1998/04/01 by Niklas Janz, Sören Nylin · 1 citation
Agricultural and Biological Sciences · Biochemistry, Genetics and Molecular Biology · #Plant Parasitism and Resistance #Lepidoptera: Biology and Taxonomy #Plant and animal studies
paper · doi:10.2307/2411084
openalex publication_date 1998/04/01 · openalex created_date 2025/10/10 · openalex updated_date 2026/06/16
-A database on host p lant records f rorn 437 ingroup taxa has been used to test a number of hypotheses on the interact ion between but ter f l ies and their host p lants using phylogenet ic methods (s i rnple character opt inr izat ion. concentrated changes test , and independent contrasts test) . The but ter f ly phylogeny was assembled f rom var ious s()urces irnd host p lant c lades were ident i f ied according to Chase et a l . 's rbcJ--based phylogeny. The ancestra l host p lant appears to be associated wi th in a h ighly der ived rosid c lade, inc luding the fami ly Fabaceae. As fossi l data suggest that th is c lade is o lder than the but ter f l ies, they must have colonized al ready divers i l ied plants. Previous studies also suggest that the pat terns of associat ion in most insect-p lant inter i rct ions are more shaped by host shi f ts , through colonizat ion and specia l izat ion. than by cospeciat ion. Consequent ly, we have focused expl ic i t ly on the mechanisms behind host shi l is . Our resul ts conf i rm, in the l ight of new phylogenet ic evidence, the pat tern reported by Ehr l ich and Raven that re lated but ter f l ies feed on re lated plants. We show that host shi f ts have general ly been rnore comrnon between c losely related plants than between nore distant ly re lated plants. This f inding. together wi th the possib i l i ty ofa h ighertendency of recoloniz ing ancestra l hosts, helps to expla in the apparent large-scale conservat ion in the pat terns of associat ion between insects and their host p lants, pat terns which at the same t in le are more f lex ib le on a more detai led level . Plant growth form '* 'as an even more conservat ive aspect of the interact ion between but ter f l ies and their host p lants than plant phylogeny. However, th is is largely expla ined by a higher probabi l i ty of colonizat ions and host shi f ts whi le 1 'eeding on t rees than on other growth fonns. Ke words.-Coevolut ion, host shi f ts , insect-host p lant interact ions, Lepidoptera. Papi l ionoidea. specia l izat ion. Received March 22, 1996. Accepted November 2.6, 1991 Few systems have played such an important role in our understanding of how species interactions evolve as butterf l ies and their host plants. To a large extent this is the result o f a s ing le in f luent ia l paper by Ehr l i ch and Raven (196.1) . Their essay inspired a f lood of paperi on dif l 'erent aspects of this associat ion, and a number of related hypotheses on the evolut ion of insect-plant interactions have emerged. However, there have been few attempts to exploit the large database on butterf ly-host plant aff i l iat ions to test such hypotheses using phylogenetic methods (Mitter and Brooks 1983; Mi l le r 1987a) . A major reason fo r th is i s tha t we l l supported phylogenies. fbr both butterf l ies and seed plants, have been unava i lab le . However , th is i s s lowly chang ing , and today i t is possible to put together reasonably robust phylogen ies fo r bo th g roups . Chase e t a l . (1993) have recent ly published a molecular phylogeny for al l seed plants, which is probably the best est imate of large-scale angiosperm phylogeny to date. Butterf ly phylogenies are also emerging and we have constructed a plausible phylogeny across the butte r f l ies by combin ing these pub l ished es t imates . Ehrl ich and Raven ( I 964) argued that the patterns of host plant associat ion that we observe today have been shaped by a s tepwise coevo lu t ionary p rocess in wh ich p lan ts evo lve defenses aga ins t na tura l enemies , and these enemies in tu rn evo lve new capac i t ies to cope w i th these de fenses . P lan ts tha t escape f iom herb ivores can d ivers i fy in the absence o l 'enemies . Insec ts tha t eventua l l y manage to co lon ize one o l ' these p lan ts w i l l en ter a new adapt ive zone and can in tu rn d ivers i fy on to the re la t i ves o l ' th is p lan t , because they w i l l be chemica l l y s im i la r . Ehr l i ch and Raven argued tha t these processes have led to the main pa t te rn they had observed, namely tha t related butterf l ies tend to f 'eed on related groups of plants. Most o r a l l p lan t d ivers i f i ca t i r )n up to the leve l o f reso lu t ion used in our analysis had probribly already taken place at the time the butterf l ies started to diversify. The oldest known butterf ly fossi l dates back to 48 M.Y.B.P and the diversif i cat ion of the butterf ly famil ies probably took place at the end of the Cre taceous, about 66 M.Y.B.P (Emmel e t a l . 1992) . A t leas t some fami l ies even in the most recent ly der ived o f the p lan t c lades used in th is ana lys is da te to th is t ime, such as , Ur t i caceae (a member o f Ros id I in Chase e t a l . , 1993) : 90 M.Y.B.P. , Rutaceae (Ros id 2 ) : 52 M.Y.B.P. , Ap iaceae (Aste r id 2 ) : 52 M.Y.B.P. , Apocynaceae (As ter id l ) : 60 M.Y.B.P. (dates f iom Eriksson and Bremer 1992). Therefore, the clades themselves must be even older. I t is reasonable to regard the evo lu t ion o fcur ren t assoc ia t ions as ar is ing genera l l y th rough butterf ly colonization of already-diversif ied hosts, and that is the approach we shal l take. This is not to say that coevolut ion is an unimportant process in the interaction between butterf l ies and their host plants. only that evidence fbr i t should be sought at other levels of resolut ion. There are two fundamental ly dif ferent approaches to comparative analyses using phylogenetic data. One approach seeks to f ind and explain general ecological or evolut ionary cor re la t ions (Fe lsens te in 1985; Gra fen 1989; Harvey and Page l 1991: Page l 1992) , wh i le the o ther seeks to recons t ruc t and exp la in par t i cu la r h is to r ica l events o r sequences o f events a long branches in a phy logeny (Mi t te r and Brooks 1983: Codd ing ton l98U; S i l l6n-Tu l lberg 1988; Madd ison 1990, Brooks and Mc lennan l99 l ) . These approaches are complernentary (Coddington 1994; Nyl in and Wedell 199,1; Pagel 1994) and we have in the present paper used bo th , depend ing