Bijdragen tot de Dierkunde, 63 (3) 149-161 (1993)
SPB Academie Publishing bv, The Hague
Phylogeny and biogeography of the Niphargus transitivus group of species
(Crustacea, Amphipoda)
Boris Sket & Jos Notenboom
of Biology, Biotechnical Faculty, University of Ljubljana, Askerceva 12, 61000 Ljubljana,
2 Slovenia; National Institute of Public Health and EnvironmentalProtection, P.O. Box 1,
3720 BA Bilthoven, The Netherlands
Keywords: Amphipoda, Niphargus, phylogeny, zoogeography
Abstract 1. Introduction
The transitivus includes 12 The transitivus introduced Niphargus group groundwater- Niphargus groupwas by
dwellingspecies. It is defined by an uprightbody position during Sket (1971) to delimitate four species of Niphargus locomotion,a small, stout body with short appendages, except Schiödte, 1849, which on the one hand were all very V-VI for long coxae and dactyli, coxae equilobateor posterolo- aberrant, but on the other hand intermediate in bate, accessory flagellum subject to reduction, and the body en-
some characters between N. kochianus Bate, 1859, dowed with adaptations for volvation ( = ability to enroll into
and Karamaniella 1962. The latter a ball). Sket, genus was
The of this southeastern European of biogeography group erected for K. pupetta Sket, 1962, a species which species is discussed and a cladistic analysis of the group is pre- did not fit withinthe Niphargus diversity spectrum sented. Some cases of parallel character evolution in ground- at the time. At present Karamaniella is considered water amphipodsare mentioned. The inference of biogeograph- with Niphargus, and K. has ic and phylogenetic information in elucidating the evolutionary synonymous pupetta
of the is ofthe data. been into the N. transitivus history group hampered by inconsistency incorporated group(cf.
One of the hypotheses about the evolutionary history of the Sket, 1972: 106, 111; Karaman, 1975). For the group is thought to be the most parsimonious. related but morphologically very distinct K. para-
doxa(Sket, 1964) the genus CarinurellaSket, 1971,
Résumé has been established.
Since establishment of the transitivus N. group, Le transitivus du 12 groupe genre Niphargus comprend espèces several additional taxa have been described by souterrains. Il est défini surtout par la positionverticale du corps Dancàu (1971), Karaman (1972, 1984a, 1989), and pendant le déplacement,par un corps petit et trapu aux appen- Karaman& Sket after dices des les (1991). Moreover, reanalysis courts (à l’exception longues coxae et dactyli), par of the known but de- coxaeV-VI equilobés ouposterolobés, par le flagelleaccessoire previously insufficiently
à réduction le doué sujet une accusée, et par corps d’adaptions scribed N. kochianus longidactylus Ruffo, 1937, it
la volvation. permettant transitivus appeared to belong to the N. group as La biogéographiede cegroupe d’espèces d’Europe du S.E. est well (cf. Sket, 1971; Karaman, 1981). discutée et une analyse cladistique est présentée. Sont men- tionnés certains cas d’évolution parallèle des caractères chez des
souterraines. L’utilisation de l’information bio- Amphipodes Altogether 12 nominal species are now consid- géographiqueet phylogénétique dans le but de l’élucidation de ered as the N. transitivus group, and we decided to l’histoire évolutive du groupe est rendue difficile par le caractère treat the systematics, phylogeny, and biogeography contradictoire des données. Une des hypothèses sur l’histoire of this of évolutive group species more properly as du groupe est considérée comme étant la plus par- Niphargus cimonieuse. a whole. 150 B. Sket & J. Notenboom - Phytogeny and biogeography of the Niphargus transitivus group
2. Definition and peculiarities of the Niphargus west Bosnia (Ruffo, 1937; D'Ancona, 1942a& b;
Sket & transitivus group Benedetti, 1942; Karaman, 1981;
Velkovrh, 1981).
2.1. List of species and distribution • N. transitivus transitivus Sket, 1971; found
directly in interstitialwaters and in wells along Pi-
the N. and Torrein northeast also The following taxa compose transitivus ave Italy, occasionally
group: in caves at the karst margins in the same area
• N. carcerarius G Karaman, 1989; two Karaman, . only speci- (Sket, 1971; 1975, 1984a, 1985).
• dissonus found mens (!) found in a small cave lake near Pljevlja N. transitivus G . Karaman, 1984;
(cave of Sljivanski potok near Djurdjevica Tara), in wells in the Bergamo region, northern Italy
northern Montenegro (Karaman, 1989). (Karaman, 1984a, 1985). Sket (1971) opened the
• N. asper G. Karaman, 1972; only one specimen question of a possible conspecifity of N. transiti-
in well (!) found a near Podgorica (= Titograd), vus with N. pupetta, according to the great varia-
Montenegro (Karaman, 1972). bility of some N. transitivus populations. Kara-
• N. factor G. Karaman &Sket, 1991; inhabitantof man (1975) denied this possibility since he found
small cave lakes in lower parts of Popovo polje no intermediatespecimens in mixed populations,
(Vjetrenica at Zavala; Baba pecina at Strujici), but we think that more material should be ex-
southeast Hercegovina (Karaman & Sket, 1991). amined before the last decision is conclusive.
This is similar to both • N. species very N. asper, may pupetta pupetta (Sket, 1962) (syn.: Karama-
be varieties of a single species. niella pupetta Sket, 1962); found in interstitial
• N. brevirostris Sket, 1971; only two specimens (!) waters and wells in Bergamo and probably in
found in interstitialwaters of the Gacko polje (at Belluno regions in north and northeast Italy, as
Licko in southwest Croatia well in northwest Lesee) Lika, (Sket, as (Bovec, new data), central,
1971). and eastern Slovenia (Sket, 1962,1972; Karaman,
• N. numerus G. Karaman & Sket, 1991; only one 1975, 1984a).
specimen (!) found in a small residual lake of an • N. pupettaparapupetta G. Karaman, 1984 (syn.:
River and active cave near Zrmanja (Cavlinska peci- N. parapupetta in Karaman, 1984a, Kara-
northern in found in in- na), near Obrovac, Dalmacija (Kara- maniellapupetta Dancâu, 1972);
man & Sket, 1991). N. brevirostris and N. numer- terstitial waters and wells near Zagreb, western
us also show many similarities pointing to their Croatia, and near the river Dunarea-Dunav
possible conspecifity. (Danube) in southwest Rumania. This taxon was
• Since it N. pectinicauda Sket, 1971; inhabitantof intersti- described as a separate species. seems to
tial of the Sava River between and be and the waters Bohinj completely allopatric to N. p. pupetta,
Ljubljana, northwest Slovenia (Sket, 1971; Sket differences are extremely small, we now suggest
& Velkovrh, 1981). that both taxa are conspecific.
• N. rostratus Sket, 1971; found in small lakes of a • Carinurella paradoxa (Sket, 1964) (syn.: Kara-
cave near Cetina River (Dragica pecina at Malj- maniella paradoxa ■ Sket, 1964); found in intersti-
kovo), central Dalmacija, Croatia (Sket, 1971). tial waters of the lower reaches of Soca-Isonzo
• N. alutensis Dancâu, 1971; probably an inhabi- and its tributaries: Vipava, Torre-Ter, and
of tant interstitial waters of the Olt River valley Nadiza-Natisone, all on both sides of the Slo-
(found in wells, village Jilbea), south-central veno-Italian boundary (Sket, 1964, 1971, 1972).
Rumania attributed this (Dancâu, 1971). Although to a separate genus, spe-
• N. longidactylus Ruffo, 1937; inhabitant of inter- cies has drawn much concern. The acceptability
stitial able build waters, to large populations in of making the genus(Niphargus) paraphyletic has
wells near rivers outside open karst areas, in been discussed elsewhere (Sket, 1990).
northeast Italy, and in central and northeast
Slovenia, northwest Croatia, central and north- In conclusion, therepresentatives of the N. transiti- - Bijdragen tot de Dierkunde, 63 (3) 1993 151
transitivus vus group are not all sufficiently characterized, 2.2. Characterization of the N. group
primarily because of lack of material. Information
about morphological variation within and between 2.2.1. Analysis of characters. - The most constant
adjacent populations is lacking and the possibility characters of the Niphargus transitivus group
nominal in- size that some of the species are merely (some apparently synapomorphies) are: body
traspecific varieties must be considered. small to very small, stout and upright-walking; cox-
The group includes taxa that are evidently mem- ae of pereiopods II—III very deep, at least twice as
of in of bers the benthos larger water bodies the deep as wide; coxa of pereiopod V equilobate or
karst; a quite remarkable feature, since in the posterolobate, that of VI posterolobate; bases of
Dinaric karst such an ecology is mostly exhibited by posterior pereiopods broad, with widely rounded
larger niphargids of a centimetre or larger. Other distoposterior lobes; dactyli of pereiopods very
of of the members the group inhabit interstitial waters in slender and long; gnathopods basically N.
river also in lon- kochianus of wider valleys, some hilly country (N. type (synapomorphy a group?).
gidactylus), but all species may build denser popu- However, there exists a number of characters,
lations in open wells - eitherin response to greater present in some of the taxa, that are functionally
richer food that above-mentioned characters and space or to supplies are usually connected by the
present there. Therefore, the group as a whole, as absent in other niphargids.
well as some of the species, appears as quite ubiqui- We will try to evaluateall specific character states
animals. tous stygobiontic Some species (N. bre- appearing in the N. transitivus group. Character
N. and found in is determined virostris, numerus, N. asper) were state polarity mainly by comparison
so few numbers that their ecology remains un- with species of the N. kochianus complex which ap-
certain. parently is the most closely related to the TV. tran-
The known distribution of the reaches and considered sister area group sitivus group as a group.
from Bergamo in northern Italy to the Olt River in Habitus: The body is small to extremely small,
Rumania; below the southern slopes of the Alps short and broad, not depressed; the stout appear-
and the of along southern parts the Danube system, ance is emphasized by deep coxae and short distal
well in the Dinaric karst. While of of with as as cave species parts appendages. N. carcerarius, a length
the Dinaric karst seem very limited in distribution, of 11 mm or more, is by far the largest member, but
ranges of interstitial species may be extraordinarily its taxonomie position within the group is uncer-
extended. Comparable zoogeographical connec- tain. All other taxa are only 2-6 mm long, being
tionsbetween northeastern Italy and southern parts amongthe smallest of the niphargids. The urosome
of the Danube system have already been noted in has been greatly modifiedin Carinurellaparadoxa.
other in groups (e.g., Sket, 1972, and prep.). They Rostrum: Some species of this group are excep-
under influence of tions obviously originated cor- among niphargids, having a conspicuous ros-
responding paleogeographical events. trum. In most taxa it is only slightly indicated. In
Thereexists number of in a taxa theJV. kochianus N. brevirostris it is well developed, and it is exceed-
with members of group, showing some similarities ingly long (equal to lateral cephalic lobes) in N. ros- the N. transitivus group (e.g., slender dactyli, see tratus. As this feature was neglected until the dis-
TV. wolfi Schellenberg, 1933, and N. “kochianus” covery of this group, it is impossible to establish its melticensis Dancâu & Andreev, 1973). Supposedly condition in other niphargids. the similarities of these taxa are homoplasies, since Epimera : The ventroposterior angles of the a high probability exists for convergent character epimera in niphargids are rounded or obtuse, to evolution in niphargids (mosaic evolution). Of acute. Either of these shapes seem to have devel- course, prior to a detailed analysis of the whole ge- oped convergently, but it is often quite constant for nus it is impossible to know whetherthe N. transiti- larger groups of taxa. In the group studied, only vus is delimited group correctly (i.e., if the group is N. carcerarius exhibits rectangular angles, whereas monophyletic). epimeral plates in all other species end in acute 152 B. Sket & J. Notenboom - Phytogeny and biogeography of the Niphargus transitivus group
angles. With the above-mentioned exception, in all of maxilla I is variable in niphargids. Most mem-
exhibit of species of the N. transitivus group the ventral (i.e. bers of the group studied a low number
of II—III with of N. carcerarius which has apical) parts epimera (sometimes I—III) setae, exception
direction. than five are strongly produced in ventrocaudal more setae.
Submarginal spines along the ventral margins of The denticulationof spines on the outer lobe of
in modest number in maxilla I be the epimera, present a most seems to a quite variable character,
of niphargids, may completely disappear. perhaps enabling even populations (subspecies)
themselves feed- Telson: The telson is always longer than wide, some species to adapt to particular
and conditions. In the with a deep incision narrow lobes, usually pro- ing analyzed group usually none vided with distal spines and dorsal pairs of pennate or only the innermost spine is comb-like, but in
there setae. In some species, the telson is extremely elon- some taxa (N. longidactylus, C. paradoxa)
three such gate, with very short spines (e.g., N. transitivus); may be or more spines. dorsal setae are sometimes moved to a nearly distal Pereiopods I—IV: Coxae of anterior pereiopods position. N. longidactylus, an exception in this (particularly II-III) are in most niphargids only up group, has comparatively long dorsal spines; we to 50 percent as long as wide. In all species of the
least twice suggest this is of secondary appearance, and there- N. transitivus group, they are at as long
wide. The wideneddistal fore an autapomorphy. as probably primarily part
Antenna I: The basal article of antenna I tends of coxa IV (in N. alutensis, N. rostratus, and N. to extend dorsodistally forming a "visor". Such a carcerarius) is in this group mostly only slightly
visor seems to be lacking in some taxa (e.g., N. pec- widened (in N. asper and N. longidactylus) or bow-
tinicauda, N. asper), while only slightly marked in shaped and parallel-sided (in remaining species).
Article 6 is and others (e.g., N. alutensis, N. rostratus), and ex- Gnathopods : originally elongated tremely developed in N. pupetta and C. paradoxa. subrectangular, with subparallel anterior and
In certain populations of N. transitivus (from the posterior margins. In some taxa of the N. transiti-
Piave River; cf. Sket, 1971; Karaman, 1985), the vus group, the margins converge in proximal direc-
the article The antenna I is highly variable. The antenna I also tion, giving a triangular shape. num-
in reduced ber is reduced tends to shorten its flagellum, resulting a of transverse posterior rows of setae number of articles. in such cases, mostly to only two. The palmar mar-
All niphargid amphipods exhibit a reduced acces- gin of gnathopod II makeswith the anterior margin sory flagellum, generally with two articles. The dis- of the article in most niphargids an acute or right tal article may be very short in several species, such angle, which angle may become obtuse by bulging
in the central and/or = of as in N. carcerarius and particularly N. pec- of posterior ( angular) part
and in In the in distal direction. tinicauda, N. longidactylus, N.J asper. palma brevirostris, it is shorter than the diameter of the In some species, the length of article 5 increased proximal article. In N. factor, it is present or miss- in relation to article 6 and may become even re-
and ing. In.N. rostratus, N. transitivus, TV. pupetta, markably longer than article 6. The originally the to be uniartic- number of article accessory flagellum seems always higher setae on the dactyli ( = 5) ulate and sometimes very short. is in this group (and in many other small species)
AntennaII\ N. rostratus and N. brevirostris seem usually reduced to 1-3.
be The of the to an exception among niphargids, having aes- Pereiopods V- VI, coxae: coxae thetascs on both antennae I and II. posterior pereiopods are in nearly all other niphar-
Mandibularpalp : Article 3 is longer than article gids anterolobate.In the taxa includedhere, coxa V
2 in most niphargids. In this group, it tends to be- is equilobate or posterolobate, while coxa VI is al-
with reduced come shorter setation. This is not ways posterolobate. The anterior lobe in coxa VI is
in N. found N. rostratus, pectinicauda, and N. car- usually lacking.
bases These articles have cerarius. Pereiopods V- VII, : a
Maxilla I: The number of the inner rounded lobe in setae on lobe distoposterior many niphargids. Bijdragen tot de Dierkunde, 63 (3) - 1993 153
"trend" of this of Some members of this group, however, have at lutionary group species, are re-
with additional lated the and the least pereiopod V developed an an- to upright poise following ability
in almost enrol less into ball terodistal lobe, resulting an rectangular to more or an irregular (volva-
basis. While the posterior margins of bases VI—VII tion). This upright position seems to be a general
this of are sometimes broadly rounded as well, in N. lon- trait of group species.
In the transition from the usual loco- gidactylus the bases V-VII are almost rectangular. niphargids,
Pereiopods V—VII, distal articles: Article 4, motion on the flank to the upright poise is connect-
flattened originally narrow and rod-shaped or with a moder- ed by widening of the previously laterally
the the ate posterior bulge, may develop a large posterior body. In N. transitivus group, upright poise
lobe, making the article very broad. In some spe- is facilitatedalso by a shortening of the entirebody.
cies, the posterior pereiopods tend to shorten their Shortening of the distal parts of some appendages,
relation distal parts (we expressed these states in the like antennae and posterior pereiopods, seems to be
between lengths of articles 2 and 4 + 5 + 6 in pereio- the next step in this direction, achieved only by
pod VII). However, most taxa of this group have some members of the group. Probably, the deep an-
particularly thin dactyli, mostly due to elongation terior coxae and the particularly slender dactyli of
of of nails inrelation to socles. The elongation of dac- the posterior pereiopods are a part this tendency
tyli in relation to article 6 is probably only a result as well.
of the nail elongation. Dactyli have only in N. pec- Posterolobate posterior coxae are extremely rare
allow tinicaudaand N. brevirostris a strong innerspine or characters in niphargids. Supposedly they
it is reduced another of the than spine-like seta; in most species to a direction pereiopod movement
fragile seta. is possible in the anterolobate ones. However, this
Uropods I—II: Only in N. pectinicauda the origi- supposition must be experimentally confirmed.
considers nally moderate number of spines along each ramus Bousfield (1983) posterolobate coxae has increased. plesiomorphous, a supposition which may be cor-
Uropods III: Originally they are moderately rect for Amphipoda in general, but within the genus long, but they tend to become particularly short, Niphargus it is correlated with so many other
of lateral that it has been resulting also in a reduced number spine peculiarities (see below), evidently
the its achieved The next in the same groups. Seemingly, exopodite developed secondarily. step shape by progressing from flattened through rod- direction has obviously been an "inversion" of the shaped to subcorneal; unfortunately, the degree of pereiopod V, or/and VI, carried out by the de-
is sometimes difficult This limb of distal lobe the anterior side of flatness to assess. velopment a on
extreme connected their bases. In this also the underwent an shortening, by context, we may explain further reduction of the terminal article, in convexity of the anterior margin and straightening
Carinurella paradoxa. of the posterior one.
The above-mentioned morphological modifica-
2.2.2. Functional relation of characters. - Some tions enable some taxa to roll into a more or less ir-
ball. is in fact characters, like pectinate spines on maxilla I and regular This volvation occasionally aesthetascs theresult of recorded in N. alutensis Volva- on antenna II, are probably (cf. Dancâu, 1972). special adaptations. Others, like the reduction of tionhas been developed furtherin N. pupetta by the
characteristic transformation of accessory flagella on antenna I and reduction of in- the proximal arti-
result of neutral cle in well of ner spines on dactyli, could be the antenna I, as as by widening proximal mutations. Some particularities in the shape of the articles and shortening of distal ones (except for gnathopods as well as reductive processes in the dactyli) in hind pereiopods. The niphargids' last
article achievement in this direction is C. with mandibular palp's terminal might be ex- paradoxa its plained by paedomorphosis, and this is supported completely modifiedposterior part of the body and by the morphology of juvenile niphargids. The the stronger incrustation of the integument. Unfor-
of other is most extraordinary characters, designating the evo- tunately, the ethology species virtually 154 B. Sket & J. Notenboom - Phylogeny and biogeography of the Niphargus transitivus group
Table I. Characters used for phylogeneticanalysis of the Niphargus transitivus group and their plesiomorphous and apomorphous states
= 1 = state. (for further explanation, see text): 0 plesiomorphous state, apomorphous
1. Body length:
a. (0): > 10 mm; (1): < 8 mm.
b. (0): > 3 mm; (1): < 3 mm.
2. Rostrum:
(0): indistinct or very small; (1) longer than 50°7o of lateral head lobe.
3. Epimeron III, angle:
(0) right; (1) acute.
In the construction ofthe cladogramit has been supposed that the seemingly plesiomorphous state in Carinurella i has been achieved
extensive of the entire abdominal in this secondarily; this supposition is supported by an remodeling part species.
4. Epimera:
(0) normal; (1) extended in ventrocaudal direction. For Carinurella as in "3".
5. Spines of ventral margins of epimeron III:
(0) 2-3 spines; (1) 0-1 spine.
6. Urosome:
(0) normal; (1) greatly modified.
7. Spinulation of telson:
(0) dorsal spines lacking; (1) dorsal spines present.
8. Pennate setae on telson:
(0) in or near the middle of the telson length; (1) apically or subapically.
9. Antenna I, basal article:
(0) without or with small "visor"; (1) visor longer than diameter of article, or variable.
10. Antenna I, number of flagellar articles:
(0) 9 or more; (1) 8 or less.
11. Antenna I, accessory flagellum:
(0) regularly with 2 articles; (1) 1 article or variable.
12. Antenna II, aesthetascs:
(0) absent; (1) present.
13. Mandibular palp article 3:
(0) length > article 2; (1) length < article 2.
14. Maxilla I, inner lobe:
(0) 1-3 apical setae; (1) 5 or more apical setae.
15. Maxilla 1, outer lobe:
(0) none or only innermost spine comblike; (1) 3 or more comblike spines.
16. Coxae II—III:
(0) twice as long as wide; (1) more than twice as long as wide.
17. Coxa IV:
(0) distoposterior lobe large; (1) coxaonly slightly wider in its distal part or parallel-sided.
18. Gnathopods, article 6:
(0) subrectangular, with subparallelanterior and posterior margins; (1) margins converge proximally, giving the article a triangular
shape.
19. Gnathopod II, article 5:
(0) up to 110% of length of article 6; (1) longer.
20. Gnathopods, dactyli:
than (0) more 5 outer setae; (1) 1-3 outer setae.
21. Gnathopod II, palmar margin:
(0) makes with anterior margin of article an acute or right angle; (1 ) mentioned angle obtuse by bulging of the central or posterior
(= angular) part of palma in distal direction.
22. Coxa V:
(0) equilobate; (1) posterolobate.
23. Pereiopod V, article 2:
(0) without an anterodistal lobe; (1) with anterodistal lobe.
24. Pereiopod V, article 2:
(0) posterior margin convex; (1) posterior margin straight. - Bijdragen tot de Dierkunde, 63 (3) 1993 155
Table I. Continued.
25. Pereiopod VII, article 2:
(0) posterior margin convex; (1) posterior margin straight.
26. Pereiopod V-VII, article 4:
Vj (0) narrow; (1) with a posterior lobe, making its width equal to or larger than of the length.
27. Pereiopods VII, relation between lengths of articles 2 and 4+ 5 + 6:
(0)art. 2 < (art. 4+ 5 + 6); (1) art. 2 > (art. 4 + 5 +6).
28. Pereiopods V-VII, dactyli:
(0) nail shorter than socle; (1) nail equal to or longer than socle.
29. Pereiopods V-VII, dactyli:
with inner with tender (0) spine or spine-like seta; (1) seta.
30. Pereiopod VII, dactylus:
(0) article 7 shorter than 'A of article 6; (1) it is longer.
31. Uropod I, endopodite:
5 6 (0) with or less spines along its upper side; (1) with or more of such spines.
32. Uropod III, proximal article of exopodite:
number of lateral side than number of such is 3 less. (0) spine groups on one more 3; (1) groups or
unknown, except for N. longidactylus, which is not to N. pupetta or even to C. paradoxa.
The withinthe able to vol vate. only partial parallel genus Niphar-
Carinurella has been observed in the N. steueri Schellen- It seems plausible that developed gus from a N. pupetta-like ancestor by some new aut- berg, 1935 "Rassenkreis". N. steueri probably apomorphies and character reversals (characters 3 belongs to the wider N. orcinus Joseph, 1869 aggre-
which All its characterized and 4, see Table I), in the modifiedbody ap- gate. races are by compara- parently received another function. tively short antennae and pereiopods, deep anterior
coxae, wide bases of posterior pereiopods, and
- of 2.2.3. Some analogies. There are a number rather narrow gnathopod article 6; making this amphipod taxa that have converged on the N. tran- comparatively large species resembleplesiomorphic sitivus group either by retaining or regaining some species of the N. transitivusgroup. In some races of characters in connection with similar movement N. steueri, however, the coxae V-VI are equilobate
of them is the and the basis of techniques. A good representative or posterolobate pereiopod VI pos-
Synurella ambulans (F. Miiller, 1846) "Rassen- sesses a small anterodistal lobe (cf. Karaman, kreis" with a similar body shape, long anterior and 1984b). In N. steueri liburnicus G. Karaman &
shortened theanterodistal lobe is posterolobate posterior coxae, uropods Sket, 1989, developed on all
III, and in some populations also shortened arti- posterior pereiopods (V-VII) and uropod III is cles 4-6 and long dactyli of posterior pereiopods. shorter than in any other known population of the
This variable An interesting convergence with the N. transitivus species (Karaman & Sket, 1989). spe- group is also shown by Parapseudoniphargus bae- cies of the centimetre category will be the most tis Notenboom, 1988 in the Pseudoniphargus- suitable subject for experimentation, which could
functional of Parapseudoniphargus complex and particularly by show us the nature the transforma-
Parasalentinella rouchi Bou, 1971 (Bou, 1971; tions in posterior pereiopods.
Notenboom, 1988a). The latter has obviously de- scended from a Salentinella-like ancestor, which resembled some taxa of the group analyzed here 2.3. Cladistic analysis in a number of characters. In P. rouchi, in con- nection with volvation, there are also "visors" on 2.3.1. Selection of character states. - An attempt the of proximal articles antenna I, and extremely was made to establish phylogenetic relationships shortened caudal it similar appendages, making within the group by cladistic analysis. According to 156 B. Sket & J. Notenboom - Phytogeny and biogeography of the Niphargus transitivus group
Table II. Distribution ofcharacter states in the Niphargus transitivus group (for description and definition of characters and their states,
= = ? = character state unknown. see Table I and text): 0 plesiomorphous state, 1 apomorphous state,
la lb 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
CA N. carcerarius 00000000000000100
AS N. asper 1001 1000000000001
FA N. factor 101 1 1000000101001
RO N. rostratus 11 111100100110001
PE N. pectinicauda 1001 1 100000000001
LO N. longidactylus 1001 1 101000001010
AL N. alutensis 11011100101000000
NU N. numerus 10011100001100001
BR N. brevirostris 10111100001010001
TR N. transitivus 17011100010101001
PU N. pupetta 11011100011101001
CP C. paradoxa 10011110011101011
17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32
CA N. carcerarius 0000000000001000
AS N. asper 0001001 100001001
FA N. factor 1001001000001101
RO N. rostratus 0001 10000001 1 100
PE N. pectinicauda 1001001 100000010
LO N. longidactylus 0011117110011100
AL N. alutensis 0111001100011701
NU N. numerus 1001001101111101
BR N. brevirostris 1001101100110101
TR N. transitivus 1111010100111101
PU N. pupetta 111111110 111110 1
CP C. paradoxa 1111110 11111110 1
the discussion in theabove sections, a set of charac- character state changes in order to explain the pat- ters was chosen and we tried to define their most tern of character state distribution among the taxa.
representative and exactly measurable states. The The most parsimonious cladograms found under
selected their characters are given in Table I and their dis- similar premises were evaluated by comparing tribution over the species in Table II. topologies and the distributionof synapomorphies.
The consistency index is used as a tool for clado-
2.3.2. Numerical cladistic analysis. - The data gram comparison; it is defined by the theoretical
divid- matrix of the N. transitivus group (Table II) with minimumnumber of character state changes the exclusion of autapomorphous (6, 7, 14, 31) and ed by the actual numberof changes. The consisten-
doubtful(la, 15, 20, 29) characters has been ana- cy index is 1 in case cladograms contain no homo-
lyzed with the MIX, PENNY, and CLIQUE pro- plasies. Rooting of cladograms is done by specify- grams of the PHYLIP phylogeny inferencepackage ing the ancestral states for each character. Strict
version (Felsenstein, 1987) 3.01. Cladograms were consensus cladograms are constructed on the basis constructed under Wagner (Eck & Dayhoff, 1966; of PHYLIP CONSENSE. Compatibility analysis
Kluge & Farris, 1969) and Camin-Sokal (Camin & (Le Quesne, 1969) is performed by PHYLIP
Sokal, 1965) parsimony options. Under the Camin- CLIQUE. This method searches for the largest col-
Sokal character criterion, reversals are not permit- of mutually compatible characters ted. The orientated parsimony programs (MIX and (clique(s)).
of PENNY) attempt to minimize the number Compatibility analysis of the data shows a high - Bijdragen tot de Dierkunde, 63 (3) 1993 157
To facilitate discussion, the third subset is
considered the N. transitivus group s. str.
Parsimony analysis of all data leads to a large
number ( + 100) of equally parsimonious cladistic
solutions. The consistency indices of the clado-
grams constructed with the Wagner method is 0.45,
and 0.40 when the Camin-Sokal method is chosen.
The differentsolutions show little agreement in the
monophyletic groups they support.
To arrive at a more consistent picture, the char-
acters not present in any of the ten largest cliques,
thus those with a high probability of not being
homologous, are excluded from analysis. The
cladograms constructed with this reduced data set
are all similar in topology and in the monophyletic
groups they support, whether obtained under
Wagner or Camin-Sokal options. They always have
a consistency index of 0.75. However, there is no
unique dichotomous solution indicating that poly-
chotomous branching patterns exist that could not
be resolved. The consensus of the cladograms
generated under both parsimony options is given in
Fig. 1A.
There in favour of are some arguments giving the
characters and 4, 5, 10, 24, 27, 28, a greater impor-
1. Strict of the N. transitivus tance in the reconstruction. Therefore Fig. consensus cladograms group phylogeny
constructed by numerical cladistic analysis (see text for explana- all numerical analyses performed have been exe-
tion). Species name abbreviations are explained in Table II. cuted with both weighed and unweighed characters. Matching synapomorphies indicated only, uniquely derived However, a weighing procedure has not resulted in characters underlined, character state reversals indicated by
a different class of cladistic solutions. The whole minus-sign. A, cladogram identical for both Wagner and in favour of for certain Camin-Sokal parsimony analysis; B, cladogram inferred by argument a greater weight
Wagner parsimony method after secondary analysis of the N. characters has therefore not been treated in more
but transitivus group s. str.; C, same as B, inferred using the detail.
Camin-Sokal parsimony method.
A of the N. transitivus secondary analysis group
s. str. has been executed with parsimony methods. level of Ten of 8 inconsistency. largest cliques The ancestral character of this states group have characters each found. Five characters are (3, 4, 12, been estimated using outgroups (other species of
19, 30) are in the intersection of all these cliques and the N. transitivus group s.l.) and parsimony, apply- 12 characters (lb, 2, 3, 4, 5, 9, 12, 18, 19, 22, 25, ing rules given by Maddison et al. (1984). The 30) in the joining of them. The characters in inter- Wagner analysis reveals two cladograms (consensus section of the cliques convincingly indicate the tree in Fig. IB) and the Camin-Sokalanalysis a sin-
of the subsets of monophyly following species: gle one (Fig. 1C); the consistency index of all three
all with of N. N. (1) species exception carceraríus, cladograms is 0.89. Noteworthy are reversals in pectinicauda, and N. (2) N. brevirostris and characters and 27 asper; 11, 16, 17, constituting syn-
N. rostratus; and (3) N. longidactylus, N. alutensis, apomorphies for the sister species N. longidactylus
N. N. and Carinurella and N. alutensis. transitivus, pupetta, doxa. para- - 158 B. Sket & J. Notenboom Phylogeny and biogeography of the Niphargus transitivus group
Fig. 2. Apomorphy indices of the species of the N. transitivus group, compared with their positions in the cladogram and their ecological and geographical ranges. Species indications as in
Table II; for explanation of the use of “apomorphy indices”, see text. Habitat: ■ = caves; □ = interstitial; ? = proper N. N. Fig. 3. Distribution area of the transitivus group (for habitat uncertain. Range: whole numbers denote length of The alutensis only an approximate position could be given).
* in km; = species known from onesingle locality only ranges lowest map shows localities of the Dinaric species and the com- in the Dinaric karst. heart of the posed track of the interstitial species. The importantkarst areas
are indicated only onthe topmost map. Species indications as in
Table II; double symbols for TR and PU for different subspecies
(see text); in the lowest map the letters indicate species known 3. Some biogeographic reflections from onesingle locality in the heart ofthe Dinaric karst: B = N.
N. brevirostris, N = N. numerus, R = N. rostratus, F = fac-
Prior to decided to C = N. A = N. a biogeographic analysis we tor, carcerarius, asper. establish the relative position of individual taxa
This is evident that taxa with along the mainevolutionary path of the group. It (Fig. 2) higher apo- is done by counting "apomorphy indices" by di- morphy indices are mostly inhabitantsof interstitial viding numbers of apomorphies by numbers of waters outside the karst regions. As much as they plesiomorphies. One has to emphasize that those approach karst regions, they always inhabit non-
relative karstic habitats their borders the indices have only a significance, only for at (in fluvio- the chosen set of characters and species, charac- karst). Cave inhabitants in the Dinaric heartland terizing mostly the developmental "direction" of exhibit mostly relatively low apomorphy. Inside the the group. The indices have not a single meaning Dinaric karst, the apomorphy indices are increasing for in direction from southeast towards northwest. the phylogenetic relationships among taxa. Bijdragen tot de Dierkunde, 63 (3) - 1993 159
Taxa with indices high apomorphy generally ex- stitial waters, particularly during the Alpino-
hibit larger distribution ranges, but this may be due Dinaric orogenesis, when most rivers had to change
to their interstitial habitat. Although N. alutensis their directions at least temporarily along the
has been mentioned from it is only one locality, thrusting fault zones; this is also the main axis of
that its is much the non-karstic very probable range larger. It may ranges of species. So the species that
extend either along the Olt River or parallel the were previously isolated became in part sympatric.
its Danube, across Transilvanian tributaries, simi- The composed track (see lowest map in Fig. 3),
lar to the situation in and Slovenia. Italy circumscribing ranges of the monophyletic non-
to the distribution of the karstic of this According group (Fig. part group, seems to be congruent
in 3) the territory between the ancient Tethys and with distribution ranges of some other aggregates
it is most thatit Paratethys, probable originated be- within the genus Niphargus as well as with other in-
fore the karst the Pliocene. terstitial development during It Crustacea (e.g. a part of the N. jovanovici
that either the cf. in is, however, possible (1) epigean an- aggregate, Sket, 1972; see Sket, prep., for dis-
cestor dispersed through freshwater connections of cussion).
the northern Aegeis, or (2) it invaded the territory
polytopically, by isolated confluents of the Parate-
thys, also where speciation started, or (3) the group 4. Summary and conclusions
formed was by step by step penetration and specia-
tion from a "centre of origin" in the southeast In the area between northeasternItaly, through the
towards the northwest the sense of the Dinaric to southern 12 nominal (in "progres- Alps Rumania, spe-
sion rule"). Increasing apomorphy indices in north- cies of the N. transitivus group are known. The spe-
western direction at first seems glance to be an argu- cies are morphologically distinct, forming an aber-
ment in favour of the third cluster option. rant within the genus. However, as descrip-
The most plesiomorphic species inhabit karstic tions of several species of the cluster are based on
cave in the southeastern Dinaric waters Alps (the very poor material, and variation within or between
somehow N. brevirostris less plesiomorphic was adjacent populations is lacking, it cannot be ex-
found in interstitial of karst and the cluded that nominal waters a polje some species are intraspecific
related N. numerus might belong to the same unities.
habitat - hesitate see above). Nevertheless, we to WithinNiphargus, the N. transitivusgroup prob-
assume that the group originated in caves. Since ably is most closely related to the N. kochianus
most species are extremely small, and since intersti- complex which includes some species with single N. tial waters existed before the karstic habitats, it is transitivus group characters. The distinctively
that the of have mosaic more probable cave species today type evolution within the whole genus
foundwithin caves a refugium from competition of makes exact delimitation of any group extremely the interstitial more highly evolved relatives. Cave difficult.The N. transitivus group is characterized
niches species are apparently adapted to particular by upright posture, and a step by step adaptation to in comparison with other karstic Niphargus spp. volvation. The accumulated result of this phyletic
Their method of locomotion probably must be un- lineage is Carinurella paradoxa. However, by ex- derstood in this context. clusion of Carinurella (along with some other es-
Species of the N. transitivus group that inhabit tablished genera) from Niphargus the latter is made karst exhibit very small distribution ranges, which a paraphyletic entity.
be due to the of isolation may large degree in small The species of the TV. transitivus group inhabit in- subunits of the of the Dinaric terstitial hydrographie system waters or caves in small areas of the Dinar-
By non-karstic have ic karst and extensive areas to the Alps. contrast, species large north. We pro-
After initial (i.e. long) ranges. an allopatric specia- pose that the initial speciation of an ancestor, arriv- tion in confluentsof the Paratethys, they may have ing from the Paratethys, occurred in its southern had much greater possibilities to disperse by inter- confluents. During the Alpino-Dinaric orogenesis - 160 B. Sket & J. Notenboom Phylogeny and biogeography of the Niphargus transitivus group
(and subsequent karstification) some conservative Benedetti, G., 1942. Prime osservazioni sopra i Niphargus della
Venezia Memorie R. Accad. Sci. Lett. Arti Padova, taxa could have escaped into isolated karstic areas Euganea.
58: 1-15. while others overlapped the areas owing to changed 16. Bou, C., 1971. Recherches sur les eaux souterraines, drainage directions. Parasalentinella rouchi n. g., n. sp., des eaux souterraines des The suffers from in- phylogenetic analysis a large Pyrénées françaises (Amphipoda, Gammaridae). Annls.
in data the results suffi- consistency the and are not Spéléol., 26: 481-494.
Bousfield, E., 1983. An updated phyletic classification and ciently robust to be convincing. Possible explana- paleohistory of the Amphipoda. In: F.R. Schram (éd.), tions for inconsistencies are discussed by Noten- Crustacean phylogeny: 257-277 (A.A. Balkema, Rotterdam boom (1988b: 188-192): (1) There are general & Boston).
when cladistic analysis is applied at low problems Camin, J.H. & R.R. Sokal, 1965. A method for deducing
levels This be taxonomie (Arnold, 1981). may a branching sequences in phylogeny. Evolution, 19: 311-326.
1971. Dancàu, E., Sur unnouvel Amphipodesouterrain de Rou- reflection of a very similar gene pool shared by the
manie, Niphargus alutensis n. sp. Trav. Inst. Spéol. Emile species, so that true parallelisms are likely to occur; Racovitza, 10: 209-215. the simplicity of the available characters, which 1972. Sur la Roumanie du Dancàu, D., présence en genre renders difficult to detect and charac- homoplasies Karamaniella (Amphipoda). Trav. Inst. Spéol. Emile Raco- ter polarity less easily to ascertain; the shortage of vitza, 11: 225-231.
Eek, R.V.&M.O. Dayhoff, 1966. Atlas of protein and available characters, with the consequence that er- sequence
structure (National Biomedical Research Foundation, Silver rors made in the recognition of synapomorphies Springs, Maryland). have a proportionally large influence on the final Felsenstein, J., 1987. PHYLIP (Phylogeny Inference Package) result; and the fact that polymorphism frequently version 3.0 manual. Available from the author, c/o Universi- increases errors made in selection of appropriate ty of Washington, Department of Genetics SK-50, Seattle, characters. (2) Development of cave-dependent Washington98195, USA. Karaman, G.S., 1972. XXXVII contribution to the knowledge structural reductions and troglomorphisms may of the Amphipoda. Niphargus asper, n. sp., Echinogamma- easily evolve independently, and this phenomenon rus foxi (Schell.) and E. stocki G. Kar. (fam. Gammaridae) can in turn result in character incompatibility. from Yugoslavia. Acta Mus. maced. Sci. nat., 13: 1-18.
(3) Underlying peripatric mechanisms of speciation Karaman, G.S., 1975. 73. Contribution to the knowledge of
Amphipoda. Three Niphargus species from Yugoslavia and can result in polychotomous patterns of descent as
Italy, N. ambulator n. sp., N. pupetta (Sket) and N. transiti- evidenced by synapomorphies (Notenboom, 1991). i vus Sket (fam. Gammaridae). Poljoprivreda Sumarstvo, There (4) are no reasons to assume that the most Titograd, 21(4): 13-34. parsimonious cladogram or the largest clique Karaman, G.S., 1981. Description and distribution of Niphar-
the best estimate of relation- Ruffo in and Gam- represent phylogenetic gus longidactylus Yugoslavia Italy (fam. ships. Thus, the inferenceof phylogenetic and bio- maridae) (Contribution to the knowledge of Amphipoda 115). Biosistematika, 7: 39-49. geographic patterns with the aim to elucidate the Karaman, G.S., 1984a. Description of several new Niphargus of the is evolutionary history N. transitivus group species (Gammaridea, Niphargidae) from southern Europe strongly handicapped by this large inconsistency in (Contribution to the knowledge of the Amphipoda 136). Pol- the data. joprivreda i Sumarstvo, Titograd, 30(2-3): 39-64.
Karaman, G.S., 1984b. Revision of the Niphargus orcinus-
group, part I. (fam. Niphargidae) (Contribution to the
knowledge of the Amphipoda 130). Crnogorska Akademija 5. References Nauka i Umjetnosti, Glasnik OdjeljenjaprirodnihNauka, 4:
7-79.
Ancona, U. d\ 1942a. Variabilité, differenziamento di razze Karaman, G.S., 1985. The taxonomy of Niphargus transitivus
locali di nel Memorie 1st. ital. 1971 with remarks N. G. 1985 e spezie genere Niphargus. Sket, to armatus Kar., (fam.
Idrobiol. Dr. Marco Marchi, 1: 145-167. Niphargidae) in Italy (Contribution to the knowledge of the
Ancona, U. d', 1942b. I Niphargus italiani. Tentativo di Amphipoda 149). Poljoprivreda i Sumarstvo, Titograd,
valutazione critica delle minori unità sistematiche. Memorie 31(2-3): 21-35.
R. 1st. Speleol. ital., (Biol.) 4: 1-225. Karaman, G.S., 1989. New species of family Niphargidae and
1981. low taxonomie localities of Arnold, E.N., Estimating phylogenies at new some other subterranean gammaridean spe-
levels. Z. zool. Syst. Evolut.-forsch., 19: 1-35. cies from Yugoslavia (Contribution to the knowledge of the Bijdragen tot de Dierkunde, 63 (3) - 1993 161
Ara- Amphipoda 166). Glasn. rep. Zav. Za5t. Prir. prir. Muz. Tito- Sket, B., 1962. Karamaniella pupetta n. g., n. sp., ein neuer
Zav. N.R. grad, 19 "1986": 15-32. phipode aus Slovenien. Izd. Ribarst. Maked.,
Karaman, G.S. & B. Sket, 1989. Niphargus species (Crustacea: Skopje, 3(2): 27-32.
Amphipoda) of the Kvarner-Velebit Islands (NW Adriatic, Sket, B., 1964. Nova aberantna vrsta postranic (Crust., Am-
Yugoslavia). Biol. Vëstn., 37(2): 19-36. phipoda) iz Slovenije (Eine neueaberrante Art der Flohkrebse
Karaman, G.S. & B. Sket, 1991. Two new Niphargus species aus Slovenija). Biol. Vëstn., 12: 147—152.
(Amphipoda) from caves in Yugoslavia. Int. J. Speleol., 19: Sket, B., 1971. Vier neueaberrante Niphargus-Arten(Amphipo-
51-66. da, Gammaridae) und einige Bemerkungen zur Taxonomie
Kluge, A.G. & J.S.Farris, 1969. Quantitative phyletics and the der Niphargus-áhnlichenGruppen. Razpr. slov. Akad. Znan.
evolution of anurans. Syst. Zool., 18: 1-32. Umet., (Cl. IV) 14(1): 1-25.
Maddison,W.P., M.J. Donoghue&D.R.Maddison, 1984. Out- Sket, B., 1972. Die Niphargus jovanovici-Gruppe (Amphipoda,
group analysis and parsimony. Syst. Zool., 33(1): 83-103. Gammaridae) in Jugoslawien und NO-Italien, taxonomisch,
und betrachtet. slov. Notenboom, J., 1988a. Parapseudoniphargusbaetis, new ge- zoogeographisch phylogenetisch Razpr.
Akad. Znan. nus, new species, a stygobiontamphipod crustacean from the Umet., (Cl. IV) 15(1): 99-141, maps 1-2.
Guadalquivirriver basin (southern Spain), with phylogenetic Sket, B., 1990. Is Niphargobateslefkodemonaki sp. n. (Crusta-
implications. J. crust. Biol., 8: 110-121. cea: Amphipoda)from Kriti (Greece) a zoogeographical enig-
Notenboom, J., 1988b. Phylogenetic relationships and biogeog- ma? Zool. Jb., (Syst.) 117: 1-10.
Distribution of subterranean raphy of the groundwater-dwelling amphipod genus Pseu- Sket, B., in prep. patterns some
doniphargus (Crustacea), with emphasis on the Iberian spe- Crustacea in the territory of the former Yugoslavia.
cies. Bijdr. Dierk., 58(2): 159-204. Sket, B. & F. Velkovrh, 1981. Phreatische fauna in Ljubljansko
Notenboom, J., 1991. Marine regressions and the evolution of polje (Ljubljana-Ebene, Jugoslavien) — ihre ökologische Ver-
groundwater dwelling amphipods (Crustacea). J. Biogeogr., teilung und zoogeographische Beziehungen. Int. J. Speleol.,
18(4): 437-454. 11: 105-121.
Quesne, W.J. le, 1969. A method of selection of characters in
numerical taxonomy. Syst. Zool., 18(2): 201-205.
Ruffo, S., 1937. Studi sui Crostacei Anfipodi VI. Un nuovo
1 Niphargus delle acque soterranee del Veneto. Boll. 1st. Ent. Received: June 1992
Univ. Bologna, 10: 24-30. Revised: 5 April 1993