Skip to main content

Full text of "Smithsonian miscellaneous collections"

See other formats



VOL. 51 


(Publication 1943) 




The present series, entitled " Smithsonian Miscellaneous Collec- 
tions," is intended to embrace all the publications issued directly by 
the Smithsonian Institution in octavo form; those in quarto constitut- 
ing the " Smithsonian Contributions to Knowledge." The quarto 
series includes memoirs, embracing the records of extended original 
investigations and researches, resulting in what are believed to be new 
truths, and constituting positive additions to the sum of human knowl- 
edge. The octavo series is designed to contain reports on the present 
state of our knowledge of particular branches of science; instructions 
for collecting and digesting facts and materials for research ; lists and 
synopses of species of the organic and inorganic world; reports of 
explorations; aids to bibliographical investigations, etc., generally pre- 
pared at the express request of the Institution, and at its expense. 

In the Smithsonian Contributions to Knowledge, as well as in the 
present series, each article is separately paged. The actual date of its 
publication is that given on its special title-page, and not that of the 
volume in which it is placed. In many cases works have been pub- 
lished and largely distributed, years before their combination into 

Secretary of the Smithsonian Institution 



Number 1 (Publication 1791). The Development of the 
American Alligator (A. misslssippiensis). By 
Albert M. Beese. 1908. Pp. 6G, Pis. 33. 

Number 2 (Publication 1803). The Taxonomy of the Mus- 
coidean Flies, including Descriptions of New 
Genera and Species. By Charles H. T. Towns- 
end. 1908. Pp. [3] +138. 

Number 3 (Publication 1807). Smithsonian Exploration in 
Alaska in 1907 in Search of Pleistocene Fossil 
Vertebrates. By Charles W. Gilmore. 1908. 
Pp. 38, Pis. 13. 

Number 1 (Publication 1869). The Mechanics of the Earth's 
Atmosphere. A Collection of Translations. 
Third Collection. By Cleveland Abbe. (Hodg- 
kins Fund.) 1910. Pp. iv + 617. 





(A. mississippiensis) 



Professor of Zoology, West Virginia University 

No. 1791 








(With 23 plates) 


With the exception of S. F. Clarke's well-known paper, to which 
freqnent reference will be made, practically no work has been done 
upon the development of the American alligator. This is probably 
due to the great difficulties experienced in obtaining the necessary 
embryological material. Clarke, some twenty years ago, made three 
trips to the swamps of Florida in quest of the desired material. The 
writer has also spent parts of three summers in the southern 
swamps — once in the Everglades, once among the smaller swamps 
and lakes of central Florida, and once in the Okefenokee Swamp. 
For the first of these expeditions he is indebted to the Elizabeth 
Thompson Science Fund ; but for the more successful trip, when 
most of the material for this work was collected, he is indebted to 
the Smithsonian Institution, from which a liberal grant of money to 
defray the expenses of the expedition was received. 

The writer also desires to express his appreciation of the numerous 
courtesies that he has received from Dr. Samuel F. Clarke, especially 
for the loan of several excellent series of sections, from which a 
number of the earlier stages were drawn. 

The present paper gives a general outline of the whole process of 
development of the American alligator (A. mississippiensis), it 
being the intention of the author to take up in detail the more spe- 
cific points in subsequent researches. 

In preparing the material several kinds of fixation were employed, 
but the ordinary corrosive sublimate-acetic mixture gave about the 
most satisfactory results. Ten per cent formalin, Parker's mixture 
of formalin and alcohol, etc., were also used. In all cases the em- 
bryos were stained in toto with borax carmine, and in most cases 
the sections were also stained on the slide with Lyon's blue. This 
double stain gave excellent results. Transverse, sagittal, and hori- 
zontal series of sections were made, the youngest embryos being cut 
into sections five microns thick, the older stages ten microns or 
more in thickness. 



The Egg 

Figures i, kt (Plate I) 

The egg (fig. i) is a perfect ellipse, the relative lengths of whose 
axes vary considerably in the eggs of different nests and slightly in 
the eggs of the same nest. Of more than four hundred eggs meas- 
ured, the longest was 85 mm. ; the shortest 65 mm. Of the same 
eggs, the greatest short diameter was 50 mm. ; the least short diam- 
eter was 38 mm. The average long diameter of these four hundred 
eggs was 73.74 mm. ; the average short diameter was 42.59 mm. 
The average variation in the long axis of the eggs of any one nest 
was 11.32 mm., more than twice the average variation in the short 
axis, which was 5.14 mm. No relation was noticed between the size 
and the number of eggs in any one nest. Ten eggs of average size 
weighed 812 grams — about 81 grams each. 

Voeltzkow (18) 1 states that the form of the egg of the Madagas- 
car crocodile is very variable. No two eggs in the same nest are 
exactly alike, some being elliptical, some "egg-shaped," and some 
"cylindrical with rounded ends." The average size is 68 mm. by 
47 mm., shorter and thicker than the average alligator egg. 

When first laid, the eggs are pure white, and are quite slimy for 
a few hours, but they generally become stained after a time by the 
damp and decaying vegetation composing the nest in which they are 
closely packed. 

The shell is thicker and of a coarser texture than that of the hen's 
egg. Being of a calcareous nature, it is easily dissolved in dilute 

The shell membrane is in two not very distinct layers, the fibers of 
which, according to S. F. Clarke, are spirally wound around the egg 
at right angles to each other. No air-chamber, such as is found in 
the hen's egg, is found in any stage in the development. 

In most — probably all normal — eggs a white band appears around 
the lesser circumference a short time after being laid. This chalky 
band, which is shown at about its maximum development in fig. 
la, is found, on removal of the shell, to be caused, not by a change 
in the shell, but by the appearance of an area of chalky substance in 
the shell membranes. Clarke thinks this change in the membrane is 
to aid in the passage of gases to and from the developing embryo. 
Generally this chalky area forms a distinct band entirely around the 
shorter circumference of the egg, but sometimes extends only partly 

1 The numerical citations throughout the article are to bibliographical refer- 
ences at the end of the paper. 


around it. It varies in width from about 15 mm. to 35 mm., being 
narrowest at its first appearance. . Sometimes its borders are quite 
sharp and even (fig. ia) ; in other cases they are very irregular. If 
the embryo dies the chalky band is likely to become spotted with 
dark areas. 

The shell and shell membrane of the egg of the Madagascar croco- 
dile are essentially the same as those just described, except that the 
shell is sometimes pierced by small pores that pass entirely through 
it. The same chalky band surrounds the median zone of the egg 

The white of the egg is chiefly remarkable for its unusual density, 
being so stiff that the entire egg may be emptied from the shell into 
the hand and passed from one hand to the other without danger of 
rupturing either the mass of albumen or the enclosed yolk. The 
albumen, especially in the immediate neighborhood of the yolk, seems 
to consist of a number of very thin concentric layers. It varies in 
color, in different eggs, from a pale yellowish white, its usual color, 
to a very decided green. 

As might be expected, no chalazas are present. 

The yolk is a spherical mass, of a pale yellow color, lying in the 
center of the white. Its diameter is so great that it lies very close to 
the shell around the lesser circumference of the egg, so that it is 
there covered by only a thin layer of white, and care must be taken 
in removing the shell from this region in order not to rupture the 
yolk. The yolk substance is quite fluid and is contained in a rather 
delicate vitelline membrane. 

The albumen and yolk of the crocodile's egg, as described by 
Voeltzkow, differ from those of the alligator only in the color of the 
albumen, which in the crocodile is normally light green (18). 

As pointed out by Clarke, the position of the embryo upon the yolk 
is subject to some variation. During the earliest stages it may occur 
at the pole of the yolk nearest the side of the egg ; later it may gener- 
ally be found toward the end of the egg; and still later it shifts its 
position to the side of the egg. It is probable, as Clarke says, that 
the position at the end of the egg secures better protection by the 
greater amount of white, at that point, between the yolk and the 
shell ; while the later removal to the side of the egg, when the vascu- 
lar area and the allantois begin to function, secures a better aeration 
of the blood of the embryo. 

Around the embryo, during the stages that precede the formation 
of the vascular area, is seen an irregular area of a lighter color and a 
mottled appearance. This area is bounded by a distinct, narrow, 


white line, and varies in size from perhaps a square centimeter to 
one-third the surface of the yolk. 

During the earliest stages of development the embryo is very trans- 
parent; so that, as there is no fixed place upon the yolk at which it 
may be expected to occur, it is often very difficult to find. Owing to 
this transparency, to the extreme delicacy of the embryo, and to the 
character of the white, the removal of an early embryo from the egg 
of the alligator is a difficult operation and is accomplished only after 
some practice. 

The Development of the Embryo 

As the writer has pointed out elsewhere (13), the embryo of the 
alligator is often of considerable size when the egg is laid. This 
makes the obtaining of the earliest stages of development a difficult 
matter ; so that the writer, as has already been said, like S. F. Clarke 
(5), made three trips to the South in quest of the desired material. 
Voeltzkow (18) experienced the same difficulty in his work on the 
crocodile, and made several trips to Africa before he succeeded in 
obtaining all the desired stages of development. 

To obtain the earliest stages, I watched the newly made nests until 
the eggs were laid, and in this way a number of eggs w r ere obtained 
within a very few hours after they had been deposited, and all of 
these eggs contained embryos of a more or less advanced stage of 
development. Gravid females were then killed, and the eggs re- 
moved from the oviducts. These eggs, although removed from a 
"cold-blooded" animal, generally contained embryos of some size, 
and only one lot of eggs thus obtained contained undeveloped em- 
bryos, which embryos refused to develop further in spite of the most 
careful treatment. Voeltzkow (18) found, in the same way, that the 
earlier stages of the crocodile were extremely difficult to handle ; so 
that, in order to obtain the earlier stages, he was reduced to the 
rather cruel expedient of tying a gravid female and periodically re- 
moving the eggs from the oviducts through a slit cut in the body 

The older embryos are hardy and bear transportation well, so that 
it is comparatively easy to obtain the later stages of development. 

For the stages up to the formation of the first four or five somites, 
I am indebted, as I have already said, to Professor Clarke, and, 
since I have had opportunity to examine only the sections and not 
the surface views of these stages, I shall quote directly Clarke's 
paper in the Journal of Morphology (5) in description of these sur- 
face views. 


Stage I 
Figures 2-2/ (Plates I, II) 

The youngest embryo that we have for description is shown in 
figures 2 and 2a. Of figure 2 Clarke says : 

"The limiting line between the opaque and pellucid areas is clearly 
marked, and within the latter is a shield-shaped area connected by 
the narrower region of the primitive streak with the area opaca. 
The blastopore is already formed near the posterior end of the shield. 

"A ventral view of another embryo of the same age (fig. 20), seen 
from the ventral side, shows that the blastopore extends quite 
through the blastoderm, in an oblique direction downwards and for- 
wards, from the dorsal to the ventral side. The thickened area of 
the primitive streak is here very prominent. There is, too, the begin- 
ning of a curved depression at the anterior end of the shield, the 
first formation of the head-fold." 

Transverse sections of this stage are shown in figures 2b— 2f. 

Figure 2b, through the anterior region of the blastoderm, shows a 
sharply defined ectoderm (ec) which is composed of three or four 
layers of cells in the median region, while it gradually thins out 
laterally. Closely underlying this ectoderm is a thin sheet of irreg- 
ular cells, the entoderm (en). 

Figure 2c is about one-fifth of the length of the blastoderm pos- 
terior to the preceding and represents approximately the same condi- 
tions, except that there is an irregular thickening of the entoderm in 
the median region (en). This thickening apparently marks the an- 
terior limit of the mesoderm, to be discussed shortly. 

Figure 2d represents the condition of the blastoderm throughout 
about one-third of its length, posterior to the preceding section. The 
somewhat regular folds in the ectoderm (ec) are probably not the 
medullary folds, but are such artificial folds as might easily be pro- 
duced in handling the delicate blastoderm. The thickening of the 
entoderm, noticed in the preceding figure, is here more sharply de- 
fined, and as we pass toward the blastopore becomes separated some- 
what from the entoderm proper as a middle layer or mesoderm (fig. 
2e, mes). It would thus seem, from a study of these sections, that 
most of the mesoderm is derived from the entoderm. In fact, all of 
the mesoderm in front of the blastopore seems to have this origin, 
for it is not until the anterior edge of the blastopore is reached that 
there is any connection between the ectoderm and entoderm (fig. 2e). 

Figure 2e is a section through the region just mentioned, where, 
medially, the ectoderm, mesoderm, and entoderm form a continuous 


mass of cells. Laterally the mesoderm (mes) is a distinct layer of 
cells of a fairly characteristic mesodermal type. The notochord is 
not yet discernible, though a slight condensation of cells in the middle 
line may indicate its position. 

Figure 2/ is one of the four sections that were cut through the 
blastopore (blp), which is a hole of considerable size that opens, as 
the figure shows, entirely through the blastoderm. Along the walls 
of the blastopore the ectoderm and entoderm are, of course, contin- 
uous with each other and form a sharply defined boundary to the 
opening. As we pass laterally from the blastopore the cells become 
less compact, and are continued on each side as the mesodermal layer 
(mes). In this series the sections posterior to the blastopore were 
somewhat torn, and so were not drawn ; but they probably did not 
differ materially from those of the corresponding region of the im- 
mediately following stages, which are shown in figures 3m and 6i 
and will be described in their proper order. 

Stage II 
Figures 3-30 (Plates II, III, IV) 

The next stage to be described is shown in surface views in figures 
3 and 3a. Of this stage Clarke says : 

"The head-fold rapidly increases in depth and prominence, as 
shown in figure 3, which is a ventral view a few hours later [than 
the preceding stage]. The time cannot be given exactly, as it is 
found that eggs of the same nest are not equally advanced when laid, 
and differ in their rate of development. The lighter curve in front 
of the head-fold is the beginning of the anterior fold of the amnion. 
The notochord has been rapidly forming, and now shows very dis- 
tinctly on the ventral side, when viewed by transmitted light. A 
dorsal view of the same embryo (fig. 3a) shows that the medullary 
or neural groove is appearing, and that it ends abruptly anteriorly 
near the large transverse head-fold. Posteriorly it terminates at the 
thickened area in front of the blastopore, which still remains open." 

Figures 3&-W are drawn from transsections of an embryo of about 
this state of development. For a short distance in front of the be- 
ginning of the head-fold, there is a mass of cells of considerable 
thickness between the ectoderm and entoderm. In figure 36 these 
cells appear as an irregular thickening of the entoderm, while in fig- 
ure 3c they form a continuous mass, uniting the upper and lower 
germ layers. This condition is seen, though in a much less striking 
degree, in the following stage of development. As to its significance 
the writer is not prepared to decide. 


Figure 3d passes through the head-fold, which in this embryo was 
probably not so far developed as it was in the embryo shown in fig- 
ures 3 and 3a. Not having seen the embryo, however, before it was 
sectioned, the water cannot be certain of this point. The ectoderm 
and entoderm are here of nearly the same thickness. 

Figure 3c is a short distance posterior to the preceding. It shows 
a marked thickening of the ectoderm in the medial region (cc), 
which is continuous posteriorly with the anterior ends of the medul- 
lary folds that are just beginning to differentiate (figs. $f-h). 

Figure 3^ passes through the anterior end of the medullary plate 
or folds (mf), whichever they may be called. The ectoderm of the 
folds is thickened and is considerably elevated above the rest of the 
blastoderm. There is scarcely any sign, in this region, of a medul- 
lary groove. The entoderm (en) is considerably thickened in the 
medial region, this thickening being continuous posteriorly, as in the 
preceding stage, with the mesoderm. 

In figure 3/;, cut in a plane at some distance posterior to the pre- 
ceding, the medullary groove (mg) is well marked; its bordering 
folds gradually thin out laterally to the thickness of the ordinary 
ectoderm. The medial thickening of the entoderm is very marked, 
but it has not in this region separated into a distinct mesoblastic 

Immediately under the medullary groove is a dense mass of cells 
(nt), apparently the anterior end of the notochord in process of 

Figure 31, still farther toward the blastopore, shows the medullary 
groove wider and shallower than in the more anterior sections. The 
mesoderm (mes) is here a layer laterally distinct from the entoderm. 
In the middle line it is still continuous with the entoderm, and at this 
place it is the more dense mass of cells that may be recognized as 
the notochord (nt). It is evidently difficult to decide whether this 
group of cells (nt) . which will later become a distinct body, the noto- 
chord, is derived directly from the entoderm or from the mesoderm, 
which is itself a derivative of the entoderm. There is here abso- 
lutely no line of demarcation between the cells of the notochord and 
those of the mesoderm and entoderm. 

In figure 3/ the ectoderm (ec) is nearly flat, scarcely a sign of the 
medullary groove appearing. The mesoderm (mes) is here a dis- 
tinct layer, entirely separate from both notochord (nt) and entoderm 
(en). The notochord is a clearly defined mass of cells, distinct, as 
has been said, from the mesoderm, but still closely united with the 
underlying entoderm, which is much thinner than the ectoderm. 


This condition of the notochord, which is found throughout about 
one-third of the length of the embryo, would give the impression 
that the notochord is of a distinctly entodermal origin. 

In figure 3/c there is no sign of the medullary groove, th< »ugh ecto- 
derm (ec) is still much thickened in the middle line. The section 
passes, posterior to the notochord, through the anterior edge of the 
ventral opening of the blastopore (blp). The mesoderm (mes) is 
here again continuous with the entoderm, around the edge of the 
blastopore, but is distinct from the ectoderm. 

Figure 3/ represents the third section posterior to the preceding. 
The blastopore, which passes upward and backward through the 
blastoderm, is seen as an enclosed slit (blp). It is surrounded by a 
distinct layer of compactly arranged cells continuous with the thick- 
ened ectoderm (ec) above, with the thin entoderm (en) below, and 
laterally with the gradually thinning and scattering mesoderm (mes). 

Figure 3;/; is the next section posterior to the one just described. 
It passes through the dorsal opening of the blastopore (blp), which 
appears as a deep, narrow cleft with thick ectodermal borders. The 
three germ layers are still continuous with each other, though the 
connection of the entoderm with the other two is slight. The sec- 
tions posterior to this one will be described in the next stage, where 
they have essentially the same structure and are better preserved. 

Figures 3« and 30 are sagittal sections of an embryo of about the 
stage under discussion. In both figures the head-fold is seen as a 
deep loop of ectoderm and entoderm, while the head-fold of the 
amnion is seen at a. 

The beginning of the foregut is seen in figure 3;/ (fg), which is 
the more nearly median of the two sections, figure 30 being a short 
distance to the side of the middle line. 

In figure 30 the thin entoderm (en) is separated from the much 
thicker ectoderm (ec) by a considerable layer of rather loose meso- 
derm (mes). In figure 3;/, which is almost exactly median in posi- 
tion, there is, of course, no mesoderm to be seen in front of the blas- 
topore, and the entoderm shows a considerable increase in thickness, 
due to the formation of the notochord (nt). The blastopore (blp) 
is the most striking feature of the figure, and is remarkable for its 
great width in an antero-posterior direction. Its anterior and pos- 
terior borders are outlined by sharply defined layers of ectoderm and 
entoderm. Posterior to the blastopore the lower side of the ecto- 
derm is continuous with a considerable mass of cells, the primitive 
streak (ps). 


Stage III 
Figures 4, 4a, 5, 5a, and 6-6i (Plates V, VI) 

"Figures 4 arf& 4a are of an embryo removed, on June 18, from an 
egg which had been taken out of an alligator two days before. Fig- 
ure 4, a dorsal view, is of special interest in that it shows a secondary 
fold taking place in the head-fold. This grows posteriorly along the 
median dorsal line, forming a V-shaped process with the apex point- 
ing backward toward the blastopore. There is quite a deep groove 
between the arms of the V. The head-fold on the ventral side, as 
seen in figure 40, made from the same embryo as figure 4, grows 
most rapidly on the mid-line, and also becomes thicker at that place. 
The medullary folds now begin to form on either side of the medul- 
lary groove, ending posteriorly on either side of the blastopore and 
anteriorly on either side of the point of the V-shaped process in the 
middle of the head-fold. This is seen in figure 5, which is a dorsal 
view of an embryo from an egg three days after it was taken out of 
an alligator. A ventral view of the same embryo (fig. 5a) repre- 
sents the thickened process on the mid-line at its greatest develop- 
ment. For some reason the notochord did not show in this embryo, 
possibly owing to particles of the yolk material adhering about the 

"In an embryo a day or two older, the V-shaped fold of the head- 
fold is seen to have broken through at the apex, and each of the 
arms thus separated from one another unites with the medullary fold 
of its respective side. This can be seen in figure 6, which is a dorsal 
view of part of an embryo a day or two older than the one repre- 
sented by figures 5 and 5a. 

"This is so unexpected a method of formation for the anterior 
part of the medullary folds that I have made use of both figures 4 
and 5. They were made from very perfect specimens, and the sec- 
tions of both of them, and of the specimen from which figure 6 was 
drawn, proves that the structure is what it is indicated to be in sur- 
face appearance. That is, the transverse sections posterior to the V, 
in the embryos shown in figures 4 and 5, show the medullary groove 
and the medullary folds; the several sections passing through the 
apex of the V show neither groove nor folds, but only a median 
thickening ; and in front of the point or apex of the V the successive 
sections discover a gradually widening groove between the arms, 
which is also much deeper than the shallow groove found posterior 
to the V. While I have not seen, and from the nature of the condi- 
tions one cannot see, the change actually proceeding from the form 


of fig. 5 to that of fig. 6, still the explanation given appears to be the 
only one possible" (5). 

A somewhat extended series of transverse sections of an embryo 
of about this age is represented in figures Oa-i. 

Figure 6a is a section through the head-fold ; it passes through the 
extreme anterior end of the secondary folds (sf) that were de- 
scribed, in surface view, above (figs. 5 and 6). The section was not 
quite at right angles to the long axis of the embryo, so that the fold 
on the right was cut further toward its anterior end than was the 
fold on the left. The pushing under of the head causes a forward 
projection of the secondary folds, so that the fold to the right ap- 
pears as rounded mass of cells with a small cavity near its center. 
On the left the plane of the section passes through the posterior limit 
of the head-fold, and shows the cells of the secondary fold contin- 
uous with the dorsal side of the ectoderm (ec). As pointed out 
above by Clarke, the secondary folds are here some distance apart, 
and gradually approach each other as we proceed toward the pos- 
terior. The entoderm (en) is here flat and takes no part in the sec- 
ondary folds. 

In figure 6b, a short distance back of the one just described, the 
secondary folds (sf) are much larger and are closer together. On 
the right the section passes through the extreme limit of the head- 
fold, so that the secondary fold of that side is still a closed circle, 
with a few scattered cells enclosed. On the left the section is pos- 
terior to the head-fold ; on this side the secondary fold is seen as a 
high arch of ectoderm, with a thick mass of entoderm beneath it. 

Figure 6c represents a section which passes back of the head-fold 
on both sides. The secondary folds (sf) are seen as a pair of ecto- 
dermal arches continuous with each other in the middle line of the 
embryo. The ectoderm of the folds is much thickened and grad- 
ually becomes thinner distally. On the right the entoderm shows 
the same thickening (en) that was shown on the left side of the 
preceding figure. This thickened appearance of the entoderm is 
due to the fact that the section passes through the anterior limit of a 
tall fold of that layer, which underlies the similar fold of the ecto- 
derm that has already been described. This secondary fold of the 
entoderm is seen on the left side of the section. It may be traced 
through several sections, but soon flattens out posteriorly. 

Figure 6d is a short distance posterior to the preceding. The sec- 
ondary folds are here much less pronouncedly arched and the deep 
groove between them is reduced to a line (/). The entoderm (en) 
is no longer markedly arched and is closely adherent, along the 
median plane, to the ectoderm, where there is seen the thickening 


(th) that has been mentioned by Clarke (see above). Springing 
from the entoderm on each side of this thickening is a small mass of 
mesoderm (mes). 

The section immediately posterior to the one just described is rep- 
resented in figure 6c. The line (/) which separated the two second- 
ary folds in the preceding section is no longer present, so that the 
ectoderm (ec) is continuous from side to side, with only a shallow 
depression (nig), which may be considered as the extreme anterior 
end of the medullary groove. The median thickening (th) is cut 
near its posterior limit and still shows a close fusion of the germ 
layers. There is no line of demarcation between the gradually flat- 
tening secondary folds of the anterior end of the embryo and the 
just forming medullary folds of the posterior end, so that it is im- 
possible to say whether the thickening of ectoblast in this figure 
should be called secondary folds or medullary folds. As a matter 
of fact, the secondary folds become, of course, the anterior ends of 
the medullary folds. The mesoblast (mes) is here of considerable 
extent, and its entodermal origin is beyond doubt, though not well 
shown in the figure. 

Figure 6f is about one-sixth of the length of the embryo posterior 
to the preceding. The medullary thickening of the ectoderm (ec) is 
still marked and the shallow medullary groove (trig) is fairly dis- 
tinct. The entoderm (en) is medially continuous with both meso- 
derm (mes) and notochord (nt), though these two tissues are other- 
wise distinct from each other. 

Figure 6g is nearly one-third the length of the embryo posterior 
to the preceding and passes through the posterior third of the em- 
bryo. The medullary thickening of the ectoderm (ec) is marked, 
but shows no sign of a medullary groove ; in fact, the median line is 
the most elevated region of the ectoderm. The notochord (nt) is 
larger in cross-section than in the more anterior regions. It is still 
continuous with the entoderm (en) and is fairly closely attached to, 
though apparently not continuous with, the mesoderm (mes) on each 

Figure 6h passes through the blastopore (blp). The appearance 
of the section is almost identical with that of figure 2.f, already de- 

Figure 6i is five sections posterior to the preceding and has about 
the same structure as the corresponding sections in the precedmg 
two stages, where this region of the embryo was injured, and hence 
not drawn. Continuous with the posterior border of the blastopore 
(seen in the preceding figure) is the deep furrow, the primitive 
groove (pg). The ectoblast (ec) bordering this groove is much 


thickened and may be called the primitive streak. The lower side 
of this primitive streak is continuous with the mesoblast (mes), 
while the entoderm (en) is here entirely distinct from the mesoderm. 
It is evident that the mesoderm posterior to the blastopore is pro- 
liferated from the lower side of the ectoblast and not from the upper 
side of the entoblast, as is the case anterior to the blastopore. The 
primitive groove gradually becomes more and more shallow, as it is 
followed toward the posterior, until it is no longer discernible; 
back of this point the primitive streak may be traced for a consider- 
able distance, becoming thinner and thinner until it too disappears, 
and there remains only the slightly thickened ectoblast underlaid by 
the thin and irregular layers of mesoblast and entoblast. The prim- 
itive streak may be traced for a distance equal to about one-third 
the distance between the head-fold and the blastopore. 

Stack IV 
Figures jn-yh (Plates VI, VII) 

No surface view of this stage was seen by the writer, and hence is 
not figured. The figures were drawn from one of the series of sec- 
tions obtained through the courtesy of Prof. S. F. Clarke. This 
series was marked "3 Urwirbeln," so that the embryo was appar- 
ently slightly younger than the youngest stage obtained by myself 
and represented in figures 8 and 8a. 

Figure ja represents a section that passed through the head-fold 
of the amnion (a) just in front of the head-fold of the embryo; the 
amniotic cavity here appears as a large empty space. 

Figure jb is several sections posterior to the preceding; it passes 
through the head-fold of the embryo, but is just back of the head- 
fold of the amnion. The deep depression of the ectoderm (ec) and 
entoderm (en) caused by the head-fold is plainly seen. In this de- 
pression lie the ends of the medullary folds, distinct from each other 
both dorsally and ventrally. Each medullary fold is made up of two 
parts — a medial, more dense nervous layer (ill), and a distal, less 
dense epidermal layer (ep). The section corresponding to this one 
will be more fully described in connection with the following stage. 

Figure yc is some distance posterior to the preceding, though the 
exact distance could not be determined because of a break in the 
series at this point. The section passes through the posterior limit 
of the head-fold. The medullary groove (mg) is very deep and 
comparatively wide; around its sides the germ layers are so closely 
associated that they may scarcely be distinguished. Ventral to the 
medullary groove the foregut (fg) is seen as a crescentic slit. 


Figure yd is about a dozen sections posterior to the one just de- 
scribed and is about one-seventh the length of the embryo from the 
anterior end. The embryo is much more compressed, in a dorso- 
ventral direction, and the medullary groove (mg) is correspond- 
ingly more shallow. Where the ectodorm (ec) curves over to form 
the medullary folds it becomes much more compact and somewhat 
thicker. The notochord (tit) is large and distinct, but is still fused 
with the entoderm (en). The mesoderm (mes) forms a well- 
defined layer, entirely distinct from both the notochord and the ento- 
derm. From this region, as we pass caudad, the size of the embryo 
in cross-section gradually decreases and the medullary groove be- 
comes more shallow. 

Figure ye is about one-third of the length of the embryo from the 
posterior end, and is only a few sections from the caudal end of the 
medullarv groove. The ectoderm (ec) is much thinner than in the 
preceding figure and the medullary groove (mg) is much more 
shallow. The notochord (nt) is of about the same diameter as 
before, but is here quite distinct from the entoderm (en) as well as 
from the mesoderm (mes). 

Figure yf is seven sections posterior to figure ye. The medullary 
groove has disappeared and the medullary folds have flattened to 
form what might be called a medullary plate (at the end of the refer- 
ence line ec), which continues to the anterior border of the blasto- 
pore. The notochord (nt) is larger in cross-section than in the 
more anterior regions ; it is still distinct from the entoderm. 

Figure yg passes through the blastopore and shows essentially the 
same structure as was described in connection with the correspond- 
ing section of stage I (fig. 2/). 

Figure yh represents the region of the primitive groove (pg) and 
primitive streak (ps). The section shows the typical structure for 
this region— a thick mass of cells is proliferating from the ventral 
side of the ectoderm (ec) and is spreading laterally to form a dis- 
tinct mesoderm (mes). The entoderm (en) is entirely distinct from 
the other layers. 

Stage A' 

Figures 8-8; (Plates VII, VIII, IX) 

On opening the egg this embryo (figs. 8 and 8a) was found to lie 
on the end of the yolk, near the center of the irregular, lighter area 
which was mentioned in connection with the description of the egg. 
The length of the embryo proper is 3 mm. This was the youngest 


und in [905, and appr quite closely the condition of 

the chick embryo after 24 hours' incubation. The dorsal aspect of 
this embryo, viewed by transmitted light, is shown in figure 8. The 
medullary folds (mf) have bent over until they are in contact, 
though apparently not fused for a short distance near their anterior 
ends. As will be described in connection with the sections of this 
stage, the medullary folds are actually fused for a short distance at 
this time, though in surface views they appear to be separated from 
each other. In the Madagascar crocodile (18) the first point of 
fusion of the medullary folds is in the middle region of the embryo, 
or perhaps even nearer the posterior than the anterior end of the 
medullary groove Throughout the greater part of their length the 
medullar) folds arc ^till wid arated; ; rly they are 

merged with the sides of the very distinct primitive streak (ps), 
which seems, owing to its opacity, to extend as a sharp point toward 
the head. Extending for the greater part of the length of the primi- 
tive streak i-. the primitive groove (pg), which, when the embryo 
is viewed by transmitted light, is a very striking feature at this stage 
of development and resembles, in a marked way, the same structure 
in the embryo chick. Clarke (5) figures the blastopore at this stage 
;i ;i small opening in fronl of the primitive streak, bul 'I 1 Men- 

tion any such condition as above described at any stage of devel 

it. Five pain of somites (s) hav< be< ted and may be 

ecu. though hut faintly outlined, in both dorsal and ventral views of 
the embryo ; thi , lie aboul half way between the extreme end- of the 
1 mhryo. The head-fold (7/, lig. 8a) shows plainly in a ventral view 
as a darker, more opaque anterior region, extending for about one- 
fourth the length of the embryo. The still unfused region of the 
medullary fold- may be seen also in the ventral view at mg. The 
I fold of the amnion (a) forms a very thin, transparent hood 
r the extreme anterior end of the embryo. The tail-fold of the 
amnion has not made its appearance, and in fact is not apparent at 
any stage in the development. This is true also of the Madagascar 
crocodile. The notochord (nt) ma; be seen , n a ventral view as a 
faint, lim liny along the middle line from the head- 

told to the primitive streak. 

Two sagittal sections of thi hown in figures 8fc and 8c. 

The embryo from which the sections were made was apparently 
somewhat crooked, so that it was not possible to get perfect longi- 
tudinal sections. For example, in figure 8b the plan.- of the section 
is almost exactly median in the extreme pO! terior and middle region--. 
h"i i- on on.- id.- of the middle line elsewhere. Thi- explains the 
enormou thickening of the ei toblast in the region of the head, wher< 


the section passes through one of the medullary folds ( mf) at its 
thickest part; and also explains the fact that the ectoblast is thinner 
in the middle region (ec), where the section passes through the 
medullar) groove, than it is farther toward the blastopore, where 
the section cuts' the edge of the medullary folds. The outline- of 
the middle and extreme posterior regions of the ectoblast are much 
more irregular and ragged than is shown in the figure. The plane 
of the section passes through the notochord (nt) in the posterior 
region, btit not in the anterior end of the embryo., where a layer of 
mesoblast (mes) is seen. The great size of the blastopore (blp) is 
well shown, as is the beginning of the foregut (fg)- Comparison of 
this figure with the more anterior transverse sections and with the 
dorsal surface view of this stage will make the rather unusual condi- 
tions comprehensible. 

Figure 8c is cut to one side of the median plane, distal to the 
medullary folds. Being outside of the medullary folds, the ecto- 
derm ( ec) is thinner and less dense than in figure 8& ; anteriorly it is 
pushed down and back as the head-fold, and posteriorly it becomes 
thin where it forms the dorsal boundary of the primitive streak (ps). 

The foregut (fg), as would be expected, is not so de*ep as in the 
median section (Sb). The most striking feature of the section is 
the presence of five mesoblastic somites (s). Each somite, especially 
the second, third, and fourth, is made up of a mass of mesoblast 
whose cells are compactly arranged peripherally, but are scattered 
in the center, where a small myocoel may be seen. 

A series of transverse sections of the embryo shown in figures 8 
and St/ is represented in figures 8d-j. 

Figure 8d is through the anterior end of the embryo; the posterior 
edge of the amnion is cut only on one side (a). The medullary 
folds (mf) are shown as two distinct masses of tissue, separated by 
a considerable space from each other, both dorsally and ventrally ; 
they are underlaid by the ectoderm of the head-fold, through which 
the section passes. A mass of yolk (y) is shown at one side of the 

Figure Se represents a section a short distance posterior to the 
one just described, and passes through the short region where the 
dorsal edges of the medullary folds have fused with each other. The 
ventral side of the medullary groove ( mg) is, as in the preceding 
section, still unclosed. An epidermal layer of ectoblast i ep) is now 
distinct from the nervous layer (;//). 

Figure 8/ is through a region still farther toward the posterior 
end. Here the medullary groove is again open above, and is still 

j — VI. 


open below. A well-marked space is seen between the epidermal 
(ep) and nervous (nl) layers of the ectoderm, but no mesoblast is 
yet to be seen. 

Figure 8g passes through the middle part of the head-fold, and 
shows that the medullary folds in this region are fused below, but 
are widely separated above, where their margins are markedly bent 
away from the mid-line. Between the epidermal and nervous layers 
of the ectoderm a considerable mass of mesoderm cells (mes) is 
seen. The curious appearance of the preceding four figures, as well 
as the first three figures of the next stage, was at first quite puzzling, 
until a model of the embryo was made from a series of sections. It 
was then plain, as would have been the case before, except for the 
unusual depth dorso-vcntrally of the head of the embryo, why the 
medullar}- canal should at the extreme anterior end be open both 
dorsally and ventrally, while a few sections caudad it is open only 
ventrally, and still farther toward the tail it is again open both above 
and below. These conditions are produced by the bending under of 
the anterior region of the medullary folds, probably by the formation 
of the head-fold. It is apparently a process distinct from the ordi- 
nary cranial flexure, which occurs later. The fusion of the medul- 
lary folds to form a canal begins, as has been already mentioned, 
near the anterior end, whence it extends both forward and backward. 
Hence, if the anterior ends of the medullary folds be bent downward 
and backward, their unfused dorsal edges will come to face ventrally 
instead of dorsally, and sections through the anterior part of this 
bent-under region will show the medullary canal open both above 
and below, as in figure S</. while sections farther caudad pass 
through the short region where the folds are fused, and we have the 
appearance represented in figure 8c. In figure 8/ is shown a section 
passing posterior to the short, fused region of the folds, and we 
again have the medullary canal open both above and below. Figure 
8g represents a section through the tip of the bent-under region of 
the medullary folds, which are here fused below and open above. 

Figure 8// passes through the posterior part of the head-fold, be- 
tween the limits of the fold of the ectoderm and that of the ento- 
derm. The medullary groove (mg) is here very wide and compar- 
atively shallow ; its walls are continued laterally as the gradually 
thinning ectoderm (ec) . The enteron (cnt) is completely enclosed, 
and forms a large, somewhat compressed, thick-walled cavity. Be- 
tween the dorsal wall of the enteron and the lower side of the medul- 
lary canal lies the notochord I nt), a small, cylindrical rod of closely 
packed cells derived, in this region at least, from the entoderm. In 
the posterior region of the embryo it is not possible to determine 


with certainty the origin of the notochord, owing to the close fusion 
of all three germ layers. Between the wall of the enteron and the 
lower side of the ectoderm is a considerable mass of mesoderm 
(incs), which here consists of more scattered and angular cells than 
in the preceding section. 

Figure 8j shows the appearance of a section through the meso- 
blastic somites, in one of which a small myocoel (myc) is seen. As 
is seen by the size of the figure, which is drawn under the same 
magnification as were all the sections of the series, the embryo in 
this region is much smaller in section than it is toward either end, 
especially toward the anterior end. The medullary groove (wig) is 
still more shallow than in the more anterior sections, and the ecto- 
derm (cc), with which its folds are continuous laterally, is here 
nearly horizontal. The mesoblast (mes) is of a more compact na- 
ture than in the preceding section and shows little or no sign of 
cleavage, although a distinct myocoel may be seen and cleavage is 
well marked in sections between this one and the preceding. 

The notochord ( nt) has about the same appearance as in the pre- 
ceding section, but is more distinctly separated from the surrounding 

Figure 8/ is through the posterior end of the embryo : it shows the 
relation of parts in the region of the primitive streak. Although not 
visible in surface views, and hence not represented in figure 8, the 
medullary groove is continued without any line of demarcation into 
the primitive groove, and the medullary folds into the edges of the 
primitive streak, so that it is impossible to set any definite boundaries 
between these structures unless the dorsal opening of the blastopore 
be taken as the point of division. The medullary groove {wig), if 
it be here so called, is proportionately more shallow than in the pre- 
ceding figure and is actually much wider. The section passes behind 
the posterior end of the notochord, so that structure is not seen. 
Though not so well indicated as might be desired in the figure, the 
three germ layers are here indistinguishable in the middle line, and 
in the center of this mass of cells the blastopore (blp) or neurenteric 
canal may be seen as a small vertical slit. As will be more fully 
described in the following stage, this canal opens dorsally a few sec- 
tions posterior to the one under discussion and ventrally a few sec- 
tions farther toward the head. 

In all the sections of this stage the ectoderm and entoderm are 
fairly thick in the region of the embryo proper, but become thinner 
iintil reduced to a mere membrane as we pass to more distal regions. 
Both layer? are composed of loosely arranged cells, with scattered 


nuclei. Where the ectoderm becomes thickened to form the medul- 
lary folds, the cells are much more compactly arranged; hence this 
region stands out in strong contrast to the rest of the ectoderm. 

Stage VI 
Figures ga-gm (Plates IX, X) 

The embryo represented by this series of transverse sections is 
intermediate in development between those represented in surface 
views by figures 8 and 10. The amnion and head-fold are nearly the 
same as in figure 8; the medullary folds are intermediate in devel- 
opment, the anterior end not showing so marked an enlargement as 
shown in figure io, v' '. There are six or seven faintly distinguishable 

Figure ga represents a section through the anterior part of the 
head-fold ; it shows one unusual condition : the head lies entirely be- 
neath the surface of the yolk. This condition is quite confusing 
when the section is studied for the first time. The pushing of the 
head under the yolk is shown at its commencement in figure II. 
The process continues until nearly the entire anterior half of the 
embryo is covered ; but when the embryo attains a considerable size 
it is seen to lie entirely above the yolk, as in the chick. According 
to Voeltzkow's figures (18), this same condition is found in the 
crocodile, and Balfour (2) also mentions it in connection with the 
lizard. The fusion of the medullary folds has made considerable 
progress, so that the entire anterior end of the canal is enclosed, 
except in the region where the folds are bent down and back, as in 
the preceding stage ; here the folds are still distinct from each other, 
leaving the medullary canal open on the ventral side, as shown in 
figures 9 and cjb. In the section under discussion the ectoderm (cc) 
is a very thin membrane on top of a considerable mass of yolk, while 
no entoderm can be distinguished. The amnion (a) completely sur- 
rounds the embryo as an irregular membrane of some thickness in 
which no arrangement into layers can be seen. The epidermal ecto- 
derm (cp) is composed of the usual loosely arranged cells, so that it 
is clearly distinguishable from the compactly arranged cells of the 
nervous layer («/), from which it is separated by only a line. 

In figure gb, which shows a section a short distance posterior to 
the preceding, the medullary canal (mc) is somewhat deeper and is 
still open ventrally. There is a distinct space between the nervous 
(nl) and epidermal (cp) layers of the ectoderm, in which space a 
few mesoblast cells (mes) may be seen. The section is cut just pos- 
terior to the edge of the amnion, so that there is now neither amnion 
nor yolk above the embryo. 


Figure gc is about ten sections posterior to figure gb. The section 
passes through the anterior wall of the bent-under part of the medul- 
lary canal ( mc'), so that the actual canal is shown only on the dorsal 
side {mc), where it is completely closed and begins to assume the 
shape of the typical embryonic spinal cord. The space between the 
superficial (ep) and nervous (nl) layers of the ectoderm is quite 
extensive and is largely filled by a fairly compact mass of mesoderm 

())ICS) . 

Figure gd, although only five sections posterior to the preceding, 
shows a marked change in structure. The medullary canal (mc) is 
here of the typical outline for embryos of this age. A large, com- 
pact mass of cells (cut) appears at first glance to be the same that 
was noted in the preceding stage at the tip end of the turned-under 
medullary canal ; it is, however, the extreme anterior wall of the 
enteron, which is in close contact with the above-mentioned tip of 
the medullary canal. Between this anterior wall of the enteron, of 
which wall it is really a part, and the medullary canal is the noto- 
chord (nt). The space surrounding the notochord and enteron is 
filled with a fairly compact mass of typical, stellate mesoblast cells. 
The depression of the ectoderm (cc) and entoderm (en) of the blas- 
toderm caused by the formation of the head-fold is here less marked, 
and the dorsal side of the embryo in this region is slightly elevated 
above the level of the blastoderm. 

Figure ge represents a section passing through the posterior edge 
of the head-fold. The epidermal ectoderm is here continuous with 
the thin layer of superficial ectoderm (ec) of the blastoderm, while 
the entoderm (en) of the blastoderm is still continuous beneath the 
embryo. The thick ectoderm of the embryo is sharply differentiated 
from the thin layer of ectoderm that extends laterally over the yolk. 
The pharynx (ent) is a large cavity whose wall is thick except at 
the dorsal side, where it is thin and somewhat depressed, apparently 
to make room between it and the medullary canal for the notochord 

Figure gf is about twenty sections posterior to the preceding sec- 
tion, and passes through the point of separation of the folds of the 
entoderm (en). From this point the entoderm gradually flattens 
out, leaving the enteron unenclosed. The medullary canal (mc) 
and notochord (nt) are about as in the preceding section, but the 
ectoderm (ep) is somewhat thinner and more flattened. The meso- 
derm (mes) on the right side exhibits a distinct cleavage, the result- 
ing body cavity (be) being a large, triangular space. 

Figure gg, the twenty-fifth section posterior to that represented in 


figure gf, shows a marked change in the form of the embryo. While 
of about the same lateral dimensions, the dorso-ventral diameter of 
the embryo in this region is less than one-half what it was in the 
head region. The epidermal ectoderm (ep) is now nearly horizontal 
in position and is not so abruptly separated laterally from the thin 
lateral sheets of ectoblast. The medullary groove (mg) is here a 
very narrow vertical slit. At this stage the fusion of the medullary 
folds has taken place over the anterior third of the embryo. For a 
short distance, represented in about thirty-five sections, the canal is 
open, as in the figure under discussion ; for the next one hundred 
sections (about one-third the length of the embryo) in the region of 
the mesoblastic somites the canal is again closed, while throughout 
the last one-third of its length the canal is widely open dorsally. The 
enteron is here entirely open ventrallv, the entoderm being almost 
flat and horizontal. The notochord (nt) is distinctly outlined and is 
somewhat flattened in a dorso-ventral direction. The body cavity 
(be) is well marked, but is separated by a considerable mass of un- 
cleft mesoblast from the notochord and the walls of the medullary 

A space of about one hundred sections, or one-third the length of 
the embryo, intervenes between figures gg and 91. This is the 
region of the mesoblastic somites, and in this region, as has been 
above stated, the medullary canal is completely enclosed. It is evi- 
dent then that the entire anterior two-thirds of the medullary canal 
is enclosed except for the short region represented in figure Sg. 
Whether or not this short open region between the two longer en- 
closed regions is a normal condition the material at hand does not 

Figure gh represents a typical section in the region of the meso- 
blastic somites just described. It shows the enclosed medullary 
canal (mc), the body cavity (be) on the right, and a mesoblastic 
somite with its small cavity (myc) on the left. The entire section 
is smaller than the sections anterior or posterior to this region, and 
seems to be compressed in a dorso-ventral direction, this compres- 
sion being especially marked in the case of the notochord. 

Figure 91 is through a region nearly one hundred sections pos- 
terior to the preceding, and cuts the embryo, therefore, through the 
posterior one-fourth of its length. The chief difference between this 
and the preceding section is in the medullary canal, which is here 
open and is in the form of a wide groove with an irregular, rounded 
bottom and vertical sides. The size of the section is considerably 
greater than in the preceding, the increase being especially noticeable 


in the notochord ( nt), which is cut near its posterior end. There is 
little or no sign of mesoblastic cleavage. 

Figure gj is about twenty sections posterior to figure gi. The 
medullary groove (mg) is considerably larger than in the more an- 
terior regions', and its folds are somewhat inclined toward each 
other, though still wide apart. The notochord and entoderm are 
fused to form a large, compact mass of tissue close under the ventral 
wall of the medullary groove. On the ventral side of this mass of 
cells a groove (blp) marks the anterior and ventral opening of the 
blastopore shown in the next figure. The mesoblast shows no sign 
of cleavage. 

Figure gk shows the medullary groove (mg) in about the same 
position as in the preceding section. The blastopore (blp) is here 
seen as a small cavity in the center of the large mass of cells that 
was noted in the last figure. The entoderm (en) is continuous from 
side to side, but is not so sharply differentiated from the other germ 
layers as is represented in the figure. 

Figure 9/ is four sections back of the preceding; the wide, dorsal 
opening (blp) of the blastopore or neurenteric canal into the medul- 
lary groove (mg) is shown. The blastopore or neurenteric canal, 
then, is still at this stage a passage that leads entirely through the 
embryo, the medullary canal being in this region unenclosed above. 
Ventrally it is seen as a narrow opening through the entoderm ; it 
then passes upward and backward, behind the end of the notochord, 
as a small but very distinct canal, which may be traced through about 
ten sections. The enclosed portion of the canal lies, as has been 
stated (figure gk, blp), in the center of the mass of cells that is fused 
with or is a part of the floor of the medullary groove. 

The above-described neurenteric canal is essentially like that de- 
scribed by Balfour (2) in the lace rt Hi a, He does not say, however, 
and it is not possible to tell from his figures, whether there is a long, 
gradually diminishing groove posterior to the dorsal opening of the 
canal, as in the alligator. He says that the medullary folds fuse 
posteriorly until the medullary canal is enclosed over the opening of 
the neurenteric canal ; also that "the neurenteric canal persists but a 
very short time after the complete closure of the medullary canal." 

In figure gm, for about thirty sections (one-tenth the entire length 
of the embryo), behind the section represented in the last figure, 
there is a very gradual change in the embryo, converting the deep 
groove, mg in figure 9/, into the shallow slit, pg, figure gm. 

There is no line of demarcation between the typical medullary 
groove region of figure 9/ and the equally typical primitive groove 


region represented in figure gm. As was noted in the preceding 
stage, the medullary folds are quite continuous with the folds of the 
primitive streak, and the medullary groove with the primitive 
groove ; so that, unless we take the dorsal opening of the neurenteric 
canal as the point of separation, there is no line of division between 
these structures. The entoderm (en) and the lateral regions of the 
ectoderm (ec) and mesoderm (mes) in figure gm are about as they 
were in figure g\, but in the middle line is seen a compact mass of 
cells forming the primitive streak (ps), with the shallow primitive 
groove (pg) on the dorsal side. The cells on each side of the prim- 
itive groove and for a short distance below it are compact, as is 
shown in the figure, but as we pass ventrally and laterally they be- 
come looser and more angular to form the lateral sheets of meso- 
blast ( mes), very much as is the case in the chick and other forms. 
For a few sections posterior to the one shown in figure gm the prim- 
itive streak may be seen, then it disappears, and only the ectoderm 
and entoderm remain as thin sheets of tissue above the yolk. 

Stage VII 

Figures 10 and 10a (Plates X, XI) 

Although of practically the same size as the preceding, this stage 
has advanced sufficiently in development to warrant a description. 

The medullary folds are apparently still slightly open for the 
greater part of their length, though they are evidently fused together 
in the head region, except at the extreme end. Transverse sections, 
however, of figure 12, in which the medullary folds, from the dorsal 
aspect, seemed open (mg) as in figure 10, have shown that these 
folds are fused throughout their length. 

The first cerebral vesicle (v') is clearly indicated as an enlarge- 
ment of the anterior end of the nervous system, and a slight enlarge- 
ment (v") posterior to the first probably represents the second 
cerebral vesicle. 

There are now eight pairs of somites (s). 

The head-fold (//) now shows in both dorsal and ventral views, 
appearing in the former, when viewed by transmitted light, as a 
lighter, circular area on either side of the body, just posterior to the 
hinder edge of the amnion. 

The head-fold of the amnion (a) has extended about twice as far 
backward as it did in the preceding stage. 

Owing to the opacity caused by the medullary folds being in con- 
tact along the middle line, the notochord is no longer visible in sur- 
face views. 


The head at this stage begins to push clown into the yolk in a 
strange way that will be described later. 

Stage VIII 
Figures 11-iiJb (Plates XI, XII, XIII) 

This stage is about one-fourth longer than the preceding. The 
medullary canal is enclosed throughout its entire length, though it 
appears in surface view (fig. 11) to be open in the posterior half 
(mc) of the embryo. An enlargement of this apparently open re- 
gion at the extreme posterior end (pg) is probably caused by the 
remains of the primitive groove or the neurenteric canal, and a 
slight opacity at the same point may be caused by the primitive 
streak. The anterior end of the neural tube is bent in a ventral 
direction (V), as in the preceding stage. The somites (s) now 
number fifteen pairs ; they are somewhat irregular in size and shape. 

The head-fold is not so striking a feature as in the preceding 
stage. The head-fold of the amnion (a) now covers nearly two- 
thirds of the embryo. The heart (lit) is seen as a dark, rounded 
object projecting to the right side of the neural canal, just anterior 
to the first somite. The vitelline blood-vessels are just beginning to 
form, but are not shown in the figure. 

The depression of the anterior region that was noted in the pre- 
ceding stage has advanced so far that a considerable part of the 
embryo now projects forward under the blastoderm. In some cases 
it is almost concealed in a dorsal view ; in other cases it may easily 
be seen through the transparent membranes, especially after clearing. 

In opening eggs of this stage one is at first apt to underestimate 
the size of the embryos, since the anterior part of the embryos cannot 
be seen until after they are removed from the yolk and are viewed 
from the ventral side. 

The embryo from which the series of transverse sections of this 
stage was made, while of the same state of development as that 
shown in figure 1 1, was more fully covered by the blastoderm than is 
shown in the surface view in question. 

Figure 11a passes through the tip of the head. Dorsal to the 
embryo is the ectoderm and a thick mass of yolk (y). The amnion 
(a) is seen as an irregular membrane which entirely surrounds the 
head. The medullary canal (mc) is entirely closed, except at the 
extreme anterior end, which is bent downward so that the opening 
is on the ventral side. The nervous (nl) and epidermal (ep) layers 
of the ectoderm are in contact throughout, but are clearly distin- 
guishable because of the difference in the compactness of their cells. 


In figure i lb is represented a section, behind the preceding, which 
passes through the posterior tip of the turned-under anterior end 
(mc'). Here the medullary canal is closed both above (mc) and 
below (mc'). The amnion (a) has about the same appearance as in 
the more anterior section, but there is here a considerable space, 
filled with mesoblast (mes), between the nervous (nl) and epidermal 
(ep) layers of ectoderm. 

Figure uc is twenty sections, about one-tenth the length of the 
embryo, posterior to the one last described. The large mass of over- 
hanging yolk (y) is still present, as is also the amnion (a), though 
the latter no longer passes entirely around the embryo ; only the true 
amnion could be made out. The thickened walls of the medullary 
canal have reduced that cavity to a narrow, Y-shaped slit (mc). 
The notochord (nt) is very slender in this region, compared to its 
diameter farther toward the posterior end. The enteron (ent) is a 
large cavity, whose wall is made up of loosely arranged cells except 
around a median, ventral depression where the cells are more com- 
pact. This depression may be traced through ten or fifteen sections 
and may represent the beginning of the thyroid gland, though this 
point was not worked out with certainty. Surrounding the noto- 
chord and enteron is a loose mass of typical, stellate mesoblast cells 
( mes), which are cleft on either side to form the anterior limit of 
the body cavity (be). Between the bod}- cavity below and the en- 
teron above, on each side, is a small blood-vessel (bv) which when 
followed caudad is found to open ventrally and medially into the 
anterior end of the heart. 

Figure lid is about a dozen sections posterior to the preceding. 
The appearance of the overhanging yolk (y) of the amnion (a) and 
of the notochord (nt) is about as in the more anterior section. The 
medullar}- canal (mc) is a straight, vertical slit, and the depression 
in the floor of the pharynx | cni ) is much more shallow. The body 
cavity (be) is much larger and extends across the mid-ventral line 
beneath the heart (ht), which is cut through its middle region. The 
heart may be traced through about twenty sections (one-tenth the 
length of the embryo) ; its mesoblastic wall (mes') is thin and irreg- 
ular, and is lined by a distinct endothelium (en) whose exact origin 
has not yet been worked out. 

Figure lie is just back of the heart, and shows in its place the 
two vitelline veins (vv). The depression in the floor of the enteron 
(ent) is entirely distinct from the one that has been mentioned 
above, and is simply the posterior limit of the head-fold of the ento- 
derm ; the fifth section posterior to this shows where this depression 


opens ventrally to the yolk sac. The other structures shown in the 
figure are not markedly different from what was seen in figure lid. 

Figure nf is about one-tenth the length of the embryo posterior 
to figure lie. The chief differences here noticed are in the enteric 
and ccelomic cavities. The former is no longer enclosed, a dorsal 
fold in the entoderm being all that remains of the cavity that was 
seen in the more anterior figures, while the latter is here reduced to 
a narrow cleft between the somatic and splanchnic mesoblast. A 
thickening of the mesoblast on either side of the notochord, espe- 
cially on the left, represents a mesoblastic somite. The medullary 
canal (mc) is more open than in the more anterior sections. 

For about one-third of the length of the embryo posterior to figure 
11/ there is a gradual flattening, in a dorso-ventral direction, with 
loss of the amnion, until the condition represented in figure ng is 
reached. The most striking feature of this region is the great 
thickness of the ectoderm (ec), which is still made up of scattered, 
irregular cells. In the middle line, directly over the medullary 
canal (here a nearly cylindrical tube), is a sort of break in the ecto- 
derm, as though there had not been a complete fusion of the epi- 
dermal layer when the nervous layer came together on the closure 
of the medullary groove. This break in the ectoderm may be fol- 
lowed back to the region of the primitive streak, and will be men- 
tioned again. As has been noted, the medullary canal (mc) is 
nearly circular in cross-section, and is closely underlaid by the noto- 
chord (fit), which is several times the diameter that it was in more 
anterior sections. The mesoblast (mes) is a comparatively thin 
layer, intermediate in thickness between the ectoderm and entoderm. 
It shows laterally a slight separation to form the body cavity. 

Figure nh is about ten sections posterior to figure 11^, and dif- 
fers from it chiefly in that the notochord (nt) is continuous with the 
lower side of the medullary canal (mc) , though still distinct from 
the underlying entoderm (en). 

Figure nt, four sections farther from the head, shows the same 
greatly thickened ectoderm (ec) with the same break (ec') in the 
middle line. The section is posterior to the notochord and passes 
through the anterior edge of the blastopore or, as it may now per- 
haps better be called, the neurenteric canal. The cells of the medul- 
lary wall are continuous with those of the entoderm. The meso- 
derm (mes) is still distinct from the other germ layers. 

Figure it/ is the next section posterior to the one just described 
and differs from it only in showing the actual opening of the neuren- 
teric canal (nc) into the medullary canal (mc). The medullary 
canal extends, with gradually diminishing caliber, for about fifteen 


sections posterior to the point at which the neurenteric canal empties 
into it. The mesoblast (vies) is so closely attached to the lower 
wall of the neurenteric canal that it seems to be actually continuous 
with it. 

For a considerable distance posterior to the end of the medullary 
canal we find the structure similar to that shown in figure 11k, 
which is about the twentieth section posterior to figure 117. The 
break (cc') in the ectoderm is here seen as a compact group of cells 
which at first glance seem to be continuous with a rounded mass of 
cells below (ps). Examination under greater magnification, how- 
ever, shows that the two groups of cells are distinct. As the sec- 
tions are followed back of this region, the upper mass of cells (ec') 
gradually disappears, and after its disappearance the lower mass 
(ps), which is already continuous with the mesoderm (mes) on 
either side, becomes continuous with the under side of the ectoderm. 
The mass of cells (ps) is apparently the primitive streak, though it 
is distinct from the ectoderm for a considerable distance posterior 
to the neurenteric canal. Just what may be the meaning of the 
thickened ridge of ectoderm (ec') it is difficult to determine. 

Stage IX 
Figures i2-i2g (Plates XIII, XIV) 

The entire length of the embryo proper is 6.5 mm. from the ex- 
treme posterior end to the region of the midbrain (v 2 ), which now, 
on account of the cranial flexure, forms the most anterior part of 
the body. Besides being slightly longer than the preceding stage, 
the embryo has increased in thickness, especially in the anterior 
region, where the enlargement of the cerebral cavity is considerable. 

Body torsion has begun (fig. 12), so that the anterior third of the 
embryo now lies on its right side, while the rest of the body is still 
dorsal side up. The direction of body torsion does not seem to be 
as definite as it is in the chick, some alligator embryos turning to the 
right side, others to the left. Clarke has illustrated this fact in his 
alligator figures. He says (5) that embryos lie "more frequently 
on the left, but often on the right side." 

The head is distinctly retort-shaped, and at the side of the fore- 
brain (V) a small crescentic thickening is the optic vesicle (e). 
The auditory vesicle, though of considerable size, does not show in 
this surface view. The head-fold (h) extends for about one-third 
the length of the entire embryo, though its exact limit is difficult to 
determine in surface view. There is no sisrn of a tail-fold. 


About seventeen pairs of somites are present. 

The amnion extends over the anterior two-thirds of the embryo. 

The above-mentioned increase in the diameter of this embryo 
over that of the preceding is evident when the first two transverse 
sections of this stage are compared with the corresponding sections 
of the earlier stage ; in the middle and posterior regions there is not 
very much difference in size. 

Figure 12a passes through the region of the forebrain. This end 
of the embryo lies on its side, as was noted above and as may be 
recognized from the relative positions of the head and the overlying 
yolk (3/). The great size of this and the following figure is due 
partly to the increase in size mentioned above and partly to the fact 
that the sections pass through the region of cranial flexure. The 
present figure (12a) represents the brain cavity as large and dumb- 
bell-shaped, with comparatively thick walls of compactly arranged 
cells. The ventral end of this cavity (fb) is cut anterior to the 
region of the optic vesicles, while the dorsal end (nib) may perhaps 
be called the midbrain. In the sections that follow this one the two 
cavities are distinct from each other. The medullary canal, as was 
stated above, is now completely enclosed, except for the ventral 
opening of the neurenteric canal, to be presently noticed. Sur- 
rounding the brain is a considerable mass of mesoblast (mcs). It 
is composed of the typical stellate cells. The ectoderm (ec) is made 
up of the same irregularly and loosely arranged cells that have been 
seen in earlier stages ; it is of unequal thickness in different regions, 
the thicker parts being at the sides. The amnion (a) has the usual 
appearance, and in this region of course completely surrounds the 

Figure 12b is ten sections posterior to the section just described. 
The width of the embryo is greater in this region, but the dorso- 
ventral diameter is about the same as in the more anterior section. 

The overlying yolk and blastoderm are not shown in any figure of 
the series except the first. In this figure the forebrain (fb) and 
midbrain (nib) are widely separated instead of being connected, as 
in the preceding figure, where the section passed through the actual 
bend of the cranial flexure. The anterior and ventral part of the 
cranial cavity, the forebrain (fb), is nearly circular in outline. It 
exhibits on one side a well-marked optic vesicle (ov), which is suffi- 
ciently advanced in development to show a rudimentary optic stalk. 
The outer wall of the optic vesicle is in close contact with the super- 
ficial ectoderm, which shows as yet no sign of the formation of a 
lens vesicle. The plane of the section being probably not quite at 
right angles to the long axis of the embryo, the optic vesicle of one 


side only was cut. The wall of this part of the forebrain is of about 
the same thickness and appearance as in the preceding stage. The 
other cerebral cavity (mb) of this section is probably the hinder part 
of the midbrain, though it may be the anterior part of the hind- 
brain ; there is no sharp line of demarcation between these regions 
of the brain. This cavity (mb) is much smaller in section than the 
forebrain ; its walls are of about the same thickness. 

Ventral to the midbrain is the anterior end of the notochord (nt), 
surrounded by the mesoblast. At various places throughout the 
mesoblast irregular open spaces may be seen ; these are blood- 
vessels. The ectoderm (cc) and amnion (a) have about the same 
appearance as in the preceding figure, though the former seems 
somewhat thinner. 

Figure 12c is just back of the bent-under forebrain represented in 
the preceding figure and in front of the main body of the heart. 
The plane of the section not being at right angles to the long axis 
of the body (as was mentioned above), the figure is not bilaterally 
symmetrical. The neural canal, since the section passes through 
the auditory vesicles, may here be called the hindbrain (hb). It has 
an almond-shaped cavity, surrounded by a wall of medium thickness. 
In close contact with the wall of the hindbrain, on each side, is the 
inner side of the auditor}- vesicle (0) , which is seen as a deep, wide- 
mouthed pit in the superficial ectoderm. On the right side of the 
section the auditory pit is cut through its middle region ; it is simply 
a thickened and condensed area of the ectoderm which has been in- 
vaginated in the usual way. Directly beneath the hindbrain is the 
notochord (nt), on each side of which, in the mesoblast, is the dorsal 
aorta (ao), or rather the continuation of the aorta into the head. 
Beneath these structures and extending from one side of the section 
to the other is the pharynx (ph) ; its lining wall is fused on each 
side with the ectoderm, but there is no actual opening to the. ex- 
terior. These points of contact (g) between entoderm and ecto- 
derm are of course the gill clefts; they are not yet visible from the 
outside. The roof of the pharynx is flat and comparatively thin, 
while the floor is thickened and depressed to form a deep, wide pit, 
traceable through six or eight sections. This pit may be the thyroid 
gland already noticed in the preceding stage. Below the main 
cavity of the pharynx and close to each side of the thyroid rudiment 
just mentioned is a large blood-vessel (tr). These two vessels 
when traced posteriorly are found to be continuous with the anterior 
end of the heart, and hence may be called the truncus. They were 
noticed in figure lie, bv. The ectoderm surrounding the lower side 


of the embryo was so thin and indistinct that it could scarcely be 
distinguished from the mesoderm of. that region. The amnion (a) 
is still a continuous envelope entirely surrounding the embryo. 

Figure \2.d, about twenty sections posterior to figure 12c, is in 
the posterior hear*t region. The spinal cord (sc), as might be ex- 
pected, is smaller than in the more anterior region, but is otherwise 
not markedly different from what was there seen. The notochord 
(lit) also has the. same appearance as before. The enteron (ent) 
shows of course in this region no gill clefts ; it is a small, irregular 
cavity with thicker walls than in the figure just described. The 
ventro-lateral depression is entirely distinct from the depression that 
was called the thyroid rudiment in the preceding figure. Dorsal to the 
enteron are the two dorsal aortse (ao), now smaller and more ventral 
to the notochord than in the preceding figure. Ventral to the enteron 
is the large heart (lit), projecting below the body cavity, which is no 
longer enclosed. The mesodermic wall (mes f ) of the heart is still 
comparatively thin and is separated by a considerable space from 
the membranous endocardium (en'). The extent and shape of the 
heart are shown in the surface view of this stage. On the right side 
of the section the body cavity extends to a point nearly opposite the 
middle of the spinal cord, considerably dorsal to the notochord, 
while on the left side the dorsal limit of the body cavity is scarcely 
level with the lower side of the notochord. Between the dorsal end 
of the body cavity and the side of the spinal cord, on the left, is a 
dense mass of mesoblast (s), one of the mesoblastic somites. A few 
sections either anterior or posterior to the one under discussion will 
show the condition of the two sides reversed — that is, the body 
cavity will extend to the greater distance on the left and will be 
interrupted by a mesoblastic somite on the right. It is evident, 
then, that the upper angle of the body cavity is extended dorsally 
as a series of narrow pouches between the somites. The mesoblast 
that lines the body cavity, the splanchnopleure (sin), and somat- 
opleure (so) is somewhat denser than the general mass of meso- 
blast, so that these layers are quite distinct, the former (sm) extend- 
ing around the enteron (ent) and heart (lit), and the latter (so) 
being carried dorsal ward as the mesoblastic part of the amnion (a). 
The amnion may be traced through about 130 of the 200 sections 
into which this embryo was cut. 

Figure I2£ is nearly one-fourth the length of the embryo posterior 
to figure I2d; it is approximately in the middle region. The diam- 
eter of the embryo has been gradually decreasing until now it is 
very much less than in the head region. The section being behind 
the head-fold the entoderm (en) is nearly flat and the enteron is 


quite unenclosed. The canal of the spinal cord (sc) is smaller in 
proportion to the thickness of its walls, and the notochord (rii) is 
somewhat larger than in the preceding sections. In proportion to 
its extent, the ectoderm is very thick. Under the notochord the 
dorsal aortse (ao) are seen as two large, round openings in the 
mesoblast. On the left side the section passes through the center 
of a somite and shows a small, round myocoel (myc). The meso- 
blastic layer of the amnion (so^ is distinct throughout from the ecto- 
blastic layer (a). 

The most important structures to be here noted are the first rudi- 
ments of the Wolffian ducts (wd). They are seen in the present 
section as lateral ridges of mesoblast projecting outward and up- 
ward toward the ectoblast, which suddenly becomes thin as it passes 
over them. These ridges or cords of mesoblast are as yet quite 
solid. They arise suddenly at about the eightieth section of the 
series of two hundred and may be traced through about forty sec- 
tions, or one-fifth of the length of the embryo. Their exact length 
is difficult to determine because, while their anterior ends are blunt 
and sharply defined, they taper so gradually posteriorly that it is 
hard to tell just where they end. They apparently originate ante- 
riorly and gradually extend toward the tail. In a slightly younger 
embryo the rudimentary Wolffian duct could be seen as a still 
smaller rod of cells extending posteriorly for a few sections, from 
the seventy-fifth section of a series of about two hundred. In the 
particular series under discussion the left rudimentary Wolffian duct 
was about one-fifth longer than the right one. 

Figure 12/ is just posterior to the head-fold of the amnion, pass- 
ing, in fact, on the left side through the extreme edge of its lateral 
fold, which is shown as an upward bend in the ectoblast and somat- 

The ectoblast (cc) shows the same remarkable thickening that 
was noted in the corresponding region of the preceding stage. The 
spinal cord (sc), notochord (nt), aorta? (ao), and entoderm (en) 
need no special mention. The mesoderm seems to be separated by 
unusually wide spaces from both ectoderm and entoderm, and is 
made up of rather closely packed cells except around the aorta?, 
where there seems scarcely enough tissue to hold these vessels in 
place. The body cavity (be) is large, and a small myocoel (myc) 
is seen on the left. 

Figure I2g is through the neurenteric canal (11c), a distinct open- 
ing through the floor of the spinal canal. The section is of course 
just back of the posterior end of the notochord. The entoderm 
(en) along the margin of the neurenteric canal is naturally contin- 


nous with the wall of the spinal cord (sc). The ectoderm (ec) is 
thicker than ever, except in the median plane, where it passes over 
the spinal cord. The mesoblast fs more abundant than in the pre- 
ceding figure, and shows on the left what appears to be a distinct 
myocoel (inyc), though in surface view the mesoblastic somites do 
not extend this far toward the tail. 

Stage X 
Figures 13-132 (Plates XIV, XV, XVI) 

This embryo (fig. 13) is about 5 mm. in length, and hence is 
slightly smaller than the preceding stage, though somewhat more 
advanced in development. The medullary canal is still apparently 
unclosed for a short distance at the extreme posterior end ; this ap- 
pearance is probably due to the neurenteric canal (nc) and to the 
thinness of the roof of the medullary canal rather than to any lack 
of fusion of the medullary folds. The optic vesicle is more distinct 
than in the preceding stage ; a somewhat similar, though smaller, 
opacity (0) marks the position of the ear. There are now about 
twenty pairs of somites, though it is difficult to determine their exact 
number on account of the torsion of the body. The amnion is at 
about the same stage of development as in stage ix. The heart 
(ht) is a large double mass, whose outlines may be dimly seen when 
the embryo is viewed by transmitted light. The vitelline vessels 
(vv) are still but faintly outlined in the vascular area; the veins and 
arteries cannot yet be distinguished from each other. The gill 
clefts, though not visible externally in the embryo drawn, may be 
seen in sections of this stage as evaginations of the wall of the 

The transverse sections of this stage are slightly more advanced 
in development than was the embryo that has just been described in 
surface view. Only those sections have been figured which show a 
decided advance in the development of some special structures over 
their condition in the preceding stage. The sections of the pre- 
ceding stages were drawn under a magnification of eighty-seven 
diameters ; those of this and the following stage were drawn under 
a magnification of only forty-one diameters. All of the figures have 
been reduced one-half in reproduction. 

Figure 13a is the most anterior section of this series to be de- 
scribed. On account of the cranial flexure, which causes the long 
axis of the forebrain to lie at right angles to that of the spinal cord, 
this section cuts the head region longitudinally. The ectoderm (ec) 
3— al 


is of varying thickness, the thickest areas being on each side of the 
forebrain ; it is more compact than in the earlier stages, and, owing 
to the low magnification under which it is drawn, it is represented 
here bv a single heavy line. Under this magnification only the 
nuclei of the mesoderm cells (mes) can be seen, so that this tissue is 
best represented by dots, more closely set in some places than in 
others. The forebrain is an elongated cavity (fb) with thick, dense 
walls. Attached to each side of the forebrain is an optic vesicle 
( ov ). which is considerably larger than in the preceding stage. The 
connection between the cavity of the forebrain and that of the optic 
vesicle is not seen in this section ; it is a wide passage that may be 
seen in several sections posterior to the one under discussion. The 
beginning of the invagination of the optic vesicle to form the optic 
cup may be seen on both sides, but more plainly on the right. On 
the right side also is noticed a marked thickening of the ectoderm, 
which is invaginated to form a small pit, the lens vesicle (lv) ; on 
the left side the section is just behind the lens vesicle. Above the 
optic stalk on each side, in the angle between the optic vesicle and 
the side of the forebrain, is a small blood-vessel (bv) . Several 
other blood-vessels may be seen at various places in the mesoblast, 
four of them, near the pharynx being especially noticeable. The 
hindbrain (hb) is wider than, but not so deep as, the forebrain; its 
walls are very thick laterally, but are thin on the dorsal and ventral 
sides. The dorsal wall is reduced to a mere membrane, which, with 
the overlying ectoderm, has been pushed into the brain cavity, as is 
generally the case with such embryos. Close to the ventral wall of 
the hindbrain the notochord (///) is seen. The character of the 
notochord has already begun to change; the cells are becoming 
rounded and vacuolated, with but few visible nuclei except around 
the periphery of the notochord. Near the center of the section, 
close to the ventral end of the forebrain, is the pharynx (ph) , cut 
near its anterior limit ; it is here a small, irregularly rectangular cav- 
ity with a comparatively thin wall. On the left side of the pharynx 
the first gill cleft (g) is indicated as a narrow diverticulum reach- 
ing toward the ectoderm. A few sections posterior to this one the 
first gill cleft is widely open to the exterior. As has been said, in 
the surface view of this stage above described none of the gill clefts 
showed ; so that in this respect at least the sectioned embryo was 
more nearly of the state of development of the embryo represented 
in figure 14, to be described later. 

Figure 13/', about forty sections posterior to figure 130, passes 
through the hindbrain in the region of the ears. Being back of the 
region affected by cranial flexure, this section is of course of much 


less area than the preceding. The ectoderm shows no unusual 
features ; it is of uniform thickness except where it becomes con- 
tinuous with the entoderm around the mandibular folds (md) ; 
there it is somewhat thickened. The most striking feature of the 
section is the presence of two large auditory vesicles (o). The sec- 
tion being not quite at right angles to this part of the embryo, the 
vesicles are not cut in exactly the same plane ; the one on the left is 
cut through its opening to the exterior, while the one on the right 
appears as a completely enclosed cavity. In a section a short dis- 
tance posterior to this one the appearance of the vesicles would be 
the reverse of what it is here. As may be seen in the figure, the 
vesicles are large, thick-walled cavities lying close to the lateral 
walls of the hindbrain. The hindbrain itself has the usual trian- 
gular cross-section, with thick lateral walls and a thin, wrinkled 
dorsal wall. Close to the ventral side of the hindbrain lies the noto- 
chord (nt), on each side of which, in the angle between the brain 
and the auditory vesicles, is a small blood-vessel (bv). Ventral to 
these structures and close to the dorsal wall of the pharynx (ph) 
are the two large dorsal aortse (ao). The ventral side of the section 
passes through the open anterior end of the pharynx (ph). On the 
left is seen the widely open hyomandibular cleft (#'), between the 
main body of the section and the mandibular arch (md). On the 
right side the plane of the section was such that the hyomandibular 
cleft was not cut through its external opening. In each mandibular 
fold a large aortic arch (ar) is seen, and also a slight condensation 
of mesoblast, the latter probably being the forerunner of cartilage. 
Figure 13c passes through the anterior part of the heart about 
seventy-five sections posterior to figure 13&. The embryo in this 
region is narrow but deep (dorso-ventrally), the depth being largely 
due to the size of the heart. The ectoderm (ec) is considerably 
thickened on each side of the pharynx (ph) ; this thickened area 
may be traced for some distance both anteriorly and posteriorly 
from this point ; its significance could not be determined. The 
spinal cord (sc) and notochord (nt) need no special description; 
the former is smaller and the latter larger than in the more anterior 
sections. The two large blood-vessels (ac) near the spinal cord 
and notochord are probably the anterior cardinal veins. The aortae 
are cut by the plane of this section just anterior to their point of 
fusion into a single vessel. A few blood corpuscles are seen in 
the right aorta. The enteron (cut), cut posterior to the region of 
the gill clefts, is a large elliptical cavity, with its long axis in a 
transverse position. Its entodermal wall is comparatively thin and 
smooth, with the cell nuclei arranged chiefly on the outer side, i. e., 


away from the cavity of the enteron. The body cavity (be) is here 
still unenclosed, and its walls, the somatic stalk, are cut off close to 
the body of the embryo. The heart (lit), the most conspicuous 
feature of this section, is nearly as large in cross-section as all the 
rest of the embryo. As seen in such a section it is entirely detached 
from the body of the embryo, and in this particular case has about 
the shape of the human stomach. The mesoblastic portion of its 
wall (mes') is of very irregular thickness; it forms a dense layer 
entirely around the outside, except for the pointed dorsal region, 
and is especially thick along the ventral margin, where it is thrown 
into well-marked folds, the heavy muscle columns. Lining the 
cavity of the heart is the membranous endothelium {en'), and be- 
tween this and the dense outer wall just described is a loose reticular 
tissue with but few nuclei. 

As the series is followed toward the tail the sections diminish in 
size until, at a point about one-third the embryo length from the 
posterior end, they are of scarcely one-fourth the area of the sec- 
tions through the region of the hindbrain. 

Figure 130? is about one hundred and twenty-five sections pos- 
terior to figure 13c. Although not so small as the sections that 
follow it, this section is considerably smaller in area than the one 
last described. The amnion (a), which was not represented in the 
last three figures, is very evident here. The spinal cord (sc) is 
considerably smaller here than in the preceding figure, while the 
notochord (nt) is not only relatively but actually larger than in the 
more anterior regions. Beneath the notochord is the aorta (ao) y 
now a single large vessel. The mesoblast on each side of the body 
is here differentiated into a distinct muscle plate (mp). These 
muscle plates have very much the appearance of the thickened ecto- 
derm seen in the younger stages of development. At about its 
middle region (i. e., at the end of the reference line ec) each muscle 
plate is separated from the overlying ectoderm by an empty space ; 
this space is still more marked in some other series. Ventral to the 
aorta, and supported by a well marked though still thick mesentery 
(ms), is the intestine. It is a small, nearly cylindrical tube with 
thick walls ; the splanchnic mesoblast which surrounds it is more 
dense than the general mass of mesoblast ; it was somewhat torn 
in the section and is so represented in the figure. The urinar) 
organs have made considerable progress since the last stage. In 
the figure under discussion they are seen as a group of tubules on 
either side of the aorta. The tubule most distant from the middle 
line, on each side, is the Wolffian duct {wd). It extends through 
the posterior two-thirds of the embrvo and varies in diameter at 


different points ; it is usually lined with a single layer of cubical 
cells which contain large nuclei. The Wolffian bodies (wt) are a 
mass of slightly convoluted tubules that may be traced throughout 
the greater part of the region through which the Wolffian duct 
extends. These tubules also vary somewhat in diameter, but they 
are usually of greater caliber than the duct. No actual nephros- 
tomes are to be seen, though the occasional fusion of a tubule with 
the peritoneal epithelium, as is seen on the left side of the present 
figure, may represent such an opening. A detailed description of 
these structures may be given in a subsequent paper. 

Figure 13c is about one hundred and forty sections posterior to 
the section just described. The embryo is here very slender, so 
that the contrast between this and the first figure (13a) of this stage 
is remarkable. Except in size, this section does not differ greatly 
from the preceding. The spinal cord, notochord, etc., are smaller 
than before, but are of about the same relative size. The mesen- 
tery (ms) in the section drawn was torn across, so that the intestine 
is not represented. Medial to the Wolffian duct is a tubule (ut ), 
which seems to be the same as those which were called Wolffian 
tubules in the preceding stage, but which may be the beginning 
of the ureter. 

Figure 13/, about two hundred and fifty sections posterior to the 
last, passes through the extreme posterior end of the embryo. The 
section is nearly circular in outline and is somewhat larger than 
the preceding. The amnion (a) completely encircles the embryo. 
The ectoderm (cc) is of fairly even thickness, and the mesoblast 
which it encloses is of the usual character. The spinal cord (sc) 
is nearly circular in outline, as is its central canal. The digestive 
tract (ent) is larger in section than it was in more anterior regions; 
it is nearly circular in cross-section and its walls are made up of 
several layers of cells, so that it resembles to a considerable degree 
the spinal cord of the same region. In the narrow space between 
the spinal cord and the hindgut is seen the notochord (nt), some- 
what flattened and relatively and actually smaller than in the pre- 
ceding figure. A few scattered blood-vessels may be seen in the 
mesoblast at various places. 

A sagittal section of an embryo of this stage, drawn under the 
same magnification as were the transverse sections, is shown in 
figure it>S- The embryo being bent laterally could not be cut by 
any one plane throughout its entire length, so that only the ante- 
rior end is represented in the figure. The amnion (a) may be 
clearly seen except at certain places where it is closely adherent to 


the superficial ectoderm. Under the low magnification used the 
superficial ectoderm cannot be distinguished from the ectoderm of 
the nervous system. The plane of the section being in the anterior 
end almost exactly median, this part of the central nervous system 
is seen as the usual retort-shaped cavity, while in the region back 
of the brain, where the neural canal is narrow, the section passes 
through the wall of the spinal cord (sc) and does not show the 
neural canal at all. The wall of the forebrain (fb) is quite thick, 
especially at the extreme anterior end ; the wall of the midbrain 
{nib), where the marked cranial flexure takes place, is somewhat 
thinner, and it gradually becomes still thinner as it is followed 
posteriorly over the hindbrain (lib). Between the floors of the fore- 
and hindbrains, in the acute angle caused by the cranial flexure, is 
the anterior end of the notochord (///), the only part of that struc- 
ture that lies in the plane of the section. Ventral and posterior to 
the notochord is a large cavity, the pharynx (/>//'), whose ento- 
blastic lining can scarcely be distinguished under this magnification 
from the surrounding tissues. The stomodeal opening being as yet 
unformed, the pharynx is closed anteriorly ; posteriorly also, owing 
to the plane of the section, the pharynx appears to be closed, since 
its connection with the yolk stalk is not shown. In the floor of the 
pharynx, almost under the reference line ph, a slight depression 
marks the position of the first gill cleft. In the mesoblast ventral 
to the pharynx and near the gill cleft just mentioned, a couple of 
irregular openings represent the anterior end of the bulbils arteri- 
osus, posterior and ventral to which is the heart (lit), a large, 
irregular cavity. The dorsal aorta (ao) may be seen as an enlon- 
gated opening in the mesoblast, extending in this section from the 
middle region of the pharynx to the posterior end of the figure 
where it is somewhat torn. Two of the eighteen or twenty pairs 
of mesoblastic somites possessed by this embryo are shown at the 
posterior end of the figure (s), where the plane of the section was 
far enough from the median line to cut them. 

Stack XI 

Figure 14 (Plate XVI) 

Only the anterior region of this embryo is shown in the figure, 
which is a ventro-lateral view. While there is some change in the 
general shape and in parts of the head, the reason for figuring this 
stage is to show the first gill cleft (g'), which lies at an acute angle 
to the long axis of the neck behind the eye (e). The cleft is narrow 


but sharp and distinct in outline; it shows, neither in this nor in 
che following stages, the branched. Y-shaped outline mentioned by 

Stage XII 
Figures 15-15/" ( Plates XVI, XVII) 

In this stage, also, only the anterior region of the embryo is 
figured in surface view. The shape of the head is about the same 
as in the preceding stage, but it is drawn in exact profile. Three 
gill clefts (g 1 ' 3 ) are now present, and are wide and distinct. The 
first cleft, as in the preceding stage, lies at an acute angle to the 
long axis of the pharynx and nearly at right angles to the second 
cleft. The third cleft sends a wide branch (g*) toward the pos- 
terior, as has been described by Clarke, from which, or in connec- 
tion with which according to Clarke, the fourth cleft will develop. 
All three clefts may be distinctly seen to open entirely through the 
pharyngeal wall. The outlines of the visceral folds, especially of 
the mandibular, begin to be apparent. The nasal pit (n) now 
shows as a round depression in front of the more definitely outlined 
eye (e). The auditory vesicle (0) is so deep beneath the surface 
that it may be seen only by transmitted light. 

Figures 15a-? represent transverse sections of an embryo of 
about this general state of development, except that the gill clefts 
are not so definitely open as in the surface view. 

Figure 15a, the most anterior section of the series, passes through 
the forebrain (fb) in the region of the eyes, and through the hind- 
brain {lib) anterior to the auditory vesicles. The forebrain is here 
a large cavity with a dense wall of a comparatively even thickness. 
Owing probably to the section not being exactly in the transverse 
plane, the eyes are cut in different regions, that on the left (ov) 
being cut through its stalk, while that on the right (oc) is cut near 
its middle region and hence does not show any connection with the 
forebrain. The almost complete obliteration of the cavity of the 
optic vesicle to form the optic cup by the invagination of the outer 
wall of the vesicle is shown on the right side of the section (oc). 
The lens vesicle (Iv) is completely cut off from the superficial ecto- 
derm (cc), which is comparatively thin. The hindbrain (lib) has 
the usual shape for that structure. Its ventral wall is dense and 
thick, while its roof is reduced to the usual thin, wrinkled mem- 
brane. Close to the floor of the hindbrain lies the notochord (lit), 
which is large and is distinctly vacuolated. To the right of the 
hindbrain a large mass of darkly stained cells (en) is one of the 


cranial nerves, which is connected with the hindbrain a few sections 
anterior to the one under consideration. The pharynx (ph), which 
is cut near its extreme anterior end, is represented by three irregular 
cavities near the base of the forebrain. Scattered throughout the 
mesoblast, which makes up the greater part of the section, are 
numerous blood-vessels (bv). 

Figure 15& is twenty sections posterior to figure 15a and passes 
through the tip of the bent-under forebrain (fb). On the left the 
section is anterior to the optic vesicle and barely touches the side 
of the optic stalk, which is seen as a small lump on the ventro- 
lateral wall of the brain. On the right the connection of the optic 
vesicle (ov) with the forebrain is shown. Dorsal to the optic vesicle 
just mentioned is a markedly thickened and slightly invaginated re- 
gion of the ectoderm ( 11) ; this is the nasal pit ; on the left side of the 
figure the thickening is shown, but the section did not pass through 
the invagination. The hindbrain (lib) is somewhat narrower than 
in the preceding figure, but is otherwise about the same ; the origin 
of a cranial nerve is seen on its left side (en). The notochord (nt) 
has the same appearance as in the preceding section. A number of 
blood-vessels may be seen, the pair lying nearest the notochord 
being the aortoe (ao), while the two other pairs, on either side of 
the fore- and hindbrains, are the anterior cardinals (ac). The first 
aortic arches are shown at ar. On the left the section passes through 
the exterior opening of the first gill cleft (g'), so that the mandibu- 
lar fold (md) on that side is a distinct circular structure, made of a 
dense mass of mesoderm surrounded by a rather thick ectoderm. 
The mesoderm of this fold is especially dense near the center, prob- 
ably the beginning of the visceral bar. Near the center is also seen 
the aortic arch that has already been mentioned. On the right the 
section does not pass through the external opening of the first gill 
cleft (g f ) so that the tissue of the mandibular fold is continuous 
with the rest of the head. It is of course the slight obliquity of the 
section that causes the pharynx (ph) to be completely enclosed on 
the right, while on the left it is open to the exterior both through the 
gill cleft and between the mandibular fold and the tip of the head. 
The superficial ectoderm shown here as a heavy black line varies 
considerably in thickness, being thickest in the region of the nasal 
pit already mentioned and thinnest over the roof of the hindbrain. 
The amnion (a) in this, as in the other sections of the series, has 
the appearance of a thin, very irregular line. 

Figure 15c is posterior to the region affected by cranial flexure 
and so shows only one region of the embryo, that of the hindbrain 
(hb), which is here of essentially the same structure as above de- 


scribed. On each side of the hindbrain is a large auditory vesicle 
(o) ; that on the left is cut through its center and shows the begin- 
nino- of differentiation, its lower end being thick-walled and 
rounded, while its upper end is more pointed and has a thin, some- 
what wrinkled' wall. The notochord (nt) is slightly larger than in 
the more anterior sections'. Numerous blood-vessels (bv, ar) are 
seen in the mesoblast. The pharynx (pli) is here open ventrally 
and also through the gill cleft of the left side ; on the right side the 
plane of the section did not pass through the external opening of the 
cleft. The mesoblast of the visceral folds is much more dense than 
that of the dorsal region of the section. 

Figure I5(/, as is evident, is a section through the region of the 
heart, which appears as three irregular cavities (ht) with fairly 
thick mesoblastic walls (mes') lined with endothelium (en'). The 
body wall, though consisting of but little besides the ectoderm (ec), 
completely surrounds the heart, and the pericardial or body cavity 
thus formed extends dorsally as a narrow space on either side of 
the foregut, giving the appearance of a rudimentary mesentery, 
though no especial development of such a structure would naturally 
be expected in this region of the embryo. The foregut (ent) is a 
moderately large cavity lined with a very distinct entoderm of even 
thickness. Dorsal to the foregut are three large blood-vessels, a 
median, and now single, dorsal aorta (ao), and a pair of cardinal 
veins (cv). The notochord (nt) is small and is flattened against 
the ventral side of the spinal cord (sc), which latter structure needs 
no special mention. The muscle plates (mp) are considerably 
elongated, so that they now extend ventrally to a point slightly 
below the upper angles of the body cavity. 

Figure 15^ is through the middle region of the embryo, and, 
owing to the curvature of the body, is not an exact dorso-ventral 
section; this accounts, in part at least, for the unusual diameter in 
a dorso-ventral direction of the aorta (ao) , which is very large in 
proportion to the other structures. The posterior cardinal vein is 
shown on the left, but not on the right. The relative sizes of the 
spinal cord (sc) and notochord (nt) are very different from what 
was seen in the preceding figure. In this section the spinal cord 
is considerably smaller than in the preceding, while the notochord 
is very much larger; in fact the notochord here seems abnormally 
large when compared to corresponding sections of other series. It 
is true, however, that while the spinal cord has been diminishing 
in diameter the notochord has been increasing. The spinal cord, 
notochord, and dorsal aorta are all so large that they are flattened 
against each other, the pushing in of the ventral side of the spinal 


cord being even more marked than is shown in the figure. ( )n 
either side of the spinal cord a large spinal ganglion (sg) is sien, 
closely wedged in between the spinal cord and the adjacent muscle 
plate ( mp ) . As in the preceding stage, there is a marked space 
between the muscle plate and the adjacent ectoderm (ec). The 
somatic mesoblast at the upper angle of the unenclosed body cavity 
is thickened on each side and somewhat bulged out by the Wolffian 
body to form what might be termed a Wolffian ridge (wr). In the 
mid-ventral line is the considerably developed mesentery ( ms ) , 
from which the intestine has been torn. The Wolffian bodies now 
consist, on each side, of a group of five or six tubules (wt) of 
various sizes, near which in a more ventro-lateral position, close to 
the upper angle of the body cavity, is the more distinct Wolffian 
duct (wd). The allantois is fairly large by this time, and may be 
seen in the most posterior sections as an irregular, thick-walled out- 
growth from the hindgnt. 

A horizontal section through the anterior end of an embryo of 
this age is shown in figure 15/. Wdiile enclosed of course in the 
same membranous amnion (a), the pharyngeal region of the section 
is >eparated by a considerable space from the more anterior region 
where the section passes through the forebrain (fb) and eyes. The 
spinal cord (so, notochord (nt), muscle plates (mp), aorta; (ao), 
and anterior cardinal veins (ac) need no special description. The 
appearance of the pharynx [pit), with its gill clefts and folds, is 
quite similar to that of the corresponding structures in the chick. 
None of the four clefts (g 1 ' 4 ) show, in the plane at which the sec- 
tion was cut, any connection with the exterior; in fact the fourth 
cleft (g*) would scarcely be recognized as a cleft if seen in this 
section alone. One or two of the more anterior clefts are open to 
the exterior. Three pairs of aortic arches are seen, and each vis- 
ceral fold has a central condensation of mesoblast. 

Stage XIII 

Figures i6-i6g (Pi.atks XVII, XVIII) 

The embryo (tig. 16) now lies on one side, body torsion being 
complete. The curvature of the body is so marked that the exact 
length is difficult to determine. The eye (c) and ear ( o) have 
about the same superficial appearance as in the preceding stage. 
The nose is not shown in this figure. About thirty somites are 
present; the exact number cannot be determined in surface view. 
The amnion is complete, though not shown in the figure, and the 
tail (/ ) is well formed. The umbilical stalk was torn in the removal 


of the embryo, so that it is not shown in the figure. The dim out- 
line of the now convoluted heart may be seen if the "cleared" 
embryo be viewed by transmitted light ; it is not shown in the 
figure. The allantois (al) is a rounded sac of considerable size just 
anterior to the' tail. Four gill clefts (g 1 * 4 ) are now present; the 
most posterior one is more faint than is represented in the figure, 
and it could not be definitely determined from a surface view 
whether or not it opened to the exterior. The mandibular fold 
( md) is now fairly well outlined, but there is as yet no sign of the 
maxillary process. 

Figure 16a is the most anterior of a series of transverse sections 
made of an embryo of the approximate age of the surface view just 
described; it passes through the tip of the forebrain (fb) and shows 
the nasal pit ( // ) of the right side. The great thickening of ecto- 
derm in the region of the nasal invagination is represented by a 
solid line. Owing to the obliquity of the section, the left nasal pit 
was not cut. The mesoblast is quite dense and contains two or three 
small blood-vessels near the roof of the brain. The plane of this 
section, owing to the cranial and body flexure, cut the embryo also 
in the region of the pharynx ; this part of the section was, as a 
matter of convenience, omitted from the drawing. 

Figure 16b is in reality more anterior in position considering the 
entire embryo, than the preceding; but the region of the embryo 
represented is most posterior, so that it is described at this point. 
The greatly elongated outline of the brain is due to its being cut 
through the region of flexure, so that the forebrain ( fb) or, per- 
haps, midbrain, is shown at one end and the hindbrain (hb) at the 
other. The walls of these cavities are somewhat wrinkled and irreg- 
ular and their constituent cells are beginning to show slight differ- 
entiation, though this is not shown in the figure. On the left side 
are seen a couple of darkly stained masses ; one is the origin of a 
cranial nerve (en) ; and the other is one of the auditory vesicles 
(o), which is still more irregular in outline than it was in the pre- 
ceding stage. The only blood-vessels to be seen are a few very 
small ones that lie close to the wall of the brain. The ectoderm is 
quite thin at all points. 

Figure i6r, the largest section of this series, passes through the 
forebrain in the region of the eyes and through the gill clefts. The 
forebrain (fb) exhibits on the left a marked thickening of its wall 
(ch ), the edge of the cerebral hemisphere of that side, which is just 
beginning to develop ; on its right side the lower part of the fore- 
brain is connected by a well-marked optic stalk ( os) with the optic 


cup (oc), ill whose opening lies the lens vesicle (Iv), now reduced 
to a crescentic slit by the thickening of its posterior wall. The 
mesoblast is more dense in those parts of the section adjacent to the 
pharynx than in the more distant regions, and the ectoderm thickens 
in a marked way as it approaches the borders of the pharynx and 
gill clefts. Only a few small blood-vessels (bv) are to be seen in 
the region of the forebrain. 

Parts of three pairs of clefts (g) are shown in the figure: one 
pair opens widely on either side, so that there is a large area of the 
section that is distinct from the two still larger portions and con- 
tains a small, thick- walled cavity (g) on the right side; this cavity 
is a gill cleft that is cut through neither its outer nor its pharyngeal 

No structures other than this small portion of a gill cleft and a 
few blood-vessels are to be seen in this middle region of the section. 
In the more posterior part of the section, in which the notochord 
(nt) is located, a pair of curved clefts may be seen, opening entirely 
through the wall on the left, but closed on the right (g). One dis- 
tinct pair of aortic arches is shown (ar) , and also the dorsal aortae 
(ao), which are of very unequal size. The spinal cord (sc) and 
muscle plates need no special description. 

Figure i6d is in the region of the heart (ht) and lungs (lu). The 
former is an irregular cavity whose walls, especially on the ventral 
side (mes r ) are becoming very thick and much folded. Although 
thin, the body wall completely surrounds the heart, as would be 
expected, since this was true of the preceding stage. The lung 
rudiments (lu) and the foregut from which they have arisen have 
the same appearance as in the chick ; they consist of three small, 
thick-walled tubes so arranged as to form a nearly equilateral tri- 
angle. They are surrounded by a swollen, rounded mass of meso- 
blast which almost completely fills the surrounding portion of the 
body cavity (be). The pleural sides of these crescentic portions of 
the body (or pleural) cavity — that is, the boundary of the mass of 
mesoblast just mentioned — is lined with a thickened layer of cells, 
shown by the solid black lines in the figure. The lung rudiments 
may be traced through about fifty sections of this series, or about 
one-twelfth of the entire series. At the dorsal angle of the part of 
the body cavity (be) just described, near the dorsal aorta (ao), are 
two dark, granular masses (ge), which, under a higher magnifica- 
tion than is here used, are seen to consist of a small group of blood- 
vessels filled with corpuscles; although several sections in front of 
the anterior limits of the kidneys, these are evidently glomeruli. 
They may be traced, though diminishing in size, far toward the 


tail, in close connection with the Wolffian bodies. At intervals they 
are connected by narrow channels with the dorsal aorta ; no such 
connection was present in the section drawn. The notochord (nt), 
spinal cord (sc), muscle plates (nip), and spinal ganglia (sg) need 
no special mention. The mesoblast is beginning to condense in the 
neighborhood of the notochord, and the ectoderm is slightly thick- 
ened laterally and dorsally. 

Figure 16c is in the region of the liver and the Wolffian bodies ; 
it also shows the tip of the ventricular end of the heart. The liver 
(/•/) is a large irregular mass, of a blotchy appearance under this 
magnification, lying between the heart (vn) and the intestine (i). 
Under greater magnification it is seen to be made up of indefinite 
strings of cells ; and its still wide opening into the intestine may be 
seen in more posterior sections. The intestine (i), which in this 
section might be called the stomach, is a fairly large cavity with 
the usual thick entodermic walls ; it is supported by a comparatively 
narrow mesentery. The body cavity on the side next this mesentery 
has the same thick lining that was noted in the region of the lungs. 
The convolutions of the thick peritoneal lining may easily be mis- 
taken in places for parts of the enteron. The Wolffian bodies may 
be seen as two groups of tubules (wt) in their usual location. The 
heart is cut through the ventricle (zw),as has been said. The section 
being at right angles to the long axes of the villi-like growths of 
the myocardium, the depressions between these mesoblastic cords are 
seen as a number of small irregular areas, each one lined with its 
endocardium. The incompleteness of the body wall below the heart 
is apparently clue to an artificial break and not to a lack of fusion. 
The only point that need be mentioned in connection with the struc- 
tures of the dorsal part of the section is that the distinctness of the 
myocoel (myc) on the right side is somewhat exaggerated. 

Figure 16/ is in the middle region of the embryo, where both 
spanchnopleure and somatopleure are unfused. Owing chiefly to 
the unclosed condition of the midgut (i) and to the increase in 
length of the mesentery (ins), the section is quite deep dorso- 
ventrally. The continuation of the amnion (a) with the somato- 
pleure is of course here evident. 

The most striking feature of the section is the marked projection 
of the Wolffian ridges, though no local enlargements of these ridges 
indicate the rudiments of the limbs. A large mass of Wolffian 
tubules (zvt) is seen projecting into the upper part of the body 
cavity on each side ; close to each of these masses is the posterior 
cardinal vein (pc), and between them is the large aorta (ao). The 
other structures are about as in the preceding section. 


Figure i6g represents a sagittal section of the anterior half of 
the body of an embryo of this or possibly a slightly younger stage 
of development. The three regions of the brain are clearly indi- 
cated, as well as the cavity of the spinal cord (sc). The roof of the 
hindbrain has been made too thick in the figure ; it should be rep- 
resented by a mere line. A little mesoblast is to be seen at places 
between the roof of the brain and the superficial ectoderm. A slight 
invagination of the epithelium (p), between the floor of the brain 
and the anterior end of the notochord, probably represents the begin- 
ning of the hypophysis. Xo indication of the pineal body is yet 
to be seen. Extending from the region of the hypophysis to the 
posterior end of the section is the notochord (nt) ; it is much vacuo- 
lated and gradually increases in thickness toward the posterior, 
though its outline is quite irregular; except at the extreme anterior 
end and at one or two other places, it lies in close contact with the 
floor of the neural tube. Directly under tin- notochord lies, in the pos- 
terior half of the figure, the large dorsal aorta (ao). The pharynx 
(ph). opening between the end of the forebrain and the thick man- 
dibular fold (across which opening the amnion (a) of course ex- 
tends), is a funnel-shaped space which passes out of the plane 
of the section toward the posterior end of the figure. Its thick 
endodermal lining extends to the mandibular fold on the ventral 
side, while on the dorsal side it gradually thins out and becomes 
continuous with the thin ectoderm that extends over the forebrain. 
Just back of tin- mandibular fold is the bulbus (b), and back of that 
is the edge of the ventricle (vn). Posterior and dorsal to the ven- 
tricle the liver ( li ) is seen as an irregular mass of cells, and dorsal 
to the liver one of the Wolffian bodies (wt) is cut through its ex- 
treme edge. 

Stack XIV 
Figures 17-17^ (Plates XVIII, XIX) 

Body flexure has increased until now the forebrain and tail are 
almost in contact (fig. 17). The eye has developed somewhat; the 
ear vesicle, which is not shown in the figure, is small and seems to 
lie nearer the ventral side; the nasal pit is much larger and is 
crescentic in shape. The hyomandibular cleft (g') still persists as a 
small crescentic slit, while the next three clefts are now represented 
merely by superficial grooves separated by distinct ridges, the vis- 
ceral folds. No indication of a fifth cleft is seen. The maxillary 
process (m.v) grows ventralward under the forebrain and is already 
longer than the manibular arch (md). 

The chief advance in development over the preceding stage, be- 


sides the formation of the maxillary process, is in the appearance of 
the appendages (aa and pa) ; they have the characteristic shape of 
the rudimentary vertebrate appendage, though the anterior pair 
seem to point in an unusual direction at this stage and to be slightly 
more developed than the posterior. The curious, anteriorly directed 
heart (lit) is, perhaps, somewhat abnormal. The umbilical stalk 
(u) is comparatively narrow and, like the allantois, was cut off close 
to the body. 

Transverse sections of an embryo of this stage are represented in 
figures ija-g, drawn under a lower magnification than were any of 
the preceding figures. 

Figure lya is in the region of the pharynx, and passes through 
the forebrain (fb ) and posterior part of the hindbrain (lib). In the 
thick walls of both of these structures histological differentiation has 
begun, so that even under low power an inner granular and an outer 
clear zone may be distinguished. Under greater magnification the 
presence of short fibers may be made out among the cells. The 
cerebral hemispheres (ch) are well-marked structures, their asym- 
metry being of course due to the obliquity of the section. Only one 
eye is cut by the plane of the section, and this one shows no con- 
nection with the forebrain. The outer wall of the optic cup (oc) is 
so thin that under this magnification it can scarcely be seen as a 
dark line surrounding the retinal wall. The lens (hi) is now a solid 
mass, of the usual type for vertebrate embryos, its front or outer 
wall being a scarcely discernible line. The hindbrain (lib) has the 
usual form for that region and does not differ particularly from 
what was noted in earlier stages except in the histological differ- 
entiation that has already been mentioned. As with the eye, it is 
only on the right side that the auditory vesicle (o) is shown. It 
shows some differentiation, but not so much as would be seen were 
it cut in another region. In the center of the section the pharynx 
( pli) forms an irregular cavity connected with the exterior on the 
left by a gill cleft (g) and by another slit which is simply the ante- 
rior margin of the stomodaeum. On the right neither of these 
openings are in the plane of the figure, though the gill cleft (hyo- 
mandibular), which lies close to the auditory vesicle, is almost an 
open passage. A few small blood-vessels are scattered through the 
section; one of these (bv), lying between the notochord (nt) and 
the floor of the brain, is noticeable from its being very closely packed 
with corpuscles, so that at first glance, under low magnification, it 
looks more like a nerve than a blood-vessel. 

Figure \yb is also through the pharyngeal region, a short distance 
behind the preceding section. The growth of the cerebral hemi- 


spheres (ch) is better shown than in the preceding figure, as is also 
the general form of the optic cup (oc). On the left the nasal cavity 
(//) is seen as an elongated slit with thick walls; it is cut near, but 
not through, its opening to the exterior. The same gill cleft (g) 
that was seen in the preceding figure is seen here as a narrow, trans- 
verse cleft, open at both ends. Between the notochord (nt) and the 
spinal cord (sc) is the same, though now double, blood-filled vessel 
(bv) that was seen in the preceding section. The other blood-ves- 
sels are larger here than in the more anterior region. There is a 
faint condensation of mesoblast in the neighborhood of the noto- 
chord, and a more marked condensation (nip) farther toward each 
side is the curiously shaped muscle plate. 

Figure 17c is through the heart region, and that organ is cut 
through the opening from the lower or ventricular into the upper or 
auricular chamber. The thickening of the wall of the ventricle, 
which was noticed in the preceding stage, has increased to such 
an extent that there is now a marked difference in the thickness of 
the ventricular and auricular walls. As in the preceding stage, the 
body wall is torn, probably in handling, so that it appears to be 
incomplete around the ventral side of the heart. Dorsal to the heart 
two small circular holes (cut) with thick walls are the cesophagus 
and trachea, cut anterior to the point of bifurcation of the latter into 
the bronchial or lung rudiments. On either side of these struc- 
tures is an elongated blood-vessel (etc), the anterior cardinal vein, 
its elongation being due to the fact that it is cut at the place where 
it turns downward to empty into the heart. Dorsal to the cesoph- 
agus are the aorta; (ao), which are here cut just at the point where 
the two vessels unite to form one ; the next section, posterior to the 
one under discussion, shows an unpaired aorta. The notochord (nt) 
and spinal cord (sc) need no description, except to note that the 
latter shows active histological differentiation, numerous mitotic 
figures being seen under higher magnification, especially in the 
cells that line the spinal canal. On the right of the cord the edge 
of a spinal ganglion (sg) is seen, in connection with which in other 
sections are seen the clearly defined nerve roots. The condensation 
of mesoblast around the notochord is quite evident, and in close 
contact with this medial condensation are two very characteristic, 
S-shaped muscle plates (mp), which extend from the level of the 
dorsal side of the spinal cord to the upper limits of the cardinal 
veins. In some sections the muscle plates even yet show slight 
remains of the myocoel at the dorsal end. 

Figure ijd is in the region of the posterior end of the heart (ht), 
which is cut through the tip of the ventricle, and the anterior end of 


the liver (//'), which has the appearance of a mass of darkly stained 
cords or strands of cells surrounding- a large blood-vessel (in?'). 
This blood-vessel may be ea.Ied the meatus venosus, though it is not 
separated by any line of demarcation from the auricle. A few 
sections anterior to this region the meatus venosus opens dorsally 
into a large vessel on each side (dc) , which at first glance seems a 
part of the body cavity, but which is in reality the ductus Cuvieri, 
formed by the union of the anterior and posterior cardinal veins. 
An irregular, crescentic cleft (be), lying medial and parallel to each 
of the Cuvierian vessels, is the body cavity. In the upper angle of 
this cavity is a granular mass, the glomerulus, that of the left side 
being accompanied by the extreme anterior end of the Wolffian duct. 
In the rounded mass of mesoblast, between the cleft-like regions 
of the body cavity, the lung rudiments {lit) and the oesophagus (oe) 
are seen as three small, circular openings ; that of the oesophagus is 
somewhat smaller than the other two. The notochord (nt), spinal 
cord (sc), and muscle plates (nip) have almost the same appear- 
ance as in the preceding section. A spinal ganglion (sg) is seen on 
each side of the spinal cord ; the one on the left shows a well-defined 
spinal nerve (sit), which may be traced ventrally as far as the end 
of the muscle plate, along whose medial side it courses. The ventral 
nerve root is plainly seen ; the dorsal root, in this section, less 
plainly. The amnion (a) and abdominal wall are, as in the pre- 
ceding figure, torn in the region of the ventricle. 

Figure ije is a short distance posterior to the figure just de- 
scribed. The liver is cut through its middle region and forms a 
large, darkly staining, reticular mass on the left side of the figure. 
The digestive tract is seen at two places to the right of the liver ; 
the smaller and more ventral of these openings (i) may be called 
the intestine, while the larger is evidently the stomach (/'). The 
body wall is here unfused and becomes suddenly thinner as it passes 
upward into the amnion (a). The Wolffian tubules (wt) form a 
very conspicuous mass on either side of the mesentery, in close con- 
nection with the posterior cardinal veins (pc). In the mesoblast 
between the dorsal aorta (ao) and the notochord are two small, 
irregular, darkly stained masses (sy). These are shown in the 
preceding two figures, but were not mentioned in the description. 
They may be traced through a great part of the length of the embryo 
back of the head region ; at intervals corresponding in length to the 
distance between the spinal ganglia they are enlarged, while between 
these enlargements they are very small in cross-section. At certain 
points a small blood-vessel is given off by the dorsal aorta to the 
immediate neighborhood of each' of these small areas. Although 

4— A I. 


they show no connection with the central nervous system, these 
structures appear to be the rudiments of the sympathetic nerves. 

Figure 17/ is in the region just in front of the hind legs. The 
abdominal walls are here unfused, and into the unenclosed body 
cavity projects the intestine (*), supported by a narrow mesentery 
and surrounded by a comparatively thick mass of mesoblast. The 
Wolffian body and duct form a mass of considerable size on each 
side of the mesentery. The Wolffian body is cut near its posterior 
end and consists of smaller tubules than in the more anterior regions. 
The Wolffian ducts (wd), on the other hand, are very large and are 
much more clearly distinguishable from the Wolffian tubules than 
in the preceding sections. The Wolffian ridges (wr) are very 
marked projections on the sides of the body, and in a region further 
caudad become especially developed as the posterior appendages, to 
be described in connection with the following section. Both spinal 
ganglia are shown in this figure (sg), and in connection with the 
left ganglion the spinal nerve (sn), extending ventrally as far as 
the level of the Wolffian duct. The sympathetic nerve rudiments do 
not extend so far caudad as the plane of this section. The dorsal 
end of each muscle plate (»ip) is seen, in this and other sections, to 
be slightly enlarged to form a round knob ; this knob contains a dis- 
tinct cavity ( not shown in the figure) , the myocoel. 

In figure ijg, owing to the curvature of the body, the plane of 
the section passes through the body at three places : through the 
region of the heart and lungs (fig. 17c?), through the region of the 
posterior appendages, and through the tail. In fact, the plane of the 
section represented by each of the preceding figures cut the embryo 
in more than one region, but for the sake of simplicity only one 
region was represented in each figure. In the figure under discus- 
sion only the leg and tail regions have been drawn, though the latter 
region (t), being cut through one of its curves, is seen as an elon- 
gated body with a section of the spinal cord, notochord, etc., at each 
end. Both regions shown in the figure are enclosed in the same fold 
(a) of the amnion. Of the structures in the dorsal side of the larger 
or more anterior part of this figure nothing need be said. The most 
striking feature of the section is the presence of the large posterior 
leg rudiments (pa). As was noted in the preceding figure, they are, 
as usual, merely local enlargements or projections of the mesoblast 
(covered, of course, with ectoblast) of the Wolffian ridge. They 
are, as shown in this section and in the surface view of this stage 
(fig. 17), bluntly pointed projections from the sides of the body. 
The anterior appendage seems to be slightly more developed than 
the posterior, as was noted in describing the surface view of the 


embryo. The digestive tract is cut through its extreme posterior 
end, in the region that may be .termed the cloaca (el), for into it at 
this point the Wolffian ducts open (zvdo). -As the narrow cloacal 
chamber is followed toward the tail, it becomes still smaller in diam- 
eter, and the ventral depression or cleft seen in this figure gradually 
becomes . deeper until its walls are continuous with the ectoderm 
that covers the ventral projection of mesoderm between the hind 
legs ; no actual opening to the exterior is present, however. There 
is a space of about twenty-five or thirty sections (in a series of eight 
hundred) between the posterior ends of the Wolffian bodies and the 
cloacal openings of the Wolffian ducts. The body cavity (be) and 
the posterior cardinal veins { he) are very small in this region, as 
might be expected. 

Stage XV 

Figure 18 (Plate XIX) 

Only the head of this embryo is represented, as the general state 
of development is about the same as in the preceding stage. 

The chief object in making the figure is to show the five gill clefts 
(g 1 '*). The fifth cleft, though small and probably not open to the 
exterior, is quite distinct in this embryo. The writer would feel 
more doubt of its being a true, though rudimentary, gill cleft had 
not Clarke (5) found a fifth pair of clefts in his material. The 
nasal pit has advanced in development somewhat and shows the 
beginning of the groove that connects it with the mouth. In this 
figure the crescentic hyomandibular cleft shows its connection with 
the groove between the mandibular and the hyoid folds. 

Stage XVI 

Figure 19 (Plate XIX) 

This embryo is only slightly more developed than the preceding. 
Body flexure is so great that the forebrain and tail nearly touch. 
Only the anterior three gill clefts are visible. The maxillary pro- 
cess (mx) is longer and more narrow; the mandibular fold has not 
changed appreciably. The nasal pit (n) is now connected by a dis- 
tinct groove with the stomodaeum. The appendages have increased 
in size, the posterior (pa) being the longer. The anterior appendage 
(aa) is distinctly broadened to form the manus, while no sign of 
the pes is to be seen at the extremity of the posterior appendage. 
The heart (ht) is still very prominent. The stalk of the Umbilicus 
(u), which is quite narrow, projects from the ventral wall in the 
region between the heart and the hind legs. The tail is of consider- 
able length and is closely coiled. 


Stage XVII 
Figures 20-20/ (Plates XX. XXI) 

The superficial changes noted in this stage chiefly concern the 
head, which has increased considerably in length (fig. 20). The 
curvature of the body is slightly more marked, and the tail is more 
tightly coiled at the end. There are still signs of three gill clefts. 
The maxillary process (m.r) is long and narrow, while the mandibu- 
lar arch ( md) is still short and broad. The fronto-nasal region has 
greatly increased and has the acquiline profile noted by Clarke. The 
nasal groove has disappeared, and there remains the small opening 
(11) at the side of the fronto-nasal region, near the end of the still 
separate maxillary process. The umbilicus is in about the same 
condition as in the preceding stage, but the heart is less prominent. 
The outline of the manus {ma) is more definite, and the extremity 
of the posterior appendage is distinctly flattened out to form the 
rudimentary pes (pe). The position of the elbow-joint in the an- 
terior appendage is seen at the end of the reference line aa. 

Typical transverse sections of this stage are shown in figures 
20a- j. 

Figure 20a is a section through the middle region of the head, 
cutting the hindbrain on one side and the forebrain on the other. 
The walls of the brain show rather more histological differentiation 
than was seen in the preceding sections, though this cannot be 
shown under the low magnification used. The hindbrain (///'). 
which is cut near its anterior border, exhibits the usual membranous 
dorsal and thick ventral walls. The forebrain is here seen as three 
distinct cavities — a median third ventricle (vt), with a thick ventral 
wall, and a thin dorsal wall extended to form a large pineal body 
(epi), and two lateral ventricles (<"/;). the cavities of the cerebral 
hemispheres, whose walls are quite thick except on the side next the 
third ventricle. The sections of this series being slightly oblique, 
the eye is here cut on the right side only, where it is seen as a large, 
semicircular cavity (<?) with thick, dense walls. The mesoblast. in 
which several blood-vessels (bv) are seen, exhibits three distinct 
areas — a median, lighter /.one. with a more dense ana on either 
side. The significance of this variation in the densit) *<i the meso 
blast is not apparent. 

Figure 20b is only a few sections posterior to the section just 
described. It is drawn chiefly to show the appearance of the fore 
brain, the other structures being about as in the preceding figure, 
except that both eyes (e) are here represented. The section passes 
through the wide opening between the third (/;) and the lateral 


ventricles (ch) and cuts the anterior edge of the pineal body (cpi). 
The pineal body is very large and is directed backward instead of 
forward, as is usually the case among the lower vertebrates (if the 
alligator may be so classed). It is shown in figure 17a of a pre- 
ceding stage and will be again shown in a sagittal section to be 
described later. The same areas of more dense and less dense 
mesoblast noted in the preceding figure are seen here. 

Figure 20c, though still in the head region, shows several features 
that were not seen in the preceding figures. On the left of the hind- 
brain (lib) the auditory vesicle (0), which is now considerably more 
advanced than in earlier figures, is seen as a larger, flask-shaped 
cavity and a smaller, round one. Between the larger cavity and 
the hindbrain is the root of a cranial nerve (en), apparently the 
eighth, since in another section it comes in close contact with the 
wall of the larger part of the auditory vesicle just mentioned. On 
the right side, ventral to the hindbrain, another and much larger 
nerve (en) is seen. Nearly in the center of the figure is seen a 
small, irregular, thick-walled cavity (p), this is the pituitary body, 
and its connection with the roof of the pharynx may easily be made 
out in another section. The mesoblast in this region of the sections 
contains numerous large and small blood-vessels and exhibits certain 
denser areas which probably represent the beginnings of the cranial 
cartilages. No sign of the forebrain is seen (the plane of the section 
passing in front of that region), except that the tip of the wall of one 
of the cerebral hemispheres (ch) is cut. The left nasal chamber (n) 
is shown : it will be noted again in the following section. The eye 
on the right side shows no remarkable features; its lens (In) is 
large and lies well back of the lips of the optic cup, which may now 
be called the iris (ir). A thin layer of mesoblast has pushed in 
between the lens and the superficial ectoderm to form the cornea, 
and the outer wall of the optic cup is now distinctly pigmented. The 
inner wall of the optic cup is beginning to differentiate into the 
retinal elements. The eye on the left side is cut farther from its 
central region and has a very different appearance from the eye just 
described. This unusual appearance is due to the fact that the sec- 
tion passed through the choroid fissure, which is very large and 
seems to be formed by the pushing in of the walls of the cup and not 
by a mere cleft in these walls. This fissure is hardly noticeable in 
the stage preceding the present, and in a stage slightly older it has 
disappeared ; so that it would seem to be a very transient structure. 
It apparently is formed at about the time that the optic stalk, as 
such, disappears. It is in the region of the choroid fissure, if not 
through it, that the optic nerve (on) enters the eye. Through the 


fissure also enters a vascular tuft of mesoblast (pt) which may be 
seen projecting into the optic cup after the disappearance of the 
fissure. This loop of blood-vessels is doubtless the pecten. 

Figure 2od represents a section through the hindbrain (hb), 
pharynx (ph), and tip of the snout. On either side of the hindbrain 
are a convoluted auditory vesicle (0), and several blood-vessels and 
nerves, while ventral to it is seen the anterior end of the notochord 
(nt), around which the mesoblast is somewhat more dense than 
elsewhere. The pharynx (ph) sends out toward the surface a nar- 
row gill cleft (g f ) in the neighborhood of each auditory vesicle. 
These clefts connect with the exterior by very narrow slits, not seen 
in the plane of this section. The opposite end of the pharynx, as 
seen in this figure, opens on the left (pn) into the nasal chamber. 
The nasal cavity on the right is cut in such a plane that it shows 
neither its external nor its pharyngeal opening. The nasal passages 
are here fairly long and nearly straight chambers ; their lining 
epithelium is quite thick in the middle region, but becomes thinner 
where it merges into the epithelium of the pharynx at one end, and 
into the superficial epithelium at the other end. The unusual appear- 
ance of the eye (e), on the right side of the figure, is due to the fact 
that the plane of the section cut tangentially through the extreme 
edge of the eye in the region of the choroid fissure. 

Figure 20c is only a short distance posterior to the preceding. On 
the left side the pharynx ( ph) is connected with the exterior through 
the stomodeaum, and on the right the hyomandibular cleft (g f ) is 
cut almost through its opening to the exterior. The auditory ves- 
icle (o) on the right is cut near its middle region, while that on the 
left is barely touched by the plane of the section. The notochord 
(nt), with its condensed area of mesoblast, is somewhat larger than 
in the preceding section. The nasal canal on the right (n) is cut 
through neither anterior nor posterior opening, while on the left 
side the canal shows the anterior opening (an). 

Figure 2of, which is in the region of the posterior part of the 
pharynx and the anterior part of the heart, shows some rather un- 
usual conditions. 

The spinal cord (sc) and notochord (nt), with the faintly out- 
lined condensations of mesoblast in their region, need no special 
description. The pharynx (ph) is here reduced to an irregular, 
transversely elongated cavity, the lateral angles of which are con- 
nected on each side with the exterior through a tortuous and almost 
closed gill cleft (g), which must be followed through many sections 
before its inner and outer openings may be determined. Dorsal to 
the pharynx numerous blood-vessels (bv), both large and small, mav 


be seen, while ventral to it is noticed a faint condensation of meso- 
blast (la), in the form of an inverted T, the anlage of the laryn- 
geal structures. The ventral portion of the figure is made up of a 
nearly circular, thin-walled cavity, the pericardium (pr). Most of 
the pericardial cavity is occupied in this section by the thick-walled 
ventricle (vn), above which is the bulbus (b) and the tip of the 
auricle (an). The bulbus is nearly circular in outline, though its 
cavity is very irregular. A few sections anterior to this, the opening 
of the bulbus into the ventricle is seen. 

In figure 20g the section represented is only a short distance pos- 
terior to the one represented by figure 20/. The mesoblastic struc- 
tures in the neighborhood of the spinal cord (sc) and notochord 
(nt) will be described in connection with the next figure, where 
they are more clearly defined. The oesophagus (oe) — or posterior 
end of the pharynx, whichever it may be called — is here a crescentic 
slit, with its convex side upward; ventrally it opens by a narrow 
glottis into the trachea (ta). The trachea is surrounded by the 
same condensed area of mesoblast (la) that was mentioned in con- 
nection with the preceding figure, but the condensation is here more 
marked. From the bulbus (b) an aortic arch (ar) extends up- 
ward for a short distance on the right side, while to the left of the 
oesophagus an aortic arch (ar) is cut through the upper part of its 
course. Ventral to the bulbus the ventricle (vn) and two auricles 
(au) are seen surrounded by the pericardial wall. 

Figure 2oh is in the region of the liver (li), which has about the 
same position in relation to the auricles (au) that was occupied by 
the ventricle in the last figure. The auricles are connected with 
each other by a wide passage. The trachea (ta) and the oesophagus 
(oe) are entirely distinct from each other; the former is a small, 
nearly circular hole, while the lumen of the latter is obliterated and 
its walls form a solid, bow-shaped mass of cells. Since there is a 
narrow space between this mass of cells and the surrounding meso- 
blast, it might be thought that the lumen of the oesophagus had been 
closed by the simple shrinkage of its walls ; higher magnification, 
however, fails to show any sign of a collapsed lumen. It is doubtless 
the problematic and temporary closure of the oesophagus that is 
noticed in other forms. On each side of the oesophagus, in close 
relation with the anterior cardinal vein (ac), is noticed a nerve 
(en) cut through a ganglionic enlargement. When traced forward 
these nerves are seen to arise from the region of the medulla, and 
when followed caudad they are found to be distributed chiefly to the 
tissues surrounding the newly formed bronchi; they are doubtless 
the tenth cranial nerves. On the right side of the figure the close 


connection of this nerve with the near-by gill cleft is seen. Above 
the paired aortas (ao) the sympathetic nerves (sy) will be noticed. 
The mesoblast surrounding the spinal cord (sc) and notochord 
(nt) is distinctly condensed (more so than the figure shows) to 
form what may be called the centrum (c) and neural arch (tun of 
the vertebrae. The arch, owing to the slight obliquity of the section, 
shows here only on one side. The spinal cord is not yet completely 
enclosed by the neural arches. The muscle plates ( mp ) are in close 
connection with the rudiments of the vertebrae just mentioned. The 
spinal cord (sc) is here differentiated into three areas — a dense, 
deeply stained area immediately around the neurocoel ; a less dense 
area of cells surrounding the inner area and extending ventralward 
as a rounded projection on each side; and an outer layer, with few 
or no nuclei, surrounding the inner two layers except on the dorsal 

In figure 20i the size and complexity of the figure are due, it will 
be easily understood, to the fact that the plane of the section passed 
through the curve of the body, thus practically cutting the embryo in 
two regions — an anterior, where the lungs (lu) and liver (li) are 
seen, and a posterior, where the Wolffian bodies (wt) are present. 
The spinal cord and the surrounding structures have almost the 
same characteristics at both ends of the figure, except that the 
primitive spinal column is rather more distinct in the posterior end 
of the section. The posterior cardinal veins (pc), Wolffian ducts 
(wd), and Wolffian bodies (wt) are also prominent structures of 
this end of the figure, the last being made up of a great number of 
tubules. The extreme anterior ends of the Wolffian bodies are seen 
in the other half of the section in the upper angles of the body 
cavity, dorsal to the lung rudiments (lu). Filling most of the body 
cavity (be) and making up the greater part of the middle of the 
figure are the liver (li), now a very large organ; the stomach (»'), 
also quite large; the pancreas (pan), a small body lying near the 
stomach; and the lungs (lu), which here consist of several thick- 
walled tubes, surrounded by lobes of mesoblast. The other features 
of the figure need no special mention. 

Figure 20; is through the base of the posterior appendages (pa), 
in which the cartilages are already being outlined by condensations 
of mesoblast. The intestine (i) is cut in two regions — at a more 
anterior point, where it is seen as a small, circular hole surrounded 
by mesoblast and hung by a narrow mesentery, and through the 
cloacal region, the larger and more ventral cavity, into which the 
Wolffian ducts (wd) open a short distance caudad to this section. 
The blood-vessels present a rather curious appearance. A short 


distance anterior to this point the aorta has divided into three, or it 
might be said that it has given off. two, large branches. These two 
branches, one on either side near the posterior cardinal vein, pass 
toward the ventral side of the embryo on each side of the cloaca and 
end at about the region represented by the present figure. The 
small portion of the aorta that remains after the giving off of the 
two 'branches just described continues, as the caudal artery (ca), 
into the tail; it is a small vessel just under the notochord, and gives 
off small, paired branches at regular intervals toward the vertebral 
region. The posterior cardinal veins (pc), posterior to the open- 
ings of the Wolffian ducts into the cloaca, unite to form a large 
caudal vein lying just ventral to the caudal artery. 

Stack XY1II 

Figure 21 (Plate XXII) 

This embryo, as may be seen, for example, by the form of the 
appendages, is slightly further developed than the one represented 
in figure 20. The figure is from a photograph of a living embryo as 
it lay in the egg, a portion of the shell and shell membranes having 
been removed. The embryo, which lies on its left side, is rather 
faintly outlined because of the overlying allantois. The allantois 
has been increasing rapidly in size, and is here so large that it ex- 
tends beneath the cut edges of the shell at all points except in the 
region in front of the head of the embryo, where its border may be 
seen. Its blood-vessels, especially the one that crosses the head just 
back of the eye, are clearly shown in the figure, and in the living 
specimen, when filled with the bright red blood, they form a most 
beautiful demonstration. As in the chick, the allantois lies close 
beneath the shell membranes and is easily torn in removing them. 

Stage XIX 

Figure 22 (Plate XXII) 

Figure 22 is a photograph of a somewhat older embryo, removed 
from the egg and freed of the fcetal membranes. The appendages 
show the position of both elbow and knee joints, and in the paddle- 
shaped manus and pes the digits may be faintly seen. The tail is 
very long and is spirally coiled, the outer spiral being in contact 
with the frontal region of the head. The jaws are completely 
formed, the upper projecting far beyond the lower. The elliptical 
outline of the eyes is noticeable, but the lids are still too little devel- 
oped to be seen in this figure. The surface of the embryo is still 
smooth and white. 


Stage XX 
Figures 23-236 (Plate XXII) 

In this surface view (fig. 23) several changes are seen, though no 
very great advance in development has taken place. The outlines 
of the digits (five in the manus and four in the pes) are now well 
defined; they even project slightly beyond the general outline of the 
paddle-shaped part. The tail has begun to straighten out, and it 
now extends across the front of the face. The lower jaw has in- 
creased in length, but is still shorter than the upper. The eyelids, 
especially the upper, are beginning to be discernible in surface view. 
Though still without pigment, the surface of the body is beginning 
to show by faint transverse lines the development of scales; these 
lines are most evident in this figure in the middle region of the tail, 
just before it crosses the nose. 

A sagittal section of the entire embryo (except the tail) of this 
age is shown in figure 23a. In the head region the section is nearly 
median, while the posterior part of the body is cut slightly to one 
side of the middle line. At the tip of the now well-developed snout 
is seen one of the nostrils (an), cut through the edge; its connection 
with the complicated nasal chamber (11) is not here seen, nor is the 
connection of the nasal chamber with the posterior nares (pn). The 
pharynx (ph) is anteriorly connected with the exterior through the 
mouth (m) and the nares, while posteriorly it opens into the oesoph- 
agus (oe) ; the trachea (ta), though distinct from the oesophagus, 
does not yet open into the pharynx. In the lower jaw two masses 
of cartilage are seen, one near the symphysis (mk) and one near the 
wall of the trachea, doubtless the rudiment of the hyoid. The deep 
groove back of the Meckel's cartilage (mk) marks the tip of the 
developing tongue, which here forms the thick mass on the floor of 
the mouth cavity. Dorsal to the pharynx a mass of cartilage (se) 
is developing in the sphen-ethmoid region. This being a median 
section, the ventricles of the fore- (fb), mid- (mb), and hindbrain 
(hb) are seen as large cavities, while the cerebral hemispheres (ch) 
appear nearly solid, only a small portion of a lateral ventricle show- 
ing. The pineal gland (epi) is cut a little to one side of the middle 
and so does not show its connection with the brain. At the base of 
the brain the infundibulum (in) is seen as an elongated cavity whose 
ventral wall is in close contact with a group of small, darkly staining 
alveoli (p), the pituitary body. Extending posteriorly from the 
pituitary body is a gradually thickening mass of cartilage (bp), 
which surrounds the anterior end of the notochord (nt) and may be 
called the basilar plate. In its anterior region, where the section is 


nearly median, the spinal column shows its canal, with the enclosed 
spinal cord, while toward the posterior end of the figure the vertebrae 
are cut to one side of the middle line, and hence show the neural 
arches (na) with the alternating spinal ganglia (sg). Near the 
posterior end "t>f the figure the pelvic girdle (pi) is seen. The 
largest organ of the embryo, as seen in this section, is the heart, of 
which the ventricle (vn) seems to be closely surrounded, both in 
front and behind, by the auricles (au). The liver (li) is the large, 
reticular mass back of the heart. Dorsal and anterior to the liver 
is the lung (lu), now of considerable size and development. The 
enteron is cut in several places (oc, i) and its walls are beginning 
to show some differentiation, though this cannot be seen under the 
magnification here used. One of the Wolffian bodies is seen as a 
huge mass of tubules (zvt) extending from the pelvic region, where 
the mass is greatest, to the region of the lungs. The Wolffian 
tubules stain darkly and the whole structure forms a very striking 
feature of the section. Dorsal to the posterior end of the Wolffian 
body is a small, oval mass of very fine tubules (k), which do not 
stain so darkly as do the Wolffian tubules ; this mass is apparently 
the beginning of the permanent kidney, the metanephros. Its 
tubules, though their origin has not been determined, seem to be 
entirely distinct from the tubules of the Wolffian body. 

A single vertical section through the anterior part of the head of 
an embryo of this age has been represented in figure 236. On the 
right side the plane of the section cut through the lens of the eye 
(In) ; on the left side the section was anterior to the lens. The 
upper (nl) and lower (//) eyelids are more evident here than in the 
surface view. Owing to the hardness of the lens, its supporting 
structures were torn away in sectioning. The vitreous humor is not 
represented in the figure. The superior (ar) and inferior (Ir) recti 
muscles are well shown on the right side; they are attached to the 
median part of a Y-shaped mass of cartilage (se), which may be 
termed the sphenethmoidal cartilage. Between the branches of this 
Y-shaped cartilage the anterior ends of the cerebral hemispheres 
(ch) — better called, perhaps, the olfactory lobes — are seen. Be- 
tween the lower end of the sphenethmoidal cartilage and a dorsally 
evaginated part of the pharynx are two small openings (pn) ; when 
traced forward these tubes are found to open into the convoluted 
nasal chamber, while a short distance posterior to the plane of this 
figure they unite with each other and open almost immediately into 
the pharynx. The rather complicated structures of the nasal pas- 
sages of the alligator have been described by the writer in another 
paper (12). In the lower jaw the cartilage (mk) is seen on either 


side and several bands of muscle are developing in the mesoblast. 
Two deep grooves give form to what may be called the rudimentary 
tongue (tn). In both jaws one or two tooth rudiments (to) may 
be distinguished as small invaginations of ectoderm. 

Stage XXI 

Figure 24 (Plate XXII) 

In this stage the curvature of the body and tail is less marked 
than was seen in the last surface view. The body has increased 
greatlv in size, so that the size of the head is relatively not so great. 
The size of the eye in relation to that of the head is much diminished 
also. The five anterior and four posterior digits are well formed, 
and their claws are of considerable size, though of course not present 
on all the digits. The outlines of scales may be traced from the 
tip of the tail to the skull ; they are especially prominent along the 
dorsal profile. The skin is just beginning to show traces of pig- 
ment, which is, however, not shown in the photograph. The umbil- 
ical stalk is seen projecting with a loop of the intestine from the ab- 
dominal wall ; this is shown more clearly in the next stage. The 
embryo now begins to exhibit some of the external characteristics 
of the adult alligator. 

Stage XXII 

Figure 25 (Plate XXI II) 

This embryo needs no particular description. It has reached in 
its external appearance practically the adult condition, although 
there is still considerable yolk (not shown in the figure) to be ab- 
sorbed, and the embryo would not have hatched for many days. 
Pigmentation, begun in the last stage, is now complete. The umbil- 
ical stalk is clearly seen projecting from a large opening in the body 
wall. The long loop of the intestine that extends down into the 
yolk sac is here evident, and it is hard to understand how it can all 
be drawn up into the body cavity when the umbilical stalk is with- 
drawn. No sharp shell-tooth at the tip of the snout, such as is 
described by Voeltzkow (18) in the crocodile, is here seen. 

Stage XXIII 

Figi-re 26 (Plate XXIII) 

This figure shows the relative sizes of the just-hatched alligator 
and the egg from which it came. It also shows the position of the 
young alligator in the egg, half of the shell having been removed for 


that purpose. The blotchy appearance of the unopened egg is due 
chiefly to stains produced by the decayed vegetation of the nest. At 
hatching the young alligator is about 20 cm. long, nearly three times 
the length of the egg ; but the tail is so compressed that, though it 
makes up about half of the length of the animal, it takes up very 
little room in the egg. 


Owing to the fact that the embryo may undergo considerable 
development before the egg is laid, and also to the unusual difficulty 
of removing the very young embryos, the earlier stages of develop- 
ment are very difficult to obtain. 

The mesoderm seems to be derived chiefly by proliferation from 
the entoderm, in which way all of that anterior to the blastopore 
arises. Posterior to the blastopore the mesoderm is proliferated 
from the lower side of the ectoderm in the usual way. No distinc- 
tion can be made between the mesoderm derived from the ectoderm 
and that derived from the entoderm. 

The ectoderm shows during the earlier stages a very great in- 
crease in thickness along the median longitudinal axis of the embryo. 

The notochord is apparently of entodermal origin, though in the 
posterior regions, where the germ layers are continuous with each 
other, it is difficult to decide with certainty. 

The medullary folds have a curious origin, difficult to explain 
without the use of figures. They are continuous posteriorly with 
the primitive streak, so that it is impossible to tell where the medul- 
lary groove ends and the primitive groove begins, unless the dorsal 
opening of the blastopore be taken as the dividing point. 

The amnion develops rapidly, and entirely from the anterior end. 

The blastopore or neurenteric canal is a very distinct feature of 
all the earlier stages up to about the time of closure of the medullary 

Preceding the ordinary cranial flexure there is a sort of temporary 
bending of the head region, clue apparently to the formation of the 

During the earlier stages of development the anterior end of the 
embryo is pushed under the surface of the blastoderm, and is hence 
not seen from above. 

Body torsion is not so definite in direction as in the chick, some 
embryos lying on the right side, others on the left. 

Of the gill clefts, three clearly open to the exterior and probably 
a fourth also. A probable fifth cleft was seen in sections and in one 
surface view. 


The first trace of the urinary system is seen as a dorsally project- 
ing, solid ridge of mesoblast in the middle region of the embryo, 
which ridge soon becomes hollowed out to form the Wolffian duct. 

The origin of the pituitary and pineal bodies is clearly seen ; the 
latter projects backward. 

No connection can be seen between the first rudiments of the sym- 
pathetic nerves and the central nervous system. 

The lumen of the oesophagus is for a time obliterated as in other 

The choroid fissure is a very transitory but well-marked feature 
of the eye. 


1. Anderson, A. : An Account of the Eggs and Young of the Gavial (G. gan- 

geticus). Proc. Zool. Soc. 1875. 

2. Balfour, F. M. : The Early Development of the Lacertilia. Quar. Jour. 

Mic. Soc, 1879, vol. xix, pp. 421-430. 

3. Balfour, F. M. : Comparative Embryology, vol. 11. 

4. Bronn, H. G. : Klassen des Thier-Reichs (vols, on reptiles). 

5. Clarke, S. F. : The Habits and Embryology of the American Alligator. 

Jour. Morph., vol. v, pp. 182-214. 

6. Dendy, Arthur: Outlines of the Development of Tuatara (Sphenodon 

punctatus). Quar. Jour. Mic. Soc, vol. xxxxn, 1899, PP- 1-87. 

7. EislER, P.: Zur Kentniss der Histologic des Alligatormagens. Archiv. f. 

Mik. Anat., vol. xxxiv, pp. 1-10, 1889. 

8. Hoffmann, C. K. : Beitrage zur Entwicklungsgeschichte der Reptilien. 

Zeit. f. wiss. Zool., vol. xxxx, 1884, pp. 214-246. 

9. Hoffmann, C. K. : Weitere Untersuchungen zur Entwicklungsgeschichte 

der Reptilien. Morph. Jahrb., vol. xi, 1886, pp. 176-219. 

10. Parker, W. K. : On the Structure and Development of the Skull in the 

Crocodile. London, 1883. Zool. Soc London, 1883, vol. xi, pp. 263-310. 

11. Rathke, H. : Untersuchungen iiber die Entwicklung und den Korperbau 

der Krokodile. Braunschweig, 1866. 

12. Reese, A. M. : The Nasal Passages of the Florida Alligator. Proc. Phila. 

Acad. Nat. Sc, 1901. 

13. REESE, A. M. : The Breedings Habits of the Florida Alligator. Smithson- 

ian Misc. Coll. (Quarterly Issue), vol. xlviii, pp. 381-387, 1907. 

14. Roese, C. : Uber die Zahnleiste und die Eischweile der Sauropsiden. Anat. 

Anz., 1892, vol. vii, pp. 248-264. 

15. Strahl, H. : Beitrage zur Entwicklung von Lacerta agilis. Archiv. f. 

Anat. u. Physiol., 1882, pp. 242-278. 

16. Strahl, H. : Beitrage zur Entwicklung der Reptilien. Ibid., 1883, pp. 1-43. 

17. Strahl, H. : Uber friihe Entwicklungsstadien von Lacerta agilis. Zool. 

Anz., vol. vi, 1883, pp. 347-350. 

18. Voeltzkow, Alfred: Biologie und Entwicklung der ausseren Korperform 

von Crocodilus madagascariensis Grand. Abhandl. Senckenberg. Naturf. 
Gesell., vol. xxvi, pt. 1, pp. 1-149, 1889. 

19. WiEdershEim, R. : Comparative Anatomy. 



a, head-fold of amnion. 
aa, anterior appendage. 
ac, anterior cardinal vein. 
al, allantois. 

an, anterior nares. 
ao, aorta. 
aop, area opaca. 
ap, area pellucida. 
ar, aortic arch. 
au, auricle. # 

b, bulbus arteriosus. 
be, body cavity. 
blp. blastopore. 

bp, basilar plate. 
bv, blood vessel. 

c, centrum of vertebra. 
ca, caudal artery. 

ch, cerebral hemisphere. 

el, cloaca. 

en, cranial nerve. 

cp, posterior choroid plexus. 

cv, cardinal veins. 

■dc, ductus Cuvieri. 

e, eye. 

ec, ectoderm. 

cc' , thickening of ectoderm. 

en, entoderm. 

en' , endocardium. 

cnt, enteron. 

cp, epidermal layer of ectoderm. 

epi, pineal body. 

es, embryonic shield. 

f, fronto-nasal process. 

fb, forebrain. 
fg, foregut. 
g x S, gill clefts. 
gf 1 - 6 , gill folds. 
gl, glomerulus. 
h, head-fold. 
hb, hindbrain. 
ht, heart. 

i. intestine. 

i', stomach. 

in, infundibulum. 

tr, iris. 

it, iter. 

fc, kidney (metanephros). 

/, remains of groove between second- 
ary folds. 
la, larynx (cartilages of). 

li, liver. 

//, lower lid of eye. 

In, lens. 

Ir, inferior rectus muscle of eye. 

lu, lungs. 

Iv, lens vesicle. 

m, mouth. 

ma, manus. 

nib, midbrain. 

me, medullary canal. 

me', tip end of medullary canal. 

md, mandibular fold. 

mes, mesoderm. 

incs', myocardium. 

mf, medullary fold. 

nig, medullary groove. 

mk, Meckel's cartilage. 

nip, muscle plate. 

ms, mesentery. 

mv, meatus venosus. 

nix, maxillary fold. 

myc, myocoel. 

n, nasal invagination or cavity. 

11a, neural arch of vertebra. 

nc, neurenteric canal. 

nl, nervous layer of ectoderm. 

nt, notochord. 

0, ear vesicle. 

oc, optic cup. 

oe, oesophagus. 

on, optic nerve. 

os, optic stalk. 

ov, optic vesicle. 

/>, pituitary body. 

pa, posterior appendage. 

pan, pancreas. 

pc, posterior cardinal vein. 

pe, pes. 

pg, primitive groove. 

p/i, pharynx. 

pi. pelvis. 

pn, posterior nares. 

pr, pericardial cavity. 

ps, primitive streak. 

pt, pecten. 

rt, retina. 

s, somites. 

sc, spinal cord. 

se, sphenethmoid cartilage. 

sf, secondary fold. 


sg, spinal ganglion. u, umbilical stalk. 

sm, splanchnic mesoblast. it!, upper lid of eye. 

sn, spinal nerve. '"'. superior rectus muscle of eye. 

so, somatic mesoblast. v'"-"', first, second, and third cere- 

st, stomodaeum. bral vesicles. 

sy, sympathetic nervous system. va, vascular area. 

/, tail. vm, vitelline membrane. 

ta, trachea. vn, ventricle of heart. 

tg, thyroid gland. w, vitelline blood-vessels. 

//;. thickening and posterior limit of sf. i^d. Wolffian duct. 

tn, tongue. wdo, opening of Wolffian duct. 

to, tooth anlage. wr, Wolffian ridge. 

tr, truncus arteriosus. wt, Wolffian tubules. 

tv, third ventricle of brain. y. yolk. 

tv' , third ventricle of brain. 


All of the figures, with the exception of the photographs and those copied 
by permission from S. F. Clarke, were drawn under a camera lucida. 

The magnification of each figure, except those from Clarke, is indicated 

The photographs were made by the author, and were enlarged for repro- 
duction by the photographic department of the Smithsonian Institution. The 
other surface views were made, under the author's direction, by Miss C. M. 

With the exception of Stage III, all of the figures of any one stage are 
given the same number, followed where necessary by a distinguishing letter, 
so that it is possible to tell at a glance which section and surface views belong 
together. The transverse sections are all cut in series from anterior to 

Figure i. Surface view of egg. X 2/3. 

id. Egg with part of the shell removed to show the chalky band in 
the shell membrane. X 2/3. 
Figures 2 and 2a. Dorsal and ventral views respectively of the blastoderm be- 
fore the formation of the notochord, medullary folds, etc. After 
2b-2f. Transverse sections of an embryo of the age represented in 
figures 2 and 2a. X 43. 

3 and 3a: Ventral and dorsal views respectively of an embryo a few 

days older than that represented in figures 2 and 2a. After 
30-3111. Transverse sections of an embryo of the age shown in figures 
3 and 3a. X 43. 
Figures 311 and 30. Two sagittal section of an embryo of the same stage as 
figures 3 and 3a. X 43. 

4 and 4c7. Dorsal and ventral views respectively of a slightly older 

embryo than the one shown in figures 3 and 3a. Figure 4a 
shows only the head region. After Clarke. 

5 and 5a. Dorsal and ventral views respectively of an embryo of 


almost the same age as the preceding, to show the further de- 
velopment of the medullary folds. After Clarke. 

Figure; 6. Dorsal view of an embrye only a day or two older than the pre- 
ceding. After Clarke. 

Figi RES 6a 6i A series nf transverse sections of this stage. X 43. 

FIGURES ~ja-~h. A series of transverse sections of an embryo slightly older 
than the one shown in figures 4-6. X 43. (No surface view of 
this stage is figured. ) 
8 and 8a. Dorsal and ventral views respectively of an embryo with 
fiye pairs of mesoblastic somites. X 20. (Drawn by trans- 
mitted light.) 
8b and 8c". Two sagittal sections of an embryo of this stage. X 43. 

FIGURES 8d-8/. A series of transverse sections of the embryo represented in 
figures 8 and 80. X 43. 
Qa-gm. A series of transverse sections of an embryo somewhat more 
advanced in development than the one represented in the last 
series. X 43. 

Figures 10 and 101;. Dorsal and ventral views respectively of an embryo with 
eight pairs of mesoblastic somites. X 20. (Drawn chiefly by 
transmitted light. ) 

Figure II. Dorsal view of an embryo with fourteen pairs of mesoblastic 
somites. The area pellucida and the developing vascular area 
are shown, the latter having a mottled appearance. The pushing 
of the head under the blastoderm is also shown. X 20. (Drawn 
chiefly by transmitted light.) 

Figures lia-llk. A series of transverse sections of an embryo of this stage. 

Figure 12. Dorsal view of an embryo with about seventeen pairs of meso- 
blastic somites. Part of the area pellucida is represented. 
( Both transmitted and reflected light were used in making the 
drawing.) X 13. 

Figures \2a-12g. A series of transverse sections of an embryo of this stage. 


FlGURE i.v Surface view of an embryo with about twenty pairs of meso- 
blastic somites. X (about) 15. (Drawn with both reflected 
and transmitted light.) 

Figures 13-13/. A series of transverse sections of an embryo slightly more 
developed than the one shown in figure 13. X 20. 

Figure I3g. A sagittal section of an embryo <>f about the age of the one 
represented in figure 13. X 20. 

14. Head of an embryo with one pair of gill clefts; ventro-lateral 

view. X 13. 

15. Profile view of the head of an embryo with three pairs of gill 

clefts. X 13. 
Figures \$a-\$c. A series of transverse sections of an embryo of about the 

age of the one represented in figure 15. X 20. 
Figure 15/. A horizontal section through the anterior region of an embryo of 

the age of that shown in figure 15. X 20. 

16. Surface view in profile of an embryo with four pairs of gill clefts. 

X (about) 12. 

5— AL 


Figures i6a-i6f. A series of transverse sections of an embryo of the approxi- 
mate age of the one represented in figure 16. X 20. 

Figure l6g. A sagittal section of an embryo of the age (possibly slightly 
younger) of the one represented in figure 16. X 20. 
17. Surface view in profile of an embryo at the time of origin of the 
limbs. X (about) 5. 

Figures I7a-i7g. A series of transverse sections of an embryo of the age of 
the one represented in figure 17. X 7. 

Figure 18. Surface view in profile of the head of an embryo slightly larger 
than, though of about the same state of development as, the one 
represented in figure 17. Reproduced here chiefly to show the 
gill clefts. X (about) 3. 
[9. Surface view of an embryo somewhat more developed than the 
one just described. X (about) 3. 

Figure 20. Surface view of an embryo older than the one represented in 
figure n>; with well developed manus and pes. X (about) 5. 

Figures 20a 20/. A series of transverse sections of an embryo of the age of 

the one represented in figure 20. X 7. 
Figure 21. A photograph of a living embryo in the egg, showing the allan- 
tois, yolk mass, etc. The embryo is somewhat more developed 
than the one shown in figure 20. X _' ,}. 

22. A photograph of a still larger embryo, removed from the shell 

and freed from the fetal membranes. X (about) 1. 

23. A photograph of a still more advanced embryo, in which the 

digits are quite evident and the scales are beginning to show. 

X (about) 1. 
230. A sagittal section of an embryo of the age of the one represented 

in figure 23; the tail has not been shown in this figure. X 

(about) 3. 
23ft. A vertical section through the head of an embryo of about the 

size (perhaps slightly smaller) of the one shown in figure 23. 

X (about) 3. 

24. A photograph of an older embryo in which the pigmentation of 

the scales is evident, though not shown in the figure. X 
( about ) 1. 

25. A photograph of an embryo in which the pigmentation and the 

development of the body form are practically complete. The 

allantois, unabsorbed yolk, etc., have been removed. X 
(aboul ) ; |. 

26. A photograph of a just hatched alligator, of an alligator egg, 

and of a young alligator in the egg just before hatching. X 
(about) 3/7. 


PL.--- a P 


■ blp 

2 a 

/, /a.— The Egg. 2, 2a.— Stage I 



2 b " 


' " ' - ~— ^* e 2 n d ^C33& r 


..•?.".. blp *~, ec 


mes en 

,.„..../• ••'•'•-^VVAVVftViv 


2b-2f— Stage I. 3, j*.— Stage II 



•'•*•:♦ .*.'•* \','/.' ^""" "'• *-'*** e n 


•.'•'» '.•.'•""" en ^ e 

"♦.* . •'«'.%•• • i A**** '**'•;• ""•£'.«.''"' •,. 

• • V.* 

jfl, ?.§•.— Stage II 



mes en 

...■• > ' . .-.■ '::"'^->:^ ; ^K';v:iV-:,v;':'*::V:-v... 

nt **- 

'' , .V.''" , ' , .:'.t<A7'/,V.;J>..^ e 

.,./.*• blp'*,.:-;V,.... 


., , ec 


mes en 

en "** S \|| 


ec .:v*|'.^;>. 'P o-i vV''> V-V:'.-'. :'•.'•'• 

3I1-30.— Stage II 



f ' 

4 a 



Stage III 


% //"cV;I;cj»«* # V;V '; 4'.*y , * / "'*''' , ec 


, th 

, ,.i mq ,, ,,, er 

.•v!*V' •'.'• •••''*•' /i ••.•/■••■'•VSI'i '*'> 

...•■••■ th en'*...','- _;.,'• 6f 


;•."'■' " nt "'•.'••.'•'•. ''••• ..'•••' :; '" ••'' :.'.'.'■ '•:':$&' ''"•'!'.{%•' '.■'-• '•''*}:% 

6 9 en'---.., ."..:■■ •;>'-' blp "•.:•>. 

i> oh eh 

••':'•: \/. : - v ' •""* -' ; --v^r?:- v."' :; v. 

/•;..•'*' 6i rr,es"..;U. 

"" ''.'. ••'••.:.'.-'V.'-:;.v.'.;... e i c 

fp-'- : r :: - 

6-6 i.— Stage III. 7a.— Stage IV 

7b ""v.. •-'•' 





£***" en 
nt m es 

T r'.+-' • me 

^UTT-*- en 

s,»«. r *.«»»»• mes en 




a 6a 


,-/'-,-/'/. -Stage IV. 8, 8a.— Stage V 

?"•••» *%':•.' 

-* - •«wg 

...» ; \ » « 

,. * . . . . en 

i 5 


. • v^L/™^. 


9 a 

en ^i 

.7 /^?£K (JZ 7 ^ 

7 :'' IW?? '?Ul i . '-''A * • 



• < ent - . . 

• . . . 

ent f ffi' 

' ' r: :''i:'-i-'''^ : > ^ 

■•'i - - '',-""* C* be 


Si, S/.— Stage V. ga-yg.— Stage VI 



it Be V 




* h ::';, »*v*?\fi ,r ' es 

'• , '''.'»*'•''.'%'••'■'•'•.'•".''•'.*/-"*. 



•Vv.N';-.-." en 



<*. .■'•.• ' • . 

.'.'■ •*/'.•' V'.'-V - *.* * -' ■ - 
,;.•;_-;, St*;// _. , 



;.— Stage VI. Stage vn 




.••■:••- &n - 



' ■'•i: jv. , 


'■■'■:•■'• ><r mc ' 


■ - c :_ » ^^ nT 



1 1 o 

ioa.— Stage VII. zza-iid.— Stage VIII 





': •'"• *$P ... L k en1 



;;> ^.^#f .;-:-.v-;..' 7. 

en •-^f?v- • -. . . 
II h nt(?i 

Stage VIII 


ec / T 10 , 

.■•■>*&■ ■■• 

■■'"•. y/--:^b v ~ ; --'{- ^^/J"/."-'-.v.'. ; .*:'.v..:-.' ; ' 

■-wNI> — ... •■ 

mes en 

Hi n'c 



.':•'•"• -:-.-. : . •:'■'.■ "$n- .>&» « » 

en nc mes 

1 1 J 

7 **> i . ec ec 

'•• • ■ • "V.' • <'M'- • • • - 

mes eh 3§ggirp S 


///-///(•.—Stage VIII. i2-i2b— Stage IX 



I2e en 

n't ao 
m * c I2F iTT 





I2c-i2g- Stage IX. ija-i^c— Stage X 




Stage X 




13 g 






Stage X. //.-Stage XI. rj-/jc— Stage XII 



mp / \ 




is/.— Stage XII. i6-/6<f. -Stage XIII 



.- ;. md 


ni , 

^. dlUC 

i s 9 


^1 s 


"■ ;. i 



i6c-i6g— Stage XIII. /;.— Stage XIV 





em/ - 





17 b 


i n 

h l j 

lya-ijg.— Stage XIV. /,?.— Stage XV. 79.— Stage XVI 



20 a 


20 b 20 c 

Stage XVII 

20 d 



20 f 


20 e 

be ao 

20 h 

Stage XVII 



23 D 

-'/.—Stage XVIII. 22— Stage XIX. 2j-2jb.— Stage XX. 24.— Stage XXI 





-Stage XXII Alligator embryo. 2(5.— Stage XXIII Alligator just hatched 
and relative .size of egg. 








In Charge of Collections of Muscoidean Flies, U. S. National Museum 




No. 1803 




f \\evrhinoton' J 





When we review the history of the classification of any highly 
specialized group of insects, provided it has attained a considerable 
degree of popularity among systematists, we find it to exhibit a 
well-marked series of oscillations between the two extreme- com- 
monly known as bunching" and splitting. This is especially true of 
the dipterous superfamily Muscoidea. 1 

The systematists of the eighteenth and nineteenth centuries, ac- 
cording to the work they did on this superfamily. mark alternate 
periods of action and reaction which fall conveniently into five his- 
torical epochs. 

Linne, Fabricins, and Latreille must be considered the pioneers. 
The system they established was followed by their immediate' con- 
temporaries. Very few others concern us here, but Geoffroy erected 
the genus Stomoxys. and Scopoli, Rossi, and Panzer did some work 
on the superfamily. As a natural result of approaching a quite new 
subject, these early workers did not always grasp the real value of 
characters. Largely because of the comparative dearth of material 
in those initial days of systematic work, they did not clearly discern 
anatomical values, and hence did not recognize many characters 
whose worth has since been well established. 

Meigen introduced a new epoch in 1804, and considerably in- 
creased the number of genera by splitting up the original ones estab- 
lished by his predecessors. Collections had become richer in mate- 

1 It is to be noted that the superfamily Muscoidea, as herein restricted, in- 
cludes but a portion of the forms to which the name was applied by it- author, 
Mr. D. W. Coquillett. As now restricted, it includes practically the old 
calyptrate Muscidae minus the Anthomyiidse, or the same group as that treated 
by Brauer and von Bergenstamm — Muscaria Schizometopa, exclusive Antho- 
myiicke. The Muscoidea is here divided into live families a- follows: (1) 
CEstridse, (2) Macronychiida? (being a part of the old Dexiidae), (3) Tachi 
(including the old Gymnosomatidae, Phaniidae, Ocypteridae, Sarcophagidae. and 
most of Dexiidae as subfamilies). (4) Muscidae, and (5) Phasiidae (including 
Rutilia and its allies). 



rial by this time, and Meigen's attention was naturally drawn to the 
discovery of further characters that could be used in classification. 
He was indorsed and followed by his contemporaries, Olivier, Fallen, 
Say, Wiedemann, who adopted his genera without proposing new 
ones, except that the last-named author erected the single genus 
Glossina for the tse-tse flies. Dumeril erected the genus Bchiiwmyia, 
and Le Peletier de Saint-Fargeau the single genus Prosena. Mei- 
gen's best work was in genera. His descriptions of species were in 
many cases faulty. On the whole, however, he is clearly to be looked 
upon as an epoch-maker. 

The first really intuitive student of the superfamily was Robineau- 
Desvoidy who. in 1830. introduced the third epoch and very greatly 
increased the number of genera, besides defining more or less natural 
taxonomic divisions for their reception. It must be understood that 
very considerable accumulations of material from the Americas, both 
North and South, had reached Europe during the early part of the 
nineteenth century, besides much material from the African, Ori- 
ental, and Australasian regions. To most of this Robineau-Desvoidy 
had access. Notable among the accumulations were the rich collec- 
tion of the Count Dejean, which had been added to constantly by 
Latreille, and the quite extensive material secured from all parts by 
the Museum of the Jardin du Roi in Paris. Palisot de Beauvois, 
Saint-Hilaire, Bosc, and many others collected in the Americas, and 
various representatives of the Jardin du Roi in other parts of the 
world. Besides these, many European entomologists sprang up who 
began to do much more thorough collecting at home. Thus a com- 
paratively great wealth of material in the Muscoidea was brought 
together from all parts of the world, both at home and abroad, which 
stimulated Robineau-Desvoidy to a detailed study of characters in 
this superfamily. His "Essai sur les Myodaires" remains to this 
day a monument to his very considerable grasp of Muscoidean rela- 
tionships. His posthumous work (1863) can not be considered as 
affecting in any way the status of the "Essai." 

Macquart, almost contemporaneous with Robineau-Desvoidy, but 
possessed of less discernment, bunched the latter's genera to a very 
considerable extent. However, it must be pointed out in defense of 
Macquart that he was eminently a general dipterist, while Robineau- 
Desvoidy was preeminently a specialist in the Myodaria. 

Zetterstedt erected only two genera in the superfamily, and prac- 
tically employed Meigen's genera for all of his work. Perty, Bouche, 
Guerin, and Bremi each erected a single genus in the superfamily. 

Robineau-Desvoidy's system, founded largely on habits, was in a 
degree faulty and insecure. Attention should be called to the fact, 


apparently long- since lost sight of, that Robineau-Desvoidy origi- 
nated the idea of including the CEstridae with his Calypterata (al- 
though renounced in his posthumous work), and the Conopidse with 
the Myodaria (Conopidae not included at all in posthumous work). 
The founding of the now obsolete division Calypterata is also to be 
accredited to him, though it is to be noted that he did not include the 
Anthomyiidas therewith. The latter family was included in that 
division by subsequent authors. In this connection, see Osten- 
Sacken for statement that the term "Acalypterata" was interpolated 
in Robineau-Desvoidy 's posthumous work by the editors (Berl. Ent. 
Zeit, 1896, pp. 329, 335-6). 

Rondani marked a fourth epoch beginning about 1850. He re- 
vised in large part the work of Robineau-Desvoidy,' still further in- 
creased the number of genera, was altogether a very close student of 
relationships, and possessed a remarkably clear insight into the affin- 
ities of the Muscoidea, in which he was essentially a specialist. His 
system was followed to some extent by his more immediate contem- 
poraries, but Schiner, with a fine grasp of dipterous characters in 
general and little conception of the needs of the Muscoidea, was espe- 
cially active in bunching his genera. 

Schiner was a splendid general dipterist, but the method of treat- 
ment adapted to other groups of Diptera fails when the attempt is 
made to apply it to the Muscoidea. That is where Schiner, Mac- 
quart, and all the other conservatists fell. And it is to be noted that 
these conservatists were always general dipterists. They tried to 
apply the same system throughout the Diptera, but the Muscoidea 
need a distinct method of treatment, as will appear further on in this 
paper under that heading. Even such conscientious students as 
van der Wulp, Loew, Osten-Sacken, Williston, and others, who fol- 
lowed Schiner largely, but were somewhat less conservative than he, 
nevertheless fell far short of reaching a requisite degree of radical- 
ism in their views as to a proper treatment of this superfamily. 

Others who entered the ranks during this fourth epoch, Walker, 
Bigot, Bellardi, Jaennicke, Thomson, Meade, von Roeder, Kowarz, 
Mik, followed Schiner more or less, adopting Rondani and Rob- 
ineau-Desvoidy at times on certain points, and gradually increased 
the stock of genera as seemed warranted along more or less con- 
servative lines. 

Robineau-Desvoidy had divided the Muscoidea into many smaller 
groups which he called stirpes, corresponding more or less in value 
to our present subfamilies. These were not recognized by Rondani, 
who grouped all into six stirpes. Neither Robineau-Desvoidy nor 
Rondani were really adopted by Schiner, who recognized eight 


stirpes, mainly founded, however, on certain of Robineau-Desvoidy's. 
Schiner thus largely adopted Robineau-Desvoidy's stirpes in those 
divisions which he did recognize, but bunched his genera along with 
those of Rondani, Robineau-Desvoidy's reviser. The eight taxo- 
nomic divisions adopted by Schiner generally obtained throughout 
the epoch. 

Rondani's system, unlike Robineau-Desvoidy's, took little note of 
habits, and, while less detailed, was more secure from being founded 
primarily on external anatomical characters. But these characters 
were liable to misinterpretation in certain cases. 

Brauer and von Bergertetarnm inaugurated the present and fifth 
epoch in [889, which is destined to hold out for a greater degree of 
radicalism than its predecessors. They approached the subject 
largely in a new way, greatl) lessening the difficulties of classifica- 
tion in the superfamily by recognizing a large number of sections 
which correspond to the subfamilies and tribes of the present paper. 
At the same time, they greatly multiplied the number of genera, 
whereby they were able to present comparatively concise diagnoses 
of these, as well as of their sections. 

They adopted Robineau-Desvoidy's plan of grouping the forms 
into many small divisions, hut they did not feel hound, as did he, to 
adhere to any definite scheme of life habits for indicating taxonomic 
limitations. In the main their divisions were made on quite original 
lines. However, many of Robineau-Desvoidy's old stirpes are still 
recognizable, now more or less revised, restricted or enlarged, and 
the\ must be considered as the original foundation of our present 
subfamilies and tribes. Brauer and von Bergenstamm's characters 
were better chosen and represent a more exhaustive study of the 
subject, as would naturally follow from their having enjoyed the 
greatly superior advantages derived from marked increase in biologic 
progress since the time of Robineau-Desvoidy and Rondani, and 
access to the greatly enriched collections of material drawn from all 
parts of the globe. 

Until quite recently Brauer and von Bergenstamm's system has 
been followed rather indifferently — in some cases enlarged upon, in 
some revised — by students of the group contemporaneous with them 
and continuing in the work since their time. The general trend of 
sentiment now, however, is strongly in their favor, recognizing, as 
it does, the necessity of a subdivision of the superfamily into many 
subfamilies, tribes, and genera, so as to allow of more careful and 
concise diagnoses. While it is true that a middle course between 
the two extremes of conservatism and radicalism is usually the best 
one to follow, the present superfamily furnishes a notable exception 


to the rule in that it can not be successfully treated on other lines 
than what are to be considered as quite radical compared with the 
treatment accorded to other superfamilies in the order. 

In this historical review, Robineau-Desvoidy, Rondani, and 
Brauer staryd forth prominently as the greatest students of the Mus- 
coidea that the world has produced. Each had a deeper insight into 
the peculiar relationships and affinities of the superfamily and a 
closer grasp of the subject as a comprehensive whole than any of his 
predecessors or contemporaries. 

The following is a tabular arrangement of the five epochs, with 
the respective students who belong to each, including the approxi- 
mate periods during which they were more or less active in work on 
the superfamily. The asterisk indicates those authors who estab- 
lished one or more genera. The plus sign indicates work continued 
to the present time : 

EPOCH I (prior to 1804). 

Redi, 1671-1712 (general insects). 

Reaumur, 1 738-1 740. 

Scopoli, 1760-1763. 
*Linne, 1761-1766. 

Poda, 1761. 
*Geoffroy, 1762 (one genus — Stomo.vys). 
*Fabricius, J. C. 1775-1805. 

De Geer, 1776. 

Schranck, 1781-1803. 

Herbst, 1789-1801 (general insects). 

Rossi, 1790. 

*Latreille, 1792-1805 (Trichopoda, Bucentes, Hypoderma, Ocyptera, 
CEdemagena) . 

Panzer, 1793-1809. 

Baumhauer, 1800. 

Illiger, 1801-1807 (general insects'). 

EPOCH II (1804-1830). 

*Meigen, 1804-1830. 

Schoenher, 1806-1817 (general insects). 

Gyllenhal, 1808-1829 (general insects). 

Dufour, 1809-1833. 

Olivier, 181 1. 

Germar, 1813-1821 (general insects). 

Fallen, 1814-1825. 
*Clark, 1815 (one genus — Cuterebra). 

Lamarck, 1815-1822 (general invertebrates). 
*Leach, 1817 (one genus — Gastrophilus) . 

Say, 1817-1832. 
*Dumeril, 18 19 (one genus — Ecliinomyia) . 
♦Wiedemann, 1821-1830 (one genus — Glossina). 
*Le Peletier de Saint-Fargeau, 1825 (one genus — Prosena). 


EPOCH III (1830-1850). 

*Robineau-Desvoidy, 1830-1863. 

*Perty, 1830-1834 (one genus — Diaugia). 

Haliday, 1832. 
*Macquart, 1834-1S55. 

*Bouche, 1835-1847 (one genus — Compsilura). 
*Guerin, 1835-1850 (one genus — Formosia). 
*Zetterstedt, 1838-1855 (IVahlbergia, Cinochira, Gymnopeza). 
*Bremi, 1846 (one genus — Amsteinia) . 

EPOCH IV (1850-1889). 

*R6ndani, 1850-1865. 

*Walker, 1850-1866 (Doleschalla, Schizotachina, Hammaxia, Saralba, 

Tor oca, Zambesa). 
*Egger, 1856 (Zelleria, Halidaya, Frauenfeldia, Microphthalma). 
*Doleschall, 1856 (Spiroglossa, Megistogaster). 
*Brauer, 1858-1889. 
*Bigot, 1850-1893- 
Bellardi, 1859-1862. 

:< AIeinert, 1860-1880 (one genus — Philomis, larva). 
*Loew, H., 1861-1872 {Stegosoma, Blcesoxipha, Euthera, Himantostoma, 

*Schiner, 1862-1868. 

*Jaennicke, 1867 (one genus — Archytas). 
*van der Wulp, 1867-1903. 
*Thomson, 1868 (Glaurocara, Tricharcca). 
*Osten-Sacken, 1877-1902 (one genus — Urodexia). 
*Pokorny, 1880-1896 (Parastaiiferia, Sarromyia, Steringomyia, Trigonos- 

*]\Ieade. 1881-1899. 
*von Roeder, 1881-1896. 
*Ko\varz, 1882-1894 (Ctcnocncmis, Mikia). 

*Mik, 1882-1901 (Crossocosmia, Zygobothria, Microtachina, Microtricha). 
*Williston, 1S86 + (Melanophrys, Acroglossa, Talaroccra, Dichocera, 

M elanodexia) . 

EPOCH V (1889 +). 

*Brauer, 1889-1899. 

*von Bergenstamm, 1889-1894 (co-author with Brauer). 

*Portschinsky, 1890-1902. 

*Schnabl, 1890-1902. 

*Giglio-Tos, 1891-1897. 

*Wachtl, 1891-1895. 

*Townsend, 1891 + 

*Girschner, 1893-190 1. 

*Meunier, 1892 + 

*Strobl, 1892 + 

Bezzi, 1892 + 
*Pandelle, 1894 + 

Becker, 1894-1901. 

Snow, 1895. 


Corti, 1895-1897. 
* Austen, 1895 + 
*Coquillett, 1895 + 
♦Hough, 1898 + 

Kertesz, 1899 + 

Robertson, 1901 + 
*Bischof, 1901 + 
*Grimshaw, 1901 + 
♦Hendel, 1901 + 
♦Hutton, 1901 + 

Villeneuve, 1902 + 

Wainwright, 1902 + 
*Speiser, 1903 + 
♦Johnson, 1903 + 


Speaking of the Muscoidea, Dr. Williston has said: "Species, 
genera, and even families, show such slight plastic or colorational 
differences that only the most patient study will define their limits. 
At the present time there is a decided tendency to base the classifica- 
tion of even the higher groups upon apparently trivial characters. 
Most naturalists have long since abandoned the idea that genera, or 
even families, represent anything but the conveniences of classifica- 
tion, and the recent writers on this family are probably right in seiz- 
ing upon any characters that will satisfactorily group the vast num- 
ber of species irrespective of their relative values. But it is very 
probable that, in the proposal of so many genera in such rapid suc- 
cession, many characters have been employed which future research 
will show to be entirely inadequate. We yet know very little about 
individual variations in this family, or the real value of many of the 
characters now used. The absence or presence of a bristle may be 
found to represent a group of species, but we should first learn how 
constant the character is in species. * * * Seriously, is not the 
stock of Tachinid genera sufficiently large for the present? Would 
it not be advisable to study species more before making every trivial 
character the basis of a new genus?" — Insect Life, vol. v (1892-3), 
pp. 238-40. 

These words, from the leading authority on American dipterology. 
written some fifteen years ago and shortly after the appearance of 
the first two instalments of Brauer and von Bergenstamm's work. 
may advantageously be taken as a text for some pertinent consider- 
ations at this time. 

While the great multitude of forms in the Muscoidea seems at 
first sight chaotic and formidable, the student soon perceives that 
standing: forth from the general mass there occur certain well- 


marked generic types, such as CBstrus, Cuterebva, Dcxia, Macrony- 
chia, P/iasia, Trichopoda, Meigenia, Masicera, PJwroccra, Tachina, 
Gonia, Belvosia, Plagia, Thryptocera, Phania, CEstrophasia, Milto- 
gramma, Pyrrhosia, Ocyptera, Gymnosoma, Bchinomyia, Hystricia, 
Dejeania, Sarcophaga, Calliphora, Musca, Stomoxys, Glossina, and 
at least a hundred others. These types correspond in value to the 
more settled genera of the older superfamilies, where intermediate 
form- are largely lacking. In the present superfamily, however, it 
is quickly seen that massed in between these many typical forms 
are numerous intermediate ones, which collectively vary in all direc- 
tions and combine certain of the characters of the various types. 
These intermediates are the bridges for the passage of genera, so to 
speak — the inevitable precursors and resultants in the process of the 
evolution of genera. The same holds good of species. Numerous 
intergrades are found to group naturally around and between the 
various species. That these intermediates and intergrades are pres- 
ent is due to the fact that the Muscoidea are now — at the present 
day, geologically speaking — in their period of greatest prolificacy, a 
period characterized by a condition of multiform development. 
After the lapse of a great space of time, many of these intermediate 
forms will have dropped out of the struggle, leaving a residue more 
or less well defined from each other and thus much more amenable 
to taxonomic treatment. This is now the case with the older dip- 
terous superfamilies. which have long since passed their period of 
greatest prolificacy. 

It should be explained that the term "intermediates" is used to 
designate forms of generic rank or higher, and "intergrades" to 
designate those which are only of specific rank. The further term 
"intergradants" may be employed to designate individuals which 
connect species, but upon which it is not practicable to bestow names. 

The Muscoidea are of very recent evolution — in fact, their evolu- 
tion is still going on. Here are species, genera, and families in the 
making. The whole superfamily is one enormous assemblage of 
thousands upon thousands of forms distinguishable from each other 
by only slight differences and exhibiting characters which intergrade 
in all directions. That such a multitude of closely similar forms is 
exceedingly difficult to classify goes without saying. These forms 
can not be classified in the ordinary way. but demand special treat- 
ment adapted to the conditions. 

The key to the whole situation, when it comes to methods of tax- 
onomic treatment in this superfamily, is that we have here the task 
of defining not only the numerous well-marked types corresponding 
to the existing forms in the older and less specialized dipterous 


superfamilies, but also a great mass of the intermediates, intergrades, 

and intergradants that have resulted during the long-continued 
process of the evolution of these types. 

Brauer and von Bergenstamm recognized these conditions in the 
Muscoidea and treated the superfamily accordingly. As being 
highly apropos of this subject, the following remarks are quoted 
frum the translation of these authors' Introduction (published in 
Psyche, vol. vi, pp. 313-16, and 329-32), the whole of which can be 
studied with much profit : 

"It is a fundamental principle in the development of the whole 
dipterous stock that, from the lowest {Orthorrhapha nematocera) 
to the most differentiated or highest (Cyclorrhapha schisometopa) , 
the actual value of the genus, and of the systematic series generally, 
becomes less and less. This proposition seems applicable to all 
groups of animals — in all cases the most recent forms are more 
closely related and more difficult to characterize .than older ones. 
The cause lies in the numerous intermediate forms occur- 
ring in a group of animals which has just reached its period of great- 
est prolificness." 

As the same authors point out farther along in their Introduction, 
it is absolutely futile to attempt a classification of these flies along 
any other lines than a separation into many comparatively restricted 
categories. The authors are also correct in maintaining that the 
classification of all animals must be based on the entire develop- 
ment — not on the adult alone. The characters of the imago are most 
important for genera and species ; those of the earlier stages are 
most important for families and higher categories, even up through 
orders and classes. In studying early stages, it may be pointed out 
that some characters will occasionally serve for generic separation, 
but much judgment must be exercised in deciding which characters 
are of value for this purpose, since conspicuous ones may in some 
cases possess less than generic value. Such are those of special 
adaptation to peculiar conditions of life. 

The fact should be recognized, as suggested in the opening text 
to this chapter and emphasized in the quotation just given, that gen- 
eric values are not necessarily uniform throughout the organic 
world. It is fallacious to attempt to set a standard whereby plant 
and animal genera, or animal genera alone, shall be gauged by a cer- 
tain fixed measure of difference. This holds good even in different 
superfamilies of the same order or suborder of insects. The de- 
mands of the group in hand must be considered in each case. A 
superfamily in the multiform stage of development, contingent upon 
its being still in process of evolution, demands a less generic value 


than an older and well established superfamily whose forms have 
become fixed through a long period of conformity to their environ- 
ment. If this be not conceded, it becomes impossible to treat the 
younger superfamilies by any satisfactory system. 

It will be alleged by some that such plan will result in multiply- 
ing genera unduly. There is, however, no doubt that the course 
adopted is warranted by the conditions. This conclusion has been 
reached after full and mature deliberation. The only possibility of 
successfully systematizing the superfamily, so that its myriads of 
forms can be designated definitely by name, lies in the recognition of 
genera founded upon comparatively slight characters — slight com- 
pared with those recognized as the standard in the older and less 
specialized superfamilies. The differences between genera are less 
pronounced in the more specialized than in the less specialized 
groups. All are genera, and of equal value systematically ; but, as 
already pointed out, they can not be measured by a standard gauge. 

The writer has always contended that a proper treatment of the 
Muscoidea demands the definition of smaller categories and more 
carefully restricted genera (see Psyche, vol. vi, p. 313, Sept., [892 I. 
As the characters of the early stages are investigated, more light 
will be thrown on higher divisions in the superfamily. Such a vast 
assemblage of closely related forms is not amenable to separation, in 
the adults, into divisions conceived on lines of mathematical pre- 
cision. Any system of classification must become more or less arti- 
ficial if it attempts, in the presence of intermediates and the absence 
of a knowledge of early-stage characters, to mark off precise lines of 
division between categories of higher value. When the interme- 
diates are lacking, or largely lacking, it becomes a comparatively 
easy matter to fix the lines of demarcation, and the system appears 
extremely natural simply through the absence of the immense mass 
of intermediate forms that at one time existed. But when these 
numerous intermediates and intergrades are extensively present, any 
attempt to apply an arbitrary system of classification to the group 
can not but result in disaster. A system can be thoroughly natural 
only in so far as it indicates natural types of families, subfamilies, 
tribes, and genera, and groups the intermediates and intergrades 
around them. Properly conceived and executed, such a system is 
the only natural one, since it must accord with the facts as known. 
At the same time the fact must not be lost sight of that taxonomy 
is at best merely a means to an end, and does not exist in nature. It 
is artificial in its original conception, because it is practically in- 
tended to ignore numerous steps in the development of life — steps 
that have been lost during the evolution of forms now existing, and 


whrch, if still present, would make a taxonomic system simply 

Taking these points into consideration, there is evidently but one 
course open. Draw lines of demarcation between the best marked 
types, and let the others, with their respective coteries of inter- 
mediate forms, fall in whatever divisions a preponderance of their 
characters in each case indicates. Definitions of characters for the 
higher divisions can not be exact, because the forms themselves in 
nature do not fall into well defined divisions. 

Such a system as outlined would recognize typical forms as genera 
and species, and would then intercalate necessary additional genera 
and species for the convenient reception of the intermediate forms, 
which group around the typical ones and connect them with each 
other. The one great difficulty here will be to arrive at the true 
relationships of the intermediate forms, for their affinities are often 
so complex that it is very hard to decide with what genus or species 
they are most closely related. The real truth will ultimately be 
attained only after many years of continued research into their 
ontogeny, combined with an exhaustive study of the geological his- 
tory of the superfamily. 

AYhat have been called typical forms, both genera and species, it is 
proposed to term typic. The additional genera and species to be 
intercalated between the typical ones it is proposed to term atypic. 
We will thus have a system of typic genera and atypic genera for the 
reception of typical genera and intermediates respectively, and typic 
species and atypic species for the accommodation of the typical spe- 
cies and intergrades respectively. This scheme accords with the 
facts, which do not conveniently admit of the employment of sub- 
genera and subspecies. The latter concepts are here inapplicable 
on account of the nature and intricate relationships of the forms. 
To include subgenera, the genera would have to be too loosely char- 
acterized. Furthermore, this scheme preserves the binomial nomen- 
clature, which is highly desirable. It can be designated in each case 
whether a genus is typic or atypic, if this is found desirable. 

All the more primary divisions — those above the subfamilies, up 
to the very subordinal divisions themselves — can at present be only 
imperfectly characterized and defined. Here is where aid will be 
derived from early stage characters, when these become known. 
Even the Cyclorrhapha and the Orthorrhapha 1 can not be sharply 

1 The writer is aware that Osten-Sacken claims there is a clearer line of 
separation between the Nemocera and Brachycera than between the Orthor- 
rhapha and Cyclcrrhapha, but this is outside our subject. 


differentiated from each other in the adults on account of inter- 
mediate forms. Less and still less grows the clearness of limita- 
tion as we descend through the series, sections, subsections and 
superfamilies to the families. Limitations clear a little in the 
families, but it is not until we get to the subfamilies and tribes that 
we can. from a study of the adults, begin to draw moderately well 
marked lines and sel fairly concise limits. A moderate degree of 
conciseness is possible here only because we arc now concerned with 
divisions sufficiently low in the taxonomic scale to allow the exclu- 
sion of refractory and disturbing elements, and if necessary put 
them alone by themselves. Many subfamilies and tribes are seen to 
stand out as natural groups of genera. 

At first sight it would appear advisable to ignore the higher divis- 
ions, and drop at once to the very considerable number of subfam- 
ilies and tribes necessary to the system outlined. But it evidently 
serves a better purpose to recognize these higher categories, however 
much their boundaries may he obscured by connectant forms. They 
are certainly present, and their existence should not be lost sight of. 
Therefore they should he retained in any taxonomic system as indi- 
cating steps in the evolution of these flies. They may be kept some- 
what in the background, with the caution that they can not be cle 
and concisely defined until the ontogeny of the intermediate forms is 

Man}- genera stand more or less apart and do not fall actually into 
any of the subfamilies. Very restricted groups of such genera, 
which may he termed refractory on account of either their complex 
relationships or their apparent neutrality with reference to the 
various subfamilies, will best be treated directly as tribes, without 
reference to any particular subfamily. 

Some few genera will prove to be quite isolated, and yet not enti- 
tled to subfamily or tribal rank. A final system should aim at the 
definition of as many well-marked subfamilies and tribes as possible 
to concisely characterize, and the consequent reduction of the num- 
ber of these isolated forms. A comprehensive table can thus be 
preparedj including the subfamilies, the non-referable tribes, and the 
non-referable genera in one synoptic treatment, which will be con- 
venient for general use. Separate tables can follow defining the 
genera within each subfamily and non-referable tribe. No attempt 
should be made to force refractory genera into any subfamily or 
tribe where they do not fall naturally, or any tribe into any sub- 
family where it does not clearly belong, or to antagonize natural 
affinities in any \\a\, or to combine refractory forms in one heter- 


ogeneous tribe or subfamily. The refractory elements should rather 
be left to stand alone. 

In such manner as the above will it be possible to work out a 
serviceable system of classification, which will indicate, so far as may 
be, the true relationships, and at the same time preserve appr< »x- 
imately the relative values of taxonomic divisions in the Cyclor- 

A very important point remains to be noticed: What is a species 
in this superfamily? The preceding remarks on intermediate forms 
apply especially to the higher divisions, but are also largely true of 
genera and species. The difficulties as to genera can be practically 
overcome by the erection of a sufficient number to accommodate all 
the intermediates. But who can tell what is a species in nature, and 
especially what is a species in the Muscoidea? It is clear that we 
must have a definition that will answer to the term. In large assem- 
blages of insects, where intergrades and intergradants have not been 
lost, there is no such thing as a species in the generally accepted 
sense. No sharp specific distinctions can be drawn in such cases. 
The term is a necessary conception in taxonomy, however, and it is 
to be noted that the only reason for its employment is the necessit} 
for being able to distinguish between assemblages of individuals that 
are unlike. Therefore it seems clear that the only safe course to 
pursue is to give a name to every assemblage that can be distin- 
guished from other assemblages. 

It is proposed to use the term "'species" in a well-restricted sense. 
Typic species are already explained. The term atypic species will 
be used for recognizable assemblages of individuals grouping around 
typic species. The term "forms" may be used interchangeably as 
referring to either or both. 

When two atypic species are connected by intergradant individ- 
uals, the former should be given names and the latter referred to as 
intergradants between the two atypic species. A few words of de- 
scriptive matter will serve to fix practically the exact taxonomic 
position of these intergradants. Such a course will afford students 
of bionomics an opportunity to attain some degree of definiteness in 
their investigations. As the names now stand in the Aldrich Cata- 
logue, this element of definiteness is totally lacking. Many distinct 
forms are bunched under one name on almost every page. Absolute 
exactness is impracticable in this phase of nature, where variation 
through pressure of environment is constantly at work in the evolu- 
tion of new forms. But a reasonable degree of definiteness is possi- 
ble of attainment. So long as we can refer by name to recognizable 
forms, we may be certain that we are not going wrong. Such forms 


should not be bunched merely because it is difficult to distinguish 
them. If it is possible to separate them, they should be separated. 

The conviction is constantly growing among biologists that we 
really do not comprehend species. Multitudes of insect forms have 
been confused under one specific name since systematic entomology 
began. The scientific concept of the invertebrate species is grad- 
ually growing less vague and more restricted. There is practically 
no doubt that in most groups of insects, the Coccidse excepted, there 
are many times more forms that will eventually be termed "species" 
than have heretofore been recognized. Every year new results ob- 
tained from a study of the early stages of insects force this convic- 
tion upon us. (The Coccidas probably form an exception. Mr. J. 
G. Sanders is authority for the statement that the species have been 
largely split on characters pertaining to different ages of the same 
stage.) Without doubt, bunching is infinitely more harmful to a 
system of classification than splitting. Splitting, even if inju- 
diciously done, does not give rise to actual error, but bunching pro- 
duces all kinds of error in the bionomic literature, which errors, 
moreover, are irremediable except through a restudy of the speci- 
mens originally referred to. It goes without saying, however, that 
forms can be properly separated only on constant structural charac- 
ters pertaining to the same age or stage of development, and on 
color, form, and size only when such are known to be constant. A 
plea is herewith entered for judicious splitting, 1 up to the limit of 
practicability. A reasonable degree of conciseness in the designation 
of forms of insects is absolutely unattainable by any other means. 

A word is not out of place here bearing upon the causes of varia- 
tion which give rise to vast multitudes of forms during the period 
of greatest prolificacy of a group in any order of life. 

Mr. W. L. Tower, in his paper on Leptinotarsa (Carnegie Institu- 
tion of Washington. Publication No. 48), has demonstrated that 
variation is not inherent in the germ plasm, but is invariably induced 
by external stimuli acting thereon. The demonstration consisted of 
several experiments in which the stimuli were directly applied to 
pregnant females of Leptinotarsa, so as to reach the germ plasm 
within the contained ova. This one point is by far the most im- 
portant contribution to science that the author makes in the whole of 
his long and highly instructive paper. All variations are directly 
caused bv the action of external stimuli — such as heat, humiditv, 

1 This term is adopted in a serious sense because it is both apt and expres- 
sive. Splitting can be accomplished only along lines of fi rmation or natural 
cleavage, and this is true of the proper division of taxonomic groups. 


atmospheric pressure, food, etc. — in other words, by the pressure of 
environment, which means all . stimuli taken together and acting 

It is thus seen that climatic or meteorologic conditions are potent 
factors in the evolution of forms of life, and that as a rule one form 
does not inhabit two widely different life zones or areas. Few, if 
any, forms inhabit both temperate and tropical regions, or both 
humid and arid regions. The external stimuli natural to the differ- 
ent zones and areas result in the modification of forms coming within 
the sphere of their influence, and the consequent production of new 
forms. Thus the progeny of individuals of one and the same form, 
spreading gradually through areas where they become subjected to 
new sets of stimuli, are gradually differentiated into distinct forms 
through the pressure of environment. Dr. Merriam's exposition of 
this law in his address before Section F of the American Association 
for the Advancement of Science, at its 55th meeting (Proc. Am. Ass. 
Adv. Sci., 1906, pp. 387-0), is an admirable one, and can be studied 
with much profit. His observations, as there given, agree perfectly 
with the results of the writer's studies in Diptera. For instance, the 
arid and humid regions of North America will be found to possess 
very few species in common. These very different life areas are 
divided and subdivided by temperature as we go north or south, or 
ascend above sea level, and again and again subdivided by various 
climatic and other environmental factors. The result is many sepa- 
rate life areas, more or less restricted, each of which exhibits a dis- 
tinct stamp of environment. Intergradations occur along the periph- 
eries of the ranges of closely related species, when such lie contig- 
uous, as they often do. These intergradants must not be confused 
with the normal specimens of the form as exhibited throughout the 
more central portions of the area of range. That the intergradants 
occur between two such forms does not invalidate the distinctness of 
the forms themselves. 

It may safely be stated as a theorem in bionomics that, given an 
arid area and a humid area contiguous to each other, both originally 
stocked with individuals of the same form — whether of Diptera or 
any other order of life — the descendants of this form will not remain 
identical in the two areas throughout any considerable period of 
time. The theorem may be enlarged to include temperate and trop- 
ical contiguous areas, and many divisions and subdivisions of these 
and of the arid and humid areas as well. The resultant differentia- 
tion is brought about by dynamic variation, incited by the respective 
sets of external stimuli acting on the germ plasm of the ova con- 


tained within pregnant females of the form, already referred to as 
demonstrated by Tower. 

A careful study of these factors and of the results produced by 
them demonstrates the fallacy of the idea that forms from the north 
Atlantic coast region of the United States and the south Gulf coast 
region of Mexico are identical. In other words, forms originally 
described from Vera Cruz are not to be identified in Massachusetts 
material. Likewise, forms from arid regions are not to be identified 
in humid region material. Furthermore, European species are not 
to be identified in American material, except in the few cases of 
forms that have been imported through the agency of man. There 
exist today practically no Muscoidean forms common originally to 
Europe and North America. The Muscoidea did not originate from 
circumpolar stock. The forms that immigrated to northern America 
from Eurasia during the warm periods that existed in the subarctic 
region in interglacial times have long since given rise to new forms, 
and no longer persist in their original state. 

There are certain more or less cosmopolitan flies, such as Musca 
domestica, Stomoxys calcitrans, Luciiia ccusar, Calliphora erythro- 
cephala, and others, which find their natural environment in the wake 
of man. These are not so amenable to the above factors, but even 
they show some effects of their agency. A considerable number of 
such species doubtless accompanied primitive man in his wanderings 
through various parts of the earth. Other species are of compara- 
tively recent dispersion through commercial agencies. Both classes 
have been involuntarily spread by man. The detection of the second 
class calls for extremely careful study ajid fine powers of perception. 
Still another and very recent class has been purposely spread by man 
for economic ends. 

A word may be said as to the difficulty of distinguishing between 
many of the distinct but closely similar forms that occur in the Mus- 
coidea. While many of these forms that closely resemble each other 
do so by virtue of their close relationship through common origin, it 
is evident that others of more diverse origin have developed a close 
resemblance through counterfeitism 1 attained by means of natural 

1 The writer herewith proposes a change from the use of the words mimic, 
model, and mimicry. The terms "mimic" and "model" have nothing", except 
usage and priority, to commend them. "Mimic" is exceptionally faulty, and 
does not nearly convey the intended meaning. In the strict sense of the word 
a mimic is one who, by sound or action, imitates another. The word does not 
imply any idea of form, color, or size. The word "counterfeit," however, 
embodies the full concept. Again, "model" does not carry the idea of size, 
and in an art sense only partially that of form; moreover, it is not necessarily 


selection and the pressure of environment. This may be termed con- 
vergent evolution. Somewhat similar are cases of parallelism, or 
recurring types of structure in nowise related to one another, which 
are to be explained by use and adaptation to external conditions. A 
thorough study of larval and puparium characters will determine 
such cases beyond a doubt, but in many instances an intimate knowl- 
edge of the adults will enable one to separate these forms quite accu- 
rately. Parallel series in the adult of forms of common origin will 
usually show their distinctness very readily to the experienced eye 
without a lens. In this way the writer has often made a preliminary 
arrangement of much material, which subsequent study demon- 
strated to be correctly separated into distinct forms, many of them 
so closely resembling each other that they were extremely liable to 
be confused. A very serviceable guide in distinguishing between 
forms of common origin is the character and color of the pollen, 
which is present to a greater or less extent in all the forms. This, 
strange to say, is extremely constant throughout series of individ- 
uals of the same form and the same sex. In those forms which pos- 
sess golden pollen on the head, the male as a rule has the golden 
shade more pronounced and extensive than the female. The color 
of the pollen of thorax, and especially that of abdomen, is very con- 
stant in both sexes. A slight difference in the shade of color of the 
abdominal pollen, such as that between a silvery cinereous and an 
ashy cinereous, will frequently serve to correctly separate closely 
related but distinct forms which might otherwise be confused. It is 
almost needless to say that reference is here made only to fresh and 
well-preserved specimens. Greased specimens must be restored be- 
fore attempting to place them. 

Illustrative of convergent evolution and parallelism, in which adults 
of tw r o or more forms closelv resemble each other through causes other 

imitated, often has no relation to color, and may even be a miniature or other 
representation and not the original at all. "Pattern," on the contrary, means 
the original, to be imitated as to form, size, and color, strictly speaking, and 
is the term used in mechanics in the exact sense of our concept. 

By using these terms — counterfeit and pattern — we can adhere strictly to the 
significance of our diction. We would thus speak of an edible counterfeit 
(species) of an inedible pattern (species), which latter has been unconsciously 
and involuntarily adopted by the former as a subject for imitation, impelled 
thereto by certain accruing advantages. Both words express the sense exactly, 
and both can be used without change as either nouns or adjectives. Deriva- 
tively, instead of the objectionable term "mimicry," we have the very sugges- 
tive and thoroughly appropriate name counterfeitism to apply to a subject of 
rapidly growing importance. It would seem that neither priority nor usage 
have any claim to consideration in a case of this kind. 


than those implied in close relationship, the following' is an excellent 
case in point among the Coleoptera. Mr. W. Dwight Pierce makes 
the statement that three species of Anthonomus (A. nigrinus, ccneo- 
lus, and albopilosus) , which breed in the flower-buds of Solomon 
spp. (S., eleagini folium, rostratum, and torreyi) in 
Texas, resemble each other so closely in the adult that they are often 
confused by experienced coleopterists. Yet Air. Pierce, who has 
studied the early stages of these species, has found that the anal 
characters of the pupas serve to readily distinguish them. A. ceneo- 
las and nigrinus belong in the same group, are distinguished in the 
pupa by a slight difference in the proportions of the posterior termi- 
nal structures of the anal segment, and in the adult only by color. 
But A. albopilosus belongs in a distinct group, is inseparable in the 
adult except by leg characters, and markedly different in anal 
characters in the pupa. A. albopilosus is thus a case of con- 
vergence toward ccneohis and nigrinus, which two are closely related 
forms. It should also be mentioned that albopilosus has been found 
recently breeding in great numbers in buds of Croton spp. Dr. 
Chittenden is authority, however, for its former breeding in Sola- 
tium spp. 

The reasons for such convergent evolution or parallelism are often 
difficult to ascertain and are outside our subject. This case is intro- 
duced from the Coleoptera merely as paralleling certain very similar 
ones in the Muscoidea. For example, the species Achcetoncura 
datanarum, A. promiscua, and Parcxorista futilis seem to form a 
group similar to the above species of Anthonomus. The first two are 
closely related, and the third furnishes a case of convergent evolu- 
tion in their direction. All three forms are entirely cinereous polli- 
nose, have the anal segment brassy, and the parafrontals and para- 
facials golden pollinose. (Achcetoncura frenchi has a different 
facies. but has been confused with the first two.) 

Similar groups will be found in the genera Tachina, Masicera, 
Phorocera, etc. Another group is probably exemplified in Myio- 
phasia spp., Phasioclista metallica, Bnnyomuia clistoides, and certain 
other species. 

Such conditions as the above explain why specimens of tachinids 
looking strongly alike and bred from the same 'caterpillar, perhaps 
issuing on the same date, are at times found to belong to different 
forms and ever to different genera. In such of these cases as are 
due to convergent evolution and parallelism, the larvae and puparia 
will be found to exhibit better differential characters than the adults. 
No work connected with the taxonomv of the Muscoidea could more 


solidly advance our knowledge of the subject than the careful and 
painstaking study and rearing of the early stages. It is a most 
promising and inviting field, and one whose problems are intimately 
woven w r ith subjects of broad biologic significance. 

It may be pointed out that the well-known promiscuity of ovi- 
position with reference to hosts in the Muscoidea is another evidence, 
and a necessary result, of the geologically recent evolution of the 
superfamily. The Microhymenoptera are of far more remote evo- 
lution, as evidenced by the fact that each genus is restricted to a 
group of hosts. Microhymenopterous parasites bred from host 
larvae belonging to different families may safely be pronounced off- 
hand to belong to different genera. This demonstrates a fixed habit 
of oviposition that has endured through a long period of time. No 
such fixed habit is to be found among those Muscoidea parasitic 
upon lepidopterous larvae, or among any of the superfamily except 
the CEstridae. 

It has been alleged that much of the so-called synonymy in this 
superfamily, as it stands in the Aldrich Catalogue, is due to a mis- 
guided erection of species on stunted specimens developed from 
underfed larvae, through a lack of acquaintance with the breeding 
habits of the species. It is well known to all students of the 
Muscoidea that the females sometimes, if not frequently, carry the 
act of oviposition to an extreme, ovipositing upon larvae that are 
already overstocked with eggs. This has been observed and recorded 
in a number of instances. It has been observed at the Gipsy Moth 
Laboratory of the Bureau of Entomology in Massachusetts that 
tachinids would oviposit at times upon larvae covered with eggs, 
while masses of unstocked larvae were abundant close by. Some of 
the unmolested larvae were dissected and found unparasitized. This, 
moreover, was in the open, outside the breeding cages. However 
puzzling this may seem, it is certainly unsafe to draw conclusions as 
to habits from observations made in the gipsy moth area, since the 
equilibrium of the various forms is in a state of extreme unrest. 
This is due not only to the enormous increase of comparatively 
newly introduced host elements in the fauna, but also to the more 
recent introductions of new parasitic species, both tachinid and 
microhymenopterous. These agencies have so disturbed the balance 
between species that the resultant conditions have become highly 
artificial. Similar conditions could hardly arise except through 
man's interference. Had the gipsy and browntail moths and their 
parasites spread into Massachusetts from a contiguous area, the 
change of equilibrium between them and the resident fauna would 


have taken place more gradually, and the balance between species 
would not have been so suddenly upset. It is not at all likely that 
tachinids oviposit so heedlessly as above observed, provided they are 
subjected to thoroughly normal conditions. 

As far as the recognition of stunted and underdeveloped individ- 
uals of a form is concerned, there is rarely any difficulty provided 
one is familiar with the characters. The stunted specimens always 
exhibit practically the same characters, and if there is any exception 
the true status of a specimen is quite recognizable. 


The following outline of the construction and development of the 
head capsule in CaUiphora, principally drawn from Lowne (Anat. 
Blowfly, pp. 114-16), forms a fitting introduction to a consideration 
of characters, inasmuch as those of the head take precedence over all 
others in the taxonomy of the Muscoidea. 

The Metacephalon comprises the segmented post-oral portion of 
the head. 

The Paracepi-ialox, which is formed of the two paracephala, or 
two lateral procephalic lobes of the nymph, comprises the prc-oral 
portion of the head. 

The paracephala bear the compound eyes and antennae. 

They are united in front and below and form the epistoma and 

The portion of the facial paracepJialon behind the epistoma shows 
three distinct parts. These are two bladder-like swellings, the 
anterior and posterior ccphaloceles, and the antennal ridge between 
them. The last is developed by a process from each of the two 
lateral procephalic lobes. 

The anterior and posterior cephaloceles correspond with the thin 
portion of the blastoderm which intervenes between the two lateral 
lobes or paracephala. 

The posterior cephalocele is the forehead (Vorderkopf) of the 
German embryologists. It bears the ocelli, and the front is devel- 
oped from it. 

The anterior cephalocele develops into the facial region. 

Behind the front there are two plates which extend forward from 
the metacephalon ; these form the epieepliah)t ( parafro)ital-occipital 

That portion of the procephalic lobe which lies in front of the 
antennal ridge unites with its fellow, and curves downward and 
backward over the mouth to form the prefacial region. 


When the posterior cephalocele is closed by plates of chitin, these 
are the triangular median epifrOntal, and the two frontals (frontalia 
or frontal vitta). 

The frontal sac or ptilinum consists of a great part of the pos- 
terior cephalocele withdrawn into the interior of the head between 
the frontals and the antennal ridge. 

The lunula is thus an anterior chitinized portion of this sac or 

The anterior cephalocele is the vesicle of the olfactory lobes. 

The posterior cephalocele is the vesicle of the cerebral hemispheres 
and their median ventricle. 

In the nymph the median parts of the head capsule lie in a deep 
cleft between the two lateral lobes or paracephala, and in close prox- 
imity to the ganglia with which they correspond, so that the head 
appears to be open on the median line. Sections show this to be a 
deep infolding of the inner edges of the paracephala ( Lowne). 

The two paracephala (two lateral procephalic lobes), having 
united on the median line, become the paracephalon of the imago. 

The paracephalon is opened transversely by a horseshoe-shaped 
suture running up from the cheek border on each side and passing 
between the antennal ridge and the frontals, bridged by a widely 
distensible membranous tissue (the ptilinum), on the forward me- 
dian portion of which is the lunula somite. This suture ends on 
each side at the cheek groove, which is formed in the integument by 
the mechanical strain on it when the suture is opened to thrust forth 
the ptilinum. The suture may be properly called the paracephalic 
suture, but the writer prefers to employ the term ptilinal suture. 

The following is a detailed statement of the external anatomical 
parts to be studied in the superfamily Muscoidea, arranged primarily 
in the order of their importance, and severally in the order of their 
relative position. The characters of the superfamily are to be found 
in the various features exhibited by these anatomical parts, and are 
pointed out so far as possible under each head. The parts preceded 
by i) afford characters of family, subfamily, tribal, and partly gen- 
eric value, and those preceded by 2) characters of mainly generic 
value. The terminology is made to conform so far as possible to 
that already in use. New terms are introduced only in such cases 
as demand their use for reasons of clearness, conciseness, and per- 
manence, and for such few parts as had no name and afford charac- 
ters of taxonomic value. 

The figure here introduced is diagrammatic and intended to show 
the main sclerites of the front aspect of the head, the characters 
afforded by which take rank over all others for taxonomic use within 
this superfamily. 



VOL. 51 

Front view of head of a Muscoidean fly (half in diagram), much enlarged. 

(Original, from drawing prepared by the Bureau of Entomology.) 

The heavy black line indicates the ptilinal suture. = Ocellar plate. 
VF = Frontalia. ~P~P — Parafrontals. PicPic : =Parafacials. CC = Cheeks. 
EE = Compound eyes. L = Lunula (postfront of larval insects). A = Antennal 
ridge (mesofront of larval insects). Fp Fp = Mesofacial plate (plus facialia 
equals prefront of larval insects). Fa Fa = Facialia. (Part^ from lunula to 
facialia both inclusive taken together constitute the homologue of the front of 
larval insects.) TZp = Epistoma. C\ = Clypcus. PI PI = Palpi. 



1) Ptilinal suture (through which is protruded the ptilinum of Robineau- 
Desvoidy) evenly rounded and widened above, narrowed above, subangular at 
top; its sides parallel, divergent, convergent; its termini high or low where 
they join the cheek grooves; position of its termini with relation to lower eye- 
border, epistoma and vibrissal angles. 


[Before ptilinal suture] 

1) Ptilinal area (area enclosed .by ptilinal suture = facial depression of 
descriptions plus antennal somite plus lunula; front of Berlese) of what form, 
width above and below compared with adjacent parts of parafacials and para- 

1) Facial plate (clypeus of Brauer and von Bergenstamm; face, facial 
plate, mcsofacial plate of Lowne plus epistoma; facial depression of authors, 
prefront of Berlese, transverse impression of face of Hough — in each case 
minus facialia and plus epistoma) produced and swollen in middle like the 
bridge of the nose, merely swollen nose-like below, tube-like, projecting for- 
ward in profile below, flat, even, elongate, reaching almost to lower margin of 
head, extending obliquely downward and posteriorly, reaching straight down 
between vibissal angles, widened below same ; shortened in front view, ending 
high above lower margin of head ; widened below, oval, triangular, compara- 
tive width above and below, narrowed high or low by the facialia or by the 
vibrissal angles. 

1) Mesofacial plate (do. of Lowne; facial plate minus epistoma). 

1) Foss.E of facial plate [fovea plus foveal sinuses) long, short, wide, 
narrow, deep, shallow, curved, straight. 

2) FovE-E (fovece of Robineau-Desvoidy ; antennal grooves of descriptions ; 
simply depressions in the facial plate) deep, shallow, elongate, short, double, 
single, and confluent. 

1) Foveal sinuses (more or less linear grooves which in certain cases 
form outlets of the fovea: anteriorly) linear, widened, deep, faint, convergent, 
divergent, etc. 

2) Facial carina (keel of descriptions) present, absent, developed only 
above, weak, strong, high, sharp, knife-like, thin, thick, flattened, rounded, 
widened, canaliculate or furrowed on its median line, or simple. 

1) Facialia (facialia of Robineau-Desvoidy and Osten-Sacken; facial 
ridges of descriptions; facial edges of paracephalon of Lowne; Vibrissenleisten 
of Brauer and von Bergenstamm; vibrissal ridges of Hough) parallel, gradually 
convergent below, short, long, bare, ciliate, narrow, sharp, widened, flattened, 
divergent, or absent. 

1) Facial bristles (those on facialia; Vibrissen of Brauer and von Ber- 
genstamm) ascending less than half way on facialia, or half way, or to point 
opposite lowest frontals, or nearly or quite to base of antennas ; in one or two 
rows, bushy, in irregular position, short, weak, long, represented by many 
rows of fine hairs, normal with hairs among the bristles, only one or two 
above vibrissa?, or wholly absent. 

1) Vibrissal angles (Vibrissenecken of Brauer and von Bergenstamm; 
angles or corners where the facialia and peristomalia meet) pronounced, 
weak, high above the lower margin of head, set low, rounded, sharp, or absent. 

2) Vibrissal papillae (Vibrissemmilste of Brauer and von Bergenstamm; 
sometimes present at vibrissal angles) prominent, pronounced, flattened, weak, 
inconspicuous, or absent. 

1) Vibrissa (the two longest or strongest bristles, one at each vibrissal 
angle; Vibrissen of Brauer and von Bergenstamm) approximated, widely 
separated ; their insertion on, close to, well removed from the oral margin, or 
on, close to the under margin of the head, or on the upper edge of the oral 
margin when this is turned up and broadened, or on or near end of facial plate, 


on a level with uppermost front edge of oral margin, or above or below same. 

i) Peristomalia (lateralia of Robineau-Desvoidy ; peristomal ridges, the 
ridges on lower edges of peristoma or checks, extending to vibrissa) with one 
or many rows of bristles, extending hi w far up; parallel above oral margin, 
divergent, convergent; parallel, divergent, convergent posteriorly below oral 
margin ; effect on epistoma. 

2) Peristomal bristles (those on peristomalia) strung, weak, in one or 
more rows, or few and with row of hairs. 

1) Epistoma (epistoma of Rob.-Desv.; Mundrand of Br. and v. Berg.; the 
portion of facial plate below vibrissal angles and enclosed between the peristo- 
malia, its point of junction with the mesofacial plate being indicated by the 
vibrissal angles) projecting nose-like, prominent in profile, retreating, set 
back or removed, produced downward or anteriorly, turned up, drawn out 
tube-like, transversely cut off, broad, narrow, thin ; thickened, widened on 
edge, callous or indurated, projecting forward and downward below vibrissas ; 
drawn up in middle to form anterior part of narrow oral slit, its sides thereby 
becoming nearly parallel; square, or curved in front outline. [The characters 
of the epistoma are usually best included in those of facial plate, of which it 
forms a part.] 

Oral margin (the anterior edge of the oral cavity, being the lower edge of 
epistoma) . 

1) Oral cavity covered over transversely in front with an oblique pos- 
teriorly-extending skin or membrane developed probably from the clypeus, 
open, elongate, short, wide, narrow, deep, shallow, slit-like, or closed. 

1) Clypeus (clypeus of Rob.-Desv., Lowne, and Berlese; the anterior or 
dorsal plate of the cephalopharvngeal skeleton, or fulcrum, of the rostrum) 
distinct, rectangular, triangular, developed into a plate closing oral cavity, or 

1) Mouth parts normal, vestigial, immovably fixed at base of shallow oral 
cavity, hidden in a narrow deep oral slit, or wanting. 

2) Proboscis short, fleshy; not longer than head height, shorter or longer 
than same ; very elongate, bristle-like, twice geniculate, once geniculate, slender 
and horny, large, stout, vestigial, or absent. 

2) Labella well developed, large, broad, small, vestigial. 

2) Palpi absent, vestigial, filiform, club-shaped, strongly elongate, normal. 

1) Longitudinal axis of head at oral margin longer than that at insertion 
of antennae, or the two equal, or the former shorter. 

1) Facial profile advancing thereby, or more or less straight or concave, 
or receding or convex. 

1) Facio-peristomal profile angular, rounded, strongly or gently convex. 

1) Antennae (arising from antennal ridge of Lowne; from antennale or 2d 
somite of front of Berlese) inserted above, on. or below a line drawn through 
middle of eyes ; above or below middle of extreme head height, widely sepa- 
rated or closely approximated. 

1) Second antennal joint strongly elongate compared with first, longer 
than shortened third joint, normal, with or without strong bristles on front 

2) Third antennal joint entire, fissiform in one or both sexes, elongate, 
narrowed, widened, enlarged, with curved point on front apical corner, normal. 

1)2) Arista bare, microscopically pubescent, hairy, pectinate, partly or 
wholly plumose, geniculate, flattened, thickened in what part of its length; first 


and second joints elongate, short, strongly elongate; or only second joint 
strongly elongate, its length compared with its width or with the third joint. 

1) Lunula (postfront of Berlese) enlarged in middle inferiorly and supe- 
riorly into a more or less diamond-shaped or rounded plate, like an extension 
of the facial plate into a secondary one; elongated below between the antennae 
into a keel-like prolongation, widely separating the antennae, or normal. 

Note. — The lunula readies its greatest development in the Syrphoidea. 

( X. B. — Mesofacial PLATE [= 2 meso facials of Lowne + carina if present, 
since latter is formed by inner edges of the two mesofacials] + 2 Facialia + 
antennal ridge + lunula = homologue of Front of larval insects [= ptilinal 

[Behind ptilinal suture] 

2) Eyes absolutely bare ; thinly microscopically hairy, sometimes distinctly 
so, sometimes indistinctly so ; thickly pubescent, sometimes more so in male, 
less so in female ; reaching as low as vibrissa?, or lower, or only to middle of 
face, or very short. [N. B. — In comparisons last mentioned, hold head in full 
profile with plane of posterior aspect of occiput perpendicular.] 

2) Vertex wide, narrow, comparative width in sexes. 

2) Vertical bristles present or absent, or present only in female; pro- 
clinate, reclinate, divergent, convergent. 

2) PostvErtical bristles (+ postocellar bristles = lesser oecllar bristles of 
Hough) large or small, separated or approximated, how many pairs. 

1)2) Front prominent in profile; flattened, or only anteriorly so; bulging, 
narrow, wide, widened anteriorly, conically produced, of equal width, or 
not so. 

1) Ocellar plate (stemmata of Rob.-Desv. ; epifrontal of Lowne) triangular, 
rounded, large, small. 

1) Ocelli separated, approximated. 

1)2) Ocellar bristles (Occllcnborstcn of Brauer and v. Berg.; greater 
ocellar bristles of Hough) strong, weak, proclinate, reclinate, divergent, 
vestigial, or absent. 

1)2) Postocellar bristles (a second or posterior pair sometimes present on 
ocellar plate just behind the two posterior ocelli; + postvertical bristles = 
lesser ocellar bristles of Hough) present, absent, or represented by fine hairs 

2) Preocellar bristles (do. of Hough; small pair on frontalia in front of 
anterior ocellus) present or absent. 

2) Frontalia (frontalia of Rob.-Desv.; frontals, mesofrontals of Lowne; 
frontal vitta of descriptions) polished, opaque, wide, narrow, long, short, equi- 
lateral ; widened or narrowed anteriorly or posteriorly, or in middle ; square in 
front, notched in front or behind. 

2) Parafrontals (optica frontis of Rob.-Desv.: parafrontals of Lowne; 
sides of front of descriptions ; geno-vcrtical plates of Hough) swollen, dilated, 
bare except for frontal and fronto-orbital bristles, hairy, bristly, short, long, 
wide, narrow, equilateral ; widened before or behind, or both ; prolonged 

1) Frontal bristles (those inserted on the inner edges of the parafrontals, 
always convergent, often extending posteriorly only to point about half way 
between ptilinal suture and vertex; transfrontal bristles of Hough) in a single 
tow, in two or more rows; descending below base of antennae, continuation 


below represented by row on parafacials descending nearly as low as oral 
margin, or about half way down, or less than half way, or not descending 
below base of antennae; or represented only by one or more rows of weak 
bristly hairs on parafrontals. 

2) Upper fronto-orbital bristles (those on posterior portion of para- 
frontals immediately in front of the vertical bristles and often appearing as a 
continuation of frontal rows posteriorly, always reclinate; ascending frontal 
bristles of Hough) in line with frontal bristles, or with middle fronto-orbital 
bristles; position, direction, number; or absent. 

2) Middle fronto-orbital bristles (Orbitalborsten of Brauer and v. Berg.; 
fronto-orbital of Osten-Sacken; orbital bristles of descriptions; they are 
usually a little nearer the orbit than the preceding, and always proclinate) 
present in both sexes, or in female only, strong, weak, divergent, convergent ; 
one, two, three, or a row, or represented only by weak hairs ; or absent in 
both sexes. 

2) Lower fronto-orbital bristles (lower fronto-orbital of Osten-Sacken 
and Williston; occurring occasionally in the Acalypterata, but rarely in the 
Muscoidea) present or absent, number. 

1)2) Parafacials (optica faciei of Rob.-Desv. ; IVangen, gence of Brauer 
and v. Berg.; sides of face of descriptions; gence of Hough) widened above. 
or not so ; bare, hairy, bristled ; widened below and narrowed above, more or 
less swollen, very wide, very narrow, elongate, short; or narrowed, shortened, 
or abbreviated below. 

2) Facio-orbital bristles (those on parafacials) present or absent, number, 
position, direction. 

1) Cheeks (peristoma of Rob.-Desv.; Back en, peristoma of Br. and v. 
Berg. ; buccce of Hough) wide, narrow, very narrow ; width equaling or ex- 
ceeding eye height, or equaling what proportion of eye height; naked, hairy, 
bristly, or so only below or behind. [N. B. — Br. and v. Berg, give apparent 
height (not width) of cheeks as seen in profile, with eyes included. Their 
actual greatest width (distance from peristomal margin to eye) should be 
compared with eye height, as seen in front view.] 

1) Cheek margins (portions bordering on parafacials and ptilinal area) 
ascending, encroaching on face, more or less circumscribing the facial plate. 

2) Cheek grooves (mediana of Rob.-Desv.) present, well defined, curved, 
wide, deep, shallow, position, vestigial. 

2) Cheek bristles (strong bristles which sometimes occur on cheeks near 
lower border, slightly outside of peristonialia) present or absent, number, 
direction, position. 

2) Posterior orbits (bare space between posterior eye margin and row of 
hairs fringing occiput) widened below, narrowed above, of even width, wide, 
or narrow. 

2) Lower margin of head (lower border as seen in profile) straight, bulged 
downward or outward posteriorly, long, short. 

2) Occiput (all the portion of the head behind the plane which defines the 
limit of the posterior orbits, as marked by the fringe-like row of small bristles 
or hairs bordering same and called by Hough and others cilia of posterior 
orbit) evenly swollen, flat ; flat above and swollen below, bulging the cheek 
profile posteriorly. 

1)2) Parafrontal-occipitai. ridge (ridge-like sclerite formed by what 
seems a continuation of parafrontals over vertex on occiput and which bifur- 


cates above great central foramen; cerebrate of Rob.-Desv. ; epicephalon of 

1)2) Occipito-cextral bristle (do. of Hough ; small bristle on parafrontal- 
occipital ridge just below inner vertical bristle before bifurcation of ridge) 
present or absen£, character of. 

1)2) Occipito-laTERal bristle (do. of Hough; small bristle on occiput just 
below outer vertical bristle) present or absent. 

2) Occipital area (the characteristic hairy area of occiput which some- 
times invades the cheeks posteriorly) invading cheeks, or restricted to occiput. 

2) Longitudinal diameter of occiput (shows its degree of swelling at any 
specified point) above or below compared with eye width in profile. 

2) Beard (pilosity arising and depending from lower portion of occiput, 
and in certain cases clearly defining a portion of cheeks invaded by occipital 
area') long, short, thick, thin. 


1)2) Sternopleural bristles one, two, three, or more, in what arrangement. 
1) HypoplEural bristles strong, weak, or represented only by hairs. 
1) PteroplEural bristles strong, weak, or hair-like. 

MesoplEural bristles very strong, or normal. 

1) ProplEural bristles strong, weak, number, direction. 

1 1 NotoplEural bristles (postliumeral of Osten-Sacken) strong, weak, 

2) Postsutural bristles (dorso central of Girschner behind suture; outer 
dorsocentral of Osten-Sacken behind suture) strong, weak, relative strength, 
number, position. 

2) Dorsocentral bristles (dorsocentral of Girschner before suture; outer 
dorsocentral of Osten-Sacken before suture) strong, weak, relative strength, 

i 12) Acrostichal bristles (2 middle rows both before and behind suture) 
strong, weak, number, position. # 

2) Humeral bristles strong, weak, number, direction. 

1) Intrahumeral bristles (postliumeral of Girschner) present or absent, 
number, position. 

1) Presutural bristles (+ postliumeral of Girschner = intrahumeral of 
Osten-Sacken) strong, weak, position in relation to preceding. 

1) Intraalar bristles strong, weak, whether one in front of suture. 

2) Supra Alar bristles (+ postalar = supraalar of Osten-Sacken) strong, 
weak, number. 

2) Postalar bristles strong, weak, number. 

1)2) Scutellar bristles strong or weak, comparative strength of the vari- 
ous pairs, number of lateral pairs ; a weaker apical pair present or absent, 
erect, suberect, directed posteriorly, decussate, or divergent; discal pairs 
present or absent. 


2) Wings broad, long, narrow, short; costal margin swollen or dilated in 
male, or in both sexes, or normal. 
2) Costal spine distinct, strong, weak, double, or absent. 
2) Longitudinal veins bristly, to what extent, or bare. 


1)2) Fourth longitudinal VEIN incomplete, straight, not forked, reaching 
neither the wing margin nor the third vein, normal, ending at or before wing- 
tip, angular or rounded at bend, bowed or not beyond bend, bend approximated 
to or removed from hind margin of wing; last section forming petiole of 
apical cell when latter is petiolate, or third vein in such case forming petiole; 
or forked and main vein represented beyond apical crossvein by only a -hurt 
stump, or by a mere wrinkle or fold in the wing-integument, or by a long 

2) Apical crossvkix (this term should be employed only when the fourth 
vein is furcate, or shows indication of previous furcation in a stump, told or 
wrinkle) bent in, straight, oblique, long, short, absent. 

1) Fifth LONGITUDINAL VEIN bent up to fourth vein, net forked; or furcate. 
giving off posterior crossvein: represented beyond latter by a short stump, or 
a long one, or only by a wrinkle, or partly by stump and wrinkle, or continu- 
ous to wingborder. 

1)2) Posterior crossvein (term not to be employed in the few cases where 
fifth vein shows no sign of furcation) oblique, in line with apical crossvein or 
with last section of fourth vein, or still more oblique than latter, or normal ; 
nearer to bend of fourth vein (or to origin of apical crossvein) than to small 
crossvein, or nearer to latter, or about in middle between the two; trisinuate, 
bisinuate, singly curved, straight. 

2) Small crossvein* on. before, "i" behind middle of discal cell; short, long, 
straight, oblique, direction. 

1)2) Apical cell (first posterior of descriptions) ending near wingtip 
far before; open, closed in margin, or long or short petiolate, or extremely 
short petiolate: wide, narrow, short, elongate, tapering equilaterally at apex. 

2) Tegul^E large, small, relative size of two scales; deeply smoky or in- 
fuscate, or while, or yellow; bare, pubescent, or hairy. 


1)2) Abdomen (shape of whole") linear, cylindrical in one or both sexes, 
widened on some portion, conical or oval in both sexes, swollen, convex dor- 
sally, concave ventrally, flattened in one or both sexes, or laterally compressed. 

1) Abdominal SEGMENTS apparently four, or how many visible from above; 
bow many actually present, which ones shortened, and relative development 
of their respective dorsal and ventral plates. [See notes on Gymnosoma. 
Trichopoda, Rhacho'epalpus, etc., under head of Descriptions. In many cases, 
at least, there are more segments in the Muscoidean abdomen than have here- 
tofore been recognized, an undeveloped basal segment being quite bidden from 
view, and only visible with difficulty on the sides below. Its dorsal and ventral 
plates are easily seen on detaching the abdomen. In order to avoid confusion, 
the old terms "first," "second,"' "third," and "fourth"' segments are retained as 
referring to those apparent from above in the undetached abdomen.] 

2) Abdominal present or absent, bristle-like, true, very strong, 
thorn-like, discal and marginal, or only marginal ; discal present on second and 
third segments (counting apparent segments from above), or only on third 
and fourth, or only on fourth ; marginal present on all, or absent on first, or 
absent on both first and second segments. 

1) Ventral membrane < membrane connecting the ventral and dorsal plates 
of the abdominal segments) visible, concealed by the sides of the dorsal 
sclerites or plates, or apparently absent. 


1) Ventral plates free, or not so; or that of second segment in both sexes 
with its edges upon and covering the edges of the corresponding dorsal plate, 
the other ventral plates free, or this true of only one sex; how many ventral 
plates, last one in male deeply or weakly Y-cleft or V-cleft, or entire. 

2) Ventral carina present in female, absent, rudimentary, more or less 
developed, emafgination of plates of same, or latter entire. 

2) Ovipositor elongate, short, tapering, stout, furnished with terminal hooks, 
appressed, exserted ; directed downward, or forward, or posteriorly; integu- 
mental, membranaceous, or horny. 

2) Hypopycium prominently exserted, elongate, appressed, directed down- 
ward, short, rounded, bulb-like, tube-like, of what formation and character- 


1)2) Legs strongly elongate, only moderately so, short, or only one pair 
elongate, relative length of pairs; bristly, bare, shaggy-haired, with or without 

2) Hind femora ciliate or not so, character and position of the cilia. 

2) Hind Tible completely and densely feather-barb-ciliate, only comb-ciliate, 
subciliate, with some longer bristles; cilia flattened and widened, scale-like, 
bristle-like, or of what character. 

2) Middle tible with or without strong bristles or macrochsetse on outer 
side, or on any portion. 

2) Tarsi slender, swollen, compressed, short, elongate, relative length of 
pairs in each sex; last joint or more of which pairs oval, thickened, swollen, 
or compressed, in one or both sexes. 

2) Metatarsi short, elongate, comparative length with relation to other 
tarsal joints of same pair, comparative length of pairs, slender, stout. 

2) Front tarsi widened in female, or widened and flattened, or only flat- 
tened, in one or both sexes. 

2) Claws and pulvilli elongate in male, or in both sexes, or short in both, 
or only anterior ones elongate in male ; claws stout, slender, curved, shape and 
character; pulvilli of what shape and character. 

While the foregoing' enumeration of anatomical parts affording 
characters of taxonomic value in the superfamily is not necessarily 
complete, it is believed that it brings out practically all the characters 
requisite to a proper separation of the forms in the adult. 

Of all these characters, those of the head take first rank. For this 
reason much space has been devoted to their consideration — in fact, 
nearly twice as much as to all the other characters together. It is 
conceded that the Schizophora are the most specialized insects, the 
most highly developed from the standpoint of ontogeny, as evidenced 
by their remarkable and practically complete reorganization of larval 
parts within the nymph. Everything points to the Muscoidea as the 
most highly organized Schizophora, and this is emphasized by their 
acute sensory development. It is therefore naturally to be expected 
that certain non-functional parts of the head, which is the chief seat 


of the specially developed senses, should afford the most important 
characters for taxonomic use. 

Here, and practically here alone in the Muscoidean anatomy, are to 
be found certain useful atavic characters pertaining to organs not of 
any functional importance in the economy at present, but possessing 
phylogenetic significance as indicating origin and relationships. 
These are of especial value for the separation of families and sub- 
families. It has long been recognized that rudimentary organs in 
recent forms bear a significant relation to those of their allied prede- 
cessors. Such are physiologically non-functional now. and appear in 
more or less developed condition only in the embryo, but were func- 
tional throughout life in the early fossil forms. They have been lost 
through disuse, involving a process of degeneration or retrogressive 
development. If, then, these organs present sufficient variation, their 
rudimentary presence is of much importance to us in the prepara- 
tion of a natural taxonomic system. Atavic characters, to be of use, 
must be exhibited by parts which vary sufficiently to offer con- 
veniently distinguishing marks. To be of use in the separation of 
higher, or family, divisions, the parts must present just enough 
variation to offer distinctive characters that will hold throughout 
considerable aggregations of forms. 

Such are the characters afforded by the facial plate in its lower 
extent, and by the facialia, vibrissa! angles, and peristomalia. The 
parts in question present sufficient variation to afford distinguishing 
characters. These are all atavic, and possess in consequence a high 
phylogenetic significance. They are connected with the portions of 
the head whose development in the nymph is not influenced by the 
coincident development of functional parts. While the development 
of the highly sensory third antennal joint affects in a degree the 
upper portion of the facial plate and determines the character of the 
fovese, its influence does not extend below the -vibrissal angles. 

Atavic characters are afforded by the wing veins in a remarkable 
degree, but the general plan of venation is too uniform to afford us 
good family characters. They can be used in higher and lower 
divisions. It may further be noted that, since the wings are so 
highly functional in a mechanical (not sensory) way, the characters 
derived from lesser variations in venation would in any event be sec- 
ondary in importance to the head characters just mentioned. It 
must be borne in mind that the wings are of great functional im- 
portance, and the veins bear the mechanical strain incident upon 
their use, while the special head characters above pointed out, whose 
importance as affording distinctions for higher divisions has been 
dwelt upon, are in nowise connected with any present function, 


either mechanical or sensory, in the economy of the adult insect. 
The type of venation furnishes atavic characters of value in sepa- 
rating higher divisions. The bristles or hairs of certain thoracic 
plates likewise furnish atavic characters of high value here. 

Atavic characters also occur to a limited extent in the abdomen, 
chiefly in the atrophied basal segment, which can be clearly made out 
only by detaching the parts. These are also too uniform to be of 
use for the separation of the larger divisions, so far as we yet know. 
But their comparative study offers promising results. 

Practically all the other portions of the Muscoidean anatomy are 
preeminently functional, even including the halteres, tegukne, etc., 
and the parts of the head other than those enumerated above. The 
frontalia and lunula may he practically non-functional, but they like- 
wise do not present sufficient variation to offer any useful characters 
for family separation. The second antennal joint is probably not 
functional, although in the Nemocera it is the seat of the so-called 
"Johnston's organ," whose function is supposed to be auditory. This 
organ does not appear to be developed in the Cyclorrhapha. Prac- 
tically the only character afforded by the second antennal joint, 
however, is that of relative length compared with the first joint, and 
this is at best available only for subfamily and generic separation. 
The arista is doubtless functional. A consideration of certain char- 
acters of functional parts, and especially of the physiological func- 
tions of certain of these parts whose characters have in the past been 
largely used in taxonomy, is now taken up. 

Antennas proper. — The first and second antennal joints are prac- 
tically non-functional. The third joint is highly functional, and 
hence does not afford reliable taxonomic characters for higher 
divisions than species, and within certain limits for genera. The 
relative length of third joint to second affords no valid character, 
and especially gives a wrong impression in those forms having the 
second joint elongate. The first joint is almost universally short, 
but the second is often more or less elongated, and in some cases 
strongly so. The relative length of second joint to first affords a 
good generic character. The third joint affords excellent specific 
characters, so far as its relative length and size go, with proper 
recognition of sexual variations. Its shape may furnish characters 
of generic, or even of tribal, value. 

The olfactory sense is very highly developed in the Muscoidea. 
Blow flies will come for miles to decaying, and even to fresh, meat 
shortly after its exposure to the air. Most other members of the 
superfamily possess this high olfactory sense, though in some it is 
developed in a varying degree. The sense of smell in these flies is 


located in the third joint of the antennae, which contains numerous 
olfactory pits communicating with the main nerve trunk by means of 
minute nerve-ends. 

According - to Gustav 1 fauser (Zeitschr. f. Wissens. Zool., xxxiv, 
pp. 307-403. 1880), who studied over sixty species of Diptera in this 
connection, the Muscoidea and other cyclorrhaphous Diptera. and 
also the Brachycera. have the olfactory pits without exception con- 
fined to the third antennal joint. Their number varies greatly in 
different forms of Cyclorrhapha. Certain syrphids, as Helophilus 
ftorens, have only one pit on each disk of the third joint, while 
Bchinomyia grossa has two hundred. In certain forms the pits are 
compound, containing from ten to one hundred olfactory hairs aris- 
ing from the coalescence of the several original pits. Xo compound 
pits occur in the Tipulidae, but only simple ones with a single olfac- 
tory hair, such as are found in the brachycerou- | s. str. 1 forms only. 
The latter have also compound pits, containing from two to ten 

The olfactory pits are sac-like invaginations of the external chit- 
inous integument, and are of various shapes in different fornix of 
diptera. They are always open externally, and never closed by a 
membrane. In the Cyclorrhapha, and the Muscoidea especially, the 
pits differ hut little in the various forms. Ilauser (1. c.) figures and 
describes at length those of Muscina stabulans as generally typical 
of not only the Cyclorrhapha, but the Brachycera s. str. as well. He 
gives a figure of the third antennal joint in longitudinal section 
showing simple and compound pits, the pits themselves being shown 
in both transverse and longitudinal section and from above. The 
main nerve trunk, accompanied by the much smaller tracheal trunk, 
passes through the second antennal joint entire and without division. 
but on entering the third joint gives off a very small branch to the 
arista, to which also runs a small branch of the trachea. The bulk 
of the nerve trunk continues undivided and undiminished into the 
mass of the third antennal joint, where it branches in all direction-., 
but especially apically and inferiorly (opposite the edge bearing in- 
sertion of arista), the main trachea following it with less branching. 
This centralization of nerve-branches, nerve-ends, and olfactory pits 
in the apical and ventral tracts of the third antennal joint — that is to 
say. outside the aristal area — bears out the conclusion that the arista 
was originally terminal and that the highly functional extra-aristal 
area of the joint has simply grown away from it as fast as more 
space was required by the advancing development of the olfactory 


It has been conclusively proved by the experiments of Hauser and 
others that the sense of olfaction is located exclusively in the an- 
tennae in Sarcophaga, Callipkora, and Cyuomyia, and not at all in 
the palpi. This has also been demonstrated in many Hymenoptera, 
Lepidoptera, Orthoptera, and Staphylinidae ; but in certain Hemip- 
tera experimented with it was found that the loss of their antennae 
did not affect in any way their sense of smell. Certain Coleoptera 
were only partially affected by the excision of their antennae. 

The olfactory organs of the Muscoidea consist of (i) a thick 
nerve trunk arising from the brain and passing into the antennae; 
(2) a sensitive apparatus at the end, consisting of rod-like modified 
hypodermis cells, connecting with the nerve-fibre terminations; (3) 
a supporting and accessory structure consisting entirely of pits. The 
same is true of the other Diptera, the Lepidoptera, Orthoptera, and 
probably the Hemiptera ; but in the Neuropteroid orders, the Coleop- 
tera, and the Hymenoptera, the accessory structure consists of peg- 
like projecting epidermal invaginations filled with a serous fluid. 
Both pegs and pits occur, however, in the Coleoptera and Hymen- 
optera, while only tactile hairs were found by Hauser in Pyrrhocoris 
of the Hemiptera, though Lespes has recorded the presence of pits 
in that order. 

It should be mentioned here that another sense, capable of distin- 
guishing between various degrees of atmospheric pressure, is be- 
lieved to reside in certain sensory structures, like the sensillum placo- 
deum, found in the antennal joints of bees and wasps. It is evident 
that insects have some means of perception, through certain sense- 
organs, of approaching changes in meteorologic conditions. 

Arista. — The arista is the persistent rudiment in the Cyclorrhapha 
of the terminal antennal joints still to be found in many of the lower 
groups of Orthorrhapha. In the development of the third antennal 
joint of the Muscoidea as a special olfactory sense organ, the arista 
has become dorsal or basal, being left to occupy a position to one side 
during the extraordinary development of the joint away from it. It 
is invariably situated close to the base of the front edge of the joint. 
Its persistent retention in this position indicates that it is to some 
extent functional. 

It is a rule in nature, which carries no exception, that there is a 
reason for everything that exists. Therefore there is some cogent 
reason for the pubescence, plumosity, and nudity of the arista, as 
well as for its presence. The arista has become subordinated to the 
third joint, but retained as an accessory. It therefore must be func- 
tional. The point is to discover its function, which must be the key 
to the explanation of its varying degrees of pubescence and plumos- 


itw The joint is mainly olfactory, and certainly highly sensory. As 
such it is highly important to the insect. The arista is directed for- 
ward, outward, and downward from its insertion on the anterior 
basal edge of the joint. This would indicate that it is primarily 
functional as a tactile sensory organ for the protection of the highly 
functional third joint. Such an indispensable organ in the economy 
of the insect as the third antenna! joint would naturally demand the 
presence of some tactile sense organ extended before its exposed sur- 
faces, to serve as a warning against contact with foreign objects. 
In other words, the arista has taken to itself the original function of 
the antenna, on account of the latter being practically turned into an 
olfactory sense organ. The bristles of the facial and frontal areas 
protect the other parts of the head from injurious contacts. 

What light does this function of the arista throw on the question 
of its nudity, pubescence, or plumosity? Simply that the separate 
hairs have a tactile function, pointing in all directions from which 
danger may come. It is to be noted that the plumosity is always 
stronger on the upper or outer than on the under or inner side. 
Those forms which have the basal joints of the arista elongated lack 
the plumosity. This elongation of the basal joints indicates an in- 
creased freedom of movement of the arista. When bare of plumos- 
ity the arista either is long and tapering, indicating a somewhat 
restricted movement in the comparatively short basal joints, or it is 
short, stout, and geniculate, with greatly elongated basal joints, indi- 
cating much freedom of movement. The nudity of the arista may 
be generally taken to indicate greater freedom of movement in its 
basal joints, and its shortening", when combined with geniculation, 
still further increase of movement. In any case, the function of the 
organ is seen to lie a tactile one, intended to guard the highly sen- 
olfactory pits and nerve-ends located in the third antennal joint. 

Those forms which have the arista more or less atrophied doubt- 
less have the third antennal joint less highly olfactory and more 
tactile in function. 

From this functional nature of the arista we can only conclude, 
in accordance with the general and almost invariable rule, that it 
possesses little value for the definition of subfamilies and higher 
groups, but that its characters may well be employed in the separa- 
tion of tribes, genera, and species. 

Eyes. — The organs of vision are with little doubt more highly 
developed in the Muscoidea than in any other superfamily of non- 
aerial insects. These flies possess, on the whole, a distinctively 
terrestrial life-habit, in contradistinction to an aerial one. The rela- 
tively small percentage of ach?etophorous and subachretophorous 


forms, and even the few of these possessing the aerial or hovering 
habit, maintain practically the same type of eye-structure, extensive 
holopticism of the type obtaining" in the Bombyliidae and Tabanidae 
being present in none of them. Partial holopticism is present in 
very few, and there is a considerable approach to this condition in 
certain others, but dichopticism is practically the rule. In no other 
group of insects of a generally terrestrial life-habit is there so rela- 
tively large an area of the head occupied by visual surface. 

This and other facts further argue for an average higher develop- 
ment in the Muscoidea of the visual sense per sc than in any other 
equally extensive group of insects, or perhaps in any other group 
whatever. The Odonata, Hymenoptera, Lepidoptera, brachycerous 
and nemocerous Diptera, and some other insects which equal or sur- 
pass them in relative visual area of the head, do so by virtue of the 
correlative evolution of visual surface and aerial life-habit. But 
their eye structure is less highly developed. While the number of 
facets in general in the Muscoidea is not nearly so great as in Odo- 
nata, certain Lepidoptera, and even Coleoptera, their eye is of a 
higher order of organization. The Muscoidea possess what is called 
the pseudocone eve, which is the most highly evolved type of the 
facetted eye. 

It is generally conceded that insects possess what may be termed 
microvision. Their ability to perceive certain minutiae approaches 
that of the human eye supplemented with the microscope. The pres- 
ence of this microvisual sense in insects is the cause of the mar- 
velous beauty of coloring and sculpture exhibited by their external 
parts, and which is revealed to us in detail only by the use of a lens. 
In other words, the facetted insect eye gains impressions from light 
rays by which the unaided vertebrate eye is unaffected. Most birds, 
especially the condor and other birds of prey, and some mammals, as 
the big-horn sheep of our western mountains, have a specially devel- 
oped far-sight, approaching in a degree the power of the human eye 
aided by the telescope. Contrasted with this is the extreme near- 
sight of insects, which do not see in general more than a few feet, 
and which see best at very close range. 

Johannes Midler's mosaic theory of insect vision, which gained 
such wide credence, especially as modified by Huxley, really seems 
untenable and quite at variance with well-known facts. It presup- 
poses a verv imperfect vision, which can not be the case. Lowne's 
dioptric theory, which indicates a perfect microvision. with sharpness 
and clearness of sight, would appear to be the correct one. Yet sub- 
sequent investigators, notably both Hickson (1885) and Hewitt 
(1907), hold that Lowne's interpretation of the functions of the 


compound eve structures is incorrect. However this may be, it 
seems certain that insects possess a clear and perfect vision. 

Mouth parts. — Kraepelin has recorded taste-pits, with hairs or 
pegs arising from them, on the proboscis of Musca (Zeitsch. f. 
Wissens. Zool., xxxix, 1883). 

The palpi are probably not generally gustatory in function in the 
superfamily. In certain of the forms they are with little doubt prac- 
tically non-functional, and some forms have in consequence more or 
less completely lost them. In others their very considerable, some- 
times extreme development, indicates some function, which may be 
either gustatory or tactile. In certain insects they are olfactory in 
function, but probably not in the Muscoidea. They furnish charac- 
ters of not more than generic value. 

Wings. — The venational characters are in the main quite constant. 
The wings themselves are highly functional, but this does not neces- 
sarily imply that the style of venation is functional. However, as 
already pointed out, the plan of venation is so comparatively uniform 
in the superfamily that it yields no characters for separation of fam- 
ilies. The venational characters are of very great importance in 
separating this superfamily from the Anthomyioidea, but do not be- 
come again available for taxonomic use in the Muscoidea until we 
descend to tribes and genera. 

It is reasonable to attach high importance to the main features of 
the dipterous wing venation, since the wing system of Diptera is a 
very highly specialized type. The hind pair has undergone atrophy, 
its rudiments being diverted to another function, and the entire 
flight function, at least so far as propelling power goes, has been 
concentrated in the front pair. As a natural consequence of this 
high wing specialization, the venation is a practically non-functional 
system of long standing, extending over a sufficient period of time to 
allow its systemic features to become well fixed and quite constant. 
There are a few minor venational characters that can not be relied 
upon in certain restricted groups. The last section of fourth vein 
(or apical crossvein) may vary in degree of curvature, but not in 
kind. The hind crossvein may vary in strength of sinuosity, but the 
double curve is never entirely lost in the same form. 

There is some ambiguity involved in the term "apical crossvein." 
as it has been used in the past. In certain genera it is impossible to 
decide its true limits. The use of the term should therefore be re- 
stricted to those cases in which its entire course is exactly defined. 
The apical crossvein lias resulted from a bifurcation of the fourth 
vein at its point of flexure. In those genera showing what has been 
called a "stump, or a wrinkle, at bend of fourth vein," the point of 


bifurcation, and therefore the true origin of the apical crossvein, is 
apparent. In such cases the term should be used. It should not be 
used in those other genera in which no point of bifurcation is indi- 
cated, this having not arisen during their development. In such 
cases the term "fourth vein" is correctly applicable to the whole, 
whereas the term "apical crossvein" can not be so applied, especially 
in Hyalomyia, Phorantha, Alophora, Beskia, Sciasma, Stomatodexia, 
and many other genera. The latter class of genera represents a 
lesser specialization than the class showing a stump or wrinkle, and 
therefore is an older type, and indicates a more ancient assemblage 
of forms. Following Williston, the last three sections of fourth vein, 
when latter exhibits no furcation, but is more or less angularly bent, 
may very appropriately be termed the antepenultimate, penultimate, 
and ultimate. 

Halteres and tcgithc. — The sense of audition is acutely developed 
in insects, at least in the majority of the forms, as evidenced by the 
sounds they produce. There is nothing to indicate that the Mus- 
coidea are in any way an exception. Air waves which produce no 
effect whatever upon our ears doubtless register impressions upon 
the auditory nerve-ends of Diptera. Auditory organs are located 
near or at the base of the wings in Diptera, Coleoptera, Lepidoptera, 
Neuroptera, Orthoptera, and Hemiptera. They are less perfect in 
the Lepidoptera, Neuroptera, and Orthoptera, and only faintly repre- 
sented in the Hemiptera, in which four orders other chordonotal 
structures have succeeded and more or less supplanted them. In the 
Culicidse, and perhaps in other of the nemocerous groups, the an- 
tennal hairs are auditory. This has been established in the male 

In the Diptera the nerve supplied to the halter is next in size to the 
optic nerve, the latter being the largest nerve in the body. At the 
base of the halter is a number of vesicles arranged in four groups, 
to each of which groups the nerve sends a branch. These vesicles 
are perforated and contain a minute hair, and the vesicles of the 
upper groups are protected by chitinous hoods. 

Sharp (Cambridge Nat. Hist., vi, p. 448) says of the halteres: 
"They possess groups of papillae on their exterior surface, with a 
chordotonal organ inside the base. Each halter is provided with 
four muscles at the base, and can, like the wings, execute most rapid 
vibrations. Seeing that they are the homologues of wings, it is 
remarkable that in no Diptera are they replaced by wings, or by 
structures intermediate between these two kinds of organs." This 
is because they have taken on special functions. 


E. Weinland (Zeitschr. f. Wissens. Zool., li. pp. 55-166) lias con- 
cluded from his studies of the halteres that these organs are func- 
tional in determining the direction of flight. They can he used to 
steer a course in the vertical plane as well as in other directions, 
lie also concluded that the chordotonal structures in the hase of the 
halteres allow the perception of the steering movements of these 
organs. But it is highly probable that the great nerve trunk sup- 
plied to the halter is not primarily subservient to this dirigible func- 
tion, hut rather to that of audition, at least in the higher families. 
In the Nemocera the halteres may be mainly dirigible or equilibratory 
in function, since the auditor}' organs are located in the antenna?. 
In the Cyclorrhapha, however, it seems safe to assume that their 
function is primarily auditory. As Lowne suggests, the halteres are 
doubtless microphones of a most efficient nature, capable of perceiv- 
ing sound waves of such low intensity that they do not affect the 
vertebrate car. They possess a function of coordination, similar to 
that of the semicircular canals of vertebrates, and thus arc organs 
combining the functions of equilibration and audition. 

The tegulse of the Schizometopa and some other Diptera are very 
likely functional in collecting sound-waves, increasing the perceptive 
power of the chordotonal organs of the halteres, thus being ana- 
logues of the external cartilaginous ear-lobes of the mammalia. 
They also doubtless serve secondarily as a protection to the highly 
sensory halteres. It seems .safe to assume that in those dipterous 
groups having no tegular the halteres perform chiefly a function of 
equilibration, but that in those groups furnished with tegulse the hal- 
teres are mainly organs of audition. In other words, the presence of 
well-developed tegulse indicate- the presence of a highly developed 
auditory sense in the halteres. Mere protection to the latter would 
not demand such structures as the tegulse, while it can not be denied 
that they arc admirably adapted to such a function as the collection 
of sound-waves. 

Whatever may be finally determined as to their functions, it is 
certain that the habere- are highly specialized organs. The tegulse, 
without doubt accessory to them, are by inference equally functional 
and of coincident evolution with some function pertaining to them. 
The latter, therefore, can not be accepted as affording characters of 
value for the separation of large groups, but are rather of decidedly 
inferior rank in this respect to the veins of the wings. They occur 
in other groups entirely outside of and removed from the Schizo- 
phora, and even from the Cyclorrhapha. Their presence in the An- 
thomyioidea is therefore not necessarily to be construed as indicating 
a close relationship between that superfamily and. the Muscoidea. 


The Acroceridae, for instance, are to be noted as an extra-cyclor- 
rhaphous group which has developed very large tegulae, wholly con- 
cealing the halteres and probably accessory to a highly developed 
auditory sense in the latter. It seems to be chiefly groups contain- 
ing a large percentage of endoparasitic forms which are provided 
with teguke. and it is possible that a greatly increased auditory per- 
ception is necessary to these forms as an aid to them in the search 
for and ultimate detection of their hosts. 

The validity of the time-honored separation of the Calypterata and 
Acalypterata on the characters of the comparative presence or absence 
of tegulae alone may well be open to serious doubt. The unflexed 
fourth vein, which from its doubtless far greater age should be a 
nuich more valid character, would indicate a closer relationship of 
the Anthomyioidea with the Acalypterata than with the Muscoidea. 
Yet this does not appear to be the proper and natural grouping. It 
rather seems preferable to adopt Brauer's names Schizometopa and 
Holometopa as founded on characters of greater valne than either 
those afforded by relative development of tegul?e or those of wing 
venation, and to recognize therefore the Anthomyioidea as a super- 
family of the Schizometopa. While the result is mainly the same, 
the divisions become founded on valid rather than on mutable char- 
acters. The tegul?e have developed, though not uniformly, in the 
Schizometopa. They have also developed to a certain extent in 
some of the Holometopa. This fact demonstrates their unfitness for 
taxonomic use in these divisions. There is a distinction between the 
characters of a functional organ and the character of the presence 
or absence of such organ. Moreover, it may be noted that Robineau- 
Desvoidy's division Calypteratae was applied by him to the super- 
family Muscoidea of the present paper in the main sense, as is 
further brought out under the head of Synopses. 

Abdomen. — The number of abdominal sclerites should be of sub- 
family significance at least, and the form of the abdomen is almost 
invariably of generic value. 

MacrochcetcE and bristles. — Chsetophorousness in the Diptera finds 
the climax of its development in the tachinoid stock of the Muscoidea. 
While chaetophorous characters are, evolutionally, of recent origin, 
yet the arrangement of the macrochaetae of the head, thorax, abdo- 
men, and legs becomes highly important in separating tribes, genera, 
and species. The characters to be derived from the macrocha?ta? of 
the head rank even higher and serve for the separation of subfamilies 
in certain cases. In one or two groups, the Gymnosomatinae and 
Phasiina?, the peculiar chaetotactic characters of the head are cor- 
related with an absence of macrochaetae on the abdomen, while in cer- 


tain other groups, as the Hystriciinae, a different type of them is 
correlated with a true spinose development of the abdominal macro- 
chaetae. The cephalic bristles are uniformly present in the super- 
family, though sometimes weakly developed, whether the abdominal 
ones are present or absent. The same is usually true of those of the 
thorax and scutellum. The function of the macrochaetae and bristles 
of the abdomen is doubtless tactile. They are capable of movement 
in life. 

In most insects the antennae, and to a less extent the palpi, are the 
main seat of the tactile sense. The cyclorrhaphous Diptera, however, 
have the antennae so modified as to preclude this function. It is 
probable that the vibrissa? are functionally tactile, and the frontal 
and vertical bristles as well. The vibrissa? project straight out in 
front near the ends of the ptilinal suture, and naturally serve as 
anterior tactile organs for the protection of the lower portion of the 
head. Likewise the frontal bristles serve as anterior and superior 
cephalic, and the vertical bristles as superior and posterior cephalic 
tactile organs. The fact that the vertical bristles are almost invaria- 
bly stronger and longer than the frontal bristles strengthens this 
view. The inner vertical bristles correspond in development to the 

The macrochaetae of the thorax, scutellum, and abdomen serve as 
lateral and dorsal tactile organs, those of anal and preanal segments 
always being the strongest of the abdomen and those of scutellum 
the strongest of the thorax. The scutellar are doubtless the main 
dorsal tactile organs, and the anal the main posterior ones. The 
abdominal macrochaetae, when dense and of spinose character, pos- 
sibly serve also as a defense against insectivorous animals, as in 
Dcjcania, Parade jeania, Bombyliomyia, Hystricia, Hystrichodexia, 
and others. 

The macrochaetae, especially those of the abdomen, constitute the 
most recent form of specialization in the Myodaria, and are especially 
characteristic of the Muscoidea. As such, and considering further 
their probable functional character as tactile sense organs, those of 
the abdomen at least can not be expected to furnish valid characters 
for the separation of higher categories in these flies than species, 
genera, and at most tribes. 

The macrochaetae of the head, thorax, and scutellum appear to be 
of far longer standing than those of the abdomen. With the excep- 
tion of most of the (Estridae, they are present not only in all Mus- 
coidea, many of which lack abdominal macrochaetae, but also in prac- 
tically all of the Myodaria except the CEstridae already named and 
Conopidae, which two families stand well apart from the other Myo- 


daria. The bristles of certain of the thoracic plates are here used as 
main atavic characters for separating the Muscoidea from the An- 
thomyioidea, as will appear later on under Synopses, accessory sup- 
porting atavic characters being furnished by the type of venation. 

An extra-tactile function is no doubt discharged by certain of the 
cephalic bristles in the Muscoidea. The orbital bristles (middle 
fronto-orbital especially) of the females, which are usually wanting, 
or of less number, in the males, have probably arisen in those forms 
where present for the purpose of enabling the males to recognize the 
opposite sex. They are especially conspicuous in profile, when the 
strongly proclinate middle fronto-orbital are prominently contrasted 
with the reclinate upper fronto-orbital bristles. A front view would 
reveal the female in the wider front in most of the forms. The fact 
that in some forms the males as well have the orbital bristles does 
not militate against this view, but is explained by a transference of 
the female character to the male through heredity. The breast nip- 
ples of male mammals furnish an example of such hereditary trans- 
fer of a female character to the male, with absolutely no functional 

The bristles of the facialia and the frontal bristles possibly serve 
for the recognition of forms among themselves. They are most 
developed in the more inconspicuously colored forms, which run 
closely together in general habitus. Further confirmatory evidence 
is found in the fact that conspicuously colored and otherwise striking 
species often have the cephalic bristles but little developed. It is to 
be noted, however, that certain of the latter lack abdominal macro- 
chsetse as well. An absence of abdominal bristles is usually cor- 
related with a weakness of cephalic bristles, doubtless due in these 
cases to the marked development of an aerial life-habit. 

Secondary Sexual Characters. — -These should be accorded generic 
rank when they can be correlated with equally constant characters in 
the opposite sex. The secondary sexual characters in the Muscoidea 
are to be found in the comparative width of front, presence or ab- 
sence of orbital bristles, size and length of third antennal joint, some- 
times form of latter, varying degrees of holopticism or dichopticism, 
comparative length of claws, ventral carina, and certain anal pro- 
cesses of abdomen ; also often in the shade of coloration and distri- 
bution of pollen, especially on the parafrontals and parafacials, less 
often on the thorax, and sometimes in the distribution of ground 
color and even of the pollen of the abdomen. 

42 smithsonian miscellaneous collections vol. 5 1 


It seems desirable to state at the outset that the subject of tax- 
onomic divisions is approached in this paper entirely without preju- 
dice. The main lines of interest in all departments of biology lie in 
problems of descent, distribution, and bionomics, and the only de- 
sirable point as regards classification is to secure a correct delimita- 
tion of forms so that they can be definitely referred to by name. 

This paper also distinctly disclaims any attempt or intention to 
present a taxonomic system that is entirely original, likewise any 
attempt to follow any particular author or authors — in either case to 
the exclusion of any useful and valid characters already pointed out 
by previous authors. This is not intended to be a revolutionary 
scheme of classification in any sense, nor one that will upset any 
previously conceived ideas of recognized taxonomic value. Rather 
have all available characters been used that could be brought together 
for a clear definition of the various divisions in each case, those of 
value being adopted wherever they were to be found, whether old, 
recent, or newly worked out. As a matter of fact, the present paper 
is naturally based largely on Brauer's extensive and careful work. 
but the latter is not by any mean- followed blindly or undeviatingly, 
and points arc at the same time drawn from Robineau-Desvoidy and 
Rondani. Brauer based his work to a very considerable extent upon 
the work of the latter authors, and Rondani drew many valuable 
ideas from the work of Robineau-Desvoidy, whose reviser he be- 
came. As already pointed out, these three students are the ones to 
whom we owe most for our present knowledge of the Muscoidea, 
and of these Brauer naturally accomplished the most, since he en- 
joyed the greatest advantages. Any one who will conscientiously 
study this superfamily can not fail of the conviction that Brauer and 
von Bergenstamm's work, while not by any means perfect, is by far 
the best that has ever been produced on these flies. The object of 
the writer of the present paper has uniformly been to sift the entire 
subject, retaining the good, discarding the useless, and adding such 
ideas of value as it lias been possible to develop independently. 

The following tabular arrangement of taxonomic divisions is in- 
tended to convey at a glance an idea of the system of classification 
adopted : 

Order Diptera. Family Tachinidae. 

Suborder Cyclorrhapha. Subfamily Tachininae. 

Series Schizophora. Tribe Tachinini. 

Section Myodaria. Genus Tachina. 

Subsection Schizometopa. Species larvarflm. 
Superfamily Muso 


The suborder Cyclorrhapha is without doubt one of the most nat- 
ural divisions of the Diptera, and yet its line of demarcation from the 
Orthorrhapha is obscured by intermediate forms. For details on the 
limitation of the suborders of Diptera the student is referred to the 
works of Brauer, Osten-Sacken, and Williston. 

As to the limits of the series Schizophora, and the final conclu- 
sions to be reached regarding the forms which naturally belong 
within its boundaries, a word may be said with special reference to 
the Pupipara. It seems quite evident that some, at least, of the latter 
are simply degradedly specialized Schizophora. There are strong 
points of resemblance, both in structure and in reproduction, between 
Ornithomyia and Glossina. The venation is fundamentally of the 
sam$ plan. In Ornithomyia the hind crossvein has been lost. The 
apical crossvein is absent, and probably never was present. In 
Trichobius the apical crossvein is not present, but the posterior one 
is, and there is even a second posterior crossvein which has been 
developed between the fifth and sixth veins. Trichobius has lost all 
but a trace of the auxiliary vein. All the winged Pupipara show a 
venation which indicates evolution from a Myodarian prototype. 
Many of them seem quite closely allied structurally with the Myo- 
daria, and it is also to be noted that we have as yet no proof of any 
pupiparous habit in either the Streblidse or the Nycteribidse. In 
fact, it is highly improbable that such exists. Kolenati, as long ago 
as 1863, stated that the larvae of Streblidse live in bats' excrement. 
If this is true, it is probable that the Nycteribidse also have a copro- 
phagous larval habit. Mtiggenburg has investigated the morphol- 
ogy of the Nycteribidse, and asserts that they possess no trace of a 
ptilinum. On the other hand, he asserts that in Braula a ptilinum 
exists, and that the mouth parts are essentially similar to those of the 
Hippoboscidse. It is probable that the Streblidse and Nycteribidse 
are derived from an extra-myodarian cyclorrhaphous stock. Miig- 
genburg states that the Hippoboscidse and Braula are descended 
from genuine muscid stock, and that the Nycteribidse are probably 
derived from some ' other stock within the Cyclorrhapha. He 
strongly indorses Brauer's judgment of the Pupipara as being nearly 
related to the Myodaria. 

Robineau-Desvoidy's name Myodaria is adopted for the Eumyidae 
of Brauer, the Muscidse s. lat. of authors. In the sense in which 
it is here used, it includes both the CEstridse and the Conopidse. 
Both Robineau-Desvoidy and Brauer were correct in their views on 
the inclusion of these families in the section. It seems that Brauer 
did not study Robineau-Desvoidy's Bssai sufficiently to know that 
the latter author had, in T830, included the CEstridse with his Calyp- 


teratae. Brauer claimed that the idea was original with him, and 
probably arrived at his conclusions on both the CEstridae and the 
Conopidae quite independently (see Psyche, vol. 6, p. 259. The au- 
thor was unaware at that time of the above facts). Brauer claimed 
to have studied Robineau-Desvoidy.' s posthumous work exhaustively, 
and probably neglected the Essai. In the former the (Estridae are 
separated entirely from the Myodaria, which would explain the 
above oversight on Brauer's part. 

Reference has already been made to the advisability of employing 
the subsection name Schizometopa. Robineau-Desvoidy, when he 
wrote his Essai, had practically the same idea of the limits of the 
superfamily as those here arrived at quite independently. He ex- 
cluded the Anthomyiidae from his Calypteratae, which division thus 
coincides in the main with the present superfamily Muscoidea, as 
here restricted. Latreille originally applied the name Creophilae to 
these flies, and Macquart and Westwood used this name. The 
division Calypteratae of Robineau-Desvoidy was later made to in- 
clude the Anthomyiidae on account of the presence of tegulae in that 
family. As has been already pointed out, the tegulae do not afford 
characters of sufficiently high value to be applied to these divisions. 
Therefore, for several very cogent reasons, which are self-evident, it 
becomes not only advisable, but necessary, to drop both Creophilae 
and Calypterata as subsection names. The superfamily name Mus- 
coidea covers the field to which they were originally applied, and the 
name Schizometopa designates the subsection. 

The failure heretofore, chiefly on the part of Schiner and his fol- 
lowers, to properly define the grand divisions of the Myodaria, and 
especially the families of Muscoidea, has been due to the attempted 
application, in a case demanding primary, constant, and approxi- 
mately well-defined characters, of two secondary and gradating char- 
acters — namely, the presence or absence of tegulae and aristal pubes- 
cence. These two characters are unserviceable, both because they 
intergradate to such an extent as to preclude the drawing of any 
natural lines of separation, and, further, because the parts exhibiting 
them are so functional that they afford characters of only secondary 
value or less. It was inevitable that a system founded on such 
characters could not stand, for the natural boundaries do not exist 
where it was endeavored to set them. 

In the present paper the Muscoidea and Anthomyioidea are sepa- 
rated in such a manner, on atavic chaetotactic and venational char- 
acters, as to throw a few forms heretofore classed with the old 
Muscidae s. str. into the Anthomyioidea, which arrangement is be- 
lieved to represent their relationships more truly. 


It is also believed that the five families into which the Muscoidea 
are divided will ultimately be found to closely correspond in value 
with the families now recognized in the other divisions of the 
Cyclorrhapha. • 

Professor *]. H. Comstock published a very able paper in the 
Wilder Quarter Century Book, setting forth certain suggestions as 
to taxonomic work. The idea is there elaborated that, in order to 
determine the proper taxonom.ic system for a given group of insects, 
the forms should be arranged independently on each one of their 
many characters in turn, and the final mean between all these sepa- 
rate arrangements should then be determined. This mean 'would 
indicate the correct taxonomic system. It is understood, of course, 
that the varying values of the various characters should be taken 
into consideration in such a procedure. It has been the aim to pre- 
sent a systematic arrangement in this paper to agree quite closely 
with the results that might be obtained from such a final average 
between characters in this-superfamily. 

In the following synoptic treatment of taxonomic categories a plan 
is followed which has been devised and perfected by Dr. A. D. Hop- 
kins, to whom thanks are due for an exposition of it. This synoptic 
plan possesses decided advantages over any scheme of the kind yet 
devised, and is really a perfected system on the lines of that used by 
Brauer and von Bergenstamm, and some other European system- 
atists. The present synopses are carried down to families only, and 
do not exhibit the plan in detail. It will be a labor of years to per- 
fect the arrangement of the forty or more subfamilies, the numerous 
tribes, the two or three hundred American typic genera and five 
hundred or more additional atypic genera in this superfamily, to say 
nothing of the multitude of typic and atypic species. The synoptic 
plan referred to is carried out in detail by employing the following 
system of characters in turn : I, II, III, etc. ; A, B, C, etc. ; ai, a2, a3, 
etc. ; bi, b2, b$, etc. : ci, C2, C3, etc. 


Lunula absent Suborder Orthorrhapha 

Lunula present Suborder Cyclorrhapha 


Ptilinal suture absent Series Aschiza 

Ptilinal suture present Series Schizophora 


Head closely united to thorax or folding- back into dorsal groove on same. 

Section Pupipara 
Head separated from thorax by a free neck Section Myodaria 



Fmnt in both sexes of equal width, or if wider in female the greater width is 
due to a widening of the frontalia and the tegtilae are absent; tegulae never 
well developed (includes Conopidae) Subsection Holometoi'a 

Front in male narrower than in female, the wider front of female never due 
to a widening of the frontalia, tegulae never absent; if the front is not 
wider in female the tegulae are well developed. .Subsection Schizometopa 


I lypopleural and pteropleural bristles and hairs always absent, fourth longi- 
tudinal vein lying partly in the hind margin of the wing behind middle of 
extreme wing-tip, proboscis never adapted for bloodsucking, if three 
sternopleural bristles present their formula is 1:2. 

Super family AnthomyioihEa 

Either hypopleural or pteropleural bristles or hairs always present, fourth 
longitudinal vein rarely continuous with hind margin of wing behind mid- 
dle of extreme wing-tip except when proboscis is adapted for bloodsuck- 
ing,' if three sternopleural bristles present their formula is either 2:1 
or 1:1:1 Superfamily Muscoidea 

Superfamily MUSCOIDEA 

Facial plate strongly produced below vibrissal angles like the bridge of the 
nose, the produced portion convex laterally and not flattened, the vibrissae 
separated by this bulging and situated high above the oral margin; the 
mesofacial plate and epistoma completely fused into one piece. 

(Phasiid Stem) Family Phasiilu-: 
Facial plate not so produced, at most projecting nose-like below with flattened 
. or if latter is somewhat convex (Gymnosomatinae) the vibrissae 
are inserted quite near oral margin. 


Facial plate always receding below vibrissal angles and oral margin never 
prominent, thus giving the facio-peristomal profile an evenly and gently 
convex outline: vibrissal angles situated at or above the lower two-thirds 
point between oral margin and base of antennae, always very much higher 
above median oral margin than length of second antennal joint, at least 
twice as high, the mesofacial plate in consequence greatly shortened, never 
widely produced downward, if not completely cut off by vibrissal angles 
then at least very strongly constricted thereby, the peristomalia either 
approximated and forming parallel lines for a considerable distance or 
bowed outwardly and more or less widely separated so as to enclose the 
epistoma as a more or less distinct sclerite of the facial plate between 
them; antennae almost always very short. . (CEsTrid-Macronychiid Stem) 


1 Mesembrina and Eumesembrina are the only exceptions known to th< 
writer, aside from the bloodsucking forms. 


Facial plate below vibrissal angles never receding conspicuously, the oral 
margin always more or less prominent, the facio-peristomal profile in con- 
sequence never evenly and gently convex ; vibrissal angles approximated 
to oral margin and never placed much higher above its median portion 
than length, of second antennal joint, distinctly below two-thirds point of 
face, the mesofacial plate elongate, never very strongly constricted, if 
constricted at all the constriction is close to oral margin; antennae usually 

long (Tachinid-Muscid Stem) 


Vibrissas and macrochaetae absent; mouthparts wanting or rudimentary, non- 
functional Family CEstrid.e 

Vibrissas and macrochaetae present, mouthparts functional. 

Family Macronychiid.e 
Macrochaetae developed, or if not (Gymnosomatinae only) then the more or 
less red abdomen highly swollen or inflated and covered with very short, 
fine, black, bristly hairs; ovipositor never Musca-like. 

Family Tachinid.£ 

Macrochaetae not developed, or if so (Reinwardtia only) then no fronto- 
orbital bristles present and ovipositor integumental, long and Muscohlike ; 
abdomen never swollen or inflated Family Muscimj 

The series Aschiza (Becher and Brauer) includes the Phoridae, 
Pipunculidae, Platypezidae, and Syrphidae. The series Schizophora 
(Becher and Brauer) includes all the rest of the Cyclorrhapha. 

The section name Pupipara might well be replaced with Nym- 
phipara (Reaumur), which has priority. The section Myodaria 
(Robineau-Desvoidy) corresponds to the Eumyidae of Brauer, and 
to the Muscoidea of Coquillett plus the Conopidae. 

The subsection Holometopa (Brauer) includes the Malacosomse, 
Palomydae, Phytomydas, etc., of Robineau-Desvoidy, and corre- 
sponds in the main to the Acalypteratae of authors plus the Cono- 
pidae. The subsection Schizometopa (Brauer) corresponds in the 
main to the Calypteratae of authors, not of Robineau-Desvoidy. 

The super family Anthomyioidea (Townsend) corresponds to the 
Mesomydae of Robineau-Desvoidy; and to the Anthomyiden of 
Girschner minus most of the Muscinen of Girschner. The super- 
family Muscoidea (Townsend) corresponds to the Creophilae of 
Latreille, Westwood, Macquart ; to the Calypteratae of Robineau- 
Desvoidy ; to the Muscaria Schizometopa (exclusive Anthomyiidae) 
of Brauer and von Bergenstamm ; and to the Tachiniden of Girsch- 
ner plus most of the Muscinen of Girschner. 

For detailed characters defining the suborders, series, and sub- 
sections, see the works of Brauer, Becher, Williston, and Girschner. 


From the section Myodaria inclusive down to the families, and in 
some cases the subfamilies, the divisions are particularly difficult of 
exact definition, from adult characters alone, on account of the 
numerous intermediates. A study of the characters of the early 
stages is needed to determine beyond question the location of certain 
intermediate forms. 

The family Macronychiidae includes those forms approaching - the 
QEstridae in the character of the facial and peristomal development 
of the head, and which have heretofore been classed partly with the 
true tachinids and parti}" with the true dexiids. It corresponds prac- 
tically to the group Macronychiidae of Brauer and von Bergen- 
stamm, but it should be noted that Megaprosopus, and not Macro- 
11 yc hia, is the real type of the family. 

The old family Dexiidae can not be maintained. With the excep- 
tion of the few just mentioned as included in the Macronychiidae, 
its forms all fall in the Tachinidae, of which they constitute several 
subfamilies and tribes. 

Concerning the three types to be distinguished in the Muscoidea, 
it may be pointed out that the most generalized type seems to be the 
Phasiid. The primeval stock was the possessor of a Phasiid-like 
facial-plate development, in all probability, more or less after the 
Syrphoidean style. From this stock sprang the three present stems. 

Phasiid — Facial plate of the primeval type practically preserved, the meso- 
facial plate and epistoma becoming solidly anastomosed into one piece, retain- 
ing the characteristic bridge-of-the-nose production below. Both antennas and 
mouthparts, especially the latter, well developed. 

Tacli inid-M itscid — Mesofacial plate much increased and epistoma more or 
less reduced from the preceding, losing the bridge-of-the-nose production, but 
retaining a more or less prominent oral margin, the mesofacial plate gaining a 
length and width sufficient to accommodate the greatly developed antennae. 
The epistomal development is largely retained to accommodate the very func- 
tional mouthparts. 

(Bstrid-Macrony child — Mesofacial plate much reduced and epistoma (except 
in Hypodermatinse) greatly narrowed and rounded off, losing the prominent 
oral margin entirely. Antennae and mouthparts approaching atrophy from 

The following detailed notes on the connectant forms appearing 
to lie more or less between the superfamilies Muscoidea and Antho- 
myioidea will be useful for comparison with the synoptic table just 
given. The former superfamily includes the bulk of the old Mus- 
cidae, the Sarcophagidae, Dexiidae, Tachinidae, et al. (Phasiidae, Gym- 
nosomatidae, Ocypteridse, Phaniidae), and the CEstridae; the latter 
superfamily includes the Anthomyiidae as herein accepted. The 


forms which are here referred to the latter, and upon which there 
has in the past been any question as to position, are: 


Myiospila meditabunda et spp. — No hypopleural nor pteropleural bristles or 
hairs. Sternopleural bristles 1. 0. 2. A weaker sternopleural bristle below first 
one, so as to appear 2. 0. 2. Venation like Stomoxys. 

Muscina stabulans et spp. — No hypopleural nor pteropleural bristles or hairs. 
Sternopleural bristles 1. 0. 2. Venation like Stomoxys. 

Muscina ccesia (det. Coquillett).— No hypopleural nor pteropleural bristles 
or hairs. Sternopleural bristles 1. 0. 2. Venation typical anthomyiid. 

Cyrtoneura podagrica, gluta, et spp. 

Pararicia pascuorum et spp. 

Clinopera frontina et spp. — No hypopleural nor pteropleural bristles or hairs. 
Sternopleural bristles 1. o. 2. 

These forms have heretofore been classed in the old Muscidse s. str. by 
Williston, van der Wulp, and Brauer. It is believed they should be excluded 
from the Muscoidea on the general averages of their characters. 


The following forms here included in the Muscoidea were referred 
"by Girschner to his Antinomy iden : 

Musca domestic®, corvina, et spp. — No hypopleural hairs. Distinct ptero- 
pleural hairs. Sternopleural bristles 1. o. 2, but in some the last two bristles 
are separated so as to appear almost like 1. 1. 1. Typical Muscoidean venation. 

Stomoxys calcitrans et spp. — Hypopleural hairs, also pteropleural hairs. 
Sternopleural bristles 0. o. 1. Fourth vein bent, arcuate, partly continuous with 
hind border. Proboscis adapted for bloodsucking. 

Lyperosia irritans et spp. — No hypopleural bristles. Pteropleural hairs pres- 
ent. Sternopleural bristles none, o. 0. 0. Aberrant venation ; fourth vein 
hardly bent, yet apical cell narrowly open at wing-tip; third vein bulged up- 
ward, convex in front or above. Proboscis adapted for bloodsucking. 

HcEinatobia stimulans et spp. — -No specimens for study. 

Graphomyia maculata, americana (det. Coquillett), et spp. — Hypopleural 
hairs present. No pteropleural bristles or hairs. Sternopleural bristles o. o. 2. 
Fourth vein arcuate at bend. 

Synthesiomyia brasiliana et spp. — Hypopleural hairs strong, quite bristly 
No pteropleural hairs. Sternopleural bristles I. 0. 2. Fourth vein arcuate at 

Glossina longipalpis et spp. — No hypopleural hairs or bristles. Distinct black 
pteropleural bristles, with yellowish hairs also. Sternopleural bristles r. o. 2. 
Venation aberrant, in CEstrid direction ; apical crossvein continuous with pos- 
terior crossvein, fourth vein deeply arcuate before small crossvein so that 
latter appears continuous with the section of fourth vein following it. 

Morellia violacea (det. Coquillett, Brazil), micans Macquart (det. Coquillett, 
Maine), et spp. — No hypopleural hairs. Pteropleural hairs present, bristly 
hairs in micans. Sternopleural bristles I. 0. 2. Fourth vein arcuate at bend. 


Mesembrma mystacea et spp. — No hypopleural hairs or pile. Pteropleural 
black pile present. Sternopleural bristles I. O. I, but often hard to distinguish 
from the black hairs or pile. Venation like Stomoxys, also like Myiospila, 
fourth vein partly continuous with hind border. This and the five following 
genera have the inner side of middle tibiae furnished with one or more strong 

Metamesembrina (gen. now ) meridiana Linne (det. Brauer and von Bergen- 
stamm, Alaska). — No hypopleural hairs or bristles. Pteropleural bristly hairs 
present. Sternopleural bristles o. o. I. Fourth vein reaching front margin of 
wing before tip and arcuate at bend. 

Bumesembrina (gen. nov.) latreillei Robineau-Desvoidy, et spp. — No hypo- 
pleural hairs. Pteropleural hairs present. Sternopleural bristles I. o. 2. 
Venation as in Mesembrina, but fourth vein more continuous with hind margin. 

Dasyphora pratorum et spp. — No specimens. Venation of Lucilia (ace. 
Brauer and von Bergenstamm ). 

Pyrellia cadaverina ( 1 spm. det. Brauer and von Bergenstamm), serena 
Meigen (det. Coquillett), et spp. — No hypopleural hairs. Pteropleural hairs 
present, bristly and short in cadaverina. Sternopleural bristles 1. o. 3 (some- 
times 4) ; in the single specimen of cada/verina 1. o. 2 on one side and 1. 0. 4 
on the other, but probably normally I. o. 3. Fourth vein arcuate at bend. 

Pseudopyrellia comicina et spp. — No hypopleural hairs. Pteropleural hairs 
present. Sternopleural bristles 1. 0. 2, but the hind pair with anterior bristle 
placed nearly as high as the posterior one. Fourth vein arcuate at bend. 

Phasiophana obsoleta et spp. 

Cyrtpneura sp. (det. Brauer, N. C. and Cala.). — No hypopleural hairs 
Pteropleural bristles present. Sternopleural bristles 1. 0. 2. Fourth vein 
arcuate at bend, apical cell narrowly open. Morellia micans (det. Coquillett) 
and hortorum have nearly these characters, and it is likely that the present 
North Carolina and California specimens belong to Morellia. 

Auchmeromyia spp. — This genus evidently belongs here. It probably has 
either hypopleural or pteropleural hairs or bristles. 

Ochromyia jejuna ]. C. Fabricius (N. W. India) et sp. (Amboyna). — 
Hypopleural bristles present. No pteropleural bristles, but yellowish pile 
present. Sternopleural bristles 1. 0. 1. Venation typical. 

It will be at once seen from the above notes that the characters of 
the presence of one or other or absence of both the hypopleural 
and pteropleural bristles or hairs are the final determining test in 
the separation of the two superfamilies. 

Metamesembrina, Graphomyia, and Synthesiomyia do not have 
the fourth vein continuous in any part of its extent with the hind 
margin of wing, but all show a more or less distinct posterior inclina- 
tion of fourth vein where it joins the wing margin, this being less 
distinct in Synthesiomyia. The genera with this venation might be 
considered by some students to form an aberrant group of the 
Anthomyioidea, exhibiting a transition toward the Muscoidean type 
of venation; but, considered from all points of view, their relation- 
ships are mainly with the Muscoidea. 


Syiithcsiomyia has strong hypopleural hairs, which can hardly be 
considered true bristles, yet they serve as a character of equal value. 
It has also a bare arista. It lacks the pteropleural hairs and bristles. 

Musca, Glossina, PseudopyreUia, Pyrellia, Morellia, and Dasy- 
phora ( ?) have the Muscoidean type of venation strongly marked 
(except Pyrellia), but possess no hypopleural bristles. Glossina and 
Musca, however, possess distinct pteropleural bristles like the other 
Muscoidea, while PseudopyreUia, Pyrellia, Morellia, and Dasy- 
phora (?) possess a tuft of more or less bristly hairs in their place, 
directly beneath the wing bases. Morellia hortorum has ptero- 
pleural bristles approaching those of Glossina and Musca in strength, 
and is doubtless not a true Morellia, which has only a tuft of ptero- 
pleural hairs. All these genera are more or less intermediate, but 
they can be distinguished by the above characters. 

Some doubt may arise with Myiospila, etc., which belong in the 
Anthomyioidea. They have neither hypopleural nor pteropleural 
hairs, which will always distinguish them, and it may be seen that 
the fourth vein is continuous with wing margin behind the middle 
point of the rather widened apex of wing. 

In connection with the characters given for the Muscoidea in the 
table, it is to be noted that the fourth vein is incomplete in certain 
genera, as Roeselia, Pliytouiyptcra, Thrixiou, Gastrophilus, Sylle- 
goptera, Euryccromyia, Dichcetoucura, etc. 

Finally it may be pointed out that certain species of the old genus 
Cyrtoneura, referred to Pararicia by Brauer and von Bergenstamm, 
and belonging to the Anthomyioidea, show the gentle removal of the 
fourth vein from the wing margin which is characteristic of the 
forms whose position has been heretofore misunderstood. These 
forms were considered by some authors as belonging to the old 
Muscidae s. str., and by others as belonging to the Anthomyiidae, but 
the characters pointed out by Girschner serve to reveal their true 
position. They are distinctly to be considered as a genealogical 
group descended from forms with a wholly straight (as far as wing 
margin) fourth vein. The extensive removal of the fourth vein 
from the wing margin in Pyrellia, Mesembrina, et al. must be con- 
sidered as a further step in the development of the venation toward 
the Muscoidean type. The Muscoidea are without question more 
specialized than the Anthomyioidea; and since the form normal in 
the latter exhibits the type of venation universal in the Holometopa 
(excepting the Conopidae), the last named subsection is less special- 
ized than the Schizometopa. The Conopidae stand evidently to one 
side as a large group rather closely related to both the Schizometopa 
and the Holometopa, but with a preponderance of affinities for the 


latter. They doubtless represent a branch which sprung from the 
proto-Myodarian stem during its period of multiform development. 
They should be considered as one of the primary divisions of the 
Holometopa, probably equal in taxonomic rank to all of the other 
Holometopa taken together. They stand in practically the same 
relation to the Holometopa as do the CEstridae to the Schizometopa, 
the QEstridse also being a group apart from the other Schizometopa 
and of older origin. Moreover, the CEstridae is a polyphyletic 
group showing affinities with various subfamilies and tribes of Mus- 
coidea, but owing to its present preponderance of characters due to 
mode of life it is best treated as a family. For similar reasons the 
Conopidae are also best treated as of family rank. 

While on the subject of the relationships and extreme specializa- 
tion of the Schizophora in general and the Muscoidea in particular, 
it becomes highly significant to note that the Muscoidean stock has 
originated three separate and distinct types of parasitism on mam- 
mals, all having the same end in view— that of nourishing their 
larvae at the expense of Mammalia — but each of the three attaining 
this result in radically opposite ways. 

Cutcrcbra and its allies attain this end by their well-known sub- 
cutaneous larval endoparasitism, in which the larva does all the 
feeding, the imago taking no nourishment whatever, this peculiarity 
being developed even to the extent of the adult mouthparts having 
become atrophied and nonfunctional. 

Glossina secures the same result by a supracutaneous imaginal 
ectoparasitism, in which the adult does all the feeding, by actual 
mechanical blood-letting, and retains and nourishes the larva within 
the oviduct until it is fully grown, when it is extruded and becomes 
a pupa almost immediately and absolutely without feeding. This is 
the exact antithesis of the preceding. 

But the Muscoidea must be credited with developing yet a third, 
and still more remarkable method, because wholly unique and unpar- 
alleled among dipterous larvae of this description, of living at the 
expense of mammals. Auchmeromyia produces a bloodsucking 
larva, and thus furnishes a case of supracutaneous larval ectoparasit- 
ism, since the larva sucks blood externally by mechanical means. 
This is the so-called Congo floor-maggot, which has recently 
attracted some attention in the literature. It possesses an extended 
range on the West African coast and has also been reported from 
Uganda. The maggot-like larva pierces the skin of sleeping per- 
sons with its small but sharp jaws, and sucks their blood. It is an 
unique habit, because the larva is a footless maggot with extremely 
small jaws and no means of attaching itself to the skin of its host 


other than by its mouthparts. It can not cling during the act* of 
piercing by any structure except its mouth-hooklets. The acquire- 
ment of such a habit has been possible through the fact that the na- 
tives of the region inhabited by it have from time immemorial slept 
on mats spread upon the earthen floors of their dwellings. The 
larvae probably originally fed on fermenting juices and liquids, as 
evidenced by the fact that they are especially common beneath the 
urine-stained mats which have been occupied by sleeping children. 
The flies are attracted by sour-smelling liquids, and doubtless ovi- 
posit beneath the sleeping-mats of young children. 

The peculiar mode of reproduction of Glossina is carried even 
farther by the Hippoboscid genera of mammal ectoparasites (Lipop- 
tena, Melophagus, Hippobosca, Ortholfersia). The larva in these 
forms is retained and nourished within the oviduct of the female 
until full grown, but upon being extruded is incapable of movement. 
The Glossina larva upon extrusion is capable of only sufficient move- 
ment to find a suitable place for pupation, whereupon its integument 
undergoes chitinization to form the pupal envelope. The Hippo- 
boscid larva upon extrusion forthwith undergoes this process of ex- 
ternal chitinization. The Hippoboscid female therefore extrudes 
the larva in a situation and position suitable for it to remain during 
its pupal period. It is thus evident that some relationship exists 
between Glossina and the Hippoboscidse, doubtless to the extent of a 
not very remote common origin. The Hippoboscidae are probably an 
offshoot from the old muscid stock on the one hand, and the CEs- 
tridae are likely an earlier offshoot in a quite opposite direction from 
several stems of the same stock. 

The QEstrid habit of parasitism seems the oldest, the Glossina and 
Hippoboscid habit next, while the Auchmeromyia mode is evidently 
verv recent. 



But little need be said in preface to the following descriptions of 
genera and species. In addition to the treatment of new forms, 
there is given considerable supplementary descriptive matter on 
forms already described. 

As a basis of operations in determining the North American Mus- 
coidea, the recent Smithsonian Catalogue of North American Dip- 
tera, by Professor Aldrich, will be found quite indispensable. Its 
value lies in its references to descriptions. It will be necessary to 
use it with much caution so far as the synonymy is concerned. It 
should also be pointed out that the sequence of genera there em- 
ployed is unnatural and misleading. This is not the fault of the 
cataloguer, but is due to the present unsatisfactory state of the litera- 
ture of the North American forms. 

The sequence of subfamilies and tribes here adopted is as nearly a 
natural one as is possible of attainment in the present state of our 
knowledge. No doubt further study will modify this arrangement 
in certain details. 

It is to be noted that the tribes which appear in center heads are 
independent of the subfamilies preceding them, except those in 
italics under the families Muscidae and Phasiida?. 


Tribe Trixodini 

Genus Trixodes Coquillett 

Trixodes Coquillett clearly exhibits in its weakly developed mouth- 
parts, peculiar facial plate, and weak macroehaeta; a close affinity 
with the CEstridse. The type species is obesa Coquillett, described 
from a specimen collected by the writer in the Sierra Madre of Chi- 
huahua. A second specimen was collected by the writer on the 
West Fork of the Gila, in New Mexico. 

Subfamily Megaprosopix.k 
Genus Microphthalma Egger. 

Microphthalmos trifasciata Say. — Tachina disjuncta Wiedemann 
may be a small specimen of this species. 

The genus Microphthalma is distinct from Dexiosoma, from which 
it differs in its relativelv small eves, almost bare and much shortened 


arista, more compressed third antennal joint, and almost bare para- 
facials. The antennae are inserted on eye middle. 

M. michiganensis Townsend is a large northern form, distinct 
from trifasciata or disjuncta in its red face and cheeks, third anten- 
nal joint hardly longer than second, facial profile more flattened, sil- 
very pollen of abdominal segments general and not restricted into 
basal fasciae. 

Tribe Neophytoini 

NEOPHYTO, gen. nov. 

The genus is like Mcgaprosopus in the formation of the facial 
plate, epistoma, and facial ridges, but the vibrissas are distinct from 
the peristomal bristles below them, and the parafacials have an 
oblique well-marked row of thinly set bristles (not thickly placed as 
in Macronychid) . Frontal bristles not strong. Peristomalia quite 
closely approximated. Cheeks more than one-half as wide as eye 
height, sometimes appearing almost as wide in female. Front prom- 
inent, facial profile strongly receding and slightly convex. Antennae 
inserted distinctly below middle of eyes. Apical cell closed in mar- 
gin considerably before wing tip ; fourth vein bent at angle, without 
stump but with slight wrinkle, hind crossvein in middle between 
small and apical crossveins. Male without, female with two middle 
fronto-orbital bristles. Type, Phyto sctosa Coquillett. 

Neophyto anomala, sp. nov. 

Syn. Phyto clesidcs Coquillett (non Walker). 

Length, 6 to 8 mm. Grayish cinereous. Facial plate narrow, 
oval, acute below, the vibrissal angles but little more approximated 
than the peristomalia below them, the facial profile strongly convex. 
Face, parafacials and parafrontals silvery ; palpi, cheeks, frontalia 
and antennae light reddish brown, third antennal joint brown. An- 
tennae very short, third joint no longer than second, second about 
three times as long as the very short first joint. Male front much 
narrower than eyes, female front wider than eyes. Male cheeks 
two-thirds eye height in width, female cheeks fully equal to eye 
height. Mesoscutum cinereous, with three dusky vittae in male, 
almost obsolete in female. Abdomen dusky cinereous, with anterior 
portions of second and third segments and most of anal segment 
silvery-cinereous. In the female especially the dusky portion is 
more variable, appearing in some only on narrow hind margins of 
second and third segments. Discal macrochaetae on all the segments 
except the first. Male abdomen long-subconical, female abdomen 


oval and flattened. Legs blackish. Wings clear, veins brown. A 
strong costal spine present, apical cell sometimes extremely short 
petiolate. Tegulae white. 

Missouri to Louisiana. 

Type.— Cat. No. 11,646, U. S. X. M. (Missouri, Riley Coll.). 


Tribe Miltogrammini 

Genus Senotainia Macquart 

Senotainia rubriventris Macquart. — The writer retains Senotainia, 
of which this species is the type, as distinct from Miltogramma in 
having a more evenly rounded facio- frontal profile, narrower cheeks, 
bare parafacials, distinct vibrissa?, and longer antennae. 

Miltogramma and Senotainia, with certain other forms yet to be 
described, constitute a tribe by themselves. The writer can not fol- 
low Brauer and von Bergenstamm in grouping Metopia, Araba, 
HilarcUa, etc., with them. The latter genera have the vibrissal 
angles close to oral margin. 

Tribe Myiophasuxi 
Genus Myiophasia Brauer and von Bergenstamm 

Myiophasia of Brauer and von Bergenstamm has the eyes bare ; 
cheeks in female more than one-third eye height in width, in male 
scarcely one-fourth eye height ; both sexes with short but strong 
claws, the front claws of male being the only ones that are somewhat 
longer than last tarsal joint; no macrochsetse on first and second 
abdominal segments ; arista thickened only at extreme base, second 
joint short. 

It is hardly possible that the Uruguayan and United States forms 
that have been referred to this species are identical. Several other 
well-marked forms have been confused here. M. cenea has the 
apical cell distinctly though narrowly open 

Myiophasia setigera, sp. nov. 

Differs from M. cenea in having a median marginal pair of macro- 
chsetae on second abdominal segment in both sexes. Male with rows 
of hairs on parafacials, female with same rows somewhat less de- 

Texas, New Mexico, Nevada, Oregon. 

Type.— Cat. No. 11,647, U. S. N. M. (Male, Beulah, New Mex- 
ico, 8,000 feet, August, Cockerell.) 


A female specimen received from the Cotton Boll Weevil Labora- 
tory (Hunter) was collected on an acorn of Quercus alba at Ruston,. 
Louisiana, October 31, and was apparently ovipositing on a weevil 
larva within. 

A female specimen from New Mexico (Santa Fe, Cockerell) has 
a pair of small macrochaetae on anterior border of second and third 
abdominal segments, and a submarginal posterior pair on third seg- 
ment. It may be a distinct form. 

These forms are placed in Myiophasia tentatively, and may need 
to be removed on further study. 

Genus Phasioclista Townsend 

The genus Phasioclista Townsend also has the eyes bare, but 
the cheeks are almost or quite one-half eye height in width in 
both sexes ; male claws long, all being distinctly longer than last 
tarsal joint; female claws very short; arista bulbous at base, indis- 
tinctly jointed; first and second abdominal segments without macro- 
chaetae, apical cell closed or sometimes very narrowly open, hind 
crossvein nearly straight. 

Myiophasia differs from Phasioclista in having a loosely set, 
oblique, fringe-like row of bristly hairs on parafacials, in addition to 
the shorter irregularly arranged hairs above them ; the cheeks are 
not so wide, as above pointed out, a double costal spine is present, 
and the antennae reach almost to insertion of vibrissse. 

Whether the specimens with apical cell open and closed represent 
different forms of Phasioclista is still a question, but the fact is re- 
corded in Psyche (June, 1893, p. 467) that specimens bred from dif- 
ferent hosts differed in this character. A specimen bred from 
Leucania unipuncta had the apical cell open, and another bred from 
Sphowphorus parvulus had same closed. The radical difference 
between a parasitic habit involving a lepidopterous larval host with 
soft skin, and one affecting an adult coleopterous host, would easily 
imply the distinctness of these forms. 

Phasioclista metallica Townsend. — Both sexes have perfectly bare 
eyes. Female with more or less suggestion of pollen on mesoscutum 
in front. No macrochaetae on first two abdominal segments. Male 
with rows of hairs on parafacials, female practically without. 

Florida, Georgia. 

Genus Ennyomma Townsend 

Bnnyomma, at least in the male, has the eyes thickly pubes- 
cent ; arista distinctly three-jointed, not so bulbous at base as in 
Phasioclista; second abdominal segment with marginal macrochaetae; 


apical cell open, sometimes narrowly so; hind crossvein strongly 
sinuate. The genus may be at once distinguished from both Myio- 
phasia and Phasioclista by its thickly hairy eyes. J\I. robusta Co- 
quillett belongs to Bnnyomma. 

Bnnyomma robusta Coquillett. — Eyes thickly pubescent (at least 
in male). Last two abdominal segments and anterior border of sec- 
ond segment thickly pollinose. Large species. 

California, Mexico. 

Bnnyomma globosa Townsend. — Eyes thickly pubescent in male, 
bare in female. Male with purplish shining mesoscutum, showing 
no pollen. Female showing pollen at least anteriorly and on humeri. 
Small species. The species was described in the male only, and 
referred to Loewia. Numerous male specimens agree perfectly 
with the description. The female is without macrochastae on first 
two abdominal segments, the male having them as in the description. 

White Mountains, New Hampshire ; Maryland, Georgia, Florida, 
Louisiana, Missouri, Sierra Madre of Chihuahua, Mexico City, 

Two specimens, male and female, bred from Anthonomus grandis, 
Alexandria, Louisiana (Hunter, No. 1326, W. 6). 

Tribe EumEGApariini 
EUMEGAPARIA, gen. nov. 

This genus may be considered intermediate between Megaparia 
and Dexia, but must be classed with the Tachinidae in the neighbor- 
hood of the Dexiinas. The oral margin is only slightly prominent 
and the facio-peristomal profile approaches that of the Megaproso- 
pinse, but the oral margin is nevertheless sufficiently prominent to 
destroy the evenly convex outline characteristic of the Megaproso- 
pine profile. The antennae are short and the mouthparts much re- 
duced, the proboscis being very short. The mesofacial plate, how- 
ever, is of good width and length ; the vibrissal angles are widely 
separated and only feebly convergent, about as high above oral mar- 
gin as length of second antennal joint. Ptilinal suture terminating 
well above vibrissal angles. Claws of male very long. Type, Meg- 
aparia ftaveola Coquillett (No. 6236, U. S. N. M.), Colorado. 

Subfamily Dexiin.e 
Genus Ptilodexia Brauer and von Bergenstamm 

Clinoneura and Ptilodexia. — Ptilodexia has parafacials hairy, 
more than one pair of discal macrochaetae on middle abdominal seg- 


ments, and male claws very long. Clinoneura has parafacials bare, 
only one pair of discal macrochaetae on middle abdominal segments. 
The species described by Robineau-Desvoidy as Bstlieria tibialis is 
neither a Ptilodexia nor a Clinoneura, since it has the apical cell 

Ptilodexia has cheeks (male) about, or slightly over, one-half eye 
height ; antennae inserted low, so as to give a long frontal profile ; 
vibrissas inserted high above oral margin ; no strong or other recli- 
nate vertical bristles; second antennal joint elongate and third 

DOLICHOCODIA, gen. nov. 

Near Myiocera, from which it differs as follows : Head conspic- 
uously elongated anteriorly, apical cell open. xA.ntennse inserted on 
or above middle of eyes ; proboscis slender and horny, with long fili- 
form palpi which are but slightly thickened apically and bear very 
long bristles ; parafacials wider ; long axis of head at antennal inser- 
tion fully equal to that at epistoma ; head longer than high. Type, 
Myiocera bivittata Coquillett, described from specimens collected by 
the writer on the Rio Ruidoso, in the White Mountains of New 

EUCH^TOGYNE, gen. nov. 

Like Chcetogyne, but proboscis rather stout and only a little longer 
than head height ; hind tibiae completely ciliate on outer edge, with no 
bristles among the cilia. It agrees with Chcetogyne in having the 
carina wide, flattened on its edge and conspicuously furrowed on 
median line. Type, Hystrichodexia roederi Williston (Kansas Univ. 
Quarterly, 11, pp. 77-78), described from Arizona (1 male). For 
purposes of comparison, the following characters are given for cer- 
tain allied genera : 

Hystrichodexia has proboscis shorter than head height. 

Paraprosena has carina narrow and thin. 

Chcetogyne has proboscis very long and slender, hind tibiae with 
long macrochaetae among the cilia. 

Phorostoma has only a weak rudimentary facial carina. 

Buchcetogyne roederi Williston. — Three males in U. S. N. M. ; 
two collected by the writer in Meadow Valley, Sierra Madre of 
western Chihuahua, head of Rio Piedras Verdes, about 7,300 feet, 
August 30 and September 2 ; and one labeled "Mexico, 400, Phoro- 

Williston says in his description : "Third, fourth, and fifth seg- 
ments opaque golden yellow." The so-called fifth segment shows 


very narrowly, being the base of the hypopygium. It is in reality 
the sixth segment, since there is a very abbreviated basal segment 
present. What appears to be the fifth segment is only the portion 
of the fourth behind the transverse row of submarginal macro- 
chaetae. The scutellum shows practically no yellowish on apex. 
The third segment (called second heretofore) has, in addition to the 
six approximated macrochaetae on hind border in middle, three or 
four (usually four) approximated lateral ones on each side. The 
second segment (so-called first) has one lateral macrochaeta on each 
side. The apical decussate pair of scutellar macrochaetae is quite as 
strong and long as any of the others of scutellum. The narrow 
linear yellow of hind margin of third segment is continued in a 
slight anterior prolongation on the median line in the two Sierra 
Madre specimens. In addition to the two large silvery spots of 
third segment of venter, there are two smaller ones on the second 
and fourth ventral segments in the above specimens. 

Genus Myxodexia Brauer and von Bergenstamm 

Syn. Tfopidomyia Brauer & vox Bergenstamm (preocc). 

Neotropidomyia Townsend, nom. nov. (Dec, 1891), Trans. Am. Ent. 
• Soc, xviii, p. 382. 

The type of this genus is M. macr onychia Brauer and von Ber- 
genstamm, of Syria. 

Subfamily Trixin/e 

EUCLYTIA, gen. nov. 

This genus is herewith proposed for the species Clytia flava 
Townsend (Tr. Am. Ent. Soc, xvin, pp. 372-373). It may be 
known by the two rows of weak frontal bristles on each side of 
frontalia. the outer row weaker and somewhat irregular. The epis- 
toma is but slightly prominent. Specimens in U. S. N. M. have 
been referred by Brauer and von Bergenstamm to Redtenbacheria, 
but the species certainly can not be included in that genus. 

It is distinct from the old genus Clytia, now to be known as 
Clytiomyia, of which the European C. helvola is to be taken as the 
type. Clistomorpha also is a very different genus. Both Clisto- 
morpha and Clytiomyia belong in the Phasiidae. 

Tribe P h asiopteryg i n i 

Genus Phasiopteryx Brauer and von Bergenstamm 

Phasiopteryx bilimeki Brauer and von Bergenstamm. — The re- 
marks on this species in Ann. and Mag. Nat. Hist., xix. pp. 33-34, 


indicate differences between Phasiopteryx and Neoptera, the signifi- 
cance of which did not appeal to the writer at the time. It seems 
quite certain that several forms are confused here. The specimens 
that the writer has seen of related forms in the CEstrophasiinae in- 
cline him to the belief that large series of material will demonstrate 
the distinctness of Neoptera and Phasiopteryx. It must be remem- 
bered that only a fraction of the neotropical fauna is yet known. 

Besides the differences, pointed out below, between CEstrophasia 
and Phasiopteryx, the following may also be noted : CEstrophasia 
and Ccnosoma have the facial plate flat or subcarinate ; antennae in- 
serted distinctly below middle of extreme head height, almost as 
low as lower margin of eyes ; arista very short and bare, and third 
antennal joint only about as long as second. Phasiopteryx has the 
facial plate more strongly, often quite strongly, carinate ; antennae 
inserted but little below middle of eyes, distinctly above middle of 
extreme head height ; arista very long, very distinctly but finely and 
thinly hairy (looks bare in some specimens, apparently from the fine 
hairs being lost or rubbed off), and the third antennal joint always 
twice as long as second. 

Subfamily CEstrophasiix.e 
Genus CEstrophasia Brauer and von Bergenstamm 

CEstrophasia clausa Brauer and von Bergenstamm. — This is a 
northern species. The specimens from Cuautla, Mexico, referred 
here bv Giglio-Tos, doubtless represent another form. Cuautla is 
thoroughly tropical, and clausa is a transition and boreal form. 

The ultimate section of fourth vein in Cenosoma signifera and 
calva is normally rather deeply bowed in, but not so in CE. setosa 
and clausa, both of which have the apical cell very short petiolate, 
while setosa has third vein bristly nearly to small crossvein. 

The antennae of CEstrophasia and Cenosoma are widely separated 
by a characteristic median enlargement of the lunula in both sexes 
of all the species. This is absent in Phasiopteryx, which has the 
antennae closely approximated. 

Genus Euoestrophasia Townsend 

Eua-strophasia aperta Brauer and von Bergenstamm. — This South 
American form seems generically distinct from the species of CEs- 
trophasia in its open first posterior cell, as pointed out in Trans. Am. 
Ent. Soc. xix (1892), p. 133. 


Genus Cenosoma van der Wulp 

Ccnosoma signifera van der Wulp. — It is likely that this tropical 
species will prove generically distinct from both Glstrophasia and 
Buastrophasia, when sufficient material is studied. Two species are 
probably confused in the catalogue under the name of signifera. 
The Canadian and New England specimens are probably a northern 
form distinct from the tropical one. CE. calva may be considered 
congeneric with signifera. 

Subfamily Paramacronychiin.e 
Genus Pachyophthalmus Brauer and von Bergenstamm 

Pachyophthalmus aurifrons Townsend.- — This species is quite 
distinct from the European signatus Meigen, which probably does 
not occur in America. It differs from signatus in the golden pol- 
linose sides of front and face, third antennal joint about the length 
of second, hind crossvein very slightly bowed, front quite strongly 
produced, etc. P. signatus has pollen of front and face silvery 
white with blackish reflections but without golden, third antennal 
joint about twice as long as second, hind crossvein strongly bowed, 
front scarcely protruded, etc. Both aurifrons Townsend and flori- 
deusis Townsend are best assigned to this genus. 

Genus Sarcomacronychia Townsend 

Sarcomacf onychia unica Townsend. — This species, S. sarcopha- 
goides, and S. trypoxylonis are to be considered as three valid 
forms. The genus Pachyophthalmus differs from Sarcomacrony- 
chia in having the ptilinal area wider in comparison with parafacials, 
being three-fifths to almost three-fourths width of face ; cheeks as 
wide as one-sixth to one-eighth eye height, or less ; eyes descending 
but little lower than vibrissas, as seen in profile. Sarcomacronychia 
has facial plate very small and restricted, being two-fifths to one- 
third width of face, parafacials proportionately wider, often nearly 
as wide as facial plate itself, but sometimes appearing narrow in 
profile; width of cheeks from little less than one-fourth to about 
one-fifth eye height ; eyes descending far below vibrissas, and even 
below epistoma, nearly as low as lateral oral margins, as seen in 
profile. Pachyophthalmus has the vibrissas inserted but little above 
epistoma, and the antennae are inserted below middle of eyes. Sar- 
comacronychia has vibrissa? inserted much farther above epistoma, 
and the antennae are inserted on eve middle. 


Tribe Melanophryonini 
Genus Atropharista Townsend 

The affinities of Mclanophrys and Atropharista are uncertain. 
One can hardly agree with Brauer and von Bergenstamm's refer- 
ence of them to the Paramacronvchiinse. They seem rather to be- 
long in the Tachinidae. 

Atropharista jurinoidcs Townsend. — The writer has previously 
considered this genus synonymous with Melanophrys, but it appears 
after all to be distinct. Melanophrys has the second antennal joint 
short, the third joint being three to five times as long as second, 
according to sex. Atropharista has second antennal joint elongate, 
the third joint same length or a little longer, probably never twice 
as long even in the male. 

The species jurinoidcs is distinct from Walker's Tachina insolita, 
if any reliance is to be placed on the description of the latter. T. 
insolita is described as having the third antennal joint fully twice as 
long as second, third aristal joint very stout, and a white oblique 
stripe on each side of head, presumably (from the connection) oppo- 
site the antenna?. The last character agrees with Mel. Havipcnnis, 
but the other characters only partially agree. None of them seems 
to agree with Atropharista, as the second antennal joint does not 
appear to be elongate in insolita. 

A. jurinoidcs differs from both in having a broad, elongate silvery 
crescent bordering the orbit on each side of the head, partly on the 
parafrontals and partly on the parafacials. It is quite certain that 
the elongate second antennal joint will prove Atropharista to be a 
valid genus, as genera will ultimately go in this superfamily. 

Subfamily Phytoix.e 
EUPHYTO, gen. nov. 

Differs from Phyto (Robineau-Desvoidy) Brauer and von Ber- 
genstamm in having parafacials absolutely naked, tegulae small and 
rounded, cheeks not widened posteriorly, apical cell quite long petio- 
late, hind crossvein in middle between small crossvein and bend of 
fourth vein, and no discal macrochsetas on abdomen. 

Differs from Stcvcnia Robineau-Desvoidy in parafacials being 
wide, same width above and below, their width being equal to that of 
cheeks, which are over one-third eye height. Cheeks bare, same as 

Type, Leucostoma subopaca Coquillett. 


Tribe Metopiini 
Genus Hilarella Rondani 

The cheeks in this genus are about one-fourth eye height, para- 
facials with a row of bristles to lower eye margin, arista pubescent 
or hairy, front sharply produced in profile. Opsidia is much closer 
to Hilarella than is Bumacronychia. 

Tri be Hi m acri > n vc h 1 1 x 1 
Genus Eumacronychia Townsend 

Bumacronychia decens Townsend. — This species is the type of 
the genus, which has cheeks about one-half eye height in width. 
parafacials bare of bristles, frontal bristles stopping at base of an- 

Genus Gymnoprosopa Townsend 

Gymnoprosopa polita, argentifrons, and clarifrons are perfectly 
distinct, in spite of the note in the catalogue. They may be recog- 
nized by the descriptions. 

SPHENOMETOPA, gen. nov. 

This genus is proposed for Araba nebulosa Coquillett. The speci- 
mens from which this species was described were collected in the 
vicinity of Meadow Valley, six or eight miles south of Colonia Gar- 
cia, in the Sierra Madre of western Chihuahua, on the head of the 
Rio Piedras Verdes, in the pine zone, about 7,000 to 7,500 feet 
(Townsend). The form is not referable to Araba. 

The genus may be known by the front being conspicuously nar- 
rowed anteriorly, the parafacials Aery narrow and bare, the vibrissa 
quite distinct, and the front not produced conically like Mctopia and 
Araba. It comes near to Metopodia in the latter character. The 
wings are slightly clouded on the veins. 

Subfamily PsEUDODEXHN^ 

EUCALODEXIA, gen. nov. 

This genus is proposed for Homode.ria flaz'ipcs Bigot. It may be 
recognized from the characters pointed out by Brauer (Sitzungsber. 
Kais. Akad. Wiss., evil, i. p. 515), who failed to give it a name. 

Genus Atrophopoda Townsend and allies 
The following new genera are here proposed : 
DIAPHOROPEZA, gen. nov. 
Type, Atrophopoda braueri 'Williston. 


CEDEMAPEZA, gen. nov. 
Type, Atroph. townsendi Williston. 

CATEMOPHRYS, gen. nov. 
Type, Vanderwulpia sequens Townsend. 

BRAUERIMYIA, nom. gen. nov. 

Type, Wulpia Brauer and von Bergenstamm (1892), preocc. by 
Bigot in Dipt. (1886). The genus is a valid one. The new name 
is proposed as a tribute to the memory of Friedrich Brauer, the 
one student who has most advanced our knowledge of the Mus- 
coidean flies. 

Below is a table of these and certain allied genera. All of them 
except Vanderwulpia, Brauerimyia, and Catemophrys have the para- 
facial bristles continuous with frontal row. This character, how- 
ever, does not seem to indicate close relationship in all cases, as it is 
probable that Ceratomyiella, Metachceta, Dichocera, and Atrppho- 
palpus, all here included, belong in other subfamilies from the rest. 
Hypertrophocera and certain other genera not included in the table 
possess this character. 

Frontal bristles stopping short at insertion of antennae, apical cell end- 
ing well before wing-tip, stump of vein at bend of fourth, hind cross- 
vein nearer bend of fourth, cheeks about one-third eye height, eyes 
bare, arista pubescent basally, abdomen slender and rather conical, 
macrochsetse only marginal 2 

Frontal bristles descending to middle of second antennal joint, apical 
cell ending well before wing-tip closed or extremely short petiolate, 
a black wrinkle but no stump at bend of fourth, hind crossvein nearer 
to bend, cheeks about one-half eye height, eves bare, arista pubescent 
basally. abdomen elongate, macrochsetse only marginal. 

Type, Vander. sequens Catemophrys, gen. nov. 

Frontal bristles descending on parafacials to lower border of eyes 3 

2. No costal spine, apical cell long petiolate, parafacials bare. 

Type, atrophopodoides I ' dhderwulpia 

A costal spine present, apical cell narrowly open or closed in margin, 
parafacials distinctly very short pilose. 

Type, Wulpia aperta Brauerimyia 

3. Palpi atrophied, minute, apical cell closed in border at wing-tip, hind 

crossvein nearer bend, eyes almost bare. 

Type, angusticornis Atrophopalpus 

Palpi normal '. 4 

4. Eyes bare, a costal spine, cheeks not over one-fourth eye height; apical 

cell ending well before wing-tip, long petiolate: claws of both sexes 
atrophied and tarsal joints compressed and swollen, arista pubescent 


in female, hind crossvein is middle between small crossvein and bend 

of fourth vein, macrochaetae only marginal. 

Type, townsendi (Bdemapeza, gen. now 

Eyes bare, a costal spine, cheeks one-half eye height 5 

Eyes hairy " 

5. Apical cell long petiolate, ending well before wing-tip; hind crossvein 

in middle between small crossvein and bend of fourth vein. 

Type, atra Metachata 

Apical cell very short petiolate, ending but slightly before wing-tip; 
hind crossvein a little nearer to bend of fourth vein. 

Type, conica Ceratomyiella 

6. Apical cell open and ending well before wing-tip, bend of fourth vein 

with long stump, male antennae with third joint lyriform cleft. 

Type, lyrata Dichocera 

Apical cell ending well before wing-tip, a costal spine, hind crossvein 
much nearer to bend of fourth vein, macrochaetae only marginal 
(except on anal segment) 7 

Apical cell ending at or but slightly before wing-tip, eyes thinly hairy. . 8 

7. Apical cell closed in border (or narrowly open, fir very short petiolate), 

eyes thinly hairy, cheeks fully one-half eye height. 

Type, mexicana Microchira 

Apical cell moderately long petiolate, eyes thickly hairy, cheeks nearly 
one-half eye height. 

Type, magnicomis Laclinoinnia 

8. Apical cell narrowly open, a costal spine, hind crossvein nearer to bend 

of fourth vein, cheeks not over one-fourth eye height, eyes very 
sparsely hairy, all the male claws and pulvilli much elongated, female 
claws atrophied and tarsal joints compressed and swollen, macro- 
chaetae only marginal. 

Type, braueri Diaphoropesa, gen. nov. 

Apical cell closed in border (or very narrowly open or very short 
petiolate). double costal spine, hind crossvein much nearer to bend 
of fourth vein, cheeks fully one-half eye height, macrochaetae discal on 
last segment only, claws and pulvilli of both sexes atrophied (?) 
and tarsal joints swollen. 

Type, singularis itrophopoda 

Apical cell open, cheeks nearly one-half eye height, hind crossvein near 
middle, no discal macrochaetae (ace. v. d. Wulp) or present on last 
segment only (ace. B. & v. B.), male claws and pulvilli of anterior 
tarsi elongated. 

Type, validinervis Paradidyma 

Note to table.— The group of Pseudomintho, Olivieria, etc., has the front 
tarsi of female plump and swollen, with very small claws. The group of 
Mintho, Actinochata, and Buantha has the last tarsal joint of all the feet in 
both sexes swollen, and claws very short. But the frontal bristles do not 
descend half way to vibrissas in any of these forms, and they are thus easily to 
be distinguished from the above genera in the table, having somewhat similar 

Cholomyia inaquipes Bigot.— One specimen bred at the Cotton 
Boll Weevil Laboratory, Dallas Texas, from Conotrachelus elegans, 
issued May 29 (Hunter). 


NEAPORIA, nom. gen. nov. 

This name is proposed for Aporia (Macquart) Brauer and von 
Bergenstamm, which is preoccupied. The type of the genus is 
qnadrimaculaia Macquart, of South America. The species lima- 
codis Townsend seems to belong here also. The latter is distinct 
from Dc.via pristis Walker in its practically bare arista. D. pristis, 
so far as can be judged from the description, is not a Macquartia 
s. str. Mr. E. E. Austen has referred it to Aporia (Ann. Mag, N. 
H., ser. 7, vol. xix, p. 344). 

RONDANIMYIA, nom. gen. nov. 

This name is proposed in honor of Camillo Rondani for his genus 
Gynniopsis (Dipt. Ital. Pr., 111, 1859, pp. 90-91), which is pre- 
occupied by Rafael in Pisces (1815). The type is Macq. chalconata 
(Wiedemann, Meigen, Zetterstedt) Rondani, of Europe. Brauer 
and von Bergenstamm retain the species in Macquartia, but it seems 
preferable to maintain it separately on the characters pointed out by 

METHYPOSTENA, gen. nov. 

This genus is proposed for the type of Hypostena barbata Coquil- 
lett, which can be referred to neither Hypostena, Tachinophyto. nor 
Pseudomyothyria. There are no bristles on the third longitudinal 
vein, the small crossvein is almost opposite to the end of the first 
vein, hind crossvein is in middle between small crossvein and bend 
of fourth vein, apical cell ends in exact wing-tip. The wings are 
narrow, their width being much less than one-half their length. The 
parafacials are narrowed below to a mere line next the lower border 
of eyes, the facial profile is very oblique and receding, the lower 
margin of head short, the arista strongly curved. 

Subfamily Pyrrhosun.t; 
Genus Leskia Robineau-Desvoidy 
Syn. Pyrrhosia pt. (Rondani) Brauer and von Bergenstanjm. 
Type, aurca Fallen. 

Genus Pyrrhosia Rondani (restricted) 

Syn. Myobia (Schiner) Brauer and von Bergenstamm. 

Type, inanis Fallen. 

Genus Anthoica Rondani 

Syn. Myobia Robineau-Desvoidy (preoce. — non Schiner, Brauer and 
von Bergenstamm). 

Type, atra Rondani. 


It is clear that the name Leskia Robineau-Desvoidy can not be 
properly substituted for Myobia Robineau-Desvoidy (preocc), inas- 
much as the species referred to Leskia by that author are not typical 
Myobia in his sense. 

Rondani proposed the name Anthoica for this very purpose, and 
it must therefore be employed. Leskia should be recognized as 

Genus Aphria Robineau-Desvoidy 

Aphria ocypterata Townsend. — One female, Massachusetts (No. 
142, Riley Coll.). Length, 7 mm. Agrees with original descrip- 
tion. The stump of fifth longitudinal vein does not quite reach 
margin of wing. The hind crossvein is nearly in middle between 
the small crossvein and bend of fourth vein, the bend being quite 
rounded. The third antennal joint is distinctly and evenly rounded 
on both apical corners. 

Aphria occidentale, sp. nov. 

One female, Colorado (No. 120, Riley Coll.) ; one female, Beulah, 
N. Mex., August (Cockerell) ; one male, Roswell, N. Mex., August 

Length of female, 7J/2 to 8 mm. ; of male, 9 mm. Differs from 
ocypterata in being more robust, larger, the abdomen more broadly 
red on sides, the red extending length of first segment and half or 
more length of third segment; third antennal joint in both sexes dis- 
tinctly angular on front apical corner, rounded on posterior apical 
corner, widened in male ; stump of fifth vein extending to margin of 
wing; hind crossvein more noticeably approximated to bend of 
fourth vein, which bend is abrupt. 

The greater size, the character of third antennal joint, and the 
more widely red abdomen will at once distinguish the species. 

Type.— Cut No. 10,900, U. S. N. M. (Colorado, Coll. Riley). 

Aphria georgiana, sp. nov. 

Two females, Georgia (Riley Coll.), (=?Ocyptera triquetra 
Olivier et ? Ervia triquetra Robineau-Desvoidy). 

Length, 10 mm. This is a distinct species from both of the pre- 
ceding. It is not so typical of Aphria as are the other species, being 
much larger and wider-bodied. Frontal bristles descend but slightly 
below insertion of antenna?, hardly more than to base of second 
antennal joint. The third vein is spined only one-half or three- 
fourths way to small crossvein, hind crossvein is nearly in middle 


between small crossvein and bend of fourth vein, stump of fifth vein 
extends to wing border; first and second antennal joints and base of 
third reddish yellow, arista and rest of third antennal joint black or 
brownish; thjrd antennal joint rounded on posterior apical corner, 
subangular on anterior apical corner. Palpi brownish yellow, or 
with a reddish tinge. Front fully one-half width of head, frontalia 
brownish yellow ; face and front silvery white, parafrontals slightly 
cinereous. Thorax, scutellum, and pleurse quite thickly silvery 
pruinose over the black ground color. Abdomen obscure light 
brownish red, obscurely blackish on median line, broadening on hind 
portions of second and third segments and nearly covering fourth 
segment ; anterior borders of second to fourth segments broadly 
silver}- pruinose, but more faintly so than thorax; legs blackish; 
wings clear, slightly tawny at base. Teguke white, very slightly 
tinged with yellowish. 

Type.— Cat. No. 10,901, U. S. N. M. 

PHOSOCEPHALA, gen. nov. 1 

Form rather L?/r/7/a-like, narrow, abdomen round-oval, head yel- 
low, wings slightly smoky, palpi absent, thorax and abdomen me- 

Head and thorax about same width, abdomen slightly wider. 
Front (female) not prominent in profile, distinctly more than one- 
half width of head, flattened anteriorly, steeply sloping on anterior 
two-thirds, ocelli marking summit, vertex lower; parafrontals wide, 
not swollen, clothed with some fine black hairs ; vertex not nar- 
rowed ; frontal bristles descending in a single row about to middle 
of second antennal joint, the four front pairs decussate and widely 
divergent below ; two strong reclinate frontal bristles next behind 
these and between them a pair of weak bristles also directed back- 
ward, the outer one outward ; two reclinate vertical bristles of equal 
strength on each side, the outer one directed also outward, these 

1 This genus and several others were purposely described in detail in order to 
furnish a forcible illustration of the length of a full generic description in 
these flies, mentioning all the characters, such as would be necessary to enable 
the student to absolutely place the form in its proper tribe or subfamily with- 
out reference to the specimen. Such a description is far too long for practical 
use, and demonstrates the inadvisability of attempting systematic work in tins 
superfamily without a great amount of previous study and access to a large 
central collection where all types are to be permanently preserved. Especial 
attention is here called to the fact that all these characters require to be 
studied and compared in order to determine the final location of a genus of 
these flies. 


being strongest of all ; postvertical bristles small, of same size as the 
black border row of occiput ; ocellar bristles strong, proclinate, diver- 
gent ; postocellar bristles represented only by weak hairs ; two strong 
proclinate orbital bristles ; lunula normal ; frontalia differentiated 
only by being bare of hairs ; facial plate elongate, ovate, widened 
below, about as high as parafacials, greatest width just above vibris- 
sal angles and taking up three-fifths the facial width at that point, 
quite flat, slightly advancing below, reaching quite to lower margin 
of head; facial carina absent, antennal grooves hardly developed at 
all; facialia divergent inferiorly to point just above vibrissas, then 
feebly convergent ; facial bristles about two above vibrissas ; vibrissas 
quite widely separated, inserted just a little above the oral margin; 
vibrissal angles only moderately pronounced, rather rounded, sit- 
uated moderately close to oral margin ; parafacials not quite twice as 
wide at base of antennas as on lower orbits, flattened anteriorly, with 
some fine black hairs next lower eye-margins: epistoma moderately 
prominent, narrowed, showing a cut-off flattened edge below 
vibrissas; mouthparts normal, proboscis when extended about as 
long as head height, moderately fleshy, only once bent, labella mod- 
erately developed; palpi entirely absent, showing no trace; oral 
cavity moderately narrow and elongate ; peristomalia with a row of 
seven or eight black bristles, which are continued around edges of 
occiput ; longitudinal axis of head at oral margin practically same 
as that at insertion of antennas, the facial profile being slightly con- 
cave, and profile of parafacials straight but obliquely receding; an- 
tenna? inserted about on a line drawn through middle of eyes and 
about on upper three-fifths of head height, closely approximated ; 
second antennal joint slightly elongate, fully twice as long as first 
joint; arista bare, moderatel} long and slender, a little thickened on 
basal one-third, basal joints short and indistinct; third antennal joint 
about twice the length of second, moderately wide and of equal 
width, rounded on both apical corners; eyes bare, not large, set 
rather high, not extending as low as vibrissas, about twice as long 
as wide; cheeks about as wide as one-half of eye height, clothed with 
very fine light hairs, cheek grooves absent ; lower margin of head 
nearly straight, but rounded behind; occiput slightly swollen on 
1< i\\ er two-thirds. 

Sternopleural bristles 3, the middle one inserted lower than the 
others and about equally distant from them ; hypopleural bristles 
moderately strong, about 5 in number : 2 pteropleural bristles, the 
posterior one very strong, curved, reclinate ; mesopleural bristles in 
a posterior fringe of 7; propleural bristles 3, curved upward and 
forward ; notopleural bristles 2, strong, curved, reclinate ; postsutural 


bristles 4, the posterior one on each side reaching beyond hind bor- 
der of scutellum, the others much less strong; 3 dorsocentral bris- 
tles ; 4 short acrostichal bristles before suture, one strong one next 
scutellum (if, more behind suture, the pin has destroyed them) ; 
6 humeral bristles, moderately short ; 3 intrahumeral bristles ; 1 pre- 
sutural bristle nearly in line with last ; 3 intra-alar bristles, one in 
front of suture ; 3 strong reclinate supra-alar bristles ; I strong post- 
alar bristle reaching middle of second abdominal segment, 2 weak 
ones below it ; scutellar bristles consisting of 3 strong and 2 weak 
pairs, an apical decussate weak pair, a weak, more separated sub- 
discal pair in front of last, a strong subapical pair reaching almost 
to base of third abdominal segment, a shorter pair outside these, and 
the strongest macrochsetse of entire body being a lateral pair inserted 
on border in front of last, and which reach nearly to base of third 
segment ; some other bristly hairs on scutellum appearing more or 
less like weak macrochaetoe. 

Wings not large, rather narrow, extending about length of anal 
segment beyond end of abdomen, normal ; costal spine distinct but 
short ; third longitudinal vein with about five bristles at base ; other 
veins not spined ; fourth vein ending in wing-tip, straight to bend, 
which is sudden (but hardly angular) and very obtuse, last section 
straight, the whole vein so gently bent as to distinctly narrow the 
apical cell, bend without stump or wrinkle and slightly more re- 
moved from hind margin than any part of the vein beyond it ; fifth 
vein running half way from hind crossvein to wing border, rest 
being wrinkle ; apical cell closed in margin, hind crossvein distinctly 
trisinuate, a little nearer to bend of fourth vein than to small cross- 
vein, but not greatly removed from middle, the axis of its anterior 
half at almost a right angle to fourth vein ; small crossvein slightly 
before middle of discal cell. 

Abdomen of 4 segments, broad-oval, almost round, strongly con- 
vex above, subflattened below, first segment shortened ; macrochaetae 
weak, only marginal except on last segment, first segment without 
any, second segment with a weak median pair and a weak lateral 
one, third segment with a marginal row of 8, anal segment with 
some marginal ones and a row of 6 subdiscal ; ovipositor withdrawn 
inside the subcircular anal orifice on ventral side of last segment. 

Legs short (only the hind pair present), femora with short black 
bristles ; hind tibiae not ciliate, with sharp bristles on posterior side 
and some shorter ones on front side ; tarsi not stout, moderatelv 
slender, short, about same length as tibiae, metatarsi fully as long as 
the other joints taken together; claws and pulvilli short, a little 
shorter than last tarsal joint. Type, the following species : 


Phosocephala metallica, sp. nov. 

One female, Tucurrique, Costa Rica, collected by Messrs. Schild 
and Burgdorf. 

Length of body, 8 mm. ; of wing, 6 mm. Head entirely pale yel- 
lowish, face and cheeks with a faint silvery bloom ; parafrontals, 
frontalia, and two basal antennal joints unicolorous with a faint 
brownish tinge ; third antennal joint, arista and proboscis pale yel- 
lowish brown ; eyes dark purplish brown. Thorax, scutellum, and 
abdomen shining metallic dark purplish, the abdomen hardly more 
of a purplish black, humeri yellowish; presutural part of mesonotum 
deep golden pruinose, through which run only two linear vitta?, the 
pruinose covering thickest on sides and in front, extending back- 
ward behind suture very faintly on sides of mesonotum ; scutellum 
faintly silvery pruinose; metanotum faintly silvery, pleura? silvery 
gray ; abdomen very faintly silvery, not obscuring the metallic sheen, 
most noticeable on bases of segments, particularly second segment, 
least so on anal segment. Wings distinctly smoky throughout, a 
little more so on costal border, extreme base of costa narrowly yel- 
lowish. Tegulae appearing almost white in some lights, but with a 
smoky yellowish tinge, much whiter than the wings, halteres pale 
yellowish. Legs brownish yellow, tarsi hardly darker, but appear- 
ing blackish from the many short black bristles, coxse lighter yel- 

Type.— Cat. No. 10,902, U. S. X. M. 

Paranaphora diademoides, gen. nov. et sp. nov. 

This new genus and species are proposed for Ervia triquetra of 
Mr. Coquillett's Revision of the Tachinidse (1897), page 66. The 
species is not to be identified with Olivier's Ocyptcra triquetra, 
which is probably an Aphria. It does not fit Robineau-Desvoidy's 
Ervia triquetra, nor does it belong to his genus Ervia. The species 
looks some like Stomatodexia diadema, from which it may be at 
once known by the bare arista, the very elongate second antennal 
joint, and the atrophied palpi. 

PARANAPHORA, gen. nov. 

The salient characters of the genus are the elongate second an- 
tennal joint and the atrophied palpi, as just mentioned. Front at 
vertex one-third width of head in female, one-fourth in male. Palpi 
extremely small, cylindrical, like a minute grass seed, with a long, 
delicate apical hair. Apical cell narrowly open a little before wing- 
tip, sometimes almost closed in margin. Bend of fourth vein angu- 


lar, with slight wrinkle, sometimes with slight stump. Hind cross- 
vein much nearer to bend of fourth vein than to small crossvein, the 
latter on middle of discal cell. First vein ending well beyond small 
crossvein. A long costal spine present. Macrochaetae only mar- 
ginal, except some submarginal on last segment. 

Second antennal joint about four times as long as first, about equal 
to third. Frontal bristles descending only two below base of an- 
tennae. Arista and eyes bare. Front prominent ; parafacials mod- 
erately wide, about one-half width of facial plate. Face receding, 
epistoma slightly prominent ; facialia bare, except two or three short 
bristles above vibrissae in male, but practically absent in female. 
Vibrissa strong and inserted a little above oral margin. Cheeks 
about one-half eye height. Antennas inserted above line drawn 
through middle of eyes. Occiput swollen inferiorly. Male with- 
out, female with two orbital bristles. 

Scutellum with a very short apical decussate pair of bristles, and 
two strong lateral pairs with a weaker bristle between them. Abdo- 
men composed of five segments, first short, second shorter than those 
following. Male abdomen elongate-conical, last segment laterally 
compressed ; female abdomen ovate with apex conical. Legs rather 
long, tarsi of male very elongate ; male claws very long. Type, 
the following species : 

Paranaphora diademoides, sp. nov. 

Five females, four males, Mississippi, Louisiana. Texas. 

Length, 7 to 12 mm. Head, thorax, and scutellum of male golden 
pollinose, most thickly so on thorax and scutellum. Same parts of 
female brassy gray, extending over abdomen. Antennae of male 
reddish yellow, those of female brownish yellow. Frontalia red- 
dish brown. Palpi minute, pale yellowish. Mesoscutum with a 
median pair of linear vittae interrupted at suture and obliterated 
shortly behind same; a lateral triangular blackish marking just in 
front of suture outside these, and a longer, narrower, posteriorly 
pointed one corresponding to it behind suture. Abdomen of male 
reddish yellow with base, median line and broad hind borders of last 
three segments brown, a golden bloom over the lighter portions. 
Female with narrow hind borders of last three segments brown, 
with brassy gray bloom, the second segment faintly yellowish. 
First segment without macrochaetas ; second with anterior and pos- 
terior lateral, and a median marginal pair; third with two lateral 
pairs and a median pair; fourth with 8; anal segment of male with 
about 8 marginal and 6 or more submarginal, those of female less in 


number and more nearly apical. Leg's reddish yellow, tarsi brown- 
ish, hind tibiae brownish, and sometimes the femora less so. Tibiae 
of female all reddish or yellowish. Wings distinctly yellow along 
narrow front border, the submarginal cell clear. Tegulae slightly 
tinged with yellowish, mostly on borders. 

Type.— Cat. No. 10,903. U. S. N. M. (Mississippi, H. E. Weed). 

It is possible that further study will show the distinctness of some 
of the above specimen-. 


This genus is proposed for the type of Demotions venatoris Co- 
quillett, which is not a Demotions. (Latter genus has the epistoma 
not at all produced, and furthermore has a short and fleshy pro- 
boscis.) The present genus approaches both D emotions, Rhino- 
tachina, and Fischeria, but agrees with neither, though it is clearly 
more closely related to the latter, as shown by its produced epistoma. 

There are orbital bristles in the male, all the claws of the male are 
short, and the epistoma is strongly produced noselike (as in Fisch- 
eria); second aristal joint is short but distinct, and arista is short- 
pubescent ; macrochaetae discal and marginal, though only weak 
discal ones (if any) are present on third abdominal segment. The 
proboscis is elongate and horny (also as in Fischeria) , the portion 
below geniculation equal to head height ( also equal to lower margin 
of head). Cheeks wide, fully one-half of eye height, hind crossvein 
nearer to bend of fourth vein than to small crossvein. Washington 
State (O. B. Johnson). 

NEOFISCHERIA, gen. nov. 

This genus is founded on the specimen from Philadelphia. Pa., 
mentioned on page 120 of Mr. Coquillett's "Revision" as Demotions 
venatoris. It is related to ParaHscheria, from which it differs as 
follows : 

Male: Discal macrochaetae well developed on last three abdominal 
segments, consisting of a transverse discal row on last segment and 
a single discal pair on intermediate segments; basal segment with a 
lateral marginal, but no median marginal ; next segment with both ; 
last two segments with a marginal, row. Hind crossvein nearly in 
middle between small crossvein and bend of fourth vein; no orbital 
(middle fronto-orbital) bristles in male, and male claws elongate. 
Cheeks about one-third eye height, proboscis elongate. Front tarsi 
(male) much longer than front tibiae. Type, the following species: 


Neofischeria flava, sp. nov. 

One male, Philadelphia, Pa. Coll. Coquillett. 

Length, n mm. General color yellowish. Antennae reddish yel- 
low ; arista and third antennal joint, except base, blackish. Head, 
thorax, and scntellum dark in ground color, thickly light golden 
pollinose, the face more silvery. Palpi light reddish yellow. Abdo- 
men and legs light reddish yellow, the tarsi quite blackish ; abdomen 
thickly light golden pollinose, under which shows faintly a broad 
median stripe that widens on next to last segment into a triangular 
marking spreading along hind border, anal segment tinged with 
blackish only on the broad median line. Venter tinged with darker 
apically. Wings clear, very slightly yellowish at base ; tegulse tinged 
with yellowish; halteres yellowish, including stalks. Pulvilli rather 
smoky ; claws brownish, with black tips. 

Type.— Cat. No. 10,904, U. S. N. M. 

EUDEMOTICUS, nom. gen. nov. 

This name is proposed for Plagiopsis Brauer and von Bergen- 
stamm (1889), which is preoccupied in Hemiptera by Bergroth 
(1883). Type, Dcmoticus soror Egger, of Europe. 

APACHEMYIA, gen. nov. 

This genus is proposed for Dcmoticus pallidus Coquillett. 

Only marginal macrochsetse, front tarsi much longer than front 
tibiae ; proboscis only moderately elongate, horny, with large labella, 
cheeks fully one-half eye height, male without orbital bristles. 
Claws of male elongate. Hind crossvein nearer to bend of fourth 
vein than to small crossvein, apical cell narrowly open before wing- 
tip, bend of fourth vein without wrinkle. Frontal bristles descend- 
ing to middle of second antennal joint, latter being more than twice 
the length of the somewhat elongate first joint, second aristal joint 
short, some fine hairs on parafrontals outside the frontal bristles, 
facialia bare, epistoma strongly produced. Palpi well developed, 
elongate, a little thickened apically, slightly curved. 

Represented in U. S. N. M. by two male specimens collected on 
the Rio Ruidoso, White Mountains, New Mexico (Townsend), on 
flowers of Rhus glabra, 6,500 to 6,700 feet, July 25 and 29, and by 
type of D. pallidus, male, Denver, Colo. The species is large and 

All three specimens may be considered as A. pallida Coq 


EUPHASIA, nom. gen. nov. 

This name is proposed for the Australian Neophasia Brauer and 
von Bergenstamm, which is preoccupied in Lepidoptera. 

Genus Drepanoglossa Townsend 

The genus Drepanoglossa ( type, lucens Townsend \ has the cheeks 
one-third or more of eye height. Bpigrimyia is distinct in having 
extremely narrow cheeks and parafacials ; the eyes long" and extend- 
ing low, full}' to insertion of vibrissas; proboscis shorter, parafacials 
hardly widened above, front not prominent, epistoma strongly pro- 
duced below, face perpendicular, and tarsal joints short. 

Drepanoglossa amydriae, sp. nov. 

Three specimens, bred from masses of pupae of a tortricid, Amy- 
dria sp., sent by Prof. A. L. Herrera, Cuernavaca. Mexico. 

Length, 6 to 7 mm. Differs from lucens in whole coloration being 
darker; wings slightly infuscate. with a faint yellow tinge in the 
marginal cell; the mesoscutum cinereous pollinose with a faint tinge 
of brass\-; abdominal segments, except anal, with a narrow hind 
margin of brown. Proboscis black on apical half. 

Type.— Cat. No. 10,905. U. S. N. M. 

Drepanoglossa lucens Townsend. — This species has the wings per- 
fectly clear, the mesoscutum pale flesh tint with silvery-white pollen, 
the abdomen pale clear yellowish except median line and more or 
less of anal segment, no dark hind margin on first segment and only 
faint ones on middle segments. 

Tribe vnxi 

Tribe Epigrimyiini is close to Phaniinae, but best retained as a 
separate tribe not actually coming within that subfamily. It in- 
cludes Bpigrimyia only. The genus Drepanoglossa clearly falls 
within the subfamily Pyrrhosiinse. 

Tribe Leucostomini 
Genus Leucostoma Meigen 

Lcucostonia nigricornis Townsend. — The species nigricomis and 
senilis are distinct, and may be recognized by the characters given in 
the descriptions. L. nigricomis is essentially a southwestern and 
western species, and senilis an eastern and northeastern species. 
The former has the antenna? more uniformlv blackish, the second 


and third joints equal in length; the latter has them more rufous, 
the third joint being distinctly longer than the second as a rule. 
L. nigricornis has the sides of abdomen somewhat reddish at base, 
and the femora and tibiae more or less so as well. Both species 
Tjelong in Lewcostoma. The genus Phyto has the cheeks and para- 
facials much widened, the cheeks about one-half eye height. 

The species atra Townsend and neomexicana Townsend are like- 
wise distinct forms, and do not belong to Phyto. It is doubtful if 
they can be properly referred to Lcucostonia. 

Lcucostoma analis Meigen. — This species should not be recorded 
from America, as van der Wulp was presumably in error in his 
determination of it in Mexican material. 

Subfamily I'haxii.wt-; 
Genus Hemyda Robineau-Desvoidy 

Hemyda an rata Robineau-Desvoidy. — The males of this species 
have the yellow of third segment wider than the females, but only 
slightly so, and nowhere nearly approaching in that respect the form 
described below from New Mexico. There are eight specimens of 
this form in the U. S. N. M., from Missouri, Kansas, Illinois, and 
Wisconsin, one of them being labeled "attracted to light, July, 1876" 
(Riley Coll.). 

It is interesting to note that one of the above specimens, from 
Milwaukee, Wis., has the small cjossvein of both wings practically 
absent; frontal bristles long, numerous, and thickly placed, and 
vibrissa? distinct, as in several others of the specimens. 

Hemyda sp. — One male, Rio Ruidoso, White Mountains, New 
Mexico (Townsend), about 6,500 feet, August 1, on flowers of 
Monardq stricta. Length, 12 mm. Differs from all the above speci- 
mens by having the yellow of third segment taking up anterior two- 
thirds of length of segment except a median triangular prolongation 
anteriorly of the black of hind portion, which stops well before 
anterior edge of segment. The femora .have only a faint trace of 
the black of aitrata in a tinge of brown before apex. The yellow of 
second abdominal segment is more extended forward also. The 
specimen shows only microscopic vibrissa?, invisible save with a 
high-power lens. 

This specimen probably represents a distinct form, but it is not 
deemed wise to name it as such without first securing a considerable 
series of specimens to substantiate its claim to distinctness. 


Hcmyda {Ancylogaster) aniuita Bigot. — It is highly probable 
that this is a good species. It may even be a good genus. Bigot 
states that the second antennal joint is longer than the third. If it 
develops that the second joint in Bigot's type is strongly elongate, 
more so than is aurata, that is to say more elongate as compared 
with the first joint (not the third), then it is likely that Ancylogaster 
should be retained. 

Genus Penthosia van der Wulp 

Penthosia satanica Bigot.— In this species the fourth longitudinal 
vein is slightly rounded at bend, and often bears a very short stump 
which can not be considered as the prolongation of the fourth vein 
beyond the apical crossvein, since no wrinkle is present in its ab- 
sence. It always points straight away from the bend, like the stem 
from the arms of a Y, and is to be regarded perhaps as indicating 
an original sharp bend of the vein back upon itself for a short dis- 
tance, the two approximated parts having later coalesced, finally dis- 
appearing more or less completely. The writer knows of no other 
tachinid which exhibits this peculiarity in the same degree. 

Genus Cercomyia Brauer and von Bergenstamm 

Synonyms are Uromyia Meigen (preocc.) and Neouromyia 
Townsend, nom. gen. now (Trans. Am. Ent. Soc, December, 1891, 
p. 382). 

Subfamily Gvmxosomatix.e 

Genus Gymnosoma Meigen 

The following description of the external anatomy of the male 
abdomen will be of interest as throwing light on the taxonomic posi- 
tion of the genus. 

The male of Gym. fuliginosa Robineau-Desvoidy has six abdom- 
inal segments besides the genitalia. The first segment is very short, 
and its width is equal to about one-half the greatest width of abdo- 
men. It consists below r of a small, much shortened, subquadrate, 
basal ventral plate, w 7 ide in front and somewhat incurvate on front 
edge where it joins metathorax, rapidly narrowed posteriorly, its 
hind margin much shorter than its front margin. The second ven- 
tral plate is a smaller replica of the first, its front edge being the 
same length as the posterior edge of first, its sides converging pos- 
teriorly on same lines, its posterior edge being correspondingly 
shortened. The first and second ventral plates together thus appear 
much like a right-angled triangle in outline, with the hypothenuse 


representing" the front margin and the right angle cut off truncate 
to represent the hind margin. 

The first segment consists above and on sides of a strip-like dorsal 
plate evenly Repressed between its lateral edges, which are tucked-up 
rounded folds of the plate, the latter ending ventrally on each side 
in a short, pointed strip that does not meet the ventral plate, the ven- 
tral membrane intervening between them. A small spiracle, smaller 
than any of the others of abdomen, is present in the extreme point of 
the first dorsal plate on each side where it joins the ventral mem- 
brane, and each one of the other five dorsal plates has a similar but 
larger spiracle on its inner edge, these being in each case quite well 
removed from the lateral margin where it joins ventral membrane. 

The third ventral plate is nearly rectangular, a little broader than 
long, about as wide as mean width of second plate. The fourth ven- 
tral plate is considerably broader than the third and much shorter, 
thus looking like a narrow transverse strip set in the ventral mem- 
brane. The fifth ventral plate is much wider than fourth, about 
same length, and its median portion (about middle one-third) ap- 
pears to be crowded under the fourth plate by the walls of the 
sheath-like genital opening, partially retracted within which lies the 
hypopygium. Thus only the lateral one-third of the fifth plate is 
visible on each side, and these two portions form the narrow visible 
strips of the curved plate, bordering the edge of the genital opening 
on each side, and each pointed at its outer extremity. 

The sixth abdominal segment is not apparent from a dorsal view. 
It is a shortened anal segment that has been pushed over and 
crowded beneath the extremity of the abdomen. It lies just under 
the posterior edge of the abdomen, is rather crescent-shaped, sub- 
semicircular on posterior (appearing anterior owing to inverted posi- 
tion) edge where it encloses the basal segment of the hypopygium, 
slightly squared on anterior lateral corners. It little more than half 
surrounds the orifice of the genital cavity, and bears a spiracle on 
each side at some distance before the pointed end of its tapering 
lateral portion. The basal sclerite or plate of the hypopygium bears 
another spiracle, which is one of the largest in the abdomen, on its 
basal edge, near the spiracle of the sixth segment and appearing as 
if it belonged to that segment. This basal plate of the hypopygium 
represents another abdominal segment, and it should be considered 
as forming a seventh segment of the abdomen rather than the base 
of the hypopygium. 

The ventral membrane is widely apparent and extensive, the ven- 
tral plates all lying free within it so far as contact with the dorsal 
plates is concerned. The area in which the ventral membrane, with 


the enclosed plates, is visible occupies more than one-third the width 
of the ventral aspect of abdomen. 

The plates, both ventral and dorsal, are at once distinguished 
throughout their extent from the membrane by being clothed with 
bristly hairs. 

The above description was drawn from a specimen collected by 
F. C. Pratt, at Poolesville, Maryland, July 9. The abdomen was 
detached and put under the microscope. 

CEDEMASOMA, gen. nov. 

This form (male) agrees with the description of Wahlbergia 
brevipennis H. Loew, except that the fourth vein is bent at a 
rounded angle, and hind crossvein is not strongly oblique. The 
hind crossvein is straight, almost at right angles to the fourth vein, 
hardly nearer to bend of latter than to small crossvein, and at right 
angles to fifth vein. The petiole of apical cell is slightly longer than 
small crossvein, but not twice as long — about one and one-fourth 
times as long. The abdomen is swollen and strongly convex above, 
wider than the thorax, exactly oval in outline from above, the wider 
end forward, absolutely without macrochaetae. Palpi are extremely 
small, almost atrophied, very slender and quite short. Antennae 
as long as face, second joint almost as long as third. No orbital 
bristles. Wings very short and narrow. The claws are about as 
long as last tarsal joint. Type, the following species : 

CEdemasoma nuda, sp. nov. 

One male, Ormsby County, Nevada, July 6, C. F. Baker, Coll. 

Length, 6 mm. ; of wing, 4 mm. Face, parafacials and para- 
frontals from above silvery white pruinose, blackish from in front, 
the silvery extending on cheeks. Frontalia silvery white pruinose, 
with a faint brassy tinge or a golden reflection. Abdomen densely 
covered with moderately short and fine brown or black hairs, and 
entirely without bristles, wholly yellowish red or brownish red. The 
mesoscutum is silvery pollinose in front of suture, but it does not 
show well in some lights. Tegulae white. Palpi pale reddish 
brownish in color. All the rest of insect is black, except the clear 
wings, which are yellowish at base. Otherwise agrees with Loew's 
description of Wahlbergia "brevipennis. 

This form apparently belongs in the neighborhood of Gymnosoma, 
indicated by the absence of macrochaetae and the possession of a 
swollen abdomen. JVahlb. brevipennis H. Loew is this genus, but 
a different species. Loew's specimen is a female from Nebraska, 


length 4§ mm., of wing 3 mm. The writer has examined the type 
in Cambridge. The hind crossvein does not form a right angle with 
fifth vein, the petiole of apical cell is fully twice as long as small 
crossvein, the head is black and shining except face, and the meso- 
scutum does not show silvery before suture. 
Type.— Cat. No. 10,898, U. S. N. M. 

Subfamily Ocypterin.e 
Genus Ocyptera Latreille 

Ocyptera euchenor Walker. — While it seems probable that this 
form and epytus Walker are the same, there can be no certainty in 
the matter until the types are compared. Probably 0. Carolina 
Robineau-Desvoidy is distinct. Some of Bigot's species may also 
prove distinct. It seems probable that Carolina is a southern form, 
and that euchenor is the more northern large form, having the 
cheeks and parafacials narrow, and the eyes elongate, descending 
low. Further study may also show the distinctness of dosiades. 

Genus Beskia Brauer and von Bergenstamm 

Beskia cornuta Brauer and von Bergenstamm and allies. — B. cor- 
nuta is the South American form. The type is from Brazil. The 
figure of the head given by the authors (fig. 276, Muse. Schiz., 1) 
is not typical of Southern States specimens in U. S. N. M. There 
is a marked difference in the third antennal joint. Williston's figure 
of his St. Vincent specimen shows the third antennal joint same as 
the Brazilian. Beskia and Ocypterosipho may be separated on this 

Genus Ocypterosipho Townsend 

Our species may be known as Ocypterosipho eclops Walker. Al- 
though Walker says "palpi black," and does not mention the slen- 
der and elongate proboscis. Mr. E. E. Austen's statement that celops 
belongs here (Ann. Mag. N. H., Ser. 7, vol. 19, p. 345) must be 
accepted. This is the Georgia and Southern States form, and has 
the third longitudinal vein bristly to small crossvein (Georgia, Lou- 
isiana, and Texas specimens in U. S. N. M.). Santo Domingo 
specimens agree with those from the Southern States in having the 
third antennal joint strongly convex on under border and concave 
on upper, presenting a curved outline like that of a pruning-knife 
blade with cutting edge upward, the anterior distal corner of the 
joint being produced in profile into a sharply pointed prolongation. 


Van der Wulp's figure (in the Biol. C. A. Dipt., n, pi. 13, fig. 12) of 
Mexican specimens gives somewhat the same impression. Two 
specimens from Mexico (one Tehuantepec, Sumichrast) show this 
character markedly, the third antennal joint not being truncate at 
tip as in the figures given by Brauer and von Bergenstamm and by 
Williston. It therefore seems evident that not only is tvillistoni a. 
good species, but the genus Ocyptcrosipho may be retained, 0. wil- 
listoni being the West Indian and Central American form, while 
O. cclops is the more northern form occurring in our Southern 
States. It is to be noted that St. Vincent belongs to the South 
American fauna, while Santo Domingo belongs to the Central 
American, which includes parts of Mexico and the Southern United 

ICHNEUMONOPS, gen. nov. 

Bearing much superficial resemblance to Ocyptcra, but differing 
radically in the structure of the basal portion of the abdomen, and 
in head characters as well. Elongate and narrowed in form. Head, 
thorax, and abdomen of almost equal width, but the head distinctly 
wider than the thorax, the abdomen constricted basally into a pedicel 
formed principally by the base of second segment, which shows 
more constriction than any other part. 

No vibrissas that can be differentiated from the bristles of peristo- 
malia. Second antennal joint rather elongate, about three times as 
long as first; third joint elongate, narrowed, about two and one-half 
times as long as second. Arista not distinctly jointed. Front at 
vertex narrower than eye width, but about equal to latter at base of 
antennae. One row of weak frontal bristles extending only to base 
of antennae. One pair of weak ocellar bristles, slightly proclinate. 
One pair of vertical bristles longer than frontal bristles, directed well 
backward. No orbital bristles (male). Cheeks about two-fifths of 
eye height. Face receding, facial profile straight, epistoma promi- 
nent. Facialia bare, not divergent below, ptilinal area about as 
wide as eye width, parafacials about half as wide above, but nar- 
rower below. Facial plate elongate, not narrowed below, produced 
on lower edge. Antennae inserted above eye middle, rather above 
three-fourths of head height. Eyes bare, descending about three- 
fourths way to lower margin of head, which is long. Seen from in 
front, the space between lower angles of eyes is more than twice that 
between upper angles, the frontofacial area evenly widening from 
middle of front to cheeks. Proboscis below geniculation hardly as 
long as antennae, labella well developed. Palpi extremely small and 
short, atrophied. Occiput convex, swollen on lower three-fourths. 


Three postsutural bristles, no acrostichal bristles either before or 
behind suture. Only one sternopleural bristle (possibly one has 
been lost anteriorly on both sides). 

Scutellum with a weak apical pair of bristles that are strongly 
decussate. Also a much longer marginal subapical pair, and a weak 
marginal pair behind last. 

Abdomen strikingly Ichneumon-like in outline, consisting of five 
segments- visible from above. Basal segment narrow, but little 
wider than base of scutellum, narrowed behind. Postbasal or sec- 
ond segment still narrower on basal portion, the greatest constric- 
tion being at about anterior three-fourths of the segment where the 
abdomen is narrower than scutellum. The second segment grad- 
ually widens posteriorly from the point of its greatest constriction, 
until on hind border it is twice its anterior width. The third seg- 
ment widens posteriorly at not quite same angle, the fourth or pre- 
anal segment narrowing posteriorly about as rapidly as the third 
segment narrows anteriorly. Anal segment still narrowing pos- 
teriorly and evenly rounded on apex. The basal and anal segments 
are about same length, the second segment nearly twice as long. 
The third and fourth segments are equal and each is a little over 
twice as long as anal. 

Second, third, and fourth segments with a median marginal pair 
of macrochsetas quite removed from posterior border of segment, 
also a lateral marginal one on each side. Anal segment with only 
an outer pair on each side near hind margin. Second segment with 
a lateral one in middle on each side. 

Ventral plates not visible, except that the basal plate shows 
plainly with adjacent ventral membrane rather widely surround- 
ing it. 

Wings elongate and narrow, reaching about to end of abdomen. 
No costal spine. Small crossvein nearly opposite end of first vein,, 
distinctly beyond middle of discal cell. Hind crossvein in middle 
between apical crossvein and small crossvein, strongly bisinuate.. 
Apical crossvein still more strongly bisinuate, quite S-shaped. 
Fourth vein produced beyond apical crossvein in a short stump. 
Petiole of apical cell half as long as apical crossvein, reaching an- 
terior wing border well before wing-tip. No veins spined, not even 
third vein at base. 

Tegulse of moderate size, inner portion subsemicircular in out- 
line, so transparent (except narrow borders) that the halteres be- 
neath are almost as clearly seen through them as through glass. 


Legs moderately long, but quite normal. Claws and pulvilli 
elongate, but not longer than the last tarsal joint, which is itself 
elongate. Type, the following species : 

Ichneumonops mirabilis, sp. nov. 

One male, Beulah, N. Mex., August 17. Prof. T. D. A. Cocker- 
ell, Coll. 

Length, 10 mm. ; of wing, about 7 mm. Wholly dull black, abdo- 
men very slightly shining. Antennas light yellowish brownish on 
second joint and base of third. Face and parafrontals slightly sil- 
very, extending on occipital orbits. Thorax and scutellum thinly 
silvery pollinose. Abdomen still more thinly silvery pollinose, the 
narrow hind margins of first three segments pale brownish with a 
yellowish tinge, that of first segment twice as broad as those of the 
others. Legs largely brownish reddish on femora, and especially 
on tibiae. Wings on costal half deeply smoky, tinged with yellow- 
ish, including basal cells. Tegulae and portion of wing behind fifth 
and sixth veins clear hyaline ; discal and apical cells faintly clouded, 
the latter more so. 

Type— Cat. No. 10,899, U. S. N. M. 


Tribe Coroximyiixi 

Genus Coronimyia Townsend 

Coronimyia and Bpigrimyia are distinct genera, belonging to and 
representing distinct tribes. Coronimyia has the arista short and 
geniculate, with very long second joint. Bpigrimyia has the arista 
elongate, with basal joints short. 

EUCORONIMYIA, nom. gen. nov. 

This name is proposed for the genus Isoglossa Coquillett (Can. 
Ent., 1895, pp. 125-126), which is preoccupied by Casey in Coleop- 
tera (Annals New York Acad. Sci., 1893, p. 304). The characters 
are sufficient to retain the genus. 

Genus Olenochaeta Townsend 

Olenochata kansensis Townsend. — This form, Pseudogermaria 
georgicv Brauer and von Bergenstamm, and Distichona varia van der 
Wulp are all generically distinct. 

Genus Chaetoglossa Townsend 

Chatoglossa nigripalpis Townsend. — This species differs from 
viola Townsend in having black palpi and discal macrochaetae on 
third abdominal segment. It is also twice the size of viola. By 


snrne error the words "palpi black" were omitted from the descrip- 
tion (Tr. Am. Ent. Soc, xix, p. 126). C. viola has palpi light 
orange, and third abdominal segment is without discal macrochastas. 

Subfamily Thryptocerati.vl: 
Genus Ceratomyiella Townsend 

Ceratomyiella conica Townsend. — This genus may be known by 
the apical cell ending but slightly before wing-tip, usually if not 
always short petiolate ; bend of fourth vein not sharply angular, 
third vein bristly not quite to small crossvein, fifth vein not at all 
bristly, and costal spine very small. The face is so elongate and 
retreating in profile below eyes as to bring the insertion of vibrissas 
nearly or quite into the transverse plane of the hind margins of 
eyes ; the cheeks are one-third to one-half eye height in width 
(nearly one-half in C. conica). 

Chcetoplagia has the apical cell narrowly open or closed in border. 

Metachceta greatly resembles Ceratomyiella in facial characters. 
The facial profile is very receding and elongate below, so as to 
bring the insertion of vibrissas close to or nearly into the transverse 
plane of hind border of eyes (as viewed in full profile). 

ACRONARISTA, gen. nov. 

Allied to Schizotachina Walker, from which it is at once distin- 
guished by the remarkable characters of the third antennal joint. 
This is biramose in the female, being split into an anterior and a 
posterior ramus, the two rami almost meeting apically and showing 
in profile like an imperfect zero. The inner or under ramus is a 
little widened apically ; otherwise the profile width of both is prac- 
tically the same throughout, even including the base of the joint 
where the rami join. The arista is inserted in the anterior edge of 
the upper ramus well before its apex, but much nearer the apex than 
the base, being at a distance from the apex equal to one-third the 
length of the joint. The second aristal joint is only about twice as 
long as wide, the first about as long as wide, and the third about 
three times as long as second. Front equilateral, about one and 
one-half times as wide as one eye, two middle fronto-orbital bristles 
in female, facial plate very wide, parafacials reduced to a mere line, 
facialia practically bare. Apical cell closed in margin near wing- 
tip, last section of fourth vein bent in ; hind crossvein distinctly 
nearer to small crossvein than to bend of fourth vein, but not nearly 
so approximated to small crossvein as in Schizotachina. Third vein 


with a few bristles at base only, costal spine present. Type, the 
following species : 

Acronarista mirabilis, sp. nov. 

One female, Palm Beach, Florida. Dr. H. G. Dyar, collector. 

Length, 4 mm. Blackish, with gray pollen. Antennae reddish 
brown, becoming more or less reddish yellow at base. Face, front, 
thorax, and scutellum silvery gray pollinose. Abdomen blackish, 
narrow anterior margin of second and third segments and all of 
fourth segment silvery gray pollinose. Legs quite blackish. Teg- 
ulse whitish. Wings faintly tawny at base. 

Type.— Cat. No. 11,685, U. S. X. M. 

It is strongly probable that the male of Acronarista has the third 
antennal joint much more elaborate in structure than that above de- 
scribed for the female, and it will be very interesting to look for the 
male in South Florida material. 

In Talarocera female the location of the arista approaches in a 
measure that of Acronarista female, but is not nearly so apical. 
Acronarista female seems to be a farther development of Talarocera 
female in this regard, in that the ramus of third antennal joint bear- 
ing the arista has become elongated and enlarged into almost the 
counterpart of the other ramus, the elongation taking place at the 
base of the ramus, and thus making the arista subapical thereto. It 
is probable that in the male of Acronarista the arista will be found 
to be apical to one of many rami, as in the male of Talarocera. 
However, this would indicate no near relationship, since Talarocera 
is a large form belonging to the Hystriciinse. 

LIXOPHAGA, gen. nov. 

Differs from Gymnostylia by having macrochsetae of abdomen 
only marginal ; parafacials and parafrontals bare except for the 
frontal and orbital bristles. Male cheeks hardly one-fourth eye 
height : no orbital bristles in male, but a row of six or seven minute 
bristles between frontalia and eye margin on the parafrontals. 
Apical cell closed in margin just before wing-tip. Hind crossvein 
in middle between small crossvein and bend of fourth vein, the latter 
rounded and bent at an obtuse angle. Front about one-third head 
width, widening on anterior portion. Face fully one-half head 
width. Type, the following species. 

Lixophaga parva, sp. nov. 

One male bred from Lisas scrobicollis. Hunter No. 219, Dallas. 
Texas, issued August 15, 1907. 


Length, 3.5 mm. Face, cheeks, parafacials, and parafrontals sil- 
very, the parafrontals tinged with cinereous. Frontalia, antennae, 
and legs blackish. Third antennal joint about three and one-half 
times as long as second. Thorax silvery pollinose with tinge of 
cinereous abflve ; four narrow linear black vittse, the outer ones in- 
terrupted at suture, the inner ones abbreviated just behind suture. 
Scutellum silvery pollinose. Abdomen blackish, the second to 
fourth segments thickly silvery-cinereous pollinose leaving a median 
vitta and the hind margins blackish or brown, the vitta not so 
marked on anal segment. The pollen of abdomen is flecked with 
numerous small dots marking insertion of bristly hairs. Macro- 
chaetse in a median marginal pair and a lateral one on first two seg- 
ments, weaker than the marginal rows on third and anal segments. 

fype.— Cat. No. 11,648, U. S. N. M. 

Subfamily BaumhaueriiNyE 
Genus Euthyprosopa Townsend 

Euthyprosopa pctiolata Townsend.— There are two pairs of 
ocellar bristles in this genus, the posterior pair being about same 
length as frontal bristles. The anterior pair is strongly proclinate, 
almost appressed ; the posterior pair is slightly reclinate, suberect, 
and inserted between the two posterior ocelli. 

Subfamily Plagiint.e 
Genus Plagia Meigen 

Plagia aurifrons Townsend. — This species is from the northeast- 
ern United States, and is not conspecific with the Mexican ameri- 
cana van der Wulp. 

Genus Plagiprospherysa Townsend 

Plagiprospherysa valida Townsend. — It is possible that the 
Presidio specimens referred by van der Wulp to his species parvi- 
palpis may be conspecific with this species, but the others are likely 
to prove distinct. 

Genus Heteropterina Macquart 

Heteroptcrina nasoni Coquillett. — This form seems, from an ex- 
amination of the type, to be quite typical of the genus Heterop- 
terina. The cheeks are very narrow, not over one-tenth eye height, 
and the few fine hairs of the normal row on parafacials are almost 
imperceptible with an ordinary low-power lens, but they are present. 


Subfamily Phoroceratin.e 
Genus Achaetoneura Brauer and von Bergenstamm 

Achatoneura. — This genus is characterized by having the second 
antennal joint but little longer than the first, and thus is easily dis- 
tinguished from Tachina s. str., to which it otherwise bears a strong 
resemblance. Type is hcsperus Brauer and von Bergenstamm, of 
North America. T. aletia Riley belongs to this genu-. 

HEMIARGYRA, gen. nov. 

Form Polleiiia-like, eyes pilose, facialia and hind tibiae ciliate. 
Ptilinal suture bent at a rounded angle in middle superiorly, its ends 
divergent inferiorly, making the ptilinal area almost triangular in 
shape and about one-third head width below. Facial plate elongate, 
not narrowed below, depressed, but not produced anteriorly on lower 
portion, fossae running full length of plate; foveae shallow, but 
marking length of third antennal joint; a low, sharp, narrow carina 
fading out before reaching inferior end of foveae. Facialia sharp- 
edged, narrow in front outline from being set on edge, with a row of 
strong bristles running fully half way up and marking the edge, 
which is nearly straight in outline save for a slight curve inward at 
lower end, and is closely approximated to suture until it begins to 
curve. Vibrissal angles quite distinctly removed from oral margin, 
very faintly pronounced, but slightly more approximated than the 
rows of bristles above them, the vibrissa? strong and decussate. 
Peristomalia subparallel in epistomal region, divergent posteriorly, 
with a row of black bristles extending to beard. Epistoma not 
prominent, not showing in profile. Proboscis very short and fleshy, 
part below geniculation hardly as long as eye width, but about as 
long as the palpi ; labella large, with long hairs ; palpi rather elon- 
gate, moderately slender, but thickened on apical half. Axis of 
head at insertion of vibrissa? distinctly less than that at insertion of 
antennae, facial profile very gently receding, but nearly straight. 
Antennas inserted about on eye middle, at about three-fifths of head 
height ; second joint rather short, but about twice as long as first ; 
third joint elongate, not wider than second, sides nearly parallel, 
subtruncate at apex. Arista thickened on hardly more than basal 
one-third, finely short-hairy, basal joints distinct but short. Eyes 
thickly pilose, extending not quite to vibrissal angles, inner outline 
appearing slightly bulged on middle by reason of a faint incurvature 
below. Front (female) not prominent in profile, at vertex (seen 
from in front) about one-fourth of head width, gradually widening 


anteriorly to distinctly more than one-third head width at base of 
antennae. No ocellar bristles, but some long fine hairs on and in 
front of ocellar area. Frontal bristles in a single row on each side 
close to frontalia, but widely divergent at an angle anteriorly, the 
foremost two being out of line with main row and the only ones 
inserted below first antennal joint. Vertical bristles consisting of a 
moderately strong inner and a very weak outer one, the latter but 
slightly longer than the row forming occipital fringe. Two strong, 
lightly reclinate upper fronto-orbital bristles inserted close to fronta- 
lia. in profile showing same strength and curvature as inner vertical, 
all three being same distance apart in profile. Two strong middle 
fronto-orbital bristles, strongly curved and proclinate, outside line 
of preceding, the posterior one being inserted midway in profile be- 
tween the two upper fronto-orbital bristles, the anterior one about 
half way between foremost frontal bristle and vertex. Parafacials 
wide, only gently narrowed below, fully two-thirds as wide on 
lower portion as opposite base of antennae, bare. Width of cheeks 
equal to about one-fourth eye height, cheek grooves well marked. 
Lower margin of head arcuate, evenly bulged and rounded. Occi- 
put considerably swollen below, behind eyes. 

Two sternopleural bristles, strong, formula 1:0: 1 ; hypopleural 
bristles in a curved row, long but slender ; one moderately strong 
pteropleural bristle with some fine hairs; three postsutural bristles, 
supra-alar bristles stronger. Scutellar bristles in three strong mar- 
ginal pairs and a weak apical decussate pair ; subapical pair longest, 
reaching nearly to base of preanal segment ; a widely separated 
discal pair about as strong as the apical. 

Wings decidedly longer than abdomen, rather broad, with very 
small but distinct costal spine. No veins spined, except a few 
bristles at base of third vein. Fourth vein bent roundly at an ob- 
tuse angle, ultimate section slightly and evenly crooked, no wrinkle 
or stump at bend. Hind crossvein bisinuate, slightly more than 
one-half as far from bend as from small crossvein, which is on 
middle of discal cell and about half way between ends of auxiliary 
and first veins. Apical cell well open, ending on front border well 
before wing-tip. Tegulae very large, antitegulae one-third as long. 

Abdomen broadly oval, rounded anally, only four segments vis- 
ible from above, with almost equally short marginal and discal 
macrochaetae. Ventral plates not visible. Anal segment with a ven- 
tral median cleft, within which is the retracted ovipositor. 

Legs not elongate, normal, the hind tibiae quite thickly ciliate with 
a slightly stronger bristle near middle. Claws and pulvilli (female) 
short, not as long as last tarsal joint. Type, the following species: 


Hemiargyra nigra, sp. nov. 

One female, San Carlos, Costa Rica, collected by Schild and 

Length, 8.5 mm. ; of wing. 8 nun. Blackish, or brownish black. 
Palpi reddish yellow, quite thickly black-hairy, blackish at base. 
Space from anterior fronto-orbital bristle to cheek grooves con- 
spicuously silvery white pruinose as seen from above, covering 
whole area of parafacials and anterior half of parafrontals, but ap- 
pearing dead black when seen from below. Facial plate silvery 
from above, blackish from below. Epistoma yellowish. Third an- 
tennal joint three and one-half times as long as second. Cheeks 
brownish, clothed with black hairs; cheek grooves and edge of 
parafacials bordering suture slightly golden in some lights, brown- 
ish or reddish in others, the golden continued on occipital orbits. 
Frontalia, thorax, scutellum, and basal abdominal segment soft 
black with slight brownish tinge, apex of scutellum silvery. Two 
middle segments of abdomen heavily golden silvery pollinose seen 
from in front, behind, qv above, but nearly lost when seen from 
side ; the coating showing broadly on venter, anteriorly on each seg- 
ment at least. Sides of middle segments slightly reddish under the 
pollen. Anal segment black, with some golden pollen on sides and 
base. Macrochaetse as follows : A median and lateral marginal pair 
on basal and postbasal segments, a median discal pair on postbasal 
and preanal segments, the latter with a marginal row, anal segment 
with only bristly and fine hairs. Wings slightly infuscate along the 
veins, chiefly on costal half, rest subhyaline. Tegulre smoky, with 
smoky yellowish borders widening in oblique lights. Pulvilli 
whitish, with a slight smoky yellowish tinge. Halteres rufous, 
knobs fuscous. 

Type— Cat. No. 10.907, U. S. N. M. 

POLIOPHRYS, gen. nov. 

Ttilinal suture rounded subangular in middle, its ends divergent 
below, giving ptilinal area an oval outline that is quite narrowed 
above and fully one-third head width below. Facial plate elongate, 
not narrowed below, depressed, produced anteriorly on lower por- 
tion ; fossae running full length of plate, foveae deep and marking 
length of third antennal joint; a distinct narrow carina between the 
foveas, with a linear median furrow on its edge. Facialia wide, but 
with rather sharp edge, latter gently curved in outline, well inside 
suture and furnished with bristles extending more than half way up. 
Vibrissal angles quite distinctly removed from oral margin, not 


sharp, conspicuously more approximated than the rows of facial 
bristles above them, the vibrissa; strong and decussate. Peristomalia 
almost straight, nearly parallel, with bristles extending to beard. 
Epistoma prominent, but not showing greatly in profile owing to 
strong depression of facial plate. Proboscis short, fleshy, part 
below geniculation about as long as eye width (in front view), 
labella well developed ; palpi elongate, rather slender, more atten- 
uate on basal one-third. Axis of head at insertion of vibrissas very 
noticeably less than that at insertion of antennae, facial profile gently 
receding, but quite straight. Antennas inserted above eye middle, 
at about three-fourths of head height ; second joint about twice as 
long as first, with a pair of bristles on lower front edge; third joint 
elongate, wider than second, sides nearly parallel. Arista bare, 
thickened on more than basal one-half, conspicuously jointed; first 
joint slightly elongate; second joint elongate, two or three times as 
long as first, and one-fourth to one-fifth as long as third joint. Eyes 
thickly pilose, not extending as low as vibrissal angles, inner out- 
line S-shaped in front view in male, straight in female. Front not 
strongly prominent in profile, but flattened and sloping straight to 
base of antennas; at vertex (seen from in front) one-third head 
width in both sexes, in male suddenly swelling in lateral outline an- 
teriorly, in female gradually widening anteriorly, almost one-half 
head width at base of antennas. A strong pair of proclinate, diver- 
gent ocellar bristles. Frontal bristles in two rows, the inner rows 
strongly curved and widely divergent below, the outer rows nearly 
straight ; the lowermost bristle on each side sometimes in line with 
both rows so as to appear (male) as belonging to either, but belong- 
ing (as shown in female) to inner row, which descends strongly to 
point somewhat below insertion of arista. Short fine hairs on para- 
frontals, long hairs on and in front of ocellar area. The usual 
strong inner and weaker outer vertical bristles, both reclinate, latter 
also divergent ; two lightly reclinate upper fronto-orbital bristles, 
the frontal bristles extending only about half way back from base of 
antennas to vertex. No middle fronto-orbital bristles in male, two 
strong, decidedly proclinate ones in female nearly in line with the 
posterior one of the upper fronto-orbital bristles. Parafacials wide, 
narrower below than opposite antennal insertion, least width about 
equal to length of second antennal joint, greatest width bordering 
parafrontals and not twice as much. Facio-orbital bristles in a 
median row of about five or six, not so strong as frontal or facial 
bristles, some fine hairs outside them. Width of cheeks equal to 
one-third eve height, cheek grooves faint. Lower margin of head 


nearly straight, about two-thirds as long as axis of head at antennal 
insertion. Occiput swollen below, behind eyes. 

Four sternopleural bristles, formula 2:1:1; hypopleural bristles 
in a curved row, strong : pteropleural bristles several, one strong ; 
four postsutural bristles. Scutellar bristles strongly developed ; 
apical pair decussate, suberect, weaker than the other marginal ones, 
a weak discal pair in front of them ; three strong marginal pairs, the 
subapical longest and reaching nearly to middle of preanal segment 
(male), or only to base of same (female). 

Wings a little longer than abdomen, moderately broad, with very 
small but distinct costal spine. No veins spined, except third vein 
with two or three bristles at base. Fourth vein bent at nearly a 
right angle, with very slight (almost imperceptible) wrinkle at 
bend, latter not sharp, apical crossvein well bowed in near origin, 
hind crossvein slightly bisinuate and approximated to bend of fourth 
vein, small crossvein half way between end of auxiliary and end of 
first vein. Apical cell widely open, ending far before wing-tip. 
Tegulse large, antitegulae overlapping them for one-third of their 

Abdomen rather broadly oval, quite pointed at apex, only four 
segments visible above, with strong marginal and short, weak discal 
macrochsetae. Ventral plates not visible. Anal segment in both 
sexes with a median ventral slit for protrusion of genitalia, which 
are retracted. 

Legs not elongate, middle tibiae with three strong bristles in mid- 
dle, hind tibiae weakly ciliate with a long bristle in middle of ciliated 
edge ; tarsi normal ; male claws and pulvilli elongate, longer than 
last tarsal joint; female claws about as long as last tarsal joint. 
Type, Polio phrys sicrricola sp. nov. 

This genus is proposed for what Mr. Coquillett has identified as 
Gcrdiopsis mexicana, represented by four male specimens from 
Organ Mountains, New Mexico, about 5,300 feet, September 4-5 
(Townsend), on flowers of Lippia wrightii; and two specimens, 
male and female, from Sierra Madre of western Chihuahua, head of 
Rio PiedrasVerdes, about 7,000 feet, July 19 (Townsend), on flowers 
of Rhus glabra. The Sierra Madre specimens are distinct from 
the Organ Mountains species, and probably represent two species. 
The male from the Sierra Madre, P. sicrricola, is made the type, 
and the Organ Mountains species is called P. orgaiicnsis. 

The genus differs from Phrissopolia chiefly in having the eyes 
pilose. The genus Gcrdiopsis differs from Clicctogcvdia chiefly in 
having the eyes hairy. 6. sctosa Coquillett belongs close to if not 
in the genus Poliophrys. 


Poliophrys sierricola, sp. nov. 

One male, Sierra Madre, Chihuahua, collected by Townsend. 

Length, 9 mm. ; of wing, y]/ 2 mm. Blackish, clothed with silvery 
cinereous. Parafrontals with a golden tinge (male), which extends 
very faintly on parafacials and cheeks. Facial plate silvery. 
Frontalia seen from in front silvery, seen from behind brownish. 
Antennae brownish, third joint blackish and three times as long as 
second, arista blackish. Palpi yellowish. Thorax silvery pollinose, 
with four moderately wide vittae, which are blackish seen from be- 
hind, but salmon colored seen from in front. Scutellum light 
brownish reddish, blackish at base, silvery. Abdomen silvery, 
slightly marmorate above, with a faint golden tinge which is strong 
on anal segment, where it is about the same as on parafrontals. 
Sides of abdomen faintly reddish under the pollen. Wings clear, 
tegulae whitish. Legs blackish, femora silvery ; tibiae reddish or 
brownish yellow, except at ends ; pulvilli slightly smoky. 

Type.— Cat. No. 10,908, U. S. N. M. 

A female from same locality differs only in second antennal joint 
being clear reddish yellow, the third joint little more than two and 
one-half times as long as second, and the sexual characters given 
above under the genus. It may prove to be a distinct species, but 
more material is needed to make sure of this. 

Poliophrys organensis, sp. nov. 

Four males, Organ Mountains, New Mexico, Townsend, Coll. 

Length, 8 to 10 mm. ; of wing, 6 to 7.5 mm. Front fully one- 
third head width at vertex, cheeks more than one-third eye height 
in width. Differs further from P. sierricola in having hardly a 
tinge of golden to the pollinose covering of head, which is silvery 
white throughout face and with only a slight tinge of golden on 
parafrontals not extending on facial plate at all. Antennae black, 
third joint nearly four times as long as second (male). Abdomen 
more noticeably reddish on sides ; anal segment less distinctly 
golden, about the same as other segments. 

Type.— Cat. No. 10,909, U. S. N. M. 

PHRISSOPOLIA, gen. nov. 

This genus is proposed for ProspJierysa crebra van der Wulp, 
which was included in Chcetogcedia by Brauer and von Bergen- 
stamm. It is characterized by a double row of frontal bristles, the 
outer row nearly or quite as strong as the other, and especially by a 
row of strong bristles on parafacials close to orbit, the facio-orbital 


bristles, of same strength as frontal bristles, and, except for their 
downward curve, appearing like a continuation of latter to lower 
eye border. The second aristal joint is long, the third much shorter 
than in Chcetogccdia, and the whole arista is widened and flattened, 
usually geniculate or subgeniculate. Eyes bare. 

Phrissopolia desertorum, sp. nov. 

Las Cruces, New Mexico, Cockerell, No. 4,952. Specimens from 
Beulah, New Mexico (Cockerell), and Santa Clara County, Cali- 
fornia, may also be referred to this species. 

Length, 9 to 10 mm. The species differs from van der Wulp's 
description of crcbra as follows: All the tibiae rufous (male) or yel- 
lowish (female). Face, including parafacials, silvery roseate white 
in male without yellowish tinge, which is confined to front ; in fe- 
male with a faint yellowish white tinge spreading over face. Third 
antennal joint of male four times as long as second, of female three 
times as long. Arista thickened nearly to end in both sexes, only 
the apical one-third or one-fourth appearing slender from certain 
viewpoints due to flattening; second joint very distinct, elongate, 
fully one-fourth as long as last joint, the articulation geniculate in 
some cases. Hind tibiae rather weakly ciliate, with a long bristle in 
middle. Wings faintly yellowish tinged at base. 

Type.— Cat. No. 10,910, U. S. N. M. (Las Cruces, N. Mex.). 

Genus Chaetogaedia Brauer and von Bergenstamm 

Chcetogccdia acroglossoides Townsend. — This is a good species. 
It is neither a Frontina nor a BanmJiaucria, but is apparently to be 
referred to Chcetogccdia. Frontina has the parafacials bare, and the 
second aristal joint is not elongate. Baumhaneria has the front 
greatly produced, the parafacials hairy and of exaggerated width, 
much wider than. the eyes, and the cheeks as wide as eye height. 

The identification of this species with Baumh. analis van der Wulp 
is quite out of the question, if the description agrees with the type. 
The second antennal joint is elongate, the third is not over four 
times as long in male and less than three times as long in female as 
second joint. The description was of the female. 

Chcetogccdia vilis van der Wulp, the type of the genus, has the 
frontal bristles in two rows, the outer row usually weaker, and the 
parafacials are clothed only with fine bristly hairs. 

Genus Gaediopsis Brauer and von Bergenstamm 

Gcediopsis cockerclli Coquillett. — This species appears to be cor- 
rectly referred to the genus Gccdiopsis. The material from which it 


was described was all collected by the writer in the White Mountain 
region of New Mexico (not New Hampshire, as given in the Cata- 
logue), at about 8,000 feet, on the head of Eagle Creek, a stream 
which takes its rise on the upper slopes of the peak known as Sierra 
Blanca (altitude, 10,050 feet). 

TREPOPHRYS, gen. nov. 

Head in profile almost half round. Antennae inserted about at 
eye middle. Front flattened, rounded in profile, showing just the 
same width beyond eye margins as do parafacials. Eyes bare, reach- 
ing quite to vibrissas. Cheeks very narrow, not over one-tenth of eye 
height. Front about one-third of head width, or slightly less, the 
inner outline of eyes but slightly divergent below base of antennae. 
Parafrontals a little wider than frontalia, parafacials gradually 
narrowing from base of antennae until they become almost linear 
at lower eye margin. Ptilinal suture inverted V-shaped, the median 
angle a little rounded. Ptilinal area elongate, about one-third head 
width below. Facial plate elongate, not narrowed below, depressed, 
with a distinct and sharp but low median carina full length, not 
produced at lower margin. Facialia edge-like, bristly more than 
half way up, vibrissal angles hardly perceptible. Vibrissae inserted 
close to oral margin, well developed. 

Frontal bristles in a single row close to frontalia and extending 
back to ocelli, all curved inward, more or less decussate, descending 
in front to insertion of arista. The usual strong inner and weak 
outer vertical bristles. Two upper reclinate fronto-orbital bristles 
set well forward, almost far enough forward to occupy the usual 
place of insertion of the middle or proclinate ones. These two 
fronto-orbital bristles are of exactly the same strength, length, 
curvature, and direction as the inner vertical bristle, and look like 
two replicas of it in profile. They are also quite in line with it, and 
the three in profile are seen to be an equal distance apart. Two 
proclinate middle fronto-orbital bristles in female, outside the upper 
ones ; none in male. 

Proboscis short and fleshy, palpi slender and normal. Second an- 
tennal joint about twice as long as first; arista indistinctly jointed 
and minutely pubescent, slightly thickened on basal one-third. Third 
antennal joint about two and one-half times as long as second. 
Occiput slightly swollen behind on lower one-fourth, the lower mar- 
gin of head short, long axis of head at vibrissae but little over one- 
half that at base of antennae. 

Three sternopleural bristles, 1. 1. 1, the middle one weakest, the 
posterior one strongest. Three postsutural bristles. Scutellar 


bristles in five pairs, one decussate apical, three lateral, of which 
anterior and posterior are longer and stronger, and one separated 
discal pair. 

Abdomen above of four visible segments, macrochaetae only mar- 

Wings reaching well beyond end of abdomen, apical cell narrowly 
open just before and almost in wing-tip, ultimate section of fourth 
vein bowed in about the middle so as to attenuate the terminal por- 
tion of apical cell. Hind crossvein slightly curved, not quite in 
middle between small crossvein and bend of fourth vein, distinctly 
nearer latter. None of the veins spined, the small crossvein slightly 
or distinctly before middle of discal cell. 

Legs normal, hind tibiae weakly ciliate, with a bristle or two 
among the cilia, claws and pulvilli very short. Type, T. cinerea, 
n. sp. 

Comes near Pseudochccta Coquillett. with which it agrees in the 
arrangement of the upper and middle fronto-orbital bristles, and 
from which it differs as follows : 

Pseudochccta Trepophrys 

Antennae inserted distinctly above Antennae inserted practically on eye 

eye middle in both sexes, but es- middle in both sexes, 

pecially so in the male. Three sternopleural bristles. 

Two sternopleural bristles. Three postsutural bristles. 

Four postsutural bristles. Apical cell ending slightly before, 

Apical cell ending not far but very almost in, wing-tip. 

distinctly before wing-tip. ' Hind crossvein distinctly nearer bend 

1 Hind crossvein almost in middle, of fourth vein, male ; almost in 

male; nearer bend of fourth vein, middle, female, 

female. ' Small crossvein well before middle 

'Small crossvein nearly in middle of of discal cell, male; nearly in mid- 

discal cell, male; before middle, die, female. 

female. Wings elongated beyond end of ab- 

Wings very short and broad, hardly domen. at least two and one-half 

more than twice as long as wide. times as long as broad. 

Head distinctly widened. Head hardly wider than thorax. 

Trepophrys cinerea, sp. nov. 

Three specimens bred from masses of pupae of a tortricid, Amy- 
dria sp., sent by Prof. A. L. Herrera, from Cuernavaca, Mexico. 

Length, 4.5 to nearly 6 mm. Blackish, parafrontals and para- 
facials golden, extending on occipital orbits ; frontalia, antennae, and 
face blackish, latter with a slight silvery reflection. Cheeks slightly 
golden. Pleurae very faintly silverv. Dorsum of thorax and abdo- 

1 These characters represent the average of the specimens. 


men cinereous pollinose, with a distinct golden tinge most noticeable 
on scutellum. First abdominal segment, narrow hind margins of 
second and third, and apical half of anal blackish, the black surface 
of anal segment shining. Wings clear, tegulae whitish with a tawny 
tinge. Legs black. 

Type.— Cat. No. 10,911, U. S. N. M. 

Subfamily MasicEratin^ 
Genus Exorista Meigen and allies 

The genus Exorista, as restricted, has the cheeks wide, one-third 
to one-half eye height; second antennal joint somewhat elongate, 
and abdominal macrochaetae discal and marginal. 

Par exorista differs from Exorista in cheeks being not over one- 
fourth eye height, second antennal joint not elongate, second aristal 
joint usually elongate ; abdominal macrochaetae usually only mar- 
ginal, but long discal bristles present, those on third segment ap- 
proaching macrochaetae in character. 

The genus Carcelia Robineau-Desvoidy may be known by hav- 
ing no long discal bristles on abdominal segments, all being short 
and of even length. Macrochaetae only marginal. Cheeks not over 
one-fourth eye height. Type, gttava Meigen. 

Nemorilla has cheeks not over one-fourth eye height, second an- 
tennal joint elongate, second aristal joint short, macrochaetae discal 
and marginal, hind tibiae weakly ciliate. 

EUSISYROPA, gen. nov. 

Proposed for Exorista blanda Osten-Sacken. Differs from Parex- 
orista in possessing regularly arranged discal macrochaetae on ab- 
dominal segments, without the erect and usually long bristly hairs 
of that genus, and especially in the peculiar form of the abdomen in 
both sexes. The latter is high, somewhat arched, slightly wedge- 
shaped ventrally in female, and obliquely truncate downward and 
forward at apex in profile. The female has these abdominal charac- 
ters more marked, but the male also possesses them in a hardly less 
degree. The female has the venter quite distinctly carinate. 

Eusisyropa blanda Osten-Sacken.- — This species has the legs and 
second antennal joint more or less deeply blackish or brownish, 
with usually only faint suggestions of yellowish or reddish. The 
palpi have a reddish tinge. Both sexes have the parafrontals with 
a slight golden tinge, and anal segment very distinctly golden. 
There are two sternopleural bristles only. 


New Jersey, New York, Massachusetts, and south to District of 

This species has been bred from Cymatophora pampinaria Guenee, 
one specimen issuing from a larva collected on cranberry at Cotuit, 
Massachusetts, by J. B. Smith (Riley Notes, Bureau of Entomol- 
ogy) ; also from Hyphantria textor, at Washington, D. C. (No. 78 03 
Riley Notes) . 

One female specimen was bred at the Gipsy Moth Parasite Labo- 
ratory, North Saugus, Massachusetts, issued July 29, 1907, which 
may have come from native Buproctis chrysorrhoca. 

Eusisyropa boarmicc Coquillett. — This is a Florida and Southern 
States form closely allied to blanda. It reaches Arkansas and Mis- 
souri. It has light reddish yellow legs and second antennal joint, 
these being quite concolorous with the reddish yellow palpi, and 
possesses a small third sternopleural bristle. 

The type specimen was bred from a larva of Aletia argillacea, 
received from Oxford, Mississippi, issued November 14, 1882 (No. 
468 L°, Riley Notes, Bureau of Entomology). The species has not 
been bred from Cymatophora (Boarmia), the Boarmia-bred speci- 
men mentioned by Coquillett being B. blanda. 

Genus Argyrophylax Brauer and von Bergenstamm and allies 

The following is a table of Argyrophylax and the forms closely 
related to it : 

1. Ocellar bristles wanting (type, albincisa Wd.) Argyrophylax B. B. 

Ocellar bristles present 2 

2. Apical pair of scntellar bristles much stronger tban those next to them, 

hind tibial cilia dense (type, pupiphaga Rdi.) Sturmia R. D. 

Apical pair of scutellar bristles much weaker than those next to them. . 3 

3. Third abdominal segment of male with two shining or pilose black spots 

1 n ventral surface, a distinct longer bristle in cilia of hind tibiae near 

middle ("type, bimaculata Htg.) Zygobothria Mik. 

Third abdominal segment of male without such spots, hind tibial cilia 
dense and without longer bristle (type, scutella ta Rdi.) .. .Blepharipa Rdi. 

Argyrophylax piperi, nom. sp. nov. 

This name is proposed for Sturmia schizurce Coquillett, which is 
an Argyrophylax. The specific name is preoccupied by Argyr. 
schizurce Townsend. 

Pullman, Washington State (Piper). Bred from Schizura ipo- 

Length, 10.5 mm. Much larger than type of A. schisurce Town- 
send, with which it at first seemed identical. Agrees with descrip- 


tion of A. schizurce Townsend except as follows: The facial plate 
shows no appreciable tinge of golden, fourth vein is quite abruptly 
bent, first abdominal segment has one lateral marginal macrochaeta, 
second segment has a lateral marginal pair of macrochaetae, pulvilli 
are smoky blackish, and size is larger. 

Genus Zygobothria Mik. 
Zygobothria nidicola, sp. nov. 

Male. — Fifteen specimens. Thirteen bred at the Gipsy Moth 
Parasite Laboratory, North Saugus, Massachusetts, as follows : 
Four bred by E. S. G. Titus, in 1906, from Buproctis chrysorrhoea 
imported from Germany (Erfurt, Munich, and Fuhlsdorf, received 
from Marie Ruhl) ; nine bred by W. F. Fiske, in 1907, from hiber- 
nated larvae of Buproctis chrysorrhoea from imported nests received 
from Vienna and other parts of Nieder-Oesterreich, and from sum- 
mer importations of larvae of same species from South Tirol and 
Carniola. Two bred at Simferopol, Russia, by S. Mokschetsky, 
from Buproctis chrysorrhoea, June 7, 1905, and July, 1907. 

Length, 7 to 9 mm. Eyes very faintly hairy, appearing bare. 
Antennae blackish, third joint more or less reddish or lighter colored 
at base, palpi light yellow. Face and front silvery, parafrontals 
darker in some lights, but not golden. No middle fronto-orbital 
bristles. Frontalia slightly, if any, wider than one parafrontal. 
Parafrontals with fine hairs outside the row of frontal bristles. 
Front anteriorly hardly as wide as one eye, half as wide as one eye 
at vertex. Facialia with some bristles extending about one-third 
way up. Arista thickened on less than proximal half, first two 
aristal joints short. Cheeks hardly one-third eye height. Thoracic 
dorsum thinly silvery pollinose. Scutellum testaceous except ex- 
treme base, apical pair of bristles weak, almost erect, decussate. 
Abdomen with more or less red on sides, first segment and narrow 
hind borders of second and third segments shining black, rest 
thickly cinereous pollinose leaving a more or less distinct median 
line. A median marginal pair of macrochaetae on first and second 
segments, also three lateral marginal ones on each side of same seg- 
ments, third segment with a marginal row of twelve or fourteen. 
Anal segment with only bristly hairs. Legs wholly black, claws 
and pulvilli very elongate, hind tibiae thickly ciliate, but with a 
longer bristle near middle. Tegulae white. Four sternopleural and 
four postsutural bristles, a few specimens showing a fifth weaker 
sternopleural bristle. 


A specimen of this series sent to Dr. K. Kertesz, at Budapest, was 
returned by him as Argyrophylax galii Brauer and von Bergen- 
stamm. As galii has the male vertex one and one-third times the 
eye width and female vertex twice the eye width, this can not be 
that species. Two specimens sent to Dr. A. Handlirsch, at Vienna, 
•were returned as unknown to him, and indicated with a query as 

Female. — Fourteen specimens. Twelve bred at the Gipsy Moth 
Parasite Laboratory. North Saugus, Massachusetts, as follows: Five 
bred by E. S. G. Titus, in 1906, from Euproctis clirysorrluva im- 
ported from Germany (Baden and Dresden, received from Marie 
Ruhl and Schopfer, respectively ) : seven bred by W. F. Fiske, in 
1907, from summer importations of Euproctis clirysorrluva from 
Germany. Two bred at Simferopol, Russia, by S. Mokschetsky, 
from Euproctis clirysorrluva, June 10, 1905, and July, 1907. 

Length. 7 to 8 mm. Differs from the male as follows: Thickly 
yellowish-cinereous pollinose all over, including front and- first ab- 
dominal segment. Face more silvery. Thoracic vittae fine, outer 
ones broken at suture and somewhat widened. Scutellum yellowish 
on margin. Middle fronto-orbital bristles two in number. Front 
from more than one-third to about two-fifths width of head, hardly 
narrowed from facial width. Abdominal macrochaetae same as male, 
but the second segment rarely has four median marginal macro- 
chaetae more or less well developed from the long marginal hairs on 
each side of the original pair. Hind tibiae sparsely but distinctly 
ciliate, a long" bristle near middle. Four sternopleural and four 
postsutural bristles. 

A specimen of this series sent to Dr. K. Kertesz was returned un- 
determined ; another sent to Dr. A. Handlirsch was returned as 
unknown to him, and probably American. The two sexes were not 
suggested by either Kertesz or Handlirsch as belonging together, 
but it seems highly probable that they are the same species. Both 
are positively European, as conclusively demonstrated not only by 
the breeding records of the Gipsy Moth Parasite Laboratory, but 
also by Mr. Mokschetsky's breeding of both at Simferopol, Russia. 

Types. — Cat. No. 11,803, U. S. N. M. (2 types: male from Nieder- 
Oesterreich, issued July 20. 1907: female from Central Europe, 
issued July 10, 1907). 

Mr. W. F. Fiske has bred this species (male specimens) from 
cages containing hibernated larvae of Euproctis chrysorrha?a under 
circumstances indicating that the female tachinids oviposit in the 
Euproctis nests in the fall, the tachinid larvae remaining through the 
winter in the nests and issuing from the host larvae or pupae in the 


summer. This is a remarkable habit of oviposition among tach- 
inids, and credit is due to Air. Kiske for the discovery of it. 

Genus Comatacta Coquillett 
Comatacta nautlana, sp. nov. 

The material upon which the genus Comatacta Coquillett was 
founded was collected at San Rafael, near Jicaltepec, Veracruz 
(Townsend). The specimens were erroneously identified with 
Brachycoma pallidula van der Wulp (Can. Ent, 1902, pp. 199-200), 
which is to be considered the type of the genus Comatacta. 

The present species differs from van der Wulp's description of 
pallidula as follows : Facial plate silvery like parafacials ; frontalia 
honey yellow, parafrontals silvery with a golden shade. Frontal 
bristles descending but one or two below base of antennae. Beard 
very short, grayish. Antenna? reaching two-thirds to three-fourths 
way to oral margin. Arista rufous, concolorous with antennae. 
Anal segment hardly at all rufous. 

Type.— Cat. No. 10,906, U. S. N. M. 

PARADEXODES, gen. nov. 

Wings longer than Dexodcs (type spcctabilis Meigen), abdomen 
very bristly like Dexodcs, with many discal macrochaetae and erect 
hairs, apical pair of scutellar bristles weak but long and markedly 
divaricate. Eyes bare ; male front narrow, about or nearly equal- 
ing eye width anteriorly, vertex about or more than one-half eye 
width. Frontal bristles in male closely placed, descending three to 
four below base of antennae. Ocellar bristles present. Vibrissas 
close to oral margin, facialia with a number of bristles above 
vibrissae. Parafacials quite narrowed below, widening above. An- 
tennas inserted on eye-middle. Second antennal joint nearly three 
times as long as first; third joint narrow, two or. more times as long 
as second, equilateral in profile, subtruncate at tip. Second aristal 
joint short but distinct, arista thickened on proximal fourth. Legs 
rather long, male claws and pulvilli elongate. Wings long and nar- 
rowed in male, apical cell open well before wing-tip, hind crossvein 
approximated to bend of fourth vein, small crossvein on middle of 
discal cell. Abdomen elongate, conico-cylindrical in male. Type, 
the following species : 

Paradexodes aurifrons, sp. nov. 

One male, North Saugus, Massachusetts (Gipsy Moth Labora- 
tory, bred in Cage E, 14 July, 1906. No. 698, E. S. G. Titus. Host 


Length, 10 mm. Blackish, gray pollinose. Entire face and front, 
even including cheeks and orbit, deep golden pollinose. Frontalia 
quite black, nearly as wide as one parafrontal. Antennae blackish. 
Palpi reddish yellow. Thorax and scutellum very thinly pollinose, 
humeri thickly so. Abdomen quite uniformly pollinose except first 
segment, thickly bristly and hairy. Legs entirely black, femora 
pollinose on under side. Claws long and black, pulvilli smoky. 
Tegulae whitish, with narrow yellowish edge. 

Type— Cat. No. 11,686, U. S. N. M. 

Paradexodes albifacies, sp. nov. 

One male, White Mountains, New Hampshire, Morrison. This 
specimen is figured in Dr. Howard's Insect Book, pi. 22, fig. 7, as 
Hypostcna variabilis Coquillett, but is not congeneric with the type 
of that species. 

Length, 9.5 mm. Face and front, cheeks and orbits silvery white. 
Frontalia reddish brown, wider than one parafrontal. Antennae 
reddish brown. Palpi yellow. Thorax, scutellum. and abdomen 
shining black, very thinly bluish silvery pollinose, most thickly so on 
humeri and bases of last three abdominal segments. Legs blackish 
brown, femora silvery beneath, pulvilli yellowish white, claws pale 
reddish brown. Tegulae white, with narrow yellowish edge. 

Type— Cat. No. 11,687, U. S. N. M. 

Genus Ceromasia Rondani 
Ceromasia aurifrons, sp. nov. 

Three females and one male. New Hampshire (2 females and the 
male from Canobie Lake, Dimmock). 

Length, 7.5 to 10 mm. Differs from the European C. florum 
Meigen (determined by Brauer and von Bergenstamm) by having 
whole of parafrontals, parafacials, and orbits deep golden in both 
sexes, even the cheeks showing golden in fresh specimens ; the pollen 
of thorax and abdomen whitish gray, without the brassy tinge of 
florum ; anal segment in both sexes with a noticeable tinge of golden, 
and scutellum testaceous only on apical half. 

Type.— Cat. No. 11,649, U. S. N. M. (female). 

Two males of this species were bred at the Gipsy Moth Parasite 
Laboratory, North Saugus, Massachusetts, by E. S. G. Titus, from 
unidentified lepidopterous larvae. 

Ceromasia auricaudata, sp. nov. 

Two females and one male, Harrison, Idaho (male and female), 
and Pullman, Washington (female, July 16). 


Length, 7 to 9 mm. Differs from C. aurifrons Townsend by hav- 
ing the anal segment wholly deep golden, same shade as para- 
frontals, etc. ; humeri with a faint, abdomen with a more distinct 
golden tinge, scutellnm hardly more narrowly testaceous, and thorax 
more distinctly vittate. 

Type. — Cat. No. 11,650, U. S. N. M. (female, Harrison, Idaho). 

EUDEXODES, gen. nov. 

This genus is proposed for Dcxodes eggeri Brauer and von Ber- 
genstamm, of Europe. The characters of the facial plate throw the 
species into a different tribe (if not subfamily) from D erodes, of 
which the type is spectabilis Meigen. 

Subfamily Willistoniin^e 
Genus Belvosia Robineau-Desvoidy and allies 

Dr. Williston published a plate of Belvosia and allies in Insect 
Life, vol. v (1893), facing p. 238, exhibiting the difficulties to be en- 
countered in separating the forms. By studying this plate, it will 
be seen that there is a correlation between length of second antennal 
joint and bristles on the facialia, also between former and distance 
of vibrissa? from oral margin. 

The more elongate the second antennal joint is, the less bristles 
there are on the facialia. Conversely, the shorter the second joint, 
the more strongly are the facialia ciliate. In all cases, the distance 
of the vibrissa? above the oral margin is about equal to the length of 
the second antennal joint. 

The forms having facialia not ciliate 'have the second antennal 
joint long, vibrissa? inserted far above oral margin, and fourth vein 
angular at bend. Those having facialia ciliate have the second joint 
much shorter and vibrissa? inserted only a little above oral margin ; 
they fall into two categories by the character of the bend of fourth 
vein. We thus have the following table : 

1. Facialia not ciliate, fourth vein angular at bend; second antennal joint 

strongly elongate, nearly as long as third joint; vibrissa? inserted high 
above oral margin, male claws normally very elongate, female claws 
less elongate. (Sto. Domingo, Brazil, California, New Mexico, 

Mexico, Jamaica.) Belvosia bicincta Robineau-Desvoidy 

Facialia ciliate 2 

2. Fourth vein bent at a sharp angle, with or without stump, but often 

V-shaped and with stump; claws of male normally elongate, of female 
not; second antennal joint not strongly elongate, vibrissa? inserted 
normally above oral margin. (Brazil.) . . . Willistonia esuriens J. C. Fabricius 


Fourth vein rounded at bend, second antennal joint short or only appre- 
ciably elongate ; vibrissse inserted close on or only appreciably above 
oral margin, about as far above as length of second antennal joint; 
male claws usually not elongate, but longer than the short female 
claws Latreillimyia, nom. nov. ( Latrcillia preocc.) 3 

3. Vibrissse inserted close on the oral margin 4 

Vibrissas inserted appreciably above the c ral margin, male claws 

strongly elongate. (Brazil, Pennsylvania.) L. (aberrant form) 

4. Second antennal joint appreciably elongate, female claws rather short. 

1 Minnesota.) L. ( intermediate form) 

Second antennal joint short, not at all elongate. ( Brazil. Mexico, New 

V rk.) Latreillimyia (typical forms) 

bifasciata J. C. Fabricius, leucopyga (van der Wulp) Williston. 

The character of the ciliate faeialia is more important than the 
venational character and the same holds good of the vibrissal charac- 
ter and the elongation of second antennal joint. As already pointed 
out in a previous section of this paper, the relative length of the sec- 
ond and third antennal joints will not hold for generic separation, 
since the length and size of the third joint in these flies is largely a 
sexual character. But the actual length of second joint taken inde- 
pendently furnishes a good character. Only a few genera have the 
second joint elongate. It may be compared in length with the first 
joint. The first two joints do not vary sexually. 

Brauer and von Bergenstamm state that the claws of male are 
elongate in Willistonia and short in Latreillimyia. It is doubtful 
how far these characters can be relied upon, since they are also 
sexual. The same authors also give as a character of Willistonia a 
stump at the angular bend of fourth vein and the angle more ap- 
proximated to hind margin of wing. These may hold good, espe- 
cially the latter, but are not necessary for the separation of the forms 
at present known to us. Further material will probably call for 
their use. 

The writer pointed out in 1892 (Trans. Am. Ent. Soc, xix, p. 
89) that bifasciata has the faeialia strongly ciliate and bicincta has 
not; that five specimens of bicincta from New Mexico had the third 
antennal joint scarcely longer than the second, which means that the 
second joint was strongly elongate; that three specimens from New 
York were easily referable to bifasciata, and one from Jamaica to 
bicincta; and that, while the parafacials are bare in both species, the 
whole anterior aspect of head is altogether more bristly in bifasciata, 
which possesses also greater hairiness of cheeks. 

The elimination of Bclvosia, argued for by Brauer and von Ber- 
genstamm, is not permissible under the rules of the International 
Code. Its maintenance fortunately does not conflict with the genus 
JJlllistonia, since bicincta differs generically from csitricns. 


Bclvosia and Latreillimyia show no ventral plates or ventral mem- 

LATREILLIMYIA, nom. gen. nov. 

This name/is proposed for Latreillia Robineau-Desvoidy (1830), 
which is preoccupied by Roux in Crustacea ( 1827) . 

GONIOMIMA, gen. nov. 

This genus is proposed for Bclvosia lutcola Coquillett. Bears a 
striking resemblance to Gonia. Second antennal joint short, third 
joint very long and narrow; arista long and flattened whole length, 
in front view appearing as a mere line, but in lateral view showing 
itself to be uniformly widened nearly to apex ; frontal bristles in one 
main inner row bordering frontalia, with a row of weaker bristles 
outside, and orbital bristles (female) outside these; second aristal 
joint very short, front not widened and swollen, facialia ciliate 
almost to base of antennae. Abdomen appearing conical from above, 
but laterally appressed on apical portion, fully as thick dorso- 
ventrally for its whole length as its greatest width, which is at base. 
The body and wing characters agree perfectly with Gonia, but the 
head characters are totally different, and it is the latter which place 
the genus in the Willistoniinse. 

The genus appears to come near Thelymorpha Brauer and von 
Bergenstamm, but is at once distinguished by having no discal 
macrochsetse on abdomen. The head is almost the same, and the 
abdomen is described as very similar. 

TRIACHORA, gen. nov. 

This genus is proposed for Latreillia unifasciata Robineau-Des- 
voidy, of which Exorista flavicauda Riley is a synonym. Differs 
from Latreillimyia in having three rows of frontal bristles on each 
side of frontalia. besides the fronto-orbital bristles of female. The 
arista is flattened, and the antennal characters are similar to those of 
Goniomima. The main or strongest row of frontal bristles, of the 
three rows on each side, is in the middle, the inner row being de- 
cidedly weaker, and the outer row but little weaker than the main or 
middle row. 

Genus Rileymyia Townsend 

(Ent. News, 1893. p. 277) 

This name was proposed by the writer in 1893 for Rileya Brauer 
and von Bergenstamm, which is preoccupied in Hymenoptera. The 
type of the genus is Blepliaripeza fulvipes Bigot, according to Brauer 


(Sitzungsber. Math.-Naturwiss. CI. k. Akad. Wiss., cvi, L p- 348), 
who says that R. americana Brauer and von Bergenstamm is a 
synonym of Bigot's species. B. adusta H. Loew is also typical of 
the genus, which may be distinguished from Blepharipeca by the ab- 
sence of apical scutellar bristles and thornlike macrochjetae. 

Riley myia albi fades Bigot. — Brauer (1. c.) says this is a synonym 
of fulvipes Bigot, but in view of the widely removed type localities 
it would seem that the point needs verification. R. aibifacies was 
founded on a specimen from Brazil, while fulvipes is from Washing- 
ton State. R. americana is from California. 

Subfamily Meigeniin^; 
Genus Viviania Rondani 

It must be noted that this genus is characterized quite fully by 
Rondani on p. 53, vol. iv, of Dipt. Ital. Prod., where the imperfectly 
erected Biomyia (1. c, vol. 1, p. 72) is given as a partial synonym. 
Biomyia does not cover the same forms, so far as any one knows, 
and its one-line characterization entitles it to no notice in the face of 
its author's subsequent rejection of it. It is therefore quite out of 
the question to attempt to use it, especially since we have no defini- 
tion of it. 

Viviania mutabilis Coquillett, etc. — Biomyia mutabilis is a Viv- 
iania. So also is B. aurigera Coquillett. B. genalis Coquillett does 
not belong anywhere near this genus. 

Viviania lachnosternae, sp. nov. 

One female, Urbana, 111. (No. 36,817, Forbes). "Supposed to 
have bred in Lachnosterna adults." 

Length, 10 mm. ; of wing, 8 mm. Gray-cinereous, more or less 
silvery. All three antennal joints and arista wholly reddish yellow, 
the frontalia same color posteriorly, but darker in front. Para- 
frontals blackish, thinly silvery. Parafacials more distinctly silvery. 
Ptilinal area blackish, the lower portion of facial plate broadly yel- 
lowish on oral margin. Palpi light reddish yellow. Mesoscutum 
with five vittse, the middle one narrow behind and obsolete in front 
of suture, the outer ones more or less triangularly widened, shorter 
and more triangular in front of than behind suture. Wings wholly 
hyaline, tegulse white. Legs wholly blackish, femora silvery below, 
pulvilli smoky. Abdomen black, thickly cinereous pollinose. 

Type.— Cat. No. 10,913, U. S. N. M. 


Genus Tachinomyia Townsend 

Tachinomyia robusta Townsen'd. — The genus differs from Tachina 
in the vibrissa; being inserted higher above oral margin, cheeks one- 
half eye height in width, and abdomen very elongate. 

Genus Emphanopteryx Townsend 

Bwphanopteryx eumyothyroides Townsend. — This genus differs 
from Cryptomeigcnia by having the abdomen large; claws and pul- 
villi of female elongate, those of male very long and strong; arista 
finely pubescent, strong subdiscal and discal macrochaetae (at least 
on second segment), fourth vein rather angular at bend and usually 
represented beyond apical crossvein by a short stump. 

Subfamily Tachixix/k 
Genus Tachina Meigen 

Tachina clisiocampcc Townsend. — -Achcetoneura femaldi Williston 
is very probably a synonym. The strongly marked wrinkle at an- 
gular bend of fourth vein, the elongate second antennal joint which 
is about three times as long as first, the frontal bristles descending 
low on parafacials, and the strongly ciliate facial ridges, whose 
bristles ascend at least to opposite the lowest frontal bristles, make 
the species typical of Tachina s. str. The third antennal joint is 
about twice as long as second. Achcetoneura has the second anten- 
nal joint hardly longer than the first, the third joint thus being easily 
five or six times as long as the second. 

This species can not be identified with T. mella Walker, if the de- 
scription of latter is to be depended on, since it states that the 
second antennal joint is ferruginous apically, third joint three times 
as long as second, arista much longer than third antennal joint, 
second aristal joint moderately long, large ferruginous spot on each 
side of second segment. The venational characters also do not 
agree. In clisiocampce there is at most in either sex only a very 
faint tinge of reddish, hardly perceptible in fact, on sides of second 
abdominal segment. The antennae are wholly blackish, the third 
joint, even in male, hardly more than twice as long as second, arista 
but little longer than third antennal joint, apical crossvein well 
bowed in, hind crossvein quite strongly sinuate. 

Tachina orgyiarum, nom. sp. nov. 

This name is proposed for T. orgyice Townsend, which is preoccu- 
pied by T. orgyice Le Baron. Both species belong in the genus 
Tacliina, as here restricted. 


Tachina utilis, sp. nov. 

Length, 6 to 8 mm. Differs from T . larvarum in it? much smaller 
size, vertex of male not wider than one eye, thoracic dorsum not so 
thickly pollinose, abdomen more shining', with pollinose bands not 
so distinct, and anal segment not thickly hairy. The male in some 
specimens shows signs of reddish on sides of abdomen. 

The anal stigmata of puparium also show important differences, 
the median slit being much abbreviated. 

Germany, Bavaria, and Carniola. Bred at the Gipsy Moth Para- 
site Laboratory, North Saugus. Massachusetts, by E. S. G. Titus, in 
1906, and W. F. Fiske, in 1907, from both Euproctis chrysorrhoea 
and Porthetria dispar larvae received as summer importations from 
above localities; and also bred by W. F. Fiske. in 1007, from native 
larva? of both species collected in field colonies near Boston (Oak 
Island and Woburn), Massachusetts, where European specimens of 
this tachinid had been previously liberated, showing that this species 
has gained a foothold. 

Type. — Cat. Xo. 11.804, U. S. X. M. (male, length 6 mm.; Dres- 
den, Germany, from Euproctis larva? collected and shipped by 

This type specimen was submitted to Dr. K. Kertesz, and by him 
determined as Tachina gldssatorum Rondani. It can not be that 
species, which is described by Rondani as having the second aristal 
joint four times as long as wide, and belongs to the genus Micro- 
tachina established on that character. Tachina, including the pres- 
ent species utilis. has the second aristal joint no longer than wide. 

Genus Euphorocera Townsend 
Euphorocera slossonae, sp. nov. 

One female, Franconia, X. H. (Mrs. A. T. Slosson). Syn. E. 
cinerea Coquillett (non van der Wulp), Rev. Tach., p. 102. 

Differs from van der Wulp's description of Phoroccra cinerea 
(Biol. C. A., Dipt., ii, pp. 81-82) as follows: Frontalia as broad or 
broader than the parafrontals. Lowest frontal bristles not close to 
the eyes. Face very distinctly yellowish. Second antennal joint 
two and one-half times as long as first, the third joint a little more 
than twice as long as second. Arista thickened on basal third only. 
Palpi somewhat swollen, evenly clothed with black hairs. No trace 
of dorsal stripe on second and third abdominal segments. Two 
discal macrochaetse on second segment as well as on third. Anal 
segment only moderately beset with bristles. Small crossvein 


slightly before the middle of discal cell. Fourth vein bent at an 
obtuse angle. Posterior crossvein gently bisinuate. 
Type.— Cat. No. 10.912, U. S.N. M. 

Subfamily Echixomyiin.e 

Genus Varichaeta Speiser 

The name Variclurta has been proposed by Speiser for Brigone 
Robineau-Desvoidy (1830), which is preoccupied by Savigny in 
Arachnida (1827). The type species is V. radicum Fallen. 

Varichaeta aid rich i Townsend.- — This species, described under 
Hystricia, belongs in the genus Brigone (Robineau-Desvoidy) 
Brauer and von Bergenstamm. and must thus be known as J ari- 
chcuta aldrichi. It is quite distinct from V. radicum. The latter has 
only three postsutural macrochsetge, while aldrichi has four or five. 
There are also differences in the abdominal macrochsetse. 

Genus Elachipalpus Rondani 

This genus is characterized by Rondani as possessing palpi, 
though small ; and having apical cell appendiculate by reason of the 
continuation of fourth vein beyond apical crossvein. The type 
cited for it by Rondani is Micropalpus longirostris Macquart. from 
the Cape of Good Hope. The species is figured by Macquart as 
having a proboscis like Spanipqlpus, but with distinct filiform palpi, 
and venation like Spanipalpus and Deopalpus, except that, instead 
of a wrinkle, there is a distinct stump representing fourth vein be- 
yond apical crossvein. Brauer and von Bergenstamm indicate B. 
longirostris Rondani as type of Blachipalpus, but throw doubt on 
Rondani's longirostris being the same as Micropalpus longirostris 
Macquart. However this may be, it is certain that the American 
species ruiicauda van der Wulp and macrocera Wiedemann do not 
belong to Blachipalpus, since they have absolutely no palpi, the pro- 
boscis is much shorter, and the venation markedly different. The 
new genus Copccrypta is therefore proposed for Schincria ruiicauda 
(van der Wulp) Williston. The species was referred to Cuphocera 
by Williston. 

COPECRYPTA, gen. nov. 

Distinguished by a characteristic narrowing of the apical cell at 
the end, the ultimate section of fourth vein being crookedly bowed 
in and for the last one-third or one-fourth of its extent parallel with 
the third vein and very closely approximated to it, thus forming a 
narrow handle-like tip to the apical cell. The proboscis beyond 


geniculation is shorter than head height. Palpi absent. Two orbi- 
tal bristles in female, none in male. Some extra bristles outside the 
frontal row, but these do not form a definite second row except 
anteriorly in some males. No ocellar bristles. Claws of female 
short, those of male as long as last tarsal joint. 

The genus differs from Trichophora by having the abdomen elon- 
gate, subcorneal or subcylindrical, reaching nearly to end of wings. 
Trichophora has abdomen much shorter than wings and rounded. 

SPANIPALPUS, gen. nov. 

This genus is proposed for Trichophora miscelli Coquillett. It 
differs from Copecrypta in possessing a strong pair of ocellar 
bristles ; proboscis long and slender, much longer than head height ; 
abdomen considerably widened (female). Male not known. Fe- 
male with two strong orbital bristles; only one row of frontal 
bristles ; inner pair of vertical bristles very long and strongly 
curved, decussate, reclinate. Apical crossvein normal, not crooked, 
evenly bowed in near origin ; apical cell widely attenuate on terminal 
portion, widely open. A distinct wrinkle at origin of apical cross- 

DEOPALPUS, gen. nov. 

Differs from Spanipalpus only as follows: No ocellar bristles. 
Two very definite rows of frontal bristles on each side of frontalia. 
No orbital bristles (male), claws of male not elongate. Parafacials, 
parafrontals, and cheeks evenly and thinly pilose with rather long 
fine black hairs. Parafrontals not metallic or blackish, silvery 
white. Venation and proboscis like Spanipalpus. Abdomen about 
like Copecrypta. The head bristles, like those of all the rest of the 
body, are strong. The inner frontal rows are decussate, extending 
only half way back between base of antennas and vertex. The outer 
row on each side is composed of lightly reclinate bristles of nearly 
equal strength, nearly as strong as the vertical bristles. Both rows 
descend well below base of antennae, the outer row slightly lower 
than the inner and to base of third antennal joint. Two facio- 
orbital bristles as strong as the frontal bristles. Facial plate strongly 
produced below. Second antennal joint elongate, about as long as 
third. Second aristal joint strongly elongate, slightly geniculate. 
Cheeks nearly equal to eye height. Type, the following species : 

Deopalpus hirsutus, sp. nov. 

One male. Meadow Valley, head of Rio Piedras Verdes, about 
7,300 feet. Sierra Madre of western Chihuahua, July 29 (Town- 


Length, 9.5 mm. Bears considerable superficial resemblance to 
Copecrypta ruiicauda, but may be distinguished therefrom by the 
generic characters above given. Head entirely silvery white, fron- 
talia showing very faintly pale brownish, first two antennal joints 
light brownish yellow, third joint hardly darker, but with anterior 
terminal border and arista blackish. Proboscis black, shining. 
Thorax cinereous pollinose, with two interrupted heavy outer dark 
vittae, and two narrow inner vittae stopping a little behind suture. 
Scutellum tawny yellowish, darker at base, silvery, with two very 
strong pairs of lateral macrochaetae reaching beyond base of third 
abdominal segment, a moderately strong but shorter apical decussate 
pair, and two lateral weak pairs besides discal bristles. Abdomen 
faintly blackish on dorsum, pale reddish or brownish yellow on 
sides, anal segment wholly reddish. All of abdomen more or less 
thickly silver}- pollinose. showing most on basal half or more of last 
three segments. Macrochaetae as follows : One lateral marginal on 
first segment ; one lateral marginal, and one median marginal pair 
on second segment ; eight strong marginal in a row on third seg- 
ment ; anal segment with about twenty in marginal, submarginal, 
and discal rows. Legs black, tibiae reddish, especially hind ones, 
pulvilli only faintly smoky. Wings clear, tegulae white, third vein 
bristly to small crossvein. 

Type.— Cat. No. 10,914, U. S. N. M. 

EUPELETERIA, gen. nov. 

Erected for Bchinomyia fera Linne, magnicornis Zetterstedt, 
prcEceps Meigen, etc. Differs from Peleteria Robineau-Desvoidy in 
lacking the two or three facio-orbital bristles (macrochaetae on para- 
facials next orbit and separated from descending frontal bristles). 
Differs from Bchinomyia Dumeril, as restricted, by having abdom- 
inal macrochaetae not closely set and thorn-like. Body Peleteria- 
like, not Jurinia-like. 

EUFABRICIA, gen. nov. 

Second antennal joint strongly elongate, fully four times as long 
as first, much longer than third; third joint strongly convex on 
front border in profile. Second aristal joint elongate, fully four 
times as long as wide. Palpi widened and flattened on distal one- 
third or so, somewhat spatulate. No ocellar bristles. Parafacials 
wide, front not specially prominent in profile. Cheeks about two- 
thirds eye height in width. Anterior tarsi of female not more 
widened than those of other legs. 


Differs from Fabricia, to which it is most nearly related, in the 
absence of ocellar bristles, and form of palpi and third antennal 

Type, the following species (to be figured in the forthcoming new 
edition of Dr. S. W. Williston's Manual of Diptera, fig. 157). 

Eufabricia flavicans, sp. nov. 

1 hie female, Brazil, H. H. Smith, Coll. Received from Dr. S. W. 

Length. 14 mm. General yellowish or rufous yellowish in ground 
color. Head silvery whitish, frontalia and first two antennal joints 
reddish yellow, third joint and arista light brown. Palpi yellow. 
Parafrontals with a faint tinge of brassy yellow. Thorax and scu- 
tellum brassy yellow pollinose. Abdomen rufous yellow, first seg- 
ment brown on depressed median portion, other segments more 
tinged with rufous on median line, third segment wholly so tinged. 
Narrow anterior margin of second and third and all of anal segment 
yellowish silvery pollinose. A median marginal pair of macrochaetse 
on second segment, a marginal one on each side of first and second 
segments, a median marginal pair and three lateral marginal ones 
on third segment (eight marginal in all), anal segment with a discal 
and marginal row. Legs blackish or brown, the tibiae more or less 
rufous, hind tibiae especially so. Claws reddish or yellowish brown, 
tips darker, pulvilli yellowish. Wing bases broadly yellow, tegulae 

Type.— Cat. No. 11,805, U. S. X. M. 

Genus Dejeania Robineau-Desvoidy and allies 

Dejecmia vexatrix Osten-Sacken and Paradejeania rutilioides 
Jaennicke. — Speaking of Dejeania vexatrix, Osten-Sacken said: "It 
is very remarkable that Dejeania, a South American and Mexican 
genus, should occur so commonly at high altitudes in the Rocky 
Mountains among alpine forms, and it would be worth the while to 
investigate on what insect (probably Lepidopterous) it preys as a 
parasite" (Western Diptera, p. 343). At the close of his paper 
(1. c, p. 354), he again referred to the same matter, and included a 
reference to P. rutilioides, not, however, mentioning it by name. 

These instances of a tropical group of tachinids developing boreal 
forms is paralleled in birds by the parrot genus Rhynchopsitta, pecu- 
liar to the pine region of the Sierra Madre of western Chihuahua. 
The tropical bird group of parrots has here developed a sub-boreal 


genus peculiar to the pine region, and which both passes the winter 
and nests there. Likewise a species of trogon occurs, belonging to 
the monotypic genus Buptilotis, also peculiar to the same region and 
breeding there. 

Parade jeania may be considered as more or less of a boreal off- 
shoot from De jeania, and D. vexatrix Osten-Sacken is a boreal and 
distinct form from the tropical corpulenta Wiedemann. Osten- 
Sacken was mistaken in taking Wiedemann's type to be the same as 

PTEROTOPEZA, nom. gen. nov. 

This name is proposed for Clicctoprocta Brauer and von Bergen- 
stamm (1891), which is preoccupied by Niceville in Lepidoptera 
(1890). Type is Blepharipeza tarsalis Schiner, of South America. 

Genus Gymnochaeta Robineau-Desvoidy 

Gymnochceta alcedo H. Loew. — This species is not typical of the 
genus Gymnochceta. The type of the genus is viridis Fallen, which 
has second antennal joint elongate, second aristal joint elongate, an- 
tennas inserted a little below middle of eyes, and cheeks one-half 
eye height. 

EUJURINIA, gen. nov. 

This genus is proposed for Hystricia pollinosa van der Wulp. An- 
tenna?, frontal bristles, arista, and palpi like Jurinia, but resembling 
Hystricia in having the eyes hairy and the cheeks not so wide. It 
differs from Jurinella in the narrower cheeks and wider parafacials, 
and from Pseudohystricia in the first of these characters and the less 
produced front. 

The cheeks of Jurinia are nearly equal to eye height, and the eyes 
are bare. Hystricia has the third antennal joint truncate at tip, 
second joint not so elongate, but attenuated at origin ; frontal bristles 
weaker, straighter, descending lower, and all directed forward ; no 
macrochaetse on lower border of cheeks, second aristal joint not 
strongly elongate. 

A female specimen in U. S. N. M., collected by the writer July 3, 
at San Rafael, near Jicaltepec, Veracruz, is apparently to be identi- 
fied as Eujurinia pollinosa, although van der Wulp says "arista 
indistinctly jointed," which, it seems, must be an error, and not 
intended by the author. The first two aristal joints in above speci- 
men are elongate and distinct. Also there are some fine strong 
bristles on under side of middle femora. Length, 16 mm. 



This genus is proposed for Saundcrsia testae ea van der Wulp. 
Mr. van der Wulp has remarked on the striking resemblance which 
this species bears to Paradejeania rutilioides. 

Rhachoepalpus olivaceus, sp. nov. 

Two specimens, male and female, collected on the head of Rio 
Piedras Yerdes, about 7,000 feet, Sierra Madre of western Chihua- 
hua (Townsend), one on flowers of Rhus glabra, July 15, the other 
August 16. 

Length of male, 18.5 mm. ; of female, 19 mm. Thorax with an 
olive green tinge. Frontalia with much the same tinge, but darker. 
Second antennal joint with a strong bristle on front border near 
distal end, sometimes a pair of them. Third joint only a little longer 
than second, hardly one and one-half times as long, same size in both 
sexes. Arista thickened on rather more than basal half, distinctly 
jointed, second joint elongate. Scutellum with at least four rows of 
spines. The male shows a median dorsal stripe on abdomen, 
widened in front on second segment, where it is marked by an area 
of spines, narrower on third segment, and narrowest, but still dis- 
tinct, on anal segment. This stripe shows only on anal segment in 
female, but the area of spines is present on second segment. The 
anal segment in both sexes is gently emarginate in middle on hind 
border, presenting a double curve like a pair of buttocks. Wings 
evenly infuscated. Color of scutellum is same in both sexes — quite 
yellowish. Abdomen of male is of a distinctly more reddish shade, 
female abdomen being of nearly same shade as scutellum, if any- 
thing, slightly lighter. Claws of female are yellow, with black tips. 
Front of female is wider, with three proclinate fronto-orbital bristles 
on one side and only two on the other. Front tarsi of female not 

Type.— Cat. No. 10,915, U. S. X. M. 

Rh. olivaceus bears the same striking resemblance to Paradejeania 
that Rh. testaccus does; perhaps even more so, since in the latter 
there seems to be no posterior emargination of anal segment on the 
median line. Mr. van der Wulp's figure shows none, and his text 
mentions none. 

Rhachoepalpus shows broad ventral plates in both sexes, but ven- 
tral membrane in female is not visible. There are five abdominal 
segments, the first very short and barely discernible from the side. 
The female shows ventral plates, bearing thick bunches of spines, 
corresponding to second to fifth dorsal plates, the lateral edges of 


latter overlapping' sides of former, and a sixth ventral plate, or 
sclerite appearing as such, at base of ovipositor. The latter bears 
only hairs. The male with second and third ventral plates bearing 
thick bunches of spines as in female, but fourth with only hairs and 
free, the ventfal membrane showing widely on sides of fourth only ; 
fifth ventral plate much narrower, longer than wide, bare, not free; 
what seems a sixth ventral plate in female represented in male by a 
paired process articulating with the hypopygium. 

EUEPALPUS, gen. nov. 

Differs from E pal pus in having third antennal joint elongate and 
convex in profile on anterior edge, front and epistoma much less 
prominent, face less deeply concave in profile. Eyes absolutely 
bare. Parafacials very wide, black-hairy. Cheeks about equal to 
eye height. Second aristal joint hardly twice as long as wide. 

Differs from Xanthozona (type, melanopyga Wiedemann) in hav- 
ing no discal macrochsetse on abdomen. 

Type, the following species (to be figured in the forthcoming new 
edition of Dr. S. W. Williston's Manual of Diptera, fig. 156). 

Euepalpus flavicauda, sp. nov. 

One female, Brazil, April, H. H. Smith, Coll. Received from Dr. 
S. W. Williston. 

Length, 15 mm. Black; face, cheeks, and beard silvery white. 
Frontalia and parafrontals blackish, quite concolorous. Antenna? 
and arista brown. Thoracic scutum metallic black with greenish 
tinge, thinly silver}' pollinose, more thickly so on anterior edge, 
humeri, and pleura?. Scutellum and abdomen metallic brown with 
a hardly purplish tinge, the anal segment with a conspicuous sub- 
triangular (from above) yellow area defined by the discal row of 
macrochaetse and extending under so as to narrowly surround the 
genital opening. A comb of median marginal thorn-like macro- 
chaeta? on ventral segments, and discal row on ventral side of anal 
segment. A single lateral marginal macrochaeta on first and second 
segments, two median marginal pairs on second, a marginal row of 
ten on third, the discal row on anal ; and only a row of weak bristles 
on posterior edge of anal, appearing like bristly hairs compared with 
the other macrochaetae. Legs brown or blackish; claws and pulvilli 
rufous yellow, tips black. Wings entirely and evenly infuscate. 
tegulae decidedly smoky. 

Type— Cat. No. 11,806, U. S. N. M. 


XANTHOZONA, gen. nov. 

This genus is proposed for Tachina melanopyga Wiedemann. 
Two female specimens in U. S. N. M., Campinas, Brazil (A. Hem- 
pel), and Sao Paulo, Brazil (Ad. Lutz), labeled "parasitic on Brass- 
olis astyra." 

The ventral plates (female) only narrowly showing, overlapped 
by edges of corresponding dorsal plates, exposed portion being 
wider behind and narrowed anteriorly owing to the posteriorly 
rounded-off shape of edges of dorsal plates overlapping them, the 
posterior ones showing more widely than anterior ones, all widen- 
ing successively from anterior to anal segments. 

Family MUSCIM 

Subfamily Calliphorix.E 

Genus Calliphora Robineau-Desvoidy 

Girschner and Hough have paved the way for a clearer under- 
standing of Calliphora and its allies, and the genera as established 
by them are accepted in this paper, with the addition of two new 

Calliphora texensis, sp. nov. 

Two males, three females, Paris. Texas. A. A. Girault, Coll. 

Length, 9 to 11 mm. Differs from C. coloradensis Hough in the 
third posterior intra-alar bristle being absent and without a trace. 
The male front at vertex is about one-fifth of head width, and nar- 
rows very noticeably in front of vertex in an even curve, widening 
at same curve on anterior portion. The male parafrontals and para- 
facials are conspicuously pale brassy. The female parafrontals are 
obscure brownish, the parafacials light russet and unicolorous with 
facialia and facial plate, which are also this color in male. In one 
female the anterior reddish portion of buccal (hairy part of cheeks) 
looks almost black in some lights, but the reddish tinge can be dis- 
tinctly seen, and the specimen should be included with this species. 
The color of abdomen varies from metallic green to purplish blue. 

Type'.— Cat. No. 10,883, U. S. N. M. 

Calliphora rubrifrons, sp. nov. 

Two females, one male, Stickeen River, British Columbia, H. F. 
Wickham, Coll. ; two females, one male, Kaslo, British Columbia, 
H. G. Dyar, Coll. 


Length of female, 9.5 to 12.5 mm. ; of male, 8.5 to 9.5 mm. Buccse 
black, beard black. The two Stickeen River females and one of 
those from Kaslo, being the three largest specimens, show the buccae 
with a good reddish tinge on anterior half, the two males and the 
other Kaslo female not. Third posterior intra-alar bristle absent. 
Frontalia bright orange red on anterior portion, in the Stickeen 
River male more of a yellowish red, in the Kaslo male a brownish 
yellow. Parafacials, facialia, epistoma, palpi, and apex of second 
antennal joint with base of third joint nearly the same color as the 
frontalia anteriorly, but sometimes a lighter shade of same color. 
Female front over one-third of head width, male front about one- 
twentieth of head width. Thorax faintly silvery white dusted, most 
thickly so on front border. Abdomen metallic green to blue, dis- 
tinctly silvery pollinose in certain lights. Wings clear, with more 
or less distinct flecks of black on humeral, small, and basal cross- 
veins, origin of third vein, and apex of auxiliary. Alulae well tinged 
with smoky, appearing quite black if resting against thorax or base 
of wing, tegulae blackish with narrow white margins. 

Type.— Cat. No. 10,884, U. S. N. M. (Stickeen River, British 

Calliphora popoffana, sp. nov. 

One female, Popoff Island, Alaska, July 16, 1899. Harriman Ex- 
pedition. T. Kincaid, Coll. 

Length, 10.5 mm. Buccal black, beard black. Front and face 
black, with a faint silvery white pollen distinctly to be seen in certain 
lights, even on facial plate, and especially on the broad frontalia and 
on the parafacials. Palpi light reddish yellow, facialia and epistoma 
darker reddish yellow, second antennal joint reddish, rest of an- 
tennae black. Front distinctly more than one-third head width. No 
trace of third posterior intra-alar bristle. Wings quite clear, even 
at base, tegulae white. Abdomen metallic green. Legs black. The 
plumosity of the arista is much shorter than in the other species. 

Type.— Cat. No. 10,885, U. S. N. M. 

A male from Bear Lake, British Columbia, 7,000 feet, R. P. 
Currie, Coll., measures 7 mm., and may be this species. The front 
is about one-eighth head width. The parafacials and narrow para- 
frontals are strongly silvery white; also facial plate. The frontalia 
are brownish. Tegulae blackish. Wings with two smoky streaks 
on costal half. Abdomen metallic blue, silvery white dusted. The 
antennae are paler on basal half of third joint. Otherwise it agrees 
with the female just described. The plumosity of the arista is quite 


normal, and this, taken with the blackish tegulse and wing- streaks, 
would point to it as a distinct form. 

Calliphora irazuana, sp. nov. 

One female, Irazu, Costa Rica, Schild and Burgdorf. 

Length, 11.5 mm. Buccae black, beard black. Third posterior 
intra-alar bristle wholly absent. Parafrontals black, with a soft 
brassy brown pollen on front half. Parafacials dark dragon's- 
blood red, facial plate blackish. Palpi reddish yellow, antennae 
blackish, inner basal portions of third joint paler. Front equilateral, 
one-third head width. Thorax and scutellum black, faintly silvery 
on front and lateral edges. Tegulae and wing bases blackish. Ab- 
domen purplish blue. Legs wholly soft black, as are also the pleurae, 
with hardly a trace of silvery. 

Type.— Cat. No. 10,886, U. S. N. M. 

EUCALLIPHORA, gen. nov. 

Proposed for Calliphora latifrons Hough. Differs from Calli- 
phora in possessing two strong pairs of ocellar bristles. This is a 
character of considerable importance in the Muscoidea, especially 
in the higher groups, and may well form a generic distinction here. 

Bucalliphora latifrons Hough. — A large series of this interesting 
species, consisting of some sixty specimens, was brought from 
Kaslo. British Columbia, by Messrs. Dyar, Caudell. and Currie. 
The character of the second pair of ocellar bristles is constant in all. 

There are also two females in the U. S. X. M., collected by H. S. 
Barber, Las Vegas Hot Springs, N. Mex., and Fieldbrook, Gal., 
which both belong to this genus and are apparently this species. 

Genus Lucilia Robineau-Desvoidy 

There are several >pecies of this genus, notably sericata (Meigen) 
Hough and sylvarum (.Meigen) Hough, which have a well-devel- 
oped second pair of ocellar bristles. The latter are remarkably 
strongly developed in these two species, and were it not for the pres- 
ence of certain intermediate forms, like pilatci Hough, and especially 
oculata, n. sp., they would constitute a well-marked new genus sep- 
arable on this character. But in pilatci the second pair in the male 
is hardly to be differentiated in strength from some of the other 
pairs of divergent ocellar hairs, and in oculata the male shows no 
second pair, though the females of both possess the character quite 
distinctly. As genera are mere matters of convenience, and these 
forms do not otherwise differ in points of generic value, the charac- 


ter in question can not be used here for the erection of a separate 
genus. This is only another illustration of the fact that a character 
of value for the separation of certain forms may be valueless for this 
purpose in certain other forms closely allied to the first. In all the 
species there'are several widely divergent pairs of weak ocellar hairs 
behind the first or regular pair of ocellar bristles. In the forms 
which have a second pair of ocellar bristles well developed, this sec- 
ond pair is always inserted just behind the two posterior ocelli, and 
not inside the ocellar triangle. In other words, it is only the pair of 
hairs inserted just behind the two posterior ocelli that ever develop 
into a second strong pair of bristles. L. ccesar is typical of the forms 
in which this pair of hairs is not developed in either sex, but it is to 
be noted that some of the bristly hairs within the ocellar triangle in 
this species often seem strong enough to be considered additional 
pairs of ocellar bristles. 

Fourteen species of Lucilia are here recognized, occurring in ma- 
terial in U. S. N.. M. They may be separated as follows: 

Table of Lucilia spp. 

1. Only one postacrostichal bristle present morrilli, sp. nov. 

Two postacrostichal bristles present 7 

Three postacrostichal bristles present 2 

2. Palpi more or less yellowish 4 

Palpi wholly black or blackish 3 

3. Second pair of ocellar bristles developed sylva/rum 

Second pair not developed nigripalpis, sp. nov. 

4. Palpi wholly yellow 5 

Palpi infuscate at tip 6 

5. Second pair ocellar bristles developed sericata 

Second pair not developed angustifrons, sp. new 

6. Second pair developed giranilti, sp. n v. 

Second pair not developed barberi, sp. nov. 

7. Abdomen nnicolorous 8 

Ba^al segment of abdomen black or blackish 9 

8. Second pair developed unicolor, sp. nov. 

Second pair not developed ccesar 

9. Whole body purplish, except basal abdominal segment, second abdom- 

inal segment with a marginal row cf bristles purpurea, sp. nov. 

Whole body not so 10 

10. Second and third abdominal segments with a purplish or blackish 

margin 1 1 

Second and third segments unicolorous 12 

11. Buccse yellow wholly or partly pilatei 

Bucoe black, not at all yellow austral is, sp. nov. 

12. Eyes normal, face black infuscata, sp. nov. 

Eyes flattened anteriorly with large front aspect, face brownish yellow. 

oculata, sp. nov. 


Lucilia morrilli, sp. nov. 

Six males, nine females, Texas, New Mexico, Arizona, California, 
British Columbia, and Missouri. 

Only one postacrostichal bristle. Male front one-seventh of head 
width, female front fully two-fifths of head width. Whole of abdo- 
men, thorax, parafrontals, and cheeks, including occiput, strongly 
metallic green. Face and frontalia black, silvery. Palpi black. 
Tegulae white. No macrochsetse on abdomen. No second pair of 
ocellar bristles. 

Type.— Cat. No. 10,887, U. S. N. M. (Victoria, Texas— Morrill '). 

Lucilia sylvarum (Meigen) Hough. — One female, Prussia. 

Three postacrostichal bristles. Male front very narrow, female 
front one-third head width. Palpi black. Two stout marginal 
macrochaetse on second abdominal segment. Second pair of ocellar 
bristles well developed. 

Lucilia nigripalpis, sp. nov. 

Two females, Cuyahoga County, Ohio. \Y. A'. Warner. 

Differs from infuscata only by having three postacrostichal bris- 
tle-;; palpi quite blackish, faintly paler basally ; antenna?, face, buccal, 
and front all more deeply black ; tegula; white. A trace of purplish 
on hind margins of second and third abdominal segments, especially 
on second. Second segment with a marginal row of weak macro- 
chsetae. No second pair of ocellar bristles. 

Type— Cat. No. 10,888, U. S. N. M. 

Lucilia sericata (Meigen) Hough. — Two males, six females, east- 
ern United States, Alabama, Hidalgo (Mexico), Kadiak Island 

Three postacrostichal bristles. Male front one-eighth to one- 
sixth head width, female front two-fifths head width. Palpi yellow. 
Abdomen unicolorous, tegulse white. A strong second pair of 
ocellar bristles in both sexes. 

Lucilia angustifrons, sp. nov. 

One male, England (Brunetti). 

Same as ccesar, but having three postacrostichal bristles. Front 
linear, eyes almost contiguous. Palpi yellow. A female, having 
front one-third head width, from Kaslo, British Columbia (Cau- 
dell), seems to be this form. No second pair of ocellar bristles. 

Type.— Cat. No. io,88q, U. S. N. M. 


Lucilia giraulti, sp. nov. 

One male, Paris, Texas. A. A. Girault, Coll. 

Three postacrostichal bristles. Male front one-eighth head width. 
Abdomen like pilatci, but no dark hind margins to second and third 
segments. Buccae and whole face and front black, palpi yellowish 
but infuscate apically. Tegulae nearly white. No strong macro- 
chretse except marginal row on third segment. Of the three post- 
acrostichal bristles, the front one is well behind the front postsutural 
bristle, and the middle one is a little behind the middle postsutural. 
A second pair of ocellar bristles present. 

Type.— Cat. No. 10,890, U. S. N. M. 

Lucilia barberi, sp. nov. 

Six males, Arizona (H. S. Barber), California (Coquillett), 
Guanajuato (Mexico), Alabama, West Virginia, and District of 

Three postacrostichal bristles. Differs from giraulti practically 
only in the second pair of ocellar bristles not being developed appre- 
ciably longer than the ocellar hairs, and the three postacrostichal 
bristles being even with the three postsutural bristles. Palpi yellow- 
ish, infuscate at tip. Buccae, face, and front blackish, facialia red- 
dish, epistoma yellowish. Tegtilae white. Basal abdominal seg- 
ment black. An even row of ten marginal macrochaetae on third 
segment above, and three on each side below. No dark margins to 
second and third segments. Male front one-eighth head width. 

Type.— Cat. No. 10,891, U. S. N. M. (Williams, Arizona). 

Lucilia unicolor, sp. nov. 

Five females, New Mexico, Mexico, and British Columbia. 

This form corresponds to cccsar, differing therefrom in having the 
second pair of ocellar bristles distinctly developed. Two postacros- 
tichal bristles. Female front a little less than one-third head width. 
Palpi yellow. Abdomen unicolorous. Tegulae white. 

Type.— Cat. No. 10,892, U. S. N. M. (Mesilla, N. Mex.— Cock- 

Lucilia cccsar Ljnne. — Numerous specimens of both sexes, Eng- 
land, eastern United States, and British Columbia. 

Two postacrostichal bristles. Male front linear, female front one- 
third head width. Abdomen unicolorous. Tegulae white. Palpi 
yellow. Second pair of ocellar bristles not developed, or only very 
weakly so. 


Lucilia purpurea, sp. nov. 

One female, Fort Wrangel, Alaska, Wickham; one male, Kadiak, 
Alaska, Kincaid (Harriman Expedition). 

Two postacrostichal bristles. Male front one-twelfth of head 
width, female front one-third head width. Palpi yellow. Basal 
abdominal segment blaekish. Whole body purplish, strongly violet 
tinged, especially in the female. Tegulse of female white, of male 
smoky. Buccae, face, and front blackish, epistoma paler. Second 
abdominal segment with a marginal row of bristles or macrochsetre, 
but not as strong as those of marginal row of third segment. No 
second pair of ocellar bristles. 

Type.— Cat. No. 10,893, U. S. N. M. (Fort Wrangel, Alaska). 

LuciHa pilatei Hough. — Two males, two females, Florida, Porto 
Rico, Guatemala, and Peru. A neotropical species. 

Two postacrostichal bristles. Male front one-eighth head width, 
female front one-fourth head width. Palpi yellow. Abdomen as 
in australis, only the purplish or black margins of segments often 
more marked. Buccae of female yellow, of male gray with yellow 
anteriorly. A second pair of ocellar bristles in female more or less 
hair-like, but distinctly larger and thicker than the other hairs of 
ocellar area ; in male very weak, not appreciably stronger than the 
other ocellar hairs. 

The purplish black hind margins of second and third abdominal 
segments are characteristic of this and one or two other species, 
added to which is the blackish basal segment. The latter in some 
females shows a little metallic green on sides, but the general opaque 
black of its dorsum is the distinguishing character. Also the hind 
margins of second and third segments are only faintly purplish in 
some specimens, but distinct traces are present in all. The white 
tegulae are characteristic of this species, and serve to separate the 
males of pilatei from the males of similar species having a black 
basal abdominal segment. 

Lucilia australis, sp. nov. 

Two females, Tennessee, Texas (Girault) ; one male, Popoff 
Island, Alaska (Kincaid, Harriman Expedition). The male is pro- 
visionally referred here. 

Two postacrostichal bristles. Male front one-twelfth head width, 
female front one-fourth head width. Palpi infuscate yellow. Basal 
abdominal segment black above, conspicuously so. the purplish or 
darker hind margins of second and third segments also showing. 


Distinguished from pilatei by the buccal being black, silvery gray 
pollinose, not at all yellow. Second pair of ocellar bristles present 
in female, not developed in male. 

Type.— Cat. No. 10,894, U. S. N. M. (Tennessee, Coll. Riley). 

Lucilia infuscata, sp. nov. 

Nine males, six females, Massachusetts, New Hampshire, Ohio, 
Missouri. New Mexico, Arizona, and British Columbia. 

Two postacrostichal bristles. Male front very narrow, female 
front two-sevenths of head width. Basal abdominal segment black 
or purplish black in female, but no dark margins to second and third 
segments. Male tegulse infuscate, female tegular more nearly white. 
Palpi yellowish. Buccae, face, and front black. No second pair of 
ocellar bristles. The female told from female ccesar only by 
narrower front and darker basal segment. 

Type.— Cat. No. 10,895, U. S. N. M. (Organ Mountains, New 
Mex., on flowers of Lippia wrightii — Townsend). 

Lucilia oculata, sp. nov. 

Six males, two females, District of Columbia, Kentucky, North 
Carolina, Mississippi, Kansas, and Cuba. 

Two postacrostichal bristles. Male front linear, eyes nearly con- 
tiguous and approximated more anteriorly than in infuscata, with 
larger front aspect than in that species. Female front one-fourth 
of head width. Tegulse nearly white, only very faintly tinged with 
yellowish. Antennae and face brownish yellow instead of black. 
Basal abdominal segment quite black. Male shows no second pair 
of ocellar bristles, but female has them developed. Otherwise like 

Type— -Cat. No. 10,896, U. S. N. M. (Cumberland Gap, Ky.— G. 

PROTOPHORMIA, gen. nov. 

Hough characterizes Phormia as having the mesonotum "some- 
what flattened caudad the transverse suture," as in ProtocaUiplwra. 
This is a mistake. P. regina, which is the type of Phormia, does 
not show this flattening at all. The species tcrrccnovce is not a 
Phormia, but differs in possessing the same conspicuous flattening 
seen in ProtocaUiplwra. The new genus Protophormia is herewith 
proposed for its reception. The characters given by Hough for 
Phormia (Ent. News, x, p. 66) all apply to P. regina except the 
character of the flattened thorax. This flattening carries with it a 
more or less complete abortion of the postacrostichal bristles except 
the hindmost one of each row. 


Subfamily MUSCIN43 

Tribe Mese m brin in i 

Two new genera are here proposed in this tribe, and the genus 
Mesembrina is restricted as follows: 

Genus Mesembrina Meigen 

Type of the genus, M. mystacea Linne. Densely pilose flies. 
Subalar pile present, representing the pteropleural bristles. Sterno- 
pleural bristles I. o. 1. Fourth longitudinal vein very deeply and 
roundly bent far before reaching margin of wing, which latter point 
is same distance behind that termination of third vein is before ex- 
treme wing-tip, the portion between bend and margin being fully 
three times that in margin. Apical cell much narrowed, its mouth 
width not over one-third its greatest width. Small crossvein dis- 
tinctly before middle of discal cell. 


Proposed for Mes. meridiana Linne. Hairy, not pilose, flies. 
Subalar bristly hairs present, representing pteropleural bristles. 
Sternopleural bristles o. o. I. Fourth longitudinal vein reaching 
front margin of wing before tip, arcuate at bend. 

EUMESEMBRINA, gen. nov. 

Proposed for Mes. latreillei Robineau-Desvoidy. Hairy, not pi- 
lose, flies. Pteropleural hairs present. Sternopleural bristles 
1. o. 2. Fourth vein very slightly and roundly bent a little before 
reaching hind margin of wing, the portion between bend and margin 
about equal to the portion in margin. Apical cell very widely open, 
its mouth width equal to about three-fourths its greatest width. 
Small crossvein distinctly beyond middle of discal cell. 

Bumesembrina latreillei Robineau-Desvoidy. — Two specimens, 
White Mountains, New Hampshire, Morrison; one, Colorado; two, 
Washington State ; two, Kaslo Creek, British Columbia, June 18, 
R. P. Currie and A. N. Caudell. All show face and parafacials sil- 
very white from above. Antennae reddish yellow to brownish. 
Palpi reddish or brownish red. 

Eumesembrina alascensis, sp. nov. 

Four specimens. — Kukak Bay, July 4; Kadiak, July 20; Saldovia, 
July 21 : Juneau, July 25. All Alaska. Collected by T. Kincaid 
(Harriman Expedition). 


These specimens are more hairy, more bristly on thorax and scu- 
tellum, and on peristomalia. They also usually show less silvery 
on face and parafacials, and the antennas are quite black. Palpi 
black. The Kukak Bay and Kadiak specimens show no silvery on 
the soft blackish facial plate, and the parafacials are tan-colored 
without a sign of silvery. The other two specimens show some 
silvery, not only on facial plate, but also on the more or less tan- 
colored parafacials. 

Type.— Cat. No. 10,897, U. S. N. M. (Kukak Bay, Alaska). 

The two Washington State and two British Columbia specimens 
mentioned under latreillei are certainly distinctly to be referred to 
that species, which is the eastern form, and which is thus seen to 
range from the Atlantic to the Pacific. Humes, alascensis doubtless 
represents rather a boreal form. 

Family PHASHM) 

Tribe Anurogynini 

Genus Hyalomyodes Townsend 

Hyalomyodes weedii Townsend. — This species seems distinct 
from Hyalomyia triangulifera H. Loew, but needs further study. 
The writer has examined the type of the latter in Cambridge. 

Hyalomyodes triangulifera H. Loew. — Ten specimens from the 
White Mountains of New Hampshire, one from Massachusetts, and 
one from Maryland agree perfectly with the description of H. weedii 
Townsend. They also agree with Loew's description, but an exam- 
ination of the type in Cambridge seemed to indicate differences. 
The front, frontalia, and parafacials are wider in the male, and the 
claws are elongate. Humeri grayish. 

Hyalomyodes robusta, sp. nov. 

Two males, North Fork of Rio Ruidoso, White Mountains, New 
Mexico, about 8,200 feet, on flowers of Solidago trincrvata, August 
17, Townsend. 

Differs from triangulifera in being more robust, and first abdom- 
inal segment with pollinose fascia same as second and third. The 
thorax is also more conspicuously pollinose. Hind crossvein quite 
straight, in one specimen much nearer to small crossvein than to 
bend of fourth vein, in both distinctly nearer. The pollen of median 
portion of thorax and abdomen has a brassy tinge, that on sides 
being silvery-whitish. Macrochsetae not so well developed, consid- 
erably weaker. Parafacials wide in both specimens. Length, 5 mm. 

Type.— Cat. No. 11,651, U. S. N. M. 


Hyalomyodes californica, sp. nov. 

Two specimens, male and female, Santa Clara county, California 
(C. F. Baker). 

Almost like triangulifcra, but distinguished by humeri being' more 
golden, extending back in a lateral stripe. 

Type.— Cat. No. 11,652, U. S. N. M. (female). 

Tribe Clistomorphini 
Genus Clistomorpha Townsend 

A synonym of Clistomorpha is Clytiomyia Coquillett (non Ron- 
dani). This genus is very distinct from Clytiomyia Rondani 
(Clytia Robineau-Desvoidy). C. hyalomoides Townsend is distinct 
from C. did y ma H. Loew (described as Xysta). The writer recog- 
nized the fact of the two being congeneric nearly fifteen years ago, 
from drawings of the type furnished by Mr. Samuel Henshaw, and 
has since examined the type of didyma in the Cambridge Museum. 

Clistomorpha didyma H. Loew. — The apical cell is very short- 
petiolate, and the hind crossvein is curved and in middle between the 
small crossvein and bend of fourth vein. 


Clistomorpha hyalomoides Townsend. — The apical cell is prac- 
tically closed in the margin. The hind crossvein is in middle and 

Xew York. 

Clistomorpha atrata Coquillett. — The apical cell is closed in mar- 
gin, or almost narrowly open. The hind crossvein is sinuate and 
nearer to bend of fourth vein than to small crossvein. 

Idaho, Washington State. 

Genus Himantostoma H. Loew 

Himantostoma sugens H. Loew. — This genus belongs in this 
tribe, as shown by an examination of the type in Cambridge. 

Subfamily Phasiin.t: 

Tribe Alophorini 

The following table will serve to separate the genera of this tribe : 

1. Front above antennae thickly beset on both sides with small bristles 2 

Front above antennae naked, only one row of frontal bristles on each 
side 3 


2. First longitudinal vein elongate, small crossvein placed opposite end of 

auxiliary vein, fourth vein very obtusely bowed, apical cell sharp- 
angled at extremity and short petiolate Alophora 

First longitudinal vein not elongate, small crossvein placed opposite end 
of same, fourth vein roundly bowed, apical cell usually long petiolate. 

Hyalo myia 

3. Second longitudinal vein ending opposite the junction of third and 

fourth veins, wings of male usually much widened Phorantha 

Second longitudinal vein elongated beyond junction of third and fourth 
veins, wings of male not widened, apical cell very long petiolate, 
fourth vein roundly bowed Paralophora 

Genus Alophora Robineau-Desvoidy 

Alophora sp. — A large species from Texas. The female shows 
ventral plates overlapped by dorsal plates. The male shows ventral 
plates free, at least those of second, third, and fourth segments, with 
membrane widely exposed on each side. 

Genus Phorantha Rondani 

The genus Alophora has the front prominent in profile above in- 
sertion of antennae. Phorantha has front flattened, and with greater 
slope so as to present in profile an almost perfectly straight line 
from insertion of antennae to vertex. 

Probably all, or nearly all, of the various forms of the Alophorini 
that have been described are distinct and entitled to recognition. 
We know practically nothing of the early stages or the mating of the 
adults, and it is premature to attempt to outline the synonymy in the 
absence of such knowledge. 

Tribe Cistogasterini 

Genus Gymnoclytia Brauer and von Bergenstamm 

The genus Gymnoclytia is distinct from Cistogaster. The pe- 
duncle of apical cell is continuous with fourth vein in Gymnoclxtia, 
but with third vein in Cistogaster. 

Gymnoclytia has ventral membrane (female) very widely visible 
and ventral plates free, much as in Gymnosoma. 

Gymnoclytia occidua Walker.- — Male. — Thorax brassy or golden 
pollinose, with two straight narrow median vittae extending from 
front margin to behind suture, and two irregularly widened vittae 
obsolete before and interrupted at suture. Abdomen more or less 
ferruginous, sometimes entirely so, but usually with a longitudinal 
fuscous stripe in connection with a median pollinose vitta, and more 
or less brown on third and fourth segments with grayish pollen. 



Female. — Thorax silvery-whitish pollinose, with two heavy shin- 
ing black vittse, sides of front silvery-white pollinose becoming black- 
ish posteriorly, abdomen black with silvery pollen in median vitta 
and two or three fascia?. 

New Hampshire, District of Columbia, North Carolina, Georgia, 
and Texas. 

Gymnoclytia occidentale, sp. nov. 

Male. — Thorax deep brassy to old-gold pollinose, with same vittse 
as in occidua. Abdomen like occidua except that pollen is golden, 
the ground color bright ferruginous and markings varying from 
none to the usual ones strongly marked. 

Female. — Colored almost like the male of occidua. Thorax brassy 
pollinose, with two broad heavy brown vittse extending from an- 
terior margin almost to scutellum, and two very narrow linear vittae 
between them. Abdomen the same as in the male, pollen being 
golden, but no specimens occur with abdomen entirely ferruginous, 
the usual markings being pronounced in all. 

Colorado and New Mexico to California. 

Type— Cat. No. 11,653, U. S. N. M. (female, Beulah, New Mex- 
ico, Cockerell, July, 1902). 

Gymnoclytia immaculata Macquart. — Male. — Fuscous stripe of 
abdomen wanting, median pollinose vitta more or less distinct. Ab- 
domen yellowish, the third and fourth segments with lateral polli- 
nose reflections. 

Female. — Thorax shining black, without pollinose markings ex- 
cept the humeri, sides of front shining black, abdomen without 
distinct pollinose vitta or crossbands, apical cell quite long petiolate 
(as in the males of the preceding species). Abdomen distinctly red 
on the sides, especially anteriorly. 

This form and Gym. occidua Walker are distinct. See Robert- 
son's and the writer's notes in T. A. E. S., xxn (1895), pp. 66-67, 
and Ann. and Mag. N. H., xx, pp. 283-284. 

Gymnoclytia femiginosa van der Wulp. — Male. — Thorax deep 
golden or old-gold pollinose, with the same stripes as occidua more 
or less apparent. Abdomen ferruginous, fuscous stripe hardly ap- 
parent, but pollinose stripe present, and third and fourth segments 
more or less pollinose, pollen being golden. 

Female. — Sides of front faintly golden-silvery, thorax shining 
black, with three faintly golden pollinose vittse. Abdomen shining 
black, with median pollinose vitta and third and fourth segments 


more or less pollinose, pollen being grayish with a hardly brassy 

Veracruz and Nicaragua. 

Tribe Xanthomelanodini 
Genus Xanthomelanodes Townsend 
Syn. Xanthomelana van der Wulp preocc. 

The name used by van der Wulp was applied by Bonaparte to a 
genus of birds in 1850. 

Xanthomelanodes arcuata Say. — Only a single vibrissa on each 

Male. — Front and face deeply golden, especially parafrontals. 
Abdomen usually with a well-defined black median vitta, last seg- 
ment and last half of penultimate segment black. 

Female. — Front and face silvery-white. Abdomen all black ex- 
cept yellow on sides of second and third segments, only covering 
anterior half of third segment, but some specimens show less black. 

New Hampshire, Kansas, Veracruz. 

Xanthomelanodes atripennis Say. — One vibrissa on each side. 

Male. — Front golden. Abdomen golden, with only some brown- 
ish shading for the median vitta. Wings quite smoky on inner 

Dixie Landing, Virginia (Townsend). 

Xanthomelanodes californica, sp. nov. 

Two vibrissas on each side. 

Male. — Front and face almost silvery, with only a faint suggestion 
of golden, in some specimens quite silvery-white. Abdomen ferru- 
ginous, more or less dusky, the brown markings not well defined as 
a rule, consisting of a broken median stripe and the usual dark 
markings of last two segments. 

Female. — Face and parafrontals silvery-white. Abdomen nearly 
same as in arcuata. 

Colorado, Nevada, California. 

Type. — Cat. No. 11,654, U. S. N. M. (male, Los Angeles county, 
California, Coquillett). 

Tribe Trichopodini 

The following is a table of the genera of this tribe : 

I. Hind tibiae without flattened cilia, with only a row of short appressed 

bristles and one or two stronger bristles among them 2 

Hind tibiae ciliate with the usual flattened cilia 3 


2. Wings infuscate on less than costal half, gray or hyaline on other 

portion Acaulona 

Wings almost wholly infuscate, but much more faintly so on inner half, 
the infuscation rather graduated into almost hyaline on inner border. 


3. Wings wholly infuscate 4 

Wings more or less widely hyaline on inner border, the hyaline 

abruptly defined 6 

4. Hind femora ciliate with closely appressed flattened bristles on one or 

opposite edges Galactomyia (males) 

Hind femora not ciliate at all 5 

5. Apical cell open Homogenia 

Apical cell closed Euomogenia 

6. Wings hyaline on more than inner half, abdomen subcylindrical in both 

sexes and largely translucent in both Pennapoda 

Wings with inner hyaline border almost as wide as the infuscate 

costal half, abdomen subcylindric in both sexes and wholly opaque 

in both Polistomyia 

Wings with hyaline border about one-third width of wing. . .Eutrichopoda 
Wings with hyaline border very narrow, not over one-fifth of wing 

width 7 

7. Wholly black form Galactomyia lanipes (female) 

Partly reddish forms 8 

8. Hind femora ciliate distally on inner edge with closely placed bristles ; 

abdomen cylindrical, reddish or orange with apical half or at least 

anal segment black Galactomyia radiata (female) 

Hind femora not at all ciliate, smaller forms with abdomen more or 
less flattened and almost wholly light reddish or yellowish in both 
sexes Trichopoda 

Genus Acaulona van der Wulp 

Acaulona costata van der Wulp. — One female, Tehuantepec, Su- 
mlchrast ; one male, Frontera, Tabasco, February 9, Townsend. The 
Tehuantepec specimen is of much lighter coloration than Veracruz 
and Tabasco specimens. Tegulae are yellowish in this genus. 

Acaulona tehuantepeca, sp. nov. 

One female, Tehuantepec, Sumichrast. Labeled "17. Homogenia 
sp." Length, 7 mm. 

Differs from A. costata in having the apical cell subfuscous, the 
abdomen with a median blackish vitta and more or less wholly 
blackish on apical half, and the hind tibiae weakly subciliate in a row 
of short, closely approximated bristles. The form is intermediate 
between Adaulona, Euacaulona, and Homogenia, but nearest to 

Type— Cat No. 10,878, U. S. N. M. 


EUACAULONA, gen. nov. 

Differs from Acaulona in having somewhat more than costal half 
of wings pronounced fuscous, the rest of wing not being clearly 
hyaline, but more or less so, the fuscous rather gradually fading out 
on inner border. There are also two distinct grayish or milky vittae 
on wings (male), one between the first and second veins, one be- 
tween the third and fourth veins, besides a short one in front of the 
auxiliary vein. Tegulse brownish or fuscous, paler in middle. 

The front at vertex is nearly as wide as either eye, and gradually 
widens anteriorly to almost width of both eyes as viewed from in 
front, the face in same view being fully three-fifths width of head. 
The frontalia are very wide, of equal width, as wide as front at 

Apical cell closed in margin. Hind tibiae not ciliate, bearing only. 
a row of short appressed bristles with one or two stronger bristles 
among them. Claws of male elongate. Abdomen of male flattened. 
Type, the following species : 

Euacaulona sumichrasti, sp. nov. 

One male, Tehuantepec, Sumichrast. Length, 9.5 mm. 

Blackish, the venter and basal half of femora, also base of hind 
tibiae, yellow ; the usual golden yellow markings on prothorax along 
and in front of suture, also extending posteriorly on the sides and 
along the scutellar suture. Frontalia black, the narrow parafrontals 
and all of face and cheeks golden yellow. Thorax, scutellum, and 
abdomen above brown or blackish. 

Type.— Cat. No. 10,879, U. S. N. M. 

Genus Homogenia van der Wulp 

Syn. Trichopododes Townsend. 

This genus has the wings wholly infuscate, those of male with 
considerable luteous and more or less of a milky bloom (latipennis 
and nigroscntellata) ; female not known. Hind femora not ciliate 
at all, hind tibiae only weakly ciliate. Apical cell open. Tegulae 
yellowish. Type, H. latipennis van der Wulp. The species rufipes 
van der Wulp evidently does not belong with the other two described 
under the genus, and will have to be separated generically. 

H. latipennis van der Wulp. — One male, Tehuantepec, Sumichrast. 
Labeled "Trichopoda luteipennis Wd." This specimen agrees with 
van der Wulp's description except that there is no trace of a black 


median abdominal vitta, and abdomen is only a little dusky on anal 
segment as seen through the golden silvery bloom. 

H. nigroscutellata van der Wulp. — One male, Cacao, Trece 
Aguas, Alta Vera Paz, Guatemala, April 18. Barber and Schwarz, 
collectors. This specimen agrees well with van der Wulp's de- 
scription except that abdomen is widely blackish on median portion, 
with only narrow lateral borders yellow. The scutellum has golden 
pollen on dorsum. 

EUOMOGENIA, gen. nov. 

Differs from Buacaulona in the wings (male) being wholly in- 
fuscatc, uniformly so, the same milky vittae being present ; and in the 
hind tibiae being ciliate with moderately developed cilia. Front like 
Homogenia, very broad. Apical cell closed in border. Tegulae 
blackish. Type, the following species : 

Euomogenia lacteata, sp. nov. 

One male, Frontera, Tabasco, March 3, Townsend. 

Length, 9.5 ram. Blackish, the usual silvery golden markings on 
mesoscutum, including the sides back to scutellar suture and along 
latter. Scutellum somewhat silvery golden on dorsum. Abdomen 
wholly fuscous, with a reddish tinge showing through the fuscous. 
The broad frontalia velvety blackish, narrow parafrontals and whole 
of face and cheeks golden. Antennae brownish. Palpi yellow, dark 
on tips. Basal half of femora yellow, least extensive on front pair, 
most extensive on hind pair, base of hind tibiae yellow, rest of legs 
black, claws and pulvilli yellow, tips of claws black. Wings black- 
ish, with the milky or golden grayish vittae described for Buacaulona. 

Type— Cat. No. 10,880, U. S. N. M. 

Genus Pennapoda Townsend 

This was described as a subgenus, in Ann. and Mag. N. H., xx, p. 
282. It is here raised to generic rank. Type, Trick, phasiana 
Townsend, loc. cit., male and female. The species Phania simillima 
Wiedemann and Trich. subalipcs Townsend may belong to this 
genus. There are no specimens in U. S. N. M. for examination. 

POLISTOMYIA, gen. nov. 

This genus is proposed for the Trich. trifasciata H. Loew group. 
The abdomen is subcylindrical in both sexes, slightly more widened 
on apical portion in male. Apical cell closed and quite long petiolate. 
Wings with but little more than costal half colored, the inner por- 


tion clear. Abdomen in both sexes wholly opaque, brown or black 
in ground color, more or less golden pollinose, never with translu- 
cent portions. Scutellum always yellow. Both sexes have yellow 
on the wings. Hind femora not at all ciliate. Tegulse white or 
yellowish. Parasitic in Acridiidae (Dissosteira), so far as known. 

The male has frontalia suddenly narrowed, presenting a curved 
outline on each side, closely followed by the frontal row of bristles, 
the width on posterior half being only one-half the width at base of 
antennre. Claws strongly elongate in male, hypopygium exserted 
and tucked up under the end of abdomen. 

The female has the frontalia but little narrowed behind, being 
evenly narrowed from anterior to posterior end, the sides and frontal 
row of bristles being quite straight. Claws somewhat elongate in 
female, even slightly longer than last tarsal joint, but very markedly 
less elongate than in male. Anal end of abdomen truncate, the ovi- 
positor more or less withdrawn within anal segments, its apex 
usually showing. 

Type, T. trifasciata H. Loew. 

The other species belonging here are histrio Walker, indivisa 
Townsend, probably umbra Walker and plumipes J. C. Fabricius ; 
also the following new species. The writer formerly suggested 
these (except plumipes) as varieties of one species, but now con- 
siders them valid forms differing in marked characters. They form 
a group apart by themselves, distinctly contrasted with the other 
members of the Trichopodini. 

Polistomyia subdivisa, sp. nov. 

One female, St. Helena, Napa County, Cal., bred by A. Koebele 
from a locust (Dissosteira venusta Stal) ; issued August 30, 1887. 

Length, 6.33 mm. Segments three and four of abdomen golden 
pollinose, segment three with a median vitta and median hind mar- 
gin brown, segment four wholly pollinose with a trace of vitta, seg- 
ment two with a large yellow spot on each side, and segment one 
with a similar smaller spot on each side. 

Type.— Cat. No. 10,881, U. S. N. M. 

Two female specimens from Las Cruces. New Mexico, collected by 
the writer, August 25 and September 2, on flowers of Solidago ari- 
zonica, are larger, measuring 7 to 8 mm., show no median vitta on 
third and fourth segments, and only a faint vitta on second segment, 
which bears a fascia rather than separated spots. They very likely 
represent another form, but more material is needed from California 
and New Mexico before separating them as distinct. They occupy 
an intermediate position between trifasciata and subdivisa. 



T. plumipes (J. C. Fabricius) Wiedemann is probably a Polis- 
tomyia, as indicated by the yellow scutellum, broad, clear inner mar- 
gin of wings, and the cylindrical abdomen. The latter is described 
as black, which would indicate a form without fasciae or pollinose 
markings, since it is hardly possible that such could have been so far 
lost as to leave no trace. We thus have the following forms of this 
genus, to be separated as below : 

Polistomyia plumipes — No pollinose fasciae on abdomen. Continuous 

black surface. 
histrio — Two pollinose fasciae, interrupted. 
trifasciata — Tbree fasciae, broadly interrupted. 
subdivisa — Four fasciae, two broadly interrupted and two 

faintly so. 
umbra — Continuous pollinose surface, interrupted by median 

indivisa — Continuous pollinose surface. 

EUTRICHOPODA, gen. nov. 

Differs from Trichopoda in the apical cell being moderately long 
petiolate, and the wings with inner border broadly hyaline, the latter 
being nearly or about one-third of wing breadth. Hind tibiae cil- 
iate, hind femora without cilia. Abdomen cylindrical in female, 
probably flattened in male. Tegulse pale or whitish yellow. Type, 
the following species : 

Eutrichopoda nigra, sp. nov. 

Syn. Trich. lanipcs van dEr Wulp (non J. C. Fabricius, Wiedemann), 
Biol. C. A. Dipt, ii, pp. 434-5. 

One female, Tehuantepec, Sumichrast. 

Length, 9 mm. ; of wing, 8 mm. Black. Parafrontals silvery 
white, with only a faint tinge of golden, which tinge is lost in view 
from above and behind. Face wholly silvery white, including para- 
facials. Transverse suture of mesoscutum marked by a golden yel- 
low linear fascia, with two golden lines running to front border of 
thorax, humeri broadly golden. Scutellum is of the same dull black 
as the abdomen, with hardly a brownish tinge. Tegulae saturated 
with a faint yellow tinge. Femora almost as black as rest of legs, 
with a faint brownish tinge. The mesoscutum behind suture is 
faintly purplish or bluish shining. The wings have no yellowish 
tinge in the black, and the inner hyaline border is hardly one-half as 
wide as the black portion. 

Type.— Cat. No. 10,882, U. S. N. M. 


This form comes nearer agreeing with Wiedemann's description 
of plumipes than anything that has turned up since Bosc's time. It 
differs therefrom only as above described, and principally in the 
black scutellum. 

Mr. van der Wulp (1. c.) has described this species from what he 
records as one male and four females, but says nothing as to whether 
the apical cell is petiolate or closed in the margin, nor does he men- 
tion the shape of the abdomen in the sexes. It seems quite certain 
that his specimens are this species, and it is likely that all five of 
them have the apical cell moderately long petiolate. 

Genus Trichopoda Latreille 

This genus, as here restricted, has the wings with inner margin 
narrowly hyaline, hind femora not ciliate at all ; only male with yel- 
low in wings, no milky radiations, apical cell very short petiolate, 
and tegula? yellowish. Type, T. pennipes J. C. Fabricius. Parasitic 
in Heteroptera (Anasa, Leptoglossus), so far as known. 

GALACTOMYIA, gen. nov. 

This genus is proposed for Trick, radiata H. Loew. Trich. 
lanipes J. C. Fabricius (description is of female; T. formosa Wiede- 
mann is the male) also belongs in this genus. 

The males have the abdomen flattened ; the wings infuscate to 
inner margin, milky radiate on a yellow or fuscous background, the 
milky radiations conspicuous and the yellow less pronounced. Hind 
femora strongly ciliate on posterior half, with flattened bristles. 

The females have the abdomen cylindrical ; the wings wholly black 
except narrow inner border, without yellow coloring, the internal 
border abruptly limpid. Hind femora at least short-ciliate distallv, 
though bristles may not be flattened. G. radiata female has the 
abdomen reddish, with at most the apical half black. G. lanipes 
female is to be distinguished by its wholly black coloration, aside 
from the usual yellow of head, thorax, claws, and pulvilli. 

It is yet uncertain what species can be referred to this genus be- 
sides radiata and lanipes (syn. formosa Wiedemann). As to the dis- 
tinctness of these two species, Loew pointed out in his description of 
radiata (male) that it has the palpi reddish yellow, abdomen purple 
black, and bases of femora yellow. G. lanipes (male) has palpi 
black, abdomen obscure rufous, and femora wholly black. 

The males of Galactomyia have ventral membrane widely visible, 
and all the ventral plates free. There are six abdominal segments, 
the first extremely short and not visible above unless abdomen is 


detached, but visible on sides below ; the second to sixth segments 
visible above, and a seventh wedged between the sides of ventral 
aspect of sixth, rounded in outline and forming the base of the 
hypopygium. This seventh segment occupies the position of a ven- 
tral plate to sixth segment and belongs to dorsum, being a dorsal 
plate. Five ventral plates, corresponding to first to fifth segments ; 
first plate rather crescent shaped, much shorter (antero-posteriorly ) 
and wider than second to fourth ; second plate long-oval, third and 
fourth long-elliptical ; fifth subquadrate and widened behind, about 
as wide as first. Immediately behind the fifth plate is the hypo- 
pygium, and behind latter is the seventh segment, with the lateral 
ends of sixth dorsal plate enclosing it on the sides. 

In the female of G. lanipes there are seven ventral plates visible, 
the first three free, with ventral membrane showing on each side, 
fourth plate showing ventral membrane only around anterior edge 
and corners, fourth and fifth plates overlapped on sides by lateral 
edges of corresponding dorsal plates, sixth and seventh plates over- 
lapping the corresponding dorsal plates, but sixth overlapped basally 
by fifth, and seventh by sixth, as is to a less extent fifth by fourth. 
Seven segments visible on sides and below, the first shortened, the 
sixth and seventh retracted with only their narrow posterior edges 
showing, the sixth being retracted within fifth and seventh within 
sixth. The seventh segment does not show at all dorsally, though 
the sixth shows equally widely dorsally and ventrally. and sixth and 
seventh show equally widely ventrally. 

Galactomyia lanipes J. C. Fabricius. — As the description of lanipes 
is earlier than that of formosa, the species must be known by the 
former name. Mr. C. W. Johnson, of the Boston Society of Nat- 
ural History, has a pair of this species taken in copula by Mr. P. 
Laurent, at Miami, Florida, March 26, 1901. This pair is men- 
tioned in But. News, November, 1901, page 294. The capture of 
these specimens in copula confirms Brauer and von Bergenstamm's 
statement as to the sexes of this species. Both specimens have the 
palpi black, and the femora wholly black. The male has the abdo- 
men obscure rufous, the female wholly black. The hind femora are 
conspicuously flattened-ciliate distally in the male, but only short- 
bristly-ciliate in the female. Apical cell closed practically in margin, 
tegulae blackish. The female is the form described by Fabricius and 
Wiedemann as lanipes. The male is the form described by Wiede- 
mann as formosa. 

Carolina, Florida, Texas. 


A small female from Costa Rica (Schild and Burgdorf) differs 
only in its smaller size and in having the apical cell rather more than 
short-petiolate. More material is needed to demonstrate its dis- 

Galactomyia tropicalis female. — This is a large robust form, with 
hind femora distinctly ciliate near tip. Body wholly black. Palpi 
lighter colored, bases of femora reddish. Apical cell closed in mar- 
gin. Male not known. Closely allied to lanipcs. (Mexico, Costa 

Galactomyia radiata H. Loew. — Mr. C. W. Johnson has males 
from New Jersey, Pennsylvania, and New York. He also has a 
female specimen collected by him at Delaware Water Gap, New 
Jersey, July 10, 1898, which is doubtless the female of this species. 
It has the palpi yellow and bases of femora yellow. The hind 
femora are short-bristly-ciliate distally. The abdomen is reddish 
yellow, except anal segment, which is wholly shining black including 
narrow posterior border of preanal segment. A female specimen in 
the U. S. N. M., and others that the writer has collected in the Dis- 
trict of Columbia, agree with this specimen in Mr. Johnson's collec- 
tion and are no doubt females of radiata. 

The writer wishes to especially thank Mr. Johnson for kindly 
placing his private collection at his disposal, and for many other 

Subfamily Ameniin^ 

Genus Amenia Robineau-Desvoidy 

Amenta leonina J. C. Fabricius (det. Coquillett). — Australia. 
Both sexes show broad ventral plates overlapped by sides of dorsal 

Subfamily Amphiboliin.e 

Genus Amphibolia Macquart 

Amphibolia fulvipes Guerin (det. Coquillett). — Australian genus. 
This species shows in both sexes posterior triangular views of ven- 
tral plates where the rounded-off posterior corners of dorsal plates 
fail to cover them from view. The male shows a very large paired 
plate-like hypopygial process similar to that of Rutilia. 

Subfamily Rutiliin.E 
Genus Rutilia Robineau-Desvoidy 

Rutilia spp. — The species are all Australian. An examination of 
specimens of both sexes of several species in U. S. N. M. reveals 


the following characters : Neither sex shows any ventral plates, but 
the males show a paired plate-like process widened on apex, occupy- 
ing the position of a ventral plate to the hypopygial segment, and the 
long fifth or anal segment is shortened to a mere margin on venter 
by reason of the hypopygial cavity being pushed strongly forward. 
There are five abdominal segments, the first being rudimentary and 
greatly shortened. 








Of the United States National Museum 

No. 1807 









I. Introduction 3 

II. Itinerary 5 

III. Occurrence of fossils 14 

1. Bonanza Creek localities 15 

2. Little Minook Creek, Junior 17 

3. Little Minook Creek 17 

4. Palisades 17 

5. Nowitna River 22 

6. Yukakakat River 23 

7. Klalishkakat River 23 

8. Discussion 24 

IV. Pleistocene fauna of Alaska 26 

1. Elephas primigenius 27 

2. Elephas columbi (?) 3° 

3. Mammut americanum 3° 

4. Equus 3 1 

5. Bison crassicomis 3 1 

6. Bison alleni 33 

7. Bison occidcntalis 33 

8. Bison priscus (?) 34 

9. Symbos tyrrelli 34 

10. Symbos cavifrons 35 

1 1. Ovibos maxim us 35 

12. Ovibos moschatus (?) 35 

13. Ovis 3b 

14. Alee 36 

1 5. Rangifer 36 

16. Ursus 37 

17. Castor 37 

18. Summary 37 

I. Introduction 

Since the discovery of extinct vertebrate remains in Alaska by 
Otto von Kotzebue, in 1815. while on "A Voyage of Discovery into 
the South Sea and Beering Straits," much interest has been manifest 
regarding the occurrence and cause of extinction of the Pleistocene 


fauna of this northern country ; and. although various expeditions 
have collected specimens and much has been written concerning 
them, it was not until 1904, when the first Smithsonian expedition 
was organized, that the subject was taken up in a systematic man- 
ner. This expedition was conducted by Mr. A. G. Maddren, whose 
report has now been some years before the public. 1 It was planned, 
at that time, to carry on the exploration for two or more consecutive 
seasons, but it was not until 1907 that the present writer was detailed 
to continue the work so well begun three years previous. The report 
herewith presented gives the results of this second trip, undertaken, 
as was the first, under a grant made by the Secretary of the Smith- 
sonian Institution at the suggestion of Dr. George P. Merrill, Head 
Curator, Department of Geology, U. S. National Museum. 
The writer's instructions were, in part, as follows : 

"You are hereby authorized to proceed to Alaska, on or about May 
22, 1907, for the purpose of exploring the regions herein described, 
with a view to securing remains of large extinct vertebrate animals 
and investigating the causes which have led to their extinction. 

"While it is expected that you will exercise your best judgment 
in carrying out the details of your itinerary, it is suggested that on 
leaving the city of Washington you proceed to Seattle, securing at 
that point the necessary outfit, excepting provisions, and arranging 
for the services of a competent assistant. 

"On leaving Seattle you will go by way of Skagway, Alaska, to 
White Horse, and thence down the Yukon River to Rampart, where 
the first stop will be made and the area explored, from which certain 
bison skulls now in the Museum collections have been obtained. You 
will then proceed to Fort Gibbon, exploring the territory in the 
direction of the Xowi River — the so-called "Bone Yard" region — 
and from this point either by steamer or canoe, to Hall Rapids, 
investigating the areas on both sides of the Yukon as far as 

"Should the explorations so far outlined not yield results warrant- 
ing your delay, it will then be advisable for you to proceed, provided 
the season be not too far advanced, by the most expeditious route 
to Kotzebue Sound, and make similar investigations in the areas 
drained by the Buckland River. 

"Should you at any point discover material of such importance as 
to justify the making of immediate excavations, you are authorized 
to undertake such work, though bearing in mind that it may be 
advisable to first make a reconnaissance of the entire field, leaving 
the work of actual excavation until the following year. This is a 
matter, however, which must be left to your discretion. 

"It is expected that the explorations herein authorized will prob- 
ably consume not more than four months of the present year." 

1 Smithsonian Exploration in Alaska in 1904 in Search of Mammoth and 
Other Fossil Remains. Smithsonian Misc. Coll., vol xnx, pp. 1-117. 


II. Itinerary 

In compliance with the above instructions the writer left Wash- 
ington, D. C, May 22, for Seattle, Washington. At this place a 
canoe and the necessary camp equipment were purchased and shipped 
to Rampart, Alaska, where the first active field work was to be done. 
Some time prior to leaving Washington the services of Mr. Benno 
Alexander were engaged. His several seasons' experience with 
various scientific expeditions in the different parts of Alaska made 
him a very desirable companion and an efficient assistant. 

The party consisted of Mr. Alexander and the writer, the plan 
being, as explained in the instructions, to employ such help from 
time to time as might be necessary. 

On May 30 we took passage on the steamer Jefferson, arriving at 
Skagway, Alaska, June 4. It was learned upon our arrival there 
that all accommodations on the first boat down the Yukon had been 
engaged and that it would be best to remain in Skagway until the 
next boat, which was scheduled to sail from White Horse June 12. 
On June 10 we left Skagway over the White Pass and Yukon Rail- 
road for White Horse, Northwest Territory, Canada, the terminus 
of the railway and head of steamboat navigation on the Yukon 
River. Here passage was secured on the river boat White Horse, 
which sailed June 12 and arrived in Dawson, Yukon Territory, 
Canada, June 14. This being a transfer point between the upper and 
lower river boats, we were again delayed because of inadequate 
accommodations, and it was not until June 22 that we left Dawson 
on the steamer Sarah for Rampart. 

The delay at Dawson was profitably spent, however, in examining 
fossils in the possession of citizens of that place ; in making inquiries 
concerning the occurrence of the fossils found in the Klondike 
region, and in visiting some of the localities on Bonanza Creek from 
which many of the specimens examined had been obtained. Scat- 
tered remains of Pleistocene mammals are commonly found in the 
diggings of this region, but the result of diligent inquiry regarding 
the finding of complete or partial skeletons in the mining operations 
conducted here were not encouraging. In only one instance were we 
told of the finding of an accumulation of bones such as would lead 
one to believe that an entire skeleton or any considerable part of the 
skeleton of a single individual had ever been found. The single case 
mentioned was that of the remains of a mammoth (Elephas primi- 
genius) disinterred while sinking a shaft on Quartz Creek in March, 
1904. The skull and tusks were recovered intact (see pi. vn), 


but, according to our informant, although surrounded by a mass of 
other bones, no attempt had been made to preserve them. 

We arrived at Fort Yukon, Alaska, the farthest point north in our 
journey, at midnight June 23, and Rampart (see pi. I, fig. 1), the 
limit of steamer travel, was reached the evening of June 24. While 
here, the area drained by Little Minook Creek, Junior, where scat- 
tered mammal remains had been found, was visited. We were shown 
a few specimens taken out by miners, but the character of their 
occurrence here did not justify a continued search; so, after over- 
hauling our camp outfit and laying in a supply of provisions, we 
loaded our canoe, and on the evening of June 28, left Rampart 
(see pi. 1, fig. 2) for our trip down the Yukon. 1 For thirty or forty 
miles below Rampart the Yukon flows between walls of the older 
rocks with a current of from five to six miles an hour, accelerating 
somewhat as the rapids are reached, near the lower end of what is 
known as the Lower Ramparts. The first alluvial deposits en- 
countered of any considerable thickness after passing the rapids 
were on the right-hand bank some twelve miles above the mouth of 
the Tanana River. Imbedded in these were myriads of small land 
shells representing the living forms, Euconulus trochiformis Mtg. 
and Succinca grosvenori Lea, as determined by Dr. W. H. Dall. No 
vertebrate remains were found. 

Fort Gibbon, a military post at the junction of the Tanana and 
Yukon rivers, was reached the evening of June 30. Here inquiry 
was made regarding localities on the lower river points and par- 
ticularly relating to the Palisades, better known locally as the 
"bone yard," 2 some thirty-five miles below. We were informed that 
scattered fossil remains were also to be found along the Tanana 
River and its tributaries ; but, as the information was somewhat 
indefinite as to exact localities, it was decided not to investigate the 
reports at this time. 

The first exposures of the elevated Yukon 3 silts were observed 
twenty miles below Fort Gibbon, where the bluffs are undermined by 
the river for a half mile or more, and although a careful examination 
was made for the presence of vertebrate fossils, none were found 
either in the face of the cliff or in the talus at its base. This point 
marks the beginning of the escarpment of which the Palisades, some 

'The day we left Rampart a small tusk of the mammoth was brought in by 
some miners from Ray River, a locality from which Pleistocene mammals had 
not been previously reported. 

2 So named because of the great number of fossil bones found here. 
Spurr, J. E. : 18th Ann. Rept. I'. S. Geol gical Survey, pt. in. p. 200. 






fifteen miles farther down, are a part. Covered with a dense vegeta- 
tion, this level-topped bluff or "plateau terrace," as called by Russell, 1 
extends along the left side of the river, only separated from it by a 
heavily timbered flood-plain at its base. The Palisades were reached 
July 3, and two days were spent in the studying of this historic 
locality. Some scattering fossil remains were found, of which a 
more detailed account will be given later. 

The evening of July 5 camp was pitched some five miles below the 
Palisades, at the mouth of "Wasikakat" River, a small tributary 
flowing into the Yukon from the south. This stream, which enters 
the river through a low alluvial flat, was ascended some distance in 
the expectation of reaching a place where it had dissected the higher 
silts of the Palisade escarpment, but we were obliged to turn back 
because of its small size and the consequent difficulty in navigating it. 

The mouth of the Nowitna River 2 was reached June 7. Inquiry 
concerning the occurrence of bones along this stream elicited the 
information from an intelligent Indian, who visited the headwaters 
of this stream occasionally on hunting excursions, that he had seen 
"big horns and other big bones" on the river bars, and a white 
trapper also told us of having picked up the "shank bone" of some 
large animal along the stream. 

The information was stimulating, for it had been planned before 
leaving Washington that this stream should constitute one of the 
principal areas of search. Before leaving Fort Gibbon, three weeks 
provisions had been purchased in the expectation of the supply being 
sufficient for us to reach the headwaters of this stream, the length 
of which, as given by Dall, 3 is one hundred miles. We ascended the 
stream for nine days, and at the farthest point reached, estimated 
to be at least one hundred and seventy to one hundred and eighty 
miles from the Yukon, found it to be a considerable stream still 
(see pi. vi, fig. 1). It may be explained, however, that in a straight 
line the distance covered might not be half of this estimate. Trap- 
pers who have ascended its entire course estimate its total length as 
being two hundred and seventy-five to three hundred miles. The 
Nowitna enters the Yukon from the southwest, about seventy-five 
miles below the mouth of the Tanana. It rises on the eastern flank 
of the Kaiyuh Mountains, and we were told its headwaters are con- 

1 Russell, I. C. : Geological Society of America, vol. 1, 1890, p. 146. 

2 By a recent decision of the United States Geographic Board, this stream, 
which has been successively designated Nozvekaket, Nozvikakat, and Nowi, 
now becomes the Nowitna. 

3 Dall, W. H. : Alaska and Its Resources, 1870, pp. 87-282. 


nected by portages with those of the Innoko and Kuskokwim rivers. 
There are no settlers living on this stream, although deserted winter 
cabins of the lonely trapper were passed several times on our journey. 
The stream flows by a tortuous, meandering course through a low 
alluvial valley covered with a dense growth of alder, willow, poplar, 
birch, and spruce. Its course forms a series of curves alternately 
sweeping from right to left, the channel being confined between 
banks of unconsolidated alluvium and silt from twelve to fifteen feet 
in height. It presents the typical effects of meandering erosion so 
well described by Maddren 1 in his description of the lower reaches 
of the Porcupine, i. c, "cutting away the banks on the concave side 
and depositing the material removed lower down on the opposite side 
as bars" (see pi. vi, fig. i). Often the water has cut in under the 
bank, which extends out over the stream like a great shelf. The 
trees growing on these undermined banks frequently lean far over 
and dip their tops in the water before being carried away. Large 
blocks of the bank, with its superincumbent vegetation, cave off into 
the stream, where they remain standing half submerged for long 
periods. Another feature of the undermined banks is the mantle of 
moss that hangs down from the top like a curtain (see pi. 11, fig. 2), 
as if to hide the destruction the waters have wrought. This blanket 
is composed of the tenacious and closely woven moss and rootlets 
which everywhere cover the ground throughout these lowlands. 

The banks are not sufficiently high to prevent their overflow by 
the spring floods, and the quantity of drift materials lodged in the 
growth on top of the banks indicates the great volume of water that 
flows down during the spring break up. Lanes through the dense 
undergrowth indicate recently abandoned watercourses, many of 
which hold ponds and sloughs. The erosional effects of ice are also 
seen in the scarred and abraded tree trunks and the deep gouges and 
gashes along the higher banks. 

The bars on the lower part of the stream are low and frequently 
covered with stranded trees and other drift materials, but on the 
upper reaches where the bends are more abrupt, they are fairly clear 
of drift and furnish a good path for the "trackers." On some of the 
upper river bars the interstratified sands and gravels have been piled 
in great heaps nearly as high as the inclosing banks. In ascending 
the stream, the first two days good progress was made with the pad- 
dle against the clear but sluggish current, but on the third day, to 
facilitate our movements against the rapidly increasing current, a 

1 Maddren. A. G. : Smithsonian Misc. Coll., vol. xux, No. 1584, 1905. pp. 



i! i 






"cache"' was made of all articles in the outfit not absolutely needed. 
Many times, in order to get over the swift places, "tracking" was 
resorted to, and a little later it was nearly all tracking and wading, as 
we alternately crossed from bar to bar at the bends in the river. 
Nearly evefy bar searched yielded something — either fragments or 
one or more complete elements of skeletons representing the mam- 
moth, horse, bison, and other extinct forms. 

The first of the older series of rocks encountered was some seventy 
to seventy-five miles above the mouth, where the stream has cut the 
end of a low-lying ridge on the right bank. This outcrop is com- 
posed of a mass of badly shattered schistose rock. Some fifteen 
miles farther up, the river again touches the end of a spur of this 
same ridge, exposing rocks of a similar nature. 

Elevated beds of silt of perhaps fifty feet in height were observed 
twice in the ascent, but appeared local in character, and no fossils 
were found in them. 

The ridge paralleling the right bank extended along the river to 
the most distant point reached by us and as far beyond as the eye 
could reach. It rises above the level of the stream from three hun- 
dred to five hundred feet, and is covered with a dense growth of 

The "Suletna," 1 the first important tributary, enters the Nowitna 
from the west ninety miles above its mouth. 

The ascent of the stream was continued until July 16, when an 
inventory of the remaining supplies showed only enough provisions 
to last until we should reach the Yukon again. On this account we 
were obliged to turn back. While the specimens found at the 
farthest point reached were not more abundant or better pre- 
served than those collected farther downstream, it was hoped we 
could reach the very headwaters, to learn, if possible, the source of all 
the scattered bones found along its course, and it was with reluc- 
tance that we abandoned the search. 

The Yukon was reached on the 19th of July, and "Mouse Point, - ' 
a small trading post, the same day. After a short stop here our 
journey was continued to Kokrines, an Indian settlement where the 
Northern Commercial Company maintains a trading post. Some 
little time was spent here in overhauling our outfit, laying in supplies, 
packing fossils for shipment, etc. 

An exposure of elevated silts on the right bank of the Yukon, 
some three miles above Melozi, a United States telegraph station, was 

1 The name by which this tributary is known to the Indians and trappers of 
this region. 


the next area visited. Here members of the Tenth United States 
Infantry had unearthed the almost complete jaws of a mammoth 
shown the writer while at Fort Gibbon. On our visit, however, 
nothing was found. 

"We had been told by Indians, who are in a position to be best 
informed concerning these out-of-the-way places, that large bones 
were to be found on the Yukakakat River, 1 a tributary entering the 
Yukon some seventeen miles above the settlement of Louden. 

The mouth of the Yukakakat was reached on July 23. The ex- 
ploration of the stream occupied the best part of a week, but was 
without especial incident. The farthest point reached was estimated 
to be ninety miles from the mouth, and although the current on the 
upper reaches was swift, it was free from serious rapids and usually 
had along its shores bars sufficiently broad to give good "tracking." 

The sluggish current of the first few miles of its meandering course 
flows through a low alluvial flat, heavily wooded and very similar in 
character to that part of the Xowitna. Farther up, however, the 
course of the stream is flanked by low ranges of hills which grad- 
ually converge and thus confine its wanderings to a shorter radius. 
On either side of the stream back of the low hills mountains were 
observed rising from one to two thousand feet in elevation. 

In many places the growth on the banks was very sparse, and 
consisted principally of scattering clumps of alders, willow, and birch 
interspersed with a few stunted spruce trees. Here and there back 
from these low banks were many shallow lakes that furnish splendid 
breeding grounds for the geese and ducks which abound there On 
the uppermost part of the stream reached by us the shores were 
more heavily timbered and there were long straight stretches of 
river flowing between banks from ten to twelve feet in height, which 
in most cases were covered with undergrowth and a tall luxuriant 
growth of grass extending nearly down to the water's edge. At the 
bends the undermining of the concave side presented features similar 
to those observed on the Xowitna River. 

The first elevated silts of any importance observed were some 
sixty miles upstream from the Yukon, where the river makes a 
right-angled bend. At a comparatively recent date the river at this 
point has changed its course, and at the time of our visit was not 
cutting the bluffs (see pi. in). It could undermine them only at an 
extremely high stage of water. These cliffs have almost perpendic- 
ular faces and are from eighty to one hundred feet in height, com- 

1 This stream appears to be known in Alaska as the Yukakakat, although 
Dall has indicated it on a map compiled by him in 1875 as the Soonkakat. 


posed mostly of fine light-colored, unstratified silts. Some sixty feet 
down from the top is a layer of coarse gravel conformable with the 
silt, which may represent the Palisade conglomerate of Spurr. 1 

This terrace at irregular intervals has been dissected somewhat by 
the drainage from above (see pi. in) . In many places along the 
cut banks of the stream the silt was underlaid by a stratum of rather 
fine reddish-colored gravel. A section of these flood-plain deposits, 
when no complications occur, presents the following divisions in 
their natural order and approximate thickness : 

Layer of peat 18 inches to 2 feet 

Layer of fine silt ' 8 feet to 10 feet 

Fine reddish gravel 4 feet to — 

DalF noted the occurrence of a similar fine reddish gravel in the 
deposits of Eschscholtz Bay. 

A few scattered bones were collected on the bars below the depos- 
its of elevated silts just described, but although continued search 
was made for two days upstream from this point, no fossils were 
found. Even though no indication of vertebrate remains were seen 
in the silts, the writer is inclined to the opinion that the few frag- 
mentary specimens picked up on the bars below may have been 
washed out of these bluffs and carried downstream by the river dur- 
ing a flood stage. This idea is strengthened somewhat from the 
fact that no mammal remains were found in the lower cut banks or 
alluvial deposits of either this stream or the Nowitna, although 
persistent and continued search was made, and from our own 
experience and that of others we do know they occur in the elevated 
lacustral phase of the silts. 

The absence of fossil evidence on the last two days of our ascent 
and the fact that little had been found previously showed that this 
stream did not cut an extensive deposit of Pleistocene mammal 
remains, and it appeared to be a waste of time to continue our 
search : so we returned by the same route we had ascended, reaching 
the Yukon on July 30. 

A short distance above Louden we met Air. R. A. Motschman, 
who. being thoroughly familiar with the region, told us of several 
localities where fossils had been found. The most important of 
these was an exposure on the Klalishkakat River, a locality visited 
by Air. Arthur J. Collier, of the V. S. Geological Survey, some five 
years previous. At the time of Mr. Collier's visit a large tusk was 

1 Spurr. J. E. : iSth Ann. Rept. U. S. Geological Survey. 1896-97, p. 199. 
"' 17th Ann. Rept. U. S. Geological Survey, pt. t, 1895-96, p. 852. 


protruding" from the bank, a picture of which is shown on plate II, 
fig. 2, in Mr. Maddren's account of his trip in 1904. 

This small stream enters a branch of the Yukon from the south 
three miles below the settlement of Louden. At the time of our 
visit there was a high stage of water, and it was with some difficulty 
that we made the comparatively short distance upstream to the point 
where the river cuts the elevated silts. That portion of the bluff 
where Collier had photographed the tusk in place had been under- 
mined and washed away. Scattered fragments of fossil ivory found 
by us on the bars below probably tell the story of its disappearance. 
A few fragmentary bones were found, some imbedded in the undis- 
turbed silt and others in the talus at its base. 

Eight miles below Louden, on the right bank of the Yukon, occurs 
a typical exposure of the Yukon silts. The bluffs extend for a dis- 
tance of perhaps two miles and present faces from two hundred to 
two hundred fifty feet in height, equal to those of the Palisade 
escarpment, which they resemble in all their stratigraphic detail. 
Mr. Motschman told us of finding fossils here, but not even a frag- 
ment was secured at the time of our visit. 

Here, it was observed that the wind is quite a factor in the erosion 
of these bluffs. The fine silt dries rapidly, and as it commences to 
sift down the precipitous face it is caught by the currents of air and 
carried away. From a distance this silt-laden air, as it poured up 
over the crest of the bluff, reminded one of an ever-ascending vol- 
ume of smoke. In places large drifts had accumulated like so 
much wind-drifted snow. 

Nulato, an important Indian village, was reached on August 2, 
and Kaltag on August 5. Here the Government telegraph line that 
extends down the river leaves the bank of the Yukon, ascends Kaltag 
River to near its head, crosses the divide to Unalaklik River, and 
descends that stream to Norton Sound, a total distance of one hun- 
dred miles. 

Inquiry here concerning localities on the Kaltag River failed to 
elicit information of enough importance to warrant investigation ; 
so canoe travel was resumed to Anvik, some two hundred miles 
below Nulato. Many stops were made to examine silt deposits, but 
in only two places were fossils found. Some five or six miles above 
Hall's Rapids, on the right bank, bones of the mammoth and bison 
were collected at the foot of the silt bluffs, and again above the old 
station of Greyling, some twenty-five miles above Anvik, where the 
silts are exposed for two or three miles by the cutting of the river. 
Here, during the summer of 1907, a fine pair of lower jaws of 
Elcphas were picked up by Mr. W. C. Chase, of Anvik. and pre- 


sented by him through the writer to the Smithsonian Institution. 
The Rev. J. W. Chapman, of the same place, also had specimens in 
his possession from this locality. 

It was planned before reaching Anvik to explore the area drained 
bv the Anvik River, as some years previous, while visiting this 
place. Mr. A. H. Brooks, of the U. S. Geological Survey, had been 
shown fossils by the Indians said to have been collected along the 
banks of this stream. Inquiry here among both the white and 
native inhabitants, many of whom are thoroughly familiar with the 
river and the country drained by it, developed the fact that, so far 
as they knew, no fossils had ever been found in the region. Never- 
theless, we ascended this stream some distance to fully satisfy our- 
selves as to the conditions prevailing there, but nothing in the nature 
of a fossil vertebrate was found. It appears quite probable that 
the specimens shown Mr. Brooks came from the deposits near 

Upon our return to Anvik we were delayed some few days by 
continued rains from resuming our journey down the Yukon. At 
Holv Cross, a Catholic mission, fifty miles below Anvik, we were 
told of the occurrence of large bones in the banks of one of the 
sloughs leading to the portage to the Kuskokwim River. Difficulty 
in securing the services of a competent guide deterred us from 
making an investigation of this locality, which was some distance off 
from the Yukon. 

Russian Mission was reached August 25, and Andreafski, where 
our canoe trip ended, on August 29. The almost incessant rains, 
accompanied by winds, during the last ten days of canoe travel were 
the most annoying feature of the whole trip. On several occasions 
it became necessary to go ashore and wait for the wind to abate, for 
fear of being swamped by the high waves encountered. 

In the two months spent upon the Yukon and its tributaries, after 
leaving Rampart, we traveled by canoe alone nearly fourteen hun- 
dred miles. 

At Andreafski passage was secured on the river boat D. R. Camp- 
bell, for St. Michael, which was reached September 1. 

Here it was learned fossils were occasionally found on the main- 
land shore across the bay, and this area was investigated, but no 
success was met with. 

Nome was reached by the local steamer Yale on September 7. 

The autumn season being too far advanced to undertake an ex- 
ploration of the Eschscholtz Bay and Buckland River localities, we 
took passage on the ocean steamer Nortlncesteni from Nome Sep- 
tember 20, and Seattle, Washington, was reached on September 29. 


We were not successful in finding that which was most desired. 
a fairly complete skeleton of a mammoth, but the expedition was by 
no means barren of results, as will be noted later. 

III. Occurrence of Fossils 

The scattered remains of Pleistocene animals occur throughout 
the ungiaciated region of Alaska and adjacent Canadian territory in 
three quite distinct deposits : First, in the black muck accumulated 
in gulches and the valleys of the smaller streams ; second, in the fine 
elevated clays of the Yukon silts and Kozvak clays; and, third, in the 
more recent fluvial and alluvial deposits. The specimens as found 
have been disinterred either through the erosive agency of the 
streams or by the work of the miner in the operations conducted 
in search of gold. 

Although so generally distributed, there have been reported, so 
far as known to the writer, but two well-authenticated occurrences of 
accumulations of bones under such conditions as to suggest an 
original entombment. While the writer was shown bones pro- 
truding from the face of the undisturbed beds in the Klondike 
region (see pi. iv, rig. i), and in other instances collected specimens 
actually imbedded in the elevated silts along the Yukon River, they 
were in all cases disarticulated and scattered, and there was no 
evidence of an association of any of the parts found. 

Diligent inquiry was made among miners, trappers, and other 
residents of Alaska, met along the route traveled, concerning what 
they knew of the occurrence of fossil specimens. While nearly all 
were familiar with the fragmental and scattered parts, very little 
information was elicited of an accumulation of bones that would 
lead one to believe a skeleton or even a part of a skeleton had ever 
been found together in any one place. 

While the scattered depositions occur as separate bones, skulls, 
teeth, tusks, horns, etc.. throughout the formations mentioned, the 
condition of the specimens found varies greatly. Some are in such 
a good state of preservation they certainly could not have traveled 
far from the original place of interment, while on the other hand 
many bones are broken, abraded, and water-worn, and show unmis- 
takable evidence of having been carried considerable distances. 
Bones representing these several phases were often found com- 
mingled and occupying relatively the same positions, whether it be 
in the muck, on a river bar. or imbedded in the undisturbed silt 

Smithsonian miscellaneous collections, vol. 51 





The best-preserved specimens coming under the observation of 
the writer were those seen at Fox Gulch, on Bonanza Creek, in 
Yukon Territory, Canada, some twelve miles distant from the city 
of Dawson. On account of the excellent state of preservation of 
many of the- specimens found here and the fact that they occur in 
what may be considered as an approach to a primary deposition, a 
somewhat detailed- description of this locality will be given. 

Bonanza Creek Localities 

Bonanza Creek empties into the Klondike River about a mile and 
a quarter above Dawson. The valley is trough-like in character and 
follows a sinuous line bending from right to left. The present valley, 
according to McConnell, 1 has been cut down through the floor of an 
older valley. At irregular intervals the sides of the valley have been 
dissected by gulches. Magnet and Fox Gulches (see pi. ix (x) ), 
on the left-hand side, are the most important from the standpoint of 
vertebrate fossils. Gold has been found in both, and at the time of 
our visit hydraulic operations were being carried on here by the 
Yukon Consolidated Gold Fields Company. In the prosecution of 
this work the content of the entire gulch to bed-rock was being 
sluiced down (see pi. vi, fig. 2), the talus spreading out fan-like into 
the creek bed below. 

In the talus from Magnet Gulch representative parts of the mam- 
moth, horse, bison, and moose were picked up. 

At Fox Gulch we were shown many fine skulls and other skeletal 
parts of Blephas, Bison, Bquus, and Alee. On the bank near the 
working face was a complete skull of the mammoth beside two bison 
skulls, recently uncovered (see pi. iv, fig. 2), and protruding from 
the face of the undisturbed muck was a large tusk (see pi. iv, fig. 1) 
and the skull of another bison. These had been uncovered the morn- 
ing of our visit. 

Fox Gulch is a short, deep gulch that has been cut down through 
the quartz drift and "White Channel" deposits and deep into the 
present bed-rock. The bed-rock is covered with a thin layer of 
rather coarse gravel on top of which is a thick layer of muck (see 
fig. 1). The gold occurs in the gravel underlying this muck, and in 
order to reach it the mass of superincumbent material is washed 
down by the powerful streams of water from the nozzles of the 

1 McConnell, R. G. : Preliminary Report on the Klondike Gold Fields. Yukon 
District, Canada. Geol. Surv. Canada, No. 687, 1900, p. 21. 




The working face as we saw it varied from twenty to forty feet in 
height. It is in the bottom part of the muck that the fossils are 
found. Those seen in place were from twelve to eighteen inches 
above the layer of gravel, and upon inquiry it was learned that all of 
the specimens taken out here had come from approximately the same 

The muck and gravel, which rest unconformably upon the under- 
lying rocks, is solidly frozen, but thaws rapidly under the heat of 
the summer sun, and large pieces were continually dropping during 
our examination of the face. This thawed material emitted the dis- 
agreeable odor of decomposing organic matter, a phenomenon 
observed by many others, particularly Dall, 1 who attributed it to 

Fig. 1. — Cross-section of Fox Gulch, Bonanza Creek, Yukon Territory, Canada. 

a. "White channel" gravels and quartz drift; b. Muck; c. Bed rock; d. Layer 

of logs, limbs, etc. ; x. Level where fossils occur. 

decaying animal flesh or to dung of the mammoth or other herbiv- 
orous animals. The present writer agrees with Mr. Maddren, 2 who 
attributes it to the gases from decaying vegetable matter, of which 
the deposits are largely composed. 

Interbedded with the muck in Fox Gulch was a layer of wood, 
represented by many fair-sized sticks (see (d), fig. 1), their ends in 
many instances being much rounded and water-worn. 

Many of the fossils found here were beautifully preserved. For 
example, several of the bison skulls had the external horn, the entire 
dentition, and the frail, delicate bones of the anterior portion of the 
face remaining intact. The conditions are unusual, for, as a rule, 
only the horn cores and the heavier and stronger parts are found, and' 
it is upon such fragmentary specimens that the descriptions of most 
of the extinct species of bison of this continent are based. Stranger 
still, however, is the fact that here no parts of these animals are 
found articulated or even so associated that skeletons might be 
assembled. All of the material is dismembered and scattered. 

1 Dall, W. H. : 17th Ann. Rept. U. S. Geol. Sun., pt. 1, 1895-96, p. 853. 

2 Maddren, A. G. : Loc. cit, pp. 64-65. 


The preservation of the horn sheaths, as in the cases of the bison 
skulls, and the completeness of many of the skulls and other elements 
show they have not been subjected to the rough usage incident to 
their removal from one place to another ; nor after death could 
they have long lain on top of the ground exposed to the vicissitudes 
of the elements. The external horn would in such case be the first 
to disappear, as all know who have visited our western plains and 
have noted the almost total disappearance of the horn sheaths from 
the buffalo skulls scattered about. Their destruction, even in a dry 
climate, has been accomplished in a comparatively few years. 

Little Minook Creek Junior 

This small stream enters Big Minook Creek from the right some 
six miles distant from the town of Rampart. Here, as in the 
Klondike region, the fossils occur in the lower part of the muck, 
which covers everything from two to twenty-five feet in depth. 
Specimens would be uncovered here only through the agency of 
mining, as the volume of water in the creek is not sufficient to cut 
away the banks. 

While sinking a shaft on claim Xo. 21, operated by Messrs. 
Bowen and Coole, a skull of Bison crassicornis (No. 5727. U. S. 
National Museum) associated with bones of Elephas was taken out 
twenty feet below the surface. 

Some years previous to our visit, we were told, the tusks of a 
large "mastodon" (mammoth) were found in a shaft sunk on a 
claim above No. 21. 

Little Minook Creek 

This creek is also a tributary of Big Minook Creek, and here as 
in other localities, the fossils found occur in the lower layers of the 

In the vertebrate fossil collection of the U. S. National Museum 
is a portion of the skull of Bison alleni (No. 2383, see plate xi) 
from this locality having the entire horn sheaths preserved. 

Mr. J. B. Duncan, of Rampart, presented to the Smithsonian Insti- 
tution, through the writer, a skull of Bison crassicornis (No. 5726. 
U. S. National Museum, see plate x) from one of the claims on 1 1 1 i s 
creek, and Mr. C. B. Allen, of the same place, presented the Institu- 
tion with the calcaneum of Blephas from claim No. 1. 


The Palisades, or ''Bone Yard," on the left bank of the Yukon, 
thirty-five miles below Fort Gibbon (see pi. v. fig. 1), has long been 


famous as a locality for vertebrate remains. This escarpment has 
been described by Russell, 1 Spurr, 2 and later by Collier 3 and Mad- 
dren. 4 

The bluff region extends for a mile or more down and around an 
almost right-angled bend in the left-hand channel of the river (see 
fig. 2). The bluffs, from one hundred and fifty to two hundred feet 
in height, and solidly frozen, are composed principally of an ex- 
tremely fine silt, greenish gray in color and showing no traces of 
stratification. Their almost perpendicular faces are being contin- 
ually undermined (see pi. v, fig. 2) by the swift current causing 
large masses to break off. many times with a startling report and 
subsequent splash, as they fall into the water below. Often during 
the two days' stay here the report sounded so like the firing of a gun 
that we were startled by the sharpness of it. 

Xcar the lower end of these exposures the bluff's have been ele- 
vated somewhat, exposing the gravels which underlie them. These 
last have been called by Spurr the Palisades conglomerate, and it 
has been suggested they may be of Pliocene age. The top of the 
bluffs extend back from the river as a level, densely wooded table- 
land. In several places small watercourses have dissected this 
table, forming deep gorges. Near their mouths, where they enter 
the Yukon, their levels are but little elevated above its high-water 

At the up-river end of the bluffs we found numerous bones of the 
mammoth in the debris from a recent slide, and a short distance 
farther down (2 on map fig. 2) the scattered elements of a bison 
were found securely imbedded in a huge block of silt not long since 
displaced from its original position higher in the face of the cliff. 
The sacrum, part of the pelvis, two dorsals, rib, and scapula were 
the parts recovered. The scapula (shoulder-blade) was quite com- 
plete (see fig. 4), which, on account of its frail nature, appeared 
rather remarkable, the heavier and stronger bones being broken and 
abraded before their interment here. 

The small streams mentioned previously as dissecting the bluffs 
were followed inland for considerable distances, and although their 
banks in many places presented very clean-cut exposures of the silt, 
no evidence of the presence of fossil remains was found. However, 

1 Russell, I. C. : Notes on Surface Geology of Alaska. Bull. Geol. Soc. Am., 
vol. 1, 1890, p. 122. 

2 Spurr, J. E. : Geology of the Yukon Gold District. 18th Ann. Rept. U. S. 
Geol. Surv., pt. 3, 1898, pp. 200-221. 

'Collier, A. J.: Bull. No. 218, U. S. Geol. Surv., 1003, pp. 18 and 43- 
4 Maddren. A. G. : Loc. cit, pp. 17-18. 






l 9 

among the debris of driftwood and other vegetable material accumu- 
lated at their mouths many disassociated bones were recovered (see 
( + )> fig- 2 )- The concentrating action of the water in carrying 
away the fine silt and leaving the heavier objects behind would 
account for thejr abundance here. 

In drifting down along the base of the cliffs in the canoe, the skull 
of an Ovibos sp. nov. (No. 5728, U. S. National Museum) was 
found on a narrow shelf just above the point (3 on map, see fig. 2) 
where the underlying gravels first appear. That the skull came 

., mm 


Fig. 2.— Sketch Map in Vicinity of "Palisades." 

1. Where section was taken, shown in Fig. 3; 2. Bison bones (Fig. 5); 

3. Musk ox skull, No. 5728, U. S. National Museum. 

down from the cliff above there can be no doubt, for it lay on a pile 
of talus accumulated since the last high stage of water. The high- 
water marks were still plainly evident on either side and above the 
heaps of detritus. Moreover, the cranial and other cavities of the 
skull were filled with the fine silt composing the bluff. This skull 
was in fairly good condition, having, beside two of the molars, some 
of the bones of the anterior part of the face in a good state of preser- 
vation. The worn and abraded appearance of most of the fossils 
here indicates that they are drift and not in a place of primary 


Maddren 1 makes the observation, "There is a little ice on top of 
these bluffs, but nothing like the extensive development exposed in 
the Old Crow Basin." Ice was also observed here by the writer, 
which on first sight appeared to represent the typical ice-bed deposits 
of many other localities. Upon closer examination, however, it was 
found to be a superficial layer on the face of the exposure and not 
a continuous ice-sheet interstratified with the muck and humus. 
The formation of this layer of superficial ice appears to be of interest 
and it may explain the presence of apparent ice-beds in some other 
places. Moreover, it does show that caution should be exercised in 
pronouncing all ice on the face of a cliff as being a section of a con- 
tinuous bed. 

At (i) on the map (fig. 2), a deep depression or basin in the top 
of the silt has been filled with alluvium and mucky material. The 
brow of the escarpment here, three or four hundred feet back from 
the edge of the stream, was estimated to be one hundred and fifty to 
one hundred and seventy-five feet above the level of the river. The 
Yukon, having cut laterally into the center of this basin, has left the 
remaining muck resting on a slope of silt inclined toward the river 
(see cross-section, fig. 3). By the undermining of the face of the 
cliff, one block after another of this frozen muck has broken away 
from its original position in the face of the escarpment and moved 
riverward. In most instances this movement has been so gradual 
that the blocks retain their upright positions and carry with them 
the superincumbent turf and vegetation undisturbed. The thawing 
of their faces and subsequent wasting away has allowed the turf to 
bend down without breaking, thus affording protection against 
further disintegration. The final destruction of the blocks, as they 
eventually fall into the stream (where several were seen half sub- 
merged), has resulted in leaving a basin-like area of an acre or more 
in extent devoid of its former covering of from thirty to forty feet of 
muck, except that here and there are masses recently detached from 
the walls of the basin. The inclosing walls or faces of the basin are 
perpendicular and from twenty to thirty feet in height. From three 
to four feet below the top of the walls was a layer of ice. Upon first 
sight it had every indication of being a section of a continuous bed. 
Some of the detached blocks standing in the center of the basin 
showed ice on both front and back faces. The top of this ice was 
straight, but the lower margins were irregular when not covered by 
the detritus at the foot. The face of the ice was also irregularly 
melted, due to the more exposed position of some part<. 

'Maddren : Loe. cit, p. 18. 



Upon ascending to the top of the escarpment at the point most 
remote from the river, it was found that a mass of frozen muck, 
estimated to be two hundred feet long and fifteen to twenty feet in 
thickness, with a vertical face of twenty to thirty feet, had moved 
outward at its, center for fully fifty feet, but had not yet become 
detached at its ends. The crevasse formed by this displacement was 
filled by water to such a depth that the bottom could not be found 
with a long pole. Back of the crevasse, in the surface of the bluff. 
were numerous parallel cracks varying from six to eighteen inches 
in width and many feet in length. These had water standing in 
them nearly to the top of the ground. The conditions observed 
here appeared to the writer to explain the presence of the ice on the 

Fig. 3. — Cross-section of "Palisades"' Escarpment, showing Formation of 

Superficial Ice. 

1-2-3. Blocks of frozen silt ; 4-5. Water level of the Yukon ; 4-6. 150-170 
feet; 7. Crevasse filled with water; 8. Ice on faces; 9. Overhanging turf; 
10. Lacustrian silts; n. Detritus (thawed muck). 

faces below. With the advent of winter, assisted by the already 
frozen ground, the water in the crevasses becomes frozen solid. A 
subsequent outward movement of the blocks would leave the ice 
clinging to the face of either the cliff, or the block, or both, and 
under the influence of the rays of the summer sun would rapidly 
smooth the broken and ragged edges. On the faces of blocks I and 
2 (see fig. 3) such layers of ice were observed, and where protected 
by the wet mantle of overhanging turf and moss were thawing very 
slowly. In places the ice was so thin the writer with a few strokes 
of his pick was able to penetrate it and into the frozen muck wall 
behind. Sections of the ice, protected by curtains of turf and falling 
debris, would persist for considerable periods. In places it had 
melted away, leaving its mould in the face of the cliff. 


Under the influence of the summer sun the blocks were gradually 
disintegrating. Large pieces were continually falling as thawing 
progressed, and all along the bases of the face and around the blocks 
were small piles of talus of the mucky material. 

The same pungent, disagreeable odor of decaying organic matter 
was noticed here as in the deposits of Bonanza and Minook creeks. 
The stench was so strong it could be easily detected on the river a 
considerable distance away. In many places on the wet muck banks 
a rusty red fungus-like plant grew in extensive patches. 

The writer does not wish to be understood that the observations 
recorded here apply to all ice deposits, but as a local phase it may 
explain the occurrence of many so-called "ice-beds." It may also 
help to explain the position of the mammoth found frozen in the 
cliff along the Berezovka River in Siberia in 1901. From the posi- 
tion in which the carcass was found it would appear as though he 
had fallen into a crevasse from which there was no escape. The 
description 1 of the locality is not so unlike the conditions observed 

Nowitna River 

The exploration of this stream added but little information con- 
cerning the occurrence or derivation of the fossils found along its 

After the first day of our ascent of this stream nearly every bar 
yielded some fossil evidence, either in the shape of a tooth, limb 
bone, vertebra, or scattered fragments. The specimens found were 
in various stages of preservation ; many broken, others entire, some 
badly water-worn, and a few as perfect as the day they performed 
their functions in the skeleton itself. Some elements, which on ac- 
count of their frail nature should by the very character of their 
structure have been broken and abraded, were found complete. 

In examining the bars we soon came to know that the up-river 
ends, where the materials composing them were coarsest, was the 
most favorable part for finding the scattered bones. The remains 
without exception were all found below the high-water level of the 
flood stages of the river, and were without question brought down 
from some source or sources of deposition, either by the water itself 
or by floating ice. 

A close examination was made of the low-cut banks and elevated 
silts, but not in a single instance were fossils actually found in place. 

The conditions on this stream differed somewhat from those found 

1 Herz, O. F. : Frozen Mammoth in Siberia. Ann. Rept. Smithsonian Inst., 
1903- PP- 611-625. 





Sluice box in the ctnter, and the muck filling of the gulch not yet sluiced out may be 
seen in the background on the right 


by Mr. Maddren on the Porcupine and Old Crow rivers, from the 
fact the fossils did not become more abundant on the bars as we 
went upstream. On some bars many fossils would be found, while 
others would yield only a single specimen. The varying degrees of 
preservation exhibited by the specimens points to the conclusion that 
the source of supply is diverse and not one large deposit. The 
writer is inclined to the opinion that the fossils found on the bars 
have been washed out of the silt banks along the stream and trans- 
ported to their present resting places largely by the action of the 

The finding of abundant remains on the bars of a stream that is 
cutting elevated silts does not necessarily lead to the conclusion that 
all of the specimens found there have come from the headwaters of 
that stream, for we know that scattered bones occur in the silt depos- 
its, and it appears that the bones brought down from far upstream 
may be augmented in numbers by those washed out of the silts along 
its course. 

The following list gives the fauna of this area as represented by 
the scattered bones collected : 

Blephas primigenius. 







Although fewer fossils were collected along this stream, the pre- 
vailing conditions as to their occurrence were found to be similar 
in most respects to those observed on the Nowitna River. 

The following forms were recognized : 

Blephas primigenius. 

Klaus hkakat River 

A locality on this stream some three miles inland from the Yukon 
was visited. Here the bluffs present nearly perpendicular faces from 
sixty to eighty feet in height, the lower parts of which are com- 
posed of reddish cross-bedded gravels, varying from fine to very 
coarse and unconformable with the overlying silt. The silt shows no 
traces of stratification and is solidly frozen. Back from the bluff 
is a level tableland, bordered on all sides, except that adjacent to the 
river, by low hills. It was at this locality that Mr. Collier in 1902 


photographed a tusk protruding from the face a few feet below the 
top of the escarpment. 

The bases of the bluffs are washed by the stream, and during 
stages of high water are undermined, causing large masses to break 
off. The tusk seen by Collier five years previous had disappeared, 
but a recent slide had exposed the distal end of a femur of Elephas 
in place about three feet above the underlying stratum of gravel. 
Other broken fragments were found in the loose clay of the talus 
along the foot of the bluffs. The silt varies in thickness from thirty 
to thirty-five feet, and broken and abraded fossil remains occur, scat- 
tered throughout. The conditions here are not favorable for the 
securing of good specimens. Bones of Elephas and Bison were col- 


After a review of the conditions prevailing in localities where fos- 
sils have been found in Alaska and contiguous territory, the writer 
feels inclined to dissent somewhat from the views expressed by 
Maddren regarding the most promising collecting grounds. 

Mr. Maddren 1 has advanced the opinion in the following state- 
ment that the old lake shores offer the greatest inducements : 

"That the fluvio-glacial Pleistocene lakes of Alaska were subject 
to annual winter freezing, at least at various stages of their existence, 
there appears no doubt, because scattered apparently indiscriminately 
through the clays, at varying depths and considerable distances from 
the former shore lines of these basins, are some mammal remains. 
Their positions can only be accounted for by supposing they were 
carried out on the waters of the lakes from the adjacent shores or 
tributary streams by ice during spring breakups and freshets, there 
to be dropped by its melting to their present positions interbedded in 
the silts. There appears no other logical way of explaining the 
presence of these bones in the lacustrine areas." . . . 

"The main point is that the remains occur in the silts as scattered 

"The animals from which they were derived probably died about 
the shores of these lakes, and it is these Pleistocene lake shores we 
must examine carefully if we are to obtain anything like complete 
remains of the mammals inhabiting the region at that time." 

There appears to be one objection to this hypothesis as applied to 
these fine-silt deposits. If the great number of isolated mammalian 
bones scattered through it were carried out from the shores and 
tributary streams by ice, it is hard to understand how they could be 

1 Maddren. A. G. : Smithsonian Misc. Coll., vol. xux, No. 1584, 1905, p. 26. 


selected for distribution in deposits from which all other large frag- 
ments of detrital materials are absent. 

It might be explained, however, on the supposition that the bones 
have been carried out from muck deposits in which there is no heavy 
detrital material. In that event many of these deposits might be 
considered older than are the silts ; or, the presence of interbedded 
layers of lignite at the "Palisades" and in the silts of Cooleen basin 
(which would indicate a local drainage or elevation of these beds at 
one time) might furnish the necessary conditions for the accumu- 
lation of animal remains, followed by subsidence and further depo- 

Up to this time the best-preserved remains have been found in 
the deposits of muck accumulated in gulches and the valleys of the 
smaller streams. Typical examples of the occurrence of this muck 
may be seen on Little Minook Creek, near Rampart, Alaska, and 
Bonanza and other creeks, near Dawson, in Canadian territory. 
Only a single skull of bison with the horn sheaths preserved is 
recorded as coming from the silt, while they are of common occur- 
rence in the muck. Their presence here may be accounted for on 
the supposition that the animals became mired in the bogs before 
they became solidly frozen as they are now. This naturally raises the 
question: If mired down in such a place, why is it that the remains 
should be so universally scattered? The writer suggests that they 
may have been separated by the creeping of the muck or peat — a 
phenomenon familiar to all students of deposits of this nature. By 
such creeping the muck may have moved considerable distances, 
particularly where the floor is inclined, as in many of the gulches. 
From the fact that most of the bones occur in the lower layers of 
the muck, no matter what the depth of the deposits may be, it is 
apparent that their specific gravity has caused them to sink to their 
present resting places. Thus it would not be necessary for the 
extermination of the fauna to have taken place at one time, as might 
be inferred by their occurrence at one level. 

It was from the muck forty-two feet below the surface that the 
skull and tusks, surrounded by other bones of the skeleton of Blephas 
primigenius shown in plate vn, was obtained. Mr. A. H. Brooks, 
of the U. S. Geological Survey, tells me of seeing a portion of a 
skeleton of Elcphas from Woodchopper Creek, Alaska, probably 
taken from a similar deposit. 

The two instances just cited undoubtedly represent places of 
primary entombment, and the manner of their occurrence appears to 
approximate the conditions found in the bogs and swamps in the 


Eastern States, from which many of the best skeletons of the 
Mastodon, have been obtained. 

From the evidence reviewed the writer believes that the deposits 
of muck represent the most likely places from which to secure 
remains of this extinct fauna. 

The writer takes this opportunity to express his appreciation for 
the assistance given him by Mr. A. H. Brooks and Mr. A. G. 
Maddren, of the U. S. Geological Survey. Many services were 
rendered by residents of Alaska along the route traveled, and favors 
were extended by agents and officials of the Northern Commercial 
Company. Mr. J. B. Duncan, of Rampart ; Air. Frank Haslund, of 
Kokrines, and Mr. Frederick, of Andreafski, were especially kind 
in many ways. My thanks are also due Mr. J. W. Gidley of the 
National Museum, for help in the identification of specimens. 

IV. The Pleistocene Fauna oe Alaska. 

Although a number of species have been described from the Pleis- 
tocene deposits of Alaska, they have for the most part been based 
on fragmentary, and therefore rather unsatisfactory, specimens. In 
many cases the principal osteological and dental characters are not 
known, and on that account it is not always possible to compare them 
intelligently with related forms. 

Only a few of the large number of localities where fossils have 
been found furnish well-defined specimens, capable of specific 
determination, and while these vertebrates are interesting from the 
standpoint of their general geographical distribution, they are of 
comparatively little aid in the interpretation of the local deposits. The 
forms have been entombed under such exceptional conditions as to 
raise some question regarding the exact age of the deposits in which 
they are many times found, although they could not have antedated 
Pleistocene time. A glance at the list of determinable species is 
sufficient to show at once that they represent a typical Pleistocene 
fauna, some of which, as the moose, caribou, musk-ox, sheep, bear, 
and beaver, have persisted down to the present day. 

To aid the student, there is given here a list of the various genera 
and species thus far reported as occurring in Alaska, followed by 
a brief review of each, with a reference to the original description ; 
the condition and present location of the type specimen (if known, 
and when based upon fossil remains) upon which these were 
founded, and in some cases figures of representative specimens from 
Alaska. Some additional information has been derived from a 
Study of specimens in the vertebrate paleontological collection of the 


U. S. National Museum, collected in Alaska by Lieutenant Hooper, 
Dr. W. H. Dall, E. W. Nelson, L. M. Turner, A. G. Maddren, and 


The Northern Mammoth 

Ekphas primigenius Blumenbach, Handb. Naturg., 1st French ed., vol. 11. 
1803, p. 407. 

Description. — "Jaw broad and rounded ; profile in front of tooth 
row almost vertical ; enamel folds narrow and compressed ; rather 
more than two folds to the inch, or twenty-four in ten inches ; enamel 
itself thin." 1 

Remarks. — This species is, geographically, the most widely dis- 
tributed of extinct elephants. It has been reported as ranging from 
Florida, Texas, and Mexico on the south and northward into Can- 
ada and Alaska. It is also found in Great Britain and nearly all 
Europe and northern Asia. 

Its remains are particularly abundant in parts of Alaska and 
Siberia. As yet no complete specimens have been found in Alaska, 
although several good skulls and nearly all parts of the skeleton 
are known from scattered but well preserved bones. Neither have 
specimens been found in the flesh, as is so often reported through 
the columns of the newspapers and even by some of the magazines. 

The size of the mammoth has been so grossly overestimated by 
the general public that a few comparisons may help to correct some 
of these false impressions. The largest mounted specimen known is 
the skeleton in the collection of the Chicago Academy of Sciences, 
obtained in 1878 from Spokane County, in the State of Washington. 
The height of this animal when alive has been estimated to be thir- 
teen feet. The African elephant "Jumbo" was eleven feet high, 
and there have been other elephants recorded as measuring twelve 
feet in height ; so, as this would indicate, there is not so much differ- 
ence in size between the mammoth and living elephants as is often 

Mr. Lucas says : 

"Tusks offer convenient terms of comparison, and those of a fully 
grown mammoth are from eight to ten feet in length, those of the 
famous St. Petersburg specimen and those of the huge specimen in 

1 Lucas, F. A. : Systematic Paleontology of the Pleistocene Deposits of 
Maryland. Maryland Geol. Surv., December, 1906, p. 163. 

The above characters are given by Mr. F. A. Lucas as distinguishing tin's 
species from all other elephants. 


Chicago measuring respectively nine feet three inches and nine feet 
eight inches. . . . Compared with these we have the big tusk 
that used to stand on Fulton Street, New York, just an inch under 
nine feet long and weighing one hundred eighty-four pounds." 

In a footnote 1 he gives the measurement of the left tusk of an 
African elephant that is ten feet three and one-half inches in length 
along the outer curve, twenty-four and one-quarter inches in cir- 
cumference, and weighing two hundred and thirty-nine pounds. 

The longest tusk reported from Alaska is twelve feet ten inches 
in length. During the summer of 1907 the writer measured a tusk 
at Fort Gibbon that was ten feet seven inches long and the greatest 
circumference was twenty-one inches. This specimen was broken at 
both ends. 

The tusks belonging to the skull shown in plate vn are seven feet 
six inches in length. 

The tusks of the mammoth, as a rule, were more curved and of 
greater length than of the living forms, although there is a great 
variety of shapes and sizes. 

Economic Importance oE Mammoth Ivory. — It appears that 
the mammoth remains found in Alaska are not in as fresh a state 
of preservation as those found in Siberia, where for a good many 
years their tusks have constituted an important article of export. 
Dr. Middendorf, who visited Siberia about the year 1840, estimated 
the annual output of this fossil ivory to be one hundred and ten 
thousand pounds and representing at least one hundred individuals. - 
From their great abundance, Dr. R. Lydekker" has suggested that 
tusks were probably developed in both sexes. 

It is seldom, if at all, that tusks are found in Alaska sufficiently 
well preserved to compete on the market with those of the African 
and Indian elephant, as is the case with the Siberian ivory ; usually 
they are found to be discolored and either badly checked or exfoli- 
ated. A curio dealer in Nome, however, told the writer, "A few 
years ago a man would not take a tusk as a gift, but of late the best 
ones had acquired a commercial value, being cut into curios for the 
tourist trade." 

In the "curio" stores at Skagway we were shown some of the 
articles manufactured for the trade from this ivory, consisting of 

' Lucas, F. A. : Annual Report Smithsonian Institution, 1899, p. 355. 

" This estimate appears rather low, as the average tusk would hardly weigh 
two hundred and fifty pounds, or five hundred pounds for the pair, which 
would give over two hundred individuals. 

3 Lydekker, R. : Annual Report Smithsonian Institution, 1899 (pp. 361-366). 
p. 362. 


sawed sections polished for paper-weights, on which were etched 
representative scenes and animals of Alaska. The life restoration 
of the mammoth with its long hair and curved tusks appeared to be 
a favorite subject. In one instance a miniature of the mammoth 
had been carved from it. This carving and etching is done by the 
Indians and Eskimo, many of whom become quite adept at this line 
of work. Similar objects were observed in the curio stores at 
Nome. The Skagway dealers obtain most of their tusks from the 
Klondike region, while the Nome dealers procure the ivory used by 
them from the Eschscholtz Bay, Buckland. and Kobuk River local- 

In 1854 Sir John Richardson said : 

"Eskimos are in the habit of employing the soundest tusks for the 
formation of various utensils ; and the American fossil ivory has for 
at least a century, and for a longer period of unknown duration, been 
an article of traffic with the Tchutche of the opposite shores of 
Beering Straits ; so that we can venture upon no calculation of the 
multitudes of mammoths which have found graves in several icy 
cemeteries of the American coast of Beering Sea." 

Dr. W. H. Dall 1 tells of obtaining "in 1880 a deep ladle as large 
as a child's head, carved, handle and all, out of a solid tusk of mam- 
moth ivory by those people," referring here to the Eskimo. 

The writer also saw pieces of tusks fashioned into sled runners, 
having holes at intervals by which they were lashed to the wooden 
framework above. On the Yukon it was observed the Indians sonic- 
times used sections of tusks as weights for sinking their salmon nets. 

An account of this fossil ivory would not be complete without a 
mention of the blue phosphate of iron sometimes formed by the 
decomposition of the tusks and used by the Alaskan Eskimo as a 

Sir John Richardson was the first to make note, in 1854, of this 
phosphate 2 (Vivianite) occurring between the plates, of the exfoli- 
ated tusks. The writer saw this blue stain on many of the tusks 
examined by him, and it was particularly noticeable on those just 
recently removed from the ground. The same iron phosphate was 
found in the metacarpal bones of the bison collected on the Nowitna 

1 Dall, W. H. : Seventeenth Annual Report, U. S. Geol. Surv., pt. 1, p. 857. 

" In this connection it is interesting to quote from Warren's report on 
Mastodon giganteas: "On burning the bone, the ash which remains is of a 
beautiful blue color, owing to the presence of phosphate of iron, which appears 
to have been formed from the iron which had penetrated into the bone from 
the marl surrounding the skeleton." 



Elephas columbi Falconer, H., 1857, Quart. Jour. Geol. Soc. of London, 
xir, p. 319. 

The only reported occurrence of E. columbi is given by Dall, 1 who 
mentions that tusks, teeth, and bones of E. primigenins and E. co- 
lumbi were collected by Wossnessenski near Topanika Creek, Norton 
Sound. We quite agree with Maddren 2 that "the identification needs 
verification before it is assigned to Alaska." 

The American Mastodon 
Ekphas americanum, Kerr, R., 1792 Anim. Kingdom, p. 116. 

Description. — It may be readily distinguished from Elephas 
primigenius by the character of the teeth, which bear simple tent- 
like ridges (see plate vin). By its low massive build and shape of 
the skull and the tooth characters just reviewed, it may be told apart 
from the mammoth by the most casual observer. 

Remarks. — This animal also has a wide distribution. Its re- 
mains have been found from New York to Florida and west to 
Texas and Washington. It extended north into Canada, and re- 
cently two teeth have been found in the Klondike region near Daw- 
son. The writer refers here to a Mastodon molar secured by Dr. 
T. W. Tyrrell on Gold Run Creek in 1902, and through him pre- 
sented to Mr. W. H. Osgood, 3 of the U. S. Biological Survey, and 
now in the vertebrate paleontological collection of the U. S. National 
Museum (see pi. vni. fig. 1). A second occurrence of this species 
in this region was noted by the writer in the summer of 1907 — a 
tooth collected during the spring of 1906 on Sulphur Creek, near 
Dawson (see map, plate ix), and now in the possession of Mr. 
Joseph Nichlas, of that city. This specimen is reproduced here from 
a photograph (pi. vni, fig. 2). It is of interest to note the occur- 
rence of the mastodon in this region and in both places associated 
with remains of the mammoth. 

In 1904 Mr. M. T. Obalski 4 mentions the occurrence of the mas- 

1 Dall, W. IT. : Seventeenth Annual Report U. S. Geological Survey, 1896, 
p. 856. 

"Maddren. A. G. : Smith. Misc. Coll., vol. xi.ix, No. 1584, 1905, p. 7. 

3 Osgood, W. H. : Proc. Biol. Soc. of Washington, November, 190 


4 Obalski, M. T. : Les grandes Fossiles dans le Yukon et 1' Alaska. Bull d< 
la Musee d'Hist. Nat., Pari?. 1904, No. 5, pp. 214-217. 



Fig 1.— UPPER MOLAR OF MASTODON (No. 5102, U S. National Museum) FOUND ON GOLD 


AIk >ut j 5 natural size 



In the possession of Mr. Joseph Nichlas, of the city of Dawson. About y 2 natural size 




a: a ~ 

o O 


todon in the placer gravels of the Klondike region. Maddren 1 
attributes this to an error. While it may have been an error in this 
particular instance, it is likely to be a very common one, for through- 
out this entire region all of the tusks, teeth, and big bones are usually 
referred to by the people as those of the mastodon. 

So far as i:he writer knows, there have been no authentic cases 
recorded of the occurrence of mastodon in Alaska. From the fact 
that its remains do occur in the Klondike region, there appears no 
logical reason why it should not be found in Alaskan territory as 
well. It is on that account that the brief review is appended here. 

Through the kindness of Mr. R. G. McConnell, of the Canadian 
Geological Survey, the writer is enabled to present a map (see plate 
ix ) of the Klondike district on which has been indicated the local- 
ities where mastodon and mammoth remains have been found. With 
three exceptions, the localities indicated are based upon specimens 
seen by the writer. 

EQUUS sp. undet. 

Scattered remains of Equus are commonly associated with the 
other Pleistocene fossils found in Alaska. These bones have been 
considered by various authorities as representing the extinct species 
Equus fossilis and E. fratcmus, and by some referred to the living 
form E. caballus. On account of the very fragmentary nature of 
the specimens upon which these determinations have been made, in 
all cases the identifications are open to question, and until better 
material is found the species should be considered undeterminable. 
Remains of horses have been found in the following localities : 
Eschscholtz Bay, Seward Peninsula, on the Kobuk and Buckland 
Rivers ; "Palisades," on the Yukon ; Nowitna River, Old Crow 
River, and in many places in the Klondike district. 


Bison crassicornis Richardson, Zool. Voy. of H. M. S. Herald, 1852-54, 
pp. 40-60, pis. ix, xi, fig. 6; pi. xii, figs. 1-4; pi. xiii, figs. 1-2, pi. xv, 
figs. 1-4. 

Type. — Poorly preserved skull in the British Museum, from 
Eschscholtz Bay, Alaska. 

Description. — "Horns long; length of horn core along upper 
curve very much- greater than circumference at base ; horn cores 
slightly flattened on superior face ; transverse diameter much greater 
than vertical ; curve of horn regular, the tip not abruptly reflected 
nor pointing decidedly backward ; horn cores raking decidedly back- 

'Maddren. A. G. : Smithsonian Misc. Coll., vol. xux, No. 1584, 1905, p. 7. 



VOL. 51 

Remarks. — 'This species heretofore has not been known outside 
of Alaska, but, as might have been anticipated, skulls of this species 
were observed at Fox Gulch, Bonanza Creek, Yukon Territory, Can- 
ada, by the writer during the summer of 1907. In the foreground 
of plate iv, fig. 1, may be seen a portion of the skull and horn cores 

Fig. 4. — Scapula of Bison crassicornis (?) 

(Cat. No. 5941) from "Palisade?,'' on the Yukon River. (See 2, Fig. 2.) 

Greatest length, 22J /• inches. 

of this species. Remains of B. crassicornis have been collected from 
the following localities: On the tundra back of Point Barrow, Ele- 
phant Point, Eschscholtz Bay, Little Minook Creek, Little Minook 
Creek Junior, and Bonanza Creek. Yukon Territory. Canada. 

5 .2 2 


This is the largest of the extinct bisons found in the deposits of 
this region, and a scapula (see fig. 4) collected by the writer at the 
"Palisades" on the Yukon River may, on account of its size, pertain 
to this species. Plate x represents a typical skull of this form col- 
lected on Little Minook Creek and presented to the Smithsonian 
Institution through the writer by Mr. J. B. Duncan, of Rampart, 


Bison alleni Marsh, Amer. Jour, of Science, vol. xiv, 1877, p. 252. 

Type. — Horn core, No. 911, Museum of Yale College, New 
Haven, Connecticut, from Blue River, near Manhattan, Kansas. 

Description. 1 — "Horn cores long, slender, much curved, slightly 
flattened above at base; transverse diameter considerably greater 
than vertical ; length along upper curve much greater than circum- 
ference at base. Bison alleni is distinguished from B. crassicornis 
by the much greater curvature of the horn cores, these being also 
more flattened and more elliptical in section in crassicornis. 

Remarks. — This species is represented in the U. S. National. 
Museum paleontological collection, by a skull, No. 2383, from Little 
Minook Creek, near Rampart, Alaska. 2 It (see pi. xi) was found in 
the frozen muck twenty-five feet below the surface, and is of more 
than usual interest on account of the excellent state of preservation 
of the horn sheaths and from its being the first of this species found 
in. Alaska. This species is also reported as occurring in Idaho. 

A skull of B. alleni from the Porcupine River is now in the Grand 
Rapids Museum, of Grand Rapids, Michigan. 


Bison occidentalis Lucas, Science, November 11, 1898, p. 678. 

Type. — Portion of skull with horn cores, No. 4157, U. S. National 
Museum, from Fort Yukon, Alaska, collected by Sir John Richard- 

Description. — "Horn cores moderate ; circumference at base 
equal to or slightly greater than length along upper curve ; sub- 
circular in section, regularly curved upward and backward." 

1 The descriptions of the Bison from Alaska is taken from Mr. F. A. Lucas' 
article, "The Fossil Bison of North America." Proc. U. S. National Museum, 
vol. xx, 1899, pp. 755-771. 

2 This specimen was presented to the Museum by Messrs. McLain and Bal- 
lon, of Rampart, through the efforts of Gen. Timothy Wilcox, U. S. A., of 
Washington, D. C. 



Remarks. — A second specimen, No. 2643 (sec pi. xn), in the U. S. 
National Museum collections, was collected by Mr. A. G. Maddren 
on the Old Crow River in 1904. A fairly complete skeleton 1 of this 
species from Gove County. Kansas, is now in the University of Kan- 
sas Museum. This species has also been reported as occurring on 
the Tatlo River and St. Michaels. Alaska. The writer doubts very 
much the authenticity of this last locality. Mr. Lucas says: "It is 
the species most nearly resembling the existing bison, with which it 
was probably for a time contemporaneous." In that event B. crassi- 
cornis was also a contemporary, as the writer recognized skulls of 
B. occidcntalis and B. crassicornis at Fox Gulch, on Bonanza Creek, 
coming from the same layers in the deposits there. 


A skull collected at Eschscholtz Cay. Alaska, was provisionally 
referred 2 to this species by Sir John Richardson. In a more recent 
paper. 3 however, Mr. F. A. Lucas has considered this specimen (No. 
24,589. British Museum) as representing an immature individual or 
"spike horn" of B. crassicornis. 

Horn cores collected by Maddren on the Old Crow River and by 
the writer at the "Palisades" on the Yukon River appear to resemble 
the figure (see pi. xni, fig. 3) given by Richardson in his report. 


Scaphoceros tyrrelli Osgood, Smithsonian Miscellaneous Collections, vol. 

xlviii. No. 1585. 1905. pp. 173-183. pi. xxxvii, fig. 2; pi. xxxviii, fig. 

2; pi. xxxix. fig. 1 ; pi. xl, fig. 2. 
Symbos tyrrelli OSGOOD, 

Type. — Fairly complete skull, Xo. 2555, U. S. National Museum, 
from Lovett Gulch, Bonanza Creek, Klondike District, Yukon Terri- 
tory, Canada (see map, plate ix). 

Description — Generic characters.* — "Similar to Ot'ibos, but horn 
cores much smaller, less compressed at base, and more divergent at 
tips ; crown of skull between bases of horn cores surmounted by a 
prominent exostosis with an anterior bounding rim and a deep 
median excavation ; orbits much less produced laterally than in 

1 Stewart, Alban : Kansas University Quart., July, 1897, Sec. A, pp. 127-135. 
Described as B. antiquus, but referred later by Lucas to B. occidentalis. 

2 Richardson, Sir John: Zoology of Voyage of II. M. S. Herald, 
Vii, fig. I, p. 34. 

8 Lucas, F. A.: The Fossil Bison of North America. Proc. U. S. Xat. Mus.. 
vol. xxi, 1899, p. 762. 
* Generic and specific characters as given by Osgood. 

o -~ 

& + 

S o 


Ovibos; facial part of skull nearly as wide as cranial ; basioccipital 
without a high median ridge ; teeth very large and relatively broad ; 
m 1 and m 2 quadrate in transverse view." 

Specific characters. — "Size smaller than in S. cavifrons 
(Leidy) ; horn cores much smaller and shorter; exostosis less exten- 
sive, but more deeply excavated ; depth of brain case and surmount- 
ing bony mass decidedly less." 

Remarks. — The only reported occurrence of this species in Alaska 
is a horn core. No. 2378, U. S. National Museum, presented by Rev. 
J. W. Chapman through Dr. Arthur Hollick. The label with the 
horn gives the locality as Anvik, on the Yukon River, but it is un- 
likely the specimen was collected in the immediate vicinity of that 
place. It is more probable that it comes from some of the silt de- 
posits along the Yukon twenty-five or thirty miles above Anvik. 


Hay 1 cites the occurrence of Or cavifrons in Alaska, due to the 
fact that he includes Richardson's indeterminate species, Ovibos 
maximus, under this head. 

This species, therefore, is not known to occur in Alaska. 


Ovibos maximus Richardson, Zool. Voy. of H. M. S. Herald, 1852-54, 
pp. 25-28, pi. xi, figs. 2, 3, and 4. 

Type. — An imperfect cervical vertebra, the axis or dentata (No. 
9 -£, Haslar Museum), from Eschscholtz Bay, Alaska. 

Remarks. — From the very fragmentary nature of the type this 
species appears indeterminable. 


This is a recent species found at present in northern North Amer- 
ica and Greenland. At present this animal is not known to range 
west of the McKenzie River, but Pleistocene remains which have not 
been distinguished from this species are found in Alaska. As in the 
case of other remains referred to living species, more complete ma- 
terial may show an extinct species separable from the living form. 

This appears more probable since a skull, collected by the writer 
at the Palisades, on the Yukon, in 1907, is being described by Mr. 
J. W. Gidley as the type of a new species, and it may be that all the 
remains formerly considered O. moschahts should be referred to this 

1 Hay, O. P. : Bulletin No. 179, U. S. Geological Survey, p. 688. 
~ Osgood now includes Ovibos cavifrons under Symbos. 


Buckland, because of the preservation of a horn sheath on a skull 
of Ovibos submitted to him from Eschscholtz Bay, considered it of 
recent origin, but now that Bison (see pi. x) skulls are known dis- 
tinct from living species having the horn thus preserved, this argu- 
ment would apply equally to the case in question. 

OVIS, sp. undet. 

A list of species occurring in the Eschscholtz Bay deposits is 
given by Seeman in his "Narrative of the Voyage of H. M. S. 
Herald in 1853, in which Ovis montana is mentioned as being found 

This list was compiled from a report 1 made by Sir John Richard- 
son, but a careful perusal of his report failed to reveal any mention 
of fossil remains, although he does describe the recent skeleton of 
Ovis montana. 

It is probably by mistake that this species was included in 
Seeman's list, although sheep remains will undoubtedly be found, 
as Mr. W. H. Osgood, of the U. S. Biological Survey, has frag- 
mentary remains of Ovis in his possession from the Klondike district, 
Yukon Territory, Canada. At present, however, the writer does not 
know of an authentic record of their occurrence in Alaska. 

ALCE, sp. undet. 

Like Rangifer, scattered remains of the moose are known from 
several widely separated localities in Alaska and adjacent territory. 
These bones have usually been referred to as representing the living 
form Alee americanus, but it appears the identifications have been 
based upon such scanty material that the assignment to this species is 
open to question. When better specimens are known, characters of 
sufficient importance to distinguish it from the living species will 
probably be found. 

Remains of Alee are known from the deposits of Eschscholtz 
Bay, on the Old Crow and Nowitna rivers, and fragmentary antlers 
were found in the muck of Magnet and Fox gulches on Bonanza 
Creek near Dawson. 

RANGIFER, sp. undet. 

Fragmentary remains representative of this genus are commonly 
found with the bones of other Pleistocene animals in Alaska. These 
scattered and fragmentary parts have been referred by various 
writers to the living species, R. caribou and R. tarandus. It appears 

1 Zoological Voyage of H. M. S. Herald, 1852-54. 


more likely, however, if referable at all to a living' form, it would 
be R. articos, the barren-ground caribou and now living in these 

As mentioned by Richardson, Zoology of the Voyage of H. M. S. 
Herald, 1854,' p. 20, fragmentary remains have been found at Esch- 
scholtz Bay, and the writer collected fragments of antlers on Little 
Minook Creek Junior and on the Nowitna River. 

So far, remains have not been found sufficiently complete upon 
■which an accurate specific determination could Tae based. 

URSUS, sp. undet. 

The finding of a scapula and astragulus of Ursus associated with 
the remains of other Pleistocene animals on the Nowitna River 
during the summer of 1907 verifies a former record of the occurrence 
■of the bear in the Pleistocene of Alaska. 

The scapula, although incomplete, indicates an animal about the 
size of the black bear (Ursus americanus), an inhabitant of these 
regions at the present time. 

Bones 1 of Ursus have also been found associated with mammoth 
remains in a cave on St. Paul Island of the Pribilof group. 

CASTOR, sp. undet. 

Among the vertebrate remains collected on the Xowitna River in 
1907 were the left pelvic bones (No. 5942, U. S. National Museum) 
of a beaver. This appears to be the first occurrence recorded of the 
finding of bones of Castor, although Mr. E. W. Nelson, 2 who visited 
Eschscholtz Bay in 1881 with the U. S. S. Corzvin, observed a 
leaver's nest imbedded in the cliffs at that place, and noted that 
many of the sticks composing it had been gnawed and others still 
retained the tooth-marks made by that animal. 

The remains found, however, are too fragmentary to admit of 
specific determination. 


From the preceding review of the extinct vertebrates reported 
as occurring in the Pleistocene deposits of x\laska, it will be seen 
that the identification of several of the forms has been based upon 
such scanty and fragmentary material that their determination is 

1 These remains, collected by the party with Dr. D. S. Jordan in 1897, are 
-now in the paleontological collection of the U. S. National Museum. 

* Maddren, A. G. : Smithsonian Misc. Coll., vol. xiix. No. 1584, 1905, pp. 


open to question. This observation is particularly applicable to those 
so long regarded as being identical with living species. The writer 
believes that when more perfect material is available it will be found, 
probably in all instances, to be quite distinct from the living forms. 
That this is in some instances the case is shown by the discovery 
this past summer of a skull of Ovibos sufficiently complete to show 
characters of enough importance to warrant its separation from the 
living form O. moschatus, to which nearly all musk-ox material 
found in this region previously had been referred. 

More persistent collecting, aided by improved methods, will un- 
doubted]}- increase the faunal list and widen the geographical dis- 
tribution of the known forms. 

Now that Mastodon and Ovis remains have been found in Cana- 
dian territory and at a comparatively short distance from the inter- 
national boundary, there appears no logical reason why both of these 
animals should not have lived in Alaska at one time. 

While in some cases we are unable to adequately define many oi 
the species, still a very good idea of the fauna as a whole is obtained. 
Its close relationships in many instances with living animals fur- 
nishes an interesting link in the development of mammalian life of 
this continent. 

The following list, based upon material sufficiently* complete for 
fairly accurate determinations, represents the Pleistocene fauna of 
Alaska as \vc know it today : 

Elephas primigeniits Ri xmexp.ach. 

lupins, sp. undet. 

. lice, sp. undet. 

Rangifrr, sp. undet. 

Ovibos, sp. nov. 

Symbos tyrrelli Osgood. 

Bison crassicornis Richardson. 

Bison occid<cntahs Lucas. 

Bison alleni Marsh. 

Vrsus. sp. undet. 

Castor, sp. undet. 



inn 200 ||m ^00 400 






£v\£-iNc p £ 




86 9 ) 






Introduction i 

1. Hadley, 1735. Concerning the cause of the general trade 

winds 5 

II. Poisson, 1837. O n tne motion of projectiles in the air, tak- 
ing into consideration the rotation of the earth 8 

III. Tracy, 1843. On the rotary action of storms 16 

IV. Braschmann and Erman, 1859-1862. The influence of the 

diurnal rotation of the earth on constrained horizontal 

motions, either uniform or variable 23 

V. Erman, 1868. On the steady motions or the average con- 
dition of the earth's atmosphere 31 

VI. Kerber, 1881. The limit of the atmosphere of the earth 43 

VII. Sprung, 1 88 1. On the paths of particles moving freely on the 
rotating surface of the earth and their significance in 

meteorology 57 

VIII. Pockels, 1901. The theory of the formation of precipitation 

on mountain slopes 80 

IX. Gorodensky, 1904. Researches relative to the influence of 
the diurnal rotation of the earth on atmospheric disturb- 
ances 105 

X. Gold, 1908. The relation between wind velocity at 1000 

meters altitude and the surface pressure distribution .... 113 
XL Guldberg and Mohn, 1876-1883. Studies on the movements 

of the atmosphere 122 

XII. von Bezold, 1892-1906. On the thermodynamics of the atmos- 
phere : fourth memoir 249 

XIII. von Bezold, 1900-1906. On the thermodynamics of the 

atmosphere: fifth communication 280 

XIV. von Bezold, 1900-1906. Theoretical considerations relative 

to the results of the scientific balloon ascensions of the 
German Association at Berlin for the promotion of 
aeronautics 285 

XV. von Bezold, 1884-1906. On the reduction of the humidity 

data obtained in balloon ascensions 326 

XVI. von Bezold, 1898-1906. On the changes of temperature in 

ascending and descending currents of air 334 

XVII. von Bezold, 1890-1906. On the theory of cyclones 344 

XVIII. von Bezold, 1901-1906. On the representation of the distri- 
bution of atmospheric pressure by surfaces of equal pres- 
sure and by isobars 366 

XIX. von Bezold, 1892-1906. The interchange of heat at the 

surface of the earth and in the atmosphere 376 

XX. von Bezold, 1901-1906. On climatological averages for com- 
plete small circles of latitude 416 


XXI. Neuhoff, 1900. Adiabatic changes of condition of moist air 
and their determination by numerical and graphical 

methods 430 

XXII. Bauer, 1908. The relation between "potential temperature" 

and "entropy" 495 

XXIII. Margules, 1901. The mechanical equivalent of any given dis- 

tribution of atmospheric pressure and the mainte- 
nance of a given difference in pressure 501 

XXIV. Margules, 1904. On the energy of storms 533 

XXV. Pockels, 1893. The theory of the movement of the air in 

stationary anticyclones with concentric circular isobars. 596 




In order to introduce English-speaking students of meteorology 
to the rapidly increasing literature bearing on the fundamental 
mechanical problems of that science, I have been encouraged to 
publish numerous translations either in the "Monthly Weather 
Review" of the U. S. Weather Bureau or in the technical journals. 
Others are collected in the "Short Memoirs on Meteorological Sub- 
jects," Smithsonian Report for 1877, pp. 376-478, and in "The 
Mechanics of the Earth's Atmosphere," Smithsonian Miscella- 
neous Collections, 1891. As our knowledge of the subject pro- 
gresses and we perceive new difficulties arising, so also we learn 
to conquer those older ones that were the ultima thule of the 
past generation. Step by step man is penetrating the complex 
maze of forces that push our atmosphere hither and thither. Its 
internal mechanism is so complex that superficial students content 
themselves with empirical rules or search for cosmical relations 
of minor importance: the very ablest investigators have as yet 
solved only the simpler problems relating to idealized conditions 
that rarely occur in nature. 

In this third collection of translations bearing on the mechan- 
ics of the earth's atmosphere I have ventured to begin with that 
elementary but classic memoir by Hadley which gave occasion to 
the Berlin Academy in 1746 to offer a prize for a mathematical 
discussion of the motions of the atmosphere. The prize was 
awarded to d'Alembert; subsequently Musschenbroek, deLuc, Euler, 
Bernoulli, Lambert, von Lindenau (in 1806) and Brandes (in 1822) 
successively contributed to the elucidation of this subject. But 
it was Poisson who, in 1837, first deduced correctly the influence 
of the earth's rotation on moving solids, and Tracy who in 1843, 
applied similar views to the rotation of storms. Poisson's works 
and ideas were generally known to the scholars of France as shown 
by the prolonged discussion, 1850-1860, of the Foucault pendulum 


and gyroscope phenomena. There are indications that Babinet 
and others at that time applied his results to the mechanics of 
the atmosphere. But the modern study of this subject is properly 
traceable to the influence of Prof. William Ferrel in America and 
Prof. William Thomson in England, both of whom cooperated to 
put our knowledge of the subject on a firmer basis than was 
before possible. Meanwhile a profound Russian scholar, Brasch- 
mann, and the equally profound German scholar, Erman, were 
independently working over the same ground, though their publi- 
cations have been scarcely noticed by technical meteorologists. 
The neglect of Erman 's work in dynamic meteorology seems 
remarkable, but has been atoned for by the enthusiastic activity 
of Sprung and his successors at Hamburg and at Berlin. 

The works of Espy, 1840, on the Philosophy of Storms; Thom- 
son, 1862, on the Convectional Equilibrium in the Atmosphere; 
Peslin, 1868, on the Thermodynamics of Moist Air; Ferrel, 1857, on 
The Motions of Solids and Fluids, and his subsequent important 
memoirs; together with Sprung's Lehrbuch, 1885, mark the tran- 
sition from ancient to modern meteorology. 

The modern sounding balloon has assured us of the intimate 
connection between the lowest stratum of air and that which is 
20 miles above us; but the conditions above this latter level are 
doubtless of equally great importance to our surface climatology 
and these can be made known to us only by the study of meteors, 
auroras, spectrum lines, and refractions. I have, therefore, included 
a memoir by Kerber on the limit of the earth's atmosphere, that 
avoids some of the difficulties attending every application to the 
outer atmosphere of our knowledge of the kinetic theory of gases. 

All students will gladly welcome the translation by Waldo of 
the memoir by Guldberg and Mohn, first published in two parts, 
1876 and 1880; it was revised by the authors in 1883 at the per- 
sonal request of Prof. Frank Waldo who expected its prompt 
publication, and to him we owe the privilege of including in the 
present collection this new edition of that classic paper. 

The series of papers by Von Bezold were revised by himself in 
1906, for publication in his collected memoirs and as thus revised 
they are now reproduced by permission of his heirs and publishers. 

The study of strictly adiabatic changes that was so greatly 
facilitated by the Hertzian diagram published in the preceding 
collection of translations is now advantageously replaced by the dia- 
grams of NeuhorT, 1900, which adapt themselves to any atmospher- 
ical condition. 


In conclusion, the two memoirs* by Margules (xxiii, 1901, and 
xxiv, 1904) introduce us to the great problems of the future, 
that is, the thermal transformations of energy persistently going 
on in the atmosphere. Margules has been the first to find methods 
of studying and solving these problems. It only remains for 
future students to combine the equations of thermodynamics 
with those of hydrodynamics so as to further elucidate the details 
of the phenomena as to time and place — a result that Ave may 
hope will eventually be attained by the analysis of fields of force 
that is now being perfected by Bjerknes of Christiania. 

Cleveland Abbe. 
Washington, D. C. 

November, igoS. 

' I 



[Phil. Trans. Vol. XXXIX, London, 1735-36, p. 5S] 

I think the causes of the General Trade Winds have not been 
fully explained by any of those who have wrote on that subject, 
for want of more particularly and distinctly considering the share 
the diurnal motion of the earth has in the production of them. 
For although this has been mentioned by some amongst the causes 
of those winds, yet they have not proceeded to show how it 
contributes to their production; or else have applied it to the 
explication of these phenomena, upon such principles as will appear 
upon examination not to be sufficient. 

That the action of the sun is the original cause of these winds, 
I think all are agreed ; and that it does it by causing a greater 
rarefaction of the air in those parts upon which its rays falling 
perpendicularly, or nearly so, produce a greater degree of heat 
there than in other places ; by which means the air there becoming 
specifically lighter than the rest round about, the cooler air will 
by its greater density and gravity, remove it out of its place to 
succeed into it itself, and make it rise upward. But it seems, 
this rarefaction will have no other effect than to cause air to rush 
in from all parts into the part where 'tis most rarefied, especially 
from the north and south, where the air is coolest, and not more 
from the east than the west, as is commonly supposed: so that, 
setting aside the diurnal motion of the earth, the tendency of 
the air would be from every side towards that part where the 
sun's action is most intense at the time, and so a NW. wind be 
produced in the morning, and a NE. in the afternoon, by turns, 
on this side of the parallel of the sun's declination, and a SW. 
and SE. on the other. 

That the perpetual motion of the air towards the west, cannot 
be derived merely from the action of the sun upon it, appears 
more evidently from this: If the earth be supposed at rest, that 



motion of the air will be communicated to the superficial parts, 
and by little and little produce a revolution of the whole the same 
way, except there be the same quantity of motion given the air 
in a contrary direction in other parts at the same time, which 
is hard to suppose. But if the globe of the earth had before a 
revolution towards the east, this by the same means must be con- 
tinually retarded. And if this motion of the air be supposed to 
arise from any action of the parts of it on one another, the con- 
sequence will be the same. For this reason it seems necessary to 
show how these phenomena of the Trade Winds may be caused, 
without the production of any real general motion of the air 
westwards. This will readily be done by taking in the considera- 
tion of the diurnal motion of the earth. For, let us suppose the 
air in every part to keep an equal pace with the earth in its 
diurnal motion; in which case there will be no relative motion of 
the surface of the earth and air, and consequently no wind, then 
by the action of the sun on the parts about the equator, and the 
rarefaction of the air proceeding therefrom, let the air be drawn 
down thither from the N. and S. parts. The parallels are each. 
of them bigger than the other, as they approach to the equator 
and the equator is bigger than the tropics, nearly in the propor- 
tion of 1000 to 917, and consequently their difference in circuit 
about 2083 miles, and the surface of the earth at the equator 
moves so much faster than the surface of the earth with its air 
at the tropics. From which it follows, that the air, as it moves 
from the tropics towards the equator, having a less velocity than 
the parts of the earth it arrives at, will have a relative motion 
contrary to that of the diurnal motion of the earth in those parts, 
which being combined with the motion towards the equator, a 
NE. wind will be produced on this side of the equator and a vSE. 
on the other. These, as the air comes nearer to the equator, will 
become stronger, and more easterly, and be due east at the equator 
itself, according to experience, by reason of the concourse of 
both currents from the N. and S. where its velocity will be at the 
rate of 2083 miles in the space of one revolution of the earth or 
natural day, and above one mile and one-third in a minute of 
time; which is greater than the velocity of the wind is supposed 
to be in the greatest storm, which according to Doctor Derham's 
observations, is not above one mile in a minute. But it is to be 
considered, that before the air from the tropics can arrive at the 
equator, it must have gained some motion eastward from the 
surface of the earth or sea, wherebv its relative motion will be 


diminished, and in several successive circulations, may be supposed 
to be reduced to the strength it is found to be of. 

Thus I think the NE. winds on this side of the equator, and 
the SE. on the other side, are fully accounted for. The same 
principle as iiecessarily extends to the production of the west 
trade-winds without the tropics; the air rarefied by the heat of 
the sun about the equatorial parts, being removed to make room 
for the air from the cooler parts, must rise upwards from the earth, 
and as it is a fluid, will then spread itself abroad over the other 
air, and so its motion in the upper regions must be to the N. and 
S. from the equator. Being got up at a distance from the surface 
of the earth, it will soon lose great part of its heat, and thereby 
acquire density and gravity sufficient to make it approach its 
surface again, which may be supposed to be by that time 'tis 
arrived at those parts beyond the tropics where the westerly 
winds are found. Being supposed at first to have the velocity 
of the surface of the earth at the equator, it will have a greater 
velocity than the parts it now arrives at; and thereby become 
a westerly wind, with strength proportionable to the difference of 
velocity, which in several revolutions will be reduced to a certain 
degree, as is said before, of the easterly winds, at the equator. 
And thus the air will continue to circulate, and gain and lose 
velocity by turns from the surface of the earth or sea, as it ap- 
proaches to or recedes from the equator. I do not think it neces- 
sary to apply these principles to solve the phenomena of the 
variations of these winds at different times of the year, and differ- 
ent parts of the earth ; and to do it would draw this paper into 
greater length than I propose. 

From whatever has been said it follows : 

First, That without the assistance of the diurnal motion of the 
earth, navigation, especially easterly and westerly, would be very 
tedious, and to make the whole circuit of the earth perhaps imprac- 

Secondly, That the NE. and SE. winds within the tropics must 
be compensated by as much NW. and SW. in other parts, and 
generally all winds from any one quarter must be compensated 
by a contrary wind somewhere or other; otherwise some change 
must be produced in the motion of the earth round its axis. 





BY M. [S. D.] POISSON 1 

In this memoir the projectile will be considered as an isolated 
and material point, that is to say, as a body whose mass is col- 
lected at the center of gravity, and the problem will be to ascer- 
tain the influence of the rotation of the earth on its motion. I 
shall present shortly another memoir to the Academy, in which 
we shall take into consideration the form and the dimensions of 
the moving body, and the object of that will be to determine, 
principally in what relates to the projectiles used in artillery, the 
influence that their own rotation can produce on their motion of 

Up to the present time the theory of the resistance which fluids 
in general, and the air in particular, offer to the motion of the 
bodies that traverse them, has received only a very imperfect 
development. We compare this force to a continual succession 
of shocks of the moving body against the particles of the fluid, 
which disappear and are annihilated, so to speak, when they have 
been struck by the body and have carried away small quantities 
of motion, proportional to their own masses and its velocity. 
Newton, to whom we owe this theory, had concluded that, ignor- 
ing the rotation of the moving body, the resistance of the air for 
a sphere, for example, is equal to the weight of a cylinder of this 
fluid having for its base the great circle of the sphere and for 
height the "full-height" due to its velocity. But the experiments 
made on the fall of bodies in the air soon showed him the inac- 
curacy of this result, and led him to reduce by one-half this 
measure of resistance; has been found that this 

1 Memoire sur le mouvement des projectiles dans 1'air en ayant egard 
la rotations de la terre. By [S. D.] Poisson. Read before the Academy 
of Sciences, Paris, November, 1837. Published in the Journal de l'Ecole 
Royale Polytechnique, Vol. XVI, Cahier 26, Paris 1838, pp. 1-68. Trans 
jated by Profs. Frank Waldo and Cleveland Abbe. 



reduction is too great, and Borda has concluded from his own 
observations that the measure of the resistance must be only 
diminished to three-fifths of its theoretical value. From the theory 
of Newton as modified by experiment, the retarding force relative 
to the unit of mass for a sphere moving through the air has for 
its expression the square of the velocity of the sphere divided by 
its diameter and by the ratio of its density to that of the fluid, 
and multiplied by a numerical coefficient concerning which the 
writers on ballistics do not agree. According to Lombard, 2 and 
relying on the experiments of Borda, this coefficient should be 
equal to about nine-fortieths. But the true law of the resistance 
as a function of the velocity is far more complex; for motions 
which are either very rapid or very slow the coefficient seems 
to deviate considerably from being proportional to the square of 
the velocity; in the case of very great velocities it increases at 
a much greater ratio, and on the contrary when it is a question 
of small velocities, such as the very small vibrations of the seconds 
pendulum 3 this coefficient is proportional to the simple velocity. 
In order to determine directly and without any hypothesis the 
law of the resistance that a body meets with in moving through 
a fluid , it will be necessary to consider at the same time both the 
motion of the body and that which the moving body communi- 
cates to the fluid; as the result of this double motion the fluid 
exerts at each instant a certain pressure at each point of the 
moving body and normal to its surface; this pressure is different 
from that which occurs in the state of rest and produces the resist- 
ance, properly so-called, that the moving body experiences, and 
to which it will be necessary also to add the force tangential to 
the surface of the body arising from the friction of this body against 
the layer of fluid in contact with it. In fact, this is what I have 
been able to do in my Memoir on the simultaneous Motions of 
the Pendulum and of the surrounding Air, 4 and which has led 
me to deduce from theory the new correction which M. Bessel 
has confirmed by experiment on the length of the seconds pendu- 
lum. Hereafter I shall try to extend that analysis to the case of 
the progressive motion of projectiles in the air and to determine, 
if it is possible for me to do so, the pressure that the fluid dis- 
placed by them exerts on their surfaces by its compression on one 
side and expansion on the other, or the resistance that they 

2 Treatise on the motion of projectiles, p. 99. 

3 Additions to the Connaissance des Temps for the year 1834, p. 18. 

4 Memoirs of the Academy of Sciences, Vol. XI [Paris]. 


meet with considered from the point of view that I shall indicate. 
I do not need to say that the exact and general knowledge of 
this law will be important in many questions, for example, in the 
problem of ballistics. But for the object which I have in view in 
this present memoir I can admit the ordinary law of the resistance 
proportional to the square of the velocity as being sufficiently 

It is Newton, also, who has given the first example of the deter- 
mination of the motion of a heavy body in a resisting medium. 
He solved the problem when the motion is vertical by assuming 
the resistance proportional either to the velocity or to its square, 
but when the projectile is projected into the atmosphere in any 
direction whatever he confined himself to considering the case 
of a resisting force proportional to the simple velocity, observing 
nevertheless that this case is not that of nature. The two equa- 
tions that Newton was obliged to integrate in order to determine 
the horizontal and vertical components of the velocity at any 
instant, are linear of the first order and with constant coefficients; 
and the two unknown quantities are so separated in them that 
these two equations are solved independently of each other, and 
their solution really implies only a simple direct integration. This 
is no longer true in the case of a resistance proportional to the 
square of the velocity; the two unknown quantities enter at the 
same time into each of the equations of motion, which are no 
longer linear, and it is only by a special combination that we 
succeed in separating the variables therein and in reducing them 
to quadratures, which we consider as the complete solution of 
the problem. 

This was done by John Bernoulli, who published it in the Acta 
Eruditorum, Leipzig, May 1719, pp. 216-226, more than thirty 
years after the solution by Newton, and at an epoch when the 
integral calculus had already made great progress. However, Euler, 
at the beginning of his memoir on this subject, 5 expresses his sur- 
prise at seeing that Newton, "who has well solved other problems 
more difficult," should stop with the case of the resistance pro- 
portional to the simple velocity, and not consider the case of 
nature. We know, however, that the question of the trajectory 
in a medium resisting in proportion to the square of the velocity 
was proposed as a challenge to the geometers of the continent 
by an Englishman named Keil, who believed the problem insol- 

1 Memoirs of the Academy of Berlin; year 1753. 

motion op projectiles — poisson 11 

uble because his illustrious countrymen had not solved it. Now 
the numerical calculation of the integrals which express the time 
and the two coordinates of the moving body, in functions of a 
fourth variable, is effected as simply as the question allows, and 
enables the approximations to be carried as far as we wish. We 
can see an example in the "Exercises du calcul Integral" of Legen- 
dre 6 in which these coordinates are calculated to within less than 
a hundred-thousandth part of their values. 

Independently of the centrifugal force arising from the rotation 
of the earth (which influences the motions of heavy bodies by 
diminishing the force of gravity by a quantity that varies with 
the latitude), this rotation also produces in these motions certain 
deviations that it is interesting to understand, either in themselves 
or in order to know to what extent they can influence the trajec- 
tory of the projectiles, and whether it is necessary to consider them 
in the practice of artillery. 

Many physicists have measured, with as much precision as has 
been possible, the small distances by which bodies that fall from 
a considerable height deviate from the foot of the vertical. La- 
place and Gauss submitted this question to the calculus, but in 
integrating the equations of this almost exactly vertical motion 
they have left out of consideration the resistance of the air, which 
can, however, sometimes have a very great influence on the result. 
1 have therefore thought it would be useful to go over this problem 
entirely and to extend the solution to the general case in which 
the projectile is projected into the atmosphere with any velocity 
and in any direction whatever. 

To this end I have in the first place formed the differential 
equations of the absolute motion in space by referring the coordi- 
nates of the moving body to fixed axes; then I have deduced from 
these the equations of apparent motion such as we observe near 
the surface of the earth, referred to fixed axes at the surface which 
participates as well as we ourselves in the rotation of the earth. 
These differential equations are very complicated, but by taking 
the second of time for the unit of time, the angular velocity of 
the diurnal motion becomes a very small fraction, which permits 
(us) to reduce them to a more simple form. From these we deduce 
some general consequences, enumerated as follows: 

(i) The motion of the earth prevents a liquid contained in a 
vase and turning with a constant velocity about a vertical axis 

8 Vol. I, p. 33 6. 


from assuming the rigorously permanent figure of a paraboloid of 
revolution as it would do if the earth were immovable. 

(2) If the body moves along a given curve that is attached 
firmly to the surface of the earth, the differential equation of its 
motion does not contain the velocity of the rotation of the earth 
and consequently this motion is the same as if the earth were at 
rest. Thus, for any given value of gravity resulting from the 
figure and the rotation of the terrestrial spheroid, the oscillations 
of the pendulum are the same in all azimuths around the vertical; 
a result that was important to demonstrate, considering the 
degree of precision that we now attain in the determination of the 
length of the seconds pendulum at different places on the earth. 
But the diurnal rotation and the direction of the plane of oscillation 
have a slight influence on the variable tension that the wire expe- 
riences during the oscillations and which is not rigorously the 
same in all azimuths. 

(3) Finally, when a projectile is sent into the air in any direction 
whatever the rotation of the earth neither increases nor dimin- 
ishes the distance that it attains at any instant from a plane 
through the point of departure and parallel to the equator. 

Before seeking the integrals of the equations of apparent motion 
in the general case of an initial velocity having any magnitude and 
any direction whatever, I have considered the simpler special cases. 

The first case is that where the moving body starts from a point 
situated at a given height above the ground without imparting 
to it any initial velocity whatever and is left to the action of 
gravity, so that it commences to fall vertically. The velocity [of 
the eastward motion] at the point of departure, due to the rota- 
tion of the earth in which it participates, being greater than that 
which belongs to the foot of the vertical, we perceive that the 
moving body when it has reached the earth must have departed 
from the foot of the vertical line, to the eastward or in the direc- 
tion of the true motion of the earth, but mathematics alone can 
give the measure of this distance, especially when we consider the 
resistance of the air; one can see that the deviation takes place 
toward the east and that it is nothing in the direction of the 
meridian. In order to compare with experience the formula which 
expresses the amount of deviation, I have chosen the observations 
of this phenomenon which were made in 1833 by Professor Reich 
in the mines of Saxony. The height of the fall was 158.5 meters 
and M. Reich concluded for the mean of 106 experiments 
that there was a deviation to the east of 28. 33 mm . He also 


found very nearly six seconds for the duration of the fall. 
By means of this latter datum I have been able to calculate with- 
out any hypothesis the coefficient of resistance of the air which 
the moving body must have experienced, and the formula gives 
27.5 mm for the deviation; which 'differs from the experiments 
by less than a millimeter. In a vacuum this deviation would 
not have exceeded by a tenth of a millimeter that which 
occurred in the air; so that in this case the resistance of the air 
has had only an inappreciable influence. 

When the projectile starts from the surface of the earth and is 
thrown vertically from below upwards with a given velocity, we 
conceive that during the time of its ascent it must be departing 
from the vertical toward the west, or in a direction contrary to 
the rotation of the earth. It would seem that afterwards during 
its fall it should approach this line and return again very nearly 
to its point of departure, but this is in fact not the case. When 
it has arrived at the highest point of its trajectory and has lost all 
its vertical velocity, the projectile by deviating towards the west 
has also acquired a horizontal velocity in the same direction, by 
virtue of which it continues to deviate in this direction, at least 
during part of its fall. The analytical difficulty which this second 
case presents is to reconcile, so to speak, the two successive motions, 
ascending and descending, of the projectile, which are expressed 
by very different formulae when we take account of the resistance 
of the air. In order to apply to an example the formula expressing 
the total deviation of the moving body when it has fallen back to 
the earth, I have assumed that this body is a spherical ball fired 
vertically from an infantry gun, with a velocity of about 400 
meters per second. The amount of this deviation varies much 
with the resistance of the air; by giving successively to the coeffi- 
cient of this resistance different values which have to each other 
the ratio of four to three, we find deviation toward the west in 
both cases but of about one and three decimeters respectively. 
In a vacuum this deviation would be about fifty-five meters, so 
that by the greater of these two resistances it is reduced to the 
fifteenth part of this value. 

I have also examined in particular the case where the initial 
velocity of the projectile is nearly horizontal, which corresponds 
to firing at a target. In my present memoir will be found the for- 
mulas that relate to this and which express all the rircumstances 
according as the firing is directed toward any given point of the 
horizon. Here I shall only stop to say that the initial velocity 


being always about 400 meters and the distance of the target, 
placed point blank, being equal to 200 meters, then the horizontal 
and vertical deviations of the ball, due to the motion of the earth, 
would amount to scarcely half a centimeter, that is to say, they 
have no sensible influence on the precision of this shooting and 
it is unnecessary to consider them in practice. These deviations 
are equally unimportant in firing a cannon, and in all motions 
which take place in a nearly horizontal direction. 

In general the effects that the motion of the earth produces on 
the motion of a projectile are: first, to increase, either positively 
or negatively, the interval of time that the moving body takes to 
go from its point of departure to the point where it falls on the 
earth; second, to increase the distance of this latter point from the 
former, which we call the horizontal range. The signs of these 
increments depend on the direction of the vertical plane in which 
the projectile is thrown; there is augmentation in one direction 
and diminution in another; their values are expressed by double 
integrals, whose numerical calculation would be very laborious. 

In addition to this the diurnal motion causes the moving body 
to leave the vertical plane in which it was initially projected. 
This gives place to a horizontal deviation, whose value is composed 
of two distinct parts, expressed also by double integrals. One 
of these partial deviations is independent of the direction of the 
vertical plane; it is always toward the right of an observer stationed 
at the point of departure and facing the trajectory. In our lati- 
tude we can consider it as being the principal effect of the rotation 
of the globe, and happily we can obtain for it limiting values that 
are easier to calculate than its own value, and which may, if we 
wish, be deduced numerically by means of the length of the range 
and the duration of the movement as given by observation, with 
an accuracy sufficient to appreciate the amount of the deviation. 
Applying, for example, these limits to such firing of shells as 
takes place in actual artillery practice, that is to say, at an angle 
of elevation of 45 , with an initial velocity of 120 meters per second, 
which gives a range of about 1200 meters, for a projectile of 27 
centimeters in diameter, and 51 kilograms in weight (the shell 
of 10 inches and 104 pounds old French measure); we find that 
the deviation of the point of impact will be between 90 and 120 
centimeters when we aim in a vertical plane, tangent to the parallel 
of latitude at the point of departure. The deviation will be toward 
the south if we fire toward the east, and toward the north if we 
fire toward the west. Calling it a meter and observing that such 


a deviation in a distance of 1200 meters corresponds to an angle 
of about three minutes of arc, it follows that in order to be more 
sure of hitting the mark it will be necessary to aim in a vertical 
plane to the left of the given plane, and making with that an angle 
of three minutes. The consideration of this result may influence 
the accuracy of the aim and the chance of striking the target in 
exercises where the gunner must seek great precision. The hori- 
zontal deviation Avill be a little less and will be toward the east 
when we fire toward the north; it will be a little more and to ward 
the west when we fire toward the south. In the firing of a shell 
at long range, for example, at a distance of about 4,000 meters 
from the mark, which supposes an initial velocity of a little more 
than one-third of 800 meters, at the elevation angle of 45 , and 
for a projectile weighing 90 kilograms and a third of a meter 
in diameter, the limits of deviation, firing either to the east or to 
the west, will be very nearly 5 meters and 10 meters, respectively. 
Estimating then its average amount at 7 or 8 meters, we see that 
in sieges some buildings and persons have been reached because of 
the deviation of a shell by the motion of the earth and others 
have not been from the same reason. 

These numbers, and those that we have before given, relate to 
a mean latitude; they vary with the latitude of the place of the 
experiment. At the equator when the firing takes place in its 
plane, the horizontal deviation vanishes while the increase in the 
duration of the trajectory and in the length of the range attain 
their maximum values. In high latitudes, on the contrary, it is 
the deviation which approaches its maximum and the increase 
of duration which diminishes. At the pole, the horizontal devi- 
ation, which is the same at this point for all vertical planes of 
firing, would exceed by very nearly one-half that which takes 
place in our latitude. Everywhere the increments of the range 
and of the time are nothing when the initial velocity is directed 
111 the plane of the meridian. 

[The preceding text is followed by a detailed mathematical 
analysis that need not be reproduced here]. 




The investigations of Mr. Redfield and Colonel Reid have accumu- 
lated a vast amount of evidence in favor of the propositions they 
maintain. The tendency of this evidence is to demonstrate, 
that in the large storms which affect extensive districts, and also 
in the violent tornadoes which devastate a brief path, there are 
two motions, the rotary and the progressive; and that the rotary 
is by far the most violent, and has an uniform direction of revolu- 
tion, being from right to left if the storm is in the northern hemi- 
sphere, and the reverse if it is in the southern hemisphere. That 
is to say, on our side of the equator the rotation is about the cen- 
ter through the points of the compass, in the order of N. W. S. E., 
or contrary to the movement of the hands of a watch lying on 
its back; and south of the equator the rotation is through the 
points in the order of N. E. S. W., or conformable to that of the 
hands of a watch. 

These propositions, although authorized by induction, have en- 
countered doubts or gained a feeble faith in many minds, for the 
want of a good cause to assign for the production of the alleged 
phenomena. Hence the occurrence of rotary storms, and the uni- 
formity of direction of revolution, have been too readily attributed 
to mere accident; and the notion that a whirlwind, once started 
by mere chance, contains the elements of growth and stability of 
motion, has been too easily admitted. An active whirlwind, 
great or small, undergoes a constant change of substance. As the 
central portions waste into the ascending column, supplies from 
the adjacent tranquil air must be drawn into the vortex and set 
in motion; and if the fresh air is neutral to the circular movement 
and must acquire velocity from the whirling mass itself, then 
since "action and reaction are equal and in opposite directions," 

1 Reprinted from the American Journal of Science and Arts, Vol. XLV, 
October, 1843, pp. 65-72. Read before the Utica Natural History Society. 
(Dated Utica, N. Y., February 27, 1843.) 



the whirling mass itself must lose just so much velocity as the 
fresh supply gains. By such a process the forces of the whirl- 
wind would be rapidly exhausted, and its existence must speedily 
cease. A stable source of momentum, adapted to originate and 
sustain the uniform rotary movement, is still required; and it is 
now proposed to develop such a source of momentum in the 
forces generated by the earth's diurnal revolution. 

The velocity of the earth's surface in the daily revolution being 
at the equator more than one thousand miles an hour, in latitude 
6o° half as much, at the pole nothing, and varying in intermediate 
places as their perpendicular distances from the earth's axis, and 
the atmosphere near the ground everywhere taking in part or 
wholly the motion of the surface it rests on, important conse- 
quences upon aerial currents must follow. A body of air set in 
motion from the equator northward maintains the equatorial east- 
ward velocity, and when it passes over regions of slower rotation 
deviates eastward from the meridian, and ultimately describes 
over the earth's surface a curved line bearing towards the east. 
A current of air from latitude 45 north, having a due south 
direction, soon reaches regions moving faster to the east, falls 
behind them and describes a curve to the west. Winds oblique 
to the meridian are similarly affected. These familiar matters are 
referred to here, and illustrated by fig. 1, to elucidate what follows. 




FIG. I. 

The influence of the figure and revolution of the earth upon 
east and west winds, must also be considered. A parallel of lati- 
tude, being a lesser circle of the globe, and at all points equally 
distant from the pole, necessarily describes upon the earth's surface 
a curved line. But a direct course, due east at the commencement 
follows a great circle and parting from the parallel reaches a lower 
latitude. The due east course continued in a right line describes 
a tangent to the curve of the latitude. The velocity of the earth's 
surface at any place, by virtue of the diurnal revolution, has for 
its direction the line of that tangent; and when the air reposing 
over any spot is transferred to a region of diverse motion, the 


direction, as well as the degree, of its previous force is to be taken 
from that of the soil on which it previously rested. Hence a wind 
from the west, if in our hemisphere, will soon be found pursuing 
a southeasterly course, and crossing successive parallels of latitude. 

The labors of Mr. Espy have been directed to the hypothesis 
of a central ascending column of rarefied air, and centripetal 
currents from every side rushing towards its base. Without pur- 
suing his reasoning, it will be safe to assume that his collection 
of facts established the existence of a qualified central tendency 
of the air, in both the general storms and the smaller tornadoes. 
He presents a theory to account for such motion, which it is not 
necessary now to examine. Dr. Hare has proposed another method 
of accounting for tornadoes — a truly brilliant suggestion — of which 
it is only to be remarked, at present, that it proceeds on the assump- 
tion of a rush of air from all quarters to a central point. It has 
been attested also, that at large clearing fires in calm weather, 
creating centripetal currents, the whirlwind and mimic tornado 
have been produced. In accounting for the whirlwind motion, 
therefore, the central tendency of the air will be presupposed. 

In the case of a large fire kindled in an open plain on a calm 
day, a small circle about the fire is first acted on by the abate- 
ment of pressure on the side next the fire, and thus receives an 
impulse toward the common center. As this moves in, the next 
titer circle loses support and begins to move. Each particle of 
is moved at first by an impulse towards the center, and during 
approach to the central region it receives fresh impulses of the 

me direction; and if it comes from some distance its velocity 
is in this way accelerated, until it reaches the space where the hori- 
zontal is broken by the upward motion. It is obvious that par- 
ticles propelled by such impulses would seek the common center 
in the lines of its radii, and their horizontal forces would be neu- 
tralized by impact, if no cause for deviation was at hand. But 
the great law of deflection which affects the course of the winds 
applies to the movements of these particles. The particles which 
seek the center from the northern points are deflected west, while 
those from southern points are deflected east. The whole rush of 
air from the northern side of the center, coming like a breeze 
bears west of the center, while an equal breeze from the southern 
side bears east of the center. The consequence is that the central 
body of air, including the fire, is acted upon by two forces which 
combine to make it turn round to the left. These forces are aided 
by the deviation of the currents from the easterly and westerly 



parts of the circle. The breeze from the west extreme inclines to 
the tangent of the parallel of latitude at its original place of repose, 
and therefore strikes south of the center into which the impulse 
it receives would otherwise carry it. The air from the east side 
also inclines toward the tangent of the parallel of latitude there, 
which is, obliquely to the north from the radius, and therefore is 
deflected northwards and strikes north of the center. The breezes 
from all quarters thus cooperate to produce the result; and all 
their forces are constant; and act with precision and at great advan- 
tage to cause and maintain a whirlwind. A diagram presenting 
the lines of approach of the particles or streams of air, will explain 
this result. The black lines in fig. 2 show the deviating currents, 
from the cardinal points alone, when the area affected by the fire 
is so small as to require no perceptible curve in those lines. 

Upon the same principle, the tornado, the typhoon, and the 
widespread storm of the Atlantic, if their currents move toward 
a central spot, must have a rotary character. The circular motion 
in the outer portions may be slight, but it is stronger near the 
center. In every such case the incoming air may be regarded as 
a succession of rings taken off the surrounding atmosphere and 
moving slowly at first, but swifter as they proceed towards the 
center. Each such ring is affected by the law of deviation during 
its passage. The particles are veering from the radii, in its northern 
quarter westward, in its southern quarter eastward, in its eastern 
quarter northward, and in its western quarter southward, and 
hence the ring begins to revolve when far from the center, turns 


more and more as it draws near it, and finally as it gathers about 
the central spot all its forces are resolved into a simple whirl. 
Ring after ring succeeds, and the whirling action is permanent. 
The deflecting power thus applied . is not small. The rotary- 
motion of the earth varies as the cosine of the latitude, and the 
differences of velocity for any differences of latitude are easily 
computed. The following are samples; being differences of velocity 
for i° or 6o£ miles of latitude. 

per hour. 

Between lat. 2° and 3 diff. of velocity 0.79 

Between lat. 3 and 4 diff. of velocity 1.11 

Between lat. io° and n° diff. of velocity 3.31 

Between lat. 23 and 24 diff. of velocity 7.25 

Between lat. 42 and 43 diff. of velocity 12.28 

The differences of velocity for one mile, or 51.84" of latitude 
are as follows: 

Difference of velocity 
for 1 mile north. 
Latitude. Feet per minute. 

io° 4 

23° 9 

42° 15-4 

43° *5-7 

45° 16.3 

The deflection of easterly and westerly breezes by reason of 
the spherical form of the earth, also, can be computed; and it is 
obviously no less important than the deflection produced in merid- 
ional winds. The angle between the courses north and east, at 
any point, is a right angle; and if two points in the same latitude 
are taken, it is evident that the obliquity of the north courses 
from the two points equals the obliquity of the east courses from 
the same points. 

These results show that in the northern states a fire large enough 
to affect the atmosphere over a few acres may possess the essential 
force for generating a whirlwind, and may produce it in fact if 
the day be calm. A large storm, covering the whole country with 
its centripetal currents, must produce a vortex about the center, 
which will combine the principal energies of the storm. The tor- 
nado and water-spout must revolve with terrific violence. 

The necessary condition, centripetal motion, may arise when- 
ever a central spot subjected to intense heat is surrounded by a 
cool atmosphere. This state of things, on a small scale, may occur 
on a summer day, upon a ploughed field surrounded by extensive 


pastures ; upon a black and charred clearing in the midst of a cool 
forest ; or at a large clearing fire. Upon a great scale — if an island 
beneath a tropical sun received upon rocks and sands the intense 
radiance of a succession of clear, calm, and hot days, and conse- 
quent sea breeees from the deep and cool ocean pressed in upon 
all its shores with the violence of a high wind, it should not cause 
surprise if these various breezes combined to generate a vast 
whirlwind; nor if the lofty revolving column should at last leave 
the place of its origin and traverse the sea, as a hurricane. The 
cause which first excited the centripetal tendencies of the storm, 
might be renewed as the upper currents of the atmosphere bore 
it over other heated spots; and the law of deflection will inevitably 
transform the central into circular motion. The destructive 
storms of our sea-coast may have such an origin among the islands 
of the West Indies, from which they appear to proceed. 

FIG. 3. 

In the southern hemisphere the same law of deflection produces 
contrary results. There the wind which first moves north bends 
to the west, and the wind which moves south at first turns towards 
the east, that from the east turns south, and that from the west 
turns north. Fig. 3 represents these effects. Hence south of 
the equator storms revolve from left to right, or conformably to 
the movement of watch hands. Fig. 4 exhibits the rotary action 
of a storm in the northern hemisphere; fig. 5 the same in the south- 
ern hemisphere. 



VOL. 51 

The relative motions of the parts of a small circular space on 
the earth's surface, by reason of the diurnal revolution, are pre- 
cisely what they would be if the same circular space revolved 
upon an axis passing through its center parallel to the axis of the 
globe. If such space be regarded as a plane revolving about such 
supposed axis, then the relative motions of its parts are the same 
as if the plane revolved about its center upon an axis perpendic- 
ular to the plane itself; with* this modification, that an entire 
revolution on the axis perpendicular to the plane would not be 
accomplished in twenty-four hours. Such plane daily performs 
such part of a full revolution about such perpendicular axis, as 
the sine of the latitude of its center is of radius. The plane itself — 
the field over which a storm or tornado or a water-spout is form- 
ing — is in the condition of a whirling table. Hence the tendency 
to rotary action in every quarter of the storm is equal and all 
the forces which propel the air towards the center cooperate in 
harmony to cause the revolution. 

FIG. 4- 

FIG. 5. 

Water discharging from a broad basin through a central orifice, 
is subject to the same law. It forms a vortex which in our hemi- 
sphere turns to the left, or against the sun, and in the southern 
hemisphere must turn to the right or contrary to the sun there. 

These rotations of the atmosphere and of water, being from 
west to east about lines inclined to parallelism with the earth's 
axis, are singularly coincident in direction with the rotation of 
the globe, and harmonize with the general mechanism of the 






[Translated from Archiv fur Wissenschaftliche Kunde von Russland 
Vol. XXI, 1862, pp. 52-96 and 325-332 1 ] 

The first volume of Braschmann's Theoretical Mechanics, Mos- 
cow, 1859, combines to such a high degree condensed and precise 
presentation with abundance and variety of problems that the 
best interests of the mathematical study of motion demands its 
broadest possible dissemination 

As an example of the treatment of the problems enumerated 
above, we choose first the theory of relative motion of a free or 
restricted mass point and then the application of this theory to 
the so-called Foucault experiment and to other physical prob- 
lems arising from the rotation of the earth, which first began to 
be considered and empirically studied in recent years 

The second practical application that Braschmann makes 2 of 
his general formula? for relative motion consists (1) in the remark 
that every mass moving on the earth's surface along a restricted 
path exerts a horizontal pressure on the boundary of its path 

1 The early publications of William Ferrel relative to the motions of 
the atmosphere, beginning with 1856, were made accessible by reprints in 
Professional Papers, Nos. VIII and XII, of the U. S. Signal Service. In 
these, as in most other memoirs on the subject, the motions of the atmos- 
phere were assumed to be uniform in velocity, but in 1854-1862 Braschmann 
and Erman gave a notable enlargement of our ideas. Since that date two 
elaborate memoirs by Dr. Joseph Finger of Vienna (I, 1877; II, 1880) have 
given further details of the motions of bodies on a rotating spheroid. 

C. A. 

" See his article in Bull. Imp. Acad. Sci. St. Petersburg, Feb. 3, i860. 




VOL. 51 

which tends toward the right-hand side in northern latitudes but 
to the left-hand in southern latitudes: 

(2) In the determination of the amount of this pressure. 

The following is Braschmann's treatment of this problem as 
given in his Memoir of Feb. 3, i860. 

Equations (1) [general equations of motion, here omitted] de- 
termine also the accelerating force that the current of a river 
exerts on its right-hand bank. 

Let M N, fig. 1, be a section of the river, b any point in this sec- 
tion at which the average velocity of the steady stream of water 
is v, that is to say, the uniform or steady velocity is the result of 
all other influences combined with that of the acceleration due 
to the rotation of the earth. Designate by x positive towards the 
east and y positive towards the north the rectangular coordinates 
of a point in the horizontal plane through b, neglecting the very 
small fraction co 2 , we then obtain from equation (1) the following: 

d 2 x n . , dy 1 

— = + 2 co sin X. — 

dt 2 dt 

d 2 y . . dx 

— — = — 2 co sin /. 

dt 2 dt 


Let a be the angle between the direction of the current and the 
positive axes of y, then we have the following integrals 3 

3 For x = o and y = o or at the origin of coordinates, the steady velocity 
v and therefore its projections on the x and y axes are 


= v sin a 


^ = v cos a 



- = A +2 co sin X.y = v sin a + 2 co sin X.y. 


■ -- = B — 2 to sin X.x = v cos a — 2 to sin /.v. 

By substituting these values in equations (2) and neglecting or 

we have 

d 2 x 

— = + 2 co v cos a sin X 

dt 2 

= — 2 co v sin a sin X 
dt 2 

These equations determine the accelerative force 


+ (f| ) 

This shows that the amount of this force is independent of th 
direction of the current of the stream and moreover that it is 
directed steadily toward the right-hand bank. 

If u is the angle made by this accelerative force with the axis 
of y or the north line counting it positively around to the right, 

d 2 x -. . 

— = F sin u 
dt 2 

d2 y Z7 

— = F cos u 
dt 2 


F = + 2 co v sin X 


sin u = + cos a 

cos u = — sin a 

u = a + 90° 

that is to say, when X is positive the direction of F is 90 farther 
toward the right than the direction of the current. 
When X is negative we have 

sin u — — cos a 
cos u = + sin a 
u= a + 270° 


that is to say, in southern latitudes the pressure F will be against 
the left hand and will be zero at the equator itself. 

In order to determine the pressure mF exerted upon the right 
bank by the whole section MX of the river, let A be the area of 
this section and p the density of the water, then 

mF = Avp 2 10 a. sin / 

This mF relates of course to the pressure exerted by the mass 
of water that flows through the section MX in a unit time or by 
the mass Avp. 

In the preceding paragraphs we have assumed a constant or 
steady and uniform velocity. It seems unnecessary to say that this 
condition is rarely fulfilled in practical cases, or that the demon- 
strated result does not hold good in such cases. But since there 
is a distinction to be made between a theorem that is not yet demon- 
strated and one that is clearly shown to be erroneous it will be of 
interest to see from the following modifications of Braschmann's 
analysis as summarized in the Comptes Rendus, Paris 1861, how 
in the case of non-uniform motions along paths restricted to the 
earth's surface, the direction of the horizontal pressure depends 
on the variations of the respective components of its velocity. 

[The original memoir in the Paris Comptes Rendus contains 
numerous typographical errors that are corrected by Erman in his 
Archiv Wiss. Kunde Russland.] 

Let X, Y, Z, be the projections (or rectangular components) of 
the accelerating force and P x , P P z the projections of the pres- 
sure exerted on a unit surface, at a point whose coordinates relative 
to a fixed system of rectangular axes are x A \\ z x ; we have then 

X - ** 


= P X 

Y - ** 


= Py 


= P 2 


If both the axes and the origin of coordinates are all assumed to be 
in motion then we have to substitute herein the values of the accel- 
eration at the point \\ y t z v Let .v, y, z be the coordinates of this 


point referred to the moving axes at the time t. Let (o t a> 2 co 3 be the 
angular velocities of this point about these axes at the same instant. 
Let oj be the resultant of these aagular velocities and let a, /?, y be 
the coordinates of the origin of the movable system of axes with 
reference to the fixed system; we have then the following expres- 
sion for the accelerations: 

d 2 x, d 2 x „ / dz dy \ dw 2 du> 3 

— l . = — + 21 a/, - cu 3 — ) + z — - - y- — - 
df- dt 2 \ dt dt dt dt 

+ io l {(0{X + oj 2 y + co 3 z) — or x + 

dt 2 ' 

d 2 y x _ d 2 y 
dt- dt 2 

+ 2 

dx dz \ dm, daj. 

j 3 — co l t- v 


dt J 



d 2 d 
+ oj-, (oj. x + co^y + u)„z) — ory + 

dt 2 

d 2 z t d 2 z I dy dx \ 

— - = — + 2 oj. - — (o 2 — ) + y 
dt 2 dt 2 \ dt ' dt 

■ (2) 

dcu. dco 9 

— x 

dt dt 

d 2 r 
+ oj 3 (w t x + w 2 y + co 3 z) - arz + — - ' 

dt 2 

[The detailed demonstration of these expressions was given by 
Braschmann in the Bulletin of the St. Petersburg Academy for 
185 1 and is repeated in Vol. I, Chap. IV of his Treatise on Theoret- 
ical Mechanics and at pp. 74-88 of Erman's Archiv XXI. 1862; 
its equivalent is found in many recent treatises on mechanics under 
"constrained motion."] 

If the last named initial point or origin of the movable system 
of coordinates be at the point on the earth's surface and movable 
with it about the earth's axis then it has a constant velocity of rota- 
tion about this axis. Hence the accelerations of its motion are 
zero and therefore we have 

d 2 a 
dt 2 



dt 2 


= f) 



Imagine the three coordinate axes movable with the earth to be 
so drawn through this point that the xy plane is horizontal, x being 
positive eastward along the tangent to the small circle of latitude 
and y positive northward along the tangent to the meridian but z 
positive downward toward the center of the earth considered as a 
sphere. Let 01 be the direction of the positive half of the momen- 
tary axis of rotation or in our case a line drawn through parallel 
to that half of the earth's axis that extends from its center to the 
south pole; then in general 

to l = to 2 cos (Ix) to 2 = — to cos (ly) to 3 = to cos (Iz) 

and in our special case 

to x = to to 2 = — to cos X to 3 = to sin X 

where X is the latitude of the place of observation (O) and to is the 
angular velocity corresponding to the diurnal rotation of the earth. 
Since in this case of steady rotation 

• dvi = d&2 = dc °3 = 

dt dt dt 

and considering the conditions expressed in equation (3), therefore 
in the present problem the general equations (2) become 

<Px x d 2 x I dz dy\ . 

— - = — — 2 tot cos X - - + sin / — to 2 . x 

dt 2 dt 2 \ dt dt I 

u?'V d?'V doc 

— = — ~ — 2 to sin X — — to 2 cos X ( — cos X.y + sin X.z) — afy I to \ 
dt 2 dt 2 dt 

c*z d^z doc 

— - = — 2 to cos X - — to 2 sin X ( — cos X.y + sin X . z) — to 2 z 

dfi dt 2 dt 

The terms in to 2 may be omitted because they are very small; but 


for the same reason the term 2 to cos X . must be omitted in all those 


cases in which the gradient of the surface (the rails or the river), 

and hence also — , is small or zero. 

Let a be the angle between the direction of the motion of the point 

x, y, z and the direction of the positive axis of y and let v be the 



velocity of this point, then 



= v sin a 

v cos a 

Let v x v v z be the components of the momentary velocity along 
the axes x y z and hence their differential quotients with respect 
to t will be the momentary accelerations in these directions and 
identical with 

d 2 x d 2 y d 2 z 

dt 2 dt 2 

Equation (2 X ) now becomes 
d 2 x x _ dv 3 
~d~f It 


2 co sin X.v . cos a 

d 2 y, dv„ , _ . , 
■ /1 = y + 2 co sin / . v . sin a > 

d* 2 

d 2 ^ 


+ 2 to cos ^ . v . sin a 


Now gravity and friction are the accelerative forces acting on a 
point in contact with the sides of the track or path of constraint. 
The projections of gravity on the horizontal plane are equal to 
zero and equally so the projections of the lateral friction on the 
direction of the lateral pressure disappear. Hence if we designate 
the acceleration of gravity by g we have the forces 

X = 

Y = 


and equation (4) gives the pressure 




+ 2 to sin X.v . cos a 

P = — — 2 — 2 co sin yLv.sin a > (5) 

y dt 

P, = e — — 2 co cos A.v. sin a 
2 * dt 


If the horizontal motion is uniform then 



*?> = and dv y = 
dt dt 

P x = +2w sin X v cos a 


Py = — 2 co sin X v sin a 

P = 2 co v sin X 

where P is the whole pressure exerted in a horizontal direction. 

If we substitute in equation (6) successively all values of a between 
a = and a = 2 71 we soon perceive that the pressure P is 
always perpendicular to the path of the moving mass and if X is 
positive the pressure is always directed to the quadrant on the 
right-hand side of the direction of motion. 

If the motion is uniform along the axis of % only, then for positive / 
the direction of the pressure P will still be always toward the right- 
hand of the direction of progress of the mass so long as the value 
of a lies between and n, i. e., so long as the progress is in a direc- 
tion between east and south and west. 

But if the motion is uniform along the axis of y only, then for posi- 
tive X the pressure of P will be directed toward the right of the direc- 
tion of progress for any value of this latter direction that lies be- 
tween a = it and a = 2z. 

If the motion is not uniform along the axis of x or the axis of y 
then the direction of the pressure P may be either toward the left 
or the right of the direction of motion depending on the current 

dv x dv 

values of — r~ and -7- to an extent and manner easilv determined from 
dt dt 

equation (5). 




In the third book of his Mecanique Celeste, Laplace has demon- 
strated that the atmosphere of a rotating planet is at rest relative 
to any point of the solid nucleus of this planet and that at the same 
time any pressure and any density can occur within any level sur- 
face of such an elastic fluid, i. e., within any surface that at any 
point is normal to the resultant of gravity and centrifugal force. 
In this he assumes that a uniform temperature prevails throughout 
the elastic fluid. 

The surface of the sea is such a level surface and apparently on 
the strength of the above demonstration by Laplace most physicists 
assume that the product of gravity by the height of the barometric 
column 2 which measures the pressure of the air must necessarily 
be the same everywhere at sea level. They grant that temporary 
disturbances of the atmospheric equilibrium are accompanied 
by temporary interruptions of this uniformity of atmospheric 
pressure, but imagine that these two exceptions are only periodical 
(viz: variations about a mean condition) which mean must be 
primarily a condition of rest relative to the earth, and secondarily 
must be that uniform mean reduced barometric height that one 
should find from measurements taken during one or many whole 
years at different points on the earth's surface. 

The falsity of every portion of these assumptions was shown 
many years ago and should have been evident a priori still earlier. 

I have found the mean reduced barometric readings for different 
localities at sea level extremely different. Among others, for in- 
stance, the pressure at the polar limits of the two trade wind belts 
is from two to three Paris lines (0.18 to 0.27 English inches) greater 

1 Translated from the Astronomische Nachrichten, No. 1680, February, 
1868, Vol. LXX, cols. 369-378. 

2 For brevity I will call this product the reduced barometric height or 
the pressure of the air. 

3 1 


than at the equator; again, on the Sea of Okhotsk and at Cape Horn 
it is six Paris lines (0.53 inch) smaller than at latitude 23. 5 . At 
any point in the interior of a continent whose altitude above sea 
level is known by geometrical measurements (e. g., railroad level- 
lings), the observed mean air pressure can be reduced to the value 
appropriate to the sea level vertically below it and thus gradually 
an empirical expression can be found for the pressure at sea level 
and, therefore, for the atmospheric pressure in general, as a func- 
tion of the longitude, latitude, and altitude above sea level, or the 
distance from the center of the earth. Individual contributions that 
I have made to this subject leave no doubt that above the solid 
land, as also above the ocean, the mean air pressure at any lev 
layer differs very much according to the latitude and longitude. 

The first of the above stated two fundamental assumptions (v 
that the atmosphere is in a state of rest relative to the globe' 
also decidedly negatived by ordinary observation. In one p 
tion of the atmosphere, lying between the parallels of +25 an> 
— 2 5 the air is at every minute and, therefore, on the averagt 
of all time, in that state of strong steady flow that we call the trad t 
wind; in other words, therefore, the average or permanent condition 
of 0.4225 of the total mass of the atmosphere is a regular flow thai 
is certainly not to be ignored. In the remaining portions of ,h 
atmosphere, however, the movements are less steady as to time. 
But when the successive motions of the air, during one or rr n 
whole years at any place are combined into one resultant, the 
general this resultant differs from zero and in such a way tha 
direction and velocity depend upon the coordinates of the loca 
and are independent of the years or number of years for which 
computation was made. 

Therefore, after eliminating the influence of periodic and accid*- 
tal circumstances and in direct opposition to the above-gi" 
assumptions of the physicists, we find that the earth's atmosph' 2 '- 
shows the following phenomena: 

(1) A current whose direction and velocity are independent of 
the time and which, therefore, at every place depend only on the 
coordinates of locality. 

(2) At any level surface (or one that is perpendicular to the result 
ant of the explicit forces) the atmosphere is under a pressure that 
varies with the coordinates of the points of this surface, but is con- 
stant as regards the time. 

These two observed facts contradict the results of the ana" sir 
of Laplace only because in place of a certain assumption fl nis 



analysis precisely the opposite holds good in the earth's atmos- 
phere. The temperature which Laplace assumed to be uniform 
throughout the fluid is in the earth's atmosphere extremely unequal 
and, indeed, not only so in respect to the periodical portion of its 
expression depending on the time, but also as to the other perma- 
nent portion, which we ordinarily call the mean temperature of the 
place. These mean temperatures, which are invariable as to time, 
are, as is well known, a function of the coordinates of the location 
to which they belong and, not only the analytical form of this 
function, but also the constants that enter it are already known 
.vith considerable approximation. 3 In accordance with these re- 
alts of experience, it is certainly worth while to investigate the 
.ollowing problem : 
What is the nature of the movement and how great is the pressure 
reduced barometric reading at any point of an atmosphere for which 
uoth the resultant of gravity and centrifugal force and, also, the tem- 
perature are expressed as functions of the coordinates of locality and 
are independent of the timet 

The remarks that follow seem to me to prove that this problem 
can be solved. 

If at any point of a liquid, or an elastic fluid or gas, at the time 
/ and with reference to three rectangular axes, we have the coordi- 
tes x, y, z\ the explicit forces X, Y, Z; the velocities of motion u, 
w, and if at this same point x, y, z, we have the density p, the 
assure p and the temperature x, then by combining the conditions 
equilibrium of this fluid with the general theorem for the move- 
nt of any system, remembering 4 that 

u = — 



v = _ ; w - 


dz . 

*ve obtain 

p dx 

du du 

— — — u — — 

a* dx 

V — — 



w — 


ld t = Y 

dv dv 

= — —u — 

dt dx 


v — — 


w — 


\ d l = z 

p dz 

dw dw 

- — — u — — 

dt dx 


V — — 


w — 


->ee for example Erman's Memoirs in the Archiv Wiss. Kunde, Russland. 
he translator has taken the liberty of substituting d for partial differ- 
and d for total differential instead of Erman's notation. — C. A. 


and, as the condition that the mass of each particle of fluid shall 
remain invariable, we have 

o = d(pu) + d(pv) + d{pw) + dp 
dx by dz dt 

In these equations the variations with the time, when x, y, z re- 
main constant, are expressed by the differential with regard to t. 
Since now in our present case of the earth's atmosphere, we 
assume that at every point the pressure, temperature, motion and, 
therefore, the density are invariable as to time, therefore, we may 
in equation (II) substitute 


and also in equations (I) 

du dv dw 

Furthermore, we have 


dt dt dt 

p.d P(l + kz) p 


P (1 + kz) d p 

where Pis that value of the pressure (or the reduced barometric 
reading that measures the pressure) for which the atmospheric 
air at the temperature o° C. is o times heavier than the mercury 
of the barometric column and k is the coefficient of expansion of 
the air for one unit of the thermometric scale that is used to measure 
the temperature or r. 

Since z is assumed to be independent of the time it can be ex- 
pressed as 

z = f {x, y, z) 

where / indicates a known function, which, as above stated, is now 
approximately known; therefore we may also write 

p P Pk . . , - , , 

~ = - + — / (x, y, z) = F (x, y, z) 

where F (x, y, z) again indicates a known function which for brevity 
we indicate by F* 


Therefore, we now have 

p p 1 F. 

p = = or - = — 

F (x, y, z) F. p p 

and the left-hand portions of the equations (I) become respectively 

1 dp = F dp =F dlogp =x 
p dx p dx dx 

1 dp = F dp = p dlogp =y 
p dy p dy dy 

I dp = Fty = F d log p _ z 

p dz p dz dz 

where log indicates the natural logarithm. 

Since it is known that the components x, y, z, of the resultant 
of gravity and centrifugal force, as also the components u, v, w 
of the velocity of a fluid particle, depend only on x, y, z or on the 
variable coordinates by which we express the location of any point 
on the sphere, therefore, as is well known, there are two functions 
of x, y, z, which I will designate by V and <p respectively 5 which are 
determined by the relations 

x = dV . V = dV • Z dV 


dx dy dz 

d<p do do 

— ; v = - ; w = — 
dx dy dz 

If these values are substituted in equation (I) and the sum is 
taken after multiplying the first, second, and third respectively by 
dx, dy, dz, and if we recall that 

5 These functions are now generally called the force potential, V, and 
the velocity potential, f, or the potential function for the external forces and 
the potential function for the resulting velocity. If such functions actually 
exist there can be no discontinuous whirls, and if the whirls exist then there 
can be no such functions. — C. A. 



VOL. 51 

dU , , dv , dw , 

u . dx + — . u . dy + — . u . dz — 

dx dx dx 





dM Bl d -f 

dx+A d lL. d y + ^. dz 

dx dx 

a 9 l»(* 


dip . (d<p 

dx \dx 
and recall that 

dx+ ydX '.dy+ W 



dy \dy I dz v dz I 


result in a similar manner from the six remaining terms of the right- 
hand side of the summation, then there results 

-dp = F .dlogp = dV - \d 

dx J \dy 

? (A) 

= F .d log p -f dF. 

Again, from equation (II), after substituting the values of u, v, w 
and dividing by p, there results 8 

/ d log p dip d log p dip d log p dip 
\ dx dx dy dy dz dz 

«V «V «V 

S dx 2 + dy 2 + dz 2 


The equation (A) can be more perspicaciously replaced by the 

d log tp 

•Because p is assumed not to vary with t, therefore — U, and this 


term drops out. — C. Abbe. 


dlogp 1 BV blogF 1 \\dxj \byj \bz I / 

bx F dx dx 2F dx 

, •{(ZHtHi)! 

b log p 1 bV b log F 
dy F dy by 2F 3;y 

Olog/j 1 3F 3logF 1 {\dx/ \byj \bz 


bz F bz bz 2F 

and we can, therefore, by simple substitution of A* in B construct 
an equation in which, in addition to the partial differential quotients 
of the first and second order of the function <p — <p(x y z) there enter 
only the known functions F = F (x, y, z) and V — V (x, y, z) and 
their first differential coefficients. 

If x, y and z are replaced by the angular coordinates of any point 
of the atmosphere so that 

x = r cos X V\ — p. 2 = r cos /? cos X 

y — r sin X Vl — fi 2 = r cos ft sin X 

z = r n = r sin /? 

where r is the distance from the center of the earth to any point 
in the atmosphere; X is the longitude of this point; /? is the latitude 
and jj. is the sine of the latitude, then, as is well known, we have 7 

y = (r * 2) -Hi)-' 2 - 2 > 2 -^ 2 

where for the surface of the earth and at the equator, we have 

r =R 

y = acceleration due to the attraction of the earth 


a = acceleration due to centrifugal force = ~i— • 


7 This expression assumes that gravity and centrifugal force are the only 
external forces and thus ignores viscosity or internal friction and the resist- 
ance of the earth's surface and the attractions of sun and moon. 


This expression for V, as well as those that must obtain for F, 
for <p and for p respectively, when we write these out as functions 
of p, X and r, all possess the properties that Laplace has demon- 
strated for all functions of this kind that have definite real values 
for constant r and for all values of X, from o° to 360 and of p from 
— 1 to + 1. That is to say, since this latter condition (the having 
a definite real value) is evidently fulfilled in the earth's atmosphere 
as to temperature r, density p, and the function ip whose differential 
coefficients are the component velocities, therefore, in accordance 
with the Laplacian theorem just referred to, we may similarly assume 
for V, for F, for <p, and for p, respectively such expressions as the 

a P° + P P' + y P" ■ ■ ■ • + v P {n) + . . . . 

in which, as we pass from one to another of these four functions, 
V, F, <p and p, there occur: 

(1) Those coefficients a, /?, y . . . u which only vary with r 
(2) the constant numbers that enter into P°, P 1 , P 2 , P n , as defined 
in the next following paragraph. 

In general P n is defined by the partial differential equation 

d { (1 - p n ) 

dp J \ dk" 

= — — - + T 2 + n (« + 1) P (n) . 

op 1 — p" 

and from this it follows explicitly that 

P (n) = 5 n ° X^ + (A' n sin / + B' n cos /) X (1 ~^^ . d ^- 

n dp 

(1 - u 2 Y /2 a' Y (2) 

+ . . . . (A} sin * X +5 n * cos i X) K - ^ . —^- + .... 

n (n — 1) (w — i + 1) dp 1 

where B n °, A n ' , B n ' are constant numbers and 

X („) = n _ n(n-l) 2 n(n-.l)(n~2)(n-3) ^ _ etc 
2(2»-l) 2.4. (2» -1) (2 » -3)' 

Since the development of each function, V, F, <p, p, in the form 

aP° + PP' + etc., 


is only possible in one manner and since it always gives a converg- 
ing series, therefore, each of the functions occurring in equation (B) 
consists of a limited and, in fact, probably a small number of terms, 

aP°,pP', r P",etc, 

which altogether constitute a series progressing according to the 
whole powers of 

p and (1 — fi 2 )* 

whose terms are multiplied into the sines and cosines of multiples 
of X. Furthermore, since the terms of this nature, in equation (B), 
resulting from the development of V and F contain respectively only 
a well-known function of r, while, on the other hand, the terms aris- 
ing from the differential quotients of <p contain the functions a, /?, y 
of this same form and the constants A, B, C — which are the only 
unknown quantities of the problem — therefore, these latter must 
be determined by equating to zero each of the sums of known and 
unknown terms that in equation (B) are multiplied by 

p q (l -p 2 )*siniX or p q {\ - pP)*cosiX. 

In order to practically execute the determination of the velocity 
function <p, for a given temperature function F, we can easily con- 
vert that form of the latter equation which results from the combi- 
nation of equations (A*) and (B) into the equivalent differential 
equation in r, p, and X whose specialization then leads directly to 
the desired end. 

The two following relations between any two functions, $ and 
<£', of the coordinates x, y, z are easily demonstrated 

00 00' 00 00' d$ dcf>' _ 00 d<p' 00 00' 1 - p 2 
dx dx by dy Oz dz Or l)r dp dp r 2 

d<f> 00' 1 



dX dX (1 - p*)r 


old- 2 ) d ±\ (?*) 
d2 ± + a V + #* = J_ dpi W/ r O 2 (r0 ) 

dx 2 dy 2 dz 2 dp 1 - a 2 dr 2 

vvuence it follows that equation (B) may be written thus: 


Q = /dV _dF\d<p /dV _dF\l-fi 2 d<p /dV _dF\ 1 d<p 

~\dr Br)~dr \djt ty / ~r*~" dp dX dXj (1 - ft^'dX 

1 J \dr J \dr J r \b\ J (1 -p 2 )r 2 J d<p 

+ 2 * dr ' dr 

f^y + /a^i-^ + /a^y i \ 

1 I \drj \dfi/ r \dl/ (l-/( ! )rj 1 - /r dp 

2 dp t dp 

1 I Vdr/ Vfy/ r 2 Vd/l/ (1 - ,« 2 ) rf (I - p 2 ) 6? 

+ 2 3/1 r 2 ~&l 

faj(i-„ 2 V^ f 3 V\ 
+ i 7 j IV / ajtc J + \ap/ + r a 2 (,^) 

L a// l — p 2 dr 

where, finally, V, F, and <p are each to be replaced by a converging 
series of the form 

aP°+ pP' . . . . + vP (n) 

and where the series for V and F will contain known functions of 
r and known constants, but the series for <p will contain similar 
terms whose functions and constants are to be determined. 

In order to obtain an approximate idea of the practical solution 
we may take the above given value of V, 

V= ( r R 2 ) i + /jLW( Ur.y-1), 
r \ZRj \2RJ 

which agrees with the form of the converging series when we put 

$ - d = £ = . . =0 

and further take for the function of F the following. 
F = a + br + c (p. 2 - $) 


which agrees with the form of the converging series by putting 
a = (a + br) 

£ = £ = 6= *. . =0 

X = + c 

In agreement with these assumptions determine the functions of 
r that I will in general indicate by 


and the constants 

A in) B (n) tc 

(to) (to) 

that enter into the following general expression for (f, 

? = fo° + h° 00 + ft 0V sin * + B x cos *) (! - /* 2 )* 

f U° W -*)+/»'(!- P 2 )*- /* ■ W sin ;, + S 2 ' cos /I) (1 - // 2 ) 
+ / 2 2 (A 2 2 sin 2 ^ -f B 2 cos 2 i) 

+ /4- 3 2 |,) + / 8 '(i-^) ! (/^)x 

X (^4 3 ' sin ^ + B 3 ' cos >i ) + . . 

From equation (C) it is evident tnat in this special case the terms 
in <p that are multiplied by functions of X must disappear and that 
therefore also the direction and velocity of the steady wind must 
be as independent of the longitude of the place as is the assumed 
distribution of temperature. 

If we have thus carried out the determination of <p, then, from 
equation (A) there follows 

log£ = 


/ /a^V , /a^yi-V /a^y i 


\ \dr/ \dnj- r \dX) (1 - n 2 )r 2 

+ Constant. 

and since the quantities under the integration sign can also be devel- 
oped in series of the kind above considered, then, for all points of the 
atmosphere, we shall know 


where b designates the mean barometric pressure, as soon as we 
have determined the constant of integration by observation of the 
barometer at only one point, for which also the force of gravity, 

uwy + /*ry + /*vy 

\\dx / V dy J \ dz / 

shall have been given. 

If the mean barometric pressures computed in this manner be 
compared with the observed pressures, we get a sharp control over 
the theory. 

It is only when in this way it shall have been shown that the 
observed steady components of the motions of the atmosphere and 
of the barometric pressures are not properly represented, that we 
shall be justified in assuming that the friction of the particles of air 
against each other and against the earth's surface exert a sensible 
influence on the phenomena. In this case, and in so far as the 
friction is assumed to be proportional to the square of the velocity 
and uniform throughout the atmosphere, we shall have to replace 
equation (A) by the following: 

Fd\ og9 -dv-dF-i* 1 (*yw*y +/ a? 

dx / \ dy I \ dz 

i*y + (*y + f»\\. 

dx/ \dy/ \dz I j 

By the introduction of the fourth term in the right-hand side of this 
equation and of the undetermined constant, C, nothing of importance 
is changed in the development above considered. 



(Dated Chemnitz, February, 1881) 

[Annalen der Physik und Chemie, New series XIV , whole series 250. Leipzig, 

1881, pp. 117-128] 


For small zenith distances the atmosphere constitutes an optical 
system of refracting media separated by centered spherical sur- 
faces of small aperture, so that the theory of such optical systems 
developed by Gauss and Mobius 1 can be applied to it. The impor- 
tant matter is the determination of the "Cardinal points" by this 
theory. It is well known that the cardinal points are : 

(1) The principal foci / and f in the first medium A and the last 
medium A' (fig. 1). At either of these points are united the rays 
that come through the opposite medium parallel to the optical axis 

(2) The nodal points k and k' having the property that every 
ray (ab) passing through the first medium in such a direction as 
would pass through k, when it reaches the second medium goes on- 
ward in a direction c a' passing through k' parallel to its initial 
direction ab. 

'Gauss: Dioptrische Untersuehungen, Goettingen, 1840. Compare also 
Helmholtz, Physikalische Optik. Braunschweig, 1861. In the present arti- 
cle I refer to the Physik of Mousson, which is widely used. 




VOL. 51 

(3) The principal points h and h' for which object and image are 

As in all complicated systems whose constitution is not known, 
so here, the location of these cardinal points can only be found 
experimentally 2 based on the astronomical determinations of the 
refraction of light by the atmosphere. 

Let m in fig. 2 be the center of the earth, c the location of the 
observer; cc' a small arc of a great circle of the earth; bb' the inter- 
section of the circle with the boundary of the atmosphere; a a fixed 
star; abc a pencil of rays from a toward c which is refracted at c 
toward the direction cf so that the apparent zenith distance 
£ determines the astronomical refraction p; am the axis of this 
optical system whose first medium is the vacuum for which the 
refractive index is n = 1, and whose last medium is the lowest 
stratum of air whose refractive index is n', and which is assumed 
to have no limit. 

The nodal points of the atmosphere coincide at the center of the 
earth. For, because of the concentric boundaries of the refracting 
media a pencil of light passes through the atmosphere in a curve 
whose tangents are never parallel to each other, so that only one 
ray, moving in the medium n' toward m, can proceed farther in the 
same direction. 

The location of the second focal point f is easily found when we 
consider the ray starting from a as originally parallel to the axis. 
It is deflected from its initial direction by the amount of astronom- 
ical refraction p and in the last medium finally proceeds toward 
the second focus of the system or f . The corresponding focal dis- 
tance F' is found from the triangle c. m. f ; in this triangle we have 
c m = R, y = r and since in this case ab is parallel to af there- 
fore a' = p hence 

F' = >-R. 


Mousson: Physik, 2d Edition, section 731; 3d Edition, section 810. 


The ratio of the astronomical refraction to the apparent zenith 
distance is - = 57.3" according to the current tables of refraction 


for the average condition of the atmosphere, 3 so that we can write 
•° =57.3" (1 + J) (2) 


where 5 7.3 "J represents any possible error of observation that we 
include in our computations in order to show the degree of accuracy 
of the results. Hence we have 

F' - R -(l +J) 

= about 22 918 400 kilometers for R = 6366 .7 kilometers. 

> (3) 

The first focal distance of the atmosphere follows from the relation 
according to which the distance of the two focal points from the 
respective nodal points is inversely proportional to the ratio of the 
refractive indices of the respective media. 4 Hence we have 

n' R 
F = n'F' = (1 + i) = about 22 924 900 ... (4) 


where we assume the coefficient of refraction for the average condi- 
tion of the atmosphere (b = 0.752 meters, t = 9.3 Cent.) to be 
n' = 1.000282 according to Ketteler's determination. 

All celestial bodies except the moon are farther removed from 
us than the focal distance F. Hence the atmosphere produces 
as a whole and at the focus inverted real images of these. The 
convergent pencil of rays coming through the atmosphere enters the 
eye so that it also unites the pencil into an inverted image within 
the focus <p of the whole eye (or it falls short of the retina) and thus 
there forms on the retina under all circumstances for a perfect far- 
sighted eye a circle of diffuse light, no sharp image, and thus the 
spread of the stellar image over the retina finds its physiological 

On the other hand, the atmosphere produces a correct upright 
virtual image of the moon because she is within the front focal dis- 

* Bruhns: Astronomische Strahlen-brechung, p. 19. 

* Mousson: 2d Edition, section 731; 3d Edition, section 809 (4). 


tance and the eye receives a diverging pencil of rays. Thus the 
inverted image produced by the two optical systems combined, 
the eye and the atmosphere, lies inside <p counting from the retina 
forward; but on account of the great distance of the objects, always 
so far from the retina that it requires an extraordinary power of 
accommodation to produce a sharply defined image on the retina. 
The muscular tension thus excited explains the apparent floating 
of the moon in the atmosphere. But even with this extraordinary 
accommodation the image of the moon will only just attain the 
nerves of the retina, and since this partial touching also occurs in 
ordinary vision for objects that are at a definite terrestrial distance 
that we may call D, therefore the eye locates the moon also at the 
same distance D because it produces the same sensation or excite- 
ment on the retina, whereby is explained the apparent floating of the 
moon at a relatively nearby point in the atmosphere. 5 

In relation to the location of the principal point, h, its distance 
from the corresponding focus is equal to the distance of the opposite 
focus from its nodal point. 6 
Therefore we get from fig. 2 

and also 

fh = F' and fh' = F 

mh = fm — fh = F — F' 

mh' = fh' - fm = F - F* 

mh = mh' = F — F' 

and thence by substituting from equations (3) and (4) 

in' — 1) R 
mh = mh' = !_ (1 - A) = about 6463 kilometers . (5) 

57 .3" 

Therefore the two principal points, like the nodal points, coincide 
in one point h, that is 6463 km. distant from the center of the earth 
or 96.3 = 6463 — 6366.7 above the surface of the earth at c (fig. 
2 or fig. 3). Therefore according to the definition of the principal 
points the object and the image coincide at the point h, that is to 

6 We determine D experimentally by measuring the distance at which an 
intense flame is sharply seen. 

8 Mousson: 2d Edition, section 731 (2); 3d Edition, section 809 (2). 



say, all rays that in vacuo are directed towards h, diverge after their 
passage through the atmosphere from this same point h. If, for 
instance we imagine a planetary nebula between any iixed star 
and the earth and the star-like image of the nebula located at the 
point h (see fig. 3), ymich is now to be considered as the luminous 

fig. 3. 

object seen through our atmosphere, then will the rays ab converg- 
ing toward h be so refracted by the atmosphere, that on their 
entrance into the last medium 11' they will appear to diverge from 
h in the direction cd, and hence an identical virtual image should be 
seen at that same point (h) at which a real star-like image must 
have existed if there had been no atmosphere. 



By reason of the nature of the curve of the beam of light abed, 
which has its concave side toward the center of the earth, it is evi- 
dent that the principal point must lie within the atmosphere. 
For if e, fig. 3, were the principal point, then a pencil of rays that in 
vacuum may have the direction ef must necessarily on its entrance 
into the medium n' go on farther in the direction egk, if an identical 
image of the object is to be formed at e; but this is impossible be- 
cause ef and gk cannot be tangents to the same curve at / and g. 
Hence therefore it follows that H > ch, that is to say, according to 
the note on equation (5), 

H > 96.3 kilometers 


But a more accurate determination results at once from the fol- 



VOL. 51 

lowing consideration. Since for small zenith distances the curved 
path of a ] encil of light is very nearly an arc of a circle of very large 
radius of curvature 7 therefore the tangents (to these curves) bh 
and ch, fig. 3, can be considered as equal and because for small 
zenith distances the tangents can be exchanged for the distances 
of the principal points from the limits of the atmosphere, therefore 
we have approximately 

H=2 ch 
or from equation (5) 

H =2R 

ri - 1 

(1 - A) - 1 

= about 192.6 kilometers 


Hence the determination of H from the observations of the twi- 
light arc as given by Alhazen (leading to the value of 79 kilometers) 
is far too small, and in fact it follows from Fresnel's formula for the 
intensity of reflected unpolarized light that the argumentation by 
Alhazen by no means excludes the existence of still higher strata 
of air. 

According to Fresnel's formula, if the incident light has the inten- 
sity unity, and e and b indicate the angles of incidence and refrac- 
tion, then the intensity of the reflected light is 


1 sin 2 (* - b) 

2 sin 2 (e + b) 

1 + 

cos 2 (e 

cos 2 (e 


Since in one case the reflection takes place at the thinner layer 


therefore the refractive index is 

n + dn 

sin b =.(1 + dn) sin e 
cos b = cos e — sin e tg e 

and by substituting these values we obtain 

and we have 





2 cos 2 e — dn 

or approximately, since e is not far from 90 

j„ = I dn 

\2 cos 2 e 

1 + (sin 2 e — cos 2 e) 7 


7 Bruhns: Astronomische Strahlen-brechung, p. 66. 



Now let mc in fig. 4 be the radius of the earth and ac the horizon 
of the place c, and 5 the sun. If the sunlight is reflected from the 
layer of the air at a whose radius is R + h toward the horizon at e 

sin e 

1 - 


cos^ e = 

2 h 

and the intensity of the beam of light seen at c will according to 
equation (8) be 

\ 4h I 

FIG. 4. 

If now the twilight ends, or the stars of the feeblest intensity J c 
become visible to the naked eye, when the sunlight is reflected from 
the layer of air at an altitude of 79 kilometers, then the only con- 
clusion that should be drawn is that at this altitude the intensity 
of the reflection 

/ dn R V 
U X 79/ 


has become less than /„. But whether there are still higher re- 
flecting layers of air and a still further diminution of the intensity 
of the twilight is beyond our power of direct observation ; however, 
the possibility cannot be gainsaid so long as J > 0. 

The fact that the limit of the atmosphere really is considerably 

higher than 79 kilometers, assuming that - = 57.3" is correct, is 


shown by the well-known differential equation for astronomical 

R , . 

n . sin £ . n' 

R +h 11 

d p - - . - ... (9) 

J n 2 - ( _? _ V n n sin 2 r 
^ V R + h ) 

which for small zenith distances reduces to the expression 

« R , 11 

d p > n' r 

R +h u- 

If for h we substitute its maximum value H so that the right-hand 
side of this expression becomes still smaller, then 

s ^ R , d n 

p > . ri r — 

R + H ^ « 2 

and after integration between the limits n = 1 for the highest layer 
and n = n' at the earth's surface we have 


in which 

p 57.3" 

H > 


C ' 


n' - 1 = 0.0002820, R = 6366 .7 kilometers 

H > 06.3 kilometers (10) 

which may be compared with the value in (6) above given. 




For brevity I speak of the two conjugate points a and a' in fig. 5 
as symmetrical wh>en the angles of divergence of the rays from the 
axis are equal to each other so that 

a = a 


Let the distances of these points from the center of the earth be 
D and D'. According to the theory of Gauss the ratio of the dimen- 

sions of the object and its image, or D/D f , multiplied by the ratio of 

the tangents of the respective angles of divergence ^— is equal to the 


inverse ratio of the indices of refraction of the first and last media 8 


D l i a „•> 

or by equation (11) 

D' tga' 



and by equation (4) 









A second equation between D and D' results from the relation devel- 
oped by Gauss between two pairs of conjugate points 9 for instance, 
the nodal point m and the symmetrical points a and a' of fig. 5. 
If we divide the distances F and F' of any pair of conjugate points 

•Mousson: 2d Edition, section 730 (7); 3d Edition, section 808 (19). 
•Mousson: 2d Edition, section 730 (4); 3d Edition, section 808 (16). 

5 2 


VOL. 51 

ra, from the focal points, by the distances D and D' of the two pairs 
from each other (see fig. 5), then the sum of the quotients is unity 
and we have therefore 

D D' 


hence from equation (12) 


= 1, D =2F, D' =2F'. 

From these we at once find the distances of the points of symmetry 
from the (upper or lower) boundaries of the atmosphere, namely, 

a' c' = 2 F' + R 

and substituting 

F = 22 924 900 km. R = 6366 .7 km. 

F' = 22 918 400 km. H = 200 km. 

there results 

a V = 45 843 233 km (14) 

a' c' = 45 843 167 km. 

Hence the points of symmetry are at approximately equal distances 
from the boundaries of the atmosphere, and since a = a' the points 
of entrance and exit, b and c, of the corresponding pencil of rays 

FIG. 6. 

are at equal distances from the axis so that be is parallel to am. By 
the zenith point of the pencil I understand the point of intersection 
d (fig. 6) of its initial direction with the prolonged radius of the place 
of observation. 


Let a and a' be the conjugate points; D and D' their distances 
from the center of the earth ; a and a' the angles of divergence of 
the corresponding pencils; then in the neighborhood of the zenith 
and according to the theorem just stated we have 

D .« = n' (16) 

D' a! 

But from the triangle mca' since y = £ there results 

D' r 

Z- = y or D' a' = R r 
R a' 

substituting this in equation (16) gives us 

D a 



Furthermore from the triangle mad, designating the altitude of the 
zenith point by h we have 

a R + h 

C +P D 


D a = (R + h) (c + p) 

so that equation (17) becomes 

i+ i)KH (i8) 

If in this we substitute for -its value from equation (2) we get 


h = [(n' - 1) - 57.3* »' (1 + J)] R = about .027 kilometers. (19) 

For zenith distances up to i° the altitude of the zenith point is 
independent of p and £; or, for the same locality, R, and the same 
condition of the atmosphere, »', the zenith point has an invariable 



By reason of equation (15) the triangles bed and amd, fig. 5, 

are similar to each other, wherefore for the point of symmetry a 

we have 

, am , TT 2 h ^ 

be = cd or H 

dm R + h 

and substituting the values from equations (4) and (19) we get 

I" n' - 1 

L 57.3" 

H = 2 R \ — (l - A) - n' 

= about 189.0 km. . . . (20) 

as compared with 192.6 in equation (7). 

The agreement of these two values shows that the error made 
by equating the distances in equation (14) is without important 
influence on the result of the computations; that therefore in 
fact the distances of the points of symmetry from the boundaries 

of the atmosphere are nearly equal to each other when - = 57.3" 

and «' — 1 = 0.000282. 

But of the two formulae (7) and (20) for H the first is more exact 
because the curvature of the beam of light through the zenith de- 
parts but infinitesimally from a circular arc. 

As to the numerical determination, H = 192.6, the assumption 
that the ratio of the refraction to the zenith distance (57.3") is cor- 
rect to within 0.001 part of itself makes A < 0.001 and this would 
lead to an error of a few kilometers in the determination of the 
height of the atmosphere. 10 


On the basis of the preceding determinations it seems natural to 
attempt a new development of the differential equation (9) for 
astronomical refraction. 

The law of diminution of refractive power with altitude may with 
great probability be 

»' - 1 

P 1 

10 1 have recently found that the numerical determinations of - , W = 1 , 


and H really do need important corrections. 



which is deduced in a manner similar to that of Bunsen's law of 
absorption. After substituting this value then the solution of the 
differential equation offers no difficulty and the equation of the 
curve of the pencil of light is easy to find. 

As to the constant exponent m, that is best found from one 
accurate determination of the refraction. I will hereafter check 
the value of m thus found against the diminution of temperature 
with the altitude, since I hope to be put in possession of the neces- 
sary observational material through the kindness of a physicist, 
a relative, who expects to remain several years in the tropics. 

In accord with my previous efforts I also believe that I shall suc- 
ceed in obtaining from the observation of the twilight colors material 
for the direct demonstration of the diminution of the refracting 
power and the determination of the constants. 

Let cm in fig. 7 be the earth's radius and ca the horizon of the 
observer at c. At t' and /" hours after sunset the sun is at s' ami 

FIG. 7. 

s" \ the altitudes of the reflecting strata of air are a' c' = h' and 
a" c" = h" \ the corresponding angles of incidence and reflection 
are e' and e" . We easily find 




24 / v > 2 24 

(1 - sin e') R h" = (1 - sin e") R 


and according to equation (8) the intensity of the reflections from 
these layers h' and h" whose indices of refraction are n' and n" 
will be given by 

2 cos 2 e' J \ 2 cos 2 e" 

If by means of a good photometer we measure the intensities J' 
and J" , then the diminution of the refracting power between the 
two neighboring layers at the altitudes h,' h," h'" can be computed 
from the equations 

3 ri = 2 cos 2 *' Vj 7 , 
d n" 2 cos 2 e" V ]» . 




(Dated Hamburg, June, 1881) 

[Published (August 188 1) in Wiedemann's Annalen der Physik und Chemie 
New series, Vol. XIV, 1881, pp. 128-14.Q. 

Translated by Thomas Russell and C. Abbe] 

Although the view expressed by Hadley in the year 1735 as to 
the influence of the rotation of the earth on the currents of the 
atmosphere has become very well known, especially through Dove's 
writings, and has been treated of in all manuals of meteorology and 
physics, still the actual construction of the path of a particle of air 
has in general only seldom been carried out according to Had ley's 
principle. There are, however, in this very "Annalen" three art- 
icles 1 in which the problem of rigorously calculating the paths of 
the winds is proposed either under the definitely expressed or readily 
recognized assumption that the particles of air are to be considered 
as freely moving points or elementary particles of mass. The ques- 
tion treated in these articles is therefore a mechanical problem that 
can be formulated exactly, namely, the free motion (motion due to 
its inertia) of a material particle which is constrained to remain on 
a rotating surface. Since the year 1858 a number of mathematicians 
have busied themselves with this problem, and about the year 1861 
a general theorem was enunciated by Coriolis 2 by which every prob- 
lem of relative motion can be reduced to one of absolute motion. 
From these analytical investigations it evidently follows that the 
Hadlerian principle gives only very imperfect expression to the 
influence of the rotation of the earth on motions parallel to its sur- 

1 Von Baeyer: Pogg. Ann., 104, p. 377, 1858. 

Ohlert: Pogg. Ann., no, p. 234, i860. 

Mbusson: Pogg. Ann., 129, 652, 1866. 
8 Coriolis: Journ. de l'Ecole Polytechnique, XV, p. 142. 



face, so that the calculations based thereon in the above-mentioned 
three articles must necessarily lead to incorrect results. 

The oldest of these, however (by von Baeyer), appears of special 
interest as in it reflections are made which betray a tendency to 
abandon the Hadlerian motion. On p. 380 von Baeyer says: "a 
particle of air put in motion at a definite angle to the meridian on 
the surface of our spheroid of rotation at rest and continuing its 
way in the direction given to it without any hindrance or disturb- 
ance under the general influence of gravity would describe a short- 
est line. . . . Let us now imagine the terrestrial spheroid put 
in motion from its condition of rest, then the particle of air when set 
in motion in the direction a will already have this motion of rotation, 
it can therefore no longer describe a shortest line but its path will 
be the development of the shortest line on the spheroidal surface 
according to the circumstances of the rotation pertaining to it." 
It is to be regretted that these fruitful ideas were completely set 
aside in the course of the mathematical discussion in favor of the 
Hadlerian theory. When I first became acquainted, this year, with 
von Baeyer's article the above lines reminded me forcibly of a pro- 
cess I had made use of in the year 1879 to set forth the origin of the 
relative paths in simpler cases of a rotating system such as the earth 
presents and to base on it also a derivation of the equations of rela- 
tive motion. 3 

My treatment was simply, as follows : If a plane disk revolves 
uniformly, then will a body or material particle influenced by no 
forces whatever or by those whose direction is perpendicular to the 
disk, progress uniformly in a right line (that is when considered 
absolutely) and the relative path on the disk will be the continuous 
series of points which come in contact successively with the body. 
This conception, which evidently agrees essentially with that of 
General von Baeyer, can be extended to the rotating system of the 
spheroid which is under discussion, but at the same time it is evi- 
dent at a glance that the first part of the above quotation from von 
Baeyer's article contains an inaccuracy. The course of the point 
on the surface of a spheroid at rest can be a shortest or geodetic 
line described with constant velocity, only when the attractive force 
of the earth is everywhere perpendicular to the surface; in reality, 
however, it is the force of apparent gravity, that is to say, the result- 

3 Sprung: Studies concerning the wind and its relations to air pressure, 
Part I; On the mechanics of the motions of the air. From the archives of 
the Deutschen Seewarte (German Marine Observatory), No. 1, 1879. Zeit- 
schrift der Oesterreichisches Gesellschaft fur Meteorologie, XV, 1880. 


ant of the attractive force of the earth and the centrifugal force 
that is perpendicular to the earth's surface; the latter is moreover 
a level surface only by virtue of its rotation. If the earth should 
come to rest without a change of form, then the body would move 
parallel to the surface under the influence of a horizontal force 
directed towards the pole, which force is a component of the force 
of attraction and whose magnitude can be easily given. If we 
denote by <p the latitude of a point on the earth's surface (which 
here and in what follows will be considered in an entirely general 
manner as a body of revolution) by r the distance from the axis, 
and by u>, the angular velocity of rotation of the earth, then the 
acceleration 4 of the horizontal poleward directed component of the 
force of attraction is equal to the expression 

r or sin <p 

which represents the horizontal component of the centrifugal force 
directed toward the equator in the case of a point at relative rest 
on the rotating surface of the earth; for in fact the condition that 
these two horizontal forces are in equilibrium determines the form 
of the surface of the rotating earth. 

At the pole and at the equator this force has the value zero, it 
reaches its maximum at <p = 45 ; over the zone from the pole to 
45 latitude it acts in a manner similar to the action of the com- 
ponent of the force of gravity in the case of the spherical pendulum 
under the influence of which for angles of elevation between o° and 
90 the pendulum makes its vibrations. Hence in general the free 
absolute motion of a body which does not take part in the rotation 
of the earth's surface, but glides on it without friction will consist 
of uninterrupted oscillations around the pole as a center; if the 
original motion began in the direction of a meridian, then the body 
would never leave it ; if the body be started in the direction of a 
specific parallel of latitude (dependent on its velocity), then it would 
forever keep moving along this parallel with constant velocity, etc. 

It cannot be doubted that it is allowable in our consideration of 
the relative motion due to inertia on the rotating surface of the 
earth to begin in the above indicated manner with the consideration 
of the absolute motion; for since we do not consider the influence 
of the rotating surface as any other than that of a rigid opposing 
shell it can therefore be considered as infinitely thin and as closely 

4 In what follows, the expression "force" will for brevity always be used 
for the accelerating force. 


enclosing the similarly formed body of the earth at rest within it, 
while the physical forces at work (force of attraction) are perfectly 
independent of the condition of motion of the mass of the earth 
It therefore appears profitable to approach the above stated prob- 
lem of the absolute motion under the influence of the force r u? 
sin <p directed towards the pole and to treat it at least approximately 
just as the problem of the oscillations of a pendulum is solved par- 
ticularly in the case of infinitely small amplitudes. 

In the vicinity of the pole sin <p changes only very slowly, but r 
very rapidly; no great error will then be committed if we give to 
sin <p the limiting value i at the pole and neglect the corresponding 
component of the motion parallel to the earth's axis, that is to say, 
the motion is to be considered as taking place in a plane; the error 
is thus purely geometrical and easily estimated inasmuch as the 
forces arising from the special form of the surface are already taken 
into account. Adopting the coordinates x and y in a tangent plane 
and the origin at the pole the differential equations of motion are 
therefore as follows: 

d 2 x x , 

= — rur - = — xor 
dt 2 r 

d2 y 2 y 2 

= — rur- = — yar 



(Strictly speaking these equations apply to the absolute motion of 
a liquid particle parallel to the level bottom of a circular vessel 
revolving with the velocity co, in which the liquid is subjected to a 
fi irce perpendicular to the bottom surface — in so far as this abso- 
lute motion can be considered as entirely unimpeded.) 

Equations (i) agree perfectly with those on which is founded the 
theory of oscillations in an elastic medium; they can (for example 
by the substitution of x or y = e xt ) be integrated separately and 
lead to the final equations* 

x = a sin tot ) 


y = b cos wt j 

U r = cud cos cot 


U y = — bco sin tot J 

5 Compare, for example, the following treatises on physics: Wullner: 3d 
Edition, I, pp. 443 and 450; Mousson: 2d Edition, II, p. 531; Miiller: 7th 
Edition, I, pp. 278 and 281. There will also be found given in these places 
the geometrical representation to be spoken of presently. 


from which it is apparent that the point moves in an ellipse whose 
semi-axes are a and b; for from (2) we get the equation of the ellipse 

x\ 2 (y 

If Ave put 

+ {V 

<» = ~y W 

then T denotes the entire time of revolution of the point in the 
ellipse, for when t = T then both the coordinates and the com- 
ponents of the velocity U x and U y attain the same values they had 
at the time t = o. Since <o denotes the angular velocity of the 
rotating surface, equation (4) shows that for the absolute motion 
(elliptical) the time of revolution agrees with that of the rotating 

The absolute motion of the point can now be easily constructed. 
Let us choose for example the time of revolution of the surface (and 
of the material particle) as T — 24 seconds (compare fig. 1, page 
62) and divide the circumference of a circle constructed on the di- 
ameter 2a into 24 equal parts and from the points of division let fall 
perpendiculars on the diameter, which in this case can be done by 
joining the points in pairs by straight lines as the points of division 
are symmetrically distributed with respect to 2a. The 12 diameters 
which join the 24 points of division divide at the same time into 24 
equal parts a circle constructed on the small axis 26; from the points 
of division of this small circumference let fall normals to the di- 
ameter 26, which is perpendicular to 2a, and prolong them on both 
sides to the larger circle. If the time t is reckoned from the moment 
at which the body is at the extremity of the radius b, then the 
points of intersection of the two systems of normals which are marked 
o, 1, 2, 3 . . . lie on the elliptical path characterized by 
equations (2), (3) and (4) for T = 24, from which the correspond- 
ing relative motion is derived in the following manner. 

A rotation of the surface about an angle to, 2co, 30* . . . cor- 
responds to the absolute motion of the body up to the points 1, 2, 3, 
; evidently therefore we can find the positions at the mo- 
ment t = o of those points I, II, III ... of the rotating sur- 
face which will come in contact with the body after 1,2,3 
seconds, by going toward them in a direction contrary to the mo- 
tion of rotation of the surface along the concentric circles through 



VOL. 51 

1, 2, 3 . . . by the arcs r^o, 2t 2 u>, ^t 3 oj . . . respectively. 
From this follows in a striking manner the important result that 
the relative "inertia path" consists of a circle which will be de- 
scribed twice in the absolute time of revolution T and in such a way 
that the direction of the rotation is opposite to that of the rotation 
of the surface. 

FIG. 1 

Introducing 6 different modifications regarding the form of the 
elliptical path and the direction in which the body traverses t, 
we can exhibit clearly in a direct geometrical manner the mutual 
dependence of the absolute and relative motions and also, for exam- 

Fig. 1, for example, differs from fig. 2 only in the circumstance that 
the ellipse is traversed in the opposite direction; the relative velocity and 
circle of inertia are in consequence about 2% times as great as in fig. 1. 



pie, convince ourselves that for the same relative velocity v the 
circle has always the same magnitude whether it passes through the 
point of rotation M or at a greater distance from it. 7 There is no 
tendency on the part of the moving body to remain in a circle of 
latitude or to move parallel to the latitude circles as assumed in the 
theories of Hadley and Dove; for every azimuth of the motion the 

FIG. 2 

tendency is precisely the same, that is, to deviate toward the right 
from the momentary current direction of motion. 

Let us now try to follow the above construction analytically. 

7 By the aid of a ball of chalk rolling about on a rotating parabolic-shaped 
blackboard an autographic representation of these inertia paths can be 
reproduced; the unavoidable friction will only be manifest in this, that 
the curves (approximately circular) become gradually narrower and nar- 
rower. I recommend the following as suitable dimensions for this appa- 
ratus: diameter of the parabolic shell, 120 cm.; depth in the middle, 10 cm. 
In this case the time of revolution, T, must be 2.7 seconds. 


The relative motion will be referred to the coordinates (£, rj), moving 
with the rotating system and which at the instant t = o (as the 
figure shows) coincide with x, y. Denote by B the angular distance 
from the y-axis at the time t in the case of the absolute motion, and 
by /? the corresponding difference from the jj-axis in the case of 
relative motion, evidently then 

p = B - cot (5) 

Further if r denotes the radius vector at the time /: 



From (5) Ave derive 

r sin 8 = r sin B cos cot — r cos B sin cot ) 

.... (8) 
r cos /? = r cos B cos cot + r sin B sin cot J 

By substituting x, y from (2) in (6) ; r sin B and r cos B from (6) 
in (8) ; and finally r sin /? and r cos /? from (8) in (7) ; we get 

? = (a — b) sin cot cos cot 

f) = b cos 2 cot -{- a sin 2 w2 = a — (a — 6) cos 2 cot. 

But for this can be written, by application of known goniometrical 

f = sin 2 arf, 

cos2 cot -{ 

2 2 


J (a — b) is the radius of the desired circle; and £(a + b) is the dis- 
tance of its center m from the center of rotation M on the >;-axis. 

Instead now of starting with the construction of the absolute 
motion, as here done, we can also follow the reverse process, assum- 
ing the arbitrary relative velocity v in an arbitrary direction as 
being given for any distance b of the body from the center of revo- 
lution M. For simplicity it will however be for the present assumed 



that v is perpendicular to the radius vector b, and that v is reck- 
oned positive in the direction of rotation of the system (from west 
to east). Hence v + bco is the absolute west-easterly velocity of 
the body at the time / = o, for which according to (3) we have the 
expression (U x ) f= aw; we have therefore the relation: 


v n + be 


a-b = 12 



by using which relations the equations (9) finally pass into the fol- 
lowing form: 

£ = -12- sin 2 cot, 
2 co 

- V ° cos 2cot + (b + ^- 
2 co \ 2 co 


Therefore we have: 

the radius of the circle of inertia, 

b + - the n-coordinate of its center 


These equations give us all desired information concerning the 
relative motion due to the inertia of the body. We first derive 


= v cos 2 cot; 

—L=v n sin 2 cot. 

But from these we have 

•-■J 3; 



= v„ 

that is to say, the relative motion due to inertia is a uniform one 
and only distinguishable from the absolute motion due to inertia 
in free space, by this, that its path is not a straight line but curved. 
For the direction of rotation of the system assumed in our figure, 
which agrees with that of the northern hemisphere, the center of 
curvature always lies on the right-hand side of the path, since the 
co6rdinate i} m of the center of the circle will be < b as soon as v 


becomes negative, that is to say, when the original relative velocity 
is directed from east to west. 

For v = o, we have t] m = b, p = o and a = b, that is to say, 
the point remains at rest relatively while its absolute path is a circle. 

For v = — bco, we have t] m = + \b and a = o; the absolute 
path consists of a pendulous oscillation in a straight line. When 
a changes sign, that is to say, for still smaller values of the 
velocity v , the ellipses are described in the opposite direction. 

For v = —2baj, we have T) m = o, p = b, and a = —b; the center 
m of the relative inertia circle coincides with the center of rotation 
M, the absolute path is again a circle as for v Q = o, but the direction 
of the rotation is opposite to what it was before. 

The angular velocity 2co of the relative inertia motion is twice as 
great as that of the rotation of the system, the relative path will 
therefore be traversed twice during the time T of one revolution 
of the system. The figures i and 2 also show directly, as has 
already been indicated above, a time of rotation of 12 seconds, when 
24 seconds is assumed for the whole system. As now our investi- 
gation may be applied with a high degree of approximation to the 
region surrounding the north pole we attain at once the interesting 
result that a body confined to the earth's surface, but otherwise free 
to move, will deviate from its original direction twice as much as 
does the plane of the Foucault pendulum. 

The value of oj is 0.00007992 m, so that the length of the radius 
p becomes about 69 km. when the velocity is v = 10 m. or that of a 
fresh breeze; at this velocity therefore the. body only passes a little 
way beyond the space between two successive parallels of latitude. 8 

Now on a plane that is not in rotation a path can be produced, 

similar to the path of inertia found for the relative motion, by intro- 

ducmg a physical force A = - always acting from left to right, per- 


pellicular to the momentary direction of motion, But if the value 

of p given by (n') is substituted in this expression, then we have 

A = 2Vco (12) 

The motion on the rotating region near the pole can therefore be 
treated as an absolute one, if in addition to all the forces customarily 
taken into consideration there is introduced this other force -A, which 
in modern meteorology is called the "deflecting force of the earth's 
rotation. " 

Namely, the difference between 1 38 and 1 r 1 km. 


A current of air in the neighborhood of the north pole can only 
flow in a straight line no matter in what direction, when a force 
to the amount of 2Vco in the opposite direction to this deflecting 
force or directed from right to left renders this departure from the 
path of inertia possible; in this case therefore the barometric or 
elastic pressure in the current of air must increase from the left 
towards the right. In the same manner in a straight channel, no 
matter in what direction it trends, the moving liquid should stand a 
little higher on the right than on the left, and in fact independently 
of the nature of the liquid it results that we must have 9 

H — H 2vcu 

~^~ = T 

where H — H denotes the difference of height between the two 
shores of the stream, and L the width of the stream assumed to 
flow everywhere with uniform velocity. 

The special case of motion due to inertia in the region of the pole 
has been treated so fully in the foregoing text because the construc- 
tion of the paths in that region can be made on the basis of certain 

9 It will be proven presently that the above expression corresponds to 
the following equation for any latitude cp: 

g (H — H ) = 2vcu sin <f>.L 

For example, if for <p = 50 , iut sin <P has the value 0.0001117; for a width 
of river L = 100 meters and a velocity v = 10 meters there results H — H =» 
0.0114 meter. Hence in every horizontal layer the pressure of the water 
is by about 1.14 cm. of a column of water, greater toward the right than 
toward the left. This amount certainly appears to be inconsiderable, but 
geologists are accustomed to see very insignificant but constant causes 
produce great results. Hence it has been attempted by the axial rotation 
of the earth to account for such gradual displacement of river beds as is 
seen in the frequently recurring and notable phenomenon that the right 
side of a river is frequently closely bordered by a range of hills while on the 
left side a tolerably broad strip of entirely flat land stretches along the 
course of the river, as, for example, on the lower part of the courses of the 
Elbe, Weser, Thames, Seine and Gironde and also on the Danube, Volga, 
and other rivers of southern Russia where this is especially noticeable. 
But recently this explanation, proposed by von Baer, has been freely con- 

The relation between the direction of the wind and its force and the dis- 
tribution of atmospheric pressure which has found empirical expression in 
Buys-Ballot's law can be easily derived from the above text. Further 
details on this subject can be found in the works of Guldberg and Mohn 
in the Zeitschrift der osterreichisches Gesellschaft fur Meteorologie, 1877, 
Vol. XII, pp. 49, 177, 257 and 273; also Sprung. Ann. d. Hydr. u. maritime 
Meteorol., VIII, Jahrg. 1SS0, p. 603, and Beiblatter, V, 1881, p. 237. 


well-known theorems of physics. For other geographical latitudes 
the problem becomes considerably more difficult; however, even 
here with the aid of our conception of the process we are enabled 
to directly attain some results. 

The condition of relative rest of a body on a horizontal plane in 
any latitude 0, consists, absolutely considered, of a circular oscilla- 
tion (diurnal rotation about the axis of the earth) under the influ- 
ence of a poleward-directed component no 2 sin of the force of 
attraction of the earth which neutralizes the local equatorial tend- 
ency. The forces required in the absolute motion are evidently the 
same whether the circle of latitude is traversed from west to east or 
east to west with the velocity rio. In the latter case, however, the 
relative motion of the body is an east-west one with the relative 
velocity v = 2rco; the latter motion is therefore, just as in the case 
of relative rest, a special case of the relative inertia motion. Only 
in two very special cases of the relative velocity is it possible for a 
free body to remain on a circle of latitude, whilst it was formerly 
assumed that the final results of every deviation due to the rotation 
of the earth consists in a motion parallel to the circle of latitude. 

It can easily be seen that the horizontal radius of curvature of 
the small circle of latitude (whose curvature must always be deter- 
mined by comparison with the geodetic line which is a great circle 


on the sphere) is equal to the slope - — of that cone which is tangent 


to the earth's surface at the latitude 0. The "deflecting force of 

the earth's rotation" is therefore in this case (2rcu) 2 ( I and 

by using the above relation v = 2rco this can be written A = 2Vco 
sin 0. The "deflecting force" at the latitude <p (at least for the 
velocity v = 2roJ) is then smaller than at the pole, where the value 
is 2Vco; its direction is the same as there, perpendicular to the path, 
towards the right in the northern, towards the left in the southern 
hemisphere, and the influence of the earth's rotation is thus repre- 
sented perfectly, because the relative motion due to inertia here 
under consideration is a uniform one. 

For the completion and generalization of this result the general 
problem of absolute mi >ti< m under the influence of the force rco 2 sin 
directed poleward will be here treated briefly. 

Denote by V the absolute velocity parallel to the surface of the 
rotating body, by 6 the azimuth of the absolute motion (counted 
positive from the north around by the east towards the south) and 


by ds the differential of the path, then will the principle of living 
force, vis viva or mechanical energy be represented by the following 
equation : 

d (£ V 2 ) = no 2 sin <p ds cos 

But since, as is evident at once geometrically, — 3 — — . = sin <p, 

b ds cos a 

then the same equation can be written: 

dQV 2 ) = - rco 2 dr 

from the integration of which results 

V 2 =D - r 2 co 2 (13) 

where D is a constant. Again the principle of the conservation of 
areas gives 

Vsinfl = - (14) 


In these two equations the general problem is contained, and to a 
certain extent already solved. From (14) there is first derived 

y 2 cos 2 0= V*- — 
r 2 

and by introducing the value of V 2 from (13) 

V cos = J D - r 2 ur - — (15) 

The expressions (14) and (15) contain the west-easterly and south- 
northerly components of the absolute velocity as functions of the 
distance r from the axis; we have only to subtract from these the 
velocity of the surface of the earth at the place in question to obtain 
the corresponding components v sin d and v cos d of the relative 
velocity; in this way we obtain 

• /i C 

v sin o = — — rco 



C 2 

v cos d = il D — r 2 co 2 — — 



By squaring and adding these equations the following is obtained: 
v 2 = D - 2 Ceo = v 2 (17) 


Hence the velocity of the relative motion due to inertia is con- 
stant at any latitude whatever on the earth's surface, just as found 
before for the region of the pole. 

By a general theorem applicable to all rotating bodies, the radius 
p of the geodetic curvature of a curve running in any direction what- 
ever on the surface of a rotating body can be expressed by 

r cos 6 ds 

P = 

d (r sin 6) 

If in 

d (r sin 6) = r cos 6 d + sin d dr 

we substitute the two values derived from the first of equations (16), 
having regard to (17) : 

cos ddO = --(~ + CL 

v \r 2 


then we have 

sin = - ( — — ru> 

v cos 6 ds 


or, since 

cos 6 ds 1 

— dr sin <p 

2 io sin (p 


This value of the radius of curvature p of the relative path due 

1 v 

to inertia corresponds perfectly to the value p = - . — found above 

2 co 

(compare equation (11') for the region of the pole, and shows that 
the path is less slightly curved the more nearly the equator is 
approached. For the equator itself (<p = o) the path becomes the 
geodetic line or great circle itself. In the southern hemisphere <p 
is negative and therefore the radius of curvature has the opposite 
sign from that in the northern hemisphere. The center of curva- 
ture in the one case lies on the right side and in the other on the 
left side of the "inertia path." The proof of this statement is 


easily deduced by a closer consideration of the expression for sin d 
in the first of equations (16) ; if, for example, we introduce the con- 
dition that sin d = o for r = r , and write 

• a (u I r o 2 
sin = - ( — — r 

v \r 

If now we consider two places on the earth's surface at the same 
distance r from the axis, one of which is in the northern hemisphere, 
the other in the southern, then sin 6 is in both cases = o, that is to 
say, the motion is to be a purely south-northerly one. In the far- 
ther course of these motions, however, r becomes smaller in the 
northern hemisphere and therefore sin d > o ; on the contrary in the 
southern hemisphere r will become greater and therefore sin #<o; 
the body therefore deviates from the meridian towards the right 
in the northern hemisphere but towards the left in the southern 

If the motion is followed still farther (in the northern hemisphere 
for example) then we have 

for the value 6 = 90° the distance r £ = — + 


6 = 270° 

/ • ■< 

1 v \ 2 

r a 2 + 



r' + 


the value of 6 becomes 360 again for r = r and therefore in the 
same geographical latitude in which d = o. But it can be easily 
seen that in this case the body has not returned to the meridian of 
the starting place but to one lying farther west; for since the curva- 
ture of the path continually diminishes wdiile d passes through its 
values from 90 to 270 , therefore the southernmost point of the 
path must lie farther westward than the preceding northernmost 
point. The motion is therefore enclosed between two definite par- 
allels of latitude and carries the body in many nearly circular con- 
volutions gradually toward the west. Presumably this progression 
is connected with a peculiarity of the corresponding absolute motion 
which the latter has in common with a peculiarity of the spherical 
pendulum ; in this latter case it is known that the successive tempo- 
rary highest positions show a regular advance in a determinate direc- 
tion on a horizontal circle. 

The correct representation of the relative (or absolute) path in 
the form of an equation between the geographical coordinates <p and 



VOL. 51 

X ]>resupposes that the form of the rotating body is known [i. e. the 
slope of the surface of revolution] therefore that r is given as a 
definite function of the latitude or r = F. (<p) : for example, in the 
case of a sphere r = R cos <p: in the case of a spheroid 

R cos <p 


e- sin- o 


Since v sin d = r \ — ) , and v cos d = - 

1 (*\ 

sin <p (1c ' 





would become 




— r-oj 



= - 

sin <p 




r'co* - 




and from this by the elimination of dt is derived the definite inte- 

X = - 



) d<p 

r sm 

<p V Dr 1 - t 



in which the constants D and C from (16) can be expressed by the 
values of v, 6, and r or from (14) and (15) by the values of V, S 
and r for the initial circumstances of the motion as follows: 

D =v 2j r'2 r { ;or + 2 vr co sin 6 ( = V 2 + r Q 2 co 2 ) 
C = ?- 2 a> + vr sin ( = V r sin d ) 

. (21) 

The solution of this problem leading to elliptic integrals does 
not seem to be worth the while, because in the first place the func- 
tion r = F. {(f) for the earth can not be given with entire certainty, 
and in the second place the careful determination of the path has 
only a subordinate interest in meteorology, since the notion, for- 
merly entertained , that the particles of air act u ally follow the " inertia 
path" has been completely refuted by the synoptic weather charts 
that show the simultaneous conditions of the atmosphere over large 
regions. It may even be asserted that in fact the direction of curv- 
ature that pertains to the inertia path is not even the more frequent ; 


in fact there are many more curved wind paths that are cyclonal 
than anticy clonal. 

On the contrary it is of the greatest interest to know that the 
tendency towards change of direction by the rotation of the earth 
is far greater ani more general than was formerly supposed. The 
"deflecting force" acting from left to right, corresponding to equa- 
tion (12) can by substituting the value of p from (18) be written 

A = 2 vco sin <p (22) 

in which <p is to be taken positive for the northern hemisphere and 
negative for the southern. Therefore, for horizontal motions on 
the rotating surface of the earth the dynamical differential equations 
in a rectangular system of coordinates, for which the positive y- 
axis extends from the positive direction of the #-axis towards the 
left, are as follows: 

d 2 x v . dy 

-=A + 2wsinco — 

d? dt 

d 2 y dx 

= y — 2 &> sin cc — 

df dt 


In fluid motions the given forces generally consist of pressures, 
so that X and Y are to be replaced respectively by 

— -I — \ and —-(-—) (where a is the density). 
o\dx } o\dy I 

With reference to the application of these equations and their 
necessary extension to motion in any direction, reference may be 
made to the theoretical investigations in the domain of meteorology 10 
whose number has lately increased to a most encouraging extent. 
Since, however, the vertical forces can be readily ascertained a short 
discussion of these may properly follow here as a conclusion to the 
preceding presentation. 

For motions parallel to the earth's surface the vertical pressure 
N directed downwards, evidently has the value 

_V 2 = Q _V 2 cos 2 6 _ V 2 sin 2 6 
R R t R 2 

10 W. Ferrel: Meteorological researches; Reports of the Superintendent 
of the U. S. Coast and Geodetic Survey for 1875 and 1878. 

C. M. Guldberg et H. Mohn: Etudes sur les mouvements de 1' atmos- 
phere; programme de l'Universite - de Christiania pour 1876 et 1880. 

J. Finger: Wien, Sitzungsberichte, Jahrg. 1877, 1880. 


in which G is the vertical component of the accelerating force of 
attraction and R the radius of curvature of that normal section 
which is tangent to the direction of the absolute motion ; R t and R 2 , 
however, denote the radii of curvature of the principal normal 
sections of the rotating body, respectively parallel to the meridian 
and to the circle of latitude for which we have in fact: 

1 . do 

— = — sin ' 

R t dr 

1 _ COS ip 

R 2 ~ r 

For 8 must be introduced the azimuth of the relative motion; 
from (14), (15) and (16) we have 

V cos 8 =v cos 

V sin 8 = v sin 6 + rco 

From all this we obtain 

. v 2 cos 2 6 , v 2 sin 2 6 \ r 2 co 2 2 rvco sin d 
N =G - - + 

R t R 2 / R 2 R2 

or, by introducing in the last two terms the preceding expression 

for R 2 : 

v 2 

N = G — rco 2 cos <p — 2 vco cos ip sin d — (24) 


in which R' denotes the radius of curvature of the normal-section 
parallel to the direction of the relative motion. The first two terms 
represent the local acceleration g of the force of gravity 11 

G — rco 2 cos <p = g (25) 

If the motion of the body has a vertical component, then the 
same equation (24) will apply if the v therein is made to denote the 
velocity of the horizontal projection of the motion, and 6 its azi- 
muth. The magnitude of the force N will be changed slightly by 
the vertical component of the motion only when this latter motion 

d 2 h 
is not uniform and in fact the change corresponds to — which is 


11 The apparent force of gravity or the vertical component of the attrac- 
tion of the earth minus the vertical component of the centrifugal force. 

C. A. 


the expression for the vertical acceleration, The entire system 
of vertical forces must be considered as a modified force of gravity 
g, and therefore it must be introduced, instead of g alone, in the 
differential equation dp = —agdhoi the barohypsometric formula 
if it is desired to take the state of motion of the atmosphere into 
account in the derivation of the hypsometric formula. According 
to the equation for the gaseous condition the density a is dependent 
on the pressure p and the absolute temperature T in the following 
manner : 

p =oKT 

where K is the gas constant for atmospheric air. 

If the change with altitude in the composition of the air is left 
out of account and the decrease of the temperature upwards, in 
accordance with the usual custom, is assumed to be constant so 
that T = T — eh then there results finally the following equation 

p T -eh\^ * R' df/ 

(Strictly speaking g is also a function of the height h and the 
geographical latitude (p.) This equation is of great importance in 
meteorology inasmuch as it gives us the means of determining the 
horizontal distribution of pressure at any altitude h, in case this 
distribution is known for any other altitude (for example, at the 
mean level of the ocean where h = o), and provided sufficient 
initial points are given for the estimation of the condition of the 
atmosphere as to temperature. and motion. 

For the purpose of illustration and investigating briefly the mag- 
nitude of the influence of the horizontal motion of the air on the 
vertical distribution of pressure, it will be assumed that the tem- 

perature everywhere = T , therefore £ = o; since also — =o, there- 
fore by integration there results: 

T Q Kl P ° = (h - h \ ( g - 2 z/wcospsin - — f V 

For an atmosphere at rest we should have 

T Kl*> = (h -h )g (27) 



If now it is assumed that p (the atmospheric pressure at the upper 
level) has the same value in both equations, then by subtraction 
there results: 

T Kli±=(h - M(2 vco cos if sin + V " 7 J . 

If in this we replace h — //„ by its value from (27) and the ratio 

p B 

, etc., by the ratio ° etc., from the recorded barometric readings 
p B 

we have finally: 

B = ,b \1{ 2vCOCOS< o s * 6+ R')' 
B \B 

For example, let B = 62o mm denote the reading of the barom- 
eter on the Schneekopfe, B = 748 mm the corresponding reading 
of the barometer at Breslau, the difference of level being about 
i45o m ); also let v = 3o m , (the velocity of a violent wind storm), 
and (p = 51 ; then by computation the exponent on the right of 
equation (28) is found to be 

0.000 280 8 sin d -f 0.000 014 4. 

The extreme values of this exponent and the corresponding values 
of B at the level of Breslau are as follows: 

Exponent. B 

for = 90 (west wind); 0.0002952 747.958 
for 6 = 270 (east wind) ; —0.0002664 748.037 

From this it follows that under the same circumstances in other 

respects, the pressure on the lower side of a stratum of air 1450 111 

thick, moving with a horizontal velocity of 30 m. p. s. and having 

an equal pressure at the upper side in the two cases will with an 

east wind be about o.o79 mm higher than with a west wind. If 

v 2 
the term — had been neglected then for the west wind there would 


have resulted B = 747.960 and the difference between the west and 
east wind would have been o.o8o mm . The influence of this term 
is therefore very inconsiderable. 

Moreover the whole effect of the horizontal movements of the 
air must be called very insignificant because a change of pressure 
of 0.08""" ran scarcely be observed with our barometers. 


Since the horizontal forces conditioned on the axial rotation of the 
earth are of the same order as the vertical ones just considered 
(equation 22) and become equal to* them at the latitude 45 , there- 
fore the question arises, how comes it that the first are of such great 
importance in nareteorology ; the reason lies simply in this that much 
greater dimensions come into play in horizontal directions. It 
frequently happens indeed that the whole region between the Alps 
and southern Scandinavia is occupied by one and the same current 
of air, in which the difference of pressure on the two sides of the cur- 
rent (measured in a horizontal direction perpendicular to the iso- 
bars) amounts to 30 or 4o mm . From this there results a "Gra- 
dient" of 2.5 to 3-3 mm (for the unit length of one equatorial degree 
or in km.), whereas for a distance of ij km. (corresponding to the 
vertical distance above considered between Breslau and Schnee- 
kopfe) there. results a proportional difference of pressure of o.o3Q mm , 
a quantity that is no longer measurable with our barometers. 
Therefore if at about latitude 50 a parallelopiped of air extending 
from west to east of 1.5 km. height and breadth and previously at 
rest were set into a condition of stormy motion then the simul- 
taneous difference of pressure for the surfaces lying opposite each other 
in a horizontal as well as a vertical direction, must change by about 
o.o4 mm . Inversely the production and maintenance of such an 
insignificant difference of pressure would suffice to gradually bring 
about these same stormy motions; but the fundamental fact is that 
the horizontal difference of pressure suffices for this purpose and 
we should conceive the processes as going on toward completion in 
the following order: 

At the start the motion of the air takes place in the direction of 
the gradient, but this is departed from more and more with increas- 
ing velocity and diminishing acceleration of the wind, until the 
direction of its motion when it lias become uniform finally stands 
perpendicular to the direction of the gradient or at least approaches 
this perpendicular direction to a certain degree, because of the 
action of frictional resistances which render necessary the introduc- 
tion of a component of the gradient parallel to the motion of the air 
and directed forward with it. The change of the vertical differ- 
ence of pressure is to be considered as primarily a consequence of 
the circumstances of the motion since the effectof thelattercan here 
be conceived of as a mere diminution of the force of gravity. 12 

12 See the exponent of equation (28) where the influence of velocity opposes 
that of gravity. — Abbe. 


Another important question is that of the relation between the 
vertical distribution of pressure and the motion of the air in the 
vertical direction. If there are no motive forces present except 
the difference of pressure, then equation (26) is to be applied in this 
case and for v = o it becomes: 

r.r* --*(, + £) (29 , 

d 2 h 

It will be assumed that - = b is a constant quantity. Proceeding 

in a method entirely similar to the preceding we finally get 

Bo -f B <>\ g (30) 

B \B 

where B is the value of the barometer-reading B at the lower level 
under the condition of uniform motion throughout the whole mass. 
If, for example, it is asked how great b will become in the case of the 
values used above for B and B when the relation B X 748.1 = 
B X 748.0 exists (when therefore the difference of pressure is about 
o.i mm greater in the condition of accelerated motion than in that 
of rest or uniform motion), the result is, there will be an upward 
directed acceleration of b = 0.007 meter per second. 

If the air has simultaneously an east-west component of velocity 
to the amount of 25 to 30 m.p.s., then the diminution of the pres- 
sure from below upwards will become about 0.10 4- 0.04 = o.i4 mm 
greater than in the condition of rest. 

If in any manner whatever an increase of the vertical difference 

icssure of o.i mm should be brought about and maintained, 

for example, b)^ an upward directed removal of air at any altitude, 

then an ascension of the air must take place and by integration of 

d 2 h 
the above equation — = b for uniformly accelerated motion, the 

velocity which a particle of air attains in passing over a distance 
h — h of 1$ km., can be deduced. As b is assumed to be con- 
stant there results: 

dh [ 

2b(h-h ) 
at * 

For the above value of b = 0.007 an d f° r h~ K = I 5°° m we obtain 
dh = 4.58 111 per second. In this vertical motion, as is known, 
there appears again a horizontal component of motion in conse- 


quence of the rotation of the earth, in case such motion is not pre- 
vented by differences of pressure. In an ascending motion the 
tendency to deviate toward the wefet is represented by the expres- 
sion 2( — ) co cos <p as can be very readily proved by the aid of the 


principle of the preservation of areas. 



School of Technology, Dresden, Germany 
[Translated from Ann. d. Physik, (^) Vol. IV, pp. ^.^g-^So. igoi] 
Reprinted from the Monthly Weather Review for April, kjoi 

It is a Avell known principle of climatology that the side of a 
mountain range which is turned toward the prevailing wind lias in 
general a greater precipitation than the plain on the windward 
side, and a still greater in comparison with the leeward side of the 
mountain range. There has been no doubt as to the explanation 
of this phenomenon since it has been recognized that the principal 
cause of the condensation of the aqueous vapor is the adiabatic 
cooling of the rising mass of air; for a current of air impinging against 
rising ground must, in order to pass over it, necessarily rise. So 
far as the author knows, however, no attempt has yet been made to 
investigate this process quantitatively, except perhaps, for the 
stratum of air immediately contiguous to the earth, whose ascension 
being equal to that of the surface itself, is thereby known directly. 
Such a quantitative treatment will be attempted in the following 
article. Even although this is only possible under special assump- 
tions, which represent nature with the closest approximation, it will, 
however, always offer a practical basis for estimating the purely 
mechanical influence exerted by the configuration of the surface 
of the earth on the formation of rain. 

In order to find the standard vertical components of the velocity 
of the air currents that determine the condensation, Ave must, first 
of all, solve the hydrodynamic problem of the movement of the air 
over a rigid surface of a given shape. In this connection Ave must 
make a series of simplifying assumptions, as follows: 

i. The current must be steady; 2, it must be continuous and 
free from whirls; 3, it must Aoav everywhere parallel to a definite 



vertical plane, and consequently depend only on the vertical co- 
ordinate (y), and one horizontal coordinate (x) ; 4, the internal fric- 
tion, as well as the external (or that due to the earth's surface), 
may be neglected; 5, at great heights there must prevail a purely 
horizontal current of constant velocity (a). As to the configura- 
tion of the ground, we must, corresponding to proposition 3, assume 
that the profile curves are identical in all vertical planes that are 
parallel to the plane of xy; 6, and further, we assume the surface 
profile to be periodic, that is to say, the surface of the earth is formed 
of similar parallel waves of mountains without, however, deter- 
mining in advance the special equation of the profile curves. 

If we designate by u and v the horizontal and vertical components 
of velocity and by e the density, then, in consequence of assumptions 
1 and 3, there follows the condition 

9 (e u) d (e v) = Q 
dx dy 

and in consequence of 2 there must exist a velocity potential, <p, 
which, according to 3, can only depend upon x and y, so that 

u=% z, = ^and ±( 9 *\ + ±( .?*) -0. 

dx dy dx \ dx J dy \ dy / 

If we consider that the density of the air (excluding large differ- 
ences of temperature at the same level) changes much more slowly 
in a horizontal than in a vertical direction, then we can regard e as 
a function of y only, and obtain for <p the differential equation — 

eJ<p=- --L- (1) 

dy dy 

The law of the diminution of density with altitude will, strictly 
speaking, be different for each particular case, because the vertical 
diminution of temperature in a rising current of air, which deter- 
mines the rate of diminution of density, depends upon the conden- 
sation. But it is allowable, as a close approximation and as is 
usually done in barometric hypsometry, to assume the law of dimi- 
nution of pressure which obtains, strictly speaking, for a constant 
temperature only, and which, as is well known, reads as follows: 


nat log = q y, 


where q is a constant and has very nearly the value of 1/8000 if 
v, the difference in altitude, be expressed in meters. In this case the 
following also holds good: 


and, consequently, 

1 de 

e dy 

hence the differential equation for y> becomes 

'*-«■£ (2) 


A solution of this differential equation that satisfies the assump- 
tions 5 and 6, is given by the expression 

<p = a (x — b cos m x. e~ ny ) (3) 

in which the following relation exists between the constants m and 

m 2 _ n 2 _ g n . 

n = — - + r, where r = V m 2 I o 2 4. I 

2 j 

In order to ascertain what profile or configuration of the ground 
corresponds to the current determined by this velocity potential, 
we must look for the lines of flow; for one of these must certainly 
agree with the profile curve. The differential equation of the stream 
lines reads as follows: 

dy.dx = ' : ' = abn cos mx.e~ n " : a (1 + bm sin nix e " !l \ 
dy ox 

The integration of this equation gives 


e~ ny . sin mx = - + Be q,J (5) 


wherein B represents the parameter of the stream lines. 

If we assume that the curve of the profile of the surface passes 
through the points x = o and y = o, then for these values B = 


m/bqn, and if its ordinates are designated by 7), its equation becomes 

or , 

n .„ **' - *" 

sin m«.« 
m </ 

As long as 77 remains so small that for both the highest and lowest 
points of the profile of the surface of the earth (q 77/ 2) 2 is negligible 
in comparison with unity — which is practically always the case 
for the mountains that come under our consideration — we • can 

= b~~ sin mx.e~ rr> \ 

i + " 

r = V'm 2 -f ^ 2 /4 


In these expressions b and m appear as parameters that can be 
chosen at will, the first of which determines the altitudes and the 
second the horizontal distances between the mountain ridges; we 
have, namely, m = 2iz\ A, if X denotes the wave length, that is to 
say the distance between two corresponding points, as for example 
the summits of neighboring mountain ranges. 

It is easy to show that the stream line determined by the velocity 
potential (3) for the configuration of the ground given by the trans- 
cendental equation (5') is the only one compatible with the general 
conditions 1 to 5. Moreover, since a potential current is determined 

single valued, for the interior, by the value of — along the bound- 
ary of a closed region, therefore, our solution in case it gives hori- 
zontal velocities that are constant, or slowly diminish with the 
altitude above the center of the valley, is also applicable to the 
specially interesting practical case in which only one single moun- 
tain range rises above an extended plain and is struck perpendicu- 
larly by a uniform horizontal current of air. To what extent this 
holds good must be established in each special case. 

The horizontal and the vertically upward velocity components 
corresponding to our solution are: 

u = a (1 + bm sin mx .e~ ny ) 0) 

v = abn cos mx.e~ ny (7) 


It would now be desirable, in order to be able to handle the cases 
actually occurring in nature, to adapt our solution to some form 
of the earth's surface arbitrarily chosen. The first thought would 
be to attempt this by the superposition of a series of velocity poten- 
tials of the form of equation (3) having different constants m and 
b, or in other words to write 

(p =^<f h = a \x -^\b h cos m h x.e- n hy\ . . . . (8) 

but we find that this solution only corresponds to a superposition 
of the profile curves, that is to say, it gives 

V = V Vh = V K — sin m h x.e~ r M (9) 

only when we can put the exponential functions e^ffland c^hVboth. 
equal to unity. In this case 77 is at once transformed into the 
simple trigonometrical series 

v =^\b h 

m h 

sin m h x (9') 

and therefore, by putting m h = h m l we can develop any arbitrary 
function, r) = f (x), into a series, proceeding for any value of x 
greater than zero and less than XI 2. But the condition that e± hmr l 
is equal to unity for any large value of the quantity h will not be 
fulfilled for any arbitrary form of the profile curve if its maximum 
altitude is assumed to be very small in comparison with the wave 
length X. Therefore, we must limit ourselves to an approximate 
representation of the desired profile curve by a definite number of 
terms of the series that enters equations (9) or (9'). Especially can 
we in this way never attain the rigid solution for a ground profile 
that has sharp angles. However, the neglected higher terms of the 
series have a proportionately smaller influence on the vertical vel- 
ocity at great altitudes and, therefore, on the resulting precipitation, 
in proportion as their serial number h is larger. 


As a first example, we choose a form of profile to correspond as 
closely as possible to a plane, broad, valley and a plateau-like moun- 
tain range, because, in this case, we may expect nearly the same 
conditions on the slope of the mountain as if it were struck by a 


uniform horizontal current of air. A profile curve of this kind, 
which rises steadily between the values x greater than — — 'and 


less than + — - and falls also with uniform gradient between the 

limits x = 5/12A and x= 7/1 2a, and in the intermediate region 

describes a horizontal straight line at the distance + H from the 

axis of Xij is obtained by means of the Fourier series 

24# ^-a 1 . hn . 2hit 

7] = > h — Sln • Sln X, 

1 ^ ^ h? 6 X 

where h has all positive uneven numbers. In order to represent a 
profile curve of the given form approximately, we take the first 
three terms of the series, and therefore have 

7? = C { % sin m l x + j sin 3 m x x + ^ sin 5 m t x\ . . . (10) 
The numerical values of the parameters are: 

X =60000 meters, hence m^ - — = . 1047 X 10 ~ s 



C = 1100 meters. 

The coefficients b h , in the expressions (8) and (9) therefore, have 
the following values : 

b, = 881, 63 = 148.3, b s = 24.8 

The profile given by equation (10) is shown in fig. 1, where the 
vertical scale is magnified five times. We perceive that the ascend- 
ing gradient is nearly all confined to the interval between 

x greater than — and less than + — 
12 12 

where, moreover, it is quite uniform, and further that the surface 
of the valley is raised a little in the center, and the surface of the 
plateau mountain is depressed by the same amount. The differ- 
ence in altitude between the center of the valley and the center of 
the mountain, which according to the adopted numerical values 
should be 900 meters, is, therefore, not the absolute maximum 
difference but is about 18 meters less. The profile curve here con- 


sidered corresponds indeed, according to what has been above said 
only approximately to the velocity potential. 

(p = a { x — 6j cos m l x.e~ n i v — b 3 cos 3 m x x .e~ n , y 

— b 5 cos 5 m v x . e~ n i V j , j 

■ .(11) 

as determined by the above coefficients, b h , but we can easily 
demonstrate that in the present example the differences could 
scarcely be observed in fig. i. 


From the preceding value of <p we derive the following values for 
the components of the velocities of the current : 

u = a I 1 + ^ b h m h e " y sin m h x \ 

{ 2 k 

= a < 1 + -_ (b x e V sin m, x + Sb 3 e V sin 3 m, x 

+ 5 6 5 e~ n * y sin 5 ra t x) > , 

> . 


v = a X ^P 6,, « A e n * v cos w A a? 
= a X .1152 > ^ e~V cos m, x + % e~ n * v cos 3 m x x 

i — » 1/ 

JL_ *? A 

+ To * * COS ^ m i X 




These equations show that when x — o, that is to say above the 
center of the slope of the mountain, u is a constant = a at all alti- 
tudes; above the valley where x is* less than o, u is smaller than a; 
and above the mountain, or plateau, where x is greater than o, 
u is larger than, a; the constant a can also be considered the mean 
horizontal velocity at any given altitude. 

For different altitudes H above the center of the valley we have 
the following values: 

H = 45o + y: 




u — a _ 

a ' 


— 0.0676 

— 0.067S 

— 0.0646 

Therefore, up to the altitude of 5000 meters, the horizontal 
velocity is sensibly constant and the vertical velocity o ; and, accord- 
ing to what is said in reference to equation (5') our solution holds 
good for the case when the profile is continued as a horizontal 
straight line indefinitely toward the negative side from the point 
x = — A/ 4, and above this there flows a truly horizontal current of 
air whose velocity is sensibly constant, namely, 0.93 a up to an alti- 
tude of 5000 meters and increases in the strata above that until 
it attains the value a. 

Above the mountain, as at the point where x = + i/4, the velo- 
cities, u, are greater than a by nearly as much as they are smaller 
above the valley. 

The distribution of the vertical velocity component which deter- 
mines the condensation of aqueous vapor is a more complicated 
matter. In order to represent it, let the values of v/a for different 
values of the coordinates x and y be as given in the following table: 



+ 1 



+ - 1 









— 0.0012 





+ 0.00226 
















Therefore, whereas there is a steady decrease of v with altitude 
above the center of the slope of the mountain, on the other hand 
these vertical velocities increase with the altitude in the neighbor- 
hood of the foot of the mountain as well as on the plateau at the 


point x = ± A./8 up to a maximum at some very great altitude. 
(The isolated negative value that occurs for x — X/6 and y = 500 
is explained by the above-mentioned slight depression of the sum- 
mit of the plateau mountain.) 

In order now, to investigate the condensation of aqueous vapor 
that occurs in consequence of the ascending currents of air forced 
upward by the upward slope of the ground, we first make the assump- 
tion that the ascending mass of air experiences an adiabatic change 
of condition and that adiabatic equilibrium prevailed already in 
the horizontal current of air advancing toward the slope of the 
mountain. In this case the air will be everywhere saturated at a 
certain altitude that can be computed from the temperature and 
humidity of the air at the surface of the valley. In a unit of time 
the quantity of air, ve, flows in a vertical direction through a space 
having a unit of horizontal surface and an altitude dy. If this 
element of space lies above the lower limit of the clouds, then in 
this quantity of air there will be as much aqueous vapor condensed 
as the difference between what it can contain in the state of satur- 
ation at the altitude y + dy and what it can contain at the altitude 
y. Therefore this quantity is 


vs. — ay, 

where F (y) is the specific humidity of saturated air at the altitude 


Still assuming a stationary condition, we have — 



= - jveF' (y)dy, (14) 

as representing the total quantity of aqueous vapor condensed in a 
unit of time in a stratum of cloud above the unit of basal area be- 
tween the altitudes y and y' . 

This would also be equal to the quantity of precipitation falling 
from that layer of cloud on to the unit of horizontal base in case the 
products of condensation simply fell vertically without being car- 
ried along by the horizontal current of air. We will make this 
assumption, since as yet we have no clue by which to frame a com- 
putation of the horizontal transportation of the falling particles of 
precipitation. It is, however, easy to foresee that the horizontal 
transportation would be of importance, especially for the slowly- 
falling particles of water or ice in the upper strata of clouds, and 


that on the other hand, the larger drops that carry down with them- 
selves the water condensed in the lower strata of clouds will fall 
at a relatively slight horizontal distance. But now, as the numer- 
ical computation sKows, the lower cloud strata contribute relatively 
far more to the condensation than the upper clouds; therefore, the 
influence of the horizontal transport will not be so very large, at 
least with moderate winds. Moreover, this influence does not affect 
the total quantity of precipitation caused by the flow up the moun- 
tain side, but only its distribution on the mountain slope and it 
consists essentially in a transfer of the location of maximum pre- 
cipitation toward the mountain. In this sense, therefore, we have 
to expect a departure of the actual distribution of precipitation 
from that which is theoretically given by the computation of W 
as a function of x, according to equation (14). This departure will, 
under otherwise similar circumstances, be considerably larger in the 
case of snowfall than in the case of rain. 

As concerns the upper limit y' , which is to be assumed in the 
integration of equation (14) in order to obtain the total quantity 
of precipitation falling upon a unit of surface, we have to substitute 
for y' that altitude at which condensation actually ceases in the 
ascending current of air. Theoretically, if from the beginning 
adiabatic equilibrium prevails up to any given altitude, then the 
condensation brought about by the rising of the earth's surface 
must also extend indefinitely high, even to the limit of the atmos- 
phere, since the vertical component of velocity diminishes asymptot- 
ically toward zero. But practically, our solution of the problem 
of flow probably no longer holds good for very high strata, and cer- 
tainly the assumption of adiabatic equilibrium does not hold good ; 
but even if the latter were the case, if therefore, the ascending cur- 
rent carried masses of air from the surface of the earth up to any 
given altitude, still, in consequence of the increasing w r eight of the 
particles of precipitation carried up by the ascending current on the 
one hand, and the increasing insolation on the other hand, an upper 
limit of cloud must be formed 2 

We will therefore assume as given some such upper limit of clouds 
at a definite altitude, and for simplicity will assume this to be the 
same everywhere. The value of this altitude, y\ is the upper limit 
of the integral (14). However, the altitude assumed for y' if it is 
large, namely, many thousands of meters, can have only a slight 

2 W. von Bezold: Sitzb. Ber. Akad. Wiss., Berlin, p. 518, 1888, and p. 303, 


influence on the value of W, since both — F'(y) and ve rapidly 
diminish with the altitude. 

For the numerical computation of W, it is advantageous to first 
bring the expression (14) by partial integration into the following 


W{x) =| veFiy) P + P F (y)—cty (14a) 


In this expression v is given by equation (13) as a function of y 
and x. F(y), or the saturation value of the specific moisture at the 
altitude y, as well as the corresponding values of the pressure and 
temperature necessary for the computation of e are most easily 
obtained with the help of the graphic diagram for the adiabatic 
changes of condition of moist air first given by H. Hertz, since 
a simple analytical expression for these quantities cannot be given. 
In using the Hertzian table 3 we have to remember that y is not the 
absolute altitude but the altitude above the axis of x in our system 
of coordinates, therefore, in order to obtain the altitude above sea 

level, it is still to be increased by the quantity — >jl x = — "7 ) and 

also by the altitude of the valley above the sea. The integral in 
equation (14a) can be evaluated with sufficient accuracy by divid- 
ing the integral from y Q to y' into parts y 9 ...y v y^—y^ y^-i ■•■■ v h 
(where y h = y'), and for each of these introducing an average 
value F mk whereby we obtain equation (15). 

Vh h 


(ev) k - M*-i 


In order to execute the complete computation of W for a special 
example, we will assume that the current of air which strikes the 
mountain having the profile shown in fig. 1 has a pressure of 760 
millimeters, temperature 20 , and specific humidity, 9.0, 4 at the 
bottom of the valley. Hence, according to our assumption of adia- 
batic equilibrium it follows that the lower limit of the clouds will 
lie at an altitude of 950 meters above the bottom of the valley, and, 
therefore, 50 meters above the center of the mountain, if v = 500; 

3 H. Hertz: Met. Zeit., Vol. I, pp. 421-431, 1884, or the preceding collec- 
tion of translations, 1891, p. 198. 

4 That is, 9.0 grams of water per kilogram of air. 


the specific humidity is at this cloud level, F {y)' = 9.0, and the 
temperature is n° C. We will further assume that the upper limit 
of the clouds is at an altitude of about 5000 meters, or y' — 4530 
meters, where the temperature has sunk to — 13. 6° and the specific 
humidity to &{y) =2.5. At the altitude of 3000 meters the tem- 
perature o° C. is attained. The application of the Hertzian tables 
assumes that for temperatures below o° C. the product of condensa- 
tion is ice; whether this is really true is at least questionable for 
moderately low temperatures, but the assumption that water below 
the freezing point is precipitated will not change the results very 
much. Since corresponding to the assumed stationary condition, 
we have to assume that all condensed water immediately falls from 
the clouds; therefore, in our computation we have to omit the hail 
stage of Hertz, in which the water that is carried along with the 
cloud is frozen. 5 

For the computation of the integral according to equation (15) 
the cloud is divided into four layers whose mutual boundaries or 
limits occur at y l = 1530, again y 2 = 2440, and y 3 = 3460 meters; 
for these altitudes we have s = 1.00 and 0.912 and 0.816, and cor- 
responding to these F(y) =6.9 and 5.35 and 3.8. 

We thus find the following values for W/a: 


x = ± — ± - ± - 

12 8 6 


— = 0.475 0.241 0.0985 0.0081 grams per second per 

square meter. 
From this table we obtain the depth of the precipitation in milli- 
meters per hour by multiplying by 3.6; the result is shown in the 
lower curve of fig. 1. The values of the precipitation for a mean 
horizontal velocity of the current of 1 meter per second are as fol- 


x = ± — ± ± — ±— ±- 

24 12 8 6 4 

W = 1.71 1.47 0.867 0.355 0.029 

Hence, the precipitation is heaviest above the middle of this slope 
of the mountain, where for the very moderate wind velocity of 7 
meters per second, it attains the very considerable rate of 12 milli- 

5 The influence upon the adiabatics of condensation, whether we assume, 
as in the Hertzian table, all condensed water to be carried with it or to imme- 
diately fall away, is of no importance in the present problem. 


meters per hour. In this connection it is, indeed, to be remembered 
that we have assumed exceptionally favorable conditions for the 
precipitation in that we have assumed the onflowing air to have been 
already fully saturated throughout the whole 4000 meters in depth 
of the layer between y and y.' 

The comparison of the curve of precipitation with the curve of 
profile in fig. 1 shows that although the maximum of precipitation 
coincides with the maximum gradient of the slope of the mountain, 
yet the depth of precipitation diminishes more slowly toward the 
plane of the valley and the plateau of the mountain than does the 
slope of the earth's surface; thus, for instance, the latter slope at 
the point where % — ± A/12, and which is given by drj/d x, amounts 
only to 1/20 of the maximum slope, while the precipitation at this 
point is more than 1/5 of its maximum value. Therefore, under the 
conditions here assumed, the effect of a mountain slope in producing 
precipitation makes itself felt in the plain lying in front of the foot 
of the slope. All of which agrees with actual experience. 8 The 
fact that in reality the maximum precipitation appears to be pushed 
more toward the ridge of the mountain is certainly partly explained, 
as well as suggested, by the horizontal transportation of the pro- 
ducts of condensation in the clouds, but also in part by the departure 
of the real distribution of temperature and moisture from that here 
assumed. (See Section IV, page 95.) 

The determination of the total quantity of precipitation caused 
by the mountain slope will be attained if we integrate the value 
of W as determined by equation (14) as a function of x between the 
limits x = — A/ 4 and x = + A/4. The result is, therefore, 

. x x 

"*" T y' + 4 

= \W(x)dx = - \eF'(y) Cvdx. . . .(16) 

\W(x)dx = - f sF' (y) Cvd 

Vo X 

In this equation, according to equation (13) we have: 

+ 4 

C vdx = a X 1100 i e~\ v - 2 e~\ v + * e~\ v 
J I 9 25 

X 1 

~ 1 

•Harm: Climatology, 2d Edition, Vol. I, p. 295; also Assmann, Einfluss 
der Gebirge auf das Klimat von Mittel Deutschland, p. 373, 1886. 


For our present example we find G = 5100a grams per second 
over a strip 1 meter wide and about 22 kilometers long. Hence, 
there follows for the average precipitation for the whole mountain 

W m ' = 0.833a millimeters per hour. 


In the example Ave have just discussed the lower limit of the clouds 
was higher than the summit of the mountain. If the reverse is the 
case, then, for that portion of the mountain slope that is immersed 
in the clouds we must take >j instead of y as the lower limit of the 
integral in the formulas (14) to (16): therefore, the theoretical dis- 
tribution of precipitation would no longer be symmetrical with 
respect to the zero point on the axis of abscissas. As an example 
of this case we will consider the flow of air above the ground profile 
that is represented by the simple equation 

r) = C sin m x.e~ r1) 

As to the constants we will adopt the following: 

C =* 1000 meters, X = 24000 meters; 

hence m = 0.262 X 10" 3 , r = 0.269 X 10" 3 , 

and for the vertical coordinate rj we find from equation (5) 

tor x = — - — - — 0+ + - + - 

4 6 12 12 6 4 

7) = - 1495 - 1194 - 585 0+444 + 715 + 805 meters. 

The resulting curve is shown in fig. 2. The altitude of the sum- 
mit of the mountain above the plain of the valley amounts to 2300 
meters. The valley may be 100 meters above sea level; the atmos- 
pheric pressure in the valley is assumed at 750 millimeters, the tem- 
perature 23 , and the specific humidity 10 grams of water per kilo- 
gram of air. From the Hertzian table we find the lower cloud limit 
at the altitude of 1220 meters, that is to say at y = — 375. The 
upper limit of the clouds is assumed at y' = 2400 and, therefore, at 
4000 meters above sea level. Therefore, for that portion of the 
clouds lying below the summit of the mountain, which is limited 
to the negative values of the abscissas up to % = — 1.35 kilometers 
approximately, since according to equation (7) 

v = Cam cos m x.e~ ny 


we have: 

VOL. 51 

w = 

= — \&vF'dy = — aCm cos m x I e F' (y) e -ny d y 


= a cos mx X 109. 


Therefore, the depth of the precipitation will here be represented 
by a simple cosine curve and, in general, corresponds to the slope 
of the mountain, which is computed from equation (5') by the 
expression : 

d rj Cm cos mx.e~ rri 
dx 1 -f Cr sin mx.e~ r 

For the region lying above the lower cloud limit y the value of 
W(x) cannot be represented by a simple function of x. We find 

X*-6 5 -4 

*Z +J <•£ +5 *-6Jc;i. 

FIG. 2 

the precipitation in millimeters per hour for a horizontal velocity 
a — 1, as follows: 

For* = -6 -5 -4 -3 -2 \ 

W = 1.01 1.96 2.78 3.40 / 

F,,r a- = -1 0+2 +4 +6 \ 

W"=3.50 2.94 1.95 0.88 f 

Beli >w the cloud. 
the cloud. 

The distribution of precipitation, as given by these figures is 
sli own in fig. 2 by the curve of dashes. The curve of dots repre- 
sents the symmetrical line that would obtain if the mountain were 
not immersed in the clouds. The location of maximum precipita- 
tion is 3.93 for x = o and is 3.68 for x = — 1.3. 


The total quantity of precipitation is computed by the formula: 


G = — a C sin mx \ e F' (y) e ny dy 

and is approximately equal to 22730; this is distributed over a hor- 
izontal strip 12000 meters in length, and therefore, for a uniform 
distribution for a = 1 the precipitation averages 1.9 millimeters. 
From the preceding expression for G, it is plain that for any given 
altitude of the mountain summit G will be smaller the shorter and 
steeper the slope becomes, that is to say, the smaller the value of k 
is, since the exponent ny increases with diminishing values of X. 
In the present case the horizontal velocity of the wind is given by 
the expression: 

3 cp I f}fi 

u = ' = a I 1 + C sin m x.e~ ny 
dx \ n 

= a(l + 0.332 sin mx.e- nj/ ); 

which attains its minimum, = 0.547a, at the bottom of the valley, 
and its maximum, 1.283 a, at the summit of the mountain, and has 
a for the mean value of all the horizontal planes. Above the center 
of the valley, it increases gradually with altitude, asymptotically 
approaching its limiting value, a; for example, at the level y — o, it 
is equal to 0.668a, and at the level y = 2400 it is already equal 
to 0.80a. Therefore, if the stream under consideration proceeds 
from a point x = — A/ 4, as a purely horizontal current of air flow- 
ing over a plain, then its velocity must diminish with the altitude 
in the ratio e~ ny . This would, of itself, be a plausible assumption, 
but there would then be a vortex motion for each horizontal cur- 
rent of air, which cannot, strictly speaking, continue steadily in 
the above assumed potential motion. 


The assumptions hitherto made by us, namely, that the distribu- 
tion of temperature in the current of air that impinges upon the 
mountain side already corresponds to the condition of indifferent 
equilibrium, that is to say that it is the same as would occur in an 
ascending current of air under adiabatic changes of condition, is in 
general not actually fulfilled. The scientific balloon ascensions at 
Berlin have recently given us reliable conclusions as to the real con- 
ditions of temperature and moisture in the free atmosphere up to 


altitudes of 8000 meters. The mean values of the temperature 
and moisture at successive levels, 500 meters apart, which von 
Bezold has deduced 7 from the observations of Berson and Suring 
show that the mean vertical diminution of temperature is slower 
than the adiabatic, and that, in general, the moisture does not attain 
the saturation value. In a horizontal current of air, in which these 
average conditions prevail, the air will, therefore, never be satu- 
rated, and, consequently, our assumption of the existence of a con- 
stant lower limit to the clouds is not allowable. Moreover, it is no 
longer the vertical component alone that controls the condensation 
that shall occur at any given point in the current of air ascending 
above the mountain slope, as was assumed in the derivation of 
formula (14). We must rather, in the computation of W, consider 
that the quantity of water condensed in a unit of space under steady 
stationary conditions is equal to the excess of the quantity of water 
vapor flowing into the space above that simultaneously flowing out. 
For one cubic meter and one second this excess is: 

d (euF) d (ev F) 
d x dy 

or since because of the equation of continuity we have approxi- 

therefore, 8 

and hence, 

de u de v _ 

o , 
d x d y 

8F dF 

— e ( u - + v 

d x dy 

W = - C s (u dF -+v d —) dy (17) 

J \ dx dy J 


where y° and y' indicate the altitudes of the limits of the clouds 
above the point under consideration. The evaluation of the inte- 
gral still demands not only a complete knowledge of the stream, but 

7 W. von Bezold: Theoretische^ Betrachtungen, etc. Theoretical consid- 
erations relative to the results of the scientific balloon ascensions of the 
German Association for the Promotion of Aeronautics at Berlin. Brunswick, 
1900, pp. 18-21. (See No. XIV of this present collection). 

8 In so far, namely, as the quantity of the aqueous vapor condensed in a 
unit of volume is inappreciably small in comparison with the total quantity 
of moist air flowing through this space. 



also the determination of the cloudy region, that is to say, that 
region in which the atmosphere is saturated and the distribution of 
temperature therein, since the latter first gives us the value of F. 
To this end we have to follow the adiabatic change of condition of 
the air along each curve of flow, starting with the given tempera- 
ture and humidity, in the vertical above the center of the valley 
where x = —A/4, where the current is truly horizontal. 

By connecting together those points in the individual stream lines 
at which saturation is just attained we find, first, the contour of the 
cloudy region. 

Since the form of the cloud is also of interest in and of itself 9 
therefore its determination will be carried through as a part of our 
second example, in that above the center of the valley, where 
x = — XI 4 first for the summer, then for the winter, we make some 
assumption as to the mean distribution of temperature in accord- 
ance with von Bezold's collected data, on page 21 of his memoir 
above quoted. In accordance with this, we have: 

For y =-1500 -600 +400 +1400 

Valley above sea 
level 100 m. 



17.7° 11.0° 
8.2 6.69 
0.2° 0.6° 
2.92 2.17 


+ 0.9° 




+ 2400 meters. 

Height above sea 
level, 4000 m. 

- 5.0° 



F = 

t = 


In place of the value of F, designated by a star, we will take that 
value (2.2) that results from the smoothing out of the protuberant 
corners which the curve for F (see von Bezold, fig. 11,) shows at 
the altitude of 4000 meters. 

According to equation 5 the lines of flow have for their expression 


e ny sin m x 

b qn 

+ Be qi 

or if r is the value of y when x = o, and y — y = t), there results, 

e -n n e ~ny gin m % = L (* «" - 1), 

b qn 

b n I / i^L - ojl\ 

_ e ~ ny ° e" rn sin m x = ( e 2 — e 
m q 

9 It seems, for example, quite possible to argue from the observed bound- 
ary of the clouds inversely to the percentage of moisture in the current of 
air flowing toward the mountain slope. 


VOL. 51 

With the same degree of approximation as before the right-h and 
side of this equation can be put equal to -q; therefore the equation 
takes the following form: 


7) = b sin mx.e' 



which differs from equation (5') of the profile curve of the ground 
only through the factor which is constant for each line of flow, which 
factor causes the amplitude of the waves to steadily diminish up- 
ward . 

TJcm Jim. Tfcm. Wffm. skm. xrf* 

FIG. 3 

If, now, the lines of flow are made through a definite point y' h 
for the vertical and % = —^4, then for this point we determine the 
appropriate value if from the transcendental equation: 

7)' - - b-e~ rn e~ n %- v) 


and then substitute y h = y\— rf in equation 18. 

In this way we have computed the four lines of flow whose initial 
and lowest points are at the altitude above sea level of 1000, 2000, 
3000, and 4000 meters, and which are drawn as curves I, II, III, 
IV, in fig. 3. The highest points of these curves are at the altitudes 
2940, 3610, 4333, 5100 meters, respectively. 


If now, by means of the Hertzian table, we determine the altitudes 
at which condensation begins at the base curve o and for the curves 
I, II, III, IV, then assuming the above given values 10 of t and F we 
find the following results : 


For the summer 930 1570 2730 4060 (5125) 

For the winter 600 2070 3100 4130 5100 

In the summer, according to this table, condensation will not 
take place on the stream line IV, since its summit lies at the alti- 
tude of 5100 meters; the summit of the clouds will, therefore, lie 
a little below this. In the winter, the summit of line IV accidentally 
agrees with the summit of the cloud. In the construction of the 
cloud limit, introduced as a dotted line in fig. 3, and indicated by 
5 for summer and W for winter, we have also used the lines of flow 
midway between o and I, and I and II, respectively. 11 

We can now, with the help of the Hertzian table, easily find the 
quantity of water condensed in every kilogram of moist air as it 
progresses along any one of the lines of flow that we have constructed , 
either in its totality or as it passes successive vertical lines : we thus 
attain the folloiving values of the total condensation: 

Curve I II III 

For the summer 2.85 2.42 1.22 0.26 grams. 

For the winter 1.5 0.74 0.34 0.14grams. 

Let g x (h) be the quantity condensed up to the abscissa x when 
moving along that line of flow whose initial point is at the altitude 
h, and let H be the initial altitude of that line of flow which at the 
given abscissa x intersects the upper cloud limit; moreover, let u' 

10 From the above numbers it follows that an elevation of any kind of 
less than 500 meters will not give occasion for condensation under average 
atmospheric conditions, neither in summer nor in winter. In the summer, 
for a mountain altitude of between 600 and 800 meters, a cloud will form 
between the altitudes 1000 and 3000 meters, but will not touch the moun- 
tain; it is only for greater mountain heights that the cloud will rest on the 

11 In an analogous way for the first example, where we have assumed a 
plateau-like mountain of 900 meters altitude, we find a region of cloud which 
for the average summer conditions begins at 40 meters below the summit of 
the plateau and reaches up to over 3000 meters; but in winter, on the other 
hand, it begins at 500 meters above the valley and rises up only about 700 
meters above the mountain top; therefore, in this season it covers the moun- 
tain like a flat cap. 


be the horizontal velocity of flow and e' the density of the air at the 
altitude h above the bottom of the valley, therefore, for the point 
whose abscissa = — A/ 4; then will the total quantity condensed 
per second above the base area one meter broad from the beginning 
of the clouds to the point x, expressed in grams, be as follows: 


G s = {e'u'g x (h)dh (20) 

The quantity of air, e u kilograms, flows in one second through a 
strip of the vertical plane at x = —A/ 4, having a unit width and the 
height d h; but an equal quantity must flow out per second through 
the vertical whose abscissa is x, and since the condition is steady, 
it therefore behaves as though the quantity of air, e u, had moved 
in one second along the line of flow from — A/ 4 up to x; but in this 
the quantity of water e u g x (h) is separated from the air according 
to our definition of g. 

If we have computed G as a function of x, according to formula 
(20), then, finally, we have 

W = dG (21) 


as the quantity of water, expressed in grams, per horizontal square 
meter per second, that falls at the place x. In this way the deter- 
mination of Wis executed more conveniently than through the 
direct formula (17). By assuming the average conditions for the 
summer in the above example for a = 1, we find that the integral 
(20), if we compute it as approximately equal to the sum of the 
intervals between the individual current curves of flow as con- 
structed, gives the following: 

G x = = 1352, £„_<,„ -2680, G x = V4 - 3460 grams. 

This last number indicates the total precipitation falling on a strip 
one meter wide in one second on the side of the slope that faces the 
wind. According to the course of the curve 55, as shown in fig. 3, 
the precipitation begins, first, in the neighborhood of x = — 0.108^ 
and therefore is distributed along a strip of the ground surface, 
whose length is 0.358,}, or 8600 meters; from this we compute the 
average precipitation per hour, as follows: 

3.6 X 3460 , AK . .. 
= 1 .45 mm. depth 


= 1.45 kg. mass 


Similarly, we find for winter: 

C x = =380, G x=>l/6 ~770, G._ w -1264; 

the total precipitation is distributed over a strip 9400 meters iong, 
so that the average precipitation is 0.485 millimeters per hour. 

From the above three values of G (x) we can graphically construct 
the course of this function approximately by considering that the 
tangent to the curve for G is horizontal at its initial point and when 
x = + A/ 4. 

The tangent to the slope of the curve is found by considering its 
measure W. Thus we recognize in our case that the maximum 
of the precipitation in summer is attained between x = o and 
x = — 1, but in winter between x = o and x — + 2 kilometers 
and amounts to a X 2.2 millimeters, or a X o.75^millimeters per 
hour, respectively, for a wind velocity of a meters at some very 
great altitude; furthermore, after passing the summit of the moun- 
tain the precipitation diminishes more slowly than was found under 
our previous assumption of a constant thickness of clouds. In 
reality, on account of the conveyance of the water or ice with the 
cloud, which we still neglect as before, the maximum of precipita- 
tion is pushed still more toward the summit of the mountain. More- 
over, since one part of the cloud floats over the summit and is there 
dissipated in the sinking or descending currents of air, the precipi- 
tation will stretch a little beyond the summit, but its total quantity 
will be less than the computed. 

The results of the preceding analysis, namely, (1) that there exists 
a zone of maximum precipitation on the windward slope of a moun- 
tain and (2) that the inclination of the surface of the earth is more 
important than its absolute elevation, in determining the quantity 
of precipitation, are confirmed by observations, at least for the 
higher mountains. 12 

12 See Hann: " Klimatologie," Vol. I, p. 2( 



VOL. 5 I 


{Reprinted from Monthly Weather Review for July, igoi) 

(1) Assuming the average vertical distribution of temperature 
and moisture for each of the four seasons of the year as it is deduced 
by von Bezold from the scientific balloon ascensions published by 
Berson and Assmann in their " Ergebnissen. " "The results of 
scientific balloon voyages," there result the following minimum 
elevations required in order that condensation may begin in a mass 
1 »f air that was originally at the absolute altitude H above sea level. 




























8 3 5 

















The smallest number in each column is also the smallest altitude 
that a mountain ridge must possess in order to cause the formation 
of clouds under the assumed conditions, but it is only in the case of 
a very broad mountain ridge that such small altitude will suffice. 
We see that in the autumn and winter a mountain of about 400 
meters in height will suffice to produce a formation of cloud in con- 
tact with the summit of the mountain whereas, in spring and summer 
the mountain must be higher (namely about 500 or 570 meters 
respectively), and when the air passes over this mountain the forma- 
tion of cloud will begin in the layer lying at 500 or 1000 meters above 
its summit. These numbers at present serve only as examples; in 
practice, however, they suggest that as soon as we observe the for- 
mation of cloud above a mountain of less altitude than the above 
given tabular minimum altitude, we may conclude somewhat as to 
the average moisture at that altitude at that time. We may also 
remark that on account of the increasing flatness of the lines of flow 
as the altitude increases, the above given minimum altitudes must be 
exceeded by so much the more in proption as the width of the sum- 
mit ridge is smaller, and the altitude of the layer in which the con- 
densation begins is higher. 


(2) The method developed by me for computing the condensa- 
tion that occurs on any given mountain slope cannot be applied to 
computing the mean value of the precipitation for any given interval 
of time, by introducing into the computation the mean values of 
the temperature* and moisture for this interval. We should in this 
way find too small a precipitation. Thus, for example, the altitude 
of the mountains might not suffice to cause any condensation at all 
for the average condition of the air, but could cause it on those 
occasions when the moisture exceeds its average value, wherefore 
the average value of the rainfall for the interval of time under con- 
sideration would be different from zero. As the variation of the 
moisture from its average value may cause rainfalls where otherwise 
there would be none, so also, with the currents of air mechanically 
forced to ascend mountain ranges, and whose effect is superposed 
upon that of the general circulation of the air in cyclonic areas ; for 
it can happen that neither one of these two causes may alone suffice 
to form rain, but that both together do. This explains why eleva- 
tions of the surface of the earth of from ioo to 200 meters increase 
the annual mean value of the total precipitation, as for instance, as 
shown by the charts in Assmann's memoir of 1886, " Einfluss, etc., " 
"On the influence of mountains on the climate of central Germany." 

(3) The examples given in my article show that in so far as con- 
densation in general takes place on the slopes of mountains, its 
intensity (therefore also, the density of the precipitation when fall- 
ing vertically) is in general greatest where the slope of the moun- 
tain is steepest. If now we consider that in the course of all the 
various conditions of the atmosphere that may occur in a long 
interval of time, the first condensation occurs most frequently above 
the upper portion of the slope, then it follows that the average den- 
sity of precipitation computed for a long interval of time, must 
increase, not only with the inclination of the slope, but also with 
the absolute altitude of the locality under consideration. To this 
case corresponds the formula for the annual quantity of precipita- 
tion expressed in millimeters deduced by Dr. R. Huber in his 
" Untersuchungen, etc. Investigation of the distribution of precipi- 
tation in the canton of Basle," namely: 

N = 793 -f 0.414/1 -f 381.6 tan a 

where h is the altitude in meters, and a indicates the gradient angle. 
(See A. Riggenbach, Verhandlungen der Naturforschenden Gesell- 
schaft. Basel, 1895. Vol. X, p. 425). 


(4) From a comparison of the effects of different broad moun- 
tain ranges of the same altitude, it results (see page 474 of my 
article, or page 95 of this translation from the Monthly Weather Re- 
view) that the smaller, and therefore steeper, mountains always 
cause a smaller total condensation than the broader and narrower 
mountain summits. Notwithstanding this, the density of precipi- 
tation on the slope of the smaller is generally larger than on the 
sl< >pe of the larger mountains because the smaller total precipitation 
is distributed over a ground surface that is relatively much smaller 
yet. In reality, however, this only obtains so long as the quantity 
of water remaining suspended in the cloud is only a small fraction of 
the total condensation; in the case of very narrow mountain ridges 
it will be more apt to happen that a considerable fraction passes on 
over and beyond the summit and is subsequently again evaporated 
[and therefore does not appear as rainfall]. 

(5) I regret to notice that in the first two figures of my original 
memoir, as also in the translation, the legend inscribed on the curves 
representing the distribution of precipitation reads "precipitation 
in millimeters per second," instead of "per hour," as is correctly 
stated in the text; the necessary correction should be made. 
[Corrections have been made in the present volume.] 

(6) A precise test of this theory cannot at present be carried out 
because we have not sufficient observations of the condition of the 
upper strata and of ground along the slope of a given mountain range. 




Memoirs of the Imperial Academy of Sciences of Saint Petersburg, igo^, 

Volume XV, section g] 

[The original memoir above quoted is published in the Russian language' 
the brief abstract, in French, communicated by the author to the " Annates" 
of the Met. Soc. of France, Vol. LIII, pp. 1 13-120, May 1904, has been fol- 
owed in the present translation] 

In order to study in detail the causes of the origin of atmospheric 
disturbances and their more important properties it is first necessary 
to resolve the fundamental problem : given a certain system of me- 
chanical forces, how will it act on the air in the different strata of the 
terrestrial atmosphere? We propose to seek the possible solution for 
the special case of the strata situated in the immediate neighbor- 
hood of the ground. 

It is easy to determine the action of any force in the midst of 
an ideal gas, whose particles move among themselves without fric- 
tion, as in a vacuum and according to the law of gravitation only. 
In order to pass from such a gas to the atmosphere it is necessary 
to know not only the special properties of the air itself but also 
the value of the friction. The influence of this friction is appre- 
ciable in two ways: (1) it diminishes the velocity of the progressive 
movement of the air, (2) it enfeebles the action of the force per- 
pendicular to this current of air, and to the same extent diminishes 
the angular velocity of the atmospheric particles. The evaluation 
of this normal friction, that is to say, the effect directed along the 
line normal to the direction of motion is the subject that we propose 
to study in this memoir. 

Fortunately there is at our disposition a very convenient agent 
that one can utilize to this end. This is a well-known constant 
force and one which is always perpendicular to the trajectory of 
any mass that is moving on a horizontal plane: it is the action of 
the diurnal rotation of the earth. 



Let us imagine an elementary small mass projected horizontally 
on a perfectly polished plane surface with an initial velocity v . 
It continues along its route with this same velocity tracing a curved 
line whose radius of curvature, as is well known, is given by the 

47r • ft\ 

sin a (1) 


where T is the duration of a complete day, i. e., one rotation of the 
earth and a the latitude of any place on the earth traversed by 
the center of the mass. The center of curvature of the path is 
always on the left-hand of the direction of motion where the mass 
is in the southern hemisphere and on the right-hand in the north- 
ern hemisphere. 

Assume the notation 

K = „ sin a 

which expresses the angular velocity of the moving point. Its 
value can be calculated from observation of the wind between two 
stations in the following manner: 

At the station A take an observation of the velocity of the wind 
blowing towards the station B. The distance between the stations 
and the velocity of the wind being known we obtain by simple divi- 
sion the interval of time required by the particle of air to reach the 
station B. By observing at this moment the direction of the wind 
(at B) we find a difference between the two observed directions, 
which difference should give us the required value K. This value 
generally differs greatly from that calculated by the theoretical 
formula because of the many accidental conditions, among which 
there is however one force that constantly and continuously influ- 
ences the movements of the atmosphere. This is the internal fric- 
tion (or viscosity) of the air and also the friction between the air 
and the surface of the terrestrial globe. If the number of observa- 
tions employed by us is sufficiently large, as well as the length of 
the period of time and the number of stations collated, then all 
anomalies neutralize each other and one obtains a resulting mean 
value for K as diminished by friction only, or k = fx K 
Now it is the coefficient n that is the desired characteristic of the 
air near the ground. 


The preceding expresses only the general scheme of the proposed 
method, for in fact Ave still have to surmount numerous difficulties, 
the more important of which are trie following: 

(i) The terrestrial surfaces separating the stations A and B 
should be as flat .and smooth as possible, not having any high obstruc- 
tions, in order that the air may pass freely from one station to the 
other. For this reason we have not utilized observations of refined 
anemometers and anemographs which are located frequently in 
large cities and have felt obliged to rely on the observations of 
stations of the second class in the meteorological system of Russia. 

(2) We do not generally find at station A a current of air flow- 
ing exactly towards station B but inclined to that direction by the 
angle /3 so that for the length of the path described by the wind 
between the two stations it suffices to take the distance 

S = A B cos p (3) 

In fact this can only be an approximation since the trajectories of 
the atmospheric particles are curves and not straight lines and the 
value of 5 is larger rather than smaller than that indicated by the 
equation (3). 

(3) The instrument by which at Russian stations we ordinarily 
measure the direction and the velocity of the wind is a wind vane 
placed at the summit of a mast and furnished with a suspended plate 
of steel which by its departure from a vertical position indicates 
the velocity of the wind. Now the iron cross-piece of this mast 
showing the cardinal points, N., S., E., W., is often oriented inac- 
curately and the wooden mast that carries it often acquires after 
awhile a twist introducing an angular error anxmnting to many 
degrees. The observations made by such a primitive apparatus 
cannot be very exact, so that one must utilize very many of them 
in order to eliminate these errors. 

(4) At stations of the second class the observations are made 
at 7 a.m., 1 p.m. and 9 p.m., and consequently we do not generally 
find at the station B any observations for the moment of time that 
we have found by our calculation. It remains then only to make 
a proportional interpolation between the two observations that 
come nearest to this moment. 

As the detailed exposition of the method now proposed cannot 
be given within the limit of this abstract we must here confine our- 
selves to giving the results of our calculation: 

The mean value of /a for 3762 cases is found to be 

fi = 0.026 (4) 


But this value has neither practical nor scientific interest because 
it corresponds to the whole scale of different velocities of the wind 
from 5 meters per second up to 20. This itself explains the impor- 
tance of the so-called mean error. 

By grouping the numbers in such a way that each group corre- 
sponds to a certain velocity we have formed the following table: 






per second. 

of cases. 



















1 1 






l 3 












l 7 










The negative values show the cases where the wind deviated to 
the left, in spite of the theory, and not to the right of the rectilinear 

After having submitted this table to a detailed examination, 
which I need not repeat here, we have obtained the following values 
of t he coefficient jj. corresponding to four different values of v. 









15.9 0.092 

Representing these figures graphically by orthogonal coordinates 
we obtain a very regular curve, a sort of parabola, whose natural 
prolongation crosses the axis of ft at some distance from the origin. 
The equation of such a parabola, as we well know, is 

fi = c v 2 +c' (5) 

where c and c' are constant parameters. The introduction of 
the parameter c' is explained by the law of Dove, according to 
which the weather vane at any meteorological station in Europe 


generally turns in a direction contrary to the motion of the hands 
of a watch when an area of low pressure is passing by the station. 
In fact by employing the observations of stations for which Dove's 
law holds good we obtain a coefficient greater than it ought to be 
by a quantity independent of the velocity of the wind. This inter- 
esting phenomenon is shown with perfect clearness in the graphical 

As the function // v characterizes the friction of the air in a direc- 
tion perpendicular to the current, one ought to be able to determine 
this function theoretically, if we knew a similar function for the 
direction parallel to the current, since the two coefficients ought to 
depend directly on each other. 

During the progressive motion of masses of air a certain friction 
is developed whose reaction, tending to reduce the linear velocity 
of the movement, is perpendicular to this velocity, according to 
the simple law of Guldberg and Mohn / = tj v where /is the reaction 
of the friction, v is the velocity of this wind and tj is a coefficient 
that depends only on the pysical state of the air and the surface of 
the earth. 

This being recognized, we have studied a regular stationary 
cyclone of large extent and without any progressive movement, 
from a purely mechanical point of view. After having examined a 
portion of this whirlwind somewhat distant from its axis we have 
obtained the following expression for the function (vpi) viz: 

1 1 

- = -7 + 1 (7) 

H sir 


sin a cos a 

£ = ~T7 : — 1? s (8) 

J (r) sin a — K cos a) 

The letters introduced into these formulae have the following 
signification : a is the angle formed by the direction of the current 
of air with the radius vector R drawn to the axis of the whirlwind 
and is counted positively starting in the direction of the motion of 
the hands of a watch from some initial radius vector; / indicates 
the product v R which I have called the expression of the intensity 
of any atmospheric disturbance ; K is given, as already stated, by 

K = T sin a (9) 


The values of K and rj being absolutely independent of v, as also 
are the values of J and a, for the given disturbance, as has been 
demonstrated in our study, therefore the coefficient e is also inde- 
pendent of v. But equation (7) expresses a certain property of the 
atmosphere, whatever may be the special phenomena by means 
of which it has been determined. It results from this that the value 
s in formula (8) is absolutely constant for any physical state of the 
atmosphere and that it cannot vary with a nor with /, but only 
with the coefficient 77; the values of e and rj characterize the friction, 
both of them, but in different directions only. 

This argumentation may seem to be erroneous, as I have already 
had occasion to convince myself on hearing the opinions of several 
experts to the effect that any such process of investigation seems 
to them doubtful and untrustworthy. 

In place of defending my logic or my honest)^ against the incredu- 
lous, I allow myself here to show in brief some properties of equation 
(8) which result from the assumption as to the invariability of the 
coefficient e. 

We can consider equation (8) as the expression of the connection 
that must exist between the angle a on the one hand and the values 
J K and rj on the other. Let us examine each of these connections 
separately by means of the special partial derivatives: 

a considered as a function of J. — This function has two different 

(1) At the moment when a disturbance originates / = o, the 

air is put into gyratory motion in the direction of the hands of a 

watch in the southern hemisphere, for which a = ~ and in the 

inverse direction in the northern hemisphere for which a = \ n. In 
proportion as the velocity of rotation increases the current of air 
deviates towards the center, that is to say, to the right hand in the 
southern hemisphere (and / a l so increases as a increases), but to 
the left-hand in the northern hemisphere, where / increases as a 
decreases. We thus have a cyclone properly so-called, and par- 
ticularly so far as regards its lower portion. 

(2) At the commencement of the phenomenon the air expands 
outward from the center along the radius for which a = x. Then 
as the disturbance develops the currents of air commence to deflect 
in the direction of the movement of the hands of a watch in the 
northern hemisphere, where a increases, and in the inverse direc- 
tion in the southern hemisphere, where a decreases. This is the 
lower portion of the anticyclone. 


For these two cases there is a certain limit that the angle a can 
only attain when / becomes infinitely large, which is determined 
by the equation (10), 


tang a = — (10) 

K as a function of a. — This function also has two branches. 

When a disturbance takes place in the equatorial regions the air 
flows along the gradients, that is to say, towards the center (for 
which a = o or away from the center in the opposite direction 
(for which a = iz). The first case corresponds to an area of low 
pressure and the second to an area of high pressure. If the center 
of the disturbance moves towards the north, the currents of air will 
deflect to the right (or a will increase with K). If the center moves 
towards the south, the current will deviate towards the left and a 
decreases with K. 

rj as a function of a. — It may be remarked that the analysis of this 
function can only be of a general character because in the form 
of equation (8) it occurs as a function of e of unknown form. If we 
rely iipon equations (i), (2) and (7) we find without difficulty that. 
e is infinitely large when r, = o; e is zero when rj is infinitely large; 
finally the derivative of e with regard to rj is always very small or 
nearly equal to zero. Moreover it is evident that if rj has positive 
values, different from zero, then e also has values greater than zero. 

A discussion of equation (8) shows moreover that the product 
erj is positive for rj = o and for rj = infinity. These peculiarities 

lead us to adopt e == - as the value of the function e, in which n is 

a constant. 

The vitality or duration of any atmospheric disturbance depends 
directly on the magnitude of the angle a between the Avind and the 
gradient; in proportion as a increases the duration increases also; 
in proportion as the wind deviates from the gradient it is more ami 
more difficult to reestablish static equilibrium. 

This being understood, let us examine some interesting me- 
chanical phenomena that we may draw from the preceding analysis. 

(1) When an anticyclone continues to develop its vitality: 
(a) increases steadily, whence it results that disturbances of this 
character ought to have a very considerable stability not requiring 
help from outside. 

(2) On the contrary, when a cyclone is developing, its vitality 
is decreasing so that a fully formed cyclone carries within itself 
the beginnings of its destruction, hence the extreme instability 


of cyclonic formations, which can only persist by the aid of such 
exterior sources as furnish the necessary energy. 

(3) The movement of a disturbance towards the pole increases 
its vitality, and vice versa. 

(4) As the friction increases, the vitality of a disturbance 
diminishes. Since friction [internal friction or viscosity] is greater 
in moist air than in dry air it follows that a disturbance should 
lose vitality when approaching moist vapors, and vice versa. 

Ordinarily barometric maxima follow this rule quite closely, but 
the minima seems to behave contrariwise and very persistently 
so. This fact shows again that the cyclone of the temperate zone 
is essentially a thermodynamic disturbance while the anticyclone 
is a mechanical disturbance. Thus we explain the profound dif- 
ference that exists in all respects between these two kinds of whirl- 
winds which are so similar in appearance. 

(5) The intensity of any disturbance, or the product v R, cer- 
tainly increases in proportion to the distance from its center, for 
the atmospheric currents become more and more nearly horizontal: 
hence follows the following very interesting theorem: 

The vitality of a cyclone diminishes in passing from its center 
towards its boundary which causes an excessive sensitiveness at 
the latter; when the cyclone is extensive with a very deep depression 
its exterior isobars vary incessantly. On the contrary the anti- 
cyclone has permanent and firm contours and its center of high 
pressure moves hither and thither without exerting any influence 
whatever on the boundaries of the whirl. 

Because of this difference the collision between these two classes 
of disturbances acts destructively upon only one of the two, that is, 
the cyclone, which eventually is destroyed or modified. 

From the preceding we see that equation (8) gives us a fairly 
probable as well as general representation of the characteristics and 
motions that belong to atmospheric disturbances, excepting only 
one of the most important of the movements, that is, the progressive 
motion of the whirlwind itself. The direction and velocity of this 
movement are determined principally by the diurnal rotation of the 
earth, which action becomes stronger in proportion as the height 
of the whirl is greater. Now Ave are not yet able to study this action 
because the law according to which the friction of the air varies with 
altitude is at present wholly unknown. However, we hope that the 
current exploration of the atmosphere with kites and sounding bal- 
loons will not fail to clear up this question, which is as interesting 
from a purely scientific point of view as it is important for the 
practical forecasting of the weather. 



Fellow of St. John's College, Cambridge 

[Communicated to the Royal Society, London, by Dr. W. N. Shaw, F.R.S. 

February 25, and Read March 5, igo8. Printed in Proc. Roy. 

Soc. Vol., 80, May 25, igo8] 

For the steady horizontal motion of air along a path whose radius 
of curvature is r, we may write directly the equation 

(tor sin X ± v) 2 _ 1 dp (cor sin X ) 2 
r p dr r 

expressing the fact that the part of the centrifugal force arising from 
the motion of the wind is balanced by the effective gradient of pres- 

In the equation p is atmospheric pressure, p density, v velocity 
of moving air, X is latitude, and co is the angular velocity of the 
earth about its axis. 

If dp/dr be negative, it is clear that v and co r sin X must have 
opposite signs: or, for motion in a path concave towards the higher 
pressure, the air must rotate in a clockwise direction, the well- 
known result for anticyclonic motion. Further, the maximum nu- 
merical value of 

1 dp (co r sin X) 2 
pdr 1S ~2~ 

and the corresponding maximum value for v is co r sin X. Therefore, 
in anticyclonic regions there are limiting values which the gradient 
and the velocity cannot exceed. This limiting value of v for lati- 
tude 50 and r = 100 miles is approximately 20 miles per hour. 

At the surface of the earth, owing to friction and eddies, the mean 
direction of the motion of the air is nearly always inclined to the 
isobars; but over the sea the inclination is very much less, and it 
seemed probable that in the upper regions of the atmosphere, if 



the motion were steady, the air would in general move tangentially 
to the isobars, and its velocity would agree with that calculated from 
the equation given above. 

The question, however, arises as to whether the pressure is likely 
to continue steady long enough for a condition to supervene in which 
the equation is applicable. We can get an idea of the time that 
would elapse before air, starting from rest, would reach a state of 
steady motion, by considering the motion of a particle on the earth's 
surface (1) under a constant force in a constant direction, corre- 
sponding to straight isobars; (2) under a constant radial force corre- 
sponding to cyclonic and anticyclonic conditions. The particle 
would begin to move at right angles to the isobars in the direction 
of the force, but as its velocity increased it would be deflected by the 
effect of the earth's rotation until it moved perpendicularly to the 

The equations of motion of a particle, referred to axes fixed rela- 
tively to the earth and having an origin on the surface in latitude 
X, are 

dH dz .dy v 

■ — — 2co cos X — - 2w sin i ~ = X , 
dt 2 dt dt 

d 2 y _ . , dx - T 

—L + 2uj sin X — = Y, 
dt 2 dt 

d 2 z dx 

— + 2u> cos k — — Z, 

dt dt 

where the axis of z is vertical and the axes of x and y are west and 
south respectively. 

If there is no vertical motion we may write the first two equations 

d 2 x dy ,. d 2 y , dx * , 

— — a — = X , — + a — = Y, 

dt 2 dt dt 2 dt 

and the form of the equations and the value of a are unaltered by 
changing to other axes in the same plane. Let us take the y axis 
to be in the direction of the constant force b. Then 


d 2 x dy 
dJ dt 

= 0, 

d 2 y dx , 
-^L + a — = b, 
dt 2 dt 

b ( * 
x = — (at- 

- sin at) , 

v = — (1— cos at) 
' a 2 



if the particle start from rest. The motion is therefore oscillatory, 
and the particle moves in a series of cycloidal-like curves, fig. 1. 
The times to the successive intersections with y = b/a? are 7z/2a, 
lizjia, etc. For latitude 50 these are about 4 and 12 hours. They 
are independent of b. If there is damping the motion will be as in 
fig 2. If the motion is resisted by a force k v proportional to the 
velocity, the path will be inclined to the .T-axis. Fig. 3 gives the 
path for the particular case k = a and for a period of time equal to 
27i/a or 16 hours. 


FIG. 2 

FIG. 3 

In the case of a constant radial force we have for the motion 

d 2 r /dd\ 2 „ dO 

- — r [ ■ — = K + ar — , 
dt 2 \dt dt 

d 2 d _ dr dd dr n 

r — + 2 - - + a — = 0, 
df dt dt pt 



If the particle start from the center, 

and we obtain 


B — and = — Jo, 

4:R 4:R 

r = (1 — cos % at) = (1 — cos 6). 

The particle therefore describes a cardioid, but if there is damping 
the motion will come to be along the circle r = ^R/a 2 . 

The time to reach the circle is 7i/a, or about 8 hours for latitude 


These times are not large meteorologically, and we may there- 
fore expect the relation between air velocity and pressure gradient 
to be that corresponding to steady motion so long as there are no 
irregularities to produce turbulent motion. 

For application to wind velocities in the upper air we require to 
know the upper-air isobars. If we have air in which the horizontal 
laj^ers are isothermal, then from the equations 

dp= - gp dz, p = gkpT, 

it follows that 

P. Jo kT 

We have, therefore, if p and p + dp are surface isobars and 
p z and p z + dp z the corresponding upper isobars, 

*-^ f so that **•-*£. 

Pz Po Pz Po To 

Therefore the velocity calculated from the surface isobars will 
apply to the upper air, except for the factor, TJT . For z =iooo 
meters the effect of this factor is to diminish the velocity by about 
2 per cent. 

If the conditions are not isothermal, but such that the isotherms 
and isobars intersect at an angle (/>, the upper isobars will have a 
different direction from the surface isobars, and the value of the 
upper gradient will also be changed (see fig. 4). 


The pressure at a height z above B the point of intersection of p , 
T , is p e ~ z,k T m, and above A, the point of intersection of p + dp , 
T - dT , is 1 

{p + dp,)e-* /k(T m- dT m>- 

If we assume the vertical temperature gradient to be the same 
over all the region considered, d T will be the same for every element 
of the above integral, and we can put dT m = dT . 

If these two pressures at height z are equal, we must have 


dPo = z _ dT o or d Po 

N\p +rfp 
FIG. 4 

In this case A B is the direction of the upper isobar and its incli- 
nation <j> to the lower isobar is given by 


tan<i = 

xd T Q cosec </» + ydy cot </» 

where xdT and y dp are the distances between the isotherms and 

Substituting for dT and dividing out by dp we get 

x T 2 
cot <f> = cot (J) + ^— cosec (p. 


Taking y and x for millimeter isobars and i° C. isotherms and 

putting z = 1000 meters and T„?/T = 2T m — T = 270 C. say, 

we find 

cot 4> = cot i[> -|- 2 '8 - cosec <p. 


x ln. this and subsequent formulas the reader will understand that the 
erms following the / belong to the exponents of e. 



VOL. 51 

To obtain the upper pressure gradient, we consider the upper 
isobars over B and A'. The difference of tern] erature between B 
and .V is assumed y/x . dp cos W = dt. 

Therefore the upper pressure difference is 

(/>» + dp*) e~'/k (T m + dt) - p e-*/kT m =e-*/kT„ 

p z . y cos c 

j P a zdt 

dp* + 


= P<T z lkT, 

1 + 


The distance between these isobars is \dp, ) cos <f> and the upper 
gradient is consequently 

, 1 

y cos <j> 


1 + 

/ i „cr cos </> 

y cos ^ 


j , gflpg^o ycos <p 


and the ratio 

1 dp, 1 a/? . 

is sec 

o 2 6r ,o dr 

1 + 

y cos 1 T z 

% J r ' 


gpz cosec <£. 

y si" v 

taking TJT to be unity, namely, 
cosec ^ 



cot ^ — cot (// sin (</> — </•>) 
In the special cases. = o or 180 , the ratios are 

T ' x 


1 ± - ), for z = 1000 meters. 


If x = 2y, which would represent a possible case, the increase or 
decrease would be about 18 ] er cent. 

For 4> = i~ tl e rotation would in the same circumstances be 
lit io°. 

During tl e year 1905 aseries of observations in theupper airwas 
made at Berlin and Lindenberg, near the time of the general 8a.m. 


morning observations. It was therefore possible to compare the 
wind velocities observed with those calculated from measurements 
of the gradient by the use of the' formula at the beginning of this 
paper, the motion being assumed tangential to the isobars. 
For purposes, of calculation the formula may be written 

v(l ±0. 00108 v cot dcosecJ) = 709 cosec /I T go 

x T B 

where <p is the angular radius of the small circle, on the earth's sur- 
face, osculating the path, v is in meters per second, x is the distance 
in kilometers between millimeter isobars, T, B axe the temperature 
and pressure, and T , B the corresponding values for air at o° C. 
and 76o mm . 

If the motion is along straight lines, cot (p = o, and the values of 
v for B = B Q , T = T , are as follows if x = 50 kilometers: 

Latitude 3C° 

v 28 . 4 

If v represent the velocity when cot <J> = o, we can most easily 
express the solutions of the equation for different values of 0, x, 
X, by taking as independent variables, </>, v Q , X. 

Taking, as an example of the dependence on <//, X = 50 , v = 40 
meters per second, we obtain the following values for v in meters 
per second in the case of cyclonic motion : 

i/> 1° 2° 3° 4° 5° 6° 7° 8° 9° 10° 

v 17 21 24 26 28 29 30 31 31 32 

For anticyclonic motion the gradient corresponding to v = 40 
meters per second is above the maximum, and we take for two ex- 
amples v = 12, and v = 30 meters per second. 

The values of v are then as follows for the two cases: 

1° 2° 3° 4° 

For v = 12 v = = - - 20 

For 7,7. = 30 v = - - - - 




















13 m. p. s, 
50 " 

Where no value is inserted for v, the gradient corresponding to 
the given value of v is above the maximum for the corresponding 
value of (p. 

To show the dependence on X, we take (p— 3 and put v = 40 
meters per second for cyclonic motion, and v = 10.5 meters per 



VOL. 51 

second for anticyclonic motion. The following table gives the 
values of v for different latitudes in the three cases: 

; 30° 



60° 70° 

For v = 40 . . 

...v = 21 



25 26 m.p.s 

Forz; = 10 . . 

. . .v — — 



17 15 

For v = 5 . . . 

...v = 7.1 



5.8 5.7 " 

By the use of tables giving values of v for different values of x, 
T, B, and of v for different values of X, v , <p, each wind observation 
at 1000 meters altitude was compared with the value deduced from 
the surface isobars. The temperature correction was not applied. 

The following table gives the result of the comparisons: 


January i5.7 

February 12.0 




May S.3 

June 6 

July 8.4 

August S.9 

September 10.3 

October 12. 1 

November 10.6 

December 1 1 . o 

Summer 8.8 

Winter 11. 6 

Year 10. 2 

The upper wind coincides in direction very nearly with the isobars 
at the surface, and the wind velocity observed agrees well with that 
calculated from the pressure distribution. The differences are not 
greater than possible errors of observation, except in spring. 

It is known that the upper wind always veers from the surface 
wind, and the numbers in Column 7 show that in 1905 the veering 
was considerably greater in winter than in summer. 

If the effect of the earth's surface were the same as if a frictional 
force opposed the motion, the relation between the wind and grad- 


ient of pressure would be as above, except that the effective gradient 
would be the maximum gradient multiplied by the cosine of a, the 
angle between the path and the isobars. The corresponding velocity- 
would be approximately v cos a, except in cases of considerable 
curvature. In^the majority of the observations the curvature was 
small, and we should therefore expect the surface wind to be nearly 
v cos a, so that the numbers in Column 8 would be nearly unity. 
This is far from being the case; but the change of the station of 
observation from Berlin to Lindenberg is accompanied by a cor 
responding change in the ratio of the surface wind velocity to v cos a. 

This suggests that the effect of the surface, apart from the purely 
frictional effect, is to reduce the velocity in a given direction in a 
constant ratio depending on the locality, and that departures in the 
observed velocities from those corresponding to this ratio are to 
be associated with unsteady meteorological conditions. 

The last column 1 gives approximately the ratio of the volume of 
air crossing the isobars at the surface to the volume crossing at 
iooo meters. 

The ratio appears to be nearly constant ; the change in December 
is probably due to the exceptional conditions which prevailed during 
part of the month, when the air was considerably warmer at iooo 
meters altitude than at the surface. 

'Or, the wind component, perpendicular to the isobars at sea level divided 
by the analogous component at iooo meters. 




(Christiania, 1876, revised 1883) 2 


Meteorological phenomena being very complicated, we shall 
attain final success in their mathematical study only by treating 
simple cases which are analogous to those of nature. The equi- 
librium and the movement of the air form a part of the mechanics of 
fluids that is as yet very little developed because there exists too 
few observations for the verification of the numerical calculations. 
Encouraged by the fine results obtained by M. M. Peslin, Reye, 
Colding, Ferrel and Hann in this new application of analysis to 
meteorology, we have applied the principles of mechanics to the 
movements of the atmosphere, and have arrived at some results 
which we think are not without importance for the development of 
meteorological science. In the first place we have found that one 
of the first things to do in order to insure the success of meteorology 
is the creation of meteorological stations at high altitudes; either 
on mountains or in balloons, and supplied if possible with self- 
registering instruments. 

The winds or the horizontal currents of air at the surface of the 
earth are intimately connected with the vertical currents; but tl e 
origin and the displacement of these latter depend not only on the 
physical state of the air at the surface of the earth, but also on the 

1 Etudes sur les Mouvements de L' Atmosphere. Par C. M. Guldberg 
et H. Mohn. Premiere Partie, Christiania, 1876. Deuxieme Partie, Christi- 
ania, 1880. [Revised by the Authors in i883-'85.] 

2 By personal interview with the authors, and correspondence during 
the years 1883 to 1886, Prof. Frank Waldo secured from Professors Guld- 
berg and Mohn a revision of the original French edition of this Memoir 
with permission to publish a translation for the use of American students. 
The delay in publication has given me opportunity for a slight revision of 
Prof. Waldo's translation. — C. Abbe. 


physical state of the air of the upper strata. Moreover the velocity 
of the wind and its direction are both eminently under the influence 
of the surface of the earth, while their values at a certain height would 
probably present the regularity that must obtain in order to be 
able to predict the progress of meteorological phenomena. 

In studying horizontal currents under simple hypotheses, we have 
introduced the friction due to the surface of the earth, and we have 
applied our theory to the winds crossing over the equator, and to 
whirlwinds. 3 

The numerical calculations accord with the phenomena of nature 
\vithin such limits as correspond to the established hypotheses. 
It follows that the exact observation of the velocity of the wind 
will be of great importance to meteorology. We hope that these 
results drawn from the mechanics of the atmosphere will show the 
necessity of more extended meteorological observations especially 
in the tropical regions and in the higher strata of the atmosphere, 
and that true progress in meteorology is founded on the develop- 
ment of the mechanics of the atmosphere. 

Chapter I 


§1. Pressure, virtual temperature 

In studying the equilibrium and the movements of the atmos- 
phere, it suffices to consider the air as a mixture of dry air and of 
aqueous vapor. The other gases forming the elements of the atmos- 
phere, of which carbonic acid gas is the most important, are found 
only in such small quantities that their action may be neglected. 
The quantity of aqueous vapor in the atmosphere is so small that 
we can accept the law of Mariotte and Gay-Lussac for moist air 
within the range of temperatures that occur on the earth. It is 
necessary, however, to consider the cases in which the vapor con- 
denses and passes into the liquid state or the solid state. 
We use the notation 

p the pressure in kilograms on a square meter. 
p the density or mass of a cubic meter. 4 
t the temperature in degrees centigrade. 

3 M. C. M. Guldberg had already, in 1872, developed a part of this theory 
in the Norwegian Polytechnic Journal (Polyteknisk Tideskrift), p. 73, 1872. 
— Editor. 

4 In the absolute system here used this mass is the weight divided by 
gravity. — Editor. 


The law of Mariotte and Gay-Lussac as applied to a kilogram of 
gas is written 

p = ap (273°*+ t) (1) 

where 273 -f x is the absolute temperature. 

Here a designates a constant which depends on the nature of the 
gas; for dry air we have 

a = 287.09 

In applying this law to a mixture containing 1 — q kilograms of 
dry air and q kilograms of the vapor of water, expressing the tension 
of the vapor of water by / and its relative density by i/e we find 

p - f - a (1 - q) p (273 + r). 
f =eaqp (273 + r). 
p = a (1 + (e - 1) q) p (273 + t) (2) 

t _ 11 (3) 

P~ 1 + (t - 1) q k > 

1 / 
q = e ' P (4) 

1 - £-1 f 

T~ ' p 

By substituting this value of q in equation (2) and putting 

273 + t 

t = nr73T-7 (5) 

e ' P 
we have for moist air 

p = apT % (6) 

We call the quantity T the virtual temperature; for dry air the 
virtual temperature is the same as the absolute temperature. 

If we consider a mixture which contains 1 kilogram of dry air 
and x kilograms of vapor of water, we shall have 

'-rr q -i j=} • • - (7) 

i = r x* (8) 

p I + £ x 


If we have 

1 £ - 1 

- = 0.623 =0.377 

e s 

Then for air at 76o mm we have the values in the following table: 







- 3 o°C 




























§2. Height of the atmosphere — mean pressure 

We can adopt either one of two hypotheses concerning the height 
of the atmosphere. We can suppose that the atmosphere is limited ; 
in this case the temperature of the exterior stratum must neces- 
sarily be absolute zero, for at this temperature the tension of a gas 
is equal to zero. The other hypothesis is that the atmosphere ex- 
tends indefinitely into space and that space is filled with a gas whose 
tension is extremely feeble. For meteorology it matters little which 
hypothesis is chosen, because in both cases the tension of the air 
at very great heights will be insensible. Suppose 76o mm be the 
pressure at the surface of the earth and suppose the temperature of 
the atmosphere constant and equal to zero centigrade, we shall find 
the pressure at the height of 200 000 meters equal to 0.000 000 oi mm . 

If the atmosphere does not contain the vapor of water its mass 
will be invariable; if we suppose, moreover, that gravity does not 
vary with elevation, the weight of this mass will be constant and 
by calculating the mean pressure on the entire surface of the earth 
it will be found to remain always the same. Considering the pres- 
ence of the vapor of water whose quantity varies from time to 
time, we shall see that the mass of the atmosphere does not remain 
constant and that, consequently, the mean pressure varies with the 

We have assumed that gravity is constant. In truth it varies 
with the altitude and consequently the pressure of the atmosphere 
depends on the law of the distribution of the mass in a vertical direc- 


tion. This distribution is a function of the temperature, and con- 
sequently even the pressure of a dry atmosphere will vary with 
the temperature in such manner that the pressure will diminish in 
proportion as the temperature increases. However, the variation 
of gravity for atmospheric strata at slight elevations is so slight that 
its action can be neglected in meteorology. 

§3. Temperature of the atmosphere 

Temperature depends on many considerations and there has 
not yet been found any function that expresses the temperature 
in terms of the coordinates of position and the time. 

The heat of the sun and of space, the absorption of the earth and 
of space, the radiation, the conductibility and the movement of the 
air, all affect the temperature. Hitherto we have sought to deter- 
mine the temperature at the surface of the earth as a function of the 
time. We shall see that the variation of the temperature with the 
height is of the greatest importance in meteorology. The observa- 
tions of this phenomenon are not numerous and it seems not to 
follow simple laws. But one can at least recognize that at slight 
elevations where the action of the sun is most energetic the layers 
of air experience equal variations of temperature, while it is very 
probable that in somewhat elevated strata the variation is slight, 
and that whatever may be the temperature at the surface of the earth, 
we shall always arrive at the same temperature at a certain height 
which will however oscillate slightly. 

We shall apply some approximate formulas. The most simple 
hypothesis is that the temperature decreases proportionally to the 
height; then we have 

t = t - az\ 

where z is the height; a, a constant and r the temperature at 
the surface of the earth. In some problems it will be more con- 
venient to introduce the above described virtual temperature and 

T = T - a z. 

These two formulae apply only to small heights; if we wish to cal- 
culate the variation of temperature for the greatest height, we can 
divide the whole elevation into layers and apply the formula to 
each stratum with different values of a. 


§ 4. Variation of pressure with altitude 

According to the theory of the equilibrium of fluids the increase 
of pressure per unit of length is equal to the force which acts on 
the unit of volume. Let g designate the force of gravity per unit 
of mass and z ttfe height, we shall have: 


T*=-s» (1) 

Introducing the value of p given by equation (6) of § i we shall 


dp gdz 

— =-*—- (2) 

p aT 

(i) The virtual temperature remains constant. 

In this case designating by p the pressure at the surface of the 
earth we shall find by integration. 

J± (3) 

p = p e T ° 

in which e is the base of the system of Napierian logarithms. 

(2) The virtual temperature decreases proportionally to the height. 

By introducing T = T — <xz in equation (2) we shall find, by 

writing m =— - — 

T =T - — (4) 


z =-m(T - T) (5) 

/-(£)" (6) 

M>- S \Y m 

p \ ami J 

As to the variation of the pressure of the vapor of water we can 
adopt various hypotheses. We shall consider only the following 
formula : 

j-gy (8) 



VOL. 51 

In which ft is a constant whose value depends on circumstances. 
Knowing p and / we shall find the temperature z by formula (5) 
of §i, namely 

e - 1 f\ 

-T~p)- 273 W 

T = T 1 - 

m = 3.441 4.0 


Variation of temperature \ gg4 Q0 ^ Q0 ^ QO ^ 

per ioo m J 

Height for a variation of \ 1Qlm n7 , n 146m ^ 

i°C. J 

z = 




1000 m 4000 ra 10000 r 

T = 273 m = 3.5 

0.88038 0.58259 0.21250 

To = 273 m = 5 

0.88094 0.59013 0.23681 

To =273 m = 10 

0.S816G 0.59841 0.26266 





Dr. Julius Hann has published a series of observations on the 
tension of aqueous vapor at different altitudes (see the Zeitschrift 
der Oesterreichischen Gesellschaft fur Meteorologie, 1874, page 195). 
In this case applying formula (8) we shall assume 

r = 20°, f - 10 mm , m = 10 and /? = 3 

The values of calculated for these constants are found in the 

third line of the following table. 

Assuming T = constant = 273 and /? = 3 we find the values 
written in the fourth line. The observed values are found in the 
second horizontal line. 



Altitude. English feet. 

Observed f/f 

Computed f/f 

Computed f/f 




















0. II 







0. 10 




§ 5. Expansion and contraction of the air 

The pressure and temperature of a mass of air that experiences 
any transformations whatsoever depend on the quantity of heat 
which it has gained or lost. We will first consider the case in which 
the air experiences a series of transformations without gaining or 
losing heat at any moment. The equation between the pressure 
and the volume represents a line that has been called the adiabatic 

In the study of meteorology it is also necessary to find the equa- 
tion between the pressure and the temperature. It is necessary to 
distinguish between several cases. The air can be dry or moist, 
and the aqueous vapor water can remain without condensation or 
it can pass into the liquid state or into the solid state. 

Representing by U the internal energy of a mixture; by V its 
volume and by A the mechanical equivalent of heat, we have 

0=d U + A pdV (1) 

(1) Dry air. 

Applying equation (1) to dry air we shall find from the mechan- 
ical theory of heat 

p /273 + r 
p = \273 +T 0/ 


m = 

A a 


where c represents the specific heat of dry air at constant pres- 
sure, whence we have, 

m = 3.441 

(2) Moist air without condensation. 

Supposing Ave have one kilogram of dry air and x kilograms of 
aqueous vapor we shall find 

1 + 2.023.* N 

m = 3.441 

1 + e x 



■where 2.023 is the ratio between the specific heat of the vapor of 
water and that of dry air. These formulas apply only in so far 
as the air is not saturated with the vapor of water. At the moment 
when the air becomes saturated, the decrease of temperature is 
accompanied by a condensation of vapor and it is necessary to dis- 
tinguish between the three cases. We will assume that the con- 
densed vapor remains suspended in the mass of air under considera- 

(3) The vapor of water is partially transformed into water. 

We will consider a mixture consisting of 1 kilogram of dry air, 
x kilograms of vapor of water, and y kilograms of water. Express- 
ing by U', U", and U'" the energies of dry air, of the vapor of water, 
and of water respectively we have the total energy of the mixture 

U = U' + xU" + yU'" 

The sum of x and y remains constant and writing 

x + y = £ 
we shall find 

d U = dU' + £ d W" + d(x{U" - £/'")) 

Designating by z/ and v" the specific volumes of the dry air 
and of the vapor of water and neglecting the volume of the water, 
we have the volume of the mixture 

V = v' + x v" 
We can then write 

p d V = (p - f) dv' + fd (x v") 

Expressing the latent heat of vaporization by /, and the specific 
heats of dry air and of water by c and c' we have approximately, 
neglecting the volume of water: 

I = {IJ" - U"') + Afv" 

(273 + r) d 


= d [x (U" - £/'")] + A fd (x v") 


A a 
cdx =d U' + - - dx 

c'dr = d U'" 

(p-f)V =- (273 + r) 

/ =606.5 - 0.696 t 


Substituting the values of d U and of p d V in equation (i)and 
introducing the values given by the equations mentioned, we shall 

= Cj d x 4- £ c' d x + (273 + x) d I 2? * + y ) 
.4 a rf (p - f) 

-T^^H (5) 

Expressing the initial values by the subscript index o, we shall 
by integration and introducing numerical values find 

/ Po-fo\ f 273 + r ] 

log ( f^f) = 3.341 [1 + 4210 £] log [ —- \ 

1 + r ~ 273 + x J ' * • 

+ 6.291 




From equation (7) of §1 we have 

* - ; ■ r^i (7) 

£ p-f 

Equations (5) and (6) apply in general, so long as the tempera 
ture remains above zero. The temperature being at zero the water 
is changed into ice. However, we can imagine the possibility of the 
vapor of water being changed into water at temperatures above 
zero. We know this phenomenon in physics; it is not water only, 
but several salts which present the phenomenon of super-saturation. 
This passage from the state of vapor to the liquid state at tempera- 
tures above the point of congealing involves a state of unstable 
equilibrium and the introduction of a crystal of ice makes the whole 
mass pass suddenly into a solid state. It is probable that this state 
of unstable equilibrium is intimately connected with the formation 
of hail. In ordinary cases congelation commences at the tempera- 
ture zero and we will now consider the passage from the liquid state 
to the solid state at zero. 

(4) Congelation at o°. 

During this stage the temperature remains constant, the water is 
transformed for the most part into ice, but a part of the water is 
vaporized because according to equation (7), section (1), any dimi- 
nution of the pressure produced by dilatation demands a greater 
quantity of vapor of water for the same vapor tension. 


Consider a mixture containing i kilogram of dry air, x kilograms 
of the vapor of water, y kilograms of water and z kilograms of ice. 
The sum x + y + z remains constant and we put 

x + y + z = £ 

Denoting by U" " the energy or specific heat of the ice and by L 
the heat of fusion of the ice, we have 

U = U' + x U" + y U'" + z U"" 
- U' + f U" + x (U" - U'") -z (U'" - U"") 

The temperature remaining constant, we have 

d U = (U" - U'") d x - (£/'" - U"") d z 
We can neglect the volumes of the water and the ice and put 

V = x v" 
p d V = (p - f) v" d x + / v" d x 

From the mechanical theory of heat we have approximately 

I = U" - U'" + A f v" 
L = U'" - U"" 

By the aid of equation (i) we find 

= / d x - L d z + A (p - f) v" d x 

Introducing from equation (7) the value of p — f and observing that 
at the temperature of zero we have 

e a 
f v" = - 273 
we shall find 

A a d x 

= Idx - Ldz + — 273 (8) 

g ^ 

At the commencement we have 

x = x , y = y , z = 
and when all the water has disappeared (by congelation) we have 

x = x; ^ = 0; z = x + y — x 
By integration and substituting / = 606.5 arR l L = 79-°6 we have 

log - = 1.822 7 - 15.80 (x - x ) (9) 



Having determined x, we shall rind p by equation (7). 

When all the water is transformed into ice and vapor, we have 
a mixture of vapor of water and of ice and from this moment on- 
ward the vapor of water is transformed directly into ice by the 
lowering of the temperature. 

(5) The aqueous vapor is partially transformed into ice. 

For this stage we will apply the formulas given in case (3) sub- 
stituting / + L for I and the specific heat of ice (c" = 0.5) for the 
specific heat of water. 

'og(^) =3.441 (1 + 2 ..05Olo g (f|^) 

fx (l + L) x(l + L) \ 

We have supposed that the water and the ice remain suspended 
in the air and take part in the thermal phenomena during the three 
periods in which the vapors of water are condensed. If we wish to 
consider the case in which the water and the ice after their formation 
separate from the mass of air, it will be necessary to consider the 
term c + £ c' or c + £ c" as variable, We can in this case give to 
£ a mean value and consider it as constant, since its value is very 
small. In this case the period of freezing at o° disappears. 

M. Peslin 1 has developed similar formulae, but he has not considered 
the variation with temperature of the latent heat of vaporization. 
This causes the difference between his formulae and ours. 

We shall apply our formulas to the case in which a mass of air rises 
in the atmosphere with a velocity so small that it can be neglected. 
Designating the height by h we can write the equation of equilib- 
rium from §4, 

V d p = - (1 + $) dh 

Combining this equation with equation (1) we find 

= d U + A d (p V) + A(i + £) dh . . . . (11) 

Applying this formula to moist air we find the formulas of section 4. 
When we consider the cases in which the vapor is condensed, we 
distinguish the following: 

x The bulletin hebdomadaire de l'association scientifique de France, No. 


(6) The vapor of water is transformed partially into water. 
Approximately we have 

a (273 + r) 
p V = (p- f)v' + f v" x = + x f v" 

By the substitution of this equation in equation (11) and by the 
aid of the formulas of case No. 3. we find 



= (c -f £ J) dx + d (x I) + A (I + £) d h 

= c + £ c') (r -t ) + x I- x l + A(l + £) h 

(7) Congelation at o°. 
In this case we find 

= Idx -Ldz + A(z + $)dh } 

= (/ + L) (x - x ) - Ly + .4(i + t)h j 

(8) The vapor of water is partially transformed into ice. 
We find 

= (c + £ c") (r - t ) + x(l + L) - x (l + L) +A(i + £)* (14) 

All these formulas tliat we have developed, pertain to the case 
in which the air experiences transformations without gaining or 
losing heat. Let us suppose that the air receives some heat and 
that the heat absorbed is proportional to the variation of the tem- 
perature. When the quantity of absorbed heat is small, the tem- 
perature decreases during the expansion and in place of equation 
(1) we write 

- bdz = d U + A pdV (15) 

Here b expresses a constant which depends on exterior circumstan- 
ces. Applying this equation to the dry air we find the law given by 
formula (2) and m is given by the equation 

m = 3.441 ( 1 +- ) 06) 

We see that we can easily apply equation (15) to other cases in 
which the vapor is condensed, but we refrain from the development 
of the formulas because there are no observations wherewith to 



check the results. Besides it is more convenient to apply the 

p I T 



and to attribute to m suitable values, which vary with the height 
of the layer of air. We shall consider then formula (17) as the 
general formula, when the air experiences a series of transformations. 
By the aid of formula (6) from § 1 we shall find 




consequently we can write 

dp J p \ 1 

= ma \ — J = - 

P \p I Po 




C P dp / p p\ \/p\ 

I = m I - — ) = m a 1 I I 

J v„ p \p Po/ L v p» 1 


• (20) 

By integration we find 

ip dp [p p 

v„ P \ P Po 

We shall make use later of formulae (19) and (20). 


Let us apply our formula to a mass of air that rises slowly in the 

p c = 760 mm ;/ = 15 mm ;r = 20° 

(1) When the air is not saturated. 
By the aid of formula (7) of § 1 we find 

x = 0.0125. 
Substituting this value of x in § 5 formula (4) we find 

m = 3.46. 
So long as the air is not saturated, the value of x remains constant 
and consequently the ratio - becomes constant. We have then 

P f 

Po fo \ 273 + r 

273 + t 



VOL. sr 

Attributing to x different values we find the point of saturation by- 
comparison with the following table, which contains the values of 
t corresponding to the maximum tensions of aqueous vapor. 


We can adopt — = 0.29 approximately and calculate a table of 

the quantity v determined by the formula 

273 +r 
v = _i 

f m 

We shall find 

v - 133.6; / - 14.4 mm ; r = 17°; p = 733.4 ram ; h = 306 m 









o. 39 mm 



4 . 6o mm 






6.5 3 

161. 3 

— 20 


259. 2 


9. 17 



1 .40 



12. 70 


— 10 






- S 

3. 11 




119. 2 

4- 60 




in. 4 

(2) The air is saturated above o°. 

By substituting in equation (6) the value of x from equation (7) 
and assuming 

$ = 0.0125 = * ; t - 17°; r = 0°; / 0= 14.4 mm , 
p = 733.4 m and / = 4.60 mra 

we shall find 

log (p-f) - 2.6005 + ■— 


p = 487. 2 mm ; x = 0.00594 

Formula (12) gives h = 3384 meters. 

If we had used formula (17) we should have given m the value 

(3) Water freezing to ice at o°. 

By substituting in equation (9) the values 
f - 0.0125, x = 0.00594, y = £ - x = 0.00656, f = f = 4-6 mm 
we shall find 

* - 0.00607 kg. and from formula (7) p = 476. 5 m,n 

Formula (13) gives h = 178 meters. 


(4) The air is saturated below 0°. • 

By substituting in formula (10) 

To = o, t - - 20°, /„ = 4.6 mm , / = 0.93 mm , p - 476.5 r 
f = 0.0125, < = 0.00607 

we shall find 


log (/>-/) =2.4613+ ^ 

/> = 312.5 mm , x = 0.00186 

Formula (14) gives h — 3239 meters. 

If we had used formula (17) we should have given to m the value 


Resulting Values 












760. O m 

2 0° 


0. oi25o 












• 476. 5 



0. 00607 



— 20 


0. 00186 

§6. Causes of the movement of the air 

When the air is in equilibrium, the active forces are the attraction 
of the earth and the centrifugal force produced by the rotation of 
the earth. These two forces have a resultant which we call the 
weight, which varies with the latitude and the altitude. In meteor- 
ology we consider only the strata of air at a slight elevation and we 
generally consider the weight constant and express its value for 
the unit of mass by g. 

In an atmosphere in equilibrium the weight is normal to the level 
surfaces, and the surface of the earth is itself a level surface. At 
the same time the density of the air and consequently its tempera- 
ture vary only from one level surface to the other; we see then that 
in the state of equilibrium the level surfaces are surfaces of equal 
pressure and isothermal surfaces. We can consider the earth as ap- 
proximately spherical, and the level surfaces as spheres concentric 
with the earth. Then the temperature can vary only with the 


altitude. In truth the temperature will never be uniform on the 
earth and equilibrium has therefore no place in nature. 

When the atmosphere is in equilibrium the law of variation of 
the temperature with the altitude has no influence on the equi- 
librium, but the stability of the equilibrium does depend on that 
law. It is necessary to distinguish between stability with reference 
to an ascending movement and to a descending movement. In 
giving to a particle of air an ascending motion the temperature of 
the particle of air can change more rapidly or more slowly than the 
variation of the temperature of the surrounding air. If the tem- 
perature of the ascending particle decreases more rapidly than the 
temperature of the atmosphere, the particle will acquire a specific 
weight greater than the surrounding air, and consequently it will 
descend when the impressed motion is consumed, and we call the 
equilibrium stable. If the temperature of the particle of air decreases 
more slowly than that of the atmosphere, the particle will attain a 
specific weight less than the surrounding air and it will continue its 
ascending movement; then the equilibrium is unstable. 

By impressing upon a particle of air a descending velocity we see 
in the same way that the equilibrium is stable if the temperature 
of the particle is increased more rapidly than that of the surround- 
ing air and that it is unstable if the temperature of the particle of air 
increases more slowly than that of the surrounding air. 

The stability of the atmosphere depends consequently on the 
law of the variation of the temperature of the atmosphere with 
the height. 

Let us suppose that in a calm atmosphere the virtual temperature 
decreases proportionally to the altitude according to the formulas 
of §4; by impressing a slight velocity upon a particle of air we can 
calculate approximately the variation of its virtual temperature 
from formula (17) of §5. Let m be the coefficient of the particle 
of air and m' that of the calm atmosphere, we see 5 that the equi- 
librium is stable for an ascending movement when m < m'. 

The general cause of the disturbances of the equilibrium of the 
atmosphere is the heat from the sun. The sun communicates heat 
to the atmosphere both directly and by the intervention of the 
surface of the earth indirectly. This quantity of heat represents 

Lei T and T' be the virtual temperatures of the particle of air and of 
the calm atmosphere respectively we have 

PS sz 

T = T - — T = r u - — 

a m a m . 


an active force which can produce the movements of the particles 
of air. The action of the heat of the sun is presented under two 
different forms; on the one hand it produces the changes of the 
temperature of the atmosphere and on the other hand by the evap- 
oration of wat^r it produces changes in the mass of the atmosphere. 
The direct action of these phenomena is to produce changes in the 
pressure of the air accompanied by movements of the particles of air 
which give rise to the currents of air. 

The currents of air, which can have any direction whatever, tend 
always to destroy the perturbations and to produce a new state of 
equilibrium. We can imagine permanent currents in the atmos- 
phere; let us suppose that a continuous heating takes place at one 
point and that a cooling takes place at another, we see that there 
will arise two currents, one carrying warm air and the other cold 

The heat set free in the atmosphere in any way whatever, produces 
currents of air. We notice the currents of air during a forest fire, 
and during the eruption of a volcano. In the last case the vapors 
and the shower of ashes set free the heat. We take this occasion to 
remark that during the eruptions of volcanoes and during earth- 
quakes it is probable that masses in the interior of the earth change 
their positions. If the masses are great enough to influence local 
gravitation we can explain the formation of the currents of air, by 
supposing that the displacement of the masses in the interior of the 
earth produces a sudden change in the force of gravity. This 
change will be accompanied by a rapid change in the pressure of the 
air which will produce currents of air. 

Chapter II 


§7. Isobars and gradients 

During the movement of the air certain new forces come in play 
and the level surfaces of § 6 are no longer surfaces of equal pres- 
sure. Let us consider a surface of equal pressure, or an isobaric 
surface, during such movement; this surface cuts the level surfaces 
into lines that are called isobaric. We shall occupy ourselves here 
principally with the isobars at the surface of the earth. 

In considering the variation per unit of length of the pressure at 
any point we perceive that the variation along the isobar is noth- 
ing and that the variation has its maximum value along the normal 
to the isobar. 



VOL. 51 

We shall find the variation of the pressure in any other direction 
whatever by projecting the maximum variation upon that direction, 
which projection is geometrically represented by the chord of a 
circle whose diameter is the maximum variation. See the fig. No. 1, 
in which / / represents the isobar, N the maximum variation, and 
O P the variation of the pressure in the direction O P. 

In meteorology we call the gradient the variation of the pressure 
normal to the isobar and expressed in millimeters of mercury per 
degree of a great circle. 

Let G be the gradient, d p infinitely slight increase of pressure, 
d 11 an infinitely slight increase of the normal, /j. a constant, we have 


,— = /j.G 






10 000 000 

= 0.00012237 


The direction of the gradient is shown by that point of the com- 
pass toward which the pressure is least. From the theory of fluids 

it is evident that the quantity - G represents the force produced by 

the variation of the pressure acting on the unit of mass. This 
force which acts in the direction toward which the pressure dimin- 
ishes, must be added to the exterior forces. We shall see later that 
the gradient is a small magnitude which even in cyclones does not 
exceed ioo mm . 

Terrestrial gravity g is equivalent at the surface of the earth to a 
gradient of 10570 mm. 


§8. Forces which act during the motion 

During the motion of the air there are two new forces that come 
into action, namely, the action of the rotation of the earth and the 
friction between the molecules of air both between themselves and 
on the surface* of the earth. The action of the rotation of the earth 
produces properly speaking two forces, the centrifugal force, which 
with the attraction of the earth produces the resultant g and the 
force called the composite centrifugal. This latter force which we 
shall call the deflecting force is perpendicular to the trajectory of 
the particle of air, and is directed to the right in the northern hemi- 
sphere and to the left in the southern hemisphere. 

Expressing by v the velocity of the air, by co the angular velocity 
of the earth and by 6 the latitude, we have the deflecting force 

= 2 co v sin 6 (1) 

The velocity is expressed in meters per second and 

°>-mk- oom72g2 

The deflecting force of the rotation of the earth is found by con- 
sidering the movement of a point relative to the earth, which is sup- 
posed to be at rest. If we do not introduce this force in all the 
dynamic problems that introduce movements relative to the earth, 
it is because this deflecting force is very feeble and the trajectories 
do not extend to considerable distances. On the contrary the 
currents of air travel over large parts of the surface of the earth, 
and the forces which produce them are very feeble. We may then 
anticipate that the deflecting force of the rotation of the earth 
plays an important part in the problems of meteorology. Let us 
add that this force being perpendicular to the trajectory has no 
influence on the velocity of the current, but tends only to change its 

On the contrary, friction is a force that tends to diminish the vel- 
ocity. The complete theory of the friction between the molecules 
of air is very complicated and will be developed in the second part 
of these studies. 

For the present Ave admit that friction is a tangential force and 
opposed to the motion. As to its magnitude we will suppose that 
it is proportional to the velocity, and expressing by k the coefficient 
of the friction, we write 

the force of friction = k v (2) 


The complete theory shows thai the value of k depends on the 
height of the current. When the height of the current increases, 
the value of k diminishes, which conforms with what we know of 
the coefficient of friction of water in open channels. For very 
broad channels the coefficient of friction is in the inverse ratio of 
the height of the current. 

In studying the movement of a particle of air it is necessary to 
add to the exterior forces the tangential forces and the centrifugal 
force produced by the motion. Expressing by 5 the distance trav- 
eled over and by R the radius of curvature of the trajectory, we 

vd v 
the tangential force = —z — (3) 

v 2 
the centrifugal force = — (4) 

Let us add that the horizontal currents move along the surface 
of the earth which is normal to gravity. Consequently we neglect 
the action of the gravity in the following problems and the acting 
forces will be (1) the gradient force, (2) the defective force of the rota- 
tion of the earth, (3) the force of friction, (4) the tangential force 
of the motion and (5) the centrifugal force of the motion. 

§9. Horizontal rectilinear and uniform motion 

When the motion is uniform and rectilinear, the tangential force 
and the centrifugal force disappear, and equilibrium is established 
between the force of the gradient, the force of friction and the deflect- 
ing force of the rotation of the earth. 

Expressing by a the angle between the gradient and the trajec- 
tory and resolving the forces along the trajectory A B, fig. 2, and 
perpendicularly to its direction we have 


- G cos a = k v (1) 



- G sin a = 2 co sin 8 v (2) 


By division we obtain 

2 oj sin 8 
tan a = 7 (3) 



In the northern hemisphere the latitude 6 is positive and in the 
southern hemisphere 6 is negative. The angle a has the same 
sign as and consequently the wind is deflected to the right in the 
northern hemisphere and to the left in the southern hemisphere. 

FIG 'i 



VOL. 51 

The ratio between the velocity and the gradient is found by equa 
tion (i) 

v fi cos a 



By formulae (3) and (4) we can determine the direction and the 
velocity of the wind : by graphical construction we find their values 
in the following manner. 

Let B, fig. 3, be the direction of the gradient and let two dis- 
tances O B and O A be in the ratio. 

OB :0 A = 2<o : k 

Describe a circle with the radius O B and erect A L perpendicular 
to O A : making the arc B D equal to the latitude and draw D N 
parallel to B O; we shall have the angle A O N = a. Describe a 


semicircle with the diameter C = — and the chord O M will be 

p k 

equal to the velocity v. 

We shall call the angle a, determined by equation (3), the normal 

angle of inclination. 


of the normal angle 

f inclination a 




0. 00004 

























14. 2 

II. 9 


62. 1 


32. 2 

2S. 3 

20. 7 











61 . 2 



36. 1 




















57- 7 

Si. 6 















5o. 1 








The relation between the velocity and the gradient depends on 
the density of the air. Supposing the temperature to be o° and the 
pressure 76o mm the density p is 



= 0.13184 



and we find the following values for the latitude 45 : 

0. 00002 





0. 00004 




O. 119 





0. 129 




7. 11 

0. 141 





0. i55 





0. 170 


v = velocity in meters per second, 

G = gradient in millimeters of mercury per degree of a great 
circle on the earth's surface. 

Supposing the temperature to be 20 and the pressure 74o mm the 

value of — will be increased in the ratio of 1 to 1.102. Supposingthe 

G v 

temperature to be — io° and the pressure 770™™ the value of -~ 

will be diminished in the ratio of 1 to 0.951. 

By the aid of the last table we have calculated the following table 
for the latitude of 45 , the velocities being always expressed in 
meters per second. The scale of wind force is that hitherto employed 
in several meteorological systems of Europe; the numbers have 
the following signification: 

o = calm; 1 = feeble; 2 = moderate; 3 = quite strong; 4 = 
strong; 5 = very strong or storm; 6 = a hurricane. The values 
of the coefficient of friction are the extreme limits which we have 
found by preliminary calculations, for the ocean and for the irregu- 
lar surface of the earth. 

SCALE 0—6 



fc = 0.00002 

fc = O.O0OI2 

2 = moderate. . . . 

3 = quite strong 

4 = strong 

6 = hurricane. . . . 

1 — 4 

4— 7 

7— 11 

11 — 17 

17 — 28 

28 — 5o 

O. I 

0. I — 0. 5 

0.5 0.8 

0.8 I. 2 

1. 2 1.9 

1.9—3. I 

O O 

O. 2 O 

0. 7 1 

I . 2 1 

I . 9 — 2 
2.9 — 4 
4.8— s 






VOL. 51 

§10. Horizontal currents of air with rectilinear isobars — the 
latitude is supposed to be constant 

When a permanent current of air flows over rectilinear isobars 
the mass of air that flows perpendicularly to the isobars in the unit 
of time must be constant. Expressing by <p the angle between the 
gradient and the tangent to the trajectory then the equation of 
continuity becomes 

v cos (/> = constant 


Decomposing the five forces (see §8) in the directions of the 
tangent and the normal we shall have 

- G cos <p = k v + v — 
o ds 

u v 2, 

-G sin 6 = 2 a) v sin 6 -\- 



2 ojy vJW, Q+ -£ 

FIG 4 

Differentiating equation (1) we have, 

d v = tan (J/ d </> 
and introducing the value of the radius of curvature 

1 <f<p 

R = ~ ~ds 

we shall have 

fl d (/> 

- G cos 6 = v (k + v tan d> , 
p T as 


G sin </» = v (2 co sin 8 


d s 



These two equations are transformed into the following 

H I 2 co sin Q \ 

- G = k v cos <p I 1 + t — ~ tan cp J (6) 

v d <p 

= k sin <p — 2 co sin 8 cos d> + -z~ . . . . (7) 

cos (p a s 


cos cp d s = d x 

where x is the distance along the gradient and s along the trajectory, 
equation (7) can be written 

v cos <!> d (tan d>) 2 co sin S 

The general integral of this equation is 

2 00 sin 6 ~ kx , 
tan 4> = + Ce » cos ^ 


where C is the arbitrary constant. In nature it is necessary to place 
C = o because the angle <p does not increase to infinity with increas- 
ing values of x. 

Thus we have (as in §9, eq. 3) 

2 co sin B 
tan (p = = tan a 

Substituting this value of cp in equation (1) we see that the velocity 
becomes constant and consequently according to equation (2) the 
gradient likewise becomes constant, provided that we suppose the 
density p constant, and in this case the isobars are equidistant. 
If we wish to consider the variation of the density p we introduce 

d p = — [iG d x 

and by the aid of formula (20) from §5 we can calculate the pres- 
sure p. In general we can introduce a mean value of the density 
and consider it as constant. 



VOL. 51 

§11. Influence of the variation of latitude on horizontal currents 
of air with rectilinear isobars 

We consider only the case in which the gradient coincides with a 
meridian. The latitude 8 is expressed by the following equation 

e = e a + x x 

x = ± - - . — 

10 6 180 



FIG. 5 

The coefficient X is positive when the gradient is directed toward 
the north and negative when it is directed toward the south. 

The equations developed in §10 now hold good by considering 
6 as variable. 

Equation (8) of §10 becomes 

d (tan (/>) 2 to sin — k tan if> 
dx v cos c> 

For the sake of abbreviation we write 

V COS (j) 

tau e = X 



Placing the arbitrary constant equal to zero, as we have done in 
§10, we find that the integral of (3) is 

2 w 

tan d) = ~ - cos e sin (6 — s) 



Substituting this value of tan </> in equation (6) of §10 we shall 

^ G = k v cos 4> I 1 + ( — V cos e sin 8 sin (8 - e) ) . . (6) 

In order to obtain the equation of the trajectory we introduce 

d y = tan <f> d x 
Substituting herein the value of tan (p and integrating, assuming 
y = o for x = o, we get 

4 oj cos s , / 8 — 8 \ ( 8 — 8 Q 

y = — . — stn[ - a[- 

Introducing [xG = — -j- into equation (6) Ave shall by integration find 

the pressure p. Finally by graphically constructing the curve of the 
gradient we shall easily determine the curve of the pressure. 

According to equation (5) the angle of inclination <p depends on 
the quantity e : e being so small that cos e does not differ sensibly 
from unity, we conclude that ^ approaches the normal angle a, 
when the latitude 8 has a large enough numerical value. As for 
the winds that cross the equator, 8 has small numerical values, and 
the angle </> can be very different from the normal angle of inclina- 
tion. It is necessary to distinguish two cases: 

(1) The gradient is directed towards the north. 

In this case X and s are positive; the angle </> is negative for south- 
ern values of 8 and for northern values until 8 = e. When 8 is 
greater than e, the angle </> becomes positive and approaches more 
and more to the normal value a. We see then that the winds that 
come from the south have turned to the left even after crossing the 
equator, that the deviation is nothing at the latitude £ and that 
beyond this point the deviation is to the right. 

We recognize this law of deviation in nature in the trade winds of 
the Atlantic and of the Indian Oceans during the summer. 

(2) The gradient is directed towards the south. 

The value of X and that of e become negative. In this case the 
angle <l> remains positive north of the equator and also south of the 
equator to latitude 8 = e; then (p becomes negative and approaches 
more and more nearly the normal value a. 

We see then that the winds that come from the north have devi- 
ated to the right even after crossing the equator and until 8 = e; 



VOL. 51 

at this latitude the deviation is nothing, then it turns to the left. 
In nature, the monsoon called the west monsoon in the Indian 
Ocean follows this law during the winter. 

We will apply our formulae to numerical examples that we can 
compare with the charts of general winds. Among these we mention 
especially the excellent charts published by the Meteorological 
Office at London, under the title "Monthly Charts of Meteorological 
Data for Square 3 : published by authority of the Meteorological 
Committee." These are in fact those charts that have led us to 
establish the theory of winds crossing the equator presented in this 
paragraph. We have developed the preceding formulae by sup- 
posing that the gradient coincides with the meridian; approximately 
we may apply them to the cases in which the angle between the 
meridian and the gradient is small. In the first example that we 
shall compute we assume this angle to be 20 as we see it in fig. 6. 


(1) The gradient is northerly {see fig. 6). 

0o - 0°, 
v cos (/> — 10 m , 

k = 0.00002, 
x = 20°; 
we shall find 

e = 4° 13'.2 
and the following values: 








- 5° 

— 5o. 2 




— 29.6 

11. 5 

0. 20 




0. 21 

- 1.3 



2. 2 



(2) The gradient is southerly {see fig. 7). 

6>„ = 0°, 
v cos <p = 5 m , 

k = 0.00002, 
t = 20°; 
we shall find 

e = -2° 15 ; 
and the following values: 


I5 1 











0. 16 





0. 10 


- 5 

— 19. 2 


0. 12 

0. 2 

— IO 

- 44-4 


0. 22 









S° ^ 2° 0° Z° ¥■' 6° 8° /<?". 











,4 r# 





■° ¥° 2° O" 2° *° 6" 8° A 


FIG. 6 

Jf" ^° 2° 0° 2" v° 6" 8° /o;„ 




















[ / 










¥" 2° O" 2° ¥° 6° 8" / 


FIG. 7 


We shall see that the law of inclination 'eq. 5) is quite conform- 
able to the observations; however the observed velocities and grad- 
ients do not follow our formulae always, which is very easily ex- 
plained when we note that in nature the currents of air near the 
equatorial calms have an ascending movement that diminishes the 
horizontal velocity and the magnitude of the gradient. We could 
easily introduce this influence into the formulas, but we shall not 
profitably extend our researches any further since we shall treat 
the problem in a more general manner in the second part of these 

§12. Horizontal currents of air with circular isobars around a 
barometric minimum. 

We shall consider the latitude as constant 8 and the isobars as 
concentric circles. The system being symmetrical with respect to 
the center of the isobars therefore the quantity of air that enters 
per unit of time must remain constant. Designating by^ the angle 
between the direction of the wind and the radius, which latter 
is the direction of the gradient, the component of the velocity in 
the direction of the radius will be v cos <p. Let r be the radius and 
h the altitude of the horizontal current, the section of the current 
will be 27r r h, and remarking that h remains constant the equation 
of continuity will give 

vr cos = constant (1) 

The acting forces are the same as in §10 and the equations 
(2) and (^ hold good by substituting 

1 sin (f) d </> 

— = - - + cos 4> 
R r dr 

cos <p d s = — dr 
By the aid of equation (1) we shall find 

fx I v cos <j> v sin <j> d <J> 

- G COS d> = v \ k + — 

p \ r 

d r 

pt ' I v sin <l> v cos </> d <j) 

G sin d) = v I 2 to sin 6 + + 

p r d r 

8 That is, uniform over the whole barometric depression. — Editor 


These two equations are transformed into 

- G = v ( k cos (p -f- 2 a) sin 6 sin (J) -\- - 

= k sin <j) — 2 w sin 6 cos ^ — i> 



The last equation can be written 
d (tan 0) & 

r dr 

tan ^ 


sin 6 

v r cos ^ 

By integration making the arbitrary constant equal to zero for 
the same reason as in §io we obtain 

tan d> = — sin 6 tan a 


FIG. 8 

The angle of inclination has therefore the normal value and the 
trajectory is a logarithmic spiral. Designating by <p the angle be- 
tween the radius and some fixed direction, the equation of the tra- 
jectory will be 

log nat r = — <p cot a -\- C (5) 

Let r and v be the values of r and of v for any point whatever, 
we can transform equations (i) and (2) into the following: 

v r = v >'„ (6) 


G- - + *- 

p cos a r 

^G = kV ° r ° l \- V ° r ° 2 



cos a r 


By introducing a mean value of the density and by expressing 
the distance r in degrees of the meridian, we can write 

a a' 

G = - + - (8) 

r r v 

in which a and a' are constants. Then the increase d b of the pres- 
sure in millimeters is equal to G d r and we shall find 

a r a' / 1 1 \ 

6 - 6 °=m 1o s-7 -2(7>~v) (9) 

The equations that we have developed demand that the altitude 
of the current of air remains constant since we assume its horizon- 
tality. We can then, therefore, only apply the equations to the 
exterior parts of a whirl about a barometric depression, for in the 
interior of the whirlwind the currents have an ascending movement 
so rapid that we cannot neglect it. 


(1) Whirlwind having a great velocity (see fig. 9). 
Let the latitude = 20°, 

k = 0.00002, 
r = 20°, 

= 0.001006 (for a mean pressure of 753 mm ) 

t; = 50 m , 
r = 0°.3. 

Expressing r in degrees of the meridian we have 

0.8 2.014 r 1.007 

G =— +— ;&-6 = 1.842 log- + 11.19- — 

<p =328 T °; a =68° 5.8' 

r = 0°.3 0.°4 0°.5 

7/ = 50 m 37. 5 m 30 m 

G =77. 3 mm 33.5 mm 17.7 mm 
6-6 o = mm 5.1 mm 7.6 mm 

-y>=0° 41° 73° 99° 140° 172* 340° 





25 m 

18.75 m 

15 m 

7.5 m 

10.6 mm 

4 gmm 

2.8 mm 



10.4 mm 






(2) Whirlwind having an average velocity (see fig. 10.) 

Let the latitude = 60°, 

k = 0.00012 
v = 15 m , 

r =7° 

- =0.0009404 (r = 0°, b = 750 m "'). 

We shall find 

19.43 10 5.5 
a == 40° 23.3'; G -- + ~j- 

r r 

r ">2.75 r n 

b-b = 44.73 log - + 1 . 070 - ; <p = 138 . 5 log . — • 
r r- T r 

r = 7° 8° 9° 10° 12° 15° 20° 

v = 15 ,n 13. l m 11. 7 m 10. 5 m 8.7 7.0 m 5.2 m 

G = 3.08 lnm 2.64 mm 2.30 mm 2.05""" 1.68 mm 1.33 mm 0.98 mm 

b - b Q = mm 2.8 mm 5.3 mm 7.5 m,n 11.2 mm 15.6 mm 21.3 mm 

- <p = 0° 8° 15° 21° 32° 46° 63° 

§13. Horizontal currents of air with circular isobars around a 

barometric maximum 

By making the same hypothesis as in § 12, Ave shall find the same 
equations, but it is necessary to write 

4> = 180° + a (1) 

and to change the sign of the gradient, on the supposition that the 
pressure diminishes with distance from the center. We shall then 

vr = v r 

P- - kv r 1 v 2 r„ 2 . 

u- = . — « (2) 

p cos a r r 


a a 
G ~-r-S < 3) 

a r a' I I i \ 

b >- b =M 1C 'Z U + 2\,<- r.>) ' ' W 


J 57 

i i i i i i i I i i i i i i i i i I i i i i i i i i i I i 
' W O' /o° 



The hypothesis that the pressure diminishes with the distance 
from the center requires that 

v < 

cos a 

In nature the wind about barometric maxima always has a slight 
velocity and the pressure diminishes with distance from the center, 
However, we can apply the formulae only to the exterior parts of 
the whirlwind, because in the interior part the currents are not 
horizontal, but have a vertical descending velocity which influences 
the horizontal movement. 


Whirlwind about a barometric maximum (see fig. u). 
Let the latitude = 45°, 

k = 0.00012, 

- 0.0009281 (t = 0°, b = 760 tni »), v = 4 m , r = 5 C 

we find 

3.406 3.879 
a =40° 35'. 9; G = - - - —j- 

r 1.94 r. 

b - h = 7.841 log - - 0.0776 + — 1~; ? = 113 log.- 

r Q r r 











4 m 

3.3 m 

2.5 m 

2.0 m 

1.3 m 

1.0 m 



0.65 mm 

. 55 mm 


. 34 mm 

0.23 m,n 



- b 



0.57 mm 

1 . 57 mm 

2 . 34 mm 

3 . 73 mm 









§14. Currents of air in the interior part of an atmospheric whirl 

In § 12 and § 13 we have considered currents of air flowing at a 
constant elevation approaching the center of the isobars or moving 
away from it. In nature the elevation of the currents does not re- 
main invariable; in atmospheric whirls around a barometric mini- 
mum the currents have an ascending movement that increases 
toward the center, and in whirls about a barometric maximum the 
currents have a descending movement that diminishes with the 
distance from the center. We shall treat the general problem in 
the second part of these studies, but at present we will consider a 


T I I I I I I i I I I I 



special case, namely, the central part of the whirl. For a system of 
circular isobars the equation of continuity can be written 

2n r h v = constant 

Supposing the height to be variable and a function of r, we can write 

v r cos (p — / (r) 

Assuming the following hypothesis: 

/ (r) = c r 2 

where c is a constant, then the equation of continuity takes the form 

v cos (/> = or 

Introducing this value of v in equations (2) and (3) of §10 and by the 
aid of the formulae of §12, we shall find 



- G cos d; = v ( k — v sin d> — — c ) 

P \ dr J 

u. „ . . /_ . _ v sin </» ,dd>\ 

- G sin <p = v [2 co sm 6 + + v cos cp — l I 

o \ r dr / 

Eliminating G we shall have 


= ^ sin <p — 2 oj sin 6 cos <J> — 2 c sin </' — z; — - . . (4) 


This equation can be written 

cr d{tang ^= (k-2c)tang</j-2cosm8 

By integration placing the arbitrary constant equal to zero we find 

tang0 = (5) 

hY k-2c 

The angle of inclination is therefore constant and the trajectory 
is a logarithmic spiral, but the angle of inclination has a value dif- 
ferent from the normal value a of § 9, eq. 3. We express this 
value by /?, and introducing the value 

tan a = — suit/ 


we shall have 

tang ^ta^« 

X- 2c 

Equations (i) and (2) may be written 

v = .r (7) 

cos /? 
tG = &- c)c r (8) 

p COS 2 /? 

Attributing to p a mean value, we can write 

G = G t r (9) 

in which G x denotes a constant magnitude and r is expressed in 
degrees of the meridian. 
Then we have 

b - b = 1/2 G t r 2 (10) 

in which b is the pressure in millimeters at the center. Thus the 
curve of pressure becomes a parabola. 

The preceding formulas apply to whirls around a barometric 
minimum; by making c negative and substituting (180 + /?) for 
/? we shall have the following formulae which apply to a whirl about 
a barometric maximum: 

tang /? = ^ (11) 

e c „(A±fL'. r (12) 

p COS 2 ft 


(1) Whirlwind of great velocity about a barometric minimum (see 
fig- 9)- 

Consider the central part of the whirlwind No. 1 in §12. 

/? = S9° 45';^ =0.00103; (r =20°, b = 735 mm ) 


we shall find 

v = 252 r; G = 566 .5 r; b - b = 283 r 2 

We apply these formulas to the central part situated between 
the center and r = 0.14 (where the current is one of ascension). 
By constructing the curves of v and G we shall also find by graphic 
interpolation their values for the region between r = 0.14 and r = 
0.3 (where the transition to horizontal motion occurs). The 
results are given in the following table : 


= 0° 





= o m 

25. 2 m 

50. 4 m 

50 . m 


— Qmm 

56.6" 1 " 1 



b - 


— Qmm 


10.0 mm 

18.6 mm 

(2) Whirlwind of mean velocity about a barometric minimum (see 
fig. 10). 

Consider the central part of the whirlwind No. 2 in § 12. 

= 55°; fi = 0.0009596 (r - 0, b = 735" 1 " 1 ) 

we shall find 

v = 3.082 r\ G = 0.583 r; b - b = 0.2915 r 

We apply these formulae to the central part (ascension) situated 
between the center and r = 5 , but we find by graphic interpolation 
the values for the part situated between r— 5 and r = 7 , as fol- 
lows : 

r = 0° 1° 2° 3° 4° 5° G° 7° 

v = m 3.1 m 6.2 m 9.2 m 12. 3 m 15. 4 m 16 m 15 m 

q _ Qmm q 5gmm \ 17mm \ 75 min 2.33 mm 2.91 mm 3 20 mm 3.08 mm 

&-,&„ ="O mm 0.29 mm 1.17 mm 2.62 mm 4.66 mm 7.28 mm 10.38 mm 13.58 mm 


132° 75° 42° 18° 0° 

(3) Whirlwind about a barometric maximum (see fig. 11). 

Consider the central part of the whirlwind of §13. 


p = 35°; 

- = 0.0009281 

we shall find 

v = 1.823 r\ G = 0.32 r; b - b = 0.16 r 2 


Applying these formulae to the central (descending) region sit- 
uated between the center and r = 1.5 and by graphic interpolation 
beyond, we find the following values: 


= 0° 







= m 

l.S m 

3.6 m 

4.6 m 

4.7 m 

4.0' n 


= Qmm 


0.62 mm 

0.68 mm 

0.69 mm 

0.65 mm 


- b 

_ Qmm 

0.16 mm 

0.62 mm 

1 . 25 mm 

1 . 93 mm 

2.60 mm 

By these formulas and examples we see that for a given latitude 
and a given coefficient of friction the whole system of a whirlwind 
is determined by the maximum velocity and the distance from the 
center of the movement of the point where that velocity is found. 

Chapter III 


§15. Rectilinear movement 

We consider a particle of air moving in the direction of the vertical 
axis of z which we suppose positive upward. We neglect the action 
of the rotation of the earth and the viscosity or resistance between 
the molecules of air. Then we have three forces, namely: the 
force produced by the variation of the pressure 

1 dp 

p d z' 

the force of the weight g, and the tangential force 

d z 

where w represents the vertical velocity. 

The equilibrium between these three forces is given by 

1 dp dw 

--JT = -g-W--j (1) 

p dz dz 

Expressing by p and w the values of p and of w at any point, the 
equation of continuity becomes 

p w = p w (2) 

This equation demands that the section of the current be of con- 
stant area. The above equations hold good for ascending move- 



VOL. 51 

ment, but by writing — w in place of w they apply to a descending 

movement. To integrate equation (1) it is necessary to know the 

density as a function of the pressure. We will assume the relation 

given by equation (18) of § 5, and then from equation (20) we shall 



w — w n 



= — z + " 



m — 1 





Here we have supposed p = p and w = iv when z = o. 
Differentiating equation (18) of §5 we have 

dp _m — 1 dp 
p m p 

If we differentiate equation (2) and eliminate dp and dp by the 
aid of equation (1) we shall find 





m — 1 p 

— w 

We conclude from the last equation that the velocity w can never 
exceed a certain limit w determined by the equation 


\ w-1 


Introducing the value of - expressed in terms of w and w and 

noting that — is equal to a T Q we find 




1 n 

m — 1 w 2 

m- 1 



If we introduce this value into equations (4) and (3) we shall 
determine the maximum value of the height z, that a vertical cur- 
rent cannot exceed, for a given value of the initial velocity w . We 
see by equation (4) that the velocity increases while the pressure 
decreases, and equation (3) shows that the pressure p diminishes 
for increasing values of w. In a vertical ascending current the pres- 


l6 5 

sure is therefore less at the same height than in the surrounding 

If we wish to apply equation (3) to a vertical current whose sec- 
tion varies, it is necessary to employ the following equation of con- 
tinuity : 


Fo(Po\ m 
F \ ~p 

his \ 

(4 bts ) 

However this formula applies only in case the section of the cur- 
rent varies very slowly. It shows that the velocity diminishes as 
the section increases, or vice versa, as at the commencement and 
at the end of the currents. 


(1) Calculation of the altitude. 

m = 5, 

T = 293° , 

we shall find 



w = 

w = i m 

W = io m 

W = 20 m 

w = 3° m 

w = 5o m 



x8 7 4 m 
1 1800 

i874 m 

i8 7 2 m 






i865 m 



i85 4 m 


1 1 244 

i8i 9 m 



(2) Maximnm altitude for a 

given initial 






T m 

Po fdw\ 
u w °\dz Jo 

i m 

170. 5 m 

2 9 530 m 


81.0 mm 




0. 02095 


















0. 1253 















The values of T m represent the virtual temperatures at the height 
v In the last column we have written numbers that represent 


the tangential force at the surface of the earth, expressed in milli- 
meters per degree of the meridian. One should compare this force 
with the gradients for horizontal movements in order to get a clear 
idea of the force that acts in the vertical ascending movements. 

§16. Conditions of the existence of ascending and descending 

currents of air 

If the vertical currents preserve a steady motion, the pressure 
within the currents and in the surrounding atmosphere must satisfy 
certain conditions which we shall now consider. 

Ascending currents 

In ascending currents (fig. 12) the air enters along the surface of 
the earth and consequently the pressure p\, of the atmosphere must 
be greater than pressure p of the lowest part of the current; this 
necessitates the existence of a barometric minimum at the surface of 


the earth. In the higher strata where the air flows out from the 
vertical current, the pressure p in the current must remain greater 
than the pressure p'oi the surrounding atmosphere and consequently 
we shall find a barometric maximum at a certain elevation. 

We remark that great velocities may perhaps modify the phenom- 
ena and that the air can flow outward even from a barometric min- 
imum 7 but it is probable that in nature we shall always find baro- 
metric maxima in such cases, because the velocities are slight at the 
boundaries of the currents. 

7 By substituting 180 + a for a in the equations of §12, we shall have 

the formulae that belong to currents flowing from the center. This hypo- 

k r 
thesis requires that v > ' 


In order to study the conditions that obtain on the inside and on 
the outside of the current between the pressures p and p ' at the 
surface of the earth and between- the pressures p and p' on the 
inside and outside of the current at a given altitude, where the ver- 
tical velocity of, the current is equal to zero, we shall assume that 
the virtual temperatures T in the current and T' in the surrounding 
atmosphere decrease proportionally to the altitude. Let m be the 
coefficient that belongs to the current and m' that belonging to the 
atmosphere, we have from §4 

IY" = (i__!M" (1) 

T / \ amT / 

*-(£)-- (1--^) 

Po \ T o ' V am' 7 ( ; 

By writing 


amT n l t T 

1 - _«!_ 

a m' T' 
we shall have 


P' Po' 

, W -(IJ. (£)-... co 

/ (*) (4) 

If the movement is ascending it is necessary to have simulta- 

^ < 1 and t > 1 

Po P' 

If the movement is descending it is necessary to have simultane- 

Po < 1 and t > 1 

Po P' 

Differentiating / (2) we shall find 

T _ T , _g m' - m 
df («) -f {z)= if {z) _ Jam^m^ ' (g) 


\ amT / \ am' 

dz a „ „,. / „ gz \ / . sz 



If f(z) is positive, f(z) increases with the altitude and remains 
larger than unity : the condition of an ascending current is then ful- 

We shall distinguish three cases : 

(1) T > T' . At the surface of the earth where z = o we shall 
have f(z) positive. We conclude then that an ascending current 
can always exist when the virtual temperature of the current is 
greater than that of the surrounding atmosphere. 

If m < m' or m = m' then f(z) increases continuously with z 
and the intensity of the current increases with the altitude. 

If m < m' then f(z) increases at first and reaches a maximum at 
an altitude given by 

h = -mm' T °~ T °' (6) 

g m' — m 

and at the same time, we have 

T = V 

(2) T = Tq. It is necessary that w> m! so that f{z) can be 
positive. This case includes the unstable equilibrium of the atmos- 

(3) ^0 < T' . At the surface of the earth f{z) is negative. If 
m > m' then f'(z) = o at an altitude h determined by equation (6) 
and f (z) becomes positive for greater altitudes. We conclude then 
that f(z) at first decreases with the altitude and reaches a minimum 
at the altitude h and then decreases with the altitude. It is therefore 
possible that an ascending current can occur even when the virtual 
temperature of the surrounding atmosphere is higher than that of 
the current. The altitude of the current must be greater than h 
and the virtual temperature of the current must decrease with the 
height more slowly than that of the atmosphere. 

Descending currents 

In descending currents (fig. 13) the air enters at the height z and 
flows outward along the surface of the earth where a barometric 
maximum occurs. Therefore the conditions of the descending 
motions are p' > p and p > p' . We will count the altitude z from 
the top downward and write 

T = T + gZ and T = T ' + gZ 

a in a m 


We may write 

' T V" = h + 

Po \ T o 

i + 

am Tq 
am' T ' 


f (z) 

1 + 


1 + 


am' T ' 

T \ m I T ' \ m ' 

tJ \ v 

FIG. 13 

we shall have 

and the conditions for descending motion are written 

t > 1 and-^ < 1 
P' Po 

and consequently we must have 

f (*) > 1 

By differentiating /(z) we shall find 



/' (a) = £ ./(z). 

+ ~ 

a mm 

T V 




We shall distinguish three cases: 

(1) T < T\. For z = o we have /'(z) positive and we conclude 
then that a descending current can always exist since the virtual 


temperature of the current is colder than that of the surrounding 

If m > m' or m = m' , f(z) increases with z and the intensity of 
the current increases with the altitude. 

If m < m' we have f'(z) = o for a height h determined by the 


t ' _ T 

h = mm' -° — ° (11) 

g m' — m 

For this value of z, f{z) reaches a maximum and at the same time 
we have 

T = T' 

If the altitude of the current is greater than h the virtual temper- 
ature of the current at the surface of the earth is higher than that 
of the surrounding calm atmosphere. 

(2) T = T '. The descending movement requires that m > m' 
and this case includes the unstable equilibrium of the atmosphere. 

(3) ^0 > ^V I n this case f'(z) is negative up to the upper 
stratum where the descending current must begin and consequently 
f(z) decreases. 

If m > m' then f'(z) becomes zero for z = h, as given by equa- 
tion (11) and f(z) attains a minimum for that value. For values 
of z greater than h, f(z) increases and it is then possible that a descend- 
ing current can even occur when the virtual temperature of the 
current is higher than that of the surrounding upper layers of atmos- 
phere. The altitude of the current must be greater than h and the 
virtual temperature of the descending current must increase more 
slowly than that of the atmosphere. 

§17. Horizontal velocity produced by a vertical current 

In nature the ascending currents produce horizontal velocities 
along the surface of the earth, which can attain very considerable 
values and which are dangerous to the obstacles they meet in their 
way. As to the descending currents we nearly always find that in 
nature the resulting horizontal velocities along the surface of the 
earth are slight, but it is probable that the horizontal velocities at 
a certain altitude where the air enters the descending current, have 
considerable values. 

Let us consider an ascending current and let v be the maximum 
horizontal velocity, p Q the minimum pressure in the current at the 
surface of the earth ; let p ' be the pressure in the calm atmosphere 


at a point so distant that we can neglect the velocity, we have ap- 
proximately for the equation of the living force or energy, assum- 
ing the density of the air to be constant, 

J^ = tl«-F (1) 


in which F expresses the energy consumed in overcoming friction 
or the work of the friction along the surface of the earth. We shall 

Po ~ Po = p(i V + F) (2) 

The work of friction [or done in overcoming friction] depends on 
the path traversed by the particles of air and on the variation of the 
velocity. There are whirlwinds where the work F is very small 
and others where it is very great. It is especially the dimension 
of the whirlwind that determines the work of friction. In every 
case we see that the horizontal velocity depends principally on the 
barometric depression p Q ' — p , which we can consider as the meas- 
ure of the force of the current. Let us denote by D the barometric 
depression at the surface of the earth. For ascending currents by 
introducing f(z) we shall have 

D = p '-p = pA\ --rj^r) (3) 

V p' f (2) / 

The depression cannot exceed the value given by this equation 
after substituting p = p' . 

Let us assume the pressure in the calm atmosphere equal to 76o mm 
and designate by D m the maximum value of the depression expressed 
in millimeters, we have 

ZP„ = 760(l- f -A_) (4) 

For descending currents, we find in the same way, 

p - p' = p'( p p \ m - 1) (5) 

D m = 760 (/(*)- 1) (6) 

Here D denotes the excess of pressure in the center of the whirl- 
wind over the pressure of the calm atmosphere and D m its maxi- 
mum value. 

I 72 


VOL. 51 

By the aid of these formulae and t he equations of § 16 we have 
calculated the following tables. 

Table I. Ascending currents 





z h 


T' D m 

273° | 7 
283 6 
293 7 
273 1 5 

273 ! 5 



Sooo' 11 

264. 5 
290. 2 

244.5 3 8.9 ,nm 
248.8 22.6 
285.2 3-7 
248.0 9.3 
238.8 7-7 


5i23 5123 
7000 5123 



240. 2 



7000 2635 

235.2 3.3 

Table II. Descending currents 









25 3 

25 3 
25 3 











267. 1 

277. 2 

272. 8° 



5 122 




In whirlwinds of small dimensions we can neglect the action of 
the rotation of the earth. Assuming the altitude / of the horizontal 
current to be very small it is necessary to attribute to the coefficient 
of friction a large value and consequently the air enters into the 
current almost radially. In this case denoting by r the radius of 
the whirl, the equation of continuity gives 

2K r I v = 71 r 2 w 

and supposing v = w we shall have r = 2 I. 

In this case the radius of the whirlwind will be equally small, as 
is proved by observations of whirls of smoke, the whirls of dust over 
roads, and whirls of sand over deserts. In order to calculate the 
horizontal velocity we neglect the work of friction, because the dis- 
tance traversed is very short, then we shall find 

^o= >j 

,Po -P< 

If we suppose the altitude z to be less than iooo m , we can develop 
f(z) as determined by equation (3) of §16 in a series and introduc- 


1 73 

ing therein this value of f(z) into equation (4) and placing p = p' 
we shall have 

^0= V 2 

g z - 

T — T ' 

1 M 

T ' 

We see therefore that for small altitudes the coefficients m and m' 
do not appear in the formulas, that is to say the latent heat of the 
vapor of water plays no part in this case. The formula is identical 
with that for the velocity of the air in a chimney (fig. 14) when we 
neglect the friction. 






FIG. 14 

Supposing the air to be dry, the virtual temperature is equal to 
the absolute temperature, that is to say T = 273 + t and T Q ' = 
2 73 + T o'- F° r this case we have calculated the following table 
assuming r = 20. ° 

Table III. 

Horizontal velocity in small whirlwinds, 
vertical current 

z = the altitude of the 


z = io m 

z = So m 

z = ioo m 

Z = 200 m 

2 5° 

i.6 m 

4 .i m 







11. 6 




11. 6 













11 . 

24. 6 




When the whirlwinds have great dimensions, we cannot neglect 
the work of friction. Assuming that the trajectories of the par- 
ticles of air are logarithmic spirals, we can calculate the barometric 
depression as we have done in paragraphs 12 and 14 (see figures 
9, 10, 11) where the whirlwind of great velocity shows a baro- 
metric depression equal to 32.9 mm for a radius equal to 2 degrees 
of a great circle and with a maximum velocity equal to 50 meters 
per second. The whirlwind of average velocity shows a barometric 
depression equal to 34.9 mm for a radius equal to 20 degrees and with 
a maximum velocity equal to 16 m. p. s. In the last case the work 
of the friction is much greater than in the first, because the distance 
traversed is ten times longer. 

Considering table I, we shall see that barometric depressions can 
be produced by different states of the atmosphere. The two whirl- 
winds, in which the barometric depressions do not sensibly differ, are 
distinguished by their maximum velocities, and it is necessary to 
seek the explanation of this difference in the lengths of the radii 
of the vertical currents that produce the horizontal velocities. 
The whirlwind of great velocity belongs to a vertical current whose 
radius is probably several tenths of a degree, but whose initial verti- 
cal velocity is very great. The other whirlwind of average velocity 
belongs to a vertical current whose radius extends over several 
degrees and whose initial vertical velocity is not great. 

The length of the radius of the vertical current, which we can 
assume proportional to the distance from the center to the point 
where the velocity attains its maximum value, plays an important 
part in the theory of whirlwinds. Comparing two whirlwinds 
having the same barometric depression, that which has the shorter 
radius has the greater velocity and consequently is the most violent. 
Comparing two whirlwinds with the same maximum velocity, that 
which has the shorter radius has the greater gradient and the smaller 

The physical cause that determines the length of the radius de- 
pends on the difference in the condition of the ascending air and of 
the surrounding atmosphere. 



(Christiania, 1880, revised 1885) 

Chapter IV 

on motion in general 

§18. Isobaric surfaces. — Vertical gradient 

In meteorology we speak of isobaric surfaces as surfaces of equal 
pressure or isopiestic surfaces or simply isopiesics. If the air is in equi- 
librium we can consider the isobaric surfaces approximately as spheres 
concentric with the earth, and, for a small part of the surface of 
the earth, we shall treat these surfaces as horizontal planes. If the 
air is in motion the isobaric surfaces differ from horizontal planes. 
In order to fix our ideas we consider a horizontal current of air 
whose isobaric lines at the surface of the earth are concentric circles. 
Let the values of the pressure for the different distances from the 
center be as follows: 

r in degrees of 

a great circle = 4.5 6.5 8 9.8 11.8 14.5 17.8 

b in millimeters 

of mercury = 725 730 735 740 745 750 755 760 

The diminution of the pressure A b for a change of altitude A z 
can be approximately calculated by the formula (see §4, eq. 2). 

A b gAz A z 

J~ ' ' ~aT ' 8200 

Giving successively to b the values 725, 730 we 

calculate the values of A z for any value of A b, and we can construct 
a vertical section of the isobaric surfaces (fig. 15). When the air is 
in equilibrium a vertical section of the isobaric surfaces will present- 
a series of straight horizontal lines (fig. 16a): supposing that we 
have a vertical current, the vertical section of the isobaric surfaces 
would also present a series of straight horizontal lines, but different 
from those of the series in fig. 16a. We shall call these lines of 
intersection vertical isobars, and if we wish to introduce the term 
vertical gradient we should be obliged to establish a definition similar 
to that of the term horizontal gradient. Assume the vertical grad- 
ient equal to the difference of the pressures shown by two isobars 
divided by their distance, we shall always and even in a state of 



VOL. 51 

equilibrium find a vertical gradient whose value will exceed mm provided that we use the millimeter and the degree of 
meridian as our units. It is evident that from this definition we do 
not get a clear idea of the force that acts during the vertical motion 
and which must be represented by the vertical gradient. 

FIG. 15 


800 o 






777 777. 



y6o 777. 777. 

fig. 1 6a 

Let the pressure be p at the height z and q the weight of the column 
of air below z and we get 

n = p + q. 


If the air is in equilibrium, the values of IJ will be equal to the 
pressure p at the surface of the earth and consequently IJ will be 
constant and independent of the height z. If the air has a vertical 
motion the value of IJ will be different from p and will vary with 



the height z. We shall call 77 the pressure reduced to the surface of 
the earth. We call the horizontal lines that correspond to the values 
of 77, the reduced vertical isobars.' We call the difference of two 
values of 77 expressed in millimeters divided by their distance ex- 
pressed in degrees of the meridian the vertical gradient. Denoting 
the vertical gradient by 77, the coefficient of reduction by /x (see §7) 
we have 

- uH 

d 77 dp da dp , 

= — — + — - = — + gp 

dz dz dz dz 


The sign minus is taken, because Ave consider the vertical gradient 
positive upward in the direction in which the pressure 77 diminishes. 
As regards the rectilinear motion we find from §15 

ft H = p .w 



Introducing d IJ we find by integration 

n = p - p w (w - w ) 

If in the formulae of §15 we substitute 

T = 290°; m = 6; p = 760 ,nm ; w = 20 m 
we shall find the results contained in the following table: 












76o mm 


20. O m 

4 O.O mm 

76o.oo mm 




43- s 







5 00 













103. 1 












By constructing the curve of 77 as a function of z we shall find the 
reduced vertical isobars as we see in fig. 1 66. The difference between 
p and 77 we shall call the vertical depression. In the same way that 

the horizontal gradient G produces a horizontal force - G (see §7) 

9 P- 
so does the vertical gradient 77 produce a vertical force - 77, which 



VOL. 51 

must be added to the exterior forces. This vertical force includes 
also the force of gravity. 



















1 / 


1 / 


1 / 


1 / 




FIG. 166 

§19. Equations of motion 

In order to study the general motion of the air we take three rect- 
angular axes of which the axes X and Y are horizontal and the 
axis Z vertical and ascending. Denoting by u, v and w the com- 
ponents of the velocity parallel to the axes and by x, y and z the 
components of the forces referred to the unit of mass and by p the 
density, the equations of hydrodynamics are written 


1 dp du 
p dx dt 


1 dp dv 

p dy dt 


1 dp dw 

p dz 





du dv dw 9 . 

dx dy dz 

the equation of continuity assumes the form 

- P + P J =0 (3) 


In the preceding equations 

du dv , dw 
— , — and — 
dt dt dt 

represent the components of the whole force; the forces produced 
by the variation of the pressure are represented by 

I dp t 1 dp 

— — and — — 

p dx p dy 


which are the components of the horizontal force - G 


and by — — — 

p dz 

which is included in the vertical force - H . 


The components A', V, and Z are the components of the exterior 
and interior forces. 

The exterior forces are the two following: 

Gravity. This is the resultant of the attraction of the globe and 
of the centrifugal force produced by the rotation of the earth. 
The direction of gravity is normal to the surface of the earth and 
is represented by the axis Z. 

Gravity has therefore only one component — g, and by introduc- 
ing the vertical gradient H, we have 

tH= -g- l J-l. 

p p dz 

We consider the force of gravity as constant, because the winds that 
we are studying are located in the lower strata of the atmosphere. 
The dejiecting force of the rotation of the earth. This compound cen- 
trifugal force is the force that we must add to the exterior forces 



VOL. 51 

in order to be able to treat a problem of relative motion as if one 
had to do with absolute motion. Denoting the angle between the 
axis O X and the direction north by a and the components of the 
deflecting force by X , Y and Z we have 

X = 2 io sin S v — 2 co cos 6 sin a w 

Y = 2 co sin u — 2co cos $ cos a w 

Z = 2 cu cos 6 sin a u + 2 co cos 6 cos a v 


Here 6 denotes the latitude considered as positive in the northern 
hemisphere and negative in the southern hemisphere and co denotes 
the angular velocity of the earth per second of mean time. 

fig. 17 

The interior forces are the components of the internal friction 
or viscosity produced by the difference between the velocities of the 
different adjacent strata of air. The surface of the earth offers a 
resistance to the currents of air, the effect of which, in diminishing 
the velocity of the lower strata, is shown by the variation of velocity 
between the different strata. The particles of air having a greater 
velocity increase the motion of the particles having a less velocity 
and, inversely, the particles having less velocity retard the motion 
of the particles having greater velocity. The resistance of the sur- 
face of the earth, therefore, transfers its influence through all the 
Strata of air and influences both the direction and the velocity of 



the motion. We shall in the following chapter consider some special 
cases of interior friction. However, the absence of observations on 
the variation of the velocity with the altitude prevents the applica- 
tion of the exact theory to the winds in general. We shall consider 
friction as an exterior force acting along the surface of the earth. 
Denoting the components of the friction by X t and Y v we write 
(see §7): 

X, = — k u ) 

i .... (5) 

Y t k v J 

in which k denotes the coefficient of friction. 

By introducing the preceding values of the components of the 
exterior and interior forces and noticing that the velocities and the 
density are functions of the four variables x, y, z and /, the equations 
of motion are written as follows: 

1 dp ,, „ du du du du 

- — = X + X 1 — — — u — v — w 
p dx dt dx dy dz 

1 dp T . .. dv dv dv dv 

= Y + Y l — ~ — u — — v - - — w 
p dy dt dx dy dz 

1 dp 

p dz 

„ dw dw dw dw 

-- Z — g — — u — — v — — w — 

dt dx dy dz 

d A + u d P- + v d E- + w d A + p A^Q . 

dt dx dy dz 

. .(6) 


The trajectory of a particle of air is determined by the equations 


= t 





v \ 

= w 


§20. Classification of the systems of wind 

Each disturbance of the eqtiilibrium of the atmosphere produces 
a motion of the air or what we in general call a system of winds. 
Considering the forces which act during the motion, we divide the 



VOL. t;i 

systems of wind into two classes. The systems of the first order 
are those which extend only over quite a limited part of the surface 
of the earth, and which at the same time exhibit variations of veloc- 
ity so great that we can neglect friction and the deflecting force 
of the rotation of the earth. For example we mention tornadoes, 
waterspouts, whirlwinds of smoke, etc. The systems of wind of the 
second order are those in which all the acting forces have some impor- 
tance. As examples we mention cyclones, the trade winds, the sea 
breezes and land breezes. 

Considering the motion of the air in the systems of wind we dis- 
tinguish between the permanent systems and the variable systems. 
In a permanent system the pressure and the velocity at any place are 
independent of time and vary only from one place to another. In 


Po +~*« ?o 

m »- 


P -« m. n> 



FIG. 1 8 

FIG. 19 

nature we never find a permanent system, but we may consider 
the systems of wind which remain nearly invariable for quite a 
long time as permanent. As examples we mention the trade winds, 
an immovable anticyclone or cyclone, with constant pressure at 
the center. The variable systems of wind are divided into movable 
systems and immovable or fixed systems. In the variable fixed 
systems the minimum or the maximum barometer does not change 
position with the surface of the earth, but its value varies with the 

In our following studies we shall consider four simple systems of 

(1) System of ascending parallel winds. 

This system (see fig. 18) has rectilinear isobars, a barometric 
minimum at the surface of the earth and a barometric maximum in 
the higher strata. The air flows inward along the surface of the 


earth from both sides toward the barometric minimum and the 
horizontal current is gradually changed into a vertical ascending 
current. At a certain altitude the vertical current is changed into 
a horizontal current which flows outward from the barometric 
maximum. Denoting by p the pressure at the lower barometric 
minimum and by p the pressure at the upper maximum and by 
p ' and p' the corresponding pressures in the exterior atmosphere, 
we shall have for the depression D which belongs to the horizontal 
current along the surface of the earth, 

D = Po - Po (1) 

The excess of the pressure D in the upper barometric maximum 
is given by the formula 

D-p-pr (2) 

Let q and q' be the weights of the columns of air of the vertical cur- 
rent and of the calm atmosphere, respectively, and let 77 be the 
reduced pressure (see §18) we have 

n =p +q (3) 

Po - P' + q' (4) 

e = p - n (5) 

in which we call E the vertical depression of an ascending current 
(see § 1 8). 

From these equations we obtain 

D + E + D = q' - q (6) 

This last equation shows us that the difference in weight of the col- 
umns of air produces the motion of the three currents. 

(2) System of descending parallel winds. 

This system (see fig. 19) has rectilinear isobars and a barometric 
maximum at the surface of the earth and a barometric minimum in 
the upper strata. The air flows in from the two sides toward the 
barometric minimum and the horizontal current in the upper strata 
changes little by little into a descending vertical current. The ver- 
tical current then changes into a horizontal current which flows 
out from the barometric maximum at the surface of the earth. 

Denoting by p the pressure at the barometric minimum and by 
p the pressure at the maximum and by p f and p' the corresponding 


pressures in the exterior calm atmosphere, we shall have the depres- 

D = Po ~ Po (7) 

which relates to the horizontal current in the upper stratum. The 
excess of the pressure D in the barometric maximum is given by the 

D = P - P' (8) 

Let q and q' be the weights of the columns of air of the vertical cur- 
rent and of the calm atmosphere respectively and let II be the reduced 
pressure, we have 

n - Po +q (9) 

p' - Po' + q' (io) 


E = n - p (11) 

we shall find 

D +E+D=q-q> (12) 

The quantity E represents the vertical depression in the descending 
current, and this last equation shows us that the difference between 
the weights of the columns of air produces the motion of the three 

(3) System of cyclonic winds. 

This system has circular isobars around a barometric minimum 
at the surface of the earth; in the upper strata it has a barometric 
maximum. The air flows in along the surface of the earth from 
all sides and the horizontal currents are changed little by little into 
vertical ascending currents. At a certain height the vertical motion 
is changed into a horizontal motion and in the upper strata the air 
flows out from the barometric maximum. By introducing the same 
notation as we have employed in the first system of parallels, equa- 
tions (1) to (6) hold good also for cyclones. 

(4) System of anti-cyclonic winds. 

Tins system has isobars circular around a barometric maximum 
at the surface of the earth: it has a barometric minimum in the 
upper strata. The upper air flows inward toward the barometric 
minimum and the horizontal currents in the upper strata are changed 
little by little into descending vertical currents. The vertical 
motion then changes into a horizontal motion and the air flows out 
from the barometric maximum at the surface of the earth. Equa- 
tions (7) to (12) hold good for anti-cyclones. 


Each of these four systems of wind has its calm space at the sur- 
face of the earth which represents the interior part where the motion 
of the air is nearly vertical and where consequently we do not feel 
any wind. In the upper strata we must also find calm spaces where 
the vertical motion changes into the horizontal motion or vice versa. 

These four systems which we have called simple systems, trans- 
port masses of air, either from the surface of the earth to the upper 
strata, or from the upper strata to the surface of the earth. When 
we consider the case in which two or several simple systems exist 
simultaneously so that their motions encroach upon each other and 
the masses of air pass from above to below and inversely, we have 
a composite system of winds of which nature offers an infinite 
number of examples. 

Chapter V 


§21. Horizontal currents of air of small extent 

We shall at first consider horizontal currents so small that we can 
neglect the effect of the rotation of the earth; we also assume the den- 
sity to be constant. Let A B and C D (fig. 20) be two horizontal 
planes that enclose the mass of air; assume that the plane C D is 
fixed and that the plane A B moves with a uniform velocity V . 
The motion of the air will therefore proceed in horizontal strata of 

' C JO 



si -jr JS 

FIG. 20 

different velocities; along A B the velocity of the air may be w . 
and along C D the velocity may be zero. Admitting the hypothe- 
sis that the internal friction or the viscosity is proportional to the 
difference between the velocities of any two strata, we conclude 
that the velocity decreases proportionally to the distance z from 
the plane A B. Let h be the distance of the two planes, the increase 

of velocity per unit of length will be — and we shall find the velocity 



u at the distance z by the formula 

u = u o( h _ z ) = u - U °z (1) 

h \ / h 

The internal friction per unit of surface which we denote by F 
will be equal to a coefficient K multiplied by the rate of increase of 
velocity and consequently 

F = K~° (2) 


The plane A B moves with the velocity V and the air along A B 
moves with the velocity u ; the resistance between the air and the 
plane .4 B is proportional to the difference V — u and to the coeffi- 
cient of friction f between the air and the plane; consequently we 
can write 

F = / (V - u ) (3) 

From these equations we find 

*-,-& <« 


In the preceding case the pressure has been supposed to be con- 
stant, We shall now consider the case where the horizontal cur- 
rent of air has a gradient ; the horizontal velocity u depends solely 
on the distance, z and the vertical velocity is zero. The increase of 

the horizontal velocity per unit of vertical distance is j- and the 

internal friction is equal to K —r- . Considering a parallelopipedon 

whose thickness is dz and whose face is a unit of area, the result of 

the frictions of the two faces will be d(K — ) and the mass of the 

element will be pdz. The force per unit of mass resulting from the inter- 
im d 2 u 

nal friction will therefore be — -7^ and this force acts in the same 

p dz 2 

direction as the force of the gradient. The equation of equilibrium 


V G= _K<Pu (g) 

p p dz 2 



The vertical motion being zero, the vertical gradient H will dis 
appear and consequently the pressure is independent of the alti- 
tude z. We conclude therefore that the horizontal gradient G is 
independent of z and constant. Integrating equation (5) we shall 



C — ji G z 




In order to determine the constant C we notice that the internal 
friction disappears at a certain value of z which we will denote by 

h; at the same time the velocity has its 
maximum value U. It is evident that 
the friction is equal to zero in the stra-~ 
turn whose velocity is a maximum be- 
cause the velocity decreases equally on 
each side of this stratum and conse- 
quently the difference between the ve- 
locities of the two strata located sym- 
metrically is equal to zero. By choosing the origin of coordinates 
at this distance h from the surface of the earth (see fig. 21) we 
shall have the constant C equal to zero and by integrating equa- 
tion (6) we shall find 


fig. 21 



Let the velocity of the air at the surface of the earth be u , then for 
z = h we shall have 

u n = U — 

2 K 



The upper limit of a free horizontal current is found by placing u = o 
and let the corresponding value of z be H, we have 

= U - 

2 K 



From equation (6) we conclude that the friction at the distance z, 
is equal to ^ G z; at the surface of the earth the internal friction is 


G h = 









equal to ptG h; the friction between the surface of the earth and the 
air is equal to f u and consequently we find 


n ill 


Here k denotes the coefficient of ordinary friction which we have 
introduced in our previous problems and we have 

k = i- (ii) 


This equation shows that the coefficient of friction k is inversely 
proportional to the depth of the current measured from the surface 
of the earth to the stratum of maximum velocity. 

By experiments on the viscosity of the air, Clerk Maxwell found 
the value of K at o° C. equal to 0.001878. Introducing this value 
in equation (7) we shall have 

u = U - 0.0033 G.z 2 (12) 

Experiments on the motion of liquids show that inequalities of 
depth produce little vortices which play an important part in the 
law of velocity. We are led to adopt the following formula: 

«* = U 2 - 0.04 G.z I (13) 

The value 0.04 is taken from experiments on the motion of water 
in straight channels. 

§22. Horizontal currents of air of large extent 

We shall consider a horizontal current of air that moves over 
so large a part of the surface of the earth that we cannot neglect 
the effect of the rotation of the earth For horizontal motion the 
deflecting force of the rotation of the earth is normal to the trajec- 
tory of the wind and its value is expressed by 2 w sin 8 U, where oj 
denotes the angular velocity of the earth, 8 the latitude and U 
the horizontal velocity of the wind. 

Assuming that the motion of the current of air is uniform, then 
the velocity and the gradient will be constant; the acting forces 
will be the deflecting force of the rotation of the earth, the force of 
the gradient and the friction. In the special case where the cur- 
rent of air moves along a surface without friction, equilibrium will 


exist between the deflecting force of the rotation of the earth and 
the force of the gradient; consequently the two forces must be oppo- 
site and their directions must be along the same straight line. 
We shall then have 

- G = 2 oj £7 sin 6 (1) 


The deflecting force of the rotation of the earth being normal to 
the path of the wind, we conclude that in the case where the friction 
is zero, the current is normal to the gradient, that is to say, the wind 
moves along the isobar. The ratio between the velocity of the wind 
and the gradient is expressed by 

^ = / 2 oj sin 8 (2) 

Let the pressure be 76o mm , the temperature o° C, and the tension 
of the vapor of water o, we shall have 





whence the following values 

6 = 10° 20° 30° 



-p. = 36.6 18.6 12.7 



50° 60° 70° 

8.31 7.31 6.77 

We have supposed that the force of friction, at the surface of the 
earth, is opposite to the motion of the particle of air. In this case 
its path will form an acute angle with the direction of the gradient. 
Since friction has its greatest value at the surface of the earth and 
diminishes with altitude, the velocity of the air and at the same time 
the angle of inclination </> in §i i must increase with the height, which 
observations also show to be the case. 

In the stratum that separates the lower current from the higher 
current, (in the systems of wind that we considered in the preceding 
chapter) the gradient must be zero and consequently the velocity 
of the air nothing. Thus the velocity of the air increases with the 
altitude in the part near the earth diminishes toward zero 
in the region near the stratum that is intermediate between the 
two currents. The velocity of the air must consequently attain its 
maximum at a certain height. 


As the resultant force of the internal friction does not necessarily 
act in the direction opposite to the motion therefore the direction of 
the motion of the air in the stratum of maximum velocity remains 
uncertain. Probably it does not sensibly deviate from a direction 
perpendicular to the gradient. 

How are the velocity and the direction of the motion related in 
the different strata of a current ? This is a problem hardly solvable 
in the present state of our knowledge of the laws of friction and in 
the absence of precise observations of the gradients and the motions 
of the upper strata of the atmosphere. 

§23. A rotary current of air 

We shall now consider a mass of air revolving about a vertical 
axis in consequence of the motion of the surrounding air. The 
exterior air moves in circular trajectories and with constant veloc- 
ity and by internal friction produces a rotation of the interior mass 
of air. We have therefore a mass of air within a cylindrical bound- 
ary whose velocity is given and which turns about a vertical axis 
by reason of internal friction. The tangential velocity U is a 
function of the distance r from the axis; the isobars are concentric 
circles; the gradient is directed along the radius. The acting 
forces are the force of the gradient, the centrifugal force and the 
deflecting force of the rotation of the earth, all of which act in the 
direction of the radius, and finally the force of the internal friction 
which acts in the direction of the tangent. We neglect the friction 
at the surface of the earth, so that the velocity is independent of the 
altitude. The resultant of the internal frictions acts tangentially 
on each element and should be equal to zero, because there exists 
no tangential force with which to establish equilibrium; the result 
is, that the internal friction along a cylindrical surface must be con- 
stant. Let the mass of rotating air be divided into cylindrical por- 
tions which rotate with different velocities. The internal friction 
is due to the differences of the velocities U, but the radius r varies 
at the same time and Avith it the frictional surface; it is necessary 
therefore, to make the friction proportional to the variation of the 
product of the velocity and the frictional surface, divided by the 
increase of the volume. We shall find then 


— = — = a — constant ( 1 ) 

r a r 


By integration we find 

r U = £ a r 2 + b (2) 

where a and b denote two constants that we can determine in the 
following manner: 

Let the given velocity of the exterior air be U l at the distance r x 
and assume the velocity of the interior mass equal to zero at the 
distance r , we find 


* a = ~2 — T7'> o = - — — 2 

r. r — rJ 

U = 1 . ~ . U t (3) 

r n- - r n - 

It is quite probable that in nature the radius r is equal to zero and 
we shall then have 

U - I • U t (4) 

' i 

Hence, the current of air rotates with a constant angular velocity (see 


In order to determine the gradient and the pressure, we distinguish 
two cases in the northern hemisphere. 

(i) Rotation contrary to the sun. 

In the cyclones of the northern hemisphere the rotation takes 
place contrary to the apparent diurnal motions of the sun, the grad- 
ient is directed toward the center, the centrifugal force and the 
deflecting force of the rotation of the earth are directed outward. 
We have then 

u U 2 

- G = — + 2 a) sin 9 . U (5) 

P r 

By writing /xG = — and introducing the value of U given in equa- 

tion (4) we find by integration, p being the pressure at the center 

where U = o, 

P ~ Po = % (U 2 + 2 co sin 6 . U r) (6) 


(2) Rotation with the sun. 

In the anti-cyclones of the northern hemisphere the rotation takes 
place with the apparent diurnal motion of the sun; the gradient 


and the centrifugal force are directed outward and the deflecting 
force of the rotation of the earth is directed toward the center. We 
shall have then 

fi U 2 

-G=2cosm6.U- (7) 

P r 


Po - P 

% {2 to sin 8 . Ur - U 2 ) (8) 

Equation (7) demands that the angular velocity — be less than 


2 co sin 8 because the gradient must be positive. 

Chapter VI 


§24. Permanent wind-systems of the first order 

In nature tornadoes and waterspouts represent examples of 
cyclones of the first order, but meteorological observations of these 
phenomena being very scarce do not suffice to show us the changes 
of pressure and velocity which take place in them. We cannot as 
yet by mathematical analysis construct a complete system of wind. 
However, we shall consider some simple cases which show analogies 
with the systems of nature and from them we shall seek to deduce 

In the general equations of § 19 we may neglect the components 
X , Y , Z , X x and Y l and we shall consider the density as constant 

The equations assume the form 

g + 


dp du 

du du 


d x d x 

— V , - — W ~r- 

dy dz 


dp d v 

d v d v 


d y d x 

dy d z 


d p dw 

dw dw 


d z d x 

— v ^— — w J— 
dy d z ) 


du d v dw 

= -T- + -r- + -r- (2) 

d x dy d z 



Considering the special case, where we have 

dw dv du dw dv du 
dy dz dz dx' dx dy 

equations (1) are' reduced to a single one. Denoting the absolute 
velocity by V, we have 


V 2 = u 2 + v 2 + w 2 

d p = — V dV - gdz, 

and consequently 

P = Po +$p(V i ? - V 2 ) + g P (z -z) 



where p denotes the pressure for V = V and z — z . 

We designate the distance from the origin of coordinates by R 
and its horizontal projection by r, the horizontal velocity by U , 
and consequently we have 

x 2 + y 2 + z 2 = r 2 + z 2 = R 2 
U 2 = u 2 + v 2 

First example. The trajectories are straight lines directed toward 
a fixed center. 

We take the fixed center (see fig. 22) at the origin of coordinates 
and put the equations of the trajectories under the form 



R n 

1 + 


R 3 


Calculation shows that these equations satisfy the conditions which 
we have introduced above. For the time t = o, we have R = R , 
from which we conclude that all the particles of air are found at first 
on the surface of a sphere whose radius is R . Differentiating x f 


y and z with respect to t and introducing u, v and w (see §19) equa- 
tion (8) and eliminating the arbitrary constants, we have 


For a< o, the air flows in to the center and for a > o the air 
flows away from the center. 

In any horizontal plane (z = constant) the gradient G is found 
by equation (3), by noticing that 

U V w 

U V 


x y z 

= 7 = R = 

= R* 

v= a 
R 2 ; 

R 2 = r 1 + Z 2 


_dp_ p 

VdV p 2a 2 r 


dr u 

~dr ~~ ju ' ^W 

... (7) 

Let us consider a horizontal plane at the distance z = — h = z Q , 
and study the phenomena along this plane, which can represent the 
surface of the earth. 

Denoting the absolute velocity at the point A by V Q , we shall have 


we shall find 


r = $ h 


= h V 1 + f 3 


(l + £ 2 )t ° 

G - 

_*p v 2 e 
tt h ' (1 4- er 


- Po = \P (V 2 - V 2 ) 

Denoting by p ' the pressure at a point so distant that we can con- 
sider the velocity V as zero [V = — from equation (6)] we place 

R 2 

d = p: - Po = i p v 2 


p —p = — . U, 

(1 + ey 


We easily see that all these formulae depend on only two con- 
stants or parameters, namely the altitude h and the maximum veloc- 
ity V . We can change the last parameter and consider the depres- 
sion D as the second parameter. Thus the function of U shows that 
the horizontal. velocity has a maximum value U for £ = V \\ the 
distance r from point A to the point where U has its maximum 
value, is 

r = h V £ and U = V V 27 
The gradient G has its maximum value G m for 


G = i o 25 V5 2V 
m p 216 fc 

We shall now choose Z) , expressed in millimeters of height of mer- 
cury, and r expressed in degrees of the meridian, as the parameters 
of the system and are thus able to establish the following formulas, 
by introducing a mean value of p (0.13 18 at the temperature o° 
and the pressure of 76o mm and for dry air) : 

The maximum horizontal velocity = U = v/30.6 D 
The maximum horizontal gradient = G m = 0.715 D a /r 
The distance from G m to the point A = r m = 0.63 r 
The height of the absolute center O = h = 1.41 r 
The absolute maximum velocity = V = 2.6 U 

By the aid of the preceding formulae we have calculated the fol- 
lowing table, in which D denotes the barometric difference: 

£ 0.5 1 2 3 4 

r :r 0.71 1.41 2.83 4.24 5.66 

U:U 0.93 0.92 0.46 0.25 0.15 

G:G m 0.99 0.48 0.06 0.01 0.003 

D:D 0.36 0.75 0.96 0.99 0.9965 

In fig. 23 we have constructed, from this table, the curve of veloc- 
ity, the curve of gradient and the curve of pressure that determines 
the system of isobars. We can compare our system of wind to the 
lower half of a cyclonic system in nature; probably in nature the 
maximum gradient occurs at the same point as the maximum veloc- 
ity. The depression D which depends on the physical state of the 
air, determines the maximum velocity; the maximum gradient 
depends on the depression and the distance r which in nature prob- 



VOL. 51 

ably represents the radius of the vertical current. The radius r 
depends on the height of the vertical current. Finally, it is neces- 
sary to remark that in our example the vertical velocity is very great; 
the velocity V at the point .4 represents the vertical velocity at this 
point. But on the other hand, in natural systems of wind the motion 
of the air differs much from this motion in our case, because the sur- 
face of the earth compels the particles of air to follow trajectories of 
a different form. 

*""-» ■ " Q 

FIG. 23 

Second example. The trajectories are parallel to a vertical plane 
and pass through a horizontal line. 

We take the plane X Z (see fig. 22) parallel to the trajectories 
and the axis Y as the horizontal line. The ordinate y disappears 
and we write 

u = U, x = r and R 2 = x 2 + z 2 

We write the equations of the trajectories under the form 



R n 


1 + 



Placing t = o, we have R = R ; consequently the particles of air are 
found at first at the surface of a cylinder whose radius is R . Differ- 
entiating with respect to t and eliminating the constants we shall 

u w V a. 

x z R R 2 

For a < o, the air flows towards the axis and for a> o, the air flows 


away from the axis. The gradient G in a horizontal plane is found 
by the formula 

G = i a = _* vdv = p_ a i^ (io) 

[i d x fi d x n R 4 

Consider a horizontal plane at the distance z = z = — h, and study 

the phenomena on this plane. 


(1 x 

V = ^ and £ = - and D = \ p V 2 

we shall have 

u = - — ; — — • V n ; G = - 

1 + f 2 ° ' // /* (1 + e 2 ) 2 

The horizontal velocity has a maximum u when £ = i and 
m = £ V . The horizontal gradient has a maximum 

£ m when £ = i/y~ 

16 ft h 

Choosing as the parameters D expressed in millimeters and the dis- 
tance of the axis x = h, where the horizontal velocity has its 
maximum, as expressed in degrees of a great circle, we shall find 

The maximum horizontal velocity U = V 51.5 D 

The maximum horizontal gradient G m = 0.65 — ° 


The distance from G m to the vertical axis x m = 0.58:r 

The height of the horizontal axis h = x 

The absolute maximum velocity 1 7 = 2 U 

By the aid of the preceding formulas we have calculated the fol- 
lowing table in which D denotes the barometric difference: 

£ 0.5 

U : U 0.80 

G :G m 0.99 

D :D 0.20 


















From this table one can construct the curves of velocity, gradient, 
and pressure, and the system of isobars and we shall find a system 
of curves analogous to those of fig. 23. 

We can compare this last wind system with the lower half of a 
system of parallels of the first order in nature and we can make the 
same remarks as on the first example. 

§ 2 5. System of parallel winds of the second order 

Mathematically speaking the systems of parallels have an infinite 
length. In nature the length is limited, but we can neglect the 
disturbances produced by the lateral limits. Along the surface of 

the earth the system of parallel as- 
cending winds presents two horizontal 
currents, which flow from both sides 
toward the barometric minimum or 
trough situated along a straight line. 
We distinguish two halves on each 
side of the barometric minimum and 
/J" 1 7 each half has its internal part whose 

-* — j /- me ! breadth may be r , and its exterior 
\ »»;wmw»,///w//w///////.'/?/, part. The horizontal current moves 

fig. 24 in the exterior part approximately at 

a constant altitude, and in the lower 
part at an increasing altitude. Consequently the horizontal veloc- 
ity has its maximum value £/„ at the distance r from the barom- 
etric minimum. 

Denoting the height of the external current by h (see fig. 24) and 
the angle between the maximum velocity and the gradient by <p , 
the quantity of air which enters per unit of length is represented 
by U cos (p h. In the interior the current changes little by little 
into a vertical current whose velocity we may indicate by w and 
consequently we have the condition 

u> r = U cos (p h (1) 

It is probable that in nature the ratio h / r is so small that we can 
neglect the vertical velocity and the vertical barometric depression 
that results from it. We shall therefore consider only the horizon- 
tal currents with either constant or variable velocity. 


Constant Latitude 

We have already in §10 discussed the systems of parallel winds 
with rectilinear isobars and constant velocity. We now write 

U cos <p = c + c x 


in which the distance x is measured along the gradient. 

Differentiating this equation and introducing the value of U 
and of d Uin place of v and d v in equations (2) and (3) of §10, we 
shall have 

G cos <p = U ( k + c + U sin <p ~ 
p \ dx 

11 G sin <p = U ( 2 to sin d - U cos </> - ) 
p \ dx) 

. . . (3) 


If we eliminate G from these equations, we shall have the equa- 


= (k + c) sin <l> — 2 co sin 6 cos <p + U 


in place of which we can write 

JL d A=U ens / (tang J) = 2 co sin 6 - (k + c) tan 4, (5) 
cos ^ dx dx 

We see that we can satisfy this equation by placing the last term 
equal to zero. Then we have 

tang $ = 

2 co sin 6 
k + c 

The angle of inclination ^ becomes constant, and the first term of 
equation (5) also becomes zero. Equations (3) and (4) become 


G sin <p = 2 co sin 6 . U 



The normal angle of inclination being expressed by the formula 

2 co sin 6 

tang a = 


we shall have 


2 co sin 8 tang a 
tang * = *_ (7) 

k + c c 


U p. cos <J> p sin (p 

G p k + c p 2 co sin 6 


From these equations we conclude: That the angle of inclination 
,W is constant for a wind of variable velocity and rectilinear isobars, 
but that it differs from the normal angle ex. The ratio between the 
velocity and the gradient remains constant and is expressed by 
the same function of the latitude and of the angle of inclination 
as for winds of constant velocity. 

The gradient increases proportionally to the velocity and con- 
sequently to the distance x. It follows that the depression between 
two isobars is found by multiplying the distance by the mean of 
the corresponding gradient. 

When c > o the wind blows with increasing velocity and the angle 
of inclination is less than the normal angle. When c < o the wind 
blows with decreasing velocity and the angle of inclination is greater 
than the normal angle. 

If we consider a station situated on the seashore and note that 
the coefficient of friction is greater on the land than on the ocean, 
we must expect that the ocean winds at such a station will have 
an angle of inclination greater than the inclination for the land 

Let us consider a system of par- 
allels in which we can represent 
the curve of velocities approxi- 
mately by two straight lines (see 
fig. 25). The curve of the gradient 
will be also represented by two 
straight lines and placing the maximum gradient equal to G we 
shall have 

D = G r (9) 


In nature the velocity is represented by a curve, and at the point 
where U = U , the variation of the velocity is zero. Consequently 
the angle of inclination is equal to the normal angle a for the maxi- 
mum velocity. Choosing D and r as the parameters of the sys- 
tem we shall f have 

The maximum gradient G = — - 

n-u • i v tt P COS a D o 

The maximum velocity U Q = - . — - 

p k r 


In the neighborhood of the equator with p = 0.1199, a = o and 

k = 0.00002, we have U = 51 — °; assuming that the system of par- 

allels has a breadth of 2r = 20 and that the total depression is 2 mm 
we shall find G n = o.2 mm and U = 10 meters. 

Variable Latitude 

We employ the same notation as in §11, and we consider only the 
case in which the gradient coincides with the meridian. Take the 
origin at the equator, and write the latitude 

& = Xx 

9 n 

X " : ± W ' 180 

Here the sign plus indicates that the gradient is directed toward 
the north and the sign minus that the gradient is directed toward 
the south. Assuming that the velocity is expressed by equation 
(2), we again find equations (3) and (4). 

Eliminating G from these equations we shall have 

d (tang (p) 
U cos <p -j = 2 o) sin (9 - (k + c) tang <[> ■ ■ ( n ) 

Introducing & in place of sin 6, we see that equation (11) is satisfied 

tan ^ = drb ( *- £) (12) 

e = -^- (13) 

k + c 


For, by differentiating (12) we have 

d (tang 4>) 2 co d S 2 co X 

dx k + 2 c dx k + 2 c 

Substituting this value and the value of U cos <p from equation (2) 
we have 

2coX 2co 

Xc + Xcx = Xc + c8 = (k + 2 c) 6 - (k + c) • (6 - e) 

h c = (k + c) e 

e = 

k + c 

By eliminating from equations (3) and (4) we find the gradient 

G = ~ • U cos 4> (k + c + 2 co 6 tang 0) . . . . (14) 

The radius of curvature of the trajectory is found by the equation 

ds dx 

R = = 

cfy cos ^ (/ ^ 

From equation (12) we deduce 

dil> 2 co dd 2coX 

cos 2 (p d x k + 2 c dx k + 2 c 
and we thus have 

R= k + 2c . _L_ (15) 

2w ^ cos 3 <f> 

It is evident that at the point of maximum velocity c changes its 
sign, and that in nature at this point the equation c = o must be 

In a system of parallels we have two horizontal currents, one along 
the surface of the earth and the other in the upper strata, For the 
last we can employ the same equations as for the first, but for want 


of observations the coefficient of friction remains unknown for the 
upper strata. Neglecting the vertical depression E, the sum of the 
two horizontal depressions, D Q at* the surface of the earth and D 
in the upper stratum, is equal to the difference of the weights of the 
columns of air [see §20, eq. (12)]. We can approximately calculate 
this difference by equation (4) of § 17. To fix our ideas we assume 
that the air at the point A has the virtual temperature T = 298°, 
and the coefficient m = 6, if there is an ascending current. At the 
point B the calm air has the virtual temperature T Q ' = 294 and the 
coefficient m' = 7. If we assume that the air moves from B to 
A and there ascends, we shall find by the formula (6) of §16, that 
the ascending current extends up to a height of 4918 111 and that the 
difference of weight of the columns of air is 3.i mm . Assume D = 
2 mm anc [ £} = II mm_ jf ^he extent of the system of wind B A is 
20 we shall find by the formula (10), U = io m . The time the 
current requires to move from B to A is expressed in hours. 

10 6 20 1 1QQ ., 

— . . — - = 123 . 3 hours 

9 \ U 3600 

If now the air from B can in 123 hours attain the physical state 
belonging to the air at A, then this system of wind is realized and, 
as we see, the parameters of the system are determined by the physi- 
cal state of the air and of the surface of the earth. If we assume the 
distance B A = 16 , we shall have U = 12.5™ and the time equals 
103 hours, but in this case the air from B will arrive at A with tem- 
perature lower than the temperature at A, and consequently the 
depression will diminish and the system of wind cannot be perma- 

§26. Cyclonic system of the second order 

We have already in § 12 and § 13 studied the cyclones of the sec- 
ond order in respect to the motion along the surface of the earth. 
We have assumed that the horizontal current has a constant height 
h in the exterior portions and that the horizontal velocity increases 
in this part toward the center and attains its maximum value U 
at the distance r from the center of the isobars. Then the current 
enters into the interior portion, where its velocity decreases at the 
same time that the motion is changed little by little into a vertical 
motion. Let the mean vertical velocity be w and the angle between 
the gradient and the maximum velocity be </> ; the condition that 



the same mass of air passes from the horizontal current to the ver- 
tical current is expressed by the formula 

7T r 2 w = 2 u r h U cos 4> 

w n = — U n cos <l> (1) 

2 h TT 

- U cos (p 


JtCax. x < 

In cyclones of the second order the ratio h/r is probably so small 
that we can neglect the vertical velocity and the vertical depression 

E which belongs to the 

accelerated motion of 
the vertical current. The 
rotation of the cyclonic 
motion is determined by 
the deflecting force T 
(fig. 26) of the rotation of 
the earth. We have as- 
sumed that the cyclonic 
system has a barometric 
minimum at the surface 
of the earth and a ba- 
nc 26 rometric maximum in 

the upper strata. It 
follows that the rotation in the upper strata is opposed to that at 
the surface of the earth. The intermediate vertical current which 
joins the two horizontal currents is consequently rectilinear. The 
phenomena are inverse to those of the anti-cyclones. However, the 
little that we know about the motions of the cirrus clouds seems to 
indicate that the axes of rotation of the lower current and of the 
upper current do not lie in the same vertical line. 

As parameters of the cyclonic system we can choose the depres- 
sion D and the radius r which 
depend on the physical state 
of the air. We can approxi- 
mately establish the follow- 
ing relations between the max- 
imum velocity U and the gra- 
dient C . Assume that the gradient curve (fig. 27) is composed 
of a straight line and of a curve whose equation is 

G= a + a ' 
r r 3 

fig. 27 


From §12 we have 

P tt k 

H cos a 


a' = P U* r > — 
a 10 8 

In the interior portion the depression is equal to 


r r G ro 

dp = I — rdr = \G 9 r 

r = '- r = 

and in the exterior portion, the depression is equal to 

\ dp= [l- dr + T* dr ) = 

*J r = r *-S r = r (1 

• - ta «--(i) + »-SM7 

Neglecting the term I ° J and noticing that 

C-- + -, 

we get the total depression between the center and the point whose 
radius is r, 

Do = a( i + log • nat • ( - ) ) + — 2 

If the radius r denotes the radius of action of the system, we can 
determine it by giving to U in the equation 

r U 

a conventionally slight value. In violent cyclones we assume U =5 
meters per second. It seems that the ratio r/r falls between 2 and 

The time that a particle of air requires to pass through the exterior 


portion, that is to say, to go from a point at the distance r to a point 
at the distance r w is found by the formula 

t = 


U cos </' U r cos </> c 

d r = 

2 U cos 4> r \ \ r< 

- 1 

By the aid of the preceding formulas we have calculated the fol- 
lowing tables.. In the sixth column h is expressed in kilometers and 
w Q in meters. We compute the value of U by successive trials. 
Cyclone of the temperate zone 

= 60° K = 0.00004 </-„ = 72.4° 








w : h 


I0 mm 


21. o m 

- mm 

4. 2° 


40 hours 






0. 17 

















0. 07 








n. 9 


0. 10 




1 1 . 







17. 2 







22. 2 







26. 7 










0. 042 


Cyclone of the temperate zone 
= 30° K = 0.00002 </<„ = 74.7 C 






Wo '■ h 


I0 mm 

0. i° 

3i m 

9 6 mm 


i.5 m 

7 hours 









0. 1 




2. 1 



0. 1 



1 . 



































4. 1 

0. 61 



Chapter VII 



§27. Variation of pressure in a stationary system of wind 

In permanent systems of wind the air that flows inward and pro- 
duces the vertical current, is always homogeneous. In variable 
systems of wind the flowing air is heterogeneous and 
consequently the motion of the system varies with 
the time. The intensity of the systems of wind de- 
pends especially on the vertical current and we call 
the air, which enters into a system of winds and has 
such a physical state that it can produce or sustain a 
vertical current either ascending or descending, the 
alimentary air. We call the air which enters a sys- 
tem of wind, but which cannot sustain a vertical cur- 
rent the supplementary air. Let us consider a vertical 
ascending current whose height is h and in the first 
place suppose the motion to be permanent. Denoting the ver- 
tical velocities by w and w x (see fig. 28) and the densities by p and 
p v at the influx and the efflux of the current respectively, the equa- 
tion of continuity assumes the form 

= p w — p r w t 

Suppose that after a certain time the inflowing air acquires the 
density p' and the temperature x' and that the vertical velocities 
remain unaltered in the first moments, the equation of continuity 
takes the form 

d (ph) = hdp = (p' w — p x Wi) dt 


FIG. 28 

Eliminating p x w v we find 





The change of density produces a change of pressure and assum- 
ing approximately 

we shall have 

dp dp p' 273 + x p 
~p ' ~p'~p = 273 - +7' ' p 

1 dp 

= a (273 + r ') 


a- h (r- O 



We conclude from equation (2) that the pressure diminishes when 
the incoming air is warmer, and the pressure increases when the in- 
coming air is colder.' Applying this result to nature we infer that 
the supplementary air is colder than the alimentary air. 

Denoting by d the change of the pressure per hour and in milli- 
meters and expressing h in kilometers we shall have 

3600.760 ap w 
3 = 10333 ' 1000 ' h ^ ~ T ^ 

and for an average value of p (0.1318) we have 

d = \0j (t - r') (3) 

Let us consider a stationary cyclone whose pressure at the center 
varies; d represents the variation of the horizontal depression D n . 
In order to introduce the relation given by equation (1) we notice 
that this can be written 

dp w 

Jt = Tz {p ' ~ p) 

In passing from these infinitesimal values of altitude to the 
finite differences, it is necessary to consider the whole height h of 
the horizontal current, because in the latter we do not know the 
variations of velocity with the height and when w expresses the 
vertical velocity in the ascending current at the height h, that is 
to say, at the level where the motion commences to be purely ascen- 
sional, we can introduce the relation given by equation (1) of §26, 
and expressing r in degrees of a great circle we shall have 

U cos O 
d = 0.18 (t-t0 (4) 

' (i 

By the aid of equation (3) we can easily calculate the variation 
of the pressure in the cyclones given in the tables of §26. 

Equation (2) applies only for the first few moments. If the 
vertical current is continually being supplied by heterogeneous air, 
the change of pressure must depend also on the humidity of the 
air. According to § 5 moist air has during ascension a mean tem- 
perature higher than dry air. If we consider r and r' as approxi- 
mately mean temperatures, we arrive at the conclusion that the 
supplementary air is colder and dryer than the alimentary air. If 


then the air flowing into a stationary cyclone changes its physical 
state and becomes colder and dryer, the horizontal barometric 
depression diminishes little by littte and the cyclone is destroyed 
after a certain time. 

§28. Instantaneous systems of wind 

Let us consider a column of air of the height / that has been heated 
so that the pressure p at the upper end (see fig. 29) exceeds the 
pressure p' of the surrounding air. The air commences to leave 
, the upper end of the column and at the same 
4, \ r f r time air enters at the lower end, but the density 
of the supplementary air filling the column up 
to the height z has a value p different from the 
value p' of the air of the calm atmosphere and 
consequently the weight of the column dimin- 
ishes so that the pressure p at the surface of 
the earth decreases and produces a depression 
I J ! ^ p ' — p Q . The pressure p diminishes at the 

p p ' same time that the vertical velocity w of the 
current increases up to a limit that corresponds 

Fir* 20 

to the maximum value of the vertical velocity, 
and after this moment the steady motion goes on. As an approx- 
imation we can neglect the variation of density due to gravity and 
consider the force that maintains the ascending motion as equal to 
Po ~ P 



The equation of motion assumes the form 

I (1) 


dw p - P 

dt pi & 

The difference p Q — p is equal to the weight of the column of air z 
having the density p and of the column I — z having the density 
p'; consequently we have 

Po ~ P = g P z + g P' ( l - z ) 
Introducing this value in eq. (1) we have 

dw (p' — p) I 

dt s ' p I 


From equation (1) we conclude that the vertical velocity increases 
up to the moment when the pressure has attained the value 

Po = P + g P l 


At this moment the column is filled with air of the density p and 
z = I and the motion is steady. Assuming approximately p' = p 
we shall find 

P'o == P +gp' I 

and consequently 

P'o ~ Po = g I (P' - P) ■ ■ • ' ( 3 ) 

d z 
Introducing w = , equation (2) will by integration give 

2=/(l-COSy/l (4) 


J Po ~ Po /KN 

w = \j .... [ft) 

is the maximum velocity. 

The duration of the current up to the moment when the steady 
motion commences, may be /„ and we shall have 

= ,.„ («) 

2 w 

Denoting by x and z f the mean temperatures of the column of the 
current and of the column of the calm atmosphere, respectively, 
we can assume approximately 

p' 273 + r 
~p = 273 + ? 

Equation (5) is now written in the form 

I gKr-t') 
y 273 4- z' 



/ = 1000 m , 
r - r' - 6°, 
273 +t / = 290°; 

we shall find 

w = 14.2 m and the duration t = 110 sec. 


The steady motion continues as long as the alimentary air remains 
unaltered. Suppose that after a certain interval of time t, the air 
flowing along the surface of the earth enters at the lower end of the 
column with the density p', then the column will little by little be 
filled with air at this density, at the same time that the velocity 
decreases to zero, and the pressure p increases to p ' , and the motion 
ceases altogether. 

The duration of the steady motion depends on the quantity of 
air that can supply the current. If for example the system can be 
regarded as a radial cyclone, then denoting by r the radius of the 
vertical current, by r the radius of the alimentary air, by h its height, 
and by t t the duration we shall have 

n r 2 w t t = n r 2 h 

u (S) 

If, on the contrary, the alimentary air can be regarded as a stratum 
whose length is very great compared with the breadth, we can 
imagine that the system of wind consists of a series of instantaneous 
systems, such that the cyclone moves along the mean or central 
line of the alimentary stratum (tornadoes, hailstorms). Let the 
breadth be L and the velocity of propagation be W, we shall have 

and consequently 

L h W = tz r£ w 


The time t that the cyclone consumes in passing any point, is 
given by equation 

*-w (10) 


r = 200 m , 
h = 100 m , 
L = 1200 m , 

we find 

W = 14.9 m and 
t = 27 seconds. 





§29. Ocean wind and land wind 

We consider the ocean winds and the land winds as variable sys- 
tems of parallel winds of the second order. The ocean winds be- 
long to an ascending system of parallels and the land winds to a 

descending system of 
p' parallels. During the 

day the land becomes 
much warmer than the 
sea and consequently the 
pressure p at the upper 
Po end of the column of air 

(see fig. 30) increases and 
exceeds the pressure p'\ 
fig. 30 the air at the upper end 

leaves the column and at 
the same time the pressure p diminishes, because the weight of 
the column diminishes, and thus produces a horizontal current which 
is the ocean wind. Approximately we can neglect the time neces- 
sary to fill the column of air from the ocean and Ave can consider the 
depression p ' — p as a function of the temperature. 

During the night the land becomes much cooler than the ocean, 
the pressure p diminishes at the same time that the weight of the 
column of air increases and a descending system of parallels obtains 
with a depression p — p '. 

The barometric depression which depends on the unequal heat- 
ing of the ocean and the land is a function of the time and of the 
place, and must be determined by observations. This depression 
produces a horizontal current which commences with a velocity 
equal to zero; the depression gradually increases, the velocity in- 
creases and the current extends more and more up to the moment 
when the depression attains its maximum value. Then the depres- 
sion and the velocity of the current decrease simultaneously up to 
the moment when the current ceases. 

Consider the horizontal current at any time and denote its max- 
imum velocity which occurs near the coast, by U , its length along 
the gradient by x, and the depression in millimeters by D ; it is 
evident that D is a function of U , of x and of the time. 
We approximately assume 

10 333 

D = p U > 


and introducing a mean value of p we shall have 

D - - itk (I) 

If we suppose that the curve of the gradient can be represented 
by two straight lines and if we express x in degrees of a great circle 
we have 

D = lG x (2) 

The gradient G is determined for the velocity U by the known form- 

G p k 

U /< ' cos a 

2 co sin 6 
tang a = ^~ 

consequently we have 

For example let 

2 U n cos a 
x = — ^ (3) 

8 = 30°, 

k = 0.00004, 

we have 


we shall have 

ex = 61.25 c 
G : U = 0.09. 

D = 0.5 mm , 

U = about 7 m , 
G = 0.63 
x = 1.6° 

§30. Movable systems of wind 

When the barometric minimum or maximum changes its position 
along the surface of the earth, the system of wind is called movable. 
The movement of the barometric minimum or maximum is accom- 



VOL. 51 

panied by a movement of the ascending or descending vertical cur- 
rent, and the cause of this is due to the heterogeneity of the air that 
enters the barometric minimum either at the surface of the earth or 
in the upper strata. The alimentary air on entering, produces a 
new vertical current at the same time that the supplementary air 
suppresses the existing current, and consequently the vertical cur- 
rent moves in advance of the barometric minimum and causes its 
change of position. When the barometric minimum is situated 
in the upper strata its movement is accompanied by the movement 
of the barometric maximum at the surface of the earth, and in- 

In any movable system of wind the pressure at any point what- 
ever varies with the time and this variation of pressure is closely 
connected with the velocity of propagation of the barometric min- 
imum or of the central calm region. 

Let x and y be the coordinates of any point, whatever; $ and rj 
the coordinates of the movable origin which represents the baro- 
metric minimum; we can generally express the pressure as a function 
of the location and the time, or 

p = f(x - $,t) - y,t) 

Differentiating we shall have 

dp dp <i£ dp drj 
dt ds dt drj dp 



FIG. 31 

Denoting the velocity of propagation the [movement of the mini- 
mum] by W and its angle with the axis X by /? (see fig. 31) the 
gradient by G and the angle of the direction of the gradient with the 


axis of X by a, we have 

d£ drj 

W cosB - — ; WsmB = -37; 
at at 

dp dp 
a G COS a = — — r- = -37 : 

« G sin a — — —r- = -3- . 

Substituting these values we shall have 

—r = fx G W cos («-/?) + 


Denote the angle between G and W by y and let d be the total 
variation of the pressure at any point (expressed in millimeters per 
hour) and d' the variation of the pressure, if the system is stationary, 
we have 

dp 10333 1 

dt ' 760 3600 
10333 1 





760 3600 

Substituting these values we shall have 

d = <5 + 0.0324 G W cos r (3) 

If the pressure at the movable origin is invariable we have 

and consequently 

d = 0.0324 G W cos r (4) 

At the front of a cyclone the angle y > tz and consequently d 
is negative and the pressure decreases; at the rear y < n and the 
pressure increases. 

Example. Let us consider a movable cyclone whose central 
pressure is constant and whose velocity of propagation is so small 
that we can consider the motion of the system as a geometrical 
movement of the isobars ; finally we suppose that the radius of action 
is so great and the maximum velocity so slight that we can apply the 
same ratio between the gradient and the velocity as in rectilinear 



VOL. 5 I 

motion. At the station A (see fig. 32) we observe the velocity 
U t — 1 2 m and the variation of pressure^ = — o.5 mm ; at the sta- 
tion B we observe U 2 = 8 m and d 2 = — 0.4 mm . Assume the mean 
latitude equal to 6o° and the coefficient of friction k = 0.00006, 
we shall have (see §9) the normal angle a = 64°.6 and the normal 
ratio G : U = 0.15. From equation (4) by substituting c7, = i.8 mm 
and G 2 = i.2 mm , we shall find W t cos y x = — 8.6 m 
and W 2 cos y = 10. 3 m . 

FIG. 32 

Let A U 1 and B U 2 (fig. 32) be the directions of the veloci- 
ties, that is to say, the true directions of the currents of air, 
which are different from the direction observed by wind vanes, 
because of the different values of the friction in the midst of the 
current of air and at the surface of the earth (see §34). Draw the 
angles U l A C = U 2 B C = a then the point of intersection C is 
the movable origin or the location of the barometric minimum. 
Lay off C a = W cos 7-, and C b = W cos y 2 and construct a circle 
through the three points a, b and C then the diameter C d represents 
the velocity of propagation W both in direction and in extent. 



§31. Velocity of propagation of a cyclone 

As we shall now explain, the movement of the barometric min- 
mum is due to the heterogeneity of the air. At the front of a cyclone 

the alimentary air whose tem- 
perature is r, enters and pro- 
duces a lowering of the pres- 
''° * sure that we denote by d v At 

. ^_ jpr the rear of the cyclone the sup- 

\(?2 £° '/ plementary air (see §24) whose 

temperature is r 2 , enters and 
produces an increase of pressure 
whose value is d 2 . The air of 
fig. 33 the central part of the cyclone 

has the temperature r and we 
have Tj > t > t 2 . In accord with equation (4) of §27 we write 

J,_0.18^i±«(r-r,) (1) 

■ ^o.lS^i^r-r,) (2) 

Designating the variation of the pressure at the center by d 2 and 
assuming that this variation occurs at all points, we can substitute 
successively y = it and y = o in equation (3) of §30, then we shall 

d x = d 6 - 0.0324 G W (3) 

d 2 = d + 0.0324 G W (4) 

Eliminating between these four equations we shall find 

J.-0JL8 y *^*'(r-?L±*) (5) 

W = 2.78 — ° cos ^ . Tl ^l^ (6) 

Go r o 

If we have r = ^(x l + t 2 ), the pressure at the center remains 
constant ; if t is greater than the mean of z x and r 2 , the pressure at 
the center increases; if r is less than this mean, the pressure at the 
center diminishes. 



VOL. 51 

For example assume 

U = 30 m , 

4>o = 72.4°, 

r = 4°, 

G = 6.5 mm 

T t -T a » 10°; 

i have W about io m . 


l/ = 50 mm , 

0o - 74.7°, 

U = 0°.l, 

£ = 246 ,nm 

T t - T, - 2° 

we have W about 3™. 

In cyclones of the temperate zones the radius r is generally so 
great that we can approximately calculate the ratio t/ : G by the 
formulae deduced for rectilinear isobars. In this case the quantity 

2.78 pr cos (p depends on the latitude S and on the coefficient of 

friction k. Denoting this quantity by B we have 

W = B 

The value of 5 is given in the following table: 


k= 0.00002 

«= 0.00004 

k = o. 00006 

fc=o. 00008 








n. 5 








FIG. 34 

to a c and 6 / parallel to a c. 

As to the direction of the pro- 
pagation of a barometric mini- 
mum that depends on the tra- 
jectory of the alimentary air. Let 
c c' in fig. 34 be the direction of 
the velocity of propagation W; 
the barometric minimum moves 
from c to c' in an infinitely short 
time d t; at the same time a par- 
ticle of alimentary air moves from 
a to a' with the velocity U. Draw 
c b parallel to c' a! '; a' b parallel 
to c c', a'd and b e perpendicular 




c c' = d a, a a' = d s, c a = r, c' a' = r + d r; 
the angles c' c a = <p and c a a' = <p. 
We have 

a c b = — d (p, 
e a = — d r 
ad = ae — bf 
a' d = b e + a' f 

By substituting the values of these quantities we shall find 


cos (p. ds 

= — dr — cos<p .d a 


sin (J>.ds - 

= — r d <p + sin <p.d a 


From these equations by substituting 

d s = Udt and d a = 


we find 

dr U cos (/> + W 

cos <p 

rdip U sin <p — W 

sin <p 

and consequently 

U sin (p.d r 

— U r cos (ft. d <p = W d (r s 


Supposing that U r cos (p is constant and that the angle <f' * s con_ 
stant and equal to a as in permanent cyclones, then by integration 
and determining the arbitrary constants so that <p = o for r = r Q 
we shall have 

U r cos a 

tang a log. nat. — — <p 
r n 

= W r sin <p . (9) 

By this equation we can determine the angle <p that alimentary 
air must describe in order to reach the interior limit of the cyclone. 

The equations that we have developed apply also to the upper 
strata of an anti-cyclone where the barometric minimum occurs. 

First example. 



= 6 ; <p = 20° 

By equation (9) we calculate the following values: 

a = 40° 50° 60° 70° 


jy = 0.44 0.57 0.68 0.77 



VOL. 51 

This case is that of cyclones that move nearly parallel to the ali- 
mentary stratum and where the alimentary air describes a very 
small angle in order to reach the interior region. 

In the northern hemisphere the wind deviates to the right and 
turns around the center aeainst the sun and consequently the cy- 




FIG. 35 

FIG. 36 

clone moves around the alimentary stratum with the sun (see 
fig- 35)- I n tne southern hemisphere the inverse phenomenon 
occurs (see fig. 37). 

When the cyclone passes any point, the temperature increases at 
first, but during the passage of the center it lowers (see figs. 36 
and 38). 

Since in general the mean or normal isotherms do not deviate 

much from the direction of the parallels of latitude of the terrestrial 

globe, we must expect the cyclone to be formed on the south of the 

supplementary air and on the north of the alimentary air and 

also that it move in general from west to east. 

Second example. Assume ~ = 10 and u> = 200 . By equation 

(9) we shall find 



U 9 

- 40° 
= 0.35 





In this case the cyclone also moves nearly parallel to the 
alimentary stratum, but the alimentary air describes about half 
a revolution around the cyclone before reaching its interior region. 


Therefore in the northern hemisphere the cyclone moves around 
the alimentary stratum against the sun and inversely in the south- 
ern hemisphere (see fig. 37). When the cyclone passes by any point, 
the temperature is lowered at first and then increases, and during 
the passage of the center it is lowered again to finally increase 
(see fig. 38). 



FIG. 38 

The cyclones of the inter-tropical regions, at least certain of them 
described by the meteorologists of the East Indies, seem to belong 
to the last class. However, the thermometric and hygrometric 
observations in the cyclones of low latitudes are unfortunately 
still too rare for it to be possible to determine the position of the 
alimentary stratum and the extent of the arc traversed by the ali- 
mentary air before it commences to ascend in the anterior portion of 
the interior circle. 

§32. Isobars of a variable cyclone 

We shall distinguish three cases: 

(1) Stationary cyclone. 

The isobars of a stationary cyclone are concentric circles that 
change their size at the same time that the barometric minimum 
varies. Consequently, the curves of equal variation of pressure are 
also concentric circles. The variation of the pressure d is a func- 
tion of the distance r. We can approximately determine this vari- 
ation by calculating two cyclones whose parameters are different. 
For example, suppose that the radius r be the same in each and that 
the maximum velocity U diminishes during a certain time. By 
the formulae of § 13 and § 14 we have calculated the following tables, 
assuming 8 = 50 and k = 0.00010: 


VOL. <l 







b -b 




o.o m 

o.oo mui 

o.oo mm 

o.o ra 

o.oo ,nm 

o.oo mm 









12. O 



1 1 . 2 





2 .90 


16. S 

2. 69 



20. O 

3 -30 



3- 17 

1 1 .00 


l8. 7 


1 5. 03 



14. 10 



2.5 5 

20. 7S 



19. 52 





11. 8 

1 .96 



10. 7 


29. 19 

10. 1 















1. 26 




1. 20 



1. 13 


Assume the radius of action of these cyclones to be about 20 and 
that at this distance the absolute pressure is 76o mm . The pressure 
b at the center is then in the first cyclone 72 2.i3 mm and in the second 
724.4o mm , and the increase of the pressure at the center is 2.27 mm 
at the same time that the maximum velocity has diminished i m 
from 20.0 to 19.0. Adding b to b — b we shall find the pressure b 
and consequently we calculate the increase of pressure at each dis- 
tance r. Assuming that the change has taken place in 4 hours, we 
find the hourly variation d by dividing the increments by 4 as 










o.57 ulm 


o. 4 i mm 


o.25 mm 

1 6° 

o.09 mm 








0. o5 


0. 5o 







FIG. 39 

By constructing a curve repre- 
senting d = f (r) we easily deter- 
mine the distance r for successive 
values of d and can construct 
the curve of equal variation (see 

%• 39)- 

(2) Moving cyclone with pressure 

constant at its center. 

When the pressure at the mov- 
ing center remains constant, the 
variation of the pressure at any 
fixed station is determined by equa- 
tion (4) of §30 and we have 

d = 0.0324 G W cos y 


Assume that W, the velocity of propagation of the center, is con- 
stant. For the value y = 71/2 we have d = o, that is to say, the 
curve of no variation is a straight line that passes through the center, 
and is perpendicular to the direction of propagation of the center. 

Assuming y = o and y = n, and also G = G Q we obtain the maxi- 
mum value of d, which consequently falls at two points at the dis- 
tance r from the center along the trajectory of the cyclone. 

The curves of equal variation are determined in general by the 

G cos y = constant. 

We can easily construct these curves by the aid of the curve of 
the gradient. 

In the interior portion, we have the equation 

G = G x r 

and the curves of equal variation assume the form 

rcos;- = constant. 

These curves are straight lines perpendicular to the direction of 

By using the values of G given in the preceding table we have 
constructed, for every o.2 mm , the curves of fig. 40, assuming W = 
io m . 

If we wish to construct curves of equal variation of pressure for 
any date whatever, we can construct two systems of isobars ap- 
propriate to the given date, and then determine graphically the 
curves of equal variation. It is evident that by choosing two appro- 
priate dates so far apart that the distance between the centers 
exceeds the diameter of action, 2 r, the curves of equal variation 
and the isobars themselves become identical and the maximum 
variations are the centers of the two systems of isobars. 

(3) Moving cyclone with variable pressure at the center. 

When the pressure at the center varies, the variation of pressure 
is determined by equation (3) of §30 and we have 

*=• d + 0.0324 G W cos y 

The variation d which is a function of r, is determined as we have 
shown in the first case where the system is stationary. Assuming 
W = io m and introducing the values of G and of d given in the 
preceding table for the first case, we shall find d as follows: 



VOL. 51 


r =o° 

= 30 

= 6o° 

= 90° 

= 120° 

= l5o° 

= 180° 


o.86 mm 

o.82 mm 

o.7i mm 

o.5S mm 

+ o. 3 9 mm 

+ o.28 mm 

+ o.2 4 mm 





0. So 

+ 0. 19 

— 0.04 

— 0. 12 


i. 3 5 

1. 22 



— 0.06 

— 0.40 

— 0.53 



1. 29 



— 0. 16 


— 0. 70 



1. 22 





— 0.69 





O. 25 

— 0. 16' 







0. 19 

— 0. iS 








— 0. i5 

— 0.36 







— O. J*4 


— 0.40 




0. 27 

0. o5 

— 0. 17 






0. 19 


— O. 19 



o. o 


- >t ~a¥ 

ft \ \ 

— H — ■ ' 1 

V / / / 

f ' I 

1 ' ' 

UP" •■. 


FIG. 40 


By the aid of this table we have constructed for every o.2 mm the 
curves of equal variation shown in fig. 41. 


FIG. 41 

We can also determine the curves of equal variation by construc- 
ting two systems of isobars for given dates. 

§33. Isotherms of systems of wind 

Permanent systems of wind demand a uniform temperature for the 
air that enters into the barometric minimum. Assuming that the 
temperature of the air varies with the pressure when nearest the 
surface of the earth, it is evident that a permanent cyclone must