Faculty of Science
IS ISLAND LIFE TURNING BIRDS INTO GIANTS? AN ASSESSMENT OF
FOSTER’S RULE ACROSS ALL BIRDS
2021 | AARON LESLIE VEALE

B.Sc. Honours thesis – Biology

IS ISLAND LIFE TURNING BIRDS INTO GIANTS? AN ASSESSMENT OF FOSTER’S
RULE ACROSS ALL BIRDS.
by

AARON LESLIE VEALE

A THESIS SUBMITTED IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF

BACHELOR OF SCIENCE (HONS.)
in the
DEPARTMENT OF BIOLOGICAL SCIENCES
(General Biology)

This thesis has been accepted as conforming to the required standards by:
Matthew Reudink (Ph.D.), Thesis Supervisor, Dept. Biological Sciences
Nancy J. Flood (Ph.D.), Co-supervisor, Dept. Biological Sciences
Eric Bottos (Ph.D.), External examiner, Dept. Biological Sciences

Dated this 26th day of April, 2021, in Kamloops, British Columbia, Canada

© Aaron Leslie Veale, 2021

ABSTRACT
The differences in the body sizes observed in island birds versus their closest mainland
relatives have puzzled biologists for decades. First noted in other vertebrate groups by Dr. Bristol
Foster, the general trend is usually summarized as small mainland species evolving larger bodies
on islands, while typically large mainland species shift towards dwarfism. With many examples in
both living and extinct fauna, the overall phenomenon became known as Foster’s rule. Herein, we
examine Foster’s rule as it applies to class Aves (the group that contains all modern birds). We
analyzed the body mass (n=9,316), body length (n=7,260) and wingspan (n=708) of avian species
from around the world. To account for phylogeny, we employed the use of multiple independent
phylogenetic generalized least squares analyses (PGLS). We first analyzed class Aves as a whole
and then looked at each individual order separately. We found support for Foster’s rule in class
Aves overall, island species heavier and longer than mainland species. Wingspan did not vary
between island and mainland species for the class as a whole. Looking at each order, we found that
body size varied between islands and mainland in Anseriformes, Accipitriformes, Charadriiformes,
Galliformes, Piciformes, Pelecaniformes, and Strigiformes. Whereas body mass in Galliformes and
Piciformes increased on islands, Anseriformes decreased. Similarly, body length in Piciformes and
Pelecaniformes increased, while Anseriformes and Strigiformes length decreased. Wingspan
increased in both Accipitriformes and Charadriiformes. Although we found support for Foster’s
rule, the direction of effects varied by order, and by body size metric, indicating that the
biogeographical pattern is not generalized across class Aves as has been previously suggested, and
underscores the idea that body size evolution may be driven by a combination of stochastic and
deterministic forces specific to bird orders.
Thesis Supervisor: Associate Professor Matthew Reudink
ii

ACKNOWLEDGMENTS

A big thank you to both Dr. Matt Reudink and Sean Mahoney for all of their much-appreciated
time and effort in the development of this project and its realization. Their feedback has helped in
every step of the way. Thank you to the Thompson Rivers University Undergraduate Research
Experience Award Program (UREAP) for funding my thesis. I would also like to thank Claudie
Pageau for her assistance in the early stages and as an academic role model. Finally, I would like
to thank Marcio Argollo de Menezes, Nathan Smith, and the BEAC lab for their assistance with
the large amount of data collection required for this project.

iii

TABLE OF CONTENTS

TITLE PAGE…………………………………………………………………………………...i
ABSTRACT…………………………………………………………………………………….ii
ACKNOWLEDGEMENTS……………………………………………………………….…....iii
TABLE OF CONTENTS AND FIGURE/TABLE LIST…………………………………...…...iv
FIGURE/TABLE LIST…………………………………………………….…………….……..v
INTRODUCTION ……………………………………….…………………………….……....1
METHODS……………………………………………………………………………….…....4
Data collection…………………………………………………………………………….......4
Phylogeny…………………………………………………………………………………......6
PGLS Analyses………………………………………………………………………………..6
RESULTS…………………………………………………………………………………..….7
Body Mass…………………………….…………………………………………………..…..7
Body Length……………………………..…………………………………………………...11
Wingspan………………………………………..…………………………………….……..14
DISCUSSION………………………………………………………………………....……...15
LITERATURE CITED………………………………………………………………………..19

iv

FIGURE LIST

Figure 1…………………………………………………………………....................8
Figure 2…………………………………………………………………....................9
Figure 3…………………………………………………………………………..…12
Figure 4…………………………………………………………………..................14

TABLE LIST
Table 1………………………………………………………………………………5
Table 2…………………………………………………………………..................10
Table 3…………………………………………………………………..................13
Table 4…………………………………………………………………..................15

v

INTRODUCTION

An intriguing biological change occurs in island populations in which the body sizes of
animals tend to differ markedly from their mainland counterparts. Dr. J. Bristol Foster first noted
that island mammals often exhibit differing body sizes compared to their closest mainland relatives.
The Island rule, also known as Foster’s rule, describes the phenomenon of large-bodied mainland
species becoming smaller (insular dwarfism) and small-bodied mainland species demonstrating the
reverse (insular gigantism) (Van Valen 1973). Beyond its initial application to terrestrial
mammalian clades, the influence of island biogeography on body size has been expanded to include
other vertebrate groups like bats (Krzanowski 1967), squamates (snakes and lizards) (Itescu et al.
2018), and even non-avian dinosaurs (Benton et al., 2010). Foster’s rule has also been tested in
birds multiple times, yet only with a limited subset of species (Grant 1965; Grant 1966; Case 1978;
Gaston and Blackburn 1995; Clegg and Owens 2002). To our knowledge, no comprehensive study
of Foster’s rule for the entire class Aves has been done to date despite the data being available (but
see Olson et al. 2009 assessment of Bergmann’s rule in birds).
The causes of insular dwarfism or gigantism are controversial and vary depending on the
species and the selective pressures they face. These include food availability, energy use, predation
risk, intraspecific and interspecific competition, and heat regulation (Clegg and Owens 2002;
Lomolino 2005). For bird species, food availability has been proposed as the driving factor behind
changes in body morphology when moving from mainland environments to islands (Keast 1968;
Mayr 1963; Abbott 1980; Grant 1998; Blondel 2000). Large-bodied species are thought to have
reduced fitness on islands where they encounter increased population densities and intraspecific
competition, leading to selection favouring decreases in overall body size as smaller individuals.
require less food and therefore less energy to survive and reproduce (Lomolino 2005). Island
1

environments are also typically species poor when compared to mainland ecosystems and often
lack the mammalian predators and interspecific resource competitors found on the mainland
(McNab 2002; Lomolino 2005). While being the largest member in a group of conspecifics may
serve as protection against predation in mainland habitats, islands without large mammalian
predators make this adaptation an energetic liability (McNab 1994; Lomolino 2005). With reduced
predation pressure, we would expect to see the largest directional change in body mass occurring
in species with the largest bodies, and therefore the highest energetic demands overall (Lomolino
2005). In addition to energy requirements, it should also be noted that Clegg and Owens (2002)
observed that body mass decline in larger birds intensified as they neared the equator. They
suggested that this was possibly an adaptation to increase heat dissipation but could not be specific
as to the exact mechanism.
McNab (1994, 2002) suggested that when small bodied species experience reduced
predation pressure on islands, they become larger and occupy similar niches to intermediate sized
herbivores on the mainland. In times of resource scarcity, individuals with larger body sizes can
dominate food sources (Lomolino 2005). As competition intensifies within species groups and
between similar species, dietary divergence through niche expansion and speciation may result;
this has been hypothesized to be the main evolutionary force behind insular gigantism in vertebrates
(Runemark et al. 2015). For species capable of high dispersal (such as birds), islands represent an
opportunity to fill terrestrial niches that are already occupied in older, mainland environments.
Consistent with Foster’s rule, body sizes in birds appear to increase following island colonization,
suggesting that larger body size is, at least in part, adaptive in island conditions (Owen and Clegg
2002; Oslon et al. 2009). Boyer and Jetz (2010) examined the predictors of body size in insular

2

bird species and found that body size is positively correlated with island size, i.e., larger birds are
found on larger islands.
On a macroevolutionary scale, body size may also, in part, be due to ancestry. In relation
to Bergmann’s rule, Olson et al. (2009) suggested that phylogenetic history may constrain body
size. They note that all penguins are relatively large in size and occupy similar niches as a result of
shared ancestry. We thus need to be cautious in assuming that ecological factors are the major
influences on body size. Observed differences between island and mainland species may be deeply
rooted in phylogeny rather than ecology. Therefore, evolutionary relationships must be accounted
for when testing broadly related taxa such as class Aves. One such approach is to use a phylogenetic
generalized least squares analysis, as outlined later on in our methods.
The generalizability of Foster’s rule in birds has previously been called into question and
thus been the focus of multiple studies (e.g., Grant 1965; Grant 1966; Case 1978; Gaston and
Blackburn 1995; Clegg and Owens 2002; Olson et al. 2009; Wright and Steadman 2012). Early
(e.g., Grant 1966; Carlquist 1974; Case 1978; Gaston and Blackburn 1995) failed to find support
for larger body sizes in birds using mass as a metric, instead noting potential trends towards
increased tarsal and beak length compared to mainland species (Grant 1965). In contrast, more
recent studies have found explicit support for Foster’s rule in birds based on body mass (Clegg and
Owens 2002; Olson et al. 2009). One of the first attempts at a broad-scale study of the island rule
in birds was conducted by Clegg and Owens (2002) and included 110 pairs of island species and
their closest-related mainland counterparts. The most inclusive study to date was conducted by
Olson et al. in 2009), in which sthey analyzed the body masses of 8270 species from across a wide
geographic range. Although they did include an island/mainland body mass analysis, the primary
focus of Olson et al. (2009) was a test of Bergmann’s rule (in which body size increases with
3

higher latitudes) in birds. They found that median body masses in island birds were higher than
expected by latitude alone (Olson et al. 2009). Our study expands Olson et al. (2009) by including
more than one body size variables and asking whether Foster’s rule is generalizable l avian orders.
Here, we tested the application of the Foster’s rule to class Aves (all modern birds),
specifically using body mass, body length, and wingspan as metrics of body size in the comparison
of island versus mainland endemic taxa. We hypothesize that birds will follow the island rule in
general terms as outlined by Van Valen (1973): small-bodied taxa will be typical of mainland
environments and large-bodied taxa will be endemic to islands, as one would expect if niche
expansion and dietary divergence on islands drives gigantism in the absence of predators and
interspecific competition, as previously hypothesized.

METHODS

Data Collection
To test for differences in body size between island and mainland birds, we collected body mass,
length, and wingspan data from Wilman et al. (2014) and from the Handbook of Birds of the World
(2020). We categorized species as either mainland or island, based on the range maps provided by
the IUCN Redlist (2020). Specifically, we classified each species as being island (n=2,056),
mainland (n=7,260) intermediate between these (n=544), or oceanic (n= 133). We described
species as discretely island or mainland if at least 70% of its range was occupied by the respective
land type. Species with more ambiguous ranges of 50-60% were termed intermediate. Lastly,
species with vast ocean ranges (roughly 70% open water, often connecting islands to mainland

4

shores) were considered as oceanic. We included Greenland as our upper limit to island size
(2,130,800 km2). Continental Australia (not including Tasmania) was considered as a mainland
environment. Birds noted as being intermediate or oceanic were subsequently excluded from the
analyses. Only extant resident and breeding ranges were used except in unique cases, such as
species that are only recently extinct in the wild (e.g., Corvus hawaiiensis) or are believed to be
extinct but have not been confirmed to be so (e.g., Campephilus principalis); in these cases, the
historical ranges were used. We excluded introduced, migratory, and non-breeding ranges from
consideration. General body size in an order was determined using methods similar to those of
Clegg and Owens (2002) where large-bodied birds were those above the mainland mean and smallbodied were below. We classified mainland taxa as either being large or small depending on
whether they were larger or smaller than the calculated median body mass for mainland birds as a
whole (Table 1). Finally, in all figures we present logarithmic transformed data to help with
interpretation.

5

Table 1. Mainland median body masses (g) depicting which orders were considered large-bodied
and which were small relative to overall median value.
order

mainland
med. (g)

All orders

32.30

Accipitriformes

body
size

order
continued

mainland
med. (g)

body
size

-

Cuculiformes

80.70

large

754.37

large

Falconiformes

261.00

large

Anseriformes

922.19

large

Galliformes

549.39

large

Apodiformes

5.20

small

Gruiformes

159.09

large

Bucerotiformes

292.00

large

Passeriformes

20.54

small

Caprimulgoformes

57.84

large

Pelecaniformes 846.00

large

Charadriiformes

158.00

large

Piciformes

54.51

large

Columbiformes

169.00

large

Psittaciformes

122.43

large

Coraciiformes

52.96

large

Strigiformes

191.00

large

Phylogeny
To generate our core phylogeny, we downloaded 1000 trees from BirdTree.org (Jetz et al.
2012) in the “HackettStage2_1001_2000” subset. We then used the package ape (Paradis and
Schliep 2018) in R (Rstudio Team 2016) to read the trees. We created maximum clade credibility
trees with 1% burn-in state and mean node height based on the remaining species within each
dataset using TreeAnnotator V.1.10.4 (Rambaut and Drummond 2018).
PGLS Analyses
To test the evolutionary associations between body size and island occupancy, we generated
three phylogenetic generalized least squares models (PGLS) for the body mass (n=9,316), body
length (n=7,584), and wingspan (n=708) datasets. PGLS analyses were generated using the R
6

packages ape (Paradis and Schliep 2018) and nlme (Pinheiro, Bates, DebRoy, Sarkar, and R core
team 2019). We first analyzed for differences in body size using all birds; in each model, land
status (island or mainland) served as a main effect, with body size measures serving as response
variables. All models included Brownian correlation and maximum likelihood methods. We used
the phytools package (Revel 2012) to isolate species by order and remove any species with
insufficient data. Only orders containing twenty or more species with at least five island and five
mainland species were included in the within-order analysis. Once these restrictions were applied,
our analysis was reduced to 19 orders for body mass, 18 for body length, and 3 for wingspan.
Because within orders, species vary in their biology and ecology, we treated each order as a
separate hypothesis, so we did not correct the P value for multiple hypothesis testing (i.e., setting
the stringency of our alpha from 0.05 to 0.0025 to avoid increased chances of making a Type II
error). In total, we performed 43 separate PGLS analyses.

RESULTS
Body Mass
Overall, birds on islands were larger, with a median body mass 16.8g heavier than their
mainland counterparts (Table 2, Figure 1 (A), df = 9314, Z = 1109.14, p < 0.01), a finding
consistent with Foster’s rule. For the order analyses, as predicted, the characteristically largerbodied Anseriformes (ducks, geese, swans and kin) showed a decrease in median body mass of ~
91.2g on islands indicating a shift to dwarfism (Table 2, Figure 2 (B; E), df = 148, Z = 179.01, p <
0.01). In contrast, Galliformes (chickens, pheasants and kin) were ~ 65g heavier on islands versus
mainland (Table 2, Figure 2 (C; F), df = 270, Z = 16.90, p = 0.01). Accipitriformes (hawks, eagles,

7

vultures, and kites) also trended towards smaller body sizes on islands; however, this difference
was marginal with island species being lighter by 179.4g (Table 2, df = 134, Z = 12.68, p = 0.10).
On the other hand, the smaller Piciformes (woodpeckers and kin) were ~ 27g heavier on islands
compared to mainland (Table 2, Figure 2 (A; E), df = 378, Z = 17.62, p = 0.01). We found no
differences in body size between island and mainland species in the remaining orders (15/19).

Figure 1. Boxplots showing differences in the log10 median body mass (A) and the log10 body length
(B) between island and mainland species for class Aves overall. Both body mass and body length
had a significant increase in island birds versus their mainland counterparts.

8

Figure 2. Top: (A) A phylogeny of 380 species (39 island and 341 mainland) examined belonging
to the order Piciformes. (B) A phylogeny of the 150 species (16 island and 134 mainland) examined
belonging to the order Anseriformes. (C) A phylogeny of the 272 species (49 island and 223
mainland) examined belonging to the order Galliformes. Island species are visually represented
by black, while red corresponds to mainland. Bar length denotes mean body mass (g) for the
species. Bottom: Boxplots showing significant differences in the log10 median body mass between
island and mainland species of (D) Piciformes versus (E) Anseriformes and (F) Galliformes.

9

Table 2. PGLS results for all bird orders together, then separated by order. Median body mass (g)
values for both island and mainland groups provided. Statistically significant results have been
bolded. Marginally significant results have been italicized.
order

n
(island/mainland)

df

Z (test stat)

P

island
mainland
median (g) median
(g)

Accipitriformes

236(49/187)

234

12.68

0.10

575.00

754.37

Anseriformes

150(16/134)

148

179.01

<0.01

831.02

922.19

Apodiformes

409(52/357)

407

10.00

0.98

11.90

5.20

Bucerotiformes

60(21/39)

58

11.85

0.18

1086.00

292.00

Caprimulgoformes

102(23/79)

100

10.06

0.80

70.30

57.84

Charadriiformes

345(48/297

343

10.08

0.77

142.50

158.00

Columbiformes

298(172/126)

296

10.08

0.77

205.65

169.00

Coraciiformes

144(63/81)

142

10.04

0.84

60.97

52.96

Cuculiformes

129(54/74)

127

10.23

0.63

160.00

80.70

Falconiformes

61(10/51)

59

10.56

0.46

145.40

261.00

Galliformes

272(49/223)

270

16.90

0.01

614.75

549.39

Gruiformes

149(37/111)

147

11.59

0.21

205.92

159.09

Passeriformes

5701(1187/4514)

5699

10.00

0.97

22.80

20.54

Pelecaniformes

100(7/93)

98

12.25

0.14

802.00

846.00

Piciformes

380(39/341)

378

17.62

0.01

81.91

54.51

Psittaciformes

339(129/210)

337

11.03

0.31

112.00

122.43

Strigiformes

184(61/123)

182

10.12

0.73

151.00

191.00

Suliformes

41(9/32)

39

11.23

0.27

2072.67

1517.55

Trogoniformes

38(9/29)

36

12.10

0.16

72.82

80.29

All orders

9316(2056/7260)

9314

1109.14

<0.01

49.10

32.30

10

Body length
Across the class Aves, species on islands had median body lengths ~2.5cm longer than
birds on the mainland (Table 3, Figure 1(B), df = 7584, Z = 1106.43, p < 0.01), which again is
consistent with Foster’s rule given the small overall median value for mainland birds. Anseriformes
on islands were also smaller, however, this is not reflected in the median lengths overall (both were
53cm) (Table 3, Figure 3 (B; F), df = 125, Z = 147.05, p < 0.01). Here we chose to include the
mean body length and the standard deviation to provide context. Island Piciformes were larger than
their mainland counterparts by roughly 8.3cm (Table 3, Figure 3 (A; E), df = 317, Z = 136.19, p <
0.01). Strigiformes species (owls) are also consistent with Foster’s rule being ~ 0.75cm smaller on
islands (Table 3, Figure 3 (C; G), df = 164, Z = 110.96, p < 0.01). Island Pelecaniformes increased
in size by ~1.5cm (Table 3, Figure (D; H) df = 94, Z = 14.24, p = 0.04). All remaining orders
(14/18) did not meet the threshold of significance.

11

A

B

C
V
V

D

E
V

F
V

G
V

H
V

Figure 3. Top: (A) A phylogeny of 319 species (24 island and 295 mainland) examined belonging
to the order Piciformes. (B) A phylogeny of the 127 species (16 island and 111 mainland) examined
belonging to the order Anseriformes. (C) A phylogeny of the 162 species (57 island and 109
mainland) examined belonging to the order Strigiformes. (D) A phylogeny of the 96 species (7
island and 89 mainland) examined belonging to the order Pelecaniformes. Island species are
visually represented by black, while red corresponds to mainland. Bar length denotes median body
12

length (cm) for the species. Bottom: Boxplots showing significant differences in the log10 median
body length between island and mainland species of (E) Anseriformes versus (F) Piciformes and
(G) Strigiformes (H) Pelecaniformes.

Table 3. PGLS analyses conducted using the body length data set, separated by order, along with
the median value (cm). Significant results have been bolded. * Island and mainland medians were
equal for Anseriformes, but the means with standard deviations were 53.72±10.72 cm for islands
and 58.41±23.12 cm for mainland. *
order

n
(island/mainland)

df

Z (test stat)

P

island
med. (cm)

mainland
med.(cm)

Accipitriformes

212(49/163)

210

12.41

0.12

47.50

52.00

Anseriformes

127(16/111)

125

147.05

<0.01

53.00

53.00

Apodiformes

349(36/313)

347

10.41

0.52

11.75

11.00

Bucerotiformes

41(12/29)

210

12.41

0.12

52.50

50.00

Caprimulgiformes

78(19/59)

76

10.26

0.61

25.00

23.50

Charadriiformes

302(42/260)

299

10.04

0.83

28.75

28.00

Columbiformes

263(149/114)

261

10.48

0.49

33.50

28.31

Coraciiformes

124(48/76)

122

10.04

0.85

23.50

25.25

Cuculiformes

108(43/65)

106

10.03

0.92

43.50

33.00

Falconiformes

59(9/50)

57

10.03

0.86

47.00

35.25

Galliformes

194(34/160)

192

10.14

0.71

34.00

40.50

Gruiformes

100(19/81)

98

10.63

0.43

29.00

28.00

Passeriformes

4519(941/3578)

4517

12.18

0.27

15.50

15.00

Pelecaniformes

96(7/89)

94

14.24

0.04

68.00

67.50

Piciformes

319(24/295)

317

136.19

<0.01

27.75

19.50

Psittaciformes

314(125/189)

312

10.68

0.41

25.00

26.00

Strigiformes

166(57/109)

164

110.96

<0.01

24.00

24.75

Trogoniformes

36(8/28)

34

12.40

0.13

29.75

29.00

All orders

7584(1660/5924)

7582

1106.43

<0.01

20.00

17.50

13

Wingspan
For wingspan, we did not detect a difference between island and mainland species for the
class Aves overall. However, diurnal raptor species (Accipitriformes) on islands had roughly
~27cm shorter median wingspans than those on the mainland (Table 4, df = 201, Z = 19.15, p <
0.01). Similarly, in Charadriiformes (shore birds, gulls, and auks) island species had wingspans
that were ~ 3.5cm shorter than those on the mainland (Table 4, df = 197, Z = 14.24, p = 0.04). The
remaining orders did not exhibit differences in wingspan between island and mainland orders.
A

B

C

D

Figure 4. Mainland bird species are indicated by red bars; island birds indicated by black bars.
(A) A phylogeny of 203 species (49 island and 154 mainland) examined belonging to the order
Accipitriformes. (B) A phylogeny of the 199 species (21 island and 178 mainland) examined
14

belonging to the order Charadriiformes. Island species are visually represented by black, while
red corresponds to mainland. Bar length denotes mean body mass (g) for the species. Boxplots
showing significant differences in the log10 median body mass between island and mainland species
of (C) Accipitriformes versus (D) Charadriiformes.

Table 4. PGLS analyses conducted using the wingspan data set, separated by order, along with
the median value (cm). Significant differences are bolded.
order

Accipitriformes
Anseriformes
Charadriiformes
Falconiformes
All orders

n (island/mainland)

df

Z (test
stat)

P

island
media
n (cm)

mainland
median
(cm)

203(49/154)

201

19.15

<0.01

95.50

122.50

66(5/61)

64

11.71

0.20

84.00

82.50

199(21/178)

197

14.24

0.04

60.50

64.00

57(9/48)

55

10.14

0.71

56.00

76.25

709(98/611)

707

12.15

0.14

75.00

85.00

DISCUSSION

Based on our analyses, we found support for Foster’s rule across class Aves, with an overall
increase in the small mainland median values (as measured by body mass and body length (Figure
1, A and B)) compared to island species; however, this pattern was largely dependent on order. In
orders of larger birds (e.g., Anseriformes and Accipitriformes), island species were usually smaller
than those on the mainland. The general support for Foster’s rule across class Aves, with larger
species more likely to be found on islands, supports the results of previous studies done by Clegg
and Owens (2002), Boyer and Jetz (2010), and Olson et al. (2009). Of the approximately 10,000
bird species, over half belong to the order Passeriformes (perching birds) which have

15

characteristically small body sizes (Clegg and Owens 2002). In pioneering studies, such as Grant’s
1965 analysis, the disproportionately high numbers of small-bodied species in their datasets were
thought to have masked support for Foster’s rule among the class Aves (see Clegg and Owens
2002). No small-bodied groups appeared overrepresented in our dataset compared to what they
should be in the phylogenetic tree created by Jetz et al. (2012) containing all known modern bird
species (at the time of their publication). We addressed overrepresentation by mirroring the
proportions of bird orders as naturally represented in class Aves. For example, of the 9,316 species
tested in our mass dataset, 5,701 were passerines (~61%). The next largest order by species was
another small-bodied taxon, Apodiformes (Hummingbirds, swifts and kin), at 409 (~7%).
Supporting earlier work by Grant (1965), wingspan did not differ between island and mainland
species across class Aves. Only in Accipitriformes and Charadriiformes did we detect differences
in wingspan, with both orders showing a reduction in median wingspan. It should also be noted
that wingspan is typically only used as a metric of body size when mass is unavailable (Hamilton
1961). Additionally, our wingspan dataset was fairly small, including fewer than 10% of all bird
species (708 compared to 9,316 for body mass and 7,584 for body length). As such, we note that
these results do not necessarily have the same statistical power as our other two response variables.
The support for Foster’s rule in all birds creates an apparent contradiction when broken
down by order. Except for taxa listed in the above results (Tables 1,2, and 3) the majority (15/19)
displayed no differences in body mass between island and mainland species—a finding consistent
with Olson et al. (2009). Similarly, we found no differences in body length between island and
mainland species for a majority (14/18) of orders. As noted earlier, body size gradients may be at
least partly adaptive (Olson et al. 2009) and could be a response to processes occurring at the
assemblage level (i.e., inter and intraspecies interactions unique to the animal life of each island

16

probably have a significant role) (Gaston et al. 2008). Bird body size has also been shown to change
after island colonization as a response to dietary expansion, competition, energetic costs, and heat
regulation (Clegg and Owens 2002) but different orders are likely to have varying degrees of
response to at least some of these factors based on differing physiologies. For example, predatory
birds such as Accipitriformes have differences in population densities compared to herbivorous
species (Juanes 1986) and therefore would experience intraspecific competition differently as a
result. With limited prey on species-poor islands, niche expansion may not be a driving force in
this case. It is likely too simplistic to say groups that are ecologically different, such as
Accipitriformes and Anseriformes, are predictable solely because they are large-bodied forms and
should therefore follow Foster’s rule without deviation. Additionally, factors like competition and
niche expansion depend on what species are present on which islands. Therefore, aspects of island
ecology are likely to play significant roles in size trends overall.
An alternate explanation may come down to stochastic events. Lomolino (2005) notes that
relatively poor flyers may become stranded, especially on remote islands, invoking a Darwinian
metaphor of a shipwrecked crew clinging to their wreck. Those that can travel back to their
mainland habitats may do so leaving a smaller population of weak flyers behind. One mechanism
could be through the founder effect, which has been shown to account for morphological and
genotypic variation in some island bird species (Clegg et al. 2002; Estoup and Clegg 2003; Spurgin
et al. 2014). In other words, changes in body size may arise from pure chance related to which
individuals get stranded and the genetic contributions they bring to the population. The influence
of genetic drift should also not be overlooked. Non-adaptive selection may partially explain why
so many orders contain species, genera, and likely even families that exhibit dwarfism or gigantism,
yet fail to demonstrate a clear trend across orders.

17

We have only scratched the surface when it comes to the exploring island biogeography.
Islands are exceptionally diverse, both in terms of geography and ecology. Wright and Steadman
(2012) noted that factors like island size and geographical region can also impact the physical traits
of the birds living on them. A study by Filin and Ziv (2004) appeared to show that the degree of
change in body size from the mainland species was inversely related to the size of the island area.
The complex associations within and among avian taxa are complicated by both living and nonliving components of their island environments, all of which may affect the evolution of bird
morphology. To summarize, our study has found that birds as whole had higher median masses
and body lengths on islands compared to mainland environments indicating a move towards
gigantism. On the level of order, this pattern begins to breakdown for the exception of only a
handful of groups. There was no significant difference in wingspan overall, except for two isolated
orders. Why and how body size changes on islands have been partly addressed, but more work is
required to reveal the nuances and specific mechanisms. When analyzed closely at higher
taxonomic power (i.e., family and genus), the causes of insular dwarfism and gigantism are likely
to be as unique and variable as the bird species they influence.

18

LITERATURE CITED
Abbott I. 1980. Theories dealing with ecology of landbirds on islands. Adv. Ecol. Res. 11, 329371.
Benton M., Csiki Z., Grigorescu D., Redelstorff R., Sander P.M., Stein K., Weishampel D.B.
2010. Dinosaurs and the island rule: The dwarfed dinosaurs from Haţeg Island. 293(3-4), 0–
454. doi:10.1016/j.palaeo.2010.01.026
BirdLife International and Handbook of the Birds of the World. 2018. Bird species distribution
maps of the world. Version 2018.1. Available at http://datazone.birdlife.org/species/requestdis.
Blondel J. 2000. Evolution and ecology of birds on islands: trends and perspectives. Vie Et Millieu
50, 205-220.
Boyer A. and Jetz W. 2010. Biogeography of body size in Pacific island birds. Ecography, 33: 369379. doi:10.1111/j.1600-0587.2010.06315.x
Carlquist, S. 1974. Island biology. New York: Columbia University Press
Case T. 1978. A general explanation for insular body size trends in terrestrial vertebrates. Ecology,
59(1), 1–18. doi:10.2307/1936628
Clegg S., Degnan S., Kikkawa J., Moritz C., Estoup A., and Owens, I. P. F. 2002. Genetic
consequences of sequential founder events by an island-colonizing bird. Proceedings of the
National Academy of Sciences, 99(12), 8127–8132. doi:10.1073/pnas.102583399
Clegg S. and Owens I.P.F. 2002. The “island rule” in birds: medium body size and its ecological
explanation. Proceedings of the Royal Society B: Biological Sciences, 269(1498), 1359–
1365. doi:10.1098/rspb.2002.2024
Estoup A, and Clegg S. 2003. Bayesian inferences on the recent island colonization history by the
bird Zosterops lateralis lateralis. Molecular Ecology, 12, 657–674.
Filin I. and Ziv, Y. 2004. New theory of insular evolution: unifying the loss of dispersability and
body-mass change. Evolutionary Ecology, Res. 6: 115-124.
Foster J. 1964. Evolution
235. doi:10.1038/202234a0

of

mammals

on

islands.

Nature,

202(4929),

234–

Gaston K. and Blackburn T. 1995. Birds, body size and the threat of extinction. Philosophical
Transactions of the Royal Society of London. Series B: Biological Sciences, 347(1320), 205–
212. doi:10.1098/rstb.1995.0022
Grant P. 1965. A systematic study of the terrestrial birds of the Tres Marias islands, Mexico. Yale
Peabody Mus. Postillo 90, 1-106.

19

Grant P. 1966. Further information the relative length of the tarsus in land birds. Yale Peabody
Mus. Postillo 98, 1-13.
Grant P. 1998 Patterns on islands and microevolution. In Evolution on islands (ed. P. R. Grant),
pp. 1-17. Oxford University Press.
Hamilton T. H. 1961. The Adaptive Significances of Intraspecific Trends of Variation in Wing
Length and Body Size Among Bird Species. Evolution, 15(2), 180. doi:10.2307/2406079
Itescu Y., Schwarz R., Donihue C. M., Slavenko A., Roussos S.A., Sagonas K., Valakos E.D.,
Foufopoulos J., Pafilis P., Meiri S. 2018. Inconsistent patterns of body size evolution in cooccurring island reptiles. Global Ecology and Biogeography, (), –. doi:10.1111/geb.12716
IUCN 2020. The IUCN
https://www.iucnredlist.org.

Red

List

of

Threatened

species.

Version

2020-1

Jetz W, Thomas G, Joy J, Hartmann K, and Mooers A. 2012. The global diversity of birds in space
and time. Nature, 491, 444-448. https://doi:10.1038/nature11631
Juanes, F. 1986. Population density and body size in birds. Am. Nat. 128: 921929.
Keast A. 1968. Competitive interactions and the evolution of niches as illustrated by the Australian
honeyeater genus Melithreptus (Meliphagidae). Evolution 22, 762-784.
Krzanowski, A. 1967. The magnitude of islands and the size of bats (Chiroptera). Acta Zoologica
Cracoviensia, 15(XI), 281– 348.
Lomolino M. 1985. Body size of mammals on islands: the island rule re-examined. Am. Nat. 125,
310-316.
Lomolino M. 2005. Body size evolution in insular vertebrates: generality of the island rule. Journal
of Biogeography, 32(10), 1683–1699. doi:10.1111/j.1365-2699.2005.01314.x
MacArthur R. H. and Wilson E. O. 1963. An Equilibrium Theory of Insular Zoogeography.
Evolution, 17(4), 373. doi:10.2307/2407089
Mayr E. 1963. Animal species and evolution. Cambridge, 1WA: Harvard University Press
McNab B.K. 1994. Energy conservation and the evolution of flightlessness in birds. American
Naturalist, 144, 628– 642.
McNab B.K. 2002. Minimizing energy expenditure facilitates vertebrate persistence on oceanic
islands. Ecology Letters, 5, 693–704.
Olson V. A., Davies R. G., Orme C. D. L., Thomas G. H., Meiri S., Blackburn T. M., Gaston K.J.,
Owens I., and Bennett P. M. 2009. Global biogeography and ecology of body size in birds. Ecology
Letters, 12(3), 249–259. doi:10.1111/j.1461-0248.2009.01281.x

20

Rambaut A. and Drummond AJ. 2018. TreeAnnotator v.1.10.4: MCMC Output analysis Available
from http://beast.community/
RStudio Team. 2020. RStudio: Integrated Development for R. RStudio, Inc. Inc, Boston, MA.
Retrieved from https://www.rstudio.com/.
Runemark A., Sagonas K., and Svensson, E. I. 2015. Ecological explanations to island gigantism:
dietary niche divergence, predation, and size in an endemic lizard. Ecology, 96(8), 2077–
2092. doi:10.1890/14-1996.1
Spurgin L. G., Illera J. C., Jorgensen T. H., Dawson D. A., and Richardson D. S. 2014. Genetic
and phenotypic divergence in an island bird: isolation by distance, by colonization or by
adaptation? Molecular Ecology, 23(5), 1028–1039. doi:10.1111/mec.12672
Valen L. 1973. Body size and numbers of plants and animals. Evolution, 27(1),
27. doi:10.2307/2407116
Wright N., and Steadman D. 2012. Insular avian adaptations on two Neotropical continental
islands. Journal of Biogeography, 39(10), 1891–1899. doi:10.1111/j.1365-2699.2012.02754.x

21