m IMAGE EVALUATION TEST TARGET (MT-3) .i^-i 1.0 I.I 1.25 ^ ^ 12.2 u 1.4 2.0 1.8 III 1.6 V <^ /; Hiotographic Sciences Corporation 23 WEST MAIN STREET WEBSTER, N.Y. 14580 (716) 872-4503 €^r i/.A ^ CIHM/ICMH Microfiche Series. CIHM/ICMH Collection de microfiches. Canadian Institute for Historical Microreproductions / Institut Canadian de microreproductions historiques Technical and Bibliographic Notes/Notes techniques et bibliographiques M p The Institute has attempted to obtain the best original copy available for filming. Features of this copy which may be bibliographically unique, which may altar any of the images in the reproduction, or which m£/ significantly change the usual method of filming, are checked below. Coloured covers/ Couverture de couleur Covers damaged/ Couverture endommag6e Covers restored and/or laminated/ Couverture restaur^e et/ou peiliculAe Cover title missing/ Le titre de couverture manque I I Coloured maps/ Cartes giographiques en couleur □ Coloured inic (i.e. other than blue or black)/ Encre de couleur (i.e. autre que bleue ou noire) □ Coloured plates and/or illustrations/ Planches et/ou illustrations en couleur Bound with other material/ ReilA avec d'autres documents Tight binding may cause shadows or distortion along interior margin/ La reliure serr^e peut causer de I'ombre ou de la distortion le long de la marge intirieure Blank leaves added during restoration may appear within the text. Whenever possible, these have been omitted from filming/ II se peut que certaines pages blanches ajoutAes lors d'une restauration apparaissent dans le texte, mais, lorsque cela 6tait possible, ces pages n'ont pas 6tA filmies. Additional comments:/ Commentaires supplimentaires; D D L'Institut a microfilm^ le meilleur exemplaire qu'il lui a At6 possible de se procurer. Les details de cet exemplaire qui sont peut-Atre uniques du point de vue bibliographique, qui peuvent modifier une image reproduite, ou qui peuvent exiger une modification dans la methods normale de filmage sont indiqu6s ci-dessous. □ Coloured pages/ Pages de couleur □ Pages damaged/ Pages endommag6es □ Pages restored and/or laminated/ Pages restaur6es et/ou pellicul6es r~~| Pages discoloured, stained or foxed/ D D Pages d6color6es, tachet^es ou piqu6es Pages detached/ Pages d6tach6es Showthrough/ Transparence Quality of prir Quality inigale de I'impression Includes supplementary materii Comprend du materiel suppi^mentaire I 1 Pages detached/ I I Showthrough/ r~~| Quality of print varies/ I I Includes supplementary material/ Only edition available/ Seule Mition disponible Pages wholly or partially obscured by errata slips, tissues, etc., have been refilmed to ensure the best possible image/ Les pages totalement ou partiellement obscurcies par un feuillet d'errata, une pelure, etc., ont 6t6 filmAes A nouveau de fagon d obtenir la meilleure image oossible. D This Item is filmed at the reduction ratio checked below/ Ce document est filmA au taux de reduction indiquA ci-dessous. 10X 14X %m 22X 26X 30X 12X 16X 20X 24X 28X 32X f e §tails 8 du lodifier r une Image The copy filmed here has been reproduced thanks to the generosity of: Library of Congress Photoduplication Service The images appearing here are the best quality possible considering the condition and legibility of the original copy and in keeping with the filming contract specifications. L'exemplaire film6 fut reproduit grdce d la g6n6rosit6 de: Library of Congress Photoduplication Service Les images suivantes ont 6t6 reproduites avec le plus grand soin, compte tenu de la condition et de la nettetd de Texemplaire filmd, et en conformity avec les conditions du contrat de filmage. Original copies in printed paper covers are filmed beginning with the front cover and ending on the last page with a printed or illustrated impres- sion, or the back cover when appropriate. All other original copies are filmed beginning on the first page with a printed or illustrated impres- sion, and ending on the last page with a printed or illustrated impression. IS Les exemplaires originaux dont la couverture en papier est imprim^e sont film^s en commengant par le premier plat et en terminant soit par la dernidre page qui comporte une empreinte d'impression ou d'illustration, soit par le second plat, selon le cas. Tous les autres exemplaires originaux sont film^s en commen9ant par la premidre page qui comporte une empreinte d'impression ou dlllustration et en terminant par la dernidre page qui comporte une telle empreinte. The last recorded frame on each microfiche shall contain the symbol •— »> (meaning "CON- TINUED"), or the symbol y (meaning "END"), whichever applies. Un des symboles suivants apparaitra sur la dernidre image de cheque microfiche, selon le cas: le symbole -^ signifie "A SUIVRE ', le symbole V signifie "FIN". I Maps, plates, charts, etc., may be filmed at different reduction ratios. Those too large to be entirely included in one exposure are filmed beginning in the upper left hand corner, left to right and top to bottom, as many frames as required. The following diagrams illustrate the method: Les cartes, planches, tableaux, etc., peuvent dtre film6s d des taux de reduction diff^rents. Lorsque le document est trop grand pour dtre reproduit en un seul clich6, il est film6 d partir de Tangle supdrieur gauche, de gauche d droite, et de haut en bas, en prenant le nombre d'images ndcessaire. Les diagrammes suivants illustrent la mdthode. errata to pelure, >n d n 32X 1 2 3 1 1 2 3 4 5 6 McGILL UNIVERSITY PAPERS FROM THE DEPARTMENT OF Botany. No. 8 — Notes on the Development of the Holdfasts OF Certain Floride^e. I ft! 1 I 'I lb: I h BY CARRIE M. DERICK, M.A. With Thseb Platks. [Reprinted from the Dotanical Gazette, Vol. xxviii, October, 1899, PP- '47*'^4-] Montreal, 1900. 'i DERICK on HOLDFASTS OF UHOD PI. A Hi XXl K on HOLDFASTS OF UHODOPHYCE/E - '•"■ RICK on HOLDFASTS OF RHODOPHYCEyE noiAxiCAi. c,A/.i:ri i:. xwm I DKRICK on HOLDFASTS OF RH 1:...-. 1 ^ICK on HOLDFASTS OF RHOUOPHYCE.'E a..„_. in "■•i^eMnpM NOTES ON T-HE DEVELOI'MENT OF THE HOLDFASTS OF CERTAIN FLORIDE/E. C A K K I K M. 1) K K I <■ K. (\V!TII rLATKS XXI-XXIII AND KIVK TKXT-KKiUKES) With the exception of passing references in various works, two articles, the one bv Borge (l), the other by Stronifelt (8), include, I believe, all that has been written upon the holdfasts of the alg.u. The former deals with a few members of the Chloro- phyceiu; the latter is very comprehensive, but it is without illus- trations and gives no specific details. Therefore, the study of the develo])mcnt of the holdfasts of some nearly related species of the Rhoclophvce.-E has seemed advisable. The observations described in this paper were made at the Marine Biological Laboratory, Woods Hole, Mass., during the summers of 1896 and 1897, '^"'^^ *^he worl: was finished in the Botanical Laboratory of McGill College. Cultures of the spores of several species were made under various conditions. Ordinary glass object-slides were placed in flat porcelain dishes, either white or painted black. The vessels were filled with filtered sea-water and in them were laid plants bearing ripe spores. The spores usually sowed themselves in twenty-four hours, in which case the plants were removed. Some of the dishes were fed by a gentle stream of running water, others were disturbed only three times a day, when the water was drawn off by a siphon and replaced by filtered sea-water. The latter method proved much the better of the two. The color of the background had no effect upon the develop- ment of the spores, nor were the plantlets heliotropic. Cultures kept in a shaded place flourished best, even a short exposure to direct sunlight killing plantlets. It is to be regretted that no record of variations in the temperature and the density of the water was kept. Oltmanns (6) has conclusively shown that such ^*Tffl999^1& - ii i wmiiiiii i« W i i» i w i iiif i . i «im.W'i>J>' 2 4.S BO TANIC.l I. GA/.E TTE |()i roiiER variations are most im])ortant factors in the life and distribution of marine alga.-, but a comi)arison of many cultures of the same species indicates that differences in the conditions which pre- vailed in the laboratory affected neither the form nor the order of development of the plantlet. It was difificult to keep the young i)lants healthy for more than three or four weeks ; and during a few davs of intense heat, in August i8g6, all ceased to grow, were attacked by bacteria, and finally died. Both carpospores and tetras|)ores were germinated and most closely resembled one another in their develoiMTient. In some cases, however, tetrasjjores attached themselves to the substratum less readilv than carpospores. This is regarded by Brannon (2) as an adajjtation to the immediate distribution of the species, Every day, slides upon which spores were growing were examined and drawings of living plantlets were made. Perma- nent mounts for com])arative study were prepared at regular intervals and mature holdfasts were preserved in alcohol or in 5 per cent, formalin in sea-water. The species selected for investigation belonged to the Rho- dvmeni;des, with the exception of one member of the Rhodo- phyllidaceit, RhuMonia tcucra J. Ag. The carpospores of this alga attach themselves firmly to the slides in nine hours; a very delicate outer layer, probably of mucilage, may be observed, but can hardly be distinguished from the cell-wall [PL XXI, fig. i) . In a few cases, an irregular layer of a coarsely granular substance surrounded the plantlets, and in one, a thick mucilaginous disk was developed at the base ; these appearances were most excep- tional, and were doubtless due to slightly abnormal conditions {PI. XXI, figs. 7, 8). After attaching themselves, the spores immediately enter upon a segmentation stage, and within twenty-two hours two divisions are made. The first separates the spore by means of a vertical wall into two equal cells (PI. XXI, figs, 2,j). Three and four-celled stages result from the successive division of the two primary cells in vertical planes at riglit angles to the first wall ; other cells are cut off from these by oblique walls, and thus an irregular spherical mass is formed. 1 ivMgw^»^' ■*• ''^•^''WitwagiifeWMiWiri^gW'Mg^i-" (oi roiiER tribution :hc same lich pre- he order cec[i the iks ; and :eased to iiid most In some bstratum Bran no II ; species, ing were Perma- t regular hoi or in the Rho- : Rhodo- s of this 5 ; a verv ;rved, bnt r,fig.i). substance nous disk St cxcep- onditions le spores d within senarates 1 |ual cells suit from :al planes rom these s formed. 1899] DEVELOPMENT OF HOLDFASTS OF FLOklDEA: 249 In about thirteen days differentiation begins. Four basal cells elongate to foim the primary root-cells, the position of which bears no definite relation to the direction of the supply of light (/v. XXI, figs. 7, 8). Occasionally only one or two pri- mary rhizoids are thus formed {PL XXI, figs, g, 10, //). Unfortunately, the plantlets died at this stage, and only by analogy could the subsequent history of the holdfasts be determined. However, the close resemblance between the pri- mary root-cells and mature holdfasts of Rhabdonia and of the two species to be considered next justifies the assumption that the intermediate stages closely agree. It is interesting to note that, according to Osterhout (7), the earliest divisions in the tetraspores which give rise to proliferations do not succeed one another in the definite order just described, but the spores divide " by means of oblique walls which do not occur in regular suc- cession." Very similar to that of Rhabdonia is the early history of Lomentaria tincinata Menegh. and of Cliampia parvida Harv. The carpospores of Lomentaria, after secreting a wall, attach them- selves to the substratum by a very thin mucilaginous secretion {PL XXI, fig. 12), which sometimes persists after the first rhizoids have been formed [PL XXI, fig. 75). The segmenta- tion proceeds rapidly, but primary root-cells do not arise until a later date than in Rhabdonia (/V. XXI, figs, ij, 14). The four basal root-cells, dividing by horizontal walls, become short, closely-appressed filaments, which soon branch pseudo- dichotomously {PL XXI, figs. 15, 16) and form a discoid holdfast. Stromfelt (8) regards the branching of the rhizoidal filaments as dichotomous and says that the increase in circumference of the mature holdfast is du'^ almost entirely to marginal growth. Vege- tative reproduction may take place not only by the mechanical separation of the various upright branches, which spring adventi- tiously from the surface of the holdfast [PL XXI, fig. 22), but by means of stolons, which are sent out in various directions and form secondary holdfasts [PL XXI, fig. 21). Davis's account (4) of the plantlets of Champia parvtda. Ifcr MWJllM »-■* w '. i -nnami i MiLj i n i ^'.'. I i jm— . .-^v- ij^jij. ii 2 5° BOTANICAL GAZETTE [OCTOBER developing under natural conditions, accurately describes those grown in the laboratory. A somewhat spherical mass composed of sixteen cells results from the segmentation of the spore, then the first indication of a permanent holdfast appears as a slight projev tion of each of the four basal cells. The primary root- cells, dividing at right angles to their longest axis, form two- celled filaments, which branch monopodially and give rise to a broad spreading holdfast. In section the mature holdfast appears parenchymatous \Pl. XXI, figs. //, i8), but it is often possible to distinguish the component filaments (/V. XXI, fig. 20). Exceptional plantlets produce one instead of four primary root- cells, but the later stages conform to the type {P/. XXI, fig. 19). In all cases the cells of the holdfast are paler than those of the frond, and the chromatophores of young specimens arc in close contact with the walls. From the foregoing it will be seen that Rhabdonia, Lomen- taria, and Champia agree {a) in passing through a segmenta- tion stage, resulting in a somewhat spherical mass of cells, {b) in the elongation of four basal cells, and {c) in the subsequent development of four primary rhizoids, which branch repeatedly and finally form a large discoid holdfast, composed of pseudo- parenchymatous tissue. In marked contrast are several members of the Rhodome laceiE. The s]jores of Chondria tenuissima (Good, et Wood.) C. A.g. and of Chondria dasyphila Ag. germinate very readily. The early divisions of the spores and the history of the development of the rhizoids are alike in the two species. The spores, which are large and have coarsely granular contents, quickly and firmly attach themselves to the substratum, doubtless by means of a thin and uniform layer of mucilage, though no secretion distinct from the cell-wall is perceptible {PI. XXI, fig. 23). They soon divide into two unequal cells sej)arated by a slightly con- cave wall ; divisions |)arallel to the first follow {PL XXI, fig. 24), and the basal cell of the resulting filament elongates into the prim.iry rhizoid {PL XXI, figs. 25, 27) . Sometimes the basal cell seems to branch dichotomously ; but, as the branches do not [OCTOBER 1800] DEVELOPMENT OF HOLDFASTS OF FLOKIDE.F. 251 )es those om posed ore, then s a slight ary root- orm two- rise to a holdfast is often \ fig. 20). ary root- . fig- 19) • se of the ; in close , Lomen- egjmenta- cells, (<^) ibsequcnt ipeatedly E pseudo- [.hodome /ood.) C. ily. The elopnient es, which ckly and by means secretion r). They htly con- \fig- 24), 5 into the basal cell :s do not arise simultaneously nor are the\ separated by a wall parallel to the longer axis of the basal cell, it is evident that the branching of the filaments is strictly monopodial {PI. XXI, figs. 26, J2, jj) . The upper cells of the plantlet soon cut off pericentral cells, and at the same period the first rhizoid divides repeatedly so as to form a multicellular, monosiphonous filament {PL XXI, figs. 28, jj). Secondary rhizoids are next developed, either as out- growths from the primary root-cell or from a pericentral cell imme- diately above the basal cell (/*/. XXI, fig. ji). Long slender rhizoids may be formed before a clasping-disk arises, but, both in plantlets developing in the laboratory {PL XXI, figs. 2g, 33,36) and in those growing in a state of nature {PL XXI, figs. 30, 38 3g, 40), the primary rhizoid generally remains very short and by means of oblique walls produces discoid holdfasts at the tip. The size and efficiency of this holdfast are increased by means of free secondary rhizoids {PL XXI, fig. 33) or by intracuticular filaments {PL XXI, fig. 34), both of which have their origin in pericentral cells near the base of the plantlet. The secondary rhizoids develop in the same maimer as the pi mary and com- bine with them to form a large, irregular, discoid holdfast, in which it is possible to distinguish few of the component fila- ments {PL XXI, fig. 37) . The mature holdfast, though adhermg closely to its host, does not penetrate it, but a cutinization of the cortical layers often results from the contact. As in Lomen- taria, etc., the large holdfast gives rise to several adventitious branches ; and great powers of vegetative reproduction are implied bv the stores of floridean starch with which the cells are often charged. The chromatophores are closely aggregated in the peripheral protoplasm, especially next to the inner radial and the lateral walls, and often indicate the plane in which a wall will soon be formed. They are large and discoid, thus resembling those of the cortical cells of the frond rather than the anastomosing filamentous chromatophores of the central cells. Very different from the short, multicellular, primary rhiz- oids of the Chondrieai are the primary holdfasts of the Poly- siphonicjE. The form taken as a type of the latter group is -fe.-. 252 BOTANICAL GAZETTE |OCI(niKR Polysiphonia violacea (Roth) Grev. In a few hours after they are sown, both carpospores and tetraspores attach themselves to the substratum by a coarsely granular, mucilaginous secretion, which completely covers the spore {PI. XXL fig. 4^) ■ I" optical section this envelope appears densely granular at the margin, while a clear amorphous area intervenes between the outer layer and the developing plantlet {PL XXI, fig. 44). The spore divides into two unequal cells, of which the smaller soon becomes slightly pointed, and finally grows into the primary root-cell {PI. XXI, fig. 42). Divisions parallel to the first ensue {PI. XXI, figs. 4^-43), and the basal-cell elongates so as to form the first rhizoid piercing the mucilaginous sheath, which finally disinte- grates and disappears, A six-celled stage is often reached before divisions in planes at an angle to the first occur. Generally, segmentation continues for some time and the siphons are clearly marked off from the central axis before a second rhizoid arises. As in Chondria, this has its origin in an unsegmented cell adja- cent to the primary root-cell. The former differs from the latter only in its brighter color and denser cell-contents ; and, remaining undivided, it forms a component part of the holdfast {PL XXII, fig. 3). The protojjlasmic connections between these cells are obvious, but they are difficult to trace between the other cells of the plantlet {PL XXI, figs. 45, 47). When three or four weeks old the young plant develops several rhizoids spring- ing either from the primary or secondary root-cell {PL XXII. fig. 7) . Occasionally, however, the segmentation stage of the sjjore ends and the growth of the frond begins before multipli- cation of the rhizoids takes place {PL XXII, fig. 10). At the base of the frond other rhizoids are sent out by corticating cells and are separated from these by a wall, a protoj)lasmic connection being maintained. No intracuticular filaments arc developed and the indejiendence of the rhizoidal constituents of the mature holdfast is practically preserved {Pi. XXII, fig. 6). In addition to the rhizoids near the base of the plant, any corticating cell of a procumbent branch may produce a secondary holdfast and thus assist in the extension of the colony. As Stromfclt (8) noted. Bt aii ! ti .» , uft * iiw sffi! a i (OClCIiKR ifter they nselves to secretion, In optical e margin, uter layer re divides becomes t-cell [PL {PL XXI, \ the first y disinte- led before jenerally, ire clearly lid arises, cell adja- froni the ;nts ; and, e holdfast veen these twcen the •n three or ds spring- TV. XXII. .ge of the ; niultipli- ) . At the iting cells ;onnection developed he mature n addition ig cell of a t and thus (8) noted. 1 899 1 DEVELOPMENT OF HOLDFASTS OF FLOA'JDE.F. 253 all the rhizoids are unicellular and unbranched ; and, although an apparent tendency to branch may be observed occasionally, the lobes are not separated from the main portion by a wall, even the protoplasmic contents being undivided [PL XXII, fig. g). All the rhizoids eventually develop terminal clasping-disks. The first indication of such a structure may appear in plantlets four days old, but, as a rule, the primary rhizoids do not undergo modification until several days later. The disks begin as a sim- ple enlargement of the tip of the rhizoid {PL XXI, figs. 48-30; XXII, fig. /), become deeply lobed, and assume a very irregular outline {PL XXI, figs. 32-36). The cell-contents of the rhizoid extend into the lobes, but no division takes place {P/. XXII, fig. 8} . Great variations in the length attained by the primary rhizoids occur both in plants grown under natural conditions and in laboratory cultures {PL XXI figs. 46, 48; XXII fig. /). The cause of such variations has not been determined, but it is prob- able that contact irritation may be the most important factor in the formation of the disks. This view is supported by the fact that disks are sometimes produced on the sides of rhizoids when these come in contact with a firm substance {P/. XXII, figs. 4, 3). The length of the secondary rhizoids depends upon the distance of the parent cells from the substratum, and as soon as contact is established broad clas]Mng-disks are formed, which mechanical] v cohering with one another and with the primary disk produce a very strong holdfast. The rhizoids are paler in color than the rest of the plantlet, having less dense contents and fewer chromatophores. The rhizoids of Polysiphonia violacea never penetrate the host-plant ; but at the point of contact the surface of the latter is often dark brown and cutinized, while the outer cortical cells are destitute of chromatophores. But incipient parasitism occurs in Polysiphonia fijstigiata Grev., growing on Ascophyllunt nodosum Stack. Gibson (5), in writing of the histology of this sjjecies, noticed that "the attachment of the epiphyte to Ascophyllum is very intimate. Root filaments given off from the base of the frond penetrate deeply into the tissue of the host, and wander 254 HOTANICAI. GAZETTE OCTOBKR amonjr the cortical cells and medullary hyph.e. The root-fila- ments have very thick cell-walls and central cells only, these being much elongated." The ends of the rhizoids are swollen and in close contact with the eel's of the host [text figs. 4, 5, p. 256), but no haustoria penetrate the walls of the latter. In one instance a unicjue variavion occurred. A few intracuticular fila- ments, descending from the corticating cells of the Polysijihonia, ran parallel to the main axis of the rhizoid throughout its length. The host suffers no serious injury, only a depression and cutini- zation of the surface with a very slight disorganization of the cortical cells at the point of penetration. Though the associa- tion of the two plants does not justify the assumption of com- plete parasitism, the symbiotic relation existing between them is much more intimate than that observed between Ascophyllum and anv of the truly ei)iphytic algae. According to Brebner (3), a similar relation exists between Dumontia filiformis anl its host, Fuciis scrratus. Dasya degans (Martens) C. Ag. was the third species of the Rhodomelaceae examined. The spores attach themselves by a mucilaginous secretion much less definite in form and less persistent than in Polysiphouia violacea [PL XXII, fig. jo). The spore elongates before the division takes place, and in many instances the very young plantlet assumes an hour-glass shape, occasionally with a delicate mucilaginous secretion at either end. The first walls are parallel to one another, forming a long fila- mentous body, of which one terminal cell becomes the apical cell of the frond, the other the primary root-cell [PI. XXII, figs. 31, J2, jj,jj, j6). Twelve or more parallel divisions may occur before the basal cell elongates and forms a rhizoid terminating in a disk ; but such a modification may a])iJcar at an earlier period {P/. XXII, fig. 34). The multicellular disks, which arise sooner or later, are like those of Chondria. The root-cell or the end of a rhizoid broadens and becomes slightly lobed ; oblicpie walls cut off the lower corners of the cell ; division is continued and a multicellular disk with a mucilaginous margin results {PL XXII, figs. 39-41 ; XXIII, figs. 2,3, 7,8,9). The primary root 1 liiiiiiini \ffi\ iiiij '1 i|iiiMmHJ!IC>ii^'"''i ill ItlHliiH^ii II T [OCTOBKR root-fila- ly, these ; swollen •• 4, 5- P- In one :ular fila- siphonia, ;s length. id cutini- n of the associa- of com- n them is jphyllum bner (3), its host, es of the lelve.s by and less u)'. The in many ss shape, ither end. long fila- ipical cell \ figs- 31 < lav occur •minating n earlier lich arise :ell or the ; obliepie :ontinued suits {PL lary root iSggJ DEVELOPMENT OF HOLDFASTS OF FLORLDEA: 255 cell often branches, each portion giving rise to a disk {PI. XXII, fig. 42; XXIII, fig. 4). Although the primary rhizoid usually remains short, in some instances it attains considerable length before undergoing division or forming a disk {PL XXII, figs, jj, j8) . While these changes are taking place, the cell adjacent to the basal cell sends out rhizoids similar to those arising from the primary root-cell {P/. XXIII, figs. /, 3, 6). These various root- filaments combine to form the primary holdfast, which is after- wards strengthened by multicellular branching rhizoids, springing from the basal corticating cells of the frond. The course of the filaments may be traced for some distance in the holdfast, but it is difficult to distinguish between those cells which have their origin in the primary disk and those which are derived from the corticating filaments. The difficulty in determining the relation- ship of the parts is increased by secondary lateral connections, which are developed between the corticating cells {PL XXII, fig. 4j). The marginal cells of the mature holdfast are larger and broader in proportion to their length than the corticating cells of the fronds, and have denser cell-contents, but the chromato- phores of both are separate disks, while those of the central siphon are the anastomosing filaments characteristic of many of the Rhodomelaceae. A creeping tendency may be exhibited at an early age, very young plantlets sometimes developing two distinct holdfasts of almost equal importance (/V. XY7//,/^. ^). As in the other species described, many branches arise from the massive rounded holdfast, probably springing adventitiously from the surface of the latter. A comparison of the three species of the Rhodomelaceae described will show that they agree in forming a primary root- cell, which elongates into a rhizoid terminating in a clasping disk ; and in developing secondary rhizoids, which are sent out by the root-cell, the cell adjacent to it, and the cortical cells at the base of the frond. But, while the rhizoids of Polysiphonia are uni- cellular, unbranched, and free, those of Dasya and Chondria are multicellular, branched, and aggregated into a compact cell-mass, which in section resembles parenchymatous tissue. lMWdM»'<«fHM«K *«etMn ■ '•'y> »ii T 256 BOTANICAL GAZETTE [OCTOBER The species remaining for consideration belong to the Cera- miaces. Cultures of Spermothamnion Turueri Arcsch. were unsuc- cessful, only one healthy plantlct having been obtained. It developed two small disks, the one a terminal primary holdfast, the other a secondary structure arising from one of the middle cells of the filament (/V. XXIII, Jig. 41). As Stromfelt has pointed out, the mature rhizoids are short, unicellular, unbranched organs, terminating in a lobed disk in which delicate threads of protojjlasm can be traced {PL XXIII, Jigs. 42, 4j). The rhizoids are especially abundant near the base of the plants, but, as is well known, any cell of a procumbent branch may give rise to these simple bodies. Now and then, an ordinary branch of the frond produces at its apex a much-lobed twisted disk, which closely embraces a neighboring branch as a tendril would a sup- port (P/. XXIII, Jig. 40). Very few spores of Griffithsia Rornctiana Farlow germinated, and the resulting plantlets were short-lived. It is, therefore, impossible to describe each step in the development of the hold- fast. Two unequal cells arise from the first division {text Jig. /), a monosiphonous filament composed of rather globose cells is formed, the basal cell of which elongates and becomes the root-cell {text Jig. 2). The structure of mature holdfasts would lead one to suppose that the primary root-cell divides repeatedly, forming a broad sjjreading mass of large-celled pseudo-parenchymatous tissue. The holdfasts of a young plant attached to a small piece of bone [text Jig. j) and of older plants growing on Zcsten? mnrina L. differed only in size, that of the latter being very large with an abundance of adventitious branches arising from its margin. Stromfelt (8) places the genus Griffithsia with Ceramium and Callithamnion in a group distinguished by branching rhizoids ; but, though many mature specimens of Griffithsia Bornetiana were examined, no such secondary formations were seen. The cells of the holdfasts are brilliantly colored and have very dense granular contents, prob- ably due to a large supply of reserve food-material. A closely allied genus, Callithamnion, differs greatly from N I -._!: 1899 1 DEVELOPMENT OF HOLDFASTS OF FL0K/DE.1: 257 Griffithsia in the form and history of the holdfast. Though the spores of Callithamnion Borreri Ai(. deveIo|)ed in the laboratory with difficulty, cultures sufficient to illustrate the order of devel- opment succeeded. After attaching themselves to the slide by an almost imperceptible secretion, the spores elongate and become pointed at both ends. The first division is parallel to the shorter axis {Pl. XXni,fig. 2g), and by subsequent partition a Flc;. I. Griffiithsia Bornetiana Karlow, germinating tetraspoie Q days after sow- ing. X 333- Fig. 2. G. Bornetiana, plantlet 9 days old. X 213. Fig. 3. G. Bornetiana, holdfast of small pl.T>t ^^rowing on bone. X S3- Figs. 4, 5. Polysipkonia fastigiaia Grevi'ie, groxrmg on Ascopliylliim nodosum Stack.; in fig. 5 the cells of P. fastigiatn are shaded in, those of Ascophyllum are merely outlined. X 213. monosiphonous filament is formed. The smaller of the two ter- minal cells is the primary root-cell {PL XXIII, fig. 30) . Judging from one set of cultures, the first cell of the holdfast, which has a very thick mucilaginous wall, is cut off from the swollen w= »,f » m *Mm aSS BOTANICAL GA/.LTTh. [OCTOBKR extremity of the root-cell and sej)arated from it by a wall {PL XXin, figs.Ji-JJ.J7,j8) . A multicellular holdfast results from the division and branching of the primary root-cell and of the oldest cell of the disk (PL XXIIl,figs. J4,j6), and is further strengthened by a few corticating rhizoidal filaments. In mature Callithamnion Baileyi Harv. the intracuticular rhizoids are more numerous. They arise from the lower angles of the central cells or from the basal cells of the branches. Thence they descend through the walls of the monosiphonous frond to the holdfast, where they branch freely and sjjread out in various directions, forming, with the filaments arising from the cells of the primary disk, a flat circular holdfast {PL XXIII, figs. J5, jg). The carpospores of Spyridia filamentosa (VVulf.) Harv. germi- nate very readily. The mucilaginous secretion, by which the spores are first fastened to the substratum, can hardly be dis- tinguished from the wall of the spore {PL XXII, fig. ii). The spore first forms two unequal cells, the smaller of which becomes the primary root-cell, the larger divides by parallel walls so as to form a short filam.ent {PL XXII, figs. 12-14). As a rule, no long primary rhizoid is formed, but occasional exceptions are found {PL XXII, fig. f8). Though variations may occur {PL XXII, fig. 75) , the first corticating cells are usually cut off from the upper angles of the central cells before the formation of the discoidal holdfast begins (/*/. XXII, fig. 16). Sooner or later, however, the primary root-cell divides in several planes parallel to the longer axis of the spore, and thus produces a flat multicellular disk, the cells of which are separated by walls {PL XXII, figs. 15, 17, ig, 20, 21). At first the primary disk is but slightly lobed, but the cells soon give rise to short filamentous outgrowths, which branch pseudo-dichotomously {PL XXII, figs. 21, 22). Rarely a lateral disk springs from the primary root-cell, which is then prolonged into a filament {PL XXII, fig. 2j). The cells of the primary hold- fast continue to divide, and with them are combined rhizoi- dal filaments, which have their origin in the corticating cells of the lower nodes of the frond and grow through the cell-walls to the substratum {PL XXII, figs. 24, 25). Free filamentous 1899 I DEVELOPXENT OF IIOLDEaSTS OF EI.OKIDE.-E 259 outgrowths ot the lander cortical cells develoi) ilisks and enter into close mechanical union with the main holdfast, ad(lin,t( ,u[reatly to its efficiency. The secondary rhizoids thus formed are destitute of cortications {PL XXII, fis;. 2q) and are easily distinguished from those ordinary branches which, coming in contact with the substratum, develop holdfasts {PL XXII, figs. 26, 28(7). Spyridia creeps not only by means of such branches but by the aid of masses of uncorticated hair-like rhizoids, which may be formed at any point {PL XXII, fig. 28b). One curious instance was noted of short rhizoidal outgrowths from each of the central cells of a trailing branch, the cortical cells being unmodified {PL XXII, fig. 2j). The differences in the chromatophores are similar to those noted in the Rhodomelacea.-. Ceramium differs in many respects from the other genera described. Both Ceramium rubnim (Huds.) C. Ag. and C. stric- tum Harv. were carefully studied and found to agree closely in so far as the early development of the plantlet and the forma- tion of the rhizoid are concerned. Shortly after it is sown, the spore produces a cell-wall {PL XXIII, fig. 24) and a well-defined temporary holdfast. The latter is a granular disk of definite thickness, and mucilaginous in character, attached to the base of the spore {PL XXIII, fig. 10) . The granules, embedded in a clear matrix, are often arranged in lines radiating from the axis to the rim of the disk {PL XXIII, fig. 25). The distances between the granules being least in vertical planes, the disk appears densest when viewed from the side. This peculiar body does not respond to the ordinary test for cellulose, and is not dissolved after pro- longed treatment with dilute potassium hydrate. It is affected neither by Hanstein's aniline blue nor safranin ; but the granu- lar portions stain deeply with haimatoxylin and with Congo-red and the whole with Bismarck-brown. The disk is, therefore, distinct from the cellulose wall of the spore, and differs materi- ally from ordinary vegetable mucilages, though it is probably closely allied to the latter. This temporary holdfast is not peculiar to plants existing in an unnatural environment, but has been found in very young plants growing on Chordaria. As the jwu s6o BOTANICAL GAZETTE [OCTOBER spore develops, the first rhizoid pierces the disk, which then becomes disintegrated and finally disappears (/-V. XXIII, fig. 26). After attaching itself in the manner described, the spore elongates and divides several times in parallel planes at right angles to its longer axis {PL XXIII, figs, ii-ij). The basal cell, growing rapidly, produces a multicellular rhizoid at an early age; but large jjlantlets, which have already cut off corticating cells, occasionally show little or no tendency to form rhizoids {PI. XXIII, figs 14, 1$, iS). The primary root-cell branches into several rhizoids, which are increased in number by outgrowths from the cell adjoining the first root-cell [P/. XXIII, figs, z/, /p, 20, ^7). Still later, the cortications near the base of the plantlet develop multicellular branching rhizoids of great length. All remain free throughout the life of the plant, and both primary and secondary rhizoids branch monopodially near the tip, and thus give rise to large multicellular disks of irregular outline (77. XXIII, figs. rj,22, 26, 28). These indented clasping-disks are closely crowded together, cohering so as to form a large rounded holdfast, in which the various elements may be clearly distinguished. As both of the species arc upright in habit, no secondary holdfasts are developed at any point of the mature frond. As in several other genera, the chromotaphores of the plantlets and of the holdfasts resemble those of the corticating cells rather than those of the central axis, the former being disks, the latter irregular branching bands [PI. XXIII, figs, ij and 21). It is evident, therefore, that the species of the Ceramiaceae examined differ greatly both in the manner of development and the form of the holdfast, agreeing only in the production of one primary root-cell. Spcrmothamnion Turneri forms at various points short unicellular rhizoids with terminal disks, branching does not occur, and cortications are not developed. Grifijithsia Bornetiana j^roduces a large spreading holdfast composed entirely of a pseudo-parenchymatous tissue arising from the primary root- cell. Callithamnion, Spyridia, and Ceramium have primary i890l DEVRI.OrMt.M' OF HOLDFASTS OF HOKIDE.F. 261 root-cells, from which spring rhizoids terminating in multi- cellular disks. Others originate in the cell adjacent to the basal cell and in the cortications. In addition the first two possess a strengthening mass of intrticuticular root-fibers, but Ceramium is quite destitute of them. Thus, whilf of some value in showing relationships, it will be seen that the chief interest in a comparative study of the devel- oping spores and holdfasts of the Floride:e would be in varia- tions dependent upon differences in light, temperature, or the density of the surrounding medium, and in adaptations to vege- tative reproduction. In closing, I would acknowledge my indebtedness to Dr. Setchell, who, in 1895, suggested the holdfasts of the Rhodo- phyceEe as a subject that would repay investigation ; to the late Dr. Humi)hrey, under whose helpful and suggestive direction the work described in this pai)er was practically begun; and to Professor Penhallow for kind advice. McGii.i University, Montrkai.. LITERATURE REFERRED TO. 1. BoRGE, O.— Ober die Rhizoidenbildung bei einigen fadenformigen Chlorophyceen. Upsala, 1894. 2. Brannon, M. a. —The structure and development of Grinnellia americana Harv. Annals of Botany 11 : 1. 1897. 3. Brebner, G.— On the origin of the filamentous thallus of Dumontia iiliformis. Jour. Linn. Soc. 30 : 436. 1895. 4. Davis, B. M.— Development of the frond of Champia parvula Harv. from the carpospore. Annals of Botany 6 : 339. 1892. 5. Gibson, R. J. H.— Notes on the histology of Polysiphonia fastigiata (Roth.) Grev. Jour, of Bot. 29 : 129. 1891. 6. Oltmanns, Fr. — Uber die Cultur- und I.ebensbedingungen der Meeresalgen. Jahrb. fiir wiss. Bot. 23 : 349- '892. 7. OSTERHOUT, W. J. v.— On the life-history of Rhabdonia tenera J. Ag. Annals of Botany 10 : 403. 1896. 8. StrOmfelt, H. F. G.— Untersuchungen iiber die Haftorgane der Algen. Bot. Centralbl. 33 :38i. 1888. 'r~ t-: 262 BOTANICAL GAZETTE [OCTOBER EXPLANATION OF PLATES XXI-XXIIL PLA TE XXI. Fig. I. Rliabdonia tenera J. Ag., carpospore 9 hrs. after sowing. X 334. Fig. 2. R. tenera, 2 days old. X 334. FiG.s. 3-1 1. A', tenera, from 4 to 17 days old. X 334. Fig. 12. Lomentaria uncinata Menegh., carpospore. X 334. Figs. 13-15. /.. ««(7'««/rt, 6 days old. X 334. Fig. 16. L. uncinata, an unusually large specimen, 6 days old. X 213. Figs. 17-19. Champia parvula Harv., plantlets resulting from the ger- mination of carpospores, 8 days old. X 334. F'lG. 20. C. parvula, holdfast of a plantlet growing on Polysiphonia. X 213. Figs. 21, 22. Lomentaria uncinata, bases of mature jjlants. Fig. 23. Cliondria tenuissima (Good, et Wood) C. Ag., tetraspore shortly after it was sown. X 233. Figs. 24-28. C. tenuissima, plantlets developed from carpospores, f'om 2 to 4 days old. X 233. Fig. 29. C. tenuissima, plantlet resulting from the germination of a tetra- spore, 10 days old. X 233. Fig. 30. C. tenuissima found growing on mature Chondria. X 233. " Figs. 31, 32. C. tenuissima, "; days old. X 233. Fig. 33. C. tenuissima, i2iiaysold. X 213. Fig. 34. C. tenuissima, 14 days old. X 213. Figs. 35, 36. C. tenuissima, 5 days old. X 213. Fig. 37. C. tenuissima, |)ortioii of mature holdfast, stained with methyl blue. X 213. Figs. 38-40. C. tenuissima, holdfasts of plants found growing on mature Chondria, stained with methyl blue. X 333. Figs. 41-56. Polysiphonia violacea (Roth) Greville. Figs. 41-45. Carpospore and ])lantlets from i to 5 days old. X 400. Fig. 46. Plantlet 7 days old. X 633. Figs. 47, 48. Plantlets found growing on Scytosiphon lomentarius Ag. X 400. Fig. 49. Plantlet, result of germinating carpospore. X 213. Figs. 50-56. Ends of primary rhizoids of plantlets 12 days old. X 400. "'."flJ^^5*^:ft?!}^S?^''' [OCTOBER f- X 334. X 213. the ger- siphonia. re shortly res, f'om if a tetra- 33- ■ ti methyl n mature 400. rius Ag. . X 400. iSOQI DEVELOPMENT OF HOLDFASTS OF FLORIDEA^ 263 PLA TE .VXD. Fic^. i-io. Polysiphonia violacea (Roth) Grev. Fig. I. Plantlet growing mi Scytosiphonlomentarius Ag. X 400. Fig. 2. Plantlet 3 days old. X 400. Fig. 3. Plantlet 9 days old. X 400. Figs. 4, 5. Pl-ntlets 12 days old. X 233. Fig. 6. Holdfast showing rhizoids springing from corticating cells. X 54- Fig. 7. Plantlet 12 days old. X 213. Fig. 8. Clasping disk of rhizoid of plant growing on Zostera marina L. X 400- . ■ . Fig.s. g, 10. Plantlets 9 and 11 days old. X 400. Figs. 11-29. Spyridiafilamentosa{\^\3M>)ViViX\. Y\G. II. Carpospore. X 400. Figs. 12-18. Plantlets from 2 to i 2 days old. X 400. Figs. 19-23. Bases of plantlets, showing primary cells and clasping disks. X 400. Fig. 24. Base of rhizoidal branch (t-. oi fig. 28). X 213. Fig. 25. Edge of mature holdfast, in optical section. X I57- Fig. 26. Transverse section of frond, showing that the origin of rhizoidal branches is in the corticating cells ; somewhat magnified. Fig. 27. Procumbent branch, near i8a, with rhizoidal outgrowths from central cell. X 333. Fig. 28. Creeping branch, slightly magnified. Fig. 29. Uncorticated secondary rhizoid with disk. X 333. Figs. 30-43. Dasya elegans {W-Axx^Vi^) C hg. Figs. 30, 31. Germinating tetraspore. X 400- Fig. 32. Segmenting carpospore 2 days old. X 400. Figs. 33-35. Segmenting tetraspore. X 400. Figs. 36-38. Plantlets 12 days old, resulting from the germination of tetraspores. X 400. Figs. 39-42. Plantlets 17 days old, resulting from the germination of carpospores; yfif.f.j9, ^.? show only the root-cell and disk. X 400. Fig. 43. Portion of mature holdfast, in optical section. X 233. PLA TE XXI ri. Figs. 1-9. Dasya elegans {y\.&xVtn^) C. Ag. Plantlets resulting from the germination of carpospores; in some cases only the primary root-cell and disk are shown; from 12 to 20 days old. X400. .*.\k. i ■ 264 BOTANICAL GAZETTE [OCTOBER I Figs. 10-23. Ceramium rubruni (Huds.) C. Ag. Fig. 10. Germinating carpospore 2 days after sowing. X 633. F'iGS. 1 1-17. Plantlets from 3 to 8 days old ; mfigs. /j, 16 only the basal cells and primary rhizoids are shown. X 400. Fig. 18. Plantlet found growing on Polysiphonia. X 400. Figs. ly, 20. Plantlets q days old. X 400. Fig. 21. Plantlet growing on Polysi[)honia. X 400. Fig. 22. Plantlet 5 days old, showing early branching of the primary rhizoid. X 400. Fig. 23. Holdfast of a rather young plant. X 213. Figs. 24-28. Ceramium strictiim (Harv). Fig. 24. Carpospore 36 hours after it was sown. X 400. Fig. 25. Plantlet 3 days old. X 233. Figs. 26, 27. Plantlets about 7 days old. X 400- Fig. 28. Primary root-cell and disk of mature plant. X 400. Figs. 29-32. Callithamnion Borreri Ag., plantlets from 2 to 6 days old. X 213. Fig. 33. C. Borreri, plantlet showing basal cell and rudimentary disk, 6 days old. X 213. Fig. 34. C /?^r/-(?r/, primary disk of plantlet 9 days old. X 333. Fig. 35. C. Baileyi Harv., mature holdfast, in optical section. X 213. Figs. 36, 37. (". Borreri, base of plantlets 6 days old, each showing pri- mary disk. X 400. Fig. 38. C. Borreri plantlet 6 days old. X 213. Fig. 39. C. Baileyi, a portion of a mature holdfast. X 100. Fig. 40. Spermothamnion Tumeri Aresch., branch with clasping disk. X 213. Fig. 41. .S". Ttirneri, plantlet 6 days old; nine cells of the filament are not represented. X 213. Figs. 42, 43. S. Tw^Tfifr/, mature holdfast. X 213. [OCTOBER ly the basal the primary 6 days old. itary disk, 6 53- X 213- howing pri- isping disk, filament are