Page 3 - Scani_et_al1995
P. 3
EVOLUTION OF HYBRID STICK-INSECTS 61
Downloaded by [31.185.101.124] at 08:47 23 March 2016(gamete multiplication, Suomalainen et al, 1987), which hybridogenetic strains. Furthermore, these have an in-
make all loci homozygous at once (Pijnacker, 1969; Scali, variant karyotype, whereas several chromosome repat-
I969). A cline from high to low polymorphism is observ- ternings have been detected in the parthenogens
ed when B. g. tnaretimi, B. g. benazzii and B. g. grandit (Manaresi et al, 1991, 1992a, b, 1993). B. whitei (= B.
are analyzed in that order (Table I). The first subspecies rossius/g. grandii) is the clonal parthenogenetic com-
appears to have kept a native genepool, as also its plex corresponding to the hybridogenetic southeastern
ecology seems to support, whereas the last is the most strains, from which it cannot be distinguished on the
differentiated, both genetically (Fig. 2) and ecologically basis of gene-enzyme systems (Mantovani et al., 1992).
(Mantovani & Scali, 1993b). The thelytokous B. atticus is The trihybrid B. lynceorum deserves a few more words
surprisingly polymorphic (Table I); the finding of very about its origin, which must have occurred in two
many clones (Nascetti & Bullini, 1988; Mantovani & Scali, hybridization steps. On the basis of gene-markers and of
1993a) can be explained in terms of both its origin from the unisexual condition of B. atticus, it has been
multiple interracial crosses and of the peculiar cytology suggested that an atticus x rossius cross occurred first to
of its parthenogenesis, allowing recombination with produce diploid hybrids, which then mated to B. gran-
production of homozygous genotypes from otherwise dii and incorporated a third genome to achive the double
fixed heterozygosities (Fig. 2) (Mantovani & Scali, 1993a; allotriploid structure of B. lynceorum (Mantovani et al.,
Marescalchi et al., 1993; Marescalchi & Scali, 1995a). 1992; Manaresi et al, 1993).
The overall picture of the genetic similarities among HETEROSPECIFIC MATE RECOGNITION
the three above-mentioned species is reported in Figure
2. The dendrogram, based on Nei's D values and ob- The ability to recognize mating partners is thought to
tained from allozyme frequences of 20 gene-enzyme be either a by-product of natural selection directly
systems, clearly shows that B. a. atticus is much closer to favouring isolation mechanisms or, alternatively, its
B. grandii (D = 0.31 - 0.36) than to B. rossius (D = 1.48 primary target. These opposite views appear to charac-
- 1.63) (Table II). Accordingly, B. grandii and B. rossius terize the biological species concept and the recognition
are highly differentiated from each other (D = 1.23 - concept of species, respectively (Paterson, 1980, 1981,
1.50) (Mantovani & Scali, 1993b). 1985; Paterson & Macnamara, 1984; Kimbel & Martin,
1993). Whatever the right theoretical background, it hap-
Repeated hybridization events involving the same three pens in practice that mating behaviour and genetic dif-
taxa led to quite different outcomes. In eastern Sardinia, ferentiation generally go in parallel. Known exceptions
extensive crosses of unisexual B. a. atticus with bisexual are rather in the direction of first developing specific
B. r. rossius originated diploid, apparently fertile, hybrids mating behaviour and later a general genetic differen-
since from their backcrosses to B. r. rossius an in- tiation, than vice versa. In nature, gene-flow between
trogression of atticus alíeles into the rossius gene-pool related groups of organisms may be cut off (and
took place (Mantovani & Scali, 1991). In northwestern reproductive cohesion within each group developed)
Sicily (Fig. IB), hybridization between B. r. redtenbacheri even with a very low genetic differentiation: it com-
and B. g. benazzii produced a hybridogenetic strain, monly happens with the so called eco-ethological
reported as B. rossius-g. benazzii (2n = 35,XX). In speciation (see Bianchi et al, 1994).
southeastern Sicily (Iblean region) (Fig. 1C), at least four
distinct hybridization events occurred, originating two Bacillus taxa appear to have opposite features: in spite
clonal parthenogenetic species and an additional of their sharp or even very high genetic differentiation, as
hybridogenetic strain, namely B. whitei (= rossius/g. judged from gene-enzyme systems (see preceding sec-
grandii; 2n = 35,XX), B. lynceorum ( = atticus/rossius/g. tion), their mate recognition system does not work very
grandii; 3n = 52,XXX) and B. rossius-g. grandii (2n = specifically. Actually, in addition to the regularly oc-
35,XX), respectively (Nascetti & Bullini, 1982, 1988; curring B. rossius-grandii x B. grandii matings that sup-
Bullini et al, 1983; Scali & Mantovani, 1989; Mantovani et port the hybridogenetic system, in the Canicattini Bagni
al, 1991a, 1992; Scali et al, 1991, 1992). area (Siracusa, Sicily) B. grandii males have been found
copulating with B. whitei females. In the same way, B.
Bacillus hybridogens are the first case recorded among rossius males have been seen several times to copulate
invertebrates; they achive a hemiclonal transmission of with B. atticus females at Laghi Alimini (Apulia), where
the maternal rossius genome (R) to the progeny by natural B. atticus/rossius F, hybrids of both sexes could
discarding the grandii haploset (G) and replacing it be found. As a matter of fact, mate acceptance can be so
through real fertilization with sperm provided by syn- loose that, on occasions, B. rossius males do mate in the
topic B. grandii host males (G'), thus renewing each field with females of the distantly related unisexual
generation the F, hybrid structure (RG') of the bacilline Clonopsis gallica 0une 1990, La Giannella,
hybridogens (Mantovani etal, 1991a; Mantovani & Scali, Grosseto, Tuscany). It seems therefore that, in Bacillidae,
1992; Tinti & Scali, 1992a, 1993). genes controlling the mate-recognition system have been
highly conserved while structural genes coding for
Both hybridogens show a «neo-» rossius chromosome
set, while both clonal parthenogenetic species embody a
«paleo-» rossius haploset; therefore the clonal taxa ap-
pear to be of more ancient origin than the related