Review Reports - Mitochondrial Genomes from New Zealand’s Extinct Adzebills (Aves: Aptornithidae: <i>Aptornis</i>) Support a Sister-Taxon Relationship with the Afro-Madagascan Sarothruridae

Round 1

Reviewer 1 Report

Article Mitochondrial genomes from New Zealand’s extinct adzebills (Aves: Aptornithidae: Aptornis) support a sister-taxon relationship with the Afro-Madagascan Sarothruridae by Boast et al.

This manuscript has the potential to be an interesting contribution to our understanding of the evolution of the enigmatic adzebills. The phylogeny provides the framework for important advances relating to evolution, classification and biogeography of the group. These aspects of the research are robust and the authors are completely forthcoming to point out the few instances where the data are inclusive.

However, the manuscript attempts to cover a wide range of taxonomical issues within Ralloidea, especially “true” rails, that seems gratuitous, tangential to the paper, and should be omitted (e.g., page 13 lines 451-498).

Line 77-80: There is not even a single avian clade whose crown divergences are directly caused by the break-up of Gondwana.

Methods: Please give more details about the alignment method. Did you partition the alignments into protein-coding genes, tRNAs, rRNAs and noncoding fragments? This will be especially useful for checking the alignments of RNA secondary structures. Please also add information about the length of the data analysed in each alignment (A, B and C).

There were papers by Ksepka et al. 2014 and Mayr 2014 that suggested that mtDNA alone generally puts older dates than it should on ‘deeper’ nodes. I think you should give a little more idea of how much confidence we should have in the old dates you get and discuss this with the similar ages obtained from Garcia-R et al 2014.

The upper bound for the colonisation of the Adzebills’ ancestor is about 39 Ma, the oldest Adzebill fossil is about 16 Ma, and the divergence between the North and South Adezebill is 1.5 Ma. There is approximately a 20 million year gap in the fossil record to infer the crown group of Adzebills and about 15 million gap of no other fossil presence for the divergence of the two Adzebills. I suggest discussing this discrepancy a bit further.

Line 599 sound as a phylogenetic tree was built but I did not find the tree. It would be appropriate to show a phylogenetic tree from the morphological dataset as well as a discussion of the synapomorphies among Rallidae and heliornithoids (lines 567-569)

Minor comments:

Line 56: Replace “animal” by “birds”

Line 57: Delete “, including the adzebill (Aptornis: Aptornithidae),”

Line 123: Please explain acronyms AMS and OSL

Lines 285 and 290: Please explain where the data is available

Line 307: It is unclear why the F84 model was chosen

Line 367: delete “the” after topology

Please show the calibration points in Figure 2

Lines 572-574: I do not think this statement is accurate. Rallidae contained members that are morphologically and ecologically varied. Coots are present in lakes and show traits such as lobed toes and frontal shields whilst other taxa can be found in tussock grass (Takahe), forest or coastal margins with variation in sizes and other traits (species of Gallirallus), etc.

Line 577-585: Also, please discuss or give a few suggestions about how the North Adzebill become smaller in such a short time after the formation of the Manawatu strait.

Line 622: missing parenthesis

Author Response

Alexander Boast

Landcare Research

PO Box 69040

Lincoln 7640

NEW ZEALAND

apboast@gmail.com

+64274775230

25 February 2019

Diversity

Dear Reviewer,

Please find attached our response to your comments on our manuscript titled “Mitochondrial genomes from New Zealand’s extinct adzebills (Aves: Aptornithidae: Aptornis) support a sister-taxon relationship with the Afro-Madagascan Sarothruridae”.

Firstly, we wish to pass on our appreciation to you for spending your time on our manuscript, and feel your comments were balanced and fair. We believe your suggestions have led to an improved article.

We have attached all comments below in this document and have provided our responses (and appropriate action on the revised manuscript, if applicable) below each statement in a blue font. All lines quoted in our response represent those of the revised manuscript and may differ from the lines quoted earlier by yourself.

Furthermore, we ourselves have made several minor corrections on our revised manuscript. All corrections have been noted as track changes.

We also wish to note that Figure 2, and Tables S1 and S3 have been revised and resubmitted.

Furthermore, we also wish to note that we have now added GenBank accession numbers for our mitochondrial consensus sequences. A figshare DOI is still pending, and can be added to the Data Availability section in the proof stage.

Yours sincerely,

Alexander Boast

Reviewer Comments:

Reviewer #1:

Open Review

(x) I would not like to sign my review report
( ) I would like to sign my review report

English language and style

( ) Extensive editing of English language and style required
( ) Moderate English changes required
(x) English language and style are fine/minor spell check required
( ) I don't feel qualified to judge about the English language and style

	Yes	Can be improved	Must be improved	Not applicable
Does the introduction provide sufficient background and include all relevant references?	(x)	( )	( )	( )
Is the research design appropriate?	(x)	( )	( )	( )
Are the methods adequately described?	(x)	( )	( )	( )
Are the results clearly presented?	( )	(x)	( )	( )
Are the conclusions supported by the results?	( )	(x)	( )	( )

Comments and Suggestions for Authors

Article Mitochondrial genomes from New Zealand’s extinct adzebills (Aves: Aptornithidae: Aptornis) support a sister-taxon relationship with the Afro-Madagascan Sarothruridae by Boast et al.

We fully understand why the phylogeny of rails may appear to be tangential the paper, therefore we wish to take this opportunity to explain our reasoning. Recent phylogenetic studies have suggested some “rails” are not close relatives of “true rails” (Rallidae) and are closer relatives of the finfoots (Heliornithidae), specifically the flufftails (Sarothruridae) and the Madagascan wood rails (Mentocrex). As our study identified that adzebills form a clade with the finfoots, flufftails, and Madagascan wood rails (a group we call the “heliornithoids”), it was necessary for us to investigate whether any other ralloids fell within the heliornithoid clade. It is also widely acknowledged that inclusion or exclusion of outgroup taxa can affect ingroup topology and rooting. All non-heliornithoid ralloids with molecular data were found to comprise a monophyletic Rallidae (“true rails”). It was the view of the authors to explain the implications of these results in detail (presently lines 473-507) for the following reasons:

1) Our analyses comprise the most comprehensive molecular dataset of Ralloidea yet assembled (including the first molecular data for the phylogenetically distinct grey-throated rail Canirallus oculeus).

2) Our large phylogeny generally recapitulated the results of earlier molecular studies, which provides support that our new results (e.g. the position of adzebills) are likely to be robust.

3) Although the Rallidae were not focal to our study, our phylogeny provided important new resolution into rail phylogenetics which we felt deserved discussion.

4) Our phylogeny identified some results that contradicted prior studies. It was therefore essential we identified where these contradictions occurred and provided hypotheses for their occurrence.

5) We identified discrepancies in the available molecular data which may reflect sampling error (for example alternative placements for sequences obtained from different studies, of the New Guinea flightless rail Megacrex ineptus). It was important all these discrepancies were identified and hypotheses for which of the sequences was “correct”.

6) We also wished to identify what is inconclusive in our analyses, and where future work should concentrate.

Overall, we acknowledge the discussion of the wider taxonomy might appear to sit outside of the primary scope of this paper. However, it remains the view of the authors that our all our discussion of ralloid and rallid phylogenetics are necessary for discussions surrounding placement of Aptornis.

Line 77-80: There is not even a single avian clade whose crown divergences are directly caused by the break-up of Gondwana.

We agree wholeheartedly. Nonetheless, it is critical that this statement remain in the manuscript. Hypotheses of Gondwanan vicaricance in avian evolution, including between New Zealand’s adzebills and New Caledonia’s kagu, have been prevalent in the literature and have appeared in recent major publications (e.g. Claramunt and Cracraft, 2015). Gondwanan vicariance is a hypothesis we reject (e.g. the abstract, lines 32-35).

Our text on lines 76-82 reads “[adzebills] have usually been placed in Gruiformes along with several other bird families found in former fragments of Gondwana (especially New Caledonia, Madagascar and South America). It has often been proposed that the ancestor of these “gruiform” bird families, including adzebills, inhabited Gondwana prior to its break up and that their present biogeography reflects continental vicariance [21]. Specifically, morphological analyses have suggested adzebills are most closely related to the New Caledonian Kagu (Rhynochetos jubatus) [21–24], consistent with the possible existence of emergent land connecting New Zealand and New Caledonia during the Palaeogene [25,26].”

We thus only make the statement that Gondwanan vicariance is a hypothesis that has appeared in the literature and has been used to connect the adzebill and kagu. We later go on to say in the same paragraph that molecular data fails to confirm monophyly of these Gondwanan “gruiformes”, and thus hypotheses on Gondwanan vicariance are unsupported.

Our alignments were partitioned into protein-coding genes, tRNAs and rRNAs, which we mention at other points in our manuscript, including saturation tests (lines 309-311) and in partitionfinder model selection (lines 323-325). We thank you for highlighting that this was unclear and have added a new sentence describing the alignment procedures in more detail (lines 269-270). We also acknowledge that it may have been unclear the lengths of all alignments were identical, and they differed only in regard to the taxa included or excluded. A new note detailing as such has been provided (lines 260-261).

We have done as you suggest and provided a new sentence in our results (lines 533-534), which references the study by Ksepka et al. (2014) and notes on the differences between mitochondrial and nuclear genomic molecular clocks. We are of the view the reference by (Mayr, 2013) may not be appropriate for this statement, as this article specifically critiqued molecular clock analyses of passerines which found Cretaceous age origins for this bird group (rather than any discussion of genomic vs mitochondrial markers). Many of the studies critiqued by Mayr also obtained their older dates due to the use of controversial calibration points, (specifically using 82 Ma, taken to represent New Zealand’s rifting from Gondwana) rather than artefacts presented by mitochondrial DNA specifically. Discrepancies between nuclear and mitochondrial clocks and the fossil record, are widely acknowledged, poorly understood, and widely debated. There exists extensive literature on the subject and we believe this debate falls outside the scope of this paper.

Nonetheless as we note (lines 535-537), our estimated dates (with HDP 95 % percentile distributions taken into consideration) overlap considerably with larger phylogenomic studies (such as Claramunt & Cracraft). Furthermore, the minimum age estimated for all calibrated nodes (exception of the adzebill-Sarothruridae split), align closely with the minimum fossil ages we provide (see Figure 2 in our manuscript). It is our view that most of our divergence estimate dates are conservative (containing appropriate errors) and should stand as the temporal context for subsequent dating efforts involving Ralloidea.

We thank you for highlighting this aspect, which is key in understanding the evolution of New Zealand’s biota and should be mentioned. New Zealand has a paucity of major Cenozoic-aged terrestrial fossil deposits outside of the late Pleistocene or Holocene, and the Miocene (19-16 Ma) aged St Bathans site is an important exception. Therefore, New Zealand taxa with a similar antiquity to adzebills also lack fossils outside of the Quaternary and Miocene eras (e.g. moa, kiwi, acanthisittid wrens, strigopoid parrots, mystacinid bats or tuatara). Therefore, this fossil gap is not unique to adzebills and is typical of NZ species in general (e.g. Jones et al., 2009; Tennyson et al., 2010; Worthy et al., 2013, 2011). As suggested by yourself we have now added a comment addressing the paucity of fossils (lines 589-590).

We thank you for your suggestion, but disagree that it is necessary to present a tree for the following reasons:

1) Our morphological analyses used character matrices from previous studies, and the results of these studies are now published (Livezey, 1998; Livezey and Zusi, 2007).

2) These morphological analyses are ancillary to our study and were only undertaken to support the results of our molecular analyses.

3) The phylogenetic trees produced from the morphological data had low posterior support for most nodes and comprised numerous polytomies. All the key results produced from these analyses are mentioned in the results.

4) The character states in these publications (i.e. the synapomorphies described) are in a coded format and require considerable additional work to identify and describe. It was the view of the authors that extensive morphological analyses fell out of the scope of our study and would be appropriate for future work to undertake (lines 520-523). We also do not believe a full description of characters is particularly insightful in the absence of functional or regional commonality among them, e.g., all cranial vs postcranial or all related to feeding.

Minor comments:

Line 56: Replace “animal” by “birds”

We deliberately elected to use “animal” over “birds”, to include NZ’s phylogenetically distinct non-avian animal lineages (including bats, reptiles, frogs and many different lineages of invertebrates). All these groups have undergone extinction in NZ.

Line 57: Delete “, including the adzebill (Aptornis: Aptornithidae),”

This sentence is key to addressing the taxonomic objectives of our study and is the norm when describing the genus and family.

Line 123: Please explain acronyms AMS and OSL

We have explained the above acronyms (lines 123-124).

Lines 285 and 290: Please explain where the data is available

We have clarified a statement at the start of this same paragraph, which directs the reader to refer to Table S3 for all sequence information.

Line 307: It is unclear why the F84 model was chosen

The F84 model was selected as it is the default setting in this analysis, and specifically distinguishes between the rate of transitions and transversions while allowing unequal base frequencies typical of mitochondrial data. Therefore, this model is especially useful for the saturation analyses described here. Other models explored demonstrated near-identical results.

Line 367: delete “the” after topology

This error has been corrected.

Please show the calibration points in Figure 2

These have now been added, and the colour scheme of this figure has been adjusted to make these points clearer.

We agree that this is a subjective statement, and we have replaced the comment “heliornithoids are considerably more [varied]” with “heliornithoids are arguably more [varied]”. Nonetheless we still wish to note the wide eco-morphological variation of heliornithoids which comprise the massive terrestrial >16kg adzebills, the fully aquatic diving finfoots and diminutive, sparrow-sized flufftails. Although rails are also diverse, they lack such a broad scope of variation in size or morphology despite a far greater species diversity. For example, the rallid coots (Fulica) lack the extreme aquatic adaptations of finfoots (Heliornthidae) (Houde, 1994), and the largest rails (the two ~4kg species of takahe in Porphyrio) are dwarfed by the >16kg adzebills. Flufftails are also probably the smallest living ralloids: for example, adult Sarothrura affinis at 25-30 g compared to the Inaccessible island rail Atlantisia rogersi (the smallest living rail) at 34-40.5 g (weights obtained from HBW checklist https://www.hbw.com/). Finally, the heliornithoid Madagascan wood rails (Mentocrex) so closely resemble “true” rails, they have often been placed in the rallid genus Canirallus. We feel the arguments above and the changes made to the manuscript address this comment.

Line 577-585: Also, please discuss or give a few suggestions about how the North Adzebill become smaller in such a short time after the formation of the Manawatu strait.

The size difference between the two adzebill species is minor, with the South Island species being roughly 20% larger on average (there is considerable overlap between specimens of both species). The South Island adzebill weighed approximately 19kg, and North Island adzebill weighed approximately 16kg (lines 60-62).

1) Such an increase (or decrease) in size is not unusually fast for a divergence of 1-2 Ma. For example, several moa species shrunk considerably more in size than 20% in response to warming climates in the past ~12,000 years (Worthy and Holdaway 2002).

2) The size difference between North Island / South Island adzebills is typical among New Zealand birds (Tennyson and Martinson, 2006; Worthy and Holdaway, 2002). For example, the South Island Goose Cnemiornis calcitrans was roughly 20% the size of the North Island goose C. gracilis. Among moa, the South Island Pachyornis elephantopus was roughly 3 times heavier is sister taxon, the North Island P. geranoides (~87kg vs ~27 kg). P. elephantopus and P. geranoides are estimated to have diverged the same time as the two species of adzebill, at ~1.5 Ma (Bunce et al., 2009).

3) Greater divergences in body mass have been seen within populations of NZ bird species across distances as short as 250km due to changes in in altitude or vegetation type (Worthy et al. 2005).

In general, this size difference is not unexpected or unusual, and instead very likely reflects Bergmann’s rule (with cooler climates in the South Island) or resource limitation. In response to your comment we have now addressed this hypothesis (see lines 668-673).

Line 622: missing parenthesis

This error has been corrected.

References

Bunce, M., Worthy, T.H., Phillips, M.J., Holdaway, R.N., Willerslev, E., Haile, J., Shapiro, B., Scofield, R.P., Drummond, A., Kamp, P.J.J., et al. (2009). The evolutionary history of the extinct ratite moa and New Zealand Neogene paleogeography. Proc. Natl. Acad. Sci. 106, 20646–20651.

Claramunt, S., and Cracraft, J. (2015). A new time tree reveals Earth history’s imprint on the evolution of modern birds. Sci. Adv. 1, e1501005.

Houde, P. (1994). Evolution of the Heliornithidae: reciprocal illumination by morphology, biogeography and DNA hybridization (Aves: Gruiformes). Cladistics 10, 1–19.

Jones, M.E., Tennyson, A.J., Worthy, J.P., Evans, S.E., and Worthy, T.H. (2009). A sphenodontine (Rhynchocephalia) from the Miocene of New Zealand and palaeobiogeography of the tuatara (Sphenodon). Proc. R. Soc. Lond. B Biol. Sci. 276, 1385–1390.

Ksepka, D.T., Ware, J.L., and Lamm, K.S. (2014). Flying rocks and flying clocks: disparity in fossil and molecular dates for birds. Proc. R. Soc. Lond. B Biol. Sci. 281, 20140677.

Livezey, B.C. (1998). A phylogenetic analysis of the Gruiformes (Aves) based on morphological characters, with an emphasis on the rails (Rallidae). Philos. Trans. R. Soc. Lond. B Biol. Sci. 353, 2077–2151.

Livezey, B.C., and Zusi, R.L. (2007). Higher-order phylogeny of modern birds (Theropoda, Aves: Neornithes) based on comparative anatomy. II. Analysis and discussion. Zool. J. Linn. Soc. 149, 1–95.

Mayr, G. (2013). The age of the crown group of passerine birds and its evolutionary significance–molecular calibrations versus the fossil record. Syst. Biodivers. 11, 7–13.

Tennyson, A.J.D., and Martinson, P. (2006). Extinct birds of New Zealand (Wellington: Te Papa Press).

Tennyson, A.J., WorTHy, T.H., Jones, C.M., Scofield, R.P., and Hand, S.J. (2010). Moa’s Ark: Miocene fossils reveal the great antiquity of moa (Aves: Dinornithiformes) in Zealandia. Rec. Aust. Mus. 62, 105–114.

Worthy, T.H., and Holdaway, R.N. (2002). The lost world of the moa: prehistoric life of New Zealand (Indiana University Press).

Worthy, T., Worthy, J.P., Tennyson, A.J., Salisbury, S.W., Hand, S.J., and Scofield, R.P. (2013). Miocene fossils show that kiwi (Apteryx, Apterygidae) are probably not phyletic dwarves.

Worthy, T.H., Tennyson, A.J., and Scofield, R.P. (2011). An early Miocene diversity of parrots (Aves, Strigopidae, Nestorinae) from New Zealand. J. Vertebr. Paleontol. 31, 1102–1116.

Worthy, T.H., M. Bunce, A. Cooper, P. Scofield, 2005. Dinornis – an insular oddity, a taxonomic conundrum reviewed. Alcover, J. A. & Bover, P. (eds), Proceedings of the International Symposium “Insular Vertebrate Evolution: The Palaeontological approach” Monographies de la Societat d’Historia Natural de les Balears 12: 337-390

Reviewer 2 Report

This is a great paper. The question is very clear and previous work is well described. The data and analyses answer the central question decisively, and another major branch of the avian phylogeny falls into place. The Discussion really helps to put the significance of these findings in context with regards to biogeography and morphological evolution. I congratulate the authors on a job well done.

The phylogenetic analyses appear to be very rigorous to me overall.

In particular, it was nice that the investigators incorporated the morphological data from Livezy and Zusi.

I was pleased to see the sources of samples attributed to museum catalog numbers (Table S1), with one odd exception as noted on line 131 (that is a strange attribution for the tissue source, would be ideal to clarify that further, only in order that this work could possibly be extended in the future).

One question that I couldn't find addressed anywhere, but is somewhat obvious: Are the two North Island sequences identical or divergent? And if divergent, how divergent, and what implications does that have?

Specific notes:

Line 123: I suggest writing out OSL and AMS rather than abbreviating, as many readers won't be familiar with these techniques.

Line 645-7: This is certainly a long shot, but is there any morphological differentiation between north and south island forms? If so, could morphological comparison with the Miocene fossil help to elucidate the biogeographical history (for example, if there has been morphological change that occurred recently on the north island following the colonization)?

Supplementary Figures: I couldn't find captions for the Supplementary figures -- maybe I missed them. Anyway, they are needed in this case to interpret Figures S1-3.

Author Response

Alexander Boast

Landcare Research

PO Box 69040

Lincoln 7640

NEW ZEALAND

apboast@gmail.com

+64274775230

25 February 2019

Diversity

Dear reviewer,

We thank you very much for your time and effort, and your recommendation of our article. Your comments are very fair and believe have helped to produce an improved manuscript.

Furthermore, we ourselves have made several minor corrections on our revised manuscript. All corrections have been noted as track changes.

We also wish to note that Figure 2, and Tables S1 and S3 have been revised and resubmitted.

Yours sincerely,

Alexander Boast

Reviewer #2:

Open Review

(x) I would not like to sign my review report
( ) I would like to sign my review report

English language and style

	Yes	Can be improved	Must be improved	Not applicable
Does the introduction provide sufficient background and include all relevant references?	(x)	( )	( )	( )
Is the research design appropriate?	(x)	( )	( )	( )
Are the methods adequately described?	(x)	( )	( )	( )
Are the results clearly presented?	(x)	( )	( )	( )
Are the conclusions supported by the results?	(x)	( )	( )	( )

Comments and Suggestions for Authors

his is a great paper. The question is very clear and previous work is well described. The data and analyses answer the central question decisively, and another major branch of the avian phylogeny falls into place. The Discussion really helps to put the significance of these findings in context with regards to biogeography and morphological evolution. I congratulate the authors on a job well done.

The phylogenetic analyses appear to be very rigorous to me overall.

In particular, it was nice that the investigators incorporated the morphological data from Livezy and Zusi.

We have provided additional details regarding the provenance of this specimen (Table S1, lines 131-135).

This is a good point. We have provided a pairwise difference in the results (lines 551-552). Our molecular clock analyses suggest that the species diverged approximately 1.1 Ma, justifying their separate species status. We also note how this age corresponds with the formation and cessation of a land-bridge between the North and South Islands in the discussion.

Specific notes:

Line 123: I suggest writing out OSL and AMS rather than abbreviating, as many readers won't be familiar with these techniques.

These abbreviations have now been expanded (lines 123-124).

The North and South Island adzebill taxa slightly differed in size but were otherwise near-identical (see new text in lines 668-673). Due to the long age gap between the St. Bathans adzebill (>16 Ma) and the recent split of the North and South Island adzebills (<2 Ma), and incomplete nature of the St Bathans adzebill, such a morphological analysis would likely be difficult and/or inconclusive.

Supplementary Figures: I couldn't find captions for the Supplementary figures -- maybe I missed them. Anyway, they are needed in this case to interpret Figures S1-3.

These captions were provided in our original manuscript submission and have been attached to the revised manuscript on the final page after the references (lines 944-970).