m^td Mt\) ¥iltrnvvr it^ m> X if' I /A y.^'' LM^^0.mLM^±Mo:^^. , :n^9 ^ 7^ / / // Digitized by the Internet Archive in 2010 with funding from Boston Library Consortium IVIember Libraries http://www.archive.org/details/transactionsofli23linn /" THE TRANSACTIONS THE LINNEAN SOCIETY LONDON. VOLUME XXIII. LONDON: PRINTED BY TAYLOR AND FRANCIS, RED LION COURT, FLEET STREET: SOLD AT THE SOCIETY'S APARTMENTS, BURLINGTON-HOUSE ; AND BY LONGMAN, GREEN, LONGMANS, AND ROBERTS, PATERNOSTER- ROW. M.DCCC.LXII. 0 rO/1 // d '7 U ^-"f- CONTENTS. PART I.— 1860. I. Synopsis Ci'escentiacearum : an Enumeration of all the Crescentiaceous Plants at present Tcnown. By Bekthold Seemann, Esq., Fh.D., F.L.S. . . . page 1 II. On the Distribtition of the Tracheae in Insects. By John Lubbock, Esq., F.B.S.^ F.L.S., 8rc 23 III. On some new Species of Musci and Hepaticse in the Herhariwm of Sir W. J. Hooker, collected in Tropical Africa, chiefly by the late Dr. Vogel and Mr. Barter. By William Mitten, Esq., A.L.S 51 IV. Further Researches on Tomopteris onisciformis, Eschscholtz. By William B. Carpenter, M.I)., F.B.S., F.L.S. , F.G.S., and Edouard Claparede, 31. B., Fellow of the Physical Society of Geneva 59 V. Further Observations on the Metamorphosis of Gasteropoda, and the Affinities of certain Genera, with an attempted Natural Bistribution of the principal Families of the Order. By John Denis Macdonald, Assistant Surgeon of S. M. S. "Herald," Captain H. M. Denham, R.N., F.B.S. Communicated by Professor Huxley, F.B.S., F.L.S G9 VI. On Sycopsis. By Daniel Oliver, Esq., F.L.S 83 VII. On tiDo Tuberiform Vegetable Productions from Travancore. By the Rev. M. J. Berkeley, M.A., F.L.S. 91 VIII. Remarks on ^clesoiivim stipitatum. Berk, et Curr., Pachyma Cocos, Fries, and some similar productions. By Prederick Currey, Esq., M.A., F.R.S., F.L.S., and Daniel Hanbury, Esq., F.L.S 93 IV CONTENTS. IX. Ow some New Zealand Verrucarise. By Charles Knight, Esq., F.L.S. . page 99 X. Contributions to the Lichenographia of New Zealand ; being an account, loith figures, of some new Species of Graphiclese amd allied Lichens. By Chas. Knight, Esq., F.L.S. , Auditor-General of New Zealand; and W. Mitten, Esq., A.L.S. . 101 XI. The Nervous System of the Asteridae ; with observations on the Structure of their Organs of Sense, and remarks on the Reproduction of lost Bays. By Henry S. Wilson, II.D., Junior Demonstrator of Anatomy in the University of Edinburgh. Commmiicated by Thomas Anderson, M.D., F.L.S 107 XII. Observations on the Netiration of the Sind Wings of Hymenopterous Insects, and on the Hooks which join the Fore and Hind Wings together in flight. By Miss Staveley. Communicated by Dr. John Edward Gray, F.B.S., F.L.S., V.P.Z.S., ^c 125 XIII. On certain Sensory Organs in Insects, hitherto undescribed. By J. Braxton Hicks, M.B. Land., F.L.S 139 XIV. niustrations of the Floras of the Malayan Archipelago and of Tropical Africa. By Joseph D. Hooker, M.D., B.N, F.B.S., L.S. Sf G.S 155 XV. On some Oceanic Entomostraca collected by Captain Toynbee. By John Lubbock, Esq., F.B.S. Sf L.S. 173 XVI. On the Anatomy and Development of Pyrosoma. By Thomas H. Huxley, Esq., F.B.S. , F.L.S., Sec.G.S., Professor of Nat%iral History in the Government School of Mines 193 PART II.— 1861. XVII. Outlines of the Distribution of Arctic Pla/nts. By Joseph D. Hooker, M.D., F.B.S., Src 251 XVIII. Further Observations on Entozoa, with Experiments. By T. Spencer Cobbold, M.D., F.L.S 349 XIX. On Proliflcation in Flowers, and especially on that Form termed Media/n Prolifi- cation. By Maxwell T. Masters, Esq., F.L.S., Lecturer on Botany at St. George's Hospital 359 CONTENTS. V XX. On the Circulation of the Blood in Pegea, as bearing on the question of a Lining to the Vascular System in the Tunicata in general. By John D. Macdonald, B.N., F.B.S. Commmiicated by George Busk, Esq., F.B.S., Sec. B.S. . . page 371 XXI. On the Physiology of the Pallial Sinuses of the Brachiopoda. By John D. Mac- donald, B.N., F.B.S. Communicated by George Busk, Fsq., F.B.S. , Sec. B.S 373 XXII. On the Nerve proceeding to the Vesicles at the Base of the Halter es, and on the Sub-costal Nermhre hi the Wings of Insects. By J. Braxton Hicks, M.B. Lond., F.L.S. 377 XXIII. On Three Oaks of Palestine. By J. D. Hooker, 3I.B., F.B.S., F.L.S., Sfc. 381 XXIV. On Fissicalyx and Prioria, two recently published Genera of Leguminosse. By George Bentham, Esq., Pres. L.S. 389 XXV. The Foot of the Fly ; its Structure and Action : elucidated by comparison with the Feet of other Insects, Sfc. — Part I. By Tufeen West, Esq., F.B.S. . . 393 PAET III.— 1862. XXVI. Note on the Structure of the Anther. By Daniel Oliver, Esq., F.B.S., Pro- fessor of Botany in University College, London 423 XXVII. Notes on the Thysanura. — Part I. Smynthurid^. By John Lubbock, Esq., F.B.S., F.L.S., 8fc 429 XXVIII. On the Geographical Belations of the Coleoptera of Old Calabar. By Andrew Murray, Esq., F.L.S., Assist. Sec. B. Hort. Soc 449 XXIX. Note on Hamamelis and Loropetalum; with a Description of a new Aniso- ])h.j\].eafrom Malacca. By Daniel Oliver, Esq., F.L.S 457 XXX. On African Anonacese. By George Bentham, Esq., Pres. L.S 468 XXXI. On Prolifioation in Flowers, and especially on that kind termed Axillary Pro- Ufication. By Maxwell T. Masters, Esq., F.L.S., Lecturer on Botany at St. George^ s Hospital 4.8I yi CONTENTS. XXXII. Contr'ibutions to an Insect Fauna of the Amazon Valley. Lepidoptera : Heliconid^. By Henry Walter Bates, Bsq. Communicated by the Secre- tary page 495 XXXIII. Observations on the Oonidia, and Confervoid Filaments of llosses, and on the relation of their Oonidia to those of Lichens and of certain freshwater Algce. By J. Braxton Hicks, 3LB. Bond., F.B.S. Sf L.S 567 XXXIV. Notes on the Thysanura. — Part II. By John Lubbock, Esq., F.B.S. , F.B.S. , FG.S 589 XXXV. On the Specific Identity of the described Forms of Tanalia. By Henry F . Blanford, Bsq. Comnmnicated by Dr. Joseph Hooker, F.B.S., F.B.S. , 8fc. 603 eeeata. Page 41, line 4 from bottom, for ' origin ' read ' organ.' ,, 101-6. In Messrs. Knight and Mitten's " Contributions to the Lichenographia of New Zealand," in all the measurements, for ' from ' and ' to ' read ' breadth' and ' length.' „ 149, line 3 from top, for ' limbs ' read ' body.' In vol. xxii. p. 381-2, in Mr. Knight's paper " On the Bitentaculate Slug, ifec. of New Zealand," the references to Plate LXVI. are wrong. Nos. 6, 7, 8, 11, 12, & 15 are dissections, &c. of the common Black Slug of the Northern Island. ERRATA. Page 41, line 4 from bottom, ybr origin read organ. Page 149, line 3 from top, /or limbs read body. TKANSACTIONS THE LINNEAN SOCIETY. I. Synopsis Crescentiacearum : an JEnumeration of all the Crescentiaceous Plants at present know7i. By Bebthold Seemann, Esq., Ph.D., F.L.S. Read November 17th, 1859. In 1853 I had the honour of laying before the Linnean Society a sliort paper on CrescentiacecB (conf. Proceedings Linn. Soc. ii. p. 268), accompanied by a promise of a more elaborate treatise on the same subject. As it was absolutely necessary to pay visits to Vienna, Berlin, and Paris, in order to compare the herbaria of those places with the notes made on the London Collections, I find myself only now in a position to fulfil my promise. I have seen eleven species of Crescentiacece in a living state, five of them wild in their native countries ; and have besides carefully examined the various herbaria, but regret that these materials are not so complete as could be wished. Of some species there are only single specimens preserved, and the friiit of most species is merely superficially known. The generic characters are therefore in some instances left incomplete, and the limits of a few species await further settlement. That the Crescentiaceous plants, or rather the genera Galea, Phyllarthron, Tancecium, Parmentiera, Crescentia, and Kigelia constitute a closely associated group of the Personatce, needs, as a universally accepted fact, no demonstration, — the views of Jussieti, who regarded them as Solanacece, not having found any modern supporters. But it will be necessary to say a few words respecting the diversity of prevailing opinions as to the affinity and rank they hold in the natural system. It is stUl a matter of debate whether this group should be regarded in the light of a separate order, or merely as a suborder of either Bigno- niacecB or Gesneracece. Bojer (Hort. Maurit. p. 220) placed it with Bignoniaceoi (in which he was followed by Don,DeCandolle,Eenzl*, Miers, and others) ; Endlicher and Miquel with * Fenzl associates three foreign elements with Crescentiacea : viz. Spatkodea, and Periblema, both Bipnoiiiacrre ; and Bravaisia, a true Acanthacea (= Onchtjcanthns Cumingianus, Nees). VOL. XXIII. B / 2 DR. B. SEEMANN ON THE CRESCENTIACEJE. Gemeracece. Gardner (Hook. Journ. ii. p. 424) erected it into a separate natural order, in whicli he was supported by Lindley (Veg. Kingdom, p. 673) and myself (Proc. Linn. Soc. ii. p. 268, and Bot. Herald, p. 181). CrescentiacetB seem to me to stand between CyrtandrecR and Bignoniacece, wbere they are placed by Lindley : they differ from Cyrtcmdrece in the transverse position of their seed ; from Bignoniacece in their indehiscent fruit and wingless seed. Respecting their claim to the rank of a separate natural order, it must be conceded that it depends entirely upon the subjective importance attached to certain characters upon which the various orders composing the great class Bersonatts {Bignoniales, Lindl.) are founded. Por instance, if Orodancliece, in vktue of their spherical pollen, capsular fruit, parietal placentae, anatropal ovule, and albuminous wingless seed, are held to be a separate order of plants, as is done in most systematic works, there is no option but to acknow- ledge the independence of the Crescentiacece, unless indeed the whole of the Bersonatce are looked upon as one great natural order, and Crescentiacece, Accmthacece, Bignonlaoece, Scrophularinece, etc. as merely so many suborders or tribes. In the former case there would be 10 separate natural orders ranged under Bersonatce, the differences of which often rest upon only a single constant character. This will be seen from the following sketch, which wiU also show that, if any of the distinctive characters pointed out be disregarded, several large groups fall together, Personate {Bignoniales et Orobanchece, Lindl.). 5 I. Pl.ace?it(S parietales. Pedaliacem. Pollen .... Ovula anatropa. Drupa vel capsula. Semina erecta v. pendula, aptera? exalbuminosa. Besleriacece. Pollen ellipticum. Ovula anatropa. Bacca. Semina pendula, aptera, albuminosa. Gesneracets. Pollen ellipticum. Ovula anatropa. Capsula. Semina pendula, aptera, albuminosa. Orohanchecs. Pollen sph^ricura. Ovula anatropa. Capsula. Semina aptera, albuminosa. Didymocarpets. Pollen .... Ovula anatropa. Capsula. Semina pendula, plerumque aptera, albuminosa. Cyrtandrece, Pollen .... Ovula anatropa. Bacca. Semina pendula, aptera, exalbuminosa. Cresceniiaceae. Pollen sphaericum. Ovula anatropa. Bacca. Semina transversa, aptera, ex- albuminosa. § II. Placenta axillares. Bignoniacece. Pollen ellipticum. Capsula. Ovula anatropa. Semina transversa, alata, exalbuminosa. Acanthacem. Pollen sphEericum. Ovula amphitropa v. campylotropa. Capsula. Semina erecta, aptera, exalbuminosa. ScrophularinecB, Pollen ellipticum. Ovula anatropa v. amphitropa. Capsula*. Semina erecta, aptera, albuminosa. § III. Placenta centrales. Lentibularia. Pollen ellipticum. Ovula anatropa. Capsula. Semina .... aptera, exalbuminosa. * Miers (Illustrations of South American Plants, i. App. p. 164) has incorporated all the baccate genera, fornierly placed in ScrophularinecB, in his order Atropacece, standing between Scrophularinece and Solanacece. DR. B. SEEMANN ON THE CRESCENTIACE.E. 3 CRESCENTIACE^. Crescentiacece, Gardn. in Hook. Journ. ii. p. 423 (1840); Lindl. Veg. Kingd. p. 673 (1846); Seem, in Proc. Linn. Soc. ii. p. 268 (1853), Bot. Herald, p. 181 (1854). Bignoniacearum et Gesneracearum tribus auct. Char. Obdinis emend. — Arbores \e\frutices stantes vel interdum scandentes, sEepe glabrae, 7'amis plus minusve angulatis. Folia alterna, fasciculata, opposita v. verticillata, petiolata, nunc simplicia, saepissime integeri-ima, nunc composita (trifoliolata vel imparipinnata). Stipules nullae, vel interdum e gemmis axillaribus foliis primariis spurise. Flores hermaphroditi, subregulares vel irregulares, terminales vel axillares, vel saepissime ex trunco aut basi ramorum orti, solitarii, racemosi vel paniculati. Calyx liber, gamophyllus, persistens vel deciduus, 5-merus, spathaceus vel bipartitus. Corolla hypogyna, gamopetala, subcampanulata, infundibuliformis vel hypocraterimorpha, limbo 5- lobo, subaequali vel subbilabiato, lobis per aestivationem duplicato-plicatis vel subplicato-imbricatis. Stamina 4, didynama, cum rudimento quinti, corollag tubo inserta, ejusdem laciniis alterna, exserta vel inclusa. Filamenta simplicia. AnthercB biloculares (abortu uniloculares), loculis discretis. Pollen sphaericum. Discus hypogynus glandulosus, obsolete lobatus, ovarii basin cingens. Ovarium liberum, 1-loculare ; ovula anatropa, indefinita. Stylus terminalis, simplex. Stigma bilobum vel bilamellatum. Fructus baccatus, 1-locularis, vel spurie 2-4- pluri-locularis. Semina plurima, aptera, transversa. Albumen nullum. Embryo rectus vel curvatus. Radicula umbilico proxima, brevis, crassa. Cotyledones plano-convesae. Obs. This definition excludes PeriUema, DeCand. Prod. ix. p. 242 {Boiitonia, De Cand.), a Madagascar genus having a bilocular ovary, and two o-\T.iles in each cell. Its fruit is unknown. It will therefore be much better placed among the genuine Blg- noniacece, — its definite number of ovules showing it to possess some affinity with Flatycarpum, H. et B., and Senriquezia, Spruce (Linn. Trans, xxii. p. 296). Geographical Distribution. The Crescentiaeece are inhabitants of the tropical and subtropical regions of both hemispheres, ranging from 30° S. to 30° X. latitude. They occiu' in the greatest - number in Madagascar, Mauritius, the Seychelles and other islands of Eastern Africa. In America they are represented by 10 species, the most northern of which (Crescentia Cnjete, Linn.) is found in Key West, Elorida. In Asia only two species have as yet been discovered, one of them (Colea Mauritiana, Boj.) having been collected in Timor, and also in Mauritius and Madagascar, the other ( C. tripimiata, Seem.) in Cochin- china. No representative of the order has hitherto been met with either in Eui'oj)e, or the continent of Australia. Numerically we may state that Africa produces 16, America 10, and Asia 2 members of this group of plants.- Considerable additions may, however, still be expected from the islands of Eastern Africa, and necessitate a careful revision of the genuine BignoniacecB and allied orders. Eour species are extensively cultivated in the tropics, viz. Crescentia Gujete, Linn, and C. alata, H. B. K. (both for the sake of their hard-shelled fruit), Colea Telfairice, Boj. (on account of its edible, agreeable-flavoured fruit), and Kigelia pinnata, DeCand. Probably the great geographical range which Cres- centia Cujete, Linn, and Kigelia pinnata, DeCand. enjoy (the former in America, the latter in Africa), is in a great measure owing to their having accompanied man in his wanderings, been planted where new homes were established, and become gradually naturalized, and to all appearance wild, in the localities where we now meet with them. b2 4 DR. B. SEEMANN ON THE CRESCENTIACE^. Properties and Uses. All CrescentiacecB may be termed ornamental plants, the fine foliage, elegant flowers, and curious fruit of which have already procured for theu' order a fair representation in our gardens. We cultivate (1859) Colea floribunda, Boj. ; two species of Fhi/llar- thron (P. Bojerianuni, DeCand. and P. Co??ior^«se, DeCand. = P. Foivrecmtim, DeCand.) ; TancBciimi parasiticum, Swartz ; all the known species of Crescentia, and Kigelia pinnata, DeCand. Farmentiera cereifera, Seem, was in 1845 at Kew, but has since been lost. Several Crescentiacece furnish excellent timber ; and considerable praise is given in this respect to the Kigelia pinnata,DeCa.n(i.., of which canoes, posts and pillars, etc. are made', and which, not only as an umbraculiferous, but also as a sacred tree, is held in high esteem in Africa. Kotschy, speaking of the kingdom of Nubia, says : " On moonlight nights the negroes celebrate their religious festivals under this tree and the Boswellia serrata, Roxb. As soon as the moon rises, they form circles under the oldest trees, and begin to dance, sing, and beat large drums, whilst the women supply them Avith the slightly intoxicating merisa (beer made of Sorghum). These festivals are repeated every month, and extend over several nights, during which time pitchers filled with merisa are placed around the trunks, and some of the same beverage poured on the roots of the trees. As symbols of special veneration, high poles made of Kigelia-wood are erected before the houses of the great chiefs '\" The genus Crescentia has a fruit with a hard woody shell, wliich in Crescentia Cujete, Linn. (=C. cuneifolia, Gardn.), the Calabash-tree of the British colonists, is so large and durable, that it admits of being converted into pails, bottles, pans, cups, sieves, ladles, spoons, and various other household articles \ In Panama I have seen milk-pans made of it, measui'ing thirteen inches across ; and Hum- boldt mentioned to me that during his travels in America, Bonpland and he commonly used one of these vessels as their wash-hand basin. Even the shell of Crescentia alata, H. B. K., the Teconiate of the Mexicans, though much smaller than that of Crescentia Cujete, Linn, is used in Mazatlan and other parts of Mexico as a drinking-cup \ The shell, or rather rind, of Kigelia pinnata, DeCand., after having been hardened by drying, serves as frames for drums in Africa ^ ; and it is not unUkely that the account given by the missionary Knoblecher, of the shell of a fruit found on the White Nile ^ and devoted to the same purpose, refers to it. The fruit of the Palo de velas, the famous Candle-tree of the Isthmus of Panama {Farmentiera cereifera, Seem.), has an apple-like smell, and fattens cattle ', whilst that of the Quauhilote [Farmentiera edulis, DeCand. = Crescentia aculeata, H. B. K.), resembling a cucumber in shape, is eaten by the Mexicans ^ The berry of Tanoscium lilacinum, Seem. (:=Fesleria ? violacea et ^. ?' ccerulea, Aubl.) is also edible ^ ; and the subacid pulp of the fruit of Crescentia Cujete, Linn, aifords food to the negroes'". Tussac believed that the fruit of Crescentia cucurbitina, Linn., a common sea-side shrub of tropical America, contained a deadly poison, and hence he ' Oswald, Mem. in Mus. Kew ; H. Earth in Bonpl. iv. p. 292. " Kotschy in Bonpl. iv. p. 304. " Seemann, Bot. Herald, p. 183. * Seemann in Hooker's Journ. and Kew Misc. vi. p. 276. ' Barth in lit. ad auct. " Kotschy in Bonpl. iv. p. 304. ' Seemann, Bot. Herald, p. 183. " DeCandolle, Prodromus, ix. p. 244; Lindley, Veg. Kingd. p. 674. ' Seemann, Bot. Herald, p. 182. " Lindley, Veg. Kingd. p. 674. DR. B. SEEMANN ON THE CRESCENTIACE^. 5 thought it necessary to warn against it by changing the specific name of the plant into that of Crescentia letUfera. " I know myself," he says, " that some English soldiers in garrison at the Merebalis, who, having foimd the fruit possessed a cucumber-like taste, boiled and ate it, were seized with dreadful colic, and nearly all perished"". As this is the only instance of poisonous properties recorded of this order, the statement must be received with caution ; possibly some fruits of the Manchineel-tree, which generally grows in company with Crescentia cucurbitina, Linn., may have been mixed with those alluded to, and caused the accident. Colea Telfairice, Boj. is extensively cultivated in Madagascar on account of its fleshy fruit, which has an agreeable flavour, and is highly esteemed as an esculent '^ Phyllarthron Bojericmum, DeCand. also yields an edible fruit '^ The natives of Guiana extract a violet colour from the fruit of " Emosse bereoy " {Tanceclmn lilacimmi. Seem.), with which they dye their cotton cloth, their bark and straw furniture ". The juice of the fruit of the common Calabash-tree dyes silk black '■\ The fruit of the " Coco de Mono " of Topo, Venezuela {Crescentia cucurbitina, Linn.), diffuses, when ripe, an agreeable odour, which attracts monkeys, birds, and other animals partial to the fruit '^ The medi- cinal properties of some species are in repute among the natives of various countries, though they have not yet been recognized in our pharmacopoeias. The Philippine Islanders consider a decoction of the leaves of Crescentia alata, H. B. K. (= C. trifolia, Blanco) an effectual remedy for haemoptysis '". The pulp of the fruit of the same species, boiled with sugar, is administered internally by the Mexicans in complaints of the chest (con- sumption ?)'^ and haK a drachm of the root oi Farmentiera eclulis, DeCand., to one pound of water, is considered by them as a remedy for dropsy '". Purgative properties reside in the pulp of Klgelia pinnata, DeCand., of which the Africans avail themselves-"; they are also found in the juice of Crescentia Cnjete, Linn., obtained in Panama by incision of the fruit ^'. The pulp of the fruit of the last-named tree is also used internally in Mexico in inflammatory and bilious diseases -", and employed, like that of TancBcimn albiflorum, DeCand. ", in various countries as poultices. With the fruit of Kigelia pinnata, DeCand. cut in halves, and slightly roasted, the nations of North-eastern Africa rub their skin, as a cure for rheumatic and syphilitic complaints'-''. Its aphrodisiacal properties are also confirmed by A. Richard ". Class-ijicatioii. I have di^dded Crescentiacece into two tribes, the one having a regular, the other an irregular calyx. I retain for them DeCandolle's names {Tanceciece and Crescentiece) and their typical genera, but reject his characters and remove Parmentiera from Tanceciece to Crescentiece, and Tripinnaria from Crescentiece to Tanceciece. No new genera are created ; but two {Schlegelia and Tripinnaria) suppressed, they having proved " Tussac, Fl. des Antilles, iv. p. 51. '= Hooker, Bot. Mag. (1830) t. 2976. '^ Bojer, Mem. in Herb. Vindob. " Aublet, Guiana, ii. p. 631. '" Seemann, Bot. Herald, p. 183. "^ W. Birschell, Mem. in Herb. Hook. ; Bonpl. v. p. 44. " M. Blanco, Fl. de Filipinas, p. 490 (1st edit.). " Seemann in Hook. Journ. and Kew Misc. vi. p. 276. " Heller, Reisen in Mexiko, p. 4)4. ^" Eduard Vogel, Mem. in Herb. Hook. ; Bonplaudia, v. p. 44. -' Seemann, Bot. Herald, p. 183. " Heller, Reisen in Mexiko, p. 414. -' Lindley, Veg. Kingd. p. 674 ; Heller, Reisen in Mexiko, p. 414. ■'* Kotschy in Bonplandia, iv. p. 304. -' A. Richard, Flora Abyssinica, ii. p. 60. 6 DR. B. SEEMANN ON THE CRESCENTIACE^. identical with TcmcBcmm and Colea ; whilst Periblema, DeCand., on account of its bilocnlar ovary and definite number of oviiles, is altogether excluded from the order. Diagnosis Genertim. Tribus I. — Tan^cie^. Calyx persistens, regularis, 5-merus. 1. Colea, Boj. Calyx obsolete 5-dentatus vel 5-partitus, ecostatus. Fnictus carnosus, cylindraccua, spurie 2-locularis. Folia opposlta vel verticillata, imparipinnata. Africa et Asia tropica. 2. Phijllarthi-on,'DeCand. Calyx 5-dentatus, 5-angulato-costatus. Fmctus carnosus, cylindraceus, spurie plurilocularis. Folia verticillata vel sparsa, lomentacea. Africa tropica. 3. Tanmcium, Swartz. Calyx obsolete 5-dentatus, ecostatus. Fructus carnosus, oblongus vel globosus, 1- vel spurie 2-locularis. Folia opposita, simplicia vel trifoliolata. America tropica. Tribus II. — Crescentie^. Calyx deciduus, irregularis (spathaceus vel bipartitus). 4. Parmew^iera, DeCand. Calyx spathaceus. Fructus carnosus, cylindraceus, epulposus. Folia opposita, trifoliolata vel simplicia. America tropica. 5. Crescentia, Linn. Calyx bilabiatus, lobis integerrimis. Fructus lignosus, rotundatus vel oblongus, pulposus. Folia sparsa vel fasciculata, simplicia vel trifoliolata. America tropica. 6. Kigelia, DeCand. Calyx bilabiatus, lobis irregulariter fissis. Fructus corticatus, ellipsoides seu cylindraceus. Folia opposita, imparipinnata. Afi-ica tropica. Tribus I. — Tan.e€ie.e. Seem, in Proc. Linn. Soc. ii. p. 269 (1853) ; Bot. Herald, p. 182 (185-1). — Calyx persistens, regularis, 5-merus. 1. Colea, Bojer. Calyx persistens, subcampanulatus, obsolete 5-dentatus, vel 5-partitus. Co7'olla infundibuliformis, tubo oblongo supra paulum ampliato, limbo 5-fido subaequali, lobis rotundatis patentibus. Stamina 4, didynama, cum rudimento quinti; antherce bi- (vel abortu uni-) loculares, loculis discretis. Discus glandulosus obsolete 5-lobus, ovarii basin cingens. Stylus elongatus ; stigma bilamellatum. Ovarium uniloculare, multioviilatum, placentis parietalibus. Ovula anatropa. Bacca oblonga cylindraceave, styli-apiculata, spurie 2-locularis. Semina exalbuminosa, imbricata, irregulariter ovata, crassiuscula. Cotyledones plano-convexse apice emarginatae ; radiciila brevissima. Frutices, arbusculce vel arbores Africa et Asice trojncce, ssepius glabrae, foliis verticillatis vel oppositis, imparipinnatis (abortu simplicibus), foliolis integerrimis, floribus racemosis vel paniculatis, termi- nalibus vel ex trunco ramisque ortis, flavidis, roseis vel albidis. Colea, Boj. Hort. Maurit. p. 220 (18.37) ; Endl. Gen. Plant. Suppl. i. n. 4171/„ (1840) ; DeCand. Prodr. ix. p. 240 (1845). Bignoniae sp., auct. Tripinna, Lour. Fl. Cochin, (ed. Ulyssip.), p. 391 (17S0). Tripinnaria, Pers. Ench. ii. p. 173 (1807) ; Endl. Gen. PI. n. 4173 (1836-40). Tripinna iripinnata, Lom\ {= Tripinnaria CocUinchinensis, Pers.), has always been numbered among the doubtful genera; and unfortunately there are no specimens of it among Lom^eiro's plants preserved at Paris and London. But I do tliink that in all essential points it agrees with Colea, and may appropriately take its place near Colea TelfairicB, Boj., with which it corresponds in its arboreous habit, terminal panicles, and undulated lobes of the corolla. Loureiro does not state in his description whether his plant has opposite leaves ; but if the systematic position now assigned to it be correct, there is reason to believe tliat, like all other Coleas with terminal flowers, the leaves are opposite. Having thus united the genera Colea and Tripinna, the necessity of adopting DR. B. SEEMANN ON THE CRESCENTIACE^. 7 the oldest name ( Tripinna) would seem forced upon me ; but in yielding to tMs I should be obliged to disturb a whole series of well-established names ; and that I cannot bring myself to do as long as no specimens of Loureiro's doubtful plant have been examined. There is another reason which prompts me to pause. The habit of all Coleas with opposite leaves and terminal flowers is very different from that of the Coleas with verticillate leaves and flowers growing out of the old wood, rendering it probable that these differ- ences may be accompanied by important generic distinctions, which, when the flowers and fruit of all the species shall have become better known, may justify a breaking up of Colea into at least two distinct genera. Loiu'eiro's Tripinna might then be restored, and the name of Colea be restricted to Colea Mauritiana and its allies. Even now it will aid us in the classification, if we divide the genus sectionally into Colece gemiince and Tripinna. I have added four new species, three of which were discovered by recent travellers, whilst one (C. JBojeri) had been misplaced under Bignonia. This will in some measure com- pensate for the reduction of species it has been necessary to make. Colea ramijlora, DeCand. and C. ohhisifolia, DeCand. have been united with C. Ilauritiana, Boj., there being no specific distinction between them. C.floribunda, Boj., like all Coleas established by Bojer, proves a good species, of which C. canlijlora, DeOand. and C. ? Commersonu are undoubtedly synonyms. C. purpu7^ascens. Seem, seems a very distinct species : I must, however, own that I entertain doubts whether it may not fall, together with C. Bojeri {Bignonia Bojeri, DeCand.), of which I have seen no specimens, and which is btit imper- fectly described. C. Seychellarum, Seem, and C. Mspidissima, Seem, are founded upon very distinctive characters, and will probably stand. C. telragona, DeCand. is also a good species, the branches of which are, however, apt to have more than four sides. C. decora is perhaps the most variable species, the leaves being simple, trifoliated, and imparipinnate, whilst the flowers (normally placed in simple racemes) are often quite isolated on a bracteate rachis. Owing to this extreme variation, the species has seldom been recognized, and a host (jf synonyms has been the consequence. Colea nitida, DeCand., C. Chapelieri, DeCand., Bignonia racemosa, Lam., and B. compressa, Lam., are the names iinder which it appears in our systematic works. C. Telfairice, Boj. and C. involucrata, Boj. are also good species, to the latter of which Bignonia bracteosa, DeCand. must be added as a synonym. § I. Folia verticillata. Flores e trunco ramisque orii (Colese genuine). 1. Colea FLOMBTTNDA ; arbuscula; ramis obtuse angulatis glabris, foliis 4-5-no-verticil- latis 5-8-jugis cum imparl, foliolis petiolulatis elliptico-oblongis vel oblongo-lanceo- latis acuminatis, utrinque glabris, racemis e caule ramisque ortis, calyce glabro, corolla (lutea) extus velutino-scabrida, intus villosa, staminibus ovario stylo fructuqu.e elongato cylindraceo subtoruloso glabris. (v. v. cult, et sp. s.) Colea floribunda, Bojer, Hort. Maurit. p. 220 (1837) ; Lindl. Bot. Reg. 1841, t. 19; DeCand. Prodr. ix. p. 241, n. 3 (1845). Colea} Commersonii, DeCand. Prodr. ix. p. 242, n. 11 (1845). Colea cauliflora, DeCand. Prodr. ix. p. 241, n. 3 (1845). Bignonia cauliflora, Sie.b. Fl. Maurit. exsic. ii. n. 284. Nomen vernaculum Madagascariense : " Rei-rei," 8 DR. B. SEEMANN ON THE CRESCENTIACE^. Geogr. Distr. Common throughout Madagascar [Commerson ! in Herb. Par., et Juss. ! Boje?- !) ; Mau- ritius [Sieber ! in Herb. Vindob.). This is the only species of Colea as yet cultivated in our gardens. It is better known than aU the others ; and we are therefore enabled to settle its synonymy with tolerable accuracy. I have seen, at Paris, Commerson' s specimens upon which DeCandoUe founded his Colea ? CommersonU, and consider them to be in every respect identical with C. flori- bimda. Why DeCandolle placed C. CommersonU among the species with terminal flowers is diificult to say, since some of the specimens prove that the racemes grow out of the old wood. Sieber's n. 284 from Mauritius, upon which DeCandolle founded his C. cauli- flora, I have examined in the Vienna Herbarium, and I find that it cannot be specifically separated from C. Jloribimda. 2. Colea Setchellarum ; arbuscula; ramis angulatis giabris, foliis 3- 4i-nove verticillatis 4-5-jugis cum imparl, foliolis petiolulatis obovato-oblongis obtusis, basi attenuatis, coriaceis utrinque giabris, racemis elongatis e caule ramisque ortis, calyce glabro, coroUa extus glabra, intus villoso-pubescente, ovario styloque glabro, fructu .... (v. sp. s.) Colea Seychellarum, Seem. MSS. in Herb. Paris. Geogr. Distr. Isle of Mahe {Bernier\, n. 38). This new species has the habit of Colea jloribunda, Boj. It is, according to Bernier, from 20-25 feet high. Leaves 1-1| foot long ; leaflets 4-5 inches long and 1^-2 inches broad. Racemes often 8 inches, and corolla about 1 inch long. At once distinguished from C. jloribunda by its very long racemes and its (outside) glabrous corolla. 3. Colea purpurascens ; . . . . , ramis angulatis pubescentibus, foliis 3-7-no verticil- latis 7-10-jugis cum imparl, foliolis petiolulatis ovato-oblongis longe acuminatis basi attenuatis, petiolis, petiolulis nervisque inferioribus purpurascentibus, racemis abbre- viatis e caule ramisque ortis, calyce glabro, corolla extus velutino-scabrida, stylo superne viUoso, fructu .... (v. sp. s.) Colea jiurpurascens. Seem. MSS. in Herb. Paris. Geogr. Distr. Ste. Marie de Madagascar {Boivin\) ; Nossi-be (Boivin !). Has the habit of Colea Jloribunda, but diifers in the pubescent branches, and pseudo- stipules, the pui'plish tinge of the petioles, petiolules, and veins of the under side of the leaflets, as vrell as in its villose style. It flowers, according to Boivin, from November to January. 4. Colea discolor ; fruticosa ; ramis (angulatis ?) apice hirsutis ; foliis (verticillatis ?) 7-jugis cum imparl, foHolis brevi-petiolulatis oblongo-lanceolatis obtuse acuminatis subtus purpureis, racemis e caule ramisque ortis, calyce . . . , coroUa . . . , ovario . . . , stylo . . . , fructu . . . Colea discolor, Seem. MSS. Bignonia discolor, Boj. in Ut. ad DeCand. 1833, non Rich. Bignonia Bojeri, DeCand. Prodr. ix. p. 165 (1845). Geogr. Distr. Madagascar, on Mount Antoungoun, prov. of Emirna [Bojer, teste DeCand.). DR. B. SEEMANN ON THE CRESCENTIACE^. 9 I have not seen specimens of this species ; and those seen by DeCandoUe were without flowers and fruit. My reasons for placing it among the genuine Coleas were twofold : — Because no true Blgnonia has imparipinnate leaves with quite entire leaflets, nor flowers growing out of the old wood, as those of Colea discolor, according to Bojer, do : these three characters combined are only met with in Colea, and in no other Bignoniaceous or Crescentiaceous genus. As an additional reason, I may plead the close resemblance this species must bear to Colea ptirpiirascens, rendering a specific distinction difiicult. Both have hairy branches, leaves with 7 pairs, and leaflets more or less purplish underneath. Should a larger number of specimens come to hand, it may be found that there is no specific distinction between Colea imrpurascens and C. discolor. 5. Colea hispidissima ; fruticosa ?, ramis subteretibus glabris, foliis oppositis (v. verti- cillatis ?) 4-5-jugis cum imparl, foliolis longe petiolulatis obovato-oblongis breviter acuminatis, racemis abbreviatis e caule ramisque ortis, calyce coroUaque extus pilis ferrugineis hispidissimis, corolla (alliida ?) intus ovario styloque hispidis, staminibus glabris, fructu .... (v. sp. s.) Colea hispidissima, Seena. MSS. Geogr. Distr. Ste. Marie, Madagascar {Boivin\ n. 1820 in Herb. Par.). Differs from all known Coleas by its very hispid calyx and corolla. The leaves are probably, in the older branches, verticillate, as in all the other species of this group of the genus. 6. Colea Mauritlana; arborea; ramis angulatis glabris, foUis 3-4-nove verticillatis 2-3-jugis cum impari, foliolis subsessilibus obovato-cvxneatis obtusis vel acuminatis coriaceis, racemis e caule ramisque ortis, calyce glabro, corolla (rosea) glabra vel tenuissime puberula, staminibus ovario styloque glabris, fructu oblongo acuminate verrucoso glabro (flavescente), seminibus ovali-oblongis. (v. sp. s.) Colea Mauritiana, Bojer, Hort. Maurit. p. 220 (1837) ; DeCand. Prodr. ix. p. 241, n. 1 (1845). Colea ramiflora, DeCand. Prodr. ix. p. 241 (1845). Bignonia Colei, Boj. MSS. ; Hook. Bot. Mag. t. 281? (1828). Blgnonia ramiflora, Decaisne, Nouv. Ann. Mus. Par. iii. p. 381. Colea obtusifolia, DeCand. Prodr. ix. p. 241 (1845). Geogr. Distr. Madagascar {Chapelier\ Petit-Thoiiarsl Co7nmerson \) ; Mauritius (5o;"er !) ; Timor [Herb. Par.). A tree, according to Bojer's MSS. in the Viennese Herbarium, and having Colea tetra- gona for its nearest ally. The leaflets of Bojer's specimen from Mauritius are quite obtuse. 7. Colea tetragona; fruticosa; ramis acute 4-7-gonis, glabris, foliis 4-7-nove ver- ticillatis 5-9-jugis cum impari, foliolis brevi petiolulatis oblongis obtusis basi in- sequalibus subcuneatis coriaceis, racemis e caule ramisque ortis abbreviatis, calyce glabro, corolla (rosea) extus puberula, ovario . . ., fructu ... (v. sp. s.) Colea} telragona, DeCand. Prodr. ix. p. 242 (1845). Geogr. Distr. Madagascar (Pe^i<-77<0Mars ! Pervillel n. 234). VOL. XXIII. C 10 DR. B. SEEMANN ON THE CRESCENTIACE^E. According to Perville, who found tliis species at St. Marie, in the N.E. of Madagascar, it is a shrub 12-15 feet liigh, with rose-coloured flowers. The speciiic name " tetragona " does not apply well to this plant, since better specimens than those at DeCandolle's disposal show that the normal number of the leaves, and the consequent edges of the branches, is seven, and that only weak specimens have square branches. The hard, acute, persistent scales occasionally met with in this species, and regarded by Bojer as stipules, appear to me (as supposed by DeCandolle) to be abortive leaves, somewhat analogous to the scales on the trunk of Cycas. § II. Fulia opposita. Flores axillares vel terminales. (Tripinna, Lour.) 8. CoLEA DECORA ; fruticosa, glabra, ramis compressis, foliis oppositis 5-jugis cum impari, trifoliolatis vel imo simplicibus, foliolis petiolulatis obovato-oblongis obtusis vel acu- minatis, basi subcuneatis, coriaceis, supra nitidis, petiolo angustissime marginato, racemis terminalibus vel asillaribus bracteatis, bracteis obovatis acutis vel obtusis basi cuneatis, pedunculis acute tetragonis, pedicellis solitariis vel 4-no verticillatis, calyce glabro, corolla (alba) staminibusque puberulis, ovario styloque glabris, fructu ... (v. sp. s.) Colea decora, Bojer, Hort. Mauvit. p. 220 (1837) ; DeCand. Prodr. ix. p. 241. n. 8. (1845). Coleal Chapelieri, DeCand. Prodr. ix. p. 242 (1845). Colea} nitida, DeCand. Prodr. ix. p. 242 (1845). Bignonia compressa, Lam. Diet. i. p. 424 (1783). Bignonia racemosa, Lam. Diet. i. p. 424 (1783) ; DeCand. Prodr. ix. p. 166. n. 139 (1845). Geogr.Distr. Madagascar (iJoJe?- ! Goudotl Commerson\ Chapelier\ Petit-Thouars\ Boivinl). Goudot col- lected it near Tamatave, and Boivin at Ste. Marie. Easily distinguished from Colea tetragona, its nearest ally, by its opposite, never verti- cillate leaves, and from all other Coleas by its racemes clothed with foliaceous bracts. By its bracts it approaches Colea involucrata, Boj. It is necessary to add, that Bojer distributed specimens of Colea floribiinda, Boj. under the name of Bignonia decora, which has given rise to some confusion, leading in some herbaria to the error of attaching the name of Colea decora to the wrong plant. 9. Colea involucrata ; fruticosa ; ramis teretibus glabris, foliis oppositis 3-G-jugis cum impari, foliolis petiolulatis ellipticis obtuse acuminatis, basi cuneatis, supra nitidis, racemis abbreviatis (subcymosis) longe peduuculatis, 3-5-floris, bracteis 4 late obo- vatis vel suborbiculatis cinctis, calyce glabro, dentibus obtusis, corolla (rubro-aurantia, marguie flavo-aurantio) limbo patente obtuse 5-lobo, extus intusque glabra, stami- nibus ovario styloque glabris, fructu (v. sp. s.) Colea involucrata, Bojer, Hort. Maurit. p. 221 (1837); DeCand. Prodr. ix. p. 242 (1845). Bignonia bracteosa, DeCand. Prodr. ix. p. 165 (1845). Nomen vernaculum : " Slfin-souki-mena-felan," teste Bojer. Geogr. Distr. Madagascar {Bojer \ in Herb. Par. et Hook.). The fruit of this species, though unknown to both DeCandoUe and myself, was pro- bably seen by Bojer, and induced him to place the plant with his genus Colea, in which DR. B. SEEMANN ON THE CRESCENTIACE^. 11 DeCandolle followed him ; but, probably by some oversight, the celebrated author of the ' Prodromus ' enumerated it a second time in his great work as Bignonia bracteosa, with which genus (if the fruit be fleshy) it can have nothing to do. The imparipinnate leaves with quite entire leaflets are also a feature not met with in any genuine species of Bignonia. 10. CoLEA Telfaiui^ ; arborea ; ramis subcompressis, minutissime pubescentibus viscosis ; foliis oppositis 6~9-jugis cum imparl, foliolis ovato-oblongis acuminatis lucidis, pani- culis terminalibus laxis, calycis dentibus puberulis longiusculis acutis, corolla (rosea, tubo ad faucem flavo fasciato) extus minute puberula, intus subglabra, lobis crenatis, staminibus ovario styloque glabris, fructu longius cylindraceo acuminato carnoso sublsevi. (v. sp. s.) Colea Telfairice, Bojer, Hort. Maurit. p. 220 (1837) ; DeCand. Prodr. ix. p. 242. n. 10. (1845). Bignonia Telfairice, Boj. MSS., in Hook. Bot. Mag. t. 2976 exclud. fig. 2 (1830). Geogr. Distr. Province of Emirna, Madagascar (Bojer !), where it is also cultivated on account of its edible fruit. Nomen vernaculum Madagascariense : teste Bojer, " Vouen-Kici-Kicika." 11. Colea tripinnata ; arborea ; ramis patentibus, foliis (oppositis ?) 3-jugis cum imparl, foliolis (petiolulatis ?) ovatis acuminatis glabris, paniculis terminalibus, calyce cyathi- formi 5-crenato, corolla (luteo-rubra) subcampanulata, laciniis ovatis undulatis villosis, ovario subrotundo styloque . . . ., fructu ovato, seminibus oblongis compressis. Colea tripinnata, Seem, in Bonplandia, iv. p. 128 (1856); Hook. Journ. of Bot. and Kew Misc. ix. p. 84 (1857). Tripinna tripinnata, Lour. Fl. Cochinch. (ed. Ulyssipol.) p. 391 (1790). Tripinnaria Asiatica, Sprengl. Syst. ii. p. 842 (1822). Tripinnaria Cochinchinensis, Pers. Ench-. ii. p. 173 (1807); DeCand. Prodr. ix. p. 248 (1845). Tanacium tripinna, Raeusch, ex Steudl. Nomen. Bot. p. 712 (1841). Nomen vernaculum Cochinchinense : " A Cay Den." Geogr. Distr. Cochinchina (teste Loureiro). As there are no specimens of this plant in Europe to supply the deficiencies of Lou- reiro's character, it may not be superfluous to add here all that the ' Hora Cochinchi- nensis ' contains about it :■ — " Cal. Perianthium cyathiforme, 5-crenatum, persistens. Cor. monopetala, subcampanulata, patens ; tubo calycem superante, limbo 5-fldo ; laciniis ovatis undulatis, villosis, summa majore. Stam. FUamenta 4 (quorum duo longiora) coroUae subsequalia. Antheroi bicornes, incumbentes. Fist. Germen subrotundum. Stylus soqualis staminibus. Stigma acutum, bifidum. Feric. Bacca ovata, carnosa, 1-locularis, polysperma. Sem. oblonga, compressa, pauca*. Nom. Tripinnam voeavi a foliis tripinnatis. — Tripinna tripinnata ; fohis 3-pinnatis, foliolis ovatis acuminatis ; racemis terminalibus. Hab. Arbor magna, ramis patentibus. Folia tripinnata t cum imparl * "pauca" is probably a misprint for "parva," as just before the berry is termed " POLYspermous" by Loureiro, and as the seeds of all Coleas are small. — B. S. f The term "tripinnata" here employed I take to mean "3-juga." — B. S. c2 12 DR. B. SEEMANN ON THE CRESCENTIACE^. maiove, foliolis ovatis acuminatis, integerrimis, glabris. Flos liiteo-ruber, corymbis race- mosis terminalibus. Habitat in sylvis montanis, Codiincbinse." 2. Phtllarthron, DeCand. Calya; persistens, 5-angulatus, 5-dentatus. Corolla infuadibuliformis, tubo oblongo supra paulum ampllato, limbo 5-fido subjequali, lobis rotundatis patentibus. Stamina 4, didynama, cum rudimento quinti ; anthercB biloculares, loculis discretis. Discus glandulosus obsolete 5-lobus, ovarii basin cingens. Stylus elongatus, stigma bilamellatum. Ovarium uniloculare, multiovulatum, placentis parietalibus. Ovula anatropa. Bacca cylindracea, spurie plurilocularis. Semina exalbuminosa, subrotunda. Arbores vel frutices ex insulis Africa austr alls orta;, ramis dichotomis compressis vel angulatis, foliis oppo- sitis verticillatis vel sparsis, lomentaceis (nempe petiole aiticulato articulis 2-5 late foliaceis), foliolis nuUis (v. in sp. dubio paucis), racemis terminalibus vel axillaribus paucifloris, floribus roseis, fruc- tibus flavis. Phyllarthron, DeCand. Prodr. ix. p. 243; Endl. Gen. Plant. Suppl. ii. p. 65. n. 4171/,. Arthroi^liyllum, Bojer, Hort. Maurit. p. 221 ; DeCand. Kev. Bign. p. 18, non Blum. Bignonixe sp., auct. The number of species here enumerated is the same as that of DeCandoUe's ' Prodro- mus,' one of DeCandoUe's species {Fh. Foivrecmum, DeCand.) having been suppressed, and a new one {Fh. Fernencmum, Seem.) added. Fh. Thouarsiauum, DeCand. will pro- bably have to be excluded from this genus, its habit being entirely different from that of the genuine Phyllarthrons. 1. Phyllarthron Noronhianum ; fruticosa, ramis angulatis, saepe trigonis, foHis 3-4- nove verticillatis vel sparsis, petiolis articulatis, articulis 2-3 ample foliaceo-margi- natis late obovatis basi cuneatis, supra lucidis subtus eveniis vel venis impressis, coriaceis, foliolis nuUis, racemis terminalibus, pedunciilis compressis, calyce cylin- draceo 5-nervio obtuse 5-deutato, corolla . . . ., staminibus . . . ., ovario styloque . . . ., fructu .... (v. sp. s.) Phyllarthron Noronhianum, DeCand. Prodr. ix. p. 243. n. 1 (1845). Bignonia articulata, Desf. in Poiv. Suppl. i. p. 636 ? Geogr.Distr. Madagascar {Boivinl Petit-Thouars \ Poivrel in Herb. Par.). 2. Phyllarthron Bojerianum ; arborea ; ramis ancipitibus petioHsque junioribus vis- cosis ; foliis oppositis, petiolis articulatis, articulis 2 late marginatis, inferiorc obovato- cuneato, super, elliptico, utroque pinnatim venoso, foliolis nuUis ; racemis termina- libus subcorymboso-trichotomis, pedunculis compressis, calyce ovato-campanulato enervio subtruncato obtuse 5-dentato, corolla (rosea) extus pulverulento-velutina, lobis 5 latis obtusissimis, staminibus , ovario styloque glabris, fructu (v. sp. s. et cult.) Phyllarthron Bojerianum, DeCand. Prodr. ix. p. 243. n. 3 (1845) ; Hook. Bot. Mag. t. 4173 (1845) ; Van Houtte, Fl. des Ser. ii. t. 7 (Mars 1846). Arthrophyllum Madagascariense, Boj. Hort. Maur. p. 221 excl. syn. (1837). Nomen vernaculum Madagascariense : " Zahane," teste Bojer. Geogr. Distr. Madagascar {Bojer \ Boivinl). Cultivated in Europe, having been introduced about 1840. DR. B. SEEMANN ON THE CRESCENTIACE^. 13 " Arbor excelsa," teste Bojer. Easily distinguished from the allied species by its vis- cous compressed branches and opposite leaves, as well as by its arboreous habit. Fruit edible, according to Bojer. Branches dichotomous. Racemes simple. Bracts opposite, ovate, acute. CoroUa witli five blunt, vravy segments. Glandular disc 5-tootlied. Flowers in August. According to information kindly furnished by Mr. Duncan in Mauritius, the fruit of this and the following species is whitish, and as long as a little finger. 3. Phyllartheon Comorexse; fruticosa; ramis acute 3-i-gonis; foliis 3-4-no verticil- latis vel oppositis, petiolis articulatis, articulis 3-5 longe cuneatis, superne obcordato- emarginatis, lateraliter prseter nervum medium parce venosis, foliolis nullis, racemis abbreviatis paucifloris, pedunculis teretiusculis, calyce tubuloso 5-nervio acute 5- den- tato lepidoto-punctulato puberulo, corolla (rosea) extus puberula, intus . . . ., staminibus . . . ., ovario styloque . . . ., fructu " Capsici formam a^mulante." (v. sp. s. et v. cult.). Phyllarthron Comorense, DeCand. Prodr. ix. p. 224 (1845). Phyllarthron Poivreaniim, DeCand. Prodr. ix. p. 243 (1845). Phyllarthron Bojerianmn, var.? /S elongatuni, DeCand. Prodr. ix. p. 244 (1845). Arthrophyllum Comorense, Bojer, Hort. Maurit. p. 221 (1837). Nomen vernaculuna : " Taumounann," teste DeCand. Geogr. Distr. Madagascar {Poivre\ Eowi?i\); Comoro Isles [Bojer\ Boivinl Grey I). Cultivated in Europe. This species, according to Boivin " a shrub 6 feet high," has the longest leaves of the whole genus, some of them measuring 12-14 inches, and having petioles normally com- posed of five joints. The width of these joints is very variable, and imparts a different look to different specimens. Hence the many synonyms of this comparatively new plant. P. Poivreamim, DeCand. is one of the broad-jointed forms, but not specifically different from the genuine P. Comorense. JP. Bojerlanum, var. ? elongatum, DeCand. is the same ; for the " beard " sometimes observable at the junction of the petioles is simply a fungus that establishes itself upon some specimens. 4. Phyllarthron Bernierianum ; fruticosa, ramis compressis glabris, foliis oppositis petiolis articulatis, articulis 2-5 anguste linearibus apice obcordato-emarginatis, nervis subparallelis, foliolis nullis, racemis terminalibus vel axillaribus usque ad 12-floris, pedunculis compressis, pedicellis oppositis, calyce glabro acute 5-dentato, corolla (rosea) extus pulverulento-velutina, staminibus . . . ., ovario styloque ........ fructu ... (v. sp. s.) Phyllarthron Bernierianum, Seem, in Bonplandia, vii. p. 224 (1859). Geogr. Distr. Madagascar {Bernierl coll. ii. n. 215); Ambongo {Pe)'ville\); Diego Suarez {Bender \). According to Bernier, a shrub 6 feet high, with rose-coloured flowers. Leaves narrower than any other known species of JPhyllarthron, 2" broad, and from 4-5 inches long. Species cliibia. 5. Phyllarthron ? Thotjarsiantjm ; fruticosa, ramis subteretibus foliisque nascentibus lepidotis demum glabris, foliis (siccitate utrinque cujireis) alternis, petiolo artieulato, articulis 2 oblongo-cuneatis sequalibus, foliolis abrupte bijugis lanceolatis utrinque attenuatis alisque petioli pinnatim venosis, racemis . . . ., fructu . . . (v. sp. s.) 14 DR. B. SEEMANN ON THE CRESCENTIACE^. PhyUarthron Thouamiammi, DeCand. Prodr. is. p. 244 (1845). Geoffr. Disir. Madagascar {Peiit-Thouarsl in Herb. Juss. et Par.). There is no eAddence to show that this plant is a genuine species of JPhyllarthron, or even a Crescentiacea, as neither the flowers nor the fruit of it are known. " Polia abrupte pinnata " only occur once among Crescentiacece, and in that instance {Kigelia pin- nata, DeCand.) by abortion. Amongst the genuine Bignoniacece such leaves are known in only one imperfectly described species, Bignonia hijuga, Yahl, which, like Bh. Thouar- sianum, is a native of Madagascar, and may possibly prove identical with it. 3. Tan.ecitjm, Swartz. Calyx persistens, globoso-cylindraceus, obsolete 5-dentatus. Corolla tubulosa, infundibuliformis, limbo 5-fido subasquali, lobis tubo brevioribus, 2 superioribus erectis approximatis paulum minoribus, inferioribus patentibus. Stamina 4, didynama, cum rudimento quinti ; antherm 2-loculares, loculis divergentibus. Discus glandulosus ovarii basin cingens. Stylus elongatus ; stigma bilamellatum. Ovarium uniloculare, raultiovulatum, placentis parietalibus. Bacca oblonga vel globosa, cortice fragiliter corticosa, 1- vel spuria 2-locularis. Semina plurima, compressa aut angulata, in pulpa nidu- lantia; albumen nullum. Frutices America tropica scandentes, ssepe radicantes ; foliis oppositis petiolatis simplicibus integerrimis, trifoliolatis vel germinatis cum cirrho intermedio; foliolis petiolulatis integerrimis; floribus race- mosis vel paniculatis axillaribus et terminalibus bracteatis bracteolatisque, corollis albis, coccineis, roseis, violaceis vel lilacinis ; baccis ferrugineis vel nigris. Tanacium, Swartz, Prodr. ii. p. 9; Fl. Ind. Occ. ii. p. 1049 ; Endl. Gen. Plant, n. 4172 ; DeCand. Prodr. ix. p. 245 ; Seem, in Bonpl. iv. p. 126, et in Hook. Journ. and Kew Misc. ix. p. 82. Schlegelia, Miquel in Bot. Zeit. ii. p. 788. BeslericB sp., Aubl. Citharexylon sp., Linn. Amongst the various species from time to time referred to this genus, there are three which must be excluded. Tanceciuni ? pcmiculatum, Sieb. is a Bignoniacea proper (Arra- biclcea paniculcda, Seem. *) ; Tcmceckt/m pinnatum, Willd. is identical with Kigelia pinnata, IDeCand. ; and T. tripinna, Raeusch, a synonym of Colea tripinnata, Seem. Having added two new ones {T. crucigerum, Seem, and T. lilaciniim. Seem.), the genus now con- sists of four well-defined species, which arrange themselves into two very natural groups, — the one having non-rooting branches, compound ecoriaceous leaves, and white pubescent corollas ; the other rooting branches, simple coriaceous leaves, and glabrous corollas of a scarlet, pink, or more or less bluish tint. The former is represented by T. crucigerum and albijlorimi, the latter by T. parasiticmn and Ulacimmi. T. cruciget^nm is the old Bignonia cmcigera of Linnseus, well figured in Burmann's edition of Plumier's work ; T. albifloriim, DeCand. is the T. Jaroba of Swartz, the alteration of its name having become neces- * Arrabida-a panindata, Seem. MSS. {Tanceciuml paniculafum, Sieb.! Flor. Martinic. n. 81 ; DeCand. Prodr. ix. p. 245 : 1845) ! ramis teretibus glabris, foliis oppositis petiolatis 3-foliolatis, foliolo medio ovali maximo, lateralibus minimis, junioribiis velutinis adultis glabris, paniculis axillaribus laxe trichotomis, pedunculis ramulisque compressis, minutissime puberulis, calyce cupnliformi, breviter 5-deiitato, sequali, corolla apice pulverulento-velutina ; fructus .... (v. sp. s. in Herb. Vindob.) Martinica (Sieb. ! Fl. Martin, n. 81). DR. B. SEEMANN ON THE CRESCENTIACE^. 15 sary, as tlie " Jaroba'' of Marcgrav, from which the older specific name was borrowed, is not, as Swartz supposed, a synonym of this plant, but, judging from the indifferent de- scription and the i*ude woodcut, a Cucurbitacea'^ . T. parasiticum is a very characteristic species (the only one hitherto introduced in our gardens), which has fortunately no con- flicting synonymy. T. Ulacinmn was originally founded by me, upon Miquel's Schlegelia lilac'inu, adding as synonyms Schlegelia elongata of the same author, and the doubtful Besler'ia ? violacea and B. ? ccendea of Aublet, original specimens of which are preserved at the British Museum. Recently I discovered that the plant described by Linnaeus in his ' Plantse Surinamenses ' as Citharexylon cinereuni (five specimens of which are pre- served under that name in Linnseus's own Herbarium) must also be referred to this species. Sir J. E. Smith had already noticed, in a memorandum to the Linnean specimens, that they could not belong to the Citharexylon cinereum of the ' Systema,' which was taken up, it would seem, from Jacquin's, Plumier's, and Pluckenet's figures, and is a terrestrial tree inhabiting the A¥est Indies, to which the brief character {^'^ parasiticnm, scandens") of the 'Plant. Sm'inam.' cannot possibly apply. § I. Caules ramique eradicantes ; folia trifoliolata vel conjugato-bifoliolata cum cirrho intermedio ; foliola ecoriacea. Corolla pubescens, alba. 1. Tan^citjm cuucigbrum ; ramis punctatis, foliis trifoliolatis vel conjugato-bifoliolatis cum cirrho intermedio (abortu unifoliolatis), foliolis ovatis acuminatis, supra glabris, subtus pubescenti-velutinis, racemis axillaribus terminalibusque 5-8-floris, bracteis ovato-lanceolatis, bracteolis subulatis, corolla (alba) longe tubulosa pubescente, bacca oblongaf. (v. sp. s.) TancBcium crucigerum. Seem, in Bonpl. iv. p. 127, et in Hook. Journ. and Kew Misc. ix. p. 83. Bignonia ci-ucigera, Linn. spec. 869 (excl. syn. omnib. except. Plum.) ; DeCand. Prodr. ix. p. 152. Bignonia foliis conjugatis cirrhosis ternaiis, Plum., ed. Burm. fasc. iii. p. 48, t. 58 (exclud. syn. omnib. ex- cept. Linn.). Geogr. Distr. Dominica [Imray !) ; St. Vincent {Guilding \) . There are no specimens of this plant preserved in Linnaeus' s herbarium; and Linnaeus seems to have taken it up solely from Plumier's figures. 2. TANiECiTJM ALBiFLORTJM ; ramis eptinctatis, foliis trifoliolatis vel conjugato-bifoliolatis cum cirrho intermedio, foliolis ovato-oblongis acuminatis utrinque glabris, racemis axillaribus 3-5-floris, bracteis . . . ., bracteolis subulatis, corolla (alba) longissima tubulosa pubescente, bacca oblonga maxima glabra, (v. sp. s.) * The description inMarcg. Hist. Reg. Nat. Brasil. lib. i. p. 25, runs as follows : — " Jarola Brasiliensibus dicta, Va- saca amargosa Lusitanis. — Altissimas arbores beec planta asceudit, caule lento, tereti, qui hinc inde in longis pedi- culis tria folia opposita habet Phaseoli modo, sunt autem plane similia foliis JNIucuna. Fructus autem similis fructui Cuete, sed ut plurimum minor, eadem pulpa, iisdem seminibus, ejusdem quoque usus." — Marcgr. Hist. Reg. Kat. Brasil. lib. i. p. 25. The figure represents a climbing stem, with alternate, tripartite or trifoliolated leaves, and obo- vate, dotted fruits. ■\ "Fructus optime reprsesentat capsulam vulgi, quam tabaco rcplent, secumque portant, uti notatur in 'Hort, Cliff.,' " Plum., ed. Burm. fasc. iii. p. 48. 16 DR. B. SEEMANN ON THE CRESCENTlACEiE. TaTUEcimn albiflorum, DeCand. Prodr. ix. p. 245 ; Seem, in Bonpl. iv. p. 12/, et in Hook. Journ. et Kew Misc. ix. p. 83. Tanacium Jaroba, Swartz. Fl. Ind. Occ. ii. p. 1050, t. 20. fig. 1 (exclud. syn. Marcgr. et idee nomine). CucurbiHfera, Sloane, Hist. ii. p. 173 ; Browne, Jamaica, p. 266. n. 6. Geogr. Distr. Jamaica {Robins ! Distan ! Wright ! Swartz !) ; British Guiana {Schomburgk ! n. 829) ; woods near Crato, prov. Ceara, Brazil {Gardner \ n. 1765). § II. Caules ramique radicantes, folia simplicia coriacea. Corolla glabra, lilacina, violacea, carulescens, rosea vel coccinea. (Schlegelia, Miq. !) 3. TANiECiUM PARASITICUM ; foiiis ovato-oblongis acuminatis iitrinque glabris, racemis axillaribus 2-12-floris, corolla (coccinea) glabra, tubo abbreviato, bacca globosa, glabra (ferriiginea), seminibus angvilato-oblongis parvis. (v. sp. s. et v. cult.) Tancecium parasitiaim, Swartz, Fl. Ind. Occ. ii. p. 1053, t. 20. f. 2 ; DeCand. Prodr. ix. p. 245 ; Seem, in Bonpl. iv. p. 127; et in Hook. Journ. and Kew Misc. ix. p. 83. Crescentia, Browne, Jamaica, p. 266. n. 5. Geogr. Distr. Woods of Jamaica {Swartzl Purdiel Alexander^) ; Cuba {Ch. Wright, n. 434); near San Gabriel da Cachoeira, Rio Negro, Brazil {Spruce ! n. 2243). Cultivated at the Royal Botanic Gar- dens, Kew, 4. Tan^cium lilacinum; foiiis ovatis ovato-elKpticis vel lanceolato-oblongis utrinque glabris, panicuKs terminalibus miiltifioris, bracteis ovato-oblongis acuminatis, brac- teolis Mneari-lanceolatis striatis pedunculisque puberulis, corolla (lilacina, rosea vel violacea) glabra, tubo abbreviato, bacca ovali glabra (nigra), (v. sp. v. et sp. s.) Tancecium lilacinum. Seem. Bot. Herald, p. 182, Bonpl. iv. p. 128, et Hook. Journ. and Kew Misc. ix. p. 84. Schlegelia lilacina, Miq, in Bot. Zeit. ii. (1844) p. 788, Stirp. Surinam, p. 116 cum icon.; DeCand. Prodr. ix. p. 564. Schlegelia elongata, Miq. in Linnaea, xxii. p. 73? Stirpes Surinam, p. 128. Besleria ? violacea, Aubl. Guian. ii. p. 630, t. 254 ; DeCand. Prodr. vii. p. 539. Besleria ? carulea, Aubl. Guian. ii. p. 631 ; DeCand. Prodr. vii. p. 539, Citharexylon cineremn, {" jmrasiticum, scandens") Linn. Plant. Surinam, p. 11. n. 92 (non Syst., non Spec. Plant.) et in Herb. prop. Nomen vernaculum Guianense : " Emosse bereoy," teste Aublet. Geogr. Distr. Darien {Seemann !) ; Surinam {Hostmann ! n. 553 in Herb, Paris. ; Herb. Linn. prop. ; Focke et Kegel, teste Miquel) ; French Guiana (Aublet! in Mus. Brit.). Authors do not quite agree respecting the colour of the corolla. In the specimens I collected in Darien it was lilac, some might call it pink. Aublet had specimens with blue, and some with violet corollas, I regard these discrepancies as indicating so many slight varieties. Species exchtsce. Tanacium'i paniculatum, Sieh. = Arrabidcea paniculata, Seem. Tancecium pinnatum, 'W\\[d. = Kigelia p)innata, DeCand. Tancecium tripinna, Raeusch = CoZea tripinnata, Seem. DR. B. SEEMANN ON THE CRESCENTIACE^. 17 Tribus II. — Crescentie.!;. Seem, in Proc. Linn. Soc. ii. p. 269 (1853) ; Bot. Herald, p. 182 (1854^). — Calyx cleciduus, irregularis (spathaceus vel bipartitus). 4. Parmentiera, DeCand. Calyx deciduus, spathaceus, longitudinaliter fissus. Corolla subcampanulata, tubo lato brevi, fauce hiante, limbo distincto 5-lobo, lobis subasqualibus patentibus. Stamina 4, didynama, cum rudiraento quinti. Antherm 2-loculai-es, loculis divergentibus. Discus glandidosus ovarii basin cingens. Stylus elongatus ; stigma bilamellatum, lamellis integerrimis. Ovarium 1-loculare, multiovulatum. Frvctns carnosus, indehiscens, teres, tuberculatus vel leevis, spurie 2-4-locularis, epulposus. Semina plurima, parva, cordata vel subrotunda. Albumen nidlum. Arbores America iropicce, ramis aculeatis vel inermibus, foliis oppositis, simplicibus vel trifoliolatis, pedunculis unifloris subcongestis ex trunco aut basi ramorum et ramulorum ortis, coroUis albidis vel virescentibus, fructibus flavis. Parmentiera, DeCand. Revis. Bignon. p. 19; Prodr. ix. p. 244. Endl. Gen. Plant. Suppl. i. p. 1409, n. 4171. 1- Seem, in Hook. Journ. and Kew Misc. vi. p. 2/2. Crescentia spec. auct. DeCandolle (Prodi", ix. p. 244!) has enumerated only one species of Parmentiera (P. edulis, DeCand.) ; and I have added two more (P. cereifera, Seem, and P. acideata, Seem.). Of the latter two, only one (P. cereifera, Seem.) holds good, P. acideata. Seem, having proved identical with P. edulis ; so that the genus consists at present of two species, both very distinct from each other. P. edulis, DeCand. has thorns, occasionally simple leaves, and a tuberculate fruit; whilst P. cereifera. Seem., the famous Candle-tree of the Isthmus of Panama, is imarmed, has always compound leaves and a smooth fruit. As synonyms of P. edtdis, DeCand., must also be regarded Crescentia acideata, H. B. K., C. edulis, Desv., and C. musoicarpa, Zaldivar. 1. Parmentiera cereifera ; ramis inermibus, foliis omnibus trifoliolatis, foUolis ovato- ellipticis vel obovato-oblongis utrinque acuminatis serratis vel integerrimis, petiolo alato, corollse (albidte) lobis emarginatis, fructu tereti bisulcato Isevi glaberrimo. (v. sp. V. et cult.) Parmentiera cereifera, Seem. iii. p. 302; Bot. Herald, p. 182, t. 32; Hook. Journ. and Kew Misc. vi. p. 273. Nomen vernaculum Panamense : " Palo de velas" {i.e. Arbor cereifera), teste Seemann. Geogr. Distr. Central parts of the province of Panama, near the villages of Cruces, Gorgona, and San Juan {Seemannl). Formerly (in 1846) cultivated at Kew. Tills singular production was first described by me in Hooker's ' Journal of Botany and Kew Miscellany,' vol. iii. p. 302, where I said : " This tree is confined to the valley of the Chagres, Isthmus of Panama, where it forms entire forests. In entering them, a person might almost fancy himself transported into a chandler's shop. From all the stems and lower branches hang long cylindrical fruits of a yellow colour and a wax-like appearance, so much resembling a candle as to have given rise to the popular appellation of Palo de Velas, or Candle-tree. The fruit is generally from two to three, but not unfrequently four feet long, and an inch in diameter. The tree itself is about 24 feet high, Avith opposite, trifoliated leaves, and large white blossoms, which appear throughout the year, but in greatest abundance during the rainy season. The fruit serves for food to nume- VOL. XXIII. D 18 DR. B. SEEMANN ON THE CRESCENTIACEJE. rous herds of cattle. Bullocks especially, if fed with Guinea grass, Batatilla (Ipomcea brachypoda, Bth. = J. variabilis, Chois.), and the fruit of this tree, soon get fat. It is generally admitted, however, that their meat partakes in some degree of the peculiar apple-like smell of the fruit ; but that is by no means disagreeable, and easily got rid of if, for a few days previous to the killing of the animal, the food is changed. As the principal harvest is at a season when all herbaceous vegetation is dried up, the culti- vation of this tree in tropical countries is especially to be recommended : a few acres of it at each farm, would effectually prevent that want of fodder always severely felt after the periodical rains have ceased." 2. Pabmentieea edulis ; ramis aculeatis, aculeis sub foliis insertis, foliis aliis simpli- cibus ovato -oblongis vel cuneato-lanceolatis, aliis trifoliolatis, foliolis ovato-oblongis utrinque attenuatis integerrimis, petiolo superne anguste alato vel nudo, corolhie (virescentis) lobis undulato-crispis, fructu angulato tuberculato. (v. sp. s.) Parmeiiiiera edulis, De Cand. Rev. Bignon. p. 19; Prodr. ix. p. 244. Seem, in Hook. Journ. and Kew Misc. vi. p. 27.3. Parmentiera aculeata, Seem, in Bot. Herald, p. 183. Crescentia aculeata, Humb. Bonpl. et Kunth, Nov. Gen. Amei\ iii. p. 158; DeCand. Prodr. ix. p. 247- Crescentia edulis, Desv. Journ. Bot. iv. p. 112; DeCand. Prodr. ix. p. 246; M09. Fl. Mex. ined. fide DC. Crescentia muscBcarpa, Zaldivar, Fl. Mex. ined. nov. ; Heller, Reisen in Mexiko, p. 414. Quauxhichotl seu arbor Cucumeris moschati, Hernandez, Mex. p. 90. Nomina vernacula Mexicana : " Quauxhichotl," " Quaunuxilotl ," " Quaxilote " et " Cuajilote," teste DeCand. et Heller. Geogr. Distr. Mexico {Schiedel n. 1207, Coulter !). According to Hernandez, it is found in Yauhetepec ; according to Desvaux, in Guaxaca; according to Humboldt and Bonpland, in Campeche, at Gona- catepic ; and according to Coulter, at Zinapan. 5. Crescentia, Linn. Calyx deciduus, bilabiatus, lobis integerrimis. Corolla subcampanulata, tubo elongato, fauce magna ventricosa, limbo inoequaliter 5-fido vel crenato aut fimbriato-laciniato. Stamina 4, cum rudimento quinti. Antherm biloculares, loculis divergentibus. Discus glandulosus ovarii basin cingens. Stylus elongatus; stigma bilamellatum. Ovarium 1-loculare, multiovulatum. Fructus globosus, ovatus vel ellipsoideus, cortice lignoso, spurie 2-locularis, intus pulposus. Semina plurima. Albumen nullum. Embryo magnus. Arbores vel frutices arborescentes in America tropica indigense, in Asia et Africa cultte, foliis alternis solitariis vel fasciculatis, simplicibus vel trifoliolatis, pedunculis e trunco ramisve ortis, coroUis rubentibus, virescentibus, vel virescentibus purpureo et flavo variegatis, fructibus virescentibus vel flavescentibus. Crescentia, Linn. Gen. n. 772 ; Endl. Gen. Plant, n. 417 ; DeCand. Prodr. ix. p. 246 excl. sp. ; Seem, in Hook. Journ. and Kew Misc. vi. p. 273. Cujete, Plum. Gen. 1. 16. Crescentia Cujete, Linn., the species upon which the genus Crescentia was founded, is distinguished from its congeners by its fasciculate leaves, all of which are simple, and its fruit, the shell of which is so hard that it can be broken only by tlie application of an axe or some other sharp instrument. To this species C. cuneifoUa, Gardn. must be added as a synonym, as the latter is in no way distinct from the former, some of the specimens in DR. B. SEEMANN ON THE CRESCENTIACEyE. 19 LinriEeiis's own herbarium liaA'ing leaves tlie under side of which is slightly pubescent, as those of C. cuneifolia, Gardn. are ; and the difference in the fruit amounts to nothing : the spots are generally observable in young fruits, and disappear in the old ones. C. acuminata, H. B. K., which (misled by the term "fragile,'' applied to its fruit by DeCandoUe, a term not contained in the original description of H. B. K.) I suggested (Bot. Herald, p. 183), might belong to C. cuciirbitina, is, according to an authentic speci- men preserved in the Berlin Herbarium, also identical with C. Cujete, Linn. C. angiisti- folia of Willdenow's Herbarium is another synonym. The second Crescentia which I con- sider a good species is C. cuciirbitina, Linn., which Linnseus published in his ' Mantissa,' and which he probably never saw, as there is no specimen of it in his herbarium ; he described it, most likely, as he has done in several other instances, from Plumier's figures. This species is distinguished by its simple isolated leaves, and the shell of its fruit, which is so fragile that it may be crushed in the hand like an e^^. C. latifolia, Linn, has justly been regarded as a synonym of this species ; and to this I have added C. oljo- vata, Benth., C. letliifera, Tussac, C. toxicaria, Tussac, and C. paliistris, Forsyth Herb. The description of C. ovata of Burmann's ' Flora Indica,' p. 132, short as it is (" C. foliis ovatis integerrimis, apice acuminatis ; folia in hac specie perfecte ovata nee attenuata, ut in Crescentia Cujete, L."), agrees perfectly well with this, and no other species. The thu'd species of Crescentia holding good is C. macrophylla, Seem., allied to, but quite distinct from C. cucurhitina. It is cultivated in some gardens under the names of C. regia and Ferdinandtisa svperha ; and its geographical range seems to be limited to a portion of Mexico. The fourth species of Crescentia, the validity of which I am ready to acknowledge, is C. alata, H. B. K., the Tecomate of the Mexicans. That species is characterized by having a hard-shelled fruit and fasciculate leaves, the central leaf of its fascicles being trifoliolated. As a synonym of it I regard C. trijolia, M. Blanco, the description of the latter in the ' Flora de Filipinas ' agreeing word for word with G. alata, H. B. K., and its Mexican origin having been traced by Blanco, the author of that ' Flora,' who says : " Tal vez habran venido de America ; . . . . Uaman in Nueva Espaha ' Tecomate.' " These are the only four genuine species of Crescentia. C. aculeata, H. B. K. and C. eclulis, Moz. are merely synonyms of Farmentiera edulis, DeCand. ; C. edulis, Desv. the simple-leaved form of the same species. C. jasmiuoides, Lam. belongs to Gardenia clusicefolia, Jacq. ; and C. pinnata, Jacq. to Kigelia pinnata, DeCand. § I. Folia alterna, solitaria. Corolla virescens. 1. Cbescentia ctjcuebitina ; arborescens, foliis alternis solitariis omnibus simplicibus lanceolato-ovatis vel obovatis breviter acuminatis, fructibus globosis ovatis vel ellipticis, cortice fragili. (v. sp. v. et cult.) Crescentia cuciirbitina, Linn. Mant. p. 250; DeCand. Prodr. ix. p. 246 ; Seem, in Hook. Journ. and Kew Misc. vi. p. 274 ; Swartz, Obs. p. 234 ; Plum, et Burm. t. 109. fig. sup. ; Pluck. Aim. 1. 17 1. fig. 2. Crescentia latifolia, Lam. Diet. i. p. 558. Crescentia obovata, Benth. Bot. Sulph. p. 130, t. 46 ; Wlprs. Rep. vi. p. 517- C7-esceniia ovata, Burm. Fl. Ind. p. 132; DeCand. Prodr. ix. p. 247- Crescentia let/dfera, Tussac, Fl. des An til. iv. p. 50, 1. 17; Wlprs. Rep. vi. p. 517- Crescentia toxicaria, Tussac, Fl. des Antil. iv. t. 17- d2 20 DR. B. SEEMANN ON THE CRESCENTIACE^. Crescentia palustris, Forsyth Herb. Nomina vernacula : in Isthmo Panamensi, teste Seemann, " Calabaza de playa ;" in Venezuela, teste Birschel, " Coco de mono" dicitur. Geogr. Dish: Jamaica {Swartz \ in Herb. Willd. n. 11486, W. Wright \ Purdie\ Distinl) ; St. Vincent {Anderson !) ; Cuba {Ed. Otto !) ; St. Tiiomas {Morits ! C. Ehrenberg !) ; Chagres {Fendler, n. 120) ; San Juan {Moritz !) ; Topo, Venezuela {Birschel !) ; St. Domingo (teste Tussac) ; Pacific shores of the Isthmus of Panama {Seemann !) ; Island of Gorgona {Barclay !). Commonly growing close to the sea, with Hibiscus tiliaceus and Hippomane Mancinella, Linn. A shrub about 15 feet high. 2. Crescentia macrophylla ; arborea ; glaberrima, foliis alternis solitariis omnibus simplicibus obovato-lanceolatis breviter acuminatis, basin versus longe cuneatis, nervo (medio) utrinque acuto, petiolis basi valcle incrassatis, corolla (virescente) sub- campanulata, tubo elongato curvato ventricoso, limbo subregulari fimbriato-laciniato, lamellis stigmatis fijnbriatis, fructu elliptico utrinque rostrato cortiee liguoso duro. (v. sp. s. et V. cult.) Crescentia macrophylla, Seem, in Hook. Journ. and Kew Misc. vi. p. 274 ; Bot. Mag. t. 4822. Crescentia regia, Hort. Ferdinanda superha, Hort. Germ. Geogr. Distr. Teapa, Tabasco {Linden, n. 1608). Cultivated in Europe. The largest plants cultivated are about 12 feet high. Young branches and petioles pvirpUsh. Largest leaves 2-2^ feet long, and from 2-6 inches broad. Peduncles 1^ inch long. Calyx and corolla furnished with glandular dots. Style and stamens as long as the tube of the corolla. Ovary unilocular. Fruit not unlike that of. 3Iartynia proboscidect ; the specimens ripened at Kew 6-8 inches long, and in the thickest part about 1 inch in diameter. The limb of the corolla probably varies like that of C. cuciirbitina, which is sometimes distinctly five-lobed, sometimes crenato-dentate or fimbriated. The leaves resemble in shape those of C. Cujete, but their size is very much larger than that of any of its congeners. § II. Folia fasciculata. Corolla rubescens vel virescens purpurea ei flavo variegata. 3. Crescentia Cujete ; arborea, foliis fasciculatis ex axilla 5, omnibus simplicibus lanceolato-obovatis breviter acuminatis basin versus longe cuneatis, supra glabris nitidis, subtus puberulis glabrisve, fructibus plerumque globosis vel globoso-oblongis, cortiee bgnoso duro. (v. sp. v. et cult.) Crescentia Cujete, Linn. Spec. 872; DeCand. Prodr. ix. p. 246; Seem, in Hook. Journ. and Kew Misc. vi. p. 275; Swartz, Obs. p. 234; Lam. Diet. i. p. 557; Jacq. Amer. p. 175, t. Ill; Hook. Bot. Mag. t. 3430 ; Veil. Fl. Flum. vi. t. 103 ; Plum, et Burm. t. 109. fig. infer. ; Com. Hort. i. t. 71 ; Tussac, Fl. des Ant. ii. 80, t. 19. Crescentia cuneifolia, Gardn. in Hook. Journ. of Bot. ii. p. 422; DeCand. Prodr. ix. p. 246, Crescentia acuminata, H. B. K. Nov. Gen. Amer. iii. p. 157; DeCand. Prodr. ix. p. 246. Crescentia angustifolia, Willd. Herb. n. 11485. Nomina vernacula: in Cuba, teste E. Otto, "Guido;" in Colon. Brit. Indiaj Occident., "Calabash-tree;" in America Hispana, teste Humboldt, Bonpland et Seemann, "Tutumo," "Turtumo," "Palo de Turtuma," "Palo de Tutuma" vel " Calabazo" vocatur. DR. B. SEEMANN ON THE CRESCENTIACE^. 21 Geogr. Distr. Jamaica (P. Browne ! in Herb. Linn, propr., W. Wriyht !) ; Cuba {Humboldt et Borqjland \ Ed. Otfol); Santa Lucia (Herb. Mus. Brit. !) ; St. Thomas {C. Ehrenberg\) ; Guiana {Aublet\). Guatemala. {Friedrichst hall n. 1186) ; Surinam {Splitzgerber \) ; Brazil {Blanchetl) ; Isthmus of Panama {Seemann !) ; Cumana (Herb. Mus. Par. !) ; Puerto Caballo {Karsten !) ; Cartago de Q.wmAm [BoussingauU \) ; Peru (Dombetj I) ; Key West, Florida {Blodgett in Herb. Torrey !). Cul- tivated throughout the Tropics and in European gardens. A tree about 30 feet high, generally crowded with OrcliidecB, Ferns, Bromeliacece, and other epiphytical plants. Eruits generally attaining the size of 32-pounder cannon-balls. 4. Crbscentia alata ; arborea ; foliis fascicnlatis ex axilla 3, medio longe petiolato trifoliolato, lateralibus simplicibus minoribus similibus, petiolo foliorum foliolatorum late alato, fructibus globosis, cortice ligno duro. (v. sp. v. et cult.) Crescentia alata, R. B. K. Nov. Gen. Amer. iii. p. 158; DeCand. Prodr. ix. p. 247; Seem, in Hook. Journ. and Kew Misc. vi. p. 275 ; Seem. Parad. Vindobonens. ii. t. 87- Crescentia trifolia, Blanco, Fl. Filipin. p. 489 ; DeCand. Prodr. ix. p. 247- Nomina vernacula : in Mexico, teste Heller et Seemann, " Guantecomate" " Tecomate " vel " Quiro ;'" in ins. Philip., teste Blanco, " Hoja cruz " appellatur. Geogr. Distr. Common on the western coast of Mexico, from Acapulco to Mazatlan [Humboldt and Bonpland\ Gregg ! n. 944 ; Seemann !) ; Itztoluca [Ehrenberg !) ; Central America [Herm. Wendlandl). Cultivated in Veraguas (Seetnann\), Mariana Islands {Gaudichaud\ in Herb. Berol.), Manilla {Pen-ottet\ in Herb. Par.), and in European gardens, having been introduced by me in 1850 to Kew. The size of the fruit is that of a large orange. Species exclusce. C. aculeata, H. B. ¥.. = Parmentiera edulis, DeCand. C. edulis, Desv., Mo2,. = Parmentiera edulis, DeCand. C. jasminoides. Lam. = Gardenia clusiafolia, Jacq. C. pinnata, Jacq. — Kigelia pinnata, DeCand. 6. Kigelia, DeCand. Flos resupinatus. Calyx deciduus, tubuloso-campanulatus, bilabiatus, labiis irregulariter fise.s et hinc calycem spurie 5-fidum constituentibus. Corollte tubo brevi, fauce lata campanulata, limbo sub- lequahter 5-lobo, lobis ovatls acuminatis. Stamina 4, didynama, cum quinto sterili. Antherce biloculares, locuhs basi longe discretis. Discus glandulosus ovarii basin cingens. Stylus staminum fere longitudine ; stigma bilamellatum. Ovarium um\ocv\ave, multiovulatum, glabri.m, placentis parietalibus. Bacca elongato-ellipsoidea seu cylindracea, 2- 4-plove latitudine longior, cortice corticato, intus pulposa, spurie 2-locularis. Semina in pulpa nidulantia, subrotunda vel ovoidea, testa fuscescente. Albumen nullum. Cotyledoties {Me DC.) rotundatae, externe longitudinaliter plicatae segregatee. Arbor Africa tropicce, ram'is divergentibus, cortice albido, fohis oppositis imparipinnatis, foliolis (5-11) ellipticis vel ovato-eUipticis, non raro insequalibus, terminali obovato, integerrimis vel subrepando- dentatis, coriaceis, utrinque glaberrimis, paniculis longissime pedunculatis, pendentibus e trunco seu ramis vetustis ortis, corollis amplis atro-rubris extus pallidioribus, baccis (2 ped. et ultra long., 5-8 unc. lat.) pendulis, albidis. Kigelia, Revis. Bignon. p. 18 ; DeCand. Prodr. ix. p. 247 ; Endl. Gen. Plant. Suppl. i. p. 1409. n. 413/ ; Seem, in Hook. Journ. and Kew Misc. vi. p. 2/6. Bignonice, Tecomce, Crescentia-, Tancecii et Tripinnarice spec. auct. Sotor, Fenzl, Vortrag liber eine Crescent.- Gattung, p. 1. 22 DR. B. SEEMANN ON THE CRESCENTIACE^. DeCandolle (Prodr. ix. p. 247) enumerates only one species of Kigelia, viz. K. pinnata, DeCand. adding as synonyms Crescentia pinnata, Jacq., Tanmcium 2^innatum, Willd., and Tripinnaria Africana, Spreng., — an arrangement in whicli I fully concur. Decaisne has, since the publication of the ninth volume of the ' Prodromus,' described as a second species K. jEtliiopica (DeLessert, Icon. 39, t. 93 A et B), and Bentham (Hook. Niger Plora, p. 4.63) subsequently a third, K. Africana, uniting with the latter the old Bignonia Africana of Lamarck, which DeCandolle enumerates amongst his doubtful Bignonias. A comparison of the various descriptions and sjoecimens of these three supposed species has convinced me that all are one and the same species, the original K. pinnata, upon which the genus was founded, and which has for nearly a century been an inmate of our gardens. K. Abyssinica, A. Rich., the genus Sotor, established by Fenzl upon Kotschy's specimens (n. 403), and Tecoma Africana, Don, are also synonyms of Kigelia pinnata. 1. Kigelia pinnata. (v. sp. s. et v. cult.) Kigelia pinnata, DeCand. Prodr. ix. p. 247; Seem, in Hook. Journ. and Kew Misc. vi. p. 27/. Bignonia Africana, Lam. Diet. i. p. 424. Kigelia ^thiopica, Dene, in DeLessert, Nov. Selec. v. p. .39. A et B ; Wlprs. Rep. vi. p. 517. Kigelia Africana, Benth. in Hook. Niger Flora, p. 463 ; Wlprs. Ann. iii. p. 93. Kigelia Abyssinica, A. Rich. Tent. Fl. Abyss, ii. p. 60, t. 75 ; Wlprs. Ann. iii. p. 93. Sotor, Fenzl, Vortrag iiber eine Crescentiaceen-Gattung, p. 1. Crescentia pinnata, Jacq. Coll. iii. p. 203, t. 18. icon floris. Tancecium pinnatum, Willd. Spec. iii. p. 312. Tecoma Africana, G. Don, Gen. Syst. iv. p. 224. Nomina vernacula : in Abyssinia, teste Richard, " Meder Deur ;" in Nubia, teste Kotschy, "Sotor;" in Mozambico, teste DeCandolle, " Kigelikeia ■" et in Africa austral!, ad ripas fluminis Zongha, teste Oswald, " Maropotla" dicitur. Geogr. Distr. Nubia { Kotschy \ n. 403) ; White Nile {Sabatier \) ; Abyssinia {Quartin Dillonl Petit \ in Herb. Par.); Mozambique (teste DeCandolle); Bornu {Ed. Vogell); Port Natal {Garden \) ; Senegal {Don, Adansonl in Herb. Juss.) ; Cape Coast Castle {Th. Vogell); River Zongha, lat. 26° S. {Oswaldl) ; and at the Great Lake of South Africa (according to specimens at Kew). Cultivated in European gardens. [ 23 ] II. On the Distribution of the Tracheae in Insects. By John Lubbock, Esq., F.B.S., F.L.S., Sfc. ReadJanuary 1 9th, 1860. Dean SWIET, m the ' Tale of a Tub,' asserts that in the world generally, and in animals in particular, the outside is generally more beautiful than the inside. " Last week," he says, " I saw a woman flayed ; and you will hardly believe how much it altered her person for the worse." The Dean, however, does Nature an injustice. We constantly find the most lovely things where we should least expect to meet with them. Many insects are very beautiful externally ; but there is nothing, I think, more pleasing to the eye than some parts of their internal anatomy ; and no one, I am sure, can see the tracheae branching over the different organs, like trees of shining quicksilver, without being both astonished and gratified. Although Lyonet, Straus-Dlirckheim, and many other entomologists have described and figured the arrangement of the larger tracheal branches, no one has yet studied the distribution of the finer branchlets. This is, no doubt, mainly owing to the fact that, more or less quickly after death, fluid penetrates into them, and that they then become very difficult to distinguish from the surrounding tissues. It is indeed true that there are a few scattered observations on this subject ; but no one has yet compared together the finer tracheae attached to the different organs of various insects, or attempted to arrive at the laws regulating their distribution, — to determine, for instance, whether the type is the same throughout each insect, or different in the various organs, and to compare with one another insects of vai'ious families and orders. I ought perhaps to apologise for presenting to the Linnean Society a memoir so imper- fect ; but if the marvellous number of insects be remembered, and if it be borne in mind that each one contains at least from fifteen to twenty different organs to be examined separately, it will be evident that no one man could hope to exhaust the subject. Having, therefore, to deal with a subject so vast in itself, I have confined myself strictly to it. Without intending to express any opinion of my own as to their homologies and functions, I allude to the various organs by the names under which they generally pass, even when the generally received opinion seems to me doubtful or incorrect. I wish, however, to say a few words respecting the homologies of the intestine of Pentatoma. It seems to be pretty generally admitted that in this group of Hemiptera, the biliary vessels, either directly or through the "vesicule biliaire" of Leon Dufour, pour their contents into the rectum, below the ileum. Leon Dufour*, indeed, while * Mem. des Savants Etransers, vol. iv. p. 153. 24 MR. LUBBOCK ON THE DISTRIBUTION admitting that this is apparently the case, seems to think that it must be an illusion, and that there cannot really be any such communication. There is however, I believe, no other group of insects kno^vn in which the biliary vessels open elsewhere than at the anterior end of the ileum (excepting perhaps the Homoptera); and Ave find in certain Hemiptera that the posterior part of the stomach is much elon- gated, and that the Malpighian vessels open at the end of it. I am therefore inclined to believe that in Tentatoma, Scutellera, &e., the so-called " cordons vahoilaires " represent the posterior part of the stomach, and that if the chamber into which the Malpighian vessels fall is really the " colon," we may infer that the ileum is not developed. In examining any given organ or membrane, it is of course necessary to have before one a surface suificiently large to give a good idea of the type of distribution, in order to feel sure that the arrangement of the tracheae which is seen is really characteristic of the organ. I have therefore chosen the largest insects I could obtain, because in them we see the same type of distribution repeated over and over again in the field of view ; and I have also compared different specimens together. The mode of branching is in many respects comparable to that of trees. As we find, for instance, no two oaks exactly alike in their mode of branching, wliile yet the species possesses a well-marked type of its own, so in the tracheae, though no two branchlets divide in exactly the same manner, still they possess a well-marked character. And though in numerous insects many organs are alike in this respect, there are others in which any fragment of an organ could at once be recognized if it were large enough to show the mode of branching of the trachetB. In the different species of one genus we generally find the trachese very similar. In comparing together, however, insects belonging to different families of the same order, this is by no means the case. In the ovarian trachese, for instance, Ilnsca (PI. III. fig. 2) much resembles Bombtis; while Tipula (PL III. figs. 4 & 5) and Tenthredo are quite different from either, but resemble one another. A third genus of Hymenoptera, OjjJdon (PI. III. fig. 7), again, is dissimilar from either Bomhus or Tenthredo, and agrees very nearly with Aclieta (PI. III. fig. 12), which, for its part, differs entirely from LoGusta. It would seem, therefore, that the distribution of the finer trachese cannot have any bearing on the question of ordinal, or even of family affinities. In some cases trachese, which at first sight are very dissimilar, present in reality no difference. Thus, in PI. Ill, the figures 4 & 5 represent the trachese on the ovary of Tipula : in fig. 5, the trachese are represented as they appear when expanded by the e^g ; while on the parts of the tube which lie between the eggs, and are much narrower, they resemble fig. 4. It may probably be stated, as a general rule, that the waved course of many trachese is a provision to allow for the expansion and movements of the organ to which they are attached. There seems to be some special tendency in the ovarian trachese to arrange themselves in tufts, perhaps because in this way they are better able to adapt themselves to the alterations which occur in the diameter of the egg-tubes during the rapid growth and passage of the eggs. In Necrophori^s, Locusta, Chrysopa, Tipula, &c., the ovaries present more decided tufts of tubules than most of the other organs ; and in Musca and OF THE TRACHEA IN INSECTS. 25 Bombus, wliere tufts are present in most of the organs, the tubules are particularly numerous on the ovary. On the other hand, this rule is not without numerous excep- tions— as, for instance, Ophion (PL III. fig. 7), Acheta (PL III. fig. 12), A-phrophora (PL III. fig. 1) ; and it would perhaps not be safe to generalize on the subject. In every insect which I have examined, the ganglia were well supplied with tracheae, and the type was very uniform, showing, I think, that the minute structure of these organs is very similar, not only in different parts of the nervous column, but also in different insects. As regards the commissures, the case is quite the reverse : not only does the distribu- tion of the tracheae difii'er much in different groups, but in some, as, for instance, in the Neuroptera and Lepidoptera, they are, at least in the perfect insects, almost entirely absent. This difference probably arises from the difference in the respective properties and func- tions of the ganglia and commissures. It is an interesting fact, that while in butterflies and moths I have generally found the commissures free from tracheae, in the larvae they are richly supplied. Is this to be accounted for by supposing that the relative functions and structures of the different parts belonging to the nervous system are not so com- pletely differentiated in the larva as they afterwards become in the perfect insect ? The presence of tracheae on the commissures is not, however, always a sign of low development, since they are present in many Coleoptera. Both the larva and perfect insect of Acheta have tracheae on the thoracic, but hardly any on the abdominal commissures. The presence of a few tracheae cannot, however, be of any great functional importance, since in insects with double commissures I have more than once seen instances in which one only was so provided. Dr. Williams*, in a paper on the respiration of insects, enu.nciates very confidently the following proj)ositions, which he considers to be true without any exception, and which he has since reiteratedf : — - 1st. That the larger tracheae never anastomose ; that " in the sinral tracheae no plexiform union of the branches ever anywhere occurs." 2ndly. " That the ' spiralled ' or larger tracheae are mere conduits, like arteries or veins, and have nothing to do with, take no part in, the ultimate act of respiration." Srdly. " That the peripheric or extreme distribution of the tracheal system is conform- able in plan to that of a blood-vascular system ; that is, the capillary or membranous tracheae are always placed intermediately between larger trunks, the branches of which they serve to connect, — standing to the larger trunks in the same relation as the capil- laries of a blood-vascular system do to arteries and veins." 4thly. " That the tracheae can be discovered in no single instance to end in caecal terminations — always in mutual inosculations." I might have passed over the fu'st assertion as a mere momentary slip of the memory if it were not repeated more than once in 1854, and again reaffirmed in 1856. Several of the figures given by Straus-Dtirckheim, Leon Dufour, and other writers, show inos- culations of the tracheae ; and there is hardly a single insect in which they do not * Annals and Mag. of Nat. Hist. 1854, vol. xiii. p. 194. f lb. 1856, vol. svii. p. 347. VOL. XXIII. E 26 MR. LUBBOCK ON THE DISTRIBUTION occur in certain organs. It will be sufficient here to mention the dorsal tracheae in certain Coleopterous larvae ; the ganglia, anterior part of the stomach, oviduct, and heart of Bomhus ; the crop and ganglia of Carabtis ; the ca^ca and stomach of Acheta ; the stomach of Tentliredo, Athalia, and lAbellula. These instances, and many others which might be given, show that inosculations of the larger tracheae, so far from never occurring, are, on the contrary, by no means uncommon. The second principle enunciated by Dr. Williams, viz. that the spiral tracheae are mere conduit pipes and have nothing to do with the act of respiration, is less easily disproved, though I believe quite as erroneous as the preceding. Of course I do not mean to deny that the absorption of air is probably more rapid at the thin ends of the tracliea3, where the spiral is absent or rudimentary and the tracheal walls thinner. Still upon this Dr. Williams probably relies, since he scarcely gives any reason for the statement. On the ovary of Aphrophora a few large tracheae are given to the posterior end, and there divide into about 200 branches. These branches (PL III. fig. 1) run the whole length of the egg-tubes, a distance of xoth of an inch, without giving off a single branchlet. They are o-o-outh of an inch in diameter at the base, and gradually taper until they become excessively fine. At the base the spiral filament is very distinct ; gradually it becomes less and less so, but it does not cease suddenly, and it can be traced for a considerable distance up the branch. If, therefore, Dr. Williams's view were correct, no respiration could take place in the ovary, except at its anterior end. Moreover, while I readily admit that in all probability very little respiration can take place through the spiral filament itself, still it must be remembered that the coils of the spiral leave between them a considerable space, which is occupied by a dehcate membrane. Why Dr. Williams should suppose that no interchange of gases can take place through this membrane, I know not ; and until some reason is given, I shall think myself justified in concluding that the spiralled trachea3 are not mere conduit pipes, and that, although respu'ation may proceed more rapidly at the fine extremities of the branchlets, still it is not confined to these parts, but is carried on also in the larger vessels. It is true (as Dr. Williams was, I believe, the first to observe) that in some cases the spiral rib ends suddenly; this is generally the case where a trachea gives off at one point several small tubules, as for instance in PI. I. fig. 8 ; but the sudden change does not occur in all insects, which led me for some time to think that there must be some error in Dr. Williams's observations. We now come to Dr. Williams's third proposition, that the capillary or membranous tracheae are always placed intermediately between larger trunks, standing to them in the same relation as the capillaries of a blood-vascular system do to arteries and veins. I must confess that I am unable in any way to confirm this statement. I have never seen the minute tubules gradually collect into efferent branches, nor have I been able to find any evidence that there are two sorts of large branches — one afferent, and the other efferent — as there ought to be according to Dr. Williams's hypothesis. In the blood- vascular system we have arteries, capillaries, and veins ; in the aeriferous system of insects we have only air-tubes, becoming, as a general rule, gradually smaller and smaller as they proceed further from their origin. In most cases also, each trachea supplies a definite area, which receives no branchlets OF THE TRACHEA IN INSECTS. ' 27 from any other source; and it seems evident tlierefore that in these cases the air must pass to and fro in the same channel. Dr. Williams's fourth assertion seems to me as little tenable as any of the three pre- ceding. I do not, of course, deny that in some organs the tubules may inosculate, or, in other words, that the trachese may end in loops ; but certainly this does not seem to me to be universally or even generally the case. E-eferring again to the trachete belonging to the ovary of Aphrophora, we see that each branch tapers from -g-^^oth do-\vn to 2000 o^ of ^'^ inch without giving off a single branchlet. Does, then, the tine end run into the similar termination of some similar branch, and thus form an elongated ellipse ? if so, we should have expected to find the calibre nearly equal throughout; but the ends were so ex- cessively delicate that any such inosculation could be of Httle importance. A glance at the trachese of the Malpighian vessels in JEschna seems to me to be almost equally conclusive, as the tracheoe there run nearly half an inch, becoming very fine, and yet sometimes only give out a single branchlet. I readily admit that these long blind tixbes seem at iirst sight but little fitted to ensure a constant supply of fresh air to the organs along their whole course : on the other hand, it must be remembered that the interchange of gases is very rapid ; and as we do not find among the larger trachese any separate afferent and efferent branches, we ought not hastily to conclude that they must certainly exist among the branchlets. It occurred to me, in reflecting upon this siibject, that the removal of the carbonic acid from the trachese, and the continual supply of oxygen, were probably effected in con- sequence of the diffusion of gases. Prof. Huxley also suggested to me that in fact the same is the case in man and the higher animals generally. Upon mentioning this to Prof. Graham, he referred me to his paper in the ' Philosoplrical Magazine ' for 1833, where he has expressed the very same opinion. As his suggestions appear to have been almost, if not altogether, overlooked by physio- logists, I may perhaps be permitted to quote a few of his remarks. He says, " I may be allowed to mention an application of the law of diffusion in explanation of the mechanism of respiration. The cavity into which air enters during respii'ation consists, first, of a large tube, the windpipe ; secondly, of smaller tubes, into which the windpipe diverges ; and thirdly, of a series of still smaller tubes, diverging from the last, themselves rami- fying to an indeterminate extent, till at last the tubes cease to be of sensible magnitude, but are believed to terminate in shut sacs. The capacity of the whole cavity cannot easily be determined ; but we may estimate it at 800 cubic inches. In a natural expiration, about 20 cubic inches or ^jth of the contents are thrown out, from the application of a general pressure to the whole ; but it is evident that these 20 cubic inches will be the 20 cubic inches nearest the outlet, or the contents of the larger tubes. The contents of the second-sized tubes will advance at the same time into the largest tubes, but no further, and will recede again into their original depositories on the next inspiration, which Avill fill the larger tubes with fresh air, wliich identical quantity will again be expelled in tlie next expiration. " This illustration is perhaps too strongly stated ; but it is evident, that, in ordinary respiration, the slight mechanical compression will have little or no effect in emptying E 2 28 MR. LUBBOCK ON THE DISTRIBUTION the most distant tubes, or the ultimate air-cells, of their contents. The bullc of the air, also, is not altered during respiration, although, for a quantity of oxygen, carbonic acid gas is substituted. This substitution, which is the great end of respu-ation, undoubtedly takes place most abundantly in the minute and distant air-cells, which present the largest surface to the blood; and the carbonic acid there produced must be moved along the snialler tubes by the diffusion-process (which we know to be extremely energetic and also inevitable), till it is thrown into the larger tubes, from which it can be expelled by the ordinary action of respiration. But the action of diffusion is always twofold : at the same time that carbonic acid is being carried outward from the air-cells, oxygen is carried inward in exchange ; and thus the necessary circulation is kept up throughout the whole lungs. " There can be no doubt that much of this quantity occupies constantly and perma- nently the most minute tubes and air-cells ; for it can scarcely be withdrawn by means of the air pump. Now the question has arisen, how these ultimate tubes and air-cells are so powerfully inflated ; for they are not distended by the action of muscular fibre, of which they are known to be destitute. This state of distension must be highly useful by exposing surface ; and the law of diffusion enables us to account for it. The heavy car- bonic acid which these minute cells may contain, is not merely exchanged for oxygen, but for a larger volume of oxygen, in the proportion of the diffusion-volumes of carbonic acid and oxygen ; namely, 81 carbonic acid are replaced by 96 oxygen. The resistance to passage through the most minute tubes is overcome by the diffusion-action, as in the case of the pores of the stucco plug ; and there follows a tendency to accumulation on the side originally occupied by the carbonic acid. This accumulation is limited by the in- creased facility with which the air-vessels can empty themselves mechanically of a portion of their contents, from their distended state. " In the law of diflPasion of gases, we have, therefore, a singular provision for the full and permanent inflation of the ultimate air-cells of the lungs. But it is in the respiration of insects that the operation of this law will be most distinctly perceived. The minute air-tubes accompanying the blood-vessels to every organ, and like them ramifying till they cease to be visible under the most powerful microscope, are kept distended during the most lively movements of the little animals, and the necessary gaseous circulation maintained wliolly, we may presume, by the agency of diffusion." Certainly, however, in many insects the respiratory movements are as well-marked as in any of the higher animals ; and even in Caterpillars and other insects* where they are absent, still the interchange of gases must be assisted by the ordinary movements of the body. Yet Prof. Graham is, no dotibt, quite correct in denying the existence of any actual current of air in the smaller tracheae; and that under these circumstances the supply of oxygen should be sufficient shows well how rapidly and forcibly the diffusion of gases takes place f ; but I am still doubtful whether the superior quantity of oxygen * Ranatra has no respiratory movements. In this genus, according to Newport, the whole respiration is effected through the caudal tuhe. It has, however, three pairs of spiracles on the under side of the abdomen ; and each spi- racle receives a moderate-sized trachea. t The membrane of the spiracle well represents the plaster plug used by Prof. Graham in his experiments. OF THE TRACHE.E IN INSECTS. 29 which enters, over the carbonic acid which escapes, is in all cases so necessary to expand the tracheae as Dr. Graham suggests, because I have found the small branchlets in a larva of Iftisca still full of air after it had been drowned by immersion for some houi's in water. On putting some larvae of Melolontha into water, I was surprised to see a considerable formation of bubbles on the skin, especially at a point below each spiracle, while no bubble ever appeared from the sjiiracles themselves. If, however, the water is first boiled, then no bubbles are produced, and the larvae very soon, say in about a quarter of an hour, become motionless, though, if pricked, they still contract a little. Their flesh is then quite soft and flabby, while generally it is tolerably firm to the touch. I expected to have found the tracheae free from air, or nearly so ; but this was not the case. Like other insects, these larvae readily recover from their suffocation when they are taken out of the water. The larvae of flies are also naked fleshy grubs ; and I expected them to behave in a similar manner; this, however, is by no means the case. They live much longer in water. When they are placed in it, no bubbles form on their skin ; nor does it seem to make any difference to them whether the water is boiled or not : I put four into some boiled water, and the same number into water' which had not been boiled ; and at the end of forty-four hours they still moved a little, gently turning their heads from one side to the other. These facts seem to me to prove that the larvae of Melolontha breathe partly by means of their skin, and that those of Musca do not. Yet these opposite states of the skin may be necessary for these larvae, living as they do under such different circumstances *. This result was quite unexpected by me ; yet it throws much light on the intermediate stages which, upon the principle of natural selection, must have existed between ordinary larvae, respiring principally through spiracles, and those which, like the larvae of Botys, of Dragonflies, Ephemerae, &c., breathe by means of foliaceous expansions of the skin. In most insects the air will be found, after death, filling the fine ends of the tracheae. In some cases, however, as in many parts of Carabus, Melolontha, Acheta, Hiiyparchia, &c., the smaller branchlets are generally, even very soon after death, filled with fluid, and can therefore scarcely be distinguished, or they even become quite invisible. This happens very frequently in Lepidoptera, Coleoptera, and Orthoptera ; but I have not noticed it so often in Hymenoptera, Diptera, or Neuroptera. It is not, however, constant in the first three orders ; and in Neerophorus, for example, the very fine tubxiles may be beautifully seen. The larvae, at least of Coleoptera and Lepidoptera, do not in tliis respect resemble the per- fect insect ; but in Acheta all the three forms are alike. * M. Lyonet, writing before the observations of Dr. Graham had thrown so much light upon the subject, and mis- led principally by the absence of special respiratory movements in many insects, doubted whether the tracheae were organs of respiration, and suggested that one at least of their uses might be " de coucourir avee les nerfs, a la contrac- tion des muscles, pour operer les mouvements," though I confess that I do not quite understand in what way the tracheae were to cooperate with the nerves. In support of this view, he mentions that if the spiracles are closed by oil, or if the insect is kept under water, after a while it loses all power of movement. This experiment, however, hardly justifies the conclusion which M. Lyonet deduces from it ; and it seems to me that the larvae lose their power of movement in the same way as M. Lyonet himself would, had he been treated in a similar manner. 30 MR. LUBBOCK ON THE DISTRIBUTION It might at first be supposed tliat the points at which the air ceases are really the ends of the trachese, and that where they can be traced further they are merely solid threads. In some cases, however, the air disappears slowly from the fine tubules ; and a case of this is represented in PI. III. fig. 8. Fig. 8 represents a smaU tracheal branchlet ramifying over two egg-tubes of MpparcUa Jctnira, examined as soon as possible (perhaps a quarter of an hour) after the death of the insect. Fifteen minutes later, the air had disappeared from the fine ends, and no more could be seen than is represented in figs. 8' & 8". In an- other quarter of an hom% still more of the tracheae had disappeared, and only the stump as it were, F, remained. In this instance all the other tracheae on the organ ended in the same abrupt manner ; but it is evident, from the preceding and many other similar obser- vations, that the apparent ends are by no means the real terminations of the tracheae. This absorption or repulsion of the air from the finer tubules depends evidently, in a great measure, on the nature of the wall of the tracheae and of the surrounding tissue, because it proceeds with different rapidity in different parts of the same insect, so that while, for instance, in Carabus I never found air in the finer tubules on the Malpighian vessels, on the ovaries the air may be seen for a short time after death ; and in the gan- glion, even after some hours, all the fine tubules still contain air. I am unable to offer any decided opinion whether this disappearance of the air from the finer tubules is owing to its absorption by the tissues, or whether the surrounding fluid forces its way through the delicate membrane of the tracheae. Probably, however, the latter is the case ; because, if the air were simply absorbed by the surrounding tissue, a fresh quantity would, I suppose, continually be supplied by the larger tracheae. Moreover I never found that the manner of death made much difference in the condi- tion of the trachese, whereas, if the disappearance of the air were caused by vital action of the surrounding tissue, this would hardly have been the case. It is difiicult, however, to understand why the small tracheae fill themselves with fluid so much qu.icker in some insects, and in some organs, than in others. Probably, however, this is o\\ing to some differences in the tracheal wall, which, though always permitting the absorption of air by the tissues, may be more easily permeable by fluid in some parts than in others. The surrounding fluid must probably have the same tendency to expel the air from the finer branchlets during life as after death ; so that, whilst the insect remains alive, some counteracting agency must be at work. In those insects which show no respiratory move- ments, Prof. Graham's suggestion, above alluded to, seems to offer the best explanation. I have not yet paid much attention to the metamorphoses of the tracheae during the change from the larva into the imago. The process seems, however, to be very curious ; and I hope ere long to be able to devote some time to the investigation of it. In a Lepidopterous pupa, I found, in various parts of the body, knotted spiral tracheae invested by a common membrane, as figured by Semper*. How these knots are formed I know not. The structure of the tracheas has generally been described as consisting of an external * Zeitschrift fur Wiss. Zool. pi. xv. fig. 10. OF THE TRACHE.E IN INSECTS. 31 and an internal membrane, enclosing between tbem a spiral filament. Marcel de Serres and Straus-Diirckheim deny the existence of the spiral thread in the air-vesicles, which, however, is afiirmed by Suckow and Sprengel. The truth seems to be that it is certainly present in some cases, while in others it is either imperfectly developed or altogether absent. In some instances the chitinous thickening has a zigzag appearance, which Leydig also has observed. Ley dig* also appears to me to have given the best description of the structure of the tracheae. He considers that the outer " peritoneal tunic " is a " connective-tissue, trans- parent, and generally colomdess membrane, formed by the union of the same cells which form the fatty tissue, and with which also they remain in intimate connexion." This description seems to me quite correct ; and any one may test its accuracy by examining the fatty tissue of a caterpillar, where he will see that the external membrane here and there leaves the trachea and encloses around it a considerable space, which is occupied by the fat globiiles. In other words, the external membrane of the trachea is continuous with that of the fatty tissue ; the trachea possesses therefore an external membrane only so long as its course is between the masses of fatty tissue, and loses it as soon as it enters one of the masses. Burmeisterf describes this membrane as structureless; but it undoutedly consists of a union of cells, whose walls can generally be perceived without much difficu.lty. The nuclei also are generally visible. Among the older writers, Sprengel appears to have had the most correct idea of the so-called spiral filament. He considers it merely as a local spiral thickening of the inner membrane, — in which view he has been followed by Burmeister and Leydig. Bur- meister, however, like DeGeer, Lyonet, Bonnet, Straus-Diirckheim, Newport, and others, describes a third, mner membrane, which, like Sprengel and Leydig, I have been unable to see. PI. II. fig. 13 represents a fragment, which shows clearly the membrane uniting the spkes of the " thread." It is, however, possible that the spirally thickened portion of the inner membrane may sometimes detach itself more or less completely from the membrane by which it is formed. Some of those naturalists who have convinced themselves of the presence of an inner membrane lining the spiral thread, appear to have arrived at this conclusion mainly on account of the inner membrane of the tracheae being shed at each change of skin;. the spiral structiu*e, however, is shed at the same time. I have not yet been aljle to investigate the mode in which tliis exuviation takes place so carefully as I could wish ; but the fact does not justify the conclusion which has been deduced from it. In the cast skin of a larva oi Dasychira pudibimda, and in a pupa-case of Deilepliila galii, I convinced myself that the spiral filament is cast at the time of moulting. Nervous System. — Ganglia. — (PI. I. figs. 1, 3, 4, 5 & 6.) Throughout the Insecta the tracheae of the nervous system and of the muscles vary less than those of almost any other part of the body, from which it may probably be * Lehb. d. Histologie, p. 386. f Handbook of Entomology, Shuckard's Transl. p. 171. 32 MR. LUBBOCK ON THE DISTRIBUTION inferred ttiat tlieir ultimate structure also varies less than that of other parts. The ganglia are always well and often very richly supplied with tracheae, differing in this respect from the commissures and nerves, which in many insects (as mentioned helow) are entirely free from them. The tuhules, excepting sometimes those which are distributed to the surface, are straight or gently curved, but not waved. They generally rise singly from the branch] ets, as in PL I. fig. 1 ; but sometimes, even in the same insect, we find tufts, as in PI. I. fig. 3. The abdominal ganglia in Bombus terrestris and JB. lapidarius, are covered over with a network of tracheae. The meshes vary much in size, from -g-^ inch in length and 7^ inch in breadth to less than -^oih. of that size. The tracheae forming the network are of considerable breadth, averaging perhaps -awo i^ich. The short commissure between the posterior and penultimate ganglia has as rich a network of air-vessels as the ganglia themselves. These latter are penetrated by niimerous branches which end in tufts (PI. I. fig. 3) of from five, or even sis, down to two or three. The tubules branch two or three times, and resemble a fishing-rod in their general proportions and curvature, running in a gentle sweep often for j^o inch, with a thickness at the base of only tto 00 inch. The ganglionic tracheae of Fanorpa, lAbellula, Carabus, and AphrojjJwra are like those oi Bombus; and so are the tracheal tubules of Vespa, Tenthredo, Athalia spincmmi. Ichneumon (probably) extensor'ms, Ophion, Clirysopa, Tipula, Campcea margaritaria, Callimorpha Jacobs (B, Pterophorus, and the larva of. Euplexia lucipara; but I am not sure whether the branches end in tufts. In none of them is there a network enclosing the ganglia, as in Bombus. In the larva of Lampyris, AmpjMmalla, Aclieta (both larva and imago), Lociista, and Gryllus, the tubules are larger and longer, but otherwise very similar. The branches, however, do not end in tufts. In Forficula the tracheae resemble those of Acheta, but are smaller. In NecropJiorus Jmmator, N. vespillo, Scarabceus, and Musca, in Sip)parcliia Janira, Pieris ncqn, Noctua gamma, the larva of Basiocampa i^iibi, of Mamestra brassiccB, of Oonepiteryx rJiamni ; the tracheae sometimes end like the preceding, and sometimes in tufts, a large branchlet ceasing suddenly and giving off from four to seven or more tubules. These latter divide and re-divide again, sometimes breaking into little secondary tufts. In the larva of Musca the mode of branching is quite unlike that of the imago, being simple and not in tufts. In Sydrous pncetis the tracheae are in tafts ; but the tubules rising from them are nearly of uniform size, so that there are no secondary tufts. In Eristalis the tracheae of the ganglia break into tufts like those prevalent throughout the body. Upon the surface they are waved and resemble those of the commissures ; but in the interior of the ganglion they seem to be straighter. In Pentatoma large wide tracheae go to the ganglia, and give off short thick branches. These again give off still shorter branchlets, which end in tufts consisting each of a few straight tubules (PL I. fig. 5). OF THE TRACHEA IN INSECTS. 33 Interganglionic Commissures. (PI. I. figs. 2 & 7.) It might have been expected that the tracheae on these organs would have resembled those on the ganglia, and at any rate that tracheae would either have been present or absent in all insects. Both suppositions, however, would have been wrong. The tubules in the ganglia are usually straight, while those on the commissures are generally waved, besides difi'ering in other respects. Again, in Fentatoma and all the Lepidopterous larvae which I have examined, the commissures are very richly supplied ; in Bombus, the Diptera, and some Coleoptera, they are less numerous ; while in some Coleoptera, the Orthoptera, Lepidoptera, and Neuroptera, the tracheae are either rare or altogether absent. This at least holds good as far as my observations go ; but the statement wiU no doubt be modified by further investigations. Between two of the thoracic ganglia in the larva of Acheta I found one of the com- missures pretty well supplied "svith trachea?, while the other was quite free from them. Indeed generally throughout insects the tracheae on the tAvo nervous columns are not arranged in exactly the same manner. In Scarabceus, Necrophorus vespillo, and N. humator, the tracheae on the commissures and on many nerves were in tufts resembling those of the ganglion, except that they are altogether smaller, that the tubules branch oftener, and that they are more waved. In Cerambyx moschcdus also the commissures were well supplied with tracheae. In Sy- drophilus piceus the tracheae were like those of Necrophorus, but in some cases I observed branchlets which did not form tufts. The tubules were numerous and very delicate indeed. In Musca and EristaUs the tracheae were in tufts, very much like those on the gangKa, but with straighter tubules. The nerves, however, were almost entirely free from tracheae. In Tipula the tracheae were numerous. In the larva of Lucanns the commissures were well supplied, but in Carabus, the larva of Lampjyris, and in Forjicula, both they and the nerves were almost entirely free from tracheae. Tn Acheta, Locitsta, and Gryllus also, the abdominal commissures had only here and there a single trachea, which seemed as it Avere to have strayed out of a neighbouring ganglion. In these cases I have not examined a suflRcient number of specimens to prove that these single tracheae are inconstant, but I have little doubt that this is the case. The thoracic commissures of Acheta, and also of its larva, were provided with decidedly more numerous tracheae, which are a little waved and give out single branchlets which again divide once or tAvice. In Noctua the tracheae are like those of Acheta, but rather more numerous, tAVO or three generally running nearly to the middle of the commissure. They are seldom, if ever, branched, and are almost straight. On those of CampcBa margaritaria, jPieris napl, Sipparchia janira, Callimorpha jacobmoi, and Tterophorus, I found no tracheae. On the contrary, in the larvae of Euplexia lucipara and Oonepteryx rhamni they Avere Avell sup- plied ; and in the larvse of Lasiocampa rubi and Mamestra they were very numerous both on the commissures and nerves. On the commissures of Bombus nmscorum they are nu- merous, waved, and generally in tufts consisting of a feAv branched tubules, sometimes even of only two. On those of Athalia spinartim, Tenthredo, Ichneumon (probably) ex- tensorius, and another small black species, I found none. On those of Ophion there VOL. XXIII. F 34 MR. LUBBOCK ON THE DISTRIBUTION were hardly any. The few that I coukl see were almost straight and branched at in- tervals. In Vespa there was only here and there a little tuft. On the commissures of Clirysopa, Panorpa, and lAbelUda, I found no tracheae, and on those of Aphrophora only one here and there. In Tentatoma jtmiperina and P. bac- carimi, on the contrary, both the commissures and the nerves were well supplied with waved tracheae, Avhose mode of branching needs no special remark. On the nerves of Insects generally I found fewer trachese than on the commissures, and in most species many even of the larger nerves were without any. Whether this difference in different nerves be constant in each species, I am unable to say. (Esophagus. (PI. I. figs. 8 & 14.) The oesophagus is generally very poorly, if at all, supplied with trachese ; indeed even when they are present they are attached to the muscular bands with which the organ is provided, rather than to the organ itself. In Tentatoma, JPanorpa, the laiwa of Lampyris, Noctua gamma, Limnepldlus vitratus, Aphrophora spumaria, Ichneumon, Opliion luteum, Vespa communis, Tenthredo viridis, Gryllus niridissimus, Tipida, and the larva of Mamestra, it appeared to have no trachese. In Ilusca there is a ring round the pharynx very richly supplied with trachese. On the upper part of the oesophagus the trachese are in tufts, with waved or straight tubules ; on the posterior half there are a few longitudinal trachese ; starting from the front end and running backwards, they give off waved, transverse, branchlets at intervals, and belong to the oesophageal muscles rather than to the organ itself. In Libellula the oesophagus is narrow in front and swollen behind. The ajiterior narrow part has no trachese ; the posterior portion is supplied as in Musca, only that the branches rise behind instead of in front. Iia Necrophorus vespillo the type is almost the same. In Cerambyx moschatus the trachese had more numerous transverse branchlets, but only their bases were visible. In Luca.mis cervus ajid Amphimalla solsti- tialis they were similar, bu.t with fewer branchlets ; the ends were invisible. In Carabus the trachese are as in Forficula and Ilusca. In the larva of Lucanns the trachese re- sembled those of the rest of the intestinal canal, while the larva of Lampyris seemed to have none. In Bombus they are few, waved, and simply branched. In the Orthoptera generally, the oesophagus is much swollen behind. In Forficula the type is the same as in Musca, but the lateral In'anchlets are larger and more branched. In Acheta domestica it is accompanied by two enormous trache88, which are vi'ide at each end and narrow in the middle. They give off large transverse branches, v.^hich are irregularly branched, somewhat like a system of rivers ; and the ends are finely and beautifully waved. Locusta is also well provided mth trachese, but the ends had become invisible. In the larva of Acheta the mode of termination of the tracheae, so far as it was seen, resembled that of the imago. In Eristalis the trachese are unlike those of Ilusca ; most of them are in tufts (PI. I. fig. 8), but are straighter and more delicate than those on the stomach. Here and there the character of tufts is almost lost, and they b i'anch simply. OF THE TRACHEA IN INSECTS. 35 Salloarii Glands. In the larva of Huplexia liicipara and of one of the Ilamestrce the sali^'^ary glands have few branchlets, vs^hich are simple and without numerous tiibules. The duct was gene- rally free from tracheae. In Clirysopa the mode of branching is similar, but the tubules are not quite so long. In Bombus liicorum the tracheae on the thoracic glands were in tufts. In Fentatoma the tracheae are very nearly like those of the stomach, but with perhaps rather fewer tubules. Eristalis has four salivary glands ; two are cylindrical with circular bosses, the other two resemble in shape the ordinary long salivary glands of caterpillars. On all four the tracheae are in tufts as on the oesophagus, and the tubules are greatly curA^ed, but in each tuft there are very few, and even sometimes the branching scarcely seems to be in tufts at all. In the larva of lliisca the branching is simple, and ]iot in tufts. It is much like the figure in PL IV. fig. 8. Cceca. (PI. I. figs. 10 & 11.) The cseca of Acheta are two thick pulpy organs, one on each side of the stomach. They have some large vascular tracheas, which divide several times and then run into the corresponding branches on the other side of the organ, as in PL I. fig. 10. The branchlets have a very rigid appearance, as they are broad at their origin, and taper very quickly. In the larva the tracheae are nearly similar, but more elongated. The character of the tracheae is nearly the same in Gryllus as in Acheta, but the in- osculations are less conspicuous, and the branchlets are rather more elongated. In Chrysopa the caecum is cylindrical, but tapering to the free end. It is divided by twelve constrictions into well-marked bosses. Large tracheae run up the organ from its base and give off transverse branches, one of which is represented in PL I. fig. 12. Each boss has also a branch or naore, which give off from their outer side long straight tubules (PL I. fig. 11). I counted 328 of these long tubules in one half of one boss, so that the total number must be very large ; and as each one is about ro^^li of an inch in length, the total length of the whole must be considerable. In the long tubes attached to the front end of the stomach in the larva of Ilusca, the mode of brandling is simple, and offers no great peculiarity. Sucking Stomach. (PL I. fig. 9.) In Bombus terrestris, Vespa vulgaris, Ophion, Tenthredo luteus. Ichneumon, Campcea margaritaria, Pleris brassicce, Tipiila, and Noctua gamma, this organ has no tracheae. In Musca, on the contrary, it is well supplied with tracheae in pretty little systems. The tracheae branch simply, and the tubules have a tendency to run a little Avay along the parent stem, and then curve out from it in a beautiful manner. In the larva of Musca it appeared to be free from tracheae. In Eristalis it is poorly supplied, but the tracheae are of the same type as those on the stomach. The duct also is supplied in a similar manner. F 2 36 MR. LUBBOCK ON THE DISTRIBUTION Crop. (PL I. fig. 13.) The walls of the crop are generally strengthened hy a thick chitinous layer which pre- vents the fine ends of the tracheae from being well seen. It is also generally surrounded by muscular tissue. In AcJieta domestica it is provided with twelve longitudinal tracheas in six pairs which are joined at the base, and send off transverse branchlets at regular intervals. In Carahus there are about ten longitudinal tracheae which send out some large branches at acute angles, and many cross-branchlets (PL I. fig. 13). At the posterior end the great longitudinal tracheae are connected by a transverse vessel. In the larva of Lampyris the tracheae are as on the stomach. In Fanorpa they form a network, as in Carabus. Stomach. (PL II. figs. 1 to 10.) In Bomhiis terrestris, B. lapidar'ms, B. muscorum, and B. hortorum, the stomach is divided into two parts : along the anterior run several large longitudinal tracheae which anastomose freqxiently, while on the posterior portion the tracheae are mostly transverse. In B. pratorum this anterior part is shorter than in B. terrestris. The tracheae give off branchlets which generally end in tufts (PL II. fig. 5), but the tubules themselves branch a good deal ; so that often the character is almost lost, as in the right-hand tuft in the figure. PL II. fig. 11 represents some of the tracheae of the larva of B. muscorimi. The branchlets are longer than is the case in the perfect insect, and the tubules are longer, straighter, and fewer, — generally, indeed, only three in a tuft. In Vespa the tracheae resembled those of Bomb us. Often, however, instead of ending in tufts, the branchlets terminated as in PL II. fig. 10, the long tubules running up a muscle. Op)Jiion (PL II. fig. 3) is characterized by very long branchlets (one-fortieth of an inch in length, without a division), which at last break up more or less dichotomously into numerous tubules. In Tentliredo, a genus belonging to the Ichneumonidae, the type was much like that of Ophioii, but altogether much smaller. The same is the case with Cynips lignicola. In Athalia spinarwn and Tenthredo, as in Borubus and Vespa, the tracheae anastomose frequently on the front part of the organ. In Ichneumon the tubules are straight, or liut little waved, long, few, and more or less at right angles with the branches from which they spring. In the Cricket the stomach is divided into two distinct parts. The walls of the first {fot Leon Dufour's fig. 19*) seem to consist of large cells, round which the tracheae run. They (PL II. fig. 9) are short and broad, and anastomose frequently. On the posterior part {h of Leon Dufour's figure) are numerous dark glandular bodies. A great riblion-like trachea runs along the organ and gives off transverse branches. On one side these branches much resemble trees (PL II. fig. 4), and the glandular bodies look like some enormous kind of fruit. On the other side of the main trachea, the mode of branching is similar, but the branches are more elongated. * " Recherches Anat. et Phys. sur les Ortliopthres, les Hymhiopteres, et haNevroptcres," ' Mem. d. Sav. Etiangers,' vol. vii. OF THE TRACHEAE IN INSECTS. 37 In Gryllus also and Forficula the mode of brandling lias much similarity to that of a tree ; but in neither did I trace the line tubules. In the larva of Acheta the termination of the tracheae, so far as it could be seen, resembled that of the perfect insect. In Melolontlia, Scarahcetis, and Lucanus I met with the same difficulty. In the larva oi Lampyris and in that of Lucanus the distribution was much the same (PL II. fig. 6). In Necrophorus vespillo and N. humator the branchlets are long, waved, not very nu- merous, and without any special peculiarity in the mode of brandling, which is generally more or less dichotomous. The organ is provided with numerous white finger-like glands, on which the waved tubules had a beautiful appearance. Gener illy one or two run up each gland. The distribution of the trachea} in Libellula is very peculiar. Two large trachese run along the stomach and give off about eight large transverse branches on each side in pairs. These lateral branches again give off branchlets at right angles, and these again others at acute angles. These last often run into one anotlier, and thus divide the organ into numerous elongated compartments. In Panorpa about ten large trachea} go to the stomach, on which they ramify in a radiating manner. There are also a few smaller systems, but by far the greatest number of tracheae originate from these large branches. The tracheae of Chrysopa and of Mamestra brassicce (larva) resembled those of Tenthredo and Cynips, but I did not notice such long unbranched tubes. Sometimes the tracheae were here and there in tufts. Those of Campcea margaritaria were still more like those of Tenthredo and Cynips, while Hipparchia again agreed with Chrysopa. In Musca and Eristalis the tracheae are in very pretty tufts. The ends of the tubules often run into the spaces occupied by other systems, but they very seldom inosculate. Sometimes the systems are more elongated, so as almost to lose the appearance of tufts. When magnified about 100 times, the tracheae have a beautiful appearance. In the larva of Musca the stomach is well supplied, principally by transverse branches, which, however, do not end in tufts. On the anterior part especially they are very pretty. In Tipida the trachea; are but very loosely attached to the stomach, so that they may easily be torn off in the dissection. They are not in tufts like those of Musca and Eristalis, but resemble those on the ilium. In Aphrophora spumaria the trachete are not very numerous, and consist of long waved branchlets, with few ramifications, much, in fact, like those of Lampyris and Necrophorus. In Pentatoma the stomach is complicated. PI. II. fig. 2 represents one of the tracheal branches magnified 60 times, and fig. 7 one of the lateral branchlets magnified 250 times, of the anterior part. The branchlets divide dichotomously, and also throw out numerous tubules from the side. The tubules often run close together for some distance. On the posterior sacculated part of the stomach are thick, broad tracheae Avhich send out large branches. These run principally between the sacculi, and appear to branch like those of the front part of the stomach. It is, however, difficult to follow them satis- factorily, on account of the folds. Finally, there is a broader, short chamber, into Avhich 38 MR. LUBBOCK ON THE DISTRIBUTION the Malpighian vessels fall ; and on this the tracheae resemble those of the anterior division of the stomach. I have already given my reasons for considering all these oro-ans as being together homologous with the stomach of other insects, though the great differences in structure certainly indicate differences of function. On the front part of the stomach two or three fine tubules rise from the large tracheae near the base of each system, and run with only one or two branches all along the main branch of the system. What makes these tubules particularly conspicuous is, that while they are nearly as delicate as the ordinary tubules, they are much longer, and as straight as the larger branches which they accompany. On the round sac to which the Malpighian vessels are attached the trachege are much like those of the front part of the stomach. The recurrent intestine of Apliropliora spmnaria is not very richly supplied. The tracheEe (PL II. fig. 16) branch more or less dichotomously, and arc waved at the ends. MalpujUan Vessels. (PI. II. fig. 15 ; PL IV. fig. 11.) In Bombiis terrestris the tracheae run along the Malpighian vessels, giving off short, broad branchlets at intervals. Each of these ends abruptly, and from the termination spring from five to ten tubules. These tubules branch generally two or three times at acute angles, and are about xiu inch long. AVhat, however, gives a very peculiar appear- ance to these tracheae, is, that the tubules of each tuft, and the branches proceeding from them, though somewhat divergent, are more or less parallel to one another. In B. mnscor-um, B. lapidarius, B. Jiortorum, Vespa vulgaris, Anthophora acervorum (PL II. fig. 15), the tracheae are like those of B. terrestris. In OjMon luteum long tracheae run along the Malpighian vessels, and give off side- liranches here and there, which form little systems. In Tentliredo they are as in Ophion, but in Atludia spinarum they seem to have hardly any tracheae. In Acheta (larva and imago), Gryllns, Locusta, and Forjicula, a very long trachea runs along each Malpighian vessel, from one end to the other, giving off minute branchlets at intervals. The ends of these branchlets were ahvays filled with fluid ; and I was there- fore unable to determine their mode of termination. In Lihellula (PL II. fig. 19) this character was carried to an extreme, and it Avas only here and there that a little branchlet could be seen proceeding from the main trachea. In Fanorpa the Malpighian vessels were so opake that the arrangement of the tracheEe could not be made out. In Chrysopa they seem to have no tracheae. In Aphrophora the tracheae were badly seen. They appeared to divide dichotomously into curved, but not waved branchlets. In the larva of Lampyris they were long, waved, and either branched dichotomously, or gave off little branchlets at intervals, very much, in fact, as on the stomach and ilium. In Necropliorus vespillo and N. Immator the l)ranchlets were more numerous, so as to have almost the appearance of being in tufts. They are less wavy than in Bampyris. In Carabiis, Cerambyx moschatus, Ampliimalla solstitialis, Ifelolontha vulgaris, and Bucanus cervus, the ends of the tracheae when examined were filled with fluid, and therefore could no longer be distingtiished. OF THE TRACHEAE IN INSECTS. ' 39 In Pentatoma the tracheae are like those on the stomach. In Camp(sa margaritaria, Pieris napi, and the larva of Eupexia lucipara, the ends became invisible so soon after death, that I was never able satisfactorily to see the finer branchlets. In the larva of Mamestra the trachea? were long, slender, and with few branchlets, at obtnse angles. In Musca and Eristalis they are in tufts, as is so generally the case with the trachea; of these insects ; while in the larva of Musca they are quite simple. A rather larger trachea runs along each Maljoighian vessel in Tipvla, and gives off little systems of branchlets at intervals (PI. II. fig. 18). In the larva of Bombiis mnscorttm the trachese of the white urinary tubes resembled those of the stomach, but the branches generally ended in only two tubules. Ilimn. (PL II. figs. 11, 12 & 17.) In Bonibus terrestris, B. miiscorwn, B. lajridaruis, and B. Itortormii the walls of the ilium are composed of quadrangular cells, and the smaller branchlets of the trachese run round and between them, so that they divide the organ as it were into quadrangular spaces. This arrangement, however, I could not see well without the assistance of acetic acid. In Vespa the arrangement of the trachese is much like that on the stomach. The same is the ease in Ophion also, though the very long branches were not so much developed, and the wliole arrangement is on a smaller scale. At the front part of the organ in Tenthredo the traehete branch frequently, and finally end in tufts (PI. IT. fig. 17). At the posterior end, the tufts are less developed, or altogether absent. In parts of the organ the larger branches anastomose frequently. In Atkalia spinarum the larger branches are like those of Tenthredo. The tufts have fewer tubules, and often the branchlets end Avithout any tuft at all. • In IcTinemnon the trachea? resemble those of Ophion. IwAcheta (imago and larva) the mode of distribution is much like that on the posterior end of the stomach, but at the lower end the tree-like branches are more elongated. In Gryllus also the branches are more elongated than on the stomach. In Forjiciila cmricularia the branches run in a wavy course along the muscles, with short transverse branchlets. In Necropliorus vespillo the ilium is very long, and covered with saccular bags. It has four longitudinal trachea?, which send a branch to each sacculus. In the larvae of Lampyrls and Imcanus the trachese are much like those of the stomacli (PI. II. fig. 6). In Panorpa, on the contrary, they are very different, but I was unable to see them satisfactorily. In Clirysopa again they are similar, but the branchlets are fewer. lArnne- pJiilus vitratus has on the ilium several large tracheae which throw out rather numerous tubules. These latter are rather long and slightly curved, l)ut not waved. The number of them is larger than is generally found on this organ. Campcea margaritaria seemed to have no trachege on this organ, nor on the caicum ; or at least they were so loosely attached, as, in the specimen examined, to have become separated from it. On the long ilium in the larva of 3£nsca the tracheae branch very simply, and resonble 40 MR. LUBBOCK ON THE DISTRIBUTION those of the salivary glands. In JEJristalis the trachere resembled those of the stomach ; in Ilusca they were perhaps rather more like those of the colon. The arrangement in Tiptda is quite different, not being in tufts. In A2oliro2)hora (PI. II. fig. 16) they are much Like those of the recurrent intestine and the Malpighian vessels. In the larva of Bombus miiscorum the tracheae resemble those of the stomach. The different systems cross one another a goo"d deal. The tracheae are stiff and wand-like, and end in small tufts of from two to four, long, straight, or gently curved tubules. These latter sometimes, though not often, branch dichotomously. In the larva of Laminjris the tracheae on the duodenum were like those on the ilium. Colon. (PL IV. figs. 1 to 7.) In Boinhus terrestris, B. muscorum, and B. Icqyidarins, the colon (PI. IV. fig. 2) was but poorly supplied, and the tracheee were long, with only a few branchlets. In Vespa the wall of the colon is divided into six parts, which have a ceUular appear- ance, and are united to one another by their membranes. These membranes have no tracheae, but a large branch runs along each of the divisions, and gives off numerous side branchlets, a part of one of which is represented in PI. IV. fig. 5. In OpJiion lateum the rectum contains a good many round glands ? which, as usual, are well supplied with tracheae. The mode of branching is much like that in Chrysopa (PL IV. fig. 4). In TentJiredo and Atlialia spjinamm the colon contains six ellip- tical sliield-like glands, which, as visual, are more richly supplied than the rest of the organ. The mode of branching is much like that of Ophion and Chrysopa. In Ichneumon also the rectal glands and their tracheae are as in Tenthredo. In Acheta the walls of the organ are divided into six compartments, each of which has a double series of tracheae, two large branches distributing themselves from near the middle, one to the front part, the other behind. The figure (PL IV. fig. 6) which represents the front part of one division will give a more correct idea than any description could do. The colon of Locitsta much resembles that of Acheta. The type is very nearly the same in Giyllns ; but the lateral branchlets are fewer and longer. On the colon of the larva of Acheta the tracheae resemble those of the imago. In Forficula aurictilaria the colon contains six round shield-like glands in two alternate rows. In Carabiis it is well supplied, and the larger branches anastomose a good deal. The muscles make it somewhat difficult to follow the ramifications of the finer branches ; but they seemed to be like those on the egg-tubes. In Ceramhyx the tracheae seemed to resemble those on the stomach, as was also the case with the larvae of Lampyris and Lucaniis. In Necropjhorus vespillo and N. hmnator they are almost as in Bombus ; but the branchlets are longer. In Libellula there are six wide longitudinal bands, connected by a membrane without any tracheae ; each band has a large trachea, which gives off about six systems of branchlets. Panorpa has also six rectal glands, each with a system of tracheee radiating from the centre. In Limnephilus vitratus there were at least twenty- five round glands, on which the distribution of the tracheae was not unlike that of Chry- sopa. Chrysopa has six round glands, as in Tenthredo, &c. The mode of tracheal dis- tribution in the glands and svirrounding membrane is represented in PL IV. fig. 4. In OF THE TRACHE/E IN INSECTS. 41 Campcea margaritaria, Hipfarchia Janira, H. TWionus, Pieris napi, and Noctua gumma^ there are many round glands, each with niiinerous tracheae. In Eristalis and Ilusca the tracheae are as on the stomach, though in the latter the systems are particularly large. In both, there are several tongue-shaped glands with numerous tracheae. In Tlpula also are several similar glands. A large trachea enters the base of each, and there divides into four or five branches, of which about three con- tinue nearly parallel, and give off branchlets from time to time, whUe the others diverge to supply the basal part of the gland. In Pentatoma the trachese resemble those of the stomach = Ilale Generative Organs.— Testis. (PL III. figs. 13, 1-1, 16, 18 & 19.) In Bombus muscorum, B. terrestris, and B. jyratormn, the trachese generally divide dichotomously, but sometimes into three. The end tubules are very long, — a character which we often meet with in the testis. In Acheta, Locusta, and Gryllus, the air had always been removed from the fine ends, which therefore had become invisible. The branches were, however, long between the branchlets, and they tapered very little. In the pupa of Acheta, the trachea?, so far as they could be seen, resembled those of the imago. In Musca the testis is a brown, sausage-shaped body, and the tracheae end in small tufts (PL III. fig. 18). In JEristalis the organ is formed on a similar type. The tufts, however, though larger, are more open, and consist of fewer branchlets. In AjyhrojjJiora the tracheae are almost like those of the ganglia. They end in long, gently curved tubules, which arise successively, and without forming tufts. In the button-like testes of Melolontha long trachese run from the centre to the circumference. They divide three or four times dichotomously, and are very long. They, however, diverge but slightly, and are a good deal twisted on themselves. In Amj^himalla solstitialis they are very similar. PL III. fig. 16 represents one of the least-complicated systems. In Necrojihorus mortuorum and N. vespillo the trachese branch simply, small tubules rise from branches of considerable size, and the end tubules are straight and of great length. In Carabus the ends of the trachese had disappeared. In Cerambyx moschatus they were much like those oi Aphrojyhora spumaria. InPanorpa the trachese were almost as in Necropliorus. Most of the ends had disappeared ; and yet the fine tubules seemed to be very long, because they remained for a considerable distance without any great diminution of size. When one of the branchlets divided, the two tubules were each almost as thick as the branchlet from which they sprang. This character gives the trachea a peculiar appearance, which is very striking. In Pentatoma the testis is covered by broad, longitudinal, saccular trachese, which give off thick branches at the ends and sides. Vas Beferens. This origin is often without trachese. When they are present, they are often very like those on the testis, as for instance in Bombus terrestris, B. muscorum, B.pratorimi, Pen- tatoma prasina, &c. In Acheta and Locusta it was thin, delicate, and apparently a's ithout tracheae. In Grytlus, on the contrary, it is large and of a yelloAV colour. I did not ob- VOL. XXIII. G 42 MR. LUBBOCK ON THE DISTRIBUTION serve tlie ends of the tracliese ; but the larger branches were like those of the testis. In Musca, Aphrophora, Tiimla, Forjicula, and Fanorpa there seemed to be no tracheae. In MHstalis the testes are almost sessile ; and in Necropliorus mortuorum the vas deferens is too short to give a good view of the tracheee. Ductus ejaculatorius. (PI. III. fig. 17.) In Bomhii,s terrestris and in Fanorpa this organ has no tracheae. In Musca the tracheee are in tufts, as in so many other parts of the body. In Necropliorus vespillo and Eristalis they are like those on the testis. In Fentatoma, finally (PL III. fig. 17), the oro-an is pear-shaped, the vas deferens being attached to the swollen end. A number of branches start from the periphery, and soon give off several branchlets at acute angles. These branchlets run towards the posterior end, and go for almost half the length of the oro-an without dividing any more. They are generally straight ; but some were in coils, which, however, may perhaps not be their natural position. Epidklijmis. (PI. III. fig. 20.) PI. III. fig. 20 represents the mode of distribution on the epididymis oi Bomhus 2yratormn. VesiculcB seminales. In Bombus terrestris, B. muscorum, and B. derhamellus, the trachese are in tufts, the tubules often being branched again two or three times : the tj^pe is much like that on the heart ; but each tuft occupies much more space. On the heart I always found the tubviles waved, which is not the case in those of the vesiculge seminales. In Necropliorus mortuorum and N. vespillo they may be seen very well. They are large, and their manner of branching is more or less dichotomous. lu ApjJiropliora spumaria they are like those of the testis, and are therefore not very diiferent from those of Necropliorus. In Fanorpa I did not see them well ; but they seemed to be like those of Necropliorus and Apliropliora. In Musca and Eristalis the tracheEe are in tufts, as is the case throughout these insects. In this organ the tufts are rather large. Female Oenerative Organs.— Ovaries. (PI. III. figs. 1 to 12.) In Bombus terrestris, B. muscormn, B. lapidarius, B. pratorum, B. hortorum, and Vespa, the egg-tubes are covered by such an immense number of trachese, tliat at first sight they look almost like organs of respiration. The mode of distribution also is very different from that found in any other part of the body. In another specimen of Vespa, belonging perhaps to another species, the tubules were fewer, so as more to resemble those of Tenthredo. The large trachese give off short stout branchlets at smaU intervals. These branchlets stop aljruptly, and give off from their end a great number of fine tubules. These latter are so excessively numerous and so much twisted and interwoven together, that I was unable to trace any one to its end, or to determine whether they are branched or not. OF THE TRACHEA IN INSECTS. 43 From analogy with other tufted tracheae, it is probable that the tubules are long, and that they branch occasionally. The impression, however, left on my mind after examining them carefully, was rather the reverse ; but, as above stated, they lie so thickly together, that I cannot speak positively. In Tenthredo the trachese were also in tiifts, like PI. III. fig. 4, which represents those of Tipula, consisting, however, only of from three to ten tubules. These latter branch several times. However, the tracheae are far less numerous than in Bomhus, and have altogether a very different aspect. In Ophion liiteum (PI. III. fig. 7) the contrast is still more striking. The tracheae are few, inconspicuous, and have entirely lost the tufted character. They give off straight branchlets at acute angles ; and the end tubules are long, slender, and straight. The branches do not keep each to a single egg-tube, but pass freely from one to another. In Athalia spinarum the tracheae generally resembled those of Tenthredo, but they were not in such well-marked tufts. In a small species belonging to the Ichneumonidse, the tracheae resembled those of Ophion, biit the branch- lets were waved and twisted instead of straight. In Acheta the tracheae are much like those of OpUon, but on a larger scale. In Gryllus I did not get a good view of the tracheae, and especially not of their terminations. They were more waved than in Acheta. In Locusta (PI. III. fig. 6) they were quite different from those of Acheta, or, so far as I could see them, of Gryllus. The large branches give off short stout branchlets, almost as in JSombns, except that they are more waved. The branchlets end abruptly, and give off tufts of tubules, like those of Tenthredo. The tubules, however, are smaller in proportion to the branchlets, so as to afford a stronger contrast. Moreover they are less frequently branched. In Forficula (PL III. fig. 9) I did not see the fine ends. The tracheae were sunply branched ; and the branchlets were long, much twisted, and of uniform diameter for considerable distances. In Carabus the trachecB branch dichotomously -towards their ends. The systems are rather large. Often two or three fine tubules spring from the side of the branchlets. In Lticamis cervus and Amphimalla soJstitiaUs I was unable to see the fine tubules. In HydropMlus piceus, Necrophorus vespUlo, and N. hnmator, the tracheae were generally in tufts, almost as in Bombus, but the tubules far less numerous. In these insects the tubules certainly di^dde. In Chrysopa (PL III. fig. 3) the tracheae end in tufts. The tubules in each are few in number, straight, and divergent like a fan. They often give off one or two still smaller tubules. Except that they are in tufts, they much resemble those of Ophion. The tracheae inPanorpa, Limnephilus, and I beHeve also in lAbellula, resemble those of Chry- sopa. In CampcBci margaritaria they are in tufts, as in Tenthredo ; but the tufts are not above a quarter as large as in that insect. In Bip)parchia I was only once able to see the tubules. They are represented in PL III. fig. 8. Pigs. 8' and 8" represent all that was visible after intervals of a quarter of an hour. In Pieris brassicce also I was unable to see the finer branchlets. In Tipula the tracheae were like those in PL III. figs. 4 & 5. Unlike as are these two figures, the mode of branching is in reality very similar, and would be seen if the tracheae of fig. 4 were extended as they are on the larger egg-germs. In Eristalis tenax and a large species of Ilmca (PL III. fig. 2), the tracheae were intermediate in character between those of Necrophorus and Bombus. g2 44 MR. LUBBOCK ON THE DISTRIBUTION In a smaller species of Ilusca, however, the tuhules in each tuft were less numerous. In Dioctria JJampes the tracheae resemhled those of T'qmla and Tenthredo. In Fentatom.a a single trachea runs up the terminal chamber of each ovary, and gives off branchlets from the side, as in PI. III. fig. 11. The trachea ends simply. In Ai^lirophora spumaria (PL III. fig. 1) we meet with a type quite different from any of the preceding. Large tracheae proceed to the posterior end of the egg-tubes, where they rapidly break up into branches about Troooth of an inch in diameter. These branches run all up the egg- tube, and half-way along the connecting filament, gradually diminishing in size tUl they can be seen no longer. They are nearly one-sixth of an inch in length, and do not give off a single branchlet. On one side of an ovary I counted sixty of these branches ; and there seemed to be as many on the other side. In Banatra there are five egg-tubes. One single very large trachea, about xoo^h of an inch in diameter, passes to the upper end of the egg-tubes, where they reunite to form the common connecting filament, which was first described by Mliller. Here the trachea divides into about fifteen large trunks, each of which is about -g^th of an inch in diameter. These trunks divide here and there dichotomously, and sometimes give off side branches ; but the distribution of air takes place principally by means of small lateral branchlets, which give off numerous long twisted tubes, and sometimes end finally in a small tuft. This con- tinues for about -^^i\). of an inch, when all the trunks except six have exhausted them- selves,— each trunk ending in the same maimer as one of the lateral branchlets ; but at these ends the tubules seem longer and more waved, and are not in tufts. The six trunks just mentioned pass along the egg-tubes for about xoth of an inch, with a diameter, excluding the so-called outer membrane, of i^jo o^lis of an inch, in which space they do not give off a single branchlet. Eor about half of this distance the egg-tubes are supplied by the other trunlcs ; but for i^otb of an inch after these have ceased, the egg- tubes are left free from trachege. Por the last gVrd of an inch, the egg-tubes, and after that the oviduct and egg-canal, are supplied by the above-mentioned six trunks, and the mode of branching is the same as at the upper end of the egg- tubes. No tracheae are given to the generative organs from any other source. The trachete on the oviduct are less numerous than those on the egg-tubes, and on the egg-canal they are fewer than on the oviduct. It is very seldom that we find a whole system of organs supplied, as in this case, ex- clusively by one great trunk ; and it is evident that here the air must pass to and fro in the same vessel. I was surprised to find six large trimks running along the five egg-tubes. The two ovaries agreed in this respect ; but I have not had an opportunity of examining other specimens, and therefore cannot say whether the number of these trunks varies or not. These six trunks run so far without a branch, and the branchlets arising from the other main trunks are so small in proportion, that the trachese have much the same appearance at first sight as those oi A2')hrophora ; in that genus, however, the main trachea is at the posterior end, while in Banatra it is at the anterior end, so that it is a case of analogy, and not of homology. At the end of October, when I examined this insect, there were no traces of egg-germs ; OF THE TRACHEA IN INSECTS. 45 SO that I am therefore unable to say what relation they would bear to that part of the egg-tube which is free from tracheae. Oviduct. In Bonibus terrestris and several other species, the tracheae of the oviduct frequently anastomose, so as to form meshes. In each mesh is one branch, ending in a tuft. The branchlets composing the tuft are waved, and branch rather often. In B. muscorum the large tracheae do not anastomose so often as in the other species I have examined. In Vespa they do not anastomose, and the branchlets often ended in tufts, with, however, only a few tubules. In Ophion the oviduct is thrown into folds, and the trachese run transversely along them. They are not very unlike those on the ovary ; but the branchlets are waved, which is probably necessary on account of the movements of the organ. In Tenthredo, Locusta, and Carahiis, likewise, the tracheae resemble those on the egg- tubes. In Cynips the tracheae end in tufts, consisting, however, of but few tracheae. In Ichneumon and Forficida I found no tracheae on the oviduct. In Necrophorus vespillo and N. hmnator the tracheae are in tufts as on the egg-tubes, . but the tubules are far less numerous and more branched. They are a good deal hidden by the irregularities of the organ, and altogether look very unlike those on the egg-tubes. In Libellula I did not get a very good view of the trachea? ; but they seemed to resemble those of the ovary. The same is the case in Panorjxt ; but I did not see the tracheae well in tliis insect. In Eristalis the tracheae are, as usual, in large tufts. In Musca and Tipula also the oviduct is very short, but I saw on it several of the usual tufts. In Aphropliora, on the contrary, they are quite unlike those of the ovary, being much smaller and more delicate. They are, however, long and withoiit very many branchlets. In Pentatoma the tracheae resemble those on the stomach. Egg-canal. (PL III. fig. 15.) In Bomhus the tracheae do not inosculate as on the oviduct. The distribution of the smaller brachlets, however, is similar ; but the tufts are fewer, more elongated, and the tubules also less numerous. In Vespa they are nearly the same. In Ichneumon it seemed to be free from tracheae. In Gryllus, Locusta, Carabus, and Necrophorus, the egg-canal is very short. In Panorpa it seemed to have very few tracheae. In Eristalis and Ilusca the tracheae were in the usual tufts. In Aphrop)hora and Pentatoma they are like those of the oviduct. Supplementary O lands attached to the Beproductke Organs. I have already stated that in this paper my attention is confined to the distribution of the respiratory organs. In the names applied to other parts, therefore, I have followed 46 MR. LUBBOCK ON THE DISTRIBUTION simply the best authorities, without wishing thereby to express any decided opinion of my own as to their fvmctions or liomologies. In Bombtis we find two sorts of appendages to the reproductive organs. One is the poison-gland, and consists of two long secreting tubes falling into an oval reservoir. On the tubes the trachea3 are in tufts, but are very unlike those on the Malpighian vessels, though the tubes themselves are very like them. On the reservoir the mode of branching is simple and successive. The tubules are straight, or at least gently curved. Between the ovaries of Aphroj)hora spimiaria lies a membranous bag with tracheae much like those of the ganglia in Aclieta. In one specimen, however, they were waved. On the spermatheca of Bombus muscoriim tracheae were few, and the mode of branching simple. In Hydropliilus jncetts there are certain ramified glands with swollen ends. On these glands the tracheae resemble those of the stomach. Heart. (PL I. figs. 15, 16 & 17.) In Bombus terrestris, muscormn (PI. I. fig. 15.), lapiclarius, pratoriim, and hortomm, Vespa mdgaris, and Apis mellifica, the larger tracheae branch freely along the margins of this organ, and anastomose frequently with one another and with the neighbouring systems, besides sending branches which run into those of the opposite side. The branches also give out short branchlets, which do not divide nor diminish in size, but terminate suddenly and give off from their end several smaller branchlets, which, again, divide more than once, so as to end in very fine, more or less waved tubules. The branchlets into which a branch divides are often of very unequal size. 'This character is, however, not common to all the Hymenoptera ; for in Opliion luteuni and in one of the Ichneumonidae, though they had in most parts been torn off, yet in one or two places where they remained they divided into more or less waved tubules, by simple successive branching. This is also the case in Imniiep)hilus vitratus, where, however, the tubules were larger ; moreover they were much obscured by the surrounding fat. For the same reason it was difficult to see them well in Ap>hrophora, wbere, however (PL I. fig. 17.), after dividing several times, and generally dichotomously, they appeared to end in a tuft of long and apparently simple tubules. In the specimens of Aclieta and Grylhis viriilissimzis which I examined, I was unable to make out the arrangement of the tracheee ; and in Locnsta I was scarcely more for- tunate, but they seemed to end in tufts, almost as in Aphrophora. In Pentatoma the distribution was almost the same as on the oviduct, but generally with fewer tubules. They were sometimes waved, sometimes nearly straight. In Necrophorus they branched simply and were straight, or rather with gentle curves. After adding acetic acid, many of them were thrown into waves. In Lucanus cerviis they were in folds, long, and with few branches, but all the ends had disappeared, as was also the case in Carabus and in Noctua gamma. In Eristalis, Miisca (PL I, fig. 16), and T'qmla, the tracheae are in small tufts consisting of from three to six tubules. These latter s^ive off smaller ones from their sides. OF THE TRACHEA IN INSECTS. 47 Fatty Tissue. (PI. IV. figs. 8, 9 & 10.) In Tenthredo the fatty tissue consists principally of large round cells. The tracheae are peculiarly straight and stiff-looking ; they branch occasionally, and finally end in small tufts of from two to six tubules. In Vespa the tracheae resemble those of Tenthredo. In Bonibiis muscorum the branches of the tracheae have the same straight, even cliaracter, and often run for some distance without branching. They end in tufts of straight tubules, which are generally not very numerous and do not branch often. In Hydrophihis picetis also the tracheae are in tufts. In the larvge of Lmnpyris and Lucanus, and in Scarcibaius, the fatty tissue consists of round masses, on which I did not ascertain the mode of distribution of the tracheae. In Forficula the fatty tissue is in large flakes, here and there running into one another. The tracheae are long and rather thin in proportion. They branch seldom, and generally at obtuse angles. Wlien the flake is broad, the trachea given to it generally emits several branchlets. When it is narrow and ribbon-like, a small branch generally runs along it, and divides only at long intervals. In the larva of Lasiocampa ruM and of Mamestra I)7'assicce, in Aphrophora spumaria, in Tijmla, and in Acheta domestica, the tracheae are nearly as in Forficula, and they break up gradually into a few, long, straight tubules. In Pentatoma the tissue consists of thimble-shaped lobules, connected by their bases. A trachea enters each, and breaks up gradually as in PI. IV. fig. 9. In Cynips Ugnicola (PI. IV. fig. 10) the arrangement is very peculiar. The tracheae are shaped lilce rolling-pins, and scarcely taper at all, being, even at then' ends, too o^h of an inch in breadth. From their ends and sides spring nu- merous fine tubules, about roooo^h of an inch in diameter, and as much as, or even more than, a^th of an inch in length. The tubules scarcely ever branch, though in one or two cases I saw a dichotomous division. Ifuscles. (PI. IV. figs. 12 to 18.) On the muscles of the abdomen in JBombus and Vespa thus end in tufts. The thoracic muscles of these two genera have a number of large, parallel, saccular, transverse tracheae, separated by intervals scarcely greater than their own width. They end suddenly in a number of very short, thick branchlets (PI. IV. fig. 14) ; and the saccular branches give off similar little systems from their sides. In Ophion the system is similar, though the branchlets are more elongated. In Bomhtis also the tracheae of the thoracic muscles are similar ; but I could follow the terminal tubules rather further. The thickness of the muscle, however, prevents them from being seen well. In Panorpa (PL IV. fig. 12) the tracheae on the abdominal muscles are straight, and the tubules expand like the rays of a fan. On the muscles of the thorax the tracheae are sometimes waved ; but the mode of branching may be best understood by referring to PI. IV. fig. 18. In Atlialia spiuarum the tracheae often end in tufts. The branchlets run across the muscles, and divide like a fan, but rather irregularly. Often the side branchlets diverge at first, and then curve round so as to become subparallel. Sometimes, 48 MR. LUBBOCK ON THE DISTRIBUTION on tlie other liancl, a side branclilet forms a separate system at riglit angles, or nearly so, to the first. On the thoracic muscles of Chrysopa the branches end in small fan-like tufts of tubules. In the larva of Lasiocamim rubi the tracheae are niimerous, simply branched, and much wared. In the larva of Mamestra brassicce they are very similar, but straighter. In Hij^parchia Janira, Pieris brassicce, and AinpMmalla solstltialis, the tracheae break up into straight tubules, which diverge from one another. In Eristalis the thoracic muscles consist of broad ribbon-like tracheae, which lie parallel, and near to one another, across the muscles. They end in finger-like prolongations, from the end of each of which rises a small tuft of tubules, which are apparently very short. In Musm the tracheae were very similar. In Tipula the tubixles were longer than in Miisca or Eristalis. In Necrophorus the saccular tracheae resemble those of Eristalis ; but the tubules are longer. In the Cricket they resemble those of Fanorpa (PI. IV. fig. 18). Some parts more resemble PI. IV. fig. 12. Locusta and Gryllus did not apparently differ from Acheta. In the larva oi Acheta the tracheae in some parts resemble those of the imago. In parts I found long ribbon-like tracheae, somewhat as in Eristalis ; only they are longer, nar- rower, and further apart. The tubules also which arise from them are long and only ln*anched hei'e and there. Apliropliora has tracheae of the usual type. In Pentatoma the parallel saccular tracheae are narrower than in Eristalis, being only about -fooo^^ of ^"^ iiich in diameter. The terminal tubules also seemed longer ; but I did not get a very good view of them. Conclusion. It would of course be rash to attempt to lay down general laAvs based on the examina- tion of the few species which as yet I have been able to compare together ; but so far as they go, my observations appear to point to the following conclusions. Eirst, that the same type of distribution regularly occurs in the homologous parts of different specimens belonging to the same species. Secondly, that the same almost always holds good of homologous parts in different species belonging to the same genus. Thirdly, that though the general type of distribution is the same in diflferent specimens of the same species, yet the individual tracheae differ very much, — just, in fact, as we find the general mode of branching is the same in different specimens of the Oak or Birch, though in no two oaks or birches are the various branches and twigs exactly alike. Fourthly, that while in some insects, as for instance in Peutatoma, the tracheae in many of the different organs have a different type of distribution, in others, as in Eristalis, one type is found in most of the different organs. Even in this case, however, the tufts (PI. I. fig. 16) which end the branchlets, are very different in size in the different parts. A small organ has, as a general rule, more systems of tubules in a given space than a large one ; but the relative proportions are, of course, not exact. Moreover, even in this case, some organs (as, for instance, the muscles) are very unlike the remainder. Eifthly, while in the larva of Orthoptera the tracheae very closely resemble those of the OF THE TRACHEA IN INSECTS. 49 imago, in those larvas which undergo a perfect metamorphosis (as, for instance, in those of Coleoptera, Lepidoptera, and Hymenoptera) the tracheae are very unlike those of the perfect insect. Sixthly, in these latter larvse the same type of tracheal distribution appears to be more widely distributed in the different organs. Thus, while in the imago of Bomhns the trachese on the stomach, ilium, and Malpighian vessels are all three very different from one another, the trachese on the corresponding organs of the larva of B. mtiscoriim are very similar to one another, and as in PL II. fig. 11. Again, in the larva of Luccmns the oesophagus, stomach, ilium, and colon, and in that of Lampyris the stomach, ilium, colon, and Malpighian vessels, had trachese respectively like one another. Seventhly, that in some cases the trachese of the larvae seem to agree in species where those of the perfect insect are different; so that, comparing this with the preceding rule, it would seem that the larval trachese are in both respects less differentiated, and more in accordance with the original type than those of perfect insects. DESCRIPTION OE THE PLATES. Plate I. Fig. 1. Trachese from ganglion of Acheta, x 125. Fig. 2. Tracheae from commissure of Carabus, x 125. Fig. .3. Trachese from ganglion of Bombus, x 125. Fig. 4. Trachete from commissure of Acheta, x 125. Fig. 5. Trachese from ganglion of Pentaioma, x 125. Fig. 6. TracheEe from ganglion of Carabus, x 125. Fig. 7- Trachese from commissure of Lasiocampa, x 125 Fig. 8. Trachese from oesophagus of Eristalis, x 125. Fig. 9. Tracheae from sucking-stomach of Musca, x 125 Fig. 10. Trachese from rsecum of Achet a, x 30. Fig. 11. Tracheae from caicum of C7i?'2/5o/)ff, x 125. Fig. 12. Trachese from c?ec\im.of Chrysopa, x 125. Fig. 13. Trachese from crop of Carabus, x 30. Fig. 14. Trache8efromcEsophagusofMMsc«, x 125. Fig. 15. Tracheffi from heart of Bombus, x 125. Fig. 16. Trachea; from heart of Musca, x 125. Fig. 17. TrachefE from heart of Aphrophora, X 125. Plate II. Fig. 1. Trachese from the stomach of Musca, x 125. Fig. 2. A few of the larger trachese from tlie stomach of Pentaioma, omitting the finer branches, x 30. Fig. 3. Tracheae from the stomach of Ophior„, x 125. Fig. 4. Trachese from the stomach of Achet a, x 60. Fig. 5. Trachese from the stomach of Bombus, x 125. Fig. 6. Trachese from the stomach of Bombus, x 125. Fig. 7- One of the lateral branches from the above figure of Pentatoma, showing the fine branchlets, x 125. Fig. 8. Trachese from the stomach of Mamestra, x 30. Fig. 9. Trachese from the stomach of Acheta, x 30. Fig. 10. Tracheae from the stomach of Vespa, x 125. Fig. 11. Trachese from the ilium of the larva of Bombus, x 125. Fig. 12. Trachese from the ilium of Tipula, x 125. Fig. 13. Small part of the inner membrane of a trachea, showing the spirally thickened rib, aaa, x 125. Fig. 14. Trachese of salivary gland of Pentatoma, x 125. Fig. 15. Trachese of Malpighian vessel of Atithophora, x 125. Fig. 16, Trachese of recurrent intestine of Aphrophora, x 30. VOL. XXIII. H 50 MR. LUBBOCK ON THE DISTRIBUTION OF THE TRACHE.^ IN INSECTS. Fig. 17. Trachcce of ilium of Tenthredo, x 125. Fig. 18. Trachese of Malpighian vessel of Tipula, x 125. Fig. 19. Trachcce of Malpighian vessel of Libelhda, x30. Plate III. Fig. 1 . Tracheae from ovary of Aphrophora, x 30. Fig. 2. Tracheae from ovary oi Musca, x 125. Fig. 3. Tracheae from ovary of Chrysopa, x 50. Fig. 4. TrachccE from ovary of Tipula, x 125. Fig. 5. Tracheae from ovary of Tipula, x 125. Fig. 6. Trachece from ovary of Zoras/a, x 125. Fig. 7- Trachete from ovary of Ophion, x30. Fig. 8. Tracheae from ovary of Hipparchia, x 125. Fig. 9. Tracheae from ovary of Forficula, x 125. Fig. 10. Tracheae from ovary of Necrophorus, x 125. Fig. 11. Tracheae from ovary of Pent at oma, x 30. Fig. 12. TrachccE from ovary of Acheta, x 30. Fig. 13. Tracheae from testis of Necropliorus, x 125. Fig. 14. Tracheae from testis of Bombus, x 125. Fig. 15. Tracheae from egg-canal of -Boniiws, x 125. Fig. 16. Tracheae from testis of Melolontha, x 125. Fig. 17- Tracheae from ductus ejaculatorius of Pent at oma, x 125. Fig. 18. Tracheae from testis of Musca, x 125. Fig. 19. Tracheae from testis of Eris talis, x 125. Fig. 20. Tracheae from epididymis of Bombus, x 125. Fig. 21. Tracheffi on a Malpighian vessel of Lib ellula. Plate IV. Fig. 1. Tracheae from the rectum of Musca, x 125. Fig. 2. Tracheae from the rectum of Bombus, x 125. Fig. 3. Tracheae from the rectum of Pentatoma, x 125. Fig. 4. Tracheae from the rectum of Chrysopa, x 125. Fig. 5. Tracheae from the rectum of Vespa, x 125. Fig. 6. Tracheae from the rectum of Acheta, x 30. Fig. 7- Tracheee from the rectum of Forficula, x 30. Fig. 8. Tracheae from the fatty tissue of the larva of Mamestra, x 125. Fig. 9. Tracheae from the fatty tissue of Pentatoma, x 125. Fig. 10. Tracheae from the fatty tissue of Cynips lignicola, x 30. Fig. 11. Tracheae from the Malpighian vessel of Musca, x 125. Fig. 12. Tracheae from the thoracic muscles of Panorpa, x 125. Fig. 13. Tracheae from the thoracic muscles of the larva of Acheta, x 125. Fig. 14. Tracheae from the thoracic muscles of Vespa, x 125. Fig. 15. Trachete from the abdominal muscles of the larva of Mamestra, x 125. Fig. 16. Tracheae from the thoracic muscles of Eris talis, x 125. Fig. 17. Tracheae from the muscles of the larva of Lasiocampa, x 125. Fig. 18. Tracheae from the thoracic muscles of Panorpa, x 125. *;,,* It has been found impossible to make the ends of the trachese as delicate as they ought to be in proportion. Iran/S.J.zTi'n,. Soc. VoZSSIL.ta/i.J,^. J~.Z,v.iiock, I G-.Jcw7ru)/rb. so. Sum/sJJinn^. Soo. Vnl, XX/// t/rh , '. -i.uiioo&, del G-Jizi) Jrans.J^i-n/fv. Soa Vol XX//I, tai. 3. ', ^ Ziohioo^, J^l. &. Ja/n7i>3.n. so. ^Wz5. Jiw-^. Soo. VolMHtah. d-.'/i. ,T. Jkihiook, Jd. 6^.Jajn7vcun',sC: [ 51 ] III. On some new species of Musci and Hepaticse in the Herbarium of Sir W. J. Hooker, collected in tropical Africa, chiefly hy the late Dr. Vogel and Mr. Barter. By William Mitten, A.L.S. Eead March loth and June 21st, 1860. Leptotrichum, Hampe. L. NITIDULUM. Dioicuin, gracillimum, humile ; foliis erecto-patentibus e basi latiore sensim angustatis lanceolatis obtusiusculis integerrimis, nervo ubique a pagina distincto percurrente, cellulis basi infima majoribus oblongis inde angustis mollibus ; pericbsetialibus obtusioribus ; tbeca in pedunculo trilineari gracillimo luteo erecta, late ovali, ore parvo, opereulo subrecte subulate, peristomio dentibus parvis rubris rugulosis solidis, annvilo lato composito trigono ; j)lanta mascula ramis brevibus floribus gemmiformibiis termiiiatis. — L. exigno simile. Hab. On the earth. River Niger [Barter). More slender than L. exiguum.= Wiessia exigua, Sch.\v. = Dicranum xanthodon, Hook., inhabiting the West Indies and South America, and with more shining pale-green leaves. Tab. V. fig. 1. 1. Plant (of the natural size) ; 2. a leaf, 3. the capsule, 4. the same after the fall of the operculum, 5. a portion of the peristome, and 6. a fragment of the annulus seen transversely (all magnified). Stereodon, Bridel ; Mitten, Muse. Ind. Orient, in Linn. Journ. Bot. Suppl. vol. i. p. 92. S. Barteri, Mitten. Monoicus ; foliis imbricatis late ovatis bre^d-apiculatis coneayis, maro-ine (sieco) recurA^o integerrimo, nervis binis brevissimis, cellulis alaribus pluri- bus quadratis viridibus, superioribus elongatis angustis ; pericbsetio magno ; foliis e basi ovata convoluta lanceolatis, internis erectis, parapliysibus exsertis; tbeca in pedunculo brevi luteo pericbsetium paululo superante ovali, peristomio dentibus ? — Habitu S. lovgiseti, sed gracilior. Hab. On trees in deep ravines. Nupe ; River Niger [Barter, 1422). Pale yellowish green, shining. Remarkable on account of its very short seta ; in other respects closely agreeing in structure with S.longketus = Neckera lonyiseta, Hook. Tab. V. fig. 2. 1. Plant (of the natural size); 2. a leaf; 3. perichfetium, male flower, and capsule (magnified). Stereophyllum, Mitten, Muse. Ind. Orient. I. c. p. 117. S. nitens, Mitten. Monoicum ; foliis elliptico-lingulatis basi contractis apice subacutis, nervo temti medio evanido, marginibus superne subserratis, cellulis basi ad angulos quadratis inde ad apicem elongatis angustis parietibus crassis ; perichajtialibus parvis ovatis acnminatis parce serrulatis ; theca in pedunculo gracili rubro parva ovali incli- nata horizontalive, opereulo brevi-rostrato, peristomio interne processibus longitudine h2 52 MR. MITTEN ON SOME NEW SPECIES OF MUSCI dentes sequantibus in membrana ad tertiam partem loiigitudinis dentium exserta. — Habitu staturaqiie ^S*. radiculosi. Hab. On decayed wood, Fernando Po [Barter). Very nearly resembling S.radiculosum — Hookeria radicuhsa, Hook. Muse. Exot. t. 5 1 ; but with leaves of a more equal width above and below the middle, and contracted at the base, the margin slightly serrulate, and the upper cells elongate. S.schistocalyx [Hypnum), C. Miillei', has shorter leaves, more sharpened above, and with the upper cells nearly round. Tab.V. fig. 3. 1. Plant (of the natural size); 2. a leaf ; 3. capsule; 4. perichffitium and male flower; and 5. a portion of the peristome (all magnified). S. MVULAKE, Mitten. B>amis elongatis ; foliis subcompressis ovato-lingulatis acutis basi contractis, nervo crasso paulo infra apicem evanido, marginibus superne erosis ; cellulis superioribus minutis ovalibus obscuriusculis, basi brevioribus subquadratis. — ^S*. radlculoso robustius. Hub. On rocks in small rivulets at the bottoms of ravines. Nupe ; River Niger [Barter, 1420). In the firmer substance of its leaves allied to ., F.B.S., F.Z.S., F.G.S., and Edouard Claparede, M.D., Felloio of the Physical Society of Geneva. ReadJanuary 19th, 1860. Having been fortunate enough to capture a considerable number of specimens of this interesting Annelid during our joint sojourn at Lamlash Bay, Arran, during the month of September last, and having been able to add much to the knowledge previously ob- tained of its organization and reproduction, we have agreed to lay the results of our observations before the Linnean Society, in the form of a continuation of the memoir already communicated to it by one of iis, and printed in its Transactions (vol. xxii. p. 353). Although we have miich to add, we find nothing to correct in any of the positive state- ments of fact which that memou- contained ; and we have the satisfaction of bearing our joint testimony, based on a renewed comparison of Mr. George West's delineations with the objects from which they were taken, in regard to the truthfulness with which they represent the conformation of this interesting animal*. In the first place, with respect to the specific diversity affirmed by Drs. Leuckart and Pagenstecher t to exist between Tomopteris onisciformis and T. quadricornis, chiefly on the ground of the presence in the latter of a pair of cephalic appendages not possessed by the former, we have to state that the conviction already expressed by one of us % as to the insufiiciency of this character has been fully borne out by our subsequent observations, which have entirely satisfied us that the presence or absence of the appendages in ques- tion depends solely upon the grade of development which the individual has attained. We have been led, however, by certain minute differences between the organization of the Tomopteris described by Leuckart and Pagenstecher, and that of the Tomopteris which we have studied, to suspect that the forms we have severally described under the designation T. onisciformis may not be specifically identical. The T. onisciformis and the T. quadricornis of Leuckart and Pagenstecher both possess that remarkable pair of " frontal horns," projecting laterally from the most anterior part of the head, which, as Mr. Huxley has remarked, give to the animal the aspect of a hammer-headed shark ; and they both possess that pair of greatly elongated appendages designated in the former memoir as the " styliform," but which we now prefer to term the second antennce. These organs are far longer, relatively as well as absolutely, in the * In transferring these figures to copper, and in reducing their scale, the engraver has imparted to them a hard- ness and stiffness which the originals do not possess ; and has also, by too deeply shading them, destroyed the effect of transparency which the artist had aimed to give. t Miiller's Archiv, 1858, p. .588. % See vol. xxii., supplemental note in p. 362. i2 60 DRS. CARPENTER AND CLAPAREDE ON TOMOPTERIS ONISCIFORMIS. advanced stage of tlie life of Tomoj)teris than in the earlier (compare figs. 1 and 6 of Plate LXII. vol. xxii.) ; and we have found that this difference results from a gradual increase in their proportional length, which bears a constant relation to the stage of deve- lopment attained by the body. In an example of this type which we believe to be one of the earliest yet met with (fig. 14, PI. VII.), these second antennae {b, b) were very short, and presented a much nearer resemljlance to the ordinary lateral appendages of the. body than they do at a later period. But the distinction between T. onisciformis and T. qua- dricornis is mainly rested by Drs. Leuckart and Pagenstecher upon the absence in the former of the pair of cephalic appendages borne by the latter, between the " frontal horns " and the " second antennse; " these, which were designated in the former memoir as the " second pair of horn-like appendages," will now be spoken of by us as the first antenn'ce, since it is clear that they have more relation to the pair of appendages immediately behind them than to those in advance of them. Por, in common with the " second an- tennge," they are setigerous, each of them bearing two setae within its terminal portion (fig. 6) ; and in the early form just now alluded to, in which the " frontal hoi'ns " were entirely wanting, these first antennse were of such a length as to be the principal append- ages of the head («, a, fig. 14). A careful study of the successive stages of the develop- ment of Tomopteris, in fact, makes it obvious that these first antennce are to be con- sidered as larval, and the second as characteristic of the adult ; for the former progres- sively diminish in relative size, and at last (in most cases) disappear altogether, whilst the latter progressively increase both absolutely and relatively. In the stage represented in fig. 1, Plate LXII. (vol. xxii.), which seems to correspond to the T. quadi'icornis of Leuckart and Pagenstecher, the setigerous portion of the first antennae is separated from the basal by a constriction which suggests the idea of an articulation (PI. VII. fig. 6. a) ; about the time when the caudal appendage begins to be developed, the setigerous portion usually detaches itself, while the basal part remains as a mere knolj or tubercle ; and in the most advanced forms, this tubercle is commonly fou.nd to have entirely disappeared. A vestige of it, however, is still to be seen in some specimens ; and it would appear, not only from the observation made by one of us last year, but also from the mention of these appendages by Mr. Huxley (p. 359), who described and figured them as " long, curved, spine-like processes arising from the ventral surface of the narrow neck," that these first antennae are occasionally retained even in the stage of advanced develop- ment. It is obvious, therefore, that no distinction between Tomoiiteris onisciformis and T. quadricornis can be drawn from the presence or absence of the first antennae ; and if no other constant and important characters of difference can be specified than such as clearly arise out of the respective ages of the specimens, the last-named species must be abolished. In stating the other additional results of our observations, we shall for the most part enumerate these, for the sake of convenience, in the order in which they occured to us in proceeding from before backwards along the body of the animal. In the advanced specimens (by which we mean those in which the caudal prolongation is beginning to show itself), the head bears on its dorsal surface a pair oi ciliated epaidettes (PL VII. fig. 5. a, a), which extend over the edges of the bilobed nervous ganglion. These, DRS. CARPENTER AND CLAPAREDE ON TOMOPTERIS ONISCIFORMIS. 61 at a certain stage of development, are fringed with long cilia Ijotli at their margins and at their base ; but the cilia are only occasionally to be seen in activity ; and we suppose it to have been either from the absence of the cilia, or from their not having been in motion, that they escaped the notice of Mr. Huxley, who describes the epaulettes as rounded elevations on either side of the narrow neck, from which a sort of band or ridge runs back upon the dorsal surface (vol. xxii. p. 359). No trace of these organs is to be seen in the earlier stages of the animal's life. The lens-like body in each of the pair of ocelli (fig. 12. b, b) which are seated upon the bilobed ganglion, is not single but double, as has been already stated by Leuckart and Pagenstecher. We were able clearly to distinguish the spheroidal cells of which the biloljed ganglion is composed ; and in its anterior part we noticed (as Leuckart and Pagenstecher had pre- viously done) a pau' of vesicles imbedded in its substance (fig. 12. a, a). These vesicles are nucleated; but we could not distinguish either otoliths or ciliary movement in their interior. How far they are to be considered as analogous to the auditory vesicles of other Invertebrata, we are therefore unable at present to determine. After a careful search for nerve-fibres in connexion with these ganglia, we feel unable to speak positively either as to their presence or their absence. There is certainly some appearance of a nerve-fibre on either side passing from the ganglion towards the frontal part of the head (as described by Leuckart and Pagenstecher), and also of one passing into the base of the second antennae (as noticed in the former memoir) ; but we are not at all sure that these appearances are to be trusted, having now satisfied ourselves that the linear trace which passes on the median line along the dorsal surface is not (as was suggested in the former memoh' that it possibly might be) a band of fibrous nerve-sub- stance, but is simply the mark of the attachment of a mesentery by which the intestine is held in its place. We have been unable, notwithstanding our careful and repeated search for it in living specimens, to detect the ventral nervous cord and oesophageal ring described by Grube (see vol. xxii. p. 361) ; and we cannot but believe that he must have been deceived by appearances prodiiced by the change which the textures of this deli- cate creature had undergone in the process of conservation to which his specimens of it had been subjected. The observation of a considerable number of specimens in various stages of develop- ment enables us now to state that the alimentary canal may present itself at any period either in the state of contraction shown in Plate LXII. fig. 1 (vol. xxii.), or in that of dis- tention shown in Plate VII. fig. 6, or in any intermediate condition. Even when the canal has been most distended, however, it has generally seemed to contain little else than water, very few solid particles being visible in its interior. In one individual, however, we ob- served in the intestine several fragments of a Beroe which had been captured at the same time and had been broken in the tow-net ; and these fragments were kept in continual agitation by the persistent movements of their large cilia. This unusual exercise of vora- city seemed to have an injurious efi'ect upon the individual; for it was found dead th^ next morning, whilst several other specimens in the same jar survived. The caudal pro- longation of the body is generally in a state of greater or less contraction longitudinally ; 62 DRS. CARPENTER AND CLAPAREDE ON TOMOPTERIS ONISCIFORMIS. and the alimentary canal is thrown by this contraction into convolutions, which are sometimes slight and easily distinguishable, as shown in Plate VII. fig. 2, but which, when the caudal appendage is greatly contracted (as shown in figs. 6 & 7 of Plate LXII. vol. xxii.), may be so close as to render it difficult to follow their course, by reason of the opacity which the part then acquires. The part of the intestine contained in the caudal prolongation has its external surface clothed with cilia; but these are most apparent along particular bands. Cilia are also distinguishable on certain parts of that innermost layer of the general integument which forms the external boundary of the perivisceral space, their action being especially apparent near the ends of those lateral appendages which give support to the pinnulse. By the agency of these cilia a more special move- ment is imparted to the corpuscles of the fluid contained within the perivisceral cavity, than that which they receive from the movements of the body generally, and from the peristaltic contractions of the alimentary canal. Although these corpuscles, which are usually of a spheroidal form and of a diameter of about -g^footh of an inch, very com- monly float singly, they are often to be seen aggregated in smaller or larger numbers into masses, which sometimes attain a considerable size (fig. 10) ; and we cannot doubt that the bodies supposed by Mr. Huxley to be young spermatozoa (p. 359, vol. xxii.) were of this nature. We have given much attention to the ciliated canals, first observed by Leuckart and Pagenstecher, which originate in two orifices near the base of the lateral appendages of each side (figs. 7, 8. a, b) on their dorsal aspect, and which then rapidly incline towards each other, so as to converge into a single canal, that runs along for some distance in the wall of the body, and then terminates in the perivisceral cavity (fig. 7. c). One of these orifices {a) is situated in the centre of a sort of rosette*, marked by radiating ridges furnished with large cilia ; the other {b) has round it a smaller rosette, which does not possess any such ciliated ridges. These ciliated canals are obviously the homologues of those which attain so much greater a complexity in the higher Annelida (the " segmental organs " of Dr. T. Williams). Although they are represented by Drs. Leuckart and Pagenstecher as existing in their specimens on every one of the pinnulated appendages, and even in the bases of the second antennae, we can state most positively, both from the observation of numerous living individuals, and from the evidence of well-preserved specimens now be- fore us, that, in the form of Tomojyteris observed by us, they do not exist in the five lateral pairs which immediately follow the second antennae. And this fact is the more remarkable, since, as Avill presently appear, there is a stage in the animal's life at which it does not possess more than five pairs of pinnulated members, and in which, therefore, it is entirely destitute of ciliated canals. The direction of the ciliary current always appeared to us to be from without inwards (as affirmed by Leuckart and Pagenstecher, in opposition to what is generally stated on this point) ; and we repeatedly witnessed the passage through the canals, in this direction, of spermatozoa which had been emitted from the testes into the surrounding water. Each of the pinnulae, in our specimens of Tomopteris, presented (fig. 7. d, cl) the peculiar * This rosette was noticed by Buscli ; but he onlv imperfectly made out the canal of which it is the entrance. DRS. CARPENTER AND CLAPAREDE ON TOxMOPTERIS ONISCIFORMIS. 63 " rosettenformiges Organ" which was noticed by Leackart and Pageustecher upon the two first pairs of appendages of their T. quadricomis, and which is affirmed by them to be entirely wanting in T. onlsciformis. This is alike remarkable for its form (fig. 9), which somewhat resembles that of a melon, and for its colour, which, being a bright sulphur- yellow, distinguishes it from the coloured spots noticed in the former memoir, these being of a reddish orange. In the very young specimen already adverted to, these organs presented themselves at the bases of the pinnulaj (tig. 14. c, d) of the two most developed pairs of members. Of their nature and purpose we have not been able to form even a probable guess. It is in regard to the caudal prolonffation of the body, and the organs which it contains, that we have the most novel and important information to offer. We can now state with certainty that the large specimen represented in Plate LXII. figs. 6 & 7 (vol. xxii.) was a male,— the eight pairs of ovoidal bodies from which the rudimentary pinnules appeared to spring being really the testes, which occupy the parts of the perivisceral cavity that are prolonged into the short lateral appendages whereon these pinnules are really borne. Each testis (PL VII. fig. 2. a, a) is an undivided sac, whose cavity, when the organ has attained its maturity, is almost entirely filled with a mass of spermatozoa. The individual parts of this mass are in continual movement upon each other, their motion being kept up chiefly, if not entirely, by the action of the cilia clothing that part of the inner wall of the testis which is near its external orifice. Each testis (fig. 3) can discharge its contents, either externally through an orifice («) in the wall of the lateral appendage within which it is lodged, or internally through another orifice {b) into the perivisceral cavity. That the external orifice is distinct from tliat of the larger, ridged rosette (c) of the cUiated canal, we feel om'selves able to affirm -wdth certainty ; but ^ve are not ec^ually sure of its distinctness from that of the smaller rosette in its neighbourhood. The internal orifice, by which the spermatozoa escape into the general cavity of the body, can only be seen occasionally ; and we are disposed to think that it is formed only when the contents of the testis are fuUy matured. Whenever we caused these to be discharged by pressure, it was through the external orifice that they escaped. The spermatozoa of Tomopteris (fig. 4), which we had abundant opportunities of ob- serving, are peculiar in having two flagella — a feature which, although general among the antherozoids of the Algae, is very rare (if not unique) among the spermatozoa of ani- mals *, — and also in being able to move in either direction with apparently equal facility. It will presently appear that the testes occupy exactly the same place in the lateral appendages of the caudal prolongation as the ovaria do in those of the body generally, — the chief difference in their apparent positions arising from the smaller size of the lateral appendages in the former region, and from the more complete occupation of the cavity of these appendages by the oval sacs of the testis. * In his 'Naturalist's Rambles on the Devonshire Coast,' Mr. Gosse mentions (p. 213) that he witnessed the exit from the mouth of Pedicellina BeJgica of numerous minute pear-shaped bodies, each having " a little tubercle at its larger end, around which are set a few (about four or five) long ciha or setse, twice or thrice as long as the body. These are not used for vibratile action, but as oars slowly waved through the waters." Although these are termed " germs " by JNIr. Gosse, yet it seems to us probable that they were multiflagellate spermatozoa. 64 DRS. CARPENTER AND CLAPAREDE ON TOMOPTERIS ONISCIFORMIS. A considerable number of males came under our notice before we met with any specimens whose female sexuality was indicated by the presence of ova : of these, how- ever, we subsequently encountered several ; and our observations on them are in accord- ance with those previously made by Huxley, Busch, and Leuckart and Pagenstecher. In a mature female (fig. 1) the ova are to be seen lying, in great numbers and in various stages of development, in every part of the perivisceral cavity of the body and of its caudal prolongation, and also in the extensions of that cavity into the lateral appendages, including even the basal portion of the second antenna?. From a compa- rison of the conditions under which they may be seen in one and the same individual, it becomes obvious that they originate in the terminal wall of the cavity of the pin- nulated appendages, — a fact that seems to have been first recorded by Drs. Leuckart and Pagenstecher, though we learn from Mr. Huxley that he had arrived at the same con- clusion ten years ago. This terminal wall, which gives support on its outer side to the bases of the pinnulse, presents on its inner surface a number of transparent tubercular elevations (fig. 7. Animal. Dentition triserial. To these may be added Brownia, and those species which are included in a cartilaginous case. J. D. MACDONALD ON THE CLASSIFICATION OF THE GASTEROPODA. 73 General Classification of the Gasteropoda. I am thus insensibly led into the subject of classification, which I must say I approach with considerable diffidence, being fully conscious of the danger of falling into a dictatorial style, where the results of my ovm researches are at variance with the views of manj worthy cultivators of the science of Malacology. It is quite foreign to my present pur- pose to enter upon the analysis of any single system. Well-informed zoologists naturally make what they believe to be truthful selections from all available systems, and thus, as it were, compound their own creed ; hence I have to deal rather with such matters as ar e generally accepted than wdth the tenets of particular men. In justice to Mr. S. P. Wood- ward, I do not hesitate to state that the soundest general work on the Mollusca that has hitherto appeared in England is his little Manual published by Weale. There is more nature, truth, and judicious conciseness in all the descriptive parts than one is accustomed to meet with in works of this kind ; and, although I am sometimes at issue with him, 1 trust that my opposition is never factious or merely for victory's sake. Mr. Woodward has availed himself of the characters aflForded by the lingual dentition more extensively, I believe, than any previous writer, if we except Professor Loven ; but much remains to be done before these can be successfully applied in the defining of the natural families, and still more before all the genera of Gasteropods can be referred to their proper places in the system. In the following Table, sho^ving the general nature of the lingual dentition and of the auditory concretions in the principal families, I have merely endeavoured to make a step in the right direction ; though I have little doubt that when the numerous errors which are always incident to attempts embracing so wide a scope shall have been cleared away, and the right exposition given of the value and bearing of assumed points of affinity, a system mil result inferior to none in the whole department of Zoology in the truthfulness of its natural arrangement. From more extended study of the sul)ject, since my former paper on the materials of classification was written, I have arrived at the conclusion that the nature of the contents of the auditory sacs is of minor importance to the primary characters of tlie lingual den- tition, as to whether it is elongated and strap-like, or broad and pavimental. Two cir- cumstances in particular suggest this view, viz. : — 1st. The close anatomical relationship existing between Cerithmm and Planaxis, and even the almost actual similarity of their lingual teeth, although the auditory sacs in the former contain otoconia and in the latter spherical otoliths. 2ndly. Several of the Eolidce (if I am not mistaken, I think I may venture to add the genus Flabellina) have otoconial particles, while the others in general have otoliths. It cannot be supposed, however, on this account that the few belong to a different family; for their whole anatomy refutes such a supposition. I may remark that amongst the Boridce also I have observed the occasional occurrence of minute otoliths instead- of otoconia. It is worthy of note that the otoliths, wherever they occur in such families as usually present otoconia, are exceedingly small, like the single primordial otoconial particle of the young Fteropod or Nudibraitch ; and in the converse examples, as in the Cerithiidce, where the presence of otoconia might a priori be quite unexpected, they are, as a rule, both few and large. VOL. XXIII. li 74 J. D. MACDONALD ON THE CLASSIFICATION OF THE GASTEROPODA. We now know enough of the distribution of otoliths and otoconia in the Molluscous Orders to perceive that there is a definite purpose in the bestowment of the one or the other, in particular cases ; and as this purpose is never the subject of caprice, but perma- nent, the resulting characters must be always significant, and therefore available in classi- fication. The division of the lingual dentition into straps and pavements, as previously defined, is not altogether free from imperfection, although it is most satisfactorily available in by far the greater number of genera. Indeed, it is only amongst the Ojnsthobrcmchiata that any difiiculty in this respect presents itself. Can anything, for example, be more diver- sified than the characters of the lingual dentition in the BnllidcB of authors ? No less than four distinct types are distinguishable in this family, thus : — Central series absent. Central series present. Laterals 1 or 2, Laterals more Laterals more Laterals 1 or 2, uncinate, » numerous, numerous, the third rudimentary, as in as in as in as in Scaphander Bullina (Risso)* Bulla (numerous species) Bulla ampulla and and and and BuLL.'EA. Aplustrum. Cylindrella. Amphisphyra. (Philinidse of Gray.) (Buhinidre and Amplustridas (Bullidse of Gray.) (Occurring respectively in of Gray.) the BuUidce and Phih- nidse of Gray.) The aggregate character of the dentition here tabularized is pavimental ; but in Bulla ampulla it is more strap-like, on account of the great breadth of the rachidian plates, and the remarkable difference between these and the laterals. It would be difficult also to form a judgment of the dentition of Bulloea taken by itself, the rachis being altogether absent, and the laterals much reduced in number. In the Nudibranchiata, moreover, the pavimental character prevails, althotxgh in the ElysiidcB and BoUcIcb the teeth are reduced to a single strap-like series ; and, considering the collateral relationships of both famiHes, I think that this view is more natural than to assume that their dentition is essentially constructed on the type of the strap. With this explanation, the consistency of the present system of determining aflinities is quite \inaffected by the retaining of ^olis and its allies in that position, which their general anatomy indicates to be the natural one. On comparing the median row of teeth in the broad pavement oi PhylUrrhoe with the single series of Glancus ov Bolls, a close observer cannot fail to be struck with their great similarity. According to my own experience, the formula of the dentition of PhylUrrhoe (3.0.3), as given by Mr. Woodward, is incorrect. I have, I trust, already sufficiently proved the impropriety of founding a classification on the physiology of respiration alone, to say nothing of tlie peculiar anatomical condi- tions respecting it, which we may presume are always present, though not sufiicieutiy known to be available. We know, from the analogy of the Crustacea in particular, that whether an animal breathe in water (fresh or salt) or in air, the general scheme of its structure exhibits no necessary restrictioii to this liabit. The principle here indicated * Cylichna (Lovcn) and Cyliyidrella (Swn.) are regarded as synonyms of Bullina (Risso), though the two former genera present a central series in the lingual pavement, while Dr. Gray gives the dental formula of Bullina as 6'(i. J. D. MACDONALD ON THE CLASSIFICATION OF THE GASTEROPODA. 76 is altogether ignored when all the air-breathing Gasteropods are associated in one so-called natural Order, Fulmonifera. Indeed the difference is so great in the two principal divi- sions of this Order, that the most superficial of all distinctions, namely, the presence or absence of an operculum, is sufiicient to distinguish them. In this way, bisexual animals with a pavimental dentition are associated with unisexual animals having a lingual rib- bon, and several more striking external differences. Indeed the characters common to both are such as apply to Gasteropods generally, affording no proof of their morphological agreement. This, therefore, must be my apology for separating the Fulmonifera opercu- lata in the subjoined Table from their supposed alliance with Lhnax and Helix, and placing them in another division, with animals having an organization in more complete harmony with their own. The Fulmonifera inoperculata together with the Opisthobran- cJiiata (M. Edwards) are retained in the position which Mr. Woodward gives them, as two natural series demanding no special change. I have, however, been obliged to place be- tween them the transitional genera Siphonaria and Amphibola, removing the former from the PatellidcB, which it resembles only in the shell, and the latter from the Apple-snails {Paludinidce), with which, even taking into account the obscure resemblance of its shell, it cannot have the remotest affinity. I refrain from the use of a family term to include those genera, because I have much reason to believe that Sipihonarla is a member of the Onchidiidcs, while Amphibola, though prosobrancliiate, exhibits an alliance with the Tectibranchs. However this may be, taking them in the order given, they render the passage from the Fulmonifera inoperculata to the Opisthobranchiate families easy and natural. The members of the second division in the Table are unisexual, and in this primary character differ from those of the first. They admit of arrangement into three sections, in one of which the dentition is pavimental ; in another altcjgether absent, while in the third it is strap-like. Prom actual observation and comparison of genera belonging to all the families cited, with one or two trifling exceptions, I am convinced that, first, the character, and, next, the number of the dental plates and processes, afford a truly natiiral test of the affinities of unisexual genera wdth strap-lUvc dentition. In studies of this kind I always bear in mind a grand principle, for which I am in- debted to Mr. W. S. Macleay, namely, that no character is natural until it has been proved to be so. No scheme, for example, however plausible from its delusive applicability to a certain number of cases, can be accepted as natural, when conditions of greater value are violated by its adoption. The employment of the mere number of longituchnal rows of teeth in the lingual ribbon as a means of classification may be compared to the method of Linnaeus, who based his Botanical System on the number and arrangement of the stamens, and would threaten at first sight to be equally artificial ; but I find that the legi- timate sway of characters of greater as well as of minor value is not interrupted or violated by its adoption. Hence it proves to be natiu-al, sustained also by the axiom, " natura non facit saltum." Errors with respect to sex may still lurk in the present Table, as I know to have been the case in a former one. I must state, however, that the sexual character of those instances which I have not been able to determine personally is supported by good l2 76 J. D. MACDONALD ON THE CLASSIFICATION OF THE GASTEROPODA. authority ; and even tliougli errors should here and there occur, it is much hetter, if possible, that nothing should be taken for granted merely because it might appear to harmonize with a pet theory. Malacologists have always been at a loss to ascertain the natural position of Janthina ; and the whole truth seems to be this — that it is not correctly referrible to any of the families already formed. Even though it should be placed by itself, as has been done by several authorities, its true relationship to other families in the series is still a matter of doubt. Dr. Gray interposes Janthmidce between the Apple-snails and the Heteropods, though the general disposition of authors is to place the solitary genus near Nerita or in some neighbouring family. Mr. Woodward has selected the HaUoiklce for it, probably conceiving that the notch at the outer angle of the shell represented that of Scissurella, Pleurotomaria, &c. Taking the anatomy of the animal into consideration, however, it is easy to show that this position is most unnatural. It has often suggested itself to me, that Janthina was in some particulars related to the Opisthobranchiate families ; and, although the aflinity is certainly not immediate, I apprehend that it will be difficult to find a better place for it than between these and the very natural family composed of the genera Sola- rium, Torinia, Scalaria, and perhaps some others. If the latter family, which may be named SolariidcB, and the Janthinid(S are really unisexual, they will form a remarkable and beautiful link or passage between the Opisthobranchiata and the unisexual Gasteropoda with a lingual ribbon. The following extracts from my notes will show that the family of the Solariidce is fouuded on no hypothesis, but on facts : — " In upwards of texi fathoms at the Feejee Islands we obtained a small species of Sola- rium, apparently quite new. The shell is orbicular, much depressed, with a wide umbi- licus, increasing with the growth of the whorls, and bounded by an angular and minutely l^eaded border, corresponding with the union of the lower and inner walls of the shell. This results from the revolution of a small lunated notch at the base of the columella, which is perfectly straight, directly joining the body-whorl. The outer lip is thin, and slightly angular at its central or more convex part. The mouth of the shell is traj)ezoidal, the angles being situate respectively at the base and summit of the pillar, at the middle of the outer lip, and at the union of this latter with the body-whorl. The most interesting part of this little shell is discovered on looking down to the bottom of the deep umbilicus, in which the sinistral nucleus, smooth, polished and completely inverted, may be distinctly seen. Operculum thin, horny, multispiral, slightly concave externally and minutely tuberculated." (The operculum of the typical Solaria is paucispiral.) " The animal, in general appearance, much resembles one of Cuvier's Tubulibranchiata. The creeping-disc of the foot is small, but the operculigerous lobe is large and subcy- lindrical. The tentacula are moderate, conical, with the eyes on the outer side of the base. " The proboscis is armed with two lateral oral plates, and the tongue with several series of tenaculiform teeth, of which the innermost are simple, the external trifid, and the in- termediate bifid. Finally, the acoustic capsules contain minute otoconia. " Solarium perspectivum. This species occurs plentifully on the sea-side of Kamba J. D. MACDONALD ON THE CLASSIFICATION OF THE GASTEROPODA. 77 point, Na Viti Levi, Feejee group ; and several specimens were selected for examination. The tentacula are stout, conical, and closely approximated above the mouth, with a longi- tudinal groove on their inner surface, and a somewhat sunken eye in a slight elevation or gibbosity at the outer side, near the liase. " The animal is furnished with a retractile proboscis of considerable length ; and Avhen retracted, the external aperture is quite round and so small as to lie scarcely perceptible. The anterior border of the foot is dilaminated, the plates or lips being separated l^y a deep transverse recess, which appears to have some communication with the sinus of the foot. The creeping-disk or mesopodium is well developed, and the lateral borders, which are produced into a simple lobe, are probably confluent with the epipodia. Its posterior ex- tremity is very thin, but rounded, and surmounted by the metapodium, which is a sub ■ cylindrical mass of muscular fibres, continuous Avith those of the great retractor, and abruptly truncated posteriorly, where it presents a subcentral recess, which lodges the internally produced nucleus of the paucispiral operculum. " The oral teeth form a narrow circular band consisting of a pavement of sharp dental cells, whose points, as in other cases, are directed forwards. " The lingual pavement is small, but elongated in form and divided into two lateral arese, supporting several series of long and gracefully curved uncinate teeth, which seem to decrease in length from within towards the lateral borders of the membrane, where they also become bifid in the vertical direction. " The auditory sacs are of comparatively large size, containing otoconia. " Torinia (Gray). The anatomical characters of this genus (distinguished by the peculiar gun-screw form of the operculum, so much resembling that of Siliquaria) agree in every essential particular with the foregoing. " Scalaria. Trom a critical examination of the anatomy of Scalaria, I have no hesi- tation in placing it in the same family with Solarium. Its principal characters are as follow : — Proboscis long, retractile, with stout muscular walls ; the oral aperture furnished with lateral plates, composed of small dental cells vni\\ their points directed forwards. Lino-ual membrane supporting a double pavement of tenaculiform teeth, rather stout, but still very similar to those of Solarium, and not, as has been supposed, lilie those of Bulla. Thus the outer teeth present one or two secondary prongs, while the inner remain simple. The eyes are placed at the outer side of the base of broadly conical tentacula ; and the auditory sacs contain vibrating otoconia. " The foot is dilaminated in front, and in general configuration resembles that of Sola- rium.'^ Erom the above facts I think it may be affirmed that it is a violation of the simplest anatomical principles to place Scalaria with the TnrritelliclcB, and Solarium with the Litorinidce, the genera in both of which families, as now received, are otherwise hetero- geneous enough. The vermetiform character of the animal of Torinia on the one hand, and the peculiar structure of the operculum of Siliquaria on the other (so closely aping that of the former genus), afford some indication that both may belong to the same family. Moreover the o-roove in the outer lip of Solarium would appear to represent the branchial slit of Silt- 78 J. D. MACDONALD ON THE CLASSIFICATION OF THE GASTEROPODA. cpiaria ; but until something more is known of tlie anatomy of the latter, I cannot hazard any further speculations on the subject of its natural position. It may now be asked, what is to be done with the FyrainidelliclcB, in which the lingual memln'ane is quite unarmed and consequently can afford us no guide in classification. The legitimate course in such a case as this Avould be to compare the whole tenor of the anatomy with that of other families, whose position has been better determined. With the hope of arriving at some successful result in this respect, I separately passed in review the anatomy of Pyramiclella, Odostomia, and Eulinia, all of which genera are rightly referred to the same family. But as tlie species of Fyrmnidella are usually more suited for examination, their study has given me the chief grounds for the conclusion I have formed, namely, that in their general anatomy they accord more closely with Solarium and its allies than with any other family that I can suggest. Prom these, however, they differ in two striking particulars, viz., the presence of minute otoliths in the ear-sacs instead of otoconia, and the absence of dental organs both labial and lingual. In these conditions, nevertheless, if my determination be correct, JPyramidella permanently repre- sents the early state of Solarium. It is therefore not improbable that, if by any change in the routine of nature dental organs were to make their appearance in Pyramidella, they would assume the pavimental arrangement. Ho^n^ever this may be, there can be little objection to the position chosen for the Pyramidellidce, immediately succeeding the SolariidcB, while there is much to sustain it. The fore part of the head (or the muzzle) of Gasteropods presents at least three different modifications which may be of service in classification. Thus, 1st, it may be simple, incapable of retraction beyond what is distinguished as mere contractility ; 2ndly, proboscidiform or retractile from the apex, invaginating itself mth one simple fold ; or, Srdly, it may be a true proboscis, retractile from or near the base, with two resulting folds. The latter form is present in all unisexual Gasteropods having a lingual ribbon with three rows of teeth and under, and in some few families in which this organ presents seven rows. The 2nd occurs in the Oyprceidce ; and all the rest, with one or two questionable exceptions, have simple muzzles with a dental armature of seven series and upwards. Pileo]^sis IIv/)igaricus is figured and described as possessing a retractile proboscis of considerable length; while the other members of Calyptrceidce have plain though more or less produced muzzles ; and it is very doubtful whether the animal in question properly belongs to this family, there being much in favour of its affinity to Vehitina, as suggested by Mr. Wood- ward. I must say that I have never discovered a true proboscis in the animals of the numerous Pileopsis-like shells which I have examined from time to time. The importance of the distinctions here indicated is also exhibited in the case of the genus Urato, which is admitted on all hands to belong to the Cyprceidce. Erato is anatomically related to Triton, or probably to Cassis, rather than to Cyprcea, as the inspection of its labial and lingual dental organs will at any time prove. Its lengthy proboscis is retractile from the base ; and the extremity of the tongue-sac can therefore never lie freely in the visceral cavity — one of the most essential characters of the Cyprceidce, though not peculiar to this family. The last case which I have to notice in this connection is that of Triforis. In a dextral J. D. MACDONALD ON THE CLASSIFICATION OF THE GASTEROPODA. 79 example of this genus *, whicli I collected myself on the Wreck Reef, I found that all my previous remarks on its anatomy were correct, save that the lingual membrane was stated to support a " multiserial pavement of minute teeth." Employing a oue-sixth-of-an- inch objective (Ross), I discovered my mistake with regard to this very small strap ; and I think I may now very safely say that it is septiserial ; and so far, indeed, Triforis agrees with CeritUum. But when we take its otoliths, retractile prolwscis, and lateral labial plates (composed of amber-tinted rhomboidal cells) into consideration, a place near Triton or Bcmella can scarcely be denied to it. I am still convinced that the early state of its shell affords only an apparent mark of aflB.nity to Cerithinm. On the other hand, three openings homologous with those of Triforis are strikingly apparent in some BcmellidcB. Having made one flagrant mistake, however, with regard to Triforis, I feel now rather disposed to leave the determination of its true status to more able malacologists, thro^v- ing out one hint — that its operculum is pointed, with an apical nucleus, as in Cerithiopsis (Eorbes), which also possesses a retractile proboscis. An error of observation is far more unpardonable than the false determination of the genus to which any particular shell may appear to belong ; for it often happens that in the short but indeiinite descriptions of the shell-characters given by authors, perspicuity is sacrificed in the attempt to attain conciseness. On this account I fell into another mistake, confounding a short and stumpy Fasciolaria, having very obscure plates on the pillar, and a crenulated outer lip, with Swainson's Tritonidea, which has now l^ecome synonymous with Fisania. The operculum in both genera is stout and claw-shaped, and the animals themselves are not at all unlike each other. The lingual dentition of Fasciolaria, Fusus, and Clavella, liowever, may be at once distinguished from that of Fisania, which I believe to belong properly to the Buccinidce and not to the Muricidce, if the Buccinmn cinctum of Quoy, and Fnsiosterna of Swainson may be taken as true examples. So far as I have been able to discover, besides Murex itself, Memifitsns (Sw.) is the only genus, in the whole of the Muricidce, having an operculum with an approach to an apical nucleus, which is essential to Fisania. The family-difference observable in the dentition of the Buccinidce, Fusidce, and Muri- cidce is so distinctive, that, when once recognized, no confusion of the species of one family with those of another can possibly occur ; and thus an unfailing guide is afforded in the determination of doubtful cases, which may be more or less conformable with the defini- tion of Fisania or any other ambiguous genus. I leave the case of the AciculidcB-f an open question untU I shall have the opportunity of comparing the anatomy of Geomelcmia and Acicula with that of Egea (Benson P), ^A'hich I think may possibly be identical with the former genus ; and if so, it wUl bring along- with it some freshwater and littoral genera, viz., Hydrohia, Syncera, and a number of beautifvil little Faludina-like shells now placed in the Bissoulce. Trimcatella is doubtless close at hand, though perhaps sufficiently distinct to form the * Having obtained a second dextral Triforis at Moreton Bay, it may be safely stated that tlie shell is not invariablv- sinistral, though commonly supposed to be so by conchologists. f As accepted by Mr. Woodward. 80 J. D. MACDONALD ON THE CLASSIFICATION OF THE GASTEROPODA. type of a separate family, with some more truly marine examples of which I am well acquainted. Many Triincatellce live far inland, exhibiting a terrestrial habit as obviously as any of the operculate Pulmonifera ; and their anatomical characters are well worthy of a rigid comparison with those of Acicula. The FateUklo} exhibit many points of affinity to Fissurella, Nerita, and the series to which they belong. Several members of the family manifest a tendency to deviate from the typical species, in their dental formula in particular ; and indeed the whole lingual armature in Fatella appears but to represent the rachidian area of Turbo or Nerita, — the pleural teeth being suppressed, as in the mooted cases of Elysia, Eolis, &c., noticed in the first part of the paper. The affinity here advocated is still further suggested by the peculiar sculpturing, olivaceous colouring, and nacreous lining of the shell, in some species of Acmma. The Fatellidce and DentcdiidcB have a broad upper-lip mandible, and their ear-sacs con- tain otoconia, but the resemblance ceases here ; while on the other hand in the Chitonidce no upper-lip mandible exists, and I have never been able to detect either visual or audi- tory organs. I must give up the attempt to class the two latter families ; but I think that the Fatellidce may be placed provisionally at least as an appendix to the Fissurellidce. I have no desire to repeat former observations, and therefore refrain from any further comment on the particular families contained in the Table. I can only say that I have endeavoiued as much as possible to make the order of their arrangement in strict accord- ance with their anatomical relations. With all its imperfections, therefore, the scheme is submitted to the judgment of the zoologist, purporting to be merely a help to the attainment of a precise conception of the natural affinities of Gasteropoda. fin 2 PQ o W O W <1 O "3 O .2 t: -gj" g I Ell 8 S £8 g- ^ H m ■^'S S cc .2 o ^ ^2 8 ■=■5-3 •2 .2 M ■ Co 5 ° 5 P^ o^S ■■as S. a 8 s 3 -g S J- s 3 to ° s o rS 8 8'Sa gS8j|jS J a.5s s^-s si c cj -S 8 -ii 1 :a S "^ 8 .;:i ; go's a. ^l§ ^ 8 5:S'C .3 ^ 'S "S " ^ o a ^ P3 H >..S °M s. oife r 5- -I VOL. XXIII. [ 83 ] . r VI, On Sycopsis. By Daniel Oliver, Esq., F.L.S. Read March 15, 1860. Sycopsis. Genus novum Samamelidearum. Char. ess. Flores abortu uni-se.xuales [an etiam nonnunquam polygami?]. Fl. S . Calyx brevissimus, irregulariter v. oblique dentatus lobatusve. Corolla 0. Stamina 8, calycis tube inserta ; filamenta brevia, in connectivum continua ; antherae basifixae, biloculares, oblongse, apice breviter productae. Gyncecium rudimentum bifidum. Fl. ? . Calyx semisuperus, tube ovarium arete cingente, limbo deciduo. Corolla 0. Ovarium bilo- culare; styli duo, distinct!, subulati, intus canalieulati, minute papilloso-stigmatosi ; ovula in loculis solitaria, pendula. S. Griffithiana, species imica. Frutex verisimiliter (v. arbor parva), ramosissimus. Hamuli cortice cinerascente, vix laevi, glabro v. ultimorum minute puberulo. Folia numerosa, alterna, minute stipulata, petiolata, Integra, ovali-lanceolata, lanceolata v. obovato-lanceolata, acute acuminata, coriacea (perennia?), discolora, glabra v. primilm pilis stellatis conspersa; costa (in spp. exsicc.) supra de- pressa, subter valde prominente, venularum reti inconspicuo. Stipulce lanceolatae, minutae, cito caducae. Inflorescentia axillaris, e glomerulis paucifloris vel racemis abbreviatis petiolum subae- quantibus composita. Flores c? et ? intermixti (an semper?), primum sessiles v. subsessiles, ante florescentiam in axillis bractearum nidulantes, omnes plus minusve pubescenti^ minuta stellata obsiti. Calyx, 3 brevissimus, triangulari-dentatus v. oblique lobatus, in ? ad faucem, stylorum basin arete cirigentem, squamis minutis pilosis instructus; limbi lobi denique curvi v. revoluti, cito decidui. Stamina nonnunquam partim imperfecta ; fertilia (cum sterilibus minoribusque non alter- nantia) filamentis brevibus, glabris, crassiusculis, calycem excedentibus, antheris (bilocularibus, dehiscentia ?) pro ratione magnis, ovato-oblongis, apice breviter et vix acute apiculatis ; stamina sterilia parva. Ovarium semi-inferum, stellato-pilosum, usque ad basin stylorum calycis tubo crasso arct^ cinctum. Styli a basi distincti, divergentes, infra stellatim pilosi. Folia lamina 1| ad 3i poll, (saepius circa 2^ poll.), petiolo 2-5 lin. longis. Turtlaer information is needed as to (1) whether the flowers are ever really hermaphro- dite, (2) if both [Luy-wan), Pun-tsaou-kang-muh, cap. xxxvii. sect. 4 (cum icone). ^C^LJDI [n~av)55Li) Carom-pallagum. The fourth tuberiform substance to which we have alluded is that named by Horaninow Mylitta Icqndescens. Specimens of this production, both from India and China, are drawn in PL IX. figs. 14-17. In a dry state they are extremely hard, so much so that, in the case of the Indian specimens, they might easily be supposed at first sight to be some stony substance from the chalk-beds out of which they are dug. There can be no doubt, how- ever, of their vegetable nature, and hut little, we think, as to their affinity with the hypogaeous fungi. Although no trace of fruit is discernible, the inner substance is marked with veins, and a microscopic section shows the division of the tissue into areolcB similar to that exhibited by the hypogseous fungi. It is not easy to say whether it is truly of parasitic growth, though the natives of India assert it to be so. It will be seen, by referring to PL IX. fig. 15, that the specimen there represented has a small root-like ex- crescence on the left-hand side ; and this excrescence, of which we have examined sections, is certainly of woody structure, although we are unable to say to what plant it belongs. Out of fifty Chinese specimens, however, only four exhibit these roots, which it is possible may be of accidental occurrence. The largest of these specimens weighs 106 grains, and the smallest 5 grains, the average weight of 50 being 24-6 grains. The largest Indian specimen weighs 268 grains. We are informed by Dr. Waring that the native Indian name, Carom-pallagum, means literally. Black pallagiim, pallagiim signifying a medicinal substance. Our specimens are all grey or brown rather than black ; but the wrinkles on the surface, which are very minute and which cannot be seen clearly without a lens, are of a very dark colour. It is possible that the specimens, when moist and fresh, may be of a darker colour than when in the hard and dry condition in which they reach this country. We think it hardly probable that the dry specimens in our possession should be brought to fructify, although the instance of Ergot of Hye producing fruit after having been baked in an oven shows that fungi may undergo a good deal of hard treatment without losing their vitality. We are experimenting with the Mylitta ; and if we succeed in causing it to produce fruit, we shall not fail to report the result. Trans, Linn, Soc. VolXXIII, Tab. 5 T 97 .<^^^ -•u^Sgj' (»'"» iC >-\ay' ■'-^»^^ TulfenWcot, sculp. Ks^ LmiM Soc. Vol XXIII. Tap. 10. E 97 ,^^^^^^^^^^SL=-. Lnwin , del . T'ilfcn'A'.-i: t; ■jculp . PACHYMA COCOS, AND MYLITTA LAPIDESCENS. 97 EXPLANATION OF PLATES. Plates IX. & X. Figs. 1, 2, 3. Specimens of Sclerotium stipitatum, Bk. & Curr., natural size. Fig. 4. Thin section of the inner substance of Sclerotium stipitatum highly magnified. Fig. 5. Specimen of Pachyma Cocos, Fr., in the Linnean Society's herbarium, natural size. Fig. 6. Longitudinal section of another specimen in the same herbarium. Figs. 7 & 8. Thin sections of portions of the latter specimen, highly magnified. Fig. 9. Portion of another specimen of P. Cocos in the Linnean Society's herbarium, showing cracks in the substance, the walls of which cracks are united by woolly fibres. Figs. 10-13. Specimens of Choo-ling, natural size. Figs. 14-17- Specimens o{ Mijlitta lapidescens, Horan., natural size. VOL. XXIIl. [ 99 J IX. Oil some New Zealand Verrucarige. By Charles Knight, Esq., F.L.S. Read Febiuaiy 16th, 1860. IhE following paper contains a description of certain corticolous Lichens of New Zealand, accompanied by figures of the spores. The size of the spores of the same Lichen is found to vary within certain limits. Those contained in the ascus are frequently much smaller than some of the loose spores found solitary on the field of the microscope. Spores of a brown colour when mature, are of a lighter colour, or even diaphanous, when young. The septa are not always visible until after the spores are perfectly mature ; frequently the number of the septa is fewer in the young spore. A. Sipor 2, .5 & 7 are copied from my Thesis. [ 125 J XII, Observations on the Neurat'ion of the Hind Wings of Hymen opterous Insects, and on the Hooks ichich join the Fore and Hind Wings together in flight. By Miss Staveley. Communicated by Dr. John Edttard Gray, F.B.S., F.L.S., V.F.Z.S., 8fc. Read June 21st, 18G0. Considerable attention has been paid to tlie distribution of the nerves on the fore wings of Hymenopterous insects, and the peculiarities which they offer have been very extensively used in the systematic arrangement of these insects, especially in the separation of nearly allied genera. After many inquiries of the naturalists best acquainted with the subject, I have failed to discover that any attention has been given to the distribution of the nerves in the lower wings. The observations which I have been able to make upon these nerves seem to show that they also offer considerable variation, and afford characters which may be valuable for the arrangement of the genera and for the distinction of species. The existence of a series of hooks on the upper edge of the hind wing, for the purpose of uniting the two wings in flight, has been long known, and the hooks were figured and described as found in the wing of the Bee, and some other insects of the order, in the works of the earlier micrographers ; but I cannot learn that any one has particularly studied them, or recorded the peculiarities which they present in different families, or in the genera and sjiecies of these groups. These hooks were generally believed to form a group near the centre of the wing, until the existence of some spines or hooks near the base of the wing, in certain species of Ichneumonidce, was pointed out in a note by the writer, inserted by Dr. Gray, of the British Museum, in the 'Annals of Natural History' for April 1860. The interest taken in those observations by Dr. Gray has induced me to examine further into the subject than appears to have been done hitherto ; and, for the purpose of enabling me to do so, he requested Mr. Erederick Smith, who has paid particular attention to the study of Hymenopterous in*sects, and who has arranged and described the very exten- sive collection of this order in the British Museum (perhaps the most extensive which exists), to furnish me with a series of wings of some of the species of the different families and genera, for examination. These wings were all named by Mr. Smith ; and, in the following observations, I have adopted the nomenclature and systematic arrangement used by him in the Museum col- lections and catalogues. The wings have been mounted in balsam, and now form part of the collection in the British Museum : they can therefore be consulted by any persons interested in the subject. The observations have all been made with great care from these mounted wings, which, although they exhibit much which would not be visible in the wings unprepared, are often 126 MISS STAVELEY ON THE NEURATION very deceptive ; each observation liad, therefore, to be repeated many times to ensure accuracy. The sketches accompanying this paper were (with the exception of two or three which were traced from the wings themselves) made Tvith the camera lucida, aided by the compound microscope or a simple lens ; and this may excuse any artistic defects that they exhibit. The hooks on the front margin of the hind wings of the Hymenoptera may be divided into two groups : those placed near the middle of the front edge, which may be called the Distal hooks, and which appear to occur in the wings of all Hymenopterous insects ; and those situated nearer to the base of the wing, which have been called the Sub-basal hooks. "^I'hen the latter are present, they are in general quite distinct from the distal series ; but in some few insects the two series seem gradually to pass into each other*. The disposition, form, and number of these hooks appear to afford, not only good characters for the separation of very nearly allied species, such as the species of the genus OpMon, but to afford also excellent characters for the distinction of genera and families ; and I have little doubt that when more attention sliall have been paid to them, their importance in the economy of the animals will be fully established. Mr. Frederick Smith, in the notice before referred to, expressed his belief that the hooks Avould be found to be more developed in those insects which have the quickest flight ; but further examination has not confirmed this opinion, as one or two of the insects marked by him as of very quick flight (as Astata Boops) have the hooks slightly developed and few in number. The neuration of the hind wings is somewhat similar to that of the fore Tidngs, and presents variations which fall naturally into three groups, according to the form of the front or costal nerA*e. I. Costal nerve divided near the base. A. Upper branch of the costal nerve marginal at least to the centre of the wing, where the upper and lower divisions are re-joined. Vespid^. "^ Vespa maculata Polistes (4 species), fig. 43 Polybia, fig. 44 . EUMENID^. Eumenes Synagris Costal nerve branching at the re-junction. Distal hooks commencing on the re-junction. * These hooks are only thus hriefly mentioned by Mr. Westwood in his ' Modern Classification of Insects' : — " Another character of the order (Hymenoptera) consists in the counexiou, during flight, of the two wiugs on each side of the body by means of a series of minute hooks, along the anterior margin of the posterior wings, which catch the hinder margin of the anterior wings, thus producing one continuous surface on each side." (Vol. ii. pp. 7Q, 77 ■) Dr. Gray has kindly sent me the following observations : — " I do not find that Mr. Westwood anywhere makes use of the hooks in the classification of the insects ; and in Curtis's ' British Entomology,' where the wings of the different genera are often figured in considerable detail, I do not find a single instance in which the hooks are either figured or noticed. I have inquired of Mr. Walker if he had ever.made use of the hooks on the wings of Hymenopterous insects to distinguish either the genera or the species of these insects, or if he could refer rae to any author who had done so ; and he said he had not himself, and, after searching in several works where he thought they might have been so used, he was not able to refer me to any." OF THE HIND WINGS OF HYMENOPTEROUS INSECTS. 127 EUMENID^. Rhynchium Odynerus, fig. 42 . ScOLIADiE. Scolia, fig. 29 ... . Sphegid^. Chlorion (2 species), fig. 33 Nyssonid^. Stizus, fig. 36 ... . Gorytes mystaceus, fig. 37 Gorytes Natalensis, fig. 38 Larrid^. Astata BoopsW, fig. 35 THYNNID.E. Thynnus, fig. 27 ... Costal nerve branching at the re-junction. .'Distal hooks commencing on the re-junction. f Costal nerve branching after the re-junction. 1 Distal hooks commencing after branching. Lower branch of costal nerve branching before the re-junction. Distal hooks commencing on the re-junction. r Costal nerve branching at the re-junction. J Distal hooks commencing after the re-junction. I Anal nervure enlarged at the base as if to branch, but not actually L branching. B. Upper brancliof the costal nerve marginal, bvit ceasing before the centre, or becoming very- thin. Lower branch rising to the margin (or nearly so) about the centre of the wing. ICHNEUMONIDiE. Ophion obscurus, fig. 17. Ophion combustus *. Ophion (N. America) *. Paniscus inquinatus t • Paniscus (England), fig. 19 Paniscus gJavxopterus . Pitnpla varicornis tt • • Piinpla Turionell(B^-\ . Piinpla (N. America) * . Pimpla (England)*, fig. 20 Tryphon^ End of upper branch of costal nerve very distinct. End of upper branch of costal nerve very indistinct. Membrane of wing appearing above the costal nerve where the distal hooks are situated, but very narrow. Ephialtes * Has several large and strong hairs irregularly placed near base under costal nerve in externo-medial cell, on membi-ane of wing. Has four similar hairs more regular and in a row. On membrane as above. Rhyssa *. Ospj'ywc/eo^Mstj fig- 22. (2 species) Membrane of wing appearing above distal hooks. There is a curious appendage (22 a) on costal nerve at base in both speci- mens ; it may be accidental, and the nerve certainly is torn, but I think it is worth looking at more specimens for. Wing bt-own. Tragus*, iig. 23 Membrane of wing appearing above the distal hooks; wm^ 6rojw«. This wing and Osprynchotus closely resemble each other in colour and veining, and in having two clear, colourless spots in the same situation. The chief difference is between t and *. Membrane of wing appearing above the distal hooks. Membrane of wing appearing above the distal hooks. Upper branch very distinct. Upper margin of wing turned over accidentally "i tt A short or faint branch from the base of the anal nervure. t ^ decided branch from the base of the anal nervure. * A slight indication of branch from the base of the anal nervure. Mesostenus t, fig. 16. . . Cryptus, (I) (Europe), fig. 14 128 MISS STAVELEY ON THE NEURATION ICHNEUMONIDiE. Cryptus (2), fig. 14 Membrane of wing appearing above the distal hooks. Upper Sphegid^. branch very indistinct. Sphex, fig. 34. POMPILID^. . . Pompilus (5 species), fig. 30. Mac7'omeris violaceus,^g.31. ' Pepsis (6 species), fig. 32. I do not know how to describe the liind wing of the ChrysididcB. The nerve divides (fig. 25) , and the division and re-junction are very indistinct, while there is merely an indication of a branch, and the only other nerve ends nowhere in particular. Perhaps some larger foreign species (if there are any) would be more decided. Ichneumon iD^E. Hemiteles The wing of my specimen is torn at the base ; and 1 think the nerve is displaced. I cannot describe it, or be sure if it should be in this division. II. Costal nerve not divided at the base ; single ; not marginal till about the centre, but margin of wing thickened above it. Apid.^. Costal nerve branching about the centre, where it becomes mar- ginal or nearly so. Distal hooks commencing on or before the branching. On costal nerve. Membrane of wing near base deeply cleft, and again nearer the centre. Xylocopa (2 species) . . Melecta armata .... Melee t a, sp Anthidiimi manicatum Anthidiwn, sp., fig. 46. . Panurgus Banksianus . Chrysanthedra smaragdina Lestis muscaria .... Andbenid^. Basypoda hirtipes, fig. 45 . FoRMICIDiE. Fo7'mica, fig. 26. CEcodoma DORYLIDiE. Dorylus CRABRONIDiE. T)-ypoxy Ion, ?ig. 41 . . , Mimesa ...... Pemphredon, fig. 40 . Nyssonid^e. Mellinus sabulosus, fig. 39 Tenthredinid^. Cejjhus pygmceus, fig. 2 MUTILLID^E. Mutilla, fig. 28 . . . . As above, but costal nerve not marginal. ^Distal hooks commencing beyond the branching. (In Mimesa and Penijihredon the anal vein is enlarged at the base as if to branch, but without actually branching.) As above,but costal nerve becoming marginal sooner than in the others. Costal nerve rising to the margin about the centre. Dipping below the margin at the commencement of the distal hooks (which grow on the membrane of the wing above it), and, when they cease, rising again and becoming marginal. Costal nerve branches at the point where it dips. OF THE HIND WINGS OF HYMENOPTEROUS INSECTS. 129 Chalcidid^. Chalcis, fig. 24 One vein only, simple ; ceasing at centre of wing end of vein. Distal hooks on III. Costal nerve not divided at the base ; marginal ; joined about the centre hy the nerve below. A. The two nerves continued after the junction as one, no longer marginal. Distal hooks commencing before the junction on costal nerve. XyPHIDRIIDyE. Xyphidria, fig. 12. . . . Tenthredinid.e. Perga, fig. 7 Pterygophorus (2 species), fig. 8. Urocerid^. Sir ex, fig. 13 Row of hooks single. ^ y Tip of wing as in Tenthredo, &c. below. J The costal nerve is divided at the base — the lower branch bent down and ceasing, whilst the upper is continued and joined hj the nerve below,and the two become one nerve nolonger marginal. Costal nerve and nerve below both reaching nearlj- to the margin at the tip of the wing without meeting there. N.B. All the other characters of Pterygophorus corresponding with this section, I have placed it here, notwithstanding the division in the costal nerve, this division being totally unlike that in sec- tion I., and, indeed, as far as I have observed, being peculiar to Pterygophorus. Row of hooks double. , Costal nerve and nerve below not joining at the tip. B, A loop formed at the junction of the nerves, by their separation and re-junction. The two nerves continued after the re-junction as one, no longer marginal. Distal hooks commencing before the junction on costal nerve. Row of hooks double. Tenthredinid^ Tip of wing: the costal nerve and nerve below join, and one is continued for a short distance ; never quite to the margin. Tenthredo (3 species), figs. 9, 10. Allantus (2 species), fig. 1. Ahia. Sciapteryx, fig. 1 1 . Hylotoma (7 species), figs. 3 to 6. In some of the Hylotomce the costal nerve is interrupted, rendering the loop exceedingly indistinct ; sometimes it is even lost, but there is always an indication of it. I now proceed to describe the two groups of hooks found on the wings. TJie Distal hooks are present on all the hind wings of Hymenopterous insects that I have examined. In the IcHNEUMONiD^ they are long and rc-curved (sometimes losing the re-curve towards the tip of the wing), and are situated on the lower division of the costal nerve VOL. XXIII. s 130 MISS STAVELEY ON THE NEURATION after that lias brauchecl and become marginal. They are generally largest towards the base of the wing. The number on various species is as follows : — Ophion combustiis, 7. Paniscus glaucopterus, 15. Ophion obscurus, 8 (fig. 1 7 Ophion (America), 6. Paniscus (England), 6. Paniscus inquinatus, 'J, Tryphon, 8. Pimpla TurionellcB, 7- Pinipla (England), 6. Pimpla varicornis, 10. Pimpla (N. America), fig. 21, has 2 strong upright hairs at the commencement of the row of hooks; 12 follow regularly; but between the 6th and 7th and the 9th and 10th, and in a Hne with the hairs, is a small hook standing up and not re-curved. In all these species of Pimpla, there are rather upright and strong hairs at the commencement of the row of hooks. Rhyssa, 13-14. Ephialtes, 6. Osprynchotus, 20, preceded by 2 upright hairs. Tragus, 2 1 . Mesoste?ms, 11. CryjJtus (England), 6. Crypius (Europe), 14. Hemiteles, 4. The distal hooks are of similar form, and begin before, on, or after the branching of the costal nerve in Sphegid^ {S2)hex) and Pompilid^e. Sphex {Qra.z\\) has .36, beginning on branching of costal nerve (7 ? sub- basal hooks). Pompilus bicolor 24, beginning on branching of costal nerve (no sub-bas.). Pompilus 44, beginning before branching of costal nerve (no sub-bas.). Pompilus nitiduliis 21, beginning after branching of costal nerve (no sub-bas.). Pompilus Br etonii 23, beginning on branching of costal nerve (7-9? sub-basal hooks). Pompilus 35, beginning on branching of costal nerve (2 sub-basal hooks ?). Macromeris violaceus .... 32, beginning after branching of costal nerve (14 sub-basal hooks). P epsis cmrulea {QvazW) . . . 24, beginning before branching of costal nerve (no sub-bas.). Pepsis speciosa 23, beginning on branching of costal nerve (no sub-bas.). Pepsis pretiosa 26, beginning on branching of costal nerve (1 sub-basal hook). Pepsis apicalis 19, beginning on branching of costal nerve (8 sub-basal hooks). Pepsis viridisetosa • . . . . 22, beginning on branching of costal nerve (5 ? sub-basal hooks). Pepsis limbata (fig. 32 a) . . . 30, beginning on branching of costal nerve (7? sub-basal hooks). They are of similar form, and begin where the upper and lower divisions of the costal nerve unite and again branch, in Vespid^, Eumenid-E, ScoLiADiE, and Sphegid.e {Chlorion). VeSPIDjE. Vespa maculata 33 Polisies 22 Polistes 21 Polistes 27 Polistes 27 Polybia 14, fig. 4 «. EUMENID^. Eumenes 26 i In ThynmiS they commence after the branching and junction, and are 15 in number. In Astata Boops, fig, 35 a, they commence after the branching and on the junction, and are 8 in number. Odynerus 23 Rhynchium 23 Synagris 23 ScOLIADiE. Scolia 17 Spiiegid^'. Chlorion (1) • ... 36 Chlorion {2) 31, fig. 33 a. OF THE HIND WINGS OF HYMENOPTEROUS INSECTS. 131 In Ntssonid^ both the veiniug and the situation of the hooks vary. Stizus (fig. 36 a) has 25 distal hooks, commencing on junction before branching. Gorytes mystaceus (fig. 37 a), 12, commencing on junction after branching. Gorytes Natalensis (fig. 38 a), 25, commencing on junction after branching. Mellinus sabulosus, 14, on costal nerve after branching (no junction). Thej^ commence on or immediately before the branching of the costal nerve in Andre- NiD^ and Apid.i;, and are large and recurved. They occasionally differ from the pre- ceding families in increasing in size after they commence at the centre of the wing. Apid^. Xylocopa 49 ? largest in the centre of the row ; very close together. Xylocopa 27, first two or three simple, the rest re-curved, diminishing in size towards the apex of the wing. Melecta 22, largest in the centre. Melecta armata .... 17, diminishing towards the apex. Anthidium, tig. 46 . . . 29, largest in the centre, broader and sharper towards the base of the wing. Anthidium manicatum . . 33 ? the wing too broken to be described. Panurgus Banksianus . . 11, diminishing towards apex. Chrysanthedra. Lestis muscana .... 20, diminishing towards apex. Andrenid^. Dasypoda hi?'tipes, fig. 45 a. 15, largest in the centre. They commence after the branching in Pormicid^, DoRYLiDiE, and Ceabronid^. FOBMlClDiE. Formica, 24 ? (Ecodoma, 25. DOBYLID.E. Doj'ylus, 22. Cbabbonid^. Pemphredon lugubris, 9. Mirnesa unicolor, 9. Trypoxylon figulus (fig. 41), 13, in two rows (of 7 & 6), growing in a line, with an interval be- tween them of the length of one row. In Mutilla, fig. 28 a, they grow on the membrane of the wing, above the costal nerve, which appears to dip down to make room for them. They are 17 in number, and commence before the branching. In the Tenthredinid^ and two allied families of Xyphidriid^ and Uroceridj^, they are in a single or a double row (one below the other), commencing on the costal nerve before the first junction with the nerve below, with occasional indications of a third series between the two principal rows. They are in a single row in XYPHIDBIIDyE. Xyphidria, which has 11. Tenthredinidje. ^ ' ■ j Radiating wrinkles above the base of each hook. Pterygophorus (fig. So), 10. J 82 132 MISS STAVELEY ON THE NEURATION And in a double or treble row in Tenthredinid^. Abia sericea, 16. Hylotoma ccerulea, 13. Hylotoma cyan, croc, 11. Hylotoma ustulata (fig. 5), 15 or 16. Hylotoma Ros<2, 12? Hylotoma (England), A. (fig. 3a), 15. Hylotoma femor alls, 12, 13? Hylotoma (England), B. (fig. 4), 15 ? Sciapteryx cost alls, 13. Urocerid^e. SireiV has 46, i. e.< 3 Tenth REDiNiD^. Tenthredo viridis (fig. 9 a), 15 {-^f}. Tenthredo viridis ? ?, 16, growing too irregu- larly to divide. Tenthredo Nothus, 13 (|). Allanius (N. America) (fig. 1 a), 16. Allantus sarophularius, 18 (xV). Cephus pyymceus has a single row of 6, beyond the branching of the costal nerve. Chalcid^, fig. 24 a. I have onlv one species of this family; it has three hooks on the termination of the only nerve. Chrysidid^. Chrysis ignita, 13 ?, 12 cJ ? Chrysis bidentata, 13. Chrysis (England), 11. Sub-basal hooks are found in the families of IcHNEXiMONiDiE, SPHEGiDiE, Chrysidid^e, PompilidtE, LARRiDiE, CrabronidyE, ANDREKiDiE, Nyssonid^, Tenthrebinid^, and THYNNiDiE. In some families, as Ichneumonid^, I have seen no species without. In others, as PoMPiLiDiE, they are sometimes present or absent in different species of the same genus ; and I have found no trace of them in the families ApiDiE, Vespid^e, Etjme- NiD^, ScoLiADxE, DoRYLiD^E, MuTiLLiD^, FoRMiciD^, and Chalcidid^. Of these, how- ever, I have (with the exception of Vespid^ and Apid^) seen very few, in some cases only one species. The sub-basal hooks vary much in situation, arrangement, form and number, being sometimes on the nerve and sometimes on the membrane of the wing. They occur at the base of the wing, or midway between the base and the distal hooks ; or a long row begins midway between the base and the distal hooks, ending only at the commencement of the distal row. Sometimes (as in some species of Hylotoma) there are one or two at the com- mencement of the row of distal hooks which can only be classed as sub-basal from their slightly differing in form from the distal hooks, while they agree with the sub-basal of other species which do not commence so near the distal. They are generally smaller and straighter than the distal hooks, 1)eiug sometimes quite fine and straight (though still, in most families, quite distinguishable from the hau's, which they then resemble), sometimes strong and straight almost to the end, where they curve suddenly — as in Paniscus. In the following species of the family IchneumoxidzE they are strong, curved at the end, and situated on the upper division of the costal nerve, near its termination. Ophion obscurus (fig. 17 6) has 2. | Ophion combustus (fig. 18) has 3. OF THE HIND WINGS OF HYMENOPTEROUS INSECTS. 133 Ophion (America) has 6. 5 grow together in a row, and 1 about the length of that row distant from them, towards the base. There is something like the scar of a missing hook in the interval. Paniscus hiquinatus has 1. Paniscus glaucopterus has 1 . Paniscus (English) (fig. 19 a) has 1, with a few very strong short hairs in a line on either side, differing from the fringe of hairs which cease when these begin. They seem intermediate in form between the hooks and hairs. Rhyssa has 1 (hook lost). Ephialtes has 1. Osprynchotus has 6. Tryphon has 1. Pimpla (N. America) has 1,2? Pinipla (England) has 1. Pimpla Turionellce has 1 \ T,. , . . , ^ } nearly straight. Pimpla varicorms has 1 J "^ ° In Tragus (fig. 23) there is one ; fine and quite straight, on the margin of the membrane of the wing, beyond tlie end of the upper division of the costal nerve, about midAvay between the base of tlie wing and the distal hooks. In Mesostenus there are 3-4 ? on the membrane as above, but not quite at the margin ; rather nearer to the base than to the distal hooks. In Crypttis (fig. 14) they are on the membrane, not marginal, midway between the base and the distal hooks, slightly curved. In another species there are 4, large and - strong, with a double curve, nearer the base than the centre of the wing. The margin of the wing is bent (accidentally ?) ; and I cannot tell whether they are on the nerve or the membrane. There are strong hairs on either side of the hooks in a line with them. In Ilemiteles there is one ; fine, straight, and erect, nearly on the margin of the membrane of the wing, above the costal nerve, rather nearer to the base than to the distal hooks. In the following species of the tribe Fossores they are large, strong, and curved at the tip, and situated on the upper branch of the costal nerve. In PoMPiLiD^, out of five exotic species of Fompihis, two have sub-basal hooks (2 ?, 7-9 r in number) ; but, owing to the darkness of the wings and the dense covering of haii', it is not easy to ascertain the number. Macromeris violaceus has 14. Pepsis (Brazil). Of six species, four have sub-basal hairs : pretiosa has 1 ; apicalis 8 ; viridisetosa 5 ? limbata 7? (fig. 32 ^i). In Sphegid^ they are of the same form as in Pompilid^e. Sphex (Brazil) has 7 or more, close together, on the margin of upper branch of costal nerve, near base. Clilorion (India, 1) has 17, near margin of wing, commencing further from the base than in Sphex, growing at irregular intervals. Ghlorion (India, 2) (fig. 33 b) has 22-23 ? near margin of wing, commencing further from the base than in Sphex, at irregular intervals. In Astata Boops (LAURiDiE) (fig. 35 a) they are 7 in niimber, growing on the costal nerve, at rather wide intervals, near the distal hooks. Very strong. Those nearest the base almost straight ; those towards the distal hooks sharply curved at the tip. A row of strong straight hairs grows in a line with them on either side, and between them. 134 MISS STAVELEY ON THE NEURATION Gorijtes Natalensis (fig. 38 h) (Nyssonh)^) lias 12, and G. mijstaceiis 8, strong sub-basal hooks on upper branch of costal nerve, nearer the base than the centre of wing. Mellinus sabulosus (Nyssonid.e) has 7-9 small sub-basal hooks on the membrane of the wing, above the costal nerve. Stiziis (Ntssokid^). I have only one specimen, much torn at the base, and an unnamed wing (? Stiztts) (fig. 36) with rather similar veining, which has the appearance of nine partitions in the cell, formed by the divided costal nerve. These may be sub- basal hooks lying down ; but if so, the manner of growth must be peculiar. They are perfectly straight. In the three following species of CKABUONiDiE (which are all that I have seen) they are on the membrane of the wing, above the costal nerve, placed at rather wide intervals, about midway between the base and the distal hooks in Mimesa imicolor, which has 6, and Pemplireclon lugiibrls, which has 4-6 ? Trypoxylon Figtdus has 2, nearer the distal hooks. Dasijpoda hirtijies (fig. 45 «) (Andrenid^) has 2 large and slightly curved hooks on the membrane of the wing, above the costal nerve and near the distal hooks. In Clmjsis ignita (fig. 25 a) and bidentata (Chbtsidid^), they are large and situated on the upper division of the costal nerve. Ignita, 8 ( ? ?) and 7 ( c? ?) ; Bidentata, 8 ? The Tentheedinid^ appear generally to be furnished with sub-basal hooks ; but it is not always easy in this family to distinguish them from the hairs, or from the distal hooks. I do not see them in Sylotoma rosce, U. ccernlea, H. cyaneo-crocea, JEL. "A.," or S. "B.," nor in Perga, unless one minute erect marginal hair close to the base of the wing on the otherwise hairless costa is to be called a " sub-basal hook," which I should be inclined to admit, from the fact that Sirex Juvencus has four or five similar hairs — the first nearly in the same situation, and the last quite at the commencement of the distal hooks — resembling the sub-basal hooks in many of the Tenthredinid^. In this Aving, however, there is a scar, not quite marginal, on the costal nerve, between the base and the distal hooks, which may have been the site of a more marked sub-basal hook. In Abia sericea there are about 7, as above, but more nearly approaching the character of sub-basal hooks. If in the above the sub-basal hooks are difficult to distinguish from the hairs, in some of the following it is as difficult to di-aw a line between them and the distal hooks. In some cases a distinction may be made by the form ; for, as the sub-basal hooks are, in nearly all families, straighter and smaller than the distal, and as the distal in nearly all families are largest and most decided in form towards the base * (diminishing in size, and in some * This fact does not always hold good with the Tenthredinidce, the distal hooks of which are sometimes smaller towards the base ; so that the distinction appears inapplicable in this family — the only one I have yet seen which requires it. I put it forward only as a quite arbitrary rule, which may be useful if it is necessary to divide the hooks into two series ; but I believe that in some species there is no real division. The hind wings of the Tenthredinidce seem to be, both in the veining and in the arrangement and form of the hooks, entirely irregular. The termination of the veins at the tip of the wing seems to me the most certain character of the group. OF THE HIND WINGS OF HYMENOPTEROUS INSECTS. 136 cases losing the re-curve as they approach the apex of the wing), I shoukl, in ill-defined cases, call any straight hook at the commencement of the distal a sub-basal hook. Thus in Fterygophorus I should reckon three (fig. 8), although there is nothing but the form of that nearest the distal, and the fact of the scars beginning nearer the base than is usual with distal hooks, to separate them. Allanius scrophularius has 5, on the costal nerve, at wide intervals, nearer the distal hooks than the base. Allantus (N. America) (fig. 1 a), 6 or 7. Tenthredo Nothus, 5 well-marked sub-basal hooks, beginning near the distal ; but the series is con- tinued almost to the base in the form of a deflexed hair, with a large scar on the lower edge of the costal nerve. The other hairs have not such " scars." Tenthredo viridis (fig. 9 a), 12. Hylotoma ustulata, 3. Tenthredo viridis} 2. Hylotoma femoralis, } Sciapteryx, 4. Cephus pygiruBus, 6, about midway between the base and the distal hooks. Xyphidria Dromedarius has 7 prominences (fig. 12 a), close to the base of the wing, on one of which is a large sub-basal hook. Probably 6 others have fallen off. Fore IViugs. — The folding over of the lower margin of the fore wing, for the reception of the hooks of the hind wing, varies much in different insects, in form, size, arrangement, and in texture, that part of the wing often differing greatly from the rest of the membrane, as in Pelecmus polycerato7% Avhere the principal roll is thick and dark, and thickly studded with short stiff spines or hairs, while the membrane of the wing is clear and thin, having large fine hairs at wide intervals. It is frequently serrated, or edged with short stiff hairs, along the upper part ; sometimes it takes a second fold, when the serration usually ceases (as in Astata Boo])s). In some insects, as in all the Tenthredinid^e which I have seen, in ScoUa, Mutilla, and Stilbum, there is but one roll ; and in some (as Mutilla) it becomes very narrow, and extends almost to the base of the wing. In Ves2)a the roll for the distal hooks is very broad, and doubled, and there is a second and narrower roll, nearer to the base, for the sub-basal hooks. In Chlorion the roll is broad at the centre of the wing, diminishes and widens again as it approaches the base, and there is a short broad turning up of the wing quite at the base. In Pimpla, Paniceus, TrypJion, Mesostenus, and Pelecimis there are two separate rolls. In Ophion two, or one prolonged roll. (Query. — Is it always a roll ?) XypJiidria Bromedarms has some hairs on the costa, near the base of the fore wing, much resembUng in appearance the sub-basal hooks in the same part of the hind wing. The only genus of four-winged insects, not Hymenopterous, on the hind wings of which I have observed a hook, is A^Ms '* (Tab. XVI. fig. 13 a). It is found in all the British species of that genus' described by Mr. Walker, of which there are specimens mounted by him in * Mr. Westwood having suggested to Dr. Gray that the hook of the Aphides was figured and described by Messrs. Ratzeburg and Curtis, he consulted their works, and has kindly furnished me with the following observations :— " Ratzeburg, in his ' Die Forst-Insecten,' figures four species of Aphides in their various states, but he does not 136 MISS STAVELEY ON THE NEURATION the Museum collection, tliougli I do not find it mentioned in the descriptions of that author. I may also observe that it is not represented in any of the 370 very much enlarged draw- ings of the species of this genus engraved in Koch's ' die Pflanzenlause, Aphiden, getreu nach dem Leben abgeleitet und beschrieben.' Nor are they mentioned in the work of MM. Amyot and Serville, ' Hist. Nat., Hemipteres,' Paris, 1843. The wing of Aphis is figured with a group of hooks, in a paper liy Mr. John Tyrrell, in the ' Jom-nal of Microscopical Science,' vol. iii. p. 230. This does not agree with any- thing that I have seen, nor does his third figure, which represents the fore and hind wings as joined casta to costa. In all the species of Aphides that I have examined, there is only a single hook placed near the middle of the front edge of the hinder wing ; and the hinder edge of the front wing is turned up for a considerable distance to receive it. In general the hook is simple, subcylindricsll, tapering to a point, with an enlarged base ; but in one or two species it looks as if it were compressed or flattened, and is slightly striated longitudinally. It is not to be observed in the wings of JLivia, Psyllce, or any of the neighbouring genera Homoptera which have come under my observation. In most of the other genera of this order of insects, the fore wings are more coriaceous or horny than in Aphis, and are to be regarded as much in the light of covers to the other wings as organs of flight. In some Lepidopterous insects there is an appendage thus described by Mr.Westwood :— " Another peculiarity consists in the apparatus by which the two wings on the same side are retained together in the same line during flight, composed of a loop in -n^hich a strong bristle plays (fig. 102. 8), The loop is formed either by an elevated portion of the mem- brane of the strong central vein of the upper wing on its under surface, or by a tuft of raised hairs. This appears to have been first noticed by De Geer (torn. i. tab. 10. f 4) and Harris (in his Essay upon the membrane of the wings), and afterwards by Giorna (Lin. Trans., vol. i. p. 135) in greater detail. M. Poey (Ann. Soc. Ent. Prance, tom. i.) has observed that the bristle is simple in the males but multiplied in the other sex, in which, according to Kirby and Spence, there is no annulus, and consequently these individuals are less fitted for flight. See also Hoeven in Perussac's Bui. Sci. Nat., March, 1828. It is chiefly among the Spingidce and Moths that we find this apparatus, the butterflies being destitute of it." (' Modern Class, of Insects,' vol. ii. p. 317.) And he adds as a foot-note at page 332 : " Prom the observation of Mr. Haworth (Lep. Brit. p. 19) upon the strong flight of the males of the Purple Emperor {Apatura Iris), it Avould appear that the males of that species possess this socket and spring, as he terms this apparatus, and of which the females, which fly but little, are destitute." represent the hook on the wing in any, although his figures of A. platanoides and A. piniphila are of a large size. He has observed and figured the hooks on the hind wing of the genus Ghernies in two species, viz., Chermes coccineus (t. 12. f. 1. o. 9), Chermes laricis (t. 13. f. 5). " Mr. Curtis, in vol. xi. of the ' British Entomology,' divides the Aphides into two genera, and figures as the types Cynara roboris (p. 576), Aphis tiliee (p. 577). He does not describe the hooks in his generic characters, nor represent them in his figures : this is the more remarkable, as he figures the hind and fore wings of Cynara roboris in the rela- tive position, of an enlarged size, to show the nerves." OF THE HIND WINGS OF HYMENOPTEROUS INSECTS. 137 I have examined this apparatus in different genera of Lepidoptera. The bristle is usually not only double or divided and forked from the base, instead of single, but also is much weaker in the females than iu the males. It has no affinity in its structure with the hooks on the hind Avings of Hymenoptera and Aphides here described ; and from its situation at the base of the wing, its flexible and elastic character, and from the imperfection of the loop, it can have little effect in keeping the wings "in the same line during Hight," or at least an effect very diGferent from that produced by the hooks on the margin of the wings of Hymenoptera and Aphides, which, wherever they may commence, always reach to near the middle of the wing, and are thus able to act with considerable power in connecting the two wings and keeping them in place. I cannot believe that the use of this apparatus in the Lepidoptera has yet been disco- vered, or that it is of the same nature with that of the hooks here described. Perhaps the name of " spriag," which Mr. Haworth gives to it, may best illustrate its use. Fig. 2. Fig. 3. Fig- 4. Fig. 5. Fig. 6. Fig. 7. Fig. 8. Fig. 9. Fig. 10. Fig. 11. EXPLANATION OP THE PLATES. Tab. 16. Hind Wings of Hymenoptera Terebrantia, Latr., and Aphis. 1. Allantus (N. America), a. the 16 distal hooks and the first sub-basal. 2. Cephus pygmxus. Hylotoma, A. a. the distal hooks, , B. The distal hooks. ustulata. Distal hooks, and, i and 2. ? sub-basal hooks. femoralis. a. tip of wing. Pe7-ffa. Pterygophorus. a. base of wing : b. i, lo. distal hooks ; i, s. sub-basal hooks. Tenthredo viridis. a. the distal and three first sub-basal hooks ; i, 2, 3. end of nerve. Tenthredo, sp. n. ? i, 2, 3. end of nerves. Sciapteryx costalis. a. end of costal nerve. Fig. 12. Xyphidria Dromedarius. a. sub-basal hooks. Fig. 13. Sir ex Juvencus. Fig. 13 a. Wing of Aphis hieracii. Fig. 14. Cryptus, Europe. Fig. 15. Cri/^^M«, England. Scar of sub-basal hook. Fig. 16. Mesostenus. Fig. 17. Ophion obscurus. a. the distal hooks ; b. sub-basal hooks. Fig. 18. combustus. sub-basal hooks. Fig. 19. Paniscus (England), a. sub-basal hooks. Fig. 20. Pimpla, England. Fig. 21. , N. America, distal hooks. Fig. 22. Osprynchotus. a. base of costal nerve. Fig. 23. Tragus. Fig. 24. Chalcis. a. distal hooks. Fig. 25. Chrysis. a. the sub-basal hooks. VOL. XXIII. T 138 miss staveley on the neuration of hind wings of hymenoptera. Tab. 17. Hind Wings of Hymenoptera Aculeata, Latr. Fig. 26. Formica. Fig. 27. Thynnus. Fig. 28. Mutilla. a. distal hooks. Fig. 29. Scolia. Fig. 30. Pompilus. Fig. 31. Macromeris violaceus. Fig. 32. Pepsis limbata. a. distal hooks; 6. sub-basal hooks : highly magnified. Fig. 33. Chlorion (No. 2). a. the distal hooks ; b. the sub-basal hooks. Fig. 34. Sphex (Brazil). Fig. 35. Astata Boops. a. some of the distal and sub-basal hooks. Fig. 36. Stizus. a. some of the distal hooks. Fig. 37. Gorytes mystaceus a. some of the distal hooks. Y\g. 38. Natalensis. a. some of the distal hooks ; b. a sub-basal hook. Fig. 39. Mellinus sabulosus. Fig. 40. Pemphredon luyubris. Fig. 41. Try poxy Ion Fiyulus. The distal hooks. Fig. 42. Odynerus. Fig. 43. Polistes. Fig. 44. Polybia. a. distal hooks. ^-~" " Fig. 45. Dasypoda hirtipes. a. the distal and the sub-basal hooks. - Fig. 46. Anthidium. a. the distal hooks. Fig. 47. ? Irons Lmn.. Soc Vol.XXUI.Uhlb^ 7/vz.«.v Zuni Soc Vol. XXIE. tah T/ FrcS't Smidi- Scid:>t [ 139 ] XIII. On certain Sensory Organs in Insects, hitherto imclescribed* . By J. Braxton Hicks, M.D. Lond., F.L.S. Read May 3rd, 1860*. BErORE proceeding to the details of the following communication, it will be weU to remind this Society of the structures I have already pointed out as existing in the Insect tribes, the consideration of which wiU serve to explain the nature of the new organs about to be described. In a paper read before the Linnean Society on the 17th June, 1856 1, 1 called attention to groups of hemispherical vesicles, disposed in rows, very regularly arranged, situated at the base of the halteres in all Diptera. These vesicles are apparently formed of cuticle, beneath which the ordinary chitine-layer is deficient, whereby a longer or shorter tube is formed, according to the thickness of that layer at the part. I showed that the number of these vesicles in some Diptera was great — sometimes as many as 360 on each halter. I also pointed out that a very considerable nerve (the largest in the body except the optic) pro- ceeded to the halter. I then, as also subsequently (Linn. Trans, vol. xxii. p. 144), showed that similar structures were to be found extending along the subcostal nervure of the wings both of Diptera and of the 4-winged tribes ; more numerously in that of the posterior wings and on the upper aspect. To these organs also nerves were clearly seen to pass from the thoracic ganglia. At the same time I showed that the elytra of the Coleoptera formed no exception in conseqiience of their peculiar structure, but that the nerve, entering their base, branched into numerous filaments, which ultimately terminated in vesicles pretty uniformly scattered on the upper surface. At the time of reading that paper I had not found these organs on the under surface of the elytron, but I now furnish a drawing of the under surface of that of Aroinia nios- chata (Musk Beetle), in which the vesicles are distributed in the course of the nervure, the nerve giving off branches to each vesicle (PL XVIII. fig. B). Although I had distinctly perceived the branches of the nerves passing to the vesicles on the subcostal nervure of the wings and on the elytra, I had not, at the time of my first notice of the subject, seen the nerves actually distributed to those more elaborate groups on the base of the halteres. I now, however, am able to show this point most clearly in the halteres of (Estrus Uqui (PL XVIII. fig. A). It will be seen that a branch passes directly to each of the three principal groups at the base, while a small nerve only proceeds up to the head of the organ. Thus it is proved that, whatever the function of these organs may be, the largest nerve, except the optic, is distributed to them. Having pursued these investigations into the different members of insects, I beg now * It should be stated, that this paper was originally presented to the Royal Society, where it was read on the 26th May, 1859 ; and an abstract of it was published in the Proceedings Roy. Soc, vol. x. p. 25. t Journ. of Proceed. Linn. Soc, vol. i. p. 136. t2 140 DR. J. BRAXTON HICKS ON UNDESCRIBED SENSORY ORGANS IN INSECTS. to draw attention to another somewliat similar group of organs observed about the joint of the trochanter and femur, chiefly on the former, and which I have found in all the insects I have examined. The organs in question consist, as in the structures above described, of a thin delicate membrane, probably the cuticle, stretching over, and closing in from the air, a tubular aperture in the chitine-layer of the part. This apertmx may be circular or oval, the tube varying in length according to the thickness of the integument, curved as in the Hornet, or forming a globular cavity as in Silpha. The delicate membrane which covers over this aperture is generally level, sometimes having a ridge or a minute papilla in its centre, as in Meloe (PI. XVIII. fig. G). In some rare instances, as in Formica rufa (PI. XVIII. fig. M), I have found that a few occur about the next joint (the tarso-femoral). The following are those which I have selected to exemplify these structures : — They are remarkably well shown in the genus Rippobosca. Fig. C, 1, represents the trochanter-femoral joint of S. Mmmdinis. The organs are found on both sides of the limb, and occur in groups of one or two rows of vesicles, each of which has a diameter of 3^5-0 inch, the rows being either straight or curved. Fig. 0, 2, shows those of S. Ovis ; they are rather more numerous than in H. Hirmi- dinis, and of the same diameter. Those of H. equina are similar. Those of Mtf^sca domestica are represented at fig. D. In all the Diptera I have exa- mined they are aUke; they are concentrated into a group of about ten to twelve on the femoral portion of the joint. The diameter of each is 3^00 iiich. In Pulex these organs can be seen, of a smaU size, about three in number, on the concave side of the trochanter. They are best shown in the SwaUow-Plea. In Silpha (PI. XVIII. fig. E) they are weU marked, being about eight on one side of the trochanter and six on the other, the centre of the tube being of much greater diameter than either the covering-in membrane or the internal opening, thus being of a globose form. The diameter of the covering-in membrane of the largest is yriVo inch. In the female of Lampijris nocUluca (Glow-worm) (PL XVIII. fig. F) they are found in a row of six to eight, extending transversely across the trochanter on each side. The branches of the nerve proceeding to them can be i)lainly seen (fig. F, 1 c). In Sleloe I have found these organs on the femur, close to the trochanter; they are about twenty in number; the covering-in membrane has a central prominence (PI. XVIII. fig. G, 2 ^», 3 6), on which the branch of the nerve (fig. G, 1 «, 3 a) ends, after passing up the tube in the integument. The diameter of this covering-in membrane is about ^tVo inch ; that of the papilla is —5^00 i^^ch, whilst that of the nerve is still smaller. In Forficula auricidaris (PI. XVIII. fig. H) they are found on each side of the tro- chanter, near the joint, on one side rather scattered and large, and on the other in a group of about six or seven. They are oval in form, and rather prolonged at one end. The diameter of the longest is jtoo inch. These structures are found in the Grasshoppers {Tetrix) on both trochanter and femur, as shown in PI. XVIII. fig. I, 3 ; they are oval, and in the common Grasshopper mea- DR. J. BRAXTON HICKS ON UNDESCRIBED SENSORY ORGANS IN INSECTS. 141 sure about i-roo inch. They are found on the same parts in Acheta and Blatta, of the form of those of the Scorpion-fly, as below. In the Scorpion-fly {JPanorpa communis) they are found in a group of ten to twelve on each side of the trochanter (PI. XVIII. fig. I, 1) each one is oval ; and has on one side a ridge running in the direction of the longest diameter (fig. I, 2 a). They are also found in Libellula in the same parts. In the genus Gerris (Hemiptera) the organs are foiind in two groups on the trochanter, about four in each group, and their shape is diflTerent from those hitherto described, being oval, with a point at each end and a papilla in the centre (PI. XVIII. fig. K, a). The sam.e organs are found in the Water-Scorpion {Nepa cinerea). In the genus Vespa they are found in a group on both the trochanter and femur, near the joint, and a few scattered over the trochanter. I have represented those of Vespa Crabro at PL XVIII. fig. L, 1 ; they are circular at the surface. The tubes leading to the sui'face are curved ; the nerve proceeding to them can be plainly seen, as is shown at fig. L, 2. In the Hive-Bee these organs are not so numerous, but are situated in the same part. \\i Formica rufa, or the Wood- Ant (PI. XVIII. fig. M, 1), the organs are very numerous, and exist both on the trochanter and femur (fig. M, 1, a a). The nerve can be seen passing to the different groups. On the femoral end also of the tibia are to be found a few small groups, to which the nerve can be traced (fig. M, 2). Their diameter is about g^ro iiich. In the various Tenthredines the same organs are to be seen (PI. XVIII. fig. N, 1), larger than the foregoing, and their form, which is shown at fig. N, 2, is somewhat conical. They exist in two groups — one small, on the coxa ; and the other larger, on the trochanter. The organs on Grylla viridissima are represented at PI. XVIII. fig. O, 1. They exist on the first and second legs, but not on the third leg ; they are in three groups, two on the trochanter and one on the femur. At this place I may mention that I have examined the dilatation on the fore leg of Grylliis and Locusta, considered by Siebold* as the " organ of hearing," and I have arrived at the same conclusion as M.Lespes in the 'Ann. des Sci. Sat.' 1858, namely, that there is no special nerve proceeding to the dilatation, but that within it there is a swelling of the trachese, which principally occupies the space, and that the nerve passes deeper down behind this tracheal swelling, being continued through the joint to the extremity of the leg. I have bleached the integuments, and have been able to see quite through the leg by adjusting the focus suitably. The next portions of the Insect to which I wish to direct attention are the palpi, both maxillary and labial. If we examine, with \ to |-inch objective, the palpus of the common Blow- fly {Musca vomitoria), it will be seen to be studded with a number of small transparent dots over more than half its surface, more thickly towards the apex (PI. XVIII. fig. P, 1). Their true nature it is difiicult to teU, but it is probably some- thing similar to that which I shall describe in Vespa Crabro. I have di-awn them enlarged (fig. P, 2), with the minute hair-like cuticular projections between them. Their diameter in this Ply is toVo inch. * Wiegm. Archiv, 1844, vol. i. p. 52. 142 DR. J. BRAXTON HICKS ON UNDESCRIBED SENSORY ORGANS IN INSECTS. In the palpi of all Diptera formed on the same plan as in Mnsca, I have found these structures existing, varying indeed in number and size. I have also found them in Eristalis tenax. In many of the Diptera (Mtisca, Mesemhrinimi, for instance, and Flies of similar con- struction), a group of organs of a somewhat similar character is to be found on the tips of the piercers ; in some there are as many as twenty on each side of the middle line, two large ones generally being situated at the apex. In Tapanus bovinus the tip of the central portion or piercers possesses as many as twenty, apparently having the same structvire as those in the Hymenoptera, to be immediately described. I have found similar organs on or about the palpi of some of the Hymenoptera, as Vespa, Odynerus, Apis, Bombus, Megachile, Nomada, &c. I have, in PI. XVIII. fig. Q, 1, figured these in Vespa Crabro. They are situated at the basal joint of the labial palpi, about twelve in number, and in the same joint of the maxillary palpi, but fewer and more scattered. In Vespa vulgaris theu' diameter is rrro iiich. In fig. Q, 2, I have given an enlarged view of one, — (2) showing a section in which the nerve («) is seen expanding on the inner surface of the covering-in membrane, (3) as seen from above. In Odynerus muraria and O. Antilope, for instance, they are foimd also on the centre of the base of the labium. In Ajns mellijica these organs, instead of being placed on the fii'st joint of the labial and maxillary palpi, are close to the origin of the palpus. I have shown these at PL XVIII. fig. R, 1 a, 2 a. If we examine the apex of the labial palj)i of this Bee, we shall find three or four structures somewhat similar, but having a few papillae at the surface. They are situated among the hairs, and are represented at PL XVIII. fig. R, 3 a. In the Megachile (Leaf-cutting Bee) they are very similarly situated, but on the base of the tongue, and on the basal joint of its palpus, and a few scattered over the second joint (PL XIX. fig. A, 1). In Nomada succincta I can find them only on the base of the labial palpi, and one on the distal end of each segment of the same palpi, which latter are of much larger size (PL XIX. fig. B, a). This will suffice to point out the kind of structure to be found on or about the palpi ; to aU of them a bundle of nerves can be seen proceeding. In Byticiis marginalis I have observed a few organs on the maxillary palpi, situated in an oblique row (PL XIX. fig. E, 1 « & 3), of apparently similar nature to the structures which I have observed on the antennae of that insect (in the second paper read before the Linnean Society, May 1859). They are shown enlarged at fig. E, 3. A group of similar structures is to be found in some of the Arachnida, as Arania, Tegenaria, on the under surface of the maxillEe, extending lengthwise along the middle on a space nearly devoid of hairs. They are the only organs which I have been able to meet with in that tribe ; but as the habits of the Arachnida show them to be chiefly dependent on sight for their guidance in obtaining supplies of food, we probably might not expect to find a large development of these organs, or others I am about to mention. In the palpi of Lepidoptera I have observed a peculiar structure, which I cannot find noticed by any one, and which differs much from the other I have described. The structure in question consists of a depression on the apex of the terminal joint. DR. J. BRAXTON HICKS ON UNDESCRIBED SENSORY ORGANS IN INSECTS. 143 the depth of which varies extremely — from a mere depression, as in the Lacquey-moth (Clisiocampa Neustria), to a tube extending inwards to half, or more, of the length of the joint. Sometimes the tube is of the same diameter throughout, or even dilated at the innermost end. What the nature of the inner surface of this tubular depression is, it is very difB.cult to ascertain from its position. It certainly possesses numerous hairs, and apparently there are some circular markings between them at the deepest part. How^ ever, upon one point there is no doubt, viz. that a nerve proceeds distinctly to it alone, and spreads out upon the apex of tliis cavity (PI. XIX. fig. C, 1 5, 2 6). It is held in its position by two or three bands which are attached to the interior of the wall of the palpus. The aperture of this structure when tubular (which is its most frequent form) is shown at fig. C, 1 a, 2 a. I have given figures of it in Argynnis Paj)hia and in a Geometer- moth. In a large number of different species examined, I have never found this struc- ture absent. In Aclierontia Atropos (Death's-head Moth) there is no depression, the part being merely marked by a circular yeUow spot. I have found something similar to this in the first joint of the palpus in Bibio Hard (Diptera), which I have drawn at PI. XIX. fig. D, 1. There are about twenty- four irregularly-shaped hollow cells or compartments on the floor of the cavity, each of which is divided from its neighbom-s by a rather stout septum; and a number of hairs spring up from the floor of each cell, and from the sides of the walls dividing them (See PL XIX. fig. D, 2). I have not as yet met with this condition in any other dipterous insect. I shall now pass to a class of structures quite dissimilar to those already mentioned. They are to be found on the palpi and those parts about the mouth whose function is that of touch, as the apex of the palpi, proboscis of Lepidoptera and of Diptera, and also on the apex of the antennae, being peculiar modifications of true hairs, and not processes of the cuticle. It is requisite to bear in mind the true nature of the hair in Insects, — namely, that it is situated in a depression caused by the absence of the inner layers of the integument, into which the cuticle is continued. In the centre of this cuticular depression a small elevation or papilla arises, which is the true root of the hair, and which rises from it of various length. By means of this arrangement, the hair itself is capable of some degree of motion. The interior of the root of each hair is in connexion with the internal parts of the member on which it is situated ; commonly fibres run to it, probably always including a branch of a nerve ; and this is decidedly the case in those hairs sitiiated near the prominent parts and extremities of the various members, as, for instance, the tips of the antennae, the palpi, pads of tarsi, &c. ; and this branch of the nerve does not run into the interior of the hair, but only to the inner aspect of the root, which separates it entirely from the interior. The difference between this structure (true hairs) and cuticular pro- cesses must be particularly borne in mind : the latter, having no root, and not being situated in a depression, evidently only spring from the surface. The spine must also be distinctly separated from the true hau', being a tapering process of the whole integument, into the interior of which the contents of the body can freely pass. The palpi have been considered by most naturalists as the organs of touch, both from 144 DR. J. BRAXTON HICKS ON UNDESCRIBED SENSORY ORGANS IN INSECTS. observations upon the habits of insects, and also from the palpi haAdng, in those instances where they are particularly used in the investigation of food, &c., an apparently thin, delicate, bladder-like membrane at their extremities. Noay, although I shall be able to show that they are endowed with a very delicate sense of touch, yet that function depends on a condition very different from that simply of a delicate membrane ; for, in the numerous palpi which I have examined, the thin bladder-lite membrane, instead of being perfectly smooth, is in most, if not in all instances, furnished with numerous bodies, which I shall be able to show are but modified hairs, sometimes of extreme minuteness and deKcacy. But even the palpi of the Orthoptera ( Gryllus, Locusta, Tetrix, for instance) bristle aU over with hairs, as will be shown below. Proceeding to each of these modified hairs, in every instance, can be plainly observed a branch of the palpal nerve ; indeed, in some Beetles, which have the extremities of the palpi dilated, the palpal nerve seems to undergo an extraordinary development, which perhaps may be only from the multitudinous sub- division to which it is subjected in order to supply each hair, or from the formation of a ganglionic enlargement of the nerve itself. This latter supposition does not seem to me improbable, as I think I have seen such a condition on the antennal nerve. The subject, however, is one of difficulty. The palpus of Timarchus (PI. XIX. fig. H) is well-fitted for examination. The palpi of many Insects are covered with large stiff hairs, such as the palpi of most of the Diptera, Hymenoptera, &c., which evidently can possess only a general sense of feeling ; but in those Insects which are remarkable for the use they make of their palpi, as the Coleoptera, Orthoptera, &c., the integument becomes more delicate towards the apex, and the large scanty hairs found on the shaft become much smaller, and altered in general appearance near the apex, while on the delicate tip itself they are further modified, and are sometimes exceedingly minute and very numerous. Sometimes, however, the shatt of the palpus is almost destitute of hairs, while those on the apex are largely developed, as in the larva of Ilelolontha. The antennae themselves are frequently used as supplemental organs of touch, and in those Insects which use them largely for that purpose, a certain modification of the hairs takes place, but not so completely as in the palpus ; also the antenna-wall undergoes considerable reduction in thickness gradually as it approaches the apex, which is well observed in the antennae of Myrmica, Formica, Vespa, Apis mellifica, and Sirex gigas ; but in no instance have I found the change of hair so weU marked, and so easily observed, as in Dyticus marginalis both in the antenna and palpus ; and by watching the habits of this insect we see that it uses the antennae in the same manner as Hydropliilits p)ioeus does its palpi, which organs they much re- semble. The form and position of these modified hairs, to which I propose to give the name " tactile hairs," will be best understood by the investigation of the different drawings here given ; and fii'st I wiU call attention to the palpi of Dyticus marginalis (PL XIX. fig. E). On the prominent portions near the joints the largest tactile hairs are to be found, as shown in PI. XIX. fig. E, 4, which may be described as foUows : a tube passes through the wall, becoming narrower in the centre, across which stretches the cuticle, which dips down the tube as far as this point and forms a conical septum ; in the centre of this DR. J. BRAXTON HICKS ON UNDESCRIBED SENSORY ORGANS IN INSECTS. 145 septum is a point or papilla, wliich is the root of the hair, which springs from it, not long and tapering, but oval like a ninepin, its apex scarcely reaching beyond the general level of the exterior. To the inner surface of the root a nerve passes. If we look with a higher power {\ or ^ inch objective) at the delicate membrane at the extremity of the apex of this palpus, it will seem to be studded all over with minute bodies, as shown in PL XIX. fig. 'E, 1 b b. These upon careful examination will be seen to be but modified hair-follicles, the space between each being depressed (fig. E, 2). The prin- cipal part of the palpal nerve proceeds to these, sending a branch to each. Compare this palpus with that of another Water-Beetle, SydrophUus jnceus (PI. XIX. fig. F, 1, 2), where, in the centre of the bladder-like apex, a small depression occurs, in which are seen two shortened hairs, whose apices just appear above the general level (fig. F, 3 a). Be- sides these is a multitude of very delicate hairs (true) which seem to be collected into two grou-ps, those towards the middle of each group being the tallest. To these, as in all others, the nerve is plainly seen proceeding. The labial palpi of Ilelanotus (PI. XIX. fig. G) present a similar structure, the smaller hairs being shorter and more distinct, as shown at fig. G, 2. Their diameter is yoVo ii^ch. The surface of the maxillary palpi has nothing but the short sunken hairs over the whole surface of the bladder-like membrane (fig. G, 1 & 3). In Timarchiis, the last segment of whose palpus is much dilated, the apex is covered by a very thin membrane, which is thickly covered by a more elongated form of hair, as shown in PL XIX. fig. H, 1, 2. The nerve (1 a), as before remarked, becomes so enor- mously enlarged or expanded on reaching the roots of these hairs, that it seems not im- probable that it is increased by the addition of some ganglionic matter ; certainly among the expanded nerve-threads globular bodies very much like ganglion-cells may be noticed. In PL XIX. fig. I, 1, 2, I have shown another slightly different form of tactile hairs covering the ^^iiole membrane of the palpi of a Carabiis. Some of them are similar to those of Dyticus marginalis. In another species of Carahis may be seen (at PL XIX. fig. J, 1 & 1 a) only a few tactile hairs ; these, however, are long and well-marked ; while in a third species they are reduced to a very small size (fig. J, 2). A similar condition is observable in the maxil- lary palpus of the Glow-worm {Lampyris noctiluca) (PL XIX. fig. K, 1 & 2), the numerous fine delicate hairs being seen at the edge. But sometimes the termination of the nerves in the surface is so delicate and minute as to have the appearance of mere points even under a |^-inch objective. This is shown in the palpi of Cyclirus rostratus (PL XIX. fig. L, 1 & 2). Some of them, however, are a little larger ; but it is impossible to decide if they possess any hair-like structure, such as appears in Goerius olens (PL XIX. fig. L, 3), where they are shown as roujided elevations of the membrane. However, between these delicate tactile hairs and the large long hairs on some palpi, there are various grades more or less approaching to the con- dition of tactile hairs. For instance, on the apex of the antennse and palpi of the larva of Melolontha vulgaris (PL XIX. fig. M, 1, 2, 3) are found from fifteen to eighteen stiff hairs, which diverge from each other at the apex, and have their extremities roixnded off, being especially blunted on the antennae. In the perfect insect these are altogether VOIi. XXIII. u 146 DR. J. BRAXTON HICKS ON UNDESCRIBED SENSORY ORGANS IN INSECTS. altered in tlie antennas ; but in the palpi they retain their features perfectly, as is shown at PL XIX. fig. M, 4. There is also another peculiarity in these antennae, that on each side is an oval space where the integument is very thin, and covered closely hy a great number of very delicate hairs, and to this part a large portion of the palpal nerve can be traced. The larva of Dyticus margmalis presents extremities somewhat on the same plan (PL XIX. fig. N, 1, 2, 3) : on the apex of each of the three palpi are short conical hau's ; on the external maxillary there are only two, on the internal maxiUary three, and on the labial palpi four. Many of the Orthoptera possess palpi, the apical segment of which is dilated, and is described by many to be terminated by a white, transparent, distended bladder, upon which the main nerve of the maxilla and tongue spreads, and distributes itself upon its superior surface, with the finest branches (Burmeister). This bladder, however, instead of being smooth, is furnished all over with numerous hairs, which difiPer from the others on the palpi in being much smaller and more delicately formed and evenly distriljuted, as is repre- sented in the palpi of Tetrix (PL XIX. fig. 0, 1). An enlarged view of the hair is shown at fig. 2. The length of each hair is about yooo i^ch, the diameter at the base is -g^oo iiich, while the diameter of the elevation from which it grows is 2^00 iiich. If the hair is compared with the smaller palpal hairs of Dyticns margmalis (PL XIX. fig. E, 2), a strong resemblance will be seen. This condition exists in all the Orthoptera I have examined, the hairs being as strong and well-marked as in Tetrkv ; they are well shown in Grylltis, Aclieta, Blatta, &c. A common form of the termination of palpi is shown in the Hive-Bee (PL XVIII. fig. R, 4), and in the IlegacUle and Nomacla (PL XIX. fig. A, 2, and fig. B), in the Scorpion-fly {Fanorpa communis) (PL XIX. fig. P, 1, 2, 3), in Formica (PL XIX. fig. Q, 1). In these the hairs are shorter than the general form of hairs, — still not so much so as those above described, but they are placed on the most prominent points of the apex, evidently for the purpose of touch. Some palpi have no rounded bladder-like membrane, but simply a pointed extremity ; this, I believe, exists chiefly in the larval state. I have figured the internal maxillary palpi of the larva of Colymhetes striatns at PL XIX. fig. R, 1 a. On the apex of the maxillary and labial palpi are a number of small papillae, as at fig. R, 2 & 3. The nerve can also be seen running up to these papiUse (fig. R, 3 a) ; the diameter of each papilla is yoTttoo i^ch. These papillae are so exposed and unprotected by any hairs that doubtless they are used as instruments of touch. Compared with those of the perfect insect, they are very different, the latter being very similar to those of Dyticus marginalis at PL XIX. fig. E. A very peculiar form of palpus is that of Forficula auricularis, in which there is no such delicate membrane ; but the centre of the apex rises up in a cylindrical form, as in PL XIX. fig. S. The apex of this cylinder is closed across by a very thin membrane, the centre of which again is elevated, so as to form a large papilla ; around this latter there can be discerned very delicate hairs. To it also the nerve passes, as at fig. S, 2. Both palpi are alike. In those palpi with a delicate bladder-like termination, the membrane commences DR. J. BRAXTON HICKS ON UNDESCRIBED SENSORY ORGANS IN INSECTS. 147 abruptly, so as to giye the edge in most cases the appearance of a ridge running round the palpus. This is well shown in Tetrix, fee, and in the Coleoptera. Closely resembling these tactile hairs, we find on the proboscis of Ilusca, Ilesembri- num, and many dipterous genera, as EnstaUs, &c., on the delicate u.nder siu-face of that suctorial organ, some nipple-like bodies, which apparently are hairs in origin (PL XIX. fig. T, 3), but which are, with their follicles, much firmer than the membrane from which they spring ; in Mnsca vomitoria and domestica they are well-marked, as in PL XIX. fig. T, 1 ; while in Eristalis, &c. the hair itself is much abbreviated, although the integu- mental tube is long, lying obliquely in the membrane (PL XIX. fig. T, 2). Their minute- ness renders it difiicult to ascertain whether a branch of the nerve proceed to them ; but, from their position on the under surface, which is that used for touching, it is probable that such is the case. And as the proboscis of the Diptera is certainly in part used for the same purposes as the palpi in other insects, it seems also highly probable that these bodies have a similar function, particularly when we consider that the palpi of the Diptera are generally devoid of those pecxdiar hairs which I have called " tactile;" Now these "tactile hairs" are found not only in Insecta, but also in Myriapoda and Arachnida, and probably in all the Articulata. In the common Wood-loixse the apex of the feeler is depressed, from the centre of which spring three shortened hairs, the tips of which appear only just above the edge of the depression. In the Grass-tick {Ixodes) there are two cup-shaped depressions close together near the end of the fore leg, from the centre of which three or four hairs arise, whose apices extend jvist beyond the level of the margin. These I have shown at PL XIX. fig. U. While considering the relation of these hairs to the sense of touch, it seems proper here to mention that on other parts of insects also, hairs, much altered from their usual form on the general integument, are to be found on those parts of the members which neces- sarily come into contact with bodies iu the natural movements; as, for instance, the pads of the tarsi, &c. That these are used as instruments of touch, some have denied ; stUl I am sure that careful investigation would lead to the conclusion that such is their vocation. On examining the tarsal pads, we shall find that they have shortened, conical and delicate hairs, sometimes closely crowded, as in the Wood- Ant (PL XVIII. fig. M, 3) ; sometimes sparingly so, as in the Tetrix. To these hairs I have observed in every case a nerve proceeding, giving off branches to their roots. This is well marked in Formica rnfa (Wood- Ant). (PL XVIII. fig. M, 3, a.) As the pads on the last joint of the tarsus of the Diptera have been closely investigated by others, I have not pursued their anatomy. rinding thus that these modified hairs are situated on those parts that are used for touch- ing, and that they become most modified in form in those parts that are employed more particularly for delicate touch, and are supjDlied with nerves proportionately to their altered form, I think we are justified in concluding that the highest form of feeling, viz. touch, resides in them, especially in those of such extreme delicacy as are found on the terminal membrane of the palpi, while those on the pads of the tarsus, &c. we may suppose to possess the sense of touch also, but in a less degree,— and further, that the hairs of the general surface possess the function of common feeling. To follow the nerve into the u2 148 DR. J. BRAXTON HICKS ON UNDESCRIBED SENSORY ORGANS IN INSECTS. hairs of the general integiiment is a matter of much diffiralty, compared with that of the members where the nerve can be easily traced throughout its course. I now wish to call attention to the anatomy of the " barrel-shaped" organs (as they are called) on the proboscis of the Lepidoptera. They are found in ahnost all genera of English Butterflies and Moths. The genera Fontia, Pieris, Gonopteryx, and the SpldngidcB, are free from them, as is also the rudimentary proboscis xa.t\\e Bomhycidoi and Tiger Moths. Perhaps they are shown best in Argynnis Paphia and Vanessa Atalanta among the Dimma, whilst among the Noctua Mormo and Trtjphcena possess them exceed- ingly well marked. In those species where they are not present, dwarfed hairs are found, in rows on the proboscis ; but whether they are in the same position as the barrel organs woidd haA^e been in, I cannot decide. The form of the barrel organs is very variable ; but they may be divided, so far as I have had the opportunity of examining in British Lepidoptera, into two divisions — the simple and the winged ; the latter I have found only in the Moths, whilst the simple are confined to the Diurna. The simple form is shown at PI. XIX. fig. V, 1, which represents that of the Argynnis Paphia. They are flattened in one direction. The winged are shown at PI. XIX. fig. V, 5, being those of Ilormo Maura (Moth). The section at 6 Indicates the position of the wings on the body of the organ. The essential part of the barrel organs is a tube more or less dilated towards the middle, contracting again towards the apex, and terixdnating in a nipple-like point, —the membrane of this papilla and for some distance around its base being very thin and deli- cate ; and, at the point where this thinning commences, a very delicate tissue stretches across the interior of the organ. A nerve can be traced up to this membrane for that certainty, as is shown at figs. V, 1 & 4, in Argynnis Paphia. I am tolerably certain that a delicate filament passes to the apex of the papilla (fig. V, 2) ; or if that is not the true interpretation of the appearance, then I think it must be produced by a very delicate tube passing from the apex of the papilla into the interior of the barrel organ. This point is difficult to settle on account of the delicacy of the parts to be investigated. In the Vanessce and many other Diurna, there is a ring of eight spines around the papilla, which spring from the part where the thinning commences ; and in the Moths there are seven or eight rings springing from the whole length of the barrel organs, ending at the point just mentioned. The upper points of the rings and of the spines, as in Vanessa, do not, I believe, ever extend to the length of the papilla. That these organs, where they exist, are used for some refined sense, one can hardly doubt, but of what nature it is very, difficult to decide. Whether it is to the sense of touch that they confine that highly- developed variety of it which wc call taste, we are not at present competent to judge. The anato- mical resemblance of these organs to that on the palpus of Foi^ficida will be apparent to all. In considering what functions we are to assign to the various structures I have just described, we have to bear in mind that a too strong dependence on analogy of position of the different parts of Insects, and of the Invertebrata in general, to those of the higher animals will tend to mislead us ; for as the ganglionic nervous system is itself diffused far beyond the state it is in the Vertebrata, so we may possibly find that the special sensations DR. J. BRAXTON HICKS ON UNDESCRIBED SENSORY ORGANS IN INSECTS. 149 are also ; and tliis view is certainly strengthened by the fact that the respiratory organs and their accompanying nerves are distriljuted in a diffused manner throughout the whole length of the Insect's limbs, instead of being concentrated anterior to the ex- tremities, as in the Vertebrata. Reasoning from analogy, therefore, will clearly not hold in this case ; indeed there can be little doubt (and this view is held by some of our best naturalists) that the true homological relations of the various organs in the Vertebrata and Invertebrata are to be decided rather by the analogy of their form, structure, and evident functions, than by that of their position. These circumstances should guide us in determining the functions of the structures I have now brought under notice; but at present I think it would be premature to attempt finally to assign a particular function to any of these organs, excepting those of the antennae, which I have described on two former occasions ; still I may venture to throw out a few hints for further investigation in the matter. That there is every reason to think that the antennal organs are those of hearing, I have stated in my papers above mentioned ; at all events I do not think it possible to suppose they can be olfactory organs, according to our present ideas of the essentials of that organ. It is difficult to understand how odorous particles can pass through, as in some Coleoi)tera, a hard spine-like membrane, then a fluid, thirdly, througli a thin mem- brane to reach the extremity of the nerve, while a wave of sound can readily be con- ceived to be capable of tlius impressing upon the nerve the required impulse. The non- existence of an otolith within these sacs cannot, I think, be considered as conclusive against their being auditory organs, — and for this reason, that as every Invertebrate hitherto described as having an auditory organ is aquatic, so it does not seem a necessary consequence that, should an air-breathing Invertebrate be discovered to have an auditory apparatus, such structure requires an otolith. The conditions of the transmission of the waves of sound are totally altered, and it is possible that the multitudinous repetition of these organs on the antennae in some Insects may preclude the necessity of such an addition within the auditory sac. Regarding the function of the organs so liberally supplied with nerves on the ncrvure of the wings, base of halter es, and elytra, I have already expressed my opinion in the Journal of the Linnean Society [l. c. p. 139) : namely, as in the Vertebrata we find the olfactory organ near the respiratory aperture, so that, by the process of breathing, a constant supply of fresh odorous particles can be brought to it, we may, I think, expect from analogy to find, in those animals where the position of the respiratory organ is altered, that the olfactory sense will accompany it : and I may ask, where should we find it more suitably placed than at the l)ase of the wings, which are so frequently in motion and so near the large thoracic spiracle, through which the air is continually passing. I think we may reasonably conclude that the organs I have above described on the palpi in Diptera and Hymenoptera are in some measure connected with the sense of taste, being situated around the mouth, perhaps supplemental to other organs ; and one can scarcely deny a similar function to the tube running inwards from the apex of the palpi in Lepidoptera. That they are in a position where the aii- is the only medium by 150 DR. J. BRAXTON HICKS ON UNDESCRIBED SENSORY ORGANS IN INSECTS. wliicli impressions can be conveyed to tliem, can easily be seen in Diptera, Hymeno- ptera, and Lepidoptera. The same must be said of those on the legs, which I have described above. If Ave observe an Insect, that joint around which the organs are placed does not touch the surface over which it is running, but remains at some distance above it : in this also the air must be the medium of communication. The delicacy with which odours are perceived by many insects argues an olfactory apparatus of considerable perfection ; and it seems to me not improbable that these latter- named organs may be in some way connected with the sense of smeU, or perhaps with some sense not to be found in the Vertebrata. In conclusion, I again ventiire to recommend the process of bleaching by chlorine, to which I have already called attention in my former papers. It is easily managed, always efficacious, and invaluable in researches of this nature, and equally adapted to decolorize Crustacea, Arachnida, and Insecta. Without it, I should have been unable to arrive at the results here detailed. DESCRIPTION OP THE PLATES. Tab. XVIII. Fig. A. Halter of (Estriis Equi. 1. Showing the groups of organs {ad) at base on one side; b b, nerves to them. 2. Section of base : a a a, organs in section ; b b b, nerves proceeding to each group. 3. Magnified view of the lateral groups: a, seen from above; b, seen in section. 4. Magnified view of the lower group : a, seen from above ; b, seen in section. Fig. B. Under surface of nervure in elytron oi Aromia moschata : a a, vesicles ; b b, nerve proceeding to them. Fig. C. Organs on legs oi Hvppobosca [a a a). 1, OiH. Hirundinis. Diameter of each vesicle ttVit inch. 2. Of H. Ovis. Fig. D. Organs on trochanter-femoral joint oi Musca domestica {a). Diameter of each -^-^^-^ inch. Fig. E. Organs on trochanter of Silpha (1 a a). 2. Enlarged view from above. Diameter of largest xaTu inch. Fig. F. Organs on trochanter oi Lampyris noctiluca (Glow-worm). 1. Under surface : a, organs; b, main nerve of leg; c, branches to organs. 2. Upper surface. Fig. G. Trochanter of Meloe. 1. b b, organs in situ; a, nerves proceeding to each. 2. Section of wall, showing the nerve («) passing to the papilla {b). 3. Magnified view of organs and papillfe from above: a a, nerve ; b, papilla. Fig. H. Trochanter-femoral joint of Forjicula auricularis (common Earwig). 1. Under surface. | ^^^ ^ ^^.g^^^_ Diameter xtVo inch. 2. Upper surface. J DR. J. BRAXTON HICKS ON UNDESCRIBED SENSORY ORGANS IN INSECTS. 151 Fig. I. Trochanter-feraoral joint oi Panorpa communis (Scorpion-i!y). 1 & 2. Both surfaces : a a, organs. 3. Trochanter-femoral joint of Tetrix: a a, organs. Diameter of each, tj\iT7 inch. Fig. K. Trochanter of Gerris : a a, oi'gan. Fig. L. Trochanter of trochanter-femoral joint of Vespa Crabro. 1. Organs in situ. 2. Nerve proceeding to them. Fig. M. Leg oi Formica nifa (Wood- Ant). 1. Trochanter-femoral joint, showing organs [a a), and nerves passing to them {bb). Diameter of organs -giVu- 2. Femoral-tibial joint : a a, organs ; b b, nerves. 3. Tibio-tarsal joint, showing the nerve («) proceeding to the numerous hairs on under surface of first tarsal joint. Fig. N. Organs on trochanter-femoral joint of Tenthredo (Saw-fly). 1. Organs in situ. 2. Section of them. Fig. O. Trochanter-femoral joint of Gryllus viridissimus. 1. Of fore leg: a, organs. 2. Organs enlarged. Fig. P. Palpus of Miisca vomitoria (Blow-fly). 1. Organs in situ. Diameter -r^j^ inch. 2. Organs magnified. Fig. Q. Organs on palpus of Vespa Crabro. 1. «, organs in situ on basal joint. Diameter in Vespa vulgaris -p^Vo inch. 2. Enlarged section of one organ : a, nerve expanding at termination. 3. Organ magnified, seen from above. 4. Apex of the palpi. Fig. R. Organs in Apis mellifica. 1. Labium and ligula, &c. : a, organs at base of labium and mandible. 2. Enlarged view of base of labium. 3. Lip of labial palpi : a, papillas. 4. Lip of maxillary palpi. 5. a, end of ligula, covered with compound hairs, enlarged at b. Tab. XIX. Fig. A. Megachile lignisecans. 1. Organs on base of tongue («) ; on base of palpi (6). 2. Apex of labial palpi. Fig. B. Palpus of Nomada succincta : a a, organs. Fig. C. Apex of palpi of Lepidoptera. 1. Of Argynnis Papihia : a, aperture of tube ; b, nerve expanding on apex of the tube. 2. Of a Geometer- moth : a, aperture of tube ; b, nerve expanding on apex of the tube. Fig. D. Palpus of Bibio Marci. 1. Second segment, showing cavity (a). 2. Enlarged view of portion of cavity. 3. Apex of palpus. 152 DR. J. BRAXTON HICKS ON UNDESCRIBED SENSORY ORGANS IN INSECTS. Fig. E. Maxillary palpus of Dyticus marginalis, 1. Apex of palpus : a, organs similar to antennffi; b, tactile organs; c c, nerves. 2. Enlarged view of tactile organs. 3. Enlarged view of 1 a. 4. Largest tactile hairs. Fig. F. Palpi of Hydrophilus piceus. 1. Maxillary palpus. 2. Labial palpus. 3. Enlarged view of 1, with the two larger tactile hairs in centre. Fig. G. Palpi oi Melanotus, 1. Apex of ma.Nillary palpus. 2. Apex of labial palpus. 3. Portion of apex of maxillary palpus magnified. Fig. H. Palpus of Timarchus. 1. Shows the nerve (a) enlarged in last joint, going to the tactile hairs in apex. 2. Tactile hairs of apex magnified. Fig. L Palpus of a Carabus. 1 . Showing the distribution of the nerve. 2. Enlarged view of apical membrane. Fig. J, Palpi of Carabus. 1. Apex: 1 a, enlarged view of tactile hairs. 2. Apex of another; hairs very small. Fig. K. Palpi of Lampyris noctiluca. 1. Apex of labial palpus, showing delicate hairs on membrane. 2. Apex of maxillary palpus. Fig. L. Palpus of Cychrus rostratus. 1. Shows the distribution of the nerve. 2. Enlarged view, a branch of nerve passing to each papilla, 3. Apex of palpus of Goerius olens. Fig. M. Palpi and antenna of Melolontha vulgaris. (1-3. Of the larva.) 1. Apex of antenna. 2. Apex of maxillary palpus. 3. Apex of labial palpus. 4. Apex of palpus of perfect insect. Fig. N. Palpi of larva of Dyticus marginalis. 1 . Apex of maxillary palpus. 2. Apex of internal maxillary palpus. 3. Apex of labial maxillary palpus. Fig. O. Apex of palpus of Tetrix. L Shows the bladder-like membrane covered with tactile hairs [a a). 2. Shows one of the hairs magnified. Fig. P. Paljjus of Panorpa communis. 1. Apical segment of palpus, showing the nerve within and its branches. 2. Exterior view of apex. 3. Sectional view of apex. Irans.Xvnm/.Soo. KZ ///// hiJi 18 h J -50 G.Jorman.so. Jrans Iinrn Sec lT/Z///i7h>i/9/, 230 'ainruuiv. so DR. J. BRAXTON HICKS ON UNDESCRIBED SENSORY ORGANS IN INSECTS. 153 Fig. Q. Formica rufa. 1. Apical joint of palpus. 2. End of maxilla. The nerves {a) are distributed to the numerous hairs on the edge. Fig. R. Palpi of larva of Cohjmbetes striatus (Striped Water-Beetle). 1 . Maxillary palpus : a, internal maxillary palpus. 2. Apex of labial palpi with papillae. .3. Apex of maxillary palpus, with papilla at apex [b] ; the nerve supplying them is distinct (a). Fig. S. Palpi of Forficula auricularis. 1. Last segment, with elevation in centre, supplied by the nerve (a). 2. Enlarged view of section of elevation, having a papilla in centre [h). Fig. T. Tactile hairs on under surface of proboscis of Diptera. 1. In Blow-fly. 2. In Eristalis (Drone-fly). 3. Enlarged view of latter. Fig. U. Tactile hairs in Ixodes Dugesii, on the fifth joint of the tarsus. Fig. V. Barrel-organs of Lepidoptera. 1. In Argynnis Paphia. A side- and end-view. 2. Enlarged view, showing nerve (a). 3. In Vanessa Atalanta. 4. Enlarged view, showing nerve [a a). 5. In Mormo Maura. ■ 6. Transverse section, showing position of wings. VOL. XXIII. [ 155 ] XIV. Illustrations of the Floras of the Malayan Archipelago and of Tropical Africa. By Joseph D. Hookee, Esq., 31.2)., B.JV., F.B.S., L.S. 8r G.S. (With Nine Plates.) Read June 21st, 1860. X HERE are few countries which possess so many new and imperfectly understood genera of plants as the Malayan Islands and Western Africa ; and the researches of recent travel- lers have added largely to our collections of them. The herharia of the late Mr. Griffith, collected ia the districts around Malacca, Singapore, and Mergui ; of the late Mr. Motley, ill the northern and southern parts of Borneo ; of Mr. Hugh Low, on the same island, and especially on the lofty mountain of Kini Balou ; and of Mr. Thomas Lobb, in Sarawak and Labuan, contain many obscure tropical forms of plants, together with interesting new genera and species of better understood families. In Tropical Africa the researches of the late indefatigable Mr.Barter, on the banks of the Niger river, during Dr.Baikie's expedi- tion, have procured the most extensive and perfect collections ever formed in those countries, which also contain many singular and interesting unpubKshed forms, Mr. Barter's suc- cessor, Mr. Gustav Mann from Kew, has been no less successful in Eernando Po, where he has ascended the Clarence Peak, and procured the first types of a temperate elevation that have hitherto been found in West Tropical Africa*. Other unpublished materials exist from the Gulf of Guinea, collected at Abeokuta by the late Dr. Irving, R.N. ; and still more recently a most valuable and extensive collection of drawings, with ana- lyses, of Eastern Tropical African plants have been received from Dr. Kirk, the accom- plished and indefatigable companion of Dr. Livingstone ; and the collections of that gentleman, which are shortly expected in England, will no doubt supply many more novel- ties of the greatest scientific importance. Erom these and other sources I propose to offer to the Linnean Society a series of papers illustrating the most interesting discoveries they contain. Nat.Ord. ANONACEiE. 1. OxYMiTEA MoTLETANA (H. f.) ; ramulis pedunculis petiolis costaque foliorum superne ferrugineo-tomentosis v. villosis, foliis oblongo-lanceolatis acutis, sepalis abrupte acuminatis, petalis coriaceis sericeo-villosis, exterioribus ovato-oblongis acutis, inte- rioribus late spathulatis. Hab, In ora septentrional! insulas Borneo ad Labuan, Motley. Rami crassitudine pennae corvinae, tomento patente induti. Folia 6-10" longa, 2h-SV' lata, sicca superne fusca V. brunnea subtus pallidiora, venis numerosis lento arcuatis, costa venisque pubescentibus. Flares solitarii, extra-axillares, pedunculis crassis, sepalis petalisque dorso dense ferrugineo-villosis. Petala suberecta, |" longa. * These comprise species of Ttudictrum, Hyjiericum, Ruhus, and Erica, generally very closely allied to Madagascar and Bourbon species. x2 156 DR. J. D. HOOKER ON THE FLORAS OF MALAYA Nov. Gen. Disepalum, H. f. Sepala 2, ampla, ovata, concava, valvata. Petala 4, anguste lineari-spathulata, incurvo-ascendentia, inter se remota, ad marginem tori depress! adnata et persistentia ! Stamina numerosa, multiseriata, con- nectivo apice truncate vix dilatato. Torus latissimus, medio leviter concavus. Carpella 10-15, strio-osa, stylo brevi conico tereti, stigmate terminali, ovulo solitario? prope basin loculi. Fructus (Jeest. — Frutex b-pedalia, glaberrimus, Borneensis, ramis gracUibus, Folia 21-3^" longa, bifaria, elliptico-obovata v. lanceolaia, utrinque angulata, ajnce caudato-acuminata, creberrime pellucido- punctata. Pedunculi terniinales, solitarii, graciles, elongati, uniflori. Flores sub P' diam., virides. 1. Disepalum ANOMALUM, H. f. Hab. In insulse Borneo locis humidis depressis prope Sarawak, T. Lobb. Of this very remarkable plant, in some respects quite as anomalous as Eupomatia, I have only two small specimens. It differs in habit from any Anonaceous plant known to me, and approaches SchizandrecB in texture and some other respects. The minute punctation of the foliage is peculiar ; the stamens and ovaries are typically Anonaceous ; but the two valvate sepals, and four organs which appear to represent the petals, are quite unlike any- thing else in the Order. Technically it will rank in the tribe Uuonecc ; but I know of no genus to which it is at all nearly allied. Tab. XX. A. fig. 1. flower; 2. the same, cut vertically; 3 and 4. front and back views of stamens; 5. carpel; 6. vertical section of the same: all magnified. Nov. Gen. Sph.^koxhalamus, H. f. Sepala 3, maxima, orbiculata, rigide membranacea, subfoliacea, erecta, imbricata. Peiala 6, biseriata, elongato-spathulata, crasse coriacea, imbricata? 3 interiora angustiora. Torus giobosus. Stamina valde numerosa, dense imbricata, obcuneata, connectivo ultra loculos dilatato truncato. Carpella numerosa, oblonga, strigosa, stylo brevissimo obtuso ; ovulis 2 ? superpositis, ventralibus. Fructus deest. — Frutex lO-pedalis, Borneensis, pulcherrimus. Folia magna, subsessilia, elongato-lineari- lanceolata, acuminata, glaberrima, 1^-2-pedalia, basi subcordata apice acuminata. Flores speciosi, rubro-aurantiaci, 3" longi. 1. SpH^ROTHALAMUS IXSIGNIS, H. f. Hab. In sylvis insulae Borneo ora septentrionali, T. Lobb. I have seen only a flower, leaf, and colom^ed drawing of this remarkable and beautiful plant. It belongs to the tribe Uvariece, and apparently to the section with more than one ovule, and will thus rank near Uvaria itself. The ovaria are very minute indeed ; and I have further had great difficulty in analysuig them, on account of the dense consistence of their walls : these are full of indurated wood-cells above, which it is difficult to discri- minate from ovules after both have been dried. Tab. XX. B. leaf; and XX. C. flower: fig. 1. outer, and 2. inner petal; 3. torus, with a few of the lower series of stamina below, and ovaries at the apex ; 4. stamen ; 5. ovary ; 6. long vertical section of the same: all, but 1, 2, and 3, magnified. AND TROPICAL AFRICA. 157 Nat. Orel. MYEISTIGE^. 1. Myristica (Ptrrhosa) grandis (H. f.) ; foliis obovato-oblongis oblongo-lanceola- tisve acutis supra scaberulis, subtus lacunoso-venosis ferrixgineo-tomentosis, pani- culis ferrugiiieo-lanatis, sepalis glandulosis. Hab, In ora septentrional! insulae Borneo, Low. Arbor vasta. Folia 12-15" longa, coriacea, rugosa, supra profunde impresso-venosa, secus costam pube- rula, subtus pilis rigidiusculis tomentosa, costa crassa, venis 12-16 lente arcuatis, margine tenuiter recurva, petiolo brevi crasso. Panicula composita, 6-S" longa, ramis ramulisque dense ferrugineo- lanatis. Flores d minimi, globosi, pedicellis gracilibus glabris; Sepala 3, glabra. Antherce in globum connatae. A magnificent timljer tree, witb leaves a good deal like those of Rhododendron Fcdconeri. Nat. Ord. NYMPH^ACE^. 1. Barclata Motleti (H. f.) ; pedunculis petiolis foliisque subtus jiubescenti-tonientosis, foliis cordato-rotundatis. Hab. In ins. Borneo australis rivulis arenosis umbrosis prope Banjoerang, /. Motley ; Sarawak, Lobb. Tbe genus Barclay a has hitherto consisted of but one species, a native of Pegu and the Tenasserim provinces, where it was discovered by Wallich, who has given an excellent detailed description of it in the Linnean Transactions (vol. xv. p. Jill, t. xviii.), and which has further been figured in the ' Icones Plantarum,' t. 809-]0. The present species is very much larger than the original one, and difl'ers in its tomentose parts, very broad leaves, fewer stamens, longer, narrower sepals and outer petals, more oblong seeds, and other characters. I am indebted to the late lamented Mr. Motley for the excellent draw- ings and dissections given here, which, together with a detailed description of the plant, were the last botanical labours of that admirable observer and collector. Shortly after their receipt at Kew, the news of his murder, and that of all his family, by the Mahom- medan settlers in Southern Borneo, was received in England ; and I have therefore replaced the name of rotimdifolia, which he had given to this species, by that of its ill- starred discoverer. Tab. XXI. fig. 1. back, and 2. front view of sepal j 3. flower with sepals removed ; 4. flower with outer series of petals removed; 5. vertical section of flower, showing the inner recurved petals, stamens, and styles; 6, 7} and 8. three views of the stamens ; 9. transverse section of anther ; 10. ovary; 11. stig- maticrays; 12. transverse section of ovary ; 13. transverse section of ripe fruit ; 14. seed; 15. trans- verse section of seed : all magnified. Nat. Ord. SAUVAGESIE^. 1. ScHUURMANSiA ANGUSTIFOLIA (H. f.); foliis angustissime lauccolatis lougc acuminatis obtuse serrulatis, racemo interrupto, floribus fasciculatis, sepalis a3qvtalibus, antheris rimis lateralibus dehiscentibus. Hab. In insula Borneo ad Sarawak, T. Lobb, 158 DR. J. D. HOOKER ON THE FLORAS OF MALAYA Frutex 10-pedalis, glaberrimus. RamuK tenues. Folia versus apices ramulorum subconferta, utrinque lucida, 5-8" longa, |" lata, petiolo brevi. Racemus terminalis, 8" long., erectus. Fasciculi florum basi bracteolis subulatis cincti. Pedicelli graciles, \" longi. Flores flavi ? A" expansi. Sepala parva, oblonga, obtusa. Petala subajqualia, sepalis quadruple longiora, oblonga, obtusa, concava. Staminodia numerosa, subulata, cum staminum filamentis brevibus in tubum connata. Ovarium anguste ovoi- deum in stylum attenuatum, placentis 3-4. Capsula coriacea, |" longa. This differs from Blume's character of the genus in the sepals and petals not being unequal, and in the dehiscence of the anthers being lateral and a little anticous ; otherwise it appears to be congeneric with the Amboyna plant of that author. 1. Neckia laxcifolia (H. f.) ; caule simplici stipulis setaceis incurvis, foliis lanceolatis duplicato-serratis, sepalis capsulam superantibus. Hab, Borneo septentrionali, T. Lohb. N. kumili affinis, difFert foliis duplo latioribus et sepalis capsulam superantibus. 2. Neckia htjmilis (H. f.) ; caule simplici, stipulis setaceis, foliis anguste lanceolatis ob- tusiusculis duplicato-serratis, sepalis capsulam sequantibus. Hab. In insula Labuan, Borneo septentrionali, T. Lobb, Motley. Caulis simplex, basi lignosus, gracilis, teres, 3-8" longus. Stipules strictse, f" longa3. Folia patentia, 2-4" ionga, i-V' lata, petiolo brevi. Pedunculi capillares, striati, gracillimi, 1-2" longi. Flores nutantes, I" expansi. Sepala lanceolata, sei-rata. Petala infequalia, oblonga, obtusa, petalis vix longiora. Staminodia ad basin tubi staminei pauca parva subulata, exteriora apice clavata. Antherm fertiles 5, oblongss, simplices, tubo insertae. Capsula oblonga, stylo terminata. Semina minima, oblonga, fusca, impresso-punctata. In the venation of the foliage all the above SaHvagesiece present a remarkable simi- larity to Eutliemis. Nat. Ord. POLYGALE^. 1, Epiehizanthus tenella (H. f.) ; caule fiUformi, calyce 5-partito. Hab. In sylvis umbrosis Borneo australis. Motley. Plantula 2-3" alt., simplex v. ramosa. Spic(B i-li" longfe, apice florentes et basi fructus maturos simul gerentes. Filamenta in tubum connata, antheris 5. I have an excellent drawing,with dissections, of this curious little plant, by Mr. Motley, who remarks that it further differs from its described congeners in its calyx cleft to the base. Nat. Ord. BUETTNEEIACE^. 1. BuETTNERiA LAyciFOLiA (H. f.) ; scandcns, ramis teretibus glabris, foliis brevi-petio- latis lanceolatis longe acuminatis integerrimis siibtus puberulis, pedunculo gracili, fructu globoso, spinis patenti-recurvis. Hab. In insulae Borneo ora septentrionali, Lobb. Rami crassit. pennse corvinse, cortice atro. FoZia 4-6" longa, coriacea, 1-U" lata, utrinque venosa, petiolo breviusculo robusto puberulo. Fructus 1" diametro, pilosulus, spinis viridibus laxiusculis \" longis, pedunculo pollicari. AND TROPICAL AFRICA. 159 Nat. Ord. DIPTEROCARPEiE. Nov. Gen. Pachtnocakptjs, H. f. Calyx 5-partitus, laciniis oblongis obtusis, tubo demum cum fructu adnato. Petala 5, linearia, obtusa, libera. Stamina 10, brevia, hypogyna, alterna lougiora, filamentis brevibus incurvis, antheris oblongis, loculis discretis, valvis insequalibus, connective in conum brevem obtusum producto. Ovarium late depresso-ovatum, stylo brevi columnari, stigmata trilobo, loculis 3, ovulis geminis ex apice loculi pendulis. Fructus crasse coriaceus, brevi- urceolatus, ultra medium tubo adnato calycis incrassato 5-lobo margine tumido inclusus, apice umbonatus, unilocularis, evalvis, 1-spermus. Semen pen- dulum, testa tenui endocarpio adhaerente, cotyledonibus convoluto-lobatis. — Arbor elata, Borneensis. Rami teretes, lignosi, cortice sulcata pallida. Folia petiolata, oblonga ovato-lanceolatave, obtuse acuminata acutave, integerrima, coriacea, glaherrima, 4-6" longa, 1^-3" lata, venis sub 6 lente arcuatis, venulis exemplarium floriferorum prominulis reticulatis, fructiferorum obsoletis. Paniculae axillares terminalesque, 1-3'' longce, laxe ramosm, ramis gracilibus, pedunculis pedicellisque cinereo- pubescentibus. Flores 5" longi. Sepala a" longa, pubescentia. Petala sepalis quadruplo longiora, dorso puberula. Ovarium obscure 5-lobum, pubescens. Fructus singularis, |" diam., pericarpio resinas alba scatente. 1. Pachtnocarpus umbonatus, H. f. Hab. In insula Labuan, oi-a septentrionali Borneo, Motley. Nom. vern. " Rasak bunga " v. " Rasak ranga." Mr. Motley states this to be a large tree with soft white wood, and deliciously fragrant flowers, and that the fruit yields a fragrant oil. The habit and inflorescence of Fachy- nocarpus resemble Vatica very closely ; but the fruit is very different, being surrounded by the adnate 5- grooved tube of the calyx. Tab. XXII, A. flowering, and B. fruiting specimen, of the nat. size ; fig. 1. unexpanded, and 2. expanded flower J 3. the same, with petals and two calyx-lobes removed ; 4 and 5. stamens ; 6. transverse section of ovary ; 7- vertical section of one half of ovary ; 8. vertical section of fruit with seed removed : all magnified. DiPTEBOCAKPUs, Gaertn, 1. D. LAMELLATUS (H. f.) ; ramulis petiolisque gracilibus dense tomentosis, foliis oblongis acuminatis subtus dense stellatim tomentosis, fructu crebre lamellato lamellis undu- latis, calycis lobis foliaceis late linearibus apice rotundatis. Hub. In insula Labuan, ora septentrionali ins. Borneo, Motley. D. Baudii, Korth., aftinis, difFert fructu lamellato. — Folia juniora 3" longa, petiolo gracili 1" longo, densissime fulvo-tomentoso, margine sericeo-piloso, pagina superiore pilosa, demum glabrata, infe- riore pilis stellatis prtesertim ad venas operta. Fructus purpureus, 6" longus. Calycis tubus globosus, I" diam.; lacinias duse, venosffi, fere 1" latae. 2. D. MoTLETANUS (H. f.) ; ramulis validis, iiltimis pulvereo-puberulis, foliis longe gracili- petiolatis ovatis basi rotundatis late cuneatisve acuminatis obscure undulatis glaber- rimis, venis utrinqiie sub 16, costa subtus sericea, calyce oblongo tetraptero, petalis dorso cinereo-pubescentibus. Hab. In ins. Labuan, ora septentrionali ins. Borneo, Motley. Ins. Philippinens., Cuming (No. 1073). Ramuli crassit. pennse olerinse, apicibus cinereis v. fulvis. Folia rigidiuscula, 8-1 0" longa, 4'' lata, venis primariis cum costa rectangulum formantibus. Pedunculi 2-3-flori, flexuosi, 3-4" longi, graciles, glaberrimi, flexuosi. Calyx glaberrimus, fere 1" longus. Corolla \V' longa, petalis |" latis. 160 DR. J. D. HOOKER ON THE FLORAS OF MALAYA 3. D. Lowii (H. f.) ; ramulis crassis crebre annulatis ultimis velutino-tomentosis, foliis coriaceis ovatis integerrimis transverse undulatis, petiolo erasso, costa yenisque subtus prominentibus dense tomentosis, pedunculo brevi erasso falyo lanato, calyce dense tomentoso tetraptero, alis undulato-erispatis. Hab. In oi'a septentrionali ins. Borneo, Low. Nom. vern. " Kaga kruning." Rannili ob stipulas delapsas annulati, crassit. digiti majoris, intei-nodiis ^-f" longis, ultimis demum glabratis. Petioli l" longi. Folia dura, 8-10" longa, venis cum costa fere rectangulum formantibus, basi rotundata v. obscure cordata, apicibus obtusis, Pedunculi 2" longi, crassi, 3-4-flori. Calyx l' longus. Corolla dense tomentosa. 1. Drtobalanops aromatic a, Gaertn. fil. Hab. Labuan, in ora septenti'ionali insulee Borneo, Motley (85). 1. Anisoptera ? MELANOXYLON (H. f.) ; ramulis teretibns glabris, foliis ovatis oblongisve apice angustato obtnso basi rotundata utrinque glabris, venis inconspicnis, paniculis folio brevioribus paucifloris, pedunculis pedicellis calycibusque dense fulvo-tomen- tosis, staminibus sub 12, connectivo apice breviter uncinate, stylo gracili. Hab. Borneo septentrionali, in insula Labuan, Motley. Nom. vern. "Rapak." Rami teretes, crass, penna; anserinse. Folia coriacea, 3" longa, petiolo i-|" longo, utrinque glaberrima, supra vix lucida, subtus venis primariis tenuibus. Inflorescentia FlcEocarpo subsimilis. Flores nutantes, ^" expansi, pedicellis gracilibus. Sepala lanceolata, dense tomentosa. Petala oblonga, basi contracta (non imbricantia). Stamina brevia, filamentis brevibus late triangularibus, antheris oblongis. Fructus globosus, basi calj'cis tubo cinctus, stellato-tomentosus, ^" diametro. Calycis fructiferi lobis omnibus accrescentibus stellato-pubescentibus, diiobus maximis lineari-oblongis, 2" longis, |" latis, 3-nerviis, pubescentibus, apicibus obtusis, cEeteris V' longis lineari-lanceolatis. This plant differs from Anisoj^tera in the short liooked apex to the anther, and slender style, and from Roi^ea in the petals not imbricating at the base. Motley describes it as being a second-class tree. Trunk almost invariably divided at a great height into two heads. Bark scaly, studded with tears of clear yellow resin. Sap-icood thin and white. Duramen, first yellow ; when seasoned, black ; bitter, and when chewed jdelding flakes of white resinous substance; easily worked and taking a beautiful polish. Next to the Dryobalanops, this is perhaps the most valuable wood in Labuan. 1. Vatica oblongifolia (H. f.) ; ramulis robustis puberulis, foliis longe petiolatis lineari- oblongis oblongo-lanceolatisve abrupte acuminatis coriaceis utrinque glaberrimis, paniculis erectis axillaribus terminaUbusque folio brevioribus, floribusque ferrugineo- pubescentibus, petalis linearibus. Hab. In Borneo ora septentrionali, Low. V. Rassak, Bl. [Retinodendron, Korth. Verhandl. t. 8), similis, differt ramis robustis, foliis latioribus magis oblongis abrupte acuminatis longius petiolatis, paniculis plurifloris ferrugineo-pubescentibus, petalisque linearibus. — Petioli 2|" longi. Folia valde coriacea, 7-9" longa, 3" lata, subtus pallida, vix glauca. Flores fere V' longi. V. Rassak and V. venulosa, Bl. (Mus. Bot. Lugd. Bat. ii. 32), are the only other de- scribed Bornean species. AND TROPICAL AFRICA. 161 1. HoPEA MICRANTHA (H. f.) ; iTiiaute lepidota, ramulis nigris, foliis coriaceis ellijitico- ovatis ovato-lanceolatisve caudato-acuminatis, subtus petiolis peduneulisque parce pul3eralis, paniculis brevibus axillaribus terminalibusque, fioribus densis secundis, calycibus cereis, petalis dorso sericeo-tomentosis. Hah, In ora septentrional! Borneo, ad insulam Labuan, Motley. Nom. vern. "Dammar mata Kucliing" (or " cats'-eye resin ") and " Dammar Batu." Folia 2-2^" longa, subgrisea, subtus cinerea, petiolo \-l" longo. Racemi ^-1" longi. Flores yL" longi. Petala oblonga, basi connata. Stamina sub 12. Motley's specimens vary a little in tlie breadtli of tlie leaf. He sends them under two names, viz., as the " Rock Dammar," a very large tree with furrowed bark, hard wood turniag brown when cut, yielding sparingly a white resin, and one of the best timl)ers ; also as the tree yielding " Dammar mata Kuching," " of no great diameter, but tall, with smooth buttresses, deeply fissured bark on the trunk, wood hard, light yellowish-brown, durable, easily worked; the resin exades iu large lumps from the truak and branches." Of other Bornean Dixiterocarpe\" latus, venosus, puberulus, demum glaberrimus, coriaceus. Semen anguste oblona'um, atrum. Trans. IjBN.SocVoi.lXlU.TAb XX. "■KididtU'lnli. A. DisepaJui7i cLnomabiiri, N. f. B. Sph£F'rothal;injiis rnsigms, // f. TBA^s.iiMN-, SocTolIHII.Ta13.XXI. arcla;ya. Moi.le}^!, E. I\ Trans l™ &cToiI}an.lAB.XXII, "Ptdi.ad jiiiOi. Pachyno carpus uiiilooiLauus, i7_. f. Trkns.Ijwi Soc^OL.IXin.3AB.XXIII WKlii,del.8;liili. MangiEera Kemanga, ^Z< iRfflsIiEff. Soc ^ypi.TXTlI. TaB.}(XIV. WHi4,ii&li4,, Peiitaspudon MotLeyi, S. f. TraT.3 Li-ni-5ac7ol._XIIIL tabylV WRtch,4aetlith. HacmatostapKis Bartrri , ///' Vincenv Broofejrnp . TTaxislTOT^Soc- .Vol XXIII tablMl. l^FiiaiidethtK TnAK s.Ij^-m. Soc:vi3], . xxi Li ;iAB XX '/1 1 ^^.^. '■i'ivii,'U,!-^ia,, TngoiiocMajiiys Criffithn, Et. Ieans Lnw SocYoL^XIII TabX'IVII ^■fiKiddiiii TieimaiLdriaa, Bomeerisis, SJ. W.Westrmp, [ 173 ] XV. On some Oceanic Entomostraca collected hy Captain Toynhee. By John Lubbock, Esq., F.B.S. Sf L.S. Read June 7th, 1860. In the present paper is a list containing fifty species of Entomostraca collected by Captain Toynbee in the years 1858 and 1859, and descriptions of fifteen which appear to be new to science, showing how very imperfect is onr knowledge of these interesting creatures. Naturalists, however, are not only indebted to Captain Toynbee for having collected these and many other marine animals, but also, and perhaps in a still higher degree, for having carefully recorded the place, date, and circumstances of each capture. Such collections as these are highly valuable ; and as Captain Toynbee fully intends to pursue the same course in his fut;u*e voyages, he cannot fail to obtain results very inter- esting in themselves, and very important, perhaps, in their application to practical navi- gation. The collections already made contain numerous Mollusca, Annelids, Zoophytes, and ThalassicoUce, besides Cru.stacea, of which, however, I have as yet confined my attention to the Entomostraca. Our knowledge of the oceanic forms belonging to this group is as yet too imperfect to justify us in attempting to draw any general conclusions. The large genera seem to be ubiquitous, — the Calani, however, growing larger in the Northern Seas, while the Pon~ tellcB flourish more in warmer latitudes. Some few species are recorded as occurring both in the Atlantic and Pacific Oceans ; but none of the cases are, I think, quite satisfactory, as there has generally been some slight difference between the specimens from the two different oceans. Other species appear to be very local ; but until they have been often captured, it would of course be unsafe to come to any conclusion. Many circumstances combine to retard our acquaintance with the geographical distribution of Oceanic species. Certain liues of ocean are repeatedly traversed, while others are almost deserted. The more delicate species can hardly be obtained except when the vessel is going slowly ; so that the calm regions near the line have been examined with (comparatively speaking) considerable care, while the more windy seas of the temperate zones are almost virgin water. The introduction of steam also has of course, in this respect, an unfavourable influence on the study of natural history. It may also happen that a species which has been very abundant for some days may suddenly disappear; and yet the change may be no indication of the geographical limit assigned to the form in question, but be caused entirely by some change in temperature, some threatening of a storm, or some other more obscure cause. The examination of Captain Toynbee' s collection has impressed me still more deeply with the difficulty, or rather the impossibility, of obtaining satisfactory generic charac- ters. I am of course aware that many naturalists are still inclined to consider genera VOL. XXIII. 2 A 174 MR. LUBBOCK ON SOME OCEANIC ENTOMOSTRACA not merely as artificial helps to classification, but as actual groups between wbicb no links are known or will ever be discovered, I have already pointed out certain species wbicb seem to prove the incorrectness of tbis opinion, and I shall have occasion in the present paper to describe more than one species apparently intermediate between two o-enera. Indeed, so far from considering such links as rare, it would be more correct to say that every species is a link between other allied forms. The same argument is ap- plicable to species. Of course, as long as any varieties remain imdescribed there will be gaps which, however, exist only in our knowledge, and not necessarily in nature. How many centuries must elapse, even under the most favourable circumstances, before all the existing animals are known to us ; and even then how small a proportion will be described of the animals which have peopled the world during the countless ages of past time ! How worthless, then, is the argument against the mutability of species which depends on the supposed absence of " links !" When every variety which now exists, and every one which ever has existed, is known, then, and not until then, can this argument be considered con- clusive. Moreover, it is admitted by every one that there are certain species which are especially variable, that is to say, which present two or more extreme forms, with all the intermediate gradations. Now we may fairly ask those who assert that no two species are connected by links, how they would separate the instances of variable animals (which they admit to occur) from the case which they say does not exist. If we were to obtain to-morrow all the links between any two species which are now considered distinct, no one can deny that the two would at once be united, and would hereafter appear in our classifications only as one variable species. In fact, therefore, they first unite into one species all those forms, hoAvever different, between which a complete series of links is known, and then argue in favour of the permanence of species because no two of them are united by links. As bearing on this point, I may also mention that there are in the collection about ten or twelve other species, represented each by very few (perhaps only one or two) speci- mens, which I can neither refer with sufficient confidence to any already known, and which yet differ so little that I cannot venture to describe them as new. I have there- fore put them aside for future examination, either when I have more specimens for examination, or when the old species in question are better known. I do not see what else I could have done; but in this way, no doubt, it comes to pass that specimens which can be decidedly determined are named, and the doubtful forms, in which perhaps many interesting series of links lie concealed, are left for re-examination at that more " convenient season " to which naturalists, like other people, are only too apt to defer any inconvenient duty. A good example of an intermediate form is presented to us by the species which I have named Calanus latus. This species possesses some of the characters of Euch(sta, and others (more numerous) which induced me to place it in Ccdanus. The maxillipeds resemble those of Eucliceta, and are quite unlike the form which prevails in the immense majority of Calani. The long setse with which the anterior antennae are provided, and the long seta at the apex, are also similar to those of Biiclmta ; but, on the other hand, the form of the front part of the cephalothorax, and the absence of long caudal setae. COLLECTED BY CAPTAIN TOYNBEE. 176 seem to show that it is in reality more nearly allied to the typical Calani. More- over, in certain Calani (as, for instance, in C. elongatus) the anterior antenna bears here and there long setse, almost as in EuchcBta. G. latus seems therefore, to me, to form an interesting link between the two above-named genera. The Calanidse, many as are the species which yet remain to be discovered, present us with a very perfect gradation of forms; and indeed the whole tribe of Cyclopoidea offers in this respect a marked contrast to the Cyproidea. This latter group is at present poor both in genera and in existing species, and it is in many respects very aberrant. The Cyclo- poidea, on the contrary, contain an immense number of species, which, as I have just remarked, form a somewhat complete series, or rather network. In this great family, then, it would seem that most of the types are still existing, that there has as yet been no very great amount of extinction, and that the type is not, geologically speaking, very ancient. The nature of the body is certainly not favourable for preservation, and nega- tive evidence is of very little value in geology ; still it is worthy of remark that the group, now so numerous, has not as yet any known extinct representatives, though, from the great diflFerences between some of the existing species, we may safely infer that the group is of considerable antiquity. The Cyproidea, on the contrary, are known to have existed as early as the Silurian period. They are therefore a very ancient type ; a great proportion of the known species are fossil; and as the soft parts are never preserved, we cannot expect to recognize among them more than a few of the links which must have connected the diflFerent genera with one another, and the whole group with what we now consider the more normal Crustacea. Diaptomus ? abdominalis, again, is obviously a link between Diaptomus and the species placed by Dana in his genus Hemicalanus and the ordinary Calani, — possessing, as it does, the second maxillae and second antennse of the latter, with the geniculating anterior antennae and abnormal posterior legs, which are characteristic of the former. Indeed, unless it is admitted to constitute such a Imk, it must be considered as the type of a new genus. I did not, however, adopt this view in 1856, nor am I disposed to do so now. Such a coiu-se, if followed in other similar cases, would lead to an immense and at present unnecessary multiplication of genera. This species, moreover, even " if called by any other name," would be just as much intermediate between the above-named genera as before. In order, however, to retain this species in the genus Diaptonms, it will be necessary to modify the generic character given by Dana, in so far as concerns the pos- terior legs of the females. The structure of the anterior antennae, again, is intermediate between that of Calamts and that of Pontella. The right antenna differs slightly from the left, and may perhaps be rightly described as geniculating ; but the prehensile power must be very small. The second pair of antennae have the two branches equal. The mandibles and first pair of maxillae resemble those of Calanus brevicornis (Trans. Ent. Soc. n. s. vol. iv. pi. 3. f . 3 & 4). The second pair of maxilla? and maxillipeds are as in Calamis; the latter has the terminal segments elongated. Mrs. Toynbee's collection included no mature males. The fifth pair of legs in the females [1. c. pi. 10. f. 6.) difl'er slightly from those originally figured by me. The present specimens, however, were rather smaller, and 2 a2 176 MR. LUBBOCK ON SOME OCEANIC ENTOMOSTRACA perhaps, therefore, immature. Some of them had fom- joints, instead of three, to the ahdomen : this is generally characteristic of the male sex ; yet these specimens agreed with the others in the form of the abdomen and in other respects, and only differed from the mature females in the form of the fifth pair of legs. We do not, however, yet know whether, as in some other animals, the characteristics of the male sex appear first when the animal is mature, or whether in the young male the abdomen, antennee, and fifth pair of legs already resemble those of full-grown specimens. I am inclined to doubt whether it be advisable to retain Dana's genus Hemiccdanus. The characters by which it is separated from Diaptonius are not, I think, of great importance, and are both some- w^hat inconvenient — the diiference of size in the fifth pair of legs in the female, from being applicable only to one sex, and the absence of the four small intermediate segments of the second pair of antennse, because the joints between these segments become fainter and fainter so gradually that in some cases it is difficult to say whether they are present or not. Mr. Darwin, in his admirable work 'On the Origin of Species' (p. 156), observes that secondary sexual characters are very variable, that " species of the same group differ from each other more widely in their secondary sexual characters than in other parts of their organization ;" and again, " that the secondary sexual differences between the two sexes of the same species are generally displayed in the very same parts of the organization in wliich the different species of the same genus differ from each other." The Entomostraca, and especially the Cyclopoidea, present remarkable examples of this law. In Fontella, for instance, the sexual characters are afforded mainly by the anterior antennse and the fifth j)air of legs. The specific differences also are principally given by these organs ; and many of the generic characters in the Cyclopoidea are taken from the same source. The genera Calanus, Fontella, Euchceta, and others are very similar in form, live together in the open sea, and probably upon nearly the same food, and might, at first sight, be supposed to have similar habits. A glance, however, at the great differences in many of their appendages shows that this cannot be the case, and proves to us how little we really understand of their habits and mode of life. CALANID^E. Calantjs. 1. Set of t-lie claw -aOTo'^ of the caudal setse -jtoo", of the caudal lamellae -jf^o" ; breadth of the caudal lamellse awo"- The setae of the appendage to the base of the abdomen were imperfect. One of the specimens had a number of eggs attached to it. Collected November 24, 1857. S. lat. 0° ,30'; W. long. 0° 30'. Pl. XXIX. fig. 18, X 15 ; fig. 19. antenna of second pair, x 30. Sapphimna parva, n. s. Feniince conspicilla contigua, prominentia. Digitus antenna- rum posticarum articulo secundo brevior, articulis valde inaequis, unguiculo dimidii digiti longitudine. Abdomen segmentis primo et secundo angustis truncatis, tertio et quarto latioribus lunatis, postico rotundato. Lamellae caudales ovatae, ad apicem rotundatae, apice interno denticulate, setis duabus terminalibus, duabus lateralibus, omnibus brevibus (lamella fere quadruplo brevioribus) . This species is nearly allied to S. detonsa, a Pacific form. It differs, however, in having longer setae to the caudal lamellae ; and the general outline is a little different. The length of the second segment of the antennae is toou"» that of the finger being -a ooo"> ^^^ of the claw awo"- The breadth of the caudal lamellse is -g-o^oo" > their length is ^^o"j ^^d that of longest seta is -^om'- The total length is about -j^s-th of an inch. Collected April 9, 1858, in S. lat. 0° 30', W. long. 0° 30'. Pl. XXIX. fig. 20, X 30; fig. 21. antenna of second pair, x 60. Sapphirina, n. s. ? The collection contains also a specimen closely resembling S. indigotica. The finger is however as long as the second joint, the claw is not half as long as the finger, and the caudal setae are barely half as long as the lamellse. It must therefore, I think, be con - sidered a new species ; but having only one specimen, I do not like to describe it. Collected in S. lat. 0° 40', W. long. 0° 20'. Sapphikika opvIentalis, Dana. I name this species with some hesitation, as it was originally collected by Dana in the Sooloo Sea, south-west of the Island of Panay. It appears, however, to agree exactly with a specimen collected by Captain Toynbee on the 5th Oct., N. lat. 12°, W. long. 20° 50'. Sapphirina Danai, Lbk. Collected in S. lat. 0° 40', W. long. 0° 20'. B. Conspicilla non conjtmcta. Sapphirina ovatolanceolata, D. Collected in S. lat. O' 30°, W. long. 0° 30'. Sapphirina Gemma, D. Collected, June 18, in S. lat. 0° 40', W. long. 0° 20'. 186 MR. LUBBOCK ON SOME OCEANIC ENTOMOSTRACA Sapphirina Thompsojs^i, n. s. Maris consi^icilla non coniuncta, fronti insita. Digitus et articulus secunclus antennarum anticarum egedem fere longitudinis, articulis duobus digiti insequis, unguiculo longiusculo (dimidium digiti longitudine supe- rante). Lamellae caudales latse, latitudine tamen longiores, apice interno producto et aciito ; sette caudales quatuor, duse apicales, aliae externee, onines breves (dimidio lamellee breviores). I have named this Sapphirina after Mr. Thompson, thinking it btit right that one species should be dedicated to the discoverer of the genus. The body consists of ten segments gradually tapering backwards,— the posterior, however, being, as usual, small and almost concealed beneath the penultimate. It is of a somewhat peculiar form ; so that I have given a separate figure of it. The first five segments are rounded at the sides, the next four have a minute spine at the posterior corners. The seventh, eighth, and ninth segments have on the under side a small toothed flap. The anterior antennee have only three apparent segments, the first and third about equal, the middle one shorter. The setee are short. The total length is -^^ths of an inch. The width of each conspicillum is Y^to". and the space between them -aoW- The length of the second segment of the pos- terior antenna is af&o", and that of the finger is the same ; the claw is ^t o" ; ^^^ length of the caudal lameUa is ^^o", and their breadth -^H-q. The caudal setse are 2/00" in length. Collected February 1, 1S58, at 3 a.m., in S.lat. 0° 30', W.long. 0° 30'. Pl. XXIX. fig. 22, X 15 ; fig. 23. antenna of second pair, x 60. MiRACIA EFFERATA, D. Collected in S. lat. 0° 40', W. long. 0° 20'. MlKACIA. POLTPHEMID^. EVADNE. EvADNE NoRDMANNi, Loven. Collected in S. lat. 0° 40', W. long. 0° 30'. HALOCYPRID^. CONCHCECIA. CONCHCECIA AGILIS, D. I am not quite sure about this species. The collection contained only two specimens. Collected in N. lat. 24° 39', W. long. 23° 28'. S. lat. 0° 30', W. long. 0' 30'. Pi.. XXIX. fig. 29. maxillas and legs, x 30. CoNCHCECiA BiROSTRATA, n. s. Superne visa elongate ovata; latere visa oblonga, sub- rectangulata, postice paulo altior, fronte instar rostri producta et angulo superno postico denticulate. Antennte anticse setis 5 intcquis, una longa, aliis crassis bre- vibus. Spieulum antenna longius, sagitticapitatum. Antennarum posticarum articulus secundus duplo longior quam sequentes simul sumti. COLLECTED BY CAPTAIN TOYNBEE. 187 This species is at once distinguishable from all the others at present known, by the posterior angle of the back being toothed. In general outline it resembles the other species oiConchoecia, but is deeper behind instead of in front. The anterior antennae resemble those of Saloci/j^rls atlantica in the form and arrange- ment of the setse ; the spiculum, h(jwever, is longer than the antenna by its whole head. The mandibles and first pair of maxillae are not unlike those of ILalocypris atlantica (Trans. Ent. Soc. n. s. vol. iv. pi. 12. f. 5 & 6). The shell is reticulated but very faintly ; and the two valves are similar in outline. Length xotli of ^^ inch. Collected April 21 and June 7- S. lat. 0° 40' ; W. long. 0° 20'. N. lat. 7° 15'; W. long. 27= 52'. S. lat. 13° 43' ; W. long. 33° 55'. Pl. XXIX. fig. 26, X 30 ; fig. 27. anterior antenna, x 30 ; fig. 28. postei'ior antenna, x 30. CONCHCECIA INTERMEDIA, n. s. Superne visa ovata, antice rotundata elliptica, postice subacuta ; latere visa oblonga, subrectangulata, fronte iustar rostri producta, dorso fere recto, angulo postero acute reetangulato. Spiculum fere cylindricum, antenna plus dimidio longius. Antennarum secundarum articulus secundus duplo longior quam ultimi simul sumti. Pes mandibularis articulo secundo elongate, tribus ultimis non inflexis, vix attenuatis. This species possesses the general outline of Conchoecia, with an almost cylindrical spiculum and short setae to the first pair of legs. Prof. Dana gives the flexure of the three terminal segments of the mandibular palpus as a generic character. Although, however, it is true that they are generally bent down, this is not always the case. Moreover their base is provided with two strong muscles^a flexor and an extensor ; so that they must have considerable play ; and I have represented a specimen (PI. XXIX. fig. 15) in which they are as little bent as is the case in JLalocyims, which also has the two muscles, and must therefore have the power of bending the three terminal segments. The present species, again, though in general outline it resembles the typical species of Conchoecia, differs from it in having the spiculum cylindrical ; and the palpus of the man- dibles is at any rate not always inflexed. The anterior antennse resemble those of H. atlantica ; but the spiciilum is longer than in that species. The five setae of the appendage to the second antennae are aU elongated, as usual; but their appearance is peculiar ; and I am uncertain whether they are perfect or not, as they scarcely taper at all and yet do not look as if they had been broken off. The same remark applies also to the setae of the anterior antennae. The terminal setae of the first pair of legs are rather short, as in Halocypris. The two valves are similar to one another. The shell is not latticed. Length -^oih of an inch. Collected May 3, in S. lat. (f 40', W. long. 0° 20'. Pl. XXIX. fig. 30,x30. 188 MR. LUBBOCK ON SOME OCEANIC ENTOMOSTRACA CoxciiOECiA CURTA, n. s. Corpus curtum. Latere visa literse forma D similis, dorso fere recto, antice rostrato, postice rotunclato; superne visa antice rotundata, postice subacuta. Spiculuui sagitticapitatum. Antenna priinse setis 3 longis, subsequis. Antennarum secundarum articulus secundus duplo longior quam ultimi simul sumti. Pes mandibiilaris articulo secundo vix elongate, articulis sequentibus subasquis, vis attenuatis. Pedes priixd setis 3 apicalibus, longis, articulo ultimo perbrevi. Surely this species must be considered as intermediate between ConchceGia and Ralo- cypris, since, with the sagittate spiculum, the first pair of maxillse, and the long hairs at the end of the first pair of legs, which belong to the former, it possesses the general outline of Salocypris. Moreover the second segment of the mandibles is not so much elongated as in CDnclicecia. Upon the whole, however, the characters which point to Conchoecia seem more important than those which would unite it to Jialocypris. The generic description, of Conchoecia must, however, be altered ; but I am unwilling to do so until we are acquainted with a greater number of species, and with the differences which may occur at different ages. The difference in general outline is very marked between the extreme forms ; but some of the new species already discovered tend to fLU up the gap, and render the rule less easy of application. The spiculum is ^Mo" in length, the anterior antenna being at to" = they are 3-jointed. The little appendage of the posterior antennae has a strong spine on the basal segment, which is opposed to another, longer, curved spine, which springs, as Avell as the four long setse, from the small apical segment. In outline the two valves are nearly, if not quite, similar to one another. The shell is reticulated, as in Halocypris rostrata. Length -g-^oths of an inch. Collected April 19, in S. lat. 0° 40', W. long. 0° 20'. Pl. XXIX. fig. S, x30; fig. 9. anterior antenna, x 30. Haxocypris, D. Haloctpris atlantica, Lbk. This species was described by me in the ' Transactions of the Entomological Society,' n. s. vol. iv. part ii. My previous figure, however, gives a somewhat incorrect idea of the organ. The two antennse were lying one exactly over the other (as I have ascertained by referring to the actual specimen copied) ; and I have consequently represented too many hairs. In fact there is one long seta, and four rather short thick rod-like hairs which end abruptly and without tapering. In my previous specimen these hairs were imperfect. The spiculum is slightly swollen at the free end. The second segment of the posterior antennje is twice as long as the succeeding joints. Collected in S. lat. 0° 40', W. long. 0° 20'. Halocypris brevirostris, D. This species appears to be very common. The anterior notch is single in one valve, and double in the other. The latter was on the left side in four specimens examined by me. Collected in S. lat. 0° 40', W. lono-. 0° 20'. COLLECTED BY CAPTAIN TOYNBEE. 189 HAloctpeis rostrata, n. s. Superne visa elliptica, postice subaeuta ; latere visa literas D formli similis, dorso fere recto, antice rostrato, postice rotundato. Antennae anticte setis insequis. Spiculum eylindricum, antenna vix longins. Antennse posticte 7-ar- ticulatge, articulo secnndo duplo longiore quam ultimi simnl sumti. This species is nearly allied to S. hrevirostrls ; and as my specimens wei'e mostly rather smaller and the two were taken together, I was at first somewhat doubtful whether the differences did not depend on age. The second segment of the posterior antennae is, how- ever, longer in proportion to the terminal portion — a difference which Prof. Dana considers of specific value (' Crustacea,' pp. 1302, 1303). Moreover some of the notched specimens were as large as S. brevirostris. The setce belonging to the longer branch of the posterior antennte are plumose ; those of the two-jointed appendage are naked. There is an indication of a minute segment at the base of the long branch ; and the terminal portion also seemed to consist of six segments rather than five. The spiculum is scarcely longer than the anterior antenna. The three terminal segments of the maoidibular palpus scarcely diminish at all in breadth. The second segment is short and broad. The two valves are similar to one another in outline. The shell is latticed by longitudinal and transverse bars. Length -g^th of an inch. Collected April 22, in S. lat. 0° 40', W. long. 0° 20'. Pl. XXIX. fig. 33, X 30 ; fig. 34. mandible, x 30. HAiiOCTPRis ToTNBEEAisfA, n. s. Superne visa, brevissime ovata ; latere visa subrotun- data, literse D forma similis, dorso fere recto, angulis rotundatis, fronte obsolete prominula. Antennae anticae 3-articulatse, setis nudis, uno longiore. Antennae posticae Q-articulatae, articulo secundo plus duplo longiore quam sequentes simul sumti. I have done myself the pleasure of calling this species after Capt. Toynbee, to whose iadustry and love for science we owe this valuable collection. It differs from H. wfiata in general outline, and in the posterior antennae. These latter are unlike those of any other species, in the arrangement and form of the large hairs belonging to the little ap- pendage ; these are six in number : four of them are simple and naked ; one is longer, and clubbed at the end ; and the other is thick, and bent upon itself in the form of a sickle. The larger branch of the antenna is 8-jointed, and the hairs belonging to it are plumose. The mandibles are like those of S. inflata in form, but differ a little in the arrangement of the hairs. The two pairs of maxillae and of legs offer no very striking peculiarities. Length -rsth of inch. The animals were very active. Collected June 25, in S. lat. 0° 40', W. long. 0° 20'. A collection made in N. lat. 7° 15', W. long. 27° 52', contained some specimens agreeing very closely with this species, except in wanting the recurved spine on the appendage of the second pair of antennae. This may perhaps be a sexual difference. Pl. XXIX. fig. 35, x 30 ; fig. 36. anterior antenna, x 30 ; fig. 37. posterior antenna, x 30 ; fig. 38. man- dible, X 30 ; fig. 39. carapace, x 30. VOL. XXIII. 2 C 190 MR. LUBBOCK ON SOME OCEANIC ENTOMOSTRACA PCECILOPODA. Baculus, n. g. Bactjlus elongatus, n. s. Cephalothorax 4-articiilatus, segmento antico maximo, sequeiitibus brevibus, subaequis. Oculi duo, parvi. Antennae quatuor, posticis pre- hensilibus. Truucus biiccalis magnus. Maxillipedes fortes. Pedes natatorii octo, biremes. Abdomen 1-articnlatura, elongatum. The interesting specimen above described is quite unlike any other witli which we are yet acquainted ; but as it bears no eggs, it is impossible to ascertain what is its sex, or whether it is yet mature. If not, it may be a young state of some little-known species ; but unless this is the case, it must, I think, be considered as the type not only of a new genus, but also of a new family. In general form it resembles a constable's staff. It is thickest near the front end, and gradually tapers to the other extremity. The cephalothorax is four-jointed; the first segment occupies more than two-iifths of the whole length. The three following segments are small, and rounded at the sides. The eyes are two in number, small, and seated on a mass of pigment. The anterior antennae are short, three- or four-jointed, and clothed with rather long hairs on the front end. The posterior antennae are shorter and stouter. They end in a large claw, which gives them a prehensile character. Immediately behind the large suctorial mouth is a pair of jaAvs, the homologies of which I have not been able to determine. The natatory feet are eight in number. The abdomen appears to consist of only a single segment ; it is ringed at the sides, and ends abruptly. At the posterior extremity are two small, slightly projecting disks. This remarkable species belongs apparently to the Ergasiloidea, and at first sight some- what resembles Monstrilla. Prom this genus, however, it differs altogether in the form of the abdomen, and in the presence of posterior antennae — a character which might seem to indicate a greater afiinity with Ergasiliis. Length Tijth of an inch. Caught April 9, in S. lat. 0° 30', W. long. 0° 30'. Pl. XXIX. fig. 40, X 30. IramsI}i,n.n,.Soo. Volu G.Ja/rma/n^. so. COLLECTED BY CAPTAIN TOYNBEE. 191 DESCRIPTION OE THE PLATE. Tab. XXIX. Fig. 1. Calanus 'lairabilis. End of anterior antenna, x 30. Fig. 2. Calanus Dance. Posterior leg, x 30. Fig. 3. Calanus DancB. End of anterior antenna, x 30. Fig. 4. Uiidina Darwinil. Fifth pair of legs of male, x 60. Fig. 5. Undina Darwinii. End of anterior antenna, x 60. Fig. 6. Ctytemnestra tenuis. Fig. 7> Clytemneira tenuis. Anterior antenna, x 60 ? Fig. 8. Corycceus Huxleyi, x 30. Fig. 9. Corycceus Huxleyi. Antenna of second pair, x 60 ? Fig. 10. Corycceus Anglicus. Antenna of second pair, x 60? Fig. 1 1 . Corycceus Anylicus. Maxilliped, x 60. Fig. 12. Setella tenuis, X 3. Fig. 13. Sapphirina cylindrica, x 15 ? Fig. 14. Sapphirina cylindrica. Antenna of second pair, x 60. Fig. 15. Sapphirina cylindrica. Caudal lamella, x 60. Fig. 16. Sapphirina nit ens. Fig. 17. Sapphirina nitens. Antenna of second pair, x 30. Fig. 18. Sapphirina elegans, x 15. Fig. 19. Sapphirina elegans. Antenna of second pair, x 30. Fig. 20. Sapphirina parva, x 30. Fig. 21. Sapphirina parva. Antenna of second pair, x 60. Fig. 22. Sapphirina Thompsoni,x 15. Fig. 23. Sapphirina Thompsoni. Antenna of second pair, x 60, Fig. 24. Onccea pyriformis, x 30. Fig. 25. Onccea pyriformis. Abdomen, seen from above, x 30. Fig. 26. Conchoecia birostrata. Outline, x 30. Fig. 27. Conchcecia birostrata. Anterior antenna, x 30. Fig. 28. Conchcecia birostrata. Posterior antenna, x 30. Fig. 29. Conchcecia agilis. Maxillae and legs, x 30. Fig. 30. Conchoecia intermedia, x 30. Fig. 31. Conchcecia curta, x 30. Fig. 32. Conchoecia curta. Anterior antenna, x 30. Fig. 33. Halocypris rostrata,x30. Fig. 34. Halocypris rostrata. Mandible, x 30. Fig. 35. Halocypris Toynbeeana,x SO. Fig. 36. Halocypris Toynbeeana. Anterior antenna, x 30. Fig. 37. Halocypris Toynbeeana. Posterior antenna, x 30. Fig. 38. Halocypris Toynbeeana. Mandible, x 30. Fig. 39. Halocypris Toynbeeana. Part of carapace, x 30. Fig. 40. Baculus elongatus, x 30. [ 193 ] XVI. On the Anatomij and Development of Pyrosoma. Bij Thomas H. Huxley, Esq., F.B.S., F.L.S., Sec. G.S., Professor of Natural Sistory in the Government School of Mines. Read December 1st, 1859. § 1. Sistory of the Genus Pyrosoma. The genus Fyrosoma was first established in 1804 by Peron, in a memoir* published in the fourth volume of the ' Annales du Museum,' and accompanied by a plate repre- senting the exterior and a longitudinal section of the animal. Peron thus defines the genus and the species which he observed : — " Pyrosoma. "Corpus liberum, subconicum, extremitate ampliore apertum vacuum, aperturje margine intus tuberculis cincto. " Pyrosoma atlanticum. J^Iquatorio-atlanticum, gregarie pelagivagum, viridissime phosphorescens, coloribus eximiis tunc efl'ulgens, in aquis viginti duobus reaumu- rianis calidioribus occurrens, 10-12, 14-16 centrimetros sequans." M. Peron's conceptions of the exigencies of a zoological diagnosis were evidently of a singular kind, and his memoir contains not a single observation calculated to throw light upon the true nature of one of the most remarkable animals that has ever been discovered. With respect to the striking property which gave rise to the name conferred on the genus, Peron asserts that the Fyrosomata exhibited movements of alternate con- traction and dilatation at regular intervals ; and that each contraction was accompanied by the development of a luminosity, which, when at its brightest, was red, but, in dying away, passed through shades of orange, green, and blue. The light was developed upon irritation, and entn-ely ceased with the animal's death. The only indication of locomotive power was the regular contraction just described, whose necessary effect was a slight retrogressive movement, in consequence of the reaction of the water forced out of the open end of the Fyrosoma. In 1815, Lesueur, having previously, as he states, described and figured a new species (P. elegans) in the ' Nouveau Bulletin de la Soci^te Philomatique ' for 1813, added a num- ber of important details to Peron's account in his " Memoire sur I'organisation des * "Memoire sur le nouveau genre Pi/rosoma," par M. Peron. Annales du Museum, tom. iv. p. 437, lSO-1. Forskal's Descriptiones Animalium (177o) contains the following passage : — " 29. Medusa BuRois. Tres ejus varietates vidi vel species. * * * * * "29c. RUFESCENS: ovato-oblonga ; ssepe 5 poll, longa ; intus prorsus vacua. G&llice, Concoml/re de la mer. In mari Mediterraneo frequens." Was Medusa Ber'Oe rufescens a Pyrosoma 1 VOL. XXIII. 2 1) 194 PROFESSOR HUXLEY ON THE ANATOMY Pyrosomes, et sur la place qu'ils doiveut occuper dans une classification naturelle*," and showed that Lamarck was in error in assigning to JPyrosoma a place near Beroe, the animal being, in reality, a mollusk closely allied to Salpa {I. c. p. 420). The species described by Lesueur was named by him F. giganteum, and was obtained in the Mediterranean, near Nice. JPyrosoma giganteum, says Lesuenr, has the general form common to the two other species ; it is transparent, of a starchy blue colom', soft and gelatinous, though slightly coriaceous; its only aperture, placed at the upper end, is bounded by tubercles and provided with a membranous expansion, which in certain cases serves to close it. The whole body is covered externally with tiibercles, but these are not disposed regularly like those of Pyrosoma elegcms ; they vary in their dimensions, some being short and indistinct, while others are greatly developed. The largest are conico- cylindrical, flat- tened and lanceolate at the extremity (while those of P. atlmiticum are simply conical), with a small aperture situated upon that side which looks towards the bottom of the sac : this lanceolate extremity is notched on its sharp edges, and presents below, between its pointed extremity and the opening of which we have just spoken, a small but very pro- minent keel. The inner surface of the Pyrosoma is smooth, and provided with a great qiiantity of little apertures, each of which corresponds with one of the tubercles, and is only the anterior end of a canal, whoso posterior aperture is placed at the free extremity of the tubercle, — a fact easily demonstrated by pouring water into the sac-like body of the Pyrosoma; for the water passes out immediately, in a multitude of distinct jets, from the extremities of the tubercles. Lesueur next proceeds to describe the internal structure of the Pyrosoma. He men- tions the internal and atrial tunics as one internal tunic, and points out their distinctness from the external, except at the aperture and over those rounded lateral bodies, which I have much reason to think are renal organs. The branchial networks are recognized as such ; the endostyle is described as " un vaisseau replie sur lui-meme ;" the testis is noted, but is interpreted as the liver. The stomach is determined as such, while the intestine is regarded as the oesophagus ; and the oesophagus is considered to be the pylorus, opening into what Lesueur regards as the intestine — " un canal assez large, glanduleux vers sa base " (p. 417), but which is, in reality, a sinus full of blood-corpuscles. The peripharyngeal ridge is accurately described as " deux petits filets qui vont en se courbant de chaque c6t6," &c. (p, 419) ; and the nature of the nervous ganglion is rightly determined. Pyrosoma is classed among the compound organisms, and the foetuses are carefully though briefly noted. Lesueur confirms P^ron's statement concerning the rhythmical contractions exhibited by the whole body in the Pyrosomata. The figures which accompany this memoir are exceedingly good. I judge from them that Lesueur observed the atrial muscles, and that he has mentioned them as the line which separates the first zone of his transverse section (fig. 13 h) from the second (p. 415) ; and again, in the description of the figures 5 & 6, as " les filets qui forment un reseau dont I'usage parait etre de lier les auimaux du Pyrosome entre eux." In n, fig. 5, I imagine * Read to the Socicte Philomatique de Paria on the 4th of March, 181.5, and published in the ' Journal de Physique * for June of the same year. AND DEVELOPMENT OF PYROSOMA. 195 I recognize an ovisac. Lesueur describes it as one of tlie ' oeufs ' or foetuses, which are well represented in figs. 8-11. Contemporaneously with Lesueur *, that great, but unfortunate anatomist, Savigny, directed his attention to the Pyrosomata, the peculiarities of whose structure found, at length, an adequate expositor in him ; and his account of the anatomy of Pyroso77ia giganteimi is at once so lucid and so concise, that I cannot do better than reproduce it, as an introduction to my own memoir. The subjects of Savigny's observations were obtained at Nice by Risso, and by him sent to Cuvier. " This Fyrosoma (P. giganteum) is a large cylindrical tube, composed of a gelatinous transparent substance, closed and rounded at one end, at the other, truncated and pro- vided with an aperture narrowed by an annular diaphragm, whicli is not without analogy with the membranous circle of the BotrylUdce. The surface of the tube presents conical and smooth eminences of different sizes, some simple and very short, others longer and terminated by a lanceolate piece. Each eminence is pierced at its apex, behind the base of the lanceolate piece, when this exists, by a little cu'cular hole, surrounded by a brown and projecting edge. This aperture, in my opinion, serves to give entrance to the water, and leads into the pharynx. " The inner Avail of the tube presents slight hemispherical enlargements, which corre- spond with the conical eminences of the external surface, and which are likewise pierced at their apices. The latter apertures, similar to the foregoing both in form and number, are situated opposite the anus, and give exit to the faeces. " This diametrical opposition of the orifices of its cells is a novel peculiarity of the Pyrosoma, and determines the form of the whole body. The functions of each of these orifices seem to me to l)e sufficiently indicated by their relative position. One is natu- rally inclined to think that in this genus, as in the foregoing t, it is the most prominent orifice which transmits the food to the pharynx and which admits the water requisite for the branchiae. Besides this, the water, incessantly renewed at the outer surface of the tube, could not be so rapidly or completely changed in its interior. The arrangement of the viscera in each animal agrees with this first indication. " To describe the animals of the Fyrosoma, we may suppose the cylinder to be placed vertically on its base — I mean, on its rounded and closed end ; for the opening of this body is evidently its summit. Each animal then represents an elliptical sac, compressed late- rally, whose great axis is horizontal, and consequently perpendicular to that of the cylinder. This sac, formed by a delicate and transparent tunic, is attached to the cell which contains it, only by the circular opposed apertures of its two ends. The extremity which is turned towards the axis of the cylinder is simply rounded : that directed towards the circumference is prolonged into a neck, whose length is proportional to the jn'o- jection which the cell makes externally, and whose orifice is provided with a festooned * The second memoir of the second part of the celebrated ' Mdmoires sur les Animaux sans Vertebres, ' entitled " Observations sur les Alcyons a deux oscules apparens, sur les Botr^vlles, et sur les Pyrosomes," bears the inscription, " Lues a la premiere classe de I'Listitut le 1" Mai 1815 ;" with the note, " Ce mt'iiioii'c a etc presente le 17 Avril ; mais les travaux de la classe en ont fait differer la lecture." ■\ [viz. BotiyUus.] 2 D 2 196 PROFESSOR HUXLEY ON THE ANATOMY membrane. The lower edge of the sac exhibits the same brown and undulating vessels as the back of the foregoing species, and ought in consequence to be regarded as the cor- responding region. The branchial cavity is very large ; it occupies those two-thirds of the tunic which lie nearest the circumference of the cylinder : its bottom, which is completely open, communicates freely with the other third, which lodges the viscera of the abdomen. These are small, and situated on the right side. The space which they leave unoccupied is commonly filled by the foetuses, which successively arrive and are developed there, as we shaU see below. The structure of the branchial sac in the Pyrosomata may lead one to believe that the water absorbed l^y the oral, makes its way out by the anal orifice. This would be a feature of resemblance with the Sal'pce, in which it is indubitalile that the water takes this course. However this may be, the network which lines the cavity is otherwise organized : it is loose, and composed of fine, undulating, opake, white vessels, some of which are longitudinal, while others are transverse and cross the former at right angles — a character which is common to all the genera of this family. The network does not occupy the whole cavity, but only its two lateral walls ; so that there are obvi- ously, in this genus, two separate and opposite branchiae, one on the right and the other on the left, and which are much narrowed, and consequently distant, at the top. In the foregoing genera, the two branchiae, although really distinct, are only separate behind. The pharynx is at the bottom of the branchial cavity, towards its upper angle. The oeso- phagus is curved sharply to be inserted into a notch of the stomach, which is placed behind the bottom of the branchial cavity. The stomach is fleshy, smooth, compressed, ovoid, or slightly cordiform. The intestine, very delicate at its commencement, suddenly enlarges; a short course brings it to the inferior edge of the tunic, where it receives the insertion* of a large organ analogous to the liver ; afterwards it retmms to the stomach, behind which it ends in a simple and rounded anus. The faeces are homogeneous, clear, yellow, and divided into little masses, the last of which is often already engaged in the atrial orifice {pscule anal), which seems to prove that the rectum has the power of elon- gating and of adapting itself to this orifice. " I must remark, that the liver, or the organ which from its position may be regarded as such, is attached to the intestine by a bundle of divergent canals ; that it is rounded, commonly opake, rose-coloured, yellow or brown, strangulated above its insertion, and divided into from eight to twelve ribs, by grooves which converge from its base to its apex ; it is very soft, and may be broken up into oblong pedunculated vesicles. I may add, as a remarkable fact, that, in many individuals, this organ is colourless, and that it resembles a cellular and transparent globule : it also varies greatly in volume ; some- times, and most frequently, it is of the size of the stomach, sometimes five or six times as large t. " The nervous system of the Fyrosomata does not appear to differ essentially from that of the foregoing animals. There are, in like manner, two tubercles, one on each side of the neck of the branchial sac. The anterior or superior tubercle seems to give oflF several filaments, * An error : the organ in question being the testis. t Savigny has here clearly confounded the testis and the ovisac together under the one name of ' foie.' What he calls the ribbed organ is the testis ; the cellular globule is an advanced ovisac. AND DEVELOPMENT OF PYROSOMA. 197 of which foTir ascend on this neck, while the others go to the opposite side. The pos- terior tubercle, wliich is here inferior, though very apparent in certain individuals, is imperceptible in most. There arise from it four opake yellowish or brown vessels, which traverse the lower side of the tunic; they are evidently the four cords of the dorsal groove of the Ascidians*. Along the upper edge, opposite the four cords of the dorsal groove, are seen two wide, short canals, of a yellow or muddy-brown colour, placed parallel, and so closely united that they might be regarded as a single canal, bent like a siphon, and extending from the middle of the branchiae to the oesophagus, where its two extremities end. The interior appears to be cellular t- " This organ, which is sometimes emj)ty and transparent, seems to me to be analogous to that which M. Cuvier regards as the ovary of the Salpce, or at least as their oviduct ; perhaps it is, at the same time, oviduct and fecundating organ. " The ovaries % are orbicular or pyriform, symmetrically opposed to one another, and placed on the sides of the neck of the branchial opening, between the tunic and the bran- chial network, which they usually overlap. They communicate with two small, sometimes coloured ducts, which embrace the neck and descend as far as the loop formed by the siphon-like canals. These ovaries contain a multitude of rounded, very small, but very distinct ova. " If I do not deceive myself, the manner in which these germs arrive at maturity is very curious. It would appear that while very small they become detached, one by one, from the ovary, and are successively lodged between the intestine and the bottom of the tunic ; there they continue to grow and to be developed, until their final expulsion. " In fact, Ave almost always find in this locality an isolated germ, which varies much in size. "V^'^hile small, it is only a perfectly white and transparent globule, in which a round aperture, like a mouth, is discernible ; when somewhat larger, this hollow globule already exhibits four little reddish spots ; and when larger still, these four spots have become a chain of four small but distinct little foetuses, which encircle the globule for three-fourths of its circumference. Lastly, Avhen it has acquired its full size, the four foetuses, provided with all their organs, are united and form a complete ring. In this state it equals a third of the size of the individual which encloses it. It is, as one sees, a new Fyrosoma already composed of four animals, and will very soon be independent of the large Fyro- soma in which it has originated. How does it escape ? I know not. If, as is j)robable, it makes its exit by the same aperture as the excrements, this opening must be capable of undergoing excessive dilatation. " These observations, taken in conjunction with those which I have made upon Botryl- liis, demonstrate that the corj)uscles contained in the ovaries of these animals are com- pound germs, not intended for the growth of systems, but for their multiplication. On * The 'anterior tubercle' is the nervous ganglion ; the posterior, merely the anterior end of the endostyle, which is described as the ' cords of the dorsal groove.' t This 'organ' is the intestine of Lesueur, and, as I have said above, is nothing but a mass of blood-corpuscles accumulated in the bypopharyngeal sinus. X These are not the ovaries, but probablv, as I have already said, renal orgnns. What Saviguy calls their ducts seem to be the lower parts of the peripharyngeal ridges. 198 PROFESSOR HUXLEY ON THE ANATOMY the other hand, if we open a Fyrosoma or Alcyonium, we find, among the adult individuals, more or less developed emhryos, which can only have proceeded from simple germs whose existence is manifested successively. These last, then, were all contained in the compound and primitive germ." It is obvious from the last paragraph that Savigny was unaware of the origin of the latter ' embryos ' by gemmation. In the ' Systeme des Ascidies ' appended to the ' M6- moires,' Savigny forms the Pyrosomata into a family — the lAicice, containing one genus, Fyrosoma, divided into two groups of species, P. verticillatum and P. paniculatum. In the former the animals are verticillate, or disposed in regular rings which project at intervals. It contains the single species Fyrosoma elegans, 15 lines long, with seven projecting rings, the first and the last terminal ; the tuberosities composing the rings are lanceolate at the ends. There is no annular diaphragm around the wide mouth of the tube. It inhabits the Mediterranean, near Nice. Under the head of this species, Savigny makes the following important remark : — " M. Lesueur has observed, that the whorl which terminates the tube at its small end is formed by four tubercles, i. e. by four animals. He thinks that this disposition is peculiar to the species in question ; but, with a little attention, the same arrangement is to be found in the following species, where these four animals seem to be the representa- tives of the four little foetuses which are developed in the eg^ before its extrusion." The Fyrosomata panlculata are species in which the animals are not verticillate, and form very irregular circles, whose apices are everywhere irregularly projecting. This division comprises F. glganteum and P. atlanticum. Fyrosoma giganteum has an almost cylindrical body, the external tuberosities being very unequal, hemispherical or conical; the most projecting having their appendage or terminal papilla lanceolate, subcarinate, and finely denticulated. The opening of the tube is commonly narrowed by an annular diaphragm. The total length of the largest tubes is 14 inches ; the opening, including the diaphragm, is 2 inches across ; the indi- viduals vary in size from 3-5 lines, according as the neck of the thorax is more or less prolonged — a circumstance which is independent of the age of the individual. The Fyrosomata of this species presented several varieties :— • a. Body strongly stained with brown, as well internally as externally, apparently in consequence of a brown substance filling the branchial cavities. Terminal papiUse wide, and for the most part obtuse. Diaphragm very narrow, and leaving a large aperture. Total length 13-14 inches. h. Body bluish or a little violet, perfectly transparent. Papillae very narrow. No annular diaphragm at the aperture, which presented only very young individuals. Total length 6 inches. c. Body bluish, perfectly transparent. Papilte longer and more pointed than in the preceding varieties. An annular diaphragm, leaving but a very narrow aperture, around which almost all the animals were adult. Total length 5, 6, 7 inches. This species inhabits the Mediterranean and Atlantic, bordering the French coasts. In the description of Fyrosoma giganteum, Savigny gives some particulars not men- tioned in his account quoted above. Thus, he says that the tunic offers but few vessels. AND DEVELOPMENT OF PYROSOMA. 199 " except upon tlie diaphragm which surrounds the aperture." The tunic is provided "helow the abdomen with two transverse muscles, and, besides, is marked by interlacing muscular nervnres, which are very fine, and hardly visible with a strong lens The festooned membrane at the entrance of the branchial sac would be exactly circular if its posterior and inferior edge were not prolonged into a point. " Branchiee wholly separated behind, divided in front as far as their bases, rounded or acuminated at their apices ; transverse vessels 18-25, increasing by degrees from the fii-st, reckoning from the top, to the fifth or even the eighth ; longitudinal vessels 11-17, the middle one only reaching the first transverse vessel, the following on each side attaining the second, and so on, the most external vessels being the shortest of all." The other species, or the Pyrosoma atlanticum, has a conical body 6 or 7 inches long, with its external protuberances terminating in subulate points, and inhabits the equato- rial seas. Mr. P. D. Bennett exhibited some specimens of Fyrosoma at a meeting of the Zoological Society on the 25th of June, 1833, and gave an account of their phosphorescence. A paper by the same author, " On marine Noctihicce," printed in the ' Proceedings ' of the Zoological Society for 1837, contains further remarks on the same subject, and the statement that the ' sphincter-hke ' membrane which surrounds the cloacal aperture is capable of contraction. In the ' Comptes Bendus ' for 1810 (tom. x. p. 285), M. Milne-Edwards published some important observations on the circulation of the Pyrosomata, by which he not only de- monstrated, for the first time, the existence of a heart, but proved that in these, as in most other Ascidians, this organ is subject to a regular reversal of its peristaltic contractions. The regular movement of the branchial cilia is also noted in this communication. With the exception of this valuable contribution to our knowledge of the genus, I am not aware that, with the exception of M. Vogt's short paper, to be noticed below, any account of observations on Pyrosoma has been published since Savigny's time, except my own memoir " On the Anatomy and Physiology of Salpa and Pyrosoma, together with remarks on Boliolum and. Apj^endicularia," contained in the ' Philosophical Transactions ' for 1851. In this memoir I have detailed the results of investigations, made under difficult circumstances and with but a few hours at my disposal, upon a single specimen of what I suppose to have been Pyrosoma atlanticum. By the publication of this essay there was added to vrhat had been already made known, an account of the tubules which envelope the intestine, and open into the stomach by a com- mon axis. The lateral circular palettes, called ' ovaria ' by Savigny, were shown not to have the function assigned to them. The blood was stated to be contained in one great sinus which extends through the whole of the body ; and the reversal of the motion of the heart, observed by Milne-Edwards, was confirmed. The ' four undulating vessels ' of Savigny were shown to be the expression of an endostyle, such as exists in other Ascidians. It was further stated that the edges of the vertical branchial bars, only, were ciliated. The ciliated fossa, the peripharyngeal ridges, the languets, and the otoliths were described. The ' liver ' of Savigny was shown to be the testis ; and the form and mode of development of the spermatozoa were described. The characters of the female organs were determined ; 200 PROFESSOR HUXLEY ON THE ANATOMY the presence of spermatozoa in the duct of the ovisac was observed ; and it was proved that the so-called ' simple embryos ' of Savigny are formed by gemmation. Is one of the com- pound embryos were observed in this specimen, however ; and hence I have always felt a great desire to re-examine Fyrosoma, for the purpose of ascertaining the real nature and origin of such singular bodies. From Savigny's habitual accuracy, I had no doubt of their existence and essential correspondence with his account ; but it seemed impossible that they should be developed in the way he describes. In his valuable memoir " Sur les Tuniciers nageants de la mer de Nice*," M. Vogt confirms my account of the structure of Fyrosoma, and adds some remarks, which are, unfortunately, very brief, upon the foetuses discovered by Savigny, and which I failed to find in my specimen. In pi. 10. figs. 9 & 10, some sketches made in 1847, and referred to in a passage of the ' Ocean und Mittelmeer ' of the same author, piiblished in 1848, are given. They are thus described at p. 89 of the present essay : — " In fig. 9 the ovisac is seen below the testicle, immediately in front of the posterior aperture of the body. It has a rounded form, and contains an enormous, yellowish- coloured and almost transparent ovum, below which again are accumulated oviform masses which exhibit a granular vitellus. I could see no further structure in these ova ; but, I must confess, I did not carry my investigations very far. In fig. 10 I have given an outline- sketch of the individual which exhibited five young in its ovisac. The latter has a rounded form, but is much larger than in the foregoing specimen, and having pushed the viscera downwards, it has extended towards the branchial cavity, in which it forms a kind of hernia." Fig. 9 represents, in fact, an ovisac with a segmented blastoderm, Avhile in fig. 10 the ' five young ' are the cyathozooid and the fom* ascidiozooids of a young foetus. I do not understand how the ovisac in M. Vogt's specimen can have occupied the position in which it is figured in fig. 10, the more especially as in fig. 9 it lies in the same place as that in which I have always found it, viz. in the large mid-atrium and altogether behind the intestine. M. Yogt concludes by putting forward the hypothesis that the ova pass down the canal of the ovisac into the neighbourhood of the intestine, where an incubatory cavity, in which their final development takes place, is formed for them in the thickness of the inner tunic. " This incubation, perhaps, takes place only during certain periods of the year, or of the life of the zooid, whence the incubatory cavity is found only in some individuals and not in all. The formation of ova in the projecting ovary would continue for a certain time. The ova would pass, as they became fecundated, into this incubatory cavity, would develope there, and would iiltimatcly be set free to lead an independent existence. Perhaps, indeed, the incubatory sac may be thrown off bodily, and thus give rise to the base of the new cylinder." {I. c. p. 90.) These observations and suggestions obviously leave much room for further inquiry, and my satisfaction will be easily understood when that opportunity of renemng my investigations which I desired, but little expected, was unexpectedly afforded me. In October of the year 1859, Rear-Admiral EitzRoy, F.R.S., the indefatigable Superin- * " Recherches sur les Animaux inferieurs de la Mediterranee," Mem. de I'lnstitut National Geuevois, torn, ii., 1854. AND DEVELOPMENT OF PYROSOMA. 201 tendent of the Marine Department of the Board of Trade, kindly forwarded to me a very beautiful specimen of Pyrosoma giganteum, taken by Captain Callow* in tlie North Atlantic, about 400 miles S. of the Cape De Verd Islands, in the month of August of that year, and admirably preserved by immersion in strong spirits. I was aware, from former experience, that the textures of Ascidians, in general, are admirably conserved in spirit- specimens which are even many years old; and I therefore commenced my inquiries with a sanguine expectation of being able to make out something about the origin of the compound embryos, which a ciu'sory inspection of the specimen proved to exist in abundance. I must confess, however, that I had no anticipation that researches con- ducted upon a preserved specimen of any animal could be followed out so far as I have been led, step by step, to carry these. And had I not had the opportunity of showing many of my preparations to observers of experience and authority, who can bear witness, at any rate, to the perfect distinctness of the most important of the appearances described, I should hardly have hoped to secure a patient reception for delicate embryological inquiries which profess to have been conducted upon thin sections of a spii-it-specimen, rendered clear by glycerine. I have already published a brief notice of the most important facts which have been developed by my investigations in a paper published in the ' Annals of Natural History ' for January 1860, and in a communication to Section D. of the Meeting of the British Association at Oxford in July 1860. § 2. The Anatomy o/PmosoMA giganteum. In the specimen of Tyrosoma under description, the ascidiariumf is a firm, hollow, conical body, 4 inches long, and about |ths of an inch wide at its broad, open end, whilst its rounded apex measures hardly more than half an inch. The translucent, colourless wall of the ascidiarium is on an average about T^g^hs of an inch thick; but it thins towards the open end, ending in a sharp ledge or rim, which is bent horizontally inwards and ends in a sort of circular valve-like lip, nearly |th of an inch wide, around the aperture of the central hollow or cloaca. In relation to the ascidiozooids, the closed, apical, end of the ascidiarium is dorsal or haemal, inasmuch as the heart is situated on that side of the body of every ascidiozooid which is turned towards the apex. The nervous ganglion, on the other hand, is on the opposite side of the body, so that the open extremity of the ascidiarium is its neural end. The outer surface of the ascidiarium is rendered uneven by conical eminences, which are scattered over it at irregular intervals, and which are elongated on their neural sides into longer or shorter processes. Among these lie similar eminences without such pro- cesses and varying in elevation, until they hardly project at all above the general level of the convex surface of the ascidiarium. Each of these eminences bears a small rounded * Finding a specimen could be procured in no other way, this gallant and skilful seaman swam for that he obtained. He informs me that it emitted a strong bluish-white light, sufficient to read small print by. t The entire body of a compound Ascidian may be conveniently termed the ascidiarmm, while the separate zooids may be called ascidiozooids. VOL. XXIII. 2 E 202 PROFESSOR HUXLEY ON THE ANATOMY aperture, tlie oral opening of an ascidiozooid ; and there are other, similar, apertures dispersed between the eminences. In the specimen under description, the ascidiozooids are almost colourless, or have at most a very pale brownish hue ; but how much of this colourlessness may be due to the action of the spirit, I do not know. Such is the general appearance of the ascidiarium. To examine its internal structure, it is expedient to make sections with a razor in various directions. Although not abso- lutely necessary, I found it extremely advantageous to treat these sections with glycerine, or with a mixture of gum and glycerine — a process which not only has the advantage of rendering the tissues extremely transparent, but of preserving the preparations for a very long time unchanged*. It might have been reasonably expected that the tissues would undergo serious distortion in such a medium, but this is not the case ; on the contrary, the most delicate structures, such, for instance, as the cilia upon the branchial sac, are most exquisitely exhibited in glycerine preparations. As I have said above, I have often had occasion to remark the perfection with which the tissues of the Aseidians generally are preserved by strong spirit, and the siibseqixent addition of glycerine seems only to increase the transparency of such preserved specimens, Avithout otherwise altering them. When a segment is cut out of the ascidiarium of Fyrosoma and examined from the inner or cloacal side, the surface presented to the eye is seen to be tolerably smooth, or at most minutely mammillated, and to present numerous small apertures, each of which corresponds with, and is opposite to, one of the apertures upon the outer surface : while the latter, in fact, is the oral, the former is the atrial t orifice of one of the ascidiozooids. In a thin vertical and radial section (PI. XXX. figs. 1 & 4), the orifices are seen to be connected together by a comparatively wide, somewhat oval cavity, composed of the branchial chamber and the atrium of the ascidiozooid, which are separated from one another only by the perforated branchial sac, stretched like a bag-net from one waU of the cavity to the other. It would be a difficult operation to perform, but a fine hair might be passed in at the oral and out at the atrial aperture, through one of the meshes of the branchial sac, without injuring any organ. From what has been said, it follows that each fully-formed ascidiozooid must be equal in length to the thickness of that part of the waU of the ascidiarium in which it occurs ; * Some which have now heen more than a year in my possession exhibit no alteration. t M. Milne-Edwards, in his "Observations sur les Ascidies Composees," 1833, describes the cavity which surrounds the branchial sac, and into which the branchial currents flow, as the ' chambre thoracique ; ' that part of it which receives the faeces and generative elements he terms the ' cloaca,' while he retains the name of ' anus ' for the external aperture of this cloaca. From experience of the inconvenience of this phraseology, I was led some years ago (" Re- searches into the Structure of the Aseidians," Reports of the British Association, 1852) to propose the term atrium to indicate the ' thoracic chamber,' and to reserve the term cloaca for the chamber common to several or many ascidiozooids, as in Botryllus, &c. The aperture of the atrium may be termed the atrial aperture. The membrane which lines it, and which was in part distinguished by Milne-Edwards in the memoir cited, is the atrial tunic. The cellulose integument of an Ascidian is for me the test. The body-wall which underlies and gives origin to this test I term the external tunic. The proper wall of the alimentary canal (with Milne-Edwards, I regard the branchial sac as a dilated pharynx) is the internal tunic of the body. For the meaning of any other terms not explained in the text, I must refer to my "Memoir on Salpa and Pijrosoma" already cited. AND DEVELOPMENT OF PYROSOMA. 203 and tlie whole ascidiaiium may be regarded as a succession of tiers of ascidiozooids enveloped in a common test. The extreme apex of the cone (PL XXX. fig. 5) is formed by only four ascidiozooids ranged round a common point. In the next tier there are at least twelve, and the number increases until, in the widest part of the ascidiarium, there are between thkty and forty in a tier. It should be understood, however, that there is nothing very regular in the arrangement of these tiers, and that the zooids in any given tier are of very various sizes and degrees of development. The Ascidiarium presents for study (1) the ascidiozooids, and (2) the common test which envelopes them*. The Ascidiozooids. — In investigating the structure of the ascidiozooids, an example from the middle region of the ascidiarium may most conveniently be selected for study. Such an ascidiozooid is represented in longitudinal section in PL XXX. fig. 1, m. trans- verse section in fig. 2, and from above and partly in section in fig. 3. It is somewhat irregularly fusiform, a good deal longer than deep, and deeper than broad. Its outer extremity exhibits the oral aperture, which lies upon the hsemal side of one of the above-mentioned conical protuberances, and is overhung by a tongue-like process of the test — the labial process, by whose outgrowth, indeed, its relations and appearance have become so completely altered, that it will be better to become acquainted with the character of the oral aperture in a less modified specimen. On examining one of those oral apertures, in fact, which are hardly, or not at all, raised above the general level of the outer face of the ascidiarium, the plane of the oral aperture is seen to be perpendicular to the axis of the body (taking a line drawn from the oral to the cloacal aperture as that axis). A circular sphincter, composed of a band of unstriped muscular fibres, surrounds the oral entrance, being attached where the lining membrane of the ali- mentary tract (inner tunic : see note, p. 202) and the integument (outer tunic) pass into one another. The inner diameter of the circular sphincter is xioth of an inch ; but the diameter of the oral passage itself is far less, amounting to not more than xwo^li of an inch. This results from the circumstance that the test is thickened at the margins of the mouth, so as to diminish its aperture to this extent ; and it is at the same time puckered, so that when viewed from without, a number of fine grooves appear to radiate from the lips of the aperture. These must not be confounded with certain fine fibres which radiate from the outer margin of the sphincter into the test, and are perhaps muscular (PL XXX. fig. 6). The test ceases to be traceable upon the walls of the oral cavity a little within the sphincter ; and where it ends, the inner tunic is produced inwards into a broad fold ^vith lobed edges, which takes the place of that circlet of tentacles which is found in this posi- tion in most other Ascidians. I shall therefore term this the tentacular fringe. It is divided altogether into thirteen lobes, of which twelve, though irregular, are tolerably similar and roughly symmetrical, while the thirteenth is situated in the middle of the hsemal half of the circlet, and is very different in form and size from the rest. It is, in * In the present memoir I propose to confine myself as nearly as may be practicable to anatomical and embryolo- gical details, reserving the many interesting histological peculiarities of Pyrosoma for a future occasion. 2 E 2 204 PROFESSOR HUXLEY ON THE ANATOMY fact, three or four times as long as they are, and is divisible into a broad trilobed base, shaped somewhat like an ace of clubs, and a narrow fringe-like terminal portion. This may be distinguished by the title of the licemal tentacle (PI. XXX. figs. 6, Qa). The form, size, and relative position of the oral aperture remain the same i]i ascidio- zooids which have the oral aperture mounted upon a very short cone ; but as the cone enlaro-es, its ha3mal grows faster than its neural side, and finally becomes prolonged into the labial process, which bends over at a right angle to the direction of the axis of the zooid*. Concomitantly with, and apparently as a result of, this development of the labial process, the plane of the oral aperture gradually shifts, until, in the first place, it lies parallel with the axis of the zooid, and then continuing to turn, as it were, on its haemal margin, it eventually takes iip a position perpendicular to the axis of the ascidiozooid again, but exactly the reverse of that which it had at first. The labial process so com- pletely overhangs the oral aperture when this stage is attained, that the free access of the water to the interior of the zooid must, one would think, be somewhat impeded. Two very delicate muscular bands, attached to the inner tunic, succeed one another at short intervals behind the aperture of the mouth, within which the buccal cavity rapidly widens, until it attains its maximiim at about the end of the first fourth of the whole length of the zooid. At this point the buccal cavity ends and the pharyngeal or branchial sac commences, the boundary-line between the two being marked by the anterior end of the endostyle and of the epipharyngeal folds, in the middle line of the hsemal side ; the peripharyngeal ridge at the sides, and the ciliated sac on the neural side. On each side, opposite the middle of the peripharyngeal ridge, is the circular patch-like yellowish organ regarded as the ovary by Savigny. The iierijiliaryngeal ridge (ciliated band, mihi, Mem. on Sal])a) is a structure which I have found in all the ordinary Ascidians which I have examined. In Fijrosoma it is a sort of ridge or inward process of the inner tunic, less than g^o^h of an inch broad, on which the epithelial lining of the tunic is peculiarly modified, so as to present the appearance of a multitude of transverse rows of elongated corpuscles, each row being set obliquely to the long axis of the band, so as to be inclined from the ha3mal side and behind, forwards and to the neural side. These corpuscles are provided with short and delicate cilia. If the peripharyngeal ridge is traced upwards on the inner tunic, it is found to reach the anterior extremity of the cleft-Hke entrance to the endostyle, and there to pass into a narrow series of similar corpuscles which runs parallel with, and indeed may be said to form the outer part of, the projecting lip or epijiharyngeal fold (' dorsal folds ' of Savigny and others) which bounds the entrance to the endostyle laterally. Arrived at the posterior extremity of the epipharyngeal fold, these prolongations of the peripharyngeal ridge, or, as they may be termed, epipharyngeal ridges, unite with one another and pass down as a single posterior epipharyngeal ridge along the middle line of the posterior wall of the pharynx to the oesophageal aperture, before reaching which the single ridge divides, and its branches soon cease to be further distinguishable. On * In the figures given by Lesueur and by Savigny, the axis of the labial process is parallel with that of the body. In most of the ascidiozooids of my specimen, the end of the process is turned towards the hsemal side, but in some it is bent the other wav. AND DEVELOPMENT OF PYROSOMA. 205 the neural side, the two peripharyngeal ridges pass on to the elevation of the inner tunic in which the ciliated sac opens, and unite upon its posterior half, widening as they do so, whence their junction forms a triangular area with its apex directed backwards. The very singular structure which I formerly termed the endostyle, and which I was at one time inclined to regard as a kind of internal shell, is, in reality, a longitudinal fold or diverticulum of the middle of the lisemal wall of the pharynx, which projects as a vertical ridge into the hsemal sinus, but remains in free communication with the pharynx by a cleft upon its neural side. In consequence of the thickness and opacity of the epithelium which lines the fundus of this fold, it appears (especially in the fresh state) like a strong hollow rod mounted upon a thin ridge-like plate. Transverse sections, however, demonstrate the true nature of this structure with perfect clearness (PI. XXX. fig. 8). The bottom of the diverticulum is seen to be occupied by two stout cords, formed of elongated epithelium-cells set perpendicularly to the axis of the cord. These cords are separated from one another by a slight interval. Externally and below they are in contact with two lateral cords of similar cells. Anteriorly the lateral pass into the middle cords, while the latter project beyond the anterior boundary of the groove-like entrance into the cavity of the endostyle (and, consequently, of the anterior ends of the lips or epipharyngeal folds which bound it) and, coated by a process of the inner tunic, constitute the free, rounded, anterior termination of the endostyle. Posteriorly, the same confluence of the median and lateral cords takes place ; but here the endostyle extends much further beyond the limit of the groove and its bounding folds, and constitutes a free, hollow cylindroid or conical process, which, as we shall see, plays a very important part in the process of gemmation, where I shall have occasion to speak of it as the enclostylic cone. The liypoflmryncjeal hand is not, as in many Ascidians, separated for the greater part of its length from the neural wall of the ascidiozooid. On the contrary, in consequence of the position of the oesophageal aperture close to the neural wall of the branchial cavity, and the non-extension of the atrium forwards in the middle line, the hypo- pharyngeal band is represented only by the inner tunic of this neural wall, which lies parallel with the outer tunic, and is separated from it only by the neural sinus, which usually contains a great aggregation of blood-corpuscles. These corpuscles are com- monly aggregated more densely in the posterior two-tliirds of the hypopharyngeal sinus, and not unfrequently are divided, more or less completely, into two lateral portions l^y a median clear space. When this state of things exists, the hypopharyngeal sinus, under a low power, presents exactly that appearance which is figured by Savigny as a siphon-like tube. The inner tunic of the hypopharyngeal band is produced in the middle line into eight slender conical processes— the languets, which are situated at tolerably equal distances from one another. Thus both the neural and the haemal walls of the pharynx are separated from the outer tunic in the middle line by nothing but the corresponding sinuses ; and the same holds good of the lateral wall in the region of the peripharyngeal ridge. But, at any point behind this, either a vertical and transverse section, or a view from above, shows that the inner tunic (or pharyngeal wall) is separated from the outer tunic by a more or less wide space, enclosed within a membrane which is totally distinct from l3oth 206 PROFESSOR HUXLEY ON THE ANATOMY the outer and tlie inner tunics, except at tlie atrial orifice, where it passes into the former, and at the anus, where it becomes continuous with the latter. Except for these two breaks, the membrane in question (which is the third tunic of Milne-Edwards, and is what I have elsewhere termed the atrkd tunic) might be compared to a closed serous sac, reflected over the viscera, on the one hand, and over a part of the external tunic, on the other, but leaving a space between itself and both these parts, which space is filled with blood, and forms a part of the general system of sinuses of the body, A careful examination of tlie side view (PL XXX. fig. 1), the upper view (fig. 3), and the transverse section (fig. 2) will render this statement intelligible. In the first, the atrial tunic is seen to be reflected over the posterior face of the stomach and first part of the intestine, and then to form the roof and the floor of the cavity, which lies between the intestine and the atrial aperture, and which I shall term the mid- atrium. In fig. 3, the atrial tunic is shown to be continued forwards at the sides of the intestinal canal on to the pharynx, united with which, it forms the branchial sac. Arrived close behind the peripharyngeal ridge, it is reflected on to the external tunic, and then passes directly backwards to the atrial aperture. The testis and ovisac, which are seen, in this view, over the alimentary canal, lie altogether above the roof of the mid-atrium (fig. 1), and therefore do not in any way interfere with the free and wide com- munication of the mid-atrium Avith the two spaces, or lateral atria, which lie between the branchial sac and the body-wall, and are Avell shown in the transverse and vertical section (fig. 2). Both in this section and in fig. 3, short cords are seen to pass between the pari- etal and the visceral layers of the lateral atria. They are hollow, and place the parietal sinuses in communication with those of the branchial sac. It follows from Avhat has been said that the Avail of the branchial sac of Pijrosoma (and, I may add, of all Ascidians Avith a similar respiratory apparatus) consists, internally, of that portion of the alimentary tract which lies in front of the oesophagus and behind the mouth (or, in other Avords, of the pharynx), and, externally, of the visceral layer of the atrial tunic. Now, these two membranes do not remain entirely separated by the inter- posed sinus, but are united at regular intervals, so as to give rise to hollow vertical bars separated by equally long vertical clefts — the branchial stigmata. Of these stigmata there are about thirty on each side. The most anterior and the most posterior ones are shorter than the others. Anteriorly, in fact, the first is not more than one-third or one-fourth, or even less, as long as the vertical height of the branchial sac. The stigmata, hoAvever, increase in length up to the sixth, and then acquire nearly the height of the sac, so as to leave only a small imperforate space on each side of the lan- guets, on the neural side, and of the endostyle, on the haemal side. Posteriorly the last four or five also gradually diminish, until the hindmost of all is not larger than the foremost. The vertical bars bounding the stigmata are fringed by a single series of elongated cor- puscles, each of which bears a row of long cilia, and (in the dead state, at any rate) all these cilia project outwards into the lateral atria *. * lu the living condition, as Milne-Edwards has hinted, and as I have shown in my memoir {I. c. p. 583), the cilia upon opposite sides of a branchial stigma move in opposite directions. AND DEVELOPMENT OF PYROSOMA. 207 The brancliial stigmata just described are subdivided into quadrate mesbes by some fifteen longitudinal bars whicli lie altogetlier on the inner side of the vertical ones, to which they are attached by their outer edges, projecting like so many narrow shelves into the pharyngeal cavity, and, as I observed in my earlier memoir on this animal*, are devoid of cilia. They terminate abruptly at their anterior and posterior ends, and they do not exhibit the small denticulations along their free edges which I have described in F. atlan- ticum. The water taken in by the oral aperture must pass with perfect ease through these meshes, and then, impelled by the cilia on the vertical branchial bars, make its way through the lateral atria and, on each side of the intestine, to the mid-atrium, whence it finds an exit by the atrial aperture. As to the proper digestive canal, the wide aperture of the oesophagus lies at the posterior, neural, angle of the pharyngo-branchial sac, and has an irregular figure ; Init whether this irregularity is normal, or arises from the collapse of its walls after death, I cannot say. The oesophagus narrows as it passes back, and then curves sharply round towards the haemal side, to open, after a very short coui'se, into the large oval stomach, which lies im- mediately behind the middle of the branchial sac, invested, everywhere but in front, by the atrial tunic, and bathed in the blood which lies between it and that tunic. At its pyloric end it gives rise to the narrow commencement of the intestine, which, after suddenly dilating and turning forwards and to the haemal side, bends back sharply upon itself, and passing backwards to the neural side and to the right, ends opposite the middle of the stomach in the abruptly truncated anus, which opens into the atrium. In my memoirs on Salpa, Pyrosoma, and BoUoliim, already referred to, I have described, in all these genera, a remarkable system of fine transparent tubes which ramify over the intestine, and eventually open by a single duct into the stomach. I have asked (/. c. p. 570), does this tubular system represent a hepatic organ ? or is it not more probably a sort of rudimentary lacteal system — a means of straining off the nutritive juices from the stomach into the blood by which these tubes are bathed ? In Pyrosoma giganteum the duct of the system is very obvious, opening into the stomach in front of the origin of the intestine, and somewhat enlarged at its opposite end. Krohn has described the structure and development of a similar system of tubuU in Phallusiai, and I have since % found an organ of the same nature in Fhallusia, Cynthia, Molgula, Perophora, Botryllus, Botrylloides, Clavelina, AjjUdium, Didemnum, and, in- deed, in all genera of Ascidians which have come under my notice, except AppendiciUaria^ In some species of Didemtmm, 1 have observed that the duct dilates almost at once into a large spheroidal sac. I suspect that Savigny was the original discoverer of this system (see his memoir on Diazona, I. c. p. 176, and the description of pi. 12). The existence of these tubuli in Salpa, Doliohim, and Pyrosoma has been confirmed by all subsequent observers. M. Vogt, however {I. c. p. 31), afiirms that the organ consists of solid branches, and that it partakes of the nature of a muscular organ, in neither of which opinions can I possibly concur. I have no doubt whatever that the apparatus is a glandular organ, and that it performs a part, at any rate, of the functions of a liver. * Loc. cit. pi. 17. fig. 3, and p. .583, line 4, where the word ' sinus ' should be ' ones.' f " Ueber die Entwickelung der Ascidien," Miiller's Archiv, 1852, p. 331. J See Reports of the British Association, 1852. 208 PROFESSOR HUXLEY ON THE ANATOMY The vascular system of Fyrosoma is exceedingly simple ; nor could we anywhere find a more convincing example of the validity (in some cases, at any rate) of Milne-Edwards's views of the circulation in the Mollusca than is offered by this animal. The heart lies close to, and apparently connected with, the right side of the posterior and haemal wall of the pharynx, between the endostyle and the bend of the intestine ; and it appears to have exactly the same structure as in P. atlanticum. There are no vessels, the whole inter- space between the inner tunic and the outer, or between these and the atrial tiinic, being one vast blood-sinus, with which the canals in the branchial bars communicate at each end. I have spoken of the haemal and of the hypopharyngeal sinuses merely as a matter of convenience ; in point of fact, the general blood-cavity is not naturally divided into distinct sinuses*. The nervous system consists, as in Fyrosoma atlanticum, of a single ganglion, of an oval shape when viewed sideways, but somewhat heart-shaped when seen from above, its nar- rower end being turned backwards. It is about xioth of an inch long, and is composed of a dark granular mass invested by a delicate structureless membrane. It lies between the inner and outer tunics, the former being raised, so as to form a slight protuberance over it. On the posterior half of this protuberance lies the broad lower median portion of the peri- pharyno-eal ridge. In its anterior half, the opening of the ciliated sac appears. The prin- cipal nerves given off from the ganglion are the following. Two, a smaller internal and a laro-er external, pass from the antero-lateral parts of the ganglion, forwards towards the oral aperture, branching as they go. I suspect that a nerve runs up on each side, be- neath the peripharyngeal ridge ; but I cannot make sure of the fact. A considerable nervous trunk is given off to the poster o-lateral walls of the body ; and, finally, two deli- cate trunks arise posteriorly, one on each side of the middle line, which run back, so as to have the languets between them, and passing up at the sides of the cesophageal aperture, are lost under the divisions of the posterior epipharyngeal ridge. In Fyrosoina atlanticum I observed a mass of deep-red otoliths in contact with the posterior end of the ganglion {I. c. p. 583), but no such structures are discernible in the pre- sent species. The ciliated sac (' tubercule anterieur ' of Savigny)— an organ of univerl^l occurrence among Ascidians — is mPyrosoma gigantemn an elongated, laterally compressed, * I have described the circulatory system of Salpa in similar terms to these, in my memoir on Salpa and Pyrosoma, and notwithstanding the criticism my statements have received both from M. Vogt and Prof. Leuckart, I must main- tain their correctness. M. Vogt affirms that I have committed ' a grave error ' in declaring the blood-canals of Salpa to be lacunse between the two layers of the mantle,— apparently supposing that I mean thereby the test and the external tunic, and forgetting my careful discrimination of test, outer tunic and inner tunic, at p. 585 of the memoir cited. In fact, nothing can be easier than to observe the entire distinctness of the inner and outer tunics in a bud or embryo of Salpa or Pijrosoma— to see that the viscera and blood-canals do really lie between these tunics, and that they are by no means, as M. Vogt states, lodged in cavities excavated in the ' inner mantle.' Prof. Leuckart has equally mistaken my meaning when {I. c. p. 14) he ascribes to me a participation in Eschricht's opinion as to the existence of a serous' sac surrounding the body of Salpa. My words in the passage cited by Prof. Leuckart are, " In very young SalpcB, this space [the interval between the inner and outer tunics] is like the cavity of a serous sac." Still less can I find in my memoir any such opinions as those ascribed to me in the note to p. 43 of Prof. Leuckart's valuable memoir. While on the subject of errors, however, I am glad to take the opportunity of painting out that several statements made at second-hand in my memoir, regarding Ascidians other than those specially described, are incorrect. The diagram of Pdonaia (pi. 19), again, is altogether erroneous— this Ascidian differing, as I have since found, in no essential respect from Cynthia, AND DEVELOPMENT OF PYROSOMA. 209 funnel-shaped bag, about xootb of an inch long, which lies in the sinus, and, passing obliquely forwards and towards the hsemal side, opens as above described. Its aperture has somewhat prominent lips, and is rather narrower than its upper portion. The poste- rior end of the sac appears to terminate caecally, and is applied against the posterior sur- face of the ganglion. The middle of the haemal side of the sac sometimes appears to be connected with a spheroidal tubercle, whose axis forms nearly a right angle with that of the sac. The muscular system is exceedingly simple in this species of Fyrosoma, consisting, besides the oral sphincter and buccal muscles already mentioned, of only an atrial sphincter and the ' mid-atrial ' muscles. The atrial aperture (fig. 7) is even smaller than the oral, not measuring more than from ■g^th to y^th of an inch in diameter. Radiating striae diverge from its margin on the surface of the test, which, as at the oral aperture, forms a thick lip, and is continued for some little distance inwards upon the wall of the mid-atrium. A sphincter formed of pale smooth fibres, and constituting a circular band xe^tli of an inch in diameter, is developed at the junction of the external and atrial tunics. There is a similar but less distinct appearance of radiating fibres to that exhibited at the oral sphincter. The mid-atrial muscles {g 2) are broad flat bands of smooth muscular fibres, which lie in close contact with, and apparently attached to, the atrial tunic. One of these bands occupies about the middle two-fourths of the height of each lateral wall of the mid-atrium, and has a direction perpendicular to the axis of the ascidiozooid. The bundle of fibres spreads out a little at each end, and then seems to be inserted by a sort of tendon into the outer tunic. Close to this tendinous insertion, at either end, a bundle of fibres (whether merely fibrous or muscular I cannot say) arises, and passes, partly to the nearest similar insertion of one of the mid-atrial muscles of the ascidiozooid above or below, partly to the same point of the mid-atrial muscle of some other ascidiozooid. Hence, when the wall of the ascidiarium is viewed from within, it presents such an interlacement of fibres as that exhibited in fig. 9. These muscles, in contracting, must tend to diminish the capacity of the atrium of the ascidiozooid to which they belong, and, if they all act together, to shorten and narrow the ascidiarium. I do not suppose that their effect in the latter direction can be very great ; but it might well be sufficient to account for the slight contraction of the whole ascidi- arium, and consequent retrogressive motion, observed by Peron and others. In my previous memoir, I have pointed out that the round, granular, yellowish patches on each side of the entrance of the branchial sac, and opposite the middle of the peri- pharyngeal ridge, are not, as Savigny imagined, the ovaries. I am greatly inclined to regard them as renal organs, but for the present defer the discussion of their structure and functions. The reproductive organs of each ascidiozooid of Fijrosoma may be divided into actual and potential — the genitalia of the ascidiozooid itself, and the blastema whence the genitalia of its buds will take their origin. I shall call this last the generative blastema ; while the genitalia proper are divisible, as I have already pointed out, into a single ovisac and a single testis. Both ovisac and testis are situated in the left five-sixths of the roof VOL. XXIII. 2 r 210 PROFESSOR HUXLEY ON THE ANATOMY of the mid-atrium ; but the testis is to the right, the ovisac to the left. Tlie size of both these organs has a definite relation to the age and advancement in development of the aseidiozooid in which they occur, being larger and more advanced as it is older and nearer perfection. Hence the early stages of both ovisac and testis can only be observed in buds and yonng ascidiozooids. The generative blastema will be most conveniently considered in connexion with the process of gemmation and in describing the most advanced condition of the foetus. The description of the ovisac will form the most fitting commencement of the history of sexual propagation in JPyrosoma (§ 3) ; all that remains, therefore, is to give in this place some account of the structure of the testis and of the character of its products. The testis lies in the hgemal sinus above the mid-atrium, and on the right side of the ovisac. It consists of about a dozen cylindroidal caeca, free at their neural ends, but connected at their hsemal extremity with the dilated upper end of a vas deferens, which passes directly to the neural side and somewhat backwards, to open by a slightly raised papilla on the roof of the mid-atrium. The cajca are -i^^ih. of an incli long, or thereabouts. Each consists of a delicate structureless membrana propria investing an aggregation of spheroidal corpuscles about 47^g^th of an inch in diameter. Near the attached end of each csecum the rod-like heads of the spermatozoa become visible, and gradually take the place of the spheroidal cells. The duct has the same structure as the caeca. It presents an upper and a middle dilatation, but is not more than Teoth of an inch wide at its termi- nation. The middle dilatation is usually full of closely packed spermatozoa. The structure which has been described is characteristic of any of the fully-formed ascidiozooids in the middle of the ascidiarium, or towards its apical end, in which regions the number of such ascidiozooids bears a large ratio to the total. But towards the open end of the ascidiarium fully-formed ascidiozooids become scarcer and scarcer, until, close to the inflected cloacal Hp, noiie are discernible. On the other hand, all those ascidiozooids which are to be found in this region possess an appendage which is not to be discovered in the others, in the shape of a long tubular diverti- culum of the external tunic, or stolon, which extends from the neural side of the body, behind the oesophageal aperture, into the lip of the cloaca, at whose free edge it ends in a ctecum. The walls of these diverticula, composed of the external tunic only, exhibit strongly marked parallel longitudinal striae, as if they were composed of mus- cular fibrillae*. The test. — The common integument, or test, in which all the ascidiozooids are enclosed, appears to the naked eye to be quite glassy and homogeneous ; but when thin sections taken in various directions are submitted to the microscope, it is found to possess marked structural peculiarities. Dispersed through its general substance are numerous cells with radiating processes, like connective-tissue corpuscles, and containing a central endo- * Savigny has figured these stolou-hke diverticula in his pi. 22. fig. 1, 1, and he speaks of tliem in the " Systeme des Ascidies," p. 208, where, in characterizing the test of Pyrosoma yiganteum, he says that it generally presents few vessels, " except in the diaphragm of the opening." He appears not to have heen acquainted with the origin of these "vessels." In describing his variety e of this species, he states (sujmi) that the opening was surrounded by animals which were almost all adult. AND DEVELOPMENT OF PYROSOMA. 211 plast or nucleus. The cells measure on an average 4o~o ott of an inch in diameter ; and their processes become very fine before they are lost in the surrounding nearly homoge- neous matrix. In two regions this general structure is departed from. At the cloacal wall of the test, there lies immediately beneath the surface a thin film of reticulated tissue, consisting of cells similar in their essential structure to those just described, but set more closely, more granular, elongated, and united together by the coalescence of their processes. Again, in a plane which would correspond Avith the peripharyngeal ridges of the ascidiozooids, and therefore near the outer surface, the test exhibits a very faint longitudinal striation, as if it were fibrillated. There is no distinct epidermic layer on either face of the test. The corpuscles are, as usual, stained dark yellowish brown by iodine ; whUe the matrix yields, though weakly, the characteristic reaction of cellulose. § 3. The Agamogeiiesis by Gemmation of Pyrosoma giganteum. Throughout the whole extent of the ascidiarium, the number of ascidiozooids appears to be undergoing a constant increase, by the development of buds from those which already exist ; at least, I have not yet met with any adult ascidiozooids devoid of a more or less advanced appendage of this kind. Gemmation always takes place from that part of the middle of the haemal side of the body of the ascidiozooid, which lies opposite the bend of the intestine and between the posterior extremity of the endostyle and the reproductive organs. At first, therefore, the bud is situated near the posterior or cloacal end of the body, and on the same side as the closed apex of the ascidiarium. Gemmation does not take place in Pyrosoma, as in so many of the lower animals (e. g. the Hyclrozoa and Folyzoa, or Salioa and Clavelina among the Ascidians), by the outgrowth of a process of the body-wall whose primarily wholly indifferent parietes become differen- tiated into the organs of the bud, but, from the first, several components, derived from as many distinct parts of the parental organism, are distinguishable in it, and each com- ponent is the source of certain parts of the new being, and of these only. Thus the body -wall or external tunic of the parent gives rise to the external tunic of the bud ; while a process of the endostylic cone of the parent is evolved into the alimentary tract of the bud, and the reproductive organs of the latter are furnished by a part of that tissue whence the reproductive organs of the parent took their origin. PI. XXX. fig. 14 represents the condition of what I may term the gemmiparous region of the body in a young ascidiozooid, in which no distinct trace of a bud is discernible ex- ternally. The outer tunic, it will be observed, passes evenly backwards, and has the same structure in the situation of the future bud as elsewhere. The endostyle is continued upwards and backwards as a celkilar cord, which contains a cavity continuous with the groove of the endostyle, is about rtoutli of an inch thick, and is rounded-off at its extre- mity. From this a thin sheet of indifferent tissue is continued downwards and backwards, so that its plane forms nearly a right angle with the direction of the end of the endostyle, and suddenly thickens to -f^o^h. of an inch. After this it tapers off gradually to its extremity, which lies free in the cavity of the blood-sinus, at a distance of TsTT'th of an inch from the ovisac of the ascidiozooid, which is -^roth of an inch in diameter, and so far 2f 2 212 PROFESSOR HUXLEY ON THE ANATOMY advanced that there can be no question as to its real nature. The total length of this mass (which, for reasons which wiU shortly appear more fully, I have termed the generative blastema) is ri^nd of an inch ; and for the greater part of its extent it has the character of indifferent tissue. Bat the sudden enlargement to which I have referred is occupied by a body which has all the characters of an ovum, consisting of a structureless yelk eioth of an inch in diameter, and of a clear germinal vesicle (xaVoth). enclosed in which is a germinal gpot (— gij^-yth). -It will appear by-and-by that this is, in fact, the solitary ovum (sur- rounded by its rudimentary ovisac) which will come to maturity in the bud to be formed at this spot ; and it is not a little remarkable that the first recognizable part of the new organism should be the foundation of that structure which will eventually develope into a creature distinct from it. In fig. 15 a more advanced condition of a bud is depicted : the backward continuation of the endostylic cone is broader, more distinctly hollow, and is so bent up as to form a more acute angle, both with the line of direction of the endostyle and with the plane of the generative blastema. In consequence of this change, and of the general enlargement of the parts, they can no longer be contained within the l}lood-sinus, whose outer wall is now elevated into a conical cap which fits over the conjoined ends of the process of the endostyle and the generative blastema. That part of the external timic which consti- tutes this cap is thickened, and exhibits the texture of indifferent tissue. The ovum in the generative blastema is now very distinct, and the tissue around it is so disposed as to mark out the walls of an ovisac which measures i^io^^ oi an inch in diameter. The clear germinal vesicle measures sToth of an inch ; and its spot has the same diameter as before. Behind the ovisac, which occupies the greater part of the cavity of the diverticulum constituting the bud, a distinct constriction marks off the rest of the generative blastema, which lies closely connected mth the external trmic of the parent, and altogether excluded from the cavity of the " cap " of the nascent bud. It is now no longer taper, but cylindrical and rounded at the end ; and near its anterior extremity a new germinal spot, surrounded by a small clear vesicle, is visible. Eig. 16 represents a bud giotb. of an inch broad by 4ioth of an inch high. The pro- cess of the endostylic cone is very distinctly hollow and somewhat thin-walled, while its axis is nearly parallel with that of the bud. In fig. 17, the bud, now sub cylindrical, has increased in length to xiTid of an inch ; and the front view of a similar bud, given in fig. 18, shows that the hollow process of the endostylic cone is slightly constricted in the middle, and that the interval between its walls and the external tunic is occupied by a granular mass. In fig. 19, a marked advance is discernible. The bud is distinguishable into a body or rudimentary ascidiozooid jT-atb of an inch long, and a much shorter stalk or peduncle. The ascidiozooid is broad at its attached end, more or less tapering at its opposite extremity. Its external tunic is distinct, but proportionably thinner than before, and is continued into the outer wall of the peduncle and thence into the external tunic of the parent. The hol- low process of the endostylic cone is about as broad as before, in the peduncle ; but after traversing this, nearer its anterior than its posterior side, it suddenly dilates into a pyri- form sac, somewhat similar in contour to the rudimentary ascidiozooid itself. The upper AND DEVELOPMENT OF PYROSOMA. . 213 taper end of this sac seems to be attached to the inner surface of the apex of the outer tunic of the ascidiozooid. Anteriorly and posteriorly, its walls appear thick, the enlarge- ment being much more marked posteriorly. The side of the sac turned towards the eye, between these thickenings, exhibits five faint rings, with comparatively clear centres. In order to avoid circumlocution, I may so far anticipate the results yielded by the investi- gation of later stages of the buds, as to state the nature of the parts which have now been described. The rings are the indications of the commencement of as many bran- chial stigmata ; the anterior apjoarent thickening is the result of the^ formation of the rudiment of the endostyle ; the posterior apparent thickening is produced by the rudiment of all that part of the alimentary tract which lies behind the branchial sac, into which almost the whole of the dilated end of the prolongation of the endostylic cone is converted. A comparatively clear space surrounds the apex of the branchial sac, below which the inner surface of the external tunic presents a band-like aggregation of indifferent tissue, the rudiment of a body which corresponds with what Krohn has called the elseoblast in the SalpcB ; and finally, projecting from the posterior wall of the external tunic, and appa- rently connected with the elgeoblast, is an elongated mass, the anterior portion of the gene- rative blastema, which has now become completely separated from the posterior part. The anterior end of the latter, in fact, extends only into the peduncle, while its posterior moiety lies, attached to the outer tunic of the parent, in the great hsemal blood-sinus.. The genera- tive blastema may therefore be now distinguished into three parts — parental, peduncular, and gemmular, — of which the two former remain connected, until a new bud is developed in the distal end of the peduncle, while the latter, now contained wholly within the bud, and separated from the others by a considerable interval, is itself divisible into three portions. The first of these is the ovum, or rather ovisac, larger and more distinct than in the last- mentioned stage ; the second is that part of the blastema near this, which will become the testis, but which, at present, has no definite form ; and the third is represented by a slender band of indifferent tissue continued up to the apex of the branchial sac (the future extremity of the endostylic cone) which is the generative blastema of the nascent ascidiozooid and Avill supply reproductive organs to its buds. The interspace between the branchial sac and the outer tunic of the bud is in free communication with the blood-sinus of the parent, by means of the interval between the endostylic cone and the wall of the peduncle ; and, in fact, this interspace is itself the foundation of all the blood- sinuses of the bud. This is as much as can be clearly made out from the inspection of side views of buds in this and earlier stages ; but much additional information is to be gained from other views of similar buds. Fig. 26 exemplifies the appearances yielded by a bud about t9tIi<^ of an inch in diameter, when seen from above. The transverse section of the wide sac-like prolongation of the endostyle thus presented to the eye is four-sided ; the lateral and posterior walls are concave, while the middle of the anterior wall is produced into a sort of fold ; so that the contour of the sac may be compared to that of a crown. Masses of indifferent tissue fill the interspace betAveen the concavity of the lateral wall of the sac and the sides of the external tunic, while the in- terval between the posterior wall and the hinder part of the external tunic is nearly filled 214 PROFESSOR HUXLEY ON THE ANATOMY up by the young ovisac. Kg. 27 represents an advance upon this condition, — tlie two principal changes to he noted being, first, the conversion of the lateral masses of indif- ferent tissue into hollow oval bodies containing a very small cavity, and, secondly, the prolongation of the posterior cornua of the alimentary sac. It is obvioiis that if either of these buds were viewed sideways, the middle fold of the anterior face of the sac would appear like an anterior thickening, while the posterior pro- longed cornua would simulate a posterior enlargement, and the whole would closely re- semble fig. 19 ; and it now becomes important to prove, by the study of more advanced stages, the nature of the sac, of its anterior median fold, of its posterior cornua, and after- wards of the lateral sacs. Fig. 28 represents a larger bud, which presents more of its anterior, than of its upper, aspect to the eye. The anterior fold, consequently, is repre- sented by only two dark streaks with a clear interval. The posterior cornua are more elongated than before. There is a rounded opake spot at the anterior end of the bud. In fig. 29, the dark streaks have become the middle bands of the endostyle; the anterior spot is obviously the nascent oral aperture ; whUe the posterior cornua have become sepa- rated from the rest of the cavity of the sac, communicating with it by only a small aper- ture. The cornua are, at present, of equal size ; but in fig. 30 the left cornu exhibits a trihedral dilatation, which is obviously the commencement of the stomach, while the right cornu and the part which joins it with the left have become, the latter the foundation of the arch of the intestine, and the former of the rectum. There can be no doubt, therefore, that the greater part of the sac into which the prolongation of the endostyle is developed becomes what would in other animals be called the pharynx, namely, that portion of the alimentary canal which lies between the oesophagus and the mouth*. Side views, such as those given in figs. 20 & 21, of buds, in similar stages of develop- ment, are equally instructive. The attached apex of the sac is seen to become the posterior ■end, or cone, of the endostyle. The small size of the gastro-intestinal, in proportion to the pharyngeal portion of the alimentary tract, and the free communication of the latter with the prolongation of the endostylic cone of the parent, which traverses the peduncle, are clearly seen ; and I entertain no doubt that, by means of the last-mentioned communi- cation, the cavity of the pharynx (or, as we shaU see it becomes, that of the branchial sac of the bud) is placed in communication with that of the parent. In favourable lateral views of buds in this stage, it is easily made out that the wall of the pharynx unites with the external tunic at its anterior end, and here gives rise to the oral aperture, whose tentacular fringe is only subsequently developed. Having traced the fate of the sac-like dilatation of the prolonged portion of the endostyle thus far, I will now direct the attention of the reader to the coincident progress of the lateral sacs, or oval hollow bodies, as I previously termed them. In fig. 27 they are very small and thick -walled ; in fig. 28 then' cavity is larger, their walls are proportionally thinner, * The development of the Ascidian pharynx, as traced out by Krohu m Phallusia, and by myself in the present memoir, appears to me to aiford ample demonstration of the justice of Milne-Edwards' s view, that the branchial sac of the Ascidian is the homologue of the pharynx of the Polyzoon. It is due to my friend Prof. Allman, however, to refer to what he has said in support of another theory, in his able essay " On the Homology of the Organs of the Tvnicata and the Pobjzoa" Transactions of the Royal Irish Academy, vol. xxii. AND DEVELOPMENT OF PYROSOMA. 215 and the sacs tliemselves are both absolutely and relatively larger. In fig. 29 they are very much larger and thinner, and their relations to other organs are especially worthy of attention. The outer layer of each is applied to the outer tunic of its side, leaving a small interspace, which communicates freely with the great posterior sinus, in wliich the intes- tine and genitalia are disposed, and with the anterior sinuses which lie between the pharyngeal wall and the external tunic. This interspace is, in fact, the parietal sinus. The internal layer, continuous with the outer anteriorly and posteriorly, but separated from it by a wide chamber for the reot of its length, is applied against the wall of the pharynx for four-fifths of the extent of the latter, and then coats the lateral portions of the gastro-intestinal tract, forming the antero-lateral boundary of the great posterior sinus. The space between the wall of the pharynx and the inner layer of the sac com- municates anteriorly with the anterior sinuses, posteriorly with the posterior sinuses, and it is interrupted at several points by the union of the pharynx and inner layer with one another. It represents the system of branchial sinuses. In side views it is not easy to make out the boundaries of the lateral sacs ; but it is most important to observe that, as has been already mentioned, in the middle of the lateral face of the pharynx, and, therefore, also in the middle of the lateral face of the inner wall of the sac, a series of opake rings with clearer centres, the rudiments of the branchial stig- mata, make their appearance (figs. 19 and 20). These correspond with the points of union of the pharynx and the inner wall of the sac. They are, at first, small, round, and very indistinct, but, by degrees, they elongate in a direction perpendicular to the long axis of the pharynx, and their real nature becomes apparent. Hence it is clear that these stigmata must eventually open into the lateral sacs, as indeed they may be seen to do in such buds as that represented in fig. 30 ; and hence also it follows, that the lateral sacs are the rudiments of the lateral atria. At first the lateral atria appear to be perfectly distuict from one another, and no atrial aperture is discernible. In buds such as that represented in fig. 29, again, they do not extend, posteriorly, further than the sides of the alimentary canal ; bixt in more advanced buds (fig. 30) they are produced backwards on each side until they pass beyond the level of the posterior margin of the stomach, so that they now Constitute the entire lateral boimdaries of the great posterior sinus. The longitudinal section (fig. 21) of a somewhat smaller bud than that represented in fig. 30 shows, however, that, in this condition, the atria are no longer distinct, but are united together below the stomach by a comparatively narrow and short canal {p), which is the mid-atrium. I have not traced out all the details of the process of coalescence of the lateral atria ; but I suppose that each branchio-parietal portion of the atrium, at first a distinct sac, is prolonged downwards and inwards, under the stomach, and that the opposed walls of the prolongation become applied to one another, coalesce, and then liecome perforated. At any rate, the mid-atrium is now surrounded by a membranous wall, continuous on all sides with the lining of the lateral atria, and ajiplied superiorly and anteriorly against the stomach and oesophagus, posteriorly and inferiorly against the external tunic, but not touching either of these parts, except for a small space on the floor of its chamber, where it becomes united with the external tunic to allow of the formation of the atrial aperture. 216 PROFESSOR HUXLEY ON THE ANATOMY In the present bud (fig. 21) this aperture is situated on the neural side of the body, in front of the posterior end, which is chiefly occupied by the genitalia; but as develop- ment goes on, the mid-atrium increases disproportionately, and encroaches upon the other organs, upwards and forwai'ds, in such a manner that its anterior wall invests the whole posterior and lateral faces of the gastro-intestinal division of the alimentary canal ; while its roof (to speak metaphorically) thrusts the genitalia altogether into the hiemal region of the body, and its posterior and inferior walls, extending backwards, carry the external tunic with them, and eventually cause the atrial aperture to take its place at that extremity of the body which is directly opposed to the mouth, and far behind the genitalia (see figs. 22-25). The communicating apertures between the mid-atrium and the lateral atria increase in size pari passu, with the growth of the parts ; and hence, in the fully formed ascidiozooid, the gastro-intestinal division of the alimentary canal is enclosed in a sort of vertical mesen- tery (formed by the anterior wall of the mid-atrium in the^ middle line, and the internal Avail of the lateral atria at the sides), whose layers are continued, on either hand, into the outer wall of the branchial sac. At the anterior boundary of the branchial sac they are reflected into the outer or jiarietal layer of the lateral atrium. The facts which I have detailed * are exceedingly important for the comprehension of Ascidian structure in general. Erom its mode of development, it is perfectly obvious that the inner wall of the branchial sac of Fyrosoma is not composed of tentacles which have coalesced, but that it is, originally, a simple imperforate dilatation of the pharyngeal por- tion of the alimentary canal. The development of the atrium adds a second or outer wall to this dilatation ; and when, by the formation of this double wall, the branchial sac is constituted, the stigmata make their appearance in its parietes — the atrial and the pha- ryngeal Avails becoming united around the margins of each stigma. When a bud has attained a length of between -xVth and -g^-th of an inch, the narrow neck connecting it witli the peduncle is obliterated, and it lies free in the general test of the parent ascidiarium. It next elongates until its oral and atrial apertures are placed in connexion with the exterior and the cloaca respectively (the latter connexion appear- ing to Idc eff'ected first), and then it increases in depth untU it acqtdres the appearance of the adult. Before it is detached, however, the portion of the peduncle nearest it enlarges and assumes the shape of a new bud ; so that the proximal end of the peduncle now passes into a small bud with whose apex a larger one is connected (fig. 22). And I suspect that this process is repeated as long as there is any reserve of generative blas- tema in the parental organism. I have, hoAvever, never actually seen more than tAvo buds thus connected together. As the buds are all develojDcd from the htemal region of the pre-existing ascidiozooids, it foUoAvs that the new ascidiozooids formed by gemmation must at fijst be thrust among the old ones, toAvards the apical end of the ascidiarium. So much in elucidation of the mode in which the buds attain the form and general arrangement of organs characteristic of the adult. I now proceed to speak of such among the minor changes which these organs undergo as call for particular remark. * The accurate Krohn, in his account of the development of Phallusia (Miiller's " Archiv," 1852), was the first to note the separate origin and subsequent confluence of the lateral atria. In this genus, however, each lateral atrium has, at first, a distinct external aperture. AND DEVELOPMENT OF PYROSOMA. 217 Of tlie outer tunic all that requires to be said is, that it becomes relatively thinner as development goes on. In buds which are situated within a certain distance of the open end of the ascidiarium, and which have attained a length of iVth of an inch (fig. 24), the outer tunic of the neural wall of the atrium is raised into a slight rounded projection (r =), and in older buds (fig. 25) this gradually elongates, and extending towards the open end of the ascidiarium, and finally into the lip of the cloacal aperture, becomes converted into one of the stolons of the test. The atrial muscular bands are visible in buds not more than -^oth of an inch in length (fig. 23) ; the pharyngeal muscular bands, only in more advanced zooids. The tentacular fringe appears first as an inward thickening of the parietes of the mouth. The hgemal tentacle is markedly the longer, even in such buds as that represented in fig. 24. The ganglion is discernible in buds T^btb of an inch long (fig. 20) as an opake oval mass situated between the peduncle and the oral end of the bud, and very much larger in proportion to the rest of the organism than afterwards. The ciliated sac appears as a short cfecal diverticulum of the pharyngeal cavity, connected with the anterior and liEemal side of the ganglion. A most curious structure is visible in buds T^oth of an inch long, and remains obvious until they have attained a length of aVth of an inch or thereabouts. For want of a better name, I will term this the ' diapharyngeal band.' In the section, fig. 21, its upper part is visible, passing obliquely downwards and backwards from between the two middle bands of the endostyle ; while, in figs. 22 & 23, its lower extremity is seen to end in the pha- rynx, immediately over the posterior moiety of the ganglion. The diapharyngeal band is hollow, and effects a communication between the htemal and neural sinuses ; and if, as is possible, the heart of the bud has at this period but little functional activity, the ex- istence of this direct channel may facilitate the circulation of the blood. However this may be, this structure becomes longer and thinner as the development of the bud advances ; and all that remains of it, in buds -jVth of an inch long (fig. 24), is a small tubercle which lies over the posterior part of the ganglion. Eventually even this disappears. I have already spoken of the origin of the branchial stigmata. Fig. 21 represents an accidental, but very fortunate, longitudinal section of a bud -7-oth of an inch long ; the razor having passed rather to the right of the middle line above, rather to the left below. As it is seen from the right side, the inner surface of the left wall of the branchial sac is exposed to view. Of the eleven stigmata, those in the middle are the longest and most oval, those at the two ends of the series shortest and most rounded. They look clear in the centre, but on careful examination they are seen to be closed, the sheet of indifferent tissue which forms the innermost wall of the pharynx being continued over them. I am strongly inclined to think that it is this sheet of indifferent tissue which gives rise to the longitudinal branchial bars, for in more advanced buds (fig. 22), in which the median stigmata have undergone much elongation, the same layer is continued over their htemal and neural ends, while it has disappeared in the interval, except along three longitudinal lines, where it evidently forms the foundatiom of as many longitudinal branchial bars. In the more advanced stages, new stigmata are added to the anterior and posterior ends VOL. XXIII. 2 G 218 PROFESSOR HUXLEY ON THE ANATOMY of the series. Those ah-eady formed elongate, and new longitudinal bars are added, until the walls of the branchial sac assume their perfect form*. As I haye explained, the sac-like alimentary tract originally ends in a conical point at that extremity which is opposite its oral end ; and this cone is connected with the external tunic. In subsequent stages the cone remains distinct, being directed at an obtuse angle to the rest of the liEemal wall of the pharynx, while the cellular bands which eventually render the endostyle so conspicuous, cease at its base. It, at first, communicates by its Avidely open base with the pharyngeal or branchial cavity ; but as development proceeds, it becomes narrower, in proportion to the endostyle, and at length is represented by that slender backward prolongation of the endostyle or ' endostylic cone ' described at the commencement of this memoir and represented in fig. 14. The languets do not appear till development has advanced a long way ; in fact, in the very young buds there is no room for them, as almost all the space between the place of the commencement of the oesophagus and the place of the ganglion, is occupied hj the aperture of communication between the prolongation of the endostylic cone and the pharynx. As growth proceeds, the distance between the ganglion and the oesophageal aperture gradu- ally increases, both alxsolutely and relatively, and in buds aVth of an inch long, one or two small tubercles are visible, projecting from the hypopharyngeal band, between the oeso- phageal aperture and that of the canal which traverses the prolongation of the endostylic cone. These gradually increase in number, elongate, and assume their adult shape and size (figs. 24, 25). The figures will sufficiently explain the further changes of form undergone by the gastro-intestinal portion of the alimentary canal. The hepatic tubular system makes its appearance in such buds as that represented in fig. 22, as a minute diverticulum of the stomach, which elongates, applies itself to the intestine and ramifies over it. Krohn {l. c. p. 331) saw it originate in a similar manner in Thallusia. The heart, similar to that of the adult in form and textiu'e, is distinctly discernible in buds not more than -^(^ilx of an inch long, attached, in its ordinary position, to the wall of the pharynx, just in front of the bend of the intestine, between it and the endostylic cone. I have not been al)le to trace out the first condition of this organ and the changes which it luidergoes in acquiring the state now described. The renal organs are plainly visible in buds not more than -g^oth of ^ii iiich long as aggregations of clear, round, almost colourless corpuscles, between the atrial and the outer tunic. In describing the first stage of the bud (fig. 14), I have spoken of a thin layer of in- difPerent tissue which passed from the end of the endostylic cone, or prolongation, into the generative blastema. In more advanced stages, this tissue forms a sort of hood over the * Krohn {I. c. pp. 324 and .327) states that the stigmata of the embryo Phalhtsia make their appearance as round apertures ; but he affirms that new ones are added, not only in front and behind, but in the neural and haemal sides of the first formed series. AND DEVELOPMENT OF PYROSOMA. 219 end of the saccular rudiment of the alimentary tract (figs. 15 and 16), and seems to be the means of connecting the end of that sac with the external tunic. After a time, however, a clear space appears around the apex of the sac, and separates this connecting mass from the rest, which now, consequently, appears as a broad zone («) surrounding the sac below its apex, but above the uppermost of the branchial stigmata (fig. 19). This zone remains as a broad, thickish girdle of indifferent tissue, closely con- nected with the outer tunic externally and in front, and with the generative blastema behind, in buds of -T^^tli — -^oth of an inch long (figs. 21 and 22) ; but in larger zooids its tissue has undergone a great change, and it has become a transparent mass, through which ramified corpuscles, like connective-tissue corpuscles, appear scattered (fig. 23). In this condition it is exactly analogous to the structure termed elgeoblast in the SalpcB by Krohn. Its bulk is now equal to a fifth or a sixth of that of the entire bud, but in subse- quent stages (figs. 24, 25) it diminishes both absolutely and relatively in size and eventu- ally it disappears. In buds sijth of an inch in diameter, the generative blastema remains in its primitive condition, except that it and the ovisac it contains, have increased in size. Its anterior pointed end is closely juxtaposed to the endostylic cone. In the zooid represented in fig. 23, which measured g-Q-th of an inch in length, the generative blastema has become divided into two parts, the smaller of which remains in close apposition to the endo- stylic cone, while the larger, retaining its connexion with the posterior and upper wall of the mid-atrium, becomes widely separated from the other. The interval between the two is occupied by the el^oblast. Even before the separation has taken place, the larger portion has become distinctly differentiated into two parts, the ovisac, on the left, sepa- rating itself from a rounded mass of indifferent tissue, on the right. This last is the rudiment of the testis. Erom rounded, it becomes pyriform, the narrowed neck of the pear remaining in connexion with the atrial wall, and eventually becoming metamorphosed into the vas deferens, while the broad end increases in size, and is directed more forwards as well as upwards. In a bud iVth of an inch long (fig. 24) the testis measures -g-s o^h of an inch in length, while its broad end is above -s^o^h of an inch thick. The apex of the vas deferens already pushes a little eminence of the atrial tunic before it. In a young ascidiozooid, somewhat more advanced than that represented in fig. 25, the vas deferens is sioth of an inch in length, and is of nearly even diameter throughout, except at its vipper end, where it is slightly dilated and plainly hollow. It is connected with the posterior part of the terminal enlargement, which is nearly ^^oth of an inch thick, and is divided into three short lobes, each about s^^njth of an inch thick. Like the previously existing pyriform enlargement, these rudimentary cpeca are solid masses of indifferent tissue. Traces of a distinct membrana propria are discernible around each CEecum. In still larger ascidiozooids the number of cseca increases, and the whole organ becomes larger, until it assumes its adult form ; and it is only \\\\en nearly in this con- dition, that spermatozoa are visible in the vas deferens and the adjacent parts of the caeca. The development of the ovisac will be described below. At first both the testis and 2g2 220 PROFESSOR HUXLEY ON THE ANATOMY the ovisac have ample room within the sinus of the zooid in which they are lodged ; but as they increase in size, the duct of the ovisac extending towards the neural side and for- wards, and the duct of the testis extending toAvards the neural side and backwards, push the atrial tunic before them, so that their openings are eventually situated on slight papillary elevations. The principal portions of the two organs, on the other hand, con- sisting of the sac of the ovisac and the ca^ca of the testis, as they enlarge, pass into cham- bers in the test, which are formed for them by the recession of the outer tunic, and whose cavities, consequently, communicate freely with the haemal blood-sinuses. With respect to that part of the generative blastema which remains in connexion with the endostylic cone, one of its endoplasts or nuclei soon acquires a larger size, and l)ecomes surrounded by a clear space, thus giving rise to a new germinal vesicle and spot, round which will eventually be formed the solitary ovum and ovisac of a new bud, developed from the zooid, Avhosc origin has just been traced, in exactly the same way as itself has arisen. Thus, if we start with a single ascidiozooid, it may give rise, to all appearance, to an indefinite succession of buds, by successive enlargements and detachments of the end of the peduncle of the first ; and each bud thus developed carries within itself, in its generative blastema and endostylic cone, provision for an indefinite succession of other buds. It must be recollected, however, that while the tissue of the rudiments of the alimentary and reproductive systems of each bud is directly descended, with compara- tively little alteration, from the blastoderm of the embryo Pyrosoma, yet this tissue cannot be said to be embryonic ; the tissue of the endostylic cone being considerably cUff'erentiated, while the outer tunic of each bud is derived from the still more modified outer tunic of the parent ascidiozooid. These facts, therefore, lend no countenance to the doctrine, whose fallacy I have demonstrated in a previous memoir, that budding depends on a retention of the primitive tissue of the germ in any part. § -i. The Gamogenesis, or Sexual Bevelopment, o/Pyrosoma giganteum (Plate XXXI.). It will conduce to intelligibility, if the somewhat complex history of this process is divided into stages, characterized partly by the size of the ovisac, partly by its structural characters. I shall describe, under each stage, a specimen or specimens, illustrating the peculiar features of that stage, but it will be understood that insensible gradations are observable between the diff'erent stages ; and, in order that the whole process of develop- ment may be viewed continiiously, it will be advisable to consider, as the first stage, that condition of the ovisac in which it is first recognizable as a completely distinct organ, a condition which it attains, as I have ali*eady stated, in buds such as that figured in PL XXX. fig. 23. First Stage. Ocisacs less than -g\-(yth of an i?ich in diameter and loitliout ducts. Eig. 1, PI. XXXI., represents an ovisac measuring 4 loth of an inch in diameter. It is ellipsoidal in form, and nowhere presents any prolongation whicli can be regarded as even the rudiment of a duct. The wall of the ovisac is comparatively thick, and obscurely AND DEVELOPMENT OF PYROSOMA. 221 cellular in structure, but it is devoid of any structureless investment or membrana propria. The contained ovum consists of a solid-looking, well-defined germinal spot oVb otb of an inch in diameter, occupying the centre of a germinal vesicle -sw^h of an inch in diameter, with a thin but well-defined wall, and perfectly clear contents. The yeUi is represented by a small zone of structureless, yellowish substance, which invests the germinal vesicle, and, on the one hand, passes into the waU of the ovisac, while, on the other, it is separated from that wall by a narrow clear space. Second Stage. Ovisacs less than ^ou^^' of an inch in diameter and unimpregnated. The ovisac represented in fig. 2. exemplifies this condition very well. It has a dia- meter of s^o^h of an inch. Its form is spheroidal, and it is produced on the side towards the atrial wall of the blood sinus in which it lies, into a short subcylindrical diverticulmn, which is directed forwards, and slightly towards the neural side of the ascidiozooid in which it lies. This diverticulum, or rudimentary duct, is 4^tli of an inch in length, and its slightly narrowed anterior extremity passes into the atrial tunic. At its opposite end, where it becomes continuous with the ovisac, it measures grotli of an inch in dia- meter. At this extremity, the cavity of the duct is in free communication with that of the ovisac, but at a little more than half way towards the opposite end, or in other words towards the atrium, the cavity ceases, the termination of the duct appearing to be a solid cellular mass. In this condition, therefore, there would appear to be no communication between the interior of the ovisac and the atrial cavity. The wall of the ovisac exhibits no distinct membrana propria, but is composed of a single layer of flattened corpuscles, about 5-^50-th of an inch in diameter, imbedded in, and connected together by, a structureless substance. The wall of the duct is similarly com- posed, but its hsemal is much thicker than its neural wall. In the cavity of the duct nothing save a clear fluid is discernible, and the same fluid seems to fill the interval observable on one side, between the wall of the ovisac and tlie ovum. The latter consists of a very finely granular, spheroidal vitelline mass ^^tti of an inch in diameter, within which lies a germinal vesicle (e^th) with perfectly clear contents, inclosing an opake, spheroidal, germinal spot (rttVoth). The yelk is in close contact with the inner side of the anterior wall of the ovisac — the germinal vesicle is close to its sur- face at the same point, and the germinal spot is applied to the inner surface of the ante- rior wall of the vesicle, so that it is as near as possible to the wall of the ovisac. I have not been able to discover a trace of a vitelline membrane in ova in this stage. It may be doubtful whether the space between the wall of the ovisac and the ovum is a natural or an artificial product. My observations upon the ovisacs of a fresh Fijro- soma (PhU. Trans. 1851) lead me to adopt the latter hypothesis. Third Stage. Ovisacs under jj^jyth of cm inch in diameter and in process of impregnation. (PI. XXX. figs. 3, 4, and 8*.) Of the two specimens in this stage which I have figured, the larger (fig. 3) is rather the less advanced. Its duet is longer than the diameter of the ovisac, and is not only 222 PROFESSOR HUXLEY ON THE ANATOMY hollow throughout its whole length, but, at its anterior end, opens into the atrium, with which, therefore, its cavity is in free communication. ■\¥hether a similar con- nexion obtains between the cavity of the ovisac and that of the duct, or not, I cannot certainly say. A marked constriction is generally observable at the point of junction between the duct and the ovisac, corresponding to an inwardly projecting lip, which greatly narroAvs the apparent aperture of communication (fig. 4) ; and in some cases, the cellular wall of the lip appeared to have grown out, in such a manner as stUl further to diminish that aperture ; but I have been unable, in any one instance, fully to assiu'c myself of the closure of the passage. If, however, as I have reason to believe, the vitellus, in the fresh state, completely fills the ovisac, the aperture will be effectually closed by its means. The ovum in the ovisac represented in fig. 3, measures Tg^ii'l of an inch in diameter, that in fig. 4, -jloth. In each case the vitellus is somewhat more opake than in the pre- vious stages ; but, as before, I have been unable to find any vitelline membrane, even when, as in fig. 8*, the ovum has been turned out of the ovisac. But I have constantly observed that while the greater part of the circumference of the yelk exhibits a well-defined dark contour, that portion which is away from the side on which the germinal vesicle lies, has a faint, hazy outline, as if it were undergoing solution. This appearance is well shoAvn in figs. 4 and 8 *, and it is worth recollecting in connexion with the subsequent fate of the yelk. In both the ova represented in figs. 3 and 4, the germinal vesicle measures 4ioth of an inch in diameter, and its contents are, as before, perfectly clear. A change of figure has accompanied its increase in size, for it is now oval, its long diameter being more or less perpendicular to the direction of the duct. Furthermore, it is situated at the surface of the ovum, at a point close to, but on one side of, the aperture of the duct ; and that face which is nearest the surface of the vitellus is not unfrequently flattened. The germinal spot retains its previous size (xiiVo") ^^^ appearance. One or two minute clear spaces are to be seen in it, occasionally, in this and in other stages, but I suspect they are accidental. In the specimen represented in fig. 3, the duct appears to contain only a clear fluid, as before, except that a few indistinct striae are visible towards its upper end. One would hardly know what to make of them, if it were not for the circumstance, that a bundle of minute filaments, a few of which would readily give rise to the striation in question, hangs from the mouth of the duct. The filaments are sticking in its atrial aperture by one end, while the remainder of their length protrudes. The filaments are exceedingly delicate, and so entangled that their individual dimen- sions cannot be estimated. The whole bundle, however, measures about 4 3 o'th of an inch in length. The ends of the filaments inserted into the aperture are thickened, and more or less rod-like. In a slightly larger ovisac (fig. 4) no such filaments are visible about the mouth of the duct, but its upper dilated end contains a conical plug, composed of pre- cisely similar bodies, and a similar plug occupied a corresponding position in every other ovisac, in this stage, which I have examined. If the ovisac is not disposed in such a AND DEVELOPMENT OF PYROSOMA. 223 manner, that the plane of the constricted junction between the duct and the ovisac is perpendicular to the stage of the microscope, so as to afford a true profile view, the broad end of the plug will appear to be in direct contact witli the vitellus, close to the germinal vesicle. But I have never met with any such absolute contact in a true profile view. On the contrary, in such a view, the end of the plug appears to be jammed in the upper aperture of the duct, and there is a small interval between it and the surface of the vitellus. But it must be remembered that (as I have already pointed out) in the fresh state, the viteUus, in aU probability, occupies the whole cavity of the ovisac, and itself stops the upper aperture of the duct ; and, if this be the case, it is exceedingly likely that the slight separation between the yelk and the plug of filaments is a post-mortem change. At any rate no filaments are ever discoverable in the cavity of the ovisac, and as I have been able to find no complete diaphragm across the upper aperture of the duct, there seems to be no reason for their absence, unless we suppose that the vitellus itself bars their entrance. But in this case the plug and the viteUus must come into direct' contact. This point is of great importance, because there can be no doubt that the filaments in question are spermatozoa. The ' plug,' and the contents of the vas deferens of the testis are precisely similar in appearance. The plug is not visible before the atrial end of the duct is open, thus providing free access for spermatozoa floating in the atrimn. As there are no cUia on the inner surface of the duct, it seems impossible to account for the pre- sence of the dense mass of filaments within it, except on the supposition that they have an inherent propulsive power ; and the only free, filamentous bodies possessed of such a power we know of, in the animal economy, are spermatozoa. Furthermore, in my former memoir on Fyrosoma {I. c. p. 584), I have recorded the following observation : — " In young specimens, when the ovum is small and the yelk pale, this gubernaculum [the duct] frequently appears to be solid ; but in fully grown specimens, when the ovum [ovisac] has its full size, and the yelk is dark and granulous, it presents the appearance of a wide tube, especially at its upper part. And here, there was frequently an appear- ance of dark strise and moving granules, prompting the belief that spermatozoa had travelled thus far. In one instance the sac of the ovum was empty, and the guber- naculum or duct widely distended ; the appearance of spermatozoa in the duct was here very strong. (Fig. 5.)" I entertain no doubt, then, that the specimens described exhibit the process of impreg- nation in Pyrosoma ; that the spermatozoa make their way up the duct and come into contact with the surface of the yelk. Whether that reciprocal action of the spermatozoon and the ovum, which constitutes the essence of fecundation, takes place immediately on the occurrence of this contact, I cannot pretend to say v/ith certainty, but I doubt it ; for, as will be seen, though very remarkable changes take place shortly after impregnation, they are not those which in other animals follow upon fecundation. It is not a little singular that, in consequence of the immature condition of the testis of zooids whose ovisacs are in the stage under consideration (^Fi/rosoma resembling Salpa 224 PROFESSOR HUXLEY ON THE ANATOMY in the much more rapid advance towards maturity of the female, than of the male, organ, in each zooid), the spermatozoa which effect impregnation must be derived from another zooid if not from another ascidiarium. The latter alternative is not S(j improbable as it looks at first sight, if we consider that a current constantly sets through the body of each zooid from the oral to the atrial aperture, and so out at the cloaca. Hence, the spermatozoa which are poured by the vas defex-ens of any given zooid, in which the testis has attained its full development, into its atrium, must be almost immediately carried into the cloaca ; and as a powerful current is setting into the cloaca from every other zooid, it does not seem possible that the spermatozoa should make their way into any one of these zooids against it. But on the other hand, as the Tyrosomata live in great troops, the spermatozoa cast out of the cloaca of any one Pyrosoma may very readily be taken in by the oral aperture of aziother, and passing with the current through the branchial stigmata into the atrium, may easily reach the aperture of the oviduct. If this reasoning is valid, Pyrosoyna affords a curious illustration of Mr. Darwin's doc- rine of the rarity of self-fertilization even among hermaphrodite animals. Fourth Stage. Ovisacs from xo^th to 4^th of an inch in diameter, in which the yelk disappears and the germinal vesicle becomes fixed to the icall of the ovisac. Eigs. 5— 8«. Figure 5 represents an ovisac -gVli of an inch in diameter, and fig. 6 another of y^th of an inch. The first thing to be observed about these ovisacs is, that they have increased in dimensions disproportionately to their ducts ; for while, in the preceding stage, the duct is longer than the transverse diameter of the ovisac, in the present stage, it, at first, hardly equals, and subsequently remains much shorter than, that diameter. The duct, in fact, does not attain a greater length than tI oth of an inch, and in the larger examples of this stage it appears shrunken and withered. The spermatozoa, however, are always visible in its upper dilated end (fig. 6 b), but sometimes they no longer form a distinct bundle, but appear scattered, and then their rod-like heads are very distinct. In the wall of the ovisac and of the duct, a differentiation has taken place into an outer structureless membrana propria, and an inner epithelial layer. The latter is pale, the cor- puscles, which lie in the wall of the ovisac in this as in earlier stages, appearing to be thin- ner and separated by wider clear interspaces. That change which arrests the attention of the observer most forcibly, however, is the entire absence in the present, as in all subsequent stages, of that vitelline mass which is so conspicuous in less advanced ovisacs. As a con- sequence of this disappearance of the yelk, the germinal vesicle lies apparently free and bare, in contact with one wall of the ovisac. There is not the slightest difficulty in observing these facts, nor the least ambiguity about the microscopical appearances ; but the. circum- stances appeared so unprecedented, that, when I first became acquainted with them, I mis- trusted the obvious interpretation of those appearances. However, I found, not only that the contour of the yelk contained in the smaller ovisacs was perfectly Avell defined, but that, by careful manipulation with needles, under the simple microscope, I could turn out the ovum entire, the vitellus being so firm and consistent as to retain its form (fig. 8*) ; AND DEVELOPMENT OF PYROSOMA. 225 aud yet I could neither observe tlie smallest trace of the yelk in entire ovisacs in this stage, nor, however carefully I opened them, discover any trace of yelk within them. I found, furthermore, not only that, hy a little pains, I could open the ovisac so as to view the germinal vesicle from Avithin (figs. 6 and 7), but that I could evert it, turn it in all directions, and even detach it entirely : and when I discovered, by these means, not merely that no vitellus surrounds the germinal vesicle in this stage, but that it is enclosed and held in place by something which is assuredly not vitellus, I was forced back into my original conclusion, that in this stage the vitellus, as such, has disappeared. There is, however, one suggestion which deserves careful consideration. It may be said, that what I have termed the germinal vesicle (represented separately in figs. 6a&S) is in fact the ovum. To meet this objection, I would beg the reader to compare figs. 8 and 8* ; the former of which represents the body whose nature is in dispute, and the latter an ovum which has not reached its full size, the two figures preserving the true relative proportions of the originals. It is at once obvious that the circular solid-looking cor- puscle, situated towards the upper end of fig. 8, is identical in all essential respects with the germinal spot of fig. 8*, the only difference being that it is slightly larger, measuring Y^g^th of an inch, while the germinal spot of the entire ovum is about xg-oo- But if this corpuscle represent the germinal spot, then the only structure Avhich corresponds with the wall of the germinal vesicle in fig. 8* is the structureless, oval, membranous sac, wrinkled on one side, wliich encloses the germinal spot in fig. 8. This sac, it must be admitted, differs a good deal from the germinal vesicle of fig. 8*, not only in size, but in form and in contents. In the first place, it is much larger, measuring siotfi of an inch in length, while the germinal vesicle of fig. 8* is only -5^0 ; next, it is oval and irregular on one side ; and thirdly (and this is the most important difference), it contains a homoge- neous yellowish deposit, which is especially accumulated around the germinal spot, but is absent under the wrinkled moiety of the vesicle. AU doubts as to the identification of the body (fig. 8) with the germinal vesicle and spot of fig. 8*, however, vanish when a series of ovisacs, intermediate in size between that which yielded the ovum, fig. 8*, and that represented in fig. 6, are studied. Thus in fig. 4s, the unquestionable germinal vesicle is oval, and its long diameter amounts to 5-aotli of an inch ; while in the ovisac represented in fig. 5, in which tlie yelk has disappeared, the body in dispute is precisely similar to the germinal vesicle of fig. 4, except that it is a little more flattened and a little longer (^^th). Its contents are quite clear, and its wall is but very slightly corrugated. But no one can question the identity of this body with that represented in place in fig. 6, and separately magnified in fig. 6a, which has a long diameter of xTT^fi of an inch, whose walls are much wrinkled, and which contains a dense yellow deposit. I have no hesitation then in regarding the body, fig. 8, which agrees in all essential respects with that represented in fig. 6a, as the germinal vesicle of the primitive ovum, stripped of its vitellus. Though devoid of any vitelline investment, however, the germinal vesicle has been neither free nor bare, in any ovisac which I have examined. It is always seen to occupy one spot of the inner face of the ovisac, a little behind and to the right of the upper VOL. XXIII. 2 H 226 PROFESSOR HUXLEY ON THE ANATOMY aperture of the duct ; and when the ovisac is opened, the germinal vesicle is found to adhere to this point with considerable tenacity. It is, in fact, held in place by a conti- nuation of the epithelial lining, which, lies between it and the cavity of the ovisac — the germinal vesicle being now situated between the epithelium and the membrana propria, so that while its outer face is covered by the latter its inner face is invested by the former. All this will be rendered easily intelligible by examining the profile views (fig. 6 and 6«), and the view from within (fig. 7), of the germinal vesicle in situ. But it has been seen that the ovum, containing the germinal vesicle, originally lay inside the wall of the ovisac, which has become metamorphosed into the epithelium, and hence it follows that the germinal vesicle, after losing its yelk, must pass through the epithelium of the ovisac. It will be recollected that the mammalian ovum becomes simi- larly related to the epithelium of the Graafian follicle, and that the germinal vesicle of the bird's egg in lilie manner passes into and through the peripheral layer of its yelk. Throughout the present stage there is not the least difficulty in observing the germinal vesicle and its spot in the uninjured ovisac. The spot, in fact, is particularly well-defined, and immediately strikes the eye when even a low magnifying power is used. But, with such a power (say 200 diameters), it is easy to fall into error as to the shape of the germinal vesicle. It constantly appears to be hemispherical, the truncated side being that which is turned away from the upper aperture of the duct. This appearance arises from the fact that, with such a power, one sees the contents and not the wall of the germinal vesicle, and as the yellow deposit fills only that moiety which lies nearer the upper aperture of the duet, it appears like a semicircular cake. Under a higher power (500 diameters) the wrinkled* membrane of the other moiety of the vesicle is always readily discernible. Fifth Stage. Ovisacs between -^th and-g^th of cm inch in diameter, in lohich the germinal spot disappears, and a number of minute granules take its ptlace. The germinal vesicle, held in place on the neural wall of the ovisac, immediately behind and rather to the right of the vipper aperture of the duct, in the manner which has just been described, next undergoes changes of very great interest and importance. I have devoted a very great deal of time and patience to the analysis of these changes, but it is only recently that I have felt satisfied with the results of my investigations ; and I must warn any one who is disposed to repeat these observations, that while everything which I have described up to the present moment may be demonstrated with the utmost readiness in almost any thin vertical section of the ascidiarium, the conditions of the germinal vesicle which I am about to describe occur but very rarely, and require the aid of high powers of a thoroughly good microscope for their complete elucidation. Out of a vast number of preparations which I have made, at intervals, during the last twelve months, not more than eight or nine have exhibited the exact features of which I am about to give an account. The germinal vesicle represented in fig. 8«, Ijelongs, strictly speaking, to the preceding * The fact that the Pyrosoma ohserved by me had been preserved in spirit should always be recollected. It is highly possible that the wrinklings are artificial. AND DEVELOPMENT OF PYROSOMA. 227 stage. But it differs from tlie characteristic germinal vesicles of that stage, in that its spot has lost its solid, opake aspect and has apparently become a vesicle with a thin, sharply- defined wall, hut so pale that under a low power it would readily escape notice. Its diameter is ^9^0^^^ of an inch. The vesicle itself measures srotlij it is much flattened, and its contents are somewhat paler than before. I have described this germinal vesicle here, because I believe that it is in that condition which constitutes the transition from the typical form of the last to the typical form of the present stage. As this last is of very great importance, I will note down the appearances presented to me by several germinal vesicles which exhibit it. If I examine a slide, at present under my microscope, I observe, under a low power, in one place, an ovisac belonging to the fourth stage. The germinal vesicle, with its yellow contents, is very obvious, and the round, sharply-defined germinal spot strikes the eye at once. If I now move the slide a little way, I bring into view a large ovisac about ^^aiif^ of an inch in diameter. In this, it is only with difficulty that I can trace the outline of the germinal vesicle, and nothing is to be seen of the germinal spot. This intUstinctness of the germinal vesicle docs not arise from want of size or clear definition ; for, if I put on a high power, I find it to have a diameter of Tg^th of an inch, and its contour is perfectly well marked. The yellow deposit occupies about half its cavity as before, but it is paler ; and partly on this account, and partly by reason of a further change in the structure of the epithelium of the ovisac, the vesicle is less obvious than previously. Of the germinal spot not a trace is to be seen anywhere, although the vesicle and its contents are quite transparent. Whether the contents exhibit any new structure or not cannot certainly be made out, on account of the interference of the wall of the ovisac, through which the germinal vesicle is seen; In another ovisac in this stage, also about ^nd of an inch in diameter, the germinal vesicle, very similar to that first described, measures -g^th of an inch in length, and is half filled with the yellow deposit. No vestige of the germinal spot is to be seen, but, on that side of the contents which in earlier stages is occupied by the germinal spot, there are a number of minute, spheroidal clear granules, none exceeding jTo^o-g-th of an inch in diameter and arranged so as to form an elongated patch on the surface of the contents, the rest of which is quite free from such bodies. In another ovisac of about the same size the germinal vesicle is axo^b of an inch in diameter with pretty nearly half that thickness, and similar granules are observable upon the face of its contents, while there is nothing to be seen of a germinal spot. But the best example of this stage is that afforded by yet another ovisac a^th of an inch in diameter, whose germinal vesicle, -aioth of an inch in diameter, is represented in fig. 85. Here the contents can be searched through and through with the greatest ease ; but not the least trace of a germinal spot is discoverable, whUe the minute clear corpuscles loooo^h to g^cToth of an inch in diameter, scattered over the face of the contents, are exceedingly distinct. Whether they are free, or whether they are imbedded in any clear substance, I cannot say certainly, but I suspect the latter to be the case. Putting the facts observed in this stage together, we find, that in ovisacs between ^th and -^th of an inch in diameter, the germinal vesicle increases in size until it attains as much as ^ j-Q-th of an inch in long diameter ; and that the germinal spot, as such, entirely 2 u 2 228 PROFESSOR HUXLEY ON THE ANATOMY disappears. On the other hand, on that side of the contents towards the wrinkled part of the membrane of the germinal vesicle, a number of minute, pale, spheroidal corpuscles make their appearance and spread over the face of the contents. Considering that, as we have seen, the germinal spot becomes pale before it ceases to be visible, and bearing in mind that the power of subdivision is one of the most characteristic properties of the class of bodies to which the germinal spot belongs, I do not think it very hazardous to assume that the corpuscles in question result from the division of the germinal spot. In all the ovisacs of this size the epithelium has undergone a very remarkable change. Instead of the thin cellular lamella which has previously lined the interior of the ovisac, a transparent substance excavated by many large spheroidal cavities of various sizes (which when the ovisac is viewed by a low power give it the appearance of being filled with numerous clear vesicles) occupies its cavity. Sixth Stage. Ovisacs about -^oih of an inch in diameter, in lohlch the germinal vesicle has disappeared but a blastodermic membrane occupies its place. Figure 9 represents an ovisac of -^rst of an inch in diameter in situ. It will be obsei'ved that the duct is now very small in relation to the sac, and that the modified epithelium of the latter manifests the vesicular appearance characteristic of the later stages. The germ- inal vesicle is no longer to be seen, but, exactly in the position it ought to occupy, there is a patch of substance which, in profile (fig. 9), is obvious as a thick, darkish yeUow line, but viewed from within or from without (fig. 9 a) is only visible under a high power, in con- sequence of the excessive paleness and delicacy of its components. It is, in fact, a very thin membrane jGotli of an incli long and about half as wide, composed of a single layer of spheroidal, or more or less polygonal, corpuscles, each of which has an average diameter of iVoo^h of an inch, though some are smaller and some are larger. Every one of these contains in or near its centre a small, apparently vesicular, more strongly refracting and hence more conspicuous body, usually not more than g-g-ooth of an inch in diameter, but sometimes attaining to fully twice this diameter. That margin of this membrane which is turned towards the upper aperture of the duct (fig. 9 a) is tolerably sharply defined, and has an evenly curved contour, so that this extremity of the patch has almost a semicir- ciilar outline. The rest of the membrane, on the other hand, has an elongated, irregular form, and less distinctly defined edges. In another ovisac -gVth of an inch in diameter, there is the same entire absence of the germinal vesicle and the same presence of a delicate membrane of precisely the same characters, but not more than xg^nd of an inch in long diameter and 2~4oth of an inch wide. In this specimen the edge of the membrane which is turned towards the duct is still more distinctly semicircular, and it is almost as well-defined as the edge of the ger- minal vesicle in its latest condition, though no distinct membrane is discernible. The irregular part of the membranous disk bears a smaller proportion to the semicircular part, than in the preceding case. In each instance the membranous disk, which has been described, lies between the modified epithelium and the membrana propria. In position, therefore, it exactly corre- f;j)onds with the germinal vesicle ; its colour, when tlie light passes through a thickness AND DEVELOPMENT OF PYROSOMA. 229 of it sufficient to give colour, is exactly tliat of the contents of the germinal vesicle ; the diameter of the semicii'cular portion is hut very slightly greater than that of the germinal vesicle in its later stages ; and finally, the minute bodies which occupy the centre of each component corj)uscle of the membrane are not a little similar in character to the small spheroidal particles which appear upon the contents of the germinal vesicle durino- the latest stages of its existence. Putting all these circumstances together, I venture to express the belief that this membrane, which the further progress of development proves to be the blastoderm out of which all the parts of this embryo take their rise, results from the metamorphosis of the contents of the germinal vesicle ; and that the curved contour which lies towards the upper end of the duct is, in fact, the contour of that side of the germinal vesicle which first becomes filled with the yellow deposit. Thus far, I feel little difficulty in interpreting the appearances presented ; but if the siu'face and the immediate edges of the blastoderm are examined with great care, minute rod-like bodies will be seen scattered about, so similar in form and size to the heads of the spermatozoa, that I have been frequently tempted to regard them as such, and the more so, as in this stage the duct looks shrunken and shrivelled, and contains but very few, if any, remains of the plug of spermatozoa so conspicuous previously. In this stage, each of the blastoderms which I have examined has presented these appearances ; btit as, in spite of long search, the total number which I have found in this state does not exceed four, I do not feel myself in a condition to pronounce positively upon the nature of the bodies in question. Seventh Stage. Ovisacs from -^th to^^th of an inch in diameter, inioMchthe blastoderm rajiidlij increases, and becomes segmented into the rudiments of five zooids. Up to this stage the ovisac lies within the sinus system of the parent, which, as I have already pointed out, becomes accommodated to its increased dimensions, partly by the thrusting of the atrial tunic into the cavity of the atrium, but, to a much greater extent, by the formation of a chamber in the test, in consequence of the extension outwards of a diverticulum of the outer tunic. In the recent condition, the blood of the parent mtist circulate in the narrow space left between the walls of the ovisac and those of its contain- ing chamber ; and it seems reasonable to suppose that the former imbibes into its interior a supply of nutritive material, which will contribute towards the subsequent development of the embryo. But during and after this stage, the ovisac bearing the embryo is to be found loose in the mid-atrivun, which, in its later stages, it fills. To arrive at this position it must necessarily break through the wall of the atrium or atrial timic, and through the duct which still connects it with that tunic. The latter process is easily intelligible, considering the very small relative size and delicacy of the duct ; but I confess I do not understand how the rupture of the atrial tunic can be effected without serious heemorrhage. Uow- ever, the zooids in which the detached ovisacs have attained a large size appear to be in as good condition as any of the rest. Henceforward I shall speak of this complex body, composed of the ovisac and the 230 PROFESSOR HUXLEY ON THE ANATOMY embryo proper, as the foetus, reserving the term embryo for the blastoderm and the results of its modification. In such a foetus as that represented in fig. 10, the blastoderm is a broad, elongated, membranous patch g^rd of an inch long by x^th of an inch wide, and so opake as at once to strike the eye when the foetus is viewed with even a very low power. It is composed of somewhat coarse, granular-looking corpuscles, and lies between the membrana propria and the modified epithelium ; but the former is separated from it by a very thin layer of structureless siibstance which extends for some little distance beyond the limits of the blastoderm on each side. The further course of development shows that this layer is the rudiment of the test of the future ascidiarium. Foetuses of very slightly increased or even of less size exhibit a marked change in the embryo, which has elongated sufficiently to extend over half the circumference of the ovi- sac and has, at the same time, become indented at opposite points of its margins, so as to be marked out into five short segments. One of the two terminal segments becomes much enlarged, spreading over and investing one pole of the ovisac like a cujd ; while the other four remain far smaller, and, the indentations between them deepening, they are eventually connected only by narrow isthmuses of blastoderm. These segments are the rudiments of as many zooids ; but the large cup-like one has a totally different fate from the rest, and for distinction's sake I shall term it the cyathozooid, while the others are, in their order of nearness to it, the 1st, 2nd, 3rd, and 4th ascidiozooids * respectively. The zooids are not merely connected with one another by the isthmuses of blastoderm above- mentioned, but the structureless test has greatly increased in thickness, and now invests them all, like a thick layer of transparent varnish. The membrana propria of the ovisac is no longer distinguishable outside this rudimentary test. The remains of the duct are often still traceable, towards the conclusion of this stage, at one end of an equatorial diameter of the foetus (supposing the cyathozooid to be situated at one of its poles) ; but later, it is no longer to be discovered. Eighth Stage. Foetuses from -^§tli of an inch tip to the largest loldcli have been met loith. In describing tliis final stage of development, it will be convenient to consider, first, the changes in general arrangement, size and form, of the different parts of the foetus ; and secondly, the special modifications which each of these parts undergoes. The cyathozooid, at fu-st, occupies but a comparatively small segment of the surface of the spheroidal foetus, and the slightly curved series of ascidiozooids stretches out from it, over about half the circumference of the uncovered portion of the ovisac (PL XXXI. fig. 11). But, by degrees, the cyathozooid extends so far as to invest nearly half the surface of the ovisac, and, at the same time, the chain of ascidiozooids (considered as a whole) gradually assumes a new direction, and applies itself closely to the face of the cyathozooid, whose circumference it half encircles (fig. 13). Tlie blastoderm of the ascidiozooids, however, remains perfectly distinct from that of the cyathozooid, the two being united only by the layer of test, which, in the earUer stages, invested both, and whose con- tiguous edges now seem to run into one another. *I have, throughout the present memoir, used the term ' ascidiozooid,' as more euphonious than 'ascidiite,' employed in my notice in the ' Annals of Natural History' for 18G0. AND DEVELOPMENT OF PYROSOMA. 231 If a line traversing the centre of the cyathozooid and the centre of tlie ovisac be regarded as the axis of the whole foetus, then, in the present condition, the longest diameters of the first and of the last ascidiozooids are parallel with that axis, and that extremity of each, at which the elseoblast is situated, is directed away from the cyatho- zooid. The long diameters of the intermediate ascidiozooids, on the other hand, cut the axis of the foetus at a high angle, their elteohlastic ends being those which are nearer the cyathozooid (fig. 13). As development advances, the fu'st and the fourth ascidiozooids retain their parallelism to the axis of the foetus, while the whole series elongates, so that the fourth comes to l^e situated close to the fii-st (Plate XXXI. fig. 14-), the four encircling the base of the cyatho- zooid completely. This elongation of the whole series is effected, mainly, at the expense of the isthmuses, which elongate so much as to be converted into slender cords, of which the first connects the cyathozooid with the neural face of the first acidiozooid ; the second connects the haemal region of the first ascidiozooid, at a point just opposite the endostylic cone, ■with the neural face of the second ascidiozooid ; the third similarly unites the second and the thu'd ; and the fourth, the third and fourth. But the elongation of the isthmuses is not merely sufiicient to allow the fourth ascidio- zooid to come close to the third ; it is also enough to permit of a movement of rotation on the part of the second and third ascidiozooids. The fkst and fourth, as has been seen, early take up such a position that their long axes are parallel with the axis of the foetus ; and, by degrees, the second and third revolve, their adjacent ends being allowed to sepa- rate by the elongation of their connecting isthmuses, until their long diameters, from being very obliquely inclined to that axis, also become parallel with it, and with the long dia- meter of the first and fourth. Thus, at last, the long diameters of all four ascidiozooids are parallel with one another and with the axis of the foetus, their similar ends being turned the same way (fig. 14), while the isthmuses slope obliquely from the neural region of one to the haemal region of the next. The long diameter of each ascidiozooid is at right angles with its proper axis (which would be a line drawn from the oral to the cloacal aper- ture), and, hence, the neural and hssmal sides of the body are at opposite ends of its long diameter. The neural side is that, which is turned in the same direction as the aperture of the cyathozooid, whUe the haemal side is the opposite. The mouth is at that end of the true axis or short diameter of the body which is turned outwards ; while the atrial aperture eventually makes its appearance at the other end of this diameter and, consequently, on that face of the ascidiozooid which is adjacent to the ovisac and cyathozooid. At the commencement of the series of changes here indicated the ascidiozooids arc, indi- vidually, much smaller than the combined cyathozooid and ovisac ; but as development advances, the latter diminish Avhile the former increase ; and as, by the increase of size of the ascidiozooids, the interval between them becomes both relatively and absolutely less, they, at last, completely hide the combined cyathozooid and ovisac from view, so that it is not always an easy matter to find the latter (Plate XXXI. fig. 15). The test increases, concomitantly with the ascidiozooids, enveloping them and filling up their intervals so as, finally, to form a spheroidal investment for the entire tetrazooidal foetus (figs. 11. & 15). 232 PROFESSOR HUXLEY ON THE ANATOMY Duriuo- the Avhole of these changes and until the foetus attains a diameter of i^jth of an inch, it remains within the mid-atrium of the parent, which, at last, it completely fills. With Savigny, I am unable to imderstand how it escapes, imless indeed it becomes freed by the destruction of its parent. For it seems quite impossible that the foetus should find a way open to it by any conceivable amount of dilatation of the atrial aperture. Nor does one ever find a fully formed ascidiozooid without a foetus in its mid-atrium. And if, at the same time, it is recollected that only one ovum ever comes to maturity in an ascidiozooid, so that when the fojtus has arrived at its full development the parent's "occupation is gone," it seems less improbable that the destruction of the latter should be involved in the maturity of its offspring. Such is a general description of the changes in the size, form, and position of the chief constituents of the foetus, in virtue of which it assumes its final characters. It now becomes necessary to trace the internal modifications which each of these constituents undergoes. 1. The Cyathozooid.—ln my brief preliminary sketch of the development of Ftjrosoma ('Annals of Natural History' for January, 1860), I have termed this part the "rudi- mentary cloaca ;" but it would have been a more accurate account of the matter, if I had called it the ' moiild ' or ' forerunner ' of the cloaca. Rudiment of the cloaca, in the strict sense of the words, it is not ; for, as we shall see, the atrial apertures of the ascidiozooids never really open into it. "When the cyathozooid is first distinguishable as a separate segment and traces of structure are discernible in it (Plate XXXI. fig. 11), it presents, when viewed from above, near that edge which is most distant from the first isthmus, a rounded depression. Viewed sideways, the blastoderm appears to be divided into two lamellge, the separation between which is most marked immediately under the depression. In a line between the depression and the first isthmus a clear streak is visible, the first rudiment of what I shall term the appendix of the ctjathozooid. As the development of the foetus progresses, the interspace between the two layers of the blastoderm enlarges and the depression becomes an opening, into which, however, the thick test is continued, projecting like a conical tongue into the interspace or cavity just mentioned, in such a manner as to leave but a narrow median passage, by which I conceive that a free communication between the cavity of the cyathozooid and the exterior must be effected (figs. 17 & 18). At the same time, the aperture is gradually shifted from the margin to the centre of the cyatho- zooid, so that, eventually, its middle corresponds to one pole of the foetus (fig. 14), and gives the latter the appearance of a cup, or of an egg with its top cut off. Contemporaneously with these changes that streak which I have mentioned takes shape as a singular appendage situated between the two layers into which the outer wall of the cyathozooid is differentiated, and a communication, which, I believe, existed from the first between the cyathozooid and the first ascidiozooid by means of the first isthmus, becomes patent and obvious. But a description of the structure of a more advanced cyathozooid will best render these changes intelligible. Fig. 14 represents a foetus -^gih. of an inch in diameter. The cyathozooid and ovisac, taken together, have the form of an ellipsoid, truncated at that end which presents the aperture of the cyathozooid, and rounded at the other. The circular aperture of the cyathozooid (/3) is tu jth of an inch across, and is bounded by a constricted perpen- AND DEVELOPMENT OF PYROSOMA. 233 dicular lip x^tli of an inch deep. The aperture leads into a wide cavity about as deep as the lip (y), into which the prolonged tongue of the test projects. The canal which traverses the centre of this tongue, and which consequently must place the cavity of the cyathozooid in communication with the exterior, appears very distinct. The appendix {h) has the form of a curved tube, with its concavity turned towards the cavity of the cyatho- zooid. Its anterior end is slightly enlarged, while its posterior end, also a little dilated, is seated upon a slight prominence : both ends seem to be open. On one side of this appendix, a canal (0), xi oth of an inch long by 93-0 tli wide, passes obliquely towards the cavity of the cyathozooid and apparently opens into it. Posteriorly it is continued, at an obtuse angle, into a similar tube having about the same length, and eventually passing into the first isthmus, now yieth of an inch long. It will be observed that, notwithstanding the advanced condition of the ascidiozooids in this foBtus, their upper extremities, do not rise so high as the level of the middle of the ellipsoid formed by the cyathozooid and ovisac. The point at which their atrial apertures will eventually be formed, consequently, can hardly be so high as the lower end of that ellipsoid. As has been akeady hinted, with the advance of the foetuses in size all their relations become changed. The ascidiozooids, instead of presenting a fraction of the length of the combined ellipsoidal cyathozooid and ovisac, and occupying only a small portion of the mass of the foetal spheroid, gradually become fully thrice as high as the ellipsoid in question, and form by far the greater proportion of the mass of the spheroid (fig. 15). The ovisac and cyathozooid, again, diminish, not only relatively but absolutely (fig. 16), inasmuch as their largest diameter does not eventually amount to more than Tijth or gV^h of an inch, while the lip and the internal cavity of the cyathozooid become less distinct structures than before. But the most curious change is that which has taken place in the test in the vicinity of the cyathozooid. It has, as it were, separated itself from the latter, following the asci- diozooids as their vertical diameter increases, whereby the central tongue of the test is pulled completely out of the mouth of the cyathozooid, as one might pull a finger out of a glove (compare figs. 18 and 19, Plate XXXL). As a consequence of this operation a cavity, which gradually increases in dimensions, is developed between the outer surface of the cyathozooid and the inner wall of the test ; and as the atrial ends of the ascidiozooids ascend in consequence of the growth of the latter, they open into this cavity, Avhich thus manifests itself as the cloaca (fig. 19). The tongue-like prolongation of the test becoming pulled out and flattened as the cloaca widens, ultimately ceases to project into the cavity of the latter, and becomes converted into the lip of its aperture. In fig. 19 it still protrudes for some distance into the cloacal cavity. 2. The Ascidiozooids. — Prom their small size, flattened form and general opacity, it is by no means so easy to trace satisfactorily the successive changes by which the other seg- ments of the blastoderm are converted into perfect ascidiozooids, as it is to follow out the development of the buds. Nevertheless, knowdng the latter process, it is not difficult to interpret the appearances presented by the segments of the blastoderm, in the course of their development. When the blastoderm first becomes marked out into those segments which eventually VOL. XXIII. 2 I 234 PROFESSOR HUXLEY ON THE ANATOMY constitute the ascidiozooids (Plate XXXI. figs. 11 and 12), each segment is about yi^tb of an inch long by as much broad, and has a thickness of less than g-^oth of an inch. Like the blastoderm whence it proceeded, the segment appears to consist of nothing but a dense, opake mass of indifferent tissue. In a somewhat more advanced condition, the first signs of organization appear in the form of a clear median longitudinal streak visible in each segment when it is viewed from above. The streak is bounded by two more-opake lines, and on each side of the whole is a more opake mass. If the foetus be turned, so as to display a transverse section of one of the segments, the clear streak is seen to correspond with a central cavity answering to the alimentary tract of a bud, while the more opake lateral masses are plainly small sacs — ^the lateral atria. The isthmus between any one segment and the next is clear in the middle, and has every appearance of a tube connecting the alimentary tracts of the two segments ; but if, as I have already said, the first isthmus enables the alimentary tract of the first ascidiozooid to communicate with the cavity of the cyathozooid, then the cavities of all the alimentary tracts of the ascidiozooids must be, indirectly, in communication with this cavity and, through it, with the exterior. In point of fact, I believe that the four primary ascidiozooids stand in the same relation to the cyathozooid, as four buds formed from the ascidiozooids in the way described above would do, if, in the process of gemma- tion as many remained connected together and with the parent ; for, as we have seen, all the branchial sacs of the buds communicate with that of the parent and, by the latter, with the exterior. And the mode of connexion of the different ascidiozooids is exactly the same in the two cases ; for, in somewhat more advanced foetuses (in which the ascidiozooids are about xoo^li of an inch long and broad), it is obvious that the clear streak above men- tioned corresponds with the interval between the bands of the endostyle, and that the end of the alimentary tract of any one embryonic ascidiozooid which is continued into the isthmus corresponds with the endostylic cone of ordinary buds ; while that part of any embryonic ascidiozooid which receives an isthmus is the interval between the oesophageal aperture and the ganglion, just as this is the place into which the peduncle of a bud opens. In ascidiozooids of this size, the nature of what I have termed the lateral atria is demonstrated by the appearance of four or five stigmata in their inner w^all, just as in buds at a corresponding stage. At the same time, that part of the indifferent tissue of the embryo which lies in the immediate vicinity of the pointed end of the alimentary tract (the future endostylic cone) becomes converted into a mass of clear reticulated tissue, the elseoblast {(b). This body is developed more largely laterally than in the middle line, so that it appears, at first, as if it were composed of two distinct portions ; but its two moieties are really continuous with one another on the heemal side of the alimentary tract. The position of the future oral aperture is just indicated in the middle of the exposed surface of the ascidiozooids in this stage ; but I could not ascertain anything definite as to the condition of the intestine. Indeed, from the flattened form of the embry- onic ascidiozooids and their close apposition to the ovisac, it is exceedingly difficult to decipher all the details of their internal structure. Ascidiozooids of --j^qUi of an inch in length exhibit a well-defined, though not open oral aperture, gioth of an inch in diameter. The l^ranchial stigmata have increased in AND DEVELOPMENT OF PYROSOMA, 235 number to nine or ten on each side and the middle ones occupy the whole depth of the branchial sac ; but there are, as yet, no longitudinal branchial bars. The mode of origin of the stigmata appears to be just the same as in the buds. The nervous ganglion makes its appearance as a thick mass of indiflPerent tissue between the isthmus and the oral aperture ; and the contour of the gastro-intestinal part of the alimentary canal is discernible on that face of the embryo which is nearest the ovisac. The isthmuses have lengthened to -jio^^i of an inch. A foetus a^-th of an inch long, whose ascidiozooids had a length of -6-3-rd of an inch, presented the clear profile view of one of the latter, which is represented in PI. XXXI. fig. IB a. The central tube or canal of the first isthmus is obvious, and it opens freely into the branchial sac of the ascidiozooid between the ganglion and the oral aperture. The central canal in question is bounded by the inner tunic of the ascidiozooid, between which and the outer tunic is an interval which is connected, on the one hand, with the sinuses of the ascidiozooid, and on the other, with the space between the two walls of the cyathozooid. The ganglion is very distinct and occupies nearly the whole interval between the oral aperture and the isthmus. The diapharyngeal band, already visible in earlier stages, is continued from above the posterior half of the ganglion to the roof of the branchial cavity ; but its proportions are more slender, as it measures axoth of an inch long, by ygVoth of an inch thick. The oral aperture is not open, but its lobed tentacular fringe may be observed, the haemal tentacle being even now distinguished from the rest of the fringe by its length and form. Nine or ten branchial stigmata are discernible ; but there are, as yet, no longitudinal branchial bars. The intestine is completely fashioned ; and the elseoblast is large, con- spicuous, and composed of a reticulated tissue. In a foetus -aVst of an inch long, with ascidiozooids -g^th of an inch long by eifth of an inch from their oral to their cloacal extremities and nearly hemispherical in form (the flat side of the hemisphere being applied to the combined ovisac and cyathozooid), the isthmuses measure -^oih of an inch in length ; and it is obvious that, while their central canals connect together the branchial sacs, the interspaces between their double walls place the sinuses of the ascidiozooids in communication. There are ten branchial stig- mata, of which the first and last are very small ; and sis or seven longitudinal branchial bars have made their appearance. The mid-atrium is distinctly developed below and behind the gastro-oesophageal part of the alimentary canal. The place of the atrial aperture is indicated by the union of the atrial and outer tunics, in a round spot at the posterior part of the mid-atrium. In fact, the whole zooid is nearly in the same con- dition as the bud represented in PI. XXX. fig. 22. The renal (?) organ has made its appearance as a patch of opake yellowish cells. Ascidiozooids ^^nd of an inch long, by ^th from their oral to their cloacal apertures, which form part of a foetus Jg-th of an inch long, and surround the lower half of the combined cyathozooid and ovisac, have ten or eleven stigmata and seven longitudinal branchial bars. The heemal tentacle is well developed, the rudiment of the ciliated sac is discernible, and between the aperture of the central canal of the isthmus and the oeso- phagus are two rudimentary languets. The diapharyngeal band is very slender. The 2 I 2 236 PROFESSOR HUXLEY ON THE ANATOMY elseoblast has a length of about yV^h of an inch. The rudiment of tlie atrial aperture (round, and about s^o^h of an inch in diameter) lies altogether below the level of the equator of the combined cyathozooid and ovisac. There is an indistinct appearance as of a small cavity between it and the latter organs. The posterior end of the endostyle appears quite distinctly to be continued back into the central canal of the isthmus. The rudiment of the heart is obvious, in close connexion \^ith, and apparently developed from the wall of the branchial sac ; and there are two slight papillary elevations in the place whence the stolons will be given off. In a foetus of about the same diameter as the preceding but whose ascidiozooids have a vertical diameter of ij-a-rd of an inch, while the combined cyathozooid and ovisac are -g-'oth of an inch long, the neural boundaries of the ascidiozooids project a little way beyond the open end of the cyathozooid. The upper edges of their atrial apertures, now yi o^h of an inch in diameter, are still fully ^xoth of an inch below the margin of the cyatho- zooid ; and although the formation of the true cloacal chamber has commenced by the separation of the test from its cyathozooidal mould, yet its depth is so slight (not more than 3^0 th of an inch) that the end of the tongue- like inward prolongation of the test still lies between the lips of the mouth of the cyathozooid. A foetus of iVth of an inch in diameter lias the combined cyathozooid and ovisac not more than -^gtli of an inch long, and cup-shaped — its upper, open end being as broad as its middle. The atrial apertures of the ascidiozooids (which measure ^th of an inch in long diameter) are vertically oval, g^th of an inch long, and lie almost wholly above the level of the upper edge of the cyathozooid. They open at once into the cloacal cavity, which, as measured from its roof, formed by the now hardly-projecting tongue-shaped process, to the upper edge of the cyathozooid, is -g-fth of an inch deep. The stolons of this foetus are -3 Jeth of an inch long, and are directed towards the aperture of the cloaca. In one of the most advanced foetuses I have met with (PI. XXXI. fig. 15), about Ysth of an inch in diameter, the greatest length of the ascidiozooids (or the diameter parallel to the foetal axis) was -g^-nd of an inch, while their antero-posterior diameter was i^.jnd of an inch. The long diameter of the combined ovisac and cyathozooid (the latter being now completely hidden between the hcemal moieties of the ascidiozooids) was only -Tij-th of an inch ; or, in other words, they had not a third of their former dimensions. Eacli ascidiozooid of this foetus has a roughly semicircular profile, the sti-aight side being turned towards the axis of the foetus. The curved contour is more convex on the haemal, more flattened upon the neural face. From side to side each ascidiozooid is much compressed, so as not to measure more than -5-Hth of an inch in this direction. The oral aperture is not yet pervious ; but a circular groove of the outer surface of the test, ^th of an inch in diameter, indicates the area in whose centre it will appear, around which centre lie the oral sphincter and the tentacular fringe. The latter, at present, not only projects into the buccal cavity but is divided into its processes ; and the hsemal tentacle, -axo^h of an inch long, exhibits its characteristic enlarged base and finger-like process. The peripharyngeal ridge exhibits its distinctive structure. Hathcr in front of its upper loop, a small process (the upper end of the diapharyngeal band) AND DEVELOPMENT OF PYROSOMA. 237 projects from the roof of tlie pharyngeal sac; and a corresponding remnant of the lower end of the same band is seen, as a small projection of the neural wall of the cavity, just above the tubercle of the ganglion. The urinary (?) organ is very distinct as a mass of pale, spheroidal, granular bodies, and occupies its normal place. The ganglion, and so much as could be made out of the ciliated sac are similar to the same structures in adults; but the ganglion has a length of only -g-ioth of an inch. In one of the ascidiozooids of this specimen the isthmus can be well studied as it passes off from the neural side immediately behind the ganglion. Where it joins the ascidio- zooid it is -^fsth of an inch wide, but, in the middle of its length, it has a diameter of not more than g loth of an inch. In consequence of its passing obliquely from the neural face of one ascidiozooid to the haemal face of the next, it is, of course, rather longer than the largest diameter of the ascidiozooid (or more than -^nd of an inch long). Viewed from the side, it looks like a clear, transparent tube, divided by a partition into two channels ; but where it bends round, and so exhibits a transverse section, this partition is itself clearly seen not to be a simple septum, but to be formed by two membranous lamellae, which stretch from wall to wall of the isthmus, and are themselves separated by an interval of :j^ijPoth of an inch. In fact, the central canal has now assumed this partition-like charac- ter. If traced up to the neural wall of the one ascidiozooid with which it is connected, the outer membrane of the isthmus obviously passes into the outer tunic of the ascidio- zooid, while the walls of its contained, inner canal are continuous with the inner tunic, or pharyngeal mucous membrane, of the same part. On the other hand, if it be followed to the haemal wall of the other ascidiozooid, the outer membrane of the isthmus passes into the outer tunic of that region, while the wall of the inner tube is continuous with the endostylic cone. It is obvious, therefore, that the composition of the isthmus is, in reality, the same as in earlier stages' and that, while its central canal connects the pharyngeal cavities of the two ascidiozooids, the interspace between this canal and the outer walls of the isthmus connects their sinuses. Between the attachment of the isthmus and the oesophageal aperture only two languets are developed from the hj^oopharyngeal region. The great sinus beneath them is full of agglomerated blood-corpuscles. The endostyle is still broad proportionally (T5~oth of an inch), but all its parts are well developed. It ends posteriorly in a short process or endostylic cone, -sio'tli of an inch long, which, as I have said above, passes into the central tube of the isthmus. A cellular mass, 5^i:oth of an inch long, is attached to the external tunic, close to the end of the endostylic cone, if not directly connected with it ; and this, T am inclined to think, is the rudiment of the generative blastema. I have not been able to detect any distinct structure (as of an ovisac or testis) in it, Aihich is remarkable when one considers the early appearance of the ovisac in the bu.ds. The branchial stigmata are altogether tvfelve in number. The anterior and posterior are rudimentary while most of the others extend across almost the whole depth of the branchial sac. The cilia are perfectly distinct upon their edges. The longitudinal bran- chial bars are nine in number. The intestine has nearly the same form as in the adult, 238 PROFESSOR HUXLEY ON THE ANATOMY and the tubular, hepatic system is well formed. The heart is visible in its place. The elseoblast is a mass of clear reticiilated tissue, causing the haemal wall to bulge a little on each side of the middle line, and occupying the interval between the endostyle and gene- rative blastema, on the one hand, and the heart and intestine on the other. The atrial aperture is enormous in proportion, occupying the greater part of the inner face of the ascidiozooid above the level of the cyathozooid and attaining a length of fully :^2nd of an inch and a breadth of xr^th of an inch. In other words, the atrial aperture is six times as large as it is in the adult ascidiozooid, though the latter is at least eight or ten times as large as one of the zooids of the foetus under description. In consequence of the great proportional size of these oval apertures, whose long diameters are parallel with the axes of the foetus, the intervening wall of the cloaca is very narrow. The cyathozooid and ovisac are -g^jth of an inch long by y^Q-th wide and more cylindrical than cup-shaped. The aperture, still distinctly visible, has a diameter of -g^oth of an inch ; and as the cloacal chamber is now j^-th of an inch deep, the margin of the aperture is but just on a level with the convex, neural margin of the oesophagus of any of the asci- diozooidsc Where the former tongue-like process existed, the roof of the cloaca now hardly projects inwards at all. The atrial muscles are visible as very delicate, straight bands, ^^^-th of an inch long by 3^00^^^ wide, which take an oblique course on each side, from a point a little below the end of the endostyle, neurad and a little forwards, to a point opposite the commencement of the oesophagus. In the middle of their course these bands lie very near the lips of the atrial aperture. The stolons are g-jotli of an inch long ; they pass almost horizontally inwards, towards the rudimentary lip of the cloaca, and are curved towards its cavity, at their blind extre- mities. The corpuscles of which their Avails are composed are more elongated than before and, sending processes into the adjacent substance of the test, cause the .ca^cal ends of the stolons to have a very pecuHar, brushlike appearance. 3. The Test. — As I propose to reserve the description of the histological changes under- gone by the embryo of Tyrosoma for another occasion, I will merely state, in this place, that the test appears, at first, to be a structureless excretion. Subsequently, cellular bodies, like connective-tissue corpuscles, are discernible in its most superficial layer, and are disposed in such a manner as to form a very regular, hexagonal network, with large meshes. The most advanced foetus has presented neither of the fibrous layers visible in the adult test. Ninth Stage. The conversion of the tetrazooidal fa-tus into the adult ascidiarkmi. The most advanced foetus which has been described differs from tlie adult ascidiarium not merely in size, in the paucity of its ascidiozooids, in the form and proportions of the latter, in the absence of buds, or ever so slightly differentiated reproductive organs, in them, and in their large atrial apertures (aU of which are peculiarities which we may easily conceive to be altered by age and growth), but in still more important charac- ters, seeing that in the adult ascidiarium I have met with no trace of the cyathozooid or the isthmuses, nor have I been able to discover any ascidiozooid with two stolons. AND DEVELOPMENT OF PYROSOMA. 239 The first tlieory of the mode of formation of the adult ascidiarium which suggests itself is obviously that which supposes that the four ascidiozooids of the foetus give rise, by budding, to all those of the adult Pyrosoma, at the same time losing the two stolons, and acqumng reproductive organs, so as to be undistiuguishable from their agamogenetic progeny. But difficulties arise when we compare this theoretical conception with the structural characters, and the ascertained laws of gemmation of Pyrosoma. In every ascidiozooid of the adult ascidiarium (and there is no reason to suppose that those of the tetrazooidal foetus constitute exceptions to the rule) budding takes place, as we have seen, from a single definite region of the body, situated in the posterior moiety of the htemal surface ; and the buds remain, in nearly the same plane as that in which they were given olT, until they have attained some distance from the parent. It has been seen, in fact, that three buds, given off successively from one ascidiozooid, may be visible, one below the other, in the same, not very thick, longitudinal section. But in the tetra- zooid, as in the adult, the hsemal side is that turned away from the aperture of the asci- diarium. If, then, the buds thrown off from the ascidiozooids of the foetus all remain on the hfemal, or apical, side of their parents, we ought, on examining the adult organism, to find the four primitive ascidiozooids close to the margin, with a series of two or three buds, in various stages of development, attached to each. As a matter of fact, however, no section taken near the margin of the aperture has ever presented an appearance essentially different from that represented in PL XXX. fig. 4. The ascidiozooids have always been young, and, on the average, younger, the nearer they were to the margin. But they have never been younger than such a bud as is represented in fig. 24i, PI. XXX. ; and those of the first three or four tiers have always possessed imperfectly developed sexual organs, and buds not more advanced than those represented in figs. 19 and 20. That the ascidiozooids which lie nearest the aperture are the result of the budding of other ascidiozooids is beyond all doubt. As I have traced the development of the stolon from such a modified bud, it is clear that the bud is not developed, as I had once imagined, from the stolon of another ascidiozooid, — these stolons being invariably trace- able, without a break, into the lip of the cloaca, where they end cajcally. There appears to me, then, to be no other course open but to suppose that these yoimg ascidiozooids which lie nearest the aperture, are buds which were originally developed from the hsemal region of ascidiozooids which lie nearer the apex, and that they have consequently passed round and to the neural side of their parents. If this migration of the buds reaUy occurs, it will follow, as Savigny supposed, that the four apical ascidiozooids of the adult are the modified zooids of the foetus,— the buds developed from their haemal waUs not remaining upon their apical side, but passing up between them on to their neural sides, and there becoming themselves new centres, whence fresh buds are thrown off, which gradually take their places in a still higher tier. I can conceive of no other mode in which the structure of the foetus, the structure of the adult and the law of budding can be reconciled ; and yet. I am reluctant to admit so seemingly artificial a process on anything short of direct evidence. Such evidence, 240 PROFESSOR HUXLEY ON THE ANATOMY however, is only to be obtained by the examination of young Pi/rosomata but little larger than foetuses, none of which have come into my possession. § 5. Smnniary and discussion of the results of the observations on the gamogenetic develojmient of Pyuosoma gigakteum. If the observations detailed in the preceding pages be correct, and no flaw be found in that interpretation of them which has been offered, it follows that — 1. The ovisac of Fyrosoma at tu'st contains an ovum, altogether similar to that of other animals, and in particular resembling that of many Ccelenterata, IfoUuscoida, and Ilollusca, in the absence of a vitelline membrane. 2. Impregnation is effected by the passage of the spermatozoa up tlie duct of the ovisac, and it would seem that these spermatozoa must immediately come into contact with the yelk ; but when, and how, the essential act of fecundation (consisting in the action of the spermatozoa upon the germ) takes place, does not appear. 3. There is neither complete nor partial, yelk-division ; but the vitellus disappears, as such, apparently becoming diffused through the contents of the ovisac, which rapidly in- creases in size. By this deliquescence of the yelk the germinal vesicle is laid bare. 4. The germinal vesicle adheres to a particiilar spot of the epithelial lining of the ovi- sac, close to the opening of its duct and, eventually traversing that epithelial lining, takes up a position between it and the membrana propria of the ovisac. 5. In the meanwhile, a turbid deposit takes place in that moiety of the germinal vesicle which lies nearest the aperture of the duct ; and the germinal spot, a remarkably obvious structure in this and preceding stages, is partially imbedded in this deposit. 6. The germinal vesicle grows and becomes more flattened ; but soon, although it is quite translucent, the germinal spot can no longer be found in it. In the place of that structure and resulting, as I suppose, from its division, a number of small, clear, sphe- roidal corpuscles are visible upon the face of the deposit in the germinal vesicle. 7. Next, the germinal vesicle, as such, is no longer visible; but, occupying the same place, preserving the same colour, having very nearly the same size and, on the side turned towards the duct, the same curved contour, there is a flat patch, consisting of a single layer of excessively delicate corpuscles, each with its clear space and central par- ticle— constituting the commencement of the blastoderm. 8. The blastoderm enlarges, assumes a band-like form, and becomes divided by con- strictions into five segments : of these, one becomes the cyathozooid — a temporary structuxe, which is especially attached to one pole of the ovisac, and, among other pur- poses, serves as a sort of precursor, or mould, of the cloaca ; the other four are converted into ascidiozooids. 9. The ascidiozooids enlarge and eventually give rise to the tetrazooidal foetus described by Savigny. The cyathozooid and the ovisac which it surmounts diminish in size and, probably, eventually disappear. The lining epithelium of the ovisac early acquires a pecu-liar vesicular structure. 10. All these changes, subsequent to the formation of the blastoderm, take place in the AND DEVELOPMENT OF PYROSOMA. 241 mid-atrium of the parent, which the fetus, at length, completely fiUs. There appears to be no placental connexion between the foetus and the parent; but the nutritive matter contained in the large ovisac may well be supposed to pass into the sinuses of the cyatho- zooid and thence into those of the ascidiozooid, and thus to subserve the nutrition of the whole foetus. In successively commenting upon the preceding paragraphs, I shall consider how far the embryogeny of Pyrosoma can be parallelled by that of other animals, and how far it offers exceptional peculiarities. 1. I do not think that any one, acquainted with the structure of the ovarian ova of other Ascidians and of the MoUusca generally, will entertain the slightest doubt that the parts called germinal spot, germinal vesicle and vitellus, respectively, in the preceding pages, really have the nature I liave assigned to them. The ovisac corresponds with a single acinus of the ovary of other Ilollusca and Molluscolda, and is altogether similar to the solitary ovisac of Sal pa. 2. The process of impregnation presents nothmg anomalous ; but, as regards the act of fecundation, it is remarkable that the spermatozoa should so long remain aggregated in a mass in the upper end of the duct, without, to aU appearance, penetrating into the cavity of the ovisac or into the substance of the yelk. StiU more singular is that appearance of scattered, rod-like bodies, not unlike the heads of spermatozoa, upon and about the very young blastodeimi. If I could feel thoroughly assured that these bodies are really the spermatozoa, I should be inclined to follow out to some length a series of considerations suggested by the fact, as to the essential natiu-e and place of occurrence of impregnation. For the present, however, I will merely remind the reader that the so-called ' disappear- ance of the germinal vesicle,' and even a certain progress in yelk-division, may take place without impregnation*; whence it may seem less strange than it ajipears at first sight, to suppose that the influence of the spermatozoa may be exerted, in some cases, not upon the yelk, nor upon the germinal vesicle as such, but upon the nascent blastoderm. 3. The only animals which, so far as I know, present a condition of the yelk at all comparable to its liqvxefied and pellucid state in Pyrosoma, are Ascaris dentata, Cucullamis elegans, and Oxyuris ambigua. In these nematoid worms, the vitellus, according to Kolliker t, is represented only by a clear, transparent fluid containing a very few granules, and it takes no direct share whatever in the formation of the embryo. The vitellus seems to play an equally subordinate part in the great majority of the Articulata, but in these animals it is commonly opake and granular. 4. If the ovisac of Pyrosoma be compared with the Graafian follicle of a mammal, the resemblance (notwithstanding their obvious difl^erences) of the two structures is marked ; and the manner in which the germinal vesicle traverses the epithelium of the ovisac of Pyrosoma is singularly like the manner in which the mammalian ovum imbeds itself * See Leuckart, art. " Zeugung," Wagner's Handworterbuch, iv. p. 958. What Leuckart says here about the Frog is not in accordance with the results of the careful experiments of Newport (Phil. Trans. 1851, p. 190), who arrives at the conclusion that segmentation certainly does not take place in the unimpregnated ovum. Vogt's case is not satisfactory, as there is no counter evidence to show that impregnated ova would have developed under the circum- stances. Bischoff's observations on the Sow (Ann. des Sci. Nat. 1844), however, appear to be unexceptionable evidence. t Beitrage zur Entwickelungs-geschichte wirbelloser Thiere. Miiller's Archiv, 1843. VOL. XXIII. 2 K 242 PROFESSOR HUXLEY ON THE ANATOMY among the cells of the proligerous disk. A still closer parallel, perhaps, is presented by the bu'd's egg, if we consider the mode in which its germinal vesicle (which at first occu- pies the centre of the future egg, and is contained in a primitive ovum surrounded by, at any rate, a rudimentary vitelline membrane) passes to the surface, and eventually lies immediately beneath the membrane which encloses the food-yelk *. 5, 6, 7. The consideration of the phenomena enumerated under these heads opens up the whole vexed question of the fate of the germinal vesicle. Since the imaginations of Dr. Martin Barry have fallen into just discredit, most physiologists have more or less distinctly adopted the doctrine that the germinal vesicle and its contents lose their identity and disappear ; and that the embryo-cells, whence the blastoderm arises, are new structures not directly derived from them. The evidence by which this conclusion is supported, however, will be found, if closely sifted, to be, for the most part, not only negative, as by the nature of the case it must be, but weakly negative. That is to say, not only is the conclusion based upon the circum- stance that, at a given period, the observer was unable to find the germinal vesicle or to identify its contents, — but, in most cases, the circumstances are such that he might very well have missed them had they existed. Even in Fyrosoma it is no easy matter, until one has had some practice, to find the germinal vesicle Avhen it is passing into the blastoderm, although, in all the earlier stages, nothing can be more obvious ; and had the ovisac of Fyrosoma been filled with even a very slightly granular yelk, I believe the discovery of the germinal vesicle, at this period, would be almost impracticable. What wonder, then, that it should be impossible to identify the germinal vesicle or its contents in the midst of the more or less opake and coarsely grantilar substance of which the yelk of ninety-nine ova out of a hundred is composed ? The only case to which this reasoning does not apply is that described by KoUiker in the paper already referred to (/. c. p. 76) : — "As regards the internal changes undergone by the eggs [oi Ascaris dentata], the most striking fact is that, immediately after fecundation, the germinal spot and the germinal vesicle have disappeared and the clear and transparent yelk contains notliing but scanty elementary granules. This is a point of great importance ; and to show that there is no possibility of being deceived about it, I add, that the ovum of Ascaris dentata, including its chorion and vitelline membrane, is so transparent that all the outlines of a body which may happen to lie beneath it are quite sharply and distinctly recognizable, and its contents are so clear and patent that hardly the smallest elementary granule of the yelk can remain hidden. Which of the two parts first disappears, the germinal spot or the germinal vesicle, I cannot as yet say with certainty ; but, in one individual, I saw two ova which had hardly traversed the seminal cells in the fundus uteri, and though they still exhibited a germinal vesicle, had no germinal spot. In another individual, I observed the same thing in an ovum imbedded in the midst of the seminal cells; so that I have some ground for the opinion that it is the germinal spot which disappears first. Further and repeated observations must decide whether this is the rule or whether, in other cases, it is not the germinal vesicle which disappears first. But I must observe, that this first stage of the development of the ova appears to be of very short duration ; * See Dr. A. Thomson's admirable article " Ovum " in Todd's Cyclopjedia. AND DEVELOPMENT OF PYROSOMA. 24-3 for, while no fecundated female Ascaris which I examined would have faUed to supply me with a complete series of all the other stages of development, it was but thus rarely that these first processes presented themselves. " As the ovum, now dejirived of its germinal vesicle and spot, is propelled downwards by the peristaltic contractions of the uterus, the first embryo-cell is formed in the middle of its clear yelk. I have never been able to detect the mode of its origin By endogenous development, the embryo-cells give rise to other cells, which become the blastodermic mjiss whence the embryo is formed. The yelk, as such, disappears." I am prepared to admit the full force of this carefully observed example of the disap- pearance of the germinal vesicle and the merging of its contents in the yelk, but it is the only case, within my knowledge, to which great weight can be attached ; while, on the other hand, independent observers have (of late years) recorded equally definite and positive observations that in some groups of animals, at any rate, the germinal vesicle does not disappear, but that it gives rise l^y division to the primary cells of the embryo. Thus, Dr. Nelson, in his memoir "On the Reproduction of Ascaris mystax" (Phil. Trans. 1852, pp. 580, 681), afiirms that the germinal vesicle of the impregnated egg of this worm bursts, and sets free the germinal spot, wliich is directly transformed into the first embryo-cell. The deservedly great authority of the late Johannes MliUer may be cited on the same side — so far, at least, as that singular moUusk, Eidoconclia mirabilis, is concerned. Dr. Gegenbaur afiirms the occurrence of a similar process to be the rule among the Calycoplioridce, Tliysophoridce, and certain other Hydrozoa, and in that singular annulose animal, Sagitta. Thus, in describing the development of Oceania armata (Zur Lehre vom Generationswechsel, 1851, p. 28), Gegenbaur says (the italics are his own) : — " Every act of division is preceded by a division of the nucleus, and consequently the first act by the division of the germinal vesicle : the transparency of the yelk allows of the most precise observation of all these phenomena, and the following of the develop^nent of the nuclei of the later embryo-cells out of the original germinal vesicle (the nucleus of the primitive ovi-cell)." Again, at p. 50 of his "Beitrage zur naheren Kenntniss der Schwimmpolypen " (1851), the same author remarks, in giving an account of the development of these Calycophoridce and JPhysophoridce : — " A process which may be here traced with particular clearness is the constant division of the germinal vesicle,' which precedes the division of the yelk ; and the products of the division of the germinal vesicle behave similarly, in relation to the subdivision of the yelk-masses. I observed this process of yelk-division in Agalmopsis, Physophora, For- skalia, I-Iip)popodins, and Biphyes, without noticing any important differences among them." Leydig expresses the same conclusion, though more guardedly, in his account of the development of the ova of Notommata Sieboldii * : — " The nuclei of the division-masses are very clear ; and it appeared to me as if the homogeneous, clear nucleus of the ripe ovum {the germinal vesicle) stood in a genetic rela- * " Ueber den Bau und die systematische Stellung der Raderthiere." Siebold und Kiilliker's Zeitschrift, 1855. 2 K 2 244 • PROFESSOR HUXLEY ON THE ANATOMY tion to the nuclei of the division-masses — i. e., gave rise to tliem by immediate division. The ovum is, in fact, more transparent than in other Rotifera ; and I have observed the absence of the germinal vesicle." In a subsequent passage Dr. Leydig adverts to these observations as having inclined him to alter his previously entertained opinions respecting the fate of the germinal vesicle. So far as the Vertehrata are concerned, such evidence as we possess as to the independent origin of the embryo-cells appears to be altogether of the weakly negative sort. I do not think it can be said that there is adequate foundation for the general assumption that the contents of the germinal vesicle take no direct share in their production ; on the contrary, as respects the Frog, I find definite evidence tending to a contrary conclusion. Prevost and Dumas, and Von Bar, as is well known, proved the existence of a canal leading from the centre of the dark part of the Prog's egg to a cavity wliich Von Bar considered to be the seat of the germinal vesicle. Newport (Phil. Trans. 1851) described and figured this canal and cavity, and showed that the germinal vesicle is, in the ovarian ovum, lodged in the cavity. The vesicle is said to lie dense, white, and opake, and its interior to be full of secondary cells*. Newport affirms that no trace of the vesicle is to be found in ova that have left the ovary, but that an accumulation of white nucleated cells sometuues occupies its place, in ova which are in the act of leaving the ovary. Ptemak (EntAvickelung der Wirbelthiere, 1855), apparently unacquainted with Newport's observations, doubts whether the cavity down to which the canal leads, and which he terms Von Bar's ' Kernhohle,' contains the germinal vesicle, though he inclines to the opinion that it does. But it is a most important circumstance that he proves (I. c. p. 137) that the division of this cavity accompanies each division of the yelk -mass, and that, eventually, these cavities become what he terms the nuclei provided with nucleoli, which occupy the centres of the division-masses of the yelk, and are the homologues of the embryo-cells of Ascaris. If both Newport's and Remak's observations are correct, it Avould seem impossible to deny that the embryo-cells of the Prog proceed from the con- tents of the germinal vesicle. I think, then, that considering the only case in which the contents of the germinal vesicle are not traceable, under circumstances in which it might be reasonably expected that, if they really exist, they should be visible, is that observed by Kolliker ; while, on the other hand, the equally definite observations of Nelson, MiiUer, Gegenbaur, and myself (and the less distinct evidence of Newport, Bemak, and of Leydig) testify to the origin of the blastoderm in one way or the other from the contents of the germinal vesicle, in various members of no less than fourf out of the five primary divisions of the animal kingdom ; the balance of the evidence is in favour of the conclusion that the em- bryo-cells are the progeny of another cell, and that here, as elsewhere, extracellular cell- development is a phenomenon of rare, if not of altogether questionable, occurrence. * Newport, it should be observed, used the term ' cell ' not very critically. But, ten years ago, cell-worship had attained its culminating point. f Calenterata, Mol/itsca, Annulosa, Vertehrata. I may add, that the first appearance of the blastoderm on the surface of the ovisac of Pyrosoma is so like that of the blastoderm in the ovum of any of the higher Articidafa, as strongly to suggest a similarity of origin. AND DEVELOPMENT OF PYROSOMA. 245 8, 9, 10. Thus far the problem has been to find a parallel for those early embryogenetic processes which are ordinarily common to large assemblages of living beings. Analogies for the more special modifications which the blastoderm undergoes may be sought for in the group of which the genus Pi/t'osoma forms a part. In the first place, it may be asked, are there, in this group, any examples of the division of the blastoderm into segments, one of which is to serve a temporary purpose, while the others become ascidiozooids ? Leaving the development of the caudal aj)pendage of ordinary Ascidians out of con- sideration, as hardly a case in point, it yet appears that even in these Ascidians, the body of the embryo is, during its locomotive stage, divided into two segments, the anterior of which gives rise to the so-called suckers (which are diverticula of its wall with involuted ends), while the posterior is the rudiment of the body. Lowig and Kolliker, in their description of the compound larva of Botrylliis (originally discovered and described by Sars), consider the three processes which are given off from the "large round 'mamelon' provided with an orifice" as the homologues of the three processes given off from the anterior division of the larval body in the simple Ascidians. In this case this ' mamelon,' which they consider to be the rudiment of the cloaca, must correspond with that anterior division. But the examination of their figures and descrip- tions renders it hardly doubtful to my mind that the ' mamelon ' is a structure homo- logous with the cyathozooid of the foetal Pyrosoma, the eight rudimentary ascidiozooids of the Botrylliis being arranged around its base, just as the four are disposed in the foetal Pyrosoma. If this reasoning be correct, it follows that the cyathozooid of Pyrosoma corresponds with the anterior division of the body in the ordinary Ascidian larva, e. g. of Clavelina. The peculiar connexion of the embryo Pyrosoma with its ovisac, and the extrusion of the latter combined with the embryo, as a single foetus, into the mid-atrium of the parent, are, however, peculiarities for which we should in vain seek a parallel among ordi- nary Ascidians. But there is one family of this class, the Salpm, which resemble Pyrosoma in having an elgeoblast, and in possessing no caudal appendage in the larval state (differing in the same respects from all other Ascidians), in which the search for analogies is more hopeful. Most SalpcB, like Pyrosoma, possess, as Krohn was the first to point out, but a single ovisac, connected by a peduncle-like duct with the wall of the mid-atrium. Prof. Leuckart, who, with a knowledge of all that had been written upon the question, subjected the reproductive processes of the SalpcB to a renewed and very careful scrutiny, some years ago * stated {I. c. p. 47) that the ovum of Salpa mticronata consists of a granular, tolerably viscid yelk, enclosing a large, vesicular germinal vesicle, with a simple germinal spot. No vitelline membrane was to be detected, — the only covering of the ovum being the ovisac, Avhich is closely applied to the surface of the vitellus, and is lined internally with a layer of small nucleated cells. The peduncle of the ovisac is a short, narrow duct, which only becomes a little thicker at its anterior end and, like the ovisac, is lined by an epithelium. Its anterior end opens into the atrium ; and in the vicinity of the aperture * Zoologische Untersuchungen, Zweites Heft, Salpen u. Verwandte. Gicssen, 1854. 246 PROFESSOR HUXLEY ON THE ANATOMY the inner tunic exhibits an elongated, discoid thickening, in which numerous small nu- cleated cells, like those in the ovisac and its duct, are to he detected. This description, it is obvious, would apply ec^ually well to the young ovisac of Fi/rosoma. It does not appear that the entrance of the spermatozoa into the duct of the ovisac has been observed in the Salp(B. I have stated in my Memoir already cited (Phil. Trans. 1851, p. 577), that in a more advanced stage, probably after fecundation, the ovisac (which I called ovum) appears like a cellular mass. H. Muller (Siebold and Kolliker's Zeitschrift, iv. p. 331) speaks of the occurrence of yelk-division at this stage, without, however, describing that process more particularly. Vogt did not observe it, nor does Leuckart add much to our information on this head : — " I can say little more about it [yelk- division] than that it begins during the change of place of the ovum (as H. MtUler has also observed), and, as in the allied Ascidians, is a total yelk-division. When it has arrived in the fcetal chamber (Brutsack), the yelk exhi- bits the well-known mulberry appearance" {I. c. p. 52). It is unfortunate that these observations are not so precise and detailed as they might have been ; for the question at once suggests itself, is this appearance presented by the ovisac reaUy due to yelk-division ? What has become of the epithelium of the ovisac ? Might not the change in the appearance of the latter be due to an alteration in the character of the epithelium, similar to that which obtains in Fyrosoma ? The next steps in the development of Salpa are, as I pointed out in 1851 {I. c. pp. 575-577), the enlargement of the ovisac, the shortening of its duct, and the consequent approximation of the ovisac to the atrial wall, and, finally, the protrusion of this part of the atrial wall into the atrium, so as to form a chamber containing the ovisac. This the German observers term the " Brut-sack," which may be rendered " foetal chamber." Arrived in the foetal chamber, I have said {I. c. p. 575) that the foetus "becomes divided into two portions, — a larger turned towards the respiratory cavity, and which projects more and more into it, and a smaller, subspherical, turned towards and lying in the cavity of the sinus, and bathed by the parental blood." The former becomes the embryo, the latter the placenta. Leuckart's description comes to the same result, but is much fuller in details (pp. 52, 53):— " When the vitelline mass has increased to about double its primitive diameter, and has become changed by continual division into numerous small division-masses about i^"' in diameter, it loses its spherical form. A circular constriction appears in it, by which its anterior end is marked off as a hiimp-like process. This constriction indicates the boundary between the foetus and the placenta. The placenta is, at first, the more con- siderable of these two parts. It is, one may say, the remains of the vitelline mass (yelk- sac) which is left after the formation of the rudiment of the embryo, and now, instead of being directly applied to developmental purposes, is metamorphosed into an accessory fcetal organ. ... It has been mentioned above that the posterior segment of the yelk, in the foetal chamber, is freely bathed by the blood of the parent. By the delimitation of the embryo, this segment has now become the posterior end of the placenta : at first, as a part AND DEVELOPMENT OF PYROSOMA. 247 of a spheroid, it naturally possessed a convex surface ; but this disappears as soon as the first traces of embryonic deyelopment are visible. The posterior end of the placenta becomes flattened, and its centre acquires a depression, Avhich penetrates deeper and deeper into its substance. The placenta loses its originally solid character, and (even before there is any marked change in the embryo) becomes rapidly metamorphosed into a cupola-like structure, whose internal cavity is connected by its posterior aperture with the circulatory apparatus of the parent, and may be regarded as a sinus for its blood. The inner walls, freely bathed by this blood, exhibit many irregular elevations, which for the most part run, like ribs, from the apex of the cupola to its entrance. Not un- commonly there is also a conical process, which projects from the roof of the cupola for a greater or less distance into the cavity." Vogt (Bilder aus dem Thierleben, p. 79 et seq.) gives an essentially similar account of the development of the placenta of Salpa pinnata *. Eventually the foetus makes its way through the wall of the atrium, and, carrying its placenta with it, lies free in that cavity, whence it must shortly be expelled. On the face of the matter, there appears to be a close analogy between this process and the development of the foetus of Fijrosoma; for the projection of the atrial wall, caused by ovisacs in which the blastoderm is just appearing, may be fairly compared with the com- mencing foetal chamber ; while, if there were only one ascidiozooid instead of four, its relation to the cyathozooid would be very similar to that which the embryo of Salpa has to its placenta. Nor is there wanting a very considerable resemblance in form and character between the cyathozooid and the placenta. But so much remains to be done before the developmental history of Salpa can be said to be fully made out, that I do not know how far these apparent resemblances may be depended upon as afPording evidence of real similarity between the developmental histories of Pyrosoma and of Salpa. Vogt, MiiUer, and Leuckart seem, as little as my- self, to have endeavoured to trace the fate of the ovisac and of its epithelium. And yet, with the development of Tyrosoma before me, it is impossible to arrive at a conclusion in the absence of information on this head. The long retention of the foetus of Salpa in connexion with the parent and nourished by its blood, in contrast to the early separa- tion of the foetus of 'Pyrosoma and the turning of its cyathozooid to account in another way, leads me to conceive that considerable differences will be found in the details of their development, though I suspect further inquiry will prove that, in essentials, they are very similar. * On the other hand, the description and figures by H. Miiller, in the ' Icones Zootomicae ' of Prof. J. V. Cams, tab. 18, lead me to suspect the existence of differences in the development of the placenta in this species. 248 PROFESSOR HUXLEY ON THE ANATOMY EXPIANATION OE THE PLATES. Plates XXX. and XXXI. The following letters and figures have the same signification throughout. a, the test; a', its oral; or, its eloacal fibrillated layer; a^, labial processes of the ascidiozooids ; «*, lip or so-called sphincter of the eloacal aperture ; w', cells of the embryonic test. b, outer tunic. c, inner tunic, or intestinal wall. d, atrial tunic. e, oral aperture. /, tentacular membrane ;/', haemal tentacle. g, g I .anterior muscles ; g^, posterior or atrial muscles. h, peripharyngeal ridge. i, endostyle. j, epipharyngeal folds. k, hypopharyngeal band and sinus ; k\ languets ; h, diapharyngeal band. I, branchial sac; U, horizontal bi'anchial bars; P, perpendicular branchial bars; P, stigmata, m, alimentary canal ; »»', oesophageal aperture ; m^, oesophagus ; m^, stomach ; m'*, intestine. n, tubular organ, probably hepatic. 0, anus. p, mid-atrium ; p^, lateral atria ; p^, atrial aperture. r, heart ; r', sinuses ; r'^, stolons of the adult ascidiarium ; r^, vascular bands, connecting the branchial sac with the parietal sinus ; r*, stolons of the embryonic ascidiarium, or tetrazooid. s, ovisac ; s', its duct ; s^, lining of the ovisac. t, testis ; /', vas deferens ; f, spermatozoa. M, yelk ; m', germinal vesicle ; u^, germinal spot ; m^, contents of the germinal vesicle. w, bud ; w', its peduncle. X, alimentary or trophic blastema of the nascent bud; ^', its generative blastema; x^, its tegumentary blastema. y, the circular cellular patch, probably a renal organ. z, the ganglion ; z^, nerves ; z'^, the ciliated sac ; z^, the tubercle. ce, elaeoblast. en, embryonic endoplasts within the germinal vesicle. ^ bl, blastoderm. cl, cloaca, - I I. II. III. IV. V. Segments of the blastoderm, i. Cyathozooid. ii.-v. Ascidiozooids. 1, 2, .3, 4, isthmuses. y8, the mouth of the cyathozooid, as formed by the test, and which becomes the future lip or ' sphincter' of the eloacal aperture; /3', aperture of the cyathozooid when separated from the foregoing. 7, cavity of the cyathozooid. 8, appendix of the cyathozooid. 6, canal connecting the first isthmus with the cavity of the cyathozooid. Tram,s.Lvrwv. Soo. Vol.JKIJItail). 30, arrrian. sa. 7ra.ns.Xi72n.SooJ^l.^XllI. Txii>.31 TSS^.aii.nai.de'l G-.Ja/rma^.so. AND DEVELOPMENT OF PYROSOMA. 249 Plate XXX. Fig. 1. A side view of a single ascidiozooid from the middle of the ascidiarium oi Pyrosoma (j'ujanteum. Fig. 2. A transverse and vertical section of the middle of the branchial region of a similar ascidiozooid, exhibiting, in addition, the oral aperture, anterior muscles, peripharyngeal ridge, ganglia, and anterior end of the endostyle. There are more vascular bands (j-^) represented than would be seen in any one transverse section. Fig. 3. A transverse and horizontal section of an ascidiozooid without any labial process. It should be understood that s and t he altogether above the intestine, and hence do not interrupt the communication between p and jo' />'. Fig. 4. A vertical section of the wall of the ascidiarium, near the cloacal aperture and including its lip. Fig. 5. View of part of the closed end of the ascidiarium, showing the four ascidiozooids (* * * *) which form its apex. Figs. 6 & 6a. The oral aperture viewed from within and from the side. Figs. 7 & 7«' The atrial aperture viewed under the same aspects. Fig. 8. A transverse and vertical section of the middle of the haemal region of the branchial sac, showing the endostyle and the parts adjacent. Fig. 9. Part of the wall of the middle of the ascidiarium viewed from within, or from the cloacal side. Fig. 10. The cloacal fibrillated layer. Fig. 11. A cell of the general substance of the test. Fig, 12. The nervous ganglion viewed from above. Fig. 13. A section of the body- wall of an adult ascidiozooid, taken through the 'urinary' organ, l'^ indi- cates a minute depression which I was at one time inclined to regard as an aperture into the parietal sinus over this organ ; but I suspect it is only a nascent stigma. Fig. 14. The youngest condition of a bud, before the external tunic is elevated. Figs. 15-20. Successive stages of development of the buds, described in the text. Fig. 21. A bud laid open by a vertical ctit, and exhibiting an interior view of the branchial sac. Fig. 22. A more advanced bud, with a second bud forming in its peduncle. Fig. 23. A still more advanced terminal bud, a second median bud in about the same stage as fig. 20, and a third proximal bud developing in the peduncle and nearly in the same stage as fig. 17, all connected together. Fig. 24. A bud naturally detached from its peduncle, and exhibiting a rudiment of the stolon, r^. Fig. 25. A bud so far advanced as to be connected with the cloaca by its atrial aperture, j3-. Figs. 26 & 27. Very young buds, viewed from the side to which their apices are turned. Figs. 28 & 29. More advanced buds viewed from the haemal side. Fig. 30. A still more advanced bud viewed from the haemal side, and given partly in section, to compare with fig. 3. Plate XXXI. Figs. 1, 2, 3 & 4. Ovisacs containing complete ova in different stages of development. The fraction above each figure gives the greatest diameter in parts of an inch. Fig. 5. An ovisac, torn at one point, but otherwise entire, and allowing the naked germinal vesicle to be seen through its wall. Fig. 6. A similar ovisac opened with needles, and the torn lower portion, to which the germinal vesicle adheres, viewed from within. Fig. 6 a. An enlarged view of the germinal vesicle. Fig. 6 6. A similarly magnified view of the dilated end of the duct of the same specimen, showing the ' plug ' of spermatozoa. VOL. XXIII. 2 L 250 PROFESSOR HUXLEY ON THE ANATOMY, ETC., OF PYROSOMA. Fig. 7. A more advanced ovisac opened and viewed from witiiin, showing the pale germinal vesicle covered by the epithelial coating of the sac. Fig. 8*. An ovum extracted from a younger ovisac than the last. Flo's. 8-Sc. Germinal vesicles containing the characteristic deposit of more advanced stages. Fig. 8 a. A germinal vesicle with a very pale spot. Its contour is rather too well defined in the figure. Fiw, 8 b. Two views of the same germinal vesicle, whose spot has disappeared. The minute vesicular cor- puscles, en, are visible. FitT. 8 c. Front and lateral views of a germinal vesicle in a condition observed once. Fig. 9. Ovisac with vesicular epithelial hning and commencing blastoderm (which is represented rather darker than in nature) in situ and causing the atrial tunic, d, to bulge as a rudimentary ' Brutsack ' or foetal chamber. Fig. 9 a. The blastoderm of a similar specimen enlarged, and viewed through the tunica propria of the ovisac. To avoid confusion, the texture of the vesicular lining is omitted. N.B. The figures 8, 8*, 8a, 8b, Sc, 9«, are drawn to the same scale. Fig. 10. FcEtus, now free in the mid-atrium, with the blastoderm much enlarged and converted into an elongated patch. Fig. 11. Two views of a foetus with the blastoderm divided into five segments, of which the cyathozooid is the largest. Fig. 12. The fourth ascidiozooid of a similar foetus, seen in section, and the fifth from above. Fig, 13. Two views of a foetus whose ascidiozooids half encircle the base of its cyathozooid. Fig. 13 a. A single ascidiozooid (the first) of a similar foetus, seen from the side. Fig. 14. A more advanced foetus, to show the stage in which the ascidiozooids (left in outline) completely encircle the cyathozooid, but still lie below the level of the equator of the ovisac. Fig. 15. One of the most advanced foetuses observed. The remains of the conjoined cyathozooid and ovisac ai-e hidden by the ascidiozooids. Fig. 16. A similar foetus viewed from above, to show the remains of the cyathozooid and the ovisac, as well as the connexion of this with the ascidiozooids, and of these with one another, by the elongated isthmuses. Figs. 17 & 18. Lateral views of the cyathozooid in foetuses J^th and J^^th of an inch in diameter respect- ively. Fig. 19. Lateral view of a foetus ^st of an inch in diameter, to show the manner in which the cloaca is developed by the separation of the test from the combined cyathozooid and ovisac. !I}ra-ns .lin^v.Sac: VoL^ W 0 R p- •A C T¥l C 0 CJEAJ^ regionsofarctic vegetation Thr. IscfOierm^ Lijiea are/ taken fromjDtyvat ToJbl&s.prnntedi for l?i4' Bnif^ Ajt^odatum- ijv 2853. The TemperaJairejt txreiirx/Tahr^^ SccuLp,. [ 251 ] XVII. Ouilines of t/ie Distribution of Arctic Plants. By Jos. D. Hooker, M.D., F.B.S., 8fc. (With a Map. Plate XXXII.) Read June 21st, 1860. I SHALL endeavour in the following pages to comply, as far as I can, with a desire expressed by several distinguished Arctic voyagers, that I should draw up an account of the affinities and distribution of the flowering plants of the North Polar Regions. The method I have followed has been, first to ascertain the names and localities of all plants which appear on good evidence to have been found north of the arctic circle in each con- tinent ; then to divide the polar zone longitudinally into areas characterized by differences in their vegetation ; then to trace the distribution of the arctic plants, and of their varieties and very closely allied forms, into the temperate and alpine regions of both hemispheres. Having tabulated these data, I have endeavoured to show how far their present distribu- tion may be accounted for by slow changes of climate during and since the glacial period. The arctic flora forms a circumpolar belt of 10° tp 14° latitude, north of the arctic circle. There is no abrupt break or change in the vegetation anywhere along this belt, except in the meridian of Bafiin's Bay, whose opposite shores present a sudden change from an almost purely European flora on its east coast, to one with a large admixture of American plants on its west. The number of flowering plants which have been collected within the arctic circle is 762 (Monocot. 214; Dicot. 548). In the present state of cryptogamic botany it is im- possible to estimate accurately the number of flowerless plants found within the same area, or to define their geographical limits ; but the following figures give the best approximate idea I have obtained : — Filices 28 Characeae .... 2 Fungi 200? Lycopodiacese . . 7 Musci 250 Algae 100 Equisetaceae . . 8 Hepaticae 80 Lichenes .... 250 Total Cryptogams 925 „ Phasnogams 762 1687 Regarded as a whole, the arctic flora is decidedly Scandinavian ; for Arctic Scandi- navia, or Lapland, though a very small tract of land, contains by far the richest arctic flora, amounting to three-fourths of the whole ; moreover upwards of three-fifths of the species, and almost all the genera, of Arctic Asia and America are likewise Lapponian, leaving far too small a percentage of other forms to admit of the Arctic Asiatic and American floras being ranked as anything more than subdivisions, which I shall here call districts, of one general arctic flora. Proceeding eastwards from Baffin's Bay, there is, first, the Greenland district, whose flora VOL. XXIII. /i, 2 m 252 DR. HOOKER ON THE DISTRIBUTION OF ARCTIC PLANTS. is almost exclusively Lapponian, having an extremely slight admixture of American or Asiatic types : this forms the western boundary of the purely European flora. Secondly, the Arctic European district, extending eastward to the Obi river, beyond the Ural range, including Nova Zembla and Spitzbergen ; Greenland would also be included in it, were it not for its large area and geographical position. Thirdly, the transition from the com- paratively rich European district to the extremely poor Asiatic one is very gradual ; as is that from the Asiatic to the richer fourth or West American district, which extends from Behring's Straits to the Mackenzie River. Eifthly, the transition from the West to the East American district is even less marked ; for the lapse of European and West American species is trifling, and the appearance of East American ones is equally so : the transition in vegetation from this district, again, to that of Greenland is, as I have stated above, comparatively very abrupt. The general uniformity of the arctic flora, and the special differences between its sub- divisions may be thus estimated : the arctic Phsenogamic flora consists of 762 species ; of these, 616 are Arctic European, many of which prevail throughout the polar area, being distributed in the following proportions through its different longitudes : — Arctic Europe . . 616: Scandinavian forms 586 ; Asiatic and American 30=1 „ Asia .... 233 „ „ 189 „ 44 = 1 „ W. America 364 „ „ 254 „ 110=1 „ E.America 379 „ „ 269 „ 110 = 1 „ Greenland . 207 ,> „ 195 „ 12 = 1 19-57 4-2 2-3 2-4 16-2 This table places in a most striking point of view the anomalous condition of Greenland, which, though so favourably situated for harbouring an Arctic American vegetation, and so unfavourably for an Arctic European one, presents little trace of the botanical features of the great continent to which it geographically belongs, and an almost absolute identity with those of Europe. Moreover, the peculiarities of the Greenland flora are not confined to these ; for a detailed examination shows that it differs from all other parts of the arctic regions in wanting many extremely common Scandinavian plants which advance far north in all the other polar districts, and that the general poverty of its flora in species is more due to an abstraction of arctic types than to a deficiency of temperature. This is proved by an examination of the temperate portion of the Greenland peninsula, which adds very few plants to the entire flora, as compared with a similar area south of any other arctic region; and these few are chiefly arctic plants and almost without exception Arctic Scandinavian species. There is nothing in the physical features of the arctic regions, their oceanic or aerial currents, their geographical relations, nor their temperature, which, in my opinion, at all accounts for the exceptional character of the Greenland flora ; nor do I see how it can be explained, except by assuming that extensive changes of climate, and of land and sea, have exerted great influence, first in directing the migration of the Scandinavian species over the whole polar zone, and afterwards in introducing the Asiatic and American species with which the Scandinavian are so largely associated in all the arctic districts except those of Europe and Greenland. It is inconceivable to me that so many Scan- DR. HOOKER ON THE DISTRIBUTION OF ARCTIC PLANTS. 253 dinavian plants should, under existing conditions of sea, land, and temperature, have not only found their way westward to Greenland, by migration across the Atlantic, but should have stopped short on its west coast, and not crossed to America ; or that so many American types should terminate as abruptly on the west coast of BaflS.n's Bay, and not cross to Greenland and Europe ; or that Greenland should contain actually much fewer species of European plants than have found their way eastwards from Lapland by Asia into Western and Eastern Arctic America ; or that the Scandinavian vegetation should in every longitude have migrated across the tropics of Asia and America, whilst those typical plants of these continents which have found their way into the arctic regions, have there remained restricted to their own meridians. It appears to me difficult to account for these facts, unless we admit Mr. Darwin's * hypotheses, first, that the existing Scandinavian flora is of great antiquity, and that previous to the glacial epoch it was more uniformly distributed over the polar zone than it is now ; secondly, that dming the advent of the glacial period this Scandina- vian vegetation was driven southward in every longitude, and even across the tropics into the south temperate zone ; and that on the succeeding warmth of the present epoch, those species that survived both ascended the mountains of the warmer zones, and also retm-ned northward, accompanied by aborigines of the countries they had invaded during thek southern migration. Mr. Darwin shows how aptly such an explanation meets the difficulty of accounting for the restriction of so many American and Asiatic arctic types to their own peculiar longitudinal zones, and for what is a far greater difficulty, the representation of the same arctic genera by most closely allied species in different lon- gitudes. To this representation, and the complexity of its character, I shall have to allude when indicating the sources of difficulties I have encountered, whether in limiting the polar species, or in determining to what southern forms many are most dkectly re- ferable. Mr. Darwin's hypothesis accounts for many varieties of one plant being found in various alpine and arctic regions of the globe, by the competition into which their common ancestor was brought with the aborigines of the countries it invaded : different races survived the struggle for life in different longitudes ; and these races again, after- wards converging on the zone from which their ancestor started, present there a plexus of closely allied but more or less distinct varieties or even species, whose geographical limits overlap, and whose members very probably occasionally breed together. Nor is the application of this hypothesis limited to this inquiry ; for it offers a possible explanation of a general conclusion at which I had previously arrived t and shall have again to discuss here— viz. that the Scandinavian flora is present in every latitude of the globe, and is the only one that is so ; and it also helps to explain another class of most interesting and anomalous facts in arctic distribution, at which I have now arrived from an examination of the vegetation of the several polar districts, and especially of that of Greenland. * This theory of a southern migration of northern types being due to the cold epochs preceding and during the glacial, originated, I believe, with the late Edward Forbes ; the extended one, of their transtropical migration, is Mr. Darwin's, and is discussed by him in his ' Origin of Species,' chap. xi. t Introd. Essay to the ' Flora of Tasmania,' p. ciii. 2m 2 254 DR. HOOKER ON THE DISTRIBUTION OF ARCTIC PLANTS. A glance at the appended chart shows how this theory bears upon the Greenland flora, explaining the identity of its existing vegetation with that of Lapland, and accounting for its paucity of species, for the rarity of American species, of peculiar species, and of marked varieties of European species. If it be granted that the polar area was once occupied by the Scandinavian flora, and that the cold of the glacial epoch did drive this veo-etation southwards, it is evident that the Greenland individuals, from being confined to a peninsula, would be exposed to very diiferent conditions to those of the great con- tinents. In Greenland many species would, as it were, be driven into the sea, that is, exterminated ; and the survivors would be confined to the southern portion of the penin- sula, and not being there brought into competition with other types, there could be no struo-o-le for life amongst their progeny, and consequently no selection of better-adapted varieties. On the return of heat, these survivors would simply travel northwards, unac- companied by the plants of any other country. In Ai'ctic America and Asia, on the other hand, where there was a free southern extension and dilatation of land for the same Scandinavian plants to occupy, these would multiply enormously in individuals, branching off into varieties and subspecies, and occupy a larger area the further south they were driven ; and none need be altogether lost in the southern migration over plains, though many would in the struggle that ensued when they reached the mountains of those continents and were brought into competition with the alpine plants, which the same cold had caused to descend to the plains. Hence, on the return of warmth, many more Scandinavian species would return to Arctic America and Asia than survived in Greenland ; some would be changed in form, because only the favoured varieties could have survived the struggle ; some of the Alpine Siberian and Rocky Mountain species would accompany them to the arctic zone ; while many arctic species would ascend those mountains, accompanying the alpine species in their reascent. Again, as the same species may have been destroyed in most longitudes, or at most elevations, but not at all, we should expect to find some of those Arctic Scandinavian plants of Greenland which have not returned to Arctic America still lurking in remote alpine corners of that great continent ; and we may account for Draba aurea being confined to Greenland and the Rocky Mountains, Fotentilla trldentata to Greenland and Labrador, and Arenaria Grcenlandica to Greenland and the White Mountains of New Hampshire, by supposing that these were originally Scandinavian plants, which on the return of warmth were exterminated on the plains of the American continent, but found a refuge on its mountains, where they now exist. It appears, therefore, to be no slight confirmation of the general truth of Mr. Darwin's hypothesis, that, besides harmonizing with the distribution of arctic plants within and beyond the polar zone, it can also be made, without straining, to account for that distribution and for many anomalies of the Greenland flora, viz., 1, its identity with the Lapponian ; 2, its paucity of species ; 3, the fewness of temperate plants in temperate Greenland, and the still fewer plants that area adds to the entire flora of Greenland ; 4, the rarity of both Asiatic and American species or types in Greenland ; and 5, the pre- sence of a few of the rarest Greenland and Scandinavian species in enormously remote alpine localities of West America and the United States. DR. HOOKER ON THE DISTRIBUTION OF ARCTIC PLANTS. 255 On the Local Distribution of Flants loitlmi the Arctic Circle. The greatest number of plants occurring in any given arctic district is found in the European, where 616 flowering plants have been collected from the verge of the circle to Spitzbergen, From this region vegetation rapidly diminishes in proceeding eastwards and westwards, especially the latter. Thus, in Arctic Asia only 233 flowering plants have been collected; in Arctic Greenland, 207 species; in the American continent east of the Mackenzie Eiver, 379 species ; and in the area westwards from that river to Behring's Straits, 364 species. A glance at the annual and monthly isothermal lines shows that there is little relation between the temperature and vegetation of the areas they intersect, beyond the general feature of the scantiness of the Siberian flora being accompanied by a great southern bend of the annual isotherm of 32° in Asia, and the greatest northern bend of the same isotherm occurring in the longitude of west Lapland, which contains the richest flora. On the other hand, the same isotherm bends northwards in passing from Eastern America to Greenland, the vegetation of which is the scantier of the two ; and passes to the northward of Ice- land, which is much poorer in species than those parts of Lapland to the southward of which it passes. The June isothermals, as indicating the most effective temperatures in the arctic regions (where all vegetation is torpid for nine months, and excessively stimulated during the three others), might have been expected to indicate better the positions of the most luxuriant vegetation : but neither is this the case ; for the June isothermal of 41°, which lies within the arctic zone in Asia, where the vegetation is scanty in the extreme, descends to 54° jST. lat. in the meridian of Behring's Straits, where the flora is comparatively luxuriant ; and the June isothermal of 32", which traverses Greenland north of Disco, passes to the north both of Spitzbergen and the Parry Islands. In fact, it is neither the mean annual, nor the summer (flowering), nor the autumn (fruiting) temperature that determines the abundance or scarcity of the vegetation in each district, but these com- bined with the ocean temperature and consequent prevalence of humidity, its geographical position, and its former conditions both climatal and geographical. The relations between the isothermals and floras in each longitude being therefore special, and not general, I shall consider them further when defining the different arctic floras. The northern limits to which vegetation extends varies in every longitude; and its extreme limits are still unknown ; it may, indeed, reach to the pole itself. Pha-nogamic plants, however, are probably nowhere found far north of lat. 81°. 70 flowering plants are found in Spitzbergen ; and Sabine and Boss collected 9 on Walden Island, towards its northern extreme, but none on Boss's Islet, 15 miles further to the north. Suther- land, a very careful and intelligent coUector, found 23 at Melville Bay and Wolstenholme and Whale Sounds, in the extreme north of Baffin's Bay (lat. 76°, 77° N.). Parry, James Boss, Sabine, Beechey, and others, together found 60 species on Melville Island, and Lyall 60 on the islands north of Barrow Straits and Lancaster Sound. About 80 liave been detected on the west shores of Baffin's Bay and Davis's Straits, between Pond Bay and Home Bay. To the north of Eastern Asia, again, Seemann collected only 4 species on 256 DR. HOOKER ON THE DISTRIBUTION OF ARCTIC PLANTS. Herald Island, lat. 7l^° N., the northernmost point attained in that longitude. On the west coast of Greenland, Scoresby and Sabine found only 50 between the parallels of 70° and 75° N. ; whilst 150 inhabit the east coast, between the same parallels. The differences between the vegetations of the various polar areas seem to be to a consi- derable extent constant up to the extreme limits of vegetation in each. Thus Hammculus glacialis and Saxifraga flagellaris, which are all but absent in West Greenland*, advance to the extreme north in East Greenland and Spitzbergen. Caltha palustris, Astragalus a,lpimts, Oxytropis JJralensis and nigrescens, Farrya arctica, Sieversia Hossii, Nardosmia corymbosa, Senecio palustris, Deschampsia C(2spitosa, Saxifraga hieraciifolia and Hirculus, all of which are absent in West Greenland, advance to Lancaster Sound and the polar American islands, a very few miles to the westward of Greenland. On the other hand. Lychnis alpina, Arabis alpina, Stellaria cerastioides, Fotentilla tridentata, Cassiopeia hypnoides, Phyllodoce taxifolia, Veronica alpina', Thymus Serpyl- lum, Luzula spicata, and Fhlenm alpimim, all advance north of 70° in West Greenland, but are wholly unknown in any part of Arctic Eastern America or the polar islands. The most arctic plants of general distribution that are found far north in aU the arctic areas are the following ; all inhabit the Parry Islands, or Spitzbergen, or both : — Ranunculus nivalis. auricomus. pygmEeus. Papaver nudicaule. Cochlearia officinalis. Braya alpina. Cardamine bellidifolia, — ■ — pratensis. Draba alpina. androsacea. hirta. muricella. incana, rupestris. Cochlearia anglica. officinalis. Silene acaulis. Lychnis apetala. Arenaria verna. arctica. Stellaria longipes. Cerastium alpinum. Potentilla nivea. frigida. Dryas octopetala. Epilobium latifolium. Sedum Rhodiola. Chrysosplenium alternifolium. Saxifraga oppositifolia. csespitosa. cernua. rivularis. nivalis. stellaris. flagellaris. Hirculus (East Greenland only). Antennaria alpina. Erigeron alpinus. Taraxacum Dens-leonis. Cassiopeia tetragona. Pedicularis hirsuta. sudetica. Oxyria reniformis. Polygonum viviparum. Empetrum nigrum. Salix herbacea. reticulata. Luzula arcuata. Juncus biglumis. Carex fuliginosa (not yet found in Arctic Asia, but no doubt there). aquatilis (do.). * Both were found by Kane's Expedition, but by no previous one. DR. HOOKER ON THE DISTRIBUTION OF ARCTIC PLANTS. 257 Eriophorum capitatum. polystachyum. Alopecurus alpinus. Deyeuxia Lapponica. Deschampsia ceespitosa (East Greenland only). Phippsia algida. Colpodium latifolium. Poa flexuosa. pratensis. nemoralis. Festuca ovina. Of the above, Saxifraga oppositifolia is probably the most ubiquitous, and may be con- sidered the commonest and most arctic flowering plant. The following are also inhabitants of all the five arctic areas, but do not usually attain such high latitudes as the foregoing : — Ranunculus Lapponicus. PolemoniumcaeruleuTn,and vars. (EastGreen- Draba rupestris. land only.) Viola palustris. Pedicularis Lapponica. Honkeneya peploides. Armeria vulgaris. Epilobium angustifolium. Betula nana. alpinum. Salix lanata. Hippuris vulgaris, glauca. Artemisia borealis, ' alpestris. "Vaccinium uliginosum. Luzula campestris. Vitis-idsea. Carex vesicaria. Ledum palustre. Eriophorum vaginatum. Pyrola rotundifolia. Atropis maritima. The absence of Gentiana and Primula in these lists is very unaccountable, seeing how abundant and very alpine they are on the Alps and Himalaya, and Gentiana on the South American Cordilleras also. The few remaining plants, which are all very northern, and almost or wholly confined to the arctic zone, are the following, f indicates those species absolutely peculiar ; ft the only peculiar genus. Ranunculus Pallasii. hyperboreus. Trollius Asiaticus. Corydalis glauca. Cardamine purpurea. Turritis mollis. Cochlearia sisymbrioides. Hesperis Pallasii. fBraya pilosa. Eutrema Edwardsii. Parrya arctica. t arenicola. Odontarrhena Fischeriana. Sagina nivalis. Stellaria dicranoides. Oxytropis nigrescens. Sieversia Rossii. Sieversia glacialis. Rubus arcticus. Parnassia Kotzebuei. Saxifraga Eschscholtzii. serpyllifolia. t Richardsoni. Ccenolophium Fischeri. tNardosmia glacialis. Artemisia Richardsoniana. glomerata. t androsacea. Erigeron compositus. Chrysanthemum arcticum. Pyrethrum bipinnatum. tSaussurea subsinuata. Campanula uniflora. Gentiana arctophila. 258 DR. HOOKER ON THE DISTRIBUTION OF ARCTIC PLANTS. Gentiana aurea. Carex nardina. Eutoca Franklinii. glareosa. Pedicularis flammea. rariflora. tDouglasia arctica. Hierochloe pauciflora. tMonolepis Asiatica. Deschampsia atropurpurea. Betula fruticosa. Phippsia algida. Salix speciosa. Dupontia Fisheri. t — _ glacialis. . Colpodium pendulinum. phlebophylla. fulvum. arctica. latifolium. Orchis cruenta. ttPleuropogon Sabini. Platantliera hyperborea. tFestuca Richardsoni. On the Distribution of Arctic Flowering Flants in various Regions of the Globe. There is but one distinct genus confined to the arctic regions, the monotypic and local Pleuy^opogon Sabini ; and there are but seven other peculiarly arctic species, together with one with which I am wholly unacquainted, viz. Monolepis Asiatica. The remain- ing 762 species are all of them found south of the circle ; and of these all but 150 advance south of the parallel of 40° N. lat., either in the Mediterranean basin, Northern India, the United States, Oregon, or California ; about 50 are natives of the mountainous regions of the tropics ; and just 105 inhabit the south temperate zone. The proportion of species which have migrated southwards in the Old and New World also bear a fair relation to the facilities for migration presented by the difi'erent continents. Thus, Of 616 Ai-ctic European species. Of 233 Arctic Asiatic species, 496 inhabit the Alps, and 210 reach the Altai, Soongaria, &c. ; 450 cross them ; ~ 106 reach the Himalaya ; 126 cross the Mediterranean; 0 are found on the tropical mountains 26 inhabit South Africa. of Asia ; 5 inhabit Australia and New Zealand. Of 379 Arctic East American, Of 346 Arctic West American species, 203 inhabit the United States (of which 274 are north temperate ; 21 are confined to the mountains). 24 on tropical mountains; 34 inhabit tropical American mountains. 37 in south temperate zone. 50 inhabit temperate south America. These tables present in a very striking point of view the fact of the Scandinavian flora being the most widely distributed over the globe. The Mediterranean, South African, Malayan, Australian, and all the floras of the New World have narrow ranges compared with the Scandinavian, and none of them form a prominent feature in any other continent than their own ; but the Scandinavian not only girdles the globe in the arctic circle, and dominates over all others in the north temperate zone of the Old World, but intrudes conspicuously into every other temperate flora, whether in the northern or southern hemisphere, or on the Alps of tropical countries. The severest test to which this observation could be put is that supplied by the Arctic DR. HOOKER ON THE DISTRIBUTION OF ARCTIC PLANTS. 259 Scandinavian forms ; for these belong to the remotest corner of the Scandinavian area, and should of all plants be the most impatient of temperate, warm, and tropical climates. The following will, approximately, express the result : — Total Arctic Scandinavian forms .... 586 Cross Aljjs, &c 480 In North United States and Canada, &c. . 360 Reach South Africa 20 In Tropical America 40 Himalaya, &c 300 In Temperate South America 70 Tropical Asia 20 In Alps of Middle Europe, Pyrenees, &c. . 490 Australia, &c 60 In one respect this migration is most direct in the American meridian, where more arctic species reach the highest southern latitudes. This I have accoimted for (Flora Antarctica, p. 230) by the continuous chain of the Andes having favoured their southern dispersion. But the greatest number of arctic plants are located in Central Europe, no fewer than 530 out of 762 inhabiting the Alps and Central and Southern Europe, of which 480 cross the Alps to the Mediterranean basin. Here, however, their further spread is apparently suddenly arrested ; for though many doubtless are to be found in the Alps of Abyssinia and the western Atlas, these are few compared with what are found further east in Asia ; and fewer still have found their way to South Africa. The most continuous extension of Scandinavian forms is in the direction of the greatest continental extension ; namely, that from the North Cape in Lapland to Tasmania* ; for no less than 350 Scandinavian plants have been found in the Himalaya, and 63 in Australia and New Zealand ; whereas there are scarcely any Himalayan and no Austra- lian or Antarctic forms in Arctic Europe. Now that Mr. Darwin's hypotheses are so far accepted by many botanists, in that these concede many species of each genus to have had in most cases a common origin, it may be well to tabulate the generic distribution of arctic plants as I have done the specific ; and this places the prevalence of the Scandina- vian types of vegetation in a much stronger light : — Scandinavian Arctic Genera in Europe . . 280 Cross Alps (approximately) 260 Found in North United States (approximately) 270 Found in South Africa (approximately) . .110 „ Tropical American Mountains „ . 100 „ Himalaya, &c. „ .... 270 „ Temperate South America „ .120 „ Tropical Asia „ .... 80 „ Alps „ . 280 „ Australia, &c. „ .... 100 The most remarkable anomaly is the absence of Primula in Tropical America, that genus being found in Extra-tropical South America; and its absence in the whole southern temperate zone of the Old World, except the Alps of Java. * The line which joins these points passes through Siberia, Eastern China, the Celebes Islands, and Australia, but the gl.arial migration has no doubt been due south from the arctic and north temperate regions in various longitudes to the Pyrenees, Alps, Carpathians, Caucasus, Asia Minor, Persian and North Indian mountains, kc. The further migration south to the distant and scattered alpine heights of the tropics, and thence to South Australia, Tasmania, and NewJZealand, is, in the present state of our knowledge, to me quite unaccounted for. Mr. Darwin assumes for this purpose a cooled condition of the globe that must have been fatal to all such purely tropical vegeta- tion as we are now familiar with. VOL. XXIII. 2 N 260 DR. HOOKER ON THE DISTRIBUTION OF ARCTIC PLANTS. Thalictrum, Delphmium, Impatiens, Frunus, Circcea, Chrysosplenium, Parnassia, Bupleurum, Jlieracleum, Viburnum, Valeriana, Artemisia, Vaccinium, Bhododeiidron, Fedicularis, and Salix, are all arctic genera found on the tropical mountains of Asia (Nilghiri, Ceylon, Java, &c.), but not yet in the south temperate zones of Asia, and very few of them in Temperate South Africa. There are, however, a considerable number of Scandinavian plants which are not found in the Alps of Middle Europe, though foiuid in the Caucasus, Himalaya, &c. ; and con- versely there are several Arctic Asiatic and American plants found in the Alps of Cen- tral Em^ope, hut nowhere in Arctic Europe. In other words, certain species extend from Arctic America through Central Asia and North India to Central Europe, which do not extend from Arctic America westward to Arctic Em'ope; and there are cer- tain other species which extend from Arctic Europe to the Caucasus and Central Asia, which do neither exist on the Alps of Central Europe nor extend eastward to Arctic America : thus, Common to Arctic Ranunculus nivalis. hyperboreus. TroUius Asiatlcus. Cardamine bellidifolia. Parrya macrocarpa. arctica. Draba alpina. muricella. hirta. rupestris. Eutrema Edwardsii. Silene turgida. Lychnis apetala. Sagina nivalis. Arenaria lateriflora. arctica. Stellaria borealis. ■ huraifusa. longipes. crassifolia. Rubus arcticus. chamaemorus. Rosa blanda. Saxifraga rivularis. nivalis. flagellaris. bronchialis. Ccenolophium Fischeri. Conioselinum Fischeri. Ligusticum Scoticum. Chaerophyllum bulbosum. Eiirope and Temperate Asia, I Cornus suecica. Galium triflorum. Valeriana capitata. Nardosmia frigida. palmata. Chrysanthemum arcticum. Pyrethrum bipinnatum. Artemisia borealis. Antennaria alpina. Seuecio frigidus. Ligularia Sibirica. Aster Sibiricus. Tataricus. Mulgedium Sibiricum. Campanula uniflora. Cassiopeia hypnoides. Cassandra calyculata. Diapensia Lapponica. Rhododendron Lapponicum, Ledum palustre. Gentiana detonsa. Pleurogyne rotata. Myosotis sparsiflora. Eritrichium villosum. Gymnandra borealis. Castilleja pallida. Veronica macrostemon. Pedicularis Lapponica. hii'suta. Sudetica. Pinguicula villosa. 'C, hut not to Alps of Mwope. Naurabergia thyrsiflora. Primula stricta. Sibirica. Koenigia Islandica. Betula alpestris. Salix lanata. myrtilloides. polaris. Picea orientalis. Larix Ledebourii. Platanthera obtusata. Calypso borealis. Sparganium natans. Calla palustris. Luzula arcuata. Juncus biglumis. Carex glareosa. Norwegica. festiva. loliacea. fuliginosa. rariflora. livida. laxa. capillaris. salina. vulgaris. csespitosa. aquatilis. globuiaris. Blysmus rufus. DR. HOOKER ON THE DISTRIBUTION OF ARCTIC PLANTS. 261 Deyeuxia Langsdorffii. Hierochloe alpina. Colpodium latifolium. Colpodium pendulinum. fulvum. Dupontia Fisheri. Alopecurus alpinus. Deyeuxia deschampsioides. Lapponica. strigosa. It is curious to remark how many of these boreal European plants, which are absentees in the Alps, have a very wide range, not only extending to the Himalaya and North China, but many of them all over Temperate North America ; only one is found in the south temperate zone. In the present state of our knowledge we cannot account for the absence of these in the Alps ; either they were not natives of Arctic Europe immediately previous to the glacial period, or if so, and they were then driven soutli to the Alps, they were afterwards there exterminated ; or, lastly, they still inhabit the Alps under disguised forms, which pass for different species. Probably some belong to each of these categories. I need hardly remark that none inhabit Europe south of the Alps, or any part of the African continent. The List of Arctic American and Asiatic species which do inhabit the Alps of Europe, but not Arctic Europe, is much smaller. Those marked t are Scandinavian, but do not enter the arctic circle. Anemone patens. alpina. narcissiflora. fRanunculus sceleratus. fAconitum Napellus. fArabis petrtea. fCardamine hirsuta. Draba stellata. fThlaspi montanum. tLepidium ruderale. tSagina nodosa. fLinum perenne. Phaca alpina. fAstragalus hypoglottis. fSpirasa salicifolia. tPotentilla fruticosa. Potentilla serieea. fCeratophyllum deniersum. Bupleurum ranunculoides. tViburnum Opulus. Galium rubioides. t saxatile. Ptarmica lalpina. Aster alpinus. Gentiana prostrata. Polygonum polymorphum. Corispermum hyssopifolium. Alnus viridis. Pinus cembra. tSparganium simplex. tTypha latifolia. Carex ferruginea. supina. stricta. t pilulifera. fScirpus triqueter. Deyeuxia varia. Sj3artina cynosuroides. tGlyceria fluitans. Hordeum jubatum. Botanical Districts within the Arctic Circle. The following are the prominent features, botanical, geographical, and climatal, of the five districts of the arctic zone : — 1. Arctic Burope. — The majority of its plants are included in the Lapland and Finland floras ; and, owing to the temperature of the Gulf Stream, which Avashes its coasts, Lap- land is by far the richest province in the arctic regions. The mean annual temperature at the polar circle, where it cuts the coast-line, is about 37°, and the June and September temperatures throughout Lapland are 40° and 37° respectively; thus rendering the climate favourable both to flowering and fruiting. Spitzbergen belongs to this flora, as do Nova Zembla and the arctic countries west of the river Obi, which forms its eastern boun- dary ; for the Ural Mountains do not limit the vegetation, any more than do the Eocky Mountains in America. GmeUn observed more than a century ago that the river Obi in lower latitudes indicates the transition longitude from the European to the Asiatic flora. 2n2 262 DR. HOOKER ON THE DISTRIBUTION OF ARCTIC PLANTS. Even in this small area, liowever, there are two floras, corresponding to the Arctic Norwegian and Arctic Russian. The latter, commencing at the "White Sea, though com- paratively excessively poor in species, contains nearly twenty that are not Lapponian, including Bratja rosea, DlantJius alpinus and Seguieri, Sx>ircea chanicedrifolia, Saxifraga Jderacifolia, Sieracleum Sibiricum, Ligularia Sibirica, Ftarmica alpina, Genticma verna, Pleurogyne rotata, and Larix Sibirica. There are further several Scandinavian plants which cross the arctic circle on the east shores of the Wliite Sea, hut do not do so in Lapland, as Athamanta Libanotis, Chrysan- themum Leucanthemum, Bidens tripartita, and others. Iceland and Greenland also hotanically belong to the Arctic Lapland province, but I have here excluded both : the former because it lies to the south of the arctic circle ; the latter Ijecause both its magnitude, position, and other circumstances, require that it should be treated of separately. As far as I can ascertain, 616 species ( ^^ ^ ' ' \ = 1:23) enter the ^ VDicotyledons 433 J / arctic circle in this region, of which 70 advance into Spitzbergen ; but no phsenogamic plant is found in Ross' Islet beyond its northern extremity. The proportion of genera to species 266 : 616 | 1 : 2-3. Of these Arctic-European plants, 453 cross the Alps or Pyrenees to the Mediterranean basin, a few occur on the mountains of Tropical Africa (in- cluding Luzida campestris and Beschampsia ccespitosa), and 23 are found in South Africa. No fewer than 264 species do not enter the arctic circle in any other longitude, and 184 are almost exclusively natives of the Old World, or of this and of Greenland ; not being found in any part of North America ; 24 are confined to Arctic Europe and Green- land. The following Arctic European plants are of sporadic occurrence in N. America : — Ranunculus acris, Rocky Mountains. Genticma nivalis, Greenland and Labrador. Arabis alpina, Greenland and Labrador. Veronica alpina, Greenland and U. States Moun- Lychnis alpina, Greenland and Labrador. tains. Arenaria arctica, Greenland and Rocky Moun- Bartsia alpina, Greenland and Labrador, tains, Pedicularis paluslris, Labrador. verna, Greenland, Arctic Islands, and Primula farinosa, Labrador. Rocky Mountains. Sali.r jjkylicifolia, U. States Mountains. AlchemiUa vulgaris, Greenland and Labrador. arbicscula, Greenland and U. States Moun- Gnaphalium sylvaticum, Greenland and Labrador. tains. siipinum, Greenland, Labrador, and U. Juncus trifidus, Greenland and U. States Moun- States Mountains. tains. Vaccinium myrtillus, Rocky Mountains only. Car ex capitata, Greenland and U. States Moun- Cassiopeia hypnoldes, Greenland, U. States Moun- tains. tains, and Labrador. Phleum alpinum, Greenland, U. States Mountains, Phyllodoce taxifolia, Greenland, U. States Moun- and Labrador. tains, and Labrador. Calainagrostis lanceolata, Labrador. There are besides a considerable number of Arctic European plants, which, in the New World, are confined to Greenland, being nowhere found in East America : these will be enumerated when treatin": of the Greenland flora. DR. HOOKER ON THE DISTRIBUTION OF ARCTIC PLANTS. 263 The plants wliich are Avidely distributed in Temperate America or Asia, l3ut almost exclusively Arctic in Europe, are the following : — Ranunculus Pallasii, Asia and America. Eritrichium arefioides, Asia and America. Trollius Asiaticus, Asia. Gymnandra Pallasii, Asia. Parrya macrocarpa, Asia and America. Castilleja pallida, Asia and America. arctica, Asia and America. Veronica mncrostemon, Asia. Stellaria longipes, Asia and America. Pedicularis Jlammea, America. Potentilla emarginata, America. Pinyuicula villosa, Asia and America. Epilobium latifolium, Asia and America. Eoeniyia islandica, Asia and America. Sedum quadrifidum, Asia. Salix polar is, Asia and America. Saxifraga bronchialis, Asia and America. Picea ori.entalis, Asia. Senecio resedcefolius, Asia and America. Larix Ledebourii, Asia. Ligularia Sibirica, Asia. Platanthera hyperborea, America. Mulgedium Sibiricum, Asia. obtusata, America. Cassiopeia tetragona, Asia and America. Beyevxia Deschampsioides, Asia and N.W. Ame- Gentiana detoiisa, Asia and America. rica. Pleurogyne rotata, Asia and America. Dvpontia Fisheri, America. The works upon which I have mainly depended for the habitats of the Arctic European plants are Wahlenberg's ' Flora Lapponica,' Ledebour's ' Elora E-ossica,' Fries' ' Summa Vegetabilium Seandinavife,' and ' Mantissge,' and various admirable treatises by Anderssou, Nylander, Hartmann, Lindblom, Wahlberg, Blytt, 0. Martins, E^uprecht, and Schrenk. For Spitzbergen plants I have depended on Hooker's enumeration of the Spitzbergen collections made dm-ing Parry's attempt to reach the north pole, Capt. Sabine's collection made in the same island, and on Lindblom and Beilschmied's ' Flora von Spitzbergen ' (Regensburg, Flora, 1842). For the southern distribution of the Arctic European plants, I have further consulted Nyman's excellent ' Sylloge,' Ledebour's ' Flora Rossica,' Grisebach's ' Flora Rumelica,' Grenier and Godi^on's ' Flore de France,' Parlatore's 'Flora Italiana,' Koch's 'Synopsis Flor^ Germanise,' Munby's ' Catalogue of Algerian Plants,' A. Richard's of those of Abyssinia, Visiani's 'Flora Dalmatica,' Delile's 'Flora yEgyptiaca,' Boissier's noble ' Voyage Botanique dans I'Espagne,' and Tchihatcheff's ' Asia Minor,' besides numerous Iftcal floras of the Mediterranean regions, Madeira, the Azores, and Canaries. 2. A7^ctic Asia. — This, which for its extent, contains by far the poorest flora of any on the globe, reaches from the Gulf of Obi eastwards to Behring's Straits, where it merges into the West American. The climate is marked by excessive mean cold ; at the Obi the isotherm of 18° cuts the arctic circle in its S.E. course, and at the eastern extremity of the province the isotherm of 20° cuts the same circle, while the centre part of the district is all north of the isotherm of 9°. The whole of the district is hence far north of the isotherm of 32°, which descends to 52° N.L. in its middle longitude. The extremes of temperature are also very great; the June isotherm of 41° ascending eastward through its Avestern half to the Polar Sea, whilst the September isotherm of 41° descends nearly to 60° N.L. ; whence the low autumn temperature must present an almost insuper- able obstacle to the ripening of seeds within this segment of the polar circle. 264 DR. HOOKER ON THE DISTRIBUTION OF ARCTIC PLANTS. The warming influence of the Atlantic currents being felt no further east than the Obi, and the summer desiccation of the vast Asiatic continent, combine to render the climate of this region one of excessive drought as well as cold ; whence it is in every way most unfavourable to vegetation of all kinds. The total number of species hitherto recorded from this area is 233 /Monocotyledons 42 1 _ , .p\ VDicotyledons 191 J ' / The proportion of genera to species is 1:2. Of the 233 species, 217 inhabit Siberia as far south as the Altai, or Japan, &c. ; 104 extend southwards to the Himalaya or moun- tains of Persia ; 4 are found on the mountains of the two Indian peninsulas, and 8 on those of Australia and New Zealand. All but the following 37 are European. Those marked with a f are almost exclusively arctic. Delphinium Menziesii (West America). fNardosmia glacialis (Arctic Asia only). tCochlearia sisymbrioides (Boreal ditto). Gmelini. Hesperis Pallasii (East and West America). fArtemisia Steveniana (Arctic Asia only). Odontarrhena Fischeriana. glomerata (West America). Cardamine macrophylla. biennis (E. and W. America). fArenaria macrocarpa (West America). Osmothamnus fragrans. laricina. Pedicularis capitata (E. and W. America). t Rossii (Rocky Mountains). euphrasioides (E. and W. America). Cerastium maximum (West America). fMonolepis Asiatica (Arctic Asia only). fOxytropis nigrescens (Boreal E. and AV. America). Rumex salicifolius (E. and W. America). Hedysarum Sibiricum. graminifolius. tSieversia glacialis (Boreal W. America). Salix ovalifolia (West America). Potentilla stipularis. Abies alba (E. and W. America.) fragiformis. Larix Americana (E. and W. America). Claytonia lanceolata. Tofieldia coccinea (E. and W. America). tSedum euphorbioides (Arctic Asia only). rritillaria Kamtchatkensis (West America). Saxifraga Escholtzii (Boi-eal W. America). Carex concinna (West America). Saxifraga serpyllifolia (W. America). Elymus mollis (E. and W. America). punctata (West America). Thus out of 37 non-European species, only 12 are confined to Asia, the remaining 25 being American. On the other hand, there are only 22 European species in Ai'ctic Asia which are not also American ; which scarcely establishes a nearer relationship between Arctic Asia with Europe than with America. These are Dianthus Segiiieri. Leontodon autumnalis. supei'bus. Hieracium alpinum. Silene inflata. Veronica longifolia. Arenaria uliginosa. Pedicularis Sceptrum. Phaca alpina. Pinguicula alpina. Hedysarum obscurum. Polygonum Sibiricum. Rubus Idseus. Salix Lapponum. Sedum quadrifidum. nigricans. Gaya simplex. hastata. DR. HOOKER ON THE DISTRIBUTION OF ARCTIC PLANTS. 265 Picea orientalis. Cypripedium Calceolus. Larix Ledebourii. Carex ferruginea. In other words, of the 233 Asiatic species, 196 are common to Asia and Europe, 22 are confined to Asia and Europe, 25 are confined to Asia and America only ; and 12 are confined to Asia, of which 3 are peculiar to the arctic circle. The rarity of Graminea3 and especially of Cyperacese in this region is its most excep- tional feature; only 21 of the 138 arctic species of these orders having hitherto been detected in it. Cryptogamic plants seem to be even more rare ; Woodsia ilvensis and Lastrea fragrans being the only Filices hitherto enumerated. Eurther researches along the edge of the arctic circle would doubtless add more Siberian species to this flora, as the examination of the north-east extreme would add American species, and possibly lead to the flora of the country of the Tchutchis being ranked with that of West America. The works which have yielded me most information regarding this flora, are Ledebour's ' Elora Rossica,' and the valuable memoirs of Bunge, C. A. Meyer, and Trautvetter, on the vegetation of the Taimyr and Boganida rivers ; and on the plants of Jenissei river in Von Middendorff's Siberian ' Travels '. Eor their southern extension Trautvetter and Meyer's 'Elora Ochotensis,' also in Middendorif's 'Travels'; Bunge's enumeration of North China and Mongolian plants ; Maximovicz's ' Elora Amurensis ;' Asa Gray's paper on the botany of Japan (Mem. Amer. Acad. N.S. vi.) ; Karelin and Kiriloff's enumeration of Soongarian plants ; Eegel, Bach, and Herder on the East Siljerian and Jakutsk coUections of PauUowsky and Von Stubendorff. Eor the Persian and Indian distribution, I have almost entirely depended on the herbarium at Kew, and on Boissier's and Bunge's numerous works. 3o Arctic West America. — The district thus designated is analogous in position, and to a considerable extent in climate, to the Arctic European, but is much colder ; as is indicated both by the mean temperature, and by the position of the June isotherm of 41°, which makes an extraordinary bend to the south, nearly to 52° N. L., in the longitude of Behring's Straits. It extends from Cape Prince of Wales, on the east shore of Behring's Straits, to the estuary of the Mackenzie river, and as a whole it differs from the flora of the province to the eastward of it by its far greater number both of European and Asiatic species, by containing various Altai and Siberian plants which do not reach so high a latitude in more western meridians, and by some temperate plants peculiar to West America. This eastern boundary is, however, quite an artificial one ; for a good many eastern plants cross the Mackenzie and advance westwards to Point Barrow, but which do not extend to Kotzebue's Sound ; and a small colony of Pvocky Mountain plants also spread eastwards and westwards along the shores of the Arctic Sea, which further tend to connect the floras ; such are Aquilegia brevistylis, Sisynibriim humile, Hutck'msia calycina, JleuQhera Micliardsonii, Crepis nana, Gentiana arctophila, Salix speciosa ; none of which are gene- rally diffused arctic plants, or natives of any other parts of Temperate America but the Rocky Mountains. The arctic circle at Kotzebue's Sound is crossed by the isotherm of 23°, and at the 266 DR. HOOKER ON THE DISTRIBUTION OF ARCTIC PLANTS. longitude of the Mackenzie by that of 12° 5' ; whilst the Jnne isotherm of 41° ascends obliquely from S.AV. to N.E., from the Aleutian Islands to the mouth of the Mackenzie, and passes south of this province ; the June and the September isotherms of 41° and 32° both traverse it obliquely, ascending to the N.E. The vast extent of the Pacific Ocean and its warm northerly currents greatly modify the climate of West Arctic America, causing dense fogs to prevail, especially throughout the summer months, whilst the currents keep the ice to the north of Behring's Straits. The shallowness of the ocean between America and Asia, north of lat. 60°, together with the identity of the vegetation in the higher latitudes of these continents, suggests the probability of the land having been continuous at no remote epoch. The number of phsenogamic plants hitherto found in Arctic West America is 364 /Monocotyledons 76 ■!_ _ \ VDicotyledons 288 J ' / The proportion of genera to species is 1 : 1'7. Of these 364 species, almost all but the littoral and purely arctic species are found in West Temperate North America or in the Eocky Mountains, 26 in the Andes of Tropical or Subtropical America, and 37 in Tem- perate or Antarctic South America. Comparing this flora with that of Temperate and Arctic Asia, I find that no less than 320 species are found on the north-western shores and islands of that continent, or in Siberia, many extending to the Altai and the Himalaya. A comparison with Eastern Arctic America shows that 281 are common to it, and the following 38 are found in Temperate, but not Arctic East America. Anemone alpina. Pennsylvanica. Hutchinsia calycina (Rocky Mountains only and Asia). Sisymbrium humile (R. M. and As.). Draba oligosperma (Rocky Mountains only). Lathyrus palustris (Europe, Asia, East and West America). Spirtea salicifolia (Eur., As., E. & W. Am.). Potentilla fruticosa (Eur., As., E. & W. Am.). Pennsylvanica (Eur., As., E. & W. Am.). Comarum palustre (Eur., As., E. & W. Am.). Montia fontana (Eur., As., & W. Am.). Saxifraga Sibirica (Asia and Labrador only). Dahurica (Asia and Rocky Mounts, only), bronchialis (Europe, Asia, & R. Mounts.). Archangelica officinalis (Europe, Asia, America). Ligusticum Scoticum (Eur., As., Am.). Cornus Suecica (Eur., As., Am.). Galium rubioides (Eur., As., Am.). Senecio resedBefolius (Eur., As., Am.). pseudo- Arnica (Asia and America). Cassandra calyculata (Europe, Asia, America). Gentiana arctophila (Rocky Mountains only). prostrata (Europe, Asia, America). tenella (Eur., As., Am.). Veronica scutellata (Eur., As., Am.). Pedicularis palustris (Eur., As., Am.). Atriplex patula (Eur., As., Am.). Corispermum hyssopifolium (Eur., As., Am.). Corallorhiza innata (Eur., As., Am.). Luzula spadicea (Eur., As., Am.). spicata (Eur., As., Am.). pilosa (Eur., As., Am.). Juncus balticus (Eur., As., Am.). Carex lagopina (Eur., As., Am.). Gmelini (America only). cryptocarpa (Europe, Asia, America). stricta (Europe, America). Hierochloe borealis (Europe, Asia, and America). These, it will be seen, are for the most part north temperate plants, common in many parts of the globe, and which are only excluded from Eastern Arctic America by the greater rigour of its climate. DR, HOOKER ON THE DISTRIBUTION OF ARCTIC PLANTS. 267 The best marked European and Asiatic species that are not found further east in Temperate or Arctic America are the following : Anemone narcissiflora. Spiraea chamaedrifolia. Atriplex littoralis. Ranunculus Pallasii. Pyrethrum bipinnatum. Pinus cetnbia. Aconitum Napellus. Gentiana prostrata. Carex Norvvegica. Parrya macrocarpa. Eritrichium aretioides. Deyeuxia strigosa. Dianthus alpinus. Pedicularis verticillata. Langsdoi-ffii. Cerastium vulgatum. Primula nivalis. Colpodium fulvum. Hence it appears that of the 364 species found in Arctic West America, 319 inhabit East America (arctic or temperate, or both), and 320 are natives of the Old World — a diiference hardly suificient to establish a closer affinity of this flora with one continent rather than with the other. The species peculiar to this tract of land are : — Braya pilosa. Artemisia androsacea. Salix glacialis. Saxifraga Richardsoni. Saussurea subsinuata. The rarity of monocotyledons, and especially of the glumaceous orders, is almost as marked a feature of this as of the Asiatic flora : of the 138 arctic species of Glumacece only 54 are natives of West Arctic America. The materials for this flora are principally the plants of Chamisso, collected during Kotzebue's voyage, and described by himself and Schlechtendahl ; Lay and Collie's col- lections, described in Beechey's voyage; the ' Elora Boreali- Americana ; ' and Seemann's plants, described in the ' Botany of the Herald.' Most of the above collections are from Behring's Straits. Eor the arctic coast flora I am mainly indebted to Richardson's researches, and to Pullen's and other collections enumerated by Seemann in his account of the flora of Western Eskimo Land. Eor the southern extension of the flora I have had recourse to the ' Elora Boreali- Americana ; ' Ledebour's ' Elora Bossica,' which includes the Sitcha plants ; the American floras of Nuttall, Pursh, Torrey, Gray, &c. ; and to the collections of Drs. Lyall and Wood formed in Vancouver's Island and British Columbia ; for the Californian, Mexican, and Cordillera floras generally, to the herbarium at Kew> the works above mentioned, and the various memoirs of Torrey and of Gray on the plants of the American Surveying Expeditions. 4. Arctic East America (exclusive of Greenland).— This tract of land is analogous to the Arctic Asiatic in many respects of position and climate, but is very much richer in species. It extends from the estuary of the Mackenzie Biver to Baffin's Bay, and its flora differs from that of the western part of the continent, both in the characters men- tioned in the notice of that province, and in possessing more East American species. The western boundary of this province is an artificial one ; the eastern is very natui-al, both botanically and geographically ; for Baffin's Bay and Davis' Straits (unlike Behring's Strait) have very deep water and different floras on their opposite shores. The arctic circle is crossed in the longitude of the Mackenzie River by the isotherm of 12°, which thence trends south-eastward to the middle of Hudson's Bay ; and in the longi- tude of Davis' Straits it is crossed by the isotherm of 18^°. The June isotherm of 41° VOL. XXIII. 2 o 268 DR. HOOKER ON THE DISTRIBUTION OF ARCTIC PLANTS. descends obliquely from the shores of the Arctic Sea, near the mouths of the Mackenzie, to the northern parts of Hudson's Bay, south of the arctic circle ; and the September isotherm of 41° is everywhere south of the circle. Hence the western parts of this pro- vince are very much warmer than the eastern ; so much so, that the whole west coast and islands of Baffin's Bay lie north of a southern inflection of the June isotherm of 32°, which passes north of all the other polar islands ; the Parry Islands have an analogous temperature of 40°. The warmth of the western portion of this tract is no doubt mainly due to the influence of the Pacific Ocean being felt across the continent of West America ; though possibly also to the presence of a comparatively warm polar ocean, or to Atlantic currents crossing the pole between Nova Zembla and Spitzbergen, of which nothing cer- tain is known *. Be this as it may, the comparative luxuriance of the flora of Melville Island is a well-known fact, and one inexplicable by considerations of temperature, if unaccompanied by a humid atmosphere. The whole region is of course far north of the isotherm of 32°, which, in the longitude of its middle district, descends to Lake Winnipeg, in lat. 52°. That portion of this province which is richest in plants is the tract which intervenes between the Coppermine and Mackenzie Rivers ; east of this, vegetation rapidly diminishes, as also to the northward. The flora of the Boothian Peninsula, surrounded as it is with glacial straits, and placed centrically among the arctic islands, is perhaps the poorest of any part of the area ; those of Banks' Land and Melville Island to the N.W. being con- siderably richer, as are those of the shores of Lancaster's Sound and Barrow's Strait, and the shores of Baffin's Bay to the north and east f. The phgenogamic flora of Arctic East America contains 379 species /Monocotyledons. 92 1 _, . o.-i \ VDicotyledons. . 287 J ' / The proportion of genera to species is 1 : 2-0. Of these 379 species, 323 inhabit Temperate North America, east of the Rocky Mountains ; 35 the Cordillera ; and 49 Temperate or Antarctic South America. Comparing this flora with that of Europe, I find that 239 (or two-thirds) species are common to the arctic regions of both continents, whilst but little more than one-third of the Arctic European species are Arctic East American. Of 105 non-European species in Arctic East America, 32 are Asiatic ; leaving 73 species confined to America, of which the following are furthermore confined to the eastward of the Rocky Mountains and Mackenzie River : — Corydalis glauca. Lathyrus ochroleucus. Vaccinium Canadense. Sarracenia purpurea. Rubus triflorus. Dracocephalum parviflorum. Viola cucuUata. Prunus Virginiana. Douglasia arctica. Silene Pennsylvanica. Heuchera Richardsoni. Elasagnus argentea. Arenaria Michauxii. Cornus stolonifera. Urtica dioica. Polygala Senega. Grindelia squarrosa. Salix cordata. * It is a well-known fact that the temperature always rises rapidly with the north (aswell as other) winds over all this Arctic American area. t Details of these florulas will be found in the volume of the ' Linnean Journal,' under the notice of Dr. Walker's Collections, made during the voyage of the Fox. DR. HOOKER ON THE DISTRIBUTION OF ARCTIC PLANTS. 269 Populus tremuloides. Spiranthes gracilis. Carex oligosperma. Picea nigra. Cypripedium acaule. Pleuropogon Sabini, Of these Douglasia and Fleuropogon are the only ones absolutely peculiar to Arctic East America. It is a noticeable fact that not one of them is found in any part of Green- land. Compared with Greenland, the Arctic East American flora is rich; =containing, besides those just enumerated, no less than 165 other species not found in Greenland. The following are found on the arctic islands, and many of them on the west coast of Baffin's Bay, but not in West Greenland :— Caltha palustris. Parry a arctica. Merkia physodes. Stellaria crassifolia. Astragalus alpinus. Oxytropis campestris. Uralensis. nigi-escens. Sieversia Rossii. Saxifraga hieracifolia. Virginiensis. Hirculus (East Greenland only). Valeriana capitata. Nardosmia corymbosa. Ptarmica vulgaris. Chrysanthemum arcticum. Artemisia vulgaris. Senecio frigidus. palustris. pulchellus. Solidago Virga-aurea. Aster salsuginosus. Crepis nana. Saussurea alpina. Andromeda polifolia. Arctostaphylos alpina. Kalmia glauca. Phlox Sibirica. Pedicularis capitata. versicolor. Androsace septentrionalis. Chamsejasme. Salix phlebophylla. Lloydia serotina. Hierochloe pauciflora. Deschampsia csespitosa (East Greenland only). Glyceria fluitans. Pleuropogon Sabini. Bromus purgans. Elyraus mollis. Castilleja pallida. There are thus no fewer than 184 of the 379 Arctic East American species (fully half) which are absent in West Greenland, whilst only 105 (much less than one-third) are absent in Europe. This alone would make the limitation of species in the mei'idian of Baffin's Bay more decided than in any other arctic longitude ; and I shall show that it is rendered still more decisive by the number of Arctic Greenland plants that do not cross to Arctic East America. Of the 379 Arctic East American species only 56 are not found in Temperate East America, of which two are absolutely confined to this area ; two others {Farrya arenicola and Festuca Richardsoni) to Arctic East and West America ; 25 are found in Temperate West America, and about 20 are Rocky Mountain species, and not found elsewhere in Temperate America. Eor our knowledge of this flora I am principally indebted to the ' Elora Boreali- Ameri- cana,' and to Richardson's* botanical appendix to Franklin's first voyage — and his 'Boat Journey through Rupert's Land.' I have also examined the materials upon Avhich the above works were founded, and the collections of almost every subsequent journey and voyage, up to those of Dr. Walker in the ' Eox.' To enumerate the numerous botanical appendices to voyages, and separate opuscules to which these have given rise, from Ross' first voyage to the present time, would be out of place here. I have endeavoured to embody in the * I am indebted to Sir John Richardson for some corrections to this list, which account for a few discrepancies be- tween his lists of Arctic American plants and my own ; these refer chiefly to genera and species introduced into his lists, but liere excluded. 2 o 2 270 DR. HOOKER ON THE DISTRIBUTION OF ARCTIC PLANTS. essay the information gleaned from all of them. For the southern distribution of these plants in the United States, &c., I have had recourse primarily to Asa Gray's excellent 'Manual of the Botany of the Northern United States,' to Chapman's 'Flora of the S.E. States,' and to the reports on the Botany of various Exploring Expeditions. 5. Arctiv Greenland. — In area Arctic Greenland exceeds any other arctic district except the Asiatic, Init ranks lowest of all in number of contained species. In many respects it is the most remarkable of all the provinces, containing no peculiar species whatever, scarcely any peculiarly American ones, and but a scanty selection of European. A further peculiarity is that the flora of its temperate regions is extremely poor, and adds very few species to the whole flora, and, with few exceptions, only such as are arctic in Europe also. Being the only arctic land that contracts to the southward, forming a peninsula, which terminates in the ocean in a high northern latitude, Greenland offers the key to the explanation of most of the phenomena of arctic vegetation ; and as I have already made use of it for this purpose, I shall be more full in my description of its flora than of any other. The east and west coasts of Greenland differ in many important features ; the eastern is the largest in extent, the least indented by deep bays, is perennially encumbered throughout its entire length by icefields and bergs, Avhich are carried south by a branch of the arctic current that sets between Iceland and Greenland ; and is hence excessively cold, barren, and almost inaccessible. The west coast, again, is generally more or less free from pack ice from Cape Farewell (lat. 60°) to north of Upernsevik in lat. 73°. It is washed by a southerly current, which is said to carry drift timber from the Siberian rivers into its fiords, and enjoys a far milder climate, and consequently has a more luxuriant vegetation. A somewhat similar contrast is exhibited between West Greenland and the opposite shores of Baffin's Bay, against which latter the northerly arctic current from Lancaster Sound drives great masses of polar ice, derived from the regions beyond that estuary, and to which the bergs that float away from the glaciers in the Greenland fiords are also drifted. It is important to bear in mind these features of the two shores of Greenland and of Baffin's Bay and Davis' Straits, because they may in some degree explain their differences of vegetation. There is also another difference between the polar islands and Greenland, inasmuch as the former are for the most part low, without moun- tains or extensive glaciers ; while the latter is exceedingly mountainous, with valleys along the shore terminating in glacier-headed fiords, and the coast is bound by glaciers of prodi- gious extent from Melville Bay northwards to Smith's Sound. The isothermal lines in Greenland all follow one course, from S.W. to N.E., running more parallel to one another in this meridian than in any other. The isotherm of 32° passes through the southern extremity of the peninsula, and that of 6° through its north extreme at Smith's Sound. The June isotherm of 41° skirts its east coast, and that of 32° passes north of Disco ; the June temperature of Disco is hence as low as that of the north of Spitzbergen, of middle Nova Zembla, and of the extreme north of Asia, and yet Disco contains quadruple their number of plants. The autumn cold is very great ; the September isotherm of 32° crossing the arctic circle on the west coast ; and to this the scantiness of the flora may to some extent be attributed. DR. HOOKER ON THE DISTRIBUTION OF ARCTIC PLANTS. 271 The Arctic Greenland flora contains 206 species, according to Lange's catalogue (in Rincke's ' Greenland ') ; or 207, according to my materials ( ^. , ' ^ =1 : 2-1 1 : the VDicot. . 140 J / proportion of genera to species being 1:2. Of these 207 species the following 11 alone are not European : — Anemone Richardsonii (Asiatic). Potentilla tridentata (Labrador only). Turritis mollis (Asiatic). Saxifraga triscuspidata (Labrador only). Vesicaria arctica (American only). Erigeron compositus (American only). Drabaaurea (Rocky Mts. and Labrador only)- Pedicularis euphrasioides (Asia). Hesperis Pallasii (Asia and America). Salix arctica (Asia). Arenaria Graenlandica (Mts. of U. S.). On the other hand, no less than 57 Arctic Greenland species are absent in Arctic East America, and the following 36 Arctic Europe and Greenland species are either absent in all parts of Eastern Temperate America, or are extremely local there : — Arabis alpina (Labrador only). Phyllodoce taxifolia (do. and White Mountains). Lychnis alpina (do.). Gentiana nivalis (Labrador only). Lychnis dioica (absent). Thymus serpyllum (absent). Spergula nivalis (do.). Veronica alpina (White Mountains only). Arenaria uliginosa (absent). saxatilis (absent). ciliata (do.). Euphrasia officinalis (N. U. States). Stellaria cerastioides (do.). Bartsia alpina (Labrador only). Alchemilla alpina (do.). Rumex acetosella (absent). vulgaris (Labrador only). Salix Arbuscula (do.). Sibbaldia pi-ocumbens (United States only). Peristylus albidus (do.). Rubus saxatilis (absent). Carex capitata (White Mountains only). Potentilla verna (Labrador only). microglochin (absent). Sedum villosum (absent). — — microstachya (do.). Saxifraga Cotyledon (Labrador and Rocky Moun- pedata (do.) . tains only). Elyna caricina (Rocky Mountains only). Galium saxatile (absent). Phleum alpinum (Labrador and White Moun- Gnaphalium sylvaticum (Labrador only). tains only). supinum, L. (do. and White Mountains). Calamagrostis lanceolata (Labrador only). Cassiopeia hypnoides (Labrador only). Deschampsia alpina (absent). When it is considered how extremely common most of these plants are throughout Europe and Northern Asia, and that some of them inhabit also N."VV. America, their absence in Eastern America is even more remarkable than their presence in Greenland. Another singular feature of both Arctic and Temperate Greenland is its wanting a vast number of Arctic plants which are European, and found also in America. The following is a list of most of these, excluding about 15, which are water-plants, or species whose range is limited. The letter I. placed before a species signifies that it is Icelandic, and I have introduced it to show not only how many are absent from this island also, but how many are present. The letter S. indicates that the species is found in the south temperate or antarctic circle. The asterisk * indicates that the species is arctic both in East America and Europe. 272 DR. HOOKER ON THE DISTRIBUTION OF ARCTIC PLANTS. Anemone alpina. nemorosa. narcissiflora. * Ranunculus Purshii. * I. Caltha palustris. * Aconitum Napellus. Actaea spicata. Nuphar luteum. Nasturtium amphibium. S. Barbarea prsecox. S. Turritis glabra. Thlaspi montanum. Sisymbrium Sophia, * I. Erysimum lanceolatum. Arabis hirsuta. I. S. Cardamine hirsuta. * Parrya arctica. I. Draba muralis. I. Subularia aquatica. * I. Drosera rotundifolia. I. longifolia. I. Viola tricolor. * I. Arenaria lateriflora. * Stellaria longifolia. I. crassifolia Linum perenne. Geranium Robertianum. Hypericum 4-angulum. Oxalis acetosella. * Phaca frigida. * Astragalus alpinus. * hypoglottis. * Oxytropis campestris. Uralensis. Lathyrus palustris. Spiraea salicifolia. S. Geum urbanum. I. rivale. * Rubus arcticus. Potentilla fructicosa. Pennsylvanica. argentea. * I.S. Fragaria vesca. I. Sanguisorba officinalis. Rosa cinnamomea. blanda. * Circaea alpina. * I. S. Epilobium tetragonum. * I.S. Epilobium alsinsefolium. S. Lythrum salicaria, * Ribes rubrum. * alpinum. * I. Parnassia palustris. Saxifraga Sibirica. * — hieraciifolia. bronchialis. * Bupleurum ranunculoides. Conioselinum Fischeri. Cicuta virosa. * I. Carum caTui. Adoxa moschatellina. Viburnum Opulus. Lonicera cffirulea. * Linn.ta borealis. * I. Galium boreale. rubioides. I. trifidum. S. aparine. * Valeriana capitata. * Nardosmia frigida. * Chrysanthemum arcticum. I. Pyrethrum nodosum. bipinnatum. * Artemisia vulgaris. S. Bidens bipartita. Tanacetum vulgare. Antennaria Carpatica. * Senecio resedaefolius. * frigidus. * palustris. * campestris. aurantiacus. * Solidago Virgaurea. * Aster Sibiricus. * alpinus. S. Erigeron acris. S. Sonchus arvensis. I. Hieracium boreale. * Saussurea alpina. I. Vaccinium myrtillus. * Andromeda polifolia. Cassandra calyculata. * I. Arctostaphylos alpina. * I. Pyrola secunda. * I. Gentiana amarella. * Myosotis sylvatica. palustris. I. arvensis. * Scutellaria galericulata. I. S. Prunella vulgaris. Glechoma hederaceum, S. Stachys palustris. * Gymnandra Pallasii. * Castilleja pallida. I. S. Veronica officinalis. S. scutellata I.S. serpyllifolia. Melampyrum pratense. sylvaticum. * I. Pedicularis palustris. * versicolor. Scrophularia nodosa. Utricularia vulgaris. * Pinguicula villosa. Glaux maritima. Trientalis Europaea. * Androsace septentrionalis * Chamtejasme.. Naumbergia thyrsiflora. I. S. Primula farinosa. I. Plantago major. lanceolata. S. Chenopodium album. I. S. Atriplex patula. Corispermum hyssopifo- lium. * Polygonium Bistorta. I. amphibium. * Myrica Gale. Betula alba. pumila. I. tenella. I. Alnus incana. I. Salix pentandra. I, myrtilloides. I. Triglochin maritimum. Scheuzeria palustris. Veratrum album. * Lloydia serotina. * Allium schsenoprasum. * Srailacina bifolia. * Platanthera obtusata. * Calypso borealis. Godyera repens. DR. HOOKER ON THE DISTRIBUTION OF ARCTIC PLANTS. 273 Cypripedium guttatum. S. Carex Buxbaumii. Eriophorum alpinum. Calla palustris. I. limosa. Rhynchospora alba. Typha latifolia. ' s. Magellanica. Alopeeurus pratensis. Narthecium osslfragum. ustulata. I. Milium efFusum. Luzula maxima. livida. S. Phalaris arundinacea. s Juncus communis. I. pallescens. I.S. Phragmites communis I — — articulatus. maritima. * I. Hierochloe borealis. I bulbosus. I. cagspitosa. * pauciflora. stygius. I. acuta. * I. Catabrosa aquatica. Carex pauciflora. stricta. * I.S. Glyceria fluitans. tenuiflora. filiformis. * I. Atropis distans. s. stellulata. I.S. Eleocharis palustris. I. Festuca elatior. 1. chordorrhiza. s. acicularis. S. Bromus ciiiaris. teretiuscula. s. Scirpus triqueter. I. S. Triticum caninum. paradoxa. s. lacustris. S. Hordeum jubatum. Altogether there are absent in Greenland upwards of 230 Arctic Eurojiean species, which are all of them American plants. The most curious feature of this list is the absence throughout Greenland of the genera Spircea, Senecio, Astragalus, Trifolium, Fhaca, Oxytropis, Androsace, Aster, Myosotis, Rosa, Sides, TJilaspi, Sisymbrium, Geranium, &c., and of such ubiquitous arctic species as Fragaria vesca, Caltha p)alustris *, Barbarea ' prcBCOx. It is remarkable that Astragalinece are also absent from Spitzbergen and Iceland. Iceland possesses 432 species / 97r )' ^-i^o^ioSt which I find about 120 Arctic European plants that do not enter Greenland ; whereas only 50 of the European plants that inhabit Greenland are absent in Iceland. The more remarkable desiderata of Iceland are Astragalinece, Anemone, Aconitum, JBraya, Ttirritis, Artemisia, and Androsace ; Alope- eurus alpinus, Luzula arcuata, Hierochloe alpina, Rubus cliamceomorus, Cassiopeia tetra- gona, Arnica montana, Antennaria dioica, and Chrysosplenium alterni^olimn. On the other hand, Iceland contains of arctic genera absent in Greenland ; Caltha (one of the most common plants about Icelandic dwellings), Cakile, Geranium, Trifoliiim, Spircea, Senecio, and Orchis. But perhaps the most remarkable fact of all connected with the Greenland flora is that its southern and temperate districts, which present a coast of 400 miles, extending south to lat. 60° N.L., do not add more than 74 species to its flora, and these are almost unex- ceptionably Arctic European plants ; and inasmuch as these additional species increase the proportion of Monocotyledons to Dicotyledons of the whole flora, Greenland as a whole is botanically more arctic in vegetation than Arctic Greenland alone is ! The only American forms which Temperate Greenland adds to its flora are, Ranun- ctdus CymbalatHce, Pyrus Americana, a very trifling variety of the Eiu'opean Aucuparia, Viola Muhlenbergii, a mere variety of V. canina, Arenaria Grcenlandica,a plant elscAvhere * This is the more remarkable because it forms a conspicuous feature iu Iceland, and is a frequent native of all the Arctic American coasts and islands. 274 DR. HOOKER ON THE DISTRIBUTION OF ARCTIC PLANTS. found only on the White Mountains of New Hampshire, and Parnassia Kotzebuei, a species which is scarcely diffei'ent from palustris. The only plants which are not members of the arctic flora elsewhere, and which are confined in Greenland to the temperate zone, besides the above American plants, are Blihim glaiicimi, JPotamogeton marinus, Sparganimn minimum, and Streptoptis amplexlfolius : the rest will all be found in the column of the arctic plant catalogue dcA-^oted to Greenland, where S. signifies that the species is found south only of the arctic circle in that country. On the other hand Temperate Greenland adds very materially to the number of Euro- pean Arctic species that do not enter Eastern America (Arctic or Temperate), amongst which the most remarkable are Cerastium viscosura. Galium palustre. Betula alpestris. Vicia cracca. Leontodon autumnale. Juncus trifidus. Rubus saxatilis. Hieracium murorum. squarrosus. Sedum annuum. -■ alpinum. Anthoxanthum odoratum. Galium uliginosum. " Gentiana aurea. Nardus stricta. Another anomalous feature in the Greenland flora is the presence, on the East Arctic coast, of some species not found on the west, nor in the temperate southern end of the peninsula. These are : — Lychnis dioica (Arctic Europe). Saxifraga Hirculus (abundant in all extreme arctic latitudes but West Greenland). Polemonium caeruleum (all arctic longitudes but West Greenland). Deschampsia caespitosa (all arctic longitudes, but also absent in Spitzbergen). Eor data, connected with the Greenland flora, I am mainly indebted to the collections of the various polar voyagers in search of a north-west passage, especially to Drs. Lyall's and Sutherland's ; to Lange's catalogue in Eincke's ' Greenland '; and to the notices of Vahl, Greville, Sir AVilliam Hooker, &c., on the plants collected by Sabine, Scoresby, Ross, Jameson, Graah, and Vahl, &c. ; to Sutherland's appendix to Penny's voyage and Durand's to Kane's voyage. There is a curious affinity between Greenland and certain localities in America, which concerns chiefly a few of the European plants common to these countries. Eirst, there are in Labrador, or on the Rocky Mountains, or White Mountains of New Hampshire, a certain number of European plants found nowhere else in the American continent. They are : — Ranunculus acris (Rocky Mountains). \ Gentiana nivalis (Labrador). Arabis alpina (Labrador). Veronica alpina (White Mountains). Lychnis alpina (Labrador). Bartsia alpina (Labrador). Sibbaldia procumbens (Rocky Mountains). Salix Arbuscula (White Mountains). Potentilla verna (Labrador). Luzula spicata (White Mountains). Montia fontana (Labrador). Juncus trifidus (White Mountains). Gnaphahum sjlvaticum (Labrador). Carex capitata (White Mountains). supinum (Labrador & White Mountains). Kobresia scirpina (Rocky Mountains). Cassiopeia hypnoides (Labr. & White Mounts.). Phleum alpinum (White Mountains & Labrador). Phyllodoce taxifblia (Labr. & White Mounts.). Calamagrostis lanceolata (Labrador). DR. HOOKER ON THE DISTRIBUTION OF ARCTIC PLANTS. 275 There are also three plants peculiar to Greenland and Labrador, or the White or Rocky Mountains, which have not hitherto been found elsewhere. They are ; — Draba aurea (Rocky Mountains). Arenaria Graenlandica (White Mountains and Labrador). Potentilla tridentata (Labrador). On the Arctic Froportions of Species to Genera, Orders, mid Classes. The observations which have hitherto been made on this subject, are almost exclusively based on data collected on areas too small to yield general results. Especially in deter- mining the influence of temperature in regulating the proportions of the great group of flowering plants, it is of the highest importance to take comprehensive areas, both because of the wider longitudinal dispersion of some orders, especially the Monocotyledons, and the effects of local conditions, such as bog land, which determine the overwhelming pre- ponderance of Cyperacece in some arctic provinces compared with others. The proportion of genera to species in the whole arctic phsenogamic flora is 323 : 762, or 1 : 2-8. j.. , ' -. ' ^.^ ; and that of orders to species 1 : 10-8 ; in the several pro- vinces as foUows : — Gen. Gen. to Sp. Orders. Ord. to Sp. Arctic Europe 277 „ Asia 117 „ West America 172 „ East America 193 „ Greenland 104 Thus Europe presents the most continental character in its arctic flora, and West America the most insular ; which may be attributable to the same cause in both ; namely, the uniformity or variety of type. In West America we have, as in an oceanic island, a great mixture of types (Asiatic, European, East and West American) and paucity of species; in Europe the contrary. The proportions of species to orders are still more various ; but here, again, Europe takes the lead decidedly. The proportions of genera and orders to species of all Greenland differ but little from those of its arctic regions ; whereas the contrast between Arctic Eiirope and this, together with Norway as far south as 60° N. lat., is very much greater. This is in accordance with the observation I have elsewhere made, that the whole of Greenland is comparativelj^ poorer in species than Arctic Greenland is. Gen. Sp. Ord. Sp. Gen. Sp. Ord. Sp. Arctic Scandinavia . . . 1:2-3 — 1 : 9*6 Arctic Greenland . . . . 1:2-0 — ^1:5-5 All Scandinavia . . . . 1 : 2-8 — 1 : 11-6 All Greenland 1:2-3—1:6-6 The proportions of Monocotyledons to Dicotyledons are : — Arctic Flora 1:2-6 Arctic Europe 1 : 2*3 „ Asia 1:4-5 „ West America 1 : 3-8 „ East America 1:3-1 „ Greenland . . . • 1:2-1 All Greenland 1:20 VOL. XXIII. . 2 P 2-3 64 9-6 2-0 38 6-1 2-1 48 • 7-6 : 2-5 56 6-8 2-0 38 . 5-5 276 DR. HOOKER ON THE DISTRIBUTION OF ARCTIC PLANTS. Tlie Froportion of largest Orders to the whole Flora. o Arctic Flora „ Europe ,j Asia „ W. America. ,, E. Am«rica . ,, Greenland.. . All Greenland . . . . 1 : 5-C'l :S0-5 1 : 502 1 :27-l|I :33-l!l: lOO'l :26-2ll: l/'Sl 1 : 5-21: 38-4 1 : 56-0,1 : 23-7,l : .30-8 1 : 12-3!l : 34-2!l : 21-2 1 1:10 1: 6 1 : 5 1: 3 1 : 3 Gil : lG-6 1 : 23-3 1 71:24-3!l:52-01 8|l :27-0il :76-0|l 8jl:29-6,l :51-7,1 71 : 34-0 1 : 42-71 16-6,1 : 21-2 1: 33-0 1 :22-7;i : ;34-5|l :23-7il; ; 23-0 1 : 17-3 1: : 24-91 :21-4jl; 9-6 1 : 9-6 1 ; 10-5 1 ; 20-71 15-0 1 15-51 19-1 1 21-0|1 17-2il 27-21 24-6,1 ; 30-8|l; 29-l'l; 28-0'l; : 27-0,1: ; 207-011 ; 149-6 1 15-0 15-4 14-5 15-9 17-2 10-3 12-4 14-1 17- 11-6 18-9 11-91: 10-9 1 12-01 1 17-7 24-6 21-2 17-3 18-9 23-0 27-2 The great differences between these proportions show how little confidence can be placed in conclusions drawn from local floras. EricecB is the only order which is more numerous proportionally to other plants in every province than in the entire arctic flora, and Cruciferoi is the only one that approaches it in this respect ; and LegimiinoscB is the only one which is less numerous proportionally in them all. East and West America agree most closely of any two provinces ; then (excluding Legiiminosce) all Greenland and Europe; next Arctic Greenland and all Greenland. The greatest differences are between Arctic Europe and Asia, and Arctic Asia and West America ; they are less between Arctic Greenland and Asia (excluding ie(/w«iMiOs«) ; they are great between Arctic Greenland and East America ; and as great between all Greenland and Arctic America. The proportion formerly deduced by Brown, &c., for the high arctic regions was a much smaller one ; the Monocotyledons being in comparison with the Dicotyledons 1:5; and this still holds for some isolated, very arctic localities, as North-east Greenland ; whereas Spitzbcrgen presents the same proportion as all the arctic regions, 1:2-7; the Parry Islands 1:2-3; the west coast of Bafiin's Bay, from Pond's Bay to Home Bay, 1:3-3; and the extreme arctic plants mentioned at p. 257., 1:3. Of the prevalent arctic plants mentioned at p. 256, the proportion is 1 : 3-f. I have dwelt more at length on these numerical proportions than their slight impor. tance seems to require ; my object being to show how little mutual dependence there is amongst the arctic florulas. Each has profited but little through contiguity with its cq- terminous districts ; though all bear the impress of being members of one northern flora. On Groujnng the Forms, Varieties, and Species of Arctic Flarbts for purposes of Comparative Study. Considering the limited extent of the arctic zone, the poverty of its flora, which is almost confined to 14° of latitude in the longitudes most favourable to vegetation, and. to only 10° in the Asiatic area, and the number of able botanists who have studied it, it might be supposed that the preliminary task of identifying the species, and tracing their distribution within and beyond the arctic circle would have been short and simple ; but this is not the case ; for owing to the number of local floras, voyages, travels, and scientific periodicals that have to be consulted, to the variability of the species, and the consequent difficulty of settling their limits, and to the impossibility of reconciling the divergent DR. HOOKER ON THE DISTRIBUTION OF ARCTIC PLANTS. 277 opinions of my predecessors regarding them, I have foimd this a very tedious and unsa- tisfactory operation. Of all these sources of douht and error, the most perplexing has been the well-known variability of polar plants ; and in the existing state of the controversy upon Mr. Darwin's hypothesis, it requires to be treated circumspectly. In several genera, I have not only had to decide whether to unite for purposes of distribution dubious or spurious arctic species, but also how far I should go in examining and uniting cognate forms from other countries, which, if included, would materially affect the distribution of the species. These questions became in many instances so numerous and complicated, that I have often resorted to the plan of treating several very closely allied species and varieties as one aggregate or collective species. This appears at first sight to be an evasive course ; but as it offered the only satisfactory method of solving the difficulty, I was obliged, after many futile attempts to find a better, to resort to it, and hence I feel called upon to enter more fully into my reasons for doing so ; premising that all my attempts to treat each variety, form, and subspecies as a distinct plant, involved the discitssion of a multitude of details from which any generalization was hopeless ; the results in every case defeated the object of this paper. Of the plants found north of the arctic circle, very few are absolutely or almost exclusively confined to frigid latitudes (only about 50 out of 762 are so) ; the remainder, as far as their southern dispersion is concerned, may be referred to two classes ; one con- sisting of plants widely diffused over the plains of Northern Europe, Asia, and America, of which there are upwards of 500 ; the other of plants more or less confined to the Alps of these countries, and still more southern regions, of which there are only about 200. Glyceria fiuitans, Atropis maritima, and Senecio campestrls are good examples of the first, as being high arctic and boreal but not alpine ; while most of the species of Saxi- fraga, Draba, and Androsace, are examples of the second*. Both these classes abound in species, the limitation of which within the arctic circle, and the identification of whose varieties with those of plants of more southern countries, present great difficulties. Those plants of the temperate plains which enter the arctic regions are often species of large, widely dispersed, and variable genera, most or all of whose species are very difficult of limitation ; as Rammmihis, of which the arctic species auricoimis, aqitatilis, and acris> are each the centre of a noeud of allied temperate species or varieties, as to whose limits no two botanists are agreed; and the same applies to the species of Viola, Stellaria, Arenaria, and Sieracinm. This has often led to the grouping of names of plants con- sidered as synonymous by some authors, varieties by others, and good species by a third class. Purthermore, such genera are often represented in the temperate regions of two or more continents (and some of them in the south temperate zone also) by closely allied groups of intimately related species. This always complicates matters extremely ; for an arctic species, being generally in a reduced or stunted state, may be equally similar to alpine or reduced forms of what in two or more of these geographically sundered groups * Conversely the only arctic genus unknown in the Alps of the middle temperate zone is Pleuropogon, and the only alpine genera containing several species which inhabit the highest Alps of the north temperate zone, but not the polar regions, are Soldanella in Europe, Swertia in Europe and the Himalaya, &c. 2 P 2 278 DR. HOOKER ON THE DISTRIBUTION OF ARCTIC PLANTS. may rank as good species, and its affinities and distribution be consequently open to doubt. Thus under the arctic Stellaria longijies are included five other arctic forms {lata, Ed- icardsii, pedunctilaris, liehecalyx and ciliatosepala) ; but amongst these forms some speci- mens approach closely the American S. longifolia, Muhl, or slight varieties of it ; while others resemble the European ;S'. Friesiana, Ser., others S. graminea, others certain Tasmanian forms, and others, again, Chilian. My own impression is, that some of these may prove but slight modifications of one common, very widely dispersed plant, between all whose varieties no constant definalile characters will eventually be found ; but in the present state of science I have abstained from including any of them, because to prove this or disprove it, the whole genus wants a far longer and closer study than it has yet received or than I can give it. Arenaria verna and its forms offer a very parallel case, and these I have included more largely, because I have the published opinion of many botanists to bear me out in doing so. Viola ejnj^sila, palustris and blanda, are thus in- cluded, though they are more constant and have to a considerable extent different distri- butions ; because I have found no differences of any moment between their normal forms, because such as exist seem to me to be too slight to attach specific value to ; and because, though well distinguished by Scandinavian botanists, they have not been so carefully col- lected and studied in other parts of the arctic zone. Viola canina, Fragaria vesca, and Sanguisorbla officinalis, afford other examples : all these arctic plants affect the temperate plains rather than the mountains of the northern hemisphere. Turning to those arctic plants that chiefly affect the Alps of the temperate or tropical zones, their limitation is quite as difficult ; alpine plants being as proverbially variable as arctic. Many alpine plants are now considered to be only altered forms of lowland ones ; and this affects the estimated distribution of every arctic species that is identified with an alpine one. As an example, Saxifraga exilis is a very slight variety of S. cermia ; botli are arctic and alpine plants, but S. cermia is considered by some botanists to be an alpine form of the lowland S. granulata, whose limits and distribution are very difficult to settle ; because it apparently passes into several oriental forms, which have been distinguished as species. In this case I have not included S. granulata with S. cernua ; because the latter is everywhere easily distinguished as a well-marked plant, having a restricted range both in area and in elevation, which S. granulata does not share. At the same time I am in favour of a hypothesis that would give these a common origin previous to the glacial epoch. Other reasons for adopting the system of including very closely allied species are the following : — When species have been founded in error ; this generally arises from their authors having imperfect specimens, or too limited a series of them ; various species founded by Brown on the first Arctic American collections come under this category, as do Adams' Arctic Siberian species; the genera Banunciilus, Draba, Arenaria, sindPoten- tilla, offer many examples : when the species, besides belonging to very variable genera, are apparently identical both in the herbarium and according to their descriptions, and present the same or a continuous distribution; of this Trientalis, Senecio, Aster, Erl- geron, Merle iisia, Sedum, Claytonia, Turritis, and many others, afford examples. It may be asked what useful scientific results can ])c obtained from the study of a DR. HOOKER ON THE DISTRIBUTION OF ARCTIC PLANTS. 279 flora whose specific limits are in so yague a condition ? the answer is, that thougli much is uncertain, all is not so ; and that if the species thus treated conjointly really express affinities far closer than those which exist between those treated separately, a certain amount of definite information, useful for my purpose, is obtained ; and it is a matter of secondary importance to me whether the plants in question are to be con- sidered species or varieties. Again, if, with many botanists, we consider these closely allied varieties and species as derived by variation and natural selection from one parent form at a comparatively modern epoch, we may with advantage, for certain purposes, regard the aggregate distribution of the very closely allied species as that of one plant. When sufficient materials shall have been collected from all parts of the arctic and sub- arctic areas, we may institute afresh the inquiry into their specific identity or difference, by selecting examples from physically differing distant areas, and comparing them with others from intermediate localities. An empirical grouping of allied plants for the pur- poses of distribution may thus lead to a practical solution of difficulties in the classification and synonymy of species. My thus grouping names must not therefore be regarded as a committal of myself to the opinion that the plants thus grouped are not to be held as distinct species ; I simply treat of them under one name, because for the purposes of this essay it appears to me advisable to do so. Every reflecting botanist must acknowledge that there is no more equivalence amongst species than there is amongst genera ; and I have elsewhere* endea- voured to show tliat, for all purposes of classification, species must be treated as groups analogous to genera, difi"ering in the number of distinguishable forms they include, and of individuals to which these forms have given origin, and in the amount of affinity botli lietween forms and individuals. My main oliject is to show the affinities of the polar plants, and I can best do this by keeping the specific idea comprehensive. It is always easier to indicate differences than to detect resemblances, and if I were to adopt extreme views of specific difference, I should make some of the polar areas appear to be botanically very dissimilar from others with which they are really most intimately allied, and from which I believe them to liave derived almost all their species. A glance at my catalogue will show that, had I ranked as different species the few Greenland forms of European plants (called generally by the trivial name Grcenlandica), I should have made that flora appear not only more different from the European than it really is,but from the American also ; and that the differences thus introduced would be of opposite values, and hence de- ceptive, in every case when the European species (of which the Grcenlandica is often not even a variety or distinct form) was not also common to America. I wish it then to be clearly understood that the catalogue here appended is intended tcj include every species hitherto found within the arctic circle, together A\ith those most closely allied forms which I believe to have branched off from one common parent within a comparatively recent geological epoch, and that immediately previous to the glacial period or since then. Further, I desire it to be understood that I claim no originality in bringing these closely allied forms together ; from the appended notes, it will be seen that there is scarcely one of them that has not been treated as a synonym, * Essay on the Australian Flora ; introductory to the Flora Tasmanica, p. v. &c. 280 DR. HOOKER ON THE DISTRIBUTION OF ARCTIC PLANTS. variety, subspecies, form, or lusiis, by one or more very able and experienced botanists, some of them by many. Eurthermore, it is curious to observe how much the botanists of each country do to a considerable extent agree amongst themselves as to the specific identity or dilference of tlie same forms — the Scandinavian agreeing with Fries, the German with Koch, and the American with Hooker's ' Flora Boreali- Americana '; also to observe, that in all these cases the authors I quote are independent observers, and not copyers or followers. I think this fact indicates that the same plant presents a diiferent aspect (pro- bably obliterated in drying) in each country. This observation is consonant with what we know of the tendency of all species to run into local varieties in isolated areas, which va- rieties are often appreciable to the eye or to the touch, but are not expressable by words. Of the 762 species enumerated, I have compared arctic or boreal specimens of all but a few which I have indicated in the appended notes, and in most cases I have compared specimens from all the southern areas indicated ; but I do not pretend to have made su.ch a critical study of all the grouped species, or of all those belonging to difficult genera (as Draba, Foa, &c.), as to enable me to say that I have given all their distribution, or satis- fied myself of all their affinities and differences. There are, on the contrary, fully 60 genera out of the 323 arctic ones enumerated, each of which requires careful monograph- ing, and months of study before the limits, systematic and geographical, of its common European species can be ascertained. In two of the largest and most difiicult of these I have been indebted to others ; namely, to Dr. Boott, who has revised my list of Carices, and to Dr. Andersson of Stockholm, who has drawn up that of the Salices : each has ex- tensively modified the conclusions of his predecessors in arctic botany ; quite as much or more so than I have done in any genus, and I have every confidence in their judgment. Colonel Munro has twice revised the list of grasses with a lilve result. In these impor- tant genera, therefore, the groups express the opinions of these acute J)otanists as to the limits of the species. With regard to the probable completeness of our knowledge of the flowering plants of the arctic zone, I think it is pretty certain that there are few or no new species to be discovered. The collectors in the numerous voyages undertaken since 1847 in search of the Eranklin expedition have not added one species to the flora of the Arctic American islands, and but one to that of Arctic Greenland. The Lapponian region is, of course, as well known as any on the globe ; but further east, and especially in Arctic Siberia, much remains to be done ; not perhaps in the discovery of new plants, but in ascertaining the southern limits of various Siberian ones that probably cross the arctic circle. Of Arctic Continental America the same may be said. The method -which I adopted in finally arranging the materials for geographical purposes was the following. I took Wahlenberg's ' Elora Lapponica,' Eries' ' Summa Vegeta- bilium Scandinavise,' Le.debour's 'Elora Eossica,' Hooker's 'Elora Boreali- Americana,' and Lange's ' Plants of East Greenland,' which together embrace in outline almost every- thing we know of arctic botany, geographical, systematic and descriptive. I put together from these all the matter they contained, and arranged it both botanically and geogra- phically into a ' Systema,' which I studied with an Admiralty north circumpolar chart ; and by this means arrived at a general idea of the position and extent of the centres of DR. HOOKER ON THE DISTRIBUTION OF ARCTIC PLANTS. 281 vegetation witliin the polar circle. I then again went through the catalogue with the herbarium, with every work treating on arctic plants that was accessible to me, and lastly revised it, verifying the habitats, comparing specimens from each province, adding new localities from more recent floras, catalogues, and voyages; tracing the extra-arctic distribution of the species, and noting all points requiring further investigation. Tabulated View of Arctic Flowering Flants and Ferns, with their Distribution. In tlie appended table of the distribution of arctic plants I have included every species which I know or believe has been found to extend north of 66^° N. lat. Of these, some few may but just cross the polar circle, and hence be scarcely entitled to the term arctic ; but, on the other hand, there are no doubt temperate species which are entitled to be so considered, but which have not yet been recorded from so high a latitude. This remark applies especially to Arctic Siberia, and possibly also in a less degree to Arctic America, where however Richardson's* conscientious researches, conducted with the view of tracing the polar limits of plants, along the valleys of the great rivers Coppermine and Mackenzie, must have left little of importance to be added or corrected. A cross in a column indicates that the species is naturalized. The columns showing the distribution of each species are so arranged and filled up as to express in the most simple manner the rather complicated directions of migration of each species or group of species. The leading idea, it will be observed, is to demonstrate the influence exerted by the glacial epoch, and^the columns selected and letters introduced are intended to express the apparent results of this influence on each species, such letters indicating physical obstacles to migrations which the species has or has not overcome. I have hence avoided all further complication than appeared to me necessary to illustrate the conclusions I have an-ived at. It would have been easy to have expressed in greater detail the southern and eastern distribution of the species, subspecies, and varieties in the European column for instance ; but I could not have done so in like manner for the same plants in the Asiatic, nor for any but a few species in the West American columns. When the species, subspecies, or varieties grouped under one in these columns shall have been disen- taneled in other countries as they have been in Scandinavia, and when their exact geographical limits also shall be ascertained as accurately in other countries as they have been in Scandinavia, then the time will have arrived for a history of the origin as well as the direction of migration of arctic plants throughout the circle and elsewhere. Meanwhile, as before explained, this essay pretends to no more than laying the foundations of this inquiry on a satisfactory basis. The globe is divided into five principal areas; or rather the species are traced in five directions, as fol- lows : — I. Arctic Distribution. — 1. Arctic European, from Lapland eastward to the Gulf of Obi. An S. in this column indicates that the plant attains the extreme northern limits of phajnogamic vegetation in this district — viz., Spitzbergen. 2. Arctic Asia, from the Obi to Behring's Straits. 3. Arctic Western Ame- rica, from Behring's Straits eastward to the Mackenzie River. 4. Arctic Eastern America, from the Mackenzie to Baffin's Bay. An M. in this column signifies that the plant extends to the islands north of Lancaster Sound, and to the Parry Islands, including Melville Island, the best explored of them. 5. Arctic Greenland. An S. in this column implies that the species has been found south onlj' of the arctic circle in Greenland ; and E. that it is found on the east coast only, the only explored portions of which lie to the north of lat. 70°. II. NoKTH AND Central European and North Asiatic Distribution. — From the arctic circle * Sir John Richardson has had the kindness to revise the list of arctic plants apjiended to the 2nd vol. of liis ' Boat Voyage through Rupert's Land,' &c. ; and I have his authority for excluding any genera and species whicii ajipcar there, and which are not included in my catalogue. 282 DR. HOOKER ON THE DISTRIBUTION OF ARCTIC PLANTS. to the Alps and Pyrenees, Carpathians, Balkan, and Tauria (inclusive) in Europe, and to the Caucasus, Altai and Soongaria (inclusive) in Asia. 1. Europe to the Alps. Here A. implies that the plant in- habits the Alps, or Pyrenees, or Caucasus ; and N. that it is a more northern plant, not advancing south of Germany or Central Russia, in many cases not south of Scandinavia. 2. Asia to the Altai, Soonyaria and Dahuria; the eastern limit of this zone is supposed to be the Jablonoi range, and the western, the Obi River. 3. Eastern Asia, from Behring's Straits to the south of Japan, including Mantchuria, Kamtschatka, the peninsula of Ochotsk, North China, and the neighbouring islands of Japan. This district presents a transition zone between the Asiatic and American floras, and were it better explored might possibly be incorporated with the latter, in the higher latitudes at any rate. III. American Distribution. — 1. North-west America includes the band of country from the Rocky Mountains to the Pacific. This is for the most part a very mountainous area, presenting two floras, — -the Columbian, which extends south to the Oregon ; and the Californian, which extends north to that river, and inland to a considerably higher but still unknown latitude. Being unable to define the limits of these two floras, I have not indicated to which the arctic plants belong, but here, as in the following column, an R. indicates that the species inhabits the Rocky Mountains. 2. North-east America. This includes all temperate North America, from the Rocky Mountains to the Atlantic. In this column the letter C. indicates that the species is Canadian, but does not enter the United States; L. that it is con- fined to Labrador, R. to the Rocky Mountains, and W. that in the United States it h'as been only found on the White Mountains of New Hampshire, or on other high mountains of the Eastern States. 3. Tro- 2)ical America. With fev/ exceptions the arctic plants found in this province are confined to the temperate and alpine regions of the Cordillera from Mexico to Peru inclusive. M. signifies that it has been found in, but not south of, Mexico. 4. Includes Extra-tropical South America and its islands. IV. South Europe and Africa. — 1. Europe south of the Alps, Asia Minor, the Ccmanj Islands, and Africa north of the tropic. In this column are included the plants of the Spanish and Italian penmsulas, Dalmatia, &c., Greece, and both Turkeys. Of these countries, Asiatic Turkey, or Asia Minor, should perhaps have been ranked in the Asiatic column with Persia, &c. ; but it contains so very many European plants that are not found further east, that I have included it here. The letter A. in this column indi- cates that the plant actually crosses the Mediterranean to North Africa, and is found in Algeria, Egypt, the Canary Islands, or the mountains of Abyssinia*. 2. South Africa, from the Cape of Good Hope to the tropic. V. Central and South Asiatic Distribution. — 1. Persia, Tibet, and tlie Himalaya Mountains. In tliis column the letter H. indicates that the species inhabits the Himalaya or Tibet. 2. Tropical Asia. This includes all India south of the Himalaya, and especially the Khasia mountains of Eastern Bengal, the mountains of both peninsulas of India, of Ceylon, and of Java. 3. The last column is confined to Australia, New Zealand, and the antarctic islands to their south. The map that accompanies this paper is reduced from the latest Admiralty chart of the north polar regions (that of 1860), by permission of Captain Washington, R.N., Hydrographer of the Admiralty. The isothermal lines are taken from Professor Dove's Vv'ork ' On the Temperature of the Globe,' translated by Major-General Sabine, and printed by the British Association for the Advancement of Science in 1853. I have selected what appeared to me the seven most instructive isotherms; namely, the annuals of 0° 5' Fahr., 5° and 32°; the two June isotherms of 32° and 41°, and the two September isotherms of 32° and 4 1° ; these two months answering respectively to the flowering and fruiting seasons within the arctic circle. * So fevv Arctic European plants have been found in Tropical Africa, that I have not devoted a column to it. About twenty, however, have been identified either by Achille Richard, by the namers of the Schimper's and Kotsehy's plants, or by myself in the Hookerian Herbarium. Even on Clarence Peak, a mountain on the isolated island of Fer- nando Po, 10,600 feet high, arctic plants have been found by Mann, the collector for Kew, Beschampsia ceespitosu, Luzula campestris, Galium aparine, and Limosella aqiiatica, together with two other boreal species, Sanicula Evropcea and Brachypodium sylvaticum. DR. HOOKER ON THE DISTRIBUTION OF ARCTIC PLANTS. 283 TABULAR VIEW OF THE DISTEIBUTION OF ARCTIC PLANTS. DICOTYLEDONES. I. EajiTuicuIaceae. Thahctrum dioicum, L I. Arctic. II. N. Europ. and Asiatic. III . American. IV. V. •? 1 ■E 1 & 1 2 3 1 S •5 15 g ■E 1 i ■E 1 ■B 1 J 2 Is g.'E 1^ II ■4 < ll "In 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 M 1 1 M 1 1 1 M 1 1 1 1 S S 1 1 s 1 1 1 A A A A A A A A A A A A A A N N A A A N N A A A 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 C 1 1 c 1 1 1 1 1 L 1 1 1 1 1 1 1 1 C E E C 1 1 1 1 1 1 1 X i 1 1 1 1 1 1 1 A 1 A 1 1 A 1 1 A 1 1 1 X H H 1 1 H H 1 H H H 1 H 1 1 1 H 9 1 H 1 1 X alpinum, L. (alp.) 1 1 1 1 TniTniSj T, Kemense, Fr. rariflorum, Fr. striotum, Led. ANEMomB patens, L Nuttalliana, DG. parviflora, Mich Eichardsoni, Hk Vahlii, Horn. alpina, L. {alp.) nemorosa, L 1 1 1 ranunoiiloides, L vernalis, L decapetala, L multifida, Pair. Pennsylvanica, L narcissiflora, L. (alp.) EANimoTJLTJS aquatilis, L 1 1 IS 1 1 1 1 confervoides, Fr. PaUasii, Schl glacialis, L. (alp.) Chamissonis, Schl. Flammula, L reptans, L. Cymbalaria, Psh. (alp.) afflnis, Br. sceleratus, L Purshii, Sich 1 1 1 1 frigidus, DC. Eschsoholtzii, Schl. STilphurens, Sol. repens, L IS 1 1 1 1 1 IS 1 1 1 1 1 polyanthemos, L nemorosus, DC. aoris, L glabrinsculus, Rvpr. Lapponieus, L. (alp.) hyperboreus, Bottb. (alp.) pygmseus, Wahl Sabinii, Br. hispidus, Mich Pennsylvanicus, L 1 1 Eicaria, L 1 1 1 Caltha palustris, L natans, Pall. : radicans, Forst. 1 arctiea, Br. VOL. XXIII. 2 Q 284 DR. HOOKER ON THE DISTRIBUTION OF ARCTIC PLANTS. DICOTYLEDONES. EanrmcTilacese {continued), DEiPHijfiuM ilenziesii, DC. iliddendorffii, Trauiv. elatiim, L. (cdp.) intermediiun, Ait. AcoNiTTTM NapeUus, L lycoctonum, L CoPTis trifolia, Sal. {(dp.) AariLEGiA canadensis, L formosa, Fisch. brevistylis, Hook AcT-EA spicata. L rubra, Willd. arguta, Nutt. alba, Bic/. Teollitjs Europseus, L Asiaticus. L II. Papaveraceaa. Papaveh alpinum, L nndicaule, meet. CoETBALis glaiica, Psh pauciflora, Pers fabacea, Ehr Chelidonitjm majiis, L III. Nymphseacese. Ntmph^a alba, L NrPHAE lutea, Sm intermedia, Led. piimila, Sm. Ealmiana, Mich. lY. Saraceniacese. Saeacenia purjiurea, L v. Cruciferae. NASTUKTiTm paliistre, DC amphibium, Br Bakdahea vulgaris, Br IDrsecox, Br. stricta, Fr. Tttreitis glabra, L mollis. Hook patida, Gr. retrofracta, Wv. Arabis HolbdUii, Hoi-n. Arabis hirsuta, L alpina, L. {alp.) petraea, Lamlc. {alp.) Ijrata, DC. Sisymbrium humifusum, J. VaJd. thpliana, L CAEBAjrofE belUdifolia, L. {alp.) Lenensis, Ad. microphylla, Willd. hirsuta, L . sylvatica, LinJc. amara, L IS n. N. Europ. and Asiatic. 1 i 1 III. American. 1 1 1 C DR. HOOKER ON THE DISTRIBUTION OF ARCTIC PLANTS. 285 DICOTYLEDONES. Cruciferae (continued). Caedamine pratensis, L digitata, Ricli. purpurea, C. 6f S macrophyUa, WiUd Paerta maorocai-pa, Br. (alp.) arctioa, Br. (n?p.) arenioola Eutrema arenioola, Hk. Vesioaria arotica, Hich Deaba alpina, L. (alp.) algida, Ad. pilosa, DC. aspera, Ad. Adamsii, Led. glacialis. Ad. pauciflora, Br. micropetala, Hk. ? trichella, Fr. androsacea, WaM. (alp.) Lapponiea, DC, an Willd.? Wahlenbergii, Hartm. lactea, Adams. nivalis, DC, non Lilj. erassifolia, Grah. ? Martinsiana, J. Gay. oligosperma, Hk. corymbosa, Br. niuricella, WahJ. («?/.).) nivalis, Lilj., non DC stellata, Jacq., non DC Johannis, Host. hirta, L. (alp.) oblongata, Br. Dovrensis, Fr. arctica, YaU. incana, L contOrta, Ehr. confusa, Ehr. (Sf DC. Magellanica, Lam. ? boreaUs, DC rupestris, Br. (alp.) altai'ca, Bunge. aurea, Valil muralis, L CoCHLEAKlA sisjTnbrioides, DC Danica, L AngUca, L oblongifolia, DC. fenestrata, Br. oflScinalis, L Pyi'enaica, DC. Grrenlandica, L. arctica, ScM. Wahlenbergii, Rupr. Lenensis, Adams. Hesperis Pallasii, T. & 0.(alp.) Hookeri, Led. ; pygma3a, Hook. minima, Torr. 6f Gr. II. N. Europ.; and Asiatic, i III. American. IS IS IS IS IS IS IS M i PhW H 2q2 286 DR. HOOKER ON THE DISTRIBUTION OF ARCTIC PLANTS. DICOTYLEDONES. Cruciferse (continued). Odontakehena Fischeriana, G. A.M. SisTMBErtrM Sophia, L sophioides, Fisch. canescens, Nutt braohycarpum, Rich. humile, 0. A. M. (al^y.) salsugineiim, Pall Turritis diffusa, HooTc. Eetsimtot Lieraciifolium, L lanceolatum, Br. alpinum, Fr. cheirantlioides, L Beata aljjina, Sternb, (alp.) glabella, SicJi. Platj^etalum pxirpiirascens, Br. dubium, Br. rosea, Bmu/e pilosa, SooJc EuTEEMA Edwardsii, Br Hutchhtsia calycina, Desv. (alp.). Thlaspi montanum, L arvense, L Capsella bursa-pastoris, L. ........ Lepidium ruderale, L Caeile maritima, L SiifAPis arvensis, L Raphanus Raphanistnim Beassica Rapa, L campestrLs, L. SuBTiLAEiA aquatica, L VI. Droseracese. Deoseea rotimdifolia, L. longifolia, L intermedia, Hayn. . VII. Violariese. Viola palustris, L epipsila, Led. blanda, Willd. microceras, Rupr. canina, L sylvatica, Fr. montana, L. (Fr.) flavicornis, Sm. lactea, Sm. Muchlonbergii, Tor sylvestris. Lam. arenaria, DO. tricolor, L arvensis, DC. bicolor, Pursh. biflora, L. (alp.) cucullata. Ait elatior, Fr M II. N. Em-op. and Asiatic. 1 1 1 1 1 1 I 1 III. American. IV. «.9 wo 1 DR. HOOKER ON THE DISTRIBUTION OF ARCTIC PLANTS. 287 DICOTYLEDONES. VIII. Caryophylleae. DlAHTHTTS alpinus, L. (alp!) repens, Willd. Seguieri, Yill dentosus, Fisch. superbus, L Sheite aoaulis, L. (ctlp.) inflata, L paucifolia, Led. maritima, L turgida, MB Tatarica, Pers rupestris, DC. Pennsylvanica, MicJi nutans, DC. Lychnis apetala, L. {alp.) afflnis, Vahl. triflora, Br. Yahlii, Bupr. angustiflora, Bupr. Magellaniea, Lam, pauciflora, Fisch. dioica, L sylvestris, Sclilc. pratensis, Bohl. alpina, L. (alp.) Flos-Cuculi, L. Spergula arvensis, L Saguta procumbens, L nodosa, E. M. nivalis, Fr CEespitosa, Vahl. intermedia, Fenzl. LinnsEi, Presl saxatilis, Wimm. Abenaria lateriflora, DO. formosa, Fisch. (alp.) nardifolia. Led. ■uliginosa, Schl Lapponioa, Spr. Alsine stricta, Wahl. Rossii, Br Miobauxii, Fenzl stricta, Mich. Tema, L propinqua, Rich. hirta, Wormslc. rubella, Br. quadrivalvis, Br. Giesekii, Horn. arctica, Stev. (alp.) PumUio, Br. biflora, Wahl. serpyUifoUa, L macrocarpa, Psh. (alp.) cUiata, L. (alp.) Norvegiea, Cfunn. laricina, Crantz (alp.) Arenaria Grosnlandica, Spr IS II. N. Europ. and Asiatic. III. American. LW IV. ^ g.' -a H 288 DR. HOOKER ON THE DISTRIBUTION OF ARCTIC PLANTS. DICOTYLEDONES. Caryophyllese (continued). HoNKENEJA peploidcs, Ehr Mekkia physodes, FiscJi Lepigonium salinum, Fr SiELLARiA borealis, Big. (alp.) . . . . crispa, Cham. huinifiisa, Bottb. (alp.) longipea, Goldie (alp.) pcdimcularis. Bye. Iceta, Rich. Edwardsii, Be. hebecalyx, Fenzl. eiliatosepala, Trautv. uliginosa, Miirr nemorum, L media, L graminea, L dici'anoidcs, Fenzl longifoKa, Fries (alp.) alpestris, Fr. longifolia, Milhl. cerastioides, L. (alp.) crassifolia, Elir Ceeastitim alpinum, L. (alp.) . . . . ■viscosiim, L somideoandriun, L. viilgatam, L Fischerianum, Ser. Behringianiim, C. 6f S. maximum, L. (alp>.) arvense, L IX. BalsaminesB. Impatiens fiilva, DO. X. Lineae. LixuM perenno, L catharticum, L XI. Geraniaceae. Geranium sylvaticum, L Eobertianum, L XII. HypericinesB. HypERicrrM perforatum, L quadi'angulum, L. XIII. Elatineae. Elatike Hydropiper, L XIV. Tamariscinese. Myeicahia Germanica, L. (alpi.) XV. OxalidsB. OxALis Acetosella, L XVI. Polygalese. PoLYGALA Senega, Willd vulgaris, L jll. N.Em-op. and Asiatic. a^^ IIT. American. IV. H 1 H H H DR. HOOKER ON THE DISTRIBUTION OF ARCTIC PLANTS. 289 DICOTYLEDONES. XVII. Leguminosse. Phaca frigida, L. {alp.) : 1 alpina, Wulf {alp.) . . aborigmorum, Hk . . AsiKAGALtis alpinus, L. {alp.) 1 hypoglottis, L . . oroboides, Horn, {alp.) ■ 1 OxTTEOPis campestris, DC. {alp.) boreaUs, DC. Middendorfii, Trautv. sordida, Pers. ; polaris, Seem. Uralensis, DC. {alp.) arctica, Br. nigresoens, Fisch. {alp.) deflexa, DC. {alp.) Lapponica, Gaud, {alp.) Hedtsaettii obscurum, L. {alp.) Sibirioum, Poir boreale, Nutt M'Kenziei, Rich TEiroLnJM medium, L pratense, L repens, L Lotus corniculatus, L Ervtjm hirsutum, L Oeobus vermis, L Lathtetjs palustris, L maritimiis, L pratensis, L ochroleuous, Hooh ViciA sylvatica, L Americana, Mdhl Sepium, L Craooa, L Anihtllis vulneraria, L LxjPiNUS percnnis, L XVIII. RosacesB. Spie.s;a chamjedrifolia, L betiila^folia. Pall salioifoKa, L Ulmaria, L pectinata, T. ^ O Alchemilla alpina, L. {alp.) vulgaris, L ' Detas octopetala, L. {alp.) integrifolia, Valil. Drummondii, Rich, {alp.) i Geum iirbanum, L ' strictum, Ait. rivale, L I SiEVEEsiA Rossii, Br ■ I humilis, Br. i glaciaUs, Br. {alp.) SiBBALDiA prooumbens, L. {alp.) Ktjbus areticns, L. {alp.) propinquus. Rich. ; acaulis, Mich. castoreus, Lcestd. Ill 1 ! 1 11 N Europ.i jjj ^^^ and Asiatic. H 290 DR. HOOKER ON THE DISTRIBUTION OF ARCTIC PLANTS. . DICOTYLEDONES. Rosaceae {continued). RuBTJS Chamsemorus, L. {alp.) Idseus, L saxatilis, L. {alp.) tiiflorus, HooTc. saxatili-castoreus, Fr. PoTENTiLLA fruticosa, L. {alp.) stipularis, L TormentOla, Sihth sericea, L Pennsylvanica, L. rubricaalis, Leiim. anserina, L nivea, L. {alp.) Vahliana, L. pulchella, Br. Jamesoniana, Grev. KeiLhaui, Som. frigida, Orev. Grcenlandica, Br. vOlosa, Pall biflora, Lehm frigida, Vill. {alp.) Eobbinsiana, Oalces. emarginata, Psh. nana, LeJim. fragifoiinis, Willd verna, L. {alp.) aurea, L. crocea, Schl. maculata, Lehm. Salisbiirgensis, Hmnke. alpestris, Hall. miiltifida, L. {alp.) argentea, L tridentata, L CoMAjRTOi palustre, L Fragakia vesoa, L SANainsoEBA officinalis, L media, L. polygama, Nyl. Canadensis, L. tenuifolia, Fisch. Rosa cinnamomea, L . majalis, Hermm. blanda, Ait aoicularis, lAndl. Carelica, Fr. fraxinifolia, Lindl. villosa, L tomentosa, Sm. CoioNEASTER Vulgaris, L Pystjs Aueuparia, L Americana, sambucifolia, Gh. 4' ScJi. PEtTNTTs Padus, L Virginiana, DC. spinosa, L AiiELANCHEiR Canadensis, Torr. £/■ Gray alnifoUa, Nutt. IS 11. N Em-op. jjj American, and Asianc. 1 A 1 1g 1 .1 i PhW DR. HOOKER ON THE DISTRIBUTION OF ARCTIC PLANTS. 291 DICOTYLEDONES. XIX. Onagrarieae. CiECiEA alpina, L Epilobium angiistifolium, L latifolium, L. (aJp.) montanum, L tetragonum, L alpinum, L alsinifolium, Vill. Hornemanni, Reich. origanifoliiim, Lam. palustre, L lineare, Miihl. XX. Halorageae. Callitrichb verna, L autumnaKs, L. Mymophtilum spicatum, L altemiflorum, BO HrppTJEis vulgaris, L montana, Led. maritima, Hellen. Ceeatophtllum demersum, L. . . submersum, L. XXI. Ljrthrarieae. Lythetim Salicaria, L XXII. Porttilacese. Clattonia laneeolata, Psh. {alp.") arctica, Ad. sarmentosa, C. A. M, MoNTiA fontana, L rivularis, Gmel. XXIII. Crassulacese. Seduji Rhodiola, DC. {alp.) elongatum, Led. villosiim, L. {alp.) quadrifidum, Pall. {alp.). . . . annmim, L acre, L euphorbioides, SM XXIV. GrossnlariesB. RiBES lacustre, Pursh rubrum, L propinquum, Turc. alpinum, L nigrum, L Hudsonianum, Rich XXV. MiTELLA nuda, L Cheisospleniijm altemifolium, L. tctrandrum, Lund. Paenassia palustris, L obtusiflora, Rupr. Kotzebuei, C. Sf 8 VOL. XXIII. 1 1 1 II. N.Europ.j and Asiatic. III. American. ti H 2 R 292 DR. HOOKER ON THE DISTRIBUTION OF ARCTIC PLANTS. XXVI. Umbelliferae. CENOLOPHiirM Fischeri, Koch BupLEUEUM ranunculoides, L. {alp.) . . CoNiosELHnjM Fischeri, Wimm. (alp.) Tartarieum, Fisch. Selinum Gmelini, De Bray. Aechajtoelica officinalis, DC littoralis, Fr. Norvegica, Tab. atropurpurea, Hoffm. Physolophium sasatUo, Turc. CoeloplemTim Gmelini, Led. Pleurospermum Gmelini, Bong. Angelica sylvestris, L DICOTYLEDONES- s Saxifragese (continued) Saxifkaga Cotj-ledon, L. {aJj}.) , 1 Aizoon, Jacq. oppositifolia, L. ((dp.) IS Esohsolioltzii, Sternh biflora, L. (cdp.) 1 csespitoaa, L. (alp.) IS venosa. Haw. Groenlandica, L. exarata, Vill. Tiniflora, Br. muscoides, Widf. sileniflora, Sternb. MageUanica, Poir. tridactylites, L adscendens, L. (alp.) controversa, Sternh. cemua, L. (cdp.) exilis, DC. Sibirica, L grannlata, L bulbLfera, L. rivularis, L byperborea, Br. Ricbardsoni, Hh nivalis, L. (cdp.) coriacea. Ad. Virgiaiensis, Mich reflexa, Hlc. bieraciifolia, W. ^ K. (alp.). . . Dabnrioa, Pcdl. ((dp.) steUaris, L. (cdp.) comosa. Fair. foliolosa, Br. Hirculus, L propinqua, Br. flageUaris, Wdld. (alp.) bronchialis, L. (alp.) serpyllifoUa, Psh trioiispidata, Retz aizoides, L. (cdp.) punctata, L. (cdp.) sestivalis, Fisch. Hetjcheea Ricbardsonii, Br. (alp.) IS IS IS IS IS IS II. N. Europ. and Asiatic. A 1 1 A ] III. American. R WL 1 H DR. HOOKER ON THE DISTRIBUTION OF ARCTIC PLANTS. 293 DICOTYLEDONES. TJmbelliferse (continued). LiGiJSTicinu: Scoticum, L Athamanta Libanotis, L arctioa, Nijm. sibirica, L. CicuTA virosa, L maculata, DC. Caeum CaiTii, L Peucedan"um palustre, Mcench HERACLErM Sibiricum, L aroticiim, Rupr. Gata simplex, Gaud, (alp.) Paehypleurum alpinum, Led. Seseli divaricatiim, Ptirsh Anthkisctjs sylvestris, L Caa)E0PHTi,LT7ii bulbosum, L XXVII. Corneae. Adoxa mosoliatellina, L. CoENUs stolonifera, Midi. Canadensis, L Suecica, L. (alp.). . . XXVIII. CaprifoliacesB. ViBUEiniM: Opnlus, L Oxycoccos, PsTi. LoNiCEEA caerulea, L Xylosteum, L LrxN^A borealis, L XXIX. Eubiacese. Galium boreale, L rubioides, L uliginosum, L triflorum, MuM trifidum, L Claytoni, Midi. palustre, L Aparine, L infestum, W. 4' K. saxatile, L XXX. Valerianeffi. Valeeiana capitata, Willd. . . . officinalis, L XXXI. Dipsacese. Knatjiia arvensis, Coidt ScABiosA suecisa, L XXXII. Composite. TussrLAGO Farfara, L Naedosmia frigida, Hlc. (alp.) . corymbosa, Hk. sagittata, HJc. glaeiaHs, Led Gmelini, DC "■A ^^'"P- III. American, and Asiatic. M ^. £a\ ^ H 2 R 2 294 DR. HOOKER ON THE DISTRIBUTION OF ARCTIC PLANTS. DICOTYLEDONES. Compositae (continued). Naedosmia laevigata, DC. straminea, Cass. palmata, HJc. {alp.) Achillea Millefolium, L PiAKMiCA alpina, DC Sibirioa, Led. borealis, Bomj. speciosa, DC. multiflora, i/7,-. oxyloba, DC. Cheisanihemum aroticum, L Sibiricum, Turc. integrifolium, Richd Leucantbemum, L PiTSETHEUM iuodonim, Sm ambigiium. Led. pulchellum, Turc. bipinnatum, L Artemisia vulgaris, L var. TiLesii, Led. arctica, Less, (alp.) Chamissoniana, Bess. Norvegica, Fr. glomerata, Led globularia, Cham. androsacea, Seem Eicliardsoiiiaiia, Bess caespitosa, Bess. biennis, Willd desertorum, Spr Canadensis, Mich. Steveniana, Bess borealis. Pall, (alp.) Heleniitm autumnale, Hk BiDENS tripartita, L connata, Mulil. Tanacetum vulgare, L Antennaeia alpina, L. (alp.) dioica, Br hyperborea, Don. Carpatbica, WaM. (alp.) GxAPHALiTTM sylvaticum, L Norvegicum, Gunn. supinum, L. (alp.) uliginosum, L pUulare, Wahl. Aejtica montana, L. (alp.) Cbamissonis, Less. angustifolia, VaJil. alpina, Last. Senecio aureus, L vulgaris, L resedaefolius, Less frigidus. Less palustris, L arcticus, Rupr. Cineraria congesta, Br. II. N. Europ. and Asiatic. III. American. 1 1 DR. HOOKER ON THE DISTRIBUTION OF ARCTIC PLANTS. 295 DICOTYLEDONES. CompositSB (contiHued). Sexecio campestiis, L lugens, liich. Hookeri, T. 4- G. am-antiacus, DC. integrifolius, Nutt. Cineraria alpina, Fr. nemorensis, L octoglossus, DO. pseudo- Arnica, Less LiGULARiA Sibirica, L SoLiDAGO Virga-aurea, L multiradiata, Ait. Canadensis Aster Tripolium, L Sibiricus, L Eicbardsonii, Spr. montanus, BicJm. salsuginosus, likJi. (alp.) pygmajus, Lindl. Tataricus, L. f. alpinus, L. {(dp.) flaccidus, Bunc/e. multiflorus, Ait falcatus, Lindl. Eeigeeon compositus, PursTi (cdp.). . trifidus, Hooh. acris, L elongatus, Led. politus, Fr. giabratus, Hlc. alpinus, L. {aJp.) uniflorus, L. pulcbeUus, DC. Pbiladelphious, L . . purpureus, Ait. GRUfDELiA squari'osa, Duval Taeaxacum Dens-leonis, Desf. . . ceratopborum, DC. palustre, DC. Scorzonera, Reich. pbymatocarpum, J. Vahl. Lapsana communis, L Tkoximon glaucum, Nutt CiJEPis tectorum, L nana. Rich, {alp.) chrysantba, Froel SoNCHUs arvensis, L maritimus, L. Leoniodon autumnalis, L Kei'etinus, Nyl. Apargia Taraxaci, Sm. MuLGEDiuM pulchcllum, Nutt alpinum. Less Sibiricum, Less HiERACiTJM murorum, L ca3sium, Fr. nigrescens, Fr. vulgatum, Fr. atratum, F-. IS n. N.Em-op, and Asiatic. III. American. 1 296 DR. HOOKER ON THE DISTRIBUTION OF ARCTIC PLANTS. DICOTYLEDONES. Compositse {continued). HiEBACiUH Auricula, L dubium, L. alpinum, L. (alp.) triste, Willcl. umbellatum, L crocatum, Fr. boreale, Fr sestivum, Fr. paludosum, L Sausstjeea alpina, L. (alp.) nuda, BO. monticola, Rich. angustifoKa, DC. subsiuuata, Led Cakditus crispus, L C'lEsniJi palustre, Scop heteropliyRmn, All XXXIII. Campanulacese. Campanula rotundifolia, L. . linifolia, Hienlc. Sclieuclizeri, ViU. uniflora, L. {cdp.) dasyantha, M.B lasiocarpa, A. DC latifolia, L XXXIV. Vaccinieae. Vacciniitm uliginosum, L pubeseens, Wormsk. Oxycoocos, L. (cdp.) .... microoarpum, Rupr. Yitis Idoea, L. {alp.) Myi'tillus, L Canadense, Kalm XXXV. Ericeae. Cassiopeia hypnoides, L. {alp.) .... tetragona, L. (alp.) Andeombda polifolia, L Cassandea calyculata, Don (alp.) .... Aeotostaphtlos Uva-ursi, Spr alpina, Spr. (alp.) CAiLirifA vulgaris, L DiAPENSiA Lapponica, L. (alp.) LoisELEUEiA procnmbens, L. (alp.) . . Ehododendeon Lapponicum, L. (alp.) Kalmia glauca, L OsMOTHAMNUs fragraus, DC. (alp.). . . Leduji palustre, L Groonlandioum, Retz. latifoliiun, Ait. dilatatum, Ait. Phtlloboce taxifolia, Sol. (alp.) .... PtPvOla minor, L. 1 1 secunda, L ' 1 II. N. Europ. and Asiatic. III. American. WL L 1 1 1 W WL 1 .. 1 M a is S H -1) DR. HOOKER ON THE DISTRIBUTION OF ARCTIC PLANTS. 297 DICOTYLEDONES. Ericeae (continued), Pybola rotundifolia, L chlorantha, 8w. elliptica, NtM. grandiilora, Bad. GrcBnlandioa, Horn. media, Sw XXXVI. Gentianese. Gentiana camjjcstris, L amarella, L lingulata, Ag. ; acuta, Mich. arctophila, Gi-iseb. (aZp.) aurea, L involuerata, Bottb. glauea. Pall propinqua, RicJi prostrata, Hcenke {aljp.) detonsa, Fr. (alp.) tenella, Fr. {alp.) verna, L. {aJp.) sestiva, B. 4' S. nivalis, L. {alp.) PtEunoGTNE rotata, ft-, {alp.) MentaniblES trifoKata, L XXXVII. HydrophyUesB. Edtoca Eranklinii, Br XXXVIII. Polemoniaceae. PoLEMOjrma: cffinileum, L capitatum, Eschscli. pulchellum, Bge. humile, W. Phlox Sibirioa, L. {alp.) Richardsoni, Rk. XXXIX. Borag^esB. Mtosotis sylvatica, Hoffm alpestris, 8m. sparsiflora, Fold caespitosa, ScJiliz palustris, L arvensis, L intermedia, Linh. Eretkichitim vUlosum, Bge. {alp.) aretioides, A. DC. latifolium, Bupr. Mehtensia maritima, Don denticulata, Don paniculata, Don. corymbosa, Lehm. ; pilosa, DC. Virginica, DC Drummondii, Don. EcHiNosPEEMTOt doflexum, Lehm AsPEEUGO procumbens, L XL. LabiatsB. Mentha arvensis, L Lapponica, WaM. II. N. Europ. and Asiatic. 1 1 III. American. M M H 298 DR. HOOKER ON THE DISTRIBUTION OF ARCTIC PLANTS. DICOTYLEDONES. Labiatse {continued). Oeiganum viilgaro, L Thtmus Serpyllum, L Deacocephalum paniflonim, Nutt. SouTELLAEiA galericulata, L Prunella vulgaris, L Galeopsis versicolor, L Tctrahit, L. Glechoma hederacea, L Stachts sjdvatioa, L paliistris, L Ajuga pyramidalis, L XLI. Orobancheffi. BoscHNiAKiA glabra, C. A. M. . . . XLII. ScrophularineaB. LiJTAEiA vulgaris LiMOSELLA aqiiatica, L GrMN"ANDKA borcalis, Pall, {alp.) . . PaUasii, C. ^ S. Stelleri, C. Sf S. Casiiileja pallida, Kth. {alp.) . . . septentrionaKs, Lindl. Veeonica alpina, L. (alp.) officinalis, L longifolia, L chamiEdrj's, L scuteEata, L macrostemon, Bge serpyllifolia, L borealis, Lcesi. sasatilis, L. {alp.) fniticulosa, L. Melampybum pratense, L Amerieanum, Mich. sylvaticmn, L Eupheasia ofiiciaalis, L EHnfANTHUs Crista-galli, L minor, Ehr. Baetsia alpina, L. {alp.) PEDicirLAKis capitata, Ad Sceptrum, L verticillata, L. {alp.) amoena, Ad. palustris, L borealis, Stev. Lapponica, L. {alp.) euphrasioides, Stev. (alp.) . . . hirsuta, L. {alp.) lanata, Willd. Sudetica, L. {al/>.) arctica, Br. Kanei, Durand. Langsdorfiii, Fisch. flammea, L. {alp.) versicolor, Wald. {alp.) ScEOPnuLAEiA nodosa, L, IS II. N. Europ, and Asiatic. III. American. WL IT. DR. HOOKER ON THE DISTRIBUTION OF ARCTIC PLANTS. 299 DICOTYLEDONES. XLIII. Lentibularineae. Utriculaeia vulgaris, L minor, L intermedia, Ehr. PiNaincTJLA vulgaris, L villosa, L alpina, L. (alp.) XLIV. PrimuIacesB. DoDECAiHEON Meadia, L integrifoHum, Mich. frigidum, C. & S. CoETUSA MatthioH, L Glatjs maritima, L Teienialis Europaea, L latifolia, Hook. ; arctica, Fiscli. Andeosace septentrionalis, L Chamsejasme, L. (cdp.) triflora, Adams. DouGLAsiA arctica, Hk. (alp.) Naitmbergia thyrsiflora, Eeioh Lysimachia vulgaris, L Pbimula stricta, Horn, (alp.) Hornemanniana, Lehm. Mistassinica, C. &S. borealis, Biiby. nivalis. Pall, (all).) cuneifoUa, Led. (a?^.) saxifragaefoLia, Lehm. farinosa, L Scotica, Hook. Sibirica, Jacq. (alp.) Finmarchica, Willd. XLV. Plumbagineae. Aemeeia vulgaris, WilJd alpina, Hoppe. elongata, Hoffm, Labradorioa, Wallr. arctica, Rupr. Macloviana, Cham. XLVI. PlantagineEB. Plantago major, L media, L lanceolata, L maritima, L LiTiouELLA lacustris, L XLVII. Polygonea. KcENiGiA Islandica, L. (alp.) . . . . OxTRiA reniformis, Hk. (alp.) . . Rdmex Acctosa, L Acetosella, L graminifolius, Lamb aquaticus, L Hippolapathum, Fr. arcticus, Trautv. domesticus, Hartm. VOL. XXIII. II. N. Europ. and Asiatic. 1 1 1 1 III. American. Mtr I fi^ 2 s 300 DR. HOOKER ON THE DISTRIBUTION OF ARCTIC PLANTS. DICOTYLEDONES. Polygoniss (continued). RuMEX salicifolius, Weinm PoLTGONUM Bistorta, L viviparum, L. (alp.) poljTaorpluim, Led. (alp.) . . . alpiniun, All. amphibiuni, L Persiearia, L lapatMfolium, L. Sibiricum, Lacnn. (alp.) aviculare, L Convolvulus, L XLVIII. Clienopodiese. Chenopodic-ji album, L maritimum, L MojToiEPis Asiatica, F. 4' M. Atkiplex littoraKs, L patula, L deltoidea. Bah. angustifoHa, Sm. Gmelini, C. A. M. hastata, L. Cop.isPEKMTJM hyssopifolium, Ant. Juss. XLIX. El^agnus argentea, L Shepheedia Canadensis, Nutt. . L. Thymeleae. Daphne ilezereum, L LI. Santalaceje. Comandra li^ida, Ru-h LII. Empetrese. EjiPETHUii nigrum, L LIII. Myrices3. Mypjca Gale, L LIV. Urtice^. TTktica urens, L dioica, L. gracilis, Ait. ■LV. Betulacese. Betula alba, L glutinosa, Wallr. intermedia, Walil. papjTacea, Ait nana, L. pumila, L glandulosa, Midi. fruticosa. Pall alpestri,s, Fr. humilis. Hart. Alnus glutinosa, Willd barbata, C. A. M. IS II. N. Eiu-op. and Asiatic. III. American. DR. HOOKER ON THE DISTRIBUTION OF ARCTIC PLANTS. 301 DICOTYLEDONES. Betolaceae {continued). AiUTTS viridis, DC. {cdp.) fruticosa, Rich. repens, Vahl. incana, Willd LVI. Cupuliferas. CoETLTJS Avellana, L LVII. Salicineae. Salix pentandra, L lucida, WilU. lanata, L. (alp.) Eichardsoni, Hool'. Barattiana, Hoolc. speciosa, H. ^ A. (cdp.) Lapponum, L. (alp.) Stuartiana, Sm. Caprea, L nigricans, Fr punctata, Wahl. phylicifolia, Sm. {alp.) myrtilloides, L pedicellaris, Pursh. . hastata, L cordata, Midil rhamnifolia, E. ^ A ovalifolia, Trautv glacialis. And Arbuscula, L. {alp.) glauca, L. {alp.) desertomm, Bich. villosa, Don. aretica, Br. {alp.) alpestris, And cordifolia, Pursh (partim). pyi-enaica, Gmtan (partim). myrsinitcs, L. (alp.) plilebophj-lla, And reticulata, L. (alp.) vestita, Pursh. nivalis, Hook. herbacea, L. (alp.) polaris, L. (alp.) PoPTJLTJS tremiila, L tremuloides, Mich balsamifera, L LVIII. Conifers. PiNTTS sylvestris, L Banksiana, Lamb Cembra, L. (alp.) Abies alba, L PiCEA nigra, L excelsa, DO orientalis, L Lasix Ledebourii, Endl Sibirica, Led. IS II. jSr.Eui-op. and Asiatic. III. American. 2 s 2 302 DR. HOOKER ON THE DISTRIBUTION OF ARCTIC PLANTS. DICOTYLEDONES. Coniferae (continued). Laeix Americana, Mich niicrocarpa, Lamb. pendula. Ait. Dahiirioa, F. 4' T. JuNiPEEtJs commiuiis, L nana, Willd. Virginiana, L prostrata, Pers. MONOCOTYLEDONES. LIX. Alismace83. Sagittaeia sagittsefolia, L variabilis, Engehn. Alisma Plantago, L LX. Fluviales. Teiglochih maritimum, L paliistre, L ScHETJCHZEEiA paliistris, L PoTAMOGEiON natans, L sparganifolius, Lcestacl. perfoliatiis, L pectinatiis, L rnfescens, ScJir prselongus, Wiilff. piisillus, L tenuissimus, M. K. gramineiis, L nigrescens, Fr. heterophyllus, Schreb. ZosTEEA marina, L LXI. Melauthacese. ToFiELDiA palustris, L. (cdjy.) .... borealis, WaM. cocoinea, Richards ZiGADENtJS chloranthus, liieh Veeatetjm album, L viride, Ait. Lobelianum, Bernh. LXII. Liliacese. Feitillaeia Kamsohatkensis, Gawl. LiOTDiA serotina, L. (alp.) Allium Schoenoprasum, L Sibirioiim, L. oleraoeum, L LXIII. Smilaceae. Paeis qiiadrifolia, L PoLTGONATTJM verticiUatum, All. SuiLACijfA bifolia, Besf. LXIV. OrcMdeae. Oechis maculataj L. omenta, Muell. . latifoHa, L III. American. H H H H H DR. HOOKER ON THE DISTRIBUTION OF ARCTIC PLANTS. 303 MONOCOTYLEDONES. Orchidese (continued). Oechis Sambucina, L Gtmnabenia Conopsea, Br Peeistyiits albidus, L viridis, L bracteatus. Ton-. Islandiciis, Lindl. PtATANTHEEA hyperborca, Lindl dilatata, Lindl. Kcenigii, Lindl. obtiisata, L bifolia, L Epipactis latifolia, Sw media, Fr. HEBMnmjM Monorchis, Br alpinum, L. {cdp.) Malaxis paludosa, L Calypso borealis, L LiSTEKA cordata, Br ovata, L GooDTEEA repens, L CoKALLOREHizA innata, L Spiranthes graciKs, Br CTPEiPEDruTj guttatum, Sw humile, Salisb acaule, Ait. Caleeoliis, L LXV. Iridea. SisTRiNCHnjM Bermudianum, L. . . . anceps, Cav. LXVI. Hydrocharidese. Stratiotes aloides, L LXVII. Aroidese. Spaeganium natans, L hyperboreum, Last. simplex, Sni Caha palustris, L Typha latifolia, L LXVIII. JuncesB. Naethecitjh ossifragum, L Americanum, Ker. LrzuLA spadicea, DC. glabrata, Hoppe. parviflora; Desv. melanocarpa, Desv. Wahlenbergii, Bupr. campestris, Sm pallescens, Wahl. multiflora, hhr. spicata, Desv. (alp.) 1 arcuata, Hook, (alp.) IS hyperborea, Br. pilosa, WiUd 1 vemalis, DC. maxima, L } 1 1 1 1 1 M II. N. Europ. and Asiatic. III. American. H ,. H 304 DR. HOOKER ON THE DISTRIBUTION OF ARCTIC PLANTS. MONOCOTYLEDONES. Junceae {continued). JuNcus bigliimis, L. {alp.) trighimis, L. {alp.) stygius, L. {alp.) castaneus, L. {alp.) communis, Ehr eonglomeratus, L. ; cifusus, L. Baltious, Wnid. (alp.) arcticiis, Willd. {alp.) filiformis, L ti'ifidus, L. {alp.) squarrosus, L bufonius. L polycephalus, Midi articulatus, L iiliginosus, Both. lampocarpus, Ehr. alpimis, Vill . bulbosus, L compressus, Jacq. Gerardi, Loisel Botlmicus, Wahl. coenosus, BicTien. atro-fuscus, Bupr. LXIX. Cyperacese. Cakex clioica, L parallela, Som. gynocrates, Wimm. rupestris, All. {alp.) nardina, Fr Hepburnii, Boott. capitata, L paueiflora, Light microstachya, Ehr seirpoidea, Mx "VVormskioldiana, Horn. tenuiflora, Wahl canescens, L ourta, Good. ; yitilis, Fr. glareosa, Wahl ursina, Dewey. Norvegica, W. elongata, L Heleonastes, Ehr marina, Dew. lagopina, Wahl. {alp.) festiva. Dew. {alp.) leporina, L steUuIata, Good echinata, Murr. loliacea, L macilenta, Fr. incurva, Light fulvicoma. Dew, stcnophylla. Light arctica, Deinh. DeinboIKana, Gay. duriuscula, C. A. M. II. N. Em-op. and Asiatic. III. American. DR. HOOKER ON THE DISTRIBUTION OF ARCTIC PLANTS. 505 MONOCOTYLEDONES. Cjrperacese (continued). Cakex chordorrhiza, Ehr teretiusoula, Good paradoxa, Wahl alpina, Sw. (alp.) Valilii, ScM. liolostoma, Drej. atrata, L. {cdj}.) Buxbaiimii, Wahl Gmelini, Hooh fuliginosa, St.(^ Hpe. (alp.) inisandi'a, Br. frigida, All. (alp.) ferruginea, Scop, (alp.) ? tristis, Bieh. rariflora, Sm. (alp.) stygia. Fries. limosa, L MageUanica, Lam, irrigua, Sm. ustulate, Walil. (alp.) j)odocai'pa, Br macrochcete, C. A. M. spectabilis, Dewey. Uvida, Wahl laxa, Wahl panioea, L sparsiflora, Steml. pheeostachya, Sm. vaginata, Tauseh. paUeseens, L supina, Wahl flava, L GEderi, Retz. pedata, Warl. (alp.) capiUaris, L hordeistichos, Vill salina, Wahl. . . . , subspathacea, Wim. reducta, Drej. cryptocarpa, G.A.M. halophila, Nyl. spiculosa, Fr. maritima. Mull vulgaris, Fr elytroides, Fr. csespitosa, L acuta, L stricta, Good rigida. Good (alp.) hyperborea, Drej. borealis, Lang. epigejos, Fr. discolor, Nyl. ? ruiina, Drej. aquatilis, Wahl stans, Drej. limula, Fr. concolor, Br. II. N. Europ. and Asiatic. III. American. 30G DR. HOOKER ON THE DISTRIBUTION OF ARCTIC PLANTS. MONOCOTYLEDONES. Cyperacese {continued), Caeex ericetorum, Poll pilulifera, L Novae- Anglise, Sc7iw. filiformis, L ornithopoda, Willd concinna, Br melanocarpa, Cham. globiilaris, L vesioaria, L compacta, Br. membranacea, Hook. saxatilis, L. pulla, Good. ampuUacea, Good rotundata, Wahl. oligosperma, Mich KoBRESiA scirpina, Willd. («7p.) . . . . Elyna spicata, Schrad. caricina, Willd. (alj).} ELiEOCHAEis palnstris, Br aciciilaris, L imiglumis, Schl SciKPUs pauciflorus, Light triqueter, L Olneyi, Gray. laoustris, L ceespitosus, L Eeiophoefm capitatum, Host {(dp.) Chamissonis, C.A.M. Scheuchzeri, Hpe. russeolum, Fr. alpinum, L vaginatum, L polystachyum, L latifolium, Hoppe. gracile, Koch. angustifolium, Rth. Rhitjchospoea alba, L Blxsmus riifiis, Huds LXX. Gramineae. AxopECTJETS alpinus, L. {alp.) ovatus, Horn. antarctious, Vcdd. pratensis, L rutbenicus, Weinm. nigricans, Horn. geniculatus, L aristulatus, Mich. fulvus, Sm. Phleum alpinum, L. {alp>.) pratensc, L Beczmannia eruciformis, Host Phalaeis anmdinacea, L MiLirji effiisum, L Ageostis Spica-venti, L rubra, L. {alp.) alpina, Wahl. (non Scop.). IS < ^ II. N. Em-op. and Asiatic. III. American. WL X C 1 1 H DR. HOOKER ON THE DISTRIBUTION OF ARCTIC PLANTS. 307 MONOCOTYLEDONES. Gramineae {cont limed). Ageosiis vulgaris, L alba, L. pol3nnorpha, HvAs. canina, L laxifolia, Rich Deteuxia Canadensis, P.B deschampsioides, Tr Lapponica, VcM neglccta, Eiipr. stricta, P.B. chalybaea, Fr. varia, P.B purpiu'aseens, Br. Groenlandica, E.M. montana, i/o.si. Hartmanniana, Fr. strigosa, Wahl aleutica, Bong. Nutkaensis, Tr. Langsdorifii, Tr purpiu'ca, Kth. elata, BIytt. Calamaqrostis epigejos, L littorea, Sclirad. lanceolata, Roth Halleriana, Gaud. phragmitoides, ffai-tm. Phragmites eommiuiis, L SpARTnfA cynosuroides, W. Atena pubescens, L Anthoxaxthum odoratum, L HiEHOOHioE boreaUs, L alpina, L pauciflora, Br raoemosa, Tr. Deschampsia csespitosa, P. B brevifolia, Br. ALra aretica, Tr. atropurpiirea, Wahl alpina, L. (alj}.) flexuosa, L Teiseittm subspicatum, P.B. (alj^.) . agrostoideum, Tr. flavescens, L Aira Euprecbtii, Oriseh. Sibii'ioum, Rupr. Melica nutans, L MoLijfiA cserulea, P. B Phippsia algida, Br monandra, H. & A. Catabrosa aquatica, P. B CoLPODiuii latifolium, Br ? paucifloruni, HJc. arundinaceum, Hook. pendulinum, Lastil Poa deflexa, Rupr. remotiflora, Rupr. similis, Rupr. VOL. XXIII. II. N. Eiiroi5. and Asiatic. III. American. 1 A H 2 T 308 DR. HOOKER ON THE DISTRIBUTION OF ARCTIC PLANTS. MONOCOTYLEDONES. Graminese (continued). CoLPOBiuM ful^iim, Tr Poa scleroclada, Rupr. latiflora, Rapr. precilantha, Rupr. DupoNTiA Fisheri, Br psilosantha, Tr, Poa pelligera, Rupjr. Gltceeia fluitans, Br aretica, Hh, Pleueopogon Sabini, Br Ateopis distans, Oriseb Poa aii'oides, Nutt. maritima, L Poa angiistata, Br. Nutkaensis, Presl. Poa annua, L alpina, L. (alp.') pratensis, L angustifolia, L. trivialis, L. nemoralis, L csesia, Sm. aspera, Gavxl. Yahliana, lAehm. serotina, Ehr. bryophila, Tr. flexuosa, WaU. {alp.) laxa, Hteiike. Cenisea, All. aretica, Br. abbreviata, Br. Dacttlis giomerata, L Festuca elatior, L Eiehardsoni, Hk ovina, L rubra, L. duriiiscula, L. sabulioola, L. Dub. arenaria, Osh. Koeleria hirsuta, Gaud. brevifolia, Br. Beojitts oiliatus, L inermis, Leyss. piirgans, L. pictus, H.f. Teiiicum repens, L violaceum, Horn. caninum, L Elyiius arenarius, L mollis, Trin HoEDEu:vi jubatum, L Naedus strieta, L IS IS IS II. N. Europ. and Asiatic. III. American. w DR. HOOKER ON THE DISTRIBUTION OF ARCTIC PLANTS. 309 ACOTYLEDONES. LXXI. FUices. Polxpodittm; Dryopteris, L Ehseticum, L Phegopteris, L vulgare, L WooDsiA Ilvensis, Br. {alp.) hyperborea, Br. {alp.) glabella, Br. {alp.) CisioPTEHis fragilis, Bernh montana, Hmnh. {alp.) La-Stkea fragrans, Sw spiniilosa, Willd Filix-mas, Sw Oreopteris, Ehr PoLYSiicHTTM angulare, Willd Lonchitis, L. {alpi.) PiERis atropurptirea, L aquilina, L CErpiOGRAJCMA aorosticholdes, R. (alp.) . crispa, Bernh SiEUTHioPTEEis Germanica, L Bleohnttm Spicant, L AsPLENiTJM Filix-foemina, L Eiita-muraria, L viride, L. {alp.) creniilatum, Fr Aspidium crenatum, Sonif. BoTEYC HTUM Lunaria, L Virginianum, Sw , Ophioglosstim vulgatum, L LXXII. Lycopodiaceae. Ltcopodium Selago, L artnotiimm, L clavatum, L complanatum, L selaginoides, L alpinum, L I80ETE3 lacustris, L LXXIII. Equisetaceae. EatrisETrM palustre, L variegatum, L arvense, L sylvaticum, L pratense, Ehr limosum, L hiemale, L Bcirpoides, Mich II. N.Em-op. ' TTT . and Asiatic. ' HI. Amen 1111 1 1 1 rv. H ADDENDUM. AsTEAGALt7S polaris, Bth., Arctic West America. See " Observations on the Species." at p. 323. 2x2 310 DR. HOOKER ON THE DISTRIBUTION OF ARCTIC PLANTS. Observations on the Species. In the preceding table I have attempted to group the arctic plants under comprehen- sive forms, founded on a careful study of the plants indicated by the names quoted : this I first did with as little reference as possible to the labours of others,— endeavouring, as far as was in my power, to suppress my preconceived ideas, whether original or adopted. Having thus arrived at the nearest possible approach to independent conclusions, when I compared my work with the descriptive Floras which I have elsewhere indicated as forming the basis of the Systematic Catalogue, I was surprised to find how many authors have directly or indirectly arrived at the same conclusions as myself with regard to the specific limits of the plants indicated. In some instances such revision corrected my previous views; but in by far the most numerous cases the summing up of this authoritative evidence afforded extraneous reasons for abiding by my own conclusions. The following notes are intended both to give these extraneous reasons, and to show to how much greater an extent than is generally supposed, the most able and experienced descriptive botanists vary in their estimate of the value of the " specific term" as applied to many of the commonest plants of the best-known countries. Prom the results of this and other most perplexing and laborious comparisons of the labours and opinions of the authors of many local and general Floras, I think I may safely afiirm tliat the specific term has three different standard values, all current in descriptive botany, but each more or less confined to one class of observers, though more or less variable with all. With the general bota- nist it is a comprehensive term, and becomes more so with age and experience ; with the monographer of large and widely diffused natural orders or genera its standard is cojitracted at first, but rapidly expands in successive revisions of his work; while the local Ijotanist, or monographer of genera or orders Avith restricted ranges, begins with a rather broad standard, which rapidly contracts. This is no question of what is right or wrong as to the real volue of the specific term. I believe each is right according to the standard he assumes as the specific ; moreover, in the majority of cases all agree with regard to the absolute and undeniable distinctness of a moiety of the plants of every area * ; all agree with regard to the permanent distinctiveness of many of the subspecies, varieties, &c. of the other or variable moiety ; and all agree with regard to the propriety and importance of tracing the characters and ranges of varieties as carefully as of species. Still the questions remain — Should the specific term ever be arbitrary ? and if so, should it be broad or narrow ? I believe it must often be arbitrarily defined, and that it shoidd be broad, because the object of botanical nomenclature is defeated by an undue multiplica- tion of names necessary to be borne in mind by the general botanist, whose convenience ought first to be considered, and also because the multiplication of specific names will demand a corresponding increase of generic ones ; moreover the daily discovery of inter- mediate forms, or new or closely allied forms, is introducing an incessant change in the nomenclature of narrowly defined species. * See Introd. Essay to Tasmanian Flora, j). v, for some ideas as to the objeetive and subjective values of the charac- ters of species, and the division thereby of all species into groups. DR. HOOKER ON THE DISTRIBUTION OF ARCTIC PLANTS. 311 Thalictbtim. For observations on the Arctic Scandinavian forms of this genus, see Fries, Summa Veg. Scand. p. 135. T. Kemense, Fr. When engaged on the ' Flora Indica,' Dr. Thomson and I made a very laborious investigation of this genus, referring the T. Kemense of Fries's ' Herb. Normale' to T. majns, Jacq., and further identifying it ^\ith the Himalayan T. Maxwellii, Pv-oyle, which also occasionally possesses stipellte. Wahlenberg referred T. Kemense to T.jiavuvi, (3 (fid. Ledebour and Ruprecht) ; and I have received from Dr. Andersson a specimen of T. Kemense that seems different from Fries's and to be identical with T. simplex, L., having the inflorescence of that plant, which is itself a form of T. Jlaviim. Of Fries's T. Kemense I have fine specimens also from Alten, gathered l)y the late W. Christy, Esq., having rather larger leaves than those of the ' Herb. Normale.' Harvey (' Flora Capensis') has referred the only South- African Thalictrnm to T. minus {Cajfriun, E. & Z., miAgracile, E. Meyer). T. rariflorum, Fr. (mmus, Fries, Herb. Norm.), appears to me almost identical in inflo- rescence with the T. stricttom, Led. {T. exaltatim, C. A. M.), which, again, resembles few- flowered specimens of T. flcwum, L. The ' Herb. Normale' specimen of T. flavum, marked "certiss." is undistinguishable from a Siberian one of Ledebour' s marked " exaltatum, C. A. M." According to Nyman, the T. rariflorum, Fr., is the same as T. Friesii, Rupr., and T. strict um-borecde of Nylander. Fries does not regard the true T. minus of Lin- naeus as Lapponian. Akemone NitttalUana, DC. This is certainly identical with A. imtens, L., and was so considered in Flor. Bor.-Am., in Torrey and Gray's Flora, and in the first edition of A. Gray's Manual. In the second edition of this last work, however, A. Gray keeps it distinct, saying that it more resembles A. pulsatiUa ihwi. jiatens : this must arise from misconception, as jmlsatilla has pinnatisect foliage, and there is no difference whatever discernible between Nuttalliana and patens, of both which I have compared large suites of specimens in all states. A. VaUii, Horn. (Flor. Dan. t. 2176), a Greenland plant, is referred by Lange to A. Richardsoni. A. alpina, L. I have seen but one Arctic Anrerican specimen ; it is much stunted. This species has not been found east of the Caucasus in the Old World, though it is not uncommon in North America on both sides of the Rocky Mouutains. Ranunculus conferooicles, Fr. This slender form of the protean R. aquatllis is the only one found within the arctic circle ; it is the U. aquatiUs /3. pantothrix of Ledebour, and, I think, also the R. aquatllis /3. heterophyllus, Fr., of Babington's ' Iceland Plants.' My Greenland, Iceland, and Lapponian specimens seem to accord well ; l}ut Durand (' Kane's Voyage,' Appendix) calls the Greenland plant " var. arcticus," and states that it is the " liederaceo-proximus" of Gieseke, having a great affinity with hederaceus, Lam,, non Linn. R. Chamissonis, Schl., according to authentic specimens, appears to be the same with R. glacialis, L. The distrilmtion of glacialis is peculiar, it having been found in East Greenland by the earliest and by all subsequent voyagers, Imt never on tlie l^.aflin's 312 DR. HOOKER ON THE DISTRIBUTION OF ARCTIC PLANTS. Bay side (where it occurs in very high latitudes only) till Kane's Expedition, when it was brought from North Proven, lat. 72" N. B. Flmnnmla, L., and reptans, L., are both arctic plants, the latter the more so. Tries keeps these distinct, as do Torrey and Gray and others, and Koch, in his Addenda et Emendanda, because of the short-beaked achenium of reptans : in the body of his ' Elora ' he made it var. jS of Flammula, as do Bentham and various other authors. I find the beak of the fruit of Flammula to be very variable and often quite undistinguishable from that of reptans. There is a United States plant very like Flammtda, but with a much more slender beak than the European plant, which is called B. Flammula by American authors. Pl. affinis, Br. Ledebour unites this with his B. amcemis. I have no hesitation in referring both to auricomus, L. B.frigidus, DC, and B. EschsclioUzii, Schl., seem to me inseparable from nivalis. B. sidfureus, Sol., is more distinct, and kept separate by Eries, but is referred to nivalis /3 by Ledebour : it is one of the most arctic plants known, and certainly passes into nivalis. B. j)olyanthenios, L. It is so difficult to distinguish herbarium specimens of this from B. acris, L., that I am not sure I have got its full distribution, for which I have relied on authentic book sources. B. nemorosus is usually combined with it. B. glabriusculus, Rupr. El. Sam., is an imperfectly described plant, referred doubtfully to B. acris. No allusion is made in the author's description to the peduncle, which is figured like that of acris ; but the receptacle is described as glabrous. It was, further, found with acris at the Gulf of Indega, east of the mouth of the White Sea. The acris j3. humilis, "Wahl., is also, according to Ruprecht, common on the Samoied shores of the same sea, and in the arctic Island of Kolgujew. B. pygmceus, Wahl. A high-arctic species, best known from hyperboreus, with which it is united in ' Elora Indica,' by wanting the creeping flagelliform stems. The Tyrol is the only European habitat south of Scandinavia. The B. Sabinii, which I have placed with it, is so named authentically, but does not differ from pygmmts ; it is, however, one of those extremely reduced forms whose origin can only be ascertained by examining an extensive suite of specimens. Of the B. Samojedarum, Rupr., I have seen no specimen ; it is an imperfectly described plant, found in the Island of Kolgujew at the mouth of the White Sea, and described as being very nearly related to B. Furshii. B. hypjerboreus, Rottb. Of this, which is a Greenland, European, and Siberian plant, I have seen no Arctic American specimen ; all so called hitherto are, I think, referable to pygmceus. The B. Gmelini, referred to hyperboreus in Elor. Bor.-Am., is reduced to Purshii by Ledebour. Caltha palnstris, L. The prevalent opinion amongst Ijotanists is to unite as varieties all the names I have placed under this. The true palustris itself inhabits the extreme north (Island of Kolgujew, Rupr.). C. natans, L., is a floating plant affecting high lati- tudes only. C. radicans and C. arctica are probably synonyms, the yellow colom- of the persistent sepals being discharged after flowering. Watson regards C. radicans as cer- tainly a reduced form ; Nyman makes it distinct, giving Scotland as its habitat, but says of C. palustris, " species ut videtur collectiva " ; Eries makes it a variety found in Lap- DR. HOOKER ON THE DISTRIBUTION OF ARCTIC PLANTS. 313 land ; Bentliam would unite them all. The absence of any form in Greenland is a most remarkable fact, the common one being most abundant and conspicuous in Iceland. Delphhstium Middendorffii. Trautvetter, the author of this species, indicates its affinity with Menziesii, but does not appear to have compared it with that plant, which, judging from the figure of Trautvetter, may not be distinguished. Whether D. Menziesii itself may not be the same with some better-known species, is a question for future deter- mination. D. intermedium. Ait., is elatum /3, Turc. (El. Baik. Dahur.). Aquilegia Canadensis, L. The very close affinity and probable identity of A. Cana- densis, L., and A.formosa, Fisch., is indicated in the 'Flora Indica;' and on re-examina- tion, with more specimens of the former to compare, I find no reason for modifying that conclusion. A. brevistylis. Hook. Originally doubtfully referred by Richardson to A. vulgaris, L. ; the styles, however, which are somewhat variable in the European plant, are always much shorter in the American. It is a very northern species in America, and not hitherto known west of the Rocky Mountains, though I have seen specimens of a Sitka plant, in an indiff'erent state, which is a great deal like it. It is also allied to the Siberian A. parvijlora. Led. AcT^^A spicata, L. The A. rubra, Willd., is referred to nigra by Fries, both being Lap- ponian. Asa Gray (in litt.) combines with them A. alba. Big., and arguta, Nutt, Papaver nudicaule is now almost universally regarded as specifically the same with alpinum, L. FuMARiA officinalis, L., occurs sporadically throughout Nordland, according to Fries and Andersson, but can hardly, I think, be considered an indigenous plant. NcPHAB, lutea, L. This, together with the species enumerated under it, and N. ad- uena, Ait., are, perhaps, rather forms of one collective or aggregate species than perma- nent undistinguishable plants ; and it is further possible that the N. sagittcefolia, Pursh, may be another state, in which the early sagittate form of leaf is retained in the adult plant. Torrey and Gray refer Kalmiana to lutea (3, and say of advena that it is not specifically distinct. Asa Gray latterly keeps up advena and Kalmiana, but adds to the latter N. intermedia, Led. ? Nyman and Ledebour keep all distinct. Watson, from his notes under pumila, seems to regard it with doubt. Fries keeps lutea and puinila di- stinct, but regards intermedia. Led., as a variety of lutea, and adds as a Lapland variety to pumila the Spenneriana, Gaud. Lastly, Koch keeps lutea, pumila, and Spenneriana, all distinct ; and Bentham unites the two first together with intermedia and minima of Engl. Bot. Barbabba vulgaris, Br. This, again, is either a collective species or several species variously discriminated. Fries distinguishes stricta, Fr., vulgaris, Br. (including under it as a variety arcuata, Reich.), and prcecox, Sm. Nyman excludes prcecox, Br., frcu Scandinavia, and reduces proicox, Sm., to arcuata. W:;tson finds himscif couipcllcil to 31i DR HOOKER ON THE DISTRIBUTION OF ARCTIC PLANTS. treat the distribution of vulgaris, Br., str'icta, Audrz., and arcuata Reich., in connexion. A. Gray considers the American stricta and arcuata as varieties of vulgaris. Bentham says that B. vulgaris, prcecox, and stricta pass through every gradation into one another. Another element of confusion in this group is the doubtful origin of the common pot-herb B. prce- cox. "Watson treats it as an alien. Torrey and Gray, on the other hand, describe it as a native of Canada, ascending to 68° N., though perhaps stricta is here meant, for A. Gray latterly refers prcecox of Fl. Bor.-Am. to vulgaris, var. stricta, and riglitly as far as the specimens I have examined go to show. Tries remarks (Summa, p. 1-16) that Smith's ovigmvil prcecox (Flora Britannica) is a cultivated arcuata, the English Botany plant being diflPerent (thus accounting for Ny man's conclusions above stated). My North Indian and Himalayan Barbareas are, again, generally referable to the European forms, though scarcely characteristic of them. I long endeavoured to keep the Australian and New Zealand form distinct ; but neither Bentham nor Mueller regard it as different from vul- garis, with which undoubted wild specimens from the Australian Alps well accord. TuEMTis ^;a^?«^rt, Graham, and retrofracta, Hook., are, I think, identical with T. mollis, Hook. The Arabis Holbollii, Horn., is another synonym. There are certainly differences in the length and breadth of the pod in both patula and retrofracta ; but these do not at all indicate the seeds being in one or two rows, which is a very variable character in both. Arabis IlolbdlUi is identified with A. retrofracta, Grab., by Lange. AiiABis lyrata, DC. This, which is the Sisymbrium humifusum, J. Vahl, and Arabis sisymbrioides, Hook., differs according to descriptions from A. petnea, Lamk., in the rather oblique radicle of the embryo, and biennial root (Torrey and Gray, i. 80, 81), neither of which characters appear to me to hold good in the rather numerous specimens I have examined, the radicle being in both variable as to direction, and the roots identical. In habit and every other respect, the plants entirely accord, as far as herbarium specimens allow me to judge. According to Eries, A. petrcea is not a native of Lapland; nor does Ledebour give any European or Asiatic arctic habitat for it ; it is, however, common in Finland, Scandinavia, Iceland, and N. Britain. Eries includes A. ambigua, DC, under it, a plant kept distinct by Ledebour and by Torrey and Gray. Cakbamine bellidifolia , L. I have sometimes been almost disposed to suspect that this may be an arctic form of hirsuta, to which it certainly is very closely allied, and to which it tends through microphylla, Willd. C. Lenensis is identical with bellidifolia, and is var. Lenensis, Trautv. (Elorula Taimyr.). C. hirsuta, L. This, though a common Icelandic plant, and one that crosses the circle in America, does not occur in Scandinavia north of Gothland and Finland. The C. sylva- tica, Link, is distinguished by its six stamens, spreading pedicels, and style as long as the siliqua is broad, — all variable characters, except that of the stamens. In the southern hemisphere and elsewliere C. hirsuta is hexandrous. Watson finds no difficulty in distin- guishing these as wild plants in England, but considers sylvatica a doubtful species, adding that the distinctions are less decided under cultivation. Bentham considers sylva- tica to be a large luxuriant hexandrous form of hirsuta. C. digitata, Bich. I have examined a large suite of excellent flowering specimens of DR. HOOKER ON THE DISTRIBUTION OF ARCTIC PLANTS. 315 this plant, in which state it differs in no respect from pratensis. I have not seen tlie fruit, however. C. umbrosa, DC, is referred by Pries to a variety of amara, L. ; both are Lapponian. EuTEEMA arenicola. Hook. I have examined this plant very carefully ; it is certainly not a JEutrema. In the few specimens known to me the radicle is slightly oblique. Turczaninoif (in Herb. Hook.) suspects that it may be his E. piliferum {Braba grancli- fiora, C. A. M. ; Fachyneurtmi grandijlorum, Bunge ; Braya 3Ieyerl, Bge. ; Barrya mi- crocarpa, Led.) ; but I find no traces of the peculiar hairs of this plant in the P. arenicola. Richardson observes that it ranges from 107°-150° W. I have seen no specimens but his own. Vesicaria arctica. Rich. Planchon , who has examined the specimens in the Hookerian Herbarium with a great deal of care, confirms the identification of the Chilian plant with the arctic. It is singular that this plant should be found as far north as 81° in Green- land, but not amongst the Polar American Islands. Draba. Of all the arctic genera, none but Salix present so many diflS.culties as Braba. Whether by variation, or hybridization, or other unknown cause, the prevalent forms seem, wherever they grow, to pass into one another by insensible gradations ; and no botanist has succeeded in bringing the arctic stunted varieties within such specific limits as to have gained the assent of others. No doubt many are, in their arrested condition of growth, absolutely undistinguishable in the present state of science ; and whoever attempts their discrimination must expect to change his opinion somewhat at every re-examination. I am fortunate in having Mr. Ball's advice upon some of the European species, which he has studied with great care, and have followed his opinion in the naming and grouping the arctic ones ; unfortunately, however, he has not completed his study of the genus, nor of all the species here enumerated ; so that the geographical data are approximate only ; nor does he speak with implicit confidence of their synonymy. In addition to the excellent critical notes he has published in the Bulletin of the Botanical Society of Paris (vol. vii. pp. 227 & 247), I have from him the following provisional grouping of the com- moner forms : — 1. B. androsacea, Wahl., 1812, an Willd. ? D. Wahlenbergii, Hartm. 1820. D. Lapponica, DC. Syst., an Willd.? D. nivalis, DC, non Lilj. 2. B, Fladnitzensis, Wulf., Jacq. Misc. D. lactea, Adams. ? D. Carinthiaca, Hoppe. D. crassifolia, Grab. 3. B. muricella, Wahl. D, nivalis, UX]., nan J}G. 4. B. rupestris, Br. D, Altaica, Bge. 5. B. hirta, L. D. Dovrensis, Fr. D. arctica, Vahl. 6. B. incana, L. D. contorta, Ehr. D. confusa, Ehr. & DC. D. Thomasii, Koch. VOL. XXIII. 2 u 316 DR. HOOKER ON THE DISTRIBUTION OF ARCTIC PLANTS. Of these, Mr. Ball is disposed to unite 1 and 2, and also 4, 5, and 6, making but three sjiecies in the above group, but distinguishing six sufficiently weU-marked forms. D. muricella is in Europe confined to Scandinavia ; neither D. hirta nor rupestris are found in the Alps. Deaba alpina, L. Regarding the numerous plants here brought together I have the following notes : — B. algida, Ad., according to specimens from Ledebour and Bunge, differs in no respect from D. alpina. B. pilosa, DC, is reduced to a synonym of D. aUjkla by Ledebour. I have no authentic specimens of it. B. aspera, Adams, is placed by Ledebour in the section Aizoopjsis, the species of which are eminently southern and oriental ; that author, however, had not seen the plant, and overlooks the fact of the distinguishing characters given by Adams being those of B. alpina. B. Adamsii, Led. {lasiocarpa, Adams), is described as differing fromZ). alpina in the pu.bescent pods, stellate hairs of leaves and scape, and smaU flowers ; but the pods of alp)ina are glabrous or pilose, and the other characters equally inconstant. B. glacialis, Adams. Of this there are five varieties described in ' Elora Boreali- Ameri- cana.' It should be distinguished by the costa prominent in the under surface of the leaf; but amongst a vast suite of specimens thus named I find the prominence of the costa to depend on the development of the leaf, and no characters that would serve to include those so distinguished, even as a permanent variety of alpina, L. Durand (Kane's Voy.) keeps glacialis j3 distinct. B. pauciflora, Br. This was proposed by Brown as a doubtful species from Melville Island ; and I find, amongst starved specimens of B. alpina from that island, some that may be the plant that Brown alkided to. B. microjjetala, Hook. This appears to me, without doubt, to be a starved form of alpina. It occurs only in very high northern latitudes, amongst the polar islands. Durand (Kane's Voy.) makes of it B. alpina, var. micropetala. B. trichella, Pr. Mr. Ball considers this to be probably a form of aljnna. DiiABA androsacea, Wahl. {Lajjponica, DC. an Willd. ?). This is a white-flowered species presenting as many puzzling forms as B. alpina, besides appearing to pass by insensible grades into hirta, rupestris, and muricella. Nor is it always possiljle to distinguish androsacea from alpAna in a dried state, nor when the flowers of the former are bleached or vary to whitish, which they occasionally do both in the arctic circle and Himalaya. Of this plant B. TFaUenbergii, Hart., and lactea, Adams, are now generally acknowledged synonyms, as is B. nivalis, DC, non Lilj. B. crassifolia, Grab., has been referred by Mr. Ball (in Herb. Hook.) to lactea. Ad. {Fladnitsensis, Wulf), and certainly rightly. The yellow-flowered plant figured in ' Flora Danica ' is referable to alpina, L. B. Martinsiana, J. Gay. Of this species I know nothing accurately. B. oligosperma, Hook., is not different, I think, from the specimens marked Fladnit- zensis from the Alps and Altai, and, further, is certainly a small form of androsacea, with DR. HOOKER ON THE DISTRIBUTION OF ARCTIC PLANTS. 317 pods short, elliptical, sharp at both ends, and few seeds. It is, on the other hand, very near muricella and rupestris, and to small forms of Jdrta. I have it both from the Rocky Mountains and Lapland. D. corijmbosa, Br. Durand doubtfully refers this to his var. corymhosa of alpirw (= alpina S, Hook.). The iigure of the Greenland plant in 'Flora Danica' appears to belong to a common arctic form of androsacea. Deaj3A muricella, Wahl. This Mr. Ball distinguishes from D. stellata, Jacq., which is nowhere found in the arctic regions. _D. muricella, "VVahl., on the other hand, has not been found in Middle or Southern Europe. It is the D. nivalis, Lilj., non DC. Durand (Kane's Voyage) refers the D. rupestris, var. j3, of Torrey and Gray to nivalis, Willd., apparently from description only. Deaba stellata, Jacq., non DC. Mr. Ball is my authority for including D. Joliannis, Host. Draba Jiirta, L. I have Mr. Ball's authority for including Dovrensis, Fr., and arctica, Vahl, under this plant. I have no confidence in its permanent distinction from D. incana. It is usually a much less leafy plant, with longer pedicels in comparison to the pods. D. oblongata, Br., is apparently founded on a small specimen of D. hirta. Draba incana, L. D. contorta, Ehr., and confusa, Ehr., are now generally considered to be forms of this ; and I find them to be scarcely distinguishable as varieties. D. Magellanica, Lamk. I have already (' Flor. Antarct.' ii. p. 233) referred this to incana ; and a fresh series of specimens, received since that time, not only confirm this opinion, but incline me to regard D. Falklandica, mihi, as a mere stunted form. J), borealis, DC. I have examined two specimens of this, communicated by Prof. Fischer ; and it appears to be nothing but a rather luxuriantly leafy form of D. incana ; the specimens have not ripe fruit, however. Dbaba rupestris, Br. It is extremely difficult to distinguish this from small states of lactea, hirta, and incana, and in some cases from D. muricella. D. Altaica, Bunge, is clearly a synonym. Durand refers here the D. hirta, E. Bot., and hirta, var. 4, of Fl. Bor.-Am. Bentham refers the British rupestris to hirta. Draba aurea, Vahl. This is a very puzzling plant, so like in its normal state to 1). incana, that I do not see how these plants can be distinguished, except by the colour of the flower, which, again, apparently varies to white. I am not aware of any yellow-flowered Draba with the habit of incana being found, except in Labrador, Greenland, and the Bocky Mountains. CocHLEARiA. This has always proved to me to be one of the most intractable boreal genera ; and I do not believe that the common littoral forms are always defineable speci- fically. Habit, pods, and leaves afford the characters hitherto made use of; and all are equally fallacious, as far as affording permanent distinctions. Buprecht, speaking of the Samoied forms (Flor. Samojed. p. 21) says of C. JFahlen- bergii, Bup. {anglica, Wahl., non L.), C. oblongifolia, DC, C. arctica, DC, C. fetiestrata, Br., C. Danica, L., " Quod reliquum est, vereor, ne Cochlearice onines hie enumerataj ad 2u 2 318 DR. HOOKER ON THE DISTRIBUTION OF ARCTIC PLANTS. unani speciem pertineant valcle vai'iabilem ; interim tamen formas clistinguere necesse fuit." In endeavouring to dispose of the arctic forms of this genus I have found myself obliged to return to the old species and definitions of Linnseus, who, as it appears to me, had a definite idea of the following four prevalent forms : — C. officinalis, with cordate subrotund radical leaves, and oblong subsinuate cauline; C. Danica, with hastate, angular, deltoid, petioled leaves ; C. Anglica, with all the leaves ovate lanceolate ; C. GrcenlancUca, with reniform, fleshy, quite entire leaves (generally small). Of these I find Gt'oenlandica so often passing directly into officinalis, both in America and Europe, that I have brought tliem together ; the others are more distinguishable pre- valent forms, though all pass into one another. CocHLEABiA sisymhrioides, DC. This is much more distinct from any of the other arctic forms than the most distinct of these latter are from one another. I know of no other which at all approaches it, though the perfectly flat pod, so opposed to the prevalent generic character, may be found in states of C. officinalis. CocHLEAKiA Danica, L. This, in its typical form of a small slender spreading plant, with hastate, petioled, small leaves, is by far the most distinct of the group to which it belongs. It is found on all the northern coasts of Evirope as far south as Brittany and Normandy, and suddenly reappears in its typical state in the Pyrenees. According to Fries it does not inhabit Lapland. CocHLEARiA Anglica, L. Watson confesses his difficulty in always distinguishing this, which to him appears to pass on one hand into Danica, and on the other into officinalis. In Western Europe it is common as far south as the Channel, but does not extend beyond Boulogne (Gren. & Godr.). Dm^and appears to restrict this in Greenland to a form with ellii^tical pods and fenestrate septa. C. oblongifolia, DC. E-uprccht, I. c, observes that the form with elliptic silicles occurs almost everywhere on the Samoicd beaches ; that with globose pods only at Cape Konuschin. C. fenestrata, Br. Nyman (' Sylloge ') includes this under WaUenbergii, Eupr. Durand (Kane's Voy.) distinguishes it by its smaller size. Fries makes it a var. of Anglica, L. CoCHLEARiA officinalis, L. Nyman (' Sylloge,' p. 198) refers C. Grcenlandica, L., and Fyrenaica, DC, to this. Watson (' Cybele ') confines it to shore plants with globose silicles, doubtfully keeping distinct the mountain form C alpina. Sweet ?, Grcenlandica, With. ?, Sm. It is a common mountain plant in Northern Europe, occurring as far south as the Pyrenees and Carpathians, but it does not extend far down the French coast. It includes Wahlenberg's q^cmaZ^* {JFalilenbergii, Eupr.), also arctica, Lenensis, and fene- strata, Br., though the latter name has, I suspect, been rather indiscriminately applied to other Cochlearice with ruptured septa. Dm-and defines the Greenland form by its glo- bose silicles. C. arctica, DC. Of this, Euprecht remarks that the Samoied specimens have entire or fenestrate septa, but that it differs from C. fenestrata, Br., in larger size (^-1 foot) and distinctly-nerved sfliquse, from C. Anglica in the ellipsoid silicles (not subrotund). DR. HOOKER ON THE DISTRIBUTION OF ARCTIC PLANTS. 319 which are longer than their pedicel, not twice as short. Fries retains arctica, Schlecht., as a species not yet sufficiently established. C. Wahleiibergii, Rupr. {Anglica, Wahl., Fl. Lap., non L.). Ruprecht remarks of the Samoied plant that the leaves are almost those of C. Danica,!^., but the siliquse- elliptic, the valves sometimes nerved, and much narrower than in true Anglica. Hesperis Fcdlasii, Torr. & Gray. Durand (Kane's Voy.) includes U. minima, T. & G., as well as pygmcea, Ilk. In Greenland it has been found only in the extreme north, by Kane's Expedition, on Washington Land, lat. 81° N. Sisymbrium sophioides, Fisch. This is nothing but an abnormal form of ^S*. Sophia, L., with a remarkably corymbose inflorescence. It has hitherto been found only in high northern regions. S. brachycarpum. Rich. This, which passes by insensible transitions into ^S*. canescens, Nutt., which inhabits the same regions to the south, often approaches flnely-cut-leaved states of S. Sophia. S. salsugineum, PaU. I have carefully compared with this the Ttirritis diffusa. Hook., and have no doubt of their absolute identity. Erysimum UeraGiifolimn, L. Ruprecht (Fl. Sam.) refers this to E. strictimi, Fl. Wett., and notices E. cheiranthoides, L., as occurring very close upon the arctic circle, almost at the termination of the woody region. I am quite unable to discriminate between these northern Erysima and their southern congeners, nor do I find any satisfactory characters for them in books. Fries does not consider either the true Meraciifoliwm, L., or cheiranthoides to be Lapponian; but E. alpiniim, Sm., is so, and is ranked by him as an insufiiciently established species. Nylander (' Flora Helsingfors,' 31) mentions E. alpinmn as a form of cheiranthoides, and keeps hieraciifoUimi distinct. Some idea of the difficulty (impossibility ?) of distinguishing these species may be obtained by reading Fries's notes (' Summa,' p. 148) and observing the synonyms quoted by Koch under E. strictum. Braya alpiina, Sternb. I think there can be no doubt that B. glabella, Rich., and dubia, Br., are both synonyms of B. alpina, Sternb., or, at the furthest, arctic forms of that plant. I am quite unable to distinguish them by any good characters. The pods are singularly variable in length and breadth, and, as with other arctic Cruciferce, are apt to be arrested by cold in early states of development, and hence to assume at later periods various forms, depending on subsequent accessions of heat and moisture. Braya rosea, Bunge. I have retained this species, though I extremely doubt its distinct- ness from B. alpina. The Arctic European habitat is introduced on the faith of Nyman's ' Sylloge,' who quotes Bunge (Cat. Sem. Dorp., anni 1839) for its being a native of Arctic Russia (Samojed.) ; but I do not find any of the genus in Ruprecht's Flor. Saraojcd. Cisural. (1845). Braya pitosa. Hook. I have repeatedly examined this form very carefully, and, though undoubtedly very near B. alpina in general appearance, I am unable to identify it with any state of that plant. 320 DR. HOOKER ON THE DISTRIBUTION OF ARCTIC PLANTS. Thlaspi montcmimi, L. I am quite unable to satisfy myself about the specific limits of this plant, and indeed of many other forms of the genus, whose species appear to be extravagantly miiltiplied. Belgium is its northern limit in Europe, whereas T. alpestre, to which the American plant may belong, reaches Gothland. Lepiditjm rtiderale, L. This widely-spread plant advances beyond the arctic circle in North-eastern America, but apparently not in West America or Europe. It is so fre- quently an introduced plant, that it is difficult to assign its indigenous limits. SuBULARiA aquatica, L., though apparently not a native of the Alps, is found in the Pyrenees. YioIjA. palustris, L. According to Eries, the typical plant of this name is rare and local in Lapland. Asa Gray identifies the White Mountain plant with it ; and it is also found in Southern Greenland. It is absent in Arctic Russia. V. epijjsila, Led. Eries, in his 'Novitise,' makes this a variety of palustris ; and in his ' Summa Veg. Scand.' it is regarded as a species found throughout Lapland. Koch keeps it distinct, but has never seen living specimens. Ruprecht (Elor. Samojed. Cisural.) mentions a variety hyperhorea, in which the sepals are narrower and sharper than in the typical states, and the peduncles and petioles more slender. In the in- terior of the arctic Island of Kolgujew it is very rare, and has glabrous leaves. V. hlanda, Wahl. I am doubtful about the specific distinctness of this plant. It closely resembles V. palustris ; and A. Gray keeps it distinct, but his differential characters appear very trifling. In ' Elora Bor.-Am.' the flowers of the two are said to present " the most perfect agreement." In ' Bot. Beechey's Voy.,' V. hlanda is said to inhabit Kotzebue's Sound; and Ledebour, who has seen specimens, retains it with a query, and says that, though allied to V. 2:)alnstris and epipsila, it differs in habit, smaller flowers, and approximate nodes of the root. Seemann does not include it in his Elora of Western Eskimo-land. The hlanda is said to be fragrant. V. tdiginosa, Schrad., is another species which is so excessively close to j^alustris as to be with difficulty distinguishable. T'. microceras, Rupr., is founded on a solitary specimen from the shores of Kolgujew Island ; its author says of it, " fades tota V. p)alustris, sed foUa margine et pagina superiore (non inferiore) pubescentia ; sepala linearia, acuta, margine rigide ciliata, serrata." Viola canina, L. The various forms and near allies of this plant are sources of infinite diversity of opinion amongst botanists ; added to which, regarding canina and its allies as a group designated Sylvestres by Eries, they seem to inosculate in many points with another and different one, the Pratenses, Eries. Thus it is not difficult to unite canina with Buppii and stagnina, through the forms of what is often called lactea. Watson has some excellent remarks on these forms; he includes lactea, Sm., under Jlavicornis, and keeps it doubtfully distinct from canina. Eries considers montana, L., a distinct variety only of canina, and gives all Lapland as the habitat for both, whereas arenaria and sylvatica, though also Lapponian, are more local : in his ' Novitise ' he makes sijlvestris, Lam., a variety of canv)ia, as does De Gingins in DeCandoUe's ' Prodromus.' Koch makes both DR. HOOKER ON THE DISTRIBUTION OF ARCTIC PLANTS. 321 Buppii and montana varieties of canina. Babington, in Henfrey's ' Bot. Gazette,' ] 850, p. 141, roughly discriminates three British forms : — sylvatica, Pr., ' Herb. Norm.' {canina (3, Fries, Novit. 272 ; sylvestris, Koch, non Lam.) ; canina, Linn, {sylvestris, Lam.), with a var. /3 including lactea, Sm. ; and stagulna. Kit. {'persicce.folia, Er. ; lactea, Reich.), Avhich I do not find quoted in any Arctic Flora, and so have not entered in my list of names. V. arenaria, DC, is very closely allied to canina; and though distinct in appearance, it is difficult to find any good characters. Bentham quotes both lactea and sylvatica as varieties under cauina. V. MuMenhergii, Torrey {Muhlenhergiana, Ging.). Torrey and Gray remark how very near this is to canina^ \ and in the ' Flora Boreali- Americana ' the difficulty of pointing out diagnostic characters is admitted ; it is, however, kept up by all authors, and is no doubt a permanently but slightly altered form. It is found in Greenland. V. hicolor, Pursh, is regarded by the American botanists as a variety of arvensis, DC, itseKa form of V. tricolor; to me it appears to be a permanently distinct plant, though having but feeble diagnostic characters. Both tricolor and arvensis are Arctic European and Siberian ; but hicolor alone is American, and found only to the eastward of the Rocky Mountains. DiANTHUS repens, "VYilld. This, which is not Scandinavian, appears to me nothing but an arctic form of alpina, L. Both are arctic. D. dentosus, Fisch., appears the same with D. Seguieri, Vill., or so nearly allied that it is difficult to distinguish them. Neither are natives of Scandinavia. SiLENE infiata, L., and S. maritima, L. I have kept these forms apart, though they are united by many authors (Bentham included). In the arctic regions they are distinguish- able. S. maritima has not been found on the Pacific coasts, though S. inflata occurs in N. E. Asia ? Watson observes that, on the coast of Britain, some forms of maritima become so very like those of inflata as to be hardly distinguishable. >S'. paucifolia, Ledeb., from the shores of the Arctic Ocean, is unknown to me ; it is a plant of Pallas, and evidently very near, if not identical with S. inflata. Lychnis apetala, L. I have no hesitation in saying that no satisfactory limits can be assigned to the six forms ranged under this name. In the Himalaya, where it is abundant at great elevations {macrorhiza and cuneifolia, Boyle), I have observed it to be singularly variable in most or all of those characters which have been employed to separate its forms in other countries. Durand (Kane's Voyage, Appendix) finds the characters oi pamiflora Fisch., and uniflora, Led., to be present in Greenland specimens. Fries keeps apetala and affinis distinct, the former being found throughout Lapland, the latter rare; he distin- guishes affinis by its flowers, erect at period of expansion, and obovate angular (not reni- form) seeds. L. sylvestris, Schkr., is doica a of Linnaeus, and diurna, Sibth. Jj. pratensi^, Spr., is dioicaj3, L., and vespertina, Sibth. Bentham keeps them distinct. Sagina c(espitosa, J. Vahl, is Arenaria ccBsintosa, Fl. Dan. 2289, according to Lange. S. intermedia, Fenzl in Bupr. Flor. Samojed., is the same with nivalis, Fr. S. saxatilis, Wimm., is the same with Idnncei, Presl, and approaches very closely both 322 DR. HOOKER ON THE DISTRIBUTION OF ARCTIC PLANTS. to S. uivalis and procmnbens. Bentham unites with it >S. subulata, Wimin. {procmnhens /3, L.). Arenaria nardifolia. Led. This and A. formosa are considered varieties of one by Fenzl (Led. M. Ross.). Torrey and Gray remark that the Arctic American plant figured in ' rior. Bor.-Am.' does not accord well with Ledehour's. To me they appear both to resemble very closely A. lyclmidea, M. Bieb, a Caucasian species; but very large suites of all these and their cognate species are required to arrive at any conclusion as to their limits. All the North Asiatic and American Alsmece are in a most unsatisfactory state. A. Lapponica, Spr., is reduced to A. uliginosa, by Tenzl in 'Flor. Ross.,' and both to Alsine stricta, Wahl. A. Bossii, Br. This, which I find often confounded with rubella in Arctic Herbaria, may prove to be a polar state of A. uliginosa, Schl. ; it has been found in very high lati- tudes only. A. stricta, Mich., is A. {Alsine) Michauxii, Eenzl, according to A. Gray (Bot. N. U. States) . A. verna, L. Under this name I have grouped four others, which represent as many arctic forms of this protean plant, and have all been reduced to it by Fenzl and other writers. Thus hirta and p)ropinqua both come under lusus 1. of Penzl's var. /3. Mrta ; rubella and qiiaclrivalcis under var. 0. glacialis. Fries, however, distinguishes rubella specifically from verna by its habit, and rugulose, not granulate seeds. A, Giesekii, Horn. (Flor. Dan. 1518), is rubella y of Lange. A. Pumilio, Br., is reduced to arctica, Stev., /3, by Torrey and Gray ; but if correctly, I do not see how arctica is to be kept distinct from biflora, AVahl. Ledehour's arctica is in- deed referred to biflora by Fenzl. According to Fenzl's descriptions (which do not however contrast), arctica is distinguished by the broader petals (sometimes narrower by deformity), and seeds " leevissimis fuscis," in opposition to " levissime rugulosis." But I find the seeds of the Lapland plant to be undistinguishable under any magnifying power from those of the American ; and the petals are, as Fenzl's description states, excessively variable in both. A. Norvegica, Gunn. Fi'ies, who includes this under A. ciliata as a variety, or, perhaps, distinct species, assigns it a rather different range, ciliata being found in Lapland and N. Finland only ; Norvegica in N. NorAvay and S. Lapland. A. ciliata alone is found in Arctic Russia, according to Ledehour's ' Flora;' but neither appear in Ruprecht's SSamoied Flora.' Watson considers their distinctness to be questionable ; and Bentham unites them. A. Grcenlandica, Spr. This has been found in Arctic Greenland by Kane's Expedition only, which brought it from Upernavik, lat. 73° N. HoNKENEJA pex)loides, Ehr. A doubtful native of Lapland according to Fries, who indicates a subspecies {oblongi folia, Torr. & Gray) as inhabiting that country. Stellaria crispa, Cham. Eenzl says of this, " A S. boreali vix distincta, facUe tamen distinguenda." I am at a loss to distinguish it, either by specimens or book characters. According to Fries, borealis is rare in Lapland, but its var. calycantha is common through- out that country. Planchon, who has studied the genus very carefully, points out (Herb. DR. HOOKER ON THE DISTRIBUTION OF ARCTIC PLANTS. 323 Hook.) that the horealis^ of 'M. Bor.-Am.' is the same with (jracllis, E-ich., I. c, and a plant of Drummoud's, called Larhrea aquatica, and further that all are referable to S. twnhellata, Turc. S. longiiies, Goldie. Of the plants brought under this name, S. Icsta, Eich., and Edward. sii, Br., are all but universally considered to be identical or states. Of Bunge's S. peduncu- laris being the same, there can I think be little question. Of S. hebecalyx, Fenzl, and ciliatosepala, Trautr., I have seen no authentic specimens ; but, judging from the drawings and descriptions, and from specimens in the Herbarium from Labrador, displaying similar characters, I have no doubt of these also being mere varieties. Fenzl says that hebecalyx is intermediate between a variety of longlpes and Dalmrica (3. Planchon (in Herb. Hook.) refers DaMmca to longipes doubtfully ; and TurczaninoflF (fid. Eupr. PL Sam.) regards Dalnirica(3 2 of Fenzl as the same with glauca, L. The smooth testa of the seed, how- ever, distinguishes all the forms of longipes. Durand (Kane's Voy.) refers S. stricta, Eich., to longipes j3. minor and 8. Iceta. Planchon (Herb. Hook.) further refers S. imhri- cata, Bimge, alpestris, Turc, and Fischeriana, Ser., to longipes, and, I think, rightly. The S. glauca of Greenland (Lyall's plants, nob. in Linn. Journ. vol. i. p. 116) is a state of this plant. S. longifolia, Fries. A very confusing plant, whether as regards its synonymy or limits. I am quite unable to distinguish the European from the American form; both have a smooth testa, and seeds otherwise identical. Fries distinguishes them ; and Koch (' Synops.' 131) says that longifolia, Muhl., is most certainly distinct, but gives no characters. S. alpestris is regarded by Fries as a variety, and perhaps a distinct species ; it ap- proaches states of idiginosa in many characters ; but the seeds are always smooth. Lepigonum saliniim, Fr. This should perhaps be regarded as a member of an aggregate species including L. marinmn, Wahl., of which Babington considers it a variety, and ruhrum, AYahl., plants which in their extreme forms differ widely, bvit seem united by innumerable variations ; Bentham unites them. Cerastium Fischer ianum, Ser. Euprecht inckuies a plant of this name, in ' Herb. Chamisso,' under vulgatum, and further doubts the possibility of distinguishing it from arctic forms of C arvense. Fenzl (in ' Led. Fl. Eoss.') reduces both this and Beeringia- num to vulgatum. Torrey and Gray, and, following these, Durand, refer Fisclierianmn to alpinum, and keep Beeringianum distinct, but next to it in order. C. arvense, L. Euprecht gives the only arctic habitat for this, viz. the extreme forest- limit in Samoiedland, but he gives its var. /3 {arvense, incanmn. Led.) as abundant in sandy places on the arctic Island of Kolgujew. PoLYGALA vulgaris, L. I hesitated long whether to amalgamate the distribution of this with the Siberian P. comosa, but have refrained. Astragalus poLxIris, Benth, ; giaber v, parce pubescens, caule breviter difFuso, stipulis postice concretis, foliolis 11-15 ovatis oblongisve retusis v. emarginato-bifidis, peduu- cuHs folio subgequalibus apice 1-i-floris, calycis dcntibus tubo paullo brevioribus, corolla calyce subtriplo longiore, carina alis vcxilloque brcviore, leguniine siibsessili VOL. XXIII. 2 X 324 DR. HOOKER ON THE DISTRIBUTION OF ARCTIC PLANTS. inflato membranaceo nigro-piloso calyce 3-4-plo longiore, sutixra carinali liatid in- trusa. — G. B. Hab. Eschscholtz Bay, in Kotzebue's Sound, Seemann. In many respects similar, especially in habit, stipules, and form, to A. aljnnus, L., but always more glabrous, with more deeply emarginate or almost bilobed leaflets, fewer flowers, longer calyx, narrower petals, shorter keel, almost sessile ovary, and widely dif- ferent pod, which is broadly linear oblong, not pendent, inflated, fully an inch long, and rounded at the ends. The introduction of a perfectly new and distinct arctic plant into the supplementary observations appended to this paper requires some notice here. I am indebted for it to to Dr. Seemann, who pointed it out to me, on his return from the Tiji Islands (after the first part of this paper was printed), as a plant omitted in his Botany of the ' Voyage of the Herald,' and as that alluded to by him as Oxytro^ris x>olaris in the narrative of that voyage. It is so like Astragalus alpinus, that it had been mixed with specimens of that plant, which is abiindant throughout Eskimo-land. I have sought in vain, through a very large suite of specimens of A. alp'uius and orohoides (which it also in many respects resembles), for another specimen of polaris ; it must therefore at present be considered as an addition (eighth) to the small list of peculiarly arctic plants men- tioned at p. 258, and the sixth to the species peculiar to Arctic N. "VY. America enume- rated at p. 267. OxTTKOPis camjyestris, DC. Many names are included under this, which represent species, varieties, and sjnionyms in the opinions of different authors. Of these, 0. borealis, DC, is referred (with 0. sordida) to campestris by Ledebour. Of O. Middendorffii, Trautv., I have seen no authentic specimens ; but Trautvetter's plate appears to identify it with a not uncommon form of the same plant. O. sordida, Pers. The plant of Eries, ' Herb. Normale,' is undoubtedly referable to campjestris ; Buprecht (Flor. Samojed.) goes at length into its characters, describes it as very near 0. campestris e. verrucosa, Led., from the Gulf of St. Lawrence. This latter, E^uprecht refers to borealis, DC, which differs from campestris " in the glandular verru- cose parts, and calyx covered densely with black hairs," — all very inconstant characters in arctic specimens. Nyman and Pries put O. sordida under campestris. Koch and Ledebour both regard it as a variety, /3. sordida. O. polaris, Seem., alluded to (but not described) in Seemann's ' Narrative of the Voyage of the Herald,' is founded in error. See Astragalus polaris, supra. O. arctica, Br. E^uprecht appears to refer the Arctic Siberian and Kotzebue Sound plant of this name to O. sordida. Ledebour refers Brown's arctica of Arctic America to a form of TJralensis. Torrey and Gray follow ' Elora Boreali- Americana,' in keeping it distinct, with the observation that it probably does not differ from TJralensis. O. nigrescens, Eisch. This is certainly a very distinct and remarkable form ; but I much doubt its permanent distinctness from O. TJralensis. In the ' Elora BoreaK- Americana ' it is regarded as probably very nearly allied to arctica {TJralensis). Spirjea chamtsdrifolia, L. The only apparently Arctic European habitat which I find DR. HOOKER ON THE DISTRIBUTION OF ARCTIC PLANTS. 325 recorded for this plant is the River Kolva in Samoiedland ; but, the precise spot not being indicated by Euprecht, it may not cross the arctic circle. Though a high northern spe- cies, it is not Scandinavian. Drtas integrifoUa, Vahl, was originally considered a variety by Chamisso and Schlech- tendal, but is kept distinct in ' Flora Boreali- Americana,' by Torrey and Gray, and Asa Gray. I have, however, vainly endeavoured to find any satisfactory characters for it ; the only one of any consequence, derived from the outline, &c. of the leaf, is not only most inconstant, but presents every transition to D. octopetala. As, however, the most narrow- and entire-leaved forms have been found only in the extreme north, no doubt due to excessive cold, the occurrence of integrifotia on the White Mountains would, if confirmed, be a curious proof of the permanence in a local variety of the most variable of all characters— the development of the leaf. Peck is (according to Pursh) the authority for the United States locality, which wants confirmation. Geum strichmi, Ait. This, though kept as a distinct species by Torrey and Gray and all authors, appears to differ from G. urbanmn only in its stout form, and petals longer than the calyx; it is described by American authors as an extremely variable plant. There are other species that might equally well be included ; but the whole genus wants working up. SiEVERSiA Immilis, Br. This is scarcely A^'orthy of ranking as a variety of S. Bossii, to which it is reduced by Ledebour and Torrey and Gray. RUBTJS jjrojnnquus, Eichardson, is B. arcticits y of Torrey and Gray, distinguished by having flagelliform stems. B. acaulis, Mich., is B. arcticus /3, Torrey and Gray, and B. pistlUatits, Smith. B. castoreus, Leest. Pries keeps this distinct from arcticus, both being Lapponian. See many excellent remarks on all these species in his ' Summa Veg. Scand.' p. 169 Wahlberg makes this a bastard between arcticus and saxatilis, occurring under two forms. B. trijlorus, Hook. This Pries regards as a hybrid between saxatilis and castoreus, if I understand aright his system of symbols and nomenclature. Torrey and Gray regard it as a distinct species, as does Asa Gray. Michaux and Seringe both ranked it as a variety of saxatilis. As castoreus and the European form of saxatilis are not found together on the American continent, Pries's view of trijlorus being a hybrid seems untenable. PoTENTiLLA Pennsijlvanica, L. Spain is the only European habitat for this plant. P. rubricaulis, Lehm., does not seem to diff'er at all from sericea, L. Torrey and Gray indi- cate numerous varieties of sericea, but keep Pennsylvanica a distinct species. When studying the Himalayan forms of Potentilla, I found it impossible to draw any line between the following varieties of sericea. Viz. Soongarica, Bge., conferta, Bge., multicaulis, Bge., holosericea, Nutt., multifida, L., Pennsijlvanica (3. sfrigosa, and y. Upinnatifula, T. and G. ; and it further appeared to me that P. sericea was connected with nirea by numerous inosculating forms. P. nivea, L. I have (under sericea) stated the difficulty of discriminating between this and sericea. The names included under it represent varieties, subspecies, or species. 2x2 326 DR. HOOKER ON THE DISTRIBUTION OF ARCTIC PLANTS. according to the different views of botanists. Thus P. Valdlana, L., is the nivea var. y of Torrey and Gray, and includes Jamesoniana, Grev. P. pulcheUa, Br., is a very arctic form, and according to Brown is Greville's sericea, L., between which plant and nivea it ap]iears quite intermediate ; all depends upon what arbitrary character is assumed to be the diagnostic one. P. Keilhami, Sommf. Of this I know nothing ; Nyman doubtfully refers the Spitz- bergen plant of this name to Brown's pidchella. P. frigida ? Grev., is referred here by Brown, and by Torrey and Gray under Vahliana, L., /3, which is also the P. Groenlmidica, Br., of the same authors. P. frigida, Vill. I am quite ptizzled with this. Arctic American and Greenland speci- mens seem to connect P. minima, Hall. {Brauniana, Hoppe), glacialis, Hall., and nana, Lehm. The United States plant, found only on the White Mountains (P. Robbinsiana, Oakes), is referred to minima j3 by Torrey and Gray, and afterwards to frigida, Vill., by Gray, who observes that the European minima is probably a variety. The Danish au- thorities refer the Greenland and Labrador plant to P. emarginata, Pursh {nana, Lehm.), whilst both Asa Gray and Durand refer it to frigida, VilL I cannot help suspecting that p. elegans, C. & S., of Eastern Siljcria, is the same thing, and that all will be found to inosculate with the varieties or forms of P. verna. I have confirmed Torrey and Gray's conjecture regarding nana, Lehm., being the same with emarginata, Pursh. P. verna, L. Under this, as one collective species or group of subspecies and varieties, I have introduced five North European arctic and alpine plants, whose forms inosculate perplexingly. The typical P. verna is not Lapponian, according to Pries ; and none of its included forms occur anywhere in N.E. America, except Labrador and Greenland (as- suming that hljlora is distinct as a species). P. aurea, L., is distinguished by Koch by its leaves silvery beneath ; the aurea of Smith he refers to aljjestrls. Hall. I find this character to be very variable. P. crocea, Schl., is referred by Koch to a variety of alpestris ; he has, however, a var. crocea of verna. P. maculata, Lehm., is, together with aurea and crocea, referred to Salisburgensis by Torrey and Gray. — P. Salisburgensis, Hgenke, is regarded by Koch as a synonym of al])estrls. Ruprecht, who observes that Salisburgensis is found throughout the arctic Island of Kolgujew, says that it is the same with '2;erra«,L., " genuina." P. alpestris, Hall. : this, which is a much larger-flowered plant than the ordinary verna, and otherwise difi'erent in habit, is kept distinct by Pries. Koch also distingxiishes it from verna by its habit and ovate stipules, and from aurea by the leaves not silvery beneath. Bentham regards both aurea and alpestris as luxuriant forms of verna. Watson suggests that the distribution of verna, alpestris, and Salisburgensis should be treated in Great Britain as that of one plant. Pragaria vesca, L. I have adopted the conclusions of J. Gay (Ann. Sc. Nat. ser. 4. vol. viii. p. 196) for the distribution of this plant, but very much suspect that it would have been more consistent to have (for distributional purposes) included that of collina, Ehr., canadensis, Mich., and Chilensis, Ehr., under it. Sanguisorba officinalis, L. The four species brought under this vary very much ; and I douljt if the majority will ever rank above local or permanent varieties. S. officinalis itself is DR. HOOKER ON THE DISTRIBUTION OF ARCTIC PLANTS. 327 European, Caucasian, and Siberian. S. media, L., which is reduced by Torrey and Gray to a var. of Canadensis, L., is N.W. American. S. Canadensis, L., is not found anywhere in Canada, though a native of both N.E. and N.W. America; it differs from officinalis in the long stamens and narrower spikes. S. tenuifoUa, Eisch., is N.E. Asian, and con- sidered by Asa Gray (Bot. of Japan, p. 387) as hardly more than a variety of Canadensis. So much for the local distribution of the common forms ; the only other is S. polycjama, Nylander, a species retained by Fries, with East Lapland as its only Scandinavian habitat. Rosa majalis, Herm. This is R. cinnamoii/ea /3, Torrey & Gray, which Ledebour and Koch do not even rank as a variety. JR. blanda, Ait. This is also B. Woodsii, Borr., of ' Elor. Bor.-Am.' according to Torrey and Gray. Eries does not include it in his Lapland column.. R. acicidaris, Lindl. (non DC), is referred by Ledebour to alpina, L., and appears to be the same with B. Gmelini, a common Siberian plant. I do not see how it is to be distinguished from hlanda. B. Carelica, E. : Ruprecht (Elor. Sam. p. 33) reduces this to a var. of B. acicidaris, as do Meyen (see Nylander, Dist. Plant. Eenn. p. 86) and Trautvetter (Elor. Boganid.). B. fraxinifola, Lindl. (and Seringa), is referred to blanda in ' Elora Boreali- Americana ' and by Torrey and Gray. B. villosa, L., tomentosa, Sm. Benthani adds to this B. mollis, E. Bot. B. canina, L., is, according to Ledebour, found in Arctic Lapland; l}ut I find no Scandinavian authority for it, and have not included it. Pyrtjs Aucicjmria, L. This, and the two plants brought under it, are species, subspecies, or local forms, according to the value assigned to those terms. It is remarkable that the Greenland one is nearer the American form than the European. P. Americana was Aucuparia |3 of Michaux. The true P. Aucuparia is, however, a native of both continents, having been found at Lake Winnipeg by Bourgeau, and in Siberia, lat. 6i° (Trautv. Elora. Jenissei). PuNUS Padns, L. I almost think that I should consistently have added P. serotina, DC, of North of America, to this. Amelaxchier alnifolia, Nutt. Torrey and Gray reduce all the American Amelanchiers to Canadensis, of which this is their var. S ; it is found on both sides of the Rocky Zvlountains. Epelobium alsinifolimn, Vill. Watson regards this as possibly a luxuriant form of alpinum, adding that intermediate forms may be assigned to either ; he also cautions us that mountain forms of E. tetragonmn occasionally pass for it. Bentham keeps it distinct ; Koch, Torrey and Gray, and Asa Gray make it the same with origanifolium. E. Sornemamii, Reich. This is alpinum j3 of 'Elora Bor.-Am.,' and y of Torrey and Gray. Ledebour keeps it distinct. K origanifolium. Lam. This is E. alpinmn -y of Torrey and Gray and Asa Gray ; Le- debour keeps it distinct, as does Eries, who gives Lapland as habitat for both. Andersson distinguishes these and lineare Ijy the seeds, which, ]iot being perfected in the majority of 3^8 DR. HOOKER ON THE DISTRIBUTION OF ARCTIC PLANTS. specimens from arctic habitats, are to me unavailable characters ; Koch also keeps them distinct, and characterizes them by their leaves. E. Uneare, Muhl. Fries keeps this distinct from jpa^H'Stre, L., both being found in Lapland. It is referred io paliistre, var. /3, in the ' Flora Bor.-Am.,' in Torrey and Gray's Flora, and in Asa Gray's Manual. Callitriche verna, L. I have treated this as one collective species, including autum- nalis, following Bentham's idea. Fries keeps this and cmtumnalis, L., distinct, giving Lapland as a habitat for both. Bentham unites them. A. Gray and most authors keep them distinct. Myriophyllum spicatnm, L., is not Lapponian according to Fries, though Lapland and Asiatic Russia are given as habitats by Ledebour. Hippums Montana, Led. I have not seen this plant, which, judging from Ledebour's description, would appear undoubtedly to be only a starved state of H. vulgaris. S. ma- ritima, Hellen, is reduced by Wahlenberg, Fries, and Koch to a variety of vulgaris ; on the other hand, it is retained as a species by Bro^vn, Ruprccht, ' Flor. Bor.-Am.,' and Torrey and Gray. Both this and vulgaris are arctic. Ceratophyllum. Owing chiefly to imperfect materials, I am obliged to treat Cerato- jjliyllmn as consisting of one collective species. The specimens I have examined do not present such definite characters in the fruit as the descriptions of authors would seem to imply. Bentham regards siihmersum and demersum as forms of one ; Torrey and Gray suggest that there is but one American species ; Fries, Koch, and Watson keep these two distinct. Neither of them are Arctic European or Asiatic, though found in Arctic America. Claytonia lanceolata, Pursh, arctica, Adams, and sarmentosa, C. A. M., are undoubt- edly trifling varieties of one plant, neither constant inter se nor as to locality. MoNTiA/oji^cmffl, L. The absence of this cosmopolitan plant in all parts of Temperate America east of the Rocky Mountains, except Labrador, is one of the most remarkable facts in botanical geography. Newfoundland is given as a habitat by Le Pylaie ; but this, like that of Calluna vulgaris, wants confirmation. 31. rivularis, Gmel., is reduced to a variety by Fries and Torrey and Gray. Sedum elongatum. Led. I do not know this plant by authentic specimens. Ledebour gives it as a doubtful species, confounded with *S'. Rhodiola by Chamisso, Hooker and Arnott, and Torrey and Gray. S. eupliorUoides, Schl. Of this I know nothing; according to Ledebour it is pro- bably a variety of the Siberian S. algidmn, Led., itself very near to quadrifidum. Pall. RiBES propinqimm, Richardson. This does not differ from R. ruhrnm, L., to which it is reduced as var. ^ by Trautvetter and Meyer in ' Flora Ochotsk.' Chrysosplenium tetrandrum, Lund, is noticed by Fries as a variety of alternifolium, also occurring in Lapland. I have not seen authentic specimens. Parnassia obtusijloi'a, Rupr., appears nothing but a small specimen of palustris, L., DR. HOOKER ON THE DISTRIBUTION OF ARCTIC PLANTS. 329 in wliicli the petals are often as short and of the same shape as in Rupreeht's plant, and the cauline leaf wanting. I doubt much if P. Kotzebwei is really distinct. Saxifraga Alzoon, Jacq. Koch, Pries, and most authors keep this distinct from S. Cotyledon ; but I must own that I cannot satisfactorily discriminate their forms or make them tally with their assigned habitats, and am thus obliged to treat them as one plant. Both, according to Fries, grow in Lapland and Western Europe generally. Cotyledon alone is stated to be Icelandic : Aizoon, which alone is Greenlandic, Labrador, and N. American, is nowhere stated to occur in the Russian dominions. S. ccBspitosa, L. It is not my purpose to enter into the disputed question of the limits of the members of this aggregate collective species. I have repeatedly examined all, and found innumerable grades connecting the most dissimilar, such as exarata and sileni- flora. Almost all are high arctic. 8. uniflora is the same as venosa, and is referred by Torrey and Gray to ccespitosa, to which ^S*. Groenlandica is reduced by almost all authors. S. exarata, Vill., is kept distinct in ' Elor. Bor.-Am.' (with an appended observation regarding the difficulty of distinguishing it), as also by Torrey and Gray, Lcdebour, and most authors. S. muscoides, AVulff., is the common Middle and S. European form. 6". sileiiiflora is confined to Arctic America ; and Jilagellanica, in a form identical with exarata, does not extend in S. America north of Peru. S. controversa, Sternb., which, according to most authors, is a synonym of adscendens, has been found on the Rocky Mountains by Bourgeau and Hector during Palliser's Expe- dition, but nowhere else in Continental America. S. exilis, DO. This, as suggested in ' Elor. Bor.-Am.,' is, so far as may be judged by specimens tallying with the description, certainly only a weedy state of cernua. S. bidUfera, L. Koch distinguishes this from granulata by its leafy cyme and other characters that appear to be clearly connected with its being a bulbilliferous condition of S. granulata, between which and bulbifera I find intermediate grades. Ledebour and Eries also keep it distinct. It is not Arctic Lapponian, but Russian. S. Ivyperborea, Br., is referred to r'wularis /3 in ' Elor. Bor.-Am.' and by Torrey and Gray ; it is often a starved high-arctic form, but not a permanent or well-marked one ; it is con- fined to Arctic America, and Greenland. S. coriacea, Adams. This appears to me, from the description, not to be distinct from nivalis. I have seen no authentically named specimens. S. reflexa. Hook. This is certainly only Virginiensis, of which several specimens from the Rocky Mountains precisely accord with the arctic. S. hieraciifolia, W. & K., though Arctic Russian, fid. Ledebovir, is not Lapponian, according to Eries. . S.foliolosa, Br. This is a monster, as pointed out by Eries {comosa), rather than a variety proper, the foliaceous cyme and bulbilli being produced at the expense of the inflorescence proper. Ledebour reduces it, and Ruprecht notices its transition to stella- ris, L. S. propinqua, Br., is identified with Hirculus, L., in 'Elor. Bor.-Am.' and by most suc- ceeding authors. 330 DR. HOOKER ON THE DISTRIBUTION OF ARCTIC PLANTS. S. cestivalis, Eiscli., is reduced to 2>iinctaia by Ledeboui-. CoNiosELiNUM Tcitaricum, Eiscli., and Selinum Gmeliiii, Dc Bray, are both synonyms of C. Fischeri, Wimm. Archangelica littoralls, Fries. Walilenberg originally regarded this as a variety, and, I think, rightly, its characters being slight and of relative value only. Ledebour doubts its claims. Fries, who keeps it distinct, discusses the species (Summa, p. 181). A. Norvegica, Tabern., is not contained in Fries's ' Summa.' Uuprecht (Flor. Samojed.) gives it as the Angelica Archangelica (3 of Linneeus, and adds that it is most highly esteemed as an article of food. A. atropurpnrea, Hoffm. This is the "Common Archangelica" of American authors, which, according to the specimen I have seen and descriptions I have compared, differs A^ery slightly from Angelica Archangelica, L. The Pltijsoloph'mm saxatile, Turcz, {Coelo- 2Jleurum [Fleurosj^enuiDn, Bong.J Gmelini, Led. fid. Flor. Ochotsk.), is, as far as I can make out from descriptions, either a form of Archangelica officinalis, or some very closely allied plant which I cannot distinguish from its forms. Athamanta arctica, Nym., is Libanotis arctica, E,upr., a plant of which but one specimen is known, differing, according to its author, from condensata, L., in the colour of the flower only. DeCandolle reduces condensata to a var. of A. Libanotis, L. ; and A. Sibirica is only another form of the same species. Heracleum arcticmn, Rupr. I have not seen this plant, which, from a careful compa- rison of its description with H. Sibiricum, I take to be an arctic state of that plant. Pachypleurum alpimmi, Led., is the same with G-aya {Neogaya) simplex, Gaud. SiUM latifolium, L., is stated by Gunner to be an inhabitant of Tromsoe (Wahlenberg) ; but I do not find so northern a habitat confirmed by Fries or Andersson, and so have not included the species. Viburnum Opuhis, L., scarcely reaches the arctic circle in Lapland. V. Oxy coccus, Pursh, is finally identified with it by A. Gray. Galium uliginosum, L. Watson remarks that states of this are frequently confounded vf\i\i palustre and jyiisillum, which may well be the case with arctic specimens. G. trijidum. L. Ruprecht calls the European plant /3. EtiropcBum, and indicates some differences, which I do not think hold in the extensive suite of sj)ecimens I have examined. G. Claytoni was thought to be distinct in ' Flor. Bor.-Am.,' but is reduced by Torrey and Gray, and by Ledebour. G. infestum, W. K. Fries considers this to be a variety of Ap>arine ; and it alone is Lapponian. G. saxatile, L., docs not advance north of Southern Norway, nor west of Uralian Siberia ; hence its being arctic in Greenland is remarkable. It is also Icelandic. Nardosmia frigida, Hk., corymbosa, Hk., and sagittata, Ilk., are clearly all forms of one rather variable plant, connected by intermediates. N. straminea, Cass. United with N. Icevigata, DC, by Ledebour ; but Huprecht con- siders it different. DR. HOOKER ON THE DISTRIBUTION OF ARCTIC PLANTS. 331 Ptabmica alpina, L. The numerous plants here brought together do not cliifer more inter se than do the forms of F. vulgaris (with which this is intimately connected), and they are further all united by intermediate forms. P. aljnna itself is found in Arctic Siberia and the Altai; P. oxyloha is only known in Arctic Siberia. P. speciosa, DC. : Ledebour rightly remarks that this is rather a form of vulgaris than a species. P. mul- tijlora, Hk., was originally referred to a var. of vulgaris, from which it differs in its deeply cut leaves, and broad-, short-ligulate flowers. P. Sibirica, Led., is confined to East Siberia and Escholtz Bay. Of the boreaUs, Bong., of Sitcha, I have seen no authentic specimens. Chrysanthemum Sibiricum, Turc, is retained by DC, Ledebour, &c., but apparently on very slender grounds. C integrifolium, Rich., is an eminently northern form, not found west or north of the Tchuktchi country, and scarcely south of the polar circle ; it may prove to be connected with leucanthemutn by intermediate forms. C. leucaiithemum, L., is not Lapponian, but Arctic Russian. C. Segetum is occa- sionally found in Nordland, but apparently rather as an alien than a genuine arctic 'plant. PYRETHmrM pulchellmn, Turc, is certainly identical with the small arctic form of ino- dormn; and Fries adds as a variety the P. ambiguum, Ledeb., correctly. Ruprecht con- sidered the woolly hairs on the peduncle ot F . pulchellum suflQ.cient to distinguish it ; but this is not the case. Artemisia Tilesii, Led. Fries ranks this as a var. of vulgaris, L., both being found in Lapland. Seemann remarks that they occur together throughout Western Eskimo-land. A. Chamissoniana, Bess., is reduced to arctica. Less., in 'Flor. Bor.-Am.,' and by most succeeding authors. A. Norvegica, Fr. : in the ' Flor. Bor.-Am.' it is stated that this is identical with A. Chamissoniana ; on comparison I find it has rather larger capitula, but is not otherwise different ; it is a very rare alpine plant in Norway, not entering the arctic circle. A. globularia, Cham. I think this is quite tlie sam.e with, glotnerata ; both are con- fined to the country around Behring's Straits. A. androsacea. Seem. This is a remarkable and very local plant, which I cannot identify with any more southern species, though, judging from analogy, I suspect that it may eventually prove only an arctic tufted variety of some better-known plant. A. ccBspitosa, Bess. I have seen no specimen of this ; but, judging both by plate and description, I do not doubt that Torrey and Gray's surmise is correct as to its being the same with Richardsoniana. A. Canadensis, Mich. This, according to Richardson, advances to the arctic circle east of the Rocky Mountains. After a careful re-examination I cannot doubt its identity with A. desertorum, Spr., as also certainly with commutata (3. Gebleriana, Bess, in Herb. Hook. Canadensis is the earliest name. BiDENs tripartita, L., is not found in Arctic Lapland, though it enters the arctic circle in Russia. I am quite unable to distinguish the American B. connata, Muhl. VOL. XXIII. 2 Y 332 DR. HOOKER ON THE DISTRIBUTION OF ARCTIC PLANTS. Antennaria liyperborea, Don, is considered by Fries to be a corymbose form of clioica, Br. Gnaphalium Nonegicimi, Gunn. Fries and Kocli consider this a distinct species from sylvaticum, both being Lapponian. Watson finds the British Norvegicum so linked to sylvaticum, that he is obliged to treat them as one. Bentham considers Norvegicum a high-arctic and alpine form of sylvaticum. - G. pilulare, Wahl., is reduced to a var. of uligiiiosnm by Fries. Abnica montana, L., certainly includes the three v.hich I have included under it, and which hardly deserve to be ranked as constant varieties. Fries keeps alpina distinct, and assigns it a more northern habitat than montana. Torrey and Gray unite angustjfolia, Vahl, with alpina, Lsest., and make of it montana /3. alp)ina. The angiisti- folia, H. & A., is by the same authors considered the same as alpina, and called angusti- folia (5. Lessingii ; they observe, at the same time, that the arctic forms, though so very closely approaching, appear quite distinct in more southern habitats. Durand (Kane's Voy.) unites alpina, Wshl., 23lantaginea and fulgens, Pursh, with montana, var. alpina, L. Senecio resedcefoUiis, Less., is Arctic Russian, but not Lapponian. S. arcticus, iiupr., is Cineraria congesta, Br., " senstt latiore ?" of Fries; and C. congesta, Br., is reduced by him to a northern var. of ^S*. palnstris. The absence of both this and campestris in Greenland is very remarkable. S. campestris, L. I have very carefuUy compared many specimens of the plants placed under this, and am quite unal:)le to discriminate between their forms. According to Koch and Fries, Cineraria alpina includes C. campestris, Retz. C aurantiaca, Hoppe, is kept distinct by Koch, but united by Fries as a var. of C. alpina. S. octoglossns, Ledeb., is referred by Fries to nemorensis, which in Scandinavia is not found south of Lapland. LiGTJLARiA Sibirica, L., is found only in East Lapland and Arctic Russia, Finland, &c., not in Nordland. SoLiDAGO mnltiradiata, Alton, is the Virga-aurea y of Torrey and Gray, and e. arctica of Ledebour. Aster Richardsoni, Spr., and montanus, Richardson, are referred to Sibiricus by Fries, and, I think, rightly ; also doubtfully by Pi-uprecht. Trautvetter (Flora Boganid.) refers Richardsoni to Sibiricus /3. subintegerrima. A.pygmceus, T. &. G. I have vainly sought to distinguish this from salsuginosus. A. alpinus, L. (with whicli A. flaccidus, Bunge, of Siberia and Dahuria appears to be identical), is not repiited to be Scandinavian, but a native of the Alps and Carpathians. A. falcatus, LindL, is certainly nothing but multiflorus. Ait. Erigeron trifidus, Hook., is identical with compositus, Psh., as is now well known. E. elongatus, Led. Fries, who keeps this distinct, says (in Herb. Norm.) that it should rather he considered a form of acris, of which it appears to me to be a mere variety, and a very slight one. E.poHtm, Fr. This also Fries keeps distinct ; he considers it most allied to alpinus, and DR. HOOKER ON THE DISTRIBUTION OF ARCTIC PLANTS. 333 includes under it glabraUis, Hook. For my own part, I find it imjoossible to separate the forms of acrls from those of al-pinus, and am rather disposed to refer this to the former. E. iiniflorus, L. This is E. alpmtts j3 of Ledebour, and alpimis y of De CandoUe, and is united with alpinus by Bentham also. According to Eries, who unites pulcliellus, DC, with it as a variety, it is a good species. I cannot regard it otherwise than as an arctic and alpine state of alpimis, with a more woolly involucre, fewer capitula, and better-developed inner-ray florets I find all intermediate forms. E. purpureus, Ait., is included under Fhiladelphiciis by Torrey and Gray, &c. Taeaxacum Bens-leonis, Desf. Eries adopts the name officinale, Weber, and includes palustre as a variety, both being Lapponian, in which Watson, Koch, and most botanists coincide. T. ceratoplioo^um is rather a form with an over-developed condition of the invo- lucral scales, than a variety properly so called. T. Scorzonera, B^eich., from Arctic Siberia (Flor. Taimyr.), is not included in Ledebour's ' Plora Rossica.' T. phymatocarpimi, J. Vahl, Flor. Dan. 2298, found in Arctic Greeland only, is a small form with the habit and involucre of T. palustre, Sm., and shorter achenia than is usual in the geniis. SoNCHUS maritimus, L., is kept distinct by Koch, bu-t included by Fries as a variety of arvensis. Leontodox autumnalls, L. I am obliged to regard Keretimts, Nyl., as the same with auhimnalis, because Ball (Ann. Nat. Hist. 1850, p. 2) observes that the character of erect and drooping capitulum (on which this species appears founded) is not to be relied on in other species of ■Jhe genus. I have never seen copious specimens of this L. Keretinus, which Ball omits in his otherwise very full account of the genus (/. c.). Fries regards it as a not yet fully developed species. Ledebour puts it in another section of the genus. Koch makes L. Taraxaci, L. {Ap>arg'ia, Smith), a variety of autimmalis ; and Watson says that the British Taraxaci graduates insensibly into autumnalis. Ball makes it a sjTionym. Fries introduces under autumnalis a variety nigro-lanatum ; he describes both as being common throughout Lapland. HiERACiTJM. I have grouped the various arctic forms discriminated by Fries, into a few aggregate species, such as are recognized by Arnott, Bentham, Watson, &c. Satjssukea alp'vna, L= I have treated this as an aggregate species, because S. mida, mon- ticola, and angustifolia all seem to graduate into it, and several of them into one another I should not be surprised if ^S*. subsimiata, which is hitherto known from one locality only, proved still another form of the same variable plant. Cakduus crispits, L. This attains a higher latitude than C. acanthoides, L., of Tiiiich Bentham and otliers consider it a variety. The two certainly pass into one another. Campanula linifolia, Hsenke, and Scheiizeri, VilL, are certainly referable to C. rotundi- folia, as held by Linnaeus and most modern authors, though ScJieuzeri is regarded by Fries as distinct. 2 y2 334 DR. HOOKER ON THE DISTRIBUTION OF ARCTIC PLANTS. Y ACCii^iVM. 2yubescens, Wormsk. (Plor. Dan. 1516), a Greenland plant, is referred to uliginosmn variety by Lange. Vaccinium (Oxyococcus) microcarpmn, E,upr., is included under O. palustrls by Ny- man, and is clearly nothing but a small-fruited state of that plant. Calluna vulgaris, L., is mentioned in DeCandoUe's ' Prodromus,' on the authority of a specimen gathered by La Pylaie, as a native of Newfoundland ; but I find no confirmation of this habitat, nor is it found in any part of the American continent. In the 01d"\\^orld it wanders no further east than the Ural district. Ledum j)alustre, L. Asa Gray and DeCandolle distinguish two species, by the stamens 5 and 10, leaves broad and narrow, and capsules ; but these do not affect diff'erent geo- graphical ranges, and Andersson, a most careful observer, says (Conspect. Veg. Lapp. 18, in note), " inter banc varietatem {laUfolitmi) et normalem tarn multi animadvertuntur transitus, ut nuUo modo distingui possint. Nostra planta foliis etiam latioribus, quam specimina Groenlandica, nonnunquamlecta." Of the three included species, L. dUatatum is ranked by Fries with palustre, to which also Grcenkmdicum is referable — L. latifolium. Ait., is merged into palustre by Michaux and in ' Plor. Bor.-Am.' Pyrola rotimdifolia, L. This I consider includes cJdorantha, Sw., which, though not found in Arctic Europe, is said by Du.rand (Kane's Yoyage) to be found in Greenland. As, however, Durand does not include grandijlora, Rad., and Lange makes no mention of chlorantha, it is possible that the same species is intended by both authors. P. Groenlandica is referred to rotimdifolia in Flor. Bor.-Am., and by DeCandolle ; it is kept distinct by Lange, who refers it to grandijiora, Had. P. elliptica, Nut., is certainly only a variety of rotimdifolia, and is in most respects in- termediate between rotundifolia and chlorantha. P. media, L. Bentham is inclined to doubt if this be permanently distinct from P. minor. Ruprecht (Flor. Samojed.) mentions P. uniflora, L., as a doubtful native of Ivolgujew Island, off the White Sea. Gentiana lingulata, Ag. Fries treats this as a distinct variety of Amarella, which alone is Lapponian. After a very careful examination of G. acuta, Mich., I am disposed to regard it as the Lapponian form of Amarella ; I cannot fix characters that will distin- guish them. Grisebach says of it, " Amarellce simillima sed bene distincta species." G. involucrata, Rottb. This very rare plant is also Icelandic. G. cestiva, R. & S., is va^na y of Grisebach ; it is not found in Lapland, any more than Fleurogyne rotata ; both, however, are Arctic Russian. PoLEMONiUM ccernleum, L. I have treated all the arctic forms of this variable plant as constituting an aggregate species. All are regarded as varieties in ' Flora Boreali- Ameri- cana.' Ledebour makes pulchellum different, and includes capitatum and humile under it ; these, however, graduate quite insensibly into cceruleum. Ruprecht finds both in the arctic island of Kolgujew. It is very remarkable that this plant inhabits no part of Greenland but the east coast only, and at a very high latitude. DR. HOOKER ON THE DISTRIBUTION OF ARCTIC PLANTS. 335 Phlox Bichardsonii, Hook. This appears to me only an arctic tufted form of P. Sibi- ricci, to which, I think, may also be referred P. Douglassii ; all have more or less, and often strongly, recurved margins to the leaves ; and the other differences are comparative and trifhng. Mtosotis alpestris, Sw. {simveolens, M. & K.). Tries and Koch merge this into a variety of s^foa^jcff, which alone inhabits Lapland. Bentham takes the same view. Watson retains it, finding that it keeps its characters well under cultivation. M. casspitosa, Schltz. This Bentham unites with palustris, and probably correctly. M. arvensis, L. {intermedia, Link.). "Watson finds this occasionally approximating to ccBspitosa ; and I find it difficult to separate northern forms of one from the other. Eretrichiitm villosum, Bunge. E. aretioldes, A. DC, is nothing hvA a dwarf arctic state of this. Ledebour unites E. latlfoUnm, Rupr. (non Kar. & Kir.), with the same, as var. (3 ; and so does Trautvetter (Elor. Taimyr.). Mertensia pilosa, DC, which includes Lith. corymbosum, Lehm., and p)cmiculatum, Don, is clearly referable to denticulata, Don, the hairy calyx being a very inconstant character. These should possibly all be united under Sibirica. M. Drummondii, Don. I find no plicse in the tube of the corolla of this plant, whence it must be removed from the section in which DeCandoUe places it to that with Virgi- nietisis, of which it appears to be a northern form, as suspected in ' Plora Bor.-Am.' It has not, however, been gathered anywhere between the Arctic Sea-coast and the United States. Mentha Lapponica, Wahl., is referred by Fries to a northern variety of arvensis. Both grow in Lapland. Galeopsis Tetrahit, L. Fries and Koch keep this distinct, as does Watson, who gives as presumptive evidence in its favour the statement that the seeds of versicolor yield plants of their own kind only. Bentham says that they graduate one into the other. ^'yikCUY^ palustris, L. According to the synonyms quoted by Asa Gray, this N. American plant includes many very divergent forms, including some that might almost be referred to S. sylvatica. Under the latter plant I have not introduced the Chilian S. chonotica, which approaches it closely. LiMOSELLA aquatica, L. I include under this L. tenuifolia, which is the more common southern form, and certainly is nothing but a variety with redu.ced foliage. L. borealis, Lessing, is another form, found in Lapland only, according to Tries. Gtmnandra borealis. Pall. I am quite unable to distinguish Stelleri and JPallasii, which were included originally by Pallas under borealis. Choisy says, in DC Prodr. xii, 24, that it is difficult to dissent from Pallas's view, who regarded the genus as monotyplc, Willdenow makes eight species, which Chamisso and Schlechtendal reduce to three. Castilleja sep)tentrionaMs, Lindl. I have no hesitation in uniting this with pallida, as suggested ' in Plor. Bor.-Am.' It advances south to Canada. 336 DR. HOOKER ON THE DISTRIBUTION OF ARCTIC PLANTS. Veronica horealis, Lsest., is included by Pries as a variety of serpylUfoUa, both, being Lapponian. V.fruticulosa, L. Koch retains this as different from saxatilis, butBentham considers it the same. Melampyrum Americanum, Mich. Asa Gray keeps this up as a species, but Bentham reduces it to a var. o^ i^ratense. The smaller, more slender corolla is the only character I can find. Bhinanthus minor, Ehr. This form of the genus is, according to Fries, the only one found in Lapland ; it is the Crista-galU, Aar. a, of Linnaeus according to Koch. Bentham does not consider it a sufficiently constant form to rank as a race even. Euphrasia officinalis, L. The varieties montana, Er., and alpestris, Koch, are found throughout Lapland according to Eries. PEDicuiiARis amcena, Ad. Retained as distinct from verticillata by Ledebour and by Bentham, who, however, on examining more specimens with me, agrees that it is only a slight variety connected by intermediates ; both are arctic plants. P. horealis, Stev., is, according to Eries, a Lapland and Einland variety of |jffZwsi;r'is,L. P. lanata, Willd., is reduced to hirsuta hj'&Gnih.sim, and, I think, rightly. P. arctlca, Br., is reduced to Langsdorffii by Bentham, with purpurascens, for an old synonym. P. Langsdorffii. This I find to pass by insensible gradations into Sndetica. P. Kanei. I have never seen authentic specimens of this, but the description agrees with forms of undoubted Sudetica gathered in Greenland. Utricularia vulgaris, L. This is unquestionably a native of Arctic Europe ; but though common in Temperate N. America, both east and west, I am not so sure of its entering the arctic circle there. I presume, however, it is one of the two alluded to by Ptichardson (' Boat Journey through Bupert's Land '). V. intermedia, Ehr.. Eries and Koch keep this distinct from U. minor. Watson also retains them, but observes that much confusion exists about them. Bentham and Oliver (who has lately worked a great deal at this genus) think that intermedia is only a form of minor. Of the two, JJ. intermedia is both Arctic European and American ; minor is Arctic European, but is also found in Tibet. Do3:)ECATHEOiSf integrifoUuin, Mich., and frigidiun, C. & S., are certainly only varieties of Meadia. Torrey indeed (Bot. Whipple's Exped., p. 62) recognizes but one species of the genus, which is found from the arctic circle to New Mexico. Trientalis latifolia, Hook., is considered a variety of Europa;a by Torrey (Bot. Whipple's Exped.) ; it extends along the Rocky Mountains to California. T. arctica, Eisch., is scarcely distinguishable from Enroptea as a variety, to which Lede- l)our refers it. I have seen no arctic specimens, though it extends to Sitcha and Kams- chatka. The T. Americana, the best-marked American species, extends from tlie U. States to Labrador. It differs in appearance from Europcea, and is certainly far more DR. HOOKER ON THE DISTRIBUTION OF ARCTIC PLANTS. 337 distinct from it than either arctica or latifoUa. Torrey at first considered it a variety ; but both he and Asa Gray have since regarded it as a species. Androsace triflora, Adams. This, which is unkno^vn to Ledebour, Duby, and myself, appears, from the description, not to differ from ChamcBJasme. A. Chamcejasme, though more properly a West Arctic American plant, extends eastward to Victoria Land. Ledebovu' (Flor. Eoss. iii.18) gives Kotzebue Sound as a habitat for villosa, quoting, erroneously, Beechey's ' Voyage.' I do not think it is anywhere an arctic plant. Primula strlcta, Horn. I find it difiicult to distinguish some arctic forms of this from others of farinosa, but think I have given its distribution correctly. P. Hornemannimia is now generally admitted to be a synonym. P. Mistasslnica, C. & S. (non Mich.), is referred by Ledebour to P. strict a ; and P. Mistassinica, Mich., was united with the same plant in the 'Elora Boreali- Americana,' with probably good reason according to Duby. P. borealis, Duby. Ledebour says of this, " planta mire varians." Ruprecht suggests that it is only a var. of strlcta, and rightly, I have no doubt. P. nivalis. Pall. The distribution of this plant is peculiar. It is found nowhere in Europe, except the Caucasian provinces be considered such ; it is, however, a native of all Siberia and N. W. America. P. saxlfragcBfolia , Lehm. Ledebour reduces this to cimeifolia, and no doubt correctly. P. Scotica, Hook. This is a form distinct enough in many places, but graduates into P. farinosa, with which Bentham joins it. Pries keeps it distinct, as does Watson, who remarks that its characters depend chiefly on its larger purple corolla. JP. farinosa itself, though a native of Pinland, scarcely extends into Lapland. P. Finmarchica, Willd. Fries says the flowers are purple, and distinguishes it from Sibirica, observing that it is arctic, and not alpine. Ruprecht, Ledebour, and Duby make of it Sibirica (3. Armeeia vulgaris, L. I do not see how the distribution of the plants named under this can be treated apart. Of these A. alpina is an inland form found in the Alps of various parts of Europe, though not Scandinavian according to Pries ; it is the uilriiieria 7 alpina of DeCandoUe. A. elongata, Hoffm., is the only one entered as Lapponian by Pries ; it is referred to Armeria a by Ledebour. A. Labradorica, Wallr., is vulgaris s of Meyen's Labrador plants. A. arctica, Rtipr., is A. vulgaris of Nyman, and vulgaris /3. arctica of Ledebour. A. Macloviana, Cham., is the same with Andina, which Torrey (who has a var. Californica) refers to vulgaris, observing tliat many of the species broken off from vulgaris had better be referred back (Bot. Whipple's Exped. p. 62). RuMEX Rippolapathu7n, Pr., is reduced to aquaticus by Meisner in DC. Prod., and in part by Ledebour. R. arcticus, Trautv. I am unacquainted with this plant, which would seem not to differ from R. aqti.aticus in any important character, or in distribution. Trautvetter (Plant. Taimyr.) observes that it is perhaps GhMnisso'sYariety oi do))iestic us with a simple whorled raceme. R. domesticus, Hartmn., is aquaticus j3 of Wahlenberg. Pries i^educes aquaticus, L., to a var. of domesticus. Both are Lapponian. Koch separates them, but by characters 338 DR. HOOKER ON THE DISTRIBUTION OF ARCTIC PLANTS. hardly tangible ; Leclebour, who keeps them distinct, thinks it would be better to unite them. Meisner refers domesticus to longifolins, DC, and makes of the Arctic American plant the var. nanus. B. salicifoUus, Weinm. I am doubtftil as to the merits of this species, and cannot come to any conclusion about it. PoLYGOXUM 23olijniorplmm, Led., is not Scandinavian ; but several of its varieties are Arctic Siberian and American, as /3. setigerum, y. lapathifolkmi, and frigldum (P. aljn- nwn, ALL). P. lapathlfoliiim, lu Kept distinct from Tersicaria by Fries, Watson, Koch, and Bentham ; but I find it impossible to distinguish North Indian specimens of one from the other, these being united by every intermediate form. P. Fersicaria is, according to Fries, very rare in Lapland, and sporadic only. Ledebour includes P. Sydropiper as Arctic Lapland on the authority of Felmann ; but I find no confii-mation in the works of Fries and Andersson. Chenopodium maritiimim, L., is doubtfully mentioned as a native of the arctic sea-coast of America by Richardson. I have examined the specimens ; they are very young, but identical with maritlmum, which is a common subarctic plant. MoNOLEPis Asiatica, Moq. I know nothing of this plant. The only recorded habitats I find are both Arctic Siberian, viz. Nishni Kolymsk and the Boganida River. Atbiplex patula, L., It is impossible to unravel the synonymy and distribution of this plant and A. liastata, if, indeed, they really be distinct, which Moquin dovibts. Fries keeps one liastata distinct, assigning it a place in the section with rugose seeds. Koch distinguishes it by the cordate-triangular (not hastato-rhomboid) sinuate-toothed perigonia. Moquin unites angustifolla with it. Bentham unites both these with deltoides, littoralls, and erecta. See, for further remarks in reference to the British species. Woods (Proc. Linn, for April 17, 1849), who observes that in several species the seeds are of two forms in the same individual,— one form shghtly depressed, smooth, black and shining ; the other (in larger lower perigonia) three times as large, more depressed, chesnut and crinkled : he admits angustifoUa, patttla, deltoidea, and perhaps erecta. On the south- east coasts of England, I recognize three very distinct forms, often growing intermixed, viz. A. littoralls, L., A. patula, L. {erecta, Huds.), and angustifoUa, Smith (all fairly well represented in English Botany). Of these the first and last are Arctic European and Arctic W. American ; and lilioralis Arctic W. American only, though common in tem- perate America. A. deltoidea, Bab. Moquin refers this to hastata; Fries makes a variety of this (prostrata) a native of Lapland ; Watson includes its distribution under that of ptatula (see remark's, ' Cybele,' ii. 324). A. angiistijolia, Sm. Moquin makes of tliis a synonym oi pattda; Watson keeps it distinct, but says nothing in its favour ; Koch brings it to patula, and both to latifolia, Wahl. A. Gmelini, C. A. M. A little-known plant, referable, according to Ledebour, to pattda, but rather, I think, to littoralis, L., or perhaps to angustifoUa, Sm. DR. HOOKER ON THE DISTRIBUTION OF ARCTIC PLANTS. 339 Urtica gracilis, Ait. Tliis is reduced to dioica, var. e, by Weddell. Betula. I am quite unable to disentangle the species, forms, and varieties of this genus, or to harmonize the views of Fries, Spach, Regel, and other botanists. The Northern species are clearly most difficult of discrimination, as a reference to Fries's notes (Summa, pp. 211 and 556) will show. Kegel's recent ' Monograph ' seems carefully and judiciously executed. B. glutinosa, Wallr., is now by Fries and others considered one with alba. Koch refers it to.pubescens, Ehr., /3. Carpatica ; Kegel (Monog. Betulac. p. 21) to alba S. glutinosa. B. intermedia, Wahl., is alba [3 of Ledebour, an Arctic Russian plant ; it is the Immilis of Hartm. and Rupr. (non Schrank). Ledebour observes that it approaches to nana very closely. B. papyracea, Ait., is reduced to a var. of alba by Spach and Regel, and possibly rightly ; but the American botanists, who know both, keep them distinct. B. nana, L., is, according to Asa Gray, distinguished by a narrowly winged fruit; but the wing of Scotch specimens is very broad, and Ledebour remarks that this is a variable character ; Regel says it is narrow or almost absent. B. glandidosa, Mich. This, according to Asa Gray, is the pmnila, L., distinguished by its cylindric catkins and broad-winged fruit. I have seen no Greenland specimens of it. It is the nana ■y. intermedia of Regel, who keeps ptmiila, L., as a distinct species ; he apparently has not consulted Asa Gray's ' Botany of the N. U. States.' B. alpestris, Fr., is the Immilis, Hartm., fid. Fries, and is so like British specimens of nana as to be scarcely distinguishable. I think it may well be included under /r?/iicos«. Pall. ; Ruprecht, however, refers Immilis, Hartm. (non Schrank), to intermedia {alba, var. intermedia, Walil.) . Regel makes of it nana S. alpestris, and possibly rightly. Alntjs barbata, C. A. M. Fries keeps this distinct, and gives as a hal3itat Lapland, where glutinosa does not grow. H is the glutinosa 7. barbata of Ledebour, and ^.pubescens of Regel. A. viridis, DC. Ledebom' distinguishes the plant of Chamisso, from Arctic America, from DeCandolle's ; but to me they seem identical. A.fruticosa, Rich. This, in the list, p. 301, is a misprint for glutinosa, Hiclmvdson (non Willd., &c.), which has been referred to viridis in 'Flor. Bor.-Am.,' and rightly. A. repens, Vahl. The specimens I have examined of this appear identical with Richard- son's glutinosa, and with A. viridis, DC. It is omitted in Regel's Monograph. Salix. I am indebted to Dr. Andersson of Stockholm for the Arctic List of this genus, who has further kindly supplied the following observations. Salix pentandra. In Lapponia et Siberia certissime occurrit, in Lapponia ad locum Enare (maxime septentrionem versus !), in terra Samojedorum, in Kamschatka infra circulum polarem. In America sub forma lucida ad Fort Franklin ct Mackenzie River. Ibi itaque non ad floram arcticam pertinet. >S'. lanata, L. In omnibus regionibus arcticis vulgaris, S. mdiardsoni, H., est S. lana tee {orraa Americana: cf. And. I. c. p. 3 3. S. speciosa, H. et A. Species elegantissima, Americce decus ! VOL. XXIII. 2 z 340 DR. HOOKER ON THE DISTRIBUTION OF ARCTIC PLANTS. S. Lapponum, L. In Lapponia et Xorvegia Aoilgaris. Terra Samojed. Siberia arctica ad Obdosck, etiam ibi I'ara. Ex America valde dubia : cf. And. I. c. p. 17. S. Stuartiana, Sni. S. Lappoimm est forma. S. eaprea. In Scandinavia ad promontorium boreale usque. In Hossia arctica (Kiola, terra Samojed.); in Siberia regiones tantum meridionales amat. In America non adest (ibi >S'. capreoides. And.). 8. nigricans. In Scandinavia omni etiam maxima boreali vulgaris. In Hossia Asiatica non circulum polarem ascendit, terra Samojedorum excepta, ubi a Ruprecht observata. In Kamschatka adest. In America nondum inventa. S. punctata^ Ldbg. Est modificatio Isevis S. nigricantis, in Lapponia tantum observata. S. i)lnjlicifolia, L, In Scandinavia omni boreali vulgaris. In Rossia arctica quidem occurrit, sed rarior. In America S. discolor, Mulil., S. plvylicoides. And., et S. ma- crocarpa in regionibus estra-arcticis earn reprsesentant : cf. And. I. c. pp. 18, 19. ^S*. myrtilloides, L. {S pedicellaris auct. Amer.). ♦S*. hastata, L. In Lapponia vulgaris, E-ossise arctica3 etiam incola. In America S. car- data liujus est forma analoga. S. rhamnifolia. Ad sinus St. Laurentii et in Unalaschka. S. ovallfolia, Trautv. Ibidem. S. Arbuscula. Vera Linnsei species Lapponise arcticse est incola. Ex Eossia Asiatica ut etiam ex America duljia ; iu formis autem nostrae analogis ibi occiu-rit, regiones autem magis meridionales amat. S. glauca, L. In regionibus arcticis omnibus vulgaris. S. desertorum, Eich. S. glaucce forma vel subspecies : cf. And. I. c. p. 23. S. villosa, Don. Salicis glancce est forma vegeta : cf. Andersson, ' Salices boreali- Ame- ricanse,' p. 22. ? S. arctica, Br. Maxima vel polymorplia species regionibus omnibus arcticis pecu- liaris. Vereor ne sub nominibus variis {8. myrsinites, S. retiisa, S. ovallfolia, S. jjokms, &c.) formse variee luijus speciei ab auctoribus sa^pe descriptse sint. And. I. c. p. 23. S. cordifolia, Psh. Partim ad S. alp)estrein. And., partim ad S. suhcordatam. And., pertinet : cf. And. /. c. p. 24. S. Pyrenaica, G. In Lapponia et America {S. cordifolia. Pal. pro parte) : cf. S. alpes- trem. And., I. c. p. 27. Verisimiliter etiam in Siberia arctica occurrit sub speciebus plurimis diu cognitis latens obscura. S. myrsinites, L. Sub plurimis formis in regione arctica ubique. ^S*. reticulata, L. In regionibus arcticis omnibus vulgaris. >S'. vestita. S. reticulata est forma fruticosa et vegeta : cf. And. I. c. p. 29. S. nivalis. S. reticidata; est forma nana : cf. And. /. c. p. 29. S. herbacea, \ . . . . . „ , . > In renjionibus omnibus arcticis vulgaris. S. polans. j The following species, though enumerated as arctic by sundry authors, are not included by Dr. Andersson, who has the following remarks on them : — DR. HOOKER ON THE DISTRIBUTION OF ARCTIC PLANTS. 3Jil S. laurina. Vere silvestris in Suecia non occumt. In Europa central! et austral! mag!s vulgaris, reglones boreales non petit : proles forsan hybrlda. S. Smithiaiia. Ab 111. Uupreclit ad ostium fl. Mesen In terra Samojedorum lecta dicitur. Speclmina nulla ex eo loco vldl, eauique ib! silvestrcm esse dubito, potius ^S*. viminalis forma latifolia. >S'. viminalis et Smit/dana in Songaria endemicse ! S. molUssima, Sm-^/S*. SmitJdana. Vix in regionibus arcticis lecta. aS*. acuminata, Sm. Auctoritate Ruprechti terrse Samojed. a Ledebourio adscripta. Mihi forma S. caprem videtur ! Species a S. caprea et S. Smithiana hybrlda, raro occurrit. S. aurita. In Scandinavia infra circulum polarem desinit. Rossite Asiaticge a Pellman adscripta, sine dubio immerito. In terra Samojed. a Schrank lecta dicitur, attamen maxime dubito ! E flora arctica certe excludenda. (In Amcjica S. hrachystachys, Benth.) S. depressa, L. Nomen in S. vagans mutandum : cf. And. I. c. p. 15, q. v. S. re2)tans (Rich. ? Ruprecht, Flora Samojed. p. 54). Speclmina authentica non vidi; vereor ne forma sit S. arcticce ? vix autem, ut opinatur Ledebour, S. glaucce. S. Fimnarchica, Er. Species a S. mijrtilloidi et S. aurita sine dubio hybrlda, non supra circulum borealem detecta ! S. Uva-ursi, Psh. Nomen excludendum ut e ^S*. rhamnifolia, ovalifolia et sequeute compositum. S. Outleri, Tuck. In civ. Ncav York : cf. And. I. c. pp. 2G, 28. S. canescens, Fr. Species a S. Lappomim cum S. caprea, cinerea et aiirita hybrida, passim in Lapponia inferiore obvia. Ex ceteris regionibus incerta. S. Taymyrensis, 1 Species hasce valde dubias judico. Potius S. phyliccefolicB v. S. Ar- S. Boganidensis. fbitscidcB formse pygmaese trunco subterraneo. Quidquam certi de lis nondum enuntiare non audeo. S. repens, L. Intra fines Lapponise non occurrit. Nee Asiam nee American! borealem inhabitat. Itaque omnino excludenda ! S. versifolia, Wahlb., est species a S. myrtilloidi et Lappomim certissime hybrida, in Lapponia sporadica inventa. E Siberia arctica, prope Obdosck, signo (?) eam aifert Ledebour. In regionibus extra Scandinavias, Livonia excepta, a me nunquam visa. 8. glabra. Scop. Nunquam extra alpes Europse meridionalis observata ! S. grandifolia vera extra alpes Helvetine et Austrise nondum observata ! S. amygdalina. Ad fluv. Tomea (infra circul. polarem) occurrit in Scandinavia. Nee in Rossia reglones subarcticas adpropinquat. S. cinerea. In Lapponia subarctica non occurrit, etiam Asiam arcticam v. subarcticam fugit. In America vix adest. S. retusa, L. In Lapponia non inventa. Species Rossicae vix certse, in S. cuneatam abeuntes. In America fit 8. pMehophylla : And. /. c. p. 27. Larix Sihirica, Led. This, which is synonymous with Ledehoiirii, Endl., is neither American nor Western European. Its western limit is Archangel. L. microcarpa, Lamb., is reduced by Asa Gray to a form oi Americana, and, no doubt, 2 z 2 342 DR. HOOKER ON THE DISTRIBUTION OF ARCTIC PLANTS. rightly. L. jJeiidula, Ait., is the same plant ; and I can find nothing to warrant the separation of L. Dahurica, P. & Turc, from these. JuNiPERTJS nana, Willd. Fries makes of this a variety of communis, L. (as do Watson, Koch, and Ledehour), which is distinguished by the broader, incurved leaves. Every in- termediate grade of leaf may be found uniting them. J. prostrata, Pers. The Arctic American plant is certainly nothing but a stunted form of Virginiana, and is with dij6B.culty distinguished from J. Sahiua, L. MONOCOTYLEDONES. Sagittaria variabilis, Engelm. This is extremely closely allied to and the represen- tative of S. sagitta^folia ; it is chiefly distinguished by the absence of purple on the claw of the petals. The latter plant is Arctic Russian, but not Lapponian. PoTAMOGETON sparganifoUus, Lsest. Pries and Ledehour both retain this as a distinct species, the former noticing its very close affinity with natans. jP. temiissimiis, M. K. Pries makes of this a Lapponian variety of j^«si/^?ts; Ledehour follows him. P. nigrescens, Pr. Under this Pries doubtfully includes lanceolatus, E. Bot. He observes its very close affinity with gramineus, and that P. coloratus, Hoffm., is wholly intermediate. P. lieterophyllus, Schreb. Pries, Koch, and Ledehour all agree in considering this a form of gramineus, L. ToFiELDiA borealis, Wahl., is the same with T. palustris, L. Veeatrum viride, Ait. Asa Gray says of this, " too near the European V. alburn^ The chief character lies in the breadth of the segments of the perianth, which, how- ever, varies considerably in both. V. Lobeliamim, Bernh., is album j3, floribus viridibus, of Chamisso and Ledehour, a very arctic form. Allium Sibiricum, L. Ledehour unites this with Schcenoprasum ; Pries keeps them distinct, regarding the latter as a cultivated plant only. Orchis sambucina, L. The distribution of this is peculiarly wide in some respects for an Orchideous plant, extending diagonally through Europe from Lapland to the confines of N.E. Persia in Karabagh. Peristylus bracteatus, Torr. This and P. Islandicus, Lindl., are so very closely allied to viridis, that they can hardly rank as anything but forms of that plant. This would appear to be the opinion of Liudley and of the ' Plora Boreali- Americana.' Platanthera hyperborea, Lindl. Iceland is the only European locality for this plant. Prom Lindley's observations and those in ' Plora Boreali- Americana,' there is no doubt that dilatata, Lindl., and Koenigii, Lindl., are states of the same. Asa Gray further adds that Huronensis is a synonym of hyperborea, and that dilatata is too near the same. Epipactis media, Pr. This its author makes a distinct species, and does not include the true latifolia in his Lapland list. Koch refers it to rubiginosa, Gaud., distinguished DR. HOOKER ON THE DISTRIBUTION OF ARCTIC PLANTS. 343 from laUfolia by the crispate callosities on the labellum. Watson includes the English media under latifolia, but doubts its being the same as Pries's plant. Bentham unites the English one, which I have seen under cultivation in Dr. Lindley's garden, where it was considered by that learned author to be a slight variety of latifolia, and essentially the same as Fries's media, though differing somewhat. Cypmpedium acaule. Ait., is a synonym of linmile, according to A. Gray. C. Calceolus, L., is introduced into the Arctic European column on Ledebour's authority ; it is not in Fries's List as an Arctic Lapland plant. SiSTBiNCHiUM Bermudianum, L. The distribution of this plant is very extensive, reach- ing from the arctic circle to the Bahamas ; and I strongly suspect that an Andean and Chilian species is identifiable with it. S, anceps, Cav., is referred by A. Gray and others to a variety of Berimidiammi, L. Sparganium liyperhoreum, Lsest., is S. nutans (5. submuticmn, Ilartmann ; it is not in Fries's ' Summa,' where S. natans appears as an Arctic Lapland species. Narthecitjm Americanum, Ker. This, though kept up in all the American Floras, appears to me to be absolutely identical with the European N. ossifragum. LtrzuLA spadicea, DC. All the species, &c., enumerated under this are considered varieties or synonyms, by Ernest Meyer in Ledebour's ' Flora Rossica.' L. glabrata, Hoppe, is distinct according to Fries, and is the spiadicea y of DeCandolle, and v Kunthii of E. Meyer ; it is distinguished by the usually single-flowered pedicels. L. parvijlora, Desv., also kept distinct by Fries, and found throughout Lapland, is spadicea s of E. Meyer. L.meloAiocarpa, Desv., is considered a var. oi piarviflora by Asa Gray. L. Wahlenbergii, Rupr., is the same with spadicea v Kunthii, E. Meyer. For remarks on its distinctness from glabrata and parvijlora see Ruprecht's 'Flor. Samojed.' Trautvetter (Plant. Taim.) observes that some of Ruprecht's Wahlenbergii probably belong to a var. of hyperborea. L. pallescens, "Wahl. Fries regards this as a species not yet fully established. Koch makes of it multijlora, var. i. E. Meyer and Bentham unite it, congesta, and multiflora with campestris. L. multiflora, Ehr., is campestris j3 of DC. and E. Meyer. Fries and Koch keep it distinct and include congesta under it. L. hyperborea, Br. E. Meyer and most authors unite this with arcuata, Hook. Fries distinguishes it specifically by its plane leaves and capitate spikes. L. vernalis, DC, is referred to pilosa, by Koch, E. Meyer, and most authors. JuNCUS biglumis, L., and triglumis, L., are united by Bentham, and are possibly states of one plant, though always so distinguishable that most other authors keep them apart. J. conglomeratus, L., and effusus, L., the two principal forms of J. communis, Ehr., arc both arctic. J. Balticus, Willd. Bentham observes that this is probably a luxuriant form of arcticus. J. uliginosus, Roth. Fries and Koch keep this (the supimis, Mcench.) distinct from articulatus. E. Meyer regards them as synonyms. 344 DR. HOOKER ON THE DISTRIBUTION OF ARCTIC PLANTS. J. lampocarpus, Ehr., is articulatus a and /3 of Linnaeus, fide Koch. E, Meyer unites it. Accordiug to Eries, it is not Scandinavian. J. alpinus, Vill. I have followed E. Meyer, Eries, and Koch, in keeping this distinct, though Kunth unites it with articulatus, and in the ' Elora B or eali- Americana ' it is put as a synonym of uliginosus. J. compresses, Jacq., is the same with bulhosus, L. J. Botlmicus, Wahl., is a synonym of GerarcU, according to Kocli and E. Meyer, as is ccenosus, Bich. fid. Eries and E. Meyer : the latter observes that J. GerarcU always grows in salt marshes, and compressus never. J. atrofuscus, Rupr., is reduced to a synonym of GerarcU by E. Meyer in Ledebour, El. Boss. Cakex. I am indebted to Dr. Boott for revising the list of Carices, which, as it stands, embodies his views as to the specific limits of the Arctic forms and their distribution. SciRPUS Olneyi, A. Gray. This appears to me nothing l)ut a very slight variety of S. triqueter. EmoPHORUM Chamissonis, C.A.M., is regarded by Eries as identical with capitatum. E. Sclieuchzeri, Hoppe {captitatum. Host.), is distinguished {vom. ca,pitatum by the leaves scabrous at the margin, subglobose spike, and stoloniferous roots ; it is not Scandinavian, but found in Arctic Greenland. JE. russeolum, Er. Kept distinct by Eries, from its opake mucronate scales and yellow- brown setas. It is found only in Lapland and West Einland. E. latifolium, L. I find it quite impossible to distinguish gracile, ancjusUfolium, and polystachijitm by any definite or constant characters, and revert to the Linnsean oj)inion that all are forms of one, in which Bentham concurs. GRAMiNEiE. — This list has been twice most carefully revised by Col. Munro, who regards all the names brought under others as undoubted varieties or synonyms. I have in the following notes invariably put foremost the views of Grisebach (Elor. Boss.) and Andersson, as those of the two best and most accurate authors on Northern Grasses with whom I am acquainted. Andersson's in particular is an excellent work in all re- spects ; he keeps up many critical species, but indicates in all cases accurately their inter- mediate positions, and recognizes the trivial nature of their characters. Alopecurus ovatus, Horn., is alpinus j3. horealis of Trinius and Grisebach, in Ledebour (El. Boss.), according to whom the a. Scoticus does not occur in Russia. A. alpinus, L., does not occur in Andersson's ' Graminese Scandinavite.' A. RutJieniciis, Wein. This, a synonym of nigricans, Hornemann, is kept distinct by Eries, Andersson, and Grisebach. Eries says of a \ wietj p)usillus, " fere A. alpiims, Sm.;" again, of nigricans, " A. prcitensi valde affinis." A. aristulatus, Mich. Kept distinct by Asa Gray ; it is the geniculatns, var. /3, of Tor- rey's Elora. A. fulvus, Sm., is kept up by Grisebach and Eries. It is considered the same as aris- tulatus in ' Elor. Bor.-Am.' Andersson says of it, "A. geniculato valde affinis ejusque \a.i'ietaii fhiitautl nimium similis." DR. HOOKER ON THE DISTRIBUTION OF ARCTIC PLANTS. 345 Beckmannia erucceformis, Host. The distribution of tliis plant is most extraordinary : in Europe it is absent in the whole north, west, and central area, but extends from N. Italy to Middle Eussia through Greece. In Asia it is found in Mesopotamia, Persia, Caucasus, Siberia, N.E. Asia, and Japan. In America it inhabits the whole west coast from Oregon to California, the Pi^ocky Mountains, the Saskatchewan, and passes down the Mackenzie to beyond the arctic circle. It is absent in the North and Central United States, but was found at St. Louis by Drummond. It is omitted in ' Mora Boreali- Ame- ricana.' Agrostis alpina, Wahl. The alpina of Scopoli is kept distinct by Grisebach. Fries includes a Lapland plant of Villars's of that name u.nder rupestris, AIL, as Lapponian. Andersson reduces alpina, WahL, to rubra, and observes that it is a very northern plant, hardly known out of Scandinavia. I follow Munro in bringing alpina under rubra, L., though I must confess I do not see why in that case A. tmpestris, AIL, should not follow. The true A. alpina of Scopoli, said to have been once found on the Loffoden Islands, is excluded by Andersson. The true A. rupestris, AIL, according to Blytt and Andersson, is not Scandinavian. A. alba, L. I follow Munro in uniting this {iwlymorplia, Huds.,) with vulgaris. Grisebach keeps it distinct. Deteuxia and Calamagrostis. I have implicitly followed Munro in these genera, being quite unable to arrive at any satisfactory conclusion regarding them. D. Canadensis, P. B. Durand is the only authority for this being a South Greenland plant ; but as he does not mention varia, strigosa, or lanceolata (all reputed Greenland plants), I hesitate to admit it. D. neglecta, Bupr. and Ehr., is a synonym of s/ricte, according to Grisebach, Eries, and Andersson. It is kept distinct from Lapponica by all. D. chalybcea, Eries, is kept distinct by its author and Andersson, who observes of it, " C. Lapponicce simillima, sed non parum diversa." It is the Lapponica (3. chalybce-a of Lgestad. D. pwpurascens, Br. Asa Gray refers this doubtfully to sylvatica, DC. (Mantial Bot. N. U. S.) ; Torrey more confidently (see ' Elor. Bor.-Am.'). D. Grcenlanclica, E. M. This is one of Kunth's unknown species (Agrost. 239) re- ferred here by Col. Munro. It is not referred to in Lange's Greenland list. It is also a Labrador plant, according to E. Meyer. B. montana. Host., is the Arundo varia of Wahlenberg. Eries gives the Baltic Islands as its only Scandinavian habitat. D. Hartmanniana, Er., is kept distinct by Andersson and Grisebach ; but Eries con- - siders it insufficiently established. Andersson says of it, " pulcherrima species, inter Sallerianam et sylvaticam fere media ;" he places it next strigosa. I follow Munro in bringing it under varia. D. aleutica. Bong. Kept up by Grisebach ; but CoL Munro considers it identical with strigosa. D. Langsdorffii, Tr. Kept up by Grisebach, but considered by Munro the same with purpurea. 346 DR. HOOKER ON THE DISTRIBUTION OF ARCTIC PLANTS. D. elata, Blytt, is the form j3. major of phragniitoides, according to Andersson, wliich approaches lanceolata. Calamagrostis litorea, DC, is a synonym of laxa, Host., according to Grisebach. According to Eries, it is an insufficiently defined species. Andersson observes its ex- tremely close affinity with Epigejos. C. Halleriana, Gaud. A distinct species, according to Grisebach, Eries, and Andersson, who, however, says (under lanceolata, var, cinerascens), " CHalleriunce proxima," and, again, " inter G. lanceolatam et ph^agmitoidem omnino media videtur." C. phragmitoides, Hartm. Also kept up by Grisebach, Eries, and by Andersson, who observes of one variety, " C. HalleriancB analoga;" of another, "ad C. lanceolatam quodammodo vergens;" and of a third, "a C. lanceolata difficile dignoscitur." HiEROCHLOE racemosa, Tr., is a synonym oi pi^f'UciJlora, according to Grisebach. Deschampsia brevifoUa, Br. This, the Aira arctica, Tr., who elsewhere regards it as a var. of ccespitosa, is kept up by Grisebach. Eries reduces it. Andersson remarks of brevifolia, Hn., that it is very remarkable from resembling a hybrid between ccBspitosa and flexuosa. Tbisetum agrostideum, Tr., is kept distinct from subspicatum by Andersson, Grisebach, and Eries. Andersson remarks that it is so entirely intermediate hetween flavescens and subspicatum, that it is extremely difficult to say to which it is best referred. T. Sibiricum, Pk^upr. {Aira MnprecMii, Griseb.), is distinguished by Grisebach by its much larger scabrous-bearded florets. Andersson observes how very near it is to agrostideum. Phippsia monandra, H. & A., is a synonym of algida, according to Grisebach. It is al(/kla (3 of Kunth. It is kept distinct in ' Elor. Bor.-Am.' Catabrosa aquatica, Br. Durand gives this as a native of Greenland, lat. 65° N. CoLPODiUM paucijlorum, Hook. Only one habitat is known for this ; it may be a reduced or starved state of latifolitim ; but much better specimens are required to deter- mine the point. C. arundhiace^im., Hook., is latifoUum (3, Trinius and Grisebach, and is reduced to a synonym by Andersson. C. fulvum; Tr. Andersson doubts whether this and penduUnum are really distinct species ; and judging from the fine specimen in Herb. Hook., I am disposed to think, with reason. Foa dejiexa, Eupr., remotljlora, Rupr., and similis, Bupr., all of Samoyedland, are all reduced by Grisebach to Colp). penduUnum — the last as a variety, the two first as synonyms. P. scleroclada, Pi/upr., latijlora, Bupr., and poecilantlia, Bupr., all of Samoyedland, are reduced by Grisebach to synonyms of Colp. fulvum. DUPONTIA Fisheri, Br., was named after Mr. Eisher, one of Parry's officers, and should not be written Fischeri. D. pisilosantha, Tr., of which Foa pelUgera, Bupr., is a synonym, according to Grise- bach, is kept by that author as a distinct species. DR. HOOKER ON THE DISTRIBUTION OF ARCTIC PLANTS. 347 Glyceria arctica, Hk,, is certainly only a starved state of Jluituns. Durand. gives this as a native of Greenland, lat. 68° N, PoA airoides, Nutt., is identical with Atrojns cUstans, Griseb. P. angtistata, Br. (Atrojns, Griseb.), is kept distinct by Grisebach with P. Nutkaensis as a synonym ; it is a very arctic form. P. angustifolia, L., is pratensis y of Grisebach ; it is rare in the arctic regions. Anders- son refers it, with a query, to serot'ma. P. trivialis, L., is kept distinct by Grisebach, Pries, Andersson, and almost all authors ; it is also arctic. Pries doubts the identity of the Linnsean plant. P. nemorcdis, L. I have followed Mum-o in bringing together the following names, which, whether species or no, are not to be discriminated by arctic stunted specimens ; he considers them all forms of one : — P. ccesia. Smith, is kept up by Grisebach and Andersson, with aspera, Gaud., as a synonym; Pries also keeps it distinct; Andersson indicates its extremely close affinity with nenioralis. P. Grcenlandica, Steud., is nothing but serotina, Ehr., which is kept up by Grisebach, Pries, Asa Gray, and Andersson. P. bryopJiila, Tr. Called arctic by Grisebach, who has never seen the plant ; but I do not find the locality (Pret. Senjawin) in any map. Munro reduces it to nemoralis. P. Vahliana, Liebm., Plor. Dan. 2401, is but a slight variety of nemoralis. P. laxa, H^nke. Grisebach, Andersson, and Pries all keep this distinct ivom flexuosa. In ' Plor. Bor.-Am.' they are treated as synonyms. Por remarks on its affinities, see Andersson (Gram. Scand.). P. Cenisia, AIL Grisebach keeps this distinct, but refers Ceuisia of Pries's Ilerb. Norm, to arctica, Br. {Jlexuosa, Wahl.). Pries vefevs flexuosa, "VYalil., to Cenisia. Anders- son considers Cenisia hardly distinct from arctica, Br. P. arctica, Br. Kept distinct by Grisebach, who Tas>k&'& jlexuosa, Wahl., a synonym of it. It b,Jlexuosa /3 of Trinius, and laxa, Br. in Ross's Voyage. P. abbreviata, Br. Unknown to Grisebach, and reduced by Munro. Pries and Anders- son reduce abbreviata, Blytt, to Cenisia, var. depavperata. Pestuca Bicliardsoni, Hook. Durand, on Torrey's authority, gives what Torrey doubtfully considers a variety of this plant as a native of Greenland. I know of no species to which Bicliardsoni is closely allied, except the Bocky Mountain var. F. scabrella, of which it may be a hirsute form. P. ovina, L. I follow Munro and Bentham in reducing r^ibra and duritiscula, and the former authority in bringing Kceleria hirsuta and F. brevifoUa here. The latter is a marked smaU form, but passes gradually into rubra. Andersson keeps up ovina, duritis- cula, and rubra, bringing arenaria under the last, and sabtilosa under the first. Grise- bach keeps up ovina, L., with duriuscula as var. S ; also rubra, L., with arenaria as var. |3, and Kceleria hirsuta (of which I know nothing) as distinct from both ; but Munro reduces all to ovina, L, Pries unites arenaria and rttbra, keeping durmscida and ovina distinct. VOL. XXTII. 3 A 348 DR. HOOKER ON THE DISTRIBUTION OF ARCTIC PLANTS. F. brevifolia, Br. Brown says of tliis (Flora of Melville Island), "forsan nimis affinis F. ovincB." Bkomus inermis, Leys., and imrgmis, L., are both kept distinct from ciliatus, L., by Grisebacli. I am indebted to Munro for their identification, together with that oi pictus, mihi, which is undoubtedly ciliahis, L. Asa Gray refers purgans to ciliatus. Durand mentions this plant as a native of Greenland, lat. G5°, under the name of B. Kalmii ? according to Dr. Torrey. Teiticum oiolaceum, Horn. Fries keeps this distinct ; but Blytt suspects it to l)e a variety of T. caniuuin. Andersson observes that it is an alpine and arctic plant, not uncommon in Greenland, altogether intermediate between cmiimmi and repiens, and con- cludes "nonne attamen modificatio alpina?" Since the printing of the foregoing paper, I have received, through the kindness of Dr. Asa Gray, particulars of the discovery of Callnna vulgaris near Tewksbury, Massa- chusetts, by Mr. Jackson Dawson, occurring rather abundantly over about half an acre of boggy ground, amongst various strictly American plants. This tends to confirm the statement of De la Pylaie, that Calluna is an inhabitant of Newfoundland. J. D. Hooker. Keii\ Oct. 11, 1801. Erratum. At (). 31 7, line 8, ufter " regions" ndd "of Eiu'opc and Asia." [ 349 ] XVIII. Further Observations on Entozoa, loith JExperiments. By T. Spencer Cobbold, M.D., F.L.S.* Read December 20th, 1860. Although, since my last communication, I have had fewer opportunities of examining Entozoa, I think it will be admitted that the new or otherwise important forms which have come under my notice, are of sufficient interest to be placed on record, in continu- ation of the series of papers previously submitted to the Society. In the present instance, the number of different animals specially examined with refer- ence to the presence or absence of internal parasites, amounts in aU to forty-three, in- cluding seven fish, two reptUes, seventeen birds and nineteen mammals. Of these only three fish, one reptile, seven bu'ds and the same number of mammals, were found infested. In so far as these numbers refer to species formerly living in the Zoological Society's menagerie, the proportion is small, and tends to confirm both the statement and explana- tion which I have before hazarded in respect of the comparative freedom from internal trematode parasites, found to obtain in foreign animals subjected to a condition of captivity, 1. DiSTOMA CONJUNCTUM (mihi). Distoma conjunctum, Cobbold, Linu. Soc. Proceed, vol. v., Zool. Div. p. 8. Bemarlcs. — As already briefly indicated in my synopsis of the DistomidEe, published in the Society's Proceedings {loc. cit.), this species of fluke infests the liver of the American Red Eox {Ccmisf ulcus). The animal in question died at the Zoological Society's Gardens, Eegent's Park, in December, 1858, and was dissected on the 24.th of that month, the carcase being still quite fresh. The biliary ducts were found thickened and enlarged in several places, forming here and there small cysts of variable size. In these cavities — partly occu- pied by purulent matter — the distomes were lodged ; and when placed in water they showed signs of life. Their extreme transparency at once displayed, under the microscope, the rela- tive position and degree of development of the various organs, as shown by the accompany- ing figure (fig. 1, Plate XXXIIL), in which I may particularly remark the great breadth of the digestive ceeca, the well-defined excretory canals passing transversely inwards from the botryoidal albumen-forming glands (Dotterstock), and also the large yelk-forming organ or ovary (Eikeimstock), placed in the centre of the body, immediately below the uterine folds. A still more noticeable feature, however, is seen in the unusual length and breadth of the contractile vesicle belonging to the excretory system. It occupies fully one-third of the longitudinal diameter of the body, and bifurcates at the summit, the course of each division being traceable almost as far forward as the bulb of the oesophagus ; below, it is suddenly constricted, and terminates, as usual, by a small central aperture. The sac itself * An abstract of this paper has been published in the Society's Proceedings, Zool. Div. vol. v. p. 255. 3 a2 350 DR. T. S. COBBOLD, FURTHER OBSERVATIONS ON ENTOZOA. is comparatively narrow and tortuous at the upper half, where it passes between, and apparently contributes to separate, the two well-developed and irregularly oval-shaped testes. The inferior half is much broader, and is occupied by numerous glittering cor- puscles, which are especially numerous at the lower end. The character of these particles has already been accurately described by Wedl, Wagner, and others ; but I did not observe any of them connected together in the form of cellular masses, such as Claparede has described as occurring in the excretory ducts of Solostomata. In connexion with the reproductive system, I may notice the circumstance of a pair of these flukes being found sexually united, as represented in fig. 2 ; an observation so rare, that it has, I believe, led some to infer that the hermaphroditic flukes were capable of self-impregnation. Even Bilharz's discovery of a trematode specially furnished with a gynsecophoric canal, afforded no direct proof of a true sexual function in the androgynous Distomata, because in his aberrant form — as obtains also in a few others — the sexes are separate. Many thousands of flukes must have now passed under my examination, and yet this is the first instance in which I have observed a true sexual union ; moreover, I am not acquainted with any similar observation on record, in so far as it affects the truly hermaphroditic species. In the instance now mentioned, the opposed suckers were so firmly adherent that it was found impossible to detach the animals without In'eaking them up piecemeal. Lastly, I may notice that the ova (/', fig. 1) contained in the u.terine folds have a longitudinal diameter of only y^o of an inch. 2. Pextastoma denticulattjm, Eudolphi. Pentastoma denticulatum, Rudolphi, Bremser, Wiegmann, Dujardin, Diesing and others. P. tctnioides, in the young state, according to Gurlt, Kiichenmeister, and especiallj' Leuckart. See Mem. in Henle and PfeifFer's Zeitsch. f. rat. Medecin, Bd. iv. p. 78, 1858. MemarJcs. — Although it is clear from the investigations of Schuliert and Van Beneden respecting the condition and development of the embryo of 'Pentastoma, that this genus is closely allied to the Acarida?, its habits are so intimately associated with those of Entozoa proper, that there can be no impropriety in offering a short notice in this place. On the 10th of Ecbruary, 1859, I obtained numerous examples from the viscera of a Bubale {Antilope Buhalis, Pallas), which had died the day previous at the Zoological Society's Gardens. The greater part of these worms occupied the surface of the lungs and intestines ; a considerable number, however, being enclosed in cysts beneath the pulmonary pleura. Eifteen or more having l^een taken away for the purposes of microscopic exami- nation and experiment, I had ample opportunity of Avatching the movements of the cephalic claws whilst the animals were still living. I may here remark, that these claws do not in any way resemble those of the Tape-worm family, but, in strict accordance with other external features, show that the Pentastomes belong to an entirely different type. Placed under the half-inch objective, with the ventral surface uppermost, the transparency of the body permits a full view of the apparatus of hooks, as displayed in the accompanying drawing (fig. 3). It will be here noticed that the points of the claws are directed towards the observer, and not towards the longitudinal axis of the body as Kiichenmeister' s figure (' Parasiten,' tab. viii. fig. 11) would lead one to suppose ; moreover, the claws ai'e placed DR. T. S. COBBOLD, FURTHER OBSERVATIONS ON ENTOZOA. 351 obliquely in reference to the central line, the angle of divergence being about 25°. During the eversion and retractation of the claws, the vacant spaces directly above them dilate and contract, so as greatly to facilitate the prehensile action of the hooks. These de- pressions are obviously inversions of the ventral integument, forming socket-like pouches for the implantation of the hook-capsules. If one of the claws be isolated and viewed laterally, it will be seen to consist of two parts — namely, a strongly curved hook («, fig. 4), and a capsule (b). The upper part of the latter forms a hood (d), for the lodgement, support, and protection of the hook ; the loAver tubular portion serving to fix the apparatus in the body, and at the same time to regulate the movements of the claw through the intervention of a muscular pulley (g), which enters tbe tube at the lower end. Acting antagonistically to this, we find an extensor muscle {/) inserted into the anterior border of the base of the hook itself. The summit or projecting point of the hood is flattened out, so as to form a three-cornered process (e), which covers and protects the sharp ex- tremity of the hook when the latter is retracted. In regard to other external appearances, my observations for the most part correspond with those already on record ; but with reference to the rows of minute integumentary spines, it will be seen from my figure (3), that the uppermost row is placed somewhat more apart from the second — at least, at the sides ; the whole forming a complete and graceful tracing, as it passes downwards on either side between the inferior and superior pair of cephalic claws. Further, I do not find the spines themselves to be simply conical, as hitherto represented, because their uniformity of outline is interrupted at the lower third, as shown at the left margin of figure 5 ; this appearance, indeed, can only be seen by a lateral view. The so-called stigmata (Athemlocher of Nordmann) display a double outline, but no vessels could be traced in connexion with them. With respect to the internal structure of this parasite, very little can 1)e made out — a cu'cumstance materially favouring the notion that we have to deal with an imperfectly developed animal. Throughout the body there exists a central clear space, and this is well defined by dark lateral masses, which narrow out on either side of the head, in the form of a horse-shoe-shaped band enclosing the mouth and the four chitinous claw-masses. These dark patches at first sight look like albumenogeue organs, but a closer examina- tion shows that they are entirely made up of large parenchymatous cellules, unconnected by any special fibres or tubes. In an isolated and highly magnified group of them, the granular contents and central nuclei at once became visible, as represented in fig. 6, where some have become flattened at the sides from accidental pressure *. Before finally quitting the consideration of this parasite, I may advert to the recently advanced opinion of Gurlt, Kllchenmeister and others, respecting its assumed larval con- dition. After the perusal of Leuckart's original memoir t, kindly placed in my hands by Mr. Busk, I naturally felt desirous of repeating that distinguished zoologist's experiments. Accordingly (as I have already stated in the Quart. Journ. Micr. Sc), on the 11th of Fe- bruary, I proceeded to infest two dogs. Into the nostrils of the first hound, stated to be * Leuckart, in his recent work, ' Bau unci Entwickelungsgescliiclite tier Peutastonien,' says they are glandular, and finds ducts in connexion with them. t Heule and Pfeiffer's Zeitsch. he. aY.— See also my translation in the ' Micr. Journal,' vol. vii. p. 182. 352 DR. T. S. COBBOLD, FURTHER OBSERVATIONS ON ENTOZOA. only ten months old, I introduced nine living Pentastomes, and in tlie second — an old dog — five examples. All the worms quickly disappeared, and I therefore fully expected a satisfactory result. I was doomed to disappointment. On the 4th of March the young animal was destroyed, and I made a most careful dissection of the head, without, how- ever, finding a single Pentastome in any of the nasal, frontal or facial cavities. I regret to add, that the old dog was subsequently killed and thrown away during my absence, and I have not since had an opportunity of repeating the experiment. Prof. Leuckart, having seen the memorandum above-mentioned, suggests that the parasites may have been dislodged by the dog's sneezing. 3. Trichocephalus affinis, Eudolphi. 7". c^BW, Rudolphi ; Miram ; Lamarck; Gurlt ; Mayer; Dujardin ; Diesing, Helminth, vol. ii. p. 296 ; Kiichenmeister, Parasiten, p. 2/5. T. Cameli, Rudolphi. T. ovis, Abildgaard. T. Giraff(B, Clot-Bey (non descript.), in Bulletin Scient. de TAcad. Imp. de St. Petersb. 1839, torn. vi. p. 94 ; also in Isis, 1839, p. 663 ; Diesing, Syst. Helminth, vol. ii. p. 294 (species inquirenda?) . T. gracilis, Cobbold, Proceed. Zool. Soc. for 1860, p. 103 (non descript.). Remarks. — In a paper, entitled " Contributions to the Anatomy of the Giraffe," which I had the honour of communicating to the Zoological Society on the 14th of February last, the circumstance of my having detected Trichocepliuli in the caecum and colon of that ruminant is already mentioned. At the time referred to, I purposely abstained from giving any particular account of the worm, but, believing it to be a distinct form, I recognized it provisionally under the title of T. gracilis. Subsequently, finding the latter title to be applicable to a Triclwcephahis discovered by Olfers and Natterer in two species of Agouti, I abandoned it ; and I have, moreover, since satisfied myself by a very careful investigation, that the species in question is identical with the T. affinis of Eu- dolphi. Under the synonym of Trichocephalus Giraffes, Diesing has grouped this worm among his doubtful forms, its presence having been previously noticed by Clot-Bey in the small intestines of a giraffe dissected at Cairo in the year 1839. Viewed with the naked eye (fig. 7), one can scarcely detect any difference between this form and the well-known T. clispar of the human subject— an observation which applies to some other species of the same genus. In my examples the females measure from l|-2 inches in length, over all ; the males reaching from 2-2^ inches. With the pocket lens, the surface of the worm appears smooth throughout, but when highly magnified, peculiar markings are seen on the anterior thin portion, which probaljly also extend over the body proper. The so-called neck presents a tolerably uniform thickness along its entire course ; it is so narrow as to measure only from rir-ii o^h of an inch transversely, whilst the finely-pointed head itself, immediately below the mouth, has a diameter less than TOToth of an inch. In the fresh state the head appeared to be lobed, or rather, I may say, fiu-nished with two alseform lobed appendages, as represented in fig, 8 ; but in preserved specimens these appearances either partially or entirely disappear, leaving one in doubt as to their true, nature. Kiichenmeister has noticed the evanescence of appa- DR. T. S. COBBOLD, FURTHER OBSERVATIONS ON ENTOZOA. 353 rently similar structures surrounding the mouth of T. dispar, and therefore supposes that the lobes in question are due to the presence of a peculiar organ, capable of eversion and inversion, and not merely the result of accidental sarcode globules. Be that as it may, I have been surprised to notice how completely other well-marked external and internal characters alter or disappear from shrinking, &c., catised by immersion in spirit. This observation especially applies to a very peculiar longitudinal band, which com- mences a little below the head, and can be traced on one side of the neck the whole way down to the beginning of the so-called body ; it is remarkably distinct in fresh specimens, but barely discernible in those preserved. This band was first discovered by Dujardin, who states it to consist of prominent and pointed papillse. Wedl has also described it as consisting of little warts and spines ; whilst Kllchenmeister goes so far as to compare these little prominences to the booklets present on the male copulatory organs. Accord- ing, however, to my own examinations, this band is made tip of projecting, bluntly pointed, polygonal, epidermal cells, which in certain adjustments of the focus refract transmitted light so strongly, that the band of them looks as if it consisted of a regularly arranged series of pigment spots (fig. 9 a); at other times the centre of each cell becomes clear (a'), and the irregularly polygonal character of each individual cell is rendered more apparent. On one side of the longitudinal band, Dujardin also figures and describes a series of minute superficial papillae, which he associates with a festooned border of the band. I have not observed these prominences ; and the festooned markings are clearly due to the subjacent convolutions of the oesophagus (i), which are singularly uniform in size and disposition. In the fresh state, the dermal rings (c c) are beautifully distinct ; they are said to extend all round the filamentary neck, but I found the transverse lines, indicating their limits, to cease at a little distance apart from either side of the longitu- dinal baud. Midway between the latter and the serrated border of the neck there exists internally a double row of oval corpuscles {d d) ; but, as no vessels or fibres were observed in connexion with them, and their contents were not visible, I will hazard no opinion respecting their nature. Turning now to the reproductive organs, the first thing that strikes one has reference to the unusual length of the penis and its membranous sheath — a character believed to be peculiar to this species. In this respect, at least, it departs very materially from what is observable in T. dkpar, where the sheath forms externally a funnel-shaped tube, which may possibly be a distinct organ. At all events, if specific differences have any existence — and unfortunately for zoological science, recent hypotheses tend to limit, if not virtu- ally to deny, their value altogether — no one can any longer reasonably entertain the notion that T. affinis and T. dispar are one and the same species. The organ to which I have just referred as present in T. affinis, is itself included in a sheath-like muscular mass, which I suppose to be concerned in the evolution of the penis. The free end of this intromittent muscle is shown in fig. 11 a. I never saw this muscular mass everted, but the anal opening {b, fig. 11) is sufficiently capacious to give it free passage, if neces- sary. The everted part of the sheath of the penis (fig. 10, and c, fig. 11) measm'es about the 1^9-th of an inch in length ; it is perfectly transparent, not always uniform in breadth, but covered throughout its entire extent with minute, conical, sharply pointed spines, 354 DR. T. S. COBBOLD, FURTHER OBSERVATIONS ON ENTOZOA. whose apices are directecl backwards towards the body of the animal. Tbe occasional absence of uniformity in the diameter of the sheath seems to me to be a point of some importance ; for, had not my examinations extended over a considerable number of examples, I might have been led to the belief that I had to deal with several distinct forms of Trichocephalus. At first, indeed, this conclusion seemed inevitable, but finding intermediate conditions between perfect uniformity and the presence of a large flask- shaped distension near the free extremity, I can only suppose the variations to be due to the degree of protrusion at which the organ has arrived. In the accompanying woodcuts I have outlined a few of these appearances (figs. 1, 2, 3), which are intended for compa- rison with the simple condition of the sheath shown in fig. 10, Plate XXXIII. The penis itself is about three times longer than the everted sheath, the exposed portion measiTring the Tjth of an inch ; it is cylindrical, ciu-ved towards the tip, and coiled within the spirally twisted tail at its superior two-thirds ; the free extremity is scimitar-shaped, and rather sharply pointed (fig. 12). I did not observe any markings on its surface, but internally there were lines indicating the presence of a groove or internal tube, such as has been described as occurring in T. dispa?'. In regard to the organs of generation in the female, Klichenmeister states that there are no external appendages in Trichocephalus comparable to those known to exist in the allied Trichosomata. So far, however, from this being the case, there is, in the present species at least, a remarkably prominent, and more or less hour-glass-shaped sheath ; this projecting vulva, if it may be so termed, being obliquely truncated at the free end, where it is also hollowed out, or rather inverted to give origin to the centrally enclosed vagina, whose orifice is somewhat constricted (fig. 13). The surface of this appendage is supplied with small spines, precisely similar to those described in connexion with the sheath of the male intromittent organ, the spines being likewise retroverted. This observation is confirmed by the statements of Mayer, who has described their occurrence at the vaginal orifice of T. disjpar. Dr. Joseph Eberth, of Wiirzburg*, however, in a recent number of Siebold and Kolliker's ' Zeitschrift,' rather incautiously denies the assertions of Mayer f. Dr. Eberth, having discovered a series of long conical processes within the vagina of T. dispar, totally unlike those described by Mayer and myself, has inferred that the structures in question are the same as those seen by us. I translate and quote his own words : — " According to Mayer, these are similar in form and size to the spines on the appendix of the penis. By means of these backwardly-directed points, the spicule when introduced into the vagina is said to be retained during the copulatory act. These statements," he adds, " are not correct : the villi of the vagina are larger than the spines of the male, and their points are, on the contrary, directed forwards." The latter part of this observation is, doubtless, quite true ; for it is evident that our little spines are entirely different from the curious villi discovered by Dr. Eberth. The ova, previous to impregnation, as in other Nematodes, are, at a certain stage, flat and irregularly triangular in outline ; the thin limiting membrane by which they are surrounded enclosing finely granular contents (fig. 14). In the perfectly developed e^^, * " Die Generationsorgane von Trichocephalus dispar," Zeitschrift for April 1860, p. 38-1. t Zeitschrift fiir wissen. Zool., Band 9, s. 367. DR. T. S. COBBOLD, FURTHER OBSERVATIONS ON ENTOZOA. 555 the external cliitinous capsule presents the same characters as in T. dispar; at either pole of the eg^, where the shell terminates abruptly, an inner transparent membrane projects, in the form of a mammillary process (fig. 15). None of the eggs I examined appeared to contain fully developed embryos, but the interior yolk-mass exhibited evi- dences of segmentation, many of them displaying two nuclei. These ova have a longi- tudinal diameter of -sioth to aio^^i of ^'^ i^ch. Fig. 2. Ym. 1. Fie. 3. Figs. 1, 2, 3. a, Epidermis; b, cutis ;c, cut ceecal end of the testis; d, seminal duct; e, intestine; f, intro- mittent muscle surrounding g, the sheath of the penis ; g', infundibuliform portion ; g", exserted portion, armed with minute retroverted spines ; pe7idix. — In this place, as on a former occasion, I have thought it useful to note down, more or less briefly, such other interesting forms of Entozoa as have incidentally come under my observation ; at least, I have here particularized a selection of them. Some of these parasites will elsewhere demand a more extended notice. Erom the Perch I have obtained two forms of Echinorhynchus, accompanied by a solitary example of the well-known viviparous Nematode, Cucullmms elegans. Being an adult female, it contained germs, eggs, and young in every possible stage of development. From a Pike I procured three very juvenile examples of THcuspidaria nodulosa. In this para- site, and in immediate connexion with the sub-epidermal calcareous corpuscles, I have discovered peculiar vessels with flask-shaped ends, the latter surrounding the corpuscles. I have akeady described these structures in the ' Quart. Journ. Micr. Sc.*' ; but I may here be permitted to add that, as they appear to be quite independent of the ordinary excretory system of vessels, I have since, on further consideration, arrived at the conclu- sion that they are rather to be regarded as special organs for the secretion and formation of the corpuscles than as a vicarious development of the excretory vascular system. Erom the intestines of a large snapping Turtle {Chehjdra serpentina) I have procured * Vol. vii. in Memoranda, p. 115 and p. 202. DR. T. S. COBBOLD, FURTHER OBSERVATIONS ON ENTOZOA. 357 some unusually slender forms of EcMnorhynchi, which were accompanied by several unimportant-looking Nematodes. Amongst birds I may mention the occurrence of numerous tape-worm proglottides infesting a black-throated species of AmacUna ; and in the intestinal caeca of the Cheer Pheasant [Fhasicmus WalUcliii), as also in those of the Black-backed Kaleege {Etiplo- comus melanotis), multitudes of the oft-recurring Ascaris vesiculcms. The last-named bird was bred in the Zoological Society's Gardens. Here also may be noticed the existence of many examples of no less than three entu'ely different species of Strongylus infesting respectively the proventriculus, intestines, and caeca of an Ashy-headed Goose {Bernicla poliocephala) . One of these parasitic forms appears to be new ; but, of the others, the second is clearly referable to Strongylus tubifex, and the third is probably S. nodularis. The intestines of a Tinamou [Tinamus ?) likewise yielded several specimens of a new Strongylus. From a Starling (Sturnus vulgaris) shot by me on the 10th of January last, I obtained a solitary example of the Tcenia farciminalis, the lowermost joints of which, though still firmly adherent to the strobila, were amply furnished with mature eggs containing sis- hooked embryos, having a diameter of about axo'th of an inch. I have figured one of these in the shell (fig. 17), to contrast with those found in the cestode infesting the chicken. The last-named parasite I suppose to have been Tcenia infundihuliformis. Turning to Mammalia, I have to notice the occurrence of multitudes of Pentastomes in the abdominal cavity and viscera of a Cape Guevi {Cephalophus pygmcBus), these being in all respects similar to those I have described from the Bubale Antelope. In a Spring-bok {Gazella Eiicliore) I also encountered two characteristic specimens of the so-called Cysticercus tenuicolUs ; and in one of the bronchial tubes of the American Red Fox {Vulpes fulvus), from which the new Distome, D. conjiinctum (mihi), was procured, there were found two specimens of Trichosoma cerophilum. The eggs of the last-named worm are scarcely distinguishable in outline from those of Tricliocephalus affinis, their longitudinal diameter being only a trifle less — that is to say, about the sloth of an inch (fig. 18). From the common Bat I have secured an example of Tcenia p)usilla, measuring upwards of 20 inches in length ; and lastly, I may mention some remarkably large Cysticerci infesting the abdominal cavity of the Ethiopian Wart Hog {Pliacoclioerns cethiopicus), and of the Bed River Hog {Totamochcerus penicillatus). These gigantic Scolices are apparently referable to two distinct forms of tape-worm ; but, as their separate description and comparison will involve many other points of interest, I do not off'er any further particulars respecting them in the present communication*. * In the recently published ' Proceedings of the Zoological Society' for 1861, Part I., I have since given a minute description of these Cysticerci (p. 93, plate xvii.) ; and in a separate paper published in the same Part of the ' Proceedings ' the new forms of Entozoa above mentioned are described under the following titles : — Echinorhynchus inflexus from the Turtle ; Cestoidemn Amadinr»u!n V, ■ / PALLIAL SINUSES OF THE BRACHTOPODA. 375 cation just noticed is cut oif, and an intestine formed, with or without an anal aper- ture. Let us now suppose tlie perivisceral cavity of the Folyzoa to be homologous with the common cavity of the Coelenterata, but giving off no canals, and the oscillation or motion of the contained blood (?) depending upon the activity of the animal's movements as well as upon ciliary action, the scheme of circulation exhibits an almost retrograde simplicity when compared with tliat of a covered-eyed Medusa. Nevertheless, having no dit'eet communication with the stomach, its superiority may be asserted as a system so far closed. Now, as this appears to be homologous with the perivisceral cavity of the Bra- chiopoda, the ramifications of the palHal sinuses of the latter, with their ciliated lining and contained corpuscles, present the highest form of this pseudo-circulatory s)'stem. The existence of a true vascular system in the Folijpi is intimated by M. Milne-Ed- wards ; but it is not improbable that the vascular ramifications observed by him, and which undoubtedly exist, are in reality derived, or, as it were, dismembered from the common cavity. May not the same doctrine be applied to the BracMopoda ? If not, the true vascular system, as described by Mr. Hancock, is altogether distinct, both in its nature and origin, from what has hitherto been regarded as the rudiment of the circulatory system in Coelenterata. In the Tiinicata, however, a single and simple heart superseding the five vesicles of the Braclilopoda offers some little approach to the more perfect organs of the Mollusca proper, though there can as yet be no distinction between arteries and veins, as the current of the blood is reversible in the same channels. Finally, it may be affirmed, that a heart with a receptive and a propulsive cavity, furnished with a valvular apparatus to deter- mine an irreversible path for the circulation, as well as the distinctive offices of artery and vein, distinguishes the true Mollusca from the Tiinicata, Brachiopoda and Folyzoa, con- stituting the Molluscoida of M. Milne-Edwards. EXPLANATION OE PLATE XXXV. Fig. 4. Portion of the pallial sinus-system of Lingula, showing the course of the ciliary currents. a, longitudinal septa dividing the ciliary currents. Fig. 5. Violet-tinted corpuscles which circulate in the pallial sinuses of Lingula. The little pit or depression, observable on one side of these corpuscles, is probably due to exosmosis ; bat in this particular case it was present in all. intelligence, at the slightesit touch, are represented as possessing the rigidity of the teeth of a comb, or, at least, of bristles. A very far more correct idea of them might be given were the artist to copy the tentacula of Vesicularia, or any well-known Polijxoon — a piece of dishonesty which, although its perpetration cannot be seriously recommended, might challenge detection in this particular case. [ 377 ] XXII. On the Neroe proceeding to the Vesicles at the Base of the Halteres, and on the Sitb-costal Nervure in the TFings of Insects. By J. Braxton Hicks, M.D. Bond., F.B.S. Read June 20th, 1861. It is unnecessary to enter into any description of tlie vesicles of the halteres and wings, as I have already fully noticed them in former j)ages of the Transactions and Journal. I showed first, also, of any observer, that to them a very large nerve proceeds, next in size to the optic, and that the divisions of this nerve pass to the base of the vesicles. The additional information, which it is my desire now to lay before the Society, is, that a bipolar ganglionic cell is to be found just beneath these organs. If, instead of viewing the parts intact, the base of the haltere be broken up, it will be found that, shortly before arriving at the position of these vesicles, the nerve enlarges in the following manner : — Each fibril is developed into an elongated ganglionic cell, gene- rally with a large nucleus {a, figs. 1, 2, 3). The distal end becomes contracted again lu for a short space, but finally and gradually dilates, until it comes in contact with the inner surface of the vesicle {b, figs. 1, 2, 3). In this terminal expansion I have sometimes observed a nucleus (c, fig. 2) ; sometimes, however, the contents have been granular ; whether there be a smaU cell at the very extremity, I am unable to say with any certainty, but comparing it with its homologue in the subcostal nervure, I am inclined to think the existence of such a structure doubtful. 378 DR. HICKS ON THE KERVE AT THE BASE OF THE HALTERES, AND ON This can be well observed in Bhingia rostrata and Eristalis tenax. (See figure.) The nerve in the subcostal nervure is disposed in a somewhat similar manner. But in many instances I have found the ganglionic enlargement to possess more than one nucleus ; two or three apparently coexist in the Coleoptera (see a, figs. 4, 5, 6). The distal end of this enlargement tapers off to a rather fine extremity, so that it may pass up through the canals which pierce the integumental layers. The mode of termination in these nerves points out a homology to the termination of the optic nerve, as seen in its development in insects, and is, indeed, probably the method by which all nerves of sensation end, as has been shown by M. N. Jacubovitsch in the Comptes Hendiis, May 7, 1860, p. 859. (See ' Microscopical Journal,' April 1861.) In some — the Saw-flies for example — the ganglionic enlargement is not elongated, but applied immediately beneath the canals ; so that, crowding as they do together, the whole seems as a compact layer of ganglionic cells. The outer nerve-sheath is attached to the edge of the skin-canal, as is shown in figs. 4, 5 & 6, reminding one of the condition of the nerve-sheath of the compound eyes by which it attaches itself to the facetted cornea. We are indebted to Leydig for the additional knowledge of the ganglionic enlargement at the termination of the nerve supplying these structures *. Upon reviewing my papers, he states also that he has found little conical "pegs or rods" (" Stifte, Stabchen") with their bases towards the integument, which are situated in the extremity of the nerve, one or two in each. These, I must confess, after a very careful search through numerous specimens (many being the same as those inv.'hich he describes them), I have been unable to discover, and am therefore under the impression that he has fallen into some error — the more so as he admits that it was only after a dozen trials that he himself succeeded. It is a difficult thing to prove a negative ; but, though I have frequently traced the nerve to its very end, I have failed to distinguish any such structures as the " Stifte " or " Stab- chen," either in specimens just killed or kept in spirit for different periods. I suspect that the method which he adopts of pulling out the nerve, by withdrawing the organized layer lining the integument internally and which passes within and lines the tube (skin-canal), has possibly led to the appearances described by him. I have noticed in the larger beetles a condition which renders this highly probable. Be this as it may, there can be no hesitation in assigning, as I have already done, an important sensory function to the organs taken as a whole. Leydig is under the impression that they are of the same nature as the organ in the hinder part of the thorax in the Acridce, &c. first pointed out ])y Miiller, Avhich he con- siders an organ of hearing. That they may be homologous to those in the Acrida, would seem to derive some confirmation from the circumstance that he has not found Miiller's organ in the same part in other insects, while I have not found those I have pointed out on the wings of the Orthoptera. But are they, therefore, auditory organs ? Have they any similarity to \vhat we know of this organ in the Invertebrata ? I confess my inability to see any resemblance beyond * Eeichert unci Du Bois-Reymond's Archiv, 18G0, No. 3, p. 299. THE SUBCOSTAL NERVURE IN THE WINGS OF INSECTS. 379 such as appertains to all sensory nerves. To what I have already advanced on this point I have nothing' to add. At the time of writing his article ahove mentioned, Leydig had not seen my second paper on the antennal organs. Nor had he been successful, as he owns, in applying the bleaching process I have recommended for their examination ; consequently neither has he, nor indeed any writer (Lespes for instance, nor his critic in this subject, Claparede), as yet noticed j^roperly the structures I have pointed out in the antennse. Under these circumstances we must consider their opinion as to the seat of the auditory function liable to revocation. EXPLANATION OF THE FIGURES. Fig. I. Termination of nerve of haltere in Rhingia rostrata, showing a row of ganglionic cells, with nuclei a a, and their terminations b, b. Fig. 2. Two ganglionic cells in situ in halteres of £j-w/a/is tenax: a a, ganglior ic enlargement; 6, terminal expansion with nucleus, c, passing to the interior of the vesicles, d. Fig. 3. Row of ganglionic cells, a a, in Eristalis tenax ; b, their termination beneath the vesicles. Fig. 4. Ganglionic cells in a Beetle (one of the Lamarice), the outer nerve-sheath attached to the skin- canal ; a a, nuclei, two or three in each : a small tube passes through the skin-canal to the closing-in membrane. Fig. 5. Shows one of the ganglionic cells with three nuclei, a a; the nerve-sheath, b b, has been attached to the inner surface of the skin-canal. Fig. 6. Shows another form, where the ganglion, a, is not so near the integument as in the preceding. Fig. 7. Organs viewed in situ from above on the nervure of Strangalia elongata : a a, ganglionic swellings with nuclei ; b, closing-in membrane and skin-canal. VOL. XXIII. 3f 381 XXIII. On Three Oaks of Palestine. By J. D. Hooker, M.D., F.E.S., F.L.S., ^c. Read June 20th, 18(J1. DURIXG a short tour in Syria, which I made last autumn in company with Mr. Han- bury, F.L.S., we paid especial attention to the oaks of that country which occurred on our route, in the hopes of being able to throw some light on their very intricate and confused history. The brief nature of our visit, and the necessity we were under of moving rapidly almost daily, and always on horseback, rendered it impossible to do more than obtain an accurate idea of the commonest species only : of these one was all but ubiquitous, from the latitude of Tripoli to that of Hebron ; and two others appeared in such great abun- dance in various localities, that we have, I think, obtained a clear conception of their limits of variation in Syria, and, by means of the specimens we procured, have identified them with those of other parts of the Mediterranean. There are Querctis psendo- coccifera, Desf., Q. JEgilops, L., and Q. infectoria, L., of which three species I find that no fewer than thirteen have been made in Syria alone, and an equal, or probably a much greater number in Asia Minor and other parts of the Levant. To the following identifications I have been guided either by good specimens or by excellent plates, amongst the very best of both kinds of which evidence I would especially indicate the Syrian and Taurian collections of Kotschy, and his superb folio fascicles entitled " Die Eichen Europa's und des Orients," the plates of which are unrivalled. 1. QuEBCUs psEtJDO-cocciFERA, Desf. Atlant. ii. p. 549. Q. Calliprinos, Webb! It. Hisp. p. 15 ; Kotschy, Herb. It. Cilic. No. 397; Willd. Sp. PL iv. p. 434. Q. Pseudo-coccifera ? Kotschy, Herb. It. Syriac. No. 265 ! Q. inops, Kotschy, fid. Herb. Reg. Berol. Lebanon, No. 83 ! Q. echinata, Kotschy, fid. Herb. Reg. Berol. Lebanon, No. 84 ! Q. PaliBstina, Kotschy, I. c. t. xix., et Herb. It. Syriac. No. 442 ! Q. Aucheri, Jaub. et Spach, 111. PI. Or. p. 11.3, t. 58 ! Q. rigida, Willd. Sp. PI. iv. p. 435 ; Kotschy, I. c. t. viii., et Herb. It. Cilic. No. 404 ! Var. Kotscliyi, Herb. It. Syriac. No. 264 ! Var. microphylla, Kotschy, It. Cilic. No. 399 ! Var. pungens, It. Cilic. 400 ! 806. Var. arborea. Herb. It. Cilic. No. 403 ! Cf. Q. Fenzelii, Kotschy, I. c. t. xxiv. et Herb. It. Cilic. No. 398 ! Cf. Q. aquifolia, Kotschy, Herb. It. Cilic. No. 402 ! Hub. Per totam Syriam, ad 6000 ped. ascendens ; copiosissime ! Fl. vere ; fr. Nov. Dist. Hispania ! et Lusitania ! Italia ! Algeria ! Cephalonia ! Grascia ! Asia minor ! Creta ! Of the above synonyms, Q. inops, echinata, and Palcestina were founded on Syrian specimens by Kotschy ; that of Calliprinos by Webb, who attempted to discriminate Labillardiere's original Syrian specimens oi pseudo-coccifera, from Desfontaines' Algerian plant of that name. The names of pseudo-coccifera and Calliprinos are also both applied 3 F 2 382 DR. HOOKER ON THREE OAKS OF PALESTINE. by Kotscliy to Lebanon specimens. Q. rigicla was fotincled on Caramanian, and Fenzelii on Taurian specimens. With tbe exception of Q. CalUprinos, there has been little attempt at diagnostic discri- mination of these species by their several authors ; and the chief characters relied on to separate them — those of form of leaf, and length of the gland and involucre— we found to fail and hold on the same tree in numberless instances. With regard to the other so-called species, long descriptions of some have been published ; but these contain no contrasting characters, and the spequnens show none. Most, if not all, of the forms indicated under the whole seven names may be found in one morning's ride along the crest of Mount Carmel, where the greater number of the series of acorns figured in Plate XXXVII. were collected, all of them from trees which, in habit of growth, foliage, &c., presented little variation. Q. pseiiclo-coccifera is by far the most abundant tree throughout Syria, covering the rocky hills, of Palestine especially, with a dense brushwood of trees 8-12 feet high, branch- ing from the base, thickly covered with small evergreen rigid leaves, and bearing acorns copiously. On Mount Carmel it forms nine-tenths of the shrubby vegetation, and it is almost equally abundant on the west flanks of the Antilebanon and many slopes and val- leys of Lebanon. Even in localities where it is not now seen, its roots are found in the soil, and dug up for fuel, as in the valleys to the south of Bethlehem. Owing to the indiscriminate destruction of the forests in Syria, this oak rarely attains its full size. We saw but few very good trees, one of which is the famous Oak of Mamre, called " Abra- ham's Oak," of which I have given a portrait (Plate XXXVI.)*, and I saw other good ones at Anturah on the Lebanon. Leaves and acorns of both these were carefully com- pared with those of the stunted form that grew around them and elsewhere, and presented no difference whatever. The " Abraham's Oak " is popularly supposed to indicate the spot where grew the oak or lentisk (for it is disputed which) under which the patriarch pitched his tent ; and is reverenced accordingly by Jews, Mahommedans, and Christians. In general habit it much resembles the Q. Ilex as grown in this country, but does not branch so much from the base ; the bark is similar in colour and Assuring, the branches in direction, and the foliao-e in colour, &c. I would draw attention to the difference shown in the portrait, between the direction of the branchlets on the two sides of the tree : on the west, where exposed to the winds of the Mediterranean, they are much stunted, denser, and more rigid than on the east, where they are more pendulous. The diameter of foliage is given, no doubt correctly, by Porter as 90 ft., and girth of trunk as 23 ft. In the winter of 1856-57, when, in the streets of Jerusalem (elev. 2200 ft.), the snow fell deep and lay for many days, the accumulation upon the Abraham's Oak was so great, that one of the finest boughs gave way under the weight and fell to tlie ground. Mr. Schembri, the highly intelligent di-agoman, of Jerusalem, informed me that he was employed by Mr. Finn, British Consul, to bring the bough to that city for him. Owing to a superstition * " David's Oak," Q. Palcestina, Kotschy, Eichcn, tab. xix. DR. HOOKER ON THREE OAKS OF PALESTINE. 383 that any person who should cut or maim the Oak would lose his firstborn son, considerable difficulty was experienced in procuring hands to saw up the timber for transportation ; these were at last brought from Jerusalem, nearly tAventy-five miles ofi", and loaded seven camels with the wood of the one limb of this fine tree. The timber of Q. pseudo-coccifera is said to be of ordinary value ; but I could procure no precise information on this head. How far it is permanently distinct from. Q. coccijera of Spain and Italy may reasonably be questioned : the characters given, even if constant, amount to no more than enough to establish a variety upon ; and it will be seen that in geographical distribution, as in botanical synonymy, the Q. pseudo-coccifera completely overleaps Q. coccifera, extending from Spain and Algeria to the borders of the Mesopo- tamian Desert, whereas the latter is not described as inhabiting any country further east than Asia Minor. With regard to the differences in the acorns of these species, they are very slight, and a long way within the limits of variation oi pseudo-coccifera observed in Syria ; and as the acorns of coccifera take two years to ripen, there is much time for variation during deve- lopment. The European Q. coccifera is described as a bushy shrulj resembling the Holly, and is browsed on by cattle : this precisely accords with the character of pseudo- coccifera in Syria, which is further called Holly by many travellers (Martineau, Porter, &c.). Again, the specimens of coccifera cultivated at Kew precisely accord with the Syrian pseudo-coccifera, whilst the Algerian specimens of the latter (which, however, are grafted) are of a rather more straggling habit, with larger leaves. Loudon and others described pseudo-coccifera as having leaves twice or thrice as large as coccifera, thicker and less wavy, with smaller and shorter spinous serratures rather than teeth. Not one of these characters holds good, with any approach to constancy, in Syria, where large and small, more or less membranous, and more or less Avaved and spiny leaves occur on individual plants, and, more conspicuously, on adjacent ones. On Carmel, just below the convent, we gathered two forms growing with interlocked branches, with the typical leaves attributed respectively to pseudo-coccifera and coccifera, but with identical acorns — those of the former plant. The Kermes was not observed by us in Syria. Q. pseudo-coccifera is an uncommon tree in English gardens. There are young Algerian specimens, which have survived the severe frosts of 1860-61, at Kew, where also the acorns I brought from Syria have germinated. Loudon (1838) mentions but one tree known to him in cultivation as bearing this name, which exists in the Horticul- tural Society's Gardens at Chiswick. 2. Qtjercus infectouia, Olivier, Voyage, 1. 10 ; Willd. Sp. PI. iv. 436 ; Kotschy, Herb. It. Cilic. No. 863 ! 371 ! var. grosse-serrata, No. 369 ! Q. Cariensis, Willd. fid. Loudon, Arboret. p. 1928. Q. Boissie7-i, Rent., Boiss. Diagn. xil. p. 119 ; Herb. It. Syriac. No. 126 ! Q. Pfmffingeri, Kotschy, Eichen Europ. und Orient, t. xxiii., et Herb. It. Cilic. No. 373 ! Q. inermis, Ehrb. fide Kotschy, Herb. It. Syriac. No. 364 ! Q. Tauricola, Kotschy ! I. c. t. x., et Herb. It. Cilic. No. 365 ! et var. fol. undul. No. 367 ! Q. leptocarpus, Kotschy, Herb. It. Cilic. No. 372 ! Q. poly car pos, Kotschy, Herb. It. Cilic. No. 370 ! 384 DR. HOOKER ON THREE OAKS OF PALESTINE. Q. Syriaca, Kotscliy, /. c. t. i. ? Q. brachyphylla, Kotschy, /. c. t. ix. ? Q. hrjpolmca, Kotschy, Herb. It. Syriac. No. 99 ! Hab. In locis petrosis Syria; alt. 1500-5000 ped., a Libano ! ad Carmel ! Fl. vere ; fr. October. Dist. Grsecia, Bosnia ! Byzantium ! Asia minor ! Q. infectoria was established by Olivier on plants found in Asia Minor. Q. Boissieri originated in specimens brought from Zcbdany, in the Antilebanon, by Boissier ; Q. inermis, founded on Lebanon specimens gathered by Ehrenberg, is also applied by Kotschy to others from the Valley of the Kedisha (Cedars). Q. Ffceffingeri is applied both to SjTian and Cilician specimens by Kotschy ; polycarpos, leptocarpos, and Tcmricola, all to Taurus specimens by Kotschy. Of Q. Syriaca I have no example ; but Kotschy's splendid plate serves to identify it absolutely with infectoria, as does the habitat. Of Q. bracliypliylla also I have no specimens ; it is a Taurus plant, also well illustrated by Kotschy. Of Q. hypoleuca I have only one specimen, without flower or fruit, Avhich exactly resembles a young shoot of infectoria ; it was collected at Zebdany in the Antilebanon. Qiierciis infectoria was met with by us in great abundance on the east slopes of Leba- non, again on the rocky hills of Galilee south of Safed, and, lastly, on the summit of Carmel. In all these localities it occurred as a small tree 15-20 feet high, or, more often, as a bush sparingly branched, with a rather slender rugged trunk, and grey deciduous foliage white on the under surface ; it was rendered very conspicuous by the abundance of those spherical galls of a deep red-brown colour and shining viscid surface which are figured in the 17th volume of our Transactions (plate 22). Another gall, of smaller size, paler colour, softer texture, precisely similar to the gall of Q. Cerris, figured by Loudon and presenting several angular conical protuberances, was sparingly found by us in Syria, and, as we believed, on this same tree ; but we had no opportunity of identifying the specimen with fruiting ones. Neither of these galls is collected in Syria, as far as we could hear : the larger and commoner is probably not different from the Aleppo gall, though it may be inferior in quality. The acorns are of a singularly elegant form and bright amber-colour. It is very rare in English nurseries : Loudon states that he had (1838) never seen it, and knew of no living specimen near London, though, according to catalogues, it was introduced in 1822. Some of those I brought home in damp earth are now growing in the Royal Gardens. The Quercus infectoria is common throughout Asia Minor. Mr. Balansa showed us trees of it near Smyrna, and informed us that it was extremely variable in foliage and pubescence. Its western limit is apparently Bosnia, and its eastern the confines of Persia ; to the north it attains the latitude of Constantinople, and its southern limit is probably the hills of Samaria. 3. QuEECUS iEGiLOPS, Linn. Sp. PL 1414 ; Willd. Sp. PL iv. 448 ; Tchihatcheff", As Min. Bot. vol. ii. p. 470, t. xli. and xlii. Q. Pyrami, Kotschy, Eich. Europ. u. Orient, t. 3 ; Herb. It. Cilic. No. 395 ! Q. macrolepis, Kotschy, I. c. t. Ifi, et Herb. No. .S09, a! Q. Ungeri, Kotschy, /. c. t. 13, et Herb. It. Cilic. No. 390 ! Tchihatcheff, I. c. p. 473. DR. HOOKER ON THREE OAKS OF PALESTINE. 385 Q. Trojana, Webb in Jaub. et Spacli, 111. i. t. 54; Kotschy, Coll. No. 391 ! Q. Gaedelii, Kotschy, Herb. It. Cilic. No. 387 ! (et 407 ! -O- Q. Vallonea, Kotschy, /. c. t. 7, et Herb. It. Cilic. No. 80 ! 391 ! 394 ! Tchihatcheff, /. c. p. 474. Q. Ithaburensis, Kotschy, I. c. t. 12. ? Q. Look, Kotschy, Z. c. t. 21 : vide sub Q. cast anecef alia - ? Q. Ehrenbergii, Kotschy, I. c. t. 15; Herb. It. Cilic. No. 393 ! TchihatchefT, L c. p. 474 (vide sub Q. Toza). Hub. In collibus Syrise in Galilee ! et Samaria ! Fl. vere ; fr. October. Dist. Hispania, Graecia ! Asia minor! Creta ! Of the above names, the original, ^gilops, is that in general use for this the Vallonea Oak. Grisebach identifies Olivier's plant {Vcdlani, Oliv., u^Scjilops, L.) with Q. Libcmi, Oliv., and Q. castanecefoUa, C. A. M. {Trojana, Webb), a plant of Lebanon which we did not observe in Palestine. TchihateheflP, As. Min. Bot. vol. ii. p. 470, however, refers the Vallani of Olivier to ^TJgilops, L. Q. Trojana was applied by "Webb to specimens from Phrygia ; Q. Ithaburensis by Decaisne to specimens from Tabor ; and Kotschy has also applied the name to the same plant from the vicinity of Nazareth, by which, no doubt, Tabor is meant. Of Q. Look, which inhabits Hermon, I have seen only one specimen, without flower or fruit, which altogether resembles small examples of Q. JEgilops, but may be referable to Q. castanecefoUa, C. A. M. The other four names are all applied by Kotschy to Taurus forms, except Ehrenbergii, of which a fine specimen is in the Kew Museum, resembling ^Egiloj^s in the acorn, but the leaves broader and more deeply cut : this species was founded on a plant gathered by Ehrenberg at Massa, in the Antilebanon. The Vallonea Oak is, like the Q. 2^seudococcifera, very gregarious in Syria, though growing in a widely different manner, never forming a bush or growth of underwood, but rising, on a stout gnarled trunk 3-7 feet in girth, to the height of 20-30 feet. Wherever we saw it, as on the hills east of Nazareth, on Tabor (where it is abundant), to the east of Caifa, and on the N.E. flank of Carmel, it forms scattered, rather round- headed, densely leafy trees, giving an open park-like appearance to the landscape. From its stout habit and sturdy limbs, and from a specimen of the wood of the Basan Oak given to the Kew Museum by Cyril Graham, Esq., I should be inclined to suppose that this was the Oak of Basan. The wood is said to be excellent ; and the tree is, like all other timber in Syria, indiscriminately cut for house-fitting and fuel. At the foot of Tabor, almost all the largest trees were mutilated by ineffectual attempts on the part of the Turks and Araljs to fell them ; none of these were more than about 6 feet in circumference. The gland of the acorn of this species often attains a very large size : some we gathered were 2^ inches long, and 3 inches in girth ; but they vary extremely on the same tree, some being so small that I have had difficulty in discriminating between them and large ones of Q. pseudococcifera. I have examples which, had they not been taken by myself from an JEgilox)s tree which also bore large acorns, I should have attributed to hybri- dization between the two. When fully ripe the gland is still green ; and in this state it germinates, the pericarp never hardening. These may be seen in all the bazaars, raw and boiled, in which state they are eaten by Turks and Arabs ; rude ornaments are also made of the shell. The seeds of this species which I brought home have germinated in 386 DR. HOOKER ON THREE OAKS OF PALESTINE. liie Royal Gardens. Loudon states that a gall similar to that of Q. iiifectoria, but more angular, is found on Q. yEgilops : this was not observed in Syria. The Vallonea Oak, though abundant in Asia Minor, &c., is almost limited in its range to the countries east of the Adriatic and west of Persia, except it should prove that Q. Fersica is a form of it, which to me appears quite possible. Nyman gives Spain as a habitat. It is rare in British gardens, though plants 50-GO feet high exist. It may be well to mention here what I know of the other Oaks said to inhabit Syria. These, which are all confined to tlie northern mountain ranges, are — 1. QuERCUS Cerkis, Linn. Sp. PL 1415 ; WiUd. Sp. PI. iv. 454. Q. Cerris var. Caramanica, Kotschy, Herb. It. Cilic. No. 80 ! & 385 ! — Var. Cilicia, Herb. It. Cil. No. 386 ! Q. crinita, Bosc, Mem. Chen. p. 19, fid. DC. Flor. Franc, vi. p. 354. Q. Austriaca, Willd. Sp. PI. iv. 454 ; Kotschy, Eich. Europ. u. Orient. 20. Q. Haliphloeos, Juss., Lamk. ! fid. Herb. Benth. Q. Pseudocerris, Boiss. at Reut. ! Kotschy, Herb. It. Syr. No. 328! The original Q. Cerris is a native of Spain, Prance, Italy, Austria, Greece, and Asia Minor, and Q. Hcdijihloeos of Asia Minor ; Q. Pseudocerris of Mount Casius in North Syria, and the Valley of the Kedisha above Tripoli. 2. QuERCUS TozA, Bosc, Journ. Hist. Nat. ii. p. 155, t. 32. f. 3, fid. DC. Flor. Pranc. vol. vi. p. 352 ; Kotschy, Eichen Europ. und Orient, t. xxii. Q. subalpina, Kotschy ! Herb. It. Syriac. No. 335. Q. Pyrenaica, Willd. Sp. PL iv. 451. Q. Tauzin, Persoon, Kotschy, Herb. It. Syriac. No. 336 ! Kotschy gives Q. Toza {Tauzin, Pers.) as a native of the Kedisha Valley ; and his spe- cimens appear in no way different from South European ones ; they do not differ from his Q. suhalpina from the same locality. His Ehrenbergii (vid. sub ^gilops) may be referable here. Q. Toza is a native of the Pyrenees and South of Prance also. 3. QuERCTJs CASTANE^EOLiA, C. A. Meyer in Eichwald Caucasus, fasc. ii. p. 9, t. 1 ; Jaub. et Spach, lUust. t. 54. ? Q. squarrosa, Kotschy, Herb. It. Syriac. No. 100 ! ? Q. Look, Kotschy, /. c. 172. ? Q. carpinea, Kotschy, /. c. No. 98 ! This handsome species (the true ^gilops of linnajus, according to Grisebach) is a native of the regions south of the Caucasus, Asia Minor and Persia, it having been ori- ginally described from Mazanderan, on the south shore of the Caspian. Kotschy's Oak No. 98, from Zebdany in the Antilebanon, appears to be the same plant, as far as can be judged by leaves alone, as may also be the Q. Look aUuded to under Q. ^gilops. 4. QuERCUS LiBANi, Oliv. fid. Kotschy, I. c. t. 5 ; Herb. It. Cilic. No. 380 ! et var. foliis obscurioribus, I. c. 384 ! A noble species, allied to Q. castanecefolia, if not a mere form of it. It is a native of Cilicia, and Kurdistan according to Kotschy, but was originally discovered in the Lebanon by Olivier, who sent specimens to Desfontaines. DR. HOOKER ON THREE OAKS OF PALESTINE. 3S7 5. QuERCUS MANNiFEiiA, LincU. ! Bot, Reg. 1840, Misc. p. 40 ; Kotscliy, Ilei-b. Syriac. nullo numero (Herb. Eeg. Berol.). Q. Dschorochensis, Koch ! Kotschy, Herb. Syriac. sine numero (Herb. Reg. Berol.). The Q. mamiifera was established by Lindley on a plant brought from Lake Van, and its close affinity with Q. sessilijlora indicated. Our specimens so called by Kotscliy (Herb. Reg. Berol.) seem identical with his Q. Dschorochensis from Syria, a plant also found in Asia Minor. EXPLANATION OF THE PLATES. Plate XXXVI. Tree of Quercus i^seudo-coccifera, Desf., near Hebron, called Abraham's Oak. Plate XXXYII. Acorns of various forms from trees of Querent pseudo-coccifera, Desf. Plate XXXVIII. Acorns of various forms from trees of Qurcus ^gilops, L. VOL. xxiir. 3 G [ 389 ] XXIV. On Fissicalyx and Prioria, ttoo recently j^uhlislied Genera o/" Leguminosee. By Geoege Bentham, Esq., Fres.L.S. Eead June 20th, 1861. 1. FissicALTX Fbndleei, Bentli. vV'HEiN I gave tlie characters of this remarkable species, forming a very distinct genus of Dalbergiese, in a recent Part of the Journal of the Society's Proceedings (vol. v. p. 78), I stated that the loose fruits distributed with the specimens were those of a Giiaiacum. In this I was misled by the fruits of the Guaiaciim arhoreum, distributed in the same set, having been accidentally exchanged with them in my parcel. I have now ascertained that the loose fruits generally sent with the Fissicalyx really belong to it, and confirm the very distinct character of the genus. Not only are the calyx and anthers exceptional among Dalbergiea?, but the broad wings do not, as is usually the case in that tribe, and in other 1- or 2-winged legumes, arise from the sutures, but from the centre of the valves themselves, like the secondary wings in the 4-winged Loti and Clitorice, or in Fsoplwcarpus, Tetrapleiira, Gognehina, and other ^-winged Leguminosse. Here, lioweA'er, the sutures are not at all dilated, and the legume has only the secondary wings, the cavity of the pod remains very narrow for its length, and the single seed in the one I opened had linear cotyledons, of which one was longer than the other and hooked over it at the extremity. The embryo is perfectly straight, as in Csesalpiniefe ; but that is the case in several Dalbergiea3. The following is the technical character of the genus, with tlie description of the species, as completed by that of the fruit. Elssicaltx, Bentli. Char. Gen. — Calycis tubus turbinatus ; limbus acuminatus, integer v. apice brevissime bidentatus, per anthesin hinc fissus, spathaceus. Petala ad apicem tubi cum staminibus inserta, papilionacea. Vexillum ovatum. Alee eo vix breviores, oblique oblongas. Carhice petala subslmilia, paullo minora, libera. Stamina 10, monadelpha, vagina supra fissa. Antherce versatiles, apice biporosae. Ovarium breviter stipitatum, biovulatum. Stylus filiformis, apice minute stigmatosus. Legumen cum alis planum, ovali-ellipticum, ipsum tamen angustum, suturis parum prominulis, valvulis medio longi- tudinaliter late alatis. Semen unicum, prope apicem affixum, oblongo-lineare. Albumen 0. Testa tenuis, hilo parvo. Embryo rectus, radicula brevi supera, cotyledonibus linearibus. Species unica, P. Pendleri, Benth. Arbor. Polia imparipinnata. Poliola (11) opposita, exstipellata, petiolulata, inferiora ovata, superiora oblonga, acuminata, 3-^-pollicaria, basi rotundata v. subcordata, glabra, membranacca. Panicula terminalis, molliter pubescens. Bractese stipulares, ad basin ramulorum parva3, sub floribus minutte. Plores conferti. Pedicelli ad singulas bracteas solitarii v. gemini, vix lineam longi. Bracteolse parvsE, acutse, persistentes. Calyx semipollicaris, pubescens, subfoliaceus, 3 g2 390 MR. G. BENTHAM ON FISSICALYX AND PRIORIA. apice longe, basi brevius attenuatus, summo apice minute penicillatus. Petala aurantiaca, calyeem pauUo superantia, breviter unguiculata. Vexillum basi com- plieatum, nee auriculatum nee appendiculatum, ereetum, lateribus reflexis, glabrum. Alte et petala carinalia basi liinc rotimclato-aiiriculata, supra basin trausversim rugu- losa, versus apicem pilosa. Legumen breviter stipitatum, puberulum, cum alis 2 poll. Ion gum, 1^ poll, latum, apice late rotundatum et emarginatum, valvulorum alis membranaceis, trausversim parallele venosis. Cotyledones in semine examinato inse- quales, longior apice uncinata. Hub. In A'^enezuela, between Turmero and Maracai, at an elevation of l700 feet above the level of the sea [Fendler, n. 2223). Tab. XXXIX. fig. 1. Flower; fig. 2. the same with the petals removed, showing the stamens; fig. 3. Vexillum ; fig. 4. one of the Alas ; fig. 5. one of the Petals'of the Carina ; fig. G. a Stamen, front view ; fig. 7- the same, back view ; fig. 8. Ovary and Style ; fig. 9. Ovary, longitudinal section showing the ovules; fig. 10. Fruit; fig. 11. the same, cross section; fig. 12. the same, longitudinal section showing the seed; fig. 13. Seed, showing the hilum ; fig. 14. Embryo; fig. 15. the same with one cotyledon removed, showing the plumula. 2. Prioiiia copaipera, Griseb. This plant, forming a distinct genus allied to Copaifera, was dedicated by Dr. Grisebacli to Dr. Alexander Prior, so well known for bis botanical researches at the Cape, in the West Indies, and Southern Europe, and was described by hun in the last Part of his ' Plora of the British West Indies,' p. 215, from specimens gathered by Mr. Wilson at Bachelor's Hall, near Bath, in Jamaica. The tree is there rare, and the fruit was unlvnown. Since then, however, we have received specimens, both in flower and in fruit, of evidently the same species, from Mr. Sutton Hayes, who found it growing in swamps in Panama, near the Barbacoa station of the Panama Railroad. He states it to be there known under the name of Cativa. The tree, which was felled to obtain the specimens, was near 100 feet high, and the trunk measured 5 feet in diameter. The pod confirms in many respects the afiinity of the genus with Copaifera ; it is, however, very much larger and flatter, woody rather than fleshy ; the single seed is very flat, without any albumen ; the cotyle- dons, of a somewhat hard fleshy consistence, adhere so closely that, after thoroughly boiling and soaking the seed, I was unable to separate them except close round the edge ; the radicle forms only a slight protuberance, at no great distance from the hilum ; and altogether the seed shows none of that approximation to the Connaraccous seed observable in Copmfera. We find in the Kew Museum some of the same fruits, deposited there by Dr. Seemann as those of the " Amanza muger " of the Isthmus of Panama, sold in the markets there for their exciting properties. They do not, however, agree with the description of the "Amanza muger" fruits given in the "Botany" of the 'Voyage of the Herald ' (p. 90), which are there referred to a Ilippocratea. The following are the technical characters of the plant. Priokia, Griseb. Char. Gen. — Calyx tubo brevi cupulato, lobis 5 orbiculatis subpetaloideis valde imbricatis, infimo ex- teriore. Petala 0. Stamina 10, libera, subfcqualia. Anthera versatiles, 2-loculares, 2-rimosae, con- TtiAiTsIiwK. Soc YoL XXIILTab. 39. Hssicalyx Peiidlen. TsANb-iJi'iiT- Soc.Yol.ISIII.Tab. 40. Priori a. copaifera. MR. G. BENTHAM ON FISSICALYX AND PRIORIA. 391 nectivo crasse apiculato. Ovarium sessile, biovulatura. Stylus breviter filiformis, apice minute stigmatosus. Legumen planum, coriaceo-sublignosum, obovato-orbiculatum, reticulato-veiiosura, tardius bivalve. Semen ab apice pendulum, orbiculatum, cavitatem Implens, planum. Albumen 0, Testa tenuis. Cotyledones carnosae, induratte, arete cohterentes ; radicuk brevissima, mammiforniis, a hilo pavum remota. Plumula inconspicua. Species unica, P. copaifeka, Griseb. Arbor procera, glabra, ramis patentibus. StipuliB squamasformes, caducre. Eolia alterna, abrupte pinnata, petiolo comnmni 1^-2-pol- licari; foliola 2-juga, rarius 1-juga, opposita, oblique v. falcato-ovata, 3-6-pollicaria, breviter et obtuse acuminata, basi rotundata v. cuueata, petiolulo 2-3 lin. longo, coriacea, utrinque reticulato-pennivenia, pellucido-punctata. Paniculse terminales, amplEe, parura ramosae, ramis iuterrupte spiciformibus 2-6-pollicaribus. Flores parvi, sessiles, dissiti v. subfasciculati. Bractete inconspicuae. Bracteolas in cupulam bilobam calycis tubum tequantem connatae. Calycis tubus vix \ lin. longus ; laciniEe concavse, lineam longse, margine minute ciliolata3. Eilamenta ad marginem tubi inserta, laciniis longiora, basi ciliato-birsuta. Antlieras ovato-oblongae, loculis curvulis; conneetivi apiculus carnoso-conicus, Humirium quodammodo referens, terminalis, sed anthera reversa inferus. Ovarium villosum. Legumen S^-i poll. Ion- gum, 8-8|- poll, latum, basi subcuneatum, facie altera convexa, altera concava, venis elevatis longitudinalibus reticulatis, ad apicem ct versus basin convergeutibus. Semen maturum 3 poll, longum, 2^ poll, latum, undique fere 2 lin. crassum. Hab. Jamaica [Wilson); Isthmus of Panama [Seemann, Sutton Hayes), Tab. XL. fig. 1. Flower; fig. 2. Stamen with a back view of the anther; fig. 3. Anther, front view; fig. 4. Ovary and Style; fig. 5. Ovary, longitudinal section, showing the ovules; fig. C. Ovule: all the above magnified. Fig. 7. Fruit, natural size; fig. 8. the same, witli one valve removed, showing the seed. [ 393 ] XXV. The Foot of the Fly ; its Structure and Action : elucidated hij comparison with the feet of of her Insects, 8fc. — Part I. By Tufpen West, I]sc[., F.L.S. Read March 21st and June Gth, 186 1. The structure and action of the Fly's foot have been so frequently treated of, and are so generally considered to be fully understood, that it may appear, at the first glance, as if nothing further could be done with so hackneyed a subject. Whilst, however, in con- sequence of some investigations which have been recently made to determine its structure, opinions jnay now be considered as pretty well fixed with regard to this portion of our subject, authorities still differ widely respecting the action of the parts of which the foot is composed. Nor has it appeared that any satisfactory progress was making toAvards a reconcilement of the differences of opinion on the subject, — each succeeding author setting aside, more or less, the opinions of his predecessor in point of time, only to be himself again contradicted by his successor. On reflection, it occiirred to me that this probably arose in great measure from the attention of observers having been confined too exclu- sively to the Flf's foot, and that perhaps, amongst the teeming myriads of insect-life, structures might, on careful search, be met with resembling those under consideration in every respect but that of their being on a larger scale. Supposing this to be the case, the advantages gained, b6th for purposes of observation and of reasoning, will be very great, as will be best appreciated by those who have had large experience in micro- scopic researches on very minute parts. The high powers necessary for investigations under such circumstances demand so many appliances in the instrument of research, with such great sldll in the use of them — all these tending to introduce new sources of error — that the student feels it a relief indeed when he can assist his mind tov,-ards forming con- clusive opinions by availing himself of larger examples of similar structures. That this advantage may be gained in the present instance will appear to some extent during the course of the following remarks, and still more strongly when I come to speak, on a subsequent occasion, of other modifications of the parts noAV to be noticed, of the types of feet in Insects which characterize, with few exceptions, the grand divisions in a syste- matic classification, and of the analogies and homologies respectively of the parts. It is intended now simply to indicate the leading divisions of the subject, and to show that parts similar to those on the Ely's foot, but on a larger scale, are to be met with abun- dantly in other insects. And, this advantage being gained, it is hoped that a fuller un- derstanding of the nature and purposes of these structures jn the insect economy will also be reached, and that the conclusions to be drawn will, fronl the wider basis on which they rest, satisfy more completely than hitherto all the conditions of the question. It will be necessary to give a brief sketch of thQ labours of previous observers on this subject, before stating the results of my own investigations. The earliest mention which I find of any minute examination of the feet of the Ely is 394 MR. TUFFEN WEST ON THE FOOT OF THE FLY. in an interesting little work by George Power (IGGi)*. He lield that the claws enabled the Fly to take hold of any roughness of the surface on which it might be moving ; and when the latter was too smooth to afford any advantage of this kind, that there was "a fuzzy kind of substance like little sponges, with which she" (Nature) "hath lined the soles of her " (tlie Fly's) "feet, which substance is always repleated with a whitish viscous liquor, which she can at pleasure squeeze out, and so sodder and be-glew herself to the plain she walks on, which otherways her gravity would hinder (were it not for this con- trivance) especially when she walks in those inverted positions." The near approach gained by Power, at one stride, towards the true structure of the minute parts in question, as well as towards what I believe to be the true explanation of their action, is indeed remarkable, when the early date of his observations is considered. Had he only seen the hairs on the Ply's cushions, nothing would liave been left for suc- ceeding observers to discover, except such minute details of structure as are revealed by the higher ])owers of our modern microscopes. The next observer was the well-known Hooke, who -wrote only three years later (1667) f. He saw, in addition to what Power had made known, the minute hairs on the under sur- face of the " soles," and recognized (as later observers hardly appear to me to have done) the importance of the grasp gained by the claws, when drawing against the strong, for- ward-pointing and sharp spinoiis hairs (one at either side of the basal line of each tarsal joint, underneath), wherever any projections, or a yielding surface, are presented by the plane on which the insect may be moving. The effect gained in such an action of the claws, when draAvn backwards and inwards against the " soles with their small bristles or tenters, which have their points looking the contrary way, that is, forwards and out- wards," is distinctly mentioned and insisted upon. " If there be any irregularity," he continues, " or yielding in the surface of the body, the Ply suspends itself very firmly and easily, without the access or need of any sponges filled with an imaginary gluten:'' Further on, he proceeds to show how these " tenters " (a better name could not be devised for them)— /^o^y these " tenters or points, whereof a Fly has about ten in each foot, to wit two in every joint," assist in the action. This was the "mechanical theory," in its purest expression, as the preceding had been the " viscid " one. It was a decided advance towards a complete explanation, although manifestly insufficient for all the facts of the case. He believed in the existence of a " smoky substance on glass," into which he thought that the minute hairs of the " soles " penetrated, and so assisted the Fly in holding on to that smooth and slippery material. This belief in a " smoky substance on glass " has been considered a mistake by every successive writer on the subject ; and yet it is certain that glass very frequently undergoes a slow decomposition on its surface in a moist atmosphere, from the excess of alkali in its composition. Such a change is speedily produced in glass exposed to the action of the weather, as in our window-panes, and con- veys the appearance as if a " smoky substance " were condensed upon it. It has been proved ])y most careful observations, which may be readily verified by any one desirous of getting at the truth, that this tarnish does very materially assist a Fly when in a weak state in maintaining its hold, and in freely moving upon the glass. To keep our windows ♦ Experimental Philosophy, p. 5. t Micrographia, p. 1/0. MR. TUFFEN WEST ON THE FOOT OF THE FLY. 395 clear for the admission of light, it requires to be constantly removed. To opticians, a similar deposition of moisture on the glasses of their lenses is but too well known ; they call it the "sweating of glass." Hooke clearly foresaw the difficulties presented by the "viscid secretion" theory, and put them in the following forcible manner : " I could not well comprehend how, if there were such a glutinous matter in those supposed sponges, as most (that have observed that object " [the Fly's cushions] " in a Microscope) have hitherto believ'd, how, I say, the Fly could so readily unglew and loosen its feet." Leeuwenhoek {circa 1690)* brought to this, as he did to every other subject he investi- gated, his usual clear-headed sagaciousness. He saw flies, " almost as large as a bee," " every year " "in the month of August, sitting on a glass, at the backside of " . . . "his house" (Uristalis tenax). "The extremities of their feet," he found, "were covered with an incredible number of hairy parts, by the help of which they are better able than other flies to climb up a glass though it be ever so free from impurities or irregularities of which they might take hold. I have therefore," he says, " often placed the feet of those flies before the microscope, in order to view the means by which they can fasten themselves to the glass, and run up it ; and I have for some years past thoiight that I could discover that these hairs were each of them provided with crooked parts like hooks, by the help of which they can take the firmer hold on glass ; but which parts have never, to my know- ledge, been described by any person, though the figiires of those hairs may be seen in many authors." Here there is a distinct reference to the peculiar flexure downwards of these parts, which will presently have to be mentioned in the description that I shall give of them from my own observations. Leeuwenhoek, thinking that the action of these hairs was purely mechanical (/. e. that +.hey acted only like so many minute hooks), illustrated his views by a description, which follows the above, of his observations on some remarkable hairs on "the hind-feet" of " laro-e crabs "...." caught among the rocks in Norway ;" which hairs are " many of them provided with a double row of parts like teeth, placed in very exact order beside each other, in like manner as if we were to imagine the back of a knife cut into a double rovv^ of teeth or notches." "This wonderful formation, I am persuaded, is intended for this pur- pose, that when the crab is climbing up the rocks, he may be enabled by this assistance to fix his feet firmly on the rocks or stones f." He also describes another form of insect-foot which bears distinctly on our present purpose, in the notice of a "Fly" infesting " the blossoms of fruit-trees, particularly apples " {Anthonomns 2^omorum ?). " Observing," he says, " that these insects " could run along or stand for a long time on any side of the glass, even with their feet upwards, I was desirous to examine accurately the formation of their feet, and, in tliis little creature, I saw such perfectly formed limbs, enabling it to adhere to the glass, and to run along upon its surface, as distinctly as I had ever seen in other larger flying animals." * Collected Works (translated by Samuel Iloole, 1800-1807), vol. ii. part 3. p. "1. t Ibid. pp. 71, 72. VOL. XXITI. 3 II 896 MR. TUFFEN "WEST ON THE FOOT OF THE FLY. A figure which is given of the part is then clescrihed at length, but it is not necessary to repeat the details here. The figure purports to be a representation of " almost a fourth part of one of the legs, consisting of four distinct joints : there " are two claws or nails, which are in some sort transparent, like a piece of horn or tortoise-shell seen with the naked eye," and "organs" (on the ' deeply-bilobed third tarsal joint'*) "by the help of which the animal can run along the smooth surface of the glass, and also hang to it a whole night. The formation of these organs is very wonderful, for all those parts witli which they are covered, and which one would conclude to be hairs, are so exactly and regularly sloped oif, and particularly of such regular lengths, as if they had been clipped with scissars, that when the animal places its foot anywhere, they all touch the place at the same time, and what is more, all these particles, which seem to the eye to be hairs, have at their extremities a hook, and at a little distance from thence, two other hooks ; but by reason of their extreme minuteness, though the limner confessed he saw them through the microscope, he declared he could not represent them through the drawing. Now if we consider, what I have always experienced, that a glass, though washed ever so clean, will have many particles adhering to it, though these are so small, that the claws on the feet of flying insects cannot take hold of them, we may easily con- ceive that these small hooks may take hold of the small particles of water or motes from the air adhering to the glass. And here we may discover the error of those, who for- merly supposed there were cavities in glass, v\^lierein flies could fix their claws, and climb upt." I form no opinion of the absolute correctness of Leeuwenhoek's description of the parts which "seem to the eye to be hairs," with "at their extremities a hook, and at a little distance from thence, two other hooks," which, "by reason of their extreme minuteness, though the limner confessed he saw them through the microscope, he declared he could not represent them through the drawing," until I have been so fortunate as to obtain living specimens for examination. I have scarcely, as yet, detected one error of observation in the descriptions by the present author, of his extended microscopical researches into objects embraced by the three grand kingdoms of Nature, and am therefore unwilling to ascribe error to this portion of the present observation, without feeling per- fectly clear that such error exists. An instatLce of an analogous structure will be men- tioned presently, which perhaps may help to explain the meaning of the appearance which he thus describes. By his careful, and at the same time lively, description of the remarkable liolding and climbing organ of the " Sea-Mussel" (byssus of Modiola vulgaris) he shows that he was fully acquainted with one instance at least of organs acting really by suction : his observations on this head will be mentioned further on, in the course of my own remarks upon such organs when presented by insects. Derham's claim (1798) t to rank as an original observer rests, so far as I can ascertain, on very slender grounds. Yet, to judge from the influence which his remarks have had in the forming of the opinions of men of the highest scientific acumen, even at the present * Westwood, «;«;»». t 0^3. C!it. part 3. pp. 186, 187. % Physico-theology, Part ii. p. 289. MR. TUFFEN WEST ON THE FOOT OF THE FLY. 397 time, tlieir importance wotild seem to have been considered paramount. He compared the Fly's foot with that of Dyticus ; but the superficial observations he made of the latter go to prove that he cannot have used a magnifier of any kind in examining the great Water-beetle ; or he could not have failed to observe parts (viz. the great suckers : to an eye practised in minute observation, the whole set of small suckers as well) clearly enough visible to the naked eye. His words shall speak for themselves. After simply saying, " I might here name divers flies and other insects, who, bes.ides their sharp-hooked-nails, have also skinny palms to their feet to enable them to stick on glass and other smooth bodies, by means of the pressure of the atmosphere," he adduces the following proof of the correctness of his explanation respecting the action of Flies' feet : — " But because the example wUl illustrate another work of Nature, as well as this, I shall chuse a singular piece of mechanism in one of the largest sorts of Hydrocanthari. Of these large ones there are two sorts, one largest all black, with antennce handsomely embossed at the ends " (SydrojyMlus). " The other somewhat lesser, hardly so black, with capillary autenncs ; the forehead, edges of the vcigliice " (elytra), " and tAvo rings on the thorax, of a tawny colour " {Dyticus marginalis). The female hath mgincs prettily furrowed; the male, smooth. But that which is most to our purpose in this male is a flap, or hollowish cap near the middle joint of the forelegs, which, when clapped on the shoulders of the female in coitu, stiqks firmly thereon. And then foUows the comparison which, for want of duly weigh- ing the value of the evidence, has misled, more or less, every succeeding author on this subject : — " after the manner as I have seen boys carry heavy stones, with only a wet piece of leather clapped on the top of the stone." Derham has not a word fm'ther on this head ! His observations and remarks evidently possess no intrinsic value ; but it was necessary to ascertain distinctly what amount of credit was their due, before proceeding further in our inquiry. Gilbert White (1788) * was no microscopist, and in writing on this subject he did but repeat the prevalent opinion of his day. He was, however, a shrewd, painstaking observer of facts in natural history, who noted the habits of flies as carefully as of the other animals that came under his eye. And because a construction has been put upon his words which was never intended by their author, and for another reason that Avill appear, it is necessary just to see what he really did say on the present subject. Joiniag, then, in the general behef, he states that " in the decline of the year this re- sistance" (the pressure of the atmosphere) "becomes too mighty for their diminished strength ; and we see flies labouring along, and lugging then' feet in windows, as if they stuck fast to the glass, and it is with the utmost difficulty they can draw one foot after another, and disengage their hollow caps from the slij)pery surface." This paragraph, when connected mth one which precedes it, points to a very curious fact in the Natural History of Flies. " As they grow more torpid " (through the advance of cold weather) " one cannot help observing that they move with difficulty, and are scarce able to lift their legs, which seem as if glued to the glass ; and by degrees many do actually stick on till they die in the place." It becomes evident from the above that the temporary and voluntary attachment of * Natural History of Selborne, edition with Bennett's noteSj 1837, pp. 474, 475. 3 h2 398 MR. TUFFEN WEST ON THE FOOT OF THE FLY. flies to the objects on which they rest or walk, was confounded by Gilbert AVhite with, and explained by a reference to their permanent and involuntary attachment when dead or dying from the attacks of the remarkable parasite called ' Empiisa Ilusca; ' by Cohn. Mr. John Blackwall (1830)* sent three communications to the Linnean Society, bear- ing upon this subject. In the first is given a description of the peculiar structures met •with on the feet of such spiders as are capable of ascending polished perpendicular sur- faces, so far as he had opportunities for making observations upon them. He at this time expressed the opinion that " the minute bristles with which the tarsal cushions of many insects, remarkable for their ability to walk up glass, are furnished, appear to possess an organization closely analogous " to that of the feet of these Spiders f. The importance of attending to the amount of development of the structu.res for holding, in connexion with the relative size of the body, and to the state of physical strength at the time of the ex- periments, is clearly pointed out. My limits will not permit me here to explain the way in Avhich the appendages to the legs of spiders having this remarkable climbing power act ; but on a future occasion I shall enter upon this branch of the subject. In his second communication the parts entering into the composition of the Ply's foot are described. The discovery that the hairs clothiug the lower surface of the tarsal cushions have their extremities enlarged in this insect was now announced. " Tlie production of a vacuum between each membrane" (tarsal cushion) "and the plane of position .... Avas at once seen by Mr. Blackwall, on examining the parts with a com- pound microscope, to be clearly impracticable, unless the numerous hairs on the under side of these organs individually perform the office of suckers." The quotation continues — " there does not appear to be anything in their mechanism which in the slightest degree countenances such a hypothesis. When highly magnified, their extremities, it is true, are seen to be somewhat enlarged ; but whether they be viewed in action or ia repose, they never assume a figure at all adapted to the production of a vacuum." I shall proceed to explain, in due course, my own view of these interesting facts and statements. With respect to the action of these parts, an air-pump experiment is mentioned, by which it was considered to have been " demonstrated, to the entire satisfaction of several intelligent gentlemen present, that the House-fly, while it retains its vital powers imim- paired, can not only traverse the upright sides, but even the dome of an exhausted receiver ; and that the cause of its relaxing its hold and ultimately falling from the station it occupies is a diminution of muscular force attributable to impeded respiration." Mr. Blackwall felt the desirability of a minute examination, for purposes of comparison, of analogous structures in other insects, and appears to have examined the cushions on the tarsi of several beetles, respecting which he says :— " If the slender bristles on the inferior surface of the pulvilli of some of the larger Coleoptera, Frionus cervicornis for example, be very highly magnified, each, beside the numerous short hairs which project from its sides, will be found to have a small dense brush of exceedingly minute hairs at its extremity ; and as the hairs on the pulvilli of flies, and many other insects belonging to various orders and genera, with which I have experimented, perform a function similar to that exercised by the bristles, and also exhibit a striking resemblance to them in external * Transactions of the Linnean Society, vol. xvi. t Ibid. p. 471. MR. TUFFEN WEST ON THE FOOT OF THE FLY. 399 appearance, it is extremely probable that they are analogous in structure, though from the smallness of then- dimensions I have not yet been able to satisfy myself that this is the case by direct observation, notwithstanding I have employed the highest magnifying power at my command*." It appears to me that an error has crept in here, through the inadequacy of the instrument employed in these investigations into structures so minute and delicate as those in question. The under surface of the enlarged ends of the hairs is always perfectly smooth in beetles, and all other insects, though the same surface is minutely serrate in the climbing spiders. There are a few small scale-like processes on the upper surface of the same part in species belonging to the Silphidse, Staphylinidas, and Cerambycidte ; but, from the difference in their structure, I am compelled to think that the hau-s on the under surface of the tarsi of these spiders and beetles must differ mate- rially in their action. In Frionus coriaceus this serration on the upper surface of the hairs forming the tarsal cushion presents an appearance somewhat like that described by Mr. Blackwall, but there are certainly no serrations on the under surface. Mr. Blackwall admits that " whoever examines the most carefully polished glass in a favourable light, with a powerful lens, will speedily be convinced that it is not free from flaws and imperfections," whence " the hold which insects are enabled to take of any roughness or irregularity of surface must be very considerable." " The mechanical theory of the movements of flies on polished perpendicular surfaces," considered to have been " established," did not prove quite satisfactory on still further experiments being tried. Accordingly Mr. Blackwall sent a third communication on the subject, which will be found amongst the " Extracts from the Minute-Book " of the Society!. In it is announced the further interesting discovery, that flies walking over well-cleaned glass- leave such indications of their steps as would be produced by the emission of a fluid from the tips of the hairs on the under surface of the tarsal cushions. On further trials with the air-pump, it was "observed" "in experimenting upon the House-fly" "that individuals frequently remained fixed to the sides of an exhausted glass receiver after they had entirely lost the power of locomotion, and an evident distension of the abdomen had been occasioned by the expansion of the aeriform fluids it contained. To detach them from these stations, the employment of a small degree of force was found requisite." With respect to the circumstance of flies remaining attached to the bell-glass of the air-pump, it may sufiice to advert to their very low specific gravity (a large Blue-bottle Fly not weighing more than a grain), which renders it evident that a very slight degree of viscidity in the fluid emitted from the tips of the hairs— a quality which it undoubtedly possesses — would effect such an adhesion as occurred in the cases mentioned. Experiments were then made to determine the effects of moisture, oil, and powders of an hinoxious nature ("flour of wheat," "finely pulverized chalk or gypsum"), in pre- venting the flies from adhering to the surface of the glass, and their results are mentioned. It appears to have been forgotten in these experiments that the particles even of what we call 'impalpable' powders are mostly quite as large as, and often vastly larger than tlie applied extremities of the hairs ; so that the attempt by a human being to wrxW up ver- tical precipices covered with rolling shingle, were such precipices possible, would not in- * Loc. cit. p. 487. t IljiJ- P- '67. 400 MR. TUFFEN WEST ON THE FOOT OF THE FLY. aptly represent the condition of a fly strnggiing ineffectually to ascend a bottle strewed witli these powders. So with watery yapoiir and oil in minute quantity ; whatever the true explanation, whether by vacuum, by viscid secretion, or by mechanical means, these substances, by preventing intimate contact between the acting parts of the foot and the fixed surface of support, i. e, the glass, would interfere equally with the Fly's ascent on all three suppositions. •The belief that the presence of an adhesive secretion was requisite to explain all the observed facts led to the discovery that "spiders, and insects in the larva and imago state, when moving in a vertical direction on clean glass, leave " a " visible track behind them." This was found invariably to be the ca^, though sometimes very difficult of detection, on account of the minute qu.antity present. The truth of this assertion can be readily veri- fied, if due precautions be observed to ensure a satisfactory result. " On submitting this secretion to the direct rays of the sun in the month of July, and to brisk currents of air whose drying power was great, I ascertained," says Mr. Blackwall, "that it did not suffer any perceptible diminution by evaporation under those circumstances." The pro- bable reason why it did not, 1 shall mention in due course. To the experiment with a noxious powder (nitrate of silver), mentioned subsequently by Mr. Blackwall in another place*, the objection, similar in kind, though still stronger in effect, must Ije made as to the other experiments with powders : the moisture (which is ad- mitted to be present, although varying in its quantity at different times and under different circumstances) must cauterize the parts after a little while, and so destroy their action, of whatever kind it consist. " Now it is reasonable to infer, from the foregoing researches, that the hau'-like append- ages constituting the brushes of spiders, and occiundng in such profusion on the inferior surface of the pulvilli of insects, are tubular," is the just inference drawn from the above. Mention is also made of the means by which larvae, some apodous, others with Ijoth true and false feet {prologs), and others again with true legs fvirnished with " hair-like appendages," are enabled to ascend perpendicular glass surfaces. In the case of the larvae destitiite of legs, a viscid mucus is emitted in great abundance, respecting which, and its action per se and as an aid in producing a vacwum, I shall have some remarks to make from independent observations. The general possession of "hair-like appendages" by climbing insects was recognized, as was also the important fact that, " in structure and function, the apparatus by whose in- strumentality " spiders " accomplish their movements" " on highly polished perpendicular sm'faces " " bears the closest analogy to the pulvilli of insects, which, if named in reference to the most important office they perform, should be termed holders or supporters." A note-worthy observation this, which I can but further exemplify by going more into detail respecting the minute structure of the parts. These remarks by Mr. Blackwall possess a very high value; his experiments and observations were made with the express desire to ascertain simply the hard, dry facts of the case, " eschewing all mere conjecture, however plausible," as " the bane of Natural History :" they were frequently repeated, and the greatest care was taken to avoid every * Auuals and Magazine of Natural History, vol. sv. p. 119. MR. TUFFEN WEST ON THE FOOT OF THE FLY. 401 possible source of error. The influence tliey have had in the formation of opinion since the date of their publication is the best test possible of the esteem in which they are held. And, in addition, the merit may fairly be claimed for them, of having stimulated other inquirers to a closer examination of the questions under consideration, E. Newman (1841)* speaks, from independent observations, of the "almost infinite number of minute hairs" composing "the soft cushions ov jndvilli, on the under sm-face of the joints of the tarsus," in insects generally, which, at their ends are slightly dilated, and constantly exude an adhesive gummy matter. By means of these gummy i^HZei^Zi, " insects possessing them are enabled to walk freely on smooth surfaces, even with their backs downwards, against the power of gravity." "Until lately discovered by Mr. Lister," it is added, " the true cause " by which they are enabled to do this "was unknown." Kirby and Spence do not appear to have used the compound microscope in their investigations into the minute structures of insects. In the earlier editions of their great work they seem to have been guided in the formation of their opinions by the weight they attached to Home's authority, and subsequently by Mr. Blaekwall's latest published observations. Mr, Kirby makes some valuable observations on the action of the parts composing the Ply's feet, as deduced from the habits of the insects, both when in what we may term the natural condition of the foot, and when it is defiled by the insect's attempting to walk on a surface slightly moistened by the breath, or coated with flour, or when the same material was powdered over them. It is added " that these observations on the way in which flies are able to support themselves against gravity cannot be considered as wholly settling the question as to the precise way in which these pulvilli and those of insects generally act in eifecting a similar mode of progression ; and my main reason for here giving these slight hints is the hope of directing the attention of entomological and microscopical observers to a field, evidently as yet so imperfectly explored." Under date " July 11th, 18-12," Kirby adds an observation of considerable interest. Shortly after writing the remarks above quoted, he observed a "fly on the window, whose motions seemed very strange. I approached it," he says, "and found that it was making \dolent contortions, as though every leg was affected with St. Vitus's dance, in order to pull its pulvilli from the surface of the glass, to which they adhered so strongly that though it could drag them a little way, or sometimes by a violent effort get first one and then another detached, yet the moment they were placed on the glass again they adhered as if their under side were smeared with bird-lime. Once it succeeded in dragging off its two fore-legs, when it immediately began to rub the pulvilli against the tarsal brushes ; but on replacing them on the glass, they adhered as closely as before, and it was only by efforts almost convulsive, and which seemed to threaten to puU off its limbs from its body, that it could succeed in moving a quarter of an inch at a time. After watching it with much interest for five minutes, it at last by continued exertions got its feet released and flew away, and alighted on a curtain, on which it walked quite briskly, * Entom. Mag. Tol. i. p. 447. Grammar of Entomology, a Familiar Introduction to the History of Insects. 402 MR. TUFFEN WEST ON THE FOOT OF THE FLY. but soon again flew back to tbe window, where it had precisely the same difficulty in pulling its imlvilli from the glass as before ; but after observing it some time, and at last trying to catch it, that I might examine its feet with a lens, it seemed by a vigorous effort to regain its power, and ran quite actively on the glass, and then flying away I lost sight of it." This observation it is not easy to explain, at first sight. Kirby thought "one Blight conjectm-e that from some cause (perhaps of disease) the hairs of i]xQ pulvilli had poured out a greater quantity of this viscid material than usual, and more than the muscular strength of the fly was able to cope with*. Revolving the subject in my mind, I have come to the belief that the fly in question was struggling against its impending death from "Empusa"; some facts I have observed tend to confirm the idea, but time has not allowed of my working it out. The great knowledge of the habits of insects possessed by the author whose labours we are now considering, and of the remarkable adaptive modifications of different structures to their varying economy, enabled him to feel the necessity of always studying these in thek mutual relations. He also recognized tJie importance, for the purposes of a Natural Classification, of a thorough knowledge of the feet in insects ; and especially in the Coleoptera. His obser- vations on this head have been so thoroughly incorporated by Mr. Westwoodt (1839-10), who has largely added to them, that the student will best consult the former through the last-named author. It does not consist witb my present purpose to quote these observa- tions, which are rather for the entomologist ; but the value of the remarks on the feet in the different orders and genera, where they appear each under its proper head, is very great, and a valuable clue is thus furnished to one desirous of a guide to these parts in this numerically immense section of Animal Life. Mr. Westwood's work is not a micro- scopical one ; yet the entomological descriptions of the feet given in it are so apt and expressive, that I shall avail myself frequently of them in the course of the following pages. The possession of a special apparatus by insects having climbing habits is mentioned as having been already skilfully pointed out ; and a remark also occurs of interest from its showing a recognition of the similarity between the feet of the Diptera and of some of the Coleoptera : — " They [the Brachinides] are also characterized in many instances by having the tarsi dilated, and the penultimate joint more or less bilobed in both sexes, a peculiarity resulting, as Bonelli has observed from the situations in which they reside, this structure being in fact analogous to the cushioned feet of flies as well as of those of the true herbivorous beetles." I shall show how completely microscopic examination confirms the justice of this analogy, by the minute details of the respective structures. In Dr. Burnett's Translation of Siebold and Stannius's ' Comparative Anatomy ' (185 i), there appears in a foot-note a brief sketch of what the authors of this work considered the then state of knowledge on this subject. " Many of the Curculionidse can grapple objects by means of the immoveable hooks on * Introduction, 7th edit., p. 451. f Introductlou to the Modern Classification of Insects. MR. TUFFEN WEST ON THE FOOT OF THE FLY. 403 their tibiae. But with the Strepsiptera and Physopoda all the hooks are wanting. The Diptera, and many Hymenoptera have moreover under these hooks softl obules {arolia) provided with numerous small papillse, by meansof which these insects can fix themselves to objects." To this, the editor of the American edition adds : — " The Tenthredinidte have a lobule of this kind on each of their legs, and the Diptera have even two to three." *' According to Blackwall the papillse of the arolia secrete a viscid siibstance, which enables the Insecta having these organs to walk on smooth and steep surfaces. But this assertion requu-es further proof, though admitted by Spence." It must not be imagined that the above is a correct representation of the knowledge possessed in this country on these structures. Not one statement is quite correct. Thus — 1. The structure of the tarsi in the Curculionidse was correctly described nearly two centuries ago ; they do not grapple objects by means of the immoveable hooks on their tibiae, chiefly, but by their tarsi, in which action the former will of course sometimes bear a part, but even then only a very subordinate one. 2. The feet of the Strepsiptera and Physopoda differ so essentially that a comparison instituted between them solely on the ground of the presence or absence of claws cannot be considered a happy one. The tarsal joints of the Strepsipterous feet somewhat re- semble a series of fly's pulvilli strung together ; a dense brush of haii's covers the under- surface of the joints. The legs of the PhysojDoda are terminated by minute, naked bladders. 3. The comparison between the feet of Diptera and of Hymenoptera is equally unfor- tunate. The structure of the former had been, as I have shown, repeatedly described with a near approach to accuracy, even before Blackwall's observations ; the latter was, I believe, correctly described for the first time by myself. Tlio observations will be given in a future communication. 4. The Tenthredinidaj are Hymenoptera, with the type of foot of their Order, and with, in addition, small cushions underneath each tarsal joint ; which peculiarity, though it at once attracts the attention of any one who looks carefully at the feet of an example of this genus, is not mentioned by the editor of the work in question. About the same time a series of brief descriptions of the microscopic structure of Insects' feet was published by Dr. Inman*, who describes the feet of not less than thii-ty-one species of Articulata, mostly of the order Insecta. Allusion Avill have to be made to these remarks presently ; from the figures given, I judge that many of the descriptions were made from preparations of the parts, mounted as " objects for the microscope." These are selected for their beauty alone ; and, from the way in which they are prepared, it not unfrequently happens that they tend to convey ideas of structure and relations of a very erroneous nature. I gladly bear my testimony to the real merits of this paper, and to the benefit I have derived from the suggestive nature of many of tlie remarks in it ; at the same time an expression of regret cannot be Avithheld, that the illustrations are so unworthy of the remarks to which they are appended. * "On the Feet of Insects," by Thomas Inman, M.D., 8vo., pi. 10, Proceedings of the Liverpool Literary and Scientific Society, No. 6, 1851. VOL. XXIII. 3 I 404 MR. TUFFEN WEST ON THE FOOT OF THE FLY. Mr. Jolin Hepworth (1854) * also, nearly at the same time, and in near neighbonrliood vnih the last-named author, was repeatedly examining the feet of some large flies, and other insects, in the living state. The terminal enlargements of the " hair-like appendages" were now for the first time clearly represented : their mode of action is compared with that of the suckers of the male Dyticus. The general form of the flaps, their thickened base, and thinned-oi\t margins were described from several flies, as they appear when viewed under the compound microscope. The scale-like hairs occasionally found on their upper surface are also noted, as well as, from independent observation, the marks left on glass by the " trumpet-shaped extremities of the appendages to their flaps." A portion of the edge of the flap, which had got turned in mounting, " is figured," which shows the suckers more distinctly. The descriptions are brief to a degree which renders it difficult sometimes to understand the exact meaning of the author ; this is, however, remedied, in some degree, by the abundance and general clearness of the illustrations ; in remarking on which, we are told that "the parts are too obvious to require description" (!). Mr. Hepworth's opinions possess much value because they were foimded on independent observations, with the microscope, on the parts as they appear during life, in action and in repose, and were repeated frequently and with care. Mr. Tyrrell (1855) t suggested the use of the " hooJes on the feet of flies," as " intended not to attach the fly to anything, but to be used as fulcra or props which the fly can j)ush against when it wishes to detach the cushions." Careful investigation proves that they are used for the purpose of detaching the foot ; biit it was a very imperfect view of the matter to think that tliis was their only use. Mr. Hepv/orth (1855) J in replying to remarks, in continuation of the above, which evince a misapprehension of his views, takes occasion to go into some detail respecting his opinions. He says that each sucker is under the influence of the will, and adduces some valuable observations by Ecker, on contractile structures in their earliest manifestations, as representing exactly what he wishes to express in this matter. An objection to the viscid-secretion theory, when carried to the excess it was by some of its advocates, and which Avill readily occur to the student, is well put. "If," he says, "the foot were to be attached for some time (twenty or thirty minutes, as I have often seen it) to the same spot, it would get so firmly fixed that, if forcibly raised by the leverage of the hooks, these exceedingly delicate structures " (the " hair-like appendages ") would be destroyed. A computation is then given of the number of " suckers " on " the flap of the Blow-fly," which I believe to be correct when the number allowed for " the triangular part, extending for attachment up to the leg " has been deducted. This deduction is necessary because none of these organs are present on a part which can never be applied to any surface for the pur- pose of holding to it ; with this allowance, the number will be about 12,000 on each foot of the fly in question. The paper continues, by explaining a " diagram," of " Hairs from the pad of the foot of a small Curculio beetle ; they expand into the form of a trumpet, and where the expansion commences they appear corrugated, and the corrugation is continued * Quarterly Journal of Microscopical Science, vol. iii. t Ibid. vol. iv. p. 230. X Ibid. P- 312. MR. TUFFEN WEST ON THE FOOT OF THE FLY. 405 to their extremities; tlie expanded parts are extremely attemiated, so much so as to require a high power and obHque light to make them out. These insects (not being aquatic) also secrete a fluid for the same purpose as the Fly : and I can imagme that if, after the ends have been attached and moistened, these folds could be ptxt upon the stretch, thereby lengthening the tubes, and consequently having a tendency to produce a vacuum, they would form an excellent apparatus for attachment." There must be some mistake here; hak-like organs of this kind, loith corrugations, are (I believe, as the result of numerous observations) only present on the tarsi of those beetles which possess them for sexual purposes. Misled by preconceived ideas, although the correct term was applied to the appearance presented, a hypothesis is started to ac- count for it, of which the least that can be said is that it is quite inconsistent with the structure of the parts concerned. Gosse (1859)* describes the " strong divergent hooks " of the Fly's foot as being them- selves well clothed with spines t ; the difference in outline of the flaps in different species of flies ; their being " thin, membranous, and transparent," so that " when a strong light is reflected through them " " their structure is seen very distinctly." The appearance of " lozenge-shaped areas " on " the inferior surface of the palm " is also mentioned, and a conjecture is started as to their real nature. " From the centre of each area proceeds a very slender, soft, and flexible peUucid filament, which reaches downwards to the surface on which the fly is walking and is there slightly hooked and enlarged into a minute fleshy bulb. Those from the areas near and at the palms arch more and more outwards, so that the space covered by the bulbs of the filaments is considerably greater than that of the palm itself. Now it is evident that the bulbous extremities of these filaments are the organs of adhesion. We notice how they di-ag and hold, as the fly draws its foot from its place ;" the marks left on the glass by these " filaments " when " the foot is suddenly removed" are supposed to show "that the adhesion is effected by means of a glutinous secretion poui^ed out in minute quantities from these fleshy tips." The necessary vitiation of the results when an insect is confined in a " nearly air-tight glass cell" is observed upon with a view to caution agamst building too much upon them. As already stated, however, the experiment is easily repeated witli flies mo\ang on plane glass surfaces, where errors from this cause cannot arise. The inquiry is extended to the "similar appendages" on the "joints of the foot" of some Beetles. Timarcha tenebricosa is adduced as the first example ; unfortunately the specialization of structure of these holding appendages is not well marked on the tarsi of this insect, though the «' velvety cushion of a rusty-brown colour," is present, and very dense. The outline of the joints composing the tarsus is then mentioned; with the absence of cushioned soles from the last, which "carries two stout hooks." "The first three" are described as "flat or even hollowed beneath into soles, something like the hoof of a horse ;" an excellent comparison, wliich may be extended even to the structure itself. It is only necessary to imagine the parts cemented together by an elastic material to have * Evenings at the Microscope, pp. 131-143. t See also Inman, p. 8, who considered this appearance of being "imbricated or covered with scales " to be owing to the claw having "its formation originally by cells." 3i2 4l06 MR. TUFFEN WEST ON THE FOOT OF THE FLY, a really good idea of the nature of tlie " frog " in a horse's foot. " The whole interior " of the joints " bristles ^nith close-set minute points, the tips of Avhich terminate at the same level and form a velvety surface. Now these points are the whitish bulbous extremities exactly answerable to those on the palms of the fly, and doubtless they answer the very same purpose. Only here they are set in closer array, and are a hundred times more nu- merous, whence we may reasonably presume a higher power of adhesion to be possessed by the beetle. The structure is best seen in the male, which may be distinguished by its smaller dimensions and by its broader feet." Two errors exist here, which, on reflection, would have at once occurred to the author from whom I quote. 1. -As to the number of hairs, allowing for the different proportionate size of the part to which they are attached in the fly, and of a single tarsal joint of the beetle. According to the above assertion, this would exceed 120,000 on each joint of the latter ; the real number I have not computed, but it would not be difficult to obtain complete accuracy on this point : I believe them to be about alike on relative arese ; if anything, somewhat fewer in number on the beetle. 2. The assertion that "a higher power of adhesion" is possessed by the beetle " (than by the fly) is contrary to easily ascertained facts. The second example of a beetle's foot which is adduced, is that of the " hand " of Dyticus. The parts constituting this organ are correctly described, and the proportionate increase in number of the holding appendages, as they diminish in size, is mentioned ; this latter interesting fact proves to be nearly universally the case. Besides the observers already cited, I might refer to numerous others who have left notices respecting the feet of insects, and the mode of progression of the Fly ; but as for the most part these observations contain nothing of real importance, I shall content myself with merely citing their names for the benefit of those who may wish to refer to their writings. Ilomberg, Recueil de I'Academie des Sciences, 1710. E-eaumur, Memoires pour servir a I'Histoire des Insectes, vol. iv. p. 259, 1738. Roesel, Entomologie, 1746-61. G. Adams, senior, 1746, Essays on the Microscope (merely a paraphrase of Hooke's description). J. C. Keller, 176 i, Geschichte der gemeinen Stubenfliege; the illustrations in which are remarkably fine. G. Adams, junior, 1787, Essays on the Microscope, new edition. And last, though not least, J. Lister, 1833, in some remarks on Mr. Blackwall's observations, in which the mode in which the hairs of the pulvillus are employed, and the greasy marks left by each, are clearly pointed out. The conclusion drawn by Mr. Lister from his observations is that the pulvilli are attached by simple adhesion of the enlarged ends of the hairs, assisted by a fluid that is probably secreted there. He also clearly describes the mode of detach- ment of the foot. My desire has been, in making this enumeration of the labours of others, to do justice to all parties. It will be seen in what an interesting way each successive observer has corrected some point in the statements of his predecessors, added his omti contribution to MR. TUFFEN WEST ON THE FOOT OF THE FLY. 407 the knowledge of the subject, and yet, from one cause or another, still left the account of the structure incomplete. It was not necessary to my purpose to mention the opinions of those who merely wrote without making observations on which to found their opinions, or who, although they did make observations, have yet added nothing to our knowledge of the structures now being treated of. Having thus cleared the way, and ascertained distinctly what has been abeady done, I shall proceed to mention the results of my observations. It will have been seen that scarcely anything was left to be discovered respecting the Fly's foot, and that what was really needed was to extend the inquiry to the structure of the feet in other insects. As the additions to the knowledge of the Ply's foot have been given above in scattered notices, when mention was made of the labours of the successive observers who have con- tributed to the building of it up, it will be necessary that I should collect them now. This shall be done in the briefest manner possible ; and in order to save repetition, I shall intercalate my own observations at the same time. The foot of a Tly, then, consists of a deeply bifid, membranoiis structure (fig. 1, PI. XLL), to which the term "pulvUlus*" has been applied; anterior to the point of attachment of this part to the fifth tarsal joint, on the upper surface, are seated two claws, or " tarsal ungues," which are freely moveable in every direction, and may be closely approximated or widely separated. These ungues differ greatly in their outline, size, and relative degree of development to the tarsi, and to the bodies of the insects possessing them, and in their covering ; most are naked over their entire surface, having, however, a hexagonal network at their bases, which indicates a rudimentary condition of minute scale-like hairs, such as are common on some part of the integument of all true insects ; these decrease in size in passing forwards towards the points of the claws, on approaching which they cease some- what abruptly. Of the tmgues, the outer one is always a little the stronger; the spines terminating the tibiae, or the calcaria attached to the same part in insects where these are present, show proofs, equally with these ungues, of the universal prevalence of the law, that of two corresponding processes on any limb, the outer is invariably the more robust. To the human or comparative anatomist, examples of this remarkable fact will occur so directly, that it will be unnecessary for me to remark further respecting it, than that the difference in the relative degrees of development of the femoral trochanters furnishes one of the best instances of it that could be adduced. It may probably be connected, in some degree, with the greater risk of injury from extraneous causes to the outer, as compared with the inner aspect, of a limb. Plexor and extensor muscles are attached to both the ungues and the flaps. These flaps, corrugated or arranged on the ridge-and-furrow plan, are in some cases perfectly smooth on their superior surface, in others this surface is covered with minute scale-like hairs. The thickness of the flaps in the Blow-fly does not exceed -0002 inch at the margin ; thence they increase rapidly in thickness in passing inwards and backwards towards the point of attachment. Projecting from their inferior * The statements of authors do not appear to me to be clear, as to whether they apply this terra in the singular or plural number to this appendage to each tarsus. Examination of the parts, conducted with due care, shows, how- ever, at once that it is, as I have stated, a single, deeply-cleft organ. 408 MR. TUFFEN WEST ON THE FOOT OF THE FLY. surface are the organs which have been so frequently mentioned as "hairs," "hair-like appendages," "trumpet-shaped hairs," &c. That these are the immediate agents in holding is now admitted by almost all ; it will be convenient to term them ' " tenent " hau'S,' in allusion to their office. The delicacy of their structure in the fly ; the bend near their estremity, after which on each supervenes an elastic membranous expansion, capable of close contact with a highly polished surface, from which a very minute quantity of a clear, transparent fluid is emitted when the fly is actively moving ; all this is now admitted by the best observers. It simply remains to add that the tubular nature of the shaft of the tenent hairs on the foot of this insect has been surmised, although its minute size and homogeneity liardly permits of the surmise being actually confirmed by visual inspection. At the root of the pulvillus, on its under-surface, is a process, which in some instances is short and stout, in others long, greatly curved, and tapering to its extremity {Scatophaga), setose {Empis), plumose {HippoboscidcB), or, in one remarkable example {Epliydra), so closely resembling in its appearance the very rudimentary pulviUus mth which it is associated, that I was for some time unable to decide whether it was a third lobe of this organ, or, with the other examples named and to be more fully described hereafter, a peculiar tactile hair, which is present, in some modification or other, in all insects, so far as my present experience goes. This tactile hair has been considered by two excellent observers* to be a spring, by the help of which the Fly is enabled to detach its cushions from any surface to which tliey have been apphed ; but I shall in due course proceed to show that this opinion is erroneous. Just at the base of the fifth tarsal joint, on its under surface, there is present, in Eri- stciUs, a pair of short, very slightly ciu'ved hairs, which point almost directly downwards. These were first shown to me by E,ichard Beck ; but the discovery of their existence is so recent that time has not yet permitted of further search for analogous structures. In the instance named, they appear as if they might be little props. It will be best now to take into consideration the large appendages to the hands (fore tarsi) of the Harpalide Beetles. These are also tenent hairs, on a comparatively gigantic scale : from this very fact of their size, and our consequent ability to dissect them, and to submit their parts to varying treatment, we shall be enabled to feel confidence in the deductions which may be drawn from them, when applied to tenent hairs even of the'minutest size or the most rudimentary character. The largest of these appendages which I have been able to study on the living insect were found on the hands of a species of Fterosticlmsf (fig. 20«, PI. XLII.). Their number * Lister; Black wall. f "Latreille and Dejean have dlTided tliis subfamily Harpai,ides into tliree groups, -uiiich, from the structure of the anterior tarsi of the males (upon which they are chiefly founded), they term Quadrumani, Simplicimani, or -pedes, and Patellimani." " Audouin and Brulle have noticed (without acknowledging the observation as that of Leon Dufour," " a peculiarity in the construction of the Patellimani (Chlseniens), the males of which have the cushion-like clothing of hairs on the under surface of the foot generally distributed over the sole of the tarsus In the Quadrumani (Harpaliens), the dilated joints of the anterior male tarsi are furnished beneath with a double series of narrow cushions, which, in H. rufipes, appear to be fleshy and transversely striated, and to be destitute of pilosity. This character, although it may be sufficient to separate the Chlajniens from the Harpaliens," , does not, however, appear to be a fixed character, being liable to much variation in the Feroniens : thus, in Ahax striola, MR. TUFFEN WEST ON THE FOOT OF THE FLY. 409 here, taking all on each fore tarsus together, is IM ; they are arranged as follows : on each side of the first joint 20 ; of the second joint 16 ; and of the third joint 12. They are the ** vesicles" of authors. Those of each series appear to nie generally to differ a little in size, the larger being generally proximal. The points of attachment form elegant, very slightly curved lines, as follows : on starting from the odontoid jDrocess at the proximal end of a joint, they are near together, but rapidly diverge; so that those which arise just within the truncated or incised anterior margia of the joint are mdely separated. When viewed from the side, these large tenent hairs are seen to leave the joint from which they arise, at a very oblique angle (fig. 22 b) : they proceed with a gentle curve forwards, until, on approaching their fore end, their direction becomes changed from nearly horizontal to almost vertical ; near the end of this bend downwards, the texture, which was previously horny, begins to alter, becoming like a very firm membrane, or even tendon, at the same time increasing rapidly in width, so as from a rounded form to assume a broad fan-shape, becomiag continually thinner from above downwards. It can be distinctly seen, with due care, that the shaft of each of these modified hairs is tubular ; but at the bend, the central hoUow is imperceptibly lost to the eye, and a granular structure becomes visible, the granules having a tendency to assume a linear arrangement. At the extreme end both the texture and the direction again change sud- denly : the part is flattened out into an extremely translucent, soft, and elastic structiu'e, of a membranous texture, and narrowly reniform in its shape when seen pressed against the glass of the live-box : faint radiating lines are visible in it, and its direction becomes, when in use by the insect, once more horizontal. In P. ni[/er, the diameter of one of the largest of the membranous expansions is, in its major axis, "0032 inch, in its minor axis -0012 inch ; entire length of the tenent hair, "0145 inch. The soft membranous expansions, when not in use by the insect, point downwards; when applied to a surface of which it is desired to take hold, their direction is at once changed, and their form becomes altered through the flattening effect of the pressure. The colour of these tenent hairs is at their origin of a clear chestnut-brown, which gradually passes into straw, and at the broadest part is so faint as to be scarcely percep- tible. This is worth remarking, because it serves as an index to the relative amount of chitinous impregnation of the different parts. It is best seen when the tarsus has been made partially transparent, and will prove of importance when the action of the different parts of the tenent hair comes to be considered. It will be well now to pause a little, and see what has, up to the present stage of the inquiry, been ascertained respecting the composition of each tenent hair on the hand of a large Geodephagous Beetle. Putting all the observations together, it becomes clear that there are : — 1. A tubular shaft, of a rounded form, which cm'ves do\^iiwards very gently imtU each joint is provided with a double hairy cushion, extending beyond the extremity of the joint ; but in Broscits cephalotes the cushion of each joint is entire and nearly circular. And Burmeister figures a tarsal joint of Zahrus, with cushions transversely striated, as in Harpalus. — Westwood, Introd. vol. i. p. 84. "Two anterior tarsi of the male alone dilated" (" Harpalides ") ; "first division, Feronidea." " Four anterior tarsi dilated in the males ;" second division (Harpalidea). — Ibid. p. 85. 410 MR. TUFFEN WEST ON THE FOOT OF THE FLY. near its end, when it suddenly bends downwards. Here tlie actual cavity ceases, and the shaft assumes a mechanical texture, such as would be capable of transmitting minute quantities of fluid by exosmosis ; its form now becomes altered into that of a broad thin blade, and its direction again suddenly changed to a horizontal one, as it expands into 2. A verj thin, elastic membranous expansion, perfectly smooth on its under surface, with faint radiating lines in its substance, capable of adjustment to a smooth, flat, or slightly undulating surface, on which, when pressed, it will take firm hold. It became desirable to endeavour to ascertain how far the structure of these append- ages, which I have called teuent hairs, agrees with that of true hairs, on which some valuable critical observations were made last year by Dr. Hicks *. His remarks cannot be improved upon ; so I quote them entire. He is discussing some " peculiar modifica- tions of true hairs," "not processes of the ctiticle." "It is requisite to bear in mind the true nature of the hair in Insects — namely, that it is situated in a depression caused by the absence of the inner layers of the integument, into which the cuticle is continued. In the centre of this cuticular depression a small elevation or papilla arises, which is the true root of the hair, which rises from it of various length. By means of this arrangement, the hair itself is capable of some degree of motion. The interior of the root of each hair is in connexion with the internal parts of the member on which it is situated ; commonly fibres run to it, probably always in- eluding a branch of a nerve ; and this is decidedly the case in those hairs situated near the prominent parts and extremities of the various members, as, for instance, the tips of the antennai, the palpi, pads of tarsi, &c. ; and this branch of the nerve does not run into the interior of the hair, but only to the inner aspect of the root, which separates it entirely from the interior. The difference between this structure (true hairs) and cuti- cular processes must be particularly borne in mind : the latter, having no root, and not being situated in a dej)ression, evidently only spring from the surface. The spine must also be distinctly separated from the true hair, being a tapering process of the whole in- tegument, into the interior of which the contents of the body can freely pass f." With these clear indications for a guide, it was not difficult to ascertain, by carefully making very thin sections, that the tenent hairs agreed in their structure with the cha- racters of true hairs as defined by Dr. Hicks : sections of some of the undoubted hairs on the lateral margins of the tarsal joints, obtained at the same time, were found to agree in structure. This proved the correctness of Dr. Hicks's observations on the struc- ture of true Insect hairs ; and, at the same time, it proved also that tenent hairs must be classed in the same category. These facts arc illustrated in sections from Carahits (fig. 21, PI. XLII.). * Trans. Linn. Soc. vol. xxiii. part i. p. 143. •j- The facts of the case could not be better put ; yet it must always be borne in mind that all organized structures are formed out of the same tissue, and that the differences they present are due solely to a greater or less amount of differentiation of that tissue. On the Cricket's leg are spines which show at their bases of attachment the earliest in- dications of that kind of differentiation which is characteristic of true hairs. So also the calcaria, which are, when greatly developed, very moveable, probably possess muscles and nerves, and in some cases have a membranous sucker- like expansion at their ends — modifications of true hairs : in many other insects they are simply spinous continuations of the tibia. MR. TUFFEN WEST ON THE FOOT OF THE FLY. 411 A few words will describe all that need be said, in tbe present place, respecting some modifications of tenent hairs for sexual purposes. An example is given from Amara (fig. 22), to show that there is no essential dif- ference between these and the large tenent hairs which have just been described at length, other than as relates to their size ; in their number, 1-11 on each hand, the two entirely agree. They are arranged thus : on each side of the first tarsal joint, 2G ; of the second, 26 ; of the third, 20. Greater diameter of the membranous expansion of a large one, -002 inch; entire length of shaft, -008 inch. The next example, from the hand of the male Carabus graniilatus (fig. 23), possesses an especial interest from the features which it presents, in relation both to the size and to the number of the tenent hairs*. In size they stand about midway between large and small ; diameter of expanded part, which is orbicular in outline, '0007 inch ; extreme length of a tenent hair from this beetle, -0055 inch. Speaking of their number, it may be said that they are very numerous; I have not yet counted them. They arise from the first three joints only of the tarsus. Towards the free end, for half their extent, they are greatly corrugated, just like stiff leather, as a strap, when much bent about, becomes covered with transverse wrinkles. This corrugation is present on all tenent hairs destined for sexual purposes ; observations on the habits of the beetles possessing them satisfac- torily account for the appearance, and prove that the mode of its formation is as above supposed. When in the live-box, beetles having such appendages to their hands press them firmly against the glass cover, and then, upon these as fixed points, the body is con- tinually swayed about in every direction. And whoever has observed the males of the large Carabi attached to their mates in copula, by whom they are di'agged about in stony or gravelly places, will see that the rude shocks to which these tenent hairs must be con- stantly exposed, under such circumstances, will readily account for the corrugations described, on parts having a leathery texture. Amongst the Staphyliuidse, the " anterior tarsi " are " often dilated in the males f." The tenent hairs for sexual purposes are often very well developed ; in some species they are very numerous, in others very few in number. Where present, they are of a softer texture than those last named, so that indications of corrugation are but very slight. In Ocypiis olens (fig. 21) and Creopldlus maxillosus, the tenent hairs form a dense cushion on the under-surface of the four basal joints of the anterior tarsi. Their free ends are especially soft, so that it is difiicult to say what should be considered their real shape. When viewed in action, from the closeness with which they arc packed, they have a ten- dency to assume a form varying from orbicular, even to slightly hexagonal ; when dragged over the glass they become lengthened into oval, elliptic, and fusiform shapes. "K. Beck has noticed that the expansions shrink and almost disappear when these insects are j)laced under the influence of chloroform : this I had not myself noticed ; but the almost complete disappearance of the expansions in dead specimens is satisfactorily accounted for by it. On the upper surface of the expansions, in the two examples above named, there are some * "CARABiDiE." "Anterior tarsi greatly dilated ia the males." — Webtwood, Introd. vol. i. p. 89. •f Ibid. vol. i. p. 163. VOL. XXIII. 3 K 412 MR. TUFFEN WEST ON THE FOOT OF THE FLY. minute, scattered, scale-like hairs. Length of tenant hairs, from the anterior inferior margin of a tarsal joint of Ocijpus, '009 inch. On the hands of a very small Staphyline, I found only about 7 or 8, proportionally very large tenent hairs ; I think they arose only from the fourth tarsal joint. The specimen was unfortunately given away before careful notes or drawings were made, and I have not vet succeeded in recovering it. It appeared worth record, however, to show what inter- esting modifications may be looked for Avhen a complete examination of insects' tarsi has been made. The Staphylinidse are very interesting objects to the microscopist, in connexion with various points in their minute structure. Nowhere will better examples of the earliest modifications of hair to suit tenent purposes be found than amongst many of the smaller examples of the family. Some of these are figured (figs. 25, 26). From observations on such of these brachelytrous insects as possess tenent hairs in the most rudimentary deve- lopment of this special type of structure, where they are not connected with sexual pur- poses, we ascertain that the specialization consists essentially in a simple flattening of the hair from above downwards, and an absence of chitinous ossification towards its extremity. The male of Cicindela campestris has cushions on its anterior tarsi, the hairs compos- ing which well illustrate the above statement by the slight degree of this modification Avhicli they present (fig. 27). The suckers of the male Dytlcus furnish a remarkable example of appendages for sexual purposes *. It now became desirable to ascertain the exact structure of these. All are circular in their outline, and concavo-convex in form (fig. S3), the concavity being on the under surface ; the two proximal ones are much larger than the remainder, which are nearly alike in size : the outer of the two great ones is -^^ inch in diameter ; these two are fringed round their outer margin with remarkable branching hairs (fig. 35), which, when the apparatus is used in the water, will by preventing too close contact also prevent the water from insinuating itself so as to destroy the vacuum ; and if the female struggle out of water, by retaining liquid for some time around the sucker, they will in like manner under these altered conditions equally tend to preserve the effectual contact. There is no direct connexion between the tarsal joints and the centre of the suckers, whether of a muscular nature or otherwise, as has been asserted. The pedicles of the two large suckers are so short that the latter appear to be sessile ; these latter pedicles are of a densely membranous texture, cutting like tendon ; and my present belief is that they are quite solid, though I have not yet had time to verify this supposition. All the suckers are siipported in an extended condition by rays of a firmer consistency than the remain- * " In many species " (of the Dyticidse) " the hasal joints of the four anterior tarsi are dilated, whilst in some of the larger species the two anterior male tarsi have the three basal joints enlarged into a broad and nearly circular shield, convex above, fringed with fine hairs, and cushioned beneath, or, rather, covered with a number of minute in- verted caps [?cups] with several larger portions resembling suckers, varying in number and size in the variou,'- species . This structure enables the male to retain his situation upon the back of the female during copulation, the rugosities upon the thorax and elytra of the latter being also similarly serviceable." — Westwood, Introd. vol. i. p. 96. The remarkable increase of holding power which is obtained by the traction of these limbs to each other, and, when acting coujointl)^ to the body of the beetle, does not appear to have been specially noticed. MR. TUFFEN WEST ON THE FOOT OF THE FLY. 413 der of the membranous expansion. In the function they perform, these rays may be aptly compared to the ribs of an umbrella. In the large suckers they are very strongly marked ; in the smaller, which are only 2^ inch in diameter, they are but faintly to be traced. An interesting indication of bilateral symmetry may be observed in both the great suckers. These are freely and to a great extent moveable on their very short membranous pedicles. The mode of attachment of the small suckers to their pedicles, by means of a very narrow tendinous cord, is also such as to admit of free motion (fig. 36). It will be seen, on reflection, that much power of adaptation to varying surfaces Avill thus be gained*. The pedicles of the smaller are strong, horny columns, '005 inch in length ; •0007 inch in thickness at the centre ; they expand a little towards either end, the lower or outer being the larger. The pedicles of the small suckers were found so closely to agree with the description already quoted of the structure of true insect hairs, that it is unnecessary to describe it or do more than refer to the figures given of it and of the base of one of the fringing " guard-hairs " (figs. 37, 38). There was hence no resisting the remarkable and unlooked-for conclusion, that the suckers of the male Dyticus are extreme modifications, in one direction, of true hairs for holding piu-poses — that therefore they must be classed with all modifications of hairs, however different in external appearance, for similar purposes. It is curious to observe how a truth, when once obtained, receives support from all other truths with which it maybe connected. It occurred to me to examine, in connexion with this part of my subject, the singular hand of the male HyclropMlus (fig. 40, PI. XLIII.), which is entirely for sexual purposes — so mtich so that the insect walks on the end of the tibia alone, and drags the tarsus after it. This portion of the limb has the usual number of five joints, the last of which is alone enlarged into the form of an irregular hollow shield. On the under surface of this shield are a few true hairs, not spines, as the papillee on which they are seated and the articulation at their points of attachment to these papillse distinctly show. The longest of these hairs are proximal ; they are all pointed, and appear loose, ill-formed, and, as it were, dragged backwards. Yet this unquestion- ably represents the hand of Dyticus, — the last the most completely suctorial .organ of its kind we are acquainted with. I now come to mention briefly the subject of tenent hairs specially subservient to climbing or holding, as distinct from sexual purposes. None of these are so large as most of the appendages which have been hitherto mentioned. They are also more slender in the shafts, generally very numerous, and horny almost to the very tip, where the small membranous expansion suddenly commences. The remarkable bending downwards near the tips attains its maximum amongst the tenent hairs of this section. Clytus elongaUis, one of the Cerambycidse (fig. 41, PI. XLIII. ), furnishes a characteristic example!. After the minute descriptions which have been already given of several forms * I have seen a male Byticus swimming about with the shell of a Paludina vivipara attached to the suckers of one of its hands. •|- " The tarsi " of the Longicornes " have the three basal joints cushioned beneath, the first and second being dilatetl, the third deeply bilobed, the fourth small and nodose and inserted between the lobes of the third, and the fifth long and slender." — Westwood, Introd.vol. i. p. 356. 3 £2 414 MR. TUFFEN WEST ON THE FOOT OF THE FLY. of tencnt liairs, it will be uunecessary to do more tliau state the size. Length -0056 inch, of subtriangulav expansion -0006 inch, breadth of ditto -0004 inch. All the Longicorn beetles Avhich I have examined have the tenent hairs composing their tarsal cushions formed like the above ; I figm-e additional examples from Aromia moschata (fig. 42) and Tnonus coriariiis (fig. 43). On the tarsi of Cantliaris resicatoria (fig. 28, PL XLII.), and Ilylabris Ciclwrice (fio-. 29, PL XLII.), a similar modification of hairs for tenent purposes is met with. I figure hairs of this kind from the tarsi of Fliosphuga (fig. 44, PL XLIIL), of Tlmarcha (fig. 45), of Telephora (fig. 46), Coccinella (fig. 47), and Forficida (fig. 48). The Curculionidte, it has been already mentioned, have their tarsi well furnished with tenent hairs (fig. 49)*. In the Chrysomelida?, these parts diflTer from the above in no respect but size (fig. 50). An elegant form is presented by the tenent hairs of some beetles, which have them bent downv.-ards, almost in a vertical direction, towards their bifurcate tips (figs. 51, 52), The hairs on the hinder part of these tarsi are simply bifurcate and terete ; and the various stages of specialization, from this to the perfectly developed tenent haks, may all be traced on these appendages to a single tarsal joint. Niptus hololeucus has a small brush of tenent hairs in the neighbourhood of the claws, mostly underneath (fig. 53). If we imagine them to be cemented together, a rude idea of the Ply's foot Avould be the result. In a species of Ualtica (fig. 54) the expansion of the tenent hairs has generally, for the purpose of obtaining additional hold, two (or sometimes three) minute claws. I think this may explain Leeuwenhoek's description of the parts in the Weevil from apple-blos- soms, to which allusion was made in p. 395. The larva of Coccinella presents, on the imder surface of its robust tibiie (the tarsi are in the most rudimentary condition, as single, very indistinctly-defined joints), a moderate number of large Avell-marked, trumpet-shaped tenent hairs (fig. 56). They are most numerous on the hind pair of legs, fewest on the anterior pair. The length of the largest is -005 inch ; of the expansion -0006 inch ; width of the latter, -0004 inch. About four or five, and these the largest, arise from the rudimentary tarsus, on its upper surface, and arch gently over the titiffuis ; those which are seated near the extremity of the joints have a very peculiar abrupt ]3end, at about one-third from their extremities. These tenent hairs are distinctly moveable ; I have repeatedly watched them (especially the long ones over the claw, in which it is most clearly to be observed) bend towards the glass of the live-box, attach themsch^es by pressing their expansions upon the glass, remove, and again fix themselves in the same Avay. Por a long time I thought I must have been deceiving myself, though it is not easy to see how those which are seated on the upper surface of the limb can be made use of unless such a power of moving them be possessed by the insect. B. Beck, however, noticed similar facts, at a time Avhen I had no idea that he was worldng in the same direction. I am compelled to the belief that a minute muscle is attached to the root of every well-developed tenent hair, as well as of every ordinary true insect-hair, and * "The structure of" the "cushioned tarsi" of the Curculionidee "indicate strong adhesive rather than cursorial powers." — Westwood, Introd. vol. i. p. . MR. TUFFEN WEST ON THE FOOT OF THE FLY. 415 that the fibre whicli can be so readily demonstrated, passing to the root of each of these structures represents chiefly muscle rather than nerve. Tenent hairs are constantly liable to be injured from extraneous causes, especially in insects of cursorial habits. Wherever met with, they are invariably found to be accom- panied by hairs arranged in series along the margins of the tarsal joints to which the tenent hairs are attached. From the office they perform in the economy of the insect, these may appropriately be termed " guard-hairs." I have ascertained, by direct obser- vation, that the male Carabidce and Harpalides, in which they are remarkably stout and strong (see figs. 20 to 23, PL XLII.), walk upon the ends of these guard-hairs ; the much- worn condition in which they are found on Ground-beetles captured in the usual Avay also testifies to the above fact. During progression, the tenent haii-s touch but slightly, if at all, the general surface of the ground ; but during the powerful excitement of the male of these beetles at the time of coitus, the guard-hairs are pressed aside, and the tenent hairs brought into close contact with the surface of the female, so as to form a powerful appa- ratus for adhesion. The habits of some amongst the Carabidse, of living upon trees, of others, " amongst the branches of umbelliferous and other plants during the autumnal months, where it is not improbable that they ascend for the purpose of feeding upon the ripe seed*," point perhaps to the possession of tenent hairs by both sexes amongst beetles having these habits. Even climbing insects, as distinguished from the ground-loving beetles, have the guard- hairs well developed, though on a slighter type of structure ; and the marginal hairs without expansions on the flaps of the Ely's foot are a still more delicate form of guard- hairs. It is these latter, which are far more readily visible than the tenent hairs on the Fly's foot, which have, I suspect, led to some errors of the earlier observers of this structure ; they are fine, sharp, slightly bent downwards towards their points, and might very readily be taken for minute hooks. Notwithstanding the protection afforded by the guard-hairs, the delicate organs which they enclose are in some insects never to be found perfect. This is the case with Necro- l^liorus (fig. 57) ; and I have found a specimen of Ocyims in Avhich they were almost equally injured (fig. 24-, i.) ; in Harpalus I have less often met with this condition of the organs. I figure the tarsi of various Insects, of a Fodura, and of Acari, in which tenent hairs are present in small number, varying from one to five ; it will be seen how generally, in these cases, they are situated above the single, double, or triple claw. The remarkable scale-like organs on the anterior and middle pairs of legs and round the body of Chelymorpha testudinaria are examples of another modification (so I'ar as is yet known, unique) of tenent hairs. In addition to these, there is present at the end of each distal tarsal joint, arching over the ungues, a pair of tenent hairs, like those mcutioued in the last paragraph (fig. 64). By favour of my friend G. Hodge, I am enabled to present a figure of some appendages to the legs of a marine Acarus (fig. 68) taken in deep water off Seaham harbour, and * Westwoodj loc. cit., passim. 416 MR. TUFFEN WEST ON THE FOOT OF THE FLY. wliicli is as yet undescribed. The specimens have got a little spoilt by the meditim in which they are mounted, lout they still show a pair of organs beneath some of the finely pectinate mnch bent ungues, which represent either a pair of tenent hairs or a rudi- mentary pulvillns. This can hardly be determined without further specimens, in better condition ; but in any light the structure is interesting as being unique, so far as I am at present aware, amongst marine Articulata. Tenent hairs are so usually present in some modification or other, that it is really difficult to name a beetle which has not some form of them ; the only one I yet know, that seems to me really to possess nothing of the kind, is a species of Helops, which lives on sandy heaths ; I suppose the dense cushion of hairs on the tarsi here to be for the protection, simply, of the joints to which they are attached. I have detected them on the tarsal joints of species of Ephyclra, and on the first basal tarsal joint of the Drone of the Hive-bee, A very rudimentary form of tenent hairs is present on the under surface of some of the Tree-bugs (Pentatomidce), which have, ia addition, a large deeply cleft organ at the extremity of the tarsus, which appears to be a true sucker, and will be afterwards described. To return now to our first subject, the Ely's foot. I do not see that it is necessary to add much to the observations which have been already made, or which will have suggested themselves, on the action of the parts of which it is composed. "When a Fly is not making use of its pulvilh, as on a surface sufficiently rough to afford it foot-hold with its claws alone, these only are made use of, — remembering of course what has been already said on the greatly increased hold gained by parts mutually opposed, drawing towards each other. When walking on such a surface, the foot represents that of a Coleopterous insect without any tenent appendages. On a smooth surface, perpendicular or horizontal, the pulvilli are brought down, and the tenent hairs applied to such surface : a slight push forwards of these, succeeded by a gentle draw backwards, at each application, removes the air between their soft, elastic expansions and then- plane of motion, and thus a firm hold is gained. Access of air is prevented by the minute quantity of moisture which exudes from the expanded tips of the tenent appendages ; and thus a vacuum is formed, on the same principle as in the "Atmosperic Hat-peg," the "Plate-holder" of the Photographer, or the "Artificial Gums " of the Dentist. When the Ely wishes to move a leg from its place of attachment, the claws are brought down and pressed against the surface ; from their position they raise the hinder part of the pulvillus, where the tenent hairs are least developed, first, and so on forwards. If the claws were attached to the fifth joint, as it has been supposed, they could not act equally well in the way I have mentioned ; and I think a fly when once stuck fast, if it had no claws, might remain so*. This introduces the curious question, " What is the pulvillus of a Ely ?" The analogy between this part and the bilobed joints of some of the beetles has been noticed by previous observers t ; it will readily suggest itself on a reference to the figures given in illustration of the present paper. * See also Tyrrell and Lister, loc. cit. t Westwood ; Inmau ; imssim. MR. TUFFEN WEST ON THE FOOT OF THE FLY. • 417 But there is a yet closer relationsliip — a true liomology. The \ohedipulvillus of a Ply is the homologue of a tarsal joint ; it is a sLvth tarsal joint, which for special purposes remains membranous, unimpregnatecl with chitine. This joint is attached to the fifth tarsal joint at an obtuse angle, and then, at some part of its length (it may be a third or one-half), it is bent upon itself, so as to bring the tenent hairs, with which this part is covered, into a proper position for their taking hold of a surface when the insect is- in motion or at rest. The ungues are always attached to the last joint of an insect's tarsus. They are not attached to the fifth tarsal joint of a Dipterous insect (see PL XLI. figs. 1, 10, 16, 19) ; neither are they attached to the fifth tarsal joint of a Hymenopterous insect, but to the terminal sucker, which, again, in this great Order, is a sixth tarsal joint, membranouSj flexible, elastic in the highest degree, retractile to almost its fullest extent within the fifth tarsal joint — a joint modified to an extraordinary degree for special purposes ! I have not had time to look much about for examples amongst other Insects in support of these views ; but the plcmtula of Lucanus, with its pair of minute claws, at once occurred as a case strictly in point. The ungues are hairs modified for special purposes *. They have the structure of true hairs. The sustentacula of Epeira, the analogous structures on the entire under surface of the last tarsal joints in Pholcus (fig. 69), the condition of the parts in the hind limbs of Notonecta in both its mature and earlier condition, as well as in Sarcoptes, Fsoroptes, and some other Acari (figs. 70-72), all contribute to the proof of this fact. I believe the fluid emitted to be neither more nor less than the ordinary sudor. It is admitted that it has two components, the one of winch is watery and evaporates imme- diately, the other is greasy and remainsf. Perspiration from the tip of a finger, pressed on clean glass, behaves in precisely a similar way ; and the marks bear much resem- blance, under the microscope (allowing for the great difference in size between the two), to the corresponding marks left by a Ply in walking over glass. Nitrate of silver in weak solution, I do not doubt, though I have had no time for trying it with the requisite degree of care, would probably equally prove the presence of a chloride in each : in human perspiration we know this to be chloride of sodium ; and, reasoning from strict analogy, it is not likely to be any other salt in an Insect. The slight degree of viscidity possessed by weak animal fluids, as saliva and sudor, would satisfactorily account for another set of the phenomena presented by the fluid from a Ply's foot. The increased amount of secretion, on exertion, in both cases, hardly requires to be pointed out. That the pressure of the atmosphere is the main agent by which a Ply is enabled to adhere to perfectly smooth surfaces cannot, I conceive, after all that has been said, be doubted. Careful experiments on the weights of numerous Plies, compared with the area of their pulvilli, both of the membranous portions and of the surface covered by the organs of holding, show the following curious facts. That atmospheric pressure, if the area of the flaps be alone considered, is equal to just one half the weight of a Ply, If the area covered by the tenent hairs be added, an increase of pressure is gaiaed, equal to about one-fourth the weight of a Ply. This still leaves one-fourth to be accounted for * This was perceived by Inman, and has no doubt been noticed by other observers as well, t Blackwall, Hepworth, &c. 418 MR. TUFFEN WEST ON THE FOOT OF THE FLY, by sliglit viscidity of tlie fluid, by the action I have so often alluded to, which may be called "grasping," by molecular attraction, and, doubtless, by other agents still more subtle, with which we have at present scarcely any acquaintance. The Orders of Insects have, for the most part, each their own type of foot. Thus there is the Coleopterous type, the Hymenopterous type, the Dipterous type, the Homopte- rous type, &c. &c. : and these types are very distinctive, so that in critical instances they will sometimes serve at once to show to which Order an Insect should be referred. Thus, amongst all the Diptera, I have as yet met with but one subdivision which presents an exception to the structure already described and figured. This exception is furnished by the Tijndidce, which have the Hymenopterous foot. With hardly an exception, then, I believe the form of foot described will be found universal amongst the Diptera, and will be found amongst the members of this Order alone. One exception I thought I had met with, really goes to prove the universality of the rule. It occurs in an example of the aberrant family IIl2)pobosciclcB ; the species I have not yet determined. Here at first sight there appeared to be no p2ilviUi at all at the end of the tarsi (the typical Coleopterous foot). But more careful examination revealed two minute rudiments of a pulviUus. And the true state of the parts proved to be, that the sixth tarsal joint was present, but in a most rudimentary condition, the part for holding being cleft in twain ; these parts, membranous in their texture, apparently removed quite away from each other, and soldered to the under surface of the fifth joint, near the roots of the ungues : these claws, which are largely developed and very powerful, appear here to be articulated unmistakeably to the fifth joint : the structure of the cleft rudimentary piilvilliis at first sight presents nothing unusual in appearance ; but more careful examina- tion shows that the hairs on its under surface are pointed, and apparently not soft at their tips, and have no membranous expansions (fig. 18, PI. XLI.). In another species, belonging to the same family, the pulvillus is present, very thin, narrow, and deeply cleft, with minute tenent hairs of the Dipterous type : that the ungues have no connexion with the fifth tarsal joint is admirably shown by this specimen (fig. 19). A foot of a Sargits in my possession is interesting in more than one respect (fig. 12, PI. XLI.). Thepiihilliis is trilobate; the central lobe the largest and subtriangular, with the base distal ; the lateral lobes are also subtriangular, but with their bases proximal. From the central lobe arises a comparatively small number of tenent hairs, which are much larger than those on the lateral lobes ; they are, in fact, the largest and strongest I have as yet met with on any Dipterous foot. From the base of the fifth tarsal joint arise 11 long, overarching setae (guard-hairs), which bend downwards towards their points. The claws have scales at their bases, and are sharply pointed and abruptly curved towards their free end like Chamois-horns. The only example which has occurred to me of a foot formed on the Dipterous type, not lielonging to that natural order, was in an Accwus, of undetermined species, but which I believe to be a Leptus or Trombidium. This had an uncleft xmlvillus, with Dipterous-like tenent hairs ; the robust ungues were furnished with several long tactile hairs on their under surface, as occurs on the same part in many of the Hymenoptera (fig. 73, PI. XLIII.). MR. TUFFEN WEST ON THE FOOT OF THE FLY. 419 It may be desiral^le to add a few words on the best plan of conducting observations on these parts. Their action should be studied in living insects under the influence of cliloroform, careful notes being taken of anything which may appear noteworthy, and accurate drawings prepared from the life as well. It is of the greatest advantage to pre- serve carefully all the parts that are examined ; for this purpose Deane's medium or glycerine jelly suits exceedingly well ; some of the very delicate preparations, however, can only be kept satisfactorily in a solution of chloride of zinc. The old plan of soaking in caustic potash, crushing, washing, putting into spirits of wine, (or pressing and drying first, and) then into turpentine, and lastly into Canada balsam, is perfectly useless, except in rare instances where points connected with the structure of the integument have to be made out. Of course the parts should be viewed from above, from below, and in profile, in order to gain exact ideas of their relations. The binocular microscope, however, promises to diminish vastly the difficulties which had until quite recently to be encountered, as by its use the parts may be clearly viewed, just as they are, without preparation of any kind. EXPLANATION OE THE PLATES. Plate XLI. DIPTERA. The figures represent the feet of different flies as named : — a indicates the view of the parts on their under surface; b, the same on their upper surface; c, profile view. The foot represented is ahvays that of the anterior leg, on the left side, unless it is otherwise distinctly stated ; the sex of the insect, when certainly observed or known, is also mentioned. R is used for a leg of the proper right side ; L, of the proper left side: 1, 2, 3, for the anterior, middle, and hind legs respectively. Fig. 1. 'S\ov;-{[y {Musca vomitoria), 2 • Fig. 2. House-fly {M. domestica), $ : d, appearance of the parts when in action, from behind. Fig. 3. Baker- fly {Sarcophaga) : L 3. Fig. 4. Drone-fly [Eristalis tenax). Fig. 5. Asilus crabroniformis, $ : L 3 Fig. 6, Gad-fly [Tabanus bovinus), 2 Fig. 7- Wiagio scolopaceus. Fig. 8. Echinomyia grossa. Fig. 9. Volucella plumata. Fig. 10. Scatophaga stercoraria, Fig. 11. Syrphus balteatus, $ Fig. 12. Sargus cuprarius, 9 Fig. 13. Bibio Marci, 2 ■ Fig. 14. Borborus eqymus. Fig. 15. Piophila casei. Fig. 16. Ephydra riparia. Fig. 17- Midge {Psychuda). Fig. 18. Sheep-tick. Fig. 19. Hippobocide fly. VOL. XXIII. ; ?> T i 420 MR. TUFFEN V/EST ON THE FOOT OF THE FLY. Plate XLII. COLEOPTERA. These figures represent feet or their appendages, in whole or in part, from different beetles ; similar letters are affixed to the parts as on the preceding plate. Fig. 20. Pterostichus niger, J : «', view of some of the appendages (" vesicles ") more magnified ; a", one of these still more enlarged; a'", extremity of one, flattened by pressure against the glass; e, section, lengthwise, of joint. Fig. 21. Carabus granulatus, 6 : f, transverse section of a tarsal joint; /'', portion of the same, more enlarged. Fig. 22. Amaru communis, 6 : e", one of the appendages (tenent hairs), in profile : additional letters as with fig. 20. Fig. 23. Caradus granulatus, J : tenent hairs projecting from the third basal tarsal joint, at its anterior edge. Fig. 24. Ocypus olens: e, section lengthwise; i, portion of tarsus of another specimen, showing injury to the tenent hairs. Fig. 25. Small Staphyline beetle : hand. Fig. 26. Another small Staphyline [Tachyporus hypnorum). This has proportionately large tenent hairs on the hands ; on the middle and hind legs these appendages are much smaller ; a few bifurcate hairs are present with the latter, at the distal extremity of the third tarsal joint (y). Fig. 27. Cicindela campestris, 6 : two of the tenent hairs from the hand. Fig. 28. Cantharis vesicatoria : tenent and guard hairs at the margin of one of the tarsal joints. Fig. 29. Mylabris cichoriis: tarsal hair, in two aspects. Fig. .30. Byrrhus fasciatus. Fig. 31. An Elateride [Lacon murhms). Fig. 32. Cychramus fungicola : hand (L 1). Fig. 33. Dyticus maryinalis : dilated portion of hand (L 1). Fig. 34. „ : sucker, in two aspects, from tarsus of middle leg (L 2). Fig. 35. „ : part of a large sucker, from below. Fig. 36. „ : a small sucker, in profile and full face. Fig. 37. „ : section through one of the small suckers, at its root. Fig. 38. ,, : section through a guard-hair, at the same part. Fig. 39. Small water-beetle [Exoletus Jmniorrhoidalis) : tarsus of L 1 : of s • This supplies a most in- teresting transition-link between the suckers of the Great Dyticus and such tenent hairs as occur on the Harpalide beetles, &c. Plate XLIII. COLEOPTERA [continued), with ACARI, &c. Fig. 40. Hydrophilus piceus, S : tarsus of L 1 . Fig. 41. Clytus elongatus : tenent hairs projecting from the third tarsal joint of one of the legs. Fig. 42. Musk beetle {Aromia moschata ) : tenent hair in two positions. Fig. 43. Prionus coriarius. Fig. 44. Carrion beetle [PJwsphuga Icevigata) : portion of L 1 of j. Fig. 45. Bloody-nose beetle [limarcha loivigata) : tenent hairs projecting from the third tarsal joint of the legs. Fig. 46. Soldier-beetle [Telephoriis rusticus) : group of tenent hairs from an anterior tarsus. Fig. 47. Lady-bird {Coccinella 1 -punctata) : tarsus of L 1 : «", e", tenent hairs more enlarged. Fig. 48. Earwig {Forficula auriculuria). Fig. 49. A Curculionide [Hylobius abietis) : k, the appearance presented by the soft globular ends of tenent hairs on another Curculionide, when not pressed upon the glass. MR. TUFFEN WEST ON THE FOOT OF THE FLY. 421 Fig. 50. A Chrysomelide [Phmdon clavicornis) . Fig. 51. A Cassidide {Oinoplata normalis) : bifid teneut and ordinary hairs from one of the tarsal joints. Fig. 52. Donacia bidens. Fig. 53. Niptus hololeucus. Fig. 54. A Halticide beetle : tenent hairs on enlarged joint of hind tarsus. These hairs have in some cases two, in others three, minute claw-like appendages at their extremity, in addition to the expanded membrane. Fig, 55. Biscuit-weevil {Anobiam paniceum). Fio-. 56. Larva of Coccinella : portion of hind leg. The tenent hairs are most developed on the hind legs of these larvee, not so much on the middle, and least on the anterior legs. Fig. 57. Burying-beetle (A'iecrop^orMs) : fore tarsus. Fig. 58. Harpalide, c? : injured tenent hairs. Fig. 59. Condijlopus clavipes, Herm., portion of leg of. Fig. 60. Scale-insect {Coccus}). Fig. 61. Aphide insect taken from a Cactus. Fig. 62. Acaride larva of a Coccus from currant-bush. Fig. 63. Tetrarynchus telarius. Fio-. 64. Maple-leaf insect [Chelymorpha testudiiiariu). Fig. 65. Coccus from orange : tenent hairs on leg of larva. Fig. 66. Acaride from a small beetle : part of the leg. Fig. 67. Podura plumbea. Fig. 68. Pachytjnatha, sp. nov. : part of a leg. Fig. 69. Pholcus plialangioides : portion of tarsus of L 3. Fig. 70. Sarcoptes scabiei, fore and hind leg of. Fig. 7^. Psoroptes equi, ,, Fig. 72. Acarus from Sand-wasp {Xylocopa), fore and hind leg of. Fig. 73. Acarus, small red, found on dried cow-dung : foot in two different positions. Trans. LiKN. Soc.A^l: XXIH, Tab: 41 . \ Tifl3ii.\ftst.adnit M It ...uV Trans. Linn. SocVol.XXIII Tab. 42. TulTen-WeaUad.na.t del, et. scvilp. Trans.Linn Soc.Vol:XMI[.Tab:43. [ 423 ] XXVI. Note on the Structure of the Anther. By Daniel Oliver, F.L.S., Professor of Botany in University College, London. Read November 7th, 1861. A GERANIUM growing in my garden, apparently the Common Meadow Crane's-bill {G. prcdense), has borne during the past autumn very numerous abnormal flowers. Many of these exhibited the stamens in various degrees of retrogression, from tolerably perfect, anther-bearing, to petaloid and anantherous forms. I examined some of the flowers while they were fresh, and sketched a curious form of anther which they pre- sented, but did not until late in the season (September) undertake a more complete examination of a larger series ; and then, unfortmiately, the remaining ones were withered and decaying. However, I was still able to make several drawings, exhibiting different stages in the development of polliniferous lobes, and I now have the honour to lay some of these before the Society. I am induced to this because these imperfect anthers seem to me to afford some additional evidence upon an obscure point in the morphology of the staminal leaf, especially interesting, being derived from an order seldom presenting gra- duated series between the floral whorls, and also one in which the anthers are, normally, versatile*. I am not aware that any clear view has been as yet generally accepted as to the mor- phological import of the sutures or lines of dehiscence of the anther-cells. I may be wrong in supposing the matter to be yet imperfectly understood ; but, at any rate, the explanation of the structure of the anther, as given in the text-books on morphology which I have seen, appears to me to be unsatisfactory, and I do not recollect to have met with any evidence bearing upon the point derived from teratological facts, further than that furnished by H. v. Mohl in his important essay, " Beobachtungen tiber die Um- wandlung von Antheren in Carpelle"t, published in 1836; and by Neumann, " Ueber Antherse anticse und posticse, und deren Uebergange in einander"{. M. Gris, in 'Annales des Sciences'§, describes and figures imperfect stamens from the " Bose verte" which correspond very nearly indeed with those which I found in the Geranium ; l)ut he does not enter upon the question of their morphological bearing further than merely to point out the confirmation which they afford to the opinion entertained by botanists, * Mr. Masters, in his paper on Prolification (Linn. Trans, xxiii. 359), does not include Gerauiaceae among the orders in which median prolification has been observed. In many of the flovrers of this Geranium the axis was pro- longed beyond the three outer whorls of the flower, bearing, at a short interval, a serond flower. Tlie abnormal stamens which I describe were formed, I believe, in both the outer and inner flower. t Bot. Zeit. 1836, pp. 513, 529, 545 ; Vermisch. Schrift. p. 28, and tab. i. ; also a translation in Ann. Sc. Nat. se'r. 2. viii. p. 50. X Bot. Zeit. 1854, p. 353. § Ser. 4. ix. 76. VOL. XXIII. 8 M 424 PROF. OLIVER ON THE STRUCTURE OF THE ANTHER. that the anther results from a metamorphosis of the lamina of the leaf*. Erom the examination of a transverse section of the form of anther generally prevailing in flower- ing plants, in its young state and prior to the absorption of the vertical septa which pass from the connective through each lobe, thus rendering it four-celled, it may reasonably be supposed that the margins of these septa, answering to the longitudinal sutures and lines of dehiscence of the anther, correspond likewise with the margins of the lamina of the stamen-leaf — the septa, however, being plates of untransformed tissue of uncertain homology. The opinion that the sutures of the anther correspond to the margins of the leaf, and that the pollen-cavities, separated by the septa of unchanged tissue, originate in its parenchyma, was advanced, according to Von Mohl, by Cassinif ; and the same view has been maintained by Roeper|, Schlechtendal§, Grisebach, and others. Among botanists holding this opinion, who have recently expressed themselves definitely upon the subject, I find Prof. Grisebach, in. his " Grundriss d. syst. Botanik"||, says, "The layers of the anther in which the pollen is formed answer to the four parenchym-layers of a leaf, in which the midrib and the vascular bundles correspond, in respect to their situation, to the parenchymatous septa in the anther between the pollen-originating layers." Dr. Asa Gray, in his ' Introduction to Botany ' (1858)'[r, says, " A transverse section of the forming anther shows four places in which the transformation of the parenchyma into pollen commences, which answer to the centre of the four divisions of the parenchyma of a leaf, — viz., the two sides of the blade, distinguished into its upper and lower stratum. So that the anther is primarily and typically four-celled, — each lobe being divided by a portion of untransformed tissue, stretching from the connective to the opposite side, which corresponds to the margin of the leaf and the line of dehiscence." And further, " Viewed morphologically, therefore, the filament answers to the petiole of a leaf, the anther to the blade ; the connective represents the midrib, the lobes or cells of the anther represent the two symmetrical halves of the blade, and the line of dehiscence is normally along the margins of the transformed leaf." Dr. Lindley, in ' Introduction to Botany**,' says, "The line of dehiscence in ordinary circumstances is the margin of the modified leaf." Neumann, in the paper before referred to, upon anticous and pos- ticous anthers, expresses the opinion very decidedly that the anterior and posterior cells of each anther-lobe correspond to opposed strata of the metamorphosed leaf, and that the suture answers to its margin. Somewhat difierent from the foregoing is the theory of Bischoff ft- He maintained that the suture of the anther does not correspond to the leaf-margin ; that the loculaments develope upon its superior sm'face, within the * He says, referring to these monstrous stamens, " Cependant ces antheres, .... n'ont pas en ge'neral la structure anatomique propre a ce systeme d'organes. Aiusi le plus souvent les renflements ne presentent pas des points ou la eouclie fibreuse interrompue pourrait permettre la libre sortie du pollen. Cependant on voit dans la figure .... que la couche fibreuse tres-amincie en a [this refers to a cross-section of one of the anthers, showing a thinning at the collateral junction of the lobes] semble inviter la lobe a s'ouvrir." Somewhat similar abnormal anthers from the Rose are figured by Kaspail (Me'ra. Soc. d'Hist. Nat. Paris, iii. tab. 2, 4.5-6). t Opus. Phytol. ii. p. 551. t Enum. Euphorb. p. 4-4. § Linnaja, i. (I82C), p. 602. In " Obs. on Monstrosity of a Garden-Tulip." Ii (1854), p. 40. II Pp. 284-5. '** Ed. 4. i. p. 349. tt Lehrbuch, i. 334. PROF. OLIVER ON THE STRUCTURE OF THE ANTHER. 425 margin, upon each side of the median nervure, as is well shown, in stages intermediate between the stamens and petals, in Atragene alpina (taf. xiii. 316 B.). Upon each of the two lobes of thickened parenchyma a longitudinal suture is found, which answers to their line of dehiscence. That this line does not correspond to the leaf-margin he further endeavours to show from a monstrous flower of Colchictmi (p. 336, taf. xiii. 319). Trevi- ranus * considers Bischoff's modification required to suit Cassini's theory to nature. He concludes his observations on anther-morphology, however, thus : — " Dass jedoch dieses nicht als allgemeines Bildungsgesetz aufgestellt werden konne, zeigt z. B, Paris, wo au- genscheinlich der Pollensack durch den Band des zu einem Trager verschmalerten Blu- menblattes, und die beiden Klappen von den beiden Oberflachen desselben gebildet werden." With regard to Paris, vide infra, p. 427. H. v. Mohl's paper, referred to above, contains an excellent exposition of the state of the question of anther morphology. The cases which he especially refers to of metamorphosis of stamens into carpels — the anther- sutures becoming ovuliferous— do not, as he himself observes, afi'ord conclusive evidence in favour either of the theory of Cassini or that of Bischoff. The circumstances alluded to by Roeper, of the red colour of the leaf-margin of certain Euphorbiacese being found again in the anther-suture, and the presence of cilia on edges of leaves and on the lines of sutm-e in the anther. Von Mohl admits to be very important, if not decisive, evidence that the latter answer to the leaf-margins in some cases; while, on the other hand, transitional forms presented by the Rose, as cited by BischoflP, by the Poppy, and Nigella damascena, incontestably show that the anther-cells in these plants do not originate opposite to each other in the leaf-parenchyma on each side of the midrib of the leaf, but at distinct points on the face of the lamina— the anterior cells nearer to the median line, the posterior to the margin. In N. damascena, described by Von Mohl, the ante- rior cells are parallel to the midvein, the posterior and marginal cells having their lower extremities contiguous to them, while their upper ends are remote. H. von Mohl does not commit himself exclusively to either view of anther- structure in this essay. Mr. Bentham has kindly called my attention to the observations upon the homology of the anther-cells contained in a review of Dr. A. Gray's " Genera of the Plants of the United States"!. The reviewer considers that the anthers are homologous with the 'glands' which frequently occur at the top of the petiole, or near the base of the limb, in several genera of Dicotyledons, remote in the order of their natural afiinity. He says, " In the leaves forming the andrcecium a partial return to the system of stem-leaves takes place, inas- much as the filament is entirely reduced to the vascular system, its glands are converted into anthers, and the cellular parenchyma is only occasionally represented in an expanded connectivum, or slight membranous expansion of the filament." I have not made the minute structure of these curious appendages my study ; I have merely inspected them with a lens in Mceococca, Aleurites, and allied genera of Euphorbiacefe, and in Pithecolo- bimn. 1 must acknowledge, however, that I am unable to view the question of their homology in the same light with the reviewer, whose theory appears to rest ixpon a curious analogy in respect chiefly to the position of these glands and anther-lobes, rela- * Phys. (1. Gewachse, i. p. 277. t Ke" Misc. vol. i. (1849), p. 3j9. 3 M 2 426 PROF. OLIVER ON THE STRUCTURE OF THE ANTHER. tive respectively to the petiole or the filament. A consideration of the transitional organs between petals and stamens normally occurring in some plants, as well as in double or abnormal flowers, as, e. g., in the subject of the present paper, compels me to regard the anther as a metamorphosed lamina ; while, further, the pollen is undoubtedly an altered condition of the internal cellular tissue of the leaf, not an external production. I am ignorant of the true nature of these remarkable petiolar " glands," which deserve complete examination. A comparison of some of the above Euphorbiaceous plants suggests the possibility that they indicate, in their case, rudimentary lateral veins, and a tendency in the leaf to become peltate. The opinion that the anther-cells are a formation distinct from lamina or petiole was also held by Link*. After referring to the views of Cassini and Bischoff, he says, " Equidem hasce explicationes nimis hypotheticas et fere mecha- nicas dixerim. Ubi petalum e stamine oritur, lamina petaloidea e filamento excrescit, connecticulum dilatatur et extenuatur, anthera vero ad latus removetur, ubi diminuitur, donee pereat. Nova igitur est formatio antherse, loco folii aut potius loco laminae folii enatse, quae cum folii structura interna vix aliquid commune habet." As confirmatory of tlie theory of Bischoff, that the sutural lines of the anther do not coincide morphologically with the margin of the leaf, I consider the evidence afforded by the abnormal stamens of this Geranium. The accompanying outline sketches show transitional forms, from a petaloid lamina slightly thickened at its upper edges only (through a series bearing, besides marginal thickenings, a pair of oblong, more or less thickened masses of parenchyma at each side of the midvein of the leaf, and considerably within its edge), to tolerably perfect stamens with longitudinally lobed anther-cells. And, further, examination of these intermediate forms, I think, clearly shows thcd the sutures of the anther answer to the lines of junction of the outer and inner thickened portions of the lamina on either side of the midrib, and that the septa of " untransformed tissue" may be regarded morphologically as resulting, in part at least, from the inflected epiderms of the adjacent anther-cells. A transverse section of one of the thickened and distinct masses of these abnormal anthers exhibits no median plate of tissue dividing the pollen-parenchyma into anterior and posterior portions, as might, on Cassini and Roeper's view of anther-structure, have been expected. Although it is true that in foliage-leaves it is usual to find the cells of parenchyma towards the upper surface closely packed and often elongated perpendicularly to it, while those of the lower layers are irregular in form and traversed by intercellular passages, yet there can hardly be said to be a clear diffe- rentiation of two layers of tissue, or of an intermediate third layer separating these, to which the septum extending from the connective to the longitudinal furrows on each side of the anther might homologically be referred. On the view which the examination of these imperfect stamens would tend to lead us to adopt of anther- structure, I conceive that many of the less usual forms are more readily explicable than on that commonly received. The anthers of Lauracese, for example, often present the simple condition normally of four lobes — two superior, two inferior — approaching some of the forms figured from the Geranium. As in this order the thickened pollen-cells are not laterally approximated, there is no collateral junction or " suture " which may offer a line of dehiscence, as in the * Elementa Phil. Bot. 1837, ii. p. 185. PROF. OLIVER ON THE STRUCTURE OF THE ANTHER. 427 usual form of anther, and we find that an exceptional mode of dehiscence, by valves opening from below, characterizes the order*. With regard to Paris, referred to by Treviranus [supra, p. 425), I have examined several species (P. quadrifolia, P. hexaphylla, P. obovata, P. incompleta, and P. polyphylla-\) ; and I certainly cannot agree with the opinion expressed by him, that the valves of each anther-cell answer to upper and lower surfaces of the leaf. The dehiscence is, it is true, apparently almost quite mar- ginal, while in Trillidium G-oveniammn it is distinctly extrorse ; but here I find no difii- culty further than may be explained by a slight dilatation of the inner face of the con- nective, or by the development of the lobes answering morphologically to the inner pair of thickenings of parenchyma upon the outer, or lower, face of the metamorphosed lamina, instead of upon the inner or upper surface. In Smilax it would seem probable that the anther-cells originate from but two thickenings of parenchyma, one on each side of the midrib of the stamen-leaf. Dehiscence is along this median line, on the face of the anther {vide PI. XLIV. fig. 18). I consider, indeed, the structure of the anthers in Paris to be as easily explicable as in Geranium itself; in fact, the anthers in the latter present, from their versatile attachment and inconspicuous connective, the more specialized structure of the two. And it is from this circumstance especially that I think the evidence which they afford on the question of anther-morphology to be important, since in Papaver, Nyniph(Ba, and some other genera, in which the anther-cells have been observed to originate upon the upper surface of a petaloid lamina, the anthers are innate or adnate, and the connective sometimes quite conspicuous. And further, as Neumann {I. c, p. 358) I'e- marks, the anther- sutures of Geranium present the dark line which Roeper noted in some Euphorbiacese, and which he considered to be confirmatory of his view, that the suture corresponded to the leaf- margin. Robert Brown, in his paper on Rafflesiat, in discuss- ing the structure of anthers generally, says, in a note, " . . . . the principal point in which the antherse and ovaria agree consists in their essential parts (namely, the pollen and ovula) being produced in the margins of the modified leaf." But though the exterior, or upper, anther-lobes originate as thickenings of the parenchyma, parallel and coinci- dent, or nearly so, with the margin of the leaf, the inner and lower lobes have no such necessary marginal relation. With regard to the origin of anthers from portions of tissue morphologically corresponding to the outer lobes only, or inner lobes only, of the antlier- * It is true that the upper pair of anther-cells (in Cinnamomum at least) have their bases slightly included between the upper points of the lower cells, and that the portions of the stamen-leaf thickened, in their case, cannot precisely correspond in position with those prevailing in the abnormal forms which I figure ; but this I believe to be quite unim- portant. In Berberideee (Be?be>-is, Epimedium, Leontice, Bonyardia), besides dehiscence along the line of junction of the posterior and anterior lobes, the walls of the former separate round their base and, dorsally, along each side the connective. The inner margins of the anterior cells remain attached, as is usual. The chambered anthers of Viseum, jSfficeras, and some species of Rhi^ophora, the pore-dehiscence of several groups, distractile, appendaged, and "one-celled" anthers, the anthers of Coniferee and Cycadese, do not offer for explanation greater difficulties on this theory than on that of Cassini, Roeper, and others; indeed some, I think, are clearly more easily explicable on this view. t The amount of depression of the line of suture previous to dehiscence may vary shghtly in the species. In P. hexaphylla I have seen it scarcely perceptible, excepting towards the extremities of the lobes ; but this was in a dried specimen. In P. incom-pleta the connective is scarcely produced, as is often the case in P. polyphylla. Conf. Hooker, 111. Ilimal. Plants, pi. xxix. X Linn. Trans, xiii. 211. 428 PROF. OLIVER ON THE STRUCTURE OF THE ANTHER. leaf, and to the development of the lobes in some cases upon the under rather than the upper surface, I am not prepared at present to enter upon any discussion. It seems probable that, in these respects, there may be many modifications of the prevailing structure. The circumstance 'that the anterior pollen-cells of anthers are generally slightly lower than the posterior ones*, and that anthers are so commonly introrse, conforms entirely to the view that anther-cells morphologically correspond to thickenings of the paren- chyma of the metamorphosed leaf in a manner essentially the same as we find indicated in the abnormal Geranium. In reference to the margin of the staminal leaf, I ought to ol)serve that, when I have been able satisfactorily to trace it down the outer thickened portion of the lamina in the Geraniimi, I have found it to lose itself on the rounded outer surface of the exterior pollen-lobes, and not to become in any way coincident with the sutures ; but this observation I have been unable to repeat easily, from the withered state of the specimens which were left me to examine in September. There has appeared to be some frequent, if not constant, connection between the inner or lower anther-lobes of the abnormal stamens and the lateral veins of the lamina, as is the case, I believe, with the thickened lobes at the base of the petals in the Barberry ; but upon the relation of the anther-cells to the vascular bundles of the stamens, I have nothing to suggest. * In some plants, e. g., Loranthus europceics, the anterior cells are remarkably lower than the posterior. In this species the line of dehiscence is, in part, along the junction of the anterior and posterior lobes, in part continued above the former, along the inner face of th« connective (vide fig. 17, « to 5). EXPLANATION OF PLATE XLIV. Figs. 1 to .3. Petaloid stamens, with marginal thickenings only. Figs. 4 to 13. Various stages intermediate between the above and nearly perfect stamens. In fig. 4 but one inner lobe is formed. Figs. 14 to 16. Transverse sections of anthers similar to those of figs. 11-13. Fig. 17- Anther oi Loranthus europmis, front view. Fig. 18. Anther oi Smilax, with cross sections of same. Tram- Tmn . Sac Vol IZII.taMf. ID. OUver rJel (ZJbrmafi.. S'c . [ 429 ] XXVII. Notes on the Thysanura. By John Lubbock, F.B.S., 8fc. Part I. Smtnthurid^. Read February 6th, 1862. iHE small but interesting" group of animals to whicli Latreille gave the name of " Thy- sanoura" has, by different systematists,' been placed in very different parts of the sub- kingdom Annulosa. Without recording all the successive places which have been assigned to these little creatures, I may mention that Pabi'icius classed them among his Synistata, which correspond in great measure to our Neuroptera, — a view as to their affinities which was also adopted by Blainville. Lamarck considered them to be allied to the Arachnida rather than to the Insecta, while Cuvier placed them between the Myriapoda and Parasita. Von Siebold does not adopt the order Thysanura, but con- siders that the two families of Lepismidse and Poduridse, together with the Pediculidse and Nirmidse, form the order Aptera, which he regards as the first among the true Insects — the Myriapoda forming, in his system, part of the Crustacea. Latreille con- sidered them as a separate order of Insects, connecting these latter Avith the Myria- pods : — " Par la masse de leurs caracteres, les Thysanoiu'es appartiennent a la classe des Insectes. La composition du thorax, des organes de la locomotion, et de la bouche, I'iu- diquent suffisamment. A I'egard meme de ces dernieres parties, et surtout de I'oviducte ext^rieur du plus grand nombre de femelles, les Thysanoures ont la plus grande afflnite avec divers Orthopteres. Mais sous d'autres considerations, comme I'absence de meta- naorphoses, les organes de la vision, les appendices abdominaux et les habitudes, ils se rapprochent aussi des Myriapodes et des Arachnides. D'apres un tel melange de rap- ports, il est naturel de conclure que ces animaux font la transition des Myriapodes aux Insectes, et que vu leur plus grande ressemblance avec ceux-ci, ils doivent etre places a leur tete. Point de transformations, abdomen termine par des soies, tel est, suivant le docteur Leach, le caractere essentiel de I'ordre des Thysanoures : mais il nous semble, par son extreme concision, un peu trop vague; et aussi d'ecarter tout embarras, nous le signalerons ainsi : point de metamorphoses, ni de stigmates apparents ; corps gene- ralement reconvert de petites ecailles, avec I'abdomen termini par trois filets ou par une queue fourchue, servant a sauter." Of these four characters, only one is general to the whole group, as Smynthurus, Petrohms, and (according to Nicolet) Podura have spiracles; Smynthurus, Achorutes, Podura, Isotoma, and some other genera have no scales ; while, finally, in the Lipurida3 the spring is absent or rudimentary, so that the power of jumping is entirely lost. Finally, Cams, in his useful work, ' Bibliotheca Zoo- logica,' places the Thysanura among the Orthoptera. For the present I will ofi'er no opinion of my own as to the position and affinities 430 MR. LUBBOCK ON THE THYSANURA. of the group; but while there is so much diflference of opinion on the subject, we mi^ht reasonably have supposed that they would have been carefully studied by our naturalists. Suspected, however, of having passed the Rubicon of Entomology, and of not being true and proper Insects, they have been not only neglected, but absolutely ignored by our entomologists; so that, to my amazement, I found that (excepting a very brief notice of FetroMus maritimus, by Dr. Dickie *) Dr. Leach's original description of that species f, and Mr. Templeton's % " description of the Irish species of Thysanura," with an introduction by Mr. Westwood, are the only British contributions to the natural history of this group §. We have thus not added a single fact to the anatomical and other details given by Continental observers, and we have at this moment only sixteen Irish and actually onhj one English species recorded. When it is considered that these animals are easily preserved, are in many cases prettily coloured, are numerous in winter, when the field-naturalist is less distracted by the number of objects than at other times of the year, and, finally, that they are so common that it is impossible to pick up a hand- ful of dead leaves, or turn over an old log of wood, without disturbing many of them, I think that every one must be astonished at the neglect they have experienced ; and I hope I need make no further apology for calling the attention of the Society to this subject. The Thysanura have generally been divided into two families : — the Lepismidse, cha- racterized principally by the presence of many-jointed appendages at the posterior end of the abdomen ; and the Podurida?, which have on the under side of the abdomen a flexible spring, by means of which they can jump to a considerable distance. M. Nicolet proposes to divide the Podurida^ into three tribes,— the " Smynthurelles," in which the body is more or less globular ; the " Podurelles," in which it is linear, and composed of eight or nine segments ; and the " Lipurelles," in which the springer is absent or rudimentary. These three groups seem to me to constitute three well-marked families, which would naturally bear the names Smynthuridse, Poduridae, and Lipuridae. Por the present, however, I confine my remarks to the Smynthuridse, and must defer the consideration not only of the natural position of the Thysanura in the animal king- dom, but also of its division into families, until I shall have had time and opportunities to make a thorough examination of the whole group. SMYNTHURID^. The characters attributed by M. Nicolet to his " Smynthurelles " are as follows :— Body globular or ovoid. Thorax and abdomen forming one mass. Head vertical or inclined. Antennae of four or eight segments ; elbowed at the middle. Sixteen eyes, eight on each side, situated on a black patch, a little behind the antennae, and on the top of the head. Legs long and slender. Saltatory appendages with a supplementary segment. * Brit. Assoc. Report, 1855. f Zool. Misc. vol. iii. % Trans. Ent. See. vol. i. § Since the above was written, Mr. Westwood has kindly called my attention to Samouelle's ' Entomologist's Useful Compendium' (1819). In this volume (pp. 140, 141) five species are recorded, viz. Lepisma saccharina, Forbicina polypoda, Petrohius maritimus, Podura phmibea, and Smynthurus fuscus. MR. LUBBOCK ON THE THYSANURA. 431 It must, however, be observed that, in a new genus about to be described, the antennae cannot properly be called " elbowed," though they are often held in a bent position. The same is the case with the Smyntlmrm viridis, figured in Cuvier's ' Eegne Animal,' pi. xiii. fig. 3, which, as M. Nicolet suggests, is not a true Smtjutlmrus, but belongs probably to my new genus. The family at present contains only two genera, Smynthurus and Bicyrtoma, to which, however, I am now about to add a third. Smynthurus, Latr. Antennae four-jointed, bent at the insertion of the fourth, which is nearly as long as the other three, and appears to consist of many small segments. No conspicuous dorsal tubercles. Of this genus I have to record a new species. DicYRTOMA, Bourlet. • Antennae eight-jointed, five before, three after the bend. Two dorsal tubercles on the abdomen. Papirius, Lbk. Antennae four-jointed, without a well-marked elbow, and with a short terminal segment offering the appearance of being many -jointed. I have to describe two new species of this genus. I may, however, be permitted to remark that this is, to my mind, far from being a natural classification of the group. The great differences in the organs of respiration existing between Smyntliim'us and Fapirius appear to me to indicate a separation of more than generic value ; and we can never hope to obtain just ideas of the afiinities of different groups of insects until more attention is paid to the internal organization. On the principles which ought to guide us in the classification of the Thysanm-a I can, how- ever, only express myself in a provisional manner. Eor the present, I have only to record three English species of Smynthuridae, all of which are, I believe, as yet undescribed. Doubtless many others exist in this country ; and I am indeed myself acquainted with two or three, of which, however, I have not yet been able to obtain a sufficient number of specimens. Smynthurus Buskii, Lbk. Body globular, with scattered hairs ; abdomen with a re- entering angle. Colour brown, with a light patch in the form of a V on the posterior half of the abdomen. Apical portion of spring and end of the tarsi whitish. Basal part of antennae three-jointed, and about of the same length as the terminal ringed portion. This short description is enough to distinguish it from any of the species hitherto described. In general colour and form of body it approaches nearest to S. slgnatus ; but the antennae of that species are longer, and have the basal part consisting of three small segments between two longer ones — a character which is at once sufficient to distingmsh them. The colour also differs a little. It must, however, be observed that S. signatus of Templeton is quite a different species from S. signatus of Nicolet, as I shall presently point out more fully. VOL. XXIII. 3 N 432 MR. LUBBOCK ON THE THYSANURA. Length ^tli of an inch. Common in Kent, under logs which have been felled and left lying among long grass and underwood. I take this opportunity of expressing my thanks to my two neighbours, Messrs. Janson and Solly, who have not only allowed me free access to their woods, but have also arranged logs of wood to serve as protection to the Smynthuridse. These facilities have enabled me to obtain plenty of specimens, and have saved me a great deal of time. They feed principally on the spores and fixst shoots of Fungi. Many specimens were infested by a small mite, which adhered to the underside of the body, and was sometimes present in considerable numbers. The hairs, which cover the head as well as the body, stand at a distance of about •0047' from one another. They are gently curved, about -0075" in length, and roughened with small asperities or projections. The larger hairs on the antennse are of the same size and structure, while those on the legs are smoother. The eyes, as in all the species of the genus, are eight in number on each side. The anterior five are arranged in a quincunx, which is not quite regular, however, one of the posterior pair being a little too far back ; the other three form a triangle. This arrange- ment agrees with that figured by M. Nicolet as characteristic of the genus {I. c. pi. ii. fig. 26) ; but in the present species the eyes are nearer together than in his figure, and agree indeed even more closely with his representation of those of Cyphodeirus {I. c. pi. ii. fig. 22), especially as in that genus he makes the eyes pretty uniform in size, whereas in Smynthurus he represents the central eye of the quincunx as much smaller than the others, which is not the case in this species. The antennae are -0475" in length, of which the terminal (so-called) many-jointed portion forms nearly one-half. The basal segment is cylindrical and quite short, being •004" in breadth, and only •OOS" in length. The two following segments are of nearly equal size, being each about •012" in length and -002" in breadth. On the whole they are cylindrical ; but their outline, and especially that of the third, is somewhat knobby. They bear a few scattered bristles, resembling those on the head and body, and also a few smaller hairs near their apices. The terminal portion of the antenna resembles a necklace of beads welded together, and gradually diminishes in size at the tip, though, for the greater part of its length, it has a diameter across the beads of •0015". At each projection is a whorl of small hairs, and at the apex are a few rod-like hairs, resembling those found on the autennee of so many Insects and Crustacea, and which are doubtless organs of sensation, though I was unable, from their minute size, to ascertain their structure in a satisfactory manner. The descriptions given by different naturalists of the mouth-parts of these animals have been very dissimilar. Pabricius mentions mandibles and maxillae, labium and two pairs of palpi, consisting respectively of five and three segments. Latreille does not hesitate to characterize this description as being " absolument fictive." Assuredly it is absolutely incorrect ; but is it not more probable that Fabricius should have mistaken the species he examined, than that this great naturalist should have voluntarily committed a fault so certain of detection ? I MR. LUBBOCK ON THE THYSANURA. 433 The account given by Latreille himself is little more satisfactory. " La levre inf^rieure," he says, " se compose de deux petites lames longitudinales, paralleles, avec trois ou quatre divisions setacees, au bord superieur de chaque, et dont I'une est peut-etre un palpe. Quelques autres pieces, et qui, a en juger par la couleur brune, ou tirant sur celle de la corne, de leur extremite, sont probablement les mandibules et les machoires, remplissent les cot6s. J'ai aper9u, a chacun d'eux, un petit corps arrondi, portant une soie, et que je presume etre un palpe maxillaire. Le centre de la bouche est mou, vesiculeux, et cintre superieurement par le labre. J'ai souvent examine, avec une grande attention, la bouche de ces insectes etant encore en vie : je n'en ai vu saiUir auctme partie, et il m'a ete impossible d'en determiner, avec certitude, I'organisation. J'ai consulte anciennement sur cet objet mon ami Savigny, et je me rappelle qu'il me repondit qu'il n'avait pas ete plus heureux que moi." This description is vague as well as inaccurate, and was evidently written in haste, as Latreille could not have intended to attribute to mandibles the possession of maxillary palpi. I quote the passage, however, because, if I have myself fallen into error, I may well be excused for having misunderstood a structure which baffled Latreille and even the illustrious Savigny himself. However this may be, the mandibles and maxilla are easy enough to identify, and this has been correctly done by Nicolet and Bourlet. The latter, whose second memoir I have as yet been unfortunately unable to obtain, distinguishes (according to M. Gervais, Suites a Buffon, Apteres, vol. iii. p. 382), 1°, ITn epistome paraissant arrondi ; 2°, un labre mem- braneux, en carre long, entier et cache ; 3°, des mandibules ; 4", des machoires ; 5°, un menton ovale ; 6°, une languette large, saillante, ciliee, a deux divisions, chacune de ces divisions quadrifide ; T, des palpes maxiUaires et des palpes labiaux, mais seulement rudimentaires." This description is not altogether correct as regards the Smynthuridse, nor does M. Gervais state which species was examined by M. Bourlet, or how far the description is considered by that author as applicable to all the Poduridse. M. Nicolet's description of the mouth in the Poduridse is as follows : — " Bouche in- complete, composee d'un labre, de deux mandibules, deux machoires et d'une levre ; point de palpes." I do not quite understand what M. Nicolet means when he calls the mouth " incomplete," nor is he correct as to the absence of palpi. He has, however, been fol- lowed by most succeeding writers. (See, for instance, v. Siebold's 'Anatomy of the Invertebrata.') The true composition of the mouth-parts appears to me to be the following : — 1st, an upper lip ; 2nd, a pair of mandibles ; 3rd, a pair of maxillse ; 4th, a pair of small palpi ; 5th, a pair of organs probably homologous with the second maxillae ; 6th, a central organ, corresponding to the so-called "langue v^siculeuse" in the Lepismidaj; 7th, the lower lip. The labrum is a simple, entire, quadrate, horny lobe, somewhat broader at the base than at the apex, with a few scattered hairs, and a row of teeth at a little distance from the free border, those on each side being parallel and opposite, so that they all point inwards. The mandibles are rather small, and differ somewhat from one another (PL XLV. 3 N 2 434 MR. LUBBOCK ON THE THYSANURA. figs. 1, 2, 3, 4), though they agree in general outline. The basal part is long and some- what cylindrical, with a large gaping orifice, through which the muscles pass into the head. The wliole organ is about -03" in length ; at a distance of about -02" from the base is a raised molar surface, which is covered with minute, rather blunt teeth, and, in conjunction with the corresponding portion of the opposite mandible, evidently serves the function of mastication. The terminal incisive portion is dissimilar in the two mandibles : the one has six teeth ; the other only three, which fit into the inter- vals of the first. This part of the mandible diminishes rapidly in thickness towards the tip; the dorsal line, however, is continued throughout in the same direction, though with three gentle curves resembhng in direction, though less abrupt than, those of a bow or a reversed S. The front line of the mandible, on the contrary, makes a sudden change of direction at the beginning of the molar portion, so that the terminal part of the mandible seems to lie in a difi^erent plane from that of its basal part. The maxillse (PL XLV. figs. 5, 6) are a little shorter than the mandibles. The width varies, being smallest towards the upper end, and greatest towards the middle. The upper extremity is terminated by a somewhat complicated system of teeth. First, there are three strong curved teeth, the outer one being the largest, and the inner one the smallest. In addition to these are two lamellae, whose inner edges are beset with a row of minute, sharp projections. Lastly, there is a large central, somewhat fan-shaped lobe, the central part of which bears some fine hairs on the margin. All these parts appear to have some power of independent movement. The body of the maxillae is strengthened by a chitinous framework ; down each side runs a strong rib, while the two are connected at the centre by a transverse piece, Avhich is also continued on the inner side a little way beyond the rib, and curves at the same time so as partly to surround the softer parts. The two lateral ribs converge to meet one another at the base as at the apex. At the former, however, they enclose a space which has the form of an acute triangle, while at the basal end they curve round to meet one another. At the base the chitine is rounded off on the outer side, but is produced into a sharp point on the inner edge. The palpi (PI. XLV. fig. 8) are short, one-jointed, rounded but not tapering at the free end, and with two longish setae and one or two minute hairs, none of them, however, attached at the apex. The second pair of maxillae are membranous and delicate. Their form (PL XLV. fig. 9) is somewhat difficult to describe. At the extremity are two teeth of unequal size ; on the basal side of these are three more or less projecting lobes ; and then follow a number of small teeth, which end at a point where the two maxillae converge towards the middle line so as to touch one another. The under lip is stronger, and consists of two more or less quadrate lobes. The feet of Sniynthurus are very peculiar, and differ in the different species. Nicolet has given two figures of a foot, seen from different sides ; but neither of them agrees altogether with any of those examined by me. In my specimens the tarsus was very short, and bore two peculiar appendages. On the outer side was an elliptic, elongated claw, transparent above, terminating in a spine, and bearing also one large tooth and three smaller ones on its underside. On the underside of the tarsus is a second claw. MR. LUBBOCK ON THE THYSANURA. 435 smaller and more slender than the first (PI. XLV. fig. 11). At its free end it is pro- duced into two processes and a small spine. The shorter process reaches about as far as the free end of the other organ, and is like a spine in form, though apparently not so stiff; the second process is somewhat longer, and whip-like. The anterior legs are the shortest, and the posterior the longest, the middle pair being intermediate in size as well as in position. The relative proportions of the segments, as well as the structure of the foot, are the same in all three pairs. The tarsus is very short ; the tibia is much the longest segment. The coxa, trochanter, and femur diminish in diameter, and increase in length. At the attachment of the leg are one or two deep folds, which, however, are not, I think, to be considered as indicating true segments ; and we may therefore consider the segments as being five in number. The legs are regularly, but not very densely, clothed with hairs. Anterior abdominal appendage. — This organ, which is attached to the sternum of the anterior abdominal segment, did not escape the attention of M. Nicolet, who, however, describes it as " une petite piece blanche, saillante, multiarticulee," while in the present species, as well as in the other two observed by me, it has at most two segments, and is of a dark colour. In Smyntlmriis Buskii it consists of a large quadrate basal piece, which is somewhat widest at its proximal end, and bears at the side two small appendages, of which the one nearest to the apex is bilobed. The terminal segment is swollen at the base, and rounded at the extremity, where are four small setse. M. Nicolet considers that this organ serves to retain the spring in its place. It is generally turned backwards, and lies in such a position as to lap over the basal part of the spring, between the two lameUse, which perhaps are in part covered by the small lateral appendages ; and, by the pressure which it thus affords to the tension of the muscles, he supposes that it adds force to the spring. However this may be, it seems to me to be homologous with the anterior abdominal appendage of Lepismidse, the basal portion being, however, very much developed, and the lamella3 only represented by the small lobes at the apex. The spring is attached, according to Nicolet, to the penultimate abdominal segment ; this, however, is apparently an oversight, since in pi. 3. f. 6, he figures it on the under- side of the great antepenultimate segment. Excepting in one case (^S*. ornatus), M, Nicolet does not allude to the form of this organ as affording any specific characters ; but, as wiU be seen by comparing our PL XLV. figs. 12, 13, and PI. XLVI. fig. 24, the structure is by no means the same in all species. The organ consists, as usual in the Smynthuridse, of a basal portion and a pair of two- jointed appendages. If we take a specimen in which the basal portion is about -01" in length, the breadth will be about -014"; the first segment of the appendage will be -0225" in length, and gradually taper from a width of -0065" at the base to -003" at the apex. It bears many scattered hairs, most of which are simple and about -005" in length. On the inner margin, however, are three hairs of very different structure. They have a length of -014", and gradually taper from the base to the apex, where, however, they suddenly swell up into a delicate oval structure, something like the button at the end of 436 MR. LUBBOCK ON THE THYSANURA. a foil, or rather perhaps a drop of dew, except that the form is more elliptic. The ter- minal portion of the organ has no hairs ; it has a broad knife-like shape, and is -009" in length by -0025" in breadth. On its inner margin it has a row of minute teeth ; the outer edge is smooth. At the base of the spring are two peculiar, thickened parts of the skin (PL XLV. fig. 13). They are constant in form and position : the shaded part seems to be rather thicker than the rest ; but the whole structure is transparent, and more refractive than the skin round it. They do not serve as attachments to muscles, and I am at a loss to understand what their function can be. On each side of the anus are two great scimitar-shaped organs (PI. XLV. fig. 14), which are no doubt gigantic and specially modified hairs. They appear, indeed, to agree with the ordinary ones in their constitution and mode of attachment. Tliey are about -008" in length, and -001" in breadth. The apical half of the inner margin, and a small part of the outer edge, are roughened by irregular teeth, which, however, are so unsymmetrical as rather to suggest the idea of the border being fretted by use. The two scimitars did not even agree in the extent to which they were thus aifected. Papirius, Lbk. The characters of this genus have been given above. PAPiKirs CURSOE, n. s. Body globular, with scattered hairs ; abdomen without a re-en- tering angle. Colour dull purple. Eyes not situated in a black patch. Ends of spring pale. Antennae fom'-jointed ; terminal segment short, with about ten distinct whorls of hairs, but without actual joints. Claw with two distinct teeth on the lower margin. Length -05". PL XLVI. fig. 25, represents a form of Oregarma frequently found in this species. This species occurred, with the preceding, under logs of wood. It appeared, however, somewhat later in the autumn ; at least, I found the first specimen towards the end of October. It was common through November, but died out again towards the end of the year, reappearing in May. The structure of the antennse at once distinguishes this species from all those described by M. Nicolet. In this respect it approaches more nearly to the Smyjithurus viridis of Templeton, from which, however, its colouring is entirely different ; and Mr, Templeton does not, unfortunately, give any anatomical details. The legs are longer than in the preceding species ; the animal is altogether more lively, not indeed so fond of walking as the following species, but more so than S. Buskii, which does not very often move except by leaps. The antennse are much longer than in the preceding species, the increase being in the two middle segments. The whole organ has a length of -055", with the comparatively small thickness of -002". The basal segment is -005" in length ; the second -0225" ; the third is a little longer than the second, and especially towards its apical end, wliich is slightly swollen, and resembles a knobby stick. The terminal segment is about as long as the basal, and is somewhat conical. The whole organ is covered with scattered hairs, which are particularly numerous on the distal half of the third and on the apical segments. MR. LUBBOCK ON THE THYSANURA. 437 On the latter they are arranged in whorls, giving the organ an appearance of being jointed, which, however, is not really the case. The mandibles much resemble those of S. JBnskii ; I found, however, in my specimens six apical teeth on one mandible, and only four on the other. The maxillse also resemble those of the preceding species. The arrangement of the teeth is, however, a little different ; and they appear to be rather more numerous ; or at least the two e and/, fig. 6, have each a large secondary lobe at the base. The palpus is simple, short, -005" in length by -002" in breadth, and rounded at the apex (PI. XLV. fig. 19) ; it bears one long and two or three short hairs, not at the apex, but almost in the middle of its length, though rather nearer the free end. The organ is membranous, and in most parts transparent. The upper lip is about -00416" in width, and is quadrate in outline, with the corners rounded off. The free edge is roughened by minute teeth, which are largest towards the centre. On each side there is also a row of larger teeth, parallel with the anterior mar- gin, and with the teeth pointing inwards. The organ is also provided with scattered hairs. The organ which I have suggested to be the representative of the second pair of maxillae is represented in PL XLV. fig. 20. In my specimens it differed slightly from that of P. Saunclersii in outline, as well as in the possession of two minute teeth. Whether these differences are constant must be decided by the examination of other individuals. The under lip resembles that of P. Smmdersii (PL XLYI. fig. 30). The legs are longer in proportion, but at the same time thinner, than those of the pre- ceding species. The proportions of the different legs are the same, the anterior being the smallest, and the second pair intermediate in size between that and the third pair. In the relative sizes of the different segments also this species agrees very closely with Smynthurus Buskii — the coxa being short, and the three following segments increasing in size almost in a geometrical ratio. The whole organ, except the tarsus, is clothed with scattered hairs, which, however, are most numerous on the tibia. The tarsus itself is short and bare. The appendages of the tarsus are manifestly (PL XLY. figs. 22, 23) arranged on the same type as in S. Buskii, though the details so far differ as to offer good specific characters. The outer spine {a) is not elliptic, but resembles in outline the blade of a clasp knife, and has two strong teeth on the imder edge. The second inner appendage {b) also presents weU-marked differences. The basal portion bears a small, cvirved spine, and in the first foot (fig. 22) gradually contracts, while in those of the third pair- (fig. 23) there is a sudden diminution immediately beyond the basal spine. The tip of the filament (PL XLV. fig. 23 c) appeared to me to be slightly swollen. In this character it presents much analogy with the so-called "tenent hairs" described by Mr. Tuffen West in the twenty-third Volume of the ' Transactions of the Linnean Society ;' see, for example, his figures of these hairs on the tarsus of Ocypus olens (PL XLII. fig. 24 i), or those from Mylabris cichorice (PL XLII. fig. 29). In both those cases, however, the hairs are very numerous, and their combined action is no doubt sufficiently efficacious ; but in this case, where we have only a single hair on each foot, its action must be very slight. The spring in this species is rather longer and slenderer than that of the preceding. 438 MR. LUBBOCK ON THE THYSANURA. The basal part gradually tapers to the extremity, which is sparingly clothed with short hairs, but has no trace of the three long, button-bearing hairs which occur in Smynthurus Buskii. The terminal segment is, as usual, free from hairs ; it scarcely tapers, but is simply rounded off at the end. On the inner side are from 25 to 35 small teeth, of uniform size ; between the two last is a rounded knob, which occupies the space of about three teeth, but does not project beyond the general outline. The ventral tentacles are very similar to those of S. Buskii. They are provided with the same gland-like papillae, and fulfil doubtless the same functions. Papirius Saundeksii, n. s. Body globular, without a re-entering angle. The great ab- dominal segment is produced backwards, so that the two terminal segments are not seen when the animal is in its natiu'al position. The penultimate segment is small, and only distinguishable on the ventral side. Variety a. — Head yellowish, with a tinge of green on the posterior part, more or less distinct brownish markings on the central line, and a curved patch of the same colour behind the antennae ; eye-disk black. Back greenish, mottled with brown : the size as well as the darkness of the markings varies in different specimens ; but the fundamental pattern is generally the same, though the proportions of the parts may differ considerably. On each side of the median line on the anterior half of the abdomen is a more or less interrupted band ; about the middle of the back (that is to say, at their lower extremity) these are joined on the outer side by a diagonal band, and the two on each side unite, then separate and pass outwards and backwards, soon however returning again to the middle line, and thus enclose a more or less quadrate, or circular, space of the paler colour. In the middle line, and at the posterior end of the great abdominal segment or segments, is a marking which is very characteristic. There is first of all a small dark square, immediately behind which are three oblongs of the same colour, and at the posterior end a dark band passing backwards. The pale spaces between the oblongs vary in width, and sometimes are almost obliterated. The oblongs also vary in length, sometimes increasing gradually in length from the front one, like the steps of a pyramid, while at others the two last, or all three, are of equal length. In different specimens I found every shade between brown and dirty olive-green, while in some the light ground preponderated over the darker markings, and in others the reverse was the case. Sometimes the darker parts were of a beautiful purple. The sides of the body, and the two basal segments of the antennae, are of the dark hue, whatever that may be, with, however, a few spots of lighter colour. The two apical segments of the antennae were purple in all my specimens, which struck me as a very curious fact, since I should have expected that the antennae would have followed the law of colouring which prevailed on the other parts of the body. The posterior segment of the abdomen, the spring, the whole underside of the body, and the legs are pale, the latter with a tinge of red. Variety b. — The dark parts more extensive, and very dark brown or purple. This variety is at first sight so different from the former, that, until I compared the nature of the markings and the structure of the different organs, I supposed that it was a different species. MR. LUBBOCK ON THE THYSANURA. 439 The two varieties occur together. The body is covered by scattered hairs, which are longest on the posterior part of the back, where they have the appearance of a tuft. The terminal segment of the antenna is short, as in P. cursor. The structure of the antennse is enough to distinguish this species from any of those described by M. Nicolet ; nor can it be confused with the Smynthurus signatus of Temple- ton, which, as already mentioned, is a different species from the S. signatus of Fabricius, to which Mr. Templeton, as I consider erroneously, refers it. In many respects it approaches Mr. Templeton' s S. viridis, which, again, I must consider to be erroneously identified with the S. viridis of Fabricius ; the antennae, however, are differently constructed, and the markings are not exactly the same. E-ather in front of the antennae, and close to the middle line, is a double dark spot, which suggests the idea of ocelli. Length 05". Found in the latter part of November, December, and January, among leaves and under logs of wood, with the two preceding sj)ecies. It is not only, however, of livelier colours than either of those, but also of more active habits, running freely, and jumping more lightly and gracefully. I have given myself the pleasure of naming this pretty little species after my kind friend Mr. W. Wilson Saunders, V.P.L.S., to whom I am under many and great obligations. The antennce closely resemble those of P. cursor, the proportions of the different segments being, indeed, almost exactly the same. The upper lip also (PI. XLVI. fig. 26) much resembles that of the preceding species. The margin is clothed with short hairs ratlier than teeth, and those at the middle are no longer than the lateral ones. The mandibles have respectively six and five teeth ; in other respects they, as well as the mawillcB and maxillary palpi, much resemble those of P. cursor, as also do The second pair of maxillce, which, however, have only one minute tooth. The margin is varied somewhat in different individuals, being occasionally almost straight, and in other cases more or less lobulated. The tongue is rounded and bilobed, having a slight notch in the middle. The lower lip is formed of two somewhat quadrate lobes, the outer angles being, if we may so say, rounded off, whUe the inner apical angle of each lobe is slightly produced. The organ has several tufts of hairs. The margins have no teeth. The three pairs of legs, and the respective segments of which they are composed, agree closely with those of P. cursor ; but the foot, though formed on the same type, differs in its details, resembling S. Biiskii in some respects, and P. cursor in others. The outer claw agrees most with that of the former species, while the lower appendage follows that of the latter. Instead of the outer claw tapering to the extremity, it is rounded off, while the lower angle is produced into a small claw (PI. XLVI. figs. 31, 32). The lower appendage is rather thicker than in P. cursor, and the filament e is some- what shorter. There are, moreover, two minute spines on the posterior foot, which, however, in different specimens, differ somewhat in size and position. In some cases, indeed, the smaller spine seemed to be absent. The small appendage attached to the ventral surface of the first abdominal segment is VOL. XXIII. 3 0 440 MR. LUBBOCK ON THE THYSANURA. shaped sometliing like a pear, and attached by its larger extremity. It bears four or five short hairs. Attached to it is a second piece, also of a pyriform shape, but with the apex divided into two lobes. When the spring is turned forwards, the base ends close to this organ, exactly as described by M. Nicolet. The basal part of the spring, that is to say, the part measured from the fold which separates (if my suggestion is correct) the first from the second abdominal segment, is ■0125" in length ; the middle part is -022" in length, with a breadth of -002" at the base, and -002" at the extremity ; while the terminal segment is -0075" in length, and -00125" in width at the base. It tapers a little towards the extremity, and has from 30 to 40 small teeth on its inner margin. The distribution of the hairs resembles that of the preceding species. The haks on the posterior part of the abdomen are of two sorts, some (PI. XLVI. fig. 34) being short and sword-shaped, while others are longer, and resemble a lady's riding- whip. The males may be distinguished from the females, so far as my observations went, only by the presence of a small papilla on the ventral surface, just in front of the anus. On the Anatomy of the SmynthuridtE. Smi/utlmms JBuskii. Digestive Organs. — The intestinal canal is a straight tube, and falls into three divisions — the oesophagus, the stomach, and the rectum. The oesophagus is ratlier long, narrow, and composed, as apparently in most of the Thysamxra, of an inner chitinous membrane and an outer layer containing muscular fibres. The stomach extends almost the whole length of the thorax and abdomen. It is a capacious sac, and is lined internally by yeUow nucleated cells, resembling the so-called liver-cells of some lower Invertebrata. They are about -005" in diameter, and tolerably uniform in size. They are themselves colom-less, but contain a great number of minute yellow globules, resembUng oil-globules. The nuclei, when freed from the surrounding matter, are clear vesicles about -002" in diameter, and contain a very few minute granules. The contents of the stomach con- sisted, in my specimens, altogether of spores and young shoots of Cryptogams. The rectum is quite short, and, like the stomach, is provided with longitudinal and transverse muscles. The former, however, are not easy to distinguish. The digestive organs of the Thysanura are very uniform; at least, in addition to the genus Smynthurm, 1 have examined specimens of Isotoma, Degeeria, Lipura, Macrotoma, and Lepidocyrtus, with- out finding any great diff'erences. M. Nicolet also gives one description of tlie digestive organs as applicable to all the Poduridse (including Smynthurus), but he takes a very diff'erent view of the subject. The digestive organs of the Poduridse, according to him, are composed of five parts, " I'esophage, le jabot, le ventricule chylifere, les vaisseaux hepatiques, I'intestin grele, et le caecum." He admits, however, that " le jabot parait n'etre qu'une simple dilata- tion de I'esophage, dont le diametre varie selon que I'insecte a plus ou moins mange;" it has also no different structure from the rest of the oesophagus, and cannot therefore be compared with the true "jabot " of other insects. M. Nicolet mentions the longitudinal muscles on the stomach, and has also noticed the transverse fibres, which, however, he MR. LUBBOCK ON THE THYSANURA. 441 simply calls " rides transyersales et irreguliferes," though in the specimens examined by me they were more distinct and regular than those of the other series. I found no posterior division of the stomach sufficiently well marked to justify M. Nicolet's name of "intestine grele;" and the rectum, which M. Nicolet describes as being " parfaitement lisse," was in my specimens distinctly muscular. He calls it "csecum," which, however, is no doubt a mere slip of the pen. The most essential difference between us, however, is as to the Malpighian vessels, of which M. Nicolet thinks that there are six, " tubuleux et filiformes ou du meme diametre dans toute leur longueur ; ils sont inseres immediatement au-dessus du retrecissement pylorique ; leur longueur egale a peine la moitie de celle du ventricule chylif ere ; du reste, leur extreme t^nuite et leur pen de consistance ne m'ont pas permis de les etudier en detail." I can only say that neither in Smyntlmrus nor in any other genus of Podm'idse which I have examined, have I been able to detect any trace of these organs. The fatty tissue consists of lobular, opaque, white masses, which owe their appearance to the presence of innumerable granules. These vary from a very minute size up to g o\) o^h of an inch, and are probably excretory products. A large band of it lies on each side of the intestinal canal, with which it has numerous connexions, though I could not convince myseK that there was any open communication between the two. Respiratory Organs. — Smyntlmrus breathes by means of tracheae. There are, however, only two large spiracles ; at least I was unable to see any others ; and all the larger tracheae, even some in the posterior part of the body, were traceable up to one or other of these. It is very unusual for an articulate animal to have only two spiracles; and then- position is stUl more extraordinary, for they open on the underside of the head, imme- diately below the antennte. No insect is known to have spiracles either in the head or between this and the prothorax ; and Btu'meister (Handbook of Entomology, p. 165) even goes so far as to state that none are ever so situated. It appears, however (see Pagen- stecher, Beitrage zur Anatomic der Milben), that, though in most Mites the tracheal ori- fices are situated at the base of the legs, yet in Trombidkim holosericetim the spiracles are two in number, and, as in Smynthti?'ti,s, are situated at the lower side of the head, though not exactly iu the same place, since in this species they open on the inner side of the basis of the mandibles. In the Thysanura, according to Nicolet *, " Les ouvertures trachedennes ou stigmates, dont je n'ai pu decouvrir que huit, sont placees par paires sur les arceaux superieurs des quatre premiers segments de I'abdomen. La couleur de leur p&'itreme, qui est la meme que celle du corps de I'insecte, lesrendtres-difficUesaapercevoir; leur forme est lunulaire ; ils occupent le milieu de chaque bord lateral des segments ci-dessus mentionnes, mais a une distance de ce bord egale au septieme environ du diametre transversal de I'insecte." M. Nicolet apparently gives this description as applicable to aU the Podurelles (a term ia which he includes the present genus) ; and it has been so tmderstood by succeeding writers (see, for instance, Gervais's ' Insectes Apteres,' vol. iii. p. 385, and Von Siebold's ' Anatomy of the Invertebrata,' translated by Burnett, p. 438). It is, hoAvever, altogether incorrect as regards SmyntJiurus ; nor does it apply to many of the true Po- * Nouv. Mem. de la Soc. Helvetique, vol. vi. p. 38. 3o2 442 MR. LUBBOCK ON THE THYSANURA. duridse, among which I have examined species belonging to most of the genera, though I have not yet been able to obtain any specimens of the true Podura, to which genus the species examined by Nicolet belonged. Close to the spiracles, the tracheae break up into a great number of thin branches, which supply the head without much more subdivision. There is also a very large trunk, which almost immediately divides into two branches, the smaller one of which soon divides again, and supplies the anterior region of the thorax, while the other gives off branches to the posterior legs and the abdominal organs. In the manner of subdivision, the trachese of Smynthurus differ from those of the true Insects, and agree more closely with the Myriapoda and tracheal Arachnida, in the fact that they do not often give off branches nor form tufts, but generally divide dichotomously, and run considerable distances without a separation. I have noticed no respiratory movements ; and the supply of oxygen is probably due therefore principally to that diffusive power of gases, the laws of which have been so well worked out by Dr. Graham, and even applied to the respiration of insects. " In the law of diffusion of gases," he says, " we have therefore a singular provision for the full and permanent inflation of the ultimate air-cells of the lungs. But it is in the respiration of insects that the operation of the law will be most distinctly perceived. The minute air-tubes accompanying the blood-vessels to every organ, and, like them, ramifying tUl they cease to be visible under the most powerful microscope, are kept distended during the most lively movements of the little animals, and the necessary gaseous circulation maintained, wholly, we may presume, by the agency of diffusion." Though we must attribute some influence to the respiratory movements exhibited by so many insects, the above explanation seems to me to throw much light on the question, which I have already treated at greater length in the ' Linnean Transactions ' for 1860. I should not have thought it necessary to allude again to the subject, but that Prof. Rathke, in a posthumous memoir " On the Respiratory Process in Insects" (See Ann. and Mag. of Nat. Hist. 3 ser. vol. ix. p. 105), appears to have overlooked these facts, and thereby to have fallen into some errors. Thus he says, " Prom the absence of all such phe- nomena, we might conclude that in the pupse of the above-mentioned insects (Coleoptera and Hymenoptera) the tracheary respiration is entirely interrupted." And further on, " In any case, it is certain that the respiration of pupse can only be very weak." It has, I think, been sufficiently shown that the mere absence of respiratory movements does not necessarily involve such a conclusion. While, however, in Smynthurus Buskii the mere presence of tracheae is easily detected, difficult as it may be to ascertain their distribution, I have, to my great astonishment, been unable to detect a trace of them in the genus Papirius. Remembering that though the great Treviranus was unable to convince himself of the existence of tracheae in Lepisma, they have since been discovered by Burmeister, and being only too well aware of the difficulties attending the dissection of these minute animals, I long attributed the ap- parent absence of tracheae to my own unskilfulness ; but this explanation is not, I think, tenable ; and even if rudimentary tracheae l)e hereafter discovered, I feel at least convinced that their arrangement and distribution will be found to differ altogether from those which characterize Smynthurus. It must be remembered that the air in the tracheae of MR. LUBBOCK ON THE THYSANURA. 443 freshly killed insects glitters like threads of quicksilver ; and however absurd it may sound, I consider the inside of an insect, with its beautiful and rich tracery of glittering tubules, to be one of the loveliest objects in nature. In searching, therefore, for tracheae we are very much assisted by the presence of the air ; and while in preserved specimens these organs are not easy to trace, at least towards their extremities, in fresh specimens they are generally very conspicuous. I have not yet discovered in Papirms any other organs of respiration ; but, considering how soon these animals dry up if placed in a dry atmosphere, it seems quite possible that they may respire through the general surface of the skin. Female Generative Organs. — The ovaries of Smyntlmrus are very simple. From the delicacy of their membranous envelope, it is not very easy to extract them entire ; but the best way is to make a section of the animal from the back at the anterior part of the cephalothorax, to the base of the spring : in this case, the ovaries will generally be found uninjured, and may be separated from the other organs without very much diffi- culty. They consist of a single, short and broad egg-tube on each side, and unite posteriorly to form a narrow vagina about -003125" in length. Having found it impossible to remove the skin in such a manner as to expose the ovaries in situ, and as the method above described destroys, of course, the natural position of the organs, I am unable to say much as to the usual position of the ovaries. They did not appear, however, to pass directly forwards, but seemed rather to curve round from the point of attach- ment to the vagina so as to lie rather in the posterior part of the body. They were, I think, on the dorsal side of the intestinal canal ; but this point I unfortunately forgot to note at the time. M. Nicolet was unsuccessful in his search for the organs of generation. " Je I'ai deja dit," he says, " les difficultes que I'estreme petitesse de ces insectes oppose a la dissection, rendant fort difficile I'etude de leur organisation interiem-e, qui peut-etre restera encore longtemps inconnue; le hasard seul, en rendant un observateur temoin d'un accouple- ment, pourra faire decouvrir leurs organes genitalis, mais un pareil hasard est difficile a prevoir." I quote this sentence as an excuse for the numerous details, in the anatomy of SmyntJmriis, which I have left unascertained ; many of which, however, want of time alone has prevented me from making out. I may add that the small size of Smynthurus gave me less difficulty than its curious form and the extreme delicacy of its internal organs. By opening the animal, however, as above mentioned, I was able several times to make out the vagina above described, and also to trace it to the vulva, which, with the anus, opens into a common cloaca at the posterior end of the body; this, however, from the great development of the segment bearing the spring, is thrown apparently on to the dorsal surface. The only accessory organs which I could find were two small, glandular, rounded bodies attached to the vagina close to its orifice. They were -0055" in width, and shaped like a kidney, with the convex side in front. On the other side of each, that is to say near the vulva, is a dark-brown rounded mass, about -003" in diameter. Beyond their relative position, I did not make out the relation which these organs bore to one another or to the vagina. 444 MR. LUBBOCK ON THE THYSANURA. My examination of Smynthurus began early in October. At this period tbe ovaries generally contained perfectly clear jelly-like nuclei, of which the largest were about "OOl" in diameter, but many were much smaller ; these nuclei were surrounded by a clear, transparent, but less brilliant substance, which sometimes formed a complete spherical envelope to a single nucleus, but generally appeared as a lobulated mass enclosing two or more. In other cases, which I suppose were more advanced, the nuclei were darkened by a deposit of granules. Even, however, early in October the development of the eggs was sometimes much more forward, aiid the ovary contained a number of yelk- masses at pretty regular intervals. In a transverse section of the ovary there were three, or at the broadest part four rows. The spaces between the yelk-masses were occupied, as before, by the nuclei and their transparent enveloping substance. The yelk-masses having at this period no bounding membrane, when I attempted to isolate them by open- ing the ovary, they immediately fell to pieces ; which made the study of theu* develop- ment somewhat difficult. On the 13th October I found a specimen with eggs apparently ready to be laid ; they were yellow, round bodies, -01" in diameter, with a single struc- tureless ovarian membrane, and the yelk was composed of oil-globules of various sizes up to -0015". I could iind no trace of Purkinjean vesicle. On the 22nd I exa- mined another female in a similar condition, and found the eggs to be about forty in number. Male Generative Organs. — SmyntMrm Buskii, as I have already observed, became much rarer in the latter part of November, and disappeared altogether early in December. Up to this time I had been occupied in the determination of species and the examination of the other parts ; so that my description of the male organs of reproduction must be taken from Fapirms Saundersii, which made its appearance later in the autumn, and lived longer through the winter. There is, however, no doubt that, if my specimens had lasted, I should have been able to lay before the Society a description of these organs in Smynthurus Buskii, which, from its greater size, was the species most favourable for dissection. In Fapirius Saundersii, as in many other apterous Articulata, the testis is formed on the same type as the ovary. On each side of the body is a simple tube (PI. XLVI. fig. 38 a), opening into a triangular reservoir with its base in front. The two posterior sides of the triangle are accompanied by a glandular accessory organ, containing a central cavity (PI. XLVI. fig. 38 c), full of minute globules, the largest of which do not exceed •000083" in diameter, while the majority were much smaller. Prom the posterior angle of the reservoir a short and narrow vas deferens runs into the ductus ejaculatorius, which is also short pyriform, and opens, like the vulva, with the anus at the posterior end of the body. The development of the spermatozoa proceeds as in Obisium (Phil. Trans. 1861). The testis contains small cells, about -00025" in length, and collected into masses of various sizes, which have probably arisen within mother-cells ; their size does not appear to depend in any way on the stage of development of the spermatozoa. The small cells which compose them are at first somewhat oval ; gradually one end elongates, and at length they resemble in form the egg of a Cynips, consisting of a thread, with MR. LUBBOCK ON THE THYSANURA. 445 a small swelling at one end, and a larger oval body at the other (PL XLVI. fig. 39) ; this, however, is not probably the mature form. The testis also contained numerous greenish, oval or somewhat quadrate bodies (PI. XLVI. fig. 41), about -0002" in length, and exactly resembling the similar bodies which I fijund in the testes of Chelifer and Obisimn, and which so curiously resemble the spermatozoa of Tolydesmus. There were also a number of larger spherical cells (fig. 42), from -0005" to -001" in diameter, and transparent or containing only a few granules. These may perhaps be compared with the similar bodies found in the testis of Oiiiscus. The walls of the triangular reservoir are rather thick, and consist of nucleated cells (PI. XLVI. fig. 38). It contains the spermatozoa and the green bright bodies, as well as some other globules, which resemble oil-globules. The walls of the accessory glands are also composed of thick cells. They are com- pletely filled by the minute granules, which form an opaque mass. The ductus ejacu- latorius has an apparently chitinous lining, which is strengthened, like the tracheae and some other chitinous parts of Articulata, by transverse ribs, which, in this case, are very delicate. The so-called "gastric tube," and the two filaments which proceed from it, are not the least curious part of the anatomy of Smyutliurus. The gastric or, as it may better be called, the ventral tube is situated on the underside of the body, in the middle line, and immediately behind the second pair of legs. From it the animal can project two long, colourless, transparent filaments, which, when extended, are nearly i^th of an inch in length, are somewhat elbowed in the middle, and at the end are covered with small vesicles or " glands," which, according to IS'icolet, secrete abundantly a mucous fluid, by which the whole organ is " continuellement invisque," but which I have not observed on the specimens examined by me. Nicolet gives a figure (pi. 3. fig. 22) of a tentacle, which does not altogether agree with my observations ; but, as he does not mention the species from which it was taken, it is impossible to say that the difi'erences may not be partly specific, though I am not disposed to think that they can all be accounted for in this manner. M. Nicolet observes that the tentacles are " doues d'un mouvement retractile exactement pareil a celui des tentacules oculaires des Limaces ;" but he does not figure nor describe the very curious muscles by which the apparatus is worked, nor has the subject been investigated by any other naturalist. In PI. XLVI. fig. 35, 1 give a figure of the whole organ in a state of retraction. At the point a, the basal portions of the two tentacles, b, b, are attached to the wall of the ventral tube ; the tentacle itself is, of course, in this position turned inside out, the outer wall, b, being the inner wall in fig. 36, and the small vesicles seen on the outer wall in that figure being here in the central part of the hollow tube. All the letters in this figure represent the same parts as in fig. 35. At the part / is attached the muscle e, which in fig. 35 lies inside the tube, while here it is necessarily external to it. The point i is the terminal part of the filament when extended ; the numerous vesicles can be seen, in an inverted position, through the transparent walls, and the muscle h, attached in fig. 34 to the inner side of the extreme end, is turned out in the same manner as the muscle i. The point/ 446 MR. LUBBOCK ON THE THYSANURA. as may be seen in fig. 36, very nearly divides the tentacle into two equal parts, though the basal portion (that is to say, the part /, b, a) is slightly the longer of the two ; although, therefore, the part i, c,fis always found bent nearly into a semicircle, still the tentacular ends of the two muscles e and h are kept about g^th of an inch apart, while their opposite ends are attached, close together, to the wall of the back, immediately above the ventral tube — a position which adds much to the difficulty of dissecting out the whole organ without injuring any of the parts. The terminal muscle h is single and cylindrical, while the other one is divided nearly in the middle into two equal branches (fig. 36, e, g). Although the arrangement and attachments of these different parts are so peculiar and intricate, a comparison of the two figures (Pl.XLVI. figs. 35, 36) will, I think, satisfactorily illustrate the mode in which the organ is extended, directed, and retracted. The whole skin of the animal is kept in a state of considerable tension by the quantity of its contents ; so that if an incision is made, part of the entrails are immediately forced out. If, therefore, we imagine the muscle e to relax, the tension of the skin acting on the place at which the ventral tube is folded inwards — in fact, on the tube b, at the end a— will begin to force the tentacle inside out, so that what was the inner wall will become the outer one, and vice versa, and the muscle e, as it elongates, from being external to the tube, will gradually liecome internal. The first effect of this upon the other section of the tentacle, c, will be to straighten it, but gradually the muscle h also must relax ; the continuance of the same pressure will invert also the tube c, and, the muscle li being carried into it in the same manner as has already happened to e, we shall at last have the state represented in fig. 35, — the whole tube being inverted from its former position, the lower half contain- ing the two muscles, the upper half only the one marked //. The protrusion of the organ, however, though probably much assisted by the tension of the skin, is not to be altogether attributed to it, as the muscles are apparently surrounded by a layer of transverse fibres, which must tend to produce the same effect. I was at fixst inclined to wonder at the presence of two muscles, and to suppose that the object might have been equally well attained by the presence of only the one marked h. The two muscles, however, acting on different parts, may enable the organ to be retracted more quickly than might otherwise be the case — an advantage evidently of much importance to a part so delicate. Whatever also may be the function of the tentacles, a muscle attached to the middle part will enable the animal to apply this portion to some parts of its body which perhaps could not otherwise be so conveniently reached. Again, it is evident that a single muscle attached to the part d could not have fully retracted the tube, because the distance between d and a is less than the length of the tentacle; but this difficulty is at once avoided by the presence of a second muscle, which throws the tentacle into a curve; moreover, greatly as the two muscles are extended when the organ is protruded to its full length, if there was only one it would require a still greater power of elongation. One function at least gi the ventral tube is easily visible. If a Smynthimis is laid on its back — a position from which it has some difficulty in recovering its feet, — and if, while it is in this attitude, a piece of glass is brought within its reach, the animal will endeavour to seize it with the feet, but at the same time it will project one or both of the ventral MR. LUBBOCK ON THE THYSANURA. 447 tentacles, and apply it or tliem firmly to the glass, emitting at the same time a drop of fluid, which no doubt gives a better hold. I have, however, known a Smynthuriis seize hold with one of its tentacles of the needle with which I was tui-ning it over on its back ; and the adherence was much firmer than the mere presence of a little fluid would account for. In the parallel case of the Poduridse, M. I'Abbe Bourlet supposes that the ventral tube acts as follows : — " 1°. Qu'il sert a ces insectes a se maintenir sur les surfaces perpendi- culaires en y faisant le vide ; 2°. Que le liquide excreta par lui sert a humecter la queue et la rainure ; 3°. Qu'il supplee a la faiblesse des pattes dans les chutes qui suivent les sauts." I am therefore disposed to agree with him in so far as he denies that the adhesive power depends altogether on the viscous fluid ; but, on the other hand, I cannot attach much importance to his two latter suggestions. De Geer well understood the use of this curious organ. He says, " quand la Podure (under which name he includes the present genus Smyntlmrus) marchait contre les parois du poudrier, il lui arrivait souvent de glisser ; c'etait comme si les pieds lui man- quaient, de fagon qu'elle etait sur le point de tomber ; dans I'instant meme, les deux filets parurent et furent lances avec rapidite hors de leur etui, s'attachant dans le moment au verre par la matiere gluante dont ils sont enduits, en sorte qu'alors la Podure se trouvait comme suspendue a ces deux filets." Nicolet gives a similar explanation of their function, and, like De Geer, attributes the adhesiveness to the glutinous matter which they secrete. Upon this point, however, I feel compelled, as already mentioned, partly to disagree with these two high authorities. The Nervous System. — M. Nicolet describes the nervous system of Smynthurus as con- sisting of four ganglia, with a double connecting-cord. Two of these ganglia occupy the head, and form the oesophageal collar. The 1st suboesophageal ganglion (Nicolet, I. c. pi. 4. fig. 1 c) gives off two small nerves on each side. The 2nd, which is situated nearly midway between the 1st and 3rd, has a third nerve on each side ; and the posterior ganglion, besides three pairs of nerves corresponding apparently to those of the pre- ceding ganglion, has at its hinder extremity three other much larger nerves, which supply the posterior part of the body. I found much difficulty in making out the nervous system at aU, but hoped that Dr. Hicks' s bleaching process might facilitate the matter. In this, however, I was disappointed, nor was Dr. Hicks himself more successful : he writes me word, " I have been as unsuc- cessful as you in the bleaching process. The thickness of the body, and the want of thickness of the chitine layer allowing shrivelKng, are obstacles which I am afraid will not be overcome. It does well for the legs, wings, and antermse ; but for the thick bodies it seldom, I thinlc, will suit, unless the integument is very firm." Dr. Hicks, however, satisfied himself that there were " three thoracic ganglia, with a posterior cord extending into the abdomen in the median line." Since that time I have been unable to obtain any specimens of S. Buskii, which, being a large species, is a favourable subject for examination. I had, however, already satisfied myself of the presence of two large ganglia in the anterior part of the body ; but the nerves proceeding from them seemed to me to be larger than those figured by Nicolet. VOL. XXIII. 3 P 448 MR. LUBBOCK ON THE THYSANURA. EXPLANATION OP THE PLATES. Fig. 1. Smynthurus Buskii. •^* 33 35 4. •'• ■) 33 6. 7. 8- 9. „ „ 10. „ „ Il- 13. 14. 15. Papirius cm 16. Fig. Fig. Fig. Fig. Fig. Fig. Fig. Fig- Fig. Fig. Fig- Fig. Fig. Fig. Fig. Fig. Fig- Fig. Fig. Fig. Fig- Fig. Fig. Fig. Fig. Fig. Fig. Fig. Fig. Fig. Fig. Fig. Fig. Fig. Plate XLV. Mandible, x 60. Tip of the same, x 250. Mandible of other side, x 60. Tip of the same, x 250. Maxilla, x 60. Tip of the same, x 250. Labrum, x 250. Palpus, X 250. Second maxilla, x 250. Leg of third pair, x 60. Foot, X 250. Spring, X 60. Part of the same, x 250. Anal appendages, x 250. Antenna, x 60. Tip of the same, x 250. 17- „ « Upper lip, X 250. 18. „ „ Tips of the two mandibles, X 250. 19. „ „ Palpus, X 250. 20. „ „ Maxilla of the second pair, x 250. 21. „ „ Leg of the thu'd pair, x 60. 22. „ „ Foot of the first leg, X 250. 2.3. „ „ Foot of the third leg, X 250. Plate XLVI. 24. „ „ Spring, x 60. 25. Gregarina oi P. cursor. 26. Papirius Saundersii. Upper lip, x 250. 27. „ „ Tip of first maxilla, X 250. 28. „ „ Tip of second maxilla, x 250. 29. „ „ Palpus, X 250. 30. „ „ Lower lip, x 250. 31. „ „ Foot of anterior leg, x 250. 32. „ „ Foot of posterior leg, x 250. 33. „ „ Anterior abdominal appendage, x 250. 34. „ „ Anal hairs, x 250. 35. Ventral tentacles retracted, Smynthurus Buskii, x 60. /. Point of attachment of basal muscle. g. Branch of basal muscle. h. Apical muscle. i. Papillae turned outside in. a. Orifice of tube. b. Basal part of tentacle. c. Apical part of tentacle. d. Point of attachment to dorsal surface. e. Basal muscle. Fig. 36. Ventral tentacles, Papirius purpureus, partially extended, x 60 Fig. 37. Female generative organs, Smynthurus Buskii. Fig. 38. Male generative organs, Papirius Saundersii, x 60. Fig. 39. Spermatozoa, x 250. Fig. 40. Earlier stages of spermatozoa, x 250. Fig. 41 . Greenish cellules occurring in considei-able numbers in the male generative organs, x 250. Fig. 42. Large transparent cells occurring in considerable numbers in the male generative organs, x 250. Fig. 43. Pajiirius Saundersii. T-ransLmnSocVol ME b 45 7 J' Lrjjjdoek del. TransLmn Soc. VolMlJl.. t : 'J LiiJjhoGh r/el'. G-.Jarmaji, s^c [ 449 ] XXVIII. On the Geographical Relations of the Coleoptera of Old Calabar. By Andrew Murbat, Esq., F.L.S., Assist. Sec. Boy. Sort. Soc. Read February 6th, 1862. xHE late Prof. Edward Forbes's spectdation or theory that the South- American continent was at one time united or in close proximity to Western Africa, in the direction of the Sargasso Sea, has met with very general favour and acceptance ; but the paper on the Orchidacese of Fernando Po, lately read to this Society by Dr. Lindley, and other recent observations show that much evidence is still needed before it can be said to be established. As a small contribution to the determination of this question and its collateral issues (such as the former points of junction or of greatest proximity of the two continents, and the lines by which the allied faunge and florae have been dispersed), I shall lay before the Society a few memoranda relating to the geographical affinities of the Coleoptera of Old Calabar, a part of West Africa of which comparatively little is known, and which I have had favourable opportimities of studying through the kindness of the Scottish missionaries who have established themselves there. I shall not attempt here to enter iuto an examination or description of the entire Co- leopteran fauna of Old Calabar. That task I have already commenced in a work partially published, and which, although for the present interrupted by more engrossing duties, I trust ere long to resume. On the present occasion I shall limit my notice to a few illustrations of the more remarkable affinities which I have observed between the Co- leoptera of Old Calabar and those of other countries, and, in doing so, shall refrain from noticing many species to which I could have referred as allied, when their form or genus could also be said to be generally distributed over the whole world. 1. As to America, and more especially South America. The first I shall particularly notice is Galerita — an American genus comprising upwards of forty species, of which nine-tenths are found in America, most of them in the United States, some in California, some in the West Indies, others in Brazil, &c., and so southward to Monte Video. The northern species have a red thorax and black or dark-blue elytra. As we go southwards, species are found with the colour of the thorax darker or black, and its shape narrower. In Caraccas and the West India Islands, the species G. unicolor found there has a black, narrow thorax, with blue elytra ; and G. Lacordairei, found at Monte Video, is whoUy black. I have four species of this genus from Old Calabar, all black, and aU having the elongated thorax, &c., of the South- American species ; one of these species is found also in Senegal. The American species Galerita unicolor is contrasted with my Old-Calabar species Galerita femoralis in PI. XLVII. figs. 1 and 1 a. There are only two Asiatic species ; so that the inference 3 P 2 450 MR. A. MURRAY ON THE GEOGRAPHICAL RELATIONS OF to be drawn is, that there is a little oyerflowing, on the one hand, from America into Asia, and a somewhat greater overflowing, on the other, into Africa. The FJieropsopM, or larger Brachini, furnish an instance the converse of what we see in Galerita. Their metropolis is in tropical Asia and West Africa ; hut a few and rare New- World species are found in Eastern South America. Three or four species only are to be found in America, to balance the fifteen or sixteen of Africa. I do not take into account the genera Galleida and Lebia, which, as at present con- stituted, are to be found all over the world. But there is one genus or subgenus of the family established by Eschscholtz, viz. Lia, confined to South America, of a peculiar facies, very smooth, very convex, and of a pale colour, which has a congener in Old Cala- bar, so like in form and general appearance, though of a finer colour (being, when alive, of a lovely delicate flesh-colour), that no one can doubt its affinity, although some of the characters given for Lia are not exactly found in it. This is a very marked instance of a purely American form, found nowhere else on the face of the globe (so far as we yet know), reappearing on the West Coast of Africa. The reader will find a South- American species Lia affinis contrasted with my Lia clavicornis in PL XLVII. figs. 2 and 2 a. Another, similar instance occurs in the genus Goniotropis. The whole of the species of that genus are American, with the exception of three African, of which one is a new species, described by me, from Old Calabar. The other two are not recognizable from the descriptions (BruUe's and Hope's), and may or may not really belong to the genus ; but there is no doubt as to my Old Calabar species being as close as can well be to G. castanea of New Granada. Compare figs. 3 and 3 a of PI. XLVII., being respectively Goniotropis castanea and my G. Wijlei. Lacordaire erroneously unites with this genus Pseudozcsna orientalis of Klug, and equally erroneously, as I think, keeps distinct from it Seller's genus Tropopsis from Chili. Dejean's genus Rypolitlms, notwithstanding the difficulty of finding appreciable characters to distinguish it from Harpalus, has a sufficiently marked facies to make it readily recognized by the eye, the upper surface having a peculiar granulated texture. West Africa and South America are the sole countries where they have been found. Old Calabar produces two species. America possesses a peculiar form of Orectochilus (of which the most readily seizable character is the want of a scuteUum) ; no species of this has hitherto been found out of America. Old Calabar has now produced one. In the Histeridse, a peculiar rounded form cUstinguishes the genus Omalodes. No species exactly of that genus has been found elsewhere ; but species of an allied genus, Contipus, forming the transition between it and the ordinary form of Sister, have been found in Yucatan, Senegal, and Old Calabar. In the Nitidiilidse, two species of the American form Prometopia have been found in Africa; and the genus Flatychora is represented by Thomson's Pherocopis, species of which have been found on both sides of the continent. Lobiopa senegalensis is scarcely distinguishable from the South -American L. contaminata. Among the Trogositidse, species of AUndria are found in 1)oth countries. In the Buprestidfe I may specially refer to the genera Actenodes and Belionota as of THE COLEOPTERA OF OLD CALABAR. 451 the closest affinity. Tliey are confined, the one (Actenodes) to America, and the other (Belionota) to Africa and the Old World. Actenodes chaUbeitarsis from Mexico may be contrasted with my Belionota Championi* from Old Calabar in PI. XLVII. figs. 5 and 5 a. Lampetis, or Psiloptera, is foimd all over the tropics ; but the Old-Calabar species is closer to some of the South- American species than to the East-African or Indian species. Psiloptera equestris from Brazil is figured in PI. XLVII. fig. 6, alongside my Lampetis piperata f from Old Calabar, fig. 6 a. The Lamellicornes in like manner present similar affinities. Among the Cetoniadse, perhaps the most striking is the almost exclusive possession respectively taken of Ame- rica by the Gymnetidae, of Africa by the Pachnodae, and of the Indian district l)y the Protsetise. This is not the place to discuss the value of the different characters used to distinguish these groups ; nor do I at all mean to indicate that I consider the three I have mentioned as of equal value. They are undoubtedly not so ; Protcetia and Pach- noda are greatly nearer to each other than Gymnetis to either of the others. Biit I place them together as alliances in a certain sense. They are examples of a not very uncommon mode in which affinity may be traced, — viz. when a general resemblance remains, while some considerable deviation has taken place in the structure. In Gym- netis and Pachnoda, for example, the colouring, texture, and peculiar dull velvety ap- pearance are preserved, while the structural characters have deviated. Every systematic naturalist will acknowledge the great assistance he derives in tracing out the true rela- tions of species from some indefinable family resemblance or trick of the visage, which suggests the idea and sets him in the right track, but which he finds it impossible to render into words. Such is the resemblance between Pachnoda and Gymnetis. At the same time it is a resemblance which, while suggesting a common parentage, according to the Darwinian theory, cannot be cited as indicating more than a very distant rela- tionship. * Nigro-fusca, capite thorace et subtus chalybeo-cuprea parum purpurata ; capite declivi, oculis magnis fere verticem capitis attingentibus, Venice linea Isevi striata, antice irvegulariter punctato, maculis parvis Isevibus, clypeo subbilobo ; thorace antice angustiore, aiigulis antice declivibus, basi recto, ante scutellum utrinque valde sinuate et exciso, magna profunda fovea semilunari transversa utrinque versus latera posita, paulo pone medium fovea et an- gulis posticis parum aciculatim lineatis ; scutello longissimo, thorace longiore (quarta parte elytrorum longitudine) ; elytris basi utrinque angulato-attenuatis, apice apiculato, ad suturam siuguhs quinquecostatis, punctatis, punctis a retro formatis, costis, punctis et textura paulo ut in Necrodi, costis tertiis et quartis conjunctis paulo pone medium : subtus lateribus punctatis et aciculatim lineatis, medio Itevi ; prothorace rotundatim et late carinato ; segmentis abdo- minis ad latera apiculatis, medio longa plana hand profunda fovea, utrinque Isevi carina metata aciculatim punctata ; femoribus aciculatim lineatis. Long. 12 lin., lat. 4 lin. t Supra aureo-cuprea, irregulariter profunde impressa vel punctata, nigro vel viridi-nigro maculata, subtus ohaly- bea ; clypeo emarginato, labro transverse, mandibulis rotundis crassis molaribus ; thorace antice angustiore, lateribus angulatis, basi siuuato, disco irregulariter spatiis Isevibus notato, punctis confluentibus et rugosis SEepius latera versus instructo ; scutello minuto, rotundato et convexo ; elytris apice attenuatis, ad humeros angulatis, lateribus sinuato- angulatis paulo pone humeros, apice attenuato emarginato rugose striato-punctatis, spatiis leevibus nigris disco irre- gulariter maculatis, lateribus magis rugosis et metallicis : subtus prothorace et mesosterno tricarinato, carina bilineata ; pedibus chalybeis, punctatis. Long. 12 lin., lat. 4h lin. 452 MR. A. MURRAY ON THE GEOGRAPHICAL RELATIONS OF On the whole, the amount of affinity in the Lamellicornes is very limited. Some of the tribes show indications of affinity ; but, with the exception of a few worldwide- distributed genera, such as Serica, Acloretum, and Onthophagus, I can scarcely point to a genus in which allied species are to be found in both countries. The great Augosoma Centaurus, as it belongs to the Dynastidse, has, of course, relations with the South- American Dynastidae ; but, according to Lacordaire, it has nearer affinities with the Indian Dynastidae than with those of South America. I have not made a special study of the Dynastidae myself, and therefore adopt Lacordaire's view in preference to my own empirical inclination rather to place it alongside of the South- American species. Scarcely a single species of epigeal Tenebrionidae has been received from Old Calabar ; and in like manner they are extremely scarce on the opposite coast of South America. Those which are non-epigeal furnish a few links, and the Taxicornes and Stenelytra a few others, as, for example, Stenochia molacea from Brazil, and S. longipennis* from Old Calabar, figs. 4 and 4 a of PI. XLVII. The Longicornes supply some very interesting and close affinities. For the first time a Farandra has been found in the Old World, at Old Calabar. See the figs, of Parandra brminea from North America, and of P. beninensisi from Old Calabar, figured in PI. XLVII. figs. 7 and 7 a. Borycera spimcornis from Old Calabar is obviously a near relation of Polyoza Lacordairei from Brazil (see PL XLVII. figs. 8« and 8). Several species of the American form of Callichroma are already known as found in Senegal ; but the number has been considerably increased by the Old-Calabar species ; at least, seven or eight new species have reached me from thence. A new (Erne has been found there, and one or two other American forms. But by far the largest proportion of the Longicornes of Old Calabar are new, and of the African facies. The same remark applies to the Curculionidae. Very few of these are described ; and among them are a number of most interesting and peculiar new forms. There is a form allied to that of Platyomus of Brazil, but their affinity is doubtful. Sphenophorus Phoenicis is, however, closely alLLed to the Pabn Calandras of South America ; and Bhina barbirostris is found in both countries. With these exceptions, I do not recall any species showing near relationship. The Old-Calabar Phytophaga are also mostly new, and peculiar to that country ; but * Nitida, omnino chalybeo-metallica ; capite punctato, inter oculos breviter longitudinaliter impresso, ante oculos ex- cavato, oculis partim divisis ut in Stenochia violacea, labro lato transverse ; thorace subquadrato, brunneo, sparsim tbrtiter punctato, marginato, lateribus rotundatis, margine laterali reflexo, antice pone caput expanse et glabro ; scutello parvo, triangulari ; elytris thorace quadruple longioribus, parallelis, striato-punctatis, interstitiis nitidissimis Icevissimis sparsim punctatis, punctis vix sub lente conspicuis : subtus lateribus sparsim fortiter punctata ; pedibuslongis. Long. 1 1 lin., lat. 3| lin. t Ferrugineo-fusca, punctata, punctis rugosis, oblongis vel quadratis sen angulatis : $ capite fronte trisulcato vel bilobato, antice transversim excavate ; clypeo prope oculos utrinque carinato, fere trilobate, lobe mediano obtuse subquadrate prominente ; mandibulis crassis, convexis, dentatis : thorace transversim subquadrato, marginato, pos- tiee angustiore, fortius et rugosius utrinque antice punctato ; utrinque bifoveolato, fovea una medium versus posita, altera delteidea ad basin ; angulis anticis projicientibus et acutis, posticis ebtusis ; marginibus lateralibus irregulariter sinuatis ; scutello glabro, impunctato ; elytris sub-tricarinatis : subtus mento rugosa, lateribus metasterni et segmen- torum abdominalium Isevissime papillosis, prosterni et femeribus punctatis. Long. 9 lin., lat. 3 lin. THE COLEOPTERA OF OLD CALABAR. 453 it affords an example of the American famUy CEcliomjchis. The Erotylidae present some resemblances between the two countries, as Bacne grandis to Dacne heros, &c. The general result of the whole undoubtedly is, that while by far the greater proportion of the species of Old Calabar have the West-African facies, a small but still perceptible proportion has the South-American facies. The natural inference from these relations is, that at some period the two continents must have been united or contiguous ; but this is not a necessary sequence. For example, no one will maintain that the Kentucky caves have ever been united or contiguous to the Carni- olan caves ; and yet the relationship and resemblance of the eyeless Coleoptera found in these two localities are greater than iu any two species which can be contrasted from Old Calabar and South America. I am willing to assume for the present that Europe and North America have been at some period united, and that the affinity between AnopUhalmm BilimeUi and A. Tellhampfii may be explained, on Mr. Darwia's theory, by assuming them to be the product of the same or of allied TrecU which have wandered into the caves, and that the like conditions have impressed a like form on their offspring. This latter assumption is of course not Mr. Darwin's ; for he repudiates the idea of physical conditions making much, if any, impress upon life. I confess I am still a believer in that exploded heresy. I can see no other way of explaining the existence of these allied blind insects in caves so widely separated ; and if it applies to the caves, it may equally apply to any district with well-marked physical conditions. May not, for instance, a North- American Frionus, the same as, or closely allied to, a European species, have wandered southAvards to the Brazil districts and produced a Polyoza, while its European fellow wandered into Old Calabar and produced a Bonjcera'i that is, supposing the physical condition of Brazil and Calabar to be the same, and that they have some resemblance m climate, moisture, &c., as is, I believe, the case. This is indicated, inter alia, by the scarcity of epigeal Tenebrionid^ in both. The distribution of that class of insects is cui'ious ; Lacordaire thus states it :— " In the old continent, the whole of Africa, the borders of the Mediterranean, and the regions in the neighbourhood of the Caspian Sea constitute three great centres in which they abound. The East Indies and Australia only possess a small number. In South America, the whole of Patagonia, the Argentine Republic, Chili, Bolivia, and Peru are the only regions of this vast continent where they are greatly multiplied. In North America, they begin to appear in the plains to the east of the Rocky Mountains ; and their number goes on increasing in proportion as they advance to the west. California and the still imperfectly explored regions of New Mexico seem to be, in that continent, the special country of these insects. But the most important fact to mention is, that, except a few Opatri, not one of these American forms enters into the genera peculiar to the old continent, and vice versa. More than that. North America and South America do not appear, so far as is yet known, to possess a single genus in common. The non-epigeal Tenebrionidse are distributed according to other laws. They are found everywhere, but more than anywhere else in the inter- tropical regions of South America." So far as regards epigeal species, this is in ac- cordance with what might be expected. Being apterous, their distribution must be less extensive and more defined. A barrier, such as a great river or an arm of the sea, 454 MR. A. MURRAY ON THE GEOGRAPHICAL RELATIONS OF which would be impassable to the one, may be easily crossed by the other. These epigeal Tenebrionidse, moreover, are specially adapted for desert and sandy, arid districts. In a very leafy or muddy country they would be out of place ; and such a country would prove as effectual a barrier to their progress as the sea. Their presence in both countries, more especially when South America seems to be separated by a barrier impassable to them from the north, would have afforded a strong presumption that they had once been united ; but their absence from both, or either, when the countries are unfitted for their mode of life, Avill not prove much either way. So far as it goes, however, its tendency would of course be against continuity. The number of American forms and alliances, although too distinct and decided and too great to allow us to attempt to explain them away by accidental introduction, as by the Longicornes being floated across the Atlantic in wood in which they may have existed in the chrysalis state, &c., is at the same time too small in proportion to the general pre- ponderance of true West-African forms to allow us to suppose that the continuity or proximity, if it ever existed, was extensive or recent. And although there are certainly some of the allied species which I have specially noticed, such as Galerita and Gonio- tropis, whose presence it is very difficult to account for upon any other principle than communication by contiguity, it cannot be said that this is beyond question. Some such machinery as I have above referred to, of parallel descent from common more northern ancestors, may, with the help of a lively imagination, ingeniously explain away ex- ceptional cases, and relieve us from the difficulty of accounting for the facts otherwise than by proximity or junction. And one important fact in favour of this is, that we can point to a good many alliances between Old Calabar and the East Indies of the same nature as those between it and South America, and these not alliances with no interruption but the sea, but interrupted on each hand by wide and distinct faunae. StiU I am bound to say that, after making every allowance, the impression which remains on my mind is in favour of some sort of communication having existed between West Africa and South America, south of, or nearly on the line of, the equator, inter- rupted not only by a long breach of time, but probably also by wide spaces of dis- continuity. 2, The relations of the Coleoptera of Old Calabar to those of the eastern coast of Africa are scarcely less interesting than they are to those of South America. They are not, however, most developed in the same families. In the latter, the Lamellicornes have occupied a very small place ; in Eastern Africa they occupy the chief place. The Goliaths especially attract attention. On the eastern coast we have representative or, rather, closely aUied species of the principal forms found on the west coast. The true Goliaths are represented by Goliatlms Fornasmii. Several distinct species of Fu- dicella are found on each coast. Biplognatha gagates and silicea are scarcely distin- guishable. Some rare Heteromera, Dinoscelis, a subgenus of or genus allied to Pezo- dontus, are found both in Mozambique and Old Calabar. The Lycidse are in some instances identical, and in others very closely allied. The related Longicornes are no less numerous ; and, in addition to actual instances of affinity, the Old-Calabar species 7}n't:.Linn..%c . WJM/.MA7. Del'cl i.-f. 1561 THE COLEOPTERA OF OLD CALABAR. 455 often exhibit a certain family resemblance to those of East Africa, which can be ap- preciated, but is difficult to express. The affinities with the Cape species are rarer than with those of Mozambique. The cgnnexion with Mozambique is a very interesting problem, which must wait unriddling until Central Equatorial Africa be explored. It would be very interesting to know whether the Goliaths stretch across the whole conti- nent, new species replacing new species until we arrive at Mozambique, or whether there is a total gap ; or if we must cast about for other explanations of the occurrence of G. cacicus on the one side and G. Fornasinii on the other. 3. The East-Indian district has a considerable number of alliances with Old Calabar. Not to speak of the PheropsopJii, the occurrence of Macrocheili, of OrtJiogonii, and a number of other Carabidge in both countries is of interest. The Heteromera furnish a number of similar cases. The Lamellicornes give a certain number of identical species ; but of those merely allied the number is not so great. Some genera of the Longicornes are represented in both countries ; but the affinity is seldom so marked as to prevent a tolerably good entomologist allotting the country to which they belong without being told it. The Curculionidse, as in South America, are nearly unrepresented. The relations with Europe are very few, and mostly in cosmopolitan genera, as Der- mestes ; and I know of no species identical with a European one. I have not met with any species showing relation to a purely Australian genus. On the whole, I would sum up the character of the Coleoptera of Old Calabar as indicating a distinct fauna, mainly impressed with the West-African character; pos- sessing none, or next to none, of the epigeal species either of Africa or any other country, such as Ant Ma, epigeal Tenebrionidse, &c. ; having very little connexion with the South-African fauna, and still less with that of North Africa and Europe, and, taking its proximity into account, a surprising want of the species common at no great distance in point of latitude along the Gold Coast, but separated from Old Calabar by the mighty floods of the Niger ; having considerable affinity with the eastern coast of Equatorial Africa, some examples of affinity with the East-Indian district, and a few but distinct and most interesting relations with South America. EXPLANATION OF PLATE XLVII. Fig. Species from America. Fig. Species from Old Calabar. 1. Galerita unicolor, i)ej. Spec. gen. des Coleopt. 1 a. Galerita femoralis, Murr. Ann. Nat. Hist. —Trinidad. 2. Lia affinis, Lap. — Brazil. 2 a. Lia clavicornis, Murr. Ann. Nat. Hist. 3. Goniotropis castanea, Dej. Spec. gen. des 3 a. Go?MO^roj9is W^y/ei, Murr. Ann. Nat. Hist. Coleopt. — New Granada. 4. Stenochia violacea, Fab. Syst. Eleuth.— Brazil. 4 a. Stenochia longipennis, n. sp. 5. Actenodes chalibeitarsis, Chev. — Mexico. 5 a. Belionota Championi, n. sp. 6. PsUo2)tera equestris, Olbf. — Brazil. 6 «. Lampetis piperata, n. sp. 7. Parandra brunnea, Fab. Syst. Eleuth. — North 7 «• Paratub-a beninensis, n. sj). America. 8. Polyosa Lacordairei, Serv. Ann. Soc. Ent. 8 a. Dorycera spinicornis, Fab. Syst. Eleuth. VOL. XXIII. 3 Q 1 [ 457 ] XXIX. Note on Hamamelis and Loropetalmn ; with a Description of a New Aniso- phyllea/rom Malacca. By Daniel Oliver, Esq., F.L.B. Read February 20th, 1862. An examination, some months ago, of the genera of Hamamelidese, in respect to the relations of a plant which I have since described in the Society's ' Transactions'* {Sycopsis), incidentally led me to the comparison of the American and Asiatic species included, pro- visionally, in the genus Hamamelis by the late Eobert Brown in his Botanical Appendix to Abel's ' China ' t- I have since re-examined both species {H. virginica and H. chinensis), and feel satisfied that they cannot be regarded as truly congeneric, and that they were rightly separated by Walpers, though without amended generic diagnoses, in his ' Annales ' (vol. i. p. 275)— the Chinese plant under the name Loropetalmn, as suggested by Mr. Brown. They differ primarily in the structure of the anthers, lobation of the calyx, and foliage. The anthers of the Chinese plant have their cells opening each by two vertical valves, and their coimective produced into a curved subulate appendage equalling or slightly ex- ceeding the cells. The calyx is quadrilobate, the scar of the limb being near the summit of the fruit, or about |ths from the base. The leaves are evergreen, entire, with a few curved and anastomosing secondary lateral veins. In IT. virginica, on the contrary, the small anther-cells open each by a single, strongly recurved valve ; they are destitute of the produced connective ; the calyx is divided almost to the base ; the deciduous leaves have, on each side the midrib, usually 5 to 7 straight, secondary veins, passing directly to the margin. These differences, I believe, constitute valid grounds for the generic separation of the two species, and the name Loropetalum, as suggested by R. Browoi, may suitably attach to the Chinese plant. The genus Hamamelis, however, is not lost to Asia: a plant grooving in Japan {H. japonica, Sieb. et Zucc.) is, as remarked by Dr. Asa Gray I, extremely near to H. virginica, and doubtless referable to the same genus with it. In the Japanese plant the fruit becomes nearly superior, the scar of the calyx-limb being about ith from the base. The foliage closely corresponds with that of the American species. Adopting the genus Loropetalum, the uniovulate genera of Hamamelideae may be briefly diagnosed thus : — * Flores apetali. Parrotia, C. a. M. Stamina 5-7 (antheris breyiter mucronatis [P. _per«eo], v. muticis [P. Jacquemontiana]). Ovarium semiinferum. Folia decidua. (Caucasus, Persia, Kashmir.) * Supra, p. 83. t Page 375. X Bot. Memoirs, from ' Mem. Am. Acad. Arts and Sciences,' n. s. vol. vi. p. 390. 3q2 458 PROF. OLIVER ON HAMAMELIS AND LOROPETALUM. FoTHERGiLLA, L. fil. Stamina circiter 24, antlieris obtusis (loculis in valvulas 2 a medio solutas dehiscentibus in F. alnifolia). Folia decidua. (United States.) DiSTYLiXJM, Sieb. et Zucc. Calyx liber, fere ad basin partitus. Stamina 2-8 (antherarum dehisceutia vix valvata in 1). racemoso), connectivo producto. Ovarium superum. Floras hermapkroditi \. abortu imisexuales. Folia persistentia. (Japan, Khasia?.) Sycopsis, Oliv. Calyx tubo ovarium arete cingente, lobis quasi superis. (Stamina in S. Griffithiana 8, antberis apice breviter productis.) Ovarium semiinferum. Flores abortu unisexuales. Folia persistentia. (Khasia.) ** Flores corolla vel squamulis cahjcis lobis alternantibiis donati. ■\ Petala elongato-Unearia, lanceolata v. spathtdata. CoRYLOPSis, Sieb. et Zucc. Flores pentameri. Antheree muticce (loculis in valvulas 2 a medio solutas dehiscenti- bus in C. himalayana *). Folia decidua. Flores amentacei. (Japan, E. Himalaya, and Khasia.) DicoRYPHE, Thouars. Flores tetrameri (v. quandoque pentameri). Stamina 4-5 (antherarum loculis in valvulam unicam ab externo solutam dehiscentibus, antheris acutiusculis, in D. stipulacea, fide Tulasne t). Calyx tubo cylindrico. (Madagascar.) Hamamelis, L. Flores tetrameri. Calyx profunde lobatus v. partitus. Antheree muticse loculis in valvulam unicam ab externo solutam dehiscentibus. Folia decidua, venulis secundariis rectis parallelis. (United States, Japan.) Trichocladus, Pers. Flores pentameri. Antheree mucronatse loculis in valvulam ab externo solutam dehiscentibus. (South Africa.) LoROPETALUM, R. Br. Flores tetrameri. Antherse loculis in valvulas 2 a medio solutas dehiscentibus, connectivo producto. Folia persistentia, venulis reticulatis. (China, Khasia.) ft Petala squamceformia, cahjcis lobis alternantia. Tetrathyrium, Benth. Antheree loculis in valvulas 2 a medio solutas dehiscentibus, connectivo producto. Stigma subulatum. (Hong Kong.) EusTiGMA, Card, et Champ. Antherse obtusaj loculis in valvulas 2 vix aequales a medio solutas dehiscentibus. Stigma latum, papilloso-lobatum. (Hong Kong.) I have not seen a specimen of any Diconjphe. Tulasne, in " Eragmenta Florae Ma- dagascariensis"J, describes several species, one (D. stipulacea, J. St.-HUaire) very fuUy. The mode of dehiscence of the anthers has been employed by aiithors as a sectional character in grouping the genera, but it does not appear sufficiently constant in closely allied forms to be thus prominently applied. I have examined the anthers of a number of species contained in Sir W. J. Hooker's herbarium, and have noted the character which the genera present in this respect in the above conspectus. The transverse separation of the two vertically opening valves above and below from the upper and lower margins of their cell is sometimes inconspicuous, or the valves may separate distinctly below, and scarcely at all above. In Farrotia Jacquemontiana, for example, the anthers * Siebold and Zuccarini describe two other species, C. paucifiora and C. spicata. The anthers in the latter are said to open "loculis longitudinaliter univalvibus." ■\ Ann. Soc. Nat. ser. iv. t. viii. p. 143. % Ann. Soc. Nat. I. c. PROF. OLIVER ON HAMAMELIS AND LOROPETALUM. 459 are small, unappendaged, and each loculament opens by a pair of valves parting like folding-doors from the median Kne. In P. persica they are much elongated, sliortly mucronate, and dehisce longitudinally without separation at top and bottom. In other respects the plants appear very near each other. Fig. 1. represents the anthers of Parrotia Jacquemontiana, and Fig. 2. those oi P. persica. Fig. 3. represents the anthers of Hamamelis virginica, and Fig. 4. those of Loropetalum chinense. The genera Hamamelis ani Loropetalum may be briefly described as follows : — Hamamelis, L. Calt/x profunde 4-lobatus v. partitus, infra ad ovarii basin adnatus. Fetala 4, linearia. Stamina 4, squamulis 4 alternantia ; filamentis brevibus ; antheris muticis, loculis valvattm dehiscentibus, valvula ab externo latere soluta persistente. Ovarvum biloculare, ovulis solitariis. Capsula semisupera v. fere supera, lignosa, apice bivalvis, loculicide dehiscens ; endocarpio soluto. Arbusculce in America boreali et Japonia, foliis decidids subcrenatis v. vncequaliter dentatis, venulis secundariis 5-7 rectis ad marginem parallelis. Loropetalum, R. Br. (App. Abel's China, p. 375). Calyx quadrifidus, tubo ad ovarium adnato. Petala 4, linearia. Stamina 4<, squamulis 4 parvis alternantia, epigyna ; filamentis brevissimis ; antherarum connectivo in acumen subulatum producto, lo- culis in valvulas 2 ab medio utrinque solutas verticaliter dehiscentibus. Ovarium biloculare, ovulis solitariis. Capsula semisupera, lignosa, loculicide dehiscens; en- docarpio corneo, soluto. Frutex in China et montibus Khasics indigena, foliis persistentibus, i/ntegerri/mis, subtus venulis anastomosantibus reticulatis. The Anisophyllea which I describe was collected by the late W. Griffith at Malacca. In his collections, now in course of distribution at Kew, there is a considerable number of 460 PROF. OLIVER ON A NEW ANISOPHYLLEA FROM MALACCA. specimens of it, apparently all of one gathering, and none in fruit. A note by Griffith accompanies the specimens, with a brief description, locality and date of collection (1842). Anisophyllea Griffithii, sp. nova. Poliis alternis petiolatis lanceolatis v. ovali-lanceo- latis,SEepe leviter acuminatis, apice acutis,coriaceis,siccitate flavescentibus,unicostatis, subtus nervis 2 a basi utrinque margini arete parallelis, venulis vix prominentibus reti- culatis ; floribus in spicis solitariis v. binis axillaribus, singulis arete sessiljbus re- motiusculis ; calyce limbo 4-lobato, lobis coriaceis triangularibus sestivatione valvatis ; petalis calycis lobis alternantibus, iisdem brevioribus, coriaceis, late quadrato-oblongis integris v. vix 3-lobulatis, stamina opposita arete foventibus et cum iisdem plus minus adnatis ; staminibus 8 epigynis, 4 sepalis 4 petalis (longitudine sequalibus) oppositis, filamentis crassiusculis, antheris parvis late rotundatis vel didymis, longitu- dinaliter dehiscentibus ; ovario infero 4-loculari, ovulis solitariis pendulis, stylis 4 liberis subulatis. — Arbor verisimiliter. Folia 3-4 unc. longa, l-l-g- lata, petiolo l\-2 lin. SjnccB 1^-2^ unc. ; flares circa 1 lin. diametro. Ager Punnus, Malacca, W. Griffith. (PI. XLVIII.) I consider this plant clearly congeneric with Anisophyllea zeylanica, Benth.*, and with A. laurina, Br., of West Africa, the original Anisophyllea, of which I have recently had the opportunity of dissecting flowers from specimens forwarded by Mr. Mann from the Gaboon Biver. The latter I find to have a 4-locular ovary, with a pendulous ovule in each cell and four free styles, as stated by Mr. Bentham in the Addenda to ' Flora Nigritana.' A. Griffithii differs from both the above in the absence of the strongly marked lateral costse of the leaves and the form of the petals, which in A. laurina and A. zeylanica are laciniate or fimbriate. Of the remarkable plant described by Jack, in the ' Malayan Miscellanies 'f, under the name of Saloragis disticha, I have not had the opportunity of dissecting flowers ; but from his description, and from the appearance of flowerless specimens, it is very probably a species of Anisophyllea, with which genus it is associated in Sh' W. J. Hooker's Herbarium. I cannot but consider that this genus was more correctly disposed of by Mr. Bentham under the order Rhizophoraceaj than it has been since by Drs. Hooker and Thomson in Hamamelidesel, or by Mr. Thwaites in Barringtonie8e§. It has no very close affinity with any described genus that I am aware of; but, except- ing in its alternate leaves, it presents much in common with some Bhizophoreous plants, though considerably removed from the true Mangroves. The divided or lobed petals (as in A. zeylanica), closely involute around their opposed stamens (as in A. Griffithii), point strongly to this relationship, which is not weakened by its inferior 4-ceUed ovary and tetramerous symmetry, with valvate aestivation of the calyx. Some of these characters, especially the latter-named, remove it from known genera of Hamamelideee. I have not * Niger Flora, p. 34cJ. f Reprint in Calc. Journ. iv. p. 336. % Linn. Proc. ii. p. 85. § En. PI. Zeyl. p. 119. WH.pLtck del fS-. JafmaTi. . sc . I PROF. OLIVER ON A NEW ANISOPHYLLEA FROM MALACCA. 461 seen fruit or seed ; so that I cannot speak to the fact noted by Mr. Thwaites respecting the embryo and its germination, which latter that excellent botanist iinds, in A. zeylanica, to resemble that of Careya and Barringtonia. Admitting this to apply to the genus, how- ever, I am far from considering it a sufficient ground for associating Anisophyllea with BarrmgtoniecB, from which group it is in other respects a1:)undantly diverse. But tlie structure and germination of the embryo, as described by Mr. Thwaites, I think not very dissimilar to that of the Mangrove, apart from the singular though comparatively unim- portant circumstance of the germination of the embryo while yet suspended in the pericarp from the parent tree. EXPLANATION OE PLATE XLVIII. Anisophyllea Griffitim, Oliv. Fig. 1. Flower. Fig. 5. Petal. Fig. 2. The same, with lobes of the calyx removed. Fig. G. Stamen. Fig. .1. The same, vertical section. Fig. 7- Cross section of the ovary. Fig. 4. Petal and its opposed stamen. [ 463 ] XXX. On African Anonacese. By GEORaE Bentham, Esq., Fres. L.S. Read April 17th, 1862. J^NONA CUJE, although evidently numerous in tropical Africa, are amongst the least- known of the plants of that continent. In the ' Niger Flora,' published in 1849, Dr. Hooker indicated twenty species as then more or less known ; but that number included three cultivated Anonas, and two or three supposed species since ascertained to be mere synonyms ; and very few have since been published. They are generally trees or woody climbers, bearing but few flowers, little attractive to the eye, and having considerable external resemblance to each other. The specimens gathered by collectors are there- fore generally few and imperfect, often with only a single flower which cannot be examined without destroying the specimen, and yet the changes in form which the parts of the flower undergo in the course of their development would often require their examination in bud and in the full-grown flower, as well as in fruit, before their generic and specific afiinities can be ascertained with precision. At the time of preparing the Order for our 'Genera Plantarum,' our African materials were still very deficient; many valuable additions were, however, received from the late Mr. Barter in time for the revision of our manuscript before going to press : but those more recently transmitted by Mr. Mann from the West Coast, and by Dr. Kirk from the Zambesi, could only be alluded to whilst correcting the last revise ; and some have only come to hand since the sheet was printed off. Having now examined them all, it has appeared to me that some were of sufiicient interest for figuring in the ' Transactions ' of the Society, and I have accordingly accompanied the plates I have had prepared by a synopsis of all the African species contained in the Kew herbaria in so far as the specimens permitted ; and I have also enumerated the few published species which I have been finable to identify, as well as such new ones as, from the imperfection of the specimens, must remain as yet of doubtful affinity. AnonacecB, with few exceptions, have a very limited geographical range. Their usually arboreous habit, slow growth, tardy maturity, and comparatively few flowers with indehiscent fruits, little attractive to birds and not endowed with any peculiar means of dispersion, give to most of the species but few chances in the general struggle for existence. Of the nearly forty African species now more or less known, all are confined to that contiaent, with the exception of the American Anona palus- tris, which appears to be really native in the swamps of the western tropical coasts ; and none are as yet known to extend across from the West to the East, excepting Anona senegalensis, which is, again, a very American form and very closely allied to several American species. In Africa, however, it is described as covering immense tracts of country in the greatest abundance, as well by Leprieur in Senegambia, as by E. Vogel in Bornou and by Kirk on the Zambesi. Of eleven African genera, three, Sexalobus, Ilonodora, and Clatlirospermnm, all very VOL. XXIII. 3 R 464 MR. G. BENTHAM ON AFRICAN ANONACE^. distinct, are confined to Africa, or at furthest extend to Madagascar. Six are common to Asia and Africa : of these, two, Popowia and Oxymitra, are small genera, of which the African species have a somewhat different character from the Asiatic ones ; and the other four, TJvaria, Artabotrys, Uiiona, and Melodortmi, are more considerable Asiatic genera, represented in Africa by one, two, or three species. One genus, Anona, is a large American one, unknown in Asia, but represented by three species in Africa and two in Madagascar ; and, lastly, Xylopia is also chiefly American, but Avith two or three Asiatic as well as African species. Anaxagorea, a very distinct genus common to America and Asia, has not yet been found in Africa ; nor do we as yet know, from the latter continent, of any Anonaceae allied to Thceantlms or Seteropetalum, two small groups, the one Asiatic, the other American, so similar in character that they might well be united in one genus. Several African Anonaceae have edible fruits (this quality is particularly attributed by collectors to the Anonas, to Po^wioia Kirhii, and to most of the Uvarias) ; and many are aromatic, which is especially noted of the Xylo])ia cethiopica or Malaghatta pepper and Monodora myristica or Calabash nutmeg ; and the woods of the arboreous ones are said to be hard and good. The climbing species, which bear in Asia so large a proportion to the whole, are but very few in Africa. 1. UvARiA, Linn. ; Benth. et Hook. f. Gen. PI. 23. This genus, after having been variously extended and reduced l^y different botanists, has been limited by Hooker and Thomson to those species in which the petals of each series, and more especially the inner ones, are imbricate in the bud as in the true American Guatterias, which the majority of Asiatic species much resemble in habit and in the shape of the flowers, differing only in their pluriovulate carpels. Three of the African species are quite conformable to the ordinary Asiatic type ; but in U. connivens and U. fusca the petals are much thicker and less spreading, and the outer ones valvate, thus approaching Melodonim, to which genus I should have been inclined to refer it, were it not that the inner petals are nearly as large as the outer ones, one of them very much overlapping the two others in the bud, and all are much more open than in any Melodora I know. 1. U. CHAMiE, P. de Beauv. PL Ow. et Ben. ii. 42, t. 83 ; ramulis junioribus foliisque ferrugineo-tomentosis glabratisve, foliis ovalibus ellipticisve obtusis subacumina- tisve breviter petiolatis subtus pilis stellatis minutis conspersis, peduncuHs 2-5-nis extra-axillaribus brevibus, sepalis basi v. ad medium coalitis ferrugineo-tomentosis, petalis obovatis patentibus, antherarum connectivo apice truncate, baccis stipitatis oblongis ferrugineo-tomentosis. — Arbor parva. Polia majora 5 poll, longa, 2^ poll, lata, nonnulla tamen multo minora. Baccse 6 ad 38, ^1 poll, longse, stipitibus 3-4-linearibus. Semina 2-6, septis horizontalibus separata, oblique oblonga, com- pressa, testa nitida, hilo incrassato umbilicato. Unona Chamoe, Guill. et Perr. Fl. Seneg. T, t. 3. f. 2. U. macrocarpa, DC. Syst. Veg. i. 489. Senegambia {Leprieur, Heudelot) ; on the Santiago in Oware (P. de Beauvois) ; at Nupe on the Niger {Barter). MR. G. BENTHAM ON AFRICAN ANONACEyE. 465 These are fruiting speciraens with the leaves at first sight glahrous, but the minute stellate hairs scattered on the under surface can generally be detected with a lens. 2. U. GLOBOSA, Hook. f. PL Nig. 210 ; ramis gracilibus novellis ferrugineo-tomentosis, foliis (parvis) oblongo-ellipticis sublanceolatisve obtusis supra nitidis subtus pilo- sulis, floribus subsessilibus solitariis geminisve ferrugineo-sericeis, petalis patentibus valde imbricatis, antheris apice truncatis, baccis subglobosis brevissime stipitatis ferrugineo-vekitinis. — Erutex ?, Z7. micrantlm Hook. f. et Thorns, affinis. Folia 2-3-pollicaria, nervis prseter costam parum conspicuis. Bractese orbiculatse, con- cav8e, deciduse. Elores parvi. Sepala lata, basi breviter connata, dense ferrugineo- sericea. Petala majora patentia extus tomentosa subsericea, in alabastro cujusve seriei valde imbricata, in speciminibus nostris vix perfecte evoluta. Stamina nu- merosa connectivo apice truncato-dilatato. Baccse 3-4 lin. diametro, 4-7-sperm8e. At Accra, in west tropical Africa [VogeT). 3. U. LTJCiDA, Benth. ; foliis ovali-oblongis acuminatis, basi late rotundatis, utrinque ramulisque glaberrimis nitidis, pedunculis terminalibus subternis brevibus flori- busqvie tomento tenui rufo-canescentibus, petalis liberis obovato-orbiculatis obtu- sissimis. — Species U. Inridce Hook. f. et Thorns, et Z7. Narum, Wall, afl&nis. Folia pleraque 8 poll, longa, 1\ poll, lata, petiolo 2-lineari, tenuiter coriacea, costa subtus prominente, nervis tenuibus. Pedicelli in specimine 3, intermeditis flore evoluto 4 lin. longus. Bractese obovatse, 2 lin. longse. Sepala orbiculata, vix 3 lin., petala fere 6 lin. longa et lata, omnia extus cano-tomentosa. Torus planus. Stamina nu- merosa, connectivo ultra loculos obovato truncato. Gynsecii carpella numerosa, linearia, arete conferta, glabra. Styli crassi, angulati, truncati. Ovula 6-8. Guatteria lucida, Boj. Hort. Maurit. 6. Island of Mombase, off the east coast of Africa {Bojer), according to the label of a specimen from a generally well-named collection of Bojer's plants formerly in the herbarium of the late Dr. Graham. The same label is affixed in the Hookerian herbarium, but in a less authentic manner, to a specimen of Polyalthia [Guatteria) Korinti. Both are probably from the Mauritius Botanic Garden, where Bojer states his G. lucida to have been raised from seeds brought from Mombase. ,4. U. CONNIVENS, sp. n. ; foliis amplis elongato-oblongis acuminatis infra medium angustatis glabris, petiolo brevissimo, pedunculis axillaribus brevibus 1-floris bracteatis floribusque minute tomenteUis, petalis conniventibus crassis latis con- cavis exterioribus subvalvatis, interiorum uno exteriore, toro convexo, antheris apice capitatis, carpellis glabris, stigmatibus oblongis. — Arbor 30-pedalis. Folia fere pedalia, supra medium 3 poll, lata, infra medium longe angustata, ima basi obtusa, petiolo crasso 3-4 lin. longo, rigide membranacea v. vix coriacea, nervis primariis arcuatis distantibus costaque subtus prominentibus. Pedunculi j— j-polli- cares, basi bibracteati et supra medium unibracteati. Sepala 4 lin. longa, orbiculata, concava, cito decidua. Petala purpurea, exteriora fere pollicem longa et lata, crassa concava acutiuscula, parum aperta, marginibus leviter imbricatis v. fere valvatis, interiora paullo minora, uno obtusissimo, cetera acutiora margine latiusculo obtegente. 3 B 2 466 MR. G. BENTHAM ON AFRICAN ANONACEiE. Torus crassus, fere liemisphsericus, sub carpellis tamen planus v. leviter excavatus. Stamina numerosissima angusta, connectivo apice capitulo parvo appendiculato. Carpella numerosa, glabra, ovulis 12-15 biseriatis, stylis oblongo-clavatis ovario ipso paullo brevioribus. Fructus non visus. Ambas Bay and Fernando Po (G. Mann). Some large, ovoid, thick glabrous carpels, full 4 inches long, containing each three or four seeds, separately gathered by Mr. Mann, may possibly belong to this species. 5. U. ? FUSCA, sp. n. ; foKis amplis elongato-oblongis acuminatis infra medium angustatis glabris, petiolo brevissimo, fioribiis subsessilibus bracteis imbricatis sepalisque ferru- gineo-sericeis, petalis latis concavis leviter imbricatis subvalvatisque, antheris apice truncatis, carpellis villosis. — Arbor 25-pedalis. Polia lis Z7. connicentis simillima, fere pedalia. Flores gemini, subglobosi, poUicem diametro. Bractese plures, arete imbricatse, interiores sepalis paullo minores, exteriores parvse. Sepala lata, basi breviter connata. Petala lis TI. conniventis subsimilia sed minus crassa et vix imbricata. Stamina ultra 2000, arctissime conferta. Carpella fere 100, villosa, 3-4- ovulata, stigmate crasso brevi truncate, post anthesin jam paullo aucta sessilia, falcato-oblonga. Fernando Po (G. Mann). The flowers of our specimen are scarcely sufl&cient to determine the affinities with precision; but I believe it to be a JJvaria, near the last. Mr. Mann describes the flowers as brown. Doubtful Species. U. ovata, DC. TJ. ovata, Schum. et Thonn. Beskr. PI. Guin. 255. TI. cylindrica, Schum. et Thonn. I. c. 256. I am unable to identify either of the above three species amongst our specimens. JJ. cylindrica, with narrow lanceolate petals, may perhaps be a TJiiona. U. ? gracilis, Hook, f Fl. Nig. 210. We have no specimen ; but from the description it is possibly a Clathrospermum. TJ. ? sp. with rather large sessile flowers, more imbricate, and without the bracts of TJ. fusca, and shorter leaves, nearer those of TJ. Chamce. Fernando Po {G. Mann). A single specimen with one imperfect flower. TJ. ? sp., named Kooharee in Sierra Leone. Fruits and one leaf only in the Kew Museum. Not unlike TJ. Chamoi ; but the carpels are more villous, shorter and thicker, on longer stalks. 2. Artabotrys, R. Br. ; Benth. et Hook, f Gen. PI. 24. This is chiefly an Asiatic genus, represented in Africa by two species, having both of them the climbing habit and remarkable hooked peduncles of the Asiatic ones ; and one of them, the W. African A. macropliylla, has the same petals, thick, concave, and closely connivent at the base, with more or less spreading oblong or narrow laminae ; MR. G. BENTHAM ON AFRICAN ANONACEiE. 467 but the other, the eastern A. brachypetala, is somewhat anomalous in the petals being almost reduced to the concave lower portion, thus bringing the plant technically near Ifelodorum or Fopowia. The habit, however, and other characters do not admit of its removal from Artahotvys, and the petals do occasionally appear to be produced into a short acumen at the top. 1. A. MACROPHYLLA, Hook. f. El. Nig. 207 ; glabra, foliis (amplis) late ovatis ellipticisve utrinque rotundatis et apice abrupte acuminatis tenuiter coriaceis supra nitidis, pedunculis multifloris, pedicellis brevissimis crassis, sepalis latis acuminatis, peta- lorum lamina oblongo-lanceolata calyce duplo longiore. — Eolia 8-10-pollicaria, j)etiolo brevissimo. Fernando Po [Vogel). 2. A. BRACHYPETALA, sp. n. ; foliis (parvis) ovali-ellipticis oblongisve obtusis novellis puberulis adultis glabratis, pedunculis bi-pluri-floris, petalis inter se sequalibus vix ultra partem concavam acuminatis, baccis obovoideis obtusissimis glaucis. — Frutex scandens ramulis tenuibus foliisque novellis ferrugineo-puberulis. Folia 1^-2^ poll, longa, f-li poll, lata, apice rotundata, basi obtusa v. rarius acutiuscula, petiolo raro lineam superante, membranacea v. chartacea. Pedunculi communes more generis breves, bifurcati, demum hamato-recurvi ; pedicelli 6-9 lin. longi, tomentosi. Sepala ovata, demum 3-4 lin. longa, extus tomentosa. Petala pauUo breviora, lata, crassius- cula, concava, circa genitalia conniventia, glabra, apice obtusiuscula v. in acumen brevissimum producta. Torus convexus, hispidus. Stamina generis, antherarum connectivo ultra loculos truncato-dilatato. Gynsecii carpella plurima, glaberrima, 2-ovulata ; styli lineares, superne patentes, decidui. Baccse |^-f -poUicares, stipite 2-3-lineari fultae, nunc late obovoidene, 2-spermae, utrinque sulco longitudinali notatae, nunc angustiores, l-spermse. Semina erecta. At Tete on the Zambesi {Kirk). The fruit is said to be pleasant eating. Doubtful Species. A. sp. ? A small tree, apparently without hooks. Inflorescence in other respects and shape of the flowers as in some species of Artabotrys ; but in our specimens they are all males. From the Gaboon River (G. Mann). 3. Hexalobus, a. DC. ; Benth. et Hook. f. Gen. PI. 24. This small genus has two African and one Madagascar species, the third African species mentioned in our ' Genera' proving to be a variety only of S. senegalensis. The African ones are both remarkable for the transverse undulations and folds of the petals (or lobes of the corolla), especially when the bud is near opening, which is not alluded to in the description of the Mascarene plant, and which was therefore omitted in the generic characters, drawn up before our H. graiidiflorus was received ; but as it occurs in both these species, it is probably characteristic of the genus. These petals 468 MR. G. BENTHAM ON AFRICAN ANONACE^. have- been described by several authors as imbricate, with one outside ; but in the only young bud I have been able to examine with care they were certainly valvate in each series, this arrangement being somewhat disguised, when the bud is ready to open, by the increased folds of the petals. 1. H. SENEGAiENSis, A. DC. Mem. Anon. 37 ; foliis ovali-ellipticis oblongisve supra gla- bris subtus puberulis, fioribus sessilibus, petalis lineari-lanceolatis acutiusculis undu- latis extus incano-puberulis. Uvaria monopetala, Guill. et Perr. Fl. Seneg. 8, t. 2. Senegambia {Leprieur and Perrottet; Heudelot, n. 360). On the Gambia {Whitfield). At Nupe on the Niger [Bai'ter). I had at first considered Barter's specimens as belonging to a difPerent species from the Senegambian ones, with longer, narrower and less obtuse leaves and narrower petals ; but a further comparison of all our specimens convinces me that they all belong to one species. The leaves are, in Perrottet's specimen, not 2 in. long and f in. broad ; in Heudelot's, 4 in. long and 1^ in. broad ; in Barter's, 4 to 5 in. long and only | to 1| in. broad ; but these agree the best in general form with the figure in the ' Flora Senegam- bise' ; and the petals alter very much in shape as they grow out. 2. H. GRANDiFLOKUS, sp. n. ; foliis petiolatis elliptico-oblongis acuminatis glabris, fiori- bus pedicellatis, pedicellis sepalisque ferrugineo-sericeis, petalis longe lanceolatis. — Arbor 60-pedalis, inflorescentia excepta glaber. Tolia 4-8 poll, longa, circiter 2 poll, lata, acumine longo, basi obtusiuscula petiolo 3-4-lineari fulta, chartacea v. tenuiter coriacea, glaberrima, costa subtus prominula, nervis primariis utrinque circiter 15. Pedunculi axillares, solitarii v. gemini, semipollicares, 1-2-flori, dense ferrugineo-sericei. Bractea ad basin pedicelli 3 lin. longa, angusta, basi convoluta, caducissima. Bracteolse in pedicello 2 oppositse, basi connatse, ferrugineo-sericese. Sepala ovata, concava, crassiuscula, 4-5 lin. longa. Petala perfecte evoluta ultra 1^ poll, longa, alba, transverse undulata, extus pube minute canescentia, basi in tubum densius pubescentem 2 lin. longum connata. Autherse numerosse, connectivo more generis ultra loculos truncato-dilatato. Carpella 10-12, villosa, 2-seriatim pluriovulata ; styli breves, reflexi, profunde bilobi, lobis dilatatis apice subtus stig- matosis. Bacca3 3-4, fere lignosae, dense ferrugineo-tomentosse, sessiles, 3 poll. longfe, 1^ poll, crassse. Semina transversa, in quoque loculo ultra 12, singula ultra pollicem longa, septis spuriis transversis (pulpa exsiccata ?) separata. Ambas Bay, West Africa (G. Mann). On the Niger River {Barter). The flowers are described from Mann's specimens ; the fruit from a single specimen of Barter's without flower, but which, from a comparison of the bark, foliage, and in- florescence, evidently belongs to the same species. Plate XLIX. Fig. 1. Corolla. Fig. 2. Stamens and gynajcium. Fig. 3. Torus bearing the gynae- cium, with the bracts; the sepals, corolla, and stamens removed. Fig. 4. Stamen, back view. Fig. 5. Carpel of the gynsecium, vertical section. Fig. 6. The same, transverse section. Fig. 7. Fruiting torus, with one carpel attached. Fig. 8. Ripe carpel, vertical section. MR. G. BENTHAM ON AFRICAN ANONACE^. 469 4. Unona, Linn. f. ; Benth. et Hook. f. Gen. PL 24. The genuine species of this genus hitherto described are all Asiatic or Australian : the two following African ones have been seen in flower only, but in that respect they agree well with some of the Asiatic ones in all essential points. The North American Asiminas might also be well considered as a section of Unona, but m.ore nearly allied to several of the Asiatic than to the African species. 1. TJ. HiRSUTA, sp. n. ; foliis breviter petiolatis ovali-oblongis obovatisve obtusis basi cordatis subtus ramulisque ferrugineo-hirsutis, pedunculis supra medium anguste bracteatis, sej)alis lato-lanceolatis, petalis tomentosis exterioribus ovato-lanceolatis 9-11-costatis interioribus minoribus. — Erutex 20-pedalis. Folia 4-6 poll, longa, 2-3| poll, lata, petiolo 2-3-lineari hispido fulta, supra glabra, subtus elevato-penni- nervia et transverse reticulato-venosa, pilis ferrugineis prsesertim ad costas hispida. Pedunculi laterales v. subterminales, irregulariter fasciculati, floridi sub- pollicares, rigidi, 1-flori, bractea lanceolata amplectente tomentosa 4-5-lineari. Sepala in floribus suppetentibus 5 lin. longa, vix basi connata, patentia. Petala exteriora 9-10 lin. longa, prope basin fere 6 lin. lata ; interiora multo minora, sed verisimiliter nondum perfecte evoluta. Torus fere planus. Stamina numerosissima, connective ultra loculos crasso planiusculo. Gynsecii carpella plurima, hispida, pauciovulata ; styli elongati, recurvi, intus sulcati. Pructus ignotus. Fernando Po (G. Blann). 2. U. OBOVATA, sp. n. ; foliis obovatis v. obovato-oblongis membranaceis utrinque ramu- lisque molliter pubescentibus, j)edunculis unifloris prope basin bractea orbicu- lato-cordata stipatis, petalis latis demum obtusissimis tomentosis. — Prutex ramulis gracUibus, novellis pilis crispulis tomentosis, annotinis glabratis. Polia obtusa, majora 3-3^ poll, longa, supra medium fere 2 poll, lata, infra m.edium angustiora, basi obtusissima v. cordata, membranacea, in sicco utrinque pallide virentia, pen- niaervia, petiolo 2-3-lineari. Pedunculi laterales v. terminales, solitarii, 1-flori, graciles, per anthesin subpoUicares. Bractea basin pedunculi amplectens, tomen- tella, 5-6 lin. diametro. Sepala lata, obtusissima, 2 lin. lata, tomentella, basi brevis- sime connata. Petala tomentosa, in alabastro acutiuscula et biseriatim valvata, in flore magis evoluto patentia, exteriora 5-6 lin. longa, fere orbiculata, obtusissima, extus subsericeo-pubescentia, tenuissime pluricostata, interiora multo minora. Torus fere planus, hirsutus. Stamina numerosa, connectivo ultra loculos crasso, truncato. Gynsecii carpella plurima, villosa, ovulis 6-8 subbiseriatis ; styli breves, recurvi, glabri. Pructus ignotus. Hot- springs at the foot of Moramballa on the Zambesi {Kirk). Amongst Asiatic TJnonce this comes nearest to U. dumosa in the orbicular bracts and obtuse petals ; but the much smaller flowers give it a very different asj)ect. The fruit is unknown ; and the ovules, more evidently arranged in two rows than in most species, indicate that the berries are most probably not moniliform but continuous as in V^. pmmosa, Dalz., and as in the closely allied genus Asimina. 470 MR. G. BENTHAM ON AFRICAN ANONACE^. PoLYALTHiA, Bliime ; Benth. et Hook. f. Gen. PL 25. An Asiatic genus, in which we include the Asiatic Guatterias as limited in Hooker and Thomson's ' Flora Indica,' and prohahly, as far as hitherto known, unrepresented in Africa. I only mention it here because in the Hookerian herbarium there is a specimen of P. {Guatteria) Korinti, with the label " Guatteria lucida, Boj. ex ins. Mombase oris Africse orientalis, Bojer." There has been, however, probably some misplacement of labels— perhaps in the Mauritius Garden ; for the Kew herbarium has another specimen, derived from the late Dr. Graham's collection, also as from Bojer, with a similar label, which is the Tlvaria lucida described above. 5. PoPOWiA, Endl. ; Benth. et Hook. f. Gen. PI. 25. A small genus, not yet sufficiently well known to be accurately characterized, con- taining two African, two Asiatic, and one Australian species ; differing from PolyaltUa in the more connivent petals, from Melodorum in the small flowers rather more open, and in the carpels with one or two ovules erect from the base of the cavity, not ventral, and from Clathrosjjermtmi in the more numerous stamens with the cells lateral, below the summit of the connectivum, not bordering the top of the connectivum. In the P. caffm, however, the anthers are not so closely packed, nor the cells so much concealed by the flattened top, as in the Asiatic and Australian species. 1. P. CAPFRA, Hook. f. et Thoms. ; foliis breviter petiolatis oblongis obtusis acuminatis acutisve basi obtusis membranaceis subtus pilosulis, costis petiolis ramulisque ferrugineo-pubescentibus, pedimculis lateralibus bifidis plurifloris, baccis stipitatis globosis subovoideisve glabris. — Rami graciles. Polia 2-4-pollicaria. Flores parvi, depresso-subglobosi, tomentelli. Sepala lata, alte connata. Petala exteriora calyce duplo longiora, lata, crassiuscula, concava, valvata, in floribus suppetentibus arete conniventia, sed forte demum subpatentia ; interiora similia, pauUo minora. Stamina vix ultra 20, brevia, connectivo crasso apice parum dilatato loculos oblongos laterales verticales vix occultante. Torus convexus, puberulus. Gynsecii carpella glabra, 1-ovulata, stylo breviter oblongo. Guatteria caffra, Sond. in Harv. et Sond. Fl. Cap. i. 9. AVoods at Omsamwubo and Port Natal {Drege, Guienzius). 2. P. ? KiRKii, sp. n. ; glabra, foliis breviter petiolatis anguste v. obovato-oblongis utrin- que obtusis membranaceis, pedunculis lateralibus brevibus paucifloris, baccis stipitatis oblongis glabris. — Habitu P. caffrce similis sed ex omni parte glaberrima. Bamuli tenues. Folia superiora cujusve ramuli 2-2^ poll, longa, ^-f poll, lata, inferiora breviora et sgepius latiora, omnia penninervia, nervis tenuibus arcuatim anasto- mosantibus, petiolo 1-2-lineari. Flores nobis desunt. Pedicelli fructiferi ad ramulos breves foliates brevissimi, uniflori, v. in pedunculo communi sen ramulo brevi aphyllo plures. Baccaj 2-6, stipite 2-3 lin. lougo fultse, ^-f poll, longse, obtu- sissimse v. vix mucronulatse, monospermss. On the Zambesi at the foot of Moramballa and opposite Senna {Kirk, who reports the fruit to be pleasant eating) MR. G, BENTHAM ON AFRICAN ANONACE^. 471 The flowers being unknowTi, the position of this species in Topowia or Clathrospermum is somewhat doubtful ; but the habit is entirely that of P. caffra. 6. OxTMiTRA, Blume ; Benth. et Hook. f. Gen. PL 26. An Asiatic genus of seven species, of which the characters had to be slightly enlarged to comprehend the three or four African ones described below. It differs from Ooniothalamus chiefly in the inner petals sessile not contracted into a claw, and from almost all the remaining Mitrephorece either in the 1- or 2-ovulate carpels, or in the outer open or spreading petals being much longer instead of shorter than the inner connivent ones. Bicliella has, however, the flowers of some of the African Oxymitras ; and we should have united it with that genus, but that the seeds are said to be 3-wiiiged, a character otherwise unknown in Anonacese. Marenteria, Nor., from the incomplete character given by Dupetit-Thouars, Nov. Gen. Madag. p. 18, ha^ been usually referred to Tlvaria or TInona ; but as he expressly says that the outer petals are spreading and the inner ones erect, it is very probably an Oxxjmitra ; and if this should prove correct, Thouars's name should in strictness be preferred. 1. O. HAMATA, sp. n. ; foliis oblongo-ellipticis acuminatis supra glabris subtus puberulis, costis petiolis ramulisque ferrugineo-tomentosis, pedicellis solitariis superne in- crassatis, sepalis minimis, petalis exterioribus longe lineari-lanceolatis apice recurvis. — Arbor parva, ramuHs gracilibus noveUis pube densa ferruginea vestitis. Folia pleraque circiter 4 poll, longa, \\ poll, lata, supra Isevia, subtus penninervia, nervis arcviatis parallelis prominulis costaque ferrugineo-tomentosis, inter nervos pube tenui conspersa. Pedicelli laterales, l-l|-pollicares. Sepala vix linea longiora, reflexa. Petala crassiuscula exteriora basi breviter concava, dein planiuscula, primum breviter linearia apice valde revoluta, et utrinque tomentosa, demum fere recta v. apice tantum recurva, usque ad 1\ poll, longa, 2-2^ lin. lata, supra glabrata ; interiora crassa, lata, vix \ poll, longa, circa genitalia arete conniventia. Torus convexus. Stamina numerosa, connectivo supra loculos incrassato. Gynsecii carpella numerosa, villosa, uniovulata, funiculo brevi ; styli Hneari-clavati glabri. Fructus ignotus. On the River Bagroo, West Africa (G. Mann). Plate L. Fig. 1. Flower with the outer petals cut short. Fig. 2. One of the inner petals. Fig. 3. Flower with all the petals removed. Fig. 4. The same with the anthers removed. Fig. 5. Anther, back view. Fig. 6. The same, inner view. Fig. 7- The same, side view. Fig. 8. Carpel and style. Fig. 9. Carpel, vertical section. Fig. 10. Carpel, transverse section. Fig. 11. Ovule. 2. O. GRACiLiPES, sp. n. ; glaberrima, foliis ellipticis obtuse acuminatis nitidulis, pedi- cellis solitariis gracilibus, petalis exterioribus ovato-lanceolatis acutiusculis paten- tibus.— Folia pleraque 3 poll, longa, 1\ poll, lata, pallide virentia, tenuiter venosa, petiolo 2-lineari. Pedicelli ultrapoUicares, glaberrimi. Sepala brevia, lata, recurva. Petala exteriora in specimine suppetente semipoUicaria, sed forte nondum perfecte evoluta, interiora dimidio minora, supra genitalia arete conniventia. Torus leviter VOL. XXIII. 3 s 472 MR. G. BENTHAM ON AFRICAN ANONACE^E. convexus. Stamiua numerosa, connectivo ultra loculos crasse capitate. Gynsecii carpella phirima, glabra, oralis 2 ventralibus superpositis. StyH breviter lineares. On the beach at Fernando Po (G. Mann, who describes it as a shrub of 25 feet, with greenish-white flowers). 3. O. ? PLATYPETALA, sp. n. ; ramuKs novellis petiolis pedunculisque tenuissime puberulis, ceterum glabra, foliis obovali-ellipticis oblongisve breviter acuminatis submembra- naceis, pedicellis solitariis gracilibus, petalis exterioribus late ovatis patentibus. — Prsecedenti primo intuitu simiKs, sed non omnino glabra, nee glaucescit, foHa tenuiora, sepala majora ovata obtusa, et petala exteriora multo latiora, 6-7 lin. longa. A small tree on the Bagroo River (G. Mann). I have some hesitation in giving a name and diagnosis to this plant, the specimen being insufficient to enable me to examine the internal structure of the carpels. It is, however, so very similar in many respects to O. gracilipes that I cannot but feel convinced that it is a congener, although the form of the petals does not admit of uniting it in the same species. 4. O. PATENS, sp. n. ; glaberrima, foliis elliptico-oblongis lanceolatisve acuminatis basi acutis, pedicellis fasciculatis, petalis exterioribus oblongis patentissimis. — Arbor elata, ramis horizontalibus. Ramuli etiam juniores uti tota planta glaberrimi. Polia majora 6-7 poll, longa, 1^ poll, lata, inferiora cujusve ramuli breviora et latiora, v. rarius omnia latiuscule oblonga, demum subcoriacea et nitidula tenuiter penninervia et reticulato-venulosa, basi in petiolum 2-3-linearem canaliculatum angustata. Pedicelli f-1-poUicares, graciles, 2-6-ni, axillares v. laterales, ramulis floridis in- terdum brevissimis 1-2-foliatis v. subaphyUis pedunculos communes simulantibus. Bractese nullse v. caducissimse. Sepala vix lineam longa. Petala exteriora demum 6 lin. longa, 2 lin. lata, obtusiuscula, prope basin parum angustata, interiora 1\ lin. longa, circa genitalia arete conniventia. Stamina ad 30, connectivo supra loculos crasso dilatato. Torus fere planus, glaber. Gynaecii carpella circiter 10, glabra, stigmate depresso-capitato. Ovula 2, e basi erecta, septo vertical! separata. Baccse in specimine nondum matures oblique subglobosse, 3 lin. -diametro, stipite brevi crasso, intits 2-loculares, loculis coUateralibus 1-spermis. Forests at Eppah on the Niger {Barter). On the banks of the Bagroo River (G. Mann). The vertical septum separating the ovules and seeds is an anomaly we have not observed in other Anonacese, but which both Mr. Pitch and myself have found at the time of flowering as well as when the fruit is further advanced. Plate LI. Fig. 1. Flower, magnified. Fig. 2. The same with the six petals removed, the sepals remaining. Fig. 3. One of the inner petals. Fig. 4. Anther, inner view. Fig. 5. The same, side view. Fig. 6. Torus and gynsecium. Fig. 7- Carpel of the gyna^cium, back view. Fig. 8. The same, side view. Fig. 9. The same, vertical section. Fig. 10. The same, transverse section. Fig. 11. Half-ripe carpel. Fig. 12. The same, vertical section. Fig. 13. The same, transverse section. MR. G. BENTHAM ON AFRICAN ANONACEiE. 473 7. MoNODOUA, Dun. ; Bentli. et Hook. f. Gen. PI. 26. The genus Monodora, entirely African, is one of the most marked among Anonacese ; for altliough placed in the tribe MitrephorecB, as having the outer petals spreading, with the inner ones connivent over the genitalia, and contracted at the base as in liitrepliora itself, with the normal anthers of the three first tribes of Anonacese, yet it is at once known by the large undulating and variegated outer petals, united at the base with the inner ones in a short ring as in Hexalobtis ; and the structure of the ovary is so peculiar, that it has been often added only at the end of the order as an anomalous genus. In my former notes on the Order (Journ. Linn. Soc. v. 72), I showed that the old idea of the ovary being monocarpeUary, with the whole inner surface Mned with ovules, as in the carpels of some Nymphseacese, was erroneous, and that the ovary consisted, in fact, of a large number of carpels united into a single unilocular ovary with parietal placentas as in some Papaveracese and in several regularly parietal Orders, but that these pla- centas are so numerous as to be absolutely contiguous and as it were blended with each other. This view has been disputed by Mr. B. Clarke, who, from the fact of the stigma of Jf. myristica being frequently oblique and even split down on one side, argues that it is evidently not a compound one, but the ordinary oblique stigma of a single carpel. Its really compound nature, however, which had been already ascertained with- out doubt by Dr. Hooker and myself in the M. termifolia, has been fuUy confirmed by the examination of the species since discovered by our intrepid African collectors, in which the stigma is still more distinctly lobed. The fruit also of 3£. gramUflora often shows a number of external longitudinal ridges and furrows, indicating the backs and sutures of the carpels ; and in a small-fruited species from the Zambesi, of which we have unfortunately neither leaves nor flowers, the sutures of the carpels are distinctly marked by prominent costae. We have now flowering specimens and fruits of four apparently perfectly distinct species, and fruits only of two more, which, with the M. angolensis, Welw., which is unknown to us, carry the number of species to seven. But there is considerable difficulty in framing diagnoses so as to make it easy to recognize them. All are perfectly glabrous, except a few tufts of hairs or partial pubescence on the inner petals ; most of them shed their old leaves before the flowering-season, so that the flowers are accompanied usually by young membranous half-grown leaves, which it is difficult to compare with the full-grown more or less coriaceous and shining ones of the fruiting specimens ; and in the case of two species, fruits alone were found on trees which had lost all their leaves. The flowers also, as in many other Anonacese, are very different in size and shape when they first expand and when they are fully developed. The most tangible characters appear to be derivable from the shape of the inner petals, from the position of the bract, and from the inflorescence on the old wood or on the young shoots. Each species has likewise its own peculiar foliage, but variable within limits very difficult to define. The fruits are also in most cases readily distinguishable from each other in size, shape, or consistence. The arrangement of the seeds in the ripe fruit is very singular. In most of the species, especially in M. grandiflom, the cavity is entirely filled by a large number of seeds fitted 3 s 2 474 MR. G. BENTHAM ON AFRICAN ANONACE^. into each otliei", and imbedded in the pulp in every direction without any apparent order. 1. M. MYRisTiCA, Dun. Anon. p. 80 ; foliis oblongis subcuneatisve vix acuminatis basi obtusis, pedunculis in ramulis hornotinis apice bracteatis, sepalis undulatis acumi- natis, petalis interioribus late ovatis basi brevissime abrupteque contractis margine dorsoque pubescentibus vix foveolatis, stigmate subintegro, fructu magno globoso. — Arbor. Folia 4-5 poll, longa, 1-2 poU. lata, rigidula. Pedunculi in ramulis brevibus laterales oppositifolii v. terminales, in speciminibus suppetentibus 2-3-pollicares, sed ex Bancroftio interdum longiores. According to Welwitsch, Journ. Linn. Soc. iii. 151, the true 31. myristica is a native of the primaeval forests of the district from the Quizembo river, north of Ambriz, to the mouth of the Coanza. I only know it from specimens cultivated in Jamaica under the name of Calabash Nixtmeg, as figured in the Botanical Magazine, t. 3059. 2. M. GBANDiFLOKA, sp. n. ; foliis obovali- v. elliptico-oblongis brevissime acuminatis basi obtusis, pedunculis in ramulis hornotinis supra medium bracteatis, sepalis undu- latis acuminatis, petalis interioribus late ovato-triangularibus subcordatis, auriculis hispidis inflexis intus foveolatis basi brevissime abrupteque contractis, stigmate lobulato, fructu magno globoso. — Folia ramulorum floralium 4-6 poll, longa, 1^-2-1^ poll, lata, tenuiora quam in M. mi/ristica ; caulina interdum pedalia, 4-5 poll. lata. Inflorescentia M. myristicce, sed pedunculi longiores graciHores, bractea foliacea acu- minata longe infra florem sita. Petala exteriora usque ad 4 poll, longa, ut in M. myristica undulata et variegata; interiora poUicaria, obtusa, concava, conniventia. Fructus usque ad 6 poll, diametro, extus longitudinaliter plus minus conspicue striatus, joericarpio crasso lignoso. Semina numerosissima, absque ordine in pulpa immersa. Xylopia undulata, P. de Beauv. Fl. Ovv. et Ben. i. 27, t. 16. excl. fr. Ambas Bay, Princes Island and St. Thomas's Island, West Africa (G. Mann). Observing a note to Dr. Bancroft's specimens of the 3£. myristica, that the flowers are often very much larger than they are there shown, I had some doubts whether the M. graiidijlora might not be the same species, as they certainly agree in many points ; but the leaves are very much larger, instead of being smaller as one would expect in the wild type, the peduncles much longer and more slender ; the bract is always from a half to two-thirds way up the peduncle, instead of close under the flower ; the shape of the inner petals, as well as can be judged in the dried state, is different, and the stigma is very decidedly many-lobed. Whether the two forms may or not be connected by Dr. Welwitsch' s specimens remains to be seen. Plate LII, (The open flower has the inner petals probably forced open in drying.) Fig. 1. Stamens and ovary. Fig. 2. The same, with a portion of the stamens removed, showing the torus. Fig. 3. Sta- men, back view. Fig. 4. Stamen, side view. Fig. 5. Pollen. Fig. 6. Ovary, transverse section. Plate LIII. Fig. 1. Fruit, natural size. Fig. 2. The same, transverse section. Fig. 3. Vertical section of a smaller specimen. Figs. 4, 5, 6. Seed, natural size. Fig. 7- The same, vertical section. Fig. 8. The same, transverse section. Fig. 9. Seed, magnified. Fig. 10. The same, with half the MR. G. BENTHAM ON AFRICAN ANONACE^E. 475 testa removed and placed on one side, showing the horizontal plates of the albumen. Fig. 11. Al- bumen with the whole testa removed. Fig. 12. Seed, transverse section. Fig. 13. Seed, vertical section, showing the embryo. Fig. 14. Embryo, back view. Fig. 15. Embryo, side view. 3. M. TENUIFOLIA, Bentli. iu Journ. Linn. Soc. v. 72 ; foliis anguste oblongis acuminatis basi angustatis ssepiusque acutis, pedunculis in ramulis hornotinis medio bracteatis, sepalis undulatis acuminatis, petalis interioribus longiuscule unguiculatis ovatis concavis apice angustatis, margine dente villoso utrinque appendiculatis, stigmate crenulato, fructu ovoideo. — Fmtex 7-pedalis. Folia in speciminibus floridis 3-4- pollicaria, sed verisimiliter nondum perfects evoluta. Pedunculi quam in if. gramli- Jlora breviores. Petala exteriora 2-pollicaria, interiora basi in unguem latum 5-6 lin. longum contracta, medio 5-6 lin. lata, uno v. ssepius utroque margine appendice dentiformi oblonga villosa aucta, acvimine apicis obtuse 3-4 lin. longo. Flores nonnulH minores, verisimiliter nondum perfecte accreti, petalis exterioribus minus undulatis. Fructus ovoidei, maximi 4 poll, longi, diametro transversal! circiter ' 3 poll., omnino ecostati, pericarpio duro sublignoso crassiusculo. Semina nitida, quam in M. gnmdiflora paullo minora, in pulpa nucis immersa, horizontalia et ad parietes arete congesta sed cavitatem ovarii non omnino implentia. At Eppah and Aghamia, on the Niger {Barter) ; Ambas Bay and Fernando Po (G. Mann). 4. M. BREviPES, sp. n. ; foliis obovati-oblongis ellipticisve breviter acuminatis basi obtusis coriaceis nitidis, pedunculis in ramulis annotinis brevibus medio bracteatis, sepalis orbiculatis, petalis interioribus latis concavis margine dense villosis, fructu globoso. — Arbor 30-40-pedalis. Folia ^1-pedalia, rigidiora et nitidiora quam in prsecedentibus, petiolo brevissimo crasso. Pedunculi e nodis vetustis 1-1^-pollicares, 1-2-flori. Bractea orbiculata, concava, obtusissima. Sepala 4 lin. longa, concava, obtusissima et vix undulata. Petala exteriora in flore ut videtur perfecte evoluto vix poUice longiora, ovata, undulato-crispa, basi cuneata ; interiora triente breviora, in unguem brevissimum abrupte contracta. Fructus globosus, vix 3 poll, diametro, ecostatus v. costis paucis obscure percursus, pericarpio crasso coriaceo-sublignoso, pedunculo 1^-polUcari. Semina cavitatem omnino implentia ut in M. grandiflora. Fernando Po and Princes Island {G. Mann). Species insufficiently known. 5. M. angolemiSjWelw. in Journ. Linn. Soc. iii. 151, from Pungo Andongo; the name only given without any indication of its characters. 6. M. sp. from the Niger {Barter). Fruits only known, which are globular, about 2 inches diameter, without longitudinal ribs or striae, and the pericarp very much thinner than in any of the preceding ones, and apparently succulent. 7. M. sp. from Shiramba, on the Zambesi {Kirk). Fruits only kno\vn, which are globular, about 1 inch diameter, and with a thin pericarp as in the last, but marked outside by 8 to 10 longitudinal raised costse. Dr. Kirk describes it as a bush which, at the time he gathered the fruits (July 1859), had neither leaves nor flowers. 476 MR. G. BENTHAM ON AFRICAN ANONACE^. 8. Anona, Linn. ; Bentli. et Hook. f. Gen. PI. 27. This is a large genus widely spread over the warmer regions of America, and repre- sented in Africa by three species, all belonging to Martius's section Guanabani, one of them identical with an American one and the two others nearly allied to corresponding American ones, and in Madagascar by two species of a peculiar type ; but none are known in a wild state in Asia. Besides the above, four species {A. muricata of the section Guanabani and A. squamosa, A. reticulata, and A. Cherimolia of the section Atta) are more or less abundantly cultivated for their fruits in Africa as well as in Asia ; and specimens are not unfrequently transmitted l)y collectors without indication of their cultivated origin, so as to have given the idea that one or other of them are indigenous to the Old World. The union of the carpels into a single fruit with numerous one- seeded cells radiating from a central fleshy torus readily distinguishes the genus from all other African or Asiatic ones. 1. A. senegalensis, Pers. Syn. ii. 95 ; ramulis ferrugineo-tomentosis, foliis ovatis ellipticisve obtusissimis subtus pube tenuissima canescentibus rufescentibusve (v. rarius glaucis subglabris ?), petalis crassis, exterioribus late ovatis, interioribus oblongo-triquetris, fructibvis demum Isevigatis areolis inconspicuis. — Prutex 2-8- pedalis v. rarius altior. FoKa ssepius 2-3-pollicaria. Pructus edulis, fiavus v. aurantiacus, magnitudine pomi. A. arenaria, Schum. et Thonn. Beskr. PI. Guin. 257- Widely spread over tropical Africa ; described as growing in great abundance over vast tracts of country in Senegambia and Sierra Leone by Leprieur, Barter, and others, in Bornou by Edward Vogel, and on the Zambesi by Kirk. 2. ? A. glaijca, Schum. et Thonn. Beskr. PI. Guin. 259. This is a somewhat doubtful species. A specimen of Heudelot's, which Guillemin thought he could identify with Thonning's plant, appears to me to be a nearly glabrous variety of A. senegalensis ; but Thonning describes the leaves as narrower and acute at the base, which I have never seen in A. senegalensis, — besides that one would infer from the expression he uses that the inner petals are broad. His plant may therefore be a distinct species as yet unknown to us. 3. A. PAiuSTRis, Linn. Spec. 757 ; glabra, foliis ovatis ellipticisve breviter acuminatis, petalis ovatis coriaceis, interioribus paullo minoribus, fructibus ovoideis demum Ise- vigatis areolis inconspicuis. — Arbor 30-40-pedalis. Polia 4-5-pollicaria, coriacea v. chartacea, reticulato-venosa, petiolo ssepius pollicari. Pedunculi ssepius poUicares. A. chrysocarpa, Guill. et Perr. Fl. Seneg. 6, ex charactere dato. In swamps at Grand Bassa Cove {E. Vogel), and, if the synonym of A. chrysocarpa be correct, in the marshes of Cape Verd and Cape Mboro in Cayor [Leprieur, his description agreeing in every respect with our plant). Dr. Hooker in the ' Niger Flora ' suggests that it may be cultivated ; but Vogel was always very exact in noting on his labels all cultivated specimens ; and in the present MR. G. BENTHAM ON AFRICAN ANONACE^. 477 instance lie certainly regarded tliis Anona as indigenous, as he refers it to the A. chrysocarpa, Lepr. 4. A. Barteri, sp. n. ; glabra, foliis brevissime petiolatis oblongo-ellipticis acnniinatis acutisve basi angustatis coriaceis nitidis, fructibus globosis areolis prominentibus umbonatis obtusissimis. — Arbor 50-pedalis. Folia 6-9 poll, longa, 2-2^ poll. lata. Flores nobis desunt. Pedunculus fructifer ^f-poUicaris. Fructus 1^-2 poll, diametro, carpellorum apicibus distinctis depresso-hemisphsericis in sicco canescen- tibus 4 lin. latis, et ssepius costis 4-6 a stigmatis cicatrice radiantibus percursis. Dissepimenta inter semina tenuia, et torus centralis parum incrassatus. Onitoka, on the Niger [Bai-ter). 5. ^. sp. ? A single specimen apparently near A. Barteri, with narrow more acumi- nate leaves and a single flower as yet very young. A small tree on the Gaboon River (G. Mann). 9. Melodorum, Dun. ; Benth. et Hook. f. Gen. PI. 28. An Asiatic genus, differing from Anona in its free carpels, and from Xylopia in the shape of the petals and the want of the peculiar hollow deciduous anther -bearing portion of the torus. The first of the two following African species does not enter precisely into either of the sections established by Hooker and Thomson, but would constitute a third with glabrous fruits and smooth arillate seeds; the second is still further removed by the want of inner petals and the narrower outer ones, but still may well be included in the genus as far as it is hitherto known. 1. M. AERiCANUM, sp. n. ; glabrum, foliis ovaU-ellipticis coriaceis nitidis subtus rubidis reticulato-venosis, pedicellis brevissimis axillaribus, carpellis 2-ovulatis, alabastris subglobosis, baccis glabris elongatis inter semina constrictis, seminibus arillatis. — Arbor 30-40-pedalis. Folia 4-5 poll, longa, 2-3 poll, lata, brevissime acuminata, basi cuneata et in petiolum brevissimum contracta, nervis primariis vix magis quam rete venularum conspicuis. Pedicelli crassi, 2-4 lin. longi, bractea basilari lanceo- lata incurva crassiuscula caducissima, bracteolis in pedicello 1-2 brevibus. Sepala in calycem trifidum connata, lobis late triangularibus, 2 lin. longa. Alabastrum globosum leviter 3-angulatum, extus sericeo-incanum v. rubiginosum. Petala exteriora lata,concava,crassa,valvata, 4-5 lin. longa; interiora minora, carinata, acute acuminata, glabra, pariter valvata. Stamina numerosa, connectivo ultra loculos truncate- dUatato. Gynsecii carpella plurima, linearia, villosa, in stylum linear em attenuata. Ovula 2, superposita. Baccse 2-poUicares, medio leviter constrictse, 2- spermse v. abortu l-poUicares 1-spermae, oblongse, stipite semipollicare fultse. Semina erecta, nitida, arillo cupulato ad medium seminis attingente in lobulos numerosissimos lineares densissime imbricates diviso. Island of S. Thomas off the West Coast, and Cameroon Mountain, at 4000 feet (G. Mann). 2. M. ? POLYCARPTJM, Benth. ; foliis ovatis v. elliptico-oblongis subacuminatis coriaceis 478 MR. G. BENTHAM ON AFRICAN ANONACEyE. nitidis subtus ramulisque molliter pubescentibus v. demum glabratis, pedicellis brevibus axillaribus, alabastris oblongo-triquetris, petalis interioribus nullis, car- pellis numerosis breviter oblongis glabris longe stipitatis. — Eolia 4-6-pollicaria, brevissime petiolata. Pedicelli solitarii v. gemini, crassi, ferrugineo-tomentosi. Sepala lanceolata, patentia, crassiuscula, 4-5 lin. longa. Alabastrum crassum, acute triquetrum. Petala exteriora pollicaria, crassissime triquetra, sibi invicem arete appressa, dorso tomeBtosa, faciebus interioribus glabris, basi circa genitalia excavata ; interiora omnino desunt. Torus convexus ? denudatum tamen non vidi. Stamina numerosa, connectivo ultra loculos depresso-capitato. Gynsecii carpella numerosa, villosa. Bacese ultra 30, glabrse, oblongse, f poll, longae, stipitibus l-l^-poUicaribus. Anonal polycarpa, DC. Syst. Veg. i. 499. Coelocline polycarpa, A. DC. Mem. Anon. 32. Sierra Leone [Afzelius) ; on the Sugar-loaf Mountain, at an elevation of 2000 feet (Dr. Daniell). From specimens in the British Museum. 10. Xylopia, Linn. ; Benth. et Hook. f. Gen. PL 28. A considerable American genus, with a few Asiatic species besides the subjoined African ones. The floral characters are nearly those of Anona and Melodoruni, differing chiefly in the very narrow long or tapering petals and in the remarkable concave almost cup-shaped torus, the outer portion, which bears the stamens outside and encloses the gynaecium, falling off as soon as the flowering is over, leaving the carjjels on an appa- rently convex torus. The fruit is nearly that of Unona. In iT. acutijlora and 2)arviflora the seeds have no arillus, which occasioned their generic separation by A. DeCandolle under the name of Coelocline ; but in X. (Bthiopica that arillus is very small ; and in no Anonaceee does its presence or absence ajjpear to be of generic value. 1. X. ^THIOPICA, A. Rich. PL Cub. 53, in adnot. ; foliis ovali-ellipticis oblongisve coriaceis nitidis subtus pilis minutis appressis puberulis, alabastris elongatis obtusis supra basin contractis, calyce tomentoso vix dentato, carpellis elongatis torulosis glaucis, seminuni arillo brevissimo. — Polia 4-6 poll, longa, 1-^2 poll. lata. Plores l^-13ollicares, extus dense sericeo-tomentosi. Baccae numerosse, sub-2-pollicares, vix 3 lin. crassse. Semina quam in sequentibus multo minora. Unona (Bthiopica, Dun. Anon. 113. Uvaria cethiopica, Guill. et Perr. Fl. Seneg. 9. XylopicB undulata fructus in P. de Beauv. Fl. Ow. et Ben. i. t. 16, excl. fl. et fol. Habzelia eethiopica, A. DC. Mem. Anon. 31. Common in the palm-woods of Senegambia [Leprieur and Perrottet) ; in forests at Eppah on the Niger [Barter) ; in Prince's Island and on the Nun River [G. Mann). 2. X. ACUTIFLORA, A. Bich. 1. c. ; foliis ovato-oblongis ellipticisve obtuse subacuminatis tenuiter coriaceis subtus glaucis adpresse pilosis, alabastris acuminatis, calyce acuto trilobo, petalis longe lineari-lanceolatis, baccis crassis oblongis vix torulosis, semi- nibus exarillatis. — Folia 2 v. rarius 3 poll, longa, |-1 poll. lata. Pedicelli calyce longiores. Plores perfecti pollicares, petalis demum ut videtur apertis v. revolutis. Baccae paucse, 1-1^ poll, longae, ^ poll, crassae. Unona acutijlora, Dun. Anon. 116, t. 22. Ccelocline acutijlora, A. DC. Mem. Anon. 32. At the confluence of the Quorra and Chadda {Barter) ; Congo [Chr. Smith). MR. G. BENTHAM ON AFRICAN ANONACE^. 479 3. X. PABViFLORA, A. Rich. 1. c. ; omnia X. acntiflorce, exceptis petalis brevioribus cras- sioribus, in speciminibus nostris lanceolatis semipollicaribus. Uvaria parviflora, Guill. et Perr. Fl. Seneg. 9, t. 3. Ccelocline parviflora, A. DC. Mem. Anon. 33. On the Niger [Vogel, Barter) ; on the Bagroo River (G. Mann). I do not feel at all confident in tlie real distinction between the above two species and the JJnona oxypetala. Dun. Anon. 114, t. 23, or Ccelocline oxijpetala, A. DC. I. c, which must also be a Xylopia, nor in my having correctly identified our specimens ; for the foliage and fruit seem to be the same in all, the differences consisting in the com- parative length of the pedicels, and especially in the length of the petals ; but that is known to change so much in Anonacese as the flowering advances, that, until we have good specimens in aU the different stages of growth from the young bud to the fading flower, the question can scarcely be decided. 11. Clathrospermum, Planch. ; Benth, et Hook. f. Gen. PL 29. A genus hitherto confined to Africa, but consisting only of one well-ascertained species, with one or two others differing in some points, but which are as yet only known from very incomplete specimens ; its position and limits, therefore, cannot be considered as settled. The size and shape of the flowers and the fruits are nearly those of Topoioia ; but the small number of stamens, either in a single row or so loosely arranged as to show the cells on the margin of their flat tops, places our plant in the tribe of Miliusiece, where it approaches nearest to Alplionsea, but differs apparently both in the shape of the anthers and in the fruit. Possibly it may hereafter be united with the Asiatic Alphonsea and two or three South- American little-known species that have been described as Bocageas but have truly valvate petals, forming a genus, like Xylopia, common to the three continents. 1. C. VoGELii, Planch, in Herb. Hook. ; ramulis foliisque novellis ferrugineo-sericeis, foliis obovali- V. elliptico-oblongis subacuminatis adultis glabris, pedicelHs ad axillas 1-3-nis, staminibus 6-9. — Prutex v. arbor parva, ramis debilibus nunc sarmentosis. Folia nunc omnia 2-3-pollicaria, nunc pleraque 4-5-pollicaria, membranacea v. tenuiter coriacea, subtus glaucescentia v. rubescentia. Pedicelli 3-4 lin. longi, bracteis minutis. Alabastra depresso-globosa, ferrugineo-sericea. Petala demum aperta v. subpatentia, late ovata, exteriora 3 lin. longa, interiora pauUo minora. Antherarum vertex planus, horizontalis, fere triangularis, ad marginem exteriorem loculos 2 parvos ostendens ; filamenta erecta, leviter complanata. Gynsecii carpella 5-8, villosa, stylo breviter filiformi. Ovula ssepius 2-3 superposita, sed interdum solitaria videntur. Carpella matura nunc omnia monosperma, in aliis specuninibus breviter oblonga 3 lin. longa et breviter stipitata, in aliis duplo longiora et longius stipitata, nunc pleraque duplo triplove longiora 2-3-sperma, inter semina plus minus contracta. Uvarial Vogelii, Hook. f. Fl. Nig. 208, t. 17- On the Quorra {Vogel) ; in Sierra Leone and on the Niger {Barter) ; on the Bagroo (G. Mann). VOL. XXIII. 3 T 480 MR. G. BENTHAM ON AFRICAN ANONACE^. Doubtful Species. 2. Uvaria ? gracilis. Hook. f. Fl. Nig. 210 ; pedicellus poUicaris, gracilis ; sepala pa- tentia, ovata ; obtusa, 2-3 lin. longa. 3. C. sp. ? A specimen from the Gaboon River {G. Mann), with much larger leaves than the C. Vogelii, rusty pubescent underneath, the flowers more numerous, of the same size and shape, but all males ; and besides the seven or eight radiating perfect stamens with marginal anther-cells as in C. Vogelii, there are round them a considerable number of small glandlike staminodia, and the centre of the flower is occupied by a hairy disk without any ovary. I Trans, Linn. SocYoi,. 2X111. Tab ' TfUKlidBletllft .HexaJobus senegalensi TRAN's.LiinM.Soc VoiJOGII.TAB^ 50. 5 -e vJHKtaiadctMi Oxymilra "h aiaat a . IbamsIinn. SggVol.X;{1II Tab 51. "WHiilckdoliilnii OxymiLra. patens. Teans.Linn.Soc.Vol. XXIII Tab. 52. Monodora graiidiflora. Trans, LiNN"- Soc.Vol.XXIILTab. 53. WHElchdijl ctlith. Monodora grandiflnr [ 481 XXXI. On Prolification in Flowers, and especially on that kind termed Axillary Pro- lification. By Maxwell T, Masters, Esq., F.L.S., Lecturer on Botany at St. George's Hospital. Read February 20th, 1862. In a paper which, is inserted in the last Part of the ' Transactions,' I had the honour of laying before the Society the results of my inquiries into the subject of median prolifica- tion. I propose now to treat especially of axillary prolification in flowers. My ma- terials bave been derived from tbe same sources that are mentioned in my previous paper ; and from them I have drawn up a list of genera in which this deviation from the ordinary rule has been observed. The list is, I believe, comprehensive enough to afford a sufficient basis for the opinions and remarks which follow, although I have no doubt many additions might be made to it by a more thorough search through the periodical botanical publications (especially those in the German language) than I have been enabled to make. Anything like a statistical record, showing the frequency with which this form of prolification has been observed in certain genera and species as compared with others, would be very difficult, if not impossible, to draw up. The approximate estimates which I have formed are, I believe, sufficiently correct for the purposes of this paper. Among the many points of interest presented by the subject, the following are par- ticularly treated of in this memoir, — viz., the nature, number, and position of the adven- titious buds, the genera in which the change is most frequently to be met with, and the inferences to be derived therefrom, the changes that occur in the flowers so affected, conjointly with the prolification, &c. There are also certain flowers whose construction is such as to render them particularly interesting at all times, and yet more so when the subjects of any deviation which illustrates their normal mode of formation ; these are, of course, not overlooked in this paper. Other flowers, that have been erroneously said to be the subjects of this malformation, also demand notice at our hands. A comparison of the two forms of prolification, axillary and median, leads to such interesting results that I have devoted some space to it. This affords me the opportunity of showing how the morphology of certain of the large families of plants may be elucidated by cases of pro- lification ; and at the same time it enables me to insert certain jiarticulars relating to median prolification, which have presented themselves to me since the publication of my paper on that subject. Axillary prolification is the term applied to those cases wherein one or more adventi- tious buds spring from the axils of one or more of the parts of the flower. Engelmann makes use of the word " ecblastesis " to denote the same condition. Both terms are open to the objection that they do not clearly enable us to distinguish prolification occurring within the flower from a similar state originating outside the flower, within the bracts of the inflorescence. This latter condition, called by Moquin-Tandon lateral 3 T 2 482 MR. M. T. MASTERS ON AXILLARY PROLIFICATION IN FLOWERS. prolification, is as truly axillary as that to which the name is restricted. In conse- quence of certain peculiarities in the structure of some flowers, to be hereafter alluded to, it is not in all cases easy to decide whether the new growth springs from the interior of the flower, or from the inflorescence beneath the flower. The accessory bud presents itself as a leaf-bud, a branch, a flower-bud, or a miniature inflorescence ; it may be sessile, but is far more frequently stalked, and in more than half the number of cases is a flower-bud or an inflorescence. There may be one or more of these buds ; if two only, then they are usually placed directly opposite one to the other, on the opposite sides of the flower. It wiU. be seen, from the appended list, that the orders and genera in which this descrip- tion of adventitious growth occurs most frequently are the following : — CrucifercB, espe- cially the genus Brassica ; Caryophyllacece, e. g., DiantJms ; Reseclacece ; LeguminoscB, e. g., Melilotus, Trifolinm, Sfc. ; Bosacecs, e. g., Bosa, Potentilla, Sfc. ; JJmhelUferce, and Cam- panulacecB. For the most part, these are groups also peculiarly liable to central pro- lification. All the parts of the flower may be thus affected ; but, as might have been anticipated from the foliaceous nature of the sepals, the new bud usually arises from within the axil of one of those organs. Next in frequency to the calyx, the pistil is subjected to this change — the carpels, however, in such a case being disunited and leaf-like. The petals rank next, and lastly the stamens ; these latter, indeed, are usually, but not invariably, absent in these instances, the new growth even occupying their position. Hence it may well be that, when such is the case, there is no real axillary prolification, but rather the substitution of a bud for a stamen. Generally, however, the position of the accessory bud is such that it may properly be referred to the axil of an undeveloped stamen. The largest number of instances of this malformation, not merely generically, but also individually, occurs in plants the members of whose floral whorls are not united one to the other : thus, it is far more common in polypetalous flowers than in gamopetalous ones. In the prolified flowers belonging to the latter group, the sepals, if not actually uncombined, are only united for a short distance. The same relationship, but in a much less degree, exists in the case of median prolification, as that aberration is likewise most . commonly met with in polypetalous flowers. Another feature of interest is the rarity with which axillary prolification is met with in irregular gamopetalous flowers. It may be that the irregular and comparatively excessive growth in some parts of these flowers, as compared with others, may operate in checking any luxuriant tendency in other directions. - As in the case of median prolification, plants having an indefinite inflorescence are more liable to be affected with ecblastesis than those having a definite one. The degree of branching of the inflorescence may be noticed, as this deformity is far more common in branched inflorescences than in those where there is either a solitary flower or a spicate inflorescence. More than two-thirds of the entire number of genera cited as the subjects of this malformation have a branched inflorescence of some form or other ; and about two-thirds of the cases occur in genera having indefinite inflorescences. If individual instances could be accurately computed, the proportion would be even higher. MR. M. T. MASTERS ON AXILLARY PROLIFICATION IN FLOWERS. 483 Pully three-fourths of the entire number of genera recorded as occasionally the subjects of this irregularity possess in their usual state some peculiarity of the thalamus ; for instance, in about a third of the whole number of genera, the thalamus is more or less prolonged between some or other of the floral whorls, e. g., Caryophyllacem, Fotentilla, Anemone, Dictamnus, Umbelliferce, &c. About one-fourth of the genera have numerous stamens or numerous carpels, or both, springing naturally from the thalamus. In others (about one-sixth) the thalamus is enlarged into a disk, or else presents one or more glan- dular swellings, e. g., Beseda, NymphcBa, Cruciferce. In the last-named family, as has been already remarked, prolification is very common. It would be interesting to ascer- tain precisely what part of an inflorescence is most liable to this affection ; but as in- formation on this point is but rarely given in the records of these cases, I can only give the results of my own observations, which go to show that, in a many-flowered inflo- rescence, those flowers at the outside, or at the lower portion, seem to be more frequently the subjects of this change than those situated elsewhere. This may probably be accounted for by the fact that the malformation is met with most generally in plants with an indefinite form of inflorescence, and therefore the lowermost or outermost flowers are most fuUy developed ; the upper flowers being in a less advanced condition, the change is more likely to be overlooked in them ; or it may be that, from the unusual luxuriance in the lower flowers, the upper ones may be either present in their ordinary condition, or may be (as indeed frequently happens) stunted in the size and proportion of their several parts. Various changes in the form or arrangement of the several floral whorls accompany this malformation ; some of these affect the particular organ or organs implicated, and these only, while in other cases some other parts of the flower likewise undergo mo- dification. The changes most commonly met with are such as may be classed under Goethe's theory of retrograde metamorphosis : for instance, if a supplementary bud be developed in the axil of a sepal, that sepal is likely to be more than ordinarily leaf- like in appearance. The dislocation of the affected sepal from its fellows is a very fre- quent occurrence ; in cases of this kind the detached sepal is placed below the others, thus approximating, in position as well as in function, to the bracts. In some of the instances of proliferous pears, on which I shall have occasion to comment, the sepals are described as sharing in the succulent character of the fruit. The petals, under such circumstances, often exist in the guise of sepals or of small leaves ; and instances are recorded wherein the place of the calyx and corolla was sup- plied by a succession of overlapping green scales, from the axils of which the new buds arose. Such instances seem to afford an extreme degree of a more common change, viz., the diminished size and contracted appearance of the sepals and petals when affected with axillary prolification. They have also a close relationship to such develop- ments as we see in the Wheat-ear Carnation, in certain species of the genus Ifasa and others, wherein the calyx is repeated over and again, to the partial or complete sup- pression of the other parts of the flower. All these cases may be in part explained by the operation of the principle of compensation. When the androecium is affected, the stamen either remains unaltered, or is present 484 MR. M. T. MASTERS ON AXILLARY PROLIFICATION IN FLOWERS, ill a more or less petal-like condition ; bnt it far more frequently happens that the stamen is entirely suppressed, the adventitious bud supplying its place ; thus was it in the Dianthus, a figure of which accompanies this paper (PL LIV. fig. 1) *. The pistil, too, is necessarily subjected to very grave alterations when affected with this malformation. It is separated into its constituent carpels ; and these assume a leaf- like aspect, and are in the great majority of instances destitute of ovules. Indeed, virescence or chloranthy is very intimately connected with this aberration, as might have been anticipated ; for if the parts of the flower assume more or less of the condition of stem-leaves or bracts, it is quite natural to expect that they will partake likewise of the attributes of leaves, even at the expense of their own peculiar functions. It occasionally happens that an adventitious bud arises from, the axil of a monocar- pellary pistil. This takes place sometimes in Leguminosce, and seems to have been more frequently met with in Trifolium repens than in other plants. The species named is, as is well known, particularly subject to a reversion of the outer whorls of the flower to leaves, and even to a leaf-like condition of the pistil. There are on record instances wherein a leaf-bud has been placed in the axil of a more or less leaf-like carpel ; while at other times a second imperfect carpel has been met with in the axil of the first f. I have myself seen numerous imperfectly developed cases of this kind. It may be asked whether such cases are not more properly referable to central prolification — whether the axis is not in such flowers terminated by two, rather than by one carpel ? It is, however, generally admitted by morphologists that the solitary carpel of Leguminos(B is not terminal, but is the sole existing member of a whorl of carpels, all the other members of which are suppressed as a general rule, though exceptional instances of the presence of two and even of five carpels have been described %. Again, the adventitious bud or carpel is placed, not laterally to the primary one, or opposite to it, on the same level, but slightly higher up — in fact, in the axil of the primary carpellary leaf. Grifiith figures and describes § an instance of the kind in a species of Melilotus. The stalk of the ovary is mentioned as having a sheathing base, bearing in its axil a prolongation of the axis of inflorescence, in the form of a short spike with hairy bracts and imperfect flowers, the latter having a well-formed calyx and rudimentary petals and stamens. Grifiith infers, from this specimen, that the legume is not to be considered as a terminal leaf. I have, in my paper on median prolification, adduced reasons for discarding the term " prolification of the fruit ;" and the instances now to be commented on supply addi- tional force to those reasons. A very frequent malformation in pears is one wherein a second pear proceeds from the centre of the first, and even a third from the centre of * This Dianthus has the more interest from its similarity to the one described by Goethe, Metam. der Pflanzen. cap. 16. sect. 105 ; but in that instance median prolification also existed. For my specimens I am indebted to Mr. T. Moore, F.L.S. t Linnaea, vol. xv. p. 266. c. ic. Caspary, Schriften d. Physik.-Oek. Gesell. zu Konigsberg, Bd. ii. p. 5, tab. 3. fig. 39, &c. X Lindley, Veg. King. p. 545 ; also Clarke on the Position of Carpels, Linn. Soc, December, 1850. § Notulse, vol. i. Dicot. p. 126. Atlas, pi. xliii. MR. M. T. MASTERS ON AXILLARY PROLIFICATION IN FLOWERS. 485 the second *. Pears are occasionally also observed arising either from the axils of the sepals of the primary pear or from the axil of leaves originating on the outer surface of the fruits — using the term fruit in its popular sense. These cases afford strong con- firmation of the view, that the outer portion of the so-called fruit in these plants is rather to be considered as an expansion and hoUowing-out of the flower-stalk, than as formed from the calyx-tube. It is noteworthy that the true carpels and seeds are frequently entirely absent in these cases t- M. Trecul has described and figured an instance in a species of Prismatocarptis, in which a second flower proceeded from the axil of a bract attached to the side of the fruit of the first flower %. I have in my previous paper al- luded to the occasional presence of leaves on the calyx-tube, so-called, of the Rose, Pear, and Apple, to which I may now add, on that of Cratcegus tcmacetifoUa. The unripe fruits of some species of Lecijthis were stated by Von Martins, at the Meeting of the German Naturalists at Carlsruhe, to produce buds when placed in the earth. The fruit of these plants is probably of the same nature as that of the Tomacece. The fruits of Oiyuntia Salmiana and O. fragilis § have been observed to form small fruit-like branches around their summits. M. Napoleon Doumet describes the fruit as ripening as usual, but as being destitute of seeds in the interior ; after a little while, the fruit begins to wither, and then a circle of small buds, like those of the stem, may be seen at the top of the fruit, each bud springing from the axil of a little tuft of wool and spines found on the fruit. These little buds elongate into long shoots, produce flowers the following year, which flowers exhibit the same peculiarity. Gas- parini and Tenore are said to have recorded the same fact as long since as 1832. The specimen from which my figure was taken produced its fruits in the Eoyal Gardens at Kew, and is now preserved in the museum of that establishment (PI. LIV. fig. 2). The adventitious growth in these cases appears to arise from the tufts of spines, which, it has been suggested, are the homologues of the sepals. There can be little doubt that fhe outer and lower portion of the fruit of Opuntia and its allies is a dilatation of the flower-stalk. This is borne out by the fruits of Pereskia, which bear leaves on their surface arranged spirally. The fruits of Pereskia Bleo are mentioned as producing buds from their summits, in the same way as the Opuntia just cited. P. Bleo is said, by M. Delavaud \\, to present this anomaly as a constant occurrence. On the summit of the primary fruit, arising apparently from the axils of the sepals, or of small leafy bracts in that situation, are a series of fruit-like branches, which, in their turn, are surmounted by others, even to the fourth generation. I have not seen an instance of this myself; but a figure is given in the work below cited. Tenore also has recorded " the transformation of the fruits of Nymphcsa alba and iV. Lotus into true tubercles, after the seeds had returned to the condition of elementary mucilage^." I have not seen the paper wherein this extra- * Cf. Moq.-Tandon, p. 384 ; also Liudl. Elements of Botany, p. 65, fig. 130 ; " Theor. Hortic." Gard. Chron. 1851, p. 67 ; Irmish, Flora, 1858, &c. t Caspary, Bull. Soc. Bot. Fr. vol. vi. 1859 ; also Payer in vol. i. 1854. X Trdcul, Ann. Sc. Nat. 2 se'r. vol. xx. p. 339. § Bull. Soc. Bot. Fr. vol. i. p. 306, vol. v. p. 115. II Bull. Soc. Bot. Fr. 1858, p. 685. % Atti della Reale Accademia delle Scienze di Napoli, t. iv. 1839, pp. 41-45. 486 MR. M. T. MASTERS ON AXILLARY PROLIFICATION IN FLOWERS. ordinary phenomenon is described ; but, in reference to it, I may cite the opinion of Prof. Alex. Brann, of Berlin, who states that he has read the memoir of Sig. Tenore with astonishment and incredulity : " His idea of the transformation of a ripe fruit, provided with seeds, seems quite inadixdssible ; and the application which he makes of it to N. Lotus, which has stolons like the Strawberry, is surely inexact : he may have con- founded with N. alba, a different species which bore stolons, or he may have seen a chloranthy of N. alba, wdth metamorphosis of the pistil into a foliar bud ; but then the flower would not be normal, still less would the metamorphosis have been preceded by the formation of a ripe fruit " *. Tetragonia expansa has been mentioned frequently as the subject of a similar adven- titious development. M. Clos has, however, shown that there is no real proMcation in this plant, or, at least, no axillary prolification, strictly speaking t- The specimens that I have examined in Sir W. Hooker's herbarium differ in some m.easure from those described by M. Clos ; and hence, as considerable interest is attached to this plant, I have deemed it advisable to speak of it here at some length. Prof. Oliver has directed my attention to the earliest notice of this plant, and of its peculiar growth, in a catalogue of plants published by Pallas %. The eminent Russian naturalist has figured and described the plant under the name of Demiclovia tetrago- noides, and seems to have had juster notions of the sti'ucture of this flower than other more recent botanists. I shaU extract such portions of his description as are necessary for the elucidation of the nature of the fruit, and of the adventitious growth attached to it ; and intercalate the observations of other botanists, as well as the results of my own examination. Pallas correctly describes the flowers as being placed on short stalks, while other authors describe them as sessile. In truth, the length and thickness of the flower-stalk are subject to considerable variation in different specimens. M. Clos says that the flower-stalk becomes gradually blended with the base of the fruit; and this is usually, though by no means universally, the case. Moreover, in some allied species, particularly in T. implexicoma, the flowers are borne on long slender stalks, which do not pass by such insensible gradations into the base of the fruit. Pallas likewise men- tions the fact that there are sometimes two flowers in the axil of one leaf, especially towards the upper part of the stem ; and I have seen instances where there were three flowers in this situation. In such cases, it is the uppermost flower that is affected in the manner hereinafter mentioned. In speaking of the flower, Pallas thus proceeds : — " Receptaculum cartilagineum, inverse conicum, compressiusculum, coronatum spinis quatuor vel quinque patentibus, et versus pedunculum, ramentis duo oppositis, minimis, sub-barbatis, notatum." Omitting such portions of the description as are not relevant to our present purpose, we come to the following passage : — " E superioribus pericarpiis (non omnibus), ad ramenta, prodeunt flores secundarii minores, &c." Here, then, we * Bull. Soc. Bot. Fr. vol. v., and ' Ueber Polyembryonie xind Keimung von Ccelobogyne.' For a figure and descrip- tion of indubitable axillary prolification in the flovrer of a cultivated sp. of Nymphcea, see Gard. Chron. August 18, 1855. t Bull. Soc. Bot. Fr. 1855. X Pallas, Enum. Plant. Hort. Demidof., Appendix, Petrop. 1781, c.ic. MR. M. T. MASTERS ON AXILLARY PROLIFICATION IN FLOWERS. 487 have the four or five horn-like processes referred to the receptacle of the flower, instead of to the calyx or to the calyx-tube, as is done by DeCandoUe, Endlicher, and others. M. Clos considers them to be processes of the upper portion of the flower-stalk, — a view which he supports by citing the fact that the supplementary flower, with its bract, is occasionally found on the top of one of the horn-like processes ; and I have myself seen a small leaf in that situation. The figure given by MM. Seringe and Heyland * shows the secondary flower as springing directly from the summit of one of the horn- like processes ; and DeCandoUe says, " cornua calycina interdum flores accessaries ge- runt " f. Reverting to Pallas's description, we find mention made of the ramenta, or small scales which are occasionally found on the fruits of this plant ; and the super- numerary flowers are stated to take then* origin from them (their axils ?). The figures given by Pallas show that the adventitious growths have nothing to do with the true sepals. Seringe and Heyland I seem to consider the ramenta or scales to be calycine lobes, as they say, "Des aisselles de quelques lobes du calice naissent, pendant la maturation, des fleurs bien conformees." Misled by these assertions, Moquin-Tandon § and others have considered the plant to afi'ord an instance of true axillary prolification. It is evident, however, from what has been stated, that the calyx is not aifected with prolification, but that the supplemental bud arises either from the extremity of one of the horn-like processes of the flower- stalk or from the axil of a small bract attached to its side (PL LIV. fig. 3). There are a few flowers, however, in which the pedicel supporting the additional flower is united to the side of the primary fruit for a considerable distance ; in these flowers, I have not been able to ascertain precisely whether the supernumerary flower-stalk arises from the base of the primary one, or is distinct from it, in the axil of the same bract. If it originates from the stalk of the primary fruit, it affords an instance of lateral prolification, or prolification afi'ecting the inflorescence. I have only to add on this point, that the pedicel of the primary fruit is sometimes provided with two rather large-stalked leaves near its base — a circumstance which would lead us to expect that the supplementary flower takes its origin from the axil of one or other of them, and thus constitutes, as just remarked, a case of lateral prolification (PI. LIV. fig. 3, b). Similar conclusions apply to the fruit of Pliiladelphus, in one species of which, P. sjj^- ciosus, M. A. Gris has observed that the so-called calyx-tube was provided -nith two small bracts, from the axil of one of which proceeded a small flower-bud 1| . As to the nature of the adventitious growth itself, but little need be said beyond what has been already stated at the beginning of this paper, — the conditions presented being, with few exceptions, of such a nature, as not to demand special comment in tliis place, albeit some of them are curious as illustrations of morphological doctrines — such, for instance, as the occurrence of tubers in the axils of the sepals of the Potato, mentioned by Knight ^. Here the leaf-lmd shows itself in the form of a tuber ; and the true nature of the latter organ is thereby elucidated. * Bull. Bot. No. 1, p. 18. t ProJromus, vol. iii. p. 4.52. % Op. cit. § Terat. Vegct. p. 373. || Bull. Soc. Bot. Fr. 1858, p. 331. % Proc. Ilort. Soc. vol. i. p. 39, fig. 2. VOL. XXIII. 3 u 488 MR. M. T. MASTERS ON AXILLARY PROLIFICATION IN FLOWERS. In the Dimitlms (PI. LIV. fig. 1), tlie adventitious growth occurred in the form of a circle of flower -stalks bearing alternate, strap-shaped, petal-like scales and one or two imperfect flower-buds, which were made up externally of leafy or petal-like scales, within which was a gamosepaloiTS calyx enclosing rudimentary petals, stamens, and carpels. In other cases, the outer scales were like carpellary leaves destitute of ovules, their margins widely separated one from the other, and their summits surmounted each by a style nearly as long as the leaf itself. A comparison of the two forms of prolification, axillary and median, leads to some interesting results, and enables me to mention a few circumstances that have occurred to me since my former paper, on median prolification, was published, or that were omitted or overlooked during its compilation. Axillary prolification is a much less frequent malformation than the central form. If only the number of orders and genera be reckoned, the truth of this statement wiU be scarcely recognized ; but if individual cases could be estimated, the difference in this respect between the two would be very much more obvious. This may perhaps be explained on the following grounds : — It is now almost universally admitted that the flower is homologous with the branch ; that, up to a certain time, the branch-bud or leaf-bud and the flower-bud do not essen- tially differ *. At a later stage, the difference between the two is manifested, not only in the altered form of the lateral organs in the flower-bud, but in the tendency to an arrest of growth in the length of the central axial portion. Now, in proHfied flowers, the functions and to a considerable extent the appearance of a leaf-bud or of a branch are assumed, and with them the tendency to grow in length. Median prolification, therefore, in this sense, is a further step in retrograde metamorphosis than is the axiUary form. To grow in length, and to produce axillary buds, are alike attributes of the branch ; but the former is much more frequently called into play than the latter ; for the same reason, median prolification is more common than the axillary form. The frequency with which " apostasis," or the separation of the floral whorls one from another, to a greater degree than usual, is met with in prolified flowers has been before aUuded to. In both forms, the adventitious growth is much more frequently a flower-bud or an inflorescence than a leaf -bud or a branch. How this is to be accounted for I can only conjecture. Perhaps it may be due to the position of the flowers on a portion of the stem of the plant especially devoted to the formation of flower-buds to the more or less complete exclusion of leaf-buds, i. e. the inflorescence. This is borne out by the com- parative rarity with which prolification has been observed in flowers that are solitary in the axils of the ordinary leaves of the plant. If the lists of genera be perused, it will be seen that nearly aU the cases occur in genera where the inflorescence is distinctly separated from the other branches of the stem. In direct proportion, then, to the degree in which one region of the stem or branches of a plant is devoted to the for- mation of flower-buds to the exclusion of leaf-buds, is the frequency with which those flower-buds may become affected with floral prolification. * Linn. Prolepsis Plant. § vii. ; Goethe, op. cit. §§ 103-106. MR. M. T. MASTERS ON AXILLARY PROLIFICATION IN FLOWERS. 489 Plowers borne upon indefinite inflorescences are liable to be affected witb either form of proliflcation more frequently than those borne upon definite inflorescences. Prolification in both varieties is also more frequently met with in branched inflorescences than in those in which the flowers are sessile ; but the degree of branching seems less material, in- asmuch as this malformation is more commonly recorded as occurring in racemes than in the more branched forms of inflorescence. Prom the similar arrest of growth in length in the case of the flower, to that which occurs in the stem in the case of definite inflorescence, it might have been expected that axillary prolification would be more frequent in plants having a cymose inflorescence than in those whose inflorescence is indefinite ; such, however, is not the case. The reason for this may be sought for in the lengthening of the floral axis, so common in prolified flowers — a condition the reverse of that which happens in the case of definite inflorescence. Median prolification occurs frequently in double flowers ; the axillary variety, on the other hand, is most common in flowers whose lateral organs have assumed more or less of the condition of leaves. The other coincident changes have either been already sufiiciently alluded to, or do not present useful points of comparison, and may therefore be passed over. The investigation of these two kinds of aberration from the usual floral arrangement brings to light many interesting facts bearing on the structural peculiarities of certain, natural orders. On some of these I propose now to speak, premising, however, that the conclusions drawn from teratological researches, should be checked by the results of a keen scrutiny into the mode of origin and progressive growth of the various flowers, and by the analogies derived from a minute and cautious comparison of one natural form with another. In the genus Anemone, the supernumerary bud has been often seen to spring from the axil of one of the leaves of the involucre, as it is generally called. If so, the proli- fication must be classed as lateral, and not axiUary. This view is borne out by the analogies presented by Uranthis, Nigella, and other genera. On the other hand, there are grounds for considering the so-caUed involucre as a calyx removed to a long distance from the corolla. Sepatica, and some species of Samadryas, in which the petals show a tendency to become tubular, may be compared with Anemone in support of this latter notion. The Criiciferce seem peculiarly liable to prolification in one or other of its varieties. When median, it usually happens that the pistil is separated into two leaves, — never into four, as might be expected were the fruit made up normally of four carpels as has been suggested*. Another common change is one which is suggestive of the relationship between this order and Capparidacece, inasmuch as the pistil is placed on the end of a lengthened thalamus or gynophore. When cruciferous flowers are affected with axiUary prolification from the region of the stamens, it almost always happens that the adven- titious buds are placed on a level with the two short stamens. This may perhaps be cited in support of the hypothesis that there are normally in this family eight stamens, the two that are usually suppressed being represented in the prolified flowers by the * Cfr. Lindl. Veg. Kingd. 3id ed. p. 355 a. 3u 2 490 MR. M. T. MASTERS ON AXILLARY PROLIFICATION IN FLOWERS. two adventitious buds. It may be here stated that there are usually (always ?) two such, buds, and two only, in this family. I have been disappointed in not having been able to discover anything in the cruciferous flowers I have examined that throws light ujion the morphology of the hypogynous glands, so common in this family. Either these bodies have been unchanged in the prolified flowers, or they have been entirely absent. The order Caryophyllacece is very liable to these malformations. This has been before alluded to, in speaking of the elongation of the thalamus, and the displacement of the members of the floral whorl. Since the publication of my paj)er on median prolification, I have been informed of the presence of that malformation in the flowers of a Geranium — a genus of an order in which such an occurrence was to have been expected, from the nature of the thalamus. Prolification among the JJmhelUferce is interesting, from the fact that frequently the calyx is completely detached from the pistil, and is separated into its constituent leaves ; at other times the structure of the calyx is less extensively interfered with. The pistil is frequently present in the guise of two disunited lance-shaped leaves. The most remarkable instance that has fallen under my notice is a specimen of Daiicus Carota, gathered by myself in Switzerland in July 1858 (PI. LIV. fig. 4). In this specimen the calyx was tubular, its limb divided into five small teeth. The carpels were leaf-like, disjoined, and unpro- vided with ovules ; between them rose a central prolongation of the axis, which almost immediately divided into two branches, each terminated by a small umbel of perfect flowers, surrounded by minute bracts. The petals and stamens were little changed ; but the calyx and the leafy carpels demand a more explicit descrij)tion. The lower part of the carpellary leaves was inseparably united to the interior of the calyx-tvibe. This latter organ was traversed by ten ribs, apparently corresponding to the primary ridges of the normal fruit ; these ribs were destitute of spines, and the bristly secondary ridges were entirely absent. Those portions of the carpels which were detached from the calyx had each three ribs, a central and two lateral ones, which appeared to be continuous with the ribs of the calyx below, — although in the case of the calyx there were ten, in the case of the carpels six ribs, three to each. This diversity in number is thus explained : — A circle of vascular tissue ran round the interior of the calyx-tube, at its junction with the limb, and at the point of insertion of the petals and stamens. This vascular circle seemed to be formed from the confluence of the ten ribs from below. Of the five ribs in each half of the calyx, the three central ones were joined together just at the point of con- fluence with the vascular circle, above which they formed but a single ril) — that traversing the centre of the carpellary leaf; the two lateral ribs of each half of the calyx seemed to be continuous, above the vascular rim, with the lateral ribs of the carpel ; these lateral ribs were connected on either side with the central one by short branches of communica- tion. The disposition of the ten ribs may be thus represented : — 111 111 32323 32323 11111 11111 The lower line of figures represents the calycine ribs, the middle row shows how each of these ribs is divided at the vascular rim, and the uppermost row shows their distribution MR. M. T. MASTERS ON AXILLARY PROLIFICATION IN FLOWERS. 491 above the rim. From this it will be seen that six of the calyciue ribs divide into three branches, one prolonged upwards as a lateral or median rib into the carpellary leaf, the other running horizontally to join with similar branches sent out from the neighbom'ing rib ; the four intermediate calycine ribs divide into two branches only, which join the side branches of the first mentioned, but have nq direct upward prolongation into the carpel. The ten ridges are placed opposite to the sepals and petals (PL LIV. fig. 6). The nature of the carpophore and of the placenta in TlmhelUfene are also illustrated by prolified flowers : thus, in some specimens described by Mr. Townsend*, the ovules were seen hanging from the margin of a leafy carpel. The carpophore, moreover, in place of being, as Dr. Lindley describes it, a forked placenta, which has become in process of develop- ment exterior to the carpels t, is in reality, as Mr. Townsend has suggested!, "^ con- tinuation of the axis, serving at first to solder the parts together, and ultimately sepa- rating to suspend the freed mericarps." That such a prolongation of the axis shoiild in some cases bear flowers (as in my specimen) is only a further proof of the real nature of the organ. It will also be obvious, from what has been before said as to the caljrx in prolified flowers of this family, that there is in this group a real calyx -tube. The proliferous flowers in BosacecB and Fomacecs have already been alluded to as throwing light on the nature of the frixit of these plants, and on the share which the expanded flower-stalk takes in their production; and similar conclusions may be drawn as to the fruit of some other orders, wherein the ovary is said to be inferior, e. g. Cactacece, FJdladelphacece, Ilyrtacece, Tetragoniacece, Campanulacece, and probably also of many other families not here alluded to. The conclusions tally well with the organogenic researches of Payer, Trecul, and others. Furthermore, the accessory buds which are occasionally found on these fruits are rather to be referred to lateral than to axillary prolification, strictly so called. They have been mentioned in this place inasmuch as they are usually recorded as cases of axillary prolification, and are to be found under that head in the works treating on such matters. Other important changes affecting the characters of certain natural orders, such as the change from an adherent to a free ovary, from perigynous to hypogynous stamens, &c., have been already mentioned, and need no further comment in this place §. * Bot. Gazette, vol. iii. p. 52. t Veg. Kingd. p. 774. t Loc. cit. § For figures and descriptions of flowers affected with prolification, which are not specially referred to in these papers, the reader is referred to the oft-cited treatises of IMoquiu-Taudon and Engelmann, where numerous references are given ; also to Dr. Lindley's 'Theory of Horticulture,' and to a memoir of Prof. C. O. Weber, in Verhandl. des Nat.-Hist. Vereins des Preuss. Rheinlandes und Westphalens, vol. vi. 1858, &c. &c. 492 MR. M. T. MASTERS ON AXILLARY PROLIFICATION IN FLOWERS. lAst of Genera in which Axillary TroUJication has been obsei'ved. Order. Genus. Leaf-bud or BrancL. Flower-bud or Inflo- rescence. From what orgiai. Ranunculaceae . . Clematis .... Flower-bud .... Sepals. Caltha Ditto Ditto. Aconitum . . . Ditto. Delphinium . . Ditto Ditto. Anemone . . . Ditto Involucre ? Nymphaeaceae . . Nymphaea . . . Fruit ? Nymphaea . . . Flower- stalked . . Petal. Cruciferas .... *Brassica . . . . Leaf-bud .... Flower-bud .... Sepals and petals f. Brassica . . . . Ditto Stamens. Brassica . . . . Ditto Ditto Pistil. Cardamine . . . Ditto Sepals. Cheiranthus . . Ditto Ditto. Erysimum . . . Ditto Sepals and pistils. Lepidium . . . Ditto Petals and stamens. Arabis Ditto Sepals. Diplotaxis . . . Flower,inflorescence Pistil, calyx and corolla. Capparidacese . . Cleome Flower-bud .... Sepals. Resedacese . . . . *Reseda Ditto Ditto. Caryophyllacese . Arenaria . . . . Branch Ditto. Agrostemma, . Leaf-bud .... Ditto. *Lychnis • . . . Ditto Ditto. Stellaria . . . , Ditto Ditto. Silene .... Ditto Ditto. *Gypsoplula . Ditto Ditto Sepals and stamens. *Dianthus . . Ditto Ditto Sepals. Dianthus . . Ditto Inflorescence . . . Petals and stamens. Malvacete . . . . Alcea .... Flower-bud .... Stamen. AurantiacefE . . . Citrus .... Ditto Ditto. Rutacese . . . . . Dictamnus . . Ditto Pistil leafy. TropEcolaceEK . . . Tropaeolum . Ditto Petals. Celastraceae . . . . Celastrus . . Ditto Ditto. Leguminosac . . . *Melilotus . . Inflorescence . . . Sepals and petals. Medicago . . Flower-bud .... Sepals. Coronilla . . Ditto Ditto. Trifolium . . Ditto i Second carpel axil- lary to first . . . 1 Pistil. Trifolium . . Flower-bud .... Sepals and petals. Rosaceae Pyrus .... Fruit? Fruit? Cerasus . . . Flower-bud .... Petals and stamens. Potentilla . . Ditto Leafy carpels. Crataegus . . Ditto Petals. *Rosa Ditto Ditto / Sepals, petals, stamens and pistil. Myrtaceae . . . Lecythis . . . Ditto Fruit? Tetragoniaceae . Tetragonia . . Ditto Ditto. Cactaceae .... Opuntia . . . . Fruit-like branch Tufts of spines. Pereskia . . . . Fruit-like branch Sepals ? Philadelphaceae. Philadelphus Ditto Sepals ? Umbellifei-ffi . . *Athamanta . Ditto Calyx. Daucus . . . Ditto Calyx and pistil. Bupleurum . Ditto Ditto ditto. .a->- Torilis .... Ditto Ditto ditto. Apium .... Ditto Ditto ditto. t A mark * is attached to those genera that are the most frequently affected with axillary prolification. Trans.Linit. Soo-Yol.XX1I1.Tab. 54. MT.MaoKTstliil KrehliSi MR. M. T. MASTERS ON AXILLARY PROLIFICATION IN FLOWERS. 493 Order. Genus. Leaf-bud or Branch. Flower-bud or Liflo- rescence. Flower-bud .... Ditto Umbel Fruit Flower-bud .... Ditto Ditto Ditto Raceme Ditto Ditto { } Flower and raceme Flower-bud .... Ditto Inflorescence . . . From what or^an. Umbelliferse . Campanulaccce . Gentiauaceae . . Convolvulaceee • Solanaceae .... Scrophulariaceas PrimulaccEe . . _.Polygonace£e . Santalaceae . . Euphorbiaceae ? LiliaccEe .... Cyperaceae. Pastinaca . . Heracleum . Angelica . . *Campanula . Prismatocarpus Gentiana . . ^Convolvulus Solanum . . Solanum . . Digitalis . . Veronica . . Anagallis . . Rumex . . . Thesium . . Euphorbia ? Herreria . . Hyacinthus. Convallaria . Leucoium . Carex .... Branch , Ditto . , Tubers . Branch Leaf-bud Ditto . Ditto . . Calyx and pistil. Ditto ditto. Ditto ditto. Sepals. Sepals, &c.? Sepals. Outer calyx. Sepals. Sepals and petals. Petals, &c. Calyx. Petals. Sepals. In place of stamens and pistils, both absent. Outer bracts? Sepals. Perianth. Ditto. Ditto. Utricle. Note. — The following genera must be added to the list of those occasionally affected with median prolification (vide p. 368 hujus voluminis) : — Ranunculacese . Leafy. Cruciferse . . Geraniaceae . Rosacese . . . Compositse . , Euphorbiaceae Liliacese ... Floral. Clematis. Delphinium. Diplotaxis. Geranium. Cerasus. Carthamus. Euphorbia. Leucoium. EXPLANATION OP PLATE LIV. Fig. 1. Fig. 2, Fig. 3, Fig. 4, Fig. Flower of Dianthus, sp. ? The calyx is removed ; the petals are reflected to show the adventitious flower-buds, &c., occupying the situation of the stamens. Opuntia Salmiana, showing the accessory buds arising from the tops of the fruits. Flowers of Tetragonia expansa, showing the position and attachment of the secondary flowers. a. Vertical section of a flower. b. Flower with two bracts on the pedicel. Flower of Daucus Carota. Petals and stamens removed. Calyx without prickles. Carpels pro- longed and developed in the form of leaves ; between the two rises a flower-bearing axis, one branch of which alone is represented. Daucus Carota. Diagrammatic sketch of the interior of the calyx and carpels, showing the ar- rangement of the ribs and the prolonged axis, the upper part of which has been removed. [ 495 ] XXXII. Contributions to an Insect Fauna of the Amazon Valley. Lepidoptera Heliconid^. By Henry Walter Bates, Esq. (Communicated by the Secretary.) * Read November 21st, 1861. "Die wissenschaftliche Untersuchung der Natur strebt in den Einzelheiten das Allgemeinezu erkennen, um endlich dem Grunde aller Dinge niiher zu komnien. Fiir diese Art Untersuchungen, die immer das Ziel der Naturforschung sein sollte, bietet wohl keine Thierclasse so reichen Stoff als die Insecten." — Karl Ernst von Baer, Address on the Opening of the Russian Entomological Society, St. Petersburg, May 1860. The family Reliconidce was established by Mr. E. Doubleday in 1847, in Doubleday and Hewitson's 'Genera of Diurnal Lepidoptera.' It was founded, on a number of Butterflies, remarkable for the elongated shape of their wings, and peculiar (with the exception of one genus, Samadryas, which the author placed provisionally in the famUy, oj). cit. p. 98) to the intertropical and subtropical zones of America. Many of them had been described by the older authors under Heliconia, Mechanitis, and several other ill-defined genera. They had been previously (in 1836) united in a tribe, Heliconides, by Dr. Boisduval iji his ' Species General des Lepidopteres ;' but this comprehended also the group Acrceidcs, which Doubleday excluded from the family. Linnaeus treated them as a section of the genus Papilio, under the name oiSeliconii. The nearest allies of the Heliconidm are the Acrmidce just mentioned and the Danaidce : all are distinguished from the true NympJialidce by the discoidal cell of the hind wings being always closed by perfect tubular nervules. Mr. Doubleday, placing more reliance on the shape of the antenna and the abdominal border of the hind wings than on the far more unportant character above named, was led to exclude the genus JEueides from the family : this rendered the definition of the two groups very difficult, if not impossible, Eueides having the wing-cells closed in the same way as the Eeliconidce. Excepting that I re-admit Eueides, and exclude Samadryas, which does not enter into the series of the American Eeliconidts, the family will be treated of in the present memoir as defined in the work above quoted. The position of the Seliconidce in the order Lepidoptera may be understood when I state that in a natural system the group would stand at the head of the whole series of families of which the order is composed. At least, this should be its place according to the view now taken of the order by many systematists, who arrange the families of Bho])alocera, or Butterflies, according to their degree of dissimilarity to the Heterocera, or Moths— in other words, according as their structure shows a lower or a higher stage in an ascending scale of organization. Eor, as the lower families of Moths are allied to other orders of insects, the further a group recedes from them in structure, the higher is the grade of perfection of the Lepidopterous type which it exhibits. The families show their degree of affinity to Moths by many characters, the principal of Avliich is the * The materials on which this memoir is founded were collected by the author during eleven years' research on the banks of the Amazons. VOL. XXIII. ^ ^ 496 MR. H. W. BATES ON THE LEPIDOPTERA structure of tlie anterior legs in tlie adult state of the insects. The Seterocera have always six perfect legs : most of the families of Rhopcdocera have the anterior pair in a more or less rudimentary condition ; and as the atrophy seems to have reached its furthest stage in the Heliconida, this group must he considered as occupying the highest rank in the order. Other characters accompany the one derived from the structure of the legs, which it is unnecessary here to enumerate. It will he seen from these remarks that the order Lepidoptera is one of those groups in the Animal Kingdom which show, heyond the many collateral branches of development that always exist, a clear linear advancement of organization. The Heliconidce, Danaidce, and Acrceidce are related to each other in a different way from that which appears in the received classifications. A few remarks on their mutual affinities are necessary, in order to exhibit the true relations of the Heliconidce to the allied groups of the Old World tropics. It has escaped the notice of all authors, that the HeliconidcB are composed of two groups, which differ very considerably in important points of structure ; in fact, the majority of the genera of which the family is composed ought to be withdrawn from it, and placed with the Danaidce. The very great superficial resemblance between the two sets of genera has led to their being united by all authors, and prevented inquiry into their real relationsliip. To avoid innovation, I will retain the family as it stands, and call the group which is allied to the Dcmaidcs, Danaoid Heli- conidjj:, and the other, which approximates somewhat to the Acrceidce, Acr^oid Heli- conidce. The Acrseoid group comprehends the genera Seliconius and Eueides; the Danaoid, the whole of the remaining HelicomdcB. The following are the distinguishing characters of the two groups : — Acrceoid Heliconidce. The hind wing-cell is very small, and the nervures are so arranged that the upper and lower radials* (discoidal nervures of Doubleday) appear to be branch and sub-branch of the subcostal nervure, the discocellulars being short and continuous with them ; the costal nervure is prolonged to the apex of the wing. The larvae are similar to those of Acrcea and Argynnis [Nymphalidce), being beset with hispid spines f. The head is broad, the palpi thick. Danaoid Heliconidce. The hind wing-cell is very large, and irregular in shape ; the two radials never appear as branches of the subcostal, but are very uncertain in position, owing to the very vacillating length and direction of the discocellulars : the costal nervure is short, and terminates on the costa, not reaching the apex of the wing. The larvse (only one species is known) are smooth, like those of the Danaidce, but are furnished with tubercles, instead of long fleshy threads. Head small, orbicular ; palpi slender J. * I have adopted the terminology of Doubleday (Doul)l., Ilewits., and Westwood's Genera of Diurnal Lepidoptera) with regard to the neuration or veining of the wings, excepting that I call the " nervules " of the subcostal and median nervures "branches," and the "discoidal nervures" "radials," these alterations appearing necessary to prevent the verbal confusion of nervule vnth nervure, and discoidal with discocellular. -j- The early states of these insects were not known to Doubleday. I reared, myself, Heliconius Erato (and Doris) and Eueides Lybia. We are acquainted, through other sources, with the larvse of //. Melpomene, H. Ricini, and Mechanitis Polymnia. X Since the above memoir was read. Dr. C. Felder of Vienna, in an article entitled, "Specimen Faunae Lepidopte- rologicse riparum Negro superioris in Brasilia septentrionali," 'Wiener Entomologische Monatschrift,' March !86'2, OF THE AMAZON VALLEY 497 This view of the affinities of the family will make a great difference in the conception of the group as regards the affiliation of the forms. Instead of heing a group isolated in its structure, and peculiar to the tropical parts of America, it results that the hulk of the genera have a very close relationship to the Danaidce, which are found in all hot countries of both hemispheres : the Acrceoid JTeliconidce alone are an isolated set of forms. The American productions, however, show a great superiority in structure and in the diversity of the forms over those of the Old World. In Africa only one genus occurs, namely, Danais. In the tropical parts of Asia four genera are found, besides Danais ; these are Ideopsis, Sestia, Euplcea, and Hamadryas, which contain together 74 species. In the hot parts of America 16 genera (of Dmiaoid Seliconidce) have been discovered, comprising 233 described species. Besides this greater diversity of generic and specific forms, the American productions show a much greater advance in organization than those of the Old World ; in other words, they recede further from what may be considered as the common type, namely, Danais. This is clear from the great and progressive modification in the position of the radial nervures and discocellulars of the hind wings, and the advanced stage of atrophy of the male fore legs reached by most of the genera. In all the Asiatic genera the fore legs are in the same condition as in Danais and the Nym- phalid(B. The following Table will show the relative value of the productions of the two hemispheres in a clearer manner. I have placed the genera in accordance with what seem to be their mutual affinities. The relative length of the lines between them is a rough expression of the degree of relationship. The collateral lines of connexion are also attempted to be expressed. Asia. Danais (31 sp.) Ideopsis (2 sp.) I I Euploea (58sp.) Hestia (12 sp.) Hamadryas (2 sp.) DaNAIDjE. Africa. Danais (10 sp.) America . Danais (8 sp.) Dan aotd Heliconid.i;. Lycorea (6 sp.) Ituna (3 sp.) Methona (3 sp.) Thyridia (3 sp.) Dircenna (8 sp.) Callithomia (3 sp.) Mechanitis (17sp.) Ceratinia (23 sp.) Sais (1 sp.) Napeogenes ( 1 7 sp.) Ithomia (107 sp.) Hymenitis (10 sp.) Eutresis (1 sp.) Olyras (1 sp.) Melinsea (13 sp.) Tithorea (7 sp.) (including Athesis) has ventured to withdraw the whole of the genera composing the group Banaoid Heliconidcs, placing them with the DanaidcB, and restricting the family Heliconidce to the two genera Helieonius and Eueides. He has discovered an excellent and constant character for the Banaidce (in which the Da)iaoid Heliconidce participate), in addition to those already known, namely, the existence of a small nervule at the base of the fore-wing median nervure which ana- 3x2 498 MR. H. W. BATES ON THE LEPIDOPTERA There is a very wide dissimilarity in minor points and in general appearance between the Asiatic set of forms and the American : the only Old World genus which at all approaches the New World group is Samadryas ; but the shape, colours, and neuration of the wings show that it has no close affinity with them. The two sets of forms seem to agree, however, in habits, and apparently occupy the same sphere in the economy of nature in their respective countries. Mr. W^allace, who has had the good fortune to observe both in their native abodes, says, the habits of the South Asian Euploecs (the most numerous genus) are precisely those of the Meliconidce. The Asiatic Danaidce are mostly above the middle size, and include some of the largest Butterflies known ; their American equivalents are in general below the middle size. Both are extremely prolific or abundant in individuals, and are amongst the most characteristic productions of their respective countries. Each set, also, are the objects of numerous mimetic resemblances on the part of other Lepidopterous insects of their own region belonging to different families, — the Asiatic mimickers being modelled after the Asiatic Danaidce, and the American after the American members of the same family. The entire dissimilarity of the two sets of forms would seem to teach us that there can have been no land com- munication east and west between the tropical parts of Asia and America since they first came into existence, and therefore that the great continents must have remained separate in those quarters from a very remote epoch to allow for such an extensive independent development of forms. They are both strictly confined to the hottest parts of their respective hemispheres. In America they are not found beyond the northern tropic, nor much further south than 30° S. lat. They are not known to occur so far from the equator as either tropic in the Old World, but are limited to the south-eastern parts of Asia and the islands of the New Guinea group. The genus Danais, with which we have seen both groups are connected, ranges as far north as 41° in Europe, and 45° in North America. It is interesting thus to find that the only genus wliich is common to the three tropical regions is the sole one of the family that occurs in high latitudes. The only means of communication between the intertropical lands of America and Asia seems to have been a circuitous route by the north (or south) ; and the essentially tropical forms do not appear to have passed along it. The fact of the peculiar equatorial Asiatic Danaidce not reaching Africa is explicable on the same grounds as their entire distinct- ness from the American ones, namely, the non-existence of an equatorial connexion of land of a nature suitable for their transit between the two continents since the remote date when the first forms of the group came into being. The habits of the Seliconidce have been described by various travellers, — Lacordaire having given a complete account of the Cayenne species, and Dyson and Gosse some interesting notes on those of Venezuela and Jamaica. The total number of species de- scribed is 284, namely, 233 belonging to the Danaoid, and 51 to the Acrseoid group. They are peculiarly creatures of the forests, and, like the Platyrrhine Monkeys, the arboreal stomoses with the median a short distance from its origin. In the systematic part of the present memoir I shall follow Dr. Felder in this altered classification. The two groups which composed the family Heliconidce are, it must be repeated, completely and widely distinct. Yet the analogical resemblance between them is so great, that some species of the one might easily be confounded (if not closely examined) with species of the other. OF THE AMAZON VALLEY. 499 Gallinacea {Penelopidce and Cracidce), and other groups of the same region, point to the gradual adaptation of the fauna, during an immense lapse of time, to a forest-clad country. I found on the banks of the Amazons 94 species (besides many local varieties, con- sidered by some authors as species) of the two groups (67 Danaoid and 27 Acrasoid), representing all the genera of the family but three. They are most numerous in those parts of the country where the forests are most extensive and the climate most sultry and humid. I found the number of species to increase in travelling from east to west, from the Lower Amazons towards the eastern slopes of the Andes. They were rare in the somewhat drier tract of country which borders the Lower Amazons about the middle of its course. I found in this large district only 26 species, namely, 10 belonging to the Danaoid and 16 to the Acrseoid group. Within an area of about the same dimensions, in the moist region of the Upper Amazons, I collected 64 species, of which 42 were Danaoid and 22 Acrseoid ILeliconidce. I should judge, from the collections received in England from those parts, that the hot Andean valleys near Bogota, or in Ecuador, contain a still larger number of species than the plains of the U|)per Amazons. In the dry forests which clothe a great part of the banks of the Tapajos I found exceedingly few : at one locality, where I collected four months, and which was rich in other families of Lepi- doptera, I saw only one species of the Danaoid and four of the Acrseoid group. Accord- ing to Dyson, many species [Ithomiae) of the lowlands in Venezuela have a vertical range of 2000 feet, and some genera {SymeniUs [^S. Bercetis], Olyras, Etttresis), which do not inhabit the Amazon region, occur at an elevation of 8000 feet. The species are exceed- ingly abundant in individuals wherever they occur : they show every sign of flourishing existence, although of slow flight, feeble structure, unfurnished with apparent means of defence, and living in places which are incessantly haunted by swarms of insectivorous birds. The pathways in the forest near towns are quite enlivened by the multitudes which fly about amongst the lower trees, in their bright dresses of orange, blue, and yellow, and red and black. The mode of flight of the members of the two groups is somewhat different. The Meliconii and Eueides move along in a sailing manner, often cu'cling round for a con- siderable time, with their wings horizontally extended. The species of the Danaoid group, for the most part, keep near the ground, and have a very slow irregular flight, settling frequently. They are all of social or gregarioiis habits. Not only do individuals of the same species congregate in masses, but the set of closely allied species which people a district keep together in one or more compact flocks. I noticed in four districts rich in Danaoid HeUconidce, where I collected, that about half the species of Ithomia flew together in one circumscribed area of the forest, and the other half in a second similar locality, the rest of the tolerably uniform wooded country, in each case, being nearly untenanted by them. The larger species {Seliconii, Lycorece) frequent flowers, probing the nectaries with their proboscides ; but the smaller kinds {Ithomice) , and the members of the Danaoid group generally, are very rarely found thtis occupied : I noticed them sometimes imbibing drops of moisture from leaves and twigs. The fine showy Helicotiii often assemble in small parties, or by twos and threes, apparently to sport together or 500 MR. H. W. BATES ON THE LEPIDOPTERA perform a kind of dance. I believe the parties are composed chiefly of males. The sport begins generally between a single pair: they advance, retire, glide right and left in face of each other, wheel round to a considerable distance, again approach, and so on : a third joins in, then a fourth, or more. They never touch : when too many are congregated, a general flutter takes place, and they all fly ofi", to fall in again by pairs shortly afterwards. The species which I have seen most frequently employed in this way is the Hellcouius Bhea, a glossy blue-black species, with bright yellow belts across its wings. The larvae of the two or three species whose transformations I observed feed together in clusters on the leaves of trees of moderate elevation, near the places where the adult uisects are found. The majority of the species have very limited ranges. I was surprised, when travelling on the Upper Amazons from east to west, to find the greater part of the species of Ithomice changed from one locality to another, not further removed than 100 to 200 miles. For instance, there were 11 of these Ithomice at a place called Eonte Boa, and 9 at St. Paulo, 180 miles distant; but only two of the total number (20) were found in both localities. This is remarkable when we consider that the whole of the country of the Upper Amazons is a nearly level plain, uniformly covered with forest, and offering no perceptible difference in soil or other physical conditions. Five only out of the 20 species have been met with in any other part of South America. The areas of distribution of most of the remaining 15 must be, in each case, a very limited tract of country. The species which inhabit other parts of Tropical America must have sunilarly contracted ranges, if we may judge from the collections received in England from different districts. Now, many of these local species have the appearance of being geographical varieties ; I could not help suspecting them to be such when I met with them in nature, the dif- ferences between the forms of one and those of another locality relating in many cases simply to the colours and colour-patterns of the wings. The marks of distinction, how- ever, are in the majority so well defined, so ordinarily common to all the individuals concerned, and there is so generally an absence of connecting links, that they are held on all hands to be good and true species. Moreover, in those cases already mentioned, where a number of very closely aUied species fly together, they keep themselves per- fectly distinct ; there are no hybrid forms (I am speaking of the IthomicB and allied genera), and on observing individuals in copula, I almost always* found the pair to be precisely the same in colours and markings. In the multiplicity, apparent distinctness, and restricted ranges of the species, this group much resembles the family of Humming- birds of the same regions. I believe, nevertheless, that the suspicion of many of the species being nothing more than local modifications of other forms has proved to be well founded. Amongst the great number of perfectly distinct and weU-marked species, a few occurred T\'liich showed great variability : these, I think, afford a key to the explanation of the origin of the rest. The details of variation will be given under the head of each species : * The exception was in the case of Mechmntis Polymnia, which, as will be seen, on referring to the account of it ill its place, is a polymorphic species, whose local varieties are in an imperfect state of segregation. OF THE AMAZON VALLEY. 501 those which supply the most decisive results are Mechanitis Fohjmnia, ItJiomia Orolina and HUnissa, Ceratinia Ninonia, and the Lycorece. The varieties of these present all the different grades between simple individual differences and well-marked local varieties or races, which latter cannot be distinguished from true species, when two or more of them are found coexisting in the same locahty without intercrossing, as takes place in ItJwmia Illiiiissa and its allies, and probably in 3£echcmitis Nescea and Lysminia. A striking case of the production of a local variety now spread over a wide area, and undistin- guishable from a true species, is afforded in ITeUcomus Thelxiope, to the details of which, given in the systematic part of this memoir, I must refer the reader. These species, when carefully studied, seem to me conclusively to show that many of the now distinct species of HeUconidce have arisen from local varieties, segregated from the variations of preexisting widely disseminated species ; for these distinct forms or species do not essentially differ from the undoubted varieties of the species cited. The genera show different degrees of susceptibility of change under altered local conditions. Thus, many species of Helicouius {H. Bhea, Clytia, Bicini, Vesta, Thelxiope, Antiocha, &c.) are unchanged over the whole of the wide country which includes the areas of several successive local races of many Ithomice and Napeogenes*. The process of the creation of a new species I believe to be accelerated in the Itho- mics and allied genera by the strong tendency of the insects, when pairing, to select none but their exact counterparts : this also enables a number of very closely allied ones to exist together, or the representative forms to live side by side on the confines of their areas, without amalgamating. The course followed by Nature in the formation of these numerous local species, I think, is clearly exhibited in Mechanitis Folymnia, to the details of which, given in its place, I must beg the reader to refer. We see here the manufacture, as it were, in process. The species is widely disseminated and variable. The external conditions in certain locaUties are more favourable to one or more of the varieties there existing than to the others ; those favoured ones, therefore, prevail over the others. We find, in this most instructive case, all the stages of the process, from the commencement of the forma- tion of a local variety (var. Egaensis) to the perfect segregation of one (var. Lysimnia, considered by all authors as a true species). In this species, most of the local varieties are connected with their parent form by individuals exhibiting all the shades of varia- tion ; and it is on this account only that we know them to be varieties. In the species allied to Ithomia Flora, the forms are in a complete state of segregation (with the excep- tion of I. Illinissa, which throws light on the rest), and therefore they are considered as species ; they are, in fact, perfectly good species, like all other forms considered as such in natural history. It is only by the study of variable species that we can obtain a clue to the explanation of the rest. But such species must be studied in nature, and mth * These axe Ithomia Flora, an inhabitant of the whole Lower Amazon region (from the Atlantic to the Rio Negro), which is represented by I. Hippodamia in Cayenne, and I. Onega, Illinissa, Gimilla, Priscilla, Ilerdina, in different areas on the Upper Amazon ; Napeoyenes Cyrianassa, which becomes N. adelphe on the banks of the Cupari (Tapajos), and N. Tunantina on the north bank of the Upper Amazons ; N. Inachia, which is changed to N. sulplmrina at Bahia, and to N. Ercilla, N. Coi-ena and N. Pharo in different areas on the Upper Amazon. Other species might be added in confirmation. Most of the species of Heliconius quoted are found unchanged over the collective areas of all these forms of Ithomia and Napeogenes. 502 MR. H, W. BATES ON THE LEPIDOPTERA strict reference to the geographical relations of their varieties. Many closet natviralists, who receive disconnectedly the different varieties in any group, treat them aU as inde- pendent species: by such a proceeding, it is no wonder that they have faith in the absolute distinctness and immutability of species. The sexes in the Hellconidce very rarely differ in colours. Secondary sexual characters of another description occur, however, very generally in the Danaoid group. The males, in all the genera but two {Lycorea and Ituna) of this section, are furnished with a pencil or fringe of long hairs near the costal edge of the hind wings on the upper sur- face. It sometimes arises from the bottom of a shallow horny cup situated between the costal and subcostal nervures ; the hairs are long, soft, and adpressed. I was unable to discover any use in this structure ; it seemed not to be under the control of the insect. There is no movement in flight, or position in repose, peculiar to the male sex, which might require an instrument to hold the wings together — a function which the position of the hairs, in the place where the fore wing overlaps the hind wing, suggests to the mind. I believe the appendage must be considered as an outgrowth of the male organization, which is not in this case appHed to any especial purpose : it may be taken to be of the same nature as the pencil of hairs on the breast of the male Turkey. Growths of one kind or other, on the surface of the wings, peculiar to the male sex, are frequent in Butterflies : in Danais the males have a small horny excrescence on the disk of the hind wings, which, considering the near relationship proved to exist between the two groups, I take to be homologically the same as the pencil of hairs in the Danaoid Seli- conidce. In the genus Favonia, belonging to the family Brassolidce, the males in some species have a fringe of hairs near the abdominal border ; in others, a long pencil of the same on the disk ; and, again, in others, instead of these appendages, a thickened plate on the inner margin of the hind wings. The most interesting part of the natural history of the HeliconidcB is the mimetic analogies of which a great many of the species are the objects. Mimetic analogies, it is scarcely necessary to observe, are resemblances in external appearance, shape, and colours between members of widely distinct families : an idea of what is meant may be formed by supposing a Pigeon to exist with the general figure and plumage of a Hawk. Most modern authors who have written on the group have mentioned the striking instances of this kind of resemblances exhibited with reference to the SeliconidcB ; but no attempt has been made to describe them fully, nor to explain them. I will give a short account of the leading facts, and then mention some circumstances which seem to throw light on their true nature and origin. A large number of the species are accompanied in the districts they inhabit by other species which counterfeit them in the way described. The imitators belong to the following groups : — JBapilio, Fieris, Euterpe, and Leptalis (fam. Fapilionida;), Protogonius {Nym- phalidce), Ithomeis {JErycinidcB), Castnia (Castniadce), Dioptis, Fericopis, Hyelosia, and other genera {BombycklcB Moths)*. I conclude that the ILeliconidcB are the objects imitated, because they all have the same family facies, whilst the analogous species are dissimilar to their nearest allies— perverted, as it were, to produce the resemblance, from * The accompanying Table, in which a number of the most striking of these are arranged in parallel columns, will 2;ive some idea of the extent to which this system of imitation prevails. OF THE AMAZON VALLEY. 503 Pntq _gfi O «4H i4 I :« r-1 t„ O ^ O . pi g ° ^ ph ^^ :^ =^;^ 1^ ^ iz; 1-1 .^ ^ jj'tSo