500 words 1.5 hrs

rsbl.royalsocietypublishing.org

Research Cite this article: Hallworth MT, Marra PP, McFarland KP, Zahendra S, Studds CE. 2018

Tracking dragons: stable isotopes reveal the

annual cycle of a long-distance migratory

insect. Biol. Lett. 14: 20180741. http://dx.doi.org/10.1098/rsbl.2018.0741

Received: 23 October 2018

Accepted: 23 November 2018

Subject Areas: ecology, behaviour

Keywords: Anax, common green darner, stable-hydrogen

isotopes, migration phenology

Author for correspondence: Michael T. Hallworth

e-mail: [email protected]

Electronic supplementary material is available

online at https://dx.doi.org/10.6084/m9.

figshare.c.4320911.

& 2018 The Author(s) Published by the Royal Society. All rights reserved.

Animal behaviour

Tracking dragons: stable isotopes reveal the annual cycle of a long-distance migratory insect

Michael T. Hallworth1,2, Peter P. Marra1, Kent P. McFarland3, Sara Zahendra3

and Colin E. Studds1,2

1Migratory Bird Center, Smithsonian Conservation Biology Institute, Washington, DC 20008, USA 2Department of Geography and Environmental Systems, University of Maryland, Baltimore County, Baltimore, MD 21250, USA 3Vermont Center for Ecostudies, White River Junction, VT 05001, USA

MTH, 0000-0002-6385-3815; PPM, 0000-0002-0508-7577; KPM, 0000-0001-7809-5503; CES, 0000-0001-5715-1692

Insect migration is globally ubiquitous and can involve continental-scale

movements and complex life histories. Apart from select species of

migratory moths and butterflies, little is known about the structure of the

annual cycle for migratory insects. Using stable-hydrogen isotope analysis

of 852 wing samples from eight countries spanning 140 years, combined

with 21 years of citizen science data, we determined the full annual cycle

of a large migratory dragonfly, the common green darner (Anax junius). We demonstrate that darners undertake complex long-distance annual

migrations governed largely by temperature that involve at least three gen-

erations. In spring, the first generation makes a long-distance northbound

movement (further than 650 km) from southern to northern range limits,

lays eggs and dies. A second generation emerges and returns south (further

than 680 km), where they lay eggs and die. Finally, a third resident gener-

ation emerges, reproducing locally and giving rise to the cohort that

migrates north the following spring. Since migration timing and nymph

development are highly dependent on temperature, continued climate

change could lead to fundamental changes in the biology for this and similar

migratory insects.

1. Introduction Insects are the most species-rich and abundant group of macroscopic organisms

on the planet, yet our understanding of the biological underpinnings of their

annual cycles and migratory behaviours drastically trails behind that of

vertebrate taxa [1]. Perhaps the best-known insect migrations are the multi-

generational migrations of the monarch butterfly (Danaus plexippus) in eastern North America [2], and the painted lady (Vanessa cardui) in Europe [3], which migrate southbound each autumn from the northern reaches of their distri-

bution into southern North America and Africa, respectively. Monarchs wait

out the boreal winter [4] and the same individuals migrate northbound the fol-

lowing spring into the southern parts of their distribution where they lay eggs,

beginning the next generation that continues the northbound progression of

their lifecycle [2,3,5,6].

Few migratory insects have had their full annual cycles sufficiently

described despite their profound cultural [7], ecological [8] and agricultural

importance [9]. This disparity is owing largely to the difficulties with tracking

small organisms over space and time, as well as their complex life histories that

often require both terrestrial and aquatic habitats [10]. Among the migratory

rsbl.royalsocietypublishing.org Biol.

Lett. 14:

20180741

2

insects, almost all are suspected to employ complex multi- generational migrations (but see [7,8]) similar to the monarch

butterfly and painted lady, where no single individual

completes the entire migratory circuit [2,3,5,6,10,11]. The con-

tinental distances travelled [12,13], small body sizes and

multi-generational aspect of their life histories pose signifi-

cant challenges to understanding the fundamental biology

of the full annual cycles of migratory insects. Adding to

this complexity, movements, timing and life-history strategies

are likely linked to photoperiod, weather, temperature and

other environmental forces [1,3,11,14 – 16].

The common green darner (Anax junius) is a ubiquitous, charismatic migratory dragonfly found in North America.

Its complex life-history includes both resident and migratory

populations. The aquatic nymphs likely give rise to resident

and migratory phenotypes because they exhibit two distinct

developmental strategies [14]. Nymphs either undergo

rapid development and emerge within a single season or

develop slowly, enter diapause and emerge the following

summer [14]. The developmental strategy is governed by the

photoperiod experienced by the egg or early instars [17] and

temperature [14]. Therefore, the seasonal timing of oviposition

likely determines whether nymphs develop, emerge within a

single growing season and migrate, or enter diapause.

Darner migration is considered an adaptive strategy

because it spreads reproductive effort across multiple,

widely separated water bodies and ameliorates the risk of

predation, competition and drought [17,18]. Migratory move-

ments are commonly observed in autumn and have been

documented using direct [12] and indirect techniques [19].

Northbound migration in spring is suspected and supported

by anecdotal observations but is not well documented [20].

Although it is known that common green darners migrate,

no study to date has linked multiple populations over time

and space to better understand the phenology of the annual

cycle and the number of generations involved.

2. Methods and materials We combined observational and stable-hydrogen isotope

data to identify the timing of migration and emergence ori-

gins of darners. Two publicly available datasets were

assembled to refine emergence assignments using stable iso-

topes. First, we acquired citizen science data on the first

observed migratory adult common green darners (first

flight dates) to determine the minimum temperature necess-

ary for darners to migrate (electronic supplementary

material, table S1). Daily mean temperature of the first

flight date was 9.178C, similar to the development threshold (8.7 + 0.18C) reported by Trottier [14]. In addition to deter- mining the temperature threshold for migratory flights, we

used occurrence records submitted by citizen scientists and

museum collections to generate a habitat suitability surface

that we included as prior information in the probabilistic

isotope-assignments [21] (see §2a).

(a) Emergence assignments We obtained contemporary and museum specimens from 681

locations from eight countries ranging from the tropics to the

northern temperate zone (17.988 N to 51.148 N) that we used to describe migration phenology and the number of gener-

ations within the annual cycle for the common green

darner. Specimens, which included exuviae, tenerals and

mature adults, were collected in every calendar month span-

ning over 100 years (electronic supplementary material, table

S2). Isotope analyses were performed on a small distal seg-

ment of the wing to determine where the captured

individuals emerged (see electronic supplementary material).

(b) Migratory distance Determining the exact emergence origin from stable isotopes

is not possible given the uncertainty in geographical assign-

ment and lack of longitudinal variation in the underlying

isoscape. Therefore, we used a conservative distance estima-

tor as an index of migration distance. We determined the

minimum distance (kilometres) between the ‘likely’ emer-

gence origin surface and the location where the specimen

was collected (i.e. shortest possible distance travelled).

Migration distances were calculated using the gDistance

function in the rgeos [22] package in program R [23].

(c) Annual cycle We summarized the emergence origins within the hardiness

zones (see electronic supplementary material, delineating

populations) for each calendar month to describe the pheno-

logy of the annual cycle. Isotope-based emergence origins

provide spatio-temporal information of adults but the

phenology of nymph development and emergence timing is

also needed to fully understand the annual cycle of darners.

3. Results We combined data on stable-hydrogen isotopes of 852

common green darner museum and contemporary specimens

spanning 140 years (1874 – 2013) from eight countries with

citizen science observations to determine the emergence

origin, continental migrations and annual cycle of the

common green darner. Using citizen science observations,

we found that northbound darner migration closely tracks a

thermocline of 9.178C (figure 1). We then used this tempera- ture-dependent flight phenology to refine stable-isotope

assignments, and found that adults originating from the

southern range margin migrate a minimum of 659.46 + 49.11 km to the northern range edge in March and April

(n ¼ 50; electronic supplementary material, figure S1a). Pre- vious estimates of darner daily migration rates indicate this

journey would take 55.42 + 4.13 days to complete [12]. There- fore, the first migratory individuals arriving in the north

likely emerge in the south in January and February. However,

prevailing wind patterns in spring favour northbound

migration, potentially allowing individuals to migrate faster

[24]. If darners use high-altitude migratory flights to take

advantage of fast-moving winds, the distances reported

above could be covered more rapidly [11,13,25].

Combining stable-isotope assignments and temperature-

related nymph development [14], we found the annual

cycle of common green darners is comprised of at least

three generations (figure 2). The first generation emerges

between January and May in the southern part of the distri-

bution and migrates north into the northern reaches of the

distribution. The first generation is complete by August, the

last month when darners with southern stable isotopic ori-

gins are captured in the north (figure 2, August). The first

60

34

GA

GA

GA

GA

ME

VT

CT

NC

NC

NC OK MD

NJ

WV

NY

OH

IN

36 38 40 42

latitude of 9°C

44 46 48

80

fi rs

t ob

se rv

ed f

li gh

t (o

rd in

al d

ay )

100

120

Figure 1. First flight date of adult common green darner and the latitude of the 9.178C thermocline. The mean latitude of the thermocline was extracted along the observation longitude from 1994 to 2014. Error bars represent the standard error in the latitude of the thermocline on the first flight date. Data represented in figure 1 can be found in electronic supplementary material, table S1. Graphic was drawn by Matthew Dodder (https://neornithes.wordpress. com) and used with permission. (Online version in colour.)

rsbl.royalsocietypublishing.org Biol.

Lett. 14:

20180741

3

cohort of the second generation—young from the previous

year that overwinter as late-stage instars and eclose between

May and July [14]—begin flying in May with 4% (2 of 53;

figure 2, May) of individuals assigned to where they were

captured. Spatio-temporal overlap exists between the first

and second generations as northbound migration continues

through July. The number of locally emerged darners in the

north increases throughout the season, peaking in September

when offspring of the first generation in April and early May

eclose (May: 4%, 2 of 53; June: 22%, 16 of 74; July: 29%, 15 of

52; August: 50%, 5 of 10; September: 86%, 6 of 7; October:

67%, 2 of 3; electronic supplementary material, figure S3).

This latter cohort of the second generation then migrates

south.

Second generation darners with northern isotopic signa-

tures are first detected in the south during August and

September (figure 2, September), suggesting southbound

migration begins in July and continues into October [12].

These southbound migrants are composed of darners that

overwinter as nymphs and emerge in June and July and dar-

ners who develop within a single growing season that eclose

in September. Southbound migration to the southern range

edge peaks in September and October, with 64% (7 of 11)

of green darners captured at least 680.97 + 178.84 km from where they emerge in the north (electronic supplementary

material, figure S1b). The second generation composed of

individuals that emerge in the northern portion of the distri-

bution ends in late October (figure 2, October). After this

time, no individuals with northern isotopic signatures were

captured in the south (figure 2, November), suggesting they

die shortly after they arrive in the south. However, south-

bound migrants from mid-latitudes were captured in

November (43%, 3 of 7; figure 2, November) and December

(57%, 4 of 7), emerging a minimum of 226.13 + 108.50 and 232.03 + 93.28 km away from where they were captured, respectively. A third, non-migratory generation occurs in the

south between November and January (figure 2, November –

January). This generation must be resident because adults

that emerge in November likely die before April and May

when the first migratory individuals are captured in the

north, based on the adult longevity reported for the species

(five to seven weeks [16]). Thus, third generation adults that

emerge in November likely give rise to the first generation of

migratory individuals that are captured in the north. Stable

isotopes, however, are unable to differentiate between the

first and proposed third generation.

4. Discussion The annual cycle of green darners is composed of at least two

migratory and one non-migratory generations. In spring, the

first generation makes a long-distance northbound movement

(659.46 + 49.11 km) from southern to northern range limits, lays eggs and dies. A second generation emerges and returns

south (680.97 + 178.84 km), where they lay eggs and die. Finally, a third resident generation emerges, reproducing

locally and giving rise to the cohort that migrates north the

following spring.

During spring, individuals with southern emergence ori-

gins arrive at northern locations prior to those with origins

from mid-latitudes (electronic supplementary material,

figure S2). A similar pattern is observed in autumn, where

individuals with northern origins are captured in the

southern portion of their distribution before green darners

that emerge at mid-latitudes. This migratory phenomenon

could arise from several alternative life-history strategies.

First, such patterns could result if individuals spread repro-

ductive risk across multiple, widely separated water bodies

and reproduce during their migratory journey north [18].

Their offspring could take advantage of warm surface

waters, mature rapidly within the same growing season

and migrate northbound (electronic supplementary material,

figure S4). Such a strategy is akin to the multi-generational

migration of the monarch and painted lady butterflies [1,2],

however, the longevity and strong direct flights of darners

allow populations to be connected over larger spatial scales

compared to the northbound multi-generational progression

of migratory butterflies. The observed pattern of shortened

migratory distances as the season progresses also could

result if individuals reproduce during southbound migration.

If darners reproduce during the journey south, their eggs

would hatch but the nymphs would enter a state of diapause

as temperatures drop, emerge and migrate north the follow-

ing spring when suitable thermal conditions allow [14,18].

A combination of these strategies is likely. Plasticity in

nymph development further complicates discerning between

the two strategies. Eggs laid during migratory periods could

either (i) undergo rapid development after hatching and

emerge within the same season or (ii) hatch, enter diapause

as nymphs and emerge the following spring, depending on

the photoperiod experienced by the egg or early instar [17].

Stable-hydrogen isotope data cannot differentiate between

these competing hypotheses. Regardless of the strategy, the

timing of migratory movements allows for spatial and tem-

poral overlap of both migratory and non-migratory

populations, explaining the lack of genetic structure in this

species [18,26].

June

M ay

A pr

M ar

Feb

Jan Dec

No v

O ct

S ep

Aug

July migr

ant and local

N or

th w

ar d

m ig

ra tio

n

non-migratory

Southw ard m

igration

generation 3

generation 2 ge

ne ra

tio n

1

Figure 2. The full annual cycle of the common green darner is comprised of three generations. The first generation is migratory, emerges in the south between February and May and dies in the north. The second generation emerges in the north, migrates south and dies. The third generation, offspring of migratory individuals in the fall, is non-migratory and emerges in the south in November. The emergence origins of adults (colour ramp; grey ¼ zero, red ¼ many) captured at the sampling locations (black dots) in each month are shown around the perimeter. Graphic was drawn by Matthew Dodder (https://neornithes.wordpress.com) and used with permission.

rsbl.royalsocietypublishing.org Biol.

Lett. 14:

20180741

4

Common green darner migration appears to be governed

in part by temperature. Northbound adult migration during

spring coincides with a daily mean thermocline of approxi-

mately 98C (9.178C) and the last observed adults in autumn coincide with a similar thermocline (9.548C). These findings are consistent with the short migratory movements of

tagged individuals following synoptic cold fronts [12]. In

addition, the 98C thermocline observed here is similar to the development threshold of the final instar (8.7 + 0.18C) [14]. Taken together, these findings suggest that temperature

is a key limiting factor to the northbound migration of

dragonflies in spring, the emergence of nymphs and south-

bound migration in autumn. Given the critical role of

temperature, a warming climate could induce earlier

spring flights, trigger later autumn flights and potentially

shorten migratory distances as the 98C thermocline moves northbound, allowing adults to reside at higher latitudes

during the winter months. Further research is needed to

determine how a changing climate may alter the migratory

schedules and annual cycle of dragonflies and other

migratory insects.

Data accessibility. Museum holdings and specimen numbers and data are available from the Dryad Digital Repository: http://dx.doi.org/10. 5061/dryad.17hf071 [27]. R code used to analyse data can be found at https://github.com/MTHallworth/Hallworth_et_al_Anaxjunius.

Authors’ contributions. M.T.H., P.P.M., K.P.M. and C.E.S. conceived and designed the study. K.P.M. and S.Z. coordinated and collected samples from field and museum specimens. M.T.H. and C.E.S. ana- lysed the data. M.T.H., P.P.M. and C.E.S. wrote the manuscript. All authors read, edited, approved the final version of the manuscript and agree to be held accountable for the content therein.

Competing interests. The authors declare no competing interests exist.

Funding. Major funding was provided by the U.S. Forest Service Inter- national Programs and the Xerces Society for Invertebrate Conservation.

Acknowledgements. This study was part of the Migratory Dragonfly Part- nership coordinated by The Xerces Society for Invertebrate

rsb

5

Conservation. Michael May, Desiree Narango and Dara Satterfield provided comments on a previous version of the manuscript. We thank Desiree Narango and Tim Guida for preparing samples, and

citizen scientists who contributed observations and samples. We thank the many natural history, public and private collections that allowed us to sample from priceless specimens.

l.royalsociety

References

publishing.org Biol.

Lett. 14:

20180741

1. Chapman JW, Reynolds DR, Wilson K. 2015 Long- range seasonal migration in insects: mechanisms, evolutionary drivers and ecological consequences. Ecol. Lett. 18, 287 – 302. (doi:10.1111/ele.12407)

2. Flockhart DTT, Wassenaar LI, Martin TG, Hobson KA, Wunder MB, Norris DR. 2013 Tracking multi- generational colonization of the breeding grounds by monarch butterflies in eastern North America. Proc. R. Soc. B 280, 20131087. (doi:10.1098/rspb. 2013.1087)

3. Stefanescu C et al. 2013 Multi-generational long- distance migration of insects: studying the painted lady butterfly in the Western Palaearctic. Ecography 36, 474 – 486. (doi:10.1111/j.1600-0587.2012.07738.x)

4. Wassenaar LI, Hobson KA. 1998 Natal origins of migratory monarch butterflies at wintering colonies in Mexico: new isotopic evidence. Proc. Natl Acad. Sci. USA 95, 15 436 – 15 439. (doi:10.1073/pnas.95. 26.15436)

5. Miller NG, Wassenaar LI, Hobson KA, Norris DR. 2012 Migratory connectivity of the monarch butterfly (Danaus plexippus): patterns of spring re- colonization in Eastern North America. PLoS ONE 7, e31891. (doi:10.1371/journal.pone.0031891)

6. Talavera G, Bataille C, Benyamini D, Gascoigne-Pees M, Vila R. 2018 Round-trip across the Sahara: Afrotropical painted lady butterflies recolonize the Mediterranean in early spring. Biol. Lett. 14, 20180274. (doi:10.1098/rsbl.2018.0274)

7. Common IFB. 1954 A study of the ecology of the adult bogong moth, Agrotis infusa (Boisd) (Lepidoptera: Noctuidae), with special reference to its behaviour during migration and aestivation. Aust. J. Zool. 2, 223 – 263. (doi:10.1071/zo9540223)

8. Warrant E, Frost B, Green K, Mouritsen H, Dreyer D, Adden A, Brauburger K, Heinze S. 2016 The Australian bogong moth Agrotis infusa: a long- distance nocturnal navigator. Front. Behav. Neurosci. 10, 77. (doi:10.3389/fnbeh.2016.00077)

9. Nagoshi RN, Meagher RL, Hay-Roe M. 2012 Inferring the annual migration patterns of fall armyworm (Lepidoptera: Noctuidae) in the United

States from mitochondrial haplotypes. Ecol. Evol. 2, 1458 – 1467. (doi:10.1002/ece3.268)

10. Alerstam T, Chapman JW, Backman J, Smith AD, Karlsson H, Nilsson C, Reynolds DR, Klaassen RHG, Hill JK. 2011 Convergent patterns of long-distance nocturnal migration in noctuid moths and passerine birds. Proc. R. Soc. B 278, 3074 – 3080. (doi:10. 1098/rspb.2011.0058)

11. Anderson RC. 2009 Do dragonflies migrate across the western Indian Ocean? J. Trop. Ecol. 25, 347 – 358. (doi:10.1017/S0266467409006087)

12. Wikelski M, Moskowitz D, Adelman JS, Cochran J, Wilcove DS, May ML. 2006 Simple rules guide dragonfly migration. Biol. Lett. 2, 325 – 329. (doi:10. 1098/rsbl.2006.0487)

13. Hobson KA, Anderson RC, Soto DX, Wassenaar LI. 2012 Isotopic evidence that dragonflies (Pantala flavescens) migrating through the Maldives come from the Northern Indian subcontinent. PLoS ONE 7, e52594. (doi:10.1371/journal.pone.0052594)

14. Trottier R. 1971 Effect of temperature on the life- cycle of Anax junius (Odonata: Aeshnidae) in Canada. Can. Entomol. 103, 1671 – 1683. (doi:10. 4039/Ent1031671-12)

15. Corbet PS. 2014 Dragonflies: Behaviour and Ecology of Odonata. See http://www.nhbs.com/dragonflies- behaviour-and-ecology-of-odonata-book (accessed on 13 September 2017).

16. May ML. 2013 A critical overview of progress in studies of migration of dragonflies (Odonata: Anisoptera), with emphasis on North America. J. Insect Conserv. 17, 1 – 15. (doi:10.1007/s10841- 012-9540-x)

17. Matthews JH. 2007 Research in motion: patterns of large-scale migration in dragonflies and birds, pp. 62 – 91. PhD Dissertation, University of Texas.

18. May ML, Gregoire JA, Gregoire SM, Lubertazzi MA, Matthews JH. 2017 Emergence phenology, uncertainty, and the evolution of migratory behavior in Anax junius (Odonata: Aeshnidae). PLoS ONE 12, e0183508. (doi:10.1371/journal.pone. 0183508)

19. Hobson KA, Soto DX, Paulson DR, Wassenaar LI, Matthews JH. 2012 A dragonfly (d2H) isoscape for North America: a new tool for determining natal origins of migratory aquatic emergent insects. Methods Ecol. Evol. 3, 766 – 772. (doi:10.1111/j. 2041-210X.2012.00202.x)

20. Russell RW, May ML, Soltesz KL, Fitzpatrick JW. 1998 Massive swarm migrations of dragonflies (Odonata) in Eastern North America. Am. Midl. Nat. 140, 325 – 342. (doi:10.1674/0003- 0031(1998)140[0325:MSMODO]2.0.CO;2)

21. Andrew Royle J, Rubenstein DR. 2004 The role of species abundance in determining breeding origins of migratory birds with stable isotopes. Ecol. Appl. 14, 1780 – 1788. (doi:10.1890/04-0175)

22. Bivand R, Rundel C, Pebesma E, Stuetz R, Hufthammer KO, Giraudoux P, Davis M, Santilli S. 2018 rgeos: Interface to Geometry Engine - Open Source (GEOS). See https://CRAN.R-project.org/ package=rgeos.

23. R Core Team. 2016 R: A language and environment for statistical computing. Vienna, Austria: R Foundation for Statistical Computing. See https:// www.R-project.org/.

24. Kranstauber B, Weinzierl R, Wikelski M, Safi K. 2015 Global aerial flyways allow efficient travelling. Ecol. Lett. 18, 1338 – 1345. (doi:10.1111/ele.12528)

25. Feng H-Q, Wu K-M, Ni Y-X, Cheng D-F, Guo Y-Y. 2006 Nocturnal migration of dragonflies over the Bohai Sea in northern China. Ecol. Entomol. 31, 511 – 520. (doi:10.1111/j.1365-2311.2006. 00813.x)

26. Freeland JR, May M, Lodge R, Conrad KF. 2003 Genetic diversity and widespread haplotypes in a migratory dragonfly, the common green darner Anax junius. Ecol. Entomol. 28, 413 – 421. (doi:10. 1046/j.1365-2311.2003.00521.x)

27. Hallworth MT, Marra PP, McFarland KP, Zahendra S, Studds CE. 2018 Data from: Tracking dragons: stable isotopes reveal the annual cycle of a long-distance migratory insect. Dryad Digital Repository. (doi:10. 5061/dryad.17hf071)