BOTANY .
Observations on the Botany of Kerguelen Island .
J. _D .
, The history of the botany of Kerguelen Island ( also called Kerguelen 's Land , and Desolation Island ) , previous to the visit of the'Rev .
Mr. Eaton , the last and most complete explorer of its flora , is a very brief one .
It commences with the visit of Oapt .
Cook during his third voyage , in the narrative of which the vegetation of the island is thus described by Mr. Anderson , the surgeon of the " Resolution : " " Perhaps no " place hitherto discovered in either hemisphere , under the same parallel of latitude , " affords so scanty a field for the naturalist as this barren spot .
The verdure which " appears , when at a little distance from the shore , would flatter one with the expec- " tation of meeting with some herbage ; but in this we were much deceived .
Eor " on landing we discovered that this lively colour was occasioned only by one small " plant , not much unlike some sorts of Sax , which grows in large spreading " tufts , to a considerable way up the hills .
" Mr. Anderson proceeds then to give some particulars of this plant { Azorella S , Hk .
f. ) , of the cabbage ( antiscorbutica , Br .
) , of two small plants found in boggy places , which were eaten as salad , one " almost like garden cress and very fiery " ( probably Ranunculus Hk .
f. ) , the other very mild and " having not only male and female , but what bota- " nists call androgynous plants " ( ?
CallitricheHe adds to these a coarse grass { Boa Cookii , Hk .
f. ) , and a smaller sort which is rarer ( probably antarctica , Hk .
) ; a sort of goose-grass ( ?
Cotula , Hk .
f. ) , and another small plant much like it ( this I do not recognise ) .
" In short , " he says , " the whole " catalogue of plants does not exceed 16 or 18 , including some sorts of moss and a " beautiful Lichen " { Neuropogon Taylori , Hk .
f. ) " which grows higher upon the " rocks than the rest of the vegetable productions .
Nor is there the least appear- " ance of a shrub in the whole country .
" The date of Cook 's visit was the summer of 1776 , and the specimens obtained by Mr. Anderson were deposited in Sir Joseph Banks ' Herbarium , which subsequently became the property of the nation , and is preserved in the British Museum .
Not having been poisoned , all the Kerguelen Island plants were , when I examined them in 1843 , much injured by insects , and many were entirely destroyed .
Erom 1776 till 1840 , when the Antarctic Expedition under Capt. ( afterwards Admiral Sir James ) Boss , anchored in Christmas Harbour , Kerguelen Island is not known to have been visited by any ship of war , or by the Discovery or Surveying ships of any nation , though it had become the frequent resort of English and B American sealers .
During the stay of the above-named expedition all tlic plants enumerated by Anderson as found by him in mid-summer were refound in midwinter , together with many more , amounting to nearly 150 , ol which 18 were flowering plants " , the other large classes being mosses and Hepaticse 3o , Dicliens 25 , and Algae 51 .
These have all been described in the botany of the voyage ( Flora Antarctica , Part II .
, 1847 ) .
The next visit of naturalists to Kerguelen 's Land was that of the " Challenger " Expedition in January and February 1874 , when Mr. Moseley collected most diligently , both in Christmas Harbour and on the east coast 60 to 70 miles south-east of it .
He found 23 flowering plants in all , including three European weeds , all annuals and doubtless imported by sealing parties ( Cerastium trimale , Poa and annua ) , and three species not in the collections of the Antarctic Expedition ( two 'Ranunculi and an Uncinia ) .
He also procured flowering specimens of the two endemic genera Rringlea and Lyallia , and made large accessions to the cryptogamic flora , especially from the southern localities visited .
Mr. Moseley had also the good fortune to land upon Marion Island , 1,650 miles to the west of Kerguelen Island ; and on Yong Island ( of the Heard group ) , about 120 miles to the south-east of it , neither of which had been previously visited by any naturalists , and in both of which he found some of the most peculiar of the Kerguelen plants .
Mr. Eaton arrived at Kerguelen Island with the Transit of Venus Expedition early in October 1874 , and left towards the end of February 1875 , during which time he collected diligently , chiefly at Boyal Sound , Swains ' Pay , and Observatory Bay .
He obtained nearly all the flowering plants of previous explorers , and added very largely in the Cryptogams , especially to the Algae .
Nearly contemporaneous with Mr. Eaton 's visit was that of the American Transit Expedition , on which Dr. Kidder was the naturalist .
He arrived in September 1874 and left in January of the following year , having explored some of the same localities as Mr. Eaton .
His collections , amounting to about 90 species , are described in the bulletin of the II .
S. National Museum , No. 3 , issued in 1876 by the Government Printing Office of Washington .
The flowering plants and ferns are revised by Prof. A. Gray ; the mosses are described by Those .
P. James ; the Lichens by Prof. E. Tuckerman , and the Algae by Dr. W. G. Farlow .
Except amongst the Lichens , there are very few novelties .
Dr. Kidder adds a list of seven plants from the Crozets , all identical with Kerguelen Island species.* The botanical results of the German Transit Expedition to Kerguelen Island are not yet published .
The three small archipelagos of Kerguelen Island ( including the Heard Islands ) , Marion and I rince Edward 's Islands , and the Crozets , are individually and collectively the most barren tracts on the Globe , whether in their own latitude or in He also mentions a small vine with blue flowers growing amongst scorias , " of which no specimens were collected .
This is probably some endemic plant unknown to botanists .
BOTANY OF KERGUELEN ISLAND.\#151 ; DR. HOOKER .
any higher one , except such as lie within the Antarctic Circle itself ; for no land even within the N. Polar area presents so impoverished a vegetation .
The chief interest attached to the flora of these archipelagos lies in the indication it affords of their being , in all probability , the remains of a much larger land area , which , though peopled with plants mainly from the southern extreme of S. America , 4,000 miles to the westward , possessed an endemic flora of its own , which included forest trees of considerable dimensions .
Before , however , proceeding to discuss the relationships of their floras , I shall describe that of the largest and the only one that is at all well known .
As pointed out in the " Plora Antarctica , " the prevalent features of the vegetation of this island as then known were Puegian ; one species of flowering plant alone , of those that are not peculiar to the island , being characteristic of any other flora , namely , the Cotula plumosa , which is found elsewhere only in the Auckland and Campbell Islands , south of New Zealand .
More recent collections have confirmed and even strengthened this Puegian affinity , for of the three additional flowering plants procured by subsequent explorers , one is Puegian ( .
trullifolius ' ) , another ( R. Moseleyi ) is closely allied to a Puegian species , and the third one , Uncinia compacta , is a native of the mountains of New Zealand and Tasmania , and this is so nearly allied to a Puegian species that it may prove to be a form of a plant common to all high southern latitudes .
Not only has a further knowledge of the Kerguelen Island flora strengthened its known affinities with the Puegian , but recent discoveries in the latter flora have done so too ; some of the Kerguelen 's grasses especially proving to be more closely allied to Puegian species than was suspected .
The discovery of the flowers of the endemic Kerguelen genus Lyallia is another instance of this affinity .
In the Plora Antarctica , judging from the fruit alone , the flowers being unknown , this remarkable plant was provisionally placed in Bortulacece , its resemblance in habit and foliage to the andine genus Bycnopliyllum being indicated .
Complete specimens collected by Moseley prove its close relationship to the latter genus , in juxta-position with which it had indeed been placed in the Genera Plantarum , where both had been referred correctly to Caryophyllece .
The elements of the Phsenogamic flora of Kerguelen Island may be thus classified :\#151 ; 1 Endemic genus , which has no near ally\#151 ; Br ingle a antiscorbutica .
1 Endemic genus allied to an Andean one\#151 ; Lyallia 6 Endemic species allied to American congeners\#151 ; Ranunculus crassipes and Moseleyi , Colobantlius herguelensis , Accena , Boa Festuca Jcerguelensis .
5 species common to Puegia but not found elsewhere : Ranunculus , Azorella Selago , Galium antarcticum , Festuca , arctica .
6 species common to America , and also to New Zealand and the islands south of it .
Tillcea moschata , Montia font ana* , * Limosella aquatica* Juncus scheuzerioides , Magellanica .
( Most of these are aquatic or marsh plants , and those marked with an asterisk are also European , and very widely dispersed .
) 2 species found elsewhere hut not in Euegia , Cotula common to Lord Auckland 's group and Campbell 's Island south of New Zealand , and compacta , a native of the mountains of Tasmania and New Zealand .
This American affinity of the Kerguelen Island flora thus clearly established by its flowering plants is very strongly manifested by its Cryptogams , amongst which , however , the only evidence of migration from South Africa occurs .
This is the case of Polypodium vulgar .
, a widely distributed fern in the north temperate zone , but known in the southern only from the Cape Colony , Marion , and Kerguelen Islands ; what is further curious respecting it is , that the Kerguelen Island individuals are referable to a variety with pellucid veins , hitherto known only from the Sandwich Islands .
As to the local grouping of the Kerguelen Island plants , that of the Phsenogams is not altogether in harmony with the Cryptogams , the former seeming to be by far the most ubiquitously dispersed of the two groups .
All the plants hitherto collected have been from two areas , one , Christmas Harbour , in the extreme north , extending about five miles either way ; the other , considerably larger , occupies the south-east coast , and following it extends for about 40 miles .
The distance between these areas is about 60 miles in a N.W. and S.E. direction .
Of the Phsenogaraic plants , 19 were found in the northern area , nearly every one of which was also found in the south-eastern one , where but two additional species were collected ; whereas of the 150 Cryptogams found in the northern area , a large proportion were not found in the south-eastern , where , however , nearly four times the number of species were obtained .
Again , whilst but one fern was found in the north , four occur in the south-east .
Of 35 and Hepaticce collected at Christmas Harbour by the Antarctic Expedition , hardly half were found at Swain 's Bay , Betsy Cove , or Boyal Sound , which localities yielded about 80 additional species .
Nearly 50 marine Algge were collected at Christmas Harbour , of which 18 did not occur in the south-eastern coasts , where ] upwards of 30 additional species were obtained .
In the case of the Lichens , the discrepancy is still more marked , but this is possibly more apparent than real , and is to be attributed in part to the difficulty of defining the species and recognizing them from descriptions ; and in part to the difficulties caused by the irreconcilable views of Lichenologists as to the limits of the species of this order .
Whatever other causes there may be for this anomalous distribution , one , Uo doubt , is the nature of the Christmas Harbour area .
This is almost occupied by transverse valleys that run east and west completely across the north tip of the BOTANY OF KERGUELEN ISLAND.\#151 ; DR. HOOKER .
island , from sea to sea , are bounded by bills 1,200 feet high , and are perennially swept by terrific blasts from the westward .
There are , hence , no shelter on land for the terrestrial flora , and no quiet bays for the proper development of a varied marine vegetation ; facts which may very well account for the paucity of Cryptogams in Christmas Harbour , but not for the presence there of nearly all the flowering plants of the island .
Turning again to the south-eastern area , its more sheltered valleys and land-locked harbours favour not only a greater development of Cryptogams , but also a far greater luxuriance of the Phsenogams than obtains in Christmas Harbour ; which last fact renders the absence of additional species of Phsenogams to the south-eastward all the more remarkable .
The question remains , granting that the great majority of the Phsenogams of Kerguelen Island are derived from South America , how was their transport effected ?
Though this question cannot be satisfactorily answered by a reference to the facilities for distant transport possessed by the fruiting organs of the Kerguelen Island plants , it is only proper to refer to these organs in some detail .
Obviously , regarding the whole flora , the plants with the most minute seeds or spores and the water-plants are the most widely distributed .
Under these categories come\#151 ; 1 .
The Pungi , of which all but 2 of the 8 species found are widely dispersed over the globe .
2 .
The marine Algae , of which only 8 out of the 74 are peculiar to the island .
3 .
The fresh water Algae , of which 28 out of 80 are regarded as endemic .
4 .
The aquatic and marsh Phaenogams , 8 in all , of which 6 are widely dispersed .
Of the Phaenogams , whether aquatic , marsh , or terrestial , none have appliances for wide dispersion except the hooked style of the Ranunculus , the reversed barbs of the Accena ( a most powerful aid ) , and the hooked organ attached to the fruit of Uncinia , also a very adequate aid .
None of the others have any aid to dispersion , though they have small seeds or fruits .
Turning to the natural agents of dispersion , winds are no doubt the most powerful , and sufficient to account for the transport of the Cryptogamic spores ; these , almost throughout the year , blow from Puegia to Kerguelen Island , and in the opposite direction only for very short periods , but appear quite insufficient to transport seeds over 4,000 miles .
Oceanic currents have , doubtless , brought the marine Algae ; but the transport of the seeds of the freshwater plants , of the grasses , and of the two plants with hooked and barbed appendages to the fruit , is not apparent in the case of a country that has no land birds but an endemic one ( the Cliionis ) , and of which the water birds come to land only or chiefly at the breeding season , and this after long periods of oceanic life in a most tempestuous ocean .
Even supposing that the sea birds which habitually breed in Kerguelen Island did visit Puegia between the periods of incubation , it is difficult to imagine that any seeds that had adhered to their beaks , feet , or bodies on leaving the latter country would not have been removed by the buffets of winds and waves over upwards of 4,000 miles of ocean .
The supposition that more land formerly existed along the parallels between Fuegia and Kerguelen Island , possibly in the form of islands , remains as the forlorn hope of the botanical geographer .
By such stepping stones the land birds , so numerous in the Falkland Islands ( which lie in the direction of such hypothetical islands ) , and of which the vegetation is identical with that of colder South America , might , favoured by the prevalent westerly gales , have passed from thence to Kerguelen Island , having adhering to them fruits and seeds .
The absence of such birds from the present Avi-fauna of Kerguelen Island offers no obstacle to such a speculation , as such immigrants would on arrival speedily be destroyed by the predatory gull and petrels of the island .
Various phenomena , of very different relative value and nature , but common to the three archipelagos , Kerguelen , the Crozets , and Marion , favour the supposition of these all having been peopled with land plants from South America by means of intermediate tracts of land that have now disappeared ; in other words , that these islands constitute the wrecks of either an ancient continent or an archipelago which formerly extended further westwards , and that their present vegetation consists of the waifs and strays of a mainly Fuegian flora , together with a few survivals of an endemic one .
The extreme southern point of South America , from lat. 52-54 ' and long .
70 ' W. comprising Fuegia , is deflected to the eastward .
Following its general direction , the Falkland Islands group is the first land met with ( in long .
60 ' W. ) ; its vegetation is comparatively rich and exclusively Fuegian ; it has , no doubt , been brought mainly by the land and freshwater birds which abound there , and are identical with Fuegian ones .
South Georgia is the next land met with to the eastward , in long .
35 ' W. and 54 ' S. ; of its vegetation nothing is known except for the scanty observations recorded in Cook 's voyage , which indicate its botanical identity with the Fuegia .
Of Bouvet Island , the assumed position of which is long .
5 ' E. and 54 ' S. , nothing is known ; it was searched for in vain by the Antarctic Expedition in 1843 .
Marion Island is 37 ' E. and 46 ' S. , and the Crozets , in 48 ' E. and 47 ' S. , are respectively about 1,650 and 1,200 miles west of Kerguelen Island , and there is no land intermediate between them .
Now , from such specimens as have been obtained of the vegetation of the first of these islands by Mr. Moseley , * it appears to be almost identical with that of Kerguelen Island ; that is , to be Fuegian with the addition of some of the peculiar Kerguelen Island types , f and the same remark applies to the Crozets , $ facts from which Mr. Moseley has drawn identically the same conclusions as those to which I had arrived thirty-five years previously from a consideration of the Kerguelen Island flora alone .
He says , speaking of Marion Island ( Linn. Journ. XV .
, * Journ. Linn. Soc. XIV .
, 387 , and XV .
, 484 .
t Marion Island contains several Fuegian species not hitherto found in Kerguelen Island , namely , Ranunculus biternatus , Hymenophyllum tunbridgense , and probably a Hierochloe ( the scented grass mentioned by Moseley ) , together with a Cape fern Aspidiu mohrioides and an olenium .
S See Kidder in Bull .
U.S. Not .
Mus. No. 3 , p. 31 .
BOTANY OF KERGUELEN ISLAND.\#151 ; DR. HOOKER .
485 ) , " the occurrence of Fringlea on the island , as also on the Crozets and Kerguelen " Island , point to an ancient land connection between these islands , which the " antiquity and extent of denudation of the lavas would appear to bear out .
It is " difficult to see how such seeds as those of Fringlea could have been transported " from one island to another by birds ; and the seeds seem to be remarkably " perishable ; besides the distinctness of the genus points to a former wide extension " of land on which its progenitors became developed .
The existence of fossil tree " trunks in the Crozets and Kerguelen Island points to similar conditions .
In the Mora Antarctica , I say , p. 220 , referring to the time required for the formation of the innumerable superimposed beds of volcanic rocks , as observed by me in Kerguelen 's Land , and for the growths and destructions of successive forest vegetations that once clothed the island , and are now imbedded in strata at great depths , that this time is sufficient " for the destruction of a large body of land " to the northward of it , of which St. Paul 's and Amsterdam Island may be the " only remains ; or for the subsidence of a chain of mountains running east and " west , of which Prince Edward 's Island , Marion , and the Crozets , are the exposed " peaks .
" And , at p. 240 , when discussing the structural peculiarities of the Fringlea , I say , " However loth we may be to concede to any of our vegetable pro- " ductions an antiquity greater than another , or to this island ( Kerguelen ) a posi- " tion to other lands wholly different from that it now presents , the most casual " inspection of the land where this plant now grows will force one of the two " following conclusions upon the mind , either that it was created after the extinc- " tion of the now buried and for ever lost vegetation , or that it spread over the " island from another and neighbouring region , where it was undisturbed during " the devastation of this , but of whose existence no indication remains."* It remains to indicate the faint traces of relationship which the Kerguelen Island vegetation presents with those of a few other spots of land in a lower latitude , and that might be supposed to share some of its peculiarities .
Of these the nearest are Amsterdam and St. Paul 's Islands , the names of which are often transposed in our best maps ( even in the Admiralty South Polar Chart of 1839 ) .
They lie about 800 * These ideas , suggesting the hypothesis that the existing distribution of plants is dependent on former geographical relations of land and sea , suggested themselves to me during my visit to Kerguelen Island in 1840 .
The first attempt to apply similar views in extenso to the conditions of a botanically well-known country was in the late Professor Edward Forbes ' paper " on the distribution of endemic plants , more espe- " cially those of the British Islands , considered with regard to geological changes .
" " Brit. Assoc. Reports " for 1845 .
" It had , however , been previously enunciated by Lyell , who thus accounted for the identity of the Sicilian animals and plants with those of the surrounding Mediterranean shores .
He supposes these to have " migrated from pre-existing lands , just as the plants and animals of the " Phlaegrean fields have colonised Monte Nuovo since that mountain was'thrown up in the 16th century , " and further on he says , " we are brought therefore to admit the curious result , that the flora and fauna of " the Yal di Noto , and some other mountain regions of Sicily , are of higher antiquity than the country " itself , having not only flourished before the lands were raised from the deep , but even before they were " deposited beneath the waters .
" Principles of Geology , Ed. v. iii .
, p. 444 , \amp ; c. miles to the N.E. of Kerguelen Island , in 78 ' E. long .
; the northernmost , Amsterdam Island , is nearly on the 38th and St. Paul 's on the 39th parallel of latitude , so they both are very little south of the latitude of the Cape of Good Hope .
I have brought together , in a paper published in the Journal of the Linnsean Society ( vol. xiv .
p. 474 ) , all the little that was then known of the^ flora of these islands , which , like Kerguelen , are volcanic .
Their scanty vegetation is on the whole more temperate than antarctic , and approximates to that of S. Africa in containing such genera as , Spartina , and Panthonia .
Their fern flora is very interesting ; one fern only is common to Kerguelen ( Lomaria ctlpina ) , one ( Nephrodium ) is peculiar , though allied to a Mauritian species , and two others ( australe and farcatum ) are natives of the Cape and other countries ; but what is most singular is , that neither the Poly podium vulgar enor Aspidium have been found in either island , though the former is common to the Cape , Marion Island , and Kerguelen 's Land , and the latter to the two first of these localities .
Tristan d'Acunha , in 12 ' W. long , and 37 ' S. lat. , and the adjacent islets called Nightingale and Inaccessible , all nearly in the latitude of Amsterdam Island and the Cape of Good Hope , are the only other islands whose vegetation demands a passing notice here.* Their flora is essentially Euegian , with an admixture of Cape genera , but with none of those characteristics of Kerguelen Island .
Of Cape types , it contains a Pelargonium and an abundance of both the and Spartina of Amsterdam Island , together with species of Oxalis and The Euegian andEalkland Island plants of Tristan d'Acunha and its islets , which have not hitherto been found in the islands south and east of them , are however more numerous than are the Cape genera even , and include Cardamine , Nertera depressa , trum nigrum , var .
rubrum , Lagenopliora Commersoniana , and Apium australe ; and it contains besides the strictly American genus Chevreulia .
Two land birds , both peculiar , are common in the Tristan group , and they possess a water hen , which has a representative in Africa and S. America .
I am not aware whether land birds are found in Amsterdam Island ; if so , they may help to account for the wonderful fact of the Tristan d'Acunha Phylica and Spartina being found also in it , though separated by 3,000 miles of ocean .
In conclusion , I have to state that no trace of the mountain flora of S. Africa has been found in any of the southern groups of islands .
* For the latest account of this group see Moseley in Journ. Linn. Soc. XIV .
, 377 .

