with small lamellae on both inner and outer distal portions.
Exopod 2 bulging medially, with a crescent-shaped sclerotization on
the midposterior face; the lateral spine distally placed, stout and
straight, its length less than that of the segment. Claw strongly curved
at middle, as long as the rest of the leg.

Left leg: Sensillum of basipod 1 stout as on the right side. Inner
part of basipod 2 produced proximally. Exopod 1 more than three
times longer than exopod 2 (50:15). The distal process short and
broad, its margins strongly serrate ; the inner process spiniform, reach-
ing to the end of the distal. Pads medial in position, the proximal the
larger.

Taxonomic position.—This new species is allied to the southwestern
species D. albuquerquensis Herrick, the known distribution of which
extends through the Rocky Mountain States from Utah to Guatemala
in Central America. The only other species of the subgenus on the
continent, D. amatitlanensis M.S. Wilson (1941), is also known from
Guatemala. There are no authentic records of the group from south-
eastern United States. Florida specimens identified by Schacht (1897)
as D. albuquerquensis are undoubtedly referable to D. floridanus
Marsh (1926). Such is also true of the specimens from Georgia listed
by Humes (1950). Specimens to which Humes referred have been
sent to me by Dr. M. S. Ferguson, of the United States Public Health
Service, and have been found to be D. floridanus.

Diaptomus saltillinus Brewer, which is closely allied to D. floridanus,
is found in Texas and some other areas where the albuquerquensis
group occurs. D. saltillinus and D. floridanus belong to the subgenus
Arctodiaptomus Kiefer (1932) and the albuquerquensis group to the
subgenus Mastigodiaptomus Light (1939). There are superficial re-
semblances between these two groups of species, but they should not
be confused with each other if careful attention is given to basic sub-
generic characters such as are found in the left exopod of the male
fifth leg, the armature of the endopod of the female fifth leg, and the

28 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

presence in Mastigodiaptomus of 2 setae on segment 11 of the female
and left male antennules, as contrasted to the single seta of this seg-
ment in saltillinus and floridanus.

D. texensis is distinguished from albuquerquensis by several easily
recognized differences in the male right fifth leg. In albuquerquensis
the lateral spine of exopod 2 is longer than the segment, the second
basipod segment has the inner proximal portion bulging upward as
does also the marginal lamella, and the distal posterior face has a
characteristic sculpturing of the cuticle that is lacking in tevxensis.
The females are very similar but can be separated by the lateral pro-
trusions of the genital segment and the usual shortness of the endopod
of the fifth leg of albuquerquensis.

ACKNOWLEDGMENTS

The extensive collections of Diaptomus that have been gathered
together by the American specialists Dr. C. D. Marsh and Dr. S. F.
Light are in the U. S. National Museum. Their use has made much
of this work possible. For the organization and selection of materials
from these collections, I am greatly indebted to Dr. Fenner A. Chace,
Je, and’ Dr--Paul': Tile.

The type lots of Diaptomus shoshone, Diaptomus leptopus, and
piscinae were lent by the Illinois Natural History Survey, through Dr.
Herbert Ross and Philip W. Smith. The type of Diaptomus mamni-
tobensis was kindly referred to me by Dr. I. G. Arnason. Collections
upon which other published records were based were lent by Dr. Frank
F. Hooper and Dr. M. S. Ferguson. Grateful acknowledgment is
also made to the following persons who have referred new collections
to me or supplied information concerning them: Charles S. Wilson,
Murray H. Voth, Dr. Irene Boliek, Dr. Joel W. Hedgpeth, Dr.
Charles C. Davis, Dr. Walter G. Moore, and Dr. L. B. Holthuis.

LITERATURE CITED
ArRNASON, I. GILBERT.
1950. A new species of diaptomid copepod from Manitoba. Journ. Elisha
Mitchell Sci. Soc., vol. 66, pp. 148-155, 23 figs.
Cart, G. CLIFFORD.
1940. The distribution of some Cladocera and free-living Copepoda in
British Columbia. Ecol. Monogr., vol. 10, pp. 55-110, 14 figs.
1944. The natural history of the Forbidden Plateau area, Vancouver Island,
British Columbia. Rep. Prov. Mus. for 1943, pp. D 18-40, 2 ills.,
I map.
Dopps, GinEon S.
1915a. Descriptions of two new species of Entomostraca from Colorado,
with notes on other species. Proc. U. S. Nat. Mus., vol. 49, pp.
97-102, 10 figs.

NO. 2 COPEPODS OF THE GENUS DIAPTOMUS—WILSON 29

1915b. A key to the Entomostraca of Colorado. Univ. Colorado Stud., vol.
II, pp. 265-208, 82 figs.
1917. Altitudinal distribution of Entomostraca in Colorado. Proc. U. S.
Nat. Mus., vol. 54, pp. 59-87, Io figs., 2 pls.
1924. Notes on Entomostraca from Colorado. The Shantz collections from
the Pikes Peak region. Proc. U. S. Nat. Mus., vol. 65, art. 18,
pp. 1-7, I fig.
Forses, S. A.
1882. On some Entomostraca of Lake Michigan and adjacent waters.
Amer. Nat., vol. 16, pp. 537-542, 640-649, 2 pls.
1893. A preliminary report on the aquatic invertebrate fauna of the Yellow-
stone National Park, Wyoming, and of the Flathead region of
Montana. Bull. U. S. Fish Comm. for 1891, pp. 207-258, 6 pls.
GuRNEY, RoseErrT.
1931. British fresh-water Copepoda. Vol. 1, 238 pp., 344 figs. Ray Society,
London.
Hooper, FranxK F.
1947. Plankton collections from the Yukon and MacKenzie River systems.
Trans. Amer. Micr. Soc., vol. 66, pp. 74-84, 1 fig.
Humes, ArTHUR G.
1950. Experimental copepod hosts of the broad tapeworm of man, Diboth-
riocephalus latus (L.) Journ. Parasitol., vol. 36, pp. 541-547.
JupAy, CHANCEY.
1914. A new species of Diaptomus. Trans. Wisconsin Acad. Sci. Arts and
Lett., vol. 17, pp. 803-805, 2 figs.
Jupay, CHANcEy, and MutrxKowskI, R. A.
1915. Entomostraca from St. Paul Island, Alaska. Bull. Wisconsin Nat.
Hist. Soc., vol. 13, pp. 23-31, 6 figs.
KIEFER, FRIEDRICH.
1932. Versuch eines Systems der Diaptomiden (Copepoda, Calanoida).
Zool. Jahrb. (Abt. Syst.), vol. 63, pp. 451-520, 88 figs.
1936. Freilebende Stiss- und Salzwassercopepoden von der Insel Haiti.
Arch. Hydrobiol., vol. 30, pp. 263-317, 126 figs., I map.
ibiveiet, Sp ae
1938. New subgenera and species of diaptomid copepods from the inland
waters of California and Nevada. Univ. California Publ. Zool.,
vol. 43, pp. 67-78, 23 figs.
1939. New American subgenera of Diaptomus Westwood (Copepoda,
Calanoida). Trans. Amer. Micr. Soc., vol. 58, pp. 473-484, 24 figs.
Marsu, CHARLES DWIGHT.
1907. A revision of the North American species of Diaptomus. Trans.
Wisconsin Acad. Sci. Arts and Lett., vol. 15, pp. 381-516, 14 pls.
1915. A new crustacean, Diaptomus virginiensis, and a description of
Diaptomus tyrelli Poppe. Proc. U. S. Nat. Mus., vol. 49, pp. 457-
462, 7 figs.
1920. The fresh water Copepoda of the Canadian Arctic Expedition 1913-
18. Rep. Canadian Arctic Exped. 1913-18, vol. 7, pt. J, pp. 1J-25J,
5 pls.
1924. A new locality for a species of Diaptomus. Science, vol. 59, pp. 485-
486.

30 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

1926. On a collection of Copepoda from Florida, with a description of
Diaptomus floridanus, new species. Proc. U. S. Nat. Mus., vol.
70, art. 10, pp. I-4, 6 figs.

1929. Distribution and key of the North American copepods of the genus

- Diaptomus, with the description of a new species. Proc. U. S. Nat.
Mus., vol. 75, art. 14, pp. 1-27, 16 figs.
Pearse, A. S.

1905. Contributions to the copepod fauna of Nebraska and other States.
Stud. Zool. Lab. Univ. Nebraska, No. 65, pp. 145-160, 5 pls.

1906. Fresh-water Copepoda of Massachusetts. Amer. Nat., vol. 40, pp.
241-251, 9 figs.

Sars, Georc OssIAn.

1898. The Cladocera, Copepoda and Ostracoda of the Jana Expedition.
Ann. Mus. Zool. Acad. Imp. Sci. St. Pétersbourg, vol. 3, pp. 324-
359, 4 pls.

1903. On the crustacean fauna of Central Asia. Pt. III. Copepoda and
Ostracoda. Ann. Mus. Zool. Acad. Imp. Sci. St. Pétersbourg,
vol. 8, pp. 195-232, 8 pls.

ScHACHT, FREDERICK WILLIAM.

1897. The North American species of Diaptomus. Bull. Illinois State Lab.

Nat. Hist., vol. 5, pp. 97-208, 15 pls.
ScHMEIL, OrrTo.

1896. Deutschlands freilebende Siisswasser-Copepoden. Pt. 3. Centro-

pagidae. Bibl. Zool., vol. 21, pp. 1-144, 12 pls.
THACKER, Mr. and Mrs, T. L.

1923. Some freshwater crustaceans from British Columbia. Can. Field-Nat.,

vol. 37, pp. 88-80.
THIEBAUD, M.

1912. Voyage d’exploration scientifique en Colombie. Copépodes de Co-
lombie et des Cordilléres de Mendoza. Mém. Soc. Neuchateloise
Sci. Nat., vol. 5, pp. 160-175, 25 figs.

TurRNER, C. H.

1921. Ecological studies of the Entomostraca of the St. Louis district.
Pt. 1. Diaptomus pseudosanguineus sp. nov. and a preliminary
list of the Copepoda and Cladocera of the St. Louis district. Trans.
Acad. Sci. St. Louis, vol. 24, No. 2, pp. 1-25, 4 pls.

Wirson, Mitprep STRATTON.

1941. New species and distribution records of diaptomid copepods from
the Marsh collection in the United States National Museum.
Journ. Washington Acad. Sci., vol. 31, pp. 509-515, I fig.

SMITHSONIAN MISCELLANEOUS COLLECTIONS
VOLUME 122, NUMBER 3

fae METAMORPHOSIS OF A
FLY Ss. HEAD

BY
R. E. SNODGRASS

Collaborator, Bureau of Entomology and Plant Quarantine,
U. S. Department of Agriculture

(PusiicaTion 4133)

CITY OF WASHINGTON
PUBLISHED BY THE SMITHSONIAN INSTITUTION

JUNE 25, 1953

The Lord Baltimore Press

BALTIMORE, MD., U. 3. A.

fib METAMORPHOSIS"OF Ay BLY’S HEAD

By R. E. SNODGRASS

Collaborator, Bureau of Entomology and Plant Quarantine
U. S. Department of Agriculture

A legless, almost headless, wormlike maggot hatches from the egg
of a fly; but the maggot is not a young fly in the sense that a kitten
is a young cat, or even in the sense that the nymph of a grasshopper
is a young grasshopper. The maggot does not grow up into a fly,
and neither does it literally transform into a fly. It is a highly special-
ized larval form of its species, which, though developed directly from
the fly’s egg, becomes a creature self-sufficient in all respects except
that of procreation. Structurally the fly larva is so different from
its parents that it cannot itself go over into the next fly generation.
Consequently nearly all the larval tissues finally go into a state of dis-
solution, and the fly is then newly generated from groups of undif-
ferentiated cells that are carried by the larva but which form no
essential part of the larval organization.

This potentiality of dual development from a single egg becomes
most accentuated among the Diptera in the cyclorrhaphous families.
It affects not only the internal organs, but also the body wall, which
is almost entirely replaced during the pupal stage from groups of
cells, known as imaginal discs, that remain undeveloped from an early
period, and at the end of the larval life begin an active growth that
forms the integument and appendages of the pupa. The cells of the
larval integument degenerate before the advancing new epidermis and
are cast into the body cavity where they become food for the develop-
ing imaginal tissues. During larval life the regenerative discs of the
thorax and head are contained in narrow-necked pouches of the epi-
dermis, closed at their outer ends beneath the cuticle. Within these
pouches the appendage rudiments may develop continuously through
the larval instars without being exposed at the larval moults. Finally,
however, during the prepupal or early pupal stage the pouches are
everted and the appendages quickly grow to the state of development
they have when the pupa is exposed by the shedding of the last larval
cuticle, while the everted pouches themselves expand by cell prolifera-
tion and construct the pupal integument.

SMITHSONIAN MISCELLANEOUS COLLECTIONS, VOL. 122, NO. 3

2 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

All this has been known for nearly a century. Weismann (1864)
said that the thorax and head of the fly, together with their append-
ages, the halteres, the wings, legs, antennae, eyes, and mouth parts de-
velop within the body of the larva, and the truth of this statement has
been verified by numerous subsequent workers. Most of the earlier
students of the structure and metamorphosis of the cyclorrhaphous
larva, however, did not understand the morphology of the larval head.
Though they correctly described facts, their identification of anatomi-
cal parts is often entirely erroneous, and later writers, taking their
statements literally, either criticize them as false, or perpetuate their
errors. In the following pages an attempt will be made first to under-
stand the nature of the head of a cyclorrhaphous larva, and then to put
together the story of the formation of the adult head as far as it can
be compiled from our present information on the subject.

In the lower nematocerous flies the metamorphic changes between
larva and adult are less intense than in the cyclorrhaphous families,
and larval tissues may go over directly into adult tissues. In the larva
of Corethra, for example, as described by Weismann (1866), the
imaginal discs of the thorax are mere groups of cells in the larval
epidermis, which begin development in the prepupal period and then
form only the pupal appendages. The general integument of the pupal
thorax in this case is a product of renewed growth activity in the cells
of the larval epidermis, which simply remodel the thorax into the form
of the pupal thorax. The same applies to changes of the head, the
pupal head being formed by alterations in shape and size of the larval
head within the unshed cuticle of the last larval instar. The imaginal
mouth parts of Nematocera have been shown by Kellogg (1902) to
be formed directly within the larval mouth parts; the adult antennae,
however, which are generally much longer than the larval organs, de-
velop with only their distal ends in the larval antennae. In some of
the lower flies, as will be shown later, the imaginal antennae grow
within pockets of the integument, and the pockets may include also
the rudiments of the compound eyes.

The structural disparity between the larva and the adult in the
Cyclorrhapha is due to the specialized form that the larva has acquired,
rather than to that of the adult fly. The larval head of these flies in
particular has become so highly modified in a specific way that it is
difficult to understand how it has been evolved from a head of more
usual structure. Only a small part of the adult head is derived directly
from the larval head.

The apparent, or functional, head of a muscoid maggot is a small,
rounded lobe at the anterior end of the body (fig. 2 A, LH) more or
less sunken into the thorax. Apically this larval head bears a pair of

NO. 3 METAMORPHOSIS OF A FLY’S HEAD—SNODGRASS 3

large papillae, on each of which are situated two small sense organs,
but there are no eyes of any kind. The under surface of the head (A)
presents a median depression from which projects a pair of strongly

re
cy

CATIT ALAN
Fits 7 CL
ARAN
MAN |
4 meg

Wu
qt un vi
ANS a a:

WN
“|B? i A a ‘J
cae S| woe
re Ai
ees
x
f

Fic. 1.—Head and proboscis of adult muscoid flies.

A, Musca domestica L., anterior. B, Calliphora vicina R.-D. (erythrocephala
Meigen), clypeus and proboscis. C, Callitroga macellaria (F.), ventral. D,
Gonia sp., with ptilinum everted.

sclerotized, decurved hooks (mh) partly covered by lateral folds of
the integument. Below the bases of the hooks is a soft median lobe
(Lb), which at least serves the larva as an under lip, and appears to
be a true larval labium. Above the labium, between the bases of the

4 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

hooks, is the food-intake orifice of the larva (Atr), but it leads im-
mediately into an atrial chamber before the mouth of the larval sucking
apparatus.

The larval organ of ingestion is a suction pump lying within the
thorax, supported by a strongly sclerotized structure commonly called
by students of cyclorrhaphous larvae the “buccopharyngeal skeleton”
or the “cephalopharyngeal apparatus” (fig. 2B). By whatever name
this complex structure is known, it is an important part of the larval
head retracted into the thorax. In details of shape it differs character-
istically in different species, but the general form and structure of the
organ is that shown here for the mature larva of Callitroga macellaria.

The dorsal part of the sucking apparatus (fig. 2 F) is a long, thin,
hyaline plate having a strongly contrasting, dark U-shaped sclerotiza-
tion around its anterior end with the arms extending posteriorly along
the lateral margins. From the edges of this sclerotized part of the
dorsal plate a strong lateral plate descends on each side (B) and ex-
pands below into a broad posterior extension. Supported between the
lower edges of the lateral plates is the sucking pump of the larva
(CbP), which is continuous anteriorly from the atrium above the
labium (Lb) and posteriorly into the oesophagus (Oc). The lumen
of the pump when contracted is crescent-shaped in cross section (D,
Cb), but on its concave upper wall are attached two rows of large
dilator muscles (dlcb) arising on the arms (Clp) of the U-shaped
sclerotization of the dorsal plate. Anterior to the lateral plates is a
smaller, independent, median, ventral plate (B, e) on which the mouth
hooks (mh) are articulated. This plate, which lies on the base of the
dorsal wall of the larval labium (C, e), is H-shaped in ventral view
(E, e). In front of its crossbar are two small sclerites bearing minute
sense organs, and a narrow anterior V-shaped sclerite. Just behind
the crossbar is the opening of the salivary duct (B, E, SIDct), which
discharges on the base of the labium.

The dorsal plate of the larval sucking apparatus is covered by a very
delicate, closely adherent membrane (fig. 2B, a). Anteriorly, how-
ever, the membrane becomes free, forming the dorsal wall of the
atrium (Atr), and is then continued into the wall of the ventral de-
pression of the external larval head (A). When the atrium is exposed
by cutting away the covering membrane (C) there is seen projecting
into it from the anterior end of the dorsal plate of the sucking appa-
ratus a small conical lobe (Lm) with a minute sclerotic tip. This lobe
is clearly the larval labrum ; in a first instar larva the sclerotized tip is
larger and forms a conspicuous tooth.

The “buccopharyngeal skeleton” of the cyclorrhaphous larva is

NO. 3 METAMORPHOSIS OF A FLY’S HEAD—SNODGRASS 5

perhaps generally regarded as a structure distinctive of the larva, since
most entomologists do not seem to have observed that it is almost a
replica of the supporting skeleton of the sucking pump of the adult
fly, which is commonly known as the “fulcrum” of the proboscis.
This structure in the fly (fig. 3 F) consists of the clypeus (C/p) and
a pair of lateral plates (f), called the paraclypeal phragmata, inflected

ay
» i >

ea
S
SS
=

=
<>

Fic. 2—Larval head structures of Callitroga macellaria (F.).

A, head lobe of larva (LH), partly retracted into prothorax, ventral. B, feed-
ing apparatus (“buccopharyngeal skeleton”) of a mature larva, lateral. C, dia-
gram of anterior end of sucking apparatus, with lateral wall of atrium (Air)
cut away, exposing the labrum (Lm). D, cross section of sucking apparatus,
showing inflection of paraclypeal phragmata (f,f) from edges of clypeus, sup-
porting the cibarial pump (Cb). , ventral surface of anterior part of cibarial
pump and H-shaped sclerite (e) supporting the mouth hooks (mh) and the
labium (Lb). F, frontoclypeal plate of sucking apparatus, dorsal. G, diagram-
matic lengthwise section of head and prothorax, mesal view of right half.

from the clypeal margins, which support between their lower edges the
sucking pump (COP) of the food tract. The dilator muscles of the
pump (G, dicb) arise on the clypeus and are enclosed between the
paraclypeal phragmata. The cross section of the “fulcrum” of the fly
(E), therefore, is an exact duplicate of a similar section of the “buc-
copharyngeal” apparatus of the larva (fig. 2 D), and there can scarcely

6 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

be any question that the two structures are merely imaginal and larval
forms of the same thing.

The clypeus of a muscoid fly is generally U-shaped or V-shaped
with the closed end dorsal. In Musca and Calliphora (fig. 1 A, B)
the clypeus (C/p) is fully exposed on the base of the proboscis; in
Callitroga (C) it is deeply sunken in a cavity on the under side of the
head; in Gonia (D) it is relatively very small. In Musca (A) the
broad base of the clypeus is closely hinged to the lower margin of the
frons (Fr) ; in Calliphora (B) a hinge plate (hpl) intervenes between
the frons and the muscle-bearing plate of the clypeus; in Callitroga
(C) the sunken clypeus is separated by membrane from the epistomal
ridge beneath the frons; in Gonia (D) the diminutive clypeus is well
removed from the frons. In any case, the proboscis, with the clypeus
and the sucking apparatus, swings back and forth below the frons in
the ample membranous connection of the clypeus with the head by
muscles attached on the supporting skeleton of the sucking pump.
The latter and the clypeus are, therefore, known as the “fulcrum” of
the proboscis.

Finally, to understand the nature of the parts that compose the “ful-
crum” of the adult muscoid fly, we must go back to the more primitive
condition in the orthopteroid insects. A median section through the
distal part of the head of a cockroach (fig. 3 A) shows that there is a
specific preoral food pocket, the cibariwm (Cb), between the epipha-
ryngeal wall of the clypeus (C7p) and the sloping basal part of the
hypopharynx (Hphy). Two suspensory rods on the cibarial floor ex-
tend up through the angles of the mouth (4) and give attachment to
muscles from the frons. On the anterior or upper wall of the cibarium
are attached thick bundles of muscle fibers (dlcb) arising on the ex-
ternal clypeal area of the head. These muscles are compressors of the
clypeus, but their contraction expands the cibarium. If, then, the mov-
able lobe of the hypopharynx is brought against the inner surface of
the labrum (Lm), the cibarium will become practically a closed cham-
ber opening anteriorly from the food meatus (fm) between the labrum
and the free lobe of the hypopharynx, and proximally into the stomo-
daeum (Stom). It is very probable that the cibarium thus serves the
cockroach as an organ for the ingestion of liquids. On its dorsal wall
are transverse compressor muscles not shown in the figure. The true
mouth of the cockroach is the opening of the cibarium (Mzth’) into the
stomodaeum. An important point to bear in mind is that the cibarial
muscles of the clypeus are separated from muscles of the stomodaeum
arising on the frons by the frontal ganglion and its brain connectives,

NO. 3 METAMORPHOSIS OF A FLY’S HEAD—SNODGRASS 7

The homology of the sucking pump of the fly with the cibarium of the
cockroach has been amply illustrated by Gouin (1949).

In various insects the cibarium becomes a permanently more or less
closed chamber by a lateral union of the epipharyngeal wall with the
base of the hypopharynx, so that the functional mouth opening may
come to lie beneath the base of the labrum. The cibarium thus becomes
more efficient as a sucking organ. Among the Diptera this condition is
fully developed in the lower families, and is well illustrated in the mos-
quito (fig. 3B). The cibarial pump of the mosquito (CbP) has a
strongly sclerotized basinlike floor ; the intake orifice lies beneath the
base of the labrum and thus constitutes a secondary mouth (Mth’).
Since the floor of the pump in the mosquito corresponds with the hypo-
pharyngeal floor of the cibarium in the cockroach, the hypopharyngeal
stylet of the mosquito (B, Hphy) represents only the free lingual lobe
of the cockroach hypopharynx (A, Hphy). A section of the sucking
pump of the mosquito (indicated by the arrow at B) shows two sets of
strong dilator muscles (dlcb) from the clypeus to the concave upper
wall of the pump.

The cibarial pump of the mosquito projects freely into the head (fig.
3 B), and, though it is strongly sclerotized and is suspended from the
frons by muscles attached on a pair of proximal processes (4), it is
still not braced against the pull of the dilator muscles. This condition
has been remedied in the higher flies. In some of them, as in the
mydas fly (C), a strong ridge is inflected from a groove on each side of
the clypeus, and the distal ends of the ridges (f) are fused with the
lateral walls of the pump, thus serving to hold the latter firmly in
place. From this simple condition it is only a step to that in the mus-
coid flies in which the clypeal ridges have been enlarged into broad
paraclypeal phragmata (F, f) supporting the full length of the pump.
The dilator muscles of the pump (FE, dich) are thus boxed in between
lateral plates (f, f), and the pump is securely braced against the clyp-
eus. As in the cockroach and the mosquito, the primary mouth of the
muscoid fly is the opening of the cibarium into the stomodaeal oesopha-
gus (F, Mth’), but the functional mouth (Mth”) is the entrance into
the cibarium from the food meatus (fm) between the labrum and the
hypopharyngeal stylet. However, in those flies in which the labellar
lobes of the labium form a broad, food-collecting disc (D), the deep
notch between the lobes (Mth’’) is the real intake aperture for liquid
food, and has been termed the prestomum.

The paraclypeal phragmata are not primarily inflections from the
extreme edges of the clypeus. In the adult male of Tabanus, as has
been shown also by Bonhag (1951), the clypeus is divided longitudi-

8 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

nally into three areas by a groove on each side well within the epistomal
sulcus. These clypeal grooves in Tabanus form merely a pair of inter-
nal ridges, but it is clear that the ridges represent the beginning of
paraclypeal phragmata in other flies. Developmentally the phragmata
are first formed in the larva, and they may be well developed in bra-
chycerous as well as in cyclorrhaphous larvae. Theoretically, however,
it seems probable that the complex sucking apparatus must have been
first evolved in the adult fly, since the larvae of lower dipterous fam-
ilies have biting and chewing mouth parts.

On returning now to the larva, it is clear that the sucking pump (fig.
2B, CbP) is the cibarium, as it is in the fly. The dilator muscles of
the larval cibarium lying in front of the frontal ganglion, therefore,
should identify the part of the dorsal plate on which they take their
origin as the clypeus (C/p), since these muscles entirely conform with
the cibarial muscles of the adult. In the larva, however, there is an
oblique posterior group of fibers just behind the frontal ganglion, at-
tached below on the stomodaeal oesophagus (G) and arising on the
posterior part of the dorsal plate of the sucking apparatus mesad of the
cibarial muscles. In the cyclorrhaphous flies the stomodaeum proceeds
from the cibarial pump as a simple tubular oesophagus (figs. 2 B, 3 F,
Oe), but in adult Brachycera it is differentiated immediately behind
the cibarium into a second, smaller pharyngeal pump, with its dilator
muscles arising on the frons, and these muscles are those represented
in the cyclorrhaphous larva by the oesophageal muscles arising on the
posterior part of the dorsal plate of the sucking apparatus. The struc-
ture and mechanism of the pharyngeal pump in the adult of Tabanus
are well described and illustrated by Bonhag (1951).

In the larvae of Stratiomydiae the pharyngeal pump has been con-
verted into a crushing organ by the transformation of its dorsal wall
into a thick plate with a convex, sometimes strongly ridged, under sur-
face that fits like a broad pestle into the concave, mortarlike ventral
wall. This pharyngeal organ is sclerotically continuous with the long,
slender cibarial pump, from the end of which it turns upward like the
bowl of a pipe from the stem (fig. 6 B, Phy). It is the Schlundkopf
of Jusbaschjanz (1910), who calls the cibarium the “pharynx” ; it is
described in the larva of Odontomyia alticola by Cook (1949), and
Schremmer (1951) gives a fully detailed account of its structure and
probable use in the larva of Stratiomys chamaeleon. The organ is
operated by dorsal muscles arising on the frontoclypeal area of the
head. A large anterior muscle inserted at the junction with the cibar-
ium is shown by Cook to lie before the frontal ganglion and its brain
connectives. This muscle, therefore, is a cibarial muscle; the other,
posterior muscles are true frontal pharyngeal muscles.

NO. 3 METAMORPHOSIS OF A FLY’S HEAD—SNODGRASS 9

The attachment of both frontal muscles and clypeal muscles on the
dorsal plate of the larval sucking apparatus should identify this plate
as a frontoclypeal element of the head skeleton, which is a well-defined,

Fic. 3.—The sucking apparatus of adult Diptera, and comparison with the
cibarium of a cockroach.

A, diagrammatic lengthwise section of head of a cockroach. B, diagram of
sucking apparatus of a mosquito. C, same of a mydas fly. D, labellar disc of a
muscoid fly. FE, cross section of sucking apparatus of adult Callitroga macellaria.
F, sucking apparatus and mouth parts of adult Callitroga macellaria, lateral.
G, lengthwise section of sucking apparatus of same, showing clypeal dilator
muscles of cibarium.

median dorsal area of the head in most nematocerous and brachycer-
ous fly larvae (fig. 6 A). Cook (1949) has called this entire area the
“clypeus,” but in so doing he disregards the evidence from muscle
attachments in insects having the clypeus separated from the frons, in

IO SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. I22

which the cibarial muscles arise on the clypeus and the postcibarial
muscles on the frons. The frontal and clypeal areas, however, are
often continuous. Ludwig (1949) says merely that this area of the
head in the larva of Calliphora “includes the clypeus and some addi-
tional part of the cranium.” The sucking apparatus of the cyclorrha-
phous larva, therefore, is a more complex structure than that of the
adult fly in so far as it includes not only the clypeus and the cibarium,
but also the frons and frontal muscles of the stomodaeum. The fron-
toclypeal plate and the sucking apparatus of the larva lie entirely
within the thorax, a position they have acquired either by retraction or
by the overgrowth of a fold from the thorax, or by both means. The
frontoclypeal plate is connected with the external larval head (fig. 4 D,
LH) by the membrane (a) extending back from the latter over the
atrium.

The H-shaped plate of the larva that lies on the base of the labium
and supports the mouth hooks (fig. 2 B, E, e) suggests by its position
the hyoid sclerite of the adult (fig. 3 F, hy), but in the larva the sali-
vary duct (SIDct) opens behind the H-shaped plate, while in the adult
it enters the hypopharyngeal stylet anterior to the hyoid. The hypo-
pharynx in the larva is represented only by the floor of the cibarium, a
free hypopharyngeal lobe corresponding with that of the cockroach
(fig. 3 A) or with the hypopharyngeal stylet of the fly (F, Hphy),
being absent in the larva. The larval labium (fig. 2 A, C, Lb) does
not become the labium of the adult ; the labium of the fly is developed
from a pair of histoblastic pouches formed inside the larval labium.

The nature of the mouth hooks of the cyclorrhaphous larva has been
a subject of much discussion, some writers contending that the hooks
are mandibles, others that they are not. The latest advocate of their
mandibular nature is Ludwig (1949). If the larval mouth hooks are
not mandibles, the question is, what are they? In the first place, it is
curious that mandibles should have their only articulations on a plate
on the base of the labium, and secondly, since the muscles of the hooks
are attached on the paraclypeal phragmata of the sucking apparatus,
it is an unusual thing for mandibular muscles to arise on any part of
the clypeus. However, since the parietal walls of a typical insect cra-
nium are obliterated in the fly larva, the phragmata offer the only
available solid support for the muscles, and muscles do change their
points of origin where efficiency demands a change. On the other hand,
if the hooks are not mandibles, they cannot be homologized with any
other structure of other insects, and it is hardly to be supposed that
such highly developed feeding organs should be developed de novo for
the express use of the larva. However, since the hooks disappear at

NO. 3 METAMORPHOSIS OF A FLY’S HEAD—SNODGRASS II

the end of larval life and the adult fly has no mandibles, the larval
hooks cannot be put to the crucial test of finding what they become in
the imago, and for this same reason we may leave the matter without
further discussion, inasmuch as the mouth hooks are not involved in
the metamorphosis of the larva into the fly.

In most nematocerous and brachycerous families the head of the
larva is more or less retracted into the thorax, so that it is at least
partly ensheathed in a fold of the prothorax. In the cyclorrhaphous
larva, however, the head appears to consist of an external part bearing
the apical sense organs of the larva, and of a retracted part that in-
cludes only the frontoclypeal area, which carries the labrum and
supports the cibarial sucking apparatus. The cyclorrhaphous larva
thus presents a cephalic condition that is difficult to understand, and
even the known facts of embryonic development do not make the
condition entirely clear.

The head of the embryo at an early stage of its development is a
simple structure. As shown by Pratt (1901) in Melophagus ovinus
(fig. 4 A) the embryonic head presents a dorsal lobe above the entrance
into the food tract (Cb) and a ventral lobe below it. The dorsal lobe,
which contains a group of compressor muscles (dich), Pratt calls the
“muscular sucking tongue,” but we can easily recognize this lobe as the
labrum and clypeus (Lm, Clip), and the muscles as the dilators of the
future cibarial pump (Cb). The ventral lobe is clearly the larval la-
bium. This stage of the embryo may be diagrammatically presented in
a more conventional form as at C of the figure. The short dorsal wall
of the embryonic head represents at least the clypeus of the larva (D,
Clp) bearing the labrum and giving attachment to the dilator muscles
of the cibarium.

The primary embryonic head now becomes covered by the forward
growth of an integumental fold (fig. 4 A, C, hf) from behind it, which
goes over the labrum (B) and forms the roof of an antechamber, the
head atrium (Atr), before the mouth of the cibarium (Cb), while the
fold itself becomes at least a part of the external head lobe of the larva
(D, LH) bearing the larval sense organs. The overgrowth of the
primitive head by this secondary dorsal head fold is well illustrated
also in Calliphora by Ludwig (1949, fig. 58). According to Pratt
(1901) there is a dorsal and a ventral fold in Melophagus (A, B, C,
hf, vf). Unfortunately Pratt’s terminology is confusing because he
calls the newly formed atrium the “pharynx,” and the cibarium the
“stomodaeum.” With the completion of the dorsal head fold the em-
bryo acquires the essential head structure of the larva, represented
diagrammatically at D of the figure. The frontoclypeal plate and the

I2 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

cibarial apparatus thus become enclosed within the thorax, and the
major part of the external larval head lobe (LH) bearing the larval
sense organs appears to be a secondary structure formed by the dorsal
head fold extended from c to d. The original space (b) beneath the
fold later becomes obliterated by the close apposition of the inner wall
of the fold (a) on the frontoclypeal plate, but the labrum (Lm) is
left projecting freely into the atrium (Afr).

Of the two sense organs on each of the apical papillae of the larva,
the dorsal one, according to Ludwig (1949), represents the larval
antenna, the ventral one the maxillary palpus. This opinion was also
that of Weismann (1864), but other authors have considered the inter-
pretation doubtful. The alleged antennal organ is shown by Ludwig to
be innervated by a long branch from the labrofrontal nerve—a most
unusual association for an antennal nerve, and neither the nerve nor
the sense organ can have any relation to the antenna of the adult. The
ventral sense organ, Ludwig says, “is the maxillary palp sense organ,”
but apparently the only basis for this statement is that the organ in
question is innervated by a branch from the “mandibular-maxillary-
labial nerve.” A sense organ wherever located must have a nerve. The
origin of the papillar sense-organ nerves from head ganglia is not
proof that the organs are either antennal or maxillary, but it is con-
vincing evidence that, whatever they are, they belong to the head, and
Ludwig shows, moreover, that the organs originate in the epidermis
of the lateral walls of the embryonic head. It becomes a problem,
therefore, to understand how these sense organs in the larva come to
be situated on the external head lobe formed by the head fold, and
their position on this lobe raises the question as to whether the fold
pertains to the thorax or to the head.

As seen in longitudinal sections the head fold of the cyclorrhaphous
embryo (fig. 4 A, B, Af) suggests the prothoracic fold that partly en-
sheaths the head of many nematocerous and brachycerous larvae (fig.
6 A, B, thf). Schremmer (1951) asserts that there appears to be no
remnant of a head in the cyclorrhaphous larva, and that as a result of
the forward growth of the dorsal fold the larval sense organs come to
be on the anterior end of the thorax. Holmgren (1904) apparently
regarded the larval head lobe as a derivative of the thorax, but he says
nothing of the sense organs. Pantel (1898) called the larval head lobe
a “pseudocephalon.” Ludwig (1949) also attributes at least a part of
the head fold to the thorax because it contains a pair of muscles in-
nervated from the prothoracic ganglion that “insert on a sclerotized
area between the mandibles.” In the Callitroga larva, however, these
muscles do not arise in the head lobe itself but on the overhanging

NO. 3 METAMORPHOSIS OF A FLY’S HEAD—SNODGRASS 13

anterior part of the prothorax, so, if the larval head lobe is a part of
the head, the muscles in question are merely prothoracic head muscles.

In a cross section of the embryonic head of Calliphora it is shown
by Ludwig that the head fold (fig. 6 C, hf) covers only a narrow space
(b) above the frontoclypeal surface (C/p). The inner lamella of the
fold (a) arches immediately over the frontoclypeus, while the outer
lamella has become continuous with the parietal walls of the head. If

IW erry case seSC ea

ARVRuaa rumen

Mth O
S Cb e

gIDet hace

Fic. 4.—Development of the larval head of a cyclorrhaphous fly.

A, lengthwise section of embryonic head of Melophagus ovinus (from Pratt,
1901), showing beginning of head fold (hf). B, later stage of same (from Pratt,
1901), in which the head fold has grown forward over the labrum and labium,
which are now enclosed in a secondary preoral atrial chamber (Air). C, dia-
grammatic expression of A. D, diagrammatic analysis of the anterior larval
structure based on B.

the fold proceeds over the head as a narrow median growth from the
thorax alone, it is difficult to understand how it becomes so intimately
a part of the head wall. In any case, it is evident that this head fold of
the cyclorrhaphous larva is something quite different from the pro-
thoracic fold that ensheaths the base of the head in a brachycerous
larva.

The head fold of Calliphora, Ludwig (1949) says, appears at about
the thirteenth hour of the developing embryo, and “‘is in the shape of a

14 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

U, with the open end pointed anteriorly.” Its lateral margins lie mesad
of the developing larval sense organs. If, therefore, the head fold
grows forward in this manner with its arms extending along the edges
of the frontoclypeal area, it would seem that, whatever its origin, the
extension of the fold must be at the expense of the head wall itself,
and that the arms of the fold close medially from the sides as the fold
advances. If this is the manner of growth of the fold, the condition
seen in cross sections of the embryonic head (fig. 6 C) becomes under-
standable. Furthermore, only by some such process of growth from
the head wall could the lateral sense organs on the embryonic head be
carried up over the labrum and finally come to be situated on the an-
terior end of the fold, which forms at least the dorsal part of the head
lobe of the larva. It is, then, certainly more rational to regard the
larval head lobe as a part of the head itself than as a derivative of the
thorax. Clearly there is need for further study of the nature of the
head fold and the manner of its growth, and Schremmer (1951,
p. 362) has promised a new investigation “tuber die Enstehung des
Cyclorrhaphenlarvenkopfes.”

When a young insect in its development takes a path widely diver-
gent from that of its parents, and acquires a head structure as extra-
ordinarily specialized as that of the cyclorrhaphous larva, it is evident
that the larval structure cannot be “transformed” into that of the
adult. The head of the fly, therefore, is practically a new structure de-
veloped without reference to the larval head. In the evolution of the
Diptera, however, the cyclorrhaphous way of forming the adult head
has been derived from a more simple method retained in some of the
lower flies.

Among the nematocerous Diptera, as has been shown by Kellogg
(1902) in Simulium and Bibiocephala, the imaginal (pupal) head may
be formed simply and entirely within the loosened cuticle of the larval
head, and the imaginal mouth parts are formed inside the cuticle of the
larval mouth parts. The antennae of the pupa, because they are much
longer than those of the larva, find space for their growth between the
pupal head and the cuticle of the larval head, but their tips are retained
in the corresponding larval organs. In Corethra, as described by Weis-
mann (1866), the long slender antennae of the pupa become sunken
into pouches of the pupal head, from which they are everted when the
larval cuticle is shed. In Corethra the compound eyes are formed on
the surface of the pupal head beneath the larval cuticle. In Tendipes
(Chironomus) Miall and Hammond (1900) showed that both the
antennae and the compound eyes of the pupa are developed within
longitudinal infoldings of the epidermis of the dorsal wall of the pupal

NO. 3 METAMORPHOSIS OF A FLY’S HEAD—SNODGRASS 15

head inside the larval cuticle. Dissection of a mature Tendipes larva
reveals a pair of long pockets converging from the larval antennae to
the posterior end of the head (fig. 5 A), each of which contains an
axial tubular antenna (Ant) and, in the wall of its basal part, the de-
veloping rudiment of a compound eye (E£). These elongate pockets

Fic. 5.—Development of the frontal sacs.

A, oculoantennal pockets from head of a mature tendipedid (chironomid)
larva, Tendipes plumosus (L.), extending posteriorly from larval antennae.
B, diagrammatic cross section of pupal head of Psychoda alternata showing open
grooves (FS) containing imaginal antennae and rudiments of compound eyes
(outline from Feuerborn, 1927). C, diagrammatic dorsal view of head of young
pupa of same, showing oculoantennal grooves extended into pockets of prothorax
(outline from Feuerborn, 1927). D, cross section of head of embryo of Me-
lophagus ovinus showing origin of frontal sacs (FS) on sides of head (from
Pratt, 1901). E, dorsal view of feeding apparatus of mature larva of Rhagoletis
pomonella, with fully developed frontal sacs (FS) extending posteriorly from
frontoclypeal plate (from Snodgrass, 1935). F, cross section of 7-hour prepupa
of Drosophila melanogaster showing united frontal sacs produced into lateral
pouches with folded walls (from Robertson, 1936).

lie immediately beneath the ecdysial cleavage grooves of the larval
head. Very similar groovelike pockets of Psychoda are described and
figured by Feuerborn (1927) as infoldings of the pupal head (B, FS)
open by narrow slits on the surface, and containing the developing an-
tennae (Ant) and compound eyes (£). In Psychoda the grooves ex-
tend into the front part of the thorax (C) as pockets, which deepen as
the pupa develops.

10 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

It is probable that similar developmental processes occur in other
Nematocera, though little attention has been given to the details of
metamorphosis in these flies. The oculoantennal pockets of the head
very clearly are equivalent to the peripodal pouches of the thorax in
which the imaginal legs are developed and to the pouches that contain
the wing rudiments.

In the higher Diptera the oculoantennal pockets, known as the
frontal sacs, are present in the larvae as long-necked pouches extend-
ing from the posterior end of the frontoclypeal plate of the sucking
apparatus into the thorax as far as the retracted brain, with which they
are connected by ocular nerves (figs. 2B, 5 E,6 B, FS). In the strat-
iomyid Odontomyia Jusbaschjanz (1910) says the pouches contain
only the histoblasts of the compound eyes (fig. 6 B, FS), the antennae
arising from the surface of the head as in Corethra. In all cyclorrha-
phous flies that have been described, however, the frontal sacs contain
the rudiments of both the eyes and the antennae. These sacs are
formed in the embryo and are present in all stages of the larva, but
reach their full development only in the last larval instar. In their early
origin, therefore, the frontal sacs of the head in the cyclorrhaphous
flies more nearly resemble the thoracic peripodal pouches of the legs
than do the oculoantennal pockets of the Nematocera, which appear
only in the prepupal stage. Because in the late embryo the sacs appear
to arise from the inner end of the passage between the inner lamella
of the head fold and the underlying frontoclypeal plate (fig. 4 B, FS),
this passage (b) has been regarded as an unpaired part of the sacs,
and the latter have been erroneously said to be invaginations from the
atrium (Air), or from the “pharynx” if the atrium is mistaken for the
pharynx. The point at which the sacs grow into the thorax (D, c) is
simply overgrown by the head fold, and the true origin of the sacs is
on the lateral parts of the embryonic head.

According to Ludwig (1949) the imaginal discs of the compound
eyes in the embryo of Calliphora arise as ectodermal thickenings on
the lateral walls of the head, but in the larva both the ocular and the
antennal rudiments are contained in a pair of membranous sacs lying
along the sides of the oesophagus. Ludwig does not explain how the
sacs are developed, or how they come to contain the histoblasts of the
eyes and antennae. In his figure 57 he shows the left sac exactly as all
other writers have depicted the frontal sacs, and yet he says “embry-
onic studies reveal no such pouches.” Furthermore, Ludwig attributes
to Pratt (1901) the absurd statement that the common opening of the
sacs “is drawn forward and downward, and then posteriorly through
the mouth,” and on this assertion he bases a criticism of Pratt’s work.

NO. 3 METAMORPHOSIS OF A FLY’S HEAD—SNODGRASS 17

However, Pratt makes no such statement, or anything like it. More
concisely than does Ludwig himself, Pratt describes in Melophagus
ovinus the origin of the frontal sacs (“dorsal head discs”) as dorso-
lateral thickenings of the epidermis of the embryonic head. Early in
their history the discs begin to invaginate in the form of crescentic
slits (fig. 5 D, FS), and later they move dorsally to the back of the
head, where their outer parts unite in a single, transverse depression,
while the inner parts increase in length and extend separately into the

Fic. 6.—Head of stratiomyid larva and sections of embryonic head of Calliphora.

A, larval head of Ptecticus trivittatus (Say) partly ensheathed in fold of pro-
thorax, dorsal. B, lengthwise section of larval head and prothorax of Odonto-
myia (combination diagram from Jusbaschjanz, 1910, relettered). C, cross
section near base of head of 15-hour embryo of Calliphora (from Ludwig, 1949).
D, cross section of head lobe of 16-hour embryo of Calliphora, overhanging the
labium (from Ludwig, 10949).

body cavity as a pair of stalked sacs that lie in contact with the cere-
bral ganglion. Now there takes place, from behind the mouth of the
sacs (fig. 4 C, c), the formation of the dorsal fold (hf), which grows
forward over the head. Since the inner lamella of the fold becomes
closely adherent to the frontoclypeal plate, it thus comes about that in
the larva the sacs appear to be attached to the posterior end of the
larval sucking apparatus (figs. 2B, 5 E, FS). Their true opening at
the posterior end of the head beneath a fold of the thorax (thf) is
shown by Jusbaschjanz (1910) in his sectional figure of a stratiomyid
larva (fig.6 B, FS).

18 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

From the description of the early history of the frontal sacs given
by Pratt it is clear that the frontal sacs of the cyclorrhaphous flies can
be correlated in their origin with the oculoantennal grooves of the
pupal head in the Nematocera. That, in the former, the sacs arise in
the embryo instead of in the pupa shows that the imaginal discs of
the head have followed the same course of evolution as have those of
the thorax, which also in the higher flies have come to be formed in the
embryo.

The further history of the frontal sacs has been followed by Wahl
(1914) in Calliphora and by Robertson (1936) in Drosophila, Weis-
mann (1864) observed that the head of the fly is formed from two
“cell masses” (the frontal sacs), which at first are in contact and later
become united. Jusbaschjanz (1910) noted that in a stratiomyid larva
there is only one frontal sac (Kopffalte) at the time of pupation, from
which fact he concluded that the two primary sacs must have united in
a single pouch. Wahl (1914) specifically describes the formation of a
single sac in the early pupa of Calliphora by a dissolution of the mesal
walls of the original two sacs followed by a union of their outer walls.
The resulting unpaired sac then increases in size by expansion of lat-
eral pouches, and its walls become thrown into numerous irregular
folds. In Drosophila Robertson (1936) says the closely appressed
frontal sacs begin to fuse two hours after the formation of the pupar-
ium. The median walls break down and the broken edges of one sac
unite with those of the other until the two sacs have completely united
(fig. 5 F, FS) except at their posterior ends where the optic concavi-
ties are applied to the cerebral ganglia.

At pupation the cephalic fold of the larva retracts (fig. 7 B, hf), the
passage (b) beneath it opens and becomes continuous with the lumen
of the now single frontal sac (FS), so that, as Wahl (1914) shows in
the early pupa of Calliphora, the frontal sac comes to open directly to
the exterior above the mouth of the cibarium (“pharynx”). The same
thing was noted by Pratt (1897) in Melophagus, but Pratt’s language
is somewhat confusing to a modern reader when he says “the lumen of
the discs and that of the pharynx become completely merged and
form together a single continuous space.”’ The “discs” are the frontal
sacs, the “pharynx” is the larval atrium. When now the pupa is first
exposed by the shedding of the last larval cuticle within the puparium,
there is to be seen at the anterior end of the body only a great hole in
the front of the prothorax (fig. 7 A). This stage is the cryptocephalic
phase of the pupa. Shortly thereafter the walls of the cavity are sud-
denly everted, and the pupa thus acquires a head (C). The pupal head
is at first relatively small and not fully developed, but it takes on its

NO. 3 METAMORPHOSIS OF A FLY’S HEAD—SNODGRASS I9

definitive size and structure (D, E) during the rest of this phanero-
cephalic stage of the pupa. When the head of Drosophila is first
everted, Robertson says, the eyes are brought to their final position, but
are not yet histologically completed, and the antennae are simple
thickenings of the front wall of the head. Bodenstein (1950) describes
in detail the development of the compound eyes in Drosophila.

Ludwig (1949) emphatically denies that there is any process of in-
vagination involved in the formation of the head of the fly. However,

Fic. 7.—The pupa.

A, cryptocephalic pupal stage of Rhagoletis pomonella, ventral. B, lengthwise
section of anterior end of 10-hour prepupa of Drosophila melanogaster showing
opening of frontal sac just before pupation (from Robertson, 1936). C, early
phanerocephalic pupal stage of Rhagoletis pomonella. D, mature pupa of Rhago-
letis, lateral. FE, same, ventral.

since the adult head is visibly everted in the pupal stage, it is not clear
how it became introverted without a previous inversion. The frontal
sacs are actually ingrowths of the embryonic integument, and an in-
growth is usually called an “invagination,” though admittedly it is
more properly an introversion. Furthermore, Ludwig criticizes a
former statement by the writer (1935, p. 313) that “the entire facial
region of the head, including the area of the frons and that of the
imaginal antennae and compound eyes, is invaginated into the thorax.”
This statment is in accord with the findings of other writers, since the
time of Weismann (1864), and all that is needed to demonstrate its
truth is a glance at a pupa in the cryptocephalic stage, whether of

20 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL, I22

Musca, Calliphora, Drosophila, or Rhagoletis (fig. 7 A). When the
frontal sac is everted, it brings with it the eyes and the antennal rudi-
ments, and its walls form the epidermis of all parts of the fly’s head
except that part derived from the sucking apparatus of the larva. The
cuticular skeleton of the latter is shed at the moult to the pupa, but the
matrix of the organ must remain to form the more simple sucking ap-
paratus (or “fulcrum’’) of the adult, though the transformation has
not been observed.

In the change from the larva to the adult the frontoclypeal plate of
the larva undergoes a very considerable modification. First, it is dis-
tinctly divided, in the fly, into frontal and clypeal elements ; the clypeal
area retains the muscles of the cibarium, the frons now carries the
attachments of the postcibarial frontal muscles of the stomodaeum.
The shape of the clypeus in the adult becomes reversed from that of
the larva, in that, though U-shaped or V-shaped in both, the open end
is distal in the adult (fig. 1, Clp). The frons of the fly is a part of the
head wall, including specifically the depressed area of the face (Fr)
in which the antennae are lodged.

Again we may point out that the frontal sacs of the cyclorrhaphous
fly larva from which the imaginal head is formed are cephalic equiva-
lents of the thoracic histoblasts, which latter not only give rise to the
legs and wings, but in the higher flies regenerate the thoracic integu-
ment as well. As an example we may refer to Robertson’s (1936)
account of the formation of the imaginal thorax in Drosophila. As the
histoblast pouches of the legs and wings open to allow the contained
appendages to evert, their edges expand by cell proliferation, while
the surrounding larval cells retreat and are gradually sloughed off into
the body cavity to be devoured by phagocytes. The newly generated
areas spread over the thorax, unite, and finally construct the entire
thorax of the fly. In describing the formation of the thorax of Me-
lophagus, Pratt (1897) says: “In proportion as the larval hypodermis
disappears under the attack of the phagocytes, the edges of the imag-
inal discs grow and take its place, forming the imaginal hypodermis.”
The idea that the larval cells are first destroyed by phagocytes, how-
ever, is not in accord with results of later investigators. The cephalic
histoblasts of the fly have no opposition from larval cells because of
the great reduction of the larval head ; the elaborate head of the cyclor-
rhaphous fly is practically a new structure with no counterpart in the
larva.

Likewise, the mouth parts of the fly owe little to those of the larva.
The larval mouth hooks are not re-formed in the pupa, and the fly has
no trace of mandibles. The adult labium is formed from a pair of his-

NO. 3 METAMORPHOSIS OF A FLY’S HEAD—SNODGRASS 21

toblastic pouches developed inside the larval labium. According to
Wahl (1914) these ventral histoblasts give rise to the entire proboscis
of the fly, including the hypopharynx, the maxillary remnants, and the
labrum, which statement suggests that the matter should be reinvesti-
gated. The cyclorrhaphous larva, as already observed, has no free
hypopharyngeal lobe, and the salivary duct opens on the base of the
labium (fig. 2C, S7Dct). In the fly, on the other hand, the salivary
outlet duct traverses a median stylet arising at the base of the labium,
which is commonly called the hypopharynx. Because this stylet gives
passage to the salivary duct, however, Ferris (1950) asserts that it is
not a hypopharynx, but a secondarily developed outgrowth containing
the salivary outlet. According to the same interpretation the Hemiptera
and Siphonaptera also should not have a hypopharynx. While it is
generally true that the salivary outlet duct of insects opens between the
base of the hypopharynx and the base of the labial prementum, the
opening is sometimes on the base of the hypopharynx, as in the cock-
roach (fig. 3 A, S/O), in dragonflies, and, as shown by Weber (1938),
in the Psocoidea. The hypopharynx is a median, postoral outgrowth
of the ventral wall of the head, principally on the maxillary segment,
but it may encroach on the labial segment. If the organ includes a
labial element, therefore, it is nonetheless a hypopharynx, and if the
salivary duct opens into a pocket on its base it might traverse its entire
length. In the larvae of nematocerous flies a hypopharynx is present,
but, as in other holometabolous insects, it is united with the labium in
a composite suboral lobe and the outlet duct of the salivary glands
opens distally between the two component parts of the latter. The an-
cestors of the Diptera, therefore, must have possessed a true hypo-
pharynx, and there would seem to be no reason why it should not be
restored in the adult, just as are the legs. Weismann (1864) called
the median mouth stylet of the fly “die Kieferborste,” and described
it as formed by the union of paired parts about a cellular strand that
became the salivary duct. Again, we can say only that the pupal devel-
opment of the mouth parts of the cyclorrhaphous flies needs further
investigation, since the ordinary criterion of correlating the adult parts
with the larval parts cannot be invoked.

A comparison of the mouth parts of the fly with those of the cock-
roach shows at least that the stylet containing the salivary outlet of the
fly (fig. 3 F, Hphy) corresponds exactly in position with the free lobe
of the hypopharynx in the cockroach (A, Hphy). Its grooved dorsal
surface, moreover, is continued into the floor of the sucking pump (F,
CbP), which represents the floor of the cibarium on the base of the
hypopharynx in the cockroach. Even the oral arms of the suspensory

22 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

rods of the cockroach hypopharynx (A, y) may be retained in the flies
as a pair of short cibarial processes (B, F, y) embracing the primary
mouth (F, Mth’).

Finally, in connection with the metamorphosis of the fly’s head, we
should mention the ptilinum, since it constitutes an example of ex-
ceptional development in the pupa of a special structure for the tem-
porary use of the fly. The ptilinum is a vesicular introversion of the
front of the head of the pupa in the schizophorous families of the
Cyclorrhapha, which is everted by the emerging fly to open the anterior
end of the enclosing puparium. After emergence, the lips of the open-
ing come together in the long groove of the head that arches over the
bases of the antennae (fig. 1 A, pts). As described by Laing (1935),
in Calliphora the ptilinum is formed in the young pupa from the head
wall just above the antennae, which on the third day of pupal life be-
gins to introvert and soon becomes a crumpled sac inside the head with
a greatly thickened cuticle. Eversion of the ptilinum in the emerging
fly is brought about by blood pressure resulting from contraction of
the abdomen. The surface of the organ in different flies may be
smooth, covered with fine spicules, or, as in Gonia (fig. 1 D), thickly
coated with coarse spines. After the ptilinum has served its purpose
it is again retracted and remains as a large though shrunken body in
the fly’s head. The retraction is caused by muscles, which are fully
described by Laing. Some of the muscles are special ptilinal retractors, -
and these muscles disappear during the first two days of the life of the
fly.

Metamorphosis in the cyclorrhaphous Diptera is a “change of form”’
in the insect as a whole, but it is not a transformation of the maggot
into a fly. The maggot represents an extreme degree to which juvenile
development among the insects has diverged from the evolutionary
course that produced the adult, until the young insect has become an
independent creature in no way structurally related to its parents. The
embryo develops directly into the form of the larva and not into that
of the insect that produced it, but certain cells of the larval tissues
retain the potentiality of reproducing the corresponding adult tissues,
while the rest of the larval tissues, after performing their temporary
function, go into dissolution and become food for the growing imaginal
tissues. The maggot is in no sense a recapitulation of any stage in the
evolution of the fly, except larval stages of its more recent ancestors.
The larval form is determined at an early period of development in the
egg, and when the larva has completed its destiny it gives way to the
ancestral development of the fly, but the manner in which the modern

NO. 3 METAMORPHOSIS OF A FLY’S HEAD—SNODGRASS 23
fly is developed has no phylogenetic significance. The larval devel-
opment and the adult development are known to be under control
of hormones, but the mechanism of dual inheritance has not been
explained.

EXPLANATION OF LETTERING ON THE FIGURES

a, membrane over frontoclypeal plate
of larva, inner wall of head fold.

Ant, antenna.

Atr, head atrium.

b, space between head fold of embryo
and frontoclypeal plate.
Br, brain.

c, posterior end of frontoclypeal plate,
origin of inner wall (a) of head
fold.

Cb, cibarium.

CoP, cibarial pump.

Clp, clypeus.

Cr, crop.

d, end of dorsal wall of head fold.
dicb, dilator muscles of cibarial pump.

e, H-shaped sclerite supporting mouth
hooks.
E, rudiment of compound eye.

f, paraclypeal phragma.
fm, food meatus.

Fr, frons.

FS, frontal sac.

Gng, ganglion.

h, hinge of clypeus on frons.
hf, head fold.

Hphy, hypopharynx.

hpl, hinge plate of clypeus.

H st, hypostome.

H stl, haustellum.

hy, hyoid sclerite.

L, legs.

Lb, labium.

Lol, labellum.

LH, external larval head lobe.
Lm, labrum.

mh, mouth hooks of larve.
Mth', primary mouth (entrance to sto-

modaeum).

Mth", secondary mouth (entrance to
cibarium).

Mth'", tertiary mouth, prestomum

(aperture to food meatus between
labella).
M-xPlip, maxillary palpus.

NC, nerve cord.
Nv, antennal nerve.

Oe, oesophagus.

Phy, pharynx.
Ptl, ptilinum.
pts, ptilinal sulcus.

Rst, rostrum of proboscis.

SIDet, salivary duct.
SIO, salivary orifice.
Stom, stomodaeum.

Th, thorax.
thf, thoracic fold.

Vent, ventriculus.
vf, ventral head fold.

W, wing.

y, oral arm of hypopharyngeal suspen-
sorium, or of floor of cibarium.

24 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. I22

REFERENCES
BovENSTEIN, D.

1950. The postembryonic development of Drosophila. In Demerec, M.,

Biology of Drosophila, pp. 275-367, 33 figs. New York.
Bonuae, P. F.

1951. The skeleto-muscular mechanism of the head and abdomen of the
adult horsefly (Diptera: Tabanidae). Trans. Amer. Ent. Soc.,
vol. 77, pp. 131-202, 31 figs.

Cook, E. F.

1949. The evolution of the head in the larvae of Diptera. Microentomology,

vol. 14, pp. 1-57, 35 figs.
Ferris, G. F.

1950. External morphology of the adult (of Drosophila). In Demerec, M.,

Biology of Drosophila, pp. 368-419, 22 figs. New York.
FEUERBORN, H. J.

1927. Die Metamorphose von Psychoda alternata Say. I. Die Umbildungs-
vorgange am Kopf und Thorax. Zool. Anz., vol. 70, pp. 315-328,
8 figs.

GouIn, F.

1949. Recherches sur la morphologie de l’appareil buccal des diptéres. Mém.

Mus. Nat. Hist. Nat., n.s., vol. 28, pp. 167-260, 64 figs., 3 pls.
Hotmcren, N.

1904. Zur Morphologie des Insektenkopfes. II. Einiges tiber die Reduktion
des Kopfes der Dipterenlarven. Zool. Anz., vol. 27, pp. 343-355,
12 figs.

JUSBASCHJANZ, S.

1910. Zur Kenntnis der nachembryonalen Entwicklung der Stratiomyiden.
Jenaische Zeitschr. Naturw., vol. 46 (N.F.38), pp. 681-736, 7 figs.,
I pl.

Kettoce, V. L.

1902. The development and homologies of the mouth parts of insects. Amer.

Nat., vol. 36, pp. 683-706, 26 figs.
LAING, JOYCE.

1935. On the ptilinum of the blow-fly (Calliphora erythrocephala). Quart.

Journ. Micr. Sci., vol. 77, pp. 497-521, 14 figs.
Lupwic, C. E.

1949. Embryology and morphology of the larval head of Calliphora erythro-
cephala (Meigen). Microentomology, vol. 14, pt. 3, pp. 75-111,
figs. 43-65.

Mratt, L. C., and Hammonp, A. R.

1900. The structure and life-history of the harlequin fly (Chironomus),

196 pp., 127 figs. Oxford.
PANTEL, J.

1898. Le Thrixion halidayanum Rond. Essai monographique sur les
caractéres extérieurs, la biologie et l’anatomie d’une larve parasite
du groupe des Tachinaires. La Cellule, vol. 15, pp. 1-285, 6 pls.

NO. 3 METAMORPHOSIS OF A FLY’S HEAD—SNODGRASS 25

Pratt, H. S.

1897. Imaginal discs in insects. Psyche, vol. 8, pp. 15-30, 11 figs.

1901. The embryonic history of imaginal discs in Melophagus ovinus L.,
together with an account of the earlier stages in the development
of the insect. Proc. Boston Soc. Nat. Hist., vol. 29, pp. 241-272,
7 pls.

Rosertson, C. W.

1936. The metamorphosis of Drosophila melanogaster, including an accu-
rately timed account of the principal morphological changes. Journ.
Morph., vol. 59, pp. 351-399, 4 pls.

ScCHREMMER, F.

1951. Die Mundteile der Brachycerenlarven und der Kopfbau der larve von
Stratiomys chamaeleon L. Osterreichische Zool. Zeitschr., vol. 3,
Pp. 326-397, 25 figs.

Swnoperass, R. E.

1924. Anatomy and metamorphosis of the apple maggot, Rhagoletis pomo-
nella Walsh. Journ. Agr. Res., vol. 28, No. 1, 36 pp., 8 figs. 6 pls.

1935. Principles of insect morphology. 667 pp., 319 figs. New York and
London.

WaRL, B.

1914. Uber die Kopfbildung cyclorapher Dipterenlarven und die postem-
bryonale Entwicklung des Fliegenkopfes. Arb. Zool. Inst. Univ.
Wien, vol. 20, pp. 159-272, 20 figs., 3 pls.

Weser, H.

1938. Beitrage zur Kenntnis der Uberordnung Psocoidea. I. Die Labial-
driisen der Copeognathen. Zool. Jahrb., Anat., vol. 64, pp. 243-286,
16 figs.

WEISMANN, A.

1864. Die nachembryonale Entwicklung der Musciden nach Beobachtungen
an Musca vomitoria und Sarcophaga carnaria. Zeitschr. wiss. Zool.,
vol. 14, pp. 187-336, 7 pls.

1866. Die Metamorphose der Corethra plumicornis. Zeitschr. wiss. Zool.,
vol. 16, pp. 45-127, 5 pls.

.
“
'
ty
ri
~
’ ¢
~ ?.
v 1
ay ore
i) ss “
4 ; « a
, ‘ ; , tad
fe! | hae y
) at ee

On
; ares

SMITHSONIAN MISCELLANEOUS COLLECTIONS
VOLUME 122, NUMBER 4

Roebling Fund

SOLAR VARIATION, A LEADING
dlc eno LE MMEIN TE

By
Cc. G. ABBOT

Research Associate, Smithsonian Institution

(PusticaTtion 4135)

CITY OF WASHINGTON
PUBLISHED BY THE SMITHSONIAN INSTITUTION
AUGUST 4, 1953

The Lord Baltimore Mress

BALTIMORE, MD., U. 8 A.

Roebling Fund
SOLAR VARIATION, A LEADING WEATHER

ELEMENT
By C. G. ABBOT

Research Associate, Smithsonian Institution

CONTENTS
Page
IntRopucTION. (Adverse contemporary opinion; three propositions contra
thereto; a suggested supporting theory; citations of pertinent litera-
[OPA Sea ee Aon One COS Pea TO DOES On oOo CORO OOOD COO ono OOD 2

Proposition I: Correlation of other solar phenomena with solar-constant
measures indicates the probable reality of solar variation. (With eight

RD HOPING GGUS.) fol siaic xcteipialepd & spoke © eines ae oia-@.so ehsseneke) suaieielar ata hahaa 5

Proposition II: Phenomena exist harmonious with a master period of 22}
years in the variation of solar-constant measures. (With six supporting
graphs and reference to Hale's discovery.) ooo cicce cca sejeocs esc nene nai 10

Proposition III: Integral submultiples of 223 years are regular periodicities
in solar variation. (With one supporting table and four graphs.)...... 18,20

INTERLUDE: On the purpose and accomplishments of the Smithsonian re-
Search on the vatiation of total! solar wmadiationy...c 4... scl «=+4-\ee a4 - 18

Proposition IV: Correlations exist between variations in solar-constant
measures and weather, not involving periodicities. (With three sup-
PORTS MORADMS Sl ovace sisic'g oe. 03 Spero else she a Meese erations, 6 wiape Sesion yo. open 25

Proposition V: Correlations exist between regular periodic changes in the
solar-constant measures and weather changes. (Interlude on lags: Lags
in weather responses to solar changes; impossibility of recognizing solar
influences without knowing the periods of variation of solar-constant
measures. With seven supporting graphs and three supporting forecasts
MANY NV EAU STANA COVANCEs) (a's ssa cv atayaye oy asst Ne OAS tis. cle eI ToT: 26

RUN ARVe Bosc ction, Lata echt ae le Ute OS oi Nadie gC ee 2 33

INTRODUCTION

On January 28, 1953, the American Meteorological Society devoted
the day to consideration of the influence of solar variation on weather.
An early speaker said he acknowledged the results of conscientious
studies of total solar variation, which had been made, as probably
sound. But the variations found appeared to be of the order of 1 per-
cent, or much less. No reasonable theory could show that these might
have important weather influences. He distrusted statistical conclu-
sions, unless grounded on sound theory. Statistics might show that it
is dangerous to go to bed, for the great majority of decedents died in
bed. The remainder of the panel appeared to agree with him that,
because percentage solar-constant variations are small, it is needless
to consider the possibility that variations of total solar radiation affect
weather importantly. The discussion was mostly confined to matters
relating to the high atmosphere, in the stratosphere and beyond. Sug-
gestions were discussed as to whether the large effects of solar changes
known to exist in the high atmosphere could be connected with
weather changes in the troposphere. No positive result was reached.

One gathered the impression that meteorologists are so firmly con-
vinced that variations of total solar radiation are of negligible weather
influence, and that statistical methods of proof are to be ignored, that
they probably do not read attentively any publications of the contrary
tendency. I do not agree that the last word has been said. I submit
several propositions.

1. Statistically derived results may be accepted, if well supported by
observation, without supporting theory. Kepler’s laws were accepted
statistically for many years before there was any supporting theory.

2. A conclusion may be accepted as a valuable working hypothesis,
without being proved in the rigid sense, e.g., that the square of the
hypotenuse of a right-angle triangle equals the sum of the squares of
the other two sides.

3. In lieu of theoretical support, to be supplied later, a statistically
derived proposition, A, may be adequately supported as a working
hypothesis, if accepted phenomena, B, C, D, E, — — — — which
stem from a related source, are harmonious with proposition A. I
propose to show that the proposition that the variations of solar radia-
tion are important weather elements is thus adequately supported.

SMITHSONIAN MISCELLANEOUS COLLECTIONS, VOL. 122, NO. 4

NO. 4 SOLAR VARIATION, WEATHER ELEMENT—ABBOT 3

Further support comes when forecasts with high correlation compared
to probable error result from such hypotheses. J depend strongly on
this paragraph in what follows.

However, I will venture a suggestion toward a theory of the matter.

1. It is commonly observed that the temperature is responsive to the
direction of the wind.

2. The direction of the wind depends on the orientation of the sta-
tion with respect to the cyclonic structure prevailing.

3. H. H. Clayton found, by his tireless statistical work, over a quar-
ter of a century ago, that the “centers of action,” about which the
cyclonic structure forms, are largely displaced in position on the earth’s
surface, as solar-constant measures rise and fall.

4. If this Clayton effect is accepted, the mystery is no longer why
large temperature and associated weather changes follow small per-
centage solar-constant changes, but rather why the “centers of action”
shift when solar-constant changes occur.

5. If meteorologists doubt the Clayton effect, they may find 30
years of 10-day solar-constant measures in paper No. 27, cited below,
which they may compare with weather maps of the period 1920 to
1950.

To provide a groundwork for reference, I first list certain pertinent
publications of the past 20 years. A book would be needed if one col-
lected all the evidence which supports the conclusion in hand. I shall
give below a few of the more telling references. Those interested may
find numerous others from the papers cited and from H. H. Clayton’s
earlier papers in the Smithsonian Miscellaneous Collections. Students
of research know that early work is often found partly erroneous as
later results come in. So it is here in some measure. Nevertheless, I
think all the papers cited here still retain features of some value and
interest. Fundamental to the whole pattern, however, is the paper
“Periodicities in the Solar-constant Measures,’’ Smithsonian Miscel-
laneous Collections, vol. 117, No. 10, 1952 (reference No. 27, below),
to which I particularly invite attention.

LITERATURE PERTAINING TO SOLAR RADIATION AND ASSOCIATED
PHENOMENA (BY ABBOT UNLESS OTHERWISE INDICATED)

1. How the sun warms the earth. Ann. Rep. Smithsonian Inst., 1933, pp.
149-179.
2. Sun spots and weather. Smithsonian Misc. Coll., vol. 87, No. 18, 1933.

1See Clayton, H. H., Solar radiation and weather, Smithsonian Misc. Coll.,
vol. 77, No. 6, June 20, 1925; also his Solar relations to weather, vol. 1, p. ix,
and vol. 2, p. 384, 1943.

16a.

20.

27.

SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

Solar radiation and weather studies. Smithsonian Misc. Coll., vol. 94, No. 10,
1035.

Weather governed by changes in the sun’s radiation. Ann. Rep. Smith-
sonian Inst., 1935, pp. 93-115.

Rainfall variations. Quart. Journ. Roy. Meteorol. Soc., vol. 61, pp. 90-92,
1935.

The dependence of terrestrial temperature on the variations of the sun’s
radiation. Smithsonian Misc. Coll., vol. 95, No. 12, 1936.

Further evidence of the dependence of terrestrial temperatures on the varia-
tions of solar radiation. Smithsonian Misc. Coll., vol. 95, No. 15, 1936.

Cycles in tree-ring widths. Smithsonian Misc. Coll., vol. 95, No. 19, 1936.

Some periodicities in solar physics and terrestrial meteorology. Zvlastni
Otisk, vol. 18, pts. 1-2 (54-55), Io pp., 1938. Prague.

Solar variation and the weather. Nature (London), vol. 143, p. 705, April
1930.

. The variations of the solar constant and their relation to weather. Quart.

Journ. Roy. Meteorol. Soc., vol. 65, pp. 215-236, 1939.

. Variations of solar radiation (Dixon). Quart. Journ. Roy. Meteorol. Soc.,

vol. 65, pp. 383-384, 19309.

. On periodicities in measures of the solar constant (T. E. Sterne). Proc.

Nat. Acad. Sci., vol. 25, pp. 559-564, 1939.
On solar-faculae and solar-constant variations (H. Arctowski). Proc. Nat.
Acad. Sci., vol. 26, pp. 406-411, 1940.

. An important weather element hitherto generally disregarded. Smithsonian

Misc. Coll., vol. 101, No. I, 1941.

On solar-constant and atmospheric temperature changes (H. Arctowski).
Smithsonian Misc. Coll., vol. 101, No. 5, 1941.

Solar relations to weather (H. H. Clayton). Vols. 1 and 2, 1943. (Privately
printed.)

. A 27-day period in Washington precipitation. Smithsonian Misc. Coll., vol.

104, No. 3, 1944.

Weather predetermined by solar variation. Smithsonian Misc. Coll., vol.
104, No. 5, 1944.

The solar constant and sunspot numbers (L. B. Aldrich). Smithsonian Mise.
Coll., vol. 104, No. 12, 1945.

. Correlations of solar variation with Washington weather. Smithsonian Misc.

Coll., vol. 104, No. 13, 1945.

. The sun makes the weather. Scientific Monthly, vol. 62, pp. 201-210, 34I1-

348, 1946.

. The sun’s short regular variation and its large effect on terrestrial tempera-

tures. Smithsonian Misc. Coll., vol. 107, No. 4, 1947.

. Precipitation affected by solar variation. Smithsonian Misc. Coll., vol. 107,

No. 9, 1047.
Solar variation attending West Indian hurricanes. Smithsonian Misc. Coll.,
vol. 110, No. 1, 1948.

. Magnetic storms, solar radiation, and Washington temperature departures.

Smithsonian Misc. Coll., vol. t10, No. 6, 1948.
Short periodic solar variations and the temperatures of Washington and
New York. Smithsonian Misc. Coll., vol. 111, No. 13, 1949.
Periodicities in the solar-constant measures. Smithsonian Misc. Coll., vol.
117, No. 10, 1952.

NO. 4 SOLAR VARIATION, WEATHER ELEMENT—ABBOT 5

28. Important interferences with normals in weather records associated with
sunspot frequency. Smithsonian Misc. Coll., vol. 117, No. 11, 1952.

29. Solar variation and precipitation at Peoria, Illinois. Smithsonian Misc.
Coll., vol. 117, No. 16, 1952.

30. Solar Aktivitat und Atmosphare (H. Koppe). Zeitschr. fiir Meteorol.,
vol. 6, Heft 12, pp. 360-378, December 1952.

31. Solar variation and precipitation at Albany, N. Y. Smithsonian Misc. Coll.,
vol. 121, No. 5, 1953.

32. Long-range effects of the sun’s variation on the temperature of Washington,
D. C. Smithsonian Misc. Coll., vol. 122, No. 1, 1953.

PROPOSITION I

Correlation of other solar phenomena with solar-constant measures
indicates the probable reality of solar variation

Applying my criterion No. 3, above, I shall now cite evidence that
variation observed in Smithsonian solar-constant measures is associ-
ated with variation (a) in solar faculae areas; (b) in sunspot num-
bers; (c) in calcium flocculi areas; (d) in incidence of magnetic
storms ; and (@) in ionospheric data.

Dr. H. Arctowski, of Poland, was attending a meeting in Washing-
ton when his savings and work were swept away by the invasion of his
country. I suggested to John A. Roebling that it would be helpful if so
eminent a European meteorologist should examine our case for the
variation of the sun and its control of weather. Mr. Roebling consented
to support this project. After several months Dr. Arctowski told me:
“T believed in neither proposition. But I determined to give them a
fair trial. When I found them unsupported, I intended to tear up my
papers and resign. I could not take money under false pretenses.” Af-
ter a brief time, however, Dr. Arctowski came to believe in both propo-
sitions, and said: “I have become more enthusiastic about them even
than Dr. Abbot himself.”

a. Referring to Dr. Arctowski’s paper, reference No. 14 above, I
reproduce his figures I, 3, 4, and 5 as figures I, 2, 3, and 4 herein.

b. Referring to L. B. Aldrich’s administrative report on the Astro-
physical Observatory for 1952 (Rep. Secretary Smithsonian Inst.,
1952, p. 131), I reproduce here as figure 5 his figure showing the
correlation of solar-constant measures with sunspot numbers.

c. Referring to my paper “Weather Predetermined by Solar Vari-
ation,” reference No. 18 above, I reproduce figure 6 of that paper as
figure 6 here. I call attention to the similarity of the full and dotted
curves of the figure. This similarity indirectly proves the correlation
claimed as c, above. Each month the curves represent means of effects
on numerous occasions.

6 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

FEB. £926 48 20 23

Fic. 1.—Variations of the solar constant and of areas of solar faculae. Daily
solar-constant values for February and March, 1926, and areas of faculae.

OCT, /929

=. Pe era
Ae

eG A Al A

. necewure
oe eee en
Pe Ot
SRR Se OR SES

Fic. 2.—Discontinuous trends in solar constant and solar faculae. Solar constants and
faculae, October, November, and December, 1929.

NO. 4 SOLAR VARIATION, WEATHER ELEMENT—ABBOT 7

-q>~
-
Ze Mik
s
4; N
id ‘

A
UN

ieee a fal
_ JESS Lal
af Ath

Vea a
yi fe NC
ine ali

2400

“4 =! ° #2 2

Fics. 3 AND 4.—Time relations between maxima and minima in the solar constant
and solar faculae. Means of faculae and solar constants for the 5 days before and the 5
days after the dates of 72 selected days of maxima and 82 days of minima of solar
constants.

1.950

& &

Solar Constant Values

4
~~

(The number near each point
indicates the number of monthly
means included in each group.)

. 7
Sunspot numbers

Fic. 5.—Monthly mean values of the solar constant compared with monthly means of
sunspot numbers for the same days.

AEE

ah
oy
va
a
a
4
aN
ry
i i]
Z mig ti
inl
Bi
ik ws ha SS e
Hil :
TT ial) oh aw
4 AA eA Se
Se a“ wea a I a
A cae
maa mks ENS pe eer
MeCN
Paleo ds (rie eat

ra i am Fo
SS=SsaecSe
= ener. A ica 2
Scar ha ee A A
A i

ate ize
Bh 9 6 a

PES
ala

ae ee ee
(ie eS,
Sal ey AS] TT TA
ak eM “a pA NSAI SEL JE NEN
PON EZ Ne TC
Pas 006s En Na a
; jae SA
IP Le OH
Cees encom oes i PRBMPEN
Se VRE 2h Se ea ee ay
RRB EE AEP EEA PESO
iio 2 ES a a No i

Fic. 6.—Average marches of temperature departures, Fahr., at Washington, DD sGe
accompanying sequences of solar change (a) of the solar constant in years 1924 to
1936; (b) of character figures for solar calcium flocculi in years I910 to 1937, for
months January to December. Ordinates are temperature departures; abscissae are
days from beginning of solar-constant sequence. Flocculi band curves are displaced 2
days to right. Temperature changes following rising solar radiation above, falling
radiation below.

NO. 4 SOLAR VARIATION, WEATHER ELEMENT—ABBOT 9

d. Referring to my paper “Magnetic Storms, Solar Radiation, and
Washington Temperature Departures,” reference No. 25 above, |
reproduce figure 2 of that paper as figure 7 here. I call attention to
the sharp depression of the solar-constant measures by 4 percent on the
day of the height of the magnetic storm. It is the mean result repre-
senting 53 great magnetic storms over the years 1923 to 1946.

Fic. 7—Depression of solar constant attending severe magnetic storms.
Abscissae, days before and after height of storm; ordinates, solar constant (to
be prefixed by 1.9).

I refer also to the note by F. E. Dixon of the Imperial College of
Science and Technology, reference No. 12 above, and to H. Koppe’s
conclusions, reference No. 30 above.

e. Referring to my paper “The Sun’s Short Regular Variation and
Its Large Effect on Terrestrial Temperatures,” reference No. 22
above, table 7 of that paper is from values of the ionospheric quantity,
Fe, furnished me by Dr. John Fleming from records of the ionospheric

10 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122
stations at Huancayo and Watheroo for the years 1938 to 1944. As
given there, the records have been cleared of average monthly march

and of sunspot influence, and are as follows:

TABLE 1.—lonospheric data, Fe. Monthly and sunspot effects removed

Months 1938 1939 1940 1941 1942 1943
Analy uaeiecriced cence 378 369 341 352 338 325
Mebruanyyee cetacean 375 360 344 351 328 330
IMianchis teens ecteen os eae 376 344 347 340 340 334
NOTH (ican vida de soe cae Oe 384 343 331 322 323 B17
May Pisa tertisnterman area eect bie 370 336 313 206 303 204
BLS RES + Ae RE oP cata Me 342 328 311 203 206 286
Jialy; S85 bd ak eats eek 342 334 313 304 300 287
TS ER US) AES TATE OE ATTRA SE 349 345 330 318 302 203
September saree as screstakis 354 366 344 325 312 209
October? Mevcicrscheiaci:s ets 361 3061 353 330 320 304
INOVemDeny eapetceeieenreties 370 352 359 332 328 314
December Gy. siicyeciacposes- 375 346 357 334 329 322

I shall show in a later section that variations in solar-constant meas-
ures, among many others, have regular periods of 6-1/30, 9-7/I0,
114, and 13-1/10 months. I do not use longer periodicities than
these here, because the ionospheric data are of too brief duration. In
figure 8° I show the mean curves representing these periods in the
ionospheric quantity Fe, computed from the table just given. The four
curves are, respectively, means of 12, 7, 6, and 5 repetitions of the
periods. Their amplitudes, respectively, are 4, 44, 9, and 63 percent of
mean Fe. The amplitudes of the corresponding curves of variation of
the solar-constant measures (see reference No. 27 above) are, respec-
tively, 12/100, 10/100, 17/100, and 11/100 percent, being means ob-
tained from 16 to 28 repetitions, according to length of period.

With these correlations shown in figures 1-8, I rest my claim that
criterion No. 3 is satisfied as regards the reality of solar variation.
Other evidence could be given, but this seems sufficient to establish as
a reasonable working hypothesis that there is really a variation in the
output of total radiation from the sun.

PROPOSITION II

Phenomena exist harmonious with a master period of 223 years in the
variation of solar-constant measures

I shall now show that (a) the features of solar-constant measures
themselves of 1924 to 1927 are approximately repeated after about 23

2 Figure 8 will be referred to again later.

NO. 4 SOLAR VARIATION, WEATHER ELEMENT—ABBOT Dal:

years in the years 1947 to 1950; (0b) this 22$-year period is also found
in sunspot frequency; (c) also in the magnetic polarity of sunspots ;
(d) also in the thickness of tree rings; (e) also in terrestrial tempera-
tures ; (f) also in terrestrial precipitation.

a. To show the master period in solar variation, I reproduce here
as figure 9 figure 4A from my paper “Periodicities in the Solar-con-
stant Measures,” reference No. 27 above. The amplitude is 0.9 percent.

Fic. 8.—Variation of Fe in solar periods of 6-1/30, 9-7/10, 114 and
13-1/10 months.

b. I reproduce here as figure 10, figure 10 of my paper “Solar Radi-
ation and Weather Studies,” reference No. 3 above. It will be found
that alternate sunspot-cycle areas, i. e., the right-hand curves of figure
10, are all greater in area included by the curves than the left-hand
areas. So the double of the usually termed ‘‘114-year cycle” in sunspot
frequency is also a sunspot period. Note that a line through sunspot
minima would incline to the left, as years increase, which shows that
period to be less than 23 years.

c. Dr. George E. Hale discovered over 40 years ago the reversal of

I2 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

polarities of sunspots with alternate recurrences of the 114-year cycles.
That is, he discovered a period of about 223 years in sunspot mag-
netism.

d. I reproduce figure 30 from my paper just cited (No. 3 above) as
figure 11 here. It shows similar features in the march of tree-ring

240

ia’)
Po

Departures from 1.90 calories in ten-thousandths.

200

Fic. 9.—Comparison of solar constants 1924-1927 (heavy lines) and
1947-1950 (light lines).

widths in southern California for four successive cycles of 23 years
each. These features stand out clearly in the mean curve at the bottom
of figure IT.

e. I reproduce here as figure 12, figure 1 of my paper “Some Peri-
odicities in Solar Physics and Terrestrial Meteorology,” reference
No. 9 above. The figure traces 23-year cycles in the temperature of
St. Petersburg, Russia, from 1752 to 1912, and also brings out the
double period of 46 years.

f. I reproduce here as figure 13, figure 22 of my paper “Weather

AC
(iol
al mem

Se
Sabi te
pad
HH

Ge
hae el a a
-CCOA ha!
A

SEA
ae inl
Cae

‘aa

fae
et
Ba
ig
x
i
i
a
a

os OS
e's 1b SAR

AMMA Unt TT TT TT
PAN SS 8 8S ea
1

Rage ale 2a

Se deals ba a ell hana al ST

Fic. 10—Wolf sunspot numbers, 1810-1933.

1875 -1897

23-YEAR PERIOD IN TREE RINGS
FROM See SO. CALIFORNIA
15

Fic. 11.—Cycles of 23 years in tree-ring widths. Individual cycles of 23
years show features which are found preserved in the mean of four cycles, or

g2 years.

1752 Ss

—

00 HY Wy pall a Ve

VOL. 122

SMITHSONIAN MISCELLANEOUS COLLECTIONS

*saoAD Sutpadeid JO uoreIapisuos
Aq apeut (gf61 wos ‘outy uly} :PE61 WorZ ‘oUT] PeHop) sysedso10,~ ‘saTIAD SAISSaDONS UL sainyeoy AvpIUNIs jUasoidot
sioyayT ‘“sopoAo steak-€z ur pasuesse ‘survow Suruuns yWUOU- -§ Aq poyjoous “TIT ‘elloag je uoneydie1g—€I “914

[25s 2S ESS ee ae ee kes Fe ee de is et ee ee
ee eee ee eee
Pe IN 7 i A
| NT Mery al UT | OR ‘ay lh CR a
[ATU es SN alN
zsh A A |
aaa | ee ee ee

oo!

ae a ik vee A eR =

szg!

4 OT A ye UT At oh A
PAA Se
ae A Pa a SS ee Re
SS Caio
a LAM TN Me te fl MAM LT ta PP ese
PPV A ie VO NIN ae ee).
Se ee
SSS SSS Sa Se SS SSeS

Le)
©

&¢

AWWYON 1N3903d

NO. 4 SOLAR VARIATION, WEATHER ELEMENT—ABBOT EF

Predetermined by Solar Variation,” reference No. 18 above. It shows
how the features of precipitation at Peoria, Ill., tend to repeat them-
selves at intervals of slightly less than 23 years.

PERCENTAGES FROM NORMAL PRECIPITATION

Fic. 133A.—Mean 223-year cycle in Southern New England precipitation. 1750
to 1931. Mean of 8 cycles.

g. I reproduce here as figure 13A, figure 1 of my paper “Rainfall
Variations,” reference No. 5 above.®

Many other harmonious phenomena might be brought forward, but
sufficient has been shown to support the working hypothesis of a 223-
year period in solar variation.

3 The New England drought of 1952 falls in timely with this curve.

18 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

PROPOSITION III

Integral submultiples of 223 years are regular periodicities in
solar variation

I shall show that (a) at least 23 such periods were found by tabu-
lating solar-constant measures of the years 1924 to 1950; (0) in tabu-
lations of the longer submultiple periods, integral submultiples of these
long periods, which, of course, are also integral submultiples of the
master period of 223 years, appeared plainly in the mean values; (c)
several of these submultiple periodicities were sought for and found
in ionospheric records; (d) by analogy to harmonics in musical
sounds, since three integral submultiples of 22% years were discovered
as supposedly isolated periods in solar variation 20 years ago, it is rea-
sonable to expect that a large number of integral submultiples of 223
years will be found to occur as regular periodicities in solar variation.

Before disclosing these evidences I insert an account of the purpose
and results of the Smithsonian solar-constant campaign.

INTERLUDE

On the purpose and accomplishments of the Smithsonian research
on the variation of total solar radiation

Measurements of the solar constant of radiation were made by Dr.
S. P. Langley at Allegheny, and then in his famous expedition to
Mount Whitney, Calif., in 1881. Becoming the third Secretary of the
Smithsonian Institution in 1887, one of his first acts was to found the
Astrophysical Observatory. After completing its first research on the
infrared line and band spectrum of solar radiation, in the year 1902
Dr. Langley directed that the measurement of the solar constant of
radiation should be undertaken, not especially for fixing that constant,
but rather, by a long series of measurements, to find if it is a variable.
His impelling thought was that in solar variation might lie a hitherto
unknown weather element of great significance.

Dr. George E. Hale cordially seconded this project, and, after the
establishment of Mount Wilson Observatory, he urged Langley to
undertake the research there. Accordingly I was sent out in 1905, and
excepting 1907, 1917, 1918, and 1919, made measurements of the solar
constant there every year up to and through 1920. L. B. Aldrich ob-
served there in 1917, 1918, and 1919. We also, following Langley’s
original suggestion, erected a tower telescope with mirrors, forming a
solar image 20 centimeters in diameter. This image was allowed to
drift across the slit of the spectrobolometer. Every day of solar-con-

NO. 4 SOLAR VARIATION, WEATHER ELEMENT—ABBOT 19

stant measurement, in the years 1913 to 1920, we made automatic
drift curves, showing the distribution of energy in many wavelengths
across the east-west diameter of the sun. This, too, was done expressly
to discover variations useful for weather forecasting. A positive corre-
lation was discovered between the solar constant and solar-contrast
measures. (See also, in that connection, paper No. 27 cited above.)

In 1917, H. H. Clayton, then chief forecaster for Argentina, in-
formed Dr. Walcott, then Secretary of the Smithsonian, that, by com-
bining into large groups the Mount Wilson solar-constant measures,
he had secured sufficient accuracy in mean values to show direct cor-
relation with weather elements. This led us to establish a solar-constant
station at Calama in the nitrate desert of Chile. Soon after, with John
A. Roebling’s aid, it was removed to Mount Montezuma, at 9,000 feet
altitude. Since 1920, when possible, daily measures of the solar con-
stant of radiation have been made there and also at other Smithsonian
observing stations on high mountains in arid lands. Mr. Clayton pub-
lished many papers showing the correlation of solar-constant measures
with weather. After his return to Massachusetts he conducted pri-
vately for many years, till his death, a long-range weather-forecasting
business, based on solar variation, and had many paying clients.

About 20 years ago, having a long series of 10-day mean values of
the solar-constant measures, I made a chart of them extending the
length of my office. Standing at a distance, I sought to discover repeti-
tions of configurations in the variations. I noted a small regular varia-
tion of slightly more than 8-months period. Proceeding similarly, I
discovered regular periods of variation of about 11} months, and of
about 39 months. It then occurred to me to find the least number of
months of which, within the errors of determination, these three peri-
ods would be approximately integral submultiples. The number 273,
seven times 39, 24 times 114, and 34 times 8, seemed best. This num-
ber, 273 months, recommended itself as a solar period, because it is
approximately twice the sunspot cycle and thus equal to Hale’s mag-
netic cycle in sunspot polarities.

Having three integral submultiples of 273 months represented in the
variation of solar-constant measures, I naturally sought for others.
This search, as completed for the present, is described in my paper
“Periodicities in the Solar-constant Measures,” published in 1952, ref-
erence No. 27 cited above. As I shall show, it would be quite tmpossi-
ble for meteorologists to discover these regular periodicities in weather
elements had they not first been found in solar variation.

In passing, I remark that it greatly strengthens our case for the va-
lidity of solar-constant work that the 10-day means, covering the 30-

20 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

year interval 1920 to 1950, which yielded the results published in
paper No. 27 cited above, rest throughout the 30 years on two sta-
tions in opposite hemispheres. Winter in California coincides with
summer in Chile. For several years the 10-day means from Mount St.
Katherine, in Egypt, also contributed to the results published in paper
No. 27.

I now proceed with the correlations promised above.
a. I quote, as table 2, part of table 1A from paper No. 27 cited

above.

TABLE 2.—Periodicities in solar-constant observations

Period Period
Period Amplitude Fraction Period Amplitude Fraction
Months Percent of 272 Months Percent of 272
2h 0.05 1/127 13-1/10 0.11 1/21
3-1/20 0.05 1/90 152 0.09 1/18
44 0.06 1/63 223 0.07 ie
5-1/18 0.05 1/54 243 0.12 I/II
6-1/30 0.12 1/45 304 0.13 4
7 0.08 1/39 345 0.15 3
8-1/14 0.06 1/34 39 0.20 +
9-1/10 0.08 1/30 454 0.13 * 4
9-7/10 0.10 1/28 544 0.13 t
10-6/10 0.06 1/26 68 0.25 4
114 0.17 1/24 Ol 0.12 4
11.43 0.11 1/24 272 Hi

12.0 0.20

* This figure for amplitude was fixed before extraneous periods were removed, as in
gure 14.

b. I now show, as figure 14, six broken curves and one smooth
curve, all relating to the period of 454 months in solar variation. Curve
A represents the direct mean of seven repetitions, from the monthly
means of the solar-constant measures, of the 454-month period. It is
plain that it contains a period of 454+3 months. This period being
removed, we have curve B. Now a period of 454~4 months is dis-
covered. Removing it from curve B we have curve C. Then a period
of 454+2 months seemed indicated. Removing it from curve C, we
have curve D. There now appears a period of 4535 months. Re-
moving it we have curve E. It discovers a period of 454+7 months.
Removing it, we have curve F. Curve F contains a period of 455+ 13
or 34 months, but I do not remove it. For it is now easy to draw the
smooth curve G, which is the real curve of the 454-month period.

As will be seen, the researcher has no option. Once started he
must follow this path. The periods discovered in solar variation by

NO. 4 SOLAR VARIATION, WEATHER ELEMENT—ABBOT 21

figure 14 are 4, 1/12, 1/18, 1/24, 1/30, 1/42, and 1/78 of 223 years.
c. I again invite attention to figure 8, which shows that periods ob-
served in solar-constant variation of 6-1/30, 9-7/10, 114, and 33-1/10
months also occur in the ionospheric data on Fe given in table 1.
There is another aspect of this matter of Fe which adds to its evi-
dential quality. From solar-constant measures, as set forth in the paper
cited above as No. 27, the times of maxima and minima for solar radia-

—

Fic. 14.—The 453-month period in solar variation.

tion in the year 1938 are as follows (December, when given, is Decem-
ber 1937) :

Peto ie sites, oa6 6-1/30 9-7/10 114 13-1/10
Mascitnan sonics 4 January December-January September March
Minimas eebian. se March April-June March June-July

From figure 8, here, the times of maxima and minima for Fe in the
year 1938 are as follows:
Reriod ia54h2 4st: 6-1/30 9-7/10 114 10-1/10

Midsximalciinee sins March April January-April August
AMA ic 2s a <te December December August-September March

22 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

Thus we find, to within the error of determinations, that for all
four subperiods maxima in radiation are simultaneous with minima in
Fe, and vice versa. This is, of course, exactly the relationship which
we should expect, if the supposed periodicities are real.

I have additional evidences of correlation of solar periods and iono-
spheric observations. From the publication of the National Bureau of
Standards entitled “Ionospheric Data,” I have tabulated the mean
monthly values of the quantity h'F2 for the hours 11, 12, and 13, from

Jan. March May July Sept. Nov.
Fic. 14A.—The yearly march in the ionospheric quantity h'F2.

September 1944 to December 1952. Taking the general mean of these
IOI mean monthly values of h'F2 for the hour of noon, it comes out
314. I computed the departures from this mean value, and arranged
them by months. Taking the means of these monthly departures over
8+ years, they are as represented in figure 14A. I then removed this
average annual march from the departures. Next, the corrected de-
partures were plotted against the appropriate sunspot monthly Wolf
numbers. The resulting graph (not shown here) was well represented
as a straight line, yielding the sunspot correction 0.22 (Wolf No.
—100). Applying this sunspot correction, I obtained the corrected

NO. 4 SOLAR VARIATION, WEATHER ELEMENT—ABBOT 23

departures of h'F2 to be compared to the subordinate periodicities in
solar variation.

In figure 14B I give graphs of 12 periodicities of the corrected iono-
spheric character h'F2 and based on January 1940. These are solar-
constant periods. It seemed desirable to remove from the periodicities

pal
§
ST

‘Tre

ny
ba
ba
mi
ra
re

i)
aa =
=

©) (3 ! v) 9

Frc. 14B—Submultiples of 223 years found as periods in this ionospheric quantity h'F2.
Periods in months.

TaBLE 3.—Characteristics of periodicities of h'F2

Beriods setataccitee.es + 44 5s 6-1/15 7 8k 9-1/10
Amplittides” 22... B77, 17 at5 2.9 4.2 2.9
No. of columns....... 23 19 16 14 12 II
Reriods;, he eats ees 92 10-I/I0 10-6/10 11% 13-I/10 152
Amplitudes ......... 5.3 3.5 6.3 5.3 4.7 12.2
No. of columns....... 10 9 9 9 8 6

of 114 and 13-1/10 months superriding periodicities of 11$+2,
13-1/10+2, and 13-1/10+3 months. In these cases the original re-
sults are shown dotted, the results cleared of superriders are shown
heavy and full. Table 3 gives the characteristics of these curves of
figure 14B. The periods are in months, the amplitudes in percentages

of 314.

24 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

As regards the comparative phases of the periodicities in the solar
radiation and in the h'F2 data, we should expect them to agree. For,
as figure 14A shows, the higher radiation of summer months brings
higher values of htF2. But a comparison of phases is uncertain for
several reasons. First, the periods of solar variation all start from
August 1920, while the basis for htF2 stems from September 1944. In
the intervening 24 years there are 67 repetitions of 43 months. An
error of I percent in the period would shift the phases by almost 3
months. Exactly the same percentage consideration (1 percent cor-
responds to 3 months) applies to all the other periods. They are none
of them certain to I percent. Second, the plots of solar-constant and
h'F2 periodicities show such ragged outlines that the phases of max-
ima and minima in both quantities are uncertain by one or several
months. The 10-1/10-month period must be omitted, in comparing
phases, for lack of solar-constant data. With these considerations be-
fore us, only the two periodicities, 13-1/10 and 15 months, are found
unreasonably discrepant from the expected agreement of the phases.
In these two cases the repetitions of the htF2 data are so few that the
mean values may not indicate the phases as they should. The other
nine periodicities show phases close to agreement, as expected.

I remark that for the shorter periods, where there are many columns
of repetitions from which to form the means shown as graphs in figure
14B, the curves are very satisfactorily smooth. When the number of
columns becomes small, naturally the curves are ragged, for each peri-
odicity is affected by the influences of all the others, including many
not shown here, and only as the means of very large numbers of repeti-
tions could these other periodic influences be eliminated. It was im-
practicable to search for longer periods than 15% months with so few
ionospheric data, but all the solar periods given in table 2, above, from
44 months to 154 months are represented in figure 14B and table 3.
Since, as I have shown, the quantities Fe and h'F2 are plainly respon-
sive to the periodicities found in the solar-constant measures, it is
probable that the other ionospheric quantities must be so also.

d. All the periods given in table 2 are integral fractions of 22# years,
to within experimental error of determining their lengths.

I conceive that criterion 3 is satisfied as regards the working hypoth-
esis of the existence of regular periodic solar variations, with periods
integrally related to 223 years.

NO. 4 SOLAR VARIATION, WEATHER ELEMENT—ABBOT 25

Of weather aspects

Hitherto I have treated only of variations in solar-radiation meas-
ures, in correlation with other phenomena, and intercorrelations among
variations of solar-radiation measures themselves.

I come now to correlations of variations in solar-radiation measures
with weather changes. These are of two kinds: A, Correlations not
involving periodicities ; B, correlations involving periodicities.

PROPOSITION IV

Correlations exist between variations in solar-constant measures and
weather, not involving periodicities

I shall cite: (a@) West Indian hurricanes correlated with depression
of solar-constant measures. (b) Rising and falling sequences in solar-
constant measures and correlated temperature changes. There are
published nearly 100 independent correlations of this sort which might
be cited, all involving temperature changes of several degrees Fahren-
heit, and, as far as their depending on solar-constant variations is con-
cerned, the result is backed up by the fact that sequences of variation
of the areas of calcium flocculi, observed at Ebro, are associated with
marches of Washington temperature nearly identical to those corre-
lated with solar-constant changes. (c) Features of precipitation re-
peated approximately in 22-year intervals.

a. I reproduce here, as figure 15, figure 1 of paper No. 24 cited
above. Counting from first reports of 45 West Indian hurricanes of
the years 1923 to 1946, the solar-constant measures dropped sharply
by + percent, on the average, on zeroth day. A solar-radiation de-
pression appears to act as a trigger to set off a hurricane when con-
ditions are ripe.

b. invite attention again to figure 6, referred to above. This shows
24 independent correlations between solar-constant changes and
Washington temperatures. The temperature changes shown in figure
6 are opposite for rising and falling solar-constant, or solar-flocculi,
measures. The temperature changes shown, which are averages for
great numbers of occasions for all 12 months of the year, range from
2° to 10° Fahrenheit. Similar correlations have been published for
several other cities, making nearly 100 independent correlations of
this kind.

c. I invite attention again to figure 13. The 223-year master period
in solar variations includes many precipitation features repeated ap-
proximately from cycle to cycle.

26 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

Criterion No. 3 appears to be satisfied as a working hypothesis re-
garding correlations of solar-constant changes with weather, both as
to temperature and precipitation, as well as with hurricanes.

002

Fic. 15—Mean solar-constant values preceding and following first reports of
West Indian hurricanes. Abscissae, days before and after report dates; ordinates,
solar-constant values, to be prefixed by 1.94.

PROPOSITION V4
Correlations exist between regular periodic changes in the
solar-constant measures and weather changes
INTERLUDE ON LAGS

Before proceeding with this section, attention must be directed to
lags in the responses of weather to changes of solar radiation. It is
common knowledge that maximum temperatures, both diurnal and
annual, lag behind the highest intensities of insolation. Such lags differ

4 This section imports a new and powerful element in meteorology.

NO. 4 SOLAR VARIATION, WEATHER ELEMENT—ABBOT 27

from place to place, and from season to season. They also differ in a
secular fashion. These differences in lag attend differences in con-
figuration of the land; differences in human occupation of the land;
differences in the transmissive conditions of the atmosphere to solar
rays; and differences in the “greenhouse” properties of the atmos-
phere.

With these facts in mind, it should be expected that weather re-
sponses to other regular periodic solar-radiation changes will alter in
phases from place to place; with the season of the year; with the
prevalence of sunspots, since the varying intensity of solar ionic bom-
bardment of the earth’s atmosphere tends to alter its transmissibility ;
and, in a secular fashion, over long spans of years, because human
occupation of the land differs.

It is not possible to fully anticipate and allow for these changes of
phases of responses of weather to regular periodic changes in solar
radiation. As a reasonable approach, I am accustomed to tabulate
weather data separately in three parts of the year, viz, January to
April; May to August; September to December. Also I tabulate
separately for sunspot numbers = 20 Wolf numbers. Also I tabulate
separately for years before and after 1900.

There is still another consideration. Weather records are custom-
arily published with respect to normal values. These normal values,
as published, are computed as monthly means of all values over a very
long span of years. It is found, however, that normal values differ im-
portantly when sunspot numbers are 2 20 Wolf numbers. Hence,
before using weather records to compare with regular periodic changes
in solar radiation, I compute two sets of normals, for sun spots 2 20
Wolf numbers, and compute two sets of departures, accordingly. (In
this connection, see paper No. 28 cited above.)

With all these variable, and not entirely controllable, factors affect-
ing phases of response of weather to regular periodic solar changes,
it is quite impossible for meteorologists to discover solar control by
mere tabulation of weather records. For in tabulations neglecting
these variable factors, all regular periodic weather changes would be
hopelessly mixed up by unknown phase changes as well as by inter-
ference between many periods. It 1s indispensable to know the solar
periods first, and to make an organization of the tabulations, such as I
have described.

Fortunately phase-changing effects are much less troublesome with
the longer solar periods. For as the period increases, fewer and fewer
numbers of repetitions of it can be found in the weather records, and
so mean results are, from that point of view, less and less satisfactory.

28 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

It would, indeed, be futile to subdivide the tabulations as extensively
as stated above, when tabulating long periods. Retaining only the
secular subdivision, before and after 1900, I give up all the other sub-
divisions for periods exceeding 20 months. Still, an embarrassment
remains for shorter periods, because, with a twelvefold subdivision of
records, there are too few repetitions to give strong means. I there-
fore make the questionable assumption that, though phases change
with time of the year, prevalence of sunspots, and years before and
after 1900, the amplitudes and forms of responses to regular periodic
solar changes will not change greatly. Hence, after computing these,
I reduce the six tabular means for one sunspot condition to the same
phase, and take the general mean, in a common phase, as representative
of amplitude and march. Though, as remarked, open to question, this
is better than using the weak individual mean values. Thus I obtain
generalized means for sunspot numbers 2 20. When I apply them to
forecasting, I readjust their phases to that proper to each of the 12
tabulations.

With these remarks, I am prepared to show the evidence that regular
periodic solar variations control weather.

a. In a paper on the temperature of Washington, D. C. (reference
No. 32), and in paper No. 31 listed above, I show that, both as to tem-
perature and as to precipitation, over 20 regular periodicities in solar
variation are also found in weather records of 86 years, 1854 to 1939,
as tabulated with regard to the principles explained above. These
numerous regular periodicities range separately to maximum ampli-
tudes of 2° F. as regards temperature and from 5 to 25 percent as
compared to normal precipitation.

b. When all known periodicities are synthesized with due regard to
phases, so as to make up ostensibly the whole weather complex, these
numerous, separately determined, regular periodicities of variation
from the normal over long terms of years exhibit approximately the
same amplitudes of variation in their syntheses as the observed
weather.

c. Such syntheses of total weather over long terms of years show
generally the same principal features, and nearly in the same phases,
as observed weather.

d. Forecasts, 50 or more years in advance, from such syntheses
show fair agreement with observed weather.

e. In brief support of these propositions, urgently referring to
original papers for further evidence, I reproduce here as figures 16, 17,
and 18, figures 1, 2, and 5 of the paper on Washington temperature,

NO. 4 SOLAR VARIATION, WEATHER ELEMENT—ABBOT 29

(No. 32, referred to above), and as figures 19, 20, 21, and 22, figures
I, 2, 3, and 5 from paper No. 31 cited above.

Figures 16 and 17 show the 12 independent determinations of the
Washington temperature periodicity of 13-6/10 months. Figure 16,
relating to sunspot numbers > 20, gives pairs of determinations from
Washington temperature records of 1854 to 1899 and 1900 to 1939,

seks a
Se
:

(

Gio ae Re

Ae
ann

ave

TL BAY

UREN

~~
;
= Z| Pid

HL Re

are

Wa \
Vila

Fics. 16 AND 17.—The periodicity of 13-6/10 months in Washington temperature
departures. Ordinates in hundredths degree Fahr. The symbols O.K., v and *%
indicate phase changes in getting means.

respectively, for the three seasons January to April, May to August,
and September to December, all adusted to a common phase and
averaged. Figure 17 shows the same for sunspot numbers <20. It
will be seen that the thick-lined mean curves for sunspot numbers
=> 20 are similar in form, but differ in phase, and have ranges of
about 14° F.

Figure 18 is a synthetic prediction, 50 years in advance, of the tem-
perature of Washington, 1950 to 1952, based on 20 regular periodici-

30 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

ties determined from monthly records of the years 1854 to 1939, cen-
tering about 1900. The prediction is in the thin line. The thick line is
the event. The two scales of ordinates, separated 2° F., indicate, as
expected, that Washington is now warmer than 50 years ago. I should
add that all the data are smoothed by 5-month running means. The
coefficient of correlation between forecast and event is 50.4 + 9.7
percent.

Figure 19 shows the 454-month period, computed as a straight mean
of all repetitions of that period, in precipitation records at Albany

BRSEESae8n roo
Ti

cH
BR SEPP SSR eS PROS Seos oF SSA eeae.
|S SSS SSSR See eee eee Ue 8 eS eeeeeel TTT ae | Fad
SER Re Bees 2X8 SP. eee eee !/ ARES Cees eee
ae 1 | JS). - —_ 48h UBER EE. Wee ER ESE PEER eSee NORE
‘BES SR SEERERee Reese 4eee H+ aH +4 SS eee SERR ion VE! —-f {Ty}
THEUREGRGRGRERUREGY 40 .we JORSEES Vea y A268!
SRSSSERSS See 2B ae LAR, Se co NT Bene pry Pr SeeES0007 GEBBRE
sass See ee as ss

GEEGE BES RE ORS Pe Bee eRe See Pee RSS ae Sh OEE Al PARRReeScAPeoToaas

lh RSe? AOR BEES ROSE See Pe ee Pe ee SER eR El PRs ee

PAT eee

Se Way 2 SSRs Ena! }
I 9SgRseaapans ape PERS CRAs Ses

| BSE GER ERS eee ee See eee eeUeea ae

8 6 ee ee Be eee Pee ee Pees Ue) eh See
rj EaBSaasaeEsaSssasvae aasraazta Gs aerTaeeaerTeeaes eet tersees TosTeereee reeeeeras tT ceriatent
fy is REBSeeeeo8 HEEEEEE EE EEE EE eee Nee

RAR ete
Vit y AN ff a

Fic. 18.—Synthetic prediction, 50 years in advance of mean basis, and verification
on Washington temperature. Computed from temperature records 1854 to 1939 with
20 regular ‘periodicities, all integral submultiples of 223 years. Correlation coefficient
50.42£0.7 percent. Forecast, lighter curve, right-hand scale. Event, heavy curve, left-
hand scale. Temperatures Fahr., 5-month running means.

over the interval of go years, 1850 to 1939. It carries several integrally
related shorter periods on its back. The curves a and c represent the
years 1850 to 1899, and 1900 to 1939, respectively. Being similar, and
in the same phase, their average, b, is used in what follows. With-
drawing the average period of 454+3 months, curve d results.
Withdrawing from it the average period 4543~+4 months, curve e re-
sults. Withdrawing from it the average period 45$~+5 months, curve
f results. Withdrawing from it the average period 453~+2 months,
curve g results. The smooth heavy curve is the 454-month period
freed from all encumbrances. It has the amplitude 7 percent of normal
precipitation at Albany.

Figures 20 and 21, relating to the periodicity of 11 months in
Albany precipitation, will be understood from the description just
given of figures 16 and 17. The heavy mean generalized curves, for
sunspots 2 20 Wolf numbers, are similar in form and amplitude, but

tea

ee
Sh
a
eae

cleared of over

Albany precipitation,
integral submultiples thereof.

Fic. 19.—The 454-month periodicity in
riding periodicities, 1

32 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

Figs. 20 (left) and 21 (right)—Fig. 20, combination of six separate determinations of
the 113-month periodicity into one general mean, for times when Wolf sunspot num-
bers exceed 20. Fig. 21, same as figure 20 for Wolf sunspot numbers less than 20. Full
curves are originals, dotted curves with phases shifted as per arrows.

NO. 4 SOLAR VARIATION, WEATHER ELEMENT—ABBOT 33

differ in phase. Their amplitude is about 9 percent of normal precipi-
tation at Albany.

Figure 22 shows predictions of precipitation at Albany for the years
1928 to 1931. The event is the heavy line. The dotted line is a predic-
tion made wholly by synthesis from the forms and amplitudes of 22
regular periodicities determined from records of 1850 to 1899, center-
ing about 1875. The correlation coefficient between this prediction and
event is 44.0+9.5 percent. The light full line is synthesized from all
records of 1850 to 1939, centering about 1900. I should add that in
this precipitation work the monthly records are smoothed by 5-month
running means. These forecasts may be claimed to be 55 and 30 years
in advance, respectively, counting from the central years of their bases.
I also computed the correlation coefficient for the light full line, repre-
senting synthesis of averages of 22 periodicities, 1850 to 1939. It is
75.6+6.9 percent. If it be urged that this is not evidential, because
1930 lies within the go-year basis 1850 to 1939, I reply that only 41
months, January 1928 to May 1931, can be of direct influence, but
1,039 other months really control the prediction.

SUMMARY

I have sought to support, as a reasonable working hypothesis, the
union of five propositions:

1. The sun’s output of general radiation is variable.

2. Solar variation has a master period of about 223 years.

3. Solar variation has numerous subordinate regular periodicities,
all integrally related to 22? years.

4. Solar variation affects weather importantly, irrespective of
periodicities.

5. Weather responds importantly to most of the regular periodic
solar variations. This is a new, powerful element in meteor-

ology.
Each of these five conclusions is supported by correlations with
several other classes of phenomena, as follows:

Conclusion 1 :
a. Areas of solar faculae.
Prevalence of sunspots.
Areas of solar flocculi.
Incidence of great magnetic storms.
Tonospheric changes.

SERS eds

“sia
oF WOIJ stsayyUAs = dAInd poop AAvay ‘sivak 06 WoOIZ stsayjUAS = AND [INF ST * + poaresqo = dAIND [NJ AAvazy “AjAATIDodsat “66Q1 0} OSgI ‘savaX

ob uo pue ‘6£61 0} OSgI ‘sivad 06 UO paseq Sat}oIporiad jo sasayyUAs 0} posedusos ‘1€61 0} gzOI ‘Aueq ry }e pedstasqo uorepdIV1g—ze “1

‘AON ip fo oF we fp aD aah - Pi 9261,

tae rs sVeaial se 2)
7 TE teas Pe ee
PrN ea alee
ee AY WT RSA PNT Tt
TLE 4 | AT

NO. 4 SOLAR VARIATION, WEATHER ELEMENT—ABBOT 35

Conclusion 2:
a. Solar-constant measures approximately repeated in form of
march of variation after about 223 years.
This period found in sunspot frequency.
Also in magnetic condition of sunspots.
Also in thickness of tree rings.
e. Also in terrestrial temperatures.
f,g-Also in terrestrial precipitation.

Conclusion 3:

a. Over 20 regular periods, submultiples of 223 years, found in
solar-constant measures.

b. The longer of these regular subperiods carry submultiple regu-
lar periods upon themselves.

c. Many of these submultiple periods are found in ionospheric
changes.

d. Analogy with sound harmonics leads us to expect many other
integral subperiods, after three of them were independently
discovered.

Conclusion 4:
a. West Indian hurricanes, a trigger effect of depressed solar

constants.

b. Very numerous temperature changes correlated to solar varia-
tions.

c. Numerous precipitation features repeated at 233-year inter-
vals.

Conclusion 5:

a. Nearly all subperiodicities found in solar-constant measures
are found strongly represented in temperature and precipi-
tation.

b. Syntheses of temperature and precipitation periodicities yield
approximate march of observed weather.

c. Forecasts 50 or more years in advance of mean years of bases,
from such syntheses, yield tolerable accord with observed
weather, with correlation coefficients from 5 to II times
their probable errors.

i eueas a in fh
dae ately

SMITHSONIAN MISCELLANEOUS COLLECTIONS
VOLUME 122, NUMBER 5

SILVER DISK
PYRMELIOMETRY.

(WitTH 1 Pate)

W. H. HOOVER
AND

A. G. FROILAND

Astrophysical Observatory
Smithsonian Institution

ows

Up ,
THSONO*. 4:
ny Cor 2°

S2eceeeee®

(PusLicaTion 4136)

CITY OF WASHINGTON
PUBLISHED BY THE SMITHSONIAN INSTITUTION
AUGUST 4, 1953

The Lord Baltimore Dress

BALTIMORE, MD., U. & A.

SELVER-DISK PYRBELIOMETRY
By W. H. HOOVER anv A. G. FROILAND

Astrophysical Observatory, Smithsonian Institution

(WitH 1 PLATE)

In June and July 1932, Dr. C. G. Abbot and L. B. Aldrich com-
pared silver-disk pyrheliometer S.I. 5yis with an improved form of
the water-flow pyrheliometer.*

The mean of 37 comparisons indicated the constant of S.I. 5pis
should be 0.3625. The original constant of S.I. 5pis (0.3715) was
determined by Dr. Abbot and W. H. Hoover in August 1931 by 24
comparisons with A.P.O. 8);;. Eight more comparisons in September
1932 by Aldrich and Hoover indicated the constant of S.I. 5nis should
be 0.3718. Silver-disk pyrheliometer A.P.O. 8pis has been used since
1912 solely for standardization at Washington. Thus the scale of the
Smithsonian revised scale of 1913 is too high by the ratio 0.3718 to
0.3625, or 1.0256—about 2.5 percent.

The results of 42 more comparisons in July 1934 by the same ob-
servers were in close agreement with the results of 1932. The mean
value of the constant of S.I. 5pis in 1932 was 0.3625, and 0.3629 in
1934.”

No comparisons were made between 1934 and 1947. In August
1947, 18 comparisons gave 0.3626 as the constant of S.I. 5nis.4 The
results of the comparisons between silver-disk pyrheliometer S.I.
Spis and the standard water-flow pyrheliometer No. 5 in 1932, 1934,
and 1947 are based on Dr. Abbot’s habit of reading the silver-disk
pyrheliometer. L. B. Aldrich made a few observations with S.I. 5n:;
in 1932.

Since there is a small personal equation in reading the silver-disk
pyrheliometer more comparisons were made in 1952 between S.I.
5pis and the standard water-flow instrument No. 5. The results of

1 Smithsonian Misc. Coll., vol. 87, No. 15, 1932.

2 Smithsonian Misc. Coll., vol. 92, No. 13, 1934.

8 Smithsonian Misc. Coll., vol. 110, No. 5, 1948.

SMITHSONIAN MISCELLANEOUS COLLECTIONS, VOL. 122, NO. 5

2 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

1952 are based on A. G. Froiland’s habit of reading the silver-disk
pyrheliometer. In order to insure comparable results with the water-
flow pyrheliometer no changes were made in the instrument.

Figure 1, A, B, and C, shows the instrument in some detail. The
two chambers a b c and a’ b’ c’ are almost exactly the same in all de-
tails. The distilled water enters at d and divides into two streams at

: TT LOLELLLE

Fic. 1.—Standard water-flow pyrheliometer.

e and e’. The water flows around the receiver at d and d’ and out of
the instrument at 7 and j’. 7 and 7’ are the thermoelectric junctions
used to determine the equality of temperature of the water streams
outflowing from the two chambers. m is the water bath for the two
receivers, water entering at p and being discharged at 0. is a wooden
case surrounding the instrument. The heating coils are indicated by
k and k’. Not shown is a shutter for alternating the chambers ex-
posed to solar and electric heating. A detailed description of one of
the receivers is given in volume 3 of the Annals of the Astrophysical
Observatory.

In order to keep the water bath surrounding the two receivers and
the distilled water entering the instrument at the same temperature,
we used a 50-gallon drum of water as a source of water for the water

NO. 5 SILVER-DISK PYRHELIOMETRY—HOOVER AND FROILAND 3

bath. A circulating pump continually stirred the water in the drum
and a bypass on the pump circulated some of the water through the
water bath. The distilled water flowed through a coil in the drum
before entering the instrument. Thus the bath water and the distilled
water were always at the same temperature when leaving the drum.

All the precautions that were taken in 1932, 1934, and 1937 to in-
sure greater accuracy were again taken in the present comparisons.
In addition, we found it very important to have the rate of flow of
water in the two receivers as near the same as possible. The water
entering the receivers may not be at exactly the same temperature as
the water bath around the receivers, thus any change in the tempera-
ture difference would produce a drift of the galvanometer.

In order to get the flow of water the same in the two circuits, we
exposed both receivers to solar radiation and adjusted the flow of
water until the galvanometer remained at the open circuit zero. Thus
the two streams of water were at the same temperature and since they
were both receiving the same amount of heat the rate of flow should
be the same.

Water currents of approximately 50 cubic centimeters per minute
in each branch of the pyrheliometer were found to give good results,
but rates as low as 35 and as high as 65 cubic centimeters per minute
were used without affecting the results. Temperature of the water
bath varied from 23° to 28° C. on different days.

Table 1 gives the results of the comparisons. The mean of 100
observations gives 0.3622 as the constant of S.I. 5yi;. The average
deviation from the mean is 0.27 percent and the maximum devia-
tion from the mean about 0.9 percent. The above value is about
.13 percent lower than the mean of the previous values.

TABLE I1.—Summary of 1952 comparisons

Calories Corrected Constant
by reading of of Deviation
Date water-flow silver-disk silver-disk from
1952 Time No. 5 S.-I. Neo 551. SL. Nowsir. mean
Sept. 28 8: 38 1.368 3-775 3624 + 2
44 1.376 3.7900 3630 + 8
50 1.380 3.700 3636 + 14
56 1.388 3.847 3609 — 13
9:05 1.400 3.877 3011 — II
II 1.411 3.889 .3627 + £5
y/ 1.413 3.895 3628 + 6
23 1.412 3.906 3016 — 6
29 1.418 3.931 3608 —I4
35 1.428 3.951 3014 — 8

(continued)

4 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

TABLE 1.—Summary of 1952 comparisons (continued)

Calories Corrected Constant

y reading of of Deviation
Date water-flow silver-disk silver-disk from
1952 Time 0.5 I. No. 5i:5 alONOsS se mean
Sept. 29 Q:22 1.405 3.861 3638 + 16
28 1.410 3.879 3638 + 16
34 1.407 3.894 .3613 =e
40 1.408 3.809 3012 — 10
46 1.415 3.918 3011 —IlI
52 1.419 3.937 -3604 — 18
58 1.425 3.947 .3609 a
10: 04 1.434 3.944 .3635 + 13
10 1.445 3.975 .3635 is
II: 14 1.447 3.986 3630 + 8
20 1.454 4.018 3620 — 2
26 1.462 4.015 .3041 + 19
32 1.461 4.011 3042 + 20
38 1.458 4.022 3624 + 2
44 1.466 4.027 3640 + 18
50 1.454 3.996 .3639 + 17
56 1.464 4.025 .3637 +15
12:02 1.454 4.005 -3629 + 7
08 1.464 4.013 -3649 + 27
14 1.457 4.004 -3639 Sa
Sept. 30 0:44 1.423 3.922 3628 + 6
50 1.421 3.930 36017 — 5
56 1.430 3.961 3010 —I2
10: 02 1.431 3.949 3624 + 2
08 1.426 3.045 3013 — 9
14 1.430 3.959 3611 —II
20 1.433 3-990 .3501 — 31
26 1.446 3.988 3026 + 4
32 1.438 3.998 .3596 —26
38 1.454 4.013 3023 + I
Oct. I TDs 0r 1.446 3.990 3625 + 3
17 1.453 4.027 .3607 — 15
23 1.458 4.029 .3620 — 2
29 1.458 4.020 3628 + 6
35 1.458 4.029 .3619 —
41 1.457 4.008 -3635 + 13
47 1.454 4.018 .3617 = 5
53 1.469 4.058 3621 — I
59 1.472 4.058 .3627 + 5
12:05 1.472 4.067 3621 — I
13 1.463 4.054 3610 —iI12
19 1.458 4.023 3623 + I
25 1.461 4.048 3009 — 13
31 1.465 4.037 3628 + 6
37 1.447 3.994 .3623 ae
43 1.451 4.007 3021 — I

(continued )

NO. 5

Och 11

SILVER-DISK PYRHELIOMETRY—HOOVER AND FROILAND

TABLE 1.—Summary of 1952 comparisons (concluded)

10:

12:

10:

Mean of 100 observations (6 days)
Average deviation

Calories
b

by
water-flow

No. 5
1.445
1.427
1.418
1.427
1.447
1.450
1.463
1.465
1.478
1.479
1.494
1.492
1.406
1.506
1.516
1.526
1.530
1.534
1.543
1.550
1.548
1.549
1.551
1.548
1.554
1.557
1.558
1.558
1.558
1.548
1.553
1.546
1.546
1.540
1.530
1.535
1.537
1.541
1.545
1.549
1.548
1.539
1.540
1.543

Corrected
reading of
silver-disk
TE. No. Shi

3.980
3.947
3.878
3.918
3-971
3-995
4.050
4.053
4.065
4.087
4.123
4.119
4.150
4.166
4.196
4.235
4.242
4.245
4.265
4.276
4.276
4.276
4.291
4.291
4.290
4.305
4.311
4.311
4.311
4.291
4.207
4.277
4.296
4.241
4.224
4.226
4.218
4.240
4.250
4.276
4.276
4.241
4.251
4.261

Constant

oO
silver-disk
PDN: Shite

3631
3616
.3657
3042
.36.44
3030
3613
3014
.3635
3019
3023
3023
3604
3016
3012
3604
.3007
3015
36017
3625
.3620
3023
3614
3607
3623
3616
3613
.3614
3613
.3607
3014
36015
3599
3032
3622
3032
-3645
-3635
3636
3023
3019

Deviation

ol
N Om 0 OO DH tNM Or NW U1

6 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL, \i22
ae Mat 3g /is- TT 5 hs SAA LIZ. fer 7~10.iP
PAE ae 10:01/a 0 20% ML & 2

F AP 32 7 Bago ve OTT T rch d
33] 27227 ies 3325 538 as aS 230% 423
vf S950 SPAS ws ¥ [3224 5935, Ge rr
WZ (3163 gu! 33560 3 3ENN4 6.76 Ee
pe ee gs Wee 96.97) ae
34-77?~ 99 SPO) —795— PIC 1S)
339 GALLO 4b/07/20 AenbeFO zs
‘ oa ——— 40 -_—/-
airy tbe syy 43364 c ie See ee
Re Ae ae eee) map ee ee sree oe
4 - SHE o
ye G3e7 it iy wie, #9 3e we Gris 7 a
wage IZ ee Ge 36.357 3 42%
- Po) . .
SAP O-S+1E Oe far ee 2975 —18%
ry F-H1- AO ax a , Z a

ZOD arg @ ~F6 4 BE) us

FG - {_ JF. 3/ LS

15° 36 93 TaD Es? pee $25
37-20 St Eb sir aa(? sl! 4902 —

af ~/9 ay. fi 33/ -Pfo EN 32g 52 igen
328 ( gy.03 ah 22/8 746 ae) ee
32
37379 or 3297 FF H i
wor + 78S td Geos Js tt2 SY
pee os. 66 ee uses “eee | 3
2392 Spee wes WIAHAVIW _-,, eile 13é
a1 ay ape ice WL D&O (2 ey
IC 22% rp aia (23.76 wo O°
333 C449 ah LG sf L9H SES Bese
Se 3007 76 9 2707 £96 3o0% A
S44 ab 35:64 ZOTs s E
SBI ggyo 7 a8 He TF,
FBO Fag Lines ee, Chee 1B2gEie Oo” ,
S677 '2O 2° fe
SBE IV-6D 197 . 4. IS-3/)
4a" ae Wag id © Ee idan) a
a4? Ay see ENTE er
Ol 88 a pee ae 1G aa Wee
s EN ) o/
37 o6 3. his is: Ay 33. Cats a7 : 3667 Wo
3799 ee 3S I2.5°3 DIS IYSS -IS 3 276
J3 3h a1 e gees : 407 nO
° pee ae of? 149 e ~/-2i0
ss -0F EES Ew) bee eae NC ee >

— 33/7 ” : 330s or

Wb”, Tag) Viens = Jeera "ee ED Pa: d 262,

ac fi we. ugase we G Y 40 ¢

c 260 39, 6 Js 170 3ot/ aE oo

2 GPE a xe aealre Co eee 357 5/ Ee
Sos 3 fee = ae oe ipl! 3Y37_ NSH ,,
-3/ i pee Luger
. FIF/LO ——
ae TIA O
i IZ 3S- aoe = —ttg
qe F22€ 12G ‘e 2V-60 ees
Ly (22 7 =} = bad Deve iz,
w ‘ f 34.09 1)
jaye 2 aay 3 Gi20 Ft
SSIS 42 26.47 pubes
22.95 -493 as 2
ea

Fic. 2—Photographic copy of Froiland’s original reading of the

silver-disk pyrheliometer.

SMITHSONIAN MISCELLANEOUS COLLECTIONS VOLE. 1227 (NO! 5, PL. 2

STANDARD WATER-FLOW PYRHELIOMETER AS MOUNTED AT
TABLE MOUNTAIN, CALIF.

NO. 5 SILVER-DISK PYRHELIOMETRY—-HOOVER AND FROILAND 7

Figure 2 is a photographic copy of A. G. Froiland’s original read-
ings of the silver-disk pyrheliometer S.I. 5nis on September 29, 1952.

A summary of all the comparisons between S.I. 5pis and standard
water-flow No. 5 is given below:

No. of values Date Mean constant S.I. 5),
37 1932 .3625
42 1934 .3629
18 1947 3626
100 1952 3022

The variations in the above results are probably within the limit
of error of the observations; thus we may assume the constant of
S.I. 5nis has remained constant since 1932.

The constant of S.I. 5yis, as detemined by 32 comparisons with the
standard silver-disk pyrheliometer A.P.O. 8pis in 1931 and 1932, was
0.3718. The mean of 64 comparisons with A.P.O. 8pis just before
S.I. 5nis was carried to Table Mountain, Calif., and 64 comparisons
just after its return to Washington gave exactly the same constant.
The mean of all good comparisons between S.I. 5pis and A.P.O. 8nis
from I93I to 1953 gives 0.3719 as the constant of S.I. 5pis. This
would indicate that A.P.O. 8,:, has remained unchanged since 1932.

The mean of all the above results would indicate that the scale of
Smithsonian revised pyrheliometry of 1913 is very nearly 2.5 percent
too high.

SOME EXPERIMENTS WITH THE SILVER-DISK
PYRHELIOMETER

In the following series of experiments with the silver-disk pyr-
heliometer the source of radiation was a 100-watt microscope lamp.
An enlarged image of the filament was focused on the silver disk by
means of a lens and the voltage on the lamp maintained constant with
a voltage regulator. Silver-disk pyrheliometer S.I. 5pis was used in
most of the tests.

A DETERMINATION OF THE CONSTANT K

A correction is added to the reading of the silver-disk pyrheliometer
which depends upon the mean bulb temperature while exposed to
radiation. The correction is [K(7-30°)] R where T is the mean
bulb temperature, the rise in temperature in 100 seconds plus the
cooling corrections, and K is a constant. The value of K in use is
0.0011. This value was determined experimentally, using two silver-

8 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

disk pyrheliometers.* The present determination was made with one
pyrheliometer.

In this experiment and all the following experiments the lamp was
turned on about an hour before making a series of readings. The
lamp was very constant during a day’s observations and changed very
little from day to day.

In one set of observations the following results were obtained:
The mean of 18 observations at a mean bulb temperature of 34.°21
was 3.388, and a mean of 12 observations at a mean bulb temperature
of 21.66 was 3.435. The above readings represent the rise in tem-
perature of the silver disk with all corrections applied with the ex-
ception of the one depending on the mean bulb temperature. If this
correction is applied, the two values should be the same. Thus 3.388+
[4.21 K (3.388) ] =3.435+[—8.34 K (3.435)], or K is equal to
0.001095. Other determinations of K with various temperature differ-
ences gave results of K between 0.00104 and 0.00118, with a mean
value approximately 0.0011. One set of observations made with pyr-
heliometer S.I. 89 gave 0.00109. Some of the determinations of K
were made with the pyrheliometer in a water-cooled chamber with a
hole in one end to admit the radiation and a slot along one side to read
the thermometer. Also an automatic shutter opening and closing device
was used for some of the work. In any set of observations individual
values were within +0.3 percent of the mean.

SERIES OF OBSERVATIONS IN WHICH THE COOLING CORRECTION DURING
THE FIRST IOO SECONDS IS ZERO OR NEAR ZERO

In making a series of readings it has been our practice to start read-
ings 20 seconds after completing an observation. Thus a 4-minute
shaded period occurs between each 2 minutes of exposure. Each value
is independent and the total rise in temperature of the silver disk is
much less than the rise in temperature of the silver disk with only a
2-minute shaded period between each 2-minute exposure.

In the following experiment sets of six readings each were taken
and for each set the cooling correction for the first 100 seconds of
the first reading was zero or near zero. Sets of readings were made
at different temperatures and some were made with the automatic
shutter opening and closing device. After the fifth reading the mean

4 Smithsonian Misc. Coll., vol. 95, No. 23, 1937.

NO. 5 SILVER-DISK PYRHELIOMETRY—HOOVER AND FROILAND 9

bulb temperature and the cooling correction remained about constant.
The mean of 30 sets is given below:

INoMotereadinges. vasa sleet. I 2 3 4 5 6
Corrected reading ..... 3.202 3.206 3.208 3.210 3.211 3.212

The above indicates some increase in the reading from 1 to 6. The
change is small, however, after the second reading. This fact was
noted about 20 years ago, when many comparisons were being made
between the silver-disk and the Angstrém pyrheliometers. Since that
time, when using the silver-disk pyrheliometer, we have preheated
the silver disk from one to three minutes before starting a series of
readings. Series of readings taken after the silver disk was preheated
gave very consistent readings. Some of the discrepancy in the read-
ings shown above may be due to a delay in opening and closing the
shutter or a time lag in reading the thermometer. The rate of move-
ment of the mercury thread is different in the first two or three read-
ings of a set from that in later readings of the set when the rate
of heating and cooling remain about constant.

EFFECT OF DELAY IN OPENING AND CLOSING SHUTTER

A set of readings were made using the regular method of opening
and closing the shutter and then a set in which there was a delay of
Io seconds in opening the shutter after the end of a shaded period
and a delay of 10 seconds in closing the shutter after an exposure
period. The results in the latter case were about 3 percent higher.
Thus a delay in opening and closing the shutter of even one second
may result in an error of 0.3 percent. Variation in the time of open-
ing and closing the shutter may explain the variation of the results with
the silver-disk pyrheliometer by different observers. The pyrheliome-
ter readings in this series of tests were made by L. B. Aldrich and
W. H. Hoover. With the regular method of reading the pyrheliome-
ter and the shutter operated by hand there was a difference of about
0.2 percent between the readings of the two observers. When the
automatic shutter-opening device was used this difference was re-
duced to 0.1 percent or less.

EFFECT ON THE PYRHELIOMETER READINGS WHEN READINGS ARE
TAKEN BEFORE THE PYRHELIOMETER CHANGES TO
AMBIENT TEMPERATURE

For this test the pyrheliometer was placed in the chamber of the
water bath and the automatic shutter device was used when readings

Io SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL, I22

were made. With the water bath at 15° and the pyrheliometer at the
same temperature, a few readings were made with the bath maintained
at 15°. The temperature of the bath water was then changed to about
35° and readings made while the pyrheliometer was changing tem-
perature. The first three readings were from 0.5 to 0.7 percent too
high. When the pyrheliometer was about 5° below the temperature
of the bath more readings were taken. These readings agreed with
the original set of readings. These results indicate that the pyr-
heliometer should be near the ambient temperature before making an
observation.

In a paper on the Abbot silver-disk pyrheliometer ° L. B. Aldrich
discussed in some detail the method of observing with the silver-disk
pyrheliometer and listed some precautions to be taken to insure greater
accuracy. One precaution should be added. The shutter should be
opened immediately at the end of the first shaded period and closed
immediately after the end of the exposed period. A delay of a few
seconds may result in an error of I percent or more.

5 Smithsonian Misc. Coll., vol. 111, No. 14, 1940.

"SMITHSONIAN MISCELLANEOUS COLLECTIONS
pVOEUME 122, NUMBER 6

|| THE EXTERNAL MORPHOLOGY OF THE

DRAGONFLY ONYCHOGOMPHUS
~ARDENS NEEDHAM

BY
HSIU-FU CHAO

Department of Entomology
University of Massachusetts

i abet Ty

(Pusication 4137)

, CITY OF WASHINGTON
PUBLISHED BY THE SMITHSONIAN INSTITUTION
SEPTEMBER 15, 1953

SMITHSONIAN MISCELLANEOUS COLLECTIONS
VOLUME 122, NUMBER 6

fae EXTERNAL MORPHOLOGY OF THE
DRAGONFLY ONYCHOGOMPHUS
ARDENS NEEDHAM

BY
HSIU-FU CHAO

Department of Entomology
University of Massachusetts

(PuBLIcATION 4137)

CITY OF WASHINGTON
PUBLISHED BY THE SMITHSONIAN INSTITUTION
SEPTEMBER 15, 1953

TBe Lord Baltimore Press

BALTIMORE, MD., U. 8 A.

the EXTERNAL MORPHOLOGY OF THE
DRAGONFLY ONYCHOGOMPHUS
ARDENS NEEDHAM *

By HSIU-FU CHAO 2
Department of Entomology, University of Massachusetts

In 1917 Tillyard brought together all the scattered information re-
garding the morphology as well as other biological studies of dragon-
flies in a book entitled “The Biology of Dragonflies.’ His discussion of
morphology in this book was based on the writings of previous
workers who were mainly interested in comparative studies of certain
organs. Since this date some morphological characters have been fur-
ther and well investigated, but other structures remain inadequately
studied. Furthermore, entomologists working on dragonflies have paid
little attention to the new interpretations given by Ferris and Penne-
baker (1939), Ferris (1940), Snodgrass (1947), and others on the
fundamental structures of certain parts of the body of insects in gen-
eral. There is not a single species of dragonfly that has been studied
critically in the light of the most recent morphological interpretations.

The purpose of the present study is fourfold: (1) To bring into
unity all the different terminologies that have been used in morpho-
logical and taxonomic work on dragonflies; (2) to apply the knowl-
edge of the most recent morphological interpretations; (3) to bring
out some new interpretations of morphological characters that the
author believes to be inadequately or erroneously treated previously ;
and (4) to serve as a contribution to the morphology of dragonflies,
especially as a foundation for future taxonomic studies.

Onychogomphus ardens Needham (Gomphidae) has been selected
for study for three reasons: (1) It belongs to the primitive family
Gomphidae of the order Odonata. This family is well represented
by genera and species in my own collection, which will be used for

1 Contribution from the Department of Entomology, University of Massa-
chusetts, Amherst, Mass.

2 The author wishes to express his appreciation for the help and advice re-
ceived from staff members at the University of Massachusetts. Sincere thanks
and appreciation are likewise extended to coworkers throughout the world who
have encouraged and helped the author during the progress of this work.

SMITHSONIAN MISCELLANEOUS COLLECTIONS, VOL. 122, NO. 6

2 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

future taxonomic studies. (2) Specimens of this species are of large
size and therefore relatively easy to study. (3) It is rather common
in South China.

HEAD
(Figures 1-6)

The head of Onychogomphus ardens Needham is hypognathous and
somewhat anteroposteriorly flattened, the anterior aspect being convex
and the posterior aspect being concave. Posteriorly it is attached to
a narrow neck. The female differs from the male in having a pair of
occipital horns (fig. 2: OCCH) on the occipital margin.

The areas generally referred to as frons, vertex, and occiput by
earlier workers are designated as such in this paper. No attempt has
been made to change their names, although they have been interpreted
differently by modern morphologists (DuPorte, 1946; Snodgrass,
1947). The old designations are used here without modification to
avoid further confusion in taxonomic work.

SUTURES OF THE CRANIUM OR CAPSULE

The principal sutures of the cranium are postocellar, epistomal,
subgenal, ocular, parafrontal, postoccipital, and clypeal sutures.

The POSTOCELLAR SUTURE (POS) is a transverse suture which
separates the vertex from the occiput. It lies between the two com-
pound eyes at their closest points. According to Lew (1933) this
suture is secondarily developed and is not homologous with the epi-
cranial suture (e.g., Garman, 1927) of other insects. This suture is
designated by Lew as postocular suture ; but unfortunately in the same
paper he created another term, postocellar suture, evidently referring
to the same structure. The latter term is probably the one he meant
to use, while the former term might be a typographical error, since it
is definitely not descriptive of its position. Snodgrass (1947), how-
ever, mentioned “The cleavage line on the head of larval Odonata is
characteristically T-shaped rather than Y-shaped, inasmuch as the
frontal arms usually go almost straight laterally.” In the present
species the postocellar suture represents the transverse bar of the T
and therefore corresponds to the frontal arms.

The EPISTOMAL SUTURE (ESS), or frontoclypeal suture, is a distinct
and almost straight line across the anterior part of the cranium. Along
this suture a strong epistomal ridge (ESR) is produced internally.

Each of the SUBGENAL SUTURES consists of two portions, the pleu-
rostomal and hypostomal sutures. The pleurostomal suture (PMS)

LATERAL VIEW

TENTORIUM,

5) TENTORIUM, VENTRAL VEW LATERAL VIEW
Fics. 1-6.—Head.

4 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. I22

marks the lateroventral margin of the capsule above the mandibular
bases and between the anterior (AAR) and posterior articulations
(CAR) of each. Along this suture the anterior tentorial arm (ATA)
is produced internally, and to it a narrow sclerite, the pleurostoma
(PM), is attached. The poorly defined hypostomal sutures (HMS)
each follows closely the posterior margin of the cranium between the
posterior articulation of the mandible (CAR) and the posterior ten-
torial pit (PTP).

The oCULAR SUTURES (OS) surround the bases of each of the com-
pound eyes except in the anterior aspect of the head where a narrow
band of sclerite, the ocular sclerite (OCS), is interposed between
them.

PARAFRONTAL SUTURES (PFS) are present, one on each side of
the frons. Apparently no name has previously been given to them
although they are shown in many drawings of various species of
dragonflies by different authors (e.g., Tillyard, 1926; Lew, 1933).
These are probably the frontogenal sulcus of DuPorte (1946) or
lateral grooves of Snodgrass (1947). Each suture extends from the
middle of the inner margin of the eye near the antenna to the anterior
articulation of the mandible, thus separating frons from gena. Each
is hidden by the lateral portion of the elevated frons and clypeus and
therefore is invisible in the anterior aspect of the head. Along this
suture a low ridge is produced internally.

The PosToccIPITAL SUTURE (POCS) closely parallels the dorsal
and lateral margins of the foramen magnum (FM).

The CLYPEAL SUTURE is absent, but its position is indicated by a
distinct line of demarcation between the sclerotized postclypeus and
the mostly membranous anteclypeus.

The antennal socket is well defined but is not circumscribed by an
antennal suture.

AREAS OF THE CRANIUM

The principal areas of the cranium are clypeus, frons, vertex,
occiput, genae, postgenae, postocciput, pleurostomae, hypostomae, and
eyes. The gula is absent in this species. However, it has been very
vaguely indicated as being present in the order Odonata by earlier
workers. Calvert (1893) said that the gula was membranous; Till-
yard (1917) probably concurred with him in this matter. Marshall
(1914) probably wrongly designated submentum as gula.

The cLYPEus is a large transverse sclerite differentiated into a light-
colored, mostly membranous anteclypeus and black-colored, sclerotized
postclypeus. The latter areas are separated by deep indentations on

No. 6 MORPHOLOGY OF THE DRAGONFLY—CHAO 5

both sides. The anteclypeus (ACL) is light-colored, laterally produced
into a lobelike structure, with a narrow sclerotized piece on each side
extending mesally from the tip of the lobe a distance of one-third the
width of the anteclypeus. The postclypeus (PCL) is black, with or
without a small transverse light-colored spot on each side. It expands
laterally and extends ventrally, thus overlapping a portion of the
lateral lobe of the anteclypeus. A small condyle, the anterior articula-
tion of mandible (AAR), is produced on each side near the base of
the postclypeus. It is covered by the laterally expanded portion of
the postclypeus so that it cannot be seen in the anterior aspect of the
head.

The Frons (FR) is a large, transverse, convex area which is
bounded ventrally by the epistomal suture, laterally by the parafrontal
sutures, and dorsally by a transverse furrow between it and the vertex.
It is differentiated by a sharp fold into an upper horizontal portion
and an anterior or vertical portion, but there is no sutural demarcation
between these regions. The upper portion is called the top of frons
(TFR) which is differentiated into two low prominences separated by
a broad median furrow. A broad, transverse, light-colored stripe
covers most of the top as well as a part of the anterior portion of the
frons. This stripe is sometimes separated in the middle along the
median furrow. In this light-colored area there are a few small black
tubercles each of which bears a minute hair.

The vERTEX (V) is a trapezoidal area bounded ventrally by a groove
between it and the frons, dorsally by the postocellar suture, and
laterally by the ocular sutures. It bears a pair of antennae and three
ocelli (OC), the latter being very large. The deeply sunken middle
ocellus is a little lower in position than the lateral ocelli. Along the
dorsal rim of the middle ocellus there is a very low but large knoblike
tubercle which bears a group of fine and wavy long hairs. External
to the tubercle and the lateral ocellus there is a subsemicircular ridge.
The dorsal tentorial pits (DTP) are present as a pair of semicircular
sutures above and lateral to the bases of the antennae (mostly obscured
by the antennae in anterior view and by the eye in lateral view). It is
interesting to point out here that the dorsal tentorial pits are present
in the adult dragonflies (Lew, 1933) whereas they are usually repre-
sented by a pair of callosities in the dragonfly nymphs and many in-
sects. Two peculiar papillae (PA), about two-thirds as long as the
third antennal segment, situated one on each side very close to the
external rim of the antennal socket, are present in both sexes. They
are small and usually obscured from view by the antennae and there-
fore are easily overlooked. Apparently they have not been reported

6 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

heretofore. They occur also in Onychogomphus micans Needham
which is very closely related to the present species, but not in Jctino-
gomphus rapax (Rambur) (Gomphidae) and Anax nigrofasciatus
Oguma (Aeschnidae) which I have examined.

The furrow which separates the frons from the vertex is called
frontal furrow or frontal suture by Tillyard (1917). These terms are
confusing since the structures they define are definitely not homologous
with the frontal sutures of other insects. Morphologically speaking
(DuPorte, 1946; Snodgrass, 1947), the areas of the frons, vertex,
and occiput described here are not homologous with those of other
insects or even with these areas as designated in certain other species
of dragonflies, although they are generally so considered by students
of Odonata.

The occiput (OCC) is situated on the top of the head between the
compound eyes. In the anterior aspect of the head it appears as a
transverse area bounded ventrally by the postocellar suture, laterally
by the compound eyes, and dorsally by the occipital border, or occipi-
tal margin (OCCM), which is almost twice as wide as the postocellar
suture, and fringed with long black hairs. In the female there is a
pair of occipital horns (OCCH) on the occipital margin. These are
not to be confused with a pair of similar horns which arise on the
vertex above the lateral ocelli and which are also, but erroneously,
called the occipital horns. Such horns occur in a number of species
of the family Gomphidae, e.g., Gomphus flavicornis Needham (Lew,
1933, pl. 8, fig. 9), Gomphus cuneatus Needham, and Davidius bi-
cornutus Selys. The posterior aspect of the occipital region is called
the rear of the occiput (ROCC). It is a subquadrate area situated
above the foramen magnum, with a large, light, yellow-colored mark-
ing in the center. Laterally it is demarked with weakly defined fur-
rows or wrinkles which indicate the dividing line between it and the
postgenae.

The GENAE (G) are small sclerites. Dorsally each gena is bounded
by the parafrontal and the ocular sutures, and ventrally to its evagi-
nated margin is attached a small transverse sclerite, the plewrostoma
(PM).

The postGENAE (PG) area pair of large sclerites, one on each side
of the posterior aspect of the head. The outer margin of each postgena
which borders the eye is notched at about the center. Mesally and
mesoventrally the postgena is bounded by the postoccipital suture and
hypostomal suture respectively. Ventrally it is fused with the gena.
This fused portion bears the posterior articulation of the mandible.

The postocciput (POOC) is a roughly horseshoe-shaped narrow

No. 6 MORPHOLOGY OF THE DRAGONFLY—CHAO Ff

sclerite surrounding the dorsal and lateral sides of the foramen mag-
num, with its ends terminating at the posterior tentorial pits (PTP)
where a pair of small transverse processes, the occipital condyles
(OCCD), are produced toward each other. On each side of the fora-
men magnum the postocciput is produced mesally into a short process.

The PLEUROSTOMAE (PM) are very small transverse sclerites, one
on each side, situated in the evaginated ventral margin of the gena
between the anterior and the posterior mandibular articulations.

The HyposToMAE (HM) are narrow bands or thickenings, one on
each side bordering the lower margin of the postgena between the
posterior articulation of the mandible and the posterior tentorial pit.

The compouNp EYES (EYE) are large, and are closest together
along the postocellar suture. Each is evaginated in the middle on its
posterior margin as shown in figures 4 and 5.

TENTORIUM
(Figures 4-6)

The tentorium consists of a corporotentorium and three pairs of
tentorial arms, namely, dorsal, anterior, and posterior.

The CORPOROTENTORIUM (CT), or tentorial body, is a transverse
bar very close to the posterior surface of the head capsule and appear-
ing as the floor of the foramen magnum. Apparently it is often mis-
taken for the gula by some students of Odonata.

The POSTERIOR TENTORIAL ARMS (PTA) arising from distinct
posterior tentorial pits (PTP) are very short and are not differentiated
from the corporotentorium.

The ANTERIOR TENTORIAL ARMS (ATA) arise from extremely elon-
gated tentorial pits lying along the entire lengths of the pleurostomal
sutures. Each anterior tentorial arm is a fanlike structure strength-
ened by three heavily sclerotized ribs radiating from the corporoten-
torium. The posterior rib (PRB) extends to the posterior mandibular
articulation, the middle rib (MRB) to the anterior mandibular articu-
lation, and the anterior rib (ARB) to the lateral end of the epistomal
ridge. On the ventral surface of the middle rib there are two proc-
esses: the anterior process is called the mandibular process (MDP)
and consists of a very large, ovoid, tendonlike structure with a short,
narrow stalk. The large ovoid portion is inserted in the heavy muscles
of the mandible. The posterior process is called the maxillary process
(MXP) and is a long and slender tendon supplying attachment for
the maxillary abductor muscles. On the posterior rib there is also a
short process.

8 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

Each of the DoRSAL TENTORIAL ARMS (DTA) consists of a simple
flattened structure arising from the middle rib of the anterior tentorial
arm. Each narrows slightly in the middle and is apically fused firmly
with the invagination of the dorsal tentorial pit.

HEAD APPENDAGES
(Figures 7-16)

The movable parts of the head are the antennae, labrum, mandibles,
maxillae, hypopharynx, and labium.

The ANTENNAE (ANT) are short, inconspicuous, setaceous, and
4-segmented. The basal segment, or scape (S), is very thick. The
second segment, or pedicel (P), is subequal in length and about half
the diameter of the preceding segment. The third and the last seg-
ments are collectively called the flagellum (FL), or distalia, which is
slender and bristle-like; the third segment being about two-thirds as
long as the pedicel ; the last segment being longer than the other three
segments combined.

The rasrum (LR), or upper lip, is a transverse subovoid sclerite,
movably attached to the anteclypeus and functions as one of the mouth
parts. It is generally regarded as not a true appendage. Great differ-
ences of opinion exist among entomologists as to its homology. For
more detailed accounts the reader is referred to recent papers by
Ferris (1947) and Henry (1948). Aborally (fig. 7) it is slightly
convex, black, with two large ovoid yellow spots, and fringed with
many long hairs along its distal and lateral margins. Adorally (fig. 8)
it has a flat surface, is black on lateral regions, and has a large clear
area called the epipharynx in middle. The epipharynx (EPX) con-
sists of a round, slightly depressed, sclerotized area in the center
surrounded by a group of small circular tubercles and hairs. These
tubercles are probably taste organs. Some hairs are grouped together
to form the brushes (BH) pointing mesad.

The mandibles (MD) are very strong unsegmented appendages
bearing strong teeth which may be divided into two groups: a large
basal mola (MO) and a distal group of three incisors (ICS).

The base of the mandible is triangular in shape, with one lateral and
two mesal angles. The mesal angles are designated as inner and outer.
The mandible is attached to the head capsule by two articulations, the
ginglymoid anterior articulation (AAR) at the outer angle and the
condylic posterior articulation (CAR) at the lateral angle. A strong
flexor tendon (FT) is attached to the inner angle and a weak re-
tractor tendon (RT) to the lateral angle.

No. 6 MORPHOLOGY OF THE DRAGONFLY—CHAO 9

The adoral side (fig. 10) of the mandible has two tuberculate and
hairy areas, one at the base of the incisors and the other parallel to
the margin of the mola, the former area being crescent-shaped. The
aboral side (fig. 9) of the mandible has a similar crescent-shaped,
tuberculate, and hairy area which is slightly depressed and joins with
a ridge extending to the anterior articulation.

The three incisors (ICS) are of unequal length; one of them being
very long, sharply pointed, slightly curved, and bearing the smaller
basal one on its mesal edge and an even still smaller medial one on its
adoral side. The mola (MO) has four cusps set on a broad base in
the shape of a Z on the right mandible and inverted Z (X) on the
left mandible. The cusps are placed one at each end and one at each
angle of the Z.

The MAXILLAE are composed of several parts, namely, cardo, stipes,
inner lobe, and outer lobe.

The cardo is an elongate structure internally strengthened by an
X-shaped ridge, the mesal arms of the X being submarginal and the
lateral arms marginal. It is divided into basicardo (BCD) and disti-
cardo (DCD) by a suture which is situated along the lateral margin
of the basolateral arm and the mesal margin of the distomesal arm.
The concave area of the basicardo between the basal arms is about
two-thirds as large as the weakly sclerotized convex area of the disti-
cardo between the distal arms; the latter area being adorned with a
few long hairs.

The stipes (STI) is a large elongate rectangular structure adorned
laterally with many short hairs on its inflected area and apically with
many long hairs. A mesal submarginal sutural groove (SG) (Snod-
grass, 1935) sets off a narrow area called parastipes (SG) (Cramp-
ton, 1923b, p. 83). An isolated sclerite is present in the membrane
which attaches along the margin of the lateral inflected area. Nothing
is known about this sclerite although it has been shown in drawings by
earlier workers (e.g., Tillyard, 1917, p. 16, fig. 4).

The inner lobe (IL) and the outer lobe (OL) are two freely mov-
able processes, the former being generally regarded as representing
the fused Jacinia and galea of more typical mandibulate insects, and
the latter the palp.

The inner lobe (IL) is a large process basally fused with the stipes
but separated from the parastipes by a narrow strip of membrane.
Basally it is expanded on its mesal portion and adorned with many
long hairs. Apically it is narrowed and gently curved, ending in three
teeth pointing mesad, the apical tooth being very long and the middle
one the smallest. The weakly sclerotized area along the mesal margin

IO SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

of the apical half of the inner lobe bears three widely spaced teeth of
subequal length, also pointing mesad.

HYPOPHARYNX

Fics. 7-12,—Mouth parts. 7, 9, and 11, Aboral views; 8, 10, and 12, adoral views.

The outer lobe (OL) is a stout, slightly curved, fingerlike structure,
about as long as the inner lobe, with the lateral portion of the basal
half weakly sclerotized and unpigmented. Basally it is situated on a
small transverse sclerite on the adoral side of the maxilla.

No. 6 MORPHOLOGY OF THE DRAGONFLY—CHAO II

The HYPOPHARNYX is a large, elongate, wedge-shaped lobe in the
preoral cavity, apparently consisting only of the lingua, with its aboral
surface (fig. 11) about half as long as its adoral surface (fig. 12),
its lateral sides slightly divergent, and its apical margin slightly
emarginate.

Adorally (fig. 12) the hypopharynx is adorned with hairs. Those
hairs along the distal margins and surrounding the subapical, de-
pressed, sclerotized area are very long and widely spaced. Basally
the hypopharynx has a heavily sclerotized transverse bar which is
fused laterally with a pair of slightly raised sclerotic structures im-
mediately distal to it ; the former with a low, transverse internal ridge,
and the latter with a number of small round nodules. A single trans-
verse, raised, somewhat wrinkled, weakly sclerotized structure, and
a pair of similar smaller ones are situated distal and lateral to the
sclerotized structures respectively.

Aborally (fig. 11) the lateral walls of the hypopharynx contain a
pair of basal plates, the apical ends of which are attenuate, whereas
the basal ends expand and extend along the basal margin of the hypo-
pharynx to the salivarium (SAL).

The Lazium consists of the following movable parts: Submentum,
mentum, middle lobe, squames, lateral lobes, and movable lobes.

The submentum (SM) is a quadrangular piece with its basal margin
slightly evaginated and its lateral edges subparallel to each other.
Basally it is bounded by membrane continuous with the neck region.
Laterally and adorally it is connected with the mesal margins of the
cardines and stipites of the maxillae and with the base of the hypo-
pharnyx by a large membrane.

The mentum (MN) is a transverse area. Adorally (fig. 16) it is
partly membranous, with a pair of large transverse subrectangular
sclerites imbedded in the membrane. These sclerites are adorned with
long hairs on their bulging lateral portions. Apically the mentum is
separated from the middle lobe by a distinct membranous fold.
Aborally (fig. 15) it is sclerotized on its basal half and weakly so on
its distal half, with distinct line of demarcation between these regions.
The basal half is fused laterally with the squames. The distal half is
unpigmented, adorned with a few scattered microscopic hairs, and
fused distally with the middle lobe. Distolaterally it is evaginated into
a socketlike structure on each side to which the mesobasal portion of
the lateral lobe is attached.

The middle lobe, or median lobe (ML), is a large subrectangular
piece which, according to Butler (1904), corresponds to the ligula of
other insects. The latter term is not to be used, because a part of

12 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122
the pronotum is also called the median lobe. Adorally (fig. 16) it is
weakly sclerotized and pigmented on its apical third and narrowly

so along its lateral portions, with long hairs on these areas: the re-

MAXILLA

LABIUM = [6

Fics. 13-16.—Mouth parts. 13 and 15, Aboral views; 14 and 16, adoral views.

maining area is unpigmented and adorned with a few scattered micro-
scopic hairs.

The squames (SQ) are a pair of convex sclerites, which, according
to Tillyard (1917), correspond to the palpigers. (The term squames

No. 6 MORPHOLOGY OF THE DRAGONFLY—CHAO 13

is rather confusing since it has been used to designate different struc-
tures in different orders of insects.) Aborally (fig. 15) they are
subrectangular in shape, mesobasally fused with the basal portion of
the mentum, laterally deflected to approach the sclerites of the mentum
on the adoral side.

The lateral lobes (LL) are a pair of hairy sclerotized structures
attached to the squames and mentum, with their mesal margins straight
and their lateral margins strongly convex. Each lobe is produced
mesoapically into a very long, bare, and sharp end hook (EH) point-
ing meso-orally. Lateral to the end hook is an even longer hairy
movable hook (MH) with its basal half about twice as wide as its
apical half.

CERVIX
(Figures 17-20)

The cervix (sometimes called neck or microthorax) is a region
between the head and the prothorax, narrow anteriorly, mainly mem-
branous, with lateral, dorsal, and ventral cervical sclerites. The lateral
cervical sclerites are the largest and serve as pivots for the head while
the other sclerites are mostly small and completely surrounded by
membranes.

Each of the LATERAL CERVICAL SCLERITES consists of a basal post-
cervicale (PC) and a distal eucervicale (EC) (Crampton, 1926)
forming a hinge at their juncture. The postcervicale is V-shaped, fit-
ting between the pronotum and the episternum, with the arms of the
V pointing anteriorly. The eucervicale is incompletely divided into a
dorsal and a ventral portion by a deep and narrow incision. The
anterior half of the dorsal portion of the eucervicale is unpigmented
and whitish. The ventral portion is somewhat twisted, produced mesad
and then anteriorly into a long process called the cephaliger (CEP) ;
the latter lies freely inside the cervical membrane, with its apex con-
nected with the occipital condyle (fig. 3, OCCD).

There are two transverse DORSAL CERVICAL SCLERITES (DC) in the
middle of the cervix, with a pair of small and weakly sclerotized
sclerites between them; the posterior transverse sclerite also being
weakly sclerotized. Another pair of dorsal sclerites is situated on the
posterior margin of the cervix: they are fairly large in size, well
sclerotized, and partly obscured dorsally by the anterior lobe of the
pronotum.

The paired VENTRAL CERVICAL SCLERITES (VC) are roughly L-
shaped, with the transverse bars of the L’s almost touching each other,
and the other ends of the L’s being in contact with the occipital con-

14 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

dyles. There is a small weakly sclerotized area attached at the angle
of the L, with many tiny tubercles each bearing a microscopic hair.

THORAX

The thorax is differentiated into two distinct parts, namely a small
prothorax and a large synthorax (the latter also called pterothorax)
representing the fused mesothorax and metathorax.

PROTHORAX
(Figures 17-20)

The prothorax is a small segment, narrow anteriorly, with its length
subequal to the vertical diameter of its anterior end which is about
two-thirds that of its posterior end. The pronotum (figs. 17, 19)
covers the dorsal half of the segment, topographically differentiated
into anterior, median, and posterior transverse lobes alternated with
two furrows where ridges are produced internally. The anterior lobe
(AB) has its anterior margin whitish. The anterior furrow is deep
and about as broad as the anterior lobe, laterally with a depression on
each side where a long, pointed apodeme is produced internally. The
median lobe (MB) is divided into two parts by a narrow median
sagittal groove. It is minutely tuberculate in the areas on both sides
of the median groove and on its lateral portions and with a similar
but smaller area between them. The lateral tuberculate areas are
adorned with fine, long, and wavy hairs. A large semicircular depres-
sion is situated at the anterior end of the median groove and gives rise
to an internal apodeme which is long, slightly curved, narrow in the
middle, and expanded distally into a discoidal structure. The posterior
lobe (PB) is somewhat like a Cupid’s bow in shape on its dorsal
aspect, dorsally minutely tuberculate, and with long, fine, wavy hairs
all over.

Each of the propLeurA (fig. 19) consists of two approximately
equal-sized sclerites, the episternum and the epimeron, separated by the
pleural suture (PLS,) which is almost perpendicular to the long axis
of the body. Along the pleural suture a low ridge or lateral apodeme
is produced internally. The episternum (ES) is a transverse piece,
narrow in the middle, with its dorsoanterior angle fused with the
sternum, and its posterior portion slightly bulging. It is minutely
tuberculate all over, with the bulging area having larger tubercules
and long wavy hairs. The epimeron (EM,) is a rectangular piece,
slightly higher than wide, tuberculate only on its ventroposterior por-
tions, dorsally produced into a narrow strip along the posterior border

No. 6 MORPHOLOGY OF THE DRAGONFLY—CHAO 15

of the pronotum to approach closely the lateral end of the posterior
lobe.

——

DORSAL VIEW

N\ eww
7

“e

LATERAL VIEW POSTERIOR VIEW

Fics. 17-20.—Prothorax and neck.

The ProsTERNUM (fig. 18) consists of an anterior, large, elongate,
rectangular piece and a posterior pair of small sclerites, the latter
collectively termed the postfurcasternum (PF ST) (Crampton, 1926).
The rectangular piece is divided into two portions, an anterior basi-

16 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

sternum (BS) and the posterior furcasternum (FS), by an anteriorly
arched sternacostal suture (SCS) which ends at the large furcal pits
(FP), or apophyseal pits (Ferris, 1940). Along the sternacostal
suture a ridge is produced internally.

The basisternum is fused on its anterior angles with the episterna.
It is minutely tuberculate, with a broad, shallow, submarginal circum-
scribing depression; the posterior course of the depression being
weakly sclerotized and unpigmented. The area between the posterior
depression and the sternacostal suture is raised, with long hairs along
its anterior margin. The central area of the furcasternum (FS) be-
tween the furcal pits and the sternacostal suture is protuberant and
rather coarsely tuberculate. The furca (F,) consists of a pair of large,
inverted, foot-shaped apodemes widely separated from each other;
each arm with a long narrow tendon at its apex.

SYNTHORAX
(Figures 21-27)

The synthorax, or pterothorax, is composed of the fused mesothorax
and metathorax, ventrally carrying two pairs of legs on its anterior
half, and dorsally two pairs of wings on its posterior half. The pleura
are very large while the terga and the sterna are very small.

TERGA
(Figures 21, 22)

The terga are connected with the pleura only by membranes. They
are not connected with the latter by prealares anteriorly or by post-
notum posteriorly, such as is the case in most other winged insects.
Thus, it would seem that the terga can move up and down without
distortion during flight.

The anteriormost part of the mesotergum is roughly a T-shaped
structure divided into the AcroTrERGITE (ATG) and the pREscUTUM

(PSC,.) by the antecostal suture (ACS) along which a pair of small
phragmata is produced internally. The ends of the transverse bar of
the T are the prealares (PRA) which serve, in the present species, as
pivots for the anterior lobes of the humeral plates (HP), and are
connected posteriorly with the detached plates of the scutum (Snod-
grass, 1935) to be described later.

The scutum (SCT,) is a large, somewhat ovoid, convex structure,
wider posteriorly, with a large, central portion weakly sclerotized and
unpigmented. Anteriorly the lateral portion of the scutum is detached

No. 6 MORPHOLOGY OF THE DRAGONFLY—CHAO 17

into a bilobed plate which is fused with the prealare. Posteriorly the
adanal sclerite, or posterior notal wing process (PWP), is narrowly
separated from the scutum by an incision and is articulated with the
axillary plate. The detached plate of the scutum has been shown by
Snodgrass (1930, fig. 11 A, a; 1935, fig. 123 B, a), but its importance
in the wing mechanics has not been well investigated. It consists of
two lobes, the anterior suralare sclerite, or anterior notal wing process
(AWP), and the posterior adnotal sclerite, separated by the notal
incision or the lateral emargination. Near the mesal margin of the
detached plate there is a groove along which an apodeme (AP) is
produced internally. The latter is a large, elongate structure, apically
expanded into an irregularly elongate plate which is constricted in
the middle. This apodeme is called cap-tendon by earlier workers
(Calvert, 1893; Tillyard, 1917). To this apodeme the principal ele-
vator muscle is attached. Two more small sclerites are present. One
of these is the first axillary (IAX), also called notal ossicle, notale,
or notopterale. It is elongate triangular, situated along the mesal
margin of the humeral plate, and mostly obscured by the latter in
dorsal view. The other small sclerite, distinct from the adnotal sclerite
in this species, is situated between the latter and the anteromesal
margin of the axillary plate. A preliminary study of Anax junius, a
common American species (Aeschnidae), shows a condition in which
the detached plates are not fused with the prealares and the axillary
sclerites are not independent from the anterior notal wing processes.
The importance of the detached plates of the scutum and the axillary
sclerites morphologically and phylogenetically in the wing mechanism
will be discussed later.

The scuTELLUM (SCL,z) is a comparatively small, convex, trans-
verse, ovoid sclerite from the posterolateral angle of which the corru-
gated axillary cords (AXC) are produced. A small transverse sclerite
is closely applied to and partly fused with the anterior margin of the
axillary cord. Along the line of fusion a low ridge is produced in-
ternally and to it the postscutellum (PSCLz) is articulated.

The PosTSCUTELLUM (PSCL2) (= acrotergite, Whedon, 1938) is
even larger than the scutum. It is a subrectangular sclerite, pigmented
laterally only, separated by an internal V-shaped ridge into three re-
gions which are probably inaccurately termed median postscutum and
lateral postscutella by Tillyard (1917).

The anteriormost part of the metatergum is a narrow transverse
sclerite, the pREScUTUM (PSC3;), with a submarginal suture along
which a low ridge is produced internally. It is mostly obscured by the

18 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

preceding postscutellum to which it is connected by a tiny linear scler-
ite on each side.

The scurum (SCT;) is a large transverse sclerite with rounded
anterior angles. Its central portion is weakly sclerotized and unpig-
mented. A small spinelike apodeme is produced internally from a pit
which is situated very close to its anterior margin. Laterally the
suralar, adnotal, and adanal sclerites are not separated from the scutum
(also true in Anax junius).

The scurELLUM (SCLs;) and the avillary cords (AXC) are similar
to those of the preceding segment, except that the latter structures are
divergent posteriorly in the metathorax.

The POSTSCUTELLUM (PSCL;) is mainly weakly sclerotized, un-
pigmented, and merged with the membrane between it and the first
abdominal tergite, except for a trace of sclerotized area on each side
posterior to the axillary cord.

PLEURA

(Figure 21)

The pleura of the synthorax are very large and greatly modified.
Laterally the synthorax has two oblique sutures, the mesothoracic
pleural (PLS,) and the metathoracic pleural (PLS:) sutures, located
between the coxae and wing base of their respective regions. On the
lower portion of the synthorax between the two pleural sutures is
a transverse ovoid spiracle, the posterior spiracle, or metastigma
(IIISP). Just anterior to the metastigma is a short slanting suture,
the middle lateral suture (MLS), which represents the remnant of
the intersegmental suture (Snodgrass, 1909).

Different names have been used by various taxonomic workers to
designate the above-mentioned sutures as follows:

Mesothoracic pleural suture

=Humeral suture (Calvert, 18903; Needham, 1903, 1930; Tillyard, 1917;
Fraser, 1933).
=First lateral suture (Needham, 1903, 1930).

Intersegmental suture (Snodgrass, 1909)

=Interpleural suture (Tillyard, 1917).

=Middle lateral suture, or middle suture (Needham, 1930).

=First lateral suture (Rambur, 1842; Calvert, 1893; Tillyard, 1917).
=Second lateral suture (Needham, 1930).

=Anterolateral suture (Fraser, 1933).

No. 6 MORPHOLOGY OF THE DRAGONFLY—CHAO I9

Metathoracic pleural suture

=Second lateral suture (Calvert, 1893; Tillyard, 1917).
=Third lateral suture (Needham, 1930).
=Posterolateral suture (Fraser, 1933).

The course of the mesothoracic pleural suture (PLSz) is crooked,
with its lower one-fourth almost perpendicular to the longitudinal axis
of the body, its upper three-fourths slightly bowed and slanting
posteriorly, and with a short portion between them smoothly curved.
The angle of skewness is 60° and the angle of tilt of wing bases is 32°
in the present species. These angles are greater than the corresponding
angles in any gomphine dragonfly measured by Needham and Anthony
(1903). Needham and Anthony defined the degree of skewness or
inclination (also called angle of humeral suture) as the acute angle
between the suture and an imaginary line perpendicular to the longi-
tudinal axis of the body, and the angle of tilt of wing bases as the
acute angle between a line drawn through the wing bases and the longi-
tudinal axis of the body. It must be pointed out that Tillyard (1917)
used the term angle of obliquity in synonymy with skewness, both of
which he defined as angle of tilt of wing bases. This must not be
confused with the angle of skewness of Needham and Anthony.

The metathoracic pleural suture (PLS;) has almost the same shape
as the preceding one except that its lower one-fourth is a little slant-
ing posteriorly instead of almost perpendicular to the longitudinal axis
of the body, and its upper three-fourths is almost straight.

Particular attention is here given to the courses of the pleural
sutures and the relative positions of the wings and the legs. The
older view as to the phylogenetic origin of the orientation of these
parts is well expressed by Tillyard (1917, 1926), Imms (1948), and
others who maintain that the great development of the mesothoracic
anepisterna “pushes” the wings backward away from the head, carry-
ing the terga with them, and that the correlated growth of the meta-
thoracic epimera “pushes” the sterna and the legs forward so that the
latter come into position close behind the mouth. However, judging
from the courses of the pleural sutures, it is believed that the vertical
positions of the lower portions of the pleural sutures (a condition usu-
ally considered to be primitive) probably indicate the primitiveness of
this region while the posteriorly slanting positions of their upper por-
tions probably indicate the evolutionary enlargement of the upper
portion of the synthorax. The static nature of the lower region indi-
cates that the legs have not been “pushed” forward, while the poste-
riorly slanting position of the upper region indicates that the wings

20 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

om SC
H ~ [:-® rR +M

G2\)\ AYl SxCoP
NL ISS:

i : 2
2 P yess

INTERNAL VIEW
OF DORSAL PORTION

Fics. 21-23.—Synthorax.

No. 6 MORPHOLOGY OF THE DRAGONFLY—CHAO 21

have moved backward away from the head, to a position at or near
the center of gravity of the greatly elongated body of the insect.

Regarding the positions of the three pairs of legs, it is interesting
to note that the knee joint between the femur and the tibia of the
prothoracic leg is directed sideward, that of the mesothoracic leg, side-
ward and backward, that of the metathoracic leg, backward. The
pleurocoxal articulation of the metathoracic leg indicates a rotation of
its axis of about 90°. The result of the rotation of the metathoracic
legs might be of definite advantage to the insect in catching prey dur-
ing flight or in perching on the twig. The legs are not fitted for walk-
ing but they serve very well for climbing when that mode of progress
is required.

The mesothoracic episternum is divided by an inverted V-shaped
suture into anepisternum and katepisternum. The anepisternum
(AES,) is greatly expanded and meets with the corresponding part
of the other side of the thorax along the middorsal line to form a
ridge, the dorsal carina (DCR), anterior to the wing bases. At the
anterior end of the dorsal carina there is a transverse ridge, the
collar (COL), which is adorned with fine wavy hairs. On the dorsal
aspect of the synthorax the area between the collar and the wing bases
is called the front of synthorax. At about the middle of the front of
synthorax the dorsal carina is raised into a sharp point. Posterior to
this point the carina is divided into two low ridges which are parallel
for a short distance and then widely divergent. These ridges are
collectively called the antealar ridge (ARG), or crest. The area poste-
rior to the ridge is called the antealar sinus (AAS). Anterior to the
collar there is a transverse sclerite, the spiracular dorsum (SPD),
which is medially invaginated to form a hornlike apodeme. The de-
flected portions of the spiracular dorsum are called mesostigmatic
laminae (MSL). Each lamina bears an anterior spiracle, or meso-
stigma (IISP). A preliminary study of a few species of gomphine
dragonfly nymphs shows that there are three pairs of small interseg-
mental sclerites: a median pair (the members of which are narrowly
separated from one another), a lateral pair bearing the spiracles, and
a ventrolateral pair anterior to the katepisternum. Some, if not all, of
these plates are referred to as prothoracic spiracle plates by Snodgrass
(1909) who states that ‘‘in the adult they unite with each other across
the back, thus forming a complete spiracular dorsum which fuses with
the mesothorax. . .”

The katepisternum (KEP.z) is a vertical sclerite ventrally separated
from the sternum by a distinct oblique suture, the sternopleural suture
(SPS), from which two large apodemes, the prefurca (PF.) and the

22 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

squame (TNz), are produced internally. Anteriorly the katepisternum
is flanged by a narrow strip of sclerite which is continuous with the
mesosternum and is probably a part of it.

The mesepimeron (EPM,) is fused with the metathoracic anepi-
sternum, except below the metastigma where the middle lateral suture
(MLS) is separated from the mesepimeron. This sclerite surrounds
the posterior half of the mesothoracic coxal cavity.

The metathoracic episternum is divided into two parts, the anepi-
sternum (AES;) and the katepisternum (KEP;), by an incomplete
and slightly undulate suture below the metastigma. Dorsally the anepi-
sternum is produced into a hairy lobe between the two wing bases.
Ventrally the katepisternum extends to a place lateral to the furcal
pit, without any suture separating it from the sternum. The meta-
thoracic epimera (EPM3) are very large, fusing ventrally with each
other to make a large unsutured area. This area bears a conspicuous
median inverted Y-shaped pigmented area that is quite in contrast to
that of the neighboring regions because of its lighter color and the
direction of pigment streaking. The Y-shaped area has longitudinal
streaks while the neighboring portions have transverse streaks. A
pair of small apodemes at the center of the Y and a low ridge along
the stem of the Y are produced internally; the latter thickening is
visible externally and was often mistakenly regarded as a suture by
earlier workers. The area between the arms of the Y is called post-
sternum or pseudosternum (PSTN) and is generally regarded as a
secondary sclerite filling the gap between the metasternum and the
first abdominal sternite. This interpretation seems very inadequate
and a more careful study of this sclerite is very desirable. The post-
coxale (PCX;) is situated on the mesal edge of the metathoracic coxal
cavity. The latter is elongated, with the coxal articulatory process
lateroposteriorly located.

STERNA

(Figure 24)

The mesosternum (IIST) is a clearly defined area between and in
front of the two furcal pits, medially keeled, laterally separated from
the katepisternum by an oblique suture, the sternopleural suture
(SPS), anterolaterally produced into a narrow piece along the an-
terior margin of the katepisternum on each side. The furcal pits
(FP.), or apophyseal pits (Ferris, 1940), are close to one another,
and are situated near the mesal margins of the coxal cavities. The
metasternum (IIIST) is medially keeled, laterally deeply invaginated
along both sides of the keel (fig. 26, posterior view of cross section

No. 6 MORPHOLOGY OF THE DRAGONFLY—CHAO 23

of metasternum posterior to furcal pits; fig. 27, anterior view of cross
section of same slightly posterior to the preceding section), and partly
obscured by the approximately raised postcoxales (PCXs3).

SYNTHORACIC ENDOSKELETON
(Figure 25)

The synthoracic endoskeletal projections are of different forms:
(1) Ridges, (2) a complicated fusion product called the neural canal
surrounding the nerve trunk, and (3) a median hornlike apodeme
on the spiracular dorsum.

The ridgelike apodemes are mesopleural and metapleural, interseg-
mental, peristigmatic, and precostal apodemes. The mesopleural and
metapleural apodemes (PLA;, PLA;) are strengthened by about eight
short ridges projecting from their posterior sides. The lower portion
of the mesopleural apodeme along the edge of the katepisternum and
the apodeme between the anepisternum and katepisternum are inap-
propriately called (due to different interpretation of sclerites) the
sternoepimeral and sternoepisternal apodemes respectively by Tillyard
(1917). The intersegmental or interpleural apodeme (IPLA) is a
simple ridge with a spinelike process at its upper end near the meta-
stigma, and with a very long fine tendon projecting from it at about
the middle of its course. The peristigmatic apodeme (PSA) lies along
the suture separating the metathoracic anepisternum from the kat-
episternum. The precostal apodemes are situated along the meso-
thoracic sternopleural sutures. From each precostal apodeme two dis-
tinct structures are produced: the prefurca (PF,.) and the squame
(TN,). The prefurca is a tonguelike structure along the edge of the
stigmatic lamina. The squame (which term is also used in maxilla)
consists of a short stalk apically expanded into an elongate flat surface,
situated anterior to the coxal cavity: this is regarded by Tillyard,
probably erroneously, as a part of the furca.

The NEURAL CANAL (figs. 26, 27, NC) is a complicated fusion
product of several invaginated processes. This fusion product is dif-
ferentiated into two portions: an anterior portion on the mesosternum
and a posterior portion on the metasternum. These two portions are
connected dorsally, but are open between them on each side.

The anterior portion of the neural canal is formed into a complete
ring by the apical fusion of the mesothoracic furcal arms. The dorsal
portion of this ringlike structure is expanded into a flat surface which
is produced anteriorly into a pair of short protuberances, a pointed
process curling ventrad (fig. 24), and lateroposteriorly into a long flat

24 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

arm which is attached to the posterior margin of the postcoxale to
act as a brace.

The posterior portion of the neural canal is an elongated structure
formed by the apical fusion of the invaginated sternal fold (fig. 26,
PLF). The dorsal of this fusion product is greatly expanded into a
shieldlike structure which is an elongate, ovoid, flat surface, with two
foci of heavy sclerotization anteriorly and posteriorly with a pair of
long, narrow, ribbonlike tendons (fig. 25). The foci of sclerotization
indicate the positions of the metathoracic furcal arms which are
lateral to the sternal folds (fig. 25). From each focus a very long,
fine tendon is produced.

WINGS

The wings are held horizontally on both sides of the body: they are
unable to fold back on the top of the abdomen. This method of hold-
ing the wings is pointed out by Crampton (1924) and his contempo-
rary workers as an important archaic characteristic of the Palaeoptery-
gota (including as living forms the dragonflies and mayflies). It
appears that this condition is accounted for by the primitive structure
of the wing base. However, it is interesting to note that great con-
fusion exists in the literature regarding the structure of the regions
at the bases of the wings of dragonflies and possibly also of mayflies.
Structures involved are (1) wing base, (2) axillary sclerite and
lateral regions of scutum, (3) pleural wing process and epipleurites,
(4) principal wing muscles, (5) articulatory points, and (6) the
mechanics of flight.

The wine BASE (fused bases of wing veins) consists of two
strongly sclerotized plates, the anterior humeral plate (HP) and the
posterior axillary plate (AXP). Dorsally the humeral plate is divided
into three lobes by transverse grooves. Ventrally the lateral edge of
the humeral plate is connected by membrane to the distal margin of
the pleural wing process. It does not seem to form an articulation with
the anterior arm of the pleural wing process such as is mentioned by
Snodgrass (1935, p. 221). The axillary plate is subquadrate in shape,
slightly convex dorsally, posteriorly fused with the axillary cord
(AXC). The costal vein (C) is articulated with a small intermediary
plate (IP) which is in turn articulated with the posterior lobe of the
humeral plate. Ventrally a short, rounded protuberance is found at
the fused bases of C and Sc. The veins posterior to R+M are firmly
fused with the axillary plate. The base of R+M is forked. Its dorsal
branch strengthens the anterior margin of the axillary plate. Its ven-
tral branch forms a process which is articulated with the pleural wing

No. 6 MORPHOLOGY OF THE DRAGONFLY—CHAO 25

INTERNAL VIEW OF
VENTRAL PORTION

SYNTHORAX,
VENTRAL VIEW

ANTERIOR VIEW OF
POSTERIOR VIEW OF X-SEC. BEHIND FP3 X-SEC. BEHIND FP3

a; a3 eean® EE
ee DTTA“S PRETARSUS,
g Ae EW ct)) VENTRAL VIEW

Fics. 24-30.—Synthorax and legs.

26 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

process and is connected with the subalare by tough membrane. Thus
the base of R+M has the same function as the second axillary of
other orders of insects. It is probable that the second axillary sclerite
is formed by detachment of a portion of the base of R+M or R.

AXILLARY SCLERITE. Regarding the axillary sclerite (s), special
attention is given to (1) the number and (2) the origin, since there
appears to be a considerable amount of confusion in the literature
regarding these considerations. In the species studied here there is
only one, i.e., the first axillary sclerite (1AX), described previously.
This sclerite has the same shape and is situated at the same position
as that illustrated by Snodgrass (1909) for Pachydiplax longipennis.
Another small sclerite between the adnotal sclerite and the anteromesal
margin of the axillary plate is probably detached from the adnotal
sclerite. Its homology is not certain.

Crampton (1924) mentioned that in Palaeopterygota “there are
frequently no alar ossicles, or at the most but one.” Forbes (1943)
maintained that in the dragonflies there are no basal sclerites dorsally,
or “no trace of dorsal axillary sclerites as separate elements.” Snod-
grass (1909) pointed out that “only one distinct axillary is present”
in the dragonflies. This axillary sclerite is clearly illustrated by him
in this paper (1909, fig. 17, IAx), but not mentioned or illustrated
in his later paper (1930) or his well-known “Principles of Insect
Morphology” (1935). The present studies are in agreement with
Snodgrass but not with Forbes.

As to the origin of the axillary sclerites, Crampton and Forbes
differ in opinion. Forbes (1943) mentioned that “the extreme bases
of the veins are modified into a series of thickened knobs, the axillary
sclerites.” Crampton (1942) maintained that “the axillary sclerites,
alar ossicles or pteralia . . . are apparently formed, in part, as de-
tached portions of the lateral region of the notum, and partly as de-
tached basal portions of the wing veins, or as sclerotized areas at the
bases of the veins.” He considered, on the basis of numerous com-
parative studies, that the first axillary or the notopterale (notale)
“probably represents a detached portion of the lateral edge of the
notum.” Crampton’s opinion is adopted in this paper since in Anax
junius the condition of fusion of the first axillary with the anterior
notal wing process is perhaps indicative of such an origin.

PLEURAL WING PROCESS and EPIPLEURITES. At the dorsal end of
each pleural suture the pleuron is produced into an inverted foot-
shaped pleural wing process (PLP). The tip of the foot (the poste-
rior longer arm of the wing process) acts as the principal pivot for
the articulation of the wing. The heel of the foot (the anterior shorter

No. 6 MORPHOLOGY OF THE DRAGONFLY—CHAO 27

arm of the process) is connected to the humeral plate by membrane
and does not seem to form an articulation with the latter. The an-
terior basalare (BA) and the posterior subalare (SA), collectively
called the epipleural sclerites or paraptera, are present. They are very
small externally and so deeply imbedded in the membrane as to be
easily overlooked. Internally each has a very large apodeme which
has a stalk greatly expanded apically into a large surface for the at-
tachment of the depressor muscle. The basalare is connected by tough
fibrous membrane to the lateral portion of the anterior lobe and, to a
lesser extent, to the posterior lobe of the humeral plate. The subalare
is connected by similar membrane to the ventral branch of the base of
R+M. It is also interesting to point out that the presence of the
basalare and the subalare in dragonflies was probably correctly de-
termined for the first time by Forbes (1943). They were considered
as “‘cap-tendons” by earlier workers (Calvert, 1893; Tillyard, 1917).
Snodgrass (1935) mentioned that “there are no epipleural sclerites in
the dragonflies.” Probably he also considered the epipleural sclerites
as tendons, since in the same works he mentioned that “in Odonata
there are two anterior wing muscles . . . inserted by long tendons
directly on the large humeral plate of the wing base,” and that “two
posterior pleural wing muscles take their origins on the ventral edge
of the epimeron in each alate segment and are inserted directly on the
axillary plate of the wing base.” However, they seem to be more
appropriately considered as basalare and subalare instead of tendons,
since (1) they are distinct, though small, sclerites imbedded in the
membrane in the same positions in which the epipleurites are found
in other orders of insects; (2) they serve for the attachment of direct
muscles as they also do in other winged insects; and (3) they are
connected to the wing base by tough membrane as in other winged
insects. Particular attention is called to the connection between the
subalare and the ventral branch of the fused bases of R+M—a
condition similar to the connection between the subalare and the second
axillary sclerite.

WING MUSCLES. Nine wing muscles in dragonflies have been recog-
nized by Berlése (1909), Calvert (1893), and Tillyard (1917), but
the apodemes of some sclerites to which the muscles are attached were
called cap-tendons, and the sclerites considered to be of no morpho-
logical importance. Of the nine muscles, three are very large: (1)
The principal elevator, (2) the anterior depressor, and (3) the poste-
rior depressor. The principal elevator (see Tillyard, 1917, p. 204,
fig. 89, A, pe: and pez) is attached to the geat apodeme (AP2, AP3)
produced internally from the detached plate of the scutum. The an-

28 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

terior and posterior depressors are attached to the basalare and suba-
lare respectively.

Snodgrass (1935) classifies flight muscles into two types, direct
and indirect. The direct muscles are the axillary and epipleural mus-
cles and the indirect are the dorsal and tergosternal muscles. The
tergosternal muscles, according to Snodgrass, “are attached dorsally
on the anterior lateral areas of the tergum, and ventrally on the
basisternum before coxae.” The present study shows, however, that
dragonflies have both types of flight muscles. This interpretation
differs from what Forbes (1943) states that in Odonata the indirect
muscles are “nonfunctional” or that “only direct wing muscles” are
present. The direct (epipleural) muscles have been discussed before.
There are several pairs of indirect muscles, the most important of
which are the principal elevator muscles, which are attached to the
detached portions of the scutum dorsally and to the “squame of furca”
(produced internally from the pleurosternal sutures) ventrally. They
have been homologized with the “‘first tergosternal” of other insects
by Berlése (1909) whose opinion is adopted by Tillyard (1917). The
dragonfly wing mechanics appear therefore to be not fundamentally
different from those of higher groups of insects.

ARTICULATORY POINTS. Each wing is articulated with the thorax
in three places: (1) The ventral branch of R+M articulates with
the posterior arm of the pleural wing process. This is the principal
pivot of the wing. (2) The humeral plate articulates with the prealare.
(3) The axillary plate articulates with the posterior notal wing proc-
ess. The prealare-humeral articulation is particularly interesting in
two respects: First, the prealare normally “extends laterad or ventrad
to the episternum and thus supports the notum anteriorly on the
pleural wall of the segment” (Snodgrass, 1935). In the present
species, the prealare is separated from the episternum by a large mem-
branous region. It does not offer any support to the notum anteriorly ;
the latter thus moves up and down freely and synchronously with the
movement of the wings. Secondly, the prealare-humeral articulation is
probably unique to dragonflies, since in other winged insects the
wings are articulated with the anterior wing processes instead of the
prealares.

MECHANICS OF FLIGHT. Since the author has not studied the mus-
cles involved in controlling the wing movements, a discussion on the
possible mechanics is based on inferences concerned with the sclero-
tized structures described on a previous page.

The wing mechanics in the dragonflies are similar to those of
higher winged insects in two fundamental respects: (1) The wings

No. 6 MORPHOLOGY OF THE DRAGONFLY—CHAO 29

are primarily controlled by the antagonistic indirect elevator muscles
and the direct depressor muscles attached to the detached lateral plates
of scutum and epipleurites respectively; (2) the subalare is con-
nected to the wing base very close to the fulcrum while the basalare
is far in front. This latter condition indicates that a pull upon the
posterior depressor muscles would strongly depress the wing while a
pull upon the anterior depressor not only would depress the wing but
also deflect the anterior part of the wing to produce the sculling type
of flight.

On the other hand, the wing mechanics in dragonflies differ from
those of higher winged insects in two important respects: (1) The
prealare-humeral articulation is probably unique in dragonflies. This
articulation forms the anterior part of the hinge line and differs from
that in other winged insects in which the anterior notal process forms
an important articulatory point. (2) Another feature probably also
unique in dragonflies is that the terga are connected to the pleura by
membrane only and the phragmata produced internally along the ante-
costal sutures are very small. The latter fact indicates that the longi-
tudinal dorsal muscles attached to the phragmata would be small, such
as illustrated by Berlése (see Tillyard, 1917, p. 204, fig. 89, A, pt).
The smail size of the dorsal muscles probably indicates that they do
not produce an effective antagonistic action to the tergosternal elevator
muscles. Judging from the above facts, it seems that the terga must
probably move up and down during flight without distortion. This
condition differs from that of other winged insects in which “the
restoration of the dorsal curvation of the back by the contraction of
the longitudinal dorsal muscles” will effect, in part, the down-strokes
of the wings (Snodgrass, 1935, p. 234).

As to the control of the direction of flight, the fore wings are
probably more important in this action than the hind ones. This con-
tention is based on the fact that the articulatory plates of the scutum
are distinctly separated from the main body of scutum (detached
plates) in the mesothorax but firmly fused with it in the metathorax.
Thus the hinge line and consequently the pitch of the fore wing can
be changed, but that of the hind wing appears to be fixed.

WING VENATION
(Figures 31, 32)
The wings are transparent and supported by numerous veins form-

ing a complicated network. The fore wings (fig. 31) are widest at
the nodus (N) which is located at the middle of the anterior margin

30 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

of the wing. The hind wings (fig. 32) are slightly shorter and about
one-fifth again as wide as the fore wings (the widest portion of the
former is a little basal to the nodus which is situated at the basal
two-fifths of the wing). In the male the inner margin of the hind
wing is excavated and fringed with a narrow white edge, the mem-
branule (mb), and the anal angle, or tornus (Aa), angulated. In the
female the inner margin of the hind wing is not excavated and the
anal angle is rounded.

Both pairs of wings have a very conspicuous pterostigma (Pt) of
the same size, shape, and color. The pterostigma is also called stigma
or anastomosis, the latter being very rarely used. The term “stigma,”
though frequently used, is rather confusing because it may refer also
to the spiracles. It is a thickened area between Sc (Fraser, 1948) and
R, near the apex of the wing, dark brown or black in color, elongate,
about four times as long as wide, with its shorter sides oblique and
parallel to each other; with its longer sides concave, surmounting
about six cells; and with a strong vein, the brace vein (br. v.), ex-
tending down from its basal side.

Several systems of notation of the venation have been proposed.
Originally, de Selys gave a name to each vein without notation in use.
Later on, Needham gave an interpretation based on the larval wing
tracheation, with a notation based on the Comstock-Needham system.
This system has been widely used in the last thirty years or so. Till-
yard (1926) gave a different interpretation, based on the study of
fossil forms. His system, with a few modifications, has been generally
accepted by later entomologists. Borror (1945) summarized in two
tables the different systems of terminologies used by different authori-
ties, such as de Selys, Kirby, Needham, Tillyard, Tillyard and Fraser,
and others. Forbes (1943) gave very different notations. Fraser
(1948) modified the costal vein and the anal veins. The present paper
uses Tillyard’s system (1926) with a few modifications by subsequent
workers. To summarize the points, a table (p. 31) is prepared to show
a comparison of terminology of the principal longitudinal veins of the
dragonfly wing.

PRINCIPAL LONGITUDINAL VEINS. The principal longitudinal veins
are costa, subcosta, radius (4 branches with 2 intercalaries), media
(a single vein), posterior cubitus, and anal.

The costa (C), or costal vein, is a simple and strong vein which,
according to Fraser (1948), extends from the base of the wing to
the nodus.

The subcosta (Sc) is a long vein posterior to the costa. According
to Fraser (1948), its course is from the base of the wing to the nodus

MORPHOLOGY OF THE DRAGONFLY—CHAO 31

No. 6

WAS
xVI

nO-+W+Sa
NI

(8£6T)
Jose1y pue

PreATIL

wny

ny inte)

VIN A

Sty SIA

eB ea

*Y aN

“YI

ey TW

et Ty

Ny NI

Wt+u Wt+u

9S Ss)

8) a)
(9761) (LT6T)

preAqihL

Joj9as [eseq-qns [eUIxO1g
9]8ULII JO 10}99S IOLIOJUT
®}S09}S0g

9] 8UeII} JO 1019s Jo1Iadnsg
snjnoie JO 10}9as JaMO'T

10}99S URIPI[

Joj9as [epouj}sog
4o}0es jediourig

snynore jo 10}09as 1addy

ueIpay

eysSOIqNS

eSsOD

ul9}SAs UPISATIS

SUIDA JeuIpNysuoy jediourid sy 7,

W+u

uorzeIAsIqqy

Gece Oo tras ene PIpaUl JOLIN}

en Orca, OnOeO sniper qyqyu-+ YjuAnoy

es AxeyeoiojUl SNIpes PITY,

@<el6 [e-:0)|6) s)18|9:.0° ollm)e eile sniper puyL

"+++ ATeTVIIIJUL SNIPeI PUOIdS

a ele te (oD wyexa Fuenama/ielers snipes ISI

"+++ (snyjnore 03 seq 3utj\)
Pep OSCiCetO Te mtacr Cat) elpoul- snipey

owen

Jaded sq} ul pesn ASojouTUTIE

uoynuar Guw2 ayouopo fo sarbojoumtsa} fo uostavgdmoj— I Adv]

32 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

where it makes a distinct curve anteriorward toward the costal margin,
at which level it straightens out again along the margin of the apical
half of the wing.

The radius and media have a common stem (R+M) which forms
the anterior border of a large cell, the basal space (bs), at the base of
the wing. From the anterior external angle of the basal space the
R+M gives off two branches, namely R, and the anterior portion of
the arculus (arc). R, extends to the apex of the wing. It is parallel
to Sc and forms the posterior border of the pterostigma (Pt). The
arculus is an oblique vein between R+M and CuP. It forms the ex-
ternal border of the basal space. It consists of two portions, namely,
the anterior portion and the posterior portion. The anterior portion
is the fusion of the radial sector and media (Rs+M). They separate
from the middle of the arculus. Rs is then called the superior or
anterior sector of arculus and media the inferior or posterior sector
of arculus. The radial sector has three branches, with two intercalary
veins between them. The three branches are R2, R3 and Ry,;. The
two intercalary veins are IR, and IRs, the former being posterior to
R, and the latter to Rs. Near the base of Rg; there is an oblique vein
(O) (=LO, lestine oblique vein of Fraser, 1944) between Rs and
IR;. The basal portion of IR; from its base to the oblique vein is
called bridge or bridge vein (br) by Needham. This bridge was re-
garded by Needham as a secondary extension of Rs backward toward
the base of the wing. Tillyard (1922, p. 7) believed that “it was never
formed backwards as a bridge vein but was always the basal portion
of a strongly formed main longitudinal vein arising from R4,; or
sometimes from R2,3 as in most recent forms.” On the other hand,
Fraser (1944) stated that “IR; never originated from a basal source
but extended inwards from a peripherial one.”

The media (MA) is a simple vein extending from the arculus to
the apical third or fourth of the posterior margin of the wing. It is
called by Lameere (1922) and Tillyard (1926) anterior median
(MA). Its basal portion forms the anterior border of the hyper-
triangle (h).

The posterior cubitus (CuP) is a crooked vein. It extends from
the base of the wing to a point beyond the arculus and then bends
abruptly posteriorly almost at a right angle, forming the common side
of the triangle (t) and the subtriangle (s). From the posterior apex
of the triangle CuP extends to a point slightly beyond or basal to the
middle of the posterior margin of the fore and hind wings respec-
tively. Lameere called this vein the posterior branch of cubitus (CuP),
and in this has been generally followed by subsequent authors on

NO.

MORPHOLOGY OF THE DRAGONFLY—CHAO

8
AROS

n
may

- i
PTC SUTC

' \
t Surc

ATG

ATG ve
34 BASAL SEGMENTS-LATERAL
VIEW, 2

STP
Siz

3 StN
STERNITES VI VI, ¢

BASAL SEGMENTS - DORSAL VIEW,9

e,
EE
(TRY
THES Ty
PRS

LS
LSOKA Ri
ay R2

See
weeps:
>

hy
RES" RS
te

a STN
36

STERNITET, ¢
STP

1
aN

4 ie a — STN
Hea 6
O~S STERNITE I,¢

Sierra STN
38
‘ STERNITE V, 2

\ '
A

3 oi Salaam
“STERNITE VI,¢

Fics. 31-39—Wings and abdomen.

34 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

dragonflies (e.g., Fraser, 1940; Borror, 1945). Forbes (1943) called
this vein the plait vein (Pl) which, according to him, corresponds to
the anal furrow of insects of other orders. The basal portion of this
vein is regarded as the common stem of the cubitus (Cu) by Needham
(1903) and as the common stem of Cu,+1A by Tillyard and Fraser
(1938).

The anal vein (A) is slightly undulated at its basal portion. It
meets with CuP at the apex of the triangle. The apical portion of the
anal vein (A,) from the apex of the triangle to the posterior margin
of the wing is parallel to CuP in the hind wing. On the other hand,
in the fore wing it is parallel to CuP for a long distance and then di-
verges from the latter. The notation A used in this paper refers to
the vein from the base of the wing to the posterior apex of the tri-
angle. A, denotes its extension. Tillyard (1926) regarded the basal
portion (A) as a backward extension of IA toward the base of the
wing. He designated the portion from the posterior apex of the
triangle to the cubito-anal cross vein (cu-a) as Ab (anal bridge) and
the portion between cu-a and the base of the wing as A’. Fraser
(1938) demonstrates that the anal vein in the Odonata has an inde-
pendent origin from the base of the wing. (The present notation A
equals Ab+A’ of Tillyard.) In the hind wing the anal area is greatly
expanded. There are three distinct anal veins, namely, Aia, As, and
As. Aia and A, become fused not far from their origins thus forming
a 2-celled anal loop (AL). The vein that separates the anal loop into
two cells is called the anal supplementary (mspl) or midrib (mr).
Forbes (1943) used a new term, “axillary” (AX), for the veins often
called “anal” but which, according to him, are distinct from the origi-
nal anal vein which is usually the Pl of his system.

CROSS VEINS. The more important cross veins are as follows:
Nodal, subnodal, antenodal and postnodal, primary antenodal, brace
vein, arculus, oblique vein, and cubitoanal vein. The brace vein, arcu-
lus, and oblique vein have been mentioned before and will not be
repeated again.

The nodus, or nodal vein (N), is a thick vein situated between the
costal margin and R,. Both notations N and u have been used but the
former is adopted in this paper. Their positions in the fore and hind
wings have already been mentioned before. The anterior portion of
the nodal vein coincides with the bending part of Sc. The extension
of the nodal vein between R, and R,z is called subnodus (sn). Be-
tween the costal margin and R, there are many cross veins, those
between the base of the wing and the nodus being antenodal cross
veins (Ax) and those between the nodus and the basal end of the

No. 6 MORPHOLOGY OF THE DRAGONFLY—CHAO 35

pterostigma being postnodal cross veins (Px). The first and the fifth
or sixth antenodal cross veins are thickened and are called primary
antenodal cross veins (Ax, and Ax,). It should be noted that Forbes
(1943) used the same notation Ax for the veins which he called
axillary veins. This has been mentioned before. The number of
the nodal veins varies. In systematic work the nodal index such as
13-16 | 16-12
I3-II | II-13
antenodal and postnodal cross veins in the left and right fore and
hind wings respectively. The nodal indices of ten specimens are shown
as follows:

has been often used and indicates the number of the

"3-16.°| 16-12. «, 11-14 || @e=13 g: T-19 | 18-2E “o. 12-15) aa

Mee eee rein Olan) JiTte 1) emg 13204" rene |i esren

fey faG-1r a. 12-16'|| To-12 g, F219 | 18-10 oe 13-16 | 15-12 9:

BERNE OCURA ll te. ae eee ALE aaa OE LE aae | ’
[sonic Vaso elm fo resto 4 Tose |) angers) ger

2-07) G12 4. £3=10 [Bar
ai eS 2 eee eel ee ee

The cubito-anal cross vein (cu-a) is a vein between CuP and A,
basal to arculus for a considerable distance. Different notations have
been given to this vein, such as ac (anal crossing, Needham), Cux
(Tillyard, 1917), Ac (Tillyard, 1926), AC (Fraser, 1940), and cu-a
(Borror, 1945) ; the last notation is used in this paper. The different
terminologies are accounted for largely by the different interpreta-
tions of the anal vein.

CELLS and SPECIAL AREAS. The important cells and special areas
are as follows: Pterostigma, basal space, triangle, hypertriangle, sub-
triangle, discoidal field, basal anal area, anal loop, anal field, and anal
triangle. The pterostigma (Pt) has been mentioned before and will
not be considered again.

The basal space (bs) is an area at the base of the wing, bounded
anteriorly, posteriorly, and apically or externally by R+M, CuP and
arc respectively, the first two being subparallel and the last slightly
slanting. It is also called by different names, such as median space,
midbasal space, sub-basal space, and basilar space. It is about 3} to
4 times as long as wide.

The triangle (t) is a distinct and almost isosceles-triangular space
formed by CuP basally and two cross veins anteriorly and apically
(or externally) respectively. A space anterior to the triangle from
arc to the apex of the triangle is called hypertriangle (h), which is
also called by different names such as supratriangle, hypertrigone, or
supratrigone. It is a narrow space bounded basally by the posterior

36 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

portion of the arculus or m-cu, anteriorly by a portion of MA, and
posteriorly by a vein which is composite. The basal portion of this
composite vein is a part of CuP and the distal portion a cross vein
between CuP and MA. This cross vein is also the anterior side of
the triangle. Its direction is such that it looks like part of a longi-
tudinal vein. The subtriangle (s), or internal triangle, is also an
isosceles triangle. It is situated basal to the triangles with the vertical
part of CuP as their common side. It is bounded externally by the
vertical part of CuP, posteriorly by a part of A, and basally by a
slanting cross vein between CuP and A. Phylogenetically the triangle
and the subtriangle are collectively called the cubital area, which is
homologous with the discoidal cell or quadrangle of Zygoptera. The
area apical to the triangle between MA and CuP is called the discoidal
field. The discoidal field in the fore wing is of almost the same width
throughout except that the portion beyond the level of nodus is slightly
widened, while that of the hind wing is considerably widened at the
posterior margin of the wing. The difference in shapes of the discoidal
fields in both wings is accounted for by the different positions of the
apical portion of the CuP.

The basal anal area is a narrow space posterior to the basal space.
It is situated between CuP and A and is limited by the cross vein cu-a.

The anal loop (AL) is a 2-celled area in the hind wing bounded
by A, Aia, and A, on its anterior, apical, and basal sides respectively.
It is bisected by a short vein, the anal supplementary, or midrib (mr).

The anal field refers to the area bounded by A, Az, Aia+Ac, and
the posterior margin of the wing. It includes the anal loop in the hind
wing. In the fore wing its apical limit is about at the level of the
posterior apex of the triangle. The anal field has two rows of cells
between A and the posterior margin of the fore wing, whereas there
are five rows in the hind wing.

The anal triangle (at) is a 4-celled space at the extreme base of the
hind wing in the male, bounded anteriorly, externally, and basally by
A, As, and the basal margin of the wing respectively. The basal mar-
gin of the hind wing is flanged by a narrow whitish membrane, the
membranule (mb). In the female the anal triangle is 8- or 9-celled.

LEGS
(Figures 28-30)
The legs are small in comparison with the size of the body. They

are strongly armed with spines. The prothoracic legs are the smallest,
being slightly smaller than the mesothoracic legs, which in turn are

No. 6 MORPHOLOGY OF THE DRAGONFLY—CHAO a7

considerably smaller than the metathoracic legs. The difference in size
of the legs is accounted for especially by differences in length of the
femur, tibia, and tarsus, since the coxa and trochanter are not con-
spicuously different in lengths in the three pairs of legs.

The coxa (CX) is of moderate size, more or less conical in shape,
with its outer surface much longer than its inner surface which is
conspicuously bulged. The basal portion of the outer surface is modi-
fied to form the pleural articular socket which is articulated with the
pleural process of the thorax. The basal end of the coxa is girdled
by a submarginal basicoxal suture (BCXS), along which a ridge, the
basicosta, is produced internally. The basicosta is enlarged on the
outer surface posterior to the pleural articular socket. The basicostal
suture sets off a marginal flange, the basicoxite (BCX), which is en-
larged on the outer surface posterior to the pleural articular socket.
Distally the coxa bears an anterior and a posterior articular socket
to which the trochanteral articular condyles are attached.

The TROCHANTER (TR) is a slender segment, about as long as the
coxa, having a short outer surface so that its distal end is obliquely
truncated. A transverse constriction gives this segment a superficially
2-segmented appearance. The basal portion is attached to the coxa by
an articular membrane. It is also articulated with the coxa by an
anterior and a posterior trochanteral articular condyle. The coxo-
trochanteral condylic hinge is a right angle with the pleurocoxal articu-
lation so that it there forms a “universal joint’ which allows a wide
range of motion of the leg. A condylic hinge is also present at the
distal end of the trochanter. This operates at a right angle to the
coxotrochanteral condylic hinge, but permits of much less freedom of
motion than the latter since nearly the entire basal rim of the femur
is closely attached to the distal end of the trochanter.

The anterior surface of the articular membrane between the coxa
and trochanter is invaginated to form a deep pouch, the posterior wall
of which is thickened and tendonlike in structure. This undoubtedly
serves for the attachment of a muscle internally.

On the inner wall of the trochanter there is a group of short and
robust spines. There are only four or five spines on the basal, and
the same number on the distal, portion of the trochanter of the pro-
thoracic leg, whereas there are many on the distal portion of the tro-
chanter of the mesothoracic and metathoracic legs. These spines are
arranged at random: they are not arranged in a definite row to form
a trochanteral brush such as reported by Cowley (1937) to occur in
some other dragonflies.

The FEMuR (FE) is the longest segment of the leg, nearly cylindri-

38 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. I22

cal, armed with short stout spines on its inner surface which faces
ventrally in its natural position. The front femur is slightly curved
and therefore fits nicely along the side of the thorax. The hind femur
is straight, slightly compressed, and is held beneath the thorax. Its
spines are more irregularly arranged on the basal portion than on the
distal portion where they are distributed in two distinct rows. Pro-
ceeding distalward the spines are progressively larger and more widely
separated. The area between the two rows of spines is smooth, flat,
or slightly grooved, fitting it for the reception of the tibia when the
latter is flexed close against the femur. The distal margin of the
femur is crowned with a few short spines.

The trp1a (TI) is a slender segment, convex on its dorsal or outer
surface, flattened on its ventral or inner surface, armed laterally with
a row of flattened spurs and dorsally with two parallel rows of short
spines. It is constricted and slightly bent near its basal end which is
articulated with the former by an anterior and posterior condyle. The
articular membrane between the femur and the tibia has the same
condition of invagination as that found in the membrane between
trochanter and femur.

The flattened lateral spurs are of two types: (a) The short and
swordlike type; and (b) the elongate type. The short swordlike spurs
number about 8 or 9 in a row, situated on the apical half of the
anterior margin of the prothoracic tibia in either sex. They are col-
lectively called tibial comb (TIC) which was first pointed out by
St. Quentin (1936) to be present on the prothoracic tibia of ail dragon-
flies. Each spur is set on a socket which is oblong in shape. The spur
is not evenly sclerotized, but has one edge unpigmented and thinner
than the other edge. The elongate type of spur is long, undulated, and
pointed at the apex. Along the anterior margin of the tibia there are
four such spurs basal to, and one apical to, the tibial comb on the
prothoracic leg, whereas there are eight on the mesothoracic or meta-
thoracic leg where the tibial comb is absent. Along the posterior mar-
gin of the tibia there are eight to ten such spurs. Proceeding distal-
ward these spurs are progressively shorter but broader.

The function of the tibial comb is unknown. Garman (1917) said
that it might be used for the cleaning of the mouth parts and antennae.
Needham and Haywood (1929) said that it might serve to hold the
dragonflies’ food. St. Quentin (1936) mentioned that it might be used
for the cleaning of compound eyes.

On the ventral surface of the tibia there is a group of bristles along
the base of the tibial comb and a nonsclerotized structure near the
distal end of the tibia. This nonsclerotized structure is elevated and

No. 6 MORPHOLOGY OF THE DRAGONFLY—CHAO 39

elongated. It is present on the prothoracic tibia of the male sex only.
As far as my knowledge goes, this special structure occurs in many
species of gomphine dragonflies which I have examined. It has ap-
parently not been reported heretofore. Its function is not obvious
to me.

The distal end of the tibia is notched and somewhat socketlike and
therefore fitted for the reception of the bulbous basal region of the
basitarsus.

The tarsus (TA) is 3-segmented; the basal segment or basitarsus
(BTA) is the shortest; the distal segment or distitarsus (DTA) is
the longest. The latter is about as long as the basitarsus and the middle
segment taken together. Each segment is armed with a few spurs on
its lateral margin. The distitarsus (DTA) bears a ventrodistal pro-
jection, the plantella (PTL), which is well developed.

The pretarsus (fig. 30) or terminal region of the leg consists of
a pair of claws, empodium, and unguitractor. The claws (CL) are
the largest parts of the pretarsus, and they articulate with a small
dorsal process of the distitarsus. Each claw bears a ventral tooth and
a narrow, wavy, ridgelike structure on each side of it. Ventrally the
bases of the claws are connected with membrane which is also closely
attached to the mesally located unguitractor. The unguitractor is a
ventral sclerite which is partially hidden by the distal projection of
the distitarsus, the plantella. An empodium (EMP) is attached to the
distal end of the unguitractor. The distal portion of the empodium is
enlarged.

ABDOMEN

(Figures 33-50)

The abdomen is composed of Io distinct segments and, according
to Tillyard (1917), also of the reduced remnants of the 11th and 12th
segments. It is long and slender; the basal two segments tumid and
slightly compressed, becoming thin, slender, and cylindrical from seg-
ments 3 to 7 (more pronouncedly so in the male) ; dilated and de-
pressed from the posterior half of segments 7 to 9 (more pronouncedly
so in the male), widest across the apical end of segment 8; seg-
ment 10 ringlike (in male) or depressed (in female). The propor-
tional lengths of the segments from base to apex are approximately
BS OMOWS:.1-65:4,5 30.02 7.0:.7.0: 6.5.5.5: 4:052:5 5 1.5.

The male differs from the female in (1) having auricles (AU) on
segment 2, (2) in having accessory sexual organs (figs. 40-44) on the
ventral surfaces of segments 2 and 3, (3) the relatively great dilation
of segments 7-9, and (4) the relatively great length of the anal
appendages (figs. 45-50) at the extreme apex of the abdomen.

40 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

MALE TERGITES

The tergites are convexly arched. They occupy not only the dorsal
region but also the whole of the lateral regions of the segments. The
lateral edges of the posterior half of tergite 3 and also of tergites 4-6
almost meet midventrally, partially hiding the sternites from view.
With the exception of segments I and 10, the tergites are distinctly
separated from the sternites by pleural membrane. Tergites 1 and 2
are weakly sclerotized middorsally ; 3-7 weakened along the midline;
and 8-10 not so weakened. Tergites 3-7 are adorned with minute
spines middorsally and also along the posterior transverse and sub-
marginal ventral carinae. The dorsal spines are absent on the tergites
of the other segments. Tergites 3-6 are similar to one another, while
those of the other segments differ from the former in various ways.
Tergite 4, to be described first, illustrates the generalized condition.

TERGITE 4 is strengthened by the formation of both internal ridges
and external carinae. Basally it is girdled by a submarginal antecostal
suture (ACS), which is very narrow and usually obscured by the
apical portion of the preceding tergite. Dorsally the tergite is weak-
ened along its midline but, conversely, is strengthened by a supple-
mentary transverse carina (SUTC) which is situated at a point one-
fourth the distance from base to apex of the segment. Posteriorly it
is strengthened by the submarginal posterior transverse carina (PTC)
which is continuous with the submarginal ventral carinae (SVC), one
on each side along the ventral margins of the tergite. The posterior
transverse carina is conspicuous and adorned with small spines. The
submarginal ventral carinae are poorly formed and weakly sclerotized.

TERGITE I is separated from the sternite by narrow pleural mem-
branes except anteriorly where it is fused with the latter. Dorsally it
is adorned with a pair of tufts of long, fine hairs each on a slightly
elevated subapical area. Lateroventrally the acrotergite (ATG) is
enlarged and produced into a pouchlike evagination on each side. All
carinae are absent except the submarginal ventral carinae (SVC)
which are very poorly developed, with or without a few minute spines.

TERGITE 2 is peculiar in having a pair of lateral outgrowths, the
auricles (AU). These are situated in an oblique position on the ante-
rior portion of the tergite, and are weakly sclerotized except along
their crests, which are denticulate ventrally. The supplementary trans-
verse carina (SUTC) is situated slightly anterior to the middle of
the tergite and extends downward on each side posterior to the auricle.
A similar but much shorter structure is present posterior to the sup-
plementary transverse carina. The submarginal posterior transverse

No. 6 MORPHOLOGY OF THE DRAGONFLY—CHAO AI

VENTRAL VIEW,
PENIS & RIGHT HAMULES REMOVED

ANTERIOR LAMINA & HAMULES PENIS, LATERAL VIEW

Fics. 40-44.—Basal abdominal segments and male accessory sexual organs.

42 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. I22

carina (PTC) is more prominent dorsally than laterally. The sub-
marginal ventral carinae (SVC) are adorned with long hairs anteri-
orly and with spines posteriorly.

TERGITE 7 (fig. 45) is peculiar in having the median area of the
submarginal posterior transverse carina raised considerably and the
submarginal ventral carinae dilated into leaflike structures, the pseudo-
lateral dilations (PLD) (Fraser, 1934).

TERGITE 8 (fig. 45) is peculiar in having a pair of submedian, low,
rounded, transversely wrinkled tubercles. The supplementary trans-
verse carina is absent. The submarginal posterior carina is raised
medially where it is deeply notched. The submarginal ventral carinae
are greatly expanded to form pseudolateral dilations.

TERGITE 9 is similar to the preceding one except that (1) it is with-
out submedian tubercles, (2) it is submedially slightly constricted on
both sides, and (3) its submarginal posterior carina is slight notched
medially.

TERGITE 10 is completely fused with the sternite to form a ring.
Dorsally it is deeply concave on its apical margin, with a pair of sub-
basal, very low, ovoid, transversely wrinkled tubercles followed by
dorsal wrinkles paralleling the edge of the apical concavity. All carinae
are absent: the position of the submarginal posterior carina is indi-
cated by a few small spines laterally.

MALE STERNITES

STERNITE I (fig. 40) is short, transversely rectangular, antero-
laterally fused with the tergite. Posteriorly it is produced into two
short processes which are inflected and obscured from view by the
main portion of the sternite. These processes are connected with the
anterior processes of the anterior lamina of sternite 2. Laterally the
sternite has a pair of ovoid spiracles and a pair of small pits meso-
posterior to the former. From these pits short hornlike apophyses are
produced internally.

Sternite 2 and the anterior portion of sternite 3 are greatly modified
into complicated accessory sexual organs which are collectively known
as the copulatory apparatus. The main structure of the apparatus con-
sists of a penis which is lodged in a membranous depression, the
genital fossa, and is protected by various organs derived from sternite
2. Different parts of the apparatus will be described in detail as
follows.

STERNITE 2 (figs. 40-43) is modified to form the following parts:
genital fossa, anterior lamina, posterior lamina, supporting framework,
sheath of the penis, anterior hamules, and posterior hamules.

No. 6 MORPHOLOGY OF THE DRAGONFLY—CHAO 43

The genital fossa (GF) is a membranous depression strengthened
by anterior lamina, posterior lamina, and lateral supporting frame-
works. The sheath of the penis (SHP) is located on the posterior
part of this membrane.

The anterior lamina (AL) is a large sclerite, situated at the anterior
third of the second abdominal segment. It is differentiated into dis-
tinct anterior and posterior portions. The anterior portion is flat,
comparatively weakly sclerotized, anteriorly produced into two proc-
esses which are connected with similar processes of the preceding
sternite. The posterior portion of the anterior lamina is strongly
sclerotized, convex, and adorned with many minute tubercles and
hairs. The posterior margin of the anterior lamina is V-shaped,
irregularly indented, and with a short median cleft, the cleft of the
anterior lamina (CAL).

The posterior lamina (PL) is a large sclerite situated at the poste-
rior end of the second abdominal segment. It is weakly sclerotized,
constricted medially, and greatly expanded laterally.

The supporting frameworks, or the anterior portions of the frame-
works (APF) (Thompson, 1908) are a pair of sclerites on the lateral
sides of the genital fossa between the anterior and the posterior
laminae. Each is an elongate sclerite, slightly convex ventrally, with
a subapical mesodorsal process which supports the base of the sheath
of the penis (SHP). Ventrally the sclerite has a low transverse ridge
to which the posterior margin of the base of the anterior hamule
(AH) is attached. Posteriorly the sclerite is emarginated to form,
together with the anterior extension of the posterior lamina (PL), a
socket to which the posterior hamule (PH) is attached.

The sheath of the penis (SHP) is a placoid structure composed of
a scooplike structure arising ventrally from the base which is imbedded
in the membranous genital fossa. The scoop is supposed to be for the
protection of the penis. The base of the sheath is five-sided, medially
with a broad shallow groove. The anterior margin of the base is
articulated with and supported by the two arms of the paired support-
ing frameworks. Each of the posterolateral angles of the base is con-
nected with a small slender sclerite, the outer end of which is articu-
lated with the posterior margin of the base of the posterior hamule.

The anterior hamules (AH) each consists of a long bifurcated proc-
ess produced ventrally from the posterior portion of an elongate
sclerite which is attached to the low ridge of the supporting frame-
work. Mesally it is attached to the lateral margin of the anterior
lamina. The two processes are of unequal length, pointed apically.
The anterior process is hooked apically, about twice as long as the
posterior process, and is subequal in length to the stem.

44 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. I22

Each of the posterior hamules (PH) is a robust structure attached
to the socket formed by the supporting framework and the posterior
lamina. It is about as long as the anterior hamule, pointed, with more
bristles on its mesal surface and apex than elsewhere.

TABLE 2.—Designations for the segments of the penis of anisopterous dragonflies

2 Williamson Kennedy Borror Fraser Present
5 (1920) (1922) (1942) (1940) author
<

Desmogomphus| Libellula Erythrodiplax | Gomphidae Onychogomphus

DRAGONFLIES
STUDIED

ardens
Vesicle Segment 1 | First or basal | Vesicle Vesicle
segment
Segment 1 Segment 2 | Second seg- Stem or first | Stem
ment joint
n | Segment 2 Segment 3 | Third seg-
= ment
Z Wee pee ale EAE ah be eels
a Fourth or Median or Median seg-
a terminal second ment
segment joint
Segment 3 (Elongation Glans or dis- | Distal segment
of distal tal joint
meatus of
segment
3

STERNITE 3 is modified anteriorly into a penis and posteriorly into
a long, narrow sclerite; the former is abutted to the truncated end of
the latter. Anteriorly the narrow sclerite bears a large, round, internal,
apodeme on each side.

The penis (fig. 44) is a complicated organ consisting of several
segments. Different designations for each of the segments of the penis
of anisopterous dragonflies have been proposed by various taxonom-
ists. A table is here given to show the different terminologies. Those
terms used by Fraser (1940) in his paper on the penes of a large
number of gomphine dragonflies are adopted in the present paper with
a few modifications.

The vEsIcLe (VS) is a robust structure, heavily sclerotized except
for its dorsal surface which is membranous, having a small sclerite
imbedded in its posterior portion. This sclerite is connected to the

No. 6 MORPHOLOGY OF THE DRAGONFLY—CHAO 45

posterior extensions of the posterior lamina by a pair of small scler-
ites. These slender sclerites apparently have not been recorded in
the literature. When viewed ventrally (fig. 40) the vesicle is slightly
constricted near the base, widened apically, slightly protruded on its
apical angles, and with a short median cleft. The cavity of the vesicle
is filled with fluid, and is continuous with those of the succeeding seg-
ments of the penis, but not with the haemocoele of the abdomen. The
latter fact is contrary to the opinion maintained by Kennedy (1922)
and probably also by Borror (1942).

The vesicle is an important organ in relation to the erection of the
penis, but the exact role that it plays remains obscure. Kennedy
(1922) mentioned that the cavity of the vesicle is continuous with the
haemocoele of the abdomen, and that the erection of the penis is
accomplished by forcing blood from the latter through the vesicle to
the cavity in the apical segments of the penis. This opinion is adopted
by Borror (1942) with modification. Fraser (1940) says that “when
pressure is raised in the vesicle . . . the penis” is “at once erected.”
But he does not mention how the pressure is raised in the vesicle.
From the present studies it appears that the raising of pressure in
the vesicle is accomplished by exerting a force on the small sclerite
imbedded in the dorsal membrane of the vesicle. This contention is
based on the observation that the cavity of the vesicle is not continuous
with the haemocoele and that the wall of the vesicle is rigid except for
the dorsal surface which is membranous. Thus a force exerted on the
small sclerite imbedded in the membrane will depress the latter and
force the fluid in the vesicle to flow into the apical segments so that
the penis is erected.

The stem (STEM) is L-shaped, attached to the dorsal membrane
at the anterior end of the vesicle. Apically it bears a large round mem-
brane which has a long narrow opening, the proximal meatus (PXM).
The presence of this proximal meatus probably indicates that the stem
is a composite segment, i.e., consisting of the second and the third
segments of Borror (1942) fused together. This contention is based
on the fact that (1) in Erythrodiplax and Libellula the proximal
meatus is always present in a small and distinct segment, i.e., the third
segment of Borror, and that (2) the apical limit of the preceding
segment is indicated by the presence of a short dorsal spur (Kennedy,
1922), or knoblike protuberance (Borror, 1942). In gomphine drag-
onflies this spur is absent in many species, such as shown by Fraser
(1940), but present in some other species, e.g., Gomphus agricola, G.
suzuki, Onychogomphus flexuosus, O. circularis, Megalogomphus
hannyngtom, Progomphus pygmaeus, Cyclophylla signata, and Stylo-

46 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

gomphus inglisit. The latter fact indicates the distal limit of a segment
proximal to the segment bearing the proximal meatus. The suture
between these two segments is generally obliterated in gomphine
dragonflies.

The MEDIAN SEGMENT (MS) is short, perpendicular to the basal
segment, and distally inflated. The inflated portion is weakly sclero-
tized and bears a large lobe, the posterior lobe (PLB) (Kennedy,
1922; Borror, 1942), also called preputial fold or prepuse (Fraser,
1940). Anteriorly the median segment is medially grooved. Dorsally
it bears a pair of very heavly sclerotized structures to which the bilobed
distal segment (DS) is attached.

The pISTAL SEGMENT (DS), or glans (Fraser, 1940), is bilobed.
Each lobe bears a curled flagellum (Fraser, 1940) or cornua (CN)
(Kennedy, 1922). The distal meatus is situated deeply between the
two lobes.

STERNITES 4 and 5 (see fig. 38, sternite V, 2) are elongate sclerites.
Each is differentiated into an anterior subquadrate area followed by
a long, narrow piece which is slightly constricted near the apex and
ending in a small piece, the sternellum (STN) (Tillyard, 1917).
Anteriorly the subquadrate area is produced into a pair of short proc-
esses. At the four angles of the subquadrate area, the sternite is pro-
duced internally and laterally into two pairs of sternal processes
(STP) for the attachment of the segmental muscles. The anterior
pair is small and linear and the posterior pair is fairly large and
scalelike.

STERNITE 6 (fig. 35) is similar to the preceding sternite except that
its sternellum is enlarged apically.

STERNITE 7 (fig. 35) is peculiar in that the anterior pair of sternal
processes is very small, and the long piece following the subquadrate
area is widened apically; without sternellum.

STERNITE 8 (fig. 46) is a large sclerite, subtrapezoidal in shape,
basally with a low median keel, and laterally slightly sinuate.

STERNITE 9 (fig. 46) is sclerotized on its basal half and very weakly
so on its apical half. The sclerotized portion is 4-lobed, two on each
side of a pair of median ovoid sclerites, the valvules (VV), which,
according to Tillyard (1917), are homologous with the lateral proc-
esses of the ovipositor of the female. The valvules cover the male
gemtal pore; the latter is guarded by a sclerotized ring.

STERNITE 10 (fig. 46) is fused with the corresponding tergite. Its
posterior margin is deeply emarginated.

MORPHOLOGY OF THE DRAGONFLY—CHAO

LATERAL VIEW,¢&

Fics. 45-50.—Terminal abdominal segments.

48 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

SPIRACLES

There are eight pairs of abdominal spiracles (SP). The first pair
is situated in the first sternite, described previously. The next six pairs
are situated in the pleural membranes near the posterior sternal proc-
esses in each of the abdominal segments 2 to 7. They are ovoid in shape
and oblique in position. The eighth pair is situated on the pleural
membrane close to the middle of the lateral margins of the eighth
sternite. It is almost twice as large as the other spiracles, elongate
ovoid, and parallel to the long axis of the body.

FEMALE TERGITES AND STERNITES

The tergites of the female are fundamentally the same as those of
the male, except for some sexual dimorphic characters mentioned be-
fore. TERGITE Io (fig. 48) differs from that of the male in that dorsally
it is not wrinkled nor tuberculate, and apically it forms a straight line
instead of being deeply concave.

Of the sternites, the first, fourth, fifth, and sixth are the same as
in the males. STERNITE 2 (fig. 36) differs from the more generalized
condition of the above in that it is rather wide, with the anterior
transverse area bearing a pair of lateral sternal processes only. STER-
NITE 3 (fig. 37) is similar to the generalized sternite except that the
anterior and lateral processes are comparatively longer. STERNITE 7
(figs. 39, 47) is comparatively broad, with very small sternellum.
STERNITE 8 (fig. 47) is large, elongate rectangular, laterally sinuate,
basally with a low median keel, subapically with a low protuberance,
and apically with a pair of subgenital plates (SGP) which are about
two-fifths as long as sternite 9. The two sclerites of the subgenital
plates are called valves (or vulvar scales, anterior processes) of the
ovipositor, probably synonymous with some other terms such as ven-
tral valves and first valvulae. STERNITE 9 (fig. 47) is broad, basally
emarginate and separated from the preceding sternite by a large semi-
circular membrane. The female genital pore is situated in this mem-
brane and is covered by the subgenital plates. At a point one-third the
distance from apex to base of the segment a low arc-shaped ridge is
produced, which is apically bordered by a narrow membrane; the latter
is constricted in the middle. This ridge might possibly be the remnant
of the median process of the ovipositor (Tillyard, 1917) (also called
inner valves or second valvulae). STERNITE Io is transversely rec-
tangular, apically not emarginated as in the male.

|
|
|

No. 6 MORPHOLOGY OF THE DRAGONFLY—CHAO 49

END SEGMENTS

The end segments, as used by Tillyard (1917), consist of various
structures apical to the segment 10. These structures differ morpho-
logically and phylogenetically in different suborders and in the two
sexes. The following table shows their homologies (modified from
Tillyard, 1917, p. 225).

TABLE 3.—Occurrence and homologies of terminal abdominal structures

Seg-

eae Name of parts Male Female
10 | Tergite - | eresent Present
Sternite Present Present
Cercoids Superior anal appendages| Anal appendages
11 Tergite Rudimentary Median dorsal appendage
Sternite Fairly large Fairly large
(bipartite)
Appendix dorsalis | Inferior anal appendages | Absent
Cerci Absent Absent
12 Tergite Rudimentary Rudimentary
Sternite (two Rudimentary Rudimentary

laminae anales)

In the male the superior and inferior anal appendages are very well
developed. The suPERIOR ANAL APPENDAGES (SAP) are elongate,
more than double the length of the segment Io, declined in their apical
halves, slightly sinuate in dorsal view, and ventrally serrate at the
apices. The INFERIOR ANAL APPENDAGES (IAP) is a bifid structure
with its two branches slightly longer than the superior anal append-
ages. Its base bends down vertically for a short distance and bears
two apposing arms which curve upward gently toward their rather
acute apices. Each branch of the inferior appendage has a minute
dorsal subapical tooth and a fairly large, low, internal ridgelike tu-
bercle just opposite the apex of the superior appendage; the tubercle
being adorned with fine hairs. The remnants of the 11TH STERNITE
(figs. 46, 50) are divided into two fairly large sclerites, collectively
called the bipartite 11th sternite, situated along the ventral and lateral
margins of the segment 10. The 12TH TERGITE and STERNITE are
represented by superior and inferior anal laminae. Both are weakly
sclerotized and adorned with minute hairs. The superior anal lamina
(SPL) is attached to the anterior surface of the base of the inferior
anal appendage. The inferior anal laminae (IFL) consist of two
pieces, one attached to each of the bipartite sternites. The anal opening
is situated at the bases of the anal laminae.

50 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

In the female the end segments are comparatively short, consisting
of anal appendages, median dorsal appendage, and the remnants of
the 11th and 12th segments. The ANAL APPENDAGES (AAP) are a
pair of slender conical structures, situated laterally above the dorsal
appendage, slightly longer than the roth tergite. The median dorsal
appendage, or supra-anal plate (SPP), is a subsemicircular sclerite,
convex above, about half as long as the anal appendages. The sB1-
PARTITE IITH STERNITE (fig. 47) consists of a pair of fairly large
sclerites which, when viewed ventrally, are triangular in shape. The
remnants of the 12th tergite and sternite are represented by superior
and inferior anal laminae. The former is weakly sclerotized, attached
to the ventral surface of the median dorsal appendage. The inferior
anal laminae (IFL) are bipartite, attached to the 11th sternite, and
exceed the length of the latter.

ABBREVIATIONS

Aa, tornus

AAP, anal appendage

AAR, anterior mandibular articulation

AAS, antealar sinus

AB, anterior lobe of prothorax

ACL, anteclypeus

ACS, antecostal suture

AES:, AESs:, mesothoracic or meta-
thoracic anepisternum

AH, anterior hamule

AL, anterior lamina

AN, adnotal sclerite

AP2, APs, apodeme of detached plate
of mesothoracic or meta-
thoracic scutum

APF, anterior portion of framework

ARB, anterior rib of anterior tentorial
arm

arc, arculus

ARG, antealar ridge

at, anal triangle

ATA, anterior tentorial arm

ATG, acrotergite

AU, auricle

AWP, anterior wing process

Ax:, Ax: primary antecostal cross
veins

1AX, first axillary sclerite

AXC, axillary cord

AXP, axillary plate

BA, basalare
BCD, basicardo
BCX, basicoxite
BCXS, basicostal suture
BH, brush

BPL, basal plate
br, bridge vein
br.v., brace vein
bs, basal space
BS, basisternum
BTA, basitarsus

C, costal vein

CAC, cleft of anterior lamina

CAR, posterior mandibular articula-
tion

CEP, cephaliger

CL, claw

CN, cornua

COL, collar

CT, corporotentorium

cu-a, cubito-anal cross vein

CuP, posterior cubitus

CX, coxa

CXC, coxal cavity

DC, dorsal cervical sclerite
DCD, disticardo

DCR, dorsal carina

DS, distal segment of penis

No. 6

DTA, dorsal tentorial arm
DTP, dorsal tentorial pit
DTTA, distitarsus

EC, eucervicale

EH, end hook

EMP, empodium
EPM, epimeron
EPX, epipharynx
ES, episternum
ESR, epistomal ridge
ESS, epistomal suture
EYE, compound eye

F, furca

FE, femur

FL, flagellum of antenna
FM, foramen magnum

FP, furcal pit

FR, frons

FS, furcasternum

FT, flexor tendon of mandible

G, gena
GF, genital fossa

h, hypertriangle

HM, hypostoma

HMS, hypostomal suture
HP, humeral plate

IAP, inferior anal appendage

ICS, incisors

IFL, inferior anal lamina

IL, inner lobe

IP, intermediary piece

IPLA, interpleural apodeme | (=inter-
segmental apodeme)

IR», IRs, intercalary radial veins

KEP, katepisternum

LL, lateral lobe
LR, labrum

M, media

MA, anterior median

mb, membranule

MB, median lobe of pronotum
MD, mandible

MDP, mandibular process
MH, movable hook

ML, middle lobe of labium

MORPHOLOGY OF THE DRAGONFLY—CHAO 51

MLS, midlateral suture (=interseg-
mental suture)

MN, mentum

MO, mola

mr, midrib of anal loop in hind wing

MRB, midrib of anterior tentorial arm

MS, median segment of penis

MSC, mesostigmatic lamina

MX, maxilla

MXP, maxillary process

N, nodus
NC, neural canal

0, oblique vein

OC, ocellus

OCC, occiput

OCCD, occipital condyle
OCCH, occipital horn
OCCM, occipital margin
OCS, ocular sclerite
OL, outer lobe

OS, ocular suture

P, pedicel

PA, papilla

PB, posterior lobe of prothorax

PC, postcervicale

PCL, postclypeus

PCX, postcoxale

PF, prefurca

PFS, parafrontal suture

PFST, postfurcasternum

PG, postgena

PH, posterior hamule

PL, posterior lamina

PLA, pleural apodeme

PLB, posterior lobe of penis

PLD, pseudolateral dilation

PLF, sternal fold

PLS, pleural suture

PM, pleurostoma

PMS, pleurostomal suture

POCC, postocciput

POCS, postoccipital suture

POS, postocellar suture

PRA, prealare

PRB, posterior rib of anterior tentorial
arm

PS, parastipes

PSA, peristigmatic apodeme

52 SMITHSONIAN MISCELLANEOUS COLLECTIONS

PSC, prescutum

PSCL, postscutellum

PSTN, pseudosternum

Pt, pterostigma

PTA, posterior tentorial arm
PTAR, pretarsus

PTC, posterior transverse carina
PTL, plantella

PTP, posterior tentorial pit
PWP, posterior wing process
PXM, proximal meatus

R, radius

ROOGC, rear of occiput

Rs, radial sector

RT, retractor tendon of mandible

s, subtriangle

S, scape

SA, subalare

SAG, subalar ridge
SAL, salivarium

SAP, superior anal appendage
Sc, subcosta

SCL, scutellum

SCS, sternocostal suture
SCT, scutum

SG, sutural groove

VOL. 122

SP, spiracle (ISP, IISP, mesothoracic
and metathoracic spira-
cles; SPs, SPs; etc, ab-
dominal spiracles)

SPD, spiracular dorsum

SPL, superior anal lamina

SPP, supra-anal plate

SQ, squame of labium

ST;, ST:, etc., abdominal sternites

IST, IIST, mesosternum, metasternum

STEM, stem of penis

STI, stipes

STN, sternellum

STP, abdominal sternal process

SUTC, supplementary transverse ca-
rina

SVC, submarginal ventral carina

t, triangle

T, tergite

TA, tarsus

TFR, top of frons

TI, tibia

TIC, tibial comb

TN, squame of precostal apodeme
TR, trochanter

UT, unguitractor

SGP, subgenital plate V, vertex

SHP, sheath of penis VC, ventral cervical sclerite

SM, submentum VS, vesicle

sn, subnodus VV, valvula
REFERENCES

BERLESE, A.

1909. Gli insetti: loro organizzazione, sviluppo, abitudini e rapporti con
l'uomo. I, Embriologia e Morfologia. x+1004 pp. Soc. Editrice

Libraria, Milan.
Borror, D. J.

1942. A revision of the libelluline genus Erythrodiplax (Odonata). Ohio
State Univ. Contr. Zool. Ent. No. 4, biol. ser., xv-+213 pp., 41 pls.

1945. A key to the New World genera of Libellulidae (Odonata). Ann.
Ent. Soc. Amer., vol. 38, No. 2, pp. 168-194, 72 figs.

Butter, H.

1904. The labium of the Odonata. Trans. Amer. Ent. Soc., vol. 30, pp.

III-134, 6 pls.
CALvERT, P. P.

1893. North American Odonata. Trans. Amer. Ent. Soc., vol. 20, pp. 153-

268, 3 pls.

No. 6 MORPHOLOGY OF THE DRAGONFLY—CHAO 53

Criark, H. W.

1940. The adult musculature of the anisopterous dragonfly thorax (Odonata,

Anisoptera). Journ. Morph., vol. 67, pp. 523-565, 7 figs.
Cow Ley, J.

1937. Tibial and femoral combs, and a trochanteral brush, in the Odonata.

Proc. Roy. Ent. Soc. London (A), vol. 12, pp. 123-125, 4 figs.
-Crampton, G. C.
1918. The phylogenetic study of the terga and wing bases in Embiids,
Plecoptera, Dermaptera, and Coleoptera. Psyche, vol. 25, pp. 4-12,
pier.
1923a. Preliminary notes on the terminology applied to the parts of an
insect’s leg. Can. Ent., vol. 55, pp. 126-132, pl. 3.

1923b. A phylogenetic comparison of the maxillae throughout the orders
of insects. Journ. New York Ent. Soc., vol. 31, pp. 77-106, pls.
12-17.

1924. The phylogeny and classification of insects. Journ. Ent. Zool., vol.
16, pp. 33-47.

1926. A comparison of the neck and prothoracic sclerites throughout the
orders of insects from the standpoints of phylogeny. Trans. Amer.
Ent. Soc., vol. 52, pp. 199-248, pls. 10-17.

1928a. The eulabium, mentum, submentum and gular region of insects.

Journ. Ent. Zool., vol. 20, pp. 1-18, 3 pls.
1928b. The basal structures of the wings of certain insects. Bull. Brooklyn
Ent. Soc., vol. 23, pp. 113-118, pl. 4.

1929. The terminal abdominal structures of female insects compared
throughout the orders from the standpoint of phylogeny. Journ.
New York Ent. Soc., vol. 37, pp. 453-496, pls. 9-16.

1932. A phylogenetic study of the head capsule in certain orthopteroid,
psocid, hemipteroid and holometabolous insects. Bull. Brooklyn
Ent. Soc., vol. 27, pp. 19-55, pls. 4-8.

1942. The external morphology of the Diptera. Guide to the insects of
Connecticut. Part VI. The Diptera or true flies of Connecticut.
Conn. Geol. Nat. Hist. Surv. Bull. No. 64, pp. 10-165, 14 pls.

CREMER, E.

1934. Anatomische, reizphysiologische und histologische untersuchungen
an der imaginalen und larvalen Flugmuskulatur der Odonaten.
Zool. Jahrb., vol. 54, pp. 191-223, 24 figs.

DuPorte, E. M.

1946. Observations on the morphology of the face in insects. Journ.

Morph., vol. 79, pp. 371-417, 7 pls.
Ferris, G. F.

1940. The morphology of Plega signata (Hagen) (Neuroptera: Manti-
spidae). Microentomology, vol. 5, pt. 2, pp. 33-56, figs. 6-20.

1942a. Some observations on the head of insects. Microentomology, vol. 7,

No. 2, pp. 25-62, figs. 10-27.
1942b. Some fundamental concepts in insect morphology. Microentomology,
vol. 8, No. I, pp. 2-7.

1943. The basic materials of the insect cranium. Microentomology, vol. 8,

No. 1, pp. 8-24, figs. 1-6,

54 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

1944. On certain evolutionary tendencies in the heads of insects. Micro-
entomology, vol. 9, pp. 78-84.

1947. The contradictions of the insect head. Microentomology, vol. 12, No.
3, Pp. 59-82, figs. 23-29.

Ferris, G. F., and PENNEBAKER, PHYLLIS.

1939. The morphology of Agulla adnixa (Hagen) (Neuroptera: Raphi-

diidae). Microentomology, vol. 4, pt. 5, pp. 121-142, figs. 59-71.
Ferris, G. F., and Ress, B. E.

1939. The morphology of Panorpa nuptialis Gerstaecker (Mecoptera:
Panorpidae). Microentomology, vol. 4, No. 3, pp. 79-108, figs.
36-51.

Forses, W. T. M.

1943. The origin of wings and venational types in insects. Amer. Midl.

Nat., vol. 29, No. 2, pp. 381-405.
Fraser, F. C.

1933. The fauna of British India, including Ceylon and Burma. Odonata,
vol. 1, xiii+423 pp., I map, 180 figs. London.

1934. The fauna of British India, including Ceylon and Burma. Odonata,
vol. Il. xxiii+398 pp., 4 pls., 120 figs. London.

1938. A note on the fallaciousness of the theory of pretracheation in the
venation of Odonata. Proc. Roy. Ent. Soc. London (A), vol. 13,
pp. 60-70, 3 figs.

1939. The evolution of the copulatory process in the order Odonata. Proc.
Roy. Ent. Soc. London (A), vol. 14, pp. 125-129, 1 fig.

1940. Comparative-study of penes of Gomphidae. Trans. Roy. Ent. Soc.
London, vol. 90, pp. 541-550, 1 fig., 6 pls.

1942. A note on the evolution of some venational structures in the dragon-
fly wing. Proc. Roy. Ent. Soc. London (A), vol. 17, pp. 64-60,
2 figs.

1944. The significance of vestigial oblique vein in the evolution of inter-
calated veins in the Odonata wing, with the description of a new
genus. Proc. Roy. Ent. Soc. London (B), vol. 13, pp. 58-67,
5 figs.

1948. A new interpretation of the course of the subcostal vein in the wings
of Odonata, with remarks on Zalessky’s notation. Proc. Roy. Ent.
Soc. London (A), vol. 23, pp. 44-50, 2 figs.

Garman, P.

1917. The Zygoptera or damsel flies of Illinois. Bull. Illinois State Lab.
Nat. Hist., vol. 12, pp. 412-587, 15 pls.

1927. The Odonata or dragonflies of Connecticut. Guide to the insects of
Connecticut. Part V. Conn. Geol. Nat. Hist. Surv. Bull. No. 30,
Pp. 1-331, 67 text figs., 22 pls.

Henry, L. M.

1948. The nervous system and the segmentation of the head in the

Annulata. Microentomology, vol. 13, No. 2, pp. 27-48, figs. 10-16.
Imos, A. D. f

1948. A general textbook of entomology. 7th ed. xii+757 pp., 624 figs.

London.
Kewnnepy, C. H.

1922. The morphology of the penis in the genus Libellula (Odonata).

Ent. News, vol. 33, pp. 33-40, 2 pls.

No. 6 MORPHOLOGY OF THE DRAGONFLY—CHAO 55

LAMEERE, A.
1922. Sur la nervation alaire des insectes. Bull. Acad. Sci. Roy. Belg.,
Bruxelles, vol. 5, No. 8, pp. 138-149. (English trans. in Psyche,
vol. 30, pp. 123-132, 1923.)
Lew, G.
1933. Head characters of Odonata. Ent. Amer., vol. 14, pp. 41-96, 12 pls.
Lucas, W. J.
1923. Labium (second maxillae) of the Paraneuroptera (Odonata). Proc.
S. London Ent. and Nat. Hist. Soc. 1922-23, pp. 57-63, 2 pls.
MARSHALL, W. S.
1914. On the anatomy of the dragonfly Libellula quadrimaculata Linn.
Trans. Wisconsin Acad. Sci., Arts, Lett., vol. 17, No. 2, pp. 775-
790, 4 pls.
Montcome_ry, B. E.
1940. Discussion of leg characters in Odonata. Lloydia, vol. 3, pp. 259-278.
MunscHeEnD, L.
1933. Die metamorphose der labiums der Odonaten. Zeitschr. Wiss. Zool.,
PP. 201-240, 44 figs.
NEEDHAM, J. G.
1903. A genealogic study of dragonfly wing venation. Proc. U. S. Nat.
Mus., vol. 26, pp. 703-764, pls. 31-54.
1930. A manual of the dragonflies of China. Zool. Sin., Ser. A, vol. 11,
No. I, pp. 1-344, 20 pls.
1951. Prodrome for a manual of the dragonflies of North America, with
extended comments on wing venation systems. Trans. Amer. Ent.
Soc., vol. 77, pp. 21-62, 7 figs., 3 pls.
NEEDHAM, J. G., and ANTHONY, MAupe H.
1903. The skewness of the thorax in Odonata. Journ. New York Ent. Soc.,
vol. II, pp. 117-124, 1 pl.
NEEDHAM, J. G., and Haywoon, H. B.
1929. A handbook of the dragonflies of North America. viii-+378 pp., many
figs. Springfield, Ill., and Baltimore, Md.

RAmBvr, J. P.
1842. Histoire naturelle des insectes. Neuroptéres ... xvii+534 pp., 12
pls. Paris.

St. Quentin, D.
1936. Der putzapparat der Odonaten. Zool. Anz., vol. 115, pp. 225-231,
2 figs.
Snoperass, R. E.
1909. The thorax of insects and the articulation of the wings. Proc. U. S.
Nat. Mus., vol. 36, pp. 511-595, pls. 40-60.
1927. Morphology and mechanism of the insect thorax. Smithsonian Misc.
Coll., vol. 80, No. 1, pp. 1-108, 44 figs.
1930. How insects fly. Ann. Rep. Smithsonian Inst., 1929, pp. 383-421,
25 figs.
1935. Principles of insect morphology. 667 pp., 319 figs. New York.
1947. The insect cranium and the “epicranial suture.” Smithsonian Misc.
Coll., vol. 107, No. 7, pp. 1-52, figs. 1-15.

56 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

Soka, R. R.

1947. The anatomy of the head capsule of the dragonfly Pantala flavescens
Fabricius (Anisoptera: Odonata). Musée Heude. Notes Ent.
Chinoise, vol. 11, fasc. 6, pp. 157-223, 5 pls.

STRICKLAND, E. H.

1946. Odonata as class-room material. Ann. Ent. Soc. Amer., vol. 30,

No. 1, pp. 28-32.
Tuompson, O. S.

1908. Appendages of the second abdominal segment of male dragonflies
(Odonata). New York State Educ. Dept. Mus. Bull., vol. 124,
PP. 249-263.

TILLYARD, R. J.

1917. The biology of dragonflies. xii+306 pp., 4 pls. Cambridge Univ.
Press.

1922. New researches upon the problem of the wing venation of Odonata.
Ent. News, vol. 33, pp. I-7, 45-51, I pl., 3 figs.

1926. Insects of Australia and New Zealand. xi+560 pp., 44 pls. Sydney.

TILLyarD, R. J., and Fraser, F. C.

1938. A reclassification of the order Odonata based on some new inter-
pretations of the venation of the dragonfly wing. Austr. Zool.,
vol. 9, pp. 125-160, 27 figs.

1939. Ibid., pp. 195-221, 11 figs.

1940. Ibid., pp. 359-396, 15 figs.

Wuepon, A. D.

1918. Comparative morphology and possible adaptations of the abdomen
in the Odonata. Trans. Amer. Ent. Soc., vol. 44, pp. 373-437, 9 pls.

1927. The structure and transformation of the labium of Anax junius.
Biol. Bull., vol. 53, No. 4, pp. 286-296, 2 pls.

1938. The aortic diverticula of the Odonata. Journ. Morph., vol. 63, No. 2,
Pp. 239-261, 6 pls.

WILLIAMSON, E. B.

1920. A new gomphine genus from British Guiana with a note on the
classification of the subfamily (order Odonata). Occ. Pap. Mus.
Zool., Univ. Michigan, No. 80, pp. 1-11, 1 pl.

GRY tee
As

“SMITHSONIAN MISCELLANEOUS COLLECTIONS
VOLUME 122, NUMBER 7

epates D. and #lary aux Walcott
Research Fund

x THE GEOLOGY OF CHACO CANYON
Po NEW-MEXICO

* .
IN RELATION TO THE LIFE AND REMAINS
| (OF Lee PREHISTORIC PEOPLES OF
- PUEBLO BONITO

(WitH 11 PLatEs) ©

a ey od, : BY GU:
eae KIRK BRYAN
Roa Bs | ; RP

| iy ah hd ih folie . ' ~eeeeeee

(Pustication 4140)

. ‘CITY OF WASHINGTON
We x PUBLISHED BY THE SMITHSONIAN INSTITUTION ds
_ FEBRUARY iy, 1954

anil Pol SUQAR }]
7 7B fALOATOSOI SJ[LIoyJeA, JO JUOUTyUeqUIO 9}

I aLV1g

He Le
ae ee

” x lay :

la ae fawfe. SD ae tag

et _ y ~ we 4 . ~4
vr. “ - -

SMITHSONIAN MISCELLANEOUS COLLECTIONS
VOLUME 122, NUMBER 7

Charles D. and Mary Waux CHalcott
Research F und

fine AG OLOGY OF CHACO’ CANYON,
NEW MEXICO

IN RELATION TO THE LIFE AND REMAINS
OR iiae PReristORIC: PEOPLES OF
PUEBLO BONITO

(Wit 11 PLates)

BY
KIRK BRYAN

(PusticaTion 4140)

CITY OF WASHINGTON
PUBLISHED BY THE SMITHSONIAN INSTITUTION
FEBRUARY 2, 1954

The Lord Baltimore Press

BALTIMORE, MD., U. 8. A.

FOREWORD

The geology of Chaco Canyon in relation to its prehistoric inhabit-
ants was a subject that greatly interested Kirk Bryan. Born and
schooled in New Mexico, he had seen hundreds of ruined Pueblo
villages, mostly abandoned before advent of the Spaniards in 1540,
and had given much thought to the reasons behind their desertion.
Warfare may have been one cause but it obviously was not the only
one,

A geologist with the United States Geological Survey and engaged
primarily in a study of groundwater resources of the Southwest, Dr.
Bryan seemed to us especially qualified to seek out the factors that had
invited, and then repelled, colonization of Chaco Canyon in the days
of Pueblo Bonito. He accepted with enthusiasm our invitation to un-
dertake this study but was able to devote only two brief vacation pe-
riods to field work, in the midsummers of 1924 and 1925. His observa-
tions in Chaco Canyon, admittedly incomplete, prompted like inquiries
in other valleys during the decade that followed.

In 1926 Dr. Bryan left the Geological Survey to accept a call from
Harvard University, and thereafter academic commitments and sum-
mers in the field allowed him but little leisure. In consequence, he
never found an opportunity to finish this report on his Chaco Canyon
researches. A first draft, dated March 1925, and written before his
second visit to the canyon, was repeatedly revised and expanded as
his continuing investigations annually provided new data. He appears
to have made no change in the text after 1940. For these several rea-
sons some sections of the report lack references to the more recent
literature.

Following Dr. Bryan’s untimely death in the summer of 1950, his
unfinished manuscript was forwarded to me by Mrs. Bryan. I have
undertaken to arrange its several parts in conformity with his original
table of contents and to eliminate repetitions of subject matter and
phraseology. The various stratigraphic columns Bryan examined and
the course he plotted for an arroyo more or less contemporary with
the decline of Pueblo Bonito are shown on the accompanying map of
Chaco Canyon. Stratigraphic sections 10 to 23 were studied in 1925,
but we have descriptions for numbers 15 and 17 only, and a third,
without number but adequately located in relation to the expedition’s
camp.

lil

iv FOREWORD

Test pit No. 3, about midway between camp and the west refuse
mound, was among those I had caused to be dug in 1922 in connec-
tion with an analysis of Chaco Canyon soils. When it was deepened
three years later at Bryan’s request and was found to penetrate the
buried channel he was then trying to isolate, a common impulse was
to extend the exploratory trench we had previously dug through the
west refuse mound and thus reveal the original surface between buried
channel and the old village dump. Pit No. 4, dug expressly for Dr.
Bryan, was so named because of its proximity to his section 4, where
the buried channel stood exposed near the southeast corner of Pueblo
del Arroyo. Thus test pits 3 and 4 and the extended west-mound
trench enabled Bryan to plot the course of that prehistoric arroyo as
it passed Pueblo Bonito, and led to his search for traces of it as far
east as Pueblo Wejegi. The extent of this ancient channel, together
with evidence of alternating periods of erosion and sedimentation,
formed the basis for Bryan’s growing conviction that a slight change
in climate was the most likely cause for disruption and dispersal of
the Chaco Canyon population in the early twelfth century. His con-
clusion is certain to exert a profound influence upon future interpre-
tation of past history in the Southwest.

I gladly acknowledge our obligation to Mrs. Kirk Bryan and to two
of Dr. Bryan’s former students, Dr. John T. Hack and Dr. Luna B.
Leopold, both of the United States Geological Survey, for their coop-
eration in the preparation of this report. Two members of my Pueblo
Bonito staff, O. C. Havens and Lynn C. Hammond, and several of
our Zufi workmen assisted Dr. Bryan in Chaco Canyon. The illustra-
tions are mostly from photographs by Mr. Havens.

It was originally intended that this paper appear as fourth in the
series reporting the results of the National Geographic Society’s
Pueblo Bonito Expeditions. But the series was discontinued after the
first number, “Dating Pueblo Bonito and Other Ruins of the South-
west,” by Dr. A. E. Douglass (1935). Early in 1953 the Society
made the present manuscript available to the Smithsonian Institution,
which proposed to publish it under the Charles D. and Mary Vaux
Walcott Research Fund.

The life and achievements of Dr. Kirk Bryan are briefly reviewed
by Frederick Johnson in American Antiquity, vol. 13, No. 3, p. 253,
January 1951.

NeiL M. Jupp.
Leader of the National Geographic
Society's Pueblo Bonito Expeditions.
Washington, D. C.
June 1953.

CONTENTS

Page
LORE WOLC ye chee aie ceca yore Hes IN Siete cRetame oun oh dl nehraye ctoubuchind bey abuiees ter steiatete ili
IEA PTCA ELON rate ter ake ow ese iacad custaveye esioheaiaiois ete oA MNED shevatele opstetoieny © relakerarets I
ATMO MEMEO DON Geico aisle w overaie ekohs wiarete opie avatay ohels llevan Shararslalavencicnate lay aierate 2
hy siockanhy sors Chacon Gatiyionkracvateyc seis oxsters ctor eletevers ere ole) atte iesaiel olekans/ovsaks 3
Generale ations: tiene ieccie oi uovsi bates ele) dele ssciekods ats chats arora wnuateieye 3
Climatic conditions affecting geologic processeS..........-.2.eeeeee 4
Excavation or Chacon Cany Oi. .veypcpeyeret atcusrotsietel ieieelel< ersisveeiedee) s.ietsyspale 8
Adluviationcol the Canyon MOOG saree ercte actos eke sislers sisiclsie Sine sieges ier eters II
Waterob necenti Stream Chem hit eye ek fesyare TN aksceys revel esi orevare) holereue as 4 stencbote 15
IP resentececolOsiciprOoCeSSeShisisa cis ss leatoiintels mis cleiclatevele Gis alain. diaiel shaele « citlebelprs 18
Wreathernmne rand erosionvon die Clittstiar stale seis levereere nelalelaie see «e's 18
Vitro eoy Ot leas esse. t 8) oes ava ciety eh nclciecct aber bore rapa vaiaatiate Cision ene aia Marth Shia, § Opals 21
PNA tayidlli aM Sees iye ssc re catelelayek s eecete les diove oe ctaters ho ceiabriees Cheat c wea eaaieeae cholate 22
PAILS RELESIO NL AE Ly eis Plale m iteic bey yee Ab cla! aCalel ga alk bide eluniore ateiove weblrar sidan 23
EOSTOMPIMat eva ht OY Orjas eins acoice sunt cleterevoreisy Ateuswh air ecorers ha ereisrey sisi sevens eiets 23
Mhesvalleysalllivitme ss mets /o tee cae rao ete oye le arco ae ieee Sate Slotiagin che 23
Material shobethne efile yi rcstetedtarcterc nireterenst oa enon om eee Merce as ott at eal et lee 23
Acrancementrot thematerial sie ves ccs eieters tect siete deeln's aemania ale Sete 25
PI COERNGTT EN ANIVID 2G crak mi teverc ce Aalst ns wile Vaalt Miele omeiars a Hades 25
Euman elses) imythenval ley itil lecterievaters ohaiceys ol <iiere orneveeels sisiaiaie ta suotoeteias 27
Complex charactermOrpthe: valleypetilltss cer crore ie «irate «ise orevaieretatere alee <raromrninte 28
Relation or human relies to! main walleyatillss 52 2 cv.c).6 21s so sielnieiso sale) ele 30
Mhewbumed channelvand its siemificances cso... ¢ css o+5 «is ule core eee sie ae nels 32
Buried channels similar to the post-Bonito channel on other streams...... 37
Geologic evidence on the means of livelihood of Chaco Canyon peoples.... 38
Pi OOUWatenecahtnini ow sep aka t Sal screvschsivcicra cisions arniole, ecoteie oie iar alsahd cals aka ata toate 39
Arrovomtormation and floodwater farming. ..cs ces «sels cmsiace ste e+ sce 45
Cause of alternate erosion and sedimentation: . 6.6.2.6 ssisseeccecenacess 47
Wetailed sections im the recent, allwavittns o.oiisscieedie'c,e see vse sales ole viene y's 51
Sechons tasche aan vale dlls. Sic bP aes bs en vs beet te ee esc ote oes 51
Sections in the buried, or post-Bonito Channel..................000. 58
SHELA TOOE West? Coa Serr eI AIALRNCAS Dis CHER TERA ICS COC SECS ELE OUR a 59
ES AGPRGS EERE uRN EOP nee ey Fa cais alia) Sato tavern tie torent: Gietane enlace afeiaidlonene ee ei sleraialweletos 62

LIST, OF ILEUSTRATIONS

PLATES

Zt, se Gena bonito from, the mOcth Clit 4s). scc.e sd sis soc cceiad eee meee Frontispiece
2: Upper: Pajada: Butte, trom) Pueblo Una Vidaid vaio sss cs den gete cee 10
Lower: Small ruin in a northern branch of Chaco Canyon............ 10
3. Left: Cleavage planes in north wall of Chaco Canyon................ 10
iight:> Puebio Lit fireplace: at SehHOm 5 ines... signe atiaesvwsees eenivla ves 10

LIST OF ILLUSTRATIONS

EESGavadavVViasit.we tres a hc tra are eee mre ia kan cane a Rem 2 ae

Lower :

5. Upper:

Lower:

6. Upper:

Lower:

7. Upper:

Lower :

8. Upper:

Lower:

9. Upper:

Lower:

10. Upper:
Lower:

11. Left: Pit house below nearby Pueblo III ruin
Right: Pueblo I pot in sand dune

“Stonelace’\ on, detached! block oki... saan avcie sees po meie
Looking down Chaco Canyon from Pueblo del Arroyo
Recently fallen. arroyo bail. '. 5 Ssiv.te opiyeis era be eis oxosverla esate
Small-house ruin, destroyed by floods........5....eeccececees
Pueblo: Tpit) Rouse list). 2. os aes os bale 2 be aise Ghee eee
Buried channel at section 16
Busiedichannelvati section: Ace a.ceiee oem seen ae ee
Mouth \of Rincon del. Caminoiwsnwsss% 1.4% silos ode coe eee
Lateral erosion caused by drainage from The Gap............
Skeleton 6 feet below surface... .io0i fon. fo vei conde dca ese bee
Hearth at depth of 12 feet 8 inches
Horses crossing thes@hacosinifood4..> cess ee seek week os oe
Middle south wall of Pueblo del Arroyo

eee eee ee

ey
ar

Ce ee

FIGURES

1. Map of Chaco Canyon showing present and post-Bonito channels......
2 Northwest portion onuNews Mexicosmcm. se embinace seen ee ee eeie
Zu Banl< or Chaco between Sections: sandy lO etraer ies sake eee eee ier

Page
4. Upper: Niches and dunes, north wall of Chaco Canyon, near mouth of

10
Io

Charles D. and Mary Waux Walcott Research Fund

TH, GEOLOGY OB; CHACO. CANYON
NEW MEXICO

IN RELATION TO THE LIFE AND REMAINS OF THE
PREHISTORIC PEOPLES OF PUEBLO BONETO

By KIRK BRYAN?

(WirTa 11 PLaATEs)

INTRODUCTION

On the initiative of Neil M. Judd, leader of the National Geographic
Society’s Pueblo Bonito Expeditions, and on the recommendation of
Dr. John C. Merriam, then president of the Carnegie Institution of
Washington, the present writer was selected to undertake an inquiry
into the geologic history of Chaco Canyon. Two brief periods were
devoted to field work: July 28 to August 9, 1924, and July 10 to
August 1, 1925. In the well-ordered camp of the expedition he was
received with gracious hospitality, and to all members of the staff he
owes much in kindness. Mr. Judd placed every facility at his dis-
posal including a number of excavations especially designed to bring
to light geologic facts and thus expedite the investigation.

Application of the stratigraphic methods of geology to archeologi-
cal problems is no longer new, and knowledge of these methods forms
a part of the equipment of every modern archeologist. Our inquiry
into Chaco Canyon geology has proved (1) that the alluvial deposits
of the canyon carry various relics of prehistoric peoples and (2) that
the deposits can be separated into divisions of differing age. In recent
years knowledge of these generalizations has become widespread and
additional data have been gathered. It appears that we are now on the
brink of establishing in the Southwest an alluvial chronology based on
a sequence of episodes of erosion and alluviation. This sequence of
geologic events gives a key to the fluctuations of climate of late geo-
logic time and yields a proximate cause for the sudden decay of the
great Pueblo communities of the San Juan country. (Bryan, 1941.)

1 Dr. Bryan died on August 23, 1950.

SMITHSONIAN MISCELLANEOUS COLLECTIONS, VOL. 122, NO. 7

SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

Previous work on the general geology of this region is referred to
hereinafter. During the summers of 1899 and Igor, Prof. Richard E.
Dodge made a geological survey of Chaco Canyon as part of the ex-
tensive plans for the Hyde Exploring Expedition. His work was done
after archeological excavation had ceased, and, unfortunately, his re-
sults were published only in skeleton outline in the report of the ex-
pedition (Pepper, 1920, pp. 23-25) and in three abstracts (Dodge,
1902a, 1902b, 1910). Even so, these brief sketches record a number
of observations of interest that are referred to in the following pages.
They indicate that Professor Dodge was on the verge of discovery
and, with more archeological help, the geological theory herein set
forth would doubtless have been advanced by him 20 years earlier. The
1877 observations of W. H. Jackson (1878) were keen and penetrat-
ing, and from exposures no longer visible he made the original dis-
covery of the buried channel whose description and interpretation
form such a large part of this report.

The long delay between initiation of this study and its publication
has not been without advantage. During the interval we have learned
that the geologic history of Chaco Canyon is not unique. Other valleys
have similar histories, as will appear from the data on these other val-
leys summarized hereinafter. Generalizations on the cause of the
alternations from erosion to alluviation and on the effect of these
events on human affairs now rest upon a foundation of fact much
larger than would have been possible in 1924 and 1925.

PLAN OF "THE REPORT

This Chaco Canyon study was begun as an isolated project. It was
an attempt to relate recent geology to the life of prehistoric peoples
in the area. The results proved so successful, however, that other
studies were subsequently undertaken. The alternate periods of allu-
viation and erosion discovered in Chaco Canyon and related to the
tree-ring dates of Douglass (1935) have been found in other local-
ities. The periods of alluviation are, so far as evidence now exists,
nearly synchronous over the whole Southwest. Thus there has been
developed an alluvial chronology still imperfect but valuable as a meas-
ure of time in the dating of archeological events. It is presumably still
more valuable as a measure of alternating periods favorable or un-
favorable to floodwater farming, an important method of agriculture
in the area. Still more important are the inferences on fluctuations in
climate parallel with alternations in the regime of streams.

The report begins with a general consideration of the area and its

=
iz
-
_
a
j ¥
eae

he TENET

hone

— wo erwin

VERON A aad he i

pes eaten e ats cy

a i canny al ial a eee Wren |
- <= Se o

oie ~~

eee S

= & a

: * Plie _ Ray! sy ; ;
jen.: op an
7 7 a

—— WH he

CENTRAL PORTION
CHACO CANY' ON NATIONAL MONUMENT

HOWING THE LOCATION OF
PUEBLO BONITO

STE AE \ / NN iS<\ We. \ Ye 7 ‘PUEBLO BONITO EXPEDITIONS
GNE ; & r A Nh - { ak OF THE _ >
\ e “NATIONAL GEOGRAPHIC SOCIETY

‘AN Topography by Robert P Anderson
=\4 1922

—s () e YZ (SM ANOS SRS = We us
TT oe YA . BY) | 2 WEL \ UY TUNG a a\: Za , \ FROM SURVEYS BY THE

t 4 2
a

Contour Interval 20 feet
» Navaho Hogans
o Ancient Ruins

DRAWN BY JAMES M DARLEY

Fic. 1.—The buried, or post-Bonito, channel in relation to the present arroyo in Chaco Canyon.

NOW 7, GEOLOGY OF CHACO CANYON—BRYAN 3

climate, with such information as is available on the age of the present
arroyo in Chaco Canyon, with rather detailed studies of geologic proc-
esses now current there, and a description of the alluvium of the
valley floor. It then presents evidence that this alluvium is divisible
into three parts: the terrace, the main valley fill, and the post-Bonito
channel. The antiquity of these divisions and their correlation with
similar alluvial formations elsewhere are also considered. The im-
portance of floodwater farming in the Southwest and the effect of the
recent epicycle of erosion on this type of agriculture are next set forth.

The cause of the alternation from alluviation to erosion in south-
western valleys is next discussed and the argument advanced that
simultaneous alternations in the regimes of widely separated streams
must be due to synchronous climatic changes. The concurrent effects
of climatic change and change in stream regime throughout the known
human history of the Southwest affords a clue to fluctuations in human
culture otherwise unattainable.

PHYSIOGRAPHY OF CHACO CANYON
GENERAL RELATIONS

Chaco Canyon lies in northwestern New Mexico on the upper
reaches of Chaco River, a tributary of San Juan River (fig. 1). Chaco
River, about 100 miles long, is an ephemeral stream such as is char-
acteristic of arid regions. Its sandy bed throughout the greater part of
the year is dry and the stream is dignified by the name of river only
because of its considerable length and the violence of its floods. The
stream begins in the high plains country north of Chacra Mesa at an
altitude of 6,900 feet and flows a little north of west for 68 miles.
Here the course changes sharply to the north and the river flows nearly
parallel to, and on the east side of, the ridge known as the Grand Hog-
back for 26 miles and thence, breaking through the Hogback in a
narrow canyon, it reaches San Juan River in 7 miles. The total length
of the stream is thus about 100 miles, of which, however, only 15 or
20 miles of the upper course lies in a canyon worthy of the name.
About 12 miles of this canyon, the portion with which we are con-
cerned, is shown on the accompanying map (fig. 1).

Chaco Canyon lies in the southwestern part of the great Plateau
province which occupies northwestern New Mexico, northern Arizona,
western Colorado, and eastern Utah. The province is noted for its
extensive flat surfaces, long lines of cliffs, and deep canyons. The
flat surfaces are in part developed on the more resistant beds of nearly
horizontal sedimentary rocks, although in part they consist of large

4 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

outflows of lava, and in part they are the remnants of extensive plains
of erosion. In northwestern New Mexico the largest unit of the Pla-
teau province is the San Juan Basin, a vast area in which the rocks
dip gently from the periphery toward the center. Chaco Canyon lies
near the southern part of this area with a dip of about 2° to the north
and east.

Sandstone and shale are the characteristic rocks. The shale is eroded
into broad, flat surfaces or gently sloping valleys ; the sandstone stands
out as ridges or plateaus, bounded, especially on the south, by cliffs.
The order and succession of these rocks have been studied by a num-
ber of geologists * interested primarily in the occurrence of coal or of
vertebrate fossils.

Chaco Canyon is cut in the Cliff House sandstone, the upper mem-
ber of the Mesaverde group. This sandstone member is 369 feet thick
as measured by Reeside on Meyers Creek, a few miles northwest of
Pueblo Bonito. It is underlain by dark shale containing thin sandstone
and coal (Menefee formation) which crops out in the cliffs on the
south side of Chaco Canyon and in a few places on the north side. The
Mesaverde group is overlain by the Lewis shale which forms the plain
north of Pueblo Alto and has a thickness of about 70 feet. Above the
Lewis shale lie the Pictured Cliff sandstone and higher formations.

The Cliff House sandstone consists of two massive sandstones sepa-
rated by relatively thin bedded sandstone. Consequently, weathering
tends to produce two cliffs separated by a bench of gentler slope. The
lower of these two massive sandstones is buff-colored and about 140
feet thick. The cliffs which make the northern wall of Chaco Canyon
are carved from this rock by processes considered more in detail on
pages 18-20.

CLIMATIC CONDITIONS AFFECTING GEOLOGIC PROCESSES

The climate of the Chaco country is arid, but such a simple state-
ment does not adequately summarize the effect of climate on the geo-
logic processes. Aridity has many gradations from the almost total
lack of rainfall characteristic of parts of the Libyan desert of Africa,
and of certain areas on the west coast of Peru, to the tempered aridity
of California where trees and grass thrive in areas having relatively
low rainfall. Aridity is thus an inclusive term embracing climates
having varying amounts of precipitation up to a quantity fixed arbi-

2 Holmes, 1877; Endlich, 1877; Schrader, 1906; Shaler, 1907; Gardner, 1909;
Sinclair and Granger, 1914; Matthew, 1897; Brown, 1910; Bauer, 1917; Bauer
and Reeside, 1921; Reeside, 1924.

NO. 7 GEOLOGY OF CHACO CANYON—BRYAN 5

trarily around 20 inches of rainfall a year. The many shadings and
gradations of aridity are dependent on such factors as the propor-
tion of the precipitation that may occur as rain or as snow, on the dis-
tribution of precipitation throughout the year, and on the incidence
of rainfall whether in hard showers or gentle drizzles. Similarly the
daily or seasonal range of temperature and the extremes of heat and
cold with their incidence and duration are all factors in aridity.

Climatic elements directly affect various subprocesses involved in
the weathering of rocks and indirectly influence the nature of streams
which act as the agents of removal and of transportation of weathered
rock. Slight differences in degree of aridity often have marked in-
fluence in the growth of a vegetative cover, one of the greatest single
factors influencing and delimiting erosive and sedimentary processes.
In the account of these processes given hereinafter it will be seen that
the scant vegetation of an arid region is a necessary prerequisite to the
relative intensity of action, or even the existence, of many of the sub-
processes. It follows, therefore, that any past or anticipated climatic
change, provided it is sufficient to alter the existing vegetation, may
have relatively large effect on geologic processes.

The available rainfall records of the Navaho Country up to the end
of 1913 were collected by Gregory (1916, pp. 51-59) and the factors
of climate in Chaco Canyon are now being recorded by the National
Park Service. Herein only such general elements of climate are de-
scribed as seem necessary for the purpose of defining climate in respect
to geologic processes.

The climate of the Plateau province may be considered moderately
arid. On the higher portion, between the valleys of San Juan River
and the Little Colorado, there is greater precipitation than in the low-
lands. In the mountains doubtless as much as 20 inches may fall each
year, but current rainfall stations are all on lower ground. St. Mi-
chaels, Ariz., altitude about 6,950 feet, has a mean of 13.72 inches
based on records for 29 years out of a period of 68 years; Crown-
point, altitude 6,800 feet, has 10.93 inches, based on an incomplete
record extending over 11 years. At lower elevations, especially to the
north and south of Chaco Canyon, the precipitation is less. Holbrook,
Ariz., altitude 5,069 feet, has 9.38 inches with 25 years of record out
of a total of 33 years. Places in the San Juan Valley have a lower
rainfall: Fruitland, N. Mex., altitude 4,800 feet, 6.38 inches with 7
years of record; Farmington, N. Mex., altitude 5,220 feet, 9.23 inches
with 7 years of record; Aneth, Utah, altitude about 4,700 feet, 4.96
inches. It seems likely the Chaco Canyon district has a precipitation
similar to that at Crownpoint with a little less rainfall on the floor of

6 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

the canyon which is 300 to 400 feet lower than the adjacent cliffs. For
the purpose of this study it will be assumed that Chaco Canyon has a
mean of about 10 inches.

A large part of this precipitation falls during the so-called summer
rainy season in July and August. This period is characterized by sharp
local or general rains from cumulus clouds or thunderheads. The rate
of rainfall is high but the storms seldom last long. The incidence of
the rains is also variable in time and space. Small areas are deluged
and adjacent areas are left dry. The rains may come as early as June
or as late as September, or may be inconsiderable in amount for a
whole summer.

Gregory (1916, p. 63) summarizes many observations as follows:

The area covered by a shower is frequently only a few square miles, and
on two occasions showers of 20 to 30 minutes’ duration resulted in wetting less
than 300 acres. Many of the showers result in a heavy downpour, and the total
precipitation for a month is not infrequently the result of a single shower... .
Generally the intense heat preceding a shower is reestablished within an hour
or two after rain has ceased, especially at elevations below 6,000 feet. . . . Light-
ning is the almost invariable accompaniment of summer showers and constitutes
a real danger to travel... . My records of thunderstorms for the Navaho
Reservation during the field seasons 1909, I910, 1911, and 1913 are 38, 26, 33,
and 23, respectively, and it is believed that the annual number exceeds 40.

The winter precipitation falls gently and is likely to be widely spaced
in time, but on the average totals nearly as much as the summer rain-
fall. At elevations above 6,000 feet there are 17 to 25 inches of snow,
and at lower elevations some snow is possible each winter.

The distribution of precipitation throughout the year and its effect
on agriculture is best expressed in the following table compiled by
Gregory and amplified in a quotation also from him (ibid., pp. 61-62) :

Precipitation
in percent

of mean

Season Months rainfall
STMIM|EHA Vateitatene chre.s trrereters July, August, “September: 2... 2.0 aparece oe 37
Early witteg rae ses cena October, November, December.............. 25
Water wintel, £2) shai, s\s,<i0/s) 0% January, Bebruaty, Marchi cgi. (sic 0ecleses 26
Spano nes ccmeniecicce eiests Ancd Maat pment. alysiee ce tixts els coc aero 12

It will be noted that the season of least rainfall, April to June, is the growing
season for most crops, and that therefore the seasonal distribution of rain is
unfavorable for agriculture or for the vigorous reproduction of many grasses.
Half an inch of rain per month for the period April, May, and June is an
unusually large precipitation for most parts of the reservation, and during many
years the combined precipitation of these three months is less than one-half inch.
Moreover, plants obtained only a portion of this meager supply, for evaporation

NOS 7 GEOLOGY OF CHACO CANYON—BRYAN Vi

is most effective during the clear, dry, hot days of early summer. The moisture
in the ground, supplied by the rains of winter supplemented by the scattered
showers of spring, is sufficient to allow seeds to germinate and to send their
stalks above ground, but is insufficient to bring a crop to maturity. The rainfall
of July becomes therefore the critical factor in the life of the Navaho. If his
prayers to the rain gods are answered his corn crop is assured, and grass springs
up from the desert floors; if his prayer is denied the crop is a failure. . . . For
a large part of the reservation corn, without irrigation, fails to mature every
second or fourth year.

The variation in rainfall from year to year is of the greatest im-
portance. The amount ranges between half the normal and twice the
normal. For the 29 years of record at Fort Defiance and St. Michaels
the year of greatest rainfall was 1854 with 22.44 inches; the year of
lowest rainfall was 1900 with 6.52 inches. It is obvious that in years
of severe drought like 1900 almost nothing grows. Such years are pe-
riods of starvation for a population dependent on agriculture or on
the pasturage of animals.

Similarly the native vegetation must be able to resist these extremes
of drought and precipitation. In general, sagebrush and scattering
grass grow in the dryer areas, and perennial grasses where precipita-
tion is more generous. With a slight additional increase of rainfall,
cedar (juniper) forms sparse groves and a total precipitation of 15
to 20 inches is adequate for the open pine forests of mountain areas.
These vegetative zones are, however, not strictly bounded by lines of
equal rainfall because slope, exposure, and soil are all factors in the
growth of plants. Near Chaco Canyon the flat parts of the plateau are
generally underlain by clayey soils derived from shale or by loams
formed by the admixture of sand from the sandstone beds with clay
from the shale areas. These soils, under the influence of the local
climate, support a fairly continuous cover of perennial grasses. The
outcrops of sandstone have a rough and broken topography without
soil or with only a thin sandy soil. Here grow scattered cedars, oc-
casional woody bushes, and patches of “sand grass” but large portions
of such areas are bare rock. The floor of Chaco Canyon supports a
growth of greasewood (chico) with, in areas overflowed by storm
water, a fair growth of perennial grass. A few cottonwood trees have
survived from the period when the stream bed was shallow and are
evidence that, with a slightly higher water table or less interference
by man, domestic animals, and floods, many of these trees would again
grow in the valley.

The temperatures of the region are, when expressed in yearly or
monthly means, those of a temperate region. Yearly means range
from 47.6° F. to 60.6° according to the altitude of the station. The

8 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

annual and daily ranges in temperature are, however, very great. The
maximum range recorded for various stations in the region is as fol-
lows: Fort Defiance-St. Michaels, 122° (98° to —24°); Fruitland,
124° (110° to —14°) ; Holbrook, 127° (106° to —21°) ; Crownpoint,
103° (98° to —5°). Temperatures exceeding 100° normally occur
for 10 to 20 days each summer and 5 to 6 days of below-zero weather
are likely each winter. The daily ranges in temperatures may amount
to as much as 40° to 50° and, although doubtless effective in produc-
ing the disruption of rocks, are somewhat mitigated in their effect on
man and beast by the low humidity of the air.

The growing season, or number of days between the last killing
frost of spring and the first killing frost of autumn, ranges at various
stations from 89 days to 161 days. In general, localities of lowest al-
titude have the longest growing season but there is at all stations a
variability from year to year in the length of the growing season that
may be shorter than the mean by as much as a month. Fort Defiance,
at an altitude of 7,000 feet, has experienced killing frost in every
month of the year except August. Obviously these variations in the
length of the growing season add an additional hazard to agriculture
in a region where rainfall is scant and also highly variable in incidence
and amount. The data also give an index of the probability of changes
in temperature that cross the frost line and these changes are the ones
effective in the disruption of rock by frost action.

EXCAVATION OF CHACO CANYON

One who climbs the north wall of Chaco Canyon to Pueblo Alto is
rewarded by magnificent views of a region that appears to be flat on
all sides. To the south, beyond the canyon, he sees a vast plain from
which rise a few low hills and, far to the southwest, high mesas that
close in the horizon south of Crownpoint. To the north, the valley of
Escavada Wash is a prominent feature bordered by ragged bluffs, but
beyond lies a plain similar to the one on which he stands. This high
level plain occurs generally on the more elevated parts of the San Juan
Basin and is more or less independent of the hardness of the under-
lying rock. Canyons divide this plain into several parts that are ob-
viously remnants of a once continuous erosion surface that formerly
extended over the entire region. The plain has been too little studied
to warrant strict definition or to hazard correlation with the Mojave
peneplain which Robinson (1907) believes to have existed over the
whole of northwestern New Mexico and northeastern Arizona.

Bryan and McCann (1936) imply that this surface is older than the

NO. 7 GEOLOGY OF CHACO CANYON—BRYAN 9

Ortiz surface and other surfaces which, in the drainage of the Rio
Puerco (of the East), are graded to the Rio Grande.

From the evidence near Chaco Canyon it seems possible to postu-
late two or more erosion cycles during each of which the region was
reduced to very low relief. Whether a single peneplain or a more com-

I
Mummy Cave

_—

eAlbuquerque

STATUTE MILES

Fic. 2.—Northwestern New Mexico showing the location of Chaco Canyon
and Pueblo Bonito.

plex series of erosion cycles will be demonstrated by further work in
the area, Chaco River and the adjacent streams gained their courses
in a region of such moderate relief that the direction of flow was more
or less independent of the distribution of hard and soft rocks. After
a general uplift of the Plateau country, the “Canyon Cycle of Ero-
sion” was initiated and the great canyons of the Colorado River
system were cut.

10 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

Chaco River, a distant and rather feeble tributary of the Colorado,
also lowered its bed. In some places it excavated canyons and in
others fairly broad valleys. That its canyon cutting was not continu-
ous is evidenced by a well-marked erosional terrace near the mouth
of Escavada Wash, a terrace capped by gravel largely derived from
the local rocks and lying at an elevation about 150 feet above Chaco
River. How important or general this pause may have been awaits
field work over a larger area.

The general course of the river appears to have been controlled by
undiscovered factors on the ancient plain already mentioned, but de-
tails of the carving of the rocks within Chaco Canyon, as we see it,
result from the interaction of forces of erosion normal to the climate
and the structure of the rocks.

The most notable feature of the canyon is its asymmetry. The north,
or rather northeast, wall is steep and but little indented; the south or
southwest wall is gentler and broken by branching canyons, Asym-
metry is not an uncommon feature of valleys and canyons in New
Mexico that have an east-west trend. For example, the relatively
smooth, boulder-strewn slope of the south wall of Canyon del Rito de
los Frijoles, near Santa Fe, contrasts strongly with the sheer cliffs of
its north wall, in which Indians carved caves for occupancy in pre-
Spanish times. Yet this canyon, cut in lava and tuffs having a slight
dip downstream, is essentially alike in the two walls. The south side,
however, is shaded for much of the day, a condition that leads to
lower evaporation both of rain and snow, and consequently plants
thrive. Small bodies of soil are held in place by grass and bushes;
chemical erosion is promoted; talus heaps become overgrown with
trees and mantle the rock slopes. In contrast, the north wall with its
slope exposed to the sun is relatively dry. The mechanical forces of
erosion are in full swing here and debris once loosened from the wall
falls clear from rock surfaces which are thereby again exposed to the
weather.

Chaco Canyon also has an almost east-west course and is subject to
the same influences. A more important factor in creating a difference
in the slopes of the canyon walls is, however, a northeasterly dip of the
rocks. This inclination varies from I to 2 degrees with the result that
the base of the Cliff House sandstone lies at or below the floor of the
canyon on the north side, whereas it is from 50 to 100 feet above on
the opposite side. In consequence the south cliff is undermined with
relative ease by sapping of the underlying soft sandstone and shale.
The fall of blocks is also assisted by a slight inclination of the bedding
planes. Consequently, numerous and relatively large branch canyons

PLATE 2

Upper: Fajada Butte from Pueblo Una Vida, with the present arroyo dimly seen beyond

the ruin, and, at the right, the treeless plateau extending southward toward Crownpoint. (Photo-
graph by Neil M. Judd, 1920.)

Lower: A small ruin in a northern branch of Chaco Canyon between Una Vida and Wejegi.
Seepage has deposited an incrustation of gypsum along the rear wall of the cave.

(Photograph
by Neil M. Judd, 1926.)

Pe ian. iheins

(¥Z0I SusAeTT DO
Aq sydeisojoyd )

“o0e FANS jussaid 9y}
MO]Aq Joos § deTdo1y

LIT ol gond Pepts-qe[s

sty}? sem SG UOl}DeS
JB aIn}ea} VW UST

“UOA
-uey ooryy) jo [eM
Y}J1OU 94} UT aUOJSpues
asnoyH YYD JMO]
oy} jo ourjd o5e
-ABII OT}STIoJOeIeYO
ayy STR I aysiu
SITY} ‘JoJBM pue pUIM
Aq AICS) ato a]

€ ALVIg

PLATE 4

Upper: Near the mouth of the Escavada Wash the lower Cliff House sandstone in the north
wall of Chaco Canyon has been scoured and blasted by wind-driven sand. Dunes have blocked
the old road to Farmington.

Lower: Rainwater percolating through sandstone often results in a type of weathering called
“stonelace.” On the rock in the background water issuing from holes has left vertical streaks.

(Photographs by O. C. Havens, 1924.)

“ah
eS

afte.

PLATE

Upper: Looking down Chaco Canyon from Pueblo del Arroyo. The irregular mass of
Penasco Blanco is seen on the horizon at left center. At the right a sunlit cliff in the middle
distance marks the mouth of Rincon del Camino; between it and the standing figure are the
broken walls of Ruin No. 8.

Lower: A newly fallen section of bank immediately west of Pueblo del Arroyo.

(Photographs by O. C. Havens, 1925.)

NO. 7 GEOLOGY OF CHACO CANYON—BRYAN II

have developed. In the vicinity of Fajada Butte a large tributary
drainage has completely destroyed the south canyon wall and enters
through a valley broader than that of Chaco River (pl. 2, upper).

On the north side of the Chaco the base of the Cliff House sand-
stone lies below the canyon floor from Escavada Wash to Mocking-
bird Canyon. This part of the escarpment is characterized by having
a sheer cliff surmounted by a bench and a more gentle cliff above. Its
tributary canyons are generally less than half a mile in length and
many are mere indentations in the cliff, of the type commonly called
“rincons.” From Mockingbird Canyon upstream the base of the sand-
stone lies above the valley floor and the lower cliff is benched rather
than sheer, the tributaries are longer and the aspect of the canyon wall
is more like that of the southwest wall.

Excavation of the canyon is a process long since interrupted, for
the main stream nowhere runs on rock today, nor is it cutting laterally
against the walls of the canyon. The process of canyon cutting was
succeeded by a period of alluviation resulting in deposition of a valley
fill to the level of the present valley floor (see below). Filling of the
canyon has also been interrupted by formation of the present arroyo

(p. 35).
ALLUVIATION OF THE CANYON FLOOR

After cutting Chaco Canyon to a depth somewhat greater than at
present, the stream changed its habit and began to deposit more mate-
rial than it removed. The gradual character of this filling and details
of the process are here recounted at some length. This change from
erosion to sedimentation was not confined to Chaco Canyon. Other
canyons of the Plateau province and other streams throughout the
Southwest were also filled and alluviation was the characteristic proc-
ess up to a time within the memory of man. The isolation of Chaco
Canyon has prevented the accumulation of definite historical data on
characteristics of the canyon during this recent period of alluviation.
However, canyons of adjacent parts of the Plateau province furnish
reliable and analogous data since they lie at similar elevation, and are
cut in like rocks under similar conditions of climate and settlement.

During the surveys of Powell and Dutton in 1878-1880, the canyons
were undergoing alluviation as attested by the following statement
(Dutton, 1882, pp. 228-229) :

Most of these lateral canyons ... are slowly filling up with alluvium at the

present time, but very plainly they were much deeper at no remote epoch in the
past. The lower talus in some of them is completely buried and the alluvium

I2 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

mounts up on the breasts of the perpendicular scarps. In some cases a smooth
floor of alluvium extends from side to side of what was originally a canyon
valley.

Such conditions no longer exist and had, even in Dutton’s time,
ceased in certain parts of the San Juan drainage. At present every
main canyon in the area is occupied by an arroyo with vertical banks
from 10 to 100 feet high. The streams now run at a level lower than
the flat floors Dutton described by an amount equal to the height of
their banks. These arroyos began at or near the lower end of canyons
and progressed headward by a receding fall. The upper (or falls)
portion of an arroyo is ordinarily marked by a vertical bank or a chaos
of jagged, vertical banks and great blocks of detached alluvium. In
many canyons the head of the arroyo has not yet completed its advance
and, generally, only in the lower and larger tributaries have branch
arroyos been formed.

In the undissected parts of canyons and in minor tributaries there
can still be seen at work the processes by which their flood plains were
built up during the period of alluviation that has so recently been
brought to a close. The flat floor of such a canyon slopes downstream
at a grade dependent on the ratio of the volume of water in floods to
the supply of sediment carried. In general there is no well-marked
central channel, but numerous small discontinuous and branching chan-
nels mark the central part of the canyon floor. Low alluvial fans con-
sisting of sediment carried in from minor tributaries diversify the
plain. Some of these fans are so large they have partly dammed flood
waters of the main canyon and thus created temporary lakes or leveled
broad stretches of the canyon floor. In such places perennial grasses
may grow in quantities great enough to be cut for hay. However, the
features of these canyon floors constantly change, since floods vary in
volume, in quantity of material carried, and in the time of their occur-
rence in the annual cycle. Obviously a flood that occurs after a long
dry season during which vegetation has been reduced to a minimum
by the dryness of the soil will readily change the minor features of the —
plain over which it flows. However, generally speaking, the parched
soil will absorb so much that only when it is present in large volume
will floodwater be able to run the full length of a canyon. On the
other hand, a heavy rain and small flood in early summer may produce
such a rapid growth of vegetation that the plain will, for the rest of
the summer, be protected against erosion by much greater floods.

Obviously, also, floods down the axis of a canyon tend to produce a
slope more or less uniform at any locality but of decreasing declivity—
the so-called graded slope of a stream. Floods in the tributaries, how-

NO. 7 GEOLOGY OF CHACO CANYON—BRYAN 13

ever, tend to dump material on the floor of the main canyon and in-
terrupt this “graded slope.” Opposite the fan of a tributary, there-
fore, the main stream may have a marked channel that is constantly
renewed and shifted in position. Such discontinuous channels must be
sharply differentiated from the arroyos characteristic of dissection.
Discrimination of these minor channels from continuous channels
that indicate a new cycle of erosion is not always easy. Both are com-
monly described in the Southwest as arroyos. The character of dis-
continuous channels can best be explained by an example. The Cafiada
del Magre, located about 90 miles southeast of Pueblo Bonito, is a
narrow, flat-floored canyon tributary to the Rio Puerco (of the East).
Its walls are of buff Cretaceous sandstone similar in color, massive-
ness, and porosity to the Cliff House sandstone. A tributary, the Ca-
fiada de Bernardo, is similar. In 1909 a steep-walled gully about 25
feet deep extended from the arroyo of the Rio Puerco across the
abandoned flood plain of that stream and a quarter of a mile within the
Cafiada del Magre where it ended in a chaos of blocks of alluvium and
an intricate dry falls. This was the head of the new arroyo represent-
ing the readjustment of grade of the Rio Puerco—a readjustment
gradually affecting all its tributaries. Upstream from this arroyo head
the floor of Cafiada del Magre was a grassy flat for a distance of
about half a mile. There the center of the flat became sandy and this
sandy stretch was enclosed in low banks which gradually increased
upstream until they were about 6 feet high and which closed in from
the sides until they were only 10 feet or so apart. Here there was an-
other dry falls, more or less grassed over. Some grass also grew on
the floor of the arroyo. Above this latter falls was another grassy flat,
succeeded at irregular distances by similar gullies and similar grassy
flats. It is evident that these discontinuous channels are merely a phase
in the process of transportation and deposition on the floor of such a
valley. If the grade is locally out of adjustment but the grass cover
prevents the easy removal of a thin sheet of alluvium over the whole
floor, more sand and clay become lodged in the grass. Alluvium thus
continues to be deposited even though the floor is above grade. Finally
the strong currents of an exceptional flood break through the grass
cover and initiate a gully which then increases headward rather easily.
The overpour at the head of the gully so increases the ability of floods
to erode that the gully is carried deeper than necessary and its lower
end begins to fill with coarser sediment derived from the upper end.
The manner in which these gullies disappear can only be inferred
from the characteristics of certain ones which are very broad and low-
banked at the downstream end and shallow and almost obliterated by

14 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

the growth of grass at the upper end. Evidently lateral erosion in the
lower and middle parts of a gully removes material lying above normal
grade, and since the head of the gully stands below this grade it is,
with the help of vegetation, gradually filled in.

That such discontinuous gullies existed during the alluviation of
Chaco Canyon there seems no reason to doubt. The kind of sediment
in the floor and the nature of rocks in the drainage area are so similar
to those of the Cafiada del Magre and other canyons of the Rio Puerco
drainage as to preclude the possibility that dissimilar conditions ex-
isted. Cross sections of the valley fill exposed in the banks of the
modern arroyo show numerous irregularly placed channel deposits
from 10 to 40 feet wide and 2 to 10 feet deep that may well have been
deposited in discontinuous gullies.

The existence of channels of this type adds much uncertainty to the
usefulness of the accounts of travelers in dating the initiation of mod-
ern arroyo systems. Thus Simpson (1850) states that in 1849 the
banks of the Rio Puerco were 10 feet high and had to be cut down to
allow the passage of artillery at a point 7 miles above Cabezon. Banks
of similar height at the crossing of Arroyo Torrejon (Torreones)
and the Arroyo Cedro are also recorded. Yet even in 1926 the upper
portions of the Cafiada (en) Medio, a tributary of the Torrejon, and
the Cafiada de Piedra Lumbre were undissected and their flat floors
were farmed by Navaho Indians. A recent investigation (Bryan,
1928a) has, however, shown that the early discontinuous channels of
the Rio Puerco were relatively shallow and narrow compared to the
existing arroyo which was initiated between 1885 and 1890.

These minor features of the canyon floor, though constantly chang-
ing, doubtless had a marked similarity from year to year for there is
a balance between the forces involved that could only have changed
gradually with the progressive fill of sediment. The experienced ob-
server can easily predict what areas of such a canyon floor will be
gently flooded and what areas will be scored and eroded by tumultuous
waters. Only gently flooded areas will be suitable for floodwater
farming.

The process of filling Chaco and other canyons was doubtless a
slow one and for each foot of material permanently added many feet
were deposited temporarily and later shifted downstream. The irregu-
larities in bedding produced by this process preclude an estimation of
the time involved in deposition by any method now known.

At various places wind is effective in shifting material and thereby
adding complexity to minor features of a canyon floor. Much of the
material dropped by floods adjacent to channels of greatest flow is

NO. 7 GEOLOGY OF CHACO CANYON—BRYAN 15

incoherent and sandy. It is easily picked up by the wind and piled in
low mounds or dunes. Such material is often in motion within an
hour after recession of a flood. Nevertheless, these piles of sand are
often effective in changing the courses of channels and in spreading
floods. Areas so affected are small and usually are confined to three
locations, here listed in the order of their relative importance: 1,
Adjacent to channels; 2, on alluvial fans of tributaries carrying sandy
debris ; 3, outcrops of sandstone.

The work of winds in the vicinity of human habitations is notable
and is described on pages 21-22.

DATE OF RECENT STREAM TRENCHING

Historical records at Chaco Canyon are meager, and it is impossible
to fix precisely the beginning of its present arroyo. That this is recent
and that the process is still continuing is self-evident. Many of the
tributary canyons, such as Mockingbird, were yet undissected in 1925,
although a falls that receded each year marked the head of their re-
spective arroyos. Since 1925, extension of these gullies probably has
destroyed the alluvial plains in the tributaries. That the main Chaco
arroyo has increased since early expeditions to the canyon was recog-
nized by Dodge (1910) and the evidence is here reviewed. Not only
are its physical features recent but they resemble in every detail those
of other arroyos in the Southwest whose date has been fixed with some
assurance (Bryan, 1925a, 1928a). It seems reasonable to assume,
therefore, that the trenching of Chaco Canyon took place at about the
same time as the trenching of other valleys.

During a military expedition against the Navahos in 1849 under
command of Col., afterward Gen., John M. Washington, Chaco Can-
yon was visited and a description of its ruins recorded by Lt. J. H.
Simpson. According to this account (Simpson, 1850, p. 78), the “Rio
Chaco” had, on August 27, 1849, a width of 8 feet and a depth of 14
feet at the expedition’s camp near Una Vida. It is evident that this
description applies to the muddy water then flowing. No mention is
made of a gully although Simpson described the steep-walled arroyos
of three other streams that were crossed on the way to Chaco Canyon.

Lt. C. C. Morrison (1879, p. 367) visited the Canyon in 1875 but
does not mention an arroyo. Oscar Loew (1879) was there about the
same time, but his description of the topography of the canyon is too
vague to be of value.

William H. Jackson, whose pioneer archeological work in the South-
west is of great accuracy as to detail, spent five days in Chaco Canyon

16 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

in 1877. Five or six miles above Pueblo Pintado the arroyo was so
shallow that Navahos had formed ‘‘water pockets” (reservoirs) by
obstructing the channel; nearer Pintado the arroyo was Io or 12 feet
deep (Jackson, 1878, p. 433). Between Pueblo Pintado and Wejegi
the depth of the arroyo almost entirely cut off communication from
one side of the canyon to the other. Numerous small cottonwoods
grew along the bank. Near Una Vida Jackson noted that the arroyo
was dry where Lt. Simpson had found running water in 1849 and ex-
plains that his own visit was made in the spring, when floods are rare,
whereas Simpson was there in August when floods are more common
(ibid., pp. 436-437). At Pueblo del Arroyo the arroyo was 16 feet
deep and 40 to 60 feet wide, as stated and also shown by his sketch
map (ibid., p. 443 and pl. 59) ; 250 yards below this ruin there were
shallow pools of stagnant water and here Jackson camped. New grass
among young willows and cottonwoods in the bed of the arroyo ex-
tended for half a mile up and down stream (ibid., p. 446). The rapid-
ity with which this channel has developed may be judged from the fact
that, at Pueblo del Arroyo where Jackson recorded a depth of 16 feet
and a width of 60 feet, the arroyo is now (1925) 30 feet deep and 150
to 300 feet wide.

Mr. Judd has diligently collected local traditions on past conditions
in Chaco Canyon. Jack Martin, a long-time freighter in the region,
said that between 1890 and Igo00 the arroyo was so shallow most
freight wagons could be hauled across without doubling the teams.
The favored crossing was just south of Pueblo del Arroyo and part of
the dugway down the north bank is still recognizable. In 1925 Padilla,
an old Navaho who lives 7 or 8 miles west of Pueblo Bonito, stated to
Mr. Judd and also to me that in his boyhood the arroyo was an “ar-
royito” not more than breast deep (about 5 feet) and that along it
grew cottonwoods and willows. Since his boyhood it has continually
widened and deepened.

Later the same year Wello, a Navaho thought to be about 75 years
old, told Mr. Judd that Chaco River had no channel when he was a
boy ; that there were cottonwood and willow trees on the flat opposite
Pueblo Bonito and grass was knee high. Water was close to the sur-
face of the ground. Padilla was present at this interview and agreed
that these things were so. He may be 5 to 10 years younger than Wello
and doubtless based his agreement on knowledge gained from older
men, but he still insisted on the truth of his previous statement that,
in his boyhood, the “‘arroyito” was only breast high. If we can accept
1860 as the period to which these elders referred then we have an ap-
proximate date for the beginning of the Chaco arroyo. It was 5 feet

NO. 7 GEOLOGY OF CHACO CANYON—BRYAN 17

deep when Padilla was a boy; about 10 years earlier there was no
channel at all, according to Wello. Simpson did not mention an arroyo
because such a feature did not exist in 1849.

The original notes and maps of the townships surveyed under con-
tract for the U. S. Land Office Survey in the early 1880’s have been
inspected. These records contain gross errors, and some of the town-
ship surveys appear to be entirely fictitious. Hence no useful informa-
tion was obtained from them.

In attempting to determine a proper date for the beginning of ar-
royo cutting in Chaco Canyon, the problem is to decide whether the
arroyo described by Jackson in 1877 is a portion of a through-going
and complete arroyo system. It may have been merely a discontinuous
arroyo with a head in the broad areas of the valley near Fajada Butte
and a fan near the entrance of Escavada Wash downstream. Our in-
formation refers only to this part of the river and we have no data on
conditions farther down the Chaco. It seems necessary, therefore, to
assume that the arroyo of 1877 was the headward portion of a new
system and that the present arroyo is its successor.

The year 1877 cannot be the beginning of this arroyo and allowance
must be made for growth of the one described by Jackson. Some
weight must also be given to statements of the two Navahos, Padilla
and Wello. If Padilla is 70 years old he cannot remember farther back
than 1860 to 1865. Wello, an older man, remembers, or remembers
statements by others, that there was once no arroyo. Balancing these
considerations, it seems that 1860 is as early a date as is possible since
it affords 17 years for cutting the arroyo Jackson saw and is consist-
ent with the stories of the two old Indians. The date must be con-
sidered as an arbitrary choice, however, and not very precise.

Elsewhere in the Southwest existing evidence indicates that the
phenomenon of arroyo cutting began at slightly different times, stream
to stream. A considerable period elapsed between initiation of a given
channel and its completion throughout the length of its valley floor.
The date of beginning is apparently earlier in southern Colorado,
northern New Mexico, and northern Arizona than elsewhere, and may
easily fall in the decade 1860 to 1870. On Rio Puerco, however, a
nearby and similar stream, the date has been satisfactorily placed
within the years 1885 to 1890, and in the Hopi country, arroyos were
not cut until after 1900.

The effect of arroyo cutting on native vegetation and on habitability
of the area affected is very considerable. In southern Arizona, mead-
ows of coarse-top sacaton with groves of cottonwood and walnut have
been drained by arroyos, and dense forests of mesquite have since

18 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

sprung up. On the Rio Puerco, natural hay fields of fine-top sacaton
and other grasses once naturally irrigated have dried up and the deep-
rooted chamiso, “wafer sage,” has replaced the grasses. Areas near San
Ignacio and San Francisco that were once irrigated by use of crude
ditches or by natural flooding are now 30 feet above the stream bed.
Since these areas can no longer be farmed, the towns have been
abandoned.

Instances of similar change in native vegetation and of the abandon-
ment of fields could be multiplied (see references given in Bryan,
1925a, 1928b). The present lack is a quantitative estimate of the de-
crease in vegetation and consequent lessened opportunity for man
under the changed conditions. Primitive man, without domestic ani-
mals, would not suffer from decrease in forage or from loss of hay
fields but, to the extent that.he was dependent on floodwater farming,
might have his very existence jeopardized by these changes. On the
other hand entrenchment of Moenkopi Wash, near Tuba City, Ariz.,
has, according to Gregory (1915), increased the low water flow of the
stream and thereby provided more water for irrigation during the criti-
cal period of plant growth. An old Hopi farming village here, after
long abandonment, was reoccupied in 1880 and is now a thriving
community of about 300 people.

PRESENT GEOLOGIC PROCESSES
WEATHERING AND EROSION OF THE CLIFFS

The asymmetric form of Chaco Canyon is due to its east-west course
and to the prevailing dip of the rocks, as explained on pages I0-II.
The processes now at work on the cliffs differ from those of the past
only because of two factors: (1) The relatively recent valley fill which
covers the lower part of certain cliffs; (2) the possible differences be-
tween present and past climates. The first factor can have little effect
on the nature of the processes; the second affects only the rate of ero-
sion, as the probable changes in climate do not involve a change to a
strictly humid climate. When past climates were wetter than the pres-
ent but still relatively arid, cliff recession was doubtless accelerated ;
when the climate was drier, the process was slowed down.

Chaco cliffs can be divided into two sorts in two locations: The
lower division of the Cliff House sandstone generally forms vertical
cliffs but these differ in detail according to whether the base of the
sandstone lies below or above the level of the alluvial plain. The upper
division has domelike forms and generally produces low cliffs either
stepped or rounded where the base is above the alluvium. In lateral

NO. 7 GEOLOGY OF CHACO CANYON—BRYAN 19

canyons the base of this upper member falls below the level of the allu-
vium and here crenulated and rounded cliffs occur.

Erosion of the lower part of the Cliff House sandstone is largely
due to differential sapping at its base. Rainwater entering the sand-
stone above emerges below. If the base is above the level of the flood
plain, the water emerges at the top of the friable sandstones, coal, or
shale of the somewhat variable underlying Menefee formation. This
material is decomposed and carried away partly by this seepage water
and partly by direct rainwash. As a result the cliff is undermined and
blocks break off along characteristic joint planes. Perhaps because the
edge of the cliff settles and these joints are open some distance back
from the face, the cliffs of these localities are less sheer than those of
the same rock where the base of the sandstone is below the level of the
alluvial plain, as in the north wall of the canyon near Pueblo Bonito.
Here water absorbed by the overlying sandstone emerges at or near
the level of the alluvial plain. It dissolves the cement of the rock and
appears as an efflorescence of a white salt. The sandstone becomes
friable and grains are loosened from the surface. These grains fall
by gravity, especially during windstorms, or are loosened and carried
off by the sheet of water that covers the face of the rock in rains.
Thus cavities or niches (Bryan, 1928c) are formed like the one shown
in plate 3, left. With the formation of these cavities the rock
splits on its characteristic vertical joints; as loosened blocks fall, the
vertical face of the cliff is renewed. Narrow slabs several hundred feet
long, partly loosened, are fairly common features, and one directly
back of Pueblo Bonito has excited much interest because massive ma-
sonry below it shows that the prehistoric peoples attempted to brace
the slab against falling. [It actually did fall, on January 22, 1941.—
N. M. J.]

The upper member of the Cliff House sandstone tends to weather
in domal forms. Widely spaced vertical joints more or less at right
angles to each other afford points of attack for the weather and the
resulting, nearly cubical blocks are then rounded on the corners. The
process of sapping takes place in this sandstone also. There is, how-
ever, a relatively irregular zone of slabby and shaly sandstone below,
at which the water may emerge. On the double cliffs back of Pueblo
Bonito the sapping takes place over such a thick and irregular zone that
the upper cliff is discontinuous and in many places is replaced by a
series of benches. In canyons and rincons that enter Chaco Canyon
from the north, this contact passes below the level of the alluvium and
the zone of seepage emergence is more or less confined. The rounded

20 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

bosses are undercut. Where temporary waterfalls cascade over the
cliffs during rains, niches are formed.*

Such a niche, in Rincon del Camino, is notable because water
emerges under the overhang throughout the year and in sufficient
quantity to constitute a spring. Other niches may have only enough
seepage to support a few green bushes or there may be merely a damp
place on the rock. It requires, however, no stretch of the imagination
to perceive that with a slightly greater rainfall springs would exist in
these and similar situations.

The rate at which cliffs recede according to the processes just re-
viewed is necessarily slow. That parts of some cliffs are newer than
others is attested by their bright, unstained surfaces and the lack of
talus. Other parts have relatively ancient, iron-stained surfaces and
have shed no fragments since completion of the alluvial plain which
laps their bases. Some cliff faces are marked by carvings or picto-
graphs; others have had holes cut in them to support the roof timbers
of abutting dwellings. Often blocks at the foot of a cliff are so like
the cliff face in color they must be of equal age. In general it can be
said that, except for fall of a few blocks here or a mass of debris there,
the Chaco cliffs are essentially the same as they were when the canyon
was inhabited by prehistoric peoples.

The blocks of sandstone that fall at the foot of a cliff also slowly
weather and disappear. There are relatively few such occurrences in
Chaco Canyon; rarely are the blocks numerous enough to form a heap
or talus. The lack of talus may usually be explained as owing to burial
of all blocks formed previous to alluviation of the valley floor. Ex-
posed blocks are mostly recent falls and some of them are so little
weathered that they can easily be correlated with scars they left on
the cliffs. Others are much weathered and disintegrated. The princi-
pal process of weathering appears to be the solution of cement by rain-
water that percolates through the blocks and emerges on the side or
near the base. Numerous fantastically shaped holes are thus produced,
as illustrated in plate 4, lower. The movement of water through
such rocks was particularly observed on August 3 and 4, 1924. A
sharp shower occurred about 2 p.m. August 3 in which 0.14 inch of
rain fell. After the shower several rocks of this type were examined.
The exposed portions were wet, and part of the dust under the larger
overhang was eroded by the splash of falling rain, but the cavities were
dry. Next morning parts wet the day before were dry but the cavities

8 Bryan’s negatives were not found, but those he intended to use here are
reproduced in Zeitschrift fiir Geomorphologie, 1928, pl. 3, A, B.

NO. 77 GEOLOGY OF CHACO CANYON—BRYAN 2I

were damp, and some were almost wet. Evidently within 12 hours
water absorbed at the top had percolated through the rock. In rain-
storms of greater duration water must pour out of such cavities in
considerable volume and carry with it sand grains resulting from
previous solution of the cement.

WIND WORK

In various places on the cliffs, especially on the bench between the
upper and lower sandstones, there are patches of windblown sand.
This sand, evidently derived from disintegration of the nearby sand-
stone, accumulates in places more or less sheltered from the wind.
Some heaps are fixed in position by the growth of grass. That such
heaps accumulate is proof that a much greater quantity of sand is
moved by the wind and either blown off the cliff altogether or into
position where it is carried away by rainwater. How much the move-
ment of this sand scours the rock is difficult to evaluate. It seems
likely that the scouring effect is small, for the rock at the surface is
soft and crumbly. At the mouth of Escavada Wash, where the Farm-
ington road left the Chaco prior to 1920, the wind has heaped up sand
out of the wash to form a group of dunes that encroaches on cliffs
similar to those upcanyon (Bryan, 1928c). Here the sandstone has
the same domes, niches, pinnacles, and other characteristic details, yet
the surface is hard and firm (pl. 4, upper). Tiny iron concretions
stand above the surface like collar buttons. The surface is continuously
scoured by sand and is in marked contrast, by reason of its firm tex-
ture, to the soft and crumbly surfaces seen elsewhere. Nevertheless,
the total erosion is obviously less than it seems, for otherwise the
domes, niches, etc., would disappear and be replaced by new details
equally dependent for their shapes on the process of wind scour.

Near Pueblo Bonito, dust is easily raised on any windy day and
windblown sand accumulates, as it has in the past, in every sheltered
nook and cranny. Between the ruin and the cliff sand had collected to
a depth of 4 to 6 feet between 1900, when the Hyde Exploring Expe-
dition concluded its excavations, and 1921, when the National Geo-
graphic Society inaugurated its researches. Similarly, windblown sand
near other ruins of the region evidences more wind work than appears
at places of otherwise similar location.

These conditions are easily understood when the activity of man
and his domestic animals is considered. Near his habitations man and
his animals continually disturb the surface soil and thus make the work
of wind easy. In addition the soil is made pulverant by abnormal

22 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

quantities of organic matter consisting of the excrement of men and
animals, the debris of crops gathered and brought in, the refuse of
building materials and fuel, and litter of all sorts swept up and carried
out of houses. These organic substances added to the surface soil
serve to make it friable in the same way that manure improves the
tilth of a field. The soil, when wet, is no longer a gummy mass that
becomes pavement-hard on drying. It is less muddy when wet, and
when dry is loose and friable and thus easily picked up and transported
by the wind.

With only dogs and turkeys to help in the processes just described,
prehistoric peoples probably did not create as much dusty ground as
the same number would today. Yet their refuse mounds testify to an
enormous quantity of rubbish discarded systematically and enable one
to picture the proportion of such litter that must have been scattered
about the inhabited areas. The quantity of windblown sand and wind-
reworked material found buried in the refuse heaps of Pueblo Bonito
evidences a considerable amount of wind work in prehistoric time that,
on the grounds set forth above, may be considered as influenced by the
habits of prehistoric man.

ALLUVIAL FANS

At the foot of the cliffs, and particularly at every indentation in
them, alluvial fans are now forming. These are composed largely of
buff and yellow sand derived from disintegration of the sandstone of
the cliffs and talus blocks. During every rain, sheets and streams of
water pour over the cliffs; the largest flows naturally occur at the in-
dentations. Alluvial fans deposited by these streams are more or less
proportional in size to the indentation and, by inference, to the area of
surface drained. In a few places the sand of alluvial fans is picked
up by the winds and formed into low heaps, but, as was pointed out
above, currently there is more wind-moved material in and around the
various ruins than at other places in Chaco Canyon.

The soil of these alluvial fans is loose and sandy and doubtless
formed the best agricultural land in prehistoric time. During and after
rains, waters that pour from recesses in the cliff could easily be
directed over the fans. The problem of directing and spreading such
runoff so it will wet without gullying the land is a difficult one which
must have taxed the ingenuity of the prehistoric farmer. It is possible,
but by no means certain, that part of these floodwaters were directed
over the more clayey areas in the middle of the valley not only for the
purpose of irrigation but also to mix water-borne sand with the less
tractable clay.

NO. 7 GEOLOGY OF CHACO CANYON—BRYAN 23

ADOBE FLATS

Adobe flats still constitute a large part of the floor of Chaco Canyon
and are added to each season by storm waters draining from adja-
cent areas. They are, however, no longer built up by the marginal
waters of main stream floods or by deposition in temporary lakes.
Progressive cutting of the arroyo has left these flats mere relics of past
conditions, but the method by which they were formed can be inter-
preted from observations in such undissected tributaries as Mocking-
bird Canyon.

EROSION IN THE ARROYO

The initiation of dissection and the formation of arroyos has al-
ready been discussed (pp. 12-13). The present (1925) arroyo varies
in width from 150 to 500 feet, and in depth from 10 feet near Esca-
vada Wash to 30 feet at Pueblo Bonito. Upstream, however, the
height of the bank again decreases to about 20 feet near Wejegi.

The vertical walls are formed by undermining the alluvium which
then breaks off in blocks parallel to an obscure jointing. Undermining
is largely due to the lateral cutting of floods in swirls and eddies on
the outside of bends. It seems probable, however, that water absorbed
in the alluvial plain seeps into the arroyo at the base of these banks,
softening and helping to undermine them. The rate of such lateral
cutting is rapid, and significant changes may occur in a single year
(pl. 5, upper and lower).

THE VALLEY ALLUVIUM
MATERIALS OF THE FILL

The alluvial fill of Chaco Canyon consists largely of sand, yet so
much of its surface is covered by a layer of dark sandy clay (locally
called “adobe’’), that the true character of the fill is not evident except
where exposed in the arroyo. Even there the degree of sandiness can
be detected only by close inspection, for rains wash mud down over
the vertical walls and this, like a film of plaster, conceals the sand and
gives the impression that the whole bank is made of clay. The relative
quantities of sand and clay will appear from the measured sections on
pages 51 to 59; in the following paragraphs, each class of material is
described separately.

Small discontinuous bodies and lenses of gravel occur sparingly.
The gravels contain much fine sand and mud and are thus similar to
the small bodies of gravel found in the present arroyo bed. Like the
latter they are obviously the deposits of tumultuous and muddy floods.

24 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

The pebbles are mostly partly rounded fragments of sandstone and of
limy concretions derived from the Mesaverde and Pictured Cliffs
sandstone, but there are also a few water-worn pebbles secondarily
derived from the Tertiary rocks. These pebbles may vary in diameter
from half an inch to 2 inches, but angular stones up to 6 or 8 inches
in diameter are also found.

The sands are of several types not wholly distinct in character that
grade into one another even in the same bed. Some sands have almost
no bedding or lamination and consist of yellowish, rusty grains such
as result from disintegration of the sandstone of the canyon walls.
These accumulations seem to have been deposited by outwash and are
parts of ancient fans similar to those now being formed at the foot of
the cliffs. Lenses of clean white sand, laminated and crossbedded,
were laid down by the stream of the main canyon, just as like lenses
are being deposited today. Somewhat similar but generally finer-
grained sands may have very minute and irregular crossbedding that
is a remnant of wind-made ripples. Such beds clearly have been sub-
ject to wind action but it seems likely that the sand was first deposited
by water and subsequently reworked by the wind. The very same proc-
ess may be seen today when sand recently deposited in the channel by
a flood is reworked by wind shortly afterward.

Gray or brown silt in beds that are minutely though irregularly
laminated make up a considerable part of the alluvium. Ripple marks,
current marks, and mud cracks are common on the surface of the
laminae; a few rusty streaks, some parallel and some at large angles
to the laminae, are doubtless the impressions of stems and roots. In
many places the silt beds are 2 to 4 feet thick, but in others silt occurs
as thin layers in sand or clay. Obviously it was deposited in outer
portions of the channel by the main stream or in nearby parts of the
flooded area.

The clay of the valley fill locally known as adobe is mostly gray or
brown and contains appreciable amounts of sand and silt. A more
realistic name would be “sandy or silty clay” or “clayey fine sand.”
The usual clay bed is from 6 inches to 3 feet thick and is nearly uni-
form in color and texture. Lamination is not always apparent to the
unaided eye but it probably is a constant feature since it has been de-
tected under the microscope and doubtless would be readily visible
were it not for vertical jointing and a tendency to “check.”

Well-laminated, less sandy clays occur in discontinuous lenses a few
feet long and varying from 2 to 12 inches in thickness. These lenses
appear to have been deposited by floodwater that was ponded in aban-
doned and shallow channels. In contrast, the more abundant sandy

NO. 7 GEOLOGY OF CHACO CANYON—BRYAN 25

clay is the result of sheetlike overflows of the main stream and
of its tributaries. These overflows formed flats or meadows of alkali
sacaton grass. The imperfect lamination and reticulated structure are
probably to be accounted for by successive drying of the mud on ex-
posure after each overflow and by the influence of grass roots. The
dark color is due to finely divided organic matter incorporated in the
clay. In such flats the rate of soil formation is rapid and the reticulated
(chernozem) soil structure is quickly attained.

ARRANGEMENT OF THE MATERIALS

The walls of Chaco arroyo and its various branches afford ample
exposure in which the several sorts of material just described are dis-
played. The dominant characteristic is lack of continuity. No bed,
however uniform over a short distance, extends very far, nor does a
like sequence of beds occur in the wall of the arroyo at any two places.

In general clay beds are the most continuous and some of them have
a length of nearly a quarter mile. Such beds rest on, and in turn are
covered by, essentially parallel beds of different composition from place
to place. These changes in lithology imply that there are many minor
unconformities between beds. In some places sharp and definite ero-
sion surfaces cut clay beds which are overlain by sand deposited in the
channel responsible for the erosion. Less definite evidences of erosion
contemporary with deposition occur, but no one of these episodes of
erosion within the formation has any time value, for such irregularities
are to be expected in the deposits of ephemeral streams. Each large
flood forms new channels of flow and destroys in part the work of its
predecessor. In such localities the measure of erosion or sedimentation
is the algebraic sum of the erosion and deposition of successive floods.

In several places a cross section of a much older channel is exposed
in the wall of the arroyo. These exposures generally show a round-bot-
tomed channel, filled with crossbedded sand to a depth of 2 to Io feet,
and the lenticular character of adjacent clay beds. Such channels are
evidently interformational for they are shallow, occur at all levels,
and are covered by beds that are conformable with others, adjacent
and contemporaneous. Evidently these channels are of the discontinu-
ous type described on pages 13-14 and are to be distinguished from
the relatively recent, buried channel discussed on pages 32-37.

DEPTH OF THE ALLUVIUM

Tangible evidence on the total depth of the valley fill is wanting in
Chaco Canyon. The well at the Society’s Pueblo Bonito camp was dug

330 Feet

9
as
h
ge
N
ars
set
ow
Oo °
Cec
ed
a
S)

1go

BANK _ OF

BETWEEN SECTIONS 5 AND 16
140
|

DIAGRAM OF

7.
rg wees

“pusble ycharco

Pre

herds 3

.. pots

Q
t
y
=
9
v

COVERED

Bed of Checo River

Bed of Chaco River

Sand with gravel lenses

as Shown.

Fic. 3.—Diagrammatic sketch of the valley fill between sections 5 and 16.

Ss.
Sot

seh 3 or
ot 7 a y
Ua es &

re) Fy

PLATE 6

Upper: Annually since 1877 Chaco floods gnawed at this small Pueblo III ruin until the
last vestige of it disappeared in 1948. Pueblo del Arroyo stands at the right, beyond the sheds.
(Photograph by Charles Martin, 1920.)

Lower: A Pueblo I pit house with roof 6 feet below the valley surface was revealed by cav-
ing of the arroyo bank during the 1921 flood season. (Photograph by Neil M. Judd, 1922.)

Upper: The buried channel as exposed at section 16. Here Late Bonitian potsherds were
found at a depth of 13 feet 10 inches. (Photograph by O. C. Havens, 1925.)

Lower: The author stands below a cross section of the post-Bonito channel at section 4, near
the southeast corner of Pueblo del Arroyo. (Photograph by O. C. Havens, 1924.)

PLATE 8

Upper: As it did in the days of Pueblo Bonito, the Rincon del Camino comes in from the
north to join the main Chaco arroyo. (Photograph by O. C. Havens, 1924.)

Lower: Lateral erosion caused by drainage from The Gap. View looking north across the
Chaco toward Pueblo del Arroyo in left middle distance. (Photograph by Neil M. Judd, 1929.)

PLATE 9

Upper: About half a mile west of Ruin No. 9 a disarticulated skeleton was found in the
arroyo bank 6 feet below the surface.

Lower: Hearth at depth of 12 feet 8 inches, section 9, at mouth of Rincon del Camino.

(Photographs by O. C. Havens, 1925.)

NO: 7 GEOLOGY OF CHACO CANYON—BRYAN 27

about 10 feet below the level of the stream bed in the arroyo, here 30
feet deep, and thus shows at least 40 feet of alluvium. About 1900 a
well said to have been 350 feet deep was drilled between Richard
Wetherill’s old home and the southwest corner of Pueblo Bonito and
was promptly abandoned because the water was brackish. No record
of the material passed through is available.

In other, somewhat similar canyons and valleys, the alluvium is
known to be relatively shallow. As recorded by Gregory (1916, p.
160), wells at Leupp, in the Little Colorado Valley, northeastern Ari-
zona, and along Oraibi Wash, farther north, show a depth to bedrock
ranging from 60 to 80 feet. In the Puerco Valley, 50 to 70 feet of
alluvium is found. At Gallup, near the Arizona line in New Mexico,
there is 175 feet of alluvium above bedrock.

From these data one may assume that the alluvium in Chaco Canyon
probably does not exceed 100 feet in depth. Thus there is exposed only
about one-third, or 30 feet, of the total. The generalizations of this
paper are based on an examination of this exposed third but we may
be sure an equally thorough examination of the lower and unexposed
portion would reveal equally interesting and significant information.

HUMAN RELICS IN THE VALLEY FILL

During investigation of the sediments exposed in the walls of Chaco
arroyo, relics of man’s activity were observed repeatedly. These relics
consisted mostly of hearths, charcoal, stones, bone fragments, and
potsherds. Remnants of masonry structures were also found, although
infrequently. Most of the hearths are not constructed fireplaces but,
rather, surfaces more or less blackened or reddened by heat and over-
lain with charcoal. They are merely places where fires had been built
and obviously represent no more than temporary camps. Some may
have been used for only a day; others, to judge from the depth of the
burnt ground and the amount of charcoal, could have been used for a
period of weeks. Some hearths have been buried by sand or clay laid
down so gently by running water that the charcoal was not eroded.
They are thus good evidences (1) that the contemporary surface at
that place was occupied by people during the period of alluviation and
(2) that alluviation was accomplished in part by gently moving sheets
of water which did not destroy previously existing surfaces.

Scattered charcoal, on the other hand, is not conclusive evidence of
the presence of man since it may have been incorporated in the sedi-
ments by erosion of areas where the vegetation was burned by fires of
natural origin. However, as previously pointed out, the character of

28 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

the sediments indicates that the region has always been an arid one.
In such a region fires of natural origin are, because of the scarcity of
burnable vegetation, likely to be small in extent and thus produce small
amounts of charcoal. Certainly charcoal is not common in the stream
beds of the present time, either in this area or in the Southwest gen-
erally. It seems more likely, therefore, that occurrences of scattered
charcoal in the Chaco fill are due to the erosion of hearths or refuse
dumps. Since charcoal is soft and friable, it cannot be carried far and
indicates human occupation at no great distance upstream.

Isolated stones in the fine-grained sediments occur in many places,
but, however others may have reached their present positions, those
found near hearths were probably brought in by man. Their presence
thus tends to confirm such evidence of onetime human occupation as
worked flints, stone and bone artifacts, and fragments of pottery.

Bone fragments are not positive proof of man’s presence unless
showing signs of human workmanship but, like scattered stones, may
be confirmatory when in association with man-made objects.

Potsherds are resistant to erosion and provide durable and unques-
tioned evidence of man’s presence in the past. They have such large
surfaces in proportion to their weight that they may be carried in cur-
rents unable to transport small stones and thus often occur in fine-
grained deposits. Generally, however, they are associated with gravels.
Some potsherds, found singly in fine-grained beds, may have been
dropped on the surface and simply buried by the mud of the next
flood. Deposition by floods on uneroded surfaces was doubtless com-
mon, as shown by the lack of erosion of charcoal in hearths. Pot-
sherds also have a genuine stratigraphic value because pottery made by
prehistoric peoples at different periods differs in texture and decoration
and can be identified with some certainty. Thus, if collected system-
atically, potsherds can be used as fossils are used to identify the age of
sediments.

COMPLEX CHARACTER OF THE VALLEY FILL

The nature of the sedimentary deposits comprising the main fill of
Chaco Canyon is clearly revealed in the walls of the present arroyo.
Here 30 feet and more of successive strata have been laid bare and one
has only to read the story they tell.

Before discussing the significance of these exposures and the meth-
ods used, some of the difficulties we encountered will be mentioned
for the benefit of future workers. Study of arroyo banks involves the
tiresome traversing of the stream bed, which is soft and sandy when

NO. 7 GEOLOGY OF CHACO CANYON—BRYAN 29

dry and may be dangerously boggy just after a flood. The banks are a
monotonous brown and in many places are covered with a film of mud
washed down from above.

Like a coat of calcimine over a fresco, this film obscures the charac-
teristics of the materials of the bank. The freshest exposures are
usually on the outside of bends for here the stream cuts laterally with
greatest activity. On the inside of bends, temporary bars or heaps of
windblown sand accumulate and may cover the bank to half its height.
Frequently there has been left adhering to the bank a remnant of such
a bar formed at some past time in the process of cutting the arroyo.
Occasionally these remnants form little terraces 3 to 20 feet wide and
4 to 6 feet lower than the top of the bank. They consist of Chaco
River deposits similar to the rest of the valley fill and are so deceptive
in appearance we came to call them “false banks.” Sometimes the
“false bank” contains sheep dung and recency of deposition is thus
definitely established. Elsewhere one must rely upon the form of the
bank or some local difference in bedding, in color, or in grain of the
clays and sands to distinguish the false from the true bank.

In full sunlight the relatively small variations in color and texture
of beds are almost invisible. Therefore each locality was scrutinized
at least twice, once in the morning and once in the afternoon so that,
as nearly as possible, every part of the exposure could be seen in the
shade.

Detailed descriptions of the valley fill at each of the 23 sections ex-
amined would merely weary the reader. A number of such descrip-
tions are offered hereinafter (pp. 51-58) for those interested prima-
rily in the geological aspect of this study, but our present purpose will
be served by a single example. Typical conditions are shown at a
locality on the south bank of Chaco River, opposite Ruin No. 8 and
approximately half a mile west of Pueblo Bonito (fig. 1). Our dia-
gram (fig. 3) was constructed by measuring the beds on verticals
spaced ro feet apart and by sketching the form of the several bodies
of material in the intervals.

On the left are shown the irregularly continuous beds of the main
valley fill at section No. 5. A conspicuous feature here is a fireplace
of nearly vertical stone slabs built when the surface was 5 feet lower
than it is now (pl. 3, right). Archeologists consider this type of fire-
place characteristic of the period of settlement known as Pueblo III.
Nearby, both in the bank and on the surface, are potsherds of that age.
A few feet to the west, however, and at a depth of 6 feet 3 inches,
sherds of indeterminate age were found. A second hearth at a depth
of 12 feet 8 inches, and charcoal at 16 feet 3 inches, are doubtless rec-

30 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

ords of the presence of people earlier than Pueblo III. Near vertical
180 a potsherd of distinctive type attributable to pre-Pueblo (P. I)
times was found in a loose block of earth. This block retained the
shape of the bank so perfectly there is no doubt the sherd came from
a depth of 11 feet 3 inches.

At the right of vertical 180 is represented the coarse, sandy filling
of an ancient arroyo at section 16 (pl. 7, upper). In the gravel lenses
of this old channel and at a maximum depth of 13 feet 10 inches, pot-
sherds of the type characteristic of the latest phases of Pueblo Bonito
were found. The sherd collection here was comparatively large be-
cause the gravel lenses could be followed for 50 feet at right angles to
the section in the lateral tributary of the modern arroyo.

Generally speaking, material within this ancient arroyo is sandier
and more crossbedded than that of the main valley fill. In other sec-
tions examined the laminations dip from both sides of the channel
toward the middle and at or near the bottom there are one or more
lenses of gravel. These lenses are, on the average, 3 to 12 inches thick
and 2 to 4 feet long. The gravel is clayey and dirty and similar to
gravel beds of the valley fill or those in the present stream bed. How-
ever, potsherds are relatively plentiful and most frequently of Pueblo
III type. The unusual number of these late potsherds is in itself sig-
nificant and, so far as my experience goes, an infallible indicator of
the presence of the buried channel. Connecting channels of con-
temporary age are commonly filled with coarser material similar to
that in the present tributary arroyos.

Typical conditions are thus recorded in the main valley fill between
sections 5 and 16. The current arroyo has exposed its predecessor and
bared evidences of human occupation in times past. To depths of 4
feet and more, rarely 6 feet, potsherds of Pueblo III type are fairly
common. Below this horizon the sherds are of definitely earlier types
or of indeterminate age. Here, as elsewhere, we detect no physical di-
vision between the lower and upper sedimentary beds, merely a sepa-
ration of the early and late pottery fragments found in them.

RELATION OF HUMAN RELICS TO MAIN VALLEY FILL

The various relics of man’s activity found exposed in the walls of
the modern arroyo indicate that man inhabited Chaco Canyon during
the latter part of the period of alluviation. The greatest depth at which
potsherds were seen during the course of our investigation was in a
bed ranging from 17 feet to 20 feet 6 inches below the surface of the
valley floor (section No. 3, fig. 1). At another site, section 8, charcoal

NO. 7 GEOLOGY OF CHACO CANYON-—BRYAN 31

was found at a depth of 21 feet. Man may have been present during
the earlier phases of sedimentation but his presence during deposition
of the last 21 feet of fill is definitely established. It seems a fair in-
ference that in the prehistoric period most streams in this region were
building up their flood plains and primitive man witnessed the process.

Chaco Canyon was by no means unique in this respect. Ashes, pot-
tery, artifacts, and like evidences of man have been recovered at vari-
ous depths in the alluvium of other Southwestern valleys. Published
records are not so numerous as might be expected but those available,
together with some of the present writer’s observations, are tabulated
herewith: *

Relics of man in the recent alluvium

Name of stream icy
State valley surface Author Reference
IATIZONA i. SE Rio de Flag 9 A. E. Douglass 1924, pp. 238-239
Santa Cruz 10 E. Huntington 1914, p. 24
Navaho Country 12 H. E. Gregory IQI5
Navaho Country Ar A. B. Reagan 1924a, pp. 283-285
Pueblo 0) A. B. Reagan 1924b, pp. 335-344
Colorado Wash
SomOfacnrccciaesls Sonoita River 12 C. Lumholtz 19I2
Coloradonaet. ai Montezuma Canyon 2 G. O. Williams 1925, pp. 201-202
New Mexico.... Chaco Canyon 17 R. E. Dodge 1920, pp. 23-25
Chaco Canyon 14 W.H. Jackson 1878, pp. 431-450
Coyote Canyon, 6 Kirk Bryan 1925a
Sandia Mts.
Rio Puerco 6 Kirk Bryan 1926b
Nutriosa Creek 6 Kirk Bryan 1926b

Zuni drainage

In Chaco Canyon, potsherds and other artifacts were collected from
noteworthy depths at many different places. Generally, sherds from
the upper 4 to 6 feet of the valley fill are of relatively late types of
pottery (Pueblo III). Earlier types are found at greater depths.
From our data it seems clear that the upper 4 to 6 feet of the main
valley fill is not only younger than the lower horizons but that the di-
vision between the prehistoric periods known as Pueblo II and III lies
at the base of this layer.

Two Pueblo I pit houses in Chaco Canyon were excavated and de-
scribed by Judd (1924). One of them, so far as critical evidence was
preserved, may have been built after alluviation had filled the valley

4It is to be emphasized that this paper was begun in 1925 and was added to
from time to time until 1940, when it was left incomplete. Hence the absence
of later references.—N. M. J.

32 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

to its present level. It was situated at the base of the talus on the south
side of the canyon, opposite Pueblo Bonito. The other and more per-
fect example, at the locality marked “Pit House” on figure 1, was evi-
dently occupied when the surface of the alluvial plain was 5 feet 11
inches or, in round numbers, 6 feet below the present surface (pl. 6,
lower). Because potsherds of precisely the same type as those recov-
ered from these two P. I pit houses have been found repeatedly in the
valley fill, one can scarcely avoid the conclusion that during the time
when strata at depths of 4 to 21 feet were being deposited, Chaco Can-
yon was occupied predominantly by people of this cultural stage.

The exact relationship between the Pit House people and later
inhabitants of the valley is not wholly clear at this writing. Pottery
similar to that from Judd’s two pit houses has been found in and below
rubbish associated with the older parts of Pueblo Bonito and portions
of a slab-lined pit house were encountered 12 feet beneath the west
court of the great ruin. I leave solution of this puzzle to the archeolo-
gists but, at the moment, it would seem as though the inhabitants of
Chaco Canyon had passed rather abruptly from a P. I to a P. III
society. So direct a transition in human culture appears physically
possible for the valley fill records no break in sedimentation ; hence the
environment must have been relatively uniform during the change.

As previously stated, numerous relics of the Pueblo III people have
been found in the upper 4 feet of the main body of the valley allu-
vium. Most distinctive of these remains are house walls which, in a
number of instances, rest on undisturbed material 3 feet or more below
the general surface. There is no better example than the small house
on the south bank of the arroyo, opposite Pueblo del Arroyo. Its re-
lationship to the underlying strata is clearly shown in plate 6, upper.
One corner had recently been undermined when Jackson first noted
its precarious situation in 1877. Since then the little house has paid
annual tribute to Chaco floods and its complete destruction is only a
matter of time.

Near section No. 2 (p. 53) the foundations of another small build-
ing reach a depth below the surface of 8 feet, but this is exceptional.
Apparently at this site the alluvial fan of the adjacent tributary has
been built up several feet above the normal level.

THE BURIED CHANNEL AND ITS SIGNIFICANCE

One of the notable discoveries of W. H. Jackson during this keen
observer’s visit in 1877 was a buried channel just south of Pueblo
del Arroyo, 14 feet deep and containing numerous potsherds, flint

NO. 7 GEOLOGY OF CHACO CANYON—BRYAN 33

chips, and other human relics (Jackson, 1878, pp. 443-444). Un-
questionably this exposure was among those later seen by Prof. Rich-
ard E. Dodge and reported in published excerpts from his field notes
of 1899 and 1901 (1m Pepper, 1920, pp. 23-25). It is perhaps only nat-
ural, therefore, that the present writer, although these earlier dis-
coveries were unknown to him at the time, should have happened upon
the same ancient channel mentioned by Jackson and Dodge.

As shown in plate 7, lower, this buried channel is round-bottomed
and crescentic in cross section. Its boundaries are sharp and clearly cut
across the horizontal bedding of the older valley fill. The materials
within the channel are such as might be expected: gravel and cross-
bedded sand 3 to 4 feet thick form the base of the deposit and are suc-
ceeded by clay which, in turn, is overlain by successive beds of sand
and sandy clay. The first clay bed seems to indicate a ponding of flood-
water in this channel while the main floods were diverted into another.
A series of sandstone blocks in the clay bed resembles stepping stones
placed to provide a passage over the mud but such an explanation of
the blocks, though plausible, can scarcely be considered as established.

Buried potsherds, broken bones, and beads were revealed also in a
reentrant of the modern arroyo which existed in 1924 and 1925 a few
yards from the south wall of Pueblo del Arroyo and a hundred feet,
more or less, west of the exposure described above. The potsherds
included some of the latest varieties of pottery made in Pueblo Bonito
and were, according to Mr. Judd, produced somewhere around A.D.
1100. Although this deposit was visible for about 75 feet its lateral
contacts were, at the time of my examination, unfortunately obscured
by a small gulley on the east and by unusually severe weathering on the
west.

With two exposures in line, extension of this buried channel east-
ward under the plain before Pueblo Bonito was obvious. To provide
a third contact, a test pit (pit No. 4) was dug 60 feet back from the
edge of the bank where the channel was first observed. At a depth of
18 feet sherds of the latest Pueblo Bonito pottery were found, proving
this pit to be located in the buried channel.

Nearer Pueblo Bonito, Judd’s test pit No. 3 was deepened and more
late pottery was discovered. Next, a trench previously dug through
the west refuse mound as a means of studying its composition was ex-
tended out into the flat. In this extension the north bank of the buried
channel was clearly profiled. The channel filling, marked by coarser
material, rests unconformably on the edge of the old refuse mound
and on strata of the main valley fill.

34 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

Farther east, an outcropping in the north wall of the present arroyo
gave still another point and thus enabled us to indicate the course of
the buried channel as it passed Pueblo Bonito. In 1936 Senter (1937)
dug pits near Pueblo Chettro Kettle and likewise exposed a buried
channel containing potsherds of late date. From the position given,
however, I would judge his buried channel to be a branch of the
arroyo produced by the stream which drained the reentrant in the
cliffs back of Chettro Kettle.

A careful reading of Jackson’s account shows that there was in
1877 an abandoned side channel 8 feet deep, or half that of the main
arroyo, between the latter and the ruins of Pueblo del Arroyo. This
side channel (“old arroyo” of his map) still existed, in part at least,
in 1925. It must be distinguished from Jackson’s buried channel which
he describes as 14 feet deep and marked by “‘an undulating stratum of
broken pottery, flint chippings, and small bones firmly embedded in a
coarse gravelly deposit.” This stratum he first observed below masonry
walls exposed by the side channel above referred to—walls that did
not show on the surface. He traced the stratum upstream “several
rods” in the vertical north bank of the main arroyo, noted its presence
also on the opposite side, and followed it thence farther upstream to
the small ruin shown in plate 6, upper. The stratum reached its
lowest level 14 feet below the surface, or 2 feet above the bed of the
main arroyo, a fact that led to Jackson’s conclusion: “Since the de-
sertion of this region the old bed has become filled to the depth of at
least 14 feet, and through this the arroyo has made its present chan-
nel.” (Jackson, 1878, p. 444.)

In 1924 and 1925 the walls of the “side channel” were much weath-
ered and waste from the Society’s excavation of Pueblo del Arroyo
had been dumped into the main arroyo southeast of the ruin in an
effort to check further erosion. The exposures described by Jackson
had been destroyed in the widening and deepening of this arroyo but
in an excavation made for a storage cellar back of the old Wetherill
“hotel,” the northern border of a channel containing late Bonito
pottery could still be seen.

When considering the significance of the buried channel described
above, I have often thought of it as the “post-Bonito channel” because
late Bonito potsherds on the bottom of it identify the channel with the
final years of Pueblo Bonito or even later. In 1925, by means of test
pits in the plain that lies in front of Pueblo Bonito, we traced this old
channel eastward from the exposure near Pueblo del Arroyo for more
than 1,000 feet. The lenses of clayey gravel and other materials found
in those pits were thoroughly characteristic of channel deposits. Late

NO. 7 GEOLOGY OF CHACO CANYON—BRYAN 35

Bonito pottery fragments were found from 10 to 18 feet below the
surface in test pit No. 3 and from 7 to 15 feet below in pit No. 4.
Additional exposures enabled us to plot the channel’s course with a
high degree of accuracy for several miles.

That this buried channel is part of a continuous arroyo once extend-
ing the full length of Chaco Canyon seems definitely established. The
existence of such a through channel is proof that there intervened dur-
ing alluviation of the valley a period of arroyo formation and dissec-
tion similar to the present. To put it another way, an early period of
alluviation, represented by the main valley fill, was followed by a pe-
riod of dissection that included our post-Bonito channel. After an
unknown interval alluviation was again resumed, the arroyo system
was filled up and some slight addition made to the valley deposits, and
alluviation, or at least a balance between alluviation and erosion, was
continued down to the year 1860 to be interrupted in the ensuing
decade by formation of the present arroyo.

Such a history appeals to the imagination by reason of its symmetry
and because of the nice correlation with human history that can be
made, as outlined in a later part of this paper. However, the impor-
tance of the issues involved requires that available evidence be ex-
amined with care.

The buried channel, as exposed in various places, is round-bottomed
although vertical side walls have been detected in a few instances.
Vertical walls are to be expected in any through-flowing arroyo. How-
ever, as previously pointed out, local conditions often prevented a criti-
cal study of side contacts of the old channel. The present arroyo, which
coincides very closely with this post-Bonito channel throughout much
of the area under examination, has, by reason of its greater width,
removed all traces of the buried channel over considerable distances.
Sections of the old channel have been discovered largely through search
for lenses of gravel containing late-Bonito potsherds, and in many
localities the side contacts of the old channel are poorly exposed. The
best evidence found that however round-bottomed the channel may
have been, its walls were generally vertical, is its known length and
presumed extension both ways.

Supporting evidence that the post-Bonito channel is part of a con-
tinuous arroyo is given by the existence of lateral channels represent-
ing its important tributaries. A mile downstream from Pueblo Bonito
there is an indentation of the north canyon wall which we called “the
Rincon del Camino.” The present road to Farmington and Aztec goes
this way. Drainage from a considerable area falls over the cliffs at the
head of this rincon and reaches the main arroyo through a half-mile-

36 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

long tributary (pl. 8, upper). The material exposed in this tributary
consists of irregularly bedded yellow sand, numerous lenses of dirty
gravel, and a few lenses, 2 to 6 inches thick, of laminated dark clay.
All this material, except the clay, is similar to that now carried by the
tributary stream and was undoubtedly derived exclusively from the
drainage of Rincon del Camino. Similar sand and gravel are exposed
in the north wall of the main arroyo for about 300 yards east of the
tributary.

There is thus a triangle of material derived from the alluvial fan of
the side stream extending from the main arroyo to the head of the rin- .
con, as outlined in figure 1. This triangular mass contains no material
from upstream except the clay beds, and evidently the area could not
have been occupied by the main stream of the canyon during deposi-
tion of the fan. All other evidence, however, leads to the conclusion
that during the period of alluviation the main stream wandered at will
over the canyon floor. If, however, an arroyo similar to the present
one had been formed, contemporary drainage from Rincon del Camino
might easily have excavated a tributary large enough to carry away
and destroy the main valley fill over fhe entire triangular area shown
on the map. If conditions of alluviation were restored thereafter and
the main arroyo began to fill up, the excavated area would also be
filled. The waters of the main arroyo could, however, enter the area
only as a gentle overflow or backwater from which clay similar to that
of the clay lenses would have been deposited. Most of the filling would
consist of sand and gravel brought in by the tributary stream. It seems
obvious, therefore, that the fan of Rincon del Camino is later than the
main valley fill and was deposited after a period of erosion.

That the area was occupied by human beings during this episode is
evident from discovery, at the place marked section 9 on figure 1, of
several hearths from 1o feet to 12 feet 8 inches below the surface
(pl. 9, lower). A short distance away two potsherds were found in
gravel lenses, but they are undecorated ware of indeterminate age. A
small sherd collection, gathered among lumps fallen from the bank of
the main arroyo but derived from the fan of Rincon del Camino, con-
sists also of fragments indeterminate as to age.

The human remains before us, therefore, do not afford reliable evi-
dence of the synchronous erosion and refilling of the post-Bonito chan-
nel and the triangular area at Rincon del Camino. However, it seems
impossible to account for the type of sand and gravel found here ex-
cept on the theory that it was deposited in an area excavated during
an earlier period of erosion. The alluvial fan of Rincon del Camino
presents general evidence, if not conclusive proof, of a period of ero-

NG. 7

GEOLOGY OF CHACO CANYON—BRYAN

sion and sedimentation contemporaneous with the cycle postulated
from a consideration of the post-Bonito channel.

The sequence of geologic events in Chaco Canyon, if the post-
Bonito episodes of erosion and sedimentation be accepted, may be
summarized as follows:

Recent geologic events in Chaco Canyon

Process Event Date
EiGFOSION Myates Werte se *‘ormation of existing arroyo 1860 to present
system
Sedimentation .... Filling of channels After completion and per-
haps after abandonment of
Pueblo Bonito
ID ROSOY Hae HeAG BES Formation of a main arroyo Probably post-Bonito
for full length of canyon
and formation of tributary
arroyos
Sedimentation ....1. Deposition of upper 4-ft. Pueblo III period

zone transitional with
lower zone

2. Deposition of zone from
21 to 4 ft. below sur-

Pre-pueblo (Pit House and
earlier?) period

face
3. Deposition of unknown Unknown, probably post-
amount Pleistocene

Possible alternations of erosion and sedimentation not as yet differentiated.

Unknown, probably Pleis-
tocene

rOStOM A. Vda ser. Formation of canyon

The twofold character of the valley fill is thus well established. The
buried channel has been traced from near Una Vida to a point about
14 miles below Pueblo Bonito where it becomes so large that rem-
nants of the main valley fill cannot be identified.

BURIED CHANNELS SIMILAR TO THE POST-BONITO CHANNEL
ON OTHER STREAMS

Discovery of the post-Bonito channel led to a search for a similar
division of the flood plains of other streams in northwestern New
Mexico. The data obtained are here summarized although it is recog-
nized that the importance of the subject requires additional field study
and more complete description.

In 1925 buried channels were detected in the walls of Arroyo
en Medio and Arroyo Cedro, both tributaries of the Rio Puerco by
way of Arroyo Chico. In 1927 a well-defined buried channel was

38 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

found and mapped for half a mile in the arroyo of Rio Puerco, be-
tween the towns of Cuba and La Ventana. In neither of these local-
ities were potsherds or other human relics found in the deposits of
the channel. On Rio Puerco I picked up two sherds of black-on-white
pottery that clearly had fallen with a portion of the arroyo bank, but
the exact position in which they were deposited is not known. They
merely indicate that part of the valley fill was laid down during a time
when prehistoric Puebloan peoples inhabited the valley.

In 1929 buried channels were also discovered in the floors of Ti-
jeras and Coyote Canyons on the west slope of Sandia Mountains.
In Coyote Canyon many evidences of human occupancy during dep-
osition of the valley fill were found. Hearths, bones, charcoal, and
potsherds occur but the localities are so disposed that none provided
the critical data that would date the channel.

In 1928 a buried channel was discovered on Galisteo Creek at the
town of Galisteo, in Santa Fe County. A similar one occurs west of
the road crossing and rock falls on San Cristobal Creek, just below
the pueblo of San Cristobal. The channel on Galisteo Creek was care-
fully traced and mapped. It obviously underlies and therefore must
be older than Pueblo Galisteo (Bryan, 1941, p. 231). This particu-
lar ruin (Nelson, 1914) contains potsherds which are all of glazed
types and no part of it appears to be older than the Pueblo IV period,
comparable to Pecos Glaze II (personal communication from A. V.
Kidder).

GEOLOGIC EVIDENCE ON THE MEANS OF LIVELIHOOD OF
CHACO CANYON PEOPLES

All peoples known to have occupied Chaco Canyon in prehistoric
times were dependent for sustenance largely upon agriculture. The
Navaho now living there are principally stock raisers, but nearly all
of them plant fields of a sort. Two small patches of corn were to be
seen near Pueblo Bonito in 1924; farther upstream, beyond Pueblo
Pintado, fields were larger and more numerous. The Bonitians, how-
ever, lacked domestic animals and although the hunting in their day
may have been better than it is now the major part of their food sup-
ply doubtless came from cultivated plots.

As pointed out heretofore, the present climate of the Chaco area
is unfavorable to agriculture by reason of the unreliability of rainfall,
particularly in June and July, and because of the comparatively short
growing season. Elsewhere irrigation makes possible the growth of
most crops common to the temperate region; corn, beans, and squash,
still staples of Indian tribes throughout the Southwest, are known from

NO. 7 GEOLOGY OF CHACO CANYON—BRYAN 39

frequent finds in excavations to have been the main crops in prehistoric
times. During the course of this investigation we were constantly on
watch for evidence of local irrigation with living water—and found
none.

The floor of Chaco Canyon has no irregularities that are not en-
tirely natural in origin except wagon roads, mounds covering ruins,
and an ill-advised ditch built some years ago by a white man. The
banks of the modern arroyo were carefully inspected for traces of
ancient irrigation ditches but none was found in the main valley fill.
Such negative evidence would be of little value were it not a fact that
in every country in which living water is used for irrigation the ditch
banks are routes of travel and are thereby compacted. It is inconceiv-
able that, if ditches had been used by the ancient inhabitants of Chaco
Canyon, all their compact ditch banks would have been destroyed,
especially when the processes of alluviation were so gentle that char-
coal in hearths was buried with little disturbance. Under such condi-
tions the hard-packed surfaces of ditch banks should also have been
preserved.

Materials of the valley fill, as already described, do not indicate that
there ever was a stream of living or perennial water in Chaco Canyon
during the two periods of alluviation determined by these investiga-
tions. Such materials all appear to have been laid down during muddy
floods similar to those now characteristic of Chaco River. Irrigation
by means of a system of ditches continuously maintained is hardly
practical on such a stream. Thus geologic interpretation confirms the
negative evidence of the ditches and indicates that the prehistoric
peoples of Chaco Canyon did not practice irrigation as we commonly
understand it. However, in selected places they could have farmed
successfully by irrigating with flood water or, as it is usually called,
“floodwater farming.” For this method the floor of Chaco Canyon,
save for the presence of its modern arroyo, is entirely suitable.

FLOODWATER FARMING

The ordinary rainfall throughout most of the southwestern United
States and northern Mexico is insufficient to grow crops. Such lands,
however, as are overflowed by the muddy water of ephemeral
streams or by rainwash from hillsides will support a hazardous agri-
culture. Floodwater farming was first adequately described by Gregory
(1916, pp. 103-105) and a rather complete account with maps of
fields has recently been published by the present writer (Bryan, 1929).
The additional data given herewith apply particularly to conditions
that once obtained in Chaco Canyon.

40 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

Gregory (1916, pp. 104-105) outlines, as follows, the methods and
results of floodwater farming in the Navaho Country, an area that
includes Chaco Canyon:

From experience and tradition the Indians have learned to know the areas
liable to be flooded during occasional showers as well as those annually inundated
by the successive rains of July and August. Along the flood plains of the larger
washes the practice is to plant corn at intermediate levels in widely spaced holes
12 to 16 inches deep. The grain germinates in the sand and rises a foot or more
above the surface before the July rains begin. With the coming of the flood the
field is wholly or partially submerged. After the water has receded parts of
the field are found to have been stripped bare of vegetation and other parts
to have been deeply buried by silt; the portion of seeded ground remaining
constitutes the irrigated field from which a crop is harvested. The Hopis, and
to a less extent the Navajos, sometimes endeavor to direct the floods and to
prevent excessive erosion within the fields by constructing earthen diversion
dams a few inches to a foot or more in height—dams which require renewal
each season. . . . Much work is done by the Indians while the flood is in prog-
ress, and an everyday sight during showers is the irrigator at work with hoe or
stick, or even with his hands, constructing ridges of earth or laying down sage-
brush in such a manner as to insure a thorough soaking of his planted field.
By these methods of flood irrigation the Navajo and Hopi together cultivate
about 20,000 acres of land widely distributed over the reservation in fields about
3 acres in average size, rarely exceeding 200 acres.

Similar methods of cultivation are in wide use throughout the pla-
teau of Mexico and in New Mexico and Arizona (Hoover, 1930, pp.
437-438 ; Hack, 1942, p. 26). The Mexican calls such a field sombrado
(planting) or temporal (temporary field). The Nahuatl word milpa
is also used. In favored localities and usually at high elevations these
terms are applied to fields dependent on rainfall alone but generally
flood irrigation from the rainwash of higher slopes or from ephemeral
streams is essential to a crop.

Since they were first visited by whites in 1698 and doubtless long
before, the Papago Indians of southern Arizona have supported them-
selves largely by floodwater farming pursued on the broad plains of
their undissected desert valleys. Corn, squash, and tepari beans are
their main crops (Lumholtz, 1912; Bryan, 1925b; Hoover, 1929).

Floodwater farming for the production of stock feed is widely prac-
ticed in northern Nevada. Quinn River valley (Bryan, 1923b) may
be taken as typical. The valley—
is bordered on the east by the Santa Rosa and Buckskin mountains and on the
west by the Quinn River Mountains. It heads in Oregon about Io miles north
of the Nevada line and extends southward about 45 miles to the Slumbering
Hills, which separate it from the broad valley of Humboldt River. The valley

is 10 to 12 miles wide and is drained by Quinn River, which runs southward
through it for about 4o miles, turns west, and, passing south of the Quinn River

NO: 7 GEOLOGY OF CHACO CANYON—BRYAN AI

Mountains, is lost in the Black Rock Desert. The river is formed about 4 miles
south of the Oregon line by the union of East Fork of Quinn River and McDer-
_ mitt, Washburn, and other creeks. Its drainage basin includes about 1,164 square
miles.

The valley floor consists of plains formed of beds of sand, gravel, and clay
deposited by the existing streams. A small part of the valley bounded by a
line of bluffs that extends from the Oregon line east of McDermitt southeast-
ward to the National mine differs from the larger part in that the streams flow
in flat-bottomed valleys that lie 50 to 100 feet below the level of sloping plains
formed by the same streams at an earlier time, when they flowed at a higher
level. In general, however, each stream, on leaving the mountains, wanders
through circuitous and branching channels over the alluvial slopes to the axial
flat, where it joins in grassy meadows the small meandering channel of Quinn
River.

The region is arid, none of the streams containing water throughout the year.
The principal streams carry considerable water or are in flood in the spring,
when the snow on the mountains is melting. The spring floods are not violent,
and the water, which may be almost clear, is easily diverted into semi-permanent
ditches to irrigate the cultivated fields. In the axial flat there are large fields
that are irrigated from Quinn River or from its tributaries. Near the mouths
of the canyons of the principal mountain streams there are smaller fields,
many of which are irrigated by the water of streams that seldom reach the lower,
larger fields. About 14,000 acres is irrigated. Native grasses, which are used
both for pasture and for hay, form the principal irrigated crops, but there are
also fields of alfalfa and small grain.

Paradise Valley lies east of the Santa Rosa Mountains and is enclosed on the
north and northeast by unnamed volcanic plateaus and mountains and on the
east by the mountain range that culminates in Hot Springs Peak. It is drained
by Little Humboldt River, which is formed by the union of Indian, Martin, and
other creeks with the east fork of Little Humboldt River. The topography of
this valley is like that of Quinn River Valley, though the central flats are wider
and the area irrigated is larger.

In both vaileys communication with centers of population elsewhere is difficult.
The distance from some of the ranches to Winnemucca, the nearest railroad
station, is more than 60 miles. Only cattle on the hoof can be readily marketed,
and the irrigated land is devoted to the raising of stock-feed. In April and May
the cattle are turned loose in the plains and lower foothills, where they browse
on the sage, weeds and grass, and as the snow melts they gradually climb higher
into the hills, reaching the summits in midsummer. As the cattle leave the
valleys the ranchers begin to irrigate their land, starting with the first floods
and continuing to use the water as long as it lasts. In August great quantities
of hay are made from the native grass and from alfalfa.

The cattle begin to come down from the hills late in August and early in
September—according to the local saying, “as soon as they hear the mowing
machines.” Late in September and during October the ranchers bring the last
of the cattle out of the hills to the owners’ fields, where they are pastured on
the still green native grass and on greasewood until it is necessary to feed them
hay. In ordinary years the cattle are brought through the winter in excellent
condition.

42 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

About 13 miles east of Pueblo Bonito and beyond Pueblo Pintado,
small Navaho fields are cultivated in the floors of valleys more or less
obstructed by sand dunes. The effect of sand dunes in spreading the
floods of ephemeral streams and in providing localities suitable for
floodwater farming is also of large importance in the Hopi country
(Hack, 1942). In canyons tributary to Arroyo Salado, in the Rio
Grande drainage and some 70 miles east of Pueblo Bonito, there is
limited floodwater farming. The Cafiada (de) las Milpas doubtless
had, before the existing deep arroyo was formed, enough fields to
justify its name. Here, as late as 1922 on hill slopes at Juan Chaves’s
ranch, there were two small fields irrigated by the runoff from high
bluffs. In 1921 the writer saw in Rincon de Lopez 40 to 50 acres of
corn and beans that had been irrigated by floodwater.

About 10 miles from Cafiada las Milpas is Bernalillito, a locality
so called because people from Bernalillo on the Rio Grande 35 miles
distant formerly moved there each summer to plant and tend their
fields. Bernalillito Wash heads on the eastern escarpment of the Mesa
Prieta and flows in a broad flat valley eroded in shales and sandstones
about 10 miles east to a narrow gap in a massive sandstone cliff.
Below this cliff there is a canyon in places broad and in others narrow.
About half a mile below the cliff is the single ranch house of Bernalil-
lito. Nearby is one field of 15 acres suitable for planting corn and
beans and to one side a smaller area where hay may be cut. The
sandy alluvial fan of a tributary gulch causes the main flood to spread
and the gentle overflow makes possible the cultivation of the field and
the growth of grass on the meadow. The flood is also caused to spread
more widely by a low dam of brush and stone (atarque) at the lower
end of the field. Just below field and meadow is the headwater falls
of an arroyo which has already dissected similar flats in the canyon
below and now threatens to destroy the remaining fields. In 1920 this
ranch came under the control of H. M. Bryan, who has planted with
the results shown in the table on the following page.

The years 1920 to 1923 were generally unfavorable in this part of
New Mexico and similar planting in years of greater rainfall might
have produced better results. However, the flood of September 1921,
which broke down and washed out part of the corn, is an incident that
might happen any year. When this flood arrived, a bean crop esti-
mated at 1,000 pounds was ready for harvesting but only 200 pounds
—the equivalent of the seed planted—were saved. Similarly, the loss
from intruding cattle that broke through the outside fence to obtain
water in 1920 and 1922 might have been avoided. Such hazards are,
however, a part of this type of farming although the prehistoric flood-

PLATE 10

Upper: The Chaco in flood. Tents of a General Land Office surveying party at left; men on
“opposite bank and rider at right stand on approaches to the wagon crossing. Pump on Wetherill’s
well, destroyed a few weeks later, is seen above the fourth horse from the right. (Photograph

by Neil M. Judd, 1921.)

Lower: Middle south wall of Pueblo del Arroyo with a partially refilled section of Jackson’s
“old arroyo” in the foreground. (Photograph by Neil M. Judd, 1920.)

PLATE II

Left: A pit house
with floor 13 feet
6 inches below the
present surface, 9
miles east of Pueblo
Jonito, had been
long forgotten be-
fore the Pubelo IIT
village owas con-
structed on an upper
level a few feet dis-
tant. (Photograph by
Neil M. Judd, 1926.)

Right: An Early
Pueblo cooking pot
covered with a sand-
stone slab lay 6 feet
below the surface of
a sand dune in Wiri-
to’s Rincon, about 14
miles southeast of
Pueblo Bonito. ( Pho-
tograph by Neil M.
Judd, 1927.)

GEOLOGY OF CHACO CANYON—BRYAN 43

NO.

jo juso10d Q¢ noe ssv1B jo puvys & Iva yey} JO [[e} OY} UT “SSCIS 0} JIDADI 0} PodeInooUs SBM PUL] 94} PUR

zI pue
6 "PO ‘oI pue ‘SI ‘g ‘sny ‘spoo],y

JNIYJIP SuIjsaAtey apeur
Jaquiejdeg pue jsnsny Ul spooy sa14 ],
Ieak a]qetoa
-e} Ing ‘pazeanpno Ajs1ado1d jou uso)
pe}a][09 sseulep
doid oo1¢ ‘ajyeo Aq uayea sem
}I jnq }1e}s poos usOD dAes pooy sunf
2OPPOy
UJOD pekOljsap pue suevaq Jo spunod
00g pofoljsap Joquia}deag Ul poo,y
Surueyd ou ‘tea poor)
Suueyd ou ‘ seak poor
SYICUIIY

painjseg
S[Si?
uvISSNy

pue Avy suo} g

suo} Z

ouON

peinjsed
Mey
dUON

SuOo} OI

Aey
JeinjeNn

powered jON
dois
poyedxe

Jo yyey noqy

oUON

suo NT

spunod 002
dUuON
poqueyd oN

suvog

peed jon

Aey
Ad] req suo} z

Jappoy ut0d

aUuON

JOPPOF
UIOD SU} S
uO N
pojueyd JON

ase10
peddossy

(uetig "JW ‘H Aq peystuiny eyed)
Cz6I 04 O16I “xa ‘N ‘071pyjyoUusag yo Ssdosad

*‘pauiryqe Useq pey purys [[ny

poqueyd JON’

do.19 [Jews ***

pesoyyes
sjeysngq oS °°

S182 poos Ma,J**

megq**
ouON”’
pojuejd }ON**

uloy

auop sem Suljuejd ou $z61 ujy—"a1ON

“SzO1

‘PzO1

*€zZ01

*zzOl

*Iz61
*OzOI
“O101

vox

44 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. I22

water farmers were saved at least the risk due to domestic animals. It
is true also that this piece of ground was poorly farmed during these
years. The owner at the time was a sheep man whose herds during
the summer season were in the mountains more than 60 miles away.
Consequently the farming was done during intervals of other, more
important work, and by hired labor. Cockleburs and Russian thistle
grow so rapidly that strict attention to cultivation is necessary. Be-
cause this attention could not be given, the owner converted the plowed
land into meadow with the intention of cutting hay for winter use.

The record of these six years indicates that, with an adequate re-
serve for lean years, this field would support a family requiring no
more than was necessary for prehistoric Pueblo peoples. But this
family would need to give the crop the constant attention which
modern Hopi lavish on their own fields.

It is obvious that the condition of aggradation that once prevailed
in the valleys and canyons of the Southwest was favorable to flood-
water farming. At that time the floods spread widely and were con-
fined to channels at only a few places. Small localized showers that
fell on the walls of canyons, if sufficient to produce runoff, flooded the
adjacent flat floor, whereas now the runoff flows into an arroyo below
the general level of the plain.

Water that soaked into the ground could not so easily drain away
and consequently underlay the valley floors at moderate depths. It
was thus available to supply the roots of trees and bushes which doubt-
less once flourished in valleys where they are now absent. At favor-
able places, lakes and swamps existed in which grew the rushes (tule)
frequently found in prehistoric ruins. Because of this higher water
table a thinner layer of earth was dried out between floods; smaller
floods were sufficient to wet the dry ground. These smaller floods
also covered larger areas since less of their flow was absorbed by the
dry ground. The result of such a regimen was a denser vegetation on
the canyon floors.

Of this denser vegetation an important part was perennial grasses.
Even today where floods spread widely over the floors of undissected
tributaries to Chaco Canyon there is a fairly dense cover of perennial
grass, usually alkali sacaton. Meadows of this type were so numerous
in the Rio Puerco region, 80 miles to the southeast, that a principal
occupation of the Mexican inhabitants in the late 1890’s was the
cutting of hay to be hauled 30 miles or more to market in Albuquerque.

The presence of these meadows was advantageous to white men,
and their destruction by the formation of arroyos has been a distinct
loss. The forage they were capable of producing naturally was of no

NO. 7 GEOLOGY OF CHACO CANYON——BRYAN 45

value to prehistoric farmers without domestic animals to feed, and
their agricultural potentialities remained untested because those same
prehistoric farmers lacked tools adequate to the task of uprooting
sod. At the present time grass is apt to grow thickest on the more
clayey soils and, to the extent that they are effectual in spreading flood-
waters, grassy areas tend to postpone arroyo cutting. The inhabitants
of Pueblo Bonito undoubtedly planted on the sandier soil bordering
such grassed areas, on alluvial fans below the cliffs, and along the
vague and meandering course of the main stream. Such soils were
more or less disturbed by successive floods but, despite annual loss of
part of the planted crop, they were doubtless recognized as the best
agricultural land because of their superior tilth and relative freedom
from deep-rooted perennial grasses.

In 1921 Mr. Judd dug a trench 20 feet in depth to study the stratig-
raphy of the west refuse mound at Pueblo Bonito. Four years later
when he extended it out into the flat fronting the mound, the trench
cut across several obviously artificial canals or ditches. They ran par-
allel to the front of the refuse mound and, therefore, essentially east
and west. They were from 4 to Io feet wide and were enclosed on the
downhill (south) bank by walls of slushed mud (adobe) laid with
care and in places supported by the dumping of house refuse. Filled
with both fine and coarse materials, including Late Bonito potsherds,
these ditches presumably had carried floodwaters from upcanyon, per-
haps from the rincon of Chettro Kettle, to fields west of Pueblo Bo-
nito. As each filled up or was washed out, a new one was constructed
along the same route. Such more or less temporary ditches for the
spreading of water are fairly common features of floodwater fields
in New Mexico. If these before Pueblo Bonito were more elaborately
constructed than is usual, it is doubtless because they lay close to the
village where labor was readily available. The data Mr. Judd has
gathered relative to comparable structures for control of floodwaters
in Chaco Canyon will be presented by him in connection with the
subsistence problem of the Bonitians.

ARROYO FORMATION AND FLOODWATER FARMING

From the foregoing it should be evident that in Chaco Canyon con-
ditions were favorable for floodwater farming from the beginning of
alluviation. The area available also gradually increased as the plain
widened with the filling of the canyon. The canyon, therefore, afforded
a locality for the initiation and development of a civilization based on
agriculture. That this agriculture was precarious and that crops might
fail owing either to lack of rain and consequently of floods, or to the

46 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

occurrence of unusually cold seasons, does not constitute a factor that
would prevent the development of a local community of relatively
high culture. These hazards merely limit the number of people and
fix their standards of living. In pre-Spanish times the Hopis are said
to have insured themselves against crop failure from these causes by
the storage of a year’s supply, or more. A like necessity may have
been one of the compelling causes which led to development of more
houses and a house cluster more elaborate than the simple pit dwelling.

The same comments may be applied to the problem of water. Lack
of a local water supply does not prevent the use of suitable floodwater
fields, for at the present time the Hopi farmer may cultivate tracts
10 or even 20 miles from his home. Moenkopi, 40 miles northwest of
Oraibi, was a farming community when seen by Ofiate in 1604 and
remained so until the Navaho forced its temporary abandonment about
a century ago. An alluvial plain in the process of sedimentation always
has local depressions which fill with water after rains or floods. These
pools or charcos (Bryan, 1920) afford a limited water supply during
the growing season. The Papagos of southern Arizona were entirely
dependent on charcos during the planting and harvesting of their crops
until, within the past few years, the United States Indian Service
drilled wells at the fields. Before coming of the Spanish, these Indians
not only cultivated the same fields while dependent on charco water
but carried the crop on their backs to winter residences located at per-
manent, or at least longer lasting, water many miles distant. Thus
lack of permanent water is not necessarily a hindrance to floodwater
farming, although it may be an obstacle to permanent residence.

In Rincon del Camino one of the Navaho workmen employed by
the Pueblo Bonito Expedition developed a small spring in a rock shel-
ter or niche under a cliff. At other similar localities on the north side
of Chaco Canyon the rock is damp and covered by an efflorescence of
salts. Here it might be possible to develop water by systematic dig-
ging, and springs of this type may have been the principal source of
domestic water to the ancient people. It is certain that a very slight
increase in rainfall over the present annual average would produce
springs in such localities.

Conditions of alluviation lead also to a relatively high water table.
At present, water may be obtained by digging about 10 feet below the
bed of the arroyo, or some 40 feet below the plain. Before the modern
arroyo was cut the stream ran in a shallow channel and ground water
must have saturated the valley fill to within 10 feet of the surface, a
distance comparable to the present depth of water below the bed of the
arroyo. As the dry season advanced, the prehistoric peoples may have
scraped holes in low places and thus formed a primitive type of well.

NOS 7 GEOLOGY OF CHACO CANYON—BRYAN 47

No evidence of such wells has been found but the digging of them
would have been entirely feasible for the inhabitants of Pueblo Bonito
even though they lacked metal tools.

The formation of an arroyo similar to the present one, or the one
we have called the post-Bonito channel, confines floodwaters to a nar-
row belt below the level of the plain. The ground-water level is also
lowered and floods from tributaries are less effective in wetting the
ground. Farming by means of floodwater is consequently impossible
over the whole plain and is limited to insignificant areas.

If, therefore, the geologic chronology tentatively outlined on page 37
be accepted, Chaco Canyon enjoyed a relatively long period of allu-
viation with conditions favorable to floodwater farming. As indicated
by finds of potsherds and other relics, people of Early Pueblo culture
occupied the valley at least during the time required to build the fill
from a level 21 feet below the present surface to 4 feet below. The
favorable conditions then existent led to development of the relatively
complex civilization of the Great Pueblo period, a culture that flour-
ished during deposition of the upper 4 feet of valley fill.

Formation of the arroyo system represented by the post-Bonito
channel may be given as the approximate cause for abandonment of
the valley by these Pueblo III people, although other factors such as
war, invasion, disease, or gradual decrease in means of subsistence
may also have had their effect. However great the changes these other
factors might have produced, it seems unlikely that any one or two
of them would have kept out of use for long a place so eminently
suitable for floodwater farming as Chaco Canyon.

CAUSE OF ALTERNATE EROSION AND SEDIMENTATION

If, as outlined in the preceding pages, the alternate erosion and sedi-
mentation in Chaco Canyon resulting in the production of a plain of
alluviation suitable for farming at successive times and the dissection
and partial destruction of this plain in the intervening intervals is an
adequate cause for the rise and destruction of human cultures, the
ultimate reason for changes in the habit of streams becomes of large
importance. A river or stream deposits sediment when it has a load
greater than it can carry on a given grade. It erodes when it can carry
more material than is furnished to it. The quantity of water, varia-
tions in this quantity, grade, supply of sediment, size of grain of the
sediment, and shape of the channel are factors which determine the
habits of a stream. The complex interrelationships of these factors
are difficult to determine quantitatively, and are summed up in the

48 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. I22

word “regimen.” The regimen of streams in the Chaco Canyon region
is such that they now erode whereas formerly they deposited material
on the canyon floors.

As this change in regimen is general, general causes must be sought.
A number of writers © attribute the present epicycle of erosion to in-
troduction of livestock. Formation of trails and destruction of vege-
tation by overgrazing are supposed to have concentrated floodwaters
and to have allowed these floods to erode the present channels. The
argument of these writers is best expressed by Duce (1918, p. 452):
“We may, therefore, summarize the effect of cattle by saying that they
increase the rapidity of the run-off and the rate of erosion by destroy-
ing vegetation, by compacting the soil, and by forming channels for
the passage of water.”

The apparent coincidence in time of the initiation of overgrazing
and the beginning of dissection is significant and more data on this
phase of the problem are needed. Coincidence of settlement and the
deepening of the channel of Rio Puerco seems well established (Bryan,
1928a). The theory that erosion is caused by a slight uplift and in-
crease in the gradient of streams has been generally rejected because
erosion of about equal magnitude affects streams of different drainage
systems that flow in all possible directions. Uplift so nicely adjusted
to the drainage pattern is inconceivable.

Huntington (1914, pp. 33-34) was doubtless the first to advocate a
climatic change as the cause of erosion, and Bryan (1923a, pp. 77-80)
has brought together available evidence that a slightly wetter climate
was characteristic of southern Arizona at the time of the first Spanish
explorations.

Gregory (1917, pp. 131-132) advocates climatic change and rejects
the argument for overgrazing in the following words:

It is important to note in this connection that the balance between aggradation
[alluviation] and degradation [erosion] is nicely adjusted in an arid region
where the stream gradients are steep, and that accordingly small changes in the
amount of rain, its distribution, or the character of storms and changes in the
amount and nature of the flora result in insignificant modification of stream habit.
Even the effect of sheep grazing is recorded in the run-off, and this influence
combined with deforestation has been considered by many investigators as the sole
cause of recent terracing in the Plateau province. For the Navajo Country these
human factors exert a strong influence but are not entirely responsible for the
disastrous erosion of recent years. The region has not been deforested; the
present cover of vegetation affects the run-off but slightly, and parts of the
region not utilized for grazing present the same detailed topographic features as
the areas annually overrun by Indian herds.

5 Dodge, 1910; Thornber, 1910; Smith, 1910; Rich, 1911; Duce, 1918; Olm-
sted, 1919.

NO. 7 GEOLOGY OF CHACO CANYON—BRYAN 49

Reagan (1922, 1924a, 1924b) presents the interesting hypothesis
that prehistoric peoples, by means of small reservoirs, check dams, and
embankments designed to spread floods, used or distributed most of
the water so that floods of the main arroyos were decreased in volume
and violence. Hence, sediment was deposited and the arroyos filled up.
When these ancients disappeared and their structures fell into decay,
erosion was resumed. Reagan also supposes that the prehistoric peo-
ples killed off the game, and thereafter the vegetative cover increased
until it gave a maximum of protection to the soil. This cover was
destroyed with introduction of domestic animals after the Spanish
conquest. Thus Reagan assumes that the structures built by prehis-
toric men were sufficient to reduce floods without complete considera-
tion of the difficulties involved. He ignores the fact that alternation
between erosion and alluviation began before entrance of the Puebloan
peoples into the area. The earlier epicycles could not have been
influenced by such causes as Reagan advances.

Olmsted (1919, p. 88) estimated the cost of check dams and bank
structures for control of Gila River above the San Carlos dam site at
a total of $6,401,029. Elaborate as these plans seem, engineers by no
means agree that they are adequate to control floods on Gila River,
much less to restore channel conditions to those obtaining before 1880.
In recent years the Soil Conservation Service has built structures for
control of erosion which exceed in magnitude anything possible to the
ancients. Success has been rare and modest. That primitive man could
erect barriers sufficient in number and size to accomplish this result
seems improbable. If such attempts had been made, remnants should
remain to be easily identified.

All the investigators mentioned above considered that they were
dealing with only one period of sedimentation and one period of ero-
sion, although Huntington thought these minor changes of the geo-
logical Recent were the latter part of a series of such changes running
back into Pleistocene time. If the post-Bonito channel represents a
valid cycle of erosion and sedimentation then three complete local cy-
cles of erosion and sedimentation and part of a fourth must be
explained. These events may be put in tabular form, as follows:

Cycles of erosion and sedimentation in Chaco Canyon

mst evele; . Si.)2 Fh. Erosion of canyon or later Sedimentation and formation
part of this erosion of terrace
BEVEL ie tos pcs. 8 Erosion of terrace Sedimentation and formation
of main valley fill
1 o/c eee Erosion of post-Bonito Sedimentation and fill of
channel post-Bonito channel

SAS CVCIC 85.5 573) 0.'s 5 0 Erosion of present arroyo Future?

50 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

Presented in this form it seems evident that a postulate of climatic
change has greater inherent possibilities as a true explanation of the
facts. Domestic animals certainly cannot be charged with inception of
the post-Bonito erosion. However great the influence of overgrazing,
therefore, it must be regarded as a mere trigger pull which initiated
an epicycle of erosion that was brought about by other causes.

Reagan (1924a, p. 285) has carried the factor of overgrazing into
prehistoric times and suggests that incoming and increasing hordes
of herbivorous animals may have overgrazed the country and thus
caused formation of arroyos. Thereafter, the animals having starved
to death or left the region, vegetation would again spring up and suffi-
ciently protect the land surface so that streams would again begin to
aggrade. Thus he postulates recurrent overgrazing with consequent
cycles of erosion and aggradation. That animals in a state of nature
would overgraze an area is an assumption without proof. It seems
best to pass this interesting postulate since there exists ample evidence
that at least one period of sedimentation in the not-distant past was
wetter than the present.

Douglass * has described briefly the valley deposits of the Rio
de Flag, a small creek near Flagstaff, Ariz. Here an arroyo dating
from 1890 to 1892 has dissected a valley fill; standing stumps of pine
trees are found from 4 to 16 feet below the surface and prostrate logs
in the upper 4 feet. Both stumps and logs belong to the living species,
Pinus ponderosa, now growing on the adjacent hillside. The stumps,
however, have wide growth rings similar to those found in trees of
humid:lands. The prostrate logs have narrow rings like living trees
of the region. It seems evident, therefore, that the zone from 4 to 16
feet below the surface was deposited under a climate much wetter
than now. In addition, human relics have been found in the fill at
depths from 4 to 9 feet but their relation to the stumps is uncertain.
Evidence in the valley fill of the presence of man seems to indicate
that the humid period demonstrated by the stumps is not very ancient.
It may represent one of the cycles of sedimentation disclosed in Chaco
Canyon or an older cycle not yet identified there.

Study of the Chaco Canyon deposits has not produced incontestable
evidence that wetter climates prevailed there in the past. From the
main valley fill—2d cycle of the table, page 49—we collected a few
fresh-water shells but similar shells were also found in the post-Bo-
nito channel. In some of the sandy and silty beds of the main valley
fill impressions that resemble rushes were noted. The adobes are dark

6 1924, pp. 238-239. This reference supplemented by an oral communication,
and by inspection of the locality by me in September 1921.

NO. 7 GEOLOGY OF CHACO CANYON—BRYAN 51

brown from included organic matter, but this might have derived
from a heavy growth of grass or other vegetation.

There is, however, nothing in the character of these sediments that
precludes their deposition under a slightly wetter climate. It is pos-
sible there may have been a sufficiently greater rainfall so that pine
trees grew in favored places on the hills, cottonwood and willows may
have bordered the river, the valley floor may have had large areas of
perennial grass and in a few places there may even have been marshy
ground with cattails. Currently existing damp places under the north
cliffs may have been small springs in times past. Such an environment
seems compatible with the type of alluvium deposited and yet suffh-
ciently favorable to have provided an adequate food supply for the
peoples of the thirteen villages of Chaco Canyon. Such a climate can-
not be inferred from the nature of the sediments themselves but, on
the other hand, those sediments present no evidence that a relatively
slight modification of climate did not exist when they were being
deposited.

[Norr.—Bryan’s speculations in the foregoing paragraphs have
been substantiated in large measure by data from our excavations and
from further exploration. Our old Navaho neighbors reported pine
stumps at various places about the valley; dead and prostrate pines
were photographed on the south cliff and a couple of dozen trees and
stumps were seen at the head of the canyon, 16 miles to the east
(Douglass, 1935, p. 46). These last few remnants of former forests
suggest that the annual rainfall in their time was considerably greater
than our postulated 10-inch average for the present. Even more sug-
gestive is the fact that thousands of logs, large and small, went into
construction of Pueblo Bonito and its neighboring villages between
A.D. 900 and 1100. The forests that furnished those logs must have
been close at hand since none of the timbers we uncovered was scarred
in transportation and such forests, at 6,500 feet, could have flourished
as they did only in a climate somewhat wetter than that of today. In-
deed, many of the old ceiling timbers from Pueblo Bonito exhibited
growth rings so uniform in width they obviously had grown where
moisture was fairly constant year after year. In addition, we know
that rushes were then abundant and readily accessible, for quantities
were utilized in the building of Pueblo Bonito.—N. M. J.]

DETAILED SECTIONS IN THE RECENT ALLUVIUM
SECTIONS IN THE MAIN VALLEY FILL

The following sections, measured at various places along the main
arroyo, record the character and thickness of the several strata from

52 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

the top of the bank to its base, and give a record of human relics found
therein. Each station is indicated on the accompanying map, figure 1,
by its corresponding number.

Section I

South bank of main arroyo, 200 yards east of expedition camp. No
potsherds were recovered in this section but opposite, in the north
bank, plain ware of coarse texture was found at a depth of about 12
feet.

Thickness Depth
Feet Inches Feet Inches

Laminated dark clay and sand, more clayey above 4 ) 4 (0)
Sil withistrealssioteclayaremiceeicesve terse tees 3 fo) vi O
Spratly SARS ate ieee RI aie 0 8 aoe a 2 () 9 0)
Eaminatediidanle silts acne nace ei eiowed ue (0) 6 9 6
Silt and sand with thin streak of clay 18 in. above

POSE Se tecvareiittas ces cee otes RAV CCE, Sh aletatoits cel areePS stata 3 4 12 10
Darkiclayawatheshellssanpraseciemencemee meee (0) 8 13 6
Sarre erway: ace te tetas cr pa ersaniis raise orsse tea sterater sake & 0) 16 6

Covered below.
PIT HOUSE

At the place on the map marked “Pit House” there was discovered
in 1922 a structure partly destroyed by erosion of the arroyo bank.
What remained was excavated and studied by Neil M. Judd, who has
described it at length (Judd, 1924). It consisted of a single circular
room with the middle of its slightly concave floor 12 feet 2 inches
below the present surface. The original excavation was 6 feet deep
(pl. 6, lower). The builders of this subterranean house lived when the
flood plain was 6 feet lower than it now is, and doubtless they or their
contemporaries are responsible for the human relics at deeper levels
shown in the sections to follow.

Section 2

In a small tributary entering the main arroyo from northwest of
Pit House.

Thickness Depth
Feet Inches Feet Inches
Soilvand ‘clay. moneysandy above aces - I 10 I 10
Sati reac otis nero ee eis) cnetere raves erehiie Gia Rev eaanatetccen iets oisite (0) 2 2 (0)
Clay ae eink chts Re ites Cole) tel ice me eclosion a (6) 2 2 2
SElins AR ontop cose a abetted brie att Sh a oan coy ea ie Oo 7 2 9
Clay acisictteee oes aa Ree SR a toe eee Lees I 7 4 4

Hard compact sandy clay with fragments of char-
COAlc and! SxpOeSHELdS ters vera ae acer I 4 5

ie)

NO. 7 GEOLOGY OF CHACO CANYON—BRYAN 53

Thickness Depth

Feet Inches Feet Inches
Clay with scattered bits of charcoal and with fire-
place at base marked by crescentic streak of
charcoal and burned ground below........... (e) 5 6 I
VIC ODOM nye a ctelt ca.e ore alcceuhe seh. cieeseheusre wistaie share 2 (6) 9 I

Near section 2 and in the same small arroyo, there is a ruined house
now almost destroyed by erosion. Its foundations reach to a depth of
8 feet below the surface, yet the walls appear to be quite similar to
masonry of the neighboring great pueblos and the potsherds are of
both early and late types.

Two alternatives must be considered, either the pit house and this
small pueblolike dwelling are contemporaneous, which is unlikely, or,
granting that the pit house is older, the pueblo-type dwelling and
section 2 are located in a later deposit.

Section 3

Section 3, on the south bank about 300 yards east of the pit house,
represents a normal succession of beds. A potsherd at 4 feet below
the surface is of undetermined type. Sherds found between depths of
17 feet and 20 feet 6 inches, are of Pit House or older age. This is the
greatest depth at which potsherds were found.

Thickness Depth
Feet Inches Feet Inches

Soil and clay, upper 8 in. the most clayey....... 2 6 2 6
STI: cae nGHE bord dela b ORD Or Oe oa TE ene inner I oO 3 6
Alternate layers of clay and silt; potsherds 6 in.

from top of this layer and 4 ft. from surface.. 1 6 5 (a)
(CIEN? (Gis BSS Ee BU H9 ae arr RAISER ERE I (0) 6 (0)
SONG dhe OOM ASH NOS OG mI SE eC ARe Seton c Ser eee 2 6 8 6
Alternate layers of clay, and) silt. .). 2.20.0. +... I 6 10 Co)
Silt and sand laminated and crossbedded and

Ea CINOMITIEON LOWE in la VEG aieye)cleveieiors) <a stare lence) ensl 7 (0) 9 (0)
Clayey sand; much scattered charcoal, sandstone

fragments, some of which are reddened by fire,

potsherds and worked cores of quartzite and

ZEISS 64 Romo bobo OND dObO oO cdo oO dOU boo OR 22 0 19 oO
Sand, with a few sandstone fragments and pot-

sherds at bottom. Marks base of culture layer. I 6 20 6
Sandy adobe darker than that above with plant

ATTA ECSSEOLIS MES si ait sysr rat ch Meslay oucho versie. ovate cha) aever ee tans 3 5 23 II

(Govered to bedvOr ALLOYO steler oa) sleieieis,. cnausie ste ws 7 I 31 (0)

54 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. I22

Section 4

In the north bank of the main arroyo and near the “store” at south-
east corner of Pueblo del Arroyo (pl. 7, lower). The section was meas-
ured in the middle of a channel deposit which lies beneath a horizontal
clay layer. The deposit itself is 13 feet 5 inches thick at the deepest
point exposed, and its base is 15 feet below the level of the adjacent
plain. Nearby, in 1877, Jackson found potsherds, bones, and a human
skull at a depth of 14 feet, as discussed on page 32.

Thickness Depth
Feet Inches Feet Inches

(Sleigh AMER ROME a heis Sh As ore nee cree I 8 I 8
Sand, with lenses of fine gravel and clay which

dip toward center of old channel. Crescentic

lense of laminated clay; depth of lowest part

12 ft., thickness 6 in. Sandstone blocks im-

bedded in this clay like steppingstones. Gravel

layer at base. Many fragments of bones, broken

rocks, a few shells and many potsherds espe-

mally, near Wase n:8 ieee os kere sist oe ease oerae 13 5 15 I
Clay and laminated silt, scattered charcoal to

epi xot 2s Ata OMIM eek aie cress ecole creaeleterern cis 7 oO 22 I

Section 5

South bank of main arroyo opposite Ruin No. 8 and a short dis-
tance east of section 16 (see pp. 29-30). Of special note is a Pueblo
III fireplace (pl. 3, right) built when the surface was 5 feet lower
than it is now, a second hearth at 12 feet 8 inches, and charcoal at 16
feet 3 inches.

Thickness Depth
Feet Inches Feet Inches

Dark clay, a fairly continuous layer that thickens

ATI CAUIANS eos) oat eheteieters aie tavn' eseus ie eiavere vere ues urea 0) 8 (a) 8
Fine-grained sand, finely laminated and cross-

bedded, with impressions of roots............ oO 8 I 4
(8) Ee BO SABIE COO TTIG SAC nr oOncia aonte 0 2 I 6
Sand, laminated and containing streaks of char-

codliand! chunks oficlaye acces cime eone re ae I 6 3 oO

Dark clay and laminated sand. Clay is in irregu-
lar thin layers which slope from south to north
and fade out in irregular broken chunks of clay
imbedded in sand. A firebox formed of nearly
vertical sandstone slabs at depth of 5 ft. is of
Pes TEM ty pes iced detec wie etne toe ORs SER eS 2 9 5 9

NO. 7 GEOLOGY OF CHACO CANYON—BRYAN 55

Thickness Depth

(a
Feet Inches Feet Inches

Compact rusty sandy clay with fragments of
stone, potsherds, and charcoal. Potsherds are
6 ft. 3 in. from surface and of indeterminate

OTE A ee SGP OB GH eae c gto te Beit a aiete ete Oo 6 6 3
Crossbedded sand to crescentic hearth 12 ft. 8 in.
below top of section; scattered charcoal at
14 ft. 8 in. and, 50 ft. south of section, at 16 ft.

RORNCAS Ee Sia Scie ice, nia Wiuhwie’ Stale <igt epi te ele siete sale 13 oO 19 3
Covered below.

Sections 6 and 7

Sections 6 and 7 are near each other and in the north branch of the
main arroyo opposite Una Vida (fig. 1). The bedding is very irregu-
lar in this vicinity, and none of the layers listed is persistent. Section
6 has potsherds considered to be of Pueblo III type between 4 feet
6 inches, and 5 feet below the surface. Section 7 shows a very large
hearth 8 feet from the surface.

Section 6

In north bank of north channel, near Una Vida.

Thickness Depth
—— oS
Feet Inches Feet Inches
Dark clay. Potsherds at 4 ft. 6 in. and 5 ft...... 5 Ce) 5 (0)
Crossbedded sand, with lenses of clay a few feet
Olt the lineror tHe SeCtOI ss... «ives sels cleversis 3 6 8 6
Pe eRe eat ink Fert s ie cicie 6 aka alo wiel dats aie wea Miauassie I 6 10 oO
ferassperded) Sand ANG Silt. < clals os-aield s ajeiecrersis aes 2 6 12 6
MBN cLy aM ve Got eh sh Gre se eit (a tcflclinytie ene Siors us apes els Siareisele 2 (a) 14 6
Wiavey silt £0. DOOM Of ATLOYOas « s+ 5.16 <o.2je'as 5 4 oO 18 6
Section 7
In south bank of north channel, Ioo yards upstream from section 6.
Thickness Depth
Feet Inches Feet Inches
(CLER] Non or oo GHEE SE ESSENCES SE Soe Sn prec ctcne 3 fo) 3 fe)
SATIN rant soe etc tote ace or Sha ctatey sl siecaciohete los ssioranatocs 3 0 6 oO
Clay, with open hearth at base. Hearth is 8 ft.
long and burnt (red) ground is nearly 1 ft.
thick wath much charcoal above. ..% ess1> «> z O 8 o
PA MAL ATAU MEULE © § clots wis sich sis, sta Sieigs Sep eSyesio cian ec cts stal< I 2 9 2
RMD cig hice cid cratatal a: sneha c/s iflava to al ute cva les teesees I 5 10 7
Sand! and!silt tovbotton Of arroyo’). ol. sii.) «1-1-1211 3 ) 14 2

56 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

Section 8

South bank of main arroyo 2 miles east of Wejegi and outside the
area of our map.

Thickness Depth
a
Feet Inches Feet Inches
Darleclaies aise srapstes aiellersia ey stotoae ue phate teemacrst ea 2 ) 2 fe)

Laminated and crossbedded sand with a few

lenses of gravel. At depth of to ft. 4 in. there

is a streak of charcoal resting on burnt (red)

earth; at 13 ft. 2 in., a second streak of char-

coal, directly below the first but no red earth;

scattered charcoal occurs to depth of 21 ft..... 25 4 27 4
Covered below.

Section 9

In arroyo of Rincon del Camino, about 300 feet south of the road
and in the fan of the rincon.

Thickness Depth
Oo FF
Feet Inches Feet Inches

Compact rusty sandy clay with fragments of

stone and charcoal; potsherds at 6 ft. 3 in..... 6 3 6 3
Crossbedded sand with impressions of plant stems

and scattered charcoal to depth of 14 ft. 8 in.

At depth of 12 ft. 8 in. there is a hearth cres-

centic in section and consisting of baked and

reddened floor 2 ft. I in. across and rising 3 in.

at the ends with layer of charcoal from 4 to

14 in. thick. South 50 ft. from this section,
scattered charcoal occurs to a depth of 16 ft.

Sil. froma top .Or bam ese cise. we elon Be ke} 0) 19 3

Section not numbered

North bank of main arroyo west of 1924 dump. [South of the
Wetherill corrals and 100 yards, more or less, west of the expedition
camp.—N. M. J.]

Thickness Depth
7) =<
Feet Inches Feet Inches

Apparently uniform indurated sand mixed with

adobe and ta: little silt .c..ct.cvteenmnetes bicas e melee 6 ) 6 te)
Laminated clay interbedded with slightly in-

durated sand. Fine interbeddings well defined

and nearly parallel. Lower 0.7 ft. has broader

Deprratitnve chet eevsrevescts, eros ancieheve ec tateteteretercteme oriole cre ete 2 7 8
Indurated sand laminated without adobe........ I 6 10 I
Fine white sand slightly indurated and finely

Devtrirmrtedo.: rsievs seratyrenaveleveretecscakeavargienarel stetouelle teranads I 0) II I

NO. 7 GEOLOGY OF CHACO CANYON—BRYAN 57

Thickness Depth
Feet Inches Feet Inches

Nonlaminated clayey sand, indurated........... (a) 5 II 6
White sand slightly and inconspicuously lami-

RUE PPP ey of chet yay eh facie Ca. aioi ote) leaistails loyaileiavatele rs ans Bialats 0 9 12 3
Well-indurated inconspicuously laminated sand

with large content of fine silt or clay......... 2 8 14 II
Gray sand slightly laminated and indurated..... 2 8 17 7
Clay containing sand and showing tendency to

granulate and fissure on shrinkage (adobe).... 2 8 20 3
Fine nonlaminated sand slightly indurated....... I 2 21 5
Clay containing some sand and showing tendency

to granulate and fissure on shrinkage......... 2 7 24 (a)
Pulverant loamy sand containing minute black

EEA tMeMtS: |(IOISE) crests ahs aleyarela de chet intel cies 3 0 27 0

Section 15

In middle of buried channel, on south bank of main arroyo and
200 yards upstream from mouth of Rincon del Camino.

Thickness Depth
Feet Inches Feet Inches

Alternate bands of sandy clay with internal

ChaCksr anlGucOmMmactySafld enisictes cis sieiele iokee oie rs 4 0 4 (0)
MOIRA ACL SAM tists au lrecchs Wtcinlalslersitvetareis' «tise ho 6 I 7 5 7
Sandy clay with internal cracks and streaks of

COMIPACE SAN ys Croversyesa eis wlel wisi elel ce ws sro leratata tatoo ove I 2 6 9
Minutely crossbedded compact sand with a few

lenses of clay 3 to 4 ft. long near base........ 3 4 10) I
Crossbedded compact sand with lense of gravel

1 ft. thick and 4 ft. long at base, containing

burnt sandstone blocks up to 6 in. across, clay

pellets, blocks of clay, and potsherds of late

type, base of buried channell.................-. 3 (0) 13 I
Crossbedded sand, crossbedding on larger scale,

streaks of pieces of black shale and coal...... 2 3 15 4
Compactycrossbeddedisandeas- a..c cscs co essa ae 2 7 17 II
GOVeredmr ies yam rlcie testes ere See ee ree ole Altay 5 O 22 II

Section 17

At the mouth of Rincon del Camino a narrow point projects into
the main arroyo from the north. On the east face of this point a well-
defined buried channel is exposed. Section 17 was measured in the
middle of this exposure.

Thickness Depth
=a SSS
Feet Inches Feet Inches

Yellow pulverant sand derived from Rincon del
ROAMING Sch avseyeee as BO OO RE CECB eee (a) 6 (0) 6

58 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

Thickness Depth
Feet Inches Feet Inches

Sandy. clay -with internal) cracks.:'....664. .cstsis ee I (o) I 6
Yellow pulverant sand derived from Rincon del

CAMAGeh Ss oes cece eas adenine Sates ee eR I 4 2 10
Sandy clay with internal cracks contains sand

lense ‘to South Of SECtiONs <i. e005 ns ede bees s 2 8 5 6
Gray, minutely laminated sand; laminae slope to

south, are interbedded there with clay layers.. 4 10 10 4
Yellow crossbedded compact sand with gravel

lenses. To the south a clay lense. Upper gravel

at 13 ft. 7 in. lower at base, contains clay

balls, pellets, and potsherds of late type in both

DENISES). heen esata cates eel ha nes daar ore TONG stake 5 8 16 Ce)
Sandy clay, upper part has internal cracks; base

oi bitcied: channels }cjaisacs/che anes Sesion, -e1s ie 2 4 18 4
Compact laminated sand with lenses of clay

having internal cracks to bed of arroyo....... 6 2 24 6

SECTIONS IN THE BURIED, OR POST-BONITO, CHANNEL

In 1922 Mr. Judd had dug a number of pits in the vicinity of Pueblo
Bonito in order to obtain soil samples for analysis. None of these pits
exceeded 10 feet in depth because that appeared to be well below the
level of fields once cultivated by the villagers and subsequently over-
lain by post-Bonito alluvial deposits. Pit number 3, located in the
plain between the expedition camp and the ruin, had been fenced and
left open for possible further tests. When this pit was deepened at
my suggestion in 1925 we were all surprised to find late Bonito pot-
sherds at depths of 10 to 18 feet. This meant we were right in the
middle of the buried channel. Some fragments showed the influence
of contact with the Mesa Verde culture, thus further identifying the
channel fill as contemporaneous with the last years of Pueblo Bonito
or even later and, of course, much later in age than the main valley fill.

Log of test pit number 3

Between Pueblo Bonito and expedition camp

Materials found Thickness Depth Found at
Feet Inches Feet Inches Potsherds Ta. eee
Dark sandy clay’ (adobe)...)..~.)...\. 2 oO 2 )
Laminated fine sand and silt........ 2 2 4 2
Dark sandy clay (adobe)’.:........ () 9 4 II
Sand with thin layers of laminated
SIP aniGlaW iis pion west. eu rnalesieri I II 6 10

Laminated layers of silt and dark

CLAY se ttape siete Goniene eS onausrers ale ws teas 4 I 10 II Io
Dark sticky ‘clay os. oo. sens eGaa wales 10 II II 10 A few to
Laminated fine Said ss). <> snrcleiers o's:0 2 5 14 3

NO. 7 GEOLOGY OF CHACO CANYON—BRYAN 59

Materials found Thickness Depth Found at
Feet Inches Feet Inches Potsherds Feet Inches
PEATE CLAY sco ci die's bivie ave resiesse te) / 14 10

Laminated and minutely crossbedded
coarse sand. This bed grades to

south into gravelly clay that ex- 16 I
tends to depth of 16 ft. Io in....... I 5 16 3 Many to
DSS (OER GAG Ma aa eC Peed Se O 5 16 8 16 8

PEIN aia aha'v ails wis, od ath linia erate co) 2 16 10)
Gravelly clay containing large and
small stones, grades into sand to 17 3
SUEUR etree a fovn miele aie eva eisacer sists I 5 18 3 Numerous to
EGUITEMS TACOS hice oye cooks a ete cterac clevettceteree fo) 3 18 6 18 3

Log of test pit number 4

Sixty feet east from arroyo bank at section 4

Materials found Thickness Depth Found at
Feet Inches Feet Inches Potsherds Feet Inches
Gray-brown sandy clay (adobe).... 1 10 I 10
Laminated sandy adobe, a lense.... 0 z 2 )
Hard brown sand with pieces of char-
EOE Se cura Sis SRN rien oes a cane 2 2 4 2

Brown sand in part finely laminated
with lenticular streaks of dark
laminated clay 4 to 1 in. thick and

spaced 2 to Io in. apart........... 3 3 z 5 \“Oneonly “7 )
“c “cc 7 2
Hard brown sand, containing at
north end of pit an irregular lense
of sandy and gravelly adobe con-
taining charcoal and bones....... 2 5 9 10 Many 9 6
Dark sandy clay (adobe) with small
: 10 6
HGMSEU OL STAVE! se ole lisld Sad.cuw e's SOCET | 7 14 5 Many in to
gravel
10
II 6
Many to
II 8
Sand (a lense) that extends to 15 ft.
Ae northend OF Pit. \ysisc seis ae eee (a) 3 14 8
mdobe to bottom: Of pit. dele. 6s (a) 9 15 5 Many 15 (a)
SUMMARY

Chaco River, a river only during occasional floods, entrenched itself
at some past time, doubtless Pleistocene, in a broad plain that then
existed in the San Juan Basin of northwestern New Mexico. In the
nearly horizontal sandstones and shales that underlie San Juan Basin,

60 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

the Chaco River flows, when it flows at all, alternately in broad valleys
and narrow canyons. To one of these latter the name Chaco Canyon
is applied almost exclusively, and here, in a stretch of about 12 miles,
there are numerous ruins of prehistoric villages, the largest of which
is Pueblo Bonito.

Chaco Canyon, after its excavation, was partly refilled with sand
and silt during a period of alluviation common to most streams of the
southwestern United States. On the flat floor of the canyon, result-
ing from this alluviation, the prehistoric peoples lived and left evidence
of their long-time occupation in hearths, scattered charcoal, potsherds,
and other relics. These remains extend to a depth of 21 feet below
the present surface of the alluvium. An ancient type of dwelling
known as a pit house has been found at a depth of 6 feet below the
surface, but the typical Pueblo III type of construction has not been
surely identified below 4 feet.

Alluviation in Chaco Canyon and generally throughout the South-
west has more recently been interrupted by the formation of an arroyo
or steep-sided gully in which the floods of the stream are now wholly
confined. The Chaco Canyon arroyo is presently 20 to 30 feet deep
and from 150 to 400 feet wide, yet a military expedition of 1849 did
not mention the gully, if it then existed. In 1877 an arroyo 16 feet
deep and 60 to 100 feet wide was reported. Available evidence indi-
cates that the arroyos of other streams were mostly formed in the
decade 1880 to 1890 and that the process is still going on. The begin-
ning of the Chaco arroyo appears to have been somewhat earlier and
the date may, with some assurance, be placed in the decade 1860 to
1870.

A study of the deposits that make up the valley fill indicates that
Chaco River never had a permanent low-water flow. No signs of irri-
gation ditches or other diversions of flowing water have been found
in the alluvial deposits. It seems probable, therefore, that the prehis-
toric inhabitants of the canyon practiced floodwater farming, a form
of agriculture still in use in the region. For this type of farming
wide-spreading floods are a prerequisite, and the beginning of erosion,
with formation of an arroyo that confines the floods and lowers the
water table, puts an end to agriculture of this type.

The main body of the valley fill is of unknown depth. Only the
upper 30 feet is exposed and of this the uppermost 21 feet contains
relics of man. Pottery made by the ancient people varies in texture
and design according to locality and age. Differences between the
kinds of pottery typical of different stages in human culture are not
wholly known, nor has a definite chronology of the stages been deter-

NOI GEOLOGY OF CHACO CANYON—BRYAN 61

mined, but broad distinctions can be made between the older and
younger civilizations.

Collections of potsherds can therefore be used as fossils in studying
the stratigraphy of the valley alluvium. Generally, only a few pot-
sherds are found at any one place and many of these are indeterminate,
hence of no diagnostic value. Somewhat meager collections of sherds
from depths of 6 to 21 feet have been examined by the expedition’s
archeologists who identify them as mostly a coarse ware characteris-
tic of the Pit House culture. On the basis of these fragments, there-
fore, we may draw the inference that people of the Pit House period
were the principal inhabitants of Chaco Canyon during the time
required for deposition of those 15 feet of alluvium.

Potsherds collected from the zone of valley fill less than 6 feet below
the surface are generally of Pueblo III type. This fact, together with
ruins whose foundations are partly buried in alluvium, indicate that
Pueblo III people occupied the valley during the period represented
by the last 6 feet of alluviation.

In the bank of the arroyo near Pueblo del Arroyo there is exposed
a buried channel which extends to a depth of 15 feet below the pres-
ent surface. At this point the channel is a well-defined ancient arroyo
that had been refilled and then buried under an additional 2 feet of
sediment in the interval between abandonment of Pueblo Bonito and
American Army penetration of Chaco Canyon in 1849. Potsherds
removed from the gravel lenses of that buried channel included frag-
ments of the latest Pueblo Bonito types. The channel, therefore, must
have been refilled late in the occupancy of Pueblo Bonito or after its
abandonment.

By means of test pits in which similar pottery was found, we traced
this buried channel for about 1,000 feet across the plain fronting
Pueblo Bonito and later discovered remnants of it both up and down
the canyon. This buried channel clearly represents a period of dissec-
tion and arroyo formation for the full length of the valley and, assum-
ing that the dissection occurred late in the occupancy of Pueblo Bonito,
an adequate cause exists for abandonment of the canyon by aboriginal
farmers whose floodwater fields were destroyed by confinement of the
floods within this channel, and by concurrent events.

Our examination of the main valley fill suggests alternate dissection
and alluviation of Chaco Canyon: three periods of dissection and two
of alluviation. If this alternation represents a true cycle, we may
expect the present arroyo to run its course and then be refilled and
perhaps covered over. However plausible it may be to attribute forma-
tion of the present arroyo to destruction of the vegetative cover by

62 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

overgrazing, the previous dissection and subsequent alluviation were in
no way affected by domestic animals. It seems probable, therefore, that
the ultimate cause of this periodic change in the regime of streams is
climatic. A slightly increased rainfall would increase the vegetative
cover and thereby both reduce the violence of floods and protect the
soil from erosion. Any decrease in rainfall would produce a reversed
effect. Although the deposits of Chaco Canyon contain no definite
evidence of a more humid climate during the two periods of their
deposition, it seems likely that an increased humidity did exist and
was a factor in development of the distinctive Chaco culture. The sub-
sequent change to more arid conditions was doubtless of less effect
until it culminated in formation of the twelfth-century arroyo that
unexpectedly became a dominant feature of this study.

BIBLIOGRAPHY

BAvER, CLYDE Max.
1917. Stratigraphy of a part of the Chaco River Valley. U. S. Geol. Surv.
Prof, Pap. 98, pp. 271-278.
BAvER, CLlypDE MAx, and REEsSIDE, JOHN B., Jr.
1921. Coal in the middle and eastern parts of San Juan County, New
Mexico. U. S. Geol. Surv. Bull. 716, pp. 155-237.
Brown, BARNUM.
1910. The Cretaceous Ojo Alamo beds of New Mexico, with description of
the new dinosaur genus Kritosaurus. Bull. Amer. Mus. Nat. Hist.,
vol. 28, pp. 267-274.
Bryan, Kirk.
1920. Origin of rock tanks and charcos. Amer. Journ. Sci., ser. 4, vol. 50,
pp. 188-206.
1923a. Erosion and sedimentation in the Papago country, Arizona, with a
sketch of the geology. U. S. Geol. Surv. Bull. 730, pp. 19-90.
1923b. Ground water in Quinn River and Paradise Valleys, Nevada. U. S.
Geol. Surv. press notice, Jan. 29, 1923.
1925a. Date of channel trenching (arroyo cutting) in the arid Southwest.
Science, n.s., vol. 62, No. 1607, pp. 338-344.
1925b. The Papago country, Arizona—a geographic, geologic, and hydrologic
reconnaissance, with a guide to desert watering places. U. S. Geol.
Surv. Water-Supply Pap. 499.
1926a. Recent deposits of Chaco Canyon, New Mexico, in relation to the life
of the prehistoric peoples of Pueblo Bonito. (Abstr.) Journ. Wash-
ington Acad. Sci., vol. 16, No. 3, pp. 75-76.
1926b. Channel erosion of the Rio Salado, Socorro County, New Mexico.
U. S. Geol. Surv. Bull. 790A, pp. 17-19.
1928a. Historic evidence on changes in the channel of Rio Puerco, a tribu-
tary of the Rio Grande in New Mexico. Journ. Geol., vol. 36, No. 3,
pp. 265-282.
1928b. Change in plant associations by change in ground water level.
Ecology, vol. 9, No. 4, pp. 474-478.

NO. 7 GEOLOGY OF CHACO CANYON—BRYAN 63

1928c. Niches and other cavities in sandstone at Chaco Canyon, N. M.
Zeitschr. Geomorph., vol. 3, pp. 125-140. Leipzig.

1929. Flood-water farming. Geogr. Rev., vol. 19, pp. 444-456.

1941. Pre-Columbian agriculture in the Southwest as conditioned by periods
of alluviation. Ann. Assoc. Amer. Geogr., vol. 31, No. 4, pp. 219-
242.

Bryan, Kirk, and McCann, FRANKLIN T.

1936. Successive pediments and terraces of the Upper Rio Puerco in New

Mexico. Journ. Geol., vol. 44, No. 2, pt. 1, pp. 145-172.
Davis, W. M.

1903. An excursion to the Plateau province of Utah and Arizona. Bull.

Mus. Comp. Zool., Harvard Coll., vol. 42, pp. 1-50.
Donce, RicHArp E.

1902a. An interesting landslide in the Chaco Canyon, New Nexico. Science,
n.s., vol. 15, p. 746.

1902b. Arroyo formation. Science, n.s., vol. 15, p. 746.

1910. The formation of arroyos in adobe-filled valleys in the southwestern
United States. (Abstr.) Rep. British Assoc. Adv. Sci., vol. 79,
PP. 531-532.

1920. See Pepper, George H.

Dovetass, A. E.

1924. Some aspects of the use of the annual rings of trees in climatic study.
Ann. Rep. Smithsonian Inst. for 1922, pp. 223-239.

1935. Dating Pueblo Bonito and other ruins of the Southwest. Nat. Geogr.
Soc., Contr. Techn. Pap., Pueblo Bonito Ser., No. 1.

Duce, JAMES TERRY.

1918. The effect of cattle on the erosion of canyon bottoms. Science, n.s.,

vol. 47, pp. 450-452.
Dutton, C. E.

1882. Tertiary history of the Grand Canyon district. U. S. Geol. Surv.

Monogr., No. 2.
Enpiicu, F. M.

1877. Geological report on the Southeastern district. 9th Ann. Rep. U. S.

Geol. and Geogr. Surv. Terr. for the year 1875, pp. 103-235.
GARDNER, J. H.

1909. The coal field between Gallina and Raton Spring, New Mexico, in

the San Juan coal region. U. S. Geol. Surv. Bull. 341, pp. 335-351.
Grecory, H. E.

1915. The oasis of Tuba, Arizona. Ann. Assoc. Amer. Geogr., vol. 5,
Pp. 107-119.

1916. The Navajo country: A geographic and hydrographic reconnaissance
of parts of Arizona, New Mexico, and Utah. U. S. Geol. Surv.
Water-Supply Pap. 380.

1917. Geology of the Navajo country: A reconnaissance of parts of Ari-
zona, New Mexico, and Utah. U. S. Geol. Surv. Prof. Pap. 93.

Hack, Joun T.

1942. The changing physical environment of the Hopi Indians of Arizona.
Pap. Peabody Mus. Amer. Archaeol. and Ethnol., Harvard Univ.,
vol. 35, No. 1.

64 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

Hoimes, W. H.

1877. Geological report on the San Juan district. 9th Ann. Rep. U. S. Geol.
and Geogr. Surv. Terr. for the year 1875, pp. 237-276.
Hoover, J. W.
1929. The Indian country of southern Arizona. Geogr. Rev., vol. 19, No. 1,
pp. 38-60.
1930. Tusayan: The Hopi country of Arizona. Geogr. Rev., vol. 20, pp.
425-444.
HuNTINGTON, ELLSworrtu.
1914. The climatic factor as illustrated in arid America. Carnegie Inst.
Washington Publ. 192.
Jackson, W. H.
1878. Report on the ancient ruins examined in 1877. 1toth Ann. Rep. U. S.
Geol. and Geogr. Surv. Terr. for the year 1876, Chaco Canyon,
PP. 431-450.
Jupp, Nei M.
1924. Two Chaco Canyon pit houses. Ann. Rep. Smithsonian Inst. for 1922,
PP. 399-413.
Loew, Oscar.
1879. Report on the ruins in New Mexico. Report upon U. S. geographical
surveys west of the 1ooth meridian (Wheeler Surveys), vol. 7,
Pp. 337-345. (Abstr. from Ann. Rep. Chief of Engineers for 1875,
App. LL, Sect. J-2, pp. 174-178.)
LuMHOLTZ, Kart SoFus.
1912. New trails in Mexico. New York.
MaAtTTHEw, W. D.
1897. A revision of the Puerco fauna.
PP. 259-323.
Morrison, C. C.

Amer. Mus. Nat. Hist. Bull. 9,

1879. Notice of the Pueblo Pintado and of other ruins in the Chaco Canyon.
Report upon U. S. geographical surveys west of the rooth meridian
(Wheeler Surveys), vol. 7, pp. 366-369. (Abstr. from Ann. Rep.
Chief of Engineers for 1876, App. JJ, Sect. E, pp. 359-360.)

Netson, N. C.

1914. Pueblo ruins of the Galisteo Basin, New Mexico. Anthrop. Pap.
Amer. Mus. Nat. Hist., vol. 15, pt. I, pp. 1-124.

OLmstTED, FRANK HENRY.

1919. Gila River flood control: Report on flood control of Gila River in

Graham County, Arizona. Dep. Interior, 65th Cong., 3d Sess.,
Senate Doc. 436.

Pepper, Grorce H.

1920. Pueblo Bonito. Anthrop. Pap. Amer. Mus. Nat. Hist., vol. 27.
REAGAN, ALBERT B.

1922. Archaeological notes on Pine River valley, Colorado, and the Kayenta-
Tuba region, Arizona. Trans. Kansas Acad. Sci., vol. 30, pp. 244-

331, 394-429.
1924a. Recent changes in the Plateau region. Science, n.s., vol. 60, pp. 283-
285.

1924b. Stream aggradation through irrigation.

Pan-Amer. Geol., vol. 42,
PP. 335-344.

NO. 7 GEOLOGY OF CHACO CANYON—BRYAN 65

REEsIDE, J. B., Jr.

1924. Upper Cretaceous and Tertiary formations of the western part of the
San Juan Basin, Colorado and New Mexico. U. S. Geol. Surv.
Protpeapwisa:

Ricu, Joun Lyon.

ro1r. Recent stream trenching in the semi-arid portion of southwestern New
Mexico, a result of removal of vegetative cover. Amer. Journ. Sci.,
ser. 4, vol. 32, pp. 237-245.

Rosinson, H. H.

1907. The Tertiary peneplain of the Plateau district and adjacent country
in Arizona and New Mexico. Amer. Journ. Sci., ser. 4, vol. 24,
pp. 109-120.

Scurapver, F. C.

1906. The Durango-Gallup coal field of Colorado and New Mexico. U. S.

Geol. Surv. Bull. 285, pp. 241-258.
SENTER, DoNovVAN.

1937. Floor deposition and erosion in Chaco Canyon. Univ. New Mexico

Bull. No. 308, App. I, pp. 134-139.
SHALER, MILvarp K.

1907. A reconnaissance of the western part of the Durango-Gallup coal
field of Colorado and New Mexico. U. S. Geol. Surv. Bull. 316,
Pp. 376-426.

Stmpson, JAMES H.

1850. Journal of a military reconnaissance from Santa Fe, New Mexico,
to the Navajo country ... in 1849. Rep. Secretary of War, 31st
Cong., ist Sess., Senate Ex. Doc. No. 64, pp. 56-168.

Srnciair, W. J., and GRANGER, WALTER.

1914. Paleocene deposits of the San Juan Basin, New Mexico. Bull. Amer.

Mus. Nat. Hist., vol. 33, pp. 207-316.
Situ, G. E. P.

1910. Ground water supply and irrigation in the Rillito Valley. Univ.

Arizona Agr. Exp. Stat. Bull. No. 64, pp. 98-99.
THORNBER, J. J.

1910. The grazing ranges of Arizona. Univ. Arizona Agr. Exp. Stat.

Bull. No. 65, pp. 335-338.
WILLIAMS, GEORGE O.

1925. Radium-bearing silts of southeastern Utah. Eng. and Min. Journ.

Press, vol. 119, pp. 201-202.

sama a 1a" Fee
; Thai ve i all ae
A ry ie ai tet)
Oye NES eat
ean th .
: ee ea nie 4 bee if i ne
‘ ; ae
; :
rie SMe | Hytoet + i 4
RE ale ay cata dae
ea - Pike ” ;
a) , ih A Wee Carle ath

WY Fidos oe
wer j sages /
\ re. a

ig
Fl Ange

NY
rr

SMITHSONIAN MISCELLANEOUS COLLECTIONS
VOLUME 122, NUMBER 8

Ser tieR ADDITIONS TO THE BIRDS
OF PANAMA AND COLOMBIA

BY,
ALEXANDER WETMORE

Research Associate, Smithsonian Institution

eS FEMA Regs
\e) Ol ie) sen
< ee ANor 0,

(PusicaTion 4142)

CITY OF WASHINGTON
PUBLISHED BY THE SMITHSONIAN INSTITUTION
DECEMBER 17, 1953

The Lord Baltimore Press

BALTIMORE, MD., U. & J.

FURTHER ADDITIONS TO THE BIRDS OF
PANAMA AND COLOMBIA
By ALEXANDER WETMORE

Research Associate, Smithsonian Institution

During work of recent months on our extensive collections of birds
from Panama and northern Colombia, several hitherto unrecognized
forms have been found that merit description to give better under-
standing of the geographic variation in the species concerned. With
these I have included records of three others that have not been re-
ported previously from Colombia.

Family CRACIDAE
ORTALIS RUFICRISSA LAMPROPHONIA, new subspecies

Characters—Similar to Ortalis ruficrissa ruficrissa (Sclater and
Salvin)? but smaller; feet smaller; tail shorter; lower breast and
abdomen whiter ; back and wings more grayish brown.

Description—Male adult, U.S.N.M. No. 368535, from the Serrania
de Macuire, above Nazaret, Guajira, Colombia, collected May 5, 1941,
by A. Wetmore and M. A. Carriker, Jr. (orig. No. 11792). Crown
chaetura drab, feathers of forehead edged with a wash of light olive-
gray; ear coverts drab; remainder of side of head mouse gray, the
feathers with fuscous shafts ; hindneck deep mouse gray, shading into
the color of the upper back ; upper back and wings between olive-brown
and deep olive ; rump and upper tail coverts slightly darker than buffy
brown; primaries light olive-brown, with a slight grayish wash on
outer webs; central rectrices deep olive, tipped indefinitely with buffy
brown; outer rectrices dark greenish olive (slightly iridescent) , tipped
widely with white ; foreneck and sides of neck light grayish olive with
a slight brownish wash; multiple line of feather shafts, extending
longitudinally down center of bare throat, black, becoming fuscous
where they merge with upper portion of feathered foreneck ; webs of

1 Ortalida ruficrissa Sclater and Salvin, Proc. Zool. Soc. London, November
1870, p. 538 (Valledupar, Magdalena, Colombia).

SMITHSONIAN MISCELLANEOUS COLLECTIONS, VOL. 122, NO. 8

2 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

feathers bordering bare throat pale smoke gray; lower breast and
upper abdomen white, washed with pinkish buff on upper portion of
breast, this buffy wash extending down onto the sides; flanks and
under tail coverts tawny ; tibiae pinkish buff to cinnamon-buff ; under
wing coverts with outer half light grayish olive, inner half clay color.
Tip of bill horn color; rest of distal half castor gray; basal half
neutral gray ; tarsus and toes fuscous; claws drab (from dried skin).

Range——The Serrania de Macuire, at the eastern end of the
Guajira Peninsula, Colombia.

Measurements—Males (2 specimens): Wing 195-202 (198), tail
227-235 (230), culmen from base 25.7-27.2 (26.4), tarsus 64.0-65.4
(64.7) mm.

Females (2 specimens): Wing 195-200 (197), tail 217 (in both),
culmen from base 25.8-26.0 (25.9), tarsus 59.5-62.3 (60.9) mm.

Type, male: Wing 202, tail 235, culmen from base 27.2, tarsus
65.4 mm.

Remarks.—Ortalis ruficrissa ruficrissa, named many years ago by
Sclater and Salvin from Valledupar, was known prior to our work in
northeastern Colombia from two specimens, the type in the British
Museum and one from Dibulla on the north coast, about 35 miles west
of Riohacha, in the Carnegie Museum. We found it first at Maicao,
Guajira, and later Carriker secured a series that extend the range of
the typical race from the western Guajira at Maicao, into northeastern
Magdalena, from La Cueva in the eastern foothills of the Sierra
Nevada de Santa Marta southward, along the western base of the
Sierra de Perija, to Casacara. This excellent series is sufficient to
demonstrate the distinctness of the isolated colony on the Serrania de
Macuire, which is cut off by many miles of barren desert, where there
is no suitable habitat for these birds, from the more-forested section
of the western Guajira inhabited by Ortalis ruficrissa ruficrissa.
Measurements of the typical race are as follows:

Males (7 specimens): Wing 206-238 (217), tail 240-272 (253),
culmen from base 26.0-29.8 (27.8, average of 6), tarsus 64.0-73.1
(69.2) mm.

Females (5 specimens) : Wing 196-217 (205), tail 230-253 (239),
culmen from base 26.0-28.2 (26.7), tarsus 62.0-67.0 (64.1, average
of 4) mm.

In the series of ruficrissa there is one bird in somewhat worn
plumage, from Camperucho, Magdalena, that is as white on the
breast as the four lamprophonia, but it has the upper breast and fore-
neck and the dorsal surface darker, the feet larger, and the tail defi-
nitely longer.

no. 8 BIRDS OF PANAMA AND COLOMBIA—WETMORE 3

The name for the new race is given because of the raucous voice
that carries for long distances.

Family BUCCONIDAE
NONNULA FRONTALIS STULTA, new subspecies

Characters.—Similar to Nonnula frontalis frontalis (Sclater),? but
somewhat grayer, less rufescent above; crown duller brown; averag-
ing very slightly duller brown on breast and foreneck.

Description—Type, U.S.N.M. No. 445077, male, El Uracillo,
Province of Coclé, Panama, February 23, 1952, A. Wetmore and
W. M. Perrygo (orig. No. 16946). Forehead, sides of head extend-
ing to area above eye, and including the lores and erectile feathers
above the anterior end of the eye, gray (dark gull gray) ; anterior
portion of crown between verona brown and warm sepia, shading to
bister on nape; rest of upper surface, including wings and tail, sepia,
with wing coverts, primaries, and secondaries edged lightly with
Saccardo’s umber, and the ends of the rectrices shading to clove
brown ; outermost rectrix drab, the second pair edged externally and
tipped rather widely with drab, the others with the drab less exten-
sive; extreme base of the feathers on chin at the base of the bill
white; throat, foreneck, and breast between cinnamon and _ sayal
brown; flanks clay color ; abdomen whitish ; under tail coverts white ;
edge of wing cinnamon; under wing coverts tawny-olive ; inner webs
of secondaries and inner primaries cinnamon-buff. Bill, tarsus, and
toes blackish slate (from dried skin).

Measurements.—Males (12 specimens): Wing 55.2-58.8 (57.1),
tail 52.3-58.7 (55.1), culmen from base 22.3-24.8 (23.5), tarsus
13.0-14.7 (13.6) mm.

Females (17 specimens): Wing 55.1-62.0 (58.1), tail 53.6-59.6
(57.1), culmen from base 22.4-25.6 (24.0), tarsus 12.2-14.2 (12.6)
mm.

Type, male: Wing 55.7, tail 54.8, culmen from base 23.7, tarsus
13.5 mm.

Range.—Panama from northeastern Coclé (El Uracillo) and the
Canal Zone (Lion Hill) through eastern Province of Panama (Tocu-
men, Pacora, Chepo), through Darién (Jesucito, Rio Esnape, El Real,
El Tigre, Boca de Cupe, Capeti, Cana) to extreme northern Choco
(Acandi), Colombia. Found mainly on the Pacific slope.

2 Malacoptila frontalis P. L. Sclater, Ann. Mag. Nat. Hist., ser. 2, vol. 13, 1854,
Pp. 479 (interior of Colombia).

4 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. I22

Remarks.—This attractive little bird, while somewhat more active
than the larger members of its family, shares with them the habit
of resting quietly for long periods. I have found it in brushy areas
or low down in tracts of gallery forest. In the main, in Panama it
ranges on the Pacific slope, thus far having been found in the Carib-
bean drainage near the head of canoe navigation on the Rio Indio in
northern Coclé, at Lion Hill (before the Panama Canal was con-
structed), and at Acandi, Colombia, on the western shore of the Gulf
of Uraba. Specimens from Unguia, Choco, on the western side of
the lower Rio Atrato, are N. f. pallescens.

This species has two color phases, one rufescent, in which the
dorsal surface is decidedly brown and where brown extends over the
entire ventral side except the center of the abdomen and the lower tail
coverts, and the other grayish, where the lower surface especially is
paler, with the white of the abdomen more extensive.

Nonnula f. stulta differs from pallescens of extreme northern
Colombia in being darker colored and also duller brown, less rufescent,
above.

Family TROCHILIDAE
COELIGENA ORINA, new species

Characters —Similar to Coeligena bonapartei (Boissoneau)* but
crown uniform, without frontal spot ; body color uniform dark green,
without bronzy sheen on lower breast, under tail coverts and rump;
under tail coverts uniform green, without cinnamon edgings; tail
dark green without bronzy reflections ; wings dull black (not fuscous) ;
spot on foreneck decidedly brighter blue; under wing coverts darker
green ; bill more slender.

Description—Type, U.S.N.M. No. 436219, male adult, Paramo de
Frontino, at 10,500 feet, Antioquia, Colombia, August 27, 1951, col-
lected by M. A. Carriker, Jr. (orig. No. 21016). Feathers of crown,
sides of head, and hindneck iridescent elm green, margined with black,
the green being evident clearly only when viewed at an appropriate
angle; back, and lesser and middle wing coverts iridescent spinach
green ; rump and upper tail coverts strongly iridescent, varying from
lettuce green to Cosse green ; remiges and greater wing coverts aniline
black, with a faint violet-purple sheen, the inner greater coverts
edged with shining lettuce green; outer primary margined lightly on
external web with avellaneous ; rectrices iridescent yellowish oil green,
with lightly indicated edgings of dull black; chin chaetura black;

3 Ornismia bonarpartei [sic] Boissoneau, Rev. Zool., 1840, p. 6 (Bogota).

no. 8 BIRDS OF PANAMA AND COLOMBIA—-WETMORE 5

foreneck and upper breast iridescent spinach green, the feathers
margined lightly with black; a spot of glittering salvia blue on fore-
neck ; lower breast, sides, flanks, and under tail coverts shining lettuce
green; center of abdomen, in a small area, dull white; tibiae cinna-
mon-buff ; under wing coverts iridescent elm green. Bill black, toes
fuscous, claws black (from dried skin).

Measurements.—Male, type: Wing 75.2, tail 44.0, culmen from
base 33.6 mm.

Range.—Known only from the Paramo de Frontino at 10,500 feet,
above Urrao, Antioquia, Colombia.

Remarks.—The single male seen appears closer to Coeligena bona-
partei than to others of the genus, and apparently is a representative
of that group in the western Andes. It is so different, however, that
I have no doubt as to its being a distinct species. The specimen
appears fully adult, so that absence of the frontal spot may not be
ascribed to immaturity. Compared to C. bonapartei the bill, in addi-
tion to being more slender, is longer.

This is one of the handsomest of the novelties obtained during the
present ornithological exploration of Colombia. Carriker noted on the
label that the bird was taken in forest below the open paramo.

Family TYRANNIDAE
MYIARCHUS FEROX AUDENS, new subspecies

Characters.—Similar to Myiarchus ferox panamensis Lawrence *
but grayer above, with the crown more nearly uniform with the back ;
slightly paler yellow below.

Description—tType, U.S.N.M. No. 443502, female, Nuqui, Choco,
Colombia, collected on March 5, 1951, by M. A. Carriker, Jr. (orig.
No. 19780). Crown and auricular area between hair brown and deep
grayish olive, with an indefinite wash of chaetura drab on central
portion along shafts; neck, back, lesser wing coverts, and upper tail
coverts slightly darker than deep grayish olive; rump grayish olive ;
middle and greater wing coverts chaetura drab, tipped rather widely
with grayish olive to produce two indistinct wing bars; primary
coverts, primaries, secondaries, and rectrices chaetura drab; sec-
ondaries margined prominently and inner primaries lightly with dull
white; outer rectrix with outer web dull buffy brown, the others
edged with grayish olive, more prominently at the base, all tipped
lightly with a wash of olive-buff; lores, and an indistinct line above

4 Myiarchus panamensis Lawrence, Ann. Lyc. Nat. Hist. New York, vol. 7,
May 1860, p. 284 (Atlantic slope of Canal Zone on the Panama Railroad).

6 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

eye pale olive-gray ; a mixture of dull white immediately in front of
eye, with whitish feathers extending back over the lower eyelid ; chin
indistinctly whitish; throat, foreneck, and upper breast light olive-
gray, lined indistinctly with dull white on throat and upper foreneck ;
lower breast and abdomen chartreuse yellow; sides and under tail
coverts sea-foam yellow; edge of wing and under wing coverts be-
tween Marguerite yellow and primrose yellow; inner webs of pri-
maries and secondaries pale olive-buff toward base. Bill, tarsus, and
feet dull black (from dried skin).

Measurements——Female (3 specimens): Wing 86.0-91.1 (89.4),
tail 78.2-86.4 (83.3), culmen from base 20.2-21.6 (20.8), tarsus 21.8-
22.7 \(22.2 jem.

Type, female: Wing 91.1, tail 86.4, culmen from base 21.6, tarsus
22.7 cin:

Range.—Known only from near Nuqui, Department of Choco,
northwestern Colombia.

Remarks.—The three skins on which this new race is based have
been compared with a large series of M. f. panamensis covering the
area from western Panama across northern Colombia. They stand
out clearly from all in the definitely gray coloration. The nearest
specimens of panamensis seen are from Jaqué, Darién, across the
border in Panama, and from Nicocli and Villa Artiaga, Antioquia,
in Colombia. It is probable that the new race ranges immediately
back of the beaches along the coast, as the species as a whole does
not penetrate into heavy forests such as are found inland in the
Choco. Its distribution, therefore, may be through a relatively narrow
belt, east and west.

It is pertinent to add here that Myiarchus ferox australis Hellmayr
is also to be included in the list of birds found in Colombia, as shown
by a male in the U. S. National Museum taken at Villavicencio, Meta,
by Hermano Nicéforo Maria in December 1939. Zimmer records
four specimens from this locality as intermediate between M. f. ferox
and australis, but nearer australis. In a later paper by Zimmer and
Phelps,® describing M. f. brunnescens, these four skins from Villa-
vicencio were, through some error in printing, included under brunnes-
cens instead of australis in their list of specimens examined. It was
this, apparently, that caused De Schauensee‘ to include Villavicencio
under the range he assigns to brunnescens, and to omit australis from
his list.

5 Amer. Mus. Nov. No. 994, June 2, 1938, pp. 12, 15.
6 Amer. Mus. Nov. No. 1312, March 12, 1946, p. II.
7 Caldasia, vol. 5, No. 24, July 10, 1950, p. 826.

no. 8 BIRDS OF PANAMA AND COLOMBIA—WETMORE 78

PHAEOMYIAS MURINA EREMONOMA, new subspecies

Characters —Similar to Phaeomyias murina incomta (Cabanis and
Heine)* but dorsal surface lighter, grayer ; slightly smaller in size.

Description—Type, U.S.N.M. No. 400534, male, taken on the
Rio Santa Maria, 4 miles north of Paris, Herrera, February 24, 1948,
by A. Wetmore and W. M. Perrygo (orig. No. 13500). Crown, sides
of head, hindneck, back, and lesser and middle wing coverts between
drab and grayish olive; the feathers of the back becoming drab at the
tips ; rump and upper tail coverts drab; superciliary stripe, extending
from the front of the eye back along the sides of the crown, and the
feathers encircling the edge of the eyelids, dull white; lores light
grayish olive; primaries, secondaries, and greater coverts dull hair
brown; lesser wing coverts edged indefinitely with dull tilleul buff,
forming an indistinct wing bar; middle and greater coverts tipped
widely with somewhat dull pale pinkish buff, forming two prominent
wing bars, in addition to the third indistinct one on the lesser coverts ;
inner secondaries margined and tipped broadly with dull white ; outer
webs of outer secondaries and primaries edged very narrowly with
pale olive-buff; rectrices dull hair brown, tipped and margined nar-
rowly on the external webs with pale olive-buff; throat and fore-
neck dull white ; upper breast and sides washed with pale smoke gray ;
lower breast and abdomen light primrose yellow; under tail coverts
Marguerite yellow; under wing coverts light primrose yellow, lined
with hair brown on bend of wing. Maxilla and tip of mandible
fuscous-black ; base of mandible light grayish olive ; tarsus, toes, and
claws black (from dried skin).

Measurements—Male (16 specimens): Wing 54.8-60.5 (56.8),
tail 48.7-56.2 (51.5), culmen from base 10.0-11.6 (10.8), tarsus 17.0-
fo.3 (17.7) mm,

Females (13 specimens): Wing 49.8-55.9 (52.8), tail 44.5-48.3
(46.7, average of 12), culmen from base 9.9-I1.5 (10.4, average
of 12), tarsus 15.8-18.3 (16.9) mm.

Type, male: Wing 55.8, tail 48.7, culmen from base 10.7, tarsus
17.5 mm.

Range.—Lowland areas of the Pacific slope of Panama, from the
valley of Rio San Pablo in southern Veraguas (Sona, Rio de Jests),
and the eastern side of the Azuero Peninsula (Los Santos, Parita,
Paris, Potuga, El Barrero) through Coclé (Aguadulce) to the western
section of the Province of Panama (Nueva Gorgona, La Campana).

8 Elainea incomta Cabanis and Heine, Mus. Hein., vol. 2, 1859, p. 59 (Carta-
gena, Colombia).

8 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

Remarks.—Since 1948, when I first found this small flycatcher in
the Provinces of Los Santos and Herrera on the eastern side of the
Azuero Peninsula, I have been assembling material for comparison
from elsewhere in Panama, and from Colombia, since it seemed
doubtful that the Panamanian birds, separated from the Colombian
group by the whole of Darién and the lower Atrato basin, were the
race incomta, named from Cartagena, to which they have been re-
ferred. It was found immediately that so many were in worn plum-
age, due to the intense light and lack of deep shade in their brushy
haunts, that there was considerable fading in color. Finally, enough
have been obtained to demonstrate the differences outlined above,
through comparison of birds in reasonably fresh dress. Panamanian
birds, in addition to being lighter, grayer above, average more yellow
below. The latter difference however is variable from specimen to
specimen, and is useful only in examining series, so that it is not
included in the formal diagnosis.

The size difference between the newly described race and incomta
is not great but is illustrated by examination of the following meas-
urements of the latter form, from Colombian specimens.

Males (23 specimens): Wing 60.1-64.4 (62.2), tail 51.0-58.9
(54.9), culmen from base 10.3-11.7 (10.9), tarsus 16.4-19.4 (18.4)
mm.

Females (16 specimens): Wing 55.5-60.1 (58.3), tail 46.5-53.5
(58.4), culmen from base 9.8-11.0 (10.4), tarsus 16.3-19.3 (17.1)
mm.

Zimmer ° has recorded one from Panama from El Villano, a locality
that I have not found on available maps.

In the field this species is liable to confusion with the beardless
flycatcher (Camptostoma obsoletum), often encountered in the same
localities, in spite of the larger size of Phaeomyias murina, because
of similar habits.

PHYLLOMYIAS GRISEICEPS QUANTULUS, new subspecies

Characters—Similar to Phyllomyias griseiceps cristatus Ber-
lepsch,!° but grayer, less greenish, on the back; pileum darker, more
brownish; under surface slightly paler, the yellow of breast and
abdomen being lighter, and the sides and upper breast paler, with
less olive wash.

9 Amer. Mus. Nov. No. 1109, May 15, 1941, p. 10.

10 Phyllomyias cristatus Berlepsch, Journ. Orn., vol. 32, April 1884, p. 250
(Bucaramanga, Santander, Colombia). This description was repeated on
page 300 in the succeeding issue for July-October.

no. 8 BIRDS OF PANAMA AND COLOMBIA—WETMORE 9

Description—tType, U.S.N.M. No. 420014, male adult, Cana, 1,800
feet elevation, Darién, Panama, June 1, 1912, E. A. Goldman (orig.
No. 15783). Crown and hindneck fuscous-black; a narrow super-
ciliary, extending from well behind eye to nostrils, dull white; lores
chaetura drab; hindneck, back, and scapulars deep olive; rump and
upper tail coverts citrine-drab; wing coverts hair brown, with very
slight paler borders; primaries and secondaries chaetura drab, the
innermost secondaries margined narrowly with pale olive-buff; rec-
trices hair brown, the outermost with slight tipping of pale olive-buff ;
sides of head chaetura drab, with the lower eyelid dull white, and
numerous thin lines of dull white across auricular area ; chin and throat
dull white; rest of under surface in general reed yellow, becoming
primrose yellow on the abdomen and under tail coverts ; sides of breast
lightly washed with citrine-drab ; axillars barium yellow; under wing
coverts primrose yellow; inner webs of primaries and secondaries
edged prominently with Marguerite yellow. Bill fuscous, tarsus and
toes fuscous-black (from dried skin).

Measurements.—Male, type: Wing 49.3, tail 43.1, culmen from
base 9.7, tarsus 13.9 mm.

Range.—Known only from near Cana, Darién.

Remarks.—The single specimen on which this form was based
was taken by E. A. Goldman toward the close of his work near
Cana. While Nelson identified it correctly to species, later the skin
was not entered in the museum catalog with the rest of Goldman’s
collection, coming to attention only recently in examining the rest of
this series. It seems appropriate to describe it, since I find no dupli-
cation of its characters in more than 30 skins of griseiceps examined,
including examples of the subspecies griseiceps, cristatus, caucae, and
pallidiceps. It is the only record for the species in Panama.

It is probable that a bird recorded by de Schauensee +! from the
Rio Juradé, across the border in Choco, Colombia, also belongs to
this new race.

The name quantulus, “how small,” is given because of the tiny size
of these little flycatchers.

Family FRINGILLIDAE
SICALIS LUTEOLA EISENMANNI, new subspecies

Characters—Male similar to that of Sicalis luteola chrysops Scla-
ter ** but clearer, brighter yellow on under surface; dark streaking

11 Caldasia, vol. 5, No. 24, 1950, p. 864.
12 Sycalis chrysops Sclater, Proc. Zool. Soc. London, 1861 (Feb. 1, 1862),

10 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

on back heavier; sides of head more greenish. Female like that of
S. 1. chrysops, but lighter, brighter yellow below and on rump; sides
of head more grayish brown.

Description—Type, U.S.N.M. No. 449369, male adult, taken 2
miles east of Anton, Province of Coclé, Panama, June 20, 1953, by
A. Wetmore (orig. No. 18159). Forehead wax yellow, extending
back over eye as an indefinitely delimited superciliary line, and shad-
ing posteriorly into the sulphine yellow of the crown; hindneck olive
lake with a slight grayish overwash; crown, behind the level of the
eyes, and hindneck streaked narrowly with chaetura drab; auricular
region and side of neck yellowish citrine; lores wax yellow, feathers
of the lower half and those behind rictus being white basally, form-
ing an indefinite whitish spot; feathers of upper back chaetura black,
edged with olive-yellow, producing heavy dark streaks outlined by
narrower lighter ones ; lower back yellowish citrine ; rump and upper
tail coverts slightly brighter than sulphine yellow; wing coverts in
general chaetura drab; lesser coverts edged with light yellowish olive,
which is more extensive on the inner feathers; middle coverts mar-
gined with deep olive-buff ; greater coverts edged with pale olive-buff ;
primaries and secondaries chaetura drab; central portion of outer
webs of primaries edged narrowly with olive-citrine, distal portion
and secondaries margined with dull white; rectrices chaetura drab,
edged basally with yellowish citrine; under surface lemon chrome,
deepening on sides of throat and foreneck to apricot yellow ; under
tail coverts wax yellow; sides of upper breast washed with strontian
yellow ; under wing coverts lemon yellow; inner webs of primaries
edged indistinctly with pale olive-buff. Maxilla and tip of mandible
chaetura black ; sides of mandible fuscous, base olive-buff ; tarsus and
toes clove brown (from dried skin).

Measurements.——Males (9 specimens): Wing 60.4-64.8 (62.8),
tail 38.5-45.0 (41.3), culmen from base 8.9-10.0 (9.5, average of
eight), tarsus 14.2-15.8 (15.0) mm.

Female (1 specimen) : Wing 60.1, tail 37.3, culmen from base 9.3,
tarsus 15.4 mm.

Type, male: Wing 62.3, tail 45.0, culmen from base 9.7, tarsus
14.9 mm.

Range.—The savannas of southern Coclé Province, Panama; re-
corded to date from west of Rio Hato to near Aguadulce, and north
to Penonomé.

p. 376. (Mexico merid. = Orizaba, Veracruz, designated by Brodkorb, Journ.
Washington Acad. Sci., vol. 33, No. 2, Feb. 15, 1943, p. 34).

no. 8 BIRDS OF PANAMA AND COLOMBIA—-WETMORE II

Remarks.—This interesting subspecies is described from Io speci-
mens taken near Anton and Penonomé. From the race of the species
found in southern México, named Sicalis luteola mexicana by Brod-
korb (in the reference cited above), which is known from Morelos
and Puebla, the form described here is distinguished by brighter color,
and by slightly smaller size. The Panamanian form furnishes an
interesting link between the races of South America and those of the
southern half of México.

This bird was first recorded through a sight observation near
Penonomé late in January 1951, by Dr. R. T. Scholes, who recog-
nized that it was unknown. Eugene Eisenmann and John Bull, in
July 1952, found it fairly common, several small colonies being
located. In crossing through this area in May 1953, I collected one
near Anton, and later, on June 20, I secured the rest of the series
from which this description was written. On the June excursion I
had the pleasure of the company of Dr. Eisenmann, in whose honor
the race is named in recognition of his studies of living Panamanian
birds.

The birds are found in the nesting season in little colonies that
may be overlooked because of the brilliant light of the savanna areas
which often obscures the yellow breast color, so that the Sicalis may
be confused with the seed-eaters that abound in the same habitat.

OTHER ADDITIONS TO THE LIST OF BIRDS RECORDED
FROM COLOMBIA
Crax daubentoni Gray:
Crax Daubentoni G. R. Gray, List Birds Brit. Mus., pt. 5, Gallinae, 1867, p. 15
(Venezuela).

M. A. Carriker, Jr., found these birds fairly common in the forested
foothills of the Sierra Negra, where he collected specimens near
Monte Elias, Magdalena, August 13, and at El Bosque, Guajira, above
Carraipia, in the Montés de Oca, June Io and 14, 1941. The occur-
rence is to be expected since the species has been taken in the drainage
of the Rio Negro on the Venezuelan side of the Sierra de Perija.

Chaetura chapmani viridipennis Cherrie:

Chaetura chapmani viridipennis Cherrie, Bull. Amer. Mus. Nat. Hist., vol. 35,
May 20, 1916, p. 183 (Doze Octobre= Doze de Outobre, Mato Grosso,
Brasil).

A female taken at El Real, on the Rio Nechi, March 10, and a
pair from Taraza, shot April 28, 1948, constitute the second report

I2 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

of this race since its description from specimens taken in Mato Grosso.
Both of the Colombian localities are in Antioquia.

Myiarchus ferox venezuelensis Lawrence:
Myiarchus veneguelensis Lawrence, Proc. Acad. Nat. Sci. Philadelphia, 1865,
p. 38 (Venezuela).
A female taken near Nazaret, Guajira, in the foothills of the
Serrania de Macuire, on May 13, 1941, is a well-marked example of
this race.

* {
oh

‘

LT sth a

yi
o
ny
£

Ny
i rey:

SMITHSONIAN MISCELLANEOUS COLLECTIONS
VOLUME 122, NUMBER 9

INSECT METAMORPHOSIS

R. E. SNODGRASS
Collaborator of the Smithsonian Institution and of the U. S. Department of Agriculture

qeoeeeesens

mE INCE S

ot RES
peu SOW MSiO,
zn O09 <NGEA ANo re
TM We nN

(Pustication 4144)

CITY OF WASHINGTON
SO he cra BY THE SMITHSONIAN INSTITUTION
APRIL 1, 1954

|
is

SMITHSONIAN MISCELLANEOUS COLLECTIONS
VOLUME 122, NUMBER 9

INSECT METAMORPHOSIS

BY
R. E. SNODGRASS

Collaborator of the Smithsonian Institution and of the U. S. Department of Agriculture

Veo
Lpatagor MON y
ITvt
ING vone

een.

(PusticaTion 4144)

CITY OF WASHINGTON
PUBLISHED BY THE SMITHSONIAN INSTITUTION
APRIL 1, 1954

TBe Lord Baltimore Press

BALTIMORE, MD., U. 8. A.

CONTENTS

Page

PME CaN ERO ine aw diovan ccs’ aio mises Pare ele raid citareiaad ean OSE I

leeWetamonphosissanduclassitication. ssiseact eae cess os ee otis sae A 10

iePElormones and) metamorphosis yates, ceisler esses om eeie eon 13

SEETetaiy Cells: OF cEe IT AINS. 6 ssc acd era fois Ween ao xise ae ee avaeoe 14

Mhescorporascandiaca Oh pakacandiacar satin vce deoue a eoteele 14

eRe sconpotay alata ancpr cto ialune canbe hn einer Aieleys Gavolslalenparsuert 16

Berm Candialer clan Secure tess pvers ershnna) slaen Sik tai eeeste Sievers ane eer 19

Wentralmeta ids} oibhemheadantarialccareverctaieve ries jiereie clei onshererstons 19

bea protnoraGicrclamdsie eves eysevrestorsheee euas ees is: anste eusioneetciorebeisie 20

The ringieland of cyclorrhaphous| Diptera. ......2¢2...06++. 2I

siiesnatine oi honmotalivacti Omen. cmicyyycciteri von eleeiei oa cere 23

EE UCEMBOLAT Cae cal, ea ec ce cesta ot nai oes dia arcie te ott dis rd ec onneNe 24

Mer apleetidire totes Sica ce ec cree em eS wie ab dvd'e 3.4 deta ee la tad 26

CY MESSI VGN GTe7a) 011) 22 RM SRO Toa a 28

Sea Mme ard at eR NR oes rat tard recto caay Ace cia arian cst Sebrota Ky ci ibia: ‘sud, & Loa! aVnanaiages As 32

WAITS LEIS isc nie re SOS aemase Selo ac otto ae On cp aC Oe Mero eens cerceiets 38

UMP SAMIDE CLA Urs ce os nat? chefarain mia, nection oie yaveiotial o/5,-n wb'Sla ot sfelnini as o'S Sale eke 44
IX. Oligoneoptera, or typical Endopterygota: Neuroptera to Hymenop-

CTPA ONS eaten Har one re eT ee alee Tee vaLet ete aretcbaha’es'el eeeeursh ae taketh eatin ca 48

NEMS atvel uM eELehOMOnPMOStS pawn eede:., Salp Gis ste Meroe oisinvc oath a Sa Meters ool 61

Bamdisites! with) a plcinidial Stage ics W100 2 4: sis a ateis, o- See euels 35006) 4) sels 62

(Cooley nwercee* Woe Reser e eircom te nein Raina Cee eee eee 62

INSTR OP et atemtee at reyeicr sect aeesieses ices esta sre tee Glas ierciccrnie oeerere 66

SEREDSIDLCH Mere aie een Re Ele erie CP Tate ey one Rrsiaiois ave sistetortle 66

ILCepopKalay op det Rese A tas cee Rh chenclen Sci SPARE Oot EERE CREM PCICR TE RoR MER ae 70

DD) Tpit paver asectatstres sts scederacekcte cassie aictiake sackevana seein eaten fe nie sarees 70

PASEO HELA hoes sail cieucins gcibeataieeecaln ake Roa wc ate aicethels ovo, es 72

Parasites without a. platidial Stage. v.05 5 cdjeiewe ae creo sos ants sae 75

DLE RA eee teveie teehee eo ctei ae ceeke aPaie one wee Kevenetene lenceria eae ature needs 75

Telia enOPtetae ae tive eaters eee Sta e erate erie ©! bos Men mee arots: Shenae ae

Nie he pupaletranstonmationy.scases oe ao. ee Bet eg ete ee MA ee 82

abl ote inte ey siaalks ap ic.e ic OO CIID Pe AO EORTC cach COE Ce eaceree 83

“bnesennoreinab yes! 3 MAIRs bo bodicdn GUC SOD ORE oon Ona ce moma cc Ger 86

PMO MAL IMEI CATV: CATIA cya src io. seteteeve 5 cin Nsvolsvere nr ateveielo orate lsusievevanerets 87

Ahem Nal protitanariDUles ee stonieure ere «cic oles sie gusnsvere (afetis ol steel tee efor 93

MRI ek Ate DOC Yeae ernst ye Ste aie ire oe eunie Rares, wis ahalietielovsiigie/ayetolasatatels 95

PEE FOETIOGYLESHE Mra reke cues cke hole te clea Lone olicvs lojate She; avelaaheb ater alareralerers 96

Ahepthacheallicvstenmy ayy yorisichavscuscs sie afetale isin focuser e ayetetenaneysueiers 07

Mitendonscale Dlomadiavesse lee nayacrcisieeiers.ocieieicicvoialeieie nerreraccueteleearerc ire 98

PES WAC NOMSUSU SERED (c)5's)- 34 0/< cicielna’s tats bpnie syeiets eis aio. aye absha mleiiahol ad 99

EMeNTMUSCIlame SySTEIM | oleic cis cvevs.s «is; opel lsle wleveresshelaralelsieyanareationaiets 101

XII. Muscle attachments and the nature of the pupa................... 107

IRELELENCES HRs Reese OR CUNT cara recs uals teins eialinierts ayavevete cle memteteitovate algietereiaaG III

ili

"! he » fh 4% aces

vane be ao ‘i nee
Te VIE WIE
ri

INSECT METAMORPHOSIS

By R. E. SNODGRASS

Collaborator of the Smithsonian Institution and of the
U. S. Department of Agriculture

INTRODUCTION

Ancient mythologies are replete with stories of the transformation
of one creature into another, called metamorphosis. So the early stu-
dents of natural history who first observed a caterpillar turn into a
butterfly had a term ready made for the phenomenon they witnessed,
and today in entomology we commonly think of metamorphosis as the
transformation of a larval insect into the imago. In so doing, however,
we overlook the fact, quite as extraordinary, that a caterpillar hatches
from the egg of a butterfly. We might truly say, then, that the real
metamorphosis in the life history of the species is that which has
changed a young butterfly into a caterpillar, the subsequent change
of the caterpillar into the butterfly being merely the return of the
metamorphosed young to the form of its parents. The transforma-
tion of the caterpillar into the butterfly is a visible event re-enacted
with each generation; the change of the young butterfly into a cater-
pillar has been accomplished gradually through the past evolutionary
history of the Lepidoptera. Today, there is not even any recapitula-
tion of the butterfly stage in the ontogeny of the caterpillar ; the but-
terfly’s egg develops directly into the caterpillar form of its species.
The idea that the caterpillar, because of its abdominal “legs,” repre-
sents a primitive stage in the ancestry of insects is quite out of har-
mony with the modern structure of the caterpillar’s head and with
the fact that the caterpillar has wings developing beneath its cuticle.
Both the caterpillar and the butterfly are modern end results of evolu-
tion, but along different lines of development.

In attaining their present distinctive forms, the butterfly has fol-
lowed out the evolutionary path adopted by its adult ancestors, and
therefore represents the adult line of descent; the caterpillar, on the
other hand, in its evolution has departed from the ancestral path and
has become a new and distinct juvenile form of its species. Since the
caterpillar leads an independent life in a very efficient manner as an

SMITHSONIAN MISCELLANEOUS COLLECTIONS, VOL. 122, NO. 9

2 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

individual, it would seem that it might be capable of developing its
reproductive organs to maturity and thus dispensing with the butter-
fly stage entirely. The caterpillar, however, has limited powers of
locomotion; the winged butterfly, therefore, retains the reproductive
function because it can widely distribute the eggs for the next genera-
tion of caterpillars and thus prevent overpopulation in any one place.
The same principle applies to all winged insects, and it is easy to see
the advantages insects have attained in acquiring wings, together with
specialized types of feeding organs and organs for mating and egg
laying. It is to be presumed then that the specialized forms and habits
of many young insects also are of some advantage to the species as a
whole. In short, we can readily perceive a reason for metamorphosis,
but how the differences between young and adult have come about,
and how two distinct creatures can develop from one egg are questions
difficult to answer.

Since there can be no doubt that the early insects lived on land and
developed without metamorphic changes, the metamorphoses that we
know among modern insects are of relatively late origin, and have no
relation to the more primitive metamorphoses of the annelids and
crustaceans. Even among the insects themselves metamorphosis has
been independently developed in several groups, though the reason
for it may be deduced from pretty much the same premises in all
cases.

Wherever a pronounced metamorphosis occurs in the life of an
insect it is generally, but not always, true that the young and the adult
lead different lives or inhabit different media, and are structurally
modified in adaptation to their individual habitats or ways of feeding.
Very probably the presence of functional wings only in the adult
stage, or conversely, the flightless state of the young, was an impor-
tant condition that led to structural differentiation between the juve-
nile and imaginal stages. The winged adult insect has opportunities
of extending its activities far beyond the range of the wingless young
insect, and, as is amply shown in modern insects, various new ways
of life are open to the winged insect if it is free to develop special
structures, particularly feeding organs, that enable it to take advan-
tage of them. Likewise, to the wingless young insect there are pre-
sented in nature various possible habitats and sources of food, some
of which might better its condition if it were free to make the ana-
tomical adjustments that would accommodate it to a new way of liv-
ing in some special environment. However, as long as the usual
mechanism of inheritance makes it necessary that newly acquired
adult characters be transmitted through the young, and that charac-

NO. Q INSECT METAMORPHOSIS—SNODGRASS 3

ters acquired in the juvenile stage must be passed on to the adult,
neither the adult nor the young could be free to develop structures
that would be detrimental to the other. Consider, for example, the
plight of the caterpillar if it had to inherit the mouth parts of the
butterfly, or that of a young mosquito equipped with blood-feeding
organs but lacking wings. The adult flea, it is true, has mouth parts
highly specialized for a blood diet, and still is wingless, but it has sub-
stituted the power of jumping for that of flight. The Hemiptera are
another exception to the rule that specialized adult mouth parts de-
pend on wings, but here the mouth parts are just as practicable to
the flightless young as to the adults.

A prerequisite of metamorphic differentiation between the young
and the adult, therefore, is the inhibition of some of the ordinary
processes of heredity. The young insect can then vary to any extent
by the development of adaptive structures for its own use so long as
its new characters are suppressed at the change to the adult; and the
adult, on its part, can acquire special feeding organs that would be
entirely impracticable to the wingless young. The individual, further-
more, thus derives whatever advantages there may be in living a
double life, or that may accrue from inhabiting successively two dif-
ferent media. Moreover, the different specializations of the young
and the adult may be mutually advantageous and therefore beneficial
to the individual as a whole. The larva, for example, usually becomes
the chief, and sometimes the only, feeding stage, thus giving the
adult a large measure of freedom for the functions of mating and
procreation.

When the divergence becomes too great between the young insect
and the adult, especially with regard to habitat, a liaison between the
two must be established by compensating instincts in order to main-
tain continuity of the individual life history. The adult female, for
example, must know where to lay her eggs so that the emerging larva
shall find its proper food, or be in its appropriate environment. The
egg-laying instinct must be more and more precise as the habits of the
larva become more restricted. The parent of an aquatic larva has
only to deposit her eggs in some suitable place in the water, but the
parent of a parasitic larva specific to some particular host must be
able to insert her eggs into a member of this same host species. It
has been shown by Thorpe (1938, 1939) that the egg-laying response
of the adult may be a result of conditioning during the feeding of
the larva. Likewise the larva, on its part, must be endowed with an
instinct that brings it to undergo its transformation at some place

4 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

appropriate for the ecdysis of the winged adult. Most aquatic larvae
come out of the water to pupate, some crawl up on rocks or plant
stems, others travel inland ; the parasitic larva emerges from its host ;
overwintering species find protection from the cold in concealed places
or within the ground, and do not transform until the return of warm
weather.

Changes of structure adaptive to environmental conditions, how-
ever, are not limited to the postembryonic stages of insects. The em-
bryo itself may acquire adaptive characters as well as the larva. The
embryo is commonly said to recapitulate ancestral stages in the evolu-
tion of its species, but, shut up in an egg shell, it can hardly be ex-
pected to follow in all ways the course of evolution that was practical
to its free-living progenitors, and it needs special features for its own
purposes. The insect embryo, for example, may have amniotic folds
for protection, perhaps a trophamnion for its nourishment, a tooth
on its head for breaking out of the egg shell. Then there are those
embryonic organs on the first abdominal segment of some insects
known as pleuropodia, but which take on special embryonic functions
quite foreign to the usual purpose of a leg. All such adaptations of
the embryo to life in an egg shell are just as truly aberrations from
phylogenetic evolution as are the adaptive characters of free-living
larvae that fit them to their particular environments, such as gills of
aquatic species or the abdominal “legs” of crawling and climbing
species.

A most interesting case of adaptive embryonic metamorphosis is
seen in scorpions of the family Scorpionidae (Mathew, 1948; Vachon,
1950, 1953), in which the eggs are provided with very little yolk.
The embryos undergo their development in follicles of the ovarial
tubes, and are nourished on material from the blood of the mother
absorbed into slender apical diverticula of the follicles. Each diver-
ticulum is traversed by an inner feeding tube reaching to the mouth
of the embryo. As a special adaptation on the part of the embryo, the
movable digits of the chelicerae take the form of flat pads or long
vesicular arms that clasp the feeding tube and bring it against the
mouth, into which the food material is sucked by the muscular pharyn-
geal pump. At birth the young scorpion retains the embryonic
chelicerae until the first moult, when these organs revert to the adult
form.

The embryonic modification of phylogenetic evolution forced upon
the embryo because of its development in the egg is well illustrated
by the manner in which the insect embryo commonly forms its stom-
ach. The food of the embryo, the yolk, is stored in the egg and thus

NO. 9 INSECT METAMORPHOSIS—SNODGRASS 5

comes to be inside the body of the embryo; consequently it cannot be
ingested in the ancestral manner by way of the mouth. In embryonic
development, therefore, the stomach grows around the yolk, a method
of “gastrulation” certainly that does not in any sense recapitulate
stomach formation in the evolution of a free-living ancestor whose
food had to be taken in from the outside. The embryo simply fol-
lows a modified process of gastrulation in adaptation to life in an egg
shell, but in the end it produces an alimentary canal the same as that
which its free-living ancestors produced by quite different evolu-
tionary methods.

The insect embryo may develop into a juvenile form resembling its
parents except in matters of immaturity, such as the rudimentary
nature of the wings and the organs of reproduction. In such cases
the young insect successively approaches the adult structure at each
moult and finally assumes the imaginal form. At the other extreme,
the embryo throws off all adult ancestral influences and develops into
a creature having no likeness to its parents. There is here no phylo-
genetic recapitulation, the young insect in its growth takes no steps
toward the adult structure; development of the adult, except for the
growth of invaginated appendage rudiments, is inhibited until the
young insect has accomplished its particular function in the life his-
tory of its species. Then the juvenile tissues disintegrate and the
imago is rapidly built up in the form of its parents.

Inasmuch as the terminology of metamorphosis is not standardized,
the same names being used in different ways by different zoologists,
it will be necessary before proceeding with a further discussion of
insect metamorphosis to explain a few common terms as they will be
applied in the following pages.

Metamorphosis—The word “metamorphosis” means merely a
“change of form.” In general zoology any pronounced change of
form during growth, such as the changes of a crustacean larva in its
development from the nauplius to the adult, or the changes of a tad-
pole in becoming a frog, is called metamorphosis. Entomologists, on
the contrary, are inclined to restrict the idea of metamorphosis to the
final change from a differentiated juvenile form to the imago, whether
the change is direct or accomplished in an intervening pupal stage.
Such a definition of metamorphosis is clearly too restrictive, since it
would eliminate the use of the term as commonly used in other
branches of zoology, and even in the insects there may be pronounced
metamorphic changes between larval instars. Students of the action
of hormones in the postembryonic development of insects commonly
refer to the change to the adult as the “metamorphosis” of the insect

6 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

regardless of the degree or nature of the change. Bodenstein (1953),
p. 879), for example, says: “We speak of metamorphosis when the
animal shows adult characters after a molt.” Wigglesworth (1953a)
finds that in Rhodnius, in addition to the full development of the
wings and reproductive organs at the final moult, the epidermal cells
now lay down an imaginal cuticle that is totally different in structure
and color pattern from that of the nymph, and hence he calls the
changes at the last moult the metamorphosis of the insect. Such final
changes as those that occur in most of the Heteroptera, however,
would appear to involve merely the completion or final acquisition of
adult characters, and are therefore not comparable to the metamorphic
changes in other insects resulting from the suppression of juvenile
aberrations in form or structure.

It will probably be useless to attempt to write a definition of
metamorphosis, since none would be generally acceptable. With nearly
all insects there is necessarily a change of some kind or degree from
one instar to the next, since the insect grows by stages, and the change
at the last moult is usually greatest because the insect now takes on
the fully developed adult characters. However, regardless of defini-
tions, we must distinguish between changes that are consequent on
growth from youth to maturity, and those that result from structural
aberrations on the part of the young insect from the direct line of
development. True metamorphic characters, as here understood, are
adaptive structures, temporarily assumed usually by the young insect
for its own purposes, that have no phylogenetic counterpart in the
adult evolution, and which are discarded at the transformation to the
imago. Metamorphic changes may take place between the immature
stages of the insect, but metamorphosis is most pronounced at the
change to the adult because it now involves the assumption of imagi-
nal characters as well as the discarding of juvenile characters. How-
ever, if the assumption of imaginal characters alone is called “meta-
morphosis,” then all insects undergo metamorphosis in some degree at
the last moult, and the term has no specific meaning.

If the metamorphic change between the young and the adult is of
small degree it is termed paurometabolism. If the young insect differs
conspicuously from the adult or has distinctive adaptive characters of
its own, but still makes the change to the adult at one moult, the in-
sect is said to be hemimetabolous. When two moults are involved
in the change and a pupal stage thus intervenes between the young
and the adult, the insect is said to be holometabolous. These terms,
of course, have no literal significance. An insect may be classed as
ametabolous if it goes through no changes during its development

NO. 9 INSECT METAMORPHOSIS—SNODGRASS 7

that are not related to growth from youth to maturity, but such
changes may be considerable and are often difficult to distinguish
from paurometabolism.

Nymph.—In its biological application this term is almost exclu-
sively entomological, but entomologists are not consistent in its usage.
According to American and most English usage a “nymph” is gen-
erally the young of an insect without a pupal stage, while with Euro-
pean entomologists a “nymph” is more commonly the pupa. In the
following discussion the term nymph will be limited to the young of
ametabolous or paurometabolous insects that in all essential respects,
except those of immaturity, resemble their parents and have no
important characters that obscure their likeness to the adult.

Larva.—tin general zoology the term “larva” is commonly used to
designate the immature stages of any invertebrate animal, or even
the tadpole stage of frogs and salamanders. Some entomologists limit
its application to the young form of insects that have a pupal stage in
their life cycle; others call any juvenile insect a “larva.” Definitions
may be arbitrary, but it is better if a scientific term has some relation
to the original meaning of the word involved. If we take the word
“larva” in one of its Latin meanings, that of a mask, it becomes an
appropriate term for any young form, particularly of an arthropod,
that differs so much from its parents that its identity is not apparent
in its structure, being “masked” under a specialized juvenile disguise.
A larva, in this sense, may be defined specifically as an immature post-
embryonic stage that has acquired for its own use adaptive characters
that its adult ancestors did not possess, and which are not carried over
into its own winged instar. Unfortunately, the insects will not always
conform with definitions. There are some young insects that are es-
sentially nymphs, and yet have a few special characters of their own.
Such borderline cases, however, only show how easily a nymph might
become a larva.

True larval forms among modern arthropods occur principally in
the crustaceans and the insects, but in these two groups the larvae are
not equivalent ontogenetic stages. The crustacean larva in most cases
is hatched at an early stage of embryonic development long before
body segmentation is completed. The earliest larval form in the Crus-
tacea is the nauplius, a minute creature without body segmentation, but
provided with three pairs of appendages, which are the first and
second antennae and the mandibles, a simple nervous system, a single
median eye, and an alimentary canal with oral and anal apertures.
The swimming nauplius serves for the distribution of its species, and,

8 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

though it is derived from an early stage of ontogeny, it is specifically
modified in adaptation to an aquatic life, and therefore in its form
and structure does not recapitulate any primitive ancestral form in
the evolution of the Crustacea. In its growth the crustacean larva
goes through subsequent stages in which body segmentation appears,
and both the segments and the appendages increase in number until
the final organization is attained. The young of a terrestrial animal
could not survive if hatched at such an early stage of development as
that of the nauplius. The insect larva, with a few exceptions among
parasitic species, is hatched with the definitive body segmentation,
and thus in its youngest stage represents a relatively late period of
development. The typical crustacean larva is anamorphic, the insect
larva is epimorphic. Here are two more terms that will need some
attention farther on.

In both the crustaceans and the insects the larva may be hetero-
morphic in that it develops through a series of different forms. In
the Crustacea the heteromorphic larva progresses toward the adult
structure ; with the insects successive larval forms are adaptations to
different functions or living conditions of the larva itself and have
no relation to the adult. Larval metamorphosis among the crustaceans,
however, especially in parasitic species, is often retrogressive and
ends in the production of a greatly modified or highly degenerate
metamorphosed adult form. With the insects, simplified, or “degen-
erate,” forms occur mostly in the early larval stages of heteromorphic
parasitic species, which have normal adults.

Pupa.—there is no ambiguity in the use of this term; the pupa is
the stage of a holometabolous insect in which the final development
of the imago takes place. There is, however, a difference of opinion
as to the nature of the pupa. A common idea is that the pupa repre-
sents the last nymphal instar of an ametabolous insect ; another is that
it is a condensation of all the former nymphal instars of its species;
a third sees in the pupa a preliminary sketch of the adult furnishing
a mold for the proper reconstruction and attachment of the adult
musculature. The respective merits of these several pupal concepts
will be discussed later.

The degree to which reconstructive processes take place in the
pupa varies with different insects. In some cases most of the larval
tissues are merely made over into corresponding parts of the adult,
in others the larval tissues go into a state of dissolution and the adult
organs are built up from special groups of undifferentiated embryonic
cells, called imaginal discs or histoblasts, which are carried by the
larva but form no essential part of the larval structure. That the larva

NO. 9 INSECT METAMORPHOSIS—SNODGRASS 9

is a “double” organism, as it is often said to be, is thus seen to be
true only in the more specialized members of some of the holometabo-
lous orders. The embryo, however, is charged with the double poten-
tiality of forming first a larva and then an imago; the larval structure
is completed in the egg, the latent adult structure is built up in the
pupa. Of all the reconstructive processes that take place in the pupa
the most important is that of the muscular system, which is perhaps
the primary reason for the pupa. The dissolution of the larval muscles
before the imaginal muscles are formed at least accounts for the im-
mobile condition of the pupa, though, since all the larval muscles are
not destroyed at the same time and some may go over intact from the
larva to the adult, various pupae retain some degree of activity.

The tmago.—Rarely does the adult insect undergo any metamorphic
changes after it emerges from the pupal skin. There is the curious
case of the streblid fly Ascodipteron, however, which is parasitic on
bats. As described by Jobling (1939) the female fly pierces the skin
of the host with her enormous proboscis and pulls her body into the
wound. The legs and wings are then cast off, while a circular fold of
the integument grows forward over the abdomen and thorax until the
body acquires a flask-shaped form. On a posterior setose knob of the
body, which alone projects from the skin of the host, are situated six
spiracles and the slitlike aperture of a chamber containing the open-
ings of the vagina and rectum.

Anamorphosis—The term anamorphosis, as usually defined, refers
to the completion of body segmentation after hatching. Though
anamorphosis thus involves a “change of form,” it should not be con-
fused with metamorphosis; it is merely a way of growing. The man-
ner by which body segments are formed in anamorphic development
is always essentially the same. Just anterior to the terminal lobe of
the body, or telson, is a mass of undifferentiated tissue, the zone of
growth, which is capable of active cell proliferation, and it is here
that the new segments are generated. As each new segment is formed
it lies between the segment before it and the zone of growth, so that
the animal extends its length posteriorly, but the anterior segments
are the oldest. This method of growth from behind forward, which
may begin in the embryo or be completed in the embryo, is in general
known as telogenesis. Anamorphosis, by definition, therefore, is telo-
genesis continued after hatching. The number of segments added by
anamorphosis depends on how many segments the young animal has
on hatching and on the number of segments it will have when mature.

Anamorphosis is characteristic of the polychaete annelids; it was
the mode of development in the trilobites ; it still prevails in most of

Io SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

the crustaceans, in two groups of chilopods, in all the diplopods, pauro-
pods, and symphylans ; and a remnant of it persists among the hexa-
pods in the Protura. It would seem, therefore, that anamorphosis
was the primitive method of growth in the arthropods, and that it is
an inheritance from their remote common ancestry with primitive
annelids.

The addition of new segments in the arthropods takes place at the
moults, and is usually accompanied or followed by the formation of
new segmental appendages. If the growing animal takes on a dif-
ferent form or distinctive characters at successive anamorphic stages,
- as is common among the Crustacea, such features are metamorphic
aberrations or adaptations superposed on anamorphosis. Anamor-
phosis, therefore, may be accompanied by larval heteromorphosis.

Epimorphosis—The development of an animal is said to be epi-
morphic when the maximum number of definitive segments is present
at hatching, though some segments may be suppressed later. The
segments in some cases are formed teloblastically as in anamorphosis
by generation from a subterminal zone of growth, but in most epi-
morphic arthropods, as in insects, the prospective body is first laid
out as an unsegmented germ band, which later becomes segmented.
Segmentation in the germ band commonly begins anteriorly and pro-
ceeds posteriorly, and the segmental appendages appear in the same
order. In this case, therefore, metamerism might appear to have no
relation to a supposedly primitive anamorphic method of growth, and
the anteroposterior progress of development has been regarded as
indicative of a “metabolic gradient” in the embryo, meaning that the
developmental processes are most intense first at the anterior pole
and proceed posteriorly. However, since in anamorphic growth the
anterior segments are the oldest, the apparent formation of segments
and appendages from before backward in an epimorphic animal may
be merely the visible results of delayed segment differentiation in the
germ band. Epimorphosis is clearly a specialized and more expedi-
tious way of growing than is anamorphosis; it delivers the young
animal into the world in a more nearly mature condition, and there-
fore in a more practical stage of development for meeting the con-
tingencies of a free existence.

I. METAMORPHOSIS AND CLASSIFICATION

Insects cannot be classified taxonomically according to the type of
metamorphosis they undergo. Hemimetabolism occurs among sey-
eral unrelated orders, and holometabolism is not limited to the group

NO. 9 INSECT METAMORPHOSIS—SNODGRASS We

of orders formerly known as the Holometabola. Even among the
orders that are typically ametabolous there may be juvenile changes
during growth sufficient to warrant the term paurometabolism.

The ametabolous and paurometabolous insects include the Aptery-
gota, and, among the pterygote orders, the Dermaptera, Orthoptera,
Embioptera, Isoptera, Zoraptera, Corrodentia, Mallophaga, Anoplura,
Heteroptera, and most of the Homoptera. Growth changes among
the ametabolous insects are often fairly conspicuous, since they may
include the acquisition of abdominal styli, developmental changes in
the mouth parts, antennae, legs, wings, and the external reproductive
organs, and furthermore they may involve changes in the shape and
proportions of the head, thorax, and abdomen, accompanied by
changes in the shape of the sclerites, and possibly in the number
and arrangement of setae. Changes of this kind, however, are for
the most part merely alterations that a young animal must go through
in attaining the adult form, and are not of the adaptive kind here
treated as true metamorphosis. Marked changes in the nymphal in-
stars of an insect, furthermore, may be due to some specialized de-
velopment of the imago, as is well illustrated by the Tingitidae, in
which the apparent metamorphoses of the nymph are merely juvenile
steps leading up to the unusual form of the adult insect, and have
no adaptive significance for the immature stages themselves.

The postembryonic development typical of ametabolous insects is
well exemplified in the nymphal growth of a cockroach or a grass-
hopper. The newly hatched insect may differ considerably in shape
from its parents, but its form is the result of its having been de-
veloped in an egg, and is not an adaptation to its juvenile life. As
the young insect grows it takes on more and more of the adult form
at successive moults ; the wings grow out as padlike extensions of the
back plates of the mesothorax and metathorax, the head becomes rela-
tively smaller, the abdomen larger, and the external genitalia develop.
There may perhaps be changes of color, or minor features found only
in the immature stages, but such characters are insignificant. The
young insect generally mingles with its parents in the same habitat,
feeding on the same kind of food with the same kind of mouth parts.
The adults on their part have taken no advantage from their wings
to lead a different kind of life. In short, it may be said of the orthop-
teroid insects in general that they lead the normal life of most other
animals instead of adopting a dual existence as do those with meta-
morphosis. They should, therefore, be the direct descendants of more
primitive winged insects, the young of which never wandered from
the parental habitat, or took on a form or characters that had to be

I2 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. I22

discarded at the moult to the imago. In a study of metamorphosis the
orthopteroid insects are thus of particular interest in that they show
us the simple course of postembryonic development in winged insects,
one that has not been complicated by the addition of juvenile charac-
ters for the specific use of the young.

The insects here classed as hemimetabolous are the Plecoptera, the
Ephemeroptera, the Odonata, and some of the Homoptera. The
metamorphoses of these insects undoubtedly have been developed in-
dependently in each group; they have nothing in common and need
no further discussion here since each order will be treated in a sepa-
rate section following.

The holometabolous insects include the males of Coccidae, the
Thysanoptera, the Neuropteroidea, Coleoptera, Strepsiptera, Tri-
choptera, Lepidoptera, Mecoptera, Siphonaptera, Diptera, and Hy-
menoptera. The presence of a pupal stage in the life history is diag-
nostic of holometabolism, but it is probable that the pupa is not in
all cases a homologous stage. It is the intensity or degree of the trans-
formation processes, particularly the reconstruction of the muscular
system, that characterizes holometabolism and makes a resting stage
necessary between the larva and the imago.

The larvae of holometabolous insects are endopterous and some of
them are endopodous, that is, they have no external wing rudiments,
and may have no functional legs. The “wingless” condition of the
larva, as well as the “legless” condition, however, is apparent rather
than real, since wing and leg rudiments are usually present but con-
cealed within pouches of the epidermis beneath the outer cuticle. A
truly apodous larva, therefore, is rare or perhaps does not exist, and
probably the only wingless larvae are those of insects that have no
wings in the adult stage. Wing rudiments, however, are sometimes
present on the pupae of wingless adults, and in such cases are pre-
formed in the larva.

The endopterous condition of the larva is not entirely characteristic
of any particular taxonomic group of insects. In the Coccidae the
wings of the male do not appear externally until the third or fourth
moult, and in Aleyrodidae they do not become external until the last
moult. The winged males of Embioptera also, as shown by Melander
(1903), develop their wings internally up to the last nymphal stage.
In the case of the male coccids and the aleyrodids a variable degree
of metamorphosis, aside from the wing and leg development, may
accompany the larval growth, but the embiids show no juvenile
changes that do not lead up to the adult structure, It is evident that

NO. 9 INSECT METAMORPHOSIS—SNODGRASS 13

the endopterous condition of the larva has been acquired independ-
ently by different insects, and it is questionable whether it is to be
regarded necessarily as a metamorphic feature, or merely as a device
for protecting the wings during the early stages of their growth.
Clearly it is an advantage to the young insect to have temporarily
useless appendages removed from the surface.

Inasmuch as the type of metamorphosis that an insect goes through,
or whether the young insect is exopterous or endopterous, does not
in all cases conform with the insect’s taxonomic relationships, it will
be more appropriate to discuss the metabolous insects according to
their usual classification rather than according to their kind of meta-
morphosis. Among the ametabolous insects special attention must be
given to the Apterygota, because certain features of hemimetabolous
and holometabolous larvae have been thought to be derived from adult
ancestral forms resembling the modern thysanurans.

II. HORMONES AND METAMORPHOSIS

The transformations of insects have long furnished a popular theme
for writers on the “marvels of insect life,’ but in recent years serious
investigators have given more and more attention to the vital mecha-
nisms that control the phenomena of metamorphosis. Though their
studies have not eliminated the mystery, they have revealed some-
thing of its nature, and insect metamorphosis has now become a sub-
ject for experimentation rather than one that merely excites our visual
curiosity. The young insect contains two opposing forces in the nature
of hormonal secretions that regulate its growth and development;
one maintains the juvenile status, the other stimulates moulting and
normal development that culminates in the production of the imago.

Though insect endocrinology is still a youthful science, it has many
devotees. The insects are excellent experimental subjects; they sub-
mit to amputations, graftings, and transplantations without complaint
and apparently without discomfort. It would go too far beyond the
scope of the present discussion to list the great number of papers now
available on the endocrine organs of insects, or to review all the ex-
perimental evidence of the action of hormones in controlling the
metamorphic processes. The student may find ample bibliographies
in the more recent papers to be cited in connection with the following
summary of what may now be regarded as known concerning the or-
gans of internal secretion and the hormones that regulate meta-
morphosis.

The endocrine organs of insects that control nymphal and larval

I4 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

growth, moulting, development, pupation, transformation of the
nymph or larva to the imago, and the ripening of the eggs in the
ovaries include the following: (1) Secretory cells in the intercerebral
part of the brain, (2) the corpora cardiaca, (3) the corpora allata,
(4) pericardial glands, (5) perhaps glands in the posterior ventral
part of the head, (6) thoracic glands, and (7) the ring gland of
cyclorrhaphous Diptera.

Secretory cells of the brain—In insects of most of the principal
orders secretory nerve cells that play an important part in moulting
and imaginal development are present in the pars intercerebralis of
the protocerebrum. According to Scharrer and Scharrer (1944) such
cells have been shown to be present in Orthoptera, Hemiptera, Neu-
roptera, Coleoptera, Trichoptera, Lepidoptera, Hymenoptera, and
Diptera. In the blattid Leucophaea, these authors observe, some of
the cells of the pars intercerebralis contain varying numbers of dis-
tinctly staining colloid inclusions, which are continued for some dis-
tance into the cell axons. The fibers from the secreting cells go down-
ward in the brain, where most of them cross from one side to the
other, and then turn backward through the nerves of the corpora
cardiaca to innervate these bodies. It has been shown by Wiggles-
worth (1940) and others that the brain secretion has to do with the
induction of moulting, but from further research it is now known
that moulting and imaginal development depend on a hormonal com-
plex derived from the brain and the prothoracic glands. According
to Williams (1948) there are two groups of secretory cells in the
larval brain of the Cecropia moth producing two different hormones,
both of which are necessary to induce moulting.

The corpora cardiaca, or paracardiaca.—The corpora cardiaca (fig.
1 A,Cc) are usually paired oval or elongate bodies lying behind
the brain, with which they have nerve connections, and are closely
attached to the sides of the aorta. They arise, however, as cellular
outgrowths from the dorsal wall of the stomodaeum at the sides of a
similar median outgrowth that becomes the hypocerebral ganglion of
the stomodaeal nervous system (hcGng). According to Pflugfelder
(1937) the corpora cardiaca in an early embryonic stage of the phas-
matid Dixippus lie against the lower surfaces of the cardioblasts, but
when the cardioblasts unite to form the aorta, they push into the aortic
wall; the lower cells remain as compact masses which are soon differ-
entiated into ganglion cells, while the others appear to be secretory.

An extensive comparative account of the corpora cardiaca in most
of the principal groups of insects is given by Cazal (1948), who more

NO. Q INSECT METAMORPHOSIS—SNODGRASS 15

appropriately calls these bodies paracardiaca, since their connection
with the heart is entirely secondary. Typically each corpus cardiacum
is connected with the back of the brain by two nerves (fig. 1 A,ccNvs),
one lateral, the other median. The lateral nerves have their roots in
the lateral parts of the protocerebrum, the median nerves arise in the

Fic. 1.—The retrocerebral endocrine organs.

A, diagram of a simple, perhaps generalized, arrangement of the corpora car-
diaca and corpora allata on dorsal surface of stomodaeum behind the brain in
association with the hypocerebral ganglion. B, diagram of ring gland of larva
of Calliphora (from M. Thomsen, 1951).

AntNvy, antennal nerve; Br, brain; Ca, corpus allatum; Cc, corpus cardiacum ;
ccNvs, corpus-cardiacum nerves; frGng, frontal ganglion; hcGng, hypocerebral
ganglion; ImNvy, labral nerve; RugCls, ring cells; rNv, recurrent nerve; Stom,
stomodaeum; Tra, trachea.

pars intercerebralis and cross each other from one side to the other.
Nerve fibers traversing the corpus cardiacum form a nerve connec-
tion between the latter and the corpus allatum of the same side.
Because of the intimate nerve relation of the corpora cardiaca to
the secretory cells of the brain, and the observation that colloid
granules similar to those in the brain can be traced along the nerve
fibers into the corpora cardiaca, Scharrer and Scharrer (1944) point
out that “the pars intercerebralis and the corpus cardiacum of insects
may be viewed as one neuro-endocrine complex rather than as two
separate sources of hormones.” The presence of a brain hormone

16 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

concerned with moulting and development is now well known, but
the specific function of the corpora cardiaca has been but little in-
vestigated. It is noted by Pfeiffer (1942) that removal of the corpora
cardiaca from nymphs of Melanopflus is followed by a delay in moult-
ing, but does not prevent moulting. This observation suggests that
a corpus-cardiacum hormone is a part of the hormone system that
activates moulting and imaginal development. The most definite in-
formation we have on the action of the corpus-cardiacum hormone,
however, has to do with its effect on crustaceans. It had been known
that extracts of the head of insects injected into blinded shrimps
would cause a contraction of the chromatophores, just as does the
hormone of the crustacean sinus gland in the eyestalk. M. Thomsen
(1943) then showed that the activating element of the insect head
comes from the corpora cardiaca, since transplantation of these bodies
into a shrimp with amputated eyestalks had the same effect as head
extract.

The corpora allata.—The corpora allata (fig. 1 A,Ca) are typically
a pair of small oval bodies lying usually behind or laterad of the cor-
pora cardiaca, with which they are connected by nerves, but in some
insects the two bodies on each side are united, and the corpora allata
themselves may be fused into a single mass. Variations in the rela-
tive position and connections of the corpora allata and cardiaca, and
their association with the hypocerebral ganglion are illustrated in
various insects by De Lerma (1937), Nesbitt (1941), and Bickley
(1942) ; an exhaustive review of the structure of the retrocerebral
organs in most of the insect orders is given by Cazal (1948). The
first general description of the histology of the organs is due to Nabert
(1913).

The corpora allata arise during embryonic development from the
head ectoderm between the mandibular and maxillary segments, and
are later transposed, as the name “allata” implies, to their definitive
position behind the brain. In most insects they come to lie above the
stomodaeum ; according to Cazal (1948) they lie below the stomo-
daeum in Ephemeroptera and Odonata. Formerly it was thought
that corpora allata are absent in the Thysanura, but Chaudonneret
(1949) has given reasons for believing that small glandular bodies
in these insects lying against the outer surfaces of the adductor
muscles of the maxillae are the corpora allata in a relatively primitive
position.

The action of the corpus-allatum hormone is better known than
that of the other incretory organs. The experiments of Wigglesworth

NO. 9 INSECT METAMORPHOSIS—SNODGRASS U7

on the hemipteron Rhodnius, of Scharrer (1946a) on the blattid
Leucophaea, of Pflugfelder (1937, 1938) on the phasmatid Dixippus,
of Bounhiol (1938) on lepidopterous larvae all go to show that the
corpus-allatum secretion in the young insect is the factor that main-
tains the juvenile status. This hormone, therefore, is known as the
juvenile hormone. Wigglesworth (1951a) summarizes the results of
his experiments on Rhodnius demonstrating the inhibitory effect of
the corpus-allatum hormone on adult development as follows: “If
the corpus allatum is removed from one of the young stages and im-
planted into the abdomen of a fifth-stage larva, when this moults it
turns into a giant or sixth-stage larva instead of undergoing meta-
morphosis to an adult. Even a seventh-stage larva has been produced
this way, and some of the sixth-stage larvae have transformed suc-
cessfully into giant adults.” Scharrer (1946a) obtained the same re-
sults from experiments on Leucophaea. Removal of the corpora allata
from the last (8th) instar of the cockroach had no visible effects on
development, but removal at earlier stages resulted in an abbrevia-
tion of development and the production of small adultlike forms, the
adult characters being more accentuated with the age of the operated
insects. Pflugfelder (1938) working with Dixippus, found that re-
moval of the corpora allata from first and second instars was followed
by a degeneration of certain tissues, including the fat bodies, the meso-
dermal sheath of the nervous system, muscles, and the Malpighian
tubules. These changes are those that normally take place at the end
of larval life in holometabolous insects, showing that it is the corpus-
allatum hormone that maintains the integrity of the juvenile tissues,
and that the dissolution of specialized larval tissues is due to the
weakening or cessation of secretory activity by the corpora allata in
the last juvenile stage. Similar results have been obtained in Lepidop-
tera by Bounhiol (1938) and other investigators (see Hinton, 1951).
Removal of the corpora allata from a young caterpillar brings on pre-
cocious pupation, but removal of the organs from a last-stage larva
has no effect on pupation. It is noted by Wigglesworth (1936),
furthermore, that “the corpus allatum also determines the characters
of each nymphal instar by limiting the degree of differentiation toward
the adult form which occurs during the moults.”

In the adult insect the corpora allata again become active, but now
their secretion is operative on egg production in the female and on
secretion by the accessory genital glands in the male. The effect of
eliminating the corpora allata from the adult insect has been studied
by various investigators, including Wigglesworth (1936, 1948), Pfeif-
fer (1939, 1942, 1945), and Scharrer (1946b). From experiments

18 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

on the grasshopper Melanoplus, Pfeiffer (1939) found that com-
plete removal of the corpora allata from adult females prevents the
production of ripe eggs in the ovaries and of secretion in the oviducts.
The eggs will develop without corpora allata until they reach the
stage at which yolk deposition normally begins, but after that time
they stop development, degenerate, and are resorbed. In later work
on the same insect Pfeiffer (1945) showed that in normal females,
during the early period of adult life before yolk formation in the
ovaries, the fatty acid content of the body increases, the fat body
hypertrophies by rapid storage of fat, nonfatty dry matter increases
in correlation with fatty acid increase, and the blood volume moder-
ately increases. These changes, however, do not take place if the cor-
pora allata are removed at the beginning of the adult stage. In the nor-
mal female the metabolic processes are reversed after yolk formation
begins. From these findings Pfeiffer concludes that “the corpora allata
control egg production principally, if not entirely, through the agency
of a metabolic hormone, and that a primary function of this hormone
is to facilitate the mobilization or production of materials necessary
for egg growth.” According to Scharrer (1946b) the corpora allata
are necessary in the blattid Leucophaea for approximately the first
third of the total period required for egg development, which time
corresponds to the period of growth and yolk deposition. Reimplanta-
tion of corpora allata into females from which these organs had been
removed caused the eggs to develop and produce normal nymphs.

It is a curious fact that the corpus-allatum hormone of the adult
seems to be the same as that which inhibits adult development in the
young insect. By implanting from two to six corpora allata from
young adult females of Melanoplus into a nymph, Pfeiffer (1942)
found that the nymphs never transformed into adults, though some
of them made one or two further moults. The same hormone appar-
ently is present in both the female and the male, since Wigglesworth
(1936) reports that in Rhodnius the corpora allata of the male will
induce egg development in the adult female, and those of the female
will activate the accessory glands of the male. He concludes (1948),
therefore, that “it is probable that the yolk-forming hormone and the
juvenile hormone are identical.” On the other hand, Wigglesworth
finds that the moulting hormone of the nymph will not induce egg
formation, nor will the egg-forming hormone of the adult induce
moulting in the nymph. A dual function of an apparently single hor-
mone, Wigglesworth notes, recalls the multiple action of thyroxin in
Amphibia. In the case of the insect, however, it now appears that the
principle of “tissue competence” emphasized by Bodenstein (1943)

NO. 9 INSECT METAMORPHOSIS—SNODGRASS IQ

and by Bounhiol (1938, 1953) plays an important part in the action
of a hormone. With respect to the corpus allatum, Bounhiol (1953)
says, it is very probable that it has only a general effect on metabolism,
and that it is the variable state of sensitivity in the different organs,
or in any one organ according to its age, that determines the varying
responses. Evidently, what is a stimulus in one case may be an inhibi-
tion in another.

Though the corpus-allatum hormone acts as an inhibitor of develop-
ment in the larval tissues between moults, the rudiments of imaginal
organs developing in the larva, such as the antennal, leg, and wing
buds, continue to grow during the larval instars. Eassa (1953) gives
measurements of the antennal growth in the larva of Pieris brassicae
between moults, and notes that mitosis may be observed in the anten-
nal cells. It would appear, then, either that the larval corpus-allatum
hormone is selective for larval tissues, or that imaginal tissues are not
affected by it.

Pericardial glands——These glands were first described by Pflug-
felder (1938) in the phasmatids Dixippus and Phyllium, but later
(1947) he reported them present also in Ephemeroptera and Plecop-
tera. The glands of Dixippus and Phyllium lie in the posterior part
of the head close above the dorsal blood vessel, mesad of the peri-
cardial cells, from which they are distinctly different. The pericardial
glands, according to Pflugfelder, arise from the lateral walls of the
head coelom, and are therefore mesodermal organs. They attain their
greatest development in the last nymphal stage, and in the adult they
soon degenerate and disappear, from which facts it is deduced that
the pericardial glands are endocrine organs, though there is no direct
evidence of their function. It is probable, as will be explained later,
that the pericardial glands compose the major part of the ring gland
of cyclorrhaphous fly larvae, and that functionally they are equivalent
to the thoracic glands of other insects.

Ventral glands of the head—These organs are small glandular
bodies lying ventrally in the posterior part of the head, described by
Pflugfelder (1938) first in Phasmatidae, but later (1947) reported
as present also in Ephemeroptera, Odonata, Plecoptera, Dermap-
tera, Acrididae, Blattidae, and Isoptera. They are of ectodermal
origin and degenerate after the last moult except in the workers and
soldiers of termites. Williams (1948) suggests that the glands may
be homologous with the prothoracic glands, but it is said by Hinton
(1951) that prothoracic glands also are now known to be present in
Odonata and Orthoptera.

20 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

The prothoracic glands.—Glands of the prothorax were described in
a caterpillar by Lyonet as “granulated vessels,” and little further at-
tention was given to them until recent times. It is now well demon-
strated that these glands are important endocrine organs, probably
present in most insects; according to Hinton (1951) they are known
to occur in Odonata, Orthoptera, Hemiptera, Lepidoptera, Hymenop-
tera, and Diptera. They are said by Toyama (1902) to arise in the
early embryonic development of the silkworm as epithelial invagina-
tions of the lateral part of the second maxillary segment and to extend
into the thorax. In lepidopterous larvae the glands are loose, branch-
ing masses of cells associated with the tracheae in the sides of the
prothorax. Their structure has been described by Williams (1948)
in the larva of Platysamia cecropia, and a well-illustrated compara-
tive account of them in various lepidopterous species is given by Lee
(1948). Prothoracic glands in the hemipteron Rhodnius are described
by Wigglesworth (1951b, 1952a).

The probable function of the prothoracic glands is best known
from experiments by Williams (1947) in connection with the pupal
diapause of Playtsamia cecropia. It appears that there is an intimate
functional relation in the caterpillar between the prothoracic glands
and the brain. The pupa of the Cecropia silkworm as soon as it is
formed goes into a prolonged state of diapause, which normally is
broken only when the pupa is exposed to low temperatures. It is
shown by Williams, however, that if the brain is removed from a
diapausing pupa, chilling has no effect and further development per-
manently ceases. On the other hand, if the brain from a chilled pupa
is implanted into a brainless pupa, normal development takes place.
It is evident, therefore, that the chilling of the brain renders it com-
petent to release its developmental hormone. However, further ex-
periments by Williams showed that a pupal abdomen severed from
the thorax will not develop even if a chilled brain is implanted into
it, but when reattached to the thorax such an abdomen proceeds with
development. The head and the thorax, on the contrary, develop when
a chilled brain is inserted. Normal development, in short, requires
besides a chilled brain the presence of the thoracic glands, which do
not need exposure to cold for activation. Thus the brain, Williams
points out, evidently exerts a controlling action on the prothoracic
glands. In other words, the resumption of normal development in
the diapausing Cecropia pupa is brought about by the interaction of
a hormone from the brain and another from the prothoracic glands,
but the gland hormone, Williams says, “most probably, has the ulti-
mate action on the tissues in terminating diapause.” The same rela-

NO. 9 INSECT METAMORPHOSIS—SNODGRASS 21

tion between the brain hormone and a prothoracic hormone has been
demonstrated in the hempiteron Rhodnius by Wigglesworth (1952b).
The brain hormone activates the thoracic gland, which latter “then
produces the factor initiating growth and moulting.”

Considering, then, the intimate relation of the secretory cells of the
brain both to the corpora cardiaca and to the prothoracic glands, it is
evident that the brain is of primary importance in the activation of
imaginal development. It is to be noted, however, that in this case
the brain does not function in the usual manner by nervous control,
but through having taken on a secondary function of hormone secre-
tion. The secretory action of the brain, however, is induced by nerve
activity.

The ring gland of cyclorrhaphous Diptera—In the larvae of cy-
clorrhaphous Diptera a glandular structure surrounding the aorta
behind the brain is known as the ring gland. Though formerly re-
garded as the corpus allatum, it is now known to be a complex en-
docrine organ that includes the corpora allata and corpora cardiaca
of other insects embedded in a ring of cells of different origin. In the
lower Diptera there is no ring gland. As shown by Cazal (1948) cor-
pora allata and corpora cardiaca are present in the usual manner in
Nematocera, either separate or united. In Tabanus and other Brachy-
cera the corpora allata are united above the aorta and are connected
by nerves going around the aorta to the ventrally placed corpora car-
diaca, which are separate. In Melophagus ovinus, according to Day
(1943), the corpora allata are paired bodies in the larva and the cor-
pus cardiacum is a single median organ.

Investigators are mostly in accord as to the structure of the ring
gland in the Cyclorrhapha, and we may follow the account of the
organ given by M. Thomsen (1951). The larval ring gland of Cal-
liphora erythrocephala as illustrated by Thomsen (fig. 1 B) is tri-
angular rather than circular ; its wide anterior part is prolonged for-
ward as a median tongue above the aorta, its narrow posterior part
lies below the aorta. A trachea (Tra) enters on each side and the two
lateral trunks are connected by a commissure through the anterior
part of the gland. The major part of the organ is formed of large
cells termed the ring cells (RngCls). Within the anterior tongue in
front of the tracheal commissure is a group of small cells (Ca) repre-
senting the corpora allata of other insects. In the posterior angle of
the ventral part of the ring is a second group of small cells (Cc)
apparently representing the corpora cardiaca. The ring cells them-
selves were formerly regarded as the corpora cardiaca, but it was sug-

22 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

gested by Ellen Thomsen (1942) that they correspond with the peri-
cardial glands described by Pflugfelder in the phasmatids, and M.
Thomsen concurs in this view, which is now generally accepted. Fur-
thermore, there is reason to believe that both the pericardial glands
and the lateral cells of the ring gland represent the thoracic glands
of other insects. Though Poulson (1950) says the lateral ring gland
cells of Drosophila arise from the roof of the stomodaeum, and would
therefore appear to be the corpora cardiaca, M. Thomsen shows that
the usual four corpora-cardiaca nerves (ccNvs) from the brain go
to the group of small cells in the posterior angle of the ring gland,
which fact would suggest that these cells alone are of corpus-
cardiacum origin. Lying behind the ring gland of Calliphora, and
connected with its posterior end by a short nerve is the hypocerebral
ganglion (icGng) of the recurrent nerve (rNv). In Drosophila
Bodenstein (1950) shows that the corpus cardiacum and the hypo-
cerebral ganglion are apparently united in the posterior end of the
ring gland.

From experimental work it is known that the ring gland of cyclor-
rhaphous larvae is necessary for the inducement of moulting and
pupation. Burtt (1938) observed that removal of the gland from
larvae of Calliphora prevents pupation and that growth of the imagi-
nal buds is arrested. Day (1943) reports that experiments on Lucilia
and Sarcophaga suggest that the ring gland produces a hormone con-
cerned with normal development; in the larva it induces puparium
formation. Bodenstein (1944) showed that larval moulting is de-
pendent on the presence of a ring-gland hormone. Possompés (1950),
however, demonstrated that the action of the ring gland as an effector
of metamorphosis depends on its stimulation by a hormone from the
brain. He suggests that the ring-gland elements thus activated from
the brain are the lateral cells (“peritracheal glands”), which thus cor-
respond at least in function with the thoracic glands of other larvae.
There appears to be no experimental demonstration of the specific
function of the corpus-allatum element of the ring gland on the larva,
but presumably it is the same as in other insects.

The ring gland of the larva moves backward in the pupa and comes
to lie in front of the proventriculus. In the newly emerged adult of
Calliphora, according to Ellen Thomsen (1942), the ring gland is
present, but in the mature fly all of it except the corpus-allatum com-
ponent disappears. In the adult of Drosophila, Bodenstein (1950)
says the lateral ring cells degenerate, but the anterior group of cells
remains as the corpus allatum, and the cells of the posterior part form

NO. 9 INSECT METAMORPHOSIS—SNODGRASS 23

an elongate body representing the corpus cardiacum and the hypo-
cerebral ganglion. The metabolic changes that the ring gland pro-
duces in the adult fly are those that are ordinarily attributed to the
corpora allata. The action of the ring gland in the adult fly according
to Day (1943) is seen “first in changes which occur during the break-
down of the larval fat body cells and subsequently in the changes
undergone by the adult fat body cells, the oenocytes, and the develop-
ment of the ovaries.” Bodenstein (1950) attributes to the corpus-
allatum remnant of the ring gland in the adult female of Drosophila
the formation of a hormone that regulates egg maturation. The neces-
sity for the presence of the ring gland in the adult fly for the ripen-
ing of the eggs is well attested by the works of Ellen Thomsen (1940,
1942) and others.

The nature of hormonal acttion—The hormones concerned with
growth and metamorphosis are not in themselves the determiners of
development ; the course of development is determined by hereditary
factors inherent in the tissues of the animal. The hormones are mere
regulators, and in most cases they are found to be nonspecific as to
species, a hormone from one insect having the same effect when intro-
duced into another, regardless of different species structure. Further-
more, the effect of a hormone depends not entirely on the nature of
the hormone, but also on the receptive state of the affected tissue.

Most of the experimental work that has been done on the hormones
of insects has had as its object the ascertaining of the effect of spe-
cific hormones. It is now coming to be recognized, however, that the
various endocrine glands and their secretions interact upon one an-
other, and that the hormonal effect at any one time may depend on
the relative amount of a particular hormone or hormone complex
present in the blood. As stated by Bodenstein (1953a) the insect is
able to keep a hormonal balance by ‘“‘a mechanism of compensating
hypertrophy or atrophy of its glands.” The glands are in constant
interaction with one another so that the amount of any hormone in
relation to the others can be changed. “It is the hormone balance at a
given time that determines the specific activity of the humoral sys-
tem.” In further work on the endocrine glands of insects Bounhiol
(1953) says “it will be necessary to study more and more the action
of the glands on one another,” or, in the words of Bodenstein (1953a),
“to disentangle the complicated relationships existing between the
various hormones and to understand their action in physiological
terms, not forgetting the vital role played in all these responses by
the reacting systems.”

24 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

III. APTERYGOTA

The insects of this group, which are wingless at all stages, as pre-
sumably were their ancestors, go through no truly metamorphic
changes in their postembryonic growth. They might, therefore, be
omitted from a discussion of metamorphosis were it not for the fact
that they have certain structures that have been thought to recur in
the larval stages of some winged insects, and which thus give them
a theoretical value in the interpretation of juvenile characters among
the metabolous insects. Of the several groups of apterygote insects,
the Thysanura are the most closely related to the Pterygota. Though
thysanurans are known in paleontology only as far back as the Ter-
tiary, while winged insects were fully developed in the Carboniferous,
a thysanuran (fig. 2A) undoubtedly gives us a concrete example of
what the wingless ancestors of the winged insects were like.

The organs of the Thysanura that are of particular interest in
connection with a study of the larvae of the higher insects are the
abdominal styli and the associated eversible vesicles. As typically de-
veloped in the Machilidae, there may be a pair of styli on the venter
of each abdominal segment from the second to the ninth inclusive
(fig. 2A), and a pair of vesicles (E,Vs) on each of the first seven
segments, or two pairs on some of the intermediate segments. In
each segment the styli and vesicles are borne on lateral plates of the
venter (C,D,E,Cx) commonly regarded as the bases of otherwise
suppressed abdominal limbs. In the embryonic development of
Lepisma, Heymons (1897) has shown that rudiments of appendages
are formed on the first ten abdominal segments, but with the dorsal
growth of the body wall they are stretched transversely and become
flattened until finally they form merely the lateral parts of the defini-
tive abdominal sterna. Eversible vesicles are absent in the lepismatids,
and styli are present only on the eighth and ninth segments, or also
on the seventh segment. In the machilids the so-called coxal plates
(C,D,E,Cx) bearing the styli and vesicles remain separated from a
median sternal plate (S). That the styli are coxal appendages and
not limb vestiges is shown by their occurrence on the coxae of the
middle and hind legs (B,Sty). The abdominal styli, therefore, are
appurtenances of former limbs, but do not themselves represent ab-
dominal legs. The same evidently is true of the eversible vesicles.
Both styli and vesicles occur also among the other groups of ap-
terygote insects, and among the pterygotes styli are present on the
ninth abdominal segment of the adult male in the cockroaches, man-
tids, and termites. We may reasonably conclude, therefore, that the

NO. 9 INSECT METAMORPHOSIS—SNODGRASS 25

immediate ancestors of both the wingless and winged insects had
abdominal styli.

In Thysanura and Diplura the abdominal styli and vesicles are
provided individually with muscles that arise on the supporting plates
(fig. 2D,E). The styli are flexibly movable on their bases; the vesi-
cles are retracted by their muscles, and protracted probably by blood
pressure. The styli are developed during postembryonic growth; ac-

Pee ie
aa =a SN ie

Fic. 2—Structural details of Machilidae.

A, Machilis sp., whole insect, showing thoracic and abdominal styli. B, Neso-
machilis maoricus Tillyard, middle leg, showing stylus on coxa. C, same, ventral
surface of first abdominal segment, vesicles retracted. D, same, ventral surface
of second abdominal segment, with vesicles and styli. E, same, ventral surface
of sixth abdominal segment, vesicles everted.

Cx, coxa; Fm, femur; Pl, pleuron; rvs, retractor muscles of vesicle; S,
sternum; Sty, stylus; Tr, trochanter; V's, eversible vesicle.

cording to Heymons (1897, 1906), Adams (1933), Sweetman and
Whittemore (1937), and Lindsay (1939) they first appear on the
fourth or fifth instar of lepismatids, or even on much later instars.
Thysanurans moult many times throughout life, the number of
moults depending on how long the insect lives. Sweetman and Whitte-
more (1937) record as many as 42 observed moults for one individual
of Thermobia domestica, and they state that both moulting and growth
continue long after the first eggs are laid. The lifelong periodic moult-
ing of the Thysanura suggests that in this respect the primitive in-
sects resembled the other wingless arthropods. With the acquisition
of wings, moulting became too arduous, and among modern winged

26 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

insects a moult in the active imaginal stage occurs only in the
Ephemeroptera.

The few structural changes that the thysanurans go through dur-
ing their postembryonic life are merely those of development from
youth to maturity. Body scales do not appear until after the first
moult, the rings of the antennae and of the caudal filaments increase
in number, the abdominal styli are formed at various moults, the leg
styli of Machilis are said by Heymons (1906) to be absent on the
first instar, there are some changes in the shape and proportions of
the parts of the body, and the external genitalia develop during late
stages. Such changes, however, do not constitute a true metamor-
phosis; they are progressive toward the adult structure, and do not
give rise to adaptive juvenile characters.

The fossil records of early insects give no evidence as to how in-
sects acquired their wings. There is no doubt that insects were hexa-
pods before wings were developed, and it seems highly probable that
wings were evolved from paranotal lobes on the thoracic segments
that first served as gliders.

1V. PLECOPTERA

Among all the “orthopteroid” insects the stoneflies are the only
ones of which the young have adopted a medium different from that
of the adults, and, though the young stoneflies live in the water, their
structural adaptation to aquatic life is relatively little. Aside from
features of immaturity, such as the unfinished development of the
wings, there is little to distinguish a young stonefly from an adult
other than the presence of gills for aquatic respiration, and differences
in the shape and proportions of the parts of the body. The stonefly
larvae have no outstanding features common to all species by which
they differ from the adults, and they could hardly be mistaken for
anything other than immature Plecoptera.

A typical stonefly larva has well-developed compound eyes and
frontal ocelli; the antennae are long, slender, and multiarticulate ; at
the end of the body are two caudal filaments representing the orthop-
teroid cerci, but no median filament; there are three subsegments in
the tarsi, and two pretarsal claws; the fully exposed wing pads un-
dergo a gradual development. In all these characters except those of
immaturity the stonefly larva is essentially like the adult, and is en-
tirely comparable to an orthopteran nymph; in short, it is simply a
nymph that has taken to the water, where most species have acquired
gills of a simple kind. If there is a difference in the mouth parts or

NO. 9 INSECT METAMORPHOSIS—SNODGRASS 27

in the length of the caudal filaments between the larva and the imago,
the difference is usually due to a reduction of these structures in the
latter. The life cycle of the larva varies, according to the species,
from one to three or four years, and there are correspondingly many
instars, as many as 33 being recorded by Schoenemund (1912) for
Perla cephalotes.

The gills of the stonefly larva are mostly tufts of delicate filaments
penetrated by tracheae; they are generally present on the sides or
sternal region of the thorax, but sometimes on the abdomen, particu-
larly at the posterior end around the anus. Gills of a different type,
however, may occur on the bases of the legs. As described by Lauter-
born (1903), these leg gills in Taeniopteryx nebulosa L. are soft, “3-
segmented,” tapering processes arising singly from the mesal ends of
the coxae, and are retractile by muscles. When retracted the three
“segments” are telescoped into each other until only a soft papilla
remains visible externally. Lauterborn compares these gills with the
coxal sacs of Diplopoda; they might be likened to the eversible vesi-
cles of Thysanura, but their position on the mesal ends of the coxae
precludes a comparison with styli. Similar tapering gill processes are
present on the sides of the first six segments of the abdomen in the
genus Eusthenia, as illustrated by Tillyard (1926) in E. spectabilis
Wwd. Though these abdominal gills are suggestive of styli, it seems
probable that all the gills of stonefly larvae are special developments
and have no relation to any other structures, including the gills of
mayfly larvae. Besides the gills there may be a subepidermal system
of tracheoles serving for respiration direct through the body wall.
Wu (1923) has described in the larva of Nemoura the presence of
numerous tufts of tracheoles on the epidermis of the submentum, the
coxae, the ventral sides of the femora, and on the first eight sterna of
the abdomen. A group of long tubular processes arising in the an-
terior end of the rectum he regards as “blood gills” because they do
not contain tracheae.

The stonefly larva generally retains the feeding habits of the adults ;
most species feed on vegetable matter (see Claassen, 1931), only
members of the family Perlidae being carnivorous. The mouth parts
are modified according to the nature of the food, and there may be
differences also in the general form of the body between vegetarian
and carnivorous species. When the larva is ready to transform into
the adult it crawls out of the water onto a stone or log, and may go
some distance from the shore, showing that it has not entirely lost
the ability to comport itself on land. The adult stonefly does not de-
pend on the larva for stored nourishment to the extent that do insects

28 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

with more specialized larvae. Though the adult mouth parts are more
or less reduced in some species (Lucy W. Smith, 1913), in others
they are well developed, and such species feed extensively in the adult
stage on vegetation. The female stonefly goes back to the water to
discharge her eggs.

In the Plecoptera we have an example of metamorphosis in its
simplest form, and one that shows very clearly that insect metamor-
phosis can have its inception in the adaptation of the juvenile stage
to a medium different from that inhabited by the adults. The higher
degrees of metamorphosis, therefore, arose from more extensive
structural modifications of the young in adaptation to a secondarily
adopted medium or way of living. Probably the nymph of the primi-
tive stonefly simply found that it could obtain a better living in the
water than on land, and natural selection then eventually furnished
it with gills for a permanent aquatic existence.

V. EPHEMEROPTERA

The young mayfly (fig. 3 A) is distinctly more specialized in its
adaptation to life in the water than is the young stonefly. Still, the
young mayfly is simply a juvenile insect of generalized structure; it
has compound eyes and frontal ocelli, well-developed legs, mouth
parts of the biting type, and during its growth it develops wing pads
that increase in size up to the last moult. In these characters the
young mayfly has the developmental status of an orthopteroid nymph,
and that it was primarily a land-inhabiting nymph may be deduced
from the presence of an elaborate tracheal system in both the adult
and the larva. Since the young of the earliest known fossil mayflies,
found in the Permian, already had gills, the mayfly larva has come
down to us with surprisingly few changes.

The larval gills of the mayflies are organs of particular interest be-
cause of their apparent likeness to the abdominal styli of Thysanura.
In modern species the gills are present on the sides of, at most, the
first seven segments of the abdomen; larvae from the lower Permian,
however, had nine pairs of gills, and some Jurassic species had eight.
The gills are highly variable in form in different species, but they are
borne singly on lateral lobes of the abdominal segments (fig. 3 B,C)
interpolated between the tergal and sternal regions. The gill-bearing
lobes fall directly in line with the bases of the thoracic legs (Cx),
and thus may be likened to the stylus-bearing plates of Machilis.
Moreover each gill is movable by muscles arising in the supporting
lobe (C,D). The movements of the mayfly gills has been made the

NO. 9 INSECT METAMORPHOSIS—SNODGRASS 29

Fic. 3.—Characters of larvae of Ephemeroptera.

A, Ephemerella sp. B, Ephoron sp., part of thorax and abdomen, showing
gill-bearing lobes in line with coxae of legs. C, diagrammatic cross section of
abdomen. D, a single gill, showing muscles arising in supporting body lobe. E,
base of gill, with tracheal trunk and muscles. F, Ephemerella sp., showing ad-
hesive disc on venter of abdomen. G, Prosopistoma foliaceus Fourcroy, dorsal
(from Vayssiére, 1890). H, same, ventral (from Vayssiére, 1890).

1b-4b, bmcls, branchial muscles; Brn, branchia, gill; Cx, coxa of leg; Cxr?,
gill-bearing lobe of abdomen; en, respiratory entrance; ex, respiratory exit; 1,
lateral body muscle; S, sternum; T, tergum; Tra, trachea.

30 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL, 122

subject of a special study by Eastham (1938, 1939). Just as the
thysanuran styli do not appear until after the first moult, so the gills
of the mayfly larva are absent in the first instar. It is said by Ide
(1935) that all the gills appear with the first moult in some species,
but that in other species most of them may be delayed until several
moults later.

Gills of the simplest form are slender processes penetrated by
tracheae, others are fringed with long filaments, some are lamelliform,
and most of them are branched. According to Ide (1935) all the
gills are at first uniramous, and some that eventually become lamel-
liform grow out first in the form of filaments. It would appear to be
true, therefore, as Spieth (1933) says, that the primitive gills of the
ancestral mayflies were simple slender tubular structures, into which
the tracheae enter, and that the compound gills of the present-day
forms have arisen as modifications of the primitive type. If the mod-
ern gills do represent styli, we may suppose that the young mayfly in
its primary terrestrial life may have had abdominal styli similar to
those of the Thysanura and Diplura, which, when it took to the water,
were readily converted into gills. That the mayfly gills have been de-
rived from styli, however, is merely a theoretical concept, but con-
sidering that the Ephemeroptera are relatively primitive insects the
concept is sufficiently reasonable to be accepted as not too improbable.
Unlike styli, however, the mayfly gills are discarded at the moult to
the subimago.

Some remarkable larval modifications occur in connection with the
gills. In the genus Baetisca, described by Vayssiére (1934), the
mesonotum is extended posteriorly to the middle of the sixth abdomi-
nal segment to form a carapace covering the gills and the meta-
thoracic wing pads, the pads of the first wings being fused with its
under surface. In Prosopistoma (Vayssiére, 1882, 1890) a carapace
is even more extensively developed (fig. 3 G) and covers a respira-
tory chamber enclosing the gills, which is shut in ventrally (H) by
the pleural regions of the thorax and lateral extensions of the first
five abdominal sterna. The Prosopistoma larva thus resembles a small
crustacean in appearance. Water has entrance to the respiratory cham-
ber by way of lateral openings (H,en) between the carapace and the
sternum, and is discharged through a median dorsal aperture (G,e7)
in the notch of the posterior end of the carapace. A preliminary stage
in the development of a carapace is suggested in the larva of Epheme-
rella (A) in which the mesonotum including the fore wing pads is
extended posteriorly over the base of the abdomen and completely
covers the hind wing pads.

NO. 9 INSECT METAMORPHOSIS—SNODGRASS 31

In various lesser ways the mayfly larva may be characterized by
special juvenile structures. In some forms the incisor processes of
the mandibles are produced into a pair of long tusks. The larvae of
Ephemerella that live in swift currents have an adhesive disc on the
under side of the abdomen (F) formed of a dense fringe of soft
marginal hairs. In the anterior part of the disc is a deep transverse
cavity behind a strong semicircular lip, which possibly has something
to do with creating a suction when the disc is applied against the
surface of a rock.

It is noted by Ide (1935) that at each moult of the mayfly larva
there is some structural change adapting the larva to environmental
changes resulting from the growth of the larva. Such changes involve
the mouth parts, the wing pads, external genitalia, the claws of the
legs, and the caudal filaments. The larva moults many times before
changing to the winged imago ; observations by Ide show that Ephem-
era simulans goes through about 30 larval moults, and Stenonema
canadense as many as 40 to 45 moults. The large number of moults
Ide attributes to the necessity for making adjustive physical changes
to the environment, rather than to growth, since the larva increases
but slowly in size. Some of these adaptive changes of the larva might
be regarded as a feeble hypermetamorphosis, but the lack of gills and
a tracheal system in the first instar and the expansion of newly formed
gills into lamellar gills, cited by Joly (1872) as examples of hyper-
metamorphosis, are simply developmental changes.

The structural adaptation of an animal to a special environment is
much easier to see as a fact, than it is to explain how it came about.
The young mayfly larva can breathe through its skin, but as it gets
larger it needs gills; the first one that entered the water, therefore,
must have suffocated if it persisted in keeping submerged. However,
if it possessed tracheated styli, it was but a simple evolutionary proc-
ess to convert these organs into gills. Adaptation can seldom be one-
sided ; in the case of an aquatic larva of a terrestrial or aerial adult,
the adult must be adapted to the way the larva lives. The female may-
fly, therefore, has an instinct for returning to the water to discharge
her eggs.

The changes that the mayfly larva undergoes in its metamorphosis
to the adult are not due entirely to the special characters of the larva.
The adult mayfly lives so short a time that it needs no food, and as
a matter of economy its mouth parts are reduced to a functionless,
condition. Murphy (1922) says the “atrophy of the mouth parts is
progressive during the aerial life of an individual,” but “varies in ex-
tent among members of species.” The ingestion apparatus and the

32 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

alimentary tract, however, are fully preserved, but for the purpose
of swallowing and retaining air. The stomach is shown by Pickles
(1931) and by Grandi (1950) to be transformed in the adult into a
thin-walled air sack. The air probably serves to make the body more
buoyant and by compression to expel the eggs.

Most mayflies undergo a moult after they have attained the state
of a winged imago, the adult stage being thus subdivided into two
winged stages, distinguished as the subimago and the imago. Con-
cerning the subimago of Cloeon dipterum, La Baume (1909) says
that it usually issues from the larval skin toward evening either on
the surface of the water or on the shore. The quickness of the change
is most noticeable, particularly the almost instantaneous spreading of
the wings. The insect now flies to vegetation along the shore, where
it remains quiet until the next moult, which, according to the species,
may occur in a few minutes, a few hours, or several days. There is
probably no specific reason why the adult mayfly should moult again ;
it is the only winged insect known to moult in the active adult stage,
and even some mayfly species omit a second moult. Evidently the
imaginal moult is simply a holdover by a primitive insect from wing-
less ancestors that shed the cuticle periodically throughout life as do
the Thysanura and most other wingless arthropods. Extraction of
the wings from the old cuticle is a difficult matter and other insects
have simply discarded a useless and dangerous habit.

It is clear that the mayfly undergoes a greater degree of metamor-
phosis than does the stonefly because the young mayfly is more exten-
sively modified in adaptation to life in the water. Inasmuch as the
larva in the two cases is differently modified for the same purpose,
metamorphosis has arisen independently in the two groups.

VI. ODONATA

The Odonata present an example of metamorphosis much more
accentuated than that of either the Plecoptera or the Ephemeroptera,
and there is no relation between the special characters of the odonate
larva and those of the other two groups, again showing that larval
structures in adaptation to aquatic life have been independently de-
veloped in these three orders. In common with other aquatic larvae,
the odonate larva has been adapted in its body form and its means
of respiration to life in the water, but in addition it has evolved a
very special modification of the labium by which this organ is greatly
enlarged and converted into an efficient device for the capture of
active prey.

NO. 9 INSECT METAMORPHOSIS—SNODGRASS 33

The Odonata are predaceous both as larvae and as adults; their
mouth parts are of the biting type of structure. The adults entrap
their insect prey on the wing by means of their hairy legs, and their
mouth parts are not unusually modified. The short body of the adult

Fic. 4.—Odonata; development and metamorphosis of the labium.

A, Anax junius Drury, labium of 17-day embryo (from Butler, 1904). B,
same, 20-day embryo (from Butler, 1904). C, Sympetrum striolatum (Charp.),
labium of pronymph containing labium of second instar (from Corbet, 1951). D,
same, free labium of second instar, expanding (from Corbet, 1951). E, same,
fully expanded labium of second instar (from Corbet, 1951). F, Anax sp.,
labium of mature larva, posterior. G, same, larval labium and early stage of
formation of imaginal labium in the prementum. H, same, later stage, the imag-
inal labium retracted into postmentum of larva and taking on the adult struc-
ture, posterior. I, same, imaginal labium from H unrolled and spread out.

iLb, imaginal labium; Lig, ligula; Plp, labial palpus; Plpg, palpiger; Pmt,
postmentum; Prmt, prementum.

labium (fig. 41) consists of a distinct postmentum and a prementum ;
the prementum bears a large median ligular lobe, and two small lateral
lobes (Plpg) that support the short, thick palpi (Plp). The larval
labium is more simple in form than that of the adult, but both the
postmentum and the prementum are greatly elongated, and are articu-
lated on each other by a freely movable elbow. The larval postmen-

34 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. I22

tum is unusual in that, instead of being as in most insects a plate on
the under side of the head, it is produced into a long, free stalk sup-
porting the prementum on its distal end. The prementum is highly
variable in form in different genera; in the common Anax junius (F)
it is a long flat lobe somewhat expanded distally where it bears the
relatively small palpi, each of which is armed with a long sharp claw.
In the passive position of the labium the postmentum is turned pos-
teriorly against the mesosternum of the thorax (fig. 5 A) ; the pre-
mentum in some species is pressed against the under surface of the
head (A), in others it is applied like a mask over the lower part of
the face (C). In action the postmentum swings downward and for-
ward on the head, the prementum is lowered (B), and the entire
labium is then projected far beyond the mandibles to seize a prospec-
tive victim. Associated with the larval labium is a long T-shaped
apodeme developed from the base of the hypopharynx that extends
posteriorly through the head, and the crossbar is embedded in the
posterior edge of the base of the postmentum. The labial muscula-
ture is surprisingly simple, but it is probable that blood pressure from
the abdomen plays an important part in the projection of the labium.
While undoubtedly the larval labium is specialized by comparison with
the adult labium, the labium of the embryo develops directly into that
of the larva, and at metamorphosis the adult labium develops within
the larval organ. The hypopharyngeal apodeme is either greatly short-
ened in the adult or reduced to a ligamentous band.

The embryonic labium of Anawx junius (fig. 4 A,B), as illustrated
by Butler (1904), has a primitive feature in the almost complete
separation of the stipital lobes of the prementum (Prmt) ; the unseg-
mented palpi (Plp) bear fingerlike processes (A) that will become
the apical hooks (B). In the pronymph of Sympetrum (C), accord-
ing to Corbet (1951), the prementum is undivided and the palpi arise
close together from its distal end, but during ecdysis of the second
instar (D) the prementum stretches transversely, and later (E) be-
comes more elongate. The embryonic labium thus goes from a primi-
tive labial structure directly into the specialized structure of the larval
labium. The labium of the adult as described by Munscheid (1933)
is first formed in the distal part of the larval labium about five days
after the larva ceases to feed. At first it takes on approximately the
form of the larval labium, but later it becomes shorter until four days
after its formation it occupies only the basal two-thirds of the larval
postmentum. A further three days now elapses before ecdysis of
the imago.

NO. 9 INSECT METAMORPHOSIS—SNODGRASS 35

At an early stage of the labial transformation in Anax junius the
imaginal labium may be seen retracted into the anterior part of the
larval prementum (fig. 4G, 7Lb). The principal changes that have
taken place affect the palpi and the ligula, which have become elon-
gated. On the palpus, the movable claw of the larval organ is re-
placed by a short setigerous lobe, as in the adult (1), and the fixed
finger has become a slender tapering median process. At the base of
each palpus a palpigerous plate is differentiated. In a later stage (11)
the imaginal labium has withdrawn into the postmentum of the larval
labium, where it is much compressed and its lateral parts are rolled
anteriorly. When the imaginal labium at this stage is removed from
the larval labium, unrolled and spread out (1), it is seen to have ap-
proximately the form of the adult labium except for the triangular
shape of the ligula and its deeper apical notch. The palpi have taken
on the form and size of the adult palpi, the prementum and postmen-
tum are distinct in the body of the labium, and the palpigers are well
defined. The triangular ligula finally becomes transversely oval.

It is of interest to note that the odonate labium begins its develop-
ment in the embryo as a labium of primitive structure (fig. 4 A,B).
In its later growth it develops directly into the specialized labium
of the larva; then finally the more generalized labium of the adult is
derived from the larval labium. It is not clear what phylogenetic de-
ductions may be made from these facts, but it seems reasonable to
suppose that the larval labium in the first place must have been evolved
from a generalized labium approximately of the adult type of struc-
ture ; if so, it carries the potentiality of reversal.

The transformation period from larva to imago is said by Mun-
scheid (1933) in Aeschna cyanea to occupy about 12 days. During
this time the structural changes of the labium are accompanied by
a total histolysis of the larval labial muscles, followed by regenera-
tion of the imaginal muscles and the formation of new tonofibrillar
muscle attachments on the imaginal cuticle. Two pairs of larval
muscles are destroyed and not replaced. The processes of muscle
histolysis and histogenesis are described in detail by Munscheid, who
points out that the transformation of the odonate labium and the re-
generation of its muscles is comparable to the pupal metamorphosis
in holometabolous insects, except that in the Odonata the process is
limited to a single organ instead of affecting the entire insect, which
otherwise is hemimetabolous. The long quiescent transformation
period apparently allows the regenerated muscles to become attached
directly on the new imaginal cuticle without the interpolation of a
second moult.

36 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. I22

Aside from the specialization of the labium, the principal adaptive
characters of the odonate larva are the organs that serve for respira-
tion. In the Anisoptera a spacious rectal sac contains six longitudinal
tracheated folds of the walls which are the larval gills. The muscular
apparatus of the rectum for the inhalation and exhalation of water be-
comes also a means of locomotion by the forcible ejection of spurts

\
\

> ; PRAIA aN KR
Papt Cer

Fic. 5—Odonata; general features of larvae.

A, Larva of Anax sp., labium in passive position. B, same, labium lowered
and partly protracted. C, Crocothemis servilis Drury, labium applied against the
face. D, anisopterous larva, posterior segments and lobes enclosing the anus. E,
Agrion virgo L., posterior end of body with gill lobes removed. F, Archilestes
grandis (Rambur), end segments of body and gill lobes. G, Agrion virgo L., end
segments and apical lobes.

An, anus; Cer, cercus (cercoid); dl, dorsal gill lobe; Eppt, epiproct; Lb,
labium; /l, lateral gill lobe; Md, mandible; Papt, paraproct; sa, supra-anal lobe.

of water. Zygopterous larvae are provided with three external gill
lobes of various forms at the end of the body, one median and dorsal,
the other two lateral, borne on basal plates surrounding the anus.
Typically these caudal gills are thin lamellae (fig. 5 F), but they may
be sacciform, and in some species they are slender horny blades (G)
that do not appear to be suitable for respiratory purposes. The gills
are weakly attached to the supporting plates so that they are easily
broken off, but they regenerate at the next moult.

The current interpretation of the zygopterous larval gills, taken

NO. 9 INSECT METAMORPHOSIS—SNODGRASS a7

from Heymons (1904), is that the dorsal gill represents a median
dorsal filament and that the lateral gills are the cerci. Arising on each
side between the bases of the gill-supporting plates is a small cercus-
like process (fig. 5 F,Cer), the ‘‘cercoid” of Heymons, who says it is
developed during larval life. The gill-bearing plates (Eppt,Papt)
surround the anus (E, An) in a manner so exactly comparable to the
epiproct and paraprocts of an orthopteroid insect that their identity
as such is hardly to be questioned, and the “cercoids”’ (Cer) have the
usual relation of cerci to these plates. The lateral gill lobes (E, Jl, re-
moved at their bases) therefore appear to be mere outgrowths from
the paraproctial plates, and as such they could hardly be cerci. In the
anisopterous larva (D) the gill-bearing plates of the zygopterous larva
are produced into long valvelike lobes enclosing the anus, and there
is no apparent reason for not identifying these lobes (Eppt,Papt)
with the usual epiproct and paraprocts in the same position. The
gills are cast off at the transformation to the adult, except as said by
Tillyard (1917) that the lateral gills (“cerci”) of the male leave a
pair of small processes developed within their bases. If the lateral
gills are cerci, it is an unusual thing for an insect to lose these organs.

In some zygopterous larvae, in addition to the caudal gills, there
are paired lateral gills in the form of tracheated filaments along the
sides of the abdomen (see Calvert, 1911; Needham, 1911; Tillyard,
1917). The tracheal system of the Odonata is present in the newly
hatched larva, but according to Calvert (1898) the tracheae do not
fill with air until the first moult. Spiracles are present in the larva
but ordinarily are not functional except for the withdrawal of the
tracheal linings at ecdysis. The early development of the tracheal
system and the presence of spiracles in the larva, Calvert points out,
attest that the immediate ancestors of the Odonata were air-breathing
insects.

The structural changes that take place during larval life of the
Odonata have been summarized by Tillyard (1917) under nine
headings. Such changes, however, as the growth of the compound
eyes, development of the ocelli, increase in the number of antennal
joints and of subsegments in the tarsi, changes in the shape of the
thorax correlated with development of the wings, progressive changes
in the nervous system, and increase in the number of Malpighian
tubules are merely stages in the postembryonic development of the
adult organs. These are not true metamorphic changes such as those
producing the general form of the larval body, the modification of
the labium, and the development and differentiation of the rectal and

38 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

caudal gills. In addition to these changes, however, there takes place
during the transformation period a radical change in the sclerotiza-
tion pattern of the abdominal segments, accompanied by an almost
total destruction of the larval abdominal musculature and the forma-
tion of a much more simple musculature for the adult. The Odonata
might almost be said to be holometabolous insects without a pupal
stage.

As in the case of other insects having aquatic larvae, the adult fe-
male of the Odonata has one instinct of responsibility to her offspring,
namely, that which impels her to go back to the water to deposit her
eggs. Some are so conscientious in this respect that they even enter
the water and insert their eggs in the stems of submerged water plants.

VII. HEMIPTERA

The Hemiptera differ from most other insects having specialized
mouth parts in the mature stage in that the adult type of mouth parts
is just as practical for the young as for the imago. The adult hemip-
teron has not evolved feeding organs useful only to an insect with
functional wings. The piercing and sucking mouth parts in Hemip-
tera, therefore, are developed in the embryo and are functional as
such in the newly hatched insect. The same is true of the Thysanop-
tera and Anoplura. If there are metamorphic changes between the
young and the adults of these insects, they do not affect the essential
nature of the feeding organs, and all instars of a species can live and
feed together in the same habitat.

Among the Heteroptera postembryonic development is principally
a succession of growth stages from the young to the adult; the
Heteroptera, as the Orthoptera, are essentially ametabolous. Though
the change between instars may be accentuated at the last moult, there
is in general little, if any, structural deviation on the part of the
young insect that must be suppressed in the imago. However, a defi-
nite case of juvenile aberration in the Heteroptera is to be seen in a
species of mirid described by China (1931) in which the nymph (fig.
6F) is armed on the head, thorax, and abdomen with large dorsal
prongs. Though the adult of the species has not been certainly identi-
fied, no adult mirid is known to possess any such armature.

Among the Homoptera there is a distinct though sporadic tendency
for the young insect to develop special characters of its own that are
not carried over into the adult stage, or to take on a form quite
different from that of its parents. The aberration of the young insect
may even become so pronounced that the final transformation to the
imago approaches or actually attains a condition of holometabolism.

NO. 9 INSECT METAMORPHOSIS—SNODGRASS 39

A good example of simple metamorphosis in the Homoptera is
seen in the structural adaptation of the young cicada to a subterranean
life by the modification of its front legs for digging (fig. 6D). The
nymphal structure of the leg is not present in the embryo (A); it
appears first on the nymph with the shedding of the embryonic cuticle
just after hatching (B) and becomes more fully developed in succes-

Fic. 6—Examples of simple juvenile metamorphic characters in Hemiptera.

A, Magicada septemdecim (L.), newly hatched nymph still in embryonic cuti-
cle. B, same, left front leg of first instar (from Marlatt, 1923). C, same, front
leg of third instar, mesal vieiw showing reduced tarsus (Tar). D, same, front leg
of mature nymph, lateral. E, same, front leg of adult. F, Paracarnus myersi
China, nymph, Heteroptera-Miridae (from China, 1931).

sive instars (C,D). The tarsus of the first instar (B) is reduced in
later stages to a small spur on the inner surface of the tibia (C,Tar),
but it is fully restored in the mature nymph (D). At the transforma-
tion to the adult, the special features of the nymphal leg are much
reduced or obliterated (E). The newly hatched cicada has a pair of
small eye spots, but in subsequent instars the eyes are lost, and func-
tional compound eyes are redeveloped only in the imago. Within a
chamber just below the surface of the ground, or built up above the
surface, the 17-year cicada at last goes through a period of recon-
struction inside the nymphal cuticle, during which the adult structure
of the insect is developed, including the compound eyes, the external
genital organs, and the sound-producing organ of the male. When

40 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

the insect emerges from its transformation chamber it is an active
adult, but it still wears the nymphal skin until it arrives at a suitable
place for ecdysis.

As another example of simple specialization in a young homopteron
we may cite the respiratory canal of the spittle bugs, Cercopidae. On
the ventral surface of the tapering posterior segments of the nymph
is a deep groove that expands anteriorly into a wide space covered
on the sides by the extended abdominal terga, which protect the
spiracles. This modification is a respiratory device necessary only to
the nymph and is discarded at the moult to the imago. As further
examples of juvenile aberration we might note the presence of large
branched spines on the back of certain membracid nymphs, and of
various minor nymphal characters in other homopterous families that
are not retained by the adult. Juvenile specialization among the
Homoptera, however, is carried progressively further in the Psyllidae,
Aleyrodidae, and Coccidae, until in the last family the transformation
to the imago attains the status of true holometabolism. In the Aley-
rodidae and the Coccidae the young insects are so different from their
parents that, following the definitions given in the introduction, we
must call them larvae, but admittedly they are nymphs that have
acquired the status of larvae by definition.

The Psyllidae go through five juvenile instars, which, except for
the flattened form of the body, in general resemble the nymphal stages
of ametabolous Hemiptera. The wings appear first in the third instar
and increase in size during the fourth and fifth instars ; the legs, how-
ever, undergo a metamorphosis, which has been fully described by
Weber (1930, 1931) in Psylla mali. The first instar is active because
the young psyllid newly hatched on the twig of an apple tree must
find an opening bud on which to feed; the legs are relatively far apart
on the under side of the body, and in their movements are fitted for
walking. After the first moult the insect becomes sessile, the legs come
closer together at their bases and are flexed transversely beneath the
thorax in order now to function as clasping organs. From the begin-
ning, however, the segmentation of the legs has been reduced by a
suppression of the femoro-trochanteral and the tibio-tarsal joints. At
the moult to the fifth instar the young insect takes on something of
the form of the adult, the body becomes deeper, the antennae longer,
and in the legs there appears a slightly marked division between the
tibia and the tarsus and an indication of two tarsal subsegments.
Finally, within the cuticle of the legs of the last juvenile instar the
imaginal legs are developed, the trochanter being now separated from
the base of the femur, the tarsus distinct from the tibia, and two

NO. 9 INSECT METAMORPHOSIS—SNODGRASS 4I

well-defined tarsomeres present. The first two pairs of clasping legs
of the young become normal walking legs in the adult, but the hind
legs are elongated and transformed into jumping organs by an en-
largement of the coxae and a lengthening of the body muscles of the
trochanters associated with a dorsal extension of the sternal apodeme
on which they are attached. The general alteration of the body form
at the last moult, Weber shows, involves changes and enlargements
of muscles in the thorax that are destined to be motors of the wings.
The mouth bristles are retracted into a crumenal pocket instead of
being looped outside the head as in the immature stages.

While the degree of metamorphosis in the psyllids is thus not
large, it is enough to show how a young nymphlike insect can be
specifically modified in adaptation to its needs, even in a different way
in successive instars. The transformation of the young psyllid into
the adult, however, is complicated by the development of special
imaginal characters along with the suppression of juvenile characters.

In the Aleyrodidae there is a juvenile metamorphosis somewhat
similar to that of the psyllids because here also the first instar is ac-
tive and the others are sessile. The young aleyrodid, however, is much
flattened, the body being of a simple, oval, scalelike form and wingless
in all immature instars; the spiracles are on the under surface, and
a wide fringe of wax filaments forms a marginal palisade that en-
closes an air space beneath the body. There are four immature stages,
the characteristics of which are described by Weber (1931, 1934) as
follows. In the active first stage the antennae and the slender, taper-
ing legs are relatively long; each leg has only three segments and
bears a stalked apical adhesive disc, representing the unguitractor
plate of the adult insect. In the second instar the antennae are much
shortened, and the legs are reduced to small, unsegmented stumps
useless for locomotion but retaining the adhesive discs. The same
leg structure is carried over into the third instar, but in the fourth
instar both the legs and the antennae become again larger, and the
legs are now 2-segmented.

From the fourth instar the adult aleyrodid is produced directly, but
by an unusual transformation process. As described by Weber (1931,
1934) in Trialeyrodes vaporrariorum, the body of the young insect
in the fourth instar becomes deeper than that of the preceding instars,
and the marginal wax palisade stands vertically below the edges. In
the early transformation stage the long, slender legs of the imago
grow beneath the larval cuticle, but for want of space they become
much folded and looped. Above the bases of the legs deep infoldings
of the body wall of the imago form large cavities, which separate the

42 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

median part of the body of the imago from the wide lateral exten-
sions of the larval body. From the median walls of these cavities the
wings are formed as outgrowths that finally extend back into the ab-
dominal region. The lumina of the lateral body lobes of the larva are
filled with fat cells, and at first are narrowly continuous above the
wing cavities with the haemocoele of the central part of the body, but,
as the wings push out, the lateral lobes become disconnected from
the central body, and at ecdysis are shed with the larval cuticle. A
cavity is similarly formed anteriorly that cuts off the precephalic mar-
gin of the nymph, while a third cavity at the posterior end of the
body provides for the growth of the external genitalia. At ecdysis,
therefore, all the superfluous marginal parts of the larval body are
cast off, and the imago is formed from the central part only.

The metamorphic characters of the young aleyrodid are thus seen
to include a flattening and simplification of the body and a suppres-
sion of the wings, together with modifications of the legs adaptive
first to active and then to sessile habits. The characteristic feature
of the final metamorphosis, however, is in the manner of transforma-
tion to the adult involving the discarding of parts of the larval body.
The aleyrodid metamorphosis has been termed allometabolism (from
allo, different), but the development of the wings beneath the cuticle
of the last larval instar is entirely comparable to the simplest form
of wing development in typical endopterygote insects; the term En-
dopterygota taken literally, therefore, would include the Aleyrodidae.

It is in the Coccidae that metamorphosis among the Homoptera
reaches its highest degree of complexity. The young scale insect is a
larva adapted to a parasitic life on plants, and in its external aspect
it is quite different from the adult. The true form of an adult coccid,
however, is known only from the winged male, since the female be-
comes sexually mature in a late larvalike stage and undergoes no
further transformation.

On hatching from the eggs the simple, flattened first-instar coccid
larvae are provided with eyes, antennae, mouth parts, and legs. They
are active crawlers whose function it is to disperse themselves over
the food plant. When the young larvae have settled down at a suitable
feeding place, they moult and enter a second larval stage in which
the legs in many species are reduced, or lost altogether, though in
some forms the legs are fully retained. With typical species (diaspine
scales) there are only two larval instars, but in some there are three
or more, and generally during the larval period there is only a slight
difference between the males and the females. At the last larval moult,
however, the sexes are differentiated. The female looks like only

NO. 9 INSECT METAMORPHOSIS—SNODGRASS 43

another and larger larval instar, since she has no vestiges of wings;
in some species the legs and antennae are retained, but in many the
legs are much reduced or suppressed. The female usually preserves
her mouth parts and alimentary canal, though the external feeding
organs may disappear. The ovaries, however, become functional and
soon the body of the female is converted into a bag of eggs. In the
reproductive stage the female scale insect thus appears to be a sexu-
ally precocious larva, but some coccidologists contend that she has at-
tained a larval form secondarily by a process of reduction or degen-
eration from a winged adult. Perhaps the only way to settle the ques-
tion would be to give the female a dose of the proper hormone and
see what happens to her.

The male coccid, after the last larval moult, goes through usually
two immobile transformation stages, and then becomes in most cases
a winged insect. In the first transformation stage, know as a propupa,
the male of winged species begins to take on the form of the adult;
the antennae, legs, and wings appear, and the eyes are fully developed,
but the mouth parts are reduced or suppressed. In the next stage,
termed the pupa, the insect assumes more closely the form of the
winged imago, the antennae and the legs increase in length, taking on
the character of the adult appendages, and the wings lengthen. In
the male of Lepidosaphes ulmi, according to Suter (1932), there is
only one pupal stage, during which the wings and legs appear and
increase in size until the moult to the adult. The adult male usually
has a pair of well-developed wings, but is devoid of feeding organs.
In some species, however, the male does not attain the typical winged
structure; the wings may be absent, the antennae and the legs much
reduced in length, while the body retains the larval form with no
constriction between head, thorax, and abdomen. The redevelopment
of the antennae and legs of the male scale has been shown by Berlese
(1896) in the Diaspinae (/ytilaspis) to take place in the early pupa
by evagination of the appendage rudiments from pouches of the
integument beneath the cuticle of the propupa.

In addition to its external transformations the male coccid under-
goes a very considerable degree of internal metamorphosis, which has
been described particularly in Pseudococcus by Makel (1942). Along
with the casting off of the mouth parts there is a great reduction of
the alimentary canal, which retains its form in the pupa, but in the
imago the mesenteron is reduced to a mass of cells without a lumen.
The oesophagus remains as a slender tube, the proctodaeum is nar-
rowed, though the rectum keeps its original dimensions, and the

44 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

Malpighian tubules increase in size. These changes are mostly retro-
gressive from the larval condition. On the other hand, the reproduc-
tive organs develop gradually to the definitive functional state, and
there is a thoroughgoing reconstruction of the larval musculature into
that of the adult.

In her account of the muscle transformation in the male of Pseudo-
coccus Makel distinguishes five different groups of muscles, as fol-
lows: (1) Larval muscles that go over with little or no change into
the imago; (2) larval muscles that undergo such changes as splitting,
uniting, or a change of position; (3) larval muscles destroyed by
histolysis and not regenerated; (4) transformation muscles formed
by addition of imaginal elements to larval muscles; (5) muscles of
the imago that arise as new muscles in the propupa. To this last group
belong four muscles of the thorax, and seven oblique intersegmental
muscles of the abdomen, together with two muscles connected with
the external genital organs. The metamorphosis of the muscular sys-
tem as given by Makel is based on a detailed comparative study of
the musculature in the larva, pupa, and adult.

It is clear that the transformation of the male coccid is a true holo-
metabolous metamorphosis, and that the larva is a specialized juvenile
stage. It may be questioned, however, that the coccid pupa is com-
parable to the pupa of the higher holometabolous insects. The pres-
ence of two pupal stages having a general resemblance to the winged
nymphal stages of other Hemiptera suggests that the so-called pupal
instars of the male coccid pertain to the juvenile period of the life
history and not to that of the imago. The work of Wigglesworth
(1948, I951a) on the hormonal control of transformation in the
reduviid Rhodnius shows that the juvenile hormone controls the
nymphal status up to the imago, and if this is true in other Hemiptera
the coccid pupa is not a part of the imaginal stage. Holometabolism
can be defined only as a type of metamorphosis ; the fact that it occurs
among the Hemiptera in the male coccid, and also in the Thysanop-
tera does not taxonomically relate these insects to each other or to
such holometabolous insects as Coleoptera, Lepidoptera, Diptera, and
Hymenoptera.

VIII. THYSANOPTERA

The Thysanoptera seem to contradict the principle that postembry-
onic metamorphosis is due to some structural aberration on the part
of the young insect that fits it to a special environment or way of liv-
ing. The active young thrips in appearance differs from the imago
little more than a young aphis differs from a winged adult aphis, and

NO. 9 INSECT METAMORPHOSIS—SNODGRASS 45

it would seem that in like manner it could grow into an adult thrips
without any radical process of transformation. However, after two
active, feeding nymplike stages (fig. 7 A,B) the young thrips becomes

SS MSW Wy Wi

LEELA e = ZB SSS BROS. SS

LALA OZ. NSAINASSSSSSS
WYO” ZA YS \\
Vif ee (pj aa Wx Nr

Fic. 7.—Life-history stages of a thysanopteron, Scirtothrips citri Moulton (out-
lines from Horton, 1918).

A, first instar. B, second instar. C, propupa. D, pupa. E, adult female.

inactive, ceases to feed, moults, and enters a quiescent stage known
as a propupa (C). The propupa in turn is followed by a second rest-
ing stage termed the pupa (D), from which after a final moult the
adult emerges (E). In the Terebrantia the wings appear in the
propupal stage as straplike outgrowths, which become still more ex-
tended in the pupa. In the Tubulifera the propupa differs little in

46 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

external appearance from the second nymph, since in this suborder
the wings do not appear until the pupal stage. In most of the Tubu-
lifera, however, there is a second pupal stage separated from the first
by a moult, making five immature instars in all, but according to
Priesner (1920) a propupal stage is absent in some species and in
others there is only one pupal stage.

The few external changes other than the growth of the wings that
take place during the postembryonic development of the Thysanop-
tera are of little consequence. In some forms the antennae are reduced
in the propupa and their segmentation becomes indistinct. In the
pupal stage the antennae elongate, their segmentation becomes distinct,
the form of the head approaches that of the imago, the compound
eyes increase in size, the ocelli appear, and the sexes are now distin-
guishable. Most of these changes are merely those that any ametabo-
lous nymph might go through in its development to maturity. The
resting stages in the life history of a thrips, however, suggest that
internal changes are going on, and, in fact, a reconstruction of some
of the internal organs takes place during the propupal and pupal
stages that is entirely comparable to the transformation processes of
holometabolous insects. These changes in the thrips affect the alimen-
tary canal, the salivary glands, the fat tissue, the muscular system,
and in a lesser degree the nervous system.

The alimentary canal of Liothrips oleae, according to Melis (1935),
does not differ essentially in external form during preimaginal stages
from that of the adult, but the cellular structure of the mesenteron
becomes highly unstable and is in a continuous state of reorganiza-
tion. On the other hand, in Parthenothrips dracaenae, as described
by Miller (1927), the alimentary canal undergoes changes in shape
and size as well as cellular reconstruction during the propupal and
pupal stages. In the two nymphal instars the long tubular ventriculus
is looped forward upon itself and then turns back to join the intes-
tine; in the propupa the whole canal becomes a simple straight tube
with no ventricular loop; in the pupa the ventricular loop reappears
but only as a short lateral fold from the middle of the tube; in the
imago the ventriculus is again bent forward on itself as in the nymph,
and there is a secondary small loop in the descending arm. Since the
alimentary canal of the adult becomes practically the same as that of
the young thrips the intervening changes might seem useless, except
that, as the insect takes no food during the propupal and pupal instars,
the ventricular changes may be simply economy adaptations to a lack of
need for a digestive organ. In Parthenopthrips, Miller says, there is
one renewal of the midgut epithelium. At the beginning of metamor-

NO. 9 INSECT METAMORPHOSIS—SNODGRASS 47

phosis in the last part of the second larval stage the regenerative
cells of the ventriculus actively multiply and later spread out to form
a new epithelial layer while the old degenerating layer is cast off into
the lumen. Elongation of the stomodaeum and the proctodaeum pro-
ceeds from cell proliferation by mitotic division in “imaginal rings”
of cells at the inner ends of these two ectodermal parts of the canal.

The Malpighian tubules of Liothrips oleae, according to Melis
(1935), undergo no appreciable transformation, being the same in
all stages. The salivary glands degenerate in the propupa and pupa,
and are reduced to long bodies crowded with large nuclei in a scant
protoplasm, but they are restored in the adult to essentially the
nymphai form. The cells of the fat body play the usual role in meta-
morphosis; they increase in size during nymphal life and store up
nutritive products in their cytoplasm, which in the propupa and pupa
are given out and consumed in the reconstruction of the muscles. The
change in the nervous system involves principally a transposition of
the brain from its nymphal position in the thorax into the head of
the adult, accompanied by development of the cerebral nerves and
their adaptation to the imaginal organs they innervate.

The reorganization of the muscular system is the most important
feature of metamorphosis in the Thysanoptera. As described by
Melis (1935) in Liothrips oleae, during the propupal stage the larval
muscles of the head go into complete histolysis, in the thorax and in
the last abdominal segment there is a partial myolysis, but most of
the abdominal muscles do not undergo any appreciable change. Dur-
ing the pupal stage there follows a total regeneration of the intrinsic
head muscles, and a reconstruction of the thoracic and abdominal
muscles to fit the needs of the adult. The processes of histolysis and
histogenesis as described in detail by Melis are the same as those in
typical holometabolous insects; muscles that are to be reconstructed
with new attachments undergo a partial dissolution, but the nuclei
persist in small fragments of cytoplasm that reassemble to form new
muscles, or attach themselves to remnants of old muscles to form
reconstructed muscles.

The internal metamorphosis of the Thysanoptera is thus seen to be
truly holometabolous, but the nymphlike form of the insects in all
the immature stages, and the small degree of external change from
nymph to pupa and from pupa to imago suggest that the so-called
pupal stages are merely the usual third and fourth instars, which have
become inactive because of the reconstructive process that takes place
within them. The immature stages of the Thysanoptera thus appear
to be comparable to the nymphal stages of ametabolous insects, with

48 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

the wings developed in the third or fourth instar. It is, therefore,
difficult to account for an internal metamorphosis for which there is
no apparent external reason.

IX. OLIGONEOPTERA, OR TYPICAL ENDOPTERYGOTA:
NEUROPTERA TO HYMENOPTERA

Whether the orders here included constitute a monophyletic group
of holometabolous insects or not will be a matter of opinion. Since
holometabolism occurs in the unrelated Coccidae and Thysanoptera,
some entomologists will contend that it may have arisen independ-
ently among other holometabolous orders. The modern larvae of the
typical endopterygote insects differ from the nymphs of ametabolous
insects and the larvae of hemimetabolous insects not only in being
endopterous but also in several other respects. They lack eyes that
are identical with the compound eyes of the adult, and usually they
have independently developed simple larval eyes; the hypopharynx,
when present, is more or less united with the labium; the body mus-
culature differs from the typical adult musculature in varying degrees ;
and metamorphosis from larva to imago in all cases involves an inter-
vening pupal stage.

Inasmuch as there can be little question that endopterygote insects
have been evolved from exopterygote ancestors, the simplest and most
reasonable view to take concerning the nature of the holometabolous
endopterygote larva is that it represents in modified form the
nymphal instars of ametabolous exopterygote insects. Both the larva
and the nymph are the active juvenile stage of the insect during which
the wings are developed. Whether the wings grow externally or in-
ternally, or may be retarded in their growth to a late instar, is a
difference of no consequence. The larva as well as the nymph has
wings in the course of development, and is not a “wingless” stage of
ontogeny. If the legs also are developed beneath the cuticle, the larva
for that reason is not “legless,” and does not represent an apodous
stage of ontogeny or phylogeny.

The principal problem concerning the origin of the endopterous
holometabolous larva involves the question: For what way of life
was the primary larva modified from an ordinary ametabolous nymph
that led to the acquisition of its distinctive features and its holo-
metabolous metamorphosis? The young cicada or the young stonefly
clearly show how, by simple structural adaptations for environments
different from those of the adult, a nymph might readily be converted
into a hemimetabolous larva, but external modifications do not account
for holometabolism.

NO. Q INSECT METAMORPHOSIS—SNODGRASS 49

The endopterous condition of the larva and the substitution of
short-sighted simple eyes for long-sighted compound eyes were con-
ceived by Lameere (1899) to have arisen as adaptations in a primary
nymphlike juvenile form to boring into plant stems. The theory, how-
ever, does not take into consideration the facts that most present-day
larvae of the boring type are specialized forms in their own orders,
and that free-living forms give no evidence of having been recon-
structed for life in the open from a primary boring type of larva. It
is hard to believe, for example, that the antecedents of the aquatic
Corydalus larva or the Dytiscus larva, or even those of terrestrial
beetle larvae lived in plant stems. As for the change of eyes, it would
seem that a boring larva would hardly need any eyes at all. Though
the Lameere theory of larval origin is thus not convincing, it is the
only theory that has been proposed to account specifically for the
characteristic external features of modern endopterygote larvae.

We can readily imagine that the suppression of external wing pads
during the nonfunctional period of their development would be a con-
venience to most any young insect regardless of its habitat. Wing-
less larvae, by comparison with winged nymphs, have certainly shown
a great superiority in ability to adapt themselves to different environ-
ments and to different ways of living.

A theory concerning the nature of the endopterygote larva, elabo-
rated by Jeschikov (1929), regards the larva as a free-living continu-
ation of the embryo; the larva has even been defined as such (Hen-
derson, 1949). First, we might ask, what animal is not a continuation
of the embryo? The theory of Jeschikov, however, contends that
the larva is an embryo, and that the nymphal stages of its ancestors
are all condensed in the pupa. However, in no other insects are the
wings developed in the embryo, at most they are represented only by
differentiated groups of cells in the embryonic epidermis. The ame-
tabolous and hemimetabolous Pterygota all show that wing develop-
ment is a function of postembryonic life. Periodic moulting is com-
mon to both nymphs and larvae, but it would be quite exceptional in
an embryo. If the larva is an embryo, cases of paedogenesis would
really be embryogenesis, and larval heteromorphosis would be embry-
onic heteromorphosis; some embryos would take to the water on
hatching, others would burrow into the ground, still others would
climb trees, and finally we should have embryos spinning cocoons and
transforming into pupae. These implications are rather too much
for the theory. When the embryo comes out of the egg and takes on
all the functions necessary for a free life, its embryonic stage is ended,

50 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. I22

though of course what we now call it is merely a matter of conven-
tional definition.

The endopterous condition of the larva very probably was not pro-
duced by a single mutation. In the simplest type of wing development
among modern endopterygote insects, as shown by Tower (1903) in
certain Coleoptera, the wing is first formed in the early pupa beneath
the cuticle of the last larval instar, and is therefore exposed only at
the moult to the pupa. If formed as a fold of the body wall at any
earlier stage the wing rudiment would be exposed at the next larval
moult. The first appearance of wing pads among exopterygote in-
sects on different instars shows that the wing growth may be retarded.
In the past history of those beetles in which the wing is not present
as a fold until the early pupa, the external growth of the wing must
have been first retarded and then suppressed until the end of larval
life, and we may conclude, therefore, that the first step in attaining
the endopterous condition was a retardation in the time of develop-
ment of the wing rudiment. The formation of a wing fold is not the
true beginning of the wing development; in earlier larval stages the
alar rudiment is present in the form of a thickening or a differenti-
ated group of cells in the epidermis, which is the wing in a state of
suppressed growth.

On the other hand, in most of the endopterygote insects the de-
velopment of the wings has been expedited by the early recession of
the growing wing rudiments into pockets of the epidermis beneath
the cuticle, which become closed and are thus not affected by the larval
moults. Within these pockets the wings can grow without being ex-
posed until they are everted at the moult to the pupa. According to
Tower (1903) the wings develop in this manner among the Coleop-
tera in Scarabaeidae, Coccinellidae, and Chrysomelidae; Patay (1939)
says the wings of Leptinotarsa develop in closed pockets toward the
end of the third instar. A familiar example of the usual recessed type
of wing development beginning in the second larval instar is that given
by Mercer (1900) for Pieris rapae.

The endopterous condition in its evolution, therefore, has probably
gone through two phases, both existing among modern insects. In
the first phase the growth of the wings presumably was suppressed
until the end of the juvenile period; in the second phase the wing
rudiments developed again at an early larval stage, but now sank into
the epidermis beneath the cuticle, thus still preserving the “wingless”
state of the young insect. It must be evident, then, that there is no
truly wingless larva of any winged insect; the wings exist in some

NO. 9 INSECT METAMORPHOSIS—SNODGRASS 5I

retarded stage of growth. The endopterygote larva, therefore, does
not represent an apterous stage of ontogeny, and much less does it
recapitulate an apterous stage of phylogeny.

Similarly, legless larvae are not truly apodous; the leg rudiments
are present in some form, though they may be greatly reduced. In
the honey bee, for example, Nelson (1915) has shown that external
leg rudiments are present on the embryo, but at the time of hatching
are reduced to discs in the epidermis, which later redevelop internally
in the larva. The leg tissue, therefore, is continuously present, though
it may not take the form of a leg bud until late in larval life.

The suppression of compound eyes during larval life, unlike the
suppression of wing pads, would not seem to confer any advantage
on a free-living young insect. The typical larval eyes are simple single
eyes, usually only a few in a group on each side of the head. They
are developed on the site of the future compound eyes and are con-
nected with the same part of the brain; but generally at the end of
larval life the larval eyes degenerate, and they never take any part in
the formation of the definitive compound eyes. In the Culicidae and
related Diptera it is shown by Constantineanu (1930) that the com-
pound eyes begin their development in an early stage of the larva,
and that the larval eyes, which are formed in the embryo, are retained
in the adult. Yet the two remain as entirely distinct organs. In the
larva of Panorpa there are 30 to 35 single eyes in a group on each
side of the head, and, as described by Bierbrodt (1942), these pan-
orpid larval eyes have attained the structure of ommatidia, and prob-
ably function as appositional compound eyes. However, the larval
eyes and their nerves degenerate during the pupal metamorphosis and
do not become the compound eyes of the adult. Here is a case, there-
fore, in which a larva has succeeded in reacquiring functional com-
pound eyes, but these larval eyes, as those of other insects, give place
to adult compound eyes newly developed in the pupa.

In discussing the origin and evolution of endopterygote larvae,
Chen (1946) contends that the primary larva, derived from an exop-
terous nymph, was aquatic, and he cites the megalopterous larvae,
particularly the larva of Corydalus, as being the closest modern repre-
sentative of the primary larva. Though it may be conceded that the
megalopterous larvae are relatively generalized modern forms, they
are nevertheless superficially modified for aquatic life, and life in the
water does not account for their more fundamental characters, which
are those of endopterygote larvae in general. The stonefly, mayfly,
and dragonfly larvae are all aquatic, and yet they have compound eyes
and external wing pads, and they transform without a pupal stage.

52 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

If we assume that the primary endopterygote larva was a modified
nymph, we might more reasonably expect it to be best represented
among modern forms by some of the simpler terrestrial larvae, such
as a raphidian larva (fig. 8B), or coleopterous larvae in the families
Carabidae, Staphylinidae (C), and Dermestidae (D). Such larvae
live in the same general habitat as the adults, feed on the same kind

Fic. 8—Examples of a generalized ametabolous nymph and of simple
holometabolous larvae.

A, Zootermopsis angusticollis (Hagen), nymph of a winged termite. B,
Agulla adnixa (Hagen), raphidian larva. C, Creophilus mavillaris Long, larva
of a staphylinid beetle. D, Attagenus piceus Oliv., larva of a dermestid beetle.

of food with the same kind of mouth parts, and have no structural
adaptations for any particular environment. Except for the lack of
external wing pads and compound eyes they resemble an ametabolous
nymph (A), and they differ least from the structure of the adults of
their species.

The larvae of the lower endopterygote groups show their closer
relation to the exopterygote insects in the possession of typical two-
clawed pretarsi. Those with paired movable claws on the feet include
the larvae of Megaloptera, Raphidioidea, most Neuroptera except
Sisyridae, and the larvae of the coleopterous families Carabidae,
Cicindellidae, Gyrinidae, Dytiscidae, Amphizoidae, and Noteridae.
The two-clawed foot evidently is primitive among the winged insects ;

NO. 9 INSECT METAMORPHOSIS—SNODGRASS 53

the single claw of the larva in the higher orders, therefore, is a
secondary larval modification, and does not represent a primitive one-
clawed pretarsus, or dactylopodite.

It would seem, therefore, that an unspecialized modern larva should
best represent the primary endopterygote larva, since from such a
larva evolution could more readily produce various specialized forms.
Yet even the simplest of modern larvae gives us no suggestion of how
or why it acquired its distinctive larval characters. The endopterous
holometabolous larva must, for the present, be accepted only as a fact ;
we have no evident explanation of its origin.

The reason for holometabolism, that is, for metamorphosis that
involves the intervention of a pupal stage between the larva and the
imago, is not to be found in the external characters of the larva. The
young mayfly or the young dragonfly differ externally from their
parents more than do the larvae of some endopterygote insects, but
yet they transform without a pupal stage. The pupal transformation
processes involve a variable degree of reconstruction of both external
and internal larval tissues, but, so far as known, they always include
at least a partial dissolution of the larval musculature accompanied
by the formation of new muscles or of new muscle attachments
for the imago. The cessation of muscular activity brings about the
quiescence of the pupal stage. Since the lesser degrees of change
in other internal organs might be accomplished direct from larva to
imago, it appears to be the disparity in the muscular system between
the young insect and the adult that constitutes the reason for holo-
metabolism. As we have seen, this is true also for male Coccidae and
the Thysanoptera. The external suppression of the wings, the absence
of compound eyes, or the presence of abdominal appendages in the
larva have nothing to do with the fact that the holometabolous larva
has a muscular system that cannot go over entirely or directly into
the adult musculature.

The somatic musculature of nearly all adult pterygote insects is
built on the same fundamental plan, though some muscles may be re-
duced or eliminated in the thoracic segments of wingless species or in
those having weak powers of flight. The musculature of exopterygote
nymphs is essentially like that of the adults, but in holometabolous in-
sects the musculature of the larva is usually very different from that
of the adult. The pattern of the larval musculature is simpler in the
less specialized larvae of each order, and it is least specialized in larvae
that differ least from the adults. Hence, we may suppose that the
simplification of the larval musculature in a primitive endopterygote

54 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. I22

larva was an economy measure correlated with the reduction of the
wings and the fewer abdominal movements that the larva had to
make as compared with the adult. Whenever the difference between
the larval musculature and the adult musculature reached a point
where new attachments for imaginal muscles became necessary, a new
moult had to intervene, and thus a pupal stage became interpolated
between the larva and the imago. With the pupa once established as
a reconstruction stage for the muscles, it served also increasingly for
the transformation of other tissues.

In most of the major holometabolous orders the larval musculature
becomes progressively more complex in the higher families. The
wormlike form assumed by so many larvae, and the consequent neces-
sity of a wormlike mechanism of movement readily accounts for the
specialization of the musculature in all vermiform types of larvae.
Since the insect larva, however, is not a worm, no matter how worm-
like it may be, its musculature is never that of a worm, it merely serves
mechanically to enable the larva to make wormlike movements.

Otherwise, the forms and structure of most modern specialized
holometabolous larvae are clearly adaptations to specific environments
or ways of living, usually different from those of the adult. Such
larvae have thus taken on temporarily structures useful only to them-
selves, which must be discarded at the final transformation to the
imago. The ordinary caterpillar with its short thoracic legs, its long
abdomen supported on leglike props, its strong biting and chewing
jaws and ample food tract is clearly made for feeding in the open and
for the storage of food reserves. A boring larva, on the other hand,
is unmistakably adapted to burrowing into wood or plant stems. The
larvae of Diptera were probably in the first place aquatic, but their
structure is readily adaptable to life in mud, fruit pulp, manure piles,
and the bodies of other animals. The grubs of wasps and bees are
incapable of self-support, but they are perfectly constructed for con-
finement in cells where they are furnished with food by their mothers
or other adult attendants. Internal parasitic larvae are usually greatly
simplified in structure because they have nothing to do but to feed
on the food in which they are immersed.

The presence of paired appendicular organs on the abdomen of
various endopterygote larvae has often been taken to be a retention
from the embryo of a stage representing a primitive polypod condition
in the ancestry of insects. Thus Chen (1946) says: “The primitive
larvae are presumably of the campodeoid-polypod type, having three
pairs of thoracic and ten pairs of abdominal legs; the latter bear each

NO. 9 INSECT METAMORPHOSIS—SNODGRASS 55
a vesicle and a stylus.” From this premise Chen concludes that the
Corydalus larva is the closest modern representative of the primary
larva, and that the latter was aquatic. In a former paper the writer
(1931) reviewed the structure of the appendicular organs on the ab-
domen of endopterygote larvae, and suggested that these appendages
represent the eversible vesicles and the styli of Thysanura.

Broad generalizations are always mentally comforting because they
relieve the mind from the confusion of seemingly unrelated facts, for
which reason also generalizations are prone to become wider than the
evidence on which they are based. A closer comparison of the ab-
dominal appendages of endopterygote larvae with the vesicles and
styli of Thysanura shows that the two sets of organs are not identical
in structure, which fact raises the suspicion that they may be in no
way related. Furthermore, there is no valid reason for supposing that
the primary endopterygote larva should have had thysanuran charac-
ters. Some exopterygote insects retain a single pair of styli on the
abdomen, but none of them has abdominal vesicles or other abdominal
appendages except cerci and the organs of copulation and egg laying,
which are a common inheritance of Thysanura and both exopterous
and endopterous Pterygota. Insects could not be encumbered with
abdominal appendages after they acquired wings. The polypod pro-
genitors of the insects are unknown; they probably became extinct
when the primitive apterygotes became hexapods. As we have seen,
even the Thysanura do not have true legs other than those of the
thorax. Possible vestiges of abdominal legs are retained among the
apterygotes only in the Protura and Collembola.

The adjectives “thysanuriform” and “campodeiform” as applied
to the more simple types of endopterygote larva can have only a de-
scriptive value. An endopterygote larva, no matter how thysanuri-
form it may be in appearance, is just as truly a winged insect as is
an exopterygote nymph, and it is much farther removed than the
nymph from its apterygote ancestors. Since it is hardly to be sup-
posed that the exopterygote orders and the endopterygote orders
represent two primary lines of divergence from primitive winged
insects, the endopterygotes must have had a long line of exopterygote
ancestry separating them from their apterygote progenitors. Exop-
terygote insects were already flourishing in Carboniferous times, en-
dopterygotes appear in the Permian; the earliest apterygotes (Col-
lembola) are known from the Devonian.

Larval abdominal appendages are most fully developed in the larva
of Corydalus (fig.9 A). Along each side of the abdomen on the first

56 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

Fic. 9.—Abdominal appendicular organs of holometabolous larvae.

A, Corydalus cornutus (L.), posterior abdominal segments and appendages.
B, same, terminal appendage of right side, mesal. C, same, cross section of gill-
bearing segment. D, Malacosoma americanum (F.), a right abdominal leg cut
open mesally to expose muscles. E, section of leg of a tardigrade, showing
muscles (from Baumann, 1921). F, panorpid larva, under surface of meta-
thorax and first abdominal segment. G, same, sternal arc of first abdominal
segment, posterior. H, Dineutes sp., larva, metathorax and anterior abdominal
segments. I, same, inner surface of an abdominal appendage-bearing lobe. J,
same, section of left side of an abdominal segment. K, same as I with inner layer
of muscles removed.

NO. 9 INSECT METAMORPHOSIS—SNODGRASS 57

eight segments is a row of lobelike projections between the terga and
sterna that fall in line with the bases of the thoracic legs. Each lobe
bears a long, tapering lateral process, and each but the last a large
ventral tubercle carrying a brush of gill filaments. On the tenth seg-
ment is a larger appendage (B) armed distally with a pair of strong
claws, and bearing on its outer surface a slender process like that of
the preceding appendages. In the larva of the related Chauliodes ven-
tral tubercles are absent, but long, tapering lateral processes are pres-
ent on the first eight abdominal segments, and the ninth segment bears
a pair of appendages similar to those of Corydalus. Ventral tubercles
are absent also in the Sialis larva (fig. 10 A), but long tapering lateral
processes are present, each of which is distinctly divided into six
segmentlike parts. The abdominal “legs” of caterpillars have a struc-
ture resembling so closely that of the gill tubercles of the Corydalus
larva as to suggest that the two are homologous organs. The same is
true of the abdominal “legs” of sawfly larvae, and of the apical ap-
pendages of trichopterous larvae, the structure of which has recently
been reviewed by Pryor (1951).

The abdominal vesicles of Thysanura are retractile by short muscles
arising on the supporting plates of the venter (fig. 2 C,D,E), and the
styli are movable by muscles arising on these same plates (D,E). The
plates are admittedly flattened remnants of abdominal limbs, and
Machilis demonstrates that the styli are coxal appendages (B,Sty)
acquired during postembryonic development. In the endopterygote
larva the abdominal tubercles are likewise retractile, but the principal
retractor is a long muscle taking its origin on the dorsal wall of the
corresponding body segment (fig. 9C,D). Whether this difference
in the musculature of the thysanuran and larval organs is significant
or not will be a matter of opinion, but the fact remains. On the other
hand, the basal musculature of the tapering lateral processes of the
megalopterous larvae (fig. 9 C) is quite comparable to that of a thy-
sanuran stylus in that it arises from the supporting body lobe, which
fact might therefore be taken as evidence that these processes truly
represent styli.

When, however, we note the occurrence of similar abdominal proc-
esses in other unrelated larvae, the interpretation of any of them as
primitive styli becomes doubtful. The aquatic larva of the gyrinid
beetle Dineutes (fig. 10 B), for example, has a pair of long, tapering
lateral filaments arising from each of the first eight abdominal seg-
ments and two pairs from the ninth segment. The single filaments
are supported on lobes of the body (fig. 9 H) that lie in a line above

58 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

the level of the pleural plates (P/) of the thorax. Each lobe is crossed
internally by two layers of vertical muscle fibers (I) enclosing a large
trachea between them (J,K,7ra) that runs out into the filament. The
filament itself is movable by two antagonistic muscles, one mesal, the
other lateral, attached on its base. Then there is the curious termi-
tophilous tineid caterpillar, Plastopolypus divisus (fig. 10C), first
described by Silvestri (1920), which has long, slender, multiarticulate

Fic. 10.—Examples of unrelated holometabolous larvae with lateral appendicu-
lar organs on the abdomen.

A, Sialis sp. Megaloptera. B, Dineutes sp. Coleoptera-Gyrinidae. C, Plasto-
polypus divisus Silv. Lepidoptera-Tineidae (from Hollande, Cachon, and Vail-
lant, 1951). D, Nymphula maculalis Clemens, Lepidoptera-Nymphulidae (from
Welch, 1916).

processes projecting from the sides of the first seven segments of the
abdomen. These appendicular structures have been shown by Hol-
lande, Cachon, and Vaillant (1951) to be sensory and not exudatory
organs, since they are covered with innervated setae and contain no
glandular tissue; but these writers, and also Silvestri, find that each
appendage is movable by a muscle inserted within its base. Must we,
therefore, interpret all these structures as representative of thysa-
nuran styli? Hollande, Cachon, and Vaillant contend that the abdomi-
nal appendages of the Plastopolypus caterpillar are merely secondary
adaptations of the larva to life in the termite colony, as are also true
exudatory lobes on the body of other termitophilous species.

NO. 9 INSECT METAMORPHOSIS—-SNODGRASS 59

Similar though nonmusculated processes are shown by Hollande,
Cachon, and Vaillant to be present on a termitophilous fly larvae,
but in this case two pairs are present on each body segment, one pair
lateral, the other dorsal, and several other fly larvae associated with
termites have simple nonarticulate lateral appendages, some very
small, others large and club-shaped. Then there are the aquatic cater-
pillars with gill filaments along the sides of the abdomen (fig. 10 D),
some of which are simple fingerlike processes and others elaborately
branched filaments. The panorpid larva is sometimes cited as an ex-
ample of a larva having abdominal leg rudiments, but an examination
of this larva shows that the supposed “legs” on the abdomen (fig.
9g G) do not fall in line with the thoracic legs, and correspond exactly
in position with seta-bearing papillae on the thoracic venter between
the legs (F).

All such examples of the presence of segmental appendicular struc-
tures on the larval abdomen only go to show the facility with which
the young insect can develop special organs for various purposes of
its own. As Pryor (1951) has pointed out, similarity of structure in
nonsegmented organs of locomotion is not necessarily a criterion of
homology. With respect to the abdominal “legs,” he says, “there is in
fact as much resemblance between a caterpillar and an onychophoran
or a tardigrade as between a caterpillar and Corydalus.” The tardi-
grade leg (fig. 9 E) has a long muscle (b) from the body wall, which,
according to Baumann (1921), is connected with each claw by a
slender tendon (f). Likewise in the Onychophora the plantar discs
and claws of the legs are retractile in the same manner by muscles,
some of which have their origins in the leg and others on the body
wall. Such cases of similarity in musculature, as noted by Pryor, are
evidently independent adaptations to the functional needs of locomo-
tor organs having the same type of structure and action.

Ideas that can be neither proved nor disproved have to depend on
circumstantial evidence for support, but when circumstantial evidence
is not conclusive they had better be dropped, or held for further in-
vestigation. This principle applies to the musculated abdominal ap-
pendages of endopterygote larvae. If these appendages are related
to the vesicles and styli of Thysanura, they are appurtenances of ab-
dominal legs and not legs themselves. On the other hand, many in-
sect embryos do have leg buds on the abdomen which usually disap-
pear, but it is certainly quite possible that embryonic remnants of
primitive legs might be retained and redeveloped in a new form for
postembryonic use. We have only to consider the extraordinary elab-

60 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

oration of pleuropodia in the embryos of some insects to see the ex-
tent to which appendages presumably equivalent to legs can be modi-
fied for a new purpose. Eastham (1930) has shown that rudiments
of abdominal appendages appear on the embryo of Pieris rapae in
line with the rudiments of the thoracic legs, and that there is no evi-
dent reason for not regarding all these appendages as serially homolo-
gous organs. Many cases might be cited from other arthropods in
which a leg rudiment develops into a very unleglike structure, and
perhaps a very good example of this is to be seen in the external geni-
tal organs of insects.

The idea usually deduced from the presence of paired movable
appendages on the abdomen of endopterygote larvae is that such
larvae are recapitulations of a polypod stage of insect ancestry. This
idea, however, is not supported by other characters of these larvae.
Take the caterpillar, for example; in no respect does it have a primi-
tive organization. In the structure of its head, its mouth parts, and
its muscular system the caterpillar is a highly specialized modern
insect form, and, most important, it is a stage of postembryonic
growth in which wings are in the course of development. The cater-
pillar has a polypod status because its abdominal appendages were
not suppressed in the embryo, but it does not represent a primitive
polypod stage of phylogeny. Wings certainly did not arise in a poly-
pod ancestor of the insects ; a winged centipede is hardly to be visual-
ized as a reality.

The holometabolous larva is an independent organism. It can de-
part to any extent from the structure of its parents, and it is under
no compulsion to recapitulate its ancestral history. The independence
of the larva begins with the embryo, which develops directly into the
larval form whatever this may be. From the experimental work of
Hegner (1911) on the eggs of Leptinotarsa, of Reith (1925) and
of Pauli (1927) on the eggs of muscoid flys, and of Smreczynski
(1938) on eggs of the beetle Agelastica alni, it is known that in these
insects the larval structure is fully determined in the preblastoderm
stage of the egg. Probably the same is true for many other holo-
metabolous insects. A few hours later, however, as shown by Geigy
(1931) in Drosophila, injury to the blastoderm causes defects in the
adult fly. In this very early stage, therefore, the egg has the poten-
tiality of producing both a larva and an imago, but the larval develop-
ment takes precedence over the imaginal development. The primary
business of the holometabolous embryo is to produce a larva; in so
doing it may entirely ignore its own phylogenetic history, and needs

NO. Q INSECT METAMORPHOSIS—SNODGRASS 61

only to conserve enough undifferentiated material for the reconstruc-
tion of the adult in the pupal stage. The embryo and the larva thus
become a single independent phase in the life history of the insect, but
this fact is not a vindication of the idea that the larva is simply a
continuation of the embryo leading a free life instead of being con-
fined to an egg shell. The reverse more nearly expresses the truth;
the specialized structure of the larva has been forced back on the
embryo until the embryo becomes a preliminary larva. A phyloge-
netic significance, therefore, cannot be attributed either to the larva
or the embryo of a holometabolous insect.

The insect larva owes its independence and its ability to take on
characters of its own to its release from the necessity of inheriting spe-
cial adult characters of its parents. The development of structures
practical only to the winged imago must be inhibited throughout the
embryonic and larval stages, and conversely, larval organs useful only
to the larva may not be transmitted to the imago. In this way both
the larva and the adult are free to become more and more specialized
in different directions, but the greater their divergence, the greater
becomes the degree of reconstruction required of the pupa. Yet the
larva, no matter how divergent it may become from the line of adult
phylogeny, must carry the adult inheritance as well as its own. The
potency for redeveloping the parent form either resides in the ability
of larval tissues to be transformed into imaginal tissues, or it is
carried by undifferentiated embryonic cells of the larva, which resume
the imaginal development in the pupa. In the more intense degrees
of pupal metamorphosis, as Tiegs (1922) has said, the changes
amount at times to an absolute death of the larva, the tissues of which
go into almost complete dissolution, and if imaginal reconstruction
cells were not present the larva would be left to decompose.

X. LARVAL HETEROMORPHOSIS

Heteromorphosis of the larva, commonly called hypermetamor-
phosis, is of frequent occurrence among predaceous and parasitic
species of insects, examples being known in Neuroptera, Coleoptera,
Strepsiptera, Lepidoptera, Hymenoptera, and Diptera. It seems re-
markable that a larva can assume two or more distinct forms during
its life history, and the fact that it may do so raises the question as
to how the juvenile hormone is able to control a succession of differ-
ent forms. However, since this hormone is nonspecific with regard
to related species of insects, it should be nonspecific with regard to
different larval forms of any one species. The hormone has nothing

62 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

to do with determining the structure of the young insect; this is the
work of hereditary factors. The hormone simply maintains the in-
tegrity of the juvenile form, whatever this may be, against the forces
of further development.

The ability of the embryo to develop into a larval form that has no
relation to the form of its parents is strange enough, but it is passing
strange that this same larva can change its form several times during
its larval life and still finally revert to the adult structure. The fact
of heteromorphosis appears to demonstrate the plasticity of larval
tissues, which seemingly can be molded and remolded by the growth
organizer to produce a succession of adaptive forms. The histologi-
cal changes that may take place in larval metamorphosis, however, as
well as the role of hormones remain yet to be investigated.

Two categories of heteromorphosis in parasitic larvae are to be
distinguished according as the adult female lays her eggs in the open,
or on or within the body or egg of the prospective host of the larva.
In the first case the newly hatched larva must be able by its own
activity either to find its appropriate host, or to attach itself to a
carrier that will transport it to the nest wherein are the host eggs or
larvae on which it is destined to feed. A first-instar larva of this
type, therefore, is constructed for an active life, and has been termed
a planidium (little wanderer), but its structure, of course, will depend
on the insect order to which its parents belong. After entry into the
nest or body of the host, however, the planidium transforms into a
second-stage larva of much simplified form and structure adapted to
a sedentary life of parasitism. With those species, on the other hand,
in which the eggs are attached on, or inserted into, the body of the
host, the young larva begins at once its parasitic existence ; it has no
need for an “active stage, and develops directly from the embryo into
a form adapted only for life and feeding usually within the host, and
accordingly in many species it is greatly reduced and simplified in
structure. During later instars, however, the simple larva takes on a
form more typical of the usual larva of its order.

PARASITES WITH A PLANIDIAL STAGE

The term planidium has a functional rather than a structural sig-
nificance, but it is remarkable how larvae in different orders have
taken on similar characters in adaptation to the requirements of
planidial life.

Coleoptera.—Larval heteromorphosis among the Coleoptera occurs
in the Carabidae, Staphylinidae, Meloidae, and Rhipiphoridae. Most

NO. 9 INSECT METAMORPHOSIS—SNODGRASS 63
familiar are the life histories of the blister beetles, Meloidae, some
of which feed on the eggs of grasshoppers, others infest the nests of
bees. The transformations of species of the American Epicauta are
well known from the early work of Riley (1876) and later papers by
Ingram and Douglas (1932) and Horsfall (1941). The European
Mylabris variabilis is described and fully illustrated by Paoli (1938).

The adult females of Epicauta and Mylabris deposit their eggs in
the ground where grasshoppers are likely to be, but not necessarily

Fic. 11.—Three larval stages of a meloid beetle, Mylabris variabilis Pall. (from
Paoli, 1938).

A, first instar, planidium. B, same, more enlarged. C, second instar. D,
fourth instar, similar to third and fifth instars.

close to a grasshopper’s nest. The first-stage larva is a planidium hav-
ing the form of an active generalized coleopterous larva with long
slender legs (fig. 11 A,B). It is commonly termed a “triungulin,”
though two of its pretarsal “claws” are merely strong spines. The
planidium runs actively about and burrows into the ground until it
finds the egg nest of an acridid. After feeding on a few eggs, it
moults and transforms into a more simple short-legged, soft-bodied
second instar (C). This larva resumes feeding, grows, and goes
through two more stages in which it becomes a thick scarabaeoid grub
(D). The next instar, which is the fifth, resembles the preceding, but
it comes out of the egg nest and burrows downward a short distance
into the ground. This larva in Mylabris, according to Paoli, does not
feed, and transforms into a sixth larva, in which the integument is
thick, rigid, and dark-colored, and the mandibles and legs are much

64 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

reduced. This sixth larva, the “ipnotica” of Paoli, is immobile and
passes the winter in a dormant condition. In the spring it sheds its
tough, protective integument. The seventh larva that emerges resem-
bles the fourth and fifth instars; it is again active, though it does not
feed, and burrows upward to near the surface of the ground, where
it transforms into the pupa, from which finally the adult emerges.
The life history of Epicauta is essentially the same as that of
Mylabris, though according to Horsfall (1941) the fifth larva of
E. pennsylvanica feeds to repletion before it burrows down into the
ground. Moreover, this fifth larva in species of Epicauta may trans-
form directly into the pupa, thus eliminating the hibernating form
and the seventh instar. Pupation following the fifth instar is said by
Ingram and Douglas (1932) to take place with individuals of Epi-
cauta lemniscata that complete the fifth stage in the spring, while
those maturing in the fall or under unfavorable weather conditions
go over into the hibernating stage, which is followed by the active
seventh instar. Horsfall (1941) reports the same thing for Epicauta
pennsylvanica. A notable feature in the metamorphosis of these
meloid larvae is its reversibility, as shown by the transformation of
the hibernating larva into an active burrowing form like that which
preceded it, though geotropically one is positive and the other negative.
Other species of Meloidae are parasites in the nests of bees in the
families Megachilidae and Andrenidae. The life history and larval
stages of one of these, Tricrania sanguinipennis, infesting the under-
ground nests of the andrenid Colletes rufithorax is fully described
by Parker and Boéving (1925). The female of Tricrania deposits her
eggs under small objects lying on the ground in the vicinity of the
nesting places of the bees. The newly hatched larvae are slender
planidia, tapering at each end, with long legs and well-sclerotized
body segments. It would appear that they might find their way di-
rectly into the nest of a prospective host, but observations show that
generally they attach themselves to a male bee seeking a female and
are thus carried into the nest. Within the brood cell the intruder first
devours the egg of the bee, thus making sure of no competition from
the bee larva that might otherwise hatch. At the first moult the
planidium transforms into a soft, smooth larva having a boat-shaped
form with the spiracles on the back, an adaptation that enables the
larva to float on the food mass of honey and pollen in the cell, which
constitutes its food from now on. At later stages, however, as the
larva grows larger and the food mass shrinks, the larva becomes a
fat scarabaeoid grub. In all, the Tricrania larva goes through six

NO. 9 INSECT METAMORPHOSIS—SNODGRASS 65

instars, but, except for a shortening of the legs and other minor modi-
fications, it makes no radical change of form after the second moult.
There is no hibernating larval stage in the life of Tricrania, but the
fifth and sixth instars remain within the unbroken fourth and fifth
skins, which serve also as a covering for the pupa. The adult beetles
are formed in the fall, but remain within the bee’s nest until the
following spring.

The metamorphosis of the carabid Lebia scapularis, which is pre-
daceous on the larvae of the elm leaf beetle, Galerucella luteola, is
described by Silvestri (1904) as follows. The young Lebia is a slen-
der, elongate larva of the planidium type, having legs adapted to run-
ning, well-developed mandibles, and a pair of long, jointed apical
processes on the abdomen. It attacks a Galerucella larva and feeds
on the viscera until its growing body becomes so large and loaded
with fat to such an extent that it can no longer move actively about.
In this condition it might fall an easy prey to other insects, but the
Lebia larva now encloses itself in a cocoon spun of silk threads from
the Malpighian tubules, and finally includes its prey in the cocoon.
When finished with feeding, and having attained its maximum de-
velopment, this first larva moults into a second form having a general
pale color, mouth parts unadapted to feeding, the legs and antennae
reduced to small conical stumps, and the caudal processes suppressed.
From this second, nonfeeding instar the larva goes into a prepupal
stage, in which the head takes on adult characters, and wing rudi-
ments are present on both the mesothorax and the metathorax. After
another moult the prepupa becomes a pupa. Silvestri makes no com-
ment on the unusual occurrence in the Coleoptera of a prepupa, which
evidently belongs to the pupal stage of the insect. Heteromorphosis
in Lebia, therefore, appears to affect the pupa as well as the larva.

Heteromorphosis again occurs in several genera of Staphylinidae
in which the larvae are parasitic on the pupae of Diptera within the
puparium. Wadsworth (1915) gives a good account of the life his-
tory and larval stages of Aleochara bilineata, a parasite of the cabbage
fly, Chortophila brassicae. The newly hatched larva is an active thysa-
nuriform planidium that must seek in the ground the puparium of
the cabbage fly. Having found a puparium it gnaws a hole in the
latter and feeds on the outside of the pupa from a puncture in the
pupal integument until it becomes much swollen. The second instar
of the larva is quite different from the first; the cuticle is soft and
white, the antennae and mouth parts are altered, the legs reduced to
vestiges, the claws are lost, and the caudal spines disappear. The third

66 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

larva resembles the second except for its larger size. Both the second
and the third larvae lie lengthwise on the back of the thorax of the
fly pupa, and obtain their food by suction. Pupation takes places in
the fly puparium, and the adult beetle gnaws its way out.

The larval heteromorphosis of the Rhipiphoridae is too well known
to need more than a brief notice. Species of Rhipidius are internal
parasites in all larval stages on cockroaches. Other species attacking
wasps and bees are endoparasitic in the first instar but take an external
position in later instars. The first larva of one of these species is an
active planidium which must find and attach itself to a carrier that
will transport it to the nest of a prospective host. Here it enters the
host larva as an internal parasite. When fully grown in the first in-
star, however, it leaves the body of the host through a puncture,
moults, and in the second stage takes a position across the back of
the host larva on the first or second thoracic segment. The external
larva loses the features of the first instar and becomes a grublike
parasite, some species being characterized by the presence of large
tubercles on the back.

Neuroptera-—Among the Neuroptera the larvae of nearly all spe-
cies are predaceous, their prey being mostly other insects which they
attack in the open. The larvae of some Mantispidae, however, feed
on spider eggs within the spider’s cocoon. The mantispid eggs being
laid on trees or bushes, the young larva must actively find its food.
The life history of Mantispa styriaca has long been known from the
work of Brauer (1869). The newly hatched Mantispa is a simple,
slender neuropteroid larva with relatively small mandibles and no
distinctive specialization. Hatched in the fall, it hibernates through
the winter ; in the spring it finds a spider’s cocoon and cuts its way
through the silken wall. With feeding in the first instar, the abdomen
becomes greatly enlarged, but at the first moult the larva changes to
a fat grub with a small head and greatly reduced legs. This change
of the active first-stage mantispid larva to a sedentary grub is cer-
tainly not a reversion to any ancestral larval form among the Neurop-
tera; it is an individual secondary adaptation to the life of ease and
plenty the larva is to lead from now on in the protection of the spider’s
cocoon, within which it finally pupates.

Strepsiptera.—The Strepsiptera are notable for the heteromorphosis
of the parasitic larval stages of both sexes, and also for the fact that
in most species the sexually mature female retains the larval form,
though with a greatly modified reproductive system, and remains
within the body of the host. In the genera Eoxenos and Mengenilla,
however, the female leaves the host and in the adult stage is found

NO. 9 INSECT METAMORPHOSIS—SNODGRASS 67

in the open, either free or enclosed in the last larval skin. The females
of these species are much less modified in structure than are the fe-
males of species that remain in the hosts. The males of all species are
free, fully winged insects.

The adult female of Eoxenos laboulbenet, as described by Parker
and Smith (1933), is broadly oval, with a distinct head and thorax
and a 10-segmented abdomen ; wings are absent, but the legs are rela-
tively long and segmented, mandibles are present though simple, and
spiracles occur on each of the first seven abdominal segments. At the
posterior border of the seventh abdominal segment is a genital aper-
ture leading into a short, open, median oviduct, but otherwise noth-
ing is to be seen of the reproductive system; in the mature female
the entire abdominal cavity is full of eggs. Females of Eoxenos
found within the last larval skin are enveloped also in a thin pupal
cuticle, showing that, in this genus at least, the female is a true imago.
In mating, as observed by Parker and Smith (1934), the male of
Eoxenos curves his abdomen beneath that of the female, but the
aedeagus pierces the integument of the female instead of entering the
genital aperture. The eggs hatch within the body of the female and
the young planidial larvae escape through the open inner end of the
oviduct. The host of Eoxenos laboulbenei was for a long time un-
known, but Carpentier (1939) finally discovered the parasitic stage
of the larva in the body of a lepismatid, Lepisma aurea.

In other strepsipteran genera, parasitic in Orthoptera, Pentatom-
idae, Fulgoridae, Cicadellidae, and particularly in Hymenoptera of
many families, both the female and the male develop to sexual ma-
turity within the body of the host. The female remains within the
host, and retains the form of the mature larva, having a large, soft
abdomen and a short, cylindrical, darkly sclerotized cephalothorax
(fig. 12 G), which latter alone is thrust out of the abdomen of the
host. Mandibles are present in some species, absent in others, but the
female is always wingless and legless. She remains enclosed in the
last larval skin, and the parasitic females are not known to have a
pupal stage. It is questionable, therefore, whether the egg-producing
parasitic female is a true adult or a sexually precocious last-stage
larva. As noted above, Parker and Smith found evidence of a pupal
stage in the nonparasitic female of Eoxenos laboulbenet.

The reproductive organs of the parasitic females are greatly re-
duced, and the abdomen is filled with a great mass of eggs. The de-
velopment of the eggs free in the body cavity is described by Brues
(1903) in Xenos peckti. Between the ventral surface of the body of
the female and the enveloping larval skin is a free space, or brood

68 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

chamber, which opens anteriorly at the base of the cephalothorax by a
median slit, or in some species by a pair of apertures. The brood
chamber is in communication with the body cavity of the female by
several funnel-shaped tubes. During mating, according to Perkins
(1918), the male of Stylops aterrima inserts the aedeagus into the an-
terior opening of the brood chamber; in Acroschismus wheeleri,
Schrader (1924) says the spermatozoa, after being discharged into
the brood chamber, find their way through the ventral ducts of the
female’s abdomen into the body cavity, where they disperse, penetrate
the egg membranes, and effect fertilization. Silvestri (1940), how-
ever, has apparently demonstrated that the male of Halictophagus
tettigometrae, in inserting the aedeagus through the ventral mem-
brane between the head and thorax, penetrates the body wall of the
female and discharges the spermatozoa directly into the haemocoele,
whence they finally migrate to the posterior extremity of the abdo-
men. The young larvae on hatching from the eggs escape from the
body of the female through the ventral funnels into the brood cham-
ber, and gain the exterior by way of the anterior opening of the
chamber.

The male strepsipteron develops also inside the body of the host,
and before emergence as an adult he is enveloped by the last larval
skin. Unlike the female, however, the male goes through a pupal
stage before transforming into a free-winged insect, leaving the pupal
cuticle behind within the larval skin.

For a good example of the larval stages of a typical strepsipteron
we may refer to the well-illustrated account by Kirkpatrick (1937)
of the life history of Corioxenos antestiae, a parasite of Pentatom-
idae infesting coffee plants in East Africa. The first-instar larva is a
planidium (fig. 12 A) of coleopterous type of structure, and attaches
itself (B) to an immature pentatomid of the genus Antestia. The
parasite remains motionless on the host until the latter moults, when
it bores into the body through the soft new skin. At its own first
moult the planidium transforms into a simple, soft-bodied, legless
scarabaeoid grub (C), in which even body segmentation is not visible.
With succeeding moults the larva goes through four more instars
(D,E,F), without any radical change of form except for the develop-
ment of a row of eight processes along the back of the abdomen. The
larva apparently feeds by the absorption of body liquids of the host
through its skin, and the dorsal protuberances are supposed to increase
the absorptive area. During the later larval stages (F) the body be-
comes differentiated into a slender cephalothorax and a large abdo-
men ; extrusion of the cephalothorax from the host takes place in the

NO. 9 INSECT METAMORPHOSIS—-SNODGRASS 69

seventh instar, after which the female larva moults to the final form
(G) within the last larval skin. The larval instars of the male resemble
those of the female, but the male goes through a pupal stage before
issuing as a winged adult.

Fic. 12.—Developmental stages of a strepsipteron, Corioxenos antestiae Blair
(from Kirkpatrick, 1937).

A, first instar, planidium in waiting attitude (length 0.25-0.27 mm.). B,
same, in position of attachment on host. C, first parasitic instar (length 0.4 mm.).
D, second parasitic instar, female. E, third parasitic instar, female. F, fifth
parasitic instar, female. G, mature female, unfertilized (length 3 mm.).

Extrusion of the cephalothorax of Corioxenos antestiae, Kirk-
patrick says, is always between the back plates of the third and fourth
segments of the host, the male in the middle and the female on one
side. In this species the female has a pair of openings into the brood
pouch, one right, the other left, a provision to insure that one or the
other will not be covered by a wing of the host. The body of the
mature larva and also of the adult female lies in the body of the host
with the ventral side of the abdomen uppermost and the cephalothorax

7O SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

bent posteriorly, so that the dorsal surface of the exposed cephalo-
thorax is uppermost when extruded. In this position of the female
copulation with a free male takes place.

An even more detailed account of the biology, life history, and
anatomy of a strepsipteron will be found in the paper by Silvestri
(1940) on Halictophagus tettigometrae Silv.

Lepidoptera.—The caterpillar in its body form is the most con-
servative of all holometabolous larvae, even predaceous species in
general preserve the eruciform type of structure. In the Epipyropidae,
however, the larvae of which are external parasites on Homoptera,
particularly Fulgoridae, there is a pronounced structural adaptation
of the larva to parasitic life and even a heteromorphosis. Since the
adult female deposits her eggs not on a prospective host but on vege-
tation, the first-stage larvae must attain a host by their own efforts,
and they resemble, as much as a caterpillar might, the planidia of
other parasites with similar habits. The newly hatched larva of Aga-
mopsyche threnodes is described by Perkins (1905) as a minute,
slender creature, tapering to the caudal extremity, and provided with
legs unusually long for a caterpillar ; it is clearly adapted for the ac-
tive life of a young predaceous insect that must find a host for itself.
The later instars, which.feed on the back of the abdomen of the host,
are very different from the first. In the mature stage the head is ex-
tremely small, the legs reduced, the mandibles minute; the body be-
comes contracted to an oval form, and the larva takes on a superficial
resemblance to a mealy bug, accentuated by the presence of a waxy
covering. A description of all the stages of Epipyrops eurybrachydis
Fletcher is given by Krishnamurti (1933). Among the Lepidoptera
various leaf-mining species also undergo a change of form during
larval life, being at first flattened for feeding within the leaf, and later,
on emergence from the mine, taking on the usual caterpillar form
for cocoon spinning.

Diptera.—Parasitic dipterous larvae, of which the first instar is of
planidium type, and which, therefore, are heteromorphic, are said by
Clausen (1940) to occur generally in the Acroceridae (Cyrtidae),
Bombyliidae, and Nemestrinidae, frequently in the Tachinidae, and
in a few species of Sarcophagidae.

The Acroceridae (Cyrtidae, Oncodidae) are parasitic in their larval
stages on spiders. The female deposits her eggs on bushes or trees,
and the young larvae by their own efforts must attach themselves to
spiders that chance to come their way, if they are to survive. The
larval stages of Oncodes pallipes as described by Millot (1938) may
be taken as typical of the family. The newly hatched planidium (fig.

NO. 9 INSECT METAMORPHOSIS—SNODGRASS 71

13 1) is not over 0.4 mm. in length, slender and elongate, with a small
head and 11 body segments. The body segments are sclerotized dor-
sally and ventrally and are armed with strong spines, a pair on the
last segment being particularly long. The mouth armature includes
a median sharp-pointed process and a pair of lateral hooks. The last
abdominal segment terminates in an attachment apparatus consisting
of three strong central hooks and a ventral semicircle of small spines.
The larva at this stage is metapneustic, having a pair of spiracles only
on the last segment. After hatching, the larva stands vertically by
means of the attachment structure on the end of its abdomen, but it
is capable of locomotion either by looping movements like those of a
measuring worm, or by small leaps of a few millimeters made by
suddenly straightening the curved body. If a young spider happens
to pass close by, the planidium springs upon it and penetrates into
its interior ; otherwise the prospective parasite will die in the course
of a few days. The parasite passes the winter without change in the
body of the spider. In the spring it moults into a second instar and
later again into a third instar (J). In these instars the larva is simply
a small fly maggot; the body is indistinctly segmented, tapering an-
teriorly, enlarged posteriorly, and ends with a small apical cone. There
are now two pairs of spiracles but those of the prothorax are not func-
tional. The infested spider remains alive and normally active almost
to the end, but at last the parasite consumes the vital organs of its
host and comes out to pupate, leaving nothing of the spider but the
empty skin.

The planidium of Pterodontia flavipes (fig. 13 H), another ac-
rocerid parasite of spiders, described by King (1916), resembles that
of Oncodes, but the body ends with a small adhesive disc between the
bases of a pair of long, slender spines. This larva, according to King,
progresses either by looping in the manner of a leech, or by jumping.
Preparatory to making a leap, the larva stands erect on the attach-
ment disc with the caudal spines extended backward; by a sudden
downward pressure of the spines the larva then throws itself a dis-
tance of five or six millimeters. When on moist surfaces, however,
King says, the larva progresses by extending and contracting its
body. This last observation is of particular interest because it shows
that the planidium still retains the common mode of locomotion of a
fly maggot. In its subsequent stages the Pterodontia larva returns to
the form of a simple, smooth-skinned maggot, which, when mature,
emerges from the body of the spider.

For another example of larval heteromorphosis among the Diptera,
we may refer to the paper by Clausen (1928) on Hyperalonia oeno-

72 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

maus, a bombyliid larval parasite of the scoliid Tiphia, which itself
is parasitic on grubs of the scarabaeid genus Anomala. The eggs of
Hyperalonia are deposited on the ground or dropped there by the
female in flight. The first larva (fig. 13 G) on hatching is a slender,
vermiform planidium 0.9 mm. in length, with a strongly sclerotized
head and 12 uniform body segments; the thoracic segments bear each
a pair of long, slender lateral spines, and on the apical segment of the
abdomen is a pair of similar but longer spines directed posteriorly.
This larva has to search through the ground for the buried cocoons
of Tiphia; after entering a cocoon it feeds on the thoracic region of
the Tiphia larva. At the first moult the planidium changes into a
much simpler larval form, which lacks the body spines and the strong
sclerotization of the head, and is characterized by deep intersegmen-
tal constrictions. In the third stage (F) the larva becomes thick-
bodied and grublike, but is not essentially different from the second
larva; it passes the winter in the cocoon of the host, and pupates the
following spring. The same species attacks also other scoliid genera,
and the Bombyliidae in general, according to Clausen, are parasitic
on Orthoptera (egg cases), Coleoptera, Lepidoptera, Diptera, and
Hymenoptera.

Hymenoptera.—Among the Hymenoptera, larval heteromorphosis
following a first-stage planidium occurs in the parasitic Perilampidae
and Eucharidae. The females in these families lay the eggs apart
from the host, and the young larvae are provided with a strongly
sclerotized integument which allows them to live a relatively long time
without desiccating and without feeding. By means of a caudal sucker
and long tail bristles the planidium is able to stand erect and to spring
at a prospective host.

The following account of the planidium of a species of Perilampus
parasitic on larvae of Chrysopa is given by H. S. Smith (1917). The
eggs are laid on the leaves of plants where the Chrysopa larvae are
looking for aphids. From the egg hatches an active planidium (fig.
13 A), which at first crawls rapidly about, but soon attaches itself to
the leaf by its caudal sucker and stands up at a right angle to the leaf
surface. In this position it may remain motionless for days at a time
until some insect comes within its reach. Then suddenly the planidium
becomes “frantically active, reaching and swaying back and forth in
its attempt to attach itself to the prospective host.” If a Chrysopa
larva comes too near, “the planidium attaches itself with lightning-
like quickness to a hair or bristle of the host. It then leisurely crawls
down the hair to the host’s body and attaches itself by its mouth
hooks.”” When the Chrysopa larva spins its cocoon and pupates, the

NO. 9 INSECT METAMORPHOSIS—-SNODGRASS 73

Fic. 13——Examples of hymenopterous and dipterous parasitic larvae with a
planidial first instar.

A, Perilampus chrysopae Crawford, planidium (from Smith, 1912). B, Peri-
lampus hyalinus Say, planidium (from Smith, 1912). C, Schizaspidia tennicornis
Ashm., planidium (from Clausen, 1923). D, same, second instar (from Clausen,
1923). E, same, third instar (from Clausen, 1923). F, Hyperolonia oenomaus
Rond., third instar (from Clausen, 1928). G, same, first instar (from Clausen,
1928). H, Pterodontia flavipes Say, first instar (from King, 1916). I, Oncodes
ie Latr., first instar (from Millot, 1938). J, same, third instar (from Millot,
1938).

74 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

planidium feeds as an external parasite on the pupa. Often, however,
. the parasite attaches itself to the stalk of a Chrysopa egg, in which
case the young chrysopid falls a sure victim to the Perilampus pla-
nidium in wait for it.

Another species of Perilampus described by H. S. Smith (1912)
is a secondary parasite on the fall webworm, Hyphantria cunea. The
planidium (fig. 13 B) enters the body of the caterpillar through the
skin and searches for the larva of a primary parasite, including both
dipterous and hymenopterous species. At the first moult the perilam-
pid larva loses the characteristic features of the planidium and changes
into an ordinary hymenopterous grub, in which form it remains
through subsequent instars with only slight modifications.

The description by Clausen (1923) of the life history of a Japanese
eucharid, Schizaspidia tenuicornis Ashm., parasitic on ant larvae, will
serve as a good example of the nature of the planidium and the
heteromorphosis of the larva in the Eucharidae. The eggs of this
species are laid by the females during the later part of summer in the
buds of trees and hatch the following spring. The newly emerged larva
(fig. 13 C) is a planidium scarcely more than one-tenth of a millimeter
in length, having suctorial mouth parts and a pair of sharp mandibles,
and is provided with a small adhesive disc at the posterior end of the
abdomen, which is armed with strong spines. Locomotion is accom-
plished by successive loopings and extensions of the body as the latter
is held to the support alternately by the mouth and the caudal disc.
When awaiting the chance arrival of a prospective host, however, the
planidium stands up at an angle of 45 degrees on its caudal sucker,
and then, when opportunity offers, it attaches itself by its mandibles
to a passing ant, and is thus transported to the ant’s nest. Here the
parasite is brushed off from its carrier and now attaches itself by its
jaws to an ant larva. At the first moult the special characters of the
planidium, together with the mandibles, are cast off with the exuviae.

The second instar of the parasite is a simple, thick, grublike larva
(fig. 13 D) having only suctorial mouth parts; it maintains its hold
on the host by the mandibles of its own cast skin. When now the ant
larva becomes a pupa, the parasite frees itself from the larval exuviae
of the ant and by means of its oral sucker attaches itself to the pupa.
It then moults to its third instar (E), in which the body segmentation
is lost and the mouth is armed with a short stiletto for puncturing the
skin of the pupa. The ant pupa is now sucked dry and soon dies, after
which the fully fed parasite pupates. Though the Schizaspidia larva
assumes a characteristically different form in each of its three instars,
the pupa is typically hymenopteran.

NO. 9 INSECT METAMORPHOSIS—SNODGRASS 75

The larval history of Schizaspidia tenuicornis shows how complex
the life of a parasite may be, and how both in its structure and its
instincts the young insect must become adapted in each instar to con-
form with the particular conditions that confront it. As noted by
Clausen, the Schizaspidia larva, in losing its mandibles, breaks with
all other hymenopterous parasites, in which the mandibles are retained
in all stages.

PARASITES WITHOUT A PLANIDIAL STAGE

Finally we come to those dipterous and hymenopterous parasites
of which the female deposits her eggs directly on or in the body of
the host or in the host egg. In these species the young larva on hatch-
ing finds itself in immediate contact with its food supply, and there
is hence no need of an active stage in its life history. The larva is
structurally adapted during its embryonic development for life in the
body of the host, and in many cases, especially with hymenopterous
parasites, the adaptive modifications result in a greatly simplified
larval form.

Diptera—Among the Diptera, modifications or special structural
developments of endoparasitic larvae appear to be related principally
to the function of respiration, but they may be superposed on a state
of simplification in which most of the usual vital organs are not yet
developed.

An extreme case of reduction or of delayed development accom-
panied by specialization in the first-stage larva of Diptera is seen in
the agromyzid Cryptochaetum, parasitic in scale insects, described by
Thorpe (1931, 1941). The eggs of the fly are inserted into the body
of a half-grown scale before the body wall has become hardened. The
first-stage larva of C. iceryae, according to Thorpe (1931), is little
more than a transparent cylindrical sac (fig. 14A), 0.3 to 0.4 of a
millimeter in length. There is no mouth or sclerotized mouth parts,
no somatic muscles, no spiracles, tracheal system, or heart. The ali-
mentary canal is complete, but the stomodaeum and the proctodaeum
are not open into the mesenteron, and no food is present in the tract ;
the parasite evidently absorbs nutriment through its integument from
the body liquid of the host. A special feature of the larva is the
presence of a pair of large, fingerlike diverticula containing blood
projecting from the posterior end of the body. In the second-stage
larva (B) the body becomes distinctly segmented, and the posterior
segments are ringed with short spines ; the mouth is open and strongly
sclerotized mouth parts are present; there is a tracheal system but no

76 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

spiracles, and a few longitudinal muscles have been developed. The
caudal diverticula of the first instar have lengthened into a pair of
tails nearly half the length of the body, and fine tracheal branches
later penetrate into their open basal parts. In the third stage the body
preserves the general form and structure of the second stage, but
the tails have increased greatly in length, being one and a half times
or more the length of the body. In its fourth stage (C) the larva

Fic. 14.—Three larval stages of an agromyzid dipteron, Cryptochaetum iceryae
(Will.), parasitic in the coccid Icerya purchasi Maskell (from Thorpe, 1931).

A, first instar. B, second instar. C, fourth instar.

becomes an ovoid, yellowish-white maggot composed of a head and
10 body segments ; the tails are greatly lengthened, slender filaments,
but have become brittle and are easily broken. Each body segment
has a belt of minute spines around its anterior end, anterior and pos-
terior spiracles are now present, but the hooklike posterior spiracles
are still closed, the alimentary canal is open, the muscular system is
fully developed.

In Cryptochaetum striatum (Thorpe, 1941) the larval stages are
said to be much the same as in C. iceryae, but in the third stage the
respiratory tails are Io times the length of the body and are filled for
at least two-thirds of their length with fine tracheal branches.

In contemplating a larva of such incomplete structure as that of

NO. 9 INSECT METAMORPHOSIS—SNODGRASS Ti.

the first instar of Cryptochaetum, the question comes up as to how it
got that way. The usual answer to the question is that the embryo
hatched at an early immature stage. Concerning the “early hatching”
idea, Thorpe (1931) says: “The theory obviously cannot be pushed
too far, for there are many truly adaptive characters which arise de
novo in insect larvae, and cannot in any way be described as embry-
onic.” There is no question that “adaptive characters” may include
the suppression of structures that are temporarily useless, as well as
the development of new structures that are only temporarily useful.
Nature is always economical where there is no need of prodigality.
A larva living in the midst of liquid food which it absorbs through its
skin has no use for a mouth, feeding organs, or a functional alimen-
tary canal, and no need of a locomotor muscular system. If also it
can get sufficient oxygen by absorption from the medium in which
it lives, there is no immediate need of a tracheal system. All these
negative conditions might be supposed to have been acquired by the
simple expedient of early hatching, but the larva, if so produced, is
not a normal early-stage embryo. The retarded state of development
very probably was early determined in the egg, and the larva must
then be what it is regardless of when it hatches. The principal new
structures of the Cryptochaetum larva, Thorpe points out, are the
respiratory tails. Otherwise the larva simply develops the other or-
gans when they are needed. The delay in development is a mere
economy, and numerous examples of various degrees of economy
might be cited from other species.

Hymenoptera.—In the Hymenoptera endoparasitic first-stage larvae
often have such strange forms that they would hardly be known for
young insects if their development had not been followed. Clausen
(1940) distinguishes, describes, and illustrates 14 different types of
first-stage parasitic larvae in the Hymenoptera, nearly all of which
but the planidium are endoparasitic. The eggs of some species are
deposited on the outside of the host, of others in the body cavity of
the host, and of still others in the host egg. The so-called “egg para-
sites,’ however, Clausen observes, are truly larval parasites, since
they feed on the larva and “their development is primarily at the
expense of that stage.” In the present discussion we are concerned
entirely with the forms of these first-instar larvae, which later take on
the more conservative structure of typical hymenopterous grubs. The
species are therefore heteromorphic, though their heteromorphism
affects principally the first instar. The change to the final form may
take place at the first moult, but often the second instar is inter-
mediate in form between the first and the following instars. As with

78 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. I22

the parasitic larvae of Diptera, these aberrant hymenopterous para-
sites present various special developments in combination with differ-
ent degrees of undevelopment of usual organs. Whatever their form
or structure may be, however, we must assume that in some way it
is fitting to the life these larvae live.

As an example of greatly simplified and specialized first-stage
larval structure in the Hymenoptera we may take the braconid Helori-
morpha antestiae, an internal parasite of the pentatomid Antestia, de-
scribed by Kirkpatrick (1937), or the similar larva of the ichneumonid
Limnerium validum, endoparasitic in the fall webworm, described by
Timberlake (1912). In each of these species the first-instar larva
(fig. 15 C) has an enormous “head” on a relatively small, simple, un-
segmented body ending in a long tapering tail. The only appendages
present are a pair of slender, incurved, sharp-pointed mandibles.
An even simpler larva of the same type is that of Platygaster
marchali (E).

In the second stage the Limnerium larva takes on a vermiform type
of structure with a small head and 12 body segments, the tail of the
first instar being greatly shortened. The third instar, as also that of
Helorwmorpha antestiae (fig. 15 D), is a typical hymenopterous larva.
The heteromorphosis of these species, therefore, results from the
extreme modification of the first instar ; in its subsequent changes the
larva merely returns to the usual form.

A somewhat more specialized type of first-instar larva occurs
among the Platygasteridae, examples of which are here illustrated at
A,B,F, and I of figure 15. The large anterior part of the body carries
the mandibles, antennal rudiments, and a pair of simple posterior
appendages. This headlike part of the larva has been shown by
Marchal (1906) to be a cephalothorax bearing the antennae, mouth
parts, and the prothoracic legs. The body region behind the cepha-
lothorax is partly or entirely segmented, and may end with tail ap-
pendages of various patterns. In their development these larvae even-
tually attain the form and structure of an ordinary hymenopterous
grub.

A curious type of first instar larva is characteristic of the Sceli-
onida; it is classed by Clausen (1940) as the “teleaform” type of
larva, but in form it suggests the embryo of a mouse (fig. 15 J).
Hadronotus ajax, an egg parasite of the squash bug, Anasa tristis,
furnishes a good example. The newly hatched larva (J) as described
by Schell (1943) is a slender creature with a sharp, tail-like caudal
horn curved anteriorly. The body is constricted between a large an-
terior part, probably a cephalothorax, and an elongate posterior part,

NO. 9 INSECT METAMORPHOSIS—SNODGRASS 79

but is unsegmented. The cephalothorax bears anteriorly a pair of
large, soft mandibles, and posteriorly a “labial projection.” The
caudal horn is a feeding accessory. This larva grows by a great in-
crease in the size of the abdomen only. In the second stage the
larva takes on an oval saclike form, still without segmentation. The
third instar, however, is a fully segmented, typical hymenopterous
larva (K).

Finally, we must note that not all parasitic hymenopterous larvae
take on queer forms in the first instar. Among the Proctotrypoidea
the larva of Phaenoserphus viator, parasitic on a carabid beetle larva,
as described by Eastham (1929), hatches in the form of a simple
grub (fig. 15 H), in which the abdomen becomes fully segmented
during the first instar. The only special character of this larva is the
presence of small paired ventral papillae on the eight body segments
following the prothorax. On the head, according to Eastham, are a
labrum, a pair of antennal papillae, a pair of sickle-shaped mandi-
bles, a pair of simple maxillary lobes, a small labium, and a sclerotized
ring supporting the mouth parts.

Advocates of the Berlese “early hatching” theory would explain
the simplicity of these first-instar parasitic larvae as products of im-
mature eclosion of the embryo. Thus Chen (1946), who discredits
the theory as applied to other larvae, says: “The precocious types
are confined to parasitic Hymenoptera and appear to have been inde-
pendently acquired by the different groups.” He then distinguishes
among these first-stage larvae a “‘vermiform polypod” type (fig.
15H), an “oligomerous protopod” type (A), and a “polymerous
protopod” type (F), supposedly representing successively earlier
stages of embryonic development. The usual implication of this theory
is that the different types of larvae correspond with phylogenetic
stages presumed to be recapitulated in embryonic development.

The presence of apparent abdominal appendages on the first-instar
larva of Phaenoserphus viator (fig. 15 H) gives this larva a polypod
appearance, but Eastham (1929) says the abdominal papillae may be
merely adaptive structures. He notes that the difficulty of ascribing
such a larva to a primitive embryonic stage “lies in comparing the
whole larva at any one stage in its life with any single embryonic
state.” The presence on the head of fully developed, typical hymenop-
terous larval mouth parts does not harmonize with the idea that the
larva respresents an early polypod stage of the embryo. The presence
of mandibles on such a simplified larva as that of Helorimorpha (C)
likewise shows that the form of this larva has no embryonic or phylo-

80 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

Fic. 15.—Examples of hymenopterous parasitic larvae without a planidial
first instar.

A, Platygaster herrickiit Packard, first instar (from Kulagin, 1898). B,
Platygaster instricator Kulagin, first instar (from Kulagin, 1898). C, Helori-
morpha sp., first instar (from Kirkpatrick, 1937). D, same, mature larva (from
Kirkpatrick, 1937). E, Platygaster marchali Kieffer, first instar (from Marchal,
1906). F, Synopeas sp., first instar (from Marchal, 1906). G, same, embryo
(from Marchal, 1906). H, Phaenoserphus viator Hal., first instar (from East-
ham, 1929). I, Tricacus remulus (Walker), first instar (outline from Marchal,
1906). J, Hadronotus ajax Girault, first instar (from Schell, 1943). K, same,
third instar (from Schell, 1943).

NO. 9 INSECT METAMORPHOSIS—SNODGRASS 81

genetic significance. The other appendages have simply been sup-
pressed as needless.

The same may be said of the so-called “‘oligomerous” and “‘poly-
merous protopod” larvae (fig. 15 A,B,F,I). They do not as a whole
have the structure of any one stage in ordinary embryonic develop-
ment, and none of them is suggestive of being a primitive embryo.
An embryo develops continuously, but these larvae maintain the
form and structure they have at hatching until the first moult, as does
any ordinary larva. In short, there is no reason for regarding them
as embryos. Just as a free, active, first-stage larva, or planidium, is
adapted to the predatory life it must lead, so these internal parasitic
larvae are adapted to an endoparasitic life. They are specialized both
in the forms they have, and in the developmental retardation of organs
they do not have and do not need. The principle of economy is in-
voked here just as with the simplified dipterous larvae.

In the first-stage Hadronotus larva (fig. 15 J) we see again an ex-
ample of early specialization in form accompanied by retardation in
the development of organs not immediately needed. If we consider
the numerous other forms of first-instar larva among the parasitic
Platygasteridae and Scelionidae, illustrations of which are assembled
by Clausen (1940, figs. 108-111, 113), it is clear there is no evident
logic in picking out any one form as representing a particular stage
of ordinary embryonic development. The development of Synopeus
rhanis within the egg from the blastula to the first larva (F), as illus-
trated by Marchal (1906, pl. 17), shows that the embryo (G) de-
velops directly from the beginning into the platygasterid larval form,
without going through any stages suggestive of those of an embryo
that develops into a typical free-living larva. Evidently the larval
form is determined in the egg, and the embryo, thus relieved from
phylogenetic influences, develops into a larva of the platygaster type.
The time of hatching has nothing to do with it.

An example of heteromorphosis affecting the first larval stage very
similar to that in the parasitic Hymenoptera occurs in the pseudoscor-
pion (Barrois, 1896; Vachon, 1938). The eggs at an early stage
of development are discharged into a brood pouch suspended below
the genital aperture of the female and are here nourished on a secre-
tion from the ovaries. On hatching, the larva breaks through both
the chorion and the wall of the brood pouch, but remains attached to
the outside of the latter by its ventral surface and the mouth region.
It is now nourished, as were the eggs, by the ovarial secretion dis-
charged into the brood pouch. At this stage the young pseudoscorpion
is a simple saclike creature with rudimentary appendages, but without

82 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. I22

body segmentation or internal organs. A deep musculated invagina-
tion on its ventral surface was regraded by Barrois as a sucking or-
gan, but Vachon has questioned this function. However, in some
manner the larva absorbs the ovarial secretion from the brood pouch
and completes its development in one instar. At the next moult it
takes on at once the adult structure in miniature. The so-called larva
might be regarded as a second embryo, but clearly it is an adaptive
form quite unlike any early stage in ordinary arachnid development.

The frequency with which larval heteromorphosis occurs among
unrelated insects shows that the larval organization is highly unstable
and that mutations make it readily responsive to the need of environ-
mental adaptation. A case of heteromorphosis among the vertebrates
would be most astonishing ; with the insects heteromorphosis is com-
monplace. The adaptational changes in the structure of heteromorphic
larvae from one instar to the next is good evidence that homomorphic
larvae are themselves merely juvenile adaptations to their various
modes of living. The ease with which the insect larva assumes a
form compatible with its living conditions is well illustrated by the
difference between a free-living planidium of one parasitic species and
the endoparasitic first larva of another related species. The planidium
is equipped for activity, for finding and attacking its prospective host ;
the endoparasite is reduced to the bare essentials needed for feeding on
an ambient food supply and for mere existence otherwise. It may be
noted here, also, that simplification of structure often occurs in the
second or following instars, as with species having a planidial first
larva, in which case “early hatching” cannot be invoked to account
for it. Whatever form the early larva may take on, however, it is
incumbent on the larva eventually to return to its parental form, and
this it does by first reverting in its later stages to the larval form
typical of its order or family.

XI. THE PUPAL TRANSFORMATION

The insect pupa is one of the most remarkable things in animate
nature; within it are intimately mingled the processes of both life
and death. With the shedding of the last larval skin the fully formed
pupa appears as a rough sketch of the future adult. The visible pupa,
however, is only an external shell; inside of it the larval tissues and
organs are being replaced by those of the imago. The juvenile hor-
mone no longer maintains the larval organization, and in consequence
the tissues of the larva have either gone into a state of dissolution, or,
under the influence of the developmental hormones, are being recon-

NO. 9 INSECT METAMORPHOSIS—SNODGRASS 83

structed into imaginal organs, while other organs of the imago are
being formed anew from undifferentiated living cells whose develop-
ment had been repressed by the juvenile hormone. Though hormones
control or regulate the transformation processes, there resides in the
pupa some mysterious organizing force that builds up the imago
from the larval material or from special cells that have been carried
by the larva from the embryo. When the imaginal structure is ac-
complished, the pupal skin is shed and the insect now reappears in the
parental form that produced the fertilized egg from which the larva
was hatched. The life cycle is thus completed, only to be indefinitely
repeated.

Various investigators have observed that the cells of tissues, par-
ticularly the epidermis, undergoing metamorphosis discharge dark-
staining globules from the nuclei. These globules are commonly
termed chromatic droplets. Earlier writers, as Pérez (1910) and
Poyarkoff (1910), regarded their discharge as a sign of “‘rejuvena-
tion” in the larval cells; by the rejection of the droplets the cells were
supposed to discard their larval ingredients and to be thus prepared
for a new growth. The same globules, however, are observed to re-
sult from the dissolution of nuclei, and Wigglesworth (1942), from
a review of the evidence, concluded that the droplets are always
formed in this way. He noted that they are present in the epidermis
during the moults of Rhodnius, and in greatest numbers where nuclei
appeared to be formed in excess of the need for new cells. Nuclei
become superfluous, he says, “either because they belong to specialized
larval structures that are being discarded, or because the exuberance
of cell division has led to their production in greater numbers than
are needed.” Several writers have observed chromatic droplets also
in the growing tissues of the embryo.

The epidermis.—Since the newly exposed pupa appears to be al-
ready perfectly formed and does not thereafter change externally, it
is the ectoderm that undergoes the first reconstruction. The change
to the pupal form, however, is not as sudden as it appears to be, since
long before the larval skin is cast off the transformation processes
had begun in the so-called prepupal stage of the larva, just as each
larval stage begins within the unbroken cuticle of the preceding instar.
Moulting and ecdysis, therefore, are not synchronous, and the two
terms are not synonymous.

The method by which the pupal epidermis is formed is not the
same in all insects. In various families of the Coleoptera in particular
it appears that the cells of the larval epidermis retain a faculty for
renewed and differential growth, and that in these insects most of the

84 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

larval epidermis goes over, with changes, directly into the pupal epi-
dermis. At the other extreme, in the higher orders the larval cells
enter a state of degeneration and are thrown into the body cavity,
while the entire pupal epidermis is generated anew from small groups
of cells, the imaginal discs, that have remained undifferentiated from
the embryo. Imaginal discs of undeveloped appendages, however, are
present in all cases, and represent adult structures whose growth is
continued during the larval life.

In the beetle Sitophilus (Calandra) oryza Murray and Tiegs (1935)
say it is usually possible to distinguish in the epidermis even of the
very young larva groups of small, more basophile cells that will form
the appendages, rostrum, and copulatory organs of the adult, but that
there is no distinct imaginal tissue to form the main part of the body
wall. Also in Leptinotarsa, according to Patay (1939), the larval
epidermis simply undergoes a renewal of developmental activity by
which it is transformed into the pupal epidermis without any process
of dissolution of its cells. Again, in the chrysomelid Galerucella,
Poyarkoff (1910) finds little evidence of destruction of larval cells in
the transformation of the epidermis from the larva to the pupa.

On the other hand, in Hymenoptera and Diptera there may be a
complete renewal of the epidermis from imaginal discs of the larva,
accompanied by a destruction of the larval cells. In the thorax the
imaginal discs of the appendages not only form the appendages them-
selves, but they spread outward on all sides to furnish new epidermis
for the thorax, and in the abdomen the pupal epidermis is likewise
proliferated from abdominal discs. As the new epithelium spreads
from the regeneration centers, the old cells of the larval epidermis go
into a state of dissolution and are forced into the body cavity where
they dissolve or are consumed by phagocytes.

According to Anglas (1901), in Vespa and Apis as the pupal epi-
dermis spreads from the proliferation centers, the old epidermis be-
comes vactiolated and separated from the basement membrane. The
new tissue advances by incorporating what remains of the larval cells,
the protoplasm of which is absorbed, digested, and assimilated by the
multiplying imaginal cells. The new epidermis is at first plastic, al-
lowing the modeling of the pupal form, but later it becomes fixed by
the hardening of the new cuticle. Anglas reports there is no phagocy-
tosis of the disintegrating larval cells, such as described in some other
insects.

In the chalcid Nasonia, Tiegs (1922) says that in the newly
hatched larva the ectoderm consists of large cells which constitute
the greater part of the integument, and of strips of small embryonic

NO. 9 INSECT METAMORPHOSIS—SNODGRASS 85

cells which are the imaginal discs of the future pupal integument. In
the last part of the final larval instar the larval cells go into a state
of cytoplasmic disintegration, which is partly chemical and partly due
to the action of phagocytes. At the same time the cells of the imagi-
nal discs of the epidermis multiply and spread out, replacing the dis-
integrating larval cells, until they re-form the entire body wall,
including such internal parts of it as the tendons of muscles, the
tentorium, and the thoracic phragmata.

The regeneration of the integument in Diptera from imaginal discs
has been described by various writers. Wahl (1901) gives a full ac-
count of the epidermal regeneration centers in the larva of Eristalis,
which on the thorax include the discs of the pupal respiratory trum-
pets, the wings, the halteres, and the legs, and on the abdomen
epithelial thickenings formed of embryonic cells. Pérez (1910) says
the newly generated epidermis of Calliphora on the thorax grows over
the larval epidermis, the cells of which are thus rejected into the body
cavity and phagocytized. In the abdomen the epidermal renovation is
progressive and slow, but here also the old cells are thrown into the
body cavity where they become the prey of phagocytes.

Finally, we may quote from the more recent paper by Robertson
(1936) on the epidermal regeneration in Drosophila. At an early
larval period the rudiments of the legs, wings, and halteres, Robertson
Says, are masses of embryonic cells sunken into pockets of the ecto-
derm, which remain open through hollow stalks. These pockets are
the imaginal discs which will regenerate the thoracic epidermis. As
development proceeds, the mouths of the stalks become wider and
the peripheral parts of the discs expand into the surrounding epi-
dermis, the cells of which gradually retreat and are sloughed off into
the body cavity, where they are taken up by phagocytes. The imaginal
discs continue to expand by cells multiplication until finally they unite
and thus replace the entire larval epidermis of the thorax with a new
epidermal epithelium, which is that of the pupa and the adult. On
the abdomen likewise the larval epithelium is replaced by a new epi-
dermis generated from islands of undifferentiated cells. On most of
the abdominal segments there are two pairs of these imaginal discs,
one pair dorsal, the other ventral, but the spiracles also are centers
of regeneration, making thus six discs on each spiracle-bearing seg-
ment. On the last segment, however, there is only a single, ventral
histoblast. During the early part of the pupal stage the cells of the
abdominal discs multiply and spread out, displacing the larval cells,
which are rejected into the body cavity and there phagocytized. At

86 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

about the thirty-sixth hour of pupal development in Drosophila the
imaginal epidermis is complete.

That the body setae may also undergo a metamorphosis is shown
by the studies of Krumins (1952) on Galleria mellonella, the wax
moth. The setal apparatus consists of the three usual cells, the tricho-
gen, the tormogen, and a sense cell. In all larval instars except the
first, the setae are slender and hairlike and are re-formed at each
moult, but in some cases the sense cell is subepidermal and in others it
is intraepidermal. At the moult to the pupa only those larval setae hav-
ing intraepidermal sense cells are re-formed on the pupa, and these
pupal setae are replaced on the imago by conical setae. The adult,
however, acquires also smaller conical setae not represented on the
pupa.

The appendages.—In the lower orders of holometabolous insects
in which the adult mouth parts are not essentially different from those
of the larva, and the larval legs are functional external organs, the
corresponding pupal appendages are formed simply within the cuticle
of the larval appendages. If the adult appendages are to be much
longer than the larval appendages, the growing organs become folded
beneath the larval cuticle until they can straighten out at the pupal
ecdysis. In some cases, however, the lengthening organs push back
into pockets of the pupal integument, but if the imaginal organ begins
its growth during larval stages it will be accommodated in a pocket of
the larval integument. Eassa (1953) very precisely describes the
growth of the imaginal antennae of Pieris brassicae, which have their
inception in the first larval stage. Each imaginal antennal rudiment
is differentiated around the sense cells and trichogenes at the base
of the corresponding short larval antenna; as it increases in size it
recedes into a pocket of the head wall, and becomes folded upon itself.
During the fifth (last) instar the antennal pocket of the forming
pupal head is open by a long slit under the yet unshed larval cuticle,
from which the antenna will be everted at the ecdysis of the pupa.
On the pupal head, however, the antenna has taken a much higher
position than that of the larval antenna. The intervening part of the
head wall, Eassa shows, is newly generated from the unfolding wall
of the antennal pocket.

The growth of the antenna and the reconstruction of the head in
Pieris is very similar to that which takes place in lower Diptera, ex-
cept that in the latter the antennal pouches may include the rudiments
of the compound eyes. In the cyclorrhaphous flies the antennae and
eyes are developed in pouches of the larval head commonly known
as the “frontal sacs,” which are ingrowths behind the frons (not “in-

NO. 9 INSECT METAMORPHOSIS—SNODGRASS 87

vaginations of the pharynx” as they are often said to be). With the
formation of the pupa the two postfrontal pouches unite, and when
everted their walls form a large part of the imaginal head bearing the
eyes and the antennae (see Snodgrass, 1953).

The imaginal legs of holometabolous insects, in which the larva
has external legs, are formed in the usual manner within the larval
legs and find space to grow beneath the larval cuticle. If the larva
is externally legless, however, the rudiments of the adult legs, which
may appear early in larval life, grow within peripodial pockets of the
larval epidermis, and are not everted until the moulting of the last
larval skin. In the same manner are developed the wings of all endop-
terous insects. In the higher orders the everting leg and wing pouches
contribute to the formation of the thoracic wall of the pupa.

The alimentary canal—Of the three constituent parts of the ali-
mentary tract of an insect, the stomodaeum and the proctodaeum are
unquestionably ectodermal since they are formed in the embryo as
ingrowths of the body wall. The embryonic mesenteron, however, is
ordinarily generated from cells proliferated at the inner ends of the
stomodaeum and proctodaeum, which, growing respectively rearward
and forward, envelop the yolk in a sac, which is the definitive mes-
enteron, or functional stomach of the insect known as the ventricu-
lus. Because of the mode of its embryonic origin, some writers have
insisted that the insect stomach also must be ectodermal. That this
interpretation is entirely unnecessary and evidently erroneous, how-
ever, has been shown by Henson (1946), who points out that the
tissue at the inner ends of the stomodaeum and the proctodaeum
represents the anterior and the posterior lips of the closed blastopore,
which, according to the rules of embryogeny, should normally gen-
erate ectoderm outwardly and endoderm inwardly. The writer (1935)
has explained the matter in essentially the same way in showing that
the anterior and posterior mesenteron rudiments are remnants of an
originally invaginated endoderm that regenerate the mesenteron. As
already noted in the introduction, the embryonic method of forming
the stomach is an adaptation to life in the egg. Inasmuch as the
embryo cannot take its food into its stomach in the manner of its
free-living ancestors, the embryonic stomach grows around the food
stored as yolk in the egg. The insect stomach, therefore, begins its
history with a metamorphic process, but it does not violate the germ-
layer theory.

Since the diet of an adult insect is often very different from
that of the larva, the alimentary canal in most holometabolous insects
undergoes a very considerable alteration during the pupal transforma-

88 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

tion, the change affecting not only the form and relative size of its
several parts, but also the nature of the epithelial wall. The degree
of reconstruction that takes place in the stomodaeum and proctodaeum
varies in different insects, but the mesenteron epithelium is probably
always renewed from the larva to the pupa, and in some cases it
undergoes a second renewal from the pupa to the adult. An interest-
ing feature in the metamorphosis of the alimentary canal, however,
is that the reconstructive changes do not proceed in the same manner
in all insects.

The stomodaeum and proctodaeum being of ectodermal origin,
their changes in the pupal metamorphosis are similar to those of the
epidermis. They may be merely remodeled by a renewal of activity
in their cells without any cell destruction, or they may be partly or
wholly regenerated from proliferation centers, accompanied by a de-
generation and elimination of the old larval cells, which are thrown
off into the body cavity. Where the proliferation centers are best dif-
ferentiated they take the form of circular bands of cells at the inner
end of the stomodaeum and the proctodaeum, termed the anterior
and posterior imaginal rings.

Reconstruction of the stomodaeal and proctodaeal epithelium by a
general renewal of developmental activity of the larval cells, without
accompanying cell destruction, has been described in some Coleoptera,
as in Galerucella by Poyarkoff (1910), Leptinotarsa by Patay (1939),
and Sitophilus by Murray and Tiegs (1935). In Tenebrio, accord-
ing to Rengel (1897), the remodeling of the stomodaeum and proc-
todaeum proceeds from their inner ends, but there are no specific
imaginal rings clearly differentiated. The old larval epithelium ap-
pears to be absorbed by the advancing newly formed cells. Dobrovsky
(1951) follows in great detail the anatomical alterations that take
place in the digestive tract of the honey bee during postembryonic
development. The stomodaeum and proctodaeum apparently are re-
modeled into the adult structure by a new growth of the larval cells.
In the wasp, according to Anglas (1901), the cells at the posterior end
of the stomodaeum and the anterior end of the proctodaeum begin at
the time of pupation an active proliferation extending respectively
backward and forward; the advancing new cells absorb the old and
thus renew the epithelium.

In Trichoptera, Litbben (1907) describes the remodeling of the
stomodaeum and proctodaeum by new growth of the larval cells. Russ
(1908), however, says that in Anabolia laevis imaginal rings are
present, though of little importance. The anterior ring serves only
for the lengthening of the stomodaeum and the formation of the

NO. Q INSECT METAMORPHOSIS—SNODGRASS 89

stomodaeal valve; the posterior ring is but weakly developed and
plays no important role in the reconstruction of the proctodaeum. A
part of the rectal region is regenerated from a circumanal zone of
proliferation.

The regeneration of the stomodaeum and proctodaeum from imagi-
nal rings is said by Tiegs (1922) to take place in the hymenopteron
Nasoma, but it is particularly in Lepidoptera and Diptera that these
proliferation centers have been observed. In the silkworm, according
to Verson (1905), the cells of the imaginal rings become active at the
change to the pupa, but they are merely centers of enlargement of the
stomodaeum and proctodaeum. Newly formed cells are added to the
larval cells already present, pushing the latter farther away without
replacing them. Otherwise the stomodaeum and proctodaeum, though
they undergo great changes in form, are remodeled by renewed ac-
tivity of the larval cells. Likewise in Malacosoma, Deegener (1908)
says the imaginal rings form only small additions to the larval stomo-
daeum and proctodaeum, and there is no degeneration or emission
of larval cells. The larval cells remain, forming the pupal epithelium
by reconstructive growth.

The imaginal rings of Calliphora erythrocephala are very precisely
described by Pérez (1910). The anterior ring is a circle of small
cells in the alimentary epithelium surrounding the base of the stomo-
daeal valve, and therefore on the dividing line between stomodaeum
and mesenteron. The posterior ring is a narrow circle of cells in the
intestinal wall just behind the bases of the Malpighian tubules. Cellu-
lar proliferation from the imaginal rings is said by Pérez to regen-
erate most of the stomodaeal and proctodaeal epithelium in Calliphora,
but the terminal parts are formed from anterior and posterior centers
of ectodermal proliferation. The degenerating replaced larval cells
are thrown off into the body cavity. In Calliphora vomitoria, accord-
ing to Van Rees (1889), there is only a partial regeneration of the
stomodaeum and proctodaeum from imaginal rings; the anterior part
of the stomodaeum is remodeled by transformation of the larval cells,
and in the proctodaeum the rectum is regenerated from behind for-
ward. In Drosophila the stomodaeal epithelium is described by Rob-
ertson (1936) as being mostly regenerated from the anterior imagi-
nal ring, but regeneration in the pharyngeal region proceeds from
the buds of the labium. “As the new epithelium forms, the old larval
cells are displaced into the body cavity where they are devoured by
phagocytes.” The proctodaeal epithelium of Drosophila is likewise
regenerated in its anterior part from the posterior imaginal ring, and

go SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

posteriorly by forward proliferation from the ectodermal imaginal
disc of the last body segment.

From these samples of the reconstruction processes that convert the
stomodaeum and the proctodaeum of the larva into the corresponding
parts of the adult, we may conclude that in the majority of insects
the larval cells of the ectodermal parts of the alimentary canal retain
the potentiality of rejuvenation. When released from the inhibition of
the juvenile hormone they proceed by renewed division and differ-
entiation with the formation of the adult organs. As with the epi-
dermis, however, there is a tendency for certain groups of cells to as-
sume more and more of the work of reconstruction, and these cells
finally take the form of specific regeneration centers, the so-called
imaginal rings. It is to be noted that the degenerating larval cells of
the stomodaeal and proctodaeal epithelia are thrown out into the body
cavity, as are those of the epidermis; the discarded epithelium of the
mesenteron, on the other hand, is ejected into the stomach lumen.

The mesenteron in its function is more specifically physiological
than are the ectodermal parts of the food tract, since it is the seat of
digestion and absorption, while the stomodaeum and proctodaeum
serve rather in a mechanical way for ingestion, storage, and elimina-
tion. The mesenteron, therefore, undergoes a more thorough renova-
tion during the pupal metamorphosis, since it must radically alter its
functional activities in response to the usual change of diet from
larva to adult. Probably in all holometabolous insects there is a com-
plete renewal of the mesenteron epithelium, but here again, as with
the epidermis, the stomodaeum and the proctodaeum, we find that the
method of renewal is not the same in all insects.

The larval epithelium of the mesenteron consists typically of two
sets of cells. Those of one set are the functional cells concerned with
secretion and absorption ; those of the other are small cells next to the
basement membrane between the bases of the functional cells, known
as replacement cells because by multiplying they form new functional
cells to take the place of those that have become exhausted and which
in a degenerating condition are thrown out into the lumen of the
stomach. In the majority of insects it is these replacement cells that
form also the entire epithelium of the pupal mesenteron, but some
Coleoptera appear to repeat the embryonic method of forming the
stomach, since they regenerate the mesenteron epithelium from cells
at the inner end of the stomodaeum.

According to Mansour (1934) the mesenteron epithelium is regen-
erated from the posterior end of the stomodaeum in representatives
of the following coleopterous families: Cucujidae, Chrysomelidae,

NO. 9 INSECT METAMORPHOSIS—SNODGRASS OI

Curculionidae, and Scolytidae. In Galerucella Poyarkoff (1910) says
the larval epithelium of the mesenteron is rejected in toto, including
the basement membrane, and that there is then formed a provisional
pupal epithelium derived from cells of the posterior face of the stomo-
daeal valve. The cells of the new pupal epithelium become differ-
entiated into ordinary epithelial cells and small replacement cells. The
pupal epithelium, however, is in turn replaced by an imaginal epithe-
lium formed by the pupal replacement cells, but the imaginal epi-
thelium is thus also derived primarily from the stomodaeum. In
Sitophilus (Calandra) oryza Mansour (1927) says that about three
days before the pupal moult, the larval epithelium of the mesenteron
collapses and degenerates, and together with the replacement cells is
thrown off into the lumen. The adult epithelium is then derived in
S. oryza and in other rhynchophorous species from the posterior end
of the transforming stomodaeum. According to Murray and Tiegs
(1935), however, the larval replacement cells of S. oryza are not
discharged with the old epithelium, but remain as a layer of scattered
cells on the outer surface of the new epithelium and eventually form
the mesenteron caeca.

The regeneration of the mesenteron of Leptinotarsa decimlineata
is described by Patay (1939) as follows. When the larva is ready
for transformation, the stomodaeal valve becomes the seat of an in-
tense proliferation, forming numerous fusiform cells of an embryonic
character. The basement membrane behind the valve soon breaks,
and the larval epithelium turns inward and rearward while the newly
formed cells from the valve extend over its outer surface. The larval
epithelium, including the replacement cells, is then soon rejected into
the lumen. The valve cells construct an entire new epithelium, includ-
ing islands of replacement cells and a basement membrane. Thus is
formed the pupal epithelium, but again at the moult to the imago the
pupal epithelium is rejected and the replacement cells reconstruct an
imaginal epithelium. The metamorphosis of the mesenteron of
Leptinotarsa as given by Patay is thus the same as that in Galerucella
as described by Poyarkoff.

Statements that the mesenteron is formed from cells of the pos-
terior end of the stomodaeum are not to be understood to mean that
these cells are ectodermal; as already noted, Henson (1946) has
shown that corresponding cells in the embryo represent the anterior
end of the blastopore, and therefore properly generate endoderm in-
ward, The imaginal ring of the larva, as said by Henson, “is not an
imaginal rudiment but a reactivated blastopore.”’

g2 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

The formation of the pupal mesenteron epithelium from replace-
ment cells of the larva is widespread among the insects, and is too
well known to need an extensive review here. According to Mansour
(1927) this method of epithelial regeneration is known to occur
among Coleoptera in Tenebrionidae, Histeridae, Hydrophilidae, Bos-
trychidae, Elateridae, Scarabidae, Buprestidae, Anoboliidae, Dytis-
cidae, and Lucanidae. It is the only method of replacement that has
been observed in Trichoptera, Lepidoptera, Hymenoptera, and Dip-
tera. At the beginning of metamorphosis in these insects, the diges-
tive cells of the larval epithelium go into a state of degeneration, while
the replacement cells enter a phase of active division, proliferating
new cells that spread out under the old epithelium and eventually re-
place it. The degenerating larval cells are cast off into the stomach
lumen, where they form a disintegrating mass of material known as
the “yellow body.”

Of particular interest are those cases in which the pupal epithelium
of the mesenteron is said to be replaced by a special imaginal epi-
thelium. However, without any renewal of the pupal epithelium, the
imaginal mesenteron may undergo changes of form and relative size.
Deegener (1904) described in Cybister the formation of a sepa-
rate epithelium for the pupa differing from that both of the larva and
the imago, the function of which he said is to digest the yellow body
resulting from the dissolution of the larval epithelium. Both the pupal
epithelium and the imaginal epithelium are generated from replace-
ment cells. We have already noted that Poyarkoff (1910) reports the
formation of a provisional pupal epithelium in Galerucella, which is
replaced by an imaginal epithelium generated from the replacement
cells of the pupal epithelium. Poyarkoff, however, contends that the
pupal epithelium of Galerucella is never functional because in the
pupal stage the mesenteron is closed at both ends. In the same way in
Leptinotarsa, according to Patay (1939), the pupal epithelium of the
mesenteron derived from the inner end of the stomodaeum is re-
placed by an imaginal epithelium regenerated from the pupal replace-
ment cells. In the coleopteron Acanthoscelides obtectus as described
by Bushnell (1936), the pupal epithelium formed from the larval re-
placement cells is later cast off into the stomach lumen, leaving only
a basal layer of cytoplasm containing the smaller nuclei, from which
there is then regenerated the definitive imaginal epithelium. The de-
generating material from the larval epithelium, Bushnell says, is prob-
ably digested and absorbed by the pupal epithelium, which is then it-
self cast off and gives place to the imaginal epithelium. Lastly, we
may note that Tiegs (1922) says the pupal epithelium of the chalcid

NO. 9 INSECT METAMORPHOSIS—SNODGRASS 93

Nasonia, which is formed from larval replacement cells, proceeds to
digest and absorb the detritus from the rejected larval epithelium,
after which it degenerates, but from its posterior part is formed the
definitive mesenteron of the adult.

Deegener (1904) contended that the presence of separate pupal
and imaginal epithelia in the mesenteron of many insects is evidence
that the pupa represents a former actively feeding stage in the life
history of holometabolous insects. Most students of insect metamor-
phosis, however, have seen a physiological reason for the formation of
a specific pupal mesenteron epithelium in the fact that the pupal
stomach must digest the disintegrating tissue of the larval mesenteron
thrown into it, in order that this material may be reutilized by the
developing imaginal organs. In this case the physiological require-
ments of the adult stomach will be very different from those of the
pupal stomach, and it is therefore but a physiological necessity that
the epithelium should be renewed for the purposes of the adult. Dee-
gener (1908) himself notes that there is no formation of a new imagi-
nal epithelium in Malacosoma; the pupal epithelium persists and goes
over directly into the epithelium of the imaginal mesenteron, but with
many changes in its cytological structure.

That there is a complete regeneration of the mesenteron epithelium
at the moults of the larva, as described by Mobusz (1897) in An-
threnus, has not generally been observed, but there is nothing improb-
able in Mébusz’s claim, since the replacement cells are active at all
times in renewing the depleted functional epithelium. According to
Henson (1929) the mesenteron epithelium of Vanessa is renovated
at each larval moult by the addition of new cells. It would be of in-
terest to know if any such change takes place in the successive forms
of heteromorphic larvae.

The Malpighian tubules——In some insects the Malpighian tubules
go over from the larva to the adult without any essential change, in
others their walls are regenerated from replacement cells while the
old cells degenerate, in still others the larval tubes completely disap-
pear and the imaginal organs grow out in their place as a new set of
tubes.

It is still an open question, or at least a disputed one, as to whether
the Malpighian tubules of insects are ectodermal or endodermal in
origin. Most investigators claim that they arise from the inner end
of the proctodaeum, others state as positively that they are outgrowths
of the posterior part of the mesenteron. In the embryo of the honey
bee Nelson (1915) says the rudiments of the tubules are formed prior
to the ingrowth of the proctodaeum as invaginations of the ectoderm

Q4 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

around the point where the proctodaeum is to appear. For a short
time, therefore, the tubules “open directly on the external surface of
the embryo.” A similar condition, according to Nelson, is known
otherwise only in Chalicodoma. If it occurred more widely we might
suspect that the Malpighian tubules were originally circumanal glands
of the integument, and that they have secondarily been carried inward
with the ingrowing proctodaeum.

The larval tubes of Hymenoptera that have been studied degenerate
and disappear, the imaginal tubes are formed anew. In the honey
bee Oertel (1930) says the larval tubes disappear apparently by
chemical means, not by phagocytosis. The imaginal tubes are then
formed as budlike outgrowths from the extreme anterior end of the
proctodaeum. According to Dobrovsky (1951) the ring of buds of
the imaginal tubes of the bee appear on the surface of the pylorus a
short distance behind the inner fold, or “diaphragm,” that separates
the lumen of the larval mesenteron from that of the proctodaeum.
These observations agree with those of Anglas (1901) that the larval
tubes of the wasp and bee arise from the front end of the procto-
daeum, disappear at metamorphosis, and are replaced by imaginal
tubes that grow out just behind their bases. In the same manner, ac-
cording to Tiegs (1922), are formed the imaginal tubules of the
chalcid Nasonia, though there are no larval tubules.

The developing imaginal Malpighian tubules of the beetle Leptino-
tarsa are described and distinctly illustrated by Patay (1939) as diver-
ticula from the anterior end of the proctodaeum. At the beginning of
pupation their cells take on an appearance of degeneration, the cyto-
plasm becoming vacuolated and the nuclei irregular, but after the
moult to the imago they soon again assume the aspect of normal func-
tional cells, and without destruction or cell division the persisting
larval tubules become the organs of the imago. In some other Coleop-
tera, however, the imaginal tubules are said to be regenerated from
small replacement cells in the walls of the larval organs. Poyarkoff
(1910) describes the imaginal tubules of Galerucella as being formed
in this manner, and Murray and Tiegs (1935) say the cells of the
larval tubules in Sitophilus (Calandra) degenerate in the pupa, while
new imaginal cells are proliferated by active mitosis of the replace-
ment cells until they form a new tube. The detritus of the larval cells
is not discharged but slowly absorbed.

A detailed account of the transformation of the Malpighian tubules
from the larva to the adult without dissolution or cell destruction is
given by Samson (1908) for the lepidopteron Heterogenea limacodes.
During the long prepupal stage of this species the Malpighian tubules

NO. 9 INSECT METAMORPHOSIS—-SNODGRASS 95

go into a degenerative state to such an extent that they appear to be
on their way to complete dissolution ; at the moult to the pupa, how-
ever, reconstructive changes begin that lead to the reformation of the
tubules into the organs of the imago. The imaginal tubules retain the
form of the larval tubules, but they have undergone an entire change
in their histological structure, which, Samson suggests, is correlated
with the change of food from the larva to the moth.

In the Diptera the Malpighian tubules, so far as observed, undergo
no essential change from larva to adult. Pérez (1910) says that the
cells of the tubules in Muscidae simply go into a resting condition
during the pupal period, and then again resume functional activity
in the imago. Robertson (1936) notes simply that the cell structure
of the tubules in Drosophila appears to be the same in the larva and
the imago. The tubules of Drosophila, Robertson says, open into the
digestive tract just in front of the posterior imaginal ring that regen-
erates the proctodaeum, from which fact it would appear “that the
Malpighian tubules of Drosophila belong to the mesenteron.” Henson
(1946) finds likewise in Calliphora that the Malpighian tubules grow
out in front of the posterior imaginal ring, so that not only the tubules
but also the pyloric region from which they arise are of endodermal
origin, and he believes that the same condition prevails in other insects.

The fat body.—The so-called fat body of the insect is a physiologi-
cal tissue; the changes its cells undergo from larva to imago are
merely the accompaniments or results of functional activities and are
not of the nature of a true metamorphosis. In the larva the fat cells
elaborate and store nutritive materials in the form of fat, albumi-
noids, and glycogen, which are utilized mostly in the pupal reconstruc-
tion, but may be carried over into the adult. In some insects there
is little or no destruction of the fat cells during metamorphosis, in
others most of the cells disintegrate in the pupa to liberate their stored
products, while a few are carried over intact to generate the fat body
of the adult. Insects such as most Coleoptera in which the pupal trans-
formation is less intense, and which feed amply in the adult stage,
have less need of larval food reserves, and show the least change in
the larval fat cells during metamorphosis. On the other hand, with
insects in which there is an extensive breakdown of larval tissues and
an almost complete reconstruction of adult tissues in the pupa, the
food material stored in the larval fat cells is of vital importance for
the reconstruction of new imaginal tissues. It is in such insects that
the fat cells most abundantly give up their contents to the pupal blood,
and perish in so doing, leaving only a few to go over into the adult
to form the imaginal fat body.

96 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

In the Muscidae, it is said by Pérez (1910), the larval fat cells dis-
integrate completely in the pupa and their remains are devoured by
phagocytes. The imaginal fat body, according to Pérez, is then re-
developed from mesenchymatous cells on the inner surface of the epi-
dermis, the abdominal fat tissue being derived from mesenchyme on
the inner surfaces of the imaginal discs of the epidermis. If the
imaginal fat body is renewed in this manner in the higher Diptera, its
formation from mesenchyme is paralleled by the renewal of the
muscles from free myoblasts in the same insects.

The oenocytes.—The oenocytes are specialized ectodermal cells de-
veloped from the epidermis in the neighborhood of the spiracles,
mostly in the abdomen. In some insects the oenocytes remain in the
epidermis, but usually they are liberated into the body cavity, where
they occur either in groups connected with the spiracular tracheae,
or freely scattered in association with the fat cells. Most students of
insect metamorphosis report that the oenocytes are renewed at the
pupal transformation, and Wigglesworth (1933) says there is in the
hemipteron Rhodnius a new generation of oenocytes formed at each
nymphal moult, though some of the old oenocytes persist. Accord-
ing to Albro (1930) the larval oenocytes of the beetle Galeruclla
nymphaeae persist very definitely up to the pupal period, but then they
undergo degeneration and histolysis. The smaller imaginal oenocytes
appear later newly proliferated from the epidermis. In Sitophilus
(Calandra) the larval oenocytes are said by Murray and Tiegs (1935)
to begin a slow distintegration in the prepupal stage, some being at-
tacked by leucocytes, but the majority later disappear without phago-
cytosis. The imaginal oenocytes are independently developed from
the epidermis of the abdomen close to the spiracles, but in the imago
they are mostly dispersed among the cells of the fat body. No budding
of imaginal oenocytes from larval oenocytes was observed by Mur-
ray and Tiegs, such as described by some earlier writers. In Leptino-
tarsa, Patay (1939) observes that the imaginal oenocytes scatter in
the body cavity by amoeboid movements.

The function of the oenocytes is still not exactly known, though the
cells are now thought to be secretory organs of some kind. It has com-
monly been observed that the appearance of secretional activity in the
cells is greatest at the times of moulting, and Albro (1930) expressed
a common opinion in her statement that secretion by the oenocytes “is
in some way, directly or indirectly, correlated with the phenomenon of
moulting seems highly probable.” Wigglesworth (1933), however,
finds that the oenocytes of Rhodnius show their greatest activity after
the new epidermis is complete. He concludes, therefore, that the oeno-

NO. 9 INSECT METAMORPHOSIS—SNODGRASS 97

cytes are concerned with the formation of the new cuticle, “that they
synthesize, and secrete into the blood, materials which go to form a
part of the cuticle.” This conclusion receives support also from the
fact that the oenocytes are specialized epidermal cells. For a good,
well-documented review of the present status of the oenocyte ques-
tion, see Richards (1951).

The tracheal system.—In most holometabolous insects the tracheal
system of the larva is carried over to the adult with little change other
than the development of new branches to accommodate the particular
needs of the imago, and the elimination of tracheae needed only by
the larva. As with other parts of the ectoderm, however, more com-
plex reconstructive processes take place in the tracheal tubes of some
insects, involving a dissolution of the larval epithelium and the regen-
eration of a new imaginal epithelium. According to Anglas (1901) the
tracheal system of the bee undergoes no true metamorphosis, the only
change being one of growth and extension by proliferation from the
ends of branches, and the enlargement of certain tubes to form the air
sacs of the adult. In the curculionid beetle Sitophilus (Calandra),
Murray and Tiegs (1935) say that the tracheal system of the adult
differs from that of the larva principally in the elaboration of the
thoracic tracheae. The larval tracheae are directly converted into the
adult tracheae, accompanied by cell division in the epithelium, but only
rarely is there any disintegration of the cells. Even terminal branches
within the metamorphosing larval muscles remain intact and become
reassociated with the newly forming imaginal fibers.

On the other hand, in the chalcid Nasonia, Tiegs (1922) finds that
there is an extensive reconstruction of imaginal tracheae from replace-
ment cells in the basal parts of the larval spiracle trunks. Partly by
disintegration and partly by phagocytosis, he says, the entire larval
tracheal system disappears, but regeneration of the imaginal epithe-
lium keeps pace with the destruction of the larval cells, so that there
is no discontinuity in the tracheal system itself. Pérez (1910) givesa
detailed account of the tracheal metamorphosis in Calliphora erythro-
cephala. Though the greater part of the larval system of the fly per-
sists into the imago with more or less extensive remodeling, certain
parts of it are destined to be totally destroyed by phagocytes, and to
be replaced by newly generated tissue. The tracheal regeneration cen-
ters, or histoblasts, are groups of small cells distributed through the
walls of the larval tubes ; they give rise to new branching trunks, and
replace the larval epithelial cells of those parts that have been de-
stroyed by phagocytes. The presence of histoblastic centers of re-
generation in the tracheal system, as in other parts of the ectoderm,

98 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

thus appears to be a specialized condition developed in only certain
groups of insects.

The dorsal blood vessel—From the descriptions of most writers
on the internal metamorphosis of insects it would appear that the
heart and aorta undergo little change from larva to adult during the
pupal transformation, and it has been observed in various insects that
the heart continues to beat throughout the pupal stage. In the wasp,
Anglas (1901) says the dorsal vessel undergoes no metamorphosis
except a change of form. According to Murray and Tiegs (1935) the
cells of the heart and aorta of Sitophilus increase in size during the
larval stage, but they do not divide, and they survive the period of
metamorphosis intact ; the alary muscles of the heart go over with little
change into the imago. In Leptinotarsa about the only change in the
heart described by Patay (1939) is the formation during the pupal
stage of the pulsatile vesicle in the mesothorax of the imago. Robert-
son (1936) says of Drosophila that “the dorsal vessel of the larva
seems to pass over directly into the adult,” and that “the alar muscles
either disappear and are re-formed in the late pupa, or they are some-
what altered, being much more delicate in the imago than in the
prepupa.”

In contrast to these accounts, Tiegs (1922) reports that the heart
of Nasonia undergoes a profound metamorphosis, beginning at the
time of larval defaecation. Just prior to this the cells of the heart
and the pericardium undergo a granular degeneration. The imaginal
heart is then regenerated mainly from scattered embryonic cells in the
heart wall. A new pericardium is formed from a mass of embryonic
cells lying below the larval pericardium, from which proliferating cells
extend forward, absorbing the elements of the larval pericardium as
they grow. Eight hours after defaecation, Tiegs says, the heart tube
of Nasonia has been completely regenerated, and below it is the re-
generated pericardium.

It seems probable that further studies on the heart of other insects
during metamorphosis will reveal greater changes than have hereto-
fore been reported, unless there is some special reason for the reno-
vation of the organ in Nasonia. On the other hand, if reorganization
in the structure of the heart is of common occurrence, it is difficult to
explain how a regular heartbeat is maintained during the pupal stage.
In the larva the heart beats continuously in a forward direction, but it
has been shown by Gerould (1924) and other investigators that dur-
ing the pupal and adult life in many insects there is a periodic reversal
in the direction of the beat. Gerould (1933) records the occurrence of
periodic heartbeat reversal in the pupa and imago of representatives

NO. 9 INSECT METAMORPHOSIS—SNODGRASS 99

of Coleoptera, Lepidoptera, Hymenoptera, and Diptera. “In general,”
he says, “normal reversal occurs independently of the central nervous
system and is essentially myogenic.” For a bibliography of the subject,
and a description of the structure and action of the heart in the pupa
and imago of Bombyx mort, see Gerould (1938).

The nervous system.—lIt is well known that changes in the gross
structure of the central nervous system commonly take place between
the larva and the adult. Ganglia are drawn forward or condensed by a
shortening of the connectives in both the thorax and the abdomen,
with the result that ganglionic masses on the nerve cords are fewer and
individual ganglia are displaced from their proper segments. On the
other hand, condensation of ganglia may be present in the larva, as in
the higher Diptera, in which all the body ganglia are united in a large
thoracic nerve mass closely connected with the brain. The significance
of these gross changes in the nervous system is not clear, but concen-
tration and anterior displacement of ganglia is always found in the
more specialized insects.

The internal reorganization of the nervous system during the pupal
transformation has been less studied than that of other tissues. Bauer
(1904) has shown that a reconstruction of the brain and the develop-
ment of the optic lobes of the adult proceeds from neuroblasts in the
larval brain, and, though he apparently made no special study of re-
organization in the other ganglia, he says that scarcely any other organ
system of the insects undergoes such a thorough metamorphosis as
does the central nervous system.

In their account of the metamorphosis of Sitophilus (Calandra)
oryza, Murray and Tiegs (1935) say that “no direct observations have
been made on the manner in which the nervous system of an insect
like Calandra becomes readjusted during metamorphosis to meet the
needs of the highly specialized imaginal musculature,” but they add
that “many new motor neurons doubtless develop from neuroblasts.”
However, “disintegration of larval cells occurs but rarely, and con-
sequently degenerating nerve trunks are never found, as in many other
insects.” In contrast to this Tiegs (1922) finds in the chalcid Nasoma
that the larval cells of the nerve cord degenerate, while the imaginal
neuroblasts begin to divide and multiply, growing at the expense of
the larval cells on which they nourish themselves. In the larval brain
there is a distinct layer of nonfunctioning neuroblasts outside the cen-
tral mass of functional cells. At the time of defaecation by the larva
the larval brain cells go into dissolution as do the nerve fibers, while
the neuroblasts become active and give rise particularly to the complex

100 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

optic lobes of the compound eyes and to the centers of the imaginal
ocelli and antennae.

A study of the developing innervation of the pupal legs of Tenebrio
molitor has been made by Sorokina-Agafonowa (1924), who describes
an elaborate definitive branching of motor and sensory nerves growing
out from the main leg nerve of the larva. The sensory branches go
to the epidermis and end in bipolar nerve cells. In a later part of the
pupal stage these end cells divide each into several cells until there are
hundreds of them which become connected with small setae of the
cuticle. The author points out that the connection between the nerve
cells and the receptor organs thus appears to be secondary and not
primary. It is generally said, however, that the sense cell of a setal
sense organ is a division product of a cell in the epidermis, and that the
sensory axon grows centrally from it (see Wigglesworth, 1953b).

A complete analysis of insect metamorphosis certainly should in-
clude a study of differences in the neuromuscular mechanisms between
larva and adult that form the basis of difference in sensory reactions
and instincts. It would seem that in many cases there must take place
in the pupa an extensive rearrangement of both sensory and motor
nerves and an almost complete reorganization of the neuron associa-
tions in the central nervous system to account for the behavioristic
differences between the larva and the adult. Since we cannot attribute
any degree of intelligence to a larva, the common act of spinning a
cocoon must be supposed to depend on some special pattern of struc-
ture in the larval nervous system that would be entirely useless to the
adult. Van der Kloot and Williams (1953a, 1953b) have made an in-
teresting analysis of the role of both external and internal stimuli in
the spinning of the cocoon by the Cecropia caterpillar.

A good example of a complex larval instinct is seen in the manner
by which the caterpillar of the bagworm moth, Thyridopteryx ephe-
meraefornus constructs its portable bag. Several hundred tiny larvae
may hatch out at the same time from the eggs of a single female moth.
After a period of dispersal they all settle down and proceed by identi-
cal methods to enclose themselves in conical bags. Each little cater-
pillar first with its mandibles cuts out a number of small oval pieces of
leaf epidermis (cork or blotting paper will do just as weil), and then
strings them together in a band with threads of its silk attached to the
leaf at each end (fig. 16 A). This done, instead of crawling beneath
the band, the caterpillar turns a complete somersault, going head first
over and under the band (B), landing on its back in reversed direction
(C). Then, righting itself (D), it cuts out more leaf bits and makes a
ventral band (E) continuous with the one over its back. It now has a

NO. 9 INSECT METAMORPHOSIS—SNODGRASS IOI

complete girdle about its thorax. Next, elevating its abdomen (F) it
lengthens the girdle downward until only its head and feet are ex-
posed below (G). Finally, when the bag encloses the whole body, the
anchoring threads break loose and the now fully clothed young cater-
pillar walks away (H_) to take its first meal on the leaf. As the cater-

Fic. 16—Construction of a bag by a newly hatched bagworm, Thyridopteryx
ephemeraeformis (Haw.).

A-H, consecutive acts of an individual larva making its bag from bits of leaf
epidermis cut out with its mandibles. I, an older specimen with later additions to
the bag, less enlarged.

pillar grows it merely enlarges the bag by leafy additions to the lower
edge (1). Such instinctive skill and methodical procedure as this of
the newly hatched bagworm must depend on the presence of a highly
developed mechanism for coordinated sensory and motor chain re-
actions in the central nervous system.

The muscular system.—In considering the metamorphosis of the
muscular system it must first be noted that all the muscles of all holo-
metabolous insects do not undergo the same degree of change. Five
categories may be distinguished: (1) Larval muscles that go over un-
changed into the adult, (2) larval muscles that are reconstructed into

102 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

imaginal muscles, (3) larval muscles that are destroyed and not re-
placed, (4) muscles newly formed for the imago replacing larval
muscles that have been completely destroyed, (5) newly formed mus-
cles not represented in the larva or which are as yet undeveloped in
the larval stage.

The histolysis and histogenesis of the muscles have been described
by many writers for various insects. The accounts are not all in entire
agreement as to the details of the processes, but a chief point of dif-
ference relates to the part that phagocytes may play in the destruction
of the larval muscles, a question which is of no concern to us in the
present discussion, and is fully reviewed by Oertel (1930). The most
important matter is the apparently well-established fact that in dif-
ferent insects the muscles are regenerated in different ways. In the
more generalized orders, such as Coleoptera, the histogenesis of a re-
organized muscle or of a replacement muscle is said to proceed from
small nuclei within the tissue of the larval muscle itself. On the other
hand, in the more specialized orders, particularly in the higher Diptera,
such muscles are remodeled or replaced by myoblasts originating owt-
side the larval muscles, probably generated from mesoderm in the
embryo. Muscles of appendages that are undeveloped in the larva are
in all cases derived from free myoblasts.

The degenerative processes in larval muscles are always pretty much
the same. The complete histolysis of a thoracic muscle of Ephestia
kiihniella is described as follows by Blaustein (1935). The advent of
degeneration appears at the beginning of the prepupal period with the
disappearance of cross striation in the muscle fibers. Lymphocytes
now enter the muscle through the sarcolemma and penetrate between
the fiber bundles, which lose their connections and separate from one
another. The sarcolemma is next broken, admitting increasing num-
bers of lymphocytes, and is finally ruptured on all sides. The lympho-
cytes, however, Blaustein says, probably do not at this time have a
phagocytic action on the muscle tissue. At the end of the third day of
pupal life the muscle nuclei begin to degenerate in large numbers, and
dissolve as the nuclear membranes disappear. The degenerating mus-
cle tissue is now attacked by phagocytic lymphocytes that penetrate be-
tween the dissociated fibrillae. By the end of the fourth day of the
pupa the histolysis of the muscle is complete, and there remains in the
place of the muscle only a great number of phagocytes engorged with
muscle fragments.

Essentially the same process of muscle degeneration has been de-
scribed by other writers for other insects. Some earlier writers re-
garded the lymphocytes penetrating the muscles as phagocytes, but it

NO. 9 INSECT METAMORPHOSIS—SNODGRASS 103

is now generally agreed that phagocytes do not initiate the destruction
of the muscles. They devour the products of the muscle disintegration,
and the greatly enlarged, engorged phagocytes may become extremely
numerous throughout the body cavity of the pupa, as in higher Dip-
tera, in which they have been called “spherules of granules,” or
“Kornchenkugeln.”

The reconstruction of the muscular system was thought by Berlese
(1902) to proceed from the nuclei of the larval muscles, which, being
set free in small masses of cytoplasm, became myocytes and were car-
ried to the places where imaginal muscles were to be formed. More
recent writers, however, find that in those insects in which the muscles
are reconstructed from intrinsic elements, the larval muscles contain
two sets of nuclei. Those of one set are the functional larval nuclei,
which are destroyed; those of the other set are converted into myo-
cytes, which form the new muscle in place of the degenerated larval
muscle. In the beetle Galerucella, for example, Poyarkoff (1910)
says that the larval muscles contain large nuclei that multiply by
amitosis, and small nuclei that multiply by mitosis. The first are the
larval nuclei, and will disappear ; the small nuclei are the regenerative
elements of the imaginal muscles. These mitotic nuclei become en-
closed in small masses of sarcoplasm to form myocytes, which asso-
ciate in long strands that eventually become the fibers of the new or
reconstructed imaginal muscle. The regeneration of muscles in Sito-
philus (Calandra) is similarly described by Murray and Tiegs (1935).
The small nuclei are at first scattered in the sarcoplasm of the larval
fibers, but as the muscle degenerates they migrate into the body of the
muscle, which becomes crowded with them. Here these nuclei form
myocytes, which unite into columns of cells that finally become the
imaginal fibers. Likewise the formation of adult muscles that replace
degenerating larval muscles is said by Patay (1939) in Leptinotarsa
to proceed from small peripheral nuclei within the tissue of the larval
muscles.

In the honey bee, Terre (1899) very concisely describes two sets of
nuclei in the larval muscles; those of one set are large nuclei in the
body of the muscle, the others are small nuclei mostly arranged in
longitudinal rows at the surfaces of the fibers. After the larva has
finished spinning its cocoon, the muscle substance degenerates and is
penetrated by the small nuclei, while the large nuclei dissolve and dis-
appear. The small nuclei become surrounded by masses of myoplasm
and thus become the myocytes that reconstruct the muscle for the
imago. On the other hand, in the account of the metamorphosis of the
muscles of the honey bee given by Oertel (1930) it would appear that

IO4 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

the myoblasts invade the muscle from the outside. Oertel does not dis-
cuss the genesis of the free myoblasts, but he says “it is commonly be-
lieved that the myoblasts are of mesodermal origin.” In the regenerat-
ing abdominal muscles he notes that the fibers in some cases are com-
pletely covered by myoblasts, and, in connection with the thoracic
muscles, that nuclei present in the larva before sealing of the comb
cell become incorporated into the new muscles. In the wasp Polistes,
according to Pérez (1912), the larval muscles have two sets of nuclei,
large larval nuclei in the body of the muscle, and small embryonic nu-
clei attached to the outside of the muscles. The muscles undergo a de-
generation and reconstruction without being entirely destroyed, but
the larval nuclei are mostly eliminated as the imaginal nuclei take their
places in the regenerating muscle. In comparing the muscle meta-
morphosis of the vespids with that of the muscid flies Pérez says the
only difference is that in the muscids the imaginal myoblasts are at first
exterior to the muscles, while in the wasps the myoblasts are attached
on the muscles they are to reconstruct and later become free in order
to proliferate outside the muscle.

In the chalcid Nasonia, according to Tiegs (1922), the adult mus-
cles are all formed from free mesodermal myoblasts, which are present
in the earliest larva. During the larval period the myoblasts are small
embryonic cells scattered in the body cavity close to the muscles. As
the larval muscles degenerate the neighboring myoblasts become active,
multiply by mitosis, penetrate the sarcolemma, and move about in the
disintegrating myoplasm by amoeboid movements. Eventually the
whole larval fiber, including the sarcolemma, disappears and the invad-
ing myocytes take its place, becoming arranged in rows that finally
form the new imaginal fibers.

The description by Blaustein (1935) of the muscle transformation
in the lepidopteron Ephestia kiihniella is not explicit as to the origin
of the myoblasts, but this author says that where a prospective muscle
is to be formed very small embryonic cells are first laid down. By
mitotic division they multiply, and by fusion with one another they
form long strands that become the imaginal muscle fibers.

The histogenesis of the muscles of the dipteron Psychoda alternata
is described by Schmidt (1929), but here again it is not clear whether
the myoblasts are intrinsic or extrinsic with relation to the larval mus-
cles. The dorsal longitudinal muscles of the metathorax of the larva
while undergoing degeneration lose their cross striation and the sar-
colemma disappears, the contractile substance and the plasma blend
into a homogeneous mass in which are imbedded many small nuclei,
which are the myoblast nuclei that will regenerate the imaginal
muscles.

NO. 9 INSECT METAMORPHOSIS—SNODGRASS 105

In the higher Diptera there appears to be no doubt that the imaginal
myoblasts are primitive embryonic cells at first free in the body cavity
of the larva. As examples of the process of muscle formation in the
higher Diptera we may cite from the paper by Pérez (1910) on the
metamorphosis of Muscidae, and from that by Robertson (1936) on
Drosophila. Both authors describe the myoblasts of the imaginal
muscles as originating outside the larval muscles. According to Pérez
the myocytes are mesodermal cells preexisting in the body cavity,
more or less in the vicinity of the epidermal histoblasts, but they are
not of ectodermal origin. They represent the precocious rudiments of
the imaginal musculature in a state of dissociation. These free myo-
cytes, Pérez asserts, are the homologues of the small regenerative
nuclei in the larval muscles of those insects in which the muscles are
re-formed from intrinsic elements. However, he does not suggest how
the free myocytes became dissociated from the larval muscles. Robert-
son does not discuss the origin of the free myocytes in Drosophila.

The adult muscles of Muscidae, according to Pérez, excepting those
that are exclusively imaginal, are mostly muscles that have been re-
constructed in the pupa from larval muscles. The larval muscle de-
generates into a homogeneous mass, which is then penetrated from the
outside by the myoblasts, which reconstruct the larval muscle tissue
into a muscle for the adult. The imaginal muscles of the muscids,
Pérez says, are thus formed from two different sources, the remains
of the larval muscles, and the embryonic myoblasts, the two being
combined in different proportions in different muscles. On the other
hand, Robertson says, “Practically all muscles of Drosophila are de-
stroyed by histolysis and consumed by phagocytes during the prepupal
and early pupal instars.” Thoracic muscles, which in Calliphora
Pérez believed were remodeled into imaginal muscles, according to
Robertson simply undergo a long-delayed histolysis. Myocytes of the
longitudinal thoracic muscles appear in the dorsal part of the pupa
of Drosophila as early as the fifth hour of the pupal period. They
surround the persisting larval muscles and increase greatly in numbers.
The larval muscles degenerate completely and disappear, leaving in
their place the myocytes, which spread out in the position of the future
imaginal muscles. Differentiation then proceeds anteriorly and pos-
teriorly from the central mass of myocytes until a new muscle is fully
formed.

Muscles newly generated in the pupa, having no representatives in
the larval musculature, are for the most part the muscles of append-
ages that are undeveloped in the larva, including the mouth parts, the
antennae, the legs, and the external reproductive organs. These mus-

106 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

cles necessarily are generated from unorganized groups of myoblasts
of mesodermal origin that are adventitious on the inner surfaces of
the ectodermal histoblasts of the appendages. If, however, an append-
age is functionally developed in the larva, it has its own normal larval
muscles, and these muscles will undergo a metamorphosis of the type
characteristic of the species. In the larva of the beetle Thymalus, for
example, Breed (1903) says the leg muscles go into a state of degener-
ation until they reach a structureless condition, but this condition is
of short duration and is followed by a phase of reconstruction.

The reason for the metamorphosis of the muscular system is not
hard to see; it is the difference between the musculature of the larva
and that of the imago. Breed (1903) argued that the larval muscula-
ture must undergo a reconstruction because of the specialized condi-
tion of the adult musculature in winged insects. The truth, however,
is clearly just the reverse. The adult musculature is essentially the
same in all insects from Ephemeroptera to Diptera, except that the
thoracic musculature is uniquely specialized in Odonata and is simpli-
fied in Blattidae, Mantidae, and Isoptera. The musculature of an
adult holometabolous insect, therefore, is in general no more special-
ized than that of a winged adult ametabolous or hemimetabolous in-
sect. It is the musculature of the holometabolous larva that has be-
come specialized for purposes of the larva. Its specialization was at
first perhaps one of simplification, but with the larval evolution the
larval musculature increases in complexity along patterns that have
little or no relation to the imaginal musculature because it becomes
adapted to the entirely different mechanism of movement in the larva.
The more different a larva becomes from the adult of its species, the
more specialized its musculature must be, and, therefore, it is in such
insects as Lepidoptera, Hymenoptera, and Diptera that the greatest
degree of muscle reconstruction occurs between larva and imago.

It is evident that the pupal transformation of the muscles is not
entirely comparable to the regeneration of any of the other tissues. The
formation of imaginal muscles from special nuclei within the larval
muscles might be likened to the regeneration of ectodermal parts from
histoblasts within the ectoderm, but the construction of muscles from
myoblasts scattered in the larval body has no counterpart in the re-
generation of other tissues. Furthermore, it is difficult to understand
how the free myocytes in one case, as Pérez contends, can be homo-
logues of the regenerative nuclei in the other, and it is quite mysterious
how mesodermal cells lying idle throughout embryonic and larval life
can be assembled in the pupa and induced to form new muscles for

NO. 9 INSECT METAMORPHOSIS—-SNODGRASS 107

the imago. Yet there seems to be no doubt that they do this very thing.
The essence of holometabolism is the muscle transformation.

XII. MUSCLE ATTACHMENTS AND THE NATURE OF THE PUPA

The somatic muscles of arthropods for mechanical reasons are
necessarily attached on the cuticle of the integument. The attachment
is by means of fine fibrils called tonofibrillae, which traverse the epi-
dermis from the cuticle and are attached to the muscle fibrillae ; their
outer ends in some cases appear to be embedded in the inner part of
the endocuticle. The nature of the tonofibrillae and the manner of
their formation have been discussed for half a century, and are still
not definitely known ; a review of opinion is given by Richards (1951)
and need not be repeated here. Probably the best explanation of the
tonofibrillae is that they are cuticular filaments formed by the epider-
mal cells where a muscle comes into contact with the integument; if
their outer ends are embedded in the cuticle we may assume that the
inner layer of the endocuticle was laid down subsequent to the forma-
tion of the tonofibrillae. The connection with the muscle fibrillae is
said to be formed by a splitting of the inner ends of the cuticular
fibers, which are thus “spliced” to the muscle fibrillae so that the two
become mechanically continuous.

It is well known that homologous muscles may have their attach-
ments at different places on the body wall in different insects. The
shift is generally attributed to “migration” of the muscles in the phylo-
genetic history of the insects; but in embryonic development and in
metamorphosis the muscles become attached where their ends come in
contact with the epidermis. It seems probable, therefore, that the for-
mation of tonofibrillae by the epidermis is evoked by the muscle con-
tact. A necessary condition for muscle attachments on the cuticle is
that the latter must be established when the epidermal cells are physio-
logically active and thus able to produce tonofibrillae while the cuticle
is in a formative state. Since most of the adult muscles of holome-
tabolous insects undergo a prolonged period of reconstruction in the
pupa, they do not make their final attachments until the end of the
pupal period when the imaginal cuticle is being formed. On the other
hand, if a muscle is ready for attachment at an early stage, as in
hemimetabolous insects, it can be attached at once on the imaginal cuti-
cle at the end of the larval stage.

The nature of the pupa has been a subject of much difference of
opinion. Perhaps the most common interpretation is that the holo-
metabolous pupa represents the last nymphal stage of insects without

108 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

metamorphosis, which would mean either that the pupa is simply a
modified last larval instar, or that the juvenile specialization that pro-
duced the larva stopped at the penultimate moult, so that the pupa is a
reversion to a nymphal stage with incompletely developed external
wings. The nymphal theory of the pupa is carried still further by
Jeschikov (1929), who contends that the larva is merely a free-living
stage of the embryo and that the pupa represents the whole period
of ancestral postembryonic development, “sie erscheint als Resultat
des Zusammenfliessens aller nymphalen Altersstufen.” The pupa it-
self sufficiently refutes this theory; it gives no evidence of being a
composite stadium since its external structure once formed remains
unchanged. (See also p. 49.)

A more reasonable theory concerning the nature of the pupa is that
of Poyarkoff (1914), which holds that the pupa is a preliminary
imaginal stage that has been separated from the final adult by an extra
moult in order to furnish a new cuticle for the attachment of muscles
reconstructed or newly formed in the pupa. Furthermore, Poyarkoff
adds, the pupa as a preliminary adult serves as a necessary mold for
the muscles forming within it, since in the larva these muscles could
not attain the size and the points of attachment appropriate for the
adult. It is only after the insect has assumed the external imaginal
form in the pupal stage that new muscles can be completed, but even
then they are still incapable of functioning because of the lack of
attachments. They cannot be attached at the beginning of pupation
since they are not yet formed, and they are not able to attach on the
pupal cuticle after the latter is hardened. Hence a new moult is
necessary to furnish the only condition in which tonofibrillae can be
formed for anchoring the muscles on the cuticle. Hinton (1948)
strongly advocates the views of Poyarkoff concerning the nature of
the holometabolous pupa. If the larval muscles had not departed
from the plan of the adult musculature, the larva might go over di-
rectly into the adult. The pupal moult is the solution on the part of
the insect to the problem of attaching new or reconstructed muscles.
The only evidence against this interpretation of the pupa that might
arise would be the discovery in some insect with a pupal stage that no
new muscle attachments are formed. At present no such condition is
known.

There can be no question that in its general form and structure the
pupa is an unfinished adult. The likeness to the adult is strikingly
seen in the relatively generalized raphidian pupa (fig. 17 B), which
has distinctly imaginal characters in the shape of the head, the long,
slender legs, the subsegmented tarsi, and the large, paired movable

NO. 9 INSECT METAMORPHOSIS—SNODGRASS 109

claws on each foot (D). When this pupa is ready to transform it
leaves the winter nest of the larva and crawls to a suitable place on
the bark of twigs of the tree, to which it tightly clings with its claws

TS; L

‘<
| i

Fic. 17.—Larva, pupa, and adult of a raphidian, and examples of pupal tarsi.

A, Agulla adnixa (Hagen), larva. B, same, pupa. C, same, adult female. D,
same, pupal tarsus. E, same, tarsus of adult. F, myrmelionid pupal tarsus. G,
Corydalus cornutus (L.), pupal tarsus. H, chrysopid pupal tarsus. I, Boreus
sp., pupal tarsus. J, Mantispa sp., pupal tarsus. K, Musca domestica (L.), pupal
tarsus.

(see Stein, 1838, Kastner, 1934). The pupa of the megalopteron
Nigronia serricornis (Say) also has paired claws, but in most of the
other neuropteroid families the end segment of the pupal leg is merely
split into two apical points (F, G), or it bears two small clawlike teeth
(H, I) within which the paired claws of the adult are formed. Ina

IIo SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

mantispid (J), however, the pupal tarsus ends with a simple expan-
sion, and in the higher insect orders, whether the larval leg is one-
clawed or two-clawed, the end of the pupal leg (K) is a simple lobe
ensheathing the pretarsus of the adult. The clawless pupal leg in the
higher orders, therefore, is a result of secondary simplification in an
appendage not yet needed for locomotion.

That the pupa is a part of the imaginal phase of the insect can be de-
duced from other lines of evidence. In the ametabolous or hemime-
tabolous insects the juvenile hormone maintains the nymphal or larval
status up to the transformation to the imago; in the holometabolous
insects the same hormone carries the larval form only up to the pupa.
Furthermore, the histoblasts of the larva, or imaginal discs, form di-
rectly not the organs of the adult but those of the pupa. The dividing
line that separates the holometabolous pupa from the larva, therefore,
is the same as that which separates the ametabolous imago from the
nymph. The holometabolous pupa and adult thus equate as a unit
\ith the ametabolous imago. Williams (1952) has shown that the
same hormone system, namely, that of the brain and the thoracic
glands, controls both pupation of the larva and the adult development
of the pupa. Finally, when we consider that all the internal organs
of the pupa are the adult organs in a state of being completed, the
pupa can hardly be regarded as anything else than a preliminary adult.
At the last larval moult, as Poyarkoff has said, the insect changes into
an imago, but the state of its internal organs does not permit it to be-
come at once an adult.

The occurrence of a moult in the imaginal stage, as Hinton (1948)
points out, is not limited to the holometabolous insects; it regularly
takes place in most Ephemeroptera, while in the apterygote insects
and the other arthropods moulting is usual throughout life. Hinton
suggests, therefore, that the pupa is equivalent to the ephemeropterid
subimago. However, it would hardly seem that there can be any real
relation between the imaginal moult of the mayfly and the moult of the
pupa in the very distantly related holometabolous insects. More prob-
ably the pupal moult was a secondary, independently developed moult
in the ancestors of the present holometabolous insects, rather than a
“throwback” to a time when adult moulting was a regular event. It
has been shown by Burks (1953) that the subimagines of Ephemer-
optera are sexually mature; their sperm and eggs mixed in normal
saline solution produce fertilized eggs, from which larvae may be
hatched. Some species, therefore, have simply eliminated the second
moult.

NO. 9 INSECT METAMORPHOSIS—SNODGRASS Lit

REFERENCES

ApAMs, J. A.

1933. The early instars of the firebrat, Thermobia domestica (Packard).

(Thysanura.) Proc. Iowa Acad. Sci., vol. 40, pp. 217-219.
Apro, HELEN T.

1930. A cytological study of the changes occurring in the oenocytes of
Galerucella nymphaeae L. during the larval and pupal periods of
development. Journ. Morph., vol. 50, pp. 527-567, 8 pls.

ANGLAS, J.

1901. Observationes sur les métamorphoses internes de la gttepe et de

Vabeille. Bull. Sci. France et Belgique, vol. 34, pp. 363-473.
Barrois, J.

1896. Mémoire sur le développement des Chelifer. Rev. Suisse Zool., vol. 3,

pp. 461-408, 2 pls.
Bauer, V.

1904. Zur innern Metamorphose des Centralnervensystems der Insekten.

Zool. Jahrb., Anat., vol. 20, pp. 123-152, 7 text figs., 1 pl.
BAUMANN, H.

1921. Beitrag zur Kenntnis der Anatomie der Tardigraden (Macrobiotus

Hufelandii). Zeitschr. wiss. Zool., vol. 118, pp. 637-652, Io figs.
Ber ESE, A.

1896. Le Cocciniglie Italiane viventi sugli agrumi. Parte III. I Diaspitti.
Riv. Patol. Veg., vol. 4, pp. 203-477.

1902. Osservazioni su fenomeni che avvengono durante la ninfosi degli
insetti metabolici. Parte II. Tessuto muscolare. Riv. Patol. Veg.,
vol. 10, pp. I-120, 32 figs.

Bickiey, W. E.

1942. On the stomodaeal nervous system of insects. Ann. Ent. Soc. Amer.,

vol. 35, PP. 343-354, 2 pls.
BIERBRODT, ERIKA.

1942. Der Larvenkopf von Panorpa communis L. und seine Verwandlung,
mit besonderer Beriicksichtigung des Gehirns und der Augen. Zool.
Jahrb., Anat., vol. 68, pp. 51-136, 33 text figs., 1 pl.

BLAUSTEIN, W.

1935. Histologische Untersuchungen iiber die Metamorphose der Mehlmotte
Ephestia kiihniella Zeller. Zeitschr. Morph. Okol. Tiere, vol. 30, pp.
333-354, 15 figs.

BopENSTEIN, D.

1943. Hormones and tissue competence in the development of Drosophila.
Biol. Bull., vol. 84, pp. 34-58, 2 pls.

1944. The induction of larval molts in Drosophila. Biol. Bull., vol. 86, pp.
113-124, I pl.

1950. The postembryonic development of Drosophila. In Demerec, Biclogy
of Drosophila, Chap. 4, pp. 275-367, 33 figs.

1953a. Endocrine control of metamorphosis with special reference to Holome-
tabola. Trans. IX Internat. Congr. Ent., Amsterdam, vol. 2, pp.
58-62.

1953b. The role of hormones in molting and metamorphosis. Jn Roeder, In-
sect physiology, pp. 879-931, fig. 244-257.

112 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

BounHIOL, J. J.

1938. Recherches expérimentales sur le déterminisme de la métamorphose
chez les Lépidoptéres. Bull. Biol. France et Belgique, Suppl. 24,
199 pp., 34 text figs., 2 pls.

1953. Role du corpus allatum dans la métamorphose des insectes. Trans.
IX Internat. Congr. Ent., Amsterdam, vol. 2, pp. 63-72.

BRAUER, F.

1869. Beschreibung der Verwandlungsgeschichte der Mantispa styriaca,
Poda. Verhandl. Zool.-Bot. Ges. Wien, vol. 19, pp. 831-840, 1 pl.
BreeD, R. S.

1903. The changes which occur in the muscles of a beetle, Thymalus mar-
ginicollis Chevr., during metamorphosis. Bull. Mus. Comp. Zool.,
vol. 40, pp. 317-382, 7 pls.

Brurs, Cra;

1903. A contribution to our knowledge of the Stylopidae. Zool. Jahrb.,

Anat., vol. 18, pp. 241-270, 3 text figs., 2 pls.
Burks, B. D.

1953. The mayflies, or Ephemeroptera, of Illinois. Bull. Illinois Nat. Hist.

Surv., vol. 26, 216 pp., 395 figs.
BURT yeas.

1938. On the corpora allata of dipterous insects. II. Proc. Roy. Soc. Lon-

don, ser. B, vol. 126, pp. 210-223, 3 figs.
BuSHNELL, R. J.

1936. The development and metamorphosis of the mid-intestinal epithelium
of Acanthoscelides obtectus (Say) (Coleoptera). Journ. Morph.,
vol. 60, pp. 221-241, 2 pls.

But Ler, HorTENSE.

1904. The labium of the Odonata. Trans. Amer. Ent. Soc., vol. 30, pp. 111-

133, 6 pls.
CALVERT Wea.

1898. The first filling of the tracheae with air in Odonata. Ent. News,
vol. 9, p. 73.

1911. Studies on Costa Rican Odonata, I. The larva of Cora. Ent. News,
vol. 22, pp. 49-64, 2 pls.

1928. The significance of odonate larvae for insect phylogeny. Proc. IV
Internat. Congr. Ent., vol. I, pp. 919-925, 7 figs.

CARPENTIER, F.

1939. Sur le parasitisme de la deuxi¢me forme larvaire d’Eoxenos laboul-
benet Peyer. Bull. Ann. Soc. Ent. Belgique, vol. 79, pp. 451-468,
12 figs.

CAZaT. PB:

1948. Les glandes endocrines rétro-cérébrales des insectes (étude morpho-
logique). Bull. Biol. France et Belgique, Suppl. 32, 227 pp., 186
figs.

CHAUDONNERET, J.

1949. A propos du corps jugal des Thysanoures. Bull. Soc. Zool. France,

vol. 74, pp. 164-167, 1 fig.
Cuen, S. H.

1946. Evolution of the insect larva. Trans. Roy. Ent. Soc. London, vol. 97,

pp. 381-404, 6 figs.

NO. 9 INSECT METAMORPHOSIS—SNODGRASS II3

Cuina, W. E.

1931. A remarkable mirid larva from Cuba, apparently belonging to a new
species of the genus Paracarnus Dist. (Hemiptera, Miridae). Ann.
Mag. Nat. Hist., ser. 10, vol. 8, pp. 283-288, 2 figs.

CLAASSEN, P. W.

1931. Plecoptera nymphs of America (north of Mexico). Thomas Say

Foundation, vol. 3, 199 pp., 35 pls.
CLAUSEN, C. P.

1923. The biology of Schizaspidia tenuicornis Ashm., a eucharid parasite
of Camponotus. Ann. Ent. Soc. Amer., vol. 16, pp. 195-217, 2 pls.

1928. Hyperalonia oenomaus Rond., a parasite of Tiphia larvae (Dip.,
Bombyliidae). Ann. Ent. Soc. Amer., vol. 21, pp. 642-659, I pl.

1940. Entomophagous insects, 688 pp., 257 figs. New York.

CoNSTANTINEANU, M. J.

1930. Der Aufbau der Sehorgane bei den in Stisswasser belebenden Dip-
terenlarven und bei Puppen und Imagenes von Culex. Zool. Jahrb.,
Anat., vol. 52, pp. 253-346, 25 text figs., 12 pls.

Corset, P. S.

1951. The development of the labium of Sympetrum striolatum (Charp.)

(Odonata). Ent. Month. Mag., vol. 87, pp. 289-296, 21 figs.
Day, M. F.

1943. The corpus allatum of the sheep ked, Melophagus ovinus L. Psyche,

vol. 50, pp. I-7, I pl.
DEEGENER, P.

1904. Die Entwicklung des Darmcanals der Insekten wahrend der Meta-
morphose, Teil I. Cybister roeseli Curtis. Zool. Jahrb., Anat.,
vol. 20, pp. 499-676, 11 pls.

1908. Die Entwicklung des Darmcanals der Insekten, wahrend der Meta-
morphose, Teil II. Malacosoma castrensis L. Zool. Jahrb., Anat.,
vol. 26, pp. 45-182, 5 pls.

De Lerma, B.
1937. Osservazioni sul sistema endocrino degli insetti (corpora allata e
corpi faringei). Arch. Zool. Italiano, vol. 24, pp. 339-368, 17 figs.
Demerec, M.
1950. Biology of Drosophila, 632 pp. New York and London.
Doprovsky, T. M.

1951. Postembryonic changes in the digestive tract of the worker honeybee
(Apis mellifera L.). Cornell Univ. Agr. Exp. Stat., Mem. 301,
45 pp., 12 pls.

IASSAU Yeni.

1953. The development of imaginal buds in the head of Pieris brassicae
Linn. (Lepidoptera). Trans. Roy. Ent. Soc. London, vol. 104, pp.
39-50, 3 text figs., 1 pl.

EastuaM, L. E. S.

1929. The post-embryonic development of Phaenoserphus viator Hal. (Proc-
totrypoidea), a parasite of the larva of Pterostichus niger (Carab-
idae), with notes on the anatomy of the larva. Parasitology, vol.
21, pp. I-21, 3 pls.

1930. The embryology of Pieris rapae—Organogeny. Philsoph. Trans. Roy.
Soc. London, ser. B, vol. 219, pp. 1-50, 9 pls.

II4 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

1938. Movements of the gills of ephemerid nymphs in relation to the water
currents produced by them. Journ. Queckett Micr. Club, ser. 4,
vol. I, pp. I-5.

1939. Gill movements of nymphal Ephemera danica (Ephemeroptera) and
the water currents caused by them. Journ. Exp. Biol., vol. 16, pp.

18-33, 8 figs.
Forp, E. B.
1937. Problems of heredity in the Lepidoptera. Biol. Rev., vol. 12, pp. 461-
503.
Geicy, R.

1931. Erzeugung rein imaginaler Defekte durch ultraviolette Eibestrahlung
bei Drosophila melanogaster. Roux’ Arch. Entw.-Mech., vol. 125,
pp. 406-447, 47 figs.

GEROULD, J. H.

1924. Periodic reversal of heart-beat in a chrysalis. Science, vol. 60, pp.
570-572.

1933. Orders of insects with heart-beat reversal. Biol. Bull., vol. 64, pp.
424-431.

1938. Structure and action of the heart of Bombyx mori and other insects.
Acta Zoologica, Internat. Tidskr. Zool. (Stockholm), vol. 19, pp.
207-352, 37 figs.

Granoi, M.

1950. Contributi allo studio dei Plecotteri. II. Morphologia comparata del
torace di alcuni specie de Plecotteri. Boll. Ist. Ent. Univ. Bologna,
vol. 18, pp. 30-57, 17 figs.

HEGNER, R. W.

191t. Experiments with chrysomelid beetles. III. The effects of killing parts
of the eggs of Leptinotarsa decemlineata. Biol. Bull., vol. 20, pp.
237-251, 17 figs.

HENnpERSON, I. F.

1949. A dictionary of scientific terms, 4th ed., 480 pp. Edinburgh and

London.
Henson, H.

1929. On the development of the mid-gut in the larval stages of Vanessa
urticae (Lepidoptera). Quart. Journ. Micr. Sci., vol. 73, pp. 87-105,
I pl.

1946. The theoretical aspect of insect metamorphosis. Biol. Rev., vol. 21,
pp. I-14, 4 figs.

Heymons, R.

1897. Entwicklungsgeschichtliche Untersuchungen an Lepisma saccharina
L. Zeitschr. wiss. Zool., vol. 62, pp. 583-631, 3 text figs., 2 pls.

1904. Die Hinterleibsanhange der Libellen und ihrer Larven. Ann. K. K.
Naturhist. Hofmus., Wien., vol. 19, pp. 21-58, 11 text figs., 1 pl.

1906. Uber die ersten Jugendformen von Machilis alternata Silv. Sitzb.
Ges. nat. Freunde Berlin, Jahrg. 1906, pp. 253-259.

Hinton, H. E.

1948. On the origin and function of the pupal stage. Trans. Roy. Ent. Soc.
London, vol. 99, pp. 395-409, I fig.

1951. The structure and function of the endocrine glands of the Lepidoptera.
Proc. South London Ent..and Nat. Hist. Soc. for 1950-51, pp. 124-
160, 19 figs.

NO. 9 INSECT METAMORPHOSIS—SNODGRASS 115

Hoiianpe, A., CACHON, J., AND VAILLANT, F.
1951. Recherches sur quelques larves d’insectes termitophile. Ann. Sci.
Nat., ser. 11, Zool., vol. 13, pp. 365-396, 30 figs.
HorsFAL., W. R.
1941. Biology of the black blister beetle (Coleoptera: Meloidae). Ann.
Ent. Soc. Amer., vol. 34, pp. 114-126, 1 pl.

Horton, J. R.
1918. The citrus thrips. U. S. Dept. Agr. Bull. 616, 42 pp., 10 figs.
Dpe; FP:

1935. Post embryological development of Ephemeroptera (mayflies). Cana-

dian Journ. Res., vol. 12, pp. 433-478, 13 figs.
IncRAM, J. W., AND Douctas, W. A.

1932. Notes on the life history of the striped blister beetle in southern

Louisiana. Journ. Econ. Ent., vol. 25, pp. 71-74.
JESCHIKOY, J.

1929. Zur Frage iiber die Entstehung der vollkommenen Verwandlung

Zool. Jahrb., Anat., vol. 50, pp. 601-652.
JoBLine, B.

1939. On the African Streblidae (Diptera Acalypterae) including the mor-
phology of the genus Ascodipteron Adens., and a description of a
new species. Parasitology, vol. 31, pp. 147-165, 4 figs.

Jory, N.

1872. Note sur un nouveau cas d’hypermetamorphose constaté chez le
Palingenia virgo a Vétat de larve. Ann. Sci. Nat., ser. 5, Zool.,
vol. 15, 5 pp.

KAstner, A.

1934. Zur Lebensweise der Kamelhalsfliegen (Raphidiina). Zool. Anz., vol.

108, pp. I-II, 6 figs.
LGU (Cad eel ls

1916. Observations on the life history of Pterodontia flavipes Gray. (Dip

tera.) Ann. Ent. Soc. Amer., vol. 9, pp. 300-321, 2 pls.
KirKPATRICK, T. W.

1937. Studies on the ecology of coffee plantations in East Africa. II. The
autecology of Antestia spp. (Pentatomidae) with a particular ac-
count of a strepsipterous parasite. Trans. Roy. Ent. Soc. London,
vol. 86, pt. 14, pp. 247-343, 40 figs.

KrISHNAMURTI, B.

1933. On the biology and morphology of Epipyrops eurybrachydis Fletcher.

Journ. Bombay Nat. Hist. Soc., vol. 36, No. 4, pp. 944-949, 1 pl.
KRuMINS, R.

1952. Die Borstenentwicklung bei der Wachsmotte Galleria mellonella L.

Biol. Zentralbl., vol. 71, pp. 183-210, 13 figs.
Kutacin, N.

1898. Beitrage zur Kenntnis der Entwicklungsgeschichte von Platygaster.

Zeitschr. wiss. Zool., vol. 63, pp. 195-235, 2 pls.
La BAUME, von W.

1909. Uber die Metamorphose der Ephemeriden. Sitzb. Ges. nat. Freunde

Berlin, Jahrg. 1900, pp. 137-153, 3 figs.
LAMEERE, A.

1899. La raison d’étre des métamorphoses chez les insectes. Ann. Soc. Ent.

Belgique, vol. 43, pp. 619-636.

116 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

LAUTERBORN, R.

1903. Tracheenkiemen an der Beine einer Perliden-Larve (Taeniopteryx

nebulosa L.). Zool. Anz., vol. 26, pp. 637-642, 2 figs.
Lee, Heren Tsui-YInG.

1948. A comparative morphological study of the prothoracic glandular bands
of some lepidopterous larvae, with special reference to their inner-
vation. Ann. Ent. Soc. Amer., vol. 41, pp. 200-205, 1 pl.

LinpsAy, EDrER.

1939. The biology of the silverfish Ctenolepisma longicaudata Esch., with
particular reference to its feeding habits. Proc. Roy. Soc. Victoria,
vol. 52, pp. 35-83, 12 text figs., 2 pls.

Lussen, H.

1907. Uber die innere Metamorphose der Trichopteren. Zool. Jahrb., Anat.,

vol. 24, pp. 71-128, 3 pls.
MAKEL, Marta.

1942. Metamorphose und Morphologie des Pseudococcus-Mannchens mit
besonderer Beriichsichtigung des Skelettmuskelsystems. Zool.
Jahrb., Anat., vol. 67, pp. 461-588, 24 figs.

Mansour, K.

1927. The development of the larval and adult mid-gut of Calandra oryzae
(Linn.): The rice weevil. Quart. Journ. Micr. Sci., vol. 71, pp.
313-352, 5 pls.

1934. The development of the adult mid-gut of coleopterous insects and
its bearing on systematics and embryology. Bull. Fac. Sci. Egyptian
Univ., No. 2, 34 pp., 5 pls.

MarcuHat, P.

1906. Recherches sur la biologie et le développement des Hymenopteéres
parasites. Les Platygasteres. Arch. Zool. Exp. Gén., ser. 4, vol. 4,
pp. 485-640, 8 pls.

Marratrt, C. L.
1923. The periodical cicada. U.S. Bur. Ent. Bull. 71, 183 pp., 68 figs.
MatTHeEw, A. P.

1948. Nutrition in the advanced embryo of the scorpion: Palamnaeus scaber

Thorell. Proc. Indian Acad. Sci., B, vol. 27, pp. 111-118, 6 figs.
MELANnpeR, A. L.

1903. Notes on the structure and development of Embia texana. Biol.

Bull., vol. 4, pp. 99-118, 6 figs.
Metts, A.

1935. Tisanotteri Italiana. Studio anatomo-morphologico e biologico del
Liothripide dell’ olivo (“Liothrips oleae’ Costa). Redia, vol. 21,
pp. 1-188, 44 text figs., 8 pls.

Mercer, W. F.

1900. The development of the wings in the Lepidoptera. Journ. New York

Ent. Soc., vol. 8, pp. 1-20, 5 pls.
Mittor, J.

1938. Le développement et la biologie larvaire des Oncodidés (Cyrtidés),
diptéres parasites d’araignés. Bull. Soc. Zool. France, vol. 63, pp.
162-181, 183-197, 14 figs.

NO. 9 INSECT METAMORPHOSIS—SNODGRASS I17

Mosusz, A.
1897. Ueber den Darmkanal der Anthrenus-Larve, nebst Bemerkungen
zur Epithelregeneration. Arch. Naturg., Jahrg. 63, vol. 1, pp. 89-
128, 3 pls.
Mutter, K.
1927. Beitrage zur Biologie, Anatomie, Histologie und inneren Meta-
morphose der Thrips-Larven. Zeitschr. wiss. Zool., vol. 130, pp.
251-303, 27 text figs., 1 pl.
MuwnscuHEr, L111.
1933. Die Metamorphose des Labiums der Odonaten. Zeitschr. wiss. Zool.,
vol. 143, pp. 201-240, 44 figs.
Murpuy, HEten EF.
1922. The metamorphosis of may-fly mouth parts. Bull. Lloyd Library,
No. 22, Ent. Ser. No. 2, 39 pp., 6 pls.
Murray, FLorence V., AnD Tiecs, O. W.
1935. The metamorphosis of Calandra oryzae. Quart. Journ. Micr. Sci.,
vol. 77, pp. 405-495, 20 text figs., 5 pls.
NABeErt, A.
1913. Die Corpora allata der Insekten. Zeitschr. wiss. Zool., vol. 104, pp.
181-358, 8 text figs., 5 pls.
NEEDHAM, J. G.
1911. Descriptions of dragonfly nymphs of the subfamily Calopteryginae.
Ent. News, vol. 22, pp. 145-154, 2 pls.

NELson, J. A.
1915. The embryology of the honey bee, 282 pp., 95 text figs. 6 pls.
Princeton.

Nesarrr, 4... Ji
1941. A comparative morphological study of the nervous system of the
Orthoptera and related orders. Ann. Ent. Soc. Amer., vol. 34, pp.

51-81, 8 pls.
OERTEL, E.
1930. Metamorphosis of the honeybee. Journ. Morph., vol. 50, pp. 205-339,
4 pls.
Pao, G.

1938. Note sulla biologia e sulla filogenesi dei Meloidi (Coleoptera). Mem.

Soc. Ent. Italia, vol. 16, pp. 71-96, 6 figs.
Parker, H. L., anp SmitH, H. D.

1933. Additional notes on the strepsipteron Eoxenos laboulbenei Peyerim-
hoff. Ann. Ent. Soc. Amer., vol. 26, pp. 217-233, 2 pls.

1934. Further notes on Eoxenos laboulbenei Peyerimhoff with a descrip-
tion of the male. Ann. Ent. Soc. Amer., vol. 27, pp. 468-477.

PARKER, J. B., AND Bovine, A. G.

1925. The blister beetle, Tricrania sanguinipennis—Biology, description of
different stages and systematic relationship. Proc. U. S. Nat. Mus.,
vol. 64, 40 pp., 5 pls.

Patay, R.

1939. Contribution a l’étude d’un Coleoptére [Leptinotarsa decimlineata
(Say)1]. Evolution des organes au cours du développement. 145 pp.,
59 figs. Rennes.

128 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

Pautt, M. E.

1927. Die Entwicklung geschniirter und zentrifugierter Eier von Calliphora
vomitoria und Musca domestica. Zeitschr. wiss. Zool., vol. 129, pp.
483-540, 64 figs.

PErEz, C.

1901. Histolyse des tubes de Malpighi et des glandes sericigénes chez la
fourmi rousse. Bull. Soc. Ent. France for 1901, pp. 307-300.

1910. Recherches histologiques sur la métamorphose des Muscides, Cal-
liphora erythrocephala Mg. Arch. Zool. Exp. Gén., ser. 5, vol. 4,
pp. I-274, 62 text figs., 16 pls.

1912. Observations sur l’histolyse et l’histogénése dans la métamorphose
des Vespides (Polistes gallica L.). Acad. Roy. Belgique, Mém., ser.
2, vol. 3, IOI pp., 10 pls.

PERKINS) Gale,

1905. Leaf-hoppers and their natural enemies (Epipyropidae). Bull. Hawaii
Sugar Planters’ Assoc. Exp. Stat., vol. 1, pt. 2, pp. 75-85.

1918. The assembling and pairing of Stylops. Ent. Month. Mag., vol. 54,
pp. I2I-I01.

PETERSON, A.
1951. Larvae of insects, Part II, 416 pp., 104 pls. Columbus, Ohio.
PFEIFFER, ISABELLE WEED.

1939. Experimental study of the function of the corpora allata in the grass-
hopper, Melanoplus differantialis. Journ. Exp. Zool., vol. 82, pp.
439-401.

1942. Suppression of metamorphosis in the grasshopper, Melanoplus differ-
entialis. Anat. Rec., vol. 84, p. 486.

1945. Effect of the corpora allata on the metabolism of adult female grass-
hoppers. Journ. Exp. Zool., vol. 99, pp. 183-233.

PFLUGFELDER, O.

1937. Bau, Entwicklung und Funktion der Corpora allata und cardiaca von
Dixippus morosus Br. Zeitschr. wiss. Zool., vol. 149, pp. 477-512.

1938. Weitere experimentelle Untersuchungen tiber die Funktion der Cor-
pora allata von Dixrippus morosus Br. Zeitschr. wiss. Zool., vol.
I5I, pp. 149-191, 30 figs.

1947. Uber die Ventraldriisen und einige andere inkretorische Organe des
Insektenkopfes. Biol. Zentralbl., vol. 66, pp. 211-235, 32 figs.

Picktes, A.

1931. On the metamorphosis of the alimentary canal in certain Ephemerop-

tera. Trans. Roy. Ent. Soc. London, vol. 79, pp. 263-279, 2 pls.
Possompés, B.

1950. Role de cerveau au cours de la métamorphose de Calliphora erythro-

cephala Meig. C. R. Acad. Sci., Paris, vol. 231, pp. 594-590.
Poutson, D. F.

1950. Development of the ring gland in Drosophila. In Demerec, Biology

of Drosophila, pp. 234-238.
PoyarkorF, E.

1910. Recherches histologiques sur la métamorphose d’um Coléoptére (La
galéruque de l’orme). Arch. Anat. Micr., vol. 12, pp. 333-474, 69
figs.

1914. Essai d’une théorie de la nymphe des insectes holomelaboles. Arch.
Zool. Exp. Gén., vol. 54, pp. 221-265.

NO. 9 INSECT METAMORPHOSIS—-SNODGRASS IIQ

PRIESNER, H.
1926. Die Thysanopteren Europas, 755 pp., 5 pls. Wien.
Pryor, M. G. M.

1951. On the abdominal appendages of larvae of Trichoptera, Neuroptera,
and Lepidoptera, and the origins of jointed limbs. Quart. Journ.
Micr. Sci., vol. 92, pp. 351-376, 30 figs.

Reiru, F.

1925. Die Entwicklung des Musca-Eies nach Ausschaltung verschiedener

Bibereiche. Zeitschr. wiss. Zool., vol. 126, pp. 181-238, 39 figs.
RENGEL, C.

1897. Uber die Veranderung des Darmepithels bei Tenebrio molitor wah-
rend der Metamorphose. Zeitschr. wiss. Zool., vol. 62, pp. 1-60, 1 pl.

1898. Uber die periodische Abstossung und Neubildung des gesammten Mit-
teldarmepithels bei Hydrophilus, Hydrous und Hydrobius. Zeitschr.
wiss. Zool., vol. 63, pp. 440-455, I pl.

Ricuarps, A. G.
1951. The integument of arthropods, 411 pp., 65 figs. Minneapolis.
RILey, C. V.

1876. On the larval characters and habits of the blister-beetles belonging
to the genera Macrobasis Lec. and Epicauta Fabr., with remarks
on other species of the family Meloidae. Trans. Acad. Sci. St. Louis,
vol. 3, pp. 544-562.

RoBErRTSON, C. W.

1936. The metamorphosis of Drosophila melanogaster, including an accu-
rately timed account of the principal morphological changes. Journ.
Morph., vol. 59, pp. 351-399, 4 pls.

Roeper, K. D.
1953. Insect physiology, 1100 pp., 257 figs. New York and London.
Russ BAR.

1908. Die postembryonale Entwicklung des Darmkanals bei den Trichop-
teren (Anabolia laevis Zett.). Zool. Jahrb., Anat., vol. 25, pp. 675-
770, 4 pls.

SAMSON, KATHARINA.

1908. Uber das Verhalten der Vasa Malpighii und die excretorische Funk-
tion der Fettzellen wahrend der Metamorphose von Heterogenea
limacodes Hufn. Zool. Jahrb., Anat., vol. 26, pp. 403-422, 2 pls.

SCHARRER, BERTA.

1946a. The role of the corpora allata in the development of Leucophaea
maderae (Orthoptera). Endocrinology, vol. 38, pp. 35-45, 4 figs.

1946b. The relationship between corpora allata and reproductive organs in
adult Leucophaea maderae (Orthoptera). Endocrinology, vol. 38,
pp. 46-55, 4 figs.

SCHARRER, BERTA, AND SCHARRER, E.

1944. Neurosecretion. VI. Comparison between the intercerebralis-cardia-
cum-allatum system of the insects and the hypothalamo-hypophyseal
system of the vertebrates. Biol. Bull., vol. 87, pp. 242-251.

ScHeEnr, S? CG:

1943. The biology of Hadronotus ajax Girault (Hymenoptera-Scelionidae),
a parasite in the eggs of the squash-bug (Anasa tristis De Geer).
Ann. Ent. Soc. Amer., vol. 36, pp. 625-635, 2 pls.

120 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

ScuHMIpT, ELsE.

1929. Die Histolyse und Histogenese der Muskulatur von Psychoda alter-

nata Say. Zeitschr. Morph. Okol. Tiere, vol. 13, pp. 117-143, 8 figs.
ScCHOENEMUND, E.

1912. Zur Biologie und Morphologie einiger Perla-Arten. Zool. Jahrb.,

Anat., vol. 34, pp. 1-56.
SCHRADER, SALLY H.

1924. Reproduction in Acroschismus wheeleri Pierce. Journ. Morph., vol.

39, PP. 157-205, 4 pls.
SILVEsTRI, F.

1904. Contribuzione alla conoscenze della metamorphosi e dei costumi della
Lebia scapularis Fourc. Redia, vol. 2, pp. 68-84, 5 pls.

1920. Contribuzione alla conoscenze dei Termitidi e Termitofili dell’Africa
Occidentale. II. Termitofili. Boll. Lab. Zool. Gen. e Agr. R. Scuola
Sup. Agr. in Portici, vol. 14, pp. 265-319.

1940. Studi sugli “Strepsiptera.” (Insecta.) II. Descrizione, biologia e
sviluppo postembryonale dell’ Halictophagus tettigometrae Silv.
Boll. Lab. Zool. Gen. e Agr. Fac. Agr. Portici, vol. 32, pp. 11-48,
34 figs.

Smiru, H. S.

1912. The chalcidoid genus Perilampus and its relation to the problem of
parasite introduction. U. S. Bur. Ent., Techn. Ser., Bull. 19, pt. 4,
pp. 33-60, 8 figs.

1917. The habit of leaf-oviposition among the parasitic Hymenoptera.
Psyche, vol. 24, pp. 63-68, 4 figs.

SmitH, Lucy W.

1913. The biology of Perla immarginata Say. Ann. Ent. Soc. Amer., vol. 6,

PP. 203-212.
SMRECZYNSKI, S.

1938. Entwicklungsmechanische Untersuchungen am Ei des Kafers Agelas-

tica alni L. Zool. Jahrb., Zool. Physiol., vol. 59, pp. 1-58, 21 figs.
Snoperass, R. E.

1931. Morphology of the insect abdomen. Part I. General structure of the
abdomen and its appendages. Smithsonian Misc. Coll., vol. 8s,
No. 6, 128 pp., 46 figs.

1935. The history of an insect’s stomach. Ann. Rep. Smithsonian Inst. for
1933, Pp. 363-387, 15 figs.

1953. The metamorphosis of a fly’s head. Smithsonian Misc. Coll., vol. 122,
No. 3, 25 pp., 7 figs.

SoROKINA-AGAFONOWA, MarIE.

1924. Das Verhalten des peripheren Nervensystems der Insekten in der
Metamorphose. Zeitschr. Anat. und Entwickl., vol. 74, pp. 318-337,
2 pls.

Spretu, H. T.

1933. The phylogeny of some mayfly genera. Journ. New York Ent. Soc.,

vol. 41, pp. 55-86, 327-390, 14 pls.
STEIN, F.

1838. Entwicklungs-Geschichte mehrerer Insectengattungen aus der Ord-

nung Neuropteren. Arch. Naturg., vol. 4, pp. 315-333, I pl.

NO. 9 INSECT METAMORPHOSIS—SNODGRASS I21I

Suter, P.

1932. Untersuchungen iiber Korperbau, Entwicklungsgang und Rassendif-
ferenzierung der Kommaschildlaus, Lepidosaphes ulmi L. Mitt.
Schweiz. ent. Ges., vol. 15, pp. 347-420, 64 figs.

SweEETMAN, H. L., AnD WHITTEMoRE, F. W.

1937. The number of molts of the firebrat (Lepismatidae, Thysanura).

Bull. Brooklyn Ent. Soc., vol. 32, pp. 117-120.
Terre, L.

1899. Contribution a l’étude de I’histolyse et de l’histogénese du tissue mus-

culaire chez l’abeille. C. R. Soc. Biol. Paris, vol. 51, pp. 896-808.
THOMSEN, ELLEN.

1940. Relation between corpus allatum and ovaries in adult flies (Muscidae).
Nature, vol. 145, pp. 28-20.

1942. An experimental and anatomical study of the corpus allatum in the
blow-fly Calliphora erythrocephala Meig. Vidensk. Medd. Dansk
Naturhist. Foren., vol. 106, pp. 317-405, 4 text figs., 12 pls.

1948. Effect of removal of neurosecretory cells in the brain of adult Cal-
liphora erythrocephala Meig. Nature, vol. 161, pp. 439-440.

THomMsEN, M.

1943. Effect of corpus cardiacum and other insect organs on the color-
change of the shrimp, Leander adspersus. K. Danske Vidensk.
Selsk. Biol. Medd., vol. 19, No. 4, 38 pp., 4 pls.

1951. Weismann’s ring and related organs in larvae of Diptera. K. Danske
Vidensk. Selsk. Biol. Skr., vol. 6, No. 5, 32 pp., 14 pls.

Tuorpe, W. H.

1931. The biology, post-embryonic development and economic importance
of Cryptochaetum iceryae (Diptera, Agromyzidae) parasitic on
Icerya purchasi (Coccidae, Monophlebinae). Proc. Zool. Soc. Lon-
don for 1931, pp. 927-971.

1938. Further experiments on olfactory conditioning in a parasitic insect.
The nature of the conditioning process. Proc. Roy. Soc. London,
B, vol. 126, pp. 370-397.

1939. Further studies on pre-imaginal olfactory conditioning in insects.
Proc. Roy. Soc. London, B, vol. 127, pp. 424-433.

1941. The biology of Cryptochaetum (Diptera) and Eupelmus (Hymenop-
tera) parasites of Aspidoproctus (Coccidae) in East Africa. Para-
sitology, vol. 33, pp. 149-168, 26 figs.

Tiecs, O. W.

1922. Researches on the insect metamorphosis. Part I. On the structure
and post-embryonic development of a chalcid wasp, Nasonia.
Part II. On the physiology and interpretation of the insect meta-
morphosis. Trans. and Proc. Roy. Soc. South Australia, vol. 46,
PP. 319-527, 16 pls.

TILLYARD, R. J.
1917. The biology of dragonflies, 396 pp., 188 text figs., 4 pls. Cambridge.
1926. The insects of Australia and New Zealand, 560 pp., 42 pls. Sydney.
TIMBERLAKE, P. H.

1912. Experimental parasitism: A study of the biology of Limnerium

validum (Cresson). U.S. Bur. Ent., Techn. Ser., Bull. 19, pt. 5,

Pp. 71-92, 10 figs.

I22 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

Tower, W. L.

1903. The origin and development of the wings of Coleoptera. Zool. Jahrb.,

Anat., vol. 17, pp. 517-572, 8 text figs., 7 pls.
Toyama, K.

1902. Contributions to the study of silk-worms, I. On the embryology of
the silk-worm. Bull. College Agr., Tokyo Imp. Univ., vol. 5, pp.
73-118.

Vacuon, M.

1938. Recherches anatomiques et biologiques sur la reproduction et la
développement des pseudoscorpions. Ann. Sci. Nat., ser. 11, Zool.,
vol. 1, pp. 1-207, 85 figs.

1950. Remarques préliminaires sur l’alimentation, les organes chélicériens,
le biberon et la tétine de l’embryon du scorpion: Ischnurus ochro-
pus C. L. Koch (Scorpionidae). Arch. Zool. Exp. Gén., vol. 86,
Pp. 137-156, 9 figs.

1953. The biology of scorpions. Endeavour, vol. 12, No. 46, pp. 80-80, 16
figs.

VAN DER Ktoot, W. G., AND WILLIAMS, C. M.

1953a. Cocoon construction by the Cecropia silkworm I. The role of the
external environment. Behaviour (Leiden), vol. 2, pp. 141-156, 8
figs.

1953b. Cocoon construction by the Cecropia silkworm II. The role of the
internal environment. Behaviour (Leiden), vol. 3, pp. 157-174,
12 figs.

VAN REEs, J.

1889. Beitrage zur Kenntnis der inneren Metamorphose von Musca vomi-

toria. Zool. Jahrb., Anat., vol. 3, pp. 1-134, 14 text figs., 2 pls.
VAYSSIERE, A.

1882. Recherches sur l’organization des larves des Ephémérines. Ann. Sci.
Nat., ser. 6, Zool., vol. 13, pp. 1-137, 11 pls.

1890. Monographie zoologique et anatomique du genre Prosopistoma, Ann.
Sci. Nat., ser. 7, Zool., vol. 9, pp. 19-87, 4 pls.

1934. Etude anatomique des larves nymphales des Bactisca obesa et caro-
lina. Ann. Sci. Nat., ser. 10, Zool., vol. 17, pp. 381-406, 2 pls.

VerRSON, E.

1905. Zur Entwicklung des Verdauungskanals bei Bombyx mori. Zeitschr.

wiss. Zool., vol. 82, pp. 523-600, 4 pls.
WapswortH, J. T.

1915. On the life-history of Aleochara bilineata Gryll., a staphylinid para-
site of Chortophila brassicae Bouché. Journ. Econ. Biol., vol. 10,
pp. I-27, 2 pls.

WAHL, B.

tool. Uber die Entwicklung der hypodermalen Imaginalscheiben im Thorax
und Abdomen der Larve von Eristalis Latr. Zeitschr. wiss. Zool.,
vol. 70, pp. I7I-I91, 4 text figs., 1 pl.

Weber, H.

1930. Biologie der Hemipteren, 543 pp., 329 figs. Berlin.

1931. Lebensweise und Umweltbeziehungen von Trialeurodes vaporrariorum
(Westwood) (Homoptera-Aleurodina). Zeitschr. Morph. Okol.
Tiere, vol. 23, pp. 575-753, 59 figs.

NO. 9 INSECT METAMORPHOSIS—SNODGRASS 123

1934. Die postembryonale Entwicklung der Aleurodinen (Hemiptera-
Homoptera). Zeitschr. Morph. Okol. Tierre, vol. 29, pp. 268-305,
14 figs.

WEISMANN, A.

1864. Die nachembryonale Entwicklung der Musciden nach Beobachtungen
an Murca vomitoria und Sarcophaga carnaria. Zeitschr. wiss. Zool.,
vol. 14, pp. 187-336, 7 pls.

WE cH, P. S.

1916. Contribution to the biology of certain aquatic Lepidoptera. Ann. Ent.

Soc. Amer., vol. 9, pp. 159-190, 3 pls.
Wuepovn, A. D.

1927. The structure and transformation of the labium of Anax junius.

Biol. Bull., vol. 53, pp. 286-301, 2 pls.
WIGGLESWorRTH, V. B.

1933. The physiology of the cuticle and of ecdysis in Rhodnius prolixus
(Triatomidae Hemiptera) ; with special reference to the function of
the oenocytes and the dermal glands. Quart. Journ. Micr. Sci., vol.
76, pp. 269-318, 15 figs.

1936. The function of the corpus allatum in the growth and reproduction
of Rhodnius prolixus (Hemiptera). Quart. Journ. Micr. Sci.,
vol. 79, pp. 91-121.

1940. The determination of characters at metamorphosis in Rhodnius pro-
lixus (Hemiptera). Journ. Exp. Biol., vol. 17, pp. 201-222, 12 text
figs., I pl.

1942. The significance of “chromatic droplets” in the growth of insects.
Quart. Journ. Micr. Sci., vol. 83, pp. 141-152, 3 figs.

1948. The functions of the corpus allatum in Rhodnius prolixus (Hemip-
tera). Journ. Exp. Biol., vol. 25, pp. I-14.

1951a. Hormones and the metamorphosis of insects. Endeavour, vol. 10,
No. 37, 5 pp., 14 figs. (Reprinted in Ann. Rep. Smithsonian Inst.
for 1951, pp. 313-318, 4 pls., 1952.)

1951b. Source of moulting hormone in Rhodnius. Nature, vol. 168, pp. 558-
559.

1952a. The thoracic gland in Rhodnius prolixus (Hemiptera) and its role
in moulting. Journ. Exp. Biol., vol. 20, pp. 561-570.

1952b. Hormone balance and the control of metamorphosis in Rhodnius
prolixus (Hemiptera). Journ. Exp. Biol., vol. 29, pp. 620-631.

1953a. Hormones and metamorphosis, with special reference to hemimeta-
bolic insects. Trans. IX Internat. Congr. Ent., Amsterdam, vol. 2,
Pp. 51-57.

1953b. The origin of sensory neurones in an insect, Rhodnius prolixus
(Hemiptera). Quart. Journ. Micr. Sci., vol. 94, pp. 93-112, 16 figs.

WituiAms, C. M.

1947. Physiology of insect diapause. II. Interaction between the pupal
brain and prothoracic glands in the metamorphosis of the giant
silkworm, Platysamia cecropia. Biol. Bull., vol. 93, pp. 89-98, 2 pls.

1948. Physiology of insect diapause. III. The prothoracic glands in the
Cecropia silkworm, with special reference to their significance in
embryonic and post-embryonic development. Biol. Bull., vol. 94,
pp. 60-65, 2 figs.

I24 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

1952. Physiology of insect diapause. IV. The brain and prothoracic glands
as an endocrine system in the Cecropia silkworm. Biol. Bull., vol.
103, pp. 120-138, 11 figs.
Wu, C. F.
1923. Morphology, anatomy, and ethology of Nemoura. Bull. Lloyd Li-
brary, No. 23, Ent. Ser. No. 3, 81 pp. 16 pls.

|| — SMITHSONIAN MISCELLANEOUS COLLECTIONS
i VOLUME 122, NUMBER 10

Charles D. and Mary Waux Talcott
Research Fund

| TWO SILICIFIED
|| CARBONIFEROUS. TRILOBITES
| _ FROM WEST TEXAS

(WirH 3 PLATES)

BY
HARRY B. WHITTINGTON

‘ : Museum of Comparative Zoology
Harvard University

Mel HONOR o/

ain NGTOW
ie (Pustication 4146)
Ri

CITY OF WASHINGTON
PUBLISHED BY THE SMITHSONIAN INSTITUTION
APRIL 22, 1954

SMITHSONIAN MISCELLANEOUS COLLECTIONS
VOLUME 122, NUMBER 10

Charles D. and flary Waux CHalcott
Research Fund

wo SICICIF IED
ea BONIFEROUS TRILOBITES
FROM WEST TEXAS

(Wir 3 PratEs)

BY
HARRY B. WHITTINGTON

Museum of Comparative Zoology
Harvard University

soe ose)
APRONS
iss
INGT TON

(Pustication 4146)

CITY OF WASHINGTON
PUBLISHED BY THE SMITHSONIAN INSTITUTION
APRIL 22, 1954

The Lord Baltimore Dress

BALTIMORE, MD., U. S. Ac

Charles BD. and Mary Waux Walcott Research Fund

TWO SILICIFIED CARBONIFEROUS TRILOBITES
FROM WEST.TEXAS

By HARRY B. WHITTINGTON
Museum of Comparative Zoology
Harvard University

(With THREE PLaTEs)

The specimens of silicified trilobites described in the following pages
were collected and prepared by Dr. Arthur L. Bowsher, of the United
States National Museum (hereafter abbreviated as U.S.N.M.). Iam
indebted to Dr. Bowsher for suggesting that I study this material and
to Dr. G. Arthur Cooper for permitting the loan of it to me. All the
specimens are in the National Museum collections and are from the
following localities :

U.S.N.M. locality 3070.—Helms formation, El Paso quadrangle,
Hueco Mountains, Tex., 24 miles west of Powwow Tanks, latitude
approximately 31°50'17” N., longitude 106°04’40” W. This locality
is stop 13 (p. 40), West Texas Geological Society Guidebook, Field
Trip No. 5, November 1949, and stop 1 on the map accompany-
ing West Texas Geological Society Field Trip of May-June 1946.
No. 3070-2 is from a limestone thought to be the same as bed 9, sec-
tion “C” of 1946 Field Trip Guidebook, and No. 3070-4 is from a
limestone thought to be the same as bed 11 of the same section.

U.S.N.M. locality 3069.—Helms formation, El Paso quadrangle,
Hueco Mountains, Tex., 1.1 miles west of Powwow Tanks, latitude
approximately 31°50'17” N., longitude 106°03'38”W. No. 3069-2 is
from about 10 feet above the base of the Helms in the saddle, from an
oolitic limestone lens, and approximately equivalent to the horizon
of No. 3070-2. No. 3069-4 is from about 25-30 feet above the base
of the Helms in the saddle, from an oolitic limestone with Archimedes,
and approximately equivalent to the horizon of No. 3070-4.

The numbers of these localities are used in subsequent references
to the specimens. The Helms formation in west Texas and adjacent
New Mexico has been described briefly by Laudon and Bowsher
(1949, pp. 19-20, 31-34), and the term is used here in the restricted

SMITHSONIAN MISCELLANEOUS COLLECTIONS, VOL. 122, NO. 10

2 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

sense of these authors. The Helms formation is stated by Laudon
and Bowsher to be of Chester (Upper Mississippian) age, and it is
of interest that the commonest trilobite in the formation, described
here as Paladin (Paladin) helmsensis, new species, is very much like
the type species of the genus from the Morrow Series (Lower Penn-
sylvanian) of Oklahoma.

The silicified specimens show the morphology of the exoskeleton in
unusual detail and perfection; hence in the next section significant
features of morphology and development are described and discussed,
and these comments are not repeated in the ensuing detailed descrip-
tions. The terminology employed follows that of previous papers
(Whittington, 1950, p. 533), except that I have used pleural region
of the pygidium rather than pleural lobes or side lobes of Warburg
(1925), and interpleural grooves rather than furrows. The additional
terms used in describing the articulation of the thorax, and the hypo-
stome, are explained on plates 2 and 3.

In order to avoid ambiguity in the terms “length” and “breadth” in
descriptions, I have used (in the abbreviated form indicated in paren-
theses) sagittal (sag.) to describe a measurement in the median line;
exsagittal (exs.), parallel to, but outside of, the median line; and
transverse (tr.), at right angles to the median line.

MORPHOLOGY AND DEVELOPMENT OF THE SILICIFIED SPECIES

An unusual feature of the silicified exoskeletons is the relatively
great thickness, as compared, for example, to those of silicified Ordo-
vician trilobites. I consider the thickness to be original and not a
result of the process of silicification. Plate 2, figures 5 and 6, and plate
3, figures 3, 5, and 6, show the thickness of the exoskeleton at the
suture lines and along selected sections. The doublure of both py-
gidium and cephalon is thicker than the immediately overlying dorsal
exoskeleton (pl. 3, figs. 3, 5), nowhere more so than at, and adjacent
to, the rostrum. The inner part of the thoracic pleurae is also thick,
at a maximum at the posterior edge, the inner surface flat and sloping
forward to the much thinner anterior edge. The thickness is such
that there is no ridge on the inner surface corresponding to the pleural
furrows on the outer surface (pl. 3, figs. 10, 13).

Four pairs of glabellar furrows have been observed in some Car-
boniferous trilobites (e.g., Stubblefield, 1948, p. 99; R. and E. Richter,
1951, pl. 5). On the inner surface of the exoskeleton (pl. 3, fig. 2)
these furrows form inwardly projecting platforms with a well-defined
edge (cf. R. and E. Richter, 1951,.p. 225). These are areas of muscle

NO. IO TRILOBITES FROM WEST TEXAS—WHITTINGTON 3

attachment, as is also the thickened and projecting outer one-third of
the occipital furrow. On the outer surface (pl. 3, fig. 1) only the
first (basal) furrow appears as a depression, the second, third, and
fourth furrows as smooth areas, in larger specimens appearing as con-
spicuous dark patches, dark perhaps because the exoskeleton is thicker
here. The articulating furrows of the thoracic axis are slightly thick-
ened at the extremity, and presumably are areas of muscle attachment.
On the pygidial axis (pl. 3, fig. 5), however, the outer parts of the
ring furrows become shallower, and the ovate areas between them, ap-
pearing darker in some specimens, are areas of muscle attachment.
The eye surface (pl. 3, figs. 4, 6) is externally almost smooth, the
facets faintly convex. On the inner surface each circular facet is
strongly convex, and the facets are close-spaced and arranged in verti-
cal and diagonal rows. The course of the cephalic sutures is revealed
in detail (pl. 2, figs. 1, 5, 6; text fig. 1), and I am not aware of any
previous descriptions of the rostrum of a Carboniferous trilobite. The
edge of the exoskeleton at the sutures is thick and flat, and the hypo-
stome fits against both the posterior edge of the rostrum and the ad-
jacent inner edge of the doublure. The wing process (at the tip of
the large anterior wing) evidently rested in the conspicuous circular
pit in the anterior boss on the inner surface of the cranidium (pl. 3,
fig. 17). Thus the hypostome was attached to the rest of the cephalon
in the same manner as in calymenids, cheirurids, and other trilobites.
Articulation between the segments of the thorax and the cephalon
and pygidium is effected by a series of devices (see pl. 3, figs. 7-13,
I5, 16, and compare Whittington and Evitt, 1954, pp. 21-24). The
ring process is a large boss situated at the outer, posterior edge of the
axial ring, and fits into a ring socket at the anterior, outer edge. Above
the ring socket, in line with the axial furrow, is a tiny, round axial
process, which fits into the axial socket in the posterior edge of the
segment at the base of the ring process. A narrow (exs.) strip along
the anterior edge of the inner part of the pleura is defined by a shallow
furrow, and the leading edge is thin and bluntly rounded. It fits into
a groove in the thick posterior edge of the inner part of the pleura,
this groove being beneath the upper, outer margin of the pleura. This
“tongue and groove” articulation extends out to the fulcrum, where
it dies out, and there are no articulation processes and sockets at the
fulcrum. The posterior edge of the cephalon inside the branches of
the facial suture, and the anterior margin of the pygidium inside the
fulcra, are shaped like the corresponding edges of the thoracic seg-
ments. The outer parts of the thoracic pleurae, and the pygidium, are
faceted to facilitate overlap in enrollment. In the doublure of each

4 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

segment is a broad V-shaped notch (pl. 3, fig. 14), the Panderian
opening, and the anterior edge of this notch is raised (inwardly pro-
jecting ), and acts to limit the amount of overlap between segments.

Only what are probably the later meraspid stages of the develop-
ment are known (pls. 1, 2). The cranidium shows a general reduc-
tion in convexity with increasing size. The glabella in the smallest
specimens is almost parallel-sided, and with increasing size the lateral
expansion of the anterior lobe takes place, the posterior part widens
between the eye lobes, and the relative convexity of the posterocentral
glabellar region, and of the basal glabellar lobes, is reduced. The eye
lobe becomes relatively shorter. The pygidium shows a considerable
reduction in convexity with increasing size, and the shallow median
notch in the posterior margin of small specimens soon disappears.
The meraspid development of Ditomopyge was described by Weller
(1935), and the smallest cranidium, 1 mm. in length, has the sub-
parallel-sided glabella, long (sag.) anterior border and eye lobe seen
in Paladin. However, the glabellar lobation is absent, in contrast to
the presence of well-marked basal lobes and furrows in Paladin.
Small pygidia of Ditomopyge show a median notch in the posterior
border (Newell, 1931, pl. 31, fig. 31; Weller, 1935, p. 508) like that
seen in Paladin. While the development of the pygidium in the two
genera has some features in common, a notable difference is that in
Ditomopyge the pleural regions increase in convexity (Weller, 1935,
figs. 4c, 5c, 7c, 8c), in contrast to the decrease in Paladin.

The meraspid specimens of Paladin do not resemble any geologically
older adult Carboniferous trilobite, and Weller (1935, p. 513) like-
wise found that the meraspid specimens of Ditomopyge resembled no
known geologically older adult trilobite. One may take these observa-
tions as further evidence of the untruth of the so-called “law” of re-
capitulation, in the strict sense of Haeckel (cf. de Beer, 1951).

SYSTEMATIC DESCRIPTIONS
Family PROETIDAE (Hawle and Corda, 1847), Salter, 1864
Subfamily PHILLIPSIINAE (Oehlert, 1886), Pribyl, 1946

A characterization of this subfamily has recently been given by
Pribyl (1946, pp. 33-34). The present material of Paladin shows that
up to four pairs of glabellar furrows may be present. Few illustra-
tions have been published of phillipsiinid hypostomes, but those avail-
able (e.g., Woodward, 1883-1884; Weber, 1937) suggest that they
are similar to each other and like that of Paladin (pl. 1, figs. 29, 30,
35; pl. 2, figs. 21, 26, 27, 32, 33). Characteristic are the large anterior

NO. I0 TRILOBITES FROM WEST TEXAS—WHITTINGTON 5

wings, lack of distinct anterior border, narrow lateral, but wider
(sag.) posterior, border, and short (sag.), crescentic, inflated posterior
lobe of the middle body. This type of hypostome is not like known
examples of hypostomes (Pribyl, 1947, figs. 12-15, 17-19) of proetid
genera in other subfamilies, and may be typical of the Phillipsiinae.
The shape of the rostrum may equally well be characteristic of the
subfamily, but little information is available.

Genus PALADIN Weller, 1936

Type species —Griffithides morrowensis Mather, 1915, by original
designation of Weller, 1936, p. 707.

Discussion—The most abundant of the two species of silicified
trilobites described below has been compared with the holotype of
Paladin morrowensis, and belongs in this genus. The second species
differs from the first notably in the greater convexity of the cephalon
and pygidium, the shorter anterior cephalic border, and the outline of
the glabella, which is less expanded between the eye lobes but more
strongly expanded anteriorly. These relatively minor differences ally
it with Kaskia chesterensis (Weller, 1936, pp. 708-711, pl. 95, figs.
4a-6), the type of the genus Kaskia Weller, 1936. K. chesterensis has
an even shorter (sag.), steeper anterior border. Weller admitted
(1936, p. 708) that Paladin and Kaskia were closely similar, and that
there were species intermediate between typical species of the two
genera. The second silicified species here described is one of these
intermediates. In view of these facts, it seems to me preferable to
regard Kaskia as a subgenus of Paladin, with P. morrowensis repre-
senting the typical subgenus Paladin (Paladin), and this procedure
has been followed below.

Reed (1942, pp. 653, 660-667, pl. 10, figs. 4-5b, pl. 11, figs. I-5a;
1943, pp. 179-184, pl. 2, figs. 6, 7, pl. 3, figs. 1-8) considered that
the forms he referred to his genus Weberides included most of, if not
all, the originals of Woodward’s (1883-1884) plate 4, and were similar
to the Russian species described by Weber (1933, pp. 33-35, 37-41,
pl. 2, figs. 2-11, 17-33, 36-41, text figs. 14-17, 19-21 ; 1937, pp. 74-75,
pl. 8, figs. 31-34, 36, 39-44, 48) under the names Griffithides lutugini
and varieties and G. transilis and varieties. Weller (1936, pp. 707-
708), however, had previously placed these Russian species and va-
rieties in his genera Paladin and Kaskia. Reed recognized this (1943,
p. 180) but did not say how Weberides differed from Paladin. It
seems that some of the species referred to above may be congeneric,
and if so ought to be placed in Paladin. Before it is concluded that

6 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

W eberides Reed, 1942, is a synonym of Paladin, however, the diplo-
type of Weberides should be reexamined, for Reed (1942, p. 663)
admits that he did not see it. The specimen in question (original of
M’Coy, 1844, pl. 4, fig. 5) is a pygidium, with a short, blunt spine on
the posterior border.

The genus Ditomopyge Newell, 1931 (as emended by Weller, 1935)
is related to Paladin (Kaskia), as Weller pointed out (1936, p. 711).
The inflation of the central region of the glabella in front of the occi-
pital ring seen in P. (K.) rarus, new species, could give rise to the
preoccipital lobe of Ditomopyge. The free check of P. (K.) rarus,
new species, is much more like that of Ditomopyge than that of P.
(P.) helmsensis, new species, which lacks the flattened upper surface
of the border. Contrary to the opinion of Weller (1936, pp. 713-714),
I regard Ameura as related to Paladin. I have examined the holotype
of Ameura sangamonensis (Meek and Worthen, 1865) and the gla-
bella is only slightly wider between the eye lobes than across the an-
terior lobe. The basal glabellar lobes do have independent convexity.
The pygidium, of length (sag.) about equal to width, recalls the
original of Woodward’s (1883-1884) plate 4, fig. 9, and the elongated
appearance is distinctive.

The aforementioned four genera, together with Sevillia (Weller,
1935, p. 506, explanation of text fig. 9, nomen nudum; Weller, 1936)
and Linguaphillipsia Stubblefield, 1948, probably form a related group
ranging from Lower Carboniferous to Lower Permian in age, wide-
spread in North America and Eurasia.

PALADIN (PALADIN) MORROWENSIS (Mather, 1915)
Plate 1, figures 1-6, 9

Holotype-—Walker Museum No. 16174, incomplete cephalon, from
Brentwood limestone, Morrow Series, lower Pennsylvanian, Sawney
Hollow, head of Indian Creek, Okla., and 33 miles south of Evansville,
Ark.

Description—The holotype is refigured here, and the following
notes are added to supplement Mather’s (1915, pp. 244-246, pl. 16,
figs. 13, 13a) original description. Basal glabellar furrow deepest at
about the midlength, disappearing before reaching axial furrow. Ad-
ditional furrows not represented by depressions in outer surface. An-
terior branch of facial suture running at first outward at about 50° to
the sagittal line, then on the border, opposite the maximum width of
the anterior glabellar lobe, curving to run inward straight to the mar-
gin. The angle between the two sections is about 100°. The doublure

NO. 10 TRILOBITES FROM WEST TEXAS—-WHITTINGTON 7

of the cephalon is convex and slopes steeply laterally, but is flattened
and slopes gently anteriorly. The rostral suture runs close to the outer
edge, the connective sutures curve inward, and the hypostomal suture
runs in a curve convex forward. The rostrum is thus similar in outline
to that of P. (P.) helmsensis, new species.

The associated pygidium is also refigured, and the border is gently
convex, not concave as stated by Mather (1915, p. 245).

PALADIN (PALADIN) HELMSENSIS Whittington, new species
Plates 2, 3; text figure 1

Holotype-—U.S.N.M. No. 116513, cranidium, original of plate 2,
figures I, 2, 5, 6; locality 3070-2.

Paratypes—U.S.N.M. Nos. 116514a-h; free cheek, rostrum, and
hypostome from locality 3070-2; two segments from locality 3069-4 ;
two segments from locality 3070-4; pygidium from locality 3069-2.

Description—Dimensions of holotype in millimeters: Length (sag. )
7.0, height 2.7; length of glabella (sag.) 6.3, width across anterior
lobe 3.9, at third furrows 3.1, of occipital ring 3.9. Length of para-
type pygidium (sag.) 6.3, width 7.8, height 2.8. Cephalon subsemicir-
cular in outline, gently convex. Glabella gently convex (sag. and tr.),
outlined by shallow axial and preglabellar furrows ; narrowing slightly
immediately in front of the occipital ring, expanding between the eye
lobes, then narrowing again forward to the minimum width opposite
the third furrows, and then expanding forward again until width
across anterior lobe is the same as, or slightly greater than, that of oc-
cipital ring. Latter moderately convex, highest point near posterior
margin, from which it slopes down to the shallow, sinuous occipital
furrow ; faint median tubercle. Four pairs of glabellar furrows (pl.
3, figs. 1, 2), the first (basal) appearing as shallow depressions, gently
curved, directed inward and backward to isolate triangular, gently con-
vex basal lobes. The basal furrows are deepest at midlength, becoming
poorly defined at the outer extremity, faint at the inner ends as they
meet the occipital furrow. The maximum width of the basal lobes is
one-third the glabellar width in front of the occipital furrow. Between
the basal lobes the central glabellar region is slightly inflated and pos-
teriorly slopes steeply. The second and third glabellar furrows are
progressively shorter and directed less strongly backward, the fourth
short, ovate, directed slightly forward and commencing a short dis-
tance inside the axial furrows. Cheeks sloping gently outward and
forward, with a broad (tr.) lateral border defined by the slight
change in slope at the faint border furrow, the anterior border nar-

8 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

rower (sag.). Posterior border defined by a deep border furrow, and
with independent convexity. Genal spine broad at base, relatively
long. Eye lobe large, length (exs.) more than one-third that of
cephalon, situated with the anterior edge about opposite the glabellar
midpoint, and close to the axial furrow, the highest point lower than
the glabellar midline between the eye lobes. Palpebral lobe without
rim, outer part horizontal, inner part sloping down to axial furrow.
Eye surface (pl. 3, figs. 4, 6) with numerous small, gently convex
facets. Anterior branch of suture runs straight outward and forward
from the eye lobe onto the border, then curves and runs straight in-
ward and forward to reach the anterior margin at a point in line
(exs.) with the inner margin of the eye lobe. The posterior branch
runs outward and backward to the border furrow, then curves, at first
more strongly outward, over the posterior border to reach the margin
just inside the base of the genal spine. Doublure laterally of less width
(tr.) than the border, gently convex and sloping steeply outward. An-
teriorly doublure becomes flattened, horizontal, and narrower (sag.).
The rostral suture runs along the outer edge of the doublure, the con-
nective suture in a curve convex outward. The anterior and posterior
margins of the rostrum are thus forwardly curved, the lateral margins
outwardly so. The rostrum (pl. 2, figs. 40-42) is also thickest along
a line midway between the anterior and posterior margins, so that
while the outer surface is flat, the inner is convex. The doublure of
the free cheek adjacent to the rostrum shows a corresponding thicken-
ing, which fades out laterally. Certain features displayed by the inner
surface of the cephalon have been discussed above. Plate 3, figure 2,
shows the doublure of the occipital ring. In the inner edge of the
doublure of the free cheek (pl. 3, fig. 6) is a shallow notch, in line
with the posterior border. I interpret this notch as the Panderian open-
ing, and as corresponding with the larger notches in the thoracic pleu-
ral doublures.

Length of hypostome (pl. 2, figs. 21, 26, 27, 32, 33) (sag.) slightly
greater than maximum width across anterior wings. Middle body
gently convex longitudinally, more strongly so transversely, not de-
fined anteriorly or separated from the anterior wings by a furrow,
but laterally and posteriorly outlined by the change in slope at the
borders. The crescentic posterior lobe, the tips at about two-thirds the
length of the middle body and opposite the lateral shoulders, has a
faint independent convexity, most marked at the tips. The anterior
sutural edge of the hypostome is thick, extending between the bases
of the wings, and fits against both the inner edge of the rostrum and
the doublure of the free cheeks (text fig. 1). The anterior wings are

NO. I0 TRILOBITES FROM WEST TEXAS—WHITTINGTON 9

broad (exs.) at the base, slope steeply upward, with a small articulat-
ing boss at the outer, anterior corner. The lateral borders narrow,
gently convex, shoulder well marked, posterior border broader, mar-
gin sinuous, posterolateral corners rounded. The interior view shows
that the doublure is narrow along the lateral borders, wider along the
posterior border, and the furrow dividing the middle body more evi-
dent. In lateral view the notch between shoulder and anterior wing
is seen, and posterior wings seem not to be developed.

Fic. 1.—Paladin (Paladin) helmsensis, new species. Outline reconstruction of the
exoskeleton of the cephalon in ventral view, approximately x 7.

Number of thoracic segments unknown. Axis moderately convex,
each ring subdivided into a short (sag.) anterior part that disappears
laterally, and a longer (sag. and exs.) posterior part. The articulating
furrow narrow and deep, the half-ring short (sag. and exs.). Inner
part of pleura horizontal, outer part bent steeply down, faceted, the
facet of the anterior segments (pl. 3, figs. 10-13) abruptly cutting off
the narrow (tr.) outer pleural part. The narrowness of these latter
enables these segments to fit between and under the genal spines of
the cephalon. Succeeding segments (pl. 3, figs. 7-9, 15, 16) have the
outer pleural parts wider (tr.). Pleural furrow narrow and deep,
situated at about half the length (exs.) at the fulcrum, and extending
to the inner edge of the facet. The interior view (pl. 3, fig. 10) shows

Io SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

the doublure of the axial ring and the low ridge, the area of muscle
attachment, formed by the outer part of the articulating furrow. The
devices which facilitate articulation between the segments have been
described above. In the doublure of the outer pleural part (pl. 3, fig.
14) is the broad notch of the Panderian opening. The doublure in
front of, and outside, this notch is gently convex.

Pygidium moderately convex, axis moderately convex and gently
tapering. In largest specimens 17 ring furrows, the inner part straight,
deep, the outer shallow, turning slightly back. Inner, anterior part of
pleural region horizontal, outer part gently convex, steeply sloping,
border distinctly separated by change in convexity, and sloping out-
ward. Ten deep pleural furrows in largest specimens, progressively
more strongly backwardly directed, ending at inner edge of border.
Interpleural grooves faint, sometimes absent, sometimes first five visi-
ble, not extending on to border. Doublure of same width as border,
inner part bent steeply up.

External surface of glabella and palpebral lobes with shallow, ir-
regular pits (pl. 3, fig. 1), largest near the median line. Small tubercles
occur along the posterior edge of the occipital and axial rings. Raised
lines, parallel to each other and the margin, on the outer part of the
cephalic and pygidial borders and the outer surface of the doublures.
Hypostome with similar lines on the middle body and borders, and
tiny, shallow, scattered pits on the middle body.

Discussion.—Comparison of the cephalon of Paladin (Paladin)
morrowensis with the type of P. (P.) helmsensis shows that the latter
differs from the former principally in characters of the glabella. That
of P. (P.) helmsensis is less inflated (as seen in longitudinal profile),
has the basal glabellar lobes and posterior part of the central glabellar
region less inflated, and has the anterior lobe less expanded trans-
versely, though between the eye lobes the glabella of P. (P.) helmsen-
sis is more markedly expanded than that of P. (P.) morrowensis. The
external surface of the glabella and palpebral lobes is tuberculate in
P. (P.) morrowensis, pitted in P. (P.) helmsensis. The lateral ce-
phalic border of P. (P.) morrowensis slopes more steeply than that
of the Texas species. The pygidia of the two species (pl. 1, figs. 4-6;
pl. 2, figs. 9, 10, 14, 15) are similar, that of P. (P.) helmsensis being
distinguished by the axis showing more rings and being more inflated
posteriorly, and by the border being relatively broader (sag.) pos-
teriorly. The axial rings of P. (P.) morrowensis are apparently with-
out the row of tubercles on the posterior margin. Evidently P. (P.)
helmsensis and P. (P.) morrowensis are closely related species, though
they differ considerably in age.

NO. 10 TRILOBITES FROM WEST TEXAS—-WHITTINGTON II

DEVELOPMENT OF Paladin (Paladin) helmsensis, NEW SPECIES

Cranidium.—Length of smallest cranidium (pl. 2, figs. 34-36)
(sag.) 1.5 mm., glabella narrowest between the anterior end of the
eye lobes, but since it lacks the anterior and posterior expansions of
larger forms it appears almost parallel-sided. Basal glabellar furrows
deep and broad, so that the basal lobes are prominent, and the posterior
part of the central glabellar region is quite strongly inflated. The sec-
ond and third glabellar furrows are ill-defined patches on the exo-
skeleton. Length of anterior border of the cranidium (sag.) about one-
eighth that of the glabella. Length of palpebral lobes (exs.) more than
one-third that of cranidium. In cranidia of increasing size that part
of the glabella in front of the third furrow becomes relatively wider
(compare figs. I and 34, pl. 2). The palpebral lobes become relatively
smaller, the length (exs.) being reduced to less than one-third that of
the cranidium. The basal glabellar furrows become shallower, and the
convexity of the basal lobes and posterior part of the central glabellar
region is reduced. Small cranidia with close-spaced tubercles on the
glabella and palpebral lobes, the tubercles on the frontomedian glabel-
lar lobe close-spaced and arranged in lines subparallel to the anterior
margin. With increasing size of the cranidium these tubercles become
less prominent, and in the largest cranidia only the reticulate pattern
of pits remains.

The smallest hypostome known (pl. 2, figs. 39, 44) is little different
from the largest—the shoulders are rather more prominent, and the
tips of the crescentic posterior lobe of the middle body are more
strongly inflated. The smallest pygidium known (pl. 2, figs. 37, 38,
43) is 1.6 mm. in length (sag.), 2.2 mm. in width. Axis of I5 rings.
Pleural region convex, inner, anterior part horizontal, outer part
steeply sloping, the border sloping outward but less steeply. Eleven
pleural furrows visible, terminating at the inner margin of the border.
First three interpleural grooves shallow, situated close to the succeed-
ing pleural furrows, and extending on to the inner part of the border.
Border broad (sag.) posterolaterally, narrow (tr.) anterolaterally,
with a shallow median notch in the posterior margin. With increasing
size the pygidium maintains about the same ratio between length and
width, and the original of plate 2, figures 30, 31, is 2.2 mm. in length
(sag.), 3.3 mm. in width. The convexity of the pleural regions is
markedly reduced, the notch in the posterior margin disappears and
the difference in the width of the border laterally and posteriorly is
reduced. Ina pygidium (sag.) 3.2 mm. long only the first interpleural
groove is visible. The tubercles on the median part of the axial rings
are visible in this and larger specimens.

I2 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

PALADIN (KASKIA) RARUS Whittington, new species
Plate 1, figures 7, 8, 10-35

Holotype-——U.S.N.M. No. 116511, cranidium, original of plate 1,
figure 7, 8, 10, 11, locality 3070-4.

Paratypes—U.S.N.M. Nos. 116512a-c, free cheek and pygidium
from locality 3070-2, hypostome from locality 3069-4.

Description.—Length of holotype cranidium 5.6 mm., height 3.2
mm. ; length (sag.) of glabella 5.1 mm., width across anterior lobe 3.6
mm., between anterior ends of palpebral lobes 2.9 mm., of occipital
ring 3.2 mm.

The cephalon of this species is similar to that of Paladin (Paladin)
helmsensis, new species, but is distinguished by (1) the greater con-
vexity ; (2) the glabella being slightly expanded between the eye lobes,
but more strongly expanded across the anterior lobe; (3) the sharper
angle in the course of the anterior branch of the facial suture on the
border (compare the antero-lateral margins of the cranidia in pl. 1,
fig. 7, and pl. 2, fig. 1) ; (4) the relatively shorter (sag. and exs.) an-
terior border ; (5) the much greater change in slope at the border fur-
row of the free cheek, resulting from the inner part of the border
being flattened. Additional ways in which the cephalon of P. (K.)
rarus differs from that of P. (P.) helmsensis are: (6) the basal gla-
bellar furrows are deeper, the basal lobes more inflated; (7) the pal-
pebral lobes are narrower (tr.) ; (8) the middle body of the hypo-
stome is more convex, with deeper middle furrows, and tiny maculae
are present. There is a sharper angle in the anterior margin between
where the hypostome fits against the rostrum and the doublure of the
free cheek, the shoulders are more prominent, and the posterior bor-
der has the three blunt spines; (9) the external surface of the gla-
bella and palpebral lobes is tuberculate rather than pitted.

Rostrum and thorax unknown.

Length of paratype pygidium (sag.) 5.0 mm., width 6.9 mm.,
height 3.0 mm. This pygidium is distinguished from that of P. (P.)
helmsensis by the greater convexity and consequent height. Both the
axis and the pleural regions inside the border are more convex in P.
(K.) rarus, and the border slopes more steeply outward. The number
of axial rings and pleural furrows is the same in the two species, but
the ribs between the furrows in P. (K.) rarus are much more convex.

Discussion.—Paladin (Kaskia) rarus is distinguished from the type
species P. (K.) chesterensis (Weller, 1936, pp. 708-711, pl. 95, figs.
4a-6), also of Chester age, by the less steep slope of the anterior part
of the glabella and the longer (sag.) projecting anterior border (com-

NO. IO TRILOBITES FROM WEST TEXAS—WHITTINGTON 13

pare pl. 1, fig. 10, with Weller, 1936, pl. 95, fig. 4c). The pleural
regions of the pygidium of the Texas species appear to be more convex
than those of P. (K.) chesterensis. Four pairs of glabellar furrows
are present in P. (K.) rarus, but only three are described as present in
P. (K.) chesterensis.

Weller (1936, pp. 708-710) pointed out that forms intermediate
between the type species of Paladin (Paladin) and Paladin (Kaskia)
occur. In the outline and convexity of the glabella, P. (K.) rarus is
more like P. (P.) morrowensis than is P. (P.) helmsensis, a further
illustration of the close relationship between these species.

DEVELOPMENT OF Paladin (Kaskia) rarus, NEW SPECIES

The smallest cranidium (pl. 1, figs. 23, 24) is 3.2 mm. in length
(sag.). Compared with the largest cranidium it is more convex as a
whole, as well as considering the frontomedian and basal glabellar
lobes separately ; the glabella is less expanded anteriorly, and the pal-
pebral lobes are longer. The development thus parallels that of Paladin
(Paladin) helmsensis, with an expansion of the glabella anteriorly, a
general reduction in convexity, and decrease in size of the palpebral
lobes. The smallest pygidium (pl. 1, figs. 26-28) is 1.7 mm. in length
(sag.), strongly convex, the outer parts of the pleural regions over-
hanging the border. There are 13 or 14 axial rings, 10 pleural furrows,
no interpleural grooves. There is no median notch in the posterior
margin of the border. The chief change with increasing size of the
pygidium is the reduction in convexity, so that the outer parts of the
pleural regions slope steeply but do not overhang the border.

REFERENCES

DE BEER, G. R.
1951. Embryos and ancestors. Rev. ed. Oxford.
Laupon, L. R., and Bowsuer, A. L.
1949. Mississippian formations of southwestern New Mexico, Bull. Geol.
Soc. Amer., vol. 60, pp. 1-87, 44 figs.
Marue_r, K. F.
1915. The fauna of the Morrow group of Arkansas and Oklahoma. Bull.
Sci. Lab. Denison Univ., vol. 18, pp. 59-284, 16 pls., 5 figs.
M’Coy, F.
1844. A synopsis of the characters of the Carboniferous limestone fossils of
Ireland. viii+207 pp., 29 pls. Dublin.
NEwELL, N. D.
1931. New Schizophoriidae and a trilobite from the Kansas Pennsylvanian.
Journ. Paleont., vol. 5, pp. 260-269, 1 pl.

I4 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. I22

PrRisyL, A.

1946. Notes on the recognition of the Bohemian Proetidae (Trilobitae).
Bull. Int. Acad. Tchéque Sci., 46th year, No. 10, pp. 1-41, 4 pls.
I fig.

1947. Aulacopleura and the Otarionidae. Journ. Paleont., vol. 21, pp. 537-
545, I pl., 19 figs.

IREED, oH ok. |G:

1942. Some new Carboniferous trilobites. Ann. Mag. Nat. Hist., ser. 11, vol.
9, pp. 649-672, 4 pls.

1943. Some Carboniferous trilobites from Scotland. Ann. Mag. Nat. Hist.,
ser. II, vol. 10, pp. 176-186, 2 pls.

RicHTER, R., and RicHTErR, Emma.

1951. Der Beginn des Karbons im Wechsel der Trilobiten. Senckenbergiana,

vol. 32, Nos. 1-4, pp. 219-226, 5 pls., 10 figs.
STUBBLEFIELD, C. J.

1948. Carboniferous trilobites from Malaya. Appendix to “Malayan Lower
Carboniferous Fossils,” by H. M. Muir-Wood, pp. 1-118, 17 pls., 3
figs. British Museum (Natural History), London.

Warsurc, Exsa.

1925. The trilobites of the Leptaena limestone in Dalarne. Bull. Geol. Inst.

Uppsala, vol. 17, pp. 1-446, 11 pls.
WEBER, V.

1933. Trilobites of the Donetz Basin. Trans. United Geol. and Prosp. Serv.
U.S.S.R., fase. 255, pp. 1-95, 3 pls., 33 figs.

1937. Trilobites of the Carboniferous and Permian system of U.S.S.R.
Centr. Geol. and Prosp. Inst., Paleont. of U.S.S.R. Mon., vol. 71,
fasc. I, pp. I-160, 11 pls., 78 figs.

WELLER, J. M.

1935. Adolescent development of Ditomopyge. Journ. Paleont., vol. 9, pp.
503-513, 31 figs.

1936. Carboniferous trilobite genera. Journ. Paleont., vol. 10, pp. 704-714,
TE pl

WuittincTon, H. B.

1950. Sixteen Ordovician genotype trilobites. Journ. Paleont., vol. 24, pp.

531-565, 8 pls., 9 figs.
WuittincTon, H. B., and Evirr, W. R.

1954. Silicified Middle Ordovician trilobites. Geol. Soc. Amer., Mem. 59,

137 pp., 33 pls., 27 figs. (Dated Dec. 18, 1953.)
Woopwarp, H.

1883-1884. A monograph of the British Carboniferous trilobites, pp. 1-86,

10 pls., 8 figs. Palaeontographical Society, London.

NO. IO TRILOBITES FROM WEST TEXAS—WHITTINGTON

EXPLANATION OF PLATES
PLATE 1

Figs. 1-6, 9 —Paladin (Paladin) morrowensis (Mather)...............05
I, 2, 3, Dorsal stereograph, left lateral, and anterior views of holo-
type, Walker Museum No. 16174, X 3. 4, 5, 6, Dorsal stereograph,
posterior, and right lateral view of pygidium, Walker Museum No.
16174, X 3. 9, Ventral view of cephalic doublure of holotype, posi-
tion of left (right in picture) edge of rostrum dotted, * 74. Brent-
wood limestone, Morrow Series, lower Pennsylvanian, Sawney Hol-
low, head of Indian Creek, Okla., and 34 miles south of Evansville.,
Ark.

Fics. 7, 8, 10-35.—Paladin (Kaskia) rarus Whittington, new species.......
7, 8, 10, 11, Dorsal stereograph, interior, right lateral, and anterior
views of holotype cranidium, U.S.N.M. No. 116511, locality 3070-4,
X 3. 12, 13, 14, Dorsal stereograph, posterior, and left lateral views
of paratype pygidium, U.S.N.M. No. 116512a, locality 3070-2, < 3.
15, 18, Anterior view, dorsal stereograph of cranidium and free
cheek, locality 3070-2, X 3. 16, 17, Dorsal and posterior views of
pygidium, locality 3070-2, X 3. 19, Right lateral view of cranidium,
original of figures 15, 18, X 3. 20, 21, 22, Dorsal, posterior, and ven-
tral views of pygidium, locality 3070-2, & 3. 23, 24, Dorsal and right
lateral views of cranidium, locality 3069-2, X 3. 25, 31, Left lateral
and dorsal views of paratype free cheek, U.S.N.M. No. 116512b,
X 3. 32, Interior view of same, X I0, locality 3070-2. 26, 27, 28,
Dorsal, posterior, and left lateral views of pygidium, locality 3070-2
xX 4. 20, 30, 35, Ventral, interior, and left lateral views of paratype
hypostome, U.S.N.M. No. 116512¢, locality 3069-4, X 3. 33, 34, Ven-
tral and left lateral views of hypostome, locality 3069-2 *.3. Helms
formation, Chester Series, upper Mississippian, Hueco Mountains,
west Tex. Locality numbers are explained on page I.

PEA 2

Paladin (Paladin) helmsensis Whittington, new species............+0.000
I, 5, 6, Dorsal stereograph, anterior view, anterolateral stereograph
of holotype cranidium (U.S.N.M. No. 116513), and paratype free
cheek (U.S.N.M. No. 116514a), locality 3070-2, & 3. 2, Left lateral
view of holotype cranidium, U.S.N.M. No. 116513, locality 3070-2,
X 3. 3, 4, Dorsal and left lateral views of free cheek, locality 3069-4,
xX 3. 7, 8, Dorsal and right lateral views of cranidium, locality
3069-4, X 3. 9, 10, 14, 15, Dorsal stereograph, interior, posterior, and
right lateral views of paratype pygidium, U.S.N.M. No. 116514b,
locality 3069-2, X 3. 11, 12, Dorsal and right lateral views of cra-
nidium, locality 3069-4, * 3. 13, 20, Dorsal and posterior views of
pygidium, locality 3069-2, * 3. 16, 17, Dorsal and right lateral
views of cranidium, locality 3069-4, < 3. 18, 19, Dorsal and pos-
terior views of pygidium, locality 3069-4, 3. 21, 26, 27, Right
lateral, ventral, and interior views of paratype hypostome, U.S.N.M.
No. 116514c, locality 3070-2, X 3. S =shoulder; n= lateral notch.

16 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

Page

22, 23, Dorsal and right lateral views of cranidium, locality 3069-4,
X 3. 24, 25, Dorsal and posterior views of pygidium, locality 3069-4,
X 3. 28, 29, Dorsal and right lateral views of cranidium, locality
3069-4, X 74. 30, 31, Dorsal and posterior views of pygidium, lo-
cality 3069-4, X 3. 32, 33, Ventral and right lateral views of hypo-
stome, locality 3069-2, X 3. 34, 35, 36, Dorsal, right lateral, and an-
terior views of cranidium, locality 3069-2, * 73. 37, 38, 43, Dorsal,
posterior, and right lateral views of pygidium, locality 3069-2, 73.
30, 44, Left lateral and ventral views of hypostome, locality 3069-2,
X 3. 40, 41, 42, Exterior, posterior, and interior views of paratype
rostrum, U.S.N.M. No. 116514d, locality 3070-2, * 6. 45, Dorsal
view of free cheek, locality 3060-4, X 3. Helms formation, Chester
Series, upper Mississippian, Hueco Mountains, west Tex. Locality
numbers are explained on page I.

PLATE: 3

Paladin (Paladin) helmsensis Whittington, new species............e.eeeee 7)
I, 2, Exterior and interior views of incomplete cranidium, showing
the four pairs of glabellar furrows and the pits in the external sur-
face, locality 3070-4, < 6. 3, Anterior view of broken edge of free
cheek, showing thickness of exoskeleton, locality 3070-4, < 6. 4, 6,
Exterior and interior views of paratype free cheek, U.S.N.M. No.
116514a, showing eye surface and Panderian notch (p), locality
3070-2, X 73%. 5, Interior view of incomplete pygidium, showing
muscle scars as dark patches between ring furrows of axis, and
thickness of exoskeleton, locality 3070-2, * 6. 7, 8, 9, Posterior,
dorsal, and left lateral views of paratype segment, U.S.N.M. No.
116514e, locality 3070-4, X 3. 10, II, 12, Ventral, posterior, and an-
terior views of paratype segment, U.S.N.M. No. 116514g, locality
3069-4, X6. rp=ring process; ap=axial process; as = axial
socket; g—= groove in posterior edge of inner part of pleura. 13,
Dorsal view of same, X 3. 14, Interior view of paratype incomplete
segment, U.S.N.M. No. 116514h, locality 3069-4, 74, showing
notch in doublure termed Panderian opening. 15, Dorsal view of
paratype segment, U.S.N.M., No. 116514f, locality 3070-4, X 6. 16,
Left lateral view of same, <3. 17, Interior view of right half of
holotype cranidium, U.S.N.M. No. 116513, showing pit in boss
formed by anterior pit in external surface, locality 3070-2, X 73.
Helms formation, Chester Series, upper Mississippian, Hueco Moun-
tains, west Tex. Locality numbers are explained on page I.

mht A

hates

SMITHSONIAN MISCELLANEOUS COLLECTIONS

PALADIN (PALADIN) MORROWENSIS AND PALADIN (KASKIA) RARUS

(SEE EXPLANATION OF PLATES AT END OF TEXT.)

SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122, NO. 10, PL, 2

PALADIN (PALADIN) HELMSENSIS

(SEE EXPLANATION OF PLATES AT END OF TEXT.)

ITHSONIAN MISCELLANEOUS COLLECTIONS
VOL. 122, NO. 10,

PALADIN (PALADIN) HELMSENSIS

EXPLANATION OF PLATES AT END OF T

MAPA
a)

nan

PA A A oe FMR SNE Cee mE er OR

SMITHSONIAN MISCELLANEOUS COLLECTIONS
VOLUME 122, NUMBER 11

A REVISION OF THE SEA-STARS OF
THE GENUS: TETHYASTER

(Wits 12 PiatEs)

BY
AILSA M. CLARK
British Museum (Natural History)
AND
AUSTIN H. CLARK

Associate in Zoology, U. S. National Museum

(PuBLICATION 4147)

CITY OF WASHINGTON
PUBLISHED BY THE SMITHSONIAN INSTITUTION
APRIL 8, 1954

SMITHSONIAN MISCELLANEOUS COLLECTIONS
VOLUME 122, NUMBER 11

Pore VISION OF (THE SEA-STARS OF
THE GENUS TETHYASTER

(Wit 12 PLATES)

BY
AILSA M. CLARK
British Museum (Natural History)
AND
AUSTIN H. CLARK

Associate in Zoology, U. S. National Museum

(PUBLICATION 4147)

CITY OF WASHINGTON
PUBLISHED BY THE SMITHSONIAN INSTITUTION
APRIL 8, 1954

The Lord Baltimore Press

BALTIMORE, MD., U. & A.

A REVISION OF THE SEA-SPARSIO“ DEE
GENUS EP EY AS LER
By AILSA M. CLARK
British Museum (Natural History)

AND

AUSTIN H. CLARK
Associate in Zoology, U. S. National Museum

(WitH 12 PLATES)

The interrelationships of the sea-stars that we regard as constituting
the genus Tethyaster have never been satisfactorily worked out. All
the species are rare—at least few specimens have been collected—and
no one museum has been able to secure a fully representative series
either of the included species or of the growth stages of any single
species. The growth stages in this genus are especially important, for
the young may present an aspect quite different from that of fully
developed individuals, and the adult characters are often late in
making their appearance.

In the preparation of this revision we have studied all the specimens
in the U. S. National Museum, in the British Museum (Natural His-
tory), and in the Museum of Comparative Zoology at Cambridge,
Mass., for the loan of which we are greatly indebted to our friend
Dr. Elisabeth Deichmann.

Two of the species (canaliculatus and vestitus) have not previously
been figured, the type specimen of another (magnificus) has not been
figured, and of one (grandis) only a few details have been illustrated.
Of the others, two (subinermis and aulophora) have been illustrated
in satisfactory detail, and the last (pacei) has been figured sufficiently
for purposes of identification.

Thomas Say in 1825 (p. 143) described a very large sea-star from
New Jersey under the name of Asterias vestita, as follows:

5. A. vestita. Disk broad, surface reticulated, covered by cylindrical promi-
nences, margin articulated; rays depressed.

The whole surface is covered by cylindrical prominences, which are placed

near each other, truncated at their summits, and each summit crowned by from
ten to eighteen small, equal, cylindrical fimbriae; wart-like tubercle [madre-

SMITHSONIAN MISCELLANEOUS COLLECTIONS, VOL. 122, NO. 11

2 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

porite] large, radiated, very conspicuous; margin articulated; each articulation
with about four very much compressed, subquadrate, truncated spines or move-
able processes, which are vertically adpressed to the surface of the segment, and
are imbricated with respect to each other.

Diameter 1 foot 2 inches.

The locality was given as Cape May, N. J. Say said that it is “Allied
to A. aranciaca Linn., but distinct by many characters, and particu-
larly by the form and number of the lateral spines. It is very rare on
this coast.”” The type and only known specimen disappeared, and the
species has since remained an enigma.

Dr. A. Philippi in 1837 (p. 193) briefly described Asterias subiner-
mis from a specimen 14 inches in diameter from the coast of Sicily.
This well-known but rare species has been recorded from both coasts
of the Mediterranean as far east as Rhodes and from the Bay of Bis-
cay southward to the Gulf of Guinea. It has never been confused with
any other species and has no synonyms, but it has been assigned to
various genera—Astropecten (Muller and Troschel, 1842), Archas-
ter (Perrier, 1875), Gontopecten (Perrier, 1885), Plutonaster (Te-
thyaster) (Sladen, 1889), and Tethyaster (Perrier, 1894), almost all
these dispositions being followed by other authors. A detailed account
of this species, under the name Plutonaster subinermis, with figures
and bibliographic references, was given by Ludwig in 1897 (p. 105).

Say’s Asterias vestita was listed, without description, as Astropecten
vestitus by Liitken in 1859 (pp. 27, 54). Verrill in 1866 (p. 339)
under Astropecten vestitus Litken said “Say’s specimen was from
Cape May, collected by Mr. J. Robbins. I am not aware of any other
being found.”

In 1882 (p. 440) Prof. F. Jeffrey Bell described Archaster magnifi-
cus from two specimens with R=207 and 138 mm., and r=50 and
37 mm., which had been presented to the British Museum some years
before by J. C. Melliss who had obtained them at St. Helena.

W. Percy Sladen in 1889 established, under Plutonaster, the sub-
genus Tethyaster (p. 101) in which he placed Philippi’s Asterias sub-
inermis and Diiben and Koren’s Astropecten parelii. He also (p. 192)
diagnosed the genus Moiraster for the reception of Archaster magni-
ficus Bell.

In 1895 Verrill (p. 133) listed Astropecten vestitus Liitken and
said “B. range, shallow water. Cape May (Say). It is not uncommon
farther south.”* In 1899 Verrill (p. 210) proposed the new genus
Sideriaster, based upon a new species, S. grandis, from Albatross
station 2378. The description was brief, but he figured the actinal side

1 Possibly here confused by Verrill with Astropecten cingulatus Sladen.

NOs, Tt SEA-STARS—CLARK AND CLARK 3

of a part of the middle of a ray, an adambulacral plate, and an
abactinal paxilla.

In 1908 Dr. René Koehler described in detail and figured a small
specimen of Motraster magnificus with R = 62 mm. from Pointe
Pyramid, Ascension, in 40 fathoms.

In 1914 (p. 21) Verrill discussed Sideriaster(?) vestitus (Say).
He said that the type of Sideriaster, S. grandis, does not agree suffi-
ciently well with vestitus to be identified as the same species, but it
seems almost certain that it is congeneric. He added that when more
specimens can be obtained it may prove to be the same species.

In 1915 (p. 191) Verrill republished his diagnosis of Sideriaster
and (p. 192) his description of S. grandis, also republishing the fig-
ures of details previously given. He also discussed (pp. 193-195)
Sideriaster (?) vestitus (Say) Verrill at considerable length. He
noted that “Probably the type is lost. It is probably not an Astro-
pecten. In having a large disk, and especially in having four appressed
spines in a transverse row on the inferomarginal plates, the Sideriaster
grandis V. agrees, perhaps, with Say’s species. But he gives too little,
as to other characters, to enable me to say whether they are related.”

In 1916 (p. 52) A. H. Clark described in detail, but did not figure,
Sideriaster canaliculata from Albatross station 2998, Gulf of Cali-
fornia, in 40 fathoms.

Dr. Walter K. Fisher in 1911 (p. 417) published a diagnosis of a
new genus, Anthosticte, based upon a new species, A. aulophora,
described from a single specimen from Albatross station 5420 in the
Philippines. In 1919 (p. 140) he republished the diagnosis of Antho-
sticte, redescribed and figured A. aulophora, and discussed the rela-
tionships of the new genus.

Dr. Th. Mortensen in 1925 (b, p. 147) described and figured
Anthosticte pacei from South Africa. He wrote that “from the type
species A. aulophora, the only species hitherto known of the genus An-
thosticte, the present species is easily distinguished through the lower
paxillae and the complete lack of pedicellariae.” In 1933 (p. 422)
Mortensen.recorded three specimens of Moiraster magnificus that he
dredged off Egg Island, St. Helena. In these R = 160-179 mm. He
gave various details of the specimens.

In 1947 Sefiorita Maria Elena Caso described and figured Moiraster
gigas from a very large specimen with R = 205-245 mm. from Santa
Rosalia, Baja California, on the western shore of the Gulf of
California.

In 1950 (p. 302) A. H. Clark recorded under the name of Moiraster
magnificus a specimen with R = 168 mm. from off the western coast

4 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

of Puerto Rico. This specimen we now consider as representing
vestitus.

As for the interrelationships of these genera, they were discussed
in some detail by Fisher in 1919 (p. 143). He said that Moiraster,
Tethyaster, Sideriaster, and Anthosticte agree in having unarmed
supermarginals, inferomarginals with a few small enlarged spines,
naked madreporite, large actinal interradial areas, and intermediate
plates far along the ray, marginal and actinal fascioles, true paxillae,
stellate abactinal plates, an astropectinid adambulacral armature, and
probably also in having the single papulae uninterrupted all over the
dorsal surface. He said that the first two seem to be a little more closely
related than either is to the last two, while Sideriaster and Anthosticte
are possibly also nearly related. He noted that unfortunately there is
but one species in each genus, and it is difficult to ascertain what char-
acters are of generic importance. He remarked that, according to the
standards used in other larger genera, Anthosticte differs from Tethy-
aster chiefly in having very deep marginal fascioles, gonads to the end
of the ray, and no midradial series of enlarged paxillae. Anthosticte
has taller and more delicate paxillae, but this may not be of generic
importance. Its special points of agreement in addition to the char-
acters listed are the deposits in the tube feet (not recorded for
Moiraster and Sideriaster) and shallow interambulacral fascioles, and
an incipient interradial series of actinal intermediate plates, less
prominent and regular than in Tethyaster.

He said that Anthosticte differs from Sideriaster in having very
deep marginal fascioles and no distally enlarged subambulacral spines.
Neither the deposits in the tube feet nor the gonads of Sideriaster are
described. He considered that the fascioles between the adambulacral
plates which he examined in the type specimen of Sideriaster grandis
form one of the most striking features of the genus. They are
densely lined with small, delicate spinelets, and are therefore similar
to marginal fascioles. Such is not the case, he said, in Anthosticte,
Tethyaster, or Moiraster.

Mortensen in 1933 (p. 424) also discussed these four genera. He
wrote that the knowledge now acquired of the characters of Moiraster
(from his three specimens from St. Helena) makes it clear that the
four genera are even more closely related than Fisher thought them
to be—so closely, indeed, that it seems scarcely possible to maintain
them all. He said that Tethyaster is well characterized by its mid-
radial row of enlarged paxillae, the shallow marginal fascioles, and
the low paxillae, so it may justly be maintained as a separate genus.
He noted that Fisher’s statement that its gonads do not extend to the

NOS DE SEA-STARS—CLARK AND CLARK 5

ends of the rays is a curious mistake, “in flat contradiction to the
description given by Ludwig.” As for Anthosticte, he said that it is
now seen that the only character by which it differs from Moiraster
is the absence of enlarged spines on the ventrolateral plates. In regard
to fascioles between the adambulacral plates he said there seems to be
a very gradual passage from Anthosticte to Moiraster and Sideriaster.
He noted that Sideriaster, which is still imperfectly known, would
likewise seem to differ from Moiraster only in lacking enlarged spines
on the ventrolateral plates. He said that it is, of course, a matter of
taste whether this character, the presence or absence of enlarged spines
on the ventrolateral plates, affords sufficient reason for generic dis-
tinction. But, he added, this is all the difference there is.

In 1950 the question of the identity of a specimen taken by the
M.V. Rosaura off the mouth of the Orinoco in 75 meters was raised
between the two present authors. Dr. Dilwyn John has provisionally
attributed the specimen to Sideriaster, but investigation seemed to
show that it also had some affinity with Bell’s Moiraster magnificus
from St. Helena. At about the same time the M.V. Oregon dredged
10 specimens of Sideriaster grandis off Corpus Christi, Tex., another
was dredged by the yacht Triton off Sombrero Key, Fla., and still
another was received by the U. S. National Museum from off the
coast of Tamaulipas, Mexico. Furthermore, the Museum acquired a
very large sea-star from the coast of North Carolina that agrees com-
pletely with the meager description of Say’s Asierias vestita, but is
slightly larger. There seems to be no doubt that it represents Say’s
long-lost species. .

With this additional material available it has seemed advisable to
review the status of Tethyaster, Moiraster, Sideriaster, and Antho-
sticte. We have personally examined specimens of all the species
described in these genera except Anthosticte pacei, which was briefly,
though sufficiently, described and figured by Mortensen.

We can see no valid reason for not considering all these species
congeneric and we therefore unite them all in the genus Tethyaster,
of which we regard Moiraster, Sideriaster, and Anthosticte as
synonyms.

Genus TETHYASTER Sladen

Asterias (part) SAy, 1825, p. 143.—Puitprt, 1837, p. 193.

Astropecten (part) MULter and TroscHeEL, 1842, p. 74, following authors.

Archaster (part) PERRIER, 1875, p. 3690, and following authors.

Goniopecten (part) PERRIER, 1885, p. 71.

Plutonaster (subgenus Tethyaster) (part) SLADEN, 1889, p. 101 (diagnosis;
genotype Asterias subinermis Philippi).

6 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

Moiraster SLADEN, 1889, p. 192 (diagnosis; genotype Archaster magnificus Bell).
—FISHER, I919, pp. 143, 144 (discussion).—MoRTENSEN, 1933, Pp. 424
(discussion).

Tethyaster PERRIER, 1894, p. 322; 1896, p. 50.—KOEHLER, 1896a, pp. 56, 57; 1896b,
Pp. 450, 451.—GREGORY, 1900, p. 251.—FISHER, I910, p. 143.— KOEHLER, 1921,
P. 53; 1924, p. 190.—RIVERA, 1930, p. 105—MORTENSEN, 1933, D. 424.

Plutonaster (part) Lupwic, 1897, p. 105—CUENOT, 1927, p. 295.—NosrRE, 1931,
figs. 42, 43.

Sideriaster VERRILL, 1890, p. 210 (diagnosis; genotype Sideriaster grandis, sp.
nov.) ; I9I4, p. 21; 1915, p. 191.—FISHER, 1919, p. 143.—MORTENSEN, 1933,
p. 424.

Anthosticte FISHER, I9II, p. 417 (diagnosis; genotype Anthosticte aulophora, sp.
nov. ).—MOoRTENSEN, 1933, Pp. 424.

Thetyaster NoBrE, 1931, p. 62.

Thethyaster Nore, 1931, p. 176.

Diagnosis —A genus of Astropectinidae with both series of mar-
ginal plates large and conspicuous, equally developed, the superomar-
ginals granulated or with numerous short spinelets, the inferomar-
ginals with a median row of usually about five enlarged and flattened
appressed spines; actinal intermediate areas large with numerous
intermediate plates arranged in definite series with an incomplete un-
paired median row; the inferomarginals separated from the adambu-
lacrals by a series of actinal intermediate plates for most of the ray ;
fascioles between the marginals, adambulacrals, and actinal inter-
mediate plates; madreporite large and bare; adambulacral armature
astropectinid; abactinal plates with paxillae having high columns;
gonads extending far along ray.

Geographical range.-—From New Jersey south to the mouth of the
Orinoco ; Gulf of Mexico; St. Helena and Ascension; Bay of Biscay
south to the Gulf of Guinea; Mediterranean east to the Aegean Sea;
South Africa; Philippines ; Gulf of California.

Bathymetrical range.—From 44 to about 1,400 (?1,425) meters.

Remarks.—Presumably the most specialized species of Tethyaster
are those with the spines on the inferomarginal and actinal intermedi-
ate plates wide, rectangular, and broadly truncated, as these depart
most widely from the generalized astropectinid type. Although our
knowledge of this genus is admittedly meager, these species appear
to be primarily American, ranging from New Jersey to Venezuela
(vestitus), occurring at St. Helena and Ascension (magnificus), and
found also in the Gulf of California (canaliculatus). This group in its
distribution would parallel roughly the genera Encope, Mellita, and
Leodia among the echinoids, the Marginatus group of Astropecten,
and Asirocaneum in the Gorgonocephalidae. It should be noted that

NO. II SEA-STARS—CLARK AND CLARK A

the crinoid genus Crinometra so very characteristic of the Caribbean
area is also represented at St. Helena.

A more generalized type with less strongly modified spines on the
inferomarginals, which only very rarely extend on to the actinal inter-
mediate plates, is widely distributed, occurring in the Gulf of Mexico
(grandis), in the Mediterranean and east Atlantic from the Bay of
Biscay to the Gulf of Guinea (subinermis), off South Africa (pacer),
and in the Philippines (aulophora). At the same time aulophora is
distinguished from the other members of this group by the relatively
tall and slender dorsal paxillar columns, such as are found also in
vestitus and magnificus. However, this character is probably less
fundamental than the shape of the inferomarginal spines.

Other differences between the species are shown in the key.

KEY TO THE SPECIES OF TETHYASTER

(This key is adapted for fully developed specimens with R = 100 mm. or more.)

a1, Actinal intermediate plates each with an enlarged, broad, flat-
tened, and broadly truncated procumbent spine directed out-
ward (if these are undeveloped the inferomarginal spines are
broadly truncated) ; enlarged and flattened spines on the in-
feromarginals broad, usually approximately rectangular or
scoop-shaped with broadly truncated ends, rarely tapering ;
R = 200-250 mm, in fully grown individuals.
b1, Enlarged spines on actinal intermediate plates fan-shaped or
scoop-shaped with divergent sides and broadly truncated
ends; spines on the inferomarginals similar (may be
tapering in young individuals: pl. 3; fig. 1, c) (Gulf of
Galton ies a he ee NOES 5 Mee Ce RS RN Te canaliculatus
b2. Enlarged spines on actinal intermediate and inferomarginal
plates rectangular, rarely scoop-shaped.
ci, Enlarged and flattened spines in fully grown individuals
6-7 mm. long, first appearing when R—=about 70 mm.
(plig) (St. Helena and Ascension) 25) ...s.2..025 000. magnificus
c?, Enlarged and flattened spines in fully grown individuals
reaching only 4 mm., first appearing when R= about
150 mm. (pl. 6; fig. 1, d) (New Jersey south to off the
MOET RAVE eae acta © Ga orci c Sasa Sco Sia wae see acpiaee a aise vestitus
e”, Actinal intermediate plates without a central enlarged spine
(there may be a very few pointed spines in some specimens
of grandis) ; enlarged and flattened spines on inferomarginals
sharp-pointed.
G4. Columns of paxillae slender, high, about 4 times as high as
thick; most of paxillae with a pedicellaria of 2-4 valves
WE MGpines Reraes soso os alck eo Paes Hose eee bae leaned cmeee aulophora
b2. Columns of paxillae stout, low, not over twice as high as
thick; no abactinal pedicellariae, so far as known.

8 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

cl. Marginals short, 68-85 in number; rays fairly broad, width
of rays at base=r; paxillae of median row on rays
sometimes larger than others (Bay of Biscay to Gulf of
Guinea and most of Mediterranean) ...........eee2eeeee subinermis
c?. Marginals longer, up to 65 in number; rays narrow, width
at base less than r; paxillae of median row on rays not
larger than others.
d1, Rays very narrow, R= 4.3 1; first series of actinal in-
termediate plates reaching to about outer third of ray,
second only in proximal third; intermarginal fascioles
deep, extending inward for about two-thirds the proxi-
mal and distal sides of marginals (South Africa)........... pacei
d?, Rays broader basally, R=3.2 to 3.5 r; first series of
actinal intermediate plates reaching to about outer
fourth of ray, second to well beyond middle; intermar-
ginal fascioles shallow, extending inward for less than
one-third the proximal and distal sides of marginals
Cols: 15,782). ACGulivok Mexico)! ii nrrss 2. Seem: ote ovis sore grandis

TETHYASTER CANALICULATUS (A. H. Clark)
Plates 1-4; text figure 1, c

Sideriaster canaliculata A. H. CiarK, 1916, p. 52 (description; Albatross station
2908) .—ZIESENHENNE, 1937, p. 212 (notes; Zaca stations 136, D-19; 142,
D-3; 146, D-1; 147, D-2; 150, D-9).—Caso, 1947, p. 225 (listed).

Moiraster canaliculata Caso, 1947, p. 225 (listed).

Motraster gigas CAso, 1947, p. 225, fig. 1, p. 226, fig. 2, p. 227, fig. 3, p. 228, fig.
4, p. 229 (description; Santa Rosalia, Baja California).

Diagnosis —Enlarged spines on the inferomarginals and actinal
intermediate plates scoop-shaped with divergent and convex sides,
broadly truncate, the outer portion commonly with a broad, shallow
groove and the distal end slightly notched ; size large, R up to 250 mm.

Type.—tIn the U. S. National Museum (No. 36951).

Type locality —Albatross station 2998, Gulf of California west of
Culiacan, Sinaloa (lat. 24°51’00” N., long. 110°39’00” W.); 73
meters ; bottom temperature 64° F.; March 16, 1889.

Additional localities —Santa Rosalia, Baja California (Caso, 1947).

Zaca station 136, D-14; Arena Bank, Gulf of California (lat.
23°29'30” N., long. 109°25’ W.) ; 82 meters; April 20, 1936 (Ziesen-
henne, 1937).

Zaca station 142, D-3; Santa Inez Bay, Gulf of California (lat.
27°04’ N., long. 111°53' W.); 73 meters; April 11, 1936 (Ziesen-
henne, 1937).

Zaca station 146, D-1; Santa Inez Bay (lat. 26°52’ N., long
111°53’ W.) ; 64 meters; April 16, 1936 (Ziesenhenne, 1937).

NO: Tt SEA-STARS—CLARK AND CLARK 9

Zaca station 147, D-2; Santa Inez Bay (26°57’30” N., long.
111°48’30” W.) ; 110 meters; April 17, 1936 (Ziesenhenne, 1937).

Zaca station 150, D-g; Gorda Banks, Gulf of California (lat.
23°04’ N., long. 109°30’30” W.); gI-109 meters; April 22, 1936
(Ziesenhenne, 1937).

Geographical range-—Central and southern part of the Gulf of
California.

5 mm

Es
y

Sy
©
vy)

Qe
gO
OL)
20,
=

090 0006 =
0 OCOO9° =
O00 oO
SURI. Ate
soQ00°08
aceecose GH
00 9 920% Ss
00992000 ar
OF99g0 co >
29,000 0 OO?
20 99 9 90° ° =
S$ 5009050 QD
2009°909 9
0°09° =
2000 02
959900 9
[a] 00009 d
é

a b c d

Fic. 1—Adambulacrals, intermediate plates, and inferomarginals of a, Tethy-
aster subinermis (R = 86 mm.) ; b, T. grandis (R = 59 mm.) ; c, T. canaliculatus
(R= 64 mm.) ; d, T. vestitus (R= 70 mm.). The adambulacral plate is at the

top in each case and is about the tenth, while the inferomarginal corresponding
is about the seventh.

Bathymetrical range.—From 64 to 110 meters.

Remarks.—Our reasons for considering Sefiorita Caso’s Moiraster
gigas a synonym of the previously described but much smaller Sideri-
aster canaliculatus are the following. The supermarginals of the type
of canaliculatus (R = 64 mm.) number 45, those of the type of gigas
(R = 205-245 mm.) 58-62. Considering the discrepancy in size this
difference is negligible. In canaliculatus R:r = 3.4: 1, in gigas (aver-
age) 3.6:1. This difference is not significant.

The paxillae on the rays in canaliculatus are in three regular alter-
nating rows in the midradial region, and from these central rows diago-

IO SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

nal rows run out at an angle of 45° to the superomarginals. Judging
from Sefiorita Caso’s figure (fig. 3) there is the same arrangement
in gigas, though here the lateral rows make a larger angle with the
median.

In canaliculatus the enlarged and flattened spines in the infero-
marginals are broad, broadly truncate, about half as broad as long, or
even shorter, with convex sides, the outer half commonly deepened
in the middle or broadly grooved. In gigas these spines are “espa-
tuladas, truncadas, aplanadas, ligeramente henidas en sus extremos
libres.”

In the type of canaliculatus most of the actinal intermediate plates
carry one or two enlarged tubercles somewhat swollen in the outer
half and frequently somewhat flattened. These resemble the less-
developed of the corresponding spines on the actinal intermediate
plates in the young specimen of vestitus from Puerto Rico.

In the type of canaliculatus all the plates bordering the adambulac-
rals and mouth plates, and a few of the other actinal intermediate
plates, bear pedicellariae with three or four valves. In gigas also
“Todas las placas que limitan a las placas adambulacrales y las placas
bucales y alguna que otra placa intermediaria, estan provistas de
pedicelarios espiniformis, trivalvados, de forma irregular ; en general,
unos son pequefios y otros grandes.”

In the largest specimen collected by the Zaca (M.C.Z. No. 36232;
pl, 3) R= 95.mm., r= 726 mm.; Rerj=32:6:1.. The ipaxilise have
cylindrical columns which are about half again as high as thick and
rather slender. The crown consists of 8-10 subcapitate peripheral
spinelets, mostly about twice as high as thick at the base, with one or
two usually much more slender than the others, and most frequently
a single central spinelet which resembles the larger peripheral.

There are 54-58 marginals. The superomarginals resemble those
of the smaller type specimen (R = 64mm.,r = 19 mm.). The infero-
marginals have usually two flattened, tapering, and pointed spines
which at the base of the rays are 3-4 mm. long and about 0.75 mm.
broad at the base.

Each actinal intermediate plate carries in the middle a strongly flat-
tened wedge-shaped or narrowly fan-shaped spine with straight sides
and a gently convex tip which is usually about twice as broad as the
base. These flattened spines, which are mostly 1-1.5 mm. long, lie
parallel with the surface of the plate, directed toward the infero-
marginals. In addition to these central spines the plates bear a few
much smaller subcapitate spines and numerous fine lateral spinelets.

The enlarged spines on the actinal surface of the adambulacral

NO: LI SEA-STARS—CLARK AND CLARK itil

plates are somewhat flaring, abruptly truncated, and broadly grooved
or chisel-shaped.

A number of the plates in the inner part of the interradial areas,
especially those adjoining the adambulacrals, bear pedicellariae of
three, sometimes four, valves which resemble stout subcapitate spines,
one of which is commonly smaller than the others.

This specimen is intermediate between the smaller type specimen
and Sefiorita Caso’s much larger type of gigas, resembling the latter
in having enlarged and flattened spines on all the actinal intermediate
plates. So far as can be judged from the published figure, these spines
resemble those of gigas. The spines on the inferomarginals, however,
are tapered and pointed and resemble those of grandis more nearly
than those of the type of gigas.

In a specimen from Arena Bank in 4o fathoms (M.C.Z. No. 3447;
pl. 4, upper) with R= 41 im, f=.15 mnt.,.Rir =2.7: 1, the paxil-
lae are very low, the thick columns being about as high as broad. The
crown consists of 6-7 peripheral granules and one central elongated
capitate granule, all similar.

There are 39 marginals. The inferomarginals are mostly on the
actinal surface, the outer ends curving upward to meet the supero-
marginals. In lateral view they are at the arm bases about half as high
as the superomarginals, but the height of the superomarginals de-
creases so that in the outer half of the ray the two series are, in
lateral view, of about the same height.

The marginals and the actinal intermediate plates have a similar
covering of granules with swollen tips which are not higher than thick,
largest and lowest in the center of the plates, becoming more slender
along the edges. The spines on the adambulacral plates resemble those
of the larger specimen. The very few pedicellariae are in the inner
part of the interradial areas.

A specimen from Santa Inez Bay in 35 fathoms (M.C.Z. No. 3448;
pi 4, lower) with R’=="32°mm:, r= rr mm.,’Rir'='2!9: 1,"and 34
marginals, resembles the preceding, but there are no pedicellariae and
the adambulacral spines, though similar, are not so stout.

A specimen (M.C.Z. No. 36251) with R = 30 mm., r= 11 mm.,
R:r=2.7:1, and 30 marginals, resembles the preceding, as does
anoter (SCZ: No, 36251) with R*="18'mm., fr ="7 mm. Rer =
2.6: 1, and 27 marginals.

In five small specimens from Gorda Bank (M.C.Z. No. 3449) R =
7-10 mm., r = 3.5-4.5 mm. The spines on the adambulacral plates are
slender, cylindrical or slightly swollen in the outer part, and little if at
all flattened ; they bear numerous fine serrations. All the granules on

I2 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

both surfaces are more slender, relatively higher, and less crowded
than in the larger specimens, and bear very numerous fine serrations.

In a specimen from Santa Inez Bay (M.C.Z. No. 3450) with R =
4 mm., r = 2.7 mm., there are 8 marginals. The terminal plate is
very large. All the spines and the slender elongate granules are
spinulose.

With decrease in size the madreporite decreases in relative size and
in the smaller specimens cannot be distinguished.

In nearly all the specimens we have seen the rays differ somewhat
in length, as was the case in Sefiorita Caso’s type of gigas.

Specimens examined.—All known specimens except the type of

gigas.
TETHYASTER VESTITUS (Say)
Plates 5-8; text figure 1, d

Asterias vestita Say, 1825, p. 143 (description; Cape May, N. J.).

Astropecten vestitus LUTKEN, 1850, pp. 27, 54 (listed).—VERRILL, 1866, p. 339
(Say’s record) ; 1895, p. 133 (Say’s record ; “not uncommon farther south.”’)

Sideriaster (?) vestitus VERRILL, 1914, p. 21 (identity of Say’s species) ; 1915,
PP. 193-195 (discussion).

Moiraster magnificus A. H. CLARK, 1950, p. 302 (off Puerto Rico) (not magnifi-
cus Bell, 1882).

Diagnosis—Enlarged spines on the inferomarginal and _actinal
intermediate plates rectangular, sometimes tapering distally, broadly
truncate, in fully grown individuals up to 4 mm. long, first appearing
when R is about 150 mm.; size large, R up to 250 mm.

Description —The paxillae are compact and in contact, with a slen-
der tall pedicel 3-4 mm. high and four or five times as high as broad,
crowned by a floriform group of usually 20-30 peripheral and 5-15 or
more central, terete, slightly tapering, blunt spinelets 0.5 to I mm.
long and 3-5 times as long as broad at the base. On the disk and arms
they are arranged in rows at an angle of approximately 75° to the
midradial line ; in the middle of the interradial areas of the disk there
are 4 or 5 usually irregular rows that do not reach the interradial
border. The paxillae are largest on the disk, slowly and gradually
becoming slightly smaller with more slender, more pointed, and rela-
tively longer spinules toward the interradial margins and on the rays.
The madreporite is very large, 17 by 16 mm., slightly sunken below
the summits of the surrounding paxillae, slightly concave with very
numerous and fine, regularly radiating striae.

The marginal plates correspond throughout the ray. They are high
and narrow with very deep fasciolar channels between them, the chan-

NO. II SEA-STARS—CLARK AND CLARK 13

nels being roughly twice as deep as the width of the summits of the
marginals or even deeper. The superomarginals, 78 in number, are in
the interradial arcs flat and 13 mm. high, decreasing in height to 6 mm.
and becoming slightly convex on about the sixth, then remaining simi-
lar to near the arm tips. They bear about four rows of elongated
granules or short spinelets. The two outer rows are regular and are
composed of slightly tapering spinelets, resembling those of the paxil-
lae in the center of the disk, and about three times as long as broad
at the base; the two median rows are irregular and are composed of
shorter and stouter spinelets. On the high, narrow superomarginals
in the interradial arcs there are only two rows of spinelets, or an
irregular single row, these spinelets resembling those of the outer rows
of the outer superomarginals. The inner rows first appear on about
the eighth superomarginal. The superomarginals are bordered on
each side by an irregular double row of from 30 to 40 slender tapering
spinelets which just meet those of the neighboring superomarginals.
The lower end of the superomarginals is bordered by a row of about
6 stout spinelets directed diagonally outward.

The inferomarginals are about the same size as the superomarginals
in the interradial angles but, being more uniform in size, are slightly
larger elsewhere. In the interradial arcs they are flat and make a
considerable angle with the superomarginals, but they soon become
convex, continuing the curve of the superomarginals to the flat actinal
surface. They bear a median row of usually four broad, flat, trun-
cated, appressed spines 3 to 4 mm. long and 1.25 to 1.50 mm. wide,
which overlap the bases of those succeeding. These are flanked by
much smaller flattened and truncated spines mixed with more or less
terete spinelets. The plates are bordered laterally by very numerous
fine, laterally directed spinules resembling those bordering the supero-
marginals. Toward the ends of the rays the enlarged spines become
very short, not much longer than broad.

The terminal plate is rather large, swollen, heart-shaped, with the
distal end deeply notched and the proximal end broadly truncated. It
overlies about 4 superomarginals.

The actinal intermediate areas are large. One series of plates ex-
tends to within about 20 mm. of the end of the ray, a second to within
about 40 mm., and a third to about the middle. An incomplete and
usually irregular row extends from a pair just beyond the mouth
plates to about one-third the distance to the inferomarginals. Between
the first inferomarginal and the second adambulacral the series con-
tains 9 or 10 plates. Deep channels lead from the marginal fascioles to
the fascioles between the adambulacrals, these being separated by single

14 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

regular rows of actinal intermediate plates. Each of these intermediate
plates bears a tall, stout, laterally compressed column somewhat
broadened at the summit, which bears a large, flattened, truncated
spine 4 mm. long and about 1.3 mm. broad resembling the large spines
on the inferomarginals. This spine lies parallel with the surface of
the plate and is directed toward the inferomarginals. Occasional plates
may carry two or even three of these spines. The median spine is
accompanied by a few much smaller, flattened spines or stout spinelets,
and the border of the summit of the plate carries a large number of
fine, laterally directed slender spinelets directed laterally and arranged
in an irregular double row. Ina single interradius seven of the actinal
intermediate plates carry a fine, somewhat scattered granulation in-
stead of the large flattened spine and the accompanying smaller ones.
Many of the actinal intermediate plates adjoining the adambulacrals
carry spiniform pedicellariae with usually two, sometimes three,
blades.

The adambulacral plates are broader than long. The inner half
forms an acute angle of roughly 60° and the outer half has parallel
sides. At the apex of the furrow angle there is a stout, sharp, pris-
matic, slightly recurved spine which in the basal part of the ray is
7 or 8 mm. long. On the sides of the angle are two, sometimes three,
similar but much flattened sharp spines of about the same dimensions.
The actinal surface of the plate carries usually 4 or 5 spines as stout
as the marginal spines but slightly shorter, flattened, broadly truncate,
and fluted in the outer half. Each adambulacral plate therefore carries
a more or less compact group of usually 9 or Io generally similar con-
spicuously large and stout spines. The plates are bordered laterally by
an irregular double row of fine spinules similar to those on the actinal
intermediate plates.

The mouth plates are densely covered with spines, larger on the
inner third (toward the mouth) than elsewhere. There are about 6
enlarged and strongly flattened marginal spines which are placed far
down on the side of the plate, with a second series parallel with them
along the edge of the plate. Beyond the marginal spines the mouth
plates carry along their border very numerous, very fine, laterally
directed spinules arranged in about three rows.

Type—Lost; the specimen described above, from the coast of
North Carolina, may be regarded as a neotype.

Type locality—Cape May, N. J.; the type was collected by
J. Robbins.

Additional localities—Twelve miles west-southwest of Diamond

Shoal, N. C.; 44 meters; February 6, 1951 (U.S.N.M. No. E.8000).

NO. II SEA-STARS—CLARK AND CLARK 15

Caroline station 35; off the west coast of Puerto Rico (lat.
18°24'45” N., long. 67°14’15” W.) ; 146-329 meters; 1933 (U.S.N.M.
No. E.3963).

Rosaura station 35 ; off the mouth of the Orinoco River ; 86 meters
(B.M. No. 1949.1.19.18).

Geographical range-—From New Jersey south to the mouth of the
Orinoco River.

Bathymetrical range.—From 44 to 146 (?329) meters.

Remarks.—We have no hesitation in identifying the specimen from
North Carolina with Say’s Asterias vestita for the following reasons.
The distinctive features in Say’s brief description are (1) the paxil-
lae on the abactinal surface ; (2) the large and very conspicuous mad-
reporite; (3) the occurrence of about four very much compressed,
subquadrate, truncated, and imbricated spines on the marginals; (4)
the size; and (5) the comparison with [Astropecten] aranciaca. All
these features are equally distinctive of the specimen from North
Carolina, but of no other species known from the western Atlantic.
The only species that might be considered in this connection is Ver-
rill’s Sideriaster grandis, but this is smaller and the spines on the
marginals are tapering and pointed, not subquadrate and truncated.

In the specimen from Puerto Rico (pls. 7, 8) R = 160 mm., r =
45 mm.; R = 3.6 r (R = 3.9 r in the large specimen from North
Carolina). The abactinal paxillae have short, stout, cylindrical col-
umns 1.25 mm. high and 0.75 mm. in diameter.

The marginals are 75 in number. The superomarginals are densely
covered with low, somewhat flattened granules, largest in the middle,
and resemble those near the tip of the rays in the large specimen in
which, however, the granules are higher. The enlarged spines on the
inferomarginals are small, mostly 1.5 to 2 mm. long by 1 mm. wide.

Many of the actinal intermediate plates, in some interradii more
than half, in others fewer, show the enlarged and flattened spines in
various stages of development ; most of them are about two-thirds the
size of those on the inferomarginals and of the same shape.

A few scattered paxillae on the disk and arm bases carry pedicel-
lariae with usually 3, occasionally 2 or 4, blades which are scarcely
more than slightly modified spines. A number of the actinal inter-
mediate plates of the inner row, especially in the second fourth of
the ray, carry a pedicellaria, sometimes two, consisting of scarcely
modified spines.

The specimen taken by the Rosaura has R= 75 mm. It has not
yet developed the enlarged spines on the actinal intermediate plates,
but the inferomarginal spines (see text fig. 1, d) are broadly truncated

16 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

or spatulate in shape. There are some actinal pedicellariae. The col-
umns of the midradial proximal dorsal paxillae are about 0.35 mm.
in maximum thickness, measuring 0.75 mm. in height, or 1-2 mm.
including the basal part. The paxillar spinelets are about 0.45 mm.
long. The marginals are 52 in number.

Specimens examined.—All those known.

TETHYASTER MAGNIFICUS (Bell)
Plates 9, 10; text figure 2, e-g

Archaster magnificus BELL, 1882, p. 440 (description; St. Helena).

Moiraster magnificus SLADEN, 1889, p. 193 (reassignment of Bell’s species) .—
KOEHLER, 1908, p. 630, pl. 12, figs. 107-110 (Ascension; notes).—MorTEN-
SEN, 10933, Dp. 422, text figure 6, pl. 21, figs. 1, 2, pl. 22, fig. 1 (Egg Island,
St. Helena; notes).—Caso, 1947, p. 225 (listed).

Diagnosis—Enlarged spines on the inferomarginal and _actinal
intermediate plates rectangular or even with divergent sides, rarely
somewhat tapering distally, broadly truncate, in fully grown individ-
uals 6-7 mm. long, first appearing when R = about 70 mm. ; size large,
R up to 220 mm. at least.

Types (2).—In the British Museum (Nos. 68.6.15.1 and 68.6.15.2).

Type locality—St. Helena, collected by J. C. Melliss; no further
details.

Additional localities —St. Helena, Egg Island; about 73 meters
(Mortensen, 1933). Ascension, Pyramid Point; 73 meters (Koehler,
1908).

Geographical range.—St. Helena and Ascension.

/———— lind =p

Imm

ed Bll

Fic. 2.—a-d, Tethyaster subinermis. a-c, specimen from Naples with R= 55
mm.; a and b, dorsal views of proximal midradial paxillae with and without
hiss c, side view of a complete paxilla; d, side view of a paxilla without
spinelets from a specimen from Senegambia with R= 72 mm.

e-g, Tethyaster magnificus, alcoholic specimen from St. Helena with R = 215
mm. e, paxillar spinelet; f, dorsolateral view of a paxilla without spinelets, and
single lobes of two adjacent ones; g, lateral view of two adjacent paxillae with-
out spinelets, showing their position in the skin. Drawings by A. M. Clark.

NO: II SEA-STARS—CLARK AND CLARK 17

Bathymetrical range-—The only definite records are 73 meters.

Remarks.—There are three specimens of this species in the British
Museum, all collected at St. Helena by J. C. Melliss, although Bell
mentioned only two. The third (B.M. No. 67.12.30.1) is preserved
in alcohol but possibly spent some time in formalin since it is in a very
flaccid state. The two types are both dry.

The larger of the types has R = 215, 213, 222, and 214 mm. on the
four entire arms (Bell gives 207 mm.). In the alcoholic specimen
R = about 215 mm.

Mortensen and Koehler have added considerably to Bell’s original
description and the only further points to be made here concern the
paxillae and the relative size of the actinal intermediate spines.

A strip of the dorsal skin with paxillae from the midradial base of
a ray of the alcoholic specimen shows considerable overlapping of the
bases of consecutive dorsal plates, although this may have been exag-
gerated by the contraction under preservation. In dorsal view the
bases of the plates are seen to have much more prolonged lobes than
in T. subinermis. The height of the columns relative to their minimum
width (4 or 5:1) is also greater than in T. subinermis, but comparable
to the proportions found in T. vestitus.

The smaller type, with R = 138 mm., has the paxillar columns only
two-and-a-half times as high as wide. Its actinal intermediate spines
are already large and overlapping, averaging 3.5 mm. in length. The
two larger specimens have these spines 5 to 7 mm. long. In Koehler’s
specimen from Ascension and Mortensen’s from St. Helena with R
about 65 mm. the spines are just beginning to make their appearance
but do not much, if at all, exceed their breadth in length.

It therefore seems that in T. magnificus the development of the
actinal intermediate spines is accelerated so that they first appear when
R is about 60 mm., whereas in T. vestitus they begin to develop only
when R is about 150 mm.

Specimens examined.—The type and paratype and a third specimen
also collected by J. C. Melliss at St. Helena.

TETHYASTER GRANDIS (Verrill)
Plates 11, 12; text figure 1, D

Sideriaster grandis VERRILL, 1890, p. 220, pl. 30, figs. 8, 8a, 8b (description;
Albatross station 2378) ; 1914, p. 21 (discussion) ; 1915, p. 192, pl. 12, figs.
5-5) (redescription, with the original figures republished).

Diagnosis —Enlarged spines on the inferomarginals narrow and
sharp-pointed, in some specimens a few also on the actinal inter-

18 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. I22

mediate plates; columns of the paxillae low, not over twice as high
as thick; no abactinal pedicellariae (on any of the specimens ex-
amined) ; rays broad at the base, tapering, the width at base equal to,
or greater than, r; first series of actinal intermediate plates reaching
to outer fourth of the ray, second to about the middle; superomargi-
nals up to 65 in number; paxillae with coarse, elongated granules ;
granulation of actinal intermediate plates coarse, coarsest in the center ;
conspicuous fascioles present; R up to 160 mm.

Description—The paxillae are compact and in contact, with a low,
stout pedicel about 1.5 mm. high, less than twice as high as thick, hav-
ing a rather strongly concave profile, crowned by a group of slightly
elongated, well-separated granules, usually 15-20 peripheral which are
about twice as long as thick, cylindrical with broadly rounded tips, and
usually 7 stouter and shorter central granules, one in the middle sur-
rounded by 6 others. The paxillae are largest on the disk, gradually
decreasing in size and with fewer and smaller central granules out-
wardly along the rays and toward the superomarginals. In the central
part of the rays they are arranged in regular longitudinal rows, on
the sides in rows at right angles to the superomarginals, three rows
to each two superomarginals. The median row on the rays is some-
times slightly larger than the others.

The madreporic body is large, approximately circular, Io mm. in
diameter, wholly exposed, flat, with numerous fine prominent radiat-
ing ridges. It is situated somewhat nearer the interradial border than
the center of the disk.

The marginal plates correspond throughout the ray. They are high
and narrow with shallow fasciolar channels between them which at
the base are about half as deep as the exterior face of the plates, be-
coming shallower distally. The superomarginals, 65 in the specimen
described, are high and narrow in the interradial arcs, 6 mm. high and
1.5 mm. wide, but gradually become wider, after the tenth being 6 mm.
high and 2.5 mm. wide, the relation of height to width then remain-
ing essentially the same to the arm tips. The superomarginals in the
interradial angles bear about 5 irregular columns of granules, those
in the middle of the plate the largest, the lateral about half as large;
on succeeding superomarginals the granules become smaller and usu-
ally more uniform, in 6-8 irregular columns, though often the central
granules are enlarged. The superomarginals are bordered on each
side by a somewhat irregular row of very fine, closely set spinelets
extending laterally over the fasciolar grooves.

The inferomarginals are confined to the actinal surface. They are

NO. II SEA-STARS—CLARK AND CLARK 19

everywhere of about the same size and shape as the superomarginals.
They bear a dorsoventrally median row of usually 4, sometimes 5,
tapering, pointed, and flattened spines about 4 mm. long which in-
crease slightly in length from the lowest to the uppermost and are ap-
pressed to the surface of the plate, each overlapping the base of the
one next above. On either side of this median row of spines, and also
between them, there are numerous much shorter, more or less flat-
tened, truncated spinules. The outer edges of the inferomarginals,
like those of the superomarginals, are bordered with an irregular row
of very fine, closely set spinules extending outward over the fascioles.

The terminal plate is of moderate size, heart-shaped, with the distal
end deeply notched and the proximal end slightly truncated.

The actinal intermediate areas are large. The innermost series of
plates extends to about the outer fourth of the ray (to about 25 mm.
from the tip), the second to well beyond the middle (to about 45 mm.
from the tip), and the third to about the tenth inferomarginal. Be-
tween the first inferomarginal and the second adambulacral the series
contains about Io plates. In each interradius there is a median un-
paired row of 1-5, commonly 3, plates. The actinal intermediate plates
bear 6-12 well-spaced, coarse, elongated granules, which are cylindri-
cal with broadly rounded ends, not over twice as high as thick, and are
bordered with an irregular row of fine spinules extending laterally
over the rather deep fascioles. On the interradial areas the elongated
granules are irregularly arranged, but on the rays they become aligned
in two or three irregular rows parallel to the axis of the ray. In some
specimens some of the actinal intermediate plates in the outer part of
the interradial areas may bear an enlarged, flattened, outwardly di-
rected appressed spine similar to those on the inferomarginals but
smaller.

The adambulacrals are at first broader than long, later becoming
squarish or even slightly longer than broad; the inner end forms an
obtuse angle ; in the middle of the inner edge there is a broad, strongly
flattened, recurved spine with a broadly rounded tip; on each side of
this, on the edge of the plate, are two flattened but straight and slightly
smaller spines. Behind these, on the actinal surface of the plate, there
is a row of usually 3 similar spines, and behind these two more, slightly
smaller. In the outer part of the ray the median spine in the row of
three gradually enlarges, and toward the tip of the ray the median
spine becomes long, stout, and conspicuous. The adambulacrals are
bordered with numerous fine spinules extending laterally.

Each mouth plate bears about a dozen short, stout, somewhat flat-
tened spines resembling those on the adambulacrals which they adjoin;

20 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

on the border abutting on the adambulacrals the mouth plates are
bordered with numerous very fine spinules.

The gonads extend almost to the tips of the rays.

The occurrence of pedicellariae in this species is very erratic. None
of the specimens have any pedicellariae on the abactinal surface, and
some have no pedicellariae at all. The one described has a few pedicel-
lariae consisting of three blades of slightly modified spines situated
on some of the interactinal plates from about the tenth adambulacral
to about the middle of the ray.

Type.—Presumably in the Yale University Museum.

Type locality—Albatross station 2378, off Mobile, Ala. (lat.
29°14’30” N., long. 88°09'30” W.) ; 124 meters; gray mud; Febru-
ary II, 1885.

Additional localities —Off Sombrero Light, Fla.; 110-128 meters ;
yacht Triton, 1951 (1 specimen).

M.V. Pelican, between Pensacola and Mobile; March 1, 1939
(1 specimen).

M.V. Oregon, southeast of Corpus Christi, Texas (lat. 27°25’ N.,
long. 96°13’ W.) ; 139 meters ; bottom temperature 60.5° F.; Novem-
ber 27, 1950 (10 specimens).

Off Tamaulipas, Mexico (lat. 24°10’ N.); 64-67 meters; Hilde-
brand, March 1951 (1 specimen).

Geographical range-—Known only from the Gulf of Mexico.

Bathymetrical range —From 67 ( ?64) to 139 meters.

Remarks.—In a specimen with R = 145 mm. some of the supero-
marginals in the second quarter of the rays carry small pedicellariae
at one or both of the lower angles. The inferomarginals from about
the fifth outward carry mostly two pedicellariae, one at each upper
angle, occasionally three or only one. The intermediate plates of the
inner row from about the eleventh to about the middle of the ray carry
usually two pedicellariae, one at each outer angle, occasionally only
one. The pedicellariae have usually three, rarely two or four, subequal
valves which resemble short spines with a swollen tip.

A small specimen from off Tamaulipas, Mexico, with R = 58 mm.
and r = 18 mm., is in general similar to the one described. The en-
larged spines on the inferomarginals are apparently just beginning to
appear. They are mostly about twice as high as the maximum diame-
ter, which is usually halfway to the tip, and are stumpy, subconical,
slightly flattened, with a subacute tip; a few have acute tips; some are
circular in cross section, and some are simply much enlarged granules.
There are no pedicellariae.

Specimens exanined.—All known specimens except the type.

NO. ITI SEA-STARS—CLARK AND CLARK 21

TETHYASTER SUBINERMIS (Philippi)
Text figures I, a, 2, a-c

Asterias subinermis Puttippi, 1837, p. 193 (description; Sicily) —LAMARCcK,
1840, p. 258 (from Philippi) —MULLER and TroscHEt, 1840, p. 324 (listed).
—PreyER, 1886, p. 32 (Naples; rare in about 100 fathoms).

Astropecten subinermis MULLER and TROSCHEL, 1842, pp. 74-75 (Sicily ).—Sars,
1850, p. 48 (Messina; 100 fathoms).—DuyarpiIn and Hupé, 1862, p. 425
(coasts of Sicily)—Prrrier, 1875, p. 369; 1876, p. 289 (Nice; Algeria;
Mediterranean).—StupeEr, 1884, p. 46 (off Liberia, lat. 4°40’ N., long.
9°10'06” W., 59 fathoms).—Carus, 1885, pp. 90-91 (summary of localities).
—PreveEr, 1886, p. 32 (Naples; rare in about 100 fathoms) —Cu£nort, 1888,
p. 134 (Banyuls)—Cotompo, 1888, pp. 47, 66 (Naples)—Sruper, 1889,
p. 28 (lat. 4°40.1' N., long. 9°10.6’ W., 108 meters).—von MARENZELLER,
1895, pp. 125, 127, 145 (Adriatic Sea, east of Pelagosa, lat. 42°23'00” N.,
long. 16°21’59” E., 131 meters, sand and mud).—NosrkE, 1903, p. 155
(Setubal) ; 1904, p. 133 (Setubal).

Astropecten crenaster (part) DujarpIn and Hupf, 1862, p. 414 (according to
Cuénot, 1912).—?FIscHER, 1869, p. 364.

Astropecten aranciaca FiscHeEr, 1860, p. 363 (not of Linné = subinermis accord-
ing to Cuénot, 1912) (Bassin d’Arcachon).

Archaster subinermis PERRIER, 1878, pp. 33, 57, 88 (Mediterranean).

Goniopecten subinermis PERRIER, 1885, p. 71.

Plutonaster (Tethyaster) subinermis SLAvEN, 1889, pp. 82, 83, 101, 102, 722.

Tethyaster subinermis PERRIER, 1894, p. 323; (Talisman station 5, Baie de Cadix,
lat. 36°26’ N., long. 8°47’ W., 60 meters, mud and shells; station 15, coast of
Morocco, lat. 33°57’ N., long. 10°47’ W., 1,283-1,425 meters, mud, coral;
station 66, off Cape Bojador, Morocco, lat. 26°13’ N., long. 17°10’ W., 175
meters, mud, coral). —KoEHLER, 1896b, pp. 450, 451 (Caudan, lat. 45°18’ N.,
long. 5°23’ W., 180 meters; lat. 45°52’ N., long. 6°03’ W., 250 meters; lat.
46°40’ N., long. 6°30’ W., 300 meters; Talisman, Baie de Cadix and coast of
Morocco, 60-1,425 meters); 1806a, pp. 56, 124 (Caudan station 17, lat.
45°18’ N., long. 5°23’ W., 180 meters, gravel and sand; station 20, lat.
45°52’ N., long. 6°03’ W., 250 meters, mud; station 27, lat. 46°40’ N., long.
6°30’ W., 300 meters, mud).—Perrier, 1896, p. 50 (Bay of Biscay, station
44, 166 meters; station 46, 155 meters)—KoEHLER, 1021, p. 54, fig. 40
(range); 1924, p. 200, pl. 7, fig. 4 (range).—MorTENSEN, 1925a, p. 178
(Atlantic coast of Morocco).—CuEnot, 1927, p. 295 (from Cuénot, 1912).—
KoEHLER, 1930, figs. I-3 (principally Mediterranean; Portugal; Cadix;
coasts of Morocco and Liberia).—RIvERA, 1930, p. 105, fig. 4, p. 106 (Cadiz).
—CuMANO, 1934, p. 138 (north of Berlengas )—NobereE, 1938, p. 55 (LeixGes,
Bacia do Tejo, Sezimbra), pl. 30 (apparently from Ludwig, 1897), p. 105
(west of Sezimbra) ; TorToNEsE, 10947a, p. 18 (Rodi [Rhodes] ).—Mapsen,
1950, p. 186 (Atlantide station 120, lat. 2°09’ N., long. 9°27’ E., 650-260
meters ; station 163, lat. 13°43’ N., long. 17°23’ W., 65-89 meters; about lat.
30°30’ N., long. 10° W., 100 to 120-500 meters).

Plutonaster subinermis Lupwtc, 1897, pp. 105-118, pl. 1, figs. 1, 2, pl. 6, figs. Io-
24 (detailed description; range).—Lo Branco, 1899, p. 473 (Gulf of Naples,
very rare, on muddy bottoms, rarely on bottoms of other types).—KoEHLER,
1909a, p. 7; 1900b, p. 22 (Princesse-Alice station 1447, lat. 45°21' N., long.

22 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

2°39’ W., 130 meters, fine sand, July 23, 1903).—-CUENOT, I912, pp. 28, 109
(range) .—GoT0, 1914, p. 3590.—CUENOT, 1927, p. 205.—NoprE, 1931, figs. 42,
43, p. 62 (probably from Ludwig).

Thetyaster subinermis NoprE, 1931, p. 62 (west coast of Portugal).

Tethyaster Tortonese, 1947b, p. 888 (Rhodes).

Diagnosis ——Enlarged spines on the inferomarginals narrow and
sharp-pointed, none on the actinal intermediate plates ; columns of the
paxillae low, not over twice as high as thick; no abactinal pedicel-
lariae; rays broad at the base, tapering, width at base equal to, or
greater than, r; first series of actinal intermediate plates to outer
fourth of ray, second to about the middle; superomarginals short,
68-85 in fully developed individuals; granulation of superomarginals
and actinal intermediate plates fine, uniform, and crowded; size large,
R up to 275 mm.

Description—This species was described and figured in detail by
Ludwig (1897, p. 105).

Type.—We have no information regarding the type.

Type locahiy.—Sicily.

Geographical range——From the Bay of Biscay (lat. 46°40’ N.)
south to the Gulf of Guinea, off Spanish Guinea (lat. 2°09’ N., long.
9°27’ E.) ; Mediterranean, east to Rhodes in the Aegean Sea.

Bathymetrical range—From about 50 to about 1,400 (possibly
1,425) meters.

Remarks.—In a specimen from Algiers (B.M. No. 1947.6.24.1)
with R = 110 mm. there are 73 superomarginals. In a specimen from
Naples (U.S.N.M. No. E.8001) with R = 86 mm., r = 21 mm., there
are 72 superomarginals; in this specimen the mouth plates are fol-
lowed by I pair of plates in three interradii, and by 2 pairs in two;
these are followed by a midinterradial unpaired row of 5 plates in
three interradii, of 4 in two, that reach to the suture between the
interradial pair of inferomarginals. In Ludwig’s figure this unpaired
median row consists of 5 plates, but reaches only to about two-thirds
the distance to the inferomarginals.

In a specimen from off Gambia (Alflantide station 163, lat. 13°43'
N., long. 17°23’ W., 69-89 meters) (B.M. No. 1950.3.18) with R =
72 mm, there are 62 superomarginals. In a specimen from off Spanish
Guinea (Altlantide station 120, lat. 2°09’ N., long. 9°27’ E., 650-260
meters) (B.M. No. 1950.7.3.26) with R= 71 mm. there are 68
superomarginals. In a specimen from Naples (B.M. No. 98.5.3.105-6)
with R = 50-57 mm. there are 48 superomarginals.

In individuals of this species the rays may be of slightly different
lengths.

NO. II SEA-STARS—CLARK AND CLARK 23

In some specimens from Naples the paxillae of the midradial row
on the rays gradually become enlarged in the outer part of the ray,
but this does not seem to be the case in specimens from Algiers or
from the Atlantic.

Specimens examined.—Five, listed above.

TETHYASTER PACEI (Mortensen)

Anthosticte pacei MortENSEN, 1925), p. 147, fig. 1, p. 148, pl. 8, fig. 3 (descrip-
tion; “Off South African Coast”).

Diagnosis.—Enlarged spines on the inferomarginals narrow and
sharp-pointed, none on the actinal intermediate plates ; columns of the
paxillae low, not over twice as high as thick; no abactinal pedicel-
lariae (in the single known specimen) ; rays narrow, width at base
markedly less than r; first series of actinal intermediate plates to
about the outer third of the ray, second only in the proximal third ;
ies E20 Sarl.

Type.—tIn the Zoological Museum, Copenhagen, Denmark.

Type locality—“Off South African Coast.”

Remarks.—This species is known only from the type specimen,
which we have not seen.

TETHYASTER AULOPHORA (Fisher)

Anthosticte aulophora FISHER, 1911, p. 417 (description; Albatross station

5420); 1910, p. 140, pl. 17, fig. 1, pl. 18, fig. 2, pl. 10, fig. 2, pl. 38, fig.
3, pl. 30, figs. 1, 1a-d (redescription).—MorTENSEN, 1925b, p. 148 (com-
parison with A. pacei).

Diagnosis —Enlarged spines on the inferomarginals narrow and
sharp-pointed, none on the actinal intermediate plates ; columns of the
paxillae slender, high, about four times as high as thick; most of the
paxillae with a pedicellaria of 2-4 valves; R = 162 mm.

Type.—lIn the U. S. National Museum (No. 28656).

Type locality—Albatross station 5420, between Cebu and Bohol,
Philippines (lat. 9°49'35” N., long. 123°45’00” E.); 232 meters;
bottom temperature 59° F.; March 25, 1900.

Remarks.—This species is known only from the type specimen
which we have examined.

BIBLIOGRAPHY

Bett, F, JEFFREY.
1882. Description of a new species of the genus Archaster from St. Helena.
Ann. Mag. Nat. Hist., ser. 5, No. 48, pp. 440-441 (December 1881).

24 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

Carus, J. V.
1885. Prodromus faunae Mediterraneae, vol. 1. Stuttgart.
Caso, MAria ELENA.
1947. Estudio sobre astéridos de México; descripcion de una nueva especie
del género Moiraster de Santa Rosalia, Golfo de California. An.
Inst. Biol., vol. 18, No. I, pp. 225-231.
CxiarkK, Austin H.
1916. Six new starfishes from the Gulf of California and adjacent waters.
Proc. Biol. Soc. Washington, vol. 29, pp. 51-62, April 4.
1950. A monograph of the existing crinoids. U. S. Nat. Mus. Bull. 82, vol.
1, the Comatulids, Pt. 4c, Superfamily Tropiometrida (the families
Thalassometridae and Charitometridae). pp. i-vii, 1-383, pls. 1-32.
Cotomso, A.
1888. La fauna sottomarina del Golfo di Napoli. Rivista marittima, Roma
(1887).
CueEnort, L.
1888. Contribution a l’étude anatomique des astérides. Arch. Zool. Exp.
et Gén., ser. 2, vol. 5, suppl.-bis, pp. 1-144, pls. 1-9.
1912. Contributions 4 la faune du bassin d’Arcachon. V, Echinodermes.
Bull. Stat. Biol. d’Arcachon, 14° année (I1QII-1912), pp. 17-116.
1927. Contributions a la faune du bassin d’Arcachon. IX, Révue générale de
la fauna et bibliographie. Soc. Sci. d’Arcachon, Stat. Biol., Bor-
deaux, vol. 24, fasc. 2™°, pp. 229-305.
CumaANo, H.
1934. Contribuigao para o estudo da fauna equinologia portuguesa. Arq.
Mus. Bocage, No. 5, pp. 137-141. Lisboa.
Dujardin, F., and Hup®, H.
1862. Histoire naturelle des zoophytes, Echinodermes, pp. 1-627, pls. 1-10.
FISCHER, PAUL.
1869. Echinodermes de la Gironde et du sud-ouest de la France. Act. Soc.
Linn. Bordeaux, vol. 27 (ser. 3, vol. 7), pp. 358-376.
FISHER, WALTER K.
1911. New genera of starfishes from the Philippine Islands. Proc. U. S.
Nat. Mus., vol. 40, pp. 415-427, May 17.
1919. Starfishes of the Philippine Seas and adjacent waters. U. S. Nat.
Mus. Bull. 100, vol. 3, pp. i-xi, 1-712, pls. 1-156.
Goto, SEITARO.
1914. A descriptive monograph of Japanese Asteroidea. Journ. Coll. Sci.,
Imp. Univ. Tokyo, vol. 29, Art. 1, pp. 1-808, pls. 1-19, December 17.
Grecory, J. W.
1900. In F. A. Bather, A treatise on zoology, ed. by E. Ray Lankester, pt. 3,
The Echinoderma, pp. i-viii, 1-344, 309 text figs. London.
GruBe, ApoLPH E.
1840. Actinien, Echinodermen und Wiirmer des Adriatischen und Mittel-
meers. K6nigsberg.
KoEHLER, RENE.
1896a. Resultats scientifiques de la campagne du “Caudan” dans le Golfe de
Gascogne. Ann. Univ. Lyon, vol. 26, pp. 33-127, pls. 1-4. [Also
published separately, Paris, 1896.]

NO. II SEA-STARS—CLARK AND CLARK 25

1896b. Rapport préliminaire sur les échinodermes. Dragages profonde ex-
écutés a bord du “Caudan” dans le Golfe de Gascogne (Aoit-
Septembre 1895). Rev. Biol. Nord France, 7 année, pp. 439-496.
Lille.
1908. Astéries, ophiures et échinides de l’expédition antarctique nationale
écossaise. Trans. Roy. Soc. Edinburgh, vol. 46, pt. 3, pp. 529-649,
pls. 1-16, December 16.
1909a. Astéries recueillies par l’investigator dans l’Océan Indien. I, Les
astéries de mer profonde. Echinoderms of the Indian Museum,
Pt. 5, Asteroidea (1). Calcutta.
1909b. Echinodermes provenant des campagnes du yacht Princesse-Alice
(astéries, ophiures, échinides et crinoides). Resultats des Cam-
pagnes Scientifiques Accomplies sur son Yacht par Albert I**
Prince Souverain de Monaco, fasc. 34, pp. 1-317, pls. 1-32.
1921. Faune de France. I. Echinodermes, pp. 1-210, figs. 1-153.
1924. Les échinodermes des mers d’Europe, vol. 1, pp. i-xiii, 1-362, pls. 1-9.
1930. Faune et flore de la Méditerranée. Commission internationale pour
l’exploration scientifique de la mer Méditerranée, No. 11, unnum-
bered loose-leaf page.
Lamarck, J. B. P. A. de.
1840. Histoire naturelle des animaux sans vertebres, ed. 2, vol. 3. Paris.
Lo Branco, SALVATORE.
1899. Notizie biologiche riguardanti specialments il periodo di maturita
seesuale degli animali del golfo di Napoli. Mitt. Zool. Stat. Neapel,
vol. 13, Heft 4, pp. 448-573, April 21.
Lupwic, HuBERT.
1897. Die Seesterne des Mittelmeeres. Fauna und Flora des Golfes von
Neapel. Monogr. 24, pp. i-x, 1-491, pls. 1-12.
LUTKEN, CHRISTIAN FREDERICK.
1859. Bidrag til Kundskab. om de ved Kysterne af Mellen- og Syd-
Amerika levende Arten af Sdstjerner I. Naturhist. Foren. Vidensk.
Medd. Kj¢benhavn.
MapseENn, F. JENSENIUS.
1950. The echinoderms collected by the Atlantide-Expedition 1945-46. At-
lantide Rep. No. 1, pp. 167-222, pls. 14-16.
MortENsEN, TH.
1925a. Echinodermes du Maroc et Mauritanie. Bull. Soc. Sci. Nat. Maroc,
vol. 5, Nos. 4-5, pp. 178-187, pls. 32-34.
1925b. On some echinoderms from South Africa. Ann. Mag. Nat. Hist.,
ser. 9, vol. 16, No. 91, pp. 146-154; fig. 1, p. 148; pl. 8, fig. 3, July.
1933. The echinoderms of St. Helena. Vidensk. Medd. Dansk Naturh.
Foren., vol. 93, pp. 401-472, 29 text figs., pls. 20-22, October 24.
Mutter, J., and Troscuet, F. H.
1840. Ueber die Gattungen der Asteriden. Arch. Naturg., Jahrg. 6, vol. 1,
pp. 318-326.
1842. System der Asteriden, pp. i-xx, 1-134, pls. I-12.
Nosre, AUGUSTO.
1903. Subsidios para o estudo da fauna marinha do sul de Portugal. Ann.
Sci. Nat. Porto, vol. 8, pp. 152-160.

26 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

1904. Fauna Portugueza. Ann. Acad. Polytechnica, Porto, 1903-1904, pp.
86-146.

1931. Echinodermes de Portugal, pp. 1-176, pls. 1-14, 74 text figs. Inst.
Zool. Univ. Porto.

1938. Echinodermes de Portugal, ed. 2, pp. 1-215, pls. 1-42, Porto.

PERRIER, EDMOND.

1875. Reévision de la collection de stellérides du Muséum d’histoire naturelle
de Paris. Arch. Zool. Exp. et Gén., vol. 4, pp. 265-450.

1876. Révision des stellérides du Muséum. Arch. Zool. Exp., ser. 2, vol.
5, Pp. 209-304.

1878. Etude sur la répartition géographique des Astérides. Nouv. Arch.
Mus. Hist. Nat., Paris, ser. 2, vol. 1, pp. 1-108.

1885. Premiére note préliminaire sur les échinodermes recueilles durant les
campagnes de dragages sousmarins du Travailleur et du Talisman.
Ann. Sci. Nat., ser. 6, vol. 19, Art. 8, pp. 1-72.

1894. Expéditions scientifiques du Travailleur et du Talisman. Echinodermes.
I. Stellérides, pp. i-iv, 1-431, pls. 1-26. Paris.

1896. Contribution a l’étude des stellérides de l’Atlantique nord (Golfe de
Gascogne, Acores, Terre-Neuve). Resultats des Campagnes
Scientifiques du Prince de Monaco, fasc. 11, pp. viii + 73, pls. i-ic,
November.

Puriprl, A.

1837. Ueber die mit Asterias awranciaca verwandten und verwechselten
Asterien der sicilianischen Kiiste. Arch. Naturges., vol. 3, pp. 193-
194.

PREYER, W.

1886. Ueber die Bewegungen der Seesterne. Mitt. Zool. Stat. Neapel, vol. 7,

Pp. 27-127.
RIVERA, VICTORIANO.

1930. Algunos astéridos de Espafia. Bol. Real Soc. Espafiola Hist. Nat.,

Madrid, vol. 30, pp. 101-112.
Sars, M.

1859. Bidrag til Kundskaben om Middelhavets Littoral-Fauna. Nyt Mag.

Naturv., vol. 10, pp. 1-99.
Say, THOMAS.

1825. On the species of the Linnaean genus Asterias inhabiting the coast of
the United States. Journ. Acad. Nat. Sci. Philadelphia, vol. 5, pp.
141-154.

SLADEN, W. PErcy.

1889. Report on the Asteroidea collected by H.M.S. Challenger during the
years 1873-1876. Report on the scientific results of the voyage of
H.M.S. Challenger during the years 1873-76. Zoology, vol. 30.

Stuper, TH.

1884. Verzeichniss der wahrend der Reise S.M.S. “Gazelle” um die Erde
1874-1876 gesammelten Asteriden und Euryaliden. Akad. Wiss.,
Berlin, Abh. 2, pp. 1-64, 5 pls.

1889. Die Forschungsreise S.M.S. Gazelle in d. Jahren 1874 bis 1876. III
Thiel. Zoologie und Geologie, pp. i-vi, 1-322, pls. i-xxxiii. Berlin.

NO. II SEA-STARS—CLARK AND CLARK 27

ToRTONESE, ENRICO.
1947a. Note intorno alla fauna e flora marina dell’Isola di Rodi (Mar
Egeo). Boll. Pesca, Piscicoltura e Idrobiologia, Roma, Ann. 23,
vol. 2 (n. s.), fase. I, pp. 13-20, January-June.

1947b. Biological investigations in the Aegean Sea. Nature, vol. 159, No.

4052, pp. 887-888, June 28.
VERRILL, A. E.

1866. On the polyps and echinoderms of New England, with descriptions of
new species. Proc. Boston Soc. Nat. Hist., vol. 10, pp. 333-357.

1895. Distribution of the echinoderms of northeastern America. Amer.
Journ. Sci., vol. 49, pp. 127-141, 199-212, February-March.

1899. Revision of certain genera and species of starfishes with descriptions
of new forms. Trans. Connecticut Acad., vol. 10, pp. 145-234, pls.
24a-30, August.

1914. Revision of some genera and species of starfishes, with descriptions
of a few new genera. Ann. and Mag. Nat. Hist., ser. 8, vol. 14, No.
79, PP. 13-22, July.

1915. Report on the starfishes of the West Indies, Florida, and Brazil.
Bull. State Univ. Iowa, Bull. from Lab. Nat. Hist., vol. 7, No. 92,
(n. s. No. 1), pp. 1-232, pls. 1-29, March 20.

VON MARENZELLER, FMI,

1895. Berichte der Commission fiir Tiefseeforschungen. XVI. Zoologische
Ergebnisse V. Echinodermen gesammelt, 1893, 1804. Denkschr.
Akad. Wien, Math.-Naturw. Cl., vol. 62, pp. 123-148, 1 pl.

ZIESENHENNE, Frep C,

1937. The Templeton Crocker Expedition. X. Echinoderms from the west
coast of Lower California, the Gulf of California, and Clarion
Island. Zoologica, vol. 22, pt. 3, No. 15, pp. 209-239, text figs. 1 and
2, October 7.

*
1

Nai

ws |

ane
,

hoe ats sty ;
) , r iy ay

}
ae i
ft ‘

SMITHSONIAN MISCELLANEOUS COLLECTIONS WO, ava, INKS atl, TPL al

Tethyaster canaliculatus (A. H. Clark), the type specimen from Albatross station 20908,
Gulf of California, in 73 meters; abactinal view, X 2. (U.S.N.M. No. 36951.)

SMITHSONIAN MISCELLANEOUS COLLECTIONS VOE. 122; NO: 11, PES

a
38 = lh ha! 2.

: bsp
rage erga?

Tethyaster canaliculatus (A. H. Clark), the type specimen from Albatross station 2908,
Gulf of California, in 73 meters; actinal view, X 2. (U.S.N.M. No. 36951.)

SMITHSONIAN MISCELLANEOUS COLLECTIONS VOLE. 222) NOS 1) PES 3

i od
PUTAS oe!
4
Lee, Se ad

“as

bet:

Tethyaster canaliculatus (A. H. Clark), the specimen with R95 mm. from Zaca
station 142, D-3, Santa Inez Bay, Gulf of California, in 73 meters; actinal view, X 2.
(M.C.Z. No. 36232.)

SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122, NO. 11, PL. 4

Tethyaster canaliculatus (A. H. Clark): Upper, the specimen with R=,t mn
from Zaca station 136, D-14, Arena Bank, Gulf of California, in 82 meters: actinal
view. (M.C.Z. No. 3447.) Lower, the specimen from Zaca station 146, D-1, Santa
Inez Bay, Gulf of California, in 73 meters; abactinal view. (M.C.Z. No. 3448.) Both
figures < 2.

SMITHSONIAN MISCELLANEOUS COLLECTIONS VOLES 1227 3NOs 22, PES

Tethyaster vestitus (Say), specimen from off Diamond Shoal, North Carolina, in 44
meters; abactinal view, natural size. (U.S.N.M. No. E.8ooo.)

SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 1227, NOP da Riese

Tethyaster vestitus (Say), specimen from off Diamond Shoal, North Carolina, in 44
meters; actinal view, natural size. (U.S.N.M. No. E.8000.)

SMITHSONIAN MISCELLANEOUS COLLECTIONS WOE. aban Nol Mabe Teles 7/

Tethyaster vestitus (Say), specimen from off Puerto Rico in 146-329 meters; abactinal
view, natural size. (U.S.N.M. No. E.3963.)

SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL i227 INO a Plas

Tethyaster vestitus (Say), specimen from off Puerto Rico in 146-329 meters; actinal
view, natural size. (U.S.N.M. No. E.3963.)

Tl bs)

NO.

‘he >

abactinal view,

VOL. 122,
ov

68.6.15.1.)

No.

(B.M.

. ot .
Stee bewtng ca ws

natural size.

hes

&
*

a
o
=
vo
—
v
3)
UY}
a
S
=}
u
Gat
v
a
jan
cal
~
ro)
=
v
on
wu
iss)
—
v
ae)
fe)
—~
=
=
vo
wy
—
YS
6
S
Se
=
>
DS
8
=
nm
a
~
~
Ka
S
(=9)
~
—
~~
aS
sy

SMITHSONIAN MISCELLANEOUS COLLECTIONS

SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL 1227 NOs sa Pie

Tethyaster magnificus (Bell), the larger cotype from St. Helena ; actinal view,
natural size. (B.M. No. 68.6.15.1.)

SMITHSONIAN MISCELLANEOUS COLLECTIONS VOES 22 INOS Tai Pies a

Tethyaster grandis (Verrill), from off Corpus Christi, Tex., in 139 meters ;
abactinal view, natural size. (U.S.N.M.)

VOLE. 122, NO. 11, (Pian

SMITHSONIAN MISCELLANEOUS COLLECTIONS

in 139 meters;

i, Tex., 1
M.)

ist
N

CUS:

1Ze€

, natural si

iew

errill), from off Corpus Chr
lv

randis (V
actina

thyaster

‘eC

oan

Sere

Persear

eee y

Pe ee

tw S

SMITHSONIAN MISCELLANEOUS COLLECTIONS
VOLUME 122, NUMBER 12

_ THE REPRODUCTION OF
COCKROACHES

(Wrre 12 Pirates)

BY
LOUIS M. ROTH

AND

EDWIN R. WILLIS

Pioneering Research Laboratories
U.S. Army Quartermaster Corps
Philadelphia, Pa.

RE INC RS ay

LSM VIG AIS

s\g0e SHON
er

Se
*)
y
§
ji
4

(Pustication 4148)

CITY OF WASHINGTON
PUBLISHED BY THE SMITHSONIAN INSTITUTION
JUNE 9, 1954

SMITHSONIAN MISCELLANEOUS COLLECTIONS
VOLUME 122, NUMBER 12

fe REPRODUCTION OF
COCKROACHES

(Wirth 12 PLATEs)

BY
LOUIS M. ROTH
AND

EDWIN R. WILLIS

Pioneering Research Laboratories
U. S. Army Quartermaster Corps
Philadelphia, Pa.

“S
Ax

S8eceece®

(PusticaTion 4148)

CITY OF WASHINGTON
PUBLISHED BY THE SMITHSONIAN INSTITUTION
JUNE 9, 1954

The Lord Baltimore Press

BALTIMORE, MD., U. 8 A.

fi REPRODUCTION OF ‘COCKROACHES*
By LOUIS M. ROTH anv EDWIN R. WILLIS

Pioneering Research Laboratories
U.S. Army Quartermaster Corps
Philadelphia, Pa.

(WITH 12 PLATES)

INTRODUCTION

Cockroaches are important for several reasons. As pests, many are
omnivorous, feeding on and defiling our foodstuffs, books, and other
possessions. What is perhaps less well known is their relation to the
spreading of disease. Several species of cockroaches closely associated
with man have been shown to be capable of carrying and transmitting
various microorganisms (Cao, 1898; Morrell, 1911 ; Herms and Nel-
son, 1913; and others). Recently there has been a resurgence of inter-
est in this subject, and some workers have definitely implicated cock-
roaches in outbreaks of gastroenteritis.

Antonelli (1930) recovered typhoid bacilli from the feet and bodies
of Blatta orientalis Linnaeus which he found in open latrines during
two small outbreaks of typhoid fever. Mackerras and Mackerras
(1948), studying gastroenteritis in children in a Brisbane hospital,
isolated two strains of Salmonella from Periplaneta americana (Lin-
naeus) and Nauphoeta cinerea (Olivier) that were caught in the hos-
pital wards. Graffar and Mertens (1950) isolated Salmonella typhi-
murium from Blattella germanica (Linnaeus) captured in a hospital
in Brussels. These latter workers were only able to check the epidemic
of gastroenteritis among children by ridding the hospital nursery of
cockroaches. Bitter and Williams (1949) have isolated three species
of Salmonella from the hind gut of P. americana captured in a hospi-
tal, private home, and sewer manholes.

It is significant that four strains of poliomyelitis virus have recently
been isolated from Periplaneta americana, Supella supellectilium
(Serville), and Blattella germanica, which were collected on the prem-

1 This study was made by the Army Quartermaster Corps as part of a re-
search program that includes the investigation of the biologies of insect pests of
economic and medical importance to the armed services.

SMITHSONIAN MISCELLANEOUS COLLECTIONS, VOL. 122, NO. 12

2 SMITHSONIAN. MISCELLANEOUS COLLECTIONS VOL. 122

ises of paralytic poliomyelitis patients (Syverton et al., 1952). In
addition to harboring bacteria and viruses, cockroaches also harbor
pathogenic protozoans and nematodes. The Surinam cockroach, Pye-
noscelus surinamensis (Linnaeus), is the vector of the eyeworm of
poultry (Fielding, 1926), and the American cockroach can carry, me-
chanically, hookworm of man (Porter, 1930) ; the latter species can
also transmit, experimentally, intestinal flagellates such as Giardia
from man to rat (Porter, 1918). Although these examples could be
multiplied, it is apparent that, as Bitter and Williams (1949) have
stated, tolerance of cockroaches around man’s habitations is unwar-
ranted ; it may even be dangerous.

There are about 450 genera and more than 3,500 species of cock-
roaches (Rehn, J. W. H., 1951). Practically nothing is known of the
biology of most species. Very little is known of the biologies of the
cockroaches associated with man, except for the more common pests
such as the German, American, and oriental cockroaches. Yet less
than 1 percent of the known species are domiciliary pests (Rehn,
J. A. G., 1945). This is a fertile field for future work.

Reproduction, enabling the individual to increase its kind many
times, is a vital factor in the biology of an insect species. Only rarely
do swarms of insects invade a locality from a distant point ; each com-
munity usually raises its own insect pests (Metcalf and Flint, 1939).
Cockroaches illustrate this principle perfectly. Hence the reproduc-
tion of cockroaches is a subject of more than academic interest. Re-
production is a phase of cockroach biology that demonstrates the di-
versity of behavior that has evolved in this relatively ancient group.
In the following pages we shall describe, among several species of
cockroaches, these aspects of reproduction: courtship, copulation, re-
productive organs and fertilization, parthenogenesis, the ootheca, ovi-
position and hatching, and egg parasites.

COURTSHIP

In general the courting behaviors of the species of cockroaches that
have been studied appear to be similar in many respects ; characteris-
tic differences, however, lend interest to the study of each additional
species. The writers (1952) have studied Blattella germamica, Blatta
orientalis, and Periplaneta americana and have analyzed the stimuli
involved in the courting behavior of the German cockroach. We could
not demonstrate distance attraction between males and females of
B. germanica; yet, when a sexually active male comes in physical con-
tact with the female he responds with a characteristic courting behav-

NO. I2 REPRODUCTION OF COCKROACHES—ROTH AND WILLIS 3

ior. The male turns around so that his terminal abdominal segments
are toward the female, and he raises both front and hind wings to an
angle of 45° to go° (pl. 1, fig. 1). In this way he exposes glandular
areas on his abdominal terga which emit a secretion that attracts the
female when she is close to him. A responsive female will feed (pl. 1,
fig. 2) on the male tergal-gland secretion, and as she does, the male
pushes backward and grasps the female genitalia. In B. germanica
the male must make contact with the female before he will court.
Mutual sparring with the antennae between the sexes and movement
by the female are important actions in stimulating the male to court.
The male courting response (i.e., raising the wings) is the overt ex-
pression of male sexual stimulation ; it can also be induced by stroking
or touching a receptive male’s antennae with antennae (pl. 1, fig. 1),
legs, abdomen, or wings, which have been removed from a female.

The raising of the male’s wings during courtship, or just prior to
copulation, apparently is characteristic of those species of cockroaches
in which the males have wings. Raising of the wings has been ob-
served in the three domestic species previously mentioned and also in
Leucophaea maderae (F.) (Sein, 1923); Blaberus craniufer Burm.
[= B. fusca Brunner (Rehn and Hebard, 1927)] (Saupe, 1928;
Nutting, 1953) ; Supella supellectiium (Roth, 1952) ; Blattella vaga
Heb. and Nauphoeta cinerea (Roth and Willis, unpublished data).

Chemical as well as mechanical stimuli are involved in the courting
behavior of Blattella germanica. This is shown by the fact that a
substance that is sexually stimulating to males can be isolated from
females. The cuticular surface of the cockroach is covered with a
freely exposed grease (Ramsay, 1935; Kramer and Wigglesworth,
1950). Presumably the sex substance is present in the cuticular
grease, because this material can easily be rubbed off from females
onto males to make the latter sexually stimulating to other males. The
available evidence indicates that sex discrimination by the male
German cockroach is mainly effected by contact chemoreception.

The sexual behavior of Blattella germanica, in terms of stimulus
response and releaser mechanisms, may be summarized as follows:

Male courts=raises wings (releaser =|

7
Male makes anten- :
mutual antennal fencing between sexes;

nal contact with|—>

chemical stimulus on the female; move-
the female
ment of the female)
Female feeds on the tergal-gland Copulation (releaser feeding by

secretion of the male (releaser—=|——*> the female on the male’s tergal-
tergal-gland secretion) gland secretion)

4 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

The male receives the female sex stimulus by means of receptors
that are present on his antennae and probably on his mouthparts.

It is highly probable that in those species of cockroaches in which
only the males have a distinctive, externally visible tergal gland, the
function of this gland is to entice the female into a position in which
mating can occur. Feeding by the female on the glandular secretion
or over the dorsal abdominal surface of the male cockroach has been
observed in Blattella germanica (Sikora, 1918; Wille, 1920; Roth
and Willis, 1952) ; Blatta orientalis (Roth and Willis, 1952) ; Ecto-
bius lapponicus Linnaeus and Ectobius sylvestris (Poda) (Koncek,
1924) ; Supella supellectilium (Roth, 1952) ; Blattella vaga, Eurycotis
floridana (Walker), Nauphoeta cinerea, and Leucophaea maderae
(Roth and Willis, unpublished data). The external appearance of the
tergal gland may vary considerably between species and genera (cf.
pl. 2, figs. 6-8); this structure has considerable taxonomic value
(Hebard, 1917; Rehn, J. A. G., 1931; Ramme, 1951). Frequently
the glandular area is a depression in one (pl. 2, figs. 7, 8) or some-
times two (pl. 2, fig. 6) of the abdominal tergites and has a mass of
secretory cells lying beneath the epidermis. Groups of setae or hairs
are often present (pl. 2, fig. 8). In B. orientalis the female moves her
mouthparts actively over the male’s dorsum; yet the source of the
male’s secretion (if any) is still unknown. Tergal glands are found
on the dorsum of the male, female, and nymphs of the oriental cock-
roach (Minchin, 1888), but they apparently have nothing to do with
sex behavior. The contents of these glands have the distinctive odor
of the oriental cockroach (Haase, 1889).

In Hawaii, Bridwell (1921), while walking in Palolo Valley at
night, saw 50 to 75 Periplaneta americana performing their “mating
dance” in the middle of the road. In this species the male sexual be-
havior is released by an odorous material secreted by the female. The
source of the sex attractant in the female is unknown, but it is a mate-
rial that readily rubs off from the female and can be perceived by the
male at a considerable distance. By keeping unmated females in con-
tainers lined with filter paper, we have collected the attractant on the
paper from which the active material was extracted in crude form
with petroleum ether. The attractive female odor alone suffices to
stimulate the male of P. americana to overt sexual activity. Males will
even attempt to mate with pieces of paper or glassware that have been
in contact with virgin females. Paper taken from jars containing
old, nonvirgin females did not stimulate males, indicating that the sex
attractant is produced chiefly by unmated females (Roth and Willis,

NO. 12 REPRODUCTION OF COCKROACHES—ROTH AND WILLIS 5

1952). However, it is possible that a mated female may again become
attractive (i.e., secrete the attractant) sometime during her lifetime.
Neither we nor Gupta (1947) have observed males of the American
cockroach courting the female prior to copulation, as is done by the
German and oriental males. The male American cockroach is much
more direct in his approach, and the female appears to be relatively
passive ; movement of the female’s mouthparts over the male’s dorsum
is not a necessary stimulus for the male to attempt to copulate, as in
the other two species.

Summarized, the sexual behavior of Periplaneta americana is as
follows:

o Male raises wings |
Male searches | =

Male in the Bee formate (re. and attempts to Copulation
H es
vicinity of|—> leaser = sex-|—> clasp female’s gen- mae (releaser =
the female Saokot teat tee italia (releaser = sex-odor
LA) sex-odor from fe- from female)
male)

In the behavior of Blattella germanica there was a succession of
releasers that alternately brought forth responses from both partners
before culmination of the sexual act. In contrast the sex odor from
the female seems to be the only mechanism involved in releasing a
chain of responses by the male of P. americana that ends in copulation.
However, the female of P. americana may or may not be receptive
to the male’s advances. Perhaps there is an as-yet-undetermined
releaser that regulates the female’s response.

The males of some species of cockroaches perform characteristic
body movements during courting behavior. The male of Leucophaea
maderae stands near the female and rapidly moves his body up and
down. According to Sein (1923) the male raises the anterior section
of his body and strikes his abdomen against the ground producing a
prolonged tapping sound. However, we have seen males of this
species move the anterior parts of their bodies up and down rather
than their abdomens. The male of Blaberus craniifer raises himself
on his legs and makes trembling movements with the abdomen (Saupe,
1928) ; we have observed the male of this species behave in a similar
manner and also butt the female with his head or pronotum. The
wingless male of Eurycotis floridana stands near the female, repeatedly
vibrates his body from side to side, and extends his abdomen slightly
revealing the light-colored, intersegmental membrane between the
sixth and seventh tergites (pl. 1, fig. 3) ; the female then behaves as
described earlier for the other species, applying her mouthparts to the
male’s dorsum starting near the end of the abdomen (pl. 1, fig. 4) and

) SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

working up to the first abdominal tergite (pl. 1, fig. 5), on which is
located a small glandular area bearing a patch of setae.

COPULATION

The terminal abdominal segments are modified to form the male ex-
ternal genitalia, which consist of genital lobes or phallomeres that are
associated with the opening of the ejaculatory duct. These structures
are described by Snodgrass (1937) and van Wyk (1952). Certain
phallomeres of the adult cockroach form highly complex structures
with horny processes or hooks. In the female the external genitalia
include the ovipositor and associated sclerites and the openings of the
oviduct, accessory glands, and spermathecae (Snodgrass, 1937).

As we mentioned earlier, while the female Blattella germanica feeds
on the male’s glandular secretion, he pushes his abdomen backward
so that the female is directly above him. This is the position (pl. 1,
fig. 5) just prior to copulation, and it is assumed by most, if not all,
species of cockroaches. We have seen this method of initiating copula-
tion, in the female superior pose, in these genera: Blatiella, Blatta,
Periplaneta, Eurycotis, Nauphoeta, Supella, and Leucophaea. Saupe
(1928) and Nutting (1953) observed that the female of Blaberus
crantfer also straddles the abdomen of the male just prior to success-
ful copulation. Because this behavior is so similar among different
genera, we are convinced that it is a regular feature in the copulation
of Leucophaea although we have seen it only once. However, Pessoa
and Correa (1928) state that the male of L. maderae “draws near
[the female] and turns his body in an opposite direction, to that of
the female, placing the posterior extremity of his abdomen against
[the] posterior extremity of the abdomen of the female.” We cannot
reconcile their statement with our observations and those of van Wyk
(1952) who saw “that the male carries the female on his back at the
beginning of copulation with their heads in the same direction and
the venter of the female resting on the dorsum of the male and that
they later assume an end to end position.”

The relationships of the external genitalia during copulation have
been studied in Blattella germanica (Khalifa, 1950), Periplaneta
americana (Gupta, 1947), and Polyzosteria limbata Burm. (Chopard,
1919). As the male German cockroach pushes backward under the
female, he extends his hooked left phallomere. This appendage clasps
a large sclerite located near the female’s ovipositor. If a hold is se-
cured on the sclerite, the male moves out from under the female, and
the couple assume the opposed position in which their heads face in

NO. 12 REPRODUCTION OF COCKROACHES—ROTH AND WILLIS 7,

opposite directions. This is the copulating position (pl. 3, figs. 9-14)
assumed by all species of cockroaches in which the act has been ob-
served ; we have also seen Nauphoeta cinerea in this position, and it
has been observed in Polyzosteria limbata (Chopard, 1919) and Par-
coblaita pensylvanica (Rau, 1940b). Once in the final opposed posi-
tion, two lateral hooks lying on either side of the anus of the male
hold the ovipositor near its base. A small crescentic sclerite, which lies
on one side near the right phallomere, grips the ovipositor firmly in
a medial position (Khalifa, 1950). In successful matings cockroaches
remain in copula in the end-to-end position for at least 30 minutes.
Usually copulation lasts more than an hour, and Nutting (1953)
noted many pairs of Blaberus cranifer that remained joined for 4
hours or more. Statements in the literature to the effect that copula-
tion is rapid, lasting only a few seconds or less, were based on
observations of unsuccessful matings.

INTERNAL REPRODUCTIVE ORGANS AND FERTILIZATION

The internal genital organs of male cockroaches consist of a pair of
testes, genital ducts, accessory genital glands, seminal vesicles, and a
phallic gland (Snodgrass, 1937). In Blattella germanica each testis
consists of four rounded sacs or vesicles which open into a common
genital duct. Each vesicle is divided into several zones which contain
sex cells in various stages of development; the spermatozoa are con-
tained in the zone nearest the genital duct (Wassilieff, 1907). The
testes of Blatta orientalis mature at the end of nymphal development
and atrophy in the adult; hence the spermatozoa must be stored in
the seminal vesicles before the testes degenerate (Snodgrass, 1937).
The diploid number of chromosomes (including the X chromosome)
has been determined during spermatogenesis for several species of
cockroaches; Loboptera decipiens Germ. has 34, Blattella germanica
24, Periplaneta americana 34, Periplaneta australasiae (Fabricius) 28,
Blatta orientalis 48, and Blaberus fusca [=B. craniifer] 74 (Suoma-
lainen, 1946). All species of cockroaches that have been investigated
are XO in the male (White, 1951).

The sperm of cockroaches is transferred to the female by means of
a capsule or spermatophore (pl. 4, fig. 17) formed from the secre-
tions of the male accessory sex glands. Spermatophores have been
found in males of Blatta orientalis (Zabinski, 1933) ; Blattella ger-
manica (Khalifa, 1950; Roth and Willis, 1952) ; Periplaneta ameri-
cana (Gupta, 1947; Roth and Willis, 1952); Leucophaea maderae
(van Wyk, 1952) ; Eurycotis floridana and Nauphoeta cinerea (Roth

8 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

and Willis, unpublished data); and Blaberus craniifer (Nutting,
1953). Presumably spermatophores are produced by other blattids as
well. Nutting (1953) observed that the spermatophore of B. craniifer
may be retained in the female’s genital pouch for several days, one
female carrying her spermatophore intact for 5 days, whereas most
of the cockroaches previously noted retain the spermatophore for a
shorter period.

The accessory glands (“mushroom-shaped gland”) in the oriental
cockroach consist of 350 to 450 small, intermediate, and large-sized
tubes. Based on their staining reactions, Jurecka (1950) distinguished
6 types of tubes in males of Blatta orientalis; their period of most
active secretion occurs for several hours following metamorphosis
into the adult. This is followed by a resting period from the time
secretion ceases until copulation, at which time the secretions and the
spermatozoa are ejected and molded into a spermatophore. The sper-
matophore consists of a number of capsules full of spermatozoa
(Qadri, 1938); the female carries it 2 or 3 days then drops it
(Zabinski, 1933).

Van Wyk (1952) described the male accessory glands in Leucophaea
maderae. These are composed of approximately 30 to 40 tubules ar-
ranged in three groups. He assumes that each group of glands is
responsible for one of the three layers of the spermatophore. The
spermatophore of L. maderae remains in the genital chamber of the
female for about a day before it dries and drops from the female.

If the males of Blattella germanica are prevented from mating, the
larger accessory gland tubes (“‘utriculi majores”) become so distended
with their chalk-white secretion (pl. 4, fig. 15) that they may fill most
of the abdominal cavity. After copulation the tubes of the accessory
glands are almost emptied of secretion (pl. 4, fig. 16) (Roth and
Willis, 1952). The following description of spermatophore forma-
tion is taken from Khalifa (1950). The spermatophore in B. ger-
manica begins to form in the male as soon as the copulating pair are
securely hooked together. In this species the layers of the sperma-
tophore are formed from three protein secretions produced by dis-
tinctly different groups of accessory gland tubules. The walls of the
tubules consist of a layer of glandular cells surrounded by a muscular
layer. The accessory glands open into the ejaculatory duct. The se-
cretions from the various accessory glands pour into the pouch of the
ejaculatory duct, and when the spermatophore is completely devel-
oped, it distends the pouch. At one point in the formation of the sper-
matophore, sperm flow from each seminal vesicle into a milky middle
layer within the spermatophore ; each of the two sperm masses forms

NO. 12 REPRODUCTION OF COCKROACHES—ROTH AND WILLIS 9

a separate sac. Following formation, the completed spermatophore
descends the ejaculatory duct and is pressed by the male’s endophallus
against three sclerites lying on the left-hand side of the spermathecal
groove in the female serve for holding the spermatophore. The tip
surround the opening of the common oviduct, and the spermathecal
groove in the female serve for holding the spermatophore. The tip
of the spermatophore, which contains the openings of the sperm sacs,
is inserted into the spermathecal groove so that the two spermathecal
pores of the female come in direct contact with the two openings of
the sperm sacs. The spermatophore remains in the genital chamber
of the female for about 12 hours, during which time the sperm mi-
grate to the spermathecae. In B. germanica the sperm have to be
chemically activated before they leave the spermatophore; probably
the activating chemical originates from a pair of spermathecal glands
which are associated with the spermathecae of the female. The empty
spermatophore dries and shrinks and is eventually dropped by the
female. During her lifetime a female may copulate and receive a sper-
matophore more than once (Khalifa, 1950), and a male may also
copulate and produce a spermatophore more than once (Cros, 1942;
Roth and Willis, 1952). The work of the male is now done.

The internal reproductive organs of the female cockroach consist
of a pair of ovaries, oviducts, spermathecae, and specialized accessory
(colleterial) glands which produce the various secretions that go to
make up the odtheca or egg case (Snodgrass, 1937, 1952). In Peri-
planeta americana each ovary usually consists of eight ovarioles. Each
ovariole is made up of an elongated egg tube and a short pedicel
which connects the basal end of the egg tube to the oviduct. The an-
terior part of the egg tube consists of a germarium made up of odcytes
or incompletely formed eggs in the early stages of differentiation, and
the remainder or vitellarium contains odcytes in various stages of
growth. In the newly emerged adult female all of the oocytes are
relatively small, although a gradation in size is noticeable, the largest
being at the base (pl. 4, fig. 18). About 8 days after copulation the
basal odcyte reaches a final size of about 3 mm. in length; it is now
encased in a chorion, and is ready for deposition (pl. 4, fig. 19). After
oviposition of the basal odcyte (pl. 4, fig. 20), the next oocyte in line
completes its growth; this cycle results in a succession of mature eggs
about every 8 days (Gier, 1936). Actually by the time an oodtheca
is completely formed and deposited, the basal oocytes of the ovary are
already well developed (pl. 4, fig. 21). Scharrer (1943, 1946) found
that the development of the eggs of Leucophaea maderae, at least for
a certain period of time, is under the hormonal control of the corpora

Io SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. I22

allata, endocrine glands situated near the brain. Apparently the hor-
mone from these glands is not required for the reproductive activity
of the male (Scharrer, 1946).

The external genital structures of the female cockroach lie con-
cealed within a cavity at the end of the abdomen that is closed poste-
riorly by the apical lobes of the seventh sternum (Snodgrass, 1952).
This cavity is divided into a genital chamber, which lies proximal to
the base of the ovipositor, and the odthecal chamber, or vestibulum,
which is the posterior part of the cavity. The ovipositor is composed
of three pairs of fingerlike valvulae and two pairs of valvifers
(Brunet, 1951). The relation of the external genitalia to the forma-
tion of the odtheca is discussed below.

Dewitz (1885, 1886) described how sperm enter the eggs of Blatta
orientalis. The eggs become bent in the oviducts and pass singly into
the genital chamber where they approach the spermathecal opening.
As the egg passes over the sensory hairs that are found mainly around
the spermathecal pore, muscles of the spermatheca contract and force
out the sperm. The numerous micropyles, through which the sperm
enter, are found at the anterior pole of the egg and come in contact
with most of the sperm. The sperm appear to be attracted to the sur-
face of the egg and move clockwise rather than in a straight line. The
eggs are fertilized as they pass along the vestibulum.

PARTHENOGENESIS

Parthenogenesis is considered to be a rare occurrence among cock-
roaches. The best-known example of this type of reproduction in
blattids is the Surinam cockroach Pycnoscelus surinamensis, which
in the Indo-Malaysia area is bisexual; but in North America and Eu-
rope, where it has been introduced, it is parthenogenetic, producing
only females (Matthey, 1948). It is generally believed by most ob-
servers that parthenogenesis does not occur in our domiciliary species
or at most is a rare occurrence in the American cockroach. However,
only recently we have found that some unfertilized eggs of four
species of our common domestic cockroaches may complete their de-
velopment, and that in two of the species some of the eggs may hatch.

Normally, none of the unfertilized eggs of Supella supellectilium or
Blattelia germanica hatch, and only a small number of the eggs in an
odtheca sometimes complete development (pl. 6, figs. 36-40). How-
ever, we dissected an egg case of S. supellectiliwm and removed a fully
developed parthenogenetic embryo ; this individual shed its embryonic
membrane while we photographed it (pl. 6, figs. 41-45) ; the nymph

NO. I2 REPRODUCTION OF COCKROACHES—ROTH AND WILLIS Il

was successfully reared, eventually becoming an adult female. On the
other hand, some unfertilized eggs of Blatta orientalis and Peritplaneta
americana do hatch normally. In fact in P. americana, which has been
claimed to be a species in which parthenogenesis is a possible but in-
frequent phenomenon (Griffiths and Tauber, 1942), we have found
that hatching of unfertilized eggs is not uncommon; of 110 unferti-
lized females, 94 (85 percent) have produced odthecae from which
some eggs hatched (pl. 6, figs. 46-48).

In our experiments, which are still in progress, we have obtained
a total of 2,433 undamaged odthecae from unfertilized P. americana
females ; from these odthecae at least some of the unfertilized eggs in
1,030 (42 percent) hatched and the nymphs left the odtheca; in 779
(32 percent) some of the embryos developed until their pigmented
eyes were visible, or to an older stage, but failed to hatch ; and the eggs
in 624 odthecae (26 percent) failed to develop. More than 500 adult
females have been reared from parthenogenetically developed eggs ; no
males have resulted from the unfertilized eggs. These parthenogeneti-
cally produced females lay relatively few eggs and these eggs usually
fail to hatch. However, if mated, these females frequently produce
eggs that hatch. Parthenogenesis in P. americana is certainly less im-
portant than bisexual reproduction in the preservation of this species
in nature. However, parthenogenesis could operate among the wild
population, and in a temporary absence of males an unfertilized
female could transmit some of her germ plasm beyond the end of
her own life span.

THE OOTHECA
FORMATION

Kadyi (1879) and others have described the formation of the
ootheca of Blatia orientalis. The secretions from the colleterial glands
flow out over the inner surface of the vestibulum or odthecal chamber
in a sheet that surrounds and is stretched by the incoming eggs. The
vestibulum is closed posteriorly by the apical lobes of the seventh
sternum. As the forming odtheca presses against these lobes, a char-
acteristic pattern (pl. 7, fig. 52) is imparted to the distal end of the
odtheca (Wheeler, 1889). After a certain number of eggs have
entered the vestibulum, the distal end of the ootheca emerges beyond
the end of the abdomen. Pryor (1940a) has described the color
changes: At first the projecting portion is an opaque white; within
3 or 4 hours it becomes transparent, changes first to pink, and then
to reddish chestnut; the odtheca continues to darken after it is laid,

I2 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

becoming almost black in about 3 weeks. During formation the an-
terior part of the odtheca remains soft and white, eggs still entering
and being pushed to the rear. The similar formation of the odtheca
by Blattella germanica has been well described by Wheeler (1889).

The organic axis of the egg when it is still within the ovary and
oviduct of the female is oriented with the cephalic pole directed toward
the head of the mother ; the egg then emerges from the oviduct caudal
end first and falls into the genital armature caudal end down (Hallez,
1885). Because the odtheca of the viviparous cockroach Diploptera
dytiscoides Serville lies on its side within the brood sac, with the mi-
cropylar ends of the eggs directed toward the left wall of the brood
sac, Hagan (1951) stated, “This fact is of considerable historic interest
since it causes the embryos to develop with an orientation contrary to
the principles of Hallez’s law.” However, Hallez postulated the orien-
tation of the eggs within the ovarioles and oviducts, en route, so to
speak, to the odtheca; hence Hallez’s law is not applicable to a sec-
ondary orientation of the eggs which depends on any future position
of the odtheca. Wheeler (1889) demonstrated this clearly with odthe-
cae of Blattella germanica in which embryos developed normally when
the odthecae were oriented in five different positions; he concluded
that gravitation has no perceptible effect on the development of the
eges of this species, and that these eggs have their constituents pre-
arranged and completely conform to Hallez’s “loi de orientation de
loeuf.”

As the eggs move posteriorly, the valvulae of the ovipositor move
them into the odthecal chamber and in some way set them on end with
their heads upward (Snodgrass, 1952). The eggs from the right
ovary pass into the left side of the odtheca and vice versa (Kadyi,
1879; Wheeler, 1889; Wille, 1920). Gier (1947) found that some of
the eggs of Periplaneta americana are placed wrong end up in the
odtheca, and though development occurs normally, the nymphs cannot
emerge from the egg case. We have seen this in Supella also. In the
completed odtheca, the eggs are placed vertically and, except at the
ends of the egg case, arranged in two rows with the axis of each egg in
one row opposite the interval between adjacent eggs in the other row
(pl. 4, fig. 24; pl. 6, fig. 35). Figures 25 to 34 (pl. 5) show the
external appearance of an odtheca of Eurycoitts as it was being formed.

From what we have seen of odtheca formation in so-called vivipa-
rous cockroaches (see pp. 25-28), it is similar to that described above.
The eggs of Pycnoscelus surinamensis, Nauphoeta cinerea, and Leu-
cophaea maderae are erected vertically in two rows in the odtheca

NO. I2 REPRODUCTION OF COCKROACHES—ROTH AND WILLIS 1

which stretches around them (pl. 11, figs. 74, 75; pl. 12, fig. 86) as in
the oviparous species. In these three species the wall of the ootheca
is relatively thin and membranous. The color varies from pale straw
to amber. These odthecae do not darken and remain quite transparent
as does the odtheca of Blaberus cranifer (pl. 4, figs. 23, 24; pl. 11,
fig. 82; pl. 12, fig. 94). The odtheca of P. surinamensis is complete,
and although there is no differential keel, such as occurs in the ovip-
arous species, there is a narrow, longitudinal slit between the thick-
ened edges of the wall of the odtheca along its dorsal surface. The
odtheca of N. cinerea is incomplete, similar to that of B. cranifer, and
usually does not cover the micropylar ends of the eggs or parts of the
sides of the last three eggs deposited. We have seen some egg cases of
N. cinerea with eggs attached along the outside, apparently rolled
back by the walls of the brood sac from the imperfectly covered ante-
rior end of the egg case, as the female retracted the odtheca into the
brood sac.

Shelford (1906) found that the ootheca of the viviparous cock-
roach Panchlora virescens is represented by a complete, thin, trans-
parent membrane. However, the membrane forming the odtheca of
B. craniifer is incomplete ; as Saupe (1928) and Nutting (1953) point
out, the edges of the odtheca are separated by the micropylar ends of
the eggs (pl. 4, fig. 24). In Diploptera dytiscoides the odtheca is re-
duced to a thin membrane that covers no more than half of the egg
mass (Hagan, 1951). Riley (1891a) dissected an “egg cluster” of
Panchlora viridis and reported that the ootheca was only a membra-
nous sheath enclosing about half the length of the eggs. He reported
that in some egg cases of this species the eggs were arranged in a
double row side by side, with no visible enveloping membrane. This
latter condition seems doubtful ; the membrane may have been so thin
and colorless as to be nearly invisible. Or perhaps this was an abnor-
mal condition ; Gould and Deay (1940) and we have seen egg masses
of Periplaneta americana deposited without an ootheca (pl. 7, fig. 58).

Among cockroaches that do not carry the odtheca internally during
embryonic development, the hardened odtheca resembles insect cuticle.
Both have been shown to be scleroproteins which are very similar if
not identical chemically (Pryor, 1940a, b). However, the oothecae of
Periplaneta americana and Blatta orientalis contain no chitin (Camp-
bell, 1929; Pryor, 1940a), a compound found in varying amounts in
insect cuticles. Most of the materials which go to make up the ootheca
are secreted by the colleterial glands (Pryor, 1940a; Brunet, 1952).
The left colleterial gland secretes a water-soluble protein (Pryor,
1940a) and an oxidase (Brunet, 1952) ; the right gland produces a

14 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

fluid containing a dihydroxyphenol, specifically protocatechuic acid
(3, 4-dihydroxy-benzoic acid) (Pryor et al., 1946). When the secre-
tions from the right and left glands mix, the phenolic substance is oxi-
dized, producing a quinonoid tanning agent ; interaction of the tanning
agent with the protein gradually hardens and darkens the ootheca
(Pryor, 1940a). The odtheca also contains crystals of calcium oxa-
late (Kadyi, 1879) ; these crystals occur mixed with protein in the
lumen of the left colleterial gland. When the diphenolic substances
of the right gland mix with the protein of the left, the calcium oxalate
may play a part in maintaining an optimum pH for the oxidation of
the phenol (Brunet, 1952).

Certain valvulae of the ovipositor of oviparous species are modified
to mold the odtheca, especially the crista or keel. Wigglesworth and
Beament (1950) have shown this clearly in Blatiella germanica. The
chorion along the upper pole of each egg is expanded and forms a
vacuolated ridge which lies below the crista of the odtheca of this
species. In the keel above each egg, overlying the vacuolated ridge of
the chorion, is a small, oval, air-filled cavity which has two lateral ex-
pansions that pierce the crista and thus connect with the outer air.
These cavities are respiratory chambers which, with associated ducts,
convey air to the membranes around the eggs. From the roof of the
genital chamber of the female, elongated fingerlike genital appendages
project downward into the soft part of the forming odtheca and hold
the latest egg in place. Near the base of these fingerlike appendages
is a pair of thumblike projections directed backward, which serve to
mold the upper cavity of the odtheca and to orient the egg within it.
At the base of the thumblike lobes is a small median lobe with a tiny
sclerotized horn projecting on either side, which has the exact form
of the respiratory chambers and is the die on which they are molded.
The colleterial glands discharge their secretions at the base of the geni-
tal appendages, and the “horned die” molds the material providing a
respiratory chamber and respiratory duct for each egg. In Periplaneta
americana the third valvulae of the ovipositor are modified to form
the “horned die” which molds the inner surface of the keel of the
ootheca (Brunet, 1951).

These respiratory structures in the odthecae of oviparous cock-
roaches, because of their relatively small connections with the outer
air, retard loss of water by the eggs. The importance of this function
is emphasized when part or all of the keel has been eaten by the cock-
roaches themselves (pl. 7, figs. 56, 57); the eggs in these damaged
odthecae usually fail to develop at room humidities, or if they do,
rarely hatch. Under these conditions death undoubtedly follows ab-

NO. 12 REPRODUCTION OF COCKROACHES—ROTH AND WILLIS 15

normal loss of water. We have found that the rate of water loss from
American cockroach eggs at low humidities is greatly accelerated after
removal of the keel. Sometimes abnormal odthecae are deposited in
which the respiratory chambers are not differentiated in the keel
(pl. 7, fig. 60).

The oothecae of different species of cockroaches are quite distinc-
tive as they may vary in size, shape, and the number of enclosed eggs
(cf. pl. 4, fig. 23; pl. 6, figs. 36, 39; pl. 7, figs. 49, 59; pl. 12, fig. 87).
Lawson (1951, 1952, 1953) has studied the structural features of the
odthecae of several species of oviparous cockroaches. Each egg cell
in the odtheca is indicated externally by an evagination (forming half
of the respiratory chamber) on each side of the upper part of the keel
(Lawson, 1951). Thus the number of respiratory chambers and their
corresponding canals (pl. 7, fig. 53), which show clearly in the keels
of certain odthecae (e.g., Periplaneta americana, Blaita orientalis,
pl. 7, figs. 49, 59, and Eurycotis floridana), is often a good criterion
for the number of eggs in the odtheca. This relationship was recog-
nized by some early workers. For example Sells (1842) described an
odtheca of B. orientalis with 22 to 24 teeth along the serrated edge,
which corresponded with the number of eggs contained within. This
is rather a large number of eggs for the oriental cockroach, and Sells
may have been dealing with another species. We have found that in
abnormally small odthecae of P. americana, usually those containing
fewer than eight eggs, the number of respiratory chambers and ducts
is frequently not the same as the number of eggs in the odtheca (pl. 7,
figs. 54, 55). Wille (1920) also found that the number of egg cells
in the odtheca of Blattella germanica did not always correspond to the
number of teeth in the keel. Occasionally, eggs may be deposited
without the formation of a protective odtheca (pl. 7, fig. 58), or an
odtheca may be formed which contains no eggs.

REPRODUCTIVE POTENTIAL

The maximum number of eggs deposited at one time by a cockroach
is largely dependent on the number of ovarioles comprising the ova-
ries. The number of eggs per odtheca varies with the species. In the
oriental cockroach, a species which normally has 8 ovarioles per ovary,
the normal number of eggs per odtheca has been stated to be 16
(Seiss, 1896; Rau, 1924; Gould and Deay, 1940). Because there are
usually 8 ovarioles in each ovary, it is often stated that Periplaneta
americana normally deposits 16 eggs. However, the odthecae of this
species frequently contain fewer than 16 eggs. Disease or some ab-

16 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

normality of one or more of the ovarioles will reduce the number of
eggs produced in an odtheca (Gier, 1947). We have records of one
female American cockroach which consistently deposited six to eight
fertile eggs per ootheca ; dissection revealed that she had one normal
ovary, whereas the other had degenerated ; several eggs had been lib-
erated into her body cavity. The number of eggs per odtheca that
have been reported for various species of cockroaches are given in
table 1. However, because this information is not available for many
species, we have included some data on the number of nymphs hatch-
ing per odtheca. Counts of nymphs are usually smaller than egg
counts because the undeveloped eggs or unhatched eggs left in the
odtheca are not included.

Among the domiciliary oviparous cockroaches at least, the number
of egg cases produced by a female during her lifetime is even more
variable than the number of eggs per egg case. There is compara-
tively little information about the egg-laying potential of other cock-
roaches. Table 2 summarizes the more comprehensive data. Certain
unique values have been included because they extend the range of
observations.

Temperature, fecundation, and age of the female influence the
rhythm of egg and odtheca production of Blatia orientalis (Ricci,
1950) ; the rhythm accelerates with an increase in temperature, re-
sulting in more odthecae in a given period of time. Diet may also
affect the reproductive ability of blattids. Chauvin (1949) found that
the fecundity of Blattella germanica decreased considerably on a diet
deficient in sterols, and the reproductive ability of these insects dis-
appeared almost entirely after two or three generations had been
reared on the experimental diet. Noland et al. (1949) also found that
the odthecae produced by German cockroaches reared on certain syn-
thetic diets were often small, deformed, or shriveled, and only a small
proportion of the eggs hatched. This nutritional effect on reproduc-
tion could not be traced to any known deficiency in the diet, and these
workers suggested that a “reproduction factor” was lacking from the
diets. In Periplaneta americana specific diets such as peptone or dex-
trose reduced the frequency of odtheca production as well as the num-
ber of eggs in the odtheca (Gier, 1947). However, female adults of
Blatta orientalis maintained on a diet containing only 2.5 percent pro-
tein deposit normal odthecae (Lafon, 1951). The lack of vitamin E
in the diet for periods of 4 to 8 months does not influence the vitality
of mature sperm of B. orientalis (Kudrjaschow and Petrowskaja,
1937).

(panunyuo?)

6061 ‘AOIjoT ss3q Z gi ZI or
qoror ‘ney sydura jy 6z o£ gI wove.
of61 ‘Aeaq, pue pynor soy see gs o Ag 2
eyep poystqndun ‘sijiAy, pue yj}0Y sso] 0z ov gz ViZee*
ZS61 ‘AAA UeA sso “ oh ete 40r °°

IS61 ‘ta1TeYyIS syduiAN Se of ‘y

L461 ‘ney syduiAN Zz ov gt qoe
ebSOr ‘SIIIAA pue YOY ss3q oor Sz SI oes"
€zZO1 ‘UIaS syduiA Ny 19 os se : ere
zS61 ‘uMoiIg ss3q II 1 6 ao
eyep poystqndun ‘sijrAq pue YOY ss3q OSI 6£ SI 1 Sipe
ov6r ‘tyseyeyxe ss3q “es ov 3 2 eas
6z61 ‘ssoxy s33q ov oS gI [As Oe
1161 yjnesry s33q 0z gb ve We
Oggi ‘I]I0 s33q ve gs Qz oy ==
eyep paystyqndun ‘siyjIAq pue y30Y $337] ore gI I Pai"
zv61 ‘soi sso ae gI 8 Se
of61 ‘Aeaq pue pjnoy sydué Ny Le cI ¢ amare
Pz61 ‘ney s33q oe ie 79 OL
€S6r ‘surjjnN S337] vI Sv &z Oe
gz6r fadnes syduré yy 8 ge II Seas Oe

a9NNOs a8eP1S aed04}0O 8 WNnuUITxXeyy WNUIIUTyy ueoyy

10° 9ONi

SS
VOXUHLOO Wd SHAWAN YO SODA AO WHANWON

sayavo4ya09 fo savsads snoraa fo ‘paayjoo 424 poyojoy sygutu 40 ‘904,00 4a4 sb6a

eee

bocce see teeeeeeeeeeere® OG

"++ (qJoa80q_) vavupajksuad 14401Q0I4Dq
trresees = (IQIAYO) dasamsa Dja0yqnD Ny

Liebe eeeeneeteeeeeeees OG
"s+ (sntorige,y) avsapou Davy gona T

(arauunIg) suusofyjagosap DAapgouy Is]

sereees (JoyyeM) Duppisoy song
Oeces WCeCscin tot. . “UYoy 1hajaa yor pAqdunjid 7

tresessseesugydayg tasupg sniqojay

ee *preqgspy php D1] 24401

‘+s (agmmssnes) Duny4aquiny 07117}40] 7
uate ee Soh kotor ee - og
been eee eeeeneeeereres OG
sree (snaeuury) Domnmutsab 0119140]
Ren see teenie ree oq

Teer me RE OS Ooo oq
sete eee e sete eeeeeeeees OG

trees sess sonopuUIy S1701uUIaId0 114017
beeen eee eee eeeeeeeeeees OG
verses sygqsrouming 4afuups2 sniaqn] g

SaIDadds

fo daqunxy—t aTav],

‘DUDIIAAUD *q Se vsourbynf{ *q JO UOIyeoyIJUepIsiM & Woz pajpnser savy Aew
anyea YySIy sty} {gz 0} O1 Wor} SURI eB YIM (SayNsded O1) evodIY4}00 Jad ssSa pz JO asvioAv uv SaAIB (O161) pjajuesoy “A[UO v[nsd¥o suo UO paseq WINUWIxeW p
“judWTIedxe 1lay} UL Soynsdvo a[ijie} (pa}e}s jou Joquinu) [je woz poyoyey syduiAu hI jo URoW B BAIS (O61) Ava PUv pPjNor 5
“eyep IINOS WiOL} Pdze[NIeD q
‘ajnsded yooyied & Ul JaquinuU IO ‘1aquINU [eENSN IO [eWI1OU 94} 2q 0} Pd}zeIS v

S161 “YJIOMSUTT]T syduiAN v oz gi qh OL “*"* Tes (SUTyIag) Sadsoy ar0jquits
eyep paysiiqndun ‘siyjrAA, pue yy S337] lz gI Il QL irre estes et eeeeeeeeees OG
orol ‘keaq pue pjnor syduréA Ny vz ee ee ZEL @eeeeeeeeeeeeeeseeeerseeeeoee og
ccO1 “Wwe}}0D sydus£Ny oe Or of" ZI Sc (apptAtag) wanyyoazjadus pvpyjadns
vyep peystjqndun ‘STEM. pue YOY ss3q oz ve ZI g'Sz eer eer eee eee eee eee eee ees eee og
6r61 ‘aqeMyosS ss3y ee gf of ee @eeteeeeee eo eee eee eee eee eneee oq
orél ‘TaSIOVY syduA Ny gf ve ¢ az 1 eoceeeee ee eee eee eee eee eee eee od
gz61 ‘adnes syduiAn 9 ze 11 Peg S80 teres ce ee a ar oe od
SzO6I qjepned sso I gv ee ee ese eeeesr eee eee eee eee ee eenre og
gi61 ‘addezZ s33q + ov VI be ****(sneeuury) Ssisuamopuisis snjassourtg
Crél ney syduisN v gz aA vz eceese eee Ce eee eee eee eee eee oq
ob61 ‘Aevaq pue pynor ssoy ot a 33 we oo (QTFAIaS) vsoubyn{ vjauvjdisag
PIO 4pneIty ss3q oz Qz 2 gz **'** (SMIDLIqey) aDIsDjDAqsnD DjauD{q14a gq
eorol Ney s337 ov gI 9 qOrlt ere ee eee eee eee eee ee ee eee eee oq
oF61 ‘Aevaq pue pyjnor Ss3y +-oo1 pvz sz 29l ‘°° (shaeuury) vupsiuamD Djaunjgi4ag
aan Os 3981S avo04}00 =wnuixeyy WMNUTUIy uvof SaIOddS
jO "ON

SS eS
VOUHLOO Wad SHAWAN UO SOOA AO WHANWON

papnjguoj—i atav L

S161 ‘YJIOMSUTT]]
‘op
oF6I ‘Aeaq pue pjnory
zzO1 ‘Wwe}}07)
OVOL ‘tass0y
or61 ‘Avaq pue pinoy
€p61 ‘1aoueds
‘op
zror ‘saqney, pue syywyyI44
xe)
ov6l ‘feaq ie pjnory
fL6r ‘Ula
or61 ‘feaq pue pynoy
qovor ‘ney
yep poysyqndun ‘sii pue yoy
OS61 ‘Iaq1a45)
QVOI “‘JassI1ojy
ov6r1 ‘Aveaq pue pynor
6z61 ‘ssoy
OzOI ‘OTT
eyep paysiqndun ‘sit pue yoy
OSOI ‘Pony

‘op
or61 ‘keaq pue pjnoy
261 ‘ney

HOXNOS

*a[eUoy JO Sy] I[Npe o11}ua

*oTEUL2} JO VFI] YNpe 911}W

dy} jou ‘shep oz Jo polsod ev suring ¢
*peZt[toFUN) 3

ay} jou ‘shep oor jo polsed & suring q
*saurlejuoOd "90 OSe Ul ydoy So[euUY 3
“s1auIv}UOD *09 OOO‘T UT jdoy Sa[vUAT x

‘geU9F JO OFT] WYNpe sue oy} Jou ‘sAep oSt jo poltad ve suring o
*sS39 O[I}J1eyuI poonpord YoryM sojeuloy UISITA p
‘QTeUay JO IFT] WNpe oatua aq} you “sABp zoe JO potted & Ssuling o

J . Jsa1vaU 0}
ao I Be . scomeeens
yurysuoz 9 /t I ee
WOO’ : ee 9 ZEr cree
ei 7 iy, C¥etacenec
o ge £ I rae
MOOR a4 oe ers Zyottsee
ee ll yl ee o. wees
o8c-41 gal Iz ¢ gol sttttt
082-41 ye Zs oi 6'°9c O10 OD or
qos? C QZ (OT QoS aiiseustats
WOO OI 06 Ol wQss cece
ae. or ov g Chics
Wo0y I bee oo sie 2 efehotaye
Se G Il v QZ ee eee
aed gs Zz 0 Gp see ees
o0c-12 VI 6 ¢ a een Or,
ode gz ZI z EL oseeeee
Wooy 9 8 a2 mS conve
oft QI zg I i ames
oce 2 Tv we uetmerelroue\testate
o4e-62 oF Lz oO Cir eeeeee
oVZ-2 TI ace oe oe eee eee
qoSe (Ss 6 p <Orsrrers
WOON ¢ gI I Noi sGuagar
he cS v I Ei ie
7), ‘dway, ts ON wnuUNIxeyy wnuuiypy = ua

AIVWaAA Wad AVOAHLOO AO WAANWON

saysvo0s4ya09 fo saraags snorapa fq apzayzo0 fo uoyInpoAq—e AAV],

enews

pa}toauod ‘J , SB UIATS ‘e}JEP 9dINOG q
‘e}ep 9dINOS WoOIF pozepNd[eD v
sees ss(suiyiag) sagsoy 290jGuhs
vee tnenecenneen ee OG
seas al
“+ (aypiares) wnyyaayadns pjjadng

“+++ (snaeuulT) sisuampurans snjaasoumk

‘+ (ayfAses) vsombynf njounjdisad

*(snioriqe,,) apispppajsinw njaun)}qidad

bh@)/e-euaene) él cbieien @ sb) eats) 8) /s ce

“og

“(snoeuury) Dupsiamp Dyoun]fi4ad
teseteres OG

a ae) wae). 00 ee 6106 2 6 8x0 6

ce eww eevee serene eet em nee

*(qa080q_) vamvaksuad 0140])Q0I4q

Sele (uvelys 1618) 0 se .678) eae

9@
od
es og
“og
rereeees OG
‘e+ (snoeuury) vammnutsob 01194101 q
: od
-- og
be eeeeeeeeees OG
teeeeeeeeeees OG
*SHoeCUUlT S1j0juUaI40 DPD] TJ

SaIOadS

eee rere ose eee

sasha) ee,» effe)e0 6. a enyerer 6

ee eeee

eee eee

eee e wees

20 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

OVIPOSITION AND HATCHING

As we shall show below, cockroaches exhibit, or have exhibited at
some time during their phylogeny, a variety of ovipositional behaviors
from single-egg oviparity through multiple-egg oviparity and ovovivi-
parity to viviparity. Unfortunately these descriptive terms have been
used so diversely in the literature that their use creates confusion (see
Hagan, 1951). We shall make no attempt to redefine these terms, but
shall point out their limitations with respect to cockroaches, as they
appear to us.

Birth in insects is characterized as oviparous, in which the birth
product is an egg covered by a chorion or shell, or viviparous, in which
the egg hatches within the mother and the birth product is an embryo
devoid of a chorion (Hagan, 1951). Viviparity may be further di-
vided, with respect to cockroaches, into ovoviviparity, in which the
egg contains enough yolk to nourish the embryo until hatching, and
pseudoplacental viviparity, in which the embryo possibly derives a
part of its nutriment from the mother by means of a pseudoplacenta
(Hagan, 1951).

THE OVIPAROUS COCKROACHES

In marked contrast to the viviparous species, which carry the
odtheca internally, the oviparous species of cockroaches always carry
the odtheca externally (pl. 5, figs. 32-34; pl. 8, figs. 63, 64). The pe-
riod during which the oviparous female carries the odtheca before
dropping it depends on the habits of the genus and external factors
such as temperature, etc. Most genera, the ovipositional habits of
which are known, deposit the odtheca on the substratum shortly after
its formation, within a few days or less. Genera in which this habit
has been observed are Periplaneta (Sein, 1923) ; Blatta (Miall and
Denny, 1886; Moore, 1900) ; Supella, Parcoblatta, Cariblatta, Eury-
cotis (Roth and Willis, unpublished data) ; and others. Many species
carry the odtheca with the keel upward, as it was formed, until deposi-
tion. This condition may be observed in Blatta orientalis, Periplaneta
americana, P. australasiae, P. fuliginosa, Supella supellectiium, Cart-
blatta lutea minima Hebard, Eurycotis floridana, and undoubtedly
others.

On the other hand, certain genera of cockroaches rotate the odtheca
go° to the left or right and carry it until, or shortly before, the eggs
hatch. This habit has been described in Blattella germanica (Wheeler,
1889; Wille, 1920; Ross, 1929; Woodruff, 1938; Gould and Deay,
1940; Pettit, 1944; Rau, 1944). It has also been observed in other

NO. 12 REPRODUCTION OF COCKROACHES—ROTH AND WILLIS 21

species of this genus: B. lituricollis (Walker) (Zimmerman, 1948) ;
B. humbertiana Saussure (Takahashi, 1940) ; B. vaga (we have ob-
served the young hatching from an odtheca that was still being carried
by the female; see pl. 8, fig. 65). The direction of rotation of the
egg case of B. germanica has been reported to be toward the right
(Wheeler, 1889; Wille, 1920; Ross, 1929) or toward the left (Gould
and Deay, 1940). Possibly the direction of rotation is of genetic
origin and is relatively constant within a particular laboratory colony.
The ootheca seems to be held in position by pressure of the encircling
genital armature (Ross, 1929). Rarely B. germanica may form a new
odtheca before the empty one is discarded, and the empty odtheca re-
mains attached to the fresh one (pl. 7, fig. 62). We have also seen
instances of Periplaneta americana depositing two odthecae attached
to each other by their ends; formation of the second was started
before the first was dropped, and the soft end of the new oodtheca
curved around the hard end of the old one.

The oviposition behavior of Ectobius pangeri Stephens combines
aspects of both the Blatta and Blattella types just described. Brown
(1952) reported that E. panzeri usually rotates the ootheca through
go° and most often to the left. Some females carry the egg case for
as long as 16 days, but the majority deposit it in less than 10 days.
This species winters in the egg, and the nymphs hatch out the next
spring (Brown, 1952). Lucas (1928) found a female of Ectobius
lapponicus Linnaeus that carried her egg case at least 12 days. In this
species the egg case is deposited in the summer and the eggs hatch
quite soon.

Some of the species that deposit the odtheca shortly after it is
formed frequently conceal it and cement it to the substrate. The fe-
male may cover the odtheca with bits of debris which she chews from
the substrate (Latreille, [1803/1804] ; Sampson, in Shelford, 1912)
and mixes with saliva (Haber, 1920). The habit of concealing the
odtheca, or covering it with debris, is practiced by Periplaneta ameri-
cana (Haber, 1920; Adair, 1923; Rau, 1943); P. australasiae (Gir-
ault, 1915; Spencer, 1943); Blatia orientalis (Qadri, 1938; Rau,
1943) ; Loboptera decipiens (Berland, 1924) ; Supella supellectilium
(Flock, 1941) ; Cryptocercus punctulatus Scudder (Cleveland et al.,
1934) ; and Eurycotis floridana (Roth and Willis, 1954a). The habit
of covering the odtheca with debris is not equally developed in all the
domiciliary species. In the laboratory, the eggs of P. americana are
covered most frequently ; those of B. orientalis are usually dropped
free, uncovered and not cemented to the substrate (Gould and Deay,
1940; and our own observations) ; those of Supella are usually ce-

22 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

mented to the substrate but are only sometimes partly covered with
bits of feces or other debris which is cemented to the odtheca. These
- females show no further care for the odtheca after it has been de-
posited, leaving the eggs to hatch, which often requires several weeks.

As mentioned earlier, with few exceptions the embryos are ar-
ranged in the odtheca with their heads directed toward the keel (pl. 9,
fig. 66) or dorsal side of the odtheca. Generally at hatching the two
halves of the keel separate, and most of the young of Periplaneta amer-
icana emerge almost simultaneously in less than 10 minutes (Gould
and Deay, 1938) or at the most within the hour (Klein, 1933). Goeze
(1782) stated that the hatching cockroaches produce a material which
softens the cement between the halves of the serrated crista enabling
the young nymphs to escape from the egg case. This view has been
repeated in many publications, but there seems to be no experimental
evidence for the statement. Internal pressure exerted by the fully
developed nymphs causes the odtheca to split along the keel (Wheeler,
1889 ; Gould and Deay, 1938; Gier, 1947). If some of the young are
late in hatching, they cannot escape from the capsule (which tends to
snap shut after the young have emerged), or they may be trapped be-
tween the two lips of the keel (Fischer, 1928) (pl. 6, figs. 46-48).
Pettit (1944) found that the odtheca of Blattella germanica splits
apart more readily as the development of the embryo nears comple-
tion. As soon as the crista opens, the young cockroaches swallow air
bubbles at the rate of two or three a second. These unite to form a
large bubble in the alimentary canal that almost doubles the volume
of the insect. By squirming and waving the upper parts of their
bodies, the young cockroaches worm their way out. We have also ob-
served the hatching nymphs of Periplaneta americana swallow air.
The emerging insects are covered by thin, transparent embryonic
membranes which are quickly shed, usually during emergence (pl. 6,
figs. 44, 47), and often eaten. Sometimes in P. americana these mem-
branes remain caught between the lips of the keel. In Ectobius panzert,
the odthecae of which are buried in sand, this first nymphal skin is
shed after the nymphs reach the surface (Brown, 1952). Brown did
not observe these exuviae to be eaten.

Numerous writers have repeated the statement that the female Ger-
man cockroach assists her young to hatch by slitting the odtheca along
the seam with her mandibles. This belief originated with a statement
by Hummel (1821); he misinterpreted the behavior of a female
which apparently was only “exploring,” with her antennae and palpi,
a recently dropped odtheca when the nymphs coincidentally began to
emerge. This egg case, incidentally, had been dropped by another fe-

NO. I2 REPRODUCTION OF COCKROACHES—ROTH AND WILLIS 23

male, and the female in question was still carrying her own odtheca.
Wheeler (1889), Ross (1929), and Pettit (1944) have thoroughly
discredited this midwifery.

Although the usual method of emergence among oviparous cock-
roaches is to split the keel of the ootheca, Terry (1910) states that
the young of Euthyrrhapha pacifica (Coquebert) gnaw an exit hole
through the capsule ; a similar method for emergence of the young is
said to occur in Nyctibora lutzi Rehn and Hebard (Wolcott, 1950).

THE VIVIPAROUS COCKROACHES

For our use viviparity as defined by Hagan (1951; see above) may
be too restrictive, in that different individuals of the same cockroach
species (e.g., Pycnoscelus surinamensis) may give birth either to
nymphs, which emerge as such from the mother, or to eggs, still en-
closed in the odtheca, which hatch immediately or shortly after the
egg case has been dropped. Hence the species may be considered ovo-
viviparous or oviparous depending upon the behavior of the indi-
vidual insect. Snodgrass’s (1935) broader definition would include
both types of behavior in viviparity, as in both, the eggs complete
development within the body of the female. By accepting a broader
definition of viviparity we avoid the anomaly of having oviparous be-
havior, by definition, in species that logically would be considered ovo-
viviparous. Such facultative oviparity, in typically ovoviviparous
species, differs greatly from the obligate oviparity of the common
domiciliary cockroaches, certainly much more than it differs from
viviparity.

Viviparity among the so-called ovoviviparous cockroaches, as we
shall show below, differs fundamentally from viviparity in other in-
sects. Among the insects considered by Hagan (1951), viviparity is
the birth of embryos from eggs retained in the mother’s genital tract.
The ovoviviparous cockroaches, however, oviposit into an odtheca, and
the eggs actually pass out of the female’s body. This fact has only re-
cently been elucidated (Chopard, 1950; Nutting, 1953; and original
observations below) and perhaps is not generally known. Oviposition
among the ovoviviparous cockroaches is very similar to that of ovip-
arous forms, except that the completed odtheca is retracted by the
ovoviviparous female and deposited in her brood sac instead of on
the substrate. Viviparity among ovoviviparous cockroaches is, there-
fore, a secondary condition. The birth product is first an egg enclosed
in a chorion, and as such satisfies the criteria for oviparity. After the
female retracts the odtheca into her brood sac, the eggs may hatch

24 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

within her body and the secondary birth product is an embryo devoid
of a chorion, thus satisfying the criteria for viviparity. However,
eschewing all theological implications, can an individual once born be
born again? If not, the so-called ovoviviparous cockroaches of which
we have the greatest knowledge are, strictly speaking, oviparous. Yet
it is a decided convenience to use the term viviparous for those species
of cockroaches which incubate the eggs in the brood sac, as opposed to
Oviparous species which do not incubate the eggs within the body.
Viviparity has been reported for several cockroach genera and spe-
cies (Chopard, 1938). It has been assumed by some observers that a
species is viviparous when, upon dissection, embryos are found devel-
oping in a brood sac, even though hatching was not observed (e.g.,
Panesthia javanica, Wood-Mason, 1878; Eustegaster micans Saus-
sure, and Oxryhaloa saussurei Borgmeier, Holmgren, 1903-1904). Re-
cently emerged nymphs found associated with a female, with no evi-
dence of an odtheca, have also been cited as evidence of viviparity
(e.g., Panchlora, Scudder, 1890; Davis, 1930). Newly hatched
nymphs of viviparous cockroaches frequently eat the odtheca (Nut-
ting, 1953; our own observations), so its absence is not surprising.
A number of workers have observed that the females of apparently
viviparous species produce odthecae that protrude externally. This
event may occur as the odtheca is formed (see below) or later on when
the ootheca is expelled either prior to or at the time of hatching. Fre-
quently, expulsion of the odtheca does not coincide with completion
of embryonic development. If the female deposits the ootheca pre-
maturely, the eggs usually fail to hatch. This has been reported for
Blaberus cranufer (Nutting, 1953), Pycnoscelus surinamensis (Saupe,
1928; Roeser, 1940), Nauphoeta cinerea (Illingworth, 1942), Grom-
phadorhina laevigata (Chopard, 1950), and Leucophaea maderae
(Scharrer, personal communication). Karny (1924) described a fe-
male of Pseudophoraspis nebulosa Burmeister carrying a weakly “chi-
tinized,” pale-yellow odtheca which contracted on drying ; this descrip-
tion together with the fact that this species has a brood sac, and has
been found with young clinging to the underside of the abdomen
(Shelford, 1906), would lead us to believe Karny was dealing with a
species similar in oviposition behavior to the above viviparous forms.
The observations of Chopard (1950) first served to explain the
carrying of odthecae by the above cockroaches. He found that Grom-
phadorhina laevigata extrudes an odtheca which after several hours
extends about 25 mm. beyond the end of the abdomen. Then when the
odtheca is held only by its end, it is slowly drawn back into a large
brood sac that extends into the metathorax. In the gravid female the

NO. 12 REPRODUCTION OF COCKROACHES—ROTH AND WILLIS 25

Ovipositor is completely inverted; Chopard states that the inversion
probably occurs at the moment the ootheca penetrates the brood sac.
Snodgrass (1952) thinks that the ovipositor of Gromphadorhina prob-
ably inserts the fully formed odtheca into the brood sac. The eggs
are incubated within this pouch for about 70 days at the end of which
time the young are born. Nutting (1953) observed a similar behavior
in Blaberus craniufer. Gurney (1953) suggested that this method of
transference of the odtheca into the brood sac may occur in other
viviparous genera which have been seen with protruding odthecae.

We have observed the extrusion of the odtheca during its forma-
tion and its subsequent retraction into the brood sac in Pycnoscelus
surinamensis, Nauphoeta cinerea, and Leucophaea maderae; it is ap-
parent that the formation and retraction of the odtheca in these species
is quite similar to that of Gromphadorhina and Blaberus. The oviposi-
tion behavior of the three former species is described below:

Pycnoscelus surinamensis—Ten females have been observed dur-
ing various stages of extruding and retracting their odthecae (pl. 11,
figs. 74-79). One female completed the entire process in 2 hours and
I5 minutes from the time the first eggs were extruded to the time they
disappeared back into her abdomen. She took 1 hour and 25 minutes
to extrude the odtheca to its fullest extent ; the odtheca then remained
more or less unchanged in this position for 40 minutes, and then the
female began to retract the odtheca into the brood sac. Retraction
time took only 10 minutes. The retraction time for the odthecae of two
other females was also about 10 minutes. When the egg case is ex-
truded (pl. 11, figs. 74, 75), the abdomen of the female is much con-
tracted, the segments being tightly telescoped. The axes of the eggs
are vertical at this time and remain so until the egg case is fully ex-
truded. As the female begins to retract the odtheca, the axes of the
eggs incline more and more to the left (pl. 11, figs. 76, 77), so that as
the last eggs disappear from view they are horizontal. When the
odtheca is finally retracted, the abdominal segments return to the nor-
mal position. Saupe (1928) was incorrect in his assumption that the
ootheca was formed and rotates 90° inside the brood sac.

Nymphs hatched from the eggs of three of the observed females 31,
33, and 35 days after formation of their odthecae. We did not observe
the hatching process. One of these females which we dissected after
the young had hatched had ovaries in which none of the eggs were
well developed. A female dissected 37 days after she had formed an
odtheca had no egg case in her brood sac; but her ovaries contained
well-developed eggs which were apparently ready to be deposited in
an o6theca. It is likely that this female had expelled and eaten the

26 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

odtheca which we had seen her retract. Another female dissected 36
days after forming an odtheca had a recently formed egg case contain-
ing undeveloped eggs in her brood sac. Apparently this female had
also expelled and eaten the egg case we had seen formed, and then
formed a new one unobserved. Obviously in determining the time re-
quired for the embryonic development of cockroaches which are in-
cubated in a brood sac, one must be certain that the odtheca which is
seen being retracted is not subsequently dropped, eaten, and replaced
by a new one without the knowledge of the observer.

Nauphoeta cinerea—More than 20 females have been observed
with odthecae in various stages of extrusion and retraction (pl. 12,
figs. 86-90). Complete extrusion of the eggs takes about 14 hours or
less. As the eggs are extruded, the female invests them with a thin,
straw-colored oodtheca. During the early stages of formation the
odtheca frequently projects upward at a slight angle. Usually by the
time the ootheca is half formed it bends slightly downward, and when
completely extruded it assumes a curve of rather short radius (pl. 12,
fig. 87). Except for the first egg, which lies with its axis on an angle
to the others, the eggs are extruded with their axes vertical; they
remain in this position until the odtheca has been completed. The fe-
male then rotates the egg case to the left rather rapidly so that within
about 5 to 10 minutes the axes of the eggs are horizontal ; the female
then retracts the odtheca, in about I to 2 hours or sometimes much
longer. One female carried her egg case with about three eggs ex-
posed for 22 hours before complete retraction occurred. Sometimes
retraction is incomplete and several eggs remain protruding from the
end of the abdomen. These exposed eggs dry up and shrivel. One fe-
male which had failed to retract her odtheca completely (about 15 eggs
had remained exposed) gave birth to 12 nymphs 34 days later. Inas-
much as the genital opening is blocked when dead eggs protrude from
the abdomen, the female must drop the odtheca or extrude it at least
partially to allow the developed nymphs to hatch. Hatching of the
eggs of four females occurred from 33.5 to 41 days after they formed
their odthecae. The odtheca in the brood sac fills almost the whole
abdominal cavity (pl. 12, figs. 91-94).

Leucophaea maderae——Two females were observed as they ex-
truded and retracted their egg cases. As a result of an accident, the
egg case of the first female broke incompletely into two halves which
were held together by a thin portion of the delicate membrane form-
ing the odtheca. This female could not retract all the eggs into her
abdomen. The eggs that protruded became dry and shrunken after 3

NO. I2 REPRODUCTION OF COCKROACHES—ROTH AND WILLIS 27

days’ exposure, but remained attached to the female. She died on the
fourth day ; the eggs within the brood sac appeared normal.

When the second female was first observed, four eggs were already
visible beyond the end of her abdomen. As more eggs were extruded,
their axes, which were originally vertical, began to tilt toward the left
side of the female. After 3 hours and 20 minutes the odtheca was fully
extruded ; it contained about 38 eggs and extended three-quarters of an
inch or more beyond the posterior end of her abdomen. After another
10 minutes the axes of the eggs were horizontal, and the female began
to maneuver the odtheca into the brood sac. When fully extruded, the
odtheca was curved so that the terminal eggs farthest from the body
were far to the left of the midline of the body of the female. While
being retracted, the odtheca stuck out straight behind the female.
After 45 minutes only one egg (the first laid) was visible. But this
egg projected for 1 hour and 25 minutes more before it disappeared.
Total retraction time was 2 hours and Io minutes, and the whole proc-
ess took more than 53 hours. -

In all species of viviparous cockroaches in which oviposition has
been observed, including Leucophaea maderae, the last eggs deposited
in the odtheca are the first to enter the brood sac as the odtheca is
retracted. This action places the most recently laid eggs in the ante-
rior end of the brood sac and the oldest eggs in the posterior end of
the brood sac. Yet van Wyk (1952) stated that in Leucophaea ma-
derae “The odtheca is contained in the vestibulum ? and the latter ex-
tends cephalad as the former increases in size. The oldest eggs are
therefore situated at the anterior end of the vestibulum.” Obviously
this description was not based on direct observation of the transfer of
the odtheca into the brood sac.

Though the formation of the odtheca is very similar in the above
three species, certain differences should be noted. In Pycnoscelus the
axes of the eggs remain vertical until the odtheca is fully extruded and
then they gradually incline toward the left as the egg case is retracted
into the brood sac; retraction itself is very rapid requiring only about
to minutes. In Nauphoeta the axes of the eggs are also vertical until
the odtheca is fully extruded. However, unlike Pycnoscelus, the
odtheca is completely rotated in about 5 or 10 minutes and is then re-
tracted in about an hour or more. In Leucophaea, as in Nauphoeta,
the axes of the eggs are horizonal when the retraction process begins.

2 The vestibulum or odthecal chamber is the portion of the genital cavity lying
above the 7th sternum (Snodgrass, 1937, 1952). Van Wyk (1952) considers the
brood sac to be an anterior expansion of the vestibulum.

28 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

The time required for the formation and retraction of the ootheca
is quite variable ; we may have induced some variation by disturbing
the females. However, the complete formation and extrusion of the
odtheca (not including retraction) by the above species occurs more
rapidly than among species in which the odthecae remain exposed
after they are formed. For example, Eurycotis floridana (Walker)
forms its egg case in about 6 hours (Roth and Willis, 1954a), and P.
americana, B. orientalis, B. vaga, and B. germanica require about the
same length of time or longer for o0theca formation ; Wille (1920) re-
ported 16 to 24 hours for complete odtheca formation in B. germanica.
Brown (1952) reported 6 hours for odtheca formation by Ectobius
panzeri. The longer period taken by oviparous cockroaches to form
their oothecae may be related to the fact that in these species the
toothed keel of the odtheca is a complicated structure containing re-
spiratory chambers and ducts. Obviously a certain amount of time is
necessary to form each of the respiratory chambers; the protein sub-
stance of the ootheca must harden around the “horned die” before the
next chamber is formed. The odthecae of Pycnoscelus, Nauphoeta,
and Leucophaea have no respiratory chambers in a specialized crista ;
this may account for the comparative rapidity with which these species
form their egg cases.

It is very difficult to ascertain just how birth occurs in the cock-
roaches that retain their eggs within a brood sac; the young may hatch
as the odtheca is extruded from the brood sac into the vestibulum or
odthecal chamber, and the nymphs then issue from the body of the fe-
male ; they may hatch as the odtheca is extruded beyond the oothecal
chamber ; or the odtheca may be dropped completely, the young hatch-
ing shortly afterward. Only direct observation can determine the
method, and inasmuch as it is practically impossible to tell the exact
moment when birth will occur, the insects have to be watched closely
and at times continuously.

Actual observations of the hatching of viviparous cockroaches are
few in number. Hatching has been observed in Panchlora viridis
(Riley, 1890, 1891a, b) ; Pycnoscelus surinamensis (Thomas, 1949;
Schwabe, 1949) ; Diploptera dytiscoides (Hagan, 1939); Blaberus
craniifer (Nutting, 1953); Gromphadorhina laevigata (Chopard,
1950) ; Leucophaea maderae (Scharrer, 1951; van Wyk, 1952) ; and
Nauphoeta cinerea (Roth and Willis, unpublished data).

We have observed birth twice in Nauphoeta cinerea. The first time,
we saw a female extrude an odtheca containing fully developed em-
bryos; while it was still attached to the female, two other cockroaches
seized the egg case, and one, managing to free the odtheca from the

NO. 12 REPRODUCTION OF COCKROACHES—ROTH AND WILLIS 29

mother, carried it off. We retrieved this odtheca; four eggs hatched
immediately, and the nymphs commenced feeding on the odtheca.
About 23 well-developed embryos failed to hatch. The second time
we observed birth, most of the nymphs had hatched before we noticed
the event. The young cockroaches, still invested with their embryonic
membranes, dropped from the female’s odthecal cavity as she ex-
truded the odtheca. When we isolated the female in a vial, she seized
and ate two nymphs immediately after they had dropped free and
before they shed their embryonic membranes. The remainder of the
odtheca, which had been extruded as far backward as the odthecal
cavity, was removed from this female when further hatching ceased.
There were six live embryos in this fragment of odtheca; these did
not hatch.

Some reports about hatching in viviparous cockroaches are conflict-
ing. For example, Scharrer (1951) states that the young of Leu-
cophaea maderae hatch from their eggs the moment they leave the
mother. Van Wyk (1952) says the eggs of this species “hatch in pairs
in the odtheca while it is still in the vestibulum * and the young im-
mediately leave the mother. . . . The remains of the ootheca and the
chorion of each egg are thrown out by the mother after all the eggs
have hatched.” Yet Pessoa and Correa (1928) state that this species
deposited the egg case, and the young hatched from the capsule 20
days after laying.

Saupe (1928) argued that the odtheca of Blaberus craniifer fills
the brood sac so completely that the nymphs have no room to hatch
within the sac. Although he did not observe the act, Saupe believed
that hatching occurred in B. craniifer as the ootheca was extruded or
shortly after it had been dropped. Nutting (1953) also believes that
it is impossible for the young to hatch in the brood sac of B. cramifer ;
he observed the hatching of nymphs from an odtheca that was being
extruded by the female and also from recently dropped odthecae of
this species. In Pycnoscelus surinamensis the ootheca completely fills
the brood sac which is stretched tightly against the egg case. Thomas
(1949) observed that in this species the nymphs were delivered alive
and individually. Schwabe (1949) stated that the nymphs of P. surt-
namensis hatch within the body of the female. Even in P. suri-
namensis, however, the female may expel the odtheca, and then the

3 As van Wyk (1952) considers the vestibulum to include the brood sac, we
do not know from his description whether or not hatching occurred within the
brood sac proper or within the posterior part of the vestibulum. After the
odtheca has been released from the brood sac, hatching most probably occurs as
the odtheca is extruded beyond the vestibulum.

30 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

young hatch outside the female ; we have observed this several times.
According to Zappe (1918) hatching of P. surinamensis may occur
within the mother, or the odthecae containing well-developed embryos
are laid in the soil, the young hatching within 24 hours. Later Zappe
(Hagan, 1951) claimed that in Connecticut this species is oviparous.
The two females of Nauphoeta cinerea that we observed giving birth
expelled their odthecae, at least partially, before the eggs started to
hatch.

The site in which the eggs of viviparous cockroaches hatch would
seem to depend upon the rapidity with which the embryos respond to
a stimulus to hatch. The site might be anywhere from the vestibulum
to a position outside of the body of the female. Presumably hatching
does not occur within the tightly investing brood sac because of spatial
limitations and the pressure exerted by the wall of the brood sac.
This pressure undoubtedly increases between the time the odtheca is
retracted and the time of hatching, as the eggs of viviparous cock-
roaches increase in size during embryonic development (Hagan, 1951;
Nutting, 1953), thereby stretching the brood sac even more. Nutting
(1953) believes that pressure exerted by the female on the odtheca
during extrusion may supply the necessary hatching stimulus. His
evidence is not convincing: he secured hatching from odthecae, that
had been extruded 2 to 6 days earlier, after manipulating the odthecae
with his fingers. Yet these same odthecae had been subjected to what-
ever pressure the female might apply prior to and during extrusion,
and the eggs had not hatched. These odthecae might have been
dropped prematurely before the embryos were ready to hatch. On
the other hand, the release of the odtheca from the tightly stretched
brood sac during extrusion might be a hatching stimulus—at least
this seems to be true in Pycnoscelus surinamensis (see below).

Pycnoscelus will drop its odtheca when exposed to temperatures
above 35° C. or below 5° C., and after being poisoned, injured, or
decapitated (Roeser, 1940). Roeser observed that seldom is an odtheca
found inside a dead female; shortly before or after death due to natu-
ral causes the female expels the ootheca from the brood sac. We de-
capitated five females 33 to 35 days after they had retracted their
odthecae. Each female extruded her odtheca, and nymphs began hatch-
ing as soon as the egg case was released from the brood sac (pl. 11,
figs. 80-85). The nymphs swallowed air, swelled up, and emerged
from the ootheca, shedding the embryonic membranes which had sur-
rounded them (pl. 11, figs. 83, 85, arrows). Some of these eggs were
obviously ready to hatch before the females were killed, and release

NO. I2 REPRODUCTION OF COCKROACHES—ROTH AND WILLIS 31

from the pressure of the brood sac was followed by immediate hatch-
ing. Just what the stimulus is for the female to extrude the odtheca
at the end of the gestation period is unknown. Some eggs were com-
pletely undeveloped and others not far enough developed to hatch;
these remained in the odtheca. Apparently some of the eggs may
develop at different rates. Schwabe (1949) recorded one instance in
which 36 hours elapsed between the birth of the first and last nymphs
of the brood. The odtheca is sufficiently developed to prevent hatch-
ing if only a few individuals in the egg case have reached the hatch-
ing stage and the egg case is dropped prematurely. Yet we noticed
two nymphs which hatched from one dropped egg case in which only
five embryos had developed ; two of the others were only able to free
part of their heads from the odtheca.

Injury or death causes the extrusion of the odtheca by females of
other species of viviparous cockroaches. From the number of genera
in which this reaction has been observed, we think that the phenome-
non is probably general among cockroaches that carry their odthecae
in brood sacs. Beebe (1925) saw a giant “woodroach,” which was
being eaten by a spider, give birth to 51 young which “had burst from
their mother.” Shelford (1906) recorded a specimen of Epilampra
burmeistert Guerin from Brazil that was preserved with two nymphs
(still partly surrounded by their embryonic membranes) emerging
from the tip of the female’s abdomen; these nymphs may have been
expelled by the female as she died. Gissler (Riley, 1891b) observed
24 young emerge from the genital orifice of a female Panchlora viri-
dis which had died. Stewart (1925) killed a female of Blabera cuben-
sis [= Blaberus discoidalis Serville (Rehn and Hebard, 1927)] ina
cyanide bottle ; during the few seconds the female survived she partly
extruded an odtheca containing 44 eggs. Heal (personal communica-
tion) observed Leucophaea maderae females with creamy-white, frag-
ile egg cases protruding from the ends of their abdomens; these
females had been killed either with cyanide or pyrethrum during dis-
infestation of buildings. We have seen Blaberus cranifer partially ex-
pel an odtheca after decapitation prior to dissection (pl. 4, fig. 22); a
pin placed against the posterior end of this odtheca prevented further
extrusion. We have also seen a female of Nauphoeta cinerea expel its
odtheca on being dropped into boiling water. Insecticidal sprays may
cause the German cockroach female to drop its odtheca prematurely
(Woodbury, 1938; Parker and Campbell, 1940). A similar occur-
rence of oviposition by injured or dying mosquitoes, Aedes sollicitans
(Walker), and the phenomenon of “death stress” have been studied
by DeCoursey and Webster (1952).

32 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

Hagan (1939, 1941, 1951) has worked extensively with the vivipa-
rous cockroach Diploptera dytiscoides. He has described evanes-
cent embryonic structures, the pleuropodia, to which, tentatively, he
ascribes nutritional or respiratory functions, or both. The pleuropodia
arise as a part of the swelling of the first abdominal segment. With
embryological development they increase greatly in length, becoming
thin-walled tubes which end near the micropyle. They are bound to
the inner surface of the chorion by the yellow serosal cuticle. Be-
cause of this modification Hagan cites D. dytiscoides as the one
example of pseudoplacental viviparity among cockroaches.

Hagan (1951) described oviposition in Diploptera dytiscoides: On
reaching the lower end of the common oviduct the eggs “are directed
by the ovipositor from the genital chamber ventrally into the open end
of the uterus.” The implication here, which has subsequently been
confirmed by Hagan (personal communication), is that the first-laid
eggs are the first to enter the brood sac and the last-laid eggs are the
most posterior in the brood sac. A pronounced central dome in the
roof of the genital pouch is presumed to facilitate “tilting of the eggs
as they pass into the uterus” (Hagan, 1941). This method of ovi-
position contrasts markedly with what has been observed in other
viviparous cockroaches, as Snodgrass (1952) has noted. For this
reason we are re-evaluating oviposition in D. dytiscoides in the light
of what is now known of oviposition in other viviparous cockroaches.

Hagan (personal communication) had not observed protrusion of
the odtheca by Diploptera dytiscoides, in the field or in rearing cages.
Other entomologists in Hawaii whom he questioned had not observed
this species with the odtheca extruded. Apparently Hagan based his
account of oviposition in D. dytiscoides on an interpretation of the
anatomical relationships in dissected specimens. D. dytiscoides is a
small insect three-quarters inch in length (Hagan, 1941) ; the eggs
it produces are also small, being 1.20 mm. long and 0.43 mm. in the
greatest dorsoventral dimension (Hagan, 1951). As only 12 odcytes
are usually matured at one time, the odtheca is very short, with a com-
puted length of about 3 mm. Possibly the posterior end of the form-
ing odtheca does not protrude far enough from the odthecal cavity to
be easily seen. As the complete process of formation and retraction
can be very rapid with much larger odthecae containing many more
eggs (e.g., Pycnoscelus surinamensis, Nauphoeta cinerea, above), the
transfer of the egg case by D. dytiscoides into the brood sac might
pass unnoticed by anyone not specifically looking for it.

It is difficult for us to visualize the formation of the odtheca of
D. dytiscoides during a direct passage of the eggs from the oviducts

NO. I2 REPRODUCTION OF COCKROACHES—ROTH AND WILLIS 33

into the brood sac. The ootheca of this species is apparently com-
parable to those of Pycnoscelus surinamensis, Blaberus cramifer, Nau-
phoeta cinerea, and Leucophaea maderae, except that it may possibly
be less extensive than those of the first two species. The delicate
odtheca of D. dytiscoides encloses the lower ends and sides of the
eggs, the micropylar ends remaining free (Hagan, 1951). This is
very similar to the condition that we have found in N. cinerea, in
which, at times, the odtheca seems to be merely a thin film of varnish
applied over the sides of the eggs. There are other close similarities
between the odtheca of D. dytiscoides and the odthecae of other vivipa-
rous cockroaches. For example, the eggs are disposed in the odtheca
in two parallel rows, the eggs of one row fitting opposite the intervals
in the other row; the ventral surfaces of the eggs in one row face the
venters of the eggs in the opposite row ; and at the ends of the odotheca
the first and last eggs lie in the midline (Hagan, 1951). These fea-
tures of the odtheca of D. dytiscoides apply to blattid odthecae in gen-
eral; hence we would expect the mode of formation of the odtheca in
this species to be similar to what has been found in other viviparous
cockroaches.

Probably the odtheca of Diploptera dytiscoides is formed around
the eggs as they are aligned in a double row in the oodthecal cavity
with their axes vertical, as has been observed in other species of cock-
roaches. Then, after the last egg is deposited, the female presumably
rotates the odtheca and retracts it into her brood sac with the last-
laid eggs coming to lie in the anterior end of the brood sac. This
procedure would place the odtheca in the brood sac with the axes of
the eggs rotated go° to the left as Hagan (1951) found in his speci-
mens. Until direct observation proves the above interpretation to be
wrong, it seems logical to identify oviposition in D. dytiscoides with
that in other viviparous cockroaches.

VARIATION IN OVIPOSITION

Wood-Mason (1878) suggested that the habit of certain species of
blattids of carrying the odtheca for a week or more before deposition
represents the retention of a vestige of a lost viviparity. However,
Shelford (1909, 1912) believed that the method of depositing eggs
in an odtheca onto a substrate is probably a more primitive behavior
than incubating the eggs in a brood sac. Not only have fossil odthecae
been found, but also fossil cockroaches which possessed elongated
Ovipositors, indicating that during the Permo-Carboniferous period
there were two categories of cockroaches; the more primitive had a

34 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

long external ovipositor and laid eggs not united in an odtheca; the
other had a much reduced internal ovipositor and made oothecae like
recent blattids (Laurentiaux, 1951).

Various adaptations of oviposition behavior and oodthecal structure
serve to some extent to protect the cockroach eggs. Two general
methods of protection have evolved; the eggs are either retained
within the mother as long as possible, or the eggs are surrounded by
a hard cover (Shelford, 1912). The method of incubating the eggs in
a brood sac within the mother possibly affords greater protection to
the species than does the deposition of a hardened oodtheca; certainly
the danger from insect parasites of the eggs would be largely elimi-
nated in this way. Sells (1842) reported that an odtheca of Blaberus
maderae [= Leucophaea maderae|, which he received from Jamaica,
contained 96 specimens of a small chalcid wasp ; some odthecae had a
round hole through the side of the capsule from which the wasps had
emerged. His statement that the odtheca had a keel with 16 denta-
tions indicates that he was dealing with an oviparous species rather
than a specimen of L. maderae. Bordage (1896) was undoubtedly
incorrect in stating that Blatta maderae [= L. maderae] was a host
of the egg parasite Evania desjardinsii [= E. appendigaster|, because
the odtheca is protected from the parasite within the body of the
cockroach.

However, the egg-laying rate is decreased among viviparous species
because the female cannot produce more egg cases during the gesta-
tion period, which usually takes more than a month. During em-
bryonic development the offspring of viviparous species are subject
to the vicissitudes that beset the mother. As we have shown, pre-
mature death of the mother may release the embryos from her brood
sac. However, if the female were killed before the embryos had de-
veloped sufficiently to maintain themselves without the protection of
the brood sac, the eggs would die with her.

The admitted success of the common oviparous cockroaches, in es-
tablishing themselves in practically every man-made niche, may in
part stem from the ability of the embryo to develop independently of
the female. Coupled with this are an increased egg-laying rate and
the ease with which egg cases may accidentally be transported far be-
yond the territory occupied by the mother. These factors may offset
any apparent advantage gained by viviparous cockroaches through
internal incubation of their eggs.

Shelford (1906) grouped several genera and species of cockroaches
according to whether the odthecal membrane is complete or incom-
plete. Hagan (1951) has suggested that an almost complete series of

NO. I2 REPRODUCTION OF COCKROACHES—ROTH AND WILLIS 35

cockroaches is available to illustrate a tendency toward the elimination
of the blattid odtheca: “The list could start with species dropping
the odtheca early, followed by species retaining a protruded odtheca
until shortly before hatching occurs. Then there are species with in-
ternally retained ootheca [sic] with varying degrees of fragility to
Diploptera whose odtheca is most delicate and imperfect, and finally
ending with species which are said to secrete none at all” (Imms,
1925).

We have arranged the types of oviposition behavior found in cock-
roaches in a similar series to show a progressive tendency toward
retention of the eggs within the body of the female until hatching.
We, like Hagan, do not imply that this is an evolutionary series, but
it serves to summarize what is known about blattid oviposition.

Ancestral types:

(1) Cockroaches with long ovipositors that presumably deposited single (?)
eggs not enclosed in an odtheca.

(2) Cockroaches that deposited eggs enclosed in an odtheca.

Present-day cockroaches:

(1) Odtheca extruded, not rotated, carried by female for only a short time,
then deposited and abandoned. (Periplaneta americana, Blatta orientalis,
Eurycotis floridana, Supella supelliectilium.)

(2) Odtheca extruded, rotated, carried by female for a longer period than in
first category, but eventually deposited a long time before hatching.
(Ectobius panzeri.)

(3) Odtheca extruded, rotated, carried by female until, or shortly before,
hatching. (Blattella germanica, Blattella vaga, Blattella humbertiana.)

(4) Odtheca extruded, rotated, and retracted into the brood sac where the
embryos develop until, or shortly before, hatching. (Nawphoeta cinerea,
Pycnoscelus surinamensis, Leucophaea maderae, Blaberus craniifer,
Gromphadorhina laevigata.)

(5) Odtheca possibly not extruded (see p. 33); eggs possibly directed from
oviduct into brood sac where they remain until hatching (Diploptera
dytiscoides.)

Coincident with the retention of the eggs within the body of the fe-
male is a reduction in the hardness, thickness, and extent of the walls
of the odtheca. In Blattella germanica the walls of the ootheca are
relatively thin, and premature dropping of the egg case may be detri-
mental to hatching, undoubtedly because of desiccation. Parker and
Campbell (1940) found that, although there may be a reduction in
hatching, some eggs in shriveled odthecae that had been removed
from the female did hatch. One such egg case hatched 24 days after
removal from the female. These workers found 36 percent complete
and 29 percent partial hatching of detached odthecae, kept under

36 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

laboratory conditions, compared to 70 percent complete and 3 percent
partial hatching in the controls.

On the other hand, the eggs of ovoviviparous cockroaches appar-
ently never hatch if the odtheca is expelled from the brood sac before
the embryos are well developed. The odthecae of these species are less
well developed than in Blattella germanica. Not only is the wall of the
odtheca thinner, but in several genera the odtheca is incomplete, being
absent along the anterior ends of the eggs. We have found that at a
very low humidity the eggs (in the odtheca) of Pycnoscelus suri-
namensis and Nauphoeta cinerea lost water much more rapidly than
the eggs of B. germanica, which in turn lost water more rapidly than
the eggs of Periplaneta americana.

We have noticed that in the laboratory Cariblatta lutea minima in-
variably deposited its odthecae on the cotton stoppers of water vials.
The only eggs that hatched were from odthecae left on the moist cot-
ton ; eggs in odthecae removed from the cotton and placed in dry vials
did not hatch.

EGG PARASITES

This, then, is reproduction in cockroaches ; with the oviparous forms
sexual behavior of the female culminates with the deposition of the
odthecae. But her entire effort to perpetuate the species may have
been in vain, for in spite of a hard odtheca which presumably pro-
tects the eggs, the eggs may be destroyed by various parasites. The
Diptera parasitic on cockroach eggs are Coenosia basalis Stein and
Megaselia sp. Edmunds (1952) reared these flies from two different
oothecae of Parcoblatta species.

In searching the literature we have found about 25 species, in about
I5 genera, of hymenopterous parasites of cockroach eggs. For ex-
ample, Tetrastichus hagenowiti (Ratz.) is a parasitic wasp which has
been reared from the odthecae of several species of cockroaches. This
small wasp pierces the odtheca with her ovipositor (pl. 9, fig. 67)
and deposits her eggs in the eggs of the cockroach. The wasp deposits
over half her eggs during the first 2 days after emergence (Roth and
Willis, 1954b). Development (pl. 9, fig. 68) is completed in from 23
to 57 days at about 30° C. or higher. Developmental time is a function
of the size of the parasite brood. The smaller the brood the longer the
wasps take to complete development. If sufficiently numerous, the
wasp larvae destroy all the cockroach eggs (pl. 9, fig. 71). With an
average of 48 wasps per odtheca all the cockroach eggs are eaten.
With an average of 18 wasps per odtheca not all the cockroach eggs are

NO. I2 REPRODUCTION OF COCKROACHES—ROTH AND WILLIS 37

eaten, but the uneaten eggs fail to hatch. When there is an average
of only eight wasps per odtheca, about eight cockroaches complete
development and hatch. Eventually the larvae pupate (pl. 9, fig. 69),
metamorphose into adults, and emerge from a hole chewed through
the side of the egg case (pl. 9, fig. 70). The wasps mate immediately
and the females soon seek out other odthecae and parasitize the eggs.

The evaniids or “ensign flies” are another group of parasites that
destroy cockroach eggs ; however, in contrast to other cockroach-egg
parasites, only one evaniid develops within each odtheca, the single
parasite larva destroying all the eggs. These wasps (pl. 10, figs. 72,
73) are frequently seen at windows, indicators of otherwise well-
hidden cockroach infestations (Edmunds, 1953).

SUMMARY

The reproductive behavior of most of the more than 3,500 species
of cockroaches is still unknown. Less than half a dozen species have
been studied intensively, and it is from these, the species most closely
associated with man, that most of our knowledge of the behavior of
the group comes. Enough additional information now exists, on some
of the less common forms, to enable us to summarize the reproduc-
tion of cockroaches of quite dissimilar habits.

Courtship—Male cockroaches prior to copulation engage in several
kinds of activity or display. In Periplaneta americana this is a re-
sponse to an odor from the female; in Blattella germanica it is a re-
sponse to a chemical substance on the body of the female which the
male detects by antennal contact. Among species in which the males
are alate, the male raises his wings, still folded and crossed, exposing
the dorsal surface of his abdomen. This activity is probably general
throughout the group; it has been seen in several genera: Blatia,
Blattella, Periplaneta, Supella, Blaberus, Leucophaea, and Nauphoeta.
Males of certain species perform characteristic body movements dur-
ing courtship. In Leucophaca maderae the male moves his body rap-
idly up and down against the substrate producing a tapping sound.
The male of Blaberus cranifer butts the female with his head and his
abdomen trembles. The male of Eurycotis floridana vibrates his body
from side to side and extends it posteriorly.

In response to the male’s display the females of many species apply
their mouthparts to the male’s dorsum; starting near the anal end of
his abdomen the female gradually works forward, apparently feeding
on a secretion on the surface of the abdomen, until she is astride the
male. This activity has been observed in these genera: Blatta, Blat-

38 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

tella, Blaberus, Ectobius, Supella, Eurycotis, and Nauphoeta. The
males of many genera have specialized tergal glands which secrete the
substance attractive to the female.

Copulation—When the female is astride the male, he pushes his
abdomen farther backward until he can make genital contact with her.
The male grasps the female with the aid of the hooked left phallomere
and moves out from under her. The cockroaches complete copulation
in the opposed position with their heads in opposite directions. This
final act of copulation has been seen in Blatia, Blattella, Parcoblatta,
Periplaneta, Eurycotis, Leucophaea, Nauphoeta, Polyzosteria, Supella,
and Blaberus.

Fertilization—Sperm transfer from male to female is accomplished
by means of a spermatophore which is elaborated by the male acces-
sory glands. The spermatophore is only formed during copulation,
and after it is transferred to the female the sperm migrate from the
spermatophore into the spermathecae of the female. Within a day or
longer the dry, shrunken, empty spermatophore drops from the fe-
male’s genital cavity. Spermatophore formation has been seen in
Blatta, Blattella, Periplaneta, Nauphoeta, Leucophaea, Eurycotis, and
Blaberus.

As each egg passes out of the oviduct into the genital chamber of
the female, sperm are forced out of the spermatheca against the micro-
pylar end of the egg. The eggs are fertilized as they pass along the
vestibulum.

Parthenogenesis.—Parthenogenesis can occur among several species
of cockroaches. It occurs regularly in Pycnoscelus surinamensis in
North America and Europe. We have found some parthenogenesis
in Supella, Blattella, Periplaneta, and Blatta. Unfertilized eggs of
only Blatta orientalis and Periplaneta americana hatched normally.

The odtheca.—The eggs of all cockroaches, so far as we know, are
invested with a covering, the odtheca. This may be hard and protec-
tive as in Periplaneta, Blatta, Eurycotis, Blattella, Parcoblatia, Cari-
blatta, Ectobius, Supella, and others, or the odtheca may be reduced
in thickness and/or enclose only part of the eggs as in Blaberus, Leu-
cophaea, Nauphoeta, Pycnoscelus, Diploptera, and others. Reduction
of the odtheca is associated with viviparity; hard protective odthecae
are characteristic of oviparous forms.

The odtheca of Blatta orientalis, and presumably of other species, is
formed from a protein, a phenol, and an oxidase secreted by the col-
leterial glands. This material stretches around the eggs as they are
erected in a double row in the forming odtheca. In oviparous species
the dorsal edge of the odtheca is modified into a series of respiratory

NO. 12 REPRODUCTION OF COCKROACHES—ROTH AND WILLIS 39

chambers, usually one per egg, which admit air to the developing
embryo.

Oviposition and hatching—The odthecae of some oviparous forms
are not rotated but are carried for a few days or less until the female
drops them (e.g., Blatta, Periplaneta, Eurycotis, Cariblatia, Supella).
Some of these cockroaches cement the odtheca to the substrate and
cover it with debris. Other oviparous cockroaches rotate the com-
pleted odtheca and retain the proximal end between the plates of the
genital chamber for some time before deposition (e.g., Ectobius pan-
geri), or retain the egg case until or shortly before the eggs hatch
(e.g., Blattella).

Nymphs hatch from the eggs of oviparous cockroaches by a con-
certed swallowing of air, thereby increasing their bulk which spreads
apart the dorsal seam of the ootheca. At hatching each nymph is in-
vested with a membrane which it sheds while emerging from, or when
free of, the ootheca.

Viviparous cockroaches, which incubate their eggs in a brood sac
within the mother’s body, extrude the odtheca as it is formed around
the eggs. However, in contrast to the oviparous forms, the viviparous
cockroaches retract the completely formed egg case into a brood sac,
where it remains until, or shortly before, hatching (e.g., Pycnoscelus,
Nauphoeta, Leucophaea, Gromphadorhina, Blaberus). Hatching of
the eggs of viviparous cockroaches apparently may occur either within
the vestibulum, while the ootheca is being extruded, or shortly after
the odtheca is dropped by the female.

The types of oviposition among cockroaches may be arranged in a
series showing a tendency toward retention of the odtheca within the
body of the female until hatching: Shortly after extrusion the odtheca
may be dropped, sometimes buried or covered, and then abandoned ;
the egg case may be carried, extending from the female’s body, for
various periods up to and including hatching; the odtheca may be
retracted into a brood sac until or shortly before hatching.

Egg parasites—Although the odtheca of oviparous cockroaches
protects the eggs from desiccation, it does not prevent destruction of
the eggs by parasites, particularly Hymenoptera. The wasp egg para-
sites insert their ovipositors through the wall of the odtheca and de-
posit their eggs in or on the eggs of the cockroach. Frequently the
wasp larvae destroy all the cockroach eggs. The eggs of viviparous
cockroaches, being protected within the brood sac, are apparently not
subject to attack by these parasites.

40 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

ACKNOWLEDGMENTS

The writers thank George Riser and Albert Levy, of our labora-
tories, for their assistance. We thank the following for supplying us
with living insect material which enabled us to establish cultures of
several species and to make some of the observations recorded in this
paper: Dr. Berta Scharrer, Medical Center, University of Colorado,
for Leucophaea maderae; Paul. T. Riherd, Texas Agricultural Ex-
periment Station, for Blattella vaga; Dr. William L. Nutting, Har-
vard University, for Blaberus cranufer; R. H. Nelson, U. S. Depart-
ment of Agriculture, Beltsville, Md., for Nauphoeta cinerea; and Dr.
Lafe R. Edmunds, U. S. Public Health Service, Mitchell, Nebr., for
Tetrastichus hagenowiu. Thanks are also extended to the following
workers of the U. S. Department of Agriculture: Dr. B. D. Burks,
for confirming the identification of T. hagenowti, Miss Louella M.
Walkley, for identifying Prosevania punctata, and Dr. Ashley B.
Gurney for his interest and for reading the manuscript.

REFERENCES

Apair, E. W.
1923. Notes sur Periplaneta americana et Blatta orientalis L. (Orthop.).
Bull. Soc. Ent. Egypte, vol. 7, pt. 2, pp. 18-38.
ANTONELLI, G.
1930. La blatta nella igiene domestica. Riv. Soc. Ital. Igiene (Milan), vol.
52, Pp. 132-142.
BEEBE, W.
1925. Jungle days. 201 pp. New York.
BERLAND, L.
1924. Observations biologiques sur les Orthoptéres. Bull. Soc. Ent. France,
No. 4, pp. 70-72.
Bitter, R. S., and Witttams, O. B.
1949. Enteric organisms from the American cockroach. Journ. Infect. Dis.,
vol. 85, pp. 87-90.
BorpacgE, E.
1896. Sur les moeurs de 1’Evania desjardinsii, Blanch. Compt. Rend. Acad.
Sci. Paris, vol. 123, pp. 610-613.
BrRIDWELL, J. C.
1921. Notes and exhibitions. Proc. Hawaiian Ent. Soc., vol. 4, p. 454.
Brown, E. B.
1952. Observations on the life-history of the cockroach Ectobius panzeri
Stephens (Orth., Blattidae). Ent. Month. Mag., vol. 88, pp. 209-212.
Brunet, P. C. J.
1951. The formation of the ootheca by Periplaneta americana. I. The
micro-anatomy and histology of the posterior part of the abdomen.
Quart. Journ. Micr. Sci., vol. 92, pp. 113-127.
1952. The formation of the ootheca by Periplaneta americana. II. The

NO. 12 REPRODUCTION OF COCKROACHES—ROTH AND WILLIS 4I

structure and function of the left colleterial gland. Quart. Journ.
| Micr. Sci., vol. 93, pp. 47-60.
CAMPBELL, F. L.
1929. The detection and estimation of insect chitin; and the irrelation of
“chitinization” to hardness and pigmentation of the cuticula of the
American cockroach, Periplaneta americana L. Ann. Ent. Soc.
Amer., vol. 22, pp. 401-420.
Cao, G.
1898. Sul passagio dei microorganismi attraverso l’intestino di alcuni insetti.
Uff. Sanit., Naples, vol. 11, pp. 385-397.
CAuDELL, A. N.
1925. Pycnoscelus surinamensis Linnaeus (Orthoptera) ; on its nymphs and
the damage it does to rose bushes. Proc. Ent. Soc. Washington,
vol. 27, pp. 154-157.
CHAUVIN, R.
1949. Les stérols dans le régime alimentaire de certains Orthoptéres. Compt.
Rend. Acad. Sci. Paris, vol. 229, pp. 902-903.
CHoparp, L.
1919. L’accouplement et le fonctionnement de l’armure copulatrice des Blat-
tidae (Orthoptéeres). Bull. Soc. Zool. France, vol. 44, pp. 347-353.
1938. La biologie des Orthoptéres. Encycl. Ent., Sér. A, vol. 20, 541 pp.
(especially pp. 207-212, 217-222).
1950. Sur lanatomie et le développement d’une blatte vivipare. Eighth
Internat. Congr. Ent., 1948, Stockholm, pp. 218-222.
CLEVELAND, L. R.; Hatz, S. R.; Sanpers, E. P.; and Coir, J.
1934. The wood-feeding roach Cryptocercus, its Protozoa, and the symbiosis
between Protozoa and roach. Mem. Amer. Acad. Arts and Sci., vol.
17, pp. 185-342, 60 pls.
Corram, R.
1922. Observations on the Phyllodromine cockroach, Blattella supellectilium
Serv., in Khartoum. Ent. Month. Mag., vol. 58, pp. 156-158.
Cros, A.
1942. Blatta orientalis et ses parasites. I. Evania punctata Brullé; II.
Eulophus sp. Etude biologique. Eos, Rev. Espafiola Ent., vol. 18,
pp. 45-67.
Davis, W. T.
1930. Rearing the young of the viviparous cockroach, Panchlora cubensis.
Journ. New York Ent. Soc., vol. 38, pp. 85-88.
DeCoursey, J. D., and Wessterr, A. P.
1952. Effect of insecticides and other substances on oviposition by Aedes
sollicitans. Journ. Econ. Ent., vol. 45, pp. 1030-1034.
DewiTz, J.
1885. Ueber die Vereinigung der Spermatozoen mit dem Ei. Pfliigers Arch.
Ges. Physiol., vol. 37, pp. 219-223.
1886. Ueber Gesatzmassigkeit in der Ortsveranderung der Spermatozoen
und in der Vereinigung derselben mit dem Ei. Pfliigers Arch. Ges.
Physiol., vol. 38, pp. 358-385.
Epmunps, L. R.
1952. Some notes on the habits and parasites of native wood-roaches in
Ohio (Orthoptera: Blattidae). Ent. News, vol. 63, pp. 141-145.

42 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

1953. Some notes on the Evaniidae as household pests and as a factor in the
control of roaches. Ohio Journ. Sci., vol. 53, pp. 121-122.
FieLpine, J. W.
1926. Preliminary note on the transmission of the eye worm of Australian
poultry. Australian Journ. Exp. Biol. Med. Sci., vol. 3, pp. 225-232.
FISCHER, O. VON.
1928. Die Entwicklung von Periplaneta americana. Mitt. Naturf. Ges.
Bern., 1927, pp. V-VII.
Frock, R. A.
1941. Biological control of the brown-banded roach. Bull. Brooklyn Ent.
Soc., vol. 36, pp. 178-181.
GERBER, C. E.
1950. Studies of the rate and type of reproduction of the German cockroach.
Master’s thesis, Ohio State Univ., pp. 1-28.
Grier, H. T.
1936. The morphology and behavior of the intracellular bacteroids of
roaches. Biol. Bull., vol. 71, pp. 433-452.
1947. Growth rate in the cockroach Periplaneta americana (Linn.). Ann.
Ent. Soc. Amer., vol. 40, pp. 303-317.
Grrautt, A. A.
1911. Standards of the number of eggs laid by insects. Ent. News, vol. 22,
pp. 14-15.
1914. Standards of the number of eggs laid by insects (Orthop.)—XI. Ent.
News, vol. 25, p. 206.
1915. New fragments on some well-known insects (Col., Orth., Hem.).
Ent. News, vol. 26, pp. 53-56.
GorzE, J. A. E.
1782. Beitrage zur Verwandlungsgeschichte der Schaben. (Blatta Linn.).
Naturforscher, vol. 17, pp. 183-180.
GouLp, G. E.
1941. The effect of temperature upon the development of cockroaches. Proc.
Indiana Acad. Sci., vol. 50, pp. 242-248.
Goutp, G. E., and Deay, H. O.
1938. Notes on the bionomics of roaches inhabiting homes. Proc. Indiana
Acad. Sci., vol. 47, pp. 281-284.
1940. The biology of six species of cockroaches which inhabit buildings.
Purdue Univ. Agr. Exp. Stat. Bull. 451, 31 pp.
GraFrar, M., and MERTENS, S.
1950. Le role des blattes dans la transmission des salmonelloses. Ann. Inst.
Pasteur, vol. 79, pp. 654-660.
GrirFiTHs, J. T., and TauBer, O. E.
1942. Fecundity, longevity and parthenogenesis of the American roach,
Periplaneta americana L. Physiol. Zool., vol. 15, pp. 196-209.
GoupPra es D:
1947. On copulation and insemination in the cockroach Periplaneta ameri-
cana (Linn.). Proc. Nat. Inst. Sci. India, vol. 13, pp. 65-71.
Gurney, A. B.
1953. Distribution, general bionomics, and recognition characters of two
cockroaches recently established in the United States. Proc. U. S.
Nat. Mus., vol. 103, pp. 39-56.

NO. I2 REPRODUCTION OF COCKROACHES—ROTH AND WILLIS 43

Haass, E.
1889. Zur Anatomie der Blattiden. Zool. Anz., vol. 12, pp. 169-172.
Haser, V. R.

1920. Oviposition by a cockroach, Periplaneta americana Linn. (Orth.).

Ent. News, vol. 31, pp. 190-193.
Haan, H. R.

1939. Diploptera dytiscoides, a viviparous roach with elongate pleuropodia.
Ann. Ent. Soc. Amer., vol. 32, pp. 264-266.

1941. The general morphology of the female reproductive system of a vivip-
arous roach, Diploptera dytiscoides (Serville). Psyche, vol. 48,
pp. I-9.

1951. Embryology of the viviparous insects. 472 pp. New York.

RVALEEZ., P.

1885. Orientation de l’embryon et formation du cocon chez la Periplaneta

orientalis. Compt. Rend. Acad. Sci. Paris, vol. 101, pp. 444-446.
Heat, R. E.
. Personal communication (1953).
HEBARD, M.

1917. The Blattidae of North America north of the Mexican boundary.
Mem. Amer. Ent. Soc., No. 2, 284 pp.

Hers, W. B., and Netson, Y.
1913. The croton bug (Ectobia germanica) as a factor in bacterial dis-
semination. Amer. Journ, Publ. Health, vol. 3, 929-934.
Hormcren, N.
1903-1904. Ueber vivipare Insecten. Zool. Jahrb., Abt. Syst., Geogr. u. Biol.
der Thiere, vol. 19, pp. 431-468.
Hume, A. D.

1821. Quelques observations sur la blatte germanique (Blatta germanica

Fabr.). Essais Entomologiques, St. Petersburg, No. 1, pp. 1-18.
ILLINGworTH, J. F.

1915. Notes on Hawaiian roaches. Proc. Hawaiian Ent. Soc., vol. 3, pp.
136-140.

1942. An outbreak of cockroaches, Nauphoeta cinerea (Olivier), in Hono-
lulu. Proc. Hawaiian Ent. Soc., vol. 11, pp. 169-170.

JurecKA, B.

1950. Morfologia gruczotu grzybowatego Blatta orientalis L. (Morphology
of the mushroom-shaped gland of Blatta orientalis L.). Sprawoz-
dania z Posiedzen, Wydziatu IV, Nauk Biologicznych, Warszawa,
vol. 41, pp. 81-102. (In Polish with English summary.)

Kapy1, H.
1879. Beitrag zur Kenntnis der Vorgange beim Eierlegen der Blatta orien-
talis. Zool. Anz., vol. 2, pp. 632-636.
Karny, H. H.
1924. Beitrage zur Malayischen Orthopterenfauna. V. Bemerkungen ueber
einige Blattoiden. Treubia, vol. 5, pp. 3-19.
Kuatira, A.
1950. Spermatophore production in Blattella germanica L. (Orthoptera:
Blattidae). Proc. Roy. Ent. Soc. London, Ser. A, vol. 25, pp. 53-61.
Kemin, SelepZ
1933. Zur Biologie der amerikanischen Schabe (Periplaneta americana L.).
Zeitschr. Wiss. Zool., vol. 144, pp. 102-122.

44 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

Koncexk, S. K.

1924. Zur Histologie der Riickendruse unserer einheimischen Blattiden.

Zeitschr. Wiss. Zool., vol. 122, pp. 311-322.
Kramer, S., and WIGGLESworTH, V. B.

1950. The outer layers of the cuticle in the cockroach Periplaneta americana
and the function of the oenocytes. Quart. Journ. Micr. Sci., vol.
OI, pp. 63-73.

KupryAscHow, B. A., and PerrowsKaja, O. A.

1937. Vitamin E und Lebensfahigkeit der Keimzellen von Insekten (Ver-
suche an Periplaneta orientalis). Bull. Biol. Med. Exp. U.R.S.S.,
vol. 4, Pp. 490-402.

LaFon, M.

1951. Essais sur l’alimentation d’un insecte: Blatta orientalis L. I. Données
quantitatives sur la nutrition azotée. Physiol. Comp. Oecol., vol.
2, pp. 224-240.

LATREILLE, P. A.
[1803/1804]. Histoire naturelle, générale et particuliere, des crustacés et
des insectes, vol. 12, 424 pp. (especially pp. 91-98). Paris.
LAURENTIAUX, D.

1951. Le probléme des blattes Paléozoiques a ovipositeur externe. Ann.

Paléont., vol. 37, pp. 185-196.
Lawson, F. A.

1951. Structural features of the oothecae of certain species of cockroaches
(Orthoptera, Blattidae). Ann. Ent. Soc. Amer., vol. 44, pp. 269-285.

1952. Structural features of cockroach egg capsules. II. The ootheca of
Cariblatta lutea lutea. (Orthoptera: Blattidae). Ohio Journ. Sci.,
vol. 52, pp. 296-300.

1953. Structural features of cockroach egg capsules. III. The ootheca of
Eurycotis floridana (Orthoptera: Blattidae). Journ. Tennessee
Acad. Sci., vol. 28, pp. 28-33.

Lerroy, H. M.
1909. Indian insect life. 786 pp. London.
MWEAS, Wal.

1928. Notes on British Orthoptera, including Dermaptera, in 1927. En-

tomologist, vol. 61, pp. 78-81.
Mackerras, M. J., and Mackerras, I. M.

1948. Salmonella infections in Australian cockroaches. Australian Journ.

Cio Vol. TO, op. 525.
MatTrHe_ey, R.

1948. La formule chromosomiale de Pycnoscelus surinamensis L. Race bi-
sexuée et race parthénogénétique. Existence probable d’une par-
thénogénése diploide facultative. Arch. Jul. Klaus-Stift. Verebf.,
vol. 23, pp. 517-520.

Metcatr, C. L., and Furnt, W. P.

1939. Destructive and useful insects, their habits and control. 2d ed. 981 pp.

New York.
Mratt, L. C., and Denny, A.

1886. The structure and life-history of the cockroach (Periplaneta orien-

talis). 224 pp. London.

NO. 12 REPRODUCTION OF COCKROACHES—ROTH AND WILLIS 45

MincHIn, E. A.
1888. Note on a new organ, and on the structure of the hypodermis, in Peri-
planeta orientalis. Quart. Journ. Micr. Sci., vol. 20, pp. 229-233.
Moore, H.
1900. How long does Blatta orientalis, Linn., 2 carry its ootheca before de-
position? Ent. Rec. Journ. Var., vol. 12, pp. 79-80.
MortsseEt, P.
1946. Comportement de la blatte germanique dans un incubateur a tempéra-
ture constante. Ann. Acfas (Assoc. Can.-Franc. Avance. Sci.), vol.
12, pp. 88-89. (Abstract. )

Moree, C. C.
1o1l. The bacteriology of the cockroach. British Med. Journ., vol. 2, pp.
1531-1532.

Notanp, J. L.; Litry, J. H.; and BAumMAnnN, C. A.
1949. Vitamin requirements of the cockroach Blatella germanica (L.). Ann.
Ent. Soc. Amer., vol. 42, pp. 154-164.
Nuttine, W. L.
1953. Observations on the reproduction of the giant cockroach, Blaberus
crantifera Burm. Psyche, vol. 60, pp. 6-14.
Parker, B. M., and Campsett, F. L.
1940. Relative susceptibility of the ootheca and adult female of the German
cockroach to liquid household insecticides. Journ. Econ. Ent., vol.
33, PP. 610-614.
Pessoa, S. B. and Correa, C.
1928. Nota sobre a biologia da Rhyparobia maderae, Fabr. Rev. Biol. Hy-
giene, Sao Paulo, vol. 1, No. 3, pp. 83-87.
leioueay Ie (CS
1944. A roach is born. Fauna, vol. 6, pp. 50-52.
Porter, A.
1918. A survey of the intestinal entozoa, both protozoal and helminthic, ob-
served among natives in Johannesburg, from June to November,
1917. South African Inst. Mem., No. 11, 58 pp.
1930. Cockroaches as vectors of hookworms on gold mines of the Wit-
watersrand. Journ. Med. Assoc. South Africa, vol. 4, pp. 18-20.
Pryor, M. G. M.
1940a. On the hardening of the ootheca of Biatta orientalis. Proc. Roy.
Soc. London, Ser. B, vol. 128, pp. 378-393.
1940b. On the hardening of the cuticle of insects. Proc. Roy. Soc. London,
Ser. B, vol. 128, pp. 303-407.
Pryor, M. G. M.; Russet, P. B.; and Topp, A. R.
1946. Protocatechuic acid, the substance responsible for the hardening of
the cockroach ootheca. Biochem. Journ., vol. 40, pp. 627-628.
Qanpri, M. A. H.
1938. The life-history and growth of the cockroach Blatia orientalis, Linn.
Bull. Ent. Res., vol. 20, pp. 263-276.
RAMME, W.
1951. Zur Systematik Faunistik und Biologie der Orthopteren von Siidost-
Europa und Vorderasien. Mitt. Zool. Mus. Berlin, vol. 27 for 1950,
431 pp.

46 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

Ramsay, J. A.
1935. The evaporation of water from the cockroach. Journ. Exp. Biol., vol.
12, pp. 373-383.
RAGE.
1924. The biology of the roach, Blatia orientalis. Trans. Acad. Sci. St.
Louis, vol. 25, pp. 57-79.
1940a. The life history of the American cockroach, Periplaneta americana
Linn. (Orthop.: Blattidae). Ent. News, vol. 51, pp. 121-124, 151-
155, 186-180, 223-227, 273-278.
1940b. The life history of the wood-roach, Parcoblatta pennsylvanica De-
Geer (Orthoptera: Blattidae). Ent. News, vol. 51, pp. 4-9, 33-35.
1943. How the cockroach deposits its egg-case; a study in insect behavior.
Ann, Ent. Soc. Amer., vol. 36, pp. 221-226.
1944. A note on the period of incubation of eggs of the cockroach Blattella
germanica L. Canadian Ent., vol. 76, p. 212.
1945. Notes of the life history of Periplaneta fulignosa Serv. Psyche, vol.
52, pp. 107-108.
1947. Life history notes on the wood-roach, Ischnoptera deropeltiformis
Brunner. Ent. News, vol. 58, pp. 1-4.
ReEuN, J. A. G.
1931. African and Malagasy Blattidae (Orthoptera).—Part I. Proc. Acad.
Nat. Sci. Philadelphia, vol. 83, pp. 305-387.
1945. Man’s uninvited fellow traveler—the cockroach. Sci. Month., vol. 61,
Pp. 265-276.
ReEuN, J. A. G., and Heparp, M.
1927. The Orthoptera of the West Indies: Blattidae. Bull. Amer. Mus.
Nat. Hist., vol. 54, pp. 1-320.
REuN, J. W. H.
1951. Classification of the Blattaria as indicated by their wings (Orthop-
tera). Mem. Amer. Ent. Soc., No. 14, 134 pp., 13 pls.

Ricci, M.
1950. Note sulla biologia di Blatta orientalis, L. Riv. Parassitol., vol. 11,
Pp. 219-231.
Ritey, C. V.
1890. A viviparous cockroach. Proc. Ent. Soc. Washington, vol. 2, pp.
129-130.

1891a. Further notes on Panchlora. Insect Life, vol. 4, pp. 119-120.
1891b. A viviparous cockroach. Insect Life, vol. 3, pp. 443-444.
RoEsEr, G.
1940. Zur Kenntnis der Lebensweise der Gewachshausschabe Pycnoscelus
surinamensis L. Die Gartenbauwissenschaft, vol. 15, pp. 184-225.
RosENFELD, A. H.
1910. Blattid notes. Journ. Econ. Ent., vol. 3, pp. 100-101.
Ross, H. H.
1929. The life history of the German cockroach. Trans. Illinois State Acad.
Sci., vol. 21, pp. 84-93.
Ror, L. M.
1952. The tergal gland of the male cockroach, Supella supellectilium. Journ.
Morph., vol. 91, pp. 469-478.
Rotx, L. M., and Wits, E. R.
1952. A study of cockroach behavior. Amer. Midl. Nat., vol. 47, pp. 66-120.

NO. 12 REPRODUCTION OF COCKROACHES—ROTH AND WILLIS 47

1954a. Anastatus floridanus (Hymenoptera: Eupelmidae) a new parasite on
the eggs of the cockroach Eurycotis floridana. Trans. Amer. Ent.
Soc. (in press).

1954b. The biology of the cockroach egg parasite, Tetrastichus hagenowit
(Hymenoptera: Eulophidae). Trans. Amer. Ent. Soc. (in press).

SAupPE, R.

1928. Zur Kenntnis der Lebensweise der Riesenschabe Blabera fusca Brun-
ner und der Gewdchshausschabe Pycnoscelus surinamensis L.
Zeitschr. Angew. Ent., vol. 14, pp. 461-500.

SCHARRER, B.

1943. The influence of the corpora allata on egg development in an orthop-
teran (Leucophaea maderae). Anat. Rec., vol. 87, p. 471. (Ab-
stract. )

1946. The relationship between corpora allata and reproductive organs in
adult Leucophaea maderae (Orthoptera). Endocrinology, vol. 38,
pp. 46-55.

1951. The woodroach. Sci. Amer., vol. 185, pp. 58-62.

——. Personal communication (1953).

ScHwaBseE, C. W.

1949. Observations on the life history of Pycnoscelus surinamensis (Linn.),
the intermediate host of the chicken eyeworm in Hawaii. Proc.
Hawaiian Ent. Soc., vol. 13, pp. 433-436.

Scunpper, S. H.

1890. [Panchlora nivea.] Psyche, vol. 5, p. 405.
SEIN, F., Jr.

1923. Cucarachas. Puerto Rico Insular Exp. Stat. Circ. 64, 12 pp.
Sess, C.F,

1896. The breeding habits of Periplaneta orientalis, in Proceedings of
Meetings, The Entomological Section, Academy of Natural
Sciences, Philadelphia, March 26, 1896. Ent. News, vol. 7, pp. 148-
150.

SELLS, W.

1842. Note respecting the egg-cases of Blattae. Trans. Ent. Soc. London,

vol. 3, pp. 103-104.
SHELForD, R. W. C.

1906. Studies of the Blattidae. VI. Viviparity amongst the Blattidae. Trans.
Ent. Soc. London, year vol. 1906, pts. 3/4, pp. 500-514.

1909. Blattidae, in Die Fauna Siidwest-Australiens. Ergebnisse der Ham-
burger siidwest-australischen Forschungsreise 1905, by W. Michael-
sen and R. Hartmeyer, vol. 2, pp. 129-142.

1912. The oothecae of Blattidae. Ent. Rec. Journ. Var., vol. 24, pp. 283-
287.

SrxorA, H.

1918. Zur Bedeutung der Riickendriise des Mannchens bei der Kiichen-
schabe (Phyllodromia germanica). Zeitschr. Angew. Ent., vol. 4,
PP. 374-375.

Snoperass, R. E.

1935. Principles of insect morphology. 667 pp. New York.

1937. The male genitalia of orthopteroid insects. Smithsonian Misc. Coll.,
vol. 96, No. 5, 107 pp.

48 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

1952. A textbook of arthropod anatomy. 363 pp. Ithaca, N. Y.
SPENCER, G. J.

1943. On the oviposition habits of the Australian cockroach, Periplaneta
australasiae (Fab.). Proc. Ent. Soc. British Columbia, vol. 40, pp.
29-30.

STEwart, A. M.

1925. Blabera cubensis (Orthoptera) and its odtheca. Entomologist, vol.

58, pp. 57-58.
SUOMALAINEN, E.

1946. Die Chromosomenverhialtnisse in der Spermatogenese einiger Blat-
tarien. Ann. Acad. Sci. Fennicae, Ser. A, sect. IV, No. 10, pp. 1-60,
6 pls.

SyvERTON, J. T.; FrscHer, R. G.; Smiru, S. A.; Dow, R. P.; and Scnoor, H. F.

1952. The cockroach as a natural extrahuman source of poliomyelitis virus.
Fed. Proc., vol. 11, p. 483.

TAKAHASHI, R.

1940. Habits and nymph of Phyllodromia humbertiana de Sauss. (Blattidae,
Orthoptera). Trans. Nat. Hist. Soc. Formosa, Taihoku, vol. 30,
pp. 26-29. (In Japanese.)

Terry, F. W.
1910. Im Entomological program. Proc. Hawaiian Ent. Soc., vol. 2, p. 145.
Tuomas, T.

1949. The life history of the greenhouse roach. Proc. Louisiana Acad. Sci.,

vol. 12, pp. 59-65.
Van Wyk, L. E.

1952. The morphology and histology of the genital organs of Leucophaea
maderae (Fab.) (Blattidae, Orthoptera). Journ. Ent. Soc. South
Africa, vol. 15, pp. 3-62.

WaASSILIEFF, A.

1907. Die Spermatogenese von Blatta germanica. Arch. Mikros. Anat., vol.

70, pp. 1-42.
WHEELER, W. M.

1889. The embryology of Blatta germanica and Doryphora decimlineata.

Journ. Morph., vol. 3, pp. 291-386.
Waite, M. J. D.

1951. Cytogenetics of Orthopteroid insects, it Advances in Genetics, ed. M.

Demerec, vol. 4, pp. 267-330. New York.
WiccLEsworTH, V. B., and BeAmeEnt, J. W. L.

1950. The respiratory mechanisms of some insect eggs. Quart. Journ.

Micr. Sci., vol. 91, pp. 420-452.
Wurtz, J.

1920. Biologie und Bekampfung der deutschen Schabe (Phyllodromia ger-
manica L.). Monographien zur angewandten Entomologie, Nr. 5.
Zeitschr. Angew. Ent., vol. 7, No. 1, 140 pp.

Wotcortt, G. N.

1950. The insects of Puerto Rico. Journ. Agr. Univ. Puerto Rico, vol. 32

for 1948, pp. 1-224,
Woopszury, E. N.

1938. Test methods on roaches. Soap and San. Chem., vol. 14, No. 8, pp.

86-90, 107, 100.

NO. 12 REPRODUCTION OF COCKROACHES—ROTH AND WILLIS 49

Woop-Mason, J.
1878. [Viviparity in a cockroach.] Proc. Ent. Soc. London, year vol. 1878,

Dipl
Woonrurf, L. C.
1938. Observations on roach reproduction. Journ. Kansas Ent. Soc., vol.
II, pp. 94-96.

ZABINSKI, J.
1933. Fonctionnement des différentes parties des appareils copulateurs
chitinés males et femelles de la blatte (Periplaneta orientalis L.).
Compt. Rend. Soc. Biol., vol. 112, pp. 598-602.
ZAPPE, M. P.
1918. A cockroach pest of greenhouses. Pycnoscelus (Leucophaea) suri-
namensis Linn. Connecticut Agr. Exp. Stat. Bull. No. 203, pp. 302-
313.
ZIMMERMAN, E. C.
1948. Insects of Hawaii. Apterygota to Thysanoptera, inclusive, vol. 2, 475
pp. (especially pp. 76-98). Honolulu.

i
i
ees
5 4 ‘ # , donee
it PC OR RICE En HOE CS RGD aT Aan
Li Bol, ah ~ ' Titi j b Wl
\ i
: : i
‘
: i i Fi ‘
Sf i h
{
sy ¢ 4 ane te ‘ U
giver tr ¥

*
Fre a BO eyey F ‘ ‘
i 4 he ROL i Rees
Pi ey? OW Pty ek ie hinges ge i
iv ; ;
:
' 4 ¥ 4 f
y AST SS i at
‘ Ble ( 5 i
; fy tg rane ; stn | } f ;
eee |
{ “ f re 7 oy i
thy ie,
¥ ‘
*
‘
q
) i
y
f
F
{
{
f
i i 7 \ i £5
Ths fee ‘ ‘ual v3 sek, y i]
iy

PVATES

PLATE I

Fig. 1. The male of Blattella germanica in a courting position (induced by
touching his antennae with the isolated antennae of a female). > 1.8.

Fig. 2. The female of Blattella germanica feeding on the male’s tergal-gland
secretion. < 1.8.

Figs. 3-5. Courting behavior of Eurycotis floridana. 3, The male (top) vibrat-
ing his body while standing near the female and exposing the intersegmental
membrane between his sixth and seventh tergites. 4, The female has com-
menced feeding on the dorsal surface of the male. 5, The female in working
her mouthparts over male’s dorsum has nearly placed herself in a position
from which male can initiate copulation. 1.1.

(Figures 1 and 2 from Roth and Willis, 1952.)

SMITHSONIAN MISCELLANEOUS COLLECTIONS VOLs 122hNOs T2naP leet

(See opposite page for legend.)

Wd, rile 72

NO.

VOL. 122,

SMITHSONIAN MISCELLANEOUS COLLECTIONS

‘ZI X ynoqy

‘(SMOIIB) SPUL]S [B919} ISIoJOeIeYD BY} SUTMOYS SayIVOIYIOD a[eu JO satdads vaI1Y} JO SusOpqe dy} JO suUOTIOd JO SMaIA [eSIOG “g-9

‘aejas Jo sdnois oxf]-ysnaq 94} soN0U Suiny1Ia}0gGns D]JaGNgy ‘g “vHvA vyJaVIwLg ‘ZL “PIUDiUW41aH vpIIn1DTg “9

Sl]

Wa, (alts 3

NO.

VOES 122;

SMITHSONIAN MISCELLANEOUS COLLECTIONS

> DDD YAO G ‘V1i-€1

: DUDpL4LoLf

SUJOINAN GT

‘Or-0

whe

O xX

YI-I1

fo XK MOLA [BIJUDA ‘VI - MOA [eIOeT “EI

“MOIA [BIJUOA “EI + MOIA [eRSIOC] ‘TI > AO fWUDAI SNAIQD]G CZI-11 ‘00 “MOTA [P19}e] “OI » MOIA [PSO] ‘O

x
“\

SoINSY UL Jo] 94} UO JAq OI puke GH SoINSY UT }YSII oY} UO ST oe OY], ‘eyNdoOD UT sayIVOIYIOD “PI-6 ‘s

Oly]

PLATE 4

Figs. 15-16. Accessory sex glands of male Blattella germanica. 15, From a vir-
gin male 37 days old. 16, From a male that copulated 3 hours prior to dis-
section. 6.8. (From Roth and Willis, 1952.)

Fig. 17. Spermatophore of Eurycotis floridana which was removed from the
female just after copulation and kept for 18 hours on moist paper before be-
ing photographed. 3.6.

Figs. 18-21. Ovaries of Periplancta americana. 18, From a newly emerged adult.
19, Shortly before oviposition. 20, During oviposition; eggs are in the ovi-
ducts (arrows); the partially formed odtheca (top) was removed during
the dissection. 21, Just after formation of an odtheca. 2.2.

Figs. 22-24. Ootheca of Blaberus crantfer. 22, Abdomen of a female dissected to
show the large brood sac or uterus which was cut open to reveal the en-
closed odtheca; the ootheca was partially extruded by involuntary contrac-
tions of the female while she was being dissected : a=ovary ; b=oviduct ; c=
colleterial glands; d=uterus or incubation pouch; e=o0theca. 2.1. 23-24,
Two views of an odtheca removed from the brood sac; the embryos are
visible through the thin transparent membrane which forms the odtheca.

SKA.

PL. 4

1225 NOe 12,

VOL.

SMITHSONIAN MISCELLANEOUS COLLECTIONS

rain en Ae 88 gy,

"pacino tte A ya

or legend.)

o
ge

(See opposite pa

PLATE 5

Figs. 25-34. A female of Eurycotis floridana in the process of forming an
oo0theca. Elapsed time during odtheca formation is given in hours (hrs.)
and minutes (min.). 25, 0 hrs. (odtheca just beginning to form; the remain-
ing values are taken from this as the starting point.) 26, 1 hr. 30 min. 27,
2 hrs. 20 min. 28, 3 hrs. 5 min. 29, 3 hrs. 40 min. 30, 4 hrs. 10 min. 31, 5 hrs.
10 min. 32, 20 hrs. 22 min. 33-34, 21 hrs. 5 min. (25-32, Dorsal views; 33,
lateral view; 34, ventral view). Natural size.

SMITHSONIAN MISCELLANEOUS COLLECTIONS VOLE L227 NOs 2 sIPicse5

(See opposite page for legend.)
( I : :

PLATE 6

Figs. 35-38. Supella supellectilium. 35, Normal ootheca (top view) containing
well-developed embryos; the black eyes of the embryos show through the
walls of the egg case. X 6.1. 36, Odtheca containing unfertilized eggs; one
(arrow ) has developed but, failing to hatch, died and turned dark. & 6.1. 37,
Ootheca shown in figure 36 with one wall removed to reveal the well-de-
veloped embryo. 6.1. 38, Embryo shown in previous two figures re-
moved from the ootheca. X 9.

Figs. 39-40. Blattella germanica. 39, Ootheca containing unfertilized eggs one
of which has developed (arrow). 6.1. 40, Developed embryo removed
from the odtheca. 9.

Figs. 41-45. A parthenogenetically developed embryo of Supella supellectilium
which, when dissected out of the odtheca (41), succeeded in shedding (42-
45) its embryonic membrane (44, arrow). This individual later developed
into an adult female. x 9.

Figs. 46-48. Hatching of a parthenogenetically developed egg of Periplancta
americana, The embryonic membrane (47, arrow) was left behind attached
to the odtheca. Two embryos were caught between the lips of the keel and
failed to hatch. > 7.7.

SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122, NO. 12, PL. 6

(See opposite page for legend.)

IPLANiNDy 9)

Figs. 49-58. Periplaneta americana. 49, Odtheca containing 16 eggs; there are
16 respiratory chambers and ducts in the keel. 50, Dorsal aspect of odtheca.
51, Anterior end of odtheca (portion held in the oothecal chamber after the
ege case is formed). 52, Posterior end of o6theca showing pattern impressed
by apical lobes of female’s genitalia. 53, Portion of the keel showing some
of the ducts and evaginated respiratory chambers (X 11.3). 54, Odtheca
which contained seven eggs; there are Io “teeth” or evaginations in the keel
and 11 ducts; the tooth (arrow) of one of the ducts is missing. 55, Ootheca
which contained 13 eggs; there are only 13 teeth but 14 ducts, one of the
teeth (arrow) being missing. 56-57, Oothecae in which the keels have been
partly (56) and completely (57) eaten by cockroaches. 58, Eggs which were
deposited without the formation of an o0theca. All figures except 53 are
X 4-5:

Figs. 59-60. Blatta orientalis. 59, Normal ootheca. 60, Abnormal egg case
which did not harden or develop normal pigmentation; respiratory chambers
and ducts were not molded in the keel.

Figs. 61-62. Blattella germanica. 61, Newly formed o6theca that partially col-
lapsed 2 days after it had been removed from the female. 62, Two attached
oOthecae ; eggs from the egg case on the left have hatched. Notice that the
first ootheca rotated to the right; the keel of the second is upward. X 4.5.

SMITHSONIAN MISCELLANEOUS COLLECTIONS WO 2, INhols. aly, IS 7/

ge HH ABASMARATRS WH
‘

BPM LE PDP LBE BEDARD PLO
; ae on," :

f..

(See opposite page for legend.)

PLATE 8

Figs. 63-65. Blattella vaga. Dorsal (63) and ventral (64) views of a female
carrying an ootheca. The keel is turned toward the female’s right. x 5.8.
65, Newly hatched nymphs clustering around and climbing over the female.
This female was still carrying the odtheca (keel to the right) at hatching,
but dropped it soon after this photograph was taken. The nymphs crawled
all over the mother and seemingly fed on the greasy material covering the
surface of her body; the female raised her wings and some of the nymphs
crawled under them on the dorsal surface of her abdomen. 5.8.

SMITHSONIAN MISCELLANEOUS COLLECTIONS VOLES 122) NO. 22) PES 8

(See opposite page for legend.)

VOL. 122, NO. 12, PL. 9

SMITHSONIAN MISCELLANEOUS COLLECTIONS

*S & ‘sdseMm oy} Aq
S889 YOROAYIOD dy} JO UOTJONAYSap ajafduiod MoYs 0} poudsdo (O04 “SY UL) BIBY}IOO ‘14 “6'O K “SurS1oiws ut sdseM zNpe vy} Aq opeur spoy sul
-Moys BoayI0Q ‘OZ *(:6°9 X ‘69) aednd pue (‘Sb X ‘gg) seAre] dseM Moys 0} pouado RIB4IOQ “69-89 “O'9 K “suljtsodiao dseM o[eulay Ao)
‘mousy snysysp.Ajo | ASeM dy} Aq pozipisesed YIeoIYIOI UvITIOUTY OY} JO 9BI9YIOO “14-49 “Sly OS & “OAIGUID YORI JO Peay IY} dAOqe

v994}00 dy} OJUL UadO sjonp ArOyerIdsat oY, “SOAIGUIa podOoJoAdp-[JaM JO SMOT OM} MOYS 0} usdo JNnd YUDIIAIWD DJIUD]G14d q JO BIIYJOC “OO “Sly

POOP SOO trto.

=
ao

%
7
6

. 9
©
#
=

-¥

bid

Boe

SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122, NO. 12, PL. 10

Figs. 72-73. The evaniid Prosevania punctata (Brullé) which parasitizes the
eggs of the American and oriental cockroaches. 72, Male with an ootheca
of Periplaneta americana. X 4. 73, Female with an o0theca of Blatta orien-
talis. X 3.3. Notice the size of these wasps in relation to the oOthecae and
compare with figure 67.

PLATE II

Figs. 74-79. Pycnoscelus surinamensis forming and retracting an egg Case,
August 18, 1953. Notice position of her head with the frons ventrad. 74,
9:30 a.m. The female is forming and simultaneously extruding the egg case;
the seam of the egg case is dorsad. 75, 9:45 a.m. Maximum extent to which
the female extruded her egg case; about 24 eggs protrude beyond the end
of the abdomen. The axes of the eggs are still vertical. 76, 9:55 a. m.
Within a period of 10 minutes the female rotated the egg case, seam to her
left, and retracted most of it into her brood sac. 77, 10:00 a.m. 78, 10:01 a.m.
79, 10:01+ a.m. The egg case was completely retracted by 10:01.5 a.m., at
which time the female raised her head and scurried off. The female was de-
capitated September 21, 1953, at which time she expelled this egg case. Five
nymphs, of 21 that developed, hatched, and there were 11 undeveloped eggs.

Figs. 80-85. Premature expulsion of an egg case by a decapitated female of
Pycnoscelus surinamensis with concurrent hatching of the eggs. Series of
photographs taken within a 5-minute period. 80, In the few seconds that
elapsed between decapitating the insect and placing it beneath the camera,
the egg case was ejected about half its length. 81, Nymphs began to emerge
through the suture along the dorsal edge of the egg case before it had been
completely expelled. 82-85, Sequence of hatching of 15 nymphs from the
dropped egg case. A few other eggs had developed but the nymphs failed
to hatch. The embryonic membrane shed by one of the nymphs is indicated
by arrows.

All figures X 1.0.

122) INO. 12s Pie, a

VOL.

SMITHSONIAN MISCELLANEOUS COLLECTIONS

(“pus

AC y

ased oysoddo 90¢ )

PLATE 12

Figs. 86-90. Nauphoeta cinerea forming and retracting an egg case, August 18

and 10, 1953. 86, 1:30 p.m., August 18. The female is forming and extruding
the egg case. The cephalic ends of the eggs are dorsad at this time. 87, 2:30
p.m. Maximum extent to which the female extruded her egg case; about
30 eggs are visible beyond end of her abdomen. The egg case curves strongly
ventrad, and its distal end twists to the side from contact with the substrate.
The axes of the eggs in the proximal end of the odtheca have been tilted
slightly to the female’s left. 88, 2:45 p.m. The female has retracted part of the
ege case. The axes of the eggs now lie in a horizontal plane. 89, 4:40 p.m.
Extent of retraction while the insect was under direct observation. 90, 7 :25
a.m., August 19. The female had completely retracted the egg case into the
brood sac during the night. The female carried this odtheca until September
II, 1953, when she expelled it. Only 10 embryos had developed; about 15
eges did not develop. X 1.4.

g. or. Female of Nauphoeta cinerea killed 30 minutes after she had begun to

retract the ootheca into her brood sac. 1.0.

g. 92. Dissection of female in figure 91 showing anterior end of odtheca lying

in the brood sac which has been cut open. X 1.9.

. 93. Female of Nawphocta cinerea dissected 27 days after she had formed an

ootheca. The o6theca fills nearly the entire body cavity. Notice the well-
developed embryos with dark eyes and mandibles which show through the
transparent wall of the odtheca. 1.0.

Vig. 94. The odtheca in figure 93 removed from the female’s brood sac. Notice

the two groups of undeveloped eggs at the left. 3.6.

NO. 12, PL. 12

122)

VOL.

SMITHSONIAN MISCELLANEOUS COLLECTIONS

(‘puaso

Ioy

35e

OHS! \ddo 99S)

SMITHSONIAN MISCELLANEOUS COLLECTIONS
VOLUME 122, NUMBER 13

Roebling Fund

Pf eoaiNGTON, BD. C.; PRECIPITATION
OF 1953 AND 1954

By
C. G. ABBOT

Research Associate, Smithsonian Institution

(Pustication 4170)

CITY OF WASHINGTON
PUBLISHED BY THE SMITHSONIAN INSTITUTION
APRIL 20, 1954

b
nee
bp
=>
= 8
Ss
2
o
a
a
»

Roebling Fund

WASHINGTON, D. C., PRECIPITATION OF
1953: AND 1954,
By C. G. ABBOT

Research Associate, Smithsonian Institution

This is the tenth year of these publications regarding precipitation
on individual days at Washington, D. C. The distribution of precipi-
tation at Washington in 1952 and 1953 was very different from that
of averages which appeared representative of 18 years preceding 1952.

In the year 1953 only the months January, April, September, Octo-
ber, and December followed the distribution of precipitation, as re-
gards “preferred” days, that prevailed in the majority of months for
all the 18 years preceding 1952. For the year 1953, the average precipi-
tation falling on “preferred” days was but 75 percent of the average
precipitation on all other days. During the 18 years preceding 1952,
that ratio averaged 146 percent, as against 142 percent expected. It is
true, however, that if the months of March and May were omitted
altogether from 1953 the ratio would be above unity, at about 110 per-
cent. In March it rained 3.42 inches from the 24th to the 26th, and in
May 5.49 inches from the 4th to the 6th. These floods upset those
months.

Last year I published a chart purporting to show the distribution of
Washington precipitation through the average 27-day cycle of 1952.
In some way, which I cannot now trace, I got the phases of that graph
completely wrong. I have now redrawn it (fig. 1, curve b) and also
one to represent the distribution that occurred in 1953 (curve c).
Along with them, I include a graph (curve a, heavy line) of the
average distribution which prevailed from 1924 to 1941, when the
basis for these forecasts was recorded. All three graphs are on the
same scale of ordinates, representing the average inches of precipita-
tion per day of the individual days of the 27-day cycle.

It is surprising to see that in both 1952 and 1953 (graphs b and c)
a high peak of precipitation occurs on the eleventh day of the cycle.
No such feature occurs in the graph a representing the years 1924 to
1941. That this high peak occurs on the identical day of the cycle for

SMITHSONIAN MISCELLANEOUS COLLECTIONS, VOL. 122, NO. 13

2 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

1952 and 1953 is the strongest proof yet appearing of the veridity of
the cycle, as an actual cosmic phenomenon. At the same time it shows
that a remarkable change of conditions of some sort occurred after
1951, as compared with the years 1924 to 1951 inclusive. And yet a
trifling decrease of the decimal .0074 would bring the highest peak of
curve a on the eleventh day in 1952 and 1953, as actually found in
curves b and c.

TABLE 1.—Washington precipitation 1954

Jan. Feb. Mar. Apr. May June
1 eats 7 3 2, 29 25 22 18
1 Loesch 8 4 3, 30 26 23 19
| ree ae 9 5 4, 31 27 2 20
1 EVAR Pen taeg 10 6 5 1, 28 25 2!
Vein: II 7 6 2,29 26 22
D4 0 Den yemers 18 14 13 9 6 2, 20
Ie ar 19 15 14 a0) i 3, 30
OV dan 21 17 16 12 fe) 5
VA ee 23 19 18 14 II 7
RV EET siete: 24 20 19 15 12 8
xox 1, 28 24 23 19 16 12
EXCXOVAl eee 5 1, 28 27 23 20 16
MEOW Eee 6 2 I, 28 24 21 17
July Aug Sept. Oct. Nov Dec
Tote 15 II 7 4, 31 27 24
1 Geet 16 12 8 5 1, 28 25
1B0 Reeser 17 13 9 6 2, 29 26
TGV posers 18 14 10 7 3, 30 27
Wise 19 15 II 8 4 1, 28
DG ae Sec 26 22 18 15 Il 8
41 1 Wee aerate 27 23 19 16 12 9
DVR ats 2,29 25 21 18 14 II
DANVAl i CEPA ee Anat 27 23 20 16 13
WAN MLSs dame 5 1, 28 24 21 17 14
OCU ees 9 5 1, 28 25 21 18
BROKE, Spr ik 13 9 5 2, 29 25 22
DED A 8 Ba 14 10 6 B30 26 23

It is difficult to decide whether to cling to the old “preferred cycle
days,” based on the data of 1924 to 1941, or to use a new set based on
consideration of the distribution of 1952 and 1953. Two reasons
incline me to use the old basis this year. First: In 1953, as stated
above, the precipitation came near giving good results on the old
basis. So it may be that conditions have returned to the old normal.
Indeed the high peak on the eleventh day of the cycle is but 0.6 as

AO uso
Ras Sass
equa

BEwaie

ya mean b

IRI et ie
Bee he A Me

Ch a Tu
ae

26 days

Fic. 1.—Observed distribution of precipitation in Washington on cycles of
27.0074 days. a, January 1, 1924, to December 31, 1941. b, January I to
December 31, 1952. c, January 1 to December 31, 1953.

Oinches precipitation

4 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

high in 1953 as in 1952, which may indicate a gradual return toward
normal. Second: Advices from E. Fraselle of Etterbeck, Brussels,
Belgium, who has used the 27.0074-day cycle in Equatorial Africa
and in Belgium, state that since 1949 there has been no failure or
change of phase in the cycle.

So I give in the accompanying table the 175 dates when higher
average precipitation in Washington may perhaps be expected in 1954
than the average precipitation of all other dates of 1954. The first
column, in Roman figures, gives the “preferred days” of the 27-day
cycle. The remaining columns give the actual dates in the 12 months
of 1954 when these preferred cycle days recur, and when higher than
average daily precipitation in Washington may be expected.

The basic tabulation, on which the table rests, began with January 1,
1924, and ended with December 1941. The “cycle” deduced from
those records is of 27.0074 days, which corresponds nearly with the
average period of the rotation of the sun.t

TEMPERATURE AT WASHINGTON
In previous papers on Washington weather, I have shown that it has

a regular period of 6.6485 days, and also of days. In previous

years I have made predictions, based on these periods, when days
would be warmer than the days immediately before and after. But
the periods are so short that, with local and temporary atmospheric
influences displacing phases of the periods frequently by one day, and
sometimes by two days, such forecasts are of doubtful interest. I
therefore discontinue them.

1See A 27-day period in Washington precipitation, Smithsonian Misc. Coll.
vol. 104, No. 3, 1944. (Publ. 3765.)

vs
ey i

i
1h
:
I
i

SMITHSONIAN MISCELLANEOUS COLLECTIONS

VOLUME 122, NUMBER 14
(Enp oF VoLuME)

fMlary Waux Talcott Fund for
Publications in Botany

A NEW GENUS AND SPECIES OF
PLANKTON DIATOM FROM
THE FLORIDA STRAITS

(WirTH Four Piates)

Be
PAUL S. CONGER

Associate Curator, Division of Cryptogams
Department of Botany, U. S. National Museum

22eee0e00?

(Pustication 4171)

ae CITY OF WASHINGTON
---—s« PUBLISHED BY THE SMITHSONIAN INSTITUTION
ai JULY 15, 1954

SMITHSONIAN MISCELLANEOUS COLLECTIONS

VOLUME 122, NUMBER 14
(Enp oF VotumeE)

Mary Waux Walcott Fund for
Publications in Botany

A NEW GENUS AND SPECIES OF
PLANKTON DIATOM FROM
THE FLORIDA STRAITS

(With Four Prares)

BY
PAUL S. CONGER

Associate Curator, Division of Cryptogams
Department of Botany, U. S. National Museum

AWE INC Re ay

(Pustication 4171)

CITY OF WASHINGTON
PUBLISHED BY THE SMITHSONIAN INSTITUTION
JULY 15, 1954

The Lord Baltimore (Press

BALTIMORE, MD., U. 8. A.

Mary Waux Walcott Fund for
Publications in Botany

A NEW GENUS AND SPECIES OF PLANKTON
DIATOM FROM*THE FEORIDA*SPRAITS

By PAUL S. CONGER 1
Associate Curator, Division of Cryptogams
Department of Botany
U. S. National Museum

(WirH Four Pirates)

The diatom here described was first found in plankton gatherings
in the summers of 1938 and 1939, while I was working at the Carnegie
Institution of Washington Marine Laboratory, Dry Tortugas, Fla.,
but its study was not completed at that time. It was taken in a No. 20
silk bolting-cloth tow net in waters adjacent to the laboratory, in sea
water of 60° to 70° F. temperature and approximately 35%o salinity.
The water was so clear that the very presence of plankton would be
doubted, and only after a haul of 20 minutes or more could an ap-
preciable quantity be secured.

To the best of my knowledge this form represents both a distinctive
new species and a new genus, which may be monotypic.

THALASSIOPHYSA RHIPIDIS Conger, gen. et sp. nov.

Frustula magna, delicatissima, in aspectu zonali hemisphaerica, di-
midio mali aurantii longitudinaliter secti similis sed sectores multo
tenuiores et numerosiores ; valvae tenues, reniformes, in plana vel fere
in plana insidentes, marginibus ventralibus vicinis, globulos uniseriatos
prope marginem dorselem ubique ferentes, eos fere deficientes in mar-
gine ventrali; volumen frustulae in regione zonali fere inclusum; ora
angusta vel tenuis.

Mature, fully grown frustules rounded-basket or cradle-shaped, re-

1] am indebted to the Carnegie Institution of Washington for sojourn at their
marine laboratory at Dry Tortugas, Fla., in the last two summers of its opera-
tion, and for opportunity and facilities for collection and study of the diatom
herein described.

SMITHSONIAN MISCELLANEOUS COLLECTIONS, VOL. 122, NO. 14

2 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

sembling in general the structure of half an orange cut end to end, a
double elbow section of stovepipe, an armadillo or chiton shell, a tropi-
cal sun helmet, or an accordion flexed so that the ends held by one’s
hands lie in the same plane, or but slightly hinged planes ; the less fully
grown individuals, resembling sectors, less than half, but always more
than a quarter, of an orange cut end to end, the intercalary bands of
the diatom simulating the sections of the fruit, these intercalary bands
being, however, much narrower (thinner) in proportion and more
numerous. Girdle-band or connective zone constituting almost the
total volume of the diatom, comprised of many intercalary bands,
these bands thin, wedge-shaped in pervalvar axis, auriculate-shaped
in apical-transapical plane. Shells thin, exceptionally delicate, auricu-
late-shaped in valvar plane, with almost no rim (mantle), bearing a
single row of very small, round beads around the edge, close to the
border, except for a clear central space about one-third to one-half
the length of the ventral edge, the dorsal edge with a deeply notched
line (simulating the notched “channel-raphe” of Epithemia in appear-
ance, but not a groove). Apical axis of the shell about one-half again
as long as the transapical axis. Shell surface virtually hyaline and
structureless, though probably ultrafinely punctate-striate with radiate
design. Over-all structure of the frustule dorsiventral, the convex side
being the dorsal aspect, the flatter (or valve) side the ventral; in the
shells (valves) the notched edge dorsal, the smooth, pore- or bead-free
edge, ventral. Pervalvar axis very extended, an arc of a circle as
viewed from the narrow-girdle-band side of the frustule, the arc at-
taining a half circle in fully grown individuals, only very rarely slightly
more; extension of the pervalvar axis achieved by interpolation of
more intercalary bands, and at the same time revolution of the valves
about the line of their adjacent ventral edges until they come to lie in
a plane or close approximation thereof, in a position presenting a right-
left-hand symmetry, daughter cells and younger stages comprising
somewhat smaller angular sectors of a half circle or half sphere, but
usually greater than a third of a circle. Frustule with a narrow, half-
round-bottomed groove, between the approximated ventral edges of
the shells and at right angles to the pervalvar axis, formed by adjunc-
tion of narrow edges of the intercalary bands, this groove shallowing
and broadening into a slightly wider zone of narrow, closely set inter-
calary bands (newly forming ones) as it rounds up the sides and over
the dorsal area.

Chromatophore single, thin, flat, pale pea green in color, broadly
rounded rectangular to broad-elliptical, about two-thirds the total

NO. I4 NEW PLANKTON DIATOM—CONGER 3

length of the frustule in the pervalvar-apical plane, suspended by
plasma threads from various points of the inner frustular surface, in
the central area of the cell, in a plane parallel to, or coincident with, a
plane segment to a central sector of the arched pervalvar-apical plane,
and cutting the transapical axis perpendicular to it and somewhat
below its center.

Nucleus approximately round to elliptical, flat, hyaline, lying adja-
cent to and dorsally, in the center of the chromatophore, about one-
third to one-half the diameter of the chromatophore in size; not visi-
ble except in cell division.

Symmetry.—Cell in pervalvar axis, dorsiventrally asymmetrical,
laterally symmetrical. Apical axis isopolar, transapical axis hetero-
polar. Shells symmetrical with respect to the transapical axis, asym-
metrical in respect to the apical axis. Plane of cell division between
the ventral edges of the shells and cutting the pervalvar axis perpen-
dicular to its center in the plane of the other two axes.

Resting spores not known.

Cell measurements.—Pervalvar length of frustule 200-300 microns ;
apical axis of shell (long diameter) 130-190 microns; transapical
axis of shell (short diameter) 80-100 microns. The greatest variation
in shape seen in proportional lengths of apical to transapical axes.

Remarks.—The hemispherical configuration of this diatom repre-
sents an extreme extension of the narrower, wedge-shaped (cuneate)
symmetry found in many groups of diatoms (e.g., Licmophora, Gom-
phonema, Cymbella, Hemidiscus).

The central, narrow region of extremely thin, closely set intercalary
bands across the dorsal surface of the frustule in apical-transapical
plane is evidently the growing zone in which new intercalary bands
are formed, although by what morphological beginnings such growth
is initiated cannot be seen.

The broad-girdle-band view of this diatom is the one that is de-
scribed as basket-shaped, the narrow-girdle-band view (end view,
looking in direction of the apical axis) is fan-shaped.

The frustules of this form are so exceptionally delicately silicified
that they are not susceptible to any customary laboratory treatment,
but instead collapse under all efforts for permanent mounting, includ-
ing drying, liquid penetration, and transfer; but they may be readily
handled for study while fresh in sea water. The frustules are also so
delicate as to be readily soluble and consequently must be handled with
great care in liquid preservation. This means not changing the osmotic
tension too rapidly, or adding any reagent that might have a tendency
to dissolve or separate the delicate intercalary bands.

4 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

Type locality —East shore of Loggerhead Key, Dry Tortugas, Fla.

Occurrence (distribution) .—This form is ever present (throughout
the summer) in rather plentiful numbers in the plankton of this area
and is one of the more common components of the phytoplankton pop-
ulation. I have no information of its occurrence through other months
of the year, having visited the area only during the summer. Also, I
have no material from other places and hence no knowledge of the
possible extent of its distribution. One might think it would be found
widely in the limy waters of the lower Florida, Gulf, and Caribbean
area, a matter which it would be desirable to ascertain. It can be got-
ten in fair numbers in a short plankton haul at any time during the
summer in the above locality, though at all times the water is so clear
as to appear practically free of plankton ; the very gossamerlike shells
of most of the plankton forms are due to the poverty of silica in these
waters. It appeared to me, on the basis of collections within a radius
of 10 miles, to be more a habitant of the shallower (6 to 15 feet), pro-
tected, coral-strand waters (certainly at least more prevalent in these),
than of the rough waters and active open-water areas. The east shore
of Loggerhead Key running off very gradually for half a mile or more
over shallow coral sand and coral-head bottoms, and protected also by
the shape and position of the island from the more vigorous sweep of
tides, current, and wave action, appeared to represent the ideal habitat
for this form. Whether or not it is a habitant of rougher waters and
wider areas needs further verification.

Relationships—This diatom is of uncertain relationships. After
considerable study and comparison, it does not fit satisfactorily in any
of several genera superficially similar, and it seems best, or necessary,
that it be put into a genus by itself. It is well, however, to call atten-
tion here to its several close simulations, in case other observers may
note affiliations which I have overlooked.

It was first thought, because of its shape and general structure, to
be an Amphora, but the rounded appearance of the valve and particu-
larly the absence of any observable raphe discouraged this connection.

In girdle view the complete frustule looks very like Hemuidiscus
Hardmanianus (Grev.) Mann (Report on the diatoms of the Alba-
tross voyages, etc., Contr. U. S. Nat. Herb., vol. 10, pt. 5, p. 316,
1907), as pictured in Schmidt’s Atlas der Diatomaceenkunde, pl. 439,
fig. 2, 1940, and in Ann. Mag. Nat. Hist., ser. 3, vol. 16, p. 2, pl. 5,
figs. 1-4, 1865 (under the name Palmeria Hardmaniana Grev.). The
valve is not greatly unlike this species in shape; it has the general
shape of members of the genus. However, the valve of Thalassiophysa

NO. 14 NEW PLANKTON DIATOM—CONGER 5

rhipidis is much more rounded at the ends, and it does not show the
radiate structure characteristic of members of the genus Hemidiscus.
The mantle is likewise different in that it is not wider on the dorsal
than on the ventral edge, in contrast to Hemidiscus, in which it is
wider on the dorsal than the ventral edge, thus giving a wedge shape
to the girdle aspect of the valve. Unlike the very slightly wedge-
shaped intercalary sections of my form, the additive result of which
is the hemispherical frustule described, the valve itself is perfectly
flat with almost no mantle, which is, as far as can be seen, of even
width all around. Hemidiscus derives much of the hemispherical
shape of its frustule from the considerable width of the wedge-shaped
mantle on the dorsal side of its shell. Members of Hemidiscus do not
exhibit the close beading around the dorsal edge of the valve as in my
form. Although these points have been noted and compared there is
really no serious question of my form belonging to this genus.

The valve shape and markings of my species more closely resemble
Auricula, and it is possible it should be placed there, but if so it is
certainly widely different in general appearance from the well-known
species of that genus. Most nearly suggestive of the valve shape of my
form is Auricula complexa Greg. as shown in Peragallo’s Diatomées
Marines de France, pl. 42, figs. 14, 15.

This new form would doubtless have been observed earlier were it
not for its gossamerlike frailty and tendency to collapse or be destroyed
in preservation, and the need, therefore, to examine it promptly in
freshly collected material. Because of this extreme frailty, and the
inability in consequence to mount or preserve it well, there is no type
specimen or type material. It must be collected and seen in fresh
material.

Morphology and development.—The vegetative reproductive or
morphological changes of this diatom are very interesting. They are
deliberate, orderly, and precise and are very delicate in visible aspect.

The nucleus, or nuclear area, is not visible in ordinary examination
of the mature cell because it is so hyaline and because it is obscured
by the broad chromatophore in close proximity to which it lies.
Judged from the morphological changes, and from its general posi-
tion in other diatoms, it is undoubtedly in the center of the cell, sup-
ported by the very delicate and almost transparent plasma threads
that radiate from this area and attach to all parts of the inner periphery
of the frustule.

It is possible that the nucleus or nuclear area might be made visible
by staining. Unfortunately I did not attempt this, when I was in

6 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122

position to obtain living material, as I have since wished I had. The
extreme delicacy of the diatom and ease of its collapse suggest that
very careful vital staining would be required, and that only staining
of very freshly collected living material would be feasible.

The nuclear area, even in its hyaline state, probably largely occupied
by the nucleus itself, looking at the cell in the direction of the apical
axis, can be seen in a certain stage of the cell development or vegeta-
tive reproductive process, namely, at that time when the nucleus and
chromatophore have divided and the new daughter chromatophores
lie in an angular position, each in its separate part of the dividing cell.
At this time the nucleus forms a small, weblike mass across the angle
of the chromatophore and is distinguished from the latter by its hyaline
character as contrasted with the green of the chromatophore.

Vegetative division, just casually mentioned above, probably takes
place, as in many diatoms, mainly at night, most of the cells found
during the day being in a more mature state. Vegetative division
occurs almost entirely only in cells that are well matured, or of a full
hemispherical shape. The first evidence of division is a readily seen
slight constriction in the middle of both sides of the broad, pale-green
chromatophore. The final resulting halves of the original chromato-
phore, assume each a right-angular shape with the adjacent faces lying
in an apical-transapical plane and extending upward toward the dor-
sal surface of the cell, the inner side of the angle thus being upward
with the nuclear area subtending its vertex. The nucleus and nuclear
mass must evidently have divided previously to, or simultaneously
with, the division of the chromatophore, to occupy this position as a
faint hyaline web across the angle of the newly formed chromatophore.
Both subsequently and slowly flatten out as the newly formed daugh-
ter cells grow and expand to maturity, attaining again the hemispheri-
cal shape. Tension and stretching of the plasma threads play a major
role, or appear to be the chief mechanism involved, in bringing these
changes about. The daughter cells finally separate as a result of ex-
panding internal pressure which forces them apart. The presence of
such internal pressure as a functional agent in the dividing cell proc-
ess is occasionally evidenced by a snapping apart of the newly formed
cells, if one is so fortunate as to be observing them just at the critical
moment or brings to bear a micro-needle upon such a cell in a state of
being just about to divide, when a slight touch will cause it to spring
apart.

Deposition of the new valves back to back in a central apical-trans-
apical plane, cutting perpendicularly the pervalvar axis, is seen to be

NO. 14 NEW PLANKTON DIATOM—CONGER 7

occurring as the divided chromatophores assume their angular posi-
tions within the new daughter halves. Formation of the new valves
becomes more and more evident as a line of increasing density in this
central section of the cell, and if the cell be turned at a slight angle,
there becomes evident in perspective two new sets of short plasma
threads connecting the edges of the new chromatophores to the edges
of the newly forming valves, as shown in the accompanying illustra-
tion (pl. 4, figs. 3, 4, 7, and 8).

Very close to the newly forming valves in this central region, and
almost imperceptible, is a condensed, seemingly striate section of very
narrow width, like a considerable number of compressed pleats or
folds of a camera bellows or accordion, which are evidently potential
or elementary intercalary bands of the prospective new cells. There
is, of course, a set of these on either side of the new valves in the cen-
tral region, one set, that is, in each of the new daughter cells.

This process is carried out with such great precision of cellular
mechanics, and this is such a beautiful and advantageous form in
which to view this sequence, that it is regrettable that the species is
not a somewhat more robust or at least a more available one. It would
seem a good one in which to study at least some phases of diatom cell
division, to the above cursory discussion of which much remains yet
to be added.

Derivation of name.—I have chosen for the generic name of this
beautiful diatom a combination of the Greek word “‘thalassios,’’ mean-
ing “of the sea,” since it is so characteristically a marine plankton
form, and the Greek word “physa,” meaning “bellows,” from the
great resemblance of its finely folded or pleated appearance to that
of an accordion, fireplace bellows, or camera bellows. Especially in
certain perspective views from both the apical end and side of the not
fully matured cell does one get a typical folded and wedge-shaped
effect of a fireplace bellows, as in plate 2, upper left figure.

As a species name for this form I have taken the Greek word
“rhipis,” meaning “fan,” from its fine resemblance to a lady’s folded
fan, in apical end view as shown in the first (upper) four figures of
plate 1. This is a very frequently observed and characteristic view.

Thus the name Thalassiophysa rhipidis seems best to indicate the
particular features of this diatom.

Importance.—This diatom, in spite of its great delicacy and watery-
like consistency, is sufficiently frequent to be an important constitu-
ent of the plankton population, and, in an area where the water is
very clear and the plankton is in general tenuous and phantomlike,

8 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. I22

it is a probably not insignificant factor in the biochemistry and economy
of the sea.

It has no really close forms in structure and appearance, and it
possesses a simple beauty which in my opinion would cause it to rate
high in this respect among the diatoms.

ATES

e
7

PLATE I

THALASSIOPHYSA RHIPIDIS

Figs. 1-4, Cell in apical aspect, various views. Fig. 5, Ventral view of ma-
ture cell (pervalvar view), showing the two halves with their ventral edges
approximate and forming the ventral groove. Fig. 6, Dorsal (pervalvar) aspect
of mature cell. Fig. 7, Single valve (in apical-transapical view). Fig. 8, Single
valve, with a few intercalary bands showing across the notch. Magnification:
All figures approximately 240 diameters.

_ SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122, NO. 14, PL. 1

(See legend on opposite page. )

VOL. 122, NO. 14, PL. 2

SMITHSONIAN MISCELLANEOUS COLLECTIONS

nee
oi

THALASSIOPHYSA RHIPIDIS
Various typical views of well-matured cells. Magnification: All figures approximately
240 diameters.

SMITHSONIAN MISCELLANEOUS COLLECTIONS

S ‘\. |
MMA XAEY :

SSS

(97 KE a4

THALASSIOPHYSA RHIPIDIS

Figs. 1-6, Various exterior views of the dividing cell, with chromatophore body dimly
outlined within. Figs. 7-8, Ventral aspects of a narrower frustule of the same species.

Magnification: All figures approximately 240 diameters.

SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 122, NO. 14, PL. 4

6 8

(See legend on-opposite page. )

PLATE 4

THALASSIOPHYSA RHIPIDIS

Fig. 1, A dividing cell in apical aspect. Fig. 2, A cell in process of division
with one side collapsed. Fig. 3, Ventral view of a cell in late stage of division;
chromatophore outlined with plasmic thread attachments. Fig. 4, Ventral per-
spective view, showing chromatophore in color, beginning central membrane of
newly forming valves, and plasmic thread attachments. Figs. 5 and 6, Apical
and pervalvar aspects of mature dividing cells respectively showing chromato-
phore bodies colored. Figs. 7 and 8, Outline sketches of the same cells as
figures 5 and 6 showing, respectively, chromatophore arrangements, nucleus,
and plasmic thread attachments. Magnification: All figures approximately 240

diameters.

a PRL ent

‘ay

eet
ft
Ne

an
eo

NMI

3 9088 01421 5081

tlh

on
uw
Cc
Be
c
o
ol
z
12
=
=)
=
Ee
”
z
4
<
z
°
”
»
=
=
7)

Oren eee
per
re IE em
ay nearer

wor
errs
vrewturnnnt

sro Pe enero
fet ware Snr ON ae RMA COE ek
i A gst Fe On TO RE

OE renter oe

srererurerenesares vases"
iceman
ocicmetereetenritarnmare te ern
o “
See engeu ime Lica er tee tr an ener
— eke
separate
ane

Internet Archive Audio

Featured

Top

Images

Featured

Top

Software

Featured

Top

Books

Featured

Top

Video

Featured

Top

Mobile Apps

Browser Extensions

Archive-It Subscription

Save Page Now

Full text of "Smithsonian miscellaneous collections"

See other formats