www.ecole ol.o g Ecology and E olu ion 2016; 6: 7574–75857574
|
© 2016 The Au ho s. Ecology and E olu ion
published by John Wiley & Sons L d.
Recei ed: 27 Ap il 2016
|
Re ised: 22 July 2016
|
Accep ed: 28 Augus 2016
DOI: 10.1002/ece3.2482
Abs ac
Adap a ion o la ge- scale spa ial he e ogenei y in he en i onmen accoun s o a
majo p opo ion o gene ic di e si y wi hin species. Theo y p edic s he e osion o
adap i e gene ic a ia ion on a wi hin- popula ion le el, bu conside able gene ic di-
e si y is o en ound locally. Gene ic di e si y could be expec ed o be main ained
wi hin popula ions in empo ally o spa ially a iable condi ions i geno ypic ank o -
de s a y ac oss con as ing mic oen i onmen al se ings. Taking ad an age o ine-
esolu ion en i onmen al da a, we es ed he hypo hesis ha empe a u e
he e ogenei y among yea s could be one ac o main aining quan i a i e gene ic di-
e si y wi hin a na u al and gene ically di e se plan popula ion. We sampled ma e nal
amilies o Boeche a s ic a, an A abidopsis haliana ela i e, a one loca ion in he cen-
al Rocky Moun ains and g ew hem in h ee ea men s ha , based on eco ds om
an adjacen wea he s a ion, simula ed hou ly empe a u e changes a he na i e si e
du ing h ee summe s wi h di e ing mean empe a u es. T ea men had a signi ican
e ec on all ai s, wi h 2–3- old inc ease in abo e- and belowg ound biomass and he
highes alloca ion o oo s obse ed in he ea men simula ing he wa mes summe
on eco d a he si e. T ea men a ec ed bi a ia e associa ions be ween ai s, wi h
he weakes co ela ion be ween abo e- and belowg ound biomass in he wa mes
ea men . The magni ude o quan i a i e gene ic a ia ion o all ai s di e ed ac oss
ea men s: Gene ic a iance o biomass was 0 in he wa mes ea men , while highly
signi ican di e si y was ound in a e age condi ions, esul ing in b oad- sense he i a-
bili y o 0.31. Signi ican geno ype × en i onmen in e ac ions ac oss all ea men s
we e ound only in oo - o- shoo a io. The e o e, empe a u e a ia ion among sum-
me s appea s unlikely o accoun o he obse ed le els o local gene ic a ia ion in
size in his pe ennial species, bu may in luence amily ank o de in g ow h alloca ion.
Ou esul s indica e ha na u al en i onmen al luc ua ions can ha e a la ge impac on
he magni ude o wi hin- popula ion quan i a i e gene ic a iance.
KEYWORDS
en i onmen al he e ogenei y, geno ype × en i onmen in e ac ion, main enance o gene ic
di e si y, empe a u e, empo al he e ogenei y
1Depa men o Bo any, Uni e si y o
Wyoming, La amie, WY, USA
2P og am in Ecology, Uni e si y o
Wyoming, La amie, WY, USA
3Depa men o Molecula
Biology, Uni e si y o Wyoming, La amie,
WY, USA
Co espondence
Ma i J. Salmela, Depa men o Bo any,
Uni e si y o Wyoming, La amie, Wyoming,
USA.
Email: [email p o ec ed]om
Funding in o ma ion
Na ional Science Founda ion, G an /Awa d
Numbe : IOS-1025965
Cu en add ess
Ma i J. Salmela, Na u al Resou ces Ins i u e
Finland, Van aa, Finland
ORIGINAL RESEARCH
Na u al quan i a i e gene ic a iance in plan g ow h di e s in
esponse o ecologically ele an empe a u e he e ogenei y
Ma i J. Salmela1 | B en E. Ewe s1,2 | Cyn hia Weinig1,2,3
This is an open access a icle unde he e ms o he C ea i e Commons A ibu ion License, which pe mi s use, dis ibu ion and ep oduc ion in any medium,
p o ided he o iginal wo k is p ope ly ci ed.
|
7575
Salmela e al.
1 | INTRODUCTION
In he wild, na i e popula ions o en exhibi a i ness ad an age in hei
home en i onmen s. This phenomenon, local adap a ion, is equen ly
documen ed among popula ions ound ac oss he e ogeneous en i on-
men s and is known o main ain gene ic di e si y in i ness and many
o he quan i a i e ai s among popula ions (He e o d, 2009; Leimu &
Fische , 2008). On a mo e local wi hin- popula ion le el, howe e , na -
u al selec ion is expec ed o educe gene ic a ia ion as nonop imal
pheno ypes a e emo ed om popula ions (Kawecki & Ebe , 2004). In
con as o hese p edic ions, subs an ial quan i a i e gene ic a ia ion
is equen ly ound wi hin popula ions, e en in ai s closely ela ed
o i ness ha a e assumed o expe ience s ong selec ion (Ba on &
Keigh ley, 2002; Hill, 2010; Houle, 1992; Johnson & Ba on, 2005).
Fo example, lowe ing ime exhibi s gene ic di e si y no jus among
bu also wi hin popula ions in he mainly sel - e ilizing model species
A abidopsis haliana (L.) Heynh. in Fennoscandia and Spain (Kui inen,
Ma ila, & Sa olainen, 1997; Méndez- Vigo, Gomaa, Alonso- Blanco,
& Pico, 2013; S enøien, Fens e , Ton e i, & Sa olainen, 2005), while
a ia ion in iming o g ow h is o en dis ibu ed in a simila ashion in
highly ou c ossing o es ees (Albe o e al., 2011; Mimu a & Ai ken,
2007; Sa olainen, Bokma, Ga cía- Gil, Komulainen, & Repo, 2004).
In spi e o ex ensi e heo e ical wo k on he po en ial o di e -
en e olu iona y and ecological o ces o main ain gene ic a ia ion
wi hin popula ions, he causes o such local di e si y emain o he
mos pa poo ly unde s ood (Johnson & Ba on, 2005). Mu a ion-
selec ion balance on i s own is conside ed unlikely o accoun o he
obse ed high le els o gene ic di e si y, while a ious o ms o bal-
ancing selec ion ha e been hypo hesized o play a causal ole (Ba on
& Keigh ley, 2002; Cha leswo h, 2015; Hill, 2010; Johnson & Ba on,
2005). Fu he , gene low among gene ically dis inc popula ions om
di e gen en i onmen s may be a sou ce o a ia ion wi hin popula-
ions (Ba on, 1999). Because a popula ion’s e olu iona y po en ial
in a changing en i onmen is de e mined la gely by he magni ude
o he i able gene ic di e si y in adap i e ai s (Falcone & Mackay,
1996), he main enance o gene ic a ia ion is ele an no only o
basic esea ch on e olu iona y p ocesses bu also o applied con-
se a ion biology and a e species’ managemen (Ho mann & Sg ò,
2011). Indeed, empi ical e idence ha a leas some popula ions in
he wild ha e been able o adap o en i onmen al changes indica es
ha adap i e gene ic di e si y exis ed in p e ious gene a ions (F anks,
Webe , & Ai ken, 2014; Me ilä & Hend y, 2014; see also Gee s e al.,
2015; I win, Finkel, Mülle - Ka ge , & Ghinaglia, 2015).
In addi ion o la ge- scale spa ial a ia ion in he en i onmen ha
o en leads o gene ically di e en ia ed popula ions, en i onmen al con-
di ions may also a y a much ine spa ial and empo al scales (Linha &
G an , 1996). Consequen ly, a popula ion may encoun e a iable con-
di ions a i s home si e, which could selec o he po en ial o exp ess
a wide ange o pheno ypes ac oss dis inc mic oen i onmen s. Such
pheno ypic a ia ion could be achie ed o ins ance by gene ic di e si y
such ha di e en geno ypes ha e he highes i ness in di e ing local
en i onmen al se ings, ha is, by geno ype × en i onmen in e ac ions
(Ellne & Hai s on, 1994; Gillespie & Tu elli, 1989; Hed ick, 1995), o ia
pheno ypic plas ici y o a gene alis geno ype (Kawecki & Ebe , 2004).
Empi ical ield s udies ha e shown ha pe o mance anks o geno ypes
may change as a esul o a ia ion in in e - and in aspeci ic compe i ion
(Ba on, Richi , Villou eix, Amsellem, & Roux, 2015; Shaw, Pla enkamp,
Shaw, & Podolsky, 1995) o dis u bance (McLeod, Scasci elli, & Vellend,
2012); geno ype × en i onmen in e ac ions in i ness ha e e en been
desc ibed on a scale o jus 10 cm wi hin a single old ield (S a on,
1994). In Be ula pendula Ro h in Finland, howe e , o es g ound he -
e ogenei y on a local le el a ec ed o e all g ow h bu was no su icien
o shi geno ypic anks (Mikola e al., 2014). Gene ic mapping s udies
on ecombinan inb ed lines and expe imen al hyb ids o model sys-
ems such as A abidopsis, Boeche a, D osophila, and Mimulus in he ield
and unde con olled expe imen al condi ions ha e p o ided insigh s
in o he gene ic basis o geno ype × en i onmen in e ac ions ac oss
highly di e gen en i onmen al condi ions. These s udies commonly
show en i onmen - speci ic QTL e ec s o nume ous quan i a i e ai s
om mo phology o ep oduc i e i ness ( e iewed in Mackay, 2009;
Sa olainen, Lascoux, & Me ilä, 2013; Weinig, Ewe s, & Welch, 2014).
Gene ically a iable popula ions o p edominan ly sel - e ilizing
plan species like he mus a d Boeche a s ic a (G aham) Al- Shehbaz
(Song, Clauss, Peppe , & Mi chell- Olds, 2006) a e sui able s udy sys-
ems o explo ing he main enance o local gene ic di e si y in a y-
ing en i onmen s because long- dis ance gene low may be easonably
excluded as he sou ce o high le els o wi hin- popula ion adap i e
gene ic di e si y (c . Yeaman & Ja is, 2006). P e iously, Salmela e al.
(2016) obse ed conside able quan i a i e gene ic di e si y in he
ci cadian clock, he endogenous imekeepe egula ing daily oscilla-
ions in nume ous ai s, and a ious g ow h ai s among ma e nal
amilies o B. s ic a sampled a a high- ele a ion si e in he cen al
Rocky Moun ains in he USA. In his popula ion, he ange o amily
means in ci cadian pe iod, one cha ac e is ic o he ci cadian clock, ac-
coun ed o o e 50% o he ange ha has p e iously been epo ed
in a global se o 150 A. haliana geno ypes by Michael e al. (2003).
Quan i a i e a ia ion in he clock was associa ed wi h g ow h so
ha amilies wi h longe ci cadian pe iods g ew mo e apidly and o
a la ge size, bu had a lowe oo - o- shoo a io. Because he ci ca-
dian clock and i s a iabili y a e gene ally hough o e lec adap a ion
o a iable ligh –da k cycles (e.g., Michael e al., 2003), Salmela e al.
(2016) hypo hesized ha local gene ic di e si y in he clock could be
main ained by pe o mance ade- o s o amilies ac oss seasonal en-
i onmen s wi h di e ing pho ope iods and empe a u es. Ye , g ow h
chambe en i onmen s simula ing di e en mon hs wi hin a g owing
season esul ed in geno ype × en i onmen in e ac ions ha we e
la ge han he o e all gene ic e ec only o oo - o- shoo alloca ion.
Highe oo - o- shoo a io may enhance su i al in s ess ul and low-
esou ce en i onmen s (Poo e e al., 2012; Wilson, 1988), an adap-
i e hypo hesis also sugges ed by he la ge oo - o- shoo a ios ound
in popula ions om highe - ele a ion en i onmen s in his egion. S ill,
he exac ac o s enabling he main enance o signi ican ine- scale
gene ic di e si y in his popula ion emain unknown.
Fluc ua ing selec ion d i en by empo al a ia ion in en i onmen-
al ac o s has been p oposed o be common in he wild (Siepielski,
7576
|
Salmela e al.
DiBa is a, & Ca lson, 2009). Tempo ally a iable en i onmen al con-
di ions ha e also been hypo hesized o shape he pa e ns o adap i e
gene ic di e si y wi hin popula ions (Salmela, 2014), bu in gene al,
heo e ical s udies sugges ha condi ions unde which empo al
a ia ion in he en i onmen can main ain gene ic di e si y a e mo e
es ic i e han hose o spa ial he e ogenei y (Hed ick, 2006). Some
e idence suppo s he ole o empo al en i onmen al he e ogenei y
in main aining gene ic a ia ion wi hin popula ions: Di e en g ow h
seasons wi hin a yea ha e been epo ed o a o di e gen geno-
ypes in Ta axacum o icinale (L.) Webe ex F.H. Wigg (Va ek, McG aw,
& Yang, 1996), while annually a ying mois u e condi ions in a des-
e may conse e lowe colo polymo phism in he annual Linan hus
pa yae (A.G ay) G eene (Schemske & Bie zychudek, 2001). In he ish
Xiphopho us a ia us, he le el o empe a u e a iabili y wi hin days
and yea s has been ound o be posi i ely associa ed wi h wi hin-
popula ion di e si y in ail spo ichness (Culumbe & Toble , 2016),
wi h he i ness ank o di e en spo ypes depending o ins ance on
he he mal en i onmen (Culumbe , Schume , Monks, & Toble , 2015).
In his s udy, we we e in e es ed in u he explo ing he capaci y
o empo al abio ic a iabili y o sus ain local quan i a i e gene ic a i-
a ion in he wild using he a o emen ioned B. s ic a popula ion om
he cen al Rocky Moun ains as ou s udy sys em. Two ea u es in
pa icula make his popula ion an in e es ing model o assessing he
possible en i onmen al causes o local gene ic di e si y: I has al eady
been shown o be gene ically di e se in mul iple quan i a i e ai s
(Salmela e al., 2016), and i s home si e is loca ed close o a wea he
s a ion wi h long- e m and ine- esolu ion empe a u e eco ds. The
popula ion expe iences o ins ance a iable summe empe a u e
condi ions; be ween 1995 and 2014, mean June and July empe a-
u es a he si e a ied be ween 8.3–15°C and 13–17.2°C, espec-
i ely. Hou ly eco ded empe a u e da a om mul iple yea s allowed
us o in es iga e he e ec s o he e ogenei y in a single abio ic ac o ,
p o iding he unique oppo uni y o simula e empo ally luc ua ing
and ecologically ele an na u al en i onmen s in con olled expe i-
men al se ings. We g ew eplica es o he same se o na u ally oc-
cu ing ma e nal amilies ac oss ea men s wi h di e ing a e age
empe a u es, hypo hesizing ha i a iable summe empe a u e e-
gimes we e capable o p ese ing quan i a i e gene ic di e si y in his
popula ion, no only would he ea men s a ec o e all means bu
also induce geno ype × en i onmen in e ac ions in g ow h and i s al-
loca ion. Mo eo e , we expec ed he magni ude o hese in e ac ions
o be la ge compa ed o he a e age e ec o amily and a la ge p o-
po ion o he in e ac ion a iance o be due o ank shi s o amilies
among simula ed g owing season empe a u e cycles.
2 | MATERIALS AND METHODS
2.1 | Sampling
We collec ed seed ma e ial o he s udy by ma e nal amily a a dis ance
o abou 300–500 m om he Sou h B ush C eek SNOTEL wea he
s a ion (41.333°N, 106.500°W; ele a ion 2,572 m) in sou heas e n
Wyoming on 26 July 2012. The inb eeding coe icien in popula ions
o B. s ic a is high (0.74–0.98 in Song e al., 2006), sugges ing ha ma-
e nal p ogeny consis mos ly o ull- sibs. All excep one o he sampled
plan s we e loca ed along a ansec o app oxima ely 300 m. Because
we we e in e es ed in na u al pa e ns o a ia ion p esen in he popula-
ion a he ime o sampling a he han cha ac e izing he gene ic back-
g ound o quan i a i e ai a ia ion, we used wild- collec ed seeds in
he expe imen . O he 21 amilies in es iga ed in his s udy, 19 we e in-
cluded in he s udy by Salmela e al. (2016) ha documen ed signi ican
wi hin- popula ion gene ic a ia ion in he ci cadian clock and g ow h.
2.2 | T ea men s
Based on eco ds om he wea he s a ion (h p://www.wcc.
n cs.usda.go /nwcc/si e?si enum=772), we p og ammed g ow h
chambe compa men s (PGC- 9/2 wi h Pe ci al Ad anced In ellus
En i onmen al Con olle , Pe ci al Scien i ic, Pe y, IN, USA) o ack
hou ly empe a u e changes a he si e du ing an 8- week pe iod on
June 1–July 25 in 2009, 2010, and 2012. Thus, each ea men con-
sis ed o o e 1,300 empe a u e s eps. In he ollowing sec ions o
he a icle, we will e e o hese ea men s by he co esponding
yea . We chose o ocus on June and July because an ea lie s udy
indica ed he mos igo ous g ow h in a ea men simula ing ea ly
g owing season condi ions (Salmela e al., 2016) and because in May
he si e egula ly expe iences eezing empe a u es as low as −15°C
ha canno be achie ed in he chambe s used. Daily mean empe a-
u es in each ea men a e shown in Figu e 1. These ea men s
ep esen ed a ange o mean June empe a u es a he si e (10.2°C
in 2009, 12.0°C in 2010, and 15.0°C in 2012) du ing a pe iod om
which hou ly eco ded empe a u e da a we e a ailable; he summe
o 2012 is he ho es on eco d a he si e and in he es o Wyoming
(NOAA Na ional O e iew o Annual 2012). Due o lack o hou ly
da a, we did no simula e he cooles June on eco d (1998, wi h a -
e age empe a u e o 8.3°C). A e age daily maximum and minimum
empe a u es in June we e 17.4°C and 3.6°C in 2009, 19.6°C and
4.2°C in 2010, and 23.4°C and 5.1°C in 2012. Unde chambe condi-
ions, mean July empe a u es we e 14.7°C in bo h 2009 and 2010,
and 16.3°C in 2012. This esul ed in o e all ea men means o ap-
p oxima ely 12.5°C in 2009, 13.4°C in 2010, and 15.7°C in 2012. Due
o chambe es ic ions on minimum empe a u e (4°C in he da k,
10°C wi h all ligh s on), o e all expe imen al condi ions we e sligh ly
(0.09–0.23°C) highe han hose in he ield.
Pho ope iod was app oxima ely 15 h , wi h g adual changes in a
simila ashion in all ea men s acco ding o local sun ise and sunse
imes o e he cou se o he 8- week pe iod. To simula e dawn and
dusk, 50% o he ligh s we e on du ing he i s and las hou o he
pho ope iod. Maximum pho osyn he ically ac i e i adiance, measu ed
wi h he ligh me e LI- 250 (LI- COR Biosciences, Lincoln, NE, USA)
in all chambe s be o e he expe imen , was app oxima ely 250 μmol
pho ons m−2 s−1 a he plan le el. We did no eplica e ea men s;
hus, amily × ea men e ec s canno be unequi ocally a ibu ed o
empe a u e. Howe e , a ia ion in o he con ollable ac o s is likely
o ha e been small compa ed o empe a u e due o e y simila ligh
le els in all chambe compa men s. Daily moni o ing o empe a u e
|
7577
Salmela e al.
by he mome e s also indica ed ha chambe condi ions adhe ed o
p og ammed se ings.
We ge mina ed and plan ed seeds om he 21 amilies, wi h
en eplica es pe amily in each ea men , ollowing he p o ocol
in Salmela e al. (2016). We did no sco e iming o ge mina ion due
o apid ge mina ion on mois pape in all he amilies wi hin 4 days
ollowing cold s a i ica ion. We di ided he seedlings in o he h ee
ea men s, wi h en blocks wi hin each and one eplica e pe amily
andomized wi hin a block. We wa e ed po s o ield capaci y e e y
2 days and o a ed la s wi hin he chambe compa men s wice
e e y week. The species equi es e naliza ion in o de o lowe ;
due o di icul ies in simula ing ecologically ele an high- ele a ion
au umn, win e , and sp ing condi ions in ou expe imen al se ings
(ai empe a u e a he si e can d op below −30°C), we measu ed he
equi alen o i s - yea g ow h only. A e 8 weeks, we sco ed he
numbe o lea es and he leng h o he longes lea in all eplica es
and sampled whole plan s o abo e- and belowg ound biomass and
oo - o- shoo a io. We d ied plan s in an o en a 65°C o 3 days be-
o e biomass measu emen s. Lea measu es we e s ongly co ela ed
wi h abo eg ound biomass ( he numbe o lea es: ≥ .807, p < .0001
in all ea men s; he longes lea leng h s. √abo eg ound biomass:
≥ .923, p < .0001 in all ea men s). Because quali a i ely simila e-
sul s we e ob ained wi h all co ela ed measu es, we will use biomass
only as a measu e o abo eg ound plan g ow h. We sampled plan s
a 8 weeks because he oldes lea es o en s a o senesce a his
poin in chambe condi ions (M.J. Salmela, pe sonal obse a ion) and
because an 8- week pe iod is likely o accoun o a la ge p opo ion o
he annual g owing season a he home si e o he popula ion.
2.3 | S a is ical analyses
We included amilies wi h a leas ou eplica es in wo ea men s
in he analysis. P io o o mal s a is ical es s, we examined whe he
ou da a me assump ions o analysis o a iance. We de ec ed a i-
ance he e ogenei y o esiduals among ea men s in all ai s, wi h
a iances a ying 6.6- old o abo eg ound biomass, 9.5- old o
belowg ound biomass, and 4.8- old o oo - o- shoo a io. These
anges o a ia ion esul ed om he 2012 ea men ha had he
mos pheno ypic a ia ion in all ai s; a iance di e ences be ween
2009 and 2010 we e less han wo old. To de e mine he in luence
o unequal a iances, we an all s a is ical analyses wi h he comple e
da ase using bo h o iginal and ans o med measu emen s. While
ans o ma ions educed a iance he e ogenei y, hey did no change
he ou come o he analyses o in e p e a ion o he esul s. This ou -
come p obably e lec s he obus ness o analysis o a iance o un-
equal a iances in he case o simila sample sizes in all ea men s.
The e o e, we used he aw alues in he analyses p esen ed he e.
We i s es ed o he signi ican e ec s o amily and block
wi hin each ea men using gene al linea models. We conside a i-
a ion among ma e nal amilies o ep esen gene ic di e si y wi hin
he popula ion, and hus, p opo ion o o al a ia ion explained by
among- amily di e si y in each ea men is an es ima e o b oad-
sense he i abili y (H2). We es ima ed a iance componen s o amily,
block, and esidual a ia ion using he REML app oach. We examined
associa ions be ween he ai s wi hin ea men s using eplica e al-
ues and Pea son’s co ela ion. Sca e plo s sugges ed he slope be-
ween abo e- and belowg ound biomass a ies among ea men s. To
es o ea men di e ences in slope (ß1), we pe o med a eg ession
analysis o abo eg ound biomass on belowg ound biomass in which
we included ea men as a ca ego ical ac o and he abo eg ound
biomass × ea men in e ac ion e m.
To in es iga e he con ibu ions o di e en sou ces o a ia ion
ac oss all h ee ea men s and be ween pai s o ea men s, we
used he model abo e- o belowg ound biomass o oo - o- shoo
a io = ea men + amily + amily × ea men + block( ea men ).
We conside ed ea men a ixed ac o . When nonsigni ican a he
FIGURE1 Va ia ion in daily mean
empe a u es in g ow h chambe
ea men s simula ing h ee summe s wi h
di e ing a e age empe a u es a Sou h
B ush C eek in sou heas e n Wyoming.
O e all ea men means we e 12.5°C
in 2009, 13.4°C in 2010, and 15.7°C
in 2012. Due o he hou ly acking o
empe a u e he e ogenei y a he si e, each
ea men consis ed o o e 1,300 di e en
empe a u e s eps
7578
|
Salmela e al.
0.10 le el, we emo ed he in e ac ion e m om he analysis. Gi en
ha among- amily a ia ion measu es gene ic di e si y wi hin he
popula ion, he amily × ea men e m de ines geno ype × en i on-
men in e ac ions. We es ima ed a iance componen s o he andom
ac o s using he REML app oach. Geno ype × en i onmen in e ac-
ions can main ain di e si y i hey a e caused by ank shi s o geno-
ypes a he han by di e ences in among- geno ype a iance among
ea men s (Mi chell- Olds, 1992). In o de o sepa a e he e ec s o
shi s in amily ank o de s. among- amily a iance on a s a is ically
signi ican in e ac ion a iance componen , we used he equa ion
(Cocke ham, 1963):
whe e VFamily × ea men is he a iance componen due o amily ×
ea men in e ac ion, σi and σj a e he squa e oo s o among- amily
a iances in ea men s i and j, GE is he gene ic co ela ion be ween
ea men s i and j, and is he numbe o ea men s. The i s pa
o he equa ion accoun s o a iance due o changes in ank o de
among ea men s, while he second pa accoun s o ea men di -
e ences in among- amily a iance. We es ima ed gene ic co ela ions
using amily a iance componen s wi hin each ea men and ac oss
pai s o ea men s (Windig, 1997):
whe e VFamily is he a iance componen due o amily ac oss he wo
en i onmen s, and VFamily, ea men i and VFamily, ea men j a e he a i-
ance componen s due o amily in ea men s i and j. We ca ied ou all
s a is ical analyses wi h IBM SPSS S a is ics Ve sion 23.
3 | RESULTS
Table 1 shows gene al linea model esul s o he h ee ai s
wi hin each ea men , along wi h ea men means and measu es
V
Family × ea men =
(∑
i<j
[2σiσj(1− GE)+(σi−σj)2]
)
∕( ( −
1),
GE =VFamily∕
√(
VFamily, ea men i×VFamily, ea men j
),
TABLE1 Gene al linea model esul s o abo e- and belowg ound biomass and oo - o- shoo a io in a popula ion o Boeche a s ic a
wi hin each ea men simula ing di e ing g owing season empe a u e condi ions a Sou h B ush C eek in sou heas e n Wyoming
Fac o d
Abo eg ound biomass Belowg ound biomass Roo - o- shoo a io
MS F- a io p- alue % MS F- a io p- alue % MS F- a io p- alue %
2009
Family 18 0.00124 1.84 * 9.96a0.0000557 1.58 .0784 6.53a0.00734 2.18 ** 11.7a
Block 9 0.00119 1.76 .0840 4.92 0.0000607 1.72 .0931 4.89 0.0111 3.30 ** 11.9
Residual 105 0.000674 85.1 0.0000352 88.6 0.00336 76.3
Mean
(±SE)
0.0557 (±0.00243) g 0.0111 (±0.000547) g 0.202 (±0.00570)
VP7.99 × 10−4 4.02 × 10−5 4.43 × 10−3
VF7.96 × 10−5 2.63 × 10−6 5.21 × 10−4
2010
Family 20 0.00460 5.35 **** 31.0a0.000179 4.52 **** 26.2a0.00590 3.24 **** 20.1a
Block 9 0.00311 3.61 *** 8.84 0.000163 4.10 *** 11.0 0.00659 3.62 *** 11.4
Residual 147 0.000860 60.2 0.0000396 62.8 0.00182 68.5
Mean
(±SE)
0.0573 (±0.00282) g 0.0105 (±0.000593) g 0.174 (±0.00384)
VP1.42 × 10−3 6.28 × 10−5 2.66 × 10−3
VF4.41 × 10−4 1.64 × 10−5 5.35 × 10−4
2012
Family 19 0.00469 0.947 ns 0a0.000377 0.957 ns 0a0.0275 3.04 *** 19.5a
Block 9 0.000855 0.173 ns 0 0.000200 0.508 ns 00.0298 3.29 ** 11.2
Residual 120 0.00495 100 0.000394 100 0.00906 69.3
Mean
(±SE)
0.115 (±0.00561) g 0.0311 (±0.00160) g 0.268 (±0.00934)
VP4.69 × 10−3 3.83 × 10−4 1.30 × 10−2
VF0 0 2.52 × 10−3
The pe cen age shows he p opo ion o o al a ia ion explained by each ac o . VP = a iance componen o pheno ypic a ia ion; VF = a iance compo-
nen o among- amily (gene ic) a ia ion.
ns = p > .10, *p < .05, **p < .01, ***p < .001, ****p < .0001.
aEs ima e o b oad- sense he i abili y (H2).

|
7579
Salmela e al.
o pheno ypic and gene ic di e si y. In 2009, he amily e ec was
signi ican o abo eg ound biomass and oo - o- shoo a io, an in-
dica ion o gene ic a ia ion in he popula ion. Fo belowg ound bi-
omass, he amily e ec was close o signi icance a he 0.05 le el
(p = .078). In 2010, he amily e ec was highly signi ican o all ai s.
Al hough 2012 was he mos pheno ypically a iable ea men , we
ound no signi ican gene ic a ia ion o abo e- o belowg ound bio-
mass in his yea ; a ia ion was en i ely esidual. Fo oo - o- shoo
a io, howe e , amily di e ences in 2012 we e highly signi ican .
Fo abo eg ound biomass, pheno ypic and gene ic a ia ion in 2009
we e 56.3% and 18.0%, espec i ely, o hose in 2010, esul ing in
highe b oad- sense he i abili y in 2010. The pa e n was simila o
belowg ound biomass: Pheno ypic and gene ic a ia ion in 2009 we e
64.0% and 16.0%, espec i ely, o hose in 2010. These di e ences
we e no due o he wo amilies in 2010 o which we lacked da a
in 2009; when hey we e excluded, we ob ained e y simila es i-
ma es o wi hin- ea men di e si y. Simila le els o gene ic di e si y
we e exp essed in oo - o- shoo a io in 2009 and 2010, bu b oad-
sense he i abili y in 2009 was lowe due o 66.5% mo e pheno ypic
a ia ion in his ea men . We obse ed he highes pheno ypic and
gene ic a ia ion in oo - o- shoo a io in 2012. Howe e , due o in-
c eases o simila magni ude in all a iance componen s, b oad- sense
he i abili y in 2012 was simila o ha in 2010.
Abo e- and belowg ound biomass we e posi i ely co ela ed in
all ea men s (Figu e 2). This associa ion was he s onges in 2010
(Figu e 2A) and he weakes in 2012 (Figu e 2B). A signi ican abo-
eg ound biomass × ea men in e ac ion in he eg ession analysis
indica ed ea men di e ences in he slope be ween he wo a iables
(F2,453 = 4.13, p < .05), wi h e y simila es ima es in 2009 and 2010
(2009: ß1 = 0.192, 95% CI: 0.171–0.212; 2010: ß1 = 0.191, 95% CI:
0.178–0.204; Figu e 2A) and a s eepe one in 2012 (ß1 = 0.233, 95%
CI: 0.206–0.260; Figu e 2B). Belowg ound biomass was posi i ely and
mode a ely co ela ed wi h oo - o- shoo a io in all ea men s (2009:
= .470, p < .0001, 95% CI: 0.330–0.605; 2010: = .592, p < .0001,
95% CI: 0.482–0.689; 2012: = .569, p < .0001, 95% CI: 0.464–0.664).
In 2010, abo eg ound biomass showed a weak posi i e co ela ion wi h
oo - o- shoo a io ( = .288, p < .001, 95% CI: 0.153–0.422). In 2009
and 2012, his co ela ion was no signi ican ly di e en om ze o.
T ea men had a highly signi ican e ec on all h ee ai s (Table 2).
Abo eg ound biomass was wice and belowg ound biomass abou h ee
imes la ge in 2012 han in 2009 o 2010, and hus, also oo - o- shoo
a io displayed i s highes ea men mean in 2012 (Table 1). Fo abo e-
and belowg ound biomass, he only signi ican andom ac o ac oss
all ea men s was amily. We ound a signi ican amily × ea men
in e ac ion e m o oo - o- shoo a io, e ealing a ia ion in pheno-
ypic plas ici y among ma e nal amilies in he popula ion. The a iance
componen due o he in e ac ion was abou 27% bigge (6.49 × 10−4)
han ha o amily (5.10 × 10−4). App oxima ely 63% o he in e ac ion
a iance was due o ank changes o amilies ac oss he ea men s
(Figu e 3A), esul ing in e y simila magni ude o a iance componen s
due o amily s. ank o de changes. In acco dance wi h he mode a e
o e all amily e ec , gene ic co ela ions o oo - o- shoo a io be-
ween ea men s we e posi i e: 0.720 be ween 2009 and 2010, 0.527
be ween 2009 and 2012, and 0.521 be ween 2010 and 2012.
Table 3 shows esul s o pai wise compa isons o ea men s. We
ound highly signi ican ea men di e ences in biomass be ween 2009
and 2012, and 2010 and 2012. Al hough he o e all e ec o ea men
was no signi ican in he compa ison o 2009 and 2010, we obse ed
a signi ican amily × ea men in e ac ion o abo eg ound biomass.
Howe e , he a iance componen due o amily was 1.73 imes bigge
han ha due o he in e ac ion, signaling ha he signi ican in e ac ion
a ises mainly om he ea men di e ences in among- amily a iance
(Table 1). Fo belowg ound biomass in he same ea men compa i-
son, he pa e n was compa able: The amily × ea men in e ac ion
FIGURE2 Va iable associa ions be ween abo e- and belowg ound biomass wi hin a popula ion o Boeche a s ic a sampled a Sou h B ush
C eek in sou heas e n Wyoming and g own in h ee ea men s simula ing di e ing summe empe a u e condi ions a i s na i e si e. 2012
di e ed om he wo o he ea men s in i s o e all ai means and a iances. Fo cla i y, we show da a o 2012 in a sepa a e igu e. (A) In
ea men s simula ing he summe s o 2009 and 2010, he co ela ions and slopes be ween he ai s we e no signi ican ly di e en om each
o he . (B) In he ea men simula ing he summe o 2012, he co ela ion be ween he ai s was signi ican ly weake han in 2010. Also, he
slope was signi ican ly s eepe (ß1 = 0.233) han he one in 2010 (ß1 = 0.191)
7580
|
Salmela e al.
was close o signi icance a he 0.05 le el (p = .0880). Fo oo - o- shoo
a io, pai wise compa isons o ea men s indica ed mode a e di e -
ences in mean be ween 2009 and 2010, and highly signi ican di e -
ences be ween 2009 and 2012, and 2010 and 2012. The di e ence
be ween 2009 and 2010, despi e he lack o signi ican ea men e -
ec s in i s wo componen s, was likely caused by small bu opposi e
changes in he wo biomass measu es: O e all abo eg ound biomass
was sligh ly la ge in 2010 han in 2009, while he opposi e pa e n was
ound o belowg ound biomass (Table 1). In he compa ison o 2009
and 2010 o 2012, we de ec ed some suppo o a amily × ea men
in e ac ion (p = .0941). In he pai wise compa isons o 2009 and 2010
o 2012, we obse ed mo e s a is ical suppo o amily × ea men
in e ac ions han o amily. Toge he wi h a ia ion in he s eng h o
gene ic co ela ions, he esul s indica e ha di e ing amily esponses
o 2012 in pa icula accoun o he signi ican in e ac ion e m in oo -
o- shoo a io ac oss all ea men s.
4 | DISCUSSION
In his s udy, we combined en i onmen al da a wi h quan i a i e
gene ic app oaches o examine whe he empo al en i onmen al
he e ogenei y could ac as one mechanism main aining gene ic a i-
a ion in g ow h and i s alloca ion wi hin a plan popula ion expe i-
encing a iable condi ions a i s home si e. In o de o ca y ou he
expe imen in as ecologically ele an chambe condi ions as possi-
ble, we ook ad an age o ine- esolu ion empe a u e da a om a
wea he s a ion ha was loca ed in he icini y o he sampled popula-
ion. This enabled us o quan i y he po en ial o na u al among- yea
he e ogenei y in g owing season empe a u es o induce geno ype ×
en i onmen in e ac ions in i s - yea g ow h ai s.
Sel - e iliza ion is expec ed o educe gene ic di e si y and conse-
quen ly limi long- e m su i al p ospec s o popula ions ( e iewed in
W igh , Kalisz, & Slo e, 2013). Ye , popula ions o p edominan ly sel -
e ilizing plan species like A. haliana and B. s ic a may ha bo con-
side able le els o local quan i a i e and molecula gene ic a ia ion
(Ba on e al., 2015; Kui inen e al., 1997; Méndez- Vigo e al., 2013;
Salmela e al., 2016; Siemens, Haugen, Ma zne , & Vanasma, 2009;
Song e al., 2006; S enøien e al., 2005). Because he a e a which
he popula ion mean o a quan i a i e ai can change is posi i ely
co ela ed wi h he amoun o gene ic di e si y in he ai (Falcone
& Mackay, 1996; Houle, 1992), unde s anding he main enance o ge-
ne ic a ia ion in he wild has become one o he key ques ions in
mode n e olu iona y and conse a ion biology (F anks e al., 2014;
TABLE2 Gene al linea model esul s o abo e- and belowg ound biomass and oo - o- shoo a io ac oss all ea men s
Fac o d
Abo eg ound biomass Belowg ound biomass
d
Roo - o- shoo a io
MS F- a io p- alue MS F- a io p- alue MS F- a io p- alue
T ea men 2 0.156 87.1 **** 0.0193 128 **** 2 0.344 17.7 ****
Family 20 0.00582 2.71 *** 0.000256 1.65 * 20 0.0215 2.29 *
Family × ea men 37 0.00950 2.07 ***
Block( ea men ) 27 0.00177 0.822 ns 0.000151 0.972 ns 27 0.0158 3.45 ****
Residual 409 0.00215 0.000155 372 0.00459
ns = p > .10, * p < .05, ** p < .01, *** p < .001, **** p < .0001.
In e ac ion e ms we e included in he models when signi ican a he p < .10 le el.
FIGURE3 (A) Reac ion no ms o oo - o- shoo a io o na u ally occu ing ma e nal amilies o Boeche a s ic a ha we e sampled a
Sou h B ush C eek in sou heas e n Wyoming and g own in h ee ea men s simula ing summe s wi h di e ing a e age empe a u es a he
popula ion’s home si e. (B) The nega i e associa ion be ween mean June empe a u e and p ecipi a ion du ing he mon h a Sou h B ush C eek
be ween 1995 and 2014 ( = −.737, p < .001)
|
7581
Salmela e al.
Me ilä & Hend y, 2014). Ou esul s e ealed no only ha he e was
quan i a i e gene ic a ia ion in he high- ele a ion B. s ic a popula-
ion om he Rocky Moun ains bu also ha he magni ude o such
di e si y a ied g ea ly among ea men s mimicking na u al g owing
season empe a u e he e ogenei y. Gene ic di e si y in abo eg ound
biomass was he highes unde he summe empe a u es o 2010,
while in he ea men simula ing he summe o 2009 ha was on
a e age only 0.9°C coole han 2010, gene ic a iance had educed
by o e 80%. Fo belowg ound biomass, we ound a quali a i ely sim-
ila pa e n, al hough in his case he amily e ec in 2009 was a li -
le weake . A di e gen pa e n o gene ic di e si y eme ged in he
ea men simula ing he wa mes summe eco ded a he home si e
o he popula ion: Va ia ion in bo h abo e- and belowg ound biomass
was comple ely esidual in 2012. Due o he di e gence in a iances,
b oad- sense he i abili y o biomass a ied be ween 0 and 0.31 ac oss
wo ea men s wi h a mean empe a u e di e ence o only 2.3°C.
He i abili ies and gene ic a iances commonly a y ac oss en-
i onmen s, which complica es p edic ing esponses o selec ion
(Cha man ie & Ga an , 2005; Falcone & Mackay, 1996). Lowe le els
o quan i a i e gene ic di e si y may be exp essed in un a o able con-
di ions (Cha man ie & Ga an , 2005; Ho mann & Me ilä, 1999). Thus,
ela i e o he in e media e empe a u es in 2010, he coole g ow h
en i onmen du ing he mon h o June 2009 may ha e con ibu ed o
he educed gene ic a iabili y in biomass in his yea . Howe e , he
subs an ial biomass inc ease obse ed in esponse o he simula ed
2012 condi ions indica es ha he unusually wa m empe a u es on
hei own we e no de imen al o ou s udy popula ion. Ins ead, he
lack o quan i a i e gene ic di e si y in biomass in his ea men may
be due o he a ypical high empe a u es, o he no el combina ion o
high empe a u e and mois u e ha his pa icula popula ion would
be unlikely o encoun e in na u e. Indeed, one limi a ion o ou s udy
is ha al hough we p og ammed ou chambe ea men s o ack
closely na u al empe a u e luc ua ions ha had occu ed a he si e,
we in es iga ed he e ec s o a single en i onmen al a iable only.
We kep plan s well- wa e ed h oughou he expe imen , bu eco ds
om he wea he s a ion show ha no p ecipi a ion occu ed a he
si e in June 2012. The egion expe ienced se e e d ough du ing he
summe o 2012 (NOAA Na ional O e iew o Annual 2012), which
would be expec ed o inc ease alloca ion o oo s bu also limi bio-
mass accumula ion in he shoo (Mokany, Raison, & P okushkin, 2006;
Poo e e al., 2012). The e o e, a di e ing esponse migh ha e been
obse ed in ou expe imen had we simula ed na u al a ia ion in mul-
iple en i onmen al ac o s simul aneously.
In he wild, na u al selec ion ope a es on o al pheno ypic a iance
shaped by bo h gene ic and nongene ic ac o s (Falcone & Mackay,
1996). Because we we e in e es ed in he esponse o a na u al pop-
ula ion o he en i onmen al he e ogenei y o i s na i e en i onmen ,
we used wild- collec ed seeds in he expe imen . As a consequence, i
TABLE3 Gene al linea model esul s o pai wise ea men compa isons
Fac o
2009–2010 2009–2012 2010–2012
d MS F- a io p- alue d MS F- a io p- alue d MS F- a io p- alue
Abo eg ound biomass
T ea men 1 0.000417 0.159 ns 1 0.208 164 **** 1 0.244 121 ****
Family 20 0.00382 2.62 * 20 0.00386 1.34 ns 20 0.00582 2.12 **
Family ×
ea men
18 0.00146 1.86 *
Block( ea men ) 18 0.00215 2.74 *** 18 0.00105 0.362 ns 18 0.00196 0.715 ns
Residual 252 0.000783 242 0.00289 286 0.00275
Belowg ound biomass
T ea men 1 0.0000261 0.217 ns 1 0.0244 165 **** 1 0.0317 166 ****
Family 20 0.000158 2.77 * 20 0.000224 0.992 ns 20 0.000287 1.42 ns
Family ×
ea men
18 0.0000569 1.51 .0880
Block( ea men ) 18 0.000112 2.95 **** 18 0.000137 0.608 ns 18 0.000190 0.935 ns
Residual 252 0.0000378 242 0.000226 286 0.000203
Roo - o- shoo a io
T ea men 1 0.0488 5.31 * 1 0.287 11.5 ** 1 0.687 29.3 ****
Family 20 0.00935 2.55 * 20 0.0213 1.62 ns 20 0.0212 1.83 .0949
Family ×
ea men
18 0.00367 1.49 .0941 17 0.0131 2.05 ** 19 0.0117 2.30 **
Block( ea men ) 18 0.00883 3.59 **** 18 0.0204 3.19 **** 18 0.0182 3.59 ****
Residual 252 0.00246 225 0.00640 267 0.00507
ns = p > .10, *p < .05, **p < .01, ***p < .001, ****p < .0001.
In e ac ion e ms we e included in he models when signi ican a he p < .10 le el.
7582
|
Salmela e al.
is possible ha ou es ima es o gene ic a iance also include a com-
ponen due o ma e nal e ec s. Va ia ion in he en i onmen may
esul in di e en ial ma e nal p o isioning du ing seed de elopmen ,
he e ec s o which could esemble hose caused by seg ega ing ge-
ne ic di e si y (e.g., Bischo & Mülle - Schä e , 2010). Fu he , ma e -
nal e ec s may be adap i e so ha hey enhance o sp ing i ness in
condi ions simila o hose expe ienced by he pa en s (Galloway &
E e son, 2007). D ough expe ienced by pa en s has been ound o
enhance d ough ole ance o o sp ing in B. s ic a (Alsdu , Ripley,
Ma zne , & Siemens, 2013), bu he gene al impo ance o ma e nal
e ec s o ai a ia ion in his species is no known. While we canno
ule ou po en ial ansgene a ional e ec s on he obse ed pheno-
ypic a ia ion, i is likely ha he seeds used in he expe imen ma-
u ed unde simila empe a u e and mois u e condi ions because all
seed collec ions ook place on he same day and wi hin a ange o a
ew hund ed me e s a a single loca ion. Rega dless o he p ecise un-
de lying causes, i is no ewo hy ha in ou expe imen , he g ea es
po en ial o e olu iona y esponses o selec ion in biomass exis ed
unde in e media e empe a u es and ha pa e ns o a ia ion unde
such condi ions we e no indica i e o hose in no el en i onmen al
se ings caused by highe empe a u e.
One ac o ha may play a ole in he main enance o mul iple
pheno ypically dis inc lineages in a popula ion is en i onmen al he -
e ogenei y on a local scale (Gillespie & Tu elli, 1989; Linha & G an ,
1996; Salmela, 2014). We hypo hesized ha he exis ence o ine-
scale gene ic a ia ion in g ow h in ou s udy sys em could be ela ed
o among- yea a ia ion in g owing season empe a u e condi ions
in he home en i onmen . Size is a equen i ness p oxy in plan s,
and shoo size has been epo ed o co ela e mode a ely wi h ep o-
duc i e e o in B. s ic a (Siemens & Haugen, 2013); hus, changes in
ank o de o biomass could be expec ed o in luence i ness anks,
oo. In compa ison wi h 2009 and 2010, we ound ha a e age abo-
eg ound biomass doubled and belowg ound biomass ipled unde
he unusually wa m condi ions o 2012. Biomass inc ease in esponse
o wa m empe a u es is common in plan s, especially when g ow-
ing in nonlimi ing condi ions (Lin, Xia, & Wan, 2010; P egi ze , King,
Bu on, & B own, 2000). Howe e , despi e he signi ican ea men
and amily e ec s, we obse ed no signi ican amily × ea men in-
e ac ions in biomass ac oss he h ee empe a u e en i onmen s. In
he pai wise compa ison o 2009 and 2010, he amily × ea men
in e ac ion e ms we e signi ican o close o signi ican , bu he a i-
ance componen s due o amily we e o g ea e magni ude han hose
o he in e ac ions. This mi o s he obse ed ea men di e ences
in gene ic a iances wi h li le change in amily ank o de ac oss em-
pe a u e egimes.
The lack o amily × ea men in e ac ions in biomass sugges s
ha empe a u e a ia ion among g owing seasons alone is unlikely
o accoun o he gene ic di e si y in g ow h in ou s udy sys em
(Mi chell- Olds, 1992). This is in acco dance wi h a p e ious expe i-
men ha explo ed he possibili y ha pho ope iodic and empe a u e
a ia ion wi hin a po en ial g owing season could main ain gene ic
di e si y in g ow h ai s in he same B. s ic a popula ion (Salmela
e al., 2016). In he ea lie s udy, mon h wi hin a g owing season (June,
Augus , o Sep embe ) had a la ge e ec on o e all g ow h, bu mo e
a ia ion was due o amily han amily × ea men in e ac ions ac oss
he h ee di e en mon hs. In hese g owing season ea men s, bo h
pho ope iod and empe a u e a ied and he ange o a ia ion in em-
pe a u e among ea men s was mo e ex ensi e han in he cu en
s udy. Ne e heless, al hough we did no de ec signi ican amily ×
ea men in e ac ions o biomass, condi ional neu ali y a he le el
o ma e nal amily may slow he e osion o gene ic a ia ion (c . F y,
Heinsohn, & Mackay, 1996; Schnee & Thompson, 1984). Speci ically,
in yea s simila o ou 2009 and 2010 ea men s, gene ic a ia ion
in size would be exp essed; bu in yea s simila o 2012, gene ic di-
e si y in he popula ion could be masked, and selec ion migh no
di e en ia e among dis inc ma e nal lineages (e.g., Me ilä, 1997). I is
also possible ha ine- scale gene ic di e si y in g ow h in his pe en-
nial species is main ained by ex ensi e empo al he e ogenei y in low
win e , sp ing, and au umn empe a u es ha could no be simula ed
in he chambe s used, o by o he abio ic o bio ic ac o s, such as ine
spa ial he e ogenei y in soil o compe i ion (Ba on e al., 2015; Shaw
e al., 1995; S a on, 1994).
Pa e ns o a ia ion exp essed o oo - o- shoo a io di e ed
om he wo o he ai s. The ai exhibi ed signi ican gene ic di e -
si y in all h ee ea men s, and in con as o biomass, he mos ge-
ne ic a ia ion was exp essed in 2012. The wo o he ea men s had
e y simila le els o gene ic a ia ion, bu di e ences in he amoun
o esidual a ia ion esul ed in b oad- sense he i abili y in 2010 being
almos wice as high as in 2009. The signi ican gene ic di e si y in
his ai in 2012 despi e he lack o such a ia ion in i s wo compo-
nen s may be due a di e ing associa ion be ween abo e- and below-
g ound biomass in his ea men : The co ela ion be ween he ai s
was weake and belowg ound biomass gain pe an inc ease in abo-
eg ound biomass la ge in 2012 han in 2010.
Simila o biomass, we de ec ed he highes o e all mean o oo -
o- shoo a io in 2012. The ai is known o be sensi i e o di e -
en en i onmen al cues such as ligh and mois u e, and limi a ions in
abo e- o belowg ound esou ces o en esul in inc eased alloca ion
o he co esponding pa s o he plan (Poo e & Nagel, 2000). In ou
expe imen , plan s g ew in well- wa e ed condi ions bu a iable em-
pe a u es. Reduced wa e up ake by oo s in coole empe a u es may
unde lie sligh ly inc eased alloca ion o oo s in 2009 (Poo e e al.,
2012). Based on among- species pa e ns o a ia ion, oo - o- shoo
a io is expec ed o be nega i ely co ela ed wi h shoo size in he -
baceous plan s (Poo e e al., 2012; Wilson, 1988; see also Mokany
e al., 2006); in ou s udy ocusing on a much ine le el o wi hin-
species di e si y ound in he wild, inc eases in abo e- and below-
g ound biomass in he 2012 ea men also led o highe oo - o- shoo
a ios. La ge alloca ion o oo s in his ea men may be ela ed o
he pa e ns o co a ia ion be ween empe a u e and p ecipi a ion in
he popula ion’s na i e en i onmen : Du ing a 20- yea pe iod (1995–
2014), a e age June empe a u e and p ecipi a ion inc emen du ing
he mon h a he si e we e nega i ely co ela ed ( = −.737, p < .001,
Figu e 3B). Du ing Junes o 2012 and 2013, he 2 yea s wi h he high-
es mean June empe a u es, no p ecipi a ion was eco ded a he
si e, condi ions unde which inc eased alloca ion o oo s would be