Contributions to Zoology, 67 (3) 197-220 (1998)

SPB Academic Publishing bv, The Hague

of theoretical Analysis a species/instars/characters table: a survey on the use

of chaetotaxy in ontophylogenetic studies

Pierre Nayrolles

Laboratoire de Zoologie, Ecobiologie des Arthropodes édaphiques, Université Paul Sabatier, 118 route de

Narbonne F-31062 Toulouse Cedex France. E-mail: , [email protected]

Keywords: Ontogeny, phytogeny, chaetotaxy, probabilistic organs, ontogenetic trajectories, arthropods,

Collembola, Symphypleona

Abstract

opposant présences et absences. Dans une population, une

soie donnée peut se révéler variable, une probabilité de présence studies of Phylogenetic on several groups arthropods, such as pouvant alors lui être attachée. Une soie variable pourrait être Acari (mites) or Collembola (springtails), make wide use of al- perçue comme la marque d’un polymorphisme (plusieurs chaetotaxy. Chaetotaxic characters, besides being morpho- lèles dans une population). En fait, cette variabilité doit être logical features (setal shape), often have a binary nature, that comprise comme la sommation de propensions individuelles, A be variable is presence vs. absence. seta can in a popula- ou en d’autres termes, comme la résultante d’une potentialité and to this seta. tion, one may attribute a presence probability à individu hasard. propre chaque exprimée au Vraisembla- In the for fact, presence probability should be defined each blement, un phénomène similaire à celui de l’inhibition laté- instar. One might think that a variable seta corresponds to a rale, phénomène qui explique comment une fluctuation au polymorphism (e.g., two or more alleles in a population); in hasard intervenue précocement peut déterminer le destin de of fact, setal variability should be regarded as the result a à des Un cellules, est l’origine organes probabilistes. tableau propensity intrinsic to individuals, i.e., a potentiality being tridimensionnel espèces/stades/caractères contenant les pro- random known expressed at among specimens. The phenomenon babilités de présence des soies peut être constitué. Deux sortes as lateral inhibition explains how an early random fluctuation de traitements, dans lesquels le problème des soies homéo- is of at the root the cell fate. Probabilistic organs are likely to types prend une place essentielle, sont alors applicables. Les from similar originate a phenomenon. A tridimensional spe- autres questions relatives à ces analyses sont la pertinence cies/instars/characters table (SIC table), containing setal pres- de l’utilisation des probabilités de soies et la réduction de be Two kinds of ence probabilities, can built up. analyses l’information redondante. Un premier traitement conduit à un may be applied to such a table. In these analyses, the problem tableau bidimensionnel dont les caractères sont en colonnes of homeotypic setae is emphasized. Related issues are the divisées et, en lignes, les espèces en stades. En utilisant des of setal and the of accuracy using probabilities problem re- statistiques multidimensionnelles, il est possible de dessiner of information. A first ducing redundancy process leads to a des et donc de les trajectoires ontogénétiques, comparer on- bidimensional in table with characters columns, and the spe- togenèses. Le but du deuxième traitement est l’obtention d’un cies divided into their instars in rows. By using multidimen- tableau être utilisé dans espèces/caractères qui pourra une sional statistics we can access ontogenetic trajectories, and in analyse cladistique. this way, ontogenetic comparisons can be achieved. The aim

of the second process is to produce a species/characters table

which can be used in a cladistic analysis.

Introduction

Résumé that several authors Andre (1988) emphasized were

surprised that so little is known about insect im- Dans l’étude de phylogénétique plusieurs groupes d’Arthro- mature stages. He explained this fact by the lack tels les podes, que Acariens ou les Collemboles, il est large- of a method for the treatmentof ontogenetic data. ment fait appel à la chétotaxie. Outre le fait qu’ils puissent The aim of this as well as those Andre, être des traits purement morphologiques (forme des soies), paper, by

les caractères chétotaxiques sont souvent de nature binaire, is to provide some theoretical ideas and several

Downloaded from Brill.com10/10/2021 04:06:52PM via free access 198 P. Nayrolles - Chaetotaxy in ontophylogenetic studies

techniques for studying arthropod ontophyloge- merit of mites and distinct from each other by “all

netics (i.e., ontogenetics and phylogenetics con- or none” characters.

sidered together) by means of chaetotaxy, i.e., The discrete characters used for differentiating

the of the of - based the of study arrangement organs setae or, stases are generally on presence se-

- considered this strictly speaking, sensilla produced by the in- tae. The characters in paper con-

tegument of arthropods. Once chaetotaxic pat- cern only idionymic setae. Idionymy (Grandjean,

have absence is the of terns been established, presence vs. 1949) “quality” a particular organ as or number of setae can be used to compare spe- distinct from other organs of the same nature;

defined cies. such an organ occupies a position so that

A general scientific procedure by which char- it can be designated by a name or a symbol, e.g., acters would be generated cannot be defined, be- our vertebrae are idionymic while our hairs are

cause of the great diversity of features used (e.g., not. Idionymy of a set of setae is established by

morphological, behavioral, biochemical features, comparisons of specimens within a species. De- etc.). Nevertheless, in some fields, a standard velopmental stage or sex should also be consid-

method be set ered because setae be in larvae can up (e.g., morphometry). may idionymic

This article focuses on the erection of charac- and not in adults. When setae are not idionymic, ters from brute chaetotaxic data with a special their positions are not constant but depend on attention paid to the problem of variable setae. their number, in other words the inter-setal dis-

is of the setal number. Ontogeny, i.e., instar of appearance of setae, is tance a function Converse- used that char- when each has as a way to distinguish species, so ly, setae are idionymic, seta a

fixed and if is the acters should embrace the ontogenetic dimension. position, a seta absent, position

Polarization of characters and phylogenetic in- of the neighboring setae is not modified.

ference and will discussed Setae that named are next steps not be appear during development are

while setae” here. Even if phylogenetics is not directly ad- “secondary setae”, “primary are

from dressed, this theme is underlying since building present the first instar on. Sometimes, at a up characters is a preliminary phase before any given instar, some setae are variable. Setal pres- phylogenetic - cladistic - work. ence probabilities must be defined at the level of

An of is taken from In this the of example data processing a population. paper, concept proba- the group of Symphypleona, one of the major bilistic organs will be emphasized.

divisions of the Collembola (Hexapoda). The onto- Analysis of a SIC table permits ontophyloge- genetic patterns of these arthropods are quite simple netic studies. For instance, it is possible to carry and constitute for the method out multidimensional statistics shed a good support pre- to light on sented. similarities dissimilarities instars of or among spe-

The of cies. A dealt with of these postembryonic development arthropods recent paper application consists of a succession of forms separated by statistical methods (Nayrolles, 1996), but missing

of life between molts the from it discussion the and molts. The period two is was a of nature possi-

“intermolt”, also called “instar”. This paper de- ble uses of chaetotaxic characters. From my study

of scribes a method for analyzing a table with three Symphypleona, it appears that data processing dimensions, these dimensions being species, their of chaetotaxic characters is rather burdensome instars, and characters, in other words a species/ because of the numerous setae that have to be instars/characters (SIC) table. In principle, my taken into account. Computer software wouldmake method can be applied to all arthropod species studies of SIC tables easier, but before undertak- exhibiting “stases” and associated discontinuous ing such a development it is important to outline

The word the these characters. “stase” was coined by Grand- purpose and methods of analyses as

for the and jean (1938) designating morphological stag- well as to specify properties of a SIC table its es succeeding one another during the develop- subsets.

Downloaded from Brill.com10/10/2021 04:06:52PM via free access 67 - Contributions to Zoology, (3) 1998 199

Three basic and concepts: species, instar, to simplify the situation, sex will be left out to

character avoid having four dimensions in the SIC table.

Nixon & Wheeler (1990a, 1990b) made a dis-

Andre (1986, discussed the of 1988) concept stase tinction between attributes that and vary attributes and proposed to extend it to other arthropods, that are constantly distributed among specimens particularly Collembola. According to Andre of an elementary division of the species-instar- each (1989a), intermolt of a mite or Col- juvenile sex The first spectrum. type was named a trait,

lembola is a and the adult of stase, Collembola, the second a character, or more precisely, a char- which molts but does discontinu- not change by Three acter state. categories were established: 1) also ous characters, corresponds to one stase. Al- the trait as a variable attribute fitting with varia-

the of and instar are differ- though concepts stase the bly distributed alleles, 2) character state as ent, I will use the word “instar” instead of “stase”, constant attribute permitting species diagnoses, because it has the of better known and advantage being 3) the character as a set of character states

is based on among entomologists. My argument derived from one another through a series of trans-

of a but Andre’s commonness term, conceptually formations. Nevertheless, many authors (e.g.. Plat- argument remains worth while. nick, 1979; Eldredge & Cracraft, 1980; Wiley, The table term “species” as used in a SIC re- 1981; Patterson, 1988) dispute the distinction be- taxon”. In fers to a “terminal fact, such a table tween characters and character states, arguing that, include can any supraspecific taxon, with one since the essence of systematics is hierarchy, char-

condition; character are states constant within every and character acters states are just a series of instar of In other taxa. words, attributes should nested increasingly modified attributes, so that a

not “elements” taken in one terminal vary among character state can become a character at a lower

taxon and in one instar. “Elements” are defined taxonomic level. According to another viewpoint,

as individuals when the terminal taxon is a spe- it can be accurate to distinguish between charac- when the cies, as a species terminal taxon is a ters and character states. De Pinna rightly points and genus, so forth. This condition should also be out that this problem is “related to recognition of the applied at specific level, e.g., if a seta is present putative independence among sources of evidence. in adults in one population and absent in another [...] Character states are attributes that can be

population, the seta should not be included in the proposed as transformations one of another (i.e.

set of characters, or, as an alternative, the state of as a series of transformations); characters, on the character for the should be treated species as un- other hand, from are putatively independent one known. Such a consideration amounts to exactly another. If the distinction between character and

a stated by Nixon & Wheeler (1990a: principle character state is not then made, theoretically any terminal 218): “Strictly speaking, lineages at any individual attribute (i.e. any character state) could level of cladistic should not inter- analysis vary be transformed into any other, through any number nally for the characters used in the analysis”. of intermediate steps” (De Pinna, 1991: 380). The The term “character” as used here is similar to difference between character states and traits re-

variables vectors in data or analysis the terminology, lies on distinction between hierarchic patterns

and and “instars” to “species” correspond objects. as and phylogeny, reticulate patterns as birth re- Consider of collected in a set specimens one or lationship (called tokogeny by Hennig, 1966). several be clustered stations; they may according Wheeler Nixon & (1990b: 122) argued “that toko- to several criteria such as species, ontogenetic genetic systems do not meet the assumptions of

levels, sex, and even castes for social insects,. cladistic analysis and cannot be considered to be The species category is the crux of the species- fully hierarchic and are therefore inappropriate

instar-sex because instar and sex spectrum, cate- for cladistic investigations”. The concept of trait gories are merely intraspecific distinctions. In order was then extended to all “attributes that are variably

Downloaded from Brill.com10/10/2021 04:06:52PM via free access 200 P. Nayrolles - Chaetotaxy in ontophylogenetic studies

the clade er females that did not same distributedwithin any grouping (population, display exactly

is their mother each other. There is or terminal lineage) that phylogenetically un- chaetotaxy as or in this and thus resolved internally” (Nixon & Wheeler, 1990a: no genetic polymorphism case,

should be alike. that 218). all specimens Considering a left variable paired seta occurs on the or right

and such is he- side at random a seta not per se

concluded that the Probabilistic organs reditary, Grandjean variability

is due to a propensity intrinsic to each individual,

that the feature lies in the Definition of the concept is, inherited presence

this probability at a stage of ontogeny. In case,

is charac- I deem that a variable seta corresponds to a prob- the “probability of presence” a genetic

of of even if I cannot this abilistic organ. The concept probability pres- ter, explain phenomenon.

needs because What is resolved is the cell fate. ence for an organ comment this partly lineage

issue is closely connected with the problem of Several authors (e.g., Doe & Goodman, 1985;

the phylogenetic informativeness of variable or- Campos-Ortega, 1988; Held, 1990a; Simpson,

which is of the central claims of this 1990; Heitzler & 1991) showed that the gans one Simpson,

The of neural and in paper (see also Nayrolles, 1995a). problem segregation epidermal lineages

this relies on cellular inter- we must then solve boils down to question: Drosophila melanogaster

are embodied in the of “lat- do variable setae correspond to simple traits? If actions that concept

it eral inhibition”. A model for- the answer is positive, and intuitively seems cybernetic was put ward in which Heitzler & positive, the implication would be that such setae chance played a part.

of and that: “a feed- do not meet the definition characters, con- Simpson (1991; 1089) suggested

mechanism cells less sequently should not be used to infer phylogenet- back whereby producing relative their will be ic relationships. The last point that will have to receptor to neighbors more

efficient In there- be dealt with relates to a practical problem, the at signaling. normal development,

if one cell less N confidence one may grant to the use of probabil- fore, produced slightly product

ity. [protein involved in the lateral inhibition], per-

A be in random paired seta may variable a population haps through physiological fluctuations, of symmetrical specimens, some having the seta it would gain an early advantage, generate a greater

others and inhibit cells.” on both sides, wholly devoid of it. Like- signal, adjacent

be variable in A related problem is the geometrical pattern in wise, a paired seta may asymmetri- often In in this should the which setae are his of cal specimens; case, one study arranged. survey

and basitarsal setae in D. Held (1990a: frequency on both the left right sides. Grand- melanogaster, stated that “the of the basitarsus jean (1939) in Acari, and myself in Collembola 61) development

be mechanism Symphypleona, have remarked that when a paired appears to governed by a hybrid

both a coordinate and seta is variable, many asymmetrical specimens involving global system

generally occur. Given a variable paired seta, let local cell interactions”. The coordinate system and/or P be the modality “presence of the seta” and A controls the arrangement of setae in rows

the modality “absence of the seta”. Grandjean in whorls, and is likely to govern the positions of

(ibid.) showed that frequencies on the left and idionymic setae (i.e., setae with fixed positions).

side similar The basitarsus of D. bears three right are (statistically not-significantly melanogaster

of bracted bract- different) and frequencies of the different types types sensory structures: bristles,

sensilla of specimens (namely PP, PA, and AA) approxi- less bristles, and campaniforma. The bract-

binomial law. observations ed bristles in mate to a My corrob- are aligned rows, evenly spaced

within and their are con- orate this fact. the rows, positions not

but function of their number. Grandjean (1948, 1971) reared a thelytokous stant a Conversely,

mite. female oth- bractless bristles and sensilla form parthenogenetic A gave birth to campaniforma

Downloaded from Brill.com10/10/2021 04:06:52PM via free access Contributions to Zoology, 67 (3) - 1998 201

two subpatterns in which the elements are few, sofar as some elements are “sturdier” than oth- aperiodic, and idionymic. Using several perturb- ers.

factors ing (e.g., gamma irradiation), Held (1990b) Lateral inhibition has been found in organisms showed that the of bracted bristles other than in the Anabaena arrangement insects, e.g., alga does not develop in a linear direction, and con- (Wilcox et ah, 1973). The “microphtalmie alea- cluded that “the failure to find development waves toire” (random microphthalmy mutation) described of that neither sensibility implies bristle nor row by Signoret & Lefresne (1969) corresponds to a positions are established by directional pattern- similar phenomenon. This mutationentails anomaly

blood ing mechanism.” Regular spacing of non-idio- for irrigation of the cephalic area during nymic setae was thus interpreted by Held as a development of the amphibian Ambystoma mexi-

which The pattern does not develop in an iterative canum. ophthalmic artery is regressed and manner its but is the limited The along major axes, probably gen- eye undergoes growth. important erated cell by short-range signaling and cell rear- point is that all mutants do not share the same

For it be rangement. idionymic setae, may as- morphology. Specimens with 0, 1, or 2 microph- sumed that setal relates cell thalmic As in variability to eyes occur. Grandjean’s observa- interactions and some underlying random fluctu- tions, the probability of microphthalmy for the ations. and right eye left eye follows a binomial law.

Probabilistic In The concluded organs are not a novelty. 1954, authors that the mutation corre-

Stern showed that the of thoracic with arrangement sponds to a gene a variably expressed reces- macrochaetae D. follows in melanogaster a pre- sive allele. Obviously, this case is exceptional,

determined in absence of and would be pattern which a macro- it incorrect to generalize so far as

chaeta’s can be a see precursor replaced by neigh- to every organ as probabilistic. This case simply

cell its become the boring that switches fate to shows that, at one moment of development, the

new between normal and conditions such precursor. Comparisons to realize a complex structure as mutant achaete flies show that the normal the gene vertebrate eye can be unstable, so that a minute leads of 11 macrochaetae to differentiation at spe- fluctuation switches the development toward one

the achaete does cific places, whereas gene not or another direction. Complex regulating systems

cause such differentiation: zero to three this for vital and regularly prevent instability organs, spec-

macrochaetae are in mutated flies. There- imens lacking in which regulations would not be firmly

in fore, a all are not iden- established would have pure lineage, specimens a very little chance of

tical. survival, but this is not true for repeated small

The effect of the number the genes on presence or organs such as setae of arthropods. of setae has been studied in detail by Held (1990a)

on the second-leg basitarsus of D. melanogaster.

Several mutations entail a decrease in the number From theory to practice of bractless bristles whereas other mutations cause

bractless bristles to If If extra appear. we compare we now turn to the commonness of probabilis-

the effects of the it tic missing-bristle mutations, ap- organs within a series of specimens, Grand- pears that all bractless bristles do not have the jean made several accurate observations. He called

variable in the wild “ecart” the same “sensibility”, e.g., a seta (deviation) case of one seta being

becomes genotype absent, setae constantly present absent in a specimen when this seta is normally in the wild become Results for in genotype variable. present the considered species and instar, or the sensilla campaniforma are similar. Therefore, reciprocally, presence of a seta when the normal it seems that each seta has an intrinsic “propensk condition is absence (Grandjean, 1939). The se-

asserted sim- ty” to develop. Grandjean (1941) a tae concerned by deviations are almost always ilar view when he the claimed that a hierarchy can be secondary ones. Studying a population of the

for a series of in- mite developed probabilistic organs Platynothrus peltifer, Grandjean (1948) not-

Downloaded from Brill.com10/10/2021 04:06:52PM via free access 202 P. Nayrolles - Chaetotaxy in ontophylogenetic studies

ed that only 2 out of 26 larvae (first mobile form probability results from different propensities.

in Oribatida) exhibited a deviation, while all tri- When the frequencies distribution fits with a bi-

the and all adults nomial there is real chance that the calcu- tonymphs (form preceding adult) law, a

least deviation. observed sim- lated to had at one I have a probability corresponds a unique propen-

ilar situation in Collembola Symphypleona: for sity.

of it is A setal the last stages development, rare to find a final argument is decisive for using

specimen whose chaetotaxy exactly matches the probabilities as character states. I showed (Nay-

standard of the species. Specimens with one or rolles, 1993a) that presence vs. absence of cer-

several deviations are the rule, but that does not tain setae are statistically correlated. Several se-

that variable. For of the fourth antennal variable in imply many setae are instance, tae segment are

Grandjean (1948) counted 655 deviations for a Collembola Symphypleona. These setae have a

of and global number 49870 setae examined (122 spec- special shape are arranged along two rows

imens observed), and more than a half of these called intergeneratrices, because each is situated

deviations originated from a few setae. In prac- between two rows of ordinary and constant setae

tice, only setae for which more than one devia- called generatrices. I studied one species, and

tion is observed will be noted as variable, other- made a histogram of the variable “number of se-

wise chaetotaxic descriptions would be cluttered tae on the intergeneratrices” and compared it with

with many exceptions (Nayrolles, 1993a). the histogram as it would have been if the setae

be viewed Since a variable seta may as a prob- were independent. The observed histogram is much abilistic calculate setal than the calculated in organ, one can a probabil- narrower one. Therefore,

ity within a population and consider this number this example the set of correlated setae amounts

as a character state. Thus, I claim that variable to only one character, and the “probabilistic na-

of setae do not correspond to simple traits. I con- ture” the “synthetic character” is far less pro-

that is nounced than cede my argument flawed by an underly- that of original characters (i.e., se-

ing assumption, that is the propensity for bearing tae considered one by one). Similar stochastic

certain observed in a seta perhaps varies among individuals, relations between organs were mites

that the real so propensity, i.e., that of each spec- (Matsakis, 1967).

imen, is beyond the reach of our observations, The stochastic development of individual items

from lines. In this is thus counterbalanced a deterministic apart breeding pure context, by pro-

comparisons between populations can be useful cess which severely restricts the range of varia-

(an example was commented on by Van der Ham- tion for the number of items. Hence, it is accurate

define chaetotaxic variables setal men, 1981; II). to as numbers

Another approach is to consider that frequen- (normal practice among entomologists) or even cies distribution of the specimens with the seta ratios of setal numbers. The SIC table analysis on both sides (PP), on one side (PA), and without chiefly aims to cluster setae belonging to the same

the seta (AA) follows a binomial law (P is the phylogenetic set, i.e., those which relate to one

modality “presence of the seta” and A the modal- character.

“absence the Let In studies ity of seta”). us consider a pop- practice, chaetotaxic on arthropods

ulation in which the probability calculated for a take a long time, e.g., in Collembola Symphy-

certain and let I used seta is 0.5, us suppose that the pleona to mount some specimens of every population is composed of individuals of two types: instar for observation, with dissection of legs and

with the furcula. The some seta, others without. We observe mounting and observation are gen- half of the specimens with the seta on both sides, erally very time-consuming, so that only a few and half devoid of the seta. The frequencies dis- specimens are examined (about ten specimens for tribution does not follow a binomial law. Thus, an instar) and comparisons of populations cannot when the distribution frequencies is different from be carried out as a matter of course in species

a binomial law, we can say that the calculated descriptions.

Downloaded from Brill.com10/10/2021 04:06:52PM via free access Contributions to Zoology, 67 (3) - 1998 203

An the setal argument against use of probabil- guish characters which are not constant between

is if ob- taken ity that, numerous specimens are not species at the same level of ontogeny. For

the of each the served, confidence interval probability is character, observed differences are as-

wide values sumed be too and consequently, the obtained to a consequence of phylogeny. For

this such are nothing more than sampling errors. Neverthe- reason, characters will be designated as less, we can use results obtained from few obser- phylogenetically variable.

the values table will have vations if are very different, because By definition, a SIC no charac-

ter both in the important thing is the comparison and not the constant ontogeny and phylogeny. It is

For defined the three values by themselves. example, consider a by following sets: seta observed with 4 and 16 - set C of characters not constant in presences absences ontophy-

C = c one ... c in and 18 logeny. {cb 2 q, ... } species, presences and 2 absences , x

- set T of taxa. T = t ... t in another these values are {t ... }. Gener- species; statistically b 2 , tj, y

different. ally each taxon is a (hence the speaking very Although one may be species expression species/instars/characters table). reluctant to use such data in phylogenetics, a more

- set O of (i.e. instars). convincing argument is that variable setae are ontogenetic stages

O = o ... o o considered (o,, , , ... }. generally not separately. Indeed, var- 2 k z Ontogenetic stages be for iable should the same all taxa. setae may be clustered in some sets and if

1 shows a SIC table. We will deal with we assume that in each set setae have statistical Fig. relations chaetotaxic characters, and a three-dimensional (i.e. compensation) we can guess that cell of the SIC table will reveal the the presence prob- setal number of every set (which amounts to

of one seta in one instar of the setal is real ability particular one sum of probabilities) a measure

taxon. of particular of the “pilosity” a body area. The variation of

Three sections can be defined within a SIC this value, if not directly measured, can be as- table. The first section cuts off characters, and sumed to be because of quite narrow the compen- each character it for makes a bidimensional table sation back phenomenon (one may go to data to T described by x O (or also O x T since taxa can estimate the variation).

be equally presented in rows or columns). With

such a section we get species/instars cards (one

I card per character). use the word “card” be- Definition and properties of a SIC table¹ cause it refers to data processing, and the tridi-

mensional table can be considered as a database. In the study of a zoological group, one can distin- The second section cuts off taxa, and for each

bidimensional taxon we get a table described by

C x O (or also O x C). the third section 1 Finally, The following mathematical symbols of set theory are cuts off instars, and for each instar bidi- used: we get a

x of set. mensional table described C T also T x { } symbol a Example: (c,, c c is a set of three by (or C). 2, 3|

elements: c c We can divide the set of into several q, 2, 3 . taxa, T, x symbol of the cartesian product of two sets. Example: T T T subsets b 2 etc. For subsets can , 3 , example, the cartesian product {c c c ) x o is the set of t , 2, 3 {0[, 2 } correspond to genera. Within C, two subsets are ordered pairs: {(q, o,), (q, o ), (c (c , o ), (c , 2 2 , o,), 2 2 3 distinguished: the set of the con- o ontogenetically °t). (c 3, 2))

stant characters constant 0 of the (i.e., during development symbol empty set.

of each written C and the set of the n of c species), symbol intersection. c n c 0 , Example: {c 1; 2 , 3 ) jq, 4 )

= lc,l phylogenetically constant characters, written C p . U of union. symbol c c u c = Example: {cj, 2, 3 ) {q, 4 } Note that C n C is the set of characters con- 0 p {c 1 ,c 2,c 3,c4 ) stant in ontophylogeny, and according to the def- c symbol of inclusion, means that a set is a subset of inition of C: C n C = 0. Let C be the set of another set. 0 n Example: (c,) c c c ) p (q, 2 , 3 V of the characters variable both in and symbol universal V T c T, is ontogeny phylog- quantificator, e.g. a

read: “for set T included in the set T” thus: C uC uC = C. any a eny, 0 p n

Downloaded from Brill.com10/10/2021 04:06:52PM via free access 204 P. Nayrolles - Chaetotaxy in ontophylogenetic studies

Figs. 2-4. Three examples ofspecies/instars cards (5 species, 4

instars). Each card corresponds to a character with one column

per species, instars being written in rows and getting older from

the the the numbers setal top to bottom; are presence

probabilities.

Fig. 2. Species/instars card ofa character belonging to the set of

ontogenetically constant characters, C . o

Fig. 3. Species/instars card of a character belonging to the set of

phylogenetically constant characters, C p . Fig. 4. Species/instars card of a character belonging to the set of

characters variable both in ontogeny and phylogeny, C . n

Fig. I. Graphical representation of a SIC table. C is the set of

characters (x characters), T the set of taxa (y taxa), and O the

V T c set of instars T, C c C false (z instars). a pa p

V c C c C T„ T, „ false 0 0

V T c C c C true 2-4 show cards for a char- a T, na Figs. species/instars n

A T acter to C C C method for subsets or or These cards good comparing a, Tp, belonging , , . 0 p n etc., is to retain their constant contain numbers which are setal presence proba- phylogenetically

characters (C C etc.), and then to bilities. pct , p p, compare

them. we define the of charac- In phylogenetic studies, the relevant characters Thereby, may set

ters which differentiate the subsets of taxa. We are those variable between taxa. By definition,

it: T is the set the variable char- write C By definition: V c T, V C of phylogenetically d( a «..«). a q

c V T c with T T T, ... T, ... acters. it is the set of C , separat- Thus, complementary Tp ffl a Tp, u p

within C. The characters of C ed subsets, C is the set of the characters are var- d(a p w) 0 obviously iable in since C does which are constant within each phylogeny, not have any phylogenetically

subset T T and different between at least constant ... ontophylogenetically character; hence: a , Tp, w

these = two of subsets. Cv Cv wu Cv 'q -'o -'ir The Let subset of and let a concrete of us consider any T, us Appendix provides example

name T on T the set of the use of the sets C C C etc. it Defined let C be , 0, , a. a, 0(X p q the ontogenetically constant characters, C the pa set of the phylogenetically constant characters,

C the set of the as well as The na phylogenetically spreading projection ontogenetically variable characters, and C the q0l of the variable In order set phylogenetically characters. to assess distances among species, or

the “three Whether a large table is cut into subsets of taxa, instars, or characters, we could use for each subset T the characters which have to mode-principal component analysis” applied to a,

tridimensional tables In be kept are those belonging to C Only these by Kroonenberg (1983). qa.

useful for such characters are differentiating species fact, a method is not necessary. Indeed, we within T can the not of considered a . study phenotype species

A each of in- character phylogenetically constant in a sub- as a whole, but species divided into set of taxa is not necessarily phylogenetically stars. That amounts to opening the tridimensional

for all SIC table and constant taxa. So, the relation C c C is getting a bidimensional table with pct p We characters columns not always true. can thus establish logical in and instars of species in

The such relations as those stated below. study of rows (of course it is equivalent to write characters relations will be a to data- in I this prerequisite develop rows or columns). name process a “spread- base application. ing projection”. Andre (1988, 1989b) proposed to

Downloaded from Brill.com10/10/2021 04:06:52PM via free access - Contributions to Zoology, 67 (3) 1998 205

into reduced the divided metrical be project a space species pattern, homeotypic relationships can into instars, and then to connect the points repre- defined with regard to chaetotaxic arrangement. senting the successive instars of each species; he For instance, setae on appendages of Symphyple- coined the phrase “ontogenetic trajectory” for such ona are arranged in a “longitudinal structure” com- a pathway. By this means, development of spe- posed of eight rows, called generatrices (Nay- cies can be compared. rolles, 1992). Furthermore, a “transversal structure”

The bidimensional table achieved is whorls of tibiotarsi. Setae by a spread- often observed, e.g., ing projection is called instars-species/characters of the same whorl or of the same generatrix are table (e.g. Tables I, II). If x, y, and z are respec- liable to have homeotypic relationships (Fig. 5).

the cardinal numbers of and then of these tively C, T, O, Moreover, the presence probabilities se- the dimensions of the obtained table are x and yz. tae can be similar; in this case, we can combine

In The character fact, only characters phylogenetically variable characters in a single one. new cor-

of that the of the are interest, so a spreading projection per- responds to sum presence probabilities.

the of the defined It is rather than formed from part SIC table on preferable to use sum average

data C is sufficient. As far as the distinction among because that does not alter initial nor dis- q the instars of species is concerned, the bidimen- tances between instars of species. sional table the As clustered must provide same information a rule, homeotypic setae are ac-

as the data: that the criteria: their content original we say cording to two 1) presence proba- instars-species/characters table must fit the “per- bilities are similar (similarity criterion), and 2) tinence tax- their principle”. For example, if, for every presence probabilities can be ordered (ordi-

all the instars have been nal is in Tables on, distinguished among criterion). An example given II

that still in and them, must be true the instars-species/ III, and a table obtained from real data is characters table. If it is not the we add to C given in the (Table XI). case, q Appendix the smallest set of characters that restores perti- Let u and v be two characters, each corresponding nence (Table I). to the presence probability of one seta. Under the

Some setae have a homeotypic relationship, e.g., first criterion u is similar to v, that we write u = v,

a seta in the same place on the three pairs of legs and according to the ordinal criterion either u is

(see Fig. 5). When setae are arranged in a geo- less than v (u < v) or u is greater than v (u > v).

Table I. Determination of the pertinent set of characters. Characters (Cj to c in columns, and t , t divided into their 8 ) species (t,, 2 3)

instars to o ) in rows. All instars are the characters ofC for and o oft Character c of C (o, distinguishedby , except o, . permits 4 q 2 2 6 p

this distinction: c restores pertinence. Hence, the pertinent set of characters (showed by a double arrow) is C u {c }. 6 q 6

C C 4 P

C C C C C C C C 1 2 3 4 5 6 7 8

0 0 0 0 1 0 ti °1 0 0

1 0 0 0 1 1 0 0 °2

°3 1 1 0 1 1 1 1 0

1 1 1 1 1 1 1 O4 1

0 1 0 0 0 0 0 0 *2 °I

0 1 0 0 0 1 0 0 °2

0 °3 1 0 1 0 1 1 0

0 1 1 1 0 1 1 1 °4

t 0 0 0 0 1 0 0 0 3 °1

°2 1 0 0 0 1 1 0 0

°3 1 0 1 0 1 1 1 0

1 0 1 1 O4 1 1 1 1

Downloaded from Brill.com10/10/2021 04:06:52PM via free access 206 P. Nayrolles - Chaetotaxy in ontophylogenetic studies

is that be mathe- Similarity a fuzzy relation can

matically defined with the help of the fuzzy sets

theory (to discuss this point is beyond the scope

of the see for an in- present paper, Zadeh, 1965,

troduction to the principles).

Characters and u v belong to the instars-species/

characters table in which characters are written in

columns and instars of species in rows. There are

We write and the i-th observations of n rows. Uj v,

i from to The u and v, varying I n. similarity or

ordinal relation between u and v can be stated as:

V i = e [1, n], U; Vj

u-v o-j n

I Uj ~ X Vj

.. i=0 i=0

Vie [1, n], Uj< v ;

u < v n

E u. < Z v i i=0 i=0 Fig. 5. Homeotypic relationships between setae. From the left

to the right are drawn the posterior side of fore, mid, and hind

tibiotarsus of the first instar of Bourletiella hortensis (Fitch, Vie [1, n], dots. are Uj> 1863), sockets of setae being represented by Setae V;

in and whorls u>v arranged generatrices (thick lines) (thin lines). „

Certain whorls are incomplete (e.g., posterior generatrix with X > X Uj V; two setae absent on midleg and three on hindleg). Three ho- > i=0 i=0

meotypic relationships (represented by double arrows) can be With regard to the ordinal criterion, it would be considered: 1) between setae ofa same whorl, 2) between setae relevant to look for gradients of of a same generatrix, and 3) between setae situated in the same anterior-posterior

the three tibiotarsi. setal place on presence probabilities on the legs (by com-

Table II. Instars-species/characters table. Characters in columns, species divided in their instars in rows.

C C C cl C C C C C 1 2 3 5 6 7 8 9

0 0 1 0 0 0 0 0 1 t, °1

0 1 1 0 0 0.4 0 0 1 °2

1 1 0.8 1 O3 1 1 1 1 1

1 1 1 1 0.8 1 1 1 1 °4

0 0 0.5 0 0 0 0 0 0 ‘2 °1

0 1 0.5 0 0 0.8 0 0 0 °2

1 1 0.5 0.8 1 0.8 0 1 1 °3

1 1 0.5 1 1 0.8 0 1 1 04

0 0 0 0 0 0 0 0 0 «3 °1

0 0 0 0 0 0 0 0 0 0 2 0.2 1 0 0 0 0 0 0 0 °3

1 1 0 0.2 0 0 1 1 0.5 °4

0 0 1 0 0 0 0 0 0 ‘4 °1

0 °2 0 1 0 0 0 0 0 0

°3 0 1 1 0 0 0 0 0 1

O4 1 1 1 0 0 0 0 0 1

Downloaded from Brill.com10/10/2021 04:06:52PM via free access 67 - Contributions to Zoology, (3) 1998 207

mid and parisons between fore, hindlegs); such antennal segment display negative statistical rela-

could also be looked for the gradients along rows tions for presence vs. absence. That means that of setae on when a seta is on the whorl the appendages. present n, proba-

setae is for Clustering justified two reasons. bilities of setae on the whorls n - 1 and n + 1 are

from a statistical it takes less than if the First, viewpoint, account seta of whorl n was absent. Setae of the initial table and does alter 2 not greatly % involved in compensation are thus clustered in

distances between instars of This latter described species. one set by a variable “number of se-

point is of interest, since the best distance in the tae”.

2 method of is the To ontogenetic trajectories % dis- cluster homeotypic setae on the basis of

tance their (Nayrolles, 1996). Second, from a biologi- relationship or on the similarity or ordinal

cal it is criteria relates viewpoint, not surprising that several se- to the same principle. Setal clus-

tae with have homeotypic relationships presence tering does not only decrease the information of either similar probabilities or in ordinal relation. data but it also improves this information because these Finally, setae convey the same information. of the reduction of redundancy and background

Thus, the setal reduces the noise the individual clustering redundancy (i.e., variability of each seta of information. setal in a set with compensation). For instance, I The idea of setal clustering lies also in the con- used the setal clustering to enhance distinctions cept of relationship between setae. I stated above between species (Nayrolles, 1995b). that for certain setae on the fourth antennal seg-

ment in Symphypleona, the observed number of

setae a much variation displays narrower range of The alphabetic and numeric projections

than the number calculated from a model of sta-

tistical independence between setae. I called this General points phenomenon “compensation” (Nayrolles, 1993a, and I also Matsakis, 1967) because absences coun- stated above that the setal presence probability terbalance has presences. Indeed, the successive vari- phyletic information content. Therefore, the able setae in a row (intergeneratrix) along the state of a variable character is not to be consid-

Table III. Cluster of several Table II: to characters of c corresponds the combination of and c c to the l 2 Cj 2, 45 corresponds of and combination c c and c to the and The , corresponds combinationof c c of initial characters are based 4 5 7>8 7 8 . relationships on

and the < = following profile comparisons: c c c c < c homeotypy c, ; ; . 2 4 5 ? 8

C c C C O l,2 3 4,5 C 6 jsl » 9

0 1 ‘l °1 0 0 0 1

1 1 0 °2 0,4 0 1

2 1 1.8 °3 1 2 1

2 °4 1 1.8 1 2 1

l 0 0.5 0 0 2 °1 0 0

1 0.5 0 °2 0.8 0 0

2 0.5 1.8 °3 0.8 1 1

2 °4 0.5 2 0.8 1 1

l 0 0 0 0 3 °1 0 0

0 0 °2 0 0 0 0

1.2 °3 0 0 0 0 0

2 °4 0 0.2 0 2 0.5

0 1 ‘4 °l 0 0 0 0

0 1 °2 0 0 0 0

1 1 0 °3 0 0 1

2 °4 1 0 0 0 1

Downloaded from Brill.com10/10/2021 04:06:52PM via free access 208 P. Nayrolles - Chaetotaxy in ontophylogenetic studies

ered it while and as polymorphic, instead corresponds to a Case (1) is very frequent, (2), (3) (4)

step in a transformation series. For example, if are very rare in Collembola. For example, (3) is

the plesiomorphic state is “presence of a certain unknown, (2) and (4) are known for only one seta

which in adults” and absence the D of Betsch seta appears (trichobothrium Symphypleona, see

apomorphic state, then species in which the seta & Waller, 1994). These features are different in

in is in interme- Acari: is of question variable adults have an (2) not rare and presence calyptosta-

diate for this character. In this several will limit evolutionary state ses poses problems. Here, we

case, values of presence probabilities are gath- the discussion to Collembola, and since cases (2), ered in few ranges, each range fitting with a state, (3), and (4) are very unusual we will only consid- e.g., probability = 1 as plesiomorphic state, [0.4 - er case (1), i.e., ontogenetic change corresponding

0.6] as intermediate, and 0 as apomorphic. to setal appearance. In other words, setal presence

Analyses of tridimensional tables aim to study probability does not decrease during ontogeny.

“ of For it We the character —” kinship species. that, is necessary to give following rules: is build up a table with two dimensions: one for used for the lack of a seta; when a seta is variable

characters, the other for taxa. If we try to give a at the instar in which it appears, this instar is

of the in written in in later graphical representation process ques- parentheses, and if a instar it

tion, this looks like a projection on the plane de- becomes constant, this instar is given as well. For

fined the characters and taxa. that with by Nevertheless, example, (o3 ) means a seta appears var-

such a table should take into account the ontoge- iability at the third instar and remains variable,

netic dimension when do not second- while )o that a seta with at species get (o3 4 appears variability

the instar. this of the third instar becomes fourth. ary setae at same Into type pro- and constant at the

dimensions obtained jection, of the table are x Table IV provides an example. and y, the cardinal numbers of C and T. Table The alphabetic projection yields a bidimensional

will hold either letters In such squares (alphabetic projec- species/characters table. a table, it is un-

tion) or numbers (numerical projection). necessary to consider the phylogenetically con-

I will support my explanations with a series of stant setae, and it is thus relevant to carry out an

These tables fit the tables (IV to VII). a didactic pur- alphabetic projection from part of the SIC

they are not real tables (real tables are much table limited by C In practice, an alphabetic ta- pose, q.

An of treatment for real data is ble is constructed in the first because it larger). example step, per-

given in the Appendix. mits us to identify the phylogenetically constant

characters and to ignore them in other analyses

(see Appendix). The alphabetic projection

If consider the absence of we presence vs. any

seta, two basic ontogenetic changes can be ob- The numericprojection

served, appearance vs. disappearance. So, during

consider the ontogeny, we can following changes: The aim of alphabetic projection was to get a

- of (1) a single change that is the appearance a standard presentation of the chaetotaxic ontogeny

secondary seta; and to provide a cursory knowledge of data. The

- a that is the (2) single change disappearance purpose of a numeric projection is to achieve a of a primary seta; cladistic matrix.

- (3) a combination of both single changes, in Values produced by this data processing should

and this way; appearance of a secondary seta then be regarded as real measures. For each final char-

its disappearance; acter, these measures provide a comparison of the

- a combination of both in (4) single changes, “intensity” with which a set of homeotypic setae

this of seta and de- way: disappearance a primary is present throughout one or several stages of

then its reappearance. velopment. An ultimate data processing, which

Downloaded from Brill.com10/10/2021 04:06:52PM via free access 67 - Contributions to Zoology, (3) 1998 209

Table IV. Alphabetic species/characters table corresponding to Table II. See text for further explanations.

C C C C l , C C C C C 2 3 4 5 6 7 8 9

°3 °2 °l °3 (O3) (°2)°3 °3 °3 °1

‘2 °3 °2 (o,) (o )o °3 (o ) 3 4 2 O3 °3

0 — (° ) — — ‘3 4 °3 (0 ) (o ) 3 4 °4 O4 4

l — — — — — 4 °4 °3 °1 °3

Table V. Numeric table species/characters corresponding to Table II. See text for further explanations.

C C 1 C C C C C C C 2 3 4 5 6 7 8 9

selected 0 +° + + °3 °2 °1 +° 0 0 +°4 0 3 4 °3+0 4 2 3 3 3+°4 0i+0 2 instars °4 0 +° 3 4

1 1 1 2 1.6 •i 2.4 2 2 4

1 l 1 0.5 1.8 2 2.4 0 2 2 2 0.2 0 ‘3 0 0.2 0 0 1 1 0,5

l 0 0 1 0 0 0 0 0 4 2

will not be dealt with delim- here, corresponds to studies on Collembola Symphypleona. Apart from

itation of character states and and cod- this which polarizing exceptional case, may deserve special

them before to achieve a cladistic the ing analysis. treatment, table obtained can be considered to

In the first for each step, character, we note be pertinent beside the original SIC table. Obvi- instars concerned the variation by among species, ously, the process of numeric projection would be

then the we sum of the se- more involved if there other of presence probabilities were types on-

lected instars. Table V, calculated from Table II, togenetic changes than the setal appearance. this corresponds to step. For instance, consider Characters will be combined in order to de- character c two instars are selected, o and o crease their number and redundant 4 , 3 4, information.

the value of c for the taxon t in Table is We 4 2 (1.8 V) will combine setae with homeotypic relation-

equal to the setal presence of c at the to the and probability 4 ship according clustering criteria perti-

instar o of t in Table the setal 3 2 (0.8 II) plus nence principle.

of c at the instar o of t (1 be presence probability 4 4 2 Characters to clustered have the same evo-

in Table II). In case of constant direction. One ontogenetically lutionary could think that this as- characters in C the first entails (included ), we only keep sertion a How would 0 superfluous assumption.

in instar, character c Tables II and one be 0[ (e.g., 3 V). able to assess a priori that several charac-

The table obtained has mention to the selected ters have the same evolutionary direction? If char- instars of each character. acters present very similar values, evolutionary

This first can be as step regarded quite objec- states of these characters (e.g., seta present at the since tive, it makes no on the fourth assumption polari- instar vs. seta present from the third instar zation of characters. Nevertheless, will adding up pres- on) have the same polarity after a cladistic ence probabilities of one seta observed in several analysis. So, characters to be clustered are corre- instars implies that this sum holds as and de much infor- lated facto have the same evolution. That mation as the data. In other that does original words, not mean that any of the correlated charac-

consider amounts to c as the same infor- be clustered. 6 having ters can Indeed, setae must share mation content for t, and t in Table in of 2 V, spite homeotypic relationships to be clustered. Ento- the fact that t[ and t in Table each divided into do 2 II, mologists not cluster clypeal setae with tarsal

have instars, different values for c such 6 . setae because setae are on distinct areas and

I have found this situation in very rarely my have no homeotypic relationship. In cases where

Downloaded from Brill.com10/10/2021 04:06:52PM via free access 210 P. Nayrolles - Chaetotaxy in ontophylogenetic studies

setal number setae are idionymic, a corresponds cess from Table VI to Table VII. However, this

to a combination of characters which be made from elementary process can directly Table II (but

the are presence vs. absence of each seta. This case is for very large tables, intermediate step of Ta-

quite frequent in literature, but authors do not bles V-VI is very useful).

always realize the underlying assumption involved

in such characters (Nayrolles, 1996). Conclusion Every new character (combined character) must

have instars’ its range which encompasses all se- A SIC table based setal on presence vs. absence lected instars of its initial component characters. provides a good tool for ontophylogenetic studies From characters a practical viewpoint, only showing on arthropods. The characters of this tridimen- the same instars’ be range can directly clustered. table setal sional correspond to presence proba- add several instars Nevertheless, we may one or bilities. In this paper, I define the main properties to characters for their instars’ matching up rang- of such a table, particularly the problem of its es. Each time one adds an instar to the instars’ division into taxonomic groups. it is range, necessary to add also its setal presence Two methods can be used for analyzing a SIC probability to the values of the character. For ex- table. The first aims to compare instars of species. in Table if we c and ample, V, want to gather t c 2 The tridimensional table is “opened out” (spread- we have to add o to c and o to c In this 2 h 3 2 . ing projection), with, for example, the characters the of process, values C[ do not change (because in columns and the instars of species in rows. for C| the setal presence probability of the added Then we gather characters with the same infor- o is to the other hand instar, 2 equal 0), on we , mation content, this information taking into ac- must add 1 to the values of c (because for c the 2 2 count statistical as well as biological factors (homeo- setal presence probability of the added instar, o 3 , this typy). By means, we get the table with the is equal to 1). Table VI shows the preparation for fewest informative characters and keeping at best the size reduction of Table V. Table VII corre- differences between instars of species of the ini- sponds to the reduced table. tial table (pertinence principle). It is then possible The similarity criterion can be used in the same to apply statistical methods such as Factor analy- in the On the other way as spreading projection. sis. By joining points representing the successive hand, to the ordinal criterion raise a apply may instars into a reduced space, we can draw ontoge-

For in Table we have: c > c problem. instance, II, 8 7 , netic trajectories (Andre, 1988, 1989b; Nayrolles, and, after the numeric (Table V), if we projection 1996). add c and c we the same value (2) for t and 7 g , get 2 A second process seems to erase the ontogenet- t It is because this value is built 3 . problematic up the ic dimension, but in fact, ontogeny, when it from two different ways, and generally that is due differs among taxa, is not forgotten. We construct to two distinct Indeed, evolutionary pathways. sup- a table with characters in two dimensions, e.g., that character of and pose the c c is so polarity 7 8 columns and taxa in rows. This is al- process an that the instar of setal is the primitive appearance phabetic projection when the achieved table is

the third intermediate evolutionary state cor- filled with one, letters symbolizing instars of setal ap- responds to the setal at the fourth in- appearance pearance and a numeric projection when the table

star (ontogenetic delay), and the fully is filled with numbers. In this last for apomor- case, each

is the setal then and phic state lacking, taxa t t we add the setal 2 3 character, presence probabilities

are is evolved for c and really separated: t 2 7 prim- of the selected instars. The obtained table can be

itive for whereas evolu- c t has an intermediate used to achieve a cladistic g, 3 analysis.

tionary state for both characters. Thus, in this ex- Klompen & O’Connor (1989) showed that the

hold and We ample, we must c c use of transformations better 7 8 separated. ontogenetic yields

characters have defined the possibility of cluster- results for cladistic analyses than a stase by stase

in Table and we have this ing II, performed pro- approach. Klompen & O’Connor worked on Acari.

Downloaded from Brill.com10/10/2021 04:06:52PM via free access Contributions to Zoology, 67 (3) - 1998 211

Table VI. for the size reduction of Preparation Table V; characters to be combined are shown by double arrows. See text for further explanations.

J

C C C C C C C C C 1 2 3 4 5 6 7 8 9

selected +o o ° +0 ° +0 ° ° +0 + 2 3 2 °3+°4 4 O2+O3+ 3 °1 3 3 +°4 3 4 o,+o 2 instars °4 °3+°4

*. 1 2 1 2 1.6 2.4 2 2 4

1 ‘2 2 0.5 1.8 2 2.4 0 2 2

l 0.2 1 0 0.2 0 0 1 1 0.5 3

l 0 1 1 0 0 0 0 0 2 4

Table VII. to Table V to the Correspondence reduced: c corresponds combination ofc and c ofTable c to the t 2 ( 2 V; 45 corresponds

combination of c and c of Table V. See text for further explanations. 4 5

C C C C C C C l,2 3 4,5 6 7 8 9

selected o +o + +0 2 3 ° °1 °3+°4 O2+O3 O3+O4 3 4 O1+O2+

instars +0 °4 °3 4

‘i 3 1 3.6 2.4 2 2 4

3 *2 0.5 3.8 2.4 0 2 2

t 1.2 0 0.2 0 i 1 0.5 3

l 1 1 0 0 0 0 2 4

Their data code than those were more difficult to joined before the setal clustering. As far as the

for Collembola. I here outline a simple method numeric projection is concerned, analyses will be

for use in Collembola (this method cannot direct- performed on each table, and two cladistic matri-

be the data of & will be thus ly applied to Klompen O’Connor). ces achieved. It can be difficult or

the two similar. In irrelevant Nevertheless, approaches are to merge these matrices and the practi-

fact, when I add setal presence of cal solution consists in each matrix probabilities analyzing sep-

several selected instars, I do not make an instar arately (cladistic analysis). Nevertheless, compar-

instar but I by out a transfor- isons between between study, really carry groups, i.e., matrices, may

mation The main difference be- be useful for but this is pattern approach. polarizing characters, an-

tween and that of & my study Klompen O’Connor other topic.

lies in the nature of the data. The biological mate- The SIC table has been defined for species al-

rial I have studied - Collembola - Symphypleona ways passing by the same number of instars. Yet,

shows transformation that homo- patterns are quite many arthropods present an extensive variability setal geneous, because, as a rule, presence proba- in the number of instars. However, several instars

either does or increases bility not change during may generally be gathered in “stases” (Andre,

development. 1989a). Every stase shows a fixed chaetotaxic

In Collembola the Symphypleona number of pattern (direct consequence of the definition of

instars is fixed within juvenile every species, but the concept of stase), so that the SIC table analy-

varies between families sis (e.g., Bourletiellidae have can also be applied providing that the number

three instars whereas Sminthuridae have of does juvenile stases not vary between taxa. Compara-

The SIC table has been for four). defined taxa tive studies of the first instar constitute a worth-

the displaying same number of instars. If we con- while introduction to ontophylogenetic tasks. Re- sider both families Bourletiellidae and Sminthuri- cently Pomorski (1996) has successfully applied dae are sister together (likely they groups), a spread- the morphology of the first instar to the generic

can be each ing projection performed on group, classification of Onychiurinae (Collembola, Ar- and the tables instars-species/characters can be thropleona).

Downloaded from Brill.com10/10/2021 04:06:52PM via free access 212 P. Nayrolles - Chaetotaxy in ontophylogenetic studies

It would be valuable to develop database soft- neurogenesis of Drosophila melanogaster. Trends Neuro-

11 400-405. make of sci., (9); ware to analyses tridimensional tables De Pinna, M. C. C., 1991. Concepts and tests of homology Such would easier. software spur phylogenetic in the cladistic paradigm. Cladistics, 7: 367-394. research based on chaetotaxy in arthropods. Doe, C. Q. & C. S. Goodman, 1985. Early events in insect about Our thinking evolution is probably too neurogenesis, II. Role of cell interactions and cell line-

vertebrates. in the determination of neural cells. Dev. influenced by studies on The nature age precursor Biol., Ill: 206-209. of characters, at the specimen level, such as pres- Eldredge, N. & J. Cracraft, 1980. Phylogenetic patterns and absence of dis- ence vs. many “petits caracteres the 1-349 evolutionary process: (Columbia University Press, continus” (i.e., setae, Grandjean, 1951), is sharp- New York). different from the of the characters ly nature usually Grandjean, F., 1938. Sur 1’ontogenie des Acariens. C. r.

in hebd. Seanc. Acad. 206: 146-150. used vertebrate analyses. Indeed, we are used Sci., Paris, Grandjean, F., 1939. La repartition asymetrique des fixed characteristics of organes to see organs as species. aleatoires. C. r. hebd. Seanc. Acad. Sci., Paris, 208 : 861- This perspective is not consistent with the study 864. of setae in it is much too determinis- arthropods; 1941. Sur la dans les Grandjean, F., priorite groupes d’orga-

tic and cannot the of rien. C. hebd. explain frequency specimens nes homeotypes qui evoluent par tout ou r. which display a chaetotaxy with one or several Seanc. Acad. Sci., Paris, 213: 417-421.

Grandjean, F., 1948. Sur les ecarts dans un clone de Platyno- differences from the standard of the species. As thrus peltifer (Acarien). C. r. hebd. Seanc. Acad. Sci., anomalies in probabilistic organs are not simply Paris, 227: 638-661. developmental genetics, systematists should pay Grandjean, F., 1949. Au sujet des variations individuelles et

more attention to this phenomenon. des polygenes de frequence. C. r. hebd. Seanc. Acad.

Sci., Paris, 229: 801-804.

Grandjean, F., 1951. Les relations chronologiques entre on-

togeneses et phylogeneses d’apres les petits caracteres Acknowledgments discontinus des Acariens. Bull. Biol., 3: 269-292.

Grandjean, F., 1971. Caracteres anormaux et vertitionnels 1 thank H, M. Andre, L. Bonnet, D. R. Brooks, K. Christiansen, rencontres dans les clones dePlatynothrus peltifer (Koch). L. Deharveng, S. Hopkin, P. Tassy, Q. D. Wheeler, and an Premiere partie, Acarologia, 13 (1): 209-237. reviewer for anonymous comments and helpful suggestions on Hammen, L. van der, 1981. Numerical changes and evolu- the manuscript. tion in actinotrichid mites (Chelicerata). Zool. Verb. Lei-

den, 182: 1-47.

Heitzler, P. & P. Simpson, 1991. The choice of cell fate in

References the epidermis of Drosophila. Cell, 64: 1083-1092.

Held, L. L, Jr., 1990a. Arrangement of bristles as a function

of bristle number on a leg segment in Drosophila mela- Andre, H. M., 1986. The stases in Collembola: a case study. nogaster. Roux’s Arch. dev. Biol., 199; 48-62. In: R. Dallai (ed.), Second international seminar on Ap- Held, L. L, Jr., 1990b. Sensitive periods for abnormal pat- terygota. Universita di Siena, Italy: 301-305. terning on a leg segment in Drosophila melanogaster. Andre, H. M., 1988. Age-dependent evolution: from theory Roux’s Arch. dev. Biol., 199: 31-47. to practice. In: C. J. Humphries (ed,), Ontogeny and sys- Hennig, W., 1966. Phylogenetic systematics: 1-263 (Univer- tematics: 137-187 (Columbia University Press, New York). sity of Illinois Press, Urbana). Andre, H. M., 1989a. The concept of stase. In: H. M. Andre Klompen, J.S.H. & B.M. O’Connor, 1989. Ontogenetic pat- & J.-Cl. Lions (eds.), L’ontogenese et le concept de stase terns and phylogenetic analysis in Acad. In; H. M. Andre chez les Arthropodes / Ontogeny and concept of stase in & J.-Cl, Lions et le de arthropods: 3-14 (AGAR Publishers, Wavre, Belgium). (eds,), L’ontogenese concept stase

chez les / and of stase in Andre, H. M., 1989b. Ontogenetic trajectories in arthropods. Arthropodes Ontogeny concept 91-103 Publishers, Wavre, In: H. M. Andre & J.-Cl. Lions (eds.), L’ontogenese et le arthropods: (AGAR Belgium)

and Kroonenberg, P. M., 1983. Tree-mode principal concept de stase chez les Arthropodes / Ontogeny component

and 1-399 concept of stase in arthropods: 83-90 (AGAR Publishers, analysis; theory applications: (DSWO Press,

Wavre, Belgium). Leiden).

Betsch, J. M. & A. Waller, 1994. Chaetotaxic nomenclature Matsakis, J. Th., 1967. Sur certaines variations quantitatives

of the head, thorax and abdomen in Symphypleona (In- affectant les appendices arthropodiens: cas des griffes de

secta, Collembola). Acta zool. Fenn., 195: 5-12. TOribate Rhysotritia ardua. Bull, Soc. Hist. nat. Tou-

Campos-Ortega, J. A., 1988. Cellularinteractions during early louse, 103; 590-631.

Downloaded from Brill.com10/10/2021 04:06:52PM via free access Contributions to Zoology, 67 (3) - 1998 213

Nayrolles, P., 1992. Aspects structuraux de la chetotaxie Appendix appendiculaire des Collemboles Symphypleones. Nouv.

Revue Ent. (N.S.), 9 (4); 345-356. I provide a concrete example of data treatment based on the Nayrolles, P., 1993a. La biometric des caracteres disconti- tibiotarsal chaetotaxy observed in 14 species of the subfamily nus d’apres le revetement appendiculaire des Collem- Sminthurinae (descriptions in Nayrolles, 1993b, 1994, 1995b). boles Symphypleones. I. Sur quelques nouveaux concepts The species and the abbreviations used in the tables are listed dans 1’analyse numerique de la chetotaxie. Bull. Mus. below: natl. Hist, nat., Paris (4e sen), 15: 79-93. GisimirusGisinurus malatestai Dallai, 1970 Gi mal Nayrolles, P., 1993b. A standardized description ofEurope- Caprainea bremondi (Delamare & Bassot, 1957) Ca bre Sminthuridae 1: an (Collembola, Symphypleona), genera Caprainea marginatamargínala (Schott,(Schott, 1893) Ca mar Lipothrix, Gisinurus, and Caprainea. Bijdr. Dierk., 63 (1): Allacma fusca {(Linnaeus,Linnaeus, 1758) A1Al fus 43-60. Allacma gallica (Carl, 1899) A1Al gal Nayrolles, P., 1994. A standardized description of European Spatulosminthurus betschi Nayrolles, 1990 Sp bet Sminthuridae (Collembola, Symphypleona), 2; review of Spatulosminthurus lesnei (Carl, 1899) Sp les four of the species genera Allacma and Spatulosminthurus. Sminthurus bourgeoisi Nayrolles, 1995 Sm boubou Bijdr. Dierk., 64 (3): 151-162. Sminthurus nigromaculatus Tullberg, 18711871 Sm nignig

1995a. retour a la defini- Nayrolles, P., Propositions pour un Sminthurus viridis (Linnaeus, 1758) Sm vir tion de de la originelle 1’homologie, d’apres Tanalyse Sminthurus leucomelanusNayrolles, 1995 Sm leuleu

chetotaxie des Collemboles, Bull. Soc. zool. France, 120 Sminthurus bozoulensis Nayrolles, 1995 Sm boz

3-20. (1): Sminthurus hispanicus Nayrolles, 1995 Sm his

1995b. A standardized of Nayrolles, P., description Europe- Sminthurus multipunctatus Schaffer, 1896 Sm mul

an Sminthuridae 3: The (Collembola, Symphypleona), descrip- 'tibiotarsus is the subapical segment of legs in Col-

tion of seven of species Sminthurus, including four new lembola. It bears eight generatrices and five whorls of primary

to science. 64 215-237. setae in between Bijdr. Dierk., (4): Symphypleona. Secondary setae appear

Nayrolles, P., 1996. Ontogenetic trajectories in Symphy- whorls during development. Fig. 6 shows the setae liable to be

J. tibiotarsi of and pleona (Collembola). Morphol., 227 (2): 127-143. present on the species observed provides a

Nixon, K. C. & Q. D. Wheeler, 1990a. An amplification of straightforward display of the setal nomenclature. A code for

the the other phylogenetic species concept. Cladistics, 6: 211-223. designating tibiotarsus (or any segment of legs) is

Nixon, K. C. & Q. D. Wheeler, 1990b. Extinction and the added to the symbol ofthe seta: TI1 for the fore tibiotarsus, TI2

for origin of species. In; M. J. Novacek & Q. D. Wheeler the mid tibiotarsus, and TI3 for the hind tibiotarsus. For is (eds.), Extinction and cladistic analysis; 119-143 (Co- example, (Tll)la the seta of the fore tibiotarsus situated on the lumbia University Press, New York), first whorl and on the anterior generatrix. When two setae in the tibiotarsi Patterson, C., 1988. Homology in classical and molecular same place on two are considered together, we write the number ofboth tibiotarsi biology. Mol. Biol. Evol., 5 (6): 603-625. separated by a comma, e.g.,

A 3” Platnick, N. L, 1979. Philosophy and the transformation of (TI 1,2)IIa. period replaces numbers “1,2, to designate a

seta all cladistics. Syst. Zool., 28: 537-546. on tibiotarsi, e.g., (Tl.)IIIa. In chaetotaxic of the instar of Pomorski, R. J., 1996. The first instar larvae of Onychiuri- descriptions Symphypleona, seta is of a given by a letter: P a - for nae a systematic study (Collembola: Onychiuridae). Genus appearance primary seta, and D, T, and C for seta the (Wroclaw), 7 (1): 1-102. respectively Q a emerging at

second, third, fourth or fifth instar. I these Signoret, J. & J. Lefresne, 1969. Description et interpreta- keep symbols (they correspond to the first letter ofthe French adjective tion d’une nouvelle mutation “microphtalmie aleatoire” “premier”,

“deuxieme”, “troisieme”, etc.) in order to provide a real chez 1’Axolotl (Ambystoma mexicanum Shaw). C. r. hebd.

on a current example based standard of description. The sign - Seanc. Acad. Sci. Paris, ser. D, 269 (4): 507-509. is for seta a absent. When a seta is variable at the instar in which Simpson, P., 1990. Lateral inhibition and the development it the letter that this instar be- of the appears, symbolizes is written sensory bristles of the adult peripheral nervous tween parentheses; if in a later instar it becomes constant, this system of Drosophila. Development, 109; 509-519. instar is well. given as For example, (Q) means that a seta Stem, C., 1954. Two or three bristles. Am. Scient., 42: 213- appears with variability at the fourth instar and remains 247.

means that with the variable; (T)Q a seta appears variability at Wilcox, M., G. J. Mitchison & R. J. Smith, 1973. Pattern third instar and becomes constant at the fourth. These symbols formation in the blue-green alga, Anabaena. I. Basic mech- are used in the alphabetic projection of the SIC table (Table anisms. J. Cell Sci., 12: 707-723. In setae in VIII). Sminthurinae, tibiotarsal appear first, third, or Wiley, E. O., 1981, Phylogenetics: the theory and practice fourth instars (letters P, T, and Q in Table VIII); the fifth instar of phylogenetic systematics: 1-439 (Wiley, New York). is the adult and shows the same tibiotarsal chaetotaxy as the Zadeh, L. A., 1965. Fuzzy sets. Information and Control*, fourth instar. A number with one digit, named occurrence 8: 338-353. (Nayrolles, 1993a), provides an estimation ofthe probability of

presence of a variable seta. Below are listed the occurrences of Received: 10 1997 September variable setae with letters of the instar of the appearance:

Downloaded from Brill.com10/10/2021 04:06:52PM via free access 214 P. Nayrolles - Chaetotaxy in ontophylogenetic studies

Gi mal 0.8 which either the third instar the fourth. (TI3)2ae (Q) setae appear: at or at

Ca bre (TI3)02pe (T)Q 0.5 From a study of ontogenetic trajectories applied to tibiotarsi

Ca bre (TI3)03pe (T)Q 0.6 of Symphypleona, it was assessed “that the long ontogenetic

of with evolution of tibiotarsal Sp les (TIl)FSpei (T)Q 0.3 trajectory S. viridis fits an the

Sp les (TI2)FSpei (T)Q 0.4 chaetotaxy, this evolution being a neochaetosis (appearance

Sm bou (Tll)Olae (Q) 0.8 of setae). All adults of the species of subfamily Sminthurinae

Sm bou 0.6 that I have observed have tibiotarsi. (TI3)3a (T)Q many secondary setae on

Sm bou (TI3)4al (T)Q 0.3 Therefore, the intensive tibiotarsal neochaetosis would be a

Sm bou (TI 1 )4i I (Q) 0.3 derived character of Sminthurinae” (Nayrolles, 1996: 133-

Sm nig (TI3)3a (T)Q 0.4 134). We can thus assume that the setae appearing at the

Sm nig (TI3)4al (T)Q 0.2 fourth instar of Sminthurinae get an earlier instar of emer-

Sm nig (TI3)FSpei (T)Q 0.3 gence in certain species of Sminthurus (especially S. multi-

Sm vir (TI2)3a (T)Q 0,8 punctatus). This hypothesis will be tested “cladistically” in a

vir future Sm (TI 1 )3a (T)Q 0.8 paper.

Sm vir (TI I )4a 1 (T)Q 0.5 The spreading projection combined with a setal cluster is

Sm vir (TI2)4al (T)Q 0.6 shown in Table XL To define setae to be clustered, the simi-

Sm vir (TI3)4aI (T)Q 0.7 larity and ordinal criteria are applied on sets of homeotypic

Sm vir (TI3)2p (T)Q 0.5 setae. For the similarity criterion, we will consider that a

Sm vir (TI 1 )3p (T)Q 0.6 difference inferior or equal to 0.3 between two values is

Sm vir 0.5 it does force the (TI2)3p (T)Q acceptable, i.e., not us to keep setae sepa-

Sm vir (TI3)3p (T)Q 0.2 rate. Comparisons between profiles and other features (e.g.,

Sm vir (TI2)4pl (Q) 0.5 setal shape) show that the setae to be clustered are those

situated Sm leu (TI 1 )4a1 (T)Q 0.4 in the same place on the tibiotarsi (relationship 3 in

Sm leu (TI2)4a 1 (T)Q 0.6 Fig. 5), and/or those situated on the same generatrix (rela-

Sm leu (TI3)4al (T)Q 0.6 tionship 2 in Fig. 5). As far as the anterior and posterior

Sm leu (TI3)2p (T)Q 0.6 generatrices are concerned, we observe that the seta in posi-

Sm leu 1 0.7 tion has the (TI )3p (T)Q 3, e.g. (TIl)3a, an occurrence superior or equal to

Sm leu 0.6 in We thus cluster the (TI2)3p (T)Q seta position 4, e.g. (TIl)4al. can setae

in 4a 1 for tibiotarsus charac- Sm leu (TI3)3p (T)Q 0.3 pairs 3a, or 3p, 4pl every (e.g.,

Sm boz 0.7 6 in Table entails for the (TI2)4al (T)Q ter XI). This process a problem

Sm boz (TI3)4al (T)Q 0.7 numeric projection: the same numeric value can be achieved

Sm boz 0.7 in different As there (TI3)2p (T)Q two ways. a consequence, are more

Sm boz (TI2)3p (T)Q 0.7 characters in the table of the numeric projection (Table XII)

Sm boz (TI3)3p (T)Q 0.6 than in the table of the spreading projection (Table XI). For

Sm boz (TI3)4pl (Q) 0.6 instance, if the setae (TIl)FSp and (TI2)FSp were clustered

Sm boz (TII)FSpei (T)Q 0.8 (characters 16 and 17 in Table XII), the numeric values would

Sm boz (TI2)FSpei (T)Q 0,8 have been the same (2) for Gisinurus and Caprainea, though

Sm bis 0.6 the of both setae is distinct between these (T12)3p (T)Q ontogeny genera

bis raised criterion Sm (TI2)4pl (Q) 0.6 (the problem by the ordinal was discussed for

the numeric projection).

build the table first in order to Differences between third instars In practice, we up alphabetic of Sminthurus can be

definethe set of constant characters, C and with The initial phylogenetically , portrayed ontogenetic trajectories (Fig. 7). p remove it from the rest of the analysis. Generic comparison matrix corresponds to Table XI with the difference that only

can then be achieved by defining the subset of characters the instars distinguishable from each other are retained (they

which differentiate the Table IX the genera. presents results correspond to Operational Semaphorontic Units, see André,

for the studied. In this the standard of setal for each the first instar genera table, 1988; Nayrolles, 1996). Thus, species,

notation is is the Hi the the and the fifth instar the used, e.g., (Tl.)IIi seta considered on (identical to second) (identical to

tibiotarsi. three pairs of Another set of characters is of inter- fourth) are removed. Correspondenceanalyses were performed

est for this is C the set of on some matrices which have been different data systematists; 0, ontogenetically produced by

characters because these characters the initial As it has been constant (Table VIII), treatments applied to array. previ- permit us to distinguish species from the first instar on. ously noticed (Nayrolles, 1996), the most accurate analysis

“Definition In the part and properties of a SIC table” it has corresponds to the splitting of characters in grade and anti-

been stated that the of characters with the same set to be retained for a grade scale grading for all characters. The

subset of taxa T is C , the set of characters inertia of the defined the two first axes is 73% (data a qa phylogenetical- plan by

variable within An of this Axis FI ly T„. example process applied to are highly structured). corresponds to ontogenetic

the Sminthurus is in Table X, The main dif- and F2 differences. Within genus provided differences, to phylogenetic the

ference relates the of for it is obvious that difference lies among species to stage development genus Sminthurus, the main

Downloaded from Brill.com10/10/2021 04:06:52PM via free access Contributions to Zoology, 67 (3) - 1998 215

6. Schematic ofthe The is and setae Fig. representation tibiotarsal chaetotaxy. figure synthetic presents all which canbe observed on

tibiotarsi ofSminthurinae. The setae are schematized as follows: an open symbol for a primary seta, a full symbol for a secondary seta,

a circle for a normal a star for a seta shaped seta, special shaped (oval organ). Ge, Gae, Ga, Gai, Gi, Gpi, Gp, and Gpe are the eight

generatrices in the following positions: external, anterior-external, anterior, anterior-internal, internal, posterior-internal, posterior,

and There are five whorls IV and V The posterior-external. numbered I, II, III, from apex to basis. symbol for a seta on a whorl

combines the numberof the whorl with the of the Ve the Ge. The named letter(s) generatrix, e.g., le. He,... for generatrix basal part is

F, and the letters FP and FS are used to and setae. first designateprimary secondary The whorl bears two setae, Ja and Jp, instead ofone

at the internal which in numbered with place of the generatrix. The secondary setae appear inter-whorls are Arab numerals.

in the position of third instars along the axis FI, i.e., along' but there are other characters (e.g., antennal chaetotaxy) and

the data of other A ontogenetic gradient. species available. future paper will deal with

Table XII could be used to perform a cladistic analysis, this issue as well as polarization of characters.

Downloaded from Brill.com10/10/2021 04:06:52PM via free access 216 P. Nayrolles - Chaetotaxy in ontophylogenetic studies

Table VIII. Alphabetic projection of the initial SIC table corresponding to the tibiotarsal chaetotaxy of 14 species of Sminthurinae.

Taxa are in columns and setae in rows. C C and C are respectively the set of phylogenetically constant characters, the set of p, 0 n

ontogenetically constant characters and the set of characters variable both in ontogeny and phytogeny. The set of interest for the

following is C = C o C q 0 n .

Gi mal Ca bre Ca mar AI fus Al gal Sp bet Sp les Sm bou Sm nig Sm vir Sm leu Sm boz Sm bis Sm mul

C (Tll)Fseî Q Q Q Q Q Q Q Q Q Q Q Q Q Q p

(TI2)FSet Q Q Q Q Q Q Q Q Q Q Q Q Q Q

(TI3)FSeî Q Q Q Q Q Q Q Q Q Q Q Q Q Q

(Tll)3ae Q Q Q Q Q Q Q Q Q Q Q Q Q Q

(TI2)3ae Q Q Q Q Q Q Q Q Q Q Q Q Q Q

(TI3)3ae Q Q Q Q Q Q Q Q Q Q Q Q Q Q

(Til)4ae1 Q Q Q Q Q Q Q Q Q Q Q Q Q Q

(TI2)4ael Q Q Q Q Q Q Q Q Q Q Q Q Q Q

(TI3)4ael Q Q Q Q Q Q Q Q Q Q Q Q Q Q

(TIl)FSa T T T T T T T T T T T T T T

(TI2)FSa T T T T T T T T T T T T T T

(TI3)FSa T T T T T T T T T T T T T T

(Tll)Vai T T T T T T T T T T T T T T

(T12)Vai T T T T T T T T T T T T T T

(TI3)Vai T T T T T T T T T T T T T T

(TI3)3ai T T T T T T T T T T T T T T

(Til Hail T T T T T T T T T T T T T T

(TI2)4ail T T T T T T T T T T T T T T

(TI3)4ail T T T T T T T T T T T T T T

(TI2)4ai2 Q Q Q Q Q Q Q Q Q Q Q Q Q Q

(Tl 1 )4ai2 Q Q Q Q Q Q Q Q Q Q Q Q Q Q

(TI3)4ai2 Q Q Q Q Q Q Q Q Q Q Q Q Q Q

(TIl)FSai Q Q Q Q Q Q Q Q Q Q Q Q Q Q

(TI2)FSai Q Q Q Q Q Q Q Q Q Q Q Q Q Q

(TI3)FSai Q Q Q Q Q Q Q Q Q Q Q Q Q Q

(Tll)Vpi T T T T T T T T T T T T T T

(TI2)Vpi T T T T T T T T T T T T T T

(T13)Vpi T T T T T T T T T T T T T T

(TI3)3pi Q Q Q Q Q Q Q Q Q Q Q Q Q Q

(TI1 )4pi 1 T T T T T T T T T T T T T T

(TI2)4pü Q Q Q Q Q Q Q Q Q Q Q Q Q Q

(T13)4pil 0 Q Q Q Q Q Q Q Q Q Q Q Q Q

1 (TI )4pi2 Q Q Q Q Q Q Q Q Q Q Q Q Q Q

(TI2)4pi2 Q Q Q Q Q Q Q Q Q Q Q Q Q Q

(TI3)4pi2 Q Q Q Q Q Q \q Q Q Q Q Q Q Q

(TIl)FSpi Q Q Q Q Q Q Q Q Q Q Q Q Q Q

(TI2)FSpi Q Q Q Q Q Q Q Q Q Q Q Q Q Q

(T13)FSpi Q Q Q Q Q Q Q Q Q Q Q Q Q Q

(TI3)3i T T T T T T T T T T T T T T

(TI2)4il T T T T T T T T T T T T T T

1 T (TI3)4i T T T T T T T T T T T T T

C (Tll)Ia P — P P P P P P P P P P P p 0

(TI2)Ia P — P P P P P P P P P P P p

(TI3)Ia P — P P P P P P P P P P P p

(Tll)lli — P P P P P P P P P P P P p

(TI2)IIi — P P P P P P P P P P P P p

(TI3)IIi — P P P P P P P P P P P P p

— — — — — — (TI2)Vp — P P P P P P p

— — — (TI3)Vp — — — — P P P P P P p

C (Tll)OIae — Q Q Q Q Q Q (Q) n Q Q Q Q Q Q

(TI2)01ae — Q Q Q Q Q Q Q Q Q Q Q Q Q

(TI3)01ae — Q Q Q Q Q Q Q Q Q Q Q Q Q

Downloaded from Brill.com10/10/2021 04:06:52PM via free access Contributions to Zoology, 67 (3) - 1998 217

Table VIII. Cont.

c o a15 U X u as < tStf) < QJ co (S oo a x> O 3 oo a 'S 00S '> 00 S — CO 00 OO a>- 00 co* x> p, w> 3 3C oO N e n g S3

— w C3 D Q Q o O O' Q Q Q Q

Q Q O O O Q Q Q Q

w rp

(N w rp C3 Q Q Q O O H T T T T

H rp o C3 Q Q Q O O po H H H H

Pp fpca Q Q Q O O po H H H H

H rp rp a Q Q Q T)Q PO H H H H H Po C3 Q oooooooooo Q oooooooooo oooooooooo oooooooooo Q O' O' PO P O H H H PP C3 Q Q Q O' O' PO P O PO H H

Prp 03 Q Q Q ;i)Q P O PO P O ■— O H H

— 1 i 1 1 1 (Q) 1 1 1 1 1 i

h rp

— h rpD, Q Q o Q o Q Q Q PO' o f—• T H

PCM rp Q. Q Q o Q cy Q Q Q O' po po (T)Q H

Prp rp E— H Cu Q Q o Q c/ Q Q Q PO' Po > o Q

P — O. Q Q o Q cy Q Q Q O' O' Q T H

P P O. OOOOOOOH Q Q 1 Q O' Q Q Q o O' Q (Q) H Prp 1 1 1 1 1 o- Q O' Q Q 1 (Q 1 H

p 1 1 u-_ 00D. O 1 1 1 1 1 1 1 1 1

p(N oo 1 u-_ O- o' O 1 1 1 1 1 1 1 1 1 i 1

P p5 D, H a H H H H H H H H H H H H

O PP o. H H H H H H H H E— H H H H o' 1

Prp O c. H E— f— t— H H H H H H H H H a 1

p — ofN H H H H H H H cL o' H H H E— H H

p fN H H H p o Qu o' H H H H H H H H H H

rp P OrN o. H PO H H H H H H H H H H H H

p — orp cL O' Q H H H H H H H H H H H

pporp o- o Q H H H t— H H H H H H H

f-H T H H H Prp orp o. O H H H H H H H H

p Tf o o. o Q H H H H H H H H (- H H

O p ■^r o Q H H H H H H H H H H H

H H H H H H H rp O a. o Q H H H H —^ P 00 ooooooooo o Q O' o Q Po Q O H H H o H H P fN Uh 00 — H H H a. > o Q o o Q Po Q O o H H

Prp Uhoo a. o o Q o o Q o Q PO H H H H H

Table IX. Distinction of The characters are constant within each and variable between genera. genus genera.

Gisinurus Caprainea Allacma Spatulosminlhurus Sminthurus

(TI.)lIi — P P P P

(TI2,3)Vp — — — — P

(TI2,3)01ae — Q Q Q Q

(TIl,2)2ae — Q Q Q Q

(TI3)2ae (Q) Q Q Q Q

T — (TIl)FSp Q —

— — — (TI2)FSp Q —

(Tll,2)03pe Q Q * T T T

(TI.)04pe Q Q T T T

Downloaded from Brill.com10/10/2021 04:06:52PM via free access 218 P. Nayrolles - Chaetotaxy in ontophylogenetic studies

Table X. Distinction of the species of Sminthurus.

Sm bou Sm nig Sm vir Sm leu Sm boz Sm his Sm mul

H OX. C3 (¡J O' — o Q Q Q Q H m (N o f— T T T T H —

H m «5 (T)Q (T)Q H H H H H

h Q Q T T T H (N— Q Q PPPgag PPPgga (T)Q T T h n (T)Q (T)Q (T)Q T T

1 1 1 1 1 — tF (Q) 1

m (N i a. O' £o o (T)Q Q

— n T T i ttt a. o £o o (N n i o. o £o o (T)Q (T)Q

n i n a, o £o o (T)Q Q

— t T i btt a. o o ttttao Q (N 't i d. o s Q (Q)

- 1 1 1 1 B D, Q o

U- C/3 H H H 1 E a (U oooooooooo Q £ — _oo t fN U- C/5 o.

p rn U- GO Ü — H H H H H a > (T)Q

Table XI. Spreadingprojection with homeotypic setae clustered on the basis of similar or ordered profiles. Each species is divided into five instars, the last is the adult. Characters with their setal components are listed below.

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18

Gi mal 3 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

3 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

3 0 0 0 0 0 0 0 0 0 0 0 0 1 6 0 0 0

3 0 0 0.8 1 2 2 2 0 1 2 2 2 1 6 6 2 1

3 0 0 0.8 1 2 2 2 0 1 2 2 2 1 6 6 2 1

Ca bre 0 3 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

0 3 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

0 3 0 0 0 0 0 0 0 0 0 0 0 0 0.5 0.6 0 0

0 3 0 6 1 2 2 2 0 I 2 2 1 2 6 6 2 1

0 3 0 6 1 2 2 2 0 1 2 2 1 2 6 6 2 1

Ca mar 3 3 0 0 0 0 0 0 0 \ 0 0 0 0 0 0 0 0 0 3 3 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

3 3 0 0 0 0 0 0 0 0 0 0 0 0 6 1 0 0

3 3 0 6 1 2 2 2 0 1 2 2 1 2 6 6 2 1

3 3 0 6 1 2 2 2 0 1 2 2 1 2 6 6 2 1

Al fus 3 3 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

3 3 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

3 3 0 0 0 0 0 0 0 0 0 0 0 0 6 6 0 0

3 3 0 6 1 2 2 2 0 1 2 I 1 0 6 6 2 1

3 3 0 6 I 2 2 2 0 1 2 1 1 0 6 6 2 1

AI gal 3 3 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

3 3 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

3 3 0 0 0 0 0 0 0 0 0 0 0 0 6 6 0 0

3 3 0 6 1 2 2 2 0 1 2 2 2 0 6 6 2 1

3 3 0 6 1 2 2 2 0 1 2 2 2 0 6 6 2 1

Sp bet 3 3 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

3 3 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

3 3 0 0 0 0 0 0 0 0 0 0 0 0 6 6 0 0

Downloaded from Brill.com10/10/2021 04:06:52PM via free access - Contributions to Zoology, 67 (3) 1998 219

Table XI. Cont.

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 (

3 3 0 6 1 2 2 2 0 1 2 2 2 0 6 6 2 1

3 3 0 6 1 2 2 2 0 1 2 2 2 0 6 6 2 1

Sp les 3 3 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

3 3 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

3 3 0 0 0 0 0 0 0 0 0 0 0 0 6 6 0.7 0

3 3 0 6 1 2 2 2 0 1 2 2 1 0 6 6 2 1

3 3 0 6 1 2 2 2 0 1 2 2 1 0 6 6 2 1

Sm bou 3 3 2 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

3 3 2 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

3 3 2 0 0 0 0 0.9 0 0 0 0 0 0 6 6 0 0

3 3 2 5.8 1 2 2 2 0.3 1 2 2 2 0 6 6 2 1

3 3 2 5.8 1 2 2 2 0,3 1 2 2 2 0 6 6 2 1

Sm nig 3 3 2 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

3 3 2 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

3 3 2 0 0 0 0 0.6 0 0 0 0 0 0 6 6 0 0.3

3 3 2 6 1 2 2 2 0 1 2 2 2 0 6 6 2 1

3 3 2 6 1 2 2 2 0 1 2 2 2 0 6 6 2 1

Sm vir 3 3 2 0 0 0 0 0 0 0 o • 0 0 0 0 0 0 0

3 3 2 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

3 3 2 0 1 1.3 1.4 1.7 0 0.5 0.6 0.5 0.2 0 6 6 2 1

3 3 2 6 1 2 2 2 0 1 2 1.5 1 0 6 6 2 1

3 3 2 6 1 2 2 2 0 1 2 1.5 1 0 6 6 2 1

Sm leu 3 3 2 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

3 3 2 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

3 3 2 0 1 1.4 1.6 1.6 0 0.6 0.7 0.6 0.3 0 6 6 2 1

3 3 2 6 1 2 2 2 0 1 2 2 1 0 6 6 2 1

3 3 2 6 1 2 2 2 0 1 2 2 1 0 6 6 2 1

Sm boz 3 3 2 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

3 3 2 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

3 3 2 0 1 2 1.7 1.7 0 0.7 1 0.7 0.6 0 6 6 1.6 1

3 3 2 6 1 2 2 2 0 1 2 2 1.6 0 6 6 2 1

3 3 2 6 1 2 2 2 0 1 2 2 1.6 0 6 6 2 1

Sm his 3 3 2 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

3 3 2 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

3 3 2 0 I 2 2 2 0 0 2 0.6 0 0 6 6 2 1

3 3 2 6 1 2 2 2 0 1 2 1.6 1 0 6 6 2 I

3 3 2 6 1 2 2 2 0 1 2 1.6 1 0 6 6 2 1

Sm mul 3 3 2 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

3 3 2 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

3 3 2 0 1 2 2 2 0 1 2 2 2 0 6 6 2 1

3 3 2 6 1 2 2 2 0 1 2 2 2 0 6 6 2 1

3 3 2 6 1 2 2 2 0 1 2 2 2 0 6 6 2 1

character set of setae number character set of setae number character set of setae number

number of setae number of setae number of setae

1 (Tl.)Ia 3 7 (T12)3a & (T12)4al 2 13 (TI3)3p& (TI3)4p 1 2

2 (Tl.)IIi 3 8 (TI3)3a & (TI3)4al 2 14 (TIl,2)FSp 2

3 (T12,3)Vp 2 9 (TI 1)4i 1 1 15 (Tl.)Olpe & (TI.)02pe 6

4 (Tl.)Olae& (TI.)2ae 6 10 (TI3)t2p 1 16 (Tl.)03pe& (TI.)04pe 6

5 (T13)2a 1 11 (Tll)3p & (TIl)4pl 2 17 (TIl,2)FSpe-L 2

6 (TIl)3a & (Tll)4al 2 12 (T12)3p & (T12)4pl 2 18 (TI3)FSpe4 1

Downloaded from Brill.com10/10/2021 04:06:52PM via free access 220 P. Naywiles - Chaetotaxy in ontophylogenetic studies

Fig. 7. Ontogenetic trajectories of Sminthurinae.Open symbols: second instars; shaded symbols: third instars; solid symbols: fourth

instars. First instars have the same chaetotaxy as second instars, and adults (fifth instars) have the same chaetotaxy as fourth instars.

Table XII. Numeric projection with clustering of homeotypic setae. Characters with their setal components are listed below

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21

selected P P P Q T T T T Q T T T T + Q T T + Q T + Q Q T T T T

instars

Gi mal 3 0 0 0.8 0 0 0 0 0 0 0 0 1 0 1 2 0 6 0 C 1 0

Ca bre 0 3 0 6 0 0 0 0 0 0 0 0 1 0 0 1 1 0.5 0.6 C 1 0

Ca mar 3 3 0 6 0 0 0 0 0 °\ 0 0 1 0 0 1 1 6 I C 1 0

Al fus 3 3 0 6 0 0 0 0 0 0 0 0 0 0 0 0 0 6 6 C I 0

Al gal 3 3 0 6 0 0 0 0 0 0 0 0 1 0 1 0 0 6 6 C 1 0

Sp bet 3 3 0 6 0 0 0 0 0 0 0 0 1 0 1 0 0 6 6 C 1 0

Sp Ies 3 3 0 6 0 0 0 0 0 0 0 0 1 0 0 0 0 6 6 0.7 0

Sm bou 3 3 2 5.8 0 0 0 0.9 0.3 0 0 0 1 0 1 0 0 6 6 C 1 0

Sm nig 3 3 2 6 0 0 0 0.6 0 0 0 0 1 0 1 0 0 6 6 C 1 0.3

Sm vir 3 3 2 6 1 1.3 1.4 1.7 0 0.5 0.6 0.5 0.5 0.2 0 0 0 6 6 2i 1

Sm leu 3 3 2 6 1 1.4 1.6 1.6 0 0.6 0.7 0.6 1 0.3 0 0 0 6 6 2! 1

Sm boz 3 3 2 6 I 2 1.7 1,7 0 0.7 1 0.7 1 0.6 0.6 0 0 6 6 1.6 1

Sm bis 3 3 2 6 1 2 2 2 0 0 2 0.6 0.6 0 0 0 0 6 6 2! 1

Sm mul 3 3 2 6 1 2 2 2 0 1 2 1 2 1 2 0 0 6 6 2; i

character set of setae number character set of setae number character set of setae number

number of setae number of setae number of setae

1 (TI,)Ia 3 8 (TI3)3a & (TI3)4aI 2 15 (TI3)4pI 1

2 (Tl.)Ili 3 9 (111)411 1 16 (TIl)FSp 1

3 (TI2,3)Vp 2 10 (TI3)2p 1 17 (TI2)FSp 1

4 (Tl.)Olae & (TI.)2ae 6 11 (TIl)3p & (TIl)4pl 2 18 (Tl.)Olpe & (TI.)02pe 6

5 (TI3)2a 1 12 (TI2)3p 1 19 (TI.)03pe & (TI.)04pe 6 6 (TI 1 )3a & (TI 1 )4al 2 13 (TI2)4pl 1 20 (TIl,2)FSpei 2

7 (TI2)3a & (TI2)4al 2 14 (TI3)3p 1 21 (TI3)FSpei 1

Downloaded from Brill.com10/10/2021 04:06:52PM via free access