Skip to main content

Full text of "Smithsonian miscellaneous collections"

See other formats







VOL. 97 



(Publication 3529) 




BALTIMORE, MD., 0. 8. A. 

VvV>r. U^-^.^opy 


The Smithsonian Miscellaneous Collections series contains, since the 
suspension in 1916 of the Smithsonian Contributions to Knowledge, 
all the publications of the Institution except the Annual Report, the 
annual pamphlet describing the Institution's field-work, 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. 

C. G. Abbot, 
Secretary of the Siiiithsouiau Institution. 



1. Strong, William Duncan; Kidder. .Vlfred, II; and Paul, 

A. J. Drexel, Jr. Preliminary report on the Smithsonian 
Institution- Harvard University archeological expedition to 
northwestern Honduras, 1936. 129 pp., 16 pis., 32 figs., 
Jan. 17, 1938. (Publ. 3445.) 

2. Johnston, Earl S. Plant growth in relation to wave-length 

balance. 18 pp., 4 pis., Jan. 12, 1938. (Publ. 3446.) 

3. Resser, Charles Elmer. Middle Cambrian fossils from Pend 

Oreille Lake, Idaho. 12 pp., i pi., Jan. 3, 1938. (Publ. 


4. ScHMiTT, John B. The feeding mechanism of adult Lepidoptera. 

28 pp., 12 figs., Jan. 10, 1938. (Publ. 3448.) 

5. Stirling, M. W. Three pictographic autobiographies of Sitting 

Bull. 57 pp., 46 pis., July 22, 1938. (Publ. 3482.) 

6. Snodgrass, R. E. Evolution of the Annelida, Onychophora, and 

Arthropoda. 159 pp., 54 figs., Aug. 23, 1938. (Publ. 3483.) 

7. Wedel, Waldo R. The direct-historical approach in Pawnee 

archeology. 21 pp., 6 pis., Oct. 19, 1938. (Publ. 3484.) 

8. Bush NELL, David I., Jr. Drawings by George Gibbs in the Far 

Northwest, 1849-1851. 28 pp., 18 pis., 4 figs., Dec. 30, 1938. 
(Publ. 3485.) 

9. Deignan, H. G. a new nuthatch from Yunnan. 2 pp., Oct. 10, 

1938. (Publ. 3486.) 

10. Resser, Charles Elmer. Fourth contribution to nomenclature 

of Cambrian fossils. 43 pp., Dec. 17, 1938. (Publ. 3487.) 

11. Weintraub, Robert L. An assay method for growth-promoting 

substances utilizing straight growth of the Avena coleoptile. 
10 pp., I pi., Dec. 31, 1938. (Publ. 3488.) 

12. Resser, Charles Elmer. The Spence shale and its fauna. 

29 pp., 6 pis., Jan. 20, 1939. (Publ. 3490.) 








(With 16 Plates) 



Anthropologist, Bureau of American Ethnology 


Peabody Museum, Harvard University 



Peabody Museum, Harvard University 

(Publication 344S) 



JANUARY 17, 1938 



•, ■ bo 

a. -x: 

tn>: .S 

U I- rt 
















(With 16 Plates) 


Anthropologist, Bureau of American Ethnology 


Peabody Museum, Harvard University 



Peabody Museum, Harvard University 




JANUARY 17, 1938 




Introduction i 

Brief geographic setting 2 

Ethnic and linguistic background 8 

Early historic contacts in northwestern Honduras 19 

Archeological explorations 27 

Chamelecon River 27 

Naco 27 

Las Vegas 34 

Tres Piedras 35 

Other sites 2>7 

Ulua and Comayagua Rivers 39 

Las Flores Bolsa 39 

Santa Rita (farm 17 ) 45 

Playa de los Muertos (farm 11) 62 

Other sites 76 

North end of Lake Yojoa 76 

Aguacate and Aguatal 80 

La Ceiba 90 

Site I 91 

Site 2 94 

Site 3 99 

Causeway and " canal " near Jaral 100 

Pyramids and stone statues near Los Naranjos 102 

Excavations on the northern border of Los Naranjos 105 

The older horizon at Los Naranjos iii 

Other sites 115 

Summary and tentative conclusions 118 

Literature cited 125 

Explanation of plates 127 



1. (Frontispiece) Processional figures on a Yojoa Polychrome vase of 

Mayoid type, site 2, La Ceiba title 

2. Various Chamelecon and Ulua River sites 130 

3. Naco sherds 

4. Naco sherds and artifacts 

5. Upper Ulua Polychrome pottery types, Las Flores 

6. Upper Ulua Polychrome pottery types. Las Flores 

7. Ulua Polychrome, Bold Geometric pottery types, Santa Rita and Naranjo 


8. Ulua Polychrome, Mayoid pottery types, Santa Rita 

9. Ulua Bichrome sherds, from deepest level at Santa Rita 

ID. Playa de los Muertos Bichrome sherds 

1 1. Playa de los Muertos Bichrome sherds and figurines 

12. Yojoa Polychrome vessels, Mayoid types 

13. Yojoa Polychrome vessels. Bold Animalistic and other types 

14. Yojoa Polychrome vessels. Bold Animalistic and other types 

15. Early ceramic types at Lake Yojoa 

16. Los Naranjos, Lake Yojoa 



1. Map of northwestern Honduras 4 

2. Map of the region around Naco 7 

3. Sketch map of the ruins of Naco 30 

4. North wall of cross-section trench through mound 3, Naco 31 

5. Map of the lower Ulua and Chamelecon Rivers 40 

6. West wall of excavation i, Santa Rita (farm 17) 47 

7. Hollow figurines, whistles, and " candelario ", from the Ulua Poly- 

chrome period, Santa Rita (farm 17) 52 

8. Ulua Polychrome, Bold Geometric tripod dish, excavation 2, Santa Rita 

(farm 17) 54 

9. Unusual Ulua Polychrome, Bold Geometric dish, excavation 2, Santa 

Rita (farm 17) 55 

10. Ulua Polychrome, Bold Geometric bowl, excavation 2, Santa Rita 

(farm 17) • • • • 56 

1 1. Ulua Polychrome bowl, excavation 2, Santa Rita (farm 17) 57 

12. Ulua Polychrome bowl, excavation 2, Santa Rita (farm 17) 58 

13. Ulua Polychrome, Mayoid vase, excavation 2, Santa Rita (farm 17) ■ • • 59 

14. Inside design from Ulua Polychrome, Lower Mayoid vases (pi. 8, a, b), 

excavation i, Santa Rita (farm 17) 60 

15. Ulua Polychrome Bowl, excavation 2, Santa Rita (farm 17) 61 

16. North wall of excavation i, showing stratification, Playa de los Muertos 

(farm 11) 64 



17. Outlines of vessels of the Playa de los Muertos culture obtained by 

Dorothy H. Popenoe /O 

18. Outlines of vessels of the Playa de los Muertos culture obtained by 

Dorothy H. Popenoe /i 

19. Polychrome vase of Ulua Lower Mayoid type 77 

20. Sketch map of archeological sites around the north end of Lake Yojoa. . 78 

21. Yojoa Polychrome bowl, Bold Animalistic type, Aguacate 82 

22. Yojoa Polychrome pot. Bold Animalistic type, Aguatal 84 

23. Yojoa Polychrome bowl, Bold Animalistic type, Aguacate 85 

24. Yojoa Polychrome tripod dish. Bold Animalistic type, Aguacate 86 

25. Yojoa Polychrome bowl, Bold Animalistic type, Aguacate 86 

26. Outline of Yojoa Polychrome pot showing " vestigial " spout, Aguacate. 87 

27. Yojoa Polychrome cooking pot, Aguatal 88 

28. Yojoa Polychrome bowl, Mayoid type, Aguacate 89 

29. Yojoa Polychrome bowl, Mayoid (?) type, Aguacate. 90 

30. Yojoa Polychrome vase, Mayoid type, Aguacate 91 

31. Cross-section of west wall, site i, Los Naranjos, showing house floor, 

burial, and superimposed cultural horizons 107 

32. Cross-section of excavation A, near site i, Los Naranjos, showing 

stratified cultural horizons 113 





Anthropologist, Bureau of American Ethnology; 

Peabody Museum, Harvard University ; 


Peabody Museum, Harz<ard University 


The present paper presents in tentative and outline form certain 
major results of the Smithsonian Institution-Harvard University 
Archeological Expedition to northwestern Honduras in 1936. The ex- 
pedition personnel included the senior author as leader and represen- 
tative of the Bureau of American Ethnology, Smithsonian Institution, 
and the two other authors as representatives of the Peabody Museum 
of Harvard University. Mrs. Strong and Mrs. Kidder accompanied 
the expedition in the field and performed invaluable services in cata- 
loging and caring for the collections. The purpose of the expedition 
was twofold : to extend the work carried on by the Smithsonian Insti- 
tution in northeastern Honduras in 1933, and to follow up earlier work 
on the Ulua River so successfully inaugurated, under the auspices of 
the Peabody Museum, by George Byron Gordon and by Mrs. Dorothy 
Hughes Popenoe. The discoveries of Mrs. Popenoe at Playa de los 
Muertos in 1928 and 1929 opened new vistas in Honduras archeology, 
and her untimely death was a sad blow to science and to all who were" 
fortunate enough to know her. In a sense our work was merely a con- 
tinuation of that which she had so ably begun. The original suggestion 
for the present expedition came from Dr. Wilson Popenoe, and it was 
due to him that the successful cooperative effort was launched and 

The expedition received cordial support from the government of the 
Republic of Honduras, and our warmest thanks are extended to the 

Smithsonian Miscellaneous Collections, Vol. 97, No. 1 


officials in Tegucigalpa and elsewhere who not only furthered the 
cause of science but put us deeply in their debt for many personal 
courtesies. Similarly, the officials of the United Fruit Company, both 
in the United States and in Honduras, furnished very material aid 
in ways too numerous to mention. Without this generous assistance 
our results would have been tremendously curtailed. Considerations 
of space prevent listing the many people who aided us in our work, but 
we cannot forbear mentioning Mr. Walter S. Turnbull, and Mr. Regi- 
nald Hamer. To them and to many other persons in Honduras we are 
deeply grateful. Later, in the final report, it will be possible to 
acknowledge more adequately our appreciation of the many cour- 
tesies, both official and personal, which we received on every hand. 

Our choice of this particular area for excavation was based on 
numerous geographic, historic, ethnographic, and archeological con- 
siderations. For this reason we have devoted considerable space to 
these important factors. Viewed against this background, it is hoped 
that a condensed account of our excavations may have value. In due 
course a final report will be prepared by the senior author for publi- 
cation by the Peabody Museum. As this may not appear for some 
time, it seems advisable to make our major results available without 
undue delay. Most of this important area still awaits adequate exca- 
vation, and it is our hope that these notes and sketch maps may be of 
value to future scientific explorers and excavators. 


The general area covered by this report includes the eastern por- 
tion of the Department of Cortes, the western border of the Depart- 
ment of Yoro, and certain places on the eastern border of the De- 
partment of Santa Barbara, all within the Republic of Honduras." 

From the archeological standpoint, however, modern political 
boundaries are of minor importance compared to factors of terrain 
and drainage, which conditioned aboriginal human occupation no less 
than they do that of the present inhabitants. Of primary importance 

'See map, fig. i. The best general maps of Honduras at present are the 
" Carta General de la Republica de Honduras, America Central, of the Pan- 
American Institute of Geography and History, 1933 ", and the " Mapa General 
de la Republica de Honduras, Levantado por el Prof. Jesus Aguilar Paz, 1933." 
These maps are far superior to any of their predecessors. Many contradictions 
still exist, however, owing to the present inadequacy of cartographic exploration 
in much of Honduras. The present map (fig. i) is primarily based on that of 
Dr. Paz. 


is the fact that at this point occurs one of the easiest passages across 
the Central American Isthmus from Tehauntepec to Panama." 

From the mouth of the Ulua River, where it enters the Gulf of 
Honduras, a series of elevated valleys extend up the Rio Blanco to 
Lake Yojoa, over the plateau of Siguatepeque, across the Plains of 
Comayagua, and down the valley of the Goascaran River into the 
Gulf of Fonseca and the Pacific. It can hardly be coincidental that it is 
at this point that the higher aboriginal cultures of the Pacific High- 
land extend north to the Caribbean Sea, in marked contrast to the lower 
cultures of the remainder of the Atlantic Lowland region in Honduras. 
The present archeological reconnaissance covers the northern half of 
this natural transition area between the Pacific Highland and the 
Atlantic Lowland regions. 

If we include the valley of the Chamelecon River, which at no very 
distant time emptied into the Gulf of Honduras through the Ulua 
River, this entire area from Lake Yojoa north may be termed the 
Ulua drainage. The lower portions of the Ulua and Chamelecon Riv- 
ers flow through the Plain of Sula, a rich alluvial valley, down to their 
respective mouths in the great mangrove swamps extending along the 
Gulf of Honduras from Puerto Cortez almost to Tela. Because of 
these swamps and their shallow, silted-up channels, neither river 
offers much inducement to modern navigation, whereas such impedi- 
ments were probably of small import to the numerous trading canoes 
of pre-Conquest times. Above the mangrove swamps, which extend 
some 20 kilometers upstream, is the rich valley floor that today is cov- 
ered with banana plantations. Formerly the valley supported a rich 
tropical flora, described by Gordon (1898) and others, but at present, 
except for isolated remnants in swamps and low areas, the great 
mahogany, ceiba, and other trees, have been replaced by the ubiqui- 
tous banana. To the northwest the Ulua valley is hemmed in by the 
great Mountains of Omoa, which reach a height of 7,000 to 8,000 feet. 
To the east occur the Mountains of Mico Quemado and Tiburon. Be- 
tween these two ranges the Ulua-Chamelecon valley reaches a breadth 
of some 45 kilometers, terminating about 75 kilometers in a direct 
southwest line from the mouth of the Ulua at Potrerillos, where the 

^ This has been pointed out time and again in the voluminous literature referring 
to the much-talked-of but never completed transoceanic railroad across Honduras. 

See Squier, 1858 and 1870, and Wells, 1857. Although perhaps unduly opti- 
mistic on some points, Squier's various reports remain the best general geographic 
descriptions of Honduras. 

Wells gives a detailed and delightful picture of Honduran life in the middle 
of the last century. So far as the remote interior is concerned, much of his 
description holds good today. 


Fig. I. — Map of northwestern Honduras (based on the General Map of the 
Republic of Honduras, by Prof. Jesus Aguilar Paz, 1933). 


bordering mountains converge and the Ulua splits up into its three 
main branches (maps, figs, i, 5). These are, from north to south, the 
Comayagua, the Rio Blanco, and the Ulua proper. The Chamelecon, 
after running parallel to the Ulua for some 50 kilometers above its 
mouth, turns north into mountains, where, as a rapid mountain stream, 
it extends almost as far southwest as Copan. Similarly, the three 
branches which form the Ulua are rapid, clear streams, in marked con- 
trast to the slow moving, muddy lower Ulua and Chamelecon. 

Owing to its configuration and the mountainous character of its 
terrain, Honduras has a wide variety of climates and seasons. In 
general, however, the dry season in the region we are considering be- 
gins in December or January and extends until June or early July. 
The temperature, which is pleasantly low in the early part of the dry 
season, increases as the wet season approaches. The rainy season is 
cooler, but more unpleasant, owing to rain, wind, and great humidity. 
Moisture brought by the northeast trade winds is deposited when they 
hit the high mountains bordering the Ulua valley. Thus, despite their 
relatively short courses, the Ulua-Chamelecon tributaries at times 
carry a tremendous volume of water. These rivers rise to their great- 
est heights about October and flood the valleys. A smaller rise occurs 
in the late Spring. As the rivers spread over the lower valleys, they 
deposit the sediment brought down from the mountains, and in this 
way the valley has been aggraded or built up. As proved by human 
occupation levels buried in situ 6 meters or more deep along the pres- 
ent channels, this lDuilding-up process has been relatively rapid and 

As the present rivers shift their channels across the valley floor, they 
thus expose in their steep banks the various human occupation strata of 
past centuries, which elsewhere in the valley are inaccessible, because 
of depth and lack of surface indication. The majority of the sites 
investigated along the Ulua by the present expedition were of this 

Whereas the lower Ulua valley was formerly covered with a luxuri- 
ant rain forest, the sites which we worked on the upper Chamelecon 

^ Yde, 1936, p. 39, in our opinion, exaggerates the difficulties due to depth of 
deposit facing the archeologist in this region. Nevertheless, there is no doubt 
that the earliest human remains in the lower Ulua valley may be buried at 
inaccessible depths. However, as the present report demonstrates, it is possible 
to obtain stratification. Gordon, 1898, shows the manner in which the rivers 
cut and shift their channels as well as the dangers of re-deposition which must 
be borne in mind by the archeologist. Neither Mrs. Popenoe nor the present 
writers encountered cultural remains at the extreme depths mentioned by Gordon 
and Yde. 


in the vicinity of Naco (map, fig. 2) are located on clear, rapid streams. 
These streams are bordered by narrow strips of tropical forest, but 
back from these are steep hills or elevated rocky plains covered with 
pines and oaks. There is no reason to believe that the environment here 
was different in aboriginal times. It is a region admirably adapted to 
maize. Abundant food supplies possessed by the numerous Indian 
pueblos, as well as the occurrence of gold in the surrounding moun- 
tains, early attracted the Spaniards to these mountain valleys. The 
climate appears to be much more healthful than that of the lower 
river valleys. 

The northern end of Lake Yojoa, where other excavations were 
carried on, offers a similar environment. This marks our most south- 
erly working point as well as the limit of the Ulua drainage in this 
direction, since Lake Yojoa in part drains through an underground 
channel into the Rio Blanco. To the south, it is said to drain into the 
Santa Barbara River by means of the Jaitique River and by sub- 
terranean channels, principally the Rio Sacapa and the Rio Salala.^ 
We did not investigate the southern end of the lake. 

Lake Yojoa is located in a small mountain valley or bolson at an 
altitude of some 2,050 feet. The auto road from the north coast to 
Tegucigalpa utilizes the lake as a water connection by means of auto- 
mobile ferries. To reach Jaral, the little town on the north shore, one 
leaves the low banana country at Potrerillos and climbs through rocky 
hills covered with oak, pine, and scrub, following the Rio Blanco River, 
which is crossed only once at the little town of that name. Just before 
the road reaches the lake, the grade increases sharply and then drops 
down onto the small triangular plain bordering the lake (map, fig. 20). 
This bush-covered plain is bounded on the east by low pine-covered 
hills, and on the north by high (5,000-6,000 feet) volcanic mountains. 
We suspect that this plain originally supported a heavy rain forest, 
but both the aboriginal and modern inhabitants have long practised 
milpa-type farming here, and today there exist few remnants of the 
original forest. With the exception of open, rolling, pine-covered hills 
on the northeast shore near Agua Azul, the remainder of the lake 
is bordered by steep slopes covered with rain forest. At the southern 
extremity of the lake, there is a considerable belt of low, swampy land, 
most of which is overflowed when the lake is full. Beyond the water- 

^ Published reports and maps of Lake Yojoa are utterly inadequate. Squier, 
1858, pp. 96-104, and i860, pp. 58-63, is still the authority. The lake has been 
studied from time to time by Honduran and American engineers, but we know 
of no up-to-date maps or reports. Amory Edwards, and Squier, i860, describe 
10 outlets. 

NO. I 



shed to the south in the vicinity of Taulebe and San Jose, there are 
fertile plains and valleys. These are separated from the Comayagua 
valley by the high plateau of Siguatepeque. 

In general, therefore, it can be said that the sites investigated by 
us in the Ulua drainage occupy two main environmental regions : First, 
those on the lower Comayagua and Ulua Rivers, located in the rain 
forests of the broad, alluvial river valleys, and second, mountain val- 
leys, as at Naco and at Lake Yojoa, where the elevation was consider- 
ably greater, the rain forest limited to the borders of stream or lake, 
and the more open pine and oak association close at hand. 


At the present time there are no obvious, aboriginal remnants of 
population in the part of Honduras considered in this report. It is 
true that the present population of the region is in considerable part 
made up of assimilated Indian groups, but the language is Spanish 
and the culture Latin American. Isolated groups of Jicaque Indians 
are reported as still living in the more remote parts of the Depart- 
ment of Yoro.'* 

It is possible that Lenca-speaking Indians may still be found in 
our region, and groups of Chorti Maya occur in the departments to 
the south and west, but, as yet, ethnographic and linguistic research 
in Honduras has received little attention. If we desire to connect the 
archeological remains with historic Indian groups, it is therefore neces- 
sary to turn back the pages of history and consider the region at the 
time of the Spanish Conquest. 

Early sources on northwestern Honduras are fairly numerous, in- 
cluding Cortez, Bernal Diaz, Palacio, Las Casas, Torquemada, Mar- 
roquin, Montejo, Palaez, Pedraza, Espino, and the historians Her- 
rera, Oviedo, and Gomara, but, with the exception of the first three, the 
grains of ethnography to be gleaned from the works of these writers 
seem rather scant. In a later paper other sources will be considered, 
but for the present we shall confine ourselves to the most important 
primary sources and more recent general studies. 

As was the case in regard to the geography of Honduras, one must 
still consult E. G. Squier's " States of Central America ", 1858, re- 
garding the ethnography of Honduras. Similarly, H. H. Bancroft's 

° Described by Habel, 1880, p. 17. In June 1933 the senior author was told by 
Mr. Acley, then American Consul at Tegucigalpa, of a very primitive group of 
Jicaque Indians he had visited that year near the town of Morale, in the Sierra 
de la Flor of the Department of Yoro, near the Olancho line. From photographs, 
they appeared very similar to those described by Habel. 


"Native Races" (1882) is a treasure house of ethnographic source 
material, and the same author's " History of Central America " ( 1883) 
not only indicates the sources but also the major trends of native and 
European contact in the period of conquest. More recently Cyrus 
Thomas and John R. Swanton (1911) have presented in brief form 
the salient facts regarding the distribution of Indian languages in 
Honduras as part of their study of the languages of Mexico and Cen- 
tral America. Preeminent in this field, however, is the voluminous 
and detailed work of Walter Lehmann, who in a preliminary report 
in 1910, and in " Zentral Amerika " (1920), his given us a wealth 
of data based on close examination of the sources as well as personal 
linguistic work in the lield. In addition to an intensive study of origi- 
nal sources and present-day Indian dialects, Lehmann has also in- 
cluded many sweeping theoretical generalizations. The latter, however, 
concern us less in the present study than do the specific references to 
linguistic and cultural distributions in Honduras at the time of the 

In general, the linguistic maps of Thomas and Swanton (1910), 
and Lehmann (1920) agree as regards the distribution of native 
languages in the Ulua drainage. Both place Choi, Chorti, and other 
Maya-speaking groups to the west of the Ulua River proper. Follow- 
ing Lehmann, we find that the Lenca occupied a large area around 
Lake Yojoa, extending north almost to the junction of the Comayagua 
and the Ulua. From here to the coast the valley of the Ulua and 
Chamelecon Rivers was Jicaque territory. To the west, Lenca and 
Jicaque territory bordered on that of the Chorti and Choi Maya, 
Copan being one center of Chorti occupation. The Lenca and Jicaque 
demesnes extended east to that of the Paya who with the Sumu, oc- 
cupied the northeastern corner of Honduras." 

To the south, peoples of Lenca speech extended to the Pacific coast. 
To the west they were bordered by the Pipil of Salvador along the 
Lempa River, and to the east by various Chiapanecan and Matagalpan 
groups (see linguistic maps, Thomas and Swanton, 191 1, and Leh- 
mann, 1920). 

Thus it appears that all of the territory investigated by the present 
expedition was occupied by Jicaque and Lenca-speaking peoples at the 
time of the Conquest — with one very important exception. This was 
the occurrence in the same region of various Nahuatl-speaking pueblos 
along what appear to have been trade routes extending from south- 
ern Mexico and from the Pipil (Nahuatl) territory in Salvador. Leh- 

* Linguistic and tribal distributions in Honduras have already been discussed 
in some detail; see Strong, 1935, pp. 7-19 and 140-172. 


mann indicates one such line of late Nahuatl influence and settlement 
which crossed the Chamelecon River in the vicinity of Naco and ex- 
tended east to the Nahuatl pueblos mentioned by Cortez, located a 
short distance south of Trujillo (Lehmann, 1920, vol. 2, p. 629 and 
map). Both Cortez and Bernal Diaz in their accounts of the traverse 
from Mexico to Honduras indicate the importance and vogue of 
these trade routes and mention the many pueblos engaged in trade 
which they visited (Bernal Diaz, 1916). Recent Nahuatl settlements 
would be thus expectable in the upper Chamelecon valley near Naco 
and probably elsewhere in the Ulua valley proper. These Nahuatl 
influences from southern Mexico were apparently quite recent, but 
the Pipil occupation of Salvador was much older. This is clearly 
indicated by Palacio (i860, pp. 21, 31, and 65), who points out the 
acquisition of the Pipil tongue by many neighboring groups originally 
of different linguistic affiliation. Moreover, Pipil cultural influences 
were obviously very active in southwestern Honduras at the time of 
which Palacio writes (i. e., 1576). 

The name of the Ulua River was apparently derived from the Ulba 
language, which Palacio ascribes to Honduras (i860, p. 21). Both 
Squier (Palacio, i860, p. 114) and Lehmann (1910, p. 747; 1920, 
vol. 2, p. 624) concur in this identification. The extension of the term 
Ulba, Ulua, or Ulvan to the Sumu is explained by Lehmann on the 
grounds that the Sumu, Jicaque and Matagalpan languages (includ- 
ing the Cacaopera and Lislique) are all basically related. This seems 
quite probable but has not yet been satisfactorily demonstrated. Since 
the Jicaque lived along the Ulua river, it is most probable that Palacio 
referred to them as the Ulba. Both Squier (Palacio, i860, p. 114) 
and Lehmann (1920, p. 624) regard Palacio's designation " Chontal " 
as a general term for non-Pipil-speaking peoples. According to Leh- 
mann, the term would include the Lenca with the Jicaque. Specifically, 
Palacio seems to refer to the Lenca when he speaks of the Taulepa. 
This is the old name for Lake Yojoa (Taulebe, according to Squier, 
i860) . In the Lenca language, Taulepa means " House of the Jaguars " 
(Lehmann, 1910, p. 747: 1920, vol. 2, p. 624). The jaguar was of 
special importance in Lenca mythology. Lehmann is convinced that 
the region around Lake Yojoa and the entire central portion of 
Honduras was occupied by the Lenca (the Taulepa of Palacio), and 
that the valley of the lower Ulua and the adjacent Department of 
Yoro was primarily occupied by Jicaque groups (the Ulba of Palacio). 
During the seventeenth century, the names Lenca and Jicaque were 
often confused, but, as indicated by the work of Thomas and Swanton, 
the general regions assigned to these groups by Lehmann seem 


It is therefore apparent that our archeological investigations were 
made in a contact area between advanced Mayan and Nahuatl peoples 
to the west, and Lenca, Jicaque, and other less advanced groups to the 
east. As Lehmann points out (1920, vol. 2, p. 625, and map), Maya 
influence, as indicated primarily by archeological objects, extended 
well into Lenca territory, including all the region west of a line drawn 
from the junction of the Ulua and Comayagua Rivers southeast to the 
Gulf of Fonseca. Similar influences were also present in the lower 
Ulua valley and in Salvador. Moreover, Palacio (i860) clearly 
indicates that cultural influences from the Nahuatl Pipil of Salvador 
extended east well into Lenca territory during early historic times. 
Whether Lehmann's assumption that the language of the Lenca is re- 
lated, on the one hand, with the Jicaque and the Paya, and on the 
other, with the Cacaopera, Matagalpa, Sumu, Ulua, Miskito, Rama- 
Guatuso, etc. (1920, vol. 2, p. 637), be accepted or not, there is little 
doubt that the majority of these languages are affiliated with major 
linguistic stocks to the south. There is at least a strong probability 
that the majority of these languages are in some degree related to the 
Chibchan linguistic stock centering in northern South America. As 
has been pointed out elsewhere (Strong, 1935, pp. 170-172), the scant 
ethnological information on certain of these groups likewise points to 
a southern derivation. On the other hand, Choi and Chorti Maya and 
Nahautl (Pipil and Aztecan) linguistic connections are clearly with 
the north. Thus the LTlua-Yojoa region comprised an important 
bufifer area between two sets of cultural traditions and linguistic 
stocks, the one derived from Mexico and northern Central America, 
the other from southern Central America and, eventually, from 
South America. A very complex archeological situation is therefore 
expectable. It is, however, a situation that, when clearly understood, 
is certain to throw much new light on the ultimate derivation and 
development of the higher civilizations of the New World. 

Since the historic occupants of our particular region were the 
Lenca and Jicaque Indians, we are particularly concerned with what- 
ever ethnological information has survived concerning their cultural 
status at the time of the Conquest. Regarding the Jicaque, little is 
on record but for the Lenca, or at least their near neighbors and cul- 
tural kin, we have the brief but excellent account of Palacio. Speak- 
ing of the plain of Jiboa in the province of San Miguel, Salvador, 
he says that here the Indians begin to speak a new language, called the 
Chontal. He states that they " are a very rude people, but had 
anciently a great reputation for valor among their neighbors." His 
description of the customs observed prior to 1576 at Micla, (Mita), 


apparently a cultural center representing a blending of Pipil and 
Lenca ceremonies and customs, is so important that we quote it in 
full, following Squier's translation (Palacio, i860, pp. 65-89). 

Three leagues distant, is the village of Micla, which anciently the Pipil Indians 
of this district held in great veneration ; it was here they came to make their 
offerings and sacrifices, as did also the Chontals, and other neighboring Indians 
of different languages. Their modes of sacrificing differed in some respects from 
those of other parts. They had cues or temples, and teupas or priests of high 
authority, of which there are still many signs and traces. 

Besides their cazique or secular lord, they had a kind of pope, called Tecti, 
who dressed in a long blue robe, and wore on his head a diadem, or sometimes a 
mitre embroidered with many colors, at the crown of which rose a cluster of 
very beautiful feathers, taken from a bird, called in this country. Quetzal. 
This pontiff carried in his hand a staff, which resembled the crook of a bishop, 
and he was obeyed in all spiritual things. After him, next in sacerdotal authority, 
was the tehu a matlini, who was the ablest diviner and the man best versed in 
their ancient books and in their arts. He it was who made auguries and foretold 
future events. After him were four priests called teupixquis, who went dressed 
in long robes, falling to their feet, each of different color, black, red, green and 
yellow. These were the counsellers of the pontiff, and directed all the super- 
stitous ceremonies and follies of their religion. Their was also a kind of mayor- 
domo, who had charge of the sacred jewels and the instruments of sacrifice. He 
also opened the breasts of the victims of sacrifice, and tore out their hearts, 
and performed such other personal services as were requisite. Besides all these 
there were other functionaries, who played on the drums, trumpets and other 
instruments used in convoking the people to the sacrifices. 


When the chief priest died, he was buried in his own house, seated in a 
painted chair, and all the people mourned for him for fifteen days, with loud 
cries and lamentations. They also fasted during this period; but when this was 
over, the cazique and the wife, man or diviner selected a new pontiff by lot. 
It was requisite that he should be one of the four priests above mentioned. When 
the choice was made, they had great feasts and dances, and he who was chosen 
drew blood from his tongue and private parts, and offered it in sacrifice to the 
idols. He also named his successor in the priesthood, who was required to be 
a son of the deceased pontiff, if he had left one, if not, the son of some other 
priest. He filled also the other offices which at any time became vacant in the 
teupas, or temples. They adored the rising sun, and had two idols, one repre- 
senting a man, whom they called Quetzalcoatl, and the other a woman named 
Itzqueye. All their sacrifices were made to them, and they had a calendar, 
with days specially set apart for each one, and on these the sacrifices were made. 


Each year they had two principal and very solemn sacrifices ; one at the com- 
mencement of summer, and the other at the beginning of winter. These were 
made in the interior of the sacred place or temple, and were of boys between 
the ages of six and twelve years, bastards, born among themselves. 



They sounded their trumpets and drums for one day and night before the 
sacrifice, and when the people were assembled, the four priests came out from 
the temple, with four small braziers in which they burnt copal and caoutchouc ; 
and the four together, turning in the direction of the rising sun, bent their knees 
to it, offering incense, and reciting words of invocation. After this they separated, 
and did the same in the direction of the four cardinal points, south, east, north and 
west, preaching and explaining their rites and ceremonies. When the sermon 
was finished, they retired within four houses or chapels which were built at the 
four corners of the temple, and there rested for a little while. They next went 
to the house of the high priest, which was close to the temple, and took thence 
the boy who was to be sacrificed, and conducted him four times around the court 
of the temple, dancing and singing. When the ceremony was finished, the high 
priest came out of his house, with the second priest and mayordomo, and 
ascended the steps of the temple, accompanied by the cazique and principal 
Indians, who, however, stopped at the door of the sanctuary. The four priests 
next seized the victim by his extremities, and the mayordomo coming out, with 
little bells on his wrists and ankles, opened the left breast of the boy, tore out 
his heart and handed it to the high priest, who put it into a little embroidered 
purse, which he closed. The priests received the blood of the victim in four 
jicaras, which are vessels made from the shell of a certain kind of fruit (the 
calabash), and, descending one after another into the court, sprinkled it, with 
their right hands, in the direction of the cardinal points. If any blood remained 
over, they returned it to the high priest, who put it back, with the purse 
containing the heart, into the body of the victim, which was interred in the 
temple itself. This was the kind of sacrifice made at the opening of the two 
seasons of the year. 

The high priest, his second, and the four priests were accustomed to meet to 
ascertain, by sorcery and enchantment, if they should make war, or if their foes 
were coming to attack them; and if it appeared that such an event was to take 
place, they called together the cazique and war chief, and advised them of the 
approach of their enemies, and whether they should go to meet them. The 
cazique then assembled the soldiers, and all went out to battle. If he was 
victorious, he despatched a messenger to the high priest, advising him of the date 
of the occurrence, and on this information the diviner ascertained to which of the 
gods sacrifice was to be made. If to Quetzalcoatl, the ceremonies lasted fifteen 
days ; if to Itzqueye, five days, and on each day they sacrificed a prisoner. These 
sacrifices were made as follows : All those who had taken part in the war, 
returned home in order, singing and dancing, and bringing with them those who 
were to be sacrificed, decorated with feathers and chalchiuites on their wrists 
and ankles, and with strings of cacao beans around their necks, the captains 
themselves conducting them in their midst. The pontiff and priests, at the head 
of the people, went out to meet the victors, with music and dancing; and when 
they encountered them, the captains delivered over the victims to be sacrificed, 
to the high priest; after which all went together to the court of their teupa, 
where they kept up the dancing night and day, for the periods above named. In 
the middle of this court was placed a block or bench of stone, on which the victim 
was stretched, the four priests holding him by the feet and hands. The sacrificer 
then came forward, loaded with plumes and bells, with a knife of flint, with 


which he opened the breast of the victim, and took out his heart, and tossed it 
in the air in the direction of the four cardinal points, and finally threw it aloft 
directly in the middle of the court, in this way soliciting the divinity to accept 
the sacrifice, in return for the victory. This sacrifice was public to all the 
Indians, great and small. 

During this period, the soldiers returning from the war, could not cohabit with 
their wives, but were obliged to sleep in certain calpules or barracks, which 
were given up to them for the occasion, by the young men who were learning 
the art of war. During the day they went to the houses of their women to eat 
and drink, and from thence to their plantations, always however, leaving a 
company to guard the town. The men sacrificed blood drawn from their private 
parts, and he who had most wounds in these was reputed to be most valiant. The 
women sacrificed blood drawn from their tongues and ears, and they sacrificed 
their entire bodies, taking up the blood with cotton and offering it to their idols— 
the men to Quetzalcoatl, and the women to Itzqueye. 

Their superstitious ceremonies, at the time of planting their fields, were as 
follows : They put in little cups of calabash the seeds which they had selected 
for the purpose, and placed them before the altar of their idols. They next 
dug a trench in the ground, in which they planted the seeds regularly, covering 
them with earth ; and over all they placed a large brazier, full of burning coals, 
on which they sprinkled copal and caoutchouc. The four priests then drew blood 
from their cars and nose, receiving it in certain large reeds, which they burnt 
before their idols. At other times they drew blood from their tongues and 
private members, and petitioned their gods to prosper the fruits of the earth, 
and give them abundant harvests. The high priest, in sacrificing, drew blood 
from the same parts, and with it anointed the feet and hands of the idols, invoking 
the demon, who spoke with him, and told him what kind of weather would follow, 
all of which was communicated to the people by the four priests, who always 
concluded by ordering the men to have connection with their wives, and then 
proceed to plant their fields. And such was the sacrifice of planting. 

We come now to their sacrifices for hunting and fishing. They took a living 
deer to the courtyard of the cue or temple which they had outside of the town, 
where they strangled and skinned him, collecting all his blood in a vase, and 
cutting in small pieces the liver, lungs and stomach. These were put aside, with 
the heart, head and feet. They next cut up and cooked the deer by itself, and the 
blood by itself, and while these were cooking they had their dances. Next the 
high priest and his assistant took the head by the ears, and each of the four priests 
one of the feet, and the mayordomo put the heart in a brazier and burned it, 
with copal and caoutchouc, as incense to the idol of the god who was held to be 
protector of hunting and fishing. When the dance was finished, the head and 
feet were scorched in the fire before the idol, as an offering, and afterwards taken 
to the house of the high-priest and eaten. The flesh and blood were then eaten 
before the idol ; and the same was done with all the animals which they offered 
in sacrifice. When they sacrificed fish, the entrails were burnt before the idol. 

When a woman was in travail, the midwives made her confess her sins ; but 
if this was not sufficient to hasten the birth, they made her husband do the same ; 
and finally, if the woman admitted illicit connection with any other man, they 
went to his house and took his clothes and placed them beneath her ; if this 
failed, as a last resort, the husband sacrificed blood from his tongue and ears. 
When the child was born, if a boy, they put in his hands a bow and arrows; 


if a girl, a spindle of cotton ; and the mother made a streak of soot mixed with 
water on the right foot of the child, which they believed would prevent it, when 
grown up, from being lost in the woods. At the end of twelve days, the child 
was taken to the priest, green branches being scattered under the feet of the 
bearers. The priest gave it the name of its grandfather or grandmother, as the 
case might be, and they presented it with cacao and fowls, which were the 
offerings made to the priest. When it was taken back to the house, the mother 
carried it to a river and bathed it, offering to the stream, cacao and copal, that 
it should never do evil to the child. 

As regards the rites for the dead ; if the defunct were a cazique or captain, 
or the wife or child of either, all the people mourned for four days and nights. 
At the rising of the sun on the fifth day, the high priest announced that the soul 
of the dead was with the gods, and that it was useless to mourn any longer. 
They buried the dead man dressed in all of his riches, in a sitting posture, and 
in his own house. Their manner of mourning during the four days and nights 
resembled a mitote, in which they chanted the lineage and deeds of the dead. 
If he were a cazique who died, the high priest and all the people, immediately 
recognized as his successor his son or daughter; or, if he had neither, his 
brother or nearest relative. 

On such occasions they had great feasts, dances and sacrifices, and the new 
chief entertained in his house all the priests and captains. If a common man died, 
only his children and relatives mourned; and if a woman lost her child, she 
reserved her milk for four days, without giving it to another ; for they believed, 
if she failed in this, the dead child would do the living one some injury. This 
sacrifice they called navitia. 

It was the office of the cacique to order the plantings, and direct the marriages. 
They always married when young; and when the affair was arranged, and the 
affianced groom met his future father-in-law he turned aside, as also did the 
affianced bride when she met her future mother-in-law. They did this, because 
the devil had made them believe that such encounters would prevent their having 
children. Marriages were celebrated in this wise : the male relatives of the 
woman sought the bridegroom and made him bathe in a river; and the female 
relatives of the woman did the same with the bride ; they then wrapped each of 
them in a new, white cloth, and took them to the house of the bride, where they 
tied them up naked, in their garments. The relatives of the young man then 
made presents to the bride of cloths, cottons, fowls and cacao, while those of 
the bride gave presents of the same kind to the bridegroom ; after which they 
all feasted together. At these marriages the high priest and cazique were obliged 
to be present. 

Concerning relationship : They have a tree painted, with seven branches, which 
represent the seven degrees of relationship in direct descent, within which no 
person might marry, excepting those who had distinguished themselves in war, 
but even these might not marry within three degrees of blood. In respect of the 
line collateral, they made use of another tree with four branches, which repre- 
sented the four degrees within which no one could marry. 

Aside from other laws which these Indians possessed in common throughout 
the province, those of this nation have the following as inviolable : 

Whoever contemned or ridiculed the sacrifices to the idols, or the ceremonies 
connected therewith, was condemned to death. 

Whoever had connection with a strange woman, was condemned to death. 


Those who had carnal intercourse with relatives, within the degrees above 
proscribed, both sufifered death. 

He who spoke libidinously with a married woman, or who made improper 
signs to her, was banished and his property confiscated. 

Whoever had commerce with a strange slave (one not his own ?) was himself 
reduced to slavery, unless pardoned by the high priest for services in war. 

Whoever wounded another, if the wound were serious, suffered death therefor. 

Whoever violated a virgin was sacrificed. 

Whoever lied was severely whipped ; and if it were in any matter concerning 
war, he was enslaved. 

Those of the people who were not soldiers cultivated the plantations of the 
cazique, pontiff and priests ; and also gave a part of their own crops for the 
support of the warriors. 

This is what I have been able to gather concerning the manners and custom of 
this people. 

Near this place, is a high rocky hill from which flow two streams of water, 
close to each other, one hot and the other cold. Here too is found an abundance 
of spices, which the Indians use in their drink and food ; and an earth which 
resembles copperas, and which it must be judging from its effects. With this 
they make a dye. 

From here to the borders of the province of Chiquimula de la Sierra, the 
country is for the most part high, of good temperature, abounding in pasturage, 
and adapted for the support of cattle, and the cultivation of all kinds of grains. 

In the portion of this province which lies in the direction of Gracios a Dios in 
Honduras, are the Chontal Indians. While there, complaint was made to me 
against a cazique of a place called Gotera, who since the time of his paganism 
had had his private member split open, as was the custom anciently, among the 
most valiant. In 1563, certain idolatrous Indians of another village called Cezori, 
got together in a neighboring forest where one of them performed the same 
operation; and afterwards they circumcised four boys of twelve years of age, 
in the Jewish manner, offering the blood to an idol of stone of a cylindrical form, 
with a double visage and many eyes, called Icelaca. They say that he is the 
god which knows the present and the past, and sees all things. Both his faces 
were anointed with blood, and they sacrificed to him deer, fowls, rabbits, peppers, 
and other things which they used in ancient times. 

Torqtiemada (1723, lib. 3, cap. 41, vol. i, p. 330) has recorded 
a Lenca myth which, he says, was told him by the old people. Ac- 
cording to them, 200 years before this time, there came to Cerquin 
(Lehmann states, 1920, vol. 2, p. 636, that this was probably Corquin 
in the Department of Gracias, Honduras), a lady, white as a Castilian, 
whose name was Comicahual, meaning " jaguar that flies ", so named 
because she was very wise and versed in supernatural arts. These 
Indians held the jaguar in high esteem. She made her abode in 
Calcoquin, which was the most fertile land in the province. Here 
there were stone " lions " which they worshiped, and a large three- 
pointed stone which had on each point three grotesque faces. Some 
said that Comicahual carried it there through the air and by its virtue 


won battles, thus extending her realm. Some said that she had three 
supernaturally conceived sons, others said they were her brothers. 
When she grew old, she distributed her territories among them with 
advice concerning the good treatment of her subjects. She then com- 
manded that her bed be taken out of the house. Lightning flashed 
and thunder roared. The people saw a beautiful bird flying across 
the sky and, as they never saw the lady again, they believed she was 
the bird and thus went to the sky. The sons (or brothers) divided 
the realm and governed it well. The people were courageous and 
warlike. They had been taught religion and enchantments by the 
Lady Comicahual. Among the many idols which they adored, there 
was one called the Great Father and another called the Great Mother. 
To these idols they prayed for their well being. Other gods were 
introduced, to whom they prayed for food, property, riches, and 
that their lands might prosper, and produce abundantly. And, " for 
many years these superstitions and deceits of Satan lasted among the 
old people." Lehmann (1920, vol. 2, p. 637) is inclined to identify 
the Calcoquin of Torquemada with the Icelaca of Palacio. He goes 
on to show that similar rites, presumably originating with the Pipil as 
indicated by Palacio, extended as far north as the Bay Islands in the 
Caribbean. This evidence, derived from Salcedo, has already been 
cited elsewhere (Strong, 1935, pp. 14, 15) and need not be repeated 
here. Sufficient for our present purposes is the fact that elaborate 
but basically similar cult observations extended from Salvador north 
beyond the mouth of the Ulua River and that many of these at the 
time of the Conquest seem to have originated in Pipil territories. 
Only the results of scientific archeology can show whether this his- 
toric Salvadorean center was actually primary or was derived from 
still earlier sources of cultural development. This will be discussed 
in relation to the results of our own archeological excavations. 

Linguistically, the Lenca and the Jicaque have since been studied 
by various travelers. This material has been summed up and amplified 
by Lehmann (1920, vol. 2, pp. 649-722). From the ethnographic 
standpoint, recent work on the Lenca and Jicaque has been pitifully 
inadequate. Habel (1880) describes various Jicaque he met in the 
Department of Yoro as follows: 

The Xicaques differ in the form of their bodies from all the other tribes of 
Central America. Their stature, on the average, being equal to that of Europeans, 
is greater than that of the other tribes. Their skin is of a lighter color, and 
their features resemble more closely those of the Caucasians, having a more 
pleasant and intelligent expression than any other tribe of this region known to 
me. Both sexes wear a kind of apron made of the inner bark of the Caoutchouc 
tree. That of the women reaches around the waist and the ends hang down 


from the hips to the knees. These two flaps are attached to the body by a strap 
of the same material fastened around the waist. By another narrower strap, tied 
around the head, they secure the long black hair, parted in front, floating down 
to the shoulders. 

According to Habel, the Jicaque had but recently been gathered 
into permanent settlements through the splendid efforts of a Spanish 
missionary. He adds that they were improvident, did not cultivate 
the soil nor raise any large domesticated animals. They had formerly 
been permitted to sell themselves into practical slavery, but this prac- 
tice had then been stopped. They traded in sarsaparilla and tobacco. 
Habel goes on to discuss the physical and other characteristics of the 
still numerous Paya, who appeared to him to be much darker in pig- 
mentation than the Jicaque. We have previously indicated that quite 
primitive groups of Jicaque survive at the present time. 

According to Otis T. Mason (1889), the Lenca of Honduras had 
an ingenious method of straightening lance shafts. A pole about 
16 feet long was suspended vertically from the limb of a tree by a 
lariat attached by half hitches to both ends of the pole. At the lower 
end, the lariat was attached to a rock weighing around 50 pounds, the 
shaft being thus held straight while seasoning. He goes on to de- 
scribe a variant of the musical bow used by the Lenca which was 
called a " bumbum." This strung bow had a small gourd on the back 
of the bend which was attached to the bow cord by another string 
running at right angles. The bow was rested on a half gourd inverted 
on the ground, which gave added resonance while playing (Mason, 
1889). Apparently this instrument was not confined to the Lenca, 
for Habel (1880, p. 31) describes an identical instrument used at 
about the same time by the Pipil of the Balsam Coast of Salvador. 
Here it was strung with wire and called the " carimba." The melody 
was produced by strumming the wires with a stick and cupping the 
hand over the gourd. Quite possibly this represents a variant of the 
musical bow, or it may be a historic borrowing from the African 
marimba so popular in Central America at present. Whether it is 
primarily of New World or of African origin, we cannot say. 

In June 1936 the junior and senior authors of the present report 
were grounded by an airplane accident at the town of San Pedro 
Sula in Honduras. While waiting for a track car, we were enter- 
tained by a small native boy who, with a short stick, strummed dole- 
fully on the identical instrument described by Mason. In this case 
the bow string was of wire and the bow rested on an empty carton 
instead of a gourd. 


Squier (1859, pp. 603-619) gives a brief but vivid picture of a 
fiesta at Comayagua in which Indians from the nearby mountains 
performed dances accompanied by much ceremonial drinking and 
native ritual. The deer and the ocelot were the symbols of the two 
main dancing groups. Their musical instruments consisted of flutes, 
the Panspipe, the marimba, and a covered pot with a string drawn 
through the bottom. At this fiesta, the Indians, the majority of whom 
were probably Lenca, visited the numerous ruined towns in the 
vicinity of Comayagua that had been occupied at the time of the 
Conquest. He also described an extremely isolated village of the 
Guajiquero Lenca and gives an amusing account of the difficulties 
involved in securing linguistic or ethnographic information from the 
Indians. As anyone knows who has attempted work with Honduras 
Indians, the repression of almost half a millennium combined with 
linguistic barriers is not an easy thing to overcome. However, it is 
obvious from Squier's account that a wealth of native custom and 
belief still survives among the more isolated groups. 

Such survivals, combined with the extremely haphazard nature 
of previous research among the living Indians, indicate that there 
is much more information available in Honduras for the trained 
ethnologist and linguist than has been generally realized. 


The fourth voyage of Columbus gives us our first historic glimpse 
of conditions on the Honduras mainland. Having visited the Bay 
Islands, Columbus landed at Punta de Caxinas (the Cape of Hon- 
duras) on August 14, 1502. The chroniclers of this voyage give a 
brief but vivid picture of the advanced agricultural life and the 
thriving coastal trade then existing on the north coast of Honduras.' 

In 1524 Gil Gonzalez named Puerto Caballos (later to become 
Puerto Cortez) and established a settlement at San Gil de Buena 
Vista. From this base he sailed down the coast and marched over- 
land into the Olancho valley, where he met and defeated a force under 
Hernando de Soto that had been exploring this region from Nica- 
ragua. Returning to Puerto Caballos he was informed of the arrival 
of a Spanish fleet under Cristobal de Olid. 

It is of interest that again Honduras becomes a buffer area and 
battleground between two earlier established southern and northern 

' Pertinent historical and ethnographic information regarding the Bay Islands 
and the adjacent mainland have been given elsewhere (Strong, 1935, pp. 7-19). 
The following historical resume of the Ulua region is primarily condensed from 
Bancroft, History of Central America, vols, i and 2, 1883. Other sources are 
cited as they occur. 


centers, the one in Panama under Pedrarias, the other in Mexico 
under Hernando Cortez. Not content with the rich spoil of the Aztec 
Empire, Cortez had already cast covetous eyes to the south where 
rumor painted the golden glories of Hibueras or Honduras. For 
this reason he despatched a trusted lieutenant with a fleet to conquer 
the province. Having already reached an agreement with Velasquez, 
Governor of Cuba and the rival of Cortez, Olid, in 1524, established 
the settlement of Triunfo de la Cruz east of Puerto Caballos and 
withdrew his allegiance from Cortez. The latter countered by dis- 
patching another fleet under Las Casas, which proceeded from 
Mexico to the Bay of Honduras. Olid promptly attacked Las Casas. 
As Bancroft says, " it was an original spectacle in these parts, Span- 
iards fighting Spaniards, in regular naval engagement ; and as the 
hissing projectiles flew out from the smoke over the still waters, 
followed now and then by a crash, the noise reverberating over the 
forest-clad hills, the dusky spectators should have been exceedingly 
grateful for this free exhibition of the wisdom and power of Euro- 
pean civilization that had come so far to instruct them in such a 

Although the honors of battle, if any, went to Las Casas, a tropical 
storm wrecked his fleet and he was forced to surrender. Along with 
Gil Gonzales, who had also been captured by Olid, Las Casas was 
taken inland to Olid's headquarters newly established in the large 
Indian town of Naco (see maps, figs, i, 2). Here, although they 
were treated as guests by their captor, the two captives plotted against 
Olid, and eventually cut his throat with a table knife. Crawling away 
into hiding, Olid sent for a priest. The latter being followed. Olid 
was dragged into the plaza at Naco and publicly beheaded. Las Casas 
returned to Mexico by an overland route through Guatemala. Even 
today, over 400 years later, a tradition still persists among the un- 
lettered inhabitants of present-day Naco that here " the king was 
killed " after being dragged in from his hiding place at El Salto, the 
falls of the beautiful little Naco river. Here, as elsewhere in the 
New World, European civilization was ushered in with blood and 

Meanwhile Cortez had had no word from his latest Honduras 
venture. Despite the advice of his other lieutenants, he decided to 
leave Mexico and proceed overland to Honduras.* 

^ The best sources on this amazing expedition are given in Maudslay's trans- 
lation of Bernal Diaz, " The True History of the Conquest of New Spain, vol. 5 ", 
1916. The pertinent letters of Cortez to the Emperor Charles V are also included 
in this volume. 


It is of interest that before starting, Cortez obtained maps from 
the Indians of the Vera Cruz region showing the entire area between 
that point and Panama. It is apparent that he was travehng along 
well-known aboriginal trade routes throughout most of his journey, 
and he mentions that nearly all the towns he stopped in were full 
of traders. The details of his Yucatan traverse do not particularly 
concern us here until he arrived at Nito near the Gulf of Dulce. Here 
he found the diseased, starving remnants of Gonzales' colony. Mak- 
ing an expedition up the Gulf of Dulce, Cortez captured a well- 
provisioned pueblo and obtained supplies for continuing his journey. 
From Nito, Cortez proceeded by sea to the vicinity of Puerto Caballos, 
where he established a settlement. He sent Sandoval overland to 
Naco. After crossing the Motagua River and visiting several pueblos, 
Sandoval's force arrived at Naco. The town had been recently de- 
serted by its native inhabitants but contained abundant provisions and 
even salt ; and here the Spaniards settled themselves, in the words of 
Bernal Diaz, " as though we were going to stay there forever." In a 
later section on the excavations at Naco, we will give more details 
regarding native conditions in the vicinity of Naco at the time of 
Sandoval's visit. 

Regarding the probable linguistic affiliations of the natives of Naco 
and the adjacent pueblos, it is significant that Lehmann lists three 
pueblos mentioned by Bernal Diaz " in the neighborhood of Naco " 
as having Nahuatl names." 

Similarly, Cortez states : 

When I first arrived at this pueblo (San Andres), I heard from the Spaniards 
who had come from Naco that the natives of that pueblo and of the neighboring 
pueblos, were somewhat disturbed, and had left their houses for the hills and 
forests, and that although some of them had been reasoned with they refused to 
be pacified from fear of the treatment that they had received at the hands of 
the followers of Gil Gonzalez and Cristobal de Olid. I wrote the Captain in 
charge there and told him to do all that he could to capture some of the natives 
by whatever means he could devise, and to send them to me so that I could 
speak to them and reassure them. This he did, and he sent me certain natives 
whom he had captured during an expedition which he had undertaken, and I 
talked to them and gained their confidence, and let them talk with some of the 
native Chiefs from Mexico, whom I had brought with me. These Chiefs told 
them who I was, what I had done in their country, and what good treatment 
they had received from me when once we were friends, and how they were 
protected and governed in justice — they and their property, their wives and 
children — and the punishment that those received who rebelled against the 
service of your Majesty, and many other things v^rhich they told them. After this, 

" Lehmann, 1920, vol. 2, p. 1018. Also see Nahuatl distributions on linguistic 


they regained confidence, although they still told me that they had some fear 
that they were not being told the truth, for those captains who had come in 
advance of me had told them the same things and more to the same effect, and 
that they had lied to them and had carried off their women when they had sent 
them to make bread, and that the men who accompanied them had been forced 
to carry loads, and they believed that I would do the same. Nevertheless, with 
the assurances which the Mexicans and the Interpreter (Marina, a Mexican 
woman) whom I had with me gave them, and seeing those of my company 
happy and well treated, they were somewhat reassured. I sent them off to speak 
to the Chiefs and people of the pueblos, and in a few days the Captain at Naco 
wrote me that some of the neighboring pueblos had become peaceful, particularly 
the chief pueblos which are : Naco, where the Spaniards are residing, Quimiztlan, 
Sula and Tholoma (Cheloma) — the smallest of these had more than two thousand 
houses — and other villages which were subject to them; and that the envoys 
said the whole country would soon be at peace, for they had sent messengers to 
pacify the people, telling them of my arrival among them and all that I had said 
to them, and also what they had heard from the natives of Mexico; they added 
that they greatly desired that I would go to Naco, as my arrival there would give 
confidence to the people. This I would have done with good will, had it not 
been very necessary for me to continue my journey in order to arrange that 
which I shall explain to your Majesty in the following chapter.'" 

From the foregoing it seems quite possible that the people of Naco 
Spoke a Nahuatl dialect understandable to the Aztec caciques and 
to Dofia Marina, Cortez' famous Mexican Indian woman interpreter. 
Had the temporary captives from Naco been Jicaque, Lenca or Maya, 
this would not have been possible. It is also possible that certain 
Nahuatl dialects served as a lingua franca in the area, due to the 
obviously extensive trade connections then in existence with Mexico 
and to the extent of Pipil influence exerted from Salvador. How- 
ever, elsewhere Cortez mentions linguistic difficulties when entirely 
alien languages were encountered by his men, but this does not seem 
to have been the case here. 

Cortez then proceeded by sea to the newly founded town of Tru- 
jillo. His settlement near Puerto Caballos was soon abandoned, owing 
to sickness and lack of food, in favor of Naco. A number of large 
and rich pueblos in this vicinity were gradually conciliated by San- 
doval, but the inhabitants of Naco, owing to the severe treatment 
they had received from Olid, refused to return to their homes. While 
Sandoval was at Naco, the caciques of two pueblos named Quespan 
and Talchinalchapa came to him to report the depredations of some 
other Spaniards who had arrived from the South." These were seized 
and proved to be a party under Garro from Nicaragua that had been 

" Bernal Diaz, 1916, vol. 5, p. 407, from the fifth letter of Hernando Cortez to 
the Emperor Charles V ; also see p. 60. 
"Bernal Diaz, 1916, vol. 5, p. 66. 


sent to claim the lands to the north for Pedrarias. They were well 
equipped with arms and horses and had handsome Nicaraguan Indian 
women with them. Sandoval sent them under guard to Cortez at 
Trujillo. Bernal Diaz, who was with this overland party, describes 
the difficulties and the Indian fights they encountered. Unless one 
has actually traveled through these mountainous, tropical countries, 
it is impossible to appreciate how truly amazing such early Spanish 
journeys were. Even today an overland trip from Naco into Nicara- 
gua would be an expedition not to be undertaken lightly. Yet in the 
time of Cortez, Spanish adventurers seem to have already traversed 
the Central American cordilleras from end to end. Cortez returned 
Garro to Nicaragua with messages of good will and mining supplies. 
For some time Cortez toyed with the idea of adding Olancho and 
Nicaragua to his conquests and even went so far as to start a road 
from Trujillo to Nicaragua! However, a mission from Mexico 
arrived with bad news, and the road is still unbuilt. 

Hearing that his holdings in Mexico had been seized by enemies, 
Cortez determined to return at once. Before departing he ordered 
Luis Marin with a number of discontented colonists from Trujillo 
to proceed to Naco where there was abundant good land. Saavedra, 
who was then campaigning in Olancho, was to remain as Governor 
of Honduras. After a hard trip Marin arrived at Naco, and the next 
day, in company with Sandoval, set out on the overland trip through 
hostile Guatemala to Mexico. Meanwhile Cortez, who, strange to say, 
appears to have been a very bad and timorous sailor, had been driven 
back by storms. Messengers were sent to Sandoval ordering him 
to stop and settle. This was a great blow to the overland party, for 
they desired above all else to return to Mexico. Sandoval hurried to 
Trujillo to plead with Cortez, that he set sail and let the overland 
party proceed. Under Marin the latter went " to some pueblos 
called Maniani and thence to another pueblo named Acalteca, where 
at that time there were many houses." ^ 

Despite Sandoval's plea, Cortez still refused to sail. Sandoval was 
dispatched to Olancho where he drove out Rojas, a lieutenant of 
Pedrarias. On Sandoval's return, Cortez sent orders to Marin to 
proceed, and he ordered Godoy, who was forming a settlement at 
Puerto Caballos, to go to Naco with all his people. Finally, in 1526, 
Cortez set out for Cuba and, eventually, Mexico. 

Finally, Pedro de Alvarado, having received orders from Cortez to 
proceed from Guatemala to Honduras, began his march. Marin, 

^' Bernal Diaz, 1916, vol. 5, p. 86. 


desperate and without orders, sent a party of lo men through Olancho 
to go to Trujillo. According to Bernal Diaz, they got as far as the 
gold-working region on the Guayape River, when they learned of 
Cortez' departure. Receiving orders from Saavedra to return, they 
did so, and, Bernal Diaz remembers, they threw stones at the country 
as they left. They met Marin at the pueblo of Acalteca and then 
proceeded to another pueblo called Maniani, where they encountered 
six of Alvarado's soldiers. In two days' marching they reached 
Alvarado " near the town called Chuluteca Malcala." This was 
probably the site of Tegucigalpa on the Choluteca River. From here 
the combined parties proceeded toward Guatemala after a difficult 
crossing of the Lempa River, which was then in flood. 

Years later, Bernal Diaz (1916, vol. 5, pp. 328, 329) thus recalled 
the country of Naco and of the Ulua River, as it was when he first 
saw it and as it soon became : 

and what I state I know, for when I came with Cortes on the expedition to 
Honduras I was present in Trujillo, which was called by the Indian name of 
Guaimura, and I was at Naco and the Rio de Pichin, and that of Balama, and 
that of Ulua, and in nearly all of the pueblos of that neighborhood, and it was 
thickly peopled and at peace (and the people were living) in their houses with 
their wives and children ; but as soon as those bad governors came they destroyed 
them to such an extent, that in the year fifteen hundred and fifty one, when I 
passed through there on my return from Castile, two Caciques who had known 
me in the old days, told me with tears in their eyes of all their misfortunes and 
the treatment (they had received), and I was shocked to see the country in such 
a condition. 

The details of this tragic and complex period in Honduran history 
cannot be considered here. The withdrawal of Cortez threw the new 
colony into turmoil and the starving colonists engaged in every form 
of intrigue. Coming from Guatemala, Pedro de Alvarado took over 
the governorship and set about pacifying the country. He built the 
town of San Juan at Puerto Caballos and founded San Pedro. For 
the Indians this was an even more tragic period. According to Ban- 
croft (vol. 7, pp. 233-234) Indian slaves were kidnapped and sold in 
Honduras by the shipload. In the vicinity of Trujillo where there had 
been villages of from 600 to 3,000 houses, there were not more than 
180 Indians left in 1547. Those not enslaved or killed had fled to the 
mountains. At Naco, where there had originally been a population 
of about 10,000, there were, in 1536, only 45 remaining. At La Haga, 
a coastal town some 9 leagues from Trujillo, there had been about 
900 houses, but of the entire population, only the daughter of the 
Cacique remained. The cruelty toward the natives was even greater 
than in Guatemala. In 1539, when Alvarado returned from Spain 


and transported the materials for building a fleet across the isthmus, 
the entire remaining Indian population fled. These evils were pre- 
sented in full detail by Bartolome de Las Casas, and the new laws 
resulting from his famous publication at least gave nominal protection 
to the oppressed natives. 

In answer to a petition from Trujillo, the Emperor appointed 
Francisco de Montejo, the former governor of Yucatan, as ruler 
of Honduras. Only a handful of starving Spanish colonists remained. 
Montejo subdued but did not enslave the Indians of the mountains 
near Trujillo. Many Indians returned voluntarily to their homes in 
this region. Montejo then visited the town of Gracias a Dios. Here, 
owing to the murder of a Spaniard, he arrested and punished the 
Lenca Indian ring leaders in the presence of the Caciques of the 
district of Cerquin previously referred to. This aroused the opposi- 
tion of the famous Lenca leader Lempira, " Lord of the Mountains." 
Lempira had previously withstood Alvarado and driven off Spanish 
attacks under Chavez, and he now opposed Montejo, The great 
Indian leader had secured allies from various interior tribes including 
several that had formerly been hostile to the Lenca, and was estimated 
to have a force of some 30,000 warriors. 

According to Lehmann (1920, vol. 2, p. 637), he had united the 
men of more than 200 towns and commanded over 2,000 " men and 
gentlemen of distinction." 

" Lempira, the last of the chiefs of Corquin, made his final stand 
against the Spaniards on the mountains of Piriera, which overlooks 
the valley of the river Lempa, in the name of which beautiful stream 
his own is commemorated." (Squier, 1858, p. 329.) 

Here for 6 months he was besieged by Caceres, a lieutenant of 
Montejo, but so greatly were the Spaniards harassed by the Indians 
that they were on the point of failure. Siege and assault having 
failed, Caceres resorted to treachery. Under a flag of truce Lempira 
came to the walls of his stronghold to parley with his enemies and 
was shot by a hidden marksman. The Lenca and their allies fled, 
and the great conspiracy soon fell to pieces. 

Mrs. Popenoe, quoting from a letter from Montejo to the King 
of Spain, June i, 1539, gives the following account of the latter part 
of this campaign against the Lenca : " 

Disturbing news reached Gracias, where Montejo was sojourning with 11 
Spanish soldiers. The Indians were preparing stubbornly to resist him. In 
Yamala, a nearby village, they were building many houses on a great, very 

" D. H. Popenoe, 1936, pp. 559-560. For the original, see Coleccion de Docu- 
mentos Ineditos, 1864, vol. 2, pp. 212-266. 


strong rock which they have, and providing them with provisions. The Spanish 
chieftain sent a Negro spy, who knew the language of the Indians, to enter the 
stronghold and bring back a report. The frightened Negro found there four 
houses built very large, and four more larger ones full of corn, and he set fire 
to the houses and to the corn. Word came of a great disaster in the valley of 
Comayagua. The Indians had risen. One Spaniard had been killed and several 
others wounded. Four horses had been lost. Unable longer to withstand the 
siege, the Spaniards had fled at night to a neighboring province where the 
inhabitants were friendly. 

Montejo realized that the time had come for desperate action. Supplies were 
brought together, and soldiers were called in from regions where the danger of 
rebellion was not imminent. Others who had been wounded but now had 
recovered sufficiently to join the colors, augmented the small band which was 
placed under the leadership of Alonzo de Caceres, recently returned from the final 
campaign against Lempira. 

When they arrived at Comayagua they found that the Indians, doubtless 
apprised of their approach, with all available supplies would fortify themselves on 
big rocks. Cattle which they could not take with them had been killed and 
eaten, so that the valley was now in a state of starvation. 

Montejo advanced into one part of the valley, Caceres into another, attacking 
and capturing a mountain fortress " which was the strongest in that region." 
The last named leader then proceeded to a village, by name Guaxerequi, where 
six Christians had recently been killed. There he found another fortress. At 
this point he was rejoined by Montejo, who describes the place in his letter. 
He says: "and (has) seen (or visited) a great rock, which was the strongest 
thing that has been seen, which, if they had time to cut a ridge of mountain, 
which they were cutting, would be impossible to capture, for they had on it 
water and wood and cultivated fields and many provisions ; they had 220 large 
houses, and certain temples and places of worship." 

It took the combined forces of Montejo and Caceres four months to conquer 
the valley of Comayagua, after which they carried the campaign into Olancho. 

Such stories as the above throw much light on the importance of fortified 
mountain tops at the time of the Conquest. Although it has been impossible to 
place Tenanipua (the famous archeological site near Comayagua, first described 
by Squier, 1858 and 1869, see map, fig. i), among the strongholds described in 
the early accounts at my disposal, it seems probable that it may have been one 
of those captured during the campaign carried out in the Comayagua region by 
Francisco de Montejo and his lieutenant, Alonzo de Caceres. It may have been 
the formidable Guaxerequi described in Montejo's letter. 

In the light of the partially cut " cuchillo " or narrow^ neck con- 
necting Tenampua with the main promontory to the northeast (D. H. 
Popenoe, 1936, pp. 562, 563 and map), I am inclined to believe that 
this identification of Guaxerequi and Tenampua is indeed very 

It is certain that a complete combing of the sources, combined with 
first-hand examination of the available archives in Honduras and 
neighboring countries, would yield a considerable mass of informa- 
tion on the Lenca and their neighbors, but this is not possible at 


present. All that has been attempted here is to suggest the main trends 
of a fascinating historical period and to indicate the probable dis- 
tribution of ethnic groups in the region under investigation. We turn 
now to outlining the results of direct archeological research. 


Chamelecon River 

Our reconnaissance of the middle Chamelecon River extended from 
May 26 until June 17, 1936. It was aimed primarily at Naco but 
several other sites were also investigated. Through the courtesy 
of the United Fruit Company we lodged comfortably at Manacal 
Ranch (map, fig. 2) which is located about a mile south of the town 
of Cofradia. Here we obtained horses and mules and were thus able 
to work at a number of archeological sites in the general vicinity. 
We first visited the San Luis site just above the confluence of the 
Naco and Chamelecon Rivers (map, fig. 2). Next we spent 2 weeks 
mapping and digging exploratory trenches at Naco. The remainder 
of the time was occupied in making stratigraphic sections and maps 
at the prehistoric Las Vegas and Tres Piedras mound sites. 


All Honduras records of the Conquest refer to Naco, first as a 
thriving Indian town and later as the site of repeated Spanish settle- 
ments. The Indian pueblo of Naco was only one of a considerable 
group in the vicinity. Montejo states that the original population of 
Naco was 10,000 persons (Colleccion de Documentos Ineditos, 1864, 
vol. 2, p. 228), an estimate that agrees reasonably well with the previ- 
ously cited statement of Cortez that the smallest of the pueblos in that 
vicinity had more than 2,000 houses. Las Casas, speaking of Hon- 
duras, says : " Tenia Pueblos innumerables, y una vega de treinta 
leguas y mas, toda muy poblada ... la ciudad de Naco que tenia 
sobre dos cientas mil animas, y muchos edificios de piedra, en especial 
los templos en que adoraban " (cited by Bancroft, Native Races, vol. 
4' P- 77^- When compared to the other authorities, as well as to the 
size of the ruins, this would seem to be an extremely exaggerated 
estimate. Similarly, his statement (Las Casas, 1822, p. 45) that be- 
tween the years 1524 to 1535 more than 2,000,000 Indians perished in 
the kingdom of Naco and Honduras, leaving only 2,000 inhabitants in 
a territory 100 leagues square, must be taken with a large grain of 
salt. Diaz, Montejo, and others give ample proof that the natives of 
Honduras were cruelly despoiled and that whole districts were de- 


populated in the early days of the Conquest. Nevertheless, the 
wholesale statistics of Las Casas seem to be those of a crusader 
rather than a historian. 

Bemal Diaz (1916, vol. 5, pp. 56-59) gives a first-hand picture 
of Naco as it was in 1525. 

At the hour of Mass we went to Naco. At that time it was a good pueblo, 
but we found it had been deserted that very day, and we took up our quarters 
in some very large courts where they had beheaded Cristobal de Olid. The 
pueblo was well provisioned with maize and beans and Chili peppers, and we 
also found a little salt which was the thing we needed most, and there we settled 
ourselves with our baggage as though we were going to stay there forever. In 
this pueblo is the best water we have found in New Spain, and a tree which 
in the noonday heat, be the sun ever so fierce, appears to refresh the heart 
with its shade, and there falls from it a sort of very fine dew which comforts 
the head. At that time this pueblo was thickly peopled and in a good situation, 
and there was fruit of the Zapotes, both of the red and small kind, and it was 
in the neighborhood of other pueblos. 

. . . When we arrived at the Pueblo of Naco and had collected maize, beans 
and peppers, we captured three chieftains in the maizefields and Sandoval coaxed 
them and gave them beads from Castile, and begged them to go and summon 
the other caciques and we would do them no harm whatever. They set ofif as 
they were ordered to do, and two caciques came in, but Sandoval could not 
induce them to repeople the pueblo, only to bring a little food from time to time ; 
they did us neither good nor harm, nor we to them, and thus we continued for 
the first days. . . . When Sandoval saw that the neighboring Indians and natives 
of Naco did not want to come and settle in the pueblo, although he sent to 
summon them many times, and that the people of the neighboring pueblos did 
not come or take any notice of us, he decided to go himself and manage to make 
them come. We went at once to some pueblos called Girimonga and Aqula, and 
to three other pueblos near Naco, and all of them came to give fealty to His 
Majesty. Then we went to Quimistan [Quimistlan in preceding chapter, Quimi- 
stan on map] and to other pueblos of the Sierra, and they too came in, so that 
all the Indians of that district submitted, and as nothing was demanded of them 
beyond what they were inclined to give, their submission did not weigh on them, 
and in this manner all was pacified as far up as to where Cortes founded the 
town which is now called Puerto de Caballos. 

Modern Naco is a small village of perhaps a dozen mud-walled 
and thatched houses on the beautiful little Naco River. Permission 
to excavate was kindly granted us by the son of the owner, Dr. Paz 
Barraona, and by Don Santiago Nolasco, the head man of the village. 
Don Santiago and the other residents of Naco were interested specta- 
tors or laborers during our work here and the children brought us 
many fragmentary specimens from the adjacent river banks. The 
heart of the site is still covered by the small but very dense shade 
trees mentioned by Bernal Diaz. These shaded our work but made 
mapping difficult. Noontime siestas spent under great jungle trees 


bordering the rapid, sparkling, Naco River made us appreciate the 
remark of the soldier-historian that " here is the best water we have 
found in New Spain." 

It is not the purpose of the present report to discuss fully the 
excavations at each site nor to analyze the archeological findings in 
any detail. Instead, a brief summary of significant excavations will 
be given, and at least one stratigraphic or horizontal artifact record 
at each site will be outlined in an effort to indicate the apparent 
trend of local cultural development. This preliminary analysis will 
be confirmed or amended in the final report in accord with the full 
statistical findings and in relation to all the excavations. Although no 
numerical record of artifact or ceramic types is given at this time, 
an effort has been made to discuss them quantitatively rather than 
selectively. In regard to ceramics, which greatly preponderate over 
any other artifact types throughout the entire Ulua drainage, we 
have here attempted to suggest the relative proportions of all wares 
at each site or in each stratigraphic section discussed. When the very 
extensive sherd collections of the expedition have been analyzed and 
the data fully presented it will be possible to check this preliminary 
analysis against the complete record. In regard to technical names 
applied to various soil layers these have been used in a very general 
sense. When our soil samples have been fully studied by experts it 
may be possible to supplement the cultural record with the detailed 
findings of soil analyst and sedimentation expert. 

As previously indicated, the ruins around Naco are extensive, and 
our detailed survey deals with only the central area. The map (fig. 3) 
gives the essential data in regard to mound orientation and elevations. 
In general, the Naco mounds are low and rounded, apparently form- 
ing the foundations of houses, but the group just northeast of the ball 
court differs in this regard. Mound 6 appears to have been the center 
of the complex. It is still the highest and was, in all probability, 
originally faced with squared stones, forming a square-faced pyramid 
with a flat top (fig. 3). A few of the cut stones are still in place. 
It has been sadly damaged by the disruptive effects of tree growth. 
According to local authorities it has also suffered by an earthquake, 
by having its stone facing removed for road foundations, and, about 
1902, by treasure-hunting excavations. It is still quite impressive, 
however. Mound 6 is flanked by mounds 3, 4, 5, all of which are 
exceptionally large. To judge by mound 3, which we cross-sectioned 
near its southern end (fig. 4), mounds 3 and possibly 5 were origin- 
ally capped by thick white plaster. This had eroded off the steep 
sides of mound 5 but was present at the base and over the flattened 




u / / 


-^ \ ^■■■,^^',■,•■'/•••---• 

Wi / 

; .- ■ 'x'.iTrr^^"--^ \ V *''- 5— .^-^-^"^ 


7 K 

> __-— — 


^; j .'-., ,v. """.'.'■" '. /'•■■' ■ 

/ \ 

§ \ i 

/ A\ 

'hb/-' '•• •• 

/ ..• / 

sS A.. ^ 

/ '■'■'•■".•'.■ / 

^^ 0) / 

t. / 

• / 

<" / 

-'1, ■.. " / 

- ''/• "^ / 


^ 'f-^ / 


1 it % / 



i-''"N, -"■ / 



I,., ■■•''■,;,,„, ,5. X ^ 


r ""''■•■,;:• "\v. : '■) ^^^ =^ ? 


^ '^1%^"^"--^ ^-— "^^^ ''^^ 


i ra 

■3 £ 

■'■i''.,\'''^-"-* '''V- ^ . ■ * « 

': " .- 

' ' : 

N ° 



.£ 0. 
£ in 




1 ^':J ^ 

1 ^, 

1 ,«. 





b o3 







/•■" p-^- "^^^>^--S'T'% .,„.. 

,Mt'/'S 5^ ^'VA*^^ ' :■ T^ ''•-'■. 


■;'<v^'" ''','> w'VVA^'' '• ^ -v 


X .■ii;i^v^v^^ii'''''^'-^^^^'''''*'^^ |J,I 

\ \ -'f ''/,'','! !''**' v)^" ^'i^'. ="-5r«-E ''■-■'■ 




„ W ''■'''- o'-' * '^•"' 

i0^ c" ^%^*w-'' ' ' 


^ ^^?' t! 



^i \^l^^ U? \/^% 

/^ " 

'-i-, ; '^ ^°6 

r ' 

;--"■"«£ 'i' 

-I Si? / 


NO. I 



/ 80 







i 0-2 
1 °S 

f- D c 
-p .S 0) 

I I I 


h -J 



top (fig. 4), Mound 3 included an inner structure the nature of 
which could not be satisfactorily determined by our one cross trench. 
This occurrence of two thin plaster walls running through the heart 
of the mound is shown in the illustrations (pi. i, 2 and text fig. 4). 
A small trench in mound 9 (fig. 3) revealed aboriginal refuse and dis- 
articulated human remains. Owing to its proximity to certain historic 
burials, work here was discontinued. 

Horizontal stripping of mound i (fig. 3) revealed considerable 
portions of the floors of two houses with massed small boulders on 
the north side and tumbled adobe blocks on the south side (Strong, 
1936, fig. 68). The plastered floors were stained a rich, dark red. 
Fragments of plaster apparently from the walls showed five succes- 
sive layers of red, yellow, red, blue gray, and red indicating the 
varying washes used in decorating the interiors of the houses. These 
colors were very fresh when uncovered but have since faded slightly. 
To judge from our test cuts, these long, low mounds north of the 
central pyramid complex consist of rows of house floors. Owing to 
the curve of mounds 2 and 14 (fig. 3) they enclose a crescentic area 
which may have been the old plaza of Naco. Our excavations, 
although very incomplete, indicate that with adequate time a whole 
series of house floors could be easily cleared. The earth covering 
them is shallow, and the floors are intact. Such work would be of 
the greatest value in revealing actual living conditions in aboriginal 
Naco. When the potsherds from inside mound i were being washed, 
we encountered two pieces of European glazed crockery. One of these 
(pi. 4, in) was obviously an early Spanish piece, the other might 
possibly have been intrusive from more recent times. Since it was in 
these houses that Olid, Bernal Diaz, and other Conquistadores lived, 
further excavations here might cast light on early historic as well as 
late prehistoric events. Certainly this association of early European 
and late Indian ceramics links the prehistoric and the early historic 
periods in Honduras. We also cross-sectioned mound 19, which is 
located about 30 meters east of mound 17 but beyond the edge of the 
map. This mound was about i meter high and 15 meters in diameter. 
It contained sherds and snail shells to a depth of 35 centimeters but 
no structural features of any sort. Here, as elsewhere in the vicinity 
of Naco, the underlying soil is hard and gravelly, making excavations 
below or beyond the artificially built or accumulated earth structures 
extremely difficult. 

There remains to be mentioned the ball court. For present pur- 
poses the general diagram (fig. 3) and the photographs (pi. 2, fig. i, 
and Strong, 1936, fig. 69) show the main features. Excavations here 


were confined to the southeastern end, owing to the presence in the 
other end of modern burials placed here under the impression that 
the structure was a colonial church. The discovery of a portion of 
one of the ball court rings in the center of the north wall was of 
especial interest (pi. 2, fig. i). A complete ring of very similar type, 
said to have been found at Naco by Sr. Roque Hernandez of San 
Pedro Sula and presented by him to Mrs. Dorothy H. Popenoe, is 
now at Lancetilla. It is possible that this specimen came from one of 
the neighboring pueblos, since none of the present Naco inhabitants 
remember its discovery and removal. As with the other structures at 
Naco, the ball court will be described in more detail at a later time. 

We searched in vain for any large refuse heaps along the Naco 
River. Scattered sherds occur where the deeply worn trail leads^ 
down the steep gravelly bank to the river at the village but we found 
no thick deposits. The children brought us various fragmentary arti- 
facts from along these banks but could show us no concentrated 
deposits. We saw little at Naco indicating any great antiquity, but 
our impressions were based on only limited study. Naco appears as a 
one-culture site, and we obtained no indication of stratigraphic 

Artifacts do not appear to be very abundant at Naco, although 
considerable broken pottery occurs in the various mounds ^nd scat- 
tered along the river bank. The bulk of the ceramic remains here as 
elsewhere in the Ulua region are from monochrome cooking vessels 
which, so far as present knowledge goes, are rarely distinctive. From 
the samples preserved, this ware appears to be primarily dull red in 
color ranging from smoke-stained black to gray. For the most part it 
is fairly well polished, but a considerable portion has artificially 
scratched and roughened surfaces. Sizes are highly variable, but 
medium to small vessels seem to predominate. Rims are usually 
direct or slightly flaring; broad strap handles, notched flanges 
(pi. 3, w) and projecting lugs occur ; and bottoms are either flat or 
slightly dimpled. Plain ware legs do not occur in our sample. The 
tempering of all the Naco wares is a fine micaceous grit. Particu- 
larly significant, though much less abundant, is the Naco painted 
ware (pi. 3). Characteristically, this ware has a white slip and 
painted, geometric or curvilinear decorations on both sides in red and 
black. One sherd (pi. 3, fl) of this ware is unusual in showing a 
portion of what appears to be a plumed figure. The painted vessels 
appear to have been small and flat-bottomed with either direct or 
slightly flaring rims. Tripod supports containing rattles are very 
common in Naco painted ware (pi. 3, s-w). A strange, four-pointed, 


bird- or animal-head foot is most the common type (pi. 3, t, u). The 
painted ware is not very well made, and the designs are usually 
badly eroded. A small proportion of unpainted and a few painted 
sherds have either heavily incised or raised geometric designs (pi. 4, 
q, s-v, x-s) in the interior. These were apparently made by some 
sort of a stamp. In one case the raised design left by the stamp was 
smoothed down and its outer border carved for emphasis. Three 
plain sherds show finely woven textile designs impressed on their 
inner surface (pi. 4, n-p). Sherds with incised designs also occur 
(pi. 4, r). On the whole, Naco ceramics consist of these two wares, 
the plain and the painted, but in excavations at mound I, two intrusive 
types occurred. The first of these, consisting of two fragments of 
European crockery (pi. 4, m), has already been mentioned. The 
second type consisted of three sherds of well-made, highly pohshed 
and painted ware which apparently belong to prehistoric ceramic series 
from other earlier sites on the Chamelecon and Ulua Rivers. 

Incensario fragments from Naco are of the usual frying pan shape 
(pi. 4, a) with the distinctive Naco painted designs. Two candelarios 
(pi. 4, iv) are crude but unique. They are made of unslipped coarse 
pottery and have tripod supports. They represent the only type found 
by us at Naco. Spindle whorls are quite common at Naco and are well 
decorated with incised designs similar to those painted on pottery 
(pi. 4, i, j). Undecorated " bobbins ", probably to hold cotton thread, 
are even more common (pi. 4, /). The occurrence of spindle whorls, 
bobbins and textile-marked pottery bears witness to the importance 
of cloth manufacture in aboriginal Naco. No distinctive type of 
figurine was noted at Naco. The various pottery heads, ranging from 
simple to complex, and the " speak no evil " monkey, are illustrated 
(pi. 4, h-f). Whistles seem rare at this site. Only one specimen 
was found (pi. 4, /i) and this animal form suggests Chiriqui, although 
the red and black paint design is in the Naco style. The only other 
artifacts noted were the ubiquitous obsidian prismatic flake knives 
(pi. 4, k), and fragmentary legged metates and manos of lava. At 
Naco, as elsewhere in Honduras, there appears to have been an amaz- 
ing emphasis on pottery in comparison with any other type of non- 
perishable artifact. Textiles and wooden implements have left only 
indirect evidences of their probable importance. 


This site, also known locally as " Potrerito de los Calpullis ", is 
located less than i kilometer in a direct line and about 2 kilometers 
by trail from Manacal (see map, fig. 2). It is a neatly arranged 


mound group and is one of the few in the Ukia-Chamelecon region 
that can be photographed to advantage (see Strong, 1937, fig. 70). 
The main features are four large mounds forming a rough square, 
with another low mound in the center. The largest mound, to the 
north, is about 2 meters in height, 27 meters in length, and 12 meters 
in width. The others are slightly smaller, those to the east and south 
being rounded rather than rectangular. The eastern mound had a 
trench, made by workmen from Manacal, in the east side. The central 
mound is about i meter in height with a diameter of 8 meters. It is 
connected with the eastern mound by a low neck. The four main 
mounds roughly correspond with the cardinal points, but there is no 
exact orientation. A low, stone-covered mound is located about 
40 meters to the west. The entire group is located on an open strip 
of high, flat land, flanked on the east by a deep gully and on the south 
and west by the steep river banks. An artificial terrace of river 
boulders borders the site to the south. Behind the site rise rolling 
pine-covered hills, and between it and the river proper is a densely 
wooded flood plain. 

No artifacts occur on the surface other than a very few sherds 
of plain brown ware. A rounded boulder in the central plaza suggested 
an ape's head somewhat similar to that shown in plate 16, figure 3, 
but we were unable to determine whether the stone had been actually 
worked. The men who had dug the deep trench in the eastern mound 
encountered nothing but stones and broken pottery. Pottery is visible 
in the cut to a depth of 2.5 meters. We ran a trench through the 
heart of the low central mound reaching a depth of i^ meters in the 
center. The upper meter consisted of soil with many large boulders, 
stones, and a few pieces of plaster ; below this was hard gravel. A 
few lava metate and mano fragments and a considerable amount of 
plain, brown pot sherds came from the upper meter. The Las Vegas 
ceramic remains are predominantly of an unslipped brown ware 
indistinguishable from cooking ware at Naco and in Ulua Polychrome 
sites. However, a few polished and slipped sherds occur, and some 
of these have linear designs in red and black. A few sherds of dull 
orange ware with red stripes, a small orange rim with red and black 
monkeys, and a hollow round tripod leg were also found. 


According to our workmen, this site received its modern name 
" because it is a place where they catch many fish ", a puzzling ex- 
planation unless one is aware that the name " Tres Piedras " may 
be given to any person or place of particular potency. In a sense 


Tres Piedras is a very miniature Copan since the Chamelecon River 
has nicely cross-sectioned it (pi. 2, fig. 4). It is located less than a 
kilometer down stream from Las Vegas and on the same or western 
bank (map, fig. 2). 

Originally, the site must have resembled Las Vegas in outward 
appearance, having four mounds enclosing a central plaza. In the 
photograph (pi, 2, fig. 4) two of these mounds can be seen in cross- 
section on the right and left of the cut ; the rear mound is visible 
to the left of the figures, but the fourth or nearest mound has been 
completely washed away except for the many boulders deposited 
in the river channel. A fifth mound, likewise cross-sectioned, occurs 
to the west. Unlike Las Vegas, the plaza at Tres Piedras had a series 
of three plaster floors, the highest at a depth of 1.5 meters below the 
present surface, the lowest at a depth of 2 meters. The material 
was a thick, white " mezcla " or plaster. The upper floor appeared 
to be flat, but the two lower floors each had one step rising to the 
east. From our limited excavations it was impossible to tell how 
extensive these floors originally may have been. Along the river bank 
they extended for about 10 meters, and a considerable amount of 
broken plaster was visible elsewhere on this general level and in the 
talus deposit at the foot of the bank. It seems probable that the entire 
court or plaza between the mounds was once paved, but until adequate 
excavations are made here this cannot be proved, nor can the nature 
of the steps or mound approaches be determined. Over the plaster to 
a depth of three-fourths of a meter is a thick deposit of large river 
boulders. These may have rolled down from the mounds or may have 
been placed here later to raise the level. 

Among the vast quantity of stones deposited in the river bed from 
the portion of the site that has been washed away are many that 
indicate human workmanship. The most tantalizing of these are a 
considerable number of large lava blocks that strongly suggest 
sculpture in the round. None, however, are definite enough for 
certainty, but they do give an impression of either a dying or a 
nascent sculptural drive. The " ape's head " from Las Vegas is of 
this type and may have been transported there from Tres Piedras. 
In addition, there are numerous squared blocks of limestone or gray- 
green schist, one circular block with abrupt edges, and several thick 
slabs with holes drilled through them. Metate and mano fragments, 
as well as lapstones without legs, occur. Stones are particularly 
concentrated in the river bed below what was once the position of 
the east mound. With them occur large fragments of plaster flooring. 
This flooring often contains boulders or shows the molds from which 


boulders came. Structurally, the Tres Piedras site appears to have 
been more pretentious than the majority of sites in the vicinity, and the 
remaining half is well worthy of complete excavation. 

We made two small stratigraphic cuts, the first west of the central 
mound group between that and the outlying mounds to the west. The 
second was in the heart of the plaza. Both sections were on the face 
of the river bank and each was 5 meters long by i meter broad, 
extending down to sterile soil. The first excavation yielded some 
pottery at 10 centimeters and reached barren soil at about 1.70 meters. 
The second cut passed through three plaster floors and reached barren 
soil at 2 meters. The first yielded the most potsherds although even 
here they were not overly abundant. Gray to red cooking ware was 
most abundant in each of the five 30-centimeter levels. A very few 
sherds of Mayoid polychrome occurred in all but the. bottom level. 
Above I meter all fragments of this type were from pots with buff 
to orange slips covered with florid, conventionalized, red, purplish, 
and black designs. Below i meter the same Mayoid types occurred, 
but in association with more realistic designs having human head 
panels. At this same level occurred fragments of an excellent Mayoid 
vessel with a panel of square, grotesque heads around the rim and, 
below this, an intricately carved design. The design had been carved 
after firing. Associated with these lower Mayoid types were a few 
sherds suggesting orange over buff negative painting ; and cooking 
ware with dull, dark red line decoration. The second excavation in 
the plaza yielded few but similar pot sherds. However, the occurrence 
of a small, restorable imitation Ulua marble bowl just above the 
upper floor at a depth of i^ meters was significant. Maya carved 
ware occurred at this same depth in the first excavation. The three 
plaster floors in this second excavation have already been mentioned. 


There are numerous mounds and other archeological sites in this 
region, but time to examine many of them was lacking. Close to 
Manacal Ranch is the site of Los Cocos, consisting of a few low earth 
and stone mounds that are being rapidly eaten away by the river 
(map, fig. 2). There is a deep 30-foot bank at this place. No notable 
structural details could be observed. The only pottery we obtained 
were some coarse, blackish brown sherds and one heavy, dull orange, 
sherd with eroded red and black designs. 

Farther upstream, beyond the mouth of the Naco River, is the site 
of San Luis (see map, fig. 2). Here in a cut of some 3 meters occur 
many river boulders and large amounts of broken pottery. The 


majority of the pottery is a coarse, brown or buff ware. There are 
also a number of heavy dull orange pieces with broad red stripes 
and some polish ; as well as a polished red incised piece and a fragment 
of a heavy platter with coarse red and black line decoration. In the 
talus below this bank were two large, square cut stones of volcanic 
origin, A few crumbling human bones were also found in the bank 
and on the talus. There are no surface mounds at San Luis, but 
broken pottery occurs from just below the present surface to a depth 
of about 3 meters. No stratigraphic changes in type occur so far 
as our very small pottery sample is concerned, but the site merits much 
more careful study than we were able to give it. 

As one rides past Cofradia on the way to Naco a few low mounds 
are visible to the south of the road just after one has crossed the 
Manchagualay River. We did not examine these in detail. Farther 
along the Naco road, about i kilometer from that village, there is a 
small Spanish colonial ruin located in dense bush about 20 meters 
north of the road. It is the foundation of a small house made of bricks 
and plaster, and the local people have tales concerning a magical 
cross of gold that was once found here. As already stated, the ruins 
of Naco extend for about i kilometer up the Naco River, and there are 
said to be numerous mounds across the river from the modern town. 
We visited the site of Quebrada Tostada, about two leagues upstream 
from Naco in a hanging valley some 400 feet above El Salto, the wild 
and beautiful falls of the Naco River. The main site at Quebrada 
Tostada includes 4 or 5 acres of pine- and thorn-covered land. Low 
stone and earth mounds are scattered over this area, and we found a 
few sherds of coarse brown pottery. Local tradition says that " the 
King ", i. e., Olid, fled to El Salto after he was wounded. We cut 
our way down the steep, rough gorge to the falls but found no signs 
of any settlement there. Our guide, Don Santiago Nolasco, said 
that there were many low mounds scattered over the hills and moun- 
tain valleys in the general vicinity of Quebrada Tostada, but he knew 
of no nearby site comparable in size to the ruins at Naco. 

There is another important site in this general vicinity which we 
had hoped to visit. This is the Bell Cave, which Blackeston (1910) 
located near the headwaters of a small stream flowing into the 
Chamelecon River, about 25 miles from the ruins of Naco. Blackes- 
ton obtained a considerable number of copper bells and a few other 
unusual artifacts at this site. We were told by Sr. Roque Hernandez 
of San Pedro Sula that the site was not yet exhausted. Just before 
we left Manacal, Sr. Juan Antonio Sarmiento of San Antonio Mahada 
offered to guide us to the cave which he said was near his home. 


Unfortunately, we were unable to make the trip. It would be very 
important to learn what types of pottery, if any, occur in association 
with these copper bells. Spinden (1925, p. 544) has suggested that 
the cache formed part of a Toltec trader's outfit. 


Our most extensive excavations on the Ulua proper were at Las 
Flores Bolsa and at Playa de los Muertos. In addition, ceramic 
samples were obtained at various river bank and mound sites between 
Naranjo Chino and the mouth of the Comayagua (see map, fig. 5). 
Our investigations were for the most part confined to the eastern 
bank of the Ulua. In a region as rich in sites as is the Ulua, it 
seemed better to confine our efforts to a few promising places rather 
than attempt too wide a survey. The depths at which cultural layers 
occur necessitated moving dirt on a very large scale for even a 
reasonable stratigraphic sample. On the Comayagua River, near 
Santa Rita, we made excavations similar to those at Las Flores Bolsa 
and at Playa de los Muertos. 


Las Flores Bolsa is located on the east bank of the Ulua River 
just south of the division line between the Las Flores and Naranjo 
Chino banana plantations. This was the farthest down-river site 
excavated by the expedition (see map, fig. 5). We worked here from 
January 20 to February 20, 1936. The site was chosen because of the 
fact that examination of the steep river bank from a dugout canoe 
revealed several human skeletons one above the other at this place. 
We therefore hoped for some sort of stratification. This was also 
the exact place where O. P. Swofford found a deformed skull with 
filed and inlaid teeth and with a jade bead in its mouth. This skull 
has been described as that of a Maya chieftain from Santa Ana (see 
Blom, Grosjean, and Cummins, 1933). It should be noted that the 
Las Flores site is actually a considerable distance downstream from 
Santa Ana (see map, fig. 5). In addition to fragmentary human 
bones there was a considerable amount of broken pottery projecting 
from the bank and on the small talus at the water's edge. 

We made two deep stratigraphic cuts paralleling the bank and ex- 
tending down to the water line. At the time of our work the almost 
vertical river bank was 5.25 meters in height. Excavation i was 
approximately 10 meters long by 4 meters wide. The top 2 meters 
was a recent sand and silt. Cultural debris, mainly broken pottery, 



Fig. 5. — Map of the lower Ulua and Chamelecon Rivers. 


occurred in the heavy clay below this for 2.25 meters ; below this was 
a sterile light clay loam extending to and below the water level. From 
the point where we struck the first artifact (at a depth of 2 meters), 
the soil was stripped off in successive layers each 25 centimeters 
thick, the first one being designated as P i (i. e., pottery level i) 
and so on through the occupation level. In all, 13 burials were en- 
countered in this excavation, 10 extended (P 5-9) and 3 bundle 
burials (P 4-6). The skeletons were in crumbly condition, and the 
skulls were badly distorted by the pressure of the earth. Only two 
complete skulls could be saved. Grave gifts were sparse, no complete 
pots occurring with any of the burials. One bundle burial had 2 clay 
spindle whorls (top of P 4), and another bundle burial (P 6) had i 
copper fishhook, 16 obsidian flake knives with needle sharp points, and 
a broken cooking pot containing bird bones. Layers of small adobe 
bricks and small baked clay basins near certain burials were the main 
structural features encountered. Broken pottery was quite abundant 
in this excavation, coming from 10 levels. The types and sequence 
represented will be discussed in the final report. 

Excavation 2 was 16 meters east of excavation i. It was roughly 
5 meters long, 4 meters wide, and 5.40 meters deep in the deepest 
portion. It contained only one extended skeleton fP 3). From the 
surface, mixed sand and silt extended down a little less than 2 
meters ; here the soil changed to a light clay. This layer of light clay, 
without artifacts, extended down slightly more than i meter. Be- 
neath this was a dark, heavy clay containing artifacts. Artifacts 
occurred throughout 7 levels or 1.75 meters. The cultural deposit 
sloped down toward the south (i. e., toward the river) so that it 
extended to the top of P 8 there, whereas on the north side of the 
excavation it terminated on top of P 4. A living level occurred 
in P 3 and it is the termination of this which sloped down to P 8 on 
the river side suggesting a refuse heap. In absolute level, P i in 
excavation 2 corresponds to P 4 in excavation i. Obviously, the 
deposition of refuse at this site had been little disturbed by burials. 
On the other hand, the occurrence of only three levels on the north 
side with what appears to be a dump heap (correlated with level P 3) 
on the south suggests that the deposit represented a relatively uni- 
form period of no very great duration. A bed of coarse sand occurred 
at a depth of 4.25 meters and below this was a light clay loam extending 
to and below the water level. Except for the sloping dump on the south 
edge, this stratum was devoid of artifacts. 

The various ceramic and artifact types from excavation 2 will be 
briefly discussed and any obvious stratigraphic changes noted. 


Pottery predominates tremendously over any other form of artifact, 
and plain or domestic wares are much more abundant than decorated 
wares. In all layers at this site the pottery shows the effects of water 
action, and the surfaces of many sherds are eroded. There is, however, 
no observable indication of re-deposition. The majority of sherds 
from all levels are of unslipped, undecorated wares ranging in color 
from a smoked or burned black, through brick red or brown to light 
buff. Sizes are highly variable. Pots with constricted and medium 
flaring lips are common, as are direct rimmed bowls. Vertical strap 
and solid round handles are most abundant. Many of these, from 
all levels, have a knob, filleting, or a crude monkey head on the 
bend. Bottoms are flat, rounded, dimpled, and annular, the first 
three types being most abundant. A few plain, hollow, conical feet 
occur. In P 6 and 7, large, thick, highly polished sherds also occur. 
Since grit tempering seems practically universal in the Ulua-Yojoa 
region, it may be taken for granted unless variants are mentioned. 
Domestic (that is coarse or household) ware with painted decoration 
is rare at this site. It occurs sparsely in P 3 and 4 where large 
vessels with high flaring necks are decorated with rayed circles, cross- 
hatches, or lines of dull red or brown paint applied in a splotchy 
fashion. Plain incised ware is rare but occurs in P 4 and 5 where 
necks are decorated with delicate, wavy, comblike patterns forming 
both vertical and horizontal patterns. 

With the wares which are both painted and incised we pass out 
of the strictly utilitarian class and find several intergrading types. 
A striking Las Flores type occurs in levels P 3-5 (pi. 5, a, h, c, d, e). 
These sherds are from thick-walled, vertical vases or bowls with high 
vertical necks, having a polished red slip, a band of black geometric 
designs below the lip and another band of incised design below this 
(compare Strong, 1935, pi. 18, fig. i, b, c, e, for similar Bay Island 
types). Another striking and distinctive ware, which occurs com- 
monly at Santa Rita (farm 17) (pi. 7, a-d) and rarely at Lake Yojoa 
(pi. 14, d), we have here called the Bold Geometric, monkey-handled 
type. It is very similar if not identical with Bay Island Polychrome 
II ware figured elsewhere (Strong, 1935, fig. 11). This is found 
in all levels at excavation 2 but undergoes some change in the two 
bottom levels. The typical vessel is large, with an orange slip and 
intricate black and red geometric designs around the neck, the body, 
and on the handles. The neck design is often of the interlocking textile 
type (compare pi. 5, c, and Strong, 1934a, fig. 54, and 1935, fig.ii), 
and the handle at the bend usually has a monkey head in relief with 
modeled or punctate features. At Las Flores, excavation 2, numerous 


sherds of this type have incised as well as painted designs around 
the neck. The more or less realistic birds and animals occurring on 
vessels of this type from the lower levels at Santa Rita (farm 17) 
(pi. 7, b-d) are lacking at Las Flores. In levels P 6-7, vessels of this 
type have lower necks, irregular handles, and incised as well as painted 

Polychrome sherds from thin-walled, vertical vases of so-called 
Mayoid type occur in all levels in excavation 2. The majority have 
florid, conventionalized, all-over designs in red, black, white or purple 
on buff, orange, black, or white slips (pi. 5, /, g, h, i, j, k, I, m). 

The majority of designs are elaborated and extremely conventional- 
ized reptilian, animal, mask, or anthropomorphic forms. They often 
cover the entire surface of the vessel and are difficult or impossible 
to reconstruct in their entirety from sherds. Crude skeuomorphic 
glyph bands occur from P 5-7, as do elaborately modeled projecting 
monkey or animal head lugs in the same levels (pi. 5, /, g). 'In some 
cases the designs are outlined with incisions. In the upper levels 
several sherds with red and purple spots occur (pi. 5,/). Bases are 
flat, dimpled or annular, and hollow cylindrical as well as solid, thin, 
rectangular, tripod legs occur in all levels. None of the isolated 
and graceful processional or " dancing " figures occur in excavation 
2, although a few sherds with this type of decoration were found in the 
deepest levels of excavation i. 

In addition to polychrome, straight-walled vases, a number of low 
bowls or small jars have similar types of designs. In P 1-2 occur 
polished red or orange sherds. In P 3 there are fragments of about 
six small jars with solid rectangular, tripod feet and eroded black and 
red designs. From P 4 to P 7, small tripod jars and low bowls with 
an orange slip, and red and black conventional or crudely realistic 
designs are common. These are in the Mayoid rather than the Bold 
Geometric tradition, though an occasional blending between these 
major styles occurs. In some instances incision is used to outline 
painted designs. In P 7 was found an unusual, restorable bowl of 
thin, polished ware, with an orange slip, and conventional, black 
and red, monkey and rosette designs outlined with incisions, a dimple 
base, and a low " vestigial " spout to one side of the direct rim 
(pi. 6, b). Three similar low "vestigial" spouts occurred in P 2-3 
as well ; hence they cannot be regarded as strictly early at Las Flores. 

From P 5-7 came a few fragments of Mayoid sculptured pottery. 

A restorable tripod vessel of this type is painted all over with an 

orange wash, except for the carved panel of elaborate Mayoid faces 

which apparently had no slip (pi. 6, d). A tiny vessel with a similar 



face panel is brown with no slip. It has an annular base and in shape 
rather suggests certain of the Ulua marble bowl types, though the 
sculptured design is Mayoid. A third fragment is the rounded 
bottom of a bowl with intricate Mayoid design in high but rounded 
relief. The slip was originally red but has disappeared except between 
the raised designs, and glittering micaceous tempering material shows 
on the surface. If it were not for Lothrop's statement (1936a, p. 142) 
that this mold-made appearance is due to delicate carving and the 
obscuring effect of the slip, one would be inclined to regard these as 
stamped or molded rather than carved. The type will repay much more 
detailed analysis than is possible here. From P 5 and 6 come three 
small restorable pots of the imitation Ulua marble bowl type (pi. 6, 
e, f). There are a few other sherds of this type. No slip is visible 
on these pieces, though all are considerably eroded. The association 
of Mayoid sculptured ware and imitation Ulua marble bowl pieces 
in the same levels may very well be significant. 

Incensario fragments occur in every level except P i and P 7. 
All seem to be of the usual perforated frying pan type with hollow, 
round handles. They lack painted decoration and range from light 
buff to brown in color. Fragments from P 2 and P 6 are very thin 
and delicate, but a fragment from P 3 is thick and crude. Candelarios, 
or small incense burners, occur in P 3 and P 5. All are of the un- 
slipped, single-hole type. That from P 3 is undecorated, whereas the 
two fragments from P 5 have crude linear incision and punctate 
ornamentation. Cassava-grinders, or round, handled, disks of coarse 
pottery, with one surface ridged with cross-hatched incisions like a 
grater, occur from P 3 to P 6. They are most numerous in P 3. 
Spindle whorls occurred only once, with burial A i in level P 4. 
Of the three, two were plain and one had neatly incised decorations. 

Figurines and whistles occur in practically all levels. They show 
little change in types from top to bottom. Solid, mold-made figurines 
of Mayoid type (like fig. 7, s) occur in P 2, 5, and 6. A portion 
of a pottery figurine mold was found in P 5. Modeled figures of thin, 
polished, brown pottery occur from P 2 to P 7. Some of these were 
originally rather pretentious (pi. 6, a), but nearly all are very frag- 
mentary and their original form often cannot be determined. Besides 
the human figurines and larger hollow statues, both solid and hollow 
animal and bird heads occur in all levels. Many of these were probably 
from whistles (like fig. 7, a, c, e). Similarly, many of the human 
figures once formed parts of whistles. Strange bulbous animal forms 
occur from P 2 to P 7. Some of these were whistles, others were not. 
A particularly interesting whistle from P 2 is in the form of a realistic 


frog with a small one on its back (compare Gordon, 1898, pi. 9, i, j). 
Pottery stamps likewise occur from top to bottom. From P i comes 
a round, stemmed stamp with a neat monkey design ; from P 5 a 
rectangular, stemmed stamp with a squirrel design and a butterfly- 
shaped stamp with two crude faces ; from P 6 an elaborate froglike 
stamp with small circles for designs, and, from P 7 a rectangular 
stamp with a geometric design. 

Compared to the amount of pottery recovered from this excavation, 
the total list of other artifacts is pitifully small. P i yielded i broken 
T-shaped drill of obsidian ; P 2, 30 fragments of obsidian flake knives, 

1 crude obsidian drill, 2 polished pebbles, i piece of crudely flaked 
quartzite ; P 3, 2 crudely chipped stones, i polished pebble, 2 pieces 
of baked clay with wattle and daub impressions ; P 4, i lump of clay ; 
P 5, 6 fragments of obsidian knives; P 6, i obsidian knife fragment ; 

2 quartzite stones, i smoothed piece of baked clay, i large alligator 
( ?) bone with 2 perforations, i tapering, cylindrical brick of baked 
clay; and P 7, the butt end of a small celt of hard green stone. This 
slim list clearly indicates what a tremendous proportion of the ancient 
material culture was perishable. Were it not for the advanced and 
abundant ceramic remains in prehistoric Ulua sites, one might reason- 
ably, but erroneously, conclude that only a very simple prehistoric 
culture had flourished there. 


Excavation work was carried on at this site by Dr. and Mrs. Kidder 
from the middle of March until the rising water level drove them out 
of the excavations on May 20, 1936. Work was also going on at 
Lake Yojoa, but all the other members of the expedition spent some 
time in the Santa Rita excavations. The site is located on the 
Comayagua River just below the little town of Santa Rita (map, 
fig. 5). It consists of refuse deposits and living levels exposed in 
the steep banks of a flood channel of the river and is only 200 meters 
west of the overseer's house on farm 17 of the Tela Railroad Co. 
The main irrigation canal for the lower valley takes out from the 
Comayagua just east of the overseer's house. Thanks to the courtesy 
of the Tela Railroad Co. and of the overseer, Mr. John Thompson, 
we were able to board comfortably at the farm house and to use its 
broad porches for sorting specimens. 

The physiographic and cultural evidences revealed by the Santa 
Rita excavations are complex and require far more detailed treatment 
than is possible here. However, certain very significant correlations 


between these factors are already apparent, and these can be briefly 
outHned. In all, three adjoining excavations were made at this site, 
the main stratigraphic cut designated as excavation i ; a northern ex- 
tension of this cut resulting from the discovery toward the close 
of the work of older and deeper cultural material; and excavation 2 
extending through a polychrome refuse heap to the east. For present 
purposes we will confine our remarks to excavation i and, to a 
lesser extent, to the northern extension. The beginnings of all these 
cuts can be seen in the illustration (pi. 2, fig. 3). 

A cross-section (fig. 6) of the west end of excavation i shows the 
outstanding stratigraphic features. This excavation was originally 
5 meters long from east to west, paralleling the cut river bank, and 
4.5 meters in width from north to south. Owing to the outward slope 
of the bank, the bottom of the excavation was 8 meters in breadth 
(fig. 6). When the May floods made further work impossible, we had 
reached a depth of 5.20 meters in excavation i and 5.40 meters in 
the northern extension. In size, the northern extension was less than 
one-third of excavation i. 

The cross-section along the west wall of excavation i (fig. 6) shows 
the various soil layers. The 2 upper meters consist of alternating 
deposits of dark silt, light silt, and sand. Below this is a thick deposit 
of dense clay which terminates at a total depth of 3.80 meters in a 
thin bed of sand or sandy silt (level 8, fig. 6). This sand layer has 
here a slight dip from the north and thins out near the southern 
edge. Beneath this layer is a sandy clay (level 9, fig, 6) which, with 
certain minor changes, extends down to the bottom of our excava- 
tion. On the extreme southern edge and in the deepest portion is a 
deposit of sand and gravel (fig. 6) which ran the length of the exca- 
vation and may represent an old stream bed. In the west wall cross- 
section the sand layer (level 8, fig. 6) seems to dip toward this sand 
and gravel deposit but, on the east wall cross-section, the sand is 
much thicker (40 cm) and extends on a level plane to the edge of 
the bank at a point i meter above the lower sand and gravel deposit. 

The first potsherds and other cultural detritus occur in the dense 
clay deposit simultaneously with a layer of river boulders (fig. 6). 
Throughout this clay deposit polychrome pottery is abundant, as are 
other cultural manifestations. The polychrome debris is thickest in 
a definite refuse deposit on the southern edge which dips slightly 
less than i meter below the main clay and pottery-bearing stratum 
(refuse heap, fig. 6). Debris extends down almost to the low sand 
and gravel deposit suggesting that it had been dumped over a low 
bank at the edge of an old water course. The refuse heap here 

NO. I 






Si ^ 


F m 


2 fc! 




< o 


terminates before the east wall of excavation i is reached, but beyond 
this point another polychrome refuse heap at the same depth occurs 
and it was in this that we made excavation 2. The sand level (level 8, 
fig. 6) was sterile of artifacts throughout all of our excavations. In 
excavation i, the same was true of the underlying sandy clay (level 9) 
except on the southern edge where the overlying polychrome refuse 
material dipped, and at the northern edge where a very few mono- 
chrome potsherds were encountered (level 9, fig. 6). These latter, 
found toward the end of our work, seemed highly significant and 
for this reason the northern extension was made extending north 
from the northwest corner of excavation i. Soil layers and pottery 
deposits were generally similar in excavation i and in the northern 
extension. However, in the northern extension more abundant pot- 
sherds differing from the polychrome type were found in the sandy 
clay (level 9, fig. 6) beneath the sterile sand stratum (level 8). This 
will be discussed subsequently. 

Cultural features, other than abundant potsherds and rare artifacts, 
were not marked in any of the excavations. River boulders occurred 
throughout the upper portions of the main clay stratum (level 7, 
fig. 6) in all. What appears to have been a roasting pit or oven is out- 
lined in figure 6. Small clay-lined fire pits and small irregular clay 
bricks also occurred. Eight burials, all in bad condition, were en- 
countered, four in excavation i and four in the northern extension. 
In all cases these occurred either in the clay stratum (level i) con- 
taining polychrome pottery or just below it and clearly intrusive 
into the sand. All were extremely friable and crumbled on exposure 
to the air. One of the burials in excavation I was extended and had 
notched upper incisors, three were flexed, only one had any grave gift 
(a ground stone knife). In the northern extension, one burial con- 
sisted merely of an immature skull, jaw, and humerus ; one was 
extended; one was flexed; and the last was the extended skeleton of 
a new-born child under a large, plain red, two-handled bowl. Excava- 
tion 2 yielded no burials. 

The succession of ceramic and artifact types from excavation i 
and the northern extension will be briefly outlined. This site was 
richer in polychrome pottery than any other we dug on the Ulua, but 
it should be remembered that even here there was much more plain 
than painted ware. The sherds from this site show little erosion 
through direct water action and the majority of the painted pieces 
are fresh and bright. We will discuss the material according to four 
major stratigraphic levels, A (P 1-3, see fig. 6), B (P 4-6), C (P 
7-9), and D (P 10-12). As indicated on the diagram (fig. 6) levels 


A include the upper portion of the clay occupation stratum. The 
southern face of the cut, including the uppermost portion of the 
refuse heap, had been removed to expose skeleton B i prior to com- 
pleting the diagram. Levels B include the lower portion of the clay 
occupation level and the upper portion of the southern dump heap. 
Levels C, the very bottom of the clay occupation level and the middle 
of the dump heap. Levels D include the lowest portion of the dump. 
With the exception of this southern polychrome dump heap, most of 
the remainder of levels C, and all of levels D, were devoid of artifacts 
except in the extreme northern portion where a very few bichrome 
sherds were encountered beneath the sand stratum (level 8, fig. 6). 
These will be discussed separately. 

Levels A contained a large amount of plain cooking ware of a 
red brown to blackish gray color. The vessels were fairly large, in- 
cluding direct bowls and pots with flaring necks and vertical handles. 
These handles are either round or flat in cross-section and, in a few 
cases, have a monkey head lug on the bend. Rounded, flat and 
dimpled bases and a few conical and round hollow feet occur. There 
are also some highly polished thin sherds tan or buff in color. The 
upper portion of B contained the same types but in the lower portions 
crudely painted ware superseded the plain cooking ware. In C and D 
plain cooking ware was very scarce except for a few very thick 
gray and brown sherds and some vertical strap handles. A portion of 
a thick, plain tray wdth horizontal handles occurs in C, and a plain 
annular base in D. Similar cooking ware but decorated with blotchy 
red or brown designs on neck and body occurs in A. These designs 
are usually rayed circles, criss-cross lines, and more or less irregular 
blotches. In B this type supersedes the plain ware in the lower levels. 
A squat, swollen pot form with flaring neck and vertical handles is 
characteristic. These are better made than in A, and the dull red, 
criss-cross line decoration on a lighter background sometimes sug- 
gests negative painting. This type also predominates in both C and 
D where undecorated domestic wares are rare. Plain incised ware 
is lacking in all levels. From both A and B levels came a few similar 
pot fragments in which the neck of the vessel is also incised with 
delicate, wavy, vertical lines and where the handle is replaced by small 
tripartite adornos. This variant of the swollen, simply-painted pot is 
more numerous in the lower levels, i. e., C and D. In D there is some 
blending of this type with the Bold Geometric, monkey-handled ware. 
Three sherds from thick-walled, vertical vases have a slip and painted 
designs with a band of heavily incised decoration around the upper 


body. They all come from A and closely resemble the more numerous 
representatives of the type from Las Flores Bolsa (pi. 5, a-e). 

The Bold Geometric polychrome type occurs in all of the lower 
major levels at farm 17. In A it occurs only in the lower third (i. e., 
level P 3, fig. 6). The vessels of the characteristic swollen olla type 
are medium rather than large in size. Textile and geometric designs 
are common (figs. 8, 9, compare Strong, 1934a, fig. 54, p. 46), but 
conventional birds and animals are lacking. In B similar designs 
occur in the upper two-thirds, and a few animal and bird designs occur 
in the lower third. In C and D animalistic designs (pi. 7, b-d, and 
fig. 10) are common, but geometric and numerous textile motifs also 
occur. The Bold Geometric vessels of the lower levels appear to have 
been slightly larger and better finished than those from the upper 
levels. Characteristic cursive, conventionalized bird, feline, bat, and 
reptile designs from the lower levels are illustrated (pi. 7, b-d, and 
fig. 10) and their association with geometric motifs indicated. In D a 
few Bold Geometric type vessels have incised patterns on the neck, 
similar to the squat, painted and incised domestic ware previously 
described. Bold Geometric monkey-handled bowls are numerous at 
this site and, with the straight-walled Mayoid vases, constitute one 
of the two most distinctive ceramic types. 

Straight-walled, vertical vases of Mayoid type are represented by 
sherds from all four major levels in excavation i. In A the pre- 
dominant, painted decorations are complex over-all designs on white, 
black, orange, or yellow backgrounds. Designs are in red, black, 
white, purple, and, in one case, blue. 

As at Las Flores Bolsa (pi. 5, f-m), the majority of the design 
motifs from A are elaborately conventionalized monster animal or 
human forms. One large fragment has a conventionalized jaguar 
with a row of conventionalized human heads above. An elaborately 
modeled and painted monkey-head lug occurs, as do hollow cylindrical 
feet and two annular bases. In B similar types occur, with the addi- 
tion of textile designs and the common occurrence of bands of con- 
ventionalized heads of several types (compare upper panel, pi. 8, a, b, 
and fig. 13). Squat, elaborated human or deity figures (pi. 8, d, and 
fig. 13) occur in this horizon and one of these is outlined with carved 
lines. An elaborate modeled monkey-head lug and a black monkey 
in low relief on a painted bowl came from levels B. One sherd with 
blue paint used as a design was noted. 

In C, panels containing paired " dancing figures " occur for the 
first time. This unique design motif, on beautifully polished and 
painted pottery, has been noted from northern Honduras to Salvador. 


Lehmann (1910, p. 740 and illus. 8, p. 736) believes that copulation, 
not dancing-, is indicated by this design and supports his view by a 
drawing of a Salvadorean example. To us, the latter seems no more 
definite than do the Ulua examples here illustrated (pi. 8, a, h, 
and fig. 14). In the light of Palacio's information regarding the cere- 
monial importance of the mutilation of male genitalia among Pipil 
and Lenca, we rather incline to connect this widespread design with 
phallic rather than procreative rites. Undoubtedly, the correlation of 
outer dancing figures with a unique design inside such vessels (fig. 14) 
is significant. This peculiar, and always slightly variable, inner design 
suggests some sort of record. It occurs inside " dancing figure " and 
certain processional vases and bowls from the Ulua River, Comayagua 
River, and Lake Yojoa. We suspect it also occurs inside Salvadorean 
vessels. This is an extremely interesting problem which at this time 
may only be mentioned in passing. Associated with the " dancers " 
are sherds decorated with isolated, processional figures. Like the 
" dancers ", these are usually well proportioned and graceful. The 
manner in which they are fitted into the simpler but more beautiful 
panels and design areas contrasts markedly with the florid, over-all 
designs of the upper levels. With these more realistic figures occur a 
variety of conventionalized human head designs (pi. 8, a, b). Simi- 
larly, the squat, conventionalized deity or priest figures (pi. 8, d, 
and fig. 13) also occur in association with the well-proportioned 
" dancers " and processional figures. Flat bases are most common in 
this level, and tripod feet are usually solid and rectangular or ovoid, 
though a few cylindrical feet occur. Lugs and annular bases do not 
occur in our sample. Levels D are identical with levels C so far as the 
Mayoid cylindrical vase shapes and designs are concerned. As was 
true of the Bold Geometric ware there is here also a slight but obvious 
development from the realistic to the conventional in painted decora- 
tions. It is significant, however, that during the time involved in 
these stylistic changes, neither the basic form of the Mayoid straight- 
walled vase or the Bold Geometric monkey-handled pot changed in 
any very marked degree. 

Fragments of Mayoid sculptured ware, as well as some examples 
of carved designs, come from B and C. From levels C there are two 
fragments from small jars in imitation Ulua marble bowl style (com- 
pare pi. 6, e, /). Here, as at Las Flores, Mayoid sculptured ware 
and imitation Ulua marble bowl incised ware are in close association. 
At Las Flores, excavation 2, these are in the lowest levels ; at Santa 
Rita, excavation i, in the two middle levels. 



Fig. 7. — ■Hollow figurines, whistles, and " candelario ", from the Ulua Poly- 
chrome period, Santa Rita (farm 17). (Specimens in National Museum of 
Honduras at Tegucigalpa.) 


Numerous small polychrome jars and vases are represented in 
excavation i. Certain of these are Mayoid, others Bold Geometric, 
and still others suggest blends between the two (compare figs, ii, 
12, 15). Any attempt to clearly delineate these two major Ulua 
polychrome styles, or to demonstrate the exact nature of their blend- 
ing, would necessitate a far more extensive analysis of design motifs 
than is possible here. Considered very generally, however, there 
are certain top-to-bottom variations which seem to be significant. In 
A, small red bowls with black geometric designs and conventionalized 
animal or anthropomorphic designs, either outside or inside, occur. 
Some of these are definitely Mayoid in feeling, having circle, diamond, 
or feather designs and dimpled bottoms. The majority, however, 
seem more closely allied to the Bold Geometric type. In C, an orange 
tone is particularly prevalent and numerous pieces show a rather 
unique blending of Mayoid and Bold Geometric styles (compare 
figs. II, 12, 15). Conventionalized birds, animals, and reptiles occur 
both outside and inside open bowls (compare the similar bat designs 
on two vessels from approximately the same levels in excavation 2, 
where one (fig. 15) has a Mayoid, the other (fig. 10) a Bold Geometric 
feeling). Both flat and dimpled bottoms occur in C. In D there are 
numerous small flat-bottomed jars of Mayoid type with processional 
figures and other elaborate anthropomorphic designs, and open bowls 
with Bold Geometric designs on the inside. These less clearly pro- 
nounced vessel forms, therefore, seem to recapitulate the tendency 
to change from realistic to geometric decoration observed elsewhere. 

A few fragments of polished gray ware came from Levels A, B, 
and C. The fragments were from small, slightly pear-shaped bowls 
without legs or handles. One fragmentary vessel had a narrow band 
of red paint around the inside of the neck. Another interesting feature 
in levels C is represented by two very definite spouts of red and brown 
polished ware. They are more similar to those from the deep layers 
at Playa de los Muertos (pis. 10, 11) than to the "vestigial " spouts 
from Las Flores (pi. 6, b). Strange to say, no Plumbate ware oc- 
curred in any of our excavations. 

At Santa Rita there are two other distinctive polychrome vessel 
types, both of which were lacking in excavation 2 at Las Flores. One 
of these, a flat plate on high tripod legs (pi. 8, e, f), may be called a 
tripod plate. The other, with somewhat higher walls, and either low 
(fig. 8) or high (compare pi. 12, /) tripod feet, may be termed a 
tripod dish. In excavation i, tripod dish fragments are lacking in A, 
fairly abundant in B, still more numerous in C, but rare in D. They 
characteristically have more or less intricate and geometric, red and 



black designs on a light red or orange back ground (fig. 8 and pi. 7, ^) . 
An unusual vessel of this type from excavation 2, which has loop 
handles and an auxiliary annular base, is also figured (fig. 9). In 
general, at Santa Rita, these vessels are in the Bold Geometric 
style, though elsev^here (as at Lake Yojoa, pi. 12, /) they may be 
more Mayoid. Tripod plate fragments or restorable pieces (pi. 8, 
e, f) are rare in levels A and B, fairly numerous in C, and rather 
abundant in D. Characteristically, the tripod legs are high with 
vertical slits and contain rattles. The plates are heavy and flat with 
slightly inward-dipping rims (pi. 8, /). The designs are often 




Fig. 8. — Ulua Polychrome, Bold Geometric tripod dish, excavation 2, Santa 
Rita (farm 17). (Specimen in National Museum of Honduras at Teguci- 

intricate, conventionalized serpents (pi. 8, ^) in black and dark red on 
a lighter red background. Although very involved, such designs are 
often very irregular in execution. The style seems rather unique 
but is more " Mayoid " than Bold Geometric in feeling. 

No incensario fragments came from levels A in excavation i. From 
B are nine unpainted incensario fragments, all of the perforated 
" frying pan " type. The handles are tubular, except one that is 
rather crude and solid. Two handles end in clutching triangular 
claws and one has slits down the side. The same number of frag- 
ments came from levels C, but half are painted with dull red and 
brown stripes or simple geometric polychrome designs. Two frag- 
ments came from D, one plain handle has a horizontal slit and another 
is painted with red and black. There is some indication here that 
painted incensarios may be relatively earlier than unpainted ones. 

NO. 1 



Candelarios are lacking in levels A. In B two were recovered, 
one decorated by an incised bird (compare fig. 7, ;") and one with 
incised lines. Both are of the single-hole variety. Levels C yielded 
one two-hole candelario decorated with a delicate incised pattern. Six 
candelario fragments came from D, one plain two-hole type and five 
single-hole specimens. One of the latter was decorated with a nicely 
executed textile design unit. There were no fragmentary cassava- 
grinders from levels A, but B yielded one, C five, and D seven. One 
of the latter is almost restorable. Like the others, it was of coarse gray 
pottery, round, with a broken strap handle on the rear and a series 




Fig. 9. — Unusual Ulua Polychrome, Bold Geometric dish, excavation 2, Santa 
Rita (farm 17). (Specimen m National Museum of Honduras at Teguci- 

of very rough-edged incisions or graters on the face. Only one plain, 
biconical, pottery spindle whorl was recovered. It came from levels D. 
Figurines and whistle fragments were, rather strangely, completely 
lacking in levels A and D, at excavation i, though they were fairly 
abundant in the two middle horizons, B and C. Excavation 2 yielded 
the finest assortment of such modeled pieces, and certain of these, now 
in the National Museum of Honduras, are reproduced here from our 
field sketches and photographs (fig. 7). All of these came from the 
polychrome horizon between pottery levels 8 and 11 in excavation 2, 
but, in general, are similar to the fragmentary pieces from levels B 
and C in excavation i . The latter types show no obvious stratigraphic 
differences ; fragments of large ornate busts and statues of polished 
brown pottery (like pi. 6, a) ; solid, mold-made figurines of Mayoid 



type (like fig. y, s) ; hollow faces with beards or ornate head dresses 
(like fig. 7, ni), bulbous human, animal or composite figures (like 
fig. 7, h, r), tubular birds (fig. 7, e), howling dogs, (fig. 7, c), and 


Fig. 10. — Ulua Polychrome, Bold Geometric bowl, excavation 2, Santa Rita 
(farm 17). (Specimen in National Museum of Honduras at Tegucigalpa.) 

a variety of squatting animals (like fig. 7, a, k, p) all occurred in both 
horizons. Some of the smaller human figures were once attached 
to whistles, but many are simply figurines, or ornate hollow statutes 
whose functions remain conjectural. Only a few exceptional pieces 

NO. I 



show any traces of painted decoration. From C came a fragmentary 
animal with a mouthpiece suggesting a spout. An incensario or pot 
cover, from P lo-ii in excavation 2, representing a deer similar to 
certain figures in the Dresden Codex, is remarkable (pi. 8, c). At 
present the distribution of these numerous products of the sculptor's 
art gives little indication of the lines of their development within 
the polychrome period on the Ulua. However, comparison with 
similar types from earlier horizons and a complete typographical 
analysis will be a large and important task. 






Fig. II. — Ulua Polj^chrome bowl, excavation 2, Santa Rita (farm 17). (Specimen 
in National Museum of Honduras at Tegucigalpa.) 

Pottery stamps are rare from excavation i , though they were fairly 
numerous in excavation 2. In excavation i, levels A yielded one 
cylindrical, roller stamp with a neat, squatting monkey design ; and 
one flat, stemmed stamp with a geometric design. Levels B and D 
yielded no stamps, but levels C yielded one flat stamp with a con- 
ventionalized serpent head design. 

As at Las Flores Bolsa, the disproportion between the abundant 
ceramic remains and all other artifact types was enormous in exca- 
vation I, Santa Rita. Levels A produced one large, conical, stone 
pestle and one obsidian flake knife ; levels B, six pieces of ground- 



down animal rib-bones and one ground stone knife ; levels C two small, 
polisbed bone needles, two obsidian flake knife fragments and one 
ground stone knife ; and levels D yielded, aside from ceramics, nothing 
but one small stone celt. 

Soil conditions in the northern extension were practically identical 
with those in the main portion of excavation i (compare fig. 6). 
The occurrence of four burials in the northern extension has already 
been noted. As in excavation i, the polychrome horizon in the north- 



Fig. 12. — Ulua Polychrome bowl, excavation 2, Santa Rita (farm 17). (Specimen 
in National Museum of Honduras at Tegucigalpa.) 

ern extension corresponded with the dense clay stratum (level 7, 
fig. 6) and was marked by a concentration of river boulders in the 
upper levels. In the northern extension the polychrome horizon 
(and burials) which were included in pottery levels A and B, termi- 
nated abruptly just above the sand layer (level 8, fig. 6). The latter 
was sterile and averaged 20 centimeters in thickness throughout this 
area. As in excavation I (excepting the polychrome dump heap on 
the southern border) the polychrome pottery horizon in excavation 2 
terminated abruptly on the sterile sand stratum. However, under this 

NO. I 






Fig. 13. — Ulua Polychrome, Mayoid vase, excavation 2, Santa Rita (farm 17). 
(Specimen in National Museum of Honduras at Tegucigalpa.) 


sand in C (P 7) there occurred a considerable number of potsherds 
of a different monochrome or bichrome type. These were in a clay 
stratum and were mainly concentrated in P 7, although they occurred 
very sparingly in D (down to P 12). Only three sherds came from 
P 12. This lowest clay stratum was sandier at the bottom of the 

Fig. 14. — Inside design from Ulua Polychrome, Lower Mayoid vases (pi. 8 a, b), 
excavation i, Santa Rita (farm 17). 

Space permits only a very brief analysis of the ceramic sequence 
in the northern extension. Dividing the 12 pottery levels (P 1-12) as 
in the main excavation, A (P 1-3) contained considerable amounts 
of undecorated cooking ware ; some fragments of straight-walled 
Mayoid vases with black slip, ilorid designs, and solid rectangular 
legs ; a few small Mayoid bowl fragments ; a few fragments of Bold 
Geometric ware ; and, finally, several sherds from thick-walled, 
painted and incised vases (Las Flores type, pi. 5, a-e). B (P 4-6) 
contained numerous sherds from excellent, thin Mayoid vases with an 
orange slip and well-executed bat and thin-line human designs in 

NO. I 



black and red (an excellent example from P 12, just above the sand, 
is illustrated in pi. 9, t). In addition, B contained a number of non- 
descript polychrome pieces. As in the main excavation, the lowest 
polychrome types in the northern extension were the best finished 
and had the most realistic and artistic designs. The sand level below 
P 12 was barren of artifacts. 

Below this sand level (.layer 8, fig. 6) potsherds were rather numer- 
ous in P 7 and occurred in very small quantities down to P 12 (i. e., 
through C and D but concentrated in the upper portion of C). All of 
these sherds are monochrome or bichrome and not a single example 




Fig. 15. — Ulua Polychrome bowl, excavation 2, Santa Rita (farm 17). 
(Specimen in National Museum of Honduras at Tegucigalpa.) 

of either Mayoid or Bold Geometric polychrome occurred. The 
sample is too limited to define the type adequately but is undoubtedly 
significant as indicating a dififerent and earlier ceramic type, here 
designated Ulua Bichrome, at this site. The majority of these lower 
sherds are monochrome ranging from highly polished red and orange 
ware to more numerous coarse brick red or sooty gray sherds. The 
highly polished red or orange sherds show examples of flat, heavily 
incised lips (pi. 9, i, n) ; swollen lips (pi. 9, /) ; flanges below the 
rim; flat bottoms; and small, solid tripod feet (pi. 9, aa, bb). They 
are from small vessels for the most part. The paste and tempering of 
these pieces is very fine and the ware is light and hard. A direct rim 
from a bowl of this type has a light gray polished interior. 

A number of the orange sherds (pi. 9, 0, p, q, r, s, u, v-z, aa, bb) 
are definitely of Usulatan ware (Lothrop, 1933, p. 50). The faded red 


or black linear designs on the bright orange background makes them 
very hard to distinguish from examples of negative painting since 
the slip at present appears to form the design, in contrast to the darker 
red or blackish overlay. Several sherds retain the black color of the 
original design, whereas in the others this has faded to a brown or 
even a dull reddish color. One very coarse potsherd, apparently 
from a flat tripod vessel, has a dull white slip on the inside with 
broad, criss-cross red lines (pi. 9, cc). Aside from the Usulatan type 
sherds this is the only painted fragment. This is similar to the red-on- 
white sherds from the old Playa de los Muertos horizon. 

Among the heavier, coarser sherds occur examples of low, flaring, 
swollen lips ; direct rims ; broad, vertical loop handles, smooth rocker 
zigzags (pi. 9, e), and both fine and coarse incised decoration (pi. 9, 
a, c). At the present stage of preliminary analysis this coarser pottery 
shows no very striking differences from the monochrome or domestic 
wares associated with the upper polychrome horizons. The polished 
orange ware and especially the Usulatan or related painted pieces are 
unique so far as this site is concerned. Aside from pottery the only 
other artifacts from these levels are a few fragmentary obsidian flakes 
(pi. 9, k, m) and a heavy, stemmed, pottery stamp with a geometric 
design (pi. 9, /). The stamp comes from P 8. The nature of the 
deposit below the sand level in the northern extension rather suggests 
the fringe of a midden whose concentration lay still farther to the 
north. Unfortunately, it was impossible to follow up this problem 
at the time, owing to the rapidly rising water level. These subsand 
layer ceramics at Santa Rita suggest definite affiliations with the oldest 
horizon at Playa de los Muertos. 


This important site is located on the east bank of the Ulua River 
close to the northwest corner of farm 11 (see map, fig. 5). In this 
general vicinity Gordon (1898) carried on extensive excavations in 
1895 and 1897, and later, in 1929, Mrs. Dorothy Hughes Popenoe 
(1934; also see Vaillant, 1934) isolated the old Playa de los Muertos 
culture at this exact spot. For this reason we visited the site on 
January 18, 1936, the day after establishing our headquarters at 
Progreso. First impressions regarding the possibility of further work 
were extremely discouraging. The terriffic flood of the preceding fall 
had removed most of the point where Mrs. Popenoe worked, as well as 
the entire island just below it (see map, Popenoe, 1934, p. 81). A 
small hard-pan or dense clay playa remained, on which we found a 
few Playa de los Muertos type potsherds. However, we found none 


in situ on the adjacent steep banks, nor did we note any traces of 
burials or of polychrome pottery deposits in the vicinity. It was 
apparent that, at most, only a tiny remnant of the area worked by 
Mrs. Popenoe remained at the small playa previously mentioned. 
This opinion, based on comparison with Mrs. Popenoe's map, was 
verified by Mr. Roberts, overseer at farm ii, who had assisted Mrs. 
Popenoe in her work. For this reason we sought other sites, hoping 
to encounter elsewhere, the older type of Playa de los Muertos 
material in direct relationship to the polychrome horizons. 

By the middle of April 1936 it was apparent that we were not 
going to find typical old Playa de los Muertos material at any of 
our other sites, despite the discovery of polychrome ware super- 
imposed on pottery suggesting the Playa de los Muertos culture at 
Santa Rita (farm 17). For this reason, while work continued at 
Santa Rita, the senior author returned to the Playa de los Muertos 
site on April 17. 

Subsequent to our first visit to the site, a large levee had been 
constructed along the river just east of the main site, and on the levee, 
and in the deep borrow pit or trench, we found numerous fragments 
of polychrome pottery. This material was concentrated at one place 
on the west wall of the borrow trench and here we later excavated 
(excavation 2). This site was only 80 meters southeast of the playa 
(with the old type sherds on its surface) which marked the eastern 
boundary of the grave area worked by Mrs. Popenoe. At this latter 
point close examination of the 4-meter bank behind the playa revealed 
a few sherds of coarse brown cooking ware and one tiny polychrome 
fragment in situ. Here excavation i was commenced. 

Excavation i, which was made on the very top of a point projecting 
out onto the clay playa, was L-shaped. The main cut was 2 meters 
wide and 6 meters from west to east. To facilitate handling the dirt 
from the deep cut, a north to south extension 4 meters long by i^ 
meters wide and slightly more than 2 meters deep was made from 
the east end of the main cut south to a steep bank on that side. 
The north wall of excavation i is illustrated (fig. 16 ; also see Strong, 
1937, fig. 79), and the position of the shallow north to south L 
extension is indicated by the shelf under skeleton i. The main 
east to west trench attained a maximum depth of 6 meters. The soil 
layers from top to bottom are well indicated in the diagram (fig. 16). 
Potsherds were first encountered at a depth of 80 centimeters at the 
west end and 1.30 meters at the east end. Scattered sherds extended 
through this layer of gray clay (P i, fig. 16) to a depth of 1.8 meters, 
where the sterile yellow clay began. The majority of these sherds 



I I I 



were from monochrome red to gray cooking ware, but enough poly- 
chrome sherds were found to establish the horizon as definitely 
belonging to the' polychrome period. Sherds were too scarce, however, 
to make 30-centimeter levels of value, so that the entire stratum 
was designated P i. Two extended skeletons occurred at the bottom 
of this horizon (fig. 16). Each was accompanied by a broken mono- 
chrome cooking pot, but definite polychrome sherds were also found 
next to each skeleton. Skeleton i had, in addition, an obsidian flake 
knife and a perforated pottery labret or ear plug. Below these 
skeletons we ran into a layer of yellow clay which was absolutely 
sterile. At a depth of 3.35 meters more sherds of a different type 
(Playa de los Muertos culture) were encountered coincident with 
our passing from the yellow clay into a hard brown clay. Owing to 
their abundance, it was now possible to work by 30-centimeter levels, 
thus P 2 (pottery level 2) began at this point. The Playa de los 
Muertos horizon (P 2-10, fig. 16) sloped down from west to east. 
As indicated in the cross-section (fig. 16) at least one and possibly 
two definite occupation or house floor levels, marked by black soil, 
concentrated charcoal, animal bones, sherds, etc., and a small deposit 
of mussel shells, were encountered. Owing to the depth of the deposit 
and to lack of time, it was impossible to work out these living levels 
beyond the walls of the excavation. No post holes were encountered, 
but baked clay with wattle and daub impressions was fairly abundant. 
In the west end sterile soil underlay P 6, but in the east end the 
occupation strata dipped to the top of P 10, terminating just above 
the then level of the river (fig. 16). 

It is obvious, both from the direct superimposition of the two 
ceramic horizons separated by a barren stratum (fig. 16) and from 
the markedly different ceramic content of each, that two distinct cul- 
tures are represented at this site. Of these, the lower or Playa de los 
Muertos horizon is the older. Since this horizon extended well to the 
west prior to the recent flood and since the main concentration of the 
upper or polychrome horizon occurred on a similar level 80 meters 
to the southeast (excavation 2), it would appear probable that only the 
edges of the two occupation levels overlap at excavation i (fig. 16). 
For the purpose of obtaining direct stratification, we were therefore 
extremely fortunate in choosing the place for our main trench. 

Excavation 2, in the west wall of the levee borrow-pit, was small 
but yielded a considerable amount of polychrome pottery. The 
excavation was 5 meters from north to south and slightly less from 
east to west. The first potsherd was encountered at a depth of 70 
centimeters and the lowest at 2.40 meters. No noticeable changes 


in polychrome pottery types were observable in this deposit, and 30- 
centimeter levels were not recorded. Abundant polychrome sherds 
were scattered throughout a gray to brown clay stratum. The pottery 
level contained concentrations of ash, charcoal, and sherds, one 
lenticular hearth, numerous small boulders, and abundant sherds. 
Below the pottery level an absolutely sterile, brown sandy clay was 
encountered. The maximum depth of this excavation was 3 meters. 
In absolute level the polychrome horizon at excavation 2 compared 
closely with the upper or polychrome horizon (P i, fig. 16) at 
excavation i. 

Excavation 3 was made on the northern side of the playa in the 
same dense brown clay level where old Playa de los Muertos material 
occurred in the main trench. At excavation 3, this level was on the 
surface, owing to the removal of the top soil by the river. An excava- 
tion 6 meters long (from northeast to southwest) and 1.5 meters 
wide was carried down to a depth of about i meter. No sherds or 
other artifacts were encountered below the surface and, as it was 
apparent that we were outside the area of ancient occupation, work 
was stopped. 

Before describing the artifact content of the various levels at exca- 
vation I, it will be well to discuss briefly the material from excavation 
2. All the ceramics (other artifacts were extremely rare) from excava- 
tion 2 correspond with those from the A (P i), the upper or poly- 
chrome level at excavation i. These two horizons are actually on the 
same level, and since material was scarce in A (excavation i) and 
abundant in excavation 2, the latter must be considered in order to de- 
fine the polychrome wares characteristic of the upper horizon. Owing 
to the apparent uniformity of all wares exposed in the cut bank at ex- 
cavation 2, it was considered as one unit. To check this, however, 
material from the very bottom portions was segregated for comparison 
with the remainder. This will be discussed after the bulk of the 
material has been analyzed. 

The domestic ware from excavation 2 is predominantly mono- 
chrome, of a dull red color. A much smaller number of sherds have 
traces of crude linear designs in brown, dark red, or black. The bulk 
of the domestic sherds appear to be from medium large vessels which 
were fairly well polished, with openings varying from heavy direct 
lips to slightly flaring rims. Vertical, solid, loop, and strap handles 
occur frequently. There are two dimpled bases and one partially hol- 
lowed, conical foot (from a unique vessel form). Six monochrome 
sherds are decorated with well-executed but simple incised geometric 


Nine rim sherds from finer vessels that were both painted and in- 
cised are of the thick Las Flores vertical -walled vase type (compare 
pi. 5, a-e) . These have a polished slip ranging from dark red to orange, 
a band of black geometric designs under the lips and below this another 
band of incised design. As indicated earlier, this Las Flores type of 
incised and painted ware is also represented on the Bay Islands 
(Strong, 1935, pi. 18, b, c, e). The sherds of this type from excava- 
tion 2 ( farm 1 1 ) also have inner and outer design elements that rather 
definitely suggest Bay Island Polychrome I pottery. One other sherd 
with more delicate painted and incised designs (similar to pi. 5, h) 
indicates the same fusion between the Las Flores painted and incised 
vase style and the Mayoid painted style that occurred at Las Flores. 
At excavation 2 ( farm 11), as at Las Flores, the Mayoid polychrome 
type of vertical vase is the more numerous. Sherds from these vases 
are very similar to those from Las Flores (compare pi. 5, f-m). They 
are relatively thick (compared to the vases from the lower levels at 
Santa Rita) with elaborate but conventionalized over-all designs in 
red and black on yellow buff. Geometric motifs such as crossed circles 
are also common. One flat bottom, one low, round, solid, tripod leg, 
and one thin, solid, rectangular, tripod leg occur. Two elaborately 
sculptured sherds have a curvilinear Mayoid design. One vestigial 
spout (identical with pi. 6, b from Las Flores) is from a painted and 
incised vessel. 

Smaller bowls with black and red designs on light red or orange 
are even more common than the Mayoid vase type. Some of these 
have conventionalized " Mayoid " figures but more have geometric 
designs such as lines and circles. They are small to medium in size 
including direct bowls, small pots, and small vases. One vertical strap 
handle, one flat bottom and numerous rounded bottoms occur. In style 
these vessels represent a blending between the Mayoid and the Bold 
Geometric with the latter style predominant. Tripod plates and dishes 
are lacking here as was true at Las Flores (excavation 2). 

Bold Geometric ware is fairly common and the large swollen vessel 
with broad strap handles occurs (like pi. 7, a). The monkey lug, 
however, is absent at this site. The slip of these pieces is a very dark 
polished red or orange with geometric designs in black. Animal design 
forms are lacking. One sherd of this type has a geometric design in 
white paint. Two typical deep dimple bottoms occur. No figurines, 
stamps, or whistles were found, but there is a brown pottery foot 
from a rather large hollow effigy. Two fragmentary prismatic flakes 
of obsidian were the only other artifacts. 


The material segregated from the lowest level in excavation 2 con- 
tains fragments of all these types and establishes the uniformity of 
the deposit. The domestic ware is identical, numerous pieces having 
blotchy dull red or brown designs. One well-polished sherd has a 
flange outside the neck with a dull red criss-cross design extending 
from flange to body. One sherd represents the Las Flores type painted 
and incised vertical vase. There are several small bowl fragments with 
conventionalized Mayoid and geometric designs, and one typical Bold 
Geometric swollen bowl fragment. A sample gathered from the sur- 
face of the borrow-pit is similar but contains several " Mayoid " verti- 
cal vase fragments rather suggesting the Bay Island Polychrome I 
type (Strong, 1935, pi. 21 and fig. 21). A portion of a very small 
tripod vase with red slip and black line decoration has an outer wall 
panel with excellently sculptured Mayoid faces in profile. In general, 
all the material from excavation 2, Playa de los Muertos, agrees very 
closely with that from excavation 2, Las Flores, and with pottery 
levels A and B in excavation i, Santa Rita. 

Returning to excavation i at Playa de los Muertos, we will first 
consider the material from P i, the upper or polychrome horizon 
(fig. 16). The fairly abundant domestic ware is identical with that 
just described at excavation 2. Two restorable vessels of this type 
accompanied the two burials in the lower portion of P I (fig. 16). 
That with skeleton i is a round-bottomed pot with a low flaring rim 
and two vertical round handles. It is of coarse brown, unslipped ware 
with triangular incised designs over the lower neck and upper body. 
The vessel with skeleton 2 is a polished black vessel with a direct 
rim and three small solid legs. In direct association with the coarser 
ware throughout P i, polychrome sherds occurred. The majority of 
these are small and some of them are eroded, but their type is definite. 
The majority come from small bowls with a red or orange slip. The 
lips of these sherds are painted red or black and similar linear designs 
occur on the body of the vessels. Ten small sherds are colored buff 
to orange and have remnants of complex red and black designs. Two 
orange sherds with red lines and large dots suggest the Bold Geometric 
ware. Two polychrome sherds are from flat bottoms, and one is a 
rounded flat bottom. One large, hollow, cylindrical leg with an orange 
slip and red and black designs is from a tripod dish. The leg has a 
vertical perforation in the lower portion and holes in the part joining 
the body. It originally contained a rattle. This type of vessel (compare 
pi. 7, e, f) was lacking in excavation 2. The only other artifact en- 
countered was a fragmentary prismatic flake of obsidian. Although 


the polychrome sherd sample from excavation i is small, it is very 
similar to the material from excavation 2. 

Playa de los Muertos culture material is abundant throughout an 
average of 2 meters in the lower portion of excavation i (P 2-9 
inclusive, fig. 16). Broken pottery comprises the bulk of the collec- 
tion, since no complete vessels were recovered by us and other artifact 
types were rare. This discovery of undisturbed refuse deposits entirely 
pertaining to the old Playa de los Muertos culture is exceptionally 
important. Not only does it give a representative and unselected 
sample of the culture but it also permits the inclusion of burial ma- 
terials obtained by Gordon and Popenoe as definitely pertaining to 
the older horizon. Although Gordon in his brief published report 
gives no data on relative depths and states that no observable strati- 
fication occured (1898, p. 38), it is undoubtedly significant, that the 
majority of complete vessels he illustrates (1898, pi. 7, a, b, c, d, 
e, h, k, n, 0, p, q, r, s, u) are characteristic of the older Playa de los 
Muertos culture. Furthermore, examination of his letters from the 
field and the occasional depths he recorded in cataloging, now on file 
in the Peabody Museum, indicates that all these complete vessels came 
from the lowest portions of his Playa de los Muertos (Largartijo) 
excavations. These undoubtedly were from burials of the old Playa 
de los Muertos period. The old burials excavated by Mrs. Popenoe 
are fully documented (Popenoe, 1934, pp. 65-79). All are from below 
4 meters in depth and contain only Playa de los Muertos materials. 
Since we found no entire vessels of the Playa de los Muertos culture, 
we have included outline sketches of vessels obtained from graves 
by Mrs. Popenoe (figs. 17, 18). Thus, each ware or ceramic subtype 
of this culture, established on the basis of our potsherd collection, 
can be illustrated in its complete form by a vessel from Mrs. Popenoe's 
collection pertaining to the same Playa de los Muertos type or subtype 
(also compare Mrs. Popenoe's illustrations, 1934, figs. 2, 6, 8, 11, 
12, and 15). The final description of the Playa de los Muertos cul- 
tural horizon must include a complete study of the abundant com- 
parable Gordon and Popenoe materials, but this is not attempted here. 
For present purposes we have grouped our 30-centimeter strati- 
graphic levels of Playa de los Muertos culture material (fig. 16, P 2-9 
inclusive) into two uneven divisions, an upper (P 2-4 inclusive), 
and a lower (P 5-9 inclusive). The lower grouping of levels, which 
we may call levels C (P 5-9), yielded almost twice as much material 
as did the upper levels, here designated as B (P 2-4), owing to the 
fact that level P 5 was unusually rich and overweighted whichever 
half it was placed with. This discrepancy can be avoided later when 



finer analyses are attempted, but for present purposes this segregation 
into a smaller upper and later grouping of Playa de los Muertos cul- 
tural materials (B) ; and a larger, lower, and earlier grouping (C) 

Fig. 17. — Outlines of vessels of the Playa de los Muertos culture obtained by 
Dorothy H. Popenoe. Not to scale, a, burial 2 ; b, c, burial 4 ; d-l, burial 5 ; 
m, burial 7. 

must suffice. Even such an arbitrary division suggests certain ceramic 
trends within the period that may well be significant. 

The ceramic materials from B and C fall into six main wares or 
ceramic subtypes based on surface finish or decorations (pis. lo, ii, 

NO. I 



Fig. 18.— Outlines of vessels of the Playa de los Muertos culture obtained by 
Dorothy H. Popenoe. Not to scale, a, burial 7; b, c, d, e, f, i, burial 8; g, h, j, 
k, I, m, burial 11. 


and figs. 17, 18). Each of these subtypes share common features of 
form and decoration uniting them into a very definite major ceramic 
type characteristic of the Playa de los Muertos culture. The most 
abundant sherds from both B and C may be described as (i) wn- 
slipped, rough, brick-red to sooty gray ware. These are often from 
large vessels with slightly flaring rims and necks of variable height, 
or from smaller vessels with low necks and swollen lips. Broad, 
vertical, strap handles are common in both B and C, but the great 
majority of handles in C have two or three vertical ridges and corre- 
sponding depressions down the outside. Two round, solid, vertical 
handles of large size from B and one from C have conical tenons on 
the ends for attachment to the body of the vessel. Of the nine basal 
fragments of this ware in B, eight are flat and one dimpled ; in C three 
of the four basal sherds are flat and are slightly rounded. One spout 
of this ware occurs. Decoration is rare, several sherds have incised 
lines forming criss-cross designs, and three sherds have a raised ridge 
with regular indentations about the greatest diameter of the vessel. 
Subtype i seems very similar to the plain or domestic wares character- 
istic of the upper or polychrome horizons at Playa de los Muertos, but 
the prevalence of flat bottoms seems rather distinctive. Owing to lack 
of space, this subtype is not illustrated here except for an outline sketch 
(fig. 17, b). 

Subtype 2, (pi. 10, a-h), slipped and polished orange-red to brown 
ware, is almost as abundant as subtype i in B, but only about one- 
fourth as abundant in C. Shapes in subtype 2 are very similar to 
those in subtype i, but the vessels were somewhat smaller (fig. 17, 
g, h, i, j, and fig. 18, e, g, i, j,k). One spout from this sub-type is from 
B and 2 from C. A rather heavy basinlike bowl or vase (fig. 17, k, I, 
and fig. 18, c) is rather common, as are direct bowls (fig. 18, g, I). 
Handles seem rare but a few vertical strap handles occur. Fluted 
sherds (pi. 10, /) are fairly abundant in B but do not occur in our 
sample of this ware from C. Fluted fragments are usually from the 
body portion of rather small rounded or swollen bowls (like fig. 18, m, 
in shape) . The fluting varies in width and is either vertical or diagonal. 
Incised lines often set ofif the fluted portions. Incised and modeled 
sherds of this type occur in B, but only incising in C. Several frag- 
ments from B have intricate and well-executed geometric and curvi- 
linear incised designs (pis. 10, 11). In C basinlike bowls with heavy 
incised designs are represented. In B a hand in high relief is the best 
modeled piece. Broad, flattened lips with deeply incised decoration (pi. 
10, h, i represents the type) are very common in B and fairly com- 
mon in C. Usually the entire rim is flattened and decorated, the rim ex- 


tending farther out on two sides, forming a handle or a definite tab 
(pi. 10, h, i). 

Subtype 3 is a dark gray to black, highly polished ware (pi. 10, 
i-n; figs. 17, m; 18, m). This is a very distinctive slipped ware with 
such a high polish and so much fire clouding that certain pieces have an 
almost purplish color. The forms are very similar to those in subtypes 
2 and 4, and fluted sherds and flat, heavily incised rims (pi. 10, /, h, i) 
are common to all. Fragments from basinlike bowls are common 
(pi. 10, j, m) and the incisions on such pieces are sometimes so deep 
as to suggest a series of outer flanges. There are no handles of this 
ware, but one small, solid, cylindrical foot occurred in B. This is the 
only foot noted in the entire Playa de los Muertos culture horizon. 
Material of subtypes 2 and 3 are about equal in amount but subtype 3 
is more abundant in C than in B. This suggests that subtype 3 is 
generally earlier than subtype 2. 

Subtype 4 is a slate-gray to buff, highly polished ware (pi. 10, o-s; 
fig. 17, d, e, f, k; fig. 18, a, b, /). In amount this ware is about equal 
to the two preceding subtypes and is slightly more abundant in C 
than in B. The majority of pieces appear to have had a light-colored 
slip and a subsequent high polishing that gives them almost a glazed 
appearance. The paste is exceptionally fine, and the pottery very hard. 
Irregular dark firing clouds are very common (pi. 10, r). In general 
the shapes are similar to those already discussed, but small bowls 
with low, slightly flaring and swollen lips are common (pi. 10, o, q, s). 
Several sherds have ridges, tabs, and- human features in relief, and a 
number of spouts of this ware occur in C. An unusual flaring, trumpet- 
like neck has been figured elsewhere (Strong, 1937, fig. y6, upper 
left). A fragment similar to this was found in the older deposits at 
Lake Yojoa by Mr. Rittenhouse and erroneously restored as a trumpet. 
Flat bottoms are common, but a few rounded bottoms occur. Handles 
are rare. 

Subtype 5 may be designated as a ware with a chalky white wash 
(pi. II, a-c). It is relatively rare in both levels but somewhat more 
abundant in C than in B. It should be noted that the majority of the 
figurines from B are of this ware (pi. ii, q, r). The majority of 
sherds come from heavy, direct bowls or from pots with low necks 
and slightly flaring lips. One extremely broad, vertical strap handle 
occurs, as well as two large spouts (pi. 11, a). The figurines and a 
few sherds of this type with painted designs will be discussed later. 

The sixth ceramic subtype from this horizon is comprised of various 
painted wares. Painted pottery is relatively very rare in the Playa 
de los Muertos culture horizon, yet forms an important and varied 


type. It was about equally divided between B and C, perhaps indicating 
that it was more abundant in later times since the smaller upper section 
yielded an equal amount of painted sherds. Red and black painted 
sherds constitute one type (pi. ii, /, g, k). Some of those sherds have 
alternate areas of black and red sometimes separated by incised lines 
(pi. II, g). In other cases these red and black areas are very irregular, 
and the colors form irregular blotches rather than controlled designs. 
Numerous flattened and incised rims of the very dark subtype 3 have 
flecks or small areas of red paint (pi. ii, /), others have red on black, 
or black on red, painted areas. The under side of such flattened 
rims are usually black. Several fluted fragments have black and red 
painted areas. A few small vertical handles and one very small 
horizontal handle (pi. ii, g, k) occur in this red and black ware. 
The red and black painted sherds are more numerous in B than in C. 
Red on bufif sherds are the next most abundant type (pi. ii, i, j, o). 
The majority of red on bufif sherds come from C, but the type is rep- 
resented in B. Most of these sherds come from small vessels with 
a red band on the inside and outside of the rim (pi. ii, /). The 
lower portions and the bottoms of these vessels were often red, and 
the remainder, except for red rims, was buff. Several have incisions 
in the red area showing the underlying buff. One rim sherd from a 
direct bowl, polished red on the outside, has faint red linear designs 
on the inside with the lighter buff showing through them like negative 
painting. A very well-modeled and painted red and buff lug comes 
from B (pi. II, 0). Several large sherds of coarse unslipped brown 
or buff ware have red bands on lip, neck, or rim (pi. ii, h). Two 
sherds in B and four sherds in C have a dull white slip with red 
lines or bands on the outside and in one case on the inside (pi. ii,m,l). 
They are from large vessels with low flaring lips and include one 
broad, vertical strap handle. Three sherds from B and one sherd 
from C have irregular white designs on a red painted background. 
These are from large, coarse vessels. In two cases the lip has a band 
of white inside and out ; in one there are broad, irregular vertical lines 
extending down the rim, and in another there are blotchy white de- 
signs on the inner surface. 

One polished, dull red rim sherd (pi. ii, n) has vertical lines of 
dull gray paint extending down the neck. This suggests negative 
painting, owing to the fading of the gray paint. It has already been 
stated that several of the other painted sherds with faded linear 
designs imitate negative painting. In passing it may be said that 
although Usulatan ware is not present in our Playa de los Muertos 
culture ceramic sample, it does occur at the site (Vaillant, 1934, p. 90). 


Gordon obtained an excellent complete vessel of this ware at a depth 
of "26 feet" (Peabody Museum C 1054), and we found sherds of 
Usulatan ware in the lowest horizon at Santa Rita (farm 17). It is 
probably one of the components of this early ceramic complex. In 
concluding this brief description, it is interesting to note that whereas 
the polished and incised ware of the Playa de los Muertos culture is 
of an advanced and mature type, the very small percentage of painted 
wares is highly variable and suggests an early, experimental interest 
in this technique. 

Of the remaining artifact types, the figurine fragments are perhaps 
the most interesting. From B come four fragments, all from hollow 
figurines and all but one with a polished white slip (pi. ii,v,r). Three 
of these, all with a polished white slip, represent a bulbous, seated 
type that is much conventionalized (pi. 11, v, r). The fourth is an 
unusually well-modeled face of polished brown ware (pi. 11, p). 
From C come 10 figurine fragments, all of which are solid and all 
but one without any slip. Three represent female torsos (pi. 11, t, 
u, v). Although very simple, they have a primitive naturalism that is 
rather pleasing. Six fragments of solid figurines are more frag- 
mentary but suggest similar types. Gordon (1898, pi. 10, d, f, g) 
shows complete examples of this solid, naturalistic type. The last 
fragment is also solid but has a dull, polished white slip like those from 
the upper level (pi. 11, s). Our sample is too small for certainty, but 
there is at least a hint that the hollow, slipped figurines were later, 
being preceded by the solid, naturalistic, hand-modeled type. 

Artifacts of materials other than clay were very rare. B yielded 
one small jade bead, four fragmentary prismatic flakes of obsidian, 
two retouched pieces of obsidian, two polishing stones (one stained 
red), one piece of pink chalk (?), and two irregular flakes of hard 
stone. C yielded one prismatic flake of obsidian, one small rectangular 
muller, and several battered hammer-stones. As is true of the later 
horizons, the proportion of perishable to nonperishable artifacts, other 
than pottery was very low in the Playa de los Muertos culture. It is 
interesting to note that animal bones were unusually abundant in this 
horizon, suggesting a considerable dependence on hunting. Over a 
dozen fragments of baked clay retain the impress of wattle and daub 
house construction. When the present midden material is considered 
in relation to the much more elaborate grave materials obtained by 
Mrs. Popenoe (1934) and by Gordon (1898), a reasonably complete 
record of the period is available. 



In addition to the sites intensively worked, numerous other mounds 
or refuse deposits were also examined. The majority of these are 
shown on the map Tfig. 5). Mound groups are abundant on both sides 
of the Ulua River but to date have yielded relatively little material. 
Aside from superficial examination and the gathering of small sample 
sherd collections, we did no work at such sites. In general, the mounds 
on the present valley floor appear to- be relatively recent and yield 
ceramics that are inferior to those of the earlier polychrome periods. 
However, until adequate work has been accomplished at such sites, 
the linkage between definitely historical sites, such as Naco, and the 
deeply buried, earlier polychrome periods will be obscure. Our failure 
to excavate mounds was due to lack of time, not of interest. Small 
excavations were made at two polychrome sites, one below Naranjo 
Chino and the other on farm 15 (see map, fig. 5). These yielded 
splendid polychrome sherds apparently contemporaneous with the 
lower levels at Santa Rita (farm 17). On farm 10 we excavated 
a shallow polychrome deposit containing pottery identical with that 
from the late polychrome horizon at the nearby Playa de los Muertos 
site (see map, fig. 5). It is probable that earlier polychrome deposits 
occur here also as indicated by the excellent, realistic Mayoid vase 
(fig. 19) which is reported as having come from farm 10. We have 
illustrated this specimen, which is in a private collection near Trujillo, 
because it is a splendid example of the best Mayoid tradition in early 
Ulua-Yojoa polychrome wares (compare pis. i and 8, a, h, also fig. 
30). It is also unique in showing ceremonial drinking among the 

North End of Lake Yojoa 

On February 22 and 23, 1936, Mrs. Strong, Dr. Wilson Popenoe, 
and the senior author stopped over at Jaral and visited sites east and 
west of that town where commercial excavations had been carried on 
(see map, fig. 20). At Aguacate we obtained a considerable amount 
of broken but restorable pottery that had been discarded by these 
earlier diggers. On March 9, Mr. Paul, Mrs. Strong, and the senior 
author returned to Jaral and remained there until April 6, carrying on 
excavations at various polychrome sites in the hope of obtaining at 
least a partial scientific record prior to their entire destruction by pot 
hunters. Our work was interrupted by " Holy Week " during which 
period neither work nor travel was practicable. Later, on May 26, 
Mr. Paul returned for a week in order to carry on deeper excavations, 

NO. I 






Fig. 19. — Polychrome vase of Ulua Lower Mayoid type, said to have been 
found downstream from Playa de los Muertos on the Ulua River bank. (From a 
private collection near Trujillo.) 




seeking an earlier type of culture at the Los Naranjos site. In this 
venture he was successful despite the very limited time available. 

Only within the last 4 or 5 years has the occurrence of prolific 
mounds around the north end of Lake Yojoa become a matter of sci- 
entific knowledge. The first of these were dug up by local farmers 
and, later with the permission of the Honduras government, J. B. 
Edwards carried on extensive excavations in the region. In 1934 Mrs. 
Doris Zemurray Stone " visited Los Naranjos and published a brief 

In 1935 Frans Blom, Dr. Jens Yde, and Prentiss Andrews spent 4 
days around Jaral in the course of the Tulane University-Danish Na- 
tional Museum Expedition." They explored around the various sites, 
worked with Mr. Edwards, and from him obtained collections of the 
polychrome ware. 

On our second visit we rented a dilapidated house in Jaral and 
boarded at the " Grand Hotel Rats' Nest ", as it was fittingly christened 
by Yde and his companions. Our genial host, " El Chino Alejandro ", 
however, made us as comfortable as his limited resources permitted. 
In Jaral we were also greatly aided by Capt. Evalyn Bush, and in the 
field by our two main workers, Paco Ramirez, of Dos Caminos, and 
Miguel Hernandez, of El Eden. Information furnished by Mr. J. B. 
Edwards proved very useful throughout our work. 

The major geographic characteristics of Lake Yojoa have already 
been touched upon. Outstanding archeological features of the plain 
at the northern end are, first of all, the great mound group and frag- 
mentary stone statues at Los Naranjos (great mound group, map, 
fig. 20) ; next the long earth mound or causeway, with its parallel 
ditch, just east of the road to Jaral (map, fig. 20) ; and finally the 
series of " ancient cemeteries ", or low house mounds containing 
burials, about 3 miles east of Jaral near the lake shore (map, fig. 20). 
In the following account we will attempt to present very briefly the 
major characteristics of certain of these features, each of which merits 
at least a full season's work and a complete report. Our primary aim 
was to determine the nature and association of the major ceramic 
wares present at such sites and, if possible, to determine whether strati- 
fication of culture horizons might be present. The apparent richness 

" 1934. PP- 123-128. Mrs. Stone, p. 128, mentions the occurrence of gold at 
this site. To the best of our knowledge, no metals of any kind have ever been 
found there. 

"See Yde, 1935, and 1936. The earlier report (fig. s) shows four typical 
Yojoa vessels; the upper has the "dancing figures" and is Mayoid in type. 
The three lower vessels are in the local Bold Animalistic style. 


and complexity of Lake Yojoa ceramics makes any brief description 
extremely difficult. However, since the majority of Yojoa collections 
in various museums are highly selective, even a preliminary account 
of the manner in which the various types of vessels and artifacts occur 
in situ, should have value. Since complete or restorable pottery ves- 
sels are more abundant in Lake Yojoa sites than in those previously 
described, we will discuss them in this preliminary report, leaving 
detailed potsherd analysis for a later time. 


Modern place names around the north end of the lake are usually 
derived from certain trees that mark favorable areas for milpa farm- 
ing. Aguacate and Aguatal (map, fig. 20) are so named, and it was 
here that the first finds of Yojoa polychrome pottery were made. Both 
sites have been sadly looted, and though we visited them on our first 
trip to the lake, we were unable to find any mounds or promising areas 
sufficiently undisturbed to merit scientific excavation. Probably a 
very large proportion of Yojoa polychrome vessels in collections inside 
and outside Honduras, have come from these sites which appear to 
have been exceptionally prolific. We reached them after a long walk 
along the trail to Dos Caminos (map, fig. 20), then cutting south 
through the generally low but very dense bush. Each site consists 
of a large (Aguacate is the larger) irregular area covered with low 
mounds ranging from barely visible eminences to some 2 meters high. 
Originally, the mounds were covered with rocks, many of large size, 
but at the time of our visit both areas were entirely covered with 
shallow, irregular burro wings and piles of rich black dirt and stones 
which obscured almost all natural contours. The dense bush added 
to the difficulty. The excavations ranged from 30 centimeters to about 
2.5 meters in depth and, starting in what originally seem to have 
been mounds, run labyrinthian courses wherever the machete-wielding 
" huaqueros " believed they were in mixed soil. Potsherds were 
abundant, polychrome pieces seeming to predominate over plain or 
cooking ware fragments, and a number of splendid and only slightly 
broken vessels lay about, indicating that they had been carelessly 
excavated and then abandoned in favor of harder and more resistant 
complete vessels. Numerous three-legged metate fragments of various 
sizes, roller pestles, rectangular mullers, and two large polished celts 
were noted. There were no human bones in sight, but our guide said 
that small fragments were occasionally encountered in association with 
complete vessels. Among the great number of rough, volcanic rocks 


that once formed these mounds we noted a few that appeared to be 
artificially squared or smoothed, and in one or two cases the disturbed 
rocks appeared to have once formed part of some simple artificial 

Although the occurrence of some domestic pottery and broken arti- 
facts, such as metates, suggests human habitation at these sites, the 
predominance of elaborately painted sherds and the reported occur- 
rence of very numerous deposits of complete polychrome vessels sug- 
gests a burial ground wherein the human bones had vanished owing 
to the damp, very humous soil. In the light of our later excavations at 
La Ceiba and Los Naranjos, it seems probable that both habitations 
and burials formerly occurred here, with the latter coming to be pre- 
dominant before the sites were finally abandoned. Except that Agua- 
cate covers a larger area than Aguatal, and that the rather closely 
adjacent sites have been given different names by modern farmers, 
the two seem to be identical in types of pottery represented, in the 
nature of the mounds, and in the complete manner in which they have 
both been looted. 

An analysis of all the vessels from, or reported to be from, these 
two sites would probably run the complex gamut of Lake Yojoa 
polychrome ware. Sherds of almost every Yojoa polychrome type 
were actually present in the old excavation pits, confirming the reports 
of various of the diggers that the majority of these vessel types occur 
in association. Our necessarily brief discussion of ceramics from these 
two sites is based, first, on 14 broken but restorable vessels which we 
ourselves picked up in the diggings at Aguacate on our first visit. Al- 
though these abandoned pieces may not represent the finest types from 
the site, they are definite as to site provenience and probably generally 
representative. Next, we were able to acquire a number of complete 
vessels obtained by local diggers at Aguacate and Aguatal, and in 
some cases to sketch and photograph other vessels from these sites 
which were not acquired (for example, fig. 30). Complete vessels 
thus obtained were sent to the National Museum of Honduras at 
Tegucigalpa, and our present illustrations were made from field photo- 
graphs and sketches. Some of these latter vessels, reported to be from 
Aguacate and Aguatal, but not excavated in our presence, may have 
come from La Ceiba. However, we talked directly to the excavator 
of each, and there is strong probability that the majority did come 
either from the place designated or from one of the adjacent south- 
eastern sites (see "ancient cemeteries", map, fig. 20). 

The 14 restored vessels we obtained at Aguacate fall into five major 
types in regard to form : first, straight-walled or only slightly flaring 



vases or jars, with or without tripod feet ; second, tripod dishes ; third, 
open bowls with dimpled bases ; fourth, two-handled pots with dimpled 
bases ; and, fifth, bichrome or monochrome pots with direct or slightly 
flaring rims and two or four vertical handles. The finest of the first 
type is a vertical-walled vase with three slightly hollow, low, cylindri- 
cal feet. In form and coloring it is almost identical with one illustrated 
(pi. i). It has the identical step and scroll design pattern on the lip, 
but the body design consists of two pairs of interlocked horned or 





Fig. 21. — Yojoa Polychrome bowl, Bold Animalistic type, Aguacate. (Specimen 
in the National Museum of Honduras at Tegucigalpa.) 

plumed serpents with spearlike flames coming from their nostrils 
(similar to pi. 13, cf). A low, flat-bottomed vase with thickened lips 
and paneled walls is too badly eroded to make out the design. A third 
piece of this general type has slighty flaring, straight walls and a 
dimpled base. Around the neck is a band of skeuomorphic glyphs, 
dark red and black on an orange background. On the orange body 
of the vessel there are two conventionalized parrots in dark red and 
black. Tripod dishes, the second type, are here represented by only 
one example (pi. 14, c). This is of medium size with hollow, cylindri- 


cal legs containing rattles. The body color is dark orange with panels 
of geometric and conventional designs in red and black. 

The third type, bowls with dimpled bases, includes five vessels ; 
the finest of these is of thin ware with a light orange background and 
elaborate design in dark red and black on the outside. Around the 
neck is a series of plumed Mayoid faces conventionally but exquisitely 
executed ; these are identical with those on a very similar vessel from 
Aguatal (pi. 12, c). The body has complex, human figures in the 
same elaborate style, but erosion prevents a clear understanding of 
the design. A smaller thicker bowl with a buff background has purple 
around the lips and on the body, enclosing buff circles in which are 
crude, conventionalized Mayoid faces. Around the neck is a buff 
band with black, skeuomorphic glyphs. A heavy bowl has a white slip 
with massive, dark red, dull orange and black panels, bands, and 
designs. On the sides are two heavy monkeys squatting in profile. 
One has a forward-sweeping plume similar to those on the priestly 
figures, the other has a backward-sweeping plume and a long tongue. 
A thick but well-executed bowl is light orange with dark orange and 
black designs. On the rim, these are geometric, but in two circular 
areas on the side are ornate birds, evidently the Muscovy duck, with 
strange, wrapped objects on their backs. The last open bowl is light 
orange with two extremely ornate black a:nd purple birds. It has iso- 
lated black stepped scrolls outside the lip (like pi. i). 

There are two vessels of the fourth type, i. e., bowls with two verti- 
cal strap handles and dimpled bottoms. One of these is light orange 
with a low straight neck and black and red geometric designs. On the 
sides these form two highly conventionalized monkeys whose raised 
faces with indented eyes project like lugs (compare fig. 22 and pi. 
13. c). 

The second vessel of this type has a low neck and more swollen 
body (like fig. 26). It is orange in color with a band of red and black 
geometric designs around the neck, a band of curvilinear designs 
around the shoulders, and three ornate concentric diamond designs 
down the body (fig. 26 had similar but more elaborate designs). 

The three vessels of the fifth type, monochrome or bichrome pots 
with direct or slightly flaring rims and two or four vertical handles, 
suggest domestic or cooking ware. The largest of these has a round 
bottom, low lip, and four solid, vertical, loop handles. It is a dull, 
slightly polished red with smoke stains on the bottom. A smaller but 
heavier vessel is similar as to handles and base. It is lower, however, 
is dark red, and has more polish. The third vessel is dull buff with 



vertical bands of dark red (like fig. 27). Unlike the above, it has a 
dimpled base and only two vertical loop handles. 

A very brief analysis of other vessels reported to be from Aguacate 
and Aguatal, probably including some from La Ceiba as well, will 
bring out the major types represented here. The majority of the 
straight-walled vases from these sites bear Mayoid designs, very 
often identical with those on similar vessels from the Ulua or Sal- 




Fig. 22. — Yojoa Polychrome pot, Bold Animalistic type, Aguatal. (Specimen in 
the National Museum of Honduras at Tegucigalpa.) 

vador regions (see pis. 12, b; 13, f, and fig. 30). Since these vases 
are the ones mostly highly valued by collectors, they are apt to pre- 
ponderate in purchased collections, disproportionately to their actual 
occurrence in the field. An exceptionally fine vase of this type, said 
to come from Aguacate, is illustrated (fig. ♦30). There are three 
design units ; two are seated priests, and the third is a monkey shown 
against a black medallion. The two priest designs are similar (fig. 30) 
except that the one not illustrated holds a five-branched scepter. This 

NO. I 



vessel was sketched and photographed in a private collection at Jaral. 
The tripod dish appears to be rare, but tripod plates, with either high, 
hollow legs (like pi. 12, /) or low, hollow feet (pi. 14, c) containing 
rattles, are rather common. The majority of these have conventional 
designs of the Bold Animalistic type, but Mayoid designs do occur 
(pi. 12, /), including skeuomorphic glyphs and "dancing figures" 
associated with textile designs. Yojoa vessels of this sort appear to 
be somewhat more variable in size than are those from the Ulua. 
Small, dull brown vessels, with or without low, solid, tripod feet and 
decorated in the imitation Ulua marble bowl technique of incising, 






Fig. 23. — Yojoa Polychrome bowl, Bold Animalistic type, Aguacate. (Specimen 
in the National Museum of Honduras at Tegucigalpa.) 

also occur (pi. 14, e). A few of these vessels, with lugs suggesting 
the Ulua marble bowl technique, have sculptured designs that appear 
more Mayoid. Particularly noteworthy at Aguacate are a few flat- 
bottomed dishes of highly polished dark brown ware having unique 
carved designs (pi. 14, /). These conventionalized designs are carved 
through the slip and, owing to the light color of the paste, stand out 
clearly. This seems to be a ware and decorative technique distinct 
from either the imitation Ulua marble vases or the Maya sculptured 
vessels. Open bowls vary in size and, as a rule, have two main types 
of design: Mayoid (often against a dark background) (pi. 12, c, e, 





Fig. 24. — Yojoa Polychrome tripod dish, Bold Animalistic type, Aguacate. 
(Specimen in the National Museum of Honduras at Tegucigalpa.) 






Fig. 25. — Yojoa Polychrome bowl, Bold Animalistic type, Aguacate. (Specimen 
in the National Museum of Honduras at Tegucigalpa.) 

NO. I 



and figs. 28, 29) ; or a combination of Bold Animalistic and geo- 
metric motifs including highly conventionalized birds (pi. 14, a, b; 
figs. 21, 24), monkeys, (pi. 13, a, b, c, and figs. 22, 23), alligators 
(fig. 25), peccary, and "dancing" jaguars (pi. 12, d). Although 
somewhat similar animal motifs occur on true Maya wares, these 
Yojoa forms are generally distinctive and are usually associated with 
other designs suggesting the Bold Geometric style of the Ulua. Mon- 


Fig. 26. — Outline of Yojoa Polychrome pot showing " vestigial " spout, Aguacate. 
(From a private collection at Jaral.) 

key designs occur commonly on the two-handled pots with dimpled 
bases (pi. 13, a, c, and fig. 22). The range of Lake Yojoa monkey 
designs is extremely wide and interesting. The Bold Geometric swollen 
vessel with monkey lug handles is not overly common at Aguacate or 
other Lake Yojoa sites but does occur (pi. 14, d). Such Yojoa vessels 
are smaller than the majority of those from the Ulua and often have 
vestigial lugs and less striking red and black designs. These vestigial 
handle-lugs are also very common on the dull bufif cooking vessels 
with dull red stripes (fig. 27). Another two-handled straight-necked 



type of pot from Aguacate is decorated with intricate dark purple 
designs on orange. One such vessel has a panel of isolated orange heads 
on purple around the shoulder and an intricate purple " mask " de- 
sign on the lower body and neck. The two lower design areas are 
separated by an ornate concentric diamond design. Figure 26 from 
Aguacate had a very similar design and shape except for the rather 
unusual but significant vestigial spout. Another vessel form, the an- 





Fig. 27. — Yojoa Polychrome cooking pot, Aguatal. (Specimen in the National 
Museum of Honduras at Tegucigalpa.) 

nular-based " salad bowl " type (pi. 14, g), also occurs at Aguacate. 
Cooking pots of dull bufif color with vertical dull red stripes and two 
handles (fig. 27) and four-handled, polished red pots are far more 
common at Aguacate than any of the selected collections would indi- 
cate, for the looters usually throw these away. 

No brief description can do justice to the diversity of Aguacate, 
Aguatal, and other Yojoa polychrome forms and decorative elements, 
yet it must be remembered that all of these come from the same sites 

NO. I 



and from depths that are rarely as much as 2 meters. Moreover, 
despite the occurrence of at least two distinct major styles, the Mayoid 
and the local Animalistic or Bold Geometric, the composition and 
even the colors of both are similar. There is, moreover, a great over- 
lapping of design elements. At Santa Rita (farm 17) the typical 
Mayoid and the Bold Geometric polychrome vessels, exclusive of the 
numerous intermediate types where they blend, seem more distinct 
than do the two major types at Lake Yojoa. Moreover, despite the 
great richness of color and design, the bulk of Lake Yojoa polychrome 








Fig. 28. — Yojoa Polychrome bowl, Mayoid type, Aguacate. (Specimen in the 
National Museum of Honduras at Tegucigalpa.) 

pottery appears to be technically inferior to either the Mayoid or 
the Bold Geometric ware at Santa Rita or other old polychrome sites 
on the Ulua. Occasional Yojoa pieces, and these are the ones eagerly 
acquired by collectors, have a fine hard paste and fast colors, but 
for every one of these, the looters discarded or ruined hundreds of 
pieces that were crumbly in texture, with faded or eroded paints. 
Had the peoples of early polychrome times on the Ulua had the arche- 
ological generosity to bury complete vessels with their dead, as did 
their contemporaries on Lake Yojoa, this comparison would be more 
obvious than it is at present. Analysis of paste, form, size, color, and 



design, and the intercorrelation of these factors in the Aguacate- 
Aguatal collections must await future pubHcation, but the foregoing, 
very brief, description may give some idea of their richness and the 
manner of their occurrence insofar as the looters have not destroyed 
all such evidence. 







Fig. 29. — Yojoa Polychrome bowl, Mayoid (?) type, Aguacate. (Specimen in 
the National Museum of Honduras at Tegucigalpa.) 


The first time we visited La Ceiba we walked in from Jaral by 
the trail to Dos Caminos and came back along the lake shore (map, 
fig. 20). The latter was an especially hard trip through the black mud 
and dense jungle of the lake shore. It was enlivened, however, by the 
profusion of orchids, animal tracks, and land and water birds we 

NO. I 



encountered. Later we always rowed down in a cranky old boat, 
forcing our way in to shore through the massed water hyacinths 
(see fig. 71, Strong, 1937) to land at the old stone steps (see map, 
fig. 20). These roughly laid tiers of unworked stone extend several 

■Tl^ V fyVfLU^/^i^ VlUUMJMW.'ALWMMM^MA^l'f.-OMlJ^y ^^.r^ 




Fig. 30. — Yojoa Polychrome vase, Mayoid type, Aguacate. 
(From a private collection at Jaral.) 

meters from the shore up over the barrier of volcanic rocks that lie 
just beyond. They api>ear to be artificially laid and of native origin. 

Site i 

Our first excavations were made in one of a series of low, rock- 
covered mounds about i kilometer southeast of the stone steps (site i, 


map, fig. 20). En route to the site one passes two sharp mounds about 
2 meters high that have been deeply pitted by pot hunters. There are 
a large number of such mounds in the general vicinity. The mound 
selected, mound i, when cleared, was 10.75 meters long from north 
to south, 10.5 meters wide from east to west, and about half a meter 
high. It had not been pitted. A trench i meter wide and 7 meters 
long was run from the eastern edge to just beyond the center of 
the mound. The soil at and above ground level was a rich black humus 
full of various-sized volcanic rocks. At a depth of slightly less than 
I meter below ground level, both stones and artifacts ran out, and 
a sterile, yellow clay was encountered. Above this, the black soil was 
flecked with the same yellow clay, indicating disturbance. The bulk 
of the sherds and charcoal occurred in the black soil below the natural 
ground level, the slightly raised mound area consisting mainly of 
rocks. No definite house floors, burials, or other structural features 
were encountered. 

Potsherds were the most numerous artifacts from mound i. No 
complete pots were encountered. The sherds were predominantly 
from unslipped brown cooking vessels with very heavy and solid 
handles which were vertical or round in cross-section. Coarse bufif 
ware with dull red stripes and similar handles was also abundant. 
Polychrome sherds were few in number, mostly from dull orange 
bowls and tripod plates with geometric red and black designs. Legged 
metate fragments, fragmentary roller pestles, rectangular manos, half 
an ovoid sandstone bowl (i m deep), and three prismatic flakes of 
obsidian were the only other artifacts. Mound i was apparently a 
habitation site judging by its contents. In all probability there were 
burials in the mound somewhere which led to the site having been 
covered with rocks, but this is problematical. 

Less than a meter south of mound i were two small rock piles which 
may once have formed one mound, since the space between appeared 
to have been pitted long ago. An L trench which sectioned both 
portions revealed similar soil conditions to those in mound i, except 
that the mixed earth extended to a greater depth. Artifacts were 
more numerous, as were also huge stones that had to be moved with 
crowbars. Five whole or restorable pots were found. The first, at 
1. 10 meters depth, was a tiny two-handled pot with its original red 
and black design almost entirely eroded. The second, 1.40 meters deep, 
was a larger, swollen-bodied red vessel with two vertical strap handles 
having knobs at the bend of each. A black geometric design on the 
neck was almost entirely eroded. Both these vessels were lying on 
their sides. Nearby, at a depth of i .30 meters, was a rather high-walled, 


dull orang-e open bowl with a simple geometric design in red around 
the neck. It is chiefly remarkable because it has wavy, vertical lines 
down the body that appeared to have been executed in negative paint- 
ing. Actually, these seem to have been formed by the disappearance 
of the dark red paint that once covered them, A particularly inter- 
esting little tripod vase came from a depth of 1.65 meters. Like the 
above, it was in an upright position. This vessel has a light orange slip, 
two broad black lines inside and below the lip and, on the outer wall, 
three prancing jaguars with raised heads and open mouths, vividly 
executed in purplish red and black. A black and red geometric panel 
separates the three identical figures (the design on pi. 12, d, is similar 
but less realistic) . Among the many sherds from about the depth of 
I meter we found enough fragments to restore an interesting bowl 
representing a bird with head, wings, and tail projecting (similar to 
but more elaborate than pi. 14, h). The basic color is light orange, 
and the rim, lip, bird head and tail, as well as the median portion, 
have dark brown, dark red, and white designs. All but the first of 
these vessels were broken when found, and only two were upright. 
It is possible that they had been placed with burials, the bones of which 
had disappeared, but it seems more probable that they had been dis- 
carded with the abundant sherds and other broken artifacts. Two 
more partially complete vessels were so soft that they crumbled to 
bits when we tried to remove them. 

Sherds were more abundant in mound 2 and of better quality. 
Cooking ware was abundant and similar to that in mound i, but there 
were also present a large number of fragments from large, straight- 
walled vases of Mayoid type. The majority of these had heavy dark 
red and black designs on buf¥' or orange, with hollow rectangular or 
cylindrical tripod feet. A rim with a painted twilled basketry design 
and several with typical Ulua conventionalized heads (on thin, hard, 
polished ware) occur. Skeuomorphic glyph bands also occur, and 
there is one painted " bird " head lug. A number of excellently painted 
tripod dish fragments and a number of very large dark red sherds 
with typical Ulua Bold Geometric designs are noteworthy. These 
include two broad straphandles with monkey face lugs. Two figurine 
heads were encountered at a depth of 1.20 meters. One of these is 
solid, suggesting a pouting " baby face " with hair but no head dress ; 
the other is hollow with an elaborate head dress, having a raised St, 
Andrew's cross above the forehead. Both are rather badly eroded. 
A few roller pestles, cylindrical manos, double-ended hammerstones, 
two crude obsidian side scrapers, one quartzite side scraper, and a 
few broken prismatic flakes of obsidian complete the artifact list. The 


occurrence of much charcoal and small amounts of animal bone bears 
out the evidence of the broken pottery in suggesting that this, too, 
was primarily a habitation site. Our workers, accustomed to under- 
cutting and burrowing in general, in their zeal, ruined our trench 
profiles at mound 2, and we decided to try another site. 

Attention may be called in passing to a small collection obtained from 
a rock mound similar to and very close to site i, at La Ceiba. These 
objects were dug up by a local pot hunter who heard us working and 
visited us. They included an excellent tetrapod dish (pi. 12, /) with 
conventionalized Mayoid designs, and feet representing an alligator's 
head and containing rattles. The colors are dark red and black on a 
yellow background. A small tripod dish with low, solid feet had a 
textile knot design with three pairs of crudely executed " dancing 
figures." These two vessels are of interest since they have Mayoid 
designs on a vessel form usually decorated in the Bold Geometric or 
Bold Animalistic style. One large broken whistle of unslipped brown 
pottery was unusually interesting since it represented a tusked monster 
almost identical to one found on the Ulua at Santa Rita (pi. 13, c, cf. 
fig- 7' P)- There were also a number of Mayoid figurine and bulbous 
animal whistles, including howling dogs, similar to those from the 
Ulua. The same mound had also yielded a rectangular and an ovoid 
bark beater, excellently made of polished gray stone. 

Site 2 

This excavation was on the southern border of the area intensively 
dug over 2 years earlier by J. B. Edwards. The rise or mound 
selected was less than i meter west of the remains of his headquarters 
shack (La Ceiba, site 2, map, fig. 20). From this point north there 
are a great number of irregular excavations both in mounds and in 
the areas between. There are numerous mounds in this immediate 
vicinity, and all of them are badly pitted. According to our men a 
very great number of pots came from this general area. The small 
rise or mound which we selected for work was not more than 30 
centimeters high and had three irregular pot holes on its surface. It 
sloped slightly from the volcanic dyke on the west, extended about 
18 meters to the east, and was 13 meters from north to south. Its 
surface was very irregular, owing to numerous volcanic rocks and to 
the old dirt heaps. We completed an east to west cross trench i meter 
wide through the center of the mound, but, finding that there were no 
regular structural details to be observed in this fashion, we carried 
out various extensions to the north and south. In cross-section the 
" mound " showed a top layer of darker soil averaging ten centimeters 


in thickness. Below this was mixed brown earth containing flecks of 
yellow clay and innumerable large volcanic boulders. At an average 
depth of 80 centimeters sterile yellow clay was encountered with still 
more volcanic boulders, many of such great size that they could not 
be moved even with crowbars. Aside from various sherds and broken 
artifacts throughout the brown earth there were no definite floors 
or other evidences of artificial structure except for five groups of 
complete pots evidently marking graves. It proved impossible to 
penetrate far into the sterile clay owing to the innumerable great 
boulders which apparently formed part of the natural volcanic dyke. 
Pottery deposit, or grave, i occurred just north of our cross trench 
on the edge of the mound. Here, at a depth of from i to 1.25 meters 
in the mixed soil just above the yellow clay and under a large number 
of great volcanic slabs, we found four pottery vessels (see Strong. 
I937> figs. 75, 'JJ^. Three were very close together (Strong, 1937, 
fig. yj^, and the fourth, an incense burner (not shown in the illus- 
tration), was 80 centimeters away. Three vessels were intact, but 
the fourth and finest (pi. i) was broken by another bowl which had 
been forcibly nested in it. The broken vase, when restored, (pi. i) 
was unusually interesting, since it depicted a processional group of 
priests calling to mind the description of Palacio (see p. 12). The first 
figure (pi. i) is the high priest with the ceremonial staff ; behind him is 
an assistant. The latter either holds a copal container or has removed 
the high priests' bustle with one hand and is reaching back with the 
other for one of the two objects carried by the third priest. These are 
probably incensarios, but they could possibly be obsidian mirrors or 
some other ceremonial objects. The three priests are followed by two 
musicians playing on wind instruments of an unusual type. From the 
attitudes of the figures, it would seem that the procession had just come 
to a halt prior to the performance of some rite. Further description of 
this vase is made unnecessary by the illustrations. The three other 
vessels are comparatively simple. The bowl nested in the broken vase 
(Strong, 1937, fig. y'] and 75, lower center) has a simple but striking 
black design on a cream-white background. Red and black designs 
occur on the lip, there is a black band inside the rim, and the under 
and inner slip is a dull orange. The small two-handled bowl (Strong, 
1937, fig. 75, lower left) is unusually interesting since it is of the 
Bold Animalistic type with geometric designs around the neck, a 
cursive and conventionalized, twice repeated animal and circle de- 
sign on the body, two handles with definite nodes on the bend, and 
a deep dimpled bottom. It has a bright orange slip with designs in 
black, dark red, and white. Thus, although the processional vase is 


definitely Mayoid in form and decoration, this accompanying vessel 
is indubitably Bold Animalistic in type. The incensario (Strong, 1937, 
fig. 75, upper) is very crudely made of coarse buff pottery with dull 
red bands around handle and rim. From their distribution it would 
seem that these vessels had been laid around a skeleton, all traces of 
which had disappeared. 

Pottery deposit, or grave, 2 occurred in our cross trench near the 
center of the mound at a depth of only 15 centimeters. It consisted 
of two vessels upright and side by side. The larger of these (pi. 13, a) 
has a bright buff' slip with geometric designs in black, dark red, and 
bright red on neck, body, and handles. On the central body it has 
an extremely conventionalized monkey face with a miniature body. 
It is an unusually conventionalized piece of the local Lake Yojoa 
Bold Animalistic style. The second vessel is smaller with swollen 
body, slightly flaring neck, two vertical strap handles, and a small, 
cross-incised node on each side of its greatest diameter. It is one- 
color bright red and, like its companion piece, very fresh in appear- 
ance. Both vessels have dimpled The larger pot contained one 
small ovoid bead of grayish jadeite or diorite, and the smaller pot a 
larger, cylindrical bead of greenish gray jadeite. The latter bead has 
a groove around one end and both have complete biconodont perfora- 
tions. The shallowness of the deposit may indicate relative recency 
and the extremely conventionalized type of Bold Animalistic design 
on the larger vessel appears to be late (pi. 13, a). The fact that 
each vessel contained a stone bead suggests deposition with the dead, 
although here again all trace of human remains had disappeared. 

One meter east of deposit 2 in the cross trench, at a depth of 35 
centimeters, there occurred two restorable little jars of chocolate- 
brown ware in close association with a larger restorable pot (pottery 
deposit 3). One of these little straight-walled jars has three low tripod 
feet and is decorated with an incised diamond and dot design. The 
other is flat-bottomed, has two vertical lugs and a carved or sculptured 
design in low relief. The lugs and form strongly suggest the small 
pottery imitations of Ulua marble bowls, but the partially restored 
sculptured design seems more Mayoid. The other small jar also sug- 
gests the imitation Ulua marble bowl type (similar to pi. 14, e). The 
third restorable vessel is a typical, two-handled, local Bold Animalistic 
pot with a striking, heavy black and red monkey design (similar to pi. 
13, c, and fig. 22). The broken condition of these vessels makes it 
uncertain whether or not they represent a grave offering. However 
the association of types at this depth is interesting. 


Pottery deposit, or grave, 4 occurred under a mass of great rocks 
and consisted of three nested pots at a depth of 72 centimeters. It 
was located 2.5 meters east of deposit 3 in the cross trench. The upper 
vessel (pi. 13, d) was inverted over an upright, smaller, two-handled 
bowl, and also contained a very crude, unslipped, and slightly shoe- 
shaped vessel with horizontal, solid, round handles. Inside the latter 
was one cylindrical, thin pottery bead. The upper vessel is a striking 
open bowl (pi. 13, d) one-half of the surface of which is eroded. 
The original slip is dull orange, but the entire outside was covered 
with black, against which a thrice repeated dull orange, dark orange, 
and red serpent design stands out. The serpent, with bulrushlike 
flames darting from its nostrils, is definitely Nicoyan in style. A band 
of conventionalized serpent heads circle the outside of the rim and 
two black bands the inside. The small two-handled pot is even more 
eroded. It has a light orange slip, two conventionalized red and black 
alligator designs (similar to fig. 25) and other geometric designs on 
the body. The vertical strap handles have definite nodes. The coarse 
brown slightly shoe-shaped pot is very badly eroded and lacks all 
surface finish. Despite the lack of skeletal remains, this pottery de- 
posit has all the earmarks of a funerary ofifering. It is particularly 
interesting since it contained only the local Bold Animalistic type of 
pottery in association with a shoe-shaped vessel. 

Pottery deposit 5 consisted of several vessels uncovered in the 
northwest quadrant of the mound. They occurred over a triangular 
area 2 by 3 meters in extent and may or may not have represented 
one or more grave ofiferings. No human remains were found here or 
elsewhere at La Ceiba. The first vessel was a small, straight-walled 
bowl with a band of red frets against a brighter background around 
the neck. The entire middle portion of the outer body is black but 
much of the surface is eroded. It was found in an upright position 
at a depth of 45 centimeters. The next is a small, swollen pot with 
slightly flaring lips and small, solid, rectangular, tripod legs. It is 
badly eroded but has traces of black and dull buff circular designs 
on a dull orange background. It was found in an upright position 
at a depth of 30 centimeters. The third vessel is a large, badly eroded 
bowl found upright and wedged in among great rocks at a depth of 
•50 centimeters. Traces remain of an intricate but conventionalized 
dark red and light orange design against a black background. Like 
many of the Yojoa pots it has two black bands inside, below the lip. 
At a depth of 32 centimeters a badly eroded straight-walled vase with 
solid, rectangular tripod support was found upright, covered with a 
broken bowl. The vase had only traces of black paint on the outer 


surface, but the bowl was slightly better preserved. It had an original 
orange slip, a band of small red horizontal chevrons outside the lip 
and two circular panels surrounded by a black background on the out- 
side. The design inside these circles was gone. These vessels had 
evidently been broken up by roots. Nearby, at a depth of 60 centi- 
meters, another upright bowl was encountered. It was badly eroded 
over its entire surface and crumbled to pieces when exposed. It seems 
probable that the original nature and finish of the individual pieces 
has more to do with their state of preservation than does their relative 
age or depth. 

About 8 meters northeast of the mound or rise described above, 
and only a short distance north of the remains of Edward's " casita ", 
was a rough stone cairn formed by about a dozen large stone slabs 
lying in rather orderly fashion. We commenced a trench at this point 
but soon ran into innumerable great boulders, laid in no particular 
sort of order. The trench yielded nothing but potsherds, fragments 
of bulbous whistles, and a few mano and metate fragments. The 
other mounds in this vicinity appeared to be similar to the one we 
cross-trenched. Many of them were higher, but all had been so badly 
pitted that further excavation seemed useless. 

Space is lacking to describe the potsherds from these two excava- 
tions. Elaborate polychrome types were abundant, a number of Ulua 
types such as rows of conventionalized heads and imitation textile and 
basketry designs occurred ; several sherds of brown engraved ware 
were noted ; and a number of large handles having monkey faces in 
relief on the bend, from red-on-buff cooking vessels, closely approxi- 
mate the Ulua Bold Geometric style. A few heavy, coarse sherds with 
rough incisions suggest graters, and a number of ground-down disks 
of polychrome pottery occur. The more localized Yojoa Animalistic 
and Mayoid polychrome types are generally the same as those described 
in the complete vessels and in the Aguacate ceramic material. Com- 
plete figurines are lacking, but a brown ware fragment, from a depth 
of 30 centimeters, depicts a woman's breasts supported by a bar or 
pendant as in certain Maya stone sculptures. A few bulbous bird 
and animal whistle fragments are present. Heavy volcanic stone 
metates, both with and without tripod supports, were fairly numerous, 
and both roller pestles and small rectangular manos occur. An ovoid 
wedge or chisel, lo centimeters in length, of hard gray-green stone is 
interesting. From a depth of 20 centimeters comes a flat slab of hard 
gray stone with a sharp, ground-down edge. Numerous prismatic 
flakes of obsidian, a few crude obsidian and flint side scrapers, and 
some red ochre, were also found. Round stone balls were fairly 


numerous at the site. The occurrence of two jadeite beads has already 
been noted. On the whole, nonceramic artifacts are more abundant 
in Lake Yojoa than in Ulua Polychrome sites. 

Site 3 

About one-third of a kilometer north of site 2 we briefly investigated 
what appears to be a quite different type of mound. To reach it one 
proceeds through the extremely dense bush past a great number of 
low, pitted, rock mounds (La Ceiba, site 3, map, fig. 20). Despite 
its relative proximity to the lake we doubt if we could have found 
it without the aid of Paco, The mound in question we called the 
" cut-stone mound ", because of the occurrence there of several large 
slabs which appeared to have been worked. The main structure is 
a rectangular platform-mound, 2.80 meters in height, with a north 
to south length of about 20 meters, and a breadth of approximately 
10 meters. The walls of this mound rise sharply, and the top, which 
measures roughly 14 by 6 meters, is rather flat. The south end, which 
faces the lake, has a more gradual slope, but the north end and east 
and west sides rise abruptly. This platform-mound is set upon a low 
circular rise, or mound, which has an estimated diameter of almost 40 
meters. It was impossible to clear this entire area with the time and 
men available ; hence these measurements are merely approximations. 

An excavation had been made near the center of the platform- 
mound which reached down to subsoil, a depth of exactly 2.80 meters. 
When cleared, the walls of this pit proved to be of brown soil con- 
taining, especially near the bottom, some potsherds and charcoal. The 
very bottom of the pit reached sterile yellow clay. No large rocks 
occurred in the walls of the pit, but we found a few just under the 
surface elsewhere on the platform. The local man who had dug the 
pit told us that he had found nothing. To the south, where the 
platform-mound rises from the low irregular substructure, we en- 
countered a row of boulders which seemed to form a lower border. 
Ten meters farther south, still on the sloping substructure, we un- 
covered a number of large, flat slabs, several of which appeared to 
have been more or less ground into shape. These were immediately 
adjacent to an old excavation containing other slightly worked, flat 
slabs. Our workers told us that four small pots had been found in 
this pit. Aside from being laid flat, none of these large slabs ap- 
peared to be in any particular arrangement. Two meters farther 
south on the outer edge of the substructure, we encountered a row 
of boulders and smoothed slabs laid end to end just under the surface. 
These slabs and boulders formed a definite border to the substructure 


which we followed for 8 meters, paralleling the south face of the 
inner platform-mound. 

Approximately 20 meters south of the " cut-stone mound " is 
another, lower mound and, running west-northwest of this, is a row 
of regularly aligned boulders barely projecting above the surface of 
the ground. These extend for about 20 meters and then take a 90°- 
turn to the south. We lacked time for further investigation, but it 
is apparent that both the " cut-stone mound " and its neighbors repre- 
sent a structural unit of some sort, the nature of which may only be 
determined by adequate clearing and excavation. The rough boulder 
and slab structures are similar to those at Agua Azul, to be mentioned 
later. Similarly, we encountered very few polychrome sherds in our 
brief work around the " cut-stone mound ", the majority being ©f 
coarse, plain ware. 


With the exception of the concentration of burial and other mounds 
near Aguacate and La Ceiba, the great bush-covered plain east of 
the Jaral-Potrerillos road appears to be without noticeable archeologi- 
cal sites. At present this is the area where most of the scattered 
milpa farming takes place, the soil being reported as very fertile. 
Just to the east of this road we discovered a great causeway and 
" canal " which separates the ancient ceremonial center near Los 
Naranjos from the main agricultural area and the burial sites farther 
to the east (see map, fig. 20) . Following up local stories of " an ancient 
canal to drain the lake ", we visited El Eden and found that the story 
had a basis in fact (Strong, 1937, fig. 73). Later, guided by Miguel, 
we followed the entire length of this structure from where it enters 
the lake to its northeastern termination on the Rio Blanco (map, 
fig. 20), an estimated distance of 5 kilometers. With the exception of 
perhaps 300 meters at El Eden which are cleared, the remainder of 
the structure is covered with dense bush, and we had to cut our way 
through. It took us about 5 hours to make the trip. 

The structure, which appears to be continuous, consists of a large, 
flat-topped causeway on the west, bordered by a definite borrow-pit 
or " canal " on the east. It enters the lake about one half a kilometer 
east of Jaral. Here the borrow-pit is 25 meters wide and the mound 
to the west about 8 meters wide and .75 meter high. To the east of 
the borrow-pit is a rise of about i meter. Where the mound crosses 
the trail from Jaral to Dos Caminos (map, fig. 20) it is about 14 meters 
wide and 2-3 meters in height. The ditch is not visible at this point. 
About three-fourths of a kilometer farther north the mound is 5 


meters across and 4 meters high. The borrow-pit or ditch is 25 to 
30 meters wide and is flanked by high ground to the east. At El 
Eden, where it crosses the road to the cemetery, the mound is 21 meters 
wide, 3.5 to 4 meters high on the ditch side, and 2-3 meters high on 
the west side. The ditch here is 9 meters wide across the bottom. 
About 100 meters north of this road is an apparently intentional break 
in the mound wall about 16 meters wide (Strong, 1937, fig. y^))- 
About 22 meters farther on is another smaller break, perhaps worn 
through by an old road. With these exceptions the mound or causeway 
appears to be continuous throughout its entire length, though the 
poor visibility due to the dense bush prevented our perceiving all 
details as we cut our way through. About half a kilometer north of 
El Eden the mound, now definitely turned to the west, crosses the 
road, where it shows in cross-section on the east side. The ditch 
here is not marked. Beyond the road both mound and ditch again 
become very definite with fairly steep walls. Here, as elsewhere, the 
mound has a flat top. Both terminate in a series of mysterious, deep, 
dry gorges which mark the underground course of the Rio Blanco. 
Miguel pointed out another series of deep pits or small gorges just 
south of here extending to the west (map, fig. 20), which he said 
marked the course of another underground stream called the Jutosa. 
At the time of our visit (April 4) no water was visible in either stream 
at this point, but during certain seasons the water level is said to rise 
to a considerable height. 

It is certain that any clear understanding of the function of this 
interesting causeway and " canal " will depend on an equally clear 
understanding of the nature and recent history of these mysterious, 
semisubterranean streams. Hidden amidst almost impenetrable bush 
and marked by precipitous canyons and yawning sink-holes, the solu- 
tion of the problem of the Rio Blanco, which apparently drains Lake 
Yojoa by some subterranean passage, is not one to be lightly attempted. 
There is probably some connection between the past level of this 
stream and the "canal" in question. If the river level was at one 
time higher than at present, the " canal " would have served to irrigate 
a large portion of the lower plain. Strange to say, local tradition 
reverses this explanation and claims that the ancient Indians sought 
to drain the lake ! Since returning to Washington, the senior author 
has also heard a story that a canal was dug in this vicinity about 1880 
by a commercial company with some similar end in view. We have 
as yet been unable to secure more definite information in this regard. 
We are unable to state positively that the causeway and ditch are 
not of historic origin, but, to say the least, this seems highly improbable. 


From the slope of the terrain we would estimate that the northern end 
of the " canal " is at least 50 to 60 feet higher than its southern termina- 
tion on the lake shore, hence any attempt to " drain the lake " would 
be absurd. This same factor, however, would favor the theory of a 
great central irrigation ditch, should geologists determine that the 
water level of the Rio Blanco was once considerably higher than at 

There is another possible explanation which emphasizes the con- 
tinuous mound or causeway and accounts for the ditch or " canal " 
as merely a bor row-pit. A glance at the sketch-map (fig. 20) will show 
that the causeway might well have been a ceremonial or defensive 
structure enclosing the great mound group west of Jaral, since it 
extends from the steep, encircling mountains all the way to the lake. 
A flat area is thus entirely enclosed and in the center of this rise 
the great mounds of Los Naranjos (fig. 20). Here we must leave 
the problem, the true answer to which must depend on the cartographer, 
the geologist, and the adequate excavation of the archeologist. 

Pyramids and Stone Statues near Los Naranjos 

The dominant archeological feature on the north shore of Lake 
Yojoa is the extremely impressive group of great mounds, or pyra- 
mids, located about 20 minutes' easy walk west of Jaral (see map, 
fig. 20). This site was first described by Mrs. Doris Zemurray .Stone 
(1934) as the southernmost known Maya city and designated Los 
Naranjos, after the little modern village to the west. J. B. Edwards 
has made what appears to be an excellent sketch map of this site 
based on his own explorations. He very kindly furnished us with 
a copy of this. So far as our own sketch map (fig. 20) is concerned, 
we have located and numbered these mounds in general accordance 
with Mr. Edwards' map, omitting, however, mounds 6 and 7, slightly 
east of the main group, which we did not examine ourselves. Since 
the Los Naranjos mounds or pyramids cover a large area and are 
all covered with dense forest or bush, the preparation of an accurate, 
surveyed map would be a considerable task. Until this is accomplished 
Mr. Edwards' map is the best available and, so far as our own limited 
explorations went, seems generally accurate. We have not reproduced 
it here, however, since it was primarily made for Frans Blom and will 
probably appear in connection with publications of Tulane University. 

The Los Naranjos mounds or pyramids are of great size, as indi- 
cated by our photograph of one of the smaller terraced mounds 
(mound i, pi. 16, fig. 4) . Yde has overdone it, however, when he shows 


a photograph (1935, fig. 4) of the sharp, natural hills behind Los Nar- 
anjos with the caption, " View of the Mounds at Jaral." His photo- 
graph, probably taken from mound i (pi. 16, fig. 4), overlooks the 
great mounds to the south which, however, are shrouded in jungle 
and do not show in the picture. As a result, the reader might easily 
assume that the natural hills which do show are the pyramids. Per- 
haps the translation should have been " View over the Pyramids ", 
rather than " View of the Pyramids." Mound i, (pi. 16, fig. 4) 
appears to be terraced, and we estimated its height at some 6 meters. 
Mound 4 (fig. 20) is much larger and higher, perhaps 8 to 10 meters. 
It is terraced and has a number of smaller mounds forming a court 
on the top. Jhere has been considerable digging here, probably by 
road workers seeking paving stones, as well as by pot hunters. Sherds 
seemed to be scarce on the surface. The other mounds appear to be 
smaller than mound 4, but several of them are terraced and all are 
worthy of careful mapping and investigation. Owing to the dense 
bush, it is impossible to make adequate observations without a great 
deal of clearing. Since we lacked facilities for this type of work or 
for any large scale excavation, we limited our own activities to smaller 
sites on the norther border of the great group (site i, Los Naranjos, 
map, fig. 20) . 

The occurrence of a number of fragmentary stone statues at the 
Los Naranjos site is particularly interesting (Stone, 1934; Yde, 1935 
and 1936; and Strong, 1937). These have all been removed from 
their original sites, probably by road workers, who have undoubtedly 
broken up and carried ofif others. Those which we located were lying 
in the great plaza between the Los Naranjos mounds at three places 
near the Jaral-Los Naranjos trail (map, fig. 20). Probably these 
had once been placed on top of mound 4, or one of its neighbors, and 
later tumbled down by the road workers. One statue represents the 
body of a man or ape, with one hand resting on the hip, the other 
crossing the body and resting on the shoulder (pi. 16, fig. 3). Feet, 
arms, and the head had all been broken ofif long ago. The material 
is a hard, gray, volcanic stone, and the body at present is i meter high 
and 50 centimeters wide across the shoulders. The neck break at 
present measures 23.5 centimeters from front to back. The body has 
a primitive simplicity and grace despite its solidity. Aside from two 
parallel incised lines on the back, suggesting a belt, there are no other 
notable features. Mrs. Stone describes a similar mutilated figure with 
a string of beads around the neck (1934, p. 126) ; hence there must 
be at least two of these figures. Next to this stone body was a large 
grotesque head (pi. 16, fig. 3) which evidently belongs to the body. 


although the uniform patination indicates that the original break oc- 
curred long ago. On the head the outline of breakage at the neck is 
very similar to that on the body. It measures 22.5 centimeters from 
front to back. The apelike head is disproportionately large for the 
body and distinctively prognathous. Anthropomorphic characteristics 
are the elongated and, presumably, decorated ears and a row of incised 
circles down the back of the head. Next to the other anthropomorphic 
torso with beads around the neck described by Mrs. Stone (1934, p. 
126) there was a similar head. We would be inclined to regard this 
as identical with the one here figured (pi. 16, fig, 3), but Mrs. Stone 
does not mention the macrocephaly, which is so outstanding in the 
head here figured, and states that it had circular ear plugs. From her 
description of it as " thick lipped " and " soft nosed " with eroded 
features, the two, if not the same object, must have been similar. 
We saw only one head and Yde and his party none. 

Mrs. Stone (1934, p. 126) calls the anthropomorphic statue a 
" stela ", and stresses the position of the hands as representing " with- 
out a question of doubt, the Mayan sign for submission." Yde (1936, 
figs. 3, 4, and pp. 27-29) also figures this statue and apparently con- 
curs with the interpretation of Mrs. Stone. In our opinion, neither 
the body nor the head are Mayan. Rather they appear to us as 
closely related to that widespread, and probably older, " Chorotegan " 
style of stone statue which occurs commonly in Costa Rica, Nicaragua, 
the highlands of Guatemala, and, rarely, on the Ulua (Lothrop, 1921). 
The archaic simplicity of the torso, plus the crudity and simian 
characteristics of the head, seem totally non-Mayan in feeling and 
technique. The position of the hands alone suggests a definite Maya 
convention, which may be relatively early but was certainly in vogue 
at a late period at Chichen Itza (see Tozzer, 1930, pp. 155-158). 
However, the same position of the arms, as well as crossed arms with 
the hands on the shoulders, occurs on a number of simian stone 
statues from Costa Rica, now in the United States National Museum. 
These statues, and others figured by Lothrop (1921) seem much 
closer to the Lake Yojoa stone figures than do the highly ornamented 
and definitely stylized Mayan bas reliefs or vase paintings. If the 
coincidence of hand position is not accidental in regard to the two 
types, it may well have some historic significance. The relationship 
of the Lake Yojoa stone carvings to the southern, " Chorotegan ", 
type is even more forcibly demonstrated by another anthropomorphic, 
cylindrical, stone carving from Los Naranjos (pi. 16, fig. i, and Yde, 
1936, fig. 6) . This type is identical with the rather common, anthropo- 


morphic, giant " roller pestles "of northeastern Honduras (Strong, 

1935. P- 148). 

With the anthropomorphic head and body, we also found the stone 
serpent head (pi. i6, fig. 2) referred to by Mrs. Stone (1934, p. 126) 
and Yde (1936, p. 29). By some strange mistake, Yde (1936, fig. 28) 
figures a side view of the " submissive figure " which he designates 
as the serpent's head. This piece is 80 centimeters long and 37 centi- 
meters wide across the broken base. It, too, is of hard gray volcanic 
stone. A short distance west of mound 4, we noted a cylindrical stone, 
95 centimeters long, smooth on one end and broken on the other, which 
apparently had once formed the base of the serpent head. It would be 
interesting to know whether the stone serpent on the " Islita ", men- 
tioned by Yde (1936, p. 30), was of the same type. He refers to a 
photograph of it in his article (1936, p, 30), but there is none. The 
style of this Los Naranjos serpent head carving (pi. 16, fig. 2) is 
very well executed and distinctive, but we cannot definitely place it. 
It would be extremicly interesting to know whether it pertains to the 
same period as do' the anthropomorphic statues. 

We did not see the various, undecorated, great stone slabs described 
and figured as " stelae " and " altars " by Mrs. Stone (1934, p. 127) 
and Yde (1936, p. 29, fig. 5). In the light of general distribution, 
however, we would be prone to relate these to similar erect stone 
slabs at Plan Grande in the Bay Islands and elsewhere on the main- 
land of northeastern Honduras (compare Strong, 1935, pi. 33 and 
pp. 160, 161) rather than to true hieroglyphic stelae of the Maya. 
As Yde (1936, p. 29) points out, the fiat rock with irregular carved 
grooves on its surface in the plaza of Los Naranjos is very similar 
to others occurring at Tenampua (compare D. H. Popenoe, 1936, pi. 5, 
fig. I ) . The adjacent fiat rock with depressions suggesting three shal- 
low bed-rock mortars seems more unique in this area. In a later report, 
it will be possible to publish adequate photographs of these interesting 
statues and carvings, but this cannot be done here. When the great 
site of Los Naranjos has been cleared, and excavations on a scale 
worthy of its size and importance have been commenced, more sta- 
tues will undoubtedly come to light. It should then be possible to 
correlate them with their exact cultural horizons and thus end the 
unsatisfactory speculation which must always center about disas- 
sociated art objects. 


Just north of mound i is a bushy field where a considerable amount 
of digging has been done in the last 3 years. We chose this place for 


work because Mr. Edwards reported deep, and possibly stratified, 
burials here, and M. K. Rittenhouse reported the finding of two pots 
of the old Playa de los Muertos type (pi. 15, a, b) amidst similar 
sherds at a depth of less than a meter. The surface of the field was 
irregular, but definite mounds were hard to find in the dense, low bush. 
However, ticks of all sorts were not. We selected and cleared a roughly 
circular mound, 18 meters wide from north to south, 21 meters wide, 
and 50 centimeters high, located about 100 meters north of the western 
end of mound i (map, fig. 20). East of the center line of the mound 
we dug a north-to-south trench 2^ meters wide and 12 meters long. 
The west wall of the trench was later extended 3.5 meters north, and 
two western side trenches were dug well beyond the center of the 

A small portion of the long western wall of this cut is shown (pi. 
16, fig. 5, and text fig. 31). A layer of dark, humous soil occurred 
just below the surface on the entire mound (fig. 31). Just below this, 
in the thick deposit of yellow-brown mixed soil we cross-sectioned the 
entire floor of a house (see house floor, fig. 31) composed of black, 
burned soil containing many sherds, metate fragments, and refuse. 
Beyond the edge of the diagram here shown (fig. 31), this occupation 
level or floor dipped, forming a level area for about 5 meters, then 
rose to the ground level, extending on into what appeared to be another 
floor or occupation level beyond the limits of the excavation. The 
same type of occupation level also occurred at the surface on the 
south end of our main trench. Our western cross trenches showed 
that the central floor area extended 2 meters to the west, where it 
again rose to the surface. No post holes occurred in our cross-section 
of the central house, but one was found extending below the occupa- 
tion level at the north end of the central trench. No special fireplaces 
were noted, but charcoal was abundant. Judging from our trenches, 
there are numerous house floors in this vicinity, on or just below the 
present level of the ground. These contain the finest Yojoa poly- 
chrome and associated cooking ware sherds, along Avith other refuse. 
Here, as at Naco, an expedition engaged in other than exploratory, 
stratigraphic work, could easily clear entire house floors and work 
out the features in detail. During the first part of our work the 
trenches were taken down below the occupation area into the sterile, 
yellow clay and gravel stratum (fig. 31) which occurred at an average 
depth of about 1.3 meters below the surface. Polychrome sherds, 
stones, charcoal, burned clay, fragments of pumice, and broken arti- 
facts occurred throughout the yellow-brown soil. The mixed soil level 
became darker just above the sterile layer (pi. 16, fig. 5, and text fig. 

NO. I 




31). Below the house floors, or occupation levels, we encountered 
several burials (see P 2, fig. 31) of the polychrome period. For the 
present we will confine ourselves to a brief discussion of features and 
artifacts from this upper or polychrome level, later discussing the 
materials below the sterile stratum. All vessels and important arti- 
facts were photographed hi situ, but these, like our complete cross- 
sections and ground plans, cannot be presented here. 

At the extreme south end of the main trench, at a depth of 94 
centimeters, we found a small orange bowl decorated with crude red 
alligators and black scrolls. It was tipped on one side. No bones 
were present. A deposit of three vessels occurred near the north 
end of the main trench at a depth of i meter. There were two super- 
imposed bowls, one with a cream-white slip on which were three dark 
red and orange designs of Mayoid type, probably conventionalized 
serpents' heads ; the other was orange with a much conventionalized, 
seated Mayoid figure. These two bowls were inverted. Next to them 
was an upright, small, but striking, efligy bowl, modeled to represent 
a frog. It was brown and unslipped. This was probably a grave, 
occurring at the base of a refuse heap. However, no bone was found. 
About 2.50 meters from the south end of the trench near the west 
wall, at a depth of 1.45 meters, we found an interesting upright bowl 
(pi. 14, d). This was small but of the typical Santa Rita Bold Geo- 
metric olla shape and color (compare pi. 14, d, and pi. 7, a). The 
present vessel has the same light orange slip, with similar red semi- 
circles inside the lips as do the monkey-handled ollas, but the geo- 
metric and conventionalized designs are in dull red with no black. 
The base has the marked dimple, and the handles have the lugs of the 
Ulua Bold Geometric olla type. 

Slightly north of this pot, at the junction of one of the western 
cross trenches, five vessels were uncovered at depths of from i to 
1.45 meters. This immediate area had been badly disturbed by arma- 
dillo burrowings (see grave P 2, fig. 30, and pi. 16, fig. 5), but all 
the vessels were evidently part of one grave ofifering. In between them 
were found a few small, crumbling fragments of human bone and 
three caps from human molar teeth, thus proving for the first time 
the presence of a burial. As can be seen from the photograph (pi. 
16, fig, 5) and diagram (fig. 30) this burial occurred under the 
southern edge of a house floor. Only one vessel shows in the cross- 
section diagram, the others being just east of the trench wall. This 
bowl was inverted. It was fluted and had a splotchy, light orange 
slip, with crude red linear and geometric designs. On the inside a 
slightly darker orange wash has been added, leaving thin vertical 


stripes of the lighter slip, thus suggesting negative painting. The bowl 
is blackened by fire on one side of the bottom, suggesting that, despite 
its thinness and fine, hard paste, it might have been used for cooking. 
Just east of this vessel, at a depth of i meter, was a crumbly red bowl 
containing a fragile little tripod vase with black and red decorations. 
Both these vessels, despite our greatest care, crumbled into tiny frag- 
ments when removed. At a depth of 1.45 meters, 30 centimeters to the 
south of these, was an upright incensario containing a considerable 
amount of charcoal. It had nine perforations and a solid, rectangular 
handle which had been completely hollowed out from the top. The 
incensario had a dirty, cream-colored slip, with both upper and lower 
edges outlined in red. The fifth bowl (pi. 13, b) was slightly to the 
north, at a depth of 1.45 meters, in an upright position. It has an 
orange slip, with a band of white below the rim and three white 
bands down the sides, dividing the outer surfaces into three panels 
(pi. 13, b). On the white bands are unusual geometric and curvilinear 
designs in dark red and, on the sides, orange. In each of the three 
panels occurs a most interesting prancing monkey, done in dark red. 
The bottom of the bowl is flat. 

In the south wall of the southerly, east to west extension trench, 
at a depth of 64 centimeters, was an upright, two-handled, cooking 
pot, 22 centimeters high and of a dull yellow color. Inside this large 
vessel was an inverted polychrome bowl with a yellow slip, a row of 
red conventionalized Mayoid designs outside the lip, and three big 
black circles on the sides. The designs were badly eroded, but the 
bowl was intact. The outer vessel barely held together while being 
uncovered and photographed, but the moment we touched it, to re- 
move it, it fell into over a hundred small pieces. Close to these two 
vessels was a dull cylindrical stone bead. On the north wall of the 
other extension trench, 1.25 meters deep, occurred a small open bowl 
of rough gray unslipped ware. At a depth of about i meter, near 
the west wall of the main trench, we found restorable fragments of 
a vertical walled vase with solid, rectangular tripod feet. It has a 
rich orange slip, divided into three parts on the sides by dark red 
and black linear designs. Red and black lines circle top and bottom, 
and each panel contains a well-executed seated Mayoid figure, with 
elaborate headdress, bustle, and outstretched hand, done with fine 
lines. Later, when this site was reopened to dig through the sterile 
layer searching for deeper cultural material, a small " salad bowl " 
type vessel (pi. 14, g) with an annular base, was found nearby right 
side up at a depth of meters. This bowl is interesting because 
of its shape and because of the darker orange wash through which 


horizontal and vertical lines of the lighter underslip stand out. More 
complicated curvilinear designs of this type occur on the inside of 
the bowl. This rather peculiar type of negative painting is well shown 
in the photograph (pi. 14, g). 

The foregoing account, in conjunction with that of our excavations 
at La Ceiba, gives an idea of the manner in which vessels representing 
the various types of polychrome ware occur in the smaller Lake Yojoa 
mounds. Later, in connection with a site at El Eden, we will discuss 
the present slender evidence regarding the apparent vertical distribu- 
tion of Yojoa polychrome pottery types in these relatively shallow 
sites. Although traces of human bones occurred with only one of 
our burials, there seems good reason to believe that the majority of 
these pottery caches were once with skeletons, all traces of which have 
now disappeared. It is further indicated that these low mounds also 
served as places of habitation during the polychrome period, and that 
burials occurred beneath the house floors. Probably, as at La Ceiba 
and Aguacate, many of these mounds were used, or came to be used, 
almost entirely for burial purposes, and it is in these that the great 
masses of rocks occur. Others, like the site we are discussing, served 
primarily for habitation, but burials also took place under and near 
the houses. Such habitation mounds seem to have relatively few large 
rocks. There remains briefly to sketch in the rest of the artifact con- 
tent of the polychrome horizon at this site, and then to describe the 
occurrence of a deeper, older, cultural horizon which was encountered 
at the very end of our stay at Jaral. 

The sherds from this one Los Naranjos mound site present a wide 
variety of Yojoa polychrome types. In addition to those already men- 
tioned among the entire vessels are Mayoid pieces with incising as 
well as painting ; Bold Animalistic sherds ; heavy Bold Geometric 
sherds ; polished brown carved fragments ; heavy gray or buff sherds 
painted only on the flat upper surface with bright black, red, and 
orange designs (one of these is flat with a small annular base) ; un- 
slipped brown grater fragments ; and two cylindrical spouts of coarse 
brown pottery. The latter may have been carried in by the natives 
from older deposits since we have seen no Lake Yojoa polychrome 
vessels with this type of spout. One candelario fragment of coarse 
brown ware has three compartments and simple incised designs. There 
are two spindle whorls, one of plain brown ware, the other a ground- 
down, painted sherd. All the above come from depths ranging from 
the surface to 1.45 meters in depth. A complete Mayoid figure form- 
ing a whistle comes from a depth of 65 centimeters, and a bird whistle 
from 1. 10 meters. In addition, there are numerous fragments includ- 


ing broken but ornate, hollow, Mayoid effigies of fairly large size. 
Figurines are varied. Solid and hollow figurines with square Mayoid 
headdresses (like fig. 7, h, i, s; also see Gordon, 1898, pi. 9, /, n, v, 
s) , as well as solid heads with pouting faces and simpler hair dresses, 
all come from depths of i to 1.45 meters. This latter type of simple, 
well modeled, solid figurine also occurs in polychrome deposits on 
the Ulua. Several of the Los Naranjos figurines are extremely crude, 
solid lumps of baked clay with grotesque, punctate faces or filleted 
" cofifee-bean " eyes. These have a decided "Archaic " appearance but 
occur in the same horizon with the polychrome pottery and ornate 
figurines. In addition to the figurines this deposit yielded a consider- 
able number of filleted or modeled fragments of baked clay. Many 
of these are quite complex but their original form is uncertain. Ground 
stone artifacts are fairly abundant, including flat ovoid lapstones ; both 
flat and tripod rectangular metates (the majority with a broad grind- 
ing groove) ; cylindrical roller pestles (including some that taper at 
both ends) ; hammerstones ; one small rock mortar ; two small, sharp, 
jadeite celts ( m deep) ; and one brown stone bead. Chipped 
stone artifacts are simple but relatively abundant. Large and small, 
fragmentary, prismatic flakes of obsidian occur. There are numbers 
of crude obsidian flakes, evidently used as scrapers, and a few flakes 
of other stone. At a depth of meters we found the only definite 
projectile point encountered, a planoconvex, obsidian dart point with 
a slight, tapering stem. Even this brief summary indicates that these 
small Lake Yojoa sites are far more prolific in nonceramic artifacts 
than are sites on the Ulua or the Chamelecon Rivers. 

The Older Horizon at Los Naranjos 

Our main work at Lake Yojoa was terminated by the advent of 
" Holy Week." It was then necessary to move on to Playa de los 
Muertos on the Ulua, and then to Naco if we were to complete the 
survey we had outlined. We had determined the general association 
of polychrome wares at Lake Yojoa but had not found any marked 
stratification of cultures, nor had we encountered the old type of Playa 
de los Muertos ware (pi. 15, a, h) discovered in the vicinity of Los 
Naranjos by Mr. Rittenhouse. Despite his statement that it had been 
found here at depths of less than a meter, we had so far been unsuc- 
cessful in locating any remains other than those of the polychrome 
period. It was obviously necessary to go deeper and penetrate below 
the sterile yellow clay stratum encountered at our Los Naranjos site. 

Therefore, in May, when the rest of the expedition went to Naco, 
Mr. Paul returned to Jaral for this purpose. Efforts to locate the 


exact Rittenhouse site through his former workmen again proved un- 
successful. Mr. Paul therefore sank a small test trench (B) 6 meters 
southwest of the mound we had excavated (fig. 20). Passing through 
the polychrome horizon, which was removed in 30-centimeter levels, 
he again encountered the sterile layer of yellow clay and gravel which 
here averaged about 50 centimeters in thickness. Digging through this, 
he encountered a brownish black clay which contained a small amount 
of cultural admixture. All potsherds from this lower occupation level 
proved to be of a crude, monochrome type. This lower cultural hori- 
zon averaged about 65 centimeters in thickness, dipping toward the 
east end of the trench, and terminating in a very hard yellow clay 
which appeared to be absolutely sterile. To check these results he 
dug another test pit (A) 8 meters to the southeast of the same mound. 
Here he again encountered the same soil and cultural conditions (fig. 
32), the brownish black clay below the sterile clay yielding only coarse 
monochrome potsherds and a few other artifacts. He then returned 
to our former excavation and sank a pit next to our old test cut into 
the sterile stratum (fig. 31). Only 25 centimeters below the lowest 
level of our former excavation he ran through this sterile layer into 
the darker clay, obtaining monochrome potsherds and a small mano. 
It is evident, therefore, that this direct superimposition of two cul- 
tural horizons, separated by a sterile stratum of yellow clay and gravel, 
extends over a considerable area. The same strata vary slightly in 
thickness and absolute level at the different pits (compare figs. 31 and 
32), but the sequence is the same in all. Material from the lower 
cultural horizon is likewise uniform and may be discussed as a unit. 
The deeper ceramic remains, some 700 sherds, are extremely crude 
(pi. 15, c-w). They are all of a crumbling ware, tempered with finely 
ground stone or sand. The apparent similarity in texture between 
this older ware and the poorer grade of Yojoa Polychrome, especially 
the cooking ware, suggests that both were made of inferior, local 
clays. This point may be determined later by microscopic analysis. 
All sherds from the old horizon seem to have come from small vessels. 
The thickest sherd is 1.4 centimeters, the thinnest .4 centimeter, and 
the majority average about .7 centimeter. Some are badly waterworn 
(pi. 15, r), and the majority have one or both faces considerably 
eroded. Of the 51 rim sherds, the great majority have low, slightly 
flaring lips (pi. 15, c, d, f,g). A small proportion of lips are swollen, 
and there are a few vertical and a few direct rims. Two sherds from 
the same vessel, seem to be parts of an annular base, but the remain- 
ing 30 basal sherds are all from small, flat-bottomed vessels (pi. 1$ s, 
u, V, w). There are no spouts, handles, lugs, or feet in the present 

NO. I 



sample. Only 12 sherds show definite traces of slip or paint. The 
Others range in color from a dull buff, through dull red, to a grayish 
black. Despite the obvious erosion on many sherds the majority do 
not appear to have ever been slipped or painted, though we cannot 
be positive of this. The painted sherds include eight that have a faded 


Polychrome y. 
Layer 1 ya 

Black Humus Soil 

a J^ 

% >^ % 

X X 

f. '*' ^ %. Dark Brown^Clay > 



•■ i 

K ^ — . 

. • S+eri le ■. Ye 1 1 ow . Clay^nd .G fa ve I'. •* ; 

0- • • . • • -^fFi re place \^'*1^P^ 
•••••• ••• • ■ -Nji ^ir* Kt ^x //. 

xj- x 

^ Brown - Black Clay Mixture" "» 
X * * X « 

X X 


X X 


liysl - Sterile Areas 
X - Approximate location of potsherds 

I meter 

Fig. 32. — Cross-section of excavation A, near site i, Los Naranjos, showing 
stratified cultural horizons. 

red or pinkish slip (pi. 15, /) ; two with a dull white slip or wash; 
and two that have definite areas painted a very dull red and black on 
the inner surface. Other decorative efforts are scant. A few sherds 
have raised ridges below the rim (pi. 15, i, I), one has such a ridge 
with regular indentations, and a few sherds have simple, linear de- 
signs incised on the outer rim or body. If the present sample is 
at all adequate, this, the oldest known Lake Yojoa pottery, appears 


to be the most primitive ceramic type yet encountered in Honduras, 
Technically, since a few sherds are painted, we should designate this 
ware as Yojoa bichrome. Actually, the great majority of sherds are 
unpainted and all of them are definitely inferior in both texture and 
finish to either the Playa de los Muertos Bichrome or the Santa Rita 
(farm 17) Bichrome wares. For this reason we have tentatively 
designated it as Yojoa " Monochrome ", subject to change when the 
results of larger excavations yielding adequate ceramic samples are 
at hand. 

Equally puzzling is the relationship of this Yojoa " Monochrome " 
to the two vessels excavated by Mr. Rittenhouse in this immediate 
vicinity. Both these vessels are well modeled, incised, and painted red 
and buff (pi. 15, a, b). Both originally had spouts. They are un- 
doubtedly closely allied, if not identical, with the old Playa de los 
Muertos Bichrome (pis. 10, 11, and figs. 17, 18). Yet not one of 
our sherds from the deep stratum is positively of this type. We have 
no reason to doubt that the general location of the two Rittenhouse 
vessels was substantially as reported. It is apparent, therefore, that 
our Yojoa Monochrome is either a strangely isolated sample of crude, 
domestic ware, actually pertaining to the Playa de los Muertos Bi- 
chrome, or else that both Yojoa Polychrome, Playa de los Muertos 
Bichrome, and a new type, Yojoa Monochrome, all occur in shallow 
deposits on the northern borders of Los Naranjos. 

Other artifacts from the older horizon at Los Naranjos included 
three figurine fragments (pi. 15, e, j). All are of solid, baked clay, 
and none are slipped. The crudely modeled little head (pi. 15, e) 
has a knot of hair on the back, and the body (pi. 15, ;') has broad 
modeled and grooved buttocks which have been smoked black. Head 
and body are from different figurines. The third figurine torso is 
also of coarse, dull buff pottery. It is somewhat similar to the old 
Playa de los Muertos horizon figurine torsos (pi. 11, t-v), but is not 
so well modeled and has no breasts. Ground stone artifacts from this 
older Yojoa horizon include one small rectangular mano with ground 
sides and battered ends; one small rectangular stone (pi. 15, n) of 
unknown use; and a fragment of ground sandstone (pi. 15, t) which 
may be from a simple metate, although it has uneven grinding surfaces 
on the two sides. There are no true prismatic flake fragments from 
this horizon. There are, however, several irregular flakes of obsidian 
(pi. 15, m), and one irregular prismatic flake with a side point which 
also shows a use retouch. There is one rather large, gray flint side 
scraper and a flint flake. Here again, definite conclusions are precluded 


by the small size of the present sample, but both the figurine and 
artifact fragments, like the pottery, show unique types. 

There can be no doubt that, just below the elaborate polychrome 
horizon at Los Naranjos, there occurs another cultural level which 
appears to be surprisingly primitive. When it is remembered that 
our deepest excavations at Los Naranjos were slightly less than 3 
meters, it can be seen that here is an area where deep excavations may 
yet furnish evidence regarding the truly simple cultures of Central 
America. On the Ulua, where we conducted our largest and deepest 
excavations, we were eventually stopped by reaching the water level. 
At Lake Yojoa this was not the case. A larger expedition, with 
adequate time and equipment, providing it is not led too far astray 
by the richer polychrome deposits, should be able to work out a most 
important sequence of human occupation in this immediate region. 


Seeking for a deep Yojoa polychrome refuse heap suitable for 
stratigraphic analysis, we conducted a small excavation about i kilo- 
meter northeast of El Eden (see site 2, near that village, map, fig. 
20) . Miguel had brought us a number of polychrome sherds from this 
place, his sample including a dark brown and highly polished tripod 
bowl fragment with delicate geometric incisions on the body, an in- 
censario fragment with rather elaborate geometric painted designs, 
and a small whistle shaped like a turtle. This sherd deposit was located 
in the abrupt face of a steep bank terminating a small, densely wooded 
arroyo. This arroyo led down toward one of the deep sink-holes which 
here mark the course of the Rio Blanco. The region is a maze of small, 
abrupt canyons or sink-holes, and is covered by unbelievably dense 

We dug a trench, 3.5 meters long and i meter wide, along the face 
of the bank, encountering our first potsherd at a depth of 40 centi- 
meters. From this point down all artifacts were segregated according 
to horizontal levels averaging 30 centimeters in depth. Unfortunately, 
only three levels were encountered when we ran into sterile yellow 
clay. The exposed surface of the bank below this point appeared 
devoid of any human detritus. The upper 30-centimeter level con- 
tained a number of well-executed fragments of polychrome ware 
with highly conventionalized Mayoid designs. A tripod leg proved 
to belong to the dark brown and incised dish fragment secured earlier 
by Miguel. A number of heavier polychrome sherds had geometric 
designs in red and black, suggesting the Bold Geometric Ulua type, 


but no definite Bold Animalistic sherds occurred. Crudely painted 
and occasionally incised sherds of domestic ware were fairly abundant 
in all levels. The middle layer yielded one conventionalized Mayoid 
sherd similar to the above, and a considerable number of red and black 
or orange sherds with designs suggesting the Bold Geometric. Defin- 
itive Bold Animalistic designs were lacking. The bottom level was 
similar but lacked both definitive Mayoid or Bold Animalistic designs, 
although several badly eroded polychrome sherds may have been of 
these types. To sum up, the El Eden polychrome site proved negative 
so far as any obvious stratification of ceramic types was concerned. 
The absence of definite Bold Animalistic type sherds is interesting 
but hardly significant, owing to the small sherd sample. 

At excavation B, Los Naranjos (near site i, map, fig. 20), a simi- 
lar stratigraphic excavation was made. Here, again, only three 30- 
centimeter levels of polychrome sherds were obtained, the depth of 
the upper Yojoa Polychrome deposit being similar to excavation A 
(fig. 32). The top level contained rim fragments of small bowls, 
many of which had thickened lips. These sherds have conventional 
and rather massive red and black designs on orange and, in one case, 
white, backgrounds. Three basal fragments, one flat, one dimpled, 
and one annular, occurred. The latter is a dark brown, almost black, 
overfired piece. The middle level contained sherds with similar, con- 
ventionalized Mayoid designs, and a few with well-executed and iso- 
lated serpent motifs. A few sherds from this level are of the Bold 
Animalistic type. Basal fragments include two flat and two dimpled 
bottoms. The lowest sherd level included a bowl fragment with an 
elaborate open-winged bat (like pi. 3, b; Gordon, 1898). Several 
rim sherds from vertical-walled Mayoid vases have rows of typical 
Ulua conventionalized faces, and a fragment of a tripod plate has 
a similar design motif. In addition, there are a number of dull orange 
sherds with more conventionalized black and red geometric designs. 
From this lowest level comes a splendid Mayoid vase (pi. 12, a) 
encountered in a broken condition at a depth of 1.25 meters in the 
original excavation B test pit. This vase, with a definite rim, marked 
entasis, and a flat bottom, has an orange slip with complex anthropo- 
morphic and glyph designs in brownish yellow, purplish red, and 
black. Thus, although the evidence is slender, there is some sugges- 
tion that the Lake Yojoa Polychrome wares exhibit the same trend 
from the more realistic to the conventional as was true of Ulua 
Polychrome pottery decoration. The occurrence at La Ceiba of both 
an extremely conventionalized Bold Animalistic vase (pi. 13, a) 
at a depth of only 15 centimeters, and a splendid, realistic Mayoid 


vase (pi. i) at a depth of 1.25 meters furthers this possibility. That 
both the more realistic and the more conventional aspects of the 
Mayoid and the Bold Animalistic tradition occur in the shallow Lake 
Yojoa Polychrome sites is certain. Their exact interrelationship, how- 
ever, remains to be demonstrated. 

Two other sites on the north end of Lake Yojoa may be briefly 
mentioned. The first of these is a little island, called merely. La 
Islita. It is near the shore between Jaral and Agua Azul (see map, 
fig, i). Yde (1936, p. 30) describes a stone serpent head from this 
place which, subsequent to their visit, was reported to have been 
smashed by natives looking for treasure. He refers to, but does not 
reproduce, a photograph of this statue. Our guide brazenly showed 
us a simple cylindrical statue, apparently anthropomorphic and about 
I meter tall, the head and face of which had been completely smashed 
by him in a futile search for treasure! Only a carved ear remained 
and the rounded pediment which was like the simplest statue at Los 
Naranjos (pi. 16, i). He did not know of the stone serpent head 
but claimed a similar anthropomorphic statue had been taken from 
the island to Tegucigalpa. Another man told us of a stone serpent 
head that had been found on this island, but said that local people 
had thrown it in the lake ! The island is very steep and densely wooded. 
On top of one of the hills are a number of low, irregular mounds, 
some of which are covered with rock slabs. Our guide had dug pits 
in several of these and claimed to have found a few pieces of painted 
pottery. The soil of the mounds is a red clay. Aside from a few 
coarse brown sherds we saw no pottery at the site. There are said 
to be other mounds of a similar nature on the island, which we did 
not see. It is sincerely to be hoped that this interesting site may be 
scientifically worked by archeologists before it is completely ruined. 

The other site is a group of three impressive mounds located in 
the open pine and savannah country about 2 kilometers northwest 
of the ranch house at Agua Azul (see map, fig. i ) . The largest mound, 
to the north, is conical with a flattened top. It is approximately 7 
meters high by 9 meters across, and is flanked to the west and south 
by a terrace edged with straight walls of large boulders about i meter 
in height. The west wall is about 5 meters from the edge of the mound 
and the south wall about 6 meters from it. The south wall is com- 
posed of several thicknesses of stone and terminates just east of the 
center of the mound. The west wall is only one stone thick and termi- 
nates just beyond the north edge of the mound. About 10 meters 
south of the edge of this terrace are two more parallel mounds. The 
mound to the east has an approximate length of 7 meters, and a 


height of 5 meters. The mound to the west is smaller and steeper, 
being about 4 meters long and perhaps 3 meters high. Local men from 
Siguatepeque have excavated a small hole in the terrace south of 
the big mound and a large trench, 5 meters wide and nearly 6 meters 
deep, on the west side of the same mound. The earth wall of this 
trench shows successive curving layers of black charcoal, suggesting 
that the mound had been built up at different times and the remains 
of fires on the top had been scattered down the sides. The small pit 
on the terrace showed nothing. We were unable to find any potsherds, 
either in the cut or on the surface. According to local report the origi- 
nal diggers encountered nothing but a very little broken pottery. This 
is a striking mound group and, as already mentioned, seems similar 
in some ways to the " cut stone mound " which is buried in the dense 
bush near La Ceiba (site 3, map, fig. 20). 

This concludes the list of sites visited by us around the north end 
of Lake Yojoa. We have since heard that local pot hunters have 
opened up a new series of ancient cemeteries between La Ceiba and 
Agua Azul. Other sites are reported in the mountains to the north 
(see map, fig. 20), at Sauce, and elsewhere around the lake, but we 
lacked time to visit these. 


The present reconnaisance of the Ulua-Yojoa region opens promis- 
ing vistas. It reveals incomplete but considerable sequences of local 
development, and it demonstrates that the interplay of northern and 
southern cultural forces, so strongly suggested by linguistic, eth- 
nographic, and historic sources, is very definitely reflected in the arche- 
ological record. 

Since ceramic remains constitute the most abundant and helpful 
guides in attaining any understanding of the development of the pre- 
historic cultures of northwestern Honduras, we may preface our brief 
summary by a table showing the sequence and groupings of Ulua- 
Yojoa ceramic types as known at present (table i). Of these the 
Naco Polychrome is definitely historic and represents, apparently, the 
late Nahuatl occupation of the region. Spinden, Tozzer, Mason, and 
Vaillant, who have examined this material, state that it appears to be 
related to certain late prehistoric wares of Mexico. Naco Polychrome 
pottery will probably be found at other sites occupied or influenced 
by these intrusive Nahuatl peoples. It may occur at Tenampua (com- 
pare Popenoe, 1936, p. 572 and fig. 2), In the same way that the 
occurrence of Spanish crockery in association with Naco Polychrome 
sherds connects the site with the historic period, so the occurrence of 

NO. I 



simply decorated Ulua Polychrome sherds (pi. 3, &) in Naco refuse 
mounds indicates that other, local cultures were contemporaneous in 
the region. Little attention has as yet been paid to the historic and late 
prehistoric cultures of the Jicaque and other local inhabitants of 
northwestern Honduras. 

Table i. — Apparent Sequence of Ceramic Types in Northzvestern Honduras. 
Ulua Yojoa 


Naco Polychrome 


(surface mounds?) 


Ulua Polychrome (including) 

Yojoa Polychrome (including) 





Mayoids _ 


Mayoid Bold Animalistic 


- Lower 


Ulua Bichrome (Santa Rita) 


Playa de los Muertos Bichrome 


Yojoa " Monochrome " 


(Los Naranjos) 



(or here?) 

Table 2. — Probable Correlation betzuecn present Ulna Polychrome Classification 
(Table i) and those of Gordon (189S) and Vaillant {1927) 

Ulua Polychrome 

Upper Mayoid 

Upper Bold 

Ulua Polychrome 

Lower Mayoid 

Lower Bold 

Gordon's B 
Gordon's C 

Gordon's A 
Gordon's C 

Vaillant's III and IV 
Vaillant's V 

Vaillant's I and II 
Vaillant's V 

The prehistoric polychrome wares of the Ulua have been classified 
on typological grounds by Gordon (1898), and by Vaillant (1927). 
In general their classifications seem to accord with our stratigraphic 
findings as above (table 2). Gordon clearly distinguished the Bold 
Geometric as Type C, and the Mayoid as Types A and B, but has 
nothing to say regarding sequence. Vaillant makes a mistake when 
he assumes, on stylistic grounds, that the Bold Geometric (Ulua Poly- 
chrome V) developed out of the Mayoid style and was therefore later. 
Strong makes the same mistake in regard to the related Bay Island 


Polychrome I (Upper Mayoid, plus southern influences) and the iden- 
tical Bay Island Polychrome II (Upper Bold Geometric) (1935, 
p. 145). Actually the Mayoid and the Bold Geometric have been 
shown to be parallel developments ; thus Vaillant's Ulua Polychrome 
V is contemporaneous with his Ulua Polychrome I and II (his 
Ulua Polychrome IV contains both Upper Mayoid and Upper Bold 
Geometric constituents), and Strong's Bay Island Polychrome I 
is, in all probability, contemporaneous with his Bay Island Polychrome 
II. Vaillant (1927, p. 266) was careful to point out the entirely tenta- 
tive nature of his assumed sequences. Further, in regard to the Ulua 
Polychrome V (and Salvador Polychrome VI) he states: "There is 
a strong suspicion of the same non-Maya factors influencing both 
these styles. The source of the influence is not discoverable in Maya 
districts, and one thinks vaguely of the south and east, of Nicaragua, 
eastern Honduras, and Costa Rica to locate a source " (1927, p. 170). 

Recently Tschopik (1937), in a brief but valuable analysis of tex- 
tile motifs on Gordon's Ulua Polychrome pottery, has independently 
pointed out this stylistic dichotomy. He groups Vaillant's Ulua Poly- 
chrome I-IV as Ware A [Mayoid], and the latter's Ulua Poly- 
chrome V as Ware B [Bold Geometric]. Tschopik points out that 
there are consistent differences in both form and decoration between 
the two, and that A is Mayoid, whereas B has a definite relationship 
in form and decoration with ceramic types from Salvador, Nicaragua, 
and Costa Rica. He, too, repeats the theory that naturalistic designs 
are apt to be earlier than geometric, suggesting that Ware A is earlier 
than Ware B, thus falling into the same error as Vaillant and Strong. 

At Santa Rita these two major styles (Mayoid and Bold Geometric) 
are intermixed throughout almost 4 meters of Ulua Polychrome de- 
posits. Although they blend in certain intermediate types of vessels, 
each style in general keeps to its own particular genius, and each shows 
a parallel development from a finer and somewhat more realistic dec- 
oration in the lower levels, to a more conventionalized and geometric 
decoration in the upper levels. Thus, the Lower Mayoid has priestly, 
processional, and " dancing " figures in open panels, whereas the Up- 
per Mayoid has florid, conventionalized, over-all designs, geometric 
motifs and, often, animal head lugs. The Lower Bold Geometric has 
intricate linear and geometric designs with remarkable, cursive ani- 
mals or birds in open panels, whereas the Upper Bold Geometric be- 
comes simpler, drops the animals, but retains textile and geometric 
designs. At Las Flores, also, both the Mayoid and the Bold Geometric 
styles occur in the same excavation, but here both are of the upper 
and later, conventionalized type. It is worth noting that the only 


metal object recovered in any of our excavations, a small copper fish- 
hook, came from these levels at Las Flores. The Ulua Polychrome 
horizon overlying the deep stratum at Playa de los Muertos is also 
late. Here only the Upper Mayoid and Upper Bold Geometric occur. 
It is undoubtedly significant that the typical, swollen, monkey-handled 
olla form of the Bold Geometric tradition (pi. 7, a-d), and the vertical- 
v/alled vase form of the Mayoid tradition (pi. 8, a-b), both persist 
practically unchanged throughout the entire Ulua Polychrome series. 
This occurrence argues rather strongly against any very considerable 
time period being assigned to the Ulua Polychrome period. 

The polychrome wares of Lake Yojoa are closely related to those 
of the Ulua. Not only does Yojoa Polychrome ware contain a large 
number of forms and motifs identical with those of the Ulua Poly- 
chrome, but it also manifests a very similar division into two major 
stylistic traditions. In general, however, Ulua and Yojoa Polychrome 
ware vessels are distinguishable. The Yojoa Mayoid type, as well 
as the Bold Animalistic type, finds many close parallels in polychrome 
vessels from eastern Salvador (see Vaillant, 1927, figs. 35-40). It 
seems strange that no Plumbate ware whatsoever was recovered in any 
of our excavations, either on the Ulua or at Lake Yojoa. The Bold 
Animalistic type from Lake Yojoa differs from the Ulua Bold Geo- 
metric in the relative rarity of monkey-handled ollas and the prevalence 
of bird, monkey, alligator, and other animal design motifs. Regard- 
ing the internal development of Yojoa Polychrome ware decoration, 
there is some very slight evidence that it parallels the trend of the Ulua 
Polychrome series from better executed realistic, to conventional 
and geometric design. However, the i^ meters of Yojoa Polychrome 
deposits so far investigated have not as yet yielded very satisfactory 
evidence in this regard. The fact that Ulua Polychrome deposits oc- 
cur throughout 3 to 4 meters of alluvial and cultural deposition, 
whereas the known Yojoa Polychrome refuse deposits are less than 
2 meters in depth, is undoubtedly significant. In our opinion, how- 
ever, this discrepancy is probably due to the very different physio- 
graphic conditions in the two regions, rather than to differences in 

Of the three wares that have been stratigraphically established as 
earlier than the Ulua-Yojoa Polychrome series, the Playa de los 
Muertos Bichrome (table i, and pis. 10, 11) is the most clearly 
defined. This is the type D of Gordon (1898). Vaillant (1934) has 
pointed out that this horizon contains a majority of traits, mainly 
ceramic, that are characteristic of the O complex. It is undoubtedly 
significant that, whereas Playa de los Muertos Bichrome ceramics 


represent an advanced pottery type so far as texture, surface finish, 
modeling, and incising are concerned, they appear to mark an experi- 
mental and inept stage in the use of surface painting. Especially char- 
acteristic of this horizon are highly polished, modeled and spouted 
forms ; flat-bottomed, vertical-walled vases ; low dishes with flaring 
incised walls or everted, flat, incised lips or both ; and solid female 
figurines, which may or may not have a white slip. There is con- 
siderable resemblance between the simple but effective modeling of 
these figurines (pi. ii, t, u, v) and the stone statue of a man or 
ape at Los Naranjos, Lake Yojoa (pi. i6, fig. 3). These traits, 
plus the occurrence of jadeite artifacts and the varied experiments 
with painted decoration, all indicate that here was an early and potent 
cultural manifestation of more than local significance. In so far as 
data are available (R. E. Smith, 1936, and Uaxactun sample sherd 
collections), we see considerable resemblance between this Playa de 
los Muertos Bichrome ware and the two earliest stratigraphic periods 
at the old Maya city of Uaxactun. These have been termed Mamon 
and Chicanel, and both precede the Maya Polychrome period. 

The determination of the northern and the southern extent of the 
Playa de los Muertos horizon is one of the important problems in 
Middle American archeology. Even more important is the determina- 
tion of the simpler ceramic horizons from which it developed. In 
Honduras we have as yet no clues to this earlier period unless the 
so-called Yojoa "Monochrome" (pi. 15, c-w) is as truly primitive 
as it superficially appears, and can be demonstrated as stratigraphically 
earlier than the developed Playa de los Muertos culture. The later 
break, between the Playa de los Muertos Bichrome and the Ulua 
Polychrome, is in part bridged by the deepest cultural horizon at 
Santa Rita containing Ulua Bichrome ware. The most outstanding 
feature of the Ulua Bichrome ceramics is the presence of Usulatan 
ware sherds. According to Lothrop this is " the earliest painted 
pottery now known from Central America ", and, although it occurs 
occasionally in the form of trade pieces at Old Empire Maya sites, 
it seems to center in Lenca territory in eastern Salvador (1933, pp. 
47-51). There is rather close resemblance between our Usulatan 
sherds with short, solid legs (pi. 9) and the early Chukumuk pottery 
from Lake Atitlan in the highlands of Guatemala (Lothrop, 1933, 
p. 49). Similarly, the tetrapod Usulatan bowl recovered by Gordon at 
a depth of " 26 feet " in his Playa de los Muertos excavations is of 
an identical type. Thus, there is a linkage in this regard between the 
deep horizons at Playa de los Muertos and at Santa Rita, despite the 
fact that our own sample of Playa de los Muertos Bichrome ceramics 


contains no definite Usulatan ware. In addition, this clear linkage 
between early Ultia and early Guatemalan highland cultural horizons 
is of great interest. We have assumed that Ulua Bichrome is some- 
what later than the Playa de los Muertos Bichrome on stylistic grounds 
and because the sterile area separating the former from the Ulua 
Polychrome is thin compared to that separating the Playa de los 
Muertos Bichrome from the overlying Ulua Polychrome (compare 
fig. 6, and fig. 16) . This, however, is at best a dubious procedure, since 
we do not as yet know the physiographic nature of either sterile 
stratum. Moreover, it must be remembered that only the Upper May- 
oid and Bold Geometric Ulua Polychrome types occur in the over- 
lying cultural stratum at Playa de los Muertos, whereas both these 
and the earlier Lower Ulua Polychrome wares occur above the Ulua 
Bichrome at Santa Rita. These details, like the cultural and temporal 
placing of the puzzling Yojoa " Monochrome " ceramics and the 
" Chorotegan " stone statues at Los Naranjos, must await further 

Tracing the relationship of the native cultures of northwestern 
Honduras backward from the known historic, we have already veri- 
fied the presence of a late Nahuatl migration from Mexico through 
the finds made at Naco. Similarly, in the Ulua Polychrome period 
we find two interlocked but distinct styles occurring in the same sites, 
the Mayoid and the Bold Geometric, which at Santa Rita persist and 
develop simultaneously over a considerable period. Lake Yojoa Poly- 
chrome is also composed of a Mayoid and a so-called Bold Animalistic 
tradition. This original fission and subsequent parallelism of both Ulua 
and Yojoa Polychrome ceramic development has obvious sociological 
as well as archeological implications. At both Ulua River and Lake 
Yojoa Polychrome sites one of these styles is Mayoid and the other 
is of southern origin. For linguistic and ethnographic reasons previ- 
ously discussed, it seems highly probable that the Bold Geometric 
element of the Ulua Polychrome was contributed by Jicaque peoples, 
whereas the very similar Bold Animalistic element in Yojoa Poly- 
chrome was due to the related Lenca. Since the Mayoid element com- 
prises about one half of the Ulua and Yojoa Polychrome ceramic re- 
mains, it can hardly be explained as due solely to trade or indirect 
influence. It seems far more logical to assume that intermixed Maya, 
Jicaque, and Lenca peoples were living together at these sites and 
that perhaps the pottery-makers of each ethnic group clung to their 
own art styles over a considerable period. The quite remarkable 
florescence and the high and complex artistic attainments of the Ulua 


and the Yojoa Polychrome periods are in all probability the direct 
results of this cultural and physical amalgamation. 

We have at present no means for dating the exact period repre- 
sented by these Lenca and Jicaque styles which apparently stem from 
Nicaraguan and Nicoyan culture centers to the south. On the other 
hand, there is in the nearby Maya city of Copan a series of dated 
monuments ranging from 9: ii.o.o.o (stela 3). to 9: 17. 12.0.0. (stela 
C) (or, roughly, according to the Goodman-Thompson-Martinez 
correlation, between 650 and 800 A. D.), in association with which 
there occur pottery vessels (Vaillant, 1927, and Lothrop, 1933, p. 66 
and 1936 b, p. 69). We have attempted to correlate our Ulua and 
Lake Yojoa Polychrome series with Vaillant's classification of Copan 
wares, but owing to the selective nature of the Copan collections, as 
well as the paucity of illustrative material, this has proved imprac- 
ticable for the present. The Copan ceramic series in the Peabody 
Museum, as a whole, seems quite distinct from the Ulua- Yojoa Poly- 
chrome wares, although numerous similarities do exist. Vaillant points 
out the occurrence of Ulua Polychrome sherds at mound 36 in Copan, 
a point we were able to verify for ourselves at the site, but there is 
reason to believe that these deposits are later than the Copan series or 
perhaps intrusive. According to Vaillant (1927, p. 271) the trend 
of the Ulua Polychrome wares " suggests the years after the fall 
of Copan." If this is the case, it may serve to point out when the 
Maya Old Empire dispersal into Salvador and northwestern Honduras 
took place and how their developed polychrome wares came to be 
grafted on to those of the Lenca, Jicaque, and, probably, the Pipil, 
with whom the various Maya groups settled. When adequate strati- 
graphic studies of the entire range of Copan ceramics have been made 
and correlated with the ceremonial series from the stelae vaults, 
described by Vaillant, there is reason to believe that the Ulua- Yojoa 
Polychrome series may also be approximately dated. 

Such excavations should also throw light on the origin or deriva- 
tion of the southern Mayoid Polychrome ceramic tradition. Did it 
arise from a groundwork similar to the Playa de los Muertos cul- 
ture in the Peten, perhaps at Uaxactun or Holmul, spread from there 
to Copan, and thence to Salvador and the Ulua ? Or are there inter- 
mediate stages between the developed Polychrome and the Playa de 
los Muertos horizons present but as yet unknown in Honduras, at 
Copan, or in Salvador? An even more basic problem concerns the 
suggested relationship between the ceramics in the oldest horizons 
at Uaxactun in the Peten and Chukumuk in the Guatemalan high- 
lands, with the Playa de los Muertos Bichrome and Ulua Bichrome 


wares respectively. Lothrop (1933, p. 62) believes that the elements 
shared in common by the earliest known cultural horizons in the high- 
lands and the Atlantic lowlands of Guatemala were derived, not from 
one another, but from' a parent culture to the south. When the 
Uaxactun materials are available, the role of the southern Playa de 
los Muertos culture as a donor or a recipient may be tested. These 
are questions for the future but, thanks to the growing scientific 
vogue of the shovel, they are questions that may soon be answered. 

Bancroft, H. H. 

1882. The native races of the Pacific States, vols. 1-5. San Francisco. 
1883-7. History of Central America, vols. 1-3. (Vols. 6, 7, and 8 of 
Bancroft's Works. San Francisco, 1882- 1890.) 
Blackeston, R. H. 

1910. Recent discoveries in Honduras. Amer. Anthrop., n. s., vol. 12, no. 4, 
pp. 536-541. 
Blom, Frans, Grosjean, S. S., and Cummins, Harold 

1933. A Maya skull from the Uloa Valley, Republic of Honduras. Tulane 
Univ. Studies in Middle America, Publ. 5, no. i. New Orleans. 
CoLECCiON DE DocuMENTOS Ineditos relativos al descubrimiento, conquista 
y colonizacion de las posesiones Espaiiolas en America y Oceania, 
tomos 1-41. Madrid, 1854-84. 
Cummins, Harold. See Blom, Frans, Grosjean, and Cummins. 
Diaz del Castillo, Bernal 

1908-16. The true history of the conquest of New Spain, vols. 1-5. Edited 
and published in Mexico by Genaro Garcia. Translated by A. P. 
Maudsley. Hakluyt Soc. Publ., ser. 2, vols. 23-25, 30, and 40. 
Gordon, George Byron 

1898. Researches in the Uloa Valley, Honduras. Mem. Peabody Mus. 
Archaeol. Ethn., Harvard Univ., vol. i, no. 4. Cambridge. 
Grosjean, S. S. See Blom, Frans, Grosjean, and Cummins. 
Habel, S. 

1880. The sculptures of Santa Lucia Cosumalwhuapa in Guatemala. Smith- 
sonian Contr. Knowl., vol. 22, art. 3, pp. 1-90. Washington. 
DE Las Casas, Barthelemi 

1922. CEuvres de Don Barthelemi de Las Casas, eveque de Chiapa. Edited 
by J. A. Llorente. Vols. 1-2. Paris. 
Lehmann, Walter 

1910. Erbebnisse einer Forschungsreise in Mittelamerika und Mexico, 1907- 
' 1909. Zeitschr. Ethnol., vol. 42, pp. 687-749. Berlin. 

1920. Zentral Amerika. Vols. 1-2. Berlin. 
Lothrop, S. K. 

1921. The stone statues of Nicaragua. Amer. Anthrop., n. s., vol. 23, no. 3, 

pp. 311-319. 
1933. Atitlan. An archaeological study of ancient remains on the borders 
of Lake Atitlan, Guatemala. Carnegie Inst, of Washington. Publ. 


1936 a. Sculptured pottery of the Maya and Pipil. Maya Research (Mexico 

and Central America), vol. 3, no. 2, pp. 140-152. New Orleans. 
1936 b. Zacualpa. A study of ancient Quiche artifacts. Carnegie Inst, of 
Washington, Publ. 472. 
Mason, Otis T. 

1889 a. How to straighten a spear shaft. Amer. Anthrop., ser. i, vol. 2, 

no. 2, p. 158. 
1889 b. Music in Honduras. Amer. Anthrop., ser. i, vol. 2, no. 2, p. 158. 
DE PALAao, Diego Garcia 

i860. Carta dirijida al Rey de Espana. Ano. 1576. (Original Spanish with 
English translation.) Squier's " Collection of rare and original 
documents and relations ", no. i. New York. 
PoPENOE, Dorothy H. 

1934. Some excavations at Playa de los Muertos, Ulua River, Honduras. 

Maya Research, vol. i, no. 2, pp. 61-86. New York. 
1936. The ruins of Tenampua, Honduras. Ann. Rep. Smithsonian Inst. 
1935, pp. 559-572. 
Spinden, Herbert J. 

1925. The Chorotegan culture area. Compt. Rend. 21st Congr. Internat. 
Amer., Goteborg, 1924, pp. 529-545. Goteborg. 
Squier, E'phraim George 

1858. The States of Central America. New York. 

1859. A visit to the Guajiquero Indians. Harper's New Month. Mag., 

vol. 19, no. 113, pp. 603-619. New York. 
i860. Some account of the Lake of Yojoa or Taulebe, in Honduras, Central 
America. Journ. Roy. Geogr. Soc, vol. 13, pp. 58-63. London. 

1869. Tongues from tombs, or the stories that graves tell. Frank Leslie's 

Illustr. Newsp. New York. 

1870. Honduras. Descriptive, historical, and statistical. London. 
Smith, H. E. 

1936. Preliminary shape analysis of the Uaxactun pottery. (Typed script, 

with illustrations reproduced photographically.) Guatemala. 
Stone, Doris Zemurray 

1934. A new southernmost Maya City. Maya Research, vol. i, no. 2, pp. 

125-132. New York. 
Strong, William Duncan 

1934 a. Hunting ancient ruins in northeastern Honduras. Explorations and 

Field-Work of the Smithsonian Inst., 1933, pp. 44-48. 
1934 b. An archeological cruise among the Bay Islands of Honduras. Ibid., 

pp. 49-50. 

1935. Archeological investigations in the Bay Islands, Spanish Honduras. 

Smithsonian Misc. Coll., vol. 92, no. 14, pp. 1-176. 

1937. Archeological explorations in northwestern Honduras. Explorations 

and Field-Work of the Smithsonian Inst., 1936, pp. 75-82. 
Thomas, Cyrus, and Swanton, John R. 

191 1. Indian languages of Mexico and Central America and their geo- 
graphical distribution. Bur. Amer. Ethnol. Bull. 44. 
DE Torquemada, Juan 

1723. Los veinte i un libros rituales i Monarchia Indiana, con el origen y 
guerras, de los Indios Occidentals. Madrid. 


TozzER, Alfred M. 

1930. Maya and Toltec figures at Chichen Itza. Proc. 23d Internat. Congr. 
Amer., New York, 1928, pp. 155-164. New York. 
TscHOPiK, Harry, Jr. 

1937. Textile motifs from Uloa Valley pottery. MS., Peabody Mus., 
Harvard Univ. 
Vaillant, G. C. 

1927. The chronological significance of Maya ceramics. Manuscript thesis 
submitted in partial fulfillment of the requirements for the degree 
of Doctor of Philosophy, Harvard Univ. 
1934. The archaeological setting of the Playa de los Muertos culture. Maya 
Research, vol. i, no. 2, pp. 87-100. New York. 
Wells, Wm. V. 

1857. Explorations and adventures in Honduras. New York. 
Yde, Jens 

1935- Foreljzfbig Beretning on Nationalmuseets og Tulane Universitetets 
Ekspedition til Mellemamerika 1935. Saertryk Geogr. Tidsskr., 
vol. 38, nos. 3-4, pp. 137-155- 
1936. A preliminary report of the Tulane University-Danish National Mu- 
seum expedition to Central America. Maya Research (Mexico and 
Central America), vol. 3, no. i, pp. 25-37. New Orleans. 


Plate i 

Processional figures on a Yojoa Polychrome vase, Mayoid type. Site 2, La 
Ceiba (13.2 cm high). 

Plate 2 

Various Chamelecon and Ulua River sites 

Fig. I. Fragment of ball court ring in sihi, Naco. 
Fig. 2. Thin plaster walls, heart of mound 3, Naco. 

Fig. 3. Excavation i and start of excavation 2 (right) at Santa Rita (farm 17). 
Fig. 4. Site at Tres Piedras, showing mounds and plaza cross-sectioned by 
Chamelecon River. 

Plate 3 

Naco sherds 

a, c-w, Naco Polychrome ; h, Ulua Polychrome sherds found at Naco. 

Plate 4 

Naco sherds and artifacts 

a, incensario ; b-g, figurine fragments ; h, whistle ; i, j, spindle whorls ; 
k, obsidian flake knives; I, clay bobbins; m, Spanish colonial crockery; 
n-p, textile-marked sherds; q, s, t, u, v, x, y, s, sherds with molded or 
carved designs ; r, incised sherd ; tv, " candelarios." 


Plate 5 

Upper Ulua Polychrome pottery types, Las Flores 

a-e, Las Flores painted and incised vase sherds ; ]-m, Upper Mayoid type 
sherds ; n, sculptured or molded Mayoid sherd. 

Plate 6 

Upper Ulua Polychrome pottery types, Las Flores 

a, human effigy (16.5 cm); h, vessel with "vestigial" spout (14 cm); c, 
Upper Mayoid jar (7 cm) ; d, sculptured or molded Mayoid jar (6.5 cm) ; 
e (5 cm), / (6.5 cm), jars of imitation Ulua marble bowl type. 

Plate 7 
Ulua Polychrome, Bold Geometric pottery types, Santa Rita 

a. Bold Geometric oUa (12 cm) ; h (22 cm) ; c (30 cm) ; d Lower Bold Geo- 
metric olla type (30 cm) ; e. Bold Geometric tripod dish (7.5 cm) ; /, Bold 
Geometric tripod dish (10 cm), (a-r, Santa Rita; /, Naranjo Chino.) 

Plate 8 

Ulua Polychrome, Mayoid pottery types, Santa Rita 

a. Lower Mayoid type vase (20 cm) ; h, Lower Mayoid type vase (20 cm) ; 
deer effigy pot cover (19.5 cm) ; e, Mayoid type vase (i7-8 cm) ; e, f, tripod 
plate, type uncertain (14.5 cm). 

Plate 9 

Ulua Bichrome sherds, deepest level, Santa Rita 

a-j, various sherds ; k, obsidian scraper ; /, pottery stamp ; m, obsidian flake 
knife fragment ; o-s, n-s, aa, bb, Usulatan ware sherds ; t, Lower Mayoid 
type sherds from just above sterile sand layer; cc, red-on-white sherd. 

Plate 10 

Playa de los Muertos Bichrome sherds 

a-h, polished orange-red to brown ; i-n, polished dark gray to black ; o-s, 
polished slate-gray to buff. (Lower cultural horizon, Playa de los Muertos.) 

Plate ii 

Playa de los Muertos sherds and figurines 

a-e, sherds with chalky white wash; /, g, k, red and black; i, j, 0, red on 
buff ; h, unslipped brown and red ; /, m, red on white ; n, gray on dull red ; 
p, polished brown face; q, r, s, polished figurines with white slip; t, u, v, 
solid brown figurines (lower cultural horizon, Playa de los Muertos). 


Plate 12 

Yojoa Polychrome vessels, Mayoid types 

a, excavation B, Los Naranjos (25 cm) ; b, Aguacate (16.3 cm) ; c (10.7 cm) ; 
d (10.5 cm), Aguatal; e, Aguacate (9.5 cm) ; /, La Ceiba (12 cm). 

Plate 13 

Yojoa Polychrome vessels 

Bold Animalistic type: a. La Ceiba (15 cm) ; b, site i, Los Naranjos (12.5 
cm) ; c, Aguatal (10 cm) ; d, La Ceiba (11 cm) ; e. Effigy (type ?, compare 
fiff- 7, P, P- 52), La Ceiba (11 cm) ; /, Mayoid type, Aguatal (11. 5 cm). 

Plate 14 

Yojoa Polychrome vessels 

a (10.7 cm) ; b (10.2 cm). Bold Animalistic type, Aguacate; c (8 cm) ; d (10 
cm), Bold Geometric or Bold Animalistic types, Aguacate and Los Naranjos, 
site i; e, imitation Ulua marble bowl type, Aguacate (6.7 cm) ; /, carved 
brown ware, Aguacate (7.3 cm) ; g, bowl with negative painting, Los 
Naranjos, site i, (6 cm) ; h, bird-shaped pot, Aguatal (7 cm). 

Plate 15 

Early ceramic types at Lake Yojoa 

a (11 cm) ; b (12 cm), Playa de los Muertos Bichrome type ( ?), Los Naranjos 
(exact provenience uncertain) ; c, d, f-i, k, I, o-s, u-w, Yojoa " Monochrome " 
sherds ; e, j, figurine fragments ; m, obsidian flake ; n, t, ground stone artifacts. 
{c-w, lower cultural horizon at Los Naranjos, site i, and excavations A 
and B.) 

Plate 16 

Los Naranjos, Lake Yojoa 

Fig. I. Crude anthropomorphic statue. 

Fig. 2. Stone serpent head. 

Fig. 3. Stone torso and head. 

Fig. 4. Mound i, from the north near site i. 

Fig. 5. Section of trench at site i, showing house floor and burial. 


VOL. 97, NO. 1, PL. 2 



Various Chamelecon and Ulua River Sites 

I, Ball court at Naco; 2, mound structure at Naco ; 3, excavation at 
Santa Rita (farm 17) ; 4, Tres Piedras site. 


VOL. 97, NO. 1, PL. 3 

Naco Sherds 

/', Ulua Pulyclirome sherds at Naco. 


VOL. 97, NO. 1, PL. 4 

Naco Sherds and Artifacts 

in, Spanish colonial sherd at Naco. 


VOL. 97, NO. 1, PL. 5 

k f 

UPPER Ulua Polychrome Pottery Types. Las Flores 


VOL. 97, NO. 1, PL. 6 

Upper Uuua Polychrome Pottery Types. Las Fuores 


VOL. 97, NO. 1, PL. 7 


/, Naranjo Chine. 


VOL. 97, NO. 1, PL. 

Ulua Polychrome. Mayoid Pottery types. Santa Rita 


VOL. 97, NO. 1, PL. 9 





,l-.v -/ 

tr? jr' 


t, Lower Mayoid sherds on sterile sand layer above Ulua Bfchrome horizon. 


VOL. 97, NO. 1, PL. 10 




VOL. 97, NO. 1, PL. 1 1 

^ / a 



VOL. 97, NO. 1, PL. 12 

YOJOA Polychrome Vessels. Mayoid types 


VOL. 97, NO. 1, PL. 13 



YojoA Polychrome Vessel.s, bold Animalistic Types 

c, uncertain type ; /. Mayoid type. 


VOL. 97, NO. 1, PL. 14 

YOJOA Polychrome Vessels, Various Types 

(7-(/, Bold Animalistic and Bold Geometric ; c, imitation Ulua marble bowl. 


VOL. 97, NO. 1 . PL. 15 


Early Ceramic Types at Lake Yojoa 

a, b, Playa de lus Muertos Bichrome (?) ; r-zi-, Yojoa "Monochrome." 


VOL. 97, NO. 1. PL. 16 

LOS Naranjos, Lake Yojoa 

I, crude human statue; 2, stone serpent head; 3. stone torso and head 
4, mound i ; 5, section of trench, site i. 




(With Four Plates) 


Assistant Director, Division of Radiation and Organiams 
Smithsonian Institution 

(Publication 3446) 



JANUARY 12, 1938 




(With Four Plates) 



Assistant Director, Division of Radiation and Organisms 
Smithsonian Institution 

(Publication 3446) 



JANUARY 12, 1938 

Z^ £ovi) <§Aitimott (preee 




By earl S. JOHNSTON 
Assistant Director, Division of Radiation and Organisms, Smithsonian Institution 

(With Four Plates) 


There can be little doubt that wave-length distribution exerts an 
enormous influence on the growth of plants. Numerous experiments 
show that stem elongation is greatly retarded under blue light, whereas 
an acceleration takes place in the red and near infrared regions. 
Chlorophyll production takes place better toward the red than toward 
the blue end of the visible spectrum. Phototropic sensitivity is great- 
est in the blue and zero in the red. For equal amounts of energy 
falling on the leaf, two maximal regions of CO2 absorption have been 
found — one in the red, the other in the blue. It thus appears that a 
wave-length region best suited to a given plant process may be entirely 
without effect upon another. 

In plant nutrition studies, experiments have shown that there is a 
general balance in the proportionate amounts of mineral elements of 
a nutrient solution that brings about a favorable growth response in 
plants. Although there may be considerable latitude in the ratio of 
amounts of elements in such a solution, it may be said that a balanced 
condition exists. 

In a somewhat analogous manner, it is possible to think of the light 
requirements of plants as a balanced condition of intensities of dif- 
ferent wave lengths which bring about good plant growth. Both light 
intensity and wave-length distribution vary to a considerable extent 
over the earth's surface. Likewise the character of the vegetation 
varies. Since plants have been growing on the earth for countless 
ages, it is reasonable to assume that their physiology is adjusted best 
to sunlight. Although there is experimental evidence to show that 
different processes go on better in some wave-length regions of the 
spectrum than in others, yet the best growth, when all the processes 
are considered simultaneously, apparently takes place in the natural 

Smithsonian Miscellaneous Collections, Vol. 97, No. 2 


light of the sun. A direct experimental comparison between sunlight 
and artificial light is, of course, difficult to make because of the great 
number of variables entering into the problem. 

Numerous experiments have been made for the purpose of growing 
plants under artificial illumination. The object of many such experi- 
ments was to find a satisfactory artificial light which could be operated 
economically on a commercial scale. In other experiments the technical 
and scientific aspects were the main objectives. So far as is known, 
there is no available light source which is like that of the sun in its 
wave-length distribution. Plants have been grown fairly successfully 
in a few instances under well-controlled laboratory conditions, but 
the problem is by no means solved. It may even be found that plants 
can be grown normally under greatly reduced intensities of light pro- 
vided a proper proportion between the intensities of its component 
wave lengths is worked out. 

The purpose of the present report is to discuss briefly some pre- 
liminary experiments dealing with the question of a wave-length 
balance of artificial light. 


In the experiments herein described, plants were grown between 
two different light sources. Three or more types of lights could be 
used, but for this preliminary survey it was thought best to limit the 
wave-length distribution to two types. All the experiments were con- 
ducted in a small room (approximately 15 x lo ft. x 8 ft. high) the 
walls and ceilings of which were painted a flat black to minimize 
scattered light effects. Both temperature and humidity were auto- 
matically controlled. The plants were grown in i-quart jars contain- 
ing nutrient solution. Each culture was placed on a small rotating 
table and usually grown for 3 weeks with a daily light period of 12 
to 18 hours. By constantly rotating the plants (3.4 r.p.m.) on an axis 
parallel to their stems, the phototropism of these stems was reduced 
to zero. The leaves in some experiments showed phototropic response. 
The wave-length distribution depended upon the light source. The 
intensity was regulated largely by the distance the culture was placed 
from the light. 

In an earlier paper Johnston (1932) found that the excess of near 
infrared of the Mazda lamp caused a distinct yellowing of tomato 
leaves. If this region of the spectrum was not actually destructive to 
chlorophyll, . it was of little or no benefit to its formation. It would 
thus appear that more nearly normal color could be obtained by re- 


ducing the infrared or by increasing the intensity of the rest of the 
spectrum. An experiment was therefore planned in which this was 
partially accomplished by building up the blue end of the spectrum. 

Experiment i. — Two i,ooo-watt projection Mazda lamps (115 v.) 
were placed i meter apart. Surrounding each lamp was a clear Pyrex 
thermos bottle blank fitted with a water inlet and outlet. The radiation 
of each lamp was thus filtered through 5 mm of water. The constant 
flow of water through this jacket was a great aid in maintaining a 
constant temperature condition in the room, since a great deal of heat 
was thus removed. 

A copper sulphate (sp. gr. 1.08, about 8 percent) filter (6 cm thick) 
was placed in front of one of the lamps. This was the added blue 
light source. The individual rotating plant cultures were located 
at positions relative to these two light sources which gave the intensity 
values expressed as watts/cm^ in table i. 

Table i. — Radiation intensities and plant data from experiment i 

Light intensity 

watts/cm^ Stem Total 

Culture . ^^ > ht. dry wt. 

no. White Blue Total cm gram 

1 0396 .0006 .0402 5.0 .086 

2 0285 .0010 .0295 7.1 .102 

3 0166 .0014 .0180 6.1 .043 

4 0064 .0027 .0091 6.3 .025 

5 0046 .0054 .0100 3.8 .019 

Marglobe tomato seeds were sprouted between moist filter paper 
at a temperature of 25° C. for 3 days. The sprouted seeds were then 
transferred to a germination net, and after about a week of growth 
five similar seedlings were selected and set out in quart jars, one per 
jar, and placed on the five small rotating tables. After 2 weeks of 
growth these plants were measured and dried in an oven at 103° C. 
to obtain the dry weight. These data are also shown in table i. 

Because of the meagerness of data, no definite conclusions can be 
drawn. The first three plants were heavier than similar ones grown 
in the north and south laboratory windows. Although the total in- 
tensity of no. I was greatest, yet maximum dry weight occurred in 
plant no. 2. Here the added blue radiation was about 3 percent of 
the total as compared to 1.5 percent in plant no. i, which was yellow- 
green in color. Plant no. 2 was a light green when compared to plants 
3, 4, and 5, whose percentage of added blue to total radiation were 
respectively 8, 30, and 54. 


Experiment 2. — The next experiment was very similar to that just 
described. Here again individual variation was too great to draw any 
accurate conclusions. 

Experiment j. — In the next experiment three duplicate sets of 
tomato plants were grown under three different sets of light condi- 
tions. In front of one lamp a Corning heat-absorbing 212 percent 
red filter was placed. In front of the other lamp a filter jar containing 
a M/2 CUSO4 solution was placed. Both filters cut off at 6040 A, the 
CuSOi solution transmitting light of shorter wave length and the 
Corning filter transmitting light of longer wave length. Two duplicate 
sets of cultures were placed between these filtered light sources. A 
third set was located to the rear of the blue filter light in such a position 
that the plants received only the full Mazda spectrum. Intensities 
were measured at the beginning and at the end of the experiment. 
These average values, together with the plant data for 3 weeks' 
growth are given in table 2. 

Table 2. — Radiation intensities and azvrage plant data from experiment j 

Average data per plant 

Radiation intensity , ^- ^ 

watts/cm^ Stem Root Total 

Culture I * > ht. length dry wt. 

nos. Red Blue White Total cm cm gram 

I and 2 0055 .0022 .0057 6.3 38 .026 

3 and 4 0028 .0011 .0039 7.4 50 .038 

5 and 6 .0056 .0056 8.0 44 .028 

The greatest amount of dry weight was produced by cultures 3 and 
4, although the total light intensity was less than under the other two 
conditions of growth. Here the blue radiation was about 28 percent 
of the total. Although these data are meager, there is an indication 
that considerable differences in growth are obtained by manipulating 
the wave-length distribution as well as the total intensity. 

Experiment 4. — In the next experiment wave-length distribution 
was further restricted by using neon and mercury grids as light 
sources. These were constructed in our laboratory by Mr. L. B. 
Clark, In order to increase the intensities a mirror was placed back 
of each. Three duplicate cultures were placed between these two 
light sources, each culture jar containing three tomato seedlings. 
This increased the number of plants per treatment to six. Because 
of reflections in the mirrors some red light came from the blue side 
and some blue light came from the red side of the cultures. As will 
be seen in table 3, the intensity of radiation was considerably less 
than in the earlier experiments. The plants were grown for 26 days 


and then harvested. The average stem height and total dry weight 
per plant for each of the three light conditions appear in the same table. 

Although the stems of plants in group 1-2 were thicker than those 
in the other groups, their leaves were quite yellow. Here again yel- 
lowing is associated with energy distribution where the greatest 
amount is found in the red end of the spectrum. The plants in group 
5-6 had the best color, even though the total amount of energy was 
about half that of group 1-2. These leaves had flat smooth surfaces, 
while those in group 1-2 were quite pointed and curled. The general 
appearance of these plants is shown in plate i. After another experi- 
ment with these lamps it was definitely indicated that the plants were 
getting insufficient illumination. 

Experiment 5. — To increase the radiation, a General Electric 400- 
watt high-pressure mercury lamp was substituted for the mercury 

Table 3. — Radiation intensities and average plant data from experiment 4 

Radiation intensity Average data 

at beginning of experiment per plant 


\ Stem Total 

Culture Neon Mercury ht. dry wt. 

no. (red) (blue) Total cm gram 

1 00023 .00002 .00025 6.6 .029 

2 00019 .00002 .00021 6.6 .029 

3 00008 .00002 .00010 3.6 .014 

4 00008 .00002 .00010 3.6 .014 

5 00005 .00007 .00012 3.4 .013 

6 00005 .00007 .00012 3.4 .013 

grid and four instead of two transformers were used with the neon 
lamp. The daily light period was increased from 12 hours to 18 
hours. Because of the marked decrease in the life of the neon lamp 
under these forced conditions, the experiment was discontinued at 
the end of 20 days. In this exploratory experiment no accurate in- 
tensity measurements were made. However, general improvement in 
growth was noted. 

Experhnent 6. — To increase further the light intensity, a 1,000- 
watt, iio-volt projection lamp housed in a water jacket as noted 
earlier (experiment i) was substituted for the neon grid lamp. Three 
plants per quart culture jar were used and the cultures run in dupli- 
cate so far as the light relations were concerned. Throughout all 
the previous experiments the plants were grown in a three-salt 
nutrient solution similar to that used by Johnston and Dore (1929). 
In this experiment cultures 2, 4, and 6 had (NH4)2S04 added to the 
former solution which contained Ca(N03)2, MgSOi, and KH2PO4 


and traces of Mn, B. Iron was added as FeS04 to all cultures from 
time to time as conditions demanded. Because the Mazda lamp was 
run at about its voltage limit its life was short, and replacements were 
necessary every 6 or 7 days. The plants were grown for 3 weeks with 
a daily light period of 18 hours. The added heat from the lamps 
caused a slight daily temperature fluctuation. The average maximum 
was 24° C. and the average minimum 21.5° C. This resulted in a 
change in humidity which averaged 57 and 51 percent for the dark 
and light periods respectively. As found in previous experiments, a 
temperature fluctuation is beneficial to the tomato plant. Better 
growth was obtained by subjecting the plants to a lower dark period 

Table 4. — Radiation measurements at beginning of experiment 6 

Foot-candles with 
Watts/cm^ small G. E. meter 

Culture ' 

nos. Mazda Mercury Total Mazda Mercury Total 

I and 2 0404 .0013 .0417 2,800 ' 200 3,000 

3 and 4 0172 .0031 .0203 1,200 600 1,800 

5 and 6 0065 .0067 .0132 550 1,000 i,5S0 

Table 5. — Plant data from experiment 6 expressed as averages per plant 

Dry wt. 
Stem Green wt. gram 

Culture ht. grams 

no. cm Tops Tops Roots Total 

1 17-5 6.6 .529 .139 .668 

2 21. 1 7.8 .671 .174 .84s 

3 18.9 5-6 .463 -091 .554 

4 20.9 6.1 .469 .074 .543 

5 23.5 s-o .384 .074 -458 

6 23.8 5-4 -371 -057 428 

temperature (about 3° C lower) than by maintaining a constant 
temperature during the dark and light periods. 

The intensity measurements which were made at the beginning of 
the experiment are presented in table 4. 

After 3 weeks of growth the plants were photographed (pi. 2) and 
harvested. Data giving average stem height, green weight of tops, and 
dry weight of tops and roots are given in table 5. 

Both the illustrations and plant data show that this group of plants 
was normal in appearance and comparable to good greenhouse plants. 
It was by far the best we have grown under the 100 percent artificial 
conditions of our laboratory. In an earlier publication, Johnston 
(1932) reported that tomato plants exposed to an intense illumination 
from a Mazda lamp grew very well but soon became yellow in color. 


The near infrared radiation was apparently destructive to chlorophyll 
or inhibited its formation. This again appeared to be the case for the 
three plants in culture i. However, one of the most interesting ob- 
servations made in this experiment was that the color of the plants in 
culture 2, which received the same radiation intensity as those of 
no. I, was much greener. This color difference is seen to some extent 
in plate 2 as differences in light and dark tones of the plants in the 
upper and lower figures. This was also true for the plants in cul- 
tures 4 and 6, as compared with cultures 3 and 5 respectively, which 
were grown under similar light conditions. All the plants grown in 
nutrient solution to which (NH4)2S04 had been added were greener 
than the corresponding ones without this additional nitrogen. This 
observation suggests the influence of the type of radiation on the up- 
take of mineral nutrients. This same solution without the (NH4)2S04 
has been used in growing tomato plants in the greenhouse but the 
characteristic chlorotic effects were not noted until the plants were 
grown under Mazda lamps. 

The percentages of added mercury radiation to total illumination 
were 3, 15, and 50 respectively for cultures 1-2, 3-4, 5-6. The green 
color of the leaves was deeper where this percentage was larger. A 
more striking color difference occurred, however, between the plants 
in cultures with and without the (NH4)2S04. 

The average total dry weight per plant for each of the three light 
conditions 1-2, 3-4, 5-6 was .757, .549, and .443 gram respectively. 
Under these three light conditions the efficiency in the production of 
dry weight per watt/cm" was 18, 27, and 34 respectively. Although 
the total intensity of 5-6 was about a third that of 1-2, on the basis 
of efficiency in producing dry weight per unit energy, the less intense 
radiation was about double that of the more intense. 

One other factor in addition to wave-length distribution must be 
recognized in an experiment of this type. One lamp (Mazda) gave 
practically continuous illumination; the other (mercury, 60-cycle), a 
fluctuating illumination varying from a minimum considerably below 
the average to a maximum much greater than the average as de- 
termined by the thermocouple and photoelectric cell. McAlister 
(1937) clearly shows that a change in efficiency of carbon dioxide 
assimilation occurs with frequency of intermittency of illumination. 
Although it may be comparatively safe to compare the different cul- 
tures in any one experiment since the " flicker " effect is doubtless 
the same, it is impossible to compare results of experiments in which 
the light is continuous with those in which it is intermittent or with 
those in which it is half continuous and half intermittent. 



Experiment 7. — An experiment varying a little from the one just 
described was next performed. In this, five cultures of tomato seed- 
lings were placed around the Mazda lamp at positions which gave 
them approximately equal light intensities from this lamp. A sixth 
culture was located at a position where the Mazda intensity was about 
half that of the other cultures. The intensities of each of the two 
lamps for each culture of these plants are best seen in table 6. 

After 18 days the plants were harvested and their dry weights 
determined. The plant data appear in tables 7 and 8. 

Table 6. — Radiation measurements^ from experiment 7 

Foot-candles with 

Watts/cm^ small G. E. meter 

Culture , *- V , * , 

no. Mazda Hg Total Mazda Hg Total 

1 0234 .0097 .0331 1,300 2,000 3,300 

2 0241 .0018 .0259 1,300 300 1,600 

3 0241 .0013 .0254 1,400 200 1,600 

4 0240 .0013 .0253 1,400 200 1,700 

5 0231 .0039 .0270 1,300 600 1,900 

6 0II2 .0090 .0202 600 1,900 2,500 

^ Since the original Mazda lamp was replaced after 6 days, these measurements were made 
on the second lamp on the loth day of the experiment. 

Table 7, — Stem height (cm) data from experiment 7 

Culture number 

. A 

Plant 1234 

a 16.5 16.2 19.0 18.5 

b 15-8 18.0 15.5 22.3 

c 14.5 19.4 19.0 20.0 

Av. ht. at harvest 15.6 17.9 17.8 20.3 

Av. original ht 2.3 2.1 2.2 1.6 

Av. increase in ht 13.3 15.8 15.6 18.7 















Table 8. — Average green and dry weights (grams) of plants from 
experiment 7 

Culture Green wt. 

no. Tops 

1 6.1 

2 4.2 

3 4-8 

4 5.8 

5 6.6 

6 6.3 


Dry wt. 







During the experiment water was added and fresh nutrient sokition 
renewed as required. Because of frequent stopping of rotating table 
no. I, these plants were slightly burned. These plants had the shortest 
internodes. Plants in culture 6 had next to the shortest internodes 
and were the best green. The leaves of plants in cultures 2, 3, 4, and 
5 were slightly chlorotic. In order of their dry weights, plants in 
cultures i, 6, and 5 were the best. It is interesting to compare the 
total dry weight per unit total energy with the percentage of energy 
received from the mercury lamp (table 9). 

Table 9. — Comparison of dry weight efficiency mith amount of radiation from 

the mercury lamp 

Cultures 12 3 4 S 6 

Ratio total dry wt. to watts/cm^ 27.6 13.7 17.8 21.0 25.2 37.4 

Percentage radiation from mercury lamp 29 7 5 5 14 45 

Plants in cultures 6, i, and 5 produced the greatest amount of dry 
weight per watt/cml These same cultures in the order given received 
the largest percentages of radiant energy rich in the blue. Total 
energy (table 6) was greatest for culture i and least for culture 6. 
Cultures 2, 3, and 4 were practically equal. Thus, plants of culture i 
had the greatest total dry weight, and those of 6 were second. How- 
ever, for greatest efficiency in the production of dry weight, plants 
in culture 6 were much better than those in culture i. This is evi- 
dently related to the greater percentage of shorter wave length in 
the one case than in the other. When light intensity as measured 
by the foot-candle meter is considered, plants of culture i are shown 
as receiving the greatest amount of light and those of culture 2 the 
next greatest amount. 

By consulting the table of stem heights it will be noted that the 
average height at harvest for plants in culture i was less than any of 
the other groups although the average original height was greatest. 
The least average stem elongation shown by this group may be cor- 
related with the greatest amount of total energy received by these 
plants. . But little difference in stem height is seen between plants 
of the other cultures. Likewise there is but little difference in total 
energy received by these same cultures. Other observations bear out 
this same point that an intense light retards stem elongation more 
than a less intense one. Although the shorter wave lengths have a 
greater retarding effect, this difference between plants of cultures 6 
and I must have been offset by the differences in total radiation 


Experiment 8. — In the last experiment of this series, the same types 
of lamps were used. Also each culture contained three tomato plants. 
The first four cultures (nos. i, 2, 3, 4) were arranged around the 
mercury lamp at approximately equal distances. The other two (nos. 
5 and 6), together with those numbered 3 and 4, were located about 
equal distances from the water-cooled 115-watt Mazda projection 
lamp. The intensity measurements taken at the beginning of the 
experiment are shown in table 10. 

Table 10. — Radiation measurements in experiment 8 

Intensity measurements 


Watts/cm^ Foot-candles 

no. Mazda Hg Total Mazda Hg Total 

1 0071 .0048 .0119 350 700 1,050 

2 0071 .0047 .0118 350 700 1,050 

3 0140 .0046 .0186 800 700 1,500 

4 0144 .0045 .0189 800 700 1,500 

5 0137 .0010 .0147 900 200 1,100 

6 0146 .0010 " .0156 900 200 1,100 

At the end of three weeks the plants were photographed and har- 
vested. Since each culture of three plants was duplicated, the average 
of the six plants is shown in table 11. 


II. — Plant 

data f 

rom experiment S expn 

^ssed as 


■ages per 


Plant data expressed as averages per 




stem ht. 


No. of 

Green wt. 

of tops 




water loss 


Dry weight (g) 




I and 2. 

.. 18.3 







3 and 4. 

.. 20.8 







5 and 6. 

.. 15.8 







Plants of cultures 3-4 were best in general appearance and had the 
thickest stems. Those of cultures 5-6 were lightest green. Plants 
with longest roots were found in cultures 5-6; those with shortest 
roots occurred in cultures 1-2. 

The general appearance of the cultures about the two lamps in this 
experiment may be seen in plate 3, and the appearance of the tops 
and roots of the plants at the end of the experiment is seen in plate 4. 

It will be recalled that in experiment 6 the ratio of dry weight to 
total energy increased with the percentage of added radiation from 
the mercury lamp. Also in experiment 7, table 9, the three cultures 


in which the greatest dry weight was produced per watt/cm° were the 
same three cultures which received the greatest percentage of radia- 
tion from the mercury lamp. In experiment 8, however, an exception 
occurred. The dry weight efficiencies for the three groups of cultures 
1-2, 3-4, 5-6, were 23.8, 28.6, 10.7 respectively, while the percentages 
of total radiation attributed to the mercury lamp were 40, 25, and 7 
for these same cultures in the order given. It is not clear from the 
data at hand why this exception occurred. 

In order to compare all these data which are fairly comparable, 
table 12 has been constructed. Since two types of solutions were used 
in experiment 6, cultures 2, 4, and 6 were selected as their solutions 

Table 12. — Average dry weight production per unit total radiation (Mazda plus 
mercury lamp) in relation to percentage of radiation from the mercury lamp 

Experiment Culture Dry wt. per radiation from 

number number watt/cm^ mercury lamp 

7 6 37.4 45 

8 3 32.7 25 

6 6 32.4 51 

8 I 28.2 40 

7 I 27.6 29 

6 4 26.8 IS 

7 5 25.2 14 

8 4 24.4 24 

7 4 21.0 5 

6 2 20.3 3 

8 2 19.4 40 

7 3 17-8 5 

7 2 13.7 7 

8 6 ii.o 6 

8 5 10.3 7 

were similar to those used in experiments 7 and 8. All plant values 
given in this table are the averages of three plants. There is a slight 
difference in the duration of the three experiments which should be 
kept in mind in making this comparison. In these experiments, 6, 7, 
and 8, the plants were grown for 20, 18, and 21 days respectively. 

The data showing dry weight produced per unit total energy in 
table 12 have been arranged from greatest to least value. The cor- 
responding values showing the percentages of total radiation that are 
obtained from the mercury lamp fall roughly into two groups. The 
first eight values are high (14 to 51 percent). The remaining seven 
with the exception of culture 2 in experiment 8 are low (3 to 7 
percent). Although there is no regular decrease in these percentage 
values with the decrease in dry weight per unit total radiation, there 


appears to be a general decrease in dry weight efficiency with illumi- 
nation containing less of the shorter wave lengths found in the 
mercury lamp. 


Plants have been grown by Harvey (1922), Hendricks and Harvey 
(1924), and others under Mazda lamps. Davis and Hoagland (1928), 
Arthur, Guthrie and Newell (1930), Garner and AUard (1931), 
Steinberg and Garner (1936), and others have conducted numerous 
experiments in which good growth was obtained with Mazda lamps 
for various lengths of daily light and dark periods. Many other in- 
vestigators both in Europe and in this country have shown that plants 
may be grown in artificial light whose wave-length distribution is 
continuous from blue-violet to red. Other investigators have de- 
termined the growth of plants in different portions of the spectrum. 
Here it was necessary to use glass or liquid filters. Others, like 
Roodenburg (1932), have used gaseous discharge lamps such as 
neon. Most of these experiments indicate the necessity of the full 
visible spectrum for normal growth. Popp's (1926) results indicate 
that the blue-violet end of the spectrum is necessary for normal 
growth although the ultraviolet may not be indispensable. Shirley 
(1929) states that " The entire visible and ultra-violet solar spectrum 
is more efficient for the growth of the plants studied than any portion 
of it used ; the blue region of the spectrum is more efficient than the 
red region." Schappelle (1936) concluded that white light is best for 
normal plant response. Either end of the visible spectrum without 
the other causes abnormal growth. Infrared, between 0.8 fi and 2.0 fi 
was ineffective in bringing about fruiting of Marchantia, while red 
and blue lights were of approximately equal effectiveness. 

Arthur and Stewart (1935) made a comparison of the growth of 
buckwheat plants under Mazda, neon, sodium, and mercury vapor 
lamps. For short periods of 8 to 10 days the sodium lamp was found 
to be most efficient in the production of dry weight. No relation was 
found between the absorption bands of chlorophyll and the emission 
bands of the various lamps. These gaseous discharge lamps produced 
plants with greener leaves than the Mazda lamps. Later Arthur 
and Harvill (1937) show that the sodium lamp alone is not ideal for 
the continuous growth of plants over long periods of time. If, how- 
ever, the continuous exposure from the sodium lamps is supplemented 
by an exposure of 2 hours per day from an 85-watt capillary mercury 
vapor lamp, excellent leaf color and flowering could be produced in 


such plants as begonia, gardenia, cotton, geranium, buckwheat, and 
snapdragon. Although this light source was not satisfactory for the 
tomato plant, the authors point out that other wave bands of light 
may be found which should be added or subtracted for the best 
growth of some plants such as the tomato. 

Dastur and Mehta (1935) determined the rate of photosynthesis in 
approximately equal intensities of red, blue, and white light. Photo- 
synthetic activity was greatest in the white light, intermediate in the 
red light, and least in the blue light. They state that both the red and 
blue regions are necessary for normal photosynthesis. 

Equally interesting are the results of Hoover's (1937) investigation 
on determining the rate of CO2 absorption as a function of wave 
length on the basis of equal incident energy. The principal maximum 
occurred at 6500 A in the red, and a secondary maximum came at 
4400 A in the blue. The greater transmission and reflection of radia- 
tion in the green region decreased the effectiveness in that portion 
of the spectrum. The limits of CO2 absorption were placed between 
7200 A and 7500 A in the red, and below 3650 A in the blue end of 
the spectrum. 

Dastur and Solomon (1937) show the importance of the blue- 
violet end of the spectrum in photosynthesis in a series of experi- 
ments in which plants are grown in the light of a carbon arc, in 
" mixed " light where the gas-filled electric lamp light has super- 
imposed upon it a beam of blue-violet light, and in the light of the 
gas-filled electric lamp alone. The " mixed " light was composed of 
two beams originating in a single source (1,000-watt flood lamp) and 
reflected to the plant by mirrors. One beam was passed through a 
copi^er sulphate filter which limited the wave-length band to the 
region 4200 A to 4720 A. These beams (white and blue) were 
reunited in the proportion 1:1 on an intensity basis. Plants grown 
in these three lights showed greatest photosynthetic activity in the 
carbon arc light, intermediate in the " mixed " light, and least in the 
gas-filled electric bulb light. This follows the order of richness in 
blue-violet light of the three sources. 

From the foregoing discussion it would appear that plants can be 
grown in artificial light, but for more or less normal growth the light 
should include those wave lengths found in the visible solar spectrum. 
An increase in intensity or the absence of a given portion of this 
spectrum brings about abnormal growth responses. Undoubtedly, 
the more nearly the artificial light resembles sunlight in its energy 
distribution, the more nearly normal are the plant growth responses. 


In the experiments reported in the present paper, a method for 
mixing artificial lights was used, but one quite different from that 
used by the above-mentioned investigators. A beginning was made 
by using two light sources, one rich in red, the other rich in blue 
light. By locating the plants on small rotating tables at different 
distances from these light sources, practically any intensity ratio of 
the two could be obtained. With this general scheme the number of 
lights could be increased, thus making it possible to study the effects 
of any given mixture of restricted wave-length regions on the growth 
of plants. With each added light, however, the interpretation of 
data becomes more difficult. By using this method it is difficult to 
grow many duplicate individuals at one time, especially if they grow 
large. This objection may be met in part by repeating an experiment 
often enough to obtain more reliable statistical date. 

The first two experiments with the Mazda light vs. the Mazda light 
filtered through a CUSO4 solution were mostly exploratory in nature. 
There is some indication that the greatest dry weight produced is 
associated with wave-length distribution and not entirely correlated 
with intensity of radiation. Although the data of experiment 3 are 
meager, a considerable difference in growth was obtained between 
plants receiving different amounts of red (wave lengths longer than 
6040 A) and blue (wave lengths shorter than 6040 A). The dry 
weight increase for the plants receiving red-blue light in the ratio 
72 : 28 was about 40 percent over those receiving white light and 
those receiving a mixture in the proportion 96 : 4, although the total 
intensities of these two cultures were over 40 percent greater. 

An attempt was made to change the type of red and blue light by 
the use of neon and mercury grids. In these experiments (nos. 4 and 
5) it was found that the intensity of radiation was too low for good 
growth. This made it impossible to draw any definite conclusion 
regarding the proportion of red to blue that gave best growth. Yellow- 
ing or lack of greenness was associated with those light mixtures 
predominant in red. 

In order to obtain lights of higher intensities, one rich in red, the 
other rich in blue, the water- jacketed projection Mazda lamp used in 
experiments i and 2 and the 400-watt high-pressure mercury lamp 
used in experiment 5 were employed. With this combination of lights 
very good growth was obtained under 100 percent artificial conditions. 
Because of this good growth and the increased number of plants per 
treatment, more weight can be attached to the data from experiments 
6, 7, and 8, than to the earlier ones. Where light and not carbon 


dioxide is the limiting factor, the dry weight increases with increased 
illumination. Hoover, Johnston, and Brackett (1933), working with 
wheat plants, found that in normal air CO2 became limiting at a light 
intensity of about 0.05 to 0.06 watts/cm\ In none of these experi- 
ments with the tomato plant was the intensity greater than these 
values. Although the two plants may not behave exactly alike, it is 
reasonable to suppose they are similar enough to assume that at no 
time was CO2 the limiting growth factor. In order to accentuate 
growth differences due to wave-length mixtures and minimize the 
effect of intensity on dry weight production, the dry weight data 
were divided by watts/cm^ This dry weight efficiency of comparable 
cultures in the last three experiments was used as a criterion of the 
effect short-wave (blue) radiation added to that of longer wave length 
had on plant growth. It would appear from the data given in table 
12 that a greater amount of dry weight is produced with a Mazda 
light by enriching it with blue from a mercury lamp to the extent of 
14 to 51 percent under the conditions of these experiments. Care 
should be exercised in drawing any far-reaching conclusions, for with 
a change in quality or wave-length distribution of the Mazda or other 
source rich in red, changes undoubtedly will be necessary in other 
portions of the spectrum. Although for good growth plants very 
probably tolerate a rather wide range in wave-length distribution, yet 
it would appear that the more nearly this distribution in artificial 
light approaches that of sunlight the better will the plants grow. 


Emphasis is placed on the importance of quality or wave-length 
distribution of light in affecting plant growth. A method and several 
experiments are described in which plants were grown in " mixed " 
lights. By placing the plants on small rotating tables between two 
light sources, one rich in red, the other rich in blue, the proportion 
of each type of radiation falling on each culture was varied by the 
position of the culture with reference to the light sources. 

As found in previous experiments, yellowing of leaves occurred 
in light rich in near infrared. Since this trouble could be corrected 
to a considerable extent by the type of nutrient solution used, it 
indicates the importance of wave-length distribution on the uptake 
of mineral nutrients. 

Excellent growth under entirely artificial conditions was obtained 
with plants grown between a 1,000-watt, water-jacketed, projection 
Mazda lamp and a 400-watt, high-pressure mercury lamp. The posi- 


tions of the plants for good growth were such that from 14 to 51 
percent of the total radiation falling on them came from the mercury 
lamp. In several cases better growth was attained in one mixture of 
wave-lengths than in another where the total intensity was higher. 
However, the relatively high growth efficiency may in part be due to 
an intermittency effect occurring in gaseous discharge tubes such as 
the mercury lamp here used. 


Arthur, John M., Guthrie, John D., and Newell, John M. 

1930. Some effects of artificial climates on the growth and chemical com- 

position of plants. Amer. Journ. Bot., vol. 17, pp. 416-482. 
Arthur, John M., and Harvill, Edward K. 

1937- Plant growth under continuous illumination from sodium vapor lamps 
supplemented by mercury arc lamps. Contr. Boyce Thompson Inst., 
vol. 8, no. 5, pp. 433-443- 
Arthur, John M., and Stewart, W. D. 

1935. Relative growth and dry weight production of plant tissue under 
Mazda, neon, sodium and mercury vapor lamps. Contr. Boyce 
Thompson Inst., vol. 7, no. 2, pp. 119-130. 
Dastur, R. H., and Mehta, R. J. 

1935- The study of the effect of blue-violet rays on photosynthesis. Ann. 
Bot., vol. 49, no. 196, pp. 809-821. 
Dastur, R. H., and Solomon, S. 

1937. A study of the effect of blue-violet rays on the formation of carbo- 
hydrates in leaves. Ann. Bot., n. s., vol. i, no. i, pp. 147-152. 
Davis, A. R., and Hoagland, D. R. 

1928. An apparatus for the growth of plants in a controlled environment. 
Plant Physiol., vol. 3, no. 3, pp. 277-292. 

Garner, W. W., and Allard, H. A. 

1931. Effect of abnormally long and short alternations of light and darkness 

on growth and development of plants. Journ. Agr. Res., vol. '42, 
no. 10, pp. 645-651. 

Harvey, R. B. 

1922. Growth of plants in artificial light. Bot. Gaz., vol. 74, no. 4, pp. 
Hendricks, Esten, and Harvey, R. B. 

1924. Growth of plants in artificial light required for blooming. Bot. Gaz., 

vol. ^^, pp. 330-334- 

Hoover, W. H. 

1937- The dependence of carbon dioxide assimilation in a higher plant on 
wave length of radiation. Smithsonian Misc. Coll., vol. 95, no. 21, 
pp. 1-13. 
Hoover, W. H., Johnston, Earl S., and Brackett, F. S. 

1933. Carbon dioxide assimilation in a higher plant. Smithsonian Misc. 
Coll., vol. 87, no. 16, pp. 1-19. 


Johnston, Earl S. 

1932. The functions of radiation in the physiology of plants. II. Some 
effects of near infra-red radiation on plants. Smithsonian Misc. 
Coll., vol. 87, no. 14, pp. 1-15. 
Johnston, Earl S., and Dore, W. H. 

1929. The influence of boron on the chemical composition and growth of 
the tomato plant. Plant Physiol., vol. 4, pp. 31-62. 
McAlister, E. D. 

1937. Time course of photosynthesis for a higher plant. Smithsonian Misc. 
Coll., vol. 95, no. 24, pp. I -17. 
Popp, Henry William. 

1926. A physiological study of the effect of light of various ranges of 
wave length on the growth of plants. Amer. Journ. Bot., vol. 13, 
pp. 706-736 
Roodenburg, J. W. M. 

1932. Kunstlichtcultuur. II. Over de noodzakelijke van planten en neonbe- 
lichting bij bloemcultures. Med. Wagen. (Nederland), vol. 36, 
no. 2, pp. 1-37. 


1929. Effect of narrow ranges of wave-lengths of radiant energy, and other 
factors, on the reproductive growth of long-day and short-day 
plants. Cornell Univ. Agr. Exp. Stat. Mem. 185, pp. 1-33. 
Shirley, Hardy L. 

1929. The influence of light intensity and light quality upon the growth 
of plants. Amer. Journ. Bot., vol. 16, pp. 354-390. 
Steinberg, Robert A., and Garner, W. W. 

1936. Response of certain plants to length of day and temperature under 
controlled conditions. Journ. Agr. Res., vol. 52, no. 12, pp. 943-960. 

Plate i 

Tomato plants grown for 26 days under the following intensities (watts/cm^). 
Daily illumination was 12 hours. 

Culture Neon Mercury 

I 00023 .00002 

3 00008 .00002 

5 00005 .00007 

Plate 2 

Tomato plants grown for 21 days under the following intensities (watts/cm^). 

Daily illumination was 18 hours. 

Mazda Mercury 

Culture (water-cooled) (400 watt) 

I and 2 0404 .0013 

3 and 4 0172 .0031 

5 and 6 0065 .0067 

The darker green leaves in cultures 2, 4, and 6, due to the added (NH4)S04, 
appear in the illustrations as a deeper shade than those in cultures i, 3, and 5. 


Plate 3 

General arrangement of cultures in experiment 8 on rotating tables placed 
about the two light sources. The Mazda lamp encased in a water jacket is on 
the left and the 400-watt mercury lamp on the right. The small rotating tables 
turned at the rate of 3.4 r.p.m. This prevented phototropic curvature of the 
stems but not of the leaves which, although turgid, appear wilted. 

Plate 4 

Tomato plants grown for 21 days under the following intensities (watts/cm*). 
Daily illumination was 18 hours. 

Mazda Mercury 

Culture (water-cooled) (400 watts) 

1 0071 .0048 

2 0071 .0047 

3 0140 .0046 

4 0144 .0045 

5 0137 .0010 

6 0146 .0010 




C Q. 




n < 


VOL. 97, NO. 2, PL. 2 

General Appearance of Plants in Experiment 6. Grown Between 
Mazda and Mercury lamps 

(For explanation, see pages 5 and 17.) 
































































(With One Plate) 



Curator, Division of Invertebrate Paleontology, 

U. S. National Museum 

(Publication 3447) 



JANUARY 3, 1938 




(With One Plate) 


Curator, Division of Invertebrate Paleontology, 
U. S. National Museum 

*S»»* •••T' 




(Publication 3+47) 



JANUARY 3, 1938 




Curator, Division of Stratigraphic Paleontology 
U. S. National Museum 

(With One Plate) 

From 1921 to 1924 Dr. Edward Sampson, of Princeton University, 
then a member of the United States Geological Survey, examined the 
Pend Oreille mining district which surrounds the southern part of 
Pend Oreille Lake, Bonner County, Idaho. A fossiliferous Middle 
Cambrian series crops out in several of the fault blocks into which 
the district is divided. This was the first Cambrian outcrop discovered 
in that part of North America. Subsequently, other Cambrian areas 
were found in the northwestern United States and the adjacent por- 
tions of Canada, in the extensive area previously thought barren of 
Cambrian strata except for the occurrence in the Lewis and Clark 
Range, west-central Montana. These occurrences were briefly dis- 
cussed in 1934.' Since the Pend Oreille Lake area is isolated from 
other Cambrian outcrops and the stratigraphic succession is clearly 
determined, description of the faunas is desirable. 

Dr. Sampson published a brief summary'' of his findings in the 
district and described the stratigraphy, naming three Cambrian 


Before discussing the Middle Cambrian formations a few words 
descriptive of the underlying Beltian strata are in order. Five Beltian 
formations, totaling more than 30,000 feet, are described beneath the 
Middle Cambrian. This enormous thickness of sediments consists of 
argillaceous sandstone, fine-grained massive sandstone, and hetero- 
geneous beds of quartzite, sandstone, and argillite, with or without a 

^ Resser, C. E., Recent discoveries of Cambrian beds in the northwestern United 
States. Smithsonian Misc. Coll., vol. 92, no. 10, 1934. 

^ Sampson, Edward, Geology and silver ore deposits of the Pend Oreille Dis- 
trict, Idaho. Idaho Bur. Mines GeoL, Pamphlet 31 (mimeographed), 1928. 

Smithsonian Miscellaneous Collections, Vol. 97, No. 3 


calcareous content. Cross-bedding, sun cracks, ripple marks and the 
Other usual Bcltian features characterize the series, but contemporane- 
ous igneous rocks are evidently lacking, w^hich is also the case in the 
Beltian strata of nearby western Montana. 


The Middle Cambrian series of the Pend Oreille region begins with 
a quartzite. which is followed in turn by argillaceous shale and calcare- 
ous formations. No mention is made of younger strata, and the pub- 
lished structure sections indicate that the Cambrian is not overlain 
by other beds. 

Gold Creek quartzite. — The Gold Creek quartzite, which is estimated 
to average 400 feet in thickness, is easily distinguished from the 
Beltian by its coarser grain. Some of the conglomeratic beds contain 
pebbles up to 3 inches in diameter and cross-bedding is a characteristic 
feature. Outcrops are conspicuous because of the resistance of the 
rock to weathering. Unfortunately, the contact of the Gold Creek with 
the Beltian rocks is not clearly exposed so that essential history is 

Rennie shale. — This formation is only 50 to 75 feet thick and con- 
sists of soft olive argillaceous shale, sometimes micaceous. Because 
it is so easily eroded, the Rennie shale seldom crops out. In fact, the 
fossils here described were collected from the bed of a brook. 

The published summary fails to mention the fossiliferous limestone 
nodules present in the Rennie shale. They evidently are about the same 
size and shape as similar nodules obtainable from most Cambrian 
shales in the Cordilleran region. Internally, however, these nodules 
are peculiar as they consist of an odd mixture of brown and blue lime- 
stone. Both sorts occur as distinct masses, sometimes sharply angular, 
but more frequently irregular in shape, and the change from one to 
the other is abrupt. The brown limestone, which looks like a fine 
sandstone, is evidently rather pure calcium carbonate, judging from 
its rapid effervescence, and, since this portion of the rock does not 
scratch steel, it is assumed to be free from silica or sand. On the 
other hand, the blue limestone masses contain sand grains or silica, 
even though they also effervesce freely. Fossils are absent from the 
brown masses but are very abundant in the blue portions. Strangely, 
there are but two species of trilobites in these nodules, the abundant 
Vanuxemella idahoensis and rare examples of Albertella sampsoni. 
Neither species has been recognized in the larger shale fauna, although 
elsewhere these trilobite forms are found together. 


The shale has the species Hsted below. Besides the names given in 
this list there is a poor specimen that seems to be Eocystites. A few 
imperfect specimens of Obolus and a fragment of a shell similar to 
Westonia clla represent the brachiopods. 

Elrathia sampsoni Resser Margaretia angustata Resser 

Elrathia longiceps Resser Schistometopiis typicalis Resser 

Glossoplcura mtcrmcdia Resser Urotheca sampsoni Resser 
Hyolithcs idahocnsis Resser 

Lakeviczv limestone. — ^Where unaltered, two rock types characterize 
this conspicuous and commercially valuable formation. One type con- 
sists of clifF-forming massive beds which vary from nearly pure lime- 
stone to nearly pure dolomite. The other beds are shaly, containing 
thin-bedded, highly fossiliferous limestone. Sampson does not state 
what relative position the two types hold with respect to each other, 
but the thin-bedded and shaly material probably forms the lower por- 
tion of the formation. Metamorphism caused different degrees of 
alteration, some of the beds becoming a crystalline marble. 

Black crystalline limestone from the shaly beds yields an abundant 
fauna which is listed below. 

Acrothcle spcciosa Resser Clavaspidella minor Resser 

Acrotreta nit ens Resser Elrathia idahoensis Resser 

Agnostiis bonnercnsis Resser Iphidclla cf. pannula (White) 

Alokistocare nodulijcrnm Resser Lingidclla idahocnsis Resser 

Alokistocare natalc Resser Pagctia fossula Resser 

Alokistocare nactum Resser Oryctocephalus walcotti Resser 

Alokistocare notatum Resser Utia curio Walcott 

Alokistocare normale Resser Zacanfhoides sampsoni Resser 
Alokistocare nothum Resser 

A small collection of altered rock, presumably from the Lakeview 
formation, contains a pygidium of Glossoplcura. Another lot of im- 
pure dark blue limestone is especially interesting because it contains, 
among other fossils, a species of Tonkinella, unfortunately too poorly 
preserved to illustrate. This is not the Tonkinella-Vike. form described 
below as the pygidium possibly belonging to Utia. 


It has already been pointed out that the faunas of the Rennie shale, 
both in the limestone nodules and in the shale, have no species in com- 
mon. Nevertheless, both must be regarded as one fauna, since else- 
where they occur together. Neither have any species been found com- 
mon between the Rennie shale and the Lakeview limestone. These 
faunas also are elsewhere found intermins:led. From these facts it 


seems that these faunas represent faunal subzones or possibly facies 
developments, but for purposes of correlation it is necessary to treat 
the faunas of the Rennie and Lakeview formations as a unit. 

The fossils in the Rennie shale are clearly related to those in the 
Stephen formation.' Margaretia, Elrathia, Glosso pleura, and the par- 
ticular form of Hyolithes are definite relatives of species in the 
Stephen. On the other hand Vanuxemella and Alhertella are more 
characteristic of the older Ptarmigan formation of the Canadian Rock- 
ies. The Lakeview is also related to the Stephen, particularly by the 
Agnostus, Oryctocephahis, and Zacanthoides. The numerous species 
of Alokistocare are found more commonly in other Middle Cambrian 
formations than in the Stephen. 

Close connection exists between the Lakeview and the Spence shale * 
of southern Idaho. Pagetia and the rare trilobite Utia curio indicate 
that these two formations are identical in age. The other genera, 
both in the Lakeview and in the Rennie, occur in the Spence also. 


The identifiable material is described and illustrated as completely 
as possible. In order to avoid unnecessary printing, locality numbers 
are given with the descriptions and in the plate legend. A full descrip- 
tion of the two localities is given below. 

Locality 37m : Middle Cambrian, Rennie shale ; headwaters North 
Gold Creek, south side of Packsaddle Mountain, east of Pend Oreille 
Lake, Idaho. 

Locality 37n : Middle Cambrian, Lakeview limestone ; cement mine 
just north of Lakeview, Pend Oreille Lake, Idaho. 

MARGARETIA Walcott, 1931 

Plate I, fig. 2 

A number of narrow flexible tubes have a surface roughened by 
elongate depressions typical of Margaretia. Compared with the geno- 
type, M. dorus, as well as species in process of publication, M. angus- 
tata is considerably smaller in size, averaging less than one-fourth the 
diameter of the smaller specimens of the other species. 

Locality 37m. 

Holotype. — U.S.N.M. no. 95019. 

* Walcott, C. D., Amount Stephen rocks and fossils. Canadian Alpine Journ., 
vol. I, no. 2, 1908. 

* Walcott, C. D., Smithsonian Misc. Coll., vol. 53, no. i, p. 8, 1908. 


UROTHECA Matthew, 1899 


Plate I, fig. I 

Long, slender tubes abundant in the Rennie shale are referable to 
this genus. The illustrated specimen appears to have a carina but is 
merely broken in the middle. Faint annulations seem to occur on some 

Locality 37m. 

Holotype. — U.S.N.M. no. 95020. 

HYOLITHES Eichwald, 1840 

Plate I, figs. 57, 58 

A species of Hyolithcs occurs in the shale; unfortunately, most of 
the specimens are poorly preserved. The species is evidently related 
to H. carinafa but is larger, the carina is less pronounced, and the 
operculum has wider wings. 

Locality 37m. 

Cotypes. — U.S.N.M. no. 95021. 

LINGULELLA Salter, 1866 

Plate I, fig. 18 

This shell is nearest to L. isse in shape, but it is a smaller brachio- 
pod. It is possible that this brachiopod is one of the Middle Cambrian 
forms now included in L. desiderata. 

Locality 37n. 

Holotype. — U.S.N.M. no. 95022. 

ACROTHELE Linnarsson, 1876 

Plate I, figs. 6, 7 

This form is most like A. colleni, from which it differs in having 
weaker ribs and growth lines, but more particularly in the narrowness 
of the false area. 

Locality 37n. 

Cotypes. — U.S.N.M. no. 95023. 


ACROTRETA Kutorga, 1847 

Plate I, figs. 3-5 

The generic reference is not certain, for this species differs from all 
descrihed forms of Acrotreta. Recently, similarly constructed species 
have been found in both Lower and Middle Cambrian collections. The 
illustrations present clearly the characteristics of the species. 

Locality 37n. 

Cotypcs. — U.S.N.M. no. 95024. 

AGNOSTUS Brongniart, 1822 

Plate I, figs. 16, 17 

This agnostid is a typical form of the Cordilleran Middle Cambrian. 
The characteristic features place it between A. nwntis Matthew of the 
Stephen formation and A. interstrictus White from the Wheeler shale 
of Utah, A. bonnerensis has also been compared with the undescribed 
species in the Spence shale fauna, from which it differs in possessing 
axial furrows on the pygidium. 

Locality 37n. 

Holotypc and parafypcs. — U.S.N.M. no. 95025. 

PAGETIA Walcott, 1916 

Plate I, figs. 8- 1 1 

P. fossiila is similar to P. clytia from the Spence shale, but dift'ers 
in having a median furrow like P. hootes. The pygidium has short 
axial spines. 

Locality 37n. 

Cotypcs. — U.S.N.M. no. 95026. 

ALBERTELLA Walcott, 1908 

Plate I, figs. 24-26 

The glabella of this species is long and is not expanded much in 
front. The pygidium is wide, much like A. Helena and has a rather 
wide concave border, with a nearly straight posterior margin. The 
spines diverge more than average. 

Locality 37m. 

Holotype and paratypes. — U.S.N.M. no. 95027. 


ALOKISTOCARE Lorenz, 1906 

Plate I, fig. 14 

This species is much Hke A. subcoronatiiin except for its larger size. 
Also the furrows, eyelines, and distribution of relief in the brim are 

Locality 37n. 

Holotypc.—V.S.^M. no. 95028. 


Plate I, figs. 52, 54 

This species has a wide brim, the test is finely and closely granu- 
lated, the brim is striated beneath the test, and two nodes are situated 
in the dorsal furrow a short distance forward of the occipital furrow. 

Locality 37n. 

Holotype and parafypc. — U.S.N.M. no. 95029. 


Plate I, fig. 53 

A. nafale is nearest like A. noduliferum. It has a brim of about the 
same size, and other proportions are similar. Small nodes are also 
present in the rear portion of the dorsal furrow. However, the sur- 
face of A. natale, which is finely granulated, has, in addition, scattered 
larger granules. 

Locality 37n. 

Holotype. — U.S.N.M. no. 95030. 


Plate I, figs. 41, 42 

A. nactum is characterized by a medium brim, on which a rather 
wide flat rim is differentiated by its upturned position. 
Locality 37n. 
Holotype and parafypc. — U.S.N.M. no. 95032. 

Plate I, figs. 51, 55 

This species is more normal than A. nactmn, which it resembles. 
A medium swelling causes the rim to be less even in width throughout. 
Locality 37n. 
Holotype and paratype. — U.S.N.AL no. 95031. 



Plate I, fig. 43 

This species departs considerably from the norm of the genus be- 
cause of its convexity. The test is finely granulated, and only a nar- 
row rim is differentiated by the upturned edge. 

Locality 37n. 

Holotype. — U.S.N.M. no. 95033. 

ELRATHIA Walcott, 1924 

Plate I, figs. 36-40 

A fullness in all parts of the cranidium characterizes this species. 
It is typical of the genus in all respects. The thorax has about 15 

Locality 37n. 

Holotype and paratypes. — U.S.N.M, no. 95034. 


Plate I, figs. 31, 35, 58 

Several cranidia of various sizes are illustrated, thus presenting the 
characteristics of the species. Compared with E. idahoensis, the spe- 
cies is somewhat narrower at the eyes, and the glabella is also tapered 

Locality 37m. 

Holotype and paratypes. — U.S.N.M. no. 95035, 


Plate I, fig. 50 

Compared with E. samps oni, this species has a longer glabella and 
relatively shorter brim; also, the brim is divided more nearly equally 
between the rim and preglabellar area. 

Locality 37m. 

Holotype. — U.S.N.AI. no. 95036. 

GLOSSOPLEURA Poulsen, 1927 

Plate I, fig. 56 

Several pygidia, libragenes, and two incomplete hypostomata, but 
no cranidia, were found in the shale collections. One pygidium is 


figured. It is more like G. hoccar than the Spence shale form because 
the doublure is not so wide. Fusion is carried nearly to the extinction 
of the rib furrows. 

Locality 37m. 

Holotype.—V.S.'NM. no. 95037. 


Plate I, figs. 22, 23 

A small fragmentary granulated cranidium, with a typical Orycto- 
cephalus glabella is tentatively referred to the species. 

The pygidium is nearest like O. reynoldsi, differing in having heavier 
spines and more clearly impressed pleural grooves. 

Locality 37n. 

Holotype and paratype. — U.S.N.M. no. 95038. 

CLAVASPIDELLA Poulsen, 1927 


Plate I, figs. 45, 49 

A number of specimens in the Lakeview limestone evidently belong 
to Clavaspidella. This species is much smaller than any other thus 
far described ; also, both the pygidial axis and the eye lobes are long. 

Locality 37n. 

Holotype and paratypcs. — U.S.N.M. no. 95039. 

UTIA Walcott, 1924 


Plate I, figs. 19-21 

Specific differences are not apparent between the Utia of the Lake- 
view limestone and of the Spence shale, consequently the Idaho form 
is identified as U. curio. 

A peculiar small pygidium characterized by its radiating furrows and 
grooves is tentatively assigned to the species. 

Locality 37n. 

Plesiotypes. — U.S.N.M. no. 95041. 

VANTJXEMELLA Walcott, 1908 

Plate I, figs. 13-15 

This species has stronger pygidial furrows than V. nortia, and also 
larger rear spines. Comparison with the Montana species, V. con- 


tract a, shows that V. idahoensis has the rear spines set wider apart, 
giving the entire pygidium a wider aspect. V. idahoensis has incom- 
pletely fused pleural furrows. 

Locality 37m. 

Cotypes. — U.S.N.M. no. 95042. 


Diagnosis. — Glabella long, occupying nearly the entire cranidial 
length ; tapered slightly. There are four pairs of glabellar furrows and 
the occipital furrow. Fixigenes about half width of glabella. Anterior 
suture slightly divergent. Posterolateral limbs rather short. Eyes 
small, situated slightly behind the midpoint. Eyelines curved back- 
ward, arising opposite anterior pair of glabellar furrows. Brim con- 
sists of rim only. Two deep furrows run forward from the anterior 
angles of the dorsal furrow and separate a central thickened portion 
from the two flat lateral portions of the rim. 

Genotype. — 5". typicalis, new species. 

Nanu\- — (TxtoTos = divided : /iero7ro?== forehead. 


Plate I, fig. 12 

During preparation the important rim was injured because the speci- 
men was thought to be an Elrathia. Fortunately, enough of the rim 
remains to show its features. It will be observed that the glabella of 
6". typicalis is like that of Elrathia sainpsoni because of the four sets 
of furrows. 

Locality 37m. 

Holotype. — U.S.N.M. no. 95040. 

ZACANTHOIDES Walcott, 1888 

Plate I, figs. 27-30 

A small form of Zacanthoides is present in the Lake view limestone. 
The pygidium has spines of nearly equal length, and the thorax has 8 
or 9 segments. Glabellar furrows are short and shallow. 

Locality 37n. 

Holotype and paratypcs. — U.S.N.M. no. 95043. 




Fig. I. Urotheca sampsoni, new species 5 

Portion of a specimen. Holotype, X 2. 

Fig. 2. Margaretia angnstata, new species 4 

Portion of the holotype, X 2. Top portion is of the outside ; 
lower portion, the impression. 

Figs. 3-5. Acrotrcta nitcns, new species 6 

3, 4. Ventral valves, X 4- 
5. Dorsal valve, X 4- 

Figs. 6, 7. Acrothclc speciosa, new species 5 

Two ventral valves, X 2. 
Figs. 8-11. Pagetia jossula, new species 6 

8, II. Cranidia, X 4- 

9, ID. Pygidia and a good cranidium, X 4- 

Fig. 12. Sch'istometopus typicalis, new species 10 

Holotype cranidium. 
Figs. 13-15. J^amixeinella idahocnsis, new species 9 

13. A cranidium. 

14. Pygidium showing the spines. 

15. Cranidia and pygidium. 

Figs. 16, 17. Agnostus botui-crcnsis, new species 6 

16. A good cranidium, X 4- 

17. Holotype pygidium, X 4- 

Fig. 18. Lingiilella idahoensis, new species 5 

Interior of a ventral valve, X 4- 
Figs. 19-21. Ufia curio Walcott 9 

19, 20. Cranidia. 

21. Pygidium referred to the species, X 4- 

Figs. 22, 23. Oryctoccphalus tvalcotti, new species 9 

22. Cranidium referred to the species, X 4- 

23. Holotype pygidium. 

Figs. 24-26. Albertella sampsoni, new species 6 

24. 25. Two cranidia. 

26. Holotype pygidium. 

Figs. 27-30. Zacauthoides sampsoni. new species 10 

27. A small cranidium, X 4- 

28. The holotype. 

29. A large pygidium. 

30. A small cranidium, X 2. 

Figs. 31-35. Elrathia sampsoni, new species 8 

3i> 32, 35. Cranidia of various sizes. 

33. Holotype cranidium (see also fig. 58). 

34. Libragene. 

Figs. 36-40. Elrathia idahocnsis, new species 8 

36-38. Cranidia of several sizes. 

39. Holotype cranidium, X 2. 

40. Entire individual. 


Figs. 41, 42. Alokistocare nactuin, new species 7 

41. Cranidium with libragene, X 2. 

42. Holotype cranidium, X 2. 

Fig. 43. Alokistocare notatmn, new species 8 

Holotype cranidium. 
Fig. 44. Alokistocare nonnale, new species 7 

Holotype cranidium. 
Figs. 45-49. Clavaspidella minor, new species 9 

45. A small cranidium, X 2. 

46. Holotype cranidium, X 2. 

47. Libragene, X 2. 

48. A pygidium, X 4- 

49. A pygidium, X 2. 

Fig. 50. Elrathia longiccps, new species 8 

Holotype cranidium. 
Figs. 51, 55. Alokistocare nothuin, new species 7 

51. Holotype cranidium, X 2. 

55. Partially exfoliated cranidium, X 2. 
Figs. 52, 54. Alokistocare nodulijerum, new species 7 

52. Holotype cranidium. 
54. A small cranidium. 

Fig. 53. Alokistocare natale, new species 7 

Holotype cranidium, X 2. 
Fig. 56. Glossopleura intermedia, new species 8 

Holotype pygidium. 
Figs. 57, 58. Hyolithcs idahoensis, new species. 5 

57. Operculum, and several tubes. 

58. A large example, and cranidia of Elrathia sampsoni. 

Figs. I, 2, 12-15, 24-26, 31-35, 56, 58, are from Loc. 37m. 
Figs. 3-1 1, 16-24, 27-30, 36-55, are from Loc. 37n. 


VOL. 97, NO. 3, PL. 1 

MIDDLE Cambrian Fossils from pend Oreille Lake, Idaho 

(For explanation, see page ii.) 





Assistant Entomologist, New Jersey Agricultural 
Experiment Station, New Brunswick 

(Publication 3448) 



JANUARY 10, 1938 






Assistant Entomologist, New Jersey Agricultural 
Experiment Station, New Brunswick 

(Publication 3448) 



JANUARY 10, 1938 

Z^i Boti) (gaitimovt (Prceef 




Assistant Entomologist, New Jersey Agricultural 

Experiment Station, Netv Brtmsivick 



Introduction I 

I. General structure of the head and mouthparts 2 

II. The proboscis 5 

Structure and musculature 5 

Mechanism of coiling and extension 8 

Comparative structure in lepidopterous families 8 

III. The sucking pump 16 

Generalized structure 17 

Pump of lepidoptera 18 

Mechanism 20 

Comparative structure in lepidopterous families 20 

IV. The labium 24 

V. Summary 26 

Abbreviations used on the figures 27 

References 28 


The mechanism of the feeding apparatus of moths and butterflies has 
been studied by a number of anatomists since Reamur and Latreille, 
but the exact means by which the proboscis is extended has not been 
determined, and it is this problem with which this paper is chiefly con- 
cerned. The morphology of the sucking pump has also engaged the 
writer's attention, as have various other parts of the lepidopterous 

The literature is not extensive and (as the general information on 
the head is contained in most textbooks) there is little need except for 
historical purposes to review the contributions previous to the work of 
Burgess (1880) who was the first worker to describe correctly the 
muscles within the proboscis. Kirbach (1883) wrote on the sucking 
pump of Vanessa io and also on the muscles within the proboscis. In 
1890 Burgess published further information on the structure of the 
head of the milkweed butterfly, followed by Kellogg (1893) on the 
same subject. In 1895 Kellogg showed that the pilifers are labral lobes 

Smithsonian Miscellaneous Collections, Vol.97, No. 4 


and not mandibles. Berlese's " Gli Insetti " (1910) contains some in- 
formation on the maxillary musculature. Tillyard (1923) demon- 
strated that the maxillary lobes forming the proboscis are probably 
the galeae. Weber (1924) has contributed to our knowledge of the 
occipital area and the cervix of certain species. Snodgrass (1935) de- 
scribed the sucking pump of a sphingid. 

This study was made possible only through the invaluable instruc- 
tion and encouragement of R. E. Snodgrass, of the Bureau of Ento- 
mology and Plant Quarantine of the United States Department of 
Agriculture, and I am therefore especially indebted to him. I am also 
indebted to Dr. T. J. Headlee, of Rutgers University, and to Dr. E. N. 
Cory, of the University of Maryland, for their cooperation while at 
their respective institutions. I also appreciate the aid of Dr. A. B. 
Klots, of the City College of New York, in determining microlepi- 
doptera. This study formed the larger part of a thesis submitted in 
partial fulfillment of the requirements for the degree of Doctor of 
Philosophy at Rutgers University. 


The cranium of the lepidopterous head is a relatively simple struc- 
ture showing very few sutures. The clypeus forms an elongate an- 
terior area and is not marked off from the frons. A suture extends 
on each side from the invagination of the anterior arm of the tentorium 
to the antenna fossa in the butterflies and most of the higher moths, 
but is usually absent in the more generalized groups. Whether this 
suture is a true frontal suture is questionable, for, as will be seen later, 
its internal ridge seems to have been developed secondarily for the 
purpose of bracing the cranium against the pull of the antenna muscles, 
which originate on the anterior arms of the tentorium. The internal 
ridge of this suture will be called the antennal ridge (fig. i. A, AR). 
The parietals are large, and in the higher Lepidoptera their size is 
further increased by the great development of the compound eyes. 
The ventral and anterior ends of each parietal are recurved mesally, 
thereby providing between them a recess for the maxillae and the 
labium (fig. 12, B). Posteriorly and dorsally, the parietals merge 
with the occiput, there being no limiting suture. The postoccipital 
suture has a well-developed internal ridge and is itself usually evident 
externally. It limits the dorsal part of the posterior edge of the occi- 
put, the ventral part being limited by the much lengthened hypostomal 
sutures. The invaginations of the posterior arms of the tentorium are 
located in the ends of the postoccipital suture, and since most of each 
hypostomal suture lies in the same dorsoventral line as the lateral part 


of the postoccipital suture, the posterior tentorial pits appear to be 
" higher " in the lepidopterous head than they are in most other insects. 
Internally, this part of the hypostomal suture is marked by a well-de- 
veloped ridge, on which are inserted the ventral intersegmental muscles 
from the thorax. The postocciput and the posterior part of the hypo- 
stoma are either poorly developed or entirely membranous. 

The ventral areas of the parietals are not marked ofi from the sub- 
genal areas by sutures, so it may be said that the pleurostomal and 
anterior part of the hypostomal sutures are nonexistent. Since the in- 
vaginations of the anterior tentorial arms of pterygote insects are 
always found in either the pleurostomal or the epistomal sutures, it 
might be supposed that the furrow extending ventrally from each 
anterior tentorial pit is the pleurostomal suture. Such, however, is not 
the case. This deep infolding is the line along which the clypeus and 
the parietal have been brought into juxtaposition, so that the true 
pleurostomal suture would necessarily be within the infolded area. In 
some groups, as in the Tineidae and the Pyralidae, this infolding is 
not pronounced, but in the butterflies it is extremely well developed. 
From a practical viewpoint, these infolded ridges are continuous with 
the ridges that brace the floor of a sucking pump, and will be de- 
scribed later. 

The cephalic endoskeleton, or tentorium, of moths and butterflies 
presents practically the same structure throughout the order (fig.i A). 
The anterior arms of the tentorium are well developed and are the most 
important part. They are without dorsal arms, and the antennal 
muscles arise directly on them. The anterior arms are attached to the 
posterior bridge, close to the invaginations of the posterior arms. In 
many cases the actual posterior tentorial pits are really large open 
depressions, so that when seen from the inside of the head the anterior 
arms and the tentorial bridge appear to have separate invaginations. 
The posterior bridge is always small and poorly developed, and no 
muscles actually arise on its span. 

The only muscles arising on the anterior arms of the tentorium are 
the antennal muscles and two pairs of muscles affecting the extension 
of the proboscis, which will be described later. In the butterflies and 
in moths having functional mouth parts the tentorial arms are often 
provided with large flanges and ridges, to allow greater attachment 
surface. In moths having degenerate or obsolete mouthparts the ten- 
torial arms are often bulging and thin-walled, especially in the anterior 

The foramen magnum is sharply constricted near the invaginations 
of the posterior tentorial arms, although the degree of constriction 


varies considerably throughout the order. However, the posterior 
bridge of the tentorium is always short. 

The mouth parts of adult Lepidoptera consist of the maxillae, the 
labium, the labrum, and the hypopharynx. The maxillae, as is well 
known, are the most important, their galeae forming the long suc- 
torial proboscis in those which have functional feeding mechanisms. 
Various degrees of degeneracy may be found, until the point is 
reached, as in the males of Thyridopteryx, where the maxillae are no 
longer recognizable as distinct appendages. The hypopharynx of 
moths and butterflies is incorporated in the floor of the sucking pump 






Prb — \- 


Fig. I. — Structural details of the head and feeding mechanism of Argynnis 
and Danaus. 

A, right half of head, mesal view, of Argynnis cybclle, showing endoskeleton 
and floor of sucking pump {SP) formed by hypopharynx (Hphy). B, left 
half of clypeus and base of left maxilla attached to parietal part of head of 
Danaus menippe, anterior view. 

and will be described under that heading. The labrum, like the hypo- 
pharynx, is really a part of the cranium, but since it plays a part in 
the mechanics of feeding it may be described as a mouthpart. The 
lateral lobes of the labrum, called the pilifers, bear against the pro- 
boscis base, and in some butterflies the labrum is sufficiently flexible 
to move as a unit with the proboscis base. In such cases (fig. i B) 
the proboscis base is provided with a knob that fits against the pilifer 
under its fringe of setae. In many moths, however, the labrum does 
not have this function. The labium is evident only as a small triangu- 
lar area bearing the three-segmented labial palpi. These palpi are 
usually so placed that the coiled proboscis can be clasped between 
them and be almost completely hidden from view. The labial palpi 


are capable of some motion, each palpus having usually one or two 
muscles at its base, which enable the palpi to clasp the coiled probos- 
cis or to release it. 

The mouth cavity, or preoral cavity, is defined by Snodgrass (1935) 
as " an external space bounded anteriorly by the epipharyngeal wall 
of the labrum and clypeus, posteriorly by the labium, and laterally by 
the mandibles and the maxillae." The hypopharynx is described as 
lying in this cavity as a tonguelike lobe. The cibarium is that part of 
the preoral cavity which is anterior to the hypopharynx ; that is, the an- 
terior surface of the hypopharynx forms its " floor." The salivary 
meatus is the portion of the preoral cavity which is posterior to the 
hypopharynx, i. e., enclosed between the hypopharynx and the an- 
terior surface of the premeijtum. The median salivary duct pours its 
secretions into this cavity from a small pocket called the salivarium, 
between the labium and the hypopharynx. 

In the Lepidoptera most of the cibarium is incorporated with a part 
of the pharynx in the sucking pump, as will be demonstrated later. 
There is a small portion of the epipharyngeal surface which is not a 
part of the sucking pump, and this part is usually applied against the 
proboscis base. The salivary meatus is practically nonexistent, as the 
hypopharynx has completely lost its lobular character, and there is no 
protrusion of a prementum beyond the salivarium. The hypopharynx, 
in fact, forms most of the " floor " of the sucking pump (fig. i A) as 
a single well-sclerotized piece. 


Savigny long ago discovered that the proboscis of Lepidoptera is 
derived from the maxillae or rather from one pair of the lobes of 
the maxillae. The cardo and the stipes are usually quite distinct, and 
form no part of the proboscis as such ; hence the term proboscis 
should be reserved for the conveying structure itself. Tillyard (1923) 
has produced evidence that the proboscis is derived from the galeae. 

Structure and musculature. — Each half or unit of the proboscis is 
therefore a tube, the lumen of which is continuous with the body 
cavity through the stipes. Each proboscis unit is rendered flexible by 
a series of fine rings separated by membrane, as described by Burgess. 
These rings are absent in nonfunctional proboscides. In the butterflies 
and higher moths the rings are made up of many small flat circles of 
hard cuticula, like small beads set in rows. The food channel is also 
lined with rings, similar but having only about one-third the width 
of the outer rings. Muscles passing obliquely between the rings were 


described by Burgess as effecting the coiling of the proboscis, a find- 
ing verified by Berlese and later writers. 

The cardo in functional maxillae is usually a small flat sclerite just 
anterior to the labial palpi. The stipes varies in shape throughout the 

Fig. 2. — Pressure-producing mechanisms of the maxillae. 

A, base of right maxilla of a swallowtail butterfly, ventral view (morpho- 
logically posterior), showing at Ai a cross-section through the line ab, giving 
appearance of stipes when proboscis is coiled. B, cross-section of stipes of 
Catocala sp., showing appearance of pressure chamber {PC) formed by stipes. 
C, same of Arcyonis alope. D, same of Hcmaris tliysbc. E, same of Danaus 
menippe. F, same of Pieris rapac. G, base of proboscis of Danaus menippe, 
lateral view, showing insertion of posterior tentorial proboscis extensor (ptp). 
H, cross-section of stipes of Hemaris thysbe near insertion of posterior tentorial 
proboscis extensor (ptp). I, cross-section of stipes of Atrytone zabulon, showing 
appearance of pressure chamber. 

families above the Tineidae, but when functional always presents 
fundamentally the same structures. The proximal portion in cross- 
section always has a median flat sclerite continuous with a tubular 
lateral part (fig. 2). This tubular part fits into the recurved ventral 
and anterior ends of the parietal, the lateral edge of the stipes being 


continuous with the parietal. In some cases the curved lateral part of 
the proximal portion of the stipes is membranous, as in Cercyonia 
alope (fig. 2 C). In others the lateral part is heavily sclerotized and 
the tubular half is modified so that in effect there are two tubes set 
side by side with membrane between. This arrangement is found in 
the Pieridae (fig. 2F). The distal portion of the stipes is also re- 
cessed under the parietal. The mesal surfaces of the maxillae bear 
against each other or against a small projection of the labium. 

In addition to the muscles within the proboscis mentioned above 
there are three pairs of maxillary muscles inserting on each stipes 
and originating within the cranium. Two of these muscles originate 



Fig. 3. — Proboscis extensor musculature. 

A, left half of head of Danmis menippe, showing interior by removal of eye. 
B, right half of head, mesal view, of Thyridopteryx ephemeracjonnis (male), as 
exposed by median sagittal cut. C, left half of head of Desiiiia juncralis, showing 
interior by removal of eye. 

on the anterior arm of the tentorium, and the third originates on the 
anterior part of the gena. Of the tentorial muscles, one arises on the 
lateral surface of the anterior arm and inserts on the distal part of 
the stipes, on the median flat sclerite. It is therefore called the anterior 
tentorial proboscis extensor (fig. 3 A, atp). The second muscle origi- 
nates on the mesal surface of the anterior arm of the tentorium and 
inserts near the distal point of the stipes. Its origin on the tentorium 
is always posterior to that of the anterior tentorial proboscis extensor, 
so that the paths of these muscles cross within the head. This second 
muscle is called the posterior tentorial proboscis extensor (fig. 3 A, 
ptp). The genal muscle originates on the anterior part of the gena 
and inserts on the flat mesal sclerite of the stipes. It is called the 
cranial proboscis extensor (fig. 3 A, cp). 


Mechanism of coiling and extension. — It should now be ixDssible to 
understand the functions of these muscles. A study of figure 2 shows 
that by their contraction, the anterior tentorial proboscis extensors 
draw the tubular part of the stipes up against the recurved end of the 
gena. (This action is shown diagrammatically in fig. 2 A.) It will be 
noticed that there is a valve arrangement between the tubular part of 
the stipes and the flat sclerite on which the muscles are inserted. As 
the muscles draw the stipes upward, the valve {vlv) closes, with the 
result that the tubular part becomes a closed cylinder. Thus pressure 
is exerted against the blood within the stipes cylinder as it is forced 
against the recurved flange of the gena. The stipes cylinder forms a 
closed point at its proximal end, and therefore the blood displaced as 
the pressure continues must move outward through the stipes, toward 
the proboscis. The stipital ridge is enlarged at the proximal end of the 
stipes and thus practically covers the lumen of the proboscis unit. The 
posterior tentorial proboscis extensor is inserted on this ridge, and con- 
traction of this muscle not only creates pressure on the blood enclosed 
within the stipes, but also moves the base of the proboscis unit upward, 
which effects a tight seal with the functional mouth (fig. 2 G, H ; 9 D) . 

The blood displaced from the stipes is thus forced out into the lumen 
of each tightly coiled proboscis unit, thereby causing the proboscis to 
unroll. The diagonal muscles within each proboscis unit, described by 
Burgess, cause the proboscis to coil. That blood pressure might be the 
agency for uncoiling the proboscis was first suggested to the writer by 
R. E. Snodgrass, who, in his " Principles of Insect Morphology " 
points out the mechanical analogy of such a mechanism with the toy 
paper snake which a child uncoils by blowing into it. The uncoiling 
action of one proboscis unit is shown diagrammatically in figure 4. 

The mechanism described above is the simplest which the writer has 
seen. In many moths and butterflies the stipital cylinder is further 
modified, but the principle is invariably the same, as may be seen in 
figure 2. The musculature concerned in the extension of the proboscis 
seems to be fundamentally the three pairs of muscles described, but 
in a large number of insects one or two pairs may be absent. However, 
functional maxillae always have at least two pairs. 

Comparative structure in lepidoptcrous families. — The maxillae of 
a number of species representing the more important families were 
examined, primarily to determine the fundamental musculature of the 
lepidoptcrous maxilla. Moths having degenerate or obsolete mouth- 
parts were also studied and, indeed, proved to be one of the most in- 
teresting phases of this investigation. To expose the proboscis ex- 
tensor musculature, a simple procedure is first to make a complete 


median sagittal cut, and then, using either half of the head, to remove 
the compound eye, the brain, suboesophageal ganglion, and the suck- 
ing pump. 

Tineidae: In the females of the common Yucca moth, Pronuha 
yuccasella, all three pairs of proboscis extensors are present, very 
much as described above. There is a single muscle at the base of 
each maxillary palpus and each maxillary tentacle. The common 
clothes moth, Tineola, lacks the cranial proboscis extensors, but other- 
wise its musculature is complete. In certain other Tineidae, determined 

Fig. 4. — Diagram of the action of the proboscis extensor muscles of the right 
halt of the head, mesal view. 

A, proboscis coiled. B, proboscis e.xtended by blood forced into it by com- 
pre.ssion of stipes (see fig. 2). 

to family only, there are no tentorial muscles but only the cranial 
proboscis extensors, a very unusual condition. 

Coleophoridae: A number of coleophorids, determined to family 
only, were found to possess both pairs of tentorial proboscis exten- 
sors, but to lack the cranial proboscis extensors. 

Limacodidae: A single representative, Eiiclea cloris indetenninia, 
was studied. This moth has practically no proboscis, but only two 
very small lobes, each a remnant of a proboscis unit. A single pair of 
tentorial proboscis extensors is all that is left of the maxillary mus- 

Oecophoridae: A species of the genus Agonopterix, with a well- 
developed proboscis, was also examined. All three pairs of extensors 


are well developed, and strands of the anterior proboscis extensor 
have migrated onto the clypeus, so that there appears to be a fourth 
pair of muscles. 

Pyralidae: In the pyralids examined, the cranial proboscis exten- 
sor is often absent. In the common wax moth, Galleria mellonella, 
both tentorial muscles are clearly evident and apparently able to act. 
The proboscis does not appear to be functional, and it is probable 
that the proboscis extensors serve only to move the maxillary palpi. 
The lesser wax moth, Achroia grisella Fab., has only remnants of the 
tentorial extensors, the proboscis being evident only as two short 
lobes, apparently nonfunctional. Two other pyralids, Nomophila 
noctuella and Ephestia kuehniella, have the proboscis well developed. 
The musculature is complete and the cranial muscle is especially well 

In figure 3 C the head of a pyralid, Desmia funeralis, is represented 
with the left eye removed. In this case all three proboscis extensor 
muscles are present, although the anterior tentorial proboscis extensor 
is very small. The cranial proboscis extensor is remarkably large, but 
since it originates on the ocular ridge, a relatively thin structure, it 
may be doubted whether it exerts much force. It may also be seen 
that a large lobe has been formed on each anterior tentorial arm in 
order to accommodate a very large antennal muscle, thereby depriving 
the proboscis extensor muscles of their usual position. Such a sacri- 
fice of feeding structures for nonfeeding structures may be found in 
many moths. 

Tortricidae: The tortricids usually possess all three pairs of probos- 
cis extensors. Figure 10 B represents the head of the common codling 
moth, Carpocapsa poikonella, as seen when opened by a median sagit- 
tal cut. The sucking pump is shown in place, but the tentorial exten- 
sors may be seen just below the pump. 

Psychidae: The male of the common bag-worm moth, Thyridop- 
teryx ephemeraeformis, was studied in this group. These moths were 
found to have an extremely degenerate proboscis (fig. 3 B) repre- 
sented only by two large lobes. Each lobe has a single proboscis ex- 
tensor muscle, arising on the anterior arm. The antennal muscles, 
however, are by far the largest muscles in the head. 

Sphingidae: The feeding mechanisms found in this family are gen- 
erally very well developed. The musculature of the head of one spe- 
cies of sphingid, Sphinx convolvuli, has already been described to some 
extent by Berlese (1910). The proboscis niuculature which he found 
homologizes thus : his no. 190 is the cranial proboscis extensor, his no. 
171 is the anterior tentorial proboscis extensor, and his no. 172 is the 


posterior tentorial proboscis extensor. However, Berlese apparently 
did not examine the insertions of these muscles and makes no refer- 
ence to their functions. In discussing the possible mode of extension 
of the proboscis Berlese followed the suggestion of Burgess, that it is 
unrolled by its own elasticity. 

The proboscis musculature of one. species of sphinx moth, Hemaris 
thysbe, may be taken as generally typical of the family. In this moth 
(fig. ID D) the tentorial muscles are equally well developed and are 
well spaced on the anterior arms of the tentorium. The cranial probos- 
cis extensor is moderately developed and in general the whole ar- 
rangement is well balanced. 

Yet, in some sphingids, there are decidedly inferior proboscis ex- 
tension mechanisms. For example, the sucking pump in Smerinthus 
geminatus is so large that there is very little space left for the probos- 
cis musculature, and the brain also is reduced in size and displaced 
posteriorly (fig. loC). In this moth there is no cranial proboscis 
extensor, and only one tentorial muscle, which appears to be the 
posterior tentorial proboscis extensor judging by its insertion. An- 
other sphingid, Darapsa pholus (fig. 6B), has an extremely large 
cranial proboscis extensor. 

Geometridae: The feeding mechanisms of members of this family 
are weak or degenerate. The proboscis musculature of Haematopis 
grataria is shown in fig. 9 C. All three stipital muscles are present, but 
very weak, especially the tentorial muscles. The cranial proboscis 
extensor is also very short. In the little green geometrids (sub- 
family Hemitheinae) the proboscis musculature resembles that of 

The geometrid Ennomos subsignarius is typical of further degen- 
eracy in this family. Only the tentorial muscles are present and they 
are very weak (fig. 5 B). Each anterior arm is very thin-walled and 
is considerably enlarged. This enlargement, of course, " lightens " 
the head by replacing blood volume with air and is of interest in view 
of the remarkable flights of this insect. Caherodes confusaria re- 
sembles Ennomos in this respect but has a functional cranial probos- 
cis extensor. 

The adults of the spring cankerworm, Paleacrita vernata, differ, as is 
well known, in that the males are winged and the females wingless. 
However, there is practically no difference between the sexes in the 
proboscis and its musculature. The proboscis itself, in both cases, is 
represented by two small lobes. Only a pair of tentorial proboscis ex- 
tensors can be found, and they are very weak. 



Noctnidae: In this family the proboscis and its musculature are 
very well developed. The cranial proboscis extensor here reaches a 
remarkable size and importance, in comparison with the tentorial pro- 
boscis muscles. This is of interest in view of the fact that the cranial 
proboscis extensor is the muscle most frequently absent in cases where 
there is not a full complement of proboscis muscles. The proboscis 
musculature of a large moth, Catocala niihilis, is shown in figure 6 A 
and is in general typical of the family. 

Arctiidae: All examined members of this family have degenerate 
feeding mechanisms. Figure 6 C shows a section through the head of 
a typical species, Apantesis virgo. Both A. virgo and A. vittata have 





Soe Gng- 


atp sip Ptp 

Fig. 5. — Proboscis extensor musculature and the sucking pump. 

A, right half of head, mesal view, of Hacmatopis grataria, as exposed by 
median sagittal cut. B, same of Ennomos stibsignarins. 

all three pairs of proboscis extensors. The posterior tentorial probos- 
cis extensor originates well back on the anterior arm of the tentorium. 
Estigmene acraca and UtetJieisa hella show about the same conditions. 

The members of the genus Haploa exhibit the strongest proboscis 
musculature seen in this family, and there can be but little doubt that 
the proboscis is functional. The members of the genus Diacrisia, on 
the other hand, have the most degenerate feeding mechanisms seen 
in this family. In D. virginica the anterior tentorial proboscis exten- 
sor has been lost and the posterior muscle is very weak. The cranial 
proboscis extensor is still evident. 

In Isia isahella all three pairs of proboscis muscles are present and 
apparently functional. The anterior tentorial proboscis extensor, how- 
ever, has migrated to the antennal ridge, above the anterior arm. 

Saturnoidea: The degeneracy of the feeding mechanism in the 
giant silk moths is so complete that there are few traces left of the 



proboscis musculature. In the large moth Samia cecropia, for ex- 
ample, the proboscis is represented by two small shapeless lobes (fig. 
II B, Prh), associated with which there is a single pair of tentorial 
muscles. The position of these muscles suggests that they may be the 
posterior tentorial proboscis extensors. 

Prh P^^^ »^P V^P /K 


Fig. 6. — Proboscis extensor musculature. 

A, left half of head of Catocala nubilis, showing interior by removal of eye. 
B, same of Darapsa pholiis. C, right half of head, mesal view, Apantesis virga, 
as exposed by median sagittal cut. D, same of Malacosoma americana. 

The tentorium of this moth has developed a peculiar secondary 
function. In addition to having the anterior part of the anterior arms 
thin-walled and bulging, displacing blood with air, the posterior part 
of the anterior arms is tubular and curved to provide a sort of cradle 
for the brain and suboesophageal ganglion. This is done by having 
each arm pass between the brain and the optic lobe on its side. If it 
were not for this the brain would be supported only by the optic lobes. 

In other saturniids remnants of one or both pairs of the tentorial 
proboscis muscles can be found, but the cranial proboscis extensor is 


invariably absent. In Basilona imperalis (Ceratocampidae) the pro- 
boscis still has a tubular shape and shows a food channel. Both ten- 
torial muscles are present, but no cranial muscle. 

In all saturniids examined, the position of the tentorium in the 
head — that is, the length of the hypostomal area below the posterior 
tentorial pits — suggests that the ancestors of these moths had power- 
ful feeding mechanisms with the tentorial muscles well developed. The 
antennal muscles in these moths are always large and powerful, and it 
may be that the great enlargement of the anterior part of the anterior 
arms results from the need for a large base for these muscles rather 
than from an effort to lighten the head. However, the latter purpose 
is unmistakably served. 

Lasiocampidae: The adult of the eastern tent caterpillar, Mala- 
cosoma americana, was studied as an example of a lasiocampid, but its 
resemblance to a saturniid was so complete that no new information 
was obtained. The antennal, proboscis, and sucking pump musculature 
is illustrated in figure 6 D. 

Boinhycidae: Similar conditions were found in the common silk 
moth, Boinbyx mori (fig. 7 A). The proboscis is represented by two 
shapeless lobes. Both pairs of tentorial muscles are present. 

Papilionoidca: Representatives of five families of butterflies were 
studied, and some interesting differences were found. In the Nym- 
phalidae and Danaidae the three pairs of proboscis extensors are most 
perfectly preserved. Figure 1 1 shows the proboscis extensor of Danaus 
menippe, known as the " Monarch butterfly." All three pairs of 
muscles are well developed. The musculature of Vanessa atalanta, the 
red admiral butterfly, closely resembles it, as does also the mourning 
cloak butterfly, Aglais antiopia, and the great spangled frittillary, 
Argynnis cyhela. In certain other Nymphalidae, however, the cranial 
proboscis extensor has been lost. Such butterflies include the viceroy, 
Basilarchia archippiis, and the common grayling, Cerpyonis alope. A 
few species of Lycaenidae were also examined. In Lycaenopsis argio- • 
Ins, the common blue, and Everes cainyntas, the tailed blue, the mus- 
culature consists of the familiar three pairs. In Chrysophanus hypo- 
phleas the cranial muscle has been lost, but both tentorial muscles are 
well developed. 

Thus, in these two families nothing unusual was found. In the 
Papilionidae and the Pieridae, however, no trace was found of the 
cranial proboscis extensor, but instead there was a remarkable migra- 
tion of part of the anterior tentorial proboscis extensor. Inserting on 
each stipes with the anterior tentorial proboscis extensor but arising 
on the clypens, between the anterior tentorial arms, there is a single 



large muscle. This is illustrated in figure 7 B, showing the muscula- 
ture of Pieris rapae. The migrant is marked atp2, and it really passes 
laterad of the posterior tentorial proboscis extensor, its strands insert- 
ing with those of the anterior tentorial proboscis extensor which origi- 
nate on the anterior arm. Papilio polyxenes, P. troilus, and P. inar- 
ccllus show similar conditions. Sometimes there is a continuous band 
of muscle from the clypeus to the anterior arm of the tentorium. 



atp — 


Fig. 7. — Proboscis extensor musculature. 

A, right half of head, mesal view, of Bornbyx niori, as exposed by median 
sagittal cut. B, same of Pieris rapae. C, left half of head of Pieris rapae, 
showing interior by removal of eye. D, right half of head, mesal view, of 
Epargyreus tityrus, as exposed by median sagittal cut. 

In the Pieridae this modification is more complete, with the " mi- 
grant " atp2 originating higher on the clypeus than in the Papilionidae. 
In Pieris protodice, the checkered white, this muscle originates just 
anterior of the antennal socket. The musculature of Pieris rapae is 
illustrated in figure 7 B. Other pier ids examined included Colias eury- 
theme and Anthocharis genutia. 

By eliminating the cranial proboscis extensor, the Papilionidae and 
Pieridae seem to have considerably narrowed the parietal area, or per- 


haps it should he said, increased the extent of the compound eye, yet 
without sacrificing muscular power. In effect, this muscle is here lo- 
cated between the anterior arm of the tentorium and the anterolateral 
surface of the sucking pump, thus utilizing what might be charac- 
terized as " waste space." At the same time, the development of this 
muscle is necessarily limited by the sucking pump and its muscles. 

A fourth pair of muscles should now be described. This pair con- 
sists of one muscle located in each proboscis unit, arising on the stipi- 
tal ridge and inserting in the proboscis base, and called the proboscis 
base muscle (FBin). In direction it is a continuation of the posterior 
tentorial proboscis extensor. Its position in the head of Papilio is 
indicated in figure 9 B, PBm, also in figure 2 G. 

Hesperiidac: Only two species of skippers have been examined: 
Epargyrens tityrus and Atrytone zabulon. In this family the probos- 
cis extensors are short but very well developed. The anterior arms are 
greatly broadened to provide greater attachment surface, while the 
cranial proboscis extensors in Epargyreus have invaded the antennal 
ridge to secure greater attachment surface. 

The remarkable simplicity of the stipital tube in Atrytone is well 
worthy of note. In figure 2 I it is represented in cross-section under 
compression. In this type, closure of the pressure chamber is effected 
directly by the cranial proboscis extensor, as it presses the mem- 
branous stipital ridge against the recurved flange of the parietal. The 
membranous fold labeled F2 becomes much larger as it approaches 
the base of the proboscis unit, at which point its outer portion is firmly 
sclerotized, while its inner lateral section (that is, its morphologically 
lateral section) remains membranous. 


Among the orders of insects equipped with sucking pumps, the 
Hemiptera and the Diptera have received considerable study. Snod- 
grass (1935) has shown that the pump of the cicada is prepharyngeal 
in origin and evolved almost entirely from the preoral cibarium. Simi- 
larly, Jobling (1929) and Snodgrass (1935) have demonstrated that 
the sucking pump in Diptera is derived from the cibarium. In respect 
to the sucking pump of the Hymenoptera, Snodgrass (1935) states 
that " while the morphology of the organ is not entirely clear, .... 
judging from the musculature, it includes without doubt the pharynx 
and the buccal cavity and perhaps the cibarium." It is, therefore, of in- 
terest to determine to what extent the sucking pump of Lepidoptera 
is preoral in derivation. 



Generalised structure. — In the more generalized insects there is 
usually a large preoral cavity bounded anteriorly by the epipharyngeal 
wall of the labrum and clypeus, laterally by the mandibles and the 
maxillae, and posteriorly by the labium. The hypopharynx is sus- 
pended between these organs (fig. 8 A, Pre) and thus divides the pre- 
oral cavity into an anterior food meatus (fm), having the anterior 
wall of the hypopharynx for its floor, and a posterior salivary meatus 
(sm) enclosed between the posterior wall of the hypopharynx and 
the anterior surface of the prementum. 

The food meatus, of course, is not part of the stomodaeum, but 
simply space enclosed by certain mouthparts. It leads to the true 
mouth which marks the beginning of the alimentary canal. A portion 

Mth Hphy 

Fig. 8. — Comparison of the orthopteroid head with the lepidopterous head. 

A, diagram of orthopteroid head (from Snodgrass). B, right half of head, 
mesal view, of Danaus menippe, as exposed by median sagittal cut. 

of the food meatus just before the mouth is used to hold food before 
swallowing and is therefore known as the cibarium (fig. 8 A, Cb). 
The dilators of the cibarium always arise on the clypeus. 

The part of the stomodaeum just inside the mouth is termed the 
buccal cavity (BuC). Beyond the buccal cavity, extending to the 
cerebral nerve connectives, we may distinguish the pharynx (Pliy). 
The dilators of the buccal cavity (dlbc) arise on the clypeus, but the 
dilators of the pharynx (dlphy) arise on the frons. The frontal gan- 
glion (Fr Gng) lies on the anterior wall of the stomodaeum between 
the buccal cavity and the pharynx. The connectives of the frontal 
ganglion (Fr Con) always pass laterad of the dilators of the pharynx. 
From figure 8 A it may be seen that the pharyngeal dilators are thus 
encircled by two nerve rings, outside of which they cannot migrate. 


Pump of Lepidoptera. — Snodgrass (1935) has shown that the suck- 
ing pump of moths and butterflies inchides at least the buccopharyn- 
geal region of, the stomodaeum. This is evidenced by the fact that the 
dilator muscles of the pump are inserted both before and behind the 
connectives of the frontal ganglion, which lies on the dorsal wall of 
the pump. The sucking pump of a butterfly, Danaus menippe, is il- 
lustrated in figure 8 B showing the dilators of the true pharynx in- 
serted on the posterior portion of the pump. Whether the cibarium or 
any portion of the food meatus is also incorporated in the lepidopterous 
sucking pump has therefore been an open question. 

The labrum {Lr) of moths and butterflies is usually described as a 
narrow transverse band at the lower edge of the clypeal region, bear- 
ing the pilifers {Plf) on its lateral extremities (fig. i B). In orthop- 
teroid insects there is a pair of muscles, the compressors of the labrum 
{cplr), originating on the anterior wall of the labrum and inserting 
on the epipharyngeal wall. If the small lobe between the pilifers is the 
labrum, as it appears to be, this pair of muscles exists in the Lepidop- 
tera (fig. 8 B and 9B), and the cibarium then necessarily forms part 
of the anterior section of the pump. 

However, there is certain other evidence that the cibarium is in- 
cluded in the pump, based on the structure of the floor of the pump. 
At the base of the salivary meatus in many generalized insects there 
is a small cuplike depression or pocket into which the median salivary 
duct pours its secretions. This pocket is known as the salivarium 
( fig. 8 A, Slv) . It is supplied with three pairs of muscles, a dorsal pair 
(is) arising on the suspensorial sclerites of the hypopharynx, and two 
ventral pairs, arising on the prementum. In the Lepidoptera, only the 
dorsal pair of muscles, arising on the hypopharynx, may be found. 
Their point of origin is on the floor of the sucking pump (fig. 9 A, B) 
showing that the anterior part of the floor is derived from the hypo- 
pharynx and therefore that this portion of the sucking pump belongs 
to the cibarium. 

In orthopteroid insects the hypopharynx has a pair of retractors 
(fig. 8 A, rhphy) originating on the tentorium. In a geometrid moth, 
Haematopis grataria (fig. 5 A), a pair of muscles was found insert- 
ing on the floor of the pump and originating on the anterior arms of 
the tentorium. Since the ventral dilators of the true pharynx in 
orthopteroid insects pass between the circumoesophageal connectives, 
they could not possibly migrate from the tentorial bridge to the an- 
terior arms. Hence, this pair of muscles in Haematopis must repre- 
sent the retractors of the hypopharynx, and although they may have 
migrated beyond the limits of hypopharynx, their presence, neverthe- 



less, is evidence that the hypopharynx is incorporated in the sucking 
pump. In an oecophorid, Agonopterix sp., a similar pair of retractors 

The highest development of the sucking pump is to be found in the 
Sphingidae, the Noctuidae, and especially in the butterflies. Dilation 
of the pump is produced by the muscles originating on the wall of the 
head ; contraction in the lower moths is produced by the intrinsic elas- 
ticity of the pump itself, but in the above-named groups, bands of 

Fig. 9. — Various structural details of the head. 

A, cross-section of sucking pump of Danaus nienippe. B, mesa! view of right 
half of head and base of right proboscis unit of Papilios'p. C, ventral view of 
salivarium of Lycaenopsis argiolus. D, base of proboscis and sucking pump of 
Papilio sp. as seen from right side. E, left half of head of Haematopsis grataria, 
showing interior by removal of eye. 

muscles passing around the pump are mainly responsible. The " floor " 
of the pump is heavily sclerotized and well braced to withstand the pull 
of the dilating muscles. Figure i shows the pump in a specimen 
cleared in KOH, with a portion of the dorsal wall of the pump re- 
moved. The infolded ridge on each side between the parietal and the 
clypeus can be seen passing under the floor of the pump (HphyR), 
thus providing support. Morphologically, these ridges are more diffi- 
cult to trace. Figure 9 D shows the bracing arrangement of the pump 
of a butterfly, Papilio sp. The infolded ridge between the clypeus 


and the parietal can be seen to merge with the hypopharyngeal ridge, 
which is apparently formed jointly by the epipharynx and the hypo- 
pharynx. In effect, the hypopharynx has contributed the median sur- 
faces of each ridge and the area in the floor of the pump between the 
ridges. The dorsal dilators (is) of the salivarium usually arise on 
these ridges, indicating that at least that much is hypopharynx. 

The dorsal wall of the pump varies greatly throughout the order 
in the particular arrangement of its dilating muscles, but shows in- 
teresting consistency in the relative development of the true dilators 
of the pharynx and the dilators of the cibarium. Moths beginning 
with the Tineidae were examined, but no means was found whereby 
muscles which might be dilators of the buccal cavity could be dif- 
ferentiated from dilators of the cibarium. Therefore, any dilator 
muscles not included in the frontal complex (i. e., encircled by the 
connectives of the frontal ganglion, hence true pharyngeal dilators) 
are labeled as dilators of the cibarium. With very few exceptions, the 
true pharyngeal dilators are restricted to the posterior part of the suck- 
ing pump. It does not follow, of course, that the portion of the pump 
derived from the pharynx is necessarily limited to this area ; it merely 
shows the extent to which the dilators of each part have contributed to 
the musculature of the pump. 

Mechanism. — In figure 12 A the complete musculature of the pump 
of Dauaus nienippe is indicated. The muscles compressing the pump 
are shown in cross-section in figure 9 A also. These muscles are 
arranged in two groups, transverse pump muscles (tpui) and longi- 
tudinal pump muscles (Ipm), with two layers in each group. Figure 
9 A was drawn from a hand-cut section of the pump imbedded in 
parafiin. Focusing through the section showed that fibers of the trans- 
verse pump muscles passed directly into the dilating muscles, indicating 
a possible origin of the compressor muscles from the dilators. 

At the anterior end of the pump, a group of transverse pump muscles 
are often arranged in a distinct group, forming what Burgess (1880) 
called the " oral valve " (fig. 8 B, OVm). Its purpose is believed to 
be to prevent the imbibed juices from escaping when the pump is 
emptied. This arrangement was found to be especially well developed 
in the butterflies and in the Sphingidae. 

Comparative stnicture in lepidopterous families. — A number of un- 
determined tineids were examined, including the common clothes 
moth, Tincola hisellieUa. In this family the axis of the sucking pump, 
that is, a straight line from the anterior to the posterior end of the 
pump, is practically perpendicular to the longitudinal body axis. Such 
a pump is illustrated in figure 10 A. A single pair of pharyngeal 



dilators is encircled by the frontal connectives. The remainder of the 
pump dilators form four or more pairs of well-developed short muscles. 
The pump musculature of the Yucca moth is poorly developed although 
the pump floor is fairly well sclerotized. 

In a coleophorid (fig. lo A) practically the same type of pump was 
observed, except that the dilators of the cibarium were grouped an- 


\ HR 

Fig. 10. — The sucking pump. 

A, right half of head, mesal view, of Coleophora coruscipcnnclla, as exposed 
by median sagittal cut. B, same of Carpocapsa pomonella. C, same of Smcrinthus 
geminatus. D, same of Hcmaris thysbe. 

teriorly into a large median band of muscles, with a large band placed 
laterally on the pump. An oecophorid, Agonoptcrix sp., has a similar 
and well-developed pump. Males of the common bagworm moth, Thy- 
ridopteryx ephemeraeformis, possess a very degenerate pump, with 
the muscles evident but very weak. 

In the Tortricidae the axis of the sucking pump is inclined to a more 
horizontal position. In the codling moth, Carpocapsa pomonella (fig. 
lo B), there is a single large pair of pharyngeal dilators, and the dorsal 


pump wall itself is well supplied with muscles. Ar chips offers nothing 
unusual in either pump or proboscis. 

Excellent development of the sucking pump may be found in such 
pyralids as Desmia funeralis and Nomophila noctuella. The cibarial 
dilators are strong and well spaced, although the pharyngeal dilators 
are limited to a single pair. In other pyralids, such as the common 
wax moth, Galleria mellonella, and the lesser wax moth, Achroia 
grisella, the pump is relatively weak, especially in the last-named 
species. Other species of pyralids were studied, but nothing unusual 
was found. 

S phingidae : Snodgrass (1935) has described the sucking pump of 
a Sphinx moth. In Hemaris thy she the structure of the pump is 
typical of this family (fig. 10 D). The pair of pharyngeal dilators is 
large and set close together, and with the cibarial dilators, provide the 
pump with powerful suction. In one species, Smerinthns geniinatus 
(fig. 10 C), this development of the sucking pump has reached such 
a point that little space is left for the brain and the suboesophageal 
ganglion. The proboscis extensor musculature is also reduced to a 
single pair of extensors. In fact, the anterior arms of the tentorium 
are curved laterally in order to accommodate the expanded pump. In 
Darapsa pholiis the pump is of more moderate proportions, although 
quite well developed. In this species there are two pairs of pharyngeal 
dilators. A mouth valve or oral valve is common in this family. The 
dorsal muscles of the salivarium are also easily found in the sphingidae. 

Geometridae: In this family the sucking pump is generally weak. 
Figure 5 B shows the head of a typical geometrid, Ennomos suhsig- 
narius. There are three pairs of pharyngeal dilators and three pairs 
of cibarial dilators, but all are relatively thin muscles. The dorsal 
salivarium muscles are also evident, although very small. Caberodes 
confusaris shows about the same pump as Ennomos. The sucking 
pump of both the males and the females of the spring cankerworm 
moth, Paleacrita vernata, is very weak, although provided with four 
pairs of dilators. 

The sucking pump of Haetnatopis grataria, in addition to possessing 
a pair of hypopharyngeal retractors (fig. 5 A; rhphy), is of interest 
because of its unusual formation. There are three pairs of pharyngeal 
dilators, the posterior pair originating posterior to the antennae and 
passing between the antenna! nerves (fig. 9E). There are also two 
pairs of cibarial dilators. 

Noctiiidae: Members of this family possess well-developed sucking 
pumps, of which that of Hcliothis ohsoleta (fig. 11 A) is typical. 
Laterally, the pump is provided with a sheet of fibers on each side, the 



posterior bundle of which is shown by the position of the frontal con- 
nective to be derived from the pharyngeal dilators. In addition to 
these sheets of muscle, there are two pairs of dilators on the anterior 
part of the pump. The dorsal salivarium muscles are well developed. 
The sucking pump of Antographa falcifcra resembles that of Heliothis. 
Arctiidae: Moths of this family are provided with poorly-developed 
feeding mechanisms. Figure 6 C illustrates the head of an arctiid, 
Apantcsis virgo. The pump muscles are mere strands, and the floor 
of the pump is but weakly sclerotized. In Isia isabcUa the pump and 




Fig. II. — The sucking pump. 

A, right half of head, mesa! view, of Heliothis obsoleta, as exposed by median 
sagittal cut. B, same of Samia cecropia. C, head of Epargyrcus tityrus, as seen 
with dorsal wall removed. 

proboscis are weak but apparently functional. In the genus Haploa 
the pump is relatively strong. Diacrisia virginica shows the most 
degenerate condition observed in this family, the pump dilators being 
mere strands. Yet, in many arctiids the dorsal salivarium muscles are 
present and probably functional. 

Saturniidae: In this family the sucking pump is extremely weak. 
Figure ii B illustrates the sucking pump of Samia cecropia. There 
is a single pair of pharyngeal dilators, still recognizable by means of 
the frontal ganglion. Laterally, there are two pairs of muscles which 
might be functional. No salivarium muscles could be found. The 


sucking pumps of other saturniids have about the same development 
as in Samia. 

Boinbycidae: The well-known silk-moth, Bombyx mari, also has 
a very feeble sucking pump (fig. 7 A). The remnants of only two 
pairs of muscles are present, one pair being pharyngeal dilators. 

C eratocampidae : The sucking pump of Basilona imperaJis is very 
weak and in general much as in the saturniids. There are two pairs of 
pharyngeal dilators. 

Lasiocampidae: The adult of the eastern tent caterpillar, Malaco- 
soma americana, was studied as an example of this family (fig. 6 D). 
The pump has a single pair of pharyngeal dilators and three pairs of 
cibarial dilators, but all are mere threads. 

Hesperiidae: The skippers have well-developed sucking pumps, not 
unlike those of the butterflies. The head of Epargyreus tityrus is 
illustrated in figures ii C and 7 D. There is only a single pair of 
pharyngeal dilators, most of the contraction being provided by the 
anterior muscles. 

Papilionoidca: The swallowtail butterflies have a large sheet of 
muscle on each side of the pump, as well as a pair of median muscles 
and a pair of pharyngeal dilators. In the family Pieridae the pump 
very much resembles that in Papilionidae, except that two pairs of 
pharyngeal dilators are usually present. The Nymphalidae and Danai- 
dae show one or two pairs of pharyngeal dilators ; in Argynnis cybele 
the dilators of the pharynx originate as two pairs but insert practically 
as one. The sucking pump of Danaus menippe is illustrated in figures 
8B and 12 A. 

However, throughout the families of the Lepidoptera it is probable 
that these pairs of pharyngeal dilators do not represent original pairs of 
muscles immediately homologous with the dilators of the pharynx of 
such insects as Dissosteira. For example, in Dissosteira there is a pair 
of retractors of the mouth angles encircled by the frontal connectives, 
but it is improbable that any of the muscles encircled by the frontal 
connectives in Haematopsis are exactly homologous with the retractors 
of the mouth angles (see fig. 8 A, rao). 


The structure and limits of the labium in adult Lepidoptera have 
been previously described by other writers, most recently by Snodgrass 
(1935). In figure 12 B the labium of Hemaris thyshe is illustrated. 
In this case the labium is limited to a median strip passing to the base 
of the proboscis, and a small area around each labial palpus. Pos- 
teriorly, the labium is supported by a hypostomal bridge (HBr.). 



In the yucca moth, Pronuha yuccasella, there is a small paired fleshy 
lobe at the distal end of the labium. The Oriental fruit moth, Grapho- 
litha inolesta, also has a pair of minute lobes at the tip of the labium, 
but it is improbable that these lobes have any significance. 

The labium of many moths and butterflies possesses a strong ventral 
ridge at the distal end (figs. 11 A and 9 C, Keel). A possible function 
of this ridge is to serve as a bearing surface for the proboscis base. 

c^? sp ji^j,^^ 

Fig. 12. — Various structural details of the head. 

A, head of Danaus menippe, as seen with dorsal wall removed. B, labium and 
basal part of maxillae, ventral view, of Heviaris thysbe. C, labial palpus mus- 
culature of Papiiio glauctis. D, same of Danaus menippe. E, same of Pieris 

The musculature of the labium is limited to the palpi muscles. Ber- 
lese and Burgess have figured palpus muscles arising on the tentorium, 
but apparently did so by mistaking proboscis extensors for palpi 
muscles. In every moth and butterfly examined by the writer there 
were never more than two pairs of palpus muscles, and these arise 
either on the labium itself or on the hypostomal bridge. The articula- 
tion of the labial palpus with the head is so formed that little or no 
blood passes out into the palpus. If the palpus of a live butterfly is 
snipped off, it will be found that the walls of both the first and the 
second joints are barely moist inside. This, of course, greatly lightens 
the palpi. 


The palpus musculature of Danaus menippe is illustrated in figure 
12 D. Each palpus is supplied with two muscles, a levator {Iplp) and 
a depressor (dplp). Nymphalidae usually have a pair of muscles for 
each palpus, although the depressor is often very weak. In the Papili- 
onidae and the Pieridae there is no depressor muscle, but the levator 
is always well developed (fig. 12 C, E). In this case the levator arises 
either on the sclerotized median plate (MP) of the labium, or on the 
hypostomal bridge. 

The presence or absence of labial palpus muscles is extremely 
variable in the other families. Pronuha yuccasella has no palpus 
muscles, nor does an Agonopterix sp. Each palpus of Galleria inel- 
lonella has a single large muscle. Geometridae show only a single 
palpus muscle or none at all. In the Arctiidae there is only a single 
muscle, usually arising on the hypostomal bridge. Sphingidae have 
either one or two pairs of palpus muscles, commonly only one. Satur- 
niids and other extremely degenerate groups usually lack any palpus 
musculature, and the trembling motion of the palpi sometimes seen in 
this family is usually caused by the remnants of the proboscis exten- 
sors. However, BasUona imperaUs has a single muscle in each palpus, 
as does also Malacosoma americana. 


1. The coiled proboscis of Lepidoptera is extended by means of 
blood pressure created in the stipes of each maxilla. This pressure is 
caused by three pairs of muscles, which by their contraction press the 
stipes against the head wall. Two pairs of these muscles arise on the 
anterior arms of the tentorium and the third pair arises on the gena. 

2. The sucking pump is a compound organ, derived from the 
pharynx, the buccal cavity, and the cibarium. This is evidenced by 
these facts : ( i ) true pharyngeal dilators are inserted only in the 
posterior part of the pump; (2) muscles homologous with the com- 
pressors of the labrum are present in some Lepidoptera; and (3) the 
dorsal salivarium muscles arise on the pump floor, showing that the 
hypopharynx forms at least the anterior part of the floor. 

3. There is no labial musculature except that of the palpi. There 
are generally two pairs of palpus muscles, but in many families only 
one pair, or none at all, may be found. 

4. The area posterior to the labial palpi is bounded by the hypos- 
toma, the hypostomal ridge offering an insertion for the ventral inter- 
segmental muscles. A hypostomal bridge is sometimes present. 

5. The anterior arms of the tentorium are well developed but lack 
dorsal arms. The posterior tentorial bridge is short and weak. The 



great length of the hypostoma in Lepidoptera elevates the tentorium 
to a higher position in the head, with respect to other cephalic struc- 
tures, than is common. 

6. The antennal muscles arise on the anterior arms of the tentorium 
and vary in number from one to five pairs. They are always well 
developed, sometimes at the expense of other head structures and, in 
moths with obsolete feeding structures, are often the only functional 
muscles within the head. 


Ant, antenna. 
ant mcl, antennal muscle. 
Ant Nv, antennal nerve. 
^7?, antennal ridge. 
AT, anterior tentorial arms. 
at, invagination of anterior arm. 
atp, anterior tentorial proboscis ex- 
BuC, buccal cavity. 
Br, brain. 
Cd, cardo. 
Clp, clypeus. 

cp, cranial proboscis extensor. 
Cv, cervix. 

dlhc, dilator of buccal cavity. 
dlcb, dilator of cibarium. 
dplp, depressor of palpus. 
dlphy, dilator of pharynx. 
E, eye. 

fm, food meatus. 
Fr, frons. 

Fr Con, frontal connective. 
Fr Gng, frontal ganglion. 
HBr, hypostoma! bridge. 
Hphy, hypopharynx. 
HphyR, hypopharyngeal ridge. 
HR, hypostomal ridge. 
hs, hypostomal suture. 
Hst, hypostoma. 
Lb, labium. 
Lb Pip, labial palpus. 
Iplp, levator of palpus. 
Ipm, lateral pump muscle. 
Lrm, labrum. 
MP, median plate. 
Mth, mouth. 

Nv, nerve. 
Oc, occiput. 
OcR, occipital ridge. 
ocs, occipital suture. 
Oe, oesophagus. 
OVm, oral valve muscle. 
PBm, proboscis base muscles. 
PC, pressure chamber. 
Phy, pharynx. 
Plf, pilifer. 
Pip, palpus. 
Poc, postocciput. 
PoR, postoccipital ridge. 
pos, postoccipital suture. 
Prb, proboscis. 
Prb Ext, proboscis extensor. 
Prtl, parietal. 

PT, posterior tentorial arms. 
pt, invaginations of posterior arms. 
ptp, posterior tentorial proboscis ex- 
rao, retractor of mouth angles. 
rhphy, retractor of hypopharynx. 
rn(RNv), recurrent nerve. 
IS, anterior salivarium muscle. 
2s, 3s, posterior salivarium muscle. 
SID, salivary duct. 
Slv, salivarium. 
sin, salivary meatus. 
Soc Gng, suboesophageal ganglion. 
SP, sucking pump. 
SR, stipital ridge. 
St, stipes. 

tpm, transverse pump muscle. 
Tr, trachea. 
vlv, valve. 


Berlese, a. 

1910. Gli insetti, vol. i. Milan. 
Burgess, E. 

1880 a. The structure and action of a butterfly's trunk. Amer. Nat., vol. 14, 

PP- 313-319- 
1880 b. Contribution to the anatomy of the milkweed butterfly. Anniv. 
Mem. Boston Soc. Nat. Hist. 


1929. A comparative study of the structure of the head and mouthparts in 
the Streblidae (Diptera Pupipara). Parasitology, vol. 21, pp. 417- 
Kellogg, V. L. 

1893. The sclerites of the head of Danaus archippus. Kansas Univ. Quart., 
vol. 2, pp. 51-59. 

1895. The mouthparts of the Lepidoptera. Amer. Nat., vol. 29, pp. 546-556. 


1883. Uber die Mundwerkzeuge der Schmetterlinge. Zool. Anz., vol. 6, pp. 
Snodgrass, R. E. 

1932. Evolution of the insect head and the organs of feeding. Ann. Rep. 

Smithsonian Inst, for 1931, pp. 443-489. 
1935. Principles of insect morphology. McGraw-Hill Book Co., New York. 
Tillyard, R. J. 

1923. On the mouth-parts of the Micropterygoidea. Trans. Ent. Soc. Lon- 

don, vol. 71, pp. 181-206, 12 figs. 
Weber, H. 

1924. Das Thorakalskelett der Lepidopteren. Ein Beitrag zur vergleichenden 

Morphologic des Insektenthorax. Zeitschr. Anat. und Entwick., 
vol. 72>, PP- 277-331, 9 figs. 




(With 46 Plates) 



Chief, Bureau of American Ethnology 




JULY 22, 1938 



rhotograph by Barry 





(With 46 Plates) 



Chief, Bureau of American Ethnology 

(Publication 3482) 



JULY 22, 1938 

€^e &ovi Q0afttmor« (prceet 




Chief, Bureau of American Ethnology 

(With 46 Plates) 


The name of Sitting Bull will probably always remain as the best 
known of any American Indian. AVhether or not this preeminent place 
is deserved, it is a fact that more has been printed about him than 
any other Indian and his name has most intrigued the popular imagi- 
nation. Part of this notoriety resulted from the fact that he was a 
prominent and influential Indian during a crucial period in the 
history of his tribe and partly from the fact that during the latter 
years of his life he was exploited, both in this country and abroad, 
in a manner calculated to bring his achievements in a highly colored 
manner before the general public/ 

It is not the purpose of this introduction to outline the career of 
Sitting Bull. This has been adequately done by a number of biogra- 
phers, and the interested reader is referred to the attached selected 

There is no question concerning the fact that Sitting Bull was a 
great man, in spite of the fact that many of his contemporaries at- 
tempted to belittle his character. It is true, however, that the promi- 
nence he later achieved in the popular mind was partly due to 
circumstances which gave unusual publicity to his career. 

In view of his status with the government, Sitting Bull was always 
reluctant to speak with white men regarding his personal adventures. 
However, in keeping with the custom of his people, he was proud 
of his war exploits and kept a careful record of them. 

^ In an luipublished manuscript on Sioux names by the famous scout and 
interpreter, E. H. Allison, the author says of the Indian name of Sitting Bull, 
Tatanka Yotanka ; "Sitting Bull's totem was a bull standing in a defiant attitude, 
which clearly expressed the meaning of his name, 'The Bull in Possession,' 
'The Conquering Bull,' 'The Bull of Occupation,' The Sitting Bull.' " 

Smithsonian Miscellaneous Collections, Vol.97, No. 5 


When Sitting Bull recounted his honors at the dance following 
the Sioux victory over the Crow in 1870, Frank Grouard, who was 
present, states that at this time Sitting Bull was entitled to 63 coups." 
About this same time Sitting Bull made his pictographic record after 
the usual manner of the Plains Indians, representing the feats which 
entitled him to special credit among the Indians. This set of drawings 
he gave to his adopted brother. Jumping Bull, who placed with them 
a pictographic record of his own. While these drawings were in the 
possession of Jumping Bull, Four Horns copied 55 of them, includ- 
ing 40 from the record of Sitting Bull and the remainder from that 
of Jumping Bull. In some manner, not yet explained, these copies 
fell into the hands of another Indian who brought them to Fort 
Buford, where they eventually came into the possession of Assistant 
Surgeon James Kimball in August 1870. The fate of the original 
drawings from which this set was made is not known, although 
Sitting Bull stated that they were still in the possession of Jumping 
Bull as late as 1881. According to Col. H. M. Morrow, his father, 
also Col. H. M. Morrow, who was with Dr. Kimball, procured an 
identical set at the same time. These copies were both drawn on 
roster sheets of the Thirty-first United States Infantry. The copy 
retained by the Morrow family was destroyed in San Francisco in 
the great fire of 1906. The copy obtained by Dr. Kimball was de- 
posited by him, together with explanations of the pictures obtained 
at the time from other Indians, with the Medical Director's Office, 
Department of Dakota, on March 14, 1871. The same year they 
were transferred to the Army Medical Museum in Washington, 
D. C. On May 15, 1915, Dr. D. S. Lamb of the Army Medical 
Museum transferred them to the archives of the Bureau of American 
Ethnology, where they are at the present time. 

From time to time this pictographic record has attracted consider- 
able attention, but it has been reproduced only in part, and the 
supporting documents concerning it have never before been published. 

Although the name of Sitting Bull had already become well known 
to the whites, he did not become a figure of outstanding national 
interest until after the annihilation of General Custer and the Seventh 
Cavalry, June 25, 1876, in which battle Sitting Bull participated. 

''De Barthe, 1894, p. 105. 


As news concerning the details of the Custer defeat slowly filtered 
in from the northern plains, newspaper men realized that they had 
the biggest news story since the Civil War. The fact that the colorful 
Custer, who had become something of a national idol, was the central 
figure in the tragic affair, made the story ideal from the standpoint 
of the journalists. In search of material which could be tied in with 
the Custer fight, a Washington correspondent learned of the existence 
of the copy of Sitting Bull's autobiography, then in the Army Medical 
Museum. Sitting Bull was known to have participated in the battle. 
Here then was the perfect nucleus for a follow-up story. On July 6, 
1876, the New York Herald published a highly colored account of 
the pictographic record, neglecting to note that the autobiography 
was not the original handiwork of Sitting Bull. The record was cited 
as proof of Sitting Bull's cruelty, lust for battle and vainglorious 
boastfulness. This story was a huge journalistic success. It was 
copied and revamped by newspapers and magazines throughout the 
United States. Sitting Bull, who heretofore in the public mind had 
been but one of a group of hostile chiefs resisting the westward ad- 
vance of the whites, now became Public Enemy Number i and a 
character of outstanding interest. 

Apparently, an introduction written by Dr. Kimball formerly ac- 
companied the explanatory index and the Williamson letter of verifi- 
cation which are now with the pictographic record. 

The news release from Washington of July 6, 1876, as published 
by the New York Herald says : 

Among the many ghastly souvenirs preserved at the Army Medical Museum 
of this city is an autobiography of Sitting Bull, gotten up in the highest style 
of the art of savage picture history, and telling, in fifty-five drawings or 
sketches, the story of his life down to 1870. Each picture is rudely outlined 
in ink, the men, horses and other objects being such as children would make. 
Many of them arc partly filled in with red and blue colors as if Sitting Bull 
had at some time got possession of f)ne of the red and blue pencils so well 
known in newspaper offices, and with it elaborated his pictorial efforts. Blood 
or a wound is indicated by a red blotch with streamers falling down from it. 
The blue is used generally in indicating the white man's pantaloons. Each 
picture is made on a sheet of paper eight by ten inches, and is pasted into a 
book of blank leaves, such as are used for a scrap book. By holding the sheets 
up to the light it is seen that they are the muster-roll blanks of the 31st United 
States Infantry, of which Col. de Trobriand was the commandant. The papers 
probably fell into Sitting Bull's hands at the evacuation of a camp, or, as is 
more likely, were stolen by him during a visit to some of our outposts. Sitting 
Bull is not- at all modest in committing to posterity the story of his great deeds. 
Whether it be the scalping of a soldier in battle or the sly theft of a mule, he 
brags equally of his prowess in his curious autobiography. This literary work, 
which is now likely to be famous, fell into the hands of Assistant Surgeon 


James C. Kimball, of the army, in the month of August, 1870, while he was 
stationed at Fort Buford, Dakota Territory. He had the pictures translated, 
and sent them, with the translation and an index, to the Curator of the Army 
Medical Museum, Washington, Surgeon George A. Otis, United States Army, 
who has filed them in book shape, among the archives of the Aluseum. The 
introduction, written by Dr. Kimball goes, on to say that the autobiography 
contains a description of the principal adventures in the life of Sitting Bull, 
who is an Unk-pa-pa chief. It was sketched by himself in the picture language, 
in common use with the Indians. Since the establishment of Fort Buford, in 
1866, Sitting Bull, at the head of from sixty to seventy warriors, had been 
the terror of mail-carriers, wood-choppers and small parties in the vicinity of 
the post and from 100 to 200 miles from it either way, up and down tlic 
Missouri River. During the time from 1866 to 1870, when the autobiography 
was written, this band had several times captured and destroyed the mail and 
had stolen and run off over 200 head of cattle and killed near a score of white 
men in the immediate vicinity of the fort. The Unk-pa-pas are a tribe of the 
great Sioux Nation, living in tlie Yellowstone and Powder River countries. 

The book was brought into Fort Buford by a Yanktonnais Sioux, and offered 
for sale and purchased for $1.50 worth of provisions. The Indian gave con- 
flicting statements regarding the manner in which he came into possession of 
the book, exciting suspicioi: that he had stolen it from Sitting Bull, who in his 
turn, undoubtedly stole the book in blank from the whites. 

In an article over the name of Porte Crayon published in the 
supplement to Harper's Weekly of July 29, 1876, the editor says: 

About the year 1870 a collection of M.S. drawings, put up in book form, 
bearing the autograph of Sitting Bull and exhibiting a record of his exploits 
and adventures, was brought into Fort Buford by a Yanktonnais Sioux and 
sold for a dollar and fift}» cents worth of provisions. When cross-questioned 
regarding the ownership of the book, the Indian shuffled and prevaricated sci 
as to confirm the belief that he had stolen it from Sitting Bull himself. The 
authenticity of the work, with its general historical accuracy, is confirmed b)- 
Assistant Surgeon James C. Kimball, U.S.A., who, with the aid of interpreters, 
Indians, and others versed in the picture-language of the Northwestern tribes, 
wrote a detailed explanation of the scenes represented, accompanied by a brief 
sketch of the warrior-artist's life. The book was then forwarded to the Superin- 
tendent of the Army Medical Museum at Washington, who placed it in the 
hands of the present editor. 

The series consists of fifty-five designs, drawn on the blank side of printed 
rosters of the Thirty-first United States Infantry, of uniform size (about eight 
by ten inches), clearly outlined with a pen and a brown ink resembling sepia. 
There is no attempt at shading, but the outlines are filled in with flat tints, 
very crudely laid on, with red and blue chalk, yellow ochre, green, and the 
same brown ink or pigment used in the outlines. The coloring, which is quite 
appropriate in the dress and trappings of the human figures, is rather florid 
in the animals. Thus while there seems to be great care in showing the 
characteristic spots and markings of the horses and mules, the sorrels are 
represented with bright yellow, the grays with blue, the bays red, and the 
browns and blacks with the aforesaid brown ink. 


This coloring, however, serves to impart Hfe and meaning to the designs, 
to relieve the groupings from confusion, and is sometimes so arranged as to 
produce quite an artistic effect of chiaro-oscuro. It may be further noted that 
there is no attempt at foreshortening, the objects and figures being all shown 
in flat profile, and without exception, all looking and moving in the same 
direction, that is, from right to left. 

Of all the objects presented by the artist, the figure of the buffalo bull is 
elaborated with the most intelligent and loving minuteness. The horses and 
mules are drawn with a free and well-assured hand, with a tendency to manner- 
ism, relieved somewhat by distinctive character in color, markings, and details. 
He is least happy in his delineations of the human figure, draperies, and 
accoutrements, although in some scenes his attitudes are spirited and his 
costumes sufficiently marked to enable us to identify the sex and country of 
those who have had the honor to sit for their portraits to this distinguished 

The information in the two foregoing newspaper accounts con- 
cerning the manner in which the pictographic record was obtained 
at Fort Buford presumably was obtained from Dr. Kimball's now 
missing introduction. The article published by Harper's Weekly 
reproduces ii of the drawings with a rather detailed description of 
the set based partly on the Kiml^all index and partly upon specula- 
tion by the editor. 

As already indicated, the pictures are drawn on the reverse side 
of loose-leaf roster pages of the Thirty-first United States Infantry. 
The numbers were subsequently placed on them arbitrarily without 
regard to the actual chronology of the events described.^ 

In 1881 the pictures, together with the Kimball index, were for- 
warded through Col. George S. Andrews to Rev. John P. Williamson, 
missionary with the Sioux, who showed them to Sitting Bull for 
purposes of verification. The results obtained from this interview 
are explained in the following letter : 

Fort Randall, Dakota Ter. 
Dec. 12, 1881. 
Col. Geo. L. Andrews, 
25 U. S. Infantry, 
Commanding Post, 

Sir : 

I have the honor to state that in connection with Capt. G. Lawson, I inter- 
viewed Sitting Bull in regard to the supposed "Hyeroglyphic Autobiography" 
of himself, contained in pictured sketches, numbered i to 55, obtained by Jas. C. 
Kimball, Ass't. Surgeon, U.S.A., in the year 1870. 

' Vestal, in describing them, has placed them in what he considers to be the 
order in which the different feats took place. 


Sittting Bull immediately recognized the pictures as scenes from his early 
life, with the exception of Nos. 39 to 51, and 53 and 54, which he said were 
not his, but were adventures of his brother Jumping Bull. 

As to the scenes from his own life, he says these are all true scenes, and 
he drew a similar set many years ago and gave them to his brother Jumping 
Bull. He saw his brother last summer and understood from him that he still had 
them. He thinks therefore that this set must be a copy of the one he made, 
and has been drawn off by some Indian, he does not know by whom. He 
could tell perhaps by seeing his brother who is at Standing Rock. 

Sitting Bull verified in the main the Index accompanying the pictures. 

No. I he says was his first feat, accomplished when he was fourteen years 
of age. 

No. ID he says was a Ree, who drops his gun and bow from fear. He was 
struck (for "coup") but not killed (no blood is shown). The scalp at the 
horses bridle, here and elsewhere, not being intended to represent the scalp 
of the enemy drawn. 

No. 55 he says is not completed — should have his "name" (as he calls the 
sitting buffalo). 

As to the particular history of each event recorded, we found Sitting Bull 
rather reserved, especially in regard to Scenes Nos. 11 to 26, and we could 
see that any narration he gave of the several events was colored by the circum- 
stances of his present situation. And I would suggest that if a more full account 
of his war deeds is desired, a better time to secure it would be at some future 
date when his status is definitely determined. 

Yours Respectfully, 

(signed) John P. Williamson, 

In reproducing the drawings, the explanation of each is given 
exactly as written in the Kimball index in 1870. It should be borne 
in mind that these interpretations were furnished by Indians familiar 
with the career of Sitting Bull but not by the Sioux warrior himself. 

For purposes of comparison these explanations are supplemented 
by the interpretations published by Vestal.* 

No. 54 is missing from the set. This picture was one of the Jump- 
ing Bull series and represented an episode in the famous battle of 
1870 between the Sioux and the Crow. The Kimball description 
says "Sitting Bull at the head of his band charges into a camp of 
Crows and kills thirty of them. (This happened in the winter of 

* Vestal says, "For Sitting Bull's interpretations of these drawings, given in 
1885, I am indebted to Mr. Seth C. Jones, Secretary, Municipal Art Com- 
mission, Rochester, N. Y." 


No. I 

"Sitting Bull, a young man without reputation and therefore wear- 
ing no feather, engages in his first battle and charges his enemy, a 
Crow Indian who is in the act of drawing his bow, rides him down 
and strikes him with a 'coup' stick. Sitting Bull's autograph — a 
buffalo bull sitting on his haunches — is inscribed over him. His 
shield suspended in front has on it the figure of an eagle which 
he considers his 'medicine' — in the Indian sense of the term." ^ — 

* See Williamson letter, p. 7. For detailed circumstances of this exploit, 
see Vestal, 1932, p. 13. 

"1846. On Red Water. The boy Sitting Bull, as yet an unfledged 
warrior, is shown on horseback, charging an enemy whom he strikes 
with a coup stick. On his blue shield a black bird is painted, anrl 
four black-tipped eagle feathers flutter from the edges of the shield." 
— Vestal. 

No. 2 

"Sitting Bull wearing a war bonnet is leader of a war party who 
takes a party of Crows consisting of three women and a man, so 
completely by surprise that the man has not time to draw his arrows 
from the quiver. Sitting Bull kills one woman with his lance and 
captures another, the man meanwhile endeavoring to drag him from 
his horse, from which it is supposed he is forced to desist by others 
of the war-party. The fate only of Sitting Bull and his victims is 
given in this history." — Kimball. 

"1858. Rainy Butte. This picture commemorates the capture of 
three Crow women, at the time when Sitting Bull's father was killed. 
Sitting Bull carries the lance made for him by his parents, and wears 
a bonnet with horns and a long trail of eagle feathers. A Crow 
warrior is represented as trying to arrest his charge." — Vestal. 



No. 3 

"Sitting Bull pursuing his enemy, a Crow Indian whom he strikes 
with his lance."- — Kimball. 

"1856. On Yellowstone River. Sitting Bull counts coup with his 
lance on a mounted Crow warrior who carries a shield and a gun. 
As required by the obligations belonging to his shield, Sitting Bull 
wears his hair in a knot like a horn on his forehead." — Vestal. 

No. 4 
"Lances a Crow woman." — Kimball. 

"i860. Sitting Bull counts coup on a Crow woman riding a mule. 
She turns to fend off his lance as he strikes at her. This happened 
when the Sioux encountered Crow hunters among the buffalo herds 
and Makes-the-Enemy killed two Crow women." — ^Vestal. 


No. 5 
"Lances a Crow Indian." — Kimball. 

"1853 (?). Sitting Bull unhorses a Crow warrior with his lance. 
The story is well known, but no eye-witnesses now live, and the 
date and place are uncertain." — Vestal. 

No. 6 

"Sitting Bull twice wounded, and unhorsed. His enemy, a Crow, 
at length killed by a shot in the abdomen and his scalp taken and 
hung on Sitting Bull's bridle." '" — Kimball. 

^^ Regarding the mention of scalps in this and succeeding pictures, see William- 
son letter, p. 7. 

"1856. On Porcupine Creek. Sitting Bull, shown wearing his 
Strong Heart bonnet and sash, crouches behind his shield and shoots 
a Crow chief through the belly, at the same time being wounded 
in the foot. Flame and smoke pour from the guns, and the wounds 
bleed freely. Sitting Bull's black war horse awaits its master in the 
background." — Vestal, 


No. 7 

"In an engagement with the Crows, Sitting Bull mortally wounds 
one of the enemy and dropping his lance rides up and strikes him 
with his whip. The lines and dashes in the picture represent the 
arrows and bullets that were flying in the air during the combat."- — 

"1861. Sitting Bull, amid a hail of enemy bullets, wounds a 
Crow warrior with his lance, then drops it and strikes him over the 
head with the heavy notched wooden handle of his quirt, which is 
decorated with a dangling kit-fox skin — the insignia of his Warrior 
Society. The Crow carries a quiver, and bleeds freely." — Vestal. 

No. 8 

"Counts 'coup' on a Gros Ventre de Prairie, by striking him with 
his lance. Gros Ventre distinguished from Crow by manner of 
wearing the hair." — Kimball. 

"1857. On the Missouri River. Winter. Sitting Bull, armed with 
a gun and wearing his Strong Heart and white blanket coat, strikes 
with his lance the Hohe lad whom he is to save and to adopt as his 
brother, named Jumping Bull, or Little Assiniboin." — Vestal. 


No. 9 
"Lances a Crow Indian." — Kimball. 

"1858. Near Rainy Butte. Sitting Bull lances and kills a Crow- 
warrior, the slayer of his father in that very fight." '" — Vestal. 

^^ For details see Vestal, p. 44. 

No. 10 

"A Crow Indian attempts to seize Sitting Bull's horse by the 
bridle. Sitting Bull knocks him down with a 'coup' stick, takes his 
scalp and hangs it to his bridle." " — Kimball. 

° Sitting Bull corrected this interpretation saying that his opponent is a Ree 
who drops his gun and bow from fear. The Ree was struck for coup but not 
killed (no blood is shown). See p. 7. 

"1859. Near Fort Berthold. A Ree enemy grabs the bridle of 
Sitting Bull's horse. Sitting Bull kills him, and takes his gun and 
bow." — Vestal. 


, ,^^^^ ^jg'jaTt'fyyy ■■._" ■ 


'>.".a»j^?H|!iWW^?- .• 



No. II 

"Sitting Bull with his brother mounted behind him kill a white 
man — a soldier." ' — Kimball. 

' When Williamson showed these pictures to Sitting Bull for verification, 
he found him unwilling to go into detail concerning his war exploits involving 
white opponents. 

"1868. In a skirmish with white men Sitting Bull rescues his 
unhorsed companion Jumping Bull, takes him up behind, and charges 
a white man armed with a rifle. Jumping Bull, being armed with 
a long lance, is able to strike the white man first. Sitting Bull has 
to be content with the second coup." — Vestal. 

No. 12 

"Counts 'coup' on a white man by striking him with a 'coup' stick." 
— Kimball. 

"1868. Sitting Bull strikes a white man. This happened on the 
same warpath as the deed recorded in Fig. ii. Circling Hawk, now 
living, was leader of this war party." — Vestal. 


No. 13 

"In a warm engagement with the whites, as shown by the bullets 
flying about, Sitting Bull shoots an arrow through the body of a 
soldier who turns and fires wounding Sitting Bull in the hip." — 

"1864. Near White Butte, on the Little Missouri River. Under 
heavy fire, Sitting Bull charges a white soldier. Though transfixed 
by an arrow from behind, and bleeding copiously from mouth and 
wounds, the brave soldier turns and shoots Sitting Bull through the 
buttocks, causing great loss of blood." " — Vestal. 

'This episode is described in detail by Vestal. See Vestal, 1932, p. 64. 

No. 14 

"Sitting Bull counts 'coup' on a white man by striking him with 
his bow. Sitting Bull wears a jacket and bandanna handkerchief 
taken from some of his victims." — Kimball. 

"1867-68 (winter). On the Montana Trail. Sitting Bull counts 
coup on a white man. In this affair Sitting Bull counted nine coups. 
This picture is followed by eight others showing the other coups 
struck. But as the drawings differ only in the details of the dress 
and persons of the white men, they have not been given here. Several 
of the white men were represented as having hair on their bodies — • 
a thing considered loathsome by the Sioux." — Vestal. 



No. 15 
'Sitting Bull counting 'coup' on a white man." — Kimball 

No. 16 
"Sitting Bull counting 'coup' on a white man." — Kimball. 


No. 17 
"Sitting Bull counting 'coup' on a white man." — Kimball. 

No. 18 
'Sitting Bull counting 'coup' on a white man." — Kimball. 






No. 19 
"Sitting Bull counting 'coup' on a white man," — Kimball. 

No. 20 
"Sitting Bull counting 'coup' on a white man." — Kimball. 


No. 21 

'Sitting Bull counting 'coup' on a white man." — Kimball. 

No. 22 

'Sitting Bull counting 'coup' on a white man."— Kimball. 




No. 23 

"Sitting Bull shoots a frontiersman wearing a buckskin shirt, 
takes his scalp which he hangs on his own bridle and captures his 
horse. Sitting Bull wears a blanket."— Kimball. 

"1863. Near Fort Totten, in the Devil's Lake country. Sitting 
Bull, wearing a red blanket, chases a mounted white man in a fringed 
buckskin coat, and shoots him between the shoulders. This was 
Sitting Bull's first white victim." — Vestal. 

No. 24 

"Sitting Bull strikes a white soldier with his 'coup' stick, takes 
his scalp and his mule. Wears a war shirt. "^ — Kimball. 

"1863, June. The skirmish with General H. H. Sibley's wagon- 
train on the Missouri River, near the mouth of Apple Creek. Sitting 
Bull, facing a heavy fire, as shown by flying bullets, charges a mule- 
skinner armed with a blacksnake whip, counts coup on him, and 
makes ofif with a saddled mule." — Vestal. 


No. 25 

"Counts 'coup' on a soldier, mounted with overcoat on, gun slung 
across his back, by riding up and striking with his riding whip." — 

"1867. On the Montana Trail. Sitting Bull overtakes a white man 
wearing an overcoat and armed with a rifle. Sitting Bull carries 
only a quirt, with which he strikes the fugitive. On his head Sitting 
Bull wears a bandanna taken from some enemy." — Vestal. 

No. 26 
"Kills a white man and takes his scalp." — Kimball. 

"1867. O^ the Niobrara River near the Missouri. Sitting Bull 
shoots a white man armed with a sawed-ofif shotgun. Sitting Bull 
carries a revolver, and is riding a rawhide saddle, made by his uncle." 





No. 27 
"Captures a mule and a scalp." — Kimball. 

"1865. North of the Black Hills. In a skirmish with the troops 
under Colonel N. Cole, of the Powder River Expedition, Sitting Bull 
runs off a slow pack-mule." — Vestal. 

No. 28 
"In a warm engagement, captures a horse and a scalp." — Kimball. 

"1865. On the Montana Trail. Under heavy fire from the soldiers, 
Sitting Bull captures a buckskin mare. He afterward gave her to 
his sister." — Vestal. 


No. 29 
"Steals a mule."- — Kimball. 

"November 6, 1867. Fort Buford. In an attack on the woodcutters 
from the post, one soldier was killed, one wounded. Sitting Bull 
captures a fine brown Army mule with a black spot on the withers, 
off side. He gave the mule to his sister."— Vestal, 

No. 30 
"Captures two horses in action." — Kimball. 

"1864. Under fire, Sitting Bull takes from the soldiers a chestnut 
and a buckskin horse. The buckskin he trained to run buffalo, and 
then gave it to his sister. These horses were captured in the Badlands 
from General Sully's troops." — Vestal. 

»*■■-;•;;.. . 

■-•:.. :■.>-. )^»:a5S'U"»?"'a:': V: 








No. 31 
"Steals a horse." — Kimrall. 

"1865. On the Montana Trail. Sitting Bull steals a fast buckskin 
war horse. He gave it to his adopted brother, Jumping Bull." — 

No. 32 

"Steals and runs off a drove of horses from the Crows." — 

"1863-64 (winter). Sitting Bull brings home nine Crow ponies: 
five bays, two blacks, one buckskin mule, and a little white mare. 
The mare he presented to his favorite sister, Pretty Plume." — 


No. S3 

"In an engagement captures a government horse, and mule, and a 
scalp." — Kimball. 

"i860. Amid a shower of bullets, which fill the air, Sitting Bull, 
riding his famous war horse Blackie, runs off two animals from a 
Crow camp. One of them is a branded Army mule, picked up or 
stolen by the Crows. These animals Sitting Bull gave for Brown 
Eyes, the girl who became his fourth wife."^ — Vestal. 

No. 34. 
"Steals a horse." — Kimball. 

"1866, On the Montana Trail. Sitting Bull takes a horse with 
a split ear from white men." — Vestal. 



No. 35 
"Captures three horses and a scalp." — Kimball. 

"1866. Wearing his Strong Heart bonnet, and riding Blackie, 
Sitting Bull captures three Crow ponies ; one bay, one black, one 
mouse-colored." — Vestal. 

No. 36 
"Steals a drove of horses from the Crows." — Kimball. 

"1862. Sitting Bull runs off a bunch of Crow ponies. Sitting Bull 
was such a noted horse-stealer that the old men say nobody can 
remember all his raids. Chief Charging Thunder stated that to his 
own knowledge Sitting Bull took horses from the Crows twenty 
times, sometimes as many as thirty head at a time." — Vestal. 


No. z-] 
"Steals a government horse." — Kimball. 

"1865. Wearing beaded leggins and a fur cap with earflaps, 
Sitting Bull runs off a horse belonging to the Powder River Expedi- 
tion." — Vestal. 

No. 38 
"Steals a drove of horses from the Crows." — 'Kimball. 

"1859-60. Wearing his Strong Heart bonnet, Sitting Bull runs off 
seven Crow ponies : two white, two black, one bay, one buckskin, 
and one mouse-colored." — Vestal. 







No. 39 

"In an engagement captures a mule. Sitting Bull first appears here 
as Chief of the Band of Strong Hearts, to which dignity his prowess 
has raised him. The insignia of his rank — a bow having on one end 
a lance head — he carries in his hand." ' — Kimball. 

" This drawing actuallj' depicts a feat of Jumping Bull tlie adopted son of 
Sitting Bull. See Williamson letter, p. 7. 

No. 40 

"Sitting Bull, Chief of the Band of Strong Hearts, captures two 
rses in an engageme 
shoulder." "* — Kimball. 

horses in an engagement in which his horse is wounded in the 

"■ This drawing actually represents an exploit of Jumping Bull. It is errone- 
ousl}^ attributed to Sitting Bull by Kimball. (See p. 7.) 


No. 41 
"Captures a horse in a fight." " — Kimball. 

"This drawing actually represents an exploit of Jumping Bull. It is errone- 
ously attributed to Sitting Bull by Kimball. (See p. 7.) 

No. 42 

'Steals a mule." '" — Kimball. 

"This drawing actually represents an exploit of Jumping Bull. It is errone- 
ously attributed to Sitting Bull by Kimball. (See p. 7.) 

AA \ 

V ./'//. 

. A. 


WX^\ l^ 


^- /o^ 

\--; c 

^r\ \ \ 

'^'nW — 


I \ 

_y^ -Yp 



\ — >^ 

=^' V 

5i4 J-^ ^- 



" 1 


•^ tT 

Vrr>^ \ 

L^ ■---__ \ 1^ 


/ ~-—' 


\ir:' ^-\ - 




^^^^'^^ )c 




-A \ / 1 







No. 43 

"Captures two horses in a fight in which his horse is wounded in 
the leg." " — Kimball. 

" This drawing actually represents an exploit of Jumping Bull. It is errone- 
ously attributed to Sitting Bull by Kimball. (See p. 7.) 

No. 44 

"Mounted on a government horse, captures a white man." " — 

" This drawing actually represents an exploit of Jumping Bull. It is errone- 
ously attributed to Sitting Bull by Kimball. (See p. 7.) 


No. 45 
"Steals two horses." '"—Kimball. 

^This drawing actually represents an exploit of Jumping Bull. It is errone- 
ously attributed to Sitting Bull by Kimball. (See p. 7.) 

No. 46 

"Captures four mules in a fight in which his horse is wounded in 
the hip." '" — Kimball. 

" This drawing actually represents an exploit of Jumping Bull. It is errone- 
ously attributed to Sitting Bull by Kimball. (See p. 7.) 

> ^ (> 





No. 47 
"Counts 'coup' on white man." '' — Kimball. 

^^ This drawing actually represents an exploit of Jumping Bull. It is errone- 
ously attributed to Sitting Bull by Kimball. (See p. 7.) 

No. 48 • 
"Counts 'coup' on white man." " — Kimball. 

^* This drawing actually represents an exploit of Jumping Bull. It is errone- 
ously attributed to Sitting Bull by Kimball. (See p. 7.) 


No. 49 
"Steals a government horse." '" — Kimball. 

" This drawing actually represents an exploit of Jumping Bull. It is errone- 
ously attributed to Sitting Bull by Kimball. (See p. 7.) 

No. 50 

"Fastens his horse to his lance driven into the earth and in a hand 
to hand fight kills a white man with his own gun. The black marks 
show the ground fought and trampled over." ''" — Kimball. 

^^ This drawing actually represents an exploit of Jumping Bull. It is errone- 
ously attributed to Sitting Bull by Kimball. (See p. 7.) 





No. 51 

"A fort into which his enemies the Crows, have retreated and from 
which they maintain a hot fire through which Sitting Bull charges 
the fort." '' — Kimball. 

'^ This drawing actually represents an exploit of Jumping Bull. It is errone- 
ously attributed to Sitting Bull by Kimball. (See p. 7.) 

This apparently represents the feat of Jumping Bull performed at Spoon 
Horn Butte, where he drew the fire of the Crows in order to empty their guns 
before the Sioux charged. This exploit is described in Vestal, 1932, p. 116. 

No. 52 

"In a fight with the Crows, Sitting Bull kills and scalps one Indian, 
and counts 'coup' on another who fired at him barely missing him." " 
— Kimball. 

■" For details concerning this exploit see Vestal, 1932, chap. 16. 

"1869. Near the Big Dry. An incident of the battle in which the 
thirty Crows were killed. Sitting Bull, wearing a horned bonnet and 
beaded leggins, charges the rocky barrier (indicated by the circle), 
and counts coup upon a Crow, who fires in his face, but misses. The 
air is full of flying lead." — Vestal. 


No. 53 
"Steals a drove of mules." ''' — Kimball. 

'^ This drawing actually represents an exploit of Jumping Bull. It is errone- 
ously attributed to Sitting Bull by Kimball. (See p. 7.) 

No. 55 

"Kills one Crow and counts 'coup' on two others, who run from 
him disgracefully." ^ — Kimball. 

^* Sitting Bull stated to Williamson that this picture is incomplete. It should 
carry his "name" glyph. (See p. 7.) 

"This, the last of the series, is incomplete, and lacks the picture 
of the seated buffalo, which should identify Sitting Bull. However, 
the shield is enough to serve that purpose. Sitting Bull himself 
explained that this unfinished sketch represented a fight with the 
Crows in which he killed one and counted coup on two others, who 
ran from him disgracefully. The date and place of this fight are 
unknown." — Vestal. 




In June 1923, through the generosity of Mr. Rohert A. Smith, 
the Bureau of American Ethnology archives were enriched by an- 
other Sitting Bull document of even greater interest. This consists 
of a later Sitting Bull pictographic autobiography drawn by the 
great Sioux warrior himself. Although it contains drawings of only 
22 exploits, it is well documented, and tlie explanations of the draw- 
ings were given by Sitting Bull at the time the pictures were made. 

The drawings were made with a pencil on the pages of an army 
ledger book, and colored by means of water-color paint. The human 
figures are rather crudely drawn in the usual Plains Indian style, 
but Sitting Bull shows his individuality even in the field of art, by 
the manner in which the horses are depicted. Departing from the 
general Plains Indian style of representing horses in a slender and 
much conventionalized fashion, he draws his horses- realistically and 
in a well rounded manner. The various horses shown are so consci- 
entiously delineated that some of them can be recognized from 
descriptions of Sitting Bull's favorite mounts given by Vestal and 

In the Four Plorns copy of Sitting Bull's autobiography of 1870, 
the warrior is always identified by his name glyph in the form of a 
seated buffalo. At the time the present picture record was made. 
Sitting Bull had learned to write his name, and his signature accom- 
panies each drawing in the place of the buffalo. 

The following letter from Mr. Smith accompanied the book of 
pictures and the documents concerning them. 

^ Bob Davis, the well known newspaper writer, informed the author that 
in an interview in 1931, he learned that Rudolph Cronau was sent to America 
by the Gartcnlanbe, a weekly periodical published in Leipzig, to cover the Indian 
wars as illustrator. In 1881 the artist made the acquaintance of Sitting Bull 
shortly after his surrender at Fort Buford. Sitting Bull was much interested 
in watching Cronau make his sketches and wishing to gain the friendship 
of the Sioux leader, Cronau spent some time in teaching him to draw. This 
very probably accounts for the sophisticated and un-Indian appearance of the 
horses drawn by Sitting Bull. 



Robert A. Smith 
430 S. Garden Street 
South BeHingham, Wash. 

June 20, 1923. 
Smithsonian Institution 
Washington, D. C. 

Gentlemen : 

I am sending with this a boolv of paintings by Sitting Bull, with interpre- 
tation of same, letters from Wallace Tear, Lieut. 25th Infantry U.S.A. to my 
father, General John C. Smith, which explain themselves. Lieut. Tear was a 
soldier in the 96th Regt. Infantry^-Illinois U.S.V. 1860-65. 

My father was Captain, Major, Lieutenant Colonel, Brevet-Colonel 96th 
Regt., and Brevet Brig. General. At the close of the war he was able to get 
a commission in the Regular Army for Tear and did him some favors after- 
wards, hence this History of Bull. 

All the people mentioned are gone and when I pass on there will be none 
that will be interested,' so I would like to have this book where perhaps it might 
interest someone — sometime. 
I am. Gentlemen, 


Robert A. Smith 

430 South Gardner Street 

South BeHingham 


Two letters of explanation from Lieutenant Tear addressed to 
Gen. John C. Smith were with the pictographic record. 

Fort Randall, D. T. 
August 10, 1882 
Dear General: 

Yours of 31th ult., with photographs of yourself and wife reed. Many thanks. 
Mrs. Smith looks younger than she did 20 years ago — fact — you don't look 
very old yourself. I may be a little prejudiced in this matter as I am "passing 
off" as a young man yet. 

I have Sitting Bull's description of the paintings, taken down when the 
pictures were made. Am copying them for you, and will send them next mail. 
Intended to send them with the book, but I had to send the book to keep it 
from being stolen. I came near losing it. Some tourists wished to look at it 
and then borrowed it for a while to show to some friends. I only got it back 
"by a scratch." They had hidden it with the intention of carrying it off. As 
soon as I got my hands on it, I put it into the mail. 

Bull is very diffident about giving any incidents of his fights with the whites. 
I have tried to have him give me a detailed description of the Custer fight 
but he seems rather timid. Once in speaking of the affair he said: "I did not 
hunt Custer. I thought I had a right to protect my own women and children. 
If he (Custer) had taken our village he would have killed our women and 
children. It was a fair fight." 

I will try and get him to make a picture of some portion of the fight. 


I will try and think of some trinket that would please him as a gift from 
you — something that will cost but little — -I can't think of anything at present. 
Am quite well. Have been out in the field a good deal this summer and am 
quite busy. I have the luck to be alone with the Co. most of the time. 

Love to all. Of course you will be elected. If you need my vote I will come 
home on election day. 

Yours truly, 
W. Tear (signed) 

Fort Randall, D. T. 

August 1 6, 1882 
Dear General: 

I send you inclosed Sitting Bull's interpretation of his paintings recently sent 

I furnished the book which contains the paintings and from time to time 
saw him at work on them. These notes were taken down by me, after the 
paintings were completed, in Sitting Bull's tipi in the same routine as given 
by himself (thro an interpreter of course) Bull having the picture before him 
while giving a description of the fight. It was impossible to locate the scenes 
with any definiteness ; "The Land of the Sioux" ; "The Land of the Crows" 
and "a long way from the Missouri" being the usual location given. In talking 
of his life Bull uses his name instead of the pronoun "T" ; that is he speaks in the 
third person. In these notes you must understand that it is Sitting Bull speaking. 
Bull made these pictures for me to show his gratitude for blankets and clothing 
furnished his children last winter before the Government supply of clothing for 
his band arrived. 

I am endeavoring to get him to complete his history up to the present time, 
and if successful you shall have it. 

Bull says he is 43 years old. I think he is nearer 50. These scenes of his 
life of course comprise his life from the time that he was able to ride a horse 
and handle a bow. The scars of the wounds he speaks of are visible now. 
Regards to your family. 

Yours truly, 

W. Tear (Signed) 
Lieut. 25th Infantry 
Gen. J. C. Smith 
Chicago, 111. 

250 W. Van Buren Street 

The list of explanations with the picture record are written in 
the hand of Lieutenant Tear on nine pages of foolscap paper. These 
explanations are here placed opposite the pictures described, exactly 
as written. 


No. O 

Assiiiniboine Chief taken prisoner by Sitting Bull in a fight l)et\veen 
Sioux and Assinniboines. lOo Sioux — whole tribe of Assinniboines 
engaged — about 27 years ago, when Sitting Bull was about 16 years 
old — Land of the Sioux. Kept the chief prisoner for while and then 
gave him the horse he ("Bull") rode and the bonnet he ("Bull") 
wore in the fight and then sent him to his people with a good heart. 

No. I 

Fight with Assinniboines — 140 Sioux — 43 Assinniboines — 23 As- 
sinniboines killed — 8 Sioux killed — 20 Sioux wounded — "Bull" 18 
years old— In land of Sioux. Bull took several prisoners. Didn't 
kill prisoners. Kept them many days. Gave them ponies and sent 
them home. Assinniboines were hunting in Land of the Sioux. 


"■,^- iT'T^P ■'i^'^^.''/ ■ ■ ''^■'VW-.l^ir' , ,J 


No. 2 

Scene in same fight. 

No. 3 

Fight with Assinniboines. "Bull" 24 years old. 50 Sioux — 200 
Assinniboines. No Sioux killed. 5 Assinniboines killed. "Bull" kills 
warrior. "Bull" 24 years old. 


No. 4 

Assinniboine woman taken prisoner by "Bull" in a fight. "Bull" 
16 years old. 

No. 5 
Same fight. 5 women captured by "Bull." 


No. 6 

Same fight. Woman captured by "Bull." The warrior toutches 
[sic] woman with his lance and she becomes a prisoner. Warrior 
never strikes a woman in a fight except to save his own life. These 
women were kept with the Sioux a short time and then sent back to 
their own people except 3 who married Sioux warriors and remained ; 
one of them here now. No one killed in this fight. Assinniboines 
passing thro Land of the Sioux. 

No. 7 

Fight with the Crows. Crow Indian killed by "Bull." 30 Sioux 
warriors — 200 Crows, men, women and children. 14 Crows killed. 
No Sioux killed. Crows were travelling with their camps. "Bull" 
20 years old. On the Little Missouri river, "Crows always fighting 
the Sioux — tried to make friends with them but they were always 
doing something bad." — Bull. 


No. 8 

Fight with Assinniboines. Warrior killed by "Bull." 350 Sioux 
run upon 10 Assinniboines and killed 2. i Sioux killed and 2 
wounded. Land of the Sioux, a little above the forks of the Missouri 
near mouth of Yellow Stone. "Bull" 25 years old. 

No. 9 
Scene in same fight described in No. 7. "Bull" kills Crow Indian. 


No. lO 

Fight with Crows. "Bull" killed 2 men and captured 2 women. 
"Bull" 25 years old. Crows were stealing ponies. Let women go 
home with presents for Crow chiefs to try and make friends. 

No. II 

Fight with Crows. 100 Sioux — whole tribe of Crows. Bull killed 
Crow Chief. 3 Crows killed, i Sioux warrior and i woman killed. 
Land of the Sioux — a little above Tongue River. Bull 22 years. 


No. 12 

Same fight. "Took long time to kill these people. Here is where 
I got wounded in leg and got off of horse and killed this man. No 
prisoners in that fight. This is 'Stand and Kill' Crow Chief. Had 
guns in this fight. The Sioux used to take the Crows prisoners 
and give them good clothes and feed them up and give them good 
ponies and then send them back so they could tell a good story of 
the Sioux to their people." ("Bull's" description of fight.) 

No. 13 

Fight with Assinniboines. "Bull" takes 2 prisoners. "Bull" 30 
years old. 300 Sioux — -20 Assinniboines — 2 Assinniboines killed— 
no Sioux killed. On the big fork of the Missouri. "Bull" took one 
prisoner, "Jumping Bull," home to his (Bull's) tipi (wigwam) gave 
him his ("Bull's") horse and war bonnet. Jumping Bulls' father 
was a Chief. Jumping Bull is now at Standing Rock (Fort Yates, 
D. T.) with my people. They call him my son. 





No. 14 

Fight with the Rees. i6 Sioux. lOO Rees. Sioux were fighting 
and retreating. Sioux turned and chased Rees. This Ree Indian, 
Chief "Bull Head," fell down dead. "Bull" took him prisoner and 
he came to life again. This is the only Ree caught in the fight. No 
Sioux killed. In the land of the Rees. "Bull" sent this prisoner home 
with presents. Made peace wuth Rees, and peace with Assinniboines. 
"Bull" 33 years old. 

No. IS 

Fight with Gen. Miles' Scouts and Crow Indians. "Bull" kills 
"Brave Indian," one of Gen. Miles Scouts. About three years ago — 
the time Gen. Miles was out after the Sioux near the Queens' land 
( Canada) "Brave Indian" was away ahead of the soldiers and was 
following up the Sioux too close. "Bull" turned and killed "Brave 
Indian." One Cheyenne Indian (scout) also killed. Sioux did not 
fight soldiers — wanted to get away from soldiers. The scouts and 
Crows killed 5 Sioux before they got to Canada. Gen. Miles' Scouts 
seemed to be from every Indian nation. The Sioux run awav. 



No. 16 

Fight with Flat Heads. "Bull" wounded in left arm and side by 
arrow. "Bull" killed Flat Head. 15 Sioux, young men, went on 
war path. Flat Heads killed them all. Sioux then went out with 
300 warriors. 40 Sioux attacked the camp of the Flat Heads ; the 
main body of Sioux being hid back from the camp ; the Flat Heads 
chased the 40 Sioux back through the main force of Sioux. The 
Sioux charged and killed 33 Flat Heads. 4 Sioux killed — good many 
wounded. 7 years ago. Bull 36 years old. Near Muscle Shell river. 

No. 17 

Crow Indian killed by "Bull." 200 Sioux run upon 7 Crows hunt- 
ing in Land of the Sioux and killed them all. Crows had guns. Sioux 
had nothing but bows and lances. Crows were crossing Missouri — 
(river). A few years ago. ("Bull" gave his age at the time of this 
fight but my notes are defaced at this point so that I am uncertain 
as to how old he said he was. — Tear.) 




No. i8 

Fight with Crows. 200 Sioux — whole tribe of Crows. "Bull" 
kills Crow warrior. 5 Crows killed in the fight. No Sioux killed. 
"Bull" ;^2 years old. (Chicken Hawk, skinned and skin stuffed, worn 
as ornament where picture of bird is seen near Bull's head.) 

No. 19 

Fight with Crows. 200 Sioux — whole tribe of Crows. "Bull" 
kills warrior. 7 Crows killed. No Sioux killed. "Bull" dressed in 
war bonnet trimmed with eagle feather. "Bull" 24 years old. Near 
mouth of Tongue River. 


No. 20 

Fight with Assinniboines. lOO Sioux— 60 Assinniboines. "Bull" 
kills warrior. 3 Assinniboines killed, i Sioux wounded. Near big 
fork of the Missouri. "Bull" 29 years old. 

No. 21 

Fight with Assinniboines. 320 Sioux attacked big winter camp 
of Assinniboines. Bull kills warrior. 5 Assinniboines killed. No 
Sioux killed. Did not get in to their (Assinniboines') camp — there 
were too many and fought too well. Sioux run off after killing these 



During the month of February, 1938, a news release appeared 
concerning the Sitting Bull autobiographies in the Bureau of Ameri- 
can Ethnology. The following letter came as a result : 

Oswego, Oregon 
March 7, 1938 

Smithsonian Institution 
Washington, D. C. 

Gentlemen : 

My uncle the late Dan'l L. Pratt of Seattle was Post Trader at Fort Randall, 
Dak. Ter. in 1882, 56 years ago and knew Sitting Bull and his band very well. 
Sitting Bull sketched for him 13 pictures of himself on horseback showing 
him in action against the Crows, Gros Ventres etc. — each one is drawn on 
paper loi by 8^ inches and marked in print— D. L. Pratt, Post Trader, Fort 
Randall, 188-. The horses are very well drawn — in Indian style — some in colors. 
These pictures came to me in book form bound in oil cloth. I now have them 
with the affidavit of Mr. Pratt in a large walnut frame — they are quite im- 

Hoping this information will be interesting, 

I am 

Yours truly, 

/s/ Mrs. G. H. Pettinger 
Oswego, Ore. 

The writer communicated at once with Mrs. Pettinger, who very 
kindly forwarded the pictures to the Bureau of American Ethnology 
so that they could be included with this publication. They belong to 
the George Howard Pettinger Collection, which contains several 
other very interesting Sitting Bull items, including the tomahawk 
surrendered by Sitting Bull to Lieutenant Ogle on the occasion of 
Sitting Bull's surrender to the Commanding Officer at Fort Buford, 
Dakota Territory. The writer wishes here to express his deep ap- 
preciation to Mr. Pettinger for making available this interesting 
addition to the Sitting Bull record. 

These pictures, like those of the Smith autobiography, were made 
at Fort Randall in 1882 and were probably painted from the same 
paint box, as the shades of the colors used are identical in the two 
sets. The Pettinger drawings are exactly the same in style but lack 
the signature. In the Pettinger drawings, Nos. 2 and 5 are unique 
in that full-face figures are shown. It seems probable that some 
pictures are missing from the series as originally drawn, for in two 
instances the descriptive sequence appears to refer to a preceding 
episode which is not shown. 



No. I 

"Winter" scene. Killing a Gros Ventres Indian — loo Sioux on 
war path killed 3 Gros Ventres whom they found hunting.'" 

^^ This is the same episode as that shown in No. 21 of the Smith autobiography. 
Tlie horse is blue-black in color. Sitting Bull is wearing black army trousers 
with a red stripe. His victim wears a green blanket coat. 

No. 2 

i860. Sitting Bull killing a Crow Indian. One hundred Sioux 
chased thirty Crows all night. Caught them in the morning and 
killed them all." 

^'This is probably the fight that took place in the winter of 1869. See Vestal, 
p. 115. 

■ U 



,<^^^^ ^^..^ y^^^£^^ ^ ^^^^ 0^^. 

^•^"^ ■ "'- .^^ ^/LSl, ^o^z^ ^^- ^uy-^'^ ^a^.^c^ i^L.^ 



No. 3 

Sitting Bull killing a Crow Indian. 40 Sioux against 7 Crows. 
Sitting Bull killed 3 Crows in this fight. 

No. 4 
Same fight. 


No. 5 
Sitting' Bull killingf a Crow Indian.'^ 

^ Sitting Bull is here shown wearing the same costume depicted in No. 2. 
These two pictures are unique in that in three instances the men pictured are 
shown in full face instead of profile. 

No. 6 
Killing a Flathead Indian in a battle. 

' Sitting Bull is wearing a shirt covered with green spots. 

/^^^^ /&i/^ ;c^«4^ ^ Siie^ 5?*'^aii^*-- 

/^i;:^ dJ £?^£r- .-^4;fc«^ ^.^^ ^^^ rz/ dc^^^ i^t!^^^.. 

% ^nv 


No. 7 
Killing a Flathead and receiving an arrow wound in left side. 

^"This is the same episode as depicted as No. i6 of the Smith series. This 
battle with the Flatheads and the particular incident here shown is described 
by Vestal, p. 125. 

No. 8 
200 Sioux in a fight with 50 Chippeways. Sitting Bull killed one. 

*^ Sitting Bull is mounted on a yellow horse, black nose and ear tips. His 
shield is painted blue. In all of the other pictures in which the shield is shown, 
the color is green. 


No. 9 

Jumping Bull, a Gros Ventres, captured in a fight by Bull — Sitting 
Bull took a fancy to him and presented him with his war bonnet and 
horse and permitted him to be free — made chief after this occur- 



^^ This represents the same episode as pictured in No. 13 of the Smith record 
and No. 5 of the Kimball record. 

No. 10 

1880. Killing a Crow Indian Scout who belonged to General 
Miles' command. Sitting Bull wears a war bonnet which once be- 
longed to Crazy Horse.^^ 

^^ This represents the same episode shown in No. 15 of the Smith record. 
The horse in each instance is colored a light purple, probably a roan. 

' /^ ^-1 






■\ \ 


/ rr~fi 



i«i<i«!»y^«.^S^> OeeM^ /^^'"'^^^ ,ii:%:^J«.i<:i<^ O^tM^ '!^^<r^ct.^.^/,^^^ /S^^ 


mt\ ^ 




No. II 

Same fight. Killing a Crow who dismounted and fought desperately 
and wounded Bull in two places.^ 

^*This does not refer to the same fight pictured in No. lo. It is the same 
episode, evidently, as shown in No. 12 of the Smith record and No. 4 of the 
Kimball record. It was in this encounter that Sitting Bull received the wound 
in his foot which crippled him for the rest of his life. 

No. 12 

^^This, evidently, does not refer to the same fight as depicted in No. 11. 
Apparently, it represents the capture of an Assiniboin woman, probably on the 
occasion shown in Nos. 5 and 6 of the Smith record. 


No. 13 
Same fieht. Killed two."' 

^' Sitting Bull is shown wearing a shirt spotted with red. His opponent is 
wearing a shirt spotted with green. This may represent the fight between the 
Sioux and the Assiniboines shown in Nos. i and 2 of the Smith record. 




Johnston, W. Fletcher 

1891. The Red Record of the Sioux Life of Sitting Bull and History of 
the War of 1890-91. Edgewood Publishing Company, Philadelphia. 
De Barthe, Joe 

1894. The Life and Adventures of Frank Grouard. Combe Printing Co., 
St. Joseph, Mo. 
MooNEY, James . 

1897. The Ghost-Dance Religion and the Sioux Outbreak of 1890. 14th 
Ann. Rep., Bur. Ethnol., pt. 2. 
McLaughlin, James 

1910. My Friend, the Indian. Houghton MifHin Co., Boston. 
Vestal, Stanley 

1932. Sitting Bull, Champion of the Sioux. Houghton Mifflin Co., Boston 
& New York. 






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

(Publication 3483) 



AUGUST 23. 193 8 





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

(Publication 3483) 



AUGUST 23, 193 8 

Z-^t Bovh QSafttmorc (^ttee 



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



I. The hypothetical annelid ancestors i 

II. The mesoderm and the beginning of metamerism 9 

III. Development of the annelid nervous system 21 

IV. The adult annelid 26 

The teloblastic, or postlarval, somites 26 

The prostomium and its appendages 32 

The body and its appendages 34 

The nervous system 39 

The eyes 45 

The nephridia and the genital ducts 45 

V. The Onychophora 50 

Early stages of development 52 

The nervous system 55 

The eyes 62 

Later history of the mesoderm and the coelomic sacs 62 

The somatic musculature 64 

The segmental appendages 67 

The respiratory organs 70 

The circulatory system 70 

The nephridia 72 

The organs of reproduction 74 

VI. The Arthropoda 76 

Early embryonic development 80 

Primary and secondary somites 82 

The cephalic segmentation and the development of the brain 89 

Evolution of the head 107 

Coelomic organs of adult arthropods 126 

The genital ducts 131 

VII. Phylogenetic conclusions 132 

References 149 


Among the simplest of the metazoic animals that lead an active, 
free existence is the planula larva of the Coelenterata. The planula 
develops into a polyp or a medusa because it is a young coelenterate, 
but, so far as its structure goes, it contains the fundamental building 

Smithsonian Miscellaneous Collections, Vol. 97. No. 6 


VOL. 97 

elements that, with the appropriate hereditary influences, might be 
fashioned into a flatworm, an annehd, an arthropod, a moUusk, or 
a vertebrate. 

The typical planula is a minute oval or elongate creature (fig. 
I A, C) consisting of an outer layer of ectoderm cells, and an in^er 
mass of endoderm cells. The planula, therefore, represents the 
gastrula stage of embryonic development, though it may have no 
enteric cavity and no blastopore. Its motor mechanism is a covering 

Fig. I. — The coelenterate planula, and two methods of endoderm formation. 

A, planula of Sympodium corraloidcs (from Kowalevsky and Marion, 1883). 
B, blastula of Carmarina fungiformis, showing dififerentiation of endoderm from 
ectoderm by delamination of blastoderm cells (from Metschnikoff, 1882). C-F, 
formation of endoderm by internal proliferation from posterior pole of planula 
(from Hatschek, 1888, after Claus). 

Blc, blastocoele ; Bid, blastoderm ; Ecd, ectoderm ; End, endoderm. 

of vibratile cilia. The embryology of the planula is very simple. The 
cleavage of the coelenterate egg produces a morula, and the morula 
becomes a blastula. In the succeeding planula stage the inner endo- 
dermal cell mass is formed, but it is not certain that gastrulation 
takes place by simple invagination in any of the coelenterates. With 
some forms the endoderm arises as an inward migration of scattered 
cells from the blastoderm ; in others the blastomeres divide regularly 
each into an outer ectoderm cell and an inner endoderm cell (fig. i B) ; 
but the most common method of endoderm formation is the internal 


proliferation of cells from the posterior pole of the blastula (D, E, F), 
and this last process suggests that it is an embryonic modification 
of gastrulation by invagination. When, shortly, the planula settles 
to the bottom of the water and becomes attached preparatory to its 
metamorphosis into a polyp, a stomach cavity appears in the endo- 
derm, and a mouth cavity breaks through at the free pole. 

The development of the planula shows clearly that there is in 
ontogeny no fixed method for the formation even of so important 
an organ as the stomach. The effective thing in embryonic develop- 
ment is the inherited organizing property resident in the egg that 
converts a mass of cells, however formed, into a definite functional 
structure. The same principle, as we shall see later, applies also to 
the development of the annelids and the arthropods, for in these 
animals there is so much apparent irregularity in the formation of 
the germ layers that attempts to interpret all observed facts in terms 
of cell genealogy lead only to confusion. Ontogeny and phylogeny, 
therefore, while they produce the same end results, may follow quite 
different methods of procedure. In phylogeny we must visualize the 
successive stages in the evolution of an animal as free-living adult 
forms, each structurally adapted for performing the functions of an 
independent animal. 

If the coelenterate planula were an adult animal instead of a 
temporary larval form, or if it had to maintain itself for any con- 
siderable length of time, it almost certainly would have a stomach 
cavity and a mouth. Thus modified, as it is later in its own develop- 
ment, the planula would be an independent, motile gastrula, having 
a stomach in the form of an open pocket of the blastoderm for the 
retention of food particles (fig. 2 A). An animal of this simple type 
of structure, we must suppose, was the actual ancestor of the polyp 
and medusa forms of the Coelenterata ; but equally well it might have 
been the progenitor of the annelids, and through the latter of the 
arthropods. Various writers on phylogeny have proposed an origin of 
the segmented worms direct from a coelenterate polyp, but it should 
be recognized as a fundamental principle in evolution that a special- 
ized type of animal does not give rise to another specialized type — if 
two forms are related, they are related through some simple common 
ancestor. This principle as applied to the coelenterate derivatives is 
expressed by Ziegler (1898), who says: 

It is to be supposed that the higher animals derived from coelenterate stock took 
their origin not from the highly specialized forms of the Coelenterata, such as 
the anthozoans and ctenophores, but from a planula-like or gastrula-like ancestral 
form of the coelenterates. 


The theoretical planulalike gastrula postulated above as the com- 
mon ancestor of the Coelenterata and the Annelida (fig. 2 A) pre- 
sumably swam habitually in one direction by means of a covering of 
cilia, and the mouth, or blastopore, was at the posterior pole where 
food particles might be swept into the stomach with the eddy of 
currents converging to the rear. 

In the ontogenetic development of the annelids, gastrulation 
generally takes place by epiboly, which is the overgrowth of the endo- 
derm by the ectoderm, and the primary open blastopore is at the 
posterior pole of the embryo. There is no reason why this ontogenetic 
stage should not represent an early phylogenetic stage, and one iden- 
tical with the gastrula ancestor of the Coelenterata (fig. 2 A). With 
the further development of the annelid embryo, however, the blasto- 
pore elongates forward on the ventral surface of the gastrula (fig. 
2 F) until its anterior end comes to be near the anterior pole (G) ; 
but, at the same time, the lips of the blastopore grow together from 
behind forward, leaving finally only the anterior end open into the 
archenteron, and this opening is the primitive mouth (H, Mth). 
Secondarily, an anal aperture {An) is formed later at the original 
posterior end of the blastopore on the caudal extremity of the embryo. 
The endodermal archenteron of the annelid thus becomes a simple 
alimentary canal having the oral aperture located ventrally near the 
anterior end of the body, and the anal aperture situated terminally 
at the posterior end. 

If we visualize the change in the position of the blastopore as an 
event in the phylogenetic history of the annelids, we must see a corre- 
lated change in the habits of the animal. The actively swimming 
gastrula (fig. 2 A) in its search for food, we may suppose, took to 
brushing over the surfaces of stones or aquatic plants (B), where 
food particles were more numerous and more easily obtained. For 
this manner of feeding, a ventrally placed blastopore would be a 
distinct advantage, or, even more efficient, a blastopore drawn out 
lengthwise on the under surface (C). With a form thus modified in 
habits and structure, there may easily have developed a creeping 
habit, and an adaptation of the ventral cilia for progression on solid 
surfaces (D). Finally, then, came a more complete adaptation to 
feeding on a subsurface, resulting in an elongate flattened body, and 
the establishment of an alimentary canal with a ventral mouth and 
a terminal anus (E) produced by the closure of the intermediate 
part of the blastoporic slit. 

A creeping mode of locomotion may be subserved entirely by a 
ciliary coating of the body wall, as is shown in the Platyhelminthes, 


but a creeping animal encounters irregularities and obstructions. A 
provision for body movements, therefore, becomes an advantageous 
adjunct to the motor mechanism, and such movements can be pro- 
duced only by an internal muscular system. Hence, the next stage 
in evolution, recorded in both the flat worms and the annelids, was 

Fig. 2. — Hypothetical evolution of a swimming planulalike creature with an 
open gastrocoele into a creeping wormlike animal with a simple alimentary 
canal, a subapical ventral mouth, and a terminal anus. 

A, primitive swimming form with posterior blastopore. B, the same having 
acquired the habit of sweeping up food particles from a solid surface. C, blas- 
topore elongated forward on surface of contact to accommodate the feeding 
habit. D, the same more fully adapted to subsurface feeding. E, final develop- 
ment of alimentary canal, with ventral mouth and terminal anus, formed by 
closure of intermediate part of blastopore, creeping habit fully established. 
F-H, three stages of elongation and closure of the blastopore, ventral view. 

AlCnl, alimentary canal; An, anus; Bpr, blastopore; Gc, gastrocoele, or 
archenteron; Mth, mouth. 

the development of contractile tissue that conferred the power of 
diversified adjustive movements on the body itself. Muscles, how- 
ever, -are not generally automatically active, as are cilia, and hence 
the development of muscle tissue is usually accompanied by the 
development of a mechanism for its activation. Furthermore, since 
a muscular system is a provision for adjustment to external con- 
ditions, the source of its stimulus must come from the environment. 
The sponges are said to have a primitive contractile tissue that is 


VOL. 97 

stimulated directly by environmental changes ; in all other animals 
there is intimately associated with the contractile muscle tissue a 
specifically receptive and conductive nerve tissue, through which 
environmental stimuli become effective on the muscles. Finally, the 
high metabolic rate of muscular activity creates the need of special 
excretory organs for the removal of waste products from the lx)dy. 
The genesis of contractile and conductive tissues, and their inte- 
gration into a neuromuscular system are best seen in the Coelenterata. 
Contractility, being a common property of protoplasm, may become 
localized and specially developed in a particular part of any cell of 
the body in a primitive animal. In the coelenterates fingerlike muscle 
processes are produced from the inner ends of cells in both the 
ectodermal and the endodermal epithelium, those of the ectoderm 
(fig. 3, iiip) taking a longitudinal course, those of the endoderm a 




SL— ' 

Fig. 3. — Diagram of the ectodermal neural and muscular elements of Hydra. 
(From Curtis and Guthrie, 1927.) 

Cnb, cnidoblast ; Ecd, ectoderm ; ntf, muscle fiber ; mp, muscle process of 
epithelial cell ; NCI, neural cell ; NSCl, neurosensory cell ; SCI, sensory cell ; 
SL, supporting lamella. 

transversely circular course. Fibrils of contractile tissue (nif) be- 
come differentiated in these processes. In the hydra, the body of the 
muscle cell remains as a part of the epithelial layer, but in some of 
the other coelenterates the entire cell may be withdrawn beneath the 
surface and converted into a muscle fiber. A primitive nerve cell is 
an epithelial cell in which the common protoplasmic properties of 
irritability and conductivity are specially developed both in the cell 
body and in branching processes given off from the latter, but the 
nerve cells become differentiated into superficial receptive cells and 
deeper-lying conductive cells. In the hydra the receptive cells (fig. 3, 
SCI) and the sensory cells (SNCl) contained in the ectoderm have 
connections, on the one hand, with the surface of the body, and, on 
the other, send branches to the strictly neural cells (NCI), which are 
distributed through the inner parts of the ectoderm, and in turn 
send branches to the muscle processes of the muscle cells. The endo- 


derm of the coelenterates, though its cells have numerous muscle 
processes, contains relatively few sensory and neural cells, and fibrous 
branches of these cells are but little developed. 

The polychaete annelid larva of the trochophore type (fig. 8) has 
a muscular system of which the elements appear to be quite analogous 
to the ectodermal muscles of the coelenterates, though the system 
itself is carried to a higher degree of development. Furthermore, 
the larval muscles are parts of a neuromuscular system, since gener- 
ally they follow the inner surfaces of nerve tracts in the ectoderm. 
The muscle fibers are formed from cells derived directly from the 
larval ectoderm, along with numerous small undifferentiated cells 
that constitute a loose layer of mesenchyme distributed through the 
haemocoele. The fibers are arranged principally in longitudinal and 
circular tracts, though some of them extend from the body wall to 
the alimentary canal. The endoderm of the larva does not produce 
directly either muscular or neural cells. The nervous system of the 
polychaete larva, when best developed, consists of longitudinal and 
circular strands of ectodermal nerve cells and fibers following the 
muscle tracts, and of ganglionic groups of nerve cells developed 
particularly in connection with sensory organs on the preoral part 
of the body. The larval elaboration of the neuromuscular system is 
largely a temporary adaptation to the specialized form and habits of 
the trochophore, for most of it is lost when the larva undergoes its 
metamorphosis into the definitive worm form ; but the preoral part 
of the larval nervous system forms the brain of the adult, and some 
of the larval muscle fibers are taken over into the definitive muscular 

If now we endow our hyix»thetical annelid ancestor (fig. 2 E) with 
a primitive neuromuscular system derived from the ectoderm, and 
provide it with a pair of primitive nephridia, it will have reached an 
evolutionary stage entirely comparable in structure with that of 
an annelid in the ontogenetic stage of the young polychaete larva. 
The usual trochopore larva of the Polychaeta (fig. 4 A), however, 
leads a purely pelagic life ; it floats upright in the water and swims 
by means of bands of cilia that encircle the body. Its radial and 
circular neuromuscular system appears to be entirely adapted to its 
upright position, and many zoologists have regarded the trochophore 
as the ancestral form of the annelids as well as of various other 
invertebrates. The lateral position of the mouth, however, just 
below the principal circle of cilia (Mth), gives us good reason for 
suspecting that the shape of the trochophore and the position assumed 
in the water are secondary adaptations to a brief swimming existence ; 



VOL. 97 

in fact, the later horizontal development of the worm form along 
the vertical axis of the larva shows clearly that the trochophoral 
position is one quite out of harmony with the general organization 
of a worm. 

The trochophore, therefore, is to be regarded as a temporary, 
specialized larval form in polychaete ontogeny, adapted to a free 
pelagic life for the purpose of disseminating the individuals of its 
species. The metamorphic alterations that it undergoes at its trans- 
formation to the worm are changes of a nature that could not have 
been a part of the phylogenetic evolution of any animal. On the 

_ Fig. 4. — The polychaete trochophore and the crustacean nauplius, two spe- 
cialized larval forms of an early ontogenetic stage, having, therefore, primi- 
tive characters, but no phylogenetic significance in their shape or general 

A, typical structure of a trochophore, diagrammatic. B, nauplius of a cirriped, 
Alcippe lampas Hancock, dorsal surface (from Kiihnert, 1935). 

An, anus ; lAnt, first antenna ; 2Ant, second antenna ; ApGng, apical ganglion ; 
ApPl, apical plate; CNv, circular nerve; Epsp, episphere; Hpsp, hyposphere; 
Ih, lateral horn; LNv, longitudinal (radial) nerve; Md, mandible; Mcnt, 
mesenteron; MsT, mesodermal teloblast; Mth, mouth; Nph, nephridium; O, 
naupliar ocellus; Pair, paratroch; Proc, proctodaeum; Prtr, prototroch; Stom, 

Other hand, inasmuch as the trochophore is an early ontogenetic stage, 
its general organization is primitive, and is repeated in the onto- 
genetic development of many other invertebrates besides the annelids. 
It should be noted, furthermore, that the trochopore is not a uni- 
versal larval form even among the annelids, for most of the archi- 
annelids, some of the polychaetes, and all the oligochaetes have a 
direct development, in which either there is no suggestion of the 
trochophore form, or a remnant of it is preserved from ancestors 
that had a typical swimming larva. 


The presence of a mesoblastic muscle system and of a mesenchyme, 
or parenchymatous layer between the ectoderm and the endoderm, 
gives the annelid larva, or the platyhelminth adult, the status of a 
triploblastic animal; but the middle layer is here only an elaboration 
of elements present also in the so-called diploblastic coelenterates. 
The young annelid larva, however, is endowed from its parents with 
hereditary influences that will mold its growing tissues into structures 
never attained by the coelenterates or flat worms. Particularly affected 
are two individualized groups of mesoblast cells, which, though they 
may be set apart in the platyhelminths, will give rise in the annelids 
to special bands of mesoblastic tissue, known as the mesoderm. 
Within the mesoderm will be formed a new body cavity, the coeloni, 
and from the walls of the latter will be produced a new muscular 
system, a more efficient excretory system, a circulatory system, and 
various tissues of special functions, to all of which is added an exten- 
sion and elaboration of the nervous system. With the formation of 
the mesodermal cavities the triploblastic annelid larva becomes a 
coelomate animal, but, shortly before the appearance of the coelom, 
there takes place a segmentation of the body afifecting the ectoderm 
and the mesoderm, so that the young annelid worm is almost at once 
a segmented and a coelomate animal. 


In the ontogeny of the articulate animals, the formation of the 
coelomic cavities in the mesoderm is so closely associated with the 
appearance of body segmentation as to give the impression that the 
two are intimately related developmental processes, and since the 
segmentation of the mesoderm is usually more conspicuous than the 
segmentation of the body, embryologists often describe metamerism 
in terms of mesoderm segmentation, as if the formation of "meso- 
derm somites" were equivalent to body segmentation. Closer atten- 
tion recently given to the sequence of events in the development of 
the Polychaeta, however, shows that metamerism begins in the ecto- 
derm and the primary ectodermal musculature, and that it secondarily 
effects a division of the coelomic mesoblast into segmental sections. 
Subsequently, the coelomic cavities are formed in the segmented 
mesoderm. That coelomic sacs do not determine metamerism is 
shown also by the formation of paired coelomic cavities in the preoral 
cephalic mesoderm of the Onychophora and Arthropoda, in which 
there is no corresponding external segmentation. 

Metamerism, therefore, probably took its origin in a subdivision 
of the primary somatic musculature into successive sections (myo- 


tomes) to give greater efficiency to body movement. The segmen- 
tation of the ectoderm and the mesoderm then followed as a result of 
the segmentation of the muscular system. The primitive coelomic 
cavities were probably spaces formed' in the mesoderm for the accu- 
mulation of waste products in the body fluid, to be discharged through 
primitive nephridial tubules. The coelomic mesoblast, however, 
formed also a secondary musculature that reinforced the primary 
musculature, and which, in the higher arthropods, has completely 
replaced the latter. Evidence that such has been the course of evolution 
in the Articulata will be shown in the following discussion of the 
early stages in annelid ontogeny ; but there still remains the question 
as to the origin and nature of the primitive mesoderm, which antedates 

A study of the growth and differentiation of the annelid mesoderm 
takes us into the later part of larval development, but to obtain light 
on the origin of the middle germ layer we must go back to an earlier 
ontogenetic stage. During cleavage of the annelid egg most of the 
yolk remains consistently in the blastomeres situated on the vege- 
tative surface of the blastula (fig. 5 A), with the result that, in the 
64-cell stage, there are 8 large, yolk-filled blastomeres at the posterior 
pole (B). These cells are designated by embryologists 4A, 4B, 4C, 
4D, and 4a, 4b, 4c, 4d, since they comprise the so-called macromeres 
of the fourth generation and the fourth quartet of micromeres. All 
of them at this stage would appear to be endodermal, and at the time 
of gastrulation they all become internal, owing to their overgrowth 
by the ectoderm. Seven of them, in fact, give rise to purely endo- 
dermal progeny, but the 4d cell will form in most cases both endo- 
derm and mesoderm. The first cleavage of 4d produces two bilaterally 
symmetrical cells, 4d^ and ^rf" (C), and these cells, in their immedi- 
ately following divisions, give rise to a few very small cells (D, end), 
usually regarded as endoderm cells, and a pair of large cells {MsT) 
that are destined to produce the coelomic mesoblast, and hence con- 
stitute the mesodermal teloblasts. (It is perhaps possible that the 
small "endoderm" cells of this generation are the primary germ cells.) 

The common occurrence in the annelids of mesodermal teloblasts 
derived from cells closely associated with the endoderm has given 
rise to the idea that the coelomic mesoblast is of endodermal origin, 
and for this reason it is often called the "endodermal mesoblast" to 
distinguish it from the larval mesoblast, which is derived from the 
ectoderm. In most animals the mesoderm is, one way or another, 
associated in its origin with the endoderm, but among the annelids 
there are many cases where its endodermal connection is not evident. 



Fig. 5. — Late cleavage stages and mesoderm formation in Annelida and 

A, diagram of posterior pole of annelid blastula showing four yolk-filled 
"macromeres" of third generation. B, posterior pole of blastula of Arenicola 
cristata Stimpson after next cleavage forming fourth quartet of "micromeres," 
showing differentiation of 4d blastomere (adapted from Child, 1900). C, same, 
after cleavage of 4d into 4(f and 4d' (adapted from Child, 1900). D, blastula 
of Podarkc ohscura Ehlers, showing mesodermal teloblasts (MsT) derived 
from blastomeres 4d^ and 4d^ (C) after separation of small endoderm cells 
(from Treadwell, 1901). E, posteroventral view of 40-hour embryo of Podarke 
obscura with mouth and anus, showing position of mesoderm bands (Msd) in 
body (from Treadwell, 1901). F, optical frontal section of embryo of Capitclla 
capitata Fabr., showing mesodermal teloblasts and rudiments of mesoderm 
bands (from Eisig, 1899). G, optical section of embryo of DinophUns sp., with 
mesoderm bands {Msd) extending forward from teloblasts (from Nelson, 
1904). H, optical section of blastula of Planoccra inquUina Wheeler (Poly- 
cladia) from right side just after division of 4d, producing 4d^ that will form 
endoderm, and -/</" that will form mesoblast (from Surface, 1907). I, same, 
later stage seen from posterior pole, showing mesoblast (Msb*) derived from 
^cP, and mesoblast (Msb^) derived from second quartet of ectodermal blastomeres 
(from Surface, 1907). 

3A-3D, 4A-4D, "macromeres" of third and fourth generations; 4a-4d, "micro- 
meres" of fourth quartet; An, anus; Bp^r, blastopore; Br, brain; 4d\ ^(f, 
daughter cells of 4d blastomere ; Ecd, ectoderm ; End, endoderm ; end, endoderm 
derived from 4d^ and ./(f blastomeres ; Msb', mesoblast derived from second 
quartet of micromeres ; Msb\ mesoblast derived from 4d cell of fourth quartet ; 
Msd, mesoderm ; Msnc, mesenchyme ; MsT, mesodermal teloblast ; Mth, mouth. 


It is claimed by both Kleinenberg (1886) and E. Meyer (1901), for 
example, that in the larva of Lopadorhynchiis the mesoderm arises 
from the ectoderm, and in Capitella, according to Eisig (1899), the 
coelomic mesoblast is produced from blastomeres other than 46. 
Furthermore, the mesoderm of the postlarval somites is said by 
Iwanoff (1928) to be formed in many polychaetes directly from the 
ectoderm, and the same is probably true in cases of regeneration. The 
mesoderm of certain other coelomate invertebrates also may have no 
genetic relation to the endoderm, as in the gastropod Paludina, in 
which the embryonic mesoblast that gives rise to the usual mesodermal 
organs is generated directly from cells of the ventral ectoderm (see 
Dautert, 1929). 

During larval life, or at the transformation of the larva to the 
worm, the annelid mesodermal teloblasts, however formed, proliferate 
within the haemocoele two masses of mesoderm cells (fig. 5 E, F, G, 
Msd), which eventually take the form of ventrolateral bands extend- 
ing forward at least as far as the sides of the mouth (fig. 6 F). These 
primary mesoderm bands are solid cell masses; they are never ob- 
served at this early stage to contain cavities, and there is no evidence 
from annelid embryology to suggest that they represent phylogeneti- 
cally a pair of open pouches. Later, with body segmentation, the 
bands are broken up into solid segmental blocks (G), and finally the 
blocks are excavated by coelomic cavities (H). The nature of the 
mesoderm and the primitive function of the coelomic cavities can be 
better discussed after we have examined the known facts concerning 
the beginning of metamerism, but it should be noted here that the 
formation of the mesoderm bands precedes body segmentation. 

Metamerism in the polychaete larva becomes first evident as a 
subdivision of the body region between the mouth and the pygidium 
into a small number of somites (fig. 7 A, I, II, III). There is ample 
reason to believe, as Iwanoff (1928) claims, that the formation of 
these primary somites, or larval segments of ontogeny, represents the 
beginning of metamerism in phylogeny, and, as we shall see, the same 
phenomenon of direct segmentation in the body of the embryo or 
young larva recurs in various arthropods. The primary somites are 
thus to be distinguished from the secondary somites later added by 
teloblastic growth in a subterminal generative zone, and which will 
constitute the major part of the adult animal. The larval somites of 
the Polychaeta, Iwanoff shows, are formed approximately simul- 
taneously in contrast with the successive, individual generation of 
the teloblastic somites. E. Meyer (1901) observes that in Lopa- 
dorhynchus metamerism takes place so rapidly as to give the impres- 


sion that a relatively large number of somites are formed all at once, 
but Sokolow (1911) says that in Ctcnodrilns the intermediate somites 
or the more anterior ones are first differentiated and the series then 
completed anteriorly and posteriorly. Segmentation may be delayed 
until the beginning of metamorphosis, as in Polygordius, or it may 
take place while the larva is still in the swimming trochophore stage. 


Fig. 6. — Transformation of the annelid blastopore, primary segmentation of 
the body, growth and segmentation of the mesoderm bands, and formation of 
the coelomic sacs, diagrammatic. 

A, blastopore and mesodermal teloblasts at posterior pole of embryo. B, 
blastopore elongating forward on ventral surface ; rudiments of mesoderm gen- 
erated from teloblasts. C, blastopore still more elongate, closing posteriorly ; 
mesoderm growing forward. D, blastopore closed posteriorly ; mesoderm bands 
extended to prostomium. E, blastopore obliterated except for mouth opening 
at anterior end ; anus formed secondarily at posterior end ; mesoderm segmented 
following metamerism of body, and extended into prostomium. F, polychaete 
trochophore before segmentation. G, same after segmentation, mesoderm cut 
into solid segmental blocks. H, same, mesoderm blocks excavated by coelomic 

AlCnl, alimentary canal; An, anus; Bpr, blastopore; Cod, coelomic cavity; 
Epsp, episphere ; Msd, mesoderm ; MsT, mesodermal teloblast ; Mth, mouth ; 
Prst, prostomium ; Pyg, pygidium ; ZG, zone of growth. 

In Polynoe, as described by Hacker (1895), seven somites are first 
marked out in the body of the trochophore, which is transformed 
while still active into a swimming "nectochaete" larva with seven 
segments and corresponding chaeta-bearing parapodia. The number 
of larval somites is always small, three or four being usual (fig. 7 A, 
B, C), the maximum not more than 13. Completion of larval meta- 
merism is followed by a pause in development. 



Observations on the beginning of embryonic segmentation in the 
annehds differ somewhat as to whether the intersegmental divisions 
appear first in the ectoderm or in the mesoderm, but most students 
of anneHd development find either that the ectoderm and the meso- 

FiG. 7. — Examples of primary segmentation in polychaete larvae. 

A, Enpamatiis micinatus, trochophore showing primary segmentation of the 
mesoderm (from Iwanoff, 1928). B, same, later larval stage, horizontal sec- 
tion showing development of chaetal sacs in primary somites, and extension of 
posterior part of body (from Iwanoff, 1928). C, Platyncrcis dunwrilii And. 
& Milne-Edw., nereidogen larva just out of egg, with four primary somites 
(from Hempelmann, 1911). D, CapitcUa capitafa Fabr., embryo before seg- 
mentation, ventral view (from Eisig, 1899). E, same, embryo with seven 
somites and zone of growth formed directly in primary body region (from 
Eisig, 1899). F, same, later stage with two additional somites formed from 
zone of growth (from Eisig, 1899). 

E, eye ; I-IX, somites ; Mth, mouth ; Prst, prostomium ; Pyg, pygidium ; SPl, 
somatic plate; Tl, tentacle; ZG, zone of growth. 

derm are segmented at the same time, or that the first signs of meta- 
merism are to be seen in the ectoderm. 

In the development of the polychaete Capitella, according to Eisig 
(1899), on the sixth day after fertilization of the egg, the cells of the 
ventral somatic plates of the larva (fig. 7 D, SPl) become arranged 
in transverse rows, and on the sixth day seven or eight somites are 
already demarked by transverse lines in the ectoderm of the larval 


body region between tbe mouth and the pygidium (E). On the same 
day, however, the mesodermal bands also become divided into seg- 
mental sections. At first the ectodermal and mesodermal somites of 
Capitella do not entirely correspond, there being several super- 
numerary mesodermal divisions in the mouth region, but by the 
twelfth or thirteenth day the larva has 13 somites with coincident 
limits in both the ectoderm and the mesoderm. 

The segmentation of the mesoderm bands as described by E. Meyer 
(1901) in Psygmobranchns, Polygordius, and Lopadorhynchus ap- 
pears to be determined by elements of the mesenchymatic primary 
mesoblast in the form of spindle-shaped cells that penetrate into the 
mesoderm bands at the intersegmental lines and cut the bands into a 
series of segmental sections. From the penetrating mesenchyme cells 
are later formed, according to Meyer, the muscles of the interseg- 
mental dissepiments. Similarly in the Serpulidae and Spionidae the 
larval segmentation is said by Iwanoff (1928) to be secondarily im- 
posed upon the mesoderm bands by metamerism in other parts of the 
body, as by the ectodermal segmentation, the ingrowth of the chaetal 
sacs (fig. 7 B), the penetration into the mesoderm of mesenchymatous 
muscle elements, or by the segmental formation of blood lacunae in 
the general mesoderm mass. 

The primary larval segments are seldom as fully developed in the 
adult worm as are the teloblastic segments, and both the segment 
limits and the differentiation of ganglia on the nerve cords may remain 
obscure. In the Spionidae, Iwanoff (1928) says, the trochophoral 
mesoderm is very weakly developed, the dissepiments are only imper- 
fectly formed, often absent, and in some species a segmentation of 
the mesoderm in the primary segments is absent even in the adult. 
Chlorogogen cells are not developed in the coelomic walls of the larval 
segments, and in these segments germ cells are never present. 

As a result of body metamerism, the mesoderm bands are divided 
each into a series of segmental sections, and these sections, as the 
bands themselves, are at first solid blocks of cells (fig. 6 G). Later the 
coelomic cavities appear as cleavage spaces within the cell blocks (H). 
Hence, just as there is no evidence that the primary mesoderm bands 
represent primitive sacs, so there is no evidence from ontogeny that 
the coelomic cavities of the annelids took their origin as a series of 
separate mesodermal pouches. The facts of development suggest only 
that the primitive mesoderm bands were continuous tracts of cells, 
and that the formation of cavities within them was a secondary 
process, subsequent to segmentation. 

With the formation of the coelomic cavities in the mesoderm, the 
young annelid becomes a coelomate animal. Before the appearance 


of the coelom, however, it might pass for the ancestor of a flatworm, 
for even in the Platyhelminthes there is a teloblastic proHferation of 
cells that appear to correspond with the mesoderm cells of the anne- 
lids, though the cells thus produced soon disperse and become a part 
of the parenchyma. It is in the development and elaboration of the 
mesoderm, or teloblastic mesoblast, therefore, that the Coelomata 
depart from the Platyhelminthes. Segmentation is a feature super- 
imposed upon the mesoderm in the Annelida as a result of body 
metamerism, in which apparently the ingrowth of the septal muscles 
plays an important part. 

The mesoderm of the adult annelid or arthropod gives rise to a 
large variety of tissues and organs, but most of the specialized deriva- 
tives of the mesoderm are formed in the secondary segments of the 
adult animal. The principal products of the primary mesoderm are 
muscle and connective tissues, and an epithelial lining of the coelomic 

According to E. Meyer (1901), the mesodermal myoblasts of the 
polychaete larva are not recognizable as such until the mesoderm 
bands have become broken up into segmental sections, and the trans- 
formation of the myoblasts into functional muscle fibers is not evident 
until after the appearance of the coelomic cavities. The myoblasts of 
each mesodermal segment, Meyer says, consist of four large cells on 
each side, two dorsal and two ventral, lying along the lines of the 
larval longitudinal muscles of mesenchymatic origin already present. 
The mesoderm fibers finally replace the mesenchyme fibers and be- 
come the definitive longitudinal muscles of the worm. The coelomic 
myoblasts, Meyer shows, are true epithelial muscle cells that form 
muscle processes from their outer surfaces, while the plasmatic bodies 
of the cells retain their places for some time in the coelomic walls. 
The parts of the coelomic walls not involved in muscle formation be- 
come thinner, and finally transform into typical peritoneal epithelium. 

The important part that the mesoderm plays in the development of 
the coelomate animals is entirely clear ; but what the mesoderm be- 
comes in the course of evolution does not explain what it was in its 
beginning. Most of the theories that have been proposed to account 
for the primitive mesoderm, it will be found, are attempts to explain 
the functional nature of the coelomic sacs rather than the origin of 
the mesoderm itself. 

The theory most widely accepted at one time as to the origin of 
the mesoderm is the enterocoele theory (Hertwigs, 1882, Sedgwick, 
1884), by which the coelomic sacs are explained as diverticula of the 
archenteron. In some animals the coelomic sacs are thus formed in 


the embryo, and the enterocoele theory has some plausibility as a 
wide generalization, considering the very common early association 
of the coelomic mesoderm rudiments with the endoderm ; but, as 
applied to the annelids and arthropods, the theory must entirely discard 
the direct evidence from embryology that the mesoderm first appears 
as solid proliferations of cells, which only in a purely hypothetical 
manner could be interpreted as representing pouches of the archen- 
teron. The only known case of the formation of the mesoderm from 
enteric pouches that might be referred to the articulates occurs in 
the Tardigrada (see Marcus, 1929), but there is much uncertainty 
concerning the relationships of the tardigrades. 

A second mesoderm theory is the gonococle theory, based on the 
almost universal association of the germ cells with the coelomic meso- 
derm in the coelomate animals. Hatschek (1877, 1894) believed that 
the mesodermal teloblasts of the annelid larva are themselves germ 
cells, and Rabl (1879, 1889) adopted this view. The gonocoele theory 
of the origin of the coelomic sacs, however, was principally elaborated 
by E. Meyer (1891, 1901). Meyer contended that the primitive 
coelomic sacs were muscular pouches, from the epithelial walls of 
which the germ cells are generated, and that, as these gonadial sacs 
expanded to increase the reproductive function, they finally preempted 
the haemocoele, and their muscles were transferred to the body wall. 
The gonocoele theory loses much of its support now that the old belief 
that the germ cells are direct products of the coelomic epithelium 
is no longer tenable, and, moreover, it entirely breaks down in view 
of the fact that the primary larval somites of the annelids do not 
contain germ cells. In the primitive annelids, as will be shown later, 
the germ cells probably were located in the zone of undifferentiated 
tissue behind the last primary somite. If so, the reproductive function 
had nothing to do with the origin of the mesoderm or the formation 
of the coelomic sacs. 

A third theory, concerned principally with the function of the 
coelomic sacs, is the nephrocoele theory (Ziegler, 1898; Faussek, 1899, 
1901). According to Ziegler, the primitive coelomic cavities were 
open pouches for the accumulation of waste products ; they were not 
diverticula of the archenteron, but were, perhaps, of the nature of 
protonephridia. The nephrocoele theory as modified by Faussek holds 
that the excretory coelomic sacs are not primitive structures in a 
phylogenetic sense, but that they have been developed for excretory 
purposes in the embryo, and are hence purely ontogenetic organs. 
Faussek supports his theory with the generalization that the open 
metanephridia constitute exits from the coelom, while the closed 


protonephridia serve for removal of waste products from the haemo- 
coele. This statement, however, is not entirely true, for in some of 
the Polychaeta protonephridia are associated with coelomic sacs, and 
the primary larval somites of the annelids do not have metanephridia. 
On the other hand, there can be no question that the coelomic fluid 
does contain waste products of metabolism. 

A fourth theory, that of Kleinenberg (1886), identifies the primi- 
tive mesoderm with muscle tissue, and is thus more satisfactory than 
the other theories because it deals with the beginning of the meso- 
derm as a functional tissue. Kleinenberg attributes the idea of a 
muscle origin for the mesoderm to Rabl, who later discarded it, but 
the theory rests principally on Kleinenberg's studies of the develop- 
ment of Lopadorhynchus. Kleinenberg claimed that in the larva of 
Lopadorhynchus the mesoderm is derived directly from the ectoderm 
at the posterior end of the body, and that the ectodermal m3^oblasts, 
and the neuroblasts of the ventral nerve cords, arise from a common 
neuromuscular rudiment. The mesoderm bands, or "muscle plates," 
become divided into segmental myotomes consequent on metamerism 
of the body, and the myotomes give rise to the body musculature, 
including, according to Kleinenberg, the dorsal and ventral longi- 
tudinal muscles, the parapodial muscles, and the circular muscles of 
the body wall. Then follows a separation of the muscle plates into 
parietal and visceral layers in each somite, producing thus the paired 
coelomic cavities, the peritoneal linings of which are formed by the 
inner cells of the myotomes. Kleinenberg's theory of the origin of 
the mesoderm thus gives to metamerism a mechanical significance, 
since it explains body segmentation as an adaptation to more efficient 
locomotion. Certainly, when once established, the chief function of 
metamerism is effective movement of the body, and to this feature 
the segmented annelids owe their superiority over the unsegmented 
flatworms. A serious weakness of the muscle theory of the origin 
of the coelomic mesoblast, however, is found in the fact that so many 
tissues other than muscle are evolved from it. Muscle fiber is a highly 
specialized tissue, and it seems hardly likely that epithelial tissue, for 
example, would be formed from muscle cells, since ordinarily it is 
epithelial tissue that gives rise to muscle fibers and to the various 
other specialized tissues of the body. Furthermore, as shown by 
Meyer (1901), muscle is not formed from the coelomic mesoblast of 
Lopadorhynchus until after the segmentation of the mesoderm bands 
and the formation of the coelomic cavities. 

The literature of annelid morphology is replete with discussions 
on the nature and difference of the "two kinds of mesoblast" ; but 


the facts concerning the ontogenetic origin of the annehd mesoblast 
apparently can be expressed in the simple statement that mesoblastic 
tissue may be formed by internal proliferation from any part of the 
blastoderm, and may, therefore, be both "ectodermal" and "endo- 
dermal." The mesoblast of the first three quartets of the blastula 
(see Torrey, 1903) gives rise to the so-called larval mesoblast, or 
mesenchyme ; from the fourth quartet ordinarily arises the coelomic 
mesoblast, or mesoderm. That these two groups of mesoblast cells 
primarily have the same morphological status is indicated by the fact 
that in the Platyhelminthes they do not become differentiated into 
separate tissues. Surface (1907), who first followed the divisions of 
the 4d cell in a flatworm, shows that in Planocera the 4d blastomere 
gives rise to both endoderm and mesoblast as it does in the annelids, 
since, of the two cells of the first division, 4d^ (fig. 5 H) forms the 
endoderm (I, End), and 4d^ gives rise to two lateral groups of scat- 
tered mesoblast cells (I, Msb'^), which are at first quite distinct from 
the mesoblast of the second quartet {Msb~), though eventually they 
intermingle with the latter to form the parenchymatous tissue of the 
adult. In Planocera the usual endodermal "macromeres" degenerate 
and almost the entire endoderm proceeds from the 4d^ cell. Finally, 
we may correlate the "double origin" of the annelid mesoblast with 
the production of muscle tissue from both the ectoderm and the 
endoderm in the Coelenterata. 

From the condition in the Platyhelminthes, it becomes evident that 
the primitive mesoblast was a parenchymatous mass of undiffer- 
entiated cells occupying the haemocoele, which had been proliferated 
internally from both the ectoderm and the endoderm. In the unseg- 
mented ancestors of the annelids, the ectodermal mesoblast must have 
formed a primary somatic muscular system, represented by the larval 
musculature of modern annelids, which is derived from the ectodermal 
quartets of the blastula. The principal part of the parenchyma, there- 
fore, came to be that part of the mesoblast proliferated in the posterior 
part of the body, chiefly, or usually, from the 4d cell of the fourth 
quartet. The persistent parenchyma thus became the embryonic 
middle layer known specifically as the mesoderm. 

Since the most important result of metamerism is the production 
of a mechanism of movement based on the division of the body into 
consecutive motor units, it can scarcely be questioned that meta- 
merism had its origin as an adaptation to more effectivt body move- 
ment. Inasmuch as the evidence from embryonic development shows 
that metamerism originates ontogenetically in the ectoderm and its 
derivatives, and is secondarily imposed upon the mesoderm, we may 


suppose that it took its inception phylogenetically from an attachment 
of the primary (ectodermal) longitudinal somatic muscles at con- 
secutive rings on the body wall, and from the accompanying ingrowth 
of fibers that formed contractile dissepiments between the myotomes. 
The ingrowth of the septal muscles cut the parenchymatous meso- 
dermal bands into segmental blocks. This modification and elaboration 
of the primitive muscular system, and the consequent segmental 
division of the mesoderm bands, give at once the essential quality of 
metamerism, and from it there follows as a necessary result the 
metamerization of other organs, such as external ectodermal struc- 
tures, the ventral nerve cords, and all structures of mesodermal origin. 

The coelomic cavities first appear in the annelid embryo or larva 
as cleavage spaces in the segmental mesoderm blocks. Since the un- 
segmented Platyhelminthes have nephridial organs, it may be assumed 
that the primitive annelids possessed simple segmental nephridia in 
the form of internally closed tubules extending into the haemocoele. 
The primitive coelomic cavities, therefore, were probably spaces 
formed in the segmented parenchyma for the accumulation of body 
fluid charged with excretory products. The inner cells of the paren- 
chyma now formed epithelial walls about the nephric cavities, which 
became the coelomic sacs ; the outer cells were converted largely into 
muscles and connective tissue. The muscle cells gave rise to fibers 
that reinforced the somatic musculature, and eventually came to 
be its principal constituents. The definitive musculatvire of modern 
annelids, therefore, is a composite of fibers derived from the larval 
ectoderm and of fibers formed from the coelomic mesoblast, but in 
the Onychophora and the Arthropoda the entire musculature appears 
to be now a coelomic product. There is no reason necessarily for 
supposing that the primitive mesodermal muscles were functional 
elements of the coelomic sacs, for, though in ontogeny the mesoderm 
usually takes the form of two-layered bands of cells, within which 
the coelomic cavities are formed, it would seem probable that the 
primitive mesoderm was a loose parenchymatous tissue. The coelomic 
sacs are specifically the epithelial walls formed about the nephric 
cavities ; the surrounding muscles were probably generated from the 
outer undifferentiated cells of the original parenchyma. 

With the later development of the teloblastic somites, into which 
the germ cells were distributed from their posterior source of pro- 
liferation, the reproductive products were discharged into the coelomic 
sacs of these somites, which thus became gonocoelic as well as nephro- 
coelic in function. Open nephridia or coelomoducts now connected 
the coelomic cavities with the exterior and served both as excretory 


and as genital outlets. Finally, in the Onychophora and the Arthrop- 
oda, the coelomic sacs have been divided into gonadial compartments 
and nephridial compartments, which have become reduced in size and 
limited to restricted parts of the body, with the result that the 
haemocoele is restored as the functional body cavity. 


The annelids and the arthropods undoubtedly have a closer bond 
of union in the structure of the nervous system than in any other 
feature of their organization, except metamerism itself. The definitive 
central nervous system of the polychaete annelids is developed from 
two distinct sources, one located in the prostomium, or episphere of 
the trochophoral larva, the other in the somatic region, or hyposphere 
of the larva. From the first is produced the brain ; from the second, 
the ventral nerve cords. The nervous system of the trochophore con- 
sists of ganglionic centers in the prostomium connected by circular 
and radial nerve tracts, from which trunks proceed into the hypo- 
sphere (fig. 4 A). This primary system centering in the prostomium 
must represent the primitive neural system of the unsegmented an- 
cestors of the annelids, adapted to the structure of the trochophoral 
larva, and is probably congenetic in its origin with the nervous system 
of the Platyhelminthes. The segmentally ganglionated ventral nerve 
cords of the postoral region of the trunk are correlated in their 
development with the development of body metamerism ; they pertain, 
therefore, to a later stage of evolution, and have no homologues in 
the unsegmented worms. The definitive connection between the pro- 
stomial and somatic parts of the nervous system is established secon- 
darily in the ontogeny of the polychaetes, but in the oligochaetes the 
two parts are said to be continuous from their inception. The funda- 
mental structure of the somatic nervous system of the articulate 
animals is an adaptation to the function of regulating the muscular 
mechanism of metameric body movement; the prostomial system is 
primarily sensory in function, except insofar as it controls the move- 
ments of prostomial appendages. 

The phylogenetic origin of the articulate nervous system can prob- 
ably be interpreted very closely from the development of the neural 
elements in the trochophore larva of the polychaete annelids, and must 
have been about as follows : The primary neurocytes were probably 
sensory cells of the ectoderm closely associated with the primary myo- 
cytes, and were thus, at first, both receptive and motor in function. 
As the muscular system became elaborated, however, the primary 
neurocytes were withdrawn to the inner surface of the ectoderm. 



VOL. 97 

while other superficial cells assumed the receptive function and trans- 
mitted secondarily the impulses from external stimuli to the first set 
of cells, which now became purely motor neurones. Finally, still other 
neurocytes gave rise to a subepidermal plexus of fibrous tracts that 
formed lines of intercommunication between the scattered motor and 
sensory elements, and thus unified and coordinated the entire nervous 
system. Then the nerve cells of the prostomial region became aggre- 
gated into a number of ganglionic centers, principally associated with 
groups of receptive cells in primitive sensory organs, and the nerve 




Fig. 8. — The nervous and muscular elements of a young trochophore larva of 
Lopadorhynchus, nerve tissue represented in white, muscle tissue in black. 
(From E. Meyer, 1901.) 

A, aboral surface of larva. B, oral surface. 

da, rudiments of so-called dorsal antennae ; dn, median dorsal nerve of hypo- 
sphere; gSo, ganglion of left ciliary organ; Ks, apical plate; mcl, muscle fibers; 
Msd, mesoderm; n, longitudinal nerves (seven pairs in episphere) ; NCls, nerve 
cells; Rn, circular nerve of prototroch ; riio^, ruo^, rno^, circular nerves of 
episphere; rnu, circular nerve of hyposphere; So, left ciliary organ; so, rudimen- 
tary right ciliary organ ; Stom, stomodaeum ; Vdn, larval stomodaeal nerve. 

tracts assumed definite courses. Thus was evolved the primary 
nervous system of the polychaete larva. The prostomial ganglia of 
this system later coalesce to form the definitive brain. The somatic 
nervous system, subsequently developed in correlation with meta- 
merism, took its origin from restricted ventrolateral tracts of the 
somatic ectoderm, became connected with the brain, and finally re- 
placed the primary system in the body region. 

The most primitive nerve center of the annelids probably is repre- 
sented by the apical ganglion of the polychaete trochophore (fig. 4 A, 
ApGng) situated beneath the ectodermal apical plate (ApPl), which 


usually bears a tuft of cilia, and with which there may be associated a 
pair of small larval tentacles, and sometimes a pair of "eye spots." 
From tne apical ganglion, nerves radiate posteriorly (LNv) on the 
inner surface of the epidermis, and these longitudinal radial nerves 
are connected by bands of circular fibers (CNv), chief of which is 
the nerve ring of the prototroch (Prtr). The nerve tracts, both radial 
and circular, closely follow the peripheral muscle bands of the larva 
(fig. 8), thus attesting that the nervous and contractile elements arose 
from common ectodermal neuromuscular rudiments. The nerve tissue 
is situated between the muscle fibers and the epidermis, the nerve cells 
being scattered individually, or condensed in small ganglionic groups. 

The nervous system of the polychaete trochophore is best known 
from the elaborate studies of Kleinenberg (1886) and of E. Meyer 
(1901) on the larval development of Lopadorhynchus, a small errant 
polychaete of the family Phyllodocidae (fig. 13 D) having two pairs 
of prostomial tentacles but no palpi. The larva of Lopadorhynchus 
is a typical trochophore (fig. 8) with an equatorial band of cilia, the 
prototroch, just above the mouth. The apical ciliary organ, however, 
does not have the usual form and position ; it is transposed to the 
anterior ventral surface, and is divided into a well-developed organ 
on the left (B, So), and a rudimentary organ on the right (so). The 
episphere contains seven pairs of longitudinal nerves (w^-n^), and is 
encircled by three nerve rings {rno^-rno^) above that of the prototroch 
(Rn). In the hyposphere there is but a single nerve ring (A, rnu). 
The largest of the longitudinal nerves are two thick lateroventral 
nerve tracts (B, «-), which anteriorly (apically) are continuous with 
each other in a wide transverse commissural arch within the episphere, 
and posteriorly are extended into the hyposphere as a pair of large 
lateral trunks {Vdn) that break up into smaller branching nerves. 

The neural cells of the Lopadorhynchus larva are described in great 
detail by Meyer. In general they lie along the fiber tracts (fig, 8 B, 
NCls), where many of them are aggregated into small ganglionic 
clumps, particularly in the episphere. In the early stages of develop- 
ment, according to both Meyer and Kleinenberg, the neurocytes are 
generated from the ectoderm in association with muscle cells, and the 
principal neuromuscular rudiments of the episphere represent larval 
sensory organs (fig. 16 A), of which the nerve cells {n) form small 
ganglionic centers. The scattered neurocytes are probably the gener- 
ative cells of the fibers in the nerve tracts. The ganglionic centers of 
the larva pertain to the apical ciliary organs, a pair of transient larval 
antennae, the two pairs of persistent tentacles, which are dorsal and 
ventral in the adult (fig. 13 D), and the nuchal organs, but include 



VOL. 97 

also two cell groups of unknown significance situated on the dorsal 
surface of the episphere (fig. 8 A, da). From some of the ganglion 
cells nerve processes go to the muscles, and from others fibers pene- 
trate centrally into the nerve tracts. 

Before the beginning of larval metamorphosis, Meyer says, the 
production of myocytes ceases in the larval neuromuscular centers, 
and during metamorphosis a large part of the larval musculature is 

Fig. 9. — -Theoretical evolution of the annelid nervous system, diagrammati- 
cally following Kleinenberg's and Meyer's accounts of the development of the 
nervous system in the larva of Lopadorhynchiis. 

A, early trochophore with diffuse nerve cells (NCls) along the radial and 
circular nerve tracts. B, nerve cells of episphere concentrated in incipient 
ganglia connected with ectodermal sensory organs along lateral radial nerves ; 
ventral nerve cords (VNC) developed from ectoderm of hyposphere. C, scat- 
tered ganglia of episphere condensed into a cerebral mass (Br) ; cerebral con- 
nectives united with ventral nerve cords. D, generalized adult nervous system ; 
podial ganglia developed at bases of body appendages. E, nervous system of 
adult polychaete, lateral view. 

AlCiil, alimentary canal; A71, anus; Br, brain; E, eye; Mth, mouth; NCls, 
nerve cells; NO nuchal organ; PdGng, podial ganglion; Pip, palpus; Prst, 
prostomium ; Tl, tentacle ; VNC, ventral nerve cord ; ZG, zone of growth. 

lost. Many of the primary muscles, however, remain, including those 
of the persistent tentacles and nuchal organs, and certain other muscles 
of the prostomium. The neural cells of the various ganglionic centers 
of the larval episphere, on the other hand, increase in number until 
they become so crowded that details of their further development 
cannot be followed. The cells thus generated, however, are massed 
upon the large lateral nerve trunks of the episphere (fig. 8 B, n^) 
and their anterior commissure. In this manner there is formed from 


numerous agglomerated centers in the larval episphere (fig. 9A, B) 
a compact cellular and fibrous body of nerve tissue (C, Br), which 
becomes the brain of the adult worm (D). Hence, as Kleinenberg 
remarks, the developmental history of the brain in Lopadorhynchus 
shows how extraordinarily complicated in its origin is the cephalic 
ganglion even in the annelids. However, that details in the probable 
phylogenetic history of the nervous system are not necessarily reca- 
pitulated in ontogeny is shown in many annelids having a direct 
development, or one in which the trochophoral stage is passed within 
the tgg, for in such forms the brain is differentiated from the 

Mth 3 2Dsp 

Fig. 10. — Median vertical sections of the anterior end of an embryo of the 
viviparous polychaete CtenodrUus branchiatus Sokolow (Cirratulidae), show- 
ing extension of the mesoderm into the prostomium, and the direct development 
of the brain from the prostomial ectoderm. (From Sokolow, 191 1.) 

A, embryo before appearance of coelom, with mesoderm (Msd) extended into 
prostomium (PMsd). B, full-grown embryo, with coelom and dissepiments, 
coelomic cavity of prostomium (PCoel) continuous with coelomic cavity of 
first postoral somite (metastomium), which in the embryo is separated from 
second somite by a temporary dissepiment (iDsp). 

Br, brain; Coel, coelom; iDsp, first (temporary) dissepiment, behind first 
postoral somite; 2Dsp, 3Dsp, second and third (permanent) dissepiments; Ecd, 
ectoderm ; mcl, muscles ; Msd, mesoderm ; Mth, mouth ; PCoel, prostomial 
coelom ; PMsd, prostomial mesoderm ; Stom, stomodaeum. 

prostomial ectoderm as a single, compact mass of neural cells 
(fig. 10, Br). 

The larval innervation of the hyposphere gives way during meta- 
morphosis to the definitive body nervous system, consisting of the 
ganglionated ventral aierve cords and their peripheral nerves. The 
rudiments of this system appear first in the embryo as continuous 
strands of neurocytes proliferated in the ventral parts of the ecto- 
dermal somatic plate as the median edges of the latter unite to close 
the blastopore. The cords later become ganglionated by the segmental 
aggregation of their cells. The neuroblasts of the somatic nerve cords, 
Meyer believes, represent the nerve cells of a series of primitive 
ectodermal sense organs. Though there are no persistent remnants of 


such sensory organs in the anneUds, the so-called "ventral organs" of 
the Onychophora, from which the nerve cords are differentiated, sug- 
gest that the latter took their origin from ectodermal structures of 
some kind. 

The final connection between the brain and the ventral nerve cords, 
according to Kleinenberg and Meyer, is established by fibers that grow 
forward from the first ventral ganglia (fig. 9 B) and unite with the 
lateral nerve trunks (fig. 8 B, Vdn) extending posteriorly from the 
arms («") of the cerebral commissure. The union thus formed pro- 
duces the stomodaeal (circumoesophageal) connectives, through which 
the prostomial and somatic nerve centers are unified in the definitive 
nervous system. 

The peripheral subcutaneous nervous system of the adult worm is 
developed directly from scattered neurocytes of the ectoderm. To 
this system Kleinenberg ascribes the parapodial ganglia (fig. 9 D, 
PdGng), which, he says, are formed quite independently of the central 
system by groups of ectodermal neurocytes situated mesad of the 
parapodial bases. Secondarily, the parapodial ganglia send connecting 
nerves to the ventral nerve cords. 


The final development of the adult polychaete annelid from the 
larva depends upon the histogenic activity in the zone of undiffer- 
entiated cells situated between the last larval somite and the pygidium 
(fig. II B, ZG). Within this zone of growth is generated anteriorly 
a series of secondary postlarval somites (C, D), which does not repre- 
sent an extension of the body, but an expansion of a small part of it, 
since the new somites are interpolated between the primarily seg- 
mented larval body and the pygidium. The more anterior somites of 
the new series, being those first formed, are the first to acquire the 
mature structure. The teloblastic growth-process is the same whether 
the larva is a typical trochophore (fig. 12 B, D), or one more nearly 
resembling the adult worm (E, F, G), but in the first case a greater 
degree of metamorphosis accompanies the formation of the new 
somites. Hence, though we may eliminate the trochophore from our 
concept of the primitive annelid, we cannot dismiss the secondary 
formation of the teloblastic somites as a purely ontogenetic process — 
it must be explained in terms of phylogeny. 


The zone of growth, as described by Lillie (1906) in Arenicola 
cristata, is a mass of large clear mesodermal and ectodermal cells, 


which are frequently to be seen in the process of mitosis. Posteriorly 
the growing zone is sharply defined from the pygidial region, but 
anteriorly it passes by gradual transition into the more fully differen- 
tiated region in front. Its ectodermal cells, Lillie says, must be derived 
from the last transverse row of cells in the ectodermal somatic plate 
produced from the 2d cell of the embryo. The space between the 
ectoderm and the endoderm is filled with a mass of mesoderm cells 
very probably generated from the mesodermal teloblasts. Anteriorly 
the mesoderm of the growing zone is shut off by a roughly defined 

Fig. II. — Diagrams illustrating the direct primary segmentation of the body 
of a larval polychaete, and the growth of the worm by successive addition of 
secondary teloblastic somites generated in the subterminal zone of growth. 

A, larva with unsegmented soma and mesoderm bands. B, larval soma and 
mesoderm directly segmented. C, D, successive formation, from subterminal 
zone of growth, of teloblastic somites interpolated between primary larval 
somites and terminal pygidium. 

AlCnl, alimentary canal; Coel, coelomic cavity; E, eye; /-///, primary larval 
somites; IV -IX, secondary teloblastic somites; Msd, mesoderm; MsT, meso- 
dermal teloblast ; Pip, palpus ; PMsd, prostomial mesoderm ; Prst, prostomium ; 
Pyg, pygidium ; Soma, body region between prostomium and pygidium in which 
somites are formed ; 77, tentacle ; ZG, zone of growth at end of soma. 

transverse partition from the coelomic cavity of the somite before it. 
The first evidence of new somite formation is the appearance of an 
irregular space in the mesodermal mass of the zone of growth, which 
enlarges upward around the alimentary canal and becomes the coelomic 
cavity of the new soinite. (Arenicola has a dorsal mesentery but 
none beneath the alimentary canal.) The anterior coelomic wall 
is pressed against the preceding partition and becomes the posterior 
lamella of the dissepiment thus formed. Longitudinal muscle fibers 
make their appearance at an early period in the somatic layers of 
the mesoderm, but the circular muscles, Lillie claims, appear much 
later, and evidently, as described by Meyer (1901), are derived from 


the inner surface of the ectoderm. According to Iwanoff (1928) the 
mesoderm of the postlarval somites is formed in Polygordins, Aricia, 
Arenicola, and the OHgochaeta from the mesodermal teloblasts that 
generate the larval bands of mesoderm, but in the rest of the Poly- 
chaeta the postlarval mesoderm is proliferated from the ectoderm of 
the zone of growth. 

The pygidial region posterior to the zone of growth retains its 
primitive characters throughout the course of development, and is 
carried continuously backward as the number of somites increases. 
When the definitive number of somites has been formed, the growing 
zone loses its distinctive features and becomes indistinguishable as 
such. Structurally the secondary, or teloblastic, somites are modeled 
according to the general plan of the primary somites before them ; 
but, though they may differ in various structural details from the 
latter, they have one distinctive feature, which is that they alone 
contain the germ cells. Germinal centers ("gonads") may occur in 
all the teloblastic segments, but in most of the polychaetes they are 
limited to a definite part of the body (the epitoke), and in the 
oligochaetes they are usually restricted to a few segments. 

The ancestral annelids necessarily were reproductive as adults 
in all their evolutionary stages, but phylogenetic forms recapitulated 
in ontogeny are generally not reproductive. Hence, it is difficult to 
study the evolution of the reproductive system from ontogenetic 
development. The germ cells of the annelids usually are not recog- 
nizable as such in the larva, and little is known of their embryonic 
origin. It is claimed by Malaquin (1925), however, that in the 
serpulid Salmacina dysteri the sex elements first appear as differen- 
tiated cells in the gastrula, and that later (Malaquin, 1924) these 
cells become localized immediately before the zone of growth in the 
posterior segments of the young larva, where they lie ventral to 
the rectum, and are distinguishable from the surrounding cells by 
their large, clear, spherical nuclei containing numerous small chro- 
matic masses. In the oligochaete Pachydrilus, Penners (1930) claims 
the germ cells arise directly from the mesodermal teloblasts, and are 
the first cells formed by the latter. The germ cells, as shown also by 
Penners and Stablein (1930) in Tubificidae, appear prior to the for- 
mation of the definitive gonad somites, and migrate in the haemocoele 
to these somites, where they penetrate the mesoderm and finally take 
their definitive positions in the dissepiments. It seems highly probable, 
therefore, that the primitive annelids, at a phylogenetic stage before 
the teloblastic somites were formed, carried the germ cells in the 
undifferentiated posterior part of the body behind the last primary 


somite. From this point the germ cells must have been distributed to 
the secondary somites when the latter began to be developed during 
the course of evolution. Hence, primarily, the entire series of telo- 
blastic somites would appear to have been genital segments. 

Fig. 12. — Examples of the growth of larval Archiannelida and Polychaeta 
by proliferation in a subterminal zone of growth of teloblastic segments added 
to the primary larval body. 

A, larva of Eupomntus uncinatus with series of teloblastic, or "postlarval," 
segments (TScgs) interpolated between the three primary larval somites (I, 
II, III, see fig. 7 A, B) and the terminal pygidium (simplified from Iwanoff, 
1928). B, larva of Polygordius ncapoUlanus Fraipont during metamorphosis, 
with series of teloblastic segments added to the trochophoral body, which is 
itself unsegmented and contains no primary mesoderm (from Woltereck, 1905). 

C, half-grown young of Ncrilla antcnnata Schmidt (from Schlieper, 1925). 

D, larva of Lopadorhynchus brevis Grube with series of teloblastic segments 
(from Kleinenberg, 1886). E-G, growth stages of "nereidogen" larva of Platy- 
nercis dumcrilii Aud. & Milne-Edw. (from Hempelmann, 191 1, see also fig. 7 C, 
first stage larva). 

AnCir, anal cirrus ; Cirl , Cirll , tentacular cirri of first two somites, united 
in peristomium; I-XI, somites; Papd, parapodium ; Perst, peristomium ; Pip, 
palpus; Prst, prostomium; Pyg, pygidium; Tl, prostomial tentacle; TSegs, 
teloblastic segments. 

A condition similar at least to that which we should expect to find 
in the primitive annelids is seen in the archiannelid Dinophihis (fig. 
13 A). The body of Dinophihis consists of six or seven somites 
clearly defined externally between the prostomium and the pygidium, 
but there are no coelomic cavities in the diffuse mesoderm of the 


somites anterior to the last one, though each of these somites has a 
pair of simple protonephridia. In the terminal somite are formed 
the reproductive organs, which, in the female, consist of a delicate 
gonadial sac, either single or double, extending forward in the body, 
and opening posteriorly by a median pore, at least at the time of egg 
laying. The gonadial sac appears to represent the coelom of the last 
segment, though, as Iwanofif (1928) points out, it may be simply a 
space accommodating the germ cells in the undifferentiated tissue 
near the end of the body. Hence, the apparent last somite is either 
a single teloblastic genital somite, or a region corresponding with 
the zone of growth of the polymerous annelids. 

A concrete example of the secondary distribution of the germ cells 
in a polymerous annelid is given by Malaquin (1924 a) in his study 
of the development of SaUnacina dysferi. The germ cells, as we have 
seen, are first localized in the growing zone of the young larva. When 
the formation of the postlarval segments begins, Malaquin says, the 
germ cells multiply, and three, four, or five of the resulting gono- 
cytes become adherent to the outer wall of each new coelomic sac. 
Thus the germ cells, proliferated from a constant source, are dis- 
tributed to the newly forming somites, and are extracoelomic both in 
their origin and in their secondary segmental positions. After a 
period of inactivity the segmentally distributed gonocytes begin to 
multiply in the coelomic walls, and here form the small masses of 
germinal cells ensheathed in peritoneal folds that are known as the 

If, now, the ontogenetic facts of annelid growth are given a phylo- 
genetic significance, it becomes evident, as claimed by Iwanofif (1928), 
that the extension of the worm by the teloblastic generation of new 
somites, in which are apportioned groups of the multiplying germ 
cells, was primarily a means of amplifying the reproductive function. 
In the course of evolution it gave rise to a type of animal from which 
have been derived the modern Annelida, the Onychophora, and the 

The teloblastic genital segments are in many respects mechanical 
improvements over the primary segments; their muscular equipment 
is stronger, the parapodia better constructed for locomotion, the 
dissepiments usually more complete, and the nephridia more efficient 
for excretory purposes. Hence, the whole worm is clearly a stronger 
and a more active animal by reason of the addition of the well- 
organized reproductive somites. At the bottom of the water the 
creeping worm is better able to force its way under stones or into 
crevices, or to burrow into sand or mud; but at the breeding season 


its new powers of locomotion come into effective service, for now 
many modern species that habitually live at the bottom rise to the 
surface in swarms of energetically swimming individuals, both males 
and females, and here discharge the matured gametes. 

That the genital segments may be of no special physiological 
importance to the worm, except for carrying, maturing, and distrib- 

Tl Prst 

Fig. 13. — Examples of annelid types. 

A, B, Dinophilns, a very simple archiannelid, perhaps a primitive form, lack- 
ing teloblastic somites, tentacles, cirri, chaetae, parapodia, and coelomic sacs ; 
with five pairs of protonephridia, reproductive organs in posterior part of body 
(A, D. gyrociUatus Schmidt, adult female; B, adult male, from Shearer, 1912). 
C, Nerilla, an archiannelid with polychaete characters, perhaps a degenerate 
form, having a coelom, open metanephridia, and direct development (fig. 12 C) 
(A^. antcnnata Schmidt, from Goodrich, 1912). D, Lopadorhynchns, an errant 
polychaete (Phyllodocidae), having typical trochophoral development (fig. 8) 
with metamorphosis (fig. 12 D) producing long series of teloblastic somites 
(L. uncinatus Fauvel). 

An, anus; AnCir, anal cirrus; Ch, chaetae; Cir, cirrus; /, //, first two 
somites ; Mth, mouth ; Nph, nephridium ; Ov, ovary ; Papd, parapodium ; Phy, 
pharynx ; Pip, palpus ; Prst, prostomium ; SlGld, salivary gland ; Tl, tentacles. 

uting the reproductive elements, is shown by the various ways in 
which the annelids can dispose of these segments without otherwise 
impairing their functional integrity. There is the well-known case 
of the palolo worms, Eunice fucata and E. vlridis, for example, 
which live in crevices of rocks at the bottom of the water, and at the 
time for spawning detach the rear parts of their bodies, already 


loaded with the mature generative elements. The reproductive tail- 
ends (epitokes) then actively swim to the surface, where myriads 
of them congregate to liberate the gametes. In their accustomed 
haunts the anterior nonreproductive sections (atokes) regenerate the 
discarded epitokes in preparation for next year's consignment to the 
breeding grounds. Various other species of Polychaeta have similar 
habits. The Syllidae are famous for the many forms of schizo- 
genesis, strobilation, and budding that take place among them, but 
here the detached piece, either before or after separation, generates 
a new head and becomes a complete worm except perhaps for the 
lack of an alimentary canal and a few other unimportant structures. 
Again, in some of the Ctenodrilidae the worm breaks up by con- 
striction into several pieces of a few segments each, and the middle 
pieces regenerate both a head and a tail. 

The periodic fragmentation of the body for reproductive purposes, 
however, cannot lead to anything in the way of constructive evolution, 
and, with the annelids in general, the tendency has been to integrate 
the entire series of somites into a mechanical and physiological unit, 
in which the reproductive cells are assigned to definite segments. In 
the Arthropoda, though the body may still be composed of freely 
movable segments, the process of integration has been carried so far, 
and the various organs so interdependently distributed, that fission 
becomes impossible without fatal results. It would seem, therefore, 
that the teloblastic somites, first added apparently for reproductive 
efficiency, have been found so useful in other ways that they have 
come to constitute not only the largest part of the body in all the 
articulate animals, but its most important part, except for the primary 
sensory and nervous elements contained in the head. 

A structural differentiation between groups of somites, forming 
distinct body regions, or tagmata, has taken place in many of the 
polychaetes, particularly in the Sedentaria, and is a characteristic 
feature of all the Arthropoda. The zone of growth, therefore, which 
presumably at first gave rise to a series of identical somites, has 
acquired the remarkable faculty of differential activity, producing 
successively, at definite segment intervals, two or more series of 
somites having often a strongly contrasting structure, while minor 
differences may be distributed throughout the entire series of 


The annelid prostomium is the part of the trunk that is not invaded 
by the blastopore as the latter elongates forward on the ventral sur- 


face of the embryo (fig. 6 D, Prst) ; in the adult it is reduced to a 
small lobe overhanging the mouth (fig. 14, Prst). Appendages of 
the prostomium are best developed in the errant polychaetes, where 
typically they include a pair of anterior tentacles, or "antennae" {Tl), 
with frequently a median tentacle between them, and a pair of more 
posterior and ventral palpi {Pip). The prostomial appendages are 
clearly not equivalent to the parapodia of the postoral body somites, 
but they have the same development in the larva as the parapodial 
cirri (cf. fig. 16, A and B). Since the prostomium usually contains 
the brain and bears the apical sense organs, it constitutes the "head" 
•of the worm. In the absence of prostomial appendages and sense 

Fig. 14. — Head and anterior body segments of Nereis virens Sars. A, dorsal ; 
B, ventral. 

Cirl, Cirll, tentacular cirri of first and second somites united in peristomium ; 
E, eye ; ///, IV, third and fourth somites ; Papd, parapodium ; Perst, peristomium 
(somites / and //) ; Pip, palpus; Prst, prostomium; Tl, prostomial tentacle. 

organs, however, the brain may be secondarily withdrawn into the 
body, as in the earthworms (fig. 17 C, D, Br). 

The prostomium is not affected by the process of metamerism that 
cuts the postoral body region into a series of somites. Since the 
mesoderm bands of the larva do not proceed anterior to the mouth 
(fig. 6F). the larval prostomium does not contain mesoderm; but 
in later stages the mesoderm of the first somite may be extended 
into the prostomium (fig. 10 A) and give rise to a cephalic coelom 
and peritoneum (B, PCoel). Ordinarily the cephalic mesoderm is 
not segmented, but according to Binard and Jeener (1928) there is 
present in the prostomium of the spionid Scolelcpis fuliginosa a pair 
of distinct coelomic sacs, which are continuous with the cavities of 
the palpi, and have no connection with the coelomic sacs of the first 
postoral somite. This fact, the authors point out, gives a new argu- 


ment in favor of the homology of the polychaete palpi with the 
tentaclelike antennae of the Onychophora ; but evidently it does not 
prove their further contention (1929) that the palpi are appendages 
of a secondarily "cephalized" somite, since it must first be demon- 
strated that coelomic cavities may not pertain to the preoral mesoderm 


The body of the annelid is the segmented part of the trunk posterior 
to the acronal prostomium, including the region of the true somites, 
the zone of growth, and the pygidium ; but the term soma, in a 
restricted sense, would apply literally only to the region of the somites 
between the prostomium and the zone of growth or the pygidium. 
In the Polychaeta the first two somites are generally united with each 
other in a double segment known as the pcristomiuni (fig. 14, Pcrst), 
the tentaclelike cirri of which {Cirl, Cirll) take an anterior position 
closely associated with the prostomium. The "cephalization" of the 
anterior segments in the polychaetes, therefore, contrasts with that 
in the arthropods, since, with the latter, the first stage of cephalization 
is a union of the first somite with the prostomium. In the oligochaetes, 
however, the first somite and the prostomium may unite to form a 
composite head as in the arthropods. 

The fundamental demarcation of the annelid somites is the attach- 
ment of the longitudinal muscle fibers of the body wall and the 
muscles of the dissepiments on transverse circular grooves of the 
integument ; but the coelomic sacs when present are strictly intra- 
segmental, and most of the ectodermal and mesodermal organs are 
segmentally repeated. The locomotor mechanism of the annelids 
consists primarily of the somatic musculature and the regulating 
nerve ganglia, which give movement to the body wall, but it usually 
includes external adjuncts in the form of bristles or chaetae, and, 
in the Polychaeta, lobelike segmental appendages, the parapodia. The 
annelid body musculature should be the basis of the derived arthropod 
musculature, but there is reason to doubt that the polychaete para- 
podia are prototypes of the arthropod legs. 

The somatic musculature of the annelids includes the muscles of 
the body wall, the muscles of the chaetal sacs, and the muscles of 
the parapodia. The muscle fibers, with possibly rare exceptions, are 
of the nonstriated type. The musculature of the body wall is of a 
very simple pattern, so far as the arrangement of the fibers is con- 
cerned, but it may attain a strong development in the rapacious 
polychaetes and the burrowing oligochaetes. The longitudinal muscles 


can produce only contraction or lateral undulatory movements of the 
body; the circular muscles are constrictors producing peristaltic waves 
of body compression, and longitudinal extension of the body by the 
creation of internal pressure. The arthropod type of body mechanism, 
involving intersegmental movement of integumental plates, can be 
derived from the intrasegmental annelid mechanism only by the 
establishment of new intersegmental divisions. 

The polychaete somatic musculature is well developed in the 
Nereidae, of which Nereis virens may be taken as an example (fig. 
15). The outermost layers of body wall muscles consist of fine cir- 
cular fibers closely adherent to the integument (A, D, /). Internal 
to these there may be bands of oblique fibers (D, 2) crossing each 
other in opposite directions. The largest of the somatic muscles, 
however, are four thick bundles of longitudinal fibers (A, j, 4) lying 
internal to the others, two dorsal and two ventral, the fibers of which 
are attached on deeply inflected intersegmented folds of the integu- 
ment (D, isf). The longitudinal muscles of the terrestrial oligochaetes 
are continuous in a thick layer around the entire circumference of 
each somite, except where they are interrupted by the intrusion of 
the four chaetal sacs. Besides the muscles of the body wall there is 
in Nereis a double series of paired, obliquely transverse ventral 
muscles, one pair anterior and the other posterior in each segment 
(D, 5^ 6), which extend outward and upward from the median ventral 
fold of the body wall (A) to the lateral intersegmental folds between 
the parapodial bases. The intersegmental folds give attachment also 
to the transverse or radial muscles of the intercoelomic dissepiments. 
Most of the other muscles of the body pertain to the chaetal sacs and 
the parapbdia, and will be described in connection with the parapodia. 

A typical polychaete parapodium is a lateral outgrowth of the body 
wall (fig. 15 A, Papd), flattened antero-posteriorly, and usually 
divided into a dorsal lobe and a ventral lobe, which again may be 
subdivided into secondary lobules. Each major lobe bears distally a 
fan-shaped group of long chaetae (B, C/z), and on its base a slender 
cirrus {dCir, vCir). The chaetae arise from the inner walls of chaetal 
sacs (C, chS), from each of which a long rod, the acicula (Acic), 
extends inward to give attachment to protractor and retractor muscles. 

The larval rudiments of the parapodia represent the cirri and the 
chaetal sacs, and are differentiated as cellular bodies within the ecto- 
derm. The rudiments of the cirri, as described by Kleinenberg 
(1886) and by Meyer (1901) in the larva of Lopadorhynchus (fig. 
16 B, dcR, vcR), consist each of an outer layer of myoblasts (m) 
and an inner core of sensory nerve cells (n). The cirri in their origin, 



VOL. 97 

„ AlCnl DV 
3 \ 

6 10 i VNC ^ 6 

Fig. 15. — The polychaete locomotor mechanism: parapodia, and somatic and 
parapodial muscles of Nereis virens Sars. 

A, transverse section of a body segment, posterior view, somewhat diagram- 
matic. B, a parapodium and its muscles, posterior view. C, chaetal apparatus 
of a parapodium. D, muscles of right half of a body segment, inner view. 

Acic, acicula; AlCnl, alimentary canal; Ch, chaetae ; chS, chaetal sac; Cir, 
cirrus ; dClt, dorsal chaetae ; dCir, dorsal cirrus ; Dsp, intersegmental dissepi- 
ment ; DV, dorsal blood vessel; isf, intersegmental fold of integument; Papd, 
parapodium ; vCh, ventral chaetae ; vCir, ventral cirrus ; VNC, ventral nerve 

I, circular muscles of body wall ; 2, oblique muscles of body wall ; 3, dorsal 
longitudinal muscles ; 4, ventral longitudinal muscles ; 5, 6, anterior and pos- 
terior lateroventral, obliquely transverse muscles ; 7, 8, dorsal motors of 
parapodium ; g, 10, ventral motors of parapodium ; //, intrinsic muscle of para- 
podium between dorsal and ventral lobes ; 12, protractors of dorsal acicula and 
chaetal sac ; 13, 14, retractor and protractor of dorsal chaetae ; 15, retractor 
of dorsal chaetal sac and acicula ; 16, protractors of ventral acicula and chaetal 
sac ; 17, 18, retractor and protractor of ventral chaetae ; ig, retractor of ventral 
chaetal sac. 


therefore, resemble the tentacular rudiments of the prostomium (A), 
and later they grow out as tentaclelike processes. The bristle sacs 
are formed as ectodermal cell masses between the cirri (B, chS), the 
outer cells of which become myoblasts, while some of the inner cells 
enlarge and produce the chaetae ; a lumen then appears in the cell mass, 
and the latter becomes an open eversible sac from which the chaetae 
protrude. Finally the cirri and the chaetal pouches are carried out- 
ward on an outgrowth of the body wall that becomes the principal 
part of the appendage. The mature parapodia of Lopadorhynchus 
are not of typical form in that each consists of a single lobe (fig. i6 C) 
with both chaetal sacs at its extremity. 

In some of the polychaetes, particularly in the Sedentaria, there 
are two rows of podial organs on each side of the body (fig. i6F), 
those of one series, the notopodia (dPd), being situated dorso- 
laterally, those of the other, the neuropodia (vPd), ventrolaterally. 
Each organ includes a cirrus (Cir) and a chaetal sac (chS), and 
is innervated separately from the corresponding podial ganglion 
{PdGng). In the Oligochaeta the podial organs are represented only 
by the chaetae, which usually are arranged in two separated rows on 
each side of the body. It is possible, therefore, that the usual two- 
branched parapodium of the Polychaeta (fig. 15 B) has been formed 
by the union of a notopodium and a neuropodium. Furthermore, the 
double composition of each notopodium and neuropodium suggests 
that the primitive polychaetes had dorsolateral and ventrolateral rows 
of cirri, and between them on each side two series of chaetal sacs. 
On the peristomial segments of adult polychaetes generally only the 
cirri are present (fig. 14 A, Ctrl, Cirll), but on the rest of the body 
segments the chaetae-bearing lobes are usually the more important 
podial elements. 

The musculature of a parapodium is somewhat complex : it includes 
extrinsic muscles that move the appendage as a whole, and intrinsic 
muscles concerned principally with the movement of the chaetae. In 
Nereis virens there are four extrinsic muscles for each parapodium, 
two dorsal (fig. 15 D, 7, S), and two ventral (p, 10). The dorsal 
muscles arise anteriorly and posteriorly on the body wall, but cross 
each other obliquely to opposite margins of the parapodial base. The 
ventral muscles take their origins on the median infold of the ventral 
wall of the body segment (A), and extend laterally and dorsally, 
above the ventral longitudinal body muscles (4), to the anterior and 
posterior margins of the base of the parapodium. If the dorsal and 
ventral muscles inserted anteriorly act in opposition to those inserted 
posteriorly, the parapodium is moved anteriorly and posteriorly on 



VOL. 97 

the vertical axis of its base, and this is the usual motion of the 
appendage; but the latter can also be lifted and depressed, and the 
up-and-down motion evidently results from an antagonistic action 

Fig. 1 6. — Development, structure, and innervation of the polychaete appendages. 

A, section through larval rudiment of persistent dorsal tentacle (tlR) of 
trochophore of Lopadorhyuchus (from E. Meyer, 1901). B, transverse sec- 
tion of larva of Lopadorhyuchus through rudiments of a pair of chaetal sacs 
(chS) and associated cirri (from Meyer, 1901). C, parapodium of adult Lopa- 
dorhynchiis. D, structure of the armature of a parapodium of Myzostomutn 
asteriae Marinzeller, diagrammatic (from Stummer-Traunfels, 1903). E, a 
myzostomid, ventral view, showing parapodia. F, diagrammatic section of an 
amphinomid, Hcnnodice canmcnlata Pallas, showing widely separated notopodia 
(dPd) and neuropodia (vPd) and their innervation (from Storch, 1913). G, 
section of Nereis virens Sars, showing innervation of parapodia (from Hamaker, 

a, lateral nerve from podial ganglion; Acic, acicula; acmcls, acicular muscles; 
h, notopodial ganglion; Brn, branchia; c, neuropodial ganglion; Cli, chaeta or 
chaetae ; chS, chaetal sac ; Coel, coelom ; dCir, dorsal cirrus ; dcR, dorsal cirrus 
rudiment; dPd, notopodium ; Ecd, ectoderm; Gng, ventral ganglion; m, pri- 
mary muscle cell ; n, primary neural cell ; NCls, nerve cells ; A''^', nerve ; Papd, 
parapodium ; PdGng, podial ganglion ; tlR, rudiment of tentacle ; vCir, ventral 
cirrus ; vcR, ventral cirrus rudiment ; VNC, ventral nerve cord ; vPd, neuro- 
podium ; 3, dorsal muscles ; 4, ventral muscles. 

between the dorsal and ventral muscles. The intrinsic muscles of the 
parapodium include protractors and retractors of the chaetal sacs. The 
principal protractors (B, C, 12, 16) converge from the parapodial 
walls upon the inner ends of the acicular processes of the sacs (B, 


C, D), but the sacs themselves are eversible by muscles in their own 
walls (C, i^, i8). A retractor {15, ip) arising within the para- 
podium is inserted on the distal part of each chaetal sac, and a muscle 
(ij, I'j) from the acicula, attached on the base of the sac, opposes 
the muscles (j^, iS') that evert the sac itself. 

The parapodia are subject to numerous structural modifications in 
the different groups of Polychaeta, and among the specialized types 
the small leglike parapodia of the Myzostomidae (fig. 16 E) are of 
particular interest because of their resemblance to the legs of Ony- 
chophora. Each myzostomid appendage, as described by Stummer- 
Traunfels (1903), contains a deep apical pouch (D, chS), from the 
inner end of which a large hooked process (C/i) projects outward, 
while from its distal wall a thick rod (Acic) extends inward and 
gives attachment to protractor muscles (acmcls) and muscles inserted 
on the base of the hook. It is evident that the hook is a single, greatly 
enlarged chaeta, and the internal arm an acicula. The myzostomid 
"leg," therefore, is only a modified parapodium adapted for clinging 
to the crinoid hosts on which the Myzostomidae live, and has only a 
superficial likeness to the appendages of Onychophora (fig. 31). 


The central nervous system of the Polychaeta, as shown in the 
larval development, is produced from separate prostomial and somatic 
rudiments, which secondarily become united (fig. 9) ; in the Oligo- 
chaeta the two parts are said to be continuous from their inception. 
The definitive brain, whether formed from discrete ganglionic centers, 
as in Lopadorhynchus, or from a single generative zone of the pro- 
stomial ectoderm (fig. 10 A, B, Br), is always a compact organ, 
though it is generally bilobed (fig. 17 A) or differentiated into several 
consecutive parts (fig. 18 C). The ventral nerve cords in the more 
primitive condition found in most of the archiannelids and in various 
polychaete and oligochaete families are entirely separate, except for 
their connection by commissures (fig. 19 A, B), and in such cases 
the nerve tissue usually preserves a close contact with the ectoderm 
from which it is derived (C). More commonly, however, the paired 
ganglia of the cords are united in single median ganglia (fig. 17 C, D), 
giving the cords themselves a median position ; but even in such cases 
the ganglia of one or more pairs carried by the divergent anterior 
ends of the cords may remain widely separated. The first pair of 
united ganglia on the cords constitutes the so-called "suboesophageal 
ganglion," but it is evident that this ganglion does not belong always 



VOL. 97 


Br StGne- 

I II \ m IV/ ^ 


Fig. 17. — Nervous system of Annelida. 

A, anterior nervous system of Nereis virens Sars, diagrammatic, showing 
nerves of cerebral and suboesophageal ganglia (adapted from Hamaker, 1898). 
B, anterior nervous system of an amphinomid, Hcrmodice carunculata Pallas, 
showing podial nerves (PdNv) from brain connecting the podial ganglia, and 
divergence of ventral nerve cords (VNC) through several somites around 
stomodaeum (from Gustafson, 1930). C, anterior nervous system of Lumbricus 
terrestris Linn., lateral view, showing retraction of brain into third somite 
(simplified from Hess, 1925). D, same, dorsal view (from Hess, 1925). 

Br, brain ; Comll, ConiV , commissures of second and fifth somites ; Gngl- 
GngVI, central ganglion of first to sixth somites; I-VIII, first to eighth somites; 
Mih, mouth ; PdGng, podial ganglion of second peristomial cirri ; PdGngI, 
FdGngVIII , podial ganglia of first and eighth somites; PdNv, podial nerve; 
Perst, peristomium ; Pip, palpus; Prst, prostomium; SoeGng, suboesophageal 
ganglion; StCoii, stomodaeal nerve connective; StGng, stomodaeal ganglion; 
Stom, stomodaeum ; Tl, tentacle. 

Nerves of Nereis (fig. A) : a, nerve to stomodaeum; b, tentacle nerve; c, d, 
nerves to muscles and prostomial integument; e, nerve to proboscis; /, palpus 
nerve ; g, tegumentary nerve ; h, i, ocular nerves ; /, nerve to nuchal organ ; k, 
commissural ganglion ; /, m, nerves to proboscis ; n, connective between peri- 
stomial ganglia; o, nerve to second peristomial ganglion (podial ganglion, 
PdGng); p, nerve to proboscis; q, r, nerves to muscles and integument of 


to the same somite, and, furthermore, it sometimes contains the 
gangHa of more than one somite. In the polychaete family Amphinom- 
idae there is, in addition to the median nerve cords, a pair of lateral 
cords extending posteriorly from the brain (fig. 17 B, PdNv), which 
unite the series of podial ganglia (PdGng) lying at the bases of the 
parapodia (see Storch, 1912, 1913, Gustafson, 1930). The tetra- 
neurous structure is regarded by Storch as representing the more 
primitive condition of the annelid nervous system, though Gustafson 
contends that it is probably secondary. According to a theory pro- 
posed by Jeener (1928) the lateral line system represents a primitive 
series of neuromuscular sensory organs, from which there has been 
preserved and developed in the Sedentaria the sensorial elements, in 
the Errantia the ganglionic elements, and in the Oligochaeta the 
muscular elements. 

The annelid brain in its simplest form probably consists of a 
homogeneous mass of neurocytes aggregated upon a fibrous com- 
missure continuous on each side with the stomodaeal connectives, and 
through the latter with the ventral nerve cords (fig. 19 A, B, Br). 
With higher development, however, specialized groups of cells appear 
in the cortex, and specific tracts of fibers are individualized in the 
neuropile. A very simple brain structure occurs in the archiannelid 
Polygordius (fig. 18 A), in which, according to Hanstrom (1929), 
a pair of glomerulous association centers {PlpGlm) receive the roots 
of the palpal nerves and are connected by a palpal commissure 
(PlpCom). The peripheral sense cells of the palpi form ganglionlike 
masses (SCls) at the bases of the appendages. Two posterior lobes 
of the brain (NL) are connected with the nuchal organs, but eyes 
and anterior tentacles are absent in Polygordius. 

In the active polychaetes, in which cephalic tentacles, palpi, eyes, 
and nuchal organs are well developed, the brain takes on a more 
complex form and may acquire a high degree of differentiation in its 
internal organization. Particularly conspicuous are the paired cellular 
and fibrous masses known as corpora pedunculata. Each of these 
bodies consists of a cap of small chromatic cells lying in the upper 
anterior part of the cortex (fig. 18 B, E, Gb), and of a stalk, or 
pedunculus (Ped), composed of the neurites of the cap cells, which 
penetrate the central part of the brain. Within the stalks the terminals 
of the neurites (B, d) form synaptic associations between fibers from 
all other parts of the brain and from the ventral nerve cords (a, h, c). 
A simple development of the corpora pedunculata is shown by Han- 
strom (1927) to occur in the Hesionidae (fig. 18 D), in which the 
caps consist each of a single globulus of cells, and the stalks are 



VOL. 97 

connected by a fibrous commissure. In most of the other errant poly- 
chaetes the corpora pedunculata are more highly developed, and the 
cap cells become segregated into two or three distinct globuli (E, F). 
On the other hand, in the sedentary polychaetes with reduced cephalic 





f fa- 




f la 

st , 



Ped Con "5 



., J 


) mb-C 








PlpNv T-^ NCom OpCom 

Fig. i8. — Structure of the brain in Archiannelida and Polychaeta. 

A, outline of head and diagram of simple brain structure in Polygordius (from 
Hanstrom, 1929). B, horizontal section of brain of Sthenelais picta Verrill 
(Aphroditidae), diagrammatic (from Hanstrom, 1927). C, brain of Eunice 
ptmctata Risso (Eunicidae), showing high degree of external differentiation 
(from Heider, 1925). D, transverse section of brain of Podarke obscnra Ehlers 
(Hesionidae) through corpora pedunculata (from Hanstrom, 1927). E, hori- 
zontal section of brain of Nereis virens Sars (Nereidae) showing internal 
structure (from Hanstrom, 1927). F, diagram of a corpus pedunculatum of 
Nereis virens (from Hanstrom, 1927). 

a, sensory fibers of palpus nerve; la, 2a, 3a, nerves of lateral and median 
prostomial tentacles ; b, c, fibers of stomodaeal connectives ; d, axons of glob- 
uli cells of corpus pedunculatum ; E, eye ; /, central neurocytes in anterior 
part of brain ; jb, forebrain ; Gb, globulus of corpus pedunculatum ; hb, hindbrain ; 
mb, midbrain; NL, nuchal lobe; NCom, nuchal commissure; OpCoin, optic 
commissure ; OpNv, optic nerve ; Ped, pedunculus of corpus pedunculatum ; 
Pip, palpus ; PlpCom, palpal commissure ; PlpGlm, palpal glomeruli ; PlpNv, 
palpal nerve ; SCls, sense cells of palpus ; StCon, stomodaeal connective. 

sense organs, the corpora pedunculata are correspondingly reduced 
or are vestigial, and in the Oligochaeta they are absent. 

The development of the corpora pedunculata in the Polychaeta is 
clearly correlated with the development of the prostomial sense organs, 
but the particular relationships of the bodies are with the sensory 
nerves of the palpi. It is shown by Hanstrom (1927, 1928, 1929) in 


the Hesionidae (fig. i8D), the Aphroditidae (B), the Nereidae 
(E, F) and other errant famihes, that the roots of the palpal nerves 
are closely associated in glomerulous bodies with the stalks of the 
corpora pedunculata, which fact, Hanstrom points out, clearly sug- 
gests that the corpora pedunculata had their inception as association 
centers for the sensory nerves of the palpi. Much importance attaches 
to a study of the corpora pedunculata in connection with annelid and 
arthropod phylogeny, because bodies very similar in position, struc- 
ture, and variations are characteristic features also of the brain of 
Onychophora and Arthropoda. 

The relative positions of the principal internal structures of the 
polychaete brain, it should be noted for later comparison with the 
onychophoran and arthropod brain, are as follows : Anteriorly and 
dorsally are the corpora pedunculata (fig. i8 B, D, E) ; closely 
associated with the stalks of the latter are the palpal glomeruli (A, B, 
F, PlpGlin), and the glomeruli are connected by a palpal commissure; 
behind the corpora pedunculata is the optic commissure (D, E, 
OpCom) ; and in the posterior part of the brain are the nerve centers 
of the nuchal organs and a nuchal commissure (E, NConi). The 
stomodaeal connectives attach to the ventral surface of the brain. 

The number of nerves given off from the brain is highly variable 
according to the development of prostomial sense organs. In Poly- 
gordius (fig. i8 A) there is but a single pair of cerebral nerves, 
which innervate the tentaclelike palpi, while in such forms as Nereis 
(fig. 17 A) an elaborate innervation of the prostomial walls, the sense 
organs, and the stomodaeum proceeds from the brain. 

The principal stomodaeal nerves of the Polychaeta arise in some 
families from the first ganglia of the ventral nerve cords, or from the 
brain connectives near these ganglia, while in others they come from 
the upper parts of the connectives or from the back of the brain. 
It is contended by Hanstrom (1927, 1928), therefore, that in the 
second case the primitive first ganglia of the cords have been drawn 
forward and united with the brain, forming thus in certain polychaete 
families a posterior part of the definitive brain corresponding with 
the tritocerebral lobes of the arthropod brain, which always have 
connections at least with the stomodaeal (stomatogastric) system 
of nerves. 

The stomodaeal innervation of the Polychaeta is most elaborate in 
those forms that have a large and eversible stomodaeal proboscis, and 
in such cases the innervation of the organ may be derived from so 
many sources that the evidence adduced in favor of Hanstrom's theory 
is not convincing. In Nereis, for example, Hamaker (1898) describes 



VOL. 97 

Fig 19. — Examples of generalized structure in the annelid nervous system, 
and the structure of the annelid eye. 

A, the "rope-ladder" type of nervous system in an oligochaete, Acolosoma 
tenebrarnm (from Brace, 1901). B, same in an archiannelid, Dinophilus conk- 
lini Nelson (from Nelson, 1907). C, cross-section of ventral body wall of 
Aeolosoma tcnebrarmn, showing nerve cords not separated from epidermis (from 
Brace, 1901 ) . D, vertical section of an eye of primitive structure in a chaetopterid, 
Ranzania sagittaria Claparede (from Hesse, 1899). E, diagram of a typical 
annelid retinal cell (based on Pflugfelder, 1932). F, cross-section of optic rods 
of retina of Heteroyicrcis sp. (from Pflugfelder, 1932). G, vertical section of 
eye of a nereid, Lycastis sp. (from Pflugfelder, 1932). 

a, outer layer of epidermis over eye ; b, inner layer of epidermis forming 
ocular vesicle (see fig. 28 C, D, E) ; Br, brain; c, optic rod of sensory retinal 
cell ; CB, crystalline body ; Com, nerve commissure ; Cor, cornea ; Ct, cuticula ; 
d, cell body of sensory retinal cell ; e, striated border of retinal optic rod ; Epd, 
epidermis ; Lji, lens ; Mtli, mouth ; nf, nerve fiber ; nfbl, neurofibrillae ; Nu, 
nucleus; Nv, nerve trunk; Ret, retina (including sensory and supporting cells) ; 
StCovi, stomodaeal connective; VNC, ventral nerve cord. 


five pairs of stomodaeal nerves, two pairs of which proceed from the 
anterior part of the brain (fig. 17 A, a, c), a third pair (/) from 
small ganglia on the upper ends of the stomodaeal connectives, a 
fourth (m) from the ganglia of the first peristomial cirri, and a 
fifth (p) from the suboesophageal ganglion. In the earthworm, 
Lumbricus, the stomodaeal innervation arises from the connectives 
between the brain and the first ganglia of the cords (fig. 17 C, StGng). 
Other examples would only show further inconsistencies iri the origin 
of the nerves that supply the annelid stomodaeum. We can, therefore, 
most readily agree with Gustafson (1930), who concludes that no 
homology exists between the stomodaeal nervous system of the Anne- 
lida and that of the Arthropoda. Gustafson points out, furthermore, 
in reference to Hanstrom's theory of transposed ganglia, that there 
is no concrete evidence of the transfer of a pair of ventral ganglia to 
the brain in any of the annelids, whereas in the arthropods there is 
conclusive proof that the tritocerebral ganglia have been secondarily 
united with the brain. In the higher arthropods, moreover, the ganglia 
of the stomodaeal nervous system are derived directly from the 
ectodermal wall of the stomodaeum itself, and their definitive nerve 
connections with the central system appear to be secondary. 


Light-receptive organs in the form of eyes are widely present in 
the Polychaeta. The polychaete type of eye is fundamentally a 
vesicular ingrowth of the integument (fig. 28 C, D, E), the retinal 
cells being epithelial cells of the vesicle wall converted into primary 
sense cells by the extension of their inner ends as nerve fibers. In the 
simpler forms of eyes the cuticula may form a mere plug in the cavity 
of the retinal sac (fig. 19 D), but usually the ingrown part of the 
cuticula is enlarged and becomes a lenslike body, either connected with 
the surface by a cuticular strand, or entirely shut in by the union of 
the lips of the retinal sac (G, Ln). The outer ends of the retinal 
cells form optic rods, converging upon the inner surface of the lens 
(E, G, c), which contain the distal parts of the neural fibrillae (E, 
w/&/), but the apposed surfaces of adjacent rods do not form rhab- 
doms (F), as they do in the Arthropoda. 


The most primitive excretory organs of the annelids are the proto- 
nephridia of trochophore larvae. These are minute tubes, one or two 
pairs, extending from the body wall into the haemocoele, where they 


end blindly, but may be branched ; each tube or each branch terminates 
with a cell that sends a long vibratile flagellum into the lumen of the 
tube. The larval protonephridia are apparently of ectodermal origin, 
being said to be formed from primary nephroblasts derived from 
cells of the third quartet of blastomeres ; their structure is essentially 
that of the "flame cell" tubes of the excretory canals of the Platy- 
helminthes. Since the larval nephridia are present before the coelomic 
sacs are formed, they lie within the primary body cavity, which later 
becomes the haemocoele. 

A type of closed nephridium resembling the larval nephridia, and 
therefore often called a protonephridium, occurs in the five pregenital 
somites of the archiannelid Dinophilus (fig. 13 A), and in the post- 
larval somites of several families of Polychaeta. The closed nephridia 
of the adult worm, however, are more highly developed excretory 
structures than the larval organs, and usually have a more complex 
end apparatus of tube-cells (solenocytes), which contain long fila- 
ments resembling the flagella of the larval nephridia, but said to be 
nonmotile. The nephridial canal has a simple structure, and its lumen 
is ciliated. These nephridia project into the coelomic cavities, but, 
inasmuch as they are ensheathed in folds of the peritoneum, they lie 
morphologically in the haemocoele. Because of the similarity of their 
structure to that of the larval nephridia, the closed nephridia of the 
adult are supposed also to be of ectodermal origin, but their develop- 
ment apparently has not been studied. 

The usual adult excretory organ, occurring in most Archiannelida 
and Polychaeta, and in all Oligochaeta, is of the type called a mcta- 
nephridimn. The characteristic feature of a metanephridium is the 
presence of an inner opening, or nephrostome, by which the nephridial 
canal communicates with the coelom. Solenocytes in this case are 
absent. The nephrostome may be a simple ciliated aperture, as in 
the archiannelids, but more commonly it has the form of a wide, open, 
ciliated funnel. Unless coelomic dissepiments are absent, the nephro- 
stome always lies in the anterior lamella of the dissepiment before 
the somite in which the canal opens to the exterior. The canal thus 
appears to traverse the coelomic cavity behind the funnel, but morpho- 
logically it is extracoelomic, since it is ensheathed in a peritoneal fold 
produced from the posterior lamella of the dissepiment bearing its 
funnel. A closed nephridium is without doubt strictly an excretory 
organ, but an open nephridium may serve both for the removal of 
excretory products and for the discharge of the gametes from the 


The reproductive elements of the anneHds are hberated in various 
ways. In some of the Archiannelida and Polychaeta there is no 
anatomical provision for the discharge of the sex products from the 
coelomic sacs, and in such cases the gametes escape by a rupture of 
the body wall or by fission of the rear part of the body. With certain 
polychaetes having closed nephridia, a funnel-shaped structure is 
developed in the genital somites on the anterior surface of the septum, 
which at maturity opens into the canal of the nephridium, and serves 
as an outlet for the gametes ; but again in others the funnel, though 
present, is a mere "ciliated organ" of the coelomic peritoneum, not 
known to acquire an opening. Special genital ducts with an internal 
funnel and an external aperture are present in only a few Polychaeta, 
as in some of the Capitellidae, but they are characteristic features of 
the genital segments of Oligochaeta and Hirudinea. In most of the 
Polychaeta the nephridial' funnels serve for the discharge of the 

The relationship of the various types of annelid excretory organs 
and genital ducts to one another is difficult to understand. According 
to the well-known theory of Goodrich (1898-1900), nephridia and 
genital ducts, or coelomoducts, originally formed two separate series 
of segmental organs, and are still retained as such in Oligochaeta, 
Hirudinea, and certain Capitellidae. In the majority of the Poly- 
chaeta, however, Goodrich claimed, the genital funnel has lost its own 
duct and its funnel has united with the mouth of the nephridium, 
intermediate stages being suggested in some forms where there is a 
partial fusion between the funnel and the nephrostome. 

The study of the development of the open nephridia has given rise 
to much difference of opinion as to the origin of the nephridial 
rudiments. The earlier investigators, such as Hatschek and Vejdov- 
sky, regarded the nephridial funnels and canals as mesodermal 
structures, but Whitman (1886) claimed that the nephridia of the 
leech Clepsine are entirely of ectodermal origin. Wilson (1889), in 
his work on the development of Lumhricus, described the nephridial 
canals as being apparently ectodermal structures, developed from 
continuous rudiments formed from the second and third rows of 
ectodermal cells of the germ band, though he admitted they might 
be mesodermal ; the funnels, however, he said are derived separately 
from the anterior walls of the coelomic septa. Staff (1910) asserted 
also that the nephridial canals are ectodermal products in Criodrilus, 
but are formed from only the second row of cells in the germ band ; 
and Tannreuther (1915) claimed the nephridia of Bdellodrilus have 
the same origin, though he did not follow their complete development. 


On the contrary, nearly all other investigators have stoutly main- 
tained that both the funnels and the canals are mesodermal, though 
some regard these two parts as derived from separate rudiments. In 
this class may be mentioned E. Meyer (1887, Psygmobranchus), 
Bergh (1888, Criodrilus, 1890, Lumhricus, 1899, Rhynchelmis) , 
Burger (1891, Nephelis, 1894, Hirudo, Aulastomum) , Michel (1898, 
Allolohophora), Lillie (1906, Arenicola), Bychowsky (1921, Clep- 
sine), Penners (1924, Tuhifex), and A. Meyer (1929, Tnhifex). 
Only Bergh is insistent that the entire nephridium is mesodermal ; 
most of the others admit that a terminal part, perhaps including the 
reservoir, may be formed from the ectoderm. 

According to Lillie, the nephridia of the polychaete Arenicola 
cristata are gradually differentiated in the somatic mesoderm, starting 
from the posterior angles between the septa and the body wall, but 
the mesoderm in early stages of somite formation presents no cell 
boundaries. The lumen of each organ appears as a minute intra- 
cellular canal, which from its inception opens through the dissepi- 
ment into the preceding coelomic cavity. Later, as the nephridial 
cells divide, the lumen becomes intercellular, and finally it opens 
posteriorly through the ectoderm. Lillie says, however, that there is 
no invagination of the ectoderm, and no specific evidence that the 
reservoir is an ectodermal formation. 

Those writers who claim that the nephridia of the Oligochaeta 
and Hirudinea are of mesodermal origin agree essentially with Bergh 
that each organ is formed from a single cell of the anterior lamella 
of an intersegmental septum. According to A. Meyer (1929), for 
example, the nephridioblasts of Tuhifex are early differentiated from 
the other cells of the septa by their large size (fig. 20 A, Nphl). By 
successive divisions of the nephridioblast a column of cells is formed 
that pushes backward within a sheath of ordinary epithelial cells 
derived from the posterior lamella of the septum (B-E). The young 
nephridium extends in a space between the somatopleure and the 
longitudinal muscles, and is thus extracoelomic. The lumen appears 
first as an intracellular canal, which later becomes intercellular by a 
radial division of the cells; it is ciliated from an early stage. Pos- 
teriorly the canal ends against an epidermal cell (G), through which 
it eventually opens to the exterior, and from which is later gener- 
ated the reservoir. The coelomic funnel is formed by the original 
teloblast, the nucleus of which divides into four nuclei, one taking 
a position in the dorsal lip of the funnel, the other three in the ventral 
lip (H, I). According to Bergh (1899) only the lower lip of the 
funnel in Rhynchelmis is derived from the nephridioblast, the upper 


lip being formed from a neighboring group of septal cells. In 
Clepsine, Bychowsky (1921) says, the first division of the nephridio- 
blast is in the plane of the dissepiment, and gives rise to an anterior 
cell that forms the funnel and the adjacent part of the canal, and a 
posterior cell that generates the rest of the canal. The latter opens 
finally to the exterior through an ectodermal invagination. Bergh 

Fig. 20. — Successive early stages in the development of the posterior nephridia 
of the oligochaete Tubifcx rhniloniiii Lam. (From A. Meyer, 1929.) 

A, a primary nephridioblast developed from a cell of the anterior lamella of 
a dissepiment. B, proliferation of nephridial cells by transverse division of 
the nephridioblast. C-F, successive extensions of the nephridial canal within 
a peritoneal sheath derived from the posterior lamella of the dissepiment ; the 
canal acquires first an intracellular lumen. G, the canal still more elongate and 
looped upon itself, attached posteriorly to an epidermal cell, through which the 
lumen penetrates to the e.xterior, and which later forms the nephridial bladder. 
H, I, two stages in the final development of the nephrostome in the primary 
nephridioblast by radial division of the nucleus. 

Dsp, dissepiment ; Luvi, lumen ; NCnl, nephridial canal ; Npbl, nephridioblast ; 
Mpr, nephropore ; A''^^, nephrostome; PSh, peritoneal sheath. 

claims that there is no ectodermal element in the nephridium of 
Criodrilns, Rhynchelmis, or Luiiibricus. 

It thus appears to be now well established that the metanephridia 
of the annelids in general are structures of the nature of coelomo- 
ducts, formed principally as outgrowths of the posterior walls of the 
coelomic sacs, but perhaps including a terminal part of variable extent 
derived from the ectoderm. They are extracoelomic, inasmuch as 
each nephridial canal is invested in a fold of the coelomic peritoneum. 
The nephridial organs have always been important subjects in dis- 


cussions of relationships between the Annehda, the Onychophora, 
and the Arthropoda. The onychophoran nephridia, however, are 
developed as simple diverticula of the ventral walls of the coelomic 
sacs, which connect with short ectodermal ingrowths of the same 
segments situated mesad of the leg bases, and the nephridial organs 
of the arthropods most probably have had the same genesis as the 
onychophoran organs. Hence, it is possible that the coelomic exits 
have had an independent origin in the higher Annelida on the one 
hand, and in the common ancestors of the Onychophora and Arthrop- 
oda on the other. 


Somewhere from a generalized annelid stock there must have 
branched off in remote pre-Cambrian time the ancestors of the group 
of animals that includes the modern Onychophora (fig. 21 A), the 
Cambrian Aysheaia (B), and the pre-Cambrian Xenusion (C). The 
primitive onychophorons undoubtedly were segmented, wormlike 
creatures, in which coelomic sacs and the basic features of the annelid 
muscular and nervous systems had long been established, and in 
which the body had been lengthened by the addition of a series of 
reproductive somites generated from the posterior zone of growth. 
A distinctive feature of the Protonychophora, however, was the pos- 
session of movable locomotor appendages having the form of small 
lobelike outgrowths of the body wall along the lateroventral lines of 
the segments. The ancestors of the lobopod Onychophora, and the 
ancestors of the chaetopod Annelida, therefore, probably constituted 
two divergent branches from a generalized annelid stock. The primi- 
tive chaetopods were creeping worms that progressed by the usual 
vermiform movements of the body, produced by the body musculature 
with the aid of integumental chaetae. The primitive onychophorons 
became distinguished as walking worms, a character well expressed 
in the name Peripatus (Guilding, 1826) given to the first-described 
modern form. The walking habit led to the adaptation of the modern 
Onychophora to life on land, but the older forms, such as Aysheaia 
and Xenusion, may have been inhabitants of the ocean. 

A typical onychophoron is a slender wormlike creature with a pair 
of tentacular antennae at the anterior end of the trunk, and a double 
row of short, conical, lateroventral legs along the length of the body 
(fig. 21 A). The trunk is cylindrical or somewhat depressed, blunt 
anteriorly, and tapering posteriorly. The rough integument is closely 
ringed, but there is no external sign of segmentation except for the 
series of appendages. The animal has no distinct head ; the anterior 


part of the trunk, however, forms a cephaHc lobe (D, E) bearing 
the antennae, a pair of small dorsal eyes (E, E) just behind the 
antennal bases, and on the ventral surface the mouth (D). The 
mouth, which is a triangular opening into the stomodaeum, is sunken 
into a preoral cavity surrounded by an integumental circumoral fold 

Fig. 21. — Onychophora, ancient and modern. 

A, Peripatoides novac-zcalandiac Hutton. B, Ayshcaia pcdunculata Walcott 
(191 1), of Middle Cambrian, British Columbia, "conjectural restoration" (from 
Hutchinson, 1930). C, Xemision aucrszualdi Pompeckj (1927), of Algonkian, 
proterozoic pre-Cambrian (from Heymons, 1928, broken lines hypothetically 
completing larking parts). D, Peripatoides noz>ac-zeaIandiac, anterior part of 
trunk, ventral view. E, same, head and anterior part of body, lateral view. 
F, same, right jaw, dorsal view, with muscles. 

a-d, jaw muscles; Ant, antenna; Ap, apodeme of jaw muscles; cof, circumoral 
fold; E, eye; /, jaw; iL, first leg; Lm, labrum; OP, oral papilla. 

(cof). Within the preoral cavity is a small anterior labral lobe 
(Lin), and a pair of flat, two-hooked jaws (/) that converge pos- 
teriorly at the sides of the mouth. On the sides of the head, laterad 
of the mouth, is a pair of oral papillae (E, D, Op) that give vent 
to a pair of large, many-branched slime glands widely spread in the 


body cavity (fig. 32 A, SlmGld). The following appendages are 
the legs, varying in number with dififerent species from a minimum 
of 13 pairs to an average of perhaps 25 or 30 pairs, though some 
species have 40 or more. Behind the last legs the body tapers to a 
terminal cone on which is situated the anus. The genital aperture 
in each sex is a median ventral opening lying either between the legs 
of the last pair, or behind the last pair present in species having one 
or two of the posterior pairs of legs absent. 


Were it not for the evidence of annelid relationships shown in the 
adult structure of the Onychophora, we should have little reason for 
believing that the onychophorons are descended from Annelida, for 
in their ontogeny we encounter none of the familiar early phases of 
development so characteristic of the annelids. Most of the Ony- 
chophora are viviparous, the embryos developing to maturity in 
uterine chambers of the oviducts (fig. 32 A, Utrs) ; only a few 
species are known to be oviparous. Eggs supplied with a large quantity 
of deutoplasm complete their development from their own store of 
yolk, but the embryos of viviparous species with small eggs receive 
nourishment from the uterine walls, and in some cases a placentalike 
growth of the blastoderm forms a large vesicular trophoblast applied 
to the walls of the uterus. 

The early stages of onychophoran development are so variable in 
diflferent species that it is impossible to give any general account of 
the processes of cleavage and germ-layer formation. Cleavage in 
some species with small eggs is holoblastic, producing first a solid 
morula and then a hollow blastula (see Sclater, 1888). Contrary to 
what we might expect, however, gastrulation in such cases does not 
take place by invagination. In Peripatus imthurni, as described by 
Sclater, an internal proliferation of cells proceeds from a definite 
point on the blastula, and the cells thus produced become dififeren- 
tiated into endoderm and mesoderm. A similar method of endoderm- 
mesoderm formation is described by Kennel (1888) in Peripatus 
edwardsi, there being here a blastoporic depression of the blastoderm 
from which an internal proliferation gives rise to endoderm and to 
ventrolateral bands of mesoderm. With eggs having much yolk, 
meroblastic cleavage is the rule. The &gg nucleus divides within the 
yolk, and the cleavage nuclei enclosed in small masses of cytoplasm 
migrate to the surface and form a blastoderm. In Peripatoidcs novae- 
zealandiae, however, according to Sheldon (1888), the blastoderm 


lies beneath a superficial layer of yolk, the early embryo in this case 
being a sac not only containing yolk, but also surrounded by it. The 
outer yolk is later absorbed. In this species the manner of germ- 
layer formation has not been definitely determined, but the endoderm 
cells appear within the yolk, and the mesoderm takes the form of 
two widely separated bands along the sides of the embryo, in which 
the coelomic sacs are formed. 

%r H 

Fig. 22. — ^Early developmental stages of Onychophora. 

A-D, successive embryonic stages of Peripatopsis capcusis Grube, showing 
elongation and closure of the blastopore except at oral and anal extremities, 
and forward growth and segmentation of mesoderm bands (from Balfour, 
1883). E, young embryo of Eoperipattis iveldoni Evans, ventral view, mouth 
covered by external yolk (from Evans, 1902). F, young embryo of Peri- 
patopsis mosclcyi Wood-Mason with open blastopore (from Bouvier, 1905). 
G, cross-section of embyro of Peripatopsis capensis through open blastopore 
(from Balfour, 1883). H, cross-section of embryo of Eoperipafiis zwldoni, 
blastopore covered with yolk (from Evans, 1902). 

An, anus ; Bpr, blastopore ; Coel, coelomic cavity ; 2Cocl, sCocl, second and 
third coelomic cavities ; Ecd, ectoderm ; E}id, endoderm ; Gc, gastrocoele ; He, 
haemocoele; Msd, mesoderm; MsT, mesodermal teloblast ("primitive streak") ; 
Mth, mouth ; Y, internal yolk ; y, external yolk. 

In Peripatopsis capensis, Sedgwick (1885) says, cleavage is complete 
but unequal, the blastomeres being differentiated into four small, dark 
ectodermal cells at the animal pole of the egg, and four large, clear 
endodermal cells at the vegetative pole. Subsequent divisions proceed 
in each group separately. The endoderm cells soon draw together 
into the center of the egg, and are here overgrown by the ectoderm 
until completely enclosed by the latter, except at one point where 


the endoderm remains exposed on the surface. A cavity now 
appears in the endodermal mass, and opens externally where the 
endoderm is not covered by the ectoderm. The opening is the blasto- 
pore (fig. 22 A, Bpr). With the growth of the embryo, the blastopore 
lengthens to an elongate slit on the ventral surface (B). The first 
observations on the development of Peripatopsis capensis were made 
by Balfour ( 1883), who believed that the mesoderm arises in the form 
of paired coelomic pouches along the edges of the elongate blastopore 
where the ectoderm and endoderm are confluent. From the subse- 
quent work of Sedgwick, however, it appears that the mesoderm in 
P. capensis is generated from an opaque area of the blastoderm 
situated behind the posterior end of the blastopore (A, B, C, MsT). 
From this area, or "primitive streak," there takes place an internal 
proliferation of cells, which, migrating forward in each side of the 
embryo, produce two ventrolateral mesoderm bands along the margins 
of the blastopore (B). The bands then break up into sections that 
mark the primitive somites of the embryo, and later are excavated by 
the coelomic cavities (G, Coel). The elongate blastopore finally closes 
by the fusion of its lips, except at the two ends, which become the 
primary mouth and the primary anus (D, Mth, An). 

The development of the endoderm of Eoperipatus zveldoni, as 
described by Evans (1902), is again different from that of Peri- 
patopsis capensis. "The endodermal elements," Evans says, "are 
derived from the lips of the blastopore and travel inward along the 
outer layers of the yolk, which is at first devoid of nuclei." Here, 
evidently, is a process suggesting invagination. On the surface of 
the yolk the endoderm cells form a complete investing layer, but later 
some of them invade the yolk, probably bringing about its partial 
digestion, and then again most of these cells return to the surface, 
where they reconstruct a permanent endodermal sac containing the 
yolk (fig. 22 H, End). A few endodermal cells, however, remain 
within the yolk. The mesoderm of Eoperipatus zveldoni, according 
to Evans, is formed in the same way as described by Sedgwick for 
Peripatopsis capensis, that is, from a proliferating area of the blasto- 
derm situated immediately behind the blastopore (E, MsT). 

Considering the various processes by which the organization of 
the onychophoron is accomplished in the embryo, it would appear 
that the manner of development has little significance. In extreme 
cases the assembling of the germ layers seems to be almost haphazard. 
Sheldon (1888) observes of Peripatoides novae-zealandiae that the 
embryo might be said to be formed "by a process of crystallizing out 
in situ from a mass of yolk, which is a protoplasmic reticulum con- 


taining nuclei." Among the early developmental phases of the Ony- 
chophora, however, we cannot fail to note two important likenesses 
to annelid development. The first is the elongation of the blastopore 
on the ventral surface of the embryo as it occurs in Peripatopsis 
capensis (fig. 22 A, B), followed by the closure of its median part 
(C), finally leaving only the persistent oral and anal apertures at the 
two extremities (D). We have here evidently a condition even more 
generalized than in the annelids, in which the anus is usually a secon- 
dary perforation. The second suggestion of annelid development, 
shown in several onychophoran species, is the forward growth of the 
mesoderm as bands of cells generated from a proliferating area of 
the blastoderm situated behind the blastopore (fig. 22 A, B, C, 
E, MsT). The mesoderm is, therefore, a teloblastic product, though 
it is not possible to identify in the generative area a primary pair of 
teloblastomeres. It would appear, however, that the onychophoran 
mesoderm may not be entirely of teloblastic origin, for Sedgwick 
(1887) finds that the forwardly growing bands in Peripatopsis 
capensis are augmented by cells proliferated from the lips of the 
blastopore along the lines where ectoderm and endoderm meet. The 
later development of the mesoderm is unquestionably a strictly 
homologous process in both the Annelida and the Onychophora, for 
in the latter, as in the annelids, the primarily solid mesoderm bands 
are first segmented corresponding with the body somites (fig. 22 B), 
and then excavated by coelomic cavities (C, G). 

Beyond the early stages of cleavage and germ-layer formation the 
course of onychophoran ontogeny is well standardized and gives a 
good basis for comparison of the Onychophora with the Annelida on 
the one hand, and with the Arthropoda on the other. It will be found, 
however, that many of the irregular earlier processes of onychophoran 
development are duplicated among the Arthropoda. 


The onychophoran nervous system includes a brain situated in the 
head above and before the decurved anterior end of the stomodaeum 
(fig. 32 A, Br), and two long, widely separated nerve cords (NC) 
extending from the brain to the posterior end of the body, where they 
appear to be continuous in an arc above the rectum. The cords are 
connected by numerous ventral commissures (Com), and they give 
ofif in each segment a series of dorsal nerves against the body wall 
(fig. 24 B) and ventral nerves that go downward to the legs and 
other ventral parts. Opposite the legs the nerve cords are slightly 



VOL. 97 

thickened, but they have no differentiated gangHa, since the neuro- 
cytes are scattered along their lengths. The brain, on the other hand, 
is a well-developed, bilobed cerebral body (fig. 25 A, B) extending 
horizontally forward from the anterior ends of the nerve cords (C). 
It bears anteriorly the large antennal nerves (AntNv), laterally a 
pair of small optic lobes supporting the eyes (E), and ventrally a 
pair of small pear-shaped bodies (B. C. iVO). Numerous other 









2V0 )Q^i -'Xn )^ 

PrC " 

Fig. 23.-— Development of the onychophoran head and anterior body region 
as shown in three embryonic stages of Pcripafus cdivardsi Blanchard, ventral 
view. (From Kennel, 1888.) 

A, young embryo with large prostomial cephalic lobes, postoral jaw appen- 
dages (/) and oral papillae {OP) resembling legs. B, older embryo with 
prostomial antennal rudiments, jaws approaching mouth and surrounded by 
circumoral fold (fo/), ventral organs (FO) becoming differentiated. C, still 
older embryo ; jaws with definitive form, retracted into preoral cavity, ventral 
organs more distinct ; head region composed of procephalic lobes, jaw somite, 
and somite of oral papillae. 

ajVO, anterior ventral organ of papillar somite; cof. circumoral fold; /, jaw; 
L, leg ; Lin, labrum ; Mth, mouth ; OP, oral papilla ; PrC, preoral mouth cavity ; 
Prsf, prostomium; P3VO, posterior ventral organ of papillar somite; SIO, orifice 
of salivary gland; iVO, ventral organ of preoral cephalic lobe; 2V O, ventral 
organ of jaw somite; 3VO, ventral organ of somite of oral papillae (subdi- 
vided into anterior and posterior parts) ; 4VO, 5VO, ventral organs of first and 
second leg somites. 

small nerves are given off from the brain (fig. 24 A), among which 
are anterior ventral nerves that go to the mouth and the circumoral 
fold, a dorsal median nerve (/) that turns downward and posteriorly 
on the dorsal surface of the stomodaeum, a pair of posterior stomo- 
daeal nerves (/), and the nerves of the jaws (;'), which arise from 
the nerve cords just behind the brain. 

The entire central nervous system of the Onychophora is developed 
in the embryo from a series of paired ventral thickenings of the 
ectoderm known as the "ventral organs" (fig. 23 B, C, VO), which 


correspond with the embryonic somites, except that the first pair 
(iVO) Hes on the preoral head region. Whether these thickenings 
represent primitive organs or are merely embryonic structures is 
open to question, but they suggest the paired tubercles on what may 
be the ventral surface of Xenusion (fig. 21 C). From the inner 
surfaces of the ventral organs of the body are differentiated the 

Fig. 24. — Nervous system of the head and of a body segment of Peripatus 
tholloni Bouvier. (From Fedorow, 1926, 1929.) 

A, diagram of brain and anterior parts of nerve cords, with bases of nerves, 
dorsal view. B, nerve cord and peripheral nerves of left side of a body segment, 
lateral view. 

a, sensory antennal nerve; aip, anterior interpedal nerve; A)it, antenna; b, 
motor nerves of antenna; Br, brain; c, optic nerve; Co)ii, nerve commissure; 
d, lateral dorsal nerve ; E, eye ; e, nerve to circumoral fold ; f, median dorsal 
nerve ; g, nerve to dorsal muscles of head ; h, commissural nerve from f to g ; 
i, stomodaeal nerve; ICom, first ventral commissure; /, k, nerves of jaw; L, 
leg ; /, nerves of oral papilla ; m, n, 0, nerves of first leg segment ; ip, 2p, first 
and second pedal nerves ; ipp, 2pp, first and second postpedal nerves ; pip, pos- 
terior interpedal nerve ; iprp, 2prp, first and second prepedal nerves ; v, ventral 

ventral nerve cords ; the outer parts are then gradually reduced in 
size until finally they disappear as distinct areas of the epidermis. 
When the nerve strands become free cords within the body they do 
not approach each other or unite as do the nerve cords of most anne- 
lids or arthropods ; on the contrary they move farther apart until they 
take positions along the sides of the body on a level with the leg 
bases (fig. 29, NC^. The definitive cords, moreover, lie laterad of 



VOL. 97 

series of dorsoventral lateral muscles (dznn) attached dorsally and 
ventrally on the body wall. A condition thus arises in the Ony- 
chophora that has no counterpart in the annelids or arthropods, for 
in the latter the nerve cords, even when laterally situated, have no 
barrier to a median approximation or union. 

The major part of the brain, from which arise the antennal and 
optic nerves, is shown by Sedgwick (1888), Kennel (1888), and 
Evans (1902) to be generated from the paired ventral organs of 



Fig. 25. — Brain of Peripatoides novae-aealandiae Hutton. 

A, dorsal surface of brain and anterior parts of nerve cords, showing posterior 
antennal commissure and dorsal position of antennal tracts. B, ventral surface 
of brain, with remnants of ventral organs. C, lateral view of brain and stomo- 
daeal connectives. 

AntCom, antennal commissure; AntNv, antennal nerve; AntT, antennal tract; 
b, motor nerves of antenna ; E, eye ; /, median dorsal nerve ; i, stomodaeal 
nerves; I Com, first ventral commissure; /, k, nerves of jaw; NC, nerve cord; 
OpL, optic lobe; StCcm, stomodaeal connective; iVO, remnant of first ventral 

the head (figs. 23 B, 27 B, iVO). Evans says that the brain includes 
also a pair of anterior "archicerebral lobes" belonging to the apical 
part of the head, but in his account of the embryonic development 
of Eoperipatus weldani he makes no mention of observing a separate 
origin of such lobes, and attributes the entire brain, except a pos- 
terior part, to the neural elements derived from the cephalic ventral 
organs. The ventral organs of the head, unlike those of the body, 
are finally invaginated as vesicles connected with the nerve tissue ; 
eventually they are reduced, but persist as the small bodies attached 
to the ventral side of the brain (fig. 25 B, C, iVO). 


The small posterior lobes of the brain from which arise the posterior 
stomodaeal nerves (fig. 25 B, i), together with the adjoining parts 
of the nerve cords that give off the nerves of the jaws (A, B, C, /), 
are said by Evans to be secondarily added to the antenno-ocular lobes 
from the ventral organs of the postoral jaw somite (fig. 23 B, 2V O), 
and Kennel clearly shows in a head section (fig. 27 A) the inclusion 
in the brain of a mass of neural cells given off from these generative 
centers (2VO). The definitive onychophoran brain, therefore, as 

Si: Con 

Fig. 26. — Internal structure of the brain of Peripatopsis capensis Grube. 
(From Holmgren, 1916.) 

AntCom, antennal commissure; AntGlm, antennal glomeruli; AntNv, sensory 
antenna! nerve ; AntT, antennal tract ; b, motor nerves of antenna ; Cc, corpus 
centrale ; e, nerve to circumoral fold ; /, median dorsal nerve ; Gb, globuli of 
corpus pedunculatum ; ;, stomodaeal nerves; ;', nerve of jaw; OpNv, optic nerve; 
Fed, peduncle of corpus pedunculatum ; StCon, stomodaeal connective. 

shown by the records of its development, and as claimed by Holmgren 
(1916) and by Hanstrom (1928, 1935) from histological evidence, 
would appear to be a syncerebrum composed of a prostomial fore- 
brain including the ocular and antennal centers, and of a postoral 
hindbrain containing the centers of the posterior stomodaeal nerves 
and the nerves of the jaw appendages. 

A quite different concept of the composition of the onychophoran 
brain is deduced by Fedorow (1929) from a study of Peripatus 
tholloni, in which he attempts to correlate the cerebral nerves with 
the nerves of a series of body segments (fig. 24 A, B). Fedorow 


concludes that the anterior part of the brain, lying before the antennal 
commissure and bearing the optic lobes, represents the prostomial 
archicerebrum of the annelids, and that the rest of the brain is of 
postoral origin, being formed of the united anterior ends of the 
nerve cords extended secondarily in front of the stomodaeum. This 
alleged postoral part of the definitive brain, Fedorovv believes, 
includes the ganglionic centers of the antennal somite, and the ganglia 
of a reduced premandibular somite that has lost its appendages. The 
jaw centers, he contends, are contained in the parts of the nerve cords 
immediately behind the brain, from which arise the nerves of the jaw 
muscles (/), and which are connected by the first postoral commissure 
(iCom). Fedorow's elaborate analysis of the brain structure and 
nerves would be more convincing if it took into account the embryonic 
development of the brain; his results are entirely unsupported by 
ontogenetic evidence, and are mostly at variance with observations on 
the brain development reported by other investigators. 

The internal structure of the onychophoran brain (fig. 26) shows 
fundamental characters of the polychaete brain, and contains certain 
arthropod features, but it presents also special modifications that are 
not found in either the annelids or the arthropods. Corpora peduncu- 
lata are well developed, each consisting of a cap of three globuli (Gb) 
of small chromatic cells lying in the anterior part of the brain, and 
of a large pedunculus (Fed) composed of three confluent groups of 
fibers springing from the globuli cells. The sensory antennal nerves 
(AntNv) coming into the anterior angles of the brain traverse the 
upper part of the cerebrum in distinct antennal tracts (AntT), which 
are united posteriorly in a broad antennal commissure {AntCoin). 
The association centers of the antennal nerve fibers, called by Holm- 
gren (1916) and Hanstrom (1928, 1935) the antennal glomeruli 
(AntGlni), lie laterad of the anterior ends of the corpora pedunculata, 
and are said by Hanstrom to be closely connected with neurites of 
the globuli cells. In this feature, Hanstrom points out, the Ony- 
chophora have a distinctly polychaete character in the brain structure, 
since the antennal glomeruli of the onychophoran brain evidently 
correspond with the palpal glomeruli of the polychaete brain (fig. 
18 B, F, PlpGlm). On the other hand, the onychophoran brain shows 
arthropodan characters in the presence of a well-developed central 
body {Co) and an antennal commissure (AntConi) . But again, the 
small optic lobes of the eyes (fig. 25 A, OpL) contain each only a 
single ganglionic center, while all arthropods have at least two. The 
optic ganglia are connected with the corpora pedunculata and with 
the central body. 


The onychophoran brain thus appears to contain, as Hanstrom 
(1935) has shown, a mixture of polychaetous and arthropodan char- 
acters. Its origin must be found in the annehd brain ; but certain 
pecuhar features of the onychophoran brain would seem to prechide 


Fig. 27. — Developmental stages of various head structures of Onychophora. 
(A, C from Kemiel, 1888; B, D, E from Evans, 1902.) 

A, cross-section of head of embryo of Pcripatus edzvardsi Blanchard through 
jaws (/), showing groups of brain cells proliferated from ventral organs (2V O) 
of jaw somite. B, cross-section of embryonic head of Eoperipatus iveldoni 
Evans, showing coelomic sacs of antenna embracing the stomodaeum, and gen- 
eration of brain (Br) from cephalic ventral organs (iVO). C, section through 
anterior part of head of embryo of Peripatus cdzuardsij showing antennal coelomic 
sac extending into antenna. D, section of embryonic head of Eoperipatus iveldoni 
with canal {d) from antennal coelom opening mesad of circumoral fold (cof). 
E, same, more anterior section, showing antennal coelomic sacs, and a pre- 
antennal sac on left side. 

Ant, antenna; AntCoel, coelomic sac of antenna; AntNv, antennal nerve; 
Br, brain ; cof, circumoral fold ; Cpr, coelomopore of antennal coelom ; E, pit 
of developi'ig eye; Ecd, ectoderm; End, endoderm ; He, haemocoele ; /, jaw; 
mcl, muscles of stomodaeum, Mcnt, mesenteron (folded forward on stomo- 
daeum) ; PrntCoel, preantennal coelom; Stom, stomodaeum; iVO, preoral 
cephalic ventral organ; 2V O, ventral organ of postoral jaw somite; Y, yolk. 

the possibihty of its having given rise to the arthropod brain. The 
superficial position of the antennal nerve tracts (fig. 25 A, C, AntT), 
which traverse the forebrain dorsal to the optic lobes (A, OpL), con- 
stitutes a condition quite at variance with that in any arthropod, for 
in all the Arthropoda the antennal nerves issue from antennal lobes 


that lie ventral to the optic lobes, showing that the antennae have 
migrated forward beneath the eyes, and not above them as in the 
Onychophora. Moreover, in the arthropod brain the antennal glomer- 
uli are not immediately connected with the corpora pedunculata. The 
onychophoran brain in its modern form, therefore, could not have 
given rise to a brain of arthropod structure, and we can assume only 
that the two types of cerebral structure have taken their origins 
separately from some common progenitor not far removed from a 
generalized annelid. Even the inclusion of the nerve centers of the 
first postoral somite in the onychophoran brain cannot be taken as 
evidence that the Onychophora are ancestral to the Arthropoda, for 
in some of the lower members of the second group the first postoral 
(tritocerebral) ganglia are not united with the brain. 


The eyes of the Onychophora resemble the eyes of annelids in 
structure and development. An eye of the annelid-onychophoran 
type is formed from an invagination of the body wall (fig. 28 C), 
which becomes closed by an approximation or union of its lips (D, E), 
thus producing an inner optic vesicle (OpV) beneath an outer layer 
of epidermis and corneal cuticula (Cor). The cavity of the vesicle 
is occupied by a crystalline lens (Ln), probably of a cuticular nature, 
and its inner wall becomes the retina (Ret). In the onychophoran 
eye (A), as described by Dakin (1921), the lens is strongly convex 
outwardly and rests on the thick retina (Ret). Each retinal cell (B) 
is differentiated into a distal cylindrical rod (c) and a basal pigmented 
part (d), which contains the nucleus (Nu), and is prolonged proxi- 
mally as a nerve fiber (;//) that enters the optic lobe of the brain. 
The rods appear to have peripheral striations (e), but, as shown in 
cross-section (F), they do not form structures between them corre- 
sponding with the rhabdoms of arthropod eyes. 


The mesoderm bands of the Onychophora in their forward growth 
(fig. 22 B, C) continue into the head, where they form a pair of 
distinct coelomic sacs in the antennal region diverging anteriorly from 
the mouth (D). The cephalic coelomic sacs are described by Sedg- 
wick (1887) m Peripatopsis capensis, by Kennel (1888) in Peripatus 
edwardsi, and by Evans (1902) in Eoperipatiis tueldoni. The sacs 
are at first of large size (fig. 2"/ B) ; posteriorly their splanchnic walls 
embrace the stomodaeum (Sfom) and give rise to a part of the 


stomodaeal musculature ; anteriorly they extend into the antennae 
(C, AntCoel), and thus show their relation to these appendages. 
According to Evans, the antennal sacs acquire temporary coelomo- 
ducts (D, d) opening ventrally to the exterior (Cpr) within the 
circumoral fold (cof). With the increase in the size of the brain, 

Fig. 28. — Structure of the onychophoran eye. (A, B, F from Dakin, 1921.) 

A, vertical longitudinal section of eye of Peripatoides occidcutalis Fletcher, 
right half of retina depigmented. B, a retinal cell, differentiated into basal 
■plasmatic part {d) and distal optic rod (c). C, D, E, diagrams of development 
of an eye of the vesicular type (see also fig. 19 G). F, tangential section 
through optic rods of retina. 

a, b, outer and inner layers of corneal epidermis ; Br, brain ; c, optic rod of 
retinal cell ; Cor, cornea ; Ct, cuticula ; d, basal plasmatic part of retinal cell ; 
e, striated border of optic rod ; Epd, epidermis ; Ln, lens ; mcl, muscle fibers ; 
«/, nerve fiber; Nu, nucleus; OpNv, optic nerve; OpV, optic vesicle; Pig, 
pigment ; Ret, retina. 

the antennal sacs become reduced until finally, Evans says, they appear 
only as two small spaces situated above the brain in front of the eyes. 
A pair of small mesoderm masses observed by Evans in an embryo 
of Eoperipatus iveldoni, lying above and before the antennal sacs, in 
one of which a cavity was present (fig. 27 E, PrntCoel), are regarded 
by Evans as representing a pair of preantennal coelomic sacs, possibly 


corresponding with a pair of transient rudiments of preantennal 
appendages mentioned by Kennel in Peripatus edwardsi. 

The coelomic sacs of the body region conform with the series of 
postoral somites. The sacs of the jaw somite soon disappear. Those 
of the following somites attain a high state of development during the 
early embryonic period, leaving thus no doubt that the Onychophora 
are descended from typically metameric ancestors. The coelomic 
cavities become connected with the exterior by ventral diverticula 
from the mesodermal walls of the sacs (fig. 32 C, c) that unite with 
ectodermal invaginations {d), and thus form ducts opening on the 
mesal aspects of the bases of the legs (D). These outlet ducts of 
the coelomic sacs (coelomoducts) probably served primarily in the 
early history of the Onychophora for the discharge of excretory 
products and the gametes (fig. 34 A) ; but the coelomic sacs of the 
somites anterior to the somite of the definitive genital outlets become 
differentiated into dorsal gonadial and ventral nephridial compart- 
ments (B, C, a, b). The gonadial compartments eventually disappear 
except in a few posterior segments where they unite to form the 
gonads ; the nephridial compartments are reduced to the form of 
delicate vesicles at the inner ends of the coelomoducts (D, b), and 
thus persist as end-sacs of the definitive nephridia. In the somite of 
the genital outlet the entire coelomic sacs (figs. 32 E, 34 E, a, b) with 
their coelomoducts (d) are converted into the lateral genital ducts. 
The sacs of the second postoral somite become the salivary glands 
that open into the preoral mouth cavity. Derivation products of the 
coelomic walls include the entire muscular system, the dorsal pulsating 
blood vessel (fig. 2g, DV), and a muscular dorsal diaphragm (DDph) 
beneath the blood vessel. 


The body musculature of the Onychophora is in general similar to 
that of the annelids in so far as it consists mostly of flat sheets or 
bands of circular, oblique, and longitudinal fibers closely applied to 
the integument throughout the length of the animal (fig. 29), but it 
includes a series of lateral dorsoventral fibers {dvni) along each side 
of the body cavity, which have no representatives in annelid muscu- 
lature. These lateral muscles divide the body cavity into a median 
compartment {niBC) containing the alimentary canal (AlCnl) and 
the slime glands (SlmGld) , and lateral compartments (IBC) enclos- 
ing the salivary glands (SIC Id), the nephridia (Nph), and the nerve 
cords (NC). The muscle fibers are all very slender, and for the 


most part are not closely grouped into bundles forming specific 
muscles as in the arthropods. Each fiber is invested in a delicate 
sarcolemma, the nuclei are superficial, and the axis is distinctly fibril- 
lated but shows no trace of cross striatio.n (see Camerano, 1897). 

The following account of the onychophoran body musculature is 
based on a study of Peripatoides novae-zealandiae . When the body 
is laid open from above there are exposed on each side three sets of 

DV dm 




Fig. 29. — Cross-section of middle body region of Peripatoides novae-zealandiae 
Hutton, showing position of principal organs, diagrammatic. 

AlCnl, alimentary canal ; Com, commissure of nerve cords ; DDph, dorsal 
diaphragm; dm, dorsal muscles; DS, dorsal sinus; DV, dorsal blood vessel; 
dmn, dorsoventral lateral muscles; 113C, lateral compartment of body cavity; 
mBC, median compartment of body cavity; NC, nerve cord; Nph, nephridium; 
Npr, nephropore, SlGld, salivary gland; SlmGld, slime gland (reservoir); 
zmi, ventral muscles. 

fibers. Dorsally is a broad, thin band of internal dorsal longitudinal 
fibers (fig. 30, /), the more median fibers beginning anteriorly at the 
bases of the antennae, the more lateral ones behind the bases of the 
oral papillae. Ventrally is a much narrower band of ventral longi- 
tudinal fibers {2) lying along the midventral line. Between the dorsal 
and ventral longitudinal muscles is a series of fiat, closely adjacent, 
straplike lateral dorsoventral muscles (fig. 29, dvm, fig. 30, j), begin- 
ning anteriorly midway between the oral papillae and the first legs. 
When fully exposed, however, these lateral muscles are seen to be 



VOL. 97 

nearly semicircular in extent (fig. 29), since they are attached dorsally 
high up on the back external to the dorsal muscles, and ventrally along 
the midline of the body external to the median ventral muscles. 

By removing a section of the lateral muscles and the more lateral 
fibers of the dorsal muscles (fig. 30, left), there will be exposed two 
flat external laterodorsal longitudinul muscles (4, 5) lying above the 
leg base, an external lateroventral longitudinal muscle (6) mesad of 
the leg base, two dorsal muscles of the leg (y, 8), and a layer of 

Pig. 30. — Muscles of body wall of Peripatoidcs novae-zealmidiae Hutton. 
The various muscle layers exposed on right side of three successive segmental 

/, dorsal longitudinal muscles ; 2, ventral longitudinal muscles ; ,?, dorsoventral 
lateral muscles ; 4, 5, internal and external laterodorsal longitudinal muscles ; 
6, lateroventral longitudinal muscles ; 7, dorsal promotor of leg ; 8, dorsal re- 
motor of leg ; 9, internal oblique muscles ; 10, external oblique muscles (9 and 
10, reversed in position between legs) ; //, ventral promotor of leg; 12, ventral 
remoter of leg ; 13, circular muscles. 

oblique muscles (p, 10). The fibers of the leg muscles penetrate 
between the oblique fibers to make attachments on the body wall. 

The oblique muscles (fig. 30, p, 10) lie external to all the other 
muscles thus far described. They consist of two thin sheets of fibers 
crossing each other at right angles in opposite directions. The fibers 
that are internal on the back (p) go from above downward and for- 
ward; those that are external dorsally (/o) go downward and 
posteriorly. Just above each leg, however, a broad band of the ex- 
ternal fibers becomes internal by crossing over a similar band of the 
otherwise internal fibers (p) going below the leg from behind. Between 


each two successive legs, therefore, the relation of the two sets of 
oblique fibers is reversed. On the venter all the fibers again take the 
same relative position that they have on the back. The two sets of 
oblique fibers arise on the integument close to the middorsal and 
midventral lines, and are hence not continuous from one side to the 
other. External to the oblique fibers may be seen the anterior and 
posterior ventral muscles of the legs (fig. 30, right, 11, 12). 

Finally, outside all the other muscles of the body wall, are the 
circular muscles (fig. 30, /?). They consist of extremely fine fibers 
closely adherent to the inner surface of the integument, and are 
apparently continuous across the middorsal and midventral lines. 

A few other body muscles occur in the region of the mouth, and 
the jaws have an elaborate musculature quite different from the 
musculature of the legs (fig. 21 F). 


The appendages of the Onychophora include the antennae, the 
jaws, the oral papillae, and the legs. Their rudiments appear in the 
embryo as conical outgrowths of the body wall (fig. 23). The an- 
tennae arise from the anterior angles of the cephalic lobes (B, Ant) 
and retain this position. The jaws, which are the appendages of the 
first postoral somite, arise posterior to the mouth (A, /), but later 
they migrate mesally and forward (B), and are finally buried in the 
preoral mouth cavity (C), where they become reduced to a pair of 
double flattened hooks (fig. 21 F) converging in a horizontal plane 
beneath the mouth (D, /). The oral papillae are the appendages of 
the second postoral somite (fig. 23 A, OP), but in the definitive state 
they take a more anterior position at the sides of the mouth (fig. 21 D, 
E, OP). The legs retain their primary lateroventral positions (fig. 
23 B, C), and show but little variation in their final structure. 

The onychophoran appendages in their development give no evi- 
dence of having been derived from polychaete parapodia; they have 
no cirri or bristle sacs, and nothing suggests that they are composite 
organs formed of notopodial and neuropodial elements. The terminal 
claws of the onychophoran leg in no way resemble parapodial chaetae, 
and the general structure and musculature of the leg has little in 
common with a parapodium. except features that adapt each appen- 
dage to forward and backward movement on its base. On the other 
hand, the segmental appendages of the Onychophora and the Arthrop- 
oda have the same manner of origin and growth in the embryo, the 
organs in each case being hollow musculated lobes of the body wall, 



VOL. 97 

and it is only in their later development that they assume the structure 
characteristic of the adult appendages in each group. 

An onychophoran leg (fig. 31 A) is a hollow, conical outgrowth 
of the body wall terminating in a small pedal lobe bearing a pair of 
decurved claws. The leg integument is thrown into permanent circular 
folds, which on the thick basal part of the limb are covered with 

19a 16 

17a c 19b ci 

J^ *^ 19b id e i- v_T i/a c lyb d e 

Fig. 31. — Structure and musculature of an onychophoran leg, Peripatoides 
novae -zealatidiae Hutton. 

A, anterior view of a leg. B, lateral view of distal part of leg. C, horizontal 
section of basal part of leg. D, section of more distal part of leg. E, diagram- 
matic vertical section of distal part of leg. F, mesal view of leg. G, section 
of entire leg in transverse plane of body. 

a, b, c, d, distal nontuberculate rings of leg; e, claw-bearing pedal lobe; Npr, 
nephropore ; Un, claws ; 14, transverse muscle of leg base ; 15, peripheral muscles 
of basal part of leg ; 16, anteroposterior septal muscles of leg ; 17, flexor muscle 
of leg ; 17a, flexor of distal leg rings ; 18, circular muscles of foot ; 19a, 19b, 
two-branched retractor of claws. 

bristle-bearing tubercles. The distal folds, however, form distinct 
segmentlike rings (A, B, a, b, c, d) and are devoid of tubercles. 
The pedal lobe (<?) appears to be a larger terminal ring bearing the 
claws (Un). 

The leg is movable anteriorly and posteriorly on the obliquely 
transverse axis of its base by the four somatic muscles (fig. 30, 7, 
8, II, 12) that converge from the body wall into its basal opening. 
These muscles undoubtedly serve principally as promotors and re- 


motors, but are probably also levators and depressors of the leg as 
a whole. Within the leg the fibers of the four somatic muscles spread 
out into a thick peripheral layer of intrinsic leg fibers (fig. 31 C, D, 
G, 75) attached on the successive rings of the thick basal part of the 
appendage. Running through the narrow axial cavity of the leg is 
an antero-posterior muscular septum (16), the fibers of which diverge 
among those of the peripheral layer to the anterior and posterior walls 
of the leg (C, D). The rest of the leg muscles, except a slender 
transverse basal muscle (G, i/f), are motors of the distal rings and 
of the claws. The former include a bundle of fibers (ly) arising 
mesally in the leg base (G), with its fibers distributed to the ventral 
walls of the distal rings (E, G), and a series of strong circular 
muscles {18) in the pedal lobe. The claws are provided with a large 
two-branched muscle (E, G, ig), the larger branch arising in the 
base of the leg (G, ipa), the other in the distal part (E, G, ipb) ; 
the short common terminal part is inserted dorsally between the bases 
of the claws. The claw muscle is, therefore, a levator, or extensor, 
of the claws and has no antagonist. 

It is quite reasonable to suppose that the onychophoran leg is a 
prototype of the arthropod limb, but if we look for structural resem- 
blances in these two sets of locomotor organs we find few such, if 
any at all. The differentiation of the onychophoran leg into a thick 
basal part and a slenderer distal part, and the individualization of the 
distal rings, on which muscle branches are separately inserted, might 
be seen as an incipient segmentation. There is, however, no actual 
parallelism between the structure of the onychophoran leg and that 
of any arthropod leg, so that all we can say of the former is that it 
suggests a mode by which segmentation might arise in an ambulatory 
appendage. We may conclude, therefore, that the appendages of the 
Onychophora and the appendages of Arthropoda have had a common 
origin as lobiform outgrowths of the body wall containing extensions 
of the somatic muscles. The common need of a mechanism for an- 
terior and posterior movement of each appendage on its base then 
brought about a differentiation of the extrinsic parts of the limb 
muscles into promotors and remotors, while the parts of the muscles 
within the leg were elaborated to give greater efficiency to movements 
of the leg itself. The further course of evolution producing segmen- 
tation and correlated musculation in the limb evidently has proceeded 
independently in the Onychophora and the Arthropoda from a very 
primitive common beginning, and has gone much farther in the 
Arthropoda than in the Onychophora. 



The Onychophora are provided with numerous fine tubular in- 
growths from the body wall, which undoubtedly serve for respiration, 
and are therefore termed tracheae, though it is possible that ana- 
tomically they are more of the nature of insect tracheoles. The 
tubules, which are only one to three microns in diameter, arise in 
dense bundles (fig. 32 B, Tra) from small flask-shaped pits (tp) of 
the integument, and extend long distances into the body cavity. The 
tracheal pits may be very numerous ; they occur on all parts of the 
body, on the head, and around the mouth, but they are most abundant 
on the back, where several may occupy the space of a square milli- 
meter. For the most part the pits are irregularly distributed, but in 
some species they are arranged in longitudinal rows. The tracheal 
bundles issuing from the inner ends of the pits contain large, con- 
spicuous nuclei in their basal parts (Nu), which probably pertain to 
the matrix cells, but the tubes themselves diverge and extend far 
beyond these nuclei. According to Dakin (1920), the tracheal walls 
are strengthened by excessively minute but perfect spiral fibers visible 
in fresh material. In their distal parts the tracheae are branched and 
go to practically all the internal organs, but their final terminations 
have not been observed. 

Since tracheal invaginations of the body wall are developed for 
respiratory purposes in nearly all groups of terrestrial arthropods, 
the mere presence of such organs can have no taxonomic significance, 
any more than has the presence of gills in diverse groups of aquatic 
animals. Inasmuch as invertebrates breathe through the skin in any 
case, evaginations or invaginations of the integument are about the 
only devices they can develop for improving their respiratory 


The blood vascular system of the Onychophora consists only of a 
tubular dorsal vessel (fig. 29, DV) extending the entire length of 
the body, said to be open anteriorly and posteriorly. The walls of 
the vessel consist of circular muscle fibers, and are perforated dorsally 
in each segment by a pair of ostia. The tube is suspended from the 
body wall by connective tissue strands, and is supported on a mem- 
branous and muscular dorsal diaphragm (DDph). The diaphragm 
muscles are fine, regularly transverse fibers medially attached on the 
ventral wall of the blood vessel ; laterally they penetrate between the 
fibers of the dorsal somatic muscles and are apparently attached on 


Fig. 32. — Internal structure of Onychophora, and later development of the 
coeloniic sacs. 

A, general view of internal anatomy of Peripatoidcs twvae-acalaiidiac Hutton, 
female, dorsal view ; muscles, nephridia, peripheral nerves, and dorsal blood 
vessel omitted. B, tracheal pit of Peripatopsis capcnsis Grube and respiratory 
tubules extending inward from it (from Schneider, 1902). C, section of embryo 
of Peripatus edwardsi Blanchard, showing constriction of coelomic cavity into 
dorsal gonadial compartment (a) and lateral nephridial compartment {h) ; rudi- 
ments of coelomoduct {c, d) not yet united (from Kennel, 1888). D, same, 
later stage, dorsal compartment (a) of coelomic sac (which later disappears 
except in genital somites) and nephridial compartment {b) entirely separated, 
coelomoduct {c, d) open to exterior (from Kennel, 1888). E, section of embry- 
onic somite of genital outlets, coelomic sacs narrowed but not divided as in 
other segments (C, D), continuous from gonads (a) through coelomoducts {d) 
to exterior (from Kennel, 1888). F, male reproductive organs of Peripatopsis 
blainvilici Gay-Gervais (from Bouvier, 1902, with accessories omitted). G, sper- 
matophore of same (from Bouvier, 1902). 

a, dorsal gonadial compartment of coelomic sac ; AcGld, genital accessory 
gland; AlCnl, alimentary canal; Ant, antenna; AntNi', antennal nerve; b, ne- 
phridial compartment of coelomic sac ; Br, brain ; c, mesodermal component of 
coelomoduct ; Com, nerve commissures ; Ct, cuticula ; d, ectodermal component 
of coelomoduct; Dej, ductus ejaculatorius; E, eye; Epd, epidermis; Epdm, 
epididymis ; mcl, muscle ; NC, nerve cord ; Nu, nucleus ; Od, oviduct ; Oe, 
oesophagus, OP, oral papilla ; Ov, ovary ; Phy, pharynx ; Res, reservoir of 
slime gland ; Rcct, rectum ; SlGld, salivary gland ; SlmGld, slime gland ; Sphr, 
spermatophore ; Tes, testis ; tp, tracheal pit ; Tra, tracheal tubules ; Utrs, uteri ; 
Vd, vas deferens; Vent, ventriculus; VO, ventral "organ"; Vsm, vesicula 



the body wall. Above the diaphragm on each side of the blood vessel 
are masses of small individual cells, probably "nephrocytes." The 
circulatory system of the Onychophora thus resembles that of the 
Chilopoda and the Hexapoda in the simplicity of its structure. Since 
many of the arthropods, in common with the annelids, have a highly 
developed blood vascular system, it would seem probable that the 
simpler forms represent reductions from a more elaborate primitive 
system such as that of the Annelida. 


The nephridialike excretory organs of the Onychophora are paired 
segmental structures usually present in all the somites between the 
somite of the oral papillae and that of the genital ducts, though they 
may differ much in size and in the relative development of their parts. 
They lie in the lateral compartments of the definitive body cavity at 
the bases of the legs (fig. 29, Nph), and open externally in grooves 
on the ventral surfaces of the leg bases (figs. 29, 31 F, 33 A, Npr), 
except those of the fourth and fifth pairs, which in most species open 
at the bases of the distal rings of the legs (fig. 33 C). 

A well-develoi>ed onychophoran nephridium consists of five distinct 
parts (fig. 33 A) : First, beginning externally, is a short outlet duct 
{Nd) ; second, a bladderlike enlargement, or reservoir {Bl) ; third, 
a tubular canal {Cnl) varying in length and usually coiled; fourth, 
a funnel-shaped enlargement of the inner end of the canal {Fun) ; 
and fifth, a thin-walled end-sac (ESc). The walls of the funnel (B) 
are relatively thick and are histologically different from the rest of 
the canal ; they are clothed with long vibratile cilia directed toward 
the nephridial exit (see Dakin, 1920, Cuenot, 1926, Zilch, 1936). 

The funnel and the canal of an adult onychophoran nephridium 
are comparable with an entire metanephridium of the annelids ; the 
end-sac is a remnant of the coelomic sac of the embryonic somite. 
The opening of the nephridial funnel into the end-sac, therefore, is 
the nephrostome (fig. 33 B, Nst). The canal is developed in the 
embryo as an exit duct of the coelomic sac, formed by the union of 
a ventral diverticulum of the sac (fig. 32 C, c) with a tubular in- 
growth (d) from the ectoderm of the same segment mesad of the 
leg rudiment (D). The primitive function of the coelomoducts un- 
doubtedly was the discharge of excretory products and, in the 
genital segments, of the gametes. Embryonic coelomoducts occur, 
according to Evans (1902), in connection with the coelomic sacs of 
the antennae (fig. 27 D, rf), and in all the postoral somites except 


the somite of the jaws. During embryonic development the coelomic 
sacs of those segments that eventually contain nephridia become each 
constricted into a dorsal section (fig. 32 C, a) and a ventral section 
{b), which soon become entirely separate compartments (D). Except 
in the genital region the dorsal compartments disappear; in the 
nephridial somites the ventral compartments become much reduced, 
but they retain their open connections with the coelomoducts, and 
persist as the delicate end-sacs of the nephridia (fig. 33, ESc). 

Fun Cil 


Fig. 22>- — Structure of onychophoran nephridia. 

A, diagrammatic transverse section of leg and nephridium of mature embryo 
of Peripatopsis capensis Grube (from Sedgwick, 1888). B, inner part of nephrid- 
ium of Peripatoidcs sp., showing ciliated funnel (Fun) with nephrostome open- 
ing into coelomic end-sac (from Dakin, 1920). C, diagrammatic transverse sec- 
tion of leg and nephridium of adult Peripatus tholloni Bouvier (from Fedorow, 

Bl, nephridial bladder ; Cil, cilia ; Cnl, nephridial canal ; ESc, coelomic end- 
sac of nephridium ; Fmi, nephridial funnel ; NC, nerve cord ; Nd, nephridal duct ; 
Npr, nephropore ; Nst, nephrostome. 

It is commonly held that the excretory organs of the Onychophora 
are homologous with the annelid metanephridia (see Glen, 1919). 
The simple development' of the canals as open ventral diverticula of 
the coelomic walls (not of the septa), the direct opening of the canals 
to the exterior on the same segment, and the occurrence of embryonic 
coelomic ducts in the head, however, are all features distinctive of 
the Onychophora. Considering, therefore, that there is little proba- 
bility on other grounds that the Onychophora have been derived from 
annelids having metanephridia, we may conclude that the open 


nephriclia of the higher Annehda and the coelomic exits of the Ony- 
chophora have been separately acquired and developed in each group. 
On the other hand, there can be little doubt that the nephridial organs 
of Arthropoda (antennal, maxillary, and coxal glands) are entirely 
comparable with the onychophoran nephridia. 


In the evolution of specific reproductive organs the Onychophora 
are far in advance of any of the polychaete or oligochaete annelids ; 
but the development and the definitive structure of the genital organs 
are so closely parallel in the Onychophora and the Arthropoda that 
we can scarcely question the probability of the genital systems in 
these two groups having had a common origin. In fact, it is the 
fundamental similarity in the genital system that would appear to 
constitute the closest bond of union between the Onychophora and 
the Arthropoda, and which most strongly suggests that the two groups 
have been derived from a common progenitor. The germinal centers 
of the Onychophora, as in the arthropods, are entirely enclosed in 
gonadial sacs of coelomic derivation, and the gametes are discharged 
through ducts whose lumina are continuous with those of the gonads. 
An approach to a closed genital system is seen in the Oligochaeta in 
the development of coelomic seminal vesicles containing the genital 
outlet funnels, and a system as completely closed as that of the Ony- 
chophora and Arthropoda is perfected in the Hirudinea ; but the 
ontogeny of the organs in these several groups shows that there is 
no possibility of the onychophoran-arthropod reproductive system 
having been evolved from that of the higher annelids. 

The primary germ cells of the Onychophora become localized at 
an early stage of embryonic development in the median dorsal parts 
of the splanchnic walls of one or several posterior pairs of coelomic 
sacs (fig. 34 A, Gnn). According to Evans (1902) there are four 
embryonic genital somites in Eoperipatus weldoni, while Kennel 
(1888) says the germ cells of Peripatus edzvardsi occur in but 
one somite. Whatever the number of genital segments may be in 
modern forms, we must suppose that the germ cells once occupied 
most of the somites, for the early embryonic relation of the germinal 
centers to the coelomic sacs is identical with the adult condition in 
the Polychaeta, and undoubtedly means that in the primitive Ony- 
chophora the gametes were discharged into the coelomic sacs (A, 
Sps), and were liberated from the latter through the coelomoducts 
(d). As we have seen, the upper parts of all the coelomic sacs between 


the somite of the oral papillae and the somite of the genital ducts 
become constricted from the ventral parts (fig. 32 C, a), and then 
separated as independent dorsal compartments (D, a). In the pre- 
genital somites the dorsal compartments disappear, but in the defini- 
tive genital somites they persist as gonadial sacs containing the 
germaria (fig. 34 C, (7). The gonadial sacs of each lateral series, 



Fig. 34. — Diagrams showing the transformation of the onychophoran coelomic 
sacs and coelomoducts into genital organs and nephridia. (From Snodgrass, 
1936, based on Sedgwick, 1885, Kennel, 1888, and Evans, 1902.) 

A, theoretical primitive stage in which excretory products and the gametes 
were discharged from the coelomic sacs through coelomoducts. B, C, D, differ- 
entiation and division of the coelomic sacs into dorsal gonadial sacs (a) and 
ventral nephi*ic sacs (6), the last finally reduced (D) to end-sacs of the 
nephridia. E, gonadial sacs of definitive genital segments united on each side 
in a gonadial tube (G) opening through undivided coelomic sac of penultimate 

a, gonadial compartment of primitive coelomic sac ; AlCnl, alimentary canal ; 
An, anus ; b, nephric compartment of coelomic sac ; BC, definitive body cavity 
(haemocoele) ; c, nephridial diverticulum of coelomic sac; Coel, coelomic cavity; 
Cpr, coelomopore; d, ectodermal part of coelomoduct ; DV , dorsal blood vessel; 
G, gonad ; Gdl, lateral gonoduct ; Gld, accessory genital gland ; Grm, germarium ; 
Msd, mesoderm; NC, nerve cord; Nph, nephridium; Spz, spermatozoa; VO, 
ventral "organ" of ectoderm. 

however, unite in a continuous tube (E, G), which becomes the 
definitive gonad with a germinal band in its ventral wall (C, D, (7). 
Furthermore, the posterior ends of the gonadial tubes open into the 
coelomic sacs of the following somite, and these sacs, which maintain 
their integrity, and their continuity with the coelomoducts (fig. 32 E, 
a, b, c, d), become the lateral genital ducts (fig. 34 E, Gdl). Eventu- 


ally the apertures of the lateral ducts come together on the midline 
of the venter, where they are carried inward at the end of an ecto- 
dermal invagination that forms a common definitive exit tube, the 
ejaculatory duct or median oviduct. 

The adult reproductive organs of the Onychophora are strikingly 
arthropodan in character. In the male, the testes retain the tubular 
embryonic form (fig. 32 F, Tes) ; each discharges into a seminal 
vesicle (Vsin) from which proceeds a long tubular vas deferens 
(Vd), the anterior part of which is thrown into an epididymislike 
mass of coils (Epdm). The ejaculatory duct (Dej) is usually long 
and irregularly looped ; its opening is on the region of the penultimate 
somite. Associated with the gonopore is a pair of tubular accessory 
glands {AcGld), said to be the reduced coelomic sacs of the last 
somite (fig. 34 E, Gld). In the female, the tubular ovaries are 
united at their extremities and lie on the dorsal surface of the ali- 
mentary canal in the posterior part of the body (fig. 32 A, Ov). The 
oviducts {Od) proceed first forward from the posterior ends of the 
ovaries, and then turn backward to unite beneath the rectum (Red) 
in a very short terminal atrium, or common oviduct, opening in the 
same position as the gonopore of the male. In viviparous species the 
intermediate parts of the oviducts are enlarged in a series of uterine 
chambers (Utrs) containing the embryos. Sperm receptacles usually 
occur on the lateral oviducts near their ovarian ends. 


The fundamental characters of the arthropods are those of the 
Onychophora and the Annelida. The three groups have in common 
the following features: (i) The ventral elongation of the blastopore 
and the closure of its intermediate part, resulting in the formation 
of a tubular enteron with a ventral subapical mouth and a terminal 
anus, and in the conversion of the preblastoporic region of the trunk 
into a prostomial cephalic lobe; (2) a definitive tripartite alimentary 
canal composed of the primitive endodermal enteron, and of a secon- 
dary ectodermal stomodaeum and proctodaeum; (3) the dififeren- 
tiation of a part of the mesoblast, originally formed in the posterior 
end of the body, into a specific mesoderm taking the form of ventro- 
lateral bands that extend forward through the entire length of the 
body and penetrate into the prostomium; (4) metamerism of the 
somatic ectoderm and mesoderm, involving a segmental repetition of 
organs derived from these germ layers; (5) the continuity of the 
acronal centers of the primary nervous system with the somatic centers 


secondarily developed in connection with metamerism; (6) internal 
cleavage of the mesoderm segments to form paired coelomic cavities ; 
(7) a somatic muscular system applied against the body wall, con- 
sisting primarily of an outer set of constrictor fibers running in 
transverse planes, and of an inner set of contractor fibers taking a 
longitudinal course, each of which may be variously amplified or 
reduced ; (8) the development of a blood vascular system from the 
mesoderm, composed essentially of a dorsal and a ventral longitudinal 
vessel connected by lateral vessels, but often reduced to a dorsal 
vessel and more or less well-defined sinuses; (9) the association of 
the germ cells with the walls of the coelomic sacs, and their discharge 
into the coelom. 

The common basic features of organization above enumerated 
attest the origin of the Arthropoda, the Onychophora, and the higher 
Annelida from a common ancestral form, which itself must necessarily 
be visualized as a generalized annelid. It is to be assumed that the 
progenitors of the three groups had already acquired a lengthened 
body by the addition of secondary genital somites proliferated from 
a subterminal zone of growth. Though teloblastic growth does not 
appear in the ontogeny of the Onychophora, it is quite as character- 
istic of certain arthropods as of the annelids. 

The Arthropoda have in common with the Onychophora the follow- 
ing nonannelid characters : ( i ) A chitinous ectodermal cuticula ; 
(2) segmental ambulatory appendages formed as simple outgrowths 
of the body wall, which in their structure and development give no 
suggestion of a community of origin with the composite parapodia 
of the Polychaeta; (3) segmental excretory organs (antennal, maxil- 
lary, and coxal glands) that resemble the nephridia of Onychophora 
in being remnants of coelomic sacs connected with the exterior by 
simple coelomoducts, but which have neither the anatomical position 
nor the development of annelid metanephridia ; and (4) closed gona- 
dial sacs of coelomic origin, containing the germarial centers in their 
walls, and connected with the exterior by a pair of coelomic sacs set 
apart to serve as genital ducts. A feature characteristic of both the 
Arthropoda and the Onychophora is the restoration of the haemocoele 
as the definitive body cavity, resulting from the reduction of the 
coelom to the cavities of gonadial and nephridial sacs, but it is not 
distinctive of them because an obliteration of the coelom occurs also 
in certain annelids. 

The small but important assemblage of characters given above as 
common to the Onychophora and the Arthropoda would seem to in- 
dicate that the two groups have been evolved from the same ancestral 


Stock, which arose from some generahzed nonchaetopodous anneHd; 
but since none of the modern anneHds has these characters it is 
evident that the annehdan progenitors of the Protonychophora- 
arthropoda have left no direct descendents. The Arthropoda differ 
in so many respects from present-day Onychophora that it is certain 
they must have branched off from the common onychophoran- 
arthropod trunk before the latter had gone far in the onychophoran 
direction. Arthropod forms were highly developed and differentiated 
in the early Cambrian period of geological history, and must, therefore, 
have had their origin in remote pre-Cambrian times, though in the 
rocks of this period there is no specific evidence of their existence. 
As an individualized group, the Arthropoda are characterized 
particularly by the development of hard plates in the cuticular layer 
of the integument, separated by areas of flexibility. In the Mandib- 
ulata sclerotization results from the presence of nonchitinous sub- 
stances in the otherwise chitinous cuticula; in the Trilobita and 
Chelicerata sclerotization may be due to a structural differentiation 
of the chitin itself, though apparently little attention has been given 
to the chemical composition of the cuticular skeleton in these groups. 
Ruser (1933) describes the physical structure of "hard chitin" and 
"elastic chitin" in the Ixodidae, but makes no determination of their 
chemical nature. 

Since the muscles are primarily attached on the body wall, the 
differentiation of the latter into hard and flexible areas at once created 
a possibility for unlimited development of skeletomuscular mecha- 
nisms, and it is through the elaboration of such mechanisms that the 
arthropods have attained their exalted position among the articulates, 
and their wonderful diversity of structure. It is true, of course, that 
some of them, particularly those that have taken up parasitic habits, 
have renounced their birthright, and among the latter we find examples 
of physical degeneration carried to such an extent that every semblance 
of arthropod structure may be lost. 

Sclerotization of the integument involved first a complete change 
in the mechanism of body movement, for if the rings of flexibility 
between segmental plates remained at the primary intersegmental 
grooves, on which the longitudinal muscles are attached, there would 
be little if any possibility of movement. Hence, each dorsal and 
ventral plate includes the primary intersegmental groove in front, 
while the areas of flexibility occupy the posterior parts of the seg- 
mental regions. The sclerotized parts of the primary intersegmental 
grooves, carrying the muscle attachments, thus come to form internal 
ridges, or antecostae, on or near the anterior margins of the definitive 


tergal and sternal plates, and the primary intersegmental grooves 
become the submarginal antecostal sutures. As a consequence, a new, 
secondary type of segmentation has been established, in which the 
functional intersegmental rings are the membranous posterior parts 
of the primary segments, and the action of the longitudinal muscles 
becomes intersegmental instead of intrasegmental. A body mechanism 
of this kind is typical of all the arthropods, but still it is by no means 
fixed, for innumerable modifications of it have been introduced 
in adaptation to the development of special structures for specific 

The acquisition of an exoskeleton necessarily limits freedom of body 
movement, such as that possessed by the highly flexible annelids, 
but at the same time it furnishes a mechanism by which movements 
may become more specific, since the development of definite hinge 
joints becomes possible, and muscles can assume more effective 
antagonistic relations to each other. The longitudinal muscles lose 
nothing of their efficiency, but their contraction now results in a 
telescoping of the body segments. The presence of dorsal and ventral 
plates, however, necessarily eliminates the constrictor effect of the 
primitive circular or semicircular muscles ; the latter, therefore, have 
become reduced to lateral tergosternal muscles, the contraction of 
which produces a flattening of the body. The primitive mechanism 
of dilation and extension by unequal distribution of internal pressure 
is still operative ; but the potentiality of developing endoskeletal 
structures gives the possibility of a new mechanism of expansion, for 
the ingrowth of apodemal arms from tergal or sternal areas, on which 
primarily compressor muscles are attached, may reverse the position 
of such muscles to the extent that they become dilators. A separation 
of contiguous plates, however, may be brought about also by the 
contraction of intersegmental muscles that have been reversed by 
the overlapping of the plates. All these mechanical devices and many 
others are variously and often highly developed in the different 
arthropod groups, and their elaboration has set the arthropods far 
above the annelids and onychophorons in the power of performing 
definite and specific acts. Even the wing mechanism of pterygote 
insects has been built up from little more than the skeletal parts and 
musculature common to the body segments. It should be observed, 
however, that although the musculature of the body segments and 
the appendages is fairly definite and fixed within the major arthropod 
groups, there seems to be no limit to the potential genesis of new 
muscles in connection with special organs, such as the male genitalia 
of insects, and, furthermore, that the entire body musculature is 


subject to adaptive changes, which may be very extensive, as in 
certain holometabolous insect larvae. 

Sclerotization of the integument has affected not only the wall of 
the body, but also the walls of the tubular segmental appendages, and 
the latter are jointed by definite rings of flexible membrane interposed 
between the resulting limb segments, or podomeres. Hence, the 
arthropod limb itself has possibilities of much variety and specificity 
of action. As a consequence, while probably the appendages in the 
first place were all simple locomotor organs, many of them have been 
converted into instruments adapted to various purposes, and those 
that still subserve the locomotor function are capable of all the kinds 
of mechanical progression except flying known among animals. 

Concomitant with the evolution of the skeletomuscular mechanisms, 
the nervous system and the sense organs have necessarily acquired a 
high state of development, and the elaboration of most intricate in- 
stincts has been possible because of the facility with which tools may 
be produced and adapted to their ends. 

The primitive arthropods, being closely related to the primitive 
onychophorons, and together with the latter derived from generalized 
annelids, must have been slender, many-segmented, polypodous crea- 
tures resembling modern centipedes. They differed from their con- 
temporaneous onychophoran relatives in having dorsal and ventral 
segmental plates and specifically jointed appendages. The Protarthrop- 
oda were early differentiated into primitive trilobites and primitive 
mandibulate forms. From the primitive trilobites were evolved the 
later Trilobita, Xiphosurida, Eurypterida, and Arachnida, while the 
Protomandibulata gave rise to the Crustacea, the Diplopoda, the 
Chilopoda, and the Hexapoda. 


The processes of cleavage and germ-layer formation are so variable 
among the arthropods that they can have little value in a phylogenetic 
study of arthropod relationships. Cleavage, whether total or partial, 
results usually in the formation of a superficial blastoderm, and the 
embryo appears as a germ band on the ventral side of the egg. 
Gastrulation in some of the Crustacea takes place by invagination, 
but more commonly both the endoderm and the mesoderm are formed 
by delamination or by proliferation from the blastoderm or the germ 
band. Manton (1928) gives a precise account of the proliferation of 
the germ layers and the primary germ cells from the blastoporic 
region in the crustacean Ilemimysis, the cells of the several groups 


being first differentiated on the surface of the germinal disk. The 
first endoderm cells in many of the arthropods scatter through the 
yolk as independent trophocytes (vitellophags) and the definitive 
enteron may then be formed either by a reassembling of the cells 
about the yolk, or by regeneration from intact endodermal rudiments. 
The mesoderm in some of the Crustacea, Chilopoda, and Chelicerata 
is proliferated forward from a posterior generative zone very much 
in the manner of the onychophoran mesoderm, and suggestive of 
the teloblastic origin of the coeloblast in the Annelida. Among the 
Crustacea there are in fact a few cases in which the mesoderm takes 
its origin, at least in part, from a single pair of teloblastomeres derived 
from the endoderm, as in the cirriped Lepas. The mesodermal telo- 

FiG. 35. — Early stages in the development of a cirriped, Lepas. (Simplified 
from Bigelow, 1902.) 

A, 8-cell stage, with large yolk-filled posterior cell. B, 30-cell stage, endoderm 
surrounded by mesoderm comprising a posterior cell (Msd) of endodermal 
origin, and four cells (msd) of ectodermal origin. C, the posterior mesoblast 
cell divided into mesodermal teloblasts (MsT). D, near close of gastrulation, 
but with mesoderm cells still exposed. 

Bpr, blastopore ; Ecd, ectoderm ; End, endoderm ; Msd. endodermal meso- 
derm ; msd, ectodermal mesoderm ; MsT, mesodermal teloblast ; vCl, yolk- 
filled cell at vegetative pole of morula. 

blasts of Lepas, according to Bigelow (1902), appear in the 32-cell 
stage on the posterior lip of the blastopore (fig. 35 C, Mst), and are 
produced from a single mesoblast cell (B, Msd) that results from 
the division of a primary yolk-filled blastomere (A, z'Cl) at the 
posterior pole of the morula. Four other mesoblast cells, however, 
are formed in Lepas from the ectodermal lips of the blastopore 
(B, C, msd), and eventually the entire mesoblast sinks into the blasto- 
pore (D). A separate destiny of the mesoblast from the two sources, 
entoblastic and ectoblastic, has not been distinguished in Lepas, but 
it is a point of much interest to note that here the mesoblast completely 
surrounds the open blastopore between the ectoderm and the endo- 
derm, a part of it being of endodermal and a part of it of ectodermal 


derivation. It is not difficult, then, to understand from this condition 
how, in forms having a closed blastopore, the coelomic mesoblast may 
arise from the entire length of the linear blastoporic area, and we 
may further see some significance in the statement by Sedgwick 
(1887) that in the onychophoron Pcripatopsis the mesoderm bands 
in their forward growth are augmented by cells derived from the 
lips of the blastopore. In the more specialized types of arthropod 
development evidence of teloblastic generation of the mesoderm is 
entirely lost, or at least obscured, and the whole of the mesoderm 
appears to be a direct product of the germ band closely associated 
with the endoderm. In its full development the arthropod mesoderm 
surrounds the blastopore anteriorly, since in the adult the lateral bands 
of the cephalic mesoderm may be continuous from side to side in 
front of the mouth. 

Segmentation of the mesoderm and the subsequent formation of 
coelomic sacs take place in the early embryonic stages of many Crus- 
tacea and Arachnida almost as completely as in the Onychophora and 
Annelida, but in the myriapods the coelomic sacs are small, and in 
the insects they are for the most part represented only by cleavage 
spaces in the lateral parts of the mesoderm. In all cases, however, 
the walls of the sacs break down, except such parts of them as are 
retained in the formation of certain organs of coelomic origin, and 
the haemocoele is restored as the definitive body cavity. Probably 
all muscle tissue of the arthropods is produced from the coelomic 
mesoblast ; though some writers have claimed that certain muscles 
are produced directly from the ectoderm, the evidence is open to 
question and needs closer scrutiny (see Needham, 1937). 


There is ample reason from arthropod ontogeny for believing that 
the arthropods have been derived, as have the annelids, from primi- 
tively unsegmented ancestral forms in which metamerism first ap- 
peared as a direct subdivision of the primary body region into a small 
number of somites, and that the subsequent increase in the number 
of somites proceeded secondarily from growth in a subterminal 
zone of undifferentiated cells. This dual method of somite produc- 
tion is recapitulated in the embryogeny of some of the arthropods, 
and teloblastic growth is of frequent occurrence in postembryonic 

In the Trilobita it seems very probable, as contended by Iwanoff 
(1933) and Schulze (1936), that the so-called head represents the 


area of primary segmentation, for there is no doubt that the post- 
cephaHc segments are produced by teloblastic growth. The youngest 
trilobite larvae known give no evidence of metamerism (fig. 46 A), 
but there soon appears in the glabellar region four pairs of lateral 
impressions or transverse grooves that divide the glabella into five 
consecutive lobes (fig. 36 A). These depressions produce internal 

Fig. 36. — Segmentation and tagmosis of Trilobita. 

A-D, four successive stages in larval development of Liostracus linnarssoni 
Brogger (from Warburg, 1925). E, Olcnclhis vermoHtanus Hall (from Wal- 
cott, 1910). F, Olenellus gilberti Meek (from Walcott, 1910). G, Schmidtiel- 
lus mickwitsi Schmidt, distal body segments (from Walcott, 1910). H, 
Asaphiscus wheelcri Meek, example of a trilobite with distal segments united 
in a caudal fan, or pygidium (from Walcott, 1916). I, Agnostis montis Matthew, 
example of the group Agnostia having only two free segments between head 
and pygidium (from Walcott, 1908). 

jg, fixed cheek, or fixigene; jrl, frontal lobe; H, head; Ig, free cheek, or 
libragene ; Pyg, pygidium ; sp, spine ; Tel, terminal lobe of body, probably the 
telson ; Th, thorax ; ZG, zone of growth. 

ridges or apodemes most probably for muscle attachments, and their 
formation, therefore, does not represent the process of segmentation 
itself, but unquestionably they mark the primary intersegmental lines 
of the segments united in the larval body. The first glabellar division, 
known as the frontal lobe (A, frl), is continuous with a pair of 
lateral areas {Ig) that become the "free cheeks" of the adult bearing 
the compound eyes (fig. 46 E, Ig). The frontal lobe, therefore, may 


be regarded as a part of the eye segment, or acron, and further reasons 
for so regarding it will be given later. The other four glabellar lobes 
must then represent four primary larval somites, the intersegmental 
lines of which should, theoretically, have extended to the lateral 
margins of the simple oval body before segmentation in the latter was 
suppressed. The postlarval somites of the adult trilobite are generated 
teloblastically (fig. 36 B, C, D) from a small region of the larva 
behind the glabella (A, ZG), and are, therefore, clearly secondary 
somites. The definitive segments of the postcephalic series remain 
distinct in some of the trilobites to the end of the body (E, F), where 
there is a small terminal lobe (E, G, Tel?) that may be the telson; 
in others the posterior segments are united in a tail-fan, or pygidium 
(H, Pyg), and in the Agnostia (I) only two segments retain their 
independence between the head and the pygidium. 

The Xiphosurida in the adult stage resemble the Trilobita in so 
many respects that we should expect to find an even closer approach 
to the trilobite structure in their developmental stages ; and, in fact, 
it has been shown by Iwanoflf (1933) that the primary segmentation 
in the embryo of Liinuhis moluccamts produces four somites (fig. 
37 A, I-IV), those of the chelicerae, the pedipalps, and the first two 
pairs of legs, which evidently represent the four postacronal head 
somites of a trilobite. Because of the large amount of yolk in the 
ectoderm, embryonic metamerism appears first in the mesoderm, 
which is early divided almost simultaneously into four sections corre- 
sponding with the four primary somites. The preoral cephalic region 
of L. moluccanuSj IwanofT says, is at first not distinctly differentiated 
from the surrounding blastoderm, but later it becomes apparent as a 
preoral head segment without appendages, and in an older embryo it 
forms a pair of definite cephalic lobes (B, Pre). Behind the fourth 
somite there is in the young embryo (A) only an unsegmented tail 
piece, but at the base of this region are later generated consecutively 
(B) the remaining segments of the adult, which are thus typically 
teloblastic in the manner of their formation. 

It would thus appear that the primary segmentation of the ancestors 
both of the trilobites and the xiphosurids produced only four somites. 
These four primary somites, united with one another and with the 
cephalic lobe, or acron, constitute the "head" in the Trilobita (fig. 
36 H, H) ; in the Xiphosurida they form the anterior part of the 
prosoma, for in this group three following somites and part of a 
fourth are combined with the four primitive somites in the anterior 
section of the body (fig, 47 E). Moreover, in the Xiphosurida a 
union has taken place between all the opisthosomatic somites, so that 


"--^ n\.\ --^ .-^ ^-^Pro— >Aor 

Fig. 37. — Embryonic and adult segmentation of Liiiuilus. (A-E from Iwanoff, 

A, Linmlns nwluccamis Linn., germ band with mesoderm divided into four 
postoral somites, cephalic lobes not yet differentiated from blastoderm. B, same, 
embryo with nine pairs of appendages, cephalic lobes (Pre) present. C, first 
instar larva, segments of opisthosoma indicated by internal mesoderm bands 
before moulting. D, first instar larva before moulting stage. E, second instar 
larva. F, Liviulus polyphemus Linn., young adult, veqtral view, prosomatic 
appendages removed, showing radial position of their bases around the central 

Acr, acron, derived from procephalic lobes of embryo; Au, anus; Chi, 
chilarium; CIil, chelicera; csp, caudal spine; dbl, doublure; dO, dorsal ocellus; 
E, compound eye; fg, fixigene ; gib, glabella; I-VI, first six somites; L, leg; 
Ig, libragene ; Lm, labrum ; Mth, mouth ; Opl, genital operculum ; Pdp, pedipalp ; 
Pre, procephalic lobe; vO, ventral ocellus. 


there is no intermediate region of free somites as in the Trilobita 
and Agnostia (fig. 36 H, I). 

The adult structure of Limulus contains evidence of the presence 
of 14 postoral somites, the last somite being behind the last gill- 
bearing segment (fig. 47 D, XIV) ; but Iwanofif (1933) says that in 
the embryo rudiments of three somites appear in the postbranchial 
region, giving thus a total of 16 somites anterior to the caudal spine. 
The caudal spine of the Xiphosurida is often called the "telson," but, 
as shown by Schulze (1936), a comparison with the subterminal spine 
of such trilobites as Mcsonacis and Olenellus (fig. 36 E, F, G), which 
arises from a segment some distance from the end of the body, sug- 
gests that the caudal spine of the xiphosurids may not be a true 
terminal structure, and that several primitive somites beyond it may 
have been lost. 

Studies on the embryogeny of Arachnida have not brought out any 
distinction between primary and secondary somites, and the arachnids 
have no postembryonic teloblastic growth. Schulze (1936), however, 
has pointed out many features in the adult structure of the arachnids, 
especially in the Acarina, that suggest the trilobite type of segmen- 
tation. The area of the four primary somites, he shows, is often 
evident as a differentiated anterior region of the prosoma, and in the 
segmentation and body form of such acarinids as Oxypleurites there 
may be seen a striking general resemblance to a mesonacid trilobite. 
The arachnid prosoma contains six postacronal somites, and in this 
respect, therefore, is intermediate between the trilobite "head" and 
the xiphosurid prosoma. 

The embryonic development of segmentation in the Crustacea 
has been particularly studied by Sollaud (1923) in the palaemonid 
Leander. The germ band of Leander is at first V-shaped (fig. 38 A), 
its two arms diverging forward on the blastoderm from a posterior 
area of proliferation (CD) in the region of the blastopore, whence 
also are proliferated forward two corresponding bands of mesoderm. 
Each mesoderm band soon becomes divided into four consecutive 
parts, which appear as four lobes on the surface (B). The germ 
bands themselves gradually become less divergent, and finally their 
anterior ends curve mesally and unite by a bridge between their 
anterior lobes (C). At the same time the rudiments of three pairs 
of appendages appear on the second, third, and fourth lobes, which 
are respectively the first antennae (B, D, lAnt), the second antennae 
{2 Ant), and the mandibles (Md). The first lobes (Pre) have no 
appendages, but they give rise to the compound eyes and the optic 
ganglia. There now appear in the ectoderm of the young embryo. 


Sollaud says, three transverse grooves which define the first seg- 
mentation (D). The most anterior groove runs between the first 
and second pairs of antennae, the next between the second antennae 
and the mandibles, and the third behind the mandibles. The body of 
the embryo is thus divided into an anterior prostomial head segment 

Fig. 38. — Early embryonic stages of a palaemonid crustacean, showing the 
development of the procephaUc lobes and the antennules from the unsegmented 
prostomial region, and the formation of four primary body somites. (From 
Sollaud, 1923.) A, Leander squilla Linn. B-F, L. serratiis Pennant. 

A, ventral surface of egg showing germinal disk and anterior proliferation 
of germ bands. B, early nauplius stage with first appearance of appendages. 
C, later stage with germ bands united anteriorly. D, nauplius stage, with 
caudal papilla (CdP) differentiated, but circle of ectodermal teloblasts (EcT) 
yet incomplete. E, older nauplius embryo with ventral groove (cf) in caudal 
papilla. F, metanauplius stage, with rudiments of first and second maxillae 
formed on posterior part of nauplius body before generation of teloblastic 
somites has begun. 

lAnt, first antenna; sAnt, second antenna; CdP, caudal papilla; EcT, ecto- 
dermal teloblasts ; GD, germinal disk ; /, //, first two somites ; Lni, labrum ; 
Aid, mandible; iMx, 2Mx, first and second maxillae; Pre, procephalic lobe; 
Prst, prostomium ; Tel, telson. 

(Prst) bearing the procephalic lobes and the first antennae, a second 
segment (/) bearing the second antennae, a third segment (//) bear- 
ing the mandibles, and a terminal unsegmented piece (CdP), which 
is the caudal papilla. The embryo is now in the nauplius stage. The 
first segment, bearing the optic lobes and first antennae, Sollaud 


claims, is the prostomium, the other two segments being the first and 
second true somites {I, II). (See also SoUaud, 1933.) 

The caudal papilla of the malacostracan embryo (fig. 38 D, E, 
CdP) projects from the blastoderm. In its distal part is a circle of 
undifferentiated cells, ectodermal (EcT) and mesodermal, which are 
the teloblasts that will generate the postnaupliar somites. Beyond 
the teloblasts is the region of the telson (Tel) containing the anus 
(An). In its development the caudal papilla bends forward (F) 
beneath the part of the embryo contained in the blastoderm. 

When the malacostracan embryo reaches the metanauplius stage 
there appear at the base of the caudal papilla the two maxillary somites 
and their appendages (fig. 38 F, iMx, 2Mx). In a study of the 
development of Hemimysis, Manton (1928) includes the two maxil- 
lary somites in the part of the body produced from the teloblasts. 
Sollaud (1923), however, asserts that in Leander both maxillary 
somites arise from the base of the caudal papilla before the beginning 
of activity in the teloblast, and that the first somite of the teloblastic 
series is that of the first maxillipeds. According to Sollaud, there- 
fore, the four somites of the metanauplius (F), namely, those of the 
second antennae, the mandibles, the first maxillae, and the second 
maxillae, are primary somites formed directly in the primitive em- 
bryonic body between the acronal prostomium and the caudal papilla. 
If so, it would seem to be more than a coincidence that the same 
number of primary somites occurs in Malacostraca, Xiphosurida, 
and Trilobita. 

In most of the entomostracan Crustacea the embryo hatches in the 
nauplius stage when only three pairs of appendages are present 
(fig. 4B). The trunk is not yet distinctly segmented, but it con- 
sists of three regions. The first region is a preoral cephalic lobe 
bearing a median eye, the first antennae (lAnt), and the labrum; 
the second carries anteriorly the second antennae {2 Ant) and the 
mandibles (Md), and includes posteriorly the area on which the first 
and second maxillae will be formed ; the third region is a terminal 
unsegmented lobe, the telson, at the base of which is the generative 
zone from which will be formed the teloblastic somites. The nauplius, 
therefore, represents an ontogenetic stage in which the body region of 
the four primary somites is present, though the appendages of the 
posterior two of these somites are as yet undeveloped. 

The crustacean nauplius has often been likened to the trochophore 
larva of the Polychaeta (fig. 4 A), and the two forms are comparable 
in so far as each represents an early stage of ontogenetic development. 
We cannot suppose, however, that the arthropods and the annelids 


are separately derived from an ancestral form represented by the 
polychaete trochophore, since the adult arthropods have too many 
features in common with adult annelids that are not yet present in 
the trochophore. The common ancestor of the two groups, therefore, 
is to be found in a much later stage of annelid development than that 
of the trochophore. The trochophore and the nauplius are specialized 
larval forms, adapted in their general shape and structure to a tem- 
porary pelagic life; but, since they represent an early stage of phylo- 
genetic development, and probably originated as larvae at an early 
phylogenetic period of evolution in their respective groups, they 
necessarily show primitive characters in their basic organization. 


The question of the number of segments that enters into the com- 
position of the arthropod "head" has been widely investigated and 
discussed, but with such lack of uniformity in the results as to lead 
to the suspicion that interpretation of the observed facts has been too 
much influenced by theoretical considerations. The writer believes 
that a more literal acceptance of the known facts of embryonic 
development in the case of the arthropod head will give a simpler 
and more satisfactory concept of the fundamental cephalic structure 
than that which has been current for several decades. 

In the first place, it should be understood that there is no specific 
"arthropod head." The cephalic structure is a variable combination 
of segments, and the number of cephalized segments may be quite 
different in different arthropod groups, or even within a single major 
group. The more complex types of head, such as occur in the Man- 
dibulata, include an anterior procephalic region bearing the labrum, 
the eyes, and two pairs of antennae, and a posterior gnathal region 
bearing the mandibles, the first and second maxillae, and in some 
forms the first maxillipeds or also the second maxillipeds. In the 
Trilobita the so-called "head" is a combination of at least four postoral 
somites with the prostomial acron, and the "prosoma" of the Chelic- 
erata is a similar composite structure, except that it contains six 'or 
eight somites. On the other hand, in many of the Crustacea, the true 
head is a primitive structure corresponding with the procephalic part 
of the head in other mandibulate groups. However, differences of 
opinion as to the number of somites involved in the head composition 
pertain chiefly to the procephalic region, since the segments of the 
gnathal region are usually distinct in the embryo, and are readily 
identified by their appendages. 


On the assumption that the Arthropoda and the Onychophora 
are derived from generalized annehds, the primary head of the 
onychophoran-arthropod ancestors must have been the prostomium. 
The prostomium, therefore, constitutes the archicephalon in the series 
of articulate animals. In the polychaete annelids the prostomium 
(%• 39 A, Prst) supports two pairs of sensory appendages, the 
tentacles {Tl) and the palpi {Pip), and often a median anterior 
tentacle, and bears dorsally the eyes and the nuchal organs, while 
between it and the first somite (/) is situated ventrally the mouth 
(Mth). The neural elements of the prostomium, probably including 
originally a median apical ganglion and several paired ganglia devel- 
oped in connection with the sensory organs (fig. 9 B), unite to form 
the composite suprastomodaeal nerve mass known as the brain, or 
archicerebrum (C, D, Br). 

The young arthropod embryo characteristically has at the anterior 
end of the body a large cephalic lobe (fig. 39 B, Acr). On this head 
lobe are developed the eyes, both simple and compound (£), the 
first antennae (lAnt), in some cases a pair of transient preantennal 
rudiments (Prnt), and the labrum (Lm). The neural elements of 
the embryonic head, which may include an anterior median ganglionic 
rudiment and as many as four paired lateral rudiments, soon unite 
to form the suprastomodaeal brain. The exact parallelism in structure 
and development between the cephalic lobe of the arthropod embryo 
and the prostomium of the polychaete worm (A) certainly suggests 
a morphological identity between the two organs. In neither is there 
ever any external mark of segmentation, or direct evidence of the 
confluence of more primitive segments. 

SoUaud (1923, 1933), from his study of the development of the 
crustacean Leander, contends that the embryonic head region (fig. 
38 D, E, Prst) on which are developed the procephalic (ocular) lobes 
(Pre) and the first antennae {lAnt) must represent the annelid 
prostomium, since the first intersegmental groove runs behind the 
first antennae, and there is no external evidence of segmentation 
before it. Moreover, in the procephalic nerve ganglia, he says, only 
a slight constriction occurs at an early stage between the ocular, or 
protocerebral, parts and the antennal, or deutocerebral, parts. SoUaud 
asserts, therefore, that there is no valid reason for the commonly 
accepted view that the first antennae are homodynamous with the 
following appendages in the sense that they are the appendages of a 
primarily postoral somite that has been secondarily incorporated with 
the prostomium. The first antennae of Leander, he shows, remain 
uniramous, while almost from the beginning the second antennae 


(D, E, 2 Ant) take on the biramous structure characteristic of the 
following somatic appendages. The postoral segment of the second 
antennae is thus, according to Sollaud's interpretation, the first true 
somite. The same view is strongly advocated by Holmgren (1916) 
and Hanstrom (1928) from a comparative study of the annelid and 
arthropod brain, but, as will be shown later, the evidence adduced by 
these authors from the brain structure must be qualified by facts 
of development. 

The principal ground for the generally accepted belief that the 
acronal region of the arthropod embryo contains one or more "cepha- 

FiG. 39. — ^Diagrams of cephalization in the Polychaeta and Arthropoda, show- 
ing the relation of the annelid prostomium to the arthropod head on the assump- 
tion that the first antennae are prostomial appendages. 

A, an adult polychaete with prostomial tentacles and palpi, first two somites 
united in the peristomium. B, an insect embrj'o in which the head (acron) is 
an archicephalon representing the annelid prostomium, and may bear two pairs 
of appendages. C, a theoretical protomandibulate arthropod, in which the head 
is a protocephalon (Prtc) composed of the acron and one somite. D, a 
chelicerate arthropod, in which the acron is extended laterally and dorsally 
over several somites united in the prosoma. 

Acr, acron (arthropod prostomium) ; lAnt, first antenna (acronal appen- 
dage) ; 2 Ant, second antenna (appendage of first somite) ; Chi, chelicera (equiva- 
lent to second antenna) ; E, lateral eye; I-VI, first six somites; Lm, labrum ; 
Md, mandible ; Mth, mouth ; iMx, 2Mx, first and second maxillae ; Pdp, pedi- 
palp ; Perst, peristomium ; Pip, palpus ; Prnt, preantenna ; Prst, prostomium ; 
Prtc, protocephalon; Tl, tentacle. 

lized somites" is the occurrence of temporary coelomic sacs in this 
region. However, it has not been shown that the presence of cavities 
in the cephalic mesoderm is necessarily indicative of somites, and it 
would seem that the burden of proof should be on the positive side 
of this question. 

The mesoderm bands of the annelids, as shown in an earlier part 
of this paper, extend forward in the sides of the body from their 
posterior centers of propagation. In the trochophore larva the meso- 
derm is arrested at the mouth, but in the later development of the 
worm the bands extend into the prostomium and may here contain a 
pair of coelomic cavities. While it is usually observed that the pro- 


stomial coeloni of the annelids is a continuation from the coelomic 
cavities of the first somite, it is claimed by Binard and Jeener (1928) 
that the prostomial cavities of the spionid Scolelepis fuliginosa belong 
to a distinct pair of mesodermal sacs associated with the palpi. In the 
Onychophora and Arthropoda the mesoderm likewise extends into 
the head region at the sides of, or before, the stomodaeum (fig. 41 A), 
and is usually excavated by a pair of well-developed coelomic sacs 
pertaining to the antennae (C, AntCS) ; but in the arthropods there 
may be formed also a pair of sacs pertaining to transitory preantennal 
appendages (fig. 42 B, PrntCS), and even a third pair in the labral 
region (D, LinCS). The position of the antennal sacs, as that of the 
antennal rudiments themselves, is somewhat variable in diliferent 
arthropods, both structures being in some cases postoral, in others 
adoral, and again preoral ; in the Onychophora the antennal sacs are 
decidedly preoral, though their posterior mesal ends embrace the 
stomodaeum and give rise to some of the stomodaeal muscles. The 
preantennal sacs are usually slightly preoral ; the labral sacs lie directly 
before the mouth. 

When we consider that the forwardly growing mesoderm bands, in 
their fullest development, should finally meet in front of the blasto- 
pore, it is evident that coelomic cavities formed in the cephalic region 
must assume adoral and preoral positions with their axes centering 
in the mouth (fig. 40 B). Being thus radial in position, the cephalic 
coelomic sacs cannot represent "somites" in the manner of the paired 
sacs lying posterior to the mouth, which are transversely opposed 
to each other. Hence, the assumption that these anterior sacs 
represent "cephalized somites" is inconsistent with the anatomical 
conditions that arise in the acronal region of the trunk. Moreover, 
as we have seen in a study of the annelids, the coelomic sacs them- 
selves do not determine metamerism ; the segmentation of the postoral 
parts of the mesoderm bands is secondary to metamerization of the 
primary somatic muscular system, and the coelomic cavities are later 
formed probably for physiological purposes. The coelomic sacs, there- 
fore, correspond with the somites in the segmented part of the trunk, 
but similar mesodermal cavities might be formed for the accumulation 
of waste products in an unsegmented region such as the prostomium. 
The usual absence of well-difYerentiated coelomic sacs in the annelid 
prostomium, and the fact that the fullest development of the head 
sacs is found in the higher arthropods indicate that the formation of 
cavities in the cephalic mesoderm is a secondary accompaniment of 
advancing organization in the prostomial lobe ; but the temporary 


nature of the head cavities might equally suggest that they are purely 
ontogenetic structures, as claimed by Faussek (1899, 1901), for 
coelomic cavities in general. 

The association of the antennal coelomic sacs with the antennae 
and the association of the preantennal sacs with preantennal appen- 
dicular rudiments suggest that in a primitive stage there may have 



Fig. 40. — Diagrams illustrating two theories of the fundamental structure of 
the Articulata. 

A, the theory of radial structure, based on a supposed origin of the articulates 
from a zoantharian polyp, according to which the coelomic sacs represent radial 
pouches of the enteron, and the nervous system a circumoral nerve ring, seg- 
mentation of the body being determined by the enteric pouches. 

B, the theory adopted in this paper, which assumes an origin of the articu- 
lates from a creeping wormlike ancestor, based on the facts that, though the 
mouth is subapical, the anus is terminal, and that in embryonic development 
segmentation precedes the formation of the coelomic sacs, which have no con- 
nection with the enteron ; the mesoderm, being teloblastic, grows forward, and, 
in its fullest development, may surround the mouth anteriorly, and thus give 
rise to a secondary radial symmetry in the prostomial region. 

Bpr, blastopore; cCom, cerebral commissure; CS, coelomic sac; Mth, mouth; 
NR, nerve ring; Prst, prostomium ; VNC, ventral nerve cord. 

been a pair of appendages in the labral region corresponding with 
the labral sacs. Some writers have contended that the labrum itself 
represents a pair of united appendages, but since the labrum is imme- 
diately preoral, a pair of "labral" appendages in an annelid would arise 
from the base of the prostomiuin. Perhaps, by a long stretch of the 
imagination, the labral sacs might better be correlated with a hypo- 
thetical pair of primitive apical prostomial tentacles (fig. 40 B), 


possibly represented by the median tentacle of certain Polychaeta 
(fig. 13 C), which, having a double nerve root in the brain (fig. 45 B, 
C, iTlNv), might be supposed to have had itself a double origin. 
However, the possibility of the median polychaete tentacle having 
been formed by the union of a pair of apical tentacles is denied by 
Binard and Jeener (1928). 

The theory here proposed to explain the occurrence of coelomic 
sacs in the prostomial region of the articulate animals has no relation 
whatever to the theory of Sedgwick (1884), Lameere (1926), and 
Binard and Jeener (1928) that the annelids and arthropods are 
derived from a coelenterate polyp form, and therefore have funda- 
mentally a radial organization (fig, 40 A). A radial structure secon- 
darily aflfects the anterior end of the articulate trunk because of the 
subapical position of the mouth (B) ; but the terminal position of the 
anus creates a quite dififerent structure at the posterior end. 

The term acron (Janet, 1899) is frequently used by students of 
arthropod embryology to designate the apical part of the arthropod 
head that lies anterior to the first true somite ; its exact application, 
therefore, differs according to each writer's interpretation of the 
head segmentation. Janet defined the acron as the preantennal part 
of the head. As the term is used in the present paper, the arthropod 
acron is equivalent to the annelid prostomium, and is represented in 
the arthropod embryo by the cephalic lobe (or lobes) bearing the 
eyes, the labrum, the preantennae, and the first antennae. The pro- 
stomium is primarily the anterior part of the trunk not invaded by 
the blastopore (fig. 6 D, Prst) ; the median part of the arthropod 
acron is always preoral, but its lateral parts may lap backward and 
extend even a considerable distance behind the mouth. The telson 
at the posterior end of the trunk is not morphologically equivalent to 
the acron. It is traversed by the alimentary canal, and has the anus 
at its extremity ; it does not contain coelomic sacs, but its represen- 
tative in the annelids, the so-called pygidium, may support a pair of 
tentaclelike appendages. 

The principal reasons for regarding the oculo-antennal region of 
the arthropod head, here defined as the acron, as representing a 
primarily unsegmented archicephalon corresponding with the annelid 
prostomium may be summarized as follows : ( i ) There is never any 
external division of the acronal region into segmental areas; (2) there 
is no specific evidence of the cephalization of primarily postoral 
somites, except in the case of the tritocerebral somite; (3) the embry- 
onic coelomic sacs of the first antennae, the preantennae, and the 
labrum are formed directly where they occur in the cephalic meso- 


derm, and give no evidence of having been drawn forward from 
behind the mouth ; (4) coelomic sacs of the acronal region, so 
far as known, are best developed in the higher arthropods, and thus 
do not appear to be primitive structures; (5) the protocerebral and 
deutocerebral parts of the brain are always connected by preoral 
commissures, the only postoral cerebral commissure being that of the 
cephalized tritocerebral ganglia; (6) the mouth and labrum are in- 
nervated from the tritocerebral ganglia, which would not likely be 
the case if several other postoral ganglia preceded the tritocerebral 
ganglia; (7) paired appendages, sense organs, and primarily discrete 
nerve centers pertain both to the annelid prostomium and to the 
arthropod acron ; (8) the first antennae of the arthropods never have 

Lm LmMsd 


Storri .£^, 


Fig. 41. — Development of the procephalic mesoderm in Orthoptera. (A, B 
from Roonwal, 1937; C from Wiesmann, 1926.) 

A, horizontal section of anterior end of 52-hour embryo of Locusta migratoria 
Linn, showing cephalic mesoderm extending to labrum anterior to stoniodaeum. 
B, same of 563-hour embryo, with coelomic cavities in labral mesoderm. C, re- 
construction of head of embryo of Carausius tnorosus Brunner, lateral view, 
with developing antennal coelom, and mesoderm extending into clypeolabral 

Am, amnion; Ant, antenna; AntCS, antennal coelomic sac; AntMsd, antennal 
mesoderm ; Br, brain ; Lm, labrum ; LmCS, coelomic sac of labrum ; LmMsd, 
labral mesoderm ; Msd, mesoderm ; Pre, cephalic lobe ; Stom, stomodaeum. 

the structure or musculature of the following appendages ; in the 
Crustacea they are never truly biramous. 

A brief review of the facts now known concerning the develop- 
ment of the procephalic mesoderm and nervous system of the arthro- 
pods will show that the facts are not inconsistent with the idea that 
both coelomic sacs and multiple nerve centers may be formed directly 
in the otherwise unsegmented acronal region, and that the phenomena 
of embryonic development pertaining to the head are most easily 
understood if they are taken approximately at their face value for 
phylogenetic recapitulations. 

The cephalic mesoderm of the arthropods is usually continuous 
with the mesoderm bands of the anterior somites. In a 52-hour 
embryo of Locusta, Roonwal (1937) says, "it is seen that a pair of 


mesoderm bands extends upward from the junction of the head-lobe 
with the trunk and meet over the stomodaeum" (fig. 41 A, Msd). 
The same is true of Caraushis (B), as shown by Wiesmann (1926), 
but in the crustacean Heinimysis, according to Manton (1928), a 
part of the preoral mesoderm has an independent origin from the 
germ band. 

Among the Chelicerata the cephahc mesoderm is less developed 
or differentiated than in the Mandibulata. In Limulus longispina, as 
described by Kishinouye (1893), the first pair of coelomic sacs in 
the embryo occupies both the cephalic lobe and the cheliceral somite. 
Later these sacs become partially divided by an incomplete septum 
into a pair of cephalic sacs and a pair of cheliceral sacs, but the latter 
soon disappear. In the scorpion, according to Brauer (1895), the 
cephalic coelom is an extension of the coelomic cavities of the chelic- 
eral somite, and is never shut off from the latter in a pair of specific 
head sacs. Likewise in the Pedipalpida {Thclyphomis) Schimkewitsch 
(1906) says the coelomic sacs of the head segment are continuous 
with those of the cheliceral segment. Kishinouye (1894) finds, on 
the other hand, in the Araneida (Lycosa and Agclena) a pair of 
coelomic sacs in the cephalic lobe that are entirely separate from the 
sacs of the cheliceral somite. The cephalic sacs are later divided each 
into two parts ; the ventral sections disappear, the dorsal sections 
elongate upward and form between them the cephalic aorta. 

In the Mandibulata coelomic cavities associated with the first 
antennae are of common occurrence in the cephalic mesoderm. A 
diverticulum from each antennal sac extends into the corresponding 
antenna (fig. 41 C, AntCS) and gives rise to the antennal muscula- 
ture. The inner dorsal parts of the sacs, as observed by the majority 
of investigators (see Wiesmann, 1926, Roonwal, 1937), grow mesally 
into the space between the stomodaeum and the brain, where they 
extend anteriorly and posteriorly and form the cephalic part of the 
aorta, including the anterior end of the tubular aorta proper, and an 
open distributing section that extends from beneath the brain to the 
clypeal region. The cephalic aorta of the crustacean Hemimysis, 
however, is said by Manton (1928) to be a product of the preantennal 

The presence of preantennal coelomic sacs associated with small 
evanescent rudiments of preantennal appendages (figs. 42 A, 43 A, 
Prnt) is recorded by Heymons (1901) in Scolopcndra (fig. 42 B, 
PrntCS), and by Wiesmann (1926) in Carausius (F, PrntCS), and 
the occurrence of coelomic cavities in the preantennal mesoderm of 
Hemimysis is reported by Manton (1928), though vestiges of pre- 


antennal appendages are not known in the Crustacea. In the diplopod 
Platyrrhacus ainanros, Pflugfelder (1932a) shows that a pair of 
coelomic sacs is formed in the cephaHc lobes of the embryo in con- 



Fig. 42. — Embryonic appendages and coelomic sacs of the procephalic region 
of an insect, a chilopod, and a crustacean. 

A, head (protocephalon) and two following somites of young embryo of 
Carausius morosus Brunner, ventral view (from Wiesmann, 1926). B, length- 
wise section through cephalic appendages and coelomic sacs of embryo of 
Scolopeiidra (from Heymons, 1901). C, lengthwise section through a coelomic 
sac of the embryonic labral rudiment of Carausiiis (from Wiesmann, 1926). 

D, cross-section of same through labral coelomic sacs (from Wiesmann, 1926). 

E, cross-section through preantennular coelomic sacs of embryo of Hemimysis 
lamornac (from Manton, 1928). F, cross-section through preantennal coelomic 
sacs of embryo of Carausius (from Wiesmann, 1926). 

Am, amnion; Ant, lAnt, first antenna; Lm, labrum ; LmCS, labral coelomic 
sac; Md, mandible; iMx, 2Mx, first and second maxillae; PntCS, postantennal 
coelomic sac; Prnt, preantemia (preantennule) ; PrntCS, preantennal (pre- 
antennulary) coelomic sac; Prtc, protocephalon (acron and first somite) ; Stom, 

nection with the protocerebral lobes of the brain (fig. 44 D, Per), 
and a second pair in connection with the deutocerebral lobes (Dcr). 
Hence, if there is any necessary homology between the cavities of 
the cephalic mesoderm in different arthropods, the "protocerebral" 


sacs of Platyrrliacus should represent the preantennal sacs of Scolo- 
pendra, Carausius, and Hemhnysis, though there are in the diplopod, 
as in the crustacean, no corresponding appendage rudiments. While, 
in most cases observed, the preantennal mesoderm is a part of the 
general mesoderm, the preantennal mesoderm of Henmnysis is said 
by Manton (1928) to have an independent origin from the germ 
band just behind the optic lobes. When the arms of the V-shaped 
germ band of Heiiiimysis later come together, the preantennal meso- 
derm rudiments are approximated immediately before the mouth. In 
their growth, Manton says, they extend posteriorly and embrace the 
lateral and dorsal walls of the stomodaeum, their cavities entirely 
disappear, and their walls give rise to a part of the stomodaeal 
("stomach") muscles, and to the cephalic aorta. 

Coelomic sacs of the labral region of the embryonic head were 
first described by Wiesmann (1926) in the stick insect, Carausius 
morosus, and have since been observed by Mellanby (1936) in the 
hemipteron Rhodnius, and by Roonwal (1937) in a grasshopper, 
Locusta migratoria. Pflugfelder (1932a) describes in the diplopod 
Platyrrhacus a pair of mesodermal cavities in the "clypeus" (fig. 44 D, 
Clp), but since these cavities lie immediately before the mouth, they 
evidently correspond with those called "labral" in the insects. In both 
Locusta (fig. 41 A) and Carausius (C) the head mesoderm extends 
into the labrum (LinMsd) anterior to the stomodaeum (Stom), and 
the cavities formed in it are thus literally preoral in position (fig. 42 C, 
LmCS) ; the mesal walls of the labral sacs of Carausius are united 
before the mouth (D). In Locusta, Roonwal says, the labral and 
stomodaeal mesoderm is loosely continuous prior to the appearance of 
the labral cavities (fig. 41 A), but when the sacs are formed the 
latter are independent structures (B, LmCS). After the disap- 
pearance of the cavities the coelomic cells remain as two bodies of 
mesoderm that suggest similar mesodermal masses found in the 
labrum of certain other insects in which corresponding cavities are 
not known to occur. 

The definitive brain of the mandibulate arthropods consists of an 
anterior bilobed part, including the protocerebrum and the deuto- 
cerebrum, which innervate respectively the eyes and the first antennae, 
and of a pair of posterior lobes, the tritocerebrum, which innervate 
the second antennae when these appendages are present. The proto- 
deutocerebral lobes are always united above the stomodaeum, and 
thus appear to belong to the prostomial part of the head ; the trito- 
cerebral lobes, on the other hand, are unquestionably derived from 
the postoral somite of the second antennae, and are connected by a 


postoral commissure. In many cases the dorsal lobes are developed 
in the embryo from a single pair of generative centers in the ecto- 
derm, just as are the corresponding lobes of the brain in the Ony- 
chophora and in some of the Annelida. Considering, however, that 
the annelid brain, as shown in the larva of Lopadorhynchus, has 
probably taken its origin from a number of discrete prostomial 
ganglionic centers corresponding with the sensory organs of the 
prostomium, we should expect that a more primitive condition in the 
arthropods would show that the definitive brain of these animals is 
likewise a composite structure formed by the union of primarily 

Clp^ ^^Prnt iST> 2g7> 3g-p ?^P Lni 

~ — iMx 


Ant ^'^-yX )E<^^^ X .Ant 

A -' 

Fig. 43. — Embryonic development of the cephalic appendages and nerve 
ganglia of a chilopod, Scolopcndra. (From Heymons, 1901.) 

A, anterior end of germ band with rudiments of appendages, including pre- 
antennal, antennal, mandibular, and maxillary lobes, but no rudiments of post- 
antennal (intercalary) appendages, though postantennal (tritocerebral) somite 
marked by a pair of ganglia (IGng). B, same, later stage (antenna removed 
on left), showing ganglionic pits (gp) of ectoderm from which ganglia are 

Ant, antenna ; Clp, clypeus ; igp, generative pit of optic ganglion ; ^gp, pit 
of protocerebral ganglion ; jgp, pit of preantennal ganglion ; 4gp, pit of antennal 
ganglion ; 5gp, pit of tritocerebral ganglion ; 6gp-8gp, pits of mandibular and 
maxillary ganglia ; iGng, tritocerebral ganglion ; iL, first leg ; Lm, labrum ; 
Md, mandible ; iMx, sMx, first and second maxillae ; Prnt, preantenna. 

separate ganglia. Various studies on the development of the arthropod 
brain, in fact, demonstrate its diffuse origin. 

The best-known example of the development of the arthropod brain 
from diffuse ganglionic centers is that described by Heymons (1901) 
in Scolopendra. The embryonic cephalic appendages of Scolopendra 
that correspond with cerebral rudiments include the persistent an- 
tennae (fig. 43 A, Ant) and a pair of transient preantennae {Prnt), 
appendages of the postantennal "intercalary," or tritocerebral, somite 
being absent. The definitive brain of Scolopendra, according to Hey- 
mons, is formed by the coalescence of an anterior unpaired ganglionic 
rudiment and five paired rudiments. The unpaired rudiment arises 


directly from the ectoderm of the clypeal region of the cephalic lobes 
(fig. 43 B, dp), and appears before any of the appendages except 
the antennae are formed. The paired rudiments are groups of gan- 
glionic cells proliferated from the inner ends of small ectodermal 
pits {igpsgp). The first of these rudiments to be formed {2gp) 
lie at the extremities of the median rudiment, and their cells become 
closely associated with the latter to produce a cellular mass that be- 
comes the lamina dorsalis of the definitive protocerebrum. Laterad 
of these rudiments are formed a pair of pits {igp) that furnish 
principally the cells of the definitive frontal lobes of the brain, and 
later when the eyes appear give rise also apparently to the small optic 
lobes. Following the lateral rudiments of the lamina dorsalis on each 
side are formed in series three other cephalic pits, which generate 
respectively the primitive ganglionic centers of the preantennae 
(Sgp), of the antennae (4gp), and of the appendageless tritocerebral 
somite (sgp). The two series of neurogenic pits are continued pos- 
teriorly on the mandibular, the maxillary, and the leg somites. 

Heymons regards the median unpaired brain rudiment as the equiv- 
alent of the annelid "archicerebrum," but it would seem rather to 
correspond with the ganglion of the apical plate of the polychaete 
larva. The two paired rudiments that combine with the median rudi- 
ment to form the definitive protocerebrum he refers also to the 
"acronal," or prostomial, part of the head, but the preantennal, 
antennal, and tritocerebral rudiments he claims represent postoral 
somites. The preantennal ganglia constitute at first a connection 
between the protocerebrum and the deutocerebrum, but later they 
merge so completely into the brain that their identity is lost in the 
definitive brain structure. The deutocerebral lobes formed of the 
antennal ganglia lie primarily behind the protocerebrum, but with 
the forward migration of the antennae they become transposed to a 
position anterior to the protocephalon and take on a conical form 
with the antennal nerves issuing from their distal ends. The trito- 
cerebral lobes lie beneath the deutocerebral lobes and are continuous 
with the stomodaeal connectives. 

The claim of Heymons that the preantennal and antennal ganglia 
represent postoral somites is not substantiated by any external evi- 
dence of segmentation in the corresponding cephalic region of the 
scolopendrid embryo, and as represented in Heymons' figure (fig. 
43 B) these ganglia appear to be actually preoral in position. In 
none of the arthropods do the true cerebral ganglia have postoral 
commissures, but the preoral position of their commissures in the 
brain mass, Heymons says, is to be explained ontogenetically by the 


fact that the commissures are not formed until after the respective 
ganglia have taken a preoral position. This proposed explanation, 
however, is merely the statement of a fact that can as well be taken as 
evidence that the ganglia themselves are morphologically preoral. 

In the Diplopoda the embryonic rudiments of the nervous system 
appear to be completely double, for no median ganglionic center has 
been observed corresponding with the "archicerebral" rudiment 
described by Heymons in Scolopendra. Preantennal appendages are 
absent so far as known, and the tritocerebral somite always lacks 
appendages, as in the Chilopoda. According to Robinson (1907) the 
nervous system of a i6-day-old embryo of Archispirostreptus consists 
of a pair of trilobed "archicerebral" rudiments situated before the 
mouth (fig. 44 E, Arc), and of two ganglionated nerve cords pro- 
ceeding posteriorly from the latter around the stomodaeum. The 
first ganglia of the cords (AnfGng), which are distinctly postoral, 
Robinson claims are the antennal ganglia. The next pair, she says, 
are the ganglia of the tritocerebral somite (TcrGiig), which has no 
appendages, and the next pair {MdGng) belong to the mandibles. 
At a later stage (F), just before hatching, the "antennal ganglia" 
{AntGng) , to which Robinson says the tritocerebral ganglia are now 
joined, still lie behind the mouth and are approximated to the median 
line. It is very surprising, however, that antennal ganglia should be 
postoral at such a late stage of development, and not yet united with 
the protocerebrum, so much so, in fact, that the relation of these 
alleged "antennal" ganglia to the antennae becomes questionable. 
Robinson gives no evidence of any nerve connection between the 
ganglia and the antennae {Ant) ; hence we might suspect that the 
antennae are innervated from the posterior ganglia of the "archi- 
cerebral" groups {Arc), and that the first postoral ganglia are the 
tritocerebral ganglia. 

Heymons (1897) gives a brief description of the embryo of 
Glomeris (fig. 44 C), in which the antennae {Ant) appear as adoral 
appendages of the cephalic lobes {Pre), whence presumably they 
derive their innervation. 

A more detailed account of the cephalic and cerebral segmentation 
of a diplopod is given by Pflugfelder (1932 a) for Platyrrhacns 
amauros, but it only adds to the difficulties of understanding the 
development and morphology of the diplopod head. According to 
Pflugfelder, the protocerebral and deutocerebral elements of the brain 
appear on the surface of the young embryo of Platyrrhacns as a 
single pair of preoral cephalic lobes (fig. 44 A, Pre). Just behind 
the mouth is the somite of the antennae {Ant), and the latter is 



VOL. 97 

BiW^'----::^0 Per 


1^'M^yM-:0^.^ — iL 

Are ^ ^ .._^ 

CIp Ant Md MxPmx 
Lm.^ ,,.;^^^-'^^^v.^,,..Arc 
^,-KC^'" <"rT7:-^- Ant 

;,.•- F 

Fig. 44. — Embryonic segmentation of the head of Diplopoda as interpreted 
by different investigators. 

A, germ band of Platyrrliacus anumros Attems before invagination (from 
Pflugfelder, 1932 a). B, longitudinal section of head of embryo of Platyrrhacus 
amauros shortly before hatching (adapted from Pflugfelder, 1932 a). C, young 
embryo of Glomcris (from Heymons, 1897). D, longitudinal section of germ 
band of Platyrrhacus amanros just before invagination, showing preoral coelomic 
sac of clypeal region and sacs of four postoral somites (from Pflugfelder, 
1932 a). E, embryo of Archispirostrcptus sp., about four days before hatching 
(from Robinson, 1907). F, same, one day before hatching (from Robinson, 

Ant, antenna; AntGng, antennal ganglion; AntNz', antennal nerve; Arc, 
archicerebrum ; CIp, clypeus (labrum) ; Dcr, deutocerebrum ; Gch, gnathochi- 
larium ; iL, first leg ; Md, mandible ; MdGng, mandibular ganglion ; Mx, 
maxilla ; iMx, sMx, first and second maxillae ; Per, protocerebrum ; Pmx, 
postmaxillary appendage, or somite ; Pre, procephalic lobe ; TcrGng, tritocere- 
bral ganglion; VNC, ventral nerve cord. 


followed directly by the mandibular (Md) and two maxillary somites 
(Mx, Pmx), there being no evidence of a tritocerebral somite. In 
sections the cephalic lobes show internally a distinct division into a 
protocerebral rudiment (D, Per) and a deutocerebral rudiment (Dcr), 
each later accompanied by a coelomic sac. It would seem to be in- 
ferred from Pflugfelder's description, though not so stated, that the 
primary antennal ganglia lie in the postoral "antennal somite" (Ant), 
and yet he says, "das Deutocerebrum tritt sehr f riih mit den Antennen 
in Verbindung durch den Antennennerv," and he clearly shows the 
antennal nerve connection with the preoral deutocerebrum (B, 


Fig. 44 G. — Germ band of a diplopod, Archispirostreptus gigas Peters, show- 
ing rudiments of appendages and ganglia. (From Silvestri, 1933.) 

An. anus; Ant, antenna; AntGiig, antennal ganglion; iL, first leg; Md, mandi- 
ble ; MdGng, mandibular ganglion ; Mth, mouth ; Per, protocerebrum ; TcrGng, 
tritocerebral ganglion. 

AntNv). The anatomical evidence here would seem to show that 
the true morphological relations of the antennae are with the deuto- 
cerebral ganglia, and we can only suppose, therefore, as in the case 
of Robinson's account of Archispirostreptus, that the postoral so- 
called "antennal" ganglia are the tritocerebral ganglia. In any event, 
the implication from Pflugfelder's statements that the antennae are 
appendages of a postoral somite, but are finally innervated from the 
preoral deutocerebrum gives the impression that there is some error 

The interpretation of the anterior cephalic ganglia of the diplopod 
embryo given by Silvestri (1933), illustrated in Archispirostreptus 
gigas (fig. 44 G), unquestionably presents the most reasonable view 


that can be taken concerning the homologies of the ganglionic rudi- 
ments, since it disposes of the latter in a manner entirely consistent 
with the evident facts in other arthropods. According to Silvestri the 
ganglia of the antennae (AntGng) are neural masses situated mesad 
of the antennal bases, and the first pair of postoral ganglia (TcrGng) 
are the tritocerebral ganglia (ganglia of the intercalary somite). It 
should be observed that the antennal ganglia, as shown by Silvestri, 
have a preoral position and are not separated from the protocerebral 
lobes of the brain (Per). 

Among the higher arthropods the more primitive stages in the brain 
development are generally not shown in embryonic recapitulation, for 
the proto-deutocerebral centers are usually proliferated from the ecto- 
derm as a unified ganglionic cell mass, just as in the Onychophora 
and in many of the Annelida. It is observed by Baden (1936) and 
by Roonwal (1937), however, that the brain of the grasshopper 
(Melanoplus, Locusta) is formed from five pairs of ganglionic 
centers, three of which give rise to the protocerebrum and the optic 
lobes, and the other two to the deutocerebrum and the tritocerebrum, 
respectively. On the other hand, Nelson (191 5) finds that in the 
honey bee the lateral surfaces of the primarily undivided cephalic 
lobes of the embryo become directly differentiated into three areas 
from which are proliferated the neural centers of the protocerebrum, 
the deutocerebrum, and the tritocerebrum. 

In view of the well-authenticated examples of a diffuse origin of 
the cerebral ganglionic centers in the arthropods, the theory of 
Holmgren (1916) and of Hanstrom (1928) that the protocerebrum 
and the deutocerebrum are secondarily differentiated parts of a primi- 
tive, undivided archicerebrum does not appear to be substantiated by 
the facts of embryogeny. However, since the definitive brain is 
evidently a conglomerate of primitively separate ganglionic centers 
in the Annelida as well as in the Arthropoda, the general contention 
of these authors is not invalidated, namely, that both the protocerebral 
and the deutocerebral parts of the arthropod brain belong to the 
preoral prostomial region of the head, and, therefore, together 
represent the annelid archicerebrum. 

The concept that all coelomic sacs and corresponding nerve centers 
represent postoral somites seemed reasonable enough, as applied to 
the arthropod head, when only antennal sacs were known; it was 
somewhat stretched, though still acceptable, when preantennal sacs 
were discovered ; but now that we must add a third pair of cephalic 
sacs lying directly before the mouth in the labral region it begins to 
look farfetched. The theory here proposed, illustrated at D of 


figure 45, accepts the embryonic facts more literally. It assumes that 
the archicephalic nervous system of the arthropods, as that of the 
annelids, has been built up from groups of ganglionic cells centering 

2Tl(Prnt) ^'^^ ^?9^^ 
Pip (Ant) 


(Pnt) — c= 

aoCom / 2pj^^ 


SGng- Mth 
- IGng- 



2TINV 1 


An 'T'^^ 

Fig. 45. — Suggestions of homologies between the prostomium of the annelids 
and the acronal region of the arthropod head. 

A, diagram of the anterior segments of a theoretical "lobopod" annelid, with 
elemental ganglia corresponding with those of the arthropod cerebrum dis- 
tributed on a preoral commissural arch of the nerve cords, and the potential 
number of coelomic sacs of the arthropod acron shown in their possible rela- 
tion to the prostomial appendages. B, diagram of the fundamental structure of 
the prostomial nervous system of a polychaete annelid (from Binard and Jeener, 
1928). C, reconstructed frontal section of dorsal fibrillar mass of brain and 
nerves arising from it in a sedentary polychaete, Sabellaria spinulosa Leuckart 
(from Binard and Jeener, 1928). D, analytical diagram of the relation of the 
coelomic sacs of an arthropod to the central nerve ganglia and the associated 

acConi, anterior cerebral commissure; Acr, acron (prostomium); An, anus; 
Aiit, antenna; ApGng, apical ganglion; Br, brain; CS, coelomic sac; cv, ventral 
fibrillar mass of brain ; E, lateral eye ; iGng, protocerebral ganglion ; 2Gng, 
preantennal ganglion ; sGng, antennal ganglion ; /, //, first and second somites ; 
IGng, first somatic (tritocerebral) ganglion; mfd, dorsal fibrillar mass of 
brain ; Msd, mesoderm ; Mth, mouth ; Pip, palpus ; PlpNv, palpal nerve ; Pnt, 
postantennal appendage ; Pmt, preantenna ; Prst, prostomium ; rd, dorsal root 
of stomodaeal connective ; SoeGng, suboesophageal ganglion ; StCon, stomodaeal 
connective ; Tel, telson ; iTl, first tentacle ; 2TI, second tentacle ; iTlNv, nerve 
roots of median tentacle ; 2TINV, nerve of second tentacle ; VNC, ventral nerve 
cord ; ZG, zone of growth ; i, labral coelomic sac ; 2, preantennal coelomic sac ; 
3, antennal coelomic sac. 

upon a fibrous commissural tract arched forward around the mouth 
and continuous posteriorly with the ventral nerve cords of the somatic 
system. The primary cephalic ganglia included a median anterior 


ganglion, paired protocerebral and optic ganglia, paired preantennal 
ganglia, and paired first antennal ganglia. That these ganglia belong 
to the preoral acron (Acr) is shown by the fact that the paired 
ganglia are always connected by preoral commissures. The cephalic 
mesoderm extends forward from the somatic mesoderm bands, and, 
in its fullest development, surrounds the mouth anteriorly ; it may 
become excavated by cavities corresponding with the first antennae 
(j), the preantennae (<?), and the labrum (/). The development of 
the prostomial nerve ganglia and mesodermal cavities is determined 
probably in all cases by external structures (appendages or sense 
organs), but the acronal neuromeres and coelomic sacs, because of 
their radial position around the mouth, cannot have the same relation 
in the body structure as have their postoral counterparts that are 
.transversely opposed to each other. For a like reason there is no 
prostomial metamerism of the muscular system. The same concept 
may be applied to the preoral lobe of the polychaete annelids (A), 
assuming that potentially the annelid prostomium might have a full 
quota of coelomic sacs corresponding with its appendages, which actu- 
ally it does not have. The tritocerebral somite of the arthropods thus 
represents the first postoral somite of the annelids. The tritocerebral 
ganglia are secondarily united with the preoral cerebrum in the Ony- 
chophora and in most of the Arthropoda, and always have a postoral 
commissure; the corresponding appendages are the jaws of the Ony- 
chophora, the chelicerae of the Chelicerata, and the second antennae 
of the Mandibulata. 

The definitive arthropod brain more closely resembles the brain of 
the Polychaeta than that of the Onychophora. Its principal part is 
the protocerebrum, formed of a median apical ganglion and the first 
pair of lateral ganglia, with which are connected the optic ganglia. 
The preantennal ganglia lose their individuality in the general cerebral 
mass. The antennal ganglia form the deutocerebral lobes, but the 
latter take a forward position beneath the protocerebrum, with the 
result that, in the definitive condition, the antennal nerves arise 
anteriorly belozv the optic lobes. In the Onychophora, on the other 
hand, though the antennae are anterior, the brain maintains a hori- 
zontal position (fig. 25 A, C) with the antennal commissure behind 
the optic region, and the antennal tracts (AntT) traverse the dorsal 
part of the brain above the optic lobes. The tritocerebral ganglia are 
united with the primary cerebrum in the Onychophora and in nearly 
all the Arthropoda, but the union would seem to have taken place 
separately in the two groups, since in some of the lower Crustacea 
the corresponding centers are independent ganglia on the nerve cords, 
as they are in most of the Annelida. 



The prostomial acron does not constitute the definitive head of any 
known arthropod ; there is always added to the acron at least one 
postoral somite, and generally the definitive head includes from four 
to six somites. A head composed of the acron and one somite, how- 
ever, recurs so frequently, either in the adult stage or in ontogenetic 
development, as to suggest that a simple head structure of this kind 
(%• 39 Q Prtc) represents the earliest stage in the evolution of the 
more complex types of arthropod head. It may hence be termed the 
protocephalon. The best example of a functional protocephalon is to 
be seen in the anostracan Branchiopoda (fig. 50 A), in which the 
definitive head is a large cephalic lobe {Prtc) bearing the eyes, both 
pairs of antennae, and the labrum. The protocephalon is unquestion- 
ably the primitive head of all the mandibulate arthropods. There is 
no direct evidence, however, that it ever occurred as a specific stage 
in the evolution of the Trilobita or the Chelicerata, and hence, in the 
ancestors of these groups, and in the protarthropods generally, the 
primitive head may have been merely the prostomial acron. 

Crampton (1928) applies the term "archicephalon" to a supposed 
stage in the cephalic evolution of the arthropods when the head con- 
sisted of the procephalic region and the mandibular somite. That 
such a stage occurred relatively late in the phylogenetic history of 
the head, however, is clearly shown in the ontogeny of the Mandibu- 
lata, in which the primitive embryonic head is always a cephalic lobe 
bearing the first antennae and usually including the second antennal 
somite, while the gnathal somites are still a part of the body region. 
Antedating this protocephalic stage, however, there must theoretically 
have been a truly primitive stage when there was no head structure 
other than the prostomium. The prostomium, therefore, which be- 
comes the acronal region of the definitive head, is the only stage in 
the evolution of the arthropod head that might properly be termed 
the "archicephalon." 

The trilobite branch of the protarthropods is characterized by a 
lateral expansion of the body, produced by an extension of the lateral 
margins of the tergal plates into long flat lobes (fig. 36 E, 48 D) . The 
dorsal surface of the body thus presents a median elevated area 
(rhachis) accommodating the alimentary canal, and broad depressed 
lateral areas (pleurae). On the under surface the true venter (fig. 
48 D, V) is the area between the leg bases, the areas laterad of the 
legs being the ventral doublure {dhl) of the dorsum. The appen- 
dages bear long coxal epipodites (Eppd) supporting branchial 
lamellae or filaments. 


The so-called "head" of an adult trilobite (fig. 36 H, H), as we 
have seen, represents the 5-segmented body of the larva (A), the 
"body" segments of the adult being formed secondarily of a series 
of teloblastic somites generated from a subterminal zone of growth 
(ZG). The very young larva (fig. 46 A) presents a broad anterior 
acronal region (Acr), and a postacronal region in which are already 
differentiated the elevated median glabella (gib), which is the cephalic 
part of the rhachis, and the broad lateral areas (fg) that become the 
fixed cheeks of the adult (E). When the glabellar impressions appear 
(B, C) the glabella is cut into five consecutive divisions, but it is 
evident that the first division, or frontal lobe (C, frl), is derived 
from the acron, and that the following four divisions represent the 
first four postacronal somites {I -IV). With successive stages of 
development (B, C, D), the lateral wings of the acron {Ig) extend 
posteriorly along the sides of the somites and eventually form the 
so-called free cheeks of the adult' (E, Ig), on which are located the 
compound eyes (E). The cephalic segmentation of the trilobite larva, 
therefore, may be represented as at I of figure 46, in which the 
intersegmental lines {is-/j.s) are theoretically extended to the lateral 
margins of the body. A median dorsal ocellus, when present, is 
always situated on the glabella, but since it must belong to the acron, 
it is placed on the frontal lobe in the diagram (I, dO). 

In the mature trilobite head of typical structure (fig. 46 E), the 
preocular part of the acronal suture (I, is) has disappeared, but the 
postocular parts become the posterior parts of the sutures known as 
the "facial sutures" (fsp), the preocular parts of which (fsa) are 
probably secondary lines of cleavage developed to facilitate moulting. 
In some forms the facial sutures end on the lateral margins of the 
head ; in others they go to the posterior margin (E), and in such cases 
the genal spines are continuations of the free cheeks. On the ventral 
surface of the head (F) the acronal surface is broadly inflected to 
form the doublure (dbl), which carries the labrum (Lm), or "hypo- 
stome," on its preoral margin. The probable dorsal segmentation of 
the adult trilobite head, therefore, may be represented diagram- 
matically as shown at J of figure 46. The acron (Acr) clearly forms 
an extensive part of the mature cephalic structure, since it must in- 
clude the frontal lobe of the glabella (frl), the free cheeks (Ig) with 
the compound eyes (E), and the doublure (F, dbl) with the labrum 
(Lm). Furthermore, since the dorsal ocellus often occurs far back 
on the glabella (J, dO), we must assume that it is contained in a 
median, tongue of the frontal lobe extended posteriorly into the 
glabellar somites, because the simple eyes as well as the compound 
eyes always belong to the acronal segment. 


The antennal appendages of the trilobites, judging from their 
filamentous form in contrast with the segmented structure of the 
following appendages, evidently represent the first antennae (anten- 

FiG. 46. — Segmentation of the trilobite "head," or prosoma. 

A-D, four consecutive stages in the larval development of Blainia grcgaria 
Walcott, showing gradual posterior extension of lateral wings (free cheeks) 
of acron against sides of anterior somites, and division of glabella (C) into four 
segmental areas behind frontal lobe of acron (from Lalicker, 1935). E, diagram 
of typical trilobite head, dorsal surface. F, diagram of ventral surface of trilobite 
head, showing labrum attached to margin of doublure. G, labrum of Pacdcumias 
transitans Walcott, example of a stalked labrum (from Walcott, 1910). H, 
head of Holotrachelus pwicHllosus, with segmentation obliterated in the large 
swollen glabella (from Warburg, 1925). I, diagram of larval trilobite, with 
head segmentation indicated. J, head of adult trilobite with probable segmenta- 
tion deduced from the larval structure (I). 

Acr, acron; dbl, doublure; dO, dorsal ocellus; E, compound eye; fg, fixigene 
(fixed cheek) ; fr/, frontal lobe (of acron) ; fsa, anterior part of facial suture; 
fsp, posterior part of facial suture; gib, glabella; I-IV, cephalic somites; Ig, 
libragene (free cheek); Lm, labrum; pi, palpebral lobe; IS-4S, intersegmental 
sutures of head. 

nules) of other arthropods; if so, according to the theory here 
followed, they should belong to the acron, and perhaps had their 
muscle attachments on the frontal lobe. The position of the antennal 


bases is not exactly known, but the antennal appendages are generally 
represented as arising at the sides of the labrum. The four following 
segmented, leglike appendages of the head clearly pertain to the four 
postfrontal somites of the dorsal shield. 

Henriksen (1926), in his analysis of the segmentation of the trilo- 
bite head, convincingly argues that the free cheeks bearing the com- 
pound eyes must belong to the "eye segment" (acron), and that the 
preocular parts of the facial sutures are secondary lines of cleavage 
to facilitate moulting ; but the median part of the eye segment he 
believes is represented only by the narrow anterior marginal rim of 
the dorsal shield before the frontal lobe. Henriksen notes, however, 
the anomalous position of the median eye far back on the glabella, 
and it is not clear why the reasoning by which he assigns the free 
cheeks to the eye segment does not demand that the eye segment 
include also the area of the median eye. The antennae, Henriksen 
contends, belong to a separate postoral somite, represented dorsally 
by the frontal lobe of the glabella. Furthermore, since he believes 
that the trilobite head must have the same segmentation as the head 
of certain Crustacea, Henriksen concludes that a second antennal 
somite has been lost by the trilobites. To the writer this theoretical 
elaboration of the trilobite head to give conformity with crustacean 
structure appears quite unnecessary, since the trilobites are non- 
mandibulate arthropods having no immediate relations with the Crus- 
tacea, and their structure clearly leads into that of the Chelicerata. 

The Xiphosurida, in the structure of the prosoma, show unmis- 
takably their trilobite derivation, for the trilobite head is carried over 
into the xiphosurid prosoma with few changes other than the inclusion 
of a few extra segments, the loss of the antennae, and a differentiation 
of the other appendages. 

A comparison of figure 47 A with figure 46 E will show at once 
the likeness of the prosomatic carapace of Limulus to the typical head 
shield of a trilobite. The segmentation of the xiphosurid prosoma is 
evident from the position of the limb bases on the ventral surface 
(fig. 47 C), where it is seen that the anterior somites lap forward at 
the sides of the labrum from behind the central mouth, while the 
posterior somites curve somewhat backward. The chelicerae (Chi) 
thus come to have anatomically a preoral position at the sides of the 
labrum, though their somite (/) is morphologically postoral, and the 
same is true of the pedipalps (Pdp) and the first legs (iL). On 
the inner surface of the prosomatic carapace the attachments of the 
limb muscles (fig. 47 B), as depicted by Benham (1885), follow the 
segmentation indicated ventrally by the limb bases. The cheliceral 





Fig. 47. — Segmental analysis of Xiphosurida (Liviulus polyphemus Linn.). 

A, young adult, dorsal surface. B, ventral surface of dorsal carapace, show- 
ing muscle attachments and series of dorsal apodemes, or entapophyses (from 
Benham, 1885, with dorsal ocelli added). C, ventral surface of prosoma, show- 
ing segmentation as indicated by position of leg bases. D, dorsal surface of 
opisthosoma, with segmentation indicated. E, theoretical approximate segmen- 
tation of prosoma. 

Acr, acron; Ap, tergal apodemes (entapophyses) ; Chi, chilarium ; Chi, chelic- 
era ; dbl, doublure (ventrally inflected part of acron); dO, dorsal ocellus; E, 
compound eye; fg, fixigene ; fs, facial suture; I-XIV, postoral somites; L, leg; 
Ig, libragene ; Lm, labrum ; Mth, mouth ; Opl, genital operculum ; p, external 
pits of tergal apodemes ; i^a, b, f, dorsal attachments of tergo-sternal muscles 
of opercular and gill somites ; 20a, b, f, dorsal attachments of anterior muscles 
of opercular and gill appendages; 25a-e, 28, dorsal attachments of tergo-coxal 
muscles of prosomatic appendages. 


muscles {2^) arise near the midline just behind the dorsal eyes; the 
muscles of the pedipalps {25a) take their origins farthest forward; 
and the muscles of the other appendages (2^b-2jc) are distributed 
on the following areas of the "fixed cheeks." Diagrammatically, there- 
fore, we may represent the segmentation of the prosomatic carapace 
as given at E of figure 47. The horseshoe-shaped acron (Acr) bear- 
ing the eyes encloses the region of the prosomatic somites (I-VIII), 
and sends posteriorly, between the lobes of the anteriorly curved 
cheliceral and pedipalp somites, a median tongue bearing the dorsal 
ocelli (dO). The structural conformity with the trilobite head (fig. 
46 J) is exact, except for the greater number of somites included in 
the xiphosurid prosoma. 

Students of the embryology of Limulus (Kishinouye, 1893, Iwanofif, 
1933) have indicated the segmental divisions of the prosoma as sub- 
tending the lateral areas of the carapace bearing the compound eyes. 
Branches of the segmental nerves, the "haemal nerves" of Patten and 
Redenbaugh (1900), extend into these parts, but, as in the case 
of the trilobites, the location of the compound eyes on the lateral plates 
of the prosoma is sufficient proof that these plates belong to the eye 
segment, or acron. Hence, they cannot be lateral extensions of the 
median somites. 

The gills of the trilobite legs, borne on coxal epipodites (fig. 48 D, 
Eppd), have not been retained on the prosomatic appendages of 
Xiphosurida, though an epipodite is present on the fourth leg (E, 
Eppd), but gill-bearing epipodites are highly developed on the opistho- 
somatic appendages, which are otherwise much reduced. 

The prosomatic appendages of Limulus, except the chelicerae, as 
shown by Benham (1885), have the typical arthropod coxal muscu- 
lature, consisting of dorsal promotor and remotor muscles (fig. 48 F, 
I, J), and ventral muscles {K, L). Of the latter, two (jj, j^) are 
promotors and remotors, but two others {32111, 32n) are united 
proximally and evidently function as adductors. The dorsal muscles 
arise on the tergal carapace (C). The ventral muscles, however, are 
attached on an internal plate, or "entochondrite" (k), suspended in 
the ventral part of the body by dorsal muscles (t-s). The same 
structure (B) is characteristic of most of the Chelicerata, and a 
similar structure occurs in the gnathal segments of many of the 
Mandibulata (figs. 50 E, H, 51 B, k). Since the ventral muscles of 
the appendages should primarily arise on the ventral body wall, the 
"entochondrite" might be supposed to be a sternal derivative, but 
Schimkewitsch (1895, 1906) claims that in the Arachnida it is pro- 
duced from transformed muscle tissue. In various mandibulate 


arthropods some of the adductor fibers of the mandibles go con- 
tinuously from one appendage to the other. 

That the prosoma of Limulus contains at least a part of the eighth 
somite is evident from several structural features, but the writer's 

Fig. 48. — Structure of Chelicerata and Trilobita 

A, Liobunum sp. (Phalangida), anterior view of body, showing secondary 
preoral position of chelicerae above base of labrum. B, same, "endosternite" 
of prosoma with adductor leg muscles (suspensory dorsal muscle not shown). 
C, Limulus polyphevms Linn. (Xiphosurida) , section of prosoma behind third 
legs, leg muscles somewhat diagrammatic. D, diagrammatic cross-section of 
a trilobite. E, Limulus polyphemus, fourth leg, with coxal epipodite. F, same, 
base of a left leg, mesal view, with muscle insertions (from Benham, 1885). 

a, dorsal articulation of coxopodite ; b, ventral end of coxal axis ; Bud, basen- 
dite ; Chi) chelicera ; Cp, carapace ; Cxpd, coxopodite ; D, dorsum ; dbl, doublure ; 
dO, dorsal ocellus ; E, lateral eye ; Eppd, epipodite ; /, tergal promotor muscles 
of coxopodite ; J, tergal remotors of coxopodite ; K, anterior ventral muscles of 
coxopodite ; k, ligamentous "endosternite" on which ventral leg muscles are at- 
tached ; L, posterior ventral muscles of coxopodite ; 3L, third leg ; Lm, labrum ; 
Pdp, pedipalp; Tlpd, telopodite; t-s, tergal suspensory muscleof "endosternite"; 
V, venter ; 26, 2/, dorsal promotor muscles of coxopodite arising on carapace ; 
2^, 28, 2g, dorsal remoter muscles arising on carapace ; 32m, 32n, anterior and 
posterior branches of coxal adductor arising on "endosternite" ; 33, 34, ventral 
remotor and promotor muscles arising on "endosternite". 

former statement (1936) that the prosoma and opisthosoma of 
Limulus are separated between segments VIII and IX is not strictly 
correct. The attachment of the muscles and the distribution of the 
nerves in this region demonstrate that the dorsal hinge between the 
prosomatic carapace and the opisthosomatic carapace lies within the 


eighth segment itself, and not behind it, a narrow anterior median 
part of this segment being incorporated into the posterior margin of 
the prosoma, while lateral parts of it form the anterior lateral lobes 
of the opisthosomatic carapace (fig. 47 D, VIII). Six following seg- 
ments of the opisthosoma are marked by the six pairs of impressions 
bordering the median elevation of the carapace, and by the six pairs 
of marginal spines. The intrasegmental division of the body into 
movable parts is not an anomalous condition ; it occurs between the 
thorax and the abdomen of many insects, and is a necessary mechanical 
adaptation resulting from the primarily intersegmental attachments 
of the longitudinal muscles. 

The six pairs of dorsal impressions on the opisthosoma of Limulus 
(fig. 47 D) and a pair of similar impressions on the posterior margin 
of the prosomatic carapace (p) form internally (B) a double series 
of tergal apodemes, the "entapophyses" of Benham (1885), of which 
the larger first pair (VIIIAp) is on the prosoma, and the other six 
pairs (IXAp-XIVAp) are on the opisthosoma. The tergosternal 
muscles of the five gill-bearing segments are shown by Benham to have 
their dorsal attachments {i2h-i2f) at the bases of the first five opistho- 
somatic apodemes, while the corresponding muscles of the opercular 
segment (VIII) arise at the bases of the corresponding prosomatic 
apodemes (12a). On the other hand, while the "external branchial 
muscles" of the gill segments have their dorsal attachments (20b- 
2of) just laterad of the first five opisthosomatic apodemes, the corre- 
sponding muscles of the operculum take their origins also on the 
opisthosomatic shield, but more laterally on the anterior lateral lobes. 
The muscle attachments, therefore, show that the dorsal part of the 
eighth segment has been divided between the prosoma and the opistho- 
soma, or, as Benham says, the first pair of tergal apophyses has been 
transferred from the opisthosoma to the prosoma. The dorsal longi- 
tudinal muscles between the prosoma and the opi-sthosoma of Limulus 
have been condensed into a single large bundle of fibers, the "arthro- 
tergal muscle" of Benham, and the attachments of this muscle (fig. 
47 B, y8) have extended somewhat anteriorly and posteriorly on the 
two body regions to acquire greater efficiency as a fiexor of the 

The innervation of the hinge region of the carapace gives the same 
evidence of division within the eighth segment as that furnished by 
the musculature. As shown by Patten and Redenbaugh (1900), the 
nerves of the genital operculum proceed from the composite ventral 
ganglion of the prosoma, while the corresponding somatic nerves 
("haemal nerves" of segment VIII) are distributed to the anterior 


lateral lobes of the opisthosoma. From a comparative study of the 
position of the cardio-aortic valve in the Chelicerata, Petrunkewitch 
(1922) found that the valve is always between segments VIII and 
IX, and he therefore claimed that segment VIII is included in the 
prosoma of Limiilus. His contention is but little affected by the 
modified view here shown to be in better accord with the facts. The 
operculum is anatomically more closely connected with the prosoma, 
from which it derives its innervation, than with the opisthososa, and 
the partition of the tergum of its segment between the prosoma and 
the opisthosoma, as above noted, is but a necessary adaptation to give 
intersegmental action to primarily intrasegmental muscles. 

The Eurypterida and Arachnida differ from the Xiphosurida in that 
the prosoma includes only six somites, and in this respect they are 
nearer to the Trilobita, which have only four prosomatic somites. 
The eurypterids and arachnids, however, lack the lateral expansions 
of the prosomatic carapace characteristic of the trilobites and xipho- 
surids, and, judging from the more anterior position of the lateral eyes 
(fig. 49 E, E), it seems probable that the acronal element of the 
prosoma is less extensive on the marginal areas of the latter, though 
medially it must include the region of the dorsal eyes (dO). 

In a typical arachnid embryo (fig. 49 A) the somites are regular 
transverse sections of the trunk behind the large prostomial acron 
(Acr), which is produced laterally into a pair of cephalic lobes. 
Ordinarily there are no appendages on the acron, but Jaworowski 
(1891) has described a pair of apparent antennal rudiments in a 
species of Trochosa (C, b), and Pokrowsky (1899) found two pairs 
of transient precheliceral lobes in an embryo of Pholcus opilionides 
(B, a, h), the second of which, he says, correspond in position with 
the "antennal" rudiments described by Jaworowski. The nature of 
these embryonic lobes may be doubtful, but since the trilobites have 
well-developed antennae, there is no reason why embryonic vestiges 
of acronal appendages might not recur in some chelicerate forms. In 
adult Solpugida there is a pair of small appendages (fig. 49 F, Antf) 
arising at the sides of the epistomal lobe, which are movable by 
muscles (G, mcl), and are, therefore, suggestive of being antennal 

The cheliceral somite of the arachnid embryo (fig. 49 A, /) lies 
transversely immediately behind the acronal lobes ; but in the adult 
this somite curves forward around the sides of the labrum from 
behind the mouth as in Liinnlus, so that the chelicerae come to have 
a preoral position above the labrum (fig. 48 A, Chi), though usually 
they are separated by a median epistomal bar extending downward 



-a Lm b 


VOL. 97 

Fig. 49. — Embryonic and adult structures of Arachnida. 

A, embryo of Agelena labyrinthica (from Balfour, 1880). B, embryo of 
PholchS opilionides Schranck with two lateral lobes (a, b) on acron (from 
Pokrowsky, 1899). C, embryo of Trochosa singoriensis Laxm., with possible 
antennal rudiments (from Jaworowski, 1891). D, longitudinal section through 
anterior end of prosoma of a phalangid (Liobiiuiiiu) , showing anterior tergal 
attachments of cheliceral muscles. E,, dorsal surface of phalangid (Liobunum) , 
legs removed from coxopodites. F, epistomal lobe and labrum of a solpugid, 
lateral view, showing movable appendage {Ant?) at side of epistoma. G, same, 
longitudinal section, showing muscles of epistomal appendage. 

a, possible preantennal rudiment of embryo; Acr, acron (cephalic lobe of 
embryo) ; Ant?, adoral (possibly antennal) appendage of adult solpugid; b, 
possible antennal rudiment of embryo ; Ap, apodeme ; Chi, chelicera ; Chlmcls, 
cheliceral muscles ; dO, dorsal ocellus ; E, lateral eye ; Epst, epistoma ; GC, 
genital chamber; Gtr, gonotreme; IIBnd, IIIBnd, basendites of second and 
third appendages; IXS, sternum of ninth somite; iL, first leg; Lm, labrum; 
nicl. muscles; Mth, mouth; Pdp, pedipalp ; Stom, stomodaeum ; I-XV, postoral 


to the labrum from the frontal region of the carapace. As in Lvmulus 
again, the chehcerae have only dorsal muscles, which arise on the 
anterior part of the carapace (fig. 49 D, Chlmcls). 

The ancestors of the modern Mandibulata were represented in the 
more generalized members of the Protarthropoda that persisted after 
the trilobite-chelicerate branch had been given ofif from the main stem 
(fig. 54). The Protomandibulata undoubtedly retained the primitive 
centipedelike form of the protarthropods, but, as shown in the embry- 
ology of modern Mandibulata, the head at this stage must have been 
a composite protocephalon (fig. 39 C, Prtc) formed by an intimate 
union of the first somite (/) with the highly developed prostomial 
acron {Acr). It carried, therefore, the eyes (£), the labrum (Lni), 
the acronal appendages, or first antennae (lAnt), and the appendages 
of the included somite, which became a second pair of antennal organs 
{2 Ant). The distinctive feature of the early mandibulates, however, 
was the presence of a pair of jaws, the mandibles (Md), developed 
from the bases of the appendages of the first postcephalic somite. 
Probably also the appendages of the following two somites were 
reduced in size and modified in a manner suggestive of their future 
transformation into maxillae ; and perhaps a pair of paragnathal lobes 
was developed between the mandibles and the first maxillary appen- 
dages, since these structures are not present in the chelicerate branch. 

The Crustacea represent the first offshoot from the mandibulate 
section of the arthropod stem that has given rise to a specialized group 
of modern forms (fig. 54). The wide recurrence among the Crustacea 
of cursorial appendages identical in segmentation with the legs of the 
trilobites can leave little doubt that the primitive crustaceans were 
polypodous walking animals, living on the bottom of the water or on 
aquatic plants along the ocean shores, and adapted to life in the water, 
as were the trilobites, by the development of branchial organs on exite 
lobes of the coxopodites. According to this view, the natatory appen- 
dages of swimming or purely pelagic Crustacea are legs that have 
been modified secondarily for swimming purposes, just as the gnathal 
appendages have been modified for feeding. It is a sound principle 
of ecology that pelagic forms in all cases have been derived from 
benthonic forms (see Hesse, Allee, and Schmidt, 1937, p. 179), and 
the fact that many of the more generalized modern Crustacea are 
pelagic is no argument that such forms are ancestral. The frequent 
biramous structure of crustacean appendages is entirely a crustacean 
feature, since the exopodite is a specially developed outer branch of 
the basipodite, and therefore has no counterpart in the Trilobita or 
in any other arthropod group. 




Fig. 50. — Cephalic structures of Crustacea in which the protocephalon (acron 
and first somite) is either the definitive head, or is united with several follow- 
ing somites to form a more extensive syncephalon. 

A, Eubranchipus vernalis Hay (Anostraca) : protocephalon a distinct head 
lobe (Prtc) separate from mandibular somite (II) ; maxillary somites (///, 
IV) united with each other. B, Apus longicandatus Le Conte (Notostraca), 
dorsal view : mandibular and maxillary somites united with protocephalon, maxil- 
lary tergum produced in large cephalic carapace (C/>). C, same, ventral view 
of head, showing labrum, antennules, mandibles, and maxillae. D, Daphnia 
p-iilex Degeer (Cladocera) : head structure as in Apus, body covered by bi- 
valved maxillary carapace. E, Eubranchipus vernalis, detached mandibular seg- 
ment, anterior view, showing mandibles suspended from tergum, and mandibu- 
lar musculature. F, Ncbalia bipcs Fabr. (Leptostraca) : bivalved carapace has 
same composition as in Apus and Daphnia. G, section of head of Ncbalia show- 
ing muscle attachments of protocephalic appendages. H, mandibles of Nebalia, 


The crustacean head is variable in structure according to the number 
of somites it contains. The most primitive crustacean head, as already 
noted, is a simple protocephalon formed by the union of the trito- 
cerebral somite with the prostomial acron. A head of this type occurs 
in some of the Branchiopoda, and in all the Malacostraca except 
Leptostraca, Amphipoda, and Isopoda. 

The best example of a protocephalic head is seen in the anostracan 
branchiopods. The head of Enhranchipus, for example (fig. 50 A, 
Prtc), is a large cephalic capsule bearing only the eyes, both pairs of 
antennae, and the labrum. Behind it is the small but distinct tergum 
of the mandibular somite (//), which supports the large mandibles 
(Md). The next following segment is evidently the two maxillary 
somites united (III + IV), since it carries the vestigial first and second 
maxillae. The muscles of the head appendages, including those of 
the eye stalks, the antennules, the second antennae (in the male), and 
the labrum, all take their origins on the walls of the head capsule. 
The mandibular muscles, on the other hand, arise on the mandibular 
tergum (E), except the adductors (KL), which are united on a 
median ligament (k) and thus form a zygomatic muscle between the 
two jaws. 

The head of most of the other Entomostraca and of Leptostraca is 
a more extensive structure than that of the Anostraca, since it includes 
the mandibular and maxillary somites united with the protocephalon. 
The maxillary region of the head is often expanded to form a large 
cephalic shield, or bivalved shell, covering the anterior part of the 
body. In Apus (fig. 50 B) the region of the protocephalon (Prtc) 
forms a distinct anterior part of the definitive head bearing the 
eyes dorsally and the antennae and labrum ventrally (C). Behind 
the protocephalon the limits of the mandibular tergum (B, //) are 
clearly marked, but the maxillary terga (III + IV) are confluent as in 

posterior view. I, Porcellio sp. (Isopoda), head, composed of protocephalon 
and four following somites (maxillae and maxillipeds removed). J, Orchcs- 
toidea californica Brandt (Amphipoda), head, same composition as in Porcellio, 
approximate division between protocephalic and gnathal regions indicated by 
broken line (2s). K, Talorchcstia longicornis Say (Amphipoda), right mandi- 
ble, mesal view. 

a, primary (dorsal) articulation of mandible; lAnt, first antenna (antennule) ; 
2Ant, second antenna ; Bud, basendite ; c, secondary (anterior) articulation of 
mandible ; Cp, carapace ; E, compound eye ; /, tergal promoter muscles of man- 
dible ; 1 1 -VI, second to sixth somites ; I IT, mandibular tergum ; /, tergal remotor 
of mandible ; k, ligament uniting ventral adductors of mandibles ; KL, ventral 
adductor muscles of mandible ; Lm, labrum ; Md, mandible ; iM.x, sMx, first and 
second maxillae ; MxGld, maxillary gland ; iMxp, first maxilliped ; Pip, palpus ; 
Prtc, protocephalon ( acron -|- somite /) ; 2S, suture between protocephalon and 
mandibular somite {B) , or theoretical line of division between protocephalic and 
gnathal regions of head (J) ; 3s, suture between mandibular and maxillary 
somites ; VStn, sternum of first maxilliped somite. 


EubrancJiipus (A). In the Cladocera (D) the general head structure 
and composition is the same as in Aptis, except for the lateral com- 
pression of the maxillary shield, which gives the latter its "bivalved" 
form, but the intersegmental lines are lost, and the limits of the proto- 
cephalon (Prtc) are marked only by the attachments of the antennal 
muscles. The Leptostraca (F) have the cladoceran type of head and 
bivalved maxillary shield, but are distinguished by the presence of a 
large frontal lobe (s) projecting above the bases of the eye stalks. 
Here again the protocephalic area of the composite head is marked 
only by the origins of the muscles of the protocephalic appendages 
(G), including those of the eye stalks and the two pairs of antennae. 
The mandibles of the Leptostraca (Nebalia) retain the palpi (H, Pip), 
but their basal structure and musculature is the same as those of 
Euhranchipus (E) and other Entomostraca. 

The Malacostraca, other than Amphipoda and Isopoda, are com- 
monly said to have a "cephalothorax," which includes the gnathal 
somites and a number of following somites up to a maximum of 12 
in all. Most of this composite structure, however, which in its fullest 
development is covered by the carapace (fig. 51 C, Cp), is more truly 
a gnathothorax, since the true head is always a distinct though small 
protocephalic lobe more or less concealed beneath the overhanging 
rostrum (r) of the mandibular somite. When the protocephalon is 
detached, as shown in the figure (C), it is seen to be a distinct cephalic 
structure bearing the stalked eyes, both pairs of antennae, and the 
labrum. The typical malacostracan head is thus identical with the 
protocephalic head of the Anostraca (fig. 50 A, Prtc). Even in the 
Brachyura (fig. 51 D, E) the protocephalon is readily identified as 
such, though dorsally (D) it is much reduced, and is concealed in a 
pocket beneath the anterior margin of the carapace; ventrally (E) 
it carries a large epistomal plate and a small labrum. In the Stoma- 
topoda, on the other hand, the protocephalon is highly developed (fig. 
51 F, G), and its integumental sclerotization is broken up into several 
distinct plates {d, e, f, g), which, however, can in no sense be re- 
garded as representing a "segmentation" of the head. The mandibles 
of the more generalized type found in the Malacostraca (B) are 
identical in their structure and musculature with the mandibles of 
Entomostraca (fig. 50 E, H). 

The Amphipoda and the Isopoda (including Apseudidae), with 
regard to the structure of the head, do not appear to be properly 
classed with the rest of the Malacostraca, since the head (fig. 50 I, J) 
is an intimate combination of the gnathal somites (II + III + IV) 
with the protocephalon (Prtc), and thus resembles in its composition 


Fig. 51. — Cephalic structures of malacostracan Crustacea in which the defin- 
itive head is the protocephalon, as in Anostraca (fig. 50 A). 

A, Anaspides tasnianiae Thomson (Syncarida), protocephalon and appen- 
dages, anterior view. B, same, mandibles and muscles, posterior view. C, 
Spirontocaris polaris (Decapoda-Macrura), showing protocephalon (Prtc) 
detached from carapace. D, Callincctcs sapidus Rathbun (Decapoda-Brachy- 
ura) protocephalon and appendages, dorsal view. E, same, protocephalon, 
anterior view. F, Chloridella panamc^isis Bigelow (Stomatopoda), protocephalon 
and appendages, ventral view. G, same, protocephalon detached from carapace 
(Cp), dorsal view. 

a, primary (dorsal) articulation of mandible; lAnt, first antenna (antennule) ; 
2 Ant, second antenna; Bnd, basendite ; c, secondary (anterior) articulation of 
mandible; Cp, carapace; Cxpd, coxopodite; d, anterior (ocular) division of 
protocephalon ; E, compound eye ; e, ocular plate of protocephalon ; Epst, epi- 
stoma ; Expd, exopodite ; /, postocular dorsal plate of protocephalon ; g, pos- 
terior (antennular) division of protocephalon; /, /, dorsal promoter and re- 
motor muscles of mandibles; II -VIII, second to eighth somites; k, adductor 
ligament of mandibles ; KL, adductor muscles of mandibles ; Lm, labrum ; Prtc, 
protocephalon; r, rostrum of mandibular somite; Tlpd, telopodite (palpus). 


the head of Nebalia and of such entomostracan forms as Apus, 
Daphnia, and others, though in form it often has a striking resem- 
blance to the head of a hexapod mandibulate. However, in both the 
amphipods and the isopods the head usually includes also the first 
maxilliped somite and its appendages (fig. 50 J, iMxp), and may in 
addition bear the second maxillipeds. The mandible acquires a secon- 
dary anterior articulation with the cranium (I, J, K, c), by which 
its action is limited to a hinge movement on a horizontal axis between 
its two articular points (K, a, c). The same mandibular mechanism 
has been independently developed in the decapod Crustacea and in the 
pterygote Hexapoda. While the head structure of the Amphipoda 
and Isopoda sets these groups apart from other Malacostraca, it does 
not necessarily relate them to any other group. 

The final type of head developed in the Arthropoda is that char- 
acteristic of the myriapods and hexapods, and must have evolved in 
the common ancestors of these groups represented in the post- 
crustacean, protomyriapodan section of the main arthropod stem 
(fig. 54). The head of all the myriapod and hexapod groups is a 
highly standardized structure, composed of the protocephalon and 
the three gnathal somites, so closely united that little evidence of the 
original segmentation remains, except in the presence of the appen- 
dages (fig. 53 A), and even here the evidence is obscured by the loss 
of the second antennae. In early ontogenetic stages, however, the 
gnathal somites are entirely distinct from a large anterior cephalic 
lobe that usually includes the second antennal somite, which may bear 
vestiges of its former appendages. The Protomyriapoda must have 
had compound eyes, since eyes of the compound type recur finally 
in the Hexapoda; they likewise must have carried paragnathal lobes 
of the head from the Crustacea to the Hexapoda, though these organs 
have disappeared in the modern myriapodous forms. The maxillary 
appendages probably were no more specialized in the protomyriapods 
than in modern Chilopoda (fig. 53 A, C). The mandibles lost the 
telopodites, but they developed a special feature of which no sug- 
gestion is to be found in the Crustacea, namely, a mobile gnathal lobe, 
the lacinia, movable by muscles arising in the mandibular base and 
on the walls of the cranium. The mandibular lacinia is retained as 
a movable lobe in modern Symphyla (fig. 52 E, Lc) and Diplopoda; 
in the Chilopoda it is not separated from the stipital region of the 
mandible (fig. 53 E, F), though it is provided with strong stipital 
and cranial muscles (F, 75, 10) ; in the Pauropoda and Hexapoda 
(fig. 52 F) apparently it has united with the stipes {St), producing 
a solid jaw of the crustacean type, and its muscles have disappeared. 


Vs Sty 

Fig. 52. — Symphyla, Diplopoda, and Thysanura. 

A, Scutigerella itnmaciilata Newport (Symphyla). B, same, maxilla. C, 
same, labium. D, labium of Machilis sp. (Thysanura). E, mandible of 
Scutigerella. F, mandible of Machilis. G, gnathochilarium of Fontaria vir- 
giniana (Drury) (Diplopoda). H, thirteenth body segment of Scutigerella, 
ventral view. I, seventh abdominal segment of Machilis, ventral view. J. thir- 
teenth body segment of Scutigerella, lateral view. K, last leg of Scutigerella, 
posterior view. L, terminal segments of Scutigerella, lateral view. 

a, a", a'", cranial articulations of mandible, maxilla, and labium; Cd, cardo; 
Cer, cercus ; Cx, coxa; dac, dactyl (clawlike remnant of dactylopodite) ; flee, 
cranial flexor muscle of lacinia ; Fm, femur ; Ga, galea^ H, head ; Lc, lacinia ; 
Pip, palpus ; Ptar, pretarsus ; St, stipes ; Stn, sternum ; Sty, stylus ; Tar, tarsus ; 
Tb, tibia; Tel, telson; iTr, first trochanter; 2Tr, second trochanter (prefemur) ; 
un, lateral claw (unguis) of pretarsus ; Vs, eversible vesicle; 1-16, body segments. 


The maxillary appendages in Symphyla and Hexapoda have acquired 
two endite lobes of the stipes (lacinia and galea), but the palpi have 
been lost in Symphyla (B). 

The last important event in the evolution of arthropod head appen- 
dages was the union of the bases of the second maxillae to form a 
single median organ, the so-called labium. The labium took its origin 
in the common ancestors of the Symphyla, Diplopoda, Pauropoda, 
and Hexapoda, which constituted the third and most prolific branch 
of the arthropod stock (fig. 54). The primitive structure of the 
labium is best preserved in the more generalized hexapods (fig. 
52 D) ; in the Symphyla (C), Pauropoda, and Diplopoda (G) it has 
lost the telopodites, or palpi, and in the diplopods it forms at least a 
part of the complex gnathochilarium (G). 

Crampton's (1928) phylogenetic conclusions drawn from com- 
parative studies of the arthropod head differ radically in some respects 
from the concept of arthropod relationships here deduced from the 
same source. Crampton believes that the first arthropods derived 
from annelid precursors were probably prototrilobites, and that from 
the latter were evolved in one direction the Trilobita and Chelicerata, 
in another the Protocrustacea, which last in turn produced the higher 
Crustacea, while finally, from the malacostracan Crustacea were 
evolved the Myriapoda and Hexapoda. 

To the writer it would seem that if the Protarthropoda are con- 
ceded to have been derived from wormlike ancestors, whether anne- 
lidan or protonychophoran, by a sclerotization of the integument and 
a jointing of the appendages, they must have taken on at once a 
centipedelike form. According to the theory here proposed, therefore, 
a long, unbroken line of slender polypodous arthropods has persisted 
from the ancient protonychophoran progenitors to the modern chilo- 
pods. Along this line have been carried the features common to all 
the arthropods, while new characters evolved in the main line itself 
have been distributed to subsequent lateral branches, where in some 
cases they have persisted in their original state, in others they have 
still further evolved, and in still others they have been lost. 

The first lateral branch from the arthropod stem was that of the 
Prototrilobita (fig. 54), which produced the Trilobita and the Chelic- 
erata. In this branch cephalization united the first four somites with 
the acron to form the trilobite "head," and continued in the Chelicer- 
ata until the "prosoma" contained six and eight somites. Meanwhile, 
in the main protarthropod stem, cephalization produced a more simple 
head (protocephalon) consisting of the acron and only the first somite, 
but the appendages of the second somite were converted into a pair 


of jaws. The protarthropods thus developed into Protomandibulata. 
At this point arose the crustacean branch, in which the simple proto- 
cephalon is still the definitive head in a large number of forms, though 

Md iMx -^1 C 

Fig. 53. — Head and mouth parts of Chilopoda. 

A, head of Scutigera forceps Raf. B, poison claws (first legs) and second 
body segment of Lithobius sp., ventral view. C, first and second maxillae of 
Lithohius, ventral view. D, head of Lithobius with maxillae removed, ventral 
view. E, right mandible and associated head structures of Lithobius, ventral 
view. F, right mandible of Lithobius, dorsal view. G, hypopharynx, hypo- 
pharyngeal suspensoria, and preoral mouth cavity of Lithobius, ventral view 
(labrum and clypeus removed). 

a', primary posterior articulation of mandible; Ant, antenna; c, secondary 
anterior articulation of mandible ; Clp, clypeus ; Cx, coxa ; E, eye ; For, foramen 
magnum; h, ventral inflection of cranium; HAp, hypopharyngeal apodeme; 
Hphy, hypopharynx (metastoma) ; HS, hypopharyngeal suspensorium (ful- 
tura) ; L, leg; Lm, labrum; Md, mandible; mr, mandibular rod; iMx, 2Mx, first 
and second maxillae ; ProC, preoral mouth cavity ; Sex, subcoxa ; Stn, sternum ; 
T, tergum ; /, 2, frontal and clypeal muscles of labrum ; 3, 4, frontal and clypeal 
muscles of hypopharyngeal suspensorium; 5, 6, 7, cranial muscles of same; 8, 
ventral dilator muscles of pharynx; 9, cranial muscle of mandibular stipes; 10, 
cranial flexor of mandibular lacinia (origin lateral on cranium) ; //, adductor 
muscle of mandible; 12, protractor muscle of mandible; 13, stipital flexor of 
mandibular lacinia. 

in several groups a more extensive head has been evolved by adding 
to the protocephalon the following three, four, or five somites. Cepha- 
lization, however, continued also in the main protomandibulate line, 
and produced here a composite head of standardized structure in 


which the three gnathal somites were intimately combined with the 
protocephalon, while the appendages of the protocephalic somite 
(second antennae) were suppressed. The Protomandibulata now be- 
came Protomyriapoda. A composite head has thus been produced 
along three separate lines of arthropod evolution, but in each case 
with characteristic differences. 

The Protomyriapoda had all the characters common to the several 
groups of arthropods finally derived from them, and also older 
characters earlier transmitted to the Crustacea, which later appear in 
one or more descendent groups, and are lost in others. From the 
protomyriapods there arose the final persistent arthropod branch, the 
Protosymphyla, while the main stem continued into the relatively 
generalized modern Chilopoda (fig. 54). The Protosymphyla de- 
veloped a labium by the union of the bases of the second maxillary 
appendages, and so characteristic is this feature of all their descen- 
dents, including the modern Symphyla, Diplopoda, Pauropoda, and 
Hexapoda, that this group as a whole might be distinguished as the 
Arthropoda labiata. The chilopods have developed few special fea- 
tures other than the conversion of the first legs into a pair of poison 
claws, but they have lost certain features of the protomyriapods. 


In no modern adult arthropod is there retained a complete series 
of coelomic sacs, but remnants of the coelom are preserved as the 
lumina of the gonads and genital ducts, of various glands having an 
excretory or accessory genital function, and perhaps of other glandu- 
lar structures. The embryonic development of the coelomic sacs of 
the Onychophora very probably recapitulates fairly closely the phylo- 
genetic history of the coelomic sacs in both the Onychophora and 
the Arthropoda. The primitive coelom of these animals undoubtedly 
consisted of a full series of paired segmental cavities, each opening 
to the exterior through a ventral diverticulum of the coelomic wall 
connected with the ectoderm mesad of the base of the corresponding 
appendage. The cavities must have served for the accumulation of 
excretory products, and for the retention of the developing germ 
cells, and the outlets gave vent to both the excreta and the 
gametes (fig. 34 A). The more primitive annelids do not have per- 
manent coelomic openings, and it seems doubtful that the simple 
coelomoducts of the Onychophora had a common origin with the 
metanephridia of the higher Annelida, since the metanephridia are 
outgrowths of the posterior walls of the coelomic sacs and each opens 
through the segment following. 


Early in the evolution of the common ancestors of the Onychophora 
and Arthropoda, judging from the embryonic development of modern 
Onychophora, the coelomic cavities were diflferentiated into dorsal 
compartments (fig. 34 B, a) containing the proliferation centers of 
the germ cells in their walls, and into ventral compartments (b) open- 
ing through the coelomoducts (c, d). With the complete separation 
of the two series of compartments (C), the dorsal compartments 
became gonadial sacs {G) and the ventral compartments (&) became 
nephridial sacs. The gonadial sacs, being deprived of outlets, united 
with one another on each side and formed a pair of longitudinal 
gonadial tubes (E, G), which retained exit passages through one 
pair of coelomic sacs that maintained their integrity and served as 
genital ducts. The ventral sacs and their respective coelomoducts 
were transformed into specific segmental excretory organs, or ne- 
phridia. It is thus clear that the genital ducts are not "modified 
nephridia," as they are often said to be, but that the genital ducts and 
the nephridia are separate products of the primitive open coelomic 
sacs, and hence, when once individually established, cannot be inter- 
changeable in function. However, because of the variable position of 
the genital ducts in the Arthropoda, it is evident that a dififerent pair 
of coelomic sacs has been retained in different groups to serve as 
genital outlets. 

Excretory organs of coelomic origin in the Arthropoda are repre- 
sented by the coxal glands of Chelicerata and the nephridial head 
glands of Crustacea, and perhaps also by certain head glands of 
Diplopoda, Chilopoda, and apterygote Hexapoda. The coxal glands 
of the Chelicerata, with one exception, consist of a single pair of 
excretory organs situated in the prosoma and opening at the bases 
of the appendages. Each gland in its fullest development is a com- 
posite structure composed of several lobes or saccules derived from 
coelomic sacs and united upon a common tubular base, the so-called 
"stolon," or "labyrinth," composed of glandular cells and tubules, and 
is connected with the exterior by one or two segmental ducts. The 
organ is, therefore, variable in features that might be supposed to 
vary in a composite structure of such a nature, as in the number of 
coelomic sacs involved, the number of segmental openings, and the 
position of the openings. The excretory head glands of Crustacea 
include a pair of antennal glands ("green glands") and a pair of 
maxillary glands ("shell glands"). The first are present in the adult 
stage only in the Malacostraca ; the second occur in the Entomostraca 
and in some Malacostraca ; both pairs are present in Nebalia (Manton, 


1934)- The maxillary glands are usually simple tubes or sacs, but 
the antennal glands may take on a highly complex structure. 

The coxal glands of Lhnulns are a pair of large brick-red organs 
lying in the sides of the prosoma. Each organ consists of four suc- 
cessive glandular lobes arising from a common longitudinal stolon 
composed of numerous connecting tubules, and of a long coiled duct 
that proceeds from an end-sac in the fourth lobe and opens behind 
the base of the fifth appendage (third leg). According to Patten and 
Hazen (1900) the nephridial lobes are developed from masses of 
mesodermal cells derived apparently from the somatic walls of the 
coelomic sacs of the second, third, fourth, and fifth somites. Similar 
masses of cells in the first and sixth somites degenerate and disappear. 
The duct arises as a tubular diverticulum of the fifth coelomic sac, 
which latter becomes the fourth nephridial lobe. A short terminal 
part of the definitive duct is formed as an ectodermal invagination at 
the external orifice of the mesodermal duct. 

The coxal glands of Arachnida are best known from the work of 
Buxton (1913, 1917, see also Petrunkewitch, 1933, and Chickering, 
1937). A relatively primitive condition is found in the araneid groups 
Liphistiomorphae and Mygalomorphae, in which each gland has two 
saccules, one in the third, the other in the fifth segment, both con- 
nected with a long convoluted tubular labyrinth, from which two 
outlet ducts proceed to the exterior, one opening behind the third 
appendage, the other behind the fifth. Such an organ would appear 
to be a composite structure formed by the union of three consecutive 
segmental glands. In certain genera of the Amblypygi group of the 
Pedipalpida the gland of the fifth segment is shown by Buxton (1917) 
to be an independent organ opening separately on the fifth segment. 
In the Uropygi each gland has two saccules but only a single opening, 
which is on the third segment. All other Arachnida have but a single 
saccule for each lateral gland and a single outlet, but the opening is 
at the base of the second appendage (pedipalp) in Solpugida and 
Palpigradida, at the base of the third appendage (first leg) in Ara- 
neida, excepting the two groups above mentioned, and at the base of 
the fifth appendage (third leg) in Scorpionida and Phalangida, as 
in Limulus. Buxton calls attention to the correspondence of the 
coxal glands of Solpugida and Palpigradida with the salivary glands 
of Onychophora, the organs in each case having their opening on 
the second postoral body somite. 

Studies on the development of the arachnid coxal gland appear to 
leave no doubt that the organs are derivatives of coelomic sacs with 
coelomoducts formed as direct diverticula from the sacs as are the 


coelomic ducts of Onychophora. Brauer (1895) has shown that in 
the embryonic development of the scorpion there are formed five pairs 
of diverticula from the coelomic sacs of somites III, IV, V , VI, and 
VIII, respectively, of which those of the fifth and eighth somites 
acquire openings to the exterior. The coelomic sacs and their diver- 
ticula in the fifth somite develop into the definitive coxal glands, the 
coelomic diverticula of the eighth somite become the genital ducts, 
and the sacs and diverticula of the other segments disappear. Accord- 
ing to Kishinouye (1894) the development of the coxal glands in 
the araneid genera Lycosa and Agelena shows that each organ is a 
composite structure formed of small parts of the coelomic sacs of 
somites ///, IV, and V, but only the first acquires an opening to 
the exterior. 

The nephridial glands of the Crustacea, being individual organs, 
resemble the simple nephridia of the Onychophora rather than the 
composite coxal glands of the Chelicerata. Each organ consists of a 
mesodermal end-sac, a mesodermal canal, which may become highly 
complex in form, and a short exit duct of ectodermal origin (see 
Cannon and Manton, 1927, and Manton, 1930). The embryogeny of 
the crustacean excretory glands, however, is in some cases compli- 
cated by an indirect course of development. 

The antennal gland of Hemimysis lamornae is said by Manton 
(1928) to be formed from two masses of cells derived from the 
antennal mesoderm, one of which produces the end-sac, the other the 
canal. The canal becomes attached distally to the ectoderm, and a 
small ingrowth from the latter forms a short ectodermal exit duct. 
Where the canal touches the wall of the sac, a compact group of 7 
or 8 cells bulges into the lumen of the canal, and at this point the sac 
and the canal become united, but the only visible communication 
between them, Manton says, is by fine rather vague channels passing 
through the nephrostome cells. According to Vogt (1935) the an- 
tennal mesoderm of My sis relict a produces only the canal and a sheet 
of connective tissue membrane in the base of the antenna, to which 
the canal becomes attached. A group of 8 cells in this membrane 
then forms the nephrostome. The true end-sac, Vogt claims, is con- 
structed from adventitious connective tissue cells that wander into 
the antenna from the thoracic segments and form the end-sac beneath 
the nephrostome membrane. Vogt contends that the development of 
the antennal gland of Mysis relicta so closely resembles the develop- 
ment of an annelid nephridium that the two organs must be homolo- 
gous structures, the nephrostome membrane of Mysis representing 
a dissepiment in the annelid. To the writer a parallelism in the two 


cases is far from evident, and the development of the mysid antennal 
gland seems better explained as a secondary modification of the 
developmental processes that give rise to the coxal glands of Arachnida 
and the nephridia of Onychophora. 

Most of the tracheate Mandibulata have a series of head glands 
pertaining to the gnathal somites, the openings of which lie mesad of 
the appendage bases, or are displaced anteriorly or posteriorly when 
the bases of the two appendages of a pair are united. Some of these 
glands have been shown to have an apparent excretory function, 
because of their property of eliminating from the blood particles of 
carmine injected into the body of the animal, and such glands also 
have a complex structure, usually described as consisting of a saccule, 
a labyrinth, and a duct. Hence, various writers have claimed that 
glands of this type represent nephridial organs corresponding with 
the excretory head glands of Crustacea, though little evidence as to 
their embryonic origin has been produced. 

The gnathochilarial glands of the Diplopoda have been shown by 
Bruntz (1903) to collect injected carmine from the blood, and they 
are said by Heathcote (1886) to be derived from the mesoderm in 
embryonic development. Likewise, according to Bruntz (1908) and 
Philiptschenko (1928), a pair of labial glands of apterygote insects 
have an excretory function and a nephridialike structure. These 
glands open either separately {Canipodca, Japyx) between the hypo- 
pharynx and the labium, or {Machilis, Lepisuia) their ducts unite in 
a common median duct, and are joined by the ducts of a pair of 
"posterior salivary glands." The labial glands of the apterygote in- 
sects, particularly those of Thysanura, would so evidently seem to be 
the same as the labial glands of pterygote insects, which are commonly 
found to be ectodermal organs, that it is difficult to believe they are 
not homologous structures, regardless of their function. In the 
Chilopoda, according to Fahlander (1938), there are present generally 
three pairs of head glands associated with the bases of the gnathal 
appendages, but in addition there is another pair having a complex 
structure suggesting an excretory function. These glands have each 
two openings, one mesad of the base of the first maxilla, the other 
behind the base of the second maxilla. Fahlander contends, there- 
fore, that each gland has been formed by the union of two nephridial 
organs pertaining to the maxillary somites. The morphological status 
of all such glands must yet be determined by a study of the embryonic 



The student of arthropod phylogeny is confronted at every turn 
with the vexing problem that arises from the different position of 
the genital outlets in the various arthropod groups, and in recent 
years much discussion has been given to the question as to hovi^ the 
heterogoneate condition came about (see Tillyard, 1930, 1932, 1935, 
Snodgrass, 1933, 1936, Reynolds, 1935, Imms, 1936). Two phases 
of the problem have been somewhat confused, namely, that pertaining 
to the position of the openings of primary lateral ducts, and that 
pertaining to the position of secondary median ducts. The opening 
of a median duct is subject to migration, usually in a posterior 
direction ; the openings of lateral ducts are closely associated with 
particular segments, since the lateral ducts themselves represent 
specific pairs of segmental coelomic sacs. 

The possible migration of lateral genital ducts is narrowly restricted 
because of the limitations imposed by the transverse segmental nerve 
trunks ; a secondary median duct formed by invagination of the ventral 
wall of the body, however, lies beneath the ventral nerve cords, and 
may, therefore, become lengthened from one segment to another by 
an extension of its connection with the body wall. There is no evi- 
dence to support Tillyard's (1930) contention that segmental gonads 
were once connected by a common duct, which has retained a single 
definitive opening on different segments in different arthropods, be- 
cause when the germaria were segmentally arranged they were con- 
tained in the dorsal parts of segmental coelomic sacs with individual 
openings to the exterior, and the serial union of the dorsal parts of 
the coelomic sacs has produced the definitive tubular gonads opening 
through a single pair of coelomic sacs, while the ventral parts of the 
other sacs discharging through the coelomoducts became nephridial 
sacs. Likewise, Tillyard's (1935) second proposal that a heterogo- 
neate condition has arisen by a variation in the number of somites 
formed before or behind the primary genital somite cannot be accepted 
for the reason that somite formation in the genital region is primitively 

Inasmuch as the primary lateral genital ducts represent specific 
coelomic sacs that have been retained to serve as genital outlets, a 
segmental difference in the position of the genital openings is to be 
explained only as the result of mutations that have been effective in 
the organizer of the zone of teloblastic growth, which determines 
what particular pair of coelomic sacs shall be utilized as genital exits. 
A branching of the embryonic lateral ducts has been observed by 


Hey mens and by Wheeler in Dermaptera, and, according to Heymons 
(1901), the definitive ducts of Scolopendra are formed from two 
united pairs of coelomic sacs. In such cases we have, perhaps, 
examples of the supplanting of one pair of exit sacs by another pair. 
The heterogoneate condition of modern arthropods, therefore, must 
be the result of mutations that occurred among ancestral forms. The 
faculty of mutation afifecting the position of the genital ducts was 
carried over into the entomostracan branch of the Crustacea, and was 
not entirely extinct in the early forms of the Hexapoda. Moreover, 
in the Chilopoda, as in the Onychophora, there still exists a variability 
as to the segment of the genital ducts, for, though the genital outlet 
is always on the subterminal segment in Chilopoda and on the ante- 
penultimate segment in Onychophora, the genital segment is not 
morphologically the same somite in all cases, since the number of 
somites preceding it may be quite different in different genera. In 
the Geophilomorpha, furthermore, the number of pregenital somites 
is said to vary even among individuals of the same species. 


I. — A planulalike creature with an open posterior blastopore was 
probably the ancestor of the Metazoa. A creeping form adapted to 
feeding on a subsurface by the forward elongation of the blastopore 
on the under side of the body might readily have evolved into a 
worm by the partial closure of the blastopore, producing thus an 
alimentary canal wdth a ventral subapical mouth and a terminal anus. 
The subapical position of the mouth differentiated the animal into an 
acronal sensory region, or prostomium, and a postoral visceral and 
motor region, the body, or sojiia in a restricted sense. 

2. — The unsegmented progenitors of the annelids were probably 
small, creeping, wormlike creatures having a simple alimentary canal, 
a mouth on the anterior part of the under surface of the body, and a 
terminal anus. Locomotion on solid surfaces was effected by a ventral 
clothing of cilia, and body movements were produced by a system of 
muscle fibers on the inner surface of the body wall, derived from 
the ectoderm. The body cavity was a blastocoelic haemocoele, and 
was largely occupied by lateral bands of a mesoblastic parenchyma 
proliferated in the gastrula stage from endodermal or ectodermal 
teloblastomeres. The nervous system consisted of longitudinal and 
circular nerve tracts centering in ganglionic cell groups of the pro- 
stomium, which latter eventually united to form a "brain." Sensory 
organs may have included tactile tentacles and photoreceptive "eye 
spots" located on the prostomium. 


S- — The annelidan progenitors acquired a more effective body 
movement by the attachment of the longitudinal somatic muscle fibers 
at several successive rings on the body wall, and by the accompanying 
formation of transverse muscular septa at the resulting integumental 
grooves. The body region of the wormlike animal in this way became 
differentiated into a small number of consecutive motor units, the 
primary somites. To regulate the new muscular mechanism of 
metameric movement, there was developed from the body surface 
of contact with the substratum a new somatic nervous system in the 
form of ventral nerve cords with ganglia corresponding with the 
myotomes. The primary and secondary nervous systems were then 
unified by a connection of the ventral nerve cords with the brain, 
and the somatic elements of the primary system disappeared. The 
ingrowth of the septal muscles cut the parenchymatous mesoblast 
bands into segmental blocks, and the latter became excavated by 
cleavage spaces (primitive coelomic cavities) for the accumulation 
of body fluid containing waste products. Excretory organs, if present 
at this stage, were simple nephridial tubules extending from the ecto- 
derm into the haemocoele, where they were associated with the meso- 
blast cavities. The inner parenchymal cells lining the cavities formed 
epithelial coelomic sacs, but the outer cells, being still an undifferen- 
tiated tissue, were converted into muscle fibers and connective tissue. 
The secondary muscles thus formed reinforced the primary somatic 
muscles already present, and eventually became the major part of 
the muscular system. The germ cells remained in a mass of undiffer- 
entiated tissue near the posterior end of the body, and the gametes 
were liberated probably through a pore or temporary rupture of the 
body wall. The primitive segmented worms evolved in this manner 
from unsegmented progenitors were the ancestors of the annelids. 

^. — To increase the reproductive function, the subterminal genital 
region of the primitive oligomerous annelids was enlarged by the 
successive generation of new somites from its undifferentiated tissue. 
A series of secondary telohlastic somites duplicating the structure 
of the primary somites was thus interpolated between the primitive 
body of the worm and a small postgenital terminal cone containing 
the anus. The multiplying germ cells spread into the haemocoele of 
the new somites, and groups of them became lodged in the walls of 
the coelomic sacs. The ripening germ cells were now discharged 
into the coelomic cavities, which latter thus became gonadial as well 
as nephric in function. Since the coelomic sacs as yet probably had 
no permanent openings, the gametes must have been liberated through 
temporary pores of the body wall, through secondary genital openings 


into the nephridia, or by the autotomous separation of the genital 
somites. At this stage the generaHzed annehds had acquired the 
fundamental characters common to the higher Annelida, the Ony- 
chophora, and the Arthropoda. 

5. — The increase in the size of the body by the addition of the telo- 
blastic genital somites created a demand for a still greater efficiency 
of locomotion, and, according to the nature of the response to this 
demand, two divergent groups of worms were evolved from the 
generalized annelids. The members of one group acquired segmental 
clusters of eversible and retractile chaetae serving as adjuncts to the 
somatic muscular system by maintaining a hold on surfaces of con- 
tact; the members of the other group developed segmental pairs of 
lobelike outgrowths of the body wall containing extensions of the 
somatic muscles, which served as primitive legs. The chaetae-bearing 
forms gave rise to the Chaetopoda ; the lobopod forms were the 
ancestors of the walking Onychophora and Arthropoda. 

6. — From the primitive chaetopods were evolved the several groups 
of modern annelids. By the extension of open tubes from the pos- 
terior walls of the coelomic cavities to the exterior, a more efficient 
type of excretory organ (metanephridium) was developed, which 
could serve also for the liberation of the gametes. The Polychaeta 
are distinguished particularly by the elaboration of external structures 
of various kinds, while the Oligochaeta and Hirudinea have achieved 
a higher development of internal organs and functions. The loco- 
motor powers of the Polychaeta were increased by the development of 
lateral lobes of the body wall supporting the segmental groups of 
chaetae, and in most forms each lateral pair of chaetigerous lobes 
eventually combined to produce a single locomotor organ, the para- 
podium. The parapodia served for progression on solid surfaces, 
and became also effective adjuncts to the undulatory motion of the 
body during swimming. 

y. — The lobopod annelids became further differentiated from their 
chaetopod relatives by a chitinization of the entire cuticula, and by 
the suppression of all the cephalic tentacles except one pair probably 
corresponding with the palpi of the Polychaeta. They also acquired 
outlets from the coelomic sacs to the exterior, but the exit ducts were 
formed as diverticula from the ventral walls of the sacs and opened 
each on the segment of its sac mesad of the base of the corresponding 
leg. The germ cells were located in the walls of the dorsal parts of 
the coelomic sacs, and the primitive coelomoducts discharged both 
excretory matter and the gametes. The coelomic sacs, however, soon 
became divided into dorsal gonadial compartments and ventral nephric 


compartments. The gonadial sacs of each lateral series united with 
each other, forming thus a pair of tubular gonads, which opened to 
the exterior through one pair of undivided coelomic sacs and their 
outlet ducts. The ventral nephric sacs now became exclusively excre- 
tory reservoirs, and, with the coelomoducts, formed a series of 
nephridial organs along each side of the body. As a result of the 
conversion of the original coelomic sacs into gonadial sacs and 
nephridial sacs, the haemocoele was restored as the definitive body 

At this stage of their evolution, the lobopod annelids assumed the 
status of Protonychophora. Some of the protonychophorons retained 
the flexible integument of the worms ; others developed a sclerotiza- 
tion in the cuticula, and thus acquired an external skeleton of cuticular 
plates. The soft-skinned forms, preserving some of the general 
aspects of their annelidan ancestors, evolved into the modern Ony- 
chophora ; the armored forms gave rise to the Protarthropoda. Since 
the members of both groups were well adapted by their leglike appen- 
dages to a walking mode of progression, many of their descendants 
found an advantageous habitat on land. 

8. — The Onychophora retained the cylindrical wormlike form, but 
they lost the segmented structure in the integument and musculature. 
The lobiform appendages became more efficient locomotor organs 
through the development of an incipient segmentation, and the acqui- 
sition of terminal claws, but the first postoral appendages were con- 
verted into a pair of jaws. The single pair of prostomial tentacles 
took an apical position by migrating forward on the dorsal surface 
of the head, but their nerve tracts were united by a commissure in the 
posterior part of the brain. The eyes retained the annelid type of 
structure. The somatic nerve cords, which presumably must have 
been ganglionated in the segmented generalized annelids, became 
simplified by a redistribution of the neurocytes, and took widely 
separated positions along the sides of the body. The ganglia of the 
jaw somite, however, united with the cerebral ganglion of the pro- 
stomium and became posterior lobes of the brain. The coelomic sacs 
of the penultimate somite, regardless of the total number of somites 
in the body, were retained intact to serve as genital outlets ; the 
persisting remnants of most of the other coelomic sacs became small 
end-vesicles of the coelomoducts, which formed nephridial excretory 

p. — The Protarthropoda, because of the hardening of the integu- 
mental cuticula, lost the flexibility and contractility of their annelidan 
ancestors and onychophoran relatives, and, to compensate, developed 


a mechanism of telescopic movement between successive body seg- 
ments by the simple device of retaining nonsclerotized areas in the 
posterior parts of the primary segments, thus establishing a secondary 
segmentation in which the longitudinal muscles became interseg- 
mental instead of intrasegmental in action. The sclerotized appen- 
dages necessarily became segmented into individually movable parts, 
and their movements became more specifically controlled by body 
muscles inserted on their bases. The protarthropods retained the 
annelid structure of the nervous system, and the independence of 
the first postoral ganglia of the ventral nerve cords. The prostomial 
appendages (antennules) assumed an anterior position by a forward 
migration below the eyes, with the result that in the arthropod brain 
the antennal lobes lie beneath the optic lobes, and the brain takes a 
vertical position by contrast with the horizontal position of the ony- 
chophoran brain. Lateral eyes of the compound type were first 
developed in the Protarthropoda. Because of the origin of the Protar- 
thropoda from Protonychophora, the protarthropods were equipped 
with a series of nephridial organs like those of the Onychophora, and 
their internal reproductive organs were of the onychophoran type. 
The segmental relations of the genital ducts, however, were subject 
to mutation, and the position of the gonopores was, therefore, dif- 
ferent in different forms, as shown by the highly variable position 
of the genital outlets in modern arthropods. 

The Protarthropoda, having an annelid ancestry, and being directly 
derived from wormlike protonychophorons by a sclerotization of the 
integument and a jointing of the appendages, could scarcely take on 
other than a centipedelike form and structure, though they did not, 
of course, have the composite head and other specialized features of 
present-day myriapods. The number of body segments was variable, 
and potentially large, since the production of new somites in the 
zone of growth was not limited. The cephalic appendages (anten- 
nules) were filamentous, the lateral eyes primitively compound. The 
body appendages were probably all ambulatory legs with little differ- 
entiation among them, each composed of seven segments. The 
dactylopodites were provided with extensor and flexor muscles aris- 
ing in the propodites. Aquatic forms probably had branchial epipo- 
dites on the coxopodites. Perhaps the majority of the protarthropods 
lived in shallow water near the ocean shore, where they inhabited the 
bottom or aquatic plants, but probably also they occurred abundantly 
in debris along the beach, and very likely some the them were to be 
found in damp places on the land. The genital openings being on 
specific body segments, propagation took place by sex mating, though 


fertilization of the eggs was probably external. Postembryonic de- 
velopment was anamorphic. The first major diversification of the 
Protarthropoda gave rise to the ancestors of the Trilobita and the 
ancestors of the Mandibulata (fig. 54). 

10. — The Trilobita preserved the uniform, generalized structure 
and segmentation of the protarthropod appendages, but otherwise they 
became highly specialized by a lateral extension of the margins of 
the body segments, taking on thus a broad, flattened form except for 
a median elevation giving passage to the alimentary canal. Further- 
more, the first four postoral segments became intimately united with 
one another and with the prostomial acron to form a solid anterior 
body section, or prosoma, the so-called "head," bearing the labrum, 
the eyes, the antennules, and four pairs of postoral ambulatory 
appendages. Basal endites of the anterior appendages may have 
served as feeding adjuncts, but the trilobites, so far as known, devel- 
oped no specific jaws. The Trilobita were entirely marine animals, 
but they lived at the bottom of the water, and their legs show few 
deviations from the ambulatory type of structure, except for the high 
development of branchial lobes from the lateral surfaces of the 
coxopodites. The extended tergal margins covering the gills probably 
formed respiratory chambers. The position of the genital openings 
in the trilobites has not been discovered, but, because of the close 
relation between the Trilobita and the Chelicerata, the genital aper- 
tures may be expected to be found on the fourth postcephalic segment. 
The Trilobita became extinct by the end of the Paleozoic period of 
geological history, but from a branch of the primitive pre-Cambrian 
prototrilobites were evolved the Chelicerata. 

II. — The Chelicerata are distinguished from the Trilobita by the 
union of several additional somites with the head to form a more 
extensive prosoma, by the loss of the acronal appendages (anten- 
nules), by a greater differentiation among the somatic appendages, 
and by the forcipate structure of the reduced first appendages. Very 
commonly, also, there is an extra podomere in at least some of the 
legs, the patella, interpolated between the femur and the tibia. In 
modern forms the nephridial organs are suppressed in most of the 
somites, but some of them are retained as coxal glands, and (except 
in Pycnogonida and some Acarinida) the genital openings occur 
always on the eighth postoral somite. The Chelicerata have become 
the most sepecialized of all the arthropods, there being little in their 
body form and general organization suggestive of the ancestral centi- 
pede type of structure, which is so evident throughout the mandibulate 


branch. The CheHcerata inckide the Xiphosurida, the Eurypterida, 
the Arachnida, the Acarinida, and very probably the Pycnogonida. 

12. — The Xiphosurida are undoubtedly the closest living repre- 
sentatives of the Trilobita. The xiphosurid prosoma has the same 
structure as the trilobite head, and the same composition except for 
the addition of three extra somites and a part of the eighth somite. 
Likewise, the opisthosoma corresponds with the trilobite pygidium 
extended forward to include all the somites behind the prosoma, so 
that in the Xiphosurida there is no intermediate "thoracic" region 
of free segments. Such fossil forms as Belinurus and Prestzvichia 
would appear to be intermediate between modern Xiphosurida and 
Trilobita, and the Middle Cambrian Naraoia (see Walcott, 193 1, 
fig. I ) must be related to the xiphosurid line somewhere close to the 
trilobites. The first six prosomatic appendages retain the leg type 
of structure, except for the reduction and chelicerate form of the 
first pair. The seventh appendages are reduced to a pair of small 
lobes, the chilaria, and the following six have the form of broad 
plates formed chiefly by epipodite lobes, those of the last five bearing 
lamellate gills. The genital openings in both sexes are on a median 
ventral fold of the eighth segment united with the bases of the 
opercular appendages of this segment. 

/J. — The Pycnogonida, judging from some of their structural 
features, such as the union of the anterior body segments, the pos- 
terior position of the dorsal eyes between the bases of the third pair 
of appendages, the presence of a patellar segment in the legs, and 
the chelicerate structure of the first appendages, are to be classed 
with the Chelicerata; but because of their many unique characters, 
including the occurrence of multiple genital openings, it is impossible 
to connect them closely with any other of the chelicerate groups. It 
may be noted, however, that species with eight pairs of legs have 
presumably the same number of somites in the prosoma as have the 

/^. — The Eurypterida and the Arachnida differ from the xipho- 
surids in having only six segments in the prosoma, and this character 
together with various other features of their organization shows that 
these two groups are more closely related to each other than is either 
group to the Xiphosurida. On the other hand, the Eurypterida have 
certain characters of the xiphosurids that leave little doubt of their 
common ancestry with the latter, and their descent from trilobite 
stock. The general resemblance of the eurypterids to scorpions sug- 
gests a relationship between the two, but the theory of Versluys and 
Demoll (1920, 1923) that the Eurypterida and Xiphosurida are 


derived from primitive aquatic scorpions cannot be maintained against 
the evidence of close relationship between the Xiphosurida and the 
Trilobita. The Arachnida, as invaders of the land, had to evolve 
organs for aerial respiration, and the lamellate gills of their aquatic 
progenitors borne on the abdominal appendages were structures 
readily convertible into "lung books" by invagination into pockets of 
the integument (see Lankester, 1885). In addition, however, tracheal 
ingrowths of the body wall were developed in the Arachnida, as they 
have been in nearly all the other terrestrial arthropods. 

75. — The Protomandibulata preserved the slender, polypodous, 
centipedelike form of the primitive protarthropods, but they acquired 
as a distinctive character a pair of jawlike feeding organs, the 
mandibles, developed from the bases of the second postoral appen- 
dages. Probably long before the evolution of the mandibles, the first 
somite had been united with the prostomial acron to form a primitive 
composite head, or protocephalon, bearing the acronal sensory organs, 
the mouth, and the first pair of postoral appendages, which last 
became a second pair of antennae. The two pairs of appendages 
following the mandibles were reduced and modified to serve as acces- 
sory feeding organs. The other appendages were probably all leglike 
in form, as in modern centipedes, and were 7-segmented, since a 
patella does not occur in the mandibulate branch of the arthropods. 
The circulatory system still retained the basic structure of that of 
the generalized annelids ; respiration probably was branchial, the gills 
being carried on epipodite lobes of the coxopodites, as in the Trilobita ; 
the nephridial organs were perhaps suppressed in most of the body 
segments, but those that remained were of the onychophoran type of 
structure. The reproductive organs were closed gonadial tubes open- 
ing in each sex through a single pair of ducts formed from a pair 
of coelomic sacs, but the segmental position of the genital openings 
varied in different forms according to what particular pair of coelomic 
sacs served as gonadial outlets. 

The primitive Protomandibulata probably inhabited both the water 
and the land, since from them were early evolved the aquatic Crus- 
tacea, while the main branch developed into the terrestrial Proto- 
myriapoda, from which have descended the modern myriapods and 
the Hexapoda. 

16. — That the Crustacea are derived from crawling, centipedelike 
protomandibulate ancestors is attested by the retention in all the higher 
forms of ambulatory appendages having the same structure as the 
limbs of terrestrial arthropods. Many forms, however, have become 
adapted in part or entirely to swimming by a modification of the 


appendages, and the special development of an exite lobe of the basi- 
podites has given rise to a characteristic biramous structure of the 
limbs. The primitive protocephalon is retained as the definitive head 
in the Anostraca and in most of the Malacostraca, but in the majority 
of the Entomostraca and in the Leptostraca, Amphipoda, and Isopoda 
from three to five gnathal somites have been united with the proto- 
cephalon to form a more extensive cephalic structure. A carapace is 
variously developed in many groups, either from the cephalognathal 
region, or from the gnathothoracic region, but there is no true cephalo- 
thorax formed by an intimate union of cephalic and thoracic somites 
as in the Chelicerata. The mandibles have no movable lobes such as 
those of the myriapods ; in most forms the jaws preserve the primitive 
monocondylic articulation with the head, but in the higher Mala- 
costraca they are secondarily dicondylic. The genital openings are 
variable in position in the Entomostraca, but are fixed with respect 
to a specific segment in the Malacostraca. The hatching of the young 
at an early embryonic stage has resulted in the development of 
specialized swimming larval forms representing more primitive an- 
cestral stages in their general structure than the immediate protar- 
thropod ancestors of the crustaceans. The great antiquity of the 
Crustacea is shown by the occurrence of highly evolved forms in the 
Cambrian period contemporaneous with the oldest known trilobites. 

I'j. — The Protomyriapoda, being the direct descendants of the 
protarthropods, perpetuated the generalized arthropod form after the 
trilobites, the chelicerates, and the crustaceans had branched off as 
side issues and taken on variously specialized forms. During their 
evolution the protomyriapods acquired the structures characteristic 
of their descendants, which include the modern Symphyla, Diplopoda, 
Hexapoda, and Chilopoda. The three gnathal somites became inti- 
mately united with one another and with the protocephalon, forming 
the standardized head of the above-mentioned groups, composed of 
the acron and four postoral segments. The compound eyes and the 
first antennae of the Protomandibulata were retained, but the second 
antennae became reduced and eventually were lost, though their 
ganglia were preserved as tritocerebral lobes of the brain. The 
mandibles lost the telopodites, but each had a strong gnathal lobe 
(lacinia) movable by a muscle arising within the coxopodite and by 
another arising on the cranial wall. The two postmandibular maxillary 
appendages were modified by a reduction of the telopodites and by 
other adaptations to serve as accessory feeding organs. Since the 
Symphyla and some of the more generalized Hexapoda have lateral 
hypopharyngeal lobes (superlinguae) resembling the paragnatha of 


Crustacea, it is possible that these structures were transmitted from 
the Crustacea to the symphyhds and hexapods through the Proto- 
myriapoda, though they have been lost in modern Diplopoda and 
Chilopoda. The legs of the protomyriapods were all alike and retained 
the generalized 7-segmented structure, but the extensor muscle of 
the pretarsus was lost, leaving only the flexor muscle, which, for 
more effective action, shifted its origin from the tarsus into more 
proximal segments of the leg. This last feature is a distinctive char- 
acter of all the descendants of the Protomyriapoda. N-ephridial excre- 
tory organs were supplemented or replaced functionally by Malpighian 
tubules of the proctodaeum. The position of the genital openings 
was probably in general posterior, but variable. Postembryonic de- 
velopment was anamorphic, the young being hatched with a small 
number of segments, and the full number acquired by teloblastic 
generation in the subterminal zone of growth. 

The Protomyriapoda undoubtedly were terrestrial, and the larger 
forms may have developed tracheal invaginations on various parts of 
the body for respiration, but there was no definitely established 
tracheal system transmitted alike to all the descendent groups of 
terrestrial mandibulates. The probable characters of the Proto- 
myriapoda are summarized as follows by Imms (1936) : 

(i) The head bore a single pair of antennae and two pairs of jaws, viz. 
mandibles and maxillae : the second maxillae were probably a subsequent acqui- 
sition. (2) The trunk was composed of a variable and indefinite number of 
sub-equal segments, each bearing a pair of legs. It is probable that anamorphosis 
was universal and was continued throughout the life of the animal. (3) The 
gonads opened to the exterior by paired apertures, and the segmental disposition 
of the orifices probably varied in different families and depended upon that of 

the coelomoducts involved (4) The alimentary canal was probably a 

simple straight tube, while the excretory organs were little more than procto- 
daeal outgrowths or pockets ; an accessory excretory function was probably 
performed by the fat-body. (5) Respiration was probably cutaneous in many 
forms and partially tracheate in others. The tracheae were presumably in the 
form of groups of unbranched tubuli devoid of taenidia and bearing a general 
resemblance to those of Diplopoda. Each group of tracheae opened laterally 
by means of simple cryptlike, segmentally arranged spiracles : in some forms a 
pair of spiracles was probably located also on the head. 

From the Protomyriapoda there emerged a specialized lateral 
branch, the Protosymphyla, from which have been evolved in one 
direction the progoneate modern Symphyla, Pauropoda, and Dip- 
lopoda, in another the opisthogoneate Hexapoda, while the general- 
ized myriapodan stock has more directly continued into the modern 


18. — Since modern Symphyla combine features of the progoneate 
Diplopoda and Pauropoda on the one hand, and of the opisthogoneate 
Hexapoda on the other, there can be httle question that they are direct 
descendants of common ancestors of these two groups. Modern 
Symphyla, however, are Hnked more closely with the progoneate 
forms by the anterior position of the gonopore, the segmentation and 
structure of the legs, and the retention of the movable laciniae of the 
mandibles. The Protosymphyla, therefore, gave rise to an opistho- 
goneate branch that became the Protohexapoda. 

In general appearance the Protosymphyla probably resembled their 
modern representatives, but retained certain features of the Proto- 
myriapoda that have been transmitted to the hexapod line, though lost 
in the progoneate descendants. The legs were all alike and had the 
/-segmented protomyriapod type of structure, but the coxopodites 
bore each, mesad of the telopodite base, a small stylus and an eversible 
vesicle, as in modern Symphyla (fig. 52 H),' which structures are pre- 
served also on the abdomen of some of the apterygote insects (I) . The 
appendages of the last body somite became reduced to styliform cerci. 
The head appendages included a pair of antennae, a pair of mandibles, 
and two pairs of maxillae. The lateral eyes must have been compound, 
because compound eyes have been transmitted along the arthropod 
line from the Trilobita to the Xiphosurida, the Crustacea, and through 
the Protosymphyla to the Hexapoda. The protosymphylan mandibles 
had the protomyriapodan structure, movable laciniae being well de- 
veloped, and palpi absent. The first and second maxillae retained 
the palpi and each acquired two basal lobes (lacinia and galea), fea- 
tures transmitted to the hexapods, though the palpi have been lost in 
the progoneate branch. The bases of the second maxillae, however, 
became united to form a single appendage, the lahhim, an organ so 
characteristic of all the descendants of the Protosymphyla that the 
group as a whole, including Symphyla, Pauropoda, Diplopoda, and 
Hexapoda, might well be designated the "Labiata" (fig. 54). 

/p. — The direct descendants of the progoneate branch of the proto- 
symphylids are the modern Symphyla, but at an early period there 
were evolved from the symphylid line the common ancestors of the 
Diplopoda and Pauropoda. The Symphyla retain the generalized 
structure of the body and appendages (fig. 52 A), but of the 16 body 
segments evident in the dorsum of most forms, 3 are without appen- 
dages. The legs (K) show the diplopod type of structure in the rela- 
tively large size of the second trochanter {■^Tr) and the smallness of 
the femur (Fin), but the coxae do not appear as typical leg segments, 
since each pair apparently is confluent in a large posterior division of 


the venter of the body segment (H, C.v), carrying mesad of the base 
of each telopodite a small stylus (Sty) and an eversible vesicle (Vs). 
The end segments of the legs are reduced to small dactyls (K, dac), 
but each has an accessory claw (un) arising from its base. The first 
legs are usually reduced in size and lack tibiae. The last body segment 
bears a pair of cerci (L, Cer), which presumably are homologues of 
the legs or possibly of the styli of the preceding segments. Com- 
pound eyes are absent. The mandibles preserve the movable laciniae 
(E, Lc) ; the maxillae have both laciniae and galeae (B), but the 
palpi are small or vestigial; the labium (C) is a simple flap without 
palpi. Lateral lobes of the hypopharynx (superlinguae) are present 
at least in Scutigerella, as shown by Hansen (1930), and a pair of 
slender apodemal arms extend into the head from the hypopharyngeal 
base. The single median genital aperture is situated on the anterior 
part of the venter of the fourth body segment, but since the paired 
gonopores of Pauropoda and Diplopoda are on the third body seg- 
ment, the median genital outlet of the symphylids might be supposed 
to have migrated secondarily into the fourth segment. 

20. — The Diplopoda are a specialized branch of the early Symphyla, 
in which the somites back of the fourth postcephalic somite are united 
in pairs to form double segments. The mandibles are well developed 
and have strong movable lacinial lobes, but there is only one post- 
mandibular appendage of the head, the gnathochilarium (fig. 52 G), 
the morphology of which is uncertain, though the organ is probably 
either a combination of the maxillae with the labium, or the labium 
alone. The legs of the first body segment are absent, and there are 
no cerci on the last somite. Body segments are numerous in most 
forms, and all but the first few are generated teloblastically in pairs 
during postembryonic development. The paired gonopores are on 
the third postcephalic somite at the bases of the second pair of legs. 
The Pauropoda are probably an early branch of the Diplopoda, in 
which a union of the somites in pairs had already taken place, and 
the first legs had been much reduced but not yet obliterated. Special 
characters of the pauropods are the lack of movable laciniae on the 
mandibles, a weak development of the gnathochilarium, and a branch- 
ing of the antennae beyond the fourth segments. 

21. — The Hexapoda resemble more closely the Symphyla than any 
other of the modern arthropods, a fact recognized by several of the 
earlier writers, and Packard (1898) first formulated a definite theory 
of the origin of insects from symphylid ancestors. Recently the 
evidence in favor of this theory has been more thoroughly reviewed 
in the light of present-day knowledge of the apterygote hexapods by 


Imms (1936), who shows that the most plausible concept of the 
ancestry of insects is that of symphylid derivation. The important 
difference between modern Hexapoda and Symphyla is in the position 
of the genital openings, the symphylids being progoneate, the hexa- 
pods opisthogoneate. It is necessary to assume, therefore, that the 
Protohexapoda were evolved from an opisthogoneate branch of the 

The Protohexapoda became differentiated as a hexapod group 
through the concentration of the locomotor function in the first three 
postcephalic segments, with the consequent division of the body into 
a motor thorax and a visceral abdomen. The abdominal appendages 
were reduced, modified for purposes other than locomotion, or sup- 
pressed, but in most cases the abdominal coxal remnants united with 
the sternal plates of the segments and preserved the styli and eversible 
vesicles inherited from the Protosymphyla, though on the thorax these 
structures were lost. The number of body segments was limited to 
14 somites and a simple terminal lobe (telson) containing the anus. 
The persistent appendicular organs of the last somite were styluslike 
cerci, as in Symphyla. It is probable that the true telopodites of all 
the abdominal segments were absent. The mandibles became solid 
jaws by a complete fusion of the lacinial lobes with the coxopodites, 
and thus came to resemble the mandibles of Crustacea, but the 
maxillae and labium retained the generalized protosymphylan struc- 
ture. The hypopharynx consisted of a median lobe and two lateral 
lobes, as in Symphyla, and had a pair of basal apodemes giving attach- 
ment to muscles of the gnathal appendages. The eyes were compound. 
The protohexapods were opisthogoneate insofar as the paired genital 
apertures were located on the posterior part of the abdomen, but the 
exact position of the ducts and their outlets was still subject to 
mutation, as shown in the variable position of the genital outlets in 
modern forms. 

The discrepancy in the position of the genital openings as between 
Symphyla and Hexapoda raises the chief difficulty in relating the 
hexapods directly to the symphylids. The opisthogoneatism of the 
Hexapoda, however, is more truly a heterogoneate condition, which 
in a broad sense applies to the entire group of labiate mandibulates, 
for the primary genital ducts open on the third postcephalic somite 
in Diplopoda and Pauropoda, on the eighth in Collembola, on the 
tenth in female Pterygota, on the thirteenth (primitively) in male 
Pterygota, and on the fourteenth in Protura. Since the primary 
gonopores of the hexapods are always fixed with specific segments, 
as in Symphyla, Pauropoda, and Diplopoda, the opisthogoneate con- 


dition in the Hexapoda is not comparable with that in the Chilopoda, 
in which the genital outlet, though always subterminal, may be on a 
quite different somite in different forms because of the variable 
number of somites that may precede it. There is reason for believing, 
therefore, that the opisthogoneate condition of the Hexapoda has been 
acquired secondarily, and that it is a derivative from the progoneatism 
of Symphyla and Diplopoda, rather than from the opisthogoneatism 
of Protomyriapoda represented in modern Chilopoda. The establish- 
ment of the genital openings on the posterior part of the body in the 
Hexapoda was very probably an adaptation correlated with the con- 
centration of the locomotor function in the thorax. 

22. — An early specialization among the Protohexapoda gave rise to 
the modern entognathous Diplura, Protura, and Collembola, small 
hexapods characterized by a retraction of the mandibles and maxillae 
into pouches of the head wall closed ventrally by the labium. The 
identity in the structure of the mouth parts would alone suggest a 
phylogenetic unity among the above-mentioned groups, but the latter 
show also a peculiarity in the development of the hypopharyngeal 
apodemes, which structures, instead of projecting as free arms into 
the head, as in myriapods and Machilidae, take the form of long 
internal ridges that, in Diplura and Collembola, diverge posteriorly 
from the base of the hypopharynx as sclerotic linear inflections of 
the membranous integument along the folds between the gnathal 
pouches and the inner surface of the labium. In Protura the two rods 
are united for a part of their length. These superficial apodemes 
give attachment to the same muscles as do the internal apodemes of 
other forms, and in Collembola they support an elaborate "tentorial" 
superstructure. In many other respects the entognathous hexapods 
are widely different from one another, and their inter-relationships 
are by no means clear. Except for the common characters above 
mentioned, they might be supposed to have had quite separate origins 
from protosymphylan or protohexapod ancestors (see Imms, 1936, 
fig. 11). They represent abortive lines of evolution that have not 
led to higher forms. 

The Diplura depart least from the thysanuran branch that has given 
rise to the winged insects, since they retain the abdominal styli and 
cerci, and have the usual hexapod position of the genital openings. 
The Protura preserve a remnant of the primitive anamorphism of 
the hexapod ancestors, inasmuch as the last two somites are formed 
during postembryonic development, but they lack antennae, styli, and 
cerci ; the small appendicular organs on the first three abdominal 
segments may be coxal remnants of limbs, with eversible vesicles in 


one family. The paired genital ducts in both sexes open on the 
eleventh abdominal segment. The CoUembola are the most aberrant 
of all the hexapods, and in some ways the most primitive. They 
have only nine body segments, and the single genital opening is on 
the fifth abdominal segment. There can be no question that the Col- 
lembola are derived from more generalized ancestors having a greater 
number of segments, but since, in their phylogenetic history, segment 
formation in the zone of growth has ceased after the establishment 
of the genital ducts in the eighth somite, it is fruitless to look for 
evidence of the ancestral segmentation in the embryogeny of present- 
day CoUembola. The three pairs of appendicular organs on the 
collembolan abdomen are unique in structure, and give little suggestion 
of homology with the abdominal appendages of Symphyla, Diplura, 
and Thysanura, though it may be supposed that the collophore is a 
pair of united eversible vesicles, and that the two paired appendages 
are highly developed styli. (For a fuller discussion of the special 
features of the CoUembola, see Imms, 1936.) 

2^. — The main evolutionary line of the early hexapods led from 
the opisthogoneate branch of the Protosymphyla directly into the 
Machilidae, since in this family are best preserved the coxal accessory 
structures of the symphylids (fig. 52 I) along with the normal ecto- 
gnathous mouth parts. Moreover, it was in the ancestors of the 
Machilidae that the characteristic ovipositor of the hexapods had its 
inception, and, therefore, from the machilid line have been evolved 
the Lepismatidae and the Pterygota. The common ancestors of these 
last two groups developed two special features in the head structure. 
One was the acquisition of a secondary anterior articulation of the 
mandible on the cranium, giving the jaw a hinge movement on a longi- 
tudinal axis, which brought about a reorganization of the mandibular 
musculature, giving the principal function of abduction and adduction 
to the dorsal muscles, and reducing the ventral muscles to a condition 
of such little importance that they have completely disappeared in 
the higher Pterygota. The other feature was the development of the 
endocranial framework known as the tentorium, characteristically 
present in Lepismatidae and Pterygota, but foretokened in Machilidae. 
The tentorium is evidently a product of the hypopharyngeal apodemes 
and of a transverse bar developed in the back of the head from lateral 
invaginations. Both structures are present in Machilidae, but are 
not united. In Lepismatidae the anterior apodemes are reflected 
directly from the cranial margins and are united posteriorly with the 
transverse bar, producing a typical tentorium. In the Pterygota the 
roots of the anterior arms take a submarginal position on the cranium, 
and in higher forms they have migrated to the facial aspect of the head. 


The hexapod structure, with the locomotor function centered in 
the thorax, apparently gave little if any advantage over the polypod 
structure for ordinary terrestrial life, but it furnished a condition 
particularly fitted for the development of wings. Hence, with the 
appearance of alar lobes on the thorax, the evolution of these lobes 
into organs of flight was readily accomplished, and the pterygote 
insects quickly achieved a great superiority over the other arthropods. 
While there is much to suggest that the winged insects are most 
closely related to the apterygote thysanurans, their direct origin from 
the latter is questionable. It is difficult to explain, for example, how 
it comes about that the pterygote Ephemeroptera and Dermaptera 
have paired genital openings while secondary median ducts are already 
established in the Thysanura, with openings on the same segments as 
in the higher Pterygota. 

2/^.. — The Chilopoda are the conservatives among the arthropods ; 
they are the least-modified descendants of the Protomyriapoda, and 
in certain phases of their embryogeny they still follow the course of 
development in the Onychophora. The gnathal appendages are prob- 
ably more generalized than in any other of the Mandibulata; though 
the bases of the mandibles are deeply sunken into pouches of the head 
wall, they have strongly musculated lacinial lobes (fig. 53 E, F), and 
the two maxillary appendages (C) are but little modified except by 
reduction of the telopodites and a partial union of the coxopodites. 
The suspensorial sclerites of the hypopharynx maintain connections 
with the cranial margins, and bear the apodemes on which the ventral 
muscles of the gnathal appendages are attached. The characteristic 
specialization of the chilopods is the conversion of the first legs into 
a pair of poison claws (B). Most of the other body appendages 
retain the structure of simple 7-segmented legs, though at the base 
of each is an extensive subcoxal sclerotization suggestive of that in 
the insect thorax. The last two pairs of legs are reduced and modified 
to serve as genital accessories, and consequently there are no terminal 
cerci. Styli and eversible vesicles are absent. The genital opening is 
always on the last somite before the telson, but since the total number 
of somites is variable, the genital segment may be a quite different 
somite in different chilopod groups. Anamorphic postembryonic 
development persists in some forms, while in others segmentation 
is complete at hatching. 

2^. — Evolution may be accepted as a fact, but the true history of 
phylogeny can never be demonstrated. Though the main branches of 
the genealogic tree of any major group of animals are fairly evident, 
an endeavor to follow in detail the phylogenetic connections between 
more closely related forms invariably leads into a maze of difficulties. 



VOL. 97 




a ^ 

c 5 

•j2 c^ 

5; "O 

^ o 



b/J rs 





for it is seldom found that all characters will fit into a scheme of 
relationship that attempts to relate every feature in one form with a 
similar feature in another. It must be recognized that various struc- 
tural adaptations have been often independently developed in approxi- 
mately the same way. A successful adaptation will be equally valuable 
in many groups, and it is, therefore, not surprising that an adaptive 
structure should independently recur either in distantly related or 
in closely related groups. To distinguish between such structures 
and those that have had an identical origin, however, is one of the 
most uncertain tasks of the phylogeneticist, but the very condition of 
uncertainty injects into the study of phylogeny the element of per- 
sonal opinion which gives to phylogeny that controversial status by 
which it never lacks in interest. Every biologist must have a working 
creed of phylogeny, but he should not too implicitly believe its tenets. 

Baden, V. 

1936. Embryology of the nervous system in the grasshopper, Melanoplus 

differentialis (Acrididae; Orthoptera). Journ. Morph., vol. 60, 

pp. 159-188, 4 pis. 
Balfour, F. M. 

1880. Notes on the development of the Araneina. Quart. Journ. Micr. 

Sci., vol. 20, pp. 1-23, pis. 19-21. 
1883. The anatomy and development of Peripatus capensis. Quart. Journ. 

Micr. Sci., vol. 23, pp. 213-259, pis. 13-20. 
Benham, W. B. S. 

1885. Description of the muscular and endoskeletal systems of Limulus. 

Trans. Zool. Soc. London, vol. 11, pp. 314-338, pis. 72-76. (See 

Lankester, Benham, and Beck, 1885.) 
Bergh, R. S. 

1888. Zur Bildungsgeschichte der Excretionsorgane bei Criodrilus. Arb. 

Zool.-Zoot. Inst. Wiirzburg, vol. 8, pp. 223-248, pis. 13, 14. 
1890. Neue Beitrage zur Embryologie der Anneliden. I. Zur Entwicklung 

und Differenzirung des Keimstreifens von Lumbricus. Zeitschr. 

wiss. Zool., vol. 50, pp. 469-526, pis. 19-21. 
1899. Nochmals iiber die Entwicklung der Segmentalorgane. Zeitschr. 

vi^iss. Zool., vol. 66, pp. 435-449, pi. 30. 

1902. The early development of Lepas. A study of cell-lineage and germ- 
layers. Bull. Mus. Comp. Zool., vol. 40, pp. 61-144, 12 pis. 
BiNARD, Alice, and Jeener, R. 

1926. Recherches sur la morphologic du systeme nerveux des Annelides. 

I. Nereis. Bull. Acad. R. Belgique, Classe Sci., sen 5, vol. 12, 

pp. 437-448, 6 figs. 

1928. Sur I'existence de cavites coelomiques dans le segment palpaire des 

Polychetes. Recueil Inst. Zool. Torley-Rousseau, vol. i, pp. 173- 

178, I fig- 

1929. Morphologic du lobe preoral des Polychetes. Recueil Inst. Zool. 

Torley-Rousseau, vol. 2, pp. 117-240, 45 text figs., pis. 5-12. 



1902. Sur I'organisation, le developement et les affinites du Peripatopsis 

Blainvillei Gay-Gervais. Zool. Jahrb., Suppl. vol. 5 (Fauna 
Chilensis, vol. 2), pp. 675-730, pis. 20-22. 
1905, 1907. Monographic des Onychophores. Ann. Sci. Nat., Zool. ser. 9, 
vol. 2, pp. 1-383, pis. 1-13; vol. 5, pp. 61-318, 191 figs. 
Brace, Edith M. 

1901. Notes on Aeolosoma tenebrarum. Journ. Morph., vol. 17, pp. 177-184, 
pi. 21. 
Brauer, a. 

1895. Beitrage zur Kenntnis der Entwicklungsgeschichte des Skorpions. II. 
Zeitschr. wiss. Zool., vol. 59, pp. 351-435, 20 text figs., pis. 21-25. 
Bruntz, L. 

1903. Sur la presence des reins labiaux et d'un organ phagocytaire chez 

les Diplopodes. C. R. Acad. Sci. Paris, vol. 136, pp. 57-59. 
1908. Les reins labiaux et les glandes cephaliques des Thysanoures. Arch. 
Zool. Exp. Gen., vol. 9, pp. 195-238, pis. 2, 3. 
Burger, O. 

1891. Beitrage zur Entwicklungsgeschichte der Hirudineen. Zur Embry- 
ologie von Nephelis. Zool. Jahrb., Anat., vol. 4, pp. 697-738, 
pis. 41-43. 
1894. Neue Beitrage zur Entwicklungsgeschichte der Hirudineen. (Hirudo 
and Aulastomum). Zeitschr. wiss. Zool., vol. 58, pp. 440-459, 
pis. 26-28. 
Buxton, B. H. 

1913. Coxal glands of the arachnids. Zool. Jahrb., Supl. 14, pp. 231-282, 

pis. 1-43- 
1917. Notes on the anatomy of arachnids. Journ. Morph., vol. 29, pp. 1-31, 
8 text figs., 3 pis. 
Bychowsky, a. 

1921. t)ber die Entwicklung der Nephridien von Clepsine sexoculata Berg- 
mann. Rev. Suisse Zool., vol. 29, pp. 41-131, 16 text figs., pis. 1-5. 
Camerano, L. 

1897. Sulla striatura traversale dei muscoli delle mandibole negli Onicofori. 
Atti R. Accad. Sci. Torino, vol. 33, pp. 589-593. 
Cannon, H. G., and Manton, S. M. 

1927. Notes on the segmental excretory organs of Crustacea. I-IV. Journ. 

Linn. Soc. London, vol. 2>^, pp. 439-456, 7 figs. 
Chickering, a. M. 

1937. Evolution in spiders. 38th Ann. Rep., Michigan Acad. Sci., Arts, 
Lett., pp. 22-51. 
Child, C. M. 

1900. The early development of Arenicola and Sternaspis. Arch. Ent- 
wicklmech., vol. 9, pp. 587-723, pls. 21-25. 
Crampton, G. C. 

1928. The evolution of insects, chilopods, diplopods, Crustacea, and other 

arthropods indicated by a study of the head capsule. Canadian Ent., 
vol. 60, pp. 129-141, pis. 8-12. 
1928. The evolution of the head region in lower arthropods and its bearing 
upon the origin and relationships of the arthropodan groups. 
Canadian Ent., vol. 60, pp. 284-301, pis. 20, 21, 22. 



1926. L'entonnoir vibratile de la nephridie des Peripates. Ann. Soc. R. 

Zool. Belgique, vol. 56, pp. 13-17. 
Curtis, W. C, and Guthrie, AIary J. 

1927. Textbook of general zoology. New York and London. 
Dakin, W. J. 

1920. Fauna of Western Australia. III. Further contributions to the study 

of the Onychophora. Proc. Zool. Soc. London, vol. for 1920, 

PP- 367-389, 5 pis. 

1921. The eye of Peripatus. Quart. Journ. Micr. Sci., vol. 65, pp. 163-171, 

3 text figs., pi. 7. 
Dautert, E. 

1929. Die Bildung der Keimblatter von Paludina vivipara. Zool. Jahrb., 
Anat., vol. 50, pp. 433-496, 33 figs. 
EisiG, H. 

1899. Zur Entwicklungsgeschichte der Capitelliden. Mitt. Zool. Sta. 
Neapel, vol. 13, pp. 1-292, pis. 1-9. 
Evans, R. 

1902. On the Malayan species of Onychophora. Part IL The development 
of Eoperipatus weldoni. Quart. Journ. Micr. Sci., vol. 45, pp. 41-88, 
pis. 5-9. 
Fahlander, K. 

1938. Beitrage zur Anatomie und Systematischen Einteilung der Chilopoden. 
Zool. Bidrag Uppsala, vol. 17, pp. 1-148, 36 text figs., 18 pis. 
Faussek, V. 

1899. tJber die physiologische Bedeutung des Coeloms. Trav. Soc. Imp. 
Nat. St. Petersburg, vol. 30, Livr. i, pp. 40-57 (in Russian), 83-84 
(resume in German). 
1901. Untersuchungen iiber die Entwicklung der Cephalopoden (VII, Was 
is das Colom?). Mitt. Zool. Station Neapel, vol. 14, pp. 83-237, 
pis. 6-10. 
Fedorov^', B. 

1926. Zur Anatomie des Nervensystems von Peripatus. I. Das Neurosomit 
von Peripatus tholloni Bouv. Zool. Jahrb., Anat., vol. 48, pp. 273- 
310, 4 text figs., pis. 10, II. 
1929. Zur Anatomie des Nervensystems von Peripatus. II. Das Nerven- 
system des vorderen Korperendes und seine Metamerie. Zool. 
Jahrb., Anat., vol. 50, pp. 279-332, pis. 6, 7. 
Glen, Edith H. 

1919. A revision of certain points in the early development of Peripatus 
capensis. Quart. Journ. Micr. Sci., vol. 63, pp. 283-292, pi. 20. 
Goodrich, E. S. 

1898-1900. On the nephridia of the Polychaeta. Parts I-III. Quart. Journ. 
Micr. Sci., vol. 40, pp. 185-195, pis. 6-9; vol. 41, pp. 439-457, 
pis. 32-35; vol. 43, pp. 699-748, pis. 37-42. 
1912. Nerilla an archianellid. Quart. Journ. Micr. Sci., vol. 57, pp. 397- 
42s, pis. 38-41. 
Guilding, L. 

1826. Molusca Carribeana. 2. An account of a new genus of Mollusca. 
Zool. Journ., vol. 2, pp. 437-444, pi. 14. 



1930. Anatomische Studien iiber die Polychaten-Familien Amphinomidae 
und Euphrosynidae. Zool. Bidrag Uppsala, vol. 12, pp. 305-471, 
64 text figs., 36 pis. 
Hacker, V. 

1895. Die spjitere Entwicklung der Polynoe-Larve. Zool. Jahrb., Anat., 
vol. 8, pp. 245-288, pis. 14-17. 
Hamaker, J. I. 

1898. The nervous system of Nereis virens Sars. Bull. Mus. Comp. Zool., 
vol. 32, pp. 89-124, 5 pis. 
Hansen, H. J. 

1930. Studies on Arthropoda HI. Copenhagen. 
Hanstrom, B. 

1926. Eine genetische Studie iiber die Augen und Sehzentren von Turbel- 

larien, Anneliden und Arthropoden. K. Svenska Vetenskapsakad. 
Handl., 3d ser., vol. 4, no. i, 176 pp., 80 figs. 

1927. Das zentrale und periphere Nervensystem des Kopflappens einiger 

Polychaten. Zeitschr. Morph. Okol. Tiere, vol. 7, pp. 543-596, 
27 figs. 

1928. Vergleichende Anatomic des Nervensystems der wirbellosen Tiere. 


1929. Weitere Beitrage zur Kenntnis des Gehirns und der Sinnesorgane 

der Polychaten. Zeitschr. Morph. Okol. Tiere, vol. 13, pp. 329-358, 

27 figs. 
1935. Bemerkungen iiber das Gehirn und die Sinnesorgane der Ony- 

chophoren. Lunds Univ. Arssk. N. F. Avd. 2, vol. 31, no. 5, 38 pp., 

22 figs. 
Hatschek, B. 

1877. Die Embryonalentwicklung und Knopsung von Pedicellina echinata. 

Zeitschr. wiss. Zool., vol. 29, pp. 502-549, pis. 28-30. 
1888-91. Lehrbuch der Zoologie. Jena. 
1894. Uber den gegenwartigen Stand der Keimblattertheorie. Verh. Deut- 

schen Zool. Ges. 3d Jahresvers., pp. 11-23. 
Heathcote, F. G. 

1886. The early development of Julus terrestris. Quart. Journ. Micr. Sci., 

vol. 26, pp. 449-470, pis. 23, 24. 
Heider, K. 

1925. Uber Eunice. Systematisches, Kiefersack, Nervensystem. Zeitschr. 

wiss. Zool., vol. 125, pp. 55-90, 16 figs. 
Hempelmann, F. 

191 1. Zur Naturgeschichte von Nereis dumerilii Aud. et Edwr. Zoologica. 
Orig.-Abhandl. Gesamtg. Zool., vol. 62, 135 pp., 4 pis. 

1 93 1. Archiannelida and Polychaeta. Kiikenthal and Krumbach's Hand- 

buch der Zoologie, vol. 2, 2d half. 
Henriksen, K. L. 

1926. The segmentation of the trilobite's head. Saert. Meddel. Dansk Geol. 

For., vol. 7, pp. 1-32, 27 figs. 
1929. Contribution to the interpretation of the cephalic segments of Arthrop- 
oda. Trans. 4th Internat. Congr. Ent. (Ithaca, N. Y.), pp. 589-594. 


Hertwig, O. and R. 

1882. Die Colomtheorie, Versuch einer Erklarung des niittleren Keim- 
blattes. Jenaische Zeitschr. Naturwiss., vol. 15 (N. F. vol. 8), 
pp. 1-150, 3 pis. 
Hess, W. N. 

1925. Nervous system of the earthworm, Lumbricus terrestris L. Journ. 
Morph., vol. 40, pp. 235-259, 7 figs. 
Hesse, R. 

1899. Untersuchungen iiber die Organe der Lichtempfindung bei niederen 
Thieren. V. Die Augen der polychaten Anneliden. Zeitschr. w^iss. 
Zool., vol. 65, pp. 446-516, pis. 22-26. 
Hesse, R., Allee, W. C. and Schmidt, K. P. 

1937. Ecological animal geography. New York and London. 
Heymons, R. 

1897. Mittheilungen iiber Segmentirung und den Korperbau der Myriapoden. 
Sitzungsbr. K. Preuss. Akad. Wiss. Berlin, vol. for 1897, pp. 915- 
923, 2 figs, 
looi. Die Entwicklungsgeschichte der Scolopender. Zoologica. Orig.- 

Abhandl. Gesammtg. Zool., vol. 33, 244 pp., 42 text figs., 8 pis. 
1928. Uber Morphologic und verwandtschaftliche Beziehungen des Xenusion 
auerswaldae Pomp, aus dem Algonkium. Zeitschr. Morph. Okol. 
Tiere, vol. 10, pp. 307-329, 7 figs. 
Holmgren, N. 

1916. Zur vergleichenden Anatomic des Gehirns von Polychaeten, Ony- 
chophoren, Xiphosuren, Arachniden, Crustaceen, Myriopoden und 
Insekten. K. Svenska Vetenskapsakad. Handl., vol. 56, no. i, 
303 pp., 12 pis. 
Hutchinson, G. E. 

1930. Restudy of some Burgess shale fossils. Proc. U. S. Nat. Mus., 
vol. 78, pp. 1-24, pi. I. 
Imms, a. D. 

1936. The ancestry of insects. Trans. Soc. British Ent., vol. 3, pp. 1-32, 
II figs. 


1928. Die Entwicklung der Larvalsegmente bei den Anneliden. Zeitschr. 

Morph. Okol. Tiere, vol. 10, pp. 62-161, 40 text figs., pis. 1-7. 
1933- Die embryonale Entwicklung von Limulus moluccanus. Zool. Jahrb., 
Anat., vol. 56, pp. 163-348, 78 text figs., pis. 1-3. 
Janet, C. 

1899. Essai sur la constitution morphologique de la tete de I'insecte, 74 pp., 
7 pis. Paris. 
Jaworowski, a. 

1 89 1. tJber die Extremitaten bei den Embryonen der Arachniden und In- 
secten. Zool. Anz., vol. 14, pp. 164-169, 173-176, 4 figs. 
Jeener, R. 

1928. Recherches sur le systeme neuromusculaire lateral des Annelides. 
Recueil Inst. Zool. Torley-Rousseau, vol. i, pp. 99-121, pi. 3. 
Kennel, J. von 

1884, 1888. Entwicklungsgeschichte von Peripatus Edwardsii Blanch, und 
Peripatus torquatus n. sp. Arb. Zool.-Zootom. Inst. Wiirzburg, 
vol. 7, pp. 95-229, pis. 5-11 ; vol. 8, pp. 1-93, pis. 1-6. 



1893. On the development of Limulus longispina. Journ. Coll. Sci. Imp. 

Univ. Japan, vol. 5, pp. S3-100, pis. S-ii. 

1894. Note on the coelomic cavity of a spider. Journ. Coll. Sci. Imp. Univ. 

Japan, vol. 6, pp. 287-296, pi. 10. 
Kleinenberg, N. 

1886. Die Entstehung des Annelids aus der Larve von Lopadorhynchus. 
Zeitschr. wiss. Zool., vol. 44, pp. 1-227, pis. 1-16. 
KowALEvsKY, A., and Marion, A. F. 

1883. Documents pour I'histoire embryogenique des Alcyonaires. Ann. Mus. 
Hist. Nat. Marseille, Zool., vol. i, no. 4, 50 pp., 5 pis. 


1935. Beitrag zur Entwicklungsgeschichte von Alcippe lampas Hancock. 
Zeitschr. Morph. Okol. Tiere, vol. 29, pp. 45-78, 24 figs. 
Lalicker, C. G. 

1935. Larval stages of trilobites from the Middle Cambrian of Alabama. 
Journ. Paleontol., vol. 9, pp. 394-399, pl- 47- 
Lameere, a. 

1926. Les organes segmentaires des Polychetes. Ann. Soc. R. Zool. 
Belgique, vol. 56, pp. 113- 11 9. 
Lankester, E. R., Benham, W. B. S., and Beck, E. J. 

1885. On the muscular and endoskeletal systems of Limulus and Scorpio, 
with some notes on the anatomy and generic characters of scorpions. 
Trans. Zool. Soc. London, vol. 11, pp. 311-384, P^s. 72-83. 
LiLLIE, R. S. 

1906. The structure and development of the nephridia of Arenicola cristata 
Stimpson. Mitt. Zool. Sta. Neapel, vol. 17, pp. 341-405, pis. 22-25. 
Malaquin, a. 

1924. Les glandes genitales et les cellules sexuelles primordiales chez 

I'Annelide Salmacina Dysteri (Huxley). La genese des ebauches 
genitales par des gonocytes extracoelomiques. C. R. Acad. Sci. 
Paris, vol. 179, pp. 1348-1351. 
1924 a. Les glandes genitales et les cellules sexuelles primordiales chez 
I'Annelide Salmacina Dysteri (Huxley). La filiation des cellules 
sexuelles (gonocytes) ; leur origine dans le bourgeon caudal. 
C. R. Acad. Sci. Paris, vol. 179, pp. 1636-1639. 

1925. La segregation, au cours de I'ontogenese, de deux cellules sexuelles 

primordiales, souches de la lignee germinale, chez Salmacina Dysteri 
(Huxley). C. R. Acad. Sci. Paris, vol. 180, pp. 324-327, i fig- 
Manton, S. M. 

1928. On the embryology of a mysid crustacean, Hemimysis lamornae. 

Philosoph. Trans R. Soc. London, ser. B, vol. 216, pp. 363-463, 

32 text figs., pis. 21-25. 
1930. Notes on the segmental excretory organs of Crustacea. V. On the 

maxillary glands of the Syncarida. Journ. Linn. Soc. London, 

Zool., vol. 37, pp. 467-472, 9 figs. 
1934. On the embryology of the crustacean Nebalia bipes. Philos. Trans. 

R. Soc. London, ser. B, vol. 223, pp. 163-238, 17 text figs., pis. 20-28. 
Marcus, E. 

1929. Zur Embryologie der Tardigraden. Zool. Jahrb., Anat., vol. 50, 

pp. 333-384, 31 text figs., pl. 8. 


Mellanby, Helen 

1936. The later embryology of Rhodnius proHxus. Quart. Journ. Micr. 
Sci., vol. 79, pp. 1-42, 22 text figs. 
Metschnikoff, E. 

1882. Vergleichend-embryologische Studien. Zeitschr. wiss. Zool., vol. 36, 
pp. 432-444- pl- 28. 
Meyer, A. 

1929. Die Entwicklung der Nephridien und Gonoblasten bei Tubifex rivu- 

lorum Lam. nebst Bemerkungen zum natiirlichen System der Oli- 

gochaten. Zeitschr. wiss. Zool., vol. 133, pp. 517-562, 31 figs. 
Meyer, E. 

1887. Studien iiber den Korperbau der Anneliden. I-III. Mitt. Zool. Sta. 

Neapel, vol. 7, pp. 592-741, pis. 22-27. 
1891. Die Abstammung der Anneliden. Der Ausprung der Metamerie und 

die Bildung des Mesoderms. Biol. Centralbl., vol. 10, pp. 296-308. 
1901. Studien iiber den Korperbau der Anneliden. V. Mitt. Zool. Sta. 

Neapel, vol. 14, pp. 247-585, pis. 12-17. 


1928. Oligochaeta. Kiikenthal and Krumbach's Handbuch der Zoologie, 
vol. 2, 2d half. 
Michel, A. 

1898. Recherches sur la regeneration chez les Annelides. Bull. Sci. France 
et Belgique, vol. 31, pp. 245-417, pis. 13-19. 
Needham, a. E. 

1937- Some points in the development of Neomysis vulgaris. Quart. Journ. 
Micr. Sci., vol. 79, pp. 559-588, pis. 23, 24. 
Nelson, J. A. 

1904. The early development of Dinophilus : a study in cell-lineage. Proc. 

Acad. Nat. Sci. Philadelphia, vol. 56, pp. 687-737, pis. 43-48. 
1907. The morphology of Dinophilus conklini n. sp. Proc. Acad. Nat. Sci. 

Philadelphia, vol. 59, pp. 82-143, pls. 12, 13. 
1915. The embryology of the honey bee. Princeton Univ. Press. 
Packard, A. S. 

1898. A text-book of entomology. New York. 
Patten, W., and Hazen, A. P. 

1900. The development of the coxal gland, branchial cartilages, and genital 
ducts of Limulus polyphemus. Journ. Morph., vol. 16, pp. 459-502, 
pis. 22-28. 
Patten, W., and Redenbaugh, W. A. 

1900. Studies on Limulus. L The endocrania of Limulus, Apus, and Mygale. 
Journ. Morph., vol. 16, pp. 1-26, pis. i, 2. 
Penners, a. 

1924. Die Entwicklung des Keimstreifs und die Organbildung bei Tubifex 
rivulorum Lam. Zool. Jarb., Anat., vol. 45, pp. 251-308, 32 te.xt 
figs., pis. 12-14. 

1930. Entwicklungsgeschichtliche Untersuchungen an marinen Oligochaten. 

IL Furchung, Keimstreif und Keimbahn von Pachydrilus (Lumbri- 
cillus) lineatus Miill. Zeitschr. wiss. Zool., vol. 137, pp. 55-119, 
40 figs. 


Penners, a., and Stablein, A. 

1930. Uber die Urkeimzellen bei Tubificiden (Tubifex rivulorum Lam. und 
Limnodrilus udekemianus Claparede). Zeitschr. wiss. Zool., vol. 
137) PP- 606-626, 13 figs. 
Petrunkevitch, a. 

1922. The circulatory system and segmentation in Arachnida. Journ. 

Morph., vol. 36, pp. 157-188, pis. I, 2. 
1933- An inquiry into the natural classification of spiders, based on a study 
of their internal anatomy. Trans. Connecticut Acad. Arts Sci., 
vol. 31, pp. 299-389, 13 pis. 
Pflugfelder, O. 

1932. t)ber den feineren Bau der Augen freilebender Polychaten. Zeitschr. 

wiss. Zool., vol. 142, pp. 540-586, 24 figs. 
1932 a. tjber den Mechanismus der Segmentbildung bei der Embryonalent- 

wicklung und Anamorphose von Platyrrhacus amauros Attems. 

Zeitschr. wiss. Zool., vol. 140, pp. 650-723, 40 figs. 
Philiptschenko, J. 

1908. Beitrage zur Kenntnis der Apterygoten. II. Uber die Kopfdrusen der 

Thysanuren. Zeitschr. wiss. Zool., vol. 91, pp. 93-1 11, pis. 5, 6. 


1899. Noch ein Paar Kopfhocker bei den Spinnenembryonen. Zooi. Anz., 
vol. 22, pp. 272-273, I fig. 
Pompeckj, J. F. 

1927. Ein neues Zeugnis uralten Lebens. Palaontol. Zeitschr., vol. 9, pp. 
287-313, pl- 5- 
Rabl, C. 

1879. Uber die Entwicklung der Tellerschnecke. Morph. Jahrb., vol. 5, 

pp. 562-660, pis. 32-38. 
1889. Theorie des Mesoderms. Morph. Jahrb., vol. 15, pp. 113-252, pis. 7-10. 
Reynolds, J. M. 

1935. A new contribution to the problem of segmentation in the Arthropoda. 
Amer. Journ. Sci., vol. 30, pp. 373-382. 
Robinson, Margaret 

1907. On the segmentation of the head of Diplopoda. Quart. Journ. Micr. 
Sci., vol. 51, pp. 607-624, pl. 37. 

1937. Studies on the embryology of the African migratory locust, Locusta 
migratoria migratorioides Reiche and Frm. II. Organogeny. 
Philos. Trans. R. Soc. London, ser. B, vol. 227, pp. 175-244, pis. 1-7. 
RusER, Magdalene 

1933. Beitrage zur Kenntnis des Chitins und der Muskulatur des Zicken 

(Ixodidae). Zeitschr. Morph. Okol. Tiere., vol. 27, pp. 199-261, 
39 figs. 


1895. t)ber Bau und Entwicklung des Endosternits der Arachniden. Zool. 

Jahrb., Anat., vol. 8, pp. 191-216, pis. 10, 11. 
1906. Uber die Entwicklung von Thelyphonus caudatus (L.), vergleichen 

mit derjenigen einiger anderer Arachniden. Zeitschr. wiss. Zool., 

vol. 81, pp. 1-95, pis. 1-8. 



1925. Zur Systematik der Gattung Nerilla. Zool. Anz., vol. 62, pp. 229- 
234, 4 figs. . 
Schneider, K. C. 

1902. Lehrbuch der vergleichenden Histologie. Jena. 

1936. Trilobita, Xiphosura, Acarina. Eine morphologische Untersuchungen 
uber Plangleichheit zwischen Trilobiten und Spinnentieren. Zeit- 
schr. Morph. Okol. Tiere, vol. 32, pp. 181-226, 39 figs. 
1938. Uber rein glabellare Karapaxbildungen bei Milben und iiber die 
Umgestaltung des Vorderkorpers der Ixodidea als Folge der 
Gnathosomaentstehung. Zeitschr. Morph. Okol. Tiere, vol. 34, pp. 
135-149, 18 figs. 


1888. On the early stages of the development of a South American species 
of Peripatus. Quart. Journ. Micr. Sci., vol. 28, pp. 343-363, pl- 24. 
Sedgwick, A. 

1884. On the origin of metameric segmentation and some other morpho- 

logical questions. Quart. Journ. Micr. Sci., vol. 24, pp. 43-82, 
pis. 2, 3. 

1885. The development of Peripatus capensis. Part I. Quart. Journ. Micr. 

Sci., vol. 25, pp. 449-468, pis. 31, 32. 

1886. The development of the cape species of Peripatus. Part II. Segmen- 

tation of the ovum and formation of the layers. Quart. Journ. Micr. 
Sci., vol. 26, pp. 175-212, pis. 12-14. 

1887. The development of the cape species of Peripatus. Part III. The 

changes from stage A to stage F. Quart. Journ. Micr. Sci., vol. 27, 
pp. 467-550, pis. 34-37- 

1888. The development of the cape species of Peripatus. Part IV. The 

changes from stage G to birth. Quart. Journ. Micr. Sci., vol. 28, 
PP- 373-396, pis. 26-29. 
Shearer, C. 

1912. The problem of sex determination in Dinophilus gyrociliatus. Quart. 
Journ. Micr. Sci., vol. 57, pp. 329-371, pls. 30-34- 
Sheldon, Lilian 

1888, 1889. On the development of Peripatus Novae-Zealandiae. Quart. 
Journ. Micr. Sci., vol. 28, pp. 205-237, pis. 12-16; vol. 29, pp. 283- 
293, pis. 25, 26. 


1933. Sulle appendici del capo degli "Japygidae" (Thysanura entotropha) 

e rispettivo confronto con quelle dei Chilopodi, dei Diplopodi e dei 

Crostacei. 5th Congr. Internat. Entom., Paris, 1932, pp. 329-343, 

7 figs. 
Snodgrass, R. E. 

1933. Morphology of the insect abdomen. Part II. The genital ducts and 

the ovipostor. Smithsonian Misc. Coll., vol. 89, no. 8, 148 pp., 

58 figs. 
1936. Morphology of the insect abdomen. Part III. The male genitalia. 

(Including arthropods other than insects). Smithsonian Misc. Coll., 

vol. 95, no. 14, 96 pp., 29 figs. 



191 1. Uber eine neue Ctenodrilusart und ihre Vermehrung. Zeitschr. wiss. 

Zool., vol. 97, pp. 547-603, pis. 27-29. 


1923. Recherches sur I'embryogenie des Crustaces decapodes de la sous- 
famille des "Palcmoninae." Bull. Biol. France et Belgique, Suppl., 
vol. 5, 234 pp., 5 pis. 
1933- Le blastopore et la question du "prostomium" chez les Crustaces. 
Assoc. Frangaise Avanc. Sci., 57th Sess., Chambery, pp. 321-327, 
2 figs. 
Staff, F. 

1910. Organogenetische Untersuchungen iiber Criodrilus lacuum (Hoff- 
nistr.). Arb. zool. Inst. Wien, vol. 18, pp. 227-256, 2 pis. 
Stephenson, J. 

1930. The Oligochaeta. Oxford. 
Storch, O. 

1912. Zur vergleichenden Anatomic der PolychJiten. Verhandl. zool.-botan. 

Ges. Wien, vol. 62, pp. 81-97, 4 figs. 

1913. Vergleichend-anatomische Polychatenstudien. Sitzungsb. Akad. wiss. 

Wien, Math.-natur. Kl., vol. 122, Ab. i, pp. 877-988, 3 pis. 
Stummer-Traunfels, v. R. R. 

1903. Beitrage zur Anatomic und Histologic der Myzostomen. I. Alyzostoma 
asteriae Marenz. Zeitschr. wiss. Zool., vol. 75, pp. 495-595, 
pis. 34-38. 
Surface, F. M. 

1907. The early development of a polyclad Planoccra inquilina Wh. Proc. 
Acad. Nat. Sci. Philadelphia, vol. 59, pp. 514-559, pis. 35-40. 
Tannreuther, G. W. 

1915. The embryology of Bdellodrilus philadelphicus. Journ. Morph., vol. 
26, pp. 143-216, 26 text figs., 8 pis. 


1930. The evolution of the class Insccta. Roy. Soc. Tasmania, Papers and 

Proc, 1930, 89 pp., 19 figs. 
1932. The evolution of the class Insecta. Amer. Journ. Sci., vol. 23, pp. 

1935. Evolution of progoneate and opisthogoneate types in the myriapod- 

hexapod group of terrestrial Arthropoda. Amer. Journ. Sci., vol. 

30, pp. 438-449. 


1903. The early embryology of Thalassema mellita (Conn.) Ann. New 

York Acad. Sci., vol. 14, pp. 165-246, 2 pis. 
Treadwell, a. L. 

1901. The cytogeny of Podarke obscura Verrill. Journ. Morph., vol. 17, 

PP- 399-486, pis. 36-40. 
Versluys, J., and Demoll, R. 

1920. Die Verwandtschaft der Merostomata mit den Arachnida und den 

anderen Abteilungen der Arthropoda. Proc. K. Akad. Wetens. 

Amsterdam, vol. 23, pp. 739-765. 
1923. Das Limulus-Problem. Die Verwandtschaftsbeziehungen der Mero- 

stomen u. Arachnoideen unter sich u. mit anderen Arthropoden. 

Ergebnisse u. Fortschritte der Zoologie, vol. 5, pp. 67-388, 123 figs. 



1935- Die Entwicklung der Antenneiidriise der Mysideen. Zeitschr. Morph. 
Okol. Tiere., vol. 29, pp. 481-506, 20 figs. 
Walcott, C. D. 

1908. Mount Stephen rocks and fossils. Canadian Alpine Journ., vol. i, 
no. 2, pp. 232-248, 4 pis. 

1910. Olenellus and other genera of the Mesonacidae. Smithsonian Misc. 

Coll., vol. 53, no. 6, pp. 231-422, 22 pis. 

191 1. Cambrian geology and paleontology. II. No. 5. Middle Cambrian 

annelids. Smithsonian Misc. Coll., vol. 57, no. 5, pp. 110-144, 

pis. 18-23. 
1916. Cambrian geology and paleontology. III. No. 5, Cambrian trilobites. 

Smithsonian Misc. Coll., vol. 64, no. S, pp. 303-456, pis. 45-67. 
1931. Addenda to descriptions of Burgess shale fossils. Smithsonian Misc. 

Coll., vol. 85, no. 3, 46 pp., 23 pis. 
Warburg, Elsa. 

1925. The trilobites of the Leptaena limestone in Dalarne. Bull. Geol. 

Inst. Univ. Uppsala, vol. 17, pp. 1-450, 11 pis. 
Whitman, C. O. 

1886. The germ-layers of Clepsine. Zool. Anz., vol. 9, pp. 171-176, i fig. 


1926. Zur Kenntnis der Anatomic und Entwicklungsgeschichte der Stab- 

heuschrecke Carausius morosus Br. III. Entwicklung und Organo- 
genese der Colomblasen, 206 pp., 86 figs. Zool.-vergl. anat. Inst. 
Univ. Zurich. 
Wilson, E. B. 

1889. The embryology of the earthworm. Journ. Morph., vol. 3, pp. 387- 
462, pis. 16-22. 


1905. Wurm "kopf," Wurmrumpf und Trochophora. Bemerkungen zur 
Entwicklung und Ableitung der Anneliden. Zool. Anz., vol. 28, 
PP- 273-322, 24 figs. 
Zacher, F. 

1933. Onychophora. Kiikenthal and Krumbach's Handbuch der Zoologie, 
vol. 3, 2d half, pp. 79-138, figs. 57-122. 


1898. liber den derzeitigen Stand der Colomfrage. Verh. Deutschen Zool. 

Ges., 8th Jahresvers., pp. 14-78, 15 figs. 
Zilch, A. 

1936. Zur Frage des Flimmerepithels bei Arthropoden. Zeitschr. wiss. 

Zool., vol. 148, pp. 89-132, II figs. 




(With Six Plates) 



Assistant Curator of Archeology 

U. S. National Museum 

(Publication 3484) 



OCTOBER 19, 1938 




(With Six Plates) 



Assistant Curator of Archeology 

U. S. National Museum 

^sf/ Q<r5P y\^l^" 


(Publication 3484) 



OCTOBER 19,1193 8 

^$e Bovb (g>afttmore (preetf 




Assistant Curator of Archeology, U. S. National Museum 

(With Six Plates) 

When the University of Nebraska Archeological Survey was 
established in 1929, its then director, Dr. W. D. Strong, envisaged 
two primary objectives. The first was a prehminary survey of the 
State, including both surface reconnaissance and sampling excava- 
tions, designed to give a general bird's-eye view of the area as a 
whole. With this was combined a second aim, namely, an effort to 
locate and work such sites as could be definitely identified with villages 
visited and recorded by the early white explorers in eastern Nebraska. 
It was believed that by isolating and clearly defining the archeological 
characteristics of the historic peoples a whole series of sites could 
soon be removed from the category of unknowns ; and furthermore, 
that a comparison of materials so identified with earlier remains in 
the region might open lines of attack which would permit the 
establishing of a time sequence extending "from the known historic 
into the unknown prehistoric." Toward this second objective a 
serious beginning had already been made by A. T. Hill, of Hastings, 
Nebr., who since 1922 had accumulated a considerable quantity of 
archeological materials from sites identified as Pawnee through criti- 
cal study of early nineteenth century maps and narratives. This 
collection, as well as numerous valuable historical leads, was 
promptly made available to Dr. Strong and his coworkers, and it 
became the starting point for the study of Pawnee archeology. In 
this paper it is proposed to review very briefly the methods and some 
results of this approach to prehistory in the Pawnee area. 

It was not chance alone that prompted selection of the Pawnee for 
the first systematic attempt at isolating a historic archeological com- 
plex in Nebraska. Aside from Hill's pioneer labors, consideration 
was given to the fact that this tribe was one of the largest, best known, 
and most powerful in the entire Plains area. Among the semi- 
sedentary so-called village tribes of the Missouri valley, including 
both Caddoan and Siouan groups, probably none shows evidence 

Smithsonian Miscellaneous Collections, Vol. 97, no. 7. 


for a longer occupancy of its historic locale than the Pawnee. 
Furthermore, of all the Nebraska peoples, the Pawnee appear to have 
offered the most effective and prolonged resistance to the host of 
alien practices introduced by the whites and to have retained longest 
their own customs. As to documentation, allusions to the Pawnee 
may be found from almost the very beginnings of recorded European 
penetration into the interior United States, although it is true that 
many of the seventeenth and early eighteenth century sources of 
information leave much to be desired. Prior to about 1800, hazy 
geographical concepts, occasional tribal shiftings, and the often hear- 
say origin of the explorer's observations made impossible the record- 
ing of village locations with the exactness necessary to permit their 
individual identification today. After that date, thanks to the lucid 
narratives and excellent maps of such men as Dulac, Pike, Lewis 
and Clark, Long, and others, the historical record has enabled us to 
correlate with reasonable certainty the native towns with known 
archeological sites. Excavations in sites so identified have revealed 
the distinguishing characteristics of historic Pawnee culture, insofar 
as these include nonperishable material traits. As the term is now 
used in Nebraska prehistory and in this paper, historic Pawnee 
archeology refers to the antiquities from documented village sites 
where the Pawnee are known to have been living in or after circa 
1800.' Needless to say, throughout this period the archeological 
picture can be greatly enriched through the ethnographic observa- 
tions of many of the white travelers. 

During the nineteenth century, the Pawnee villages with but two 
or three apparent exceptions were centered about the confluence of 
the Loup with the Platte River. Both of these streams flow in a 
general easterly direction through broad flat-floored valleys inclosed 
on either side by lofty bluff's. Above the mouth of Shell Creek the 
native towns stood on terraces or second bottoms well out of reach 
of floods; below this point suitable terraces are mostly lacking and 
the sites are situated on the bluffs with the river sweeping past their 
bases. The tree- fringed watercourses are in marked contrast to the 
dry rolling, formerly grass-covered, uplands which lie beyond the 
valley margins. To the natives the latter were suited only for hunting 
and it was the fertile river bottoms, with an abundance of wood, 
water, arable ground, and shelter, that determined the location of 
their villages. 

^ For a discussion of historic Pawnee archeological remains see Wedel, 1936, 
and Strong, 1935, pp. 55-61. 


The extreme limits of the known Pawnee settlements were, to the 
west, near St. Paul on the Loup and Central City on the Platte; to 
the east, downriver, they ran to Leshara or Yutan on the Platte 
(see fig. I for location of all sites discussed herein). Within this 
120-mile stretch of river valley they shifted back and forth as fancy 
or circumstance dictated, leaving it only for their seasonal hunting 
excursions. The exceptions, it may be noted, included two sites on 
the Republican near the Kansas-Nebraska line and one on the Blue 
near Blue Springs, Nebr. That this nineteenth century restriction 
of habitat was in effect long before will become apparent presently 
when certain additional historical and ethnographic facts are con- 
sidered. Here it is desired to add only the observation that all of 
these village sites, in addition to a somewhat decadent aboriginal 
material culture, yield also many articles of iron, copper, brass, and 

Within this same area, but of even more limited distribution, are 
found other sites whereon the native remains are far more abundant, 
of superior quality, and associated with much smaller quantities of 
white contact material. These sites extend along the Platte-Loup 
riverway from Schuyler on the east to the vicinity of Genoa on the 
west, a distance of approximately 50 miles ; they are mostly on the 
north bank, but one is also known on the south side. Generally, the 
sites are large (from 15 to 100 acres or more) and compactly 
arranged ; not infrequently they seem to have been located on bluffs 
or hilltops with an eye to defensibility and in a few instances they 
were further protected by earth walls and ditches. To date about a 
dozen have been placed on record. The sites are particularly abun- 
dant from Monroe westward, where for more than 8 miles remains 
occur almost continuously along the Loup and on the lower portion 
of Beaver Creek. In the aggregate these antiquities cover many 
hundreds of acres, and prior to introduction of modern farming 
operations, innumerable house circles, middens, and artifacts were 
to be found. Because of their occurrence in the very heart of the 
historic Pawnee habitat and since they yielded smaller amounts of 
contact material than the identified nineteenth century Pawnee sites 
while exhibiting many similarities to the latter, it was thought that 
they might prove to be an earlier, if still post-European, phase of 
Pawnee culture. Consequently, in 1931, as a sequel to the study of 
the historic Pawnee, two of these protohistoric ^ sites were partially 

^ Protohistoric sites yield limited amounts of glass and metal trade wares, 
indicating their occupancy, at least in part, since the arrival of Europeans. They 


examined by parties from the University of Nebraska. About 8 weeks 
were devoted to excavation of houses and middens at the Burkett 
site near Genoa and at the Gray- Wolfe site north of Schuyler. All 
but one week of this field-work was in direct charge of the present 
writer, under the supervision of Dr. Strong and with much active 
assistance in the field from Mr. Hill. A detailed description of the 
findings has been published recently by the University, and the 
remains have been assigned to the "Lower Loup Focus of an 
unnamed aspect of the Upper Mississippi Phase." ^ A wealth of 
additional information has since been gathered by Mr. Hill for the 
Nebraska Historical Society at three other protohistoric sites near 
Genoa. This latest work, completed in 1936 and as yet unpublished, 
included the opening of 10 houses, a number of large and prolific 
caches, and the collecting of several thousand artifacts, all at sites 
lying within 4 or 5 miles of the Burkett site. Pending future analysis 
and detailed comparison, it must suffice to say that preliminary exami- 
nations indicate a close similarity between this material and that 
already described in print from the Burkett and Gray- Wolfe sites. 
In passing it may be noted also that extensive surface collections 
from most of the other protohistoric sites in the immediate locality 
diverge in no significant respect. In short, a fairly uniform and 
consistent cultural complex seems to be manifested at the sites 
designated on the map as belonging to the Lower Loup Focus. 

Plistoric archeology in Nebraska received added stimulus in the 
summer of 1935, when Hill explored the large protohistoric Leary 
site on the Nemaha River in the extreme southeastern corner of the 
State. This has been elsewhere described and identified as Oneota. 
Midwestern archeologists are inclined to view the Oneota culture 
in Iowa and adjacent States as possibly early Siouan.* There are 
indications that the Leary site was inhabited contemporaneously with 
or possibly slightly earlier than the known sites of the Lower Loup 

differ from historic sites in that the written records are too general to permit 
their individual identification with villages actually visited by white men. In 
time they antedate 1800. 

^ Dunlevy, 1936, pp. 147-248 (quot. p. 216). A discussion of the placing of 
the Lower Loup Focus in the McKern taxonomic system is beyond the scope 
of this paper. However, it may be pointed out that at least four of the nine 
Upper Mississippi Phase determinants listed by Deuel (F. C. Cole and T. Deuel, 
Rediscovering Illinois, table 2, p. 214, 1936) are unreported from the Lower 
Loup Focus and incidentally from the historic Pawnee as well. The present 
writer regards as debatable the assignment of either complex, or of a hypo- 
thetical aspect which might include both, to the Upper Mississippi Phase. 

*Hill and Wedel, 1936; Griffin, 1937. 


Focus but no documentary record exists as to the tribe which 
inhabited it. It definitely antedates the historic Pawnee sites of the 
nineteenth century. This is of some interest because there are Pawnee 
traditions pointing to early residence of the tribe somewhere in this 
section of southeastern Nebraska, suggesting the possibility of a 
generic connection with the Oneota. 

As regards the relation of these three postcontact archeological 
complexes to one another, dissimilar conclusions have been reached 
by different field and laboratory workers. Strong expressed the 
belief that the sites now labeled collectively as the Lower Loup Focus 
probably represented a very early historic horizon directly ancestral 
to the somewhat simpler and decadent Pawnee culture of the nine- 
teenth century. His use of the term "protohistoric Pawnee" in speak- 
ing of these remains reflects a view with which the present writer 
has elsewhere indicated his general agreement.' . Dunlevy, on the 
other hand, dissenting after her detailed analysis of material from 
two of these sites, was persuaded that the Lower Loup Focus is 
more closely related to the Oneota than to the historic Pawnee." 
Since these differences of viewpoint occur among individuals dealing 
with substantially the same materials, it seems worthwhile to re- 
examine the data on which they rest. 

In the accompanying table the presence or apparent absence of 
traits has been indicated for each of the three cultural complexes 
above mentioned. The traits, totaling 120, have been grouped in 
seven categories which, with exception of ceramics and miscellaneous 
items, are based upon function rather than on form or substance. 
Traits for the historic Pawnee and the Lower Loup Focus have been 
compiled largely but not exclusively from published sources. In the 
absence of complete analyses for the recently worked sites, the data 
therefrom have been incorporated in and added to a check list based 
on the published studies. Actually, this somewhat superficial treat- 
ment involved no changes in the list other than its slight expansion 
to include a larger number of traits. Data on the Oneota Aspect, 
including three Wisconsin variants or foci, have been drawn from a 
list furnished by W. C. McKern, of the Milwaukee Public Museum, 
which has been supplemented by the published report on the Leary 
site in Nebraska. No attempt has been made to weight the various 
elements or' to determine the degree to which a particular trait may 
be present in one or another of the groups. It has not always been 

^ Strong, op. cit., pp. 68, 297 ; Wedel, op. cit., pp. 38-42, 74. 
° Dunlevy, op. cit., p. 216. 


possible to refine the traits as fully as desired, owing to differences 
in terminology in the sources used and to inability to examine all the 
material at first hand. It is believed, however, that the data are 
sufficiently extensive and representative to be strongly indicative of 
trends, at least. 

Table i. — Presence or Absence of Traits in the Historic Paimiee, the Lower 
Loup Focus, and the Oneota Aspect 


Historic Loup Oneota 
Pawnee Focus Aspect 

I. Architecture and Village Complex 

1. Large, intensively occupied sites x x x 

2. Walled or defensively located x x 

3. Numerous outside caches x 


4. Shallow semisubterranean circular earth-covered... x x 

5. Vestibule entrance in east or south x x 

6. Unlined central firepit x x 

7. Bison-skull shrine opposite door x x 

8. Four main central posts x x 

9. More than four central posts x x 

10. One or two rows of widely spaced outer posts x x 

11. Inside caches x x 

12. Numerous small, closely set, slanting wall posts x x 

II. Ceramic Complex 

13. Grit X X 

14. Shell X X 


15. Fine to medium coarse x x x 


16. Flaky X X X 

17. Granular • x 


18. 1-4, softer predominating x 

19. 3-6, 4-5 predominating x x 

Surface finish 

20. Irregularly smoothed x x x 

21. Polished (imperfectly) x x x 

22. Marked by grooved paddle x x 


23. Light to dark gray and buff, dull terra cotta x x x 


24. s\-M inch range x x x 

Lip form 

25. Squared x x x 

26. Rounded x x x 

NO. 7 


Table i. — Presence or Absence of Traits in the Historic Panmee, the Lower 
Loup Focus, and the Oneota Aspect (continued) 


Historic Loup Oneota 
Pawnee Focus Aspect 
Rim form 

27. Plain high direct flaring x x x 

28. Collar or braced x x 

29. Cloistered x 

Neck form 

30. Line of juncture between rim and body x x x 

31. More pronounced x x 


Z2. Broad x x x 

Z2. Round x x x 

34. Oval X X 

Shoulder form 

35. Round X X X 

Basal form 

36. Rounding x x x 

37. Subconical x x 


38. Narrow to broad, flat, straplike, paired x x 

39. Loop X x 

40. Alternate collar tabs form handles x 

41. Multiple X X 


42. Lip X X X 

43. Shoulder area to lip, neck plain x x x 

44. Incised rectilinear parallel line motifs x x x 

45- Opposed series of parallel lines x x x 

46. Herringbone and chevron on rim x x 

47. Concentric pendent chevrons inside rim x 

48. Concentric circle motif and/or cross x 

49. Geometric series of lines and dots x 

50. Trailed or fluted decoration x 


5 1 . Small bowls x x 

52. Small decorated "fishtail" figurines x x 

53. Use of red wash or pseudo-slip x x 

54. Perforated pottery disks ; x . x 

55. Pot lids with handles x 

56. Cut sherds and bisected vessels x 

in. Horticulture and Food-Gathering 

57. Intensive horticulture, with maize, beans, etc x x x 

58. Hoes made of bison scapulae x x x 

59. Wooden mortar x 

60. Stone mortar : irregular, shaped, flattened surface. .. x x x 


Table i. — Presence or Absence of Traits in the Historic Paimiee, the Lower 
Loup Focus, and the Oneota Aspect (continued) 


Historic Loup Oneota 
Pawnee Focus Aspect 

IV. Military and Hunting Complex 

61. Arrowpoints, small triangular unnotched x x x 

62. Knives : diamond-shaped, beveled x x 

63. Knives : oval and/or flake x x x 

64. Scrapers : small to'medium planoconvex x x x 

65. Scrapers: large elliptical quartzite or sandstone.... x x 

66. Drills x x x 

67. Abraders : paired longitudinally grooved sandstone. . x x x 

68. Abraders : amorphous pumice lumps x 

69. Mauls : grooved x x x 

70. Axes : grooved ? x" 

71. Celts : polished diorite or hematite x x 

72. Hammerstones, pitted x x 

72. Adz-shaped elkhorn hide scrapers x 

74. Deerhorn "cylinders" or tapping tools x 

75. Deerhorn tip flakers x x 

76. Deerhorn projectile points, conical, socketed x 

77. Bone projectile points, socketed, square or conical. . . x 

78. Bone projectile points, stemmed x 

79. Bundles of cane (arrourshafts ?) x 

80. Perforated ribs (arrowshaft straighteners) x x x 

81. Notched fleshing tools or grainers x x 

82. Shoulder blade scrapers x 

83. Celtlike antler scrapers x 

84. Metapodial beamers x 

85. Bone fishhooks x 

V. Dress, Textiles, and Adornment 

86. Bison-hair cloth and/or cordage x x 

87. Awls X X X 

88. Eyeleted needles x 

89. Plume holder x 

90. Roach spreader x 

91. Combs X 

92. Bracelets and/or gorgets x x x 

93. Paint bones ( "brushes" ) x x 

94. Polished bone tubes x x x 

95. Rush matting x 

96. Flat polished-bone mat needles x 

97. Twined bags of vegetal material x 

98. Shell ornaments, variously shaped x x x 

VI. Ceremonial Complex 

99. Primary extended burials x 

100. Primary flexed burials x x x 

lOi. Secondary bundle burials x* 

* Rare, probably atypical. 


Table i. — Presence or Absence of Traits in the Historic Pazmiee, the Lower 
Loup Focus, and the Oneota Aspect (continued) 


Historic Loup Oneota 
Pawnee Focus Aspect 

VI. Ceremonial Complex — Continu-cd 

102. Grave furniture x x 

103. Burial in diig pits or caches, x x x 

104. Burial in or under mounds x 

105. Gaming stones ? ; bun-shaped, flat pitted face x 

106. Gypsum crystals, worked x 

107. Shaped balls of crystalline stone (grave finds) x 

108. "Whetstones" (grave finds) x 

109. Pipes of pohshed stone x x x 

I ID. Pipes of clay x 

111. Pipes : elbow-shaped or equal-armed x x x 

112. Pipes : "Siouan" type, stem projects beyond bowl. . . . x 

1 13. Pipes : disk bowl x x 

1 14. Pipes : "Micmac" x 

115. Ornamented animal skulls x x 

VII. Miscellaneous 

116. Incised stone tablets x x x 

117. Bison horn spoons x 

1 18. Tanged mussel shell spoons x 

119. Ulna "picks" x x x 

120. Tally bones (scored ribs) x x 

'80 82 74 
Analysis of Table: Summary 

Total number of traits — 120 

Historic Pawnee has 80, or 66.6 percent of total 
Lower Loup Focus has 82, or 68.3 percent of total 
Oneota Aspect has 74, or 61.6 percent of total 

"Universal" traits — 39, or 32.5 percent of total 

39 universals in 80 historic Pawnee elements 48.8 percent 

39 universals in 82 Lower Loup Focus elements 47.6 percent 

39 universals in 74 Oneota Aspect elements 52.7 percent 

Out of total of 120 traits — 

26 occur only in historic Pawnee and Lower Loup Focus 21.7 percent 

9 occur only in Lower Loup Focus and Oneota Aspect 7-5 percent 

3 occur only in historic Pawnee and Oneota Aspect 2.5 percent 

On basis of 81 nonuniversal traits these percentages become 
respectively 32, 11, and 3.7. 

Traits occurring in only one complex — 

Historic Pawnee 12 

Lower Loup Focus 8 

Oneota Aspect 23 


Analysis of the table shows first that out of the total of 120 
different elements the historic Pawnee and the Lower Loup Focus 
have, respectively, 80 and 82 (66.6 and 68.3 percent), and the Oneota 
Aspect has 74 (or 61.6 percent). Of the 120 traits, furthermore, 
39 are common to all three culture complexes. Since this represents, 
respectively, 48.8, 47.6, and 52.7 percent of those found in each 
complex, it is evident that there is a strong underlying relationship. 
These "universals" include elements in practically all of the categories, 
but occur least commonly under the "Architecture and Villages" 
heading.' As regards specific relationships between any two of the 
three complexes, we find that 26 traits, or 21.7 percent, occur only in 
historic Pawnee and the Lower Loup Focus ; ** 9, or approximately 7.5 
percent, only in the Lower Loup Focus and the Oneota ; " and 3, or 
2.5 percent, only in historic Pawnee and Oneota. Since it is these 
relationships within the defined universe of three which are the 
principal concern here, we may reduce our totals and sharpen the 
above differentiations by omitting the "universal" traits. Thus, using 
the 81 nonuniversals as our basis, the percentages become, respectively, 
32, II, and 3.7. Whichever set of figures is taken, it is apparent that 
the table indicates very nearly three times as many traits in common 
between the historic Pawnee and the Lower Loup Focus (and in no 
other) as in the Lower Loup Focus and the Oneota.^" Evidently the 
suggested connection between the first two complexes, considered on 
purely archeological grounds alone, is considerably closer than that 
between the second pair. This is the more striking in view of the 
previously indicated fact that the Lower Loup Focus flourished at 
the very beginning of European contact and approximately con- 

' The single rectangular earthlodge floor found at the Leary site has not 
been included in the present table since there seems to be general agreement 
among field workers that this type of structure is not characteristic of the 
Oneota. I am inclined to agree with McKern's suggestion that the occurrence 
of earthlodges in the western Oneota sites "may be due to the taking on of foreign 
traits after leaving the area of earlier occupation." (Letter of Oct. 28, 1937.) 

* Including among others nine in architecture, besides such elements as de- 
cided predominance of grit tempering, use of grooved paddle in surfacing pottery, 
small decorated "fishtail" figurines of clay, large elliptical quartzite hide scrapers, 
bone paint "brushes," notched fleshers, ornamented animal skulls (rare), etc. In 
the trait list these are Nos. 2, 4-13, 19, 22, 28, 31, ZT, 41, 46, 51-53, 65, 72, 81, 

93. 115- 

* Including five in ceramics, besides diamond-shaped beveled knives, platform 
disk pipes, scored ribs (tallies?), and antler tip flakers, Nos. 14, 34, 38, 39, 54, 
62, 75, 113, 120. 

*" Cf. Dunlevy, op. cit., p. 216. 


temporaneously with the Oneota, whereas the Pawnee traits are based 
on sites inhabited one or more centuries later toward the close of the 
tribe's residence in Nebraska. The conclusion seems inescapable 
that the Lower Loup Focus stands in very much closer and more 
direct relationship genetically to the later historic Pawnee than to the 
contemporaneous Oneota peoples." 

With the Oneota culture and its probable Siouan connections we 
shall not further concern ourselves here. Its role in the development 
of later native civilization west of the Missouri is not yet clear, 
although it probably introduced into the Pawnee area various innova- 
tions in ceramics, pipe-making, stone-working, and certain other 
fields of activity. At the moment, there is no reason to regard it as 
in any sense basic to historic Pawnee culture, since its contributions 
seem to have been rather in matters of detail. 

Bearing directly on the question of the nineteenth century Pawnee 
and their postulated descent from the Lower Loup Focus are certain 
noteworthy nonarcheological considerations. These seem to have been 
generally overlooked by those who challenge such a correlation on 
grounds (i) that the Pawnee have no legends concerning the sites, 
and (2) that the recent occupancy of the region by that tribe proves 
nothing as to its connection with the older remains. Both points 
can be met squarely with recorded data. Thus, to take up the first, 

" The kinds of traits comprising similarities and dissimilarities in the respective 
pairings is perhaps of as much significance as the absolute numbers. For ex- 
ample, while many of the hunting and skin-dressing practices were similar 
throughout, important differences are probably implied in the presence of fish- 
hooks and metapodial (split leg bone type) beamers in the Oneota. Both the 
latter items are widespread throughout the eastern United States, incidentally 
occurring also in prehistoric cultures in the Plains. The Pawnee and Lower 
Loup peoples apparently did not fish, and the outstanding feature of their skin- 
working industry was its distinctly Plains character; e. g., large elliptical 
quartzite scrapers, the notched flesher, bone paint "brushes," and probably the 
adzlike elkhorn hide scraper. At least a part of the subsistence economy of 
the Oneota, as well as the supposed bark or thatch house type, mound burials, 
extended use of woven mats, and a number of other items which this group alone 
of the three possesses, all tend to link them with eastern peoples and stamp them 
as comparatively recent arrivals west of the Missouri. The Pawnee and Lower 
Loup Focus peoples, on the other hand, resemble each other closely in virtually 
every fundamental respect and such common elements among them as the earth- 
lodge, pottery, horticulture, and other less distinctive items clearly have con- 
siderable historic depth in the eastern Plains. Onto this horticultural base they 
had grafted a hunting complex of western type, differing considerably but evi- 
dently well attuned to the peculiarities of the former. The successful integration 
of the two modes of life, both involving local ingredients, would in itself suggest 
a considerable period of adjustment in loco. 


on at least two occasions, Pawnee Indians have claimed certain of 
the protohistoric sites as the former dwelling places of their tribe. 
In 1867 Hayden collected a number of potsherds from "a Pawnee 
village site on Beaver Creek, Nebraska ....," some of which were 
subsequently figured by Holmes. ^^ Hayden nowhere records the 
exact location of' his finds, but Hill has since shown that two very 
large and almost contiguous protohistoric sites occur on the right 
bank of Beaver Creek a short distance above its mouth, while 2 or 
3 miles to the southwest is the Burkett site (fig. i, nos. 16-18). The 
ceramic and other remains from the three are very similar, and they 
were undoubtedly inhabited by the same people and at about the 
same time. In all probability Hayden's specimens which are of Lower 
Loup Focus type were picked up on one of these locations. It is, 
therefore, noteworthy that he says: 

No Pawnee Indian now living knows of the time when this village was in- 
habited. Thirty years ago [i. e., about 1837] an old chief told a missionary that 
his tribe dwelt there before his birth, but he knew nothing of the use of stone 
arrowheads, though, he said, his people used them before the production of iron. 

When the "production of iron" here began is not known, but the 
old chief's story tends to imply habitation of the site in question prior 
to the middle of the eighteenth century. The claim gains support from 
another tradition recorded by Bruce in his account of the North 
brothers and their Pawnee scouts." This is much more explicit and 
telling. It alludes to a battle which took place long ago between the 
Pawnees and the Poncas, when 500 of the latter made a treacherous 
but unsuccessful attack on a Skidi Pawnee village on Shell Creek 
north of Schuyler. The time of this alleged raid is wholly unknown, 
but it could not have taken place recently because there is no historic 
record to indicate that the Skidi, or for that matter any other Pawnee 
band, dwelt on Shell Creek as late as 1775 or after. Interestingly 
enough, at the precise locality where the old Skidi village is said 
to have stood, is the Gray- Wolfe site, one of the first of the Lower 
Loup Focus to be intensively studied and also one of the two on 
which the complex as defined is based. (See fig. i, nos. 24 and 25.) 
Finally, in a myth explaining the formation of the Skidi federation, 
Murie locates by streams two of the ancient villages. One of these 
was on the Elkhorn River, the other on Looking Glass Creek." This, 
if far less definitive, is still suggestive, since the lower course of the 
latter is sprinkled with not one but several related protohistoric sites. 

"Holmes, 1903, pp. 200-201 and pi. 177; Hayden, 1872, pp. 411-412. 
" Bruce, 1932, pp. 42-43. 
" Murie, 1914, p. 554. 

NO. 7 




Insofar as they are any clue, legends are thus seen to point toward a 
Pawnee authorship for at least some of the sites. 

It is unnecessary to stress the fact that mere areal concurrence of a 
nineteenth century tribe and a certain archeological complex is, per se, 
no proof of direct relationship. In the case of the Pawnee this par- 
ticular argument has never been used except as a possible corroborative 
circumstance. However, a careful study of the documentary history 
of the tribe tends to strengthen rather than weaken its force. Here 
it is possible to pass in review only a few of the more significant 
points ; for further details the reader is referred to recent publica- 
tions on the Pawnee and citations therein. Prior to the last quarter 
of the seventeenth century the sources are inconclusive as to the 
location of the tribe. Coronado, in 1541, places the province of 
Harahey, tentatively identified as Pawnee territory, north of Quivira. 
Later Spanish documents locate Quivira somewhere in central 
Kansas and its people are believed to have been the Wichita. If these 
identifications are correct, they suggest the presence of the Pawnee 
in southern or central Nebraska at this early date. A century and a 
quarter after, in 1666, Perrot mentions the Panys but without defining 
their habitat.'' Bandelier notes their presence as captives in New 
Mexico in the seventeenth century observing that they were not 
uncommonly ransomed from the Yutes and Apaches." By 1673, 
however, they had become sufficiently well known to be shown on 
Marquette's map, as also on that of Hennepin in 1678. Before 1680 
the Spanish in New Mexico heard rumors of Frenchmen among the 
Pawnees, and, wherever the location is given, subsequent narratives 
consistently place the Pawnee on the Rio Jesus Maria, north of 
Quivira. This stream is identified by historians with the Platte." 
For the eighteenth century there are many more records, as well as 
numerous maps showing ethnic distributions in the Missouri drain- 
age. Curiously enough, with all the unrest and tribal movements 
manifested therein from time to time, the Pawnee are almost always 
shown as a relatively stable group localized west of the Missouri 
on streams identifiable with the Loup, Platte, and possibly Republican 
Rivers. Particularly interesting in this connection is the 1718 Delisle 
map of Louisiana and the Mississippi River," because it depicts 
with remarkable accuracy the geographical details of the present 
Nebraska region (fig. 2). It shows the Pani (Pawnee) in 12 villages 

^^ Wisconsin Hist. Soc, Coll., vol. 16, pp. 15, 27, 1902. 

"Bandelier, 1890, p. 185, n. 4. 

" Thomas, 1935, pp. 12, 37. 

"Delisle, G., Carte de la Louisiane et du Cours du Mississippi. Paris, 1718. 

NO. 7 



on the "Riv des Panis," unquestionably the Platte, about the mouth 
of a large unnamed tributary entering from the north. Comparison 
with modern maps leaves Httle room for doubt that this tributary 
denotes the Loup, on whose banks the Panimaha (Skidi Pawnee) 
are represented, also with 12 villages. This is the first really con- 
vincing cartographic evidence that the Pawnee were established in 

Fig. 2. — Portion of the Delisle map (1718) showing the 
Pawnee towns on the Loup and Platte Rivers in east-central 

the Loup-Platte region in considerable numbers in the first quarter 
of the eighteenth century. Taken in conjunction with the data 
gleaned from earlier narratives, it adds strength to the view that this 
tribe has occupied its historic nineteenth century locale since the 
very beginning of white explorations. 

Of much concern to the archeologist using the so-called direct 
historical approach is the question of when European manufactures 


first began to reach his area. The discovery of such materials may 
offer an opportunity to determine approximately the time of occu- 
pancy of the sites or levels in which they occurred. Sometimes it is 
possible to identify beads or other trinkets with types known to have 
been made at certain stated periods in Europe. There are, of course, 
limitations to the method, and it must be used with due caution. Such 
objects as glass beads, copper bells or ornaments, and other small 
trinkets may have, and probably very often did, spread from village to 
village and from tribe to tribe, wholly independent of the trader 
after their original acquisition by the natives. They might thus 
precede the white man by several years. Also it is possible that the 
earliest traders left no written records, or that such as they may have 
left were lost or for other reasons remain unknown today. Still, 
where trade goods occur in small but consistent amounts in several 
related and neighboring sites, it seems reasonable to believe that a 
steady and direct, if perhaps limited, traffic had been established, and 
that historical records may offer valid clues as to the approximate time 
involved. It is theoretically possible that stray pieces reached the 
central Plains indirectly from New Mexico through the expeditions 
of Coronado (1541), Bonilla and Humana (1594), Ofiate (1601), and 
others, or as a result of raids against the Spanish settlements or their 
Apache and puebloan proteges. These, however, must have been 
of minor consequence. As a matter of fact, the Spaniards credit the 
rival French from Canada with introducing firearms, metal kettles, 
axes, and the like to the Pawnee," but it is not certain just how early 
this trade began. The first Frenchman to penetrate the region west 
and south of the Great Lakes is generally believed to have been 
Nicolet, who in 1634 visited the Winnebago and Illinois in what is 
now southern Wisconsin and northern Illinois.'" Owing to the 
hostility of the Iroquois and for other reasons, this voyage of explora- 
tion was not immediately followed up. It seems extremely doubtful 
that there was any appreciable commerce with tribes west of the 
Missouri prior to about 1650. By 1680 the Spanish had reports of 
French trade goods among the Pawnee on the Platte and in 1706 
their Apache allies killed a French couple somewhere in what is now 
northeastern Colorado. All this leads to the inference that regular 
trade was established in the central Plains region sometime between 
1650 and 1700. It is worth noting that from the first the Spanish 
records relating to French activities in this area uniformly link with 

^"Thomas, 1935, pp. I2ff. 
^Butterfield, 1881. 


them the Pawnee who seem to have been in firm possession of the 
Platte valley. 

Archeological findings leave no room for doubt that some at least 
of the sites belonging to the Lower Loup Focus were inhabited 
during a period when commercial intercourse was still comparatively 
limited in volume. Moreover, the European beads and other ma- 
terials so far studied from these sites, insofar as they can be dated, 
appear to be of types used in the Indian trade not prior to the latter 
seventeenth or eighteenth centuries. Finally, no early contact sites 
have been found in the region, other than those belonging to this 
complex, which could possibly be connected with the Pawnee or 
which can be viewed as the residence of settled Indians in contact 
with early traders. 

The historical background as here reviewed sheds significant light 
on the contention that the Lower Loup Focus may represent some 
group other than the Pawnee, not necessarily ancestral or even 
related to them. In the latest published work on this complex, it is 
suggested that "possible migration could account for the settling of 
different peoples in the same locality." "^ Early in this discussion it 
was pointed out that the village sites of the Lower Loup Focus, 
although of comparatively restricted distribution, are both numerous 
and very large. Moreover, since all those so far excavated have 
consistently yielded limited quantities of copper, glass beads, and 
(rarely) iron, it follows that they must have flourished for a time 
after white influences had penetrated into their locality. Even 
granting that all were not inhabited simultaneously, they undoubtedly 
indicate the presence here in protohistoric times of a populous, firmly 
established, and presumably potent ethnic group. Let us assume for 
the moment that this group was not ancestral nor even related to the 
Pawnee. We then have the somewhat difficult situation of a numerous 
and powerful tribe, resident for many years (witness the innumerable 
middens, earthlodge sites, etc.) in the very heart of the Pawnee 
territory, clinging to it until after trade contacts had been established 
with Europeans (circa 1650 or later), and then emigrating so un- 
obtrusively and so completely that the Pawnee, who must have 
followed closely on their heels so as to be firmly settled in the region 
by Delisle's time (1718), retained no tradition of their existence. 
This would not only do violence to Pawnee traditions linking that 
group with the protohistoric Lower Loup Focus, but would also 
require an explanation for the apparent absence of any legends of 

"' Dunlevy, op. cit., p. 215. 


an earlier tribe, unrelated but with very similar culture, whom the 
Pawnee could reasonably be thought to have displaced since estab- 
lishment of European contacts. Such a theory, furthermore, would 
presumably postulate a comparatively late incursion for the Pawnee, 
which is at variance with the ethnographic indications. Pawnee ma- 
terial culture of the nineteenth century, as has been stated, is pretty 
clearly a composite based essentially on two distinct and funda- 
mentally divergent economies — one horticultural and sedentary, the 
other hunting and nomadic. The significant constituents of the former, 
irrespective of their ultimate origin, are now known to have been well 
established west of the Missouri in prehistoric times. Those of the 
latter, in part rooted .in the very remote past, were shared with 
numerous other historic tribes of the Plains and particularly with 
the western bison hunters. The Pawnee seem to have combined 
the two in harmonious fashion, and so far as adjustment to environ- 
mental and ethnic conditions goes, give no evidence whatever of 
having been recent arrivals in the Nebraska region. 

There are other clues. Dunbar has shown how the placement of 
villages relative to one another has modified certain linguistic usages 
in accord with local geography .^^ During the later years of their 
residence in Nebraska there were seldom more than three or four 
villages — in other words, usually one for each of the four bands. 
At times two or more bands might occupy a single town, but the Skidi 
seem always to have remained more or less aloof. Both Murie and 
Grinnell present evidence supporting the view that subgroups within 
each of the main bands formerly constituted separate villages.** 
Murie credits the Skidi with 13 of these originally. This interesting 
observation may partially explain the general tendency of the early 
explorers to assign, usually from hearsay, as many as a score or 
more towns to the Pawnee nation. Incidentally, too, it may have 
archeological implications since the Pawnee locality abounds with 
small and Widely scattered precontact earthlodge villages which appear 
to have a number of features in common with the later ones. The 
sudden disappearance of the many small prehistoric villages and the 
presence of a few very large fortified towns in protohistoric times 
is an archeological puzzle which still awaits solution. Finally, the 
mythology of the Pawnee is replete with local Nebraska place names 
such as the Platte, the Loup, the Republican, Nemaha, and others.^* 
There are migration legends, to be sure, but none which afford any 

^^ Dunbar, 1880, p. 251. 

^ Murie, 1914, pp. 549-556; Grinnell, 1893, PP- 231-239. 

"* Dorsey, 1906. 


proof of recent arrival. Three of the five "sacred places" of the tribe 
were on the Loup and Platte within 50 miles of their junction ; the 
other two were in southern Nebraska and northern Kansas "* ; and 
a number of their myths and tales relate directly to this neighborhood. 

It must be apparent by this time that there exists little else than 
academic grounds for questioning the presence of the Pawnee as 
a firmly ensconced tribe in the Platte-Loup region since at least the 
coming of the whites. The data of tradition, history, ethnography, 
and mythology all support this inference. Moreover, the numerous 
archeological similarities between the historic Pawnee and the earlier 
Lower Loup Focus reflect essentially the same dual mode of life. 
Viewed in the light of history, the differences in materials from the 
two complexes are not so great as to strain the probability of a 
common authorship. They involve details rather than fundamentals. 
The greater richness, abundance, and variety of remains on the proto- 
historic sites indicate a general level of cultural achievement far 
above that of the historic Pawnee. If, as is very probable, this 
superiority extends to the nonmaterial side of life as well, then the 
protohistoric period may be regarded as the climax of social, cere- 
monial, and political development in the Pawnee area. The culmina- 
tion must have been reached before 1750. Thereafter came a steady 
decline which left the nineteenth century peoples in possession of a 
much simpler and clearly decadent cultural heritage, though the 
recorded myths as well as many political and ceremonial survivals 
hark back to the older and better days. Such a regression is perfectly 
in keeping with the contemporary history of the area : increased pres- 
sure from hostile tribes, growing commercial intercourse and terri- 
torial quarrels with the whites, new diseases, and a generally more 
desperate struggle for sheer existence, all of which left scant 
leisure for cultural advancement. 

The leads for future research on this problem are very clear. 
It is imperative first of all that thorough analyses be made of all 
available archeological materials from sites of the Lower Loup Focus. 
These should be carefully compared with similarly detailed studies of 
collections and data from documented sites of the nineteenth century. 
Needless to say, identities are not to be expected in all details, since 
individual, village, and probably band preferences were undoubtedly 
active factors. The element of time, too, must ever be borne in mind, 
for over a period of two or three centuries considerable changes are 
expectable. Another line of attack which has so far been totally 

' Grinnell, op. cit., pp. 358-359- 


neglected in this connection is physical anthropology. Skeletal re- 
mains either supposedly or certainly attributable to the Pawnee are 
by no means plentiful, as the early cemeteries remain undiscovered, 
and the later ones have suffered woefully at the hands of vandals. 
There is a disturbing possibility that scaffold burial and subsequent 
dismemberment may have been practised in the early period. Still, 
careful examination of the material thus far recovered might further 
illuminate the issue. For obvious reasons, it will probably never be 
possible to prove empirically that the inhabitants of any one of the 
Lower Loup Focus sites spoke a Pawnee dialect, since the individual 
sites cannot be linked with recorded towns. Thus the identification 
made on other grounds must remain a probability — a very high one, 
it is true, but still a probability. To maintain from this that the sites 
are not Pawnee, however, seems a captious argument, particularly 
in face of the very strong circumstantial evidence in every other 
respect. On the whole, it may be soundest and perhaps least confusing 
to retain a nonlinguistic designation for these protohistoric remains, 
at any rate for the present. For this purpose the term suggested by 
Dunlevy and used in this paper is as appropriate as any. 


In the foregoing pages the relationships between one historic and 
two protohistoric archeological complexes in Nebraska have been 
briefly discussed. These are respectively the Pawnee of the nineteenth 
century, the Lower Loup Focus, and the Oneota Aspect. From the 
evidence of archeology, history, tradition, mythology, and ethnography, 
as outHned herein, the following major facts emerge : 

(i) Village sites assignable to the Lower Loup Focus, 10 or more 
in number, occur only in the very heart of the historic Pawnee region 
about the confluence of the Loup and Platte Rivers. 

(2) These sites nearly all yield limited amounts of historical ma- 
terials, indicating their occupancy at least into very early contact 

(3) Historic maps and documents show that the Pawnee villages 
since virtually the earliest contact times were localized in and about 
this region. 

(4) On the basis of available archeological evidence alone, sites 
of the Lower Loup Focus show a much closer relationship to the 
later historic Pawnee culture than they do to the contemporaneous 
Oneota sites. 

(5) Pawnee traditions link that tribe directly with several of the 
protohistoric Lower Loup Focus sites. 


(6) Neither history, ethnography, nor recorded traditions offer 
any proof that another sedentary horticultural tribe inhabited this 
locality since the arrival of Europeans. 

These six points sufficiently refute the objections so far raised 
against identification of the Lower Loup Focus with the Pawnee 
tribes. There is, therefore, no reason whatsoever for abandoning the 
hypothesis outlined by Strong wherein the Lower Loup Focus is 
considered a protohistoric phase of Pawnee culture. 

Bandelier, a. F. 

1890. Contributions to the history of the southwestern portion of the United 
States. Cambridge. 
Bruce, Robert. 

1932. The fighting Norths and Pawnee scouts. 


1881. History of the discovery of the Northwest by Jean Nicolet. Cincinnati. 

1906. The Pawnee Mythology, pt. i. Carnegie Inst, of Washington. 
Dunbar, J. B. 

1880. The Pawnee Indians : their history and ethnology. Mag. Amer. Hist., 
vol. 4, no. 4, pp. 241-281. 

1936. A comparison of the cultural manifestations of the Burkett (Nance 

County) and the Gray- Wolfe (Colfax County) sites. Chapters in 
Nebraska Archeology, vol. i, no. 2, pp. 149-247- Univ. Nebraska. 
Griffin, J. B. 

1937. The archeological remains of the Chiwere Sioux. Amer. Antiquity, 

vol. 2, no. 3, pp. 180-181. 
Grinnell, G. B. 

1893. Pawnee hero stories and folk-tales. New York. 
Hayden, F. V. 

1868. Notes on Indian history, etc. Ann. Rep. Smithsonian Inst. 1867, pp. 
Hill, A. T., and Wedel, W. R. 

1936. Excavations at the Leary Indian village and burial site, Richardson 
County, Nebraska. Nebraska Hist. Mag., vol. 17, no. i. 
Holmes, W. H. 

1903. Aboriginal pottery of the eastern United States. 20th Ann. Rep., 
Bur. Amer. Ethnol., pp. 200-201 and pi. 177. 
Murie, James. 

1914. Pawnee Indian Societies. Anthrop. Pap., Amer. Mus. Nat. Hist, 
vol. II, pt. 7. 
Strong, W. D. 

1935. An introduction to Nebraska archeolog3^ Smithsonian Misc. Coll., 
vol. 93, no. 10. 
Thomas, A. B. 

1935. After Coronado. Univ. Oklahoma Press. 
Wedel, W. R. 

1936. An introduction to Pawnee archeology. Bur. Amer. Ethnol., Bull. 112. 


VOL. 97, NO. 7, PL 1 

Scenes in the Pawnee village on the Loup River near Genoa, Nebr., in 1871. 
This was the last northern settlement of the tribe prior to its fmal removal 
to the Indian Territory circa 1875. (Photographs by W. H. Jackson.) 


VOL. 97, NO. 7, PL. 2 

The Wright site near Genoa, JNiehr., sliuwmg type ot blutt tup village location 
preferred by the Pawnee in protohistoric times ; Beaver Creek valley at 
right. (Courtesy of the Nebraska State Historical Society.) 

_'. J'.xcavated floor of prntdhistoric Pawnee earthlodge sliownig circular outline, 
central firepit, postholes, and short vestibule doorway; Wright site. (Cour- 
tesy of the Nebraska State Historical Society.) 


VOL. 97, NO. 7, PL. 3 

Excavated fiuor of prutohisturic Pawnee earthludge at Larsen site, on Looking- 
glass Creek ; showing central firepit, surrounded by four primary and three 
circles of secondarj- post molds. Note the peculiar arrangement of postholes 
at the rear of the floor, opposite the entrance, where the family shrine was 
traditionally placed. (Courtesy of the Nebraska State Historical Society.) 

Excavated floor of late historical Pawnee earthly "l,m- mar 1 .c.shara, occupied 
probably after 1850. This lodge had eight central roof supports, a raised 
altar platform at the rear directly opposite the doorway, and a sill of baked 
clay across the inner end of the entrance passage. Another house floor may 
be seen in the background. (Courtesy of the Nebraska State Historical 


VOL. 97, NO. 7, PL. 4 

T. Restored pot of late Pawnee type froin Arclier, 
Nebr. ; height 9 inches. ( Courtesy of the 
Nebraska State Historical Society.) 



" '"^t^t^i^S^^^^^^A 



t- M 


,•' : .ii£.L-^._, 

^ «^ 

2. Restored vessel of pidluhislnric Pawnee type from the Wolfe 
site near Schuyler: height 4k inches. (Courtesy of the 
Nebraska State Historical Society.) 




• «-• 

















u o !- 

o 1- "^ 

ID — r/i 








■r- <U O 

Oh 0) 

2 ° 







(With 18 Plates) 


(Publication 3485) 



DECEMBER 30. 193 8 


VOL. 97, NO. 8, PL. 1 

The Tish-rawa Village, and the Klamath. Below the Entrance 
OF the Salmon* 

Drawn liy Capt. Setli Eastman, Irom original sketch I.iy George Gil)bs, 

October 185 1. 






(With 18 Plates) 



(Publication 3485 



DECEMBER 30, 1938 




1. "The Tish-rawa village, and the Klamath, belmv the entrance of the 

Salmon." Drawn by Capt. Seth Eastman from original sketch by 
George Gibbs. October 1851. (Frontispiece) 

2. Shoshonee Falls of Snake River, August 15, 1849. i. The Canon below 

the falls. 2. The falls 6 

3. Ravine in mountains of Burnt River, Baker County. Oregon. Sep- 

tember 4, 1 849 6 

4. On the trail in Oregon. Ascending hill from Deschutes River. Octo- 

ber 2, 1849 6 

5. Burial canoe on bank of the Columbia at mouth of La Camas . Creek. 

October 30, 1850 6 

6. Canoe on the Columbia River near Oak Point. October 1850 6 

7. Chiefs of tribes in the valley of the Willamette, i. Slacum. Chief of 

tribe at the falls of the Willamette, probably the Clowwewalla. May 
1851. 2. Joe or Alquema. Chief of the Santiam band of the Calapooya. 6 

8. ^^alley of the Willamette, i. Champoeg and French Prairie. April 1851. 

2. The Willamette at Champoeg. May 1851 10 

9. At Oregon City, 1851. i. Oregon City and the Falls of the Willamette. 

2. Indians taking salmon at the falls of the Willamette. June 1851.. 10 

10. Hudson's Bay Company's post at Fort Vancouver. July 185 1 10 

ir. At the Hudson's Bay Company's Post. i. Fort \'ancouver, July 2, 

1851. 2. Catholic Chapel at Fort Vancouver, July i, 1851. 3. Officers 

Quarters, Columbia Barracks, July 2, 1851 10 

12. Sonoma, California. August 11, 1851 14 

13. Upper end of Clear Lake, California. August 19, 185 1 14 

14. Redwood Tree, 52 feet in circumference. September 6, 185 1 16 

15. Sketches in the valley of the Klamath, i. Woman and child, at junction 

of Klamath and Trinity Rivers. October 6, 1851. 2. Young married 
woman, at junction of Klamath and Trinity Rivers. October 6, 1851. 

3. Young girl, Salmon River. November 12, 1851 16 

16. Views in the valley of the Klamath, i. The Klamath above the fish 

dam. October 9, 1851. 2. The Klamath. Signal tree of the Indians. 
October 13, 185 1 20 

17. The Shaste Butte and A'alley. October 27, 185 1 20 

18. Young Chief of the Weit-spek tribe, probably Mec-ug-gra. Drawn by 

Capt. Seth Eastman from original sketch by George Gibbs. October 
1851 26 


1. Carvings in wood and bone from lower Columbia River 8 

2. Wooden bowl from Shoalwater Bay 9 

3. Bow, hats, and headband from Klamath River, California 16 

4. Baskets from Klamath River, California 18 

5. Spoons and trays from Klamath River, California 19 

NORTHWEST, 1849-1851 


(With i8 Plates) 


George Gibbs, whose drawings form the subject of this article, 
was born July \J, 1815, at Suns wick, Long Island, New York, near 
the present Astoria. He died in 1873. At the age of 17, failing to 
receive the desired appointment to West Point, he accompanied an 
aunt to Europe, where he devoted 2 years to travel and study. Return- 
ing to New York, he soon entered Harvard, where he was graduated 
in law in 1838, and subsequently entered the law office of Preston 
Hall. However, the profession did not appeal to him, and during the 
next few years he wrote several works on historical sul^jects which 
were highly acclaimed. 

The Far West was now becoming of interest, and the mystery of 
the wilderness appealed to many. Gibbs was among those who were 
thus attracted, and in 1849 he accompanied the Mounted Rifle Regi- 
ment to Oregon, where he arrived early in October. That same 
autumn he became deputy collector of customs at Astoria and was 
later attached to the Indian Commission in Oregon. In 185 1 he was 
a member of the McKee party and visited the northwestern part of 
California. During the journey he learned much concerning the 
various native tribes with whom he came in contact, especially those 
who were encountered in the valley of the Klamath. Later he settled 
near Fort Steilacoom, Washington, where he devoted much time 
to the study of the languages of the different tribes and prepared 
extensive vocal)ularies which, together with brief lists of words, are 
now in the Smithsonian Institution. Although interested primarily 
in linguistic studies, Gil:)bs collected ethnological material, much of 
which is preserved in the collections of the National Museum, being 
among the earliest specimens gathered in the country beyond the 
Rocky Mountains. 

Gibbs was preparing to return to New York and so wrote to his 
friend, Prof. S. I"". Baird of the Smithsonian Institution: "North 
West Boundary Survey, Fort Walla- Walla. Nov. 16, i860. I arrived 

Smithsonian Miscellaneous Collections, Vol. 97, No. 8 


here yesterday on my way to WashingtcJii." He next wrote to i'rof. 
Baird from "261 Greene Street, New York, Feb. 5. i(S6i/' soon 
after he reached his home. 

The correspondence between (iil)l)s and Professor Baird ctjntinued 
through many years, botli while Gibl)s was in the west and after his 
return to New York, and later when he lived in Washington. The 
letters are most interesting, and many refer, in addition to the work 
in which both were engaged, to places and persons now known only 
in history. 


As previously mentioned, Gibbs accompanied the Mounted Rifle 
Regiment to Oregon in 1849, being one of many civilians who 
reached the valley of the Columbia at that time.' The regiment was 
under command of Brevet-Col. W. W. Loring, and the expedition 
started "with about 600 men, 31 commissioned officers, several women 
and children, the usual train agents, guides, and teamsters, 160 
wagons, 1,200 mules, 700 horses, and subsistence for the march to 
the Pacific." 

An interesting account of the trip has been preserved ; ' it was 
presented as (p. 126) : "A report, in the form of a journal, to the 
Quartermaster General, of the march of the regiment of mounted 
riflemen to Oregon, from May 10 to October 5, 1849, t>y JMajor 
O. Cross, c[uartermaster United States army." 

Excerpts from the journal will shed light on the dangers and 
difficulties with which all were confronted : 

Major Cross left St. Louis May 10, 1849, ^^^^ ascended the 
Missouri to Fort Leavenworth where he arrived 9 days later. "On 
inquiring at the fort I learned that the troops were ten days in 
advance of me, which was a very long start, as my mode of travelling 
was the same as that of the regiment." The next day he left for 
Fort Kearny. "My outfit was as indift'erent a one as ever left for 
any station, much less the Rocky mountains." 

It is not known with which of the groups Gibbs was then traveling, 
Init he was probably with the troops that had left Fort Leavenworth 
about the time Cross was departing from St. Louis. 

' Bancroft, Hubert Howe, The works of . . . vol. 30, History of Oregon, 
vol. 2, 1848-1888, p. 81. San Francisco, 1888. 

"The report was made by Maj. Osborne Cross to Maj. Gen. T. S. Jesup, 
Quartermaster General, and was incorporated in the report of the latter to 
C. M. Conrad, Secretary of War. 31st Congr., 2d Sess., Senate Ex. Doc. 
No. I, pt. 2, Washington, 1850. 


Again the journal (p. 143) : "June 5. — Large trains could be seen 
this morning wending their way along on both sides of the Platte. 
The river here is nearly three miles wide, interspersed with islands, 
some of which are thinly covered with very small cottonwood and 
willow." That day the wagons, 160 in number, were overhauled and 
many were repaired. 

June 7. — "To-day buffalo were seen for the first time, which 
created no little excitement." 

Ju>ic ig. — 'T visited Chimney Rock this morning, as the command 
wended its way along the river." 

June 22. — Arrived at Fort Laramie. "Fort Laramie is situated 
on Laramie's creek, a rajiid stream, about 60 yards wide, with a firm, 
pebbly bottom. This stream rises among the Black Hills to the west, 
and falls into the North Platte, about half a mile below the fort. 

"This fort is built in the form of a quadrangular figure, and of 
unbaked clay, or adobes ; the wall is about twenty feet high, with 
a small palisading on a part of it. There are two block-houses at 
the corners, diagonally from each other . . . Over the main entrance, 
which faces the river, there is also another small block-house. The 
buildings are made inside, the wall forming a part of them." There 
were no trees about the fort. Game was formerly plentiful, but. 
"has greatly diminished since emigrants have made it the great 
thoroughfare to Oregon and California." Fort Laramie is 639 miles 
from Fort Leavenworth. 

August I. — "It was at the side of the river, and at this place, 
that I saw the celebrated spring generally known as the Steamboat 
spring . . . This place is immediately at the point where the two 
trails turn off for California and Oregon, and within a short distance 
of the Salt lake ..." 

August 7. — "We descended a long hill, which brought us into a 
sandy plain, which extends to b^irt Hall, and on the banks of the 
Port Neuf ..." 

August 10. — "Our encampment last evening seemed to be the 
terminus of Snake River valley, as the appearance of the river 
entirely changed after a march of about five miles, which brought 
us to the American falls . . . The scene was truly magnificent ..." 
Manv rapids in the river, islands and masses of rock in the stream. 

August II. — "We crossed Ogden's river about 12 o'clock. The 
road turns oft' to the south for California, which was taken by the 
Calif ornians who were still along ..." 


/liiijusi ij.- -A (lay <it much I'likTcst for (iil)l).s. Left camp at 4 
in the moniiiii^' and, as llie journal continues (p. 196) : 

We travelled, however, rapidly for about eight miles . . . until we arrived 
at the creek again. At this place we waited for our wagons, which soon came 
up ; and, having assisted them out of the canon, which was no easy work, we 
continued on until the middle of the day, when we again came to the banks of 
the river, which were at least two or three hundred feet in height. I attempted 
to descend into the valley through which the river ran, for the purpose of 
procuring water, but it was so fatiguing, both for myself and horse, that I 
returned without being able to accomplish it. 

It was at this place we could easily hear the sound of a waterfall, which, 
from the noise, we at first supposed might have been the Little Falls of Snake 
river ; but, as we were still twenty miles from that point, we were soon 
satisfied that it did not proceed from there, or the small cascade on the opposite 
bank, which is mentioned by Colonel Fremont as the Subterranean river ; and 
we were much surprised to learn, the next day, that within ten miles of this 
place there is a cascade, which, in height, is not surpassed by the Niagara 
Falls. The guide, who was with the command, having travelled this route very 
often, was shown the place by an Indian, and took Mr. Gibbs, of New York, 
and Lieutenant Lindsay to the place, who pronounced it one of nature's great 
wonders. The river here becomes a little contracted, and passes through a 
chasm of soHd rock ; it commences to fall about a quarter of a mile above the 
last pitch, and, after forcing itself among loose rocks which lay in its way, 
takes a perpendicular pitch of at least 160 feet, and it is even thought to be a 
greater height. They descended to the foot of the falls, and after much difficulty 
and some length of time, where they were better able to judge more accurately 
of its great height ; and there seems to be but one opinion, that it equalled in 
grandeur, in proportion to the column of water, the Niagara Falls. Having 
been the first who had ever taken the trouble to examine them carefully, and 
wishing to change the name said to have been given by a priest many years 
since, they decided on that of the Great Shoshonie falls, instead of Canadian, 
as being the most appropriate. 

The road does not pass there, and probably its nearest point is not less than 
eight or ten miles, which is probably the reason why it is so little known, for 
I have never seen it mentioned by those who have trapped in this country for 
years . . . 

We continued our jtiurney until sundown, when we came to the foot of the 
little falls on Snake river, commonly called the Little Salmon Falls, and 
encamped for the night immediately on the banks of the river. 

The drawings of the falls made hy (lihbs that day are reproduced 
in plate 2. This was the first mention of (iibbs by Major Cross. 
The expedition continued through the mountains until (p. 210) : 

September 4. — Mountains were to be seen all around, and it appeared a 
mystery how we had extricated ourselves from those left behind us with so 
little difficulty, or how we were to pass those ahead of us. This brought us 
again on Burntwood creek, where we encamped for the night . . . 

The ravine through which the Burntwood passes is too narrow to be culti- 
vated, but the soil is rich and ought to yield well. The evening was spent in 


reaching the tops of some of tlie highest mountain hills, where the view of 
the adjacent country well rewarded us for our trouble; a few scattering hem- 
locks were seen in the ravine where we made our encampment, and the distant 
hills and ravines beyond were interspersed with several groves of cedar and 
pine. Our encampment lay in a fork formed by Burntwood creek and a little 
brook which falls into it . . .^ [PI. 3.] 

The party was now moving in several groups, and it is evident 
that Gibbs was not always with Major Cross; this explains the 
difference in the dates that often appear on the ^sketches made bv 
Gibbs from those of the entries in the journal. 

On September 22, Gil)bs made a sketch of the Columliia from the 
mouth of Deschutes River. From this point the wagon train, with 
which Gibbs must have been traveling, moved southward up the right 
l)ank of the Deschutes River. During the morning of Octoljer 2, 
the train ascended the steep cliff' near the river. A sketch made 
at that time reveals the wagons, each drawn by eight mules, form- 
ing a long line extending from the camp at the foot of the cliffs to 
the summit. It is an interesting drawing of a subject seldom recorded 

(Pl. 4). 

Leaving the Deschutes River, the expedition passed through the 
Cascade Range, and, on October 5, Gibbs made several sketches of 
the forest scenery, to which he attached the legends: "Burnt forest 
in Cascade Mts.," and "Cascade Mts. Cedar & firs," and again on 
October 9, "Forests of the Cascade Mts. Cedar & fir." 

The expedition had now arrived at its destination. Gibbs continued 
on to Astoria where he became Deputy-collector of Customs, soon 
to become attached to the Indian Commission. 


Two drawings of exceptional interest, made by Gil)bs on the banks 
of the Columbia during the month of Octol)er 1850, are reproduced 
in plates 5 and 6. 

The first of these shows the "Prow of dead Canoe on Bank of 
Columbia river, at mouth of Chamus Creek," and is a beautiful 
example of Gibbs' work. Chamus Creek is believed to have been the 
stream now known as La Camas Creek, which Hows into the Columbia 
River near the southeast corner of Clarke County, Washington, 
about 15 miles above Vancouver. This was within the Chinookan 
country. Whether this canoe was placed on a scaffold or rested on 

^ Burnt River flows eastward and joins .Snake River in the southern part of 
Baker County, Oregon. 


the ground is not known, but as so little has lieen recorded concerning 
the burial customs of the people of the region this sketch is of 
special interest. 

A brief reference to the strange form of burial was made by 
lieutenant Broughton, of the Vancouver Expedition/ who explored 
the lower Columl)ia during the autumn of 1792. He was near Cape 
Disappointment, on the AA^ashington side of the mouth of the 
Columbia, and wrote (vol. 2, p. 54) : "At this place was found 
the remains of a deserted Indian village, and near it three large 
canoes supported from the ground, each containing dead human 
bodies. These canoe coffins were decorated at the head and stern 
with rude carved work, and from their decayed state seemed to have 
been thus appropriated for a great length of time." 

Soon the Lewis and Clark party reached the valley of the Columbia. 
They encountered the same peculiar burials and left a more detailed 
account of the manner in which the canoes were placed, and of 
the various objects deposited in theuL' They stated (p. 429) : 

The Chinnooks, Clatsops, and most of the adjoining nations, dispose of the 
dead in canoes. For this purpose a scaffold is erected, by fixing perpendicularly 
in the ground four long pieces of split timber. These are placed two by two, 
just wide enough apart to admit the canoe, and sufficiently long to support its 
two extremities. The boards are connected by a bar of wood run through them 
at the height of six feet, on which is placed a small canoe, containing the body 
of the deceased, carefully wrapped in a robe of dressed skins, with a paddle, 
and some articles belonging to the deceased, by his side. Over this canoe is 
placed one of a larger size, reversed, with its gunwale resting on the crossbars, 
so" as to cover the body completely. One or more large mats of rushes or flags 
are then rolled round the canoes, and the whole secured by cords usually made 
of the bark of the white cedar. On these crossbars are hung different articles 
of clothing, or culinary utensils. The method practised by the Killamucks 
differs somewhat from this ; the body being deposited in an oblong box, of 
plank, which, with the paddle, and other articles, is placed in a canoe, resting 
on the ground. 

Later accounts of the curious form of l^urial are to be found, 
but the earlier descriptions are usually the more interesting. How- 
ever, as remarked in the Lewis and Clark journal (p. 429), "Those 
who first visit the ground, can only be expected to furnish sketches 
rude and imperfect." 

* Vancouver, Captain George, Voyage of Discovery ... 3 vols. London, 1798. 

^ Lewis and Clark, Travels to the source of the Missouri river and across 
the American continent to the Pacific ocean ... in the years 1804, 1805, and 
1806. London, 1814. 


VOL. 97, NO. 8, PL. 2 

I. The canon below the Falls. 

M m rw' 

i v., t ' ( 

iky> Knut. JaOt t^' ^.^ fit^ Uu^uaf n'- . 

2. The Falls. 
Shoshonee Falls of Snake river, August 15. 1849 


VOL. 97, NO. 8, PL. 3 I 


i'pj^M- ■''^PM^:.f'''<I'. 

1\ .''. A Av' 

Ravine in Mountains of Burnt River, Baker County, Oregon, 
September 4, 1849 





't- < 


VOL. 97, NO. 8, PL. 5 

r — '-^ 


W K'J 



^^-<> IMi.. 



Burial Canoe on Bank of the Columbia River at Mouth of 
La Camas Creek, October 30,|1850 







fv '1^ 






-C ^ 



As previously mentioned, it is not known whether the canoe, the 
prow of which was sketched by (iibbs, was placed on a scaffold 
when in use or had always rested on the ground as shown in the 
drawing. The drawing suggests that the prow was rather massive 
and heavy, but there is no way to judge its size. 

The Chinookan tribes who occupied both banks of the lower 
Columbia excelled in carving wood and bone. On January 20, 1806, 
when near the mouth of the Columbia on the south side, the Lewis 
and Clark party were among the Clatsops with whom they maintained 
a friendly intercourse. The narrative of the expedition refers to the 
skill of the natives in making many articles used in and about their 
houses, described as "large wooden buildings, varying in length from 
twenty to sixty feet, and from fourteen to twenty in width." The 
narrative continues (p. 432) : 

They are . . . very dexterous in making a variety of domestic utensils, 
among which are bowls, spoons, skewers, spits, and baskets. The bowl or 
trough is of different shapes, sometimes round, semicircular, in the form of a 
canoe, or cubic, and generally dug out of a single piece of wood, the larger 
vessels having holes in the sides by way of handles, and all executed with great 
neatness. In these vessels they boil their food, by throwing hot stones into 
the water, and extract oil from different animals in the same way. Spoons 
arc not very abundant, nor is there any thing remarkable in their shape, except 
that they are large and the bowl broad . . . The usual plate is a small mat of 
rushes or flags, on which every thing is served. 

Later, when the expedition was at the Cathlamah village, also on 
the Columbia and not far from the Clatsops, certain customs of the 
people were recorded in the narrative (p. 493) : 

This village we have already described, as situated opposite to the seal 
islands : on one of these the Indians have placed their dead in canoes, raised on 
scaffolds, above the reach of the tide. These people seem more fond of carving 
in wood than their neighbours, and have various specimens of their taste about 
the houses. The broad pieces supporting the roof and the board through which 
doors are cut, are the objects on which they chiefly display their ingenuity, 
and are ornamented with curious figures, sometimes representing persons in 
a sitting posture supporting a burden. 

Beautiful examples of the work of the people near the muuth of 
the Columbia are shown in figure i. Three of the carvings are in 
wood and one in bone. The latter, a knife handle, has on the end a 
remarkable representation of a raccoon, Procyon loior, with the eyes 
indicated by copper inlays. The club is made of cedar and is rather 
light for the purpose indicated. All were collected by George Gibbs 
probably in 1850 or 185 1. Another bowl obtained by him in the 
vicinity of Shoal water Bay, on the coast a short distance north of 


the niunth of the Colunihia, is reproduced in figure 2. Similar pieces 
were undoubtedly seen by Lewis and Clark a generation earlier. 

The second sketch made during the autumn of 1850, plate 6, bears 
the legend "Columbia River near Oak Point, Oct. 1850." The point 
is on the right bank of the Columbia about midway between the mouth 
and Vancouver, and was so named by Lieutenant Broughton in 1792." 
When going up the river they arrived at a spot "where, for the first 

Fig. I. — Specimens collected by George Gibbs on the lower Columliia. 

a, bone knife handle, length 8^ inches, U.S.N.M. no. 708 ; b, club for killing 
fish, wood, length 18 inches, U.S.N.M. no. 651; c, spoon, wood, length of figure 
on handle 3! inches, no number ; d. bowl, wood, diameters 6 and 8 inches, 

U.S.N.M. no. 691. 

time in this river, some oak-trees were seen, one of which measured 
thirteen feet in girth; this obtained the name of Oak Point." 

The canoe is the most interesting feature of the sketch. To quote 
again from Lewis and Clark" (pp. 433-434) : 

The industry of the Indians is not confined to household utensils : the great proof 
of their skill is the construction of their canoes. In a country, ipdccd. wlierc so 

° Vancouver, op. cit. vol. 2, p. 61. 
'Oo. cit. 


much of the intercourse hctwecn different tribes is carried on by water, the in- 
genuity of the people would naturally direct itself to the improvement of canoes, 
which would gradually become, from a mere safe conveyance, an elegant ornament. 
We have accordingly seen, on the Columbia, canoes of many forms, beginning 
with the simple boats near the mountains, to those more highly decorated, 
because more useful nearer the mouth of the Columbia. Below the grand 
cataract there are four forms of canoes : the first and smallest is about fifteen 
feet long, and calculated for one or two persons : it is, indeed, by no means 
remarkable in its structure, and is chiefly employed by the Cathlamahs and 
Wahkiacums among the marshy islands. The second is from twenty to thirty- 
five feet long, about two and a half or three feet in the beam, and two feet in 
the hold. It is chiefly remarkable in having the bowsprit, which rises to some 
height above the bow, formed by tapering gradually from the sides into a 
sharp point. Canoes of this shape arc common to all the nations below the 
grand rapids. 

Fig. 2. — Wooden bowl collected by George Gibbs at Shoalwater Bay. 
Diameters 8\ and lol inches, U.S.N.M. no. 692. 

The other types of canoes, larger than those just described, need 
not be mentioned. Evidently the canoe sketched by Gibbs belonged 
to the second group mentioned by Lewis and Clark, those which were 
"common to all the nations below the grand rapids." There is no 
allusion in the early narratives to the use of sails and masts in the 
native craft. The mast and sail shown in the sketch had 1:)een adopted 
after contact with Europeans. 

OREGON, 1851 

A letter from the Acting Commissioner of Indian Affairs, dated 
October 25, 1850, addressed to J. P. Gaines, A. H. Skinner, and 
Beverly S. Allen, stated' (p. 114): 

"In Annual Report of the Commissioner ni Indian Affairs . . . 1850. Wash- 
ington, 1850. 


"Gentlemen: I li;iv<.' ])vvu orticially notified of yonr ;i])])ointnicnl as 
'Commissioners to neyotiale treaties with the several IncHan trihes 
in the Territory of Oregon, for the extinguishment of their claims 
to lands lying west of the Cascade Mountains, under the act of 5th 
June last' ; and am directed by the Hon. Secretary of the Interior to 
prepare appropriate instructions for your observance in the discharge 
of the duties of your office." The region was briefly described, the 
tribes were mentioned in a vague manner, and the letter then con- 
tinued : "It will probably be best for you to treat first with the 
Indians in the white settlements, particularly in the Willammette 
Valley — and to treat separately with each tribe ..." 

Evidently the three commissioners were active during the ensuing 
months. In a j(jint communication to the Commissioner of Indian 
Atfairs, dated Champoly, April 19, 185 1," they transmitted (p. 205) : 
"a treaty concluded, on the i6th instant, with the Santian band of 
the Callapooya tribe of Indians, l)y which they cede to the United 
States a portion of the Willamette valley, about eighty miles in 
length and about twenty in width. And also a treaty, concluded this 
day, with the Twallalty band of the same tribe, including a country 
about fifty miles in length and about thirty in width . . . Their 
numbers are, of the Santian band, 155, and of the Twallaltys, 65." 

Gibbs was associated with the Commissioners when the treaties 
were made. 

Among the Gibbs material in the Bureau of /Vmerican Ethnology, 
Smithsonian Institution, is a manuscript designating where and when 
he prepared many of the vocabularies. One record is of interest at 
this time as it refers to events at Champoeg in April 185 1 : '" 

Kalapiiya. — Mj^ own vocabulary of this language was obtained April 4, 1851. 
while the Commission was engaged in a treaty with them at Champoeg. It 
is of the Si-yam-il, or as generally called Yamhill band, living on the river of 
that name, which empties into the Willamet from the coast range. The Twallatys 
(Twalati), and the Luckamukes (Luk-a-mai-yuk) speak the same dialect. The 
Santiam band, on the east side of the Wilamet, a rather different one. It was 
given by Thomas and Antoine, Chiefs. 

Hlolelc. — Obtained at the same place. This was received from an Indian of 
the band inhabiting the upper waters of the Santiam. 

Alany drawings were made at this time, four of which are now 
reproduced. Others show different parts of the valley as it appeared 
during the spring of 185 1. 

"In Annual Report of the Commissioner of Indian Affairs . . . 1851. Wash- 
ington, 1 85 1. 
"• Bur. Amer. Ethnoi. Manuscript Catalog No. 742. 


VOL. 97, NO. 8, PL. 8 

ii» » a iiiftim»ri ii >r* i«ii*« ii «l »» ^^^^^M^. 

I. Champoeg and French Prairie, April 1851. 

2. The \\'illamette River at Champoeg, Alay 1851. 
Valley of the Willamette 


VOL. 97, NO. 8, PL. 9 

I. Oregon City and the Falls of the Willamette. 

2. Indians taking salmon at the Falls of the Willamette, June 185 1. 
At Oregon City. 1851 


VOL. 97, NO. 8, PL. 1 1 

I. Fort ^^ancouvcr, July 2, 185 1. 

2. Catholic Chapel at Fort A^ancouver, July i, 185 1. 

3. Ofliccr,s' quarters, Columbia Barracks, July 2, 185 1. 
At the Hudson's Bay Company's Post 


Portrait sketches of two chiefs are given in ])latc 7, one heing 
that of a chief of the Santiani l)an(l of the Callapooyas, Init not one 
who had contrihuted to the vocahulary. This is well drawn, and the 
sloping forehead reveals the effect of artificial flattening. It was 
probably a good likeness. 

The second portrait is that of "Slacum, Chief of tribe at Falls of 
Willamette (Upper Chinooks)," and was drawn a few days after 
the treaty was made with the Callapooya. Slacum may have been a 
chief of the Clowwewalla, belonging to the Chinookan family, a 
tribe that occupied the region bordering the falls of the Willamette 
River, the site of Oregon City. The name Slacum was probaljly 
derived from that of an x\merican naval officer who visited the 
region in 1836 "to obtain information in relation to the settlements 
on the Oregon river." '^ He prepared an interesting, although Ijrief, 
account of the native tribes then living on the Willamette, part of 
which follows. When ascending the river : 

The first tribe of Indians are the Kallamooks, on the left bank, on a small 
stream of the same name, 30 miles from its mouth : 2d are Keowewallahs, alias 
Tummezvatas or Willhanietts. This tribe, now nearly extinct, was formerly 
very numerous, and live at the falls of the river, 22 miles from its mouth, on 
the right bank. They claim the right of fishing at the falls, and exact a 
tribute from other tribes who come hither in the salmon season (from May 
till October). Principal chiefs deceased. This river at the present day takes 
its name from this tribe. 3d. "Kallapooyahs" occupy lodges on both sides of 
the river. 4th. "Fallatrahs" on a small stream of same name, right or west bank. 
5th. Champoicho — west bank. 6th. Yamstills — west bank. 7th. Leelahs — both sides. 
8th. Hanchoicks. All these 5 tribes speak Kallapooyah dialects, and are doubt- 
less of that tribe, but at present are divided as designated, and governed by 
chiefs as named. All these tribes do not exceed 1,200. 

The Willamette was a beautiful stream, as Slacum wrote, "even 
in midwinter, you find both sides clothed in evergreen, presenting 
a more beautiful prospect than the Ohio in June . . . On the right 
the land rises gradually from the water's edge, covered with firs, 
cedar, laurel, and pine. The oak and ash is at this season covered 
with long moss, of a i)ale sage green, contrasting finely with the 
deeper tints of the evergreens." 

Gibbs appreciated the beauty of the region and revealed it in two 
sketches reproduced in plate 8. 

The first is a view of "Champoeg and the prairies beyond," 
looking over the Willamette ; the second shows the banks of the 
stream with the variety of trees and shrubs, with spring foliage. 

" Slacum, William A., Memorial of . . . praying compensation for his ser- 
vices . . . 25th Congress, 2d Session, Senate Doc. 24. Washington, 1838. 


C,'hani|)(»cy was on the vit;lil liaiik (if the Willamette, at the northern 
end of French I'rairie, the origin of which name was explained by 
Bancroft/' who wrote (pp. 70-73) : 

As their terms of contract expired, the Hudson's Bay Company began to 
retire its servants, giving them choice lands not too far removed from its 
benign rule. This was the origin of the French Canadian settlements in the 
beautiful Valley Willamette . . . French Prairie, the tract where the servants 
of the fur company began tiieir planting in the Willamette Valley, extended 
from the great westward bend of that river south to Lac La Biche about 
twenty-five miles . . . The landing at the crossing of the Willamette on the 
east side was known as Campement du Sable, being a sandy blufif and an 
encampment at the point of arrival or departure for French Prairie. Two 
miles above this point was Champoeg, the first settlement. 

The falls of the Willamette, when snrrounded by the primeval 
forest and in its natural condition, was a place of great beauty. And 
as it was here that many Indians from the scattered villages would 
come to get their supply of salmon, it was likewise a place of great 
importance to the native inhabitants of the valley. But about the 
year 1840 a settlement was begun at the falls, Oregon City, and soon 
all was changed, although the few remaining Indians continued to 
take salmon at the falls, as others had done through generations. 

Two sketches by Gibbs, reproduced in plate 9, show the falls, and 
Oregon City as it appeared in June 185 1. An Indian is portrayed 
spearing fish from a canoe, another is seen standing on a fishing 
stage, in the right center of the sketch, using a net at the foot of the 
falls. Fish are also shown leaping from the water. The upper di"aw- 
ing is a view of Oregon City with the falls just beyond. 

Oregon City was visited by Major Cross on October 5, 1849, 
after the completion of the trip from Fort Leavenworth " and was 
described in these words (pp. 227-228) : 

The city of Oregon is not a very prepossessing place in its appearance, for, 
like all new places in the western country, the stumps and half-burnt trees lie 
about in every direction. It is immediately at the Willamette Falls, hemmed 
in by the river in front, and a ledge of rocks immediately in rear and very 
close to the city. 

Leaving Oregon City, Gibbs evidently continued down the Wil- 
lamette and next visited Fort Vancouver, which had been erected 
during the years 1824-1825. Sketches of the fort, and one of 
Columbia Barracks a short distance away, are shown in plates 10, 11. 

^"Bancroft, Hubert Howe, The works of . . . vol. 29, History of Oregon, 
vol. I. San Francisco, 1886. 

"Op. cit. In 31st Congr., 2d Sess., Senate Fx. Doc. No. i, pt. 2, Washington, 


A concise description of Fort A'ancouver, printed in 1840," 
explains many of the details of the drawings. To quote (pp. 19-20) : 

On the north side of the Columbia, and a quarter of a mile from it, stands 
Fort Vancouver, the principal establishment of the Hudson's Bay Company 
west of the Rocky Mountains. It consists of a number of wooden buildings 
within a stockade, serving as dwelling-houses, stores, magazines, and work- 
shops ; and near it are other small buildings inhabited by the laborers, together 
with a saw-mill and grist-mill. The whole number of residents at the place 
is about eight hundred, of whom a large proportion are Indians or half-breeds. 
Several hundred acres of land near the fort are under cultivation, producing 
wheat, barley, oats, pease, potatoes, &c., in abundance; and the stock of cattle 
is also considerable. 

It was a place of great activity, surrounded by many tribes who 

spoke diiTerent languages and had strange manners and ways of life. 

Maj. Osborne Cross mentioned Fort Vancouver in his journal:^'* 

Fort Vancouver, which is the head(|uarters of the Hudson's Bay Company, 
is on the right bank of the river. It is situated on a beautiful plain, about five 
miles long, and probably is three quarters of a mile wide. The country gradually 
rises, and runs back for ten or fifteen miles, passing through several plains, 
some of which are cultivated. On one of these plains there is an excellent 
seminary, where the children from the fort and the neighborhood are educated. 

Immediately in rear of the fort, and on the rising ground, the company of 
artillery under Brevet Major Hatheway have put up temporary quarters, and 
have made themselves very comfortable. 

The latter became Columbia Barracks, and the temporary quarters 
were soon replaced by others of a more permanent nature. The 
Officers' Quarters at the barracks, as they appeared 2 years later, 
were sketched by Gibbs July 2, 185 1 (pi. 11, fig. 3). At that time 
they formed an attractive group of buildings facing Fort Vancouver, 
with the Columbia beyond, while a short distance in the rear was 
the edge of the forest which extended olif to the north. 

Gibbs did not remain many days in the vicinity of the fort, but 
turned southward to California where he joined the McKee party 
and soon set out to explore the northwestern part of the State and 
to visit the many native tribes some of whom may never before 
have come in contact with the white man. 


The journal of the expedition into northwestern California, pre- 
pared by Gibbs and later mentioned by McKee in his letter of March 

" Greenhow, Robert, Alemoir, historical and political, on the Northwest coast 
of North America . . . 26th Congr., ist Sess., Senate Doc. 174. Washington, 

'" Op. cit., p. 228. 


13, 1852, was published Ijy Schoolcraft the following year.'' It is 
a valuable account of a journey through a i)art of the country never 
before carefully studied, and describes briefly the native inhabitants 
of the rough, mountainous region who f)ccupied secluded valleys in 
the vicinity of the rivers, often difficult to discover. 

Pencil sketches made by Gibbs of scenes along the route reveal 
much of interest and beauty encountered in the wilderness and are 
now reproduced for the first time. Statements in the journal which 
tend to describe or explain the drawings will be quoted, although 
much of equal value, but not referring to the sketches, must neces- 
sarily be omitted. 

This will be followed by excerpts from McKee's account of the 

Journal ok the Expedition of Colonel Redick M'Kee, United States 

Indian Agent, through North-Western California. Performed 

During the Summer and Fall of 1851. By George Gibbs. 

j\Io)iday, Ai((/. 11. — Colonel M'Kce and party, escorted by Major Wessells, 
and a detachment of thirty-five nionnted riflemen, left Sonoma this morning, 
and moved over to .Santa Rosa, encamping a Httle beyond Carillo's ranch . . . 
The general route proposed to be followed by the expedition, was up Russian 
river to its sources, down Eel river to Humboldt bay, and thence over to the 
Klamath, ascending that to the neighborhood of Shaste Valley, should the season 
permit. [PI. 12.] 

Continuing northward, the party soon reached Clear Lake where 
they remained several days. Large groups of Indians assembled, 
and a treaty was entered into. "In personal appearance, many of 
the Clear Lake Indians are of a very degraded caste ; their fore- 
heads naturally being often as low as the compressed skulls of the 
Chinooks ... A vocabulary of this language was obtained from 
the Indian who accompanied us, and who spoke Spanish sufBciently 
to be enabled to interpret with his ])eople." On the next day, 
August 19, the proposed treaty was explained to the assembled 
Indians. A region of great natural l)eaut}' (p. 109): 

Surrounded on every side by mountains, this valley is completely isolated 
from the adjoining country, there being no access except by difficult trails . . . 
The principal valley upon the lake is that upon which we encamped, lying on 
the western side, and extending from mount Af'Kce towards the head. The 
extent of this may be stated at ten miles in length, by an average width of 
four. A more beautiful one can hardly be pictured. Covered with abundant 

^' Schoolcraft, Henry R., Information respecting the history, condition and 
prospects of the Indian tribes of the United States, pt. 3, pp. 99-177- Phila- 
delphia, 1853. 






■, 'nI 








' p 

«» L' 


y. ,t 




■<■ , ^ 

c .' 


'k " 









' ',' 









ri' .-ii;-- 








































^ ■ 


. l^l 


grass, and interspersed with groves of superb oaks of the most varied and 
graceful forms, with the lake and its green margin of tule in front, and the 
distance bounded everywhere by precipitous ranges, it combines features of 
surpassing grandeur and loveliness. Plowers of great variety and elegance 
abound, the woods are filled with game, and in the season innumerable flocks 
of water-fowl enliven the shores. [PI. 13.] 

]Vedncsday, Aug. 20. — The council was again assembled, and the treaty 
explained to them as engrossed ... As regards the suitableness of the reserva- 
tion for its purpose, there can hardly be a doubt. The spot is isolated to a 
degree unusual even on the Pacific ; abounds in all that is necessary for a large 
number of people in their savage state, and is capable of being made in the 
highest degree productive by cultivation. 

Saturday, Aug. 30. — This valley, called by the Indians Ba-tem-da-kai, we 
supposed to be on the head of the south fork of Eel river, and so were informed 
by our guide and other mountaineers ; but a belief exists, as we afterwards 
found . . . that it is, on the contrary, the head of the river before spoken of 
as entering the coast to the westward ... A few Indians visited us, and were 
directed to call in the adjacent tribes. 

The entire party remained in canip the following day. 

Sunday, Aug. 31. — Quite a number of Indians were assembled and presents 
distributed, but no treaty attempted ; for our Clear Lake interpreter, although 
able to comprehend them, could not explain freely in turn. Their language, 
however, is clearly of the same family as that of the tribes at the head of 
Russian river, and those last encountered. The total number in the vicinity, as 
near as could be ascertained, was about six hundred souls . . . They pluck their 
beards, and some of them tattoo. Many had their hair cut short, but others 
wore it turned up in a bunch in front, or occasionally on the back of the head 
. . . The average height of these men was not over five feet four or five 
inches . . . We saw no women . . . 

I took the opportunity of to-day's halt, to ascend the hills on the eastern 
side of the valley. The view from this point was beautiful, the stream winding 
in serpentine form along the margin of the plain, fringed with oaks and firs, 
and the long slopes beyond diversified with forest and prairie. To the east 
rose heavy ranges of mountains, between which and the yet more distant 
Sacramento chain, a wide and deep gap indicated another valley, supposed to 
be the source of the main fork of Eel river. 

The next day the trail led through a moinitainous section, "crossing 
deep arroyas and then ascending a hroken ridge hetween the waters 
of the south and middle forks [of Eel river]." Some Indians were 
encountered who '"had rohes of deer skin, dressed with the hair on, 
over their shoulders. They helonged to a wild mountain tribe, the 
terror of the valley Indians ... Of their language and affinities, 
nothing is known." 

September 5. — The trail crossed the river and passed a grove of 
redwoods. During the day a few Indians were encountered, and 
(P- I -'3) : 

twd or three of them were of larger stature than usual, and one was reall}' 
a fiiic-iooking young fellow. They wore the deer-skin robe over the right 





shoulder, and carried the common short bow, backed with sinew [a, fig. 3!, 
and arrows pointed with stone, both tolerably well made. With all these 
Indians, the arrow-points are fastened into a short piece of wood, which in 
turn is fixed, though but loosely, into the shaft. The quiver, of dressed deer- 
skin, holds both bow and arrows. They had also, suspended round the neck, 
small nets, neatly made after the fashion of the common game-bag ; the twine, 
which was very even, being of course their own work. 

The last part (if otn- niarcli led us into a thick redwood forest. 

_1 > V ^ 

•U0 4*'4*^'^>0'^ 


Fig. 3. — Specimens collected by George Gibbs on the Klamath River, California. 

a, sinew-backed bow, length 34 inches. U.S.N.M. no. 649; /', basketry hat, 
diameter 7 inches, depth 4 inches, U.S.N.M. no. 7556; c. basketry hat, diame- 
ter 7I inches, depth 4 inches, U.S.N.M. no. 7558; d, headband, U.S.N.M. 
no. 7520; c. two sections of d, X approximately 2.5. 

Saturday, Scptcuihcr 6. — Indians visited the camp but they were 
of little interest, and (p. 124) : 

I endeavored in vain to get from them the names of articles at hand, parts 
of the body, t&c, as they either could not or would not understand the object 
of the inquiry ; nor was our Clear Lake Indian more successful after his 
method . . . 


VOL. 97, NO. 8, PL. 14 


_ u 

Qi "■ 











































S 5^' 
5 G , 

►^ >,/' -^ 



Our camp was a very pretty one, the little prairie being level and rich, 
and encircled by a magnificent redwood forest. One tree near the tents I 
measured, and found it to be fifty-two feet in circumference, at four or five feet 
from the ground, and this although the bark and a portion of the wood were 
liurned away . . . [PI. 14.] 

From September 29 until the morning of October g, the party 
occupied a camp estabH.shecl at the junction of the Klamath and 
Trinity Rivers. Gibbs did not make a separate entry for each day 
spent at the camp, but between the days mentioned devoted much 
time to the study of the Indians with whom he was in contact. Many 
tribes were represented at the gathering, possessing similar manners 
and ways of life. To quote briefly (p. 139) : 

With regard to their form of government, at least that of the Klamath and 
Trinity tribes, the mow-ce-ma, or head of each family, is master of his own 
house, and there is a sci-as-lau, or chief, in every village . . . The lodges of 
these Indians are generally very well built ; being made of boards riven from 
redwood or fir, and of considerable size, often reaching twenty feet square. The 
roofs are pitched over a ridge-pole, and sloping each way ; the ground being usually 
excavated to the depth of three or four feet, and a pavement of smooth stones 
laid in front. The cellars of the better class are also floored and walled with 
stone. The door always consists of a round hole in a heavy plank, just 
sufficient to admit the body ; and is formed with a view to exclude the bears, 
who in winter make occasional and very unwelcome visits. 

The people were descril)ed as being superior to an}- previously met, 

and with countenances denoting greater force and energy of character, as well 
as intelligence . . . The superiority, however, was especially manifested in the 
women, many of whom were exceedingly pretty ; having large almond-shaped 
eyes, sometimes of a hazel color . . . their only dress the fringed petticoat, or 
at most, a deer-skin robe thrown back over the shoulders, in addition. The 
petticoat with the wealthier, or perhaps more industrious, was an affair on 
which great taste and labor were expended. It was of dressed deer-skin ; the 
upper edge turned over and embroidered with colored grasses, the lower cut 
into a deep fringe, reaching nearly to the knee, and ornamented with bits of 
sea-shell, beads, and buttons . . . The same round basket-cap noticed before, 
is worn by the Klamath women [/». c, fig. 3], figures of different colors and 
patterns being worked into it. They tattoo the underlip and chin in the manner