Genetic and morphological evidence for two species of Leucocarbo shag (Aves, Pelecaniformes, Phalacrocoracidae) from southern South Island of New Zealand Author Nicolas J. Rawlence Author R. Paul Scofield Author Hamish G. Spencer Author Chris Lalas Author Luke J. Easton Author Alan J. D. Tennyson Author Mark Adams Author Eric Pasquet Author Cody Fraser Author Jonathan M. Waters Author Martyn Kennedy text Zoological Journal of the Linnean Society 2016 177 676 694 journal article 10.1111/zoj.12376 b49e437b-c2a1-4e2a-a683-84c637a0535b 270312 LEUCOCARBO STEWARTI FOVEAUX SHAG PHALACROCORAX STEWARTI OGILVIE-GRANT, 1898 Diagnosis. A species of Leucocarbo most closely related to L. chalconotus and L. onslowi but distinguished from these species by the plumage characters and allometries outlined in Table 1 . Distribution . Restricted to Foveaux Strait and Stewart (based on historical museum skins and modern specimens). Leucocarbo stewarti bones have been recorded from Late Quaternary and archaeological deposits from this region (e.g., Worthy, 1998b ). Rare archaeological and modern beach wrecks in Otago ( Rawlence et al ., 2014 , 2015 ). TYPE LOCALITIES The type locality of Gracalus chalconotus G. R. Gray, 1845 is currently considered to be Otago Province. However, the Otago Province included Southland and Foveaux Strait until 1861. We consider that the likely type locality is in fact Karitane, where the type specimen was likely collected by Percy Earl in 1843 ( Scofield et al ., 2012 ). Although Earl spent the majority of his time at Karitane, Earl only travelled as far south as the Clutha River, which is still within the range of the Otago lineage, but may have had Maori ‾ collect for him elsewhere ( Scofield et al ., 2012 ). The type locality of Phalacrocorax stewarti Ogilvie-Grant, 1898 is Bluff (a town in Southland), where specimens were collected by Baron A. von Hugel on 13 February 1875 ( Ogilvie-Grant, 1898 contra Stewart, Gill et al ., 2010 ). Ogilvie-Grant (1898) designated three syntypes ( NHMUK 1880.5.3.1, 1880.5.3.2, and 1880.5.3.6), all pied morphs. Warren (1966) only segregated and listed the syntype 1880.5.3.6 for inclusion in the NHMUK type collection, but this action does not affect the status of the remaining unselected syntypes . NHMUK 1880.5.3.1 and 1880.5.3.2 are currently labelled as Phalacrocorax campbelli huttoni (reflecting a previous taxonomic treatment) and we did not attempt to obtain DNA from them. As all three syntypes were collected from the same locality at the same time, and 1880.5.3.6 clusters with individuals from Foveaux Strait ( Figs. 4 , 6 , Appendix 1), we refer 1880.5.3.1 and 1880.5.3.2 to L. stewarti . PLUMAGE EVOLUTION AND PHYLOGEOGRAPHICAL CONSIDERATIONS Phylogenetic relationships amongst the Otago, Foveaux, and Chatham shags suggest a geologically recent colonization of the Chatham Islands from a mainland South source population, consistent with numerous phylogeographical studies from the region ( Goldberg, Trewick & Paterson, 2008 ; Mitchell et al. , 2014 ; Wood et al. , 2014 ). All New Zealand /sub-Antarctic Leucocarbo shags have subtle differences in plumage patterns (e.g. scapulars, prenuptial head crest; Marchant & Higgins, 1990 ). However, the Otago and Foveaux shags are the only Leucocarbo species with adult dimorphic plumage patterns – pied (black and white) and bronze (wholly dark) ( Figs 1 , 2 ). Given our phylogeny ( Fig. 4 ), we can hypothesize that the common ancestor of L. chalconotus , L. onslowi , and L. stewarti evolved dimorphic plumage. Divergence between L. chalconotus and L. stewarti is probably a result of isolation in refugia during the Pleistocene glacial cycles, when the majority of offshore islands and rock stacks in Foveaux Strait and around Stewart favoured by L. stewarti were landlocked because of lower sea levels. The Chatham Islands were colonized by progenitors of L. chalconotus , which evolved into L. onslowi . Colonization was either by pied L. chalconotus individuals or a mixture of pied and bronze individuals, with subsequent genetic drift resulting in the loss of the bronze morph. Given the lack of well-supported structure within the Otago shag we cannot definitively test evolutionary scenarios for the bronze morph, but it seems much simpler to assume that the bronze morph evolved once and was subsequently lost in L. onslowi than that it evolved independently twice (in the Otago shag and in the Foveaux shag). Table 1. Species diagnoses for the Otago shag Leucocarbo chalconotus (G. R. Gray, 1845) , Foveaux shag Leucocarbo stewarti (Ogilvie-Grant, 1898) , and Chatham Island shag Leucocarbo onslowi (Forbes, 1893)

Chatham Island shag
Feature	( L. onslowi )	Otago shag ( L. chalconotus )	Foveaux shag ( L. stewarti )	References
Prehistoric distribution	Chatham Islands	Eastern South Island	Foveaux Strait	Rawlence et al. (2015)
Modern distribution	Chatham Islands	Otago	Foveaux Strait	Lalas (1983), Lalas &
Perriman (2009),
Rawlence et al. (2014)
DNA [fixed mtDNA control	Monophyletic clade ‘daughter’	Monophyletic clade,	Monophyletic clade basal to	Rawlence et al. (2014,
region 1 ( CR1 ) differences,	to L. chalconotus .	distinguished from	L. chalconotus / L. onslowi .	2015)
248 bp; full CR , 1040 bp.	Distinguished from	L. stewarti by C-T (92), G-A	Distinguished from L. chalconotus
Numbers in parentheses	L. chalconotus by C-T (40),	(110), C-T (148), T-C (229),	by T-C (92), A-G (110), T-C (148),
represent the base pair	T-C (50), and C-T (105).	G-A ( 493 ), and T-C ( 628 ).	C-T (229), A-G ( 493 ), and C-T
position (e.g. 40) in the	Distinguished from	Distinguished from	( 628 ). Distinguished from
CR1 sequence where the	L. stewarti by T-C (24), C-T	L. onslowi by T-C (40), C-T	L. onslowi by C-T (24), T-C (40),
fixed difference (e.g. C-T)	(40), C-T (92), C-T (105),	(50), and T-C (105)	T-C (92), T-C (105), A-G (110), C-T
occurs.	G-A (110), T-C (117),	(117), T-C (148), C-T (229), A-G
C-T (148), T-C (229), G-A	( 493 ), and C-T ( 628 ).
( 493 ), and T-C ( 628 ).
Plumage morphs	Pied (black and white)	Pied (black and white),	Pied (black and white), bronze	Lalas (1983); Marchant
bronze (wholly dark)	(wholly dark)	& Higgins (1990);
Rawlence et al. (2014)
Plumage morph	100% pied	~20–30% pied	~50–60% pied	Lalas (1983); Rawlence
percentages	et al. (2014)
Carunculation (breeding	Pronounced bright orange	Equal frequencies of small,	Scattered dark orange papillae	Lalas (1983); Rawlence
plumage)	facial caruncles	bright orange facial	et al. (2014)
caruncles and dark orange
scattered papillae
Gular pouch (breeding	Bright orange	Bright to dark orange or	Bright to dark orange	Lalas (1983); Rawlence
plumage)	purple	et al. (2014)
Morphometics/allometries	Smaller on average than both	Larger on average than both	Smaller on average than	Lalas (1983); Rawlence
L. chalconotus and	L. stewarti and L. onslowi	L. chalconotus , and similar in size	et al. (2014); this study
L. stewarti	to L. onslowi
Osteology	The foramen vasculare	The foramen vasculare	The foramen vasculare distale on	Worthy (2011); this
distale on the distal	distale on the distal	the distal tarsometatarsus is small	study
tarsometatarsus is small	tarsometatarsus varies in
size
Onset of breeding season	September–February	May–September	September onwards	Lalas (1983)

CONSERVATION IMPLICATIONS Prior to our splitting of the Stewart shag into two distinct taxa, the International Union for Conservation of Nature Red List categorized this species as Vulnerable with population estimates varying between 1600 and 1800 pairs to fewer than 5000 birds and threats including fisheries interactions and nest disturbance ( Birdlife International, 2012 ). Given their revised status, Otago and Foveaux shags should be managed separately, with further population-genetic research to determine levels of genetic variation at faster-evolving nuclear loci and potential inbreeding (e.g. Calderon et al. , 2014 ). The Foveaux shag has been characterized by population and range stability, in contrast to the Otago shag that has undergone a pronounced population bottleneck and range contraction ( Rawlence et al. , 2015 ). Although the Otago shag has recolonized part of its former distribution since the European era, its population size continues to decline ( Lalas & Perriman, 2009 ).