BULLETIN DE L'INSTITUT ROYAL DES SCIENCES NATURELLES DE BELGIQUE, BIOLOGIE, 71-SUPPL.: I7 -33, 200I BULLETI N YAN HET KONINKLIJK BELGISCH INSTITUUT YOOR NATUURWETENSCHAPPEN, BIOLOGIE, 71-SUPPL.: 17-33, 200 I

Rodent taxonomy at the end of the 20th century

Delectatur musica

by Marco CORTI

Why rodent taxonomy? Rodents are so numerous and di­ Pius, the plague was widespread ail over the Mediterranean verse so that they have been able to colonise most of habitats basin causing the death of millions of people. The Black on the Earth. They are widespread in ail continents and in the Death in the 141"century determined an enmmous number of past they even reached areas of the world which were previ­ deaths reaching two-thirds or three-fourths of entire ously unoccupied. Rodents therefore represent one of the populations in the first epidemic in some parts of Europe. most suitable animal groups for the study of biological Other species of rodents today cause severe damage in devel­ mechanisms promoting biodiversity and for biodiversity oping countries where agricultural practices need to expand management, a primary task today. due to an increase in the human populations and their needs, Furthermore, alterati ons of natural habitats which occurred so that different programmes have started concerned with the management of rodent populations and a new science, called during Neolithic agriculture have determined a close associa­ "rodent pest control", has developed (BUCKLE & SM ITH, tion between many rodents and humans as can be seen in the 1994). extraordinary success of Mus musculus (BROTHWELL, 1981) . Others, such as the rat, have caused plagues that repeatedly How many rodent species have been identified? The total heavily reduced human populations in Europe and Asia in number of mammals has grown from the original 210 found historical times. During 165 BC, at the time of Antoninus in the tenth edition of Systema Naturae (LINNAEUS, 1758), to

600 Q) " 5000 ,. .- .- " "Cro ... .- (..) - - - ... (/) Q) 500 - --- Q) "C - ,, ,, " -- 'ü '- , 4000 Q) Q) c.. c.. If (/) "C 400 / Q) 0 / " 0 .0 I - ·c - 3000 c (..) I - I Q) (/) - - Q) 300 I :.;::::;> "C " " Il ro (/) ... ,. ,, " 11 :J Q) 2000 - I - - E T5 I - :J Q) 200 - / - c.. I ü (/) - ,, ... ~ - 0 - - ,, Il - 1000 -0 ,, 100 - ,. - c ,, - 11 - - - - -::-n·1· 11 n n 0 0 1 1 1 1770 182 0 1870 1920 1970 Fig. J. - T he increase in the number of mammal species from the tenth edition of Systerna Naturae up to 1993. Bars represent the nurnber of species of rnarnmals described for each decade and the li ne represent the cumulative number of species up to J 993 (redrawn from W ILSON & REEDER , 1993). 18 MARCO CORTI

4629 (WILSON & REEDER, 1993), 2021 of which are rodents. LINNAEUS . In the Editio Decima ( 1758), LINNAEUS was able This constitutes . nearly 50% of the entire class, including to determine 36 species in the Glires, although he also in­ monotrems and marsupial$ (Fig. 1). How was it possible to cluded lagomorphs and the rhino in the group. He classified arrive at such a figure? What k.ind of taxonomy has been used the six genera forming the order providing detailed descrip­ to depict this outstanding scenario? This review presents cur­ tions of the fore and front limbs, tail, pelage, ears, dentition, rent developments in rodent taxonomy at the turn of the mil­ ail characters which are still widely accepted in creating a lennium by revising the major scientific events that have reliable classification system, at least at the family Jevel. contributed to cunent researcher attitude. Most of the ad­ LINNAEUS also indicated a series of behavioural characters in vances in rodent taxonomy have occurred during the second his key, with a premonitory sense for a classification that half of the 20'h century, many of which during the last two should also include traits that are not directly measurable decades. This development is still in progress in many labo­ from the morphology of the organism, an attitude that has ratories world-wide and much of this review reflects a per­ been neglected for a long time. For the description of Mus sona) attitude towards the present, near past and future of this musculus he wrote: "M. cauda elongata subnuda, palmis discipline. To fulfil the task, the following topics will be con­ tetradactylis, plantis pentadactylys. Mus cauda nudiuscula, sidered: corpore cinereo-fusco, abdomine subalbescente . ... Habitat in domibus. Unguis pelliciaris palmarum nu/lus, quo a Ratto !. The Rodents' Systema Naturae. How has the knowledge diffe1: Delectatur m.usica 1, 11011 facile incarceratm; on the rodent fauna evolved to the present day, and how polyphagus ". was the cunent classification system developed? The question also concerns the base unit of classification - the After LINNAEUS, the nineteenth century was characterised by species - and the different philosophies of hierarchical an effort to produce catalogues of regional faunas, a series of grouping, from the species level up to the order. works which reached maturity during the second half of the 2. The origin and development of biometry, morphometrics l 9'h century 'and the beginning of the 20'h century. and the new geometric morphometrics. The most comprehensive and modern systematic catalogue 3. The impact of Numerical Taxonomy, as the explicit defi­ for Western Europe (excluding Russia), was published in nition of the different taxonomie philosophies which de­ 1912 by MILLER, one of the major figures amongst the term i ned classification. founders of modern rodent taxonomy. MILLER personally ex­ 4. Rodent cytotaxonomy and speciation. amined 11500 specimens from the British Museum, the United States National Museum, and from private collec­ 5. Biochemical systematics. tions. The catalogue comprises 26 genera and 88 species of Taxonomy is not a trivial question, as different Iines of rodents, making up a total of 139 taxa including subspecies. thought can produce very different results. For example, hi­ After identification keys, MILLER provided tables of indi­ erarchies can be constructed on the basis of similarities, cer­ vidual skull meas urements, such as condylobasal length, tain aspects of the phenotype or by ancestor - descent rela­ zygomatic bread th, interorbital constriction, and others in­ tionships. tended for use in taxon diagnosis. This represents one of the first attempts in providing a reference to study character vari­ Systematics, taxonomy and classification are often used as ati on between and within species. However, the data was set interchangeable tenns, and it was SIMPSON (196 1) who de­ out poorly in a fi at table with no analysis and poor geo­ fined their meaning definitively, in order to avoid confusion. graphic sampling. This Jack has in part been covered by de­ However, they are always connected in our day to day work. velopments during the last 50 years, but much has still to be According to SIMPSON (1961 ), systematics is "the scientific done. study of the kinds and diversity of organisms and of any and al! relationships among them.". Classification is defined as At the turn of the 20'h century a modern attitude for the stud­ "the ordering of organisms i11to groups (or sets) on the basis ies on rodents has been established, Oldfield THOMAS be­ oftheir relationships ". Taxonomy is "the theoretical study of coming one of the prominent fi gures. Among his papers, classification. including ils bases, principles, procedures and there is an article of just six pages published in 1905 in which rules". he presents the precise description and nomenclature for the main Iinear measurements to be collected on the rodent skull The term "relationships" could obviously cause some confu­ for taxonomie purposes (Fi g. 2). The need for such a defini ­ sion because it may imply or not an ancestor - descendant tive description arose from the increasing amoun t of species relation exclusively. Taxonomy must therefore fo llow pre­ description at that lime (Fig. 1) . Many of the comparisons cise rules and aims. were made from skull measurements, such as MILLER' s Since SIMPSON 's time, there has been an evolution and tun­ (1912), and there was an obvious need for an unequivocal ing in classification. It took time to agree on the fact that bio­ description of characters to be measured. logical species exist (according to the Modern Synthesis The outstanding work of MILLER, THOl'vIAS and other schol­ definition) and that taxa must be defined on the basis of their ars brought the ori ginal 36 species of Glires of LI NN AEUS to ancestor - descendant relation (see MAYR & ASHLOK, 199 1, for a general review). 2021. This was the number counted by WILSON & REEDER in Taxonorny concepts changed through time after the modern ori gin of the rodent Systema Natu rae cstabli shed by Carolus 1. lt loves music. Rodent taxonomy at the end of the 20'h century 19

~o.-{~ ------

-- ~~ - -- --·--

J J

j

Fig. 2. - The precise defi nition made by THOMAS ( 1905) for the main measurements of the rodent skull .

their 1993 second edition of Mammal Species of the World five species by CORBET & HILL (199 1) and MUSSER & (Fig. 1) . There has been a 4.2% increase in the number of CARLETON (1993), but with sorne contradictions between the species during the 11 years between the first edition in 1982 two classifications. Further studies on this genus carried out by HONAKI et al. and the second in 1993 by WILSON & during the last ten years (VOLOBOUEV et al., 1988; AFEWORK REEDER. Much of this increase can be explained by a better BEKELE et al., 1993; CORTI et al., 1995; CORTI & FADDA, tuning of taxonomists, i.e. taxa finally accepted as good spe­ 1996; CAPULA et al., 1997; DUCROZ et al., 1997; FADDA & ,ri.es, rather than by new discoveries. CORTI, 1998) have shown Lhal there is even a higher nurnber of species occurring in the complex, and that science has to Nonetheless, revisions through the decades have been char­ wait for the definitive taxonomy of the genus as well as for acterised by alternate attitudes of lumping and splitting. This others. attitude accepted the fact that rodents include groups that are characterised by a particularl y high number of cryptic or sib­ ling species, i.e. distingui shed by a hi gh rate of speciation, Higher level systematics which however are diffi cult to identify. Taxonomy and sys­ tematics therefore need a continuous revision due to continu­ Rodent taxonomy it is not just a matter of species identifica­ ous new fi ndings which suggests that many species are in tion and classification. Higher rank systematics is a current fact species complexes. matter under discussion which has nol yet completely been solved. As an example, one may cite Lhe African Arvican.this LESSON 1842, which Lhrough the years reached the conspicuous The 32 exlanl fami li es of rodents (HA RTENBERGER, 1985) number of 44 species, later red uced by HONA K I el al. ( 1982) show a considerable amount of shared morphological char­ to Lhe single species A. niloticus, and Lhen incremenled Lo acters, incl uding (and above ail) those relati ve to the mastica- 20 MARCO CORTI ------. 1 1 Other caviomorphs 111111111 1 caviomorphs Erethizontidae mll ITT"11 l lrrnl Il

i---~~tfï~~r~lë:j~Ei------~ï ------1 1

•1 1 Hystricog nath i : Thryonomyidae 111111111 phiomorphs Petromuridae ~111~111~111

Hystricidae 111111111 ~------Ctenodactyloidea 111111111 Pedetoidea 111111111

Anomaluroidea 111111111 ,------, Gliroidea Aplodontoidea 5#1 sciuromorphs Sciuroidea 111""""""""...... I Sciurognathi Castoroidea 138811 •------,­ Geomyoidea Muroidea =-~=-~=-~:; myomorphs

Dipodoidea 111111111 •------

"L . protrogomorph hystricomorph myomorph sciuromorph ...... ~ mm ...... Ill

Fi g. 3. - The division of Rodents into the two suborders Hystricognathi and Sci urognathi, and the furt her subd ivision of the fam ili es in to the four groups of Caviomorphs, Ph iomorphs, Sciuromorphs and Myomorp h s~ Boxes represent th e four zygomasseteric co nditi ons (s hown at the bottom) . Note Lhat Anomaluridae , Ctenodactyli dae and Pedetidae have an intermecliate condition of both sciurognath y and hys tri comorph y. Rodent taxonomy at the end of the 201h century 21 tory apparatus, which strongly support a monophyletic hy­ phylogenetic reconstruction through the identification of pothesis for the order. primitive and derived states in characters. They also, without doubt, introduced the scientific world to the ideas of The original classification (BRANDT, 1855) recognised three phylogenetic systematics or cladistics which, after their major suborders, i.e. Sciuromorpha, Myomorpha, and original fomiulation by HENNIG (1966), had remained ig­ Hystricomorpha, ptevalently;on the basis of the insertion of nored for a few years before being rediscovered and gener­ I massetere muscles in relation to the infraorbital foramen, the ' ally accepted (see HULL, 1990, for an historical discussion). '' rostrum and the zygomatic bar (Fig. 3). The generally ac­ cepted classification by TULLBERG ( 1899) recognises two Much of the development in rodent taxonomy we will see major suborders: Sciurognathi and Hystricognathi, following hereafter has its roots in SOKAL & SNEATH 's original work. the insertion of massetere muscles (Fig. 3).

However, there is a high degree of convergence and parallel­ Morphometrics and multivariate morphometrics ism between families of these suborders, suggesting that they may not represent natural groups (HARTENBERGER , 1985; Metric (continuous and qualitative) characters have, for con­ JAEJER, 1988). From a palaeontological viewpoint, siderable time, been the only characters used to study within­ sciurognathy in the lower jaw, and the protrogomorphous and between-species variation since MILLER 's (1912) and condition relative to the zygomatic bar is a primitive condi­ colleagues' times. Morphometrics and multivariate tion (JAEJER, 1988). Nonetheless, it is present in the morphometrics (REYMENT et al., 1984; MARCUS , 1991) is Bathyergidae and Aplodontidae, although perhaps this is a the science which deals with a kind of data which investi­ secondary condition. Families in the Hystricognathi show gates variation in size and shape. Theriologists (and rodent both hystricomorphous and hystricognathous features; and students) have traditionally showed a major interest in the Sciurognathi are protogomorphous, sciuromorphous, multivariate morphometrics, as can be seen by ail the work­ hystricomorphous and myomorphous, although ail shops, books and primers developed ad hoc during mamma­ sciurognathous (Fig. 3) . This resulted in several different lian meetings (e.g. NEFF & MARCUS , 1980; MARCUS & suborder classifications (MILLER & GIDLEY, 1918; CORTI , 1989; CORTI , 1992) and by the considerable number EL_LERMAN , 1940; SIMPSON , 1945; WOOD, 1955, 1965, of articles published in Mammalian journals. 1985; CHALINE & MEIN, 1979). However there has only been a limited number of papers It was therefore to be expected that paraphily of the Sciuro­ which have explicitly concentrated on the central concept of gnathi (HARTENBERG, 1985) has been investigated. There is allometry, i.e. on how variation in size and shape, the relative a series of other cliaracters, which do not produce a consist­ proportions, varies across organisms. ent scenario of natural groups. Among the others, the same incisor enamel microstructure is shared by Hystricognathi Allometry was defined formally by HUXLEY in 1932, and al­ and the families Pedetoidea and Ctenodactyloidea but not by though his book has been quoted on thousands of occasions Anomaluroidea and Dipodoidea (MARTIN, 1993). The over the years, it has probably only been read from cover to growth of cytogenetics followed by DNA/DNA hybridisa­ cover only by few peopl e. A more precise definition of al­ ti on and DNA sequencing has brought further insight and lometry has been made recently by GOULD ( 1966) and clarification in this complex matter (see below). KLINGENBERG (1996). Allometry can be considered either with respect to development, i.e. comparing individuals at different ages (growth allometry), or to variation within the The impact of Numerical Taxonomy same age class ( static allomeuy), or with respect to evolu­ tion, i.e. comparing the same age classes of related species It should be mentioned that the attitude to rodent taxonomy ( evolutioncuy allometty ) (HERSH, 1941). and systematics during the course of last 30 years has To build up a rodent classification scheme, we need to com­ changed - like for students of many other animal and plant pare allometric patterns between taxa. This is what morpho­ groups - considerably si nce the development of Numerical metrics is intended for. Taxonomy. The idea of treating taxa numerically (Opera­ ti onal Taxonomie Units, OTUs, in the original fonnul ati on) One of the first papers was published in 1952 by VON was presented by two scientists, Philip SNEATH and Robert BERTALANFFY & PIROZINSKY. Although limited to a single SOfAL, in a paper published in Nature in 1962; and followed species (albino rats), they were able to show how body pro­ later by two books: Principles of Numerical Taxonomy, in portions change during growth. They produced a series of 1963, and Numerical Taxonomy. The principles and practice plots between body weight and organ weight and compared of numerical classifïcation, in 1973. them with known data for other m ammals and showed how ontogenetic allometry does not COJTespond with interspecific Many of the ordination and c lu st~ ring techniques to identify allometry. homogeneous groups or patterns of variation that we still use today derive or have their roots in those ori gi nal texts. It is However, many studies on morphometri c and crani ometric peculiar, however, that those ideas are generally regarded as variation have been hampered by scanty data sets (this is fre­ "phonetics", where the taxonomie rules are the general and quent when collecting in the field), Jack of computing instru­ un weighted simil arity between OTUs. There is a good part ment , and by a consequent inadequacy, in the partitioning of reported by SNEATH & SOKAL that defined rigorous bases fo r nongeographic ( ex and growth) and geographic effects in 22 MARCO CORTI the data. Consequently this has limited most analyses (either approach went through an initial experimental phase during univariate or multivariate) to simple discrimination pur­ the mid-nineties before being perfected (BOOKSTEIN, 1998). poses. Even when computers became progressively avail­ Geometric morphometrics has received a close attention able, they were only used to produce discriminant fonction frorn rodent students as it represents an almost definitive so­ scores to classify specimens, The attitude (out of discrimi­ lution to the study of size and shape variation across organ­ nant fonctions) towards disçrimination was exemplified in isms. Severa! aspects have been attempted, from ontogeny the pioneer study of Michael FESTING in 1973. In laboratory (ZELDITCH et al., 1992; MONTEIRO et al., 1999; HINGST­ ' strains of mice, he was able to distinguish between strains ZAHER et al. , 2000), sexual dimorphism in size and shape 1' using mandible measurements. Further insight into non-ge­ (HOOD, 2000), variation within (AUFFRAY et al., 1996; netic effects (ontogeny, sexual dimorphism) on CORTI et al., 1996; CORTI & FADDA , 1996; FADDA & CORTI , (multivariate) shape and size changes came from a series of 1998) and between species (CORTI et al. , 1998, 2000; FADDA papers by LEAMY (1975, 1977). & CORTI, 2000). In natural population studies, however, the major aim is the The study on three species of the African genus Lophuromys identification of patterns and causes of species differences (CORTI et al. , 2000) is a good example. These species are (history and !or ecology) or, in within-species studies, proc­ representative of the two subgenera Lophuromys s.s. (L. esses of divergence (clines or step clines) with the eventual flavopunctatus and L. sikapusi) and Kivumys (represented by occuITence of races or subspecies (ENDIER, 1986). This is L. woosnami) and have been studied by means of a three di­ particularly important in comparisons of closely related spe­ mensional approach, which allowed the separate study of cies, because morphometric characters are influenced both size and the reconstruction and visualisation of shape differ­ by historical (i.e. phylogenetic) and adaptive effects ences in the skull. They found not only significant differ­ (THORPE, 1983). The assessment of morphometric differ­ ences in size among species, but also between age classes ences due to cladogenesis can be seriously hampered by (Fig. 4). Furthermore, size is sexually dimorphic in L. homoplasies, i.e. adaptation to similar changes in current flavopunctatus and L. woosnami. The comparisons of trajec­ conditions or sampling from similar areas. Moreover, non­ tories of size increase during growth for males and females geographic effects, such as allometric growth, can alter the also indicated a common and parallel pattern retained after assessment of taxa similarities; this has generally been ne­ cladogenesis by the species, females at age class 0 are glected in rodent multivariate morphometrics, although a smaller in size than males, but at age class 1 they become general multivariate solution to the problem was produced larger. From age class 2 up to age class 4 males become by B URNABY (1966) and ROHLF & BOOKSTEIN (1987). larger. Discrimination attitudes, leading to the identification of ho­ A principal component analysis of the shape components al­ mogeneous taxa (which is classification but in statistics only) lowed the distinction of the two subgenera (Fig. 5). There is a has always been the main goal for many rodent high statistical significance between both the subgenera and morphometrics practitioners. Even if it has generally been among the species. Changes in shape are visualised as three possible to assign specimens to their correct group, many re­ dimensional reconstructions with rendering, to provide an sults are only indications for a proper use in taxonomy be­ aid to the human eye for further interpretation. This repre­ cause they are limited to small areas of the range and there­ sents one of the main features of geometric morphometrics fore cannot be interpretable in a more general context of geo­ compared to traditional morphometrics where one had to im­ graphic variation. agine such changes simply by looking at a list of scores. There is a clear distinction in the shape of the rostrum, which GENOWAYS & JONES (1971) carried out one of the first is shorter and larger in Kivumys and longer and thinner in multivariate morphometrics studies where a clear distinction Lophuromys s.s. between non-geographic effects (sex and age) on linear measurements was made in which an adequate sample was Authors interpreted size and shape differences taking into used. They studied the systematics of the kangaroo rats of the account both phylogeny (i.e. the separation into the two Dipodomys phillipsii group in Central America. Their factor subgenera) and a diet specialisation. This was because the analysis showed that ail the populations examined should be genu s is characterised by a strong tendency to in sectivority considered as a sin gle species, with a clinal geographic vari­ with Kivu.m.ys being m ore in sectivorous than Lophuromys ation, with .Dipodom.ys omatus being a geographic race of S.S. Dipodomys phillipsii and not a separate species as previously coqsidered. The establishment of Rodent cytotaxonomy: a chromosomal Systema Naturae? Geometric morphometrics BUSH et al. ( 1977) and CAPANNA & CORTI ( l 991) have high­ Traditional morphometrics (MARCUS , 199 1) is based on lin­ li ghted the strict relati onship between variation in chromo­ ear measurern ents, whereas this new approach looks at the somal diploid number and the number of species in the major geornetric properties of the form, by recovering them marnmalian orders (Fig. 6). The observations concern dip­ through two- or tlu·ee-dimensional Cartesian co-ordin ates loid numbers only, so that only a part of the karyotypic vari­ (ROHLF & M ARCUS , 1993; M ARCUS et al., 1996). This new ability is described, i. e. not those resulting from inversions, Rodent taxonomy at the end of the 2Q1h century 23

6.0

5.5 L. sikapusi '' Q) . . . . . N en L. f/avopunctatus "'C . . L. woosnami 0 5.0 ...... ~ c Q) () 4.5 ..

4.0 0 1 2 3 4 age class

Fig. 4 - Comparisons of mean growth in size among the species fro m age class 0 to age class 4, in the two sexes. The solid lines are fo r males and the dotted lines for fe males. deletions, etc. which is revealed by banding patterns only. This quote is taken from a chapter of the book "Cytotax­ However these observations reveal a di rect relati onship be­ onomy and vertebrate evolution" edited by CHIARELLI & tween the structural reordering of the genome and the prolific CAPANNA in 1973. The ti tle includes the term "cytotax­ speciation occurring in rodents. onomy" rather than "cytology'', and clearl y represents the Robert MATTHEY in 1949 provided the firs t review on the general attitude during the earl y seventies, when a certain karyotypes of Vertebrates, and reported the high number of amoun t of literature on chromosomal polymorphism was al­ studies on rodent cytogeneti cs, which had been investigated ready available. It became clear that many species groups are up to then. He also hypothesised on the use of these data in a recognisable only on the basis of their karyotypes and that phylogenetic context. However when superimposed on the chromosomal rearrangement is not just a new taxonomie phylogeny of redents by WINGE (1924), the variability character to examine but a possible cause of the high shown both in diploid and fundamental numbers was so high bi odiversity and speciation characterising rodents. Conse­ and overlapping between the families that no conclusion quently fo llowed a suggestion that chromosomes needed to could be drawn. be primarily investigated in every qualified study on rodent species di versity. However, the extensive work carried out in the fo llowing years on the pigmy mice Leggada and on other rodents by BUS H et al. ( i 977) estimated that there are 0.43 1 karyotypic MATTHEY himself (1964) and by many other scholars, changes per lineage per million years, and chromosomal showed the occurrence of ex tensive karyotype speciation became one of the hottest areas in evolutionary polymorphisms within genera and species (see CHIARELLI & biology. Examples from Spalax (NEVO, 1991 ), Mus CAPANNA, 1973, fo r a review). (CAPANNA, 1982), Rattus (YONG, 1969; RAMAN & SHARMA, In hi s chromosome fo rmul ae of Eutherian Mammals, 1977; YOS HIDA, 1980), Thomomys (PATTON, 1972), sug­ MATTHEY (1 973) introduced the Leggada as "in this group, gested that karyotype rearrangements occur and establish the knowledge of the kmJ·otype daes notfit with the views of themselves in natural populations and favour speciation by the Taxonomists. There are tao many original km)'otypes decreasing relative fe rtil ity in structural heterozygote hy­ which cannot be referred to a described sp. or ssp. and that I brids. The book of Michael WH ITE "Modes of speciation" quot e as f.c . (from French, '.fo rmulae chromosomique '). The ( 1978) champio!'led these ideas and provided exhaustive poly11101phism. rests more frequently Oil Robertsonian /HOC­ samples on redents, one of the most karyotypically variable esses (muscul oides) sometimes on. pericentric in versions order of mammals. Moreover, WHITES ' mode! strongly op­ (oubanguii) or multiple translocations (gundae)." posed the classical all opatric mode! of speciati on by MAY R 24 MARCO CORTI

PC 1 - PC 1 +

Kivumys Lophuromys s.s.

D

6. 6. D 6. D 0.04 D 6. D D D D 6. & Qi &-116. DO it::.. 6. D ~~OO ~~ QJ ~D D "'U 0.00 6.~ ~6. () 6. D N ~6. ~ 6.6. ~ o~~ OQ)BO 6. 0 o~ o L. flavopunctatus -0.04 CO o o o L. sikapusi 0 0 ~ t::.. L. woosnami

1 1 1

-0.08 -0.04 0.00 0.04

PC 1

Fig. 5. - Three dimensional stud y of the shape differences between the two subgenera Lophuro 111 ys s.s. and Ki v11111rs. A principal component scatter pl ot is shown at the botto m. The shape changes in th e upper part are represented as a lateral and dorsal three­ dimensional surface reconstru cti on of th e sk ull relative to th e hi ghest (PC l +, Loplwro111 ys s.s.) and lowest (PC l - , Kiv11111ys) va lues of th e first vec tor. Shape changes have been intenti onall y exaggerated by a factor of 4. . Rodent taxonomy at the end of the 201h century 25

0 Rodentia OO • '1 0 ·Primates >. 0"'" :!: c.o ..a Carnivora• CO 0 ï::::::: LO. e Insectivora CO • Chiroptera > 0 -q- •• Q) Artiodactyla C> c 0 CO ("") L... c 0 Lagomorpha C\I C\I • 0 Pinnipedia T""" •e Cetacea 200 400 600 800 1000 1200 1400 1600

Number of species

Fig. 6. - The relationship between diploid number vari ability and the number of species included in some of the major mamma!ian orders (redrawn from C APANNA & CORTI, 1991) .

~ •~ ,,., • ij \ J .( ~ 2.17 1.18 6.7 4.11 3.13 5.15 8.14 10.12 9.16 ,_ f ( r. ! ,4 V 1· ( >\ ~ ~ -..!!'\ ~ 1.7 3.8 4.15 19.11 '5.17 2.18 6.13 12.14 9.16 Fig. 7. - Left side: G-banded chromosomes of the Apenni ne races of the house mouse C B (upper row) and CD (lower row). Numbers represent the acrocentric chromosomes fused to give ri se lO a melacentric. Note lhat the onl y homology in fu sion fou nd in the two races is between acrocentrics 9 and 16. Right side: meiosis of a structural helerozygote hybrid between the CB and CD races. T he long ring is fo rmed by 8 metacenrric showing monobranchial homology; the three bivalents are chromosomes 19, 9. 16 and the heterochromosomes. 26 MARCO CORTI

( 1963) (the "stasipatric" mode of WHITE, 1978). and reticulate evolution in the house mouse by CORTI et al. ( 1986) (Fig. 8). The case of the chromosomal races of the One of the most cited examples by WHITE concerns the chro­ house mouse not only documents how fast speciation can mosomal races or species in status nascendi of the bouse occur in rodents, but also provides a taxonomie tool which mouse, which have been extensively studied by CAPANNA on can be used in the study of the phy logeny of these races. The co-workers (GROPP et al., 1970; CAPANNA, 1982; CORTI et G-banding patterns were used to identify homologies in al., 1986). chromosomal rearrangements (see Fig. 7 for an example). Figure 7 shows a diakynesis of a male hybrid between two of From that, a phylogenetic hypothesis was then built (Fig. 8). the many chromosomal races occurring in Italy as well as There are some homoplasies in the tree that could only be their G-banded karyotype (the Apennine races CB x CD). explained by reticulate evolution (in the form of dashed lines During the meiosis of these structural heterozygote hybrids a in Fig. 8). long multivalent ring is formed, which does not segregate. Meiosis is blocked and the resulting hybrid is sterile. There is At this point the synthesis between systematics and tax­ a post-mating and sometimes a pre-mating reproductive bar­ onomy seemed complete: good, reliable taxonomie charac­ rier acting between some of these races (CAPANNA et al., ters, i.e. chromosomal reaITangements identified on the basis 1984). Chromosomal speciation has been very fast in the of diploid and ann numbers and different banding tech­ house mouse. In fact it did not start in these areas until 7- niques, to classify species and to build up reliable schernes of 9000 years ago (CAPANNA, 1982). speciation and phylogeny. A debate arose concerning chromosomal vs. allopatric Nonetheless, there is strong evidence suggesting that when speciation models (see WHITE, 1978, and M AYR 199 1) in comparing chromosomal reaiTangements above the genus promoting biodiversity. From this taxonomie and systematic level there are a lot of hornoplasies which cannot be detected use of karyotype rearrangements was discussed. During the by usual banding, mainly due to subsequent repattering of the eighties a good set of qualified studies documented how karyotype (see for example WIENBERG & STANYON, 1987). chromosomal rearrangements can be used as a reliable taxo­ Major criticisms were made on some atternpts to reconstruct nomie character in rodent taxonomy. These include a study Rodent phylogenies through karyotype rearrangernents (see on Nearctic Arvicolidae by MODI (1987) and on speciation VIEGAS-PÉQU!GNOT et al. , 1986).

Fig. 8. - Phylogenetic tree of the Alpine chromosomal races of the house rnouse (from CORTI, 199 1) . The occurrence of chro mosoma\ fu sion is indicated by two numbers separated by a dot whi ch correspond to the two acrocentrics which are fu sed into a metacentric (foll owing the Standard Nomenclature for the House Mause Karyotype). Homoplas ies (repeated occurrence of the same fu sion into independent branches) are expl ained through hybridisation events (indicated by dashed lines). There were only 1wo homoplasies in the tree which are enclosed in circ les (metacentrics 6.7 and 7. 18). Rodent taxonomy at the end of the 2Q•h century 27

1 1

.... ••••••• ••• · .... •••• :• M. m. museul us • • • • .... ·•.: M. spicilegus •• •• •. M2 •. ••. .•• M. m. domesticus • . •• . . • •• •• M1 ••• . • ••• .• •••• •• •••• •

M. m. domesticus M. spretus ...... ·········· ·••··········· ...... ········ \

Fig. 9. - Map whi ch shows the fi ve species and subspecies of the genus Mu s occurring in Europe and the Mediterranean basin and their corresponding biochemi cal group according to B ONHOMME el al. (1978). Dotted li nes roughl y correspond to the li mit of the ranges. The grey li ne indicates the hybridisation zone between M. 11111scu/us do111 es1ic11s and M. 111usculus musculus.

Protein electrophoresis cies may have been maintained by a combination of environ­ mental factors and post-mating isolating mechanisms. Ge­ Many of the new insights into rodent taxonomy came from netics was fin ally able to provide a stronger biological basis the applicati on of mul ti locus protein electrophoresis, which to the previously described taxa on the basis of morphologi­ became widely di ffused in the mid sixties (see AYALA, 1975, cal and ecological characters (SCHWARTZ & SCHWARTZ, 1976). The technique all owed the estimati on of gene flow 1943; URS IN, 1952; ZIMMERMANN, 1949). 'between populations and an estimate of the time of genetic Species boundari es could be then identified tlu·ough classical di vergence, assuming neutral evolution of structural genes population genetics as allozyme variation is inherited in a over time. For the first time there was a tool to investigate simple Mendelian fas hion. Furthermore processes of within species systematics and taxonomy which was structurall y a species differentiation (i .e. geographic vanatton and central part of the modern Sy nthesis, i.e. the bi ological spe­ raciation) could be highlighted and hierarchies based on rela­ cies concept (MAYR, 1963). · tive genetic dissimil arity/divergence established.

One of the first and complete studies came from SELANDER By the late sixties and the early seventies there was an expo­ et al. (1969) and HU NT & SELANDER (1973) on the two sub­ nenti al growth of studies in rodent taxonomy and systematics species Mus musculus musculus and Mus musculus using protein electrophoresis. domesticus, which have been hybridising in Central Europe Examples include the studies by JOHNSON & SELANDER and in the Jutland peninsul a for at least 5 000 years. They ( 197 l) on the kangaroo rats Dipodomys, by NEVO & SHAW were able to show the occurrence of a hybrid zone and of an ( 1972) on Spalax ehrenbergi, and by BONHOMM E et al. almost unidirec ti onal gene introgression from c/0111.esticus ( 1978, 1983) on the European and Mediterranean specic · and into musculus, suggesting that the separation of the subspe- subspecies of the genus Mus, which led to the identi fication 28 MARCO CORTI of at least five biochemical groups, corresponding to the taxa the first attempts to salve the complex problem of rodent Mus musculus domesticus (M 1), Mus musculus musculus (M family phylogeny was the summary paper by CATZEFLIS et 2), Mus spretus (M 3), Mus macedonicus (M 4A) and Mus al. ( 1992) through the application of scnDNA hybridisation. spicilegus (M 4B) (Fig. 9). They found that the rate of DNA evolution is faster in rodents than in any other order of mammals and birds, and used In 1975 AYALA examined data from allozymic studies on DNA-DNA hybridisation distances to produce an overall seven genera of rodents and found that the average genetic evolutionary tree of the muroid rodents together with a time similarities between species vary from group to group and scale. Their phylogeny seriously contradicted others derived for some cases they are similar to those found between sub­ from karyotype rearrangements (e.g. VIEGAS-PEQUINOT et species. Nonetheless, much of the earlier work in this field al. , 1986), and suggested that the relationship between nucle­ proved that electrophoretical determination produce classifi­ otide substitution and time is probably curvilinear and not cations which are in many respects similar to those based on linear, i.e. speciation events dated by palaeontologists are morphological characters. However, it also gave insights into older than those inferred through a linear molecular clock relationships which were not apparent in earlier classifica­ hypothesis. tions. We could quote the case of 15 species of Peromyscus (AVISE , 1974 ), where the allozyme relationships "promoted" Severa) general studies followed the pioneer work by some subspecies to species, and moved some species from CATZEFLIS and co-workers, using the increasing amount of one subgenus into another. data made available. These led to many attempts, of variable quality, to salve the complicate malter of rodent families sys­ After establishing the occurrence of gene flow, one should tematics. A good body of literature attempted to question the question how reliable allozymes are in the study of species real monophyly of rodents (GRAUR et al., 1991 ), contradict­ relationships. The strong controversy between the selectio­ ing the hypothesis of the uniquely shared characters of the nist versus the neutralist hypothesis on structural gene evolu­ masticatory apparatus of the group. Many fall into the "is the tion went on for over ten years (KIMURA, 1968; KIMURA & guinea-pig a rodent?" category (GRAUR et al. , 1991 ; LI et al. , MARUYAMA , 1971 ; NEVO et al., 1974; ÜHTA & KIMURA , 1992; MA et al., 1993; D' ERCHIA et al. , 1996). This consti­ 1975; STEBBINS , 1982). tutes a clear example of how inappropriate data analysis - Historically, the applicalion and growth of the technique of even if the data, i.e. mtDNA sequences, are complete and protein electrophoresis has been an extremely powerful tool correctly aligned - may cause erroneous interpretations. in solving complex problems at the genus level. However an Although the story involved an apparent waste of time spent adequate number of individuals from populations must be in arguing by supporters of rodent monophyly or polyphyly analysed in classical Mendelian population genetic frame­ (together with an increase in Impact Factor), it provides how­ work. It soon became evident however that the technique ever a clear example of how easily one can violate rigorous could not be of great systematic value beyond genus level. assumptions, i.e. ail sites are equally variable and evolve at the same rate. This however is not the case of Mammalian mtDNA (YANG et al., 1994; SULLIVAN et al., 1995). Are­ Nuclear and mitochondrial DNA examination of the data by omitting the third-codon position (SULLIVAN & SWOFFORD, 1997) through a maximum-likeli­ By the end of the seventies, the use of digestion hood mode! clearly established the guinea-pig as a rodent so endonucleases had opened up a new area in rodent taxonomy rejecting the polyphyletic origin of the order. and systematics (PARKER & WATSON , 1977; AVISE et al., A recent analysis of the 12S RNA (rRNA) by NEDB AL et al. 1979), closely fo ll owed by DNA-DNA hybridisation (HUNT ( 1996) bas fin ally cast some light on the complex et al. , 1981 ) and DNA and RNA sequencing using the PCR phylogenetic relati onships between sciurognathous families technique (GUYER & KOSHLAND, 1989; ARNHEIM et al. , and Hystricognathi (Fig. 10). They found that the monophyly 1990). It became possible to include a higher number of geo­ ofHystricognathi is supported by ail analyses (i.e. parsimony graphic localities or species represented by fewer specimens and maximum-likelihood, neighbour joining), with the in the analy sis, as comparisons at least in a phylogenetic or guinea pig failing within the Caviidae family. Furthermore, phylogeographic context cou Id be made not on gene frequen­ the analysis also supports a sister-group relationship between cies but bel ween cleavage patterns, distances or sequences, Sciuroidea and Aplodontoidea, and for Muroidea and requiring a small er amount of individuals. Geomyoidea (with definitive evidence for monophyly for There was a dramatic appearance in the scientific literature these latter two families). The analysis however on the of papers dealing with the phylogenetic interpretation of ro­ polytomous relationships among Muroidea, Dipoidea, dent species below and beyond the genus level, and the work Geomyoidea, and Pedetoidea, and the unclear placement of by YONEKAWA et al. (1981) on cleavage patterns ofEuropean Gliroidea and Castoroidea led to some ambiguous results. and Asiàn species of Mus constitutes one of the first case An important aspect of Rodents compared to other mamma­ studies. Others, later attempted to study evolutionary histo­ li an orders is their high heterogeneity between lineages, re­ ries of groups (e.g. the South American Akodontinae, SMITH sulting in different branch lengths, with a hi gh rate of among­ & PATTO N, 1993). site variation. This can seriously affect analyses as it may be Molecular techniques also became an obvious tool which the result of either differential rates of molecular evolution or could chall enge hi gher level systematics in Rodents. One of ancient divergence, the so called "branch length e ffect" Rodent taxonomy at the end of the 2Q'h century 29

;• • • • • ·• • • • • • • • • • • • • • • • • • Outgroup • 111111111 • Caviomorpha 111111111111111111111111111111111111111111 l } Hystricognathi • 111111111 Phiomorpha • • ___,~ Sciuroidea • • - Aplodontoidea • - Ctenodactyloidea • - .. - - ....··~ • - ~-:,··~.. , .. , ...... -:.-..-... .-.i:.-... .-.i:..-... .-... .-.-..-... .-...... -... ,... GI'1ro1 'd ea • - .~.~-~-~· ~·~- ~-~-~-~-~-~-~-~-~-~-~·~ • - Castoroidea • Sciurognathi • - ~-.-=... .;-.-.-.-:.-..-:... -.-=..-:... -.-..-:.·• .-.-...... -... .-...... -:... .-.-.. Muro1'dea •• llC ·~-~-~-~-~-~-~-~-~-~-~-~.~-~- ~-~ -~-~ ~-:.::··~ - - 111111111111111111111111111111111111 Di podoidea r- Geomyoidea

111111111111111111111 Pedetoidea

11111111111 hystricomorphous myomorphous sciuromorphous protrogomorphous

Fig. 1O . - The phylogeny of rodent families based on the analysis of the sequence of the l 2S RNA (rRNA), after NEDBAL et al. ( J996) . Branch patterns follow the zygomasseteric condition (shown below the tree).

(FELSENSTEIN, 1978; SWOFFORD et al., 1996; but see, for a the southern part of our globe, where there is the highest de­ discussion on the guinea pig, SULLIVAN & SWOFFORD, mand for rodent taxonorny. There is an obvious practical 1997). problem in species identification in the fi eld, as rnany rats and rnice represent real pests for agriculture and are vectors Perspectives of parasites and micro-organisrns which affect livestock and humans. Biodiversity however is also severely disturbed in Rodent taxonomy and systematics at the turn of the century is that part of the world where the deepest environrnental crisis a well established and mature science. Different disciplines since the ori gin of Neolithic agriculture is taking place. Hu­ including morphology, morphometrics, cytogeneti cs, genet­ rn an popul ation growth and the consequent expansion of ag­ ics, molecular gènetics are extensively widespread in many riculture and of other economic acti vities determine the laboratories worldwide and integrated more and more often shrinking and alteration of natural habitats at a faster rate. with each other to produce a consistent scenario of rodent Rodents represent therefore a key group to be used as a tool diversity. for eventual conservation plans of biodiversily.

Many of the recent multidisciplinary approaches adopted by scientists corne from the studies that have carried out for over LITERATURE ClTED a century on the rodent faunas of the northern hemisphere. AFEWORK BEKELE, CAPANNA, E., CORTI, M., MARCUS, L.F. , & These have provided insights into the richness of species and SCHLITTER, D.A., 1993. Systematics and geographic variation of on man y of the biological and geographic factors which have Ethi opian Arvicanthis (Roclentia, Muriclac). Jaumal of Zoologr determined such a hi gh diver ity. Much still has to be done in London, 230: 11 7- 134. ' · ' 30 MARCO CORTI

ARNHEIM, N., WHITE, T. & RAYNEY, W.E. , 1990. Application of the (Rodentia: Murinae). Biochemical Systematics and Ecology, 25: polymerase chain reaction to organismal and population biology. 403-409. Bioscience, 40: 174-182. CATZEFUS, F.M., AGUILAR, J.-P. & JAEJER, J.-J., 1992. Muroid ro­ AUFFRAY, J. C., AUBERT, P. , LAT!EULE, c. & DOD, B., 1996. Rel a­ dents: phylogeny and evolution. Trends in Ecology & Evolution, 8: tive warp a·nalysis of skull shape across the hybrid zone of the hou se 122-126. mouse (Mus musculus) in Denmark. Journal of Zoology, London, CHAUNE, J. & MEIN, P., 1979. Les rongeurs et /'évolution. Dain, 240:441-455. Paris. A VISE, 1.C ., 1974. Systematic value of electrophoretic data. System­ CH!ARELU, A.B. & CAPANNA, E., 1973. Cytotaxonomy and verte­ atic Zoology, 23: 465-481. brate evolution. Academic Press, N.Y. AVISE, J.C., GIBUN-DAV!DSON, C. , LAERM , J., PATTON, J.C. & LANSMAN, R.A., 1979. Mitochondrial DNA clones and matriarchal CORBET, G.B. & HILL, J.E., 1991. A world list of mammalian spe­ phylogeny within and among geographical populations of the cies. 3rc1 edition. NHM Publications, Oxford Uni versity Press. pocket gopher, Geomys pine1is. Proceedings of the National Accul­ CORTI, M., 1991 . Chromosomal speciation and reticulate evolution: emy of Sciences of 1he U.S.A., 76: 6695-6698. Testing phylogenetic hypotheses with compatibility and parsimony. AYALA, F.J., 1975. Genetie differentiation during the speciation Bol/e11ù10 di Zoologia, 58: 307-319. process. Evolutionary Biology, 8: 1-78. CORTI, M., 1992. Data analysis in systematics: a workshop and a AYALA, F.J., 1976. Molecular evolution. Sinauer Associates !ne. manual to introduce geometric 111 017Jho 111 etrics. Primera Congreso MA. Latinoamericano de Teriologia. BERTALANFFY VON, L. & PIROZINSKY, W.J., 1952. Ontogenetic and CORTI, M., ESTABROOK, G.F. & CAPANNA, E., 1986. Micro­ evolutionary allometry. Evolwion, 6: 387-392. evolutionary sequences in house mouse chromosomal speciation. Systematic Zoology, 35: 163-175. BONHOMME, F., BRITTON-DAV!DIAN, J., THALER, L. & TRIANTAPHYLUDIS, C., 1978. Sur l 'existence en Europe de quatre CORTI , M., CIVITELU, M.V., CASTIGUA, R., AFEWORK BEKELE & groupes de souris (genre Mus L.) du rang espèce et semi-espèce par CAPANNA, E., 1994. Cytogenetics of the genus A111icamhis (Roden­ la génétique biochimique. Comptes rendus des séances de tia, Muridae). 2. The chromosomes of three species from Ethiopia: /'Académie des Sciences, Paris, 287, Sé1: D: 63 1-633. A. abyssi11ic11s, A. dembeensis and A. blicki. Zeitschrift fiir Sii11getierku11de, 61: 339-351. BONHOMME , F. , CATALAN , J., GERASIMOV, S. , ORSINI, P. & THALER, L., 1983. Le complexe d'espèces du genre Mus en Europe CORTI, M., FADDA, C., SIMSON, S. & NEVO, E., 1996. Size and Centrale et Orientale. 1. Génétique. Zeitschrift fiir Saugetierkunde, shape variation in the mandible of the fossori al rodent Spalax 48: 78-85. ehrenbergi. A Procru stes analysis of three dimensions. Pp. 303-320 BOOKSTE IN , F. L. , 1998. A hundred years of morphometrics. Acta in: Advances in 111orpho111 etrics, MARCUS, L.F. , CORTI, M., LOY, A., Zoologica Academiae Scientiarum Hungaricae, 44:7-59. NAYLOR, G. & SUCE D. (eds.), N.A.T.0. AS! Series. Plenum Press Co., New York. BRANDT, J.F., 1855. Beitrage zur nahern Kenntniss der Saugetiere Russland. Mémoires de l'Académie royale de St-Pe1ersbo11rg. Se i: CORTI, M. & FADDA , C. , 1996. Systematics of Arvicanthis (Roden­ 69: 1-375. ti a, Muridae) fro m the Horn of Africa: a geometric morphornetri cs evaluation. ltalian Journal of Zoology, 63: 185- 192. BROTHWELL, D., 1981. Th e Pleistocene and Holocene Archaeology of the hou se mouse and related species. Pp. 1-13 in: Th e bio/ogy of CORTI, M. , AGUILERA, M. & CAPANNA, E., 1998. Phylogeny and the house mouse, BERRY, R.J. (ed.). Symposia of the Zoological size and shape changes in the skull of the South American rodent Society of London. Academic Press. Proechùnys. Acta Zoologica Academiae Scientiarwn Hungaricae, 44: 139-150. BUCKLE, A.P. & SMITH, R.H ., 1994. Rodenl pests and their contro /. CAB International. Wallingford. CORT!, M., Dl GIUUOMARIA, c. & VERHEYEN, W. , 2000. Three­ dimensional geometric morphometrics of the African genus BURNABY, T.P., 1966. Growth-invariant discriminant fun ctions and l ophuromys (Rodentia Muridae). Hystrix, ltalian Journal of generali zed distances. Biometrics, 22: 96- 11 O. Mammalogy, Il: 145- 154. BUSH, G. L., CASE, S.M. WILSON, C.A. & PATTON , J.L., 1977. COURANT, F. , DAV ID, B., LAU RI N, B. & CHAU NE, J. , 1997. Quanti­ Rap id speciati on and chromosomal evolution in mammals. Pro­ ceedings of the National Academy of Sciences of 1he U.S.A. , 74: fication of cranial convergences in arvicolids (Rodentia). Biologica/ 3942-3946. Journal of the l innean Society, 62: 505-5 17. CAPANNA, E., 1982. Robertsonian nwnerical variation in ani111al D' ERCHIA, A. M., GRISS I, C. , PESOLE, G., SACCONE, C. & speciation: Mus musculus, an emb/e111atic mode/. Pp. 155-1 77 in: ARNASON, U., 1996. The guinea-pig is not a rodent. Nature, 381: Mechan isms of speciation, BARICOZZI, C. (ed.), Alan Li ss, N.Y. 597-600. CAPANNA, E., CORTI M., MAINARDI D., PARM IGIANI S. & BRAIN DUCROZ, J. F., L. GRANJION, P., CHEV RET, J.M., DUPLANTIER P.F. , 1984. Karyotype and intermale aggression in wild bouse mice: LOMBARD, M. & YOLOBOUEV, Y., 1997. Characterization of tw~ ecology and speciati on. Behavior Generics 14: 195-208. distinct species of Arvicanthis (Rodenti a, Muridae) in West Africa: cytogenetic, molecul ar and reproductive evidence. Journal of Zool­ CAPANNA, E. & CORTI , M., 199 1. Chromosomes, speciation and ogy, London, 24 1. 709-723. phylogeny in Ma111 111als. Pp. 355-370, in Symposium on the evo /11.­ tion of terrestrial vertebrates. Ed. by GH IARA, G. , Coll ana U.Z. I., ELLERMAN, J.R., 1940. The fcunilies and genera of living rodel!ts, Selected symposia and monographs. Mucchi , Modena. Vols. 1- 1 l. British Museum (N .H.), London. CAPULA, M., CIV ITELU, M.V, CORTI, M., AFEWORK BEKELE & END IER, J. A., 1986. Natura/ selection in the ivild. Monograph in CAPANNA, E. , 1997. Geneti e divergence in the genus Arvicanth is popul ation biology 2 1. Princeton University Press. Rodent taxonomy at the end of the 20'h century 31

FADDA, C. & CORTI, M., 1998. Geographic variation of Arvicalllhis JAEJER, J.J. , 1988. Rodent phylogeny: new data and old problems. (Rodentia, Muridae) in the Nile Valley. Zeitschrift fiir In: Th e phylogeny and classification of the Tetrapods, vol. 2, Mam­ Saugetierkunde, 63: 104-113. mals. BENTON, M.J. (ed.), pp. 177-1 99, Clarendon Press, N.Y. FADDA, C. & CORTI, M., 2000. Three dimensional geometric JOH NSON, W.E. & SELANDER, R.K. , 197 1. Protein variation and morphometric study of the Ethiopian Myomys - Stenocephalemys systematics in the kangaroo rats (genus Dipodomys). Syste111atic Zo­ Il complex (Murinae, Rodentia). Hystrix, /taliall Journal of ology, 20: 377-405. Mammalogy, 11 : 131-143. KIMURA, M. , 1968. Evolutionary rate at the molecular level. 1. FELSENSTEIN, J., 1978. Cases in which parsimony and compatibility Nature, 217: 624-626. methods will be positively misleading. Systematic Zoology, 27: KIMURA , M. & MARUYAMA, T., 1971. Pattern of neutral polymor­ 401- 4 10. phism in a geographically structured population. Genetica/ Re­ FESTING, M., 1973. Mause strain identification. Nature, 238: 351- search, Cambridge, 18: 125-131. 352. KLI NGENBERG, C.P. 1996. Multivariate allometry. Pp. 23-49 ln: GENOWAYS, H.H. & JONES, J.K. jr., 1971 . Systematics of southern MARCUS, L.F., CORTI, M., LOY, A. , NAYLOR, G. & Slice, D. (eds.), banner-tailed kangaroo rats of the Dipodomys phillipsii group. Adva11ces in 111 011Jho111etrics. Plenum Press, N.Y. Journal of Mammalogy, 52: 265-287. LEAMY, L., 1975. Component analysis of osteometric traits in ran­ GOULD, S.J., 1966. All ometry and size in ontogeny and phylogeny. dom-bred house mice. Systematic Zoo/og); 24: 176-190. Biologica/ Reviews, 41: 587-640. LEAMY, L. , 1977. Genetie and environmental correlations of GRAUR, D. , HIDE, W.A. & LI, W.-H., 1991. ls the guinea pig a ro­ morphometric traits in random-bred house mice. Evolution, 31: dent? Nature, 35 1: 649-652. 357-369.

GROPP, A. TETTENBORN, Y. & VON LEHMAN, E. , 1970. LI, W. -H., HIDE, W.A., ZHARKIKH , A. , MA, 0.-P. & GRAUR, D. , Chromosomenvariationen vom Robertson 'schen Typus bei der 1992. The molecular taxonomy and evoluti on of the guinea pig. Tabakmaus M. poschiavi1111s, und ihren hybriden mit der Journal of Heredity, 83 : 174-181. Laboratoriumsmaus. Cytogenetics, 9: 9-23. LINNAEUS, C. , 1758. Systema Nat11rae. Regnum Animale. Editio GUYER, R.L. & KOSHLAND, D.E. , 1989. The molecule of the year. decima, Lipsia. Science, 246: 1541 - 1546. MA, D.-P., ZHARKIKH, A. , GRAUR, D., V ANDEBERG , J.L. & LI, W.­ HARTENBERGER, J.-L., 1985. The order Rodentia. Major questions H., 1993. Structure and evolution of opossum, guinea pig, and por­ on thei r evolutionary ori gin , relationshi ps, and suprafami li al sys­ cupine cytochrome b genes. Journal of Molec11/ar Evo/111io11, 36: tematics. ln : Evo /11tionary relationships among rodents: a 327-334. multidiscip/inary ana/ysis. LUCKETT, W.P. & HARTENBERG ER, J.-L. (eds.), pp 1-33, Plenum Press, N.Y. MARCUS, L.F. , 1991. Tradi tional Morphometrics. In: Proceedings of the Michigan 111011Jl10metrics W orkshop. Ed. by ROHLF, F.J. & HENNIG, W., 1966. Phylogenetic syste111atics. Un iversity of Illi nois BOOKSTEIN, F.L. University of Michigan, Museum of Zoology: Press, Urbana. Special Publication 2, 95-130. HERSH, A.H ., 1941. Al/0111etric gro1vth: Th e ontogenetic and MARCUS, L.F. & CORTI, M., 1989. Data a11a /ysis in syste111atics. phylogenetic significance of differential rates ofg ro1vth. 3'd Growth A 111a11ual for workshop W3. Fifth International Theriological Symposium: 11 3- 145. Congress, Rome.

HINGST-ZA HER, E., MARCUS, L.F. & CERQUEIRA, R., 2000. Appli­ MARCUS, L. F. , CORTI , M., LOY, A., NAYLOR, G.J.P. & SUCE, D. E. cation of geometric morphometrics to the study of postnatal size 1996. Advances in morphometrics. Plenum, New York. and shape changes in the sku ll of Ca /omys expu/sus. Hystrix, /talian MARTIN, T. , 1993. Early rodent incisor enamel evolution: Journal of Mammalog); 11 : 99-11 3. phylogenetic implications. Journal of Ma111 111alia11 Evolutio11, I: HONAKI , J.H., KINMAN, K.E. & KOEPPL, J.W. , 1982. Mammal spe­ 227-254. cies of the world. A taxo11 0111ic and geographic reference. Allen MATTHEY, R. , 1949. Les chromosomes des Vertébrés. Rouge, Press and Association of Systemati c Collections. Lawrence, K.A. Lausanne. Hooo, C.S., 2000. Geometri c morphometric approaches to the MATTH EY, R. , 1964. Evolution chromosomique et spéciati on chez study of sexual size dimorphism in mammals. Hystrix, ftalian Jour­ les Mus du sous-genre Leggada Gray 1837. Experientia, 20: 657- nal of Ma111111alogy, 11 : 77-90. 7 12. HULL, D.L., 1990. Science as a Process. An Evolutionary Account MATTHEY, R. , 1973. The chromosome fom111/ae in Eutheria11 11 w 111 - of the Social and Conceptual Development of Science. Chi cago mals. Pp. 35 1-6 16 ln : Cytotaxonomy and vertebrate evo/11tio11 University Press, USA. CH IARELLI , A.8 . & CA PANNA, E. (eds.), Academic Press, N.Y. ' HUNT, W.G. & SELANDER" R.K., 1973. Biochemi cal genetics of MAYR, E., 1963. Animal species and evo/111io11 . Harvard University hybridisati on in European bouse mi ce. 1-/eredity, 31: 11 -33. Press, Harvard.

HUNT, J. A., HALL, T.J. & BRITTEN, R.J., 198 1. Evolu tionary dis­ MAYR, E. & ASHLOK, PD., 1991. Princip/es of S."stematic zoo/ogy. tances in Hawaii an Drosophila measured by DNA reassociation. 211ct editi.on. McGraw-Hill. . Journal of Molecular Evolution, 17: 361-367. MILLER, G.S., 19 12. Catalogue of the Mammals o.f Western Europe HUX LEY, J.S., 1932. Problems of relati ve growth. Methuen and Co., (Europe exclusive of Russia) in the col/ectio11 of 1h e British Ltd, London. Museum. Trustees of the Brit ish Museum (N.H.), London. 32 MARCO CORTI

MILLER, G.S. & GIDLEY, J.Y., 191 8. Synopsis of supergeneric SMITH, M.F. & PATTON, J.L., 1993. The di versification of South groups of rodents. Journal of the Washington Acade111y of Sciences, American murid rodents: evidence fro m mitochondrial DNA se­ 8: 431-448. quence data for the akodontine tribe. Biological Journal of the Linnean Sociel)', 50: 149-177. MODI, W.S., 1987. Phylogenetic analysis of chromosomal banding pattern aniong the Nearctic Arvicolidae (Mammalia: Rodentia). SNEATH , PH.A. & SOKAL, R.R. , 1962. Numerical taxonomy. Systematic Zoology, 36: 109-136. Na111re, 193: 855-860. MONTEIRO, L.R., LESSA L. G. & ABE, A.S., 1999. Ontogenetic vari­ SNEATH, PH.A. & SOKAL, R.R., 1973. Nwnerical Taxonomy. 'The ation in skull shape of Thricho111ys apereoides (Rodentia: principles and practice of nu111erical classifications. W.H . Freeman Echimydae). Journal of Mammalog)', 80: 102-111. and Co., San Francisco. MUSSER, G.G. & CARIETON, M.D., 1993. Family Muridae. Pp. 576- SOKAL, R.R. & SNEATH, PH.A., 1963. Principles ofnumerical tax­ 578 in: Mammal Species ofthe World. A taxonomie and geographic onomy. W.H. Freeman and Co. , San Francisco. reference. WILSON, D.E. & REEDER, D.M. (eds.). Smithsoni an STEBBINS , G.L., 1982. Modal 1hemes: a new fra111 e1vork for Institution Press, Washington D.C. evolwionary syn1heses. Pp. 1-1 4 in: Perspec1ives on evolution, NEDBAL, M.A. , HONEYCUTT, R.L. & SCHLITTER, D.A., 1996. MILKMAN, R. (ed.), Sinauer Associates lnc., MA. Higher-level systematics of rodents (Mammalia, Rodentia): evi­ SULLIVAN, J., HOLSINGER, K.E. & SllvlON, C., 1995. Among-site dence from the mitochondrial I 2S rRNA gene. Journal of Ma111111a­ rate variation and phylogenetic analysis of l 2S rRNA data in lian Evolution, 3: 201-237. sigmodontine rodents. Molec11lar Biology and Evolution, 12: 988- NEFF, N. & MARCUS, L. F., 1980. A survey of multivariate methods 1001 . for systematics. New York. SULLIVAN, J. & SWOFFORD, D.L., 1997. Are guinea pigs rodents? NEVO, E., 199 1. Evolutionary theory and processes of active The importance of adequate models in molecular phylogenies. speciation and adaptive radiation in subterranean mole rats, Spalax Journal of Ma111111alia11 Evol111ion, 4: 77-86. ehrenbergi superspecies, in Israel. Evo l111ionmy Biology, 25: 1-125. SWOFFORD, D.L., ÜISEN , G.P , WAD DEL, P.J . & HILLIS, D.D., 1996. NEVO, E. & SHAW, C.R., 1972. Genetie variation in a subterranean Phylogenelic lnference. Pp. 407-514 in: Molec11lar syslemalics, 2"ù mammal, Spalax ehrenbergi. Bioche111ical Genetics, 7: 235-341. edition, HI LLIS, D.M. , MORITZ, C. & MABLE, B.K. (eds.), Sinauer, MA. NEVO, E., KIM , Y.J. , SHAW, C.R. & THAELER, C.S.jr., 1974. Genetie vari ati on, selection and speciation in Th o111omys talpoides, pocket THOMAS, O. , 1905. Suggesti on for the nomenclature of the cranial gophers. Evolution, 28: 1-23. length measurements and of the cheek-teeth of Mammals. Proceed­ ings of the Biological Society Washington, 18: 11 -16. ÜHTA, T. & KIMURA, M., 1975. Theoretical analysis of electro­ phoretically detectable polymorphisms: models of very sli ght del­ THORPE, R.S., 1983. A review of the numerical methods fo r recog­ eterious mutations. American Naturalist, 109: 137-145. nizing and analysing racial differentiation. Pp. 404-423 in: N11111 eri­ cal Taxonomy, FELSENSTEIN , J. (ed.), Springer Verlag. PARKER, R.C. & WATSON, R.M., 1977. Restriction endonuclease cleavage maps of rat and mouse mitochondrial DNAS. N11cleic TULLBERG, T. , 1899. Ueber das system der Nagetiere: eine phylo­ Acids Research, 4: 129 1-1299. geneti sche studie. Nova Acta Reg. Soc. Sei. Uppsala, Ser 3, 18: 1-5 14. PATTON, J.L. , 1972. Patterns of geographic variation in the karyo­ type of the pocket gopher, Thomomys bo11ae (EYDOUX and URS IN, E., 1952. Occurrence of voles, mice and rats (Muridae) in GERVOIS) . Evolution, 26: 574-586. Denmark, with a special note in a zone of intergradation between two subspecies of the bouse mouse (Mus 11wsc1.ilus L.). RAMAN, R. & SHARMA, T. , 1977. Karyotype evolution and speciation in genus Rai/us, FISHER. Journal ofSci entific and lndus­ Videnskabelige Meddelelserfra den dansk 11at11rhistorisk Forening, 11 4: 217-244. trial Research, 36: 385-532. REYMENT, R.A., BLACKHITH, R.E. & CAMPBELL, N. A. , 1984. YIEGAS-PÉQUIGNOT, E., PETIT, 0 ., BENAZZOU, T., PRUD' HOMME, M. , LOMBARD, M., HOFFSCHIR, F. , DESCAILLEAUX, J. & Multivariate m01pho111 etrics. 211d edition. Academi c Press. DUTRILLAUX, B., 1986. Phylogéni e chromosomique chez les ROHLF, F. J. & MARCUS, L.F., 1993. A revolution in morphometrics. Sciuridae, Gerbillidae et Muridae, et étude d'espèces appartenant à Trends in Ecology & Evolution, 8: 129- 132. d'autres fa milies de Rongeurs. Mammalia, 50: 164-202. . ROHLF, F. J. & BOOKSTE IN, F.L. , 1987. A comment on shearing as a YOLOBOU EV, Y. , VIEGAS- PÉQUIGNOT, E. , LOMBARD, M. , PETTER, method for "size correction". Sys1e111atic Zoology, 36:356-367. F. , DUPLANTI ER, J.M. & DUTRILLA UX , B., 1988. Chromosomal evi­ "!- . dence for a polyty pic structure of Arvicawhis nilo 1icus (Rodentia, SCHWARTZ, E. & SCHWARTZ, H.K., 1943. The wild and commensal Muridae). Zeitschrift fiir Zoologische Systematik und Evol111io11s­ stocks of the house mouse Mus 11111sculus LINNAEUS . Journal of forscl111n g, 26: 276-285. Ma111111alog)', 24: 59-72. WI ENBERG, 1. & STANYON , R., 1987. Fluorescent heterochromatin SELANDER, R.K. , HUNT, W.G. & YANG, S.J ., 1969. Protein poly­ staining in primate chromosomes. Human Evol111ion, 2: 445-457. morphism and genetic )1eterozygosity in two European subspecies of the house mou se (Mus musculus). Evolution, 23: 379-390. WH ITE, M.J. D. , 1978. Modes of speciation. WH. Freeman and Co. SIMPSON , G.G. , 1945. The principles of classification and a classifi­ WILSON , D.E. & REEDER, D.A.M. , 1993. Ma111111al species of. the cation of mammals. Bulletin of the American Museum of Natural World. A taxonomie and geographic reference. 211ct edition. HistOI)\ 85 : 1-350. Smithsonian Institution Press, Washington and London. SIMPSON, G.G. , 196 1. Principles of animal 1axonomy. Columbia WI NGE, H., 1924. Rodentia. ln: Pattedyr-Slaegter, 2, Haherups, H. Uni versity Press, N. Y. Forlag, Copenhagen. Rodent taxonomy at the end of the 2Q1h century 33

WOOD, A.E., 1955. A revised classification of the rodents. Journal YONG, H.S., 1969. Karyotypes of Malayan rats (Rodentia, Muridae, of Mammalog); 36: 165-187. genus Rat/us, FISHER). Chromosoma, 27: 245-267. WOOD, A.E., 1965 . Grades and clades among rodents. Evolution, YOSHID A, T.H. , 1980. Cytogenetics of black rat: kaJ )'otype evolu­ 19: 115-130. tion and species differentiation. University of Tokyo Press, Tokyo. ', WOOD, A.E., 1985. The relationships, origin, and dispersal of ZELDITCH , M.L., 800KSTEIN, F.L. & LUNDRIGAN , 8.L., 1992. On­ the hystricognath rodents. In: EvolutionaJ)' relationships among togeny of integrated skull growth in the cotton rat Sigmodon rodents: a multidisciplinary analysis. LUCKETT, W.P. & fulviventer. Evolution, 46: 1164-1180. HARTENBERGER, J.-L. (eds.), pp 475-513, Plenum Press, N.Y. ZIMMERMAN, K., 1949. Zur Kenntnis der mitteleuropaischen YANG , Z. , GOLDMAN, N. & FRIDAY, A.E., 1994. Comparison of Hausmause. Zoologische Jahrbücher: Abteilung für Systematik , models for nucleotide substitution used in max imum-likelihood 78: 301-322. ph ylogenetic estimation. Molecular Biology and Evolution, 11 : 316-324. YONEKAWA, H., MORIWAKI, K., GOTOH , 0. , HAYASHI , J.-1., Marco CORTI WATANABE, J. , MIYAS HITA , N., PETRAS , M.L. & TAGAS HIRA , Y. , Dipartimento di BiologiaAnimale e dell' Uomo 1981. Evolutionary relationships among five sub species of Mus Università di Roma 'La Sapienza' musculus based on restriction enzyme cleavage patterns of mito­ Via Borelli 50, 00161 Roma, Italia chondrial DNA. Genetics, 98: 801-816. E-mail : [email protected]