ON THE SE

ATTERN OF

CATERPIIXARS AND PUPAE.

LIBRARY

t.V2 29 1952

UNIVERSITV OF CALIFORNIA

A. SCHIERBEEK.

ON THE SETAL PATTERN OF CATERPILLARS AND PUPAE.

ON THE SETAL PATTERN OF
CATERPILLARS AND PUPAE.

PROEFSCHRIFT TER VERKRIJGING VAN DEN GRAAD
VAN DOCTOR IN DE PLANT- EN DIERKUNDE
AAN DE RIJKS-UNIVERSITEIT TE GRONINGEN, OP
GEZAG VAN DEN RECTOR-MAGNIFICUS Dr. C. VAN
WISSELINGH. HOOGLEERAAR IN DE FACULTEIT DER
WIS- EN NATUURKUNDE, TEGEN DE BEDENKINGEN
DER FACULTEIT TE VERDEDIGEN OP ZATERDAG
20 JANUARI 1917, DES NAMIDDAGS TE 3 UUR, DOOR
ABRAHAM SCHIERBEEK, GEBOREN TE LEEUWARDEN.

N.V. BOEKHANDEL EN DRUKKERIJ
vooRHEEN E. J. BRILL, Leiden 1917.

BOEKDRUKKERIJ VOOrheen E, J. BRILL. LEIDEN.

AAN MIJNE VROUW.

Aan het einde gekomen van de bewerking van mijn proofschrift,
zie ik nog gaarne eons torug naar den 14d«n Augustus 1914. Midden
in den pas uitgebroken wcreldoorlog schoen het mij haast een bc-
spotting om met nicuwen moed aan een mij tota^I onbekend
onderwerp te beginnen. Ilet kwam mij soms wel wat dwaas voor,
om, waar tallooze, de gansche werold beroerende problemen ook mijn
aandaeht bezighielden, al de cnergie, die niet verbruikt werd door
myn meer dan 35 lesuren per week, te bestedcn aan de studie van
de plaatsing der harm op de rupsen.

Mocht ik hierdoor wel eens wat terneergodrukt worden, dan kwam
te juister tijd een Uwer opwekkende brieven, Hooggeleerde van
Bemmelen, om mij weer met lust aan het werk te doen gaan. Nu
ik mijn arbeid, naar ik hoop voorloopig, heb afgesloten, en ik de
met U gevoerde correspondentie nog eens doorlees, nu wordt het
mij weer duidelijk, dat veel wat ik geheel als eigen werk beschouwde,
toch zijn ontstaan te danken heeft aan Uw belangstellende vragen.
Ik hoop, dat Gij niet alleen als Promotor bij de bewerking mijner
dissertatie, maar ook later, mij Uw welgemeende raadgevingen en
hartelijke belangstelling zult blijven schenken. Voor de hulp, mij
bij dit w^erk verleend, voor Uw uitstekende raadgevingen bij het
vertalen in het Engelsch, en voor de gelegenheid, mij door U ge-
schonken, om zelf te ondervinden „hoe wetenschap gemaakt wordt"
ben ik U ten zeerste dankbaar.

Hooggeleerde Bonnema, als gymnasiast van nauwelijks dertien
jaren kwam ik reeds onder Uw leiding. De groote liefde voor de
natuurlijke historic in haar geheelen omvang, die Gij bij Uw leer-
lingen opwekte, heeft velen ertoe gebracht om deze wetenschap tot
het vak hunner keuze te maken. Waar ik nu reeds eenige jaren
als leeraar Uw lessen in praktijk breng, daar streef ik ernaar het

zelfde te verkrijgen, wat Gij bereikt hebt. Ik hoop, dat Gij mij
Uwe vriendschap ook verder niet zult onthouden. Het vele, dat ik
U verschuldigd ben, als gymnasiast, als student en als leeraar, zal
ik steeds in dankbare herinnering houden.

Hooggeleerde Moll, ik hoop, dat Gij in dit proefschrift iets van
de methodische en zuiver wetenschappelijke behandeling van een
onderwerp terug moogt vinden, welke Gij steeds als eerste eisch
voor degelijk en vruchtdragend onderzoek aan ons hebt voorgehouden
en waarvoor Gij U zooveel moeite getroost hebt om ze Uwen leer-
lingen bij te brengen.

U, overige Hoogleeraren der philosophische faculteit, wier colleges
en practica ik heb mogen volgen, ben ik zeer erkentelijk voor alles
wat ik van U heb geleerd.

Hooggeleerde Heijmans, Kapteijn en van Wijhe, veel van mijn
vorming ben ik U verschuldigd. Als gast op Uw colleges heb ik
veel geleerd en ik hoop, dat Gij hiervoor een woord van dank wel
wilt aanvaarden.

Mevrouw Kuenen, Uw zoo welwillend aangeboden hulp bij de
correctie van den Engelschen tekst, is door mij op hoogen prijs
gesteld.

Uit vriendschap hebt Gij, Emil Gkuno, de moeilijke taak op
U genomen mijn studie te vertalen. Met voldoening kunt Gij op
Uw arbeid terugzien en Gij weet hoe zeer ik Uw werk waardeer.

Aan mijn Vrouw heb ik mijn proefschrift opgedragen. Tusschen
de beslommeringen van huishouding en artsen-praktijk hebt Gij
tijd gevonden mij op velerlei wijzen te helpen en een gedeelte
van mijn manuscript te vertalen. Ook op deze plaats past hier-
voor een woord van dank.

'sGravenhage, 7 October 1916.

CHAPTER I.

Introduction, Material and Method.

§ 1. Intrmluctioii.

The well-known treatise of Weismann (1876) on the Sphingid-
caterpillars has given rise to many recent investigations of the
Lfpidoptera. It very soon became evident that many important
discoveries could still be made about these well-known insects,
although they have been observed for centuries, even if one con-
fines oneself to external characteristics only. The studies of
Weismann on the seasonal dimorphism (1876), the rediscovery
of Ratzeburo's observations (1840) on the external sexual charac-
teristics of the pupae by Jackson (1890) and Poulton (1890),
the studies by Poulton on the antennae and wingsheaths of the
nymphae, the enlargement of our knowledge of the primary colour-
pattern on the wings of the butterflies by J. F. van Bemmelen
(1890), the investigations by Spuler (1892) of the wingveins and
by Walter (1885) and Chapman (1893 B) of the active mandibles
of Micropteryx are the most striking proofs, of how many im-
portant and successful investigations could still be made in the
morphology of the Lepidoptera, in the last decades.

W. MCller's (1886) treatise showed us the constancy in the
arrangement of the so-called primary hairs of the Nymphalid-cater-
pillars. In a supplement this writer points out, that the same
pattern occurs also in other families.

1

At Prof. J, F. VAN Bemmelen's suggestion I decided to inves-
tigate how far this assertion holds good for different families.
My purpose was to find an answer to the following questions :

1". Does a conformity exist between the different colours,
the pattern and the skin-relief of the caterpillars of the Rhopa-
locera, perhaps also of the Ileterocera?

2°. Is it possible to deduce from this conformity a general plan
of the caterpillar pattern, and if so, does this possess a metameric
character ?

3". Can any connection be found between the pattern of cater-
pillars and that of pupae, perhaps even with the colourpattern on
the body of the imagines?

In thinking these questions over, I soon found that the following
points are connected with the former three, viz:

4". Is the arrangement of the hairs on all segments the same
or is it different; and in the latter case, what are the relations
of these differing segments in other anatomical respects?

b'^. Is the arrangement on one individual in all stadia the same ?
6". Have all the individuals of one species the same pattern?
Not until I had solved these last questions, could I expect to
find an answer to the first three. A priori I might expect a
certain constancy of the setal pattern. The investigations by
DE Meyere (1894) of the hairs of the mammals, by Megusar and
Werner of the spots on Salamandra maculosa^ gave rise to the
supposition that here also a constancy might be expected. In the
beginning of August 1914, I was not yet acquainted with the
extensive literature on the setae of the caterpillars. Spuler's remark
(1910, p. VII) did not make me think that the primary pattern
would be maintained with such pertinacity.

The 6tt question I could drop very soon, as it appeared that
the fluctuation or individual variation (Plate, 1914, p. 148) is
so insignificant, that it might practically be neglected. This
fact apparently so simple, indicates already the great constancy
of the pattern of caterpillars (see however Papilio machaon). During
my investigation it turned out that not all the colours of cater-

8

pillars can stand the influence of alcohol. I therefore had to conHno
myself to the arrangement of the setae. For the solution of the
third question I could only collect a few data which will be
mentioned under the species of which I have exaniined the pupae.
In chapter VII I intend to come back to this point. As only tlioso
naturalists, who have a very extensive collection at their disposal
can fully solve this problem, I am obliged to leave the further
investigation to others.

In working out the different questions I became convinced that
the 4''» and the 5»l> are the cardinal points of the investigation.
Through the 4^^ question especially I came into contact with pro-
blems, which entirely differed from those I had originally thought
of. Here too I wish to thank Prof. J. F. vax Bemmelen for the
manner in which he encouraged me to enlarge the subject of my
investigations, and for the way in which he inspired me with
interest for it and also for his kindness in assisting mo with his
knowledge of the extensive literature.

§ 2. Material and Method.

There are a great number of books about caterpillars, with
coloured and uncoloured figures, but most of them are unsuitable
for my purpose. In general the figures of caterpillars are still
more inaccurate than those of their imagines, of which van
Bemmelen said (1913, p. 107), that he found one great difficulty
in his work on the colourpattern of the body of Lepidoptera', viz:
that neither most of the existing figures of butterflies and moths,
nor the dried specimens of the collections were suitable for a
more exact analysis of this part of their colour-pattern, and that
this was the same with the very few figures of the nymphs and
caterpillars in the entomological illustrated publications. I cannot
but agree with these words, making an exception for Packard's
standard-work on the Bombyces (1895 — 1914). As it would be
too expensive to make such large photographs of each caterpillar
as VAN Bemmelen did for the nymphs of Papilio machaon and
podalirius, Vanessa to, Pieris brassicae and napi, Aporia crataegi^

Euchloe caydatnines, Gotio])teri/x rhamni and Thais polyxena^
I have only made Indian ink drawings, with the aid of the
„Zeichenapparat nach Abbe (Zeiss. Jena)." For the composition of
the plates, these enlarged drawings were all reduced to the same
size: +10 cm., by photographic reproduction. In consequence the
smaller and younger specimens are represented on a much larger
scale than the bigger and full-grown animals, but their actual size can
always be accurately ascertained by the scale given with each figure.

These figures have the advantage that they reproduce the real
aspect of the animals.

The implantation and in most cases also the length of the setae
has been traced exactly from the fresh or conservated specimens.
Except where a homogeneous spreading of the setae occurred and
I had to work according to a scheme, I could always use this
method. This method of drawing from life with a magnifying
apparatus, and of reducing these enlarged figures to a certain
standard length, seems to me preferable to that followed by Tsou
(1914). This writer describes a method by which the length and
the breath of the setae can be determined. In chapter II I shall
refer to his work.

In my method, which has also been applied in Packard's work
on the Bombyces (1895 — 1914), the growth in length becomes
so to speak eliminated, and therefore the variations of the pattern
become more distinctly visible. At the same time these figures
can be used to study the growth in thickness of the successive
segments in the different instars, as the correlations of growth
fully deserve to be studied further. I believe that up till now only
the head has been studied in this manner. The well-known leaps
in the changes of the size of the head have often assisted me to
determine, the moment of the moult, in cases where this was
not mentioned with the preserved material. Many investigators
have only turned their attention to the fullgrown caterpillars. It
therefore seemed to me an interesting subject to examine the
placing of the setae in all instars as accurately as possible. The
difference of the results of my investigations and of those of many

others must bo attributed fur a great part to their not having
examined the younger instars.

I should have liked to study the living animals only, but this
proved to be too difficult. Therefore the animals were nearly
always preserved in alcohol of 96'Y„. As I have already said, this
causes the cplour to disappear for the greater part. Nearly the
wliole material has been cultivateii in the Zoological laboratory
in the University of Oroningen. Mr. E. TlIEY^^8EN, the attendant
of the I^aboratory, was charged with the care of the living animals
and with preserving them; and here I wish to thank him for the
trouble he has taken. IJefore the investigation the preserved cater-
pillars were stwked in glycerine-gelatine. A certain quantity of
this substance in a solid state was placed on the object-glass, on
which the caterpillar also lay, moistened by alcohol. The glycerine-
gelatine on the object-glass was somewhat heated till it became
entirely liquid, when it mixinl with the alcohol. Thus I obtained
preparations which did not shrivel up and which on the whole
kept very well. Very thick caterpillars had to be examined in
a dry state or lying in a watch-glass with alcohol, sometimes after
the hairs had been cut very short.

Besides this material I could dispose of the magnificent collec-
tion of Dr. F. W. O. Kallenbach at Apeldoorn. This very rich
collection has been presented by the collector to the Zoological
Laboratory in the University of Groningen. Besides the numerous
imagines it contains the mounted and dryed caterpillars of most
of the species and of many even more or less complete series
of the stages of development. The specimens which I have used
are indicated by Coll. Kail. Many entomological plates I have
looked through, and though the objections already mentioned could
not be discarded, the figures sometimes gave valuable indications
as to the direction in which the further investigation had to be
made. The works indicated with an asterisk give the most correct
figures and descriptions. As a proof of the unreliability of the
figures in scientific entomological works, I draw attention to the
figure Taf. Ill, fig. 38 in Weismann's before mentioned study

of Deilephila euphorbiae, which bears a stigma on the meso-
thorax as well as on the metathorax.

A list of the illustrated works I consulted follows, the exact
titles are to be found in the bibliography; the date given refers
to the beginning of the publication.

* Buckler 1886.

SCUDDER 1888.

HoFMANN 1893 see Spuler.

* Packard 1895.
Beutenmuller 1900.
Forrester 1907.
TONGE 1907.

* Spuler 1910.

De Reaumur 1737.
Sepp 1762.
HCbner 1786.
Ratzeburq 1840.
duponchel 1849.
HoRSFiELD and Moore 1857.
*MlLLlfeRE 1858.

Wilde 1861.

In chapter VI I have given a systematic synopsis of the cater-
pillars which I have examined. Of those families, of which I had
no specimens at my disposal, I have given an account taken from
the literature on the subject.

CHAPTER 11.

Literature.
The studies by Weismann which I have mentioned before, were
made to prove the correctness of the ideas introduced by Darwin
on the transformation of the organisms. He had two reasons for
his choice of the pattern of caterpillars as a test-object for his
theoretical conceptions:

1. because with them sexual selection is out of the question.

2. because only the colouring of caterpillars was considered to
be of value for the life of their bearers and not the pattern, which
has nothing to do with the colour.

For special reasons Weismann confined himself to the Sphin-
gidae. His terminology is as follows:

1. Linea dorsalis, placed in the middle of the dorsum.

2. Linea stigmalis or linea suprastigmalis and infrastigmalis.

3. Linea subdorsalis, just between 1 and 2.

It 18 well known how Weismasx brought back the eye spots
and the ringed spots to the linea subdorstilis. He distinguishes four
ontogenetic and phylogenetio stagi^s in the course of development
of the colourpattern during larval life.

Stage I. Qreen, without any pattern.

Stage II. Subdorsal line, sometimes also a dorsal and a stigmal line.

The biological value of this stripe was that it divided the
strikingly large body of the caterpillar into parts and in that way
made it less conspicuous.

Species showing this longitudinal striation lived on grasses and
conifers.

Stage III. Cross stripes I.e. p. 127.

New characters arise only during the ultimate stages of the
larval life and when new ones are developed, they disappear from
the last stage and arise in the former one. The character of the
cross stripes becomes completed by accompanying coloured borders
(shadow).

Stage IV. The eye spots (with a dark pupil = central spot, a
bright shining spot and a dark ring) and the ringed spot (without
central spot) arise from or in connection with the subdorsal line,
on the fourth and fifth segment of Chaerocampa (1. c. p. 97) and
on the eleventh caudal-horn segment of Deilephila. The spotted
pattern is a warning colour. Weismann was able to point out a
biological meaning for these three principal elements of the sphin-
gidal pattern and thus he could explain their origin by natural
selection. For the explanation of the repetition of a locally origi-
nating pattern on the other segments, however, he had to refer
to the rule of correlation (I.e. p. 136).

Weismann declares positively that the first stage has no
pattern e. g. Deilephila euphorhiae (1. c. p. 25). „When, however,
the youngest larvae of this species are scrutinized with a high
power, it is seen that from the beginning they are dark-green,
while the horn is black, so also are the head, the feet and a
semi-circular chitinous shield on the dorsum of the prothorax and
one paired and two unpaired chitinous shields on the last segment.

8

As yet no trace is to be found of the pattern, which appears
later on. The stigmata are visible as white spots. On each segment
there is a number of warts (in most cases ten) each of which
bears a simple bristle. When the small caterpillars have obtained
a length of 7 mm. they are olive-green and no longer form such
a great contrast with the green Euphorbia leaves as before; still
they do not possess any definite pattern. After five days the first
moulting takes place and with it a very complicated pattern sud-
denly appears (fig. 38—39)."

Weismann also mentions a full-grown Smerinthus{?) species
in the museum in Berlin of which he says, „that it is sparsely
covered with bristles but does not show a trace of any pattern, and
agrees all the more with the youngest stage of most of the now
living Spinr/idae as it also has short bristles thinly spread over
the surface of the animal. This ^living fossil" had a length of 6 cM."

For the rest the pattern of the hairs resembles that which I
found amongst others in Sphinx liguMri and Smerinthtis tiliae.

Weismann's classical treatise has rightly met with much appre-
ciation but, unfortunately, has found too little imitation.

For a more exact insight into the system of the Lepidoptera a
complete knowledge of caterpillars will without doubt prove to be
of great value.

The accuracy of Weismann's investigations and the great keen-
ness with which he has deducted very comprehensive theories
from apparently unimportant facts, guarantee to his work a pro-
minent place in zoological literature.

In his next study Weismann (1876 11,) discusses the so-called
„ parallel rows". He starts from the following argument (1. c. p. 141) :

„If the development of the organic world depends upon a
phyletic vital power, there must have taken place and still be taking
place what I call „phyletic parallelism", i. e. the development of
the two stages of metamorphic species must have taken place in
exactly parallel direction; each transformation of the butterfly
would have been accompanied or followed by a transformation of
the caterpillar, and the systematic groups of the butterflies would

0

be found again in just the same way in a system of the larvae,
or in other words: the relation of forms of the caterpillars must
harmonise precisely with that of the butterflies."

„If the development is only the reaction of the specific or^^a-
nism to the influence of the outer circumstances, dissimilarities
in the phyletic development of the different stages of life might
be expei'ttMl" ... \ congruency might be the consequence of
correlation.

Weismanx goes on to say (I.e. p. 157) "that the primitive
cause of variations when coming from outer circumstances must
occur far oftener with larvae than with butterflies."

Darwin points out the heridity in corresponding ages or as
Hakckki. calls it: homochronic heridity.

Wkismaxn thinks he has here found an explanation for the great
differences between larva and imago (p. 168) „a8 the acquirements
of the separate stegw in the following generations are always
transferred to those stages themselves but the other stages remain
untouchetl".

After having discussed the different families, to which I shall
return later on in discussing the groups, Weismann comes to the
following conclusion : „there is a great congruency between the
system of larvae and that of imagines, especially where the genera
are concerned, but the incongruencies appear mostly with varieties
and families". To the questions, what may be the cause of the
difference in form of butterflies and moths being so much greater
than that of their caterpillars, and why the imagines of the
Rhopalocera have so many characteristics in common which their
caterpillars do not possess, Weismann gives the following answer
(1. c. p. 195) : „that this might be explained by the great differences
in the manner and duration of life of the imagines."

"Weismann's study was followed by a great many others, some
of which I intend to discuss with the families. Here I will only
give a short index of those which are of general importance.

Wilhelm Muller is the first (1886) who pays special attention

10

to the arrangement of the „hair8" on caterpillars. It is true that
DE REAUMUR had pointed out in 1736 the peculiar warts of different
caterpillars and the plumed hairs they bear, also that Milliere
added a drawing of the back part of the body with the hairs on
the last three segments (1858 I, PI. I, fig. 4) to his description
of Coccyx junipera Mil., but no systematic investigation of the
arrangement of the hairs had, as far as I know, taken place before
W. MOller.

Before going further I wish to observe that the „hairs" of the
insects are mostly developed as offshoots of a hypodermis-cell.
They are absolutely different in construction from the hairs of
mammals and therefore the word setae has been introduced for
them by Lankester. Consequently I shall only use the word ^hairs"
in this study, in cases where the writers quoted do not mention
the word setae for some reason or other. Fracker (1915, p. 38)
thinks that the setae are sensory in function. As with all kinds
of other organs the form of the setae of the caterpillars often
gets more intricate after each moult, so that it is of great impor-
tance to examine all the succeeding stages, and thus to get a good
insight into the covering of the skin of the full-grown cater-
pillars. To avoid confusion between the two meanings of the
word stage (the caterpillar or larval stage, the pupal stage, and
the period between two moults of caterpillars), Fischer has intro-
duced the term instar for the last mentioned meaning. The larval
stage therefore consists of several instars.

MuLLER began to pay attention to the first instar of the Nymphalid-
larvae. In them he discovered a constant arrangement of the so-called
„primary bristles" which he gave the numbers 1 — 6 (see Nomen-
clature). Bristle 6 only occurs on the segments 2 — 5 and 10 — 12.

The segments 1, 2 and 3 (the thoracic rings) are different and
so is segment 12. A comparison of the bristles proves that
there is a special segment 12a, but this only exists during the
first instar. It is very easy to homologize the bristles on the
segments 4 — 12 (the abdominal ones) but the mesothorax and meta-
thorax are widely different, a shifting may perhaps have taken

11

place in their arrangement in connection with the development
of the wings and the disappearance of the stigmata. To this I
shall refer again in chapter V.

Just before the first moult white spots glimmer through the
skin, some of which will grow into secondary bristles (scoli) and
others are destined to become the white warts which are the
cause of the j>attern.

The starting point of these secondary bristles does not coincide
with that of the primary ones (I.e. p. 110 and fig. 14, PI. 3),
very often they are median and consequently unpaired, which is
never the case with the primary ones (See PI. 1, fig. 1).

The pattern of the caterpillar often passes on to the pupa, the
cuticular pigments disappear and the subcuticular ones remain
(l.c. p. 231).

la contradistinction to WEisMJOnr, MClleb (I.e. p. 232) says
that the new characteristics which have ap|)eared during the larval
stage are shifted on to the pupal stage and the other way about.

With each moult the scoli of the Nymphalinae get more and
more intricate. He calls them after the line on which their base
is fixed. In an appendix MCller mentions the observations on
the origin of the scoli known up to that time (1. c. p. 250), and
also shows that the pattern of the primary hairs is found again
in other families. He then says:

Spines arise:

1. As independent elevations, without any relation to setiferous
warts: horns of the Xymphalidae, pseudo-spines of Caligo and
DatMis, gills of Cataclysta and Paraponyx.

2. By transformation of warts bearing bristles ; viz. :

a. Of the warts of primary bristles, probably the most common
method (forked tail of the Satyridae, pseudo-spines of the Papi-
lionidae, spines of the Saturniadae, tail-horn of the Sphingidae).

h. Of the warts of the secondary bristles (spines of the Nym-
phalinae).

The biological meaning of the spines cannot be for defence
against caterpillareaters (1, c. p. 93), for caterpillars with large

12

and numerous spines are persecuted just as much as those
without them.

In his derivation of the spines from the secondary bristles and
not from the primary ones, Muller is in contradiction with Gtruber
(1884). In a discussion of the American Papilionidae and Nijm-
phalidae Gruber says (I. c. p. 476). „The Papilionidae often
possess in the first instar warts with spoon-shaped setae and some-
times with forked ones which become smaller during the moults
and which have sometimes totally disappeared after the first moult.

„The Nijmp>halidae have small bristles in the first instar, each
apart on an elevation and later on we see large warts with numerous
bristles, completely agreeing in number and place with the elevations
of the first instar. (See Papilionidae and Nt/mphalidae, Chapter VI)".

ScuDDER (1888) too gives a description of the caterpillars.

He names the setae, spines etc. after the lines which connect
them with each other and he distinguishes twelve paired and two un-
paired lines viz. the dorsal or mediodorsal^ subdorsal, laterodorsal,
supralateral, lateral^ infralateral, laterostigmatal, suprastigmatal,
sti(/matal, infrastigmatal, ventrostigmatal, lateroventral, subventral,
ventral or medioventral.

The italicized lines divide each side of the body in three parts
which are about equally large.

ScuDDER (p. 12) observes: „A11 of our butterfly caterpillars are
clothed with hairs . . . ., their arrangement affords admirable generic
characteristics which have not hitherto been sufficiently appro-
priated. It should be stated that juvenile caterpillars in their first
stage may be safely said to differ generically from themselves at
a mature epoch. The hairs, spines etc. are placed in transversal
and longitudinal rows, the former are subordinate to the latter".

On p. 235 ScuDDER says his opinion on the ancestors is that
the surface of the body was profusely covered with little papillae
from each of which sprang a minute simple hair. In harmony with
this he says that the wings too of the first butterflies were uni-
formly dark brown. In the course of this paper I hope to prove
that a homogeneous spreading of the setae is not a primary

feature, whiUt for the colour of the wings of Lopidoptera I refer
to J. F. VAN Bemmelkn (1889—1916), J. Botkk (1916), J. H. dk
Meyere (1916).

The next study which is of great importance for the know-
ledge of the setae is that of Packard (1890). Without paying
much attention to the arrangement of the setae, Packard especially
devoted himself to their different shapes. What struck him especially
with the Bombyces, was the intricate shape of the setae of the
full-grown caterpillars. He examined their ontogenesis and thereby
was led to the establishment of a scries of types (1. c. p. 512 sqq.)
of tubercula and setae which are often used even now in des-
criptions. Packard's list, which has also been included in the
great work (1895) on the Bombyces, may be cited here:

y,A. Tabeirles.

a. Simple and minute, due to a slight thickening of the hypo-
dermis and a dei-ided thickening of the overlying cuticle; the
hypodermis contains a large unicellular gland either for the secre-
tion of the seta or for the production of poison.

1. Minute piliferous warts (Most Tineid^ Tortricid and Noctuid
larvae).

2. Enlarged smooth tubercles, bearing a single seta.
(Many Geometnd and Bomhycine larvae).

3. Enlarged, spherical tubercles, bearing a number of setae,
either radiated or subverticillate (Ardians, Lithosians, Zygaenidae,
including some Glaucopinae).

4. High, movable, smooth tubercles, having a terrifying func-
tion {Schizura, Xylinodes, Notodonta, Nerice).

5. Low and broad, rudimentary, replacing the „caudal horn"
{Choerocampa, the European Pheosia dictoea and dictoeoides).

h. More or less spinulose or spiny (disappearing in some Sphingids
after stage /).

1. Long and slender, usually situated on top of the eighth
abdominal segment, with microscopic spinules in stage /. (Most
Sphingidae and Sesia).

14

2. Smooth, subspherical warts (Zi/gaenidae e. g. Chalcosia, East
Indies) or elongated but still smooth {Attacus atlas and a species
from South-western territories U. S. A.).

3. Subspherical or clavate, spiny tubercles of many Attaci,
the spinules usually short.

4. Spinulated spines or elongated tubercles of Ceratocampidae
and Hemilucidae {H. io and H. maia etc.).

5. Spikelike hairs or spines (Samia cynthia, Anisota, East-Indian
Hypsa, Anagnia).

6. Antler-like spines. Early stages of Heterocampa biundata,
(juttivitta and obliqua.

B. Hetae (^hairs", bristles etc.).

1. Simple, fine, short or long, microscopic or macroscopic setae,
tapering hairs, scattered or dense, often forming pencils (Many
Bombyces, Zygaenidae, Noctuobombyces, Apatelae).

2. Glandular hairs, truncate, spindle-shaped or forked at the
end and secreting a more or less viscid fluid (Many in stage I
and II of Notodontians, many butterfly-larvae and in the last
stages of Pterophoridae).

3. Long, spindle-shaped hairs of Apatelodes, Apatela americana,
and the European Tinolius eburneigutta Walk.

4. Flattened, triangular hairs in the tufts or on the sides of
the body of Gastropacha americana, or flattened, spindle-shaped
scales in the European G. quercifolia.

5. Spinulated or barbed hairs. Most Glaucopides^ etc. Arctians,
Lithosians and lAiJaridae and many other Bombyces.

C. Pseudo-tubercles.

1. The filamental anal legs (stemapoda) of Cerura and Hetero-
campa morthesia.

2. The long suranal spine of Platyptericidae."

Packard's view on the origin of these different forms is the
following (1890 1. c. p. 560) :

1. The more prominent tubercles and spines or bristles arising

15

from them, are hypertrophied piliferous warts, the warts with the
seta or hair which they bear being common to all caterpillars.

2. The hypertrophy was probably primarily due to a change
of station from herbs to trees, involving better air, a more equable
temperature, perhaps a different and better food.

3. The enlarged and specialized tubercles developed more rapidly
on certain segments than on others, especially the more prominent
segments, because the nutritive fluids would tend to more freely
supply parts most exposed to external stimuli.

4. The stimuli were in great part due to the visits of insects and
birds, resulting in a mimicry of the spines and projections on the
trees, the colors (lines and spots) were due to light or shade, with
the general result of protective mimicry or adaptation of tree-life.

5. 6, 7, 8 and 9. Through heredity these first steps in the
evolution, in the beginning due to primary factors of evolution
(Xeo-lamarckism) became constant, due to segregation and natural
selection, bi'oause intercrossing with low feeders would cease.

As the probable time of the origin of the large setae and warts
Packard mentions : „the critical time attending or following the
close of the Palaeozoic or the early part of the Mesozoic age,
the time when deciduous trees and flowers probably began to
appear" (1. c. p. 506).

In 1893 Packard refers again to this subject, in which he
thinks he has found a basis for a natural classification of the
Bombyces (see chapter VI).

Dyar independently of W. MCller, examined the primitive
pattern in 1894 and has even made an analytical list for deter-
minations according to the setae.

As for the nomenclature I refer to Chapter V. His table
follows here.

Synopsis of the Families of Lepidopterous Larvae.
A. More than one tubercle on the third annulet and more than
three on the base of the leg. — Jugatae, Hepialidae.
A A. Not more than one tubercle on third annulet and only six
above the base of the leg. — Frenatae.

16

B. Three tubercles on middle annulet, none on the third.

Tubercles IV and V approximate, two thoracic shields,
Psychidae.
BB. No more than two tubercles on middle annulet and usually
one on the third annulet, one thoracic shield (prothoracic).
C. Tubercles IV and V approximated or consolidated.
Generalized Frenatae.

D. Tubercles simple, single haired. Cossidae^ Pyralidina,
ToHricina^ Tineina (in part), Lacosomidae, Sesiidae.
DD. Tubercles absent, as well as legs. Tineina in part.
DDD. Tubercles modified, many haired.

E. All present but tubercle I. Pterophoridae.
EE. Subventral tubercles also reduced, only three
left. Pyromorphidae^ Megalopyyidae.
EEE. Substigmatal tubercles absent, only two left.
Eucleidae.
,CC. Tubercles IV and V remote (sometimes IV disappeared
and then essentially the same arrangement as in EEE.)
Specialised Frenatae.

F. Tubercles all present or with a slight tendency to
unequal reduction, setiferous or equally reduced.
G. Simple with a single seta: Noctuidae (in part).
Ayaristidae, JVotodontidae, Geometridae, Drepa-
nidae, Lithosidae (in part).
GG. Tubercles with many hairs.

H. Without any development of hairs from
the skin. Noctuidae (in part), Pericopidae,
Arctiidae, Euchoniidae, Zyyaenidae, Lyman-
tridae.
HH. Tubercles greatly reduced, abundant hair
from the skin. Lasiocampidae.
FF. Tubercles with marked unequal reduction or greatly
modified or absent.
I. Tubercles still wartlike, hairy. (The young larvae
of many Papilionidae will also come in here).

17

//. Tubercles greatly modified or absent.
J. Tubercle I normal (when present).

A". Tubercles productnl into naked Heshy horns or repre-
sented by coloured spots. Papilionuiae, Xyniphalidae
(in part).
KK. No trace of tubercles. Xymphalidite (in part), Pieridae^
Hesperidae.
JJ. Tubercle / consolidated with its fellow on the dorsum.
L. No unpaired dorsal tubercle anterior to abdo-
minal segment 8.

M. Tubercles largely present. Saturnia.
MM. Only the dorsal tubercle on segment 8.
Sphingidae.
LL. A line of unpaired dorsal tubercles throughout
the length of the abdomen, anterior to segment
8f or largely so. Nymphalidae (in part).

What strikes us in the first place in this list is that several
families belong partly to one group, partly to quite a different one.
This might indeed be expected in an artificial system like this.
In the second place we see that the determinations of many families
with the aid of this table will prove to be difficult e.g. the Pieridae^
which are densely covered with setae, and which, though their
primitive pattern remains visible for a long time, are said by
Dyar to present: „no trace of tubercles".

The artificiality of such a classification is evident.

I think that Dyar in this case was under the influence of
Comstock's suggestive paper (1893), and that he exerted himself
to find a characteristic in the caterpillars, which allowed him to
apply Comstock's division of Lepidoptera into Jugatae and Frenatae
to the larvae also.

My third and main objection is that except in group I no
attention has been paid to the ontogenesis. Where the setal pattern
undergoes rather important modifications during the larval stage
and some members of the family remain on a lower scale of

2

18

development (as a classical example I cite the Sphingid with a
complete setal pattern, mentioned by Weismann), there it is
absolutely certain that a classification like this must prove to be
inefficient as soon as a great number of different forms of one
family are compared with each other. Fracker who in 1915 once
more tried to compose an analytical list for the determination
of caterpillars, has been obliged to use other characteristics, such
as the rows of crochets on the abdominal legs. At the same time
Fracker very logically begins with the most generalized families
and gradually passes to the most specialized ones (1. c. p. 49 — 59).
Fracker's main classification is no longer based on the setae but
on other characteristics. He introduces a completely new nomen-
clature, against which I intend to raise my objections in a following
chapter. It is a great pity that, where he apparently had extensive
material at his disposal, he paid so little attention to the onto-
genetic changes of the setal pattern. During the discussion of
the different families I shall have to point out some mistakes
in Fracker's work (see e. g. Piendae^ Bomhyx inori, Porthesia
chrijsorrhoea etc.).

Tsou who worked at about the same time as Fracker, published
a method in 1914 for determining the length and the breadth
of a seta on the segment. He examined almost exclusively a full-
grown Cossus cossuSj Hepialus humuli and Jaspedia celsia. He
chooses the prothorax as point of issue for his deductions, on the
not quite scientific ground that it is the first segment of the body.
(1. c. p. 228). He also groups the setae in a peculiar way of
which he himself admits that it is more or less artificial. I cannot
but agree with him in this qualification. As I do not agree with his
method of comparing fullgrown caterpillars of different families
with each other without attending to the first instars at all, nor
with his taking the prothorax as the starting-point, nor with his
uniting the setae to arbitrary groups, I think that here it is
sufficient simply to mention his work.

0. HoFMANN in 1898 devoted a study to the caterpillars of the
Pterophoridae. Dyar had brought forward the great systematic

i

19

value of the setiil pattern, but O. Hoffmann came to the con-
clusiou that in a perfectly natural family like the above,
the primitive pattern showed great divergences. In discussing the
family I shall return to this. A general importance is granted to
this study by A. Spi'LER (1910), who writes on p. VII of the
fourth volume of his well-known book on butterflies and cater-
pillars, that the setal pattern is evidently not of such great
value as Dyar believes.

J. Tn. Oi'DEMANs on p. 384 of his excellent work on the Dutch
Insects (1897 — 1900) remarks, in general terms, on the hairs as
they usually occur on most caterpillars.

He does not touch on their systematic value.

Ambrose Quail's notes on Cojjsf/iae 1904b, [he uses these animals
i. a. to determine the number of the abdominal segments (10),J
are followed by some general remarks. lie divides the caterpillars
into three groups. 1. c. p. 269.

I. single seta-tubercles in all stages. Hepialidae^ Cosftidae, Xoc-

tuae, Geometrae.
II. single seta-tubercles only in first stage: Pieridae, Spkingidae^

Nymphalidae, Arctiidae.
III. More rarely in first larval stage some tubercles with more
than one seta. Liparidae etc.

After his attention had been drawn to it by Mr. A. Bacot
(footnote p. 95) he says in his second treatise 1. c. p. 270 :

„I submit the homologue of II B of the thorax is a minute
anterior supraspiracular tubercle of the abdomen called by me
III 5, that Dyar's III of the thorax = a sub-spiracular tubercle
of the abdomen and so on." I shall refer to this in chapter IV.

Forbes in 1910 and 1911 also gave attention to the problems
I have just discussed. I regret not having been able to obtain
these papers. In the chapter on nomenclature I have discussed
what information I got about them from the quotations of other
writers.

It is peculiar that though Dyar as early as 1894 homologized
setae and tiibercula with the pigmental spots, it was not before

20

1912 that J. F. VAN Bemmelen succeeded in proving this sug-
gestion. This writer found that the pattern of the caterpillars of
Pieris brassicae might be retraced in the design of the pupae.
He then could recognize the same pattern on pupae of several
other Pieridae and also of Papilionidae and Nymphalidae and
even succeeded in discovering it on the bodies of the imagines.

In a following chapter (VII) containing the discussion of the
pupal pattern, I shall return to this important question.

It is perhaps due to the influence of Weismann and Eimer,
who considered the linear-pattern as the original one, that in-
sufficient attention is paid to theories which regard the spotted
pattern as the most primitive.

It is certain that Schroder (1894) in his study of the Geome-
tridae has been too much influenced by this preconceived idea.
This investigator has paid still less attention than Weismann to
the setae bearing tubercles and consequently has quite overlooked
the pigment accumulation at the base of the setae. For the dis-
cussion of the origin of the linear pattern and the primitive cha-
racter of the spotted pattern I refer to chapter IX.

J. C. II. DE Meyere in his recent paper (1916) arrives at con-
clusions which are in general the same as mine. In chapter VI
and VII I shall return to his paper.

CHAPTER III.

On the structure of the thoracic segments.

In several respects the construction of the thorax differs from
that of the abdomen, a difference which very early attracted the
attention of entomologists. Before discussing my nomenclature, I
prefer to investigate, which segments are the most primitive
ones, in other respects than the covering with setae. There are
reasons for supposing that segments of primitive construction in
other respects, will also display this original character in
their setal pattern. I cannot avoid mentioning many facts which

21

are generally known. As however the investigators, who have
written on the setal pattern, have not paid any attention to these
familiar facts, I think it may be useful to recall them to memory
and point out their connection with the arrangement of the setae.

The first who, I think, remarked the wing-rudiment in cater-
pillars, was Jan Swammerdam, who described it in Biblia Naturae
II, p. 615 and in 1668 showed it to the Duke of Toscane, to
Thevenot and to Magalloti. Ho described it under the title of:
„Animal in animali or the butterfly hidden within the caterpillar."

The same thing is found in Ilistoria generalis p. 202 as: „In-
sectum in Insecto seu Papilio in Eruca". In both cases Piens
brassicae was the object of the investigation. The next investi-
gator who mentioned the wing-rudiment was P. Lijonet who in
1 760 wrote his famous ,Traite anatomique" and on p. 592 says :
,The isolated figure towards the bottom of the middle of PI. XI
is a mass of white satin-like stuff, placed in fat without sticking
to it and which is attachtnl in B to the inner membrane of the skin.

There are four such lumps within the caterpillar (Cosxits), they
are found on either side of the 2nd and 3rd rings. They might
be the origin of the wings of the moth". And on p. 449: „It
is attached to the skin in the deep fold which it makes".

With his well-known accuracy Lijonet draws many more
muscles in the thorax than in the abdomen and many of them
are not connected with the thoracic feet. — I think therefore,
that other differences besides these legs have been established be-
tween the thorax and the abdomen, so that Henneguy's con-
tention (1904, p. 442), that in legless larvae all the segments
have exactly the same constitution, can only refer to external
features.

The presence pf the wing-rudiment combined with the absence
of the stigmata has induced many students to consider the wings
to be modified tracheae.

But a good many investigators who later on have examined
the development of the wing, have come to the conclusion that
the first change does not start from a trachea but from the skin.

22

As the principal studies on this subject I may mention those
of Landois 1871, Ganin 1876, Pancritius 1884, Dewitz 1887,
VAN Bemmelen 1889, Bugnion 1892, Gonin 1894, Mayer 1896,
Mercer 1900, Bauer 1904, v. Voss 1912.

Gonin's work especially is often cited. He examined in the
first place Piens brasslcae and says (1894) of the origin of the
so-called disque imaginal that it is better to use the term repli
imaginal.

I wish to point out the concurrence between him and Lijonet.
„The repli imaginal originates from evagination of the hypoder-
mis, preliminarily invaginated. The part which the tracheae and
the nerves play in this formation is a secondary one. The tracheae
are neither the cause of the duplication nor of the extension of
the walls of the wing. The rudiments of the wings are developed
from the first larval age but do not participate in the larval
moults, their surface does not produce a cuticle till towards the
end of the last stage".

Mercer who I think, has been one of the latest investi-
gators of these organs, has found in Pieris also that the rudiments
of the wing grow.

In a few words his results may be summarized as: instar /,
over against a trachea a thickening of the hypodermis is found,
in the middle of which lies a cavity, and round the trachea
some detached cells (= lymphocytes ?). Instar 7/, a chitinous plug
penetrates into the cavity ; Instar 7/7, the wingbud arrives in
the body cavity, but remains connected with the hypodermis by
the „peripodal membrane" (v. Rees). The trachea becomes larger
and grows into the two layers of the hypodermis. Instar 7F,
the tracheoli enclosed in the wing-bud reach the edge. Instar F,
the tracheoli grow further, the „ wing-rudiment" reaches the leg
and becomes folded. Pupa, the larval tracheoli disappear and are
replaced by a quite different pupal system.

Bauer (1904) adds to this that the form of the wing-folds is
only governed by mechanical forces. If the thick larval cuticle
is an obstacle, the fold becomes invaginated, but directly evagi-

23

nates and becomes an elevation as soon as this obstacle is removed.
This explains the so-called ^vorzeitige Entwicklung'' (premature
development).

Consequently in the beginning the wing-rudiment is not connected
with the tracheae and in instar 1 the rudiment of the wing is
only a minute thickening of the hypodermis. This knowledge will
be uf much use to us in discussing the so-called rudimentary
stigmata on the mesothorax and roetathorax.

The way in which the thorax is provided with stigmata is a
second point which I wish to discuss. There are many different
opinions on this subject.

In his description, which for the rest is exceedingly accurate, of
CoAsun (Traite de la chenille, 1760) Lijonet does not mention the
rudimentary thoracic stigmata. The remaining stigmata he points
out very accurately and also their connection with the „bronchi"
as he calls the tracheae. It is peculiar that he does not fully
believe in their respiratory function in consequence of the
experiments mentioned by him on p. 78.

A. C. OuDEMAKs observed (1886, p. 19) that the Myriapodae
and Hexapodae always bear the stigmata within the limits
of the segment itself, but he admits that in adults they can be
shifted either to the front part of the segment or to its back part
and in so doing may even get into the intersegmental membrane.

In direct opposition to this stands Henneguy's contention (1904)
that the stigmata generally occur intersegmentally, but that later
on this position may be altered. He ascribes this to the compli-
mentary segments which Kolbe discovered. This writer says that
in the beginning no stigmata occur on the head and the protho-
rax (1. c. p. 20). With this the models of Hydrophilus piceus L.
(manufactured in the studio of Ziegler according to K. Heider's
treatise, 1889) are in perfect harmony. Here it is clearly visible
that the first stigma appears on the mesothorax (Stage 9, model 9).

For a student, who has not yet undertaken any special investi-
gations in this direction, it is exceedingly difficult to find out
the truth. A priori one would be inclined to think that both

24

opinions can be true for different objects. In the literature I found
some papers which specially treat this question. It is a pity
that the last-mentioned writer Kummeth, apparently was not
acquainted with Boas' paper.

Boas devotes p. 390 of his article (1899) to the question of the
thoracic stigmata. Relying on his examinations of Cossus ligni-
perda and Ergates faher (a Cervicornid) he says that there are
two thoracal stigmata. The first is shifted to the prothorax, the
second forms a closed, rudimentary stigma in the intersegmental
membrane between the mesothorax and the metathorax. This rudi-
mentary stigma is situated lower than the ordinary ones. The so-called
rudimentary stigmata, lying in the stigmal line, are in reality the
origins of wings. In the imago the closed larval stigma becomes open.

Tower (1906) found in Leptitiotarsa „that the wing in develop-
ment starts from a minute invagination of cells in the region of
the wing spots, which is an area, as shown by Verson, myself
and others, homologous to the spiracular centre of other segments
(I.e. p. 163)."

G. C. Crampton (1914) does not enter into this question. He is
unwilling to accept the assumption of subsegments in the thorax,
since he is convinced that all theories about the compound-segment
are unfounded (p. 56).

Janet (1909) gives a very interesting list of the origin of the
different segments and subsegments.

From this I only quote: (see also Chapter IV, p. 27).

Ordre Ordre

ontogenique. anaturaique.

13

10

prothorax

15

11

mesothorax

14

12

metathorax

spir. prothorac. des Diptera.
spir. mesothoracique.
spir. metathoracique.

We therefore see that Boas and Janet agree with Henneguy
in the supposition that the prothorax has no stigma of its own
and that a shifting of the stigma in an oral direction has taken
place when a prothoracal stigma occurs.

25

KfMMETH again went into this question thoroughly in 1914 and
has added very accurate figures to his text (PI. V, fig. 1 — 25). He
has exaniiniHl various orders of insects, but unfortunately, mostly
as imagines. His principal results are:

The thoracic shield is formed by three thoracic segments and
one abdominal segment, which, however, retains an abdominal
structure. These four rings bear three pairs of stigmata, one ab-
dominal i>air as usual situated in the praesegmental zone (some-
times more dorsal : Pulicidae, sometimes more ventral : Khynchotae)
and two thoracic pairs, situated postsegmentally or intersegmen-
tally. The first pair always breaks through the connective mem-
brane between ,the prothorax and mesothorax, sometimes more
prothoracally (Coleoptera, Uhynchota, Hymenoptera), sometimes
more or less in the intersegmental connective membrane (Plecop-
tera, Lepidoptera). In Odonata, Neuroptera and Panorpata it is
forced by the strong reduction of the prothorax against the prae-
segmental edge of the mesothorax. It is able to move consi-
derably in the dorso-ventral direction.

The second pair of thoracic stigmata is found between the
mesothorax and metathorax, mostly on a line with the first pair.
In Hymenoptera and Lepidoptera it is situated directly under
the root of the back-wing. This is confirmed by the larvae.

Zander found in 1910 that the first pair belongs to the post-
segmental zone of the prothorax, the second pair to the inter-
segmental membrane of the mesothorax and metathorax. An
exception is formed by Di/tiscus and Ergates, where the first pair
is situated in the praesegmental part of the 2Qd segment.

There is not the least indication that formerly there were three
thoracic pairs of stigmata, and that in one group one of these
became obliterated, in another group another. The only thing which
is certain, is that the functionizing stigmata of different insects
belong to diflferent segments. The second stigma mostly lies a
little ventrad.

Without making any claim to finality in this important ques-
tion, I think it may be accepted as quite certain:

26

1st that Ergates^ as examined by Boas, cannot be considered
as showing the general rule.

2nd that the prothoracic stigma of recent insects originally belongs
to the intersegmental membrane or to the praesegmental zone of the
mesothorax and that it can be shifted unto the prothorax.

3>"d that the second thoracic stigma was originally situated in
the intersegmental membrane between the mesothorax and meta-
thorax, and that commonly it lies more ventrally than the other
stigmata.

4th that the abdominal segments (1 — 8) possess a praesegmen-
tal stigma.

Janet (1909) mentions a 9th abdominal stigma for Lepisma,
Brauer (1851) for Panorpa a stigma on all the 13 segments,
except on the mesothorax and metathorax.

Accepting the probability, that the homoiomery, which shows
itself 80 strongly in Insects, originally also ruled the tracheal
system, the conclusion logically follows, that in the beginning the
stigmata on the thorax were also situated praesegmentally. But then
we must also assume that the second pair of thoracic stigmata in
reality belongs to the metathorax and the prothoracic stigma to
the mesothorax. From this follows a shifting of the stigmata in
the direction of the head sometimes over a considerable distance.
Through this shifting the metathoracic stigma is pressed a little
towards the ventral side, whilst the first thoracic stigma may at
the same time turn to the dorsal side.

If this hypothesis is correct, then the whole thorax has been
modified and therefore cannot have preserved a primitive struc-
ture. This assertion is supported by the different arrangement of
the muscles, already described by Lijonet. Von Voss also has
pointed out the secondary structure of the thorax in his papers
1911, 1912, 1913. If therefore the primitive structure of a segment
is the object of our research, we must study the abdominal instead
of the thoracic segments. Herein lies a strong argument against
Fracker and Tsou, who take the prothorax as a starting point.

27

CHAPTER IV.
On the number of the seqments and on the abdominal legs.

According to Kowalewsky'b observations on the development
of Stnerinthm populi (1871) the abdomen of insects originates from
10 somites, all of which possess a tendency to form abdominal
legs. (p. 53, PI. XII fig. 8 and 10). He was followed by TiCHO-
MIROFF (1879) who counted 11 abdominal souuU^s \n lionibyx nwn,
likewise provided with pedes spurii, except the first. These ab-
dominal legs also occur in other orders of insects. Ratiikk showed
them in 1846 for Melolontha, Heidkr (1889) for HydrophUus.
The first abdominal legs are remarkably large (Zieqler's Model).
Wheeler observed in 1893 also 11 abdominal somites in Xiphi-
dium etuiferum.

Janet in 1909 comes to a total of 27 metameres of which 9
pass into the head and 3 into the thorax, the other 15 form the
abdomen. The three posterior ones which appear immediately
after the first three head-metameres constitute the proctenteron.
The metameres appear in triads, the first and last member of
each triad always showing themselves before the middle one.
The last triad is formed after the first, the others are regularly
developed from the oral to the caudal side. Janet also distin-
guishes 1 2 abdominal ganglions and 3 proctentrical ones in accor-
dance with the 15 metameres.

When W. Miller occupied himself in 1886 with the setal
pattern, he clearly saw that the 12th segment (= the 9th abdo-
minal) consisted of two parts which were separated by a furrow
(1. c. p. 106 — 107). The first part develops into a nearly complete
normal segment, the second part, though in fact also a special
segment, is called in his description 12a. As he adds „accord-
ing to tradition and owing to the circumstance that the
value of 12a as an independent ring can only be proved during
the first stage".

As we are obliged to agree with Henneguy's contention

28

(1904, I.e. p. 423): „The moment of the hatching does not cor-
respond to an exact point of the embryonic evolution. This mo-
ment is of a purely physical nature and depends on the smaller
or larger quantity of nutritive reserves contained within the egg^\
we may expect that in other families segment 12'^ will remain
independent till after the first instar. The more so as in 1896
Chapman was able to show so great a difference between the
eggs of Lepidoptera.

Therefore we need not be surprised if some writers are con-
vinced that there are more than ten abdominal segments, and others
that there are only ten or less.

For my part I think that it entirely depends on the specimens
examined.

PouLTON (1890) says that „behind the 7tli abdominal segment
most writers only detect a somewhat confused mass of segments,
but a careful comparison with the pupa proves that it is certainly
made up of three segments." He arrived at this conclusion through
the homologizing of the setae and through Jackson's investi-
gations (in 1890) on the pupae, which harmonized with the older
(1875) observations of that writer on the so-called cremaster of
the pupae, proving it to be the same as the anal-flap of the
caterpillars. Poulton's conclusion is I.e. p. 195: „In the pupa,
this ninth abdominal segment, although small, is as distinct as
any of the others. The part behind this segment in the larva
forms a tenth abdominal segment. This segment is separated into
a dorsal portion (X') of which the posterior and lower part form
the anal-flap and a ventral portion (X), of which the anal claspers
form the posterior and lower part, between the latter is the anus".
On p. 196 he says that for a long time he has considered that
X consisted of two segments.

Spuler (1910) also believes he can discern a 14th segment,
fig. 4, p. XXVII. The last three segments he takes together as the
anal segments. They often bear an „After-klappe". „That the last
part should be looked upon as the dorsal part of a 14th ring,
consequently of the 11th abdominal segment, is proved by the

29

warts, but we cannot make out by the aid of the warts whether
the lobes at the back of the Xlllth segment which are separated by
an incision, are to be considered as remnants of the ventral part of
a XlVth segment, because the regular succession of the legless
rings on the ventral side has been interrupted by the back feet."

Sharp (1901) however says on II p. 323: "The caterpillar is
composed of a head and thirteen divisions or segments of the
body, the first three of the latter are called thoracic, the other
ten abdominal segments, in most caterpillars the terminal two or
three abdominal segments are more or less run together, and the
ninth may be very small, so that the true number is indistinct.!'
Thus he leaves this highly important question of the number
of primary segments unsolved.

J. Th. O1UKMAX8 thinks (1897—1900, p. 63) that the number
of the abdominal segments is ten and that those who take it to
be eleven are wrong.

Fracker (1915) says of a full-grown Hepialua (p. 29): '*We
may consequently conclude that the setae give no evidence for
considering the anal segment to be composed of more than one
metamere either in its dorsal or ventral portions. Those who
have asserted that the setae show that this segment consists of
more than one somite, have not studied the data carefully on which
their opinions were based."

Against this we may say that Fracker does not speak of the
arrangement on the newly hatched larva and that as early as 1886
Miller observed that the first instar only showed this last
segment distinctly, whilst at the same time we may contend that
features, holding good for a certain form, may prove to be falla-
cious for another.

Haxdlirsch (1908) on the basis of morphological, embryological
and especially palaeontological investigations, came to the con-
clusion that insects, except the CoUembola, possess eleven abdominal
segments and a telson. In all the primitive insects this 11th segment
ends in two cerci.

Neither do writers agree on the primary number of abdominal

30

legs (pedes spurii). Some of them, in imitation of Kowalewsky
(1871) and Tichomiroff (1879) seek to provide all the abdo-
minal segments with them. This conception is supported by em-
bryology and comparative morphology. On the other hand
Deegener (1909, p. 3) considers these pedes spurii to be „ secon-
dary, adaptive, provisorial organs". With a view to this contro-
versy it is worth while to point out how, as early as 1886, Muller
called attention to the primary seta N°. 6, which occurs on the
legless segments, but which is absent on the segments 6, 7,
8 and 9 (= abdominal 3, 4, 5 and 6).

Fracker who has a different conception of these setae, does
not go into this question at all. We see therefore that there are
many different opinions on the number of the abdominal segments
and on the question whether the abdominal legs are primitive or
not. I think that I have adduced some proofs in Chapter VI of
the presence of more than ten abdominal segments, and I believe
that the regular presence of the seta pedalis also on those segments
which bear no legs, should be taken as a proof that originally
all the abdominal segments were provided with legs. This con-
ception is also supported by embryology.

CHAPTER V.

Nomenclature and primitive pattern.

With a view to making the descriptions as clear as possible
if is desirable to introduce a well defined nomenclature.
Weismann (in 1876) made a first attempt, but as I pointed out in
chapter II, he considered the setae of no importance and only
gave names to the stripes.

Schroder also confined himself to these in his study on the
Geometridae (1894). He found the same stripes as Weismann on
the Sphingidae but in larger number. For the intermediate stripes,
he introduced particular names which, however, can be brought
back to Weismann's nomenclature.

31

W. MCller (in 1886) gave a very useful terinininology which
altM) stands in connection with that of Weismann. He distinguishes
the following ^primary bristles" (see PI. I, fig. 1, 2, 3. after
W. MCller).

On an abdominal segment.

1. on the back, next to the dorsal line.

2. a little ventrad and caudad of 1.

3. above the stigma.

4. behind the stigma.

5. under the stigma.

6. where we would expect the leg on the legless segments.
On the thorax:

1 , 5, 6 as on the abdomen ; 3 and 4 blended into one, 2 is wanting.

The spines have names and are called after the stripes along
their base. The dorsal line can be single or double. lie distinguishes :

Dorsalia (D. s.) viz. D. s. ant. if they are situated in front of,
and I), s. post.^ if they are lying behind the connecting line be-
tween the right and left subdorsal.

Subdorsalia (S. d. s.) the spines which are situated half way
between the dorsalia and the stigma.

Suprastigmalia (Sst.) and Infrastigmalia (Ifst,) are determined
by the situation of the stigma.

Pedalia under the infrastigmalia.

Though Muller makes a distinction between the spines of the
Saturnidae which arise from the primary setae and those of the
Nymphalidae which do not come from those setae, he thinks
(1. c. p. 246) that the names given in the first case may be kept.

Scudder (1889) did not give names to the setae but only to the
stripes (see chapter II).

Independently of Muller, Dyar proposed a new nomenclature
in 1894, in which he purposely neglected the first instar, as it is
„a generalized condition of tubercles and setae", and it is not at
all certain „that the character of presence or absence of this
generalized first stage has any special phylogenetic significance"
(I.e. p. 196).

32

He distinguishes two types: 1°. Hepiahoi. „This type consists
(abdom. segm.) of five tubercles above the spiracle on each side,
three in a transverse row about the middle of the segment and
two behind, below the spiracle are two oblique rows, containing
respectively two and four tubercles (I.e. fig. 2 p. 197. See PI. I, fig. 4).
2^. The second type contains two dissimilar lines of modification
of the first type. The fundamental arrangement is as follows : On
each side above the spiracle three tubercles, below or behind the
spiracle and above the base of the leg three more, on the base
of the leg three (or four) on the outside and one on the
inside near the midventral line. I propose to designate thus,
counting from the dorsal line down the side : Tubercles I, II, III
above the spiracle, IV, V, VI below it, the group of three on the
outside of the leg as VII and the single one on the inside of the
leg as VIII. VII and VIII are also present on the legless abdo-
minal segments in the corresponding position" (I.e. p. 196 — 197,
fig. 5, p. 198. See PI. I, fig. 5).

In the Psychidae the three tubercles are retained on the middle
annulet, while both are lost on the posterior one (1. c. p. 198,
^g. 3). See PI. I, fig. G.

Other deviations also occur so that he separates the Psychidae
from all the rest of the Frenatae.

The thoracic segments differ a great deal, the 7a+ b and 7/a
-f- b occurring there, are not homologous with the abdominal / and
// but they are simply called thus, because there often occur
two tubercles, one above the other, each bearing two setae.

In 1901 Dyar came to different conclusions, especially through
0. Hofmann's criticism. He accepted Hofmann's opinion about
the homology of the thoracic setae.

O. HoFMANN (1898) found that in the Pterophoridae the protho-
rax deviates strongly from the rest. During instar / the meso-
thorax and the metathorax bear six setae and so does the abdomen.
They are homologous but not in the way Dyar thought. A better
homology runs thus:

33

Dyar la, lb, Ila, lib, III, IV, V, and VI.

O. HoFMANN I, II, III, IV, V, and VI.

HoFMANX considers the setae called by Dyar III and V to be
secondary or subpriinary ones.

The setae I — IV are usually arranged more or lesa in a straight
line. On- PI. I, fig. 8 I have indicated the place of the secon-
dary setae by an •, in O. Hofmann's figure of TaeniocaniiHt
ijothiea L. (I.e. fig. 2, p. 129).

From a comparison of fig. 7 and 8 we see what a confusion
of numbers has been producetl here.

In Beutknmuller's monograph on the SesiUiae (1900), Dyar
described the caterpillars and here still used his old system. In
the many descriptions, given by Dyar, attention must always be
paid to the year of publication. In 1901 he proposed to call III
and V on the thorax, which Hofmann considered to be secondary
setae, \a and \b (these notations I have also used in fig. 8),
whereby at the same time VI of the abdomen became \h and
VII became VI.

Quail (1900) usually speaks of Dyar's setae I and II as
trapezoidal tubercles, in the same way as Hofmann had done
before. For the rest he uses names for the setae: supraspiracular,
subspiracular, basal setae. For his description of the Hepialidae
see chapter VI. In 1904 he laid stress on the study of the first
larval stage. In that year Quail described the first instar
of Cossus cossus and compared it with Zeuzerapyrina. Mr. A. Bacot
pointed out to him „a minute free spiracular point of very general
occurrence on the abdominal segments of lepidopterous larvae"
(1. c. p. 95). Quail believes this point to be III B and he
sees here already „that the elimination of spiracles probably is
the chief cause of the altered positions of the tubercles on thoracic
segments". In a second article of the same year (19046) Quail
comes to the conclusion, that II B (of Dyar) on the thorax is not
similar to IV on the abdomen, as Dyar and Hofmann take it to
be, but that „the homologue of II B of the thorax is a minute
anterior supraspiracular tubercle of the abdomen called by me

34

III B, that Dyar's III of the thorax is the homologue of a sub-
spiracular tubercle of the abdomen and so on".

His terminology is best understood from the fig. 9 and 10, which
are drawn after Quail's fig. 1, 2, 3, 4, PI. IX 19046.

We see that the confusion between the different writers becomes
worse.

Forbes (1910) wrote a study which unfortunately I was
not able to read. According to Fracker(1915 p. 14), „he did not
cover the subject of the homotypy of the setae. The few figures
he labels, include errors for which he was not responsible, as he'
had not given the subject consideration" and (p. 35) „most of these
associations would be very difficult to explain and they are wholly
unnecessary. The mistakes (confusion of primary and subprimary
setae) are due, not to errors in observation but to a failure to
take the primitive first stage into account." In the table on p. 40
Fracker says that the setae are named by Forbes in about the
same way as I have done in fig. 11 and 12 of my PI. I in
accordance with his indications. It should be observed that the
labelling of the Jugatae slightly differs in the numbers 4 — 6 viz:

Frenatae absent IV V VI

Jugatae IV V VI absent

according to Fracker & a? j^ /x

By his studies of the pattern of the pupa and imago, J. F. van
Bemmelen (1889, 1912, 1913, 1914, 1915, 1916) was led to an
examination of the pigment spots of the caterpillars. In 1912
I.e. p. 115 he gives a synopsis of the spots on Pier is brassicae,
larva pupa, and imago and of the pupa of Aporia crataegi. Following
Weismann and W. Muller he calls the spots after the rows in
which they lie. He distinguishes : the dorsal, dorsolateral, epistigmal,
stigmal, hypostigmal, ventrolateral and ventral rows. The first
and the last are median, the others are paired. The number of spots
in each row is either one, two or three, a group may replace
one single spot. By the blending of the spots occurring on con-
secutive segments, stripes are brought about. See PI. I, fig. 13.

35

As , stigma" originally is a greek word, much is to bo said in
favour of using the prepositions epi- and hypo-, instead of supra-,
sub- or infra-. In connection however with the other terms
and with the existing names used by Weismann and MCller,
I think that the words suprastigmal and infrastigmal might
be retained. For the connection found by van Bemmelen be-
tween the pattern of the pupa and that of the larva see chapter
I and VII.

Tsou (1914) has a very peculiar way of indicating the setae.

His groups are: A = anterior, D = dorsal, S = subdorsal,
C = circumstigmatal, L = lateroventral, P = pseudopodal, M =
niidventral. Each individual seta of a group is numbered as D, ,
Pf etc. The setae belonging to the above groups are regarded as
primary setae.

In chapter II, I have already expressed my objections to his
method. For the sake of completeness I have copied on PI. I,
fig. 14 Tsou's figures of Hepialm humuli, the metathorax and
the first abdominal segment (1. c. PI. X, fig. 1 c. d.).

Fracker (1915) has examined the setae of the caterpillars on a
large scale. As appears from the synopsis given, the confusion in
the numbering of the setae had become very serious. Fracker
has therefore rightly felt that he could not once more propose
a new indication with the use of numbers. He began to pay
attention to a certain segment and tried to find out in how far
the same setae ocurred on the same segment of the members of
other families of the suborder. This he calls homology. In the
second place he tried to compare the different segments of one
caterpillar with each other and this he calls homotypy.

The use of this term might give rise to confusion.

In the widely spread "Lehrbuch der Zoologie von Claus-
Grobben" 2nd edition (1910, on p. 12) is given:
Haniologous : = moTT^hologicaWj equivalent.

Homodynamous = homologous organs, which repeat themselves
in the longitudinal axis of the animal (e. g. verte-
brae, pairs of legs of the Arthropoda etc.).

36

Homofypical = homologous organs which form the reflected images
of each other, hence antimers ; e. g. the right and
left hands and the rays of a star-fish.

Fracker therefore considers as homotypical what Claus-Grobben
calls homodynamous. He arrives at the following definition
1. c, p. 15: „Two organs on different segments of the same animal
are homotypic, regardless of their positions at the present time,
when they have developed from homotypic organs of a generalized
ancestor. In a generalized type two similar organs on different
segments are homotypes, when they bear the same relations to
the other organs of their respective segments".

On the whole I agree with this definition, but I wish to point
out the hypothetic element which is hidden in it. It will often
be difficult to tell how a certain seta is placed in a generalized
ancestral type, so that in most cases it will be better to trust
the second part of the definition rather than the first. Here we
meet with a great number of difficulties, which Fracker places
under three headings:

1. Absence of intermediate stages between radically different
conditions.

2. The lack of developmental series.

3. Apparently a lepidopterous larva has three or more entirely
distinct types of arrangement of the setae (prothoracic, thoracic,
abdominal, anal).

Fracker obviates these difficulties in the following manner,
1. c. p. 17: "The setae of the prothorax, metathorax and abdomen
of the generalized members of both sub-orders of Lepidoptera were
plotted, one segment over the other, as if all were on the same '
segment. The number was about fifteen (fig. 1) and they were
in approximately the same position as on the prothorax of the
most generalized forms of the order." (in casu Hepialus mustelinus).
These primary setae Fracker indicates by the characters of the
Greek alphabet, p. 28, because:

1. A special letter can be introduced for a subprimary seta in
a limited group without disarranging the system.

37

2. The alphabetical order is not so fixed in the mind as to
prejudice one in regard to homology.

He distinguishes:
Pi-imary setae on the newly hatched larvae.
Subprimnry,, appearing after one moult, but fairly constant (/t«, &).
Secondary^ no constant position but scattered, very rare in the
first instar.

PI. I fig. 15 shows Frackeu's indication best.

The above-mentioned writer thinks it justifiable to conclude
from the setae on the prothorax of the Tortncidae (1. c. fig. 39),
Aegeriidae and Yponomeutidae (1. c. fig. 35), that e remains in its
place and ; moves forward towards it, whilst in the Macrolepi-
doptera the opposite movement is to be noted. In the latter e has
migrated back to ; on the fullgrown larva, whilst they are far
away from each other on the newly hatched larva (1. c. p. 34).

He therefore thinks that, for instance, in instar / the first seta
over the stigma of Feltia ylandaria^ is not y or f but p (see PI,
I, fig. 17).

On the mesothorax and the metathorax also he arrives at con-
clusions, which differ from those of former writers.

These differences have been expressed in his figures, which
are kept very diagrammatic. Seta p as well as seta /3 is always
turned caudally, whilst x and e point in an oral direction. These
figures already suggest the hypothesis proposed by Fracker.
For the formation of an unprejudiced opinion it is therefore pre-
ferable to pay exclusive attention to the points of implantation
of the setae on the skin.

I also wish to draw attention to PI. I, fig. 16 which agrees
with Fracker's PI. V fig. 36, the mesothorax of Atteva aurea
( Yponomeiitidae).

Fracker sometimes unites the setae into several groups:

B = x + (3. K = ^-\-x-\->;. P = s-\- p.

n ^ y + r (on thorax), y + tt + r (on abdomen).

T = T + ^ -f «.

Before proceeding to the nomenclatures I should like to

38

propose and which I have used in the following descriptions, I
have to subject to criticism the systems which have been so far used.

Concerning the indication of the setae by means of cyphers, as
used by W. Muller (1886), Dyar (1894 and 1901), Quail
(1904 and 1904 b), Forbes (1911), these writers have made so many
changes, that it would be a hopeless task to try to improve it.

Fracker has grouped together the opinions of the different
investigators into a table (1. c. p. 40). For a single slight mo-
dification I wish to quote a striking case from it. Quail indicates
the large seta over the stigma with III or III A and the small
one in front of it with III B, and not as Fracker does : with III
and III A respectively. If we look at the seta over the stigma on
the abdomen, we see that it has been named in the following ways :

Mesothorax and Metathorax (Frenatae).

Muller 1886, Dyar 1895, Hofmann 1898, Dyar 1901, Quail 1904, Forbes 1910.
4. lib. IV. IV. III. * II B.

Abdomen.

Muller 1886, Dyar 1895, Hofmann 1898, Dyar 1901, Quail 1904, Forbes1910.
3. III. III. III. Ill A. III.

So, whilst on the abdomen at least all the writers have given
the same cypher to this seta, Fracker considers it to be p, to
which in Hepialus the others certainly would not have given
this index, as p belongs to the caudal row, and the abdominal
seta III has been placed by the majority of writers in the oral
row. Anyone can find other examples from the figures placed
side by side on PI. I.

Tsou's system I have already discussed. Of the systems in
which no names are used, that of Fracker is the most important,
and against this system I have serious objections.

In the first place Fracker's method of concentrating all the
setae occurring on the several segments of different larvae into
one segment and of declaring this hypothetical segment to be the
original form, is wrong.

For it is probable that a certain organ of a given original

39

form gets specialized in very different directions and in so doing
gives rise to numerous new forms, one group acquiring this, an-
other group that improvement.

In a case like this, that organ is the most primitive which does
not show any new modifications, hence has preserved the charac-
teristics which they have all in common in the pure forms, but
certainly not such an organ, as we might artificially compose by
summarizing all the new formations.

It is my opinion that this way of looking at the question can
be justified equally well, and that it is certainly oftener used in
comparative anatomy than the method of Fracker, of which he
asserts that it is the ordinary one in problems of this kind (1. c. p. 17).

Fracker tries to give a solid basis to this summarizing hypo-
thesis by the description of the prothorax of Hepialus lupulinus.
In the first place the drawing on which this description has been
founded is not Fracker's but Dtar's work. And though I have
a great respect for the exactness with which this writer generally
works, it still remains exceedingly dangerous to take another
man's drawing like this, as the chief basis of a hypothesis which
upsets all former ideas. This, however, is not the greatest ob-
jection which I have to Fracker's opinions. The point I am
going to treat now is of a more general nature.

I think that it is not quite scientific to raise one segment,
picked out at random in an arbitrarily chosen family, to the rank
of the most generalized type. Such a procedure could only be
justified by adducing a number of facts to prove that all or at least
nearly all the members of the family possess the same foundation.
This is not at all the case here and Quail's descriptions of the Hepia-
lidae (1900) might have taught Fracker (1915) as much. Though
the family of the Hepialidae is justly considered to be a primitive
one, this does not include the necessity that all the features of
all the members have to show a primitive character and that they
cannot possibly have undergone any secondary modifications. A
study of the existing literature would have taught Fracker that
in 1914 and 1915 (and in 1916), J. F. van Bemmelen found

40

very primitive qualities in the pattern of the wings of the Hepia-
lidae, but nevertheless could show at the same time great secon-
dary alterations in it.

It does not seem advisable to me to look upon the pattern of
one segment of one single representative, as the generalized type
of the Jugatae, from which the generalized type of the Frenatae
has descended. And this applies in a high degree to the prothorax.

In a preceding chapter I think, I have proved sufficiently that
the thorax on the whole is not built primitively, that the stigmata
on its surface have been shifted and that the wing-rudiments
very soon bring about changes in these segments.

By nearly all the writers the prothoracic stigma is considered
to have been shifted towards the oral side and it is strange to
choose this very segment as a starting-point.

It seems to me a very iinfortunate accident that the setae on
the prothorax of Hepialus lupulinus (1. c. p. 1 7) are also about
fifteen and that they were in approximately (!) the same position
as on the hypothetical segment.

Another objection I have to Fracker's assertion is, that x and b
on the prothorax should be homologous with x and b of the
abdomen (c.f. on this point my PI. I, fig. 15).

On the prothorax we find these setae in front of the stigma,
on the abdomen behind it, without there being any change in the
position in regard to the other setae. There we get the impres-
sion as if the stigma had passed under these setae, a kind of
dislocation, the possibility of which I cannot understand. I am
convinced that a seta which is situated in front of the stigma must
remain prostigmal, and that it will either disappear in case the
stigma is shifted, or that it will display the traces of the shifting
of the stigma in its situation on the segment. Therefore I think
that the seta called III B by Quail may agree with a prostigmal
seta, even if this seta is sometimes placed a little higher.

The "proofs" mentioned by Fracker of the shifting of £ and
p do not appear to me to be convincing. The upshot of these
arguments is always the preconceived idea, that we must consider

41

the thoracic segments in general and the prothoracic one in par-
ticular, to bear a primary character. I think I have shown suffi-
ciently in chapter III that this a-priori view does not deserve
general acceptance.

In criticizing the different systems and especially that of Fracker,
I have had the opportunity of explaining my opinion on the
nomenclature.

Therefore it will not cause surprise that I begin by first
taking the abdominal pattern and that, as neither the use
of jiumbers nor of Fracker's Greek letters appeared possible, I
have come back to the use of names which at the same time
express the place of the seta on the segment. In composing my
nomenclature I have tried to keep in agreement as far as possible
with the following writers :

Weismanv (1876), MCller (1886), Scudder (1889), Schroder
(1894), Quail (1900), J. F. van IJemmelen (1912).

In accordance with Dyar and Fracker I think that in the
composition of a primitive pattern a verruca (wart), a scolus (spine),
a tiibercuhim (elevation which mostly bears one or more setae) and
a seta ought to be considered homologous with each other.

I should like to add to this series of homologous organs a
pigmental spot. I think that the seta is the original part, and
that the other organs are its secondary modifications, A homo-
geneous spreading of the setae, as also their complete absence,
must be considered as secondary features.

The reduced patterns of the anal segments have also arisen in
a secondary way, though they often lead to pseudo-primitive
conditions.

In discussing the results I intend to refer again to these facts
(chapter IX).

To be able to shorten the descriptions I have also given cyphers
to the patterns as a whole : Type I, la etc. The deviations from
this pattern can be easily indicated so that with a few words the
place of all the setae can be exactly mentioned. In connection with
my remarks in Chapter lY it should be pointed out, that the anal

42

segments deserve special mentioning. For them I cannot give any defi-
nite rules, as in some cases they are far more reduced than in others.

Type I. This pattern consists of the following setae, tubercula
etc. (PI. I, fig. 19).

Seta dorsalis on the oral and at the same time on the dorsal
edge of the segment.

S. subdorsalis superior more caudal and also a little more ventral
than the former.

S. suprastig mails lying over the stigma, abont in a line with
the stigma and s. dorsalis.

S. prostigmalis is usually very short and stands right in front
of the stigma or has been shifted a little upwards.

S. poststigmalis caudal and in most cases somewhat ventral of
the stigma.

S. infrastigmalis under the stigma.

S. basalts anterior and

S. basalts posterior situated between s. infrastigmalis and the
place where the leg is implanted, or where this is wanting,
between s. infrastigmalis and s. pedalis.

S. pedalis at the beginning of the leg and, if the leg is wanting,
on the place where we might expect it.

In connection with the remarks in chapter IV the presence of
this seta seems to me a proof of the secondary disappearance of
the legs on the abdominal segments 1, 2, 7,8,9.

S. propedalis on the ventral side, in front of the beginning of
the leg.

S. ventralis between the inner side of the leg and the ventral
median line.

We see that this Type I almost completely agrees with the
pattern chosen as the fundamental plan by W. Muller (1886),
Dyar (1894), 0. HoFMANN (1898) and Quail (1904). Where,
in my investigation independently of them, I was led to the same
type, I think I am justified in attaching great value to this result.

As rather frequently occurring extensions of this type, there
also occur:

43

S. subdorsalis inferior lying ventral of 8. subdorsalis superior.

8. dorsolateralis implanted on the oral edge of the segment
between s. dorsalis and s. suprastigmalio.

In the descriptions I think it better to mention, whether these
setae are present or not, whilst I consider it advisable to make
a s()eoial notice of s. prostigmalis and s. propedalis as they often
show important deviations in size.

Typ© Irt. A simplification of frequent occurrence is that the
seta subdorsalis disappears, whilst seta poststigmalis is united with
seta infr(isti(jtnali<A.

Consequently we get one series in which the stigma is also
situated. This pattern, which at first sight makes a primitive im-
pression, is found i. a. in Saturnia pavonia. The setae have already
chang^ in instar / into rerrucae^ later on in scoli^ but one seta
dorsalis remains on the prothorax as a proof that here as every-
where else the verrucae have arisen by modification of the setae
(c. f. PI. I, fig. 20).

Type \h. At the end of larval life the conditions of the pattern
are nearly the same in the Lymantridae = Liparidae, Here
however verruca subdorsalis does not disappear, but verruca dor-
salis does, whilst r. poststigmalis is blended with v. suprastigmalis
and not with p. infrastigmalis. This pattern I call type lb.

Apparently the types la and 16 are almost alike, in reality
they have arisen independently of each other.

The coalescence of v. poststigmalis and v. infrastigmalis re-
mains during all the larval instars and even in the pupa it is
represented by a furrow which separates the two halves from
each other (PI. I, fig. 21).

Type II. At first sight the meso- and the metathorax seem to
possess a setal pattern which is entirely different from that of
the abdominal segments (PI. I, fig. 18).

They agree, however, so much with the prothorax, that it does
not seem advisable to establish a special type for it but to
describe it as : type I without seta .... or : type II with
seta , etc.

44

The opinions of the investigators differ very much on the
meso- and the metathorax, as may be seen from Fracker's list
(1915 1. c. p. 40), from the figures arranged by me on plate I
and from the example on page 38.

I think that the cause of the difference between type II and
type I must be sought in the shifting of the stigmata and in
their final disappearance later on.

On the meso- and the metathorax one generally finds three
setae at the oral border of the segment, mostly arranged in one
vertical line.

About in the same row, but sometimes a little in front of it
or behind it, is another seta, and behind this one we often find
a peculiarity of the surface of the skin. Sometimes in the shape
of a pigmental spot (e.g. Porthesia chrysorrhoea^ Zeuzera pyrina)^
another time in that of a verruca (e. g. Arctia caja, Sericmus
telamon), or as a distinct dilatation of the system of air-tubes,
which is seen through the transparent skin (e. g. Pieris brassicae
and P. napt). These variations of the surface of the skin I take
to be caused by the wing-rudiment (Boas 1899), but, as it is situated
on the place originally occupied by the stigma, the seta which is
placed in front of it keeps its character as .s. prostigmalis. On
this point I agree with Quail, who calls it III B. Under this
seta stands another one, and over the base of the leg, one more.
All these six setae are situated in one vertical row. This type seems
therefore to be very primitive. But having seen that type la and \h
are pseudo-primitive and after having explained in chapter IV that
the thorax is generally of a secondary construction, I think an-
other point of view may also be taken.

The uppermost seta does not give any difficulties, and may
safely be considered to be s. dorsalis, and the fourth s. prostig-
malis. In this case it is clear, that the two last setae are
s. infrastigmalis and s. basalis and that the third of the row is
s. suprastigmalis. There remains still one seta, namely the second
of the row.

It can be easily understood that this seta is considered by many

45

iiivestigatorH to be s. subdot-ndlis, muovod to the front-border
of the segment. But against this explanation I have an objection
for the following reasons :

1". In different families there appear on the abdomen tliree
setae above tlie stigma : e. g. Phaleia hucephala, Hepialus, Pterin
tuipi and the Pstjchithte.

2". On the prothorax we very regularly find s. dorsolateralis
together with s. sitbdorsalis.

3". On all the pupae of Rhopalocera, which I have examined
and which possess a pattern, I could show three elevations or
pigment-spots on the oral side of the abdominal segments, above
the stigma. At the same time there exists a spot or an eleva-
tion, agreeing with .««. !<ufnlorsalis.

In all three cases it is evident, that the middle one of the
three setae is not .»». suhdorsalis. This seta I have called s. dor-
solateralis.

At first I hesitated whether we could possibly explain this
seta in another way, namely by calling the upper one s. dorsalis,
the next one s. suprastigmalis and by considering the third one
as agreeing with III B of Quail. I thought at first that III B
had changed its place and had come above the stigma and by
this transgression had caused a dorsad motion of s. suprastigmalis.
I think, however, that the seta before and not that above the
transparent wing-rudiment, viz. s. prostigmalis, agrees with III B.

A third possible explanation of the thoracic arrangement fol-
lows here:

Starting from the idea that the prothoracic stigma really be-
longs to the praesegmental zone of the mesothorax, and that the
metathoracic stigma is removed to the intersegmental membrane,
it is easy to imagine, that the setae lying in the neighbourhood
have also taken part in this shifting. Further taking it for granted
that the setae on the thorax are originally arranged in the same
way as on the abdomen, we also find there s. suprastigmalis, s.
prostigmalis (III B), s. poststigmalis and s. infrastigmalis. If a
stigma is shifted from the oral side of a segment to the caudal

46

edge of the former, then s. poststigmcdis becomes prostigmal in
respect to this stigma.

In that case there must exist two s. jiTostig males. If s. infra-
stigmalis also takes part in this shifting, there are three s. pro-
stigmales. This then would agree with Fracker's bisetose and
trisetose K.-group.

On the mesothorax and the metathorax s. poststigmalis can be
removed so much, that it comes in one vertical line with s. su-
prastigmalis.

It is easy to imagine, that hereby s. prostigmalis is pushed
upwards and thus occupies the place of s. suprastigmalis. This
one is also shifted and appears between its old place and s. dorsalis,
i. e. on the place of s. dor solater alls.

It is also possible that s. infrastigmalis is removed downwards
and so appears on the place of s. basalts. I think that this expla-
nation is very artificial, but still it seems to me better than that
given by Fracker. That a stigma can be shifted in such a manner
that it plunges under the setae, without exerting any influence
on their position, seems to me highly improbable. I think however
that the cases mentioned sub 1 , 2 and 3 form an argument against
this explanation, which is given here for the sake of com-
pleteness only.

Besides it seems to me entirely incorrect to introduce in a
nomenclature hypothetic views on homology. In human anatomy
we might as well call the muscles and bones of the arm and
legs by the same names, if they were considered homologous for
some more or less probable reason. In my opinion a nomenclature
ought to be a means for a short and clear description and it
should not be the expression of hypotheses on which various
investigators have different opinions. Through this, it is made
difficult for outsiders to understand the terminology, and the in-
vestigators of the setal pattern themselves will, each with the
same right, claim their own hypothesis to be the right one.

47

CHAPTER VI.
Systematic synopsis of the setal pattern of caterpillars.

Not much harmony is to be found in the classification of the
Lepidoptcni. Some of the latter-day writers reject the sub-order
of the Jugatae, but even if this sub-order is accepted, the sub-
order of the Frenatae remains an extraordinarily difficult problem
for the systematists.

The principal newer systems are:

Chapman (1893), especially founded on the pupae.

CoMSToCK (1893), especially founded on the wing-venation.

Dyar (1894 and 1894 6), especially based on ihe setal pattern
of caterpillars.

Packard (1895), starts from the three preceding systems and
the geogpraphical distribution.

Spuler (1895 and 1910), especially founded on the wing-venation.

Handlirsch (1908), based on palaeontological data and on the
preceding systems.

I have followed the last-mentioned writer, because he has worked
into his system all the previous ones and because I think that
the palaeontological data have been neglected too much by the
entomologists. Handlirsch's great knowledge of the morphology,
embryology and palaeontology of the insects, makes him an
authority deserving confidence. Therefore I have not followed
ScHARp's manual, as I did in my preliminary note (1916).

As far as possible I have mentioned the literature with each
family. It is more than possible that I have not noticed some of
the very scattered articles in the entomological periodicals. I there-
fore wish to draw attention to the fact that owing to the war, I
have not been able to procure some of the foreign periodicals.
In the first sub-order I have explained the data known to me
rather in detail, because in this way I hoped to assist in the
solution of one of the most difficult problems in the systematic
arrangement of Lepidoptera.

My work, however, would become too comprehensive if I had
tried to do the same for all the families.

48

Sub-order I. Jiigatae.
Family I. Eriocephalidae.

In many respects this family is very primitive but with respect
to the setae it is exceedingly specialized. I do not possess any
material for investigations and am therefore obliged to confine
myself to literature, especially to one article of Chapman's (1894).
It is a great pity that no more is known about these highly
interesting larvae, which possess the full set of functionising abdo-
minal legs (1 — 8) and give rise to imagines, which, as Walter
found in 1885 for Micropteryx {Eriocephala = Eriocrania) calthella,
retain the mandibles as biting organs and the first pair of maxil-
laries fully developed.

Dyar (1893 and 18946) does not give any characteristics.

Chapman (1894) gives a very interesting description, with figures
of the larva of Micropteryx calthella R. which has been checked
and copied by Packard (1895) and which also occurs (with
figure) in the manual of Oudemans (1900).

On the newly hatched larva there are ten rows of globular
rough warts, each of them on a little stem. They are placed on
four elevated ridges, and become smaller and smooth when the
larva grows.

Handlirsch (1908, p. 1254) says that there are several rows of
tubercles on the back. Like the other writers, he draws attention
to the abdominal legs at the segments 1 — 8.

Fracker (1915) does not discuss this family.

Family II. Micropterygidae.

In the pupae of this family Chapman (1893) discovered the
gigantic mandibulae. Their larvae, on the other hand, are more
secondarily modified, in so far as they lack the abdominal legs.

For want of material I could not study their setal pattern.

Dyar (1894 J, p. 49) describes Micropteryx purpurella: "The
arrangement of the setae corresponds with that of Hepicdus^ except
that the four setae on the base of the leg are absent .... It has
a double dorsal shield on every segment which may account for

49

tlu' uiiUHually posterior position of tubercle I (fig. 7 1. c. p. 49).
The uniform position on the thoracic segments of the Frenatae
is lacking." In my opinion the arrangement of the setae might
equally well be conceiveil as follows : .<». dorsatis absent, s. supmstig-
malis or s. dorsolateral is on the oral side, s. subdorsalis snj>.
and inf. on the caudal side, s. j)08tsti(/malis and s. infrastigmalis^
two 8. basales.

Packard (1895, p. 62) does not mention the larvae of this
family but draws a few setae on the nympha.

Fkacker (1915, p. 24) only says that the setae have been
reduced so far by leaf-mining habits that conclusions cannot be
based upon them.

Family III. Hepialidae.

For this family which interested me especially in connection
with the investigations of J. F. van Bemmelen on the pattern
of the wings (1912 — 1916), I had material of two species at my
disposal.

Dyar (1894, p. 197) described Hepialus lupulinus, full-grown
specimen and later on (1895, p. 66 sqq) instar / of Hepialus
musteliniis.

Dyar especially emphasized the differences between the Juga-
tae and the Frenatae, and did not describe the prothorax.

Packard (1895) described Hepiahis tnustelinus instar I and full-
grown larvae of H. humuli and H. hectui. These latter he
figured together with the pupa of Oncopora intricata on p. 72
and 73 of the first part of his work on the Bombyces. I wish
to draw special attention to these pictures, because Fracker
does not mention them. The prothorax of all three agrees much
more with my description than with that of Fracker, which has
been selected as the fundamental scheme of the pattern of all
caterpillars. H. hectus agrees best with it, but just this species
is considered by Packard to be a specialized one.

Quail (1900, PI. VI, fig. 11, 12, 13, 14) gives a description
and a figure of the newly hatched larva of Porina cervinata

4

50

Walk, and of full-grown larvae of P. umhracalnta Gn. and Cha-
ragia virescetis Dbld. He also describes the pupa of Porina cer-
vinata Walk, and figures, moreover, the last segments with the
setae, which are in the same position as on the caterpillar.

These descriptions too differ in important respects from Fracker.

Tsou (1914) minutely describes: Hepialus humuli L. and gives
the setal maps of many segments. This is the same kind which
Fracker examined, but Tsou draws a few more setae. Among
other things he puts C.^ on the prothorax on the shield, whilst
Fracker draws it as b on the outside. On abdominal segment 1
he places S, between /3 and p and in front of Sj (= e) a
special seta A.

Fracker (1915) who considered the prothorax of Hepialus
mustelinm as the primeval type of the pattern should, by a tho-
rough study of the literature, have compared different descriptions
and illustrations. I am convinced that the list of the drawings
and descriptions of caterpillars of the Hepialids already cited
by me, is far from complete. When a student attaches so much
importance to a certain family, as Fracker does, I think it only
right, that he should make as far as possible a complete perusal
of the existing literature and that he should not confine himself
to one special type, which accidently proves to be suitable for a
certain hypothesis. And the more so where the other Jugatae : the
Eriocephalidae and Microj)terygidae show so many deviations from
this type. Fracker himself described full-grown larvae of H. humuli^
H. hectiis and H. lupulinm. I have already discussed his investi-
gations in the chapter on the nomenclature.

Hepialus hedus Linn.

Instar I. Length 1 mm. Duration? Material in alcohol, col-
lected at Groningen 1914, from eggs, bought in Germany. Plate
I, fig. 22, 23, 24.

The head of this caterpillar, the smallest, which I have examined,
is relatively large, to wit more than '/4 part of the length of the
body. The upper-jaws are strongly developed. There are many

51

setae on the head which I drew with accuracy, but which I did
not study further. The ocelli are arranged in two vertical row8,
each containing the number of three, between which stands a seta.
No distinct prothoracic shield. On each of the tubcrculu there is
one si'ta which, on being magnified 400 times, still shows no
plumes and which has a length of 50 — 150 ^.
The tubercula are not coloured.

Prothorar. Seta dot'salis, s. subfiorsalis sup. and inf., 8. dor-
solateralis, s. prostigmalis, s. infntstiynialis, s. banal is, s. propedalis,
s. Fenfraliif. On the leg several s. jmialrn and behind the leg on the
ventral side a 8, po8tpedali8, which only occurs on the thorax.
Menothorax. S. dot'salis and s. pvostigmalis on the oral edge of
the segment. Behind these, approximately in a line, ». suhdorsalis
sup. and inf., further .f. infrastigmalis, s. propedalis, s. basalis,
8. poatpedalis and .s. centralis.
Metatkorax = mesothorax.

Abdomen 1. S. dorsalis, s. subdorsalis sup. and inf., s. supra-
stigmalis, 8. j>ost8tigMaliSj s. infrastigmalis, s. basalis, s. pedalis
or s. centralis.
Segm. 2 = 1.

„ 3, 4, 5, 6 are as 1 and 2, but 8. poststigmalis is situ-
ated a little lower down than usually; there are two s. ba-
sales, one s. propedalis and one s. ventralis.
,, 7, 8= 1, but s. basalis lies a little higher.
, 9. S. dorsalis, s. subdorsalis sup., s. suprastigmalis, two s.

basales, s. propedalis, s. ventralis.
„ 10. S. dorsalis. s. subdorsalis sup., s. suprastigmalis, s. pro-
pedalis or s. basalis, s. ventralis.
„ 11. 5. dorsalis and further on the caudal side of the pro-
legs two setae, in my opinion : s. subdorsalis and s. supra-
stigmalis', one s. propedalis and on the ventral side of the
claspers one seta, in my opinion to be put on one line
with s. ventralis.
I clearly see here an eleventh abdominal segment, whereas
Fracker in H. mustelinus only distinguishes ten.

52

Hepialus sipec. cf. lupulinus L. Plate II, fig. 4,

Material. Through the kindness of Mr. Claassen, Instructor of
the Government Horticultural School at Boskoop, Prof. Dr. J.
F. VAN Bemmelen procured material from the nurseries there.
Some years they are so abundant, that they damage the roots of
the lilacs, but in 1915 they were so scarce, that only a few
specimens could be obtained. The smallest measured was 2 cm.,
the biggest 3 cm. (preserved in alcohol). Judging by the size of
the head they seem to belong to the same instar. The head is flat,
large, covered with setae. The ocelli are arranged in two vertical
rows of three each, between which there is one seta. The upper-jaws
are large. The setae of the body are placed on tubercula which are
not coloured. The setae are 1000 — 2500 /*. long and not plumed.

Prothorax. This segment is almost completely hidden by the
head. No prothoracic shield. Over the stigma there is only one
seta, probably s. mhdorsalis. Further there occur s. infrastigmalis,
s. poststigmalis, s. basalis and on the leg many s. pedales.

Mesothorax. S. dorsalis, s. dorsolateralis, s. suhdorsalis superior.,
s. poststigmalis, s. infrastigtnalis, s. basalis, many s. pedales.

Metathorax. S. dorsalis and s. dorsolateralis between which there
are still three other setae, which in all specimens occur on the
metathorax, but not on the mesothorax; s. suprastigmalis,
s. suhdorsalis sup., s. poststigmalis, s. propedalis, s. basalis and
many smaller s. pedales.

Abdomen 1. S. dorsalis, s. dorsolateralis, s. suprastigmalis, of
which the two last ones are close together and s. dorsolateralis
is a little more oral, s. suhdorsalis sup. and inf., s. poststigmalis,
two s. infrastigmales of which the foremost might be compared with
s. prostig mails ; s. basalis, s. propedalis, s. pedalis or s. ventralis.
Segm. 2 = 1, but lacks s. suhdorsalis inf.

^ 3 = 2, but has mostly five s. 'basales.

„ 4, 5, 6 ^ 3 but only three s. basales.

« 7 = 2.

„ 8, stigma extraordinarily large, three s. basales, which are
placed higher than usual.

58

Setjm, 9 .S. dorsalis^ 8. dorsohteralis, tt. sHprastif/maliK in one lino,

s. subdorsalig sup., 8. jMststigmalis, two or three «. infra-

stigmales, s. basalts, one ». pi'ojyedalis.
„ 10. S. (iot'satis and s. dorsolateralix are wanting, «. supra-

stig malts, s. sttbdorsalis sup., s. jwststifftnalis, two .s. infra-

stigmales, s. propedalis and five or seven «. pedales with

very long setae.
^ 11. As there are no s. f>edalej< on the abdominal legs lam

inclined to ascribe the setae pedales of segment 10 to a

reduced segment 1 1 .

Stfttopsis of the sub-order of the Jugatae.

It is not possible to build up a pattern which they have in
common, as each family certainly possesses its own pattern and it
may be so with each genus. This may be expected as the Jugatae
most probably are not the immediate ancestors of the Frenatae.
Only in some respects they have preserved the primitive character
of the Ij<'pidopterou8 tribe in a better way. Where Uaxdlikpch
(1908) places the separation of the Frenatae from the Jugatae as
early as the Lias, because in Dogger and Malm the Palaeotineidae
already belong to the Frenatae (I.e. p. 1253 sqq.), this result
need not surprise us.

When the setae alone are concerned I would rather conclude that
the Eriocephalidae descend from the Micropterygidae than the
reverse, but I admit that a system, only based on the form and the
number of the setae would be exceedingly artificial. Moreover there
is a possibility, that in this respect the Micropterygidae have
remained in a more primitive state, especially by their mining
habits, instead of the Eriocephalidae, which are exposed to all kinds
of exterior influences, by which the form of the setae might be highly
modified. The Eriocephalidae are the most specialized forms in a
progressive line, the Micropterygidae are perhaps reduced. The
different Hepialidae deviate very much from each other, the pro-
thorax is often highly reduced. It is of importance that a s. dor-
solateralis and a s. subdorsalis inf. frequently occur.

54

With a good deal of probability 1 1 abdominal segments can be
counted in instar / of H. hectus.

For these families I have collected as much literature as pos-
sible, and I think I am justified in my conclusion that Fracker
made a bad choice in raising the prothorax of one of the
numerous species to the rank of a fundamental form for his gene-
ralized . type.

Sub-order II. Frenatae.

As it was Fracker's (1915) intention to compose an analytic
list of determinations of the caterpillars, he had to examine as
many different families as possible. In some respects I have
started from a different point of view and I had by no means
such an extensive material at my disposal as he had. Though I
am convinced that it will be advantageous for a good system of
the Lepidoptera to be acquainted with the setal pattern of all
the caterpillars, yet I do not believe that the setal pattern in
itself is a reliable guide for the limitation of families etc. My
own experiences of the Sphingidae^ the Hepialidae, the Cossidae^
as well as Hofmann's observations of the Pterophoridae etc.
have increased my doubts upon that point. This result which
later on I intend to discuss more fully, added to my lacking
complete series in many families, and on the other hand the nu-
merous data given by Dyar, Packard, Fracker and others, al-
lowed me to confine my work to certain selected cases.

For those families which I have not examined I will confine
myself to mentioning the literature I have collected, generally
without entering into criticism.

The arrangement of the groups has been made according to
Handlirsch's plan (1908), but those families of the so-called
Microlepidoptera, which he does not mention, I have entered
into the series according to Fracker's system. This is based
on a compilation from Walsingham and on published and un-
published work of August Busck (1. c. p. 48, 61 sqq.). As this
system differs in many instances from that of Handlirsch, I have

55

placed the families, not mentioned by him, next to those with
which, according to Fracker, they harmonize best.

Family Xepticiilidae. Fracker (1915, p. 64) does not state any-
thing about the setal pattern, neither does Wood (1894).

Family Prodoxidae. Fracker (1915, p. 64) does not say any-
thing of the setal pattern.

Family Incurvariidae. Fracker (1915, p. 65) does not mention
the setal pattern. Packard (1895, p. 63, fig. 7) describes Adela
viHdella and Xematois violellui<. The setal pattern is not distinct,
but differs much in the two kinds mentioned.

Family Tischeriidae. Fracker (1915, p. 66) does not say any-
thing of the setal pattern. These last four families are taken to-
gether as the A c u 1 e a t a.

Family Acrolophidae. Fracker (1915, fig. 7, 8; I, II; 1, 2, 3;
p. 66) Pseudanaphora areanella. The abdomen bears : «. dor-
mlu< (of), s. subdorsalis (p), k. suprastigm(di}< (;) s. prostUjmalin (f ),
two .N". infrastigmah» of which the hindermost may agree with
«. posUHgmalis\ s. h<imlis^ three s. pedales, tt. centralis.

Family Tineidae. Quail (19046, PI. IX) gives a figure of
Tinea pellionella. There is an extraordinarily small s. prostigtnaUs
and under the stigma two s. infrai^tigmalex in one vertical line.

Fracker (1915, p. 67) says that the s. dorsales (x) are further
from each other than the s. subdorsales (/3). It may therefore be
possible, that here his x agrees with s. dorsolateralis.

Family Bticcalatrigidae. Fracker (1915, p. 67) says the same
of the segments 8 and 9 as of the Tineidae, further are x and yj
widely separated, thereby corresponding with s. infrastigmalis
and s. poststigmalis.

Family Lyonetiidae. Fracker (1915, p. 67) gives a statement
of the pattern but he himself does not attach much value to it.

Family Helioniidae. Fracker (1915, p. 68) says only that s.
poststigmalis and s. infrastigmalis are placed in diflferent ways on
the body of the genera examined. Lithariopteryx abrioniaella
carries them on diiFerent tubercles, Antispila nyssaefoliella on
the same tubercle.

56

Family Yponomeutidae. Fracker (1915, p. 69) mentions a great
difference between the several genera, the only characteristic in
common being (and even this does not always hold good) that
s. suhdorsalis (/3) is lower on the prothorax than s. dorsalis {x)^
that the K. group (s. prostigmales and s. infrastigmales) consists
of three setae on the prothorax, and that on the abdomen s. post-
stigmalis and s. wfrastigmalis are separate from each other.

Family Gracilaridae. Fracker (1915, p. 70) does not give a
setal pattern, neither does Chapman (1902).

Family Tortricidae. Fracker (1915, p. 71 — 74) gives a sum-
mary of the genera, which differ very much from each other.
Moreover the characteristics are not absolutely constant. Changes
occur with the following setae : n ; 5, /3, x, >;, ^a, K. It is of im-
portance that X and >j, i. e. s. poststigmalis and s. infrastigmalis^
may be arranged on the abdomen in a horizontal line as well
as in a vertical line and that f , i. e. s. prostigmalis is placed on
one tubercle with s. suprastigmalis (p).

Family Cossidae.

Material: Only a certain number of full-grown specimens
of Cossus cossus L. {ligniperda) were at my disposal.

Zeuzera pyrina L. from the Coll. Kail.

In connection with the recent views on the system of the
Lepidoptera the Cossidae have often attracted attention.

LiNTNER (1885) (nympha).

Dyar (1894 and 1894 &).

.Packard (1895).

Quail (1904 a. b).

Tsou (1914).

Fracker (1915).

The remarkable ribbon-shaped, twisted setae have been des-
cribed and illustrated by Lijonet in 1760 in his famous work
on the anatomy of Cossus.

The most accurate description is that of Quail, who also exa-
mined a newly hatched caterpillar of Cossus cossus. In doing

57

#

this, he discovered the existence of trumpet-like setae, which as
he believes, can open and shut. He also discusses Zeuzera pyrina
and CuluttMt expre»sa.

Tsor found two punctures upon the prothorax of Cossm, iden-
tical in ap|)eaninoe with the hiise of a seta.

Boas (1899) examined the thorax of this kind and found a
rudimentary thoracic stigma on the intersegmental membrane be-
tween the m«»«othorax and the metathorax, whilst he takes the
little 8pot«, generally considered as rudimentary stigmata, for the
rudiments of the wings.

COSSWS COtHtUK.

Prothorax : S, dontalis^ a, <hrsolateralis^ s. supraMigmalis^ x. stib-
dorsntis superior and .«. KHbdorsalis inferior and a special seta as
third in this Maries; in a little group before the stigma there are
three setae, probably agreeing with s. prostigmalis and two », in-
fraMigmnhs. Two ». ftasalea, two .<». propedale^, and s. ventralis.

A prothoracic shield is wanting but a dark-coloured spot occurs
in its place.

Mesothorax. S. dorsalis, s. dorsolateralis, 8. suprastigmalis, 8. post-
stigmalis, a little higher than usual ; s. prostigmalis or «. infra-
stigmalis, s. basalis, s. propedalts, s. postpedalis and 8. ventralis.

I draw attention to the fact that in Quails figure of the first
instar four setae are marked above the rudimentary stigma (?),
and moreover one before, one behind, and two below it, making
eight in all, which consequently entirely agrees with my statement,
because s. ventralis has not been mentioned by him. Tsou dist-
inguishes many more.

Metathorax = mesothorax.

Abdomen 1. S. dorsalis (smaller than the others) s. siibdorsalis,
s. suprastigmalis., s. poststigmalis (higher than usual) and one
s. prostigmalis which can only be seen with a higher power;
two 8. infrastigmales, s. basalis, three s. jyropedales, s. ventralis.
Segm. 2 :^ 1 .

On none of my specimens could I find A, of Tsou.

58

♦

Seym. 3, 4, 5, 6 = 1, but instead of the three s. pedales^ there

are three s. propedales, of which two are placed a little nearer
to the caudal part than the third. On the leg no setae
are found.

„ 7, 8 = 1.

„ 9, S. dorsalis, s, subdorsalis, s. suprastigmalis, s. poststigmalis
more ventral than on the other segments, two s. infra-
stigmales^ one of which is so much more ventral than on
the other segments, that it comes to lie almost exactly under
the other; s. basalts, s. (pro)pedalis and s. ventralis.

„ 10. Above the anal opening: s. dorsalis, s. subdorsalis,
s. suprastigmalis ; under it : lateral three setae = s. post-
stigmalis and two s. in frastig males ; ventral but above the
anal legs two setae = .s. basalis, and s. propedalis ; and
ventral under the anal legs: s. ventralis.

Nympha. Except the vertical rows of little hooks on the dorsum
there are no setae on the nympha.

Zeuzera pyrina L.

Coll. Kail. PI. I fig. 26 and 27. Full-grown specimen, length
55 m.m. The last segments have been a little damaged in stuffing.

Prothorax. Large paired prothoracic shield. On this I see only
two setae = s. subdorsales. There may have been more, which have
been broken off during the mounting. S. suprastigmalis is very
small and at the base of it there is no pigment, s. prostigmalis is
very large, possessing much pigment at its base; right over the
stigma a little pigmental spot without seta, s. propedalis, s. basalis,
s. postpedalis, s. ventralis.

Mesothorax. S. dorsalis, s. dorsolateralis, s. suprastigmalis is
doubled, s. prostigmalis in front of a dark pigmental spot, which
was plainly visible when the empty skin was examined, s. infra-
stigmalis, s. propedalis, two s. basales, s. postpedalis, s. ventralis.
Moreover on the left side, oral of s. prostigmalis a seta is found,
the base of which is surrounded by much pigment and on the
right side a seta caudal of the s. dorsolateralis without a pigmental

59

spot. On the intersegmental membrane between the mesothorax
and the nietathorax there is a little pigmental spot without a seta.
At first I looked upon the spot behind 8. prostigmalis as a
rudimentary stigma and supposinl the last-mentioned little spot to
be accidental. After having read Boas, it seems to me a more
probable solution that the large spots agree with the wing-rudiment
(Pt.) and that the little spot is the rudimentary stigma (St.).

Metathorax = mesothorax, but one «. baaalis is absent.

Abdomen 1, 2, 7, 8. S. dorsalis^ s. subdofsalits, «. suprastigma-
lis, 8. infrastigmalis all with a large pigmental spot at the base,
a small s. prostigmalis which is oral and dorsal of the stigma,
8. poststignmlis is absent; s. basalts^ 8. pedalis, 8. vetUmlis.
Segm. 3, 4, 5, 6 = 1 but on the leg are three 8. ptvpedales
implanted on the oral side.

, 9, 10 not clearly visible.

I draw attention to the fact that the imago of Zemera pyrina L.
shows an extraordinarily primitive pattern.

The agreement between Coxxm cos^us L. and Zeuztra pyrina L.
is exceedingly great, only the prothorax being different. Cossus
also has two s. in fra^tig males and Zeuzera has only one. I think
we may speak of a definite fundamental plan, but are obliged to
consider the full-grown specimens at least, as somewhat modified.
S. poststigmalis occurs in Cossus but not in Zeuzera^ so that Cos-
sus has evidently got more setae than there should be according
to the fundamental plan and therefore is more modified.

QuA.lL (1904) gives two s. infrastigmales on Zeuzera eucalypti^
instar 7, but does not indicate a s. poststigmalis. Neither does he
draw 8. prostigmalis.

Family Psychidae. Dyar (1899), Fracker (1915). „The setae are
very minute in later stages, in many cases it is impossible to
find them without a prolonged search. Abdomen with x, (3, §, p
(s. dorsalis, s. dorsolateralis, s. suprastigmalis ?) in an almost
straight line above the spiracle" (PL I, fig. 6 after Dyar
1. c. p. 179).

60

Family Elachistidae. Fracker (1915) does not mention the setal
pattern.

Family Coleophoridae. Fracker (1915, p. 80): „setae almost
indistinguishable, apparently in the normal micro-lepidopteran
arrangement".

Family Ethmiidae. The two larvae, examined by Fracker
I.e. p. 81, differed as to the abdomen, the thorax was the same.

Family Stenotnidae. Fracker I.e. p. 81, says that p and s are
situated below y, evidently s. dorsolateralis, s. suprastigmalis and
8. prostigmalis. Segment 9 is slightly different from the others.

Family Ilenierophilidae. According to Fracker (1915, p. 82) this
family differs a good deal from the Yponomeutidae with which
formerly it was sometimes connected. On the prothorax a (s. dor-
salis) is more lateral than (3 (s. subdorsalis), on the abdomen it
is the other way about, except on segment 9, which again resem-
bles the prothorax. The four genera examined differ slightly
from each other.

Family Gelechiidue. Fracker (1915), mentions as a distinction
from the Pyralidae, the three setae of the K-group ; and the distance
of seta /3 on segment 9 as a distinction from the Tortricidae.
He goes on to say that the genera differ and that the species of
Gelechia vary greatly (1. c. p. 84).

Family Oecophoridae. Fracker (1915, p. 85) cannot find a
satisfactory characteristic to distinguish them from the Gelechiidae.

Family Blastobasidae. This family differs from all the others
in p (= 8. suprastigmalis ?) on the abdominal segment 8 being
caudodorsad of the spiracle and x cephaloventrad ; therefore a
8. prostigmalis must apparently be present. Fracker (1915, p. 86).

Family Cosmopterygidae. Fracker gives on p. 86 : /3, 5, p in
a transverse line, apparently the same arrangement, which 0. Hof-
MANN (1898) found in some Pterophoridae.

Starting from the TiscJieriidae Fracker takes the families to-
gether as the TINEOID SERIES of the NON-ACULEATA. If
we try to find out the constant characteristics, we only can say that
the tubercula are monosetal and that there are no secondary setae.

«1

The usual type of the Micro's is to be found everywhere, but in
all groups niodifioations in place as well as in number of setae occur.

From this point of the series Fracker passes on to the Pyra-
lidoidea^ which Handlirsch places before the Thyrididae and
after the (ieomeiridae.

Handlirsch on the contrary goes on to the Zygaenidae, and I
therefore prefer to insert the ZYGENOID SERIES of the NON-
ACULEATA in this place, with the observation that IIan'dliksch dis-
cusses the family of the Meyalopyyidae,he\ong'mg to it, a little earlier.

Family Chalcoaidae and

Family Dttlceridae were not examined by Fracker (1915).

Family Pyromorphidae. Fracker (1915, p. 95) finds that these
larvae form the transition from a typical Micro into a Slug-cater-
pillar. These larvae possess verrucae to wit : r. dorsal is grown
together with v. subdorsalis. According to Fracker's fig. 59 be-
side these large verrucae occur : r. supr(istigtnali'<j v. infrastigmalis,
V. basalis, r. pedalL'i.

The pattern reminds us a little of the Saturnidae or the
LymaHtridae.

In the various genera there is some difference in the arrangement.

Family Epipyridae. Fracker says on p. 96 that there is no
sign of verrucae and that secondary setae are sparsely scattered
over the entire body. In this respect the family differs very
much from the former.

Family Megalopygidae. Fracker (1915) considers this family
as the transition from the Zygaenidae to the Cochlidiidae.

There are cernicae, on the abdomen: v. dorsalis united with
V. subdorsalis, v, poststignialis consolidated with v. infrastigmalis,
r. basalis and besides on abdomen 1 also one v. propedalis and
V. postpedalis. The thorax is a little diflFerent.

Family Cochlidiidae. Dyar (1899) has given a synopsis of the
Slug-caterpillars.

Fracker (1915, p. 97) only says that the verrucae some-
times have the form of scoli and that some of the genera are
entirely smooth.

62

If we try to find whether this ZYGENOID SERIES possesses a
definite characteristic, we see that except the Epipyrtdae^ which live
as parasites (Fracker 1915, p. 96), the families are characterized by
verrucae which sometimes grow out to scoli and which sometimes
disappear entirely. These verrucae are mostly grown together,
80 that a pattern arises strongly resembling that, which I have
called type la. From this it differs in some respects, so that no
direct descent of the Saturnidae from these Zygaenoidae can be
assumed, but only a parallel development of the pattern.

Family Zygaenidae. Fracker (1915) no statement. In Kallen-
bach's collection there are many species represented by larvae
which, however, are all fullgrown.

The striking spots do not take origin from the verrucae. On
the abdomen are : v. dorsalis (grown together with v. subdormlis ?)
V. suprastigmalis, v. infrapitigmalis (united with v. poststigmalis?)
V. haaalis. The pigmental spots which occur oftenest are: 1st a
spot mediad of v. dorsalis^ 2nd one in front of it, 3>'d one behind
it, 4th one behind the stigma, 5th one between v. infrastigmalis
and t\ hasalis. These spots may be doubled, they may also grow
together so that they develop into stripes.

The examination of these caterpillars in the youngest instars might
prove important. They are closely connected with the Microlepidpp-
tera, viz. with the ZYGENOID SERIES of the NON-ACULEATA.

Bomhycinae.

This gigantic group of caterpillars has been studied in detail
by Packard, (1895, 1905, 1915). It is very much to be regretted
that through his death the third part of his work has not been com-
pleted; the writer would probably have added a general synopsis.
All we have now is the very interesting introduction which
preceeds volume I and which was written without the experience
obtained during his study of the enormous material. His work
is the only monograph on caterpillars known to me in which all
or many instars of nearly all kinds of caterpillars have been
described and figured. The magnified setal pattern mostly has

63

been drawn next to them as a separate figure. I think that wo
owe this for a great part to Dyar.

Packard's classification is most easily studied from his: "genea-
logical tree of Lepidoptera" (1895, p. 83). He thinks all the
Honihycinae descend from the Lithoaiidae, and these from the
Tineina. The Sphingidctf descend like the Saturniidae from the
Ceratocampidae. This classification so far concurs with that of
IIandi.irsch (1008). This writer also places the families mentioned
in close contact, but thinks they have developed from different
ancestors. Against Packard's opinion may be adduced, that the
Bombycidae — Satuniidae^ the Sphingidae, the Lithosiidae have all
been found for the first time in the beginning of the Caenozoicum,
just as the Xoctuidae, the fteometridae,' the Hesj>eri)ine and the
Papilionidae s. 1., which according to Packard have all des-
cended with more or less intermediate groups from the Litho-
siidae and which for the greater part form the extreme branches
of his genealogical tree. In 1905 (p. 46) Packard gives another
classification which is slightly different but yet in principle the
Siime. lie thinks the Xotodontidae descend from the Thyatiridae,
which Handlirsch places next to the Hesperidiie. This clashes
with the palaeontological data, the Notodontidae are the youngest
family and are only known from the Quartair.

The series of Packard's families is: (according to 1905, p. 46)
Xotodontidae, Ceratocampidae, Satumiidae, Hemileucidae, Sphin-
gidae and Cerncinae as Syssphingina, opposed to which are
the Symbombycina with Dataninae, Apatelodinae, Euptero-
tidae, Ichthyurinae, Liparidae, Lasiocampidae, Endromidae, Bom-
bycidae and Brahmaeidae.

I have arranged the families according to Packard and wish
to point out Hajidlirsch's series: Bombycidae — Saturniidae,
Lasiocampidae, Sphingidae, Liparidae, Xotodontidae. At the end
of this discussion I shall return to this subject.

Family I. Xotodontidae. This large family contains seven subfami-
lies, of which according to Packard (1895) some (the first four) are

64

to be considered as the original forms of the Si/mhomhi/cinae,
others (the last three) on the other hand belong to the Sifssphingina.
The general pattern of setae is according to type I, in one sub-family
there are verrucae even in instar /, in others they appear later on.

1. Sub-family Gluphisinae is difficult to separate from some of
the Notodontinae. Some larvae are smooth, GlupJmia septentrio-
nalis possesses glandular hairs (Packard 1895, PI. VIII, p. 91)
which are very shortly forked and afterwards disappear.

According to fig. 1 b, the pattern on the abdomen is : s. dorsalis,
8. dorsolateralis, s. suprastigmalis, s. subdorsalis sup. and inf.
s. prostigmalis^ s. infrastigmalis which is placed very orally,
six s. basales, in instar 7, whilst in instar II s. dorsolateralis and
8. 8ubdorsalis inferior disappear. The presence of these two setae
in instar 7 is very important.

2. Sub-family Apatelodinae. The young larvae are covered with
long white setae, which are standing on verrucae according to
Type I, with v. subdorsalis inferior (Packard 1895, PI. IX).

3. Sub-family Pygaerinae. Packard (1895) makes a distinction
between the colour of the primary and secondary setae in the
full-grown larvae. Packard (p. 105) thinks that this sub-family
is the most generalized one of the family. As far as I can see
on PI. X — XIV the pattern of the Datana species agrees with
Phalera bucephala. In connection with the origin of the stripes,
which I was able to observe in a Pygaerine, this is of much
importance, just as the presence of an 11th abdominal segment.

To outline the family in an easier way, I have put the description
of Phalera bucephala L. after the discussions of the sub-families
p. 65 sqq. (On the origin of the stripes see chapter VII and VIII).

Fracker (1915) does not discuss the setal pattern.

4. Sub-family Ichthyurinae. Ichthyura apicalis (Packard 1895,
PI. XV) has setae, I. inclusa and I. albosigma (1. c. PI. XVI)
verrucae according to type I.

5. Sub-family Notodontinae. According to Packard (1895, PI.
XVII — XXIII) the larvae possess setae according to type I.

6. Sub-family Heterocampinae. Larvae sometimes with stema-

65

tojx)dtt (Marnirocampa). Larvae (Packard 1895, PI. XXIV —
XXXV) with setae according to type I, sometimes with a small
s. prostiymalis: {Hyparpax aurora, Schizura unicornis).

It is remarkable that Packard in 1905 (p. 44), while discussing
the setal pattern does not figure the .s. prostifftnalis on Schizitra.

Exceedingly peculiar setae resembling antlers are born by Jfe-
terocampa during instar /, later on they disappear. They seem to
agree with .*<. dormlut of the prothorax (1. c. PI. XXX).

7. Sub-family Cerurinae. I^ong stematopoda, setae according to
type 1 (Packard 1895. TM. XXXIV— XXXVII).

Sub-family 3. Pygaerinae,

Phaiera bucephala Linn. Plate III, fig. 7 — 13.

Material in alcohol. Collected at Groningen, summer 1915.

In this caterpillar the head becomes coloured last, while in
nearly all others the head is black immediately after moulting
and the tubercles become coloured afterwards.

Instar 1. Duration 9 days. Length 2"j mm. The tubercula
are not black. The setae are not plumed.

Prothorax. V. dorsalit in front of the prothoracic shield, con-
sisting of three setae, about */, mm. long, they are not united.

V.suhdorsalis. Three setae which are longer than 1 mm., inplanted
on a tuberculum, which is clearly higher than the prothoracic
shield. V. suprastig malts has also three setae of a length of not quite
7, mm., concentrated on a tuberculum which in most individuals
has coalesced with the prothoracic shield.

No s. dorsolateralis, v. poststigmalis with some (mostly 3 or 4)
setae, if at least they have not become combined with the pro-
thoracic shield, seta in frastig malts, v. basalis with two setae. No
s. pedalis and no s. ventralis.

Mesothorax and metathorax. V. dorsalis with three setae, v. dorso-
lateralis, V. suprastig malis and v. infrastigmalis, each of them
with two setae. V. basalis is not distinctly outlined, there are
two setae which are not joined together.

5

66

No 8. pedalis or s. ventralis.

Abdomen 1, 2. V. dorsalia with two setae, so that I presume
s. dorsokiteralis has been added to them, s. subdorsalis, s. supra-
.ttigtnali!^, s. infrastigmalis and s. poststigmalis each of them with
one seta. V. basalis with two or three setae and on the place
where there is a leg in the segments 3 — 6 one setn = s. pedalis.
Seym. 3, 4, 5, 6 = 1, but one s. basalis is missing, whilst there
are along the outer and the lower edge of the leg mostly five
or six rather long setae (+'/2mm.) I consider these to be
the s. pedales. It may be possible, however, that the s. basales
of 1 agree with these s. pedales of segm. 3 — 6 and that
the seta there called s. pedalis really is a s. ventralis. This
supposition appears especially probable, when the caterpillar
is looked at from the ventral side and, starting from seg-
ment 3, we try to explain the arrangement of 4. If, on
the other hand, the caterpillar is looked at from the lateral
side and segment 1 is started from, we hesitate to give
the above explanation for segment 3.
, 7,8,9 = 1.

„ 10^ 1. «S. dorsalis, close to it lies s. suprastigmalis. These
is a s. basalis and the setae forming together the s. pedales
are very strong and are + '/4 mm. long.
„ 11. Behind the setae of 10 there are on the anal flap three
more setae which are arranged in the following way:
s. dorsalis. s. subdorsalis, s. sujrrastigmalis.

Instar //. Duration 10 days. Length 7 mm. The arrangement
of the setae is almost the same as during instar I. The setae
are not plumed either, but the tubercula are black and on
the front and the back edge of nearly all the segments there is
a black spot bearing many setae, which are smaller than
those on the primary tubercula. In the figure I have marked the
median spots with a darker colour than the paired ones. The
tubercula of the other side, which are visible on PI. Ill, fig. 10,
lateral aspect, are striped.

67

Prothorax. The prothoracic shield has disappeared.
V. (iormlis very hirge and protruding, provided with 3 — 8 setae,
more than 1 mm. long. V. subdorsalis small, consisting of some
setae behind the large r. dorsalis. V. suprastigmalis with three setae
just as r. piostigmalis. Of this last wart the setae are smaller.
Under the stigma one seta ^= s. itifnistigmaliK. V. Inisalis with
three setae. On the leg some smaller setae.

Mesothorcu. V. dorsalis, v. dorsolaterals, r. suprastigmalis, all
of them with two setae. Seta prostigmalis and seta bimtUs with
one seta. At the back edge of the segment a black dorsal spot
with very short setae.

Metathorax = mesotfiorax but v. prostignuilis has two setae and
r. basalts four. Further there are two small setae behind r.^ros////-
tnalis and just over the leg on the front side is r. pedalis.

Abdomen 1, 2. Two nunlian shields occur at the oral and caudal
edge, of which especially the latter bears many small setae.

V. dorsalis with two setae, seta subdorsalis, s. suprastigmalis,
s. poststigmalis and .<». infrastignmlis, all of them + '/, mm. long.
One row mostly of five .s. basale.<, and one large s. pedalis.
Segm. 3, 4, 5, 6 = 1 , but the large s. pedales of instar / are
absent here, and instead there are 4 — 6 large setae above
the leg = .s. basales. On the leg itself there is a large spot
with small Bet&e = s. pedales. One gets the impression that
the 8. pedales of instar / are placed higher. Beneath s. sub-
dorsalis a spot or a seta.
. 7,8 = 1.
^ 9 = 1 but r. suprastigmalis has two setae. S. poststigmalis

is absent. No median dorsal shields are found.
„ 10. All the setae situated above the stigma from the left
to the right side are placed on an anal shield. These setae
are ± 1 mm. long. Further there are on this segment no
setae, except a few on the outside and inside of the leg,
which are + ^It nim. long.

Instar III. Length + 15 mm. The number of spots with small

68

setae between the primary warts and setae has increased. More-
over there occur some scattered setae. The setae on the primary
tubercula are ± 1 — I'/a mm. long. The longest which are at the
same time the biggest, are found on the prothoracic shield and
on the anal shield.

Prothorax. On the prothoracic shield three setae, which in my
opinion represent s. dorsaUs, s. dorsolaterals and behind it a small
one = s. subdorsalis. V. suprastigmalis has three setae just as v.pro-
Htif/malis. The seta behind v. suprastigmalis is also plumed. V. basalts
with three setae. The small setae on the leg are not feathered.

Mesothorax and metathorax. On the front-edge of the segment
a double row of spots, on the hind edge one row.

V. dorsnlis has three setae, v. dorsolateralis just as v. suprastig-
malis has two. There is a single seta j>'f'ostigmalis. One large
V. basalis and many small setae which, however, are not fixed on
black tubercula.

Abdomen 1 — 10 as in instar /, but:

Seta infrastigmalis is single. Three setae, which are not united
and not implanted on a tuberculum, stand instead of v. poststig-
malis. The number of ,s. basales amounts to +10.

Instar IV, Length ± 25 mm.

Apparently this instar completely resembles the full-grown form.
Small setae, not plumed, are to be found everywhere. The spots
have become more numerous and bear setae of ± 1 mm. long.
Still more than in instar /// they are arranged in horizontal and
vertical rows and thereby give the impression of forming stripes.

The primary tubercula may be recognised by the seta which
are longer (mostly 1 '/2 mm.), and generally thicker too. They
have short plumes. On each wart there are more setae than be-
fore, which is rather striking as during instar /, // and ///
the number is almost constant.

As examples of the most intricate pattern of spots I chose
the prothorax and abd. 5, the others agree with these mu-
tatis mutandis.

6ft

Frothorax. There is ii large protlioracio shield which on each
side has two very long setae (3 mm.) and behind them two or three
shorter ones ('/^ mm.). These taken together I consider to be
V. (iormlix and apart from them is one seta corresponding with
s. subdorsdlix. T. ttupnuftigiHalis and r. dorsolateralis over the
clearly visible v. pi-ostigtnttlis. V. pro»tigmalh projects a good
deal and has three setae, v. basaliA four. Over the stigma are
two tubercula with five and two setae. They t^ike their origin
from the spots which have appeared in instar ///.

Abdomen 5. Both the median dorsal shields, arisen during in-
star //, are still present. The verrucae dorsalea, r. subdorsalis,
r. supr(k'<tigniali'<^ i\ jMjshtigmalis are clearly visible. The per-
manency of this last wart is interesting as it showed an inclination
to dissolution in instar ///, V. infrastigmalis consists of some
scattered setae. There are less setae bamles than in instar ///,
there is also a r. pedalis present in the form of a spot. Other
spots are: on the front edge under the median shields a large one and
two smaller ones of which the 2nd lies in front of r. HuproHtig-
malis^ next come a large and a small one about on the same height
as the stigma and r. jMststigmalis. Behind the first smaller spot of
this row we find two a little larger, and behind these in the same
row V. subdor!<alis. Over r. poststigmalis is a very large spot,
which lies therefore in one row with the 2nd large spot of the
1st vertical row. Under v. poststiginaUs are small spots with long
setae which I first took to be primary ones because of their
length. They lie in a row with r. infrnstigmalis. Over the setae
basales is a long-drawn spot. In different individuals and in diffe-
rent segments of one individual, these spots and tubercula vary
a little in form and size. Most of them resemble the above example
and one can always recognize the same primary tubercula.

Instar V. Length 40 m.m.

The primary tubercles can not be found. All the setae are
somewhat plumed and are + 2 m.m. long. The spots have almost
grown together to stripes. Round each seta is a little area free
from pigment. Over the stigma are, the dorsal median included,

70

three stripes agreeing with the row of spots running through
V. subdorsal^, v. suprastigmalis and the spots over the stigma.
There is no clear stigmatic stripe. The infrastigmatic stripe is distinct
on the front and back edges of the segment, the basal stripe is
narrow, and there is a clear, broad pedal stripe.
Pupa. I could not find any sign of a pattern.

Family Kupterotidae. Fracker (1915) includes Apatoledes which
deviates in respect to the tubercula. I think that he is right as
regards the larva.

Family Liparid(ie = Lymantndae. The tussock-moths can be
divided into two groups, those with and those without pencil-shaped
setae. Several of them have long since drawn the attention of
investigators because of their peculiar pattern, the thick tussocks
of setae and because of the sexual dimorphism of the cater-
pillars.

Literature:

SwAMMERDAM (1737), Orgijia 9 and cf caterpillar, pupa with setae.

HCbner (1766), development of the setae.

Riley (1885), Orgyia.

Waciitl and Kornauth (1893), particular setae of Ocneria.

Packard (1889 and 1893), development of Orgyia.

Fracker (1915), synopsis of the family.

The last mentioned writer says (1. c. p. 104 sqq.) that Porthetria,
Gynaephora and Euproctis have the ordinary verrucae like the
Ardiidae, except that there are three verrucae on the mesothorax
and the metathorax above the K-group.

He thinks that the coalescence of x with p which occurs in
Porthetria = Lymantria (therefore of v. poststigmalis with v. supra-
stigmalis) is unique. He examined Euproctis = Porthesia but did not
observe the same arrangement there. The second group has pencils,
to this belong Olene^ Hemerocampa and Notolophus = Orgyia.

Fracker does not attach much importance to the question whether
X and /3 (i. e. verruca dorsalis and subdorsalis) coalesce or not.

I examined three complete series, my principal results are:

71

1. The coalescence of i\ snprmtiymalia with r. poKhtiyninlig,

2. The disappearance of r. dursalis in diflFerent ways.

3. The structure of the setae becoming more intricate in the
course of the successive moultings.

4. The presence of peculiar setae on LymantrUt UisjHir.

5. The disap])earance of the prothoracic shield.

6. The presence of verrucae on the pupa, to which Swammerdam
(1737) already drew attention.

7. The verrucae on the larva are arranged according to type I
or to type 16.

Family Lipan'dae Linn. Plate II, fig. 8 — 11.
LytiMntria (Ocnena) dittpar L.
Material in alcohol and living animals (summer 1915). In the
successive instars of caterpillars the same pattern is always found,
the tubercles are warts with many setae.

Instar /. Duration 12 days. Length 4 mm. The setae are diffe-
rent in form viz. :

1. Very short feathered, light coloured ones, with a length of
± 1 '/j — 2 mm. (the side-branches are ± 7 (i). The base is
enclosed in the black case of the wart.

2. Not feathered ones, mostly 400 — 500 (i. long and coloured deep
brown (PI. II, fig. 10). The lower part is ± 100 /c* long and a little
narrower than the elevation of the wart on which the seta stands.
A little bladder follows, with almost colourless wall. The diameter
is + 20 fi. Through the presence of this bladder movements of
the top part of the seta are rendered possible. This part is
+ 300 fi long and tapers into a point. At first I thought that
these peculiar setae were of use during hatching, as I imagined
that before and during the removal of the skin, they were folded
up. The examination of newly hatched larvae showed me, however,
that there these setae already stand upright. The number of the
bends is decidedly not larger than afterwards. I surmise that the
same organs occur in the closely related Lijmatitria monacha.
According to Sharp (II, p. 407), Wachtl and Kornauth (1893)

72

think that there they serve for the dispersion of the larvae by
the wind. In that case, however, they should have been much
longer, for these small vesicles with a radius of ± 10 /!/., cannot
excercise any influence on the specific weight. Moreover, the long
setae are more suitable for this purpose. It is certainly remark-
able that these setae only occur . in the first instar both on
Lijmantria dispar and on Lymantria monacha. Their function is a
riddle to me; for the time being, I identify these setae with the
glandular hairs which often occur in a first instar. Directly after
hatching, the head is coloured black, the tubercula become
coloured later on.

The following tubercula are visible:

Prothorax. Small v. dorsalis with short setae, rather large v.
xKbdorsalix on which the setae mentioned sub. 2 are present
(this verruca is connected with the preceding one by an indistinct
prothoracic shield); a far projecting v. suprastigmalis with which
V. prostigmalis is connected, v. infrastigmalis, v. basalis and one
V. ventralis.

Mesothorax and metathorax. Very large v. dorsalis^ small v.
dorsolateralis, large v. suprastigmalis, a long drawn v. infrastig-
malis, V. propedalis, v. basalis, v. ventralis.

Abilomen 1. Very small v. dorsalis and an extraordinarily large
V. subdorsalis. The verruca over the stigma is divided into two
parts by a furrow. The front part is v. suprastigmalis, the back
part V. poststigmalis ', v. infrastigmalis is elongated, v. basalis
small, no v. pedalis.

This pattern I have called type lb.
Segm. 2=1, but v. infrastigmalis consists of two parts, one be-
hind the other.

„ 3, 4, 5, 6 = 2, but V. basalis is absent.

„ 7, 8, 9 = 2, but on 9 the verrucae are much higher than

on the other segments.
„ 10 is strongly reduced, the verrucae are not distinct.

Instar //. Duration 12 days. Length 8 mm. The same as
instar /, but the peculiar setae have disappeared. The setae bear

73

Mutncwhiit longer plumes, r. (loi'snlis becomes smaller. The pio-
thuracio shield has disappeared.

Instar ///. Duration 8 days. Length 15 mm. As instar //.

Instar I\\ Duration 6 days. Length 24 n>in. As instar //, the
sides of the head become grey.

Instar V. Duration 11 days. Inrngth 35 mm. As instar //
but some secundary setae appear on the body and the v. in/ni-
stitjmalfs are sometimes divided into many pieces, on the thora-
cic segments the v. pedales are wanting.
Pupa.

On the abdomen of the pupa the verrucae are arranged in a
distinct way according to type I b. The furrow between v. supm-
^tiijmalis and r. i)Oststi(jinalis has grown larger.

The setae on the pupa too are plumed. Distinct are : v. dorsatis
even larger than on the caterpillars, v. subdorsalis, v. suprastig-
omUs 4- V. poststig malts, r. infrastigmalis, v. basalis and a far
projecting r. pedalis,

Euprodis (Porthesia) chrysorhoea Linn., Plate II, fig. 12 — 15.

Material in alcohol, from Groningen (spring 1915) and living
material from the Hague (autumn 1914).

I have not been able to obtain the newly hatched caterpillars
of this species. The youngest are of December 19th 1914. They
were found in a nest on Hippophaes rhamnoides Linn, and as
in the course of their development they cast the skin four times,
and were only 4 a 5 mm. long, I have described them as
instar I.

Instar /. Duration at least 130 days, they hibernate/ The first
moulting took place on April 27th. Length 4 mm. Tubercula
formed like warts.

The setae are not plumed, ± 1 mm. long.

The colour pattern is very intricate, there is a distinct median
dorsal line.

Prothorax. Both the v. dorsales from the left to the right are
connected by a distinct prothoracic shield. The v. subdorsales are

74

also connected by a shield which is smaller, however, than the former.
The V. suprastigmalis projects very much.
Small V. infrastifj/nialisj very large v. hasalis. No v. pedalis.
Mesothorax and Metathorax. V. domdis, v. dorsolateralis and
V. siqmtHtifjmalis all equally large. No v. infrastigmalis. Very large
V. hasalis. On the median side, close to the leg a small v. pedalis.
Abdomen 1,2. V. dorsalis large, united with that of the other
side. They have a dense bundle of coloured setae, length ± ^l^mnx.
V. subdorsalis smaller than on the other segments. V. supra-
stigmalis, and V. infrastigmalis are very large, v. basalis small
and V. pedalis also.

Segm. 3, 4, 5, 6. V. dorsalis is wanting in most individuals. Very
large v. subdorsalis, large r. suprastigmalis and v. infrastig-
malis. Elongated but very slender v. basalis, and on the front
edge of the leg a small v. propedalis, on the leg some setae.
„ 4 and 5. V. dorsalis is present and is even rather large.
„ 6. V. dorsalis is small and between the v. subdorsales is

a median brown elevation.
ri 7 = 6. V. pedalis, however, is placed more backwards, about
on the spot where the legs are joined to the segments 3 — 6.
„ 8 = 7, without the median elevation. V. basalis is closer

to V. pedalis.
„ 9. V. dorsalis is absent and v. infrastigmalis is small ; the
V. subdorsales of the left and right sides are united into
one shield. For the rest as 8.
„ 10. Small but strongly projecting v. subdorsalis, v. suprastig-
malis, V. infrastigmalis. No v. basalis and very small v, pedalis.
Instar II = I but the prothoracic shield has disappeared and
the setae are distinctly feathered except the scattered ones on
the legs. The v. dorsales of the segments 6 and 7 have disappeared.
V. suprastigmalis is divided into two parts as in Lymantria dispar.
This I consider to be a coalescence of v. sujjrastigmalis and
V. poststigmalis and I think I could also observe this in Instar/.
Instar ///, IV, r= instar //.
During instar V a dark spot is generally present on the meso-

75

and metnthorax, on the place where the stigma should have occurred,
if it had been placed on the back edge of the segment, as is the
case on the prothorax. At first I considered this spot to be n
rutlimentary stigma, but after having read Boas (1895)) and after
a repeated examination, I take it to be a transparant wing-rudiment.
The appearance of this spot in Instar V proofs the latter
suggestion to be the right one.

Orf/i/ia antiqua Linn.

Most writers only mention the beautiful c«)lours, the pencils
and the tussocks of this larva, these attributes being generally
found in the members of this family.

The principal statements are:

Buckler (III, p. 11 and 12) gives an excellent description of the
fullgrown caterpillar, but of course without mentioning the posi-
tion of the tubercula. Of the first stages, however, he says very
little. Thus for instar / he only mentions that the tubercula
are black, and that after the fourth moulting (instar T) they are
red. The tussocks on the segments 5 and 6 in instar IV show
a kind of black hue, those on 7 and 8 a sort of white. After the
4th moulting they are all white, later on, during this stage, they
become brown. In instar / a broken subdorsal line was seen.

His drawings (PI. XXXIX, fig. 1, la, 16) give different varia-
tions of colour. My specimens agree for the greater part with
fig. 1. It seems to me that his fig. 16 shows the same mistake
as that of Hubner.

HoFMASJf-SpuLER (1910) says that the cT caterpillars are smaller
and that they have yellow bristles on the back, whilst the large
9 caterpillars are provided with yellow-brown bristles. In this
he agrees with Sw^ammerdam (1737). His drawings (PI. 15, fig. 25a b)
are very bad.

Jacob Hubner gives as a frontispiece to his book in four volumes
on caterpillars, a drawing of Orgyia antiqua^ of which the length
of the body without the setae is 11 cm. and with the setae 21 cm.

The drawing is certainly large enough to justify the expec-

76

tation of a great accuracy in the arrangement of the tubercula. On
examining it closely, however, one sees that on each segment a
verruca is left out. V. suprastigmalis was drawn excessively large
and thereby the v. in frastig males were placed too low, so that no
room was left for a v. basalts^ though the legs are plainly visible
from the side and it should therefore have been represented.
On the prothorax v. subdorsalis has not been drawn, though
in reality it is present, and the false impression is given that
V. dorsalis is as much developed as v. suprastigmalis. In fact
the latter projects a considerable distance from the body and the
former is visible as a median protuberance on the base of it.

I have treated this case somewhat in detail, because it shows
that the confidence, which I originally placed in existing illustra-
tions, was misplaced.

Dyar (1894) defines the Lymantridae with the following words:

"Not more than one tuberculum on the third annulet and only
six above the base of the leg.

"Not more than two tubercula on the middle annulet, and
generally one on the third; one prothoracal shield.

"Tubercula IV and V {v. poststigmalis and v. infrastignialis) far
from each other or IV has disappeared. Tubercula with many
setae, no setae on the skin."

Packard (1889, p. 55 — 59) gives the fullest description.

As has already been observed, he describes only four stages,
whilst Buckler gives five; my investigation has also shown me
that there are five different instars.

Packard gives as the most important results:

Instar I. Duration 7 — 8 days. Length 4 mm., tubercula black,
the middle ones on the thoracal segments smaller than the lateral
ones. Setae thinly spinose, very long. The two glands on the
abdominal segments 6 and 7 not clear.

Instar //. Duration 4 days. Length 6 — 8 mm. Tubercula black ;
except the two large lateral ones on the 1st thoracal segment
which are red at the base. The glands are coral-red. There is
a subdorsal line which is not quite complete. Towards the end

77

of this stage there appear some plunuH] setae on the dorsal tuber-
cula of the 8th (abd.) segment and sometimes on the same tuber-
cula of the l«t and 2n(l (abd.) segments.

In»tar ///. Dunition 5 days. Length 10 nun. The lateral tubrr-
cula of the prothorax are of a pale Indian red, with black in between
and form a pencil of plume-like setae which grow thicker towards
the end and which are as well developed proportionally as those
of the fuU-finl larva.

The four median dorsal tufts are well developed, the two front
ones are deep brown, the two back ones are white. The 8th (abd.)
segment also has a long pencil of plume-like setae. All the
lateral tubercula are of a bright flesh-coloured red. With some
the colour of the dorsal tufts changes.

Instar IV (=: last). Duration 7 — 14 days. Length 17 — ? mm.
The cJ* ones sooner develop into pupae than the 9" The dorsal
tufts become pale buff-yellow. The tubercula are of a bright
coral-red, except the dorsal ones of the segments 2 and 3 which
are of a bright yellow.

The data which Packard gives make it appear that this larva
only possesses four instars in America. The report of the 2nd moult
on May 22nd (p. 55) ig probably due to a mistake, apparently
the 1st moult is meant, as the hatching took place on May 15th.

Orgyia antiqua Linn. Plate III, fig. 1 — 6.

Material in alcohol from June till August 1915 and of in-
star 7, II and III also of June 1914. Collected at the Hague
and Groningen.

Instar 7. Duration 6 days. Length 3 mm. At the hatching the
head is black, the tubercula are light grey, they get coloured
half an hour later and become black.

The colour-pattern is very intricate, but the colours are not
proof against the influence of alcohol. This is also the case
during the following stages, so that I do not mention them here.

See Packard (1889, p. 55—59).

The r. subdorsales are triangular and the setae which are fixed

78

on them point in all directions. The other verrucae are mostly
elliptical with a long horizontal axis. The setae are arranged more
or less in one line, the middle seta is mostly the longest and the
others get symmetrically smaller in regard to this one. The longest
setae are born by the v. suprastigmales. The setae are not plumed.

Prothorax. There is a small v. dorsalis, a very large v. suh-
dorsalis which is connected with the corresponding verruca of the
other side by a black coloured prothoracic shield. By this the
small 4). dorsales get united. Moreover, there is a rather large
V. suj>rastigmaliH which projects a good deal, a v. infrastignialis,
which has an offshoot pointing towards the dorsum, which per-
haps corresponds to the missing r. prostigmalis and a v. basalts
just over the beginning of the leg.

Mesothorax and metathorax. There occur: v. dorsalis, v. dorso-
latemlis, a little larger than the v. dorsalis, v. supmsfigmalis,
V. infrastigmalis and a small v. basalts. The setae on these seg-
ments are shorter than those on the other. One might be inclined
to think that v. dorsolateralis owed its origin to the shifting of
r. subdorsalis. This verruca on the mesothorax and metathorax
is placed in front of v. dorsalis and on all the abdominal segments
V. subdorsalis lies behind it. Besides in other families the same
arrangement occurs as in this one, but then there is also a ver-
ruca or seta in the place where v. subdorsalis lies. For these
two reasons I think that the name of v. dorsolateralis deserves
the preference. (See Chapter IV).

Abdomen 1, 2, 3, 4. There are a small v. dorsalis, a large
V. subdorsalis which is prolonged in a triangle towards the oral
side, beneath the v. dorsalis a v. siiprastigrnalis which has the
longest setae, a v. infrastigmalis and a small v. basalis.
Segm. 5. V. dorsalis is very small in some individuals and totally
absent in others, for the rest as on the abdomen 1.

„ 6. V. dorsalis is very small and generally bears only one
short seta, for the rest = 1 .

„ 7 = 1, In some individuals v. dorsalis has been partly
split into separate setae.

79

Seym. 8=1. Two s. basalcs.

^ 9. No V. dorsalis, r. subilorsalis and v. suprastigmalis very
large and strongly projecting, r. infnistiffmaUs about as
large as on 8, a wreath of s. hasales.

J, 10. It is difficult to reduce the verrucae to the pattern of
the former segments. I consider the wart under the anal
opening as a r. infriisi'ujmalis.

It seems to me that the two setae hasales of segni. 8 are pseudo-
primary ones. If the newly-hatched larvae are examined many
individuals with a well developed wart are found. Most parts do
not become coloured, but only the two extreme setae. I do not
think therefore that this case should be taken as a proof of the
thesis that warts have taken their origin from simple tubercula,
but I tliink we have to deal with the secondary dissolution of a
wart into simple setae.

The two dorsal glands which are so obvious in the following
instar on the segments 6 and 7 are faintly visible.

Instar II. Duration 10 days. Length 6 mm.

All the setae are short-plumed, the side branches arise at irre-
gular distances. The form of most warts is about the same as of
those of instar /. The setae are not very different in length.

Prothorax. V. dorsalis is small and is connected with that of
the other side by an indistinct prothoracic shield; a very small
r. sttbdorsalis which is remarkable in connection with the size it
possesses in instar 7, neither is it connected anymore with the pro-
thoracic shield. V. suprastigmalis is very large and projects con-
siderably. On it are the longest setae (2 mm., short feathered)
and between these are mostly five short ones, which are much
thicker and have longer plumes. It is an exception when they
are longer than '/^ mm. V. infrastigmalis is small and is no
longer so tall, v. basalis too is small.

Mesothorax and Metathorax. We find the following small warts :
r. dorsalis, v. dorsolateralis, v. suprastigtnalis, v. infrastigmalis,
r. basalis.

Abdomen 1. V. dorsalis is small, v.suhdorsalis very large with

80

longer setae (l'/2 mm.) than in instar /, v. stiprastigmalis, v. in-
frastigmalis and v. basalts about as in instar /.
Segni. 2 = 1 ^ but v. subdorsalis is still larger and grows along
the ventral edge of v. dorsalis in the oral direction.
„ 3, 4 := 1, but on the leg we find a black spot with a
number of small setae which are only partly plumed. This
spot I consider to be a modified v. pedalis.
jj 5=1, but V. dorsalis is wanting.

^ 6 = 3, but V. dorsalis is absent and in the median line a
red elevated gland is visible, which has the form of a
champagne-cork.
^ 7 = 6, but there is no v. pedalis.

^ 8 = 7. Without the gland. The two v. subdorsales of the

left and right sides are placed very closely together and

also have the particular setae mentioned in segment 1.

„ 9, 10 =: 7. Moreover, the wart of segment 10 of instar /

which was considered to be v. infrastigmalis is well developed.

Instar ///. Duration 11 days, length 10 mm.

Prothorax. There is no longer a prothoracic shield. F. dorsalis

is small, lying against the median side of the very large, far

projecting v. suprastigmalis. On the top of the last wart we find

short plumed setae of a length of ± 2 mm. Amidst these are shorter

ones (± 1 '/i mm.) of which the side branches are a little longer.

Towards the end of the setae these side branches grow somewhat

longer and thicker and are placed more closely together. Each

of these peculiar setae reminds one a little of a French plumeau

(feather mop) or of one of the feathers on the head of a Goura

victoriae. V. subdorsalis is small, v. infrastigmalis large, v. basalis

very small.

Mesothorax and Metathorax as in instar //.
Abdomen 1, 2. V. dorsalis and v. subdorsalis have united with
each other and also with those of the other side. This method
of origin is still clearly visible. They bear one tuft of setae,
+ 1 mm. long, which show longer and thicker side-branches than
the setae generally possess, and they are of a deep brown colour.

81

r. .tujttastiffmalis is large, just as r. infrastigmalis. The setae
of these tubercula are ±2'/, "im. long. V. httsalis is small.
Seym. 3, 4. V. dorsalis is absent, the r. subilorsales of the left
and right sides have coalesced. They have setae as those
of 8t»gnient 1, but are a little shorter and are white in
colour. For the rest as in instar //.
„ 5. V. (iorsalh is absent, the other tubercula as in instar II.
y, 6, 7 as in instar //.

„ 8. The two r. auMorsales of the left and right sides arc
united, but not coalesced. They project higher than in instar //.
The setae closely resemble those of the segments 1 and 2.
„ 9, 10. r. infrastigmalis of 10 remains large.
Instar IV. Duration 8 days. I^ength 15 mm.
Prothorax. V. «lormlis is small and ifl now to bo found as a
protuberance placed* in the median line on the base of the strongly
developed r. suprastigmalis. On its top is a black spot on which
are implanted short plumed setae, ± 3*/, mm. long and two
tufts of plumose setae viz. ± 1 — I'/j mm. and ±2'/, — 3 mm.
long. Each group consists of +6 — 10 setae. There are no tran-
sitions between these two, so that they clearly form two storeys.
The longest ones also have the largest side branches. The other
tubercula are as in instar ///.

Mesothorax and Metathorax. As in instar /// but v. dorso-
lateralis has grown smaller.

Abdomen 1, 2, The four warts which in instar III began to

coalesce, now form a whole. For the rest as in instar III.

Segm. 3, 4 = 1. It should be observed that here r. dorsalis has

not been united with the median shield, but has disappeared.

The colour and the form of the setae now agree with

those of segment 1.

The other warts on this segment and on the following
segments correspond with those of instar III but:
„ 8. The two v. subdorsales are almost united and between
the many plumed setae some plumose ones are found as
on the prothorax.

6

82

Instar V. Duration 10—20 days. Length 23—30 mm. For
this stage I made an exception where the drawings are concer-
ned. I did now draw the objects themselves under the microscope
with the aid of a camera, but took the outline of Jacob Hubner's
old frontispice (1766) and in this entered the different tubercula.
The changes are not so great as in instar IF.
Packard does not describe this stage (see p. 76).
Prothorax. The plumose setae have a length of 5 and 10 mm.
Bnd stand out above the short feathered setae.
Mesothorax and Metathorax. As in instar IV.
Abdomen 1. V. infrastiymalis has a tuft of bright yellow setae,
± 7 mm. long, short plumed, for the rest as in instar IV.
Segm. 2=1. V. infrastigmalis is provided with a little group
of plumose setae, the length of which is + 7 mm.
y, 3 to 7 as in instar IV.

„ 8. The V. subdorsales have a tuft of setae of different forms.
Between the short plumed setae which are more than 7
ram. long, we find plumose ones of a length of + 10
mm. and long feathered ones, ± 5 mm. long. Further as
in instar IV.
y, 9. The V. subdorsales are not united and have each a tuft
of very long, short plumed setae and some (2 to 4) plu-
mose ones, the length of which is 5 mm.
The differences between Packard's views and mine follow
here in a short synopsis.
Packard.
Setae are already spinose in instar I.

The feathered setae towards the end of instar II on segment 8.
The plumed setae in proportion as large as on the full fed
caterpillar in instar III. They appear at the same time on the
prothorax and segment 8.

Whole duration of all the-instars as caterpillar: 23 — 30 days.
Three moults.

SCHIERBEEK.

In instar II spinose or plumed setae for the first time.

83

Iinnu'diatt'ly after the moult on prothorax and segment 8 setae
which differ from the rest. The plumose ones shorter than the
ordinary setae in instar /// and IV. Those of segment 8 only
come in instar IV and are still very short.

Whole life as caterpillar is 45 — 55 days. Four moults.

I hope that these differences may give rise to new investigations.

It also seems to me that the histological structure of these
setae and the manner of their origin within the more simply orga-
nized ones of the f«irmer instar, are worth while examining.

I draw attention to the fact, that J. U. Kruimkl in his inves-
tigations on the feathers of the Gallinao (1916) also found a
more composite structure in the successive ^editions of feathers"
whifh apjM'ar after the different moults.

Family Ijasiocampidae. Fracker (1915, p. 103) thinks there
are too many setae to be able to describe them. Some genera
are characterised by a dorsal horn on segment 8.

Dyar (1893 h) has described different kinds.

Lasiocamjxi nihi L., Plate II, fig. 6, 7.

Material in alcohol of the three last instars of the cater-
pillars (i. e. instar ///, /F, F), collected at Groningen, in the
summer of 1913.

Instar ///. Length 27 mm.

The whole body is covered with widely spread setae of about
'/i mm. long. The tubercula are black, have the form of warts
and possess many unfeathered setae, which are about \^j^ mm. long.

Prothorax. No prothoracic shield, but there is an elevation between
the three following verrucae. This elevation, however, is not black
and the setae are as large as on the remaining surface of the body.

There are: v. dorsalis, r. subdorsalis, v. supraatig mails ^ placed
very high ; v. jirostigmalis, v. basalis, smaller v. pro2)edaUs and
r. postjiedalis.

84

Mesothorax and metathorax, V. dorsalis and in front of it a
verruca, which is not coloured black and bears setae, which
are about */<, mm. long. V. suhdorsalis, v. suprastigmalis, v. 2^f'o-
stigmalis, v. basalts.

Abdomen 1 — 9. V. dorsalis, v. subdorsalis, v. suprastigmalis
which is placed very low; v. jirostigmalis, ik infrastigmalis; v. ba-
salts very large. On the fore-part of the segment v. propedalis
is located. The proleg is coloured dark and bears setae of 1 mm. in
length. All the abdominal segments bear the same pattern, with
the understanding that segm. 1, 2, 7 and 8 bear a v. pedalis
as well.

Segm, 10. The arrangement in my specimens is not very clear.
I do not believe that there is a segment 11.

Instar IV. Length 33 mm. The whole body is now covered
with irregularly spread verrucae between which there are also
placed setae on the skin. I could not distinguish the primary
verrucae. The segment is now divided into four rings. The first
three especially bear the verrucae.

Instar V as IV.

Recapitulation. The third instar of the caterpillars bears
a regular pattern of warts. The setae are not feathered. Between
these verrucae are spread shorther setae. The pattern agrees with
that of other families. There is no v. dorsolateralis, but we also
find V. subdorsalis on the meso- and metathorax.

In the IVth instar no pattern is discernable. The number
of verrucae has increased considerably. The appearance of i\ sub-
dorsalis on the mesothorax and metathorax may perhaps be
ascribed to the same secondary augmentation of verrucae. An
examination of the first two instars may possibly give an expla-
nation of this fact.

Family Endromidae. Fracker (1915) does not describe this family.

Grote (1896) describes the verrucae as those of Bombyx mori
instar /, i. e. type I.

Packard (1905, p. 40) says that the fullgrown larva of En-

85

<//-o//i»> resiciilosa is smooth, without any hairs or only minute
ones. The ontogenesis is therefore probably like that of Boin-
hyx mori.

Family Boinhijcidae. The caudail horn does not resemble that of
the Cemtommimlae according to Packard (1905, p. 20).

This family is to be considered as originating from the Lasio-
rampidae.

Dyar (1896/>, p. 140) says that '^Bombij.r mori hjis true warts
of the typical lasiocampid pattern".

Grote (1896) pointed out, that the warts of instar / resemble
those of Endromis.

Sasaki (1898, p. 33 sqq.) says that the dorsal horn in instar /
is already a single median wart and this not only on B. mori^
but also on Theojihila mandarina^ which is considered to be the
primitive wild form.

Packard (1905, p. 40 sqq.) says that stage 1 has warts, later
on the body is smooth or with minute hairs.

The genus Ocitiam has a horn. (See i. a. Horsfield and Moore).

Fracker (1915, p. 102) says that the setae are so reduced as
to be of little value in identification.

In the literature the assertion is very often met with that Bomhyx
mori is naked. It is remarkable that this species, which has been
cultivated in such large quantities and is one of the few
insects which have become domesticated, has been observed so
insufficiently. Although indistinct, the old pattern of the verrucae
remains visible to the last moment of the larval stage.

Bombyx mori L., Plate II, fig. 1 — 5.

Material in alcohol at Groningen 1914, many specimens col-
lected every day, so that I had a very extensive collection at my
disposal.

Instar /. Duration 40? days. Length 4 mm. The setae are not
feathered, about 400 ^ long. They are only placed on tubercula.
Most of the swellings are verrucae with four or five setae, but
s. subdorsalis, s. poststigmalis^ s. infrastigmalis remain separately

86

visible. At about the time of the first moulting the vernicac dor-
sales of the abdominal segment 9 begin to stretch themselves,
and rise to the "caudal horn" which reminds us somewhat of
that of the Sphingidae.

Prothorax. Verruca dorsalis large, with four or five setae; seta
subdorsalL^, v. suprastigmalis^ not distinctly confined, but there
are three to six setae placed together; v. prostigmaliis with three
setae, v, basalis with three setae, s. propedalis^ s. postpedalis, some
small setae on the leg, and s. ventralis.

Mesothorax. V. dorsalis and v. dorsolateralis large, each with

five or six setae, s. suprastigmalis^ v. prostigmalis with four setae,

V. basalis with four setae. S. propedalis^ s. postpedalis, s. ventralis.

Metathorax = mesothorax'^ but sometimes there is a median

dorsal seta.

Abdomen 1,2. Here too this medial dorsal seta is sometimes found.

V. dorsalis large with four setae, s. subdorsalis, v. suprastig-

inalis large with four setae, s. poststigmalis, s. infrastigmalis,

V. basalis with four setae; two or three s. {pro- et 2^ost) pedales;

8. ventralis.

Segm. 3, 4, 5, 6 = 1 and 2, but v. propedaUs bears two setae, and
on the leg one finds at the front and the back edge two
pigmental spots with one or more small setae ; s. poststigmalis
bears very often two setae, but this is not a constant feature,
neither is it so on the successive segments of one individual.
„ 7, 8 = 1 and 2, but s. poststigmalis has often been doubled.
The newly-hatched larva directly bears a v. dorsalis on
segment 8, which is higher and larger than the other ver-
rucae. That it is a v. dorsalis appears from the s. sub-
dorsalis^ placed behind it. The two v. dorsales of the right
and left sides are united. When the length of the cater-
pillar is 4 mm. this verruca is 50 — 60 f/.. high. It soon
grows to the double height, whereas the length of the
caterpillar increases only very little.

When the caterpillar is a fortnight old, it is about 6 mm.
long and the v. dorsales have already attained a height of

87

about 500 11. In other words: while the caterpillar only
grows 50°/p, the r. dofsales, which soon become united,
have attained a length of ten times their original one.
SefftH. 9. \'. ilorsdliit is very large with four setae, t\ snprmti)jinali8
is reducinl to one seta; s. subdorsalis has disappeared, as
also have s. poststigmalis and s. infrastigmalis ; v. basaliti
with four setae; two setae on the ventral side I should like
to call .>«. proffedulix and s. vcntralis.
, 10. K. doraalis upon the anal flap with four setae, below
it a very large verruca, which has possibly taken origin
from a consolidation of r. suprastigmalis and v. basalts
(five setae); two it. projmiahs^ s. ventrales.
No 11 til Abdominal segment can be detected.
Int<tar //. Duration a fortnight. Length 8 mm.
On the whole the same verrucae as in instar /, but whilst
there they are distinctly confineil and have dark pigments, here they
consist of a small elevation of the surface, of the same colour
as the rest of the body. Upon this elevation stand a number of
setae, which are about 400 /z long and which therefore have
not grown. The whole surface is covered by irregularly spread
secondary setae, about 100 /x long and therefore distinctly con-
trasting to the primary ones.

Prothorax. Verruca dorsalis with five setae, s. subdorsalis^ s.
dorsolateralis, v. suprastigmalis with two setae, v. prosiigmalis^
with four setae, v. basalis with five setae, s. propedalis, s. post-
Ijedalis, s. ventralis.

Mesothorax. V. dx)rsalis with six setae, v. dorsolateralis with
four setae, r. suprastigmalis with three small setae or already to-
tally vanished, v. basalis with five setae, s. propedalis, s. post-
pedalis, s. ventralis.

Metathorax = tnesothorax, but v. suprastigmalis has almost dis-
appeared.

Abdomen 1. V. dorsalis with two to four setae, s. subdorsalis,
r. suprastigmalis with four setae, no v. or s. poststigmalis and
infrastigmalis, v. {pro)pedalis with four setae, s. ventralis.

88

Segm. 2=1, but s. subdorsalis is absent.

„ 3, 4, 5, 6 ^ 1 . but 8. subdorsalis is absent, v. infrastigmalis
often consists of one or two setae; and on the segments 3
and 4 there is often a v. poststigmalis with four setae;
V. basalts with six setae; the pigmental spots on the leg
have increased in size and bear many short setae.
. 7 = 1.

„ 8. The ''caudal horn" has become entirely median, and its
formation by the blending of the two v. dorsales is no
longer clearly visible. It is now 600 f^ high, perhaps s. sub-
dorsalis is united with it, else it has disappeared ; further
there occur v. suprastigmalis with three setae, v. basalis with
four setae, v. ventralis with two setae, of which the most
lateral oae is perhaps s. propedalis.
„ 9. V. dorsalis is formed by the union of the right and the left
one. It is now 300 f/,. high and resembles that of segment 8,
but is smaller ; no s. subdorsalis, s. suprastigmalis^ v. basalis
with four setae, s. propedalis^ s. ventralis (see abd. 8).
„ 10. F. dorsalis, with four setae; s. subdorsalis; one spot
with setae under the anal opening, probably agreeing with
V. basalis, s. propedalis and on the leg three s. pedales.
Instar ///. Length 9 mm.

In these and the following instars all verrucae and setae remain
entirely as during instar//. To prove this, I drew the abdominal
segment 5 of instar V, i. e. after the last moulting (Length 40 mm.).
Distinguishable are still : v. dorsalis, with three setae, placed imme-
diately beside that of the other side and behind this a black
spot; s. subdorsalis, v. suprasfigmalis with three setae; v. basalis
with five setae. Further the whole leg is covered with long setae ;
namely more than 1000 y. long, whereas the remaining primary
setae are about 500 jct., and the secondary ones about 100 [x.
So we see that a very insignificant growth of the primary setae
has taken place, namely of 400 — 500 (z., whereas the body has
grown from 4 to 40 mm. and during instar V to about 80 mm.

89

SHmtrtary.

Ill instair / distinct, black verrucae occur, most of them
with throe or four setae, some of the tubercles bear one seta
and are probably already reduced , as they sometimes have two
setae. Th«' pattern on the abdomen agrees with type I, on the
thorax with type II.

In instar // we see the secondary setae, and the verrucae are
juirtially dissolved, but they are still visible. This is the case during
all the instars of the caterpillars. The primary setae grow very
little, the secondary do not grow at all, except on the abdo-
minal legs.

The caudal horn is formed by the left and right v. dorsalea
of segment 8.

Family lirahmaeulae.

Packard (1905, p. 43) thinks that this family is the most
siK-cialized of the SYMBOMBYCINA. They have in instar /
multisetose warts, which they lose after the first moult, as is
clearly seen in B. japonica. Packard (1915, PI. XXXIV).

In the first instars there is a caudal horn, which later on dis-
appears. I had no material for investigation.

According to Packard (1893, 1905) the families we dis-
cussed just now have descended from the hairy Notodontidae, in
jjarticular from the Ichthyaritiae and Apntelodinae. Packard says
that in the first instar they all have warts which in a few cases
they retain, but nearly always lose after instar /. They also
agree in the pupae and imagines.

If this view is right, we have here a case of the development
of setae into verrucae which, however, are soon supplanted by
secondary setae. These can be very long (Lasiocampidae) but also
very short. {Bomhycidae).

As in the other group, the SYSSPHINGINA, a caudal horn is
developed on segment 8, but here it is soft and fleshy. And as
in the last mentioned group, this horn may disappear again, as is
the case with the Brahmaeidae.

90

Consequently a complete parallelism exists between these groups.
Of course another arrangement may also be conceived, in which
the "horned" ones are placed together.

In this case, however, difficulties would be met with in regard
to the warts.

The SYSSPHINGINA descend from forms (Packard 1 905, p. 44)
which bear tubercles not producing more than a single setae. The
primary groups of the Notodontidae viz. Notodontinae, Heterocam-
pinae and Cerurinae have already been discussed with the family.

Family Ceratocampidae. Two of the three volumes of Packard's
work have been devoted to this family. Of a great many cater-
pillars complete descriptions are given (1905 — 1914).

The caudal horn deserves to be specially mentioned. In Adelo-
cephala it has arisen [as is clearly visible in Packard's figure,
1905, PI. XLV, fig. 3] from the s. dorsalis of segment 8.
It ends in two setae, exactly as in the Sphingidae. On the meso-
thorax and the metathorax too, the s. dorsales have increased
enormously and are bifurcated. The ordinary setae change during
the following instars into spinose ones, whilst the caudal horn
loses the two setae at the end and then forms one single, purely
median projecting part.

Syssphinx and Fades have the same ontogenesis. In Ani-
sota a median dorsal horn is developed on the metathorax from
the s. dorsales and a median tuberculum on the 9th abdominal
segment apparently from the s. subdorsales. atheroma has many
setae developed into large scoli, as well on the thoracic as on the
last abdominal segments.

Except in the Ceratocampidae these scoli only occur on the
Hemileucidae, the Saturniidae, the Nymphalidae and the Heli-
coniidae. Packard thinks with Fracker (1915, p. 120 — 126) that
this is of great importance and I am inclined to agree with them. It
seems to me that the dorsal horn on the 8th abdominal segment
too is of great importance in judging of the relationship. The
number of the scoli varies greatly, secondary setae may be found
or not. See Fracker (1915, p. 121—123).

91

Family Hemileuddm. Packari)(1914, p. 77-151, PI. XX-XXXI).

Fracker (1915, p. 122). The setae are partly bifurcated and
8ituatiH] on very long and thin tubercula, some of them bear a
«pino8e character and are short and forked. The latter are developed
fronj the former e.g. Heniilcuca inaiu. (Packard, 1914, PI. XXII).

They are arranged as on the Saturniidae i. e. Type la. S. dor-
sitlis is still to be found on the abdomen of Pseudohasis eglanterina.

Family Saturnidae. Weismann (1876) drew attention to the
spots which on the larvae are different according to their places
of birth. I therefore thought it right to draw these spots in their
exact shape. Plate III, fig. 14, 15.

Packard (1914, PI. XXVI— XXXIII and p. 151—271). Tracker
(1915, p. 121 — 122). As early as in instar / the larvae have
verrucae which later on become scoli. Secondary setae sometimes
make the arrangement a little indistinct. Generally the scoli
consist of a conical tuberculum on the top of which some (2 — 10)
setae take their origin. The setae often end in a knob and they
are glandular hairs. In the successive moults the number of setae
often diminishes. On abdominal segment 8 the sc. dorsales are
placed close together. Naked forms are also found, but the first
instars have the above mentioned scoli e. g. Rhodia fugax (Packard
1914, PI. XXVIII— XXIX). — PouLTON (1890) mentions depressed
scars on the pupa of Satuniia carpina. To me it seems that the Hemi-
leucidae and the Satinitidae are side-branches, not directly connected
with the Sphingidae. The description of Saturnia pavonia follows.

Saturnia pavonia. Plate III, fig. 14, 15.

Material in alcohol, cultivated at Groningen in the summer of 1915.

Instar /. Length 3'/2 m™- Head black. The tubercula are
warts, mostly with 5 — 7 not plumed setae, which have a length
of ± 700 fji.. On the skin there are no other setae.

Prothorax. V, dorsalis, seta subdorsalis, v. suprastigmalis, s. pro-
stigmalis^ v. hasalis with two setae.

Mesothorax and Metathorax. V. dorsalis, v. suprastigtnalis,
V. poststiginalis, v. hasalis with two setae, s. dorsolateralis.

92

Abdomen 1, 2. V. dorsalis is the largest of the tuberciila; no
V. subdorsalis, v. suprastigmalis a little smaller than the other
verrucae, v. infrastigmalis lying a little further towards the tail
than usually. It may be that it has been derived from v. post-
stigmalis and v. infrastigmalis. Two s. basales close to each other
and sometimes blended. Generally no s. pedalis.
Segm. 3, 4, 5, 6 = 1. But v. suprastigmalis is larger and v. hasalis
bears four setae, next to each other. On the leg is a dark
spot and quite at the end is a seta pedalis.
„ 7, 8 = 1. But V. suprastigmalis is about as large as the
others, v. hasalis is well developed and there is a d. pedalis
with two or three setae.
„ 9 = 1. But V. suprastigmalis is not clearly defined and
generally bears only one long seta and two or three short ones.
V. hasalis has two setae which are sometimes separated. No
distinct segment 10.
Instar // and the following instars.

All the tubercula are arranged as in instar / and grow into
scoli, but V. hasalis is dissolved into several setae, v. pedalis dis-
appears. Between the verrucae setae appear on the skin directly
after the first moulting. When the caterpillar is full-grown they
are ± '/i °^™* l^^g? (on the abdominal legs a little longer). The
setae on the verrucae are + 1 '/.^ mm., exceedingly thick and not
plumed. Therefore three scoli occur on each segment : sc. dors(dis,
sc. suprastigmalis, sc. infrastigmalis. Neither is the s. subdorsalis
of the prothorax any longer visible.

A dark colour in the drawing of abdominal segment 5 (Plate III,
fig. 15) accurately shows the spots. They do not take their origin
from the primary verrucae.

Recapitulation. The pattern hardly changes. The tubercula
are scoli. There is no sc. subdorsalis except a very much reduced
one on the prothorax. Probably it has also disappeared on the other
segments and for this I refer to the Hemileucidae. On the meso-
thorax and metathorax there is no sc. dorsolateralis. Sc. basalts
becomes dissolved later on.

93

Between the 8Coli we see in instar II small setae. In instar 7
no t«pot8, but in instar II we find spots along the front and
back edges of the segments, they are connected with each other
by a line above the stigma. T^ater on they are irregular.

Family Sphingidae.

Weismann's important study (1876) has been mentioned before
in detail. Chapter II, p. G sqq.

Mkldola translated this work in 1882 and drew attention to the
resemblance with the flagellate organ Papilio and Dicranura possess.

W. MOller (1886, p. 250) relates among other things that
the horn is composed of the two primary hairs 1 = Ds. of tlie
8th abdominal segment and gives a drawing of it.

Poi'LTON (1884, 1888, 1890) gave his attention to the colours
of the pupae which agree with those of the caterpillars. He thinks
that the horn arises from two tubercula. I intend to discuss
Poi'LTOx's opinion on the colours in the chapter on pupae.

Dr. Jur. M. C. Piepers published in 1889 a treatise in Dutch
and in a more extensive form in German in 1897. Piepers, who lived
for many years in the Dutch East-Indies, raised numerous Sphin-
gidae from the e^^. He found that the horn is very movable.
[J. Th. OrDEMANs discovered the particular muscles which cause
this movement]. Then he discusses the hypothesis of Th. Goossens
(1873) who thought that the horn serves to protect the glands
by which the urine is secreted. Piepers thinks it is used to
drive away the Ichneumonidae and Tachinhiae. The granulation is
the consequence of a secondary disappearance of prickles so that
the organ may have been a poisonous prickle formerly.

Piepers found also that in the beginning the horn possesses
two points.

Packard (1890, p. 513) is remarkably short in treating this
organ, but in 1905 he discusses it at length. With it he draws
figures of Ceteromia amyntor. His PI. XXXIV and XLII agree
almost entirely with mine of Sphinx Ugustri, in so far as there
are no secondary setae except on the horn, and with Smerinthiis

94

spec, as these secondary spinules are glandular setae which are
bifurcated. In 1905 (p. 17 — 21 and p. 34—45) Packard worked
out the relationship between the Ceratocampidae and the Sphingidae.
He believed to have found a proof in the horn, which is built in
a similar way, on the 8th abdominal segment.

Dyar (1894, p. 204) says only that on the 8th abdominal
segment there is the dorsal tuberculum and draws attention to
the position of IV and V. Tuberculum IV is placed under the
stigma, tuberculum V in front of it. Packard (1905) follows him
here, as Fracker does in 1915.

I think there is not a single proof of a rotation of 90 degrees like
that, round the stigma and I consider IV of Dyar to be in this case
s. infrastigmalis (i. e. V) and his V as a s. 2irostigmalis i. e. as III B
of Quail. Further I draw attention to the peculiar secondary setae.

I was not able to obtain Forbes' work of 1911.

Sphinx ligmtri.

Material in alcohol. July 1915. Plate IV, figure 1.

Instar /. Length 6 mm.

Prothorax. There are : s. dorsalis, s. suhdorsalis^ s. suprastig-
malis, s. dorsolateralis, two s.prostigmales, two s. basales, some small
s. pedales. Beneath s. siibdorsalis, at the same height as s. dorso-
lateralis, is a s. stibdorsalis inferior.

Mesothorax. We find two s. dorsales, perhaps arisen from the
blending of s. dorsalis and s. subdorsalis; s. dorsolateralis united
with s. suprastigmalis, one s. prostigmalis in front of a wing
rudiment, which at first I thought to be a rudimentary stigma,
two s. bascdes, some s. p>edales which are small.

Metathorax == Mesothorax but s. dorsolateralis is simple, and
the wing-rudiment is not distinct.

Abdomen 1, 2. S. dorsalis, s. subdorsalis, s. suprastig mails, s.
jirostigmiilis, s. infrastigmalis, s. basalis and ventral of it two
setae probably agreeing with the s. pedalis and s. ventralis.

Segm. 3, 4, 5, 6 = 1. There is a s. propedalis and a s. ventralis.
„ 7 = 1 but only one s. projjedalis.

95

Segm. 8. On the place of the s. donutlc^t of the left and right
sides, we find the median caudal horn which is white directly
after the birth and gets hluck later on. The skin has totally
changeil on this spot and is covertnl with numerous, irre-
gularly plaeetl small setae.

I do not think that the caudal horn has exclusively been
developed from the two s. donalrs^ which are distinctly
to be seen on the top, as two knobs each with a seta,
but I am incline<l to consider it as a protubcnuue which
has lifted up the «. dorsaUs on its top.

Besides the horn we find in the usual order: a. sub-
doi'saiiSy ». siHjtrdiftif/inalis, n. prostif/malis, a. infrastujmalis^
s. bfwalis, and a. proitrdalin.
„ 9 = 1. But s. infrastigmalis is absent, there is only one

8. boMiis and one ». propedalis.
„ 10. iS>. dorsaiis, s. subdorsalis and s. suprustujmnUs on the
anal-flap; there is a s. prostiynialix, s. infrastigmalis is
absent, one s. i>ed(ili<i.
„ 11. Just under the anal-flap a large tuberculum with setae.
The seta under it we might take aas. basalis of segment 10,
but in connection with the reduced number of setae of
segment 9, this does not seem probable to me. I consider this
, tuberculum to belong to the ventral part of 11.
Instar // and the following instars.

The primary tubercula and setae have disappeared. The skin
is totally covered with small tubercula placed in vertical rows,
each of which has a very short seta.

Recapitulation. Instar / has a primitive setal-pattern, which
is nearly the same as that of the other families. The caudal horn
has arisen on the place of the left and right setae dorsales which
remain visible on the top. "With some exceptions (meso- and meta-
thorax) which it is easy to explain, the tubercula bear one seta.

There is no s. poststigmalis, but we find on the abdominal
segment a 8. jjrostignialis. In instar II this pattern is lost and

96

then the caterpillar acquires secondary setae which are so short
and homogeneous that it seems to be naked.

Smerinthus tiliae.

Material in alcohol. July 1915. Plate IV, Fig. 2a, h.

Instar /. Length 3 mm.

The head is very large. Besides the tubercula described below,
the whole body and the caudal horn too, are covered with irre-
gularly plUced small setae, the length of which is + 50 ^a. With
a low power it looks as if they are bifurcated at the end as has
been figured by Sharp (II, p. 359) for instar / of Etichloe carda-
mines and by Packard (1905) for that of Ceteromia amyntor.
Highly magnified they appear to consist of a couple of bigger
rays and a number of thinner ones, all protuding from the
upper part of the seta. The whole thing might be compared
with an umbrella turned upside down of which some of the ribs
are thicker than the other. Plate IV, fig. 2a.

The primary tubercula are not black in reality : the black in
the figure is only intended to draw attention to them.

Prothorax. There are s. dorsalis, s. subdorsalis, s. supmstigmalis
or s. dorsolatendis, as it stands a little lower than one might expect
at first sight, s. prostigmalis and s. in frastig mails , two s. basales
placed next to each other, some smaller s. pedales.

Mesothorax and Metathorax. S. dorsalis, s. suprastigmalis or
s. dorsolateralis, s. j^^'ostigmalis, two s. basales, a few smaller
s. pedales.

Abdomen 1. S. dorsalis, s. subdorsalis, s. suprastigmalis, s.pro-
stigmalis, s. in frastig mails, two s. basales, one of which agrees
with s. 2)edalis.
Segm. 2 = 1. One s. basalls.

„ 3, 4, 5, 6. 7 = 2 i. e. one s. basalis and one s. iiedalis.

„ 8. We must consider the caudal horn as in the case of
Sphinx ligiistri. The skin is continued in unaltered con-
dition over the horn and bears the same umbrella-shaped
setae, for the rest as with 2.

97

Se^m. 9 = 2, but 8. infrastiginalis fails.

„ 10. S. dorsalis, a. sufhiontaluf and s. HNpntstigmalis on the
anal Hap, under it s. prmtigtiMlis; %, infrastigmalis is ab-
sent, one «. peditlis.
„ 11. Behind the anal flap is one tuherculuin helonj^inu: to

the dorsal part of 8(>gincnt 11.

Instar //. The pattern described above is totally absent here. The

skin is covered with the tubercula describwl for Sphinx ligustn.

Recapitulation. The setal pattern of instar / quite agrees

with that of Sphiiw lufustn^ only the a. tlormlaternles or the

». Huprfuitigmdhs on the thoracic st^gments are wanting.

The caudal horn just as in Spkit$j' lii/ustrl originates by a
niiHiian protuberance t»f aM. wgm. 8 under the left and right

lk«tween thest* primary tubercula we find small umbrella-shaptnl
setae.

In instar // this whole {lattern has disappeared just as the par-
ticular setae which have been replaced by small ones of the
ordinary form.

Smerinthus populi Linn.

Material on alcohol. July 1914. Plate IV, fig. 3 a, b.

Instar /. I^ength 4 mm.

The whole surface is covered with the small umbrella-shaped
setae, which I described for Stnerinthus tUiae. The primary
tubercula and setae are wanting except:

Prothorfix. S. in frost ignialis^ one s. basalts.

Mesothorax and metathorax. One s. hasalis.

Abdomen 2. S. in frastig malts ^ s. basalis^ s. pedalit.
Segm. 3, 4, 5, 6. S. basalts, s. pedalis.

„ 8. The caudal horn bears no other setae but the umbrella-
shaped ones, and is very short. Further one s. hasalis,
just as on 9 and 10.

Instar II. We only find the small tubercula and setae described
for S)nennfhus tiliae and Sphinx ligiistn.

7

98

Recapitulation. The homogeneous setae together with the
umbrella-shaped ones, which in Sinerinthns tiliae are found by
the side of the primary pattern, have almost entirely replaced
the old pattern in instar /. In instar // it is replaced by short setae.

Pterogon proserpina Pall, and P. c/orgoniades Hb. in the col-
lection Kallenbach are homogeneous in the two last stages and
are covered with thin and small setae.

Macroglossa stellatarum L, M. croatica Esp., Hemaris scabiosae
Z. and H. fuciformis which are all in the collection Kall. as
the last instar of caterpillars, are naked.

Family Lithosiidae. The setae and verrucae are arranged accord-
ing to type I, the same as in Arctiidcie.

Fracker (1915, p. 118) saw no other species than those with
setae instead of verrucae. This family is therefore probably rather
primitive. Uho on the abdomen and Pi on the mesothorax and
metathorax are double and from this Fracker concludes the
reduction of the verrucae to setae.

The collection Kall. had no material for investigation.

Family Arctiidae. In this family the arrangement of the ver-
rucae is very distinct, so that Dyar (1894) gave the name of
Arctian type to an arrangement which almost completely agrees
with Type I.

Fracker (1915, p. 114 — 118) gives a table of genera with
which he himself is not satisfied.

This writer thinks that the genera Doa and Utetheisa are reduced,
as they only bear setae and no verrucae. I might add Hipocrita
{Euchelia) but think that this condition should rather be considered
as something primitive.

To prove his opinion he says (discussing the Noctuidae, 1. c, p. 113)
that Doa possesses a multisetiferous leg-plate and that the Pi-
group on the metathorax of Utetheisa is bisetose.

I am not convinced by these arguments, as the setae on the
legs of primitive caterpillars are often numerous and the thorax
in general often bears two s. basales.

99

Tftpocritu (Kuchelia) jitcohaea Linn.

Mattriiil 1915 (thi' Hague) and roll. Kaij,. I'latt- IV, fig. 7.

Tilt' Hpociinens colltH^ttHl by me in tlie middle of April measured
10 mm. I did not succeetl in cultivating them, therefore I do nut
know to which instar they belonged. The arrangement of the
setae of theue specimens is the same as of those of coll. Kall.
The full-grown specimens are 35 mm.

Prothoi'djr. S. dot-inilis, s. suhdorsalisj .s. prostiyimdis^ two ». basales
on one tuberculum, .<«. proprdaliSy «. postpedalis.

Mesothorar and Metathorojc. S. dormlis^ a. dorsoldtendis, 8. supra-
ittigmalis, s. prostitjtnidis and on the place of tlic wing-rudiment
one seta. One might be tempteil to look for the rudimentary
stigma on this spot, as in this place the air-tubes arise from the
main trachea. Two s, basaUs^ s. propedalus^ 8. postpedalis.

Abdomen 1, 2. <Si. dorstdis, ». suMorxalis^ n. supraatignmlis^
8. po8t8titjmidi8, 8. infru^tiijimduf^ two m. htutalijt on one tuber-
culum, 8. ptdidi8.

Segm. 3, 4, 5, 6=1, but instead of s. pedalis there is a seta
on the front side of the leg ^ 8. propedalis.

r, 7,8,9 = 1.

, 10 = 1, but somewhat reduced, for s. poststigmalis and
8. infrastigmalis are absent. On the anal legs three setae.

Pupa. Coll. Kall. no setae.

Ardia caja Linn. Plate IV, fig. 4, 5 a, b.

Material on alcohol. Groningen 1914.

Instar /. Length 6 — 7 mm. The head is black, the tubercula
are brown, they soon begin to colour. The tubercula are warts,
the setae are plumed.

Prothorax. There is no prothoracic shield, but the two v. dorsales
of the left and right sides approach one another and they have
progressed a great deal over the segment in ventral direction. V. suh-
dorsalis is very small. V. suprastigmalis is large and situated
a little lower than usually is the case in other families of cater-

100

pillars. There is also a large i\ hasalis^ and some non-plumed
setae on the leg.

Mesothorax and metathorax. We find v. dorsalis, v. suprastig-
malis and v. prostiymalis which are about equally large. The
last mentioned wart lies on the same height as the very high
placed V. poststigmalis on the abdominal segments. V. dorsolate-
ralis is wanting.

Behind v. prostif/tnalis is a small tuberculum with three or four
setae. It lies therefore on the place of the wing-rudiment. Besides
we find in many individuals a tuberculum with one or two setae.

V. bdsalis is large. On the leg occur some setae which are not
feathered.

Abdomen 1, 2. V. dor.taJis is a little smaller than v. auhdor-
mlis^ V. Huprmfif/malis is large. Large and far projcfting is v.
jiosfsfu/nKdis which is partly continued under the stigma towards
the head. V. infrmtigmalis is situated lower than in most of tiie
other families of caterpillars. V. basalts and v. i^edalis on the
place where on the abdominal segments the leg is fixed.

Segm. 3, 4, 5, 6=1, except that v. pedalis is absent. But on
the leg are one or two strong setae.

„ 7, 8 = 1.

„ 9. Very small v. dorsalis, very large v. siibdorsalis which
has far extended in the direction of the missing v. suprastigmalis,
this being perhaps united with it. For the rest as 3. Moreover
there is a rather large tuberculum under the anal flap. I con-
sider this to be v. infrastigmalis of segm. 10.

Consequently the 10th segment is very much reduced.

Instar //. Duration 30 days. Length 12 mm.

The whole animal looks exactly as in instar /, but between
the tubercula a few setae (generally not plumed) are found.

Prothorax. V. dorsalis is still larger than in instar I and v.
subdorsalis too has grown longer. Under it is a new verruca.
For the rest as in instar I.

On the abdominal segments we see some small tubercula in
the neighbourhood of v. infrastigmalis. V. poststigmalis now grows

101

uearly entirely under the stigma tiu that the following warts are
lying almost in one line, which also passes through the stigiim
and the middle of the leg. V. (hrsatis^ r. supnistigtnalis^ v, post-
stigmalisy v. infrastigmalis and p. hasalis.

The other instars remain unchanged.

Recapitulation. The wart-shaped tubercula bear in all in-
stars plumed setae. The pattern (Type I) is very simple and
hardly changes during the different instars. All the different
tubercula are about equally large.

There is no v. dorsolateral is. V. poststujmalU is shifted a little
under the stigma, r. infrastigmaliH is situated on a low level.

Spiloitoma {Ocnogyna) lubricipeda Linn. Plate IV, fig. 6.

Entirely as Arctia ciya^ but in the beginning the setae are
not feathennl. V. poststigHtalis is also shifted from under the
stigma. The setae on the leg are very densily feathered. On the
ventral side mediad of the leg a tuberculum without setae = v.
centralis occurs.

In instar IV a linea dorsalis is present.

Family Syntomidae. The verrucae completely agree with type I,
but there is only one verruca on the mesothorax and metatho-
rax over r. prostigmalis. This only occurs in the Pericopidae.
The veri^cae change a great deal in form, size and number.
Fracker (1915, p. 118) found that segment 7 has the same pat-
tern as the other abdominal segments. The setae are plumed and
often form pencils. Mostly secondary setae.

Coll. Kall. no material.

Family NoUdae. Fracker puts this family (1915, p. 98) with
the Lacosoinidae as Microlepidoptera of uncertain position after
the ZYGAENOID series. The caterpillars bear verrucae which
remind us of those of the Arctiidae.

Family Agm-istidae. Fracker (1915, p. 114) examined different
kinds and comes to the conclusion, that it is right to unite this
family with the Noctuidae. This harmonizes with the fact that
Handlirsch (1 908) places these families close together.

102

Coll. Kall. no material.

Family Noduidae. Whilst Packard (1895, p. 83) entirely sepa-
rated this family from the Bomhyces^ he says (1905, p. 41)
"that in the SYMBOMBYCINAE the noetuiform characters are
crowded back in the phylogeny of the group."

Dyar (1899) describes some Hijdroeciae and proves the exis-
tence of a large «. prostigtnalis.

Fracker (1915) devotes some pages (p. 111—118) to the owlet-
moths and describes the difficulties he met with.

He divides the family into four groups, "they are, however,
purely for convenience and do not constitute a natural arrange-
ment." His groups are:

/. Larvae with primary setae only. The pattern agrees with
that of Mamestra and Bepressaria (see below).

To this not a single Aa'onycta species belongs, otherwise the
greater part of the genera.

II. Larvae with well developed verrucae, arranged as in Arctiidae
but X (r. prostigmalis) is often small and on segment 7 it stands on a
lower level than on 6 and 8. To these belong the Acronyctinae in part.

///. Larvae with verrucae which are obscured by the deve-
lopment of secondary setae. Acronycta in part.

IV. Verrucae reduced to single setae, although proceeded by well
developed tufts in earlier stages ; f^ {=v. basalts) remains present
as verruca, which is very peculiar as a proof of the pseudo-primitive
character. Acronycta in part.

Fracker (1915, PI. Ill and lY) also describes Feltia glandaria
instar / and full-fed. We find here the following peculiarities:

a. instar / has no s. infrastigmalis (>) on the prothorax, though
we do find it on the full-grown form.

b. instar / has no s. i^rostigmalis {s) on the abdomen, but gets
one in maturity.

In the last case, supposing it really has been stated rightly, I
presume that Fracker's labeling is not correct. For then the
seta on the abdomen is either subprimary or secondary and on
the thorax he calls it a primary seta, whilst he s^ys on p. 21

103

that it is not allowed to honiologize a primary with a secondary
seta. The many caterpillars in the coll. Kall. agree with Fracker's
groups. As examples of the 4th group 1 mention Acroni/cta ahii,
where the s. snlidorsales reach a great length and Diloba cocni-
locephala L.

I should like to propose a 5th group viz.:

V. The setae or the verrucac have disappeared, the pigmental
spots on the tubercula remain, e.g. Nonayvia {Deptrsmria) nervosa.
Coll. Kall. PI. IV, fig. 10.

Mamestra binssicae Linn.

This species belongs to group /. Plate IV, fig. 8 a, b.

Prothorax. S. dorsalis^ s. subdorsalis, s. supnistigmalis^ s. pro-
stif/imilis, three s. basales, s. projfedalis, s. postpedalis, s. ventralis.

Mesothonur and Metathorax. S. dorsalis, s. dorsolateralia, s. su-
prttstifjiiKiliit^ .<. prostigitialis, s. jH)sMigmalis, two s. basales, s.pro-
prdnlis^ s. postpedalis, s. rentr(dis.

Abdomett. 1 — 9. S. dorsalis, s. dorsolateralis, s. supntstigmalis^
s. prostigmtilis (very small); s. poststiymalis^ s. infraMujmalis,
two s. basales^ s. propedalis, s. ventralis, some setae pedales
which are also developed in 1, 2, 7, 8, 9.

It is therefore quite like type I.

Acronycta pai Linn. Plate IV, fig. 9 a, b.

I only want to draw attention to the following segments which,
side by side of the verrucae I am going to mention, also have
secondary setae, especially on the ventral half.

Metathorcur. V. dorsalis^ v. dorsolateralis, v. suprastigtnalis, some
*\ prosiigmales, v. basalis, some s. propedales.

Abdamen I. V. dorsalis has grown enormously and has blen-
ded with that of the other side to a fleshy stump. V. subdorsalis
small, but still recognisable, v. suprastigmalis, s. prostigmalis^
s. basales^ v. pedalis.

Segm. 2=1, but without a horn, v. dorsalis is smaller than v.
subdorsali§.

„ 8. V. dorsalis, i\ subdorsalis, (united with that of the other

104

side to a fleshy horn) v. suprastigmalis, v. prostigmalis,
two s. poststigmales, some s. basales.

The caudal horn has in this case taken origin from v. subdorsalis^
in the Sjihingidae on the contrary from r. dorsalis, this is a mor-
phological proof of the fact that these horns are not homologous.

Family Brephidae. Fracker (1915, p. 101) ranges Brephos etc.
with the Geometridae. The presence of the first three pairs of
ventral legs, even if they have become rudimentary, seems to
me to be of sufficient importance to make this group a family apart.

Family Epiplemidae. Fracker (1915, p. 100) gives a normal
setal type, with the s. poststigmalis and s. infrasiigmalis placed close
together on some segments.

Family Geometridae. Fracker (1915) only says that this family
of the loopers or inch-worms differs very much in armature, setae
etc. In the coll. Kall. there are only full-grown specimens (e. g.
Abraxas grossulariata, Amphidasis betidaria), to my regret, as I
should like to examine first instars.

The setae are placed on the mesothorax and metathorax accor-
ding to type II, on the abdomen according to type I. There is,
however, a s. subdorsalis inferior and s. suprastigmalis is placed
a little more caudal than usually.

Family CymatopJioridae. Spuler figures these caterpillars as quite
naked, but probably small setae are present. Fracker (1915) unites
the three following families, founded on the investigations of
BuscK and Walsingham, to the PYRALOID series of the MICRO-
LEPIDOPTERA-NONACULEATA ; Handlirsch (1908) on the
other hand puts these families just in front of the Thyrididae
and Hesperidae^ as the nearest relatives of the Papilionidae s. 1.

I think that the verrucae which do not occur in other Micro-
lepidoptera, point to a considerable difference, though I must
acknowledge that in several families verrucae arise from setae.

Family Pyralidae.

Dyar (1894) described the Pyralidina as Generalized Frenatae
with the tubercula IV and V (i. e. s. poststigmglis and s. infra-
siigmalis) approximated and single haired.

105

In 1895 he referred to it again and added: I and II i.e.
s. dorsaliSj s. subdorsalis remote, (opposed to I and II consolidated:
A rtkroeerina).

Chapman (1896) thought that the Pyraloids belong to Tineina
with obtect pupae.

lIoFMANN (1898) thought that the Pyralidae like the Tineidae
etc. remain primary.

Frackkr (1915) says: Kappn is bisetose on prothorax and there
is a close association of K and >f on the abdomen ; (i. e. s. post'
stigiitaliix and 8. iHfmstiyitialis). With the different genera there
is some dissimilarity in Pi (». haaales), but further the distinction
on p. 87 — 94 is made by other characteristics than the setae.

Family Orneodidae. Kappa- and Pi-groups bisetose (Fracker
1915, p. 94), for the rest as the former family.

Family Pteropkondae.

Dyar (1894) states that tuberculum I = v. dorsalis is absent,
and the tubercula are many-haired.

In 1895 he says that I and II are consolidated.

Chapman (1896) figures some Pterophondae, but adds that the
hairs have been represented rather too diagranimatically. He says
on p. 135: "As regards its panoply of hairs, spines, bristles and
other appendages, the different species of Pterophortis present
immense variety, some being very smooth and plain and with a
delicate shell, others most elaborately clothed with hairs and spines
of various arrangements."

Quail (1904) says that the trapezoid tubercles (s. dorsales and
*'. subdorsales) do not only occur on the abdomen, but also on the
thorax, just as in the Hepialidae.

Fracrer (1915) does not mention any literature and says on
p. 94 "that the prolegs are long and stemlike. No other caterpil-
lars possessing verrucae and secondary setae have prolegs of this
shape, although a few lower micros with primary setae show
similar structures. All of the latter, however, have a trisetose
Kappa-gTOVL^ on the prothorax, while that of the Pterophoridae
is bisetose as in other Pyraloidae^\

106

0. HoFMANN (1898) wrote au excellent article on this family,
with twelve figures.

Having no material for independent investigation, I quote
HoFMANN as follows — from which we see that in this well
defined, natural family much difference occurs in the pattern,
though it shows the fundamental type I. —

Taeniocainpa gothica L. (I.e. fig. 1, 2) has on the abdomen:

s. dorsalis (I), s. suhdorsalis (II), s. suprmtigmulis (III), s. post-
sjigmulis (IV), s. infmstigmalis (V), s. pedalis (VII) and on the
mesothorax : I — IV in an oblique row as mentioned before. I pro-
bably would not label them in this way. I take V as s. prostig-
inalis, VI then is s. infrastigmalis^ VII = s. pedalis.

Euaiemidophonis rhododactylus S. V., (1. c. fig. 4) has on tlfe
abdomen :

8. dorsalis, s. suhdorsalis, s. suprastigmalis, s. infrastigmalis,
s. poststigmalis, s. basalts, two s. pedales.

Platyptilia gonadactyla S. V. (1. c. fig. 8) almost entirely agrees
with it, but here, as Dyar would have it for all Pterophoridae,
s. infrastigmalis is placed on the same tuberculum as s. p>oststig-
inalis. Between these setae and the two s. pedales there are three
s. basales of which two are placed higher than the third.

Pterophorus monodactylm L. (1. c. fig. 12) resembles it very
closely but VI (i. e. s. hasalis) is absent.

Leioptilns carphodactylus Hb. var. buphthalmi Hfm. (1. c. fig. 6)
agrees with Leioptilns distinctus H. S. (1. c. fig. 9), through the
possession of a secondary seta over the stigma. Therefore we
might describe this pattern: s. dorsalis, s. infrastigmalis, s. suh-
dorsalis sup. and inf. The last-mentioned species has, moreover,
one more seta under the stigma, whilst IV and V are not united.

Aciptilia tetradactyla L. (1. c. fig. 5) has verrucae, but a secon-
dary seta suhdorsalis inf.

On Oxyptilus leonari Stange (1. c. fig. 7) on the other hand, I
and II are united to one verruca.

Platyptilia gonodactyla S. V. (I.e. fig. 11) bears on the me-
sothorax: I -f II or according to my view two s. dorsales;

107

III + IV or two «. suprustitjuMles^ )>erhap8 «. dursulattralis is united
with I or III, V is doubltHl with a secondary seta = s. prostig-
malin, moreover a secondary seta over the wing-rudiment; two
s. bwiales, s. itednlis.

StenoptUin jtrliiioiiactyla Stein (I.e. fig. 10) agrees with it in
so far as I + II together form a verruca, as also do III and IV.
There is mortH)ver a secondary verrucu over the wing-rudiment.

I do not understand why we are to accept a doubling for V
and a consolidation for I + II.

In both cases there are two sotac in the place where usually
there is one; if one couple is interpreted as a consolidation,
the same ought to be done for the other.

Family TkyndUiae.

The place of this family seems to be very uncertain. Packakd
considered it in 1895 as a very primitive side-branch of the
SrohpUlopteru far from the BomhycitM (p. 83) but in 1905 (p. 46)
he could derive the Xutodontulae directly from them.

Sharp places it (1901, II) between the Meijalopygulae and the
Lasiocampidae which last group he sharply separates from the
Bombyces^ whilst Fracker (1915) ranges it between the Liparidae
and the Xotodontitlae. 1{ani)LIR.sch thinks that its place is just in
front of Hesperid<ie and PapUionidae s. 1. together with the Pyra-
lidae, Pterophoridae and 0nte4>nidae.

This short account which might be enlarged a great deal, shows
at least that the family has a rather primitive character.

In the coll. Kall. is a beautiful, full-grown specimen of Thy-
ris fenestrella of which the drawing is to be found on PI. IV,
fig. 11, 12, 13.

Thyiis fenestrella Linn.

Prothorcur. A paired prothoracic shield with s. dorsalis, s. sub-
dorsalis, s. suprastiymalis, two s. prostig males ^ of which the last is
perhaps s. infrastigninlis^ two s. basales on one tubercle, s. prope-
dalis, s. postpedalis and one median s. ventralis.

Mesothorax and nietathorax. S. dorsalis, s. dorsolateralis together
with s. siqn-a^figtnalis on one tubercle, on the place of s. pro-

108

stigmalis a black spot without a seta, s. poststif/malis, a. infra-
stigmalis, two s. basales on one tubercle, s. propedalis, s. postpedalis
or s. ventralis.

On the metathorax a small seta only on the left side on the place for
s. subdorsalis occurs. Where there are no other secondary setae
on the caterpillar, I think, I may conclude that s. subdorsalis in
other cases has disappeared from the thorax. A remnant like
that might be expected in the first place in a primitive animal.
On the mesothorax is a black spot in front of the s. basales.

Abdomen 1. »S'. dorsalis^ s. subdorsalis, s. suprastig mails, s. in-
frastigmalis, two s. basales on one tubercle, between these and
the others a black spot without a seta, s. pedalis, s. ventralis.
(In connection with the pattern of the Tineidae it is not quite
improbable that the seta under the stigma agrees with s. post-
stigmalis and that the black spot represents the vanished s.infra-
stigmalis. T prefer, however, not to bring hypothetical suppositions
like these to expression in nomenclature).

8egm. 2 = 1, but between the s. basales and s. pedalis is another
seta = s. propedalis (?)
„ 3, 4 = 2, but the black spot under s. infrastigmalis is wan-
ting and next to the already mentioned s. propedalis we
find on the base of the leg two s. pedales and also
s. ventralis.
„ 5, 6 = 3, but s. propedalis of segm. 2 is double, the s. pedales
are placed on the outside of the log, and s. basalis is
not doubled.
„ 7 = 2, but s. projyedalis is double and s. basalis single.
„ 8 = 2, but s. basalis is single.

„ 9 = 2, but the s. dorsales and the s. subdorsales from the
left and right sides are situated on a median dorsal shield,
s. basalis is single.
„ 10. On the anal shield we find s. dorsalis, s. subdorsalis,
s. suprastigmalis from the left and right sides, further
s. infrastigmalis, s. basalis, s. pedalis, s. ventralis are present.
There is no trace of an 11th segment.

109

Family AeyenUhe (= Sesiultw). This family shows a certain
relationship with the Sphingidae.

In Bkitkn'mCller's enormous work (1900) Dyar has described
the caterpillars. Fkackek (1915) ad«>pt8 those results with some
modifications. Dyar says (p. 228) that all the tubercles are single,
there never being any development of warts or secondary hairs.
The abdomen has type I with s. posMifftnalis and s. infrastiij-
malts united.

It is the same on the prothorax^ but there is a s. guhUorsalis
inf.j and rerruca aupraatigntalis and r. prostitjnitdis ea«'h bear
thrt« setae, the two s. Inualea are on one tuberculuin.

Mr!»)thorax and Mrttithorwc according to Dyar: ''I^ and !'• uni-
ted, Ib» and III* likewitu>, but IV and V well separated, IV being
even nearer to III than to V, a curious circumstance''.

I think there are : .<». donalii -\- s. donolateraliit, h. suprastiy-
malis -f -"*' pt'ostiiftnaliSy 8. posWUjinalU^ s. infrttstigmalis, 8. basalis,
8. fwdalis.

Khopaloeera.

Fracker believes that the butterflies arose from the Microlepi-
iioptera in a time when x and )f had not yet become adjacent.

Family Hesperidae. No material in the coll. Kall.

Fracker (1915, p. 127) follows the descriptions of Scudder
(1889) not only in this family but in all the Bhoj)alocera.

This family deviates a great deal from the others which Kand-
LiRSCH takes together under the name of Papilionidae s. 1.

Fracker says : "Secondary setae numerous, small flattened plates
sometimes present, possibly showing position of primary setae,
setae on the head often plumose."

Family Megathymidae with the last-mentioned one united to
the Hespen'oidae.

Fracker (1915, p. 128) says, "no setae on dorsal half, nume-
rous on ventral half of the first two thoracic segments, rare or
absent on abdomen except prolegs."

Family Lycaenidae. These larvae resemble ZYGAENOIDEA a
little (Fracker 1915, p. 128).

no

There are many secondary setae, sometimes in tufts or pencils.

It seems to me that as full-fed larvae most kinds possess a
homogeneous distribution of the setae.

The verrucae of Tkestor halhis F. (coll. Kall.) I consider to
be r. dorsalis^ v. infrastigmalis^ v. hasalis.

Family Piendae.

Although these insects are very numerous and have long ser-
ved for investigations (even SwaMmerdam directed his attention
to F. hrassicae)^ still opinions differ a great deal concerning them.

J. F. VAN Bemmelen confined himself in 1912 to a comparison
between the pattern of a full-grown caterpillar of P. hrassicae and the
pupae of various Piendae, Vanessa spec, and Papilionidae and found
patterns which harmonized fairly well (compare chap. Ill and YII).

Forbes (1910?) thinks, according to Fracker (p. 136) that
the chalazae — large spots bearing the setae — have come from
primary hairs.

Fracker (1915, p. 136) denies this.

Buckler (1886, Part I p. 148 sqq. IM. II sqq.) gives long
descriptions of various Piendae, with illustrations of different instars.

Very conspicuous is on PI. Ill fig. 16. P. daplidice, which also
by HiiBNER (1786,. Vol. I) has been represented as possessing a
setal pattern just like that of P. hrassicae in instar /.

The only one, who as far as I know, has occupied himself
with a similar investigation about the ontogenesis of the chalazae,
is Froiiawk (1914). It is a pity that this careful study will
probably be unattainable for most entomologists.

Frohawk also draws this caterpillar, but in the last instar sub-
stitutes this primary pattern by a homogeneous distribution of
the setae. Probably the first-mentioned writers have studied a
younger instar or elsewise have met with deviating individuals
keeping the old pattern.

HoRSFiELD and Moore give (1857, Vol. I, PI. I, fig. 13, 14)
a similar drawing of Pieris eucharis Drury and P. helisama Cramer.

Sharp gives (1901, II p. 358) a drawing of Euchloe cardamines
instar I with bifurcated glandular hairs in the primitive arrange-

Ill

ment, later on the .v. dorsales only remain in existence in an
unaltered condition.

P. tlaplidict also poesemes these glandular setae.

Ql' AIL (U>04) discovered III Bss.*. ftfOsttujmaliK on P. brassirae.

In opposition with these opinions we find others e.g.:

Packard (1890, p. 495). "The true Pieriwu all live on herbs,
sometimes on low bushes and none of them is provided with hairs,
bristles or spinules."

Dyak (1894, p. 204) says: no trace of tubercles.

DiXKY (1894) devoted an article to the phylogeny of the Pie-
rinae^ Paying attention to the wings only. It would be very inte-
resting to repeat this study for the larvae.

As it si>emed im|)ortant to me to examine this question c1o8(t,
I chose the caterpillars of two kinds, which are often found on
the same plant {Branitira), and are very similar as egg and as
imago, but of which one bears bright warning colours (Poulton,
1890) and the other is in near accordance with the surroundings
(protective coloration).

Pieris brassicae Linn. Plate V, fig. 1 — 7.

Material. Eggs and larvae of the 2"d generation, laid on cab-
bage and Tropaeolum.

Efftjs laid in groups on the lower side of leaves, 1 mm. high,
shape of a pitcher, with 15 — 18 (mostly 18) vertical ribs, cross-
striped. The head of the larva can be seen at the side during
the last two days. Duration 4 — 6 days.

Instar /. Duration 4 days. Length 2' .^ mm. The larvae bite
a little hole in the side of the egg-shell, eat the top of the egg
and crawl out. Then they eat the whole egg-shell. Tubercula black,
skin yellow-green, transparent, no trace of stripes. Head imme-
diately black, at' first a little transparent, after ten minutes
pitch-black. Setae near the eyes.

Prothorax. There occur: s. dorsalis, s. subdorsalis, s. suprastig-
maliSj two s. dorsolaterales on one tuberculum, a minute s. prostig-
malis^ mostly two s. basales.

The pro thoracic shield appears half an hour or two hours

112

after the hatching, a fact which deserves attention as the head
and the tubercles are black from the beginning.

Mesothorax and Metathorax. S. dorsalis, s. suprastigmalis, s. dor-
solateralis, a small s. j^^ostignialis, no rudimentary stigma, though
the tracheae are seen through the skin ; s. bamlis, sometimes a
8. pedalis.

On the border of the mesothorax and metathorax is a rudi-
mentary stigma.

Abdomen 1, 2. S. dorsalis^ s. subdorsaUs, s. snprasUgmalis, s.
podstigmalis^ a. infrastif/malis, three s. hasales.

Even after repeated examination I could not find in my ma-
terial a 8. prostigmalis.
Seym. 3 — 6 = 1, but two s. hasales on one tuberculum.

^ 7 = 1, two s. hasales.

„ 8 = 1, two s. hasales^ large stigma.

„ 9 = 1, no 8. infrastigmalis, in the beginning a rudimen-
tary stigma (?).

„ 10. S. dorsalis, s. suhdorsalis and s. awprastigmalis with
that of the other side on one median anal shield which
gets black from half an hour to two hours after the hat-
ching, whilst the tubercles are immediately black, two
8. hasales, rather far from each other.

„ 11. Behind the anal shield is a black spot with a single
seta. I think that a part of the s. basales of 10 belongs
to 11, as s. infrastigmalis fails already on 9.

Towards the time of moulting brown-red spots appear between
the primary setae, which are mostly ring-shaped. They also arise
as a broad border round the primary tubercles, and are due to
the transparency of the skin which allows the colours of instar II
to be visible.

Instar //. Length 4 mm.. Duration 4 days. The skin bursts just
behind the head. The caterpillar creeps out at the front, as when
leaving a bag, the skin of the head remains for a short time as
a shield on the head. The caterpillar does not eat the old skin.

In the middle of the back we see a rather sharply confined,

113

bright-yellow linea dursalis, beginning between the two prothoracic
shields and ending near the anal shield; the skin is green. The
arrangement of the tubercula is as in instar /, but the setae are
much longer and between the primary ones numerous small,
M'oondary tubercula, each with one seta, have inserted themselves.

The prothoracic shield has now four tubercula, one ventrad of
«. sulHloi-salis ha« been added = .«♦. itub<lorsali.f inf. V

On the distal border is a row of red spots. On all the segments
the easily recognisable primary tubercula are round the edges
still a little red. Segment 11 is no longer to be seen distinctly.

Instar ///. length 6 — 8 mm. Duration 4 — 5 days.

Ik'sides the characteristics of instar //, we see a bright-yellow
lini*a stigmalis and many small hairs between the primary ones
which remain distinctly visible. Instead of the mono* and bisetose
tubercula basalia we find on the segments 3 — 6 a row of s, basales^
curving over the beginning of the abdominal legs and over the
places where legs might have stood on the segments 1, 2, 7, 8.

On the abdominal and anal legs there is a black spot covered
with short hairs, which spot might perhaps be taken as a highly
modified jf. ftedalin.

Instar IV. Length 12 mm. Duration 5 days.

The smaller setae have grown a great deal, especially one under
i«. subdorsal in of the abdominal segments. Between ». dorsalis and
j(. suprastuj mails on one side and s. subdorsalis on the other
we see a new distinct row and also one spot caudal of s. sub-
dorsalis.

We can easily recognize the primary tubercula: s. dorsolate-
ralis of the prothorax for instance is double as in instar /.

They nearly all possess, however, a ring of smaller setae round
the larger ones. The stigmal line gets very broad.

Instar V. Length 23 mm. (to 45 mm.) Duration 9 — 14 days.

The tubercula are very large and conical, they bear many
setae (= chalazae).

The linea stigmalis is continued on the head as a white stripe.
The arrangement of newly arisen tubercula in vertical rows is

114

distinctly visible. On the meso- and metathorax chal. dorsalis^
dial, dorsolateralis and chal. suprastigmalis are united.

We can still recognize the primary tubercula on the abdominal
segments, though chal. sujvastiffnmUs has become double. The
arrangement in cross-rows is more striking than that in horizontal
ones. Distinct are e. g. on abd. segment 5 : chal. dorsalis, chal.
subdorsalis, a double chal. suprastigmalis (oral of it a large tubercle),
s. poststigmalis and s. infrastigmalis. Many s. hasales, and under
chal. subdorsalis a large new chalaza.

Instar VI = Chrysalis. As J. F. van Bemmelen (1912, p. 114
and fig. 6) has shown, we can compare the spots of the pupa with
the chalazae of the caterpillars.

This is especially clear when the stigmata are examined. I feel
justified in designating the spots as follows: the spot under the
stigma as macula infrastigtnalisj that behind it as m. poststigmalis and
taking this for granted, the lowest of the four spots over the stigma
as m. suprastigmalis. I see the proof of this in the oral and ventral
enlargements which this spot shows on some segments, agreeing
with the two chalazae, so distinctly developed in instar V.

The uppermost of the same row seems to me to be m. dorsalis
and the double spot in the row caudal to the one I have just
discussed, I consider to be m. subdorsalis, which is also double
in instar V.

From this point of view the spot under it may obviously be
considered as the large one under m. subdorsalis which has
appeared in instar V.

The rows oral of the first-mentioned and caudal of the last-
mentioned one agree entirely with those of the caterpillar.

We must still explain the spot, generally large, ventral of the
stigma. In my opinion this is the m. bctsalis, which has nearly
become irrecognisable in instar F; and ventral of it a narrow,
elongated spot agrees with the row of the s. pedales.

It is more difficult to explain the two spots between m. dorsalis
and m. suprastigmalis. The more dorsal one of these two we may
take as a part of m. dorsalis, which also in the larva often bears

115

two Umg setae, and with which indwHl on tho last abdominal
segments of the chrysalis this spot becomes united.

The spot in between might corri^spond in this case to chal.
donolateraliif, which only tKVurs on tlie thoracic segments of tho
larva. If this is true, it would appear that this spot belongs to
the abdominal segments as well as to the thoracic and has dis-
appeared from them in a secondary way. If the last abdominal
segments of the chrysalis did not show us p8eud«>primitive con-
ditions, this conception would become more probable. In the same
way we might try to find the origin of the spot under m. .tuh-
tloi'nalis in «. gubdorsalis inferior.

Instar VII. Imago. J. F. van Bemmelen could trace the same spots
on the body of imagines which had not yet emerged. Here especially
III. infnistigmalig and in. posMitjinalis are distinct, the rows on the
oral and caudal edges of the segments and the double in. suhdorsalis.

Recapitulation.

1. It has been proved to be possible to reduce the intricate
pattern of the last instar of caterpillars to the pattern, as it oc-
curs in newly-hatched larvae.

2. This pattern of instar / agrees in the main with that of
the caterpillars of other families. (Type I).

3. The pattern of the pupa and imago is more like this pri-
mitive pattern than that of the last instar of caterpillars.

4. In instar // the linea dorsalis, in instar III the linea
stigmalis arises spontaneously i. e., without any stage of transi-
tion in the preceding instars.

Pieris napi. Linn. Plate V, fig. 8 — 13.

Egg laid apart on the lower and upper sides of cabbage-leaves
and Tropaeolum majus L. It resembles that of P. hrassicae but
is 1'/^ mm. high and has 15 vertical ribs. Duration 6 days.

Instar 7. Length 2'/; mm. Duration 4 or 5 days.

The colour is transparent, bright-yellow with numerous copper-
brown spots in vertical rows. The intestinal canal and the air-
tubes shine through. Head provided with setae.

lie

Prothorax. S. dorsalis, s. subdorsalis, s. suprastigmaUs^ s. dorso-
latemlis, s. prostigmalis, s. basalis, smaller s. ped(des. No protho-
racic shield.

Mesotkorax. S. dorsalis, s. suprastigmalis, behind which a small
.<?. suhdorsalis, s. prostigmalis, s. hasalis, smaller s. pedcdes. If
a rudimentary stigma occurred, I certainly should have found
it, as the tracheae were very distinct. The two lateral main-stems
are connected with each other near the caudal edge of the seg-
ment (see Plate V, fig. 8 of instar /). Here the rudimentary stigma
lies in the intersegmental membrane.

Metathorax ^ mesothorax but tlie above-mentioned s. subdorsalis
is absent.

Abdomen 1, 2. S. dorsalis^ 8. subdorsalis, s. suprastigmalis, s.
infrdstignmlis and s. poststigmalis, mostly small s^basales.
Segm. 3, 4, 5, 6 = 1, but the three s. busales are larger.

J, 7, 8 = 1, but one s. hasalis.

y, 9. No 8. infrastigmalis, one s. hasalis, for the rest as 1.

„ 10 = 1. No anal shield, but on the place which in P. bras-
sicae agrees with this shield, three setae are found. Two
8. basales.

„ 11. Behind the three setae of 10, on the area of the anal
shield of P. hrassicae, three small setae arise agreeing
with s. dorsalis, s. subdorsalis and s. prostigmalis. Ventral
of these is a large one, which I consider to be s. hasalis.

Instar //. Length 6 mm. Duration 4 days.

Between the primary setae we find many smaller ones and
setae without a tuberculum arise on the leg. The primary ones
remain distinct, they are much larger than the secondary ones,
the setae are longer and thicker.

Of the secondary setae the vertical rows along the oral and
caudal edges of the segment and many s. basales are conspicuous.
Segment 11 is still clearly visible.

Instar ///. Length 9 mm. Duration 8 days.

The whole body is now covered with numerous little knobs
and setae between which the primary setae are clearly visible.

117

One seta situated over the stigma and s. supntstiymaliit is quite
like a primary tuberculum and therefore agrees with s. dorso-
Idteratis on the thoracic segments. .S'. iHfr<k'<i'ujin(tlis has become
double, but for the rest we hardly find under the stigmata any long
si^tae with a well-developetl tuberculum. There are numy vS. bu.sales,

Instar /K. Length 13 mm. Duration 4 days.

The number of secondary tubercula inorwues, the primary ones
e. g. on abd. segment 5 : .s. dorsalis, a. subdorMlis, s. suprast'uj-
malist^ it. (tornolateralis (arisen in instar ///), s. infraKtit/maliH and
n. poittittitjfnalis remain distinct, the s. ba:<ale^'< are indistinct. Under
«. subdorsalix we find a new tuberculum which in form and size
agrees with the primary ones and which therefore perhaps agrees
with it. siibdorsalis inf.

There is no real dorsal line, but exactly in the median plane
the tubercula and setae are wanting, so that there the skin has
a different aspect.

Instar ^\ Length 16 — 25 mm. Duration 8 days.

A dorsal line is faintly visible and coloured bright-yellow. In
front of the stigma we find one yellow spot, behind it are two
spots. There is no question of a stigmal line, but the pro-
longation of the spots is of a somewhat different colour to the
rest of the skin.

The primary tubercula and setae are not distinctly visible, as
many secondary ones resemble them very much. The pattern is
lost and the animal makes the impression of possessing a coat
of irregularly homogeneously spread setae. The tubercula are now
coloured somewhat black. Under the stigma they are almost absent.

Instar VI. Chrysalis, Smaller than P. brassicae (±20 mm.).

I had at my disposal the variety which Buckler described
(Vol. I, p. 1 58) i. e, the bright, tender green one with many spots.

The pupa quite agrees with that of P. brassicae, it shows dots
corresponding with the tubercula of the larva mentioned below
and the last segments are reduced in the same way:

Manila dorsalis, under it sometimes a small spot which in my
opinion belongs to it. ♦

118

macula dorsolateralis, sometimes double.

m. stiprastigmalis.

m. siibdorsalis, double.

in. infnistigmaUs and m. poststiynudh.

two in. bascdes in one horizontal line.

one very tiny, elongated m. pedalis.

Here too we find an excellent correspondence between the spots
of the pupa and the primary tubercula of the larva, whicli fact
the more deserves our attention, as they had become totally irre-
cognisable in the last larval instar. This is of great importance in
connection with the conception of the colour pattern of the chrysalis.

Recapitulation:

I. The apparently homogeneous setae have been derived from
the primitive pattern, as it occur in instar /.

II. This pattern almost completely resembles that of P. hrassicae
and other caterpillars. (Type I).

III. In instar /// s. dorsolateralis^ which at first only occurred
on the thoracic segments, appears on the abdominal segments.

IV. The pattern of the pupa can be traced to that of instar i,
not to that of instar F, but there is besides a spot in the position
of s. dorsolateralis.

V. Linea dorsalis and stigmalis are indistinct, only in instar V
they are represented by coloured lines and the last line is made
up of spots near the stigma.

Family Riodinidae. As far as I know, no description of the
larvae mentions a distinct setal pattern.

Family Lyhytheidae. According to Edwards, quoted by Fracker,
the larvae resemble Pieridae as far as their setae are concerned.

Family Nytnphalidae. This family is here taken in the same
wide sense as W. Muller did in 1886. Therefore the L?/mwa(/«c?a^,
Ithomiidae, Helicotiidae, Agapetidae are also included in it, together
with the Satgrinae, Danainae and the Nymphalinae s. str.

Gruber (1884) describes the development of the scoli from the
primary setae, W. Muller (1886) contradicts this. He says that

119

tho scoli are formed from socondary setae, placed between _tho
primary ones (c. f. Chapt. II and my PI. I, fig. 1).

For this large group I refer to Wkismann's studies (187(5) and
to \\. MCllkk (1886). Where this family is concerned, Fkackkh
(li)15) follows W. MCllkk, but thinks that the expressions mudio-
dorsal, subdorsal etc. can be applied exclusively to ''secondary"
stu)li and even then only when they are placed in one transverse
row. This clashes with MCli.kr's own ideas (I.e. p. 250).

The pattern of the primary setae is Type I on the abdomen.
Type II on the thorax. I was only able to examine V^anessa
urtieae in instar I. This specimen already possesses scoli and
secondary setae.

As I have cited in Chapt. II MCller's ideas on the Ni/m-
pluUUUte^ I do not describe this caterpillar any further, but refer
to ^1. V, fig. 14,15, where instar / and abdominal segment 5 of
the full-grown caterpillar are figured. I would rather draw
attention to the pattern of the pupa which van Bemmelkx
(1912) described. There is a small spine on the place of «c. (/o/'»rt/w,
and pigmental spots agreeing with sc. (lorsolateralis (which is not
developed on the abdomen of the caterpillar) 8C. suprmtigmaliSj
sc, infrastigtnalis, sc. basalis, sc. pedalis.

Family Papilionidae.

In 1884 Gruber described the larvae of the Swallowtail
butterflies, instar /. To this paper too little attention has been
paid by Schulze (1912), Fracker (1915) and others.

Gruber had a complete material of Papilio asterias^ P. turniis,
P. troilusj P. ajax^ P. philetwr at his disposal. All these cater-
pillars agree with each other in the fact that during instar / they
possess verrucae with many setae, which in P. ajax are bifurcated,
but in the others terminate in a knob. In P. philenor the tubercles
bear only one seta.

During the ontogenesis these verrucae become smaller and at
last are replaced by colour-spots, as can be distinctly seen in
P. asterias. As far as I am able to see the arrangement on the

120

abdomen is : v. dorsalu small, v. subdorsalis largo, i\ supfastiyinalis^
V. infrastigmalis very large, v. dorsalis. The arrangement on the
thorax is not distinct. Verruca subdorsalis maintains itself longer
than the rest and this pattern reminds us of that of the Lijmantridae.

W. MuLLER (1886) is of opinion that the hairs of these P«^////o-
nidae may be considered as primary.

PiEPERS (1888) examined several Javanese Papilionidae. He
says of Papilio agatnemnon amongst others that in the beginning
it has bifurcated hairs which have disappeared before the first
moult. He thinks that they are bitten oif by the caterpillars
themselves.

Chapman (1895, p. 88) says: "The young larva of Papilio
machaon seems obviously reminescent of an adult Vanessa-larwa,.
Yet it is certain that, whether Vanessa be or not be derived from
a Pajyilio'lilie form, nor is there any probability that any adult
Papilio-\a,T\& ever was spinous in precisely that manner. The
spines are a special development of the young Pa^nlio-larva, for
protective objects affecting itself. They have not been derived
from spinous full-grown larvae amongst their ancestors, and are
not passed on to the present adult larva, because it does not
require them. The processus on the adult larva of Ornithoptera
may be derived from the spines of the first stage and are not
ancestral to those of the young Machaon.^'

Fracker (1915) does not seem to know this. He only says on
p. 138; "No unpleasant spines or horns are present to discourage
the observer and no discordant colors to offend him."
. ScHULZE (1912) discusses the Osmateria (Nackelgabel) and says
they are derived from "dorsal protuberances". He points out the
meaning of the ellipsoid glands as secretory organs, and thinks
that they have no use in dispersing the Ichneumonidae.

Sericinus telamon Don., Coll. Kall., is covered with secondary setae
which however are absent round many verrucae. Plate V, fig. 14.
The back part of the body has been a little damaged, so that
I was not able to examine the anal segments. There are:

121

Prothurujr. In tho position of tho r. dormle^ the ogmuteria;
f. Kub^lormliit, v. protttiginalut as a long spinule, r. basaliif^ t\ ftost-
jtedaiis with only two setae.

Memt/iota.r. 1'. </o/"8rt/w, r. sHbtiuratilitt^ r. /n'ostitjttmlix^ r. Ixmalis^
r. ftostfteiialLs with three setae and one verruca above the wing-
rudiment.

Mftdthorax = tnesothorojr but the last nientioni'd verruca is only
a pigmental spot, and such a spot is also found in the position
of r. inft'(tstiijniali.<(.

Abdometi 1 — 9. V. (.-<m6 fyorita/w, v. infriuitigmalis^ c. bmulis,
p. pedaiis.

It seems to me that this form is very primitive.

I saw in some unpublished draw^ings of Prof. J. F. van
Bkmmelen after specimens of Pajtilio fjodalinus from the collec-
tion Kallenbach (which I could compare with the original objects)
that the pigmental spots of the full-grown caterpillars are arranged
in a definite pattern. In my opinion they agree with v. dorsalu^
r. suproidigmalis, r. siibdormlis sup. and im/., r. prostigmalis,
r. jtoststigmalLt and r. infrantigimdis. The presence of v. subdor-
gaiis inferior and of v.prosfigmalis is remarkable, as they do not
occur in Gruber's figures of instar I.

On the pupa of Papilio machaon (J. F. van Bemmelen 1912,
fig. 5) a tubercle stands in the position of v. dorsalLt. Between the
stigma and this tubercle there are two spots which therefore
agree with r. suprastigmalis and r. dorsolateral is. So here too a
r. dorsolateralis occurs on the pupa which is not developed on the
abdomen of the caterpillars. Further there are spots on the pupa
which agree with r. prostig malts, v. jwststigmalis ^ v. basalts and
often V. pedaiis.

As a larva, P. machaon varies very much in pattern. In
most cases, however, the above mentioned spots are present.

122
CHAPTER VII.

Comparison between the patterns of caterpillars and those
OF THE Pupa, the Imago and of other Larvae of Insects.

From early days the pupae of butterflies have had a great
attraction for the human mind.

The word chrymli^ was already used by Aristotle to indicate
the goldcoloured pupae of some butterflies.

The conception of the pupal-stage has not always remained the
same however. W. Harvey for instance considered the pupa to be
a perfect egg;.

J. Swammerdam who was the first to discover the wing-rudiment
in the caterpillar, suspected that the pupa was not a new creation
but that the result of another moult (1737, Bihlia Naturae p. 567).
He attached much value to the fact, that the setae of the cater-
pillars also occur on the pupa.

The majority of later investigators were convinced that the
caterpillar is the primitive state and the pupa a secondary pheno-
menon. When the Darwinian ideas gained ground, it therefore
became an important problem, how this resting stage had arisen.

The first who tried to give a solution was John Lubbock (1871),
in a paper which became very popular. He thought that the pupa
was a form of transition between the larva and imago, necessitated
by the great difference in the mouth parts. I think that the con-
ditions of Micropteryx and Eriocephala definitely dispose of this
theory.

Lubbock's opinion has been propagated by many treatises, but
from the first another view has maintained itself against it.

Brauer (1869) defended the opinion that it is the larva which
has become modified and this conception has recently gained
ground again.

Ivatzeburg's discovery of the external genital organs of the
pupae (1840) was not paid attention to at the time, and thus
it may be explained that Jackson and Poulton thought they
had described these organs for the first time (1890) after

123

having diwovcrod them again respectively in 1875 and 1883.
Though convinced that the larva is u phylogenetieally old form,
PouLTON has still contributed much to propagate a now idea of
the pupal stage. Many argumonta are used by him to confirm
his assertion that the pupi is a subimaginal stage, which has
become immovable.

As his chief proofs I mention the well-doveloiKKl external ge-
nital organs, the simpler structure of the wings, the large antennae
and wings of those female forms, which in the imaginal state
possess short ones, etc. The agreement in colour between cater-
pillar and pupa has been observed by him, but that feature he
considered of little value.

J. F. VAX Bemmelen discovered in 1889, that the definite
colour pattern of the imagines is preceded in the pupal state by
a simpler one. More recent investigators confirme<l this and Bryk
(1914) demonstrated its persistancc in anaberrative imaginal form.

Packard (1889) discovered that Boinbun, before emerging from
the pupa, moults once more, so that this reminds uh somewhat
of the subimaginal stage of the Ephemeridae.

In all these ways of considering the pupal instar it was taken
for granted that all the Holometabola originated monophyletic-
ally. To this idea Swammerdam has certainly contributed a great
deal, and Lubbock too was of the same opinion.

Brauer (1884, p. 318) directed attention to this problem
without solving it, however. As it is closely connected with an-
other question, namely : is it necessary that the setal pattern of
the caterpillars agrees with that of the other larvae or can it be
derived from it, I thought I had better give a short synopsis of
the literature on this problem.

Miall (1895) considers the incomplete metamorphosis of the
Orthoptera to be a primitive one.

Boas (1899) thinks the meaning of the pupal stage is to pro-
cure the animal the opportunity of developing its wings. His
decisive assertion (p. 397) is incomprehensible to me: "Es ist
demnach ausgeschlossen, dass das Insekt vor der letzten Hau-

124

tung seino Flugel entwickele". As he himself cites Swam-
merdam's discovery of the subimago of the Ephemeridae^ it must
also be known to him that Swammerdam described (Biblia Naturae
I, p. 269 and Historia Generalis p. 87): "how in June 1670 in the
neighbourhood of the village of Slooten, the subimagines flew upon
his coat, that they moulted there and returned directly afterwards
to the water." By this observation the argument of Boas is refuted,
ilk my opinion, more than two hundred years before it was used.

De Lameere (1900) tries to find the origin of holometabolism
in the habit, insects acquired by penetrating into vegetable tis-
sues. The different Holometabola might therefore be derived mono-
phyletically from the Neuroptera. Handlirsch raises many im-
portant objections to these conceptions (See p. 126).

P^Riiz (1903) mentions several causes, but, as Henneguy
(1904, p. 692) rightly observes, these are not explanations, but
only statements of facts.

Heymons (1907) lays emphasis upon the great changes, which
the so-called Ametabola undergo (e. g. Machilis). He says on
p. 160, that the pupa is a new stage. Naturally we must not
think that no trace of it is to be found in the lower insects :
„E8 kann erstens die Holometabolenpuppe das umgewandelte
letzte Larvenstadium oder die 8umme der letzten Larvenstadien
der Hemimetabolen repriisentieren, oder es kann zweitens die
Puppe weiter nichts als eine unvolkommene Imago selbst, ge-
wissermassen eine vorlaufige noch unfertige Ausgabe der Imago
sein." Heymons adopts the latter alternative and is therefore an
adherent of the subimaginal theory. He thinks besides that the
Lepidopterous pupa has become secondarily movable again. In
this connection I think attention should be drawn to the investiga-
tions of Chapman (1893) who demonstrated that it is exactly
the pupa incompleta which occurs in the lower families. He also
directed attention to the fact that the pupae are generally much
more movable than is usually supposed. In my opinion the mova-
bility of the Lepidopterous pupa is to be considered a primitive
characteristic.

125

Deeoekrk (1909) supposes the rigidity of the ehitinous
skin to be the cause of the moultings and he thinks (p. 19
sqq.): ''that the higher specializeil insects possess a tendency and
a will to diminish the number of these moultings. This can only
bo achieved by retarding the development between two moul-
tings, and so attaining by one eedysis what otherwise could only
be attained by many. The more these moultings were reduced
in number, the less recapitulations of phylogenetic stages had to
be passed by the larva, and so it makes the, impression that
phylogenetically its development is retarded, and ontogenetically
it rcnmins all the longer in a primitive state, by retardation of its
development. Through this, the larva has the opportunity of
specializing itself, according its own desire and character. The
great difference has ariseii in consequence of a different manner
of life, especially the aquatic one. Pupa and subimago are
not identical, but they are both primary stages of development,
which have remained preserved."

Bastis (1918) and Cakpester (1913) are both adherents of
the subimaginal theory, especially the latter, who has published
a short but very interesting treatise, which contains a great many
facts and is written with grejit conviction.

In constructing a theory on the origin of the pupal state it is neces-
sary to study the subject from as many different points of view as
possible. By Weismann's investigations especially (1863, '64, '66),
later on continued by i.a. J. van Kees (1888), MESNiLand Metsch-
MKOFF (1900), Bauer (1904), Janet (1909) and Poyarkoff (1910),
the attention of the investigators has been turned more to the
histological processes and in the first place to the histolysis in the
pupal stage, than to the purely morphological problems. Histolysis
may be considered to be a secondary phenomenon, but from that
it does not follow that the pupa itself is secondary. The larva of a
frog is not secondary because its tail disappears by phagocytosis!
It is therefore desirable to direct the attention to palaeontology,
which has been greatly neglected by many entomologists. On this
point I refer especially to the last of the three following writers :

126

Scudder(1886),Brongniart (1894) and Handlirsch(1903,'06, '10).
Zittel-Broili (1915) agrees with the latter in the main points.

The number of fossil insects has now become so considerable,
that the existence of a palaeo-entomology can no longer be denied.
The principal lines of the genealogical tree can already be drawn.

Handlirsch (1903, p. 720 sqq.) directs attention to the fact,
that originally the abdomen consists of 11 segments and a telson,
the 11th segment bearing a pair of cerci.

In the Palaeozoic era there are only Ametabola ; all the different
Metabola appear at about the same time, namely in the Permian
formation.

In the Palaeozoicum the relatives of the Polynephria are found,
in Perm and Trias the Oligonephria appear. From the fact of
their synchronistic appearance it follows, that perhaps a poly-
phyletic origin of the higher orders may be assumed, which orders
therefore appeared as parallel lines. In the Perm the branch of
the Neti rapt eroi (lea and that of the Coleopteroidea are already
separated from that of the Panorpoidea. To the latter belong the
Phrt/f/dHoideUj Panorpata, Lejndoptera, Sucto7'ia and Diptera. This
classification differs entirely from that of Lameere.

Handlirsch adduces against this writer (1900, p. 174):

1°. The reduction of the cross-veins. Very often the recent
holometabolic Nearoptera and Sialidae still show a more original
vein-course than the carbonic Megaseco])tera (the primitive group
of the Panorpoidea)^ besides the reduction of the veins is a
common feature in lleteromctabola.

2°. The habit of carrying the wings in a horizontal position
is only found in Heterometabola, these alone have well-developed
cerci; the Mega^ecopterous-\a,v\diQ d\m had wing-sheaths like that,
and therefore were probably heterometabolic too.

3°. The Lepidoptera are originally phytophagous and they only
become secondarily endophagous. In this case organisms would have
been driven in a progressive direction by the influence of parasitism
according to Lameere, whereas as a rule they degenerate . by it.

4°. If we adhere to a monophyletic origin of all Holometabola,

127

the holuinetabulie Paiaewlicttfoptem must have been the ancestors,
but then the lleterometabohi nuist have been developtnl polypliy-
letii^Uy, and what was always considereil as being primitive,
wouhi turn out to be seeondary.

None of the different e^iuses of holometaboly, given by the
other writers which I have already quoted, can be the reason
according to Haniu-irsch, because many larvae have remained
phytophagous, and nevertheless are still heterometabolic. Neither
can endophagism be the reason, because nearly all the Holometabola
are carnivorous or phytophagous, lleterometabola as well as Holo-
metabola can lead an aquatic or subterrani>ous life, so that only
meteorological causes remain. The beginning of many groups
of the Holometabola at the same geological period, namely in the
transitory |>eriod between Palaeozoicum and Mesozoicum, also
indicates a heterophyletical origin. Perhaps the glacial age of the
Perm has been the decisive factor. However this may be, it is
certain, that Hasuliiu^c'H, on account of his extraordinary knowledge
of fossil insects, does not think that the monophyletic origin of all
Holometobola is |M>ssible. For the problem which I am endeavouring
to solve, this means that the skinreliefs of the larvae and nymphae
of the different orders need not neccssjirily agree with each
other. I will soon refer to this point again, after having compared
the pupa of the Lepidoptera with the caterpillar.

PouLTON (1890, p. 193) drew attention to the fact that the pupae
of the Sphingulae exhibit for a short time the same pattern which
the caterpillars possess in the last instar. The stripes, however, are
secondarily hidden by a brown colour, in regard to which it should
be noticed that the wings and other new-formed parts adopt this
colour later- than the organs already present in the larva. Poulton
did not attach much value to these stripes, for he says: ''The
persistence of such colours depends upon the fact, that the hypo-
dermis-cells of larva and pupa are the same, so that any pigment
contained in them during larval life, may remain unchanged after
the pupal period has begun".

Though I have great confidence in Poulton's knowledge of

128

morphology, yet I think I am allowed to oppose Lameere's opinion
(1900, p. 623) to his, viz. that the hypodermis is entirely renewed
by histolysis. Besides the colour-pattern of the pupae cannot be a
simple copy of the larval pattern.

J. F. VAN Bemmelen (1912, p. Ill — 117), has observed that the
hypodermic pigment has also a morphological importance and he
succeeded in demonstrating, that the same pattern exists on the pupae
of Pieris brassicae, Aporia crataegi and Euchloe cardamines^ and
that this pattern agrees to a great extent with that of Papilio machaon
and Vanet^sa urticae and V. to. The caterpillars of these butterflies
differ a great deal in their pattern. Van Bemmelen thought, that a
great resemblance was to be remarked between the pupal markings
and the larval pattern of Pieris brassicae. He took this to be
the colour-pattern of the once movable chrysalis.

(Jn this idea I have based my investigations. In the main I agree
with van Bemmelen's opinions, with this restriction however, that
I should not compare the pattern of the pupae with that of a full-
grown caterpillar of Pieris brassicae^ but with the pattern of instar I.

In chapter VI I have shown that the colour-pattern of Pieris
brassicae and that of P. tiajn instar J, closely resemble that of the
pupa. Compared with the last larval instar of the first species a reduc-
tion of the number of pigment spots has taken place, with that of the
last mentioned on the contrary, a strong accrescence of the number.

This is a convincing proof of the inexactness of the statement
that the pupal pattern should be a simple copy of existing hypo-
dermal pigment.

J. F. VAN Bemmelen demonstrated the same pattern on the
pupae of different Pieridae, Papilionidae and Vanessa spec. In
this article I have given further details for many families. The
arrangement of the verrucae on the pupae of Ocneria dispar and
Orgyia antiqua is also the same. Swammerdam already indicated
this last fact {Biblia Naturae, 1737) and Poulton (1890, p. 193)
also called attention to it.

Quail described (1900, p. 416, PI. V) the pupae of Porina
cervinata Walk. (Hepialidae) and showed here the same setal

129

(Mtterii as on the caterpillar. The pupae of Hepialns lupuUntua
in the coll. Kali, are a little damaged, yet they show setae ar-
ranged, in my opinion, as in type I, augmented by ». dorsolateralin.

It seems to me, that these are remnants of a formerly common
pupal |>attern, consisting of setae which had accumulations of
pigment at their baM>s. Just as is the case with the caterpillars,
the pigment spots can remain after the disappearance of the setae.
I cannot but think that this pattern of the pupae has taken origin
on a movable animal. Therefore I believe that I am allowed to
consider these remains of a pupal pattern as a proof of the theory,
that the pupa is a subimaginal stage which has secondarily become
immovable. ConstMjuently the pu|>a is not a phylogenetically younger
form, but a preserve*! primitive form which has become secon-
darily immovable. The agreement between the pattern of the
caterpillar instar / and the pupa is so striking, and the differences
between the pattern of the last larval instar and the pupa
are often so considerable, that it becomes interesting to try to
solve this problem.

I believe that this can only be explained by accepting the
first larval instar as well as the pupa as primitive forms, but
the following instars as newly acquired ones. The latter instars
are all specialized in diflFerent ways. Dekoener (1909) has also
advocated this hypothesis. Some instars become bearers of
warning colours, others obtain long thick tussocks, a third group
retains the primitive type pretty well, because it lives in hidden
places, but, when the pupal stage has begun, the old pattern
returns, to be sonietimes overspread by a homogeneous colour.

Even on the imagines (J. F. van Bemmelex, 1912) the old
pattern is sometimes to be seen. W. Muller (1886) thought that
the pattern could pass from the caterpillar on to the pupa, and
also the other way about. I think I have proved by the detailed
account of the Pierids, that in this case it certainly cannot be
true. Therefore we must return to the opinion of Weismann
(1876), who concluded from all these phenomena: "dass die
Errungenschaften der einzelnen Stadien in den folgenden Gene-

9

130

rationen iminer nur auf diese Stadien selbst wieder iibertragen
werden, die aiiderii Htadien aber unbehelligt davon bleiben"
(see p. 6 sqq. chapter II).

If the first larval instar as well as the pupa both show primitive
conditions, it may be that the colour pattern borne by both, is
80 old that it also appears in other orders of insects. In that
case it has already been obtained before the separation of the
different orders, i. e. in the under-carbonic period. Considering
that the families of Lepidoptera were separated, according to
Handlirsch (1906), in the cretaceous period, and hence the genera
and species still later, no great result can be a priori expected from
such an investigation. In the following I intend to discuss the orders,
which are in some way related, although I think that such a dis-
cussion has only a very relative value, if it is not supported by a
very extensive investigation. But for that I lacked time and material.

Handlirsch thinks that the Panorpata are in some respects
to be considered as the ancestral form of the Lepidoptera, It is
certainly of great value, that J. Botke (1916) came to the same
conclusion through his investigation of the colour-pattern of the
Lepidojftera, differing in this from de Meijere (1916).

Having explained in chapter VI that in different families, in-
dependently of each other, verrucae have appeared, and directed
attention to the fact that sometimes within the precincts of one
family rather important differences in the setal pattern occur
(e. g. Hepnalidae^ Pterophoridae^ Pieridae), it is not to be expected
that other orders should show the same pattern.

The existing illustrations of the larvae of the Panorpatae^
Neuropteridae^ etc. are still less exact than those of the Lepi-
dopterous larvae. Besides I think that in the preceding lines I
have sufficiently pointed out, how only complete series can give
us a good idea of the real character of the setal pattern and
that conclusions, reached by comparing the full-grown forms
only, can easily lead to wrong hypotheses. It is therefore with
the greatest reserve that I submit the following remarks on
the setal pattern of insect-larvae.

131

Bkaukr dovoted some articles to tho larvae of Pa nor pa com-
Munin (1851, 1852, 1863). He found that the larvae bear setae
in instar /, later on verrucae. According to the figures, on all
the segments three setae «)cour arranged in a row above the
stigma. The arningement of the verrucae cannot be clearly seen
in his figures. It is certain that rerrucne occur (braune hornige
Warzen) with rather short setae.

Felt (1895) described the Si'oi'pionfiirs^ but in his work the
setal pattern is not very distinctly indicatwl either. In connection
with the fact that there are verrucae on the Eriocephalidous larvae
and also on numerous other families, one would almost be inclined
to consider the verrucae as being primary. The simple setae of
the higher families might in that case be taken to be a secon-
dary characteristic. The disappearance of the verrucae on the
PapilioHuiae, the Bamhycidae^ the Kttdrotnidae and the Brah-
maeidae could then be used as an argument in this direction.
However, it seems to me that this hypothesis should not be ac-
cepted. As far as I can judge, the verrucae in all the families
are formed from simple setae. The verrucae of the Eriocephulidae
differ too much to be a strong proof of the hypothesis and they
take origin from single setae; the disappearance of the warts
in the three above-mentioned families of the SYMBOMBY-
CI DAE is easily explained as a reduction of the verrucae by the
development of a homogeneous setal cover. It is the same with
the Acronyctinae and the Papilionidae. In spite of Chapman's
statement (1902) I consider the verrucae of the first instar to
be rudiments of scoli which formerly were more strongly developed.
Dyar (1894) thought that the setal pattern of the Tenthredidae
was the ancestral pattern of the Lepido^dera. This writer adheres
to the monophyletic origin of all the Holometabola. There are nine
setae on either side of an abdominal segment and they are placed
in three rows each containing three, of which the middle one is right
over the stigma. He supposes, that in Lepidoptera the first of
these rows has disappeared, except perhaps s. prostigmalisj the
second row should agree with s. dorsah's, s. dorsolateralis and

132

s. snprastigmalis. S. suhdorsalis sup. and inf., as well as s. post-
stigmalis might perhaps be derived from the last row. Such an
explanation seems to me a little farfetched; besides the anato-
mical differences between the Hymenoptera and the Lepidoptera
are too great to accept such a near relationship between these
orders. Neither do the palaeontological data harmonize with
Dyar's opinion.

At present several writers defend a nearer relationship of the
Lejndoptera with the Neuroptera, amongst others Chapman (1896)
and DE Meyere (1916). With Handlirsch, I believe that the
palaeontological data do not agree with this hypothesis. I could
not get any proper data about the larvae of the Neuroptera.
According to the figures in the manuals the larvae have very
different forms; some are naked, others are covered with long
setae. The arrangement also seems to be very different (Oudemans
1897, p. 317—323).

The figures of the Trichoptera, the caddice. worms, are much
better. Siltala (1907) studied them accurately. In this order
also there appears a secondary augmentation of the setae during
the ontogenesis. In the first instar the setae of the larvae are
only very sparse. He could not find an agreement with the setal
pattern of the Lepidoptera and by studying his figures I came
to the same conclusion. The pupae are sometimes also covered
with setae and some larvae (e, g. Hydrojisyche) bear verrucae,
though mostly simple setae.

As the Trichopterous larvae have certainly undergone profound
secondary modifications, I think that too much value must not
be attached to the arrangement and form of their setae, though
they belong to the Panorpoidea and though Chapman (1896 c)
associates the Phryganeidae and the Micropterygidae together.

Although I do not believe that the Coleoptera and the Lepido-
ptera are closely related, I still think it necessary for the sake
of completeness to conipare the setal pattern of Jjeptinotarsa
with that of the Lepidoptera,

The classical investigations by Tower (1906) have drawn great

133

attention to this genus, and from his very accurate figures it is
easy to study the pattern. Tower distinguishes two rows of spots
on the abdominal segments, the anterior and posterior band of
tergal spots, each consisting of three spots, phiced in a vertical
row, namely the inner, middle, and outer tergals. Behind them
ootaies the spimcular spot and then the basopleural one. Ventral
of these are placed two rows each consisting of three spots, the
outer, middle and inner sternal spots.

On the mesothorax and metathorax these spots are partly united,
and in the position of the spiracular spot there is a wing-spot.
The prothorax differs a great deal and possesses a prothoracic
shield, with an anterior and a (Msterior pronotal band.

I think it an important fact that in these Coleoptera two
rows of three spots occur above the stigma. The Tenthredinidae^
which are certainly not so closely related, have three rows, each
consisting of three setae; on the Panorjtata one row of three is
found and on the Lepidoptera I think that the row above the
stigma altk) consists of three, including the 8. dorsolateral^.

In this connection the spot on the pupae agreeing with the
seta dorsolateralis acquires a greater importance.

It seems to me that these three spots or setae, placed in a
vertical row, have been acquired in very remote periods and that
the meso- and metathorax, though they have suffered profound
secondary modifications, have best preserved them.

De Meijere (1916) recently published an interesting study on
the wing-markings of Diptera and Lepidoptera. He has also made
a study of the larval pattern. His paper reached me too late to
consider it in dealing with the different species, and therefore
I may quote his main result here.

On p. 63, the author says, that it seems to him as if, when
a depositing of pigment has become physiologically necessary, it
is indifferent where that process takes place. It is only restricted
to the sixteen places or patterns given by him. I cannot agree
with this in so far as the larval and pupal body are concerned.

When he says p. 64 that in one and the same family the

134

patterns belong to different evolutionary rows, I can refer to pg. 137
and 139 where I have stated the same opinion. The pattern itself
comes back in different families (1. c. p. 75) and de Meijere is of
opinion that this has been caused by parallel evolution, whereas
I am convinced that generally spread patterns are phylogenetically
old ones.

On p. 132—138 de Meijere discusses the colour pattern of
the abdomen of Lepidopterom imagines. He agrees with J. F. van
Bemmelen in considering the spotted ones as primitive.

De Meijere (1, c. p. 136 — 143) compares the larval pattern with
the pupal one. He states the fact that the colour first appears at
the bases of the sensory setae (Sinnes-borsten) e. g. on Diloba,
Zeuzera, Hydroecia^ Pieris instar /, Abraxas. The last mentioned
species is highly interesting as Schroder (1894) says that the
stripes appear first, (see p. 20)

So far I agree with de Meijere. This writer however rejects
the hypothesis of the primary pattern of Lepidopteroiis pupae,
and liis chief argument is that the Neuroptera are the common
ancestors, from which the Trichoptera, Panorpata, Diptera and
Lepidoptera were diflFerentiated, after having acquired the holo-
metabolism. All these primitive forms and also the lower Lejn-
doptera, as e. g. Micropteryx^ have but slightly coloured pupae,
which live hidden in the earth or in cocoons.

I refer in the first place to Handlirsch and in the second
instance I think that in the foregoing pages I have given several
proofs of my thesis, that the pattern of pigment spots is the same as
the setal pattern. This setal pattern however is widely spread
amongst the uncoloured pupae, and so I suppose that this setal
pattern (type I) is an old phylogenetical one, and that the pigment
spots, in larvae as well as in pupae, follow this arrangement. The
pupae of the Rhopalocera are secondary in so far as they have lost
the setae, but have only retained the pigment spots. The pupae
which have become immovable and therefore often remain in the
earth or in a cocoon, have secondarily lost the pigment, but have
often preserved the setae.

i;i5
CHAPTER VIII.

QeMEKAL considerations and synopsis of the RE8ULTS.

In the preceding pages I have tried to lay duwn a general
fuundation for the armature of caterpillars.

In consequence of the shifting of the stigniuta over niesothorax
and metathorax and the development of the wingH on them, T
take it for granted, that these segments are to be considered us
being secondarily moditied.

On account of the anatomical differences with the abdomen,
this result might a priori be expected, and the chaetotaxy on
these s(>gment8 j)rovide8 proofs that the setal pattern also has
undergone secondary modifications. The prothorax too has taken
part in these changes by obtaining the stigma which originally
was placed on the niesothorax. Hence it is not desirable to start
from th(>se segments in reconstructing the primitive ])attern, as
Tsou and Fkackek have done (Chapters III and V).

The anal segments too differ in structure and even vary in
number. This probably happens in connection with a process of
reduction which in some species has farther advanced than in
others (Chapter IV).

Among the remaining eight or nine abdominal segments a few
occur bearing legs and others without them. On account of embryo-
logical facts as well as of the presence of the setae pedales, I
think it allowable to consider the segments with legs as the
more primitive ones (Chapter IV).

From the literature, Chapter II, it appears that all the different
investigators of the setal pattern have introduced a nomenclature
of their own, in which many made use of numbers. As some of
them indicated totally different setae by the same number and
as the same setae are indicated by different investigators by widely
different numbers, a great confusion has arisen, as is best illus-
trated by Plate I, fig. 1—21.

I have therefore been led to use a nomenclature which agrees with
that of Weismann, W. Muller, Scudder and J. F. van Bemmelex.

136

The setae on an abdominal segment are indicated by names
which refer to the place of the implantation.

An identical system has been applied to the thorax in which the
same names as far as possible are used.

Where the homology with the abdomen is not very clear, the
changes in the position of the setae have been indicated by
other names, because I think that a nomenclature should be a
means of describing a thing in a short and clear way, and not
an expression of more or less probable hypotheses.

The shortest method of description certainly is to indicate the
whole of the setae on a certain segment as Type I.

The other ways of arrangement on the abdomen, Type la
and lb, can be derived from Type I, by assuming reduction.

The thoracic segments differ most of all and are called Type II.
Very often a reduction of the number of setae has taken place
on their dorsal side, but at their oral border there is one seta
more than usually occurs on the abdomen. This seta I have
called s. dorsolateralh and in so doing I disagree with other
writers. By especially studying the setae and the pigmental spots
on the pupa and by an accurate comparison of the prothorax
with the abdomen, I have come to the conclusion that this seta
does not correspond to s, subdorsalis as is generally accepted.

On the mesothorax and metathorax a pigmental spot is often
to be found in the place where we might expect the stigma, if
this were exactly situated as on the prothorax. Most investigators
have taken this spot to be the rudimentary stigma. It is by
studying Boas, that I have come to the conviction that this is
not the case, but that this spot agrees with the wing-rudiment.

A shifting of the stigmata must have taken place and by means
of this fact I have tried to explain Type II (Chapter IV). Like
Quail I consider the seta in front of the wing-rudiment to be
s. prostigmalis (III B.).

For their bearers the setae may be useful in several ways, but
it is difficult to assume, that any correlation could possibly exist
between usefulness and arrangement of pattern. The consequence

137

18, that we may expect changes in the pattern to possess a cer-
tain systeniatical value. Though this rule is not always adhered
to, it generally holds good.

Of the setal pattern of caterpillars wo may wty what Chapman
says of the pupa: "The Lepidoptera certainly cannot be arranged
in one line by their pupae, but the Lepidoptera of one line can
be arranged by their pupae." In studying the pattern we get
the impression of many lines of development which often run
parallel. This completely harmoni/.es with Hanih.iksch's opinion
and with that of i>e Meijkrk.

Before passing on to the discussion of the families, I wish to
devote a few words to the biological signification of the setae,
Fkacker assumes (1915, p. 38) that the setae are sensory in
function.

Quail (1900) thinks that the setae of the llepialidae can open
and shut, Wachtl and Koknauth say that the special setae of
Psilum serve to facilitate the spreading of the caterpillars by
the wind. Usually it is thought, however, that the setae serve
as a means of defence against enemies, especially Ichneiimonidae
and Tachinuiae. This opinion has particularly been propagated by
Packard and Poultox. I think I may call it into doubt. The
experiments of the last-mentioned writer give us a right to assume
that a dense covering of setae or tufts and long pencils, form a
means of defence against some vertebrates, but the results he
obtained cannot be directly transferred to enemies of the insect-
tribe. In structure the eyes of the insects differ so much from
those of the vertebrates and are so absolutely different in their
sensibility to colours, that we may not treat the problems which
here present themselves, from a point of view so anthropomorphic,
as for instance Poultox does, in his well-knowm and interesting
book: Colours of animals. On page 87 he says: "A person un-
accustomed to the observation of the animals (the light-coloured
trout) would certainly fail to detect any trout except the black
ones, which were blind and did not vary their colour". I must
confess that I fail to see the value of this argument, as I am

138

convinced that the enemies of the trouts are certainly accustomed
to observe and detect them.

The results of my experiments in cultivating Acrowjcta pst
and Fieris napi — only to mention two widely differing forms —
brought me to the same conclusion as W. Muller came to, after
an experience of many years viz: that naked forms are as much
afflicted by Ichneumonids as the species which bear large spines.

In the first instar many setae are so-called glandular hairs.

The systematic signification of these generally bifurcated setae
cannot possibly be very great in my opinion. They occur in numerous
families: Papilionidae (Gruber, 1884), Nymphalidae (W. Muller,
1886), Notodontidae and Pterophondae (Packard, 1890), Piendae
(Sharp, 1901) and Sphingidae (Packard, 1905). For this last
family I have given an accurate description of the form and it
was only later on that I studied Packard's drawing which dif-
fers in some respects from mine.

Also outside the order of the Lepidoptera we may find these
setae, i. a. on Pen'clista melanocephala F. (Tenth redinidae). Like
the setae of Psilura, described by Wachtl and Kornauth, which
I found again in Ocneria and the peculiar elevations of Hetew-
campa (Packard, 1895), I consider them to be rudimentary or-
gans which are disappearing and which now do not possess any
important function for the welfare of their bearers.

I should not be astonished if it were found that the monosetal
tubercula originally have been tactile organs.

In numerous families the monosetal tubercula developed into
warts (verrucae), without it being possible to attribute any sy-
stematical importance to this feature.

In my opinion more importance should be attached to the setae
being plumose or not. As far as I know plumed setae only occur
on those caterpillars which possess verrucae, but not even on all
of them. The only exception known to me is the family of the
Hesperidae of which Fracker says on p. 127 "The head is cove-
red with numerous secondary setae, often plumose but never long,
sometimes borne on chalazae."

139

This case excepted, it seems to me that we can prove that the'fea-
ture of plumose setae has been obtained later than that of verrucae.

If this observation should also be confirmed for other species
than those 1 had at my disjwsal or for thoHc of which I was
able to collect data from the literature, this mif?ht add to a more
aceuratu insight into the phylogeny of the Lepidoittem.

Scudder's opinion that a homogeneous spreading of the setae
over the segment is a primitive ({uality, is decidedly wrong. Na-
ked forms and species with a dense covering of setae have always
arisen from species with a definite sotal |)attern. (Chapter VI
i. a. BombjfjTy Sphinx^ Pieris mtpi).

The verrucae of some families are reduced again to setae. Now
in palaeontology the law of irreversibility holds, which Doli.o (1893)
formulates in the following few words: "The development goes
on with leaps, is irreversible and limited."

In discussing the families I have drawn attention to the fact
that it can si>metimes be seen from the whole pattern, but often
not from the separate setae, whether they have arisen primitively
or by reduction of the verrucae.

In any case therefore for the separate organs we have to do
with a reversible development, a fact which deserves our atten-
tion in connection with the interest which from the palaeontolo-
gical side especially is paid to this problem [compare for instance
the exceedingly clever expositions by D^p^ret (1908) and by his
critic HoERXES (1911)].

The objections which Fracker (1915) makes to my opinion
explained above, do not appear to me to be quite convincing.
For particulars I refer to the Xoctuidae, Chapter VI.

Concerning the separate series of development I can sum up
my results in the following way (Chapter VI):

The Hepialidae differ rather from the FRENATAE, but at the
same time present such important differences from the other
JUGATAE and even amongst themselves, that it is impossible to
fix a definite, strictly circumscribed pattern for this sub-order.
Verrucae occur on the Eriocephalidae.

140

There is very often a s. dorsolateralis on the abdomen.

Of the so-called Micro's the PYRALOIDEA differ from the others
by a slightly altered arrangement of the setae and by the for-
mation of verrueae.

The BOMBYCES seem to be descended from forms with mono-
setal tubercula which are developed into verrueae. In the more
specialized families these verrueae disappear and they are only
distinct during instar /. A reduction of the number of setae is
often to be found.

The Noctuidae too originally possess monosetal tubercula which
are transformed into verrueae and afterwards are again reduced
to simple setae.

The Sphivfiidae i. a. differ by the presence of a s. prostigmalis
and the absence of s. podstigmalis on the abdomen.

The RHOPALOCERA in so far agree with each other that the
primitive setal pattern becomes supplanted during the ontogenesis
by another arrangement of the setae. On the pupa, however,
type I appears again. The presence of verrueae during instar I
of the Papilionidae can be explained as a last remnant of the
dermal armature they formerly possessed.

A comparison with the rest of the orders of insects did not yield
many results. I could find however an indication of a general
groundform which consisted in an arrangement of the setae in
rows of each three on either side (Chapter VII).

I have the impression, that it is under Eimer's (1874, 1889)
influence that Weismann (1876) came to attach such a par-
ticular value to the stripes. Later on Escherich (1892) and
Schroder (1894) advocated the same hypothesis. In opposition
to it J. F. VAN Bemmelen (1889 sqq.) tried to introduce his
opinion, that not stripes but spots compose the primary pattern.
J. Botke (1916) in his studies comes to conclusions which in the
main agree with this opinion. De Meijere (1916) comes to the same
conclusion. J. H. Kruimel also rejects Eimer's hypothesis, after
his study of the feathers of the Gallinae (1916). Tower (1906,
p. 226) says "in ontogeny and in evolution (species foundation)

141

(H)lor appears firet in centres which upon the body are metame-
rioally repeated spots.'*

These writers, however, belong to the few who do not consider
the stripes to be the most primitive element of the pattern.

From my investigation it appears to me that the pigment first
accumulates n)und the bases of the setae or in the verrucae, so
that the primary pattern oonsists of pigmental spots arranged
according to type 1. This {>attern is repeated on each segment
and hence has a metamerical chan»cter.

The strip(>s arise in the ontogeny either simultaneously with
or later than these spots and are therefore a new characteristic.

1 have tried to find a form in which I could trace the development
of a stripe and I think I have succeeiled in Phalera hurephuht (see
p. C5 sq«|.). In instar / the ordinary pattern (type I) is present.
In the cours(> of the development the number of spots increases
a great deal whilst the original pattern gets less distinct. The
secondary spots are situated in vertical rows but by a consolida-
tion of some primary and secondary, spots a horizontal stripe
arises. This stripe, however, is less sharply confined than is usu-
ally the case on caterpillars, so that I am not quite sure whether
all the stripes are developed in the same way. It may also be
that the stripes have suddenly arisen, perhaps as mutations.

It is however a fact, that the pigment spots arranged like
type I, form a phylogenetic element of the pattern which is
older than a stripe.

Under Eimer's influence we have entirely forgotten that a stripe,
i. e. an alteration of a certain p|irt of a segment over the whole
breadth, is altogether a different thing from the series of spots
arranged on the segment in a certain pattern.

A group of spots like this, will bo repeated on all the segments,
because of the strong homoiomery which governs the structure
of the body of caterpillars, but a continuous stripe is quite an
other thing, for it is an alteration of a certain part of the skin
over the total breadth of the segment.

Such an alteration does not happen on the other organs either.

142

It quite agrees with this that a stripe appears later than a certain
pattern of the spots. In the descriptions of the Pieridae and
Phalera amongst others, proofs have been given of the fact that
Eimer's hypothesis does not deserve adherence and that J. F. van
Bemmelen and Tower are right in defending their opinion that
a pattern of spots is more primitive than a stripe. The agree-
ment of the pupal pattern with that of the caterpillar instar I
was proved to be so great, that an accidental agreement is out
of the question. The differences of the pupal pattern with that
of the last larval instar are often so great that there is no pos-
sibility of the pupal pattern resulting from the remaining parts
of the larval hypodermic pigment.

On these facts I have based the theory developed in Chapter
VII, that the pattern of the larval instar / as well as those of
the pupa and imago are primitive characteristics.

The differing armatures of the other larval instars have arisen
from specialisation in connection with the mode of life of the
caterpillars.

The later larval instars have arisen from a retardation of the
development and with it the setal pattern has had the opportunity
of differentiating in various directions. From this we see that I
arrive at the same conception of the larval instars as Deegener
did on the ground of totally different investigations.

Summarizing my results, I come to the following conclusions:

1. The organisation of the thorax is secondary.

2. The anal segments change in number in the various species
of larvae.

3. Originally all abdominal segments were provided with a pair
of legs.

4. In connection with earlier writers a new nomenclature has
been given for the arrangement of the setae : type I, !«, Ih and II.

5. These various types can be derived from each other.

6. A metamerically repeated pattern of pigment spots is more
primitive than a pattern of stripes.

7. The change of setae into verrucae is a reversible process.

143

8. From the agrt»oinent of the pupal p«itteru with inHtar / and
the difference with the lattt larval inntar, the hypothesis has been
developeil, that the pupa and the first caterpillar instar are both
primitive states.

9. The other larval instars arc to be considered as secondary
adaptations.

10. The pupa is to be considered as a subimaginal stage which
secondarily has become immovable.

11. The various caterpillar families have for the greater part
developed themselves independently of or parallel to each oth(>r.

12. A general larval pattern for the Holometabola is ns yet
not to Im» established with certainty.

BIBLIOGRAPHY.

1. Bastin (H.). Insects, their Life-histories and Habits. London 1913.
'<?. Bauer (V.). Zur inneren Metamorphose des Centralnervensystems der
Insect en. S. 132—152, Taf. 8.

Zool Jahrb. Ahl. Anat. Bd. XX, 1904.
3. B e m m e I e n (J. F. van). Over de ontwikkeHng van de kleuren en de
aderen der vUndervleugels in de pop.

Ilandel. v. h. 2e Ned. Nat. en Gcncesk. Com/res, Leiden 1889.

4. Die Entwicklungder Farben und Adern aufden Schmetterlingsflugeln.

Tijdschr. Ned. Dierk. Ver. Deel 3, 1889.

5. De ontwikkeling der Viindervleugels in de pop.

Natuurk. Tijdschr. Ned. Indie, N°. 6, 1890.

6. Ober die Phylogenie der FlUgelzeichnung bei Tagschmetteriingen.

Zool. Jahrb. SuppL Bd. XV, 3 Dd. (Festschrift Spcngel), 1912, mit 1 Taf.

7. On the phylogenetic significance of the Wingraarkings of Hhopalocera.

Transact. 2'"' Entom. Congres, 1912 6. with 3 PI.

8. Die phylogenetische Bedeutiing der Puppenzeichnung bei den Rho-

palocera und ihre Beziehungen zu derjenigen der Raupen und Imagine.s.
Verh. Deiitsch. Zool. Ges. 23. Jahresvers. zu Bremen 1913.

9. Onderzoekingen over de ontwikkeHng van bet kleurenpatroon op de

vleugels der Nymphaliden, Pieriden en Papilioniden en vergelijking met
dat der Hepialiden.

Tijdschr. Ned. Dierk. Ver. 2e Ser. Deel XIV, 1914.

10. Ilet kleurenpatroon van Zelotypia stacyi. Ibidem. 1914 6.

11. Ontwikkeling van het kleurenpatroon op vleugels en lichaam der

Lepidoptera.

Hand. 15e Ned. Nat. en Geneesk. Congres, Amsterdam 1915.

12* '). Over de phylogenetische beteekenis van het kleurenpatroon der

Hepialiden.
Versl. Verg. Wis- en Nat. Afd. Kon. Akad. v. Wetensch. Amsterdam.
Deel XXIV, 1916.

1) The articles marked with aa asterisk are translated into English, in: Proceed,
of the Meetings. Kon. Akad. v. Wet. Amsterdam.

Number 53, 58, 59, 88, 98, 107, 112, 124, 143, 144, 145 and 146 I have not seen
myself, but was obliged to study them after citations from different writers.

145

13. BeuteninUller (Win.). Monograph of the SMiidae of North America
north of Mexico.

Mem. Am. Mut. Nat. HUt. Vol. 1, |>t. VI, 19(X).

14. Boas (J. E. v.). Einige Bemerkungen Qbertlie Metamorphose derlnsekton.

Zool. Jahrb. Abl. System Btl. XII, 1899.

15. Borner (C). Die Verwandlungen der Insekten. Vorl. Mitt.

Sir:. Ber. G«». Naturf. Berlin, Mai 1909.
16' Kotke (J.). Bijdraj^re tot de keniiiti van de phylogenie der vieugel-
teekening bij de Lepidopteni.

Versl. Very. Wis- -n V-i/ ifil Kon. Akait. Wrt. Amsterdam, Deel
XXIV, 1910.

17. Les motifs priiniiii.s >iii licssin dee ailes des L^pidopt^res et leur

origine phyletique.

Tijdschr. Nett. Dierk VVr. 2de reeks, Deel XV, 1916 en Onderzoe-
kimjen ivrrichl in het Zool. Laborat. der Hijks'Universiteit le
Gronitigen, N°. V.

18. Brauer (F.). Entwicklungsgeschichte der IVinorpa communis.

Si/?. Ber. d. K. K. Akcul. d. Wiss. W'ien. Bd. VII, 1851.

19. Cber die Lanre von Panorpa communis. <

Verh. d. Zool. Bot. Ver. Wien, Bd. I. 1852.

90. Beitrftge lur Kenntnis der Panorpiden-Larven.

Verh. d. Zool. Bot. Get. Wien, Bd. XIII, 1863.

21. System. Zool. Studien 2. Die unvermittelten Reihen in der Classe

der Insecten.

Siiz. Ber d. K. K. Akad. d. Wiss. Wien, Math.-Nat.w. Classe,
Bd. XCI, 1884

22. Brongniart (C). Recherches pour servir a I'histoire des Insectes fos-
siles des Temps Priraaires. St. Elienne 1894.

23. Bryk (Fel.). Ein Citronenblatt mit einer ursprtinglicheu Weiszlings-
leichnung.

Zool. Anzeiger, Bd. XLIV, 1914.

24. Buckler (W.). The larvae of the British Butterflies and Moths.

Edited hy H. F. Stainton, Vol. I— IX, London, Ray Society 1886— 1899.

25. B u g n i o n (EI.) Resume des recherches de M. J. Gonin sur la meta-
morphose des Lepidopt^res.

Bull. d. I. Soc. Entom. Suisse, Vol. VR, 1892.

26. Busck (Aug.). Notes on Microlepidoptera. with description.s of new
North American species.

Proc. Ent. Soc. Wash. Vol. XI, 1909.

27. On the classification of the Microlepidoptera.

Ibidem, Vol. XVI, 1914.

28. Carpenter (G. H.). The life-story of Insects.
Cambr. Manuals of Sc. and Lit. 1913.

29. Chapman (T. A ). On some neglected points in the structure of the
Pupae of Heteroceroiis Lepidoptera and their probable value in classifi-
cation, with some as.sociated observations on larval prolegs.

Trans. Ent. Soc. Land. 1893—1894—1896.

10

146

30. Chapman (T. A.). On a lepidopterous pupa (Micropteryx purpurella)
with functionally active mandibles.
Trans. Enl. Soc. London, 1893 h.

31. Note.s on Butterfly-Pupae, with some remarks on the Phylogenesis

of Rhopalocera.

Entom. Record. Vol. VI, 4895.

32. Notes on pupae.

Trans. Enl. Soc. 18966.

33. On the phylogeny and Evolution of the Lepidqptera from a pupal

and oval standpoint.

Trans. Ent. Soc. London, 1896c.

34. The cla.ssification of Gracillaria and allied Genera.

Eniomolo(jist, Vol. XXXV, 1902.
35. Com 8 took (J. H.). Evolution and Taxonomy. An essay on the appli-
cation of the natural selection in the classification of animals and plants,
illustrated by a study of the wings of insects and by a contribution to
the classification of the Lepidoptera.
Ithaca N.Y. 1893.
• 36. Cramp ton (G. C). The Groundplan of a typical thoracic segment in
Winged Insects.
Zool. Anzeig. Bd. XLIV, 1914.
37. Deegener (P.). Die Entwicklung des Darracanals der Insecten wahrend
der Metamorpho.se.

Zool. Jahrb. Alt. Anal. Bd. XX, 1904.

38. Metamorphose der Insekten.

Tcubner. Leipzig 1909.

39. Dep6ret (Ch.). Les transformations du Monde animal.

Paris Tme Mille 1908.

40. Dewitz (II). Beitrage zur posterabryonalen Gliedmassenbildung bei
den Insekten.

Zeils. Wiss. Zool. Bd. XXX. Suppl. 1887.

41. Dixey (Fred. A.). On the Phylogeny of the Pierinae as illustrated by
their wingmarkings and geographical distribution.

Trans. Ent. Soc. London, 1894.

42. D u p 0 n c h e 1 (M. P. A. J.). Iconographie et Histoire naturelle des Chenilles.

Tome I, II, Paris, 1849.
D y a r (11 arr. G.). Numerous articles in Canad. Entom. 1893 and in Psyche.
43. A classification of Lepidopterous larvae.

Ann. N.Y. Acad, of Sc. Vol. VIII. 1894.
44. Additional Notes on the Classification of Lepidopterous Larvae.

Trans. N.Y. Akad. of Sc. Vol. XIV, 1895.
45. A Classification of Lepidoptera on Larval Characters.

Am. Nat. Vol. XXIX, 1895 b.
46. Dyar (Harr. G.). Relationship of Pyralidae and Pterophorina from
the Larvae.

Enl. News. 1895.

147

47. Djar (Harr. G.). On the larvae of the higher Bombyow.

Proc. Boat. Soe. Nat. Hist. Vol. XXVII, 1896.
48. Life Historiee of New York Slug-caterpillars.

Journ. NY. Ent. Soc. Vol. VIII, 1899.
49. Note on two Hydroeciu larvae.

Journ. N. Y. Ent. Soc. Vol VIM. 1899.

1900 See Beutenroailei

50. A Century of Larval l)esciiplioiis>.

Entomologiata Recvitl. Vol. XIII, 1901.

51. A list of North Anoerican Lepidoptera.

Bull. 52 U. S. Nat. Mm*. 1902.

52 A deecriptive list of a collection of early stages of Japaneae Lepi-
doptera.

Smiths, hutil. Proc. U. S. Nat. Hist. Mus. Vol. XXVIII, 1905.

53. The Pericopid Larvae in the National Museum.

Insecutor Inacitiae Menstruus, Vol. II, 1914.
54. Eimer (Th.) LaoarU muralis coerulea. ZooL Studien auf Capri U.
Leiptig 1874.

55. Die Artbildung und Verwandtschafl bei den Schmetterlingen.

Jena 1889.
50. Escherich (K.). Cber die Gewtimkssigkeit im Ab&ndern derZeiclmung
bei Inaekten.

Deutsche Enlotn. Zcitsch. 1892.

57. Felt (E. P.). The Skorpion- Flies.

Ilej}. State Entom. N. Y. Vol. X, 1895.

58. Forbes (W. T. M.). A structural study of some Caterpillars.

Ann. Entom. Soc. Am. Vol. Ill, 1910.

59. A structural study II. The Sphingidae. '

Ann. Entom. Soc. Am. Vol. IV, 1911.

60. Fracker (Stan I. B.). The Classification of Lepidopterous Larvae.

lUin. Biol. Monogr. Vol. II, 1915.

61. Frohawk (F. W.). Natural History of British Butterflies.

The Field. The Country-Gentleman's Newspaper 1914.
See also: The Entomologist; numerous articles.

62. G a n i n (M.). Materiaux pour I'histoire du d^veloppement posterabryonaire
des Insectes.

4°. 76 pg. 4 pi. Varsovie 1876.

63. G 0 n i n (J.). La metamorphose des Lepidopteres.

Bull. Soc. Vaudoise d. Sc. Nat. T. XXX, 1894.

64. Grote (A. .Red cliff). Die Saturniiden.

AJitt. a. d. Roemer Museum. Hildesheim, N°. 6, 1896.

65. G r u b e r (A.). Ub. Nord-Amerikan. Papilioniden- und Nyraphaliden-Raupen.

Jena Zeitsch. Bd. XVII, 1884.

66. Handlirsch (A.). Zur Phylogenie der Hexapoden.

Sitz. Bei'. K. K. Akad. d. Wiss. Wien, N. W. Abt. Bd. I, CXIF, 1903.

148

67. Handlirsch (A.). Die fossilen Insekten und die Phj'Iogenie der rezenten
Formen. Ein Ilandbuch f. Palaeont, u. Zool. Leipzig, 1906—1908.

68. Einige interessante Kapittel der Paiao-Entomologie.

Verh. Zool. Bol. Ges. Wien, 1910.

69. H eider (K.). Die Entwickinng des Hydrophilus piceus. Jena, 1889.

70. Henneguy (L. F.). Les Insectes. Paris, 1904.

71. Heymons (R). Die verschiedenen Formen der Insectenmetaniorphose
und ihre Bedeutung im Vergleich zur Metamorphose anderer Arthropoden.

Ergebn. u. Forlscht'. d. Zool. Bd. I, 1907.

72. Hoernes (R.). Das Aussterben der Arten und Gattungen. Graz, 1911.
Hofmann (E). 1893, see Spui.er.

73. Ilofmann (0.). Uber die Anordnung der borstentragenden Warzen
bei den Raupen der Pteroplioriden.

111. Zcilschr. j. Enlom. Bd. Ill, 1898.

74. Beobachtungen iiber dieNaturgeschicbteeiniger Pterophoriden-Arten.

Ibidem.

75. Hors field (T.) and F. Moore. A catalogue of Lepidopterous Insects
in the Museum of Nat. Hist, at the East-India House. Vol. I, II, London,
1857— '59.

76. H (i b n e r (Jacob). Geschiclite Europ. Schmetterlinge. Bd. I — IV,
Augsburg, 1786—1791.

77. Jackson (W. Hatchet). Morphology of Lepidoptera. ,

Tram. Linn Soc. 2"'' Series. Zool. Vol. V, 1890.

78. Janet (Ch ). Sur I'ontogenese de I'insecte. Limoges, 1909,
Kornauth see Wuclitl.

79. K 0 w a I e w s k y (A ). Embryol. Stud, an Wiirmern u. Arthropoden.

Mem. Acad. Imp. St. Pelersboicrg, T. XVI, 1871.

80. K r u i m e I (J. H.). Onderzoekingen over de veeren bij hoenderachtige
Vogels.

Disserlatie,, 1910, Amsterdam.

81. Kiimmeth (F.). I>ie Stigmenversorgung des Insektenthorax.

Zeits. Wiss. Zool. Bd. CXII, 1914.

82. Lameere (Au g.). La raison d'etre des metamorphoses chez les Insectes.

Ann. Soc. Enl. Belg. T. XLIII, 1900.

83. Landois (H.). Beitrage z. Entw. gesch. d. Schmetterlingsflagel in den
Raupen und Puppen.

Zeits. Wiss. Zool. Bd. XXI, 1871.

84. L i n t n e r (J. A). Second Annual Report on the injurious and other
insects of the State New York. Albany, 1885.

85. Lubbock (Sir John). Origin and Metamorphosis of Insects. 1871.

86. L i j 0 n e t (P.). Traite anatomique de la chenille qui ronge le bois de saule.

La Haye, 1760.

87. Malpighi (M). Dissertatio epistolica de Bombyce.

Opera Omnia. Lugd. Batav. 1687.

88. Mayer (A. G.). The development of the Wing-scales and their pigment

in Butterflies and Moths.
Bull. Mus. Comp. Zool. Vol. XXIX, 1896.

149

89. Mercer (W. F.). The development of the Wings in the Lepidoptei-u.

Journ. N.Y. Ent. Soc. Vol. VIII, 1900.

90. liesDil et Metschnikoff. Qiielquee remarques au sujet dii deter-
miiUBine de la metamorphose.

C. n. Soc. ISioL, Pari*. T. LII, 1900.

91. Meyere (J. C. U. de). Cber die Haare der S&ugetiere, bettonders Uber
ihre Anordnung.

Morph. Jahrb. lid. XXI, 1894.
92. Teekening der Dipteren-vleugeU.

T^iischr. V. Kntvmologie, Deel LVllI, 1915.

Zur Zeichnung dee Insekten-, im beeonderen des Dipteren- uiid

LepidopterenflOgels.

Tijilachr. v. Enfomologie Deel LIX, 191C.

93. M i a 1 1. Transformations of Insects.

Nature 1895.

94. M i 1 1 i ^ r e (P.). Iconographie et Deecription de Chenilles et Lepidoptires
inedits Vol. I, II. 1858—1868.

tjrir. d. Ann. d. I. Soe. Linn, de Lyon. N. S. Tome V.

95. M Q 1 1 e r ( W i I h.). SOdamerikaniacbe Nymphalidenmupen.

Zool. Jahrb. Bd. I, 1886

96. Oudemans ^A. C). Die gegenaeitige Verwantschaft, Abstammung und
Classification der sogenannten Arthropoden.

Tijdachr. Ned. Dierk. Ver. 2e Serie, Deel I, 1886.

97. Oudemans (J. Th.). Nederlandsche Insekten.

Thiemo, Zutphen, 1897—1900.

98. Packard (A. S.)- Guide to the study of Insects. (9lh Ed.) 1889.

99. Hints on Evolution of bristles, spines, tubercles of certain caterpillars.

Proc. Bout. Soc. Nat. HUt. Vol. XXIV, 1890.
100. Notes on the phylogeny of lepidopterous larvae.

Ibidem Vol. XXV, 1891.
101. Phylogeny of lepidopterous larvae.

Journ. Roy. Microsc. Soc. London, 1891 b.

102. Studies on the life history of some Bombycine Moths.

Ann. N. Y. Akad. of Sc. Vol. VIII, 1893.
103. Monograph of Bombycine Moths, I. Notodontidae.

Mem. Nat. Acad. Sci. Vol. VII, 1895.
104. Idem II. Ceratocampidae.

Ibidem. Vol. IX, 1905.
105. Idem III (posthumous) Ceratocampidae, Saturnidae, Heraileucidae.

Ibidem. Vol. XII, 1914.
see numerous articles in Psyche, Americ. Naturalist, etc.

106. Pancritius (P.). Beitrage zur Kenntniss der Fliigelenlwicklung bei
den Insecten.

Inaug. Diss. 1884.

107. Perez. Signification phyletique de la nyrapbe chez les Insectes metaboles.

Bull. Sc. Franco-Belg. 7e Serie. T. XLIV, 1910.

150

108, Pie per s (Dr. Jur. M. C). Uber die Entwicklungsgeschichte einiger
Javanischen Papilioniden-Raupen.

Tijdschrift v. Enlomologie Deel XXXI, 1888.
109. Over den hoorn der Sphingiden.

Ibidem, Deel XXXII, 1889.
110. Ober das Horn der Sphingiden-raupen.

Ibidem, Deel XL, 1897.

111. Plate (L.). Prinzipien der Systematik m. bes. Berucksicht. d. Syst.
d. Tiere.

Die Kult. d. Gegenwart, T. 3, Abt. IV, 4. Teubncr 1914.

112. Poyarkoff (E.). Recherches histologiqu6s sur la metamorphose d'un
coldoptfere.

Arch. d'Anat. microsc. T. XII, 1910.

113. Poulton (Edw. B.). Notes in 1887 upon lepidopterous larvae.

Transact. Ent. Soc. London, 1888.
114. The colours of Animals.

The Intern. Scient. Ser. Vol. LXVIII, 1890 a.

115. The external Morphology of the Lepidopterous Pupa; its relations

to the other Stages and to the Origin and History of Motamoiphosis.

Trans. Linn. Soc. London, 2nd Ser. Zool. Vol. V, 1890—91.
IIG. Quail (Ambrose). Life Histories in the Ilepialid Group of Lepidop-
tera, with description of one new Species and Notes on Imaginal structure.

Trans. Ent. Soc. London, 1900.
117. Notes on Cossidae.

Entomologist, Vol. XXXVII, 1904.
118. On the tubercles of thorax and abdomen in first larval stages of

Lepidoptera.

Ibidem 1904.

119. Ratzeburg (J. T. Ch.). Die Forst-Insecten. Th. I u. II. Berlin, 1840.

120. Reaumur (M. d e). Memoire.s pour servir a I'liistoire naturelle et a
I'anatomie des Inseotes. Paris, 1736 — 1742 et Amsterdam, 1737.

121. Rees (J. van). Beitrage zur Kenntniss der inneren Metamorphose von
Musca voraitoria 134 S. 2 T.

Zool. Jahrh. Bd. Ill, 1888.

122. Riley (C. V.). The Philosophy of Pupation.

Am. Entomol. Vol. HL 1880.

123. Note on the eversible glands of larvae of Orgyia and Parorgyia.

124. Sasaki. On the affinity of our wild and domestic silkworms.

Annotat. Zool. Japon. 2 Pt. Vol. II, 1898.
125*. Schierbeek (A.). Over het setale patroon der Rupsen.

Versl. Verg. Wis. Nat. Afd. Kon. Akad. v. Wet. Amsterdam,
Deel XXIV, 1916.
126. Schroder (Chr.). Die Entwicklung der Raupenzeichnung und Abhan-
gigkeit der letzteren von der Farbe der Umgebung.

Inaur. Diss>. Berlin, 1894.

151

127. Schulie (P.). Die Nackengabel der Papilionidenraupen.

Zool. Jahrb. Abl. f. Anal. ltd. XXXII, 1912.

128. Sc udder (S. H.). A systematic Review of our present kuowledge of
loHil Insects, including Myriapods and Arachnids.

Bull. U.S. Geol. Sut-v. Vol. XXXI, 1886.

139. The butterflies of the Eastern United States and Canada with

special reference to New-England. 1888 — 1889.

130. Sepp (J. C). Beechouwingen der wonderen Gods in de roinst geacbte
schepzelen of Nederlandticlie Insecten. Amsterdam, 1702.

131. Sharp (D ). Insects I, II. Cambridge, 1901.

132. Siltala. Trichopterologische Untersuchungen 2.

Zool. Jahrb. Suppl. Hd. IX, 1907.

133. Spuler (A.). Zur Phylogenie und Ontogenie des FlUgelge&ders der
Schmetterlinge.

Zeitschr. Wiss. Zool. Bd. LIU, 1892.

134. Die Schmetterlinge u. Raupen Europa's. Stuttgart, 1910.

(3e Aufl. V. E. Hofmann, Die GrossachmcUcrlinyc).

135. Stobbe (R.). Die abdominalen Duflorgane der miinnlichen Sphingiden
und Noctuiden.

Zool. Jahrb. Abl. f. Anat. Bd. XXXII, 1912.

136. Swammerdam (J). Historia Insectorum generalis. Lugd. Bat. 1685.
137. Bijbel der Natuure. Diblia Naturae. V4. J. Boerhaave, 1737.

138. Tichom irof f (A.). Ober die Entwickelungsgeschichte dea Seiden-
wurms.

Zool. Anz. Bd. II, 1874.

139. Tonge (A. £.). Some Moths and Butterflies and their Eggs. 1907.

140. Tower (W. L ). The development of the colors and coiorpatterns of
Coleoptera, with observations upon the development of color in other
orders of Insects.

The decennial publ. of the Univ. of. Chicago. Iser. Vol. X, 1905.

141. An Investigation of Evolution in Cbrysoraelid Beetles of the

Grenus Leptinotarsa.

Cartu Listit. Publ. N°. 48, 1906.

142. Tsou (Y. H.). The body setae of Lepidopterous Larvae.

Trans. Am. Microsc. Soc. Vol. XXXUI, 1914.

143. V 0 s 8 (Fr. v.). Cber den Thorax von Gryllus doraesticus.

Zeits. Wiss. Zool. Bd. C, 1911.

144. Ibidem, Bd. CI, 1912.

145. Vergleichende Untersuchungen uber die Flugwerkzeuge der Insekten.

Verhandl. deutsch. Zool. Ges. 23. Vers. 1913.

146. Wachtl u. Kornauth.

Mitt, forst. Versuchwesens Oeslerreichs. Bd. XVI, 1893.

147. Walter (A.). Beitrage z. Morphologie d. Schmetterlinge.

Jenaisch. Zeilsch. f. Naturw. Bd. XVIII, 1885.

148. W a 1 s i n g h a m (Lord). Biologica Centrali-Americana. Lepidoptera
Heterocera. 1911.

152

149. We is m an n (Aug.). Die Entwicklung der Dipteren im Ei.

Zeitsch. Wiss. Zool. Bd. XIII, 1863.
150. Die nachembryonale Entwicklung der Musciden.

Ibidem. Bd. XIV, 1864.
151. Die Metamorphose d. Corethra plumicornis.

Ibidem. Bd. XVI, 186C.
452. Studien zur Descendenztheorie. Leipzig, 1875 — 1876.

(Translated into English by Meldola).
II. Uber die letzten Ur.sachen der Transmutationen.

1, Die Entstehung der Zeichnung bei den Schmetterlingsraupen.

2, Ober den phyletischen Parallelismus bei metarnorphischen Arten.

153. Wood (J. H.). Notes on the earlier stages of Nepticulae.

The Ent. Month. Magaz. 1894.

154. Z i t te 1-B r 0 i 1 i. Grundzikge der Palaeontologie. Bd. I, Munchen, 1915.

EXPLANATION OF THE PLATES.

PLATE I. Synopsis of Nomenclature, 1886-1916.

Fig. 1. Myacelia orsit (Nymphalidae). Instar /, Immediately before rooultin^.
To show the arrangements of the primary setae and the place of
the secondary scoli. After W. MCller (1886, Taf. 3, Hg. 14).

» 2 and 3. Acraea pellenea. HObn. (Nymphalidae). Instar /. To show
the prinoary setae on the metathorax and the 2nd and 3rd abdo-
minal aegmenU, after W. MtlLLER (188C, Taf. 1, flg. 1).

» 4. Hepialus lupulinus, after H. G. Dyar (1894, p. 197). Observe the
three setae above the stigma.

» 5. The arrangement of vomicae is of the "Arctian type", marked
according to H. G. Dvar's system (1894, p. 198).

» 6. An abdominal segment of a Psychid larva. Adapted from a figure
by H. G. DvAR (1884, p. 198). Observe the three setae above the stigma.

» 7. Thoracic scheme, marked according to Dyar's system (1894 b).

t 8. Thoracic scheme after 0. ElomiNN (1898). The subprimary setae are
marked with an asterisk. H. G. Dyar himself agreed with this
system in 1901. The difTerences between his opinion at this date
and that of 1894 are given in Roman cyphers.

» 9 and 10. Melanchria nutans. (Noctuidae). Instar 11. The setae on a
metathoracic and an abdominal segment, after the ideas of A. Quail
(1904 b). Mark seta III H.

» 11 and 12. Metathorax and abdomen with primary setae, according to
the system of W. T. M. Forbes (1910), cited by St. B. Fracker (1915).

» 13. Pieris brassicae L. Instar V. The rows of pigment-spots, with the
names given by J. F. van Bemmelen (1913, p. 115).

» 14. Hepialus humuli. Metathorax and 1st abdominal segment of a
mature larva. Adapted from a figure by Y. H. Tsou (1914, Pi. X,
fig. 1 c, rf)- Compare figure 4, 22, 23, 24. 25 of this plate.

» 15. Hypothetical type showing twelve primary setae. The three usual
subpriraaries are dotted in. The spiracle is shown in both protho-
racic (thor.) and abdominal (abd.) positions. After St. B. Fracker
(1915, PI. I, fig. 1).

154

Fig. 16. Atteva aurea {Yponomeutidae). Metathorax. To show the thoracic

setae of a typical Micro. After St. B. Fracker (1915, PI. V, fig. 36).
» 17. Feltia glandaria (Nocluidae). Instar /. 6th abdominal segment.

Marked according to St. B. Fracker's system (1915, Pi. IV, fig. 29).

Note seta p in this figure and in figure 15.
» 18. Mesothoracic and raetathoracic scheme, according to my view.

Type II. Note the wing-rudiment.
» 19. The primary setae on a typical abdominal segment. Type I.
» 20. Type la, the usual arrangement of the Saturniidae.
» 21. Type 16, the armature of some Lymantridae.

For fig. 18, 19, 20, 21 see Chapter V, p. 42 sqq.

» 22. Hepialus hectus Linn. Instar i, dorsal aspect,

» 23. » » » » », lateral aspect.

» 24, » » » » », ventral aspect.

» 25. Hepialus of. lupulinus Linn. Instar F, lateral aspect. Compare lig. 4

and 14.
» 26. Zeuzera pyrina Linn. (Cossidae), mature larva. Coll. Kall. Pro-,
meso- and metathorax, and abd. 1, 2, 3, lateral aspect.
» 27. » » Pro-, meso- and metathorax, abd. 1, dorsal aspect.

PLATE II.

» 1. Bombyx mori. Instar/, newly hatched, dorsal aspect.
» 2. » » » » » » lateral aspect.

» 3. » » » » just before moulting, lateral aspect.

» 4. » » » II, just after moulting » •»

Mark the numerous secondary setae.
» 5. » » Abdominal segment 5 of mature larva.

Obsei've the primary verrucae between the secondary setae.
» K)a, b. Lasiocatnpa riibi {Lasiocampidae). Prothorax, mesothorax and

abd. segment 5 of Instar HI.
» 1. » » Abdominal segment 5 of mature larva.

» 8. Ocneria dispar (Liparidae). Instar i, dorsal aspect.

» 9. » » » » » lateral aspect.

» 10. » » » » » the curious setae, which dis-

appear after moulting.

» 11a, 6. » » » mature larva. Prothorax and 5th ab-

dominal segment (see Type I b,
Plate I, fig. 21).

» 12. Porlhesia chrysorrhoca {Liparidae) Instar i, dorsal aspect.

»13. » » » »» lateral aspect.

» 14. » » » mature larva. Pro- and mesothorax.

» 15. » » » » » 6th abdominal segment.

155

PLATE III.

Fig. 1. Onjyia antitpia (Liparitlae). Instur /, ventnil a»i»ect.
» 2. » • » » A lateral asiiect.

» 3. » » » »//,»»

» A. • • • » ///, • »

» 5. > » > » IV, pro-, mew- and metathoiux,

abdominal segm. 1.
» 6. » > > > v. mature larva, partly urtor

J. HObner (1786).
Fig. i — C show the ontogenv ut the plumose setae and Type lb.

7. Phalera{Pyijaeitt)bucephala{NottHiuulidae). Instar /, lateral aufwct.

8. • » » Instar i, dorsal aspect.
U. » » » » • ventral aspect.

10. » » » » /i, lateral aspect.

11. » » » » ///, prothorax, mesothorux, ubd.

segment 5.

12. • • • • IV, pi-othorax, abd. segment 5.

13. » > » » V\ mature larva, abd. segment 5.
Fig. 7 — 13 show the development of strii)es.

14. Saturnia jtatonia, lostar 1. Type la.

15. • » * V, mature larva, 5th abd. segment.

PLATE IV.

1. Sphinx liguslri. Instar /, lateral aspect.

2a, b. Smerinthus tiliae. Instar /, lateral aspect.

3a, b. Smerinthus populi.lnst&v I, » »

Fig. 1—3 show the primary setae and the secondary forked ones.
Observe seta prostigmalis..
4. Arctia caja. Instar /, lateral aspect.
5a, b. » » » III, prothorax and 5th abdominal segment.

6. Ocnogyna lubricipeda (Arctiidae), mature larva.

7. Euchelia jacobaea. (Arctiidae) Instar /?, lateral aspect.
Fig. 4 — 7 show the development of the plumed setae and the

occurrence of verrucae and primary setae in this family. Cora-
pare Plate I, fig. 5.

8«, b. Mamestra brassicae {Nocluidae), prothorax, raesothorax and 5th

abdominal segment of raatnre larva.
9rt, b. Acronycla psi (Noctuidae), metathorax, Ist and 8th abdominal
segments of mature larva.
Observe the origin of the fleshy horns from v. dorsalis and v. subdorsalis.
10. Depressaria nervosa {Noctuidae). Coll. KALf.. Note the arrangement
of the pigment-spots, according to Type I and
the absence of setae.
Compare with fig. 8, 9, 10, PI. I, fig. 9, 10, 17.

3.

»

4.

»

5a,

by>

Ga,

b»

7.

»

156

Fig. 41. Thyris fenestrella, mature larva. Coll. Kall., dorsal aspect.
» 12. » » » » » » ventral aspect.

* iS. » » » » » » lateral aspect.

» 14. Vanessa urlicae (Nymphalidae). Instar /.
"15. » » mature larva, 5th abd. segment.

PLATE V.

» 1. Pieris brassicae. Instar /, newly hatched, dorsal aspect.
» 2. » » » » immediately before the moulting, lateral

aspect.
» » II, partly dorsal, partly lateral aspect.

» » 7//, lateral aspect.

> » /K, pro-, meso- and metathorax, 1st abd.

segment ; abd. segm. 2, 3, 4, 5, G.
» » ^) pro-» meso- and metathorax, abd. segm.

5 of mature larva.
» Pupa, abdomen.

Compare Plate I, fig. 13. Observe the development of the chalazae
from the primary setae, the origin of the stripes, the reduction
of the pupal pigment-spots in comparison with the mature larva,
and V. dorsolateralis on the pupa.

» 8. Pieris napi. Instar I.

» 9. » » w II.

» 10a, 6» » » ///, pro- meso- and metathorax, 5th abd. segm.

» 11. » » D IV, abdom. 5.

» 12. » » » V, abdom. 5 of mature larva.

9 13. » » Pupa, abdomen.

Observe the development of the secondary setae and the disappea-
rance of the primitive arrangement.
The pupa has the same pattern as Instar /, with a well developed
s. dorsolateralis. Note the increase of the pigment-spots in com-
parison with the mature larva.
» 14. Scricinus telamon (Papilionidae), mature larva.

Coll. Kali. Pro-, meso- and, metathorax, abdom. 1.
Observe the osmaterium on the place of v. dorsalis.

CONTENTS.

Page.
Chapter I. Intixxiuction, Maienal and Melhcxl 1.

$ 1 Introduction 1.

S S Material and Method 3.

Chapter 11. Literature G.

Chapter 111. On the Structure of the Thoracic Segments 20.

Clmpter IV. On the Number of the Segments and on the Abdominal Legs. 27.

Chapter V. Nomenclature and Primitive Pattern 30.

Chapter VI. Systematic Synopsis of the Setal Pattern of Caterpillars . 47.

(The following caterpillars are described in detail and
figured on Plate 1— V).

tiepialus hectus 50.

Hepialtts spec, cf luf»illinit 52.

Coasus co$au». . 57.

Zetuera pyrina 58.

Pygaera bueephata 05.

Lymantria dispar 71.

Euproctia ckrytorhoea 73.

Orgyia antiqua 77.

Leuiocampa rubi 83.

Bombyx mori 85.

Salurnia pavonia 91.

Sphitur ligustri 94.

Smerinthus titaie 96.

Smerinthus populi 97.

Hypocrita jacobaea 99.

Arctia caja 99

Spilosoma lubricij)eda 101.

Nonagria nervosa 103.

Mamestra brassicae 103.

Actxtnycta psi 103,

Thyris fmestrella 107,

Pieris brassicae 111.

Pieris napi 115.

Vanessa urticae 119,

Sericinus telamon 121.

Chapter VII. Comparison between the Patterns of Caterpillars and those

of the Pupa, the Imago and other Larvae of Insects. 122.
Chapter VIII. General Considerations and Synopsis of the Results. , . 134.

Bibliography 144.

Explanations of Plates 152.

2t>

23

24

25

'2(i

i2

d3

//mm

9a

10

II

12

13

14

J

Z fft.«t

/mm.

14

STELLINGEN.

STELLINGEN.
I.

De excreti(M>rganen van Amphioxus zijn geen protonuphri(lir>n.

II.

I)t' l)ewt'ring van von Hk-/., tlat de bijen kleurenblind /ijii, is
door VON Frisch afdoende weerlegd.

III.

De meening van Dickel, dat jonge bijenlarven hermaphrodiet
zijn en door de voeding veranderd kunnen worden in darren of
wijfjes, is onjuist.

IV.

De redenen, door Fabre aangevoerd om de doof heid der Cicaden
aan te toonen, zijn geen bewijzen.

V.

De laterale kelkbladeren der Cruciferen zijn de buitenste.

VI.

De alcoholgisting van Saccharomyces is slechts voor een gering
gedeelte toe te schrijven aan een zuivere fermentwerking.

VII.

Het voorkomen van chitine bij verscliillende planten wijst niet
op een systematische verwantschap.

YIII.

Do bewering van Steinmanx, dat de Delphinidae afstammen van
do Ichthyosauria, is ongegrond.

IX.

De wet van de toeneniing der lichaamsgrootte is in strijd mot do
waarneming.

X.

Drijftillen zijn gunstig voor de veenvorming.

XI.

Het is wenscholijk, dat zoo spoedig mogelijk het ius promovendi
aan de bezitters van eon einddiploma eener Hoogere Burgerschool
mot vijf-jarigen cursus wordt verleend.

XII.

Het normaalprogramma voor de Rijks Hoogere Burgerscholen mot
vijfjarigen cursus is in strijd met de wet van 1863, art. 17t\

XIII.

Bij het onderwijs op een Hoogere Burgerschool met vijfjarigen
cursus dientde historische geologie door den leeraar in de natuur-
lijke historic, de algemeene geologie door den loeraar in do aard-
rijkskundo to worden bohandold.

XIY.

Het is wenscholijk do hoofdpunten der physiologie en der mikros-
kopischo anatomie van planten on dioren op een Hoogere Burgerschool
met viifiarigen cursus to behandelen.