Science - USA (2022-06-03)

other sexual species, and only 3% were fixed
for one allele in the parthenogens and for a
different allele in both sexual species. More-
over, of the 238 SNP loci fixed for an allele in

the parthenogens, most (2:1 ratio) skewed to

W. whitei. Such patterns likely reflect gene

conversion events. If we assume that 238 of

the (620 + 238) formerly heterozygous SNPs

have converted at a rate of 10−^3 , then this

gives an age of ~230,000 years (this species is

annual), which is consistent with a previous

age estimate of ~300,000 years based on mito-

chondrial DNA for bothW. virgoandW. ngadju

( 14 ) (see also fig. S11). Gene conversion rates two

orders of magnitude slower have been estimated

fromDrosophilainversion data ( 19 ). Thus, we

consider our age estimate to be conservative

and the species is probably older.

To test whether a hybrid origin of the

parthenogens has imposed fitness costs or

benefits,wecompared14lifehistoryand

physiology traits linked to fitness, including

heat and cold tolerance, desiccation resistance,

metabolic rates, thermal sensitivity, reproduc-

tive output, maturation times, and longevity

(Fig. 4 and tables S4 to S16). Of the physiolog-

ical traits, only critical thermal maximum (Fig.

4A) varied between the sexual progenitors, and

W. virgowas intermediate, consistent with an

additive genetic model.W. virgoalso had a

marginally lower water loss rate than its

progenitors but there was considerable overlap

(Fig. 4C). For the life history traits, lifetime

reproductive output per female and egg devel-

opment time differed betweenW. whiteiand

W. flavolineata,withW. virgoin both cases

matching the phenotype ofW. whitei(Fig. 4, K

and L). We also found thatW. virgohad a

smaller clutch size and a larger egg mass,

hatchling mass, and clutch mass compared

with the sexual species, although again with

substantial overlap (Fig. 4, G to J). A lower

clutch size inW. virgocompared withW. whitei

has previously been noted ( 20 ), but fecundity is

clinal inW. whiteiin association with a body

size cline [fig. S12 and see also ( 20 )]. In fact,

W. virgohad a slightly higher clutch size

than the two southernmost populations of

W. whitei(Fig. 4I, red dots).

Parthenogenetic species are often referred

to as“fugitive,”with distributions that escape

interactions with sexual taxa ( 21 , 22 ). The

geographic distribution ofW. virgorelative

to its sexual progenitors provides a natural

experiment to test whether this species has a

competitive advantage over the sexual spe-

cies by comparing its geographic range limit

with and without sexual taxa present (Fig. 1).

We constructed species distribution models

forW. virgo(including separate models for

eastern and western populations) and its sex-

ual progenitors (Fig. 1 and figs. S13 and S14),

confirming suitable habitat forW. whiteiand

W. flavolineatain the east despite their ab-

sence there. Model projection based only on

easternW. virgopopulations correctly infers

the distribution of the western populations,

implying that sexual populations are not lim-

itingW. virgo.

In summary, our data show thatW. virgo

comprises a single clone having persisted

through major environmental changes. Its very

Kearneyet al., Science 376 , 1110–1114 (2022) 3 June 2022 2of4

115 120 125 130 135 140 145 150

−38

−36

−34

−32

−30

−28

−26

longitude

latitude

0 200 400 600 km

W. whitei

W. flavolineata

W. virgo

W. ngadju

Nullarbor Plain

Flinders Ranges

Fig. 1. Distribution of the parthenogenetic grasshopperW. virgorelative to its sexual progenitors
W. whiteiandW. flavolineata. Background gray shading indicates elevation. Green shading shows the
projection of a species distribution model (SDM) forW. virgo. Yellow and blue shading show the projections
of SDMs for the sexual species to the east of their present range. Purple shading shows a projection of
a SDM for the eastern populations ofW. virgoto the western part of its range. Not considered in the
present study isW. ngadju, another parthenogenetic lineage that also evolved through hybridization between
W. whiteiandW. flavolineata.

flavolineata virgo whitei

0

10000

20000

30000

Ita

Ivanhoe Malcolm

ReidySwmp1

Netley

Laverton

BrokenHill WentworthBullabllng

Nowingi

Monia Gap ReidySwmp2 Arrow Lake MountMagnt

Malcolm Kalgoorlie

DeptSprngs

Menzies Hazel Vale ShuttletonSpargovill MenziesPtrDnnyLkt

MountLeonr DiortKngMn

Leinster StirlingVl

Site

Eigenvalue distance

Comparison Species

W. flavollineata

W. whitei

W. virgo

Fig. 2. Eigenvalue distances (first 16 principal components) of SNP variation between allWarramaba
spp. individuals.Points are grouped by collection site and colored by species with which the collection site
individuals are compared. Comparisons of individuals within species fall in the lower band, comparisons
between individuals from the sexual species and parthenogens fall in the middle band, and comparisons
between individuals from the two different sexual species fall in the upper band.

RESEARCH | REPORT